Carbon Monoxide Poisoning

Citation preview

CARBON MONOXIDE POISONING

8417: “8417_c000” — 2007/9/25 — 20:28 — page i — #1

8417: “8417_c000” — 2007/9/25 — 20:28 — page ii — #2

CARBON MONOXIDE POISONING EDITED BY

DAVID G. PENNEY

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

8417: “8417_c000” — 2007/9/25 — 20:28 — page iii — #3

8417: “8417_c000” — 2007/9/25 — 20:28 — page iv — #4

Dedication ——————— I wish to dedicate this book first to my mother, Gertrude Ellen (Goodhew) Penney, always a source of support and encouragement, to my grandchildren, and to all of those victims of carbon monoxide poisoning who sought but did not find professional help for their suffering.

8417: “8417_c000” — 2007/9/25 — 20:28 — page v — #5

8417: “8417_c000” — 2007/9/25 — 20:28 — page vi — #6

Preface Carbon Monoxide Poisoning is a new title covering further areas of the expansive field of carbon monoxide (CO) toxicology that were not covered in the first two books, Carbon Monoxide and Carbon Monoxide Toxicity, both edited by David G. Penney, PhD. Both were published by CRC Press, the first in 1996 and the second in 2000. This book is designed to be complementary to both earlier books, forging into new areas and following new themes. The scope of this book is even broader than the earlier two. The first book took a very scholarly approach, presenting the latest basic and medical science of CO toxicology in 13 chapters. The contents of that book remain current. The second book was broader in its approach, and while extending presentations of basic and medical science, added discussions of human CO exposure under specialized conditions and in geographic locations other than the United States, in 23 chapters. It also presents a large body of new data on both acute and chronic CO poisoning. This present, third book, Carbon Monoxide Poisoning, further extends these presentations both to new areas such as the law, rehabilitation, personal experience with CO poisoning, education of the public about CO using the World Wide Web, and so forth, and adds further new data on chronic CO poisoning. This book contains some unique features: 1. A critical look at the efficacy of hyperbaric oxygen therapy in decreasing the damage caused by CO poisoning 2. The use of exciting new scanning techniques in revealing damage from CO poisoning 3. The introduction of a handheld pulse-oximeter that reads COHb directly and noninvasively 4. New data showing the persistent health damage that can be caused by chronic CO poisoning 5. The dangers of CO poisoning possible in motor homes, recreational boats, and so on 6. The levels of ignorance regarding CO on the part of the general public Interest in the effects of carbon monoxide on human health has grown rapidly during the past 20+ years. Governmental agencies, private groups, and the public are concerned. While an old and familiar poison, CO remains the number one “poison” in our environment in terms of its “brain-killing” potential, and its potential for overall immediate and long-term health harm. The public and the medical community need to obtain quality information about the risks from CO and need the means to identify and manage victims of CO poisoning successfully. It is hoped that this book, and its two previous companions, will in some way be of value in meeting these challenges.

8417: “8417_c000” — 2007/9/25 — 20:28 — page vii — #7

8417: “8417_c000” — 2007/9/25 — 20:28 — page viii — #8

Author David G. Penney, PhD, is a retired professor of physiology, who taught and conducted research on carbon monoxide at the School of Medicine, at Wayne State University, Detroit, Michigan. He was at one time adjunct professor of occupational and environmental health in the School of Allied Health Professions at Wayne State University. He is also a retired director of general surgical research at Providence Hospital in Southfield, Michigan, where for 12 years he directed the scholarly activities of surgical residents and attending surgeons. Dr. Penney obtained his BSc degree from Wayne State University in 1963, and his MSc and PhD degrees from the University of California, Los Angeles, in 1966 and 1969, respectively. Before coming to Wayne State University in 1977, he was a faculty member at the University of Illinois, Chicago. With his wife, Linda Mae Penney, the couple have six children. Dr. Penney’s professional interests have been focused on carbon monoxide for over 37 years, in both animal models and in humans. His special interests center around chronic CO poisoning, education of the public about the dangers of CO poisoning, the diagnosis and management of CO poisoning victims, and the medicolegal aspects of CO toxicology. Dr. Penney has assisted many national and international government and nongovernment agencies in matters involving carbon monoxide. He was among the earliest consultants to the US Environmental Protection Agency (EPA) in setting CO standards for outside air. He assisted the World Health Organization (WHO) in the late 1990s in setting similar standards for the world. He has worked with the Australian Medical Association (AMA) and with other concerned groups in Australia to attempt to stem the tide of suicides involving CO. Currently, Dr. Penney assists Underwriters Laboratory (UL) as a medical expert on CO in establishing standards for CO alarms and other gas-monitoring equipment, and major gas distributing companies in educating the public about the dangers of CO poisoning. Dr. Penney’s published works on CO include over 65 peer-reviewed research articles, several dozen other articles and abstracts, a number of review articles, book chapters, and three other books in print. At last count, Dr. Penney had more research articles and books published on the topic of CO toxicology than anyone else in the world. He has also published several other books on medical education and on Royal Oak history, and for some years wrote a column on local history for a hometown newspaper.

8417: “8417_c000” — 2007/9/25 — 20:28 — page ix — #9

8417: “8417_c000” — 2007/9/25 — 20:28 — page x — #10

Acknowledgments I wish to thank all the authors, former patients, and all who have contributed to this book. It has been a long road and at times it seemed impossible. Now it is done. Thanks to everyone. I also wish to thank Wayne State University School of Medicine and my Department of Physiology chairman, Dr. Joseph Dunbar, for granting me the time off in 2005 to get the book off the ground. I of course thank CRC – Taylor and Francis Publishers and all their employees who have been wonderful to work with these past 12 years, in developing my three books on carbon monoxide. Finally, I wish to thank my wife Linda for her constant support in developing these books, hearing my complaints, providing inspiration and also some perspiration in getting the work done. I of course thank my mother, Gertrude, for her support, encouragement and even late night help with data entry.

8417: “8417_c000” — 2007/9/25 — 20:28 — page xi — #11

8417: “8417_c000” — 2007/9/25 — 20:28 — page xii — #12

Contributors Rob Aiers

Kosmas Galatsis, Ph.D.

Envirotec (UK) Ltd. Hampshire, U.K.

Microelectronics Advanced Research Corporation Center on Functional Engineered Nano Architectonics University of California Los Angeles, California

Carol L. Armstrong, Ph.D., A.B.P.N. Division of Oncology The Children’s Hospital of Philadelphia Department of Neurology University of Pennsylvania Medical School Philadelphia, Pennsylvania

Steve N.M. Collard, B.A., M.E.D. Adjunct Professor of Law Nova Southeastern University Valrico, Florida

David C. Cone, M.D. Division of Emergency Medicine Yale New Haven Hospital Yale University School of Medicine New Haven, Connecticut

Joseph A. Cramer Wyoming, Michigan

James F. Georgis, O.D. Optometry Clinic Pueblo, Colorado

James M. Gracey, Ed.D. Colorado Institute for Injury Rehabilitation, Inc. Denver, Colorado

Neil B. Hampson, M.D. Center for Hyperbaric Medicine Section of Pulmonary and Critical Care Medicine Virginia Mason Medical Center Seattle, Washington

Jacqueline L. Cunningham, Ph.D. Department of Psychology The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania

Michael F. Hanzlick

Thomas M. Dydek, Ph.D., D.A.B.T., P.E.

Alastair W.M. Hay, Ph.D.

Dydek Toxicology Consulting Austin, Texas

Gas Dynamics Corporation St. Paul, Minnesota

Molecular Epidemiology Unit LIGHT Laboratories School of Medicine University of Leeds Leeds, U.K.

Peter G. Flachsbart, Ph.D., A.I.C.P.

Dennis A. Helffenstein, Ph.D.

Department of Urban and Regional Planning University of Hawaii at Manoa Honolulu, Hawaii

Colorado Neuropsychological Associates Colorado Springs, Colorado

Robert E. Engberg, B.S., P.E.

Hanzlick & Associates Highland Ranch, Colorado

8417: “8417_c000” — 2007/9/25 — 20:28 — page xiii — #13

Contributors

xiv

Gunnar Heuser, M.D., Ph.D.

Joshua A. Mott, Ph.D.

Neuromed and Neurotox Associates Santa Barbara, California Clinical Assistant Professor University of California, Los Angeles Los Angeles, California

CDC/CCID/NCIRD Air Pollution and Respiratory Health Branch National Center for Environmental Health Atlanta, Georgia

S. Gregory Hipskind, M.D., Ph.D. Department of Anthropology Western Washington University Bellingham, Washington

Ramona O. Hopkins, Ph.D. Psychology Department and Neuroscience Center Brigham Young University Provo, Utah

Gary Hutter, Ph.D. Meridian Engineering & Technology, Inc. Glenview, Illinois

David G. Penney, Ph.D. Wayne State University School of Medicine, and Providence Hospital and Medical Centers (retired) St. Augustine, Florida & Beulah, MI

Linda M. Penney St. Augustine, FL & Beulah, MI

Kevin J. Reilly, Jr.

Richard Karg, B.S., M.S.

Training Operations & Firefighter Diversified Security Solutions, Inc. Saddle Brook, New Jersey

R.J. Karg Associates Topsham, Maine

James W. Rhee, M.D.

Michael E. King, Ph.D. CDC/CCEHIP/NCEH Air Pollution and Respiratory Health Branch National Center for Environmental Health Atlanta, Georgia

Jerrold B. Leikin, M.D. Rush Medical College Chicago, Illinois and Evanston Northwestern Health Care Glenview, Illinois

Section of Emergency Medicine & Medical Toxicology The University of Chicago Chicago, Illinois

Frank Ricci New Haven City Fire Department New Haven, Connecticut

Carlos D. Scheinkestel, M.D. Monash University Melbourne, Australia

Jane Brown McCammon, B.S., M.S.

Robert E. Schreter, B.S., P.E.

Double Angel Foundation Broken Circle M Consulting, LLC Littleton, Colorado

R. Schreter and Associates, Inc. Roswell, Georgia

Peter Tikuisis, Ph.D. Ian L. Millar Monash University Melbourne, Australia

Human Modeling Group Defence Research & Development Canada Toronto, Ontario, Canada

8417: “8417_c000” — 2007/9/25 — 20:28 — page xiv — #14

Contributors

Christian Tomaszewski, M.S., M.D., F.A.C.E.P., F.A.C.M.T. Department of Emergency Medicine University of Pittsburgh Medical Center Hannad Medical Corporation Doha, Qatar

Suzanne R. White, M.D., F.A.C.M.T., F.A.C.E.P. Children’s Hospital of Michigan Regional Poison Control Center Department of Emergency Medicine Wayne State University School of Medicine Detroit, Michigan

xv

Stephen P. Willison, B.S., J.D. Willison & Hellman, P.C. Grand Rapids, Michigan

Wojtek B. Wlodarski, D.Sc., Ph.D., M.Sc. E.E. School of Electrical and Computer Systems Engineering Royal Melbourne Institute of Technology University Melbourne, Victoria, Australia

8417: “8417_c000” — 2007/9/25 — 20:28 — page xv — #15

8417: “8417_c000” — 2007/9/25 — 20:28 — page xvi — #16

Table of Contents Table of Contents for Carbon Monoxide Toxicity, 2000 . . . . . . . . . . . . . . . . . . . . . . . . .

xxi

Table of Contents for Carbon Monoxide, 1996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xxv

Chapter 1 Introduction to and Overview of the Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David G. Penney

1

Chapter 2 Exposure to Ambient and Microenvironmental Concentrations of Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peter G. Flachsbart Chapter 3 Carbon Monoxide Build-Up in Houses and Small Volume Enclosures. . . . . . . . . Robert E. Engberg Chapter 4 Formation and Movement of Carbon Monoxide into Mobile Homes, Recreational Vehicles, and Other Enclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert E. Schreter Chapter 5 Carbon Monoxide Emissions from Gas Ranges and the Development of a Field Protocol for Measuring CO Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard Karg Chapter 6 Investigating Carbon Monoxide-Related Accidents Involving Gas-Burning Appliances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Hanzlick Chapter 7 Carbon Monoxide Dangers in the Marine Environment. . . . . . . . . . . . . . . . . . . . . . . . . Jane McCammon

8417: “8417_c000” — 2007/9/25 — 20:28 — page xvii — #17

5

43

57

99

129

157

xviii

Contents

Chapter 8 Application of Warnings and Labels for Carbon Monoxide Protection . . . . . . . . . Gary Hutter Chapter 9 Public Health Surveillance for Carbon Monoxide in the United States: A Review of National Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael E. King and Joshua A. Mott Chapter 10 Carbon Monoxide Sensors and Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kosmas Galatsis and Wojtek Wlodarski Chapter 11 Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond: Use of the World Wide Web in Saving Lives . . . . . . . . . . . . . . . . . . . . . . . . . . Rob Aiers

197

233

251

271

Chapter 12 Investigating Carbon Monoxide Poisonings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thomas M. Dydek

287

Chapter 13 Carbon Monoxide Detectors as Preventive Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . James W. Rhee and Jerrold B. Leikin

305

Chapter 14 Misconceptions About Carbon Monoxide. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David G. Penney

313

Chapter 15 A Survey Study of Public Perceptions About Carbon Monoxide . . . . . . . . . . . . . . . David G. Penney and Linda M. Penney

325

Chapter 16 Treatment of Carbon Monoxide Poisoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Suzanne R. White

341

Chapter 17 The Case for the Use of Hyperbaric Oxygen Therapy in Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christian Tomaszewski

8417: “8417_c000” — 2007/9/25 — 20:28 — page xviii — #18

375

Contents

xix

Chapter 18 Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning: Useful Therapy or Unfulfilled Promise? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carlos D. Scheinkestel and Ian L. Millar Chapter 19 A Challenge to the Healthcare Community: The Diagnosis of Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David G. Penney Chapter 20 Neuroimaging after Carbon Monoxide Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gunnar Heuser

391

437

449

Chapter 21 Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning 457 S. Gregory Hipskind Chapter 22 Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning . . . . . . Ramona O. Hopkins Chapter 23 Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide Poisoning: A Retrospective Study and Case Presentation . . . . . . . . . . . . . . . . . . . . . . . Dennis A. Helffenstein Chapter 24 Chronic Carbon Monoxide Poisoning: A Case Series . . . . . . . . . . . . . . . . . . . . . . . . . . . David G. Penney Chapter 25 Functional and Developmental Effects of Carbon Monoxide Toxicity in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carol L. Armstrong and Jacqueline Cunningham Chapter 26 Issues in Rehabilitation and Life Care Planning for Patients with Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James M. Gracey Chapter 27 Treatment of Carbon Monoxide Poisoning with Yoked Prism Lenses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James F. Georgis

8417: “8417_c000” — 2007/9/25 — 20:28 — page xix — #19

477

495

551

569

591

619

xx

Contents

Chapter 28 Firefighters and Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kevin J. Reilly, Jr., Frank Ricci, and David Cone Chapter 29 The Purpose and the Process of Litigation in a Carbon Monoxide Poisoning Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stephen P. Willison

643

655

Chapter 30 Offering Expert Opinions in a Carbon Monoxide Case . . . . . . . . . . . . . . . . . . . . . . . . . Stephen P. Willison

671

Chapter 31 Injury Caused by Carbon Monoxide Poisoning: Defining Monetary Damages Steve Collard

683

Chapter 32 My Carbon Monoxide Poisoning: A Victim’s Story . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joseph A. Cramer

725

Chapter 33 Noninvasive Measurement of Blood Carboxyhemoglobin with Pulse CO-Oximetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neil B. Hampson Chapter 34 Chronic Carbon Monoxide Exposure: How Much Do We Know About it?—an Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alastair W.M. Hay

739

745

Chapter 35 Essential Reference Tables, Graphs, and Other Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . David G. Penney

753

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

765

8417: “8417_c000” — 2007/9/25 — 20:28 — page xx — #20

Table of Contents for Carbon Monoxide Toxicity 2000

Chapter 1 History of Carbon Monoxide Toxicology Dieter Pankow Chapter 2 Carbon Monoxide in Breath, Blood, and Other Tissues Hendrik J. Vreman, Ronald J. Wong, and David K Stevenson Chapter 3 Carbon Monoxide Detectors Richard Kwor Chapter 4 The Setting of Health-Based Standards for Ambient Carbon Monoxide and Their Impact on Atmospheric Levels James A. Raub Chapter 5 Effect of Carbon Monoxide on Work and Exercise Capacity in Humans Milan J. Hazucha Chapter 6 The Interacting Effects of Altitude and Carbon Monoxide James J. McGrath Chapter 7 Interactions Among Carbon Monoxide, Hydrogen Cyanide, Low Oxygen Hypoxia, Carbon Dioxide, and Inhaled Irritant Gases David A. Purser Chapter 8 Carbon Monoxide Poisoning and Its Management in the United States Neil B. Hampson

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxi — #21

Contents

xxii

Chapter 9 Death by Suicide Involving Carbon Monoxide around the World Pierre Baume and Michaela Skopek Chapter 10 Carbon Monoxide as an Unrecognized Cause of Neurasthenia: A History Albert Donnay Chapter 11 Update on the Clinical Treatment of Carbon Monoxide Poisoning Suzanne R. White Chapter 12 Treatment of Carbon Monoxide Poisoning in France Monique Mathieu-Nolf and Daniel Mathieu Chapter 13 Acute Carbon Monoxide Poisonings in Poland - Research and Clinical Experience Jerzy A. Sokal and Janusz Pach Chapter 14 Treatment of Carbon Monoxide Poisoning in the United Kingdom Martin R. Hamilton-Farrell and John Henry Chapter 15 Carbon Monoxide Air Pollution and Its Health Impact on the Major Cities of China Qing Chen and Lihua Wang Chapter 16 Use of Scanning Techniques in the Diagnosis of Damage from Carbon Monoxide I.S. Saing Choi Chapter 17 Low-Level Carbon Monoxide and Human Health Robert D. Morris Chapter 18 Chronic Carbon Monoxide Poisoning David G. Penney Chapter 19 Chronic Carbon Monoxide Exposure: The CO Support Study Alistair WM. Hay, Susan Jaffer, and Debbie Davis

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxii — #22

Contents

Chapter 20 Neuropsychological Evaluation of the Carbon Monoxide-Poisoned Patient Dennis A. Helffenstein Chapter 21 Pediatric Carbon Monoxide Poisoning Suzanne R. White Chapter 22 Carbon Monoxide Production, Transport, and Hazard in Building Fires Frederick W. Mowrer and Vincent Brannigan Chapter 23 Approaches to Dealing with Carbon Monoxide in the Living Environment Thomas H. Greiner and Charles V. Schwab

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxiii — #23

xxiii

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxiv — #24

Table of Contents for Carbon Monoxide 1996

Chapter 1 Carbon Monoxide Analysis Roger L. Wabeke Chapter 2 Carbon Monoxide Formation Due to Metabolism of Xenobiotics Dieter Pankow Chapter 3 Modeling the Uptake and Elimination of Carbon Monoxide Peter Tikuisis Chapter 4 Cerebrovascular Effects of Carbon Monoxide Mark A. Helfaer and Richard J. Traystman Chapter 5 Pulmonary Changes Induced by the Administration of Carbon Monoxide and Other Compounds in Smoke Daniel L. Traber and Darien W Bradford Chapter 6 Effects of Carbon Monoxide Exposure on Developing Animals and Humans David G. Penney Chapter 7 Carbon Monoxide - From Tool to Neurotransmitter Nanduri R. Prabhakar and Robert S. Fitzgerald Chapter 8 Toxicity of Carbon Monoxide: Hemoglobin vs. Histotoxic Mechanisms Claude A. Piantadosi

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxv — #25

Contents

xxvi

Chapter 9 Carbon Monoxide-Induced Impairment of Learning, Memory, and Neuronal Dysfunction Masayuki Hiramatsu, Tsutomu Kameyama, and Toshitaka Nabeshima Chapter 10 Behavioral Effects of Carbon Monoxide Exposure: Results and Mechanisms Vernon A. Benignus Chapter 11 Delayed Sequelae in Carbon Monoxide Poisoning and the Possible Mechanisms Eric Kindwall Chapter 12 Treatment of Carbon Monoxide Poisoning Suzanne R. White Chapter 13 Options for Treatment of Carbon Monoxide Poisoning, Including Hyperbaric Oxygen Therapy Stephen R. Thom

8417: “8417_c000” — 2007/9/25 — 20:28 — page xxvi — #26

1

Introduction to and Overview of the Field David G. Penney

CONTENTS References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4

I have designed this book to complete the series on carbon monoxide (CO) begun with Carbon Monoxide, 19961 and continued with Carbon Monoxide Toxicity, 2000.2 This and the second book are NOT new editions of the first book, as has often been assumed. While CO may seem a very narrow subject area, it finds its way into many diverse disciplines and its literature is vast. This third book, Carbon Monoxide Poisoning, completes the trilogy and should become a standard reference source on CO for years to come. The new book covers areas not previously presented, including rehabilitation, education of the public using the WWW, litigation involving CO poisoning, economic loss assessment, and firefighting. There are areas of update, such as the chapter by Dr. Suzanne White, on diagnosis and management. There are two chapters in which the authors take opposing views, one stating the case for use of hyperbaric oxygen therapy (HBOT) by Dr. Christian Tomaszewski and against the use of HBOT by Dr. Carlos Scheinkestel. One chapter deals with toxicology investigation (i.e., forensic) procedures. A series of chapters detail the risk of CO poisoning from kitchen ranges, recreational trailers and motor homes, and recreational powerboats. Three chapters cover the very important area of neuropsychological evaluation of adults and children following CO poisoning. The chapter by Dr. Dennis Helffenstein presents new data on a case series of patients that had sustained chronic CO poisoning. I have written a companion chapter to his in which a retrospective review of 61 chronically CO-poisoned patients were symptomatically evaluated (Chapter 24). Better Education of Physicians: This is essential if CO-poisoned patients are to be properly diagnosed and treated in the future. Several years ago I overheard a prominent emergency room physician say, “the standard of care for CO poisoning in the U.S. is less than the standard of care.” I estimate that in 80% of CO cases reviewed, one, two, or more mistakes were made in diagnosing and/or treating CO-poisoned patients. This may involve misdiagnosis, dependence on faulty pulse-oximetry data, administering NBO with the wrong 1

8417: “8417_c001” — 2007/9/11 — 12:17 — page 1 — #1

2

Carbon Monoxide Poisoning

equipment (e.g., nasal prongs), failure to order HBOT when it was needed and possible, discharge of patients while still symptomatic, and so forth. Physicians should also be informed of the possible serious permanent health harm that chronic or lower-level acute CO poisoning can cause if not diagnosed immediately and treated fully. From my perspective, the CO cases that result from acute poisoning and those most likely to reach the media are actually the smaller fraction of all CO poisonings, while the chronic (i.e., occult) CO poisonings make up by far the largest fraction, and probably result in the most injuries, but they are the very group that physicians are least trained to properly deal with. Better Education of the Public: The public too needs education about the dangers presented by CO exposure. My chapter presenting the results of surveys of public perceptions of CO in Michigan and Florida shows this. While almost everyone knows that CO is a deadly poison, substantial fractions (sometimes most) of the adult and juvenile population cannot intelligently evaluate the risk of CO from automobiles, propane radiant heaters, generators, and recreational powerboats. In some situations people are overly cautious in a given situation, but in other situations people vastly underestimate the risk of injury and death. Youth, as opposed to adults, are particularly uninformed. New Approaches to Treating Acute, Severe CO Poisoning: There appears to be very little new in treating acute severe CO poisoning. We cannot decide for sure whether HBOT is more effective in reducing neurologic sequelae, even though it has been used for approximately 50 years. The pros and cons chapters on HBOT provide detailed discussions of many aspects of the situation, and a few ideas about possible new approaches. The bright light in this area is almost certainly the new generation of pulse-ox devices that read COHb directly and noninvasively. See Dr. Neil Hampson’s chapter about the testing of this device. Requiring Proper Warnings on Combustion Equipment: The lack of proper and adequate warnings on equipment that do, or under foreseeable conditions might, emit harmful or lethal amounts of CO remains a real problem. People continue to die because warnings on combustion devices are not obvious, explicit and direct, and not on the device itself. Warnings in operating manuals should be continued and improved, but they alone are insufficient because manuals are usually separated from the device. Warnings on the device must tell the user what might occur in using the device a certain way. Statements such as “provide adequate ventilation” are useless. The warning must specify where and when not to use the device, and for how long. Warnings in the United States should be written in both English and Spanish, along with standard prohibition symbols. Some of the devices this applies to include portable generators, cement saws, lawn mowers, pressure washers, scissor-lifts, kerosene heaters, propane-radiant heaters, lamps and cook stoves, powerboats, charcoal grills and hibachis, and so forth. See Dr. Hutter’s excellent chapter on warnings.

8417: “8417_c001” — 2007/9/11 — 12:17 — page 2 — #2

Introduction to and Overview of the Field

Rethinking Work Guidelines for Carbon Monoxide that Reflect the Science: Another area that needs immediate attention is threshold limit standards for inhalation of CO. Environmental Protection Agency (EPA) and World Health Organization (WHO) after extensive study and deliberation some years ago set the 8-h standard at 9 ppm, or 10 mg/m3 for outside air. On the other hand, National Institute for Occupational Safety and Health (NIOSH) and Occupational Safety and Health Administration (OSHA) have set different standards for the work environment. We know that people with coronary artery disease, congestive heart failure, asthma, and a state of fetal development are often members of the workforce, and represent a more sensitive, higher risk subgroup of the general population. It is also well known that people, even those with no obvious risk factors, vary widely in their tolerance of CO. Why then should the standards be so different—9 ppm (EPA, WHO) versus 50 ppm (OSHA) for the same species? This is a 5-1/2 fold difference! I believe it is time we in the toxicology community re-examine ambient air CO concentration work standards, and make decisions for new standards based only on the best science. Realization that Brain Damage Resulting from CO Poisoning Is not Dependent on COHb and/or Severity of Poisoning: New studies make it clear what many of us have believed for some time based on experience, that brain damage from CO poisoning is only very poorly correlated with the severity of the poisoning by whatever criteria, even the COHb saturation. Some people with very severe poisoning and/or high initial COHb values make remarkable recoveries, while some with what appears to be minimal poisoning, and even on occasion, near normal COHb when measured, incur substantial damage. Clearly, loss of consciousness is not required for the development of neurologic sequelae, although I sometimes hear the less well-informed say that it is. I believe everyone agrees that a longer (i.e., soaking) exposure is more detrimental than a short one, but strangely some still insist that chronic, lower-level CO exposure can cause no permanent harm. Other markers for brain damage have been proposed such as acidosis, gait/balance/clumsiness on presentation, and release of cellular enzymes, but it remains unclear how useful they are. Carbon Monoxide Disaster Management: Release of CO during certain kinds of disasters could pose significant problems. Fire almost invariably gives off CO as an incomplete combustion product, and fire is a larger or a smaller component in most disasters, especially those that might be instigated by terrorists. It is unclear to me whether any overall planning has been done by agencies of the government with respect to CO. Firefighters regularly encounter CO in the work they do, and have equipment such as self-contained breathing apparatus to deal with it. See the chapter on firefighting by Mr. Reilly, Mr. Ricci, and Dr. Cone. Other concerns for human health that may not yet be fully addressed include indoor car, monster truck, and motocross events, work in coal mines, warehouses, enclosed construction sites, and so forth where significant CO is generated by combustion devices. Since it is dangerous and

8417: “8417_c001” — 2007/9/11 — 12:17 — page 3 — #3

3

Carbon Monoxide Poisoning

4

foolhardy to run a portable generator inside a garage or house, why is it not dangerous to human health to run several monster trucks with huge gasolinefueled engines lacking catalytic converters inside a covered sports stadium? The recent Sago Mine disaster where a dozen men died slowly, mainly from CO poisoning, points out the need for adequate emergency equipment that would allow men to live for extended periods of time in the presence of lethal CO air concentrations. Autostarters—are they safe? How often will car engines be started inadvertently (by children, otherwise by mistake) in a closed garage, leading to injury or death from CO poisoning? The wonder is that CO has been with man since prehistory, probably since we first began using fire. Other scourges such as plague, cholera, typhus, smallpox, and so forth are gone, at least from the developed world, whereas this simple, small molecule, CO, continues to afflict us, and probably will, at least as long as we are wedded to the “carbon energy cycle.” I hope you enjoy this book and will use it with its earlier brothers, Carbon Monoxide (1996)1 and Carbon Monoxide Toxicity (2000).2

References 1. Penney, D.G., ed. Carbon Monoxide, CRC Press, NY, 1996, 296 pp. 2. Penney, D.G., ed. Carbon Monoxide Toxicity, CRC Press, NY, 2000, 560 pp.

8417: “8417_c001” — 2007/9/11 — 12:17 — page 4 — #4

2

Exposure to Ambient and Microenvironmental Concentrations of Carbon Monoxide Peter G. Flachsbart

CONTENTS 2.1 2.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standards and Guidelines for Exposure to Ambient Concentrations of Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Trends in Carbon Monoxide Emissions and Ambient Air Quality . . . . . . . 2.4 Human Exposure to Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Microenvironmental Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1 Residential Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1.1 Nonfatal Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1.2 Fatal Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.2 Occupational Exposures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.3 Shopping Center Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.4 Recreational Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.4.1 Exposures on Recreational Vehicles . . . . . . . . . . . . . . . . . . . . 2.5.4.2 Exposures at Indoor Sporting Events . . . . . . . . . . . . . . . . . . . 2.5.5 Commuter Exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.1 Defective Exhaust Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.2 Parking Garages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.3 Service Stations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.4 Drive-Up Facilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.5 Airbag Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5.6 Motor Vehicle Emission Standards. . . . . . . . . . . . . . . . . . . . . . 2.6 Exposure to Methylene Chloride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1 Nonoccupational Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.2 Occupational Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 7 8 10 17 17 18 19 20 22 23 23 24 25 26 29 29 29 30 30 31 31 32 32 34 34 5

8417: “8417_c002” — 2007/9/11 — 12:16 — page 5 — #1

Carbon Monoxide Poisoning

6

2.1 INTRODUCTION Carbon monoxide (CO) is a gas commonly produced by incomplete combustion of fuels containing carbon atoms. Many people use these fuels (i.e., coal, gasoline, kerosene, natural gas, oil, propane, and wood) around the globe. As a result, CO is ubiquitous in the atmosphere. However, without sophisticated instruments, a person is unable to detect CO, because the gas is not irritating and has no color, odor, or taste. Moreover, the gas is a potential health hazard, because exposure to CO can starve critical body organs, especially the brain and heart, of oxygen. Once inside the lungs, CO molecules pass easily into the bloodstream and compete with oxygen for hemoglobin (Hb) in the red blood cells. About 95% of the absorbed CO readily binds with Hb to form carboxyhemoglobin (COHb), because the affinity of Hb for CO is over 200 times stronger than it is for oxygen. Thus, the percentage of total Hb in the blood that is in the form of COHb is a biomarker of CO exposure.1 The health effects of CO, which are a function of its concentration and the duration of exposure, range from subtle to severe. They include neurobehavioral, cardiovascular, and developmental effects, observed at low levels of CO exposure, to unconsciousness and death, which occur after acute exposure to high CO concentrations. Lethal CO exposures are usually linked to CO concentrations greater than 1000 parts per million (ppm) by volume. Coma, convulsions, cardiopulmonary arrest, and death have been observed when COHb levels reach 50%, although death from CO poisoning is frequently reported at far lower COHb saturations. The exact COHb concentrations that trigger acute and chronic health effects in different people differ widely. Sublethal levels of CO may cause neurological-type symptoms, including fatigue, headache, nausea, vomiting, deficit in short-term memory, to name a few. Exposures to these CO concentrations are often misdiagnosed as viral illness, clinical depression, and so forth. Still lower CO exposures (i.e., those producing less than 10% COHb) may not be associated under certain circumstances with overt symptoms.2,3 This broad range of effects makes CO relevant to people concerned with ambient air quality management as well as officials responsible for protecting public health and safety. Ambient air quality standards are the foundation of air quality management programs in many countries worldwide. Such standards typically specify maximum permissible concentrations in ambient air for certain pollutants. To achieve ambient standards in the United States, the U.S. Environmental Protection Agency (EPA) implemented progressively tighter tailpipe emission standards for motor vehicles. As a result, these standards have substantially reduced ambient CO concentrations in most metropolitan areas of the U.S. and have had other collateral benefits. For example, the nation had 11,667 fewer deaths from accidental CO poisoning between 1968 and 1998, according to a study by the Centers for Disease Control and Prevention (CDCP).4 Still, an average of 480 U.S. residents died each year during 2001–2002 from nonfire-related unintentional CO poisoning. In addition, an estimated 15,200 persons, that is, people with confirmed or possible nonfire-related CO exposure or poisoning, were treated annually in U.S. hospital emergency rooms.5 In fact, more than 50% of all fatal poisonings reported in many countries may be attributable to CO, because these cases are under-reported or misdiagnosed by medical professionals.2

8417: “8417_c002” — 2007/9/11 — 12:16 — page 6 — #2

Exposure to Ambient and Microenvironmental Concentrations

7

This chapter explores reasons behind the paradox of declining ambient CO concentrations in urban areas of the United States coupled with persistent injuries and fatalities from CO poisoning. The chapter is organized around the superposition principle of CO exposure, which may help to explain this paradox. This principle holds that CO concentrations at any given point in time and space consist of both ambient and microenvironmental components. The next section takes a closer look at ambient CO concentrations in urban areas. The chapter then describes how the development of portable monitors enabled measurements of personal exposure to CO concentrations in places where people perform routine daily activities. Since these activities often occur in specific microenvironments, the chapter then describes typical CO exposures where people live, work, shop, play and commute, and factors that affect these exposures. The last section offers some concluding thoughts.

2.2 STANDARDS AND GUIDELINES FOR EXPOSURE TO AMBIENT CONCENTRATIONS OF CARBON MONOXIDE The Clean Air Act (CAA) of 1963 was amended by the U.S. Congress in 1970, 1977, and 1990. The 1977 and 1990 versions largely reaffirmed the course set by the 1970 amendments.6 Under the 1970 CAA amendments, the U.S. EPA established the National Ambient Air Quality Standards (NAAQS) and set deadlines for their attainment. The current NAAQS reflect EPA’s scientific judgments about maximum allowable ambient concentrations and averaging times for certain “criteria” air pollutants including CO. Criteria air pollutants are those that could reasonably endanger public health or welfare. The 1970 and 1990 CAA amendments also mandated stringent motor vehicle emission standards as a means to achieve the NAAQS for CO and other air pollutants that have been linked to mobile sources. Air pollutant concentrations can be expressed either as ppm or as milligrams per cubic meter (mg/m3 ) of air. Many of the studies reviewed in this chapter refer to the NAAQS for guidance on allowable limits of CO exposure. The EPA promulgated identical primary and secondary NAAQS for CO on April 30, 1971. The primary standards specify a level of air quality sufficient to protect public health and the secondary standards are intended to protect public welfare. The standards include “an adequate margin of safety” to reflect scientific uncertainties related to measurement of the effects of air pollutant exposure in the population. In 1985, EPA rescinded the secondary standard for CO, but retained two primary standards: 9 ppm (10 mg/m3 ) as an 8-h average and 35 ppm (40 mg/m3 ) as a 1-h average. Each standard may be exceeded once per year in an air quality control region (AQCR) without violating the standard.7 The NAAQS for CO are designed to keep COHb levels below 2% in the blood of 99.9% of nonsmoking healthy adults and people who belong to probable high-risk groups. Smokers are excluded because they may exhale more CO into the air than they are inhaling from the ambient environment. The high-risk groups include the elderly; pregnant women; fetuses; young infants; and those suffering from anemia or certain other blood, cardiovascular, or respiratory diseases. People at greatest risk

8417: “8417_c002” — 2007/9/11 — 12:16 — page 7 — #3

Carbon Monoxide Poisoning

8

from exposures to ambient CO levels are those with coronary artery disease. Some of these people suffer myocardial ischemia as identified by ST-segment depression, during exercise when their COHb levels ≥ 2.4%.7 The symptoms of this disease are spasmodic attacks of chest pain (angina pectoris) caused by insufficient oxygen in the heart muscles. Controlled laboratory studies are needed to observe these health effects, because the COHb levels are at or near the lower margin of detection of current instruments.8 Although annual death rates from heart disease have been declining since 1980, heart disease is still America’s leading cause of death.9 Coronary artery disease reduces a person’s circulatory capacity, which is particularly critical during exercise when muscles need more oxygen. Given the widespread prevalence and lack of awareness of coronary heart disease, Godish3 argues that a significant number of people still may be at risk from CO exposure, even if ambient CO concentrations do not exceed the 8-h NAAQS for CO. The World Health Organization (WHO) guidelines for CO (see below) are also relevant to this discussion. Relative to the NAAQS for CO in the U.S., these guidelines have an identical 8-h concentration but a lower 1-h concentration. Unlike the NAAQS, the guidelines specify maximum concentrations for two shorter time spans (30 min and 15 min).10,11 Maximum Concentrations

Averaging Times

9 ppm (10 mg/m3 ) 25 ppm (30 mg/m3 ) 50 ppm (60 mg/m3 ) 90 ppm (100 mg/m3 )

8h 1h 30 min 15 min

WHO’s guidelines are intended to prevent blood levels of COHb from exceeding 2.5–3% in nonsmoking populations even when a person engages in relatively heavy work. Romieu12 reported that average COHb levels are about 1.2–1.5% in the general population and from 3% to 4% in the blood of cigarette smokers.

2.3 TRENDS IN CARBON MONOXIDE EMISSIONS AND AMBIENT AIR QUALITY The CAA amendments have substantially reduced nationwide CO emissions, even as other socio-economic indicators of growth have increased. For example, between 1970 and 2002, nationwide emissions of CO fell 48%, despite national increases of 38% in population, 155% in vehicle miles of travel (VMT), and 164% in gross domestic product.13 The rapid growth of VMT has been attributed to the decentralization of jobs and housing within urban regions during the post World War II era.14 The CAA amendments have also reduced ambient CO concentrations in urban areas of the United States. The U.S. EPA determines compliance with the NAAQS based on measurements of ambient air quality made by a nationwide network of fixed-site monitoring (FSM) stations. Ambient concentrations of air pollutants are

8417: “8417_c002” — 2007/9/11 — 12:16 — page 8 — #4

Exposure to Ambient and Microenvironmental Concentrations

9

typically measured in air “external to buildings, to which the general public has access.”15 In urban areas, most ambient stations that measure CO concentrations are located near roadways.16 These stations use nondispersive infrared reference (NDIR) spectrometry to measure ambient CO concentrations. Monitoring instruments based on the NDIR method are large, complex and expensive, and require a vibration-free, air-conditioned facility for the production of accurate and reliable data. The nationwide network consists of state and local air monitoring stations (SLAMS), which send data to EPA’s Aerometric Information Retrieval System (now Air Quality System) within six months of acquisition.17 Several stations within the SLAMS network belong to a network of national air monitoring stations (NAMS) to enable national assessments of air quality. FSM stations typically reveal two peaks in ambient CO concentrations. These peaks usually coincide with periods of congested rush-hour traffic.7 For that reason, some exposure analysts consider CO to be a signature air pollutant for mobile sources. On a nationwide basis, the EPA’s annual emissions inventory revealed that highway vehicles accounted for 62.8% of the 93.7 million tons of CO emitted from all sources except fires in 2003.13 Regional inventories show that motor vehicles account for even higher percentages of all CO emissions released into the ambient air. For example, CO emissions from motor vehicles ranged from 78% of all CO emissions in Fairbanks, Alaska, to 96% of all CO emissions in Phoenix, Arizona. EPA classified both cities as having “serious” levels of ambient CO concentrations in 1999.18 An ambient station is considered to be in violation (i.e., nonattainment) of the NAAQS for CO, if it records a nonoverlapping average concentration that exceeds either the 1-h or 8-h standard more than once per calendar year. The historical record shows that cities have had more difficulty satisfying the 8-h standard than the 1-h standard. For the 8-h standard, the nonoverlapping average omits other high values that occur within 8 h of the first value. Also, values of 9.5 ppm, or greater, are counted as exceeding the 8-h standard due to the standard’s rounding convention. Maximum 8-h average CO concentrations typically exceeded 30 ppm when continuous monitors were first installed in some U.S. cities in the early 1960s. When EPA promulgated the NAAQS for CO in 1971, 91.4% of 58 ambient monitors recorded violations of the 8-h standard and 12.1% of 58 stations recorded violations of the 1-h standard.16 In 1996, EPA’s Office of Air Quality Planning and Standards (OAQPS) reported that CO levels exceeded the NAAQS in seven counties, which had a combined population of more than 12.7 million people.19 The CAA amendments require states to develop plans to achieve and maintain ambient air quality that satisfies the NAAQS. To prepare these plans, states inventory emissions in each AQCR for a baseline year and determine the necessary emission reductions to achieve the NAAQS. Pursuant to the 1977 CAA amendments, many states established inspection and maintenance (I/M) programs as required by their plans for those regions that were in nonattainment of the NAAQS. An AQCR must satisfy the CO NAAQS for two consecutive years to be considered in attainment by EPA. States must submit plans to EPA showing how the region will maintain that attainment for at least 10 years. As of March, 2006, OAQPS reported that there were 38 CO “maintenance areas” in the United States encompassing 89 counties with a combined population of 46.8 million people.20

8417: “8417_c002” — 2007/9/11 — 12:16 — page 9 — #5

10

Carbon Monoxide Poisoning

Since 1995, 11 cities have reported violations of the 8-h standard and no monitor has reported a violation of the 1-h standard.16 Violations appear to persist in areas with meteorological and/or topographical handicaps. Meteorological handicaps make it particularly difficult for cities to satisfy the standard, because of the stochastic nature of ambient air pollutant concentrations. For example, violations of the NAAQS for CO in Fairbanks, Alaska, have been attributed to stagnant air masses during winter months. The atmosphere is more stable in Fairbanks during winter, because less sunlight causes ground-level temperature inversions to occur more frequently. Also, less air pollutant dispersion occurs in winter, because winds are milder and mountains surrounding the city hinder horizontal dispersion.16 As of March 2006, the EPA had classified five American cities (El Paso, Texas; Las Vegas and Reno, Nevada; the Los Angeles South Coast Air Basin, California; and Missoula, Montana) as urban areas that were in nonattainment of the NAAQS for CO. They represented eight counties with a total population of about 15.4 million people.21 Compared to 1996, there was one more county in nonattainment of the NAAQS for CO by 2006, and the total population living in such areas had increased by 21.3%.

2.4 HUMAN EXPOSURE TO CARBON MONOXIDE The study of population exposure is multidisciplinary and the definition of personal exposure has evolved over time. A recent definition states that exposure is the contact between an agent and a target at a specified contact boundary, defined as a surface in space containing at least one exposure point, that is, a point at which contact occurs. According to this definition, an inhaled CO molecule (the agent) reaches a human (the target) at the lining of the lung (the contact boundary), where CO exchange takes place between air and blood.22 Actual studies of CO exposure use small-scale portable monitors to measure CO concentrations within a few feet of a person’s nasal and oral cavities. These studies assume that the air surrounding the person is well mixed and that measured CO concentrations in that air represent the person’s actual exposure from CO inhalation. Besides inhalation exposure to CO, metabolic degradation of many drugs, solvents (e.g., methylene chloride), and other compounds of CO can elevate levels of COHb in a person’s blood. Because the endogenous production of CO from drugs and solvents may continue for several hours, it can prolong any cardiovascular stress from COHb. Moreover, the maximum COHb level from endogenous CO production can last up to twice as long as COHb levels caused by comparable exposures to exogenous CO.23,24 Hence, this chapter also discusses the literature on exposure to methylene chloride. Figure 2.1 illustrates an individual’s air pollutant exposure over time. In this figure, the function Ci (t) describes the CO concentration to which an individual i is exposed at any point in time t. Ott defined this event as the instantaneous exposure of an individual.25 The shaded area under the curve represents the accumulation of instantaneous exposures over some period of time (t1 − t0 ). This area also is equal to the integral of the air pollutant concentration function, Ci (t), between t0 and t1 . Ott defined the quantity represented by this area as the integrated exposure. An exposure analyst can derive the average concentration to which a person is exposed

8417: “8417_c002” — 2007/9/11 — 12:16 — page 10 — #6

Exposure to Ambient and Microenvironmental Concentrations

11

Concentration

Ci (t )

t0

t1 Time

FIGURE 2.1 Exposure of person i to air pollutant concentration (C) as a function of time t.

by dividing the person’s integrated exposure by the period of integration (t1 − t0 ). This average concentration is sometimes referred to as the average exposure. To compare the average concentration with an established air quality standard, the period of integration should equal the averaging period of the standard. This concept of exposure thus combines two parameters, the air pollutant concentration and the time duration of exposure to the air pollutant. An exposure analyst assumes that these two parameters are directly proportional to the dosage of CO in the body, as represented by the level of COHb in the blood stream, and ultimately to health outcomes. Prior to the use of portable monitoring devices, exposure analysts relied on ambient data from FSM stations in urban areas to estimate population exposure. Such data were used to provide crude estimates of population exposure to CO concentrations that violated the NAAQS. For example, the President’s Council on Environmental Quality (CEQ) did a crude estimate of population exposure to CO in the late 1970s.26 This method is based on data collected for each county in the United States. The estimate is derived as follows: TPE =

n


(pi )(di )

(2.1)

i=1

where TPE = the total population exposure in the United States (person-days) = the resident population of county i (persons) pi di = the number of days in a calendar year that violations of the NAAQS for CO are observed in county i (days) n = the total number of counties in the United States in a given year

8417: “8417_c002” — 2007/9/11 — 12:16 — page 11 — #7

12

Carbon Monoxide Poisoning

For 1978, the CEQ estimated that the nation’s total population exposure above the 1-h NAAQS for CO of 35 ppm was 2.80 billion person-days. This represented about 3.7% of the total possible exposures of 75.555 billion person-days, which was derived by multiplying an estimated 207 million U.S. residents in 1978 times 365 days in a year.26 CEQ’s method was a crude estimate of total population exposure, because the estimate rested on three major assumptions: 1. The CO concentrations measured by the county’s fixed-site monitors were representative of concentrations to which residents were actually exposed. If a county had more than one fixed-site monitor, then the monitor with the worst CO concentration represented the exposure of all residents in the county. 2. Residents did not leave the county on days that violations of the NAAQS for CO occurred for that county. 3. There were no violations of the NAAQS for CO in counties that were not monitored (e.g., rural counties). Several scientific studies have questioned the validity of the first two assumptions as discussed below. The first assumption implies that ambient CO concentrations are spatially homogeneous throughout a county. A study in the early 1970s questioned the ability of FSM stations to accurately represent human exposure to ambient CO concentrations. Using large Tedlar™ bags filled by a constant flow pump over 5-min periods, Ott collected 1128 CO concentrations at “breathing height” at outdoor locations in San Jose, California, on weekdays between October 1970 and March 1971.27 Of 438 samples collected on 21 dates while walking along sidewalks of congested downtown streets, 60% were above values measured concurrently at the nearest FSM station. The correlation between the “walking samples” and FSM values was positive, but low (r = 0.20). On 2 of 7 days, the sidewalk concentrations (13 and 14.2 ppm) averaged over an 8-h period were well above the corresponding concentrations (4.4 and 6.2 ppm, respectively) reported for the FSM station. Overall, the 8-h average CO concentrations for the 7 days ranged between 1.4 and 3 times the values observed simultaneously at the FSM station. The highest values were in late December when streets were heavily congested with traffic due to Christmas shopping.28 The San Jose study confirmed the hypothesis that FSM stations did not represent the CO exposures of a person walking in outdoor settings of a major city. Since the San Jose study measured the CO exposure of only one person, other studies tested the hypothesis for larger groups of people. Some of these studies focused on commuters, because CO is a signature air pollutant of motor vehicle tailpipe emissions. For example, Cortese and Spengler29 recruited 66 nonsmoking volunteers who lived in different parts of the metropolitan area of Boston, Massachusetts. Each volunteer carried an Ecolyzer monitor attached to a Simpson recorder for 3–5 days between October 1974 and February 1975. The study reported that the mean of all commuter exposures (11.9 ppm) was about twice the concurrent concentration measured at six FSM stations (6 ppm). Automobile commuters had exposures nearly twice that of

8417: “8417_c002” — 2007/9/11 — 12:16 — page 12 — #8

Exposure to Ambient and Microenvironmental Concentrations

13

transit users, and about 1.6 times that of people who did “split-mode” commuting, which involved both auto and transit. The first assumption of the CEQ method of exposure estimation also raises the following question: To what extent do ambient CO levels reflect the CO concentrations to which people are exposed indoors? Early studies of human activity patterns in America30 and other industrialized countries31 have consistently shown that people spend most of their time indoors. Several studies published in the early 1970s addressed this question.32−34 Each study examined the relationship between indoor and outdoor CO concentrations. In the absence of indoor sources, these studies found that indoor CO concentrations of office buildings tended to follow outdoor CO levels with some degree of time lag and with a tendency not to reach either the extreme high or low values that were found outdoors. The General Electric study also reported that CO concentrations were larger indoors than outdoors at heights greater than 100 ft above the roadway, due to entrapment of CO within the building.33 To simplify the exposure calculation, the CEQ assumed that each person spent 24 h at home, because the method relied on household data provided by the U.S. Census Bureau. The CEQ method also assumed that people did not travel outside the area represented by the FSM station. People who live in cities spend a significant amount of personal time in pursuits away from the home. In a study of metropolitan Washington, DC, residents in 1968, Chapin found that the hours spent away from home, on the average, ranged from 6.33 h on Sunday to 10.64 h on Friday.30 In other words, people spent between 26.4% and 44.3% of their day away from home. Chapin also reported that people travel an average of 14.2 miles per day, which suggested that people moved between areas represented by different FSM stations over the course of the day. By the early 1980s, additional studies raised further questions as to the ability of FSM stations to represent the CO exposures of the public.35,36 As stated previously, FSM stations use the NDIR method to measure ambient CO concentrations. Because NDIR monitors are not portable, they cannot be used to measure CO exposure as a person performs routine daily activities. The advent of microelectronics during the 1970s enabled considerable progress to be made in the development of reliable, compact, mobile air quality monitoring instruments. The most dramatic of these were the new miniaturized personal exposure monitors (PEMs) as described by Wallace and Ott.37 These instruments could go nearly anywhere, as they were equipped with batteries and shoulder straps. The utility of PEMs for measuring personal CO exposure was demonstrated in several early studies. These included a study of automobile commuters in Los Angeles during the summer of 1973,38 and field surveys of personal exposure in many commercial settings of several California cities between November 1979 and July 1980.39 Continued technical improvements of PEMs stimulated scholarly interest in how to use them in studies of personal and population exposure. Both direct and indirect methods in the use of PEMs have evolved. The direct method distributes PEMs to ordinary people who record their CO exposures and activities directly using either a paper diary or an electronic data logger. For example, Cortese and Spengler used this method in their survey of Boston commuters.29 A second method relies on indirect estimates of population exposure, based on the

8417: “8417_c002” — 2007/9/11 — 12:16 — page 13 — #9

Carbon Monoxide Poisoning

14

following model advocated by Fugas40 and Duan.41 Ei =

n


ck (tik )

(2.2)

k=1

where Ei = the total integrated exposure of person i over some time period of interest (e.g., 24 h) ck = the air pollutant concentration in microenvironment type k tik = the amount of time spent by person i in microenvironment type k n = the number of microenvironment types encountered by person i over the period of interest. The total exposure of a population can be determined by summation of the integrated exposures of individuals who are members of that population. The indirect method assumes that an individual’s total exposure to air pollution is a function of location and time. It follows that variation in the total exposure of an individual occurs because air pollution concentrations vary from one location to another and because time spent in different locations varies substantially from person to person. The indirect method postulates the existence of “microenvironments” in which a person is exposed to an air pollutant at a given concentration over a fixed period of time. Duan described a microenvironment as a “chunk of air space with homogeneous pollutant concentration.”41 Because of the potentially large number of microenvironments, Duan suggested that similar ones should be grouped, either by location (e.g., indoor or outdoor) or activity performed at a location (e.g., residential or commercial) into “microenvironment types.” Under the CAA, the U.S. EPA has authority to perform periodic reviews of the criteria that support the NAAQS. In the early 1980s, EPA scientists developed a riskanalysis framework to support its reviews of the NAAQS for CO.42,43 This framework required estimates of the percentage of an urban population that was exposed to CO concentrations that exceeded the NAAQS. The need for these estimates was partly related to EPA’s proposal to change the form of the primary standard from deterministic to statistical.44 In response to this need, the EPA supported development of several large-scale population models of CO exposure in the early 1980s. These models included the Simulation of Human Activity and Pollutant Exposure (SHAPE) model and the NAAQS Exposure Model (NEM). Subsequently, the EPA supported development of the probabilistic NEM for CO (pNEM/CO) and the Air Pollutants Exposure Model (APEX). To provide data for these models, the EPA funded both direct and indirect studies of urban population exposure. The indirect studies included measurements of CO in various microenvironments and a nationwide survey of human activity patterns as described later. EPA’s direct studies took measurements of the daily CO exposures of the noninstitutionalized, nonsmoking adult populations (ages 18–70 years) living in two metropolitan areas of the United States. The surveys were performed using portable CO exposure monitors equipped with data loggers to reduce the data-collection

8417: “8417_c002” — 2007/9/11 — 12:16 — page 14 — #10

Exposure to Ambient and Microenvironmental Concentrations

15

burden on the population. Field surveys of 454 residents of Denver, Colorado, and 714 residents of Washington, DC, occurred during the fall of 1982 and winter of 1983. Study participants carried PEMs equipped with data loggers. They also kept diaries for 48 h in Denver and for 24 h in Washington. The studies revealed that ambient CO levels at the composite network of fixed-site monitors were able to track variation in personal exposures. However, the network overestimated the 8-h exposures of people with low-level personal exposures and underestimated the 8-h exposures of people with high-level personal exposures. With respect to the underestimates, over 10% of the daily maximum 8-h exposures in Denver exceeded the NAAQS of 9 ppm, and about 4% did so in Washington. The end-expired breath CO levels were in excess of 10 ppm, which was roughly equivalent to about 2% COHb in about 12.5% of the Denver participants and about 10% of the Washington participants. Recall, that the NAAQS for CO are designed to keep COHb levels below 2% in the blood of the general public including probable high-risk groups. During the survey period, the composite CO concentrations at fixed-site monitors exceeded the 8-h NAAQS for CO (9 ppm) only 3% of the time in Denver, and never in Washington, DC. Hence, the results of these two surveys raised further doubts as to the ability of fixed-site monitors to represent the total CO exposure of urban populations.45 The results of the population exposure studies in Denver and Washington did not persuade the U.S. EPA to change the form of the NAAQS for CO from deterministic to statistical. In September, 1985, the EPA announced that it would retain the existing primary NAAQS for CO, but that it would rescind the secondary standard to protect public welfare, because there was no evidence to support it.7 However, one could argue that the results of these two studies provided some justification for the tighter CO emission standards that had taken effect for new cars sold outside of California during the 1980 model year. (Under the CAA, the state of California has authority to set its own tailpipe emission standards.) For example, Table 2.1 shows results for selected microenvironments of the Denver study. The table shows that higher CO concentrations were associated with commuting by motor vehicles (i.e., motorcycle, bus, car, and truck). It also shows that indoor CO concentrations in excess of the NAAQS were observed in public garages and in service stations or vehicle repair facilities. In the Washington study, participants who commuted 6 h or more per week had higher CO exposures than those who commuted fewer hours per week. Likewise, participants who had occupations that involved motor vehicles (e.g., professional drivers of trucks, buses, and taxis; automobile mechanics; garage workers; and policemen) had a mean CO exposure (22.1 ppm) that was over three times higher than the average exposure of those who did not work with motor vehicles (6.3 ppm). The Denver population exposure study provided raw data to test the SHAPE model and the probabilistic version of the NAAQS Exposure Model (pNEM/CO). The evaluation of SHAPE showed close agreement between the observed and predicted arithmetic means of the 1-h and 8-h maximum average CO exposures. However, SHAPE over-predicted low-level exposures and under-predicted high-level exposures.48 Likewise, an evaluation of pNEM/CO showed relatively close agreement between simulated and observed exposures for CO concentrations near the

8417: “8417_c002” — 2007/9/11 — 12:16 — page 15 — #11

Carbon Monoxide Poisoning

16

TABLE 2.1 Carbon Monoxide Concentrations of Selected Microenvironments in Denver, Colorado, 1982–1983 (Listed in Descending Order of Mean CO Concentration) n

Meana (ppm)

Standard Error (ppm)

22 76 3, 632 405 619 9

9.79 8.52 8.10 7.03 3.88 1.34

1.74 0.81 0.16 0.49 0.27 1.20

29 22 12 61 126 16 74 29 21

8.20 7.53 3.68 3.45 3.17 1.99 1.36 0.97 0.69

0.99 1.90 1.10 0.54 0.49 0.85 0.26 0.52 0.24

Indoor Public garages 116 Service stations or vehicle repair facilities 125 Other locations 427 Other repair shops 55 Shopping malls 58 Residential garages 66 Restaurants 524 Offices 2, 287 Auditoriums, sports arenas, concert halls 100 Stores 734 Health care facilities 351 Other public buildings 115 Manufacturing facilities 42 Homes 21, 543 Schools 426 Churches 179

13.46 9.17 7.40 5.64 4.90 4.35 3.71 3.59 3.37 3.23 2.22 2.15 2.04 2.04 1.64 1.56

1.68 0.83 0.87 1.03 0.85 0.87 0.19 0.002 0.48 0.21 0.23 0.30 0.39 0.02 0.13 0.25

Microenvironment In-Transit Motorcycle Bus Car Truck Walking Bicycling Outdoor Public garages Residential garages or carports Service stations or vehicle repair facilities Parking lots Other locations School grounds Residential grounds Sports arenas, amphitheaters Parks, golf courses

a An observation was recorded whenever a person changed a microenvironment, and on every

clock hour; thus each observation had an averaging time of 60 min or less. Source: Johnson, T. A Study of Personal Exposure to Carbon Monoxide in Denver, Colorado, Report No. EPA-600/4/84-014, Environmental Monitoring Systems Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC, 1984 as reported in U.S. EPA. Air Quality Criteria for Carbon Monoxide, Report No. EPA 600/8-90/045F, Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, U.S. Environmental Protection Agency, Research Triangle Park, NC, 1991.

8417: “8417_c002” — 2007/9/11 — 12:16 — page 16 — #12

Exposure to Ambient and Microenvironmental Concentrations

17

average exposure (i.e., within the range of 6–13 ppm for the 1-h standard and within 5.5–7 ppm for the 8-h standard). Like SHAPE, pNEM/CO also over-predicted lower exposures and under-predicted higher exposures for both standards.49

2.5 MICROENVIRONMENTAL EXPOSURES This section describes studies of CO exposures in microenvironments that people use to live, work, shop, play, and commute. In particular, it identifies factors that may contribute to high-level exposures in these microenvironments. High-level exposures deserve more attention, because population exposure models (i.e., SHAPE and pNEM/CO) appear to underestimate high-level exposures in these microenvironments, as stated earlier. These microenvironments were also chosen, because they were identified by the U.S. EPA’s National Human Activity Pattern Survey (NHAPS) as locations relevant to air pollution exposure. That survey collected 24-h retrospective diary data on activities and their locations from 9196 respondents interviewed in the 48 contiguous states from late September 1992 through September 1994. Table 2.2 summarizes the minutes spent on the diary days in six locations for respondents to the survey.50

2.5.1 RESIDENTIAL EXPOSURES Exposure to CO in the home is an important component of a person’s total daily exposure, because an estimated 68.7% of one’s time on average is spent inside a residence.50 Major sources of CO concentrations inside the home include unvented or poorly vented furnaces, gas appliances, fireplaces, wood stoves, kerosene space heaters, and charcoal grills and hibachis. Other CO sources include motor vehicles inside an attached garage. Studies of exposures to nonfatal concentrations are discussed first, followed by studies of unintentional deaths caused by high indoor CO concentrations.

TABLE 2.2 Time Spent in Different Locations by 9196 Participants of The National Human Activity Pattern Survey (NHAPS), October 1992–September 1994 Location In a residence Office-factory Bar-restaurant Other indoor In an enclosed vehicle Outdoors

Overall Mean (min)

Doer %

Doer n

Doer Mean (min)

990 78 27 158 79 109

99.4 20.0 23.7 59.1 83.2 59.3

9153 1925 2263 5372 7596 5339

996 388 112 267 95 184

Source: Modified from Klepeis et al., J. Expo. Anal. Env. Epid., 11, 231, 2001.

8417: “8417_c002” — 2007/9/11 — 12:16 — page 17 — #13

Carbon Monoxide Poisoning

18

2.5.1.1 Nonfatal Exposures Many Americans have appliances that emit CO in the home. These homes often use gas (natural gas and liquid propane) for cooking, heating water, and drying clothes. Of all participants of the NHAPS study, 38.3% had a gas range or oven at home, and 23.7% said that it had a burning pilot light. The same study showed that some people still use small appliances to heat a room: fireplace (10%); wood stove (6%); and kerosene space heater (2%). In the presence of indoor sources such as gas appliances, indoor CO concentrations often exceed outdoor levels.47 In a 1985 Texas study, CO concentrations were greater than or equal to the NAAQS for CO of 9 ppm in 12% of surveyed homes. Residential CO concentrations were high in cases where unvented gas space heaters were used as the primary heat source.51 According to the Barbecue Industry Association, 44 million American households owned a charcoal grill in 1989, and an estimated 600 million charcoal-barbecuing events took place annually.52 An earlier study showed that the air stream from charcoal grills contained 20–2000 ppm of CO, with 75% of grills emitting 200 ppm and above.53 Another study reported COHb levels ranging from 6.9% to 17.4% in a family of four people in northern California who had been exposed to smoke from cooking indoors on a barbecue grill, which was found by firefighters in the middle of the living room.54 Based on data for ten counties, a study in Washington state reported features of unintentional CO poisoning cases that occurred between 1982 and 1993.52 Most cases occurred when electrical power was interrupted during fall and winter months, because of either regional storms or unpaid utility bills. Of 509 patients treated with hyperbaric oxygen, 79 (16%) were exposed to CO emissions when charcoal briquets were burned for heating or cooking in 32 separate incidents. Non-English speaking Hispanic whites and Asians were disproportionately represented among the cases. The COHb levels of these 79 people averaged 21.6% and ranged from 3.0% to 45.8%. Two studies assessed CO exposure to emissions from unvented portable kerosene heaters in eight small mobile homes with no gas appliances and low air exchange rates.55,56 Each home was monitored for an average of 6.5 h per day for 3 days per week for 4 weeks. For 2 weeks the heater was on, and, for 2 weeks, it was off. When the heater was turned on, it was in use for an average of 4.5 h. When the heater was in use, study participants (all nonsmokers) spent most of their time in the family room or kitchen. Sampling took place in the living area about 1.5–3 m from the heater. The mean 8-h CO concentrations were 7.4 ppm (1-h peak = 11.5 ppm) when the heater was on and 1.4 ppm (1-h peak = 1.5 ppm) when it was off. Peaks usually were observed at the end of the combustion period. The ambient CO level measured 0.5 h prior to heater use ranged from 0 to 8 ppm. When the heater was on, three of the eight homes had 8-h average CO levels that exceeded the NAAQS, and one home routinely had levels of 30–50 ppm. A California study reported CO exposures for a random sample of homes that used gas appliances during a 48-h period from December 1991 to April 1992.57−59 For periods of 48 h, the median CO concentration was 1.2 ppm (indoors) and 0.8 ppm (outdoors), and the median of the maximum 8-h average CO concentration was 2.0 ppm (indoors) and 1.4 ppm (outdoors). Of surveyed homes, 13 of 286 homes (4.5%)

8417: “8417_c002” — 2007/9/11 — 12:16 — page 18 — #14

Exposure to Ambient and Microenvironmental Concentrations

19

had indoor CO concentrations above the NAAQS of 9 ppm for 8 h, and 8 of 282 homes (2.8%) had outdoor CO concentrations above this standard. Although most of the exceedances occurred in the Los Angeles basin, these percentages could be low because the basin was under-represented in the statewide sample. The study did not translate these percentages into statewide estimates. The study suggested that a small percentage of California homes would still have indoor CO problems even if outdoor CO levels at these homes complied with federal ambient standards. For a common sample of 277 homes, 17 homes (6.1%) had 1-h maximum concentrations indoors that were at least 5 ppm higher than outdoor levels, and 10 homes (3.6%) had 8-h maximum CO concentrations indoors that were at least 5 ppm higher than outdoor levels. Using univariate regression analysis, outdoor CO concentrations explained approximately 55% of the indoor CO variation. Higher net indoor CO levels (indoor minus outdoor CO concentrations) were traced definitively to space heating with gas ranges and gas-fired wall furnaces, use of gas ranges with continuous gas pilot lights, small home volumes, and cigarette smoke. However, several other factors also may have contributed to the higher CO levels: malfunctioning gas furnaces, automobile exhausts leaking into homes from attached garages and carports, improper use of gas appliances (e.g., gas fireplaces), and improper installation of gas appliances (e.g., forced air unit ducts).59 2.5.1.2 Fatal Exposures Factors contributing to unintentional deaths from CO poisonings were identified by studies in California and New Mexico. In California, two studies collected data for the 1979–1988 period. In the first study, 59 of 444 deaths (13.3%) were caused by improper use of charcoal grills and hibachis.60 Of the 59 deaths, 54% occurred inside motor vehicles (e.g., vans, campers) and 46% in residential structures (e.g., homes, apartments, shacks, tents). Relative to their share of the state’s population, higher death rates occurred amongAsians, blacks, males, and people aged 20–39. The second study identified specific factors that contributed to unintentional deaths caused by CO from several combustion sources (e.g., charcoal grills and hibachis, other heating and cooking appliances, motor vehicles, small engines, camping equipment).61 In this study, there was a strong association between alcohol use and CO poisoning from motor vehicles. Typically, motorists under the influence of alcohol would pull into their garages, leave the engine running while listening to cassette tapes, and then fall asleep. Faulty heating equipment used during winter months was implicated in about 50% of all unintentional deaths in both the California study61 and the New Mexico study.62 The National Center for Health Statistics (NCHS) and the U.S. Consumer Product Safety Commission (CPSC) estimated that 212 deaths in 1992 could be attributed to fuel-burning appliances used in the home. Of these deaths, 13 involved use of gasoline-powered appliances.63 An estimated 3900 CO injury accidents occurred in 1994, of which about 400 were associated with the use of gasoline-powered engines or tools. In response to the problem, several federal government agencies issued a joint alert concerning exposure to CO emitted by these sources.64 These sources involved use of pressure washers, air compressors, concrete-cutting saws, electric generators,

8417: “8417_c002” — 2007/9/11 — 12:16 — page 19 — #15

Carbon Monoxide Poisoning

20

floor buffers, power trowels, water pumps, and welding equipment. Unintentional CO poisonings frequently happened indoors even when people took precautions to ventilate buildings. Power outages following hurricanes and tropical storms often create demand for alternate sources of electricity (e.g., portable gasoline generators) to run air conditioners and refrigerators. But these generators can be a significant source of CO exposures if they are placed in garages or outdoors near windows. The majority of exposures occur overnight when generators are used to run air conditioners and other appliances. Hurricanes Katrina and Rita, which struck the U.S. Gulf Coast in the late summer of 2005, caused 10 deaths from CO poisoning in 18 storm-affected counties of Alabama and Texas.65 All of the fatalities were caused by gasoline-powered generators placed either inside the home or in a fully enclosed space outside the home. Very few homes had functioning CO detectors. In four hurricanes that hit Florida in 2004, some victims of CO poisoning placed generators inside their homes or garages to protect the devices from weather damage or to prevent theft.66

2.5.2 OCCUPATIONAL EXPOSURES The NHAPS study reported that 20% of Americans spent nearly 6.5 h per day on average working inside an office or factory.50 The National Institute for Occupational Safety and Health (NIOSH) estimated that 3.5 million workers who work in the private sector potentially are exposed to CO primarily from motor exhaust. The number of persons potentially exposed to CO in the work environment is greater than that for any other physical or chemical agent.67 In 1992, there were 900 work-related CO poisonings resulting in death or illness in private industry according to the U.S. Bureau of Labor Statistics as cited in a NIOSH report.64 Three risk factors affect industrial occupational exposure: (1) the work environment is located in a densely populated area that has high background (i.e., ambient) CO concentrations; (2) the work environment produces CO as a product or by-product of an industrial process, or the work environment tends to accumulate CO concentrations that may result in occupational exposures; and (3) the work environment involves exposure to methylene chloride (i.e., dichloromethane), which is metabolized to CO in the body. Proximity to fuel combustion of all types elevates CO exposure for certain occupations: airport employees; auto mechanics; small gasoline-powered tool operators (e.g., users of chainsaws); charcoal meat grillers; construction workers; crane deck operators; firefighters; forklift operators; parking garage or gasoline station attendants; policemen; taxi, bus and truck drivers; toll booth and roadside workers; and warehouse workers.47 Table 2.3 shows results for several occupational studies (typical CO values and/or ranges), averaging periods, and the measured or estimated percent COHb levels for nonsmokers, if reported. The CO exposure of office building workers has received less attention. Flachsbart and Ott83 developed a “rapid method” for surveying CO concentrations inside several high-rise buildings. In one case, they observed CO concentrations in excess of 9 ppm, which is the 8-h NAAQS for CO, on four visits to a 15-story office building in Palo Alto, California. A survey in April 1980, showed that CO concentrations in the building’s underground garage averaged 40.6 ppm. The CO levels ranged from

8417: “8417_c002” — 2007/9/11 — 12:16 — page 20 — #16

5.0–13.6 (0.25 h) 5–300 (0.1–1.7 h) (INT) 5.8–12.5 TWA (0.5–1 h) NA >200 ( 35 (1 h) 2.7 (8 h) 1–4.3 (8 h) 5–42 (ENV)

Airport workers NA NA NA 9.2–75.6 in 5 farmers NA > 4 in 10 NS 5–22 for 4 NS 4.2–28.7 for 7 NS 21.1 ± 0.7 6.3–13.3 for 4 NS > 3.25 in 5% of NS > 5 in 45% of NS NA NA 9 ppm) when people ride certain types of recreational vehicles (e.g., snowmobiles, power boats), gather indoors to barbecue food (sometimes to cope with electrical power outages after severe storms), and watch sporting events held at indoor arenas. High-level exposures (>25 ppm) may occur inside arenas when they are used for ice skating or motocross, monster-truck, and tractor pull competitions. Vehicles in these competitions often lack any type of emission controls. At some events, ventilation did not sufficiently lower CO concentrations to safe levels ( 20%

Is patient symtomatic?

YES

NO

Transport to nearest medical facility for continued evaluation, treatment, and monitoring (2)

PATCH

Consider field treatment with no transport. Continue treatment with oxygen until COHb is 18 years, versus 0.4 per million for children aged 17 years or younger) and men (2.8 per million men versus 0.9 per million women). The average daily number of CO-related deaths was greatest during the months of January and December (277 per month) and lowest during the months of July and August (91 per month). For the period from 1999 to 2002, only 32 states had a sufficiently large number of UNFR CO-related deaths to calculate stable mortality rates (Figure 9.1). Although numerous states had rates above the national four-year average annual rate, the state with the highest statistically reliable UNFR CO mortality rate was Nebraska (5.6 deaths per million person-years) and the state with the lowest reliable rate was California (0.6 average deaths per million) (Figure 9.1). Reporting of acute CO poisoning by healthcare providers was mandated for 13 states, although there was no clear pattern of differences in CO-related mortality between states with mandated reporting and those without.17 In 2005, the National Center for Injury Prevention, in conjunction with the National Center for Environmental Health, used multiple-cause-of-death mortality data to estimate the annual incidence of fatal and nonfatal unintentional, NFR CO poisoning.18 This study provided national crude annual UNFR CO death rates by demographic characteristics and season for 2 years: 2001 and 2002. In addition, case-fatality rates were estimated by dividing the number of deaths by the sum

UR

UR (NH) UR (VT)

UR

UR

UR UR

MR MR

UR MR

UR

UR (ME) UR/MR (MA)

MR

MR

UR (RI) UR (CT) MR (NJ) UR/MR (DE) UR (DC)

MR UR MR

UR UR(HI)

CO Death Rate* > U.S. Average = U.S. Average

*Rate per 1,000,000 population per year, age-adjusted to 2000 US Census population . UR = Rate unreliable MR = CO death is a mandated reportable condition

< U.S. Average

FIGURE 9.1 Average rates of unintentional, nonfire, carbon monoxide-related death, by state, United States, 1999–2002.

8417: “8417_c009” — 2007/9/11 — 12:15 — page 238 — #6

Public Health Surveillance for Carbon Monoxide in the United States

239

of UNFR CO deaths and nonfatal exposures multiplied by 100. Fatal UNFR CO poisoning deaths were identified as those with ICD-10 code T58 as a leading or contributing cause of death and an underlying cause of death code indicating either accidental poisoning (X47) or poisoning of undetermined intent (Y17). This study found that the crude annual UNFR CO death rate for the United States was 0.17 per 100,000 with an average of 480 deaths per year. The annualized incidence of fatal UNFR CO exposures was greatest during the fall and winter months, with more cases occurring during December (56 deaths) and January (69 deaths). The crude death rate from CO was highest for adults over 65 years (0.32 per 100,000), males (0.24 per 100,000), and nonHispanic whites and blacks (0.17 per 100,000). The case-fatality rate increased with age, ranging from 0.6% for children under 4 years to 5.5% for adults aged 55–64 years. Overall, males had a 2.3 times higher casefatality rate than women and 23.5% of deaths occurred among adults aged 65 years or more.18 In 2002, Mott et al.19 conducted an ecological analysis of national CO death data from the NVSS and annual CO emissions estimates for light-duty vehicles obtained from the US Environmental Protection Agency. This study estimated the percent change in CO emissions and CO mortality rates by intent and mechanism for the years 1968–1998. The main outcome from this study was US resident deaths from 1968 to 1998 coded with ICD-8 or ICD-9 code N986 (toxic effect of CO) as a contributing cause -of -death or those records with an ICD external cause of injury code exclusive to CO poisoning (Table 9.2). From 1968 through 1998, this study identified 116,703 NFR deaths due to CO in the United States. During this same time period, crude mortality rates associated with NFR CO declined from 20.2 deaths to 8.8 deaths per million person-years in the United States. There were 2.2 intentional deaths for every one unintentional NFR CO-related death. Between 1968–1998, motor vehicles were identified as the mechanism for 70.6% of deaths and, following the introduction of the catalytic converter in 1975, annual estimates of NFR CO emissions decreased by 76.3%. In addition, unintentional motor vehicle-related NFR CO death rates declined by 81.3% and rates of motor vehicle-related NFR CO suicides declined by 43.3%.19 In 1996, the CDC published a summary of findings from an investigation of deaths associated with multiple motor-vehicle related CO poisonings in Colorado and New Mexico, together with national estimates of CO deaths from 1979–1992.20 This report presented the geographic pattern of national unintentional CO mortality associated specifically with stationary motor vehicles (UMVR). Deaths were identified using ICD-9 code E868.2, a code specific to deaths due to accidental poisoning by CO or another utility gas from the following sources: farm tractor not in transit, gas engine, motor vehicle not in transit, or any type of combustion engine not in a watercraft. This analysis found that crude death rates for UMVR CO were highest in states in the northern regions of the United States from 1979 to 1992, although no specific national or state rates were provided.20 In the early 1990s, Cobb and Etzel2 published a descriptive analysis of unintentional NFR CO-related deaths (UNFR) in the United States. This study calculated crude and age-adjusted death rates by demographic characteristic, season, and state for the years 1979–1988. All US resident deaths coded using ICD-9 code N986 were

8417: “8417_c009” — 2007/9/11 — 12:15 — page 239 — #7

240

Carbon Monoxide Poisoning

identified between 1979–1988, excluding those coded as intentional (codes E950– E959 and E960–E969), intent undetermined (E980–E989), and those owing to fire (E837, N940–N949, E923). The 10-year average crude UNFR CO death rate for males (0.78 per 100,000 persons) was found to be 3 times higher than that for females (0.26 per 100,000 persons). Overall, 83% of UNFR CO-related deaths occurred among whites, yet “race-specific death rates were more than 20% higher for blacks (0.63 per 100,000) than for whites (0.51 per 100,000).”2 Annual incidence of unintentional CO-related death was found to be highest in the month of January, with an average of 181 deaths, and lowest in the month of July, with an average of 44 deaths. The state with the highest age-adjusted UNFR CO death rate for the 10-year period was Alaska (2.72 per 100,000) and the lowest rate was found in Hawaii (0.05 per 100,000).2 In addition to the above-mentioned surveillance reports produced by the CDC, national estimates of CO-related mortality produced by the Consumer Products Safety Commission (CPSC) may provide useful national surveillance for CO-related mortality. These reports combine data from multiple sources, such as the NVSS and National Electronic Injury Surveillance System (NEISS), with proprietary CPSC datasets21 on the basis of death certificate data purchased directly from states and in-depth follow-up investigations of select deaths. These data are used to produce annual national estimates of unintentional NFR CO exposures and fatalities by product, victim age, and incident location. However, the reports currently available are limited to estimates associated specifically with the use of consumer products under the jurisdiction of the CPSC (e.g., engine-powered tools, charcoal grills, gas ovens, and other appliances). CPSC reports do not capture fatalities associated with motor vehicles, despite the fact that CO in motor-vehicle exhaust has been found to account for the majority of poisoning deaths in the United States.22 The most recent year for which a complete CPSC report of CO-related mortality is available is 2002.23 In 2002, CPSC identified 188 unintentional, nonfire CO-related deaths and the average annual estimate from 1999 to 2002 was 141 deaths. Heating systems were associated with 55% of deaths and engine-powered tools were associated with 28% of deaths in 2002. An estimated 71% of CO deaths occurred in the home. Overall, 81% of fatal CO incidents involved a single death, with adults over 45 years of age accounting for 55% of all unintentional, nonfire CO deaths in 2002.23 Although CPSC memoranda have been published covering both non-fatal CO exposure incidents and fatalities for the periods 1990–2004 and 2002–2005, respectively, CPSC notes that the counts for recent years contained in these reports may not be complete.24 Briefly, from 1990 to 2004, there were 318 nonfire CO-related deaths identified by CPSC, 274 of which were associated with the use of generators. Most of these deaths (39%) occurred during winter and 77% took place in the home. From 1990 to 2004, an estimated 33% of deaths occurred among adults aged 45–64 years and 75% of all “engine-driven tool” CO-related deaths were males.24 From 2002 to 2005, preliminary counts from CPSC indicate that 253 nonfire CO-related deaths were associated with engine-driven tools in the United States. An estimated 218 (86%) of CO deaths were associated with gasoline-powered generators.20

8417: “8417_c009” — 2007/9/11 — 12:15 — page 240 — #8

Public Health Surveillance for Carbon Monoxide in the United States

241

9.2.3 CARBON MONOXIDE-CAUSED MORBIDITY National surveillance for nonfatal CO exposures is challenging, given the limited number of data sources currently available (Table 9.1). One potential source for national nonfatal CO poisoning data is the NEISS, operated by the CPSC since 1971.25 The NEISS has been used to monitor consumer product-related injuries resulting from consumer products under the regulatory jurisdiction of the CPSC. In 2000, the surveillance system was expanded to collect data about all types and causes of injuries and poisonings treated in hospital EDs, whether or not they were associated with consumer products. This expanded system is called the NEISS All Injury Program (NEISS-AIP). NEISS data are currently collected from a nationally representative sample of 100 EDs, selected from a stratified probability sample of all US hospitals that have at least 6 beds and provide 24-h emergency services, while NEISS-AIP data are collected from a 66 hospital subsample. Data from approximately 500,000 injuryrelated ED cases are collected annually by NEISS and coded for cause and intent of injury using guidelines consistent with coding guidelines in the ICD-9-CM.25 Use of these data for CO surveillance is limited, however, they exclude incidents owing to occupational exposure, those associated with motor vehicles, and cannot provide state or local estimates.25 Another source of national nonfatal CO poisoning surveillance data is the National Hospital Ambulatory Medical Care Survey (NHAMCS) conducted by the National Center for Health Statistics of the CDC. The NHAMCS collects data on the utilization and provision of ambulatory care services in hospital emergency and outpatient departments. Findings are on the basis of a national sample of visits to the emergency and outpatient departments of noninstitutional general and short-stay hospitals, exclusive of federal, military, and veterans administration hospitals, located in the 50 States and the District of Columbia. Annual data collection began in 1992. Similarly, the national Toxic Exposure Surveillance System (TESS) operated by the American Association of Poison Control Centers has demonstrated capability to serve as a source of CO data for both state-based26 and national studies.27 Implemented in 1983, TESS is a comprehensive database that contains detailed toxicological information about over 24 million poisoning incidents reported to 61 poison control centers in the United States.28 The calculation of rates of calls to TESS for a given exposure is facilitated because participating poison control centers also report the size of the population they serve. However, the use of TESS data for CO surveillance may be limited by the fact that most reports pertain to the treatment of nonfatal poisonings. Therefore, TESS data may underestimate the incidence of fatal CO exposures. As of October, 2006, annual reports summarizing TESS data are available online for the years 1983–2004 (http://www.aapcc.org/annual.htm). Finally, recent studies have suggested that data obtained from hyperbaric oxygen (HBO) therapy centers may serve as an indicator of CO burden.27,30 HBO is typically used to treat those patients with the most severe CO poisoning,27 comprising approximately 6% of all patients seen for CO in EDs.30 According to Hampson, “as long as the spectrum of severity of the condition and treatment practices has not been recognized to have changed significantly during the period surveyed, the national rate of HBO therapy should serve as a qualitative marker of total disease incidence”.27

8417: “8417_c009” — 2007/9/11 — 12:15 — page 241 — #9

Carbon Monoxide Poisoning

242

Using data from HBO therapy facilities for national CO surveillance is limited by the lack of general availability and the fact that these data do not include less severe cases, those not referred for HBO, or those treated at EDs using HBO. Likewise, a patient’s state of residence may differ from that in which they received the HBO therapy, complicating the calculation of valid state rates. 9.2.3.1 Recent National Studies of Nonfatal Carbon Monoxide Poisoning A 2005 study by Hampson27 used TESS data and information from a recent survey of HBOT facilities29 to compare trends in the annual incidence of nonfatal CO poisoning in the United States. Hampson27 searched TESS records from 1985 to 2002 for all calls received by poison control centers regarding cases of CO exposure, and calculated annual call rates per million person-years. This study found that annual call rates for CO poisoning increased from 31.1 per million in 1985 to 95.4 per million in 1996. Although the rate decreased to 54.5 per million persons from 1996 to 2002, there was a significant increase in calls for the entire 18 year period (p = .0022).27 Rates of HBOT for CO poisoning and rates of calls to poison control centers were also found to correlate strongly (r = 0.82, p = .0002).27 Also in 2005, the CDC published national estimates of unintentional, NFR CO exposures for 2001–2003 using NEISS-AIP data.18 Nonfatal cases of CO poisoning were considered to be those coded as “CO exposure” or “CO poisoning” in NEISS-AIP hospital data. Additional criteria used to identify nonfatal CO poisonings included: (1) intent of injury unintentional or undetermined, (2) principle diagnosis of “poisoning” or “anoxia”, and (3) additional narrative information indicative of CO. This study estimated that 15,200 patients were treated in EDs annually for nonfatal, unintentional, NFR CO poisoning during 2001–2003. Overall, the rate of nonfatal exposure was similar for males and females, although the rate was highest for children under 4 years of age (8.2 per 100,000 person). While most nonfatal CO exposures (64.3%) occurred in the home, only 9.3% of patients reported owning a CO detector.18 This study also provided national estimates of CO-related mortality discussed earlier in this chapter. Although not pertaining to CO specifically, a CDC report published in 1999 used data from the NHAMCS to describe poisoning-related ED visits from 1993 to 1996 in the United States.31 This study used a variety of ICD-9-CM codes to identify all injury-related ED visits, but did not provide sufficient detail to describe the etiology, intent, or mechanism of CO exposures. From 1993 to 1996, toxic effect of CO (ICD-9-CM code N986) was the sixth leading principle diagnosis identified among poisoning-related ED visits in the United States An annual average of 34,000 ED visits were attributed to CO during the 4 year study period.31 In an earlier study, Hampson30 reported a synthetic national estimate of nonfatal CO poisoning on the basis of a survey of EDs located in Washington, Idaho, and Montana. This survey collected data about the total number of patients seen for acute CO poisoning in 1994, although details regarding the etiology, source, or intent of the poisoning were not collected. This analysis extrapolated data from the three states to the United States as a whole and used state-specific age-adjusted CO death rates

8417: “8417_c009” — 2007/9/11 — 12:15 — page 242 — #10

Public Health Surveillance for Carbon Monoxide in the United States

243

to adjust for the fact that CO poisoning is more common in the Pacific Northwest region. The study estimated 42,890 ED visits for CO poisoning occurred in 1994 in the United States, resulting in a national ED visit rate of 16.5 per 100,000 persons for CO poisoning.30 The public health literature describing surveillance of CO poisoning included in this chapter is characterized by its variability. Although most studies focus on fire-related CO poisoning, different data sources and definitions of poisoning are used for case ascertainment, thus limiting the comparability of national estimates. In addition, other national estimates of nonfatal CO poisoning are available in the literature, although many of these focus only on poisonings resulting from a specific mechanism (i.e., consumer products) or are very old. For example, recent reports from the CPSC have used data from the NEISS to estimate that approximately 5000 nonfatal CO poisonings associated with motor-driven appliances occur annually in the United States.32 Likewise, a 1974 study estimated that 10,000 nonfatal CO poisonings occurred annually in the United States.33 The variability among national studies and estimates of CO poisoning underscores the need for a national surveillance system for CO.

9.3 TOWARD A NATIONAL SURVEILLANCE SYSTEM FOR CARBON MONOXIDE The lack of information linking environmental hazards to health outcomes has contributed to what has been labeled the “environmental health gap” by The Pew Environmental Health Commission.34 To address this gap and coordinate the development of CO surveillance in the United States., the CDC established The National Workgroup on Carbon Monoxide Surveillance, a partnership of public health professionals and agencies spanning private and federal jurisdictions from environmental health and injury prevention to emergency response.5 This workgroup has noted that the need for nationwide CO surveillance “is recognized in the Healthy People 2010 goal for the United States of ‘increasing the number of Territories, Tribes, and States, and the District of Columbia that monitor carbon monoxide poisoning from 7 to 51.’”5 One indicator of progress toward the Healthy People 2010 goal for CO is the number of states that mandate reporting of CO poisoning as part of the National Notifiable Disease Surveillance System. The National Council of State and Territorial Epidemiologist (CSTE) serves as a partner in the National Workgroup for CO Surveillance and monitors patterns of nonnotifiable diseases and conditions; currently, 15 states mandate reporting of CO poisoning (http://www.cste.org/NNDSSSurvey/ 2004NNDSS/NNDSSstatechrreporcondnona2005.asp). Similarly, although a number of states have mandated the installation of CO detectors, surveillance data are needed to evaluate the effectiveness of ongoing legislative interventions for CO.

9.3.1 CASE DEFINITIONS Although there is a growing recognition of the need to develop a national surveillance system for CO, methods of case identification and ascertainment remain a

8417: “8417_c009” — 2007/9/11 — 12:15 — page 243 — #11

Carbon Monoxide Poisoning

244

serious challenge for injury and death due to CO. The ability to accurately describe CO-related morbidity and mortality depends heavily on the methods used to identify and classify cases, the rules for coding data, and the type of data available. Effective public health surveillance relies on the establishment of a widely adopted, clear, and reliable case definition that includes criteria describing person, place, and time.35 To date, no such consensus has been reached in the case of CO poisoning. According to the National Workgroup on Carbon Monoxide Surveillance, 2 national public health organizations have published their own version of a case definition for public health surveillance of CO poisoning; The CSTE and the State and Territorial Injury Prevention Directors Association (STIPDA). The STIPDA definition is more conservative than the CSTE definition, that is, if both definitions are applied to the same dataset, the STIPDA definition will identify fewer records as having CO poisoning than the CSTE definition. The CSTE definition has been modified by at least one state to make it a more conservative definition.36 The case definitions suggested by CSTE and STIPDA are listed in Table 9.3. A formal evaluation of these case definitions is the focus of a current study by the National Workgroup for CO Surveillance.36

9.3.2 MORTALITY LIMITATIONS In addition to achieving a uniform case definition for tracking CO poisoning, understanding the limitations of available data is vital to improve national CO surveillance efforts. For mortality specifically, the use of death certificate data for national CO

TABLE 9.3 Two Case Definitions for the National Surveillance of Carbon Monoxide in the United States Carbon Monoxide Case Definitions CSTE Confirmed case: ICD-9 Coded Data: (1) a record in which the Nature of Injury code N-986 “Toxic effect of CO” is listed or (2) a record in which an External Cause of Injury (E-Code) indicating exposure to carbon monoxide (exclusively) is listed such as E868.3, E868.8, E868.9, E952.1, or E982.1. Probable case: ICD-9 Coded Data: A record in which an E-code indicating acute carbon monoxide poisoning inferred from motor-vehicle exhaust gas exposure is listed, ie. E868.2, E952.0, or E982.0. STIPDA Records must have the N-code for CO poisoning (986) in the principal diagnosis field. Because this case definition relies only upon the presence or absence of the N-code, it does not define classification of cases, such as confirmed and probable. Source: National Workgroup on Carbon Monoxide Surveillance. Project to evaluate carbon monoxide surveillance CSTE and STIPDA case definitions with hospital data. 2006 Unpublished report. Obtained October, 2006 from the Air Pollution and Respiratory Health Branch, U.S. Centers for Disease Control and Prevention.

8417: “8417_c009” — 2007/9/11 — 12:15 — page 244 — #12

Public Health Surveillance for Carbon Monoxide in the United States

245

surveillance is limited by recent changes in the way deaths are coded and classified. As noted earlier, the Tenth Revision of the ICD (ICD-10) was implemented in 1999 and is currently used to categorize causes of death in the NVSS. Earlier versions of ICD coding provided external cause of injury codes exclusive to CO poisoning (ICD 8 = E874, E875, and E952.1; ICD 9 = E868.3, E868.8, E868.9, E952.1, and E982.1) or poisoning from motor vehicle exhaust (ICD8 = E873 or E952.0; ICD9 = E868.2, E952.0, and E982.). Since CO is the only acutely poisonous gas in motor vehicle exhaust, in the past these latter codes could also be used to indicate CO poisoning. Although numerous ICD-10 codes mention CO, the Tenth Revision of ICD has only one code specific to CO:T58. While the latest revision contains codes that indicate the manner or intent of injury or can identify fire-related exposures, much detail has been lost in the ability to describe certain etiologic mechanisms (such as motor vehicle-related injuries) that could be described previously using E-codes related to CO. Hence, estimates of CO deaths classified using ICD-9 are not directly comparable to estimates derived from ICD-10 coded data. In addition to limitations regarding case ascertainment, data from NVSS do not become available to the public in a timely manner and lack other etiologic detail necessary for CO surveillance. For example, NVSS data do not contain sufficient detail to identify multiple victims of the same incident of CO poisoning. Estimating CO-related death from the NVSS requires a thorough understanding of the ICD coding rules and the limitations for case identification imposed by selection criteria for specific years of data. To summarize, the most important implication of the implementation of ICD-10 is that the already limited etiologic detail in the NVSS has been further reduced, thereby substantially decreasing its usefulness as national CO surveillance system. The use of the NVSS for national mortality surveillance indicates that there are roughly 500 unintentional, NFR CO deaths per year in the United States. Follow-up epidemiologic investigations of these deaths, however are needed to provide us with enough etiologic information to suggest meaningful public health interventions. Such investigations remain a high priority until alternate sources of surveillance data can be made available.

9.3.3 MORBIDITY LIMITATIONS Data currently available for the surveillance of nonfatal CO poisoning are limited by difficulties with case ascertainment. CO poisoning is characterized by nonspecific signs and symptoms that are often ignored or attributed to another condition by the public and medical professionals alike.11 In addition, those who experience nonfatal CO poisoning but do not present for medical treatment are not counted in morbidity estimates from medical records. Hence, all estimates of nonfatal poisoning are likely to underestimate the actual incidence of CO poisoning. The variation of populations and methods evident among datasets currently available suggest that no single source can serve as a comprehensive surveillance source for CO poisoning morbidity in the United States at this time. Despite these limitations, the continuing morbidity and mortality associated with acute CO poisoning in the United States necessitates the establishment of a national surveillance system. The goals of a national surveillance system

8417: “8417_c009” — 2007/9/11 — 12:15 — page 245 — #13

Carbon Monoxide Poisoning

246

for CO range from planning for rapid public health responses following disasters and tracking the burden of CO poisoning over time, to improving our understanding of exposure sources, related hazards, and facilitating research.5,36 To promote and facilitate surveillance at the local, state, and federal levels, the National Workgroup for CO Surveillance has produced a summary report describing the attributes of CO as a model environmental public health indicator (http://www.maine.gov/dhhs/eohp/epht/documents/CO_White.pdf).5

9.4 CONCLUSIONS Although the overall national rate of death from NFR CO declined steadily from 1968 to 1998 in the United States,19 the average annual rate of unintentional nonfire-related (UNFR) CO-related death has remained relatively stable over the past 4 decades.17–19 From 1968 to 1998, the crude death rate for UNFR CO was 7.06 deaths per 1 million person-years.19 The most recent analysis of mortality data from 2002 reported a crude UNFR CO death rate of 1.8 deaths per 1 million person in the U.S.17 There are an average of 494 accidental, NFR deaths and approximately 1,747 intentional deaths due to CO-poisoning each year in the United States.19 Evidence from a recent ecological study suggests the decrease in all CO-related deaths in the United States was driven primarily by a reduction in deaths from exposure to motor-vehicle exhaust.9 This reduction has been attributed to the national implementation of the 1970 Clean Air Act. In contrast to trends in national CO mortality, the incidence of nonfatal CO poisonings rose, while rates of death fell from 1985 to 1996 in the United States27 The annual number of nonfatal poisonings then decreased from 1996 to 2002.27 Published estimates of ED visits due to UNFR CO poisoning suggest there are approximately 15,200 ED visits per year, with 1,676 resulting in subsequent hospitalization.18 Similarly, unpublished estimates from the 2002 National Hospital Discharge Survey using the CSTE-proposed case definition identified 1496 hospitalizations due to UNFR CO poisoning in the U.S.37 Across studies, men and older adults (ranging from over 45 to over 65 years) are most at risk for unintentional death or injury from CO.17–19 Despite the lack of uniform national surveillance data, the burden of CO poisoning in the United States may be summarized using estimates from a variety of related data sources (Figure 9.2). The substantial health burden of unintentional CO poisoning illustrated in Figure 9.3 suggests the need to put CO morbidity and mortality under ongoing public health surveillance. The de facto national surveillance system for CO poisoning in the United States has been the NVSS. However, this system has several limitations for the surveillance of CO poisoning. Although the data have taken two years to process following collection, the recent implimentation of ICD-10 codes has further limited the utility of the NVSS for CO surveillance due to the removal of several important ICD codes that denote etiologic mechanism of poisoning. Public health professionals must consider these recent changes in the ICD coding system when assessing CO-related mortality derived from the NVSS. When estimating nonfatal CO poisoning, data from multiple sources should be used to cross-validate estimates of morbidity, to reduce the likelihood of under-estimation from any single source of data. To obtain a complete picture of

8417: “8417_c009” — 2007/9/11 — 12:15 — page 246 — #14

Public Health Surveillance for Carbon Monoxide in the United States

247

120

10

100

8

80

6

60

4

40

1975– catalytic converters required on all new passenger cars

2

20

0

CO Emissions control timeline Grams of CO emited per VMT

CDR per1,000,000 personyears

Suicides 12

1975: Catalytic converter introduced on new passenger cars to meet new CO emissions standard of 15 g/mile. 1978:1975 and newer model year cars make up 34% of the U.S.passenger vehicle fleet.

0

5

120

4

100 80

3 60 2 40

98 19

94

96 19

92

19

90

19

88

CDR: Nonmotor vehicle

19

86

19

84

19

82

19

80

19

78

CDR: Motor vehicle

19

76

19

74

19

19

19

19

19

72

0 70

20

0 68

1

COEmissions (g/mile)

Gramsof CO emited per VMT

Unintentional deaths CDR per1,000,000 personyears

1970:Congress enacts Clean Air Act. CO emissions standard at 34.0 g/mile.

1980: All new passenger cars required to meet new CO emissions standard of 7.0 g/mile. 1975 and newer model year cars make up 50% of US.passenger vehicle fleet. 1981:All newcars required to meet new CO emissions standard of 3.4 g/mile. 1990:1975 and newer model year cars make up 91% of the U.S.passenger vehicle fleet. 1992: Standards setting emission limits for carbon monoxide at temperatures CO2 Similar reactions take place for all other toxic gases that are capable of being electrochemically oxidized or reduced. From the reaction at the counter electrode, it is evident that oxygen is required for the current generation process to take place. This is usually provided in the sample stream by air diffusing to the front of the sensor, or by diffusion through the sides of the sensor (a few thousand ppm is normally sufficient).

8417: “8417_c010” — 2007/9/11 — 12:15 — page 262 — #12

Carbon Monoxide Sensors and Systems

263

However, continuous exposure to an anaerobic sample of gas may result in signal drift, despite the oxygen access paths which may cause the sensor to be poisoned. Similar to SMO gas sensors, electrochemical gas sensors are also affected by temperature variations. The baseline signal of most electrochemical sensors tends to increase exponentially with temperature, approximately doubling for every 10◦ C rise in temperature which proves problematic for domestic applications as the baseline shift with temperature could seriously affect the ability to measure these gases accurately and result in false alarming. Nevertheless, by compensating for this drift either in the hardware or software, such temperature influences can be reduced.

10.3 SENSOR SYSTEMS The CO sensor is the main component within all domestic CO detectors. Support electronics are also required to provide the sensor with intelligence so that it will actuate alarms according to compliant standards. Most detectors incorporate at least one microprocessor that allows them to be quickly reprogrammed and the behavior of the alarm to be altered to suit various applications or standards. For domestic applications, CO alarm design and alarm requirements are well defined by associated performance specifications. However, in emerging CO and air quality monitoring applications such as monitoring vehicle cabin air quality, specifications, and standards have yet to evolve. Vehicle cabin air quality concerns are usually generated by the following four scenarios: (1) Pollutant gases entering the vehicle through the ventilation system, (2) A lack of fresh airflow resulting in low oxygen and high carbon dioxide concentrations due to occupant respiration, (3) Pollutant gases entering from the external environment through window openings, imperfect seals, and other holes, and (4) Toxic gases entering the vehicle cabin by redirected exhaust fumes for self-harm (i.e., suicide) purposes. Currently, no system or aftermarket product addresses all four vehicle AQM concerns. Only two commercial AQM solutions currently exist for vehicles: (1) The most common are AQM systems controlling HVAC ventilation flaps, and (2) Less common are aftermarket toxic gas alarms for vehicle cabin applications, such as that commercialized by the Quantum Group (U.S.). Currently, the demand for AQM systems is driven by the increasing concern for passenger safety, health, comfort, and by automakers aiming for features and attributes that differentiate their vehicles. In turn, this growth has increased demand for reliable automotive air quality sensors. Figure 10.7 shows a simplified view of an AQM system controlling the HVAC ventilation flap. External gases enter the vehicle cabins through the ventilation system. Mounted in the air intake of the HVAC system, the AQM sensor sends a signal to the fresh air inlet flap to close when pollutant gases are detected and automatically reopen when the external air quality returns to an acceptable level. Although a driver could close the air inlet manually, forgetting to reopen it could cause the oxygen concentration in the cabin to decrease and carbon dioxide levels to increase. Therefore, a compromise must be reached. One way of tackling the problem to implement with the system an air quality factor. For instance, the absolute concentration of particular gas (Cx ) in the vehicle cabin is dependent on the exhaust flow rate (F), time (T ), cabin

8417: “8417_c010” — 2007/9/11 — 12:15 — page 263 — #13

Carbon Monoxide Poisoning

264

CO2

O2 OPEN

3. Ventilation system/flap 2. Electronics/algorithms

CLOSED

1. Sensor response

CO NOx HC Particulates

FIGURE 10.7 An overview of a typical automobile air quality monitor (AQM) employed within a heating, ventilation and air conditioning (HVAC) system. When the ventilation flap is open, dangerous pollutants such as CO and NOx may enter the cabin. To mitigate this, electronics automatically close the ventilation flap. High carbon dioxide and low oxygen concentrations may result through occupant respiration. High carbon dioxide and low oxygen concentrations are dangerous because they induce fatigue and drowsiness, reducing driver attention and response times.

volume (V ), and cabin seal (S). Therefore, Cx = f (F, T , V , S) Concentrations of carbon monoxide (CCO ), and oxygen (CO2 ) have been identified as important gas species contributing to poor cabin air quality. The summation of each absolute gas species concentration gives rise to an air quality factor (AQcabin ) such as: AQcabin = αCO + δ(CO2 )−1 Where α, and δ are proportionality coefficients. It should be noted that other gas species such as hydrocarbons and nitrogen oxides have been ignored. Absolute threshold limits could then be set for scenarios such as suicide (AQsuicide ) and driver fatigue (AQfatigue ). For increased reliability and effective suicide attempt identification, the change of air quality with time (dAQcabin /dt) should also be incorporated into the driver fatigue and suicide detecting algorithms: ∂CCO ∂(CO2 )−1 dAQcabin =α +δ dt ∂t ∂t An alarm threshold, dAQsuicide /dt, could also be incorporated as done so by Quantum Group. Therefore, the cabin gas-sensing system should include both absolute and changing air quality factors, to determine if alarms need to be activated. Software solutions to improve CO detectors are commonplace. In addition, rate of change, humidity compensation (through humidity sensors) and temperature

8417: “8417_c010” — 2007/9/11 — 12:15 — page 264 — #14

Carbon Monoxide Sensors and Systems

265

compensation (through temperature sensors) are also common within sensor systems. However, compensation for environmental variables increases the cost of the CO detector. Data referenced above showing that 79% of alarms failed when tested at 5% RH3 is compelling evidence of performance standards not meeting real-life long-term requirements for adequate domestic CO monitoring.

10.4 CONCLUSIONS Detection of CO has gone far beyond the primitive approach of Claude Bernard and others. CO sensors and detectors employ advanced materials, electronics, and software to ensure reliable and selective performance while maintaining economic sensitivity and feasible for the domestic market. This chapter has discussed the three major sensing techniques employed in mainstream CO detector/alarms. SMO sensors depend on chemi-absorption between the oxide and CO molecules for CO detection. Various methods are employed by industry to increase selectivity through the introduction of catalysts such as Pt and Pd, and filters using activated carbon. Optical sensors depend on CO energy absorption by incident photons. Humidity, temperature, and pressure are environmental factors that may affect sensor components. Electrochemical sensors are also vulnerable to cross sensitivity, temperature, and humidity variations. The intrinsic deficiencies of materials and electronic components that make up commercial CO sensors and systems have been documented and are well known. These issues have plagued manufacturers and the research community for many years and continue to be areas of active scientific interest. Nevertheless, economic forces, government legislation, competition, and customer demand drive CO detector products to be sold at the lowest possible prices, while high customer standards, product superiority, competitive advantage and market reputation drive product quality and innovation. Hence, these forces lead to the classic economic balance between price and performance.

10.5 ACKNOWLEDGMENTS The authors kindly thank all contributors including Dr Mark Goldstein from Quantum Group, Inc. (USA), Dr Herve Borrel from MiCS (Switzerland), Dr Nobuaki Murakami from FiS (Japan), Dr Jürgen Schilz from PerkinElmer Optoelectronics (Germany), and City Technology Sensors (UK).

10.6 APPENDIX 1 (D.G. PENNEY) Suicide in Australia, especially that of young men, had attained an alarming rate in recent years, higher than that in the USA and most other countries. The use of motorvehicle exhaust gas for this purpose was the most popular method. For this reason, in 1998 the Australian Medical Association in cooperation with other governmental and industrial groups as well as various individuals in Australia, invited me to provide conceptual solutions for reducing this tragic loss of young life.

8417: “8417_c010” — 2007/9/11 — 12:15 — page 265 — #15

266

Carbon Monoxide Poisoning

Almost immediately, it was felt that by limiting the availability of the lethal component of motor-vehicle exhaust gas, carbon monoxide (CO), the use of this method of commiting suicide would decline, possibly saving several hundred lives per year. The use of CO was often chosen because of its availability, ease of use, and supposed painless induction of unconsciousness. Making CO unavailable would defeat this approach. Several solutions were considered: (1) Accelerate the rate of installation of effective catalytic converters on Australian motor vehicles, possibly by instigating a retro-fit program. (2) Require the sale or retroinstallation of CO detectors on all motor vehicles in Australia that would warn drivers/occupants of the danger, and/or, immediately shut-down the engine and prevent restarting. (3) Place a distinctive odorant in petrol/gasoline that would give motor-vehicle exhaust an unpleasant odor and thus discourage/warn potential suicide attempters. (4) Design ignition systems that would prevent motor vehicles from remaining in an “idle” mode for more than a short time. Catalytic converters are expensive, eventually wear-out, are slow to come on-line in Australia owing to a long mean vehicle life, and retro-fitting would be difficult and place financial burdens on people least able to pay. Also, current catalytic converters still permit exhaust gases to contain lethal CO concentrations. Finally, the fact that catalytic converters only become effective in reducing CO at elevated temperatures means that exhaust gases would continue to contain supra-lethal concentrations of CO during the “warm-up” period. CO detectors that produce engine “shut-down” would have to be carefully designed so as not to exacerbate traffic problems due to elevated ambient CO concentration. This approach appeared to be the best overall solution, and could have provided some additional health benefits separate from the suicide issue. Solutions involving odorants in motor-vehicle fuel might cause public discomfort and complaints and undesirable environmental pollution. Most motor vehicles need the capability to idle, for example, waiting for traffic or stop lights, taxis, and vehicles being repaired. Australia represented just 1% of the world motor-vehicle market. The average age of Australian motor vehicles (8 × 106 ) in 1997 was 12–14 years. It was my charge in visiting Australia in early April, 1998, to recommend to the Australian Medical Association (AMA) and the Working Group on Motor Vehicle Exhaust Suicide, a CO concentration that might be set as the threshold at which engine shut-down would occur. Mathematical modeling of motor vehicle exhaust gas revealed a “unique signature” that might be used to quickly and unequivocally identify a suicide attempt, distinct from simple leakage of outside gases into the vehicle. Motor vehicles would be equipped with a sensor array in the passenger compartment that was sensitive to: carbon monoxide, carbon dioxide, and oxygen. Sensor output would be directed to a microchip with an embedded program such that: (1) measured CO concentration was integrated over time in a manner modeling human CO uptake, and thus provides a Low warning alarm at 35 ppm (7% COHb), and a High warning alarm at 100 ppm (14% COHb), and (2) A CO concentration at 100 ppm and above, as well as rapidly rising CO2 concentration and rapidly falling O2 concentration would immediately trigger engine shut-down.

8417: “8417_c010” — 2007/9/11 — 12:15 — page 266 — #16

Carbon Monoxide Sensors and Systems

267

There were several advantages to this scheme: Low and high alarms give warning of CO presence at levels shown to impair psychometric performance (“drowse alarm”) and known to produce health damage in at-risk groups (congestive heart failure (CHF), coronary artery disease (CAD), fetus). These would give warning of elevated ambient CO and/or abnormal exhaust gas leaks into the motor-vehicle driver/passenger compartment. With concentration × time computer integration, neither heavy cigarette smoking, auto tunnels, nor congested roads would be likely to trigger even the low CO alarm. Use of CO2 concentration and O2 concentration changes along with CO concentration would prevent “false positives,” that is, inappropriate engine shut-down. Changes in the concentrations of these three gases would provide a unique “signature” of the suicide attempt. A second sensor array might be placed outside the motor-vehicle, preferrably near the rear tailpipe. This would cause engine shut-down in those instances where people attempt to commit suicide outside of the car, behind the tailpipe (in a garage, outside, etc.). If the cost of the three-sensor array proved too great, only one sensor responding to CO might instead be used. In this event, threshold CO concentration might be set somewhat higher in order to avoid inappropriate engine shut-downs. CO detectors are standard equipment in households in the USA, warning of furnace malfunction, etc. They are also required in motor homes, recreational powerboats, and other devices where people are fully or partially enclosed and in proximity to an internal combustion engine. Why shouldn’t such devices now become standard equipment in motor vehicles, considering that cars are such prodigious generators of CO and in such close proximity to the driver and passengers, and the fact that cars already incorporate minimally several microcomputers in their normal operation. For further details of the proposed scheme, see www.coheadquarters.com/CO1.htm.

References 1. C. Bernard, Introduction a l’etude de la medecine experimentale. Paris: J.B. Bailliere et Fils, 1865, pp. 85–92, 101–104, 107–112, 265–301. 2. Revised Standards to Improve Carbon Monoxide Alarm Performance, Gas Research Institute Digest (GRID), Chicago, vol. 21, pp. 24–25, 1998. 3. P. K. Clifford, Evaluating the Performance of Residential CO Alarms, Mosaic Industries, Inc., Newark GRI–02/0112, 2002. 4. K. Galatsis, W. Wlodarski, Y. X. Li, and K. Kalantar-zadeh, Vehicle cabin air quality monitor using gas sensors for improved safety, presented at COMMAD 2000 Proceedings. Conference on Optoelectronic and Microelectronic Materials and Devices (Cat. No. 00EX466). IEEE. 2000, pp. 65–68. Piscataway, NJ. 5. K. Galatsis, W. Wlodarski, L. Yongxiang, and K. Kalantar-zadeh, Ventilation control for improved cabin air quality and vehicle safety, presented at IEEE VTS 53rd Vehicular Technology Conference, Spring 2001. Proceedings (Cat. No. 01CH37202). IEEE. Part vol. 4, 2001, pp. 3018–3021, Piscataway, NJ. 6. K. Galatsis, B. Wells, and S. McDonald, Vehicle cabin air quality monitor for fatigue and suicide prevention, SAE Transactions, vol. 2000-01-0084, 2000. 7. S. Sato, R & D Review of Toyota CRDL 39, vol. 1, 36, 2004.

8417: “8417_c010” — 2007/9/11 — 12:15 — page 267 — #17

268

Carbon Monoxide Poisoning 8. V. H. Routley and J. Ozanne-Smith, The impact of catalytic converters on motor vehicle exhaust gas suicides, Med. J. Australia, vol. 168, pp. 65–67, 1998. 9. M. A. Skopek and R. Perkins, Deliberate exposure to motor vehicle exhaust gas: the psychosocial profile of attempted suicide. Australian and New Zealand J. Psychiatry, vol. 32, pp. 830–838, 1998. 10. Motor Vehicle Census, Australian Bureau of Statistics, Canberra Cat. 0309.0, 2001. 11. W. Gopel, New materials and transducers for chemical sensors, presented at Sensors and Actuators B (Chemical), vol. B18, no. 1–3, March, 1994, pp. 1–21. Switzerland. 12. N. Yamazoe, New approaches for improving semiconductor gas sensors, presented at Sensors and Actuators B (Chemical), vol. B5, no. 1–4, Aug.–Dec., 1991, pp. 7–19. Switzerland. 13. G. Sberveglieri, Recent developments in semiconducting thin-film gas sensors, presented at Sensors and Actuators B (Chemical), vol. B23, no. 2–3, Feb. 1995, pp. 103–109. Switzerland. 14. M. J. Madou and S. R. Morrison, Chemical Sensing with Solid State Devices: Academic Press, San Diego, 1989. 15. S. R. Morrison, In Semiconductor Sensors, S. M. Sze, ed., New York: J. Wiley, 1994. 16. H. Meixner, J. Gerblinger, and M. Fleischer, Sensors for monitoring environmental pollution, presented at Sensors and Actuators B (Chemical), vol. B15, no. 1–3, Aug., 1993, pp. 45–54. Switzerland. 17. G. Kiriakidis, N. Katsarakis, M. Katharakis, M. Suchea, K. Galatsis, W. Wlodarski, and D. Kotzias, Ultra sensitive low temperature metal oxide gas sensors, presented at 2004 International Semiconductor Conference. CAS 2004 Proceedings (IEEE Cat. No. 04TH8748). IEEE. Part vol. 2, 2004, pp. 325–331, vol. 2. Piscataway, NJ, USA. 18. K. Galatsis, Y. Li, W. Wlodarski, C. Cantalini, M. Passacantando, and S. Santucci, MoO3 , WO3 single and binary oxide prepared by sol-gel method for gas sensing applications, J. Sol-Gel Sci. Tech., vol. 26, pp. 1097–1101, 2003. 19. S. Kaciulis, L. Pandolfi, S. Viticoli, G. Sberveglieri, E. Zampiceni, W. Wlodarski, K. Galatsis, and Y.X. Li, Investigation of thin films of mixed oxides for gas-sensing applications, Surface and Interface Analysis, vol. 34, pp. 672–676, 2002. 20. L. M. Cukrov, P. G. McCormick, K. Galatsis, and W. Wlodarski, Microcharacterisation and gas sensing properties of mechanochemically processed nanosized iron-doped SnO2 , presented at Proceedings of IEEE Sensors 2002. First IEEE International Conference on Sensors (Cat. No. 02CH37394). IEEE. Part vol. 1, 2002, pp. 443–447, vol. 1, Piscataway, NJ. 21. Y. X. Li, D. Wang, Q. R. Yin, K. Galatsis, and W. Wlodarski, Microstructural characterization of sol-gel derived Ga2 O3 -TiO2 thin films for gas sensing,” presented at COMMAD 2000 Proceedings. Conference on Optoelectronic and Microelectronic Materials and Devices (Cat. No. 00EX466). IEEE. 2000, pp. 363–366. Piscataway, NJ. 22. Y. X. Li, K. Galatsis, W. Wlodarski, J. Cole, S. Russo, J. Gorman, N. Rockelmann, and C. Cantalini, Polycrystalline and amorphous sol-gel derived WO3 thin films and their gas sensing properties, presented at COMMAD 2000 Proceedings. Conference on Optoelectronic and Microelectronic Materials and Devices (Cat. No.00EX466). IEEE. 2000, pp. 206–209. Piscataway, NJ. 23. L. Yongxiang, W. Wlodarski, K. Galatsis, S. H. Moshli, J. Cole, S. Russo, and N. Rockelmann, Gas sensing properties of p-type semiconducting Cr-doped TiO2 thin films, presented at Transducers ’01. Eurosensors XV. 11th International Conference on Solid-State Sensors and Actuators. Digest of Technical Papers. Springer-Verlag. Part vol. 1, 2001, pp. 840–843, vol. 1. Berlin, Germany.

8417: “8417_c010” — 2007/9/11 — 12:15 — page 268 — #18

Carbon Monoxide Sensors and Systems

269

24. K. Galatsis, L. Yongxiang, W. Wlodarski, E. Comini, G. Sberveglieri, C. Cantalini, S. Santucci, and M. Passacantando, Comparison of single and binary oxide MoO3 , TiO2 and WO3 sol-gel gas sensors, presented at Transducers ’01. Eurosensors XV. 11th International Conference on Solid State Sensors and Actuators. Digest of Technical Papers. Springer-Verlag. Part vol. 1, 2001, pp. 836–839, vol. 1, Berlin, Germany. 25. For availability: www.coheadquarters.com/CO1.htm

8417: “8417_c010” — 2007/9/11 — 12:15 — page 269 — #19

8417: “8417_c010” — 2007/9/11 — 12:15 — page 270 — #20

11

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond: Use of the World Wide Web in Saving Lives Rob Aiers

CONTENTS 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8

My Introduction to CO Alarms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Department of Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Department of the Environment, Transport, and the Regions . . . . . . . The Department of Trade and Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Health and Safety Executive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

271 274 277 281 281 282 284 285

11.1 MY INTRODUCTION TO CO ALARMS I have been involved with carbon monoxide (CO)-related issues for 8 years. My work and indeed my passion for getting the message out has resulted in our website becoming the number one CO-related issues website based on Google searches. From time to time we swap places with David Penney’s website. My background is in marketing, although I am an aircraft engineer by trade, serving 13 years in the Royal Air Force (RAF) and culminating in the first Gulf War. On leaving the RAF, I completed a diploma in sales and marketing management. I have worked for a number of large organizations and have been consulting in my own marketing business for 5 years. 271

8417: “8417_c011” — 2007/9/21 — 22:04 — page 271 — #1

272

Carbon Monoxide Poisoning

My entre’ to the problem of CO exposure came when working for the world’s largest water utility company, as a consultant, with a remit for new and innovative products. My goal was to find a product that could be known outside the water company’s geographical area (water utilities in the United Kingdom cover a geographic area for their core business, but can own a number of such businesses in various areas) and thus make nonregulated profits (core business profits are regulated in the United Kingdom). The first product that I was asked to market was a product called Water Fuse. It was a water-related product that the water company took on before I joined. Although Water Fuse was a good product, it had a number of problems. It measured water usage and if a leak was detected, would shut off the water. We did a number of commercial trials for the product and found it to be very reliable, however that was not the whole story. We found that the cost of the product combined with the installation costs made it prohibitively expensive, around $550. Most people in the market for this type of product (i.e., the domestic user) were happy if they had a problem their insurance would cover. Indeed most people’s insurance premiums were less than the cost of Water Fuse and covered all elements of liability including fire, theft, and so forth. I set out to find a replacement for Water Fuse that would be cheaper and easier to install. It was at this time that we came across many wacky products. One that springs to mind is Eco-Balls. This was a product that one used in place of detergent. Its marketing blurb read, “unleash the ionic power of your washing machine.” The premise of this product was that when you put the balls in with your washer, the power of ions would clean your clothes. When tested in the lab, it was found to be only marginally more effective than water alone. On my travels I met with a company that had done a trial of a product that was similar to Water Fuse, but was much less expensive and easier to install. This product could also be expanded to cover small leaks in all parts of a property by utilizing the electrical ring main of the house to send a signal back to the master unit, turning off the water. At the end of the meeting I asked what other products they were working on. “We have a product that we think could be a major lifesaver, but it does not fit the water company profile.” The product it turns out was a CO detector. We spent more time talking about this product than about the Water Fuse replacement. I was amazed how little I knew about CO. I spent the next few weeks researching the topic and discovered I was by no means alone in my ignorance. I was further astounded at the conflict in the numbers of deaths and injuries from CO exposures/poisonings. The figures ranged from an official government statistic of around 50 deaths a year to as many as 1500 deaths per year from nonofficial sources. I was appalled that there wasn’t more information available and that the government did not and still doesn’t, recognize or promote CO safety. Every now and then, late at night on TV, there is a puny campaign about CO exposure consisting of a public relations commercial. I am quite cynical about why the government does not publicize the CO problem. I believe that if they made people more aware of CO, the government would be compelled to follow up with a plan to resolve the problem. The government in the

8417: “8417_c011” — 2007/9/21 — 22:04 — page 272 — #2

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

273

UK owns about 2.5 million homes, mostly “social housing.” If it publicized the problem, it would have to protect the residents from danger—the costs would be great. A CO alarm costs around $20.00. Multiply that by 2.5 million homes and the total is roughly $50 million. I saw the CO alarm to be a perfect public relations vehicle for a water company. It could be a true win-win situation. The water company had around 11 million customers and had access to the rest of the 60 million UK population through other water utilities. They could publicize the dangers of CO and also sell a product that could protect its users from future exposure. This might drastically reduce deaths. The company who first had the CO alarm was small, but it was innovative. It also had all the problems inherent to small companies, namely cash on hand. The product had been prototyped, with a short manufacturing run in Taiwan. However, the company had no money to move into full production. We concluded a deal where we had exclusive access to the product. The CO alarm was inserted directly into the wall electrical power receptacle. It incorporated some clever technology that allowed it to predict carboxyhemoglobin (COHb) levels, that is, blood CO content. We agreed to pay for an increase in product run once we had seen the product being built in a factory in Taiwan and were convinced that all was okay. Some colleagues and myself flew out to Taipei, Taiwan to verify the situation. This took a total of 5 days, either on the road or at manufacturing facilities. At every location we were dressed in full medical whites (“clean garb”), as the facilities also produced computer and automotive industry components. My view at that time was that Taiwan was just a producer of cheap toys for Christmas, crackers, and so forth. Nothing was further from the truth. All the facilities we saw were high-tech—at the top end of the computer industry. The costs of the CO alarms were a problem. If the selling price was too high, we’d be unable to sell many units. Not only would we not make projected profits, but more importantly, people would not be able to protect themselves from CO poisoning. Clearly the profit element was important and getting unit costs down was vitally important. We had to have the best product at the best price, so that when we advertised, we had an affordable solution for all. Ideally I would have liked to get the product down to a price of around £10 ($17). That is about the same price as a smoke alarm. We soon realized that the CO alarm was a much more complicated product and it may never be possible to make it as cheap as a smoke alarm. The smoke alarm industry had become a mature market, unlike that for CO alarms. It is worth drawing some parallels with what had allowed the smoke alarm industry to gain so much ground. In the 1980s, UK market penetration for smoke alarms was around 9% of households. At that time the government mounted a big campaign through the fire service and TV advertising. The “Smoke kills” campaign was responsible within a 9-year period of raising market penetration to around 65%. Recently, government legislation has caused that penetration to be near 100%. By law now, all new houses must have a smoke detector hard-wired into the construction, and it is now considered the norm to have such a device in place. The smoke kill’s campaign still goes on, but the focus is now more on maintaining existing devices. I think that the world would be a safer place if the same regulations existed for CO alarms.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 273 — #3

Carbon Monoxide Poisoning

274

We had to look for manufacturing facilities that would get CO alarms to market at a price that would ensure that most people would buy them. Many of the Taiwanese factories also had Chinese manufacturing facilities. These involve much reduced cost, since labor costs in the Republic of China are less than that of Taiwan. This would have been the next move had not something happened. It became clear that the price that I was trying to achieve was not realistic and that CO alarms would never be as cheap as smoke alarms. As stated above, this is because smoke alarms detect danger as the danger is occurring (by either optical sensors or ion detection), but do not require the complex circuitry required in a CO alarm. It is possible to have a large release of CO that causes an immediate audible alarm, or to have much lower levels of CO released over prolonged periods of time. The best type of CO alarm is one that can make the correct calculations regarding human CO uptake. As stated above, I also wanted to sell the best alarm, so it had to be a unit that measured CO concentration over time. Inexpensive card-type detectors are available that change color in the presence of CO, but these can be contaminated by chemicals, don’t give satisfactory low-level indications, and have no audible alarm. I am not totally opposed to this type of CO detector, as they play a role in raising CO awareness. Nonetheless, they are not sufficiently reliable for constant usage. We had by this stage fully evaluated the CO alarm market and were excited to get on with marketing the product and making the public aware that there was a problem. We hit on a major problem at this time—the water company was likely to be sold. This was happening at a high level in the company so none of us was aware of the impending sale. Orders came down from on high that we were to stop all unnecessary marketing and to get back to the core business. My vice-president at that time was a very talented woman who was not held in high regard by the company board. There followed a free for all with junior managers stepping on each others heads to try to gain political high ground. In particular, while I was on leave, an alliance was formed between two managers who were determined to steal our thunder, or better still, to totally kill our work. In the end they won and all of the people who had been working for me were let go or moved to other departments. With that, all of our efforts were dashed. I then left the company to move on to other things, but the company had no idea what contracts had been signed. Consequently a legal dispute ensued, because contracts had been signed for the water company (by me) that tied them to marketing and supply agreements. The company with whom we had been dealing needed the agreements in order to move forward, but foolishly had signed an agreement with a third party to sell the product before the legal wrangle was resolved and they ultimately lost everything.

11.2 USE OF THE INTERNET I resumed my marketing business at this time, after working initially on a consulting basis for the water utility. I worked for a number of large clients, but also felt the government could and should publicize the CO issue. It was around this time that I spawned the idea of a CO information website, that is, www.carbonmonoxidekills.com. I immediately began getting it set up.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 274 — #4

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

275

Unfortunately I knew nothing about the internet, so I needed to team up with someone who was familiar with its complexity. It was at this time I met Frank Tiarks, a Danish chap who had been educated in the UK. Frank had a very good understanding of what was required to make website content user friendly, and of the content necessary to cause the major web search engines to list it first. I soon learned that it is no use just having a website—it has to be seen. I would liken it to having a shop that sells great products that people really need, but the shop is located down a dead-end road or in the depths of the countryside where no one will find it. One of the most important factors for success here is having a “Frank” who knows how to do it. There are many reasons why the world wide web (WWW) is an excellent way of getting your message out. First, the WWW is democratic. You have the same chance of getting your message across to the public if you are a small corner toy shop, as Toys-R-Us does. From that perspective it didn’t matter where you were geographically located as long as you could speak English. Anyone who has visited our website and has acted on the information there owes a debt to Frank Tiarks who has worked tirelessly to get CO-related information to the public. We needed to have a website with immediate impact, so we came up with the name, www.carbonmonoxidekills.com. As a capitalist, I had to generate a profit from the website simply to be able to continue running it. Over the past 5–6 years, some $80,000 has been spent on it, of which only a small percentage has been recouped. I am proud of the fact that we have become one of the most highly utilized, web-based CO information sites in the world. Approximately 5500 people per day look at the information on www.carbonmonoxidekills.com (see Figure 11.1). If we have saved one life or helped one sufferer improve his life, it has been money well spent. Some of the respondents have been referred to Dr. Penney for evaluation. More recently we have added some law firms as consultants, so some of our viewers have been helped with their legal problems. As you might guess, the website is quite comprehensive. Where questions are not answered, we pass the visitor on to an expert to answer the question (see Figure 11.2). It should also be noted that what you can see here is only the front page of the website. There are many additional layers, so almost any subject relating to CO poisoning can be found. It took some time to get us where we are now in terms of a web presence. We constantly re-examine the content of the website to insure that it has the best information for our visitors. We now get around 2 million visitors a year, many of whom e-mail us to thank us for the website. It is a sad indictment of those official government sources that should be doing this job and not be relying on individuals like me to do the job for them. Unfortunately we only get visitors who already are aware that there is a problem involving CO. In the words of Donald Rumsfeld, “we don’t know what we don’t know.” A clumsy way of saying that is, if you have never heard of CO, why would you look for information about it. That is the main issue here. If governments would just get the word out, people would be better able to decide for themselves whether there is a problem. The United States is much better at this than the United Kingdom. It may have to do with the fact that America is a more litigious society. We do not have a punitive damages system in the United Kingdom, so where someone may get an award of

8417: “8417_c011” — 2007/9/21 — 22:04 — page 275 — #5

276

Carbon Monoxide Poisoning

FIGURE 11.1 The front page of www.carbonmonoxidekills.com

several hundred-thousand dollars in the United States, in the United Kingdom the award is more likely to be £10,000 at best. The questions we are asked through the website are diverse to say the least. In some cases they are downright stupid. In one case an American youth asked me if it was dangerous to travel in the trunk of his friend’s car. The great bulk of the questions from an uninformed public are completely justified (see Figure 11.3). About 75% of the website’s visitors are American—15% are from the United Kingdom (we have a lot of UK links from other websites, especially government sites), with the remaining visitors coming from the rest of the world. The reasons for this are several-fold. First, the United States has a larger population than the United Kingdom. Second, America is more web-savvy than most of the world. However the most important reason is that Americans are more aware of the dangers of CO, and are thus better able to seek out information. It should be noted that our website is in English alone, so other language speakers may not be able to use it. While there are charities and experts who work hard to get the message out, the same cannot be said of everyone. People with vested interests in CO issues have not,

8417: “8417_c011” — 2007/9/21 — 22:04 — page 276 — #6

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

277

FIGURE 11.2 The law connection to the Web.

in my view, been ready to sing from the same song sheet. There are three or four government departments in the United Kingdom who have some responsibility with regard to CO. From a public health perspective there is the Department of Health, from an environmental point of view (i.e., housing, etc) there is the Department of the Environment, and from a business perspective there is the Department of Trade and Industry. There is also the Health and Safety Executive who looks at the industrial side of safety.

11.3 DEPARTMENT OF HEALTH This department’s charge is the state of the nation’s health, with its remit covering doctors, nurses, and the general health infrastructure. Health concerns cover every part of people’s lives and some areas are covered better than others. Let me draw a parallel if I may. Meningitis affects about the same number of people in the United Kingdom as those who are affected by CO, based on government statistics. It is a

8417: “8417_c011” — 2007/9/21 — 22:04 — page 277 — #7

278

Carbon Monoxide Poisoning

FIGURE 11.3 The forum associated with www.carbonmonoxidekills.com.

very serious disease that can kill or maim. Unchecked, CO poisoning can also kill, or damage people’s lives either through an acute poisoning episode or through chronic exposure. You would be hard pressed to find a parent in the United Kingdom who was not aware of the symptoms of meningitis and what to do if he/she suspected his/her child was suffering from it. This is just as it should be and is so because the government and the media have spent a good deal of time and effort publicizing its dangers. There was scarcely a time 3 years ago when you turned on the TV that this issue was not being aired. Please excuse me if I shout, “CO EXPOSURE IS AS DANGEROUS AS THE MANY STRAINS OF BACTERIA AND VIRUS THAT CAUSE MENINGITIS.” It is about time that politicians and health experts in the United Kingdom took a lead from some of the US states and began taking the CO exposure/poisoning issue more seriously. It is my opinion that most medical practitioners here and in the United States are only slightly better than useless when it comes to diagnosing and treating their

8417: “8417_c011” — 2007/9/21 — 22:04 — page 278 — #8

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

279

patients exposed to CO. If you visit your physician with flu-related symptoms, CO poisoning is usually the last cause he/she will include in the differential diagnosis, if it is included at all. Physicians are poorly trained about the dangers of CO poisoning— after all, the concern is primarily toxicological, not internal medicine. Physicians generally don’t know the correct questions to ask. For example, they usually fail to ask how your house is heated? Whether several people and even animals are sick at the residence? I very much doubt it. However, this is not an excuse for complacency. We must train our medical personnel better to recognize the symptoms of CO exposure. There are few ways for a physician to investigate the living circumstances of the person who has presented for treatment. That patient could be residing in a property where someone previously suffered similar symptoms. It is possible that over an extended period of time many people suffered CO poisoning in that building because no medical professional was alert. This scenario could go on for years. Early in 2006 I had a meeting at the Department of Health with a number of people. This included a lawyer who had represented victim’s interests in CO cases that had gone to court in the United Kingdom. There were a couple of representatives from a CO charity and the head of the Health Protection Agency, whose responsibility it is to deal with preventive health measures. I found the way the meeting was conducted to be astounding. We were informed that a leaflet had been sent to all UK physicians regarding CO. The leaflet was fairly brief. We asked if the department had followed up the leaflet to make sure that physicians had received it, and whether they had read and understood the information. The answer was no. We were also told that if there was a high demand for the leaflet there would be a reprinting, and that it was also available on the Department of Health’s website. If the physicians had not read the information in the first place how was there going to be a high demand? I told department personnel that I would help them in whatever way I could to get the message out, but that they needed to help me. One of the ways that this could be done quite simply and at very little cost, was for them to call TV stations to alert them to the problem. In the United Kingdom we have a morning program called GMTV. They spearheaded a brilliant campaign for meningitis and could do the same for CO. I have been unsuccessful in getting through to them. However one call from the chief medical officer would spur them into action at the cost of a phone call. At the meeting, a vice-president of CO Awareness said that she had tested the department’s system by calling “NHS Direct” (a phone hot line that gives health advice). She was told that the best people to speak with about CO was CO Awareness, her own charity. When questioned about the computer menu on their help line which advises what should be done when certain syptoms are mentioned, the official said that unexplained headaches and drowsiness would NOT be related to CO exposure by the operator. Of course this is incorrect. The problem goes much deeper. I asked the chief medical officer what would be the recommendation from physicians should they suspect CO poisoning? He replied that “the patient should be tested.” Do they have the appropriate equipment I asked? “I’ll have to get back to you,” he said. “So let’s accept the fact that they have the relevant testing equipment,” I said. “When do you recommend that the patient be tested?” “Well, during their consultation” was his reply. I had to point out that if the

8417: “8417_c011” — 2007/9/21 — 22:04 — page 279 — #9

280

Carbon Monoxide Poisoning

FIGURE 11.4 www.codetection.com, which is lined to form carbonmonoxidekills.com

patient had been away from the source of the CO poisoning for more than a few hours, then they would likely give a negative test. In this event, someone who was suffering an episode of CO poisoning, but who happened to get an appointment at the end of the day might have CO dismissed as a potential cause of his/her illness, and then might go on being exposing needlessly to a life threatening situation. In some ways this is much worse than not being diagnosed at all. This problem was dismissed almost out of hand, likely because I was not a medical professional. The Department of Health should in my view be recommending that all homes be fitted with CO alarms (see Figure 11.4). There are a number of different views about CO alarms and their efficacy, but I think they are a good weapon in the arsenal for combating CO poisoning. They should also be recommending that all combustion appliances be checked annually. Only landlords have to perform an annual safety check. I am sorry to be bashing the Department of Health, because on the whole they do a great job. They are often hog-tied by a lack of government funding.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 280 — #10

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

281

11.4 THE DEPARTMENT OF THE ENVIRONMENT, TRANSPORT, AND THE REGIONS This department is responsible for a numbers of areas. It is concerned with amongst other things, the air we breathe. Many people think that CO is only caused by vehicular emissions. There are major crossovers with other departments and I am loathe to list the full responsibility of this department, since by the time this book goes to print, it may have changed or there may be a new administration. This Department is responsible for dangers that lurk in our homes. It is responsible for planning regulations and for new building laws, for example, requiring newly built homes to have hard-wired smoke detectors. In my view, it is remiss in this respect in not also requiring CO detectors. The Department of the Environment requires landlords to have a safety check each year of accommodations where fossil or wood burning appliances are used. Because many people in the United Kingdom own there own homes, it means that a large fraction of residences are not covered by any regulation. It is unfortunately the way of the world that we buy “stuff” for which we can see a direct benefit. A CO detector costs about the same as a take-away meal, and at the end of that meal we are happy to part with our cash. Because it costs a bit more to get our furnace/boiler serviced (i.e., $100/200), many of us might shy away from the cost. It usually doesn’t make us feel good—in fact quite the reverse. We will only pay this charge if compelled to do so. We do not like paying our house or car insurance, but do it because we are compelled to. New York has the right idea in that landlords must by law install a CO alarm, and in turn can collect $25 from each tenant to help pay for it. Do yourself a favor. Purchase a recommended CO alarm to protect yourself and your family. Better yet, buy two CO alarms. By its nature CO is colorless, odorless, and has no taste. It does not occur to us that there will ever be a problem. You would not leave small children near a staircase lacking a gate to prevent them from falling.

11.5 THE DEPARTMENT OF TRADE AND INDUSTRY This department deals with the interests of trade, but also regulates some of the peripheral issues related to CO. One of the main areas of authority regarding CO are the gas supply companies, which until recently was reduced to just one supplier, “British Gas.” British Gas lost its monopoly and now there are many suppliers in the market place. These companies came into being as a result of denationalization, a trend in the United Kingdom. The majority of gas users still use British Gas. Some credit has to be given to British Gas for recent initiatives, but they are not on the whole completely altruistic in their nature. I will paraphrase a conversation that I had with one senior British Gas executive. He said that “if we publicize the fact that gas could be dangerous and that in certain circumstances gas appliances will give off CO, then we denigrate our product.” This is a bit like saying that we should not install seat belts in automobiles because it makes them appear unsafe. From a trade point of view they are missing the point. Their product is not the problem, it is what is done with that product when it leaves the pipeline that is key.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 281 — #11

Carbon Monoxide Poisoning

282

British Gas now sells CO detectors. In fact, they are the biggest UK distributor for the SF-350 CO detector, the best selling unit in Europe. Most of these have been given away with service, which is sold on a commercial basis. British Gas has a large service division that must make a profit. In my opinion, if they played the white knight and told people that there was a major problem, they would be seen as a responsible company with their customer’s interest at heart. They would benefit from their customers coming to them and asking for service. Also, people would in turn buy the CO detectors, and not expect them to be free. The company would not be held up as irresponsible as they are now by some groups for not publicizing the CO danger. In this way, they could benefit financially from making CO a safety and health issue.

11.6 THE HEALTH AND SAFETY EXECUTIVE This department, HSE, is responsible primarily for safety. I recently attended a meeting of their Gas Safety Committee. I had been invited by the head of a leading charity to present my views to the committee. A number of things came up in that meeting that I found astounding and frustrating. The Department had a £100,000 surplus at the end of the financial year which they had to spend quickly, otherwise it would disappear from the following year’s budget. They didn’t ask the charities who endeavor to get the message about the best course of action. They didn’t call on our firm for suggestions. Instead they called CORGI. CORGI is the regulatory body for heating engineers. You cannot work on gas appliances in the United Kingdom unless you are qualified and a member of CORGI. CORGI’s suggestion was to give it to them and they would do a survey; so they did. The HSE could not provide us with a qualitative or quantitative evaluation of what they were going to get for their money. I guess they will get a survey back which will form part of the departments future strategy that says, “CORGI are really good guys and are in the opinion of the people who where surveyed, a top organization?” The HSE Department has tried to develop a cohesive plan to alert people to CO dangers. They have not called on me. I have on occasion called them, but they seem to have no clue as to what is going on in the world outside government. An independent enquiry stated in 2000 that it would be desirable to impose a fuel levy on gas, oil, and coal of about 0.25%, which would be used to fund CO awareness publicity. The HSE has now said that they will not implement the tax after being lobbied hard by the gas companies. They can see this going the way of all taxes, that is, once introduced, the taxes will only go up. Consequently, they don’t want even a small tax that they could absorb. Instead they have said to the HSE, “We have spent £2 million this year on CO awareness.” In actual fact, they had a campaign about service that was high profile, with many UK celebrities in ads. Unfortunately the ads never mentioned CO. The ads were mainly about getting the British public to buy their service—service which I might add, will go on making millions for years because it is annual service. The net effect of this is that less people will be harmed by CO, but don’t for a minute think that they have done it for purely altruistic reasons, they want your buck.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 282 — #12

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

283

All of these government departments have budgets which in and of themselves are not really that much, maybe a couple hundred thousand pounds each. If however, they combined their resources, they could do some real good. Also, if they had one team looking at this rather than many disparate groups, they would not be duplicating effort and could reduce spending. The effort could also be better coordinated so that we have something tangible to show for it. People would readily be able to see the dangers immediately, not months or years later. People would also be more aware about protecting themselves in the first place. In the last few months I have been to meetings at the House of Lords and subsequently with Lord McKenzie who is tasked with getting all the information on CO together in one place with one agency responsible. It seemed at last that there was some momentum, however that was some months ago and I have heard nothing since. We have created a website brand www.carbonmonoxidekills.com which is well recognized throughout the English-speaking world. Since I am a capitalist, our website is linked to a number of others. I submit that most problems in the world could be resolved by entrepreneur’s looking at them as opportunities to earn revenue. The more money that we make in this venture, the more we are able to plow it into CO awareness. It’s a win-win situation. Also, the more we get out there, the more information that we acquire, the more we can educate people about the issue. Dr. Penney asked me to write this chapter partly because we have a relationship going back a few years. He had kindly agreed to be our online “doctor” some years ago. As such, he responds to enquiries regarding CO from the public. We feel that he is an integral part of our online organization. We also have relationships with law firms who are able to answer legal questions. Therefore, we are able to get the best representation for people who have suffered CO poisoning episodes. As mentioned above, we sell CO detectors through another link on our website, that is, www.carbonmonoxidekills.com. This has been hard work and it has taken a great deal of effort to get where we are. People often contact me, asking how they can set up there own website. They are more than welcome to try, but we are the experts and that is why over the years we have arrived at the position that we hold. To achieve this, you have to get as much information about your subject on the website. That information needs to be updated regularly so that the search engine spiders can recognize new information. This keeps the rankings up. A spider is a computer routine that runs on search engines like Google, Yahoo, and so forth. Working for a high ranking is a bit of a catch 22 situation. Unless you have a good ranking, you can’t get one. Another factor in terms of search engine rankings, is how many other websites link to yours and the type of organizations they are. The best types of organizations are government departments, the next are educational establishments such as universities, and so forth. Media institutions are also good in that they attract lots of visitors on a given subject such as CO that appears in the news. Thus, you will get renewed interest at the time of newsworthy items. One has to be very careful when approaching these kinds of websites. They will not link to you if your information is commercial or inflammatory. Many charities suffer for the latter reason, because they will often hold someone accountable for an incident making them on occasion being seen as biased and having a narrow manifesto. No government department will link to them. It is also sometimes difficult to get

8417: “8417_c011” — 2007/9/21 — 22:04 — page 283 — #13

Carbon Monoxide Poisoning

284

a government body—and most educational institutions are government influenced— to link to you unless another government institution has linked first. No one wants to be first to link because they are scared that it will impact their impartiality and position. The first one to come on board is the hardest to get. If I call an institution now and ask them to link to us, they are only too willing because we have 100s of other links. There is now a comfort factor for them. There has not been a willingness in the United Kingdom to grasp the nettle and do something about the issue of CO. Proper unbiased research need to be done, and we need to have a better understanding of the issue. Statistics on CO need to be consistent. Some of the country’s top toxicologists state numbers that differ greatly from the government’s official figures. I mentioned above that most physicians are ill-prepared to diagnose and treat CO poisoning. Moreover, if a physician has not identified the connection to CO, then people may never know that they had CO-exposure. They may have relocated from their house, their job, left home, etc. They may exhibit effects of CO in later life that will never be attributed to CO. Death certificates may not tell the whole story either for very similar reasons. The physician may list just one of the symptoms as the cause of death. The coroner’s office in the United Kingdom is not attached to the Department of Health—it is part of the Home Office. I was amazed to learn that until recently, the job of coroner was a part time job. I am glad to say that we are now going to get a full-time chief coroner. This may clarify the position and he/she will liaise closer with the Health Department. In short, we need to know who and how many people are affected. Only then can the problem be properly addressed. In Paris, France all cadavers are tested for poisoning of all types. As a result, more people are recorded to have died from CO than in all of the United Kingdom. This is probably not because Paris has a worse problem, but just that better monitoring is in place. Take for example the murderer Harold Shipman, who was a UK general practitioner. No one is quite sure how many people he killed. His story has been recognized as one shortcoming of the coroner’s office. As a result, certain remedial measures were suggested. The one I am most interested to see implemented is the checking of all bodies for poisons, including CO, as a matter of course. This could prove a real breakthrough and give us proper statistics so that the problem will be highlighted and solutions found.

11.7 CONCLUSIONS There is much work to be done. Combining the budgets of disparate government departments, along with some coordinated thinking would result in a better understanding. I have offered government departments use of our website facility. If they wish, we are able to run studies and research at little cost through our valued visitor stream. We are very well-placed to be able to assist. While recognized for the efforts that we have made, we are never seen as part of the solution. My ambition is that www.carbonmonoxidekills.com will be the world-wide marketing tool for all CO-related issues. I want to make myself redundant, in as much as

8417: “8417_c011” — 2007/9/21 — 22:04 — page 284 — #14

Marketing of Carbon Monoxide Information and Alarms in Europe and Beyond

285

we have done the job so well, there is no more need for publicity because the whole world has got the message and deaths and injury from CO are a thing of the past. I would like people to know as much about CO as about meningitis. If the government picked up the ball and ran with it, the issue of CO may actually pay for itself. Get the message out and people will be empowered to protect themselves. People would not be taking up the valuable time of physicians with mysterious illnesses or have to be hospitalized. I make no apologies for the fact that this chapter lacks high academic content. Instead, I hope you will appreciate the journey that we at www.carbonmonoxidekills.com have made over the last few years. We have tried to be the most open information platform possible, and not an exclusive high brow intellectual website that turns visitors off, indeed those that need us most. It has sometimes been tough, and I have thought about giving it up at times. It takes a lot of my time. Then I get an e-mail from a distraught mother, father or other relative or friend, that thanks us for our help and it all seems worthwhile. It is sad to say that unless governments do more, we are the only outlet for some people. I only wish that we could get through to more people, but language barriers make that difficult. It would be difficult for us to translate our website into other languages, such as Chinese. The paradox is that as the markets in the developing world get bigger, it is they who will need it most.

11.8 APPENDIX 2 Carbon Monoxide Headquarters known as COHQ, or “coheadquarters.com,” is one of the oldest carbon monoxide (CO) information sites on the web. It was started in 1996, so is now more than 11 years old. That is very old in terms of the history of the web! Initially COHQ ran on an old Macintosh computer in my research laboratory at the Medical School at Wayne State University in Detroit, MI. It had a long URL because it did not have its own registered domain name. In the late 1990s the website received its own domain name and I migrated it over to a commercial server in Texas. The new URL and the one used today is “coheadquarters.com/CO1.htm.” From a collection of a few dozen linked pages in the early days, COHQ has grown to a site containing hundreds of pages dealing with many subtopics in the field of CO toxicology: chronic CO poisoning, dangers to high risk groups, neuropsychological effects, FAQs, CO alarms/detectors, and so forth. COHQ was begun as a way to provide the public with straightforward, unbiased, information about CO and its effects on humans. It has been my operation from the beginning. With the exception of a few dozen pages written at first by my student Amy Derusha, all of the content, architectural design, art (whether good or bad), and maintenance was been done by me. The website was produced by writing HTML in “text” by hand—it never involved using a web-editor. As crude and lacking in flash as COHQ is, it still comes up in the top 15 sites in doing a Google search using the terms “carbon monoxide poisoning.” The goals of COHQ were: (1) To act as a platform for public information about CO, (2) To act as an educational resource for all viewers, including medical professionals,

8417: “8417_c011” — 2007/9/21 — 22:04 — page 285 — #15

286

Carbon Monoxide Poisoning

(3) To act as a data resource—“look-up” and retrieval of information, and (4) To present new ideas and information at the forefront of research in the field. Like other good expert sites, it (1) Provides reliable, in-depth information in a narrow, defined span of knowledge, (2) Is accessible, understandable, and useful to the lay public as well as professionals in field, (3) Remains current as the field progresses and as viewers needs change, and (4) Is unbiased; that is not supported or influenced by a commercial interest in the field. The web was a very different place a decade ago. The number of websites was miniscule as compared to today. Google was not yet operating. “My Space” was years away. There were search engines running there, but they didn’t have the power and speed of today’s. It is hard to believe that in the middle 1990s it was difficult for us to even find sources of instruction for the HTML language. Much of the early writing was hit or miss—whatever worked was used. The objective of COHQ was never to sell new furnaces, to provide furnace maintenance, and so forth as so many thousands or tens of thousands of websites found by a “carbon monoxide” search pulls up today. I have been accused of using the site as an advertising vehicle. Any good information source, whether a book, storefront, or website inadvertently advertises the author/owner when people go there. Nonetheless, unlike most CO sites, COHQ has no axe to grind. It provides both simple and technical information free about CO to whoever wishes to look at it. The constant theme is public health, for groups and for individuals—educating and protecting people from this age-old poison [see other chapters in this book on misconceptions about CO (Chapter14) and on public perceptions of CO (Chapter15)]. The moto on the home page reads, “CARBON MONOXIDE HEADQUARTERS, (to) provide information, public service, help people, maintain health, save lives.” I was pleased when “carbonmonoxidekills.com” came along in the late 1990s as another major CO information source, in this instance, emanating from the United Kingdom. As you have now read, its major objective is to safe-guard everyone by everyone having CO alarms. Carbonmonoxidekills.com is currently number 1 when the same search noted above for COHQ is run on Google. So I was thrilled some years ago, when Rob Aiers, the owner, asked me to answer questions on CO from his website. I have continued in that role to this day. Thus, it was natural when this book was being planned that I would ask Mr. Aiers to tell us about his experience in developing a very successful website similar to COHQ.

8417: “8417_c011” — 2007/9/21 — 22:04 — page 286 — #16

12

Investigating Carbon Monoxide Poisonings Thomas M. Dydek

CONTENTS 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Signs and Symptoms of Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . 12.3 Assessments of Carbon Monoxide Exposure Level and Duration. . . . . . . . 12.3.1 Carboxyhemoglobin Levels as a Measure of Carbon Monoxide Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.2 Occupational Exposure Standards for Carbon Monoxide . . . . . . . . 12.3.3 Community Exposure Standards and Guidelines for Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.4 Carbon Monoxide Exposure Duration Assessments . . . . . . . . . . . . . . 12.4 Treatments for Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.5 Other Factors to Consider in Investigations of Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.6 Case Study of Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

287 288 289 290 291 291 292 295 296 297 299 300

12.1 INTRODUCTION Carbon monoxide (CO) is a colorless, odorless, tasteless, and potentially toxic gas. These properties have earned it the title of “the silent killer”.1 CO poisoning is responsible for more than one half of the poisoning fatalities reported in this country every year. It is the leading cause of death in industrial accidents as well. Fatalities and CO-related injuries are also common throughout the world. Another factor that makes CO an especially dangerous toxin is that the early symptoms of poisoning are easily confused (and often misdiagnosed) as the onset of a cold or the flu, stomach virus, or other common diseases.2 CO is produced by natural sources and by man-made sources. Natural sources include forest fires, oxidation of nonmethane hydrocarbons, and oxidation of methane. Plants can also emit CO as a metabolic by-product.3 Anthropogenic sources of CO are mostly associated with incomplete combustion of organic materials such as

287

8417: “8417_c012” — 2007/9/11 — 12:15 — page 287 — #1

Carbon Monoxide Poisoning

288

gasoline, fuel oil, natural gas, wood, or plastics. Common man-made sources of CO include improperly vented cooking or heating devices, tobacco smoke, agricultural burning, and internal combustion engines. Exposure to CO is one of the chief dangers associated with the fighting of fires in buildings and forest fires. Workers in occupations in which routine exposures to CO occur include truck and bus drivers, mechanics, highway toll takers, garage attendants, and police officers. Everyone who drives a car or truck, especially in areas of congested traffic, has some exposure to CO.4 This chapter covers topics related to the investigation of CO-poisoning events. These investigations can occur in a number of different situations. Oftentimes these investigations are undertaken as part of legal proceedings. Many people have commented that we in the United States today are the most litigious society in the history of the world. While this may be a somewhat dramatic statement, there is a large element of truth in it. Many people in this country feel that pursuing legal remedies for perceived injuries which they believe they have suffered should be a first, rather than a last, resort. Expert witnesses in engineering, industrial hygiene, and toxicology are often called to investigate CO-poisoning incidents and to render opinions about whether or not the CO exposure reported was of at a sufficient level or duration to have caused the health harm alleged. The topics covered in this chapter include signs and symptoms of CO poisoning, exposure level and duration assessments, treatments for CO poisoning, differences in susceptibility between people, how the above factors affect the acute effects of exposure, and the long-term prognosis for CO victims, and other toxic exposures or conditions that mimic CO toxicity (differential analysis). All of these factors are important in the investigation of CO poisonings.

12.2 SIGNS AND SYMPTOMS OF CARBON MONOXIDE POISONING CO poisoning results in a decreased level of oxygen in the body. The brain and other parts of the central nervous system are the areas of the body which are among the most sensitive to oxygen lack.5 When oxygen levels in tissues fall, aerobic metabolism decreases and lactic acid accumulates. Neurons begin to break down, leading to cell death and brain damage.6 The longer the brain is deprived of adequate oxygen, the more widespread the damage will be. Similar effects occur in muscle tissues deprived of oxygen. This is of special concern when the muscle involved is in the heart. For many years most toxicologists believed that COs toxicity was fully explained by this hypoxic effect. More recent research has shown, however, that CO exerts direct toxic effects by inhibiting the activity of cytochrome a3 oxidase and by causing lipid peroxidation. These latter findings help to explain the clinical experience that carboxyhemoglobin (COHb) levels (an indicator of the risk of tissue hypoxia) are a very poor predictor of a patient’s medical condition and his or her prognosis.2,5,7

8417: “8417_c012” — 2007/9/11 — 12:15 — page 288 — #2

Investigating Carbon Monoxide Poisonings

289

The acute symptoms seen with CO-poisoning depend on the concentration of CO and the duration of the exposure. At low levels of exposure there may be subtle changes in time discrimination, visual vigilance, and choice response. Exposure to higher levels will aggravate preexisting angina pectoris. Symptoms seen in people with higher level CO exposure include severe headache, dizziness, nausea, vomiting, mental confusion, visual disturbances, reddening of the skin (not always), compartment syndrome, loss of muscle tissue, fatigue, hypotension, and coma. Severe exposure can of course be fatal.8 The COHb levels measured in the CO-poisoned patient are sometimes used to classify CO exposures in terms of mild, moderate, or severe poisonings. In this system, COHb levels of less than 30% are termed “mild” poisonings. “Moderate” poisonings are those in which the victim has a COHb level of from 30% to 40% and “severe” poisonings occur when COHb levels are greater than 40% . This construct, while sometimes useful at a simplistic level,7 should be used with great caution since other signs and symptoms are known to be far more important in determining how and when to treat a patient poisoned by CO. Depending on the level and duration of CO exposure, delayed or prolonged symptoms can also occur in CO-exposed individuals. Some of these effects can be severe and can last for years. These symptoms can include sleep disturbances, vision problems, hearing loss, tinnitus, peripheral neuropath, mental deficits, memory problems, and difficulty concentrating, to name but a few. While there is often some recovery of mental function in CO poisoned individuals over time, many brain tissues damaged by CO show little, if any capability for regeneration. Similarly, the neurological or other damage done in cases of compartment syndrome is not generally reversible. Therefore, many of the chronic medical conditions brought on by severe CO poisonings are likely to be permanent.5,9,10

12.3 ASSESSMENTS OF CARBON MONOXIDE EXPOSURE LEVEL AND DURATION In any poisoning case the investigator must try to determine the amount of exposure an individual has had. The cornerstone of toxicology is the “dose makes the poison.” Knowledge of the “dose” as reflected by the exposure a person has experienced helps to assess the potential for adverse health effects the person may exhibit, guides treatment that the individual will require, assists in assessing the patient’s follow-up care needs, and determines his/her prognosis. In the legal arena, a large part of job of the expert witness as a poisoning incident investigator is to determine what the exposure level was. Exposure assessments are also required in the workplace to ascertain whether or not occupational exposure standards were exceeded. Community exposure standards and guidelines have also been established by federal and state environmental agencies and by private organizations. Monitoring ambient air levels of CO and keeping outdoor CO levels below applicable standards and guidelines is one of the jobs of federal and state air pollution control agencies.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 289 — #3

290

Carbon Monoxide Poisoning

12.3.1 CARBOXYHEMOGLOBIN LEVELS AS A MEASURE OF CARBON MONOXIDE EXPOSURE When measurements of the CO level in the air have not been made or if made are not representative of the actual CO exposure levels a poisoning victim has experienced, other methods have been used to estimate CO exposure levels. One such method is to rely on COHb levels in the victim’s blood. When CO is inhaled, it is readily absorbed into the blood. Once there, the vast majority of the CO binds to hemoglobin in the erythrocytes to form COHb. CO binds to hemoglobin with an affinity more than 200 times that of oxygen. Thus CO displaces the oxygen from the hemoglobin, impairing oxygen delivery to critical body tissues such as the central nervous system, heart, and other organs.5 COHb levels in people exposed to CO reach a peak or plateau if the exposure lasts for 5–10 h. Further exposure to CO will not increase the COHb saturation of the hemoglobin if the air CO concentration remains constant. This is referred to as the “equilibrium” (or steady-state) COHb level. Estimates have been made as to what COHb level will be reached on the basis of concentration of CO in the air to which a person is exposed. For example, a person exposed to 30 ppm CO for an extended period of time will eventually have about 5% of their hemoglobin as COHb. An exposure to 100 ppm CO will yield an equilibrium COHb level of 20%. Exposures to 600 ppm gives a COHb level of more than 50%.11 In the latter case 50% of the hemoglobin is in the “carboxy” form. Whether more subtle toxic effects occur after this “plateau” COHb level is attained is the object of ongoing research. Various investigators have attempted to correlate health effects with COHb level. Such data nearly always show a huge variability, presumably because it is not just the effect of CO on hemoglobin that is important, but also the effects of CO on tissue and cells (e.g., on the cytochromes) and the effects of lipid peroxidation. Some studies have shown decreased vigilance in subjects with only 2–3% COHb.12,13 Others have shown no effects on vigilance or other health endpoints at COHb levels up to 12.6%.14−27 Conversely, there are studies that showed effects in addition to decreased vigilance at relatively low (less than 13%) COHb levels. These effects included small changes in the electrocardiogram, increased minute volume, reduction in exercise stamina, driving skill deficit, increased heart rate, visual sensitivity decrement, fatigue, and increased reaction time.28−41 Mild headaches are reported in people with COHb levels of 13–20%,16 but other effects have not generally been reported until COHb levels exceed about 30%.31,42−49 Symptoms seen at COHb levels of from 30% to 40% have been associated with severe headache, dizziness, difficulty concentrating, nausea, vomiting, polycythemia, and loss of consciousness.50,51 Subjects with COHb levels from 40% to 45% were unable to perform any tasks requiring even minimal physical exertion.52 Coma and convulsions usually occur at COHb levels of 50–60%, or below.51 Acute COHb levels near 70% are almost always lethal. Note that the median COHb saturation of people dead from CO-poisoning is near 53%. While such crude relationships between COHb levels and health effects may serve as a general guide, the toxicologist must be cognizant of the fact that different individuals are affected differently by CO. This issue is discussed more fully below.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 290 — #4

Investigating Carbon Monoxide Poisonings

291

12.3.2 OCCUPATIONAL EXPOSURE STANDARDS FOR CARBON MONOXIDE Many CO exposures (and poisonings) occur in occupational settings. There are two main types of occupational exposures to CO: accidental exposures and routine exposures. In the case of accidental poisonings, the challenge for the investigator (who is usually not present at the accident itself) is to determine what the exposure levels might have been. It is rare for there to be air quality measurements in such situations and the actions of emergency response personnel may complicate the exposure level assessment. Obviously, the first duty of safety personnel is to assist the CO-poisoning victim. In this effort, first responders may open doors or windows or turn on fans or institute other ventilation efforts before they take measurements of ambient CO levels. Measurements made after ventilation has taken place will be lower that those responsible for the poisoning. CO levels during routine industrial operations are easier to obtain and interpret. Airborne CO levels are routinely monitored in some industrial settings, but in some cases where CO intoxication is suspected, monitoring may not be available. A typical function of the occupational safety investigator is to go to the place of business in question and to obtain CO levels in the air under normal plant operating conditions. These measured CO levels can then be compared to the existing occupation exposure standards. Occupational exposure standards have been set to protect worker’s health. There are three major types of occupational exposure standards and guidelines for CO in this country. The current Occupational Safety and Health Administration (OSHA) standard (which carries the force of law) is an 8-h average of 50 ppm. The National Institute of Occupational Safety and Health (NIOSH) recommended standard is an 8-h average of 35 ppm.53 The American Conference of Governmental and Industrial Hygienists threshold limit value (TLV) over 8 h is 25 ppm.54 OSHA originally proposed lowering their standard to that recommended by NIOSH, but this rule was remanded by the U.S. Circuit Court of Appeals.55 These occupational standards and guideline limits were supposedly set to protect against adverse cardiovascular, respiratory, and neurobehavioral effects. These limits were set to also be protective of pregnant workers and their unborn children, and other workers at high risk, although whether this is actually true is open to question. NIOSH has also established a “ceiling” exposure limit for CO of 200 ppm. This level is not to be exceeded at any time during a working day. The “Immediately Dangerous to Life and Health” (IDLH) for CO is 1200 ppm. Exposure to levels of CO greater than the IDLH “is likely to cause death or immediate or delayed permanent adverse health effects or prevent escape from such an environment.”53

12.3.3 COMMUNITY EXPOSURE STANDARDS AND GUIDELINES FOR CARBON MONOXIDE Air pollution investigators are often called upon to access the levels of CO in community air. CO levels are typically highest near highways or major industrial facilities having combustion sources. As in the case of the industrial environment,

8417: “8417_c012” — 2007/9/11 — 12:15 — page 291 — #5

Carbon Monoxide Poisoning

292

there are methods available to measure the levels of CO in community air. These measured levels can then be compared to air quality standards or guidelines set by federal or state agencies or by other organizations. A variety of community exposure guidelines have been set up to protect public health from air pollutants, including CO. Table 12.1 summarizes the current community exposure standards and guidelines for this chemical. The agencies and organizations that have set community exposure limits for CO include the American Industrial Hygiene Association (AIHA),56 the U.S. Environmental Protection Agency (EPA),57 and the California Air Resources Board (CARB).58,59 Like the occupational exposure standards mentioned in the previous section, these limits have been set to protect members of the general population from the adverse effects of CO exposure.60 Since these standards and guidelines are subject to change, the reader is urged to consult with the agency or organization involved to get the most current exposure levels of interest.

12.3.4 CARBON MONOXIDE EXPOSURE DURATION ASSESSMENTS As mentioned briefly above, the CO exposure duration as well as the exposure level has to be considered in investigations of CO poisonings. In occupational and in some nonoccupational settings, the frequency of exposure is also important. One general rule in toxicology that reflects the importance of both exposure level and duration is referred to as Haber’s Law or Haber’s Rule. This rule can be stated as follows: C×t =k where C is the toxicant concentration or exposure level, t is the time of exposure, and k is a constant reflecting the severity of the toxic effect. For example, if a toxic substance obeys Haber’s Law, a 30 min exposure to 100 ppm of the chemical should give a similar level of toxic effect that a 60 min exposure to 50 ppm would (30 min × 100 ppm = 3000 ppm-min = 60 min × 50 ppm).61 The formation of COHb in CO-poisoned people does not seem to follow Haber’s Law. Data from research studies and from clinical experience is summarized in Table 12.2.62 Three different ranges of COHb levels are shown below: 2.0–2.7%, 7.0–8.5%, and 11.0–12.6%. These data show that the product of CO exposure level and exposure duration is in no way indicative of COHb levels. This may be partially explained by the fact that COHb levels reach an equilibrium concentration at some point in time and do not increase even when exposure duration does. If C remains constant and t is increasing, k, the COHb level in this case will not increase. The deviation from Haber’s Law may also explained by the fact that responses to CO exposures are highly variable from one individual to another. In any case, the duration of exposure is important to the CO investigator. Unless CO levels are very high, an exposure of a few minutes will rarely if ever cause adverse effects. On the other hand, long-term exposures to quite low levels of CO for extended periods of time (months to years) can lead to serious health consequences.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 292 — #6

Investigating Carbon Monoxide Poisonings

293

TABLE 12.1 Summary of Existing Community Air Quality Standards and Guidelines Agency or Organization AIHA AIHA AIHA AIHA AIHA AIHA AIHA EPA EPA CARB CARB CARB

Type of Standard TEEL-0 TEEL-1 TEEL-2 TEEL-3 ERPG-1 ERPG-2 ERPG-3 NAAQS-1 NAAQS-2 AQS-1 AQS-2 REL

Carbon Monoxide Level (ppm) 50 83 83 330 200 350 500 35 9 0.25 0.04 20

Averaging Time 15 min 15 min 15 min 15 min 1h 1h 1h 1h 8h 1h 24 h 1h

The AIHA is the American Industrial Hygiene Association. The EPA is the U.S. Environmental Protection Agency. The CARB is the California Air Resources Board. TEEL-0, the threshold concentration below which most people will experience no appreciable risk of health effects. TEEL-1, the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing other than mild transient adverse health effects or perceiving a clearly defined objectionable odor. TEEL-2, the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing or developing irreversible or other serious health effects or symptoms that could impair their abilities to take protective action. TEEL-3, the maximum concentration in air below which it is believed nearly all individuals could be exposed without experiencing or developing life-threatening health effects. ERPG-1, the maximum concentration in air below which it is believed nearly all individuals could be exposed for up to 1 h without experiencing other than mild transient adverse health effects or perceiving a clearly defined objectionable odor. ERPG-2, the maximum concentration in air below which it is believed nearly all individuals could be exposed for up to 1 h without experiencing or developing irreversible or other serious health effects or symptoms that could impair their abilities to take protective action. ERPG-3, the maximum concentration in air below which it is believed nearly all individuals could be exposed for up to 1 h without experiencing or developing life-threatening health effects. NAAQS-1, the 1-h average National Ambient Air Quality Standard as established by the US EPA. NAAQS-2, the 8-h average National Ambient Air Quality Standard as established by the US EPA. AQS-1, the 1-h average air quality standard as established by the CARB. AQS-2, the 8-h average air quality standard as established by the CARB. REL, the Reference Exposure Level as established by the CARB. RELs are levels at or below which even the most sensitive members of the community would not suffer any adverse health effects.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 293 — #7

Carbon Monoxide Poisoning

294

TABLE 12.2 Carboxyhemoglobin Levels for Different Carbon Monoxide Exposure Levels and Durations CO Exposure Level (C in ppm) 50 12 50 100 50 650 75 100 200

Exposure Duration (t in h) 1.33 192 0.42 2.5 192 0.75 168 8.0 2.67

COHb Level (%)

C × t (ppm-h)

2.0 2.4 2.7 7.0 7.1 8.5 11.0 12.0 12.6

67 2304 21 250 9600 488 12600 800 534

Another factor that comes into play in evaluating a person’s risk of CO exposure is the frequency of exposure as compared to the half-life of CO excretion in the body. For example, in the occupational world, workers normally work 8 h shifts. Unless a worker has significant CO exposures outside of the work place (a possibility, especially if the worker has a long daily commute through heavy traffic) he or she will have 16 h away from work to recover from CO exposure on the job. If a person takes in CO more quickly than it can be eliminated from the body, elevated COHb will tend to persist during off-work times. The half-life (expressed in units of time) of an exogenous chemical in an organism is the time it takes for a given level of that chemical to be reduced by one-half. In two half-lives, the chemical level would be 25% of the original level, after three half-lives, it would be 12.5%, and so on. In CO poisonings, it is sometimes useful to know what initial COHb level in a particular individual. This can be done if the COHb level is measured a short time after the CO-poisoning, and then back-calculating to get the initial level known at to , based on how many COHb half-lives have transpired. The half-life for COHb of a person breathing ambient or room air is roughly 4–5 h.63 Next consider the previously mentioned example of a worker exposed for 8 h on the job and then having little or no CO exposure for the following 16 h. Sixteen hours is approximately four half-lives for COHb for people breathing ambient air. After four half-lives have passed, COHb levels should be reduced by 94%. While this is a large reduction, it should be pointed out that the COHb levels would not return to the baseline levels before the worker went back to work the next day, again to be exposed to CO. On the second day of exposure the COHb level would start at a higher baseline and would then reach a higher level than reached on the first day. The intervening 16 h “rest” periods would decrease the COHb levels by 94% each day, but there would be some accumulation over the work week. Fortunately, the 63 h between 5:00 p.m. Friday afternoon and 8:00 a.m. Monday morning affords almost 16 half-lives, during which the COHb level would be reduced to less than 0.002% of the level on Friday afternoon.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 294 — #8

Investigating Carbon Monoxide Poisonings

295

12.4 TREATMENTS FOR CARBON MONOXIDE POISONING Medical treatments administered to CO-poisoned patients are designed to reverse the effects that CO has on the body; namely, the effects on the blood oxygen carrying capacity, the binding to myoglobin in muscle, the interference CO exerts on the cytochrome oxidase system, effects on lipid peroxidation in the brain and other tissues, and so forth. The type of treatment provided can influence the investigator’s conclusions concerning the nature and extent of the damage caused by CO poisoning. Emergency medical technicians (EMTs) and other “first responders” are trained to administer 100% oxygen to people they suspect have been poisoned by CO. This treatment has the effect of reducing the COHb levels in the patient’s blood at a faster rate than would be expected without such treatment. As mentioned earlier, the halflife of COHb without supplemental oxygen is 4–5 h. The half-life of COHb in patients breathing 100% oxygen centers around 60–80 min.63,64 Oxygen treatment functions to promote the dissociation of COHb65 and to reduce tissue hypoxia. A more controversial treatment regime for the CO-poisoning victim is hyperbaric oxygen treatment (HBOT). See in-depth discussions of this treatment modality elsewhere in this book. HBOT is accomplished by placing the patient in a large sealed chamber, giving the patient 100% oxygen, and gradually increasing the pressure inside to levels several times greater than of atmospheric. It is clear is that HBOT can greatly decrease the half-life of COHb—to roughly 20–30 min.63,65,66 HBOT has also been shown to promote CO dissociation from cytochrome a3 oxidase in animal studies and to reduce brain lipid peroxidation.65 The efficacy of HBOT in humans has not been conclusive. Some studies have shown marked reductions in the number and severity of both acute symptoms and the incidence of delayed neuropsychological sequelae.67,68 Other investigators have reported either no advantage to HBOT69 or have even found that HBOT worsened the patient’s condition.70 One explanation for these conflicting findings, besides the variability in susceptibility expected in those CO-poisoned, is that the groups of patients studied came from a wide variety of CO poisoning situations. Some HBOT was done at less than the optimal pressures of from 2.5 to 3.0 atm., while in other studies there had been a delay in initiating the HBOT. Some groups of CO-poisoned individuals studied, included individuals who had lost consciousness, while others did not and some studies had flaws in design and execution.5,65,71 The current consensus seems to be that HBOT should be applied selectively. There are risks of side-effects to this treatment and the transport of critically injured patients to the nearest HBOT center also poses risks. HBOT may not be the best approach for all patients. HBOT is more likely to have benefits outweighing its risks for patients having one or more of the following characteristics: 1. Severe intoxication as evidenced by coma, seizures, focal neurological deficits or cardiac effects65 2. Those who can be given HBOT within 6 h of the poisoning event67

8417: “8417_c012” — 2007/9/11 — 12:15 — page 295 — #9

296

Carbon Monoxide Poisoning

3. Those presenting with COHb levels in excess of 25%.5,71 4. Some studies suggest that patients with less severe CO poisoning may do just as well after normobaric treatment (i.e., sea level) with 100% oxygen.66

12.5 OTHER FACTORS TO CONSIDER IN INVESTIGATIONS OF CARBON MONOXIDE POISONING The investigator of CO-poisoning events needs to be aware of other causes of the symptoms that are associated with overexposure to CO. The process by which other causes are dealt with is called the “differential diagnosis.” Other conditions that can have symptoms in common with CO poisoning include viral infections, food poisoning, depression, anxiety, transient ischemic attacks, coronary artery disease, cardiac arrythmias, Parkinson’s disease, meningitis or encephalitis, epilepsy, migraine, drug overdose, ethanol intoxication, pneumonia, and sinusitis.10,65 Clues exist by which to discern whether the case being investigated is actually a CO-poisoning case. For example, if all of the occupants of a residence or business establishment are affected, many of the above alternate explanations can be eliminated. If symptoms improve when the victims depart the site, this usually points the finger at suspected site as the likely culprit. Indications of incomplete combustion, such as yellow flames on gas appliances or heaters provide an indication that CO exposure is possible/likely.70 In the occupational environment, CO exposures can occur wherever there is a source of combustion. Some examples include propane or gasoline powered forklifts (i.e., hi-lows, lift trucks), generators, bobcats, scissor lifts, cherry pickers, concrete saws and chain saws, floor strippers and polishers, and so forth. The toxicity of CO is enhanced at high altitudes, at elevated temperatures, and in people having increased ventilation or metabolic rates. The human fetus is at particularly high risk from CO poisoning because of its normal development in a somewhat oxygen-deprived environment. Children are at higher risk of CO poisoning because they generally have higher metabolic rates than adults. People with preexisting diseases such as anemia, cardiovascular, or cerebral vascular disease, hypovolemia, or those with increased endogenous CO production are also at higher risk of adverse effects from CO exposure.10 Exposures to some other chemicals either in the home or in the workplace can mimic CO exposures and are easy to confuse with those exposures. For example, exposure to methylene chloride (i.e., dichloromethane) can cause the same types of effects as exposure to CO. This is because methylene chloride is metabolized to CO and carbon dioxide in the body. An 8-h exposure to 150 ppm of methylene chloride produces the same elevated carboxyhemoglobin level that a 35 ppm exposure to CO would produce over the same time period. While causing less of an effect per ppm of exposure, it has been found that the half-life of methylene chloride-induced COHb is greater than that of CO-induced COHb, so effects may last for longer periods of time.7 Exposure to cyanide can mimic CO intoxication, mainly because cyanide also disrupts the functioning of cytochrome a3 oxidase.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 296 — #10

Investigating Carbon Monoxide Poisonings

297

12.6 CASE STUDY OF CARBON MONOXIDE POISONING The following is an account of an actual case in which three individuals were severely poisoned by CO. This particular situation involved a faulty swimming pool heater at a hotel. Malfunctions of heaters are a common cause of CO-related poisonings and fatalities. The names of the victims, the other individuals involved in this incident, and the name of the hotel have been omitted to maintain privacy. The three individuals involved were staying at a hotel near Denver, Colorado. Victim #1 went jogging the evening of October 31 and returned to his room at about 7:00 p.m. He soon felt ill and probably passed out around 8:00 p.m. that evening. Housekeeping staff entered his room the next morning at 8:20 a.m., but seeing him on the bed, assumed he was just sleeping soundly and left the room. A colleague of Victim #1 who was staying at another hotel called Victim #1’s hotel at 9:45 a.m. on November 1 and asked the desk clerk on duty to check on his friend. The clerk went to Victim #1’s room, knocked on the door, but did not enter. He stated that it was against hotel policy to enter a room unless the guest gave them authorization to do so. Later that morning after repeated calls from the colleague, the hotel staff did enter Victim #1’s room. Hotel records show that Victim #1’s door was opened at 11:04 a.m. and 11:15 a.m. by the desk clerk trying unsuccessfully to rouse him. The colleague finally came to the hotel to check on Victim #1 at about 11:30 a.m. and went to his room. Finding him unconscious, the colleague asked the desk clerk to call 911 and to summon emergency medical personnel. Victim #1 was finally removed from the room at about 11:55 a.m. Victim #1 therefore had an approximately 17-h exposure to CO. In this case the duration of exposure could be obtained from what is known as an “audit trail.” At some hotels each time the card key is used to open the door, the time and whose card was used (guest, housekeeping, desk clerks, etc.) is recorded on the hotel computer. This victim suffered severe brain damage. Victim #2 has stated that she got into her room at the hotel at about 5:00 p.m. on October 31. After 15–20 min she felt confused and nauseated. Maids working at the hotel entered her room at about 8:30 a.m. the next day and found her passed out on the floor and thought that she had had too much to drink (example of a “misdiagnosis”). Upon reporting this to the hotel management, the maids apparently were told to leave Victim #2 alone. Nothing was done to aid her until the emergency personnel (who had been summoned to assist Victim #1) arrived and removed her from her room just before noon. She was therefore exposed to the CO for a total of about 19 h. This victim incurred cognitive deficits and had to have a leg amputated because of compartment syndrome she suffered caused by the CO exposure. Victim #3 entered her room at about 5:00 p.m. on October 31. She believed she passed out by about 6:00 p.m. Victim #3 awoke at about 6:30 a.m. the next morning, showered, got dressed, and passed out again. She called some friends to come and get her at about 8:00 a.m. Sometime shortly thereafter, she also called the front desk to let them know she had vomited and that the carpet needed to be cleaned. Her friends arrived at about 9:00 a.m. and seeing her condition took her to the hospital. Therefore, she was in her room and exposed to the CO for about 15.5 h. Her outcome involved limited neuropsychological deficits.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 297 — #11

298

Carbon Monoxide Poisoning

Blood tests done after these three victims reached the hospital showed that each had elevated COHb levels. This confirmed that they had been exposed to CO. All three people were diagnosed by physicians at the hospitals as having CO-poisoning. Their symptoms at the emergency room (ER) included severe headache, dizziness, nausea, vomiting, reddening of the skin, compartment syndrome, loss of muscle tissue, fatigue, hypotension, and coma. These are consistent with what would be expected from CO exposures. In addition, all three victims suffered ongoing health damage involving sleep disturbance, vision problems, hearing loss, tinnitus, peripheral neuropathy, mental deficits, memory problems, and difficulty concentrating. During their investigations of this incident, local fire department personnel measured CO concentrations in the rooms occupied by the three individuals. They were all in excess of 200 ppm. These readings were taken after the doors to the rooms had been opened and some ventilation of these rooms had occurred. Because of this, the CO levels to which the victims were exposed would certainly have been higher than that recorded by the fire department. The source of the CO was traced to the swimming pool heater in the hotel. CO levels in this mechanical room were found to be “extremely high” according to fire department personnel. Police investigating this incident looked for, but did not find any evidence of the presence of drugs, drug paraphernalia, or alcohol in the Plaintiffs’ hotel rooms. Furthermore, there was no evidence of any violence or trauma to the victims. Victim #1 suffered occasional headaches and visual problems prior to the incident, but these became much worse afterwards. Otherwise, the Plaintiffs’individual medical histories prior to this incident were unremarkable. Thus, it was possible to rule out other possible causes for the victims’ adverse health effects. The conclusion reached in this CO-poisoning investigation was that the adverse health effects suffered by the victims were caused by their exposures to CO at the hotel. The bases for this conclusion are as follows: 1. Elevated levels of CO in excess of 200 ppm were found in the hotel rooms occupied by the individuals even after some ventilation of these rooms had occurred 2. The victims were exposed to high levels of CO for periods of time ranging from 15.5 to 19 h 3. The presence of elevated levels of COHb in the blood of these three individuals confirms that they did sustain an exposure to CO 4. Both the acute and the long-term symptoms exhibited by the three victims were entirely consistent with those associated with an overexposure to CO 5. Finally, the three victims in this case had no significant medical or psychiatric conditions or problems prior to the incident, and no other likely explanations could be found for the health harm suffered by the individuals. Another conclusion was that the victims’ injuries would have been less severe if they had been removed from their hotel rooms earlier. The housekeeping staff went into each of the three victims’ rooms at about 8:30 a.m. on November 1. Although the first two victims were found unconscious at that time, they were not taken out of their rooms until about noon. This resulted in each of them sustaining an extra three and

8417: “8417_c012” — 2007/9/11 — 12:15 — page 298 — #12

Investigating Carbon Monoxide Poisonings

299

one-half hours of CO exposure. The odds of more serious injuries go up with each additional hour of exposure. Victim #3 was fortunate to have been able to summon aid on her own. Notably, she suffered fewer and less severe effects than the other two individuals who had 2–3 h more exposure to CO at the hotel. The basis for this finding is that neurological damage from CO poisoning is progressive. It becomes more severe the longer tissues are without adequate oxygen as explained above. If Victim #1 had been taken out of his room earlier and had received medical treatment sooner, he would have probably suffered less severe brain damage. If Victim #2 had been rescued earlier, she may not have suffered such a severe injury to her leg. Victim #3’s continuing injuries (problems with memory loss and other mental deficits), while not as great as those suffered by the other two victims, would most likely not have been as severe or taken as long to overcome had she been able to get to the hospital sooner.

12.7 SUMMARY CO poisoning is a leading cause of accidental injuries and fatalities in this country and throughout the world. Exposures to CO are possible wherever there is a source of combustion; for example, heating systems, fires, petroleum product fueled vehicles, and industrial equipment, to name a few. This chapter has been an overview of how investigations of CO-poisoning incidents are carried out and the types of information required in such investigations. Investigations include evaluations of the signs and symptoms of the intoxication, assessments of exposure level and duration, reviewing the medical treatments that may have been administered, and the ruling out of other factors that may have caused the poisoning. The investigator first of all should be familiar with the signs and symptoms of CO poisoning. Many of the symptoms are common to other conditions such as viral infections, alcohol intoxication, coronary disease, and other disease states. This can complicate a positive determination of CO’s involvement. Other chemicals can cause similar symptoms and should be also investigated as possible causes. It is crucial in an investigation of CO poisonings to determine how great a “dose” of CO the victim received. This may be done using CO measurement instruments, or by determining COHb levels subsequent to the poisoning incident. Measured airborne CO levels can be compared to occupational and community exposure standards and guidelines, to assess health risk in more routine human exposure scenarios. The medical treatment that an individual received can influence the investigation of CO poisonings. Whether a victim received no supplemental oxygen, was given 100% oxygen to breathe, or had HBOT, will influence the immediate health condition and the long-term prognosis for that individual. CO poisonings have occurred for tens of thousands of years, and probably even before humans inhabited the earth (from volcanic activity, forest fires, etc.). The advent of man’s use of controlled fires for warmth and cooking no doubt signaled a sharp rise in the risk and incidence of CO poisonings. Even with all of the technological advances that have been made over the millennia, CO poisoning is still common

8417: “8417_c012” — 2007/9/11 — 12:15 — page 299 — #13

Carbon Monoxide Poisoning

300

and continues to be a major threat to public health. One objective of this chapter is to assist those who are involved in investigations of CO poisoning. This chapter and the others in this book also serve to raise awareness of this threat and to hopefully reduce the number of CO poisonings.

References 1. Shephard, R.J. Carbon Monoxide, The Silent Killer, Charles C. Thomas Publisher, Springfield, Illinois, 1983. 2. Kindwall, E.P. Carbon monoxide, In: Occupational Medicine, 3rd ed., Zenz, C., Dickerson, O.B., and Horvath, E.P., eds., Mosby, St. Louis, 1994, Chapter 29. 3. U.S. Environmental Protection Agency, Air Quality Criteria for Carbon Monoxide, USEPA Publication No. EPA 600/P-99/001F, U.S. Government Printing Office, Washington, DC., 2000. 4. Lipsett, M.J., Shusterman, D.J., and Beard, R.R. Inorganic compounds of carbon, nitrogen, and oxygen, In: Patty’s Industrial Hygiene and Toxicology, 4th ed., Vol. II, Part F, Clayton, G.D., and Clayton, F.E., eds., John Wiley & Sons, Inc., New York, 1994, pp. 4523–4552. 5. Tomaszewski, C. Carbon monoxide, In: Goldfrank’s Toxicological Emergencies, 6th ed., Goldfrank, L.R., et al., eds., Appleton & Lange, Stamford, Connecticut, 1998, pp. 1551–1563. 6. Victor, M., and Adams, R.D. Metabolic Diseases of the Nervous System, In: Harrison’s Principles of Internal Medicine, 9th ed., McGraw-Hill Book Company, New York, 1980, pp. 1978–1979. 7. Hathaway, G.J., Proctor, N.H., and Hughes, J.P. Carbon monoxide, In: Proctor and Hughes’ Chemical Hazards of the Workplace, 4th ed., Van Nostrand Reinhold, New York, 1996, pp. 113–116. 8. Smith, R.P. Toxic responses of the blood, In: Casarett and Doull’s Toxicology, The Basic Science of Poisons, 5th ed., Klaassen, C.D., Amdur, M.O., and Doull, J., eds., McGraw-Hill, New York, 1996, Chapter 11, pp. 343–344. 9. Smith, J.S., and Brandon, S. Morbidity from acute carbon monoxide poisoning at three-year follow-up, Br. Med. J. 1, 318, 1973. 10. Respiratory Toxicology, In: Ellenhorn’s Medical Toxicology, 2nd ed., Ellenhorn, M.J., et al., eds., Williams & Wilkins, Baltimore, pp. 1465–1476, 1997. 11. Seinfeld, J.H. Air Pollution, Physical and Chemical Fundamentals, McGraw Hill Book Company, New York, 1975, pp. 22. 12. Fodor, C.G., and Winneke, C. Effect of Low CO concentrations on resistance to monotony and on psychomotor capacity, Staub-Reinhalt Luft, 32, 46, 1972. 13. Beard, R.R., and Grandstaff, N.W. Carbon monoxide and human functions, In: Environmental Science Research, Vol. 5, Behavioral Toxicology, B. Weiss and V.G. Laties, eds. New York, Plenum Press, 1975,pp. 1–27. 14. Sievers, R.F., Edwards, T.I., and Murray, A.L. A Medical Study of Men Exposed to Measured Amounts of Carbon monoxide in the Holland Tunnel for 13 Years, Public Health Bulletin No. 278, U.S. Government Printing Office, Washington, DC., 1942. 15. Mikulka, P.R., et al. The effect of carbon monoxide on human performance, Ann. N.Y. Acad. Sci. 174, 409, 1970. 16. Stewart, R.D., et al. Experimental human exposure to carbon monoxide, Arch. Environ. Health 21, 154, 1970.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 300 — #14

Investigating Carbon Monoxide Poisonings

301

17. Horvath, S.M., Dahms, T.E., and O’Hanlon, J.F. Carbon monoxide and human vigilance, a deleterious effect of present urban concentrations, Arch. Environ. Health 23, 343, 1971. 18. O’Donnell, R.D., et al. Low level carbon monoxide exposure and human psychomotor performance, Toxicol. Appl. Pharmacol. 18, 583, 1971. 19. Raven, P.B., et al. Effect of carbon monoxide and peroxyacetyl nitrate on man’s maximal aerobic capacity, J. Appl. Physiol. 36, 288, 1974. 20. Ettema, J.H., et al. Effects of alcohol, carbon monoxide and trichloroethylene exposure on mental capacity, Int. Arch. Occup. Environ. Health 35, 117, 1975. 21. O’Hanlon, J.F. Preliminary studies of the effects of carbon monoxide on vigilance in man, In: Behavioral Toxicology, Weiss, B., and Laties, G., eds., Plenum Press, New York, 1975, pp. 61–75. 22. Benignus, V.A., et al. Lack of effects of carbon monoxide on human vigilance, Perception and Motor Skills. 45(3, Pt 1), pp. 1007–1014, 1977. 23. Luria, S.M., and McKay, C.L. Effects of low levels of carbon monoxide on vision of smokers and nonsmokers, Arch. Environ. Health 34, 38, 1979. 24. Davies, D.M., et al. The effects of continuous exposure to carbon monoxide on auditory vigilance in man, Int. Arch. Occup. Environ. Health 48, 25, 1981. 25. DeLucia, A.J., Whitaker, J.H., and Bryant, L.R. Effects of combined exposure to ozone and carbon monoxide (CO) in humans, In: Advances in Modern Environmental Toxicology, Vol. 5, Lee, S.D., Mustafa, C.G., and Mehlman, M.A., eds., Princeton Scientific Publishers, Princeton, N.J., 1983, pp. 145–159. 26. Mihevic, P.M., Gliner, J.A., and Horvath, S.M. Carbon monoxide exposure and information processing during perceptual-motor performance, Int. Arch. Occup. Environ. Health 51, 355, 1983. 27. Benignus, V.A., et al. Effect of low level carbon monoxide on compensatory tracking and event monitoring, Neurotoxicol. Teratol. 9, 227, 1987. 28. McFarland, R.A., et al. The effects of carbon monoxide and altitude on visual thresholds, J. Aviat. Med. 15, 381, 1944. 29. Ray, A.M., and Rockwell, T.H. An exploratory study of automobile driving performance under the influence of low levels of carboxyhemoglobin, Ann. N.Y. Acad. Sci. 174, 396, 1970. 30. Bender, W., Goethert, M., and Malorny, G. Effect of low carbon monoxide concentrations on psychological functions, Staub-Reinhalt Luft, 32, 54, 1972. 31. Ekblom, B. and Huot, R. Response to submaximal and maximal exercise at different levels of carboxyhemoglobin, Acta. Physiol. Scand. 86, 474, 1972. 32. McFarland, R. Low-level exposure to carbon monoxide and driving performance, Arch. Env. Health 27, 355, 1973. 33. Ramsey, J.M. Effects of single exposures of carbon monoxide on sensory and psychomotor response, Am. Ind. Hyg. Assoc. J. 34, 212, 1973. 34. Wright, G., Randell, P., and Shephard, R.J. Carbon monoxide and driving skills, Arch. Env. Health 27, 349, 1973. 35. Drinkwater, B.L., et al. Air pollution, exercise, and heat stress, Arch. Env. Health. 28, 177, 1974. 36. Gliner, J.A., et al. Man’s physiologic response to long-term work during thermal and pollutant stress, J. Appl. Physiol. 39, 628, 1975. 37. Horvath, S.M., et al. Maximal aerobic capacity at different levels of carboxyhemoglobin, J. Appl. Physiol. 38, 300, 1975.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 301 — #15

302

Carbon Monoxide Poisoning 38. Putz, V.R., Johnson, B.L., and Setzer, J.V. Effects of CO on Vigilance Performance. Effects of Low-Level Carbon monoxide on Divided Attention, Pitch Discrimination, and the Auditory Evoked Potential, Publication Number DHEW (NIOSH) 77–124, U.S. Department of Health, Education, and Welfare, National Institute of Occupational Safety and Health, Cincinnati, Ohio, 1976. 39. Putz, V.R., Johnson, B.L., and Setzer, J.V. A comparative study of the effects of carbon monoxide and methylene chloride on human performance, J. Env. Pathol. Toxicol. 2, 97, 1979. 40. Davies, D.M. and Smith, D.J. Electrocardiographic changes in healthy men during continuous low-level carbon monoxide exposure, Env. Res. 21, 197, 1980. 41. Bunnell, D.E. and Horvath, S.M. Interactive effects of heat, physical work and CO exposure on metabolism and cognitive task performance, Aviat. Space Env. Med. 60, 428, 1989. 42. Chevalier, R.B., Krumholz, R.A., and Ross, J.C. Reaction of non-smokers to carbon monoxide inhalation, cardiopulmonary responses at rest and during exercise, JAMA. 198, 1061, 1966. 43. Pirnay, F., et al. Muscular exercise during intoxication by carbon monoxide, J. Appl. Physiol. 31, 573, 1971. 44. Vogel, J.A. and Gleser, M.A. Effect of carbon monoxide on oxygen transport during exercise, J. Appl. Physiol. 32, 234, 1972. 45. Vogel, J.A., et al. Carbon monoxide and physical work capacity. Arch. Environ. Health 24, 198, 1972. 46. Parving, H.H. The effect of hypoxia and carbon monoxide exposure on plasma volume and capillary permeability to albumin, Scand. J. Clin. Lab. Invest. 30, 49, 1972. 47. Stewart, R.D., et al. Effect of carbon monoxide on time perception, Arch. Environ. Health 27, 155, 1973. 48. Hudnell, H.K. and Benignus, V.A. Carbon monoxide exposure and human visual detection thresholds, Neurotoxicol. Teratol. 11, 363, 1989. 49. Schulte, J.H. Effects of mild carbon monoxide intoxication, Arch. Environ. Health 7, 524, 1963. 50. DiMarco, A. Carbon monoxide poisoning presenting as polycythemia, N. Engl. J. Med. 319, 874, 1988. 51. Chiodi, H., et al. Respiratory and circulatory responses to acute carbon monoxide poisoning, Am. J. Physiol. 134, 683, 1941. 52. Stewart, R.D. The effect of carbon monoxide on humans, Ann. Rev. Pharmacol. 15, 409, 1975. 53. NIOSH Pocket Guide to Chemical Hazards, Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, DHHS (NIOSH) Publication Number 97–140, U.S. Government Printing Office, Washington, DC, 2003. 54. 2005 TLVs and BEIs, American Conference of Governmental Industrial Hygienists, ACGIH Signature Publications, Cincinnati, Ohio, 2005. 55. Department of Labor, Occupational Safety and Health Administration, 29 CFR Part 1910, Air Contaminants, Federal Register, 54, 2651, 1989. 56. Current values for the American Industrial Hygiene Association can be found at the Internet Web Site http://www.eh.doe.gov/chem_safety/teel.html. 57. Current National Ambient Air Quality Standards can be found at the USEPA Internet Web Site http://www.epa.gov/air/criteria.html. 58. Current values for California Air Quality Standards can be found at the CARB Internet Web Site http://www.baaqmd.gov//pln/air_quality/ambient_air_quality.htm.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 302 — #16

Investigating Carbon Monoxide Poisonings

303

59. Current values for the California Reference Exposure Limits can be found at the CARB Internet Web Site http://www.oehha.ca.gov/air/pdf/acuterel.pdf. 60. Craig, D.K. Derivation of temporary emergency exposure limits (TEELs), J. Appl. Toxicol. 20, 11, 2000. 61. Pastenbach, D.J. The History and Biological Basis of Occupational Exposure Limits for Chemical Agents, In: Patty’s Industrial Hygiene and Toxicology, 5th ed., Volume 3, VI, Harris, R.L., ed., John Wiley & Sons, Inc., New York, 2000, pp. 1953–1954. 62. National Research Council, Carbon monoxide, In: Review of Submarine Escape Action Levels for Selected Chemicals, National Academy Press, Washington, DC, 2002, pp. 69–96. 63. Varon, J. and Marik, P.E. Carbon monoxide poisoning, J. Emerg. Int. Care Med. 1, 1, 1997. 64. Weaver, L.K., et al. Carboxyhemoglobin half-life in carbon monoxide-poisoned patients treated with 100% oxygen at atmospheric pressure, Chest. 117, 801, 2000. 65. Gill, A.L. and Bell, C.N.A. Hyperbaric oxygen: Its uses, mechanisms of action, and outcomes, Q. J. Med. 97, 385, 2004. 66. Leach, R.M., Rees, P.J., and Wilshurst, P. ABC of oxygen: hyperbaric oxygen therapy, Brit. Med. J. 317, 1140, 1998. 67. Thom, S.R., et al. Delayed neuropsychological sequelae after carbon monoxide poisoning: Prevention by treatment with hyperbaric oxygen, Ann. Emerg. Med. 25, 474, 1995. 68. Ducasse, J.L., Celis, P., and Marc-Vergnes, J.P. Non-comatose patients with acute carbon monoxide poisoning: Hyperbaric or normobaric oxygenation?, Undersea Hyperb. Med. 22, 9, 1995. 69. Weaver, L.K., Hopkins, R.O., and Larson-Lohr, V. Neuropsychological and functional recovery from severe carbon monoxide poisoning without hyperbaric oxygen therapy, Ann. Emerg. Med. 27, 736, 1996. 70. Scheinkestel, C.D., et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: A randomized controlled clinical trial, Med. J. Aust. 170, 203, 1999. 71. Harper, A. and Croft-Baker, J. Carbon monoxide poisoning: undetected by both patients and their doctors, Age Ageing 33, 105, 2004.

8417: “8417_c012” — 2007/9/11 — 12:15 — page 303 — #17

8417: “8417_c012” — 2007/9/11 — 12:15 — page 304 — #18

13

Carbon Monoxide Detectors as Preventive Medicine James W. Rhee and Jerrold B. Leikin

CONTENTS 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Carbon Monoxide Detector Technology: A Brief Review . . . . . . . . . . . . . . . . 13.3 Setting an Initial Standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4 The Chicago Experience. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 Clinical Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.6 Some Data Regarding Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.7 Current and Future Directions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

305 306 306 307 308 308 309 310

13.1 INTRODUCTION Carbon monoxide (CO) is odorless, colorless, tasteless, and nonirritating. As such, CO has no warning properties that can alert unwary individuals to its presence. When people become ill owing to CO toxicity, the symptom complex can be nonspecific and variable making the diagnosis of CO poisoning difficult. Given its lack of warning signs and lack of distinct clinical features, CO has been labeled by some as a “silent killer.” Data analyzed by Cobb and Etzel from the National Center for Health Statistics, attributed 53,133 deaths to CO from 1979 to 1988—making it the most common cause of acute poisoning deaths.1 When Shepherd and Klein-Schwartz examined the mortality data from 1979 to 1994 for persons aged 10–19 years, they found 3034 out of 7936 poisoning deaths were attributable to carbon monoxide.2 As these numbers were derived from databases, they are likely to be significant under representation of the true number of deaths caused by CO. They also do not take into account the significant morbidity that may occur from severe CO exposures. Given the “silent” nature of CO and its significant impact on public health, a need for residential-based CO detectors was identified. In 1989, the Consumer Product Safety Commission (CPSC) urged Underwriters Laboratories (UL) (Northbrook, Illinois) to develop a standard that would serve as a guideline for residential 305

8417: “8417_c013” — 2007/9/11 — 12:15 — page 305 — #1

Carbon Monoxide Poisoning

306

CO detectors. This standard (UL 2034) applying to single and multiple station CO detectors was published in April, 19923 —subsequently, the first two CO detectors were listed later that year.4

13.2 CARBON MONOXIDE DETECTOR TECHNOLOGY: A BRIEF REVIEW Most of the residential CO detectors utilize one of the two methods of sensing the presence of CO—either a biomimetic sensor or a metal oxide sensor. The types of technology usually involved in residential sensors include the “Gel cell” metal oxide and electrochemical sensor. The nondispersive infrared technology (NDIR) is usually utilized for industrial purposes. The “Gel cell” based sensor sends a light beam through a biomimetic sensor to a photosensitive component which alarms at appropriate set points. This colorimetric sensor essentially mimicked the hemoglobin uptake of CO, thus changing the spectral response. These types of sensors exhibit proper sensitivity to CO and can achieve a unique accumulation of CO. There is some interference with other gases and vapors along with humidity, and these types of sensors take a longer time to recover in the setting of CO removal (slower response reversibility). The “Gel cell” detectors are rarely equipped with digital displays. The metal-oxide (usually tin-oxide) sensor detects CO by measuring the resistivity of the metal component through oxidation of CO to carbon dioxide, which reduces the resistance of the sensor. These sensors usually require main (i.e., AC) power and cannot use batteries as a primary power source. Upon exposure to CO, the electrochemical sensors will generate electricity through an acid electrolyte (usually sulfuric acid or phosphoric acid). This technology has been utilized as portable or fixed gas monitors in industry owing to its resolution (down to 1 ppm) and stability, and is probably the primary type of sensor used in residential detectors. These sensors can also be affected by cross-contamination of gases. The NDIR can measure CO concentrations with an accuracy of ±2 ppm, but are expensive and therefore usually not utilized in residential detectors. Metallocorroles utilizing cobalt (111) can selectively absorb CO gas on a molecular complex and may provide a future matrix in sensor development.5

13.3 SETTING AN INITIAL STANDARD UL is a well-known organization located in Northbrook, Illinois that develops standards and test procedures for materials, components, assemblies, tools, equipment, and procedures, chiefly dealing with product safety and utility. The CPSC worked closely with UL to develop standards for CO detectors. Part of setting this standard required UL to create thresholds at which the detectors would alarm. The threshold at which the residential CO detectors would alarm presented a unique challenge. Initially, set points were generated on the basis of extrapolating what levels of CO over what period of time would generate a carboxyhemoglobin (COHb) concentration of 10% in a nonsmoking individual. These values were based

8417: “8417_c013” — 2007/9/11 — 12:15 — page 306 — #2

Carbon Monoxide Detectors as Preventive Medicine

307

TABLE 13.1 UL 2034 (1992–1996). Carbon Monoxide Concentration Versus Time for Alarm Test Points Air Concentration (ppm)

Maximum Response Time (min)

100 200 400

90 35 15

on the Coburn–Forster–Kane equation. (In addition, the CO detector should alarm at an exposure of 6000 ppm within 3 min.)6 This initial standard (UL 2034) was set in place in 1992.3 Soon afterwards CO detectors were sold across the United States. Table 13.1 illustrates the original set points for activating the audible alarm in the CO detectors. Other standards, such as loudness of the alarm and the response to other gases needed to be clarified as well. Similar to smoke detectors, the CO detector was set to emit an 85-dB alarm (at 10 ft.), which is loud enough to wake most people when sounding outside a bedroom through a closed door. However, unlike smoke detectors, CO detectors have computer processor-based software which allow it to alarm at certain set points. Sensors were set to not alarm when exposed for 2 h to methane (at 500 ppm), butane (at 300 ppm), heptane (at 500 ppm), ethyl acetate (at 200 ppm), isopropyl alcohol (at 200 ppm), carbon dioxide (at 5000 ppm), toluene (at 200 ppm), and acetone (at 200 ppm).7 Other major features include a red lightemitting diode (LED) as a trouble signal, a green LED indicating normal operation and a reset mechanism for testing and resetting purposes. Some detectors utilize a digital display (over 30 ppm).

13.4 THE CHICAGO EXPERIENCE The state of Illinois, in the past had one of the highest fatality rates from CO— accounting for 8.7% of all unintentional CO deaths nation-wide, with a rate of about 0.9 deaths per 1000 people between 1979 and 1988.1 In light of this statistic, Chicago, Illinois, passed a mandatory CO detector ordinance in March, 1994 requiring CO detectors to be present in all homes, apartments, class “B” and “C” buildings with heat sources that generate CO.8 During the first 3 months of the ordinance implementation, 68 individuals were transported with suspected CO poisoning attributed to CO detector alarms.9 However, during this time period (October 1–December 31, 1994), the Chicago Fire Department received over 12,000 calls of CO detector alarming. In about 85% of these cases, less than 9 ppm of CO was reportedly measured by the Chicago Fire Department.9–12 This was climaxed on December 21–22, 1994, when the Chicago Fire Department was involved in 3464 CO investigations.11 It was subsequently determined that a thermal air inversion (upper atmosphere warmer than lower atmosphere) had caused a fivefold

8417: “8417_c013” — 2007/9/11 — 12:15 — page 307 — #3

Carbon Monoxide Poisoning

308

elevation of ambient CO concentration, of 13–20 ppm, thus setting off multiple CO alarms. CO detector sensitivity or resistance specifications were 15 ppm over an 8-h period at that time.9,10 In response to the inordinate number of “nuisance alarms,” UL increased the 15 ppm resistance specification from 8 h to 30 days in the 1996 revised standard.7 The UL standard 2034 is under a continuous maintenance process and was last revised in 2006.

13.5 CLINICAL IMPLICATIONS Given the variable presentation of CO toxicity—it is apparent that healthcare providers could use assistance in diagnosing occult CO poisoning. Patients with CO poisoning are often misdiagnosed as having a viral syndrome or food poisoning.13 Earlier studies have shown that among adult patients presenting to an emergency department during the winter months with complaints of headache or dizziness, 3–5% have COHb levels greater than 10%.14 Because CO is a colorless, odorless, tasteless, and nonirritating gas, the physician and the patient have few clues as to the contribution of CO to the illness (see other chapters in this book). The CO detector can act as a useful screening tool to identify CO exposure. The CO detector should allow the clinician to obtain a history of exposure and prompt the clinician to begin investigating the potential for CO poisoning.

13.6 SOME DATA REGARDING EFFECTIVENESS CO detectors/alarms have made death from CO poisoning completely preventable. CO detectors can save lives. Despite the high number of false alarms in Chicago after the initial mandatory CO detector ordinance was established in 1994, only one CO-related death was reported in the Chicago media between September 1994 through February 1998. Compared to other cities during this same time period, Chicago (the only city to have a CO detector ordinance during the study period) had the lowest case fatality rate.15 And despite the loudly voiced criticisms and objections to the ordinance from various sources, a Chicago Sun-Times poll (though hardly scientific) reported that 77% of respondents supported such a mandatory ordinance. Krenzelok et al.16 found that in an investigation of emergency responses to possible CO poisoning, residences with CO alarms had lower CO concentrations (18.6 ppm with CO detectors versus 96.6 ppm without CO detectors) and fewer symptomatic patients (11.7% with CO detectors vs 63.4% without CO detectors).16 They concluded that audible CO detectors were effective in alerting people to the presence of abnormal levels of CO, thus resulting in less exposure to CO. This subsequently lead to a lower incidence of CO-related symptoms.16 A study conducted by Bizovi et al.10 found similar results where only about 5% of individuals at a site where a CO detector had activated displayed signs of CO poisoning, suggesting that the CO detector may have prevented more serious exposure.10

8417: “8417_c013” — 2007/9/11 — 12:15 — page 308 — #4

Carbon Monoxide Detectors as Preventive Medicine

309

Yoon et al.17 estimate that 78 out of 136 unintentional deaths due to CO poisoning that occurred during a 15-year period in New Mexico, may have been prevented if audible CO detectors were in use.17 In North Carolina, a couple of studies evaluated the impact of a local CO detector ordinance (which exempted all-electric heated homes) was conducted.18,19 While, the ordinance did not seem to decrease the amount of CO exposures, the relative incidence of severe poisoning requiring hyperbaric oxygen treatment was diminished.18 Another study evaluating a CO poisoning outbreak during a winter storm in 2002 demonstrated that the North Carolina county CO detector ordinance did not eliminate a CO poisoning outbreak, but it did mitigate its effects.19 Of note, the study found that none of the patients who developed symptoms of severe CO poisoning had a functioning CO detector.19 Despite the apparent effectiveness of CO detectors at saving lives, only 29% of respondents to a survey conducted by Runyan et al.20 reported the presence of CO detectors in their homes.20 The lack of widespread use of these detectors intuitively limits the effectiveness of these devices to have a profound impact on mortality and morbidity due to CO poisoning.

13.7 CURRENT AND FUTURE DIRECTIONS Newer aspects of CO detector specifications include an alarm reset within 6 min if the CO concentration exceeds 70 ppm, secondary power supply considerations, and alarm tests point revision (revised November, 2001—see Table 13.2). Usage in recreational vehicles/marine units and unconditioned areas was added in 1997.7 The revised UL 2034 standard is similar to that of the International Approved Services (IAS) and the British Standards Institute (BSI). While these new standards and new sepcifications for CO detectors facilitate the design and development of newer and better CO detectors, the potential impact these CO detectors can have on public health is limited by the prevalence of their use.

TABLE 13.2 UL 2034, revised November, 2001. Carbon Monoxide Concentration Versus Time for Alarm Test Points Based on 10% Carboxyhemoglobin (COHb) A. Carbon Monoxide Concentration (ppm) and Response Time Concentration, ppm Response time, min 70 ± 5 60–10 150 ± 5 10–50 400 ± 10 4–15 B. False Alarm-Carbon Monoxide Concentration Resistance Specifications Concentration, ppm Exposure time, (no alarm) 30 ± 3 30 days 70 ± 5 60 min

8417: “8417_c013” — 2007/9/11 — 12:15 — page 309 — #5

Carbon Monoxide Poisoning

310

This problem can potentially be remedied by CO detector ordinances set at the state and local levels, since the ordinances already in place have had a demonstrable impact on public health.

References 1. Cobb, N. and Etzel, R. A. Unintentional carbon monoxide-related deaths in the United States, 1979 through 1988, JAMA, 266, 659, 1991. 2. Shepherd, G. and Klein-Schwartz, W. Accidental and suicidal adolescent poisoning deaths in the United States, 1979–1994, Arch. Pediatr. Adolesc. Med., 152, 1181, 1998. 3. Standard for Safety UL 2034 Single and Multiple Station Carbon Monoxide Detectors, Underwriters Laboratories, Northbrook, IL, 1st ed., 1992. 4. Hrones, T. L. and Patty, P. E. Carbon monoxide detectors: protection against the silent killer, In Poisoning and Toxicology Compendium, Leikin, J. B., and Paloucek, F. B., eds., Lexicomp, Hudson, Ohio, 1998, 630. 5. Barbe, J. M., Canard, G., Brandes, S., Jerome, F., Dubois, G., and Guilard, R. Metallocorroles as sensing components for gas sensors: remarkable affinity and selectivity of cobalt(III) corroles for CO vs. O2 and N2 , Dalton Trans., 1208, 2004. 6. Coburn, R. F., Forster, R. E., and Kane, P. B. Considerations of the physiological variables that determine the blood carboxyhemoglobin concentration in man, J. Clin. Invest., vol. 44(11), 1899–1910, 1965. 7. Standard for Safety UL 2034 Single and Multiple Station Carbon Monoxide Detectors, Underwriters Laboratories, Northbrook, IL, 2nd ed., 1996. 8. City Council of the City of Chicago: Amendment of Title 13, Chapter 64 of the Municipal Code of Chicago by addition of new sections 190 through 300 requiring CO detectors in various buildings. Meeting of March 2, 1994. 9. Leikin, J. B. Carbon monoxide detectors and emergency physicians, Am. J. Emerg. Med., 14, 90, 1996. 10. Bizovi, K. E., Leikin, J. B., Hryhorczuk, D. O., and Frateschi, L. J. Night of the sirens: analysis of carbon monoxide-detector experience in suburban Chicago, Ann. Emerg. Med., 31, 737, 1998. 11. Eversole, J. M. Carbon Monoxide Detector Ordinance: Review of the Chicago Fire Department experience, In Poisoning and Toxicology Compendium, Leikin, J. B., and Paloucek, F. B., eds., Lexicomp, Hudson, Ohio, 1998, 633. 12. Leikin, J. B., Krenzelok, E. P., and Greiner, T. H. Remarks to the Illinois House of Representatives Executive Committee hearing regarding state carbon monoxide detector act (HB 603), J. Toxicol. Clin. Toxicol., 37, 885, 1999. 13. Barret, L., Danel, V., and Faure, J. Carbon monoxide poisoning, a diagnosis frequently overlooked, J. Toxicol. Clin. Toxicol., 23, 309, 1985. 14. Heckerling, P. S., Leikin, J. B., and Maturen, A. Occult carbon monoxide poisoning: validation of a prediction model, Am. J. Med., 84, 251, 1988. 15. Clifton, J. C. N., Leikin, J. B., Hryhorczuk, D. O., and Krenzelok, E. P. Surveillance for carbon monoxide poisoning using a national media clipping service, Am. J. Emerg. Med., 19, 106, 2001. 16. Krenzelok, E. P., Roth, R., and Full, R. Carbon monoxide: the silent killer with an audible solution, Am. J. Emerg. Med., 14, 484, 1996. 17. Yoon, S. S., Macdonald, S. C., and Parrish, R. G. Deaths from unintentional carbon monoxide poisoning and potential for prevention with carbon monoxide detectors, JAMA, 279, 685, 1998.

8417: “8417_c013” — 2007/9/11 — 12:15 — page 310 — #6

Carbon Monoxide Detectors as Preventive Medicine

311

18. Tomaszewski, C., Lavonas, E., Kerns, R., and Rouse, A. Effect of a carbon monoxide alarm regulation on CO poisoning, J. Toxicol. Clin. Toxicol.,41(5), 167–168, 2003. 19. Lavonas, E., Tomaszewski, C., Kerns, W., and Blackwell, T. Epidemic carbon monoxide poisoning despite a CO alarm law, J. Toxicol. Clin. Toxicol.,41(5), 711–712, 2003. 20. Runyan, C. W., Johnson, R. M., Yang, J., Waller, A. E., Perkis, D., Marshall, S. W., Coyne-Beasley, T., and McGee, K. S. Risk and protective factors for fires, burns, and carbon monoxide poisoning in U.S. households, Am. J. Prev. Med., 28, 102, 2005.

8417: “8417_c013” — 2007/9/11 — 12:15 — page 311 — #7

8417: “8417_c013” — 2007/9/11 — 12:15 — page 312 — #8

14

Misconceptions About Carbon Monoxide David G. Penney

CONTENTS 14.1 Properties, Presence, and Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Physiology of Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.3 Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.4 Treatment and Outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

313 316 318 320 322 323

14.1 PROPERTIES, PRESENCE, AND DETECTION There are many many misunderstanding about carbon monoxide (CO), even today, by the general public, healthcare professionals, and others who ought to be better informed. The first grouping of these misconceptions are shown in Table 14.1. CO is known as the “Silent Killer” or the “Stealthy Poison” because it is impossible to detect by humans—we cannot smell, taste, or see CO gas, no matter what the concentration. For this reason gas supply companies often put smelly tracers in fuel gases to assist us in detecting their presence. With school groups and nonscientists I often refer to CO as a “smart poison,” as opposed to “dumber poisons” like hydrogen sulfide and others that we have warning of by our chemoreceptors (i.e., taste, smell). CO is also smart in that it enters our bodies simply by our breathing, and does not have to be “picked up” by ingestion through the mouth as lead largely is. Another element of “smartness” is that it leaves the body quickly, although damage may already have been done, while dumb poisons remain (lead, mercury, etc.) to be detected days, weeks, even years later. CO is of course slightly lighter than air, being made up of one carbon atom and one oxygen atom. It has about the dimensions of an oxygen molecule (i.e., diatomic oxygen), contributing to several others of its properties (i.e., ability to diffuse easily, attach to hemoglobin where oxygen sits). The molecular weight of CO is roughly 28 (see Reference Data, Chapter 35). Calculating the molecular weight of air, a mixture of nitrogen (28), oxygen (32), argon (40), carbon dioxide (44), and water vapor (18), gives a weighted average of approximately 29. Thus, CO is about 4% lighter than

313

8417: “8417_c014” — 2007/9/11 — 12:15 — page 313 — #1

314

Carbon Monoxide Poisoning

TABLE 14.1 Misconceptions: Properties, Presence, and Detection • • • • • • • • • • • • • •

CO is easy to detect. CO is lighter than air and therefore rises (to the ceiling) and stays there. CO is not combustible. CO adsorbs and absorbs to fabric, crockery, walls and thus remains long after it has left the air. CO and natural gas are the same thing. Natural gas contains considerable amounts of CO. You can always tell if CO is present because of a peculiar odor that is present. A brand new, well designed, perfectly “tuned” heating/cooking device cannot produce toxic/lethal amounts of CO. Diesel engine exhaust never contains enough CO to cause harm. Heating, ventilation, and air conditioning (HVAC) and gas company service personnel always check for CO when performing maintenance/service on home heating systems. CO will be detected immediately by service personnel if it is present in a home heating system. When your home CO detector shows low levels of CO, it is probably just an instrument malfunction. Cracks in heat exchangers are responsible for the production of CO. Home CO detectors/sensors are the best devices to ferret out CO because they react to very low levels of the gas.

air. However, the assertion that CO rises to the ceiling when released into the air of a structure is incorrect. The CO is always mixed with vast quantities of excess air, mixes thoroughly and completely, and cannot unmix lest it violate the Second Law of Thermodynamics. The key to observations of higher CO concentration near ceilings stems from the fact that hot air rises, and exhaust containing CO is often warmer than the surrounding air. CO is of course combustible and has been used as a fuel gas. It will burn to carbon dioxide, which it does with the release of additional heat. CO was used as a fuel source in past years, for both lighting and heating. This property should be kept in mind when working around high concentrations of CO, where ignition may occur. We may forget about this property because usually the toxic effects are of greatest significance at far lower concentrations. CO does NOT attach in any way to fabric, glass, crockery, and so forth. I heard that a physician when seeing a patient who had sustained acute CO poisoning, told his patient’s mother to clean everything in the apartment after the incident because CO “sticks.” She did this as well as giving away all the furniture and clothing that had been in the apartment. This was of course entirely needless; the result of inadequate education in toxicology. Natural gas as it is used in the United States, contains little CO. Its main component is methane, CH4 . I understand there are also small amounts of other gases too, of the aliphatic series (e.g., ethane). CO may only be present around a combustion device using natural gas after the gas is combusted, and when that combustion process is incomplete, that is, not all of the gas is burned to form carbon dioxide and water vapor. Other possible incomplete combustion products include soot (i.e., elemental carbon).

8417: “8417_c014” — 2007/9/11 — 12:15 — page 314 — #2

Misconceptions About Carbon Monoxide

315

All common combustion devices involve some amount of incompelete combustion, although under certain circumstances this may be very small. So, all such devices produce some CO. Usually the concentration of CO generated is below levels set as safe by regulatory standards. But even these devices may under certain conditions produce extraordinarily high, even lethal concentrations of CO if something else in the heating system is inadequate or fails (compromised combustion air source, broken exhaust system), or if slow or catastrophic failure of the heating device itself takes place through corrosion, lack of cleaning, physical impact by outside forces, and so forth. Diesel engines generally produce lower CO concentrations in exhaust gases than gasoline-fueled engines. On the other hand, the Environmental Protection Agency (EPA) over the past several decades has required gasoline-fueled vehicles in the United States to have catalytic converters that work on the exhaust gases. This device converts the CO to carbon dioxide using various catalysts. They only work effectively when they are hot, that is, they are not effective at cold start. The key point here is that diesel engine equipped vehicles lack catalytic converters, so ALL the CO generated by the engine goes out the tailpipe. Thus, the CO emissions from many diesel powered vehicles are above those from gasoline-fueled vehicles. Heating and cooling technicians and gas company personnel should always monitor for CO when checking a combustion device. Unfortunately, this is not always done. “Sniffing” for natural gas or propane (LPG) is more common, to detect leaks that could lead to fire or explosion. Checking for CO that can lead to injury and/or death is sadly often not done. Unfortunately, even when CO is checked, a problem may be missed. A high CO reading is proof-positive of a problem with a combustion device or the system it is operating in. Failure to find CO on one attempt does not entirely rule out the possibility of a problem during that time frame and under slightly different environmental conditions. Of course measurements must be made when the combustion device is operating, and for sufficient time to catch possible delayed build-up. Often measurements will have to be made several times, and under slightly different conditions, for example, when other combustion devices are also operating, when exhaust fans are being run, when outside wind conditions change, and so forth. One of the most common, possibly lethal mistakes made by people is to assume that a residential CO alarm (i.e., detector) showing low, or even high levels of CO, is malfunctioning. “It never did before,” “it must be the battery,” “The smoke alarm is not sounding,” “take the battery out and go back to bed.” I recommend that people have a minimum of two CO alarms in their residence, one located very near where they sleep, and another near where they are when awake, reading, watching TV, and so forth. Two alarms sounding simultaneously make it much less likely that a person will come to the conclusion that the warning is due to malfunction. If there are two or more floors to the residence, an additional alarm should be installed for each additional floor. Home CO alarms are not the best devices to discover where CO is coming from in a breathing space. After all, they are built as “alarms,” not as instantaneous sensing devices. They have mathematical algorithms built in to simulate human CO uptake, and thus to provide warning before we have taken very much of the CO up into our blood and tissues. Also, their cost is minimal, making them affordable to the general

8417: “8417_c014” — 2007/9/11 — 12:15 — page 315 — #3

Carbon Monoxide Poisoning

316

public, but by no means professional grade in accuracy. The best devices to ferret out CO sources and monitor it accurately are what we call “monitors.” They measure CO almost instantaneously and produce a reading in parts per million (ppm). They are usually portable and the readings are highly reproducible. As one might imagine, these kind of devices are more costly than home CO detectors.

14.2 PHYSIOLOGY OF CARBON MONOXIDE CO binds to hemoglobin reversibly (Table 14.2). While this binding is avid, as we know it is some 230 times that of oxygen, it is nonetheless a competitive process with oxygen. As the oxygen partial pressure rises, the CO is forced off the site on the hemoglobin molecule and oxygen takes its place. This mechanism is one of the major reasons for using hyperbaric (high pressure) oxygen in treating CO poisoning. CO-induced hypoxia is far more serious than hypoxic hypoxia. The presence of CO in the blood occupying the oxygen sites on the hemoglobin decreases the oxygen transport capability of the cardiovascular system, not unlike increasing high altitude does by desaturating arterial blood as lung partial pressure of oxygen falls. In addition to this, however, CO causes the remaining oxyhemoglobin to hold its oxygen more strongly, thus shifting the oxygen dissociation curve (ODC) to the left and making it harder to unload what oxygen remains in the blood to the tissues. People tend to forget this second action of CO on the ODC. Hypoxic hypoxia does just the opposite, shifting the ODC to the right and enhancing the unloading of oxygen to the tissues. Thus, CO inflicts a “double-whammy” on the body in terms of oxygen delivery capability that is much more serious than is that of hypoxic hypoxia. CO poisoning is also far more serious than a comparable simple anemia, where it may appear that there is the same amount of total hemoglobin able to carry oxygen. Like hypoxic hypoxia, anemia causes a right-shift in the ODC, which enhances oxygen

TABLE 14.2 Misconceptions: Physiology • • • • • • • • •

CO binding to hemoglobin is irreversible. CO (caused) hypoxia is no more serious than any other type of hypoxia. CO poisoning is no more serious than an anemia in which there is a comparable amount of hemoglobin able to carry oxygen. Small animals (birds, mice, etc.) die more quickly because their hemoglobin binds CO more avidly than that of humans, thus they were used as alarms for CO in mines. The fetus is protected from CO by the maternal body. Good COHb measurements can be obtained 1 day to a week after a person leaves the site of the CO poisoning. Breathing “clean” air for 2–3 h will eliminate all CO from the body. Breathing 100% oxygen for 20–30 min will eliminate all CO from the body. Breathing (filter) masks protect the wearer from the inhalation of CO.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 316 — #4

Misconceptions About Carbon Monoxide

317

unloading to the tissues, whereas CO does the opposite. A condition where the body has only one-half of its normal hemoglobin concentration (i.e., anemia) is welltolerated, whereas a condition where half of the hemoglobin has CO attached [i.e., 50% carboxyhemoglobin (COHb) saturation] is usually quickly lethal. Small birds and mammals were at one time used as CO alarms in mines and other closed places where CO occurred. Canaries and other like small birds were especially popular in this capacity because they would fall off their perches when incapacitated, making it obvious to the observer there was a problem. The reason for their usefulness was not that the small birds or mammals were necessarily more sensitive to CO, but because their warm-bloodedness and high surface to volume ratio caused their metabolic rate to be necessarily much higher than that of humans, making their ventilation rate greater than that of humans. Hence they take up the CO much faster. This provided a living “early warning” system for CO. The fetus residing inside the maternal body is subject to the CO breathed by the mother. Being a nonirritating gas, unlike chlorine or nitrogen dioxide, it is taken up silently and without notice through the maternal lungs. The CO reaches the placenta through the circulating maternal blood, and again, it passes silently and readily across the membranes into the fetal circulation where it attaches to the fetus’ hemoglobin. While CO compromises oxygen delivery to tissues in the maternal circulation by the mechanisms described above, the situation is much more serious for the fetus, which normally is operating in an oxygen-depleted environment. The presence of CO makes the situation worse by further degrading the oxygen carrying capacity of the fetus. The right time to draw blood for accurate measurement of COHb is one of the biggest misunderstandings by lay people, and even members of the medical community. The half-life of COHb in humans breathing sea level ambient air centers around 4–5 h. That is, one-half of the COHb will be gone in 4–5 h, and half of what remains will be gone in another 4–5 h (i.e., two half-lives). Within one day, COHb level in the blood will be at or near background (0.4–1.4%) no matter how high it was to start with. For accurate measurement, blood sampling should be done within 2–4 h of leaving the site of the CO poisoning. Physicians must not say when contacted by a patient on Friday or on the weekend, “just come to the office on Monday and we’ll do the COHb test.” It will be too late and the results will be useless. On the basis of what we know of COHb half-life, it is clear that breathing “clean” air for 2–3 h. will not eliminate all the CO from the body—approximately 24 h is required. In the same way, breathing 100% oxygen for 20–30 min at sea-level pressure, known as normobaric oxygen therapy, will not eliminate all the CO from the body, since the half-life of COHb when breathing 100% oxygen under normal conditions centers around 50–70 min. A number of hours will be required to do this. Another misconception involves the use of masks in polluted environments. Dust masks or more complicated filtering respirators will remove particulates of various sizes. Virtually none of these devices will remove CO. Donning a mask that does not remove CO when the wearer believes it will, can be a very dangerous, even deadly mistake. In order to safely enter an atmosphere containing CO, self-contained breathing apparatus (SCBA) is required, such as that used by firefighters and hazmat personnel.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 317 — #5

Carbon Monoxide Poisoning

318

14.3 SYMPTOMS While CO poisoning is classically described as causing the skin, nail beds, and exposed mucous membranes to turn pink or bright red, this is rarely seen in actual fact. Often dead CO victims appear gray or yellow. The COHb saturation must be high enough for the COHb to show; people dying in the water from acute CO poisoning usually reach only 40–45% COHb, the level at which incapacitation occurs. Breathing then ceases below the water’s surface. The pink color is of course difficult to see in darkly pigmented people (i.e., negroids), those with severe sunburn, and so forth. (see Table 14.3) Fever is rarely associated with CO poisoning, especially chronic CO poisoning. Nonetheless, thermoregulatory dysfunction is a common outcome of CO poisoning, and is thought to be due to brain damage caused by the poison. Victims with this condition usually feel cold in thermally neutral surroundings, and occasionally report being hot, however, in the latter instance actual body temperature is rarely above normal. Fever is reported following severe acute CO poisoning, probably associated with cerebral tissue damage and edema, gastrointestinal damage, and so forth. The lungs along with the nasal passages, throat, and trachea are generally unaffected by CO. Most of the CO uptake by the body occurs through the former route, and the CO goes directly into the blood. While congestion, cough, and hoarseness are not caused by CO inhalation, other substances that often accompany CO, particulates, nitrogen oxides, sulfur oxides, aldehydes, and so forth will do this. A recent book on CO (Dwyer et al., 2003, p. 5)1 lists “wheezing” or bronchial constriction” and “persistent cough” as “signs and symptoms of CO poisoning.” In my experience this is incorrect. Such symptoms may be due to other components of

TABLE 14.3 Misconceptions: Symptoms • • • • • • • • • • • • • • •

The skin, nail beds, and so forth of people with CO poisoning are invariably red or pink in color. Fever is a common symptom of CO poisoning. Nasal congestion, cough, and hoarseness are symptoms of CO poisoning. The lungs are inflammed by low to moderate levels of CO and appear abnormal by x-ray. Hyperventilation is a response to low and moderate CO poisoning Symptom clusters involving prolonged headache, dizziness, nausea, and fatigue of the whole family should be blamed on viruses, bad food, or group craziness. Everyone responds to CO in the same way, that is, all people show the same symptoms. Depression is not a residual effect of CO poisoning. Loss of short-term memory capability is not a residual effect of CO poisoning. Muscle and/or joint pain is not a residual effect of CO poisoning. Difficulty with attention and concentration is not a residual effect of CO poisoning. Blurry vision is not a residual effect of CO poisoning. Personality change is not a residual effect of CO poisoning. More people experience acute CO poisoning than the chronic type There is a good dose–response relationship between CO in the air and COHb and immediate symptoms or long-term health damage.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 318 — #6

Misconceptions About Carbon Monoxide

319

exhaust gases (e.g., Particulates, aldehydes, sulfur, and/or nitrogen oxides), but not to the CO specifically. As noted above, the lungs are virtually transparent to CO. In life-threatening CO poisonings, the lungs become edematous and begin to fill with fluid. This congestion is coughed up as a light to more darkly tan fluid, often mistaken for vomitus. It is usually observed at scenes of lethal CO poisonings, combined or not combined with vomitus. Tachypnea is not a symptom of low to moderately severe CO poisoning, since the carotid chemoreceptors are insensitive to percent saturation of the arterial blood, responding only to changes in arterial PO2 and pH, which generally change little until the CO poisoning becomes very severe. That is why hyperventilation with simple CO poisoning is so rare, and its presence suggests a more complicated poisoning. As discussed at length earlier, CO poisoning usually presents with a number of nonspecific symptoms; few if any of the symptoms are specific (i.e., pathognomonic) to CO poisoning. We often speak of symptom clusters, since CO usually induces symptoms involving so many organ systems. “Flu-like” is the way the symptoms are often described. Yet the “flu” may continue for weeks or months, affect everyone in the same breathing space (i.e., house, apartment) at the same time, subside when an individual leaves the space, etc. characteristics which are very “un-flu-like.” Historically, chief among misdiagnoses, has been viral or bacterial flu, food poisoning, psychosomatic behavior, and so forth. Sometimes medical providers will continue to insist on traditional causes for weeks and months, when they are clearly impossible, and fail to recognize the ear-marks of a site-specific poisoning. See my chapter on misdiagnosis of CO poisoning (Chapter 19). Although there is a commonality of responses to CO, not all people respond exactly the same, or have the same overall sensitivity to CO. This is the basis for the problems with tables of symptoms seen in articles and books on CO, as they occur at increasing concentrations of CO or COHb. I’ve often said that “people are not lab rats.” We vary in age, gender, race, height, body weight, and so forth and in ways that are difficult to describe or at present unknown, that influence our sensitivity (or tolerance) to CO. In any group of people exposed to what appears the same CO concentration for the same period of time, a diversity of responses develop. Some individuals suffer few immediate or even long-term health effects, while others can be severely affected, or even die during the initial CO exposure. Then there is a third group who respond in an intermediate manner, with mild initial symptoms and mild long-term effects. Thus, we usually see a “normal distribution” of responses, approximating a bell-shaped curve. This pattern appears to occur whether the CO poisoning is extremely mild, or extremely severe. Clinical depression is seen with a high frequency in people who have suffered long-term health damage from CO exposure. This is documented by the Utah group, principally Dr Ramona Hopkins (see chapter 22 in this book). The depression appears to have at least two components, that caused by the realization of the loss of functionality by the CO-injured individual, and by direct CO-induced damage to the brain. The latter pathway is often overlooked by those health professionals unfamiliar with CO poisoning. The depression is usually treatable with standard antidepressant medications.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 319 — #7

Carbon Monoxide Poisoning

320

Along with chronic fatigue, decrement in short-term memory capability is probably the most common long-term damage caused by CO poisoning. It is described by victims as involving problems in remembering appointments and what was already said in a conversation, losing keys, wallet/purse, finding the right word, and so forth. It is embarrassing, affects a person’s confidence and self-esteem, and often forces that person to become less social, and even a recluse. Damage to longterm or remote memory, is far less common. In fact, it is rarely seen by this author, although it is reported in the literature. In the CO Support Study (see Hay et al., Carbon Monoxide Toxicity, 2000),2 muscle and joint pain were the most common long-term outcome of CO poisoning. My studies of chronic CO poisoning in particular, verify the frequency of these symptoms. Because clinical testing of the complaining muscles and joints rarely show evidence of local damage, it is my working hypothesis that the damage is wholey in the brain, and that the pain is referred, ie. sited in the brain. I don’t suggest that it is imagined or fabricated, that it is in fact real, but is simply another product of the CO-induced brain damage, not unlike that that causes the well-recognized cognitive and memory deficits. Like short-term memory damage, problems with attention–concentration are extremely common after CO poisoning. People become more distractible, with even minor auditory or visual distractions. Combined with this, and closely related to attention–concentration, is usually a vastly reduced ability to do more than one task at a time (i.e., multitasking). Having to have the TV off when talking on the phone, the radio off when entering a freeway, and having to study in a totally quiet environment, is often reported. While blurry vision is a common complaint during CO poisoning, it can continue for some time after poisoning has ended. It is one of several dozen visual complaints that may be persistent and residual. See Dr Helffenstein’s chapter in this book (Chapter 23). One of the most common statements by family and friends of a CO victim is that he/she is no longer the same—there has been “a personality change.” This usually occurs in the direction of increased irritability, anxiety, depression, apathy, and so forth. CO poisoning does not cause personalities to improve! I have written elsewhere (Chapter 1) that the most frequent kind of CO poisoning is the chronic type, that it is the least likely to be diagnosed, and probably results in the largest amount of total injuries. New studies make it clear that brain damage from CO poisoning is only very poorly correlated with severity of poisoning, whether that be gauged by air CO concentration, COHb saturation, loss of consciousness, and so forth.

14.4 TREATMENT AND OUTCOME I often see cases where nasal prongs were used following CO poisoning. Just because the oxygen in the tank is 100%, that doesn’t mean the oxygen coming through the nose to the lungs will be 100%. It could be just a bit over 21%, that is, room air, owing to dilution with copious amounts of fresh air (see Table 14.4).

8417: “8417_c014” — 2007/9/11 — 12:15 — page 320 — #8

Misconceptions About Carbon Monoxide

321

TABLE 14.4 Misconceptions: Treatment and Outcome • • • • • • • • •

Inhalation of 100% oxygen from a rebreathing mask or from nasal prongs are the best immediate means of removing CO from the body. Victims of CO poisoning should be released from medical care immediately following 1–2 h of oxygen treatment, whether or not their symptoms have disappeared. There is no need for repeat COHb measurements, psychometric tests, or other clinical tests following medical treatment for CO poisoning. People who recover from CO poisoning are always completely normal. Use of 100% oxygen therapy at normal atmospheric pressure (NBO) is a proven approach to eliminate the residual effects of CO poisoning. Use of 100% oxygen at increased atmospheric pressure (HBO) is a proven approach to eliminate the residual effects of CO poisoning. CO exposure never produces brain damage unless there is a period of unconsciousness. Low/moderate CO exposure cannot produce brain damage or significant changes in functional performance. Venous blood measurement of COHb is not as accurate as arterial.

CO poisoning victims should be kept in-hospital for observation, usually for a minimum of one night. All too often patients are seen for 2–3 h, given oxygen, then released and continue to be symptomatic—headache, nausea, dizziness, and so forth. Patients should never be discharged until they are completely symptom-free. An extra night in the hospital gives staff a chance to observe them. Delayed sequelae are such well-known complications of CO poisoning, even in patients who appear to be doing very well patient. Discharge documents should clearly warn of possible delayed effects, that is, sequela and how to recognize them. CO washout rates vary tremendously among people, even when body mass, gender, age, and so forth are taken into consideration. Second COHb measurements should be done to confirm that normal or near-normal COHb levels exist after treatment. This is an area that needs great improvement. COHb should be measured increasingly by the newer methods (breathlyzer, multiwavelength pulse-oximeters). Their advantage is that they can be done as often as necessary until COHb is sufficiently reduced. Psychometric testing should also be done on a more frequent basis in the emergency room (ER), especially in those patients who presented with symptoms of altered consciousness or gait/balance problems. It is recognized that a fraction of people who recover from CO poisoning are left with decrements in function. Mainline medicine has been slow to recognize the physical symptoms that often persist, and also the sensory, motor, cognitive, and psychological deficits that many patients are left with, instead only chosing to focus on the gross neurologic problems. The use of 100% normobaric oxygen (NBO) has been used for many years to treat CO poisoning. On a mechanistic basis, it should be efficacious simply because it more quickly removes CO from the body than breathing room air. In recent studies, it was the therapy to which hyperbaric oxygen (HBO) treatment was compared. The use of

8417: “8417_c014” — 2007/9/11 — 12:15 — page 321 — #9

Carbon Monoxide Poisoning

322

NBO is proven only in the sense that aside from HBO, it appears to be the best approach available. See the chapters by Drs Tomaszewski (Chapter 17) and Scheinkestel and Millar (Chapter 18) on the use of NBO and HBO therapies. Studies during the past 20 years make it abundantly clear that loss of consciousness is not required for development of neurologic sequelae. Even more recent studies demonstrate the lack of correlation between brain damge and air CO concentration, COHb saturation, and many other traditional markers of poisoning severity. For the measurement of blood CO alone, a venous sample is just as good as an arterial sample. The use of noninvasive methods (breathylizer, new generation pulse-oximeters) for this purpose is encouraged.

14.5 MISCELLANEOUS Medical students and residents receive virtually no training in the diagnosis and treatment of CO poisoning, and usually have little or no contact with CO-poisoned patients in their practice (see Table 14.5). This is a large part of the reason for the high misdiagnosis rate of CO poisoning by physicians. As a result few physicians have an adequate index of suspicion for CO poisoning when presented with appropriate symptoms and when they take a situational history. The gold standard for the cognitive-memory problems that occur with such a frequency following CO poisoning is neuropsychological testing. This should be done by a neuropsychologist who understands the pathology of CO poisoning and who has evaluated many patients with this condition. Psychiatry is well-suited to assist medically in the treatment of emotional and personality problems that often arise after CO poisoning. Neurology can be brought to bear for patients with gross neurologic problems (i.e., aphasia, gait disturbance) caused by CO poisoning. It must

TABLE 14.5 Misconceptions: Miscellaneous • • • • • • • • •

Physicians receive adequate training in the diagnosis and treatment of CO poisoning in medical school and residency/fellowship. Physicians get adequate experience with CO poisoning in treating their patients. Physicians generally have a high enough index of suspicion relative to CO poisoning in order to diagnose it reliably. Psychiatrists and neurologists are the medical professionals of choice to determine the extent of central nervous system (CNS) damage caused by CO. High-tech imaging devices (CT, MR, SPECT) always show areas of brain damage from CO poisoning, if they exist. Two wavelength pulse oximeters are excellent devices for determining oxygen saturation and also whether a person is suffering from CO poisoning. In environments containing CO, the levels of CO2 , oxygen and other gases are unimportant in the degree of poisoning. The presence of green plants in a closed space will remove CO from the atmosphere. Toxicology is central to medicine.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 322 — #10

Misconceptions About Carbon Monoxide

323

be kept in mind however, that professionals in neither of these disciplines of medicine have training in toxicology. Computed tomography [CT, Computerized Axial Tomography (CAT)] is a form of computer-enhanced x-ray scanning. Magnetic resonance imaging (MRI) shows internal anatomy without the use of ionizing radiation. Both of these techniques allow us to look at structure only, not function. Single photon emission computed tomography (SPECT) permits the imaging of function in living tissue (see Chapters by Drs. Heuser [20] and Hipskind [21]). Pulse oximeters currently in use are “blind” to COHb, thus giving incorrect values for oxygen saturation in the presence of COHb. These pulse oximeter devices should never be used when there is suspicion of CO poisoning. In that case arterial blood should immediately be drawn for measurement of total blood gases, which includes COHb and oxygen saturation. See the chapter by Dr. Hampson [33] and mine [19], regarding the new generation of pulse CO-oximeters marketed by Masimo. Elevated carbon dioxide concentrations causes hyperventilation and an increased rate of CO uptake. Depressed oxygen (i.e., depletion) does the same, and also increases the final COHb saturation by shifting the Haldane equation. A recent article on residential CO poisoning published in the January, 2007 issue of “Woman’s Day” magazine, states that the presence of green plants will remove CO from the atmosphere. This was reportedly told to a patient by a treating physician. There is in fact no basis for this statement. Maybe he was confused by the fact that plants take up carbon dioxide (not carbon monoxide), converting it to oxygen through the process of photosynthesis. Toxicology might be likened to a poor relation to medicine. For most physicians the normal circles of training and practice experience in internal medicine rarely intersect with the toxicology circle. Thus, the constant complaints I receive from CO-victims, saying that few if any physicians they see are helpful in the diagnosis, evaluation, or treatment of their conditions.

References 1. Dwyer, Leatherman, Manclark, Kimball, Rasmussen. Carbon Monoxide: A Clear and Present Danger, ESCO Press, USA, 3rd ed., 2003 (www.escoinst.com). 2. Hay, A.W.H., Jaffer, S., Davis, D. Chronic carbon monoxide exposure: The CO Support study. In: Carbon Monoxide Toxicity, D.G. Penney, ed., CRC Press, NY, 2000, pp. 419–438.

8417: “8417_c014” — 2007/9/11 — 12:15 — page 323 — #11

8417: “8417_c014” — 2007/9/11 — 12:15 — page 324 — #12

15

A Survey Study of Public Perceptions About Carbon Monoxide David G. Penney and Linda M. Penney

CONTENTS 15.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.2 Electric Generator Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.3 Propane Radiant Heater Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.4 Recreational Powerboat Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.5 Conditions Affecting Exhaust Gas Danger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.6 Safe Use Indoors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.7 The Greatest Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.8 The Worst Poison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.9 The Best Advice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.10 Duration of Time in the Body . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.11 Properties of CO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.12 Experience with CO Poisoning? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.13 Use of CO Emitting Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

325 327 329 331 332 333 334 335 336 338 338 338 339 339 339

15.1 BACKGROUND On the basis of the large numbers of accidents involving carbon monoxide (CO), recent studies have asked questions about the public’s knowledge of the dangers of this gas.1 Our survey was conducted in the latter half of 2006 by sampling the opinions of people in Michigan (MI) and Florida (FL). The data derived from this survey are shown in Tables 15.1–15.13. Adult respondents belonged to civic and athletic social clubs. These data were analyzed by age groupings and by gender. Youth data were obtained by surveying students in both junior high and senior high school classes. The Michigan school was a large public school in Benzie County (first set of four rows). In Florida, one school was a large public high school in St. John’s County (first set of four rows), while data from two smaller parochial schools in the county were combined (second set of four rows). The youth data were also analyzed 325

8417: “8417_c015” — 2007/9/11 — 12:15 — page 325 — #1

Carbon Monoxide Poisoning

326

TABLE 15.1 Assume the Electric Power is Off, but Can be Restored by the Use of a Gasoline-fueled, Portable Generator. If Started in a Room, How Much Ventilation is Adequate for the Generator to be Run Safely? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

13.2 28.6 0 12.5 7.1 11.6 13.2 12.1 7.7

10.5 18.4 37.9 22.9 14.3 14.0 10.5 10.8 19.2

26.3 24.5 17.2 27.1 3.6 7.0 5.3 6.0 3.9

26.3 18.4 24.1 31.3 21.4 18.6 36.8 26.5 23.1

23.7 10.2 20.7 6.3 53.6 48.8 34.2 44.6 46.2

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

15.8 21.4 8.3 7.1 22.2 7.7 4.7 2.6 1.4 5.6 2.9 0 4.4

13.2 19.1 0 28.6 22.2 15.4 4.7 15.4 2.8 11.1 5.7 0 8.7

23.7 14.3 0 21.4 0 30.8 9.3 20.5 4.2 16.7 7.1 5.3 8.7

23.7 28.6 41.7 35.8 33.3 15.4 46.5 33.3 15.7 22.2 14.3 31.6 21.7

23.7 16.7 50.0 7.1 22.2 30.8 34.9 28.2 75.0 44.4 70.0 63.2 56.5

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = One window open; B = Two windows open; C = Three or more windows open; D = All windows and the door open; E = None of the above. ∗ “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

by age group and gender. The questions put to respondents in the survey are printed above each table, and the “correct” answer is indicated in bold. The survey asked only for age, gender, and occupation to be written in. Before the survey was administered, respondents were instructed to answer quickly, not to agonize over their responses, and told the survey was not an intelligence test. When we could not be present for administration of the survey, the survey takers (teachers) were provided instructions to be read to the class prior to starting. There was an initial un-numbered fill-in question that asked yes or no, whether respondents (or if students, their parents) had rented or owned various small combustion devices (Table 15.12). A final question on the survey asked respondents about their experience with CO poisoning incidents. This

8417: “8417_c015” — 2007/9/11 — 12:15 — page 326 — #2

A Survey Study of Public Perceptions About Carbon Monoxide

327

TABLE 15.2 Portable Propane Radiant Heaters are Available to Provide Heat, for example, when Camping. Which One of the Situations Below is Safe? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

0 2.0 0 2.1 0 2.3 2.7 2.5 0

7.9 14.3 3.5 4.2 3.6 2.3 10.8 6.2 3.7

34.2 49.0 41.4 66.7 10.7 11.6 13.5 12.4 11.1

18.4 4.1 13.8 8.3 17.9 11.6 24.3 21.0 7.4

39.5 30.6 41.4 18.8 67.9 72.1 48.7 58.0 77.8

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

5.3 0 7.7 0 0 0 2.3 0 1.3 0 1.4 0 0

21.1 18.6 0 15.4 0 7.1 9.1 15.0 0 0 0 0 4.2

44.7 44.2 30.8 53.9 55.6 64.3 36.4 37.5 22.7 55.6 25.0 45.0 16.7

5.3 28.6 15.4 7.7 11.1 0 4.6 5.0 5.3 0 4.2 5.0 8.3

21.7 16.7 46.2 23.1 33.3 28.6 47.7 42.5 70.7 44.4 69.4 50.0 70.8

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Use in a closed tent while sleeping; B = Use in a tent with the door flap loose/open; C = Use for a short time before going to sleep in order to warm up the closed tent; D = Use in a tent with two windows open; E = None of the above. ∗ “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

question had no “correct” answer. Sources of CO involved in CO poisonings were written in on the survey by respondents, and are compiled in Table 15.13.

15.2 ELECTRIC GENERATOR USE (QUEST. 1) In most cases a majority of respondents believed that a gasoline-fueled electric generator could be operated inside a closed space (i.e., room) if some ventilation were provided. This is of course false—no amount of ventilation will make this scenario safe. The generator MUST be operated outside, preferably 20–25 ft. from the nearest structure. In general, adults responded with the highest percent of correct answers,

8417: “8417_c015” — 2007/9/11 — 12:15 — page 327 — #3

Carbon Monoxide Poisoning

328

TABLE 15.3 When Using a Recreational Powerboat, What is Safe (one only)? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

7.9 6.3 6.9 4.4 3.7 0 2.8 1.3 3.9

5.3 2.1 3.5 4.4 0 0 0 0 0

5.3 8.3 0 4.4 0 0 0 0 0

79.0 66.7 89.7 82.6 81.5 97.6 91.7 91.1 92.3

2.6 16.7 0 4.4 14.8 2.4 5.6 7.6 3.9

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

13.2 7.1 6.7 14.3 33.3 28.6 13.6 7.7 2.8 0 1.5 5.3 0

2.6 0 6.7 0 0 0 0 7.7 0 0 0 0 0

7.9 9.5 0 0 0 0 2.3 0 0 0 0 0 0

68.4 73.8 73.3 71.4 55.6 71.4 68.2 82.1 93.1 93.8 94.1 89.5 96.2

7.9 9.5 13.3 14.3 11.1 0 15.9 2.6 4.2 6.3 4.4 5.3 3.9

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Sitting on the transom with the boat engine idling; B = Teak surfing (hanging on to the swim platform behind the boat); C = Wake surfing (body surfing in the wake immediately behind the boat); D = Water skiing 100 ft. behind the boat; E = Sitting on the boat moored near other boats whose engines are idling. ∗ “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

youths the least. People in healthcare-science fields answered correctly no more frequently that other adults generally. On the basis of the responses of many people, and especially so youths, who thought that opening one window will be enough, many of them would not have survived that scenario if it had really happened. In fact, for many of the cohorts of teenagers, the vast majority of the individuals would be at risk for death or injury from CO poisoning from a generator based on their faulty knowledge and perceptions of this gas. People die on a regular basis from CO poisoning because of improper use of generators, and especially after severe storms and hurricanes when main electrical power is not available. The study by Hampson and Zmaeff1 cited a number of incidents of injury and death from improper use of generators following storms and hurricanes. This included

8417: “8417_c015” — 2007/9/11 — 12:15 — page 328 — #4

A Survey Study of Public Perceptions About Carbon Monoxide

329

TABLE 15.4 What Weather Condition Might Greatly Affect the Levels of Exhaust Fumes on Board a Boat? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

13.9 4.2 0 2.0 8.0 2.6 0 2.8 3.7

16.7 10.4 17.2 10.2 8.0 0 2.9 2.8 3.7

8.3 10.4 3.5 6.1 0 7.7 0 2.8 3.7

22.2 31.3 20.7 14.3 28.0 33.3 25.7 27.8 33.3

38.9 43.8 58.6 67.4 56.0 56.4 71.4 63.9 55.6

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

16.2 7.3 7.1 0 22.2 14.3 6.8 7.5 2.9 13.3 4.6 5.6 0

8.1 9.8 14.3 21.4 0 14.3 18.2 12.5 7.1 6.7 9.1 0 0

5.4 9.8 7.1 0 0 21.4 2.3 5.0 4.3 0 3.0 5.6 4.2

37.8 26.8 35.7 50.0 33.3 28.6 29.6 25.0 34.3 26.7 31.8 33.3 25.0

32.4 46.3 35.7 28.6 44.4 21.4 43.2 50.0 51.4 53.3 51.5 55.6 70.8

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Sun out/overcast; B = Air temperature; C = Rain; D = Humidity; E = Wind direction. ∗ “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

26 people in January, 1993 (Washington), 74 in January, 1998 (Maine), 71 in December, 2002 (North Carolina), and 2 in September, 2003 (Hurricane Isabel). Lovanas, at the Centers for Disease Control and Prevention (CDC), found that about one-third of generator incidents involved non-English speaking patients.2

15.3 PROPANE RADIANT HEATER USE (QUEST. 2) A small percentage of respondents believed that it is safe to sleep in a tent with an operating propane radiant heater. While these heaters produce little CO when burning in the open air, they can produce prodigious amounts of CO if oxygen in the ambient air becomes depleted and limiting to complete combustion. A somewhat larger percentage, but usually still far less than a majority, realized that when using these much smaller (i.e., than generators) sources of CO, the secure opening of two windows

8417: “8417_c015” — 2007/9/11 — 12:15 — page 329 — #5

Carbon Monoxide Poisoning

330

TABLE 15.5 Which One of the Following May Safely be Used Indoors? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

5.3 2.1 17.2 6.3 0 2.3 0 0 3.6

76.3 60.4 72.4 68.8 76.9 84.1 84.6 82.7 82.1

2.6 4.2 0 0 0 0 2.6 0 3.6

2.6 0 3.5 2.1 0 4.6 0 2.5 0

13.2 33.3 6.9 22.9 23.1 9.1 12.8 14.8 10.7

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

5.3 9.3 0 0 11.1 0 9.1 10.0 4.1 6.7 2.9 10.0 8.0

73.7 58.1 73.3 78.6 55.6 50.0 79.6 60.0 82.4 66.7 79.4 80.0 72.0

2.6 2.3 6.7 0 0 7.1 4.6 2.5 0 0 0 0 4.0

0 2.3 6.7 0 0 0 0 0 0 0 0 0 0

18.4 27.9 13.3 21.4 33.3 42.9 6.8 27.5 13.5 26.7 17.7 10.0 16.0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Hibachi grill; B = A dozen candles; C = Gasoline-powered generator; D = Gasoline-powered pressure washer; E = I don’t know. *“medical”—all adults, MI and FL, with self-reported medical, biological occupations.

of a tent, guarantees safe operation. This requirement is specified by manufacturers. Many people believed that warming up the tent prior to going to bed was safe. This, in fact, may be one of the most unsafe practices that has resulted in may deaths: (1) CO can accumulate in the closed tent during warm-up, which then poisons the person after he/she comes inside to sleep, or (2) the heater is operated while people are inside the tent in order to heat up, the people then fall asleep, CO accumulates, and death occurs silently. The warmed inside environment, combined with fatigue, CO uptake, and possible alcohol use readily induces sleep. An opened door flap is not safe, because unless it is secured it could inadvertently be brushed closed or be closed by wind. In either case, the end result is predictable. One manufacturer’s propane radiant heaters are known to have killed over 75 people during the past 16 years.

8417: “8417_c015” — 2007/9/11 — 12:15 — page 330 — #6

A Survey Study of Public Perceptions About Carbon Monoxide

331

TABLE 15.6 Which One of Those Below is Known on Average to Represent the Greater Risk of Immediate Injury or Death from Carbon Monoxide Poisoning? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

50.0 38.6 79.3 76.6 76.0 88.6 91.9 88.8 80.8

15.8 6.8 0 8.5 8.0 0 0 2.5 0

7.9 20.5 3.5 4.3 16.0 9.1 8.1 8.8 15.4

2.6 6.8 0 2.1 0 0 0 0 0

23.7 27.3 17.2 8.5 0 2.3 0 0 3.9

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

68.4 73.2 53.9 64.3 55.6 64.3 65.1 59.0 77.5 62.5 75.0 72.2 96.0

5.3 4.9 15.4 14.3 0 0 7.0 2.6 2.8 12.5 2.9 11.1 0

15.8 9.8 15.4 7.1 22.2 28.6 11.6 15.4 14.1 18.8 14.7 16.7 4.0

2.6 0 7.7 0 11.1 0 9.3 10.3 0 6.3 1.5 0 0

7.9 12.2 7.7 14.3 11.1 7.1 7.0 12.8 5.6 0 5.9 0 0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = An automobile; B = A gasoline-powered lawn mower; C = A ski-boat with inboard gasoline engines; D = A snow thrower; E = An electric water heater *“medical”—all adults, MI and FL, with self-reported medical, biological occupations.

15.4 RECREATIONAL POWERBOAT SAFETY (QUEST. 3) While the vast majority of respondents believed that the safe activity was water skiing 100 ft. behind the powerboat, some respondents believed that sitting on the boat transom (the side to side structure at the back of a ship or boat) with the engine idling, teak surfing (holding onto the swim platform and being dragged through the water), wake surfing (being pulled behind the boat on a board via a very short rope), and sitting on a boat near other boats whose engines are idling was also safe. All four of the latter behaviors of course expose individuals to the potential for injurious or lethal CO poisoning. In general, younger people were more likely to endorse these

8417: “8417_c015” — 2007/9/11 — 12:15 — page 331 — #7

Carbon Monoxide Poisoning

332

TABLE 15.7 Which Poison Debilitates or Accidentally Kills the Most People Each Year in the USA? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

0 6.3 0 2.1 0 2.3 0 1.2 0

0 2.1 0 0 0 0 0 0 0

97.4 83.3 92.9 79.2 100 97.7 97.4 97.5 100

2.6 2.1 3.6 8.3 0 0 0 0 0

0 6.3 3.6 10.4 0 0 2.6 1.2 0

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

2.6 4.9 7.1 0 0 0 4.6 2.5 1.4 6.7 3.0 0 0

5.3 4.9 0 7.1 0 0 4.6 5.0 0 0 0 0 0

68.4 78.1 85.7 85.7 100 78.6 81.8 87.5 90.0 93.3 91.0 88.2 96.0

5.3 7.3 0 0 0 0 4.6 5.0 0 0 0 0 0

18.4 4.9 7.1 7.1 0 21.4 4.6 0 8.6 0 6.0 11.8 4.0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Chlorine; B = Hydrogen sulfide; C = Carbon monoxide; D = Cyanide; E = Carbon dioxide *“medical”—all adults, MI and FL, with self-reported medical, biological occupations.

risky behaviors, especially, sitting on the transom with the engine running by the Florida youth.

15.5 CONDITIONS AFFECTING EXHAUST GAS DANGER (QUEST. 4) There was considerable confusion among respondents about what conditions will affect exhaust fume distribution on a boat. Greater fractions of adult respondents thought wind direction would be the major factor, while fewer young people made this choice. Approximately one-third of respondents thought humidity would be most important. Some even thought sunshine versus overcast, air temperature, or rain might be major influences. Studies show that wind velocity and wind direction, combined

8417: “8417_c015” — 2007/9/11 — 12:15 — page 332 — #8

A Survey Study of Public Perceptions About Carbon Monoxide

333

TABLE 15.8 If Carbon Monoxide Exposure is Suspected, What is the Best Advice? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

36.8 30.4 37.9 16.7 44.4 34.1 51.3 48.8 25.0

7.9 4.4 3.5 0 0 0 0 0 0

52.6 63.0 51.7 75.0 55.6 65.9 48.7 51.2 75.0

2.6 2.2 6.9 6.3 0 0 0 0 0

0 0 0 2.1 0 0 0 0 0

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

31.6 23.8 21.4 38.5 33.3 21.4 25.0 32.5 35.7 31.3 34.3 38.9 44.0

7.9 2.4 0 7.7 11.1 0 9.1 0 1.4 0 1.5 0 0

52.6 69.1 78.6 46.2 55.6 78.6 59.1 65.0 62.9 68.8 64.2 61.1 56.0

5.3 2.4 0 7.7 0 0 2.3 2.5 0 0 0 0 0

2.6 2.4 0 0 0 0 4.6 0 0 0 0 0 0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Go outside and breathe some fresh air; B = Ignore it and it will go away; C = Go immediately to an emergency center and be evaluated by a physician; D = Call your doctor for an appointment in a few days; E = Take an aspirin or Motrin for headache. *“medical”—all adults, MI and FL, with self-reported medical, biological occupations.

with boat speed and direction, greatly influence the concentrations of CO found on boat decks, in cabins and immediately behind the boat.

15.6 SAFE USE INDOORS (QUEST. 5) A high percentage of respondents recognized the burning of a dozen candles as being the safe choice indoors. Candles are small combustion sources and produce very little CO. Nevertheless, a significant number of people didn’t know the answer (in one instance >40%), and some even endorsed the use of a hibachi or an internal combustion engine indoors. Younger respondents appeared to be the less knowledgeable, and in some cases significant percentages of the youth (>10%) thought very risky, even lethal behaviors were okay (e.g., use of a hibachi indoors).

8417: “8417_c015” — 2007/9/11 — 12:15 — page 333 — #9

Carbon Monoxide Poisoning

334

TABLE 15.9 How Long Does Carbon Monoxide Stay in the Human Body? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

10.5 6.4 10.3 4.2 0 4.6 5.1 3.6 3.7

5.3 2.1 17.2 18.8 3.7 11.4 0 4.8 7.4

10.5 8.5 10.3 4.2 11.1 9.1 5.1 8.4 7.4

15.8 2.1 17.2 6.3 7.4 6.8 10.3 7.2 11.1

57.9 80.9 44.8 66.7 77.8 68.2 79.5 75.9 70.4

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

7.9 11.9 20.0 0 0 7.1 11.4 10.0 5.6 6.7 6.1 5.3 0

15.8 4.8 13.3 7.1 0 0 11.4 10.0 8.5 6.7 10.6 0 12.5

13.2 9.5 6.7 7.1 0 7.1 13.6 5.0 12.7 6.7 12.1 10.5 16.7

10.5 2.4 6.7 7.1 0 0 11.4 7.5 15.5 6.7 16.7 5.3 20.8

52.6 71.4 53.3 78.6 100 85.7 52.3 67.5 57.8 73.3 54.6 79.0 50.0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Once breathed, it stays always; B = For 65 days; C = For 1 week; D = For 1 day; E = I don’t know. * “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

15.7 THE GREATEST RISK (QUEST. 6) The vast majority of respondents viewed the automobile as the greater risk from CO poisoning. This of course is out of line with the fact that automobiles produce only extremely low levels of CO today when properly used and maintained, thanks to the US Environmental Protection Agency (US EPA) and the catalytic converters required at manufacture. Internal combustion engine-driven devices without catalytic converters represent far greater hazards today for the individual user. The greatest of these are those with the largest engines, therefore the greatest CO generating capacity. That would be gasoline-powered powerboats—ski-boats, cabin cruisers, fishing boats, and so forth, especially those with inboard engines. Generally, the bigger the boat and engines, the more dangerous. Only 10–20% of respondents were aware of this fact. A few thought that lawn mowers or snow throwers were most

8417: “8417_c015” — 2007/9/11 — 12:15 — page 334 — #10

A Survey Study of Public Perceptions About Carbon Monoxide

335

TABLE 15.10 Which One of the Following is a Property of Carbon Monoxide? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

23.7 6.4 13.8 18.8 7.1 20.5 7.7 12.1 14.3

5.3 21.3 27.6 25.0 42.9 18.2 38.5 38.6 10.7

34.2 4.3 6.9 2.1 0 2.3 0 1.2 0

21.1 21.3 27.6 20.8 28.6 29.6 25.6 31.3 17.9

15.8 46.8 24.1 33.3 21.4 29.6 28.2 16.1 57.1

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

7.9 9.8 20.0 21.4 22.2 21.4 13.6 15.0 22.2 12.5 18.8 27.8 32.0

23.7 17.1 33.3 14.3 11.1 0 20.5 15.0 25.0 37.5 30.4 16.7 12.0

21.1 19.5 6.7 7.1 0 7.1 9.1 10.0 1.4 0 1.4 0 0

15.8 12.2 20.0 21.4 11.1 7.1 34.1 22.5 27.8 6.3 2.7 33.3 28.0

31.6 41.5 20.0 37.7 55.6 64.3 22.7 37.5 23.6 43.8 27.5 22.2 28.0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = It is much lighter than air; B = It is much heavier than air; C = It has a strong odor of burning materials; D = It is about the same density as air; E = I don’t know. * “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

dangerous in terms of CO. Certainly they too can be dangerous because even though they have much smaller engines, they also lack catalytic converters. The basis for CO risk posed by electric water heaters, which as many as 27% of one student group chose, is unclear. It may represent a total misunderstanding of how CO is generated. School teaching programs should take note of this abberation.

15.8 THE WORST POISON (QUEST. 7) This question makes it clear that while people of all ages may have confusion on other issues relating to CO, high percentages know that CO debilitates and kills people. Instances of this are frequently reported in the newspapers and on TV. Some groups, especially young people chose carbon dioxide. Confusion between these two carbon oxides could be serious. The one we breathe out is not toxic until very high

8417: “8417_c015” — 2007/9/11 — 12:15 — page 335 — #11

Carbon Monoxide Poisoning

336

TABLE 15.11 Have YOU Experienced the Toxic Effects of Carbon Monoxide Poisoning? Percent of Total Male Michigan

Female

Mean Age

A

B

C

D

E

38 — 14–15 — 49 14–15 29 — 16–18 — 48 16–19 20 8 30–59 31 14 60–69 34 7 70–81 85 — 31–81 — 29 30–79 Florida (rows 1–4 public; 5–8 private)

14.4 14.4 16.7 16.8 48.5 64.9 74.2 64.9 62.3

5.3 4.3 13.8 4.2 10.7 11.1 0 6.0 10.3

5.3 4.3 0 2.1 3.6 4.4 5.1 4.8 3.5

60.5 59.6 55.2 75.0 64.3 55.6 79.5 63.9 72.4

5.3 14.9 69.9 8.3 7.1 11.1 5.1 8.4 6.9

23.7 17.0 24.1 10.4 14.3 17.8 10.3 16.9 6.9

38 — 15 — 9 — 44 — 56 16 72 — *20

15.6 15.5 17.1 17.2 13.9 13.5 16.5 16.6 44.8 71.1 50.8 46.9 57.5

7.9 0 0 0 11.1 7.1 4.6 7.5 4.1 6.3 5.8 0 8.0

2.6 0 0 0 0 0 0 0 6.9 0 5.8 5.3 4.0

76.3 71.4 66.7 100 55.6 64.3 77.3 52.5 67.1 75.0 65.2 79.0 64.0

0 4.8 0 0 11.1 0 9.1 7.5 5.5 6.3 7.3 0 16.0

13.2 23.8 33.3 0 22.2 28.6 9.1 32.5 16.4 12.5 15.9 15.8 8.0

— 43 — 14 — 14 — 40 18 2 — 20 *5

Range

14–16 15–16 17–18 17–18 13–15 11–15 16–18 16–18 18–59 60–83 19–83 26–69 26–80

A = Yes, once; B = Yes, more than once; C = No; D = No, but I know of someone who has; E = I don’t know. * “medical”—all adults, MI and FL, with self-reported medical, biological occupations.

levels are reached, while its sibling with one less oxygen atom is deadly at very low concentrations. Science classes in school could do better in teaching these important distinctions.

15.9 THE BEST ADVICE (QUEST. 8) The majority of respondents recognized that the best action to take when CO exposure is suspected, is prompt presentation at an emergency center for evaluation and treatment by a health professional. A small but substantial percentage of respondents believed that simply going outside and breathing fresh air was adequate to conteract the effects of CO. This is the general wisdom of many older individuals in the medical community, even though it is wrong. CO poisoning is not to be trifled with.

8417: “8417_c015” — 2007/9/11 — 12:15 — page 336 — #12

A Survey Study of Public Perceptions About Carbon Monoxide

337

TABLE 15.12 Sources of Carbon Monoxide for Survey Respondents Michigan Automobile—7 In trunk—1 In garage with door shut—1 Exhaust—2 Suicide attempt—2 In house—1 Furnace—4 Heater—2 Pilot light failure—1 Kerosene heater—2 Water heater (electric)—1 Fire (electrical)—1 Plugged chimney—1 Lawn mower—1 Pool heater/AC—1 Gas refrigerator—1 Golf cart—1 Boat—1 Exhaust gases—1 Race car with bad ventilation system—1 U.S. Army tank + 2-1/2 tn truck—1

Florida Automobile—3 Repair facility—1 In garage—1 Exhaust—1 Tractor—1 Gas oven—1 Gas heater—1 Coal heater—1 Heating system—1 Camp fires—1 Fire—1 Fireplace leakage—2 Lawn mower—1 In garage—1 Boat (anchored)—2 Inside cabin—1 Generator—1 Generator during hurricane—1 Commercial diving—1 Drag race—1 In chemistry class—1

TABLE 15.13 Rental and Ownership of Selected Combustion Devices by Survey Respondents in Michigan and Florida, Adults and Juveniles∗ Rental

Ownership

Chain Press, Kerosene Chain Press, Kerosene Saw Washer Heater Generator Saw Washer Heater Generator Michigan Men (85) Women (29) Boys (67) Girls (97) Florida Men (72) Women (20) Boys (106) Girls (111)

4.7 0 16.4 9.3

16.5 6.9 11.9 11.3

1.2 0 7.5 8.3

5.9 0 16.4 8.3

65.9 48.3 68.7 54.6

27.1 20.7 35.8 21.7

17.7 20.7 28.4 18.6

21.2 20.7 37.3 22.7

25.0 0 22.6 12.6

34.7 25.0 34.9 36.0

6.9 0 7.6 1.8

13.9 5.0 21.7 17.1

44.4 25.0 50.0 45.1

31.9 20.0 45.3 33.3

6.9 0 15.1 8.1

27.8 25.0 40.6 36.0

∗ It is assumed that this reflects rental and ownership by the juvenile’s families.

8417: “8417_c015” — 2007/9/11 — 12:15 — page 337 — #13

Carbon Monoxide Poisoning

338

A substantial fraction of those poisoned develop long-term health damage. It must be treated immediately and correctly! Likewise, ignoring it as a few respondents thought, will not guarantee that it will go away. Finally, action must be prompt— calling a physician for an appointment days to weeks hence is of no value relative to maintenance of health after CO poisoning, and is equivalent to doing nothing.

15.10 DURATION OF TIME IN THE BODY (QUEST. 9) Probably the greatest confusion related to the length of time CO remains in the body. Being a very smart poison, as it has been characterized, it “washes out” quickly when one breathes clean air. The half-life of carboxyhemoglobin (COHb) is approximately 4–5 h, so within 24 h (5–6 half-lives) body CO concentration (i.e., COHb) is back at or very near background level (0.4–1.4% for nonsmokers). The healthcare-scientific subgroup did only slightly better on this question than lay people in any age group. Fifty percent of the people in that group didn’t know the answer. This is consistent with my experience with physicians, nurses, and so forth. Physicians will often ask patients to come to their office or a hospital on Monday to have blood drawn for COHb measurement, when the fact of CO exposure was discovered the previous Friday. Consequently, measurements made in this way are useless, and when used as a forensic tool may be directly injurious to the CO victim, since the finding of a normal COHb value may cause the patient to be sent back to the same site for continued CO exposure, injury, and even death.

15.11 PROPERTIES OF CO (QUEST. 10) This question gave respondents almost as much trouble as number 9. Only 15–30% knew that CO has about the same density as air (actually 4% lighter). An approximately equal or even larger percentage indicated they didn’t know CO’s properties. Of the healthcare-scientific subgroup, 28% got the question correct and 28% said they didn’t know the answer. Significant fractions of people thought CO was either much lighter or much heavier than air. An elementary knowledge of chemistry (molecular weights of the gases) allows one to easily compute density without recourse to pencil and paper, providing the information necessary to answer this question. Again, the educational establishment must take note. Of the select group of healthcare-scientific respondents, 32% believed CO was much lighter than air. CO mixed with air only rises when the air is warmer than surrounding air—that is, hot air rises! As stated elsewhere, CO cannot separate from air and rise as cream may separate from milk. That would violate the Second Law of Thermodynamics. Gases containing CO may have a strong or a weak odor, but CO plays no part in that. It is odorless and tasteless.

15.12 EXPERIENCE WITH CO POISONING? (QUEST. 11) As many as 10–15% of respondents believed they had experienced CO poisoning once or more than once. Up to another 15% knew of someone who had. That means

8417: “8417_c015” — 2007/9/11 — 12:15 — page 338 — #14

A Survey Study of Public Perceptions About Carbon Monoxide

339

that as many as one-fifth of respondents had some direct or indirect experience with CO poisoning. Another significant fraction of respondents didn’t know whether they had or not. Table 15.12 lists the sources of CO poisoning that were written in by respondents. Vehicles and heaters were prominent sources, both in Michigan and in Florida. Other sources included water heaters, fires, lawn mowers, boats, generators, and so forth. The percentages of respondents indicating poisoning at least once was not different between Michigan and Florida, and almost equal numbers of sources were given in both areas of the country. This tends to repudiate assertions that CO poisoning is much more prevalent in the cold north as opposed to the warm south.

15.13 USE OF CO EMITTING DEVICES Table 15.13 provides data about rental and ownership of CO emitting devices by respondents in Michigan and in Florida. Chain saws were owned by the largest percentages of respondents, both adults and juveniles or their parents, whether in Michigan or in Florida. Ownership was higher in the north than in the south. In all cases, a higher percentage of males owned chain saws than females. Ownership of generators was higher in Florida than in Michigan. A male–female difference was not clear-cut. Ownership of pressure washers was somewhat higher in the south than the north, but for ownership of kerosene heaters this difference was reversed. Chain saw, pressure washer, and generator rental tended to be higher in the south than in the north. Clearly, whether in the north or in the south, a significant percentage of people own or rent one or more small CO emitting devices that potentially expose them to CO. Improper use of these devices because of misperceptions and/or lack of education could result in injury or death.

ACKNOWLEDGMENTS In Benzie County Michigan, we wish to thank the Rotary club of Frankfort, the Benzie Sunrise Rotary Club, and the Benzie Bicycle Club. Our many thanks go to Mr. Gary Waterson, science teacher at Benzie County Central High School in Benzonia. In St. Johns County Florida, we wish to thank the St. Augustine Rotary Club, the St. Augustine Sunrise Rotary, and the Ancient City Road Runners. Many thanks also to St. Augustine science teachers Karen Ford, Ph.D. at Pedro Menendez High School (Public), Becky Melton, M.D. at St. John’s Academy (Christian School), and Mr. Peter Bugnet at St. Joseph Academy (Catholic High School).

References 1. Hampson, N.B., Zmaeff, J.L. Carbon monoxide poisoning from portable electric generators. Am. J. Prev. Med., 28, 123–125, 2005. 2. Centers for Disease Control and Prevention. Use of carbon monoxide alarms to prevent poisonings during a power outage—North Carolina, December, 2002. Morb. Mortal. Wkly. Rept., 53, 189–192, 2004.

8417: “8417_c015” — 2007/9/11 — 12:15 — page 339 — #15

8417: “8417_c015” — 2007/9/11 — 12:15 — page 340 — #16

16

Treatment of Carbon Monoxide Poisoning Suzanne R. White

CONTENTS 16.1 16.2 16.3 16.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Historical Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanisms of Carbon Monoxide Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Currently Available Neuroprotective Treatments . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.1 Normobaric Oxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.2 Hyperbaric Oxygen Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.3 Allopurinol and N-acetylcysteine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.4 Insulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.5 NMDA Receptor Antagonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.6 Brain-Derived Neuropeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.7 Hypothermia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.5 Approach to the Patient with Carbon Monoxide Poisoning . . . . . . . . . . . . . . 16.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.5.1 Neuroimaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.5.2 Pregnancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.5.3 Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.6 Management of the Sequelae of Carbon Monoxide Poisoning . . . . . . . . . . . 16.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

341 342 344 348 348 348 351 352 354 355 355 355 355 358 360 360 361 362 362

16.1 INTRODUCTION Carbon monoxide (CO) poisoning is the leading cause of toxicologic death in the United States of America, with 5600 fatalities reported annually.1 Worldwide CO remains the most lethal toxin in every community in which it has been studied.2 While mortality rates associated with acute exposure to CO may have declined over the past two decades, the total public health burden has not decreased.3 Most notably, delayed neuropsychiatric sequelae in a significant percentage of CO-poisoning survivors continues to pose an enormous challenge.4−9 The following chapter will focus primarily on the various treatment aspects of acute CO poisoning. It should be kept in mind that our present knowledge regarding 341

8417: “8417_c016” — 2007/9/11 — 12:15 — page 341 — #1

Carbon Monoxide Poisoning

342

therapy for CO poisoning is limited for several reasons. First, effective medical treatment is ideally guided by diagnosis of either positive or negative outcomes following exposure to toxic substances. For example, blood or urine levels of toxins, combined with characteristic signs or symptoms of toxicity often aid in the institution of appropriate therapy. Unfortunately, symptoms relating to CO exposure are notoriously vague, and some studies estimate that the diagnosis is missed in 30% of cases presenting to the emergency department.10 For instance, when all neurologic admissions over a 5-month period were screened, it was determined that three out of 29 patients admitted with impaired consciousness and no lateralizing neurological signs had serious CO intoxication.11 A further toxicological challenge is that carboxyhemoglobin (COHb) levels neither correlate with toxicity nor predict the risk for development of long-term effects.12,13 Although other predictors of long-term neuropsychiatric sequelae have been proposed (i.e., loss of consciousness,14 cerebral edema on brain computed tomography,15 elevated blood glucose16 or a history of a “soaking” type exposure,17 )their sensitivity and specificity are largely unproven. As a result, how best to treat patients with these clinical warning signs and symptoms remains controversial. Second, appropriate therapy for poisoned patients is ideally guided by an understanding of the toxic mechanisms of that poison. Unfortunately, even though CO has most likely been present since the beginning of time, and has been studied clinically for over 100 years, an adequate understanding of its toxic mechanisms continues to elude us. Lastly, treatment guidelines should ideally be on the basis of prospective, well-controlled, peer-reviewed studies. There is a dearth of such studies relating to CO-poisoning treatment in the literature. Despite these limitations, a general approach to treatment will be described. As an overview, treatment is on the basis of cessation of tissue hypoxia, the removal of CO from the body, the consideration of potential neuroprotective interventions, and management of the long-term sequelae of CO poisoning. First, a review of historical, often failed treatments for CO poisoning will be presented, followed by a discussion of promising neuroprotective agents. Finally, a clinical approach to the CO-poisoned patient will be outlined.

16.2 HISTORICAL PERSPECTIVE Oxygen therapy has been the mainstay of treatment for CO poisoning since it was first used therapeutically by Linas and Limousin in 1868.18 Haldane subsequently was able to experimentally demonstrate that mice exposed to “carbonic oxide” were unaffected if oxygen was provided during the exposure. In this seminal work, Haldane concluded that “the higher the oxygen tension the less dependent an animal is on its red corpuscles as oxygen carriers, since the oxygen simply dissolved in the blood becomes considerable when the oxygen tension is high.”19 Indeed, sea level oxygen decreases the half-life of CO from 320 to 80 min. Unfortunately, sea level oxygen alone has not been entirely effective in the treatment of CO poisoning, particularly with regard to the prevention of delayed neuropsychiatric sequelae. This realization has prompted researchers and physicians to search for yet other treatment modalities.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 342 — #2

Treatment of Carbon Monoxide Poisoning

343

The use of resuscitative gasses other than oxygen has been proposed. Studies by Killick demonstrated more rapid clearance of COHb with 5% carbogen (5% carbon dioxide, plus 95% oxygen), which was thought to be related to increased ventilatory drive.20−21 Schwerma et al.,22 exposed dogs to 0.3% CO until near-respiratory arrest occurred. Upon removal from exposure, 36% survived with fresh air alone. The survival rate increased to 50% when mechanical ventilation was used, 69% when ventilation with 100% oxygen was applied, and 66% when mechanical ventilation was combined with 7% carbogen. There was no clinical advantage to the use of carbogen in terms of improved survival, normalization of pH or lactate, or decreased incidence of neurologic sequelae in animals relative to breathing 100% oxygen alone.2 Thus, this method of treatment has subsequently fallen out of favor. Several other fascinating drug therapies for CO poisoning have not proven to be effective, and are mentioned here only for historical interest. Methylene blue, succinic acid, persantine, iron and cobalt preparations, and ascorbic acid have all been tried, without benefit.23 In animals, cytochrome c, theorized to activate cytochrome oxidase upon supplementation, has not been associated with clinical improvement.24 Hydrogen peroxide infusions do reduce COHb content in experimental animals, but the absence of human experience with this chemical and the danger of air embolism preclude its clinical use.25 While ultraviolet radiation was proposed to facilitate the dissociation of COHb from red blood cells during transit through skin capillaries and to decrease mortality in animals,26 these results were not able to be duplicated in a subsequent animal trial.27 Intravenous procaine hydrochloride does not improve the anoxia of CO poisoning in humans.28 Intravenous lidocaine was advocated, on the basis of its facilitation of neuronal recovery after cerebral ischemia in experimental animals, but has not yet been employed in CO-poisoned humans.29 Dipyridamole pretreatment in rats with inhalational CO toxicity inhibited ultrastructural changes in capillary endothelial cells, myocardial mitochondria, and myocardial myofilament arrangement,30 but follow-up studies were never peformed. Exchange transfusion has been reported to improve survival following CO poisoning in an animal model.31 Despite the fact that this method has been utilized in only a single patient,32 it is still promoted by some clinicians as an alternative to hyperbaric oxygen therapy (HBOT).33 While exchange transfusion does in fact lower COHb levels, given the complex mechanisms for CO toxicity that extend well beyond the toxicity of COHb, this technique is not likely to be effective as a sole therapy. Furthermore, given the potential for exchange transfusions to deplete valuable blood product resources and place the patient at risk for blood-borne pathogen infections, this treatment modality can no longer be recommended. Perfluorochemical infusions have been used in animal models as treatment for CO toxicity.34,35 Recently, pyridoxalated hemoglobin-polyoxyethylene conjugates (PHPs) have been developed. These agents act as blood substitutes capable of transporting oxygen through chemical modification of hemoglobin derived from human blood erythrocytes whose shelf-life has expired. Affinity of PHP for oxygen is almost identical to that of whole blood. PHP use in rabbits poisoned by CO was associated with prolonged survival time, temporary recovery of PO2 and PCO2 , and elevations in pH and blood pressure in comparison with animals treated with saline.36 Beyond the fact that human use of this product has not yet been reported, its efficacy as

8417: “8417_c016” — 2007/9/11 — 12:15 — page 343 — #3

344

Carbon Monoxide Poisoning

a sole therapy is unlikely, for the same reasons as discussed above with exchange transfusion. HBOT was first suggested as treatment for CO poisoning in 1901 by Mosso.37 The first clinical use of HBOT in the treatment of human CO poisoning, occurred in 1960.38,39 This modality has subsequently become the mainstay of therapy for severe CO poisoning in an average of nearly 1500 patients treated annually.40

16.3 MECHANISMS OF CARBON MONOXIDE TOXICITY To gain an understanding of the available methods of treatment for CO poisoning, a review of what is known about mechanisms for CO toxicity, albeit an incomplete comprehension of the problem, is presented here. Hypoxic ischemia plays a significant role in the neurotoxicity of CO, which binds slowly to hemoglobin, but with extremely high affinity (240 times that of oxygen). Oxygen binding sites are occupied by CO at very low partial pressures of the gas, decreasing the oxygen carrying capacity of the blood and subsequently decreasing the usual facilitation phenomenon for further unloading of oxygen at the tissue level. The net result is an abnormally hyperbolic oxygen dissociation curve that is shifted to the left. Those tissues most susceptible to the hypoxic effects of CO are those that are the most metabolically active. Oxygen delivery may further be impaired through the alteration of erythrocyte diphosphoglycerate concentration.41 In adults, COHb half-life is dependent upon the concentration of inspired oxygen, and is most commonly reported to be approximately 4.5 h on room air, 90 min on normobaric 100% oxygen, and 20 min with oxygen applied at hyperbaric concentrations. It should be noted that reported half-lives are extremely varied in the literature. In children, the half-life of COHb has not been well-studied, but is reported by one author42 to be 44 min on 100% oxygen at normobaric pressure, on the basis of measurements performed on 26 school-aged children. The half-life of COHb in the fetus is approximately 7 h.43 Over and above hypoxia, CO induces ischemia secondary to hypotension. Hypotension may be mediated through CO binding and activation of guanylyl cyclase which increases cGMP and triggers cerebral and peripheral vasculature smooth muscle dilatation or myocardial suppression. It may further be compounded by CO triggered release of nitric oxide from platelets with subsequent central and peripheral vasodilation. The degree of central nervous system (CNS) damage observed following poisoning correlates well with the degree of hypotension noted.44 Additional mechanisms of toxicity have been proposed given observations that (1) COHb levels did not correlate with toxicity, (2) COHb formed by noninhalational routes did not produce the same lethal consequences as inhalational exposure to CO, and (3) that delayed neuropsychiatric sequelae were common after apparent complete recovery from the initial CO insult. Early researchers such as Haldane,45 Brabkin,46 and Goldbaum47−49 suggested intracellular uptake of the gas as a possible mechanism for toxicity. In competition with oxygen, CO will bind iron or copper scontaining proteins such as myoglobin, mixed-function oxidases, and cytochrome c oxidase in vitro. The binding to cytochrome c oxidase (a, a3 ) has been proposed as the mechanism

8417: “8417_c016” — 2007/9/11 — 12:15 — page 344 — #4

Treatment of Carbon Monoxide Poisoning

345

for intracellular CO toxicity and has been demonstrated in animals.50 It has also been demonstrated that during recovery, the ultimate restoration of mitochondrial function lags behind clearance of COHb.51 However, the Warburg constant for cytochrome oxidase is unfavorable for CO binding relative to the other hemeproteins.52 Furthermore, only reduced cytochrome a, a3 binds CO. It is likely, then, that other hemeproteins act as “CO buffers,” thus preventing significant binding to cytochrome c oxidase at COHb levels of less than 50%. At high levels of COHb, depletion of high energy stores and intracellular neuronal acidosis occurs, which may favor CO-cytochrome binding. On the other hand, in vivo data from Rivers53 supports hemoglobin binding with impaired oxygen delivery, rather than mitochondrial poisoning as the etiology of the metabolic acidosis in CO poisoning. In this model, even at extremely high COHb levels, dogs were able to fully extract and utilize oxygen, indicating a lack of mitochondrial effect. In addition work by Ward54 demonstrated that expression of the heat shock proteins 72 and 32 (sensitive markers of acute neuronal stress) did not occur following CO poisoning in rats who were maintained normotensive throughout the exposure. This caused the authors to question the role of CO as a direct neurotoxin, and to suggest that neuronal injury results from hypotension-induced ischemia. The role of iron as a promoter or attenuator of CO toxicity is not clear. Iron deficiency, results in lowered hemoglobin, cytochrome and myoglobin levels in the animal model.55 These combined effects could potentially predispose to CO toxicity. Conversely, neuronal tissues high in iron content, such as the basal ganglia seem particularly vulnerable to the effects of CO. In fact, limiting iron availability confers neuroprotection from CO in the developing auditory system.56 Myoglobin binding may play a role in CO-mediated toxic effects. Myoglobin is a hemeprotein with similar three-dimensional configuration to hemoglobin that can bind CO reversibly. Myoglobin binds CO more slowly and with greater affinity than does hemoglobin in vivo. Normally, myoglobin is an O2 carrier protein that facilitates oxygen diffusion into skeletal or cardiac muscle cells and serves to place oxygen stores in close proximity to mitochondria. Cardiac and skeletal muscle injury could theoretically result from impaired myoglobin function. While the clinical significance of COHb formation is not yet clear, there is increasing emphasis on cardiac injury related to CO poisoning in the literature. Cardiac injury has historically been observed at COHb levels of 20–40%.57 Recently, Aslan reported on 83 young, healthy patients with severe CO poisoning. These victims had loss of consciousness in 63% and an average COHb level 34.4%. They were evaluated with echocardiogram (ECHO) and myocardial perfusion single-photon emission computed tomography (SPECT) scanning. Findings included diagnostic ischemic electrocardiogram (EKG) changes in 14.4% and abnormal SPECT results in 11%. Six of the latter group had confirmatory and corresponding ECHO abnormalities.58 Henry and colleagues noted ischemic EKG changes in 30% of 230 patients with moderate to severe CO poisoning. Cardiac biomarkers were elevated in 35% of these patients and in-hospital mortality was 5%.59 In a subsequent outcome study, these investigators noted an association between moderate to severe CO exposure and myocardial injury, finding this injury to be a strong predictor of mortality. It was further suggested that patient subsets include those with a “stunned myocardium” and others with “unmasking” of underlying coronary artery disease.60 Longer durations of CO exposure predisposed people to myocardial

8417: “8417_c016” — 2007/9/11 — 12:15 — page 345 — #5

346

Carbon Monoxide Poisoning

injury, but not mortality in this study. Defining the earliest and most sensitive cardiac markers of injury is clearly an important area of future investigation.61,62 The mechanisms underlying the delayed effects of CO poisoning have been longstanding toxicological conundrums. An increasingly complex body of literature suggests that brain ischemia/reperfusion injury, lipid peroxidation, vascular oxidative stress, neuronal excitotoxicity, apoptosis, and immunotoxicity all play significant roles. After removal from the CO environment, animal models demonstrate marked changes in neutrophil structure and function. Abnormal adherence to brain endothelial cell receptors quickly occurs, possibly as a result of endothelial damage. Up-regulation of endothelial intercellular adhesion molecules (ICAMs) is demonstrated on endothelial cell surfaces as a result of activation by inflammatory mediators. ICAMs bind beta 2 integrins located on PMN surfaces, resulting in aggregation of polymorphonuclear cells (PMNs) onto endothelial cells in the neuromicrovasculature. Subsequent degranulation of PMNs results in release of destructive proteases, which cause oxidative injury and trigger further inflammatory responses. Thom’s work63−65 has been instrumental in elucidating the CO-induced perivascular oxidative changes that occur during recovery from CO poisoning and ultimately lead to superoxide formation, prolonged lipid peroxidation reactions, reactive oxygen species (ROS) generation, vascular injury, and neuronal death. Even lower level CO exposure can produce vascular oxidative stress as evidenced by platelet-mediated nitric oxide release and deposition of peroxynitrate, a highly oxidative substance.66 Peroxynitrite, which forms from NO released from platelets and endothelial cells, can further inactivate mitochondrial enzymes and damage vascular endothelium of the brain.67,68 ROS generation can be attributed to several other sources including mitochondria and cycloxygenase. ROS production increases notably during CO hypoxia, with the highest oxidative stress occurring in the most vulnerable brain regions.69 This stress may result from lower antioxidant capacity or higher tissue concentrations of iron. In fact, limiting iron availability confers neuroprotection from CO.56 In mitochondria, decreased ratios of reduced oxidized glutathione are seen following CO poisoning, and may reflect decreased ability to detoxify ROS.70 As in animal models, acute CO poisoning in humans has resulted in intravascular platelet-neutrophil interactions and neutrophil activation. Thom71 demonstrated these phenomena in 50 patients by measuring actual aggregates and myleoperoxidase (MPO) concentrations. It was noted that patients with exposures of greater than 3 h duration had increased neutrophil expression of CD18 surface receptors and MPO. In animal models, MPO was deposited along the brain vascular lining and colocalized with nitrotyrosine. Changes did not occur in thrombocytopenic models or those using platelet–neutrophil interaction inhibitors such as tirofiban. Lastly, theses changes were not noted when l-nitroargninine methyl ester, a nitric oxide synthesis inhibitor, was given or in knock-out mice lacking MPO.71 Reactive products of lipid peroxidation like malonylaldehyde can cause adduct formation with neuronal myelin basic protein (MBP). The altered cationic state MBP triggers an adaptive immunological cascade that includes antibody-mediated degradation of MPB over the course of days. This triggers a secondary influx of macrophages and CD-4 lymphocytes that exhibit an autoreactive, proliferative

8417: “8417_c016” — 2007/9/11 — 12:15 — page 346 — #6

Treatment of Carbon Monoxide Poisoning

347

response to the altered MPB. Brain microglial activation ultimately occurs. These neuropathological changes are associated with decrements in learning that are not seen in rats immunologically tolerant to MBP.72 CO-mediated oxidative stress leads to the release of excitatory amino acids (EAA).73 Subsequent neuronal excitation leading to cell death may also play a role in the development of delayed toxicity following CO poisoning. These effects have been extensively reviewed by Piantadosi,74 and are paraphrased here. Excitatory amino acids such as glutamate, accumulate in synaptic clefts during neuronal depolarization owing to both excessive presynaptic release and failure in ATP-dependent reuptake mechanisms.75 Interstitial glutamate concentration increases in the hippocampus during and after CO exposure. Postsynaptic binding to at least three glutamate receptors including N-methyl-d-aspartate (NMDA), Kainic acid (KA), and α-amino3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) occurs with a secondarily increased influx of calcium into postsynaptic neurons. This hypercalcemia is associated with neuronal death. ROS production follows increased excitatory amino acid (EAA) release as well. NMDA receptor antagonism attenuates delayed neuronal degeneration in the hippocampus after CO poisoning, and is discussed further below. Physiological amounts of CO generated by heme breakdown seem to act to impact neurologic processes such as long-term potentiation of memory and neurotransmitter release. Conversely, in toxic doses, CO alters the modulating influence of local NO on soluble guanylate cyclase, an effect that is most evident at 7 days after exposure.76 Despite its role in vasodilation and peroxynitrite formation, NO has therefore been proposed to also mediate the toxic effects of CO.77 Given these disparate data, the exact role of NO as an attenuator or mediator of CO toxicity is still under investigation.78 Catecholamine excess may also be detrimental following CO exposure. EAA release causes excessive surges of norepinephrine and dopamine release and synaptic accumulation of these neurotransmitters. This effect appears to somehow be linked to NO production, as both events can be prevented by nitric oxide synthase (NOS) inhibition. CO induces both heme oxygenase and NOS in cortical pyramidal neurons. While it is unclear whether the resulting altered cerebral blood flow is a pathological or protective response, worsened outcome in sheep treated with hemoxygenase (HO) and NOS blockers suggests the latter.79 Therefore, EAA release, catecholamine release, and NO production are under tight regulation in vivo and can be influenced by COinduced oxidative stress. Furthermore, auto-oxidation or oxidative deamination of catecholamines occurs during ischemia and reperfusion by type B monoamine oxidase and contribute to further ROS formation.80 ROS production after CO exposure can be inhibited by partially blocking type B monoamine oxidase, located predominantly in glia.81 Gliosis, a known neuronal response to injury, and a condition found in Alzheimer’s disease, develops and may play a role in the delayed toxicity seen after CO hypoxia. ROS are capable of triggering programmed cell death or apoptosis. Apoptotic cell death requires activation and/or expression of specific cellular processes, some of which may act through oxidant pathways. In animals, CO-induced neuronal loss was slight at Day 3, increased at Day 7 and persistent at Day 21 following exposure. Neuronal apoptosis was observed to be present upon histopathological examination

8417: “8417_c016” — 2007/9/11 — 12:15 — page 347 — #7

Carbon Monoxide Poisoning

348

of the animals in this model.82 Others have not demonstrated apoptosis in rats, despite moderately severe poisoning,83 but have observed gliosis. Caspase-1 and NOS inhibitors both block CO induced apoptosis.84 As eloquently summarized by Piantadosi,74 “impaired mitochondrial energy provision in CO hypoxia/ischemia leads to neuronal depolarization, EAA and catecholamine release, and failure of re-uptake until energy metabolism is restored during reoxygenation. These processes, normally modulated by NO production, could contribute to degeneration of neurons in vulnerable regions, possibly by enhancing mitochondrial ROS generation which can initiate apoptosis.”

16.4 CURRENTLY AVAILABLE NEUROPROTECTIVE TREATMENTS 16.4.1 NORMOBARIC OXYGEN There are limited data to suggest that normal neuropsychiatric outcome is possible after treatment with normobaric oxygen (NBO). In one study, 33 patients with acute CO poisoning (mean COHb 29.4%, 10 patients above 40%, 7 comatose on arrival) were treated with 100% NBO. Recovery was reportedly rapid, with no neurological deficits at discharge. Formal neuropsychiatric testing and follow-up were not performed, however.85 Similarly, four patients presenting comatose from CO poisoning who did not receive HBOT were identified retrospectively, but then evaluated with formal neuropsychological testing at 6 and 12 months after exposure. All had normal neuropsychiatric examinations.86 Meert looked retrospectively at the outcome of children treated with NBO, and concluded that acute neurologic manifestations resolve rapidly without HBOT.87 However, neurologic outcome was assessed from nursing and physician records, physical and occupational therapy evaluations, and unspecified neuropsychological examinations in an unspecified number of patients. Therefore, the major limitation of this study is the lack of detailed neurologic assessment both at presentation and at follow-up that would allow detection of those specific neuropsychiatric changes known to result from CO poisoning. Nonetheless, the authors noted “gross” neurologic abnormalities in nine (8.5%) survivors, with seven of those persistent at various stages of follow-up (2 months to 3.3 years). Three patients developed delayed neurologic syndromes including tremors, hallucinations, seizures, occipital lobe infarctions, defects in cognitive, and interpersonal skills. The presence of serious comorbidities, such as smoke inhalation, burns, need for mechanical ventilation, and need for surgical procedures certainly confounds the outcomes reported by these investigators, which are not reflective of pure CO exposure. Overall, given the small size of these trials and their significant methodological limitations, they do not significantly add to our understanding of predictors for good outcome after treatment with NBO therapy.

16.4.2 HYPERBARIC OXYGEN THERAPY In addition to the aforementioned effects of increasing both the amount of dissolved oxygen in the blood and the rate of displacement of CO from hemoglobin, HBO may

8417: “8417_c016” — 2007/9/11 — 12:15 — page 348 — #8

Treatment of Carbon Monoxide Poisoning

349

have other beneficial effects. In animal models, HBO at 3 atmospheres absolute (ATA) also prevents functional neurological impairment88 and HBO at 2.5–3.0 ATA reversibly inhibits PMN CD18 beta 2 integrin activation therefore decreasing adherence of PMNs to endothelial cells.89,90 This effect has also recently been noted in patients with acute severe CO poisoning, in whom HBO modulated neutrophil generation of ROS and surface expression of CD18 receptors.91 Moreover, HBO is known to regenerate inactivated cytochrome oxidase, and may thereby restore mitochondrial function.92 An astroglial structural protein S100B that is a proposed marker for neuronal injury is elevated in CO-poisoned rats treated with ambient air or NBO, but not in those treated with HBO.93 Other proposed beneficial effects of HBOT include decreased production of ROS,94 protection against cerebral edema and increased cerebrospinal fluid (CSF) pressure,95,96 induction of production of protective stress proteins (SP72), and antagonism of NMDA excitotoxic neuronal injury. Conversely, HBO has not been effective in animal models as a treatment for nonCO mediated acute cerebral ischemia with reperfusion.97 Anecdotal human case reports suggest significant clinical improvement from CO poisoning during HBOT.98−105 In addition, numerous other case series report on the beneficial effects of HBOT, but these are limited by either their retrospective nature, or prospective design without the use of randomization, double-blinding, or controlled methodology.106−112 Three pioneering research efforts attempted to discern potential benefit from HBOT, through randomization of CO-poisoned patients into HBO and NBO treatment groups. Raphael113 studied 343 mildly poisoned CO-patients (those who had not lost consciousness) and found no difference at 1 month follow-up in neurologic outcome between HBO and NBO-treated groups, and no benefit to multiple HBO treatments in a more severely poisoned group (those who had lost consciousness), randomized to either one or two HBO treatments. Criticisms of this study regarding the method of neurologic evaluation used, the application of inadequate doses of oxygen therapy, and potential delays in HBO treatment have been raised.114 Ducasse115 randomized noncomatose patients to HBO and NBO groups and found significantly improved differences in quantitative electroencephalogram (EEG) and cerebral vascular responsiveness to acetazolamide in the HBO group during the first 24 h. Differences at 3 week follow-up are not reported consistently. Early improvements in clinical signs and symptoms, such as headache, reflex impairment, and asthenia were significant in the HBO group. Thom116 randomized patients with mild to moderate CO poisoning into NBO versus HBO groups. The incidence of delayed neuropsychiatric sequelae (DNS) was 23% after treatment with oxygen at ambient pressures and 0% in the HBO group. DNS persisted for an average of 41 days following exposure. The author concluded that HBOT decreased the incidence of DNS. Limitations of this study117 are the lack of double-blinding, inconsistent NP testing methods used, and exclusion of severely poisoned patients. Recently, several additional more rigorously designed trials have been carried out. All are prospective, randomized, double-blinded, controlled clinical trials assessing the benefit of HBO versus NBO in the treatment of CO poisoned patients. Two have been completed and one has resulted in publication of the interim data analysis. These are reviewed below.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 349 — #9

350

Carbon Monoxide Poisoning

Scheinkestel et al. peformed the first large prospective, randomized, double blinded, controlled clinical trial investigating the neurologic sequelae in 191 patients with all grades of CO poisoning after treatment with HBO and NBO.118 Sham HBO treatments were given to those randomized to the NBO group. Pregnant women, children, and burn victims were excluded. Higher doses of oxygen were utilized than reported in most previous studies, averaging approximately 37 COHb-dissociation half-lives in the HBO group and 28.5 in the NBO group (up to three daily treatments in those not improving). Neuropsychiatric evaluations were performed at completion of treatment and at discharge. No benefit and possible adverse effects of HBO were found. Overall mortality was 3%, with persistent neurologic deficits in 71% at hospital discharge and 62% at one-month follow-up. All five patients with delayed onset of neurologic deficit occurred in the HBO group. Limitations of the study are that 44% of victims had ingested other drugs, there was a mean treatment delay of 7.1 h, 56% of patients were lost to 1-month follow-up, and 76% were suicidal, which could impact neuropsychological testing. These results led the investigators to conclude that HBOT should no longer be recommended for CO poisoning treatment. A similarly designed longitudinal follow-up study employing sham treatments, found approximately 30% of patients with acute CO poisoning have neurocognitive problems 1 year after poisoning. Of these patients, approximately one-third have the delayed neuropsychiatric syndrome and two-thirds have persistent neurocognitive problems, mainly difficulties with memory and executive function. 119 Patients in the treatment arm received three HBO sessions. This randomized control trial demonstrated a 46% reduction in cognitive sequelae in HBO-treated patients at 6 weeks following poisoning, which was maintained at 1 year after poisoning.120 It should be noted that nonspecific symptoms were the primary determinant of a statistical difference between treatment groups. A third randomized controlled trial (RCT) performed on noncomatose CO-poisoned patients showed no benefit between HBO and NBO at 1 month, 6 months, or 1 year postexposure. Since the study is published only in abstract form, details are minimal.121 Interim results from a 4th RCT trial employing one HBOT session in patients with moderate to severe CO poisoning showed no difference in outcome.122 Assessments were made by questionnaire and a blinded neurology examination. A high rate of symptoms in both groups was noted (39–42%) at 1 month after poisoning. The use of repetitive HBOT is controversial but common with 23% of US HBOT facilities automatically giving more than one HBO treatment per CO-poisoned patient.40 Some investigators perform additional treatments if lack of improvement is noted after the 1st HBO treatment.104,123−125 Others report no benefit to multiple HBO treatments.113 Turner et al.124 recently proposed the use of the initial hydrogen ion concentration (degree of metabolic acidosis) as a marker for repetitive HBO treatment requirement, on the basis of a retrospective analysis of 48 patients. McNulty et al.126 found impairment of short-term memory for verbal material to be predictive of the number of HBO treatments needed. Finally, the use of peak alpha frequency on EEG as an indicator of need and efficacy of repetitive HBO treatments has been proposed.127 While generally safe, unusual risks of HBOT include complications arising from transport to an HBOT facility, barotrauma, oxygen toxicity with resultant seizures in 1–3% of CO-poisoned patients128 and fire or explosion hazard. Historically, the lack of definitive results from RCTs have rendered the indications for HBOT in CO poisoning

8417: “8417_c016” — 2007/9/11 — 12:15 — page 350 — #10

Treatment of Carbon Monoxide Poisoning

351

arbitrary and tremendously varied. Recommendations have primarily been derived from patient history, presenting neurologic and cardiovascular signs and symptoms, and laboratory data such as glucose, lactate, arterial blood gases, electrocardiography, and COHb levels. Unfortunately, none of these clinical or laboratory findings at presentation are entirely predictive of long-term outcome following CO exposure. Those that show future promise, such as SB100 and neutrophil response, are not yet clinically available. Of note, recent preliminary work suggests that HBOT benefit may be limited to reducing the incidence of persistent neuropsychiatric sequelae (PNS) but not DNS, perhaps through modulating early but not later mechanisms for brain injury.129 Overall, the evidence favoring HBOT as protective against DNS and PNS remains under fire as experts continue to debate the results and limitations of each of the randomized controlled trial performed to date.130−138 These uncertainties definitely suggest that greater future research emphasis be placed on non-HBOT methods of therapy.

16.4.3 ALLOPURINOL AND N-ACETYLCYSTEINE As described above, there is considerable evidence that reactive oxygen metabolites mediate neurologic injury in models of CO poisoning. Lipid peroxidation is documented in rats after CO exposure at concentrations sufficient to cause unconsciousness. Products of lipid peroxidation are increased by 75% over the baseline values 90 min after CO exposure. Unconsciousness is associated with a brief period of hypotension, so brief that in itself it causes no apparent insult. Lipid peroxidation occurs only after the animals are returned to CO-free air; and there is no direct correlation between the degree of lipid peroxidation and COHb level.63 This suggests that ischemia and reperfusion may play a role in the ultimate neurologic injury. Xanthine oxidase also has a central role in this toxicity. During the COinduced PMN degranulation described above, released proteases convert xanthine dehydrogenase to xanthine oxidase. Xanthine oxidase generates superoxide free radicals and lipid peroxidation occurs.65 The xanthine oxidase enzyme is an NADdependent dehydrogenase that under ischemic conditions converts to an oxidase, utilizing molecular oxygen rather than nicotinamide adenine dinucleotide (NAD) as an energy source and generates the superoxide radical and hydrogen peroxide. These products in turn cause tissue injury, the brain being particularly susceptible with its low content of catalase and glutathione peroxidase.139 The restoration of xanthine dehydrogenase functional activity is accomplished through the use of xanthine oxidase inhibitors (allopurinol)140 and sulfhydryl donors [N-acetylcysteine (NAC)]141 in nonCO-mediated neuronal injury. Fechter et al.142 noted that acute CO poisoning produces preferential high-frequency hearing impairment, noted to be a consequence of other types of anoxic exposure. These investigators also discovered that either allopurinol or phenyl-n-tert-butyl-nitrone (PBN), a free radical scavenger blocked the formation of characteristic compound action potential threshold elevation and cochlear microphonic amplitude. Therefore, both agents were effective in blocking loss of CO-mediated auditory threshold if given prophylactically in the guinea pig model. Allopurinol and N-acetylcysteine have been used in the treatment of CO-induced neuronal injury. Thom63 demonstrated decreased conversion of xanthine dehydrogenase to xanthine oxidase with decreased lipid peroxidation in rats pretreated with

8417: “8417_c016” — 2007/9/11 — 12:15 — page 351 — #11

Carbon Monoxide Poisoning

352

allopurinol. Only one human case report demonstrated effectiveness of such combined therapy for the treatment of CO poisoning. A 26-year-old male with a COHb level of 25%, 40 h postexposure, who was comatose for 4 days with cerebral edema on computed tomography (CT) was treated with both a xanthine oxidase inhibitor (allopurinol) and a sulfhydryl donor, NAC. NAC was given intravenously over a 20-h period and allupurinol was given orally for 2 weeks. Eight hours after the completion of this regimen, the patient became responsive and gradually improved over the next three weeks. “Neurological and mental examination at six weeks followup were normal.”143 Although no formal neuropsychiatric testing was reported, this type of therapy may perhaps provide a basis for further study. Since allopurinol theoretically prevents the formation of free radicals, it remains to be seen whether any benefit exists for postexposure administration. Other antioxidants such as dimethyl sulfoxide and disulfiram have been shown to prevent learning and memory deficits in CO-poisoned mice in preliminary reports.144 Such agents may someday serve as useful adjuncts in limiting free radical mediated injury.

16.4.4 INSULIN In humans and animals numerous studies have shown that elevated blood glucose is associated with worsened neurologic outcome after brain ischemia caused by stroke or cardiac arrest. Acute severe CO poisoning is characterized by hyperglycemia and this elevation has been linked to increased severity of brain dysfunction in the rat.145 Indeed, animal studies show that CO exposure raises blood glucose in a dosedependent manner, and is an independent predictor of neurologic outcome. A few similar observations have been made in CO-poisoned patients. Penney146 observed that elevated admission blood glucose was associated with worse neurologic outcome after CO poisoning in patients. Leikin found elevated blood glucose in most patients presenting with COHb saturation above 25%.147 Furthermore, anecdotal evidence in the human literature suggests that the neurologic outcome in diabetics poisoned with CO is generally worse than in nondiabetics.148 Considerable basic research has been directed at identifying molecular mechanisms of tissue injury and potential interventions to allow the preservation or rescue of neurons after stroke and cardiac arrest. On the basis of the aforementioned association between elevated blood glucose and poor outcome, insulin has recently been investigated as a potential therapeutic agent in various models of brain and spinal cord ischemia, and indeed appears to substantially ameliorate neuronal death induced by ischemia in these studies. Surprisingly, this neuron-sparing effect is known to be independent of insulin-induced reductions in blood glucose, and is hypothesized to be mediated through cell signal transduction mechanisms, in common with other growth factors. The question of whether insulin ameliorates neuronal injury secondary to CO toxicity was investigated in rats who were exposed to a CO LD50 of 2400 ppm for 90 min. Survivors were treated for 4 h with (1) normal saline infusion (2) continuous infusion of glucose to clamp blood glucose levels at 250–300 milligrams per deciliter (mg/dL) and (3) continuous infusion of glucose to maintain blood glucose levels at 250–300 mg/dL with intraperitoneal (IP) injections of 4 units/kg of regular insulin.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 352 — #12

Treatment of Carbon Monoxide Poisoning

353

Neurologic scoring was performed at time 0, 5.5, 24, 48, 72, and 96 h. It was noted that significant neurologic deficit occurred in all groups after the CO exposure and treatment period. Induced hyperglycemia after CO exposure was associated with significantly worsened neurologic scores as compared to saline-treated controls. Insulin therapy simultaneous with induced hyperglycemia significantly improved neurologic scores at all times despite maintenance of comparable hyperglycemia with respect to the group treated only with glucose. No significant difference in mortality was found between treatment groups.149 Several theories have evolved regarding the postreceptor binding protective effect of insulin on the neuron. Insulin has been shown to provide neuromodulatory inhibition of synaptic transmission in vivo and in vitro. As an inhibitor of glial uptake of gamma amino butyric acid (GABA), insulin may increase the availability of this inhibitory neurotransmitter and may decrease neuronal firing, beneficially reducing cell metabolism.150 The additional effect of sodium extrusion from the cell which affords subsequent protection against water accumulation may prevent neuronal swelling.151 Furthermore, it has been suggested that an insulin-induced elevation of brain catecholamines through both inhibition of catecholamine uptake and stimulation of release might be a contributory neuroprotective mechanism, since catecholamines have been found to attenuate ischemic brain damage. Of these theories regarding insulin’s neuroprotective activity, however, the most recent highlights its role in stimulating second messengers, and emphasizes its potential genomic effects, that is, the regulation of protein synthesis, enzymatic activity, and the signaling of cell proliferation. It is well established that the neonatal brain is rich in insulin-like growth factor receptors. Indeed, insulin is similar in structure to other growth factors such as platelet-derived growth factor (PDGF), epidermal growth factor (EGF), insulin-like growth factor-1 (IGF-1). Such peptides are involved in basic neuron development and differentiation. Once bound to its receptor, like other growth factors, insulin triggers signal transduction by internal autophosphorylation of tyrosine on the insulin receptor, which subsequently enhances further phosphorylation reactions of other tyrosine containing substrates by tyrosine kinase, located on the insulin receptor. This type of tyrosine phosphorylation is important in signaling pathways for such growth factors and products of proto-oncogenes. Insulin is a progression growth factor in replication G0G1 phase and works synergistically with other growth factors to generate both competence and progression of cells. Through tyrosine phosphorylation of phosphokinase C, other second messengers such as diacyglycerol are formed, which increase intracellular calcium, activate the sodium–hydrogen pump, and increase intracellular pH. This pH change in turn activates the sodium potassium ATPase pump, which signals cell proliferation.152 Insulin also regulates specific mRNA levels through diacylglycerol153 and may increase mRNA efflux from the nucleus through nuclear triphosphatase activation.154,155 More importantly, insulin stimulates lipid neogenesis. It appears therefore, that the effects of insulin are fundamental with regard to cell signaling, proliferation, replication, and repair following injury. These processes are crucial to cells such as neurons which normally are terminally differentiated and contain little if any capacity to replicate or to synthesize repair lipids.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 353 — #13

Carbon Monoxide Poisoning

354

Certainly, anatomic correlates to the above proposed mechanisms are in place. For example, it has been demonstrated that the location of insulin and IGF-1 receptors correlates with phosphotyrosine products in the brain.156 Moreover, the basal ganglia, areas typically found to be damaged by CO, possess low levels of insulin receptors. Although the initial results of animal studies such as those above may provide building blocks for clinical work, any recommendations regarding the use of insulin in humans as treatment for CO poisoning will of course await further studies.

16.4.5 NMDA RECEPTOR ANTAGONISTS Recent evidence implicates the endogenous excitatory amino acids such as NMDA in ischemic neurodegeneration.157−159 Moreover, the NMDA receptor antagonist, MK801, prevents nonCO induced ischemic neurodegeneration in animal models.160 Successive CO exposures induce a consistent pattern of degeneration of hippocampal CA1 pyramidal neurons, a selective neuronal death that resembles that seen with other models of cerebral ischemia. This observation has prompted the study of NMDA receptor antagonists in CO poisoning in mice. Ishimaru pretreated animals with a competitive NMDA antagonist, CPP; a noncompetitive NMDA antagonist, dizocilpin (MK-801); a glycine binding site antagonist, 7-CK; a polyamine binding site antagonist, ifenprodil; glycine; and saline. Seven days postexposure the number of hippocampal CA1 pyramidal cells was quantified using an image analyzer. A decrement of 20% in the number of hippocampal CA1 pyramidal cells was noted relative to the control group. Those animals receiving high doses of MK-801, 7-CK, and CPP had significant reductions in neuronal damage. No clear protective effect was obtained with ifenprodil. Interestingly, glycine, a facilitory neurotransmitter at the NMDA receptor complex did not exaggerate the CO-induced neuronal damage as might be expected.161 Although no neurologic outcome correlates or survival data are reported, this work may provide valuable mechanistic and possibly future therapeutic insights. Similar work by Lui162 suggested beneficial effects of MK 801 when administered either systemically or directly to the cochlea in protecting against CO-induced ototoxicity. Glutamate is another excitatory neurotransmitter acting at the NMDA receptor complex. Postexposure treatment of mice with glutamate antagonists prevents CO-induced learning and memory deificits.163 Finally, NOS inhibitors prevent NMDA receptor activation and were protective of learning deficits in CO poisoned mice.73 Ketamine is a widely used dissociative anesthetic agent, which is known to have NMDA receptor-blocking properties.164 It has been shown to be neuroprotective in various animal models of ischemic and anoxic neuronal injury. It has also been observed to blunt hypotension, a condition known to worsen CO-mediated neuropathologic changes. Promising work by Penney165 demonstrated significantly reduced cerebral edema, more rapid recovery from hypotension, and suppressed lactate formation following CO-poisoning when 40 mg/kg ketamine was administered to rats before and during CO exposure.165 This same study did not yield positive results with the use of verapamil, which could theoretically block NMDA-mediated postsynaptic calcium uptake in neurons.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 354 — #14

Treatment of Carbon Monoxide Poisoning

355

16.4.6 BRAIN-DERIVED NEUROPEPTIDES Cerebrolysin, a drug,produced by enzymatic breakdown of lipid free proteins of porcine brain is a putative neuroprotective agent of unknown mechanism. Proposed neurotrophic effects are supported by reports that cerbrolysin-treated rats had increased brain protein synthesis,166 prevention of neuronal degeneration,167 and enhanced neuronal growth in tissue culture.168 Effects on the blood brain barrier have also been noted. Interestingly, cerebrolysin increases expression of the blood brain barrier glucose transported gene in brain endothelial cell cultures. It is hypothesized that cerebrolysin may accelerate repair of the blood brain barrier in regions compromised by hypoxia.169 Recently, a model of acute CO poisoning combined with spreading depression-induced metabolic stress was used to examine the protective effects of cerebrolysin on the development of electrophysiological, behavioral, and morphological signs of hypoxic damage in rats. Spreading depression waves reflect the recovery of cerebral cortex in the peri-ischemic areas, or penumbra zone. After a 90 min exposure to 0.8–5% CO, microinjections of 5% KCl into the cortical and hippocampal areas were performed and the duration of spreading depression was noted. At 9 and 18 day follow-up, repeat spreading depression measurements were taken, and a decrease in amplitude was used as an index of brain damage. Postexposure cerebrolysin-treated animals had significantly improved hippocampal recovery. Better performance was also noted on behavioral testing, and no apparent histological damage was apparent in the hippocampus as compared to controls.170 This very promising neuroprotective agent, which appears to be effective even if given postexposure, certainly deserves further study to elucidate any possible beneficial role in humans.

16.4.7 HYPOTHERMIA Hypothermia was found to be beneficial in the management of CO poisoning by Sluijter,171 an effect that was thought to be secondary to increased dissolved oxygen in the blood at lower temperatures. Peirce et al.,172 howeve, was unable to demonstrate any synergistic effect when hypothermia was used in conjunction with HBO in a dog model.172 An interesting report of the use of mechanical ventilation and hypothermia in patients with abnormal motor activity or coma to treat CO toxicity, noted complete reversal of these manifestations in three patients when therapy was initiated within the first 24 h. No beneficial effects were noted in a fourth patient who did not receive hypothermic treatment until 5 days after exposure. HBO was not available to these patients.173

16.5 APPROACH TO THE PATIENT WITH CARBON MONOXIDE POISONING 16.5.1 GENERAL The most critical step in managing the patient poisoned with CO is the cessation of tissue hypoxia. This involves supplementation with 100% oxygen, delivered

8417: “8417_c016” — 2007/9/11 — 12:15 — page 355 — #15

356

Carbon Monoxide Poisoning

by either a tight-fitting continuous positive airway pressure (CPAP) mask or by endotracheal intubation. Intubation may be necessary in the patient with chronic obstructive pulmonary disease (COPD), to avoid carbon dioxide retention secondary to high concentrations of oxygen. NBO should be initiated as soon as the diagnosis is entertained, and should not be delayed for confirmatory COHb levels. As discussed above, the use of 5% carbon dioxide mixed with 95% oxygen (carbogen) has been proposed by some to facilitate the release of CO from hemoglobin by increasing ventilatory response. This therapy is of questionable value, and has fallen out of favor. Even though not demonstrated in animals,174 the potentially life-threatening possibility of carbogen-induced carbon dioxide retention and subsequent worsening of an already existing acidosis would contraindicate it use in the patient with COPD, concurrent poisoning with respiratory depressants, or altered mental status. The duration of oxygen therapy is guided by a knowledge of the COHb half-life and allows for a margin of safety. Generally this would involve at least 6 h of therapy on 100% oxygen, longer if the patient is gravid or an infant. An early chest x-ray is mandatory to assess for evidence of pulmonary edema resulting from CO or other inhaled toxins. Once airway control and oxygenation are assured, attention should be directed toward the cardiovascular system. Continuous cardiac monitoring is advisable and a 12 lead EKG should be obtained to assess for subclinical cardiac ischemia. Myocardial enzymatic changes with or without EKG changes, are increasingly described in adults with CO poisoning as outlined above. It is prudent therefore to perform a cardiac evaluation on patients with CO poisoning. It is not clear, however, whether CO-induced myocardial injury, particularly the “stunned myocardium” is a predictor of poor clinical outcome or mortality. Should arrhythmias, ischemia, or hemodynamic instability occur despite therapy with 100% oxygen, the patient could be considered a candidate for HBOT. Myocardial depression and arrhythmias may occur secondary to extremely low arterial pH, such as has been noted in patients with severe lactic acidosis. Severe acidosis should therefore be treated aggressively. However, correction of mild acidosis with sodium bicarbonate is not advisable as this could result in a further shift of the oxyhemoglobin dissociation curve to the left, and impair the unloading of oxygen to hypoxic tissues. A novel calcium sensitizer, levosimendan, improves myocardial contractility and increases coronary artery flow. Its use in a single patient with cardiogenic shock from CO-induced myocardial stunning was associated with improved hemodynamics relative to dobutamine. Improvement was assessed by cardiac magnetic resonance imaging.175 In those patients with altered mental status, causes of rapidly reversible causes of coma should be considered and treated by the bedside assessment of a fingerstick glucose and the administration of thiamine and naloxone, a narcotic antagonist. Supplemental glucose may be needed to correct hypoglycemia, but every attempt should be made to maintain euglycemia, and avoid iatrogenic hyperglycemia. A thorough physical assessment for burns, odors, toxidromes, skin findings, and signs of smoke inhalation, trauma, or abuse is indicated. Acareful history regarding the circumstances surrounding the exposure must be obtained once the patient is stabilized. Considerations should be made to gastric or skin decontamination and activated charcoal

8417: “8417_c016” — 2007/9/11 — 12:15 — page 356 — #16

Treatment of Carbon Monoxide Poisoning

357

administration in the setting of suspected intentional drug abuse, suicide attempt, or dermal chemical exposure (methylene chloride). Removal of CO from the body is best accomplished through displacement by oxygen, either normobaric or hyperbaric. Other less conventional therapies have been used anecdotally with favorable outcomes, and include both exchange transfusion and extracorporeal oxygenation.176 These invasive approaches however, would be recommended only in unusual circumstances, for example if HBOT was not available in a deteriorating or moribound patient. Should the patient be CO-poisoned from smoke inhalation, numerous other products of combustion may be contributing to the metabolic and pulmonary derangements seen. Of particular concern is cyanide, a lethal combustion product, that is commonly elaborated when plastics or synthetic materials burn. The patient with CO poisoning and evidence of smoke inhalation who remains significantly acidotic despite treatment with oxygen should be suspected to have concomitant cyanide toxicity. Specifically, cyanide poisoning is associated with enclosed space fires, the presence of soot in mouth or sputum, altered consciousness, hypotension, and an elevated lactate >8–10 mmol/L without significant burns.177 Some authors advocate empiric treatment of fire victims with suspected cyanide poisoning with the sodium thiosulfate component of the cyanide antidote kit. The other components, methemoglobin-forming agents such as amyl nitrite and sodium nitrite, have been traditionally withheld if CO poisoning is suspected in order to avoid further hemoglobinopathy and worsened hypoxia. Moore et al.178 demonstrated a 25% increase in mortality in sodium nitrite-treated animals with CO-poisoning compared to the untreated controls. Despite this, a human study demonstrated that five of seven patients with CO poisoning were safely treated with the antidote kit in its entirety.179 These patients however, had only moderate COHb levels, with a mean level of 26%. Coupled with the fact that only a small number of patients were studied, this form of empiric treatment for cyanide poisoning in victims of smoke inhalation cannot yet be widely recommended. Another cyanide antidote, hydroxycobalamin is under Food and Drug Administration (FDA) consideration for approval. While not yet readily available in the United States, this new antidote will not harbor the risk of methemoglobinemia found with nitrites and will ultimately offer a safer option for victims of smoke inhalation. For now, given the relative safety of the sodium thiosulfate component of the cyanide antidote kit, its sole use may be advisable for the treatment of the patient dually poisoned with CO and cyanide. Once the patient has been stabilized, consideration of the use of possible neuroprotective agents including HBO should be made. If the patient is awake, a mental status examination should be performed. Abbreviated neuropsychologic tests (CONSB) have been developed specifically for the CO poisoned patient, but often are not practical for use in the emergency department or in those with moderate or severe acute intoxication. Examples of tasks performed by the patient during administration of the CONSB include placing pegs in a board, complete rapid finger tapping, memorization, construction, number processing, and subjective stress response.180 Memory impairments are the most frequent cognitive impairment noted following CO poisoning, some improving over time.181 Other common deficits include visual tracking,

8417: “8417_c016” — 2007/9/11 — 12:15 — page 357 — #17

Carbon Monoxide Poisoning

358

visuomotor skills, visuospatial planning, and abstract thinking. Impaired executive function, information processing speed, attention and concentration are also common.

16.5.1 NEUROIMAGING Neuroimaging studies can be valuable adjuncts to the neuropsychiatric assessment. They should be considered in CO-poisoned patients with altered mental status, abnormal, or lateralizing neurologic examinations, or a history of head trauma. Increasingly, correlates are described between cognitive impairment and neuroimaging findings (see other chapters in book), particularly relative to basal ganglia atrophy, fornix atrophy, and white matter hyperintensities.182 Moreover in adults, CT abnormalities have been prognostic with regard to neurologic outcome in several studies.183−185 Pathologic lesions seen on CT owing to CO intoxication are variable, including cerebral edema, symmetrical low density areas in the basal ganglia, symmetrical and diffuse white matter low density areas, and, as late changes, ventricular dilation, and sulcal widening. The classic finding of bilateral symmetrical hypodensities in the basal ganglia, especially the globus pallidus, most typically becomes evident within 24–48 h of exposure. However, such abnormalities have been reported to appear anywhere from the first day to 5 years following CO exposure. Remarkably, such basal ganglia lesions have been reported to occur in 32% to 86% of CO-poisoned patients.186 These lesions are not pathognomic for CO poisoning, however, and when encountered, the differential diagnosis includes methanol, cyanide, or hydrogen sulfide toxicity; hypoxia; hypoglycemia; the hemolytic uremic syndrome; osmotic myelinolysis; encephalitis; inborn errors of metabolism; and Huntington’s disease.187 Likewise, white matter lesions are frequent and include hyperintesities in the periventricular and centrum semiovale or deep regions, generalized white matter degeneration, and generalized atrophy. Several studies suggest that white matter lesions occur even more commonly than do basal ganglia lesions.188 Magnetic Resonance Imaging (MRI) may be superior to CT in detecting white matter, cerebral, cerebellar, substantia nigral, and basal ganglia lesions following CO poisoning.189−195 Quantitative MRI may be more sensitive in evaluating the hippocampal regions in patients with DNS following CO poisoning. Some authors report a good correspondence between MRI and memory deficits on neuropsychological evaluation in adults.196−198 Conversely, Prockop199 noted a significant percentage of patients with normal MRI examinations had intellectual impairment on neuropsychological testing. Functional imaging such as SPECT scanning provides an indicator of the severity of cerebral damage and correlates with outcome.200 The combination of EEG with SPECT scanning may provide greater sensitivity for detecting anomalies than EEG alone.201 In a cohort of adult patients with acute severe CO poisoning, treated with HBO, positron emission tomography (PET) scan findings of globally increased oxygen extraction ratios and decreased blood flow in the frontal and temporal cortex were most severe in those patients with DNS or PNS. These changes are temporary in patients who appear normal following CO exposure and in those with temporary neurological and psychiatric deficits. This suggests that ischemia is ongoing after

8417: “8417_c016” — 2007/9/11 — 12:15 — page 358 — #18

Treatment of Carbon Monoxide Poisoning

359

CO intoxication, even after apparent normalization of the clinical status.202 Others have noted prolonged perfusion and metabolic abnormalities in patients with neurologic deficits following exposure.203 In a small case series, PET abnormalities in the basal ganglia of CO victims were undetected by CT or MRI, suggesting possible greater sensitivity with PET.204 Newer techniques in SPECT scanning may allow for even earlier detection of regional CO-induced anomalies.205 Specifically, decreased perfusion in the basal ganglia and cortex correlated with parkinsosian symptoms and cognitive deficits, respectively. More recently, magnetic resonance spectroscopy (MRS) detected decreased n-acetyl aspartase in the basal ganglia bilaterally in one-third of CO-poisoned patients studied.199 Should the patient perform abnormally on the CO neuropsychiatric screening battery (CONSB), have a history of a soaking-type exposure or loss of consciousness, have abnormal neurologic findings (particularly cerebellar findings,119 exhibit cerebral edema on CT scan, or have evidence of cardiac ischemia, HBOT should be recommended. Despite this general practice, clinical predictors of DNS or PNS remain elusive and controversial. Even syncope unreliably predicts the need for HBOT.119 Similarly, laboratory markers such as COHb level, lactate level, or base deficit are unreliable factors in predicting DNS or PNS. Preliminary animal evidence points us toward the potential future use of laboratory markers such as peripheral lymphocyte cytochrome c oxidase, cyclic GMP, cholinergic muscarinic receptors and S100 B protein for determining patient prognosis.206,207 As suggested above, the use of COHb levels to guide therapy is controversial. A survey of medical directors of US and Canadian facilities indicated that 62% use a specific COHb level as the sole criterion for asymptomatic patients. The same survey found that when a specific COHb level was used as the indication for HBOT, 25% was the most common level chosen.208 See Table 16.1. Others suggest that HBOT is prescribed on the basis of COHb level in 40% of patients and in 60% on the basis of central nervous system or cardiac dysfunction.209 The patient’s clinical findings and history are of equal importance relative to COHb in determining the need for HBOT. Patients who could be considered candidates for HBOT include the pregnant patient with a COHb level greater than 10–15%, the patient with a history of coronary artery disease and a COHb level greater than 20%, the asymptomatic patient with a level

TABLE 16.1 Proposed Indications for Hyperbaric Oxygen Therapy in Pregnancy 1. 2. 3. 4. 5. 6. 7.

Abnormal CONSB (CO neuropsychiatric screening battery) Neurologic abnormalities (particularly cerebellar findings) Loss of consciousness (syncope) COHb > 25–40% Ongoing myocardial ischemia Worsened, recurrent,or refractory symptoms on NBO Relative considerations: soaking exposure, cerebral edema on CT scan

8417: “8417_c016” — 2007/9/11 — 12:15 — page 359 — #19

Carbon Monoxide Poisoning

360

greater than 25% to 40% COHb, or the patient with recurrent or persistent symptoms despite 6 hours of therapy with NBO. If HBOT is indicated, treatment within 6 h is desirable.210 Patients should undergo a full neuropsychiatric evaluation prior to discharge. Close follow-up is necessary with repeat neuropsychiatric examinations at 6 weeks, 6 months, and 12 months.

16.5.2 PREGNANCY The effects of CO on the fetus has been extensively reviewed by Penney.211 The fetus is particularly vulnerable to the effects of CO, which readily crosses the placenta, and, in animal models, is even more tightly bound to fetal hemoglobin than adult hemoglobin. The fetus also reaches higher peak COHb levels than does the mother. Fetuses that survive a significant CO poisoning may be left with limb malformation, hypotonia, areflexia, persistent seizures, mental and motor disability, and microcephaly.212,213 The only prospective, multicenter study of acute CO poisoning in pregnancy recently reported adverse outcomes in 60% of children whose mothers suffered severe CO toxicity. Of those babies born to mothers with mild to moderate CO exposure, normal physical exams and neurobehavioral development were reported.214 Since CO elimination from the fetus is prolonged (7–10 h), it is generally accepted that HBOT is indicated at lower maternal COHb levels than would be acted upon in the nongravid patient. In addition, surface oxygen therapy should be extended to four to five times the normal duration. Although controversial, HBO has been reported to be safe in the pregnancy,215 despite theoretical dangers of fetal hyperoxia in animal models.216−218 (Such animal models exceeded the time and pressure routinely used in clinical therapy). A recent report of 44 women undergoing HBOT during pregnancy for CO exposure suggests that it is safe and should be considered, although miscarriages did occur, and six patients were lost to follow-up.219 It should be noted that HBO was implicated in the induction of labor in one pregnant patient, the pregnancy however, was near term when the CO exposure occurred.220 Proposed indications for HBOT in the pregnant patient are listed in Table 16.2, although these are not well-studied.

16.5.3 CHILDREN Pediatric CO poisoning has been reviewed by White.221 Younger children have traditionally been viewed to be more susceptible to CO poisoning on the basis of more

TABLE 16.2 Proposed Indications for Hyperbaric Oxygen Therapy in Pregnancy 1. 2. 3. 4.

Maternal COHb level > 10–15% at any time during the exposure Any neurological signs or symptoms other than headache Evidence of fetal distress (fetal tachycardia, decreased beat-to-beat variability, late decelerations) If maternal neurologic symptoms or fetal distress persist 12 h after initial therapy, additional HBO treatments may be necessary

8417: “8417_c016” — 2007/9/11 — 12:15 — page 360 — #20

Treatment of Carbon Monoxide Poisoning

361

rapid metabolic rate and higher oxygen demands. They may also have more atypical presentations relative to adults. Both persistent and delayed sequelae are described in the children, however, formal neuropsychiatric testing is generally difficult and not well-documented in such case series.222 The use of HBOT in the treatment of pediatric CO poisoning is controversial, and recommendations vary, even among pediatric toxicology experts. Until further knowledge and experience is gained in this area, children are likely be treated as aggressively as adults who are CO-poisoned.

16.6 MANAGEMENT OF THE SEQUELAE OF CARBON MONOXIDE POISONING Delayed sequelae from CO poisoning is devastating and occurs in 10–43% of persons recovering from the acute exposure. Parkinsonism, the most dramatic long-term neurologic complication has a grim prognosis. Fortunately, most cases of DNS associated with mild CO poisoning resolve within two months.119 Unfortunately, only one-third of severely CO-poisoned patients surviving to HBOT have resolution of DNS. Conventional therapy of DNS-related parkinsonism with l-dopa has been disappointing. Use of another centrally acting dopaminergic agonist, bromocriptine has been reported. Nine patients (mean age 61 years) suffering from CO-induced parkinsonism who were given bromocriptine (5–30 mg/day), displayed improvement in Webster’s scores while under treatment.223 Clearly no definitive conclusions regarding bromocriptine therapy can be made on the basis of small study, but perhaps it will provide a basis for future investigations. Treatment with ziprasidone, a newer atypical antipsychotic agent resulted in improved neuropsychiatric symptoms and cognitive function in a patients with CO-induced severe DNS refractory to HBOT, bromocriptine, conventional antipsychotics, and other atypical antipsychotics, risperidone and quetiapine.224 Similar success was noted using aripiprazole in managing CO-induced psychotic symptoms and parkinsonism.225 One report involving hyperpyrexia and muscle rigidity as sequelae of CO poisoning was treated successfully with a prolonged course of dantrolene sodium, a peripheral skeletal muscle relaxant.226 Given that the patient manifested signs characteristic of severe hypoxic/ischemic encephalopathy, this therapy was symptomatic for that condition, and not specific to CO poisoning. Dantrolene would not likely provide any benefit beyond other safer sedatives, such as benzodiazepines, in treating such complications. Another common sequelae from CO poisoning is memory impairment. Recent work by Hiramatsu et al.227 focused on treating delayed amnesia in mice. The investigators treated mice with documented amnesia 5 days postexposure with dynorphin A (1–13). They found this treatment regimen to be effective in reversing CO-induced memory impairment. Nor-binaltrophimine (kappa opioid receptor antagonist) blocked the effect of dynorphin A (1–13), suggesting that kappa receptors mediated the reversal of impairment in memory seen from CO poisoning in this animal model.227 The authors reported similar findings with a second kappa receptor agonist, U-50488H, which appeared to additionally activate the cholinergic neuronal system, known also to play an important role in cognitive deficits associated with

8417: “8417_c016” — 2007/9/11 — 12:15 — page 361 — #21

Carbon Monoxide Poisoning

362

other conditions such as aging and neurodegenerative diseases.228 These agents may hold promise for the future in treating the persistent or delayed detrimental effects of CO on acquisition and consolidation of memory. Delayed, sometimes repetitive, HBOT has been advocated by some to improve the long-term neurologic deficits from CO toxicity, even if instituted weeks after the initial CO exposure.229−235 Such practice which is advocated by several treatment centers in the United States lacks validation by well-controlled, blinded clinical studies that utilize neuropsychiatric testing data. Interestingly, behavioral treatment has been successful when guided by formal neuropsychiatric testing. In certain patients, indirect measures of learning are better predictors of treatment efficacy.236 Patients who present to health care facilities late or have suffered recurrent or chronic lower level CO exposures pose particular treatment challenges to the clinician. Any proposed therapeutic approach to such patients should be considered carefully given the fact that no definitive clinical or animal studies in this area exist.

16.7 CONCLUSION Much remains to be learned about CO, including the mechanisms of toxicity, predictors of outcome after poisoning, and best treatments. Further research is needed to formulate clear-cut clinical indications for the use of potentially neuroprotective agents (i.e., insulin, sulfhydryl donors, allopurinol, ketamine, brain-derived peptides, kappa receptor agonists). Given that multiple pathways are involved in the ultimate neuronal injury, how to best use these agents, perhaps synergistically as a “cocktail” approach, remains to be seen. Other areas that deserve further study are the clarification of risk factors for adverse fetal outcome following CO exposure during pregnancy, the delineation of the true incidence of PNS and DNS in children, and the best test for indicators of risk for these sequelae. The future for HBOT in CO poisoning remains to be seen. Given the disparate results from randomized clinical trials using HBOT, we are compelled to continue to carefully select patients for this therapy, and to promote further study to delineate subpopulations such as children and pregnant women who may potentially benefit.

References 1. Cobb, N., Etzel, R.A. Unintentional CO-related deaths in the United States, 1979 through 1988, JAMA, 266, 659,1995. 2. Runciman, W.W., and Gorman, D.F. Carbon monoxide poisoning: from old dogma to new uncertainties, Med. J. Aust., 158, 439, 1993. 3. Hampson, N.B. Trends in the incidence of carbon monoxide poisoning in the United States, Am. J. Emerg. Med., 23, 838, 2005. 4. Smith, J.S., and Brandon, S. Morbidity from acute carbon monoxide poisoning at three-year follow up, Br. Med. J., 1, 279, 1983. 5. Lugaresi, A., Montagna, P., and Morreale, A. "Psychic akinesia" following carbon monoxide poisoning, Eur. Neurol., 30, 167, 1990. 6. Klawans, H.L., Stein, B.W., and Tanner, C.M. A pure parkinsonian syndrome following acute carbon monoxide intoxication, Arch. Neurol., 39, 302, 1982.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 362 — #22

Treatment of Carbon Monoxide Poisoning

363

7. Werner, B., Back, W., Akerblom, H., Barr, P.O. Two cases of acute carbon monoxide poisoning with delayed neurologic sequelae after a ‘free’ interval, Clin. Toxicol., 23, 249, 1985. 8. Hart, I.K., Kennedy, P.G., Adams, J.H., et al. Neurological manifestations of carbon monoxide poisoning, Postgrad. Med. J., 64, 213, 1988. 9. Sauk, G.M., Watson, C.P., Tebragge, K., et al. Delayed encephalopathy following carbon intoxication, Can. J. Neurol. Sci., 8, 77, 1981. 10. Barret, L., Danel, V., and Faure, J. Carbon monoxide poisoning, a diagnosis frequently overlooked, Clin. Toxicol., 23, 309, 1985. 11. Balzan, M.A., Agius, G., and Debono A.G. Carbon monoxide poisoning, easy to treat but difficult to recognise, Post grad. Med. J., 72,470,1996. 12. Sokal, J.A., Majka, J., and Palus, J. The content of carbon monoxide in the tissues of rats intoxicated with carbon monoxide in various conditions of acute exposure, Arch Toxicol, 56, 106, 1984. 13. Lasater, S.R. Carbon monoxide poisoning, Can. Med. Assoc. J., 134, 991, 1986. 14. Olson, K.R. Carbon monoxide poisoning: mechanisms, presentation, and controversies in management, J. Emerg. Med., 1, 233, 1984. 15. Ikeda, T., Kondo, T., Mogami, H., Miura, T., Mitomo, M., Shimazaki, S., and Sugimoto, T. Computerized tomography in cases of acute carbon monoxide poisoning, Med. J. Osaka Univ., 29, 253, 1978. 16. Penney, D.G. Hyperglycemia exacerbates brain damage in acute severe carbon monoxide poisoning, Med. Hypotheses, 27, 241, 1988. 17. Bogusz, M., Cholewa, L., Pach, J., and Mlodkowska, K. A comparison of two types of acute carbon monoxide poisoning, Arch. Toxicol., 33, 141, 1975. 18. Linas, A.J., and Limousin, S. Bul. Mem. Soc. Ther., 2, 32, 1868. 19. Haldane, J. The relation of the action of carbonic oxide to oxygen tension, J. Physiology, 18, 201, 1895. 20. Killick, E.M., and Marchant, J.V. The effect of barbiturates on the resuscitation of dogs from severe acute CO poisoning, J. Physiol., 180, 80, 1965. 21. Killick, E.M., and Marchant, J.V. Resuscitation of dogs from severe, acute carbon monoxide poisoning, J. Physiol., 147, 274, 1959. 22. Schwerma, H., Ivy A.C., Friedman, H., and Brosse, E.L. A study of resuscitation from the juxtalethal effects of exposure to carbon monoxide, Occup. Med., 5, 24, 1948. 23. Jain, K.K. Carbon Monoxide Poisoning, Warren H. Green, Inc., St Louis, 1990, 140. 24. Gros, J.F., and Leandri, P. Traitment de l’intoxication oxycarbonée par le cytochrome, Presse Méd, 64, 1356, 1956. 25. Bentolila, P., Tran, G., and Olive, G. Essai de traitment de l’intoxication oxycarbonée par perfusion de solutions diluées de paroxyde d’hysrogene, resultats obtnus chez le lapin, Thérapie, 28, 1043, 1973. 26. Koza, F. Die kohlenmonoxidevergiftungund deren neurartige therapie mit bestrahlung, Med. Klinik, 26, 422, 1930. 27. Estler, W. Experimentelle untersuchungen uber die anwendung der bestrahlung mit ultravoilettem licht zur behandlung der kohlenoxidvergiftung, Arch. Hyg. Bakt., 115, 152, 1935. 28. Amyes, E.W., Ray, J.W., and Brockman, N.W. Carbon monoxide anoxia; intravenous administration of procaine hydrochloride in the treatment of acute and chronic effects, JAMA, 142, 1054, 1950. 29. Evans, D.E., Catron, P.W., McDermott, J.J., et al. Effect of lidocaine after experimental cerebral ischemia induced by air embolism, J. Neurosurg., 70, 97, 1989.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 363 — #23

364

Carbon Monoxide Poisoning 30. Ojeda-Sahagun, J.L. Accion protectora del dipiridamol frente a la intoxicacion por gas del alumbrado, Rev. Esp. Fisiol., 27, 305, 1971. 31. Agostini, J.C., Ramirez, R.G., Albert, S.N., Goldbaum, L.R., Absolon, K.B. Successful reversal of lethal carbon monoxide intoxication by total body asanguineous hypothermic infusion, Surgery, 75, 213, 1974. 32. Yee, L.M., and Brandon, G.K. Successful reversal of presumed carbon monoxideinduced semicoma, Aviat. Space Envir. Md., 54, 641, 1983. 33. Shumate, M.J. Carbon monoxide poisoning, Chest, 107, 1474, 1995. 34. Geyer, R.P., Haggard, H.W. Review of perfluorochemical-type blood substitutes, In Proc. 10th Int. Cong. Nutrition-Symposium on Perfluorochemical Artificial Blood, Igakushobe, Osaka Japan, 1976, 3. 35. Levine, E.M., and Tremper, K.K. Perfluorochemical emulsions: potential clinical uses and new developments, Int. Anesthesiol. Clin., 23, 211, 1985. 36. Uchiyama, T., and Harafuji, K. Life-saving effect of pyridoxalated hemoglobinpolyoxyethylene conjugate on carbon monoxide intoxication of rabbits, Artif. Organs,18, 576, 1994. 37. Mosso, A. La mort apparente du coer secours l’empoisonneement par l’oxyd de carbonee, Arch. Ital. Biol., 35, 75, 1901. 38. Smith, G., and Sharp, G.R. Treatment of carbon monoxide poisoning with oxygen under pressure, Lancet, ii, 905, 1960. 39. Sluitjer, M.E. The treatment of carbon monoxide poisoning by administration of oxygen at high pressure, P. Roy. Soc. Med., 56, 123, 1963. 40. Hampson, N.B., and Little, C.E. Hyperbaric treatment of patients with carbon monoxide poisoning in the United States, Undersea Hyper. Med., 32, 21, 2005. 41. Dinman, B., Eaton, J., and Brewer, G. Effects of carbon monoxide on DPG concentrations in the erythrocyte, Ann. NY Acad. Sci., 1974, 246, 1970. 42. Klasner, A.E., Smith, S.R., Thompson, M.W., and Scalzo, A.J. Carbon monoxide mass exposure in a pediatric population, Acad. Emerg. Med., 5, 992, 1998. 43. Decker, J., and Wiley J. F. Carboxyhemoglobin elimination half-life in children, presented at the North American Congress of Clinical Toxicology, Salt Lake City, Utah, September, 1994. 44. Okeda R., Funata, N., Song, S.J., et al. Comparative study of selective cerebral lesions in carbon monoxide poisoning and nitrogen hypoxia in cats, Acta. Neuropathol., 56, 265, 1982. 45. Haldane, J.B.S. Carbon monoxide as a tissue poison, Biochem. J., 21, 1086, 1927. 46. Brabkin, D.L., Lewey, F.H., Bellet, S., et al. The effect of replacement of normal blood by erythrocytes saturated with carbon monoxide, Am. J. Med. Sci., 205, 755, 1943. 47. Goldbaum, L.R. Studies on the relationship between carboxyhemoglobin concentration and toxicity, Aviat. Space Envir. Med., 48, 969, 1977. 48. Goldbaum, L.R. Mechanism of the toxic action of carbon monoxide, Ann. Clin. Lab. Sci., 6, 372, 1976. 49. Goldbaum, L.R., Orellano, T., and Dergal, E. What is the mechanism of Carbon monoxide toxicity? Aviat. Space Envir. Med., 46, 1289, 1975 50. Brown, S., and Piantadosi, C. In vivo binding of CO to cytochrome oxidase in rat brain, J. Appl. Physiol., 68, 604, 1990. 51. Brown, S., and Piantadosi, C.A. Recovery of energy metabolism in rat brain after CO hypoxia, J. Clin. Invest., 89, 666, 1992. 52. Coburn, R.F., and Forman, H.J. Carbon monoxide toxicity, In: Fishman, A.P., Farki, L.E., and Geiger S.R. eds., Handbook of Physiology, The Respiratory System, Sect 3, Vol. IV, Chapt. 21, 439, Baltimore: Williams & Wilkins, 1987.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 364 — #24

Treatment of Carbon Monoxide Poisoning

365

53. Smithline, H.A., Rivers, E.P., Chiulli, D.A., Rady, M.Y., Baltarowich, L.L., Blake, H.C., et al. Systemic hemodynamic and oxygen transport response to graded carbon monoxide poisoning. Ann. Emerg. Med., 5, 203, 1993. 54. Ward, K.R., Lin, J., Zhang, L., and Chopp, M. Carbon monoxide poisoning does not cause neuronal heat shock protein 72 or 32 expression, Acad. Emerg. Med., 5, 487, 1998. 55. Dallman, P., and Schwartz, H. Distribution of cytochrome C and myoglobin in rats with dietary iron deficiency, Pediatrics, 35, 677, 1965. 56. Webber, D.S., Lopez, I., Korsak, R.A., Hirota, S., Acuna, D., and Edmond, J. Limiting iron availability confers neuroprotection from chronic mild carbon monoxide exposure in the developing auditory system of the rat, J. Neurosci. Res., 80, 620, 2005. 57. Wittenberg, B.A., and Wittenberg, J.B. Effects of carbon monoxide on isolated heart muscle cells, Research Report-Health Effects Institute, 62, 1, 1993. 58. Aslan, S., Uzkeser, M., Seven, B., Gundogdu, F., Acemoglu, H., Akasakal, E., and Varoglu, E. The evaluation of myocardial damage in 83 young adults with carbon monoxide poisoning in the East Anatolia region in Turkey, Hum. Exp. Toxicol., 2006, 25, 439, 2006. 59. Satran, D., Henry, C.R., Adkinson, C., Nicholson, C.I., Bracha, Y., and Henry, T.D. Cardiovascular manifestations of moderate to severe carbon monoxide poisoning, J. Am. Coll. Cardiol., 45, 1513, 2005. 60. Henry, C.R., Satra, D., Lindgren, B., Adkinson, C., Nicholson, C.I., and Henry, T.D. Myocardial injury and long-term mortality following moderate to severe carbon monoxide poisoning, JAMA, 295, 398, 2006. 61. Yucel, M., Avsarogullari, L., Durukan, P., Akdur, O., Ozkan, S., Sozuer, E.M., Muhtaroglu, S., and Ikizceli, I. The relationship between the severity of myocardial injury and the levels of brain natriuretic peptide and cardiac troponin-I in rats experimentally exposed to carbon monoxide, Eur. J. Emerg. Med., 13, A11 (abstract), 2006. 62. Aurora, T., Chung, W., Dunne, R., Martin, G., Ward, K., Rivers, E., Knoblich, B., Nguyen, H.B., and Tomlanovich M.C. Occult myocardial injury in severe carbon monoxide poisoning, Acad. Emerg. Med., 6, 394 (abstract), 1999. 63. Thom, S.R. Carbon monoxide-mediated brain lipid peroxidation in the rat, J. Appl. Physiol., 68, 997, 1990. 64. Thom, S.R. Leukocytes in carbon monoxide-mediated brain oxidative injury, Toxicol. Appl. Pharmacol., 123, 234, 1993. 65. Thom, S.R. Dehydrogenase conversion to oxidase and lipid peroxidation in brain after carbon monoxide poisoning, J. Appl. Physiol. ,73, 1584, 1992. 66. Thom, S.R., Garner S., and Fisher D. Vascular oxidative stress from carbon monoxide exposure, Undersea Hyperbar. Med., 25 (Suppl), 47, 1998. 67. Thom, S.R., Xu, Y.A., and Ischiropoulos, H. Vascular endothelial cells generate peroxynitrite in response to carbon monoxide exposure, Chem. Res. Toxicol., 10, 1023, 1997. 68. Ischiropoulos, H., Beers, M.F., Ohnishi, S.T., Fisher, D., Garner, S.E., and Thom, S.R. Nitric oxide production and perivascular nitration in brain after carbon monoxide poisoning in the rat. J. Clin. Invest., 97, 2260, 1996. 69. Hall, E.D., Andrus, P.K., Althaus, J.S., et al. Hydroxyl radical production and lipid peroxidation parallels selective post-ischemic vulnerability in gerbil brain, J. Neurosci. Res. 1993, 34, 107. 70. Zhang, J., and Piantadosi, C.A. Mitochondrial oxidative stress after carbon monoxide hypoxia in the rat, J. Clin. Invest., 90, 1193, 1992.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 365 — #25

366

Carbon Monoxide Poisoning 71. Thom, S.R., Bhopale, V.M., Han, S-T., Clark, J.M., and Hardy, K.R. Intravascular neutrophil activation due to carbon monoxide poisoning, Am. J. Respir. Crit. Care Med., 2006, 174, 1239–1248. 72. Thom, S.R., Bhopale, V.M., Fisher, D., Zhang, J., and Gimotty, P. Delayed neuropathology after carbon monoxide poisoning is immune-mediated, PNAS, 101, 13660, 2004. 73. Thom, S.R., Fisher, D., Zhang, J., et al. Neuronal nitric oxide synthase and N-methyl-D-aspartate neurons in experimental carbon monoxide poisoning, Toxicol. Appl. Pharmacol., 1914, 280, 2004. 74. Piantadosi, C.A. Toxicity of carbon monoxide, hemoglobin vs. histotoxic mechanisms, In Carbon Monoxide, Penney, D.G., ed., Boca Raton, FL, CRC Press, Inc, 1996, Chapt. 8, pp. 163–186. 75. Amara, S.G. Neurotransmitter transporters, Nature, 360, 420, 1992. 76. Hernandez-Viadel, M., Castoldi, A.F., Coccini, T., Manzo, L., Erceg, S., and Felip, V. In vivo exposure to carbon monoxide causes delayed impairment of activation of soluble guanylate cyclase by nitric oxide in rat brain cortex and cerebellum, J. Neurochem., 89, 1157, 2004. 77. Meilin, S., Rogatsky, G.G., Thom, S.R., Zarchin, N., Guggenheimer-Furman, E., and Mayevsky, A. Effects of carbon monoxide on the brain may be mediated by nitric oxide, J. Appl. Physiol., 81, 1078, 1996. 78. Shinomura, T., Nakao, S., and Mori, K. Reduction of depolarization-induced glutamate release by hemoxygenase inhibitor, possible role of carbon monoxide in synaptic transmission, Neurosci. Lett., 166, 131, 1994. 79. Gorman, D., Lin, H.Y., and Williams, C. Blockade of haemoxygenase and nitric oxide synthetase causes cortical dysfunction in sheep exposed to carbon monoxide, Toxicology, 209, 237, 2005. 80. Zhang, J., and Piantadosi, C.A. Prevention of H2 O2 generation by monoamine oxidase protects against CNS O2 toxicity, J. Appl. Physiol., 71, 1057, 1991. 81. Piantadosi, C.A., Tatro, L., and Zhang, J. Hydroxyl radical production in the brain after CO hypoxia in rats, Free Radical Bio. Med., 18, 603, 1995. 82. Pianttadosi, C.A., Zhang, J., and Levin, E.D., et al. Apoptosis and delayed neuronal damage after carbon monoxide poisoning in the rat, Exp. Neurol., 147, 103, 1997. 83. Ward, K.R., Junmin, L, Zhang, L, and Chopp, M. Moderately severe carbon monoxide poisoning does not cause neuronal apoptosis, Acad. Emerg. Med., 5, 487, 1998, abstract. 84. Thom, S.R., Fisher, D., Xu, Y.A., et al. Adaptive responses and apoptosis in endothelial cells exposed to carbon monoxide, P. Natl. Acad. Sci. USA, 97, 1305, 2000. 85. Van Der Hoeven, J.G., Compier, E.A., and Meinders, A.E. Goede resultaten van behandeling met 100% zuurstop wegens acute koolmonoxide-intoxicatie; voorlopig geen indicatie, voor hyperbare zuurstoftoediening, Ned. Tijdschr. Genees., 137, 864, 1993. 86. Weaver, L.K., Hopkins, R.O., and Larson-Lohr, V. Neuropsychologic and functional recovery from severe carbon monoxide poisoning without hyperbaric oxygen therapy, Ann. Emerg. Med., 27, 736, 1996. 87. Meert, K.L., Heidemann, S.M., and Sarnaik, A.P. Outcome of children with carbon monoxide poisoning treated with normobaric oxygen, J. of Trauma, 44, 149, 1998. 88. Tomaszewski, C., Rudy, J., Wathen, J., Brent, J., Rosenberg, N., and Kulig, K. Prevention of neurologic sequelae from carbon monoxide by hyperbaric oxygen in rats, Ann. Emerg. Med., 21, 631–632, 1992.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 366 — #26

Treatment of Carbon Monoxide Poisoning

367

89. Thom, S. Functional inhibition of neutrophil beta 2 integrins by HBO in CO-mediated brain injury, Toxicol. Appl. Pharmacol., 123, 248–256, 1993. 90. Thom, S.R. Mendiguren, I., Hardy, K., et al. Inhibition of human neutrophil beta2integrin-dependent adherence by hyperbaric oxygen, Am. J. Physiol., 272, C770, 1997. 91. Schnittger, V., Rosendahl, K., Lind, F., Pamblad, J. Effects of carbon monoxide poisoning on neutrophil response in patients treated with hyperbaric oxygen, J. Invest. Med., 52, 523, 2004. 92. Brown, S.D., and Piantadosi, C.A. Reversal of carbon monoxide-cytochrome c oxidase binding by hyperbaric oxygen in vivo, In Adv. Exp. Med. Biol., K. Rakusan, G.D. Biro, T.K. Goldstick, Eds, Plenum Press, New York, Vol. 248, 1989, pp. 747–754. 93. Brvar, M., Finderle, Z., Suput, D., and Bunc, M. S100B protein in conscious carbon monoxide-poisoned rats treated with normobaric or hyperbaric oxygen, Crit. Care Med., 34, 2228, 2006. 94. Thom, S.R., Antagonism of carbon monoxide-mediated brain lipid peroxidation by hyperbaric oxygen, Toxicol. Appl. Pharmacol., 105, 340, 1990. 95. Sukoff, M.H., and Ragatz, R.E. Hyperbaric oxygen for the treatment of acute cerebral oedema, Neurosurgery, 10, 29, 1982. 96. Jiang, J., and Tyssebotn, I. Cerebrospinal fluid pressure changes after acute carbon monoxide poisoning and therapeutic effects of normobaric and hyperbaric oxygen in conscious rats, Undersea Hyperbar. Med., 24, 245, 1997. 97. Roos, J.A., Jackson-Friedman, C., and Lyden, P. Effects of hyperbaric oxygen on neurologic outcome for cerebral ischemia in rats, Acad. Emerg. Med., 5, 18, 1998. 98. Kokame, G.M., and Shuler, S.E. Carbon monoxide poisoning, Arch. Surg., 96, 211, 1968. 99. Myers, R.A.M., Snyder, S.K., Linberg, and S., Cowley, R.A. Value of hyperbaric oxygen in suspected carbon monoxide poisoning, JAMA, 246, 2478, 1981. 100. Winter, A., and Shatin, L. Hyperbaric oxygen in reversing carbon monoxide coma, New York State J. Med., 1, 880, 1970. 101. Thomson, L.F., Mardel, S.N., Jack, A., and Shields, T.G. Management of the moribund carbon monoxide victim, Arch. Emerg. Med, 9, 208, 1992. 102. Welsh, F., Matos, L., and DeTreville, R.T.P. Medical hyperbaric oxygen therapy: 22 cases, Aviat., Space Envir. Md. , 51, 611, 1980. 103. Gibson, A.J., Davis, F.M., and Ewer, T. Delayed hyperbaric oxygen therapy for carbon monoxide intoxication- two case reports, New Zeal. Med. J., 104, 64, 1991. 104. Dean, B.S., Verdile, V.P., and Krenzelok, E.P. Coma reversal with cerebral dysfunction recovery after repetitive hyperbaric oxygen therapy for severe carbon monoxide poisoning, Am. J. Emerg. Med., 11, 61, 1993. 105. Ersanli, D., Yildiz, S., Togrol, E., Ay, H., and Qyrdedi, T. Visual loss as a late complication of carbon monoxide poisoning and its successful treatment with hyperbaric oxygen therapy, Swiss Med. Wkly., 134, 650, 2004. 106. Roche, L., Bertoye, A., Vincent, P., et al. Comparison de deux groupes de vingt intoxications oxycarbonees traitees par oxygene normobare et hyperbare, Lyon Med., 49, 1483, 1968. 107. Mathieu, D., Nolf, M., Durocher, A., Saulnier, F., Frimat, P., Furon, D., and Wattel, F. Acute carbon monoxide poisoning risk of late sequelae and treatment by hyperbaric oxygen, Clin. Toxicol., 23, 315, 1985. 108. Gorman, D.F., Clayton, D., Gilligan, J.E., and Webb, R.K. A longitudienal study of 100 consecutive admissions for carbon monoxide poisoning to the Royal Adelaide hospital, Anaesth. Intens. Care, 20, 311, 1992.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 367 — #27

368

Carbon Monoxide Poisoning

109. Myers, R.A.M., Snyder, S.K., and Emhoff, T.A. Subacute sequelae of carbon monoxide poisoning, Ann. Emerg. Med., 14, 1163, 1985. 110. Hsu, L.H., and Wang, J.H. Treatment of carbon monoxide poisoning with hyperbaric oxygen, Chin. Med. J-Peking, 58, 407, 1996. 111. Gozal, D., Ziser, A., Shupak, A., and Melamed, Y. Accidental carbon monoxide poisoning, Clin. Pediatr. 24, 132, 1985. 112. Norkool, D.M., and Kirkpatrick, J.N. Treatment of acute carbon monoxide poisoning with hyperbaric oxygen: a review of 115 cases, Ann. Emerg. Med., 14, 1168, 1985. 113. Raphael, J.,C., Elkharrat, D., Jars-Guincestre, M.C., et al. Trial of normobaric and hyperbaric oxygen for acute carbon monoxide intoxication, Lancet, 1, 414, 1989. 114. Weaver, L.K. Carbon monoxide poisoning, Crit. Care Clin., 15, 297, 1999. 115. Ducasse, J.L., Izard, P.H., Celsis, P., et al. Moderate carbon monoxide poisoning: hyperbaric or normobaric oxygenation? Proceedings of the Joint Meeting of the 2nd European Conference and the 2nd Swiss Symposium on Hyperbaric Medicine, Basel, Switzerland, 1990, pp. 289–297. 116. Olson, K.R., and Seger, D. Hyperbaric oxygen for carbon monoxide poisoning: does it really work, Ann. Emerg. Med., 25, 535, 1995. 117. Weaver, L.K., Hopkins, R.O., and Larson-Lohr, V. Hyperbaric oxygen and carbon monoxide poisoning, Ann. Emerg. Med., 2693, 390, 1995. 118. Scheinkestel, C.D., Bailey, M., Myles, P.S., Jones, K., Cooper, D.J., Millar, I.L., and Tuxen, D.V. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial, MJA, 170, 203, 1999. 119. Weaver, L.K., Hopkins, R.O., Howe, S., Larson-Lohr, V., and Churchill, S. Outcome at 6 and 12 months following acute CO poisoning, Undersea Hyperbar. Med., 23S, 9, 1996. 120. Weaver, L.K., Hopkins, R.O., Chan, K.J., et al. Hyperbaric oxygen for acute carbon monoxide poisoning, N. Engl. J. Med., 347, 1057, 2002. 121. Mathieu, D., Wattel, F., Mathiew-Nolf, M., et al. Randomized prospective study comparing the effect of HBO versus 12 hours NBO in non-comatose CO poisoned patients: results of the interim analysis, Undersea Hyperbar. Med., 23 (Suppl.), 7, 1996. 122. Raphael, J.C., Chevret, S., Driheme, A., and Annane, D. Managing carbon monoxide poisoning with hyperbaric oxygen (abstract), J. Toxicol. Clin. Toxicol., 42, 455, 2004. 123. Gorman, D.F., Clayton., D., and Gilligan, J.E. A longitudinal study of 100 consecutive admissions for carbon monoxide poisoning to the Royal Adelaide Hospital, Anesth. Intens. Care, 20, 311, 1992. 124. Turner, M., Esaw, M., and Clark, R.J. Carbon monoxide poisoning treated with hyperbaric oxygen: metabolic acidosis as a predictor of treatment requirements, J. Accid. Emerg. Med. 16, 96, 1999. 125. Tibbles, P.M., and Edelsberg, J.S. Hyperbaric oxygen therapy, N. Engl. J. Med., 334, 1642, 1996. 126. McNulty, J.A., Maher, B.A., Chu, M., and Sitnikova, T. Relationship of short-term verbal memory to the need for hyperbaric oxygen treatment after carbon monoxide poisoning, Neuropsy., Neuropsy. Be., 10, 174, 1997. 127. Murata, M., Suzuki, M., Hasegawa,Y., Nohara, S., and Kurachi, M. Improvement of occipital alpha activity by repetitive hyperbaric oxygen therapy in patients with carbon monoxide poisoning: a possible indicator for treatment efficacy, J. Neurological Sciences, 235, 69, 2005.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 368 — #28

Treatment of Carbon Monoxide Poisoning

369

128. Hampson, N.B., Simonson, S.G., and Kramer, C.C. Central nervous system oxygen toxicity udring hyperbaric treatment of patients with carbon monoxide poisoning, Undersea Hyperbar Med., 23, 215, 1996. 129. Weaver, L.K., Hopkins, R.O., Valentine, K.J., and Churchill, S. Persistent and delayed carbon monoxide cognitive sequelae, Abstract presented at the Undersea and Hyperbaric Medical Society Annual Scientific Meeting, June, 2006, Las Vegas, NV. 130. Scheinkestel, C.D., Jones, K., Myles, P.S., Cooper, J., Miller, I.L., and Tuxen, D.V. Where to now with carbon monoxide poisoning? Emerg. Med. Australas., 16, 151, 2004. 131. Weaver, L.K., Hopkins, R.O., Chan, K.J., Thomas, F., Churchill, S.K., Elliott, C.G., and Morris, A. Carbon monoxide research group, LDS Hospital, Utah in reply to Scheinkestel et al. and Emerson: The role of hyperbaric oxygen in carbon monoxide poisoning, Emerg. Med. Australas., 16, 394, 2004. 132. Emerson, G. The dilemma of managing carbon monoxide poisoning, Emerg. Med. Australas., 16, 101, 2004. 133. Juurlink, D.N., Buckley, N.A., Stanbrook, M.B., Isbister, G.K., Bennett, M., and McGuigan, M.A. Hyperbaric oxygen for carbon monoxide poisoning (review), Cochrane Db. Syst. Rev. 2005, Issue 1, Art. No.:CD002041.DoI:10.1002/14651858. CD002041. pub 2. 134. Buckley, N.A., Isbister, G.K., and Juurlink, D.N. Hyperbaric oxygen for carbon monoxide poisoning: evidence versus opinion, Toxicol. Rev., 24, 159, 2005. 135. Thom, S.R. Hyperbaric oxygen therapy for carbon monoxide poisoning: is it time to end the debates, Toxicol. Rev., 24, 157, 2005. 136. Bentur, Y. Hyperbaric oxygen for carbon monoxide poisoning, Toxicol. Rev., 24, 153, 2005. 137. Olson, K.R. Hyperbaric or normobaric oxygen? Toxicol. Rev., 24, 151, 2005. 138. Henry, J.A. Hyperbaric oxygen therapy for carbon monoxide poisoning, To treat or not to treat, that is the question, Toxicol. Rev., 24, 149, 2005. 139. Marklund, S.L., Westman, G., Lundgren, E., and Roos, G. CuZn superoxide dismutase, Mn superoxide dismutase, catalase and glutathione peroxidase in normal and neoplastic human cell-lines and normal human tissues, Cancer Res., 42, 1955–1961, 1982. 140. Toledo-Pereyra, L.H., Simmons, R.L., and Najarian, J.S. Effect of allopurinol on the preservation of ischemic kidneys perfused with plasma or plasma substitutes, Ann. Surg., 180, 780–782, 1974. 141. Stewart, J.R., Blackwell, W.H., Crute, S.L., Loughlin, V., Hess, M.L., and Greenfield, L.J. Prevention of myocardial ischemia / reperfusion injury with oxygen-free radical scavengers, Surg. Forum, 33, 317–320, 1982. 142. Fechter, L.D., Liu, Y., and Pearce, T.A. Cochlear protection from carbon monoxide exposure by free radical blockes in the guinea pig, Toxicol. Appl. Pharm., 142, 47, 1997. 143. Howard, R.J., Blake, D.R., Pall, H., Williams, A., and Green, I.D. Allopurinol/nacetylcysteine for carbon monoxide poisoning, Lancet, 2, 628–629, 1987. 144. Gilmer, B., Tomaszewski, C., and Watts, J.A. The neuroprotective effects of dimethyl sulfoxide on memory following actue carbon monoxide poisoning in mice, Ann. Emerg. Med., 35, S69, 2000. 145. Penney, D.G., Helfman, C.C., Hull, J.C., Dunbar, J.C., and Verman, K. Elevated blood glucose is associated with poor outcome in the carbon monoxide poisoned rat, Toxicol. Lett., 54, 287, 1990.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 369 — #29

370

Carbon Monoxide Poisoning

146. Penney, D.G. Hyperglycemia exacerbates brain damage in acute–severe carbon monoxide poisoning, Med. Hypoth., 27, 241–244, 1988. 147. Leikin, J.B., Goldenberg, R.R., Edwards, D., and Zell-Kantor, M. Metabolic predictors of carbon monoxide poisoning, Vet. Hum. Toxicol., 30, 40–42, 1988. 148. Pulsinelli, W., Waldman, S., Sigsbee, B., Rawlinson, D., Schere, P., and Plum, F. Experimental hyperglycemia and diabetes mellitus worsen stroke outcome, T. Am. Neurol. Assoc., 105, 21–24, 1980. 149. White, S.R., and Penney, D.G. Effects of insulin and glucose treatment on neurologic outcome after carbon monoxide poisoning, Ann. Em. Med., 23, 606, 1994. 150. Bouhaddi, K., Thomopoulos, P., Fages, C., Khelil, M., and Tardy, M. Insulin effect on GABA uptake in astroglial primary cultures, Neurochem. Res., 13, 1119, 1988. 151. Stahl, W. The Na-K-ATPase of nervous tissue, Neuro. Int., 8, 449, 1988. 152. MaCara, I.G. Oncogenes, ions, and phospholipids, Am. J. Physiol., 248, C3–C11, 1985. 153. Standaert, M.L., and Pollet, R.J. Insulin-glycerolipid mediators and gene expression, FASEB J., 2, 2453, 1988. 154. Goldfine, I.D. The insulin receptor: molecular biology and trans membrane signaling, Endocrinology, 8, 235, 1987. 155. Goldfine, I.D., Purrello, F., Vigneri, R., and Clawson, G.A. Direct regulation of nuclear functions by insulin: relationship to mRNA metabolism, in Molecular Basis of Insulin Action, Czech, M.P., ed., Plenum, New York, 1985, 329. 156. Moss, A.M., Unger, J.W., Moxley, R.T., and Livingston, J.N. Location of phosphotyrosine-containing proteins by immunocytochemistry in the rat forebrain corresponds to the distribution of the insulin receptor, Proc. Natl. Acad. Sci., 87, 4453, 1990. 157. Beneviste, H., Drejer, J., Schousboe, A., and Diemer, N. H. Elevation of extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischaemia monitored by intracerebral microdialysis, J. Neurochem., 43, 1369, 1984. 158. Jorgensen, M.B., and Diemer, N.H. Selective neuron loss after cerebral ischaemia in the rat: possible role of transmitter glutamate, Acta. Neurol. Scand., 66, 536, 1982. 159. Simon, R.P., Swan, J.H., Griffith, T., and Meldrum, B.S. Blockade of N-methyl-Daspartate receptors may protect against ischaemic damage in the brain, Science, 226, 850, 1984. 160. Gill, R., Foster, A.C., and Woodruff, G.N. Systemic administration of MK-801 protects against ischemia-induced hippocampal neurodegeneration in the gerbil, J. Neurosci., 3343, 1987. 161. Ishimaru, H., Katoh, A., Suzuki, H., Fukuta, T., Kameyama, T., and Nabeshima, T. Effects of N-methyl-D-aspartate receptor antagonists on carbon monoxide induced brain damage in mice, J. Pharm. Exp. Ther., 261, 349–352, 1991. 162. Liu, Y., and Fechter, L.D. MK-801 protects against carbon monoxide induced hearing loss, Toxicol. Appl. Pharm., 132, 196, 1995. 163. Gilmer, B., Thompson, C., Tomaszewski, C., and Watts, J.A. The protective effects of experimental neurodepressors on learning and memory following carbon monoxide poisoning, J. Toxicol. Clin. Toxicol. 36, 606, 1999. 164. Anis, N.A., Berry, S.C., Burton, N.R., and Lodge, D. The dissociative anaesthetics, ketamine and phencyclidine, selectively reduce excitation of the central mammalian neurons by N-methyl-aspartate, Br. J. Pharmacol., 79, 565, 1983. 165. Penney, D.G., and Chen, K. NMDA receptor-blocker ketamine protects during acute carbon monoxide poisoning, while calcium channel-blocker verapamil does not, J. Appl. Toxicol., 16, 297, 1996.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 370 — #30

Treatment of Carbon Monoxide Poisoning

371

166. Piswanger, A., Paier, B., and Windisch, M. Modulation of protein synthesis in a cell free system from rat brain by Cerebrolysin during development and aging, In Amino Acids, Lubec, G., and Rosenthal, G.A., eds., ESCOM Science Publishers, 1990, p. 651. 167. Akai, F., Himura, S., Sato, T., Iwamoto, N., Fufimoto, M., Ioku, M., and Hashimoto, S. Neurotrophic factor-like effect of FPF 1070 on septal cholineric neurons after transections of fimbria-fornix in the rat, Histol. Histopathol., 7, 213, 1992. 168. Satou, T., Imano, F., Akai, S., Himura, S., Hashimoto, T., Itoh, T., and Fujimoto, M. Morphological obeservation of effects of Cerebrolysin on cultured neural cells, In Alzheimer’s Disease and Related Disorders, Nicolini, M., Zatta P.F.,and Coraine, B., eds., Pergamon, Oxford, 1993, pp. 195–96. 169. Boado, R.J. Brain-derived peptides increase the expression of a blood-brain barrier GLUT1 glucose transporter reporter, Neurosci. Lett., 220, 53, 1996. 170. Koroleva, V.I., Korolev, O.S., Mares, V., Pastalkova, and E., Bures, J. Hippocampal damage induced by carbon monoxide poisoning and spreading depression alleviated by chronic treatment with brain derived polypeptides, Brain Res., 816, 618, 1999. 171. Sluitjer, M.E. The treatment of carbon monoxide poisoning by administration of oxygen at high pressure, P. Roy. Soc. Med., 56, 1002–1008, 1963. 172. Peirce, E.C., Zacharias, A., Alday, J.M., Hoffman, B.A., and Jacobson, J.H. Carbon monoxide poisoning: experimental hypothermic and hyperbaric studies, Surgery, 72, 229–237, 1972. 173. Boutros, A.R., and Hoyt, J.L. Management of carbon monoxide poisoning in the absence of hyperbaric oxygenation chamber, Crit. Care Med., 4, 144–147, 1976. 174. Norman, J.N., and Ledingham, I.McA. Carbon monoxide poisoning: investigations and treatment, P. Roy. Soc. Med., 56, 101, 1963. 175. Rocco, M., Carbone, I., Morelli, A., Palantonio, P., Rossi, S., Spadetta, G., Passariello, R., and Pietropaoli, P. The calcium sensitizer levosimendan improves carbon monoxide poisoning related stunned myocardium: a cardiac magnetic resonance study, Acta. Anaesth. Scand., 50, 897, 2006. 176. Radushevich, V.P., and Koroteeva, E.L. Parallel blood circulation with oxygenation of blood in severe poisoning with carbon monoxide fumes, Vestn. Khir., 116, 131–134, 1976. 177. Baud, F.J. Value of lactic acidosis in the assessment of the severity of acute cyanide poisoning, Crit. Care Med., 30, 2044, 2002. 178. Moore, S.J., Norris, J.C., Walsh, D.A., et al. Antidotal use of methemoglobin-forming cyanide antagonists in concurrent carbon monoxide/cyanide intoxication, J. Pharm. Exp. Ther., 242, 70–73, 1987. 179. Kirk, M.A., Gerace, R., and Kulig, K.W. Cyanide and methemoglobin kinetics in smoke inhalation victims treated with the cyanide antidote kit, Ann. Emerg. Med., 22, 1413–1418, 1993. 180. Seger, D., and Welch, L. Carbon monoxide controversies: neuropsychologic testing, mechanism of toxicity, and hyperbaric oxygen, Ann. Emerg. Med., 24, 242–248, 1994. 181. Hopkins, R.O., and Woon, F.L.M. Neuroimaging, cognitive, and neurobehavioral outcomes following carbon monoxide poisoning, Behavioral and Cognitive Neuroscience Reviews, 5, 141, 2006. 182. Pulsipher, D.T., Hopkins, R.O., and Weaver, L.K. Basal ganglia volumes following CO poisoning: a prospective longitudinal study, Undersea Hyperbar. Med., 33, 245, 2006.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 371 — #31

372

Carbon Monoxide Poisoning

183. Sawada, Y., Ohashi, N., Maemura, K., et al. Computerized tomography as an indication of long-term outcome after carbon monoxide poisoning, Lancet, 1, 783, 1980. 184. Pracyk, J.B., Stolp, B.W., Fife, C.E., et al. Brain computerized tomography after hyperbaric oxygen therapy for carbon monoxide poisoning, Undersea Hyperbar. Med., 22, 1, 1995. 185. Lee, M.S., and Marsden, C.D. Neurological sequelae following carbon monoxide poisoning clinical course and outcome according to the clinical types and brain computed tomography, Movement Disord., 9, 550, 1994. 186. Hopkins, R.O., Fearing, M., Weaver, L., and Foley, J. Basal ganglia lesions following carbon monoxide poisoning, Brain Injury, 20, 273, 2006. 187. Ho, V.B., Fitz, C.R., Chuang, S.H., et al. Bilateral basal ganglia lesions, Pediatric differential considerations, Radiographics, 13, 269, 1993. 188. Parkinson, R.B., Hopkins, R.O., Cleavinger, H.B., Weaver, L.K., Victoroff, J., Foley, J.F., et al. White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning, Neurology, 58: 1525, 2002. 189. Horowitz, A.L., Kaplan, R., and Sarpel, G. Carbon monoxide toxicity, MR imaging in the brain, Radiology, 162, 787, 1987. 190. Taverni, N., Dal Pozzo, G., Bartolozzi, C., Caramelli, L., and Boddi, P. Magnetic resonance imaging in the study of brain changes due to carbon monoxide poisoning, Radiol. Med., 76, 289–292, 1988. 191. Silverman, C.S., Brenner, J., and Murtagh, F.R. Hemorrhagic necrosis and vascular injury in carbon monoxide poisoning , MR demonstration, AJNR, 14, 168, 1993. 192. Uchino, A., Hasuo, K., Shida, K., Matsumoto, S., Yasumori, K., and Masuda, K. MRI of the brain in the chronic phase of carbon monoxide poisoning, Neuroradiology, 36, 399, 1994. 193. Mascalchi, M., Petruzzi, P., and Zampa, V. MRI of cerebellar white matter damage due to carbon monoxide poisoning, case report, Neuroradiology, 38, S73, 1996. 194. Kawanami, T., Kato, T., Kurita, K., and Sasaki, H. The pallidoreticular pattern of brain damage on MRI in a patient with carbon monoxide poisoning, J. Neurol. Nuerosur. Ps., 64, 282, 1998. 195. Tuchman, R.F., Moser, F.G., and Moshe, S.L. Carbon monoxide poisoning: bilateral lesions in the thalamus on MR imaging of the brain, Pediatr. Radiol., 20, 478, 1990. 196. Gale, S.D., Hopkins, R.O., Weaver, L.K., Bigler, E.D., Booth, E.J., and Blatter, D.D. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning, Brain Injury, 13, 229, 1999. 197. Hopkins, R.O., Weaver, L.K., and Kesner, R.P. Long-term memory impairments and hippocampal magnetic resonance imaging in carbon monoxide poisoned subjects, Undersea Hyperbar. Med., 20 (Suppl.), 15 (abstract), 1993. 198. Hopkins, R.O., Weaver, L.K., and Kesner, R.P. Quantitative MRI analysis of the hippocampus corresponds with persistent memory impairments in carbon monoxide poisoned subjects, Brain Cognition, 28, 215 (abstract), 1995. 199. Prockop, L.D. Carbon monoxide brain toxicity: clinical, magnetic resonance imaging, magnetic resonance spectroscopy, and neuropsychological effects in 9 people, J. Neuroimaging, 15, 144, 2005. 200. Turner, M., and Kemp, P.M. Isotope brain scanning with Tc-HMPAO: a predictor of outcome in carbon monoxide poisoning? J. Accid. Emerg. Med., 14, 139, 1997. 201. Denays, R., Makhoul, E., Dachy, B., Tondeur, M., Noel, P., Ham, H.R., and Mols, P. Electroencephalopgrahic mapping and 99mTc HMPAO single-photon emission

8417: “8417_c016” — 2007/9/11 — 12:15 — page 372 — #32

Treatment of Carbon Monoxide Poisoning

202.

203.

204.

205.

206.

207.

208.

209.

210. 211.

212. 213. 214.

215.

216. 217. 218.

373

computed tomography in carbon monoxide poisoning, Ann. Emerg. Med., 24, 947, 1994. DeReurck, J., Decoo, D., Lemahieu, I., et al. A positron emission tomography study of patients with acute carbon monoxide poisoning treated by hyperbaric oxygen, J. Neurol., 240, 430–434, 1993. Shimosegawa, E., Hatazawa, J., Nagata, I., and Okudera, T. Cerebral blood flow and glucose metabolism measurements in a patient surviving one year after carbon monoxide intoxication, J. Nucl. Med., 33, 1696, 1992. Hon, K.E., Yeung, W., Assunta, C., Leung, W.A., Li, A.M., Chu ,W.C., and Chan, Y. Neurologic and radiologic manifestations of three girls surviving acute carbon monoxide poisoning, J. Child Neurol., 21: 737, 2006. Kao, C.H., Hung, D.Z., ChangLai, S.P., Liao, K.K., and Chieng, P.U. HMPAO brain SPECT in acute carbon monoxide poisoning, J. Nucl. Med., 39, 769–772, 1998. Castoldi, A.F., Coccini, T., Randine, G., Hernandez-Viadel, M., Felipo, V., and Manzo, L. Lymphocyte cytochrome c oxidase, cyclic GMP and cholinergic muscarinic receptors as peripheral indicators of carbon monoxide neurotoxicity after acute and repeated exposure in the rat, Life Sci., 78, 1915, 2006. Brvar, M., Mozina, M., Osredkar, J., Suput, D., and Bunc, M. Prognostic value of S100B protein in carbon monoxide-poisned rats, Crit. Care Med., 32, 2128, 2004. Hampson, N.B., Dunfored, R.G., Kramer, C.C., et al. Selection criteria utilized for hyperbaric oxygen treatment of carbon monoxide poisoning, J. Emerg. Med., 13, 227, 1995. Sloan, E.P., Murphy, D.G., Hart, G., Cooper, M.A., Turnbull, T., Berreca, R.S., and Ellerson, B. Complications and protocol considerations in carbon monoxide-poisoned patients who require hyperbaric oxygen therapy: report from a ten-year experience, Ann. Emerg. Med., 18, 629, 1989. Goulon, M., Barois, A., Rapin, M., et al. Carbon monoxide poisoning and acute anoxia due to breathing coal gas and hydrocarbons, J Hyperbaric Med., 1, 23, 1986. Penney, D.G. Effects of carbon monoxide exposure on developing animals and humans, In Carbon Monoxide, Penny, D.G., ed., Boca Raton, FL, CRC Press, Inc, 1996, Chap 6, pp. 109–144. Ginsberg, M.D., and Myers, R.E. Fetal brain injury after maternal carbon monoxide intoxication, Neurology, 26, 15–23, 1976. Ginsberg, M.D., and Myers, R.E. Fetal brain damage following maternal carbon monoxide intoxication, Acta. Obstet. Gynecol., 53, 309–317, 1974. Koren, G., Sharav, T., Pastuszak A., Garrettson, L.K., Hill, K. Samson, I., Rorem, M., King A., and Dolgin J.E. A multi-center, prospective study of fetal outcome following accidental carbon monoxide poisoning in pregnancy, Reprod. Toxiol., 5, 397–403, 1991. Brown, D.B., Mueller, G.L., and Golich, F.C. Hyperbaric oxygen treatment for carbon monoxide poisoning in pregnancy: a case report, Aviat. Space Envir. Md., 63, 1011–1014, 1992. Ferm, V.H. Teratogenic effects of hyperbaric oxygen, P. Soc. Exp. Biol. Med., 116, 975–976, 1964. Fujikura, T. Retrolental fibroplasia and prematurity in newborn rabbits induced by maternal hyperoxia, Am. J. Obstet. Gynecol., 90, 854–858, 1964. Miller, P.D., Telfored, I.D., and Haas, G.R. Effects of hyperbaric oxygen on cardiogenesis in the rat, Biol. Neonatorum, 17, 44–52, 1971.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 373 — #33

374

Carbon Monoxide Poisoning

219. Elkaharrat, D., Raphael, J.D., Korach, J.M., Jars-Guincestre, M.C., Chastang, C., Harborn, C., and Gajdos, P. Acute carbon monoxide intoxication and hyperbaric oxygen in pregnancy, Intens. Care Med., 17, 289–92, 1991. 220. Farrow, J.R., Davis, G.J., Roy, T.M., et al. Acute carbon monoxide intoxication and hyperbaric oxygen in pregnancy, Intens. Care Med., 17, 289–292, 1991. 221. White, S.R. Pediatric Carbon Monoxide Poisoning. In: Carbon Monoxide Toxicity, Penney, D.G., ed., CRC Press LLC, Boca Raton 2000, Chapt. 21. 222. Seger, D., and Welch, L. Carbon monoxide controversies: neuropsychologic testing, mechanism of toxicity, and hyperbaric oxygen, Ann. Emerg. Med., 24, 242, 1994. 223. De Pooter, M.C., Leys, D., Godefroy, O., DeReuck, J., and Peter, H. Parkinsonian syndrome caused by carbon monoxide poisoning. Preliminary results of the treatment with bromocriptine, Rev. Neurol., 147, 399–403, 1991. 224. Hu, M.C., Shiah, I.S., Yeh, C.B., Chen, H.K., and Chen, C.K. Ziprasidone in the treatment of delayed carbon monoxide encephalopathy, Prog. Neuro-psychoph. , 30, 755, 2006. 225. Pae, C.U., Kim, T.S., Lee, C., and Paik, I.H. Effect of aripiprazole for a patient with psychotic symptoms and parkinsonism associated with delayed sequelae of carbon monoxide intoxication, J. Neuropsych. Clin. Neurosci., 18, 130, 2006. 226. Ten Holter, J.B.M, and Schellens, R.L.L.A.M., Dantrolene sodium for treatment of carbon monoxide poisoning, Brit. Med. J., 296, 1772, 1988. 227. Hiramatsu, M., Sasaki, M., Nabeshima, T., and Kaemyama, T. Effects of dynorphin A(1–13) on carbon monoxide-induced delayed amnesia in mice. Pharmacol. Biochem., 56, 73, 1997. 228. Hiramatsu, M., Hyodo, T., and Kameyama T. U-50488H a selective k-opioid receptor agonist, improves carbon monoxide-induced delayed amnesia in mice, Eur. J. Pharmacol., 315, 119, 1996. 229. Myers, R.A., Snyder, S.K., and Emhoff, T.A. Subacute sequelae of carbon monoxide poisoning, Ann. Emerg. Med., 14, 1163–1167, 1985. 230. Smith, J.S., and Brandon, S. Morbidity from acute carbon monoxide poisoning at three-year follow up, Brit. Med. J., 1, 318–321, 1973. 231. Myers, R.A., Mitchell, J.T., and Cowley, R.A. Psychometric testing and carbon monoxide poisoning, Disaster Med., 1, 279–281, 1983. 232. Myers, R.A., Snyder, S.K., Linberg, S., et al. Value of hyperbaric oxygen in suspected carbon monoxide poisoning, JAMA, 246, 248, 1981. 233. Coric, V., Oren, D.A., Wolkenberg, A., and Kravitz, R.E. Carbon monoxide poisoning and treatment with hyperbaric oxygen in the subacute phase, J. Neurol. Neurosur. Ps., 65, 245, 1998. 234. Maeda, Y., Kawasaki, Y., Jibiki, I., Yamaguchi, N., Matsuda, H., and Hisada, K. Effect of therapy with oxygen under high pressure on regional cerebral blood flow in the interval form of carbon monoxide poisoning: observation from subtraction of Technetium-99m HMPAO SPECT brain imaging, Eur. Neurol., 31, 380, 1991. 235. Samuels, A.H., Vamos, M.J., and Taikato, M.R. Carbon monoxide, amnesia and hyperbaric oxygen therapy, Aust. NZ. J. Psychiat., 26, 316, 1992. 236. Heinrichs, R.W. Relationship between neuropsychological data and response to behavioral treatment in a case of carbon monoxide toxicity and dementia, Brain Cognition, 14, 213–219, 1990.

8417: “8417_c016” — 2007/9/11 — 12:15 — page 374 — #34

17

The Case for the Use of Hyperbaric Oxygen Therapy in Carbon Monoxide Poisoning Christian Tomaszewski

CONTENTS 17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2 Carbon Monoxide Effects at the Cellular Level . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.1 HBOT Reverses the Cellular Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3 Clinical Efficacy of HBOT in Carbon Monoxide Poisoning . . . . . . . . . . . . . 17.3.1 Negative Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3.2 Positive Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.4 Indications for HBO Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.5 Delayed Administration of HBOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.6 Repeated Treatment with HBOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.7 Future Directions in Better Targeting HBO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

375 376 377 378 378 379 382 383 384 384 385 385

17.1 INTRODUCTION Carbon monoxide (CO) is a serious but complex poison. If one is lucky enough to survive the acute hypoxic event from avid binding of hemoglobin, the patient still has to contend with a potential of up to 40% chance of delayed and/or persistent neurological deficits.1–6 These effects can be debilitating including dementia, amnestic syndromes, parkinsonism, movement disorders, and cortical blindness.7–9 The problem is that patients may appear initially well, and following several days to weeks, develop delayed neurological sequelae (DNS).9 These problems can last for a year or longer.10 The main issues in treating CO poisoning is identifying those that are at risk for neurological sequelae and referring these patients to a treatment modality that will prevent such sequelae. These issues are important once the patient has reached the 375

8417: “8417_c017” — 2007/9/11 — 12:15 — page 375 — #1

Carbon Monoxide Poisoning

376

hospital, having been resuscitated and stabilized. Traditionally 100% oxygen was advocated as initial first aid to enhance removal of CO from hemoglobin and reverse any concomitant hypoxia. But starting in the 1960s, especially in the United States, hyperbaric oxygen therapy (HBOT) had started to be advocated for treating all such poisonings.11 Anywhere between 450 and 2500 HBO treatments are done for CO poisoning each year in the US alone.12,13 According to recent review articles, the weight of evidence does not appear to support the use of HBO for CO poisoning.14–17 But most of these critical review articles give equal weighting to all controlled studies. Almost all the studies failing to show benefit from HBO in CO poisoning have critical flaws in treatment regimen or follow-up. They also contradict consistent studies showing benefit in animal models, in which the pathophysiology for improvement is now being elucidated. All these studies, along with some recent randomized trails, suggest that a safe modality like HBO has the potential to offer hope in a poisoning with often unforeseen devastating neurological consequences.

17.2 CARBON MONOXIDE EFFECTS AT THE CELLULAR LEVEL The simplest explanation for the utility of HBO is that it accelerates the removal of CO from hemoglobin. Normally, the half-life of carboxyhemoglobin (COHb) with 100% oxygen averages 74–92 min, on the basis of three studies.6,18,19 With HBO, COHb half-lives are reduced to as low as 20 min.20,21 Should someone be hypoxic, HBO also increases dissolved oxygen by about ten times, sufficient to supply metabolic needs.22 But on entry into the HBO chamber, in Weaver’s trial of ill CO-poisoned patients, the mean COHb was less than 5%.6 So in most cases of CO poisoning, COHb clearance is not an issue because of aggressive treatment with normal pressure oxygen prior to ability to arrange HBO treatment. The consensus is that there must be an alternative mechanism for HBOT efficacy in CO poisoning. This is why the percentage of COHb does not always correlate with degree of acute toxicity and with eventual neurological outcome.2,10,23,24 Once a patient is stabilized, the real target organ of interest is the brain. CO is delivered intracellularly, where it can bind to heme proteins other than hemoglobin. At high levels, CO interferes with cellular respiration by binding to and inhibiting mitochondrial cytochrome oxidase.25 This is particularly exaggerated during episodes of hypoxia and hypotension. The inhibition of cellular respiration may contribute to the ischemic reperfusion that occurs in rat brains during CO poisoning ultimately leading to lipid peroxidation.26 These free radicals cause endothelial damage, which allows lymphocytes to attach and release proteases that promote more production of oxygen free radicals.27 The end result of this process is delayed central nervous system (CNS) damage that is accompanied by learning decrements. Part of the neuronal damage may also be mediated by excitatory amino acids.28,29 Glutamate binds to N-methyl-d-aspartate (NMDA) receptors causing intracellular calcium release, thus causing delayed neuronal cell loss. Accompanying this may

8417: “8417_c017” — 2007/9/11 — 12:15 — page 376 — #2

The Case for the Use of Hyperbaric Oxygen Therapy

377

be apoptosis.30,31 Some of the most sensitive areas for neuronal cell loss from these processes are the basal ganglia and hippocampus, resulting in impaired learning and memory.32 All these rodent studies demonstrate that CO is not a simple chemical asphyxiant. The CO molecule sets in a motion a cascade of cytochemical events that days later result in CNS cell loss. So in spite of COHb clearance from the bloodstream early on with normal pressure, or normobaric oxygen (NBO), HBO may still have an essential role in most cases of CO poisoning in preventing these delayed neurological events.

17.2.1 HBOT REVERSES THE CELLULAR EFFECTS HBO has been shown to reverse many of the biochemical effects of CO in the same rodent models that elucidated CO’s neurochemical and immune effects. First, HBO accelerates regeneration of inactivated cytochrome oxidase.25 This reduction in potential for oxidative stress is accompanied by prevention of lipid peroxidation. Rat models show a dose response effect from HBO, maximal at 3 atm. absolute (ATA), in decreasing the lipid peroxidation products in the brains of rats exposed to CO.26 The critical event that precedes lipid peroxidation is the vascular abnormalities that CO induces in cerebral endothelium. Adherence of neutrophils amplifies CO-mediated oxidative stress.27,33 The end result of this is necrosis and apoptosis of critical areas of the brain for learning and memory, particularly the hippocampus.28,30,34 HBOT is able to prevent neutrophil adherence to the brain microvascular endothelium, an essential step for amplification of CNS damage from CO.35 This blockage of lipid peroxidation prevents the precipitation of abnormalities in myelin basic proteins. Therefore, the CO-mediated oxidative stress that causes alteration in myelin basic protein and leads to immunologic effects is blocked by HBO.36 The final outcome is that HBO reduces mortality in rodent models of serious CO poisoning. This appears to be due to protection against cerebral edema. All control rats died in one study from cerebellar herniation as opposed to 100% survival in those that received HBO.38 The end result is that cognitive decrements from CO are prevented by HBO, as demonstrated in radial maze performance.36 The histological manifestation of this protection is lack of lesions in the globus pallidus and hippocampus in rats, which are seen after CO poisoning.32 One recent study does not support this, but there were issues with excessive toxicity from concomitant hypoxic stress.39 In conclusion, animal models show that the mechanism for CO toxicity is more than hypoxia. The poisoning initiates an immune cascade in the brain that includes changes in vascular endothelium, adherence of leukocytes, and lipid peroxidation. The end result of these events is neurological deficits in learning and memory. HBO, but not NBO, prevents all of these events, usually in a dose-related fashion. None of these mechanisms are related to the continued presence or clearance of COHb. This lends support to the use of HBO in symptomatic CO-poisoned patients once they have been stabilized regardless of the lack of remaining COHb. These patients are still at risk for delayed or persistent neuropsychological sequelae and therefore could benefit from HBO.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 377 — #3

Carbon Monoxide Poisoning

378

17.3 CLINICAL EFFICACY OF HBOT IN CARBON MONOXIDE POISONING HBOT has been advocated as the treatment of choice for patients with significant CO exposures.11,40 However, clinical studies have not consistently demonstrated efficacy for HBO in preventing neurological damage from CO, as basic science studies discussed above suggest. Initially, interest in HBO for CO poisoning was fueled by uncontrolled human clinical series. In such studies, the incidence of persistent neuropsychiatric symptoms, including memory impairment, ranged from 12% to 43% in patients treated with 100% oxygen, and was as low as 0–4% in patients treated with HBO.4,5,40–42 Such early studies were obviously biased. So although HBO was advocated widely for CO poisoning, no randomized controlled trials (RCTs) existed to support such a stance.

17.3.1 NEGATIVE TRIALS The first truly randomized study in acutely poisoned CO patients did not come until 1989.2 The study failed to show the beneficial effect of HBO in over 300 patients. Patients who presented without loss of consciousness (LOC), n = 343, were randomized to one HBO session or 100% oxygen at room pressure by mask. Outcome criteria were not strict and were mainly based on questionnaire responses regarding symptoms. At 1 month, there was no difference in both groups with respect to self-reported neurological symptoms: 32% in the HBO group, 34% in the room pressure oxygen group. Another outcome, return to prior occupation, was 97% in both groups. Major flaws in this study, besides the lack of blinding and poor endpoints, were suboptimal HBO pressure used, 2.0 ATA, and almost half of the patients receiving HBO over 6 h from end of CO poisoning, which was a major delay.43 After the Raphael trial described above2 , a larger, double-blinded trial was hailed by many as conclusive proof that HBO was not useful in CO poisoning.44 This trial involved 191 patients who up to 24 h after poisoning were randomized to HBO (2.8 ATA maximum) versus sham HBO treatments. Patients were treated aggressively at a maximal pressure of 2.8 ATA. They received daily treatments for 3 days, and up to 6 additional daily sessions if they did not recover. HBO provided no benefit in this trial with the majority of each group having adverse neurological outcomes at 1 month: 74% in HBO-treated patients, and 68% in controls (reported odds ratio (OR) = 1.7; 95% confidence interval (CI) = 0.8–4.0; P = 0.19, not significant (NS)). The major flaw in this study was a mean delay of over 6 h to treatment with HBO. In addition 54% of the original subjects were lost to follow-up. Disproportionate numbers of suicide cases (about two-thirds) and drug toxicity (44%), with accompanying excessive neuropsychological defects, could have confounded any beneficial effect from HBOT. Other work shows that depression can contribute to poor neuropsychological testing.45 Because of these multiple flaws, it is not surprising that HBO failed to show any benefit in what was initially a well designed study. The most recent study showing no beneficial effect from HBO in CO poisoning is only available in abstract form.3 One hundred and seventy-nine patients that had transient LOC, but who were not comatose, were randomized to one session of HBO

8417: “8417_c017” — 2007/9/11 — 12:15 — page 378 — #4

The Case for the Use of Hyperbaric Oxygen Therapy

379

at 2 ATA versus 6 h of normal pressure oxygen. Outcome was self-assessed symptoms and blinded neurological examination at 1 month. Recovery was the same in both groups: HBO 46/79 (58%) versus control 45/74 (61%). As in Raphael’s previous study2 , lack of sufficient pressure, at 2.0 ATA, may have undermined the efficacy of HBO. There are consistent themes in the negative studies failing to show beneficial effects of HBO in CO. The first is delay in treatment. Animal studies show that the biochemical effects begin within hours after exposure to CO.36 The issue is whether delayed treatment can reverse this once it has started.43 The other problem is that treatment may have not been optimal. Three ATA is the pressure at which animal studies show the most benefits on tissue effects and neurological outcomes after CO poisonings.25,46,47 In fact, to prevent human neutrophil adherence, only marginal inhibition occurs at 2.0 ATA; 2.8 ATA gives almost complete inhibition.46 Other issues with these studies include poor outcome data due to both quality and quantity of follow-up.

17.3.2 POSITIVE TRIALS As the critics of HBO mounted, studies began to show the beneficial effects of HBO with CO poisoning. The first randomized positive study used mildly poisoned patients.1 Sixty-five patients without LOC were randomized to one session of HBO or 100% oxygen by face mask at room pressure. A decrease in any one of six neuropsychological tests immediately after poisoning versus at 4 weeks defined outcome. No patient [95% C.I. 0–12%] in the HBO group versus 23 [95% C.I. 10–42%] in the control group showed deterioration. Critics point out that the neurological deterioration was seen in only one test, Trail-Making; but neurological sequelae persisted for a mean of 41 days, and 10% of the controls had difficulties with daily activities. The success of this trial can be partially attributed to the adequate pressure used, maximum 2.8 ATA, and that all patients were treated within 6 h of discovery. A subsequent study by Mathieu et al.48 examined the efficacy of HBO at 2.5 ATA for 90 min versus NBO for 12 h. All patients had to be within 12 h of poisoning termination and noncomatose on presentation. At one and three months there was a lower incidence of persistent neurologic manifestations in the HBO group. This was significant at 3 months: HBO 9.5% versus NBO 15%, p < .02. This difference resolved over the following year. This study suffers from lack of detail and the fact that it is an interim analysis with no final analysis published. Regardless, early aggressive treatment, using pressures above 2.0 ATA, confirm the efficacy of HBO in CO poisoning. The landmark randomized control trial showing efficacy of HBO in CO poisoning differs from the other positive trials, in that all patients, including intubated cases (13%), were enrolled. Patients, n = 152, who presented with symptoms from acute CO poisoning were randomized to three sessions of HBO, with the first hour at 3.0 ATA, versus 100% oxygen at room pressure.6 Rather than looking at the absolute difference between the two groups on neurological testing, neurological sequelae were defined a priori. Aggregate performance on six neuropsychological tests at least one standard deviation below published norms, or an aggregate score greater than

8417: “8417_c017” — 2007/9/11 — 12:15 — page 379 — #5

380

Carbon Monoxide Poisoning

or equal to two standard deviations below expected, was defined as neurological sequelae. At 6 weeks, HBO caused an absolute reduction of 21% [95% C.I. 6–34%] in the proportion of neurological sequelae: 24% in the HBO group; 46% in control, which equates to a number needed to treat 5. At 1 year, there was still some benefit, albeit less, with an absolute reduction of 15% [95% C.I. 1–28%]. As stated earlier, the patients were seriously ill on presentation, with a mean initial COHb level of 25% and half having suffered LOC. Although patients could be entered up to 24 h post CO-poisoning, the mean time to treatment was less than 2 h. The combination of high pressure and early treatment probably contributed to the success of this trial. The main criticisms with Weaver’s study6 lie in the statistical definition of neurological sequelae.14 Although there were individual differences in the neuropsychological test of Trail-Making, there was no difference at 6 weeks in the parametric comparison of mean neuropsychological score. In addition, there was no difference in activities of daily living at 6 weeks and 12 months. In defense of the study and HBO, patients had decreased self-reported memory problems at 6 weeks (28% versus 51%), and the beneficial effect on cognitive sequelae lasted up to one year. Another criticism leveled on the Weaver study was that patients in the control group were more ill, with a higher incidence of cerebellar dysfunction on presentation, which ended up being a predictor for cognitive sequelae.17 But, when doing a logistic regression to adjust for this incidental finding, HBO was still protective (OR=0.45, 95% C.I. 0.22–0.92).6 In conclusion, Weaver’s study appears methodologically sound, especially when considering that outcome measures were determined a priori and the fact that the trial was double-blinded including sham HBO treatments in the control group. The Cochrane Review recently examined all six controlled studies published as of 2006 that examined the effect of HBO versus NBO in CO poisoning.15 (Table 17.1). The common outcome was the presence of neurological symptoms or signs at time of primary analysis 4–6 weeks postpoisoning. With a collective 1335 participants, the incidence of persistent signs and symptoms was 29% (202/691) in the HBOT group versus 34% (219/444) in the normobaric group. The overall odds ratio favored HBO at 0.78 [95% C.I. 0.54–1.12]. Because of the lack of statistical significance their conclusion was that existing trials do not support use of HBO in CO-poisoned patients in order to reduce neurological sequelae. Similarly, the American College of Emergency Physicians, on the basis of the same data, formulated a similar conclusion in their 2007 Clinical Policy for the management of acute CO poisoning.49 They concluded that although HBO should not be mandated for CO poisoning, it “remains a therapeutic option to potentially reduce the incidence of neurological sequelae.” The consensus among all these expert reviews is that further RCTs are needed, particularly with regard to identifying which CO-poisoned patients are most likely to benefit from HBO. Using the same studies, the Underwater and Hyperbaric Medical Society recommends HBO treatment for any CO-poisoned patients with signs of serious intoxication.11 With little risk,50 almost 1500 patients are treated with HBO for CO poisoning in the United States each year.51 The main risk with HBO is related to barotraumas such as tympanic perforation.6 Less likely, and not usually associated with

8417: “8417_c017” — 2007/9/11 — 12:15 — page 380 — #6

HBO 2 h vs. 100% NBO until asymptomatic HBO 1 h (min × 3) vs. NBO 100 min HBO 2 h (×3) vs. NBO 2 h (×1)

Thom 19951

2.0 ATA

3.0 ATA

2.8 ATA

2.8 ATA

2.0 ATA

2.5 ATA

Maximum HBO Pressure

∗ Odds ratio less than 1.00 favors treatment with HBO.

Raphael 20043

Weaver 20026

HBO at 2.0 ATA 60 min vs. 6 h NBO

HBO 2.0 h vs. 6 h NBO

Raphael 19892

Scheinkestel 199944

HBO 90 min vs. 12 h NBO

Design

Mathieu 199648

Study

< 12 h

Mean 5.6 h

Mean 7.1 h

Mean 2.0 h

Mean 7.1 h

25% (Adapted from Tomaszewski, C., Goldfrank’s Toxicologic Emergencies, Goldfrank, L.R., Flomenbaum, N.E., Hoffman, R.S., Howland, M.A., Lewin. N,A., Nelson, L.S (Eds), McGraw-Hill, New York, 2006.)

8417: “8417_c017” — 2007/9/11 — 12:15 — page 382 — #8

The Case for the Use of Hyperbaric Oxygen Therapy

383

injury in animal models.60,61 However, syncope is not entirely predictive for cognitive sequelae.6 Patients with long exposures, or “soaking” periods, are also at greater risk for neurological sequelae.62,63 Animal models and human cases suggest that “soaking,” that is, prolonged exposure to high levels of CO, is a factor that actually predicts final neurological outcome, rather than any particular COHb level.64,65 The presence of a significant metabolic acidosis may be a surrogate marker for this. A base excess lower that −2 mmol/L was an independent predictor for cognitive sequelae and the potential for benefit from HBO in Weaver’s study.6,58 Other studies confirm the utility of metabolic acidosis as a predictor of HBO requirement.10,66,67 It is still unclear if mild neurological symptoms (e.g., confusion, headache, dizziness, visual blurring) or abnormal mental status testing on initial presentation after CO poisoning is prognostic for cognitive sequelae. These symptoms simply represent CO poisoning, which, at COHb levels approaching 10% in volunteers, can cause temporary impairment of learning and memory.68 To date, neuropsychological screening tests have not been found to be reliable indicators of the need for HBO because they do not consistently predict neurological sequelae.1 In a recent prospective clinical trail of CO poisoning, the incidence of cerebellar dysfunction portended a higher incidence of cognitive sequelae (odds ratio 5.7 [95% C.I. 1.7–19.3]).6 Therefore, difficulty with finger-to-nose, heel-to-shin, rapid alternating hand movements, or even ataxia, should all be considered indications for HBO. Some authors recommend selective use of HBO because of cost and difficulties in transport if the primary facility lacks a chamber.69 However, complications that may make such transfers and treatment unsafe are rare.50 Although HBO cannot be recommended for every patient with CO poisoning, it is a relatively safe treatment that should be considered in all serious exposures. Post hoc analysis of Weaver’s data showed that the patients most likely to benefit were those who presented with a base deficit greater than 2 mmol/L, unconsciousness, age ≥50 years, and COHb level greater than 25%.6,70 Therefore, most clinicians refer any CO-poisoned patient with any of these criteria for HBO (Table 17.2). One group of patients who could probably be excluded from HBOT are those who have had a cardiac arrest from CO; these cases have been universally fatal.71

17.5 DELAYED ADMINISTRATION OF HBOT The optimal timing of HBO treatment for CO poisoning is unclear. Patients treated later than 6 h after exposure tend to have worse outcomes in terms of delayed sequelae (30% versus 19%) and mortality (30% versus 14%).72 Randomized trials with longer times to treatment have generally failed to show benefit from HBO.73 In the recent randomized trial showing beneficial effects from HBO, with a number needed to treat of 5 in order to prevent one case of neurological sequelae, patients were entered into treatment up to 24 h post-poisoning.6 In fact 38% of the patients were treated later than 6 h. Studies purporting beneficial effects much later are anecdotal and lack controls.4 Therefore, at this time it seems reasonable to attempt treatment up to 24 h post-exposure.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 383 — #9

Carbon Monoxide Poisoning

384

17.6 REPEATED TREATMENT WITH HBOT Repeated HBO treatments have been advocated for patients that do not initially improve from CO poisoning, particularly those in coma.74 In the recent randomized study showing beneficial effects from HBO, all patients received three HBO treatments within 24 h of presentation.6 A retrospective review of records on patients who received two versus one treatment showed a reduction in DNS from 555 to 18%.75 Prospective studies comparing single versus multiple courses of HBOT have failed to confirm any benefit from repeated HBO treatment.2,3 Therefore, multiple HBO treatments cannot be recommended at this time. The most recent clinical guidelines from the Underwater and Hyperbaric Society state that the optimal number of HBO treatments for CO poisoning is unknown, reserving multiple treatments for patients who fail to recover after the initial treatment.11

17.7 FUTURE DIRECTIONS IN BETTER TARGETING HBO A search has been made for plasma markers for CO exposure that predict degree of toxicity and therefore, perhaps, the potential for neurological sequelae and the need for HBO. Many plasma markers of oxidative stress, such as glutathione release from erythrocytes, increase within hours of CO poisoning.76 Multiple peripheral vascular cells have been implicated in CO poisoning: erythrocytes, platelets, leukocytes, and endothelial cells.27,46,77 They actually appear to mirror what is going on in the target organ of concern, the CNS.78,79 One recent study showed that CO exposure in rats resulted in a 50% decrease in cyclic guanosine monophosphate (GMP) activity, as measured in leukocytes, that started at 24 h post-exposure.80 No human series of plasma markers are available to show if they are predictive of final outcome and therefore a potential need for HBOT. Plasma markers are still only surrogates for the real target organ of CO poisoning, the brain. S100B is a neurobiochemical marker of brain damage. It is found in astroglial cells and is released into the blood with brain injury. Rat studies show that it is elevated after severe CO poisoning and is a better predictor of death than level of consciousness.81 In patients presenting with Glasgow Coma Scores (GCS) of less than eight after CO poisoning, but who did not experience LOC, immediate blood samples show elevated peripheral blood levels of SB100B.82 HBO prevents this rise in SB100B when given immediately after CO poisoning in rats.83 A recent clinical study, however, showed no increase in S100B levels after CO poisoning.84 Therefore, it is unclear if SB100B can be used as a consistent marker of CO poisoning and the need for HBO. As an alternative to serum markers, various types of neurological imaging are available that show early changes from CO poisoning. Except for xenon enhancement, computed tomography does not show early changes from CO.85 More sensitive is Magnetic Resonance Imaging (MRI), which can show changes within the first day post-exposure.86 Diffusion-weighted is even more sensitive, detecting changes in subcortical white matter within hours of CO poisoning.87 However, MRI changes have not been shown to correlate with eventual outcome.57,87

8417: “8417_c017” — 2007/9/11 — 12:15 — page 384 — #10

The Case for the Use of Hyperbaric Oxygen Therapy

385

A more sensitive radio imaging modality is single-photon emission computed tomography (SPECT), which gauges regional blood flow noninvasively. It shows perfusion changes within watershed regions that are associated with delayed neuropsychological impairment months after CO poisoning.88,89 With HBO, an improvement in oxygen extraction of the frontal and temporal cortices was demonstrated on positron emission tomography (PET) scan.90 Because of practicality considerations, along with lack of prospective data, no neurological imaging technique can be advocated at this time to predict who needs HBO treatment.

17.8 SUMMARY HBOT may originally have been a “therapy in search of a disease” with respect to CO poisoning. But now several controlled studies show early benefits in preventing cognitive sequelae from CO poisoning. Although these studies are not perfect, the alternative negative clinical controlled studies suffer from serious flaws. In addition, the positive clinical results corroborate the findings of animal studies that show that HBO can prevent the cascade of neurochemical events that occur during recovery from acute poisoning. The biggest challenge though will be to decide who can really benefit from this therapy. To date, there is no consistent marker to predict who will suffer DNS and therefore, who has the most to gain. Until more studies are available confirming HBO’s utility, and owing to the inherent delay with such, we probably should not be subjecting patients to extraordinary transports to receive such therapy. But based on the safety of the procedure, and relative low resource utilization, there is no reason not to at least attempt one HBO treatment in any seriously ill CO-poisoned patients.

References 1. Thom, S.R., and Taber, R.L., Mendiguren II et al. Delayed neuropsychologic sequelae after carbon monoxide poisoning: prevention by treatment with hyperbaric oxygen. Ann. Emerg. Med. 25, 474–480, 1995. 2. Raphael, J.C., Elkharrat, D., Jars-Guincestre, M-C et al. Trial of normobaric and hyperbaric oxygen for acute carbon monoxide intoxication. Lancet 2, 414–419, 1989. 3. Raphael, J.C., Chevret, S., Driheme, A., and Annane, D. Managing carbon monoxide poisoning with hyperbaric oxygen. J. Toxicol. Clin. Toxicol. 42, 455–456. 2004. (Abstract) 4. Myers, R.A.M., Snyder, S.K., and Emhoff, T.A. Subacute sequelae of carbon monoxide poisoning. Ann. Emerg. Med. 14, 1167, 1985. 5. Mathieu, D., Nolf, M., and Durocher, A. Acute carbon monoxide poisoning: risk of late sequelae and treatment by hyperbaric oxygen. J. Toxicol. Clin. Toxicol. 23, 315–324, 1985. 6. Weaver, L.K., Hopkins, R.O., Chan, K.J et al. Hyperbaric oxygen for acute carbon monoxide poisoning. NEJM 347, 1057–1067, 2002. 7. Lee, M.S., and Marsden, C.D. Neurological sequelae following carbon monoxide poisoning: clinical course and outcome according to the clinical types and brain computed tomography scan findings. Movement Disorders 9, 550–558, 1996.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 385 — #11

386

Carbon Monoxide Poisoning 8. Ginsberg, M.D. Carbon monoxide intoxication: clinical features, neuropathology, and mechanisms of injury. J. Toxicol. - Clin. Toxicol. 23, 281–288, 1985. 9. Choi, I.S. Delayed Neurological Sequelae in Carbon Monoxide Intoxication. Arch. Neurol. 40, 433–435, 1983. 10. Sokal, J.A., and Kraldowska, E. The relationship between exposure duration, carboxyhemoglobin, blood glucose, pyruvate and lactate and the severity of intoxication in 39 cases of acute carbon monoxide poisoning in man. Arch. Toxicol. 57, 196–199, 1985. 11. Thom, S.R., and Weaver, L.K. Carbon monoxide poisoning, Hyperbaric Oxygen 2003 Indications and Results: The Hyperbaric Oxygen Therapy Committee Report, Feldmeier, J.J., Ed, Chap 2. Dunkirk, M.D., Undersea Hyperb. Med. Soc., 2003, pp. 11–17. 12. Hampson, N., Dunford, R.G, Kramer, C.C et al. Selection criteria utilized for hyperbaric oxygen treatment of carbon monoxide poisoning. J. Emerg. Med. 13, 227–231, 1995. 13. Watson, W.A., Litovitz, T.L., Rodgers, G.C et al. 2004 Annual Report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am. J. Emerg. Med. 23, 589–666, 2005. 14. Buckley, N.A., Isbister, G.K., Stokes, B et al. Hyperbaric oxygen for carbon monoxide poisoning: a systematic review and critical analysis of the evidence. Toxicol. Rev. 24, 75–92, 2005. 15. Juurlink, D.N., Buckley, N.A., Stanbrook, M.B., Isbister, G.K., Bennett, M., and McGuigan, M.A. Hyperbaric oxygen for carbon monoxide poisoning. Cochrane Database of Systemic Reviews 4(2), CD002041, 2006. (Abstract) 16. Silver, S., Smith, C., and Worster, A. Should hyperbaric oxygen be used for carbon monoxide poisoning? Can. J. Emerg. Med. 8, 43–46, 2006. 17. Judge, B.S., and Brown, M.D. To dive or not to dive? Use of hyperbaric oxygen therapy to prevent neurological sequelae in patients acutely poisoned with carbon monoxide. Ann. Emerg. Med. 46, 462–464, 2005. 18. Levasseur, L., Galliot-Guilley, M., Richter, F et al. Effects of mode of inhalation of carbon monoxide and of normobaric oxygen administration on carbon monoxide elimination from the blood. Hum. Exp. Toxicol. 15, 898–903, 1996. 19. Weaver, L.K., Howe, S., Hopkins, R et al. Carboxyhemoglobin half-life in carbon monoxide-poisoned patients treated with 100% oxygen at atmospheric pressure. Chest 117, 801–808, 2000. 20. Myers, R.A., Jones, D.W., and Britten, J.S. Carbon monoxide half-life study, In: Proceedings of the 9th International Congress on Hyperbaric Medicine, Kindwall, E.P., Ed, Flagstaff, AZ, Best Publishing, 1987, pp. 263–266. 21. Sasaki, T. On half-clearance of carbon monoxide hemoglobin in blood during hyperbaric oxygen therapy. Bull. Tokyo Med. Dent. Univ. 22, 63–77, 1975. 22. Boerema, I., Meyne, I., Brummelkamp, W.H et al. Life without blood. Arch. Chir. Neer. 11, 70–83, 1959. 23. Seger, D., and Welch, L. Carbon monoxide controversies: neuropsychologic testing, mechanism of toxicity, and hyperbaric oxygen. Ann. Emerg. Med. 24, 242–248, 1994. 24. Myers, R.A.M., and Britten, J.S. Are arterial blood gases of value in treatment decisions for carbon monoxide poisoning? Crit. Care. Med. 17, 139–142, 1989. 25. Brown, S.D., and Piantodosi, C.A. Recovery of energy metabolism in rat brain after carbon monoxide hypoxia. J. Clin. Invest. 89, 666–672, 1991. 26. Thom, S.R. Antagonism of carbon monoxide-mediated brain lipid peroxidation by hyperbaric oxygen. Toxicol. Appl. Pharmacol. 105, 340–344, 1990.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 386 — #12

The Case for the Use of Hyperbaric Oxygen Therapy

387

27. Thom, S.R. Leukocytes in carbon monoxide-mediated brain oxidative injury. Toxicol. Appl. Pharmacol. 123, 234–247, 1993. 28. Ishimaru, H., Katoh, A., Suzuki, H et al. Effects of N-methyl-D-aspartate receptor antagonists on carbon monoxide-induced brain damage in mice. J. Pharmacol. Exp. Ther. 261, 349–352, 1992. 29. Thom, S.R., Fisher, D., Zhang, J et al. Neuronal nitric oxide synthase and N-methyl-Daspartate neurons in experimental carbon monoxide poisoning. Tox. Appl. Pharmacol. 194, 280–295, 2004. 30. Piantadosi, C.A, Zhang, J., Levin, E.D et al. Apoptosis and delayed neuronal damage after carbon monoxide poisoning in the rat. Exper. Neurology 147, 103–114, 1997. 31. Thom, S.R., Fisher, D., Xu, Y.A et al. Adaptive responses and apoptosis in endothelial cells exposed to carbon monoxide. Proc. Natl. Acad. Sci., USA 97, 1305–1310, 2000. 32. Thom, S.R. Learning dysfunction and metabolic defects in globus pallidus and hippocampus after CO poisoning in a rat model. Undersea Hyperb. Med. 23 (Suppl.), 20. 1997. (Abstract) 33. Ischiropoulos, H., Beers, M.F., Ohnishi, S.T et al. Nitric oxide and perivascular tyrosine nitration following carbon monoxide poisoning in the rat. J. Clin. Invest. 97, 2260–2267, 1997. 34. Nabeshima, T., Katoh, A., Ishimaru, H et al. Carbon monoxide-induces delayed amnesia, delayed neuronal Death and change in acetylcholine concentrations in mice. J. Pharmacol. Exp. Ther. 256, 378–384, 1991. 35. Thom, S.R. Functional inhibition of leukocyte B2 integrins by hyperbaric oxygen in carbon monoxide-mediated brain injury in rats. Toxicol. Appl. Pharmacol. 123, 248–256, 1993. 36. Thom, S.R., Bhopale, V.M., and Fisher, D. Hyperbaric oxygen reduces delayed immune-mediated neuropathology in experimental carbon monoxide toxicity. Tox. Appl. Pharmacol. 213, 152–159, 2006. 37. Jiang, J., and Tyssebotn, I. Normobaric and hyperbaric treatment of acute carbon monoxide poisoning in rats. Undersea Hyperb. Med. 24, 107–116, 1997. 38. Jiang, J., and Tyssebotn, I. Cerebrospinal fluid pressure changes after acute carbon monoxide poisoning and therapeutic effects of normobaric and hyperbaric oxygen in conscious rats. Undersea Hyperb. Med. 24, 245–254, 1997. 39. Gilmer, B., Kilkenny, J., Tomaszewski, C et al. Hyperbaric oxygen does not prevent neurologic sequelae after carbon monoxide poisoning. Acad. Emerg. Med. 9, 1–8, 2002. 40. Norkool, D.M., and Kirkpatrick, J.N. Treatment of acute carbon monoxide poisoning with hyperbaric oxygen: a review of 115 cases. Ann. Emerg. Med. 14, 1168–1171, 1985. 41. Goulon, M., Barios, A., Raphin, M et al. Carbon monoxide poisoning and acute anoxia due to breathing coal gas and hydrocarbons. Ann. Med. Intern. 120, 335–349, 1969. 42. Smith, G.I., and Sharp, G.R. Treatment of carbon monoxide poisoning with oxygen under pressure. Lancet 2, 905–906, 1960. 43. Brown, S.D., and Piantadosi, C.A. Hyperbaric oxygen for carbon monoxide poisoning. Lancet 2, 1032, 1989. 44. Scheinkestel, C.D., Bailey, M., Myles, P.S et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial. Med. J. Aust. 170, 203–210, 1999.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 387 — #13

388

Carbon Monoxide Poisoning 45. Schiltz, K.L. Failure to assess motivation, need to consider psychiatric variables, and absence of comprehensive examination: a skeptical review of neuropsychologic assessment in carbon monoxide research. Undersea Hyperb. Med. 27, 48–50, 2000. 46. Thom, S.R., Mendiguren, I., Hardy, K.R et al. Inhibition of human neutrohil B2 integrin-dependent adherence by hyperbaric oxygen. Am. J. Physiol. (Cell Physiol) 272, C770–C777, 1997. 47. Thom, S.R. Carbon monoxide-mediated brain lipid peroxidation in the rat. J. Appl. Physiol. 68, 997–1003, 1990. 48. Mathieu, D., Wattel, F., Mathieu-Nolf, M., Durak, C., Tempe, J.P., Bouachour, G., and Sainty, J.M. Randomized prospective study comparing the effect of HBO versus 12 hours of NBO in non comatose CO poisoned patients. Undersea Hyperb. Med. 23(Suppl.), 7–8. 1996. (Abstract) 49. Wolf, S.J., Lavonas. E.J., Sloan. E.P et al. Clinical Policy: critical issues in the management of adult patients presenting to the emergency department with acute symptomatic carbon monoxide poisoning. Ann. Emerg. Med. 49, in press, 2007. 50. Sloan. E.P., Murphy. D.G., Hart, R et al. Complications and protocol considerations in carbon monoxide-poisoned patients who require hyperbaric oxygen therapy: report from a ten-year experience. Ann. Emerg. Med. 18, 629–634, 1989. 51. Hampson, N.B., Little, C.E. Hyperbaric treatment of patients with carbon monoxide poisoning in the United States. Undersea Hyperb. Med. 32, 21–26, 2005 52. Hampson, N.B., Simonson, S.G., Kramer, C.C et al. Central nervous system oxygen toxicity during hyperbaric treatment of patients with carbon monoxide poisoning. Undersea Hyperb. Med. 23, 215–219, 1996. 53. Tomaszewski, C. Carbon Monoxide, In: Goldfrank’s Toxicologic Emergencies, Goldfrank, L.R., Flomenbaum, N.E., Hoffman, R.S., Howland, M.A., Lewin. N,A., Nelson, L.S (Eds): Chap 120, New York, McGraw-Hill, 2006, pp. 1689–1704. 54. Jones, J.S., Lagasse, J., Zimmerman, G. Computed tomography findings after acute carbon monoxide poisoning. Am. J. Emerg. Med. 12, 448–451, 1994. 55. Benignus, V.A., Kafer, E.R., Muller, K.E et al. Absence of symptoms with carboxyhemoglobin levels of 16–23%. Neurotoxicol. Teratol. 9, 345–348, 1987. 56. Davis, S.M., Levy, R.C. High carboxyhemoglobin level without acute or chronic findings. J. Emerg. Med. 1, 539–542, 1984. 57. Parkinson, R.B., Hopkins, R.O., Cleavinger, H.B et al. White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning. Neurology 58, 1525–1532, 2002. 58. Thom, S.R. Hyperbaric-oxygen therapy for acute carbon monoxide poisoning. NEJM 347, 1105–1106, 2002. 59. Smith, J.S., and Brandon, S. Morbidity from acute carbon monoxide poisoning at three year follow-up. BMJ 1973, 318, 1973. 60. Thom, S.R., Bhopale, V.M., Fisher, D et al. Delayed neuropathology after carbon monoxide poisoning is immune-mediated. Proc. Natl. Acad. Sci., USA 101, 13660– 13665, 2004. 61. Okeda, R., Runata, N., Takano, T et al. The pathogenesis of carbon monoxide encephalopathy in the acute phase—physiological and morphological conditions. Acta. Neuropathol. 54, 1–10, 1981. 62. Bogusz, M., Cholewa, L., Pach, J et al. A comparison of two types of acute carbon monoxide poisoning. Arch. Toxicol. 33, 141–149, 1975. 63. Wara-Wasoweki, J., Myslak, Z., Graczyk, M et al. An attempt at comparing the results of carboxyhemoglobin level in blood and gasometric determinations in capillary blood

8417: “8417_c017” — 2007/9/11 — 12:15 — page 388 — #14

The Case for the Use of Hyperbaric Oxygen Therapy

64.

65. 66.

67. 68. 69. 70.

71.

72. 73.

74.

75.

76.

77.

78.

79.

80.

389

in cases of carbon monoxide poisoning when treatment began at the place of the accident. Anaesthesia, Resusc. Inten. Therapy 4, 249, 1976. Burney, R.E., Wu, S.C., and Nemiroff, M.J. Mass carbon monoxide poisoning: clinical effects and results of treatment in 184 victims. Ann. Emerg. Med. 11, 399, 1982. Piantadosi, C.A. Diagnosis and treatment of carbon monoxide poisoning. Respir. Care. Clin. N. Am. 5, 183–202, 1999. Turner, M., Esaw, M., and Clark, R.J. Carbon monoxide poisoning treated with hyperbaric oxygen: metabolic acidosis as a predictor of treatment requirements. [see comments.]. J. Accid. Emerg. Med. 16, 96–98, 1999. Larkin, J.M., and Moylan, J.A. Treatment of carbon monoxide poisoning: prognostic factors. J. Trauma 16, 111–114, 1976. Amitai, Y., Zlotogorski, Z., Golan-Katzav, V et al. Neuropsychological impairment from acute low-level exposure to carbon monoxide. Arch. Neurol. 55, 845–848, 1998. Olson, K.R., Carbon monoxide poisoning: mechanisms, presentation, and controversies in management. J. Emerg. Med. 1, 233–243, 1984. Thom, S.R., Hyperbaric Oxygen, In: Goldfrank’s Toxicologic Emergencies, Flomenbaum, N.E., Goldfrank, L.R., Hoffman, R.S., Howland, M.A., Lewin, N.A., Nelson, L.S., Eds, Chap A34. New York, McGraw-Hill, 2006, pp. 1705–1711. Hampson, N.B., and Zmaeff, J.L. Outcome of patients experiencing cardiac arrest with carbon monoxide poisoning treated with hyperbaric oxygen. Ann. Emerg. Med. 38, 36–41, 2001. Goulon, M., Barios, A., and Rapin, M. Carbon monoxide poisoning and acute anoxia due to breathing coal gas and hydrocarbons. J. Hyperbar. Med. 1, 23–41, 1986. Scheinkestel, C.D., Bailey, M., Myles, P.S et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomized controlled clinical trial. Undersea Hyperb. Med. 27, 163–164, 2000. Dean, B.S., Verdile, V.P., and Krenzelok, E.P. Coma reversal with cerebral dysfunction recovery after repetitive hyperbaric oxygen therapy for severe carbon monoxide poisoning. Am. J. Emerg. Med. 11, 616–618, 1993. Gorman, D.F., Clayton, D., Gilligan, J.E et al. A longitudinal study of 100 consecutive admissions for carbon monoxide poisoning to The Royal Adelaide Hospital. Undersea Hyperb. Med. 20, 311–316, 1992. Thom, S.R., Kang, M., Fisher, D et al. Release of glutathione from erythrocytes and other markers of oxidative stress in carbon monoxide poisoning. J. Appl. Physiol. 82, 1424–1432, 1997. Thom, S.R., Ohnishi, S.T., and Ischiropoulos, H. Nitric oxide release by platelets inhibits neutrophil B2 integrin function following acute carbon monoxide poisoning. Toxicol. Appl. Pharmacol. 128, 105–110, 1994. Costa, L.G., and Manzo, L.Biomarkers in occupational neurotoxicology, In: Occupational Neurotoxicology, Costa LG, Manzo L., Eds, Boca Raton, CRC Press, 1998, pp. 75–100. Minana, M.D., Corbalan, R., Montoliu, C et al. Chronic hyperammonemia in rats impairs activation of soluble guanylate cyclase in neurons and in lymphocytes: a putative peripheral marker for neurological alterations. BBRC 257, 405–409, 1999. Castoldi, A.F., Coccini, T., Randine, G et al. Lumphocyte cytochrome c oxidase, cyclic GMP and cholinergic muscarinic receptors as peripheral indicators of carbon monoxide neurotoxicity after acute and repeated exposure in the rat. Life Sciences 78, 1915–1924, 2006.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 389 — #15

390

Carbon Monoxide Poisoning 81. Brvar, M., Mozina, H., and Osredkar, J. Prognostic value of S100B protein in carbon monoxide-poisoned rats. Crit. Care Med. 32, 2128–2130, 2004. 82. Brvar, M., Mozina, H., Osredkar, J et al. S100B protein in carbon monoxide poisoning: a pilot study. Resuscitation 61, 357–360, 2004. 83. Brvar, M., Finderle, Z., Suput, D et al. S100B protein in conscious carbon monoxidepoisoned rats treated with normobaric or hyperbaric oxygen. Crit. Care Med. 34, 2228–2230, 2006. 84. Rasmussen, L.S., Poulsen, M.G., Christiansen, M et al. Biochemical markers for brain damage after carbon monoxide poisoning. Acta. Anaesthes. Scand. 48, 469–473, 2004. 85. Sesay, M., Bidabe, A.M., Guyot, M et al. Regional cerebral blood flow measurements with Xenon-Ct in prediction of delayed encephalopathy after carbon monoxide intoxication. Acta. Neurol. Scand. 166 (Suppl.), 22–27, 1996. 86. Kanaya, N., Imaizumi, H., Nakyama, M et al. The utility of MRI in acute stage of carbon monoxide poisoning. Intensive Care Med. 18, 371–372, 1992. 87. Kawada, N., Ochiai, N., and Kuzuhara, S. Diffusion MRI in acute carbon monoxide poisoning. Intern. Med. 43, 639–640, 2004. 88. Gale, S.D., Hopkins, R.O., Weaver, L.K et al. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning. Brain Inj. 13, 229–243, 1999. 89. Choi, I.S., Kim, S.K., Lee, S.S. Evaluation of outcome of delayed neurologic sequelae after carbon monoxide poisoning by technetium-99m hexamethylpropylene amine oxime brain single photon emission computed tomography. Eur. Neurol. 35, 137–142, 1995. 90. De Reuck, J., Decoo, D., Lemahieu, I et al. A positron emission tomography study of patients with acute carbon monoxide poisoning treated by hyperbaric oxygen. J. Neurology 240, 430–434, 1993.

8417: “8417_c017” — 2007/9/11 — 12:15 — page 390 — #16

18

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning: Useful Therapy or Unfulfilled Promise? Carlos D. Scheinkestel and Ian L. Millar

CONTENTS 18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2 Hyperbaric Oxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.1 Hyperbaric Oxygen Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.2 Basic Science Overview in Use of Hyperbaric Oxygen . . . . . . . . . . 18.2.3 HBO for Carbon Monoxide Poisoning: Human Evidence. . . . . . . . 18.2.3.1 “Trial of Normobaric and Hyperbaric Oxygen for Acute Carbon Monoxide Intoxication” by Raphael . . . . 18.2.3.2 “Non-Comatose Patients with Acute Carbon Monoxide Poisoning: Hyperbaric or Normobaric Oxygenation?” By Ducasse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.3.3 “Delayed Neuropsychologic Sequelae after Carbon Monoxide Poisoning: Prevention by Treatment With Hyperbaric Oxygen” by Thom . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.3.4 “Randomized Prospective Study Comparing the Effect of HBO Versus 12 h NBO in Noncomatose CO Poisoned Patients: Results of the Interim Analysis” by Mathieu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.3.5 “Managing Carbon Monoxide Poisoning with Hyperbaric Oxygen” by Raphael . . . . . . . . . . . . . . . . . . . . . . . . 18.2.3.6 “Hyperbaric or Normobaric Oxygen for Acute Carbon Monoxide Poisoning: a Randomised Controlled Clinical Trial” by Scheinkestel . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.3.7 “Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning” by Weaver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

392 393 394 395 398 400

401

402

404 405

406 412 420 391

8417: “8417_c018” — 2007/9/19 — 13:22 — page 391 — #1

Carbon Monoxide Poisoning

392

18.4 Where to Now? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

422 426

18.1 INTRODUCTION Superficial consideration suggests that hyperbaric oxygen (HBO) should be the optimal antidote for acute carbon monoxide (CO) poisoning. It is a well-established form of therapy, albeit with a limited distribution of facilities, and the side-effect and complication profile is well established, manageable and rarely associated with longterm sequelae. By generating the highest tolerable intravascular partial pressure of oxygen, HBO provides the most rapid means available of simultaneously reversing cellular hypoxia and accelerating the elimination of CO, not only from its binding with hemoglobin, but also from intracellular binding sites. Human clinical use of HBO for this purpose appears to have first occurred in 19421 and was proposed again by Pace in the publication “Acceleration of CO elimination in man by high pressure oxygen” in 1950. This US Navy study measured acceleration of elimination of CO in volunteers treated with 2.5 atmospheres absolute (ATA) pressure oxygen after a brief loading exposure with CO, producing carboxyhemoglobin (COHb) in the range of 20–30% 2 . CO poisoning subsequently became established as one of the principal indications for HBO therapy (HBOT) as clinical use of HBO grew in the 1960s. Acute CO poisoning has subsequently been listed as an indication for HBO in the guidelines produced by multiple international hyperbaric medicine societies. Despite this, HBO has failed to become the standard of care for CO poisoning. For instance, in the US northwest, it was estimated in 1994 that only 6.9% of CO poisoning patients presenting to emergency departments received HBO.3 The historical reasons for this were probably related to the limited availability of hyperbaric chambers, limited awareness of this form of therapy, a reluctance to expose patients to the hazards of transfer for treatment, as well as a degree of general scepticism regarding the therapy. In recent times, clinical trial results and debate surrounding these have questioned the effectiveness of HBO in producing significant improvements in outcome. In the US, the number of CO-poisoned patients treated with HBO has remained relatively constant since 1992 at about 1500 per annum4 despite a more than doubling of hyperbaric facilities in that time. The lack of increase in patients treated in this environment may be a result of increasing doubt as to the effectiveness of HBOT, although any change in usage must be interpreted in the light of changes to the rates of poisoning. In the US, survey results and toxicology service data indicate that while the CO related death rate has fallen, nonfatal CO emergency calls have remained relatively stable. This correlates with the situation regarding HBO use.4 In Australia, by contrast, the number of CO poisoning patients treated by hyperbaric units fell from 240 per annum to 60 over the last 10 years. In Australia the most common cause of CO poisoning has been automobile exhaust suicide attempt. The incidence of this has fallen in association with depression recognition, suicide prevention campaigns and the reduction in CO production mandated for newer model vehicles.5,6 At the same

8417: “8417_c018” — 2007/9/19 — 13:22 — page 392 — #2

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

393

time, it is likely that public health measures have reduced occupational and domestic exposures. The decline in CO poisoning numbers treated with HBO may thus be as much a result of reduced incidence of poisoning as it is of reduced provision of treatment arising from implementation of evidence from clinical trials. Meanwhile, our understanding of the biology, pharmacology, and toxicology of both CO and HBO has become much more complex, and it is clear that both these low molecular weight gases (i.e., CO and oxygen) are two edged swords—essential elements of normal physiology but toxic at higher doses, and thus potentially useful therapeutically yet capable of harm at doses which overlap with those required for therapeutic effect.7−13 Alongside the sometime impassioned debate about interpretation of clinical trials, there is thus a fascinating and rapidly evolving stream of basic science research which will hopefully soon deliver us a better basis for designing future clinical trials aimed at answering the question of whether HBO has a place in the routine treatment of CO poisoning, and if so, in what clinical settings, and at what dose.

18.2 HYPERBARIC OXYGEN HBO provides a means of elevating oxygen concentration in the body to the maximum tolerable, in order to seek therapeutic effects not achievable with administration of oxygen in the normal ward environment, with its 1 ATA pressure at sea level, or less at altitude. Hyperbaric chambers are used to expose patients to pressures that are normally 2–3 times atmospheric, that is, 2–3 ATA. Combined with 100% inspired oxygen breathing, this can deliver arterial partial pressures as high as 15–20 times normal. As a result, not only is the volume of oxygen delivered to tissues increased but more importantly, intracellular partial pressures of oxygen increase well beyond normal, providing potentially useful effects both directly and by inducing a response to the oxidant stress generated. The technology for providing HBOT is well established. Pressure vessels for human occupancy have been built for clinical use for over 150 years, although the early proponents of HBOT mostly believed that it was pressure that was therapeutic and increased levels of oxygen were not used, limiting the therapeutic effects possible. During the 1930s, Cunningham used an air pressurized hyperbaric chamber for patients with severe acute respiratory disease, probably maintaining life through a crisis period by effectively increasing the partial pressure of oxygen available to patients, albeit in a much more complicated fashion than achievable now via administration of facemask (normobaric) oxygen (NBO) therapy. The clinical combination of increased oxygen with increased pressure was first promoted in the 1950s as a means of increasing the time window for cardiovascular surgery requiring circulatory arrest and in the 1960s a significant number of large hyperbaric chambers were constructed in hospitals around the world. Many of these are still in use today although surgery is now rarely performed under pressure. These operating theater chambers and smaller chambers evolved from commercial diving industry recompression chambers. They have been the basis for the development of the modern “multiplace” chamber. Over a similar period, single person, oxygen filled

8417: “8417_c018” — 2007/9/19 — 13:22 — page 393 — #3

Carbon Monoxide Poisoning

394

“monoplace” chambers have evolved as an alternative, originally used to enable coadministration of HBO and radiotherapy, but now widely used for both hospitalized and ambulatory patients. Ideally, hyperbaric facilities should be located inside the building structure of a hospital, close to relevant patient care areas to minimize transport difficulties, and they should have good critical care support. The most versatile multiplace chambers have multiple compartments, are quiet and well lit, have doors that will readily admit a trolley or even better a full sized patient bed and have a floor level matching that of the surrounding building. In recent years, a number of manufacturers have provided rectangular chambers that have further advanced the aim of making hyperbaric chambers as close to a normal clinical room as possible. Multiplace chambers are pressurized with air to minimize fire risk and cost. Oxygen must be administered to patients via a mask or hood, or in the case of intubated patients, via a suitable ventilator. In most cases a hyperbaric nurse, doctor or paramedic attends patients undergoing treatment which typically lasts from 1.5 to 2 h overall. Although there are safety and functional limitations on what equipment can be used inside a chamber, most major hospital facilities can provide intensive care including positive pressure ventilation, arterial pressure monitoring, and inotrope infusion. There are many variations in chamber size and configuration but the predominant alternative type of chamber in use is the single person, horizontal acrylic cylinder style monoplace chamber. These are much smaller, cheaper, and easier to install than properly designed clinical multiplace chambers and are thus in widespread use in many parts of the world, especially in facilities that concentrate on using HBO for problem wound cases and the late side effects of radiotherapy. While most hyperbaric centers with a routinely used critical care capability have multiplace chambers, monoplace chambers can be used for critical care patients by experienced teams.14

18.2.1 HYPERBARIC OXYGEN EFFECTS HBO provides it’s effects via multiple mechanisms. These can be grouped into those related to gas dynamics, to restoration of function by normalization of oxygenation and those that are a therapeutic result of the oxidative challenge achieved by very high pO2 . Some, such as elimination of gas bubbles and those underlying HBO’s anti-infective and wound healing promotion effects, are probably not directly relevant to CO poisoning. It is interesting to note, however, that in many cases, effects previously thought to be brought about merely by reversing hypoxia are now understood to result from HBO triggering intracellular signaling via various oxidative and nitric oxide related biochemistry.15−18 Other mechanisms of HBO that would seem to have particular potential in CO poisoning include accelerating elimination of CO,2,19 reversing cellular hypoxia,19,20 reducing edema,21−25 up-regulating antioxidant systems,26,27 inhibiting reperfusion injury18,28 and possibly modulating other elements of the secondary injury cascade.29,30 Some models show reduced cellular necrosis and apoptosis,31−37 although the mechanisms for these effects are still emerging.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 394 — #4

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

395

A key element of HBO therapeutics is that hyperbaric sessions are of limited duration—usually a maximum of around 2 h. The limited duration and intermittent nature of HBO is an essential factor in the safety profile of this therapy. Pulmonary, optic, and central nervous system (CNS) toxicity are well recognized complications of excess exposure to oxygen at pressure and the threshold of toxicity overlaps with the therapeutic dose range.38,39 Of particular interest is acute CNS oxygen toxicity, manifested most frequently by short duration loss of consciousness (LOC) and tonicclonic seizure activity which self-terminates as toxic levels of oxygen fall in the temporary absence of respiration. The incidence of acute CNS oxygen toxicity can be very low in comparatively well, ambulatory patients, with Yildiz40 reporting only two cases in 80,000 treatments,40 although most recent author’s estimates are in the order of 3/10,000.39,41 The incidence rises significantly with higher pressure treatments and results of this are seen in the treatment of acute patients such as those suffering decompression illness.42 In CO poisoning patients suffering neurological injury, the rate of acute CNS toxicity can be significant,43 with Sloan38 reporting a 5% incidence in 297 patients seen over 10 years.38 As applies to CO poisoning, the rapidly evolving field of free-radical and oxidant/antioxidant research has major implications for our understanding of the therapeutic and toxic effects of HBO. Some years ago, there were significant concerns that short-term benefits seen with HBO might be associated with at least some burden of oxidative stress related problems such as acceleration of cardiovascular disease or even premature aging. Reassuringly, the balance of findings to date suggests that although HBO is an oxidative stressor, it up-regulates antioxidant systems sufficiently to avoid significant net damage or even, in some cases, to provide a paradoxical and beneficial net antioxidant effect. An important caution must remain, however, that the net result of HBO probably depends upon the physiology of the host receiving HBO. Without adequate nutritional substrates for antioxidant production, when faced with excessive total oxidative stress or when specific processes are in train such as lipid peroxidation, HBO might well exacerbate the problem rather than ameliorate it, at least in some doses or at some time points.

18.2.2 BASIC SCIENCE OVERVIEW IN USE OF HYPERBARIC OXYGEN The history of using HBO for CO poisoning is sometimes claimed to date back to Haldane’s classic 1895 paper: “The relation of the action of carbonic oxide to oxygen tension”.44 This paper in fact demonstrates that, in a mouse model, the high levels of oxygen achievable in the hyperbaric environment can make normally lethal levels of CO exposure tolerable without mortality. Confusing this demonstration of the competitive inhibition of CO uptake with the therapeutic use of HBO after CO poisoning underscores one of the major challenges to date in trying to draw clinically useful conclusions from animal studies. Most human CO poisoning patients experience a delay of many hours between the termination of CO poisoning and the start of HBO, while most animal researchers have initiated HBO soon after termination of CO exposure, and often immediately. The effectiveness of HBO and indeed the entire basis for interaction between HBO and CO poisoning pathology are likely to

8417: “8417_c018” — 2007/9/19 — 13:22 — page 395 — #5

396

Carbon Monoxide Poisoning

be very different at different time points after exposure: from the acute rescue and resuscitation phase, through the period when it may be possible to prevent secondary injury, to modulating established secondary injury processes, attempting to prevent delayed neurological sequelae (DNS) and even using HBO to accelerate recovery or for delayed treatment for established residuae. Although there is some debate regarding the degree to which the acute manifestations of CO poisoning are mediated by a shortfall in intracellular oxygen availability, there is no doubt that hypoxia is a significant factor, at least in severe cases.45,46 . It may be that the level of hypoxia that is critical for any particular individual’s physiology is a key determinant of susceptibility to toxicity at any given level of CO exposure. The elderly, collapsed patients and those affected by certain drugs and co-toxins will have limited capacity for compensatory cardiac output and cerebral blood flow increase as the oxygen carrying-capacity of blood falls. This may underlie the susceptibility of these populations to CO poisoning. High altitude (hypoxic/hypoxia) exposure increases the toxicity of CO.47−52 HBO dissolves sufficient oxygen in plasma to meet physiological demands, meaning that oxygen can be immediately delivered to cells despite the presence of high COHb levels. Clinically this can be demonstrated by resolution of abnormal electrocardiogram (ECG) activity and rapid recovery of consciousness, which are sometimes but not always seen when HBO is used for CO poisoning. HBO also reduces COHb levels much more rapidly than can be achieved by NBO. Although there are interindividual and inter-study differences in actual numbers reported, the human half-life reduction is in the order of 4.5 h on air, to 90 min on 100% oxygen, to 20 min with HBO at 2.5–3 ATA.53 In the patient with significant cellular hypoxia resulting from CO poisoning, HBO can therefore expedite resolution of this problem and this would be expected to not only provide immediate benefit but to reduce the risk of ongoing hypoxia compounding existing injury. Because the uptake of CO within cells is slower than the association of CO with hemoglobin, it could be hoped that early HBO might prevent the onset of cellular toxicity.46 HBO will also speed the dissociation of CO from intracellular sites, although the proportional acceleration is different for different biochemical processes and varies with species studied. All of this would seem to provide a case for the immediate use of HBO for CO poisoning where this is can be made available. Some high risk industrial site and paramedic rescue systems have been established using portable hyperbaric chambers for this purpose.54,55 Whether very early HBO actually produces more survivors and better outcomes than NBO has not been definitively established, given that all major clinical trials to date have had insufficient numbers of patients treated within, say, the first hour after CO poisoning. Of the randomized human studies, only that of Ducasse56 came near to this with all patients treated within 2 h of CO exposure. However, this study was small and used outcome measures that are difficult to interpret as will be discussed. Most animal studies to date do seem relevant to the question of early therapy benefit, however with therapy usually provided to exposed animals commencing either immediately after cessation of poisoning or after a relatively short delay—most usually between 15 and 60 min.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 396 — #6

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

397

In many animal models of CO poisoning, HBO has shown benefit on a range of measures including survival, recovery of consciousness, recovery of brain function and reduction in postmortem neuropathology, including reductions in cellular necrosis and apoptosis. Despite early initiation of treatment, it is noteworthy that not all animal studies show benefit of HBO over NBO, however, or even of any kind of oxygen therapy over air breathing. Amongst recent studies looking at mechanisms underlying longterm sequelae, Brvar et al.57 did demonstrate that in rats poisoned without loss of consciousness (LOC), 30 min of 3 ATA HBO, but not 30 min of NBO, significantly reduced an immediate post-CO rise in blood levels of S100B, an astroglial structural protein that shows promise as a marker of CO poisoning severity.57 The CO exposure was 3000 ppm for 60 min. The same group has also reported however, in the same model, that both NBO and HBO resulted in a dramatic but effectively equal reduction in pyknotic cells in the hippocampus when the rats were sacrificed two weeks later.58 By contrast, in a mouse model of more severe poisoning with LOC, Gilmer et al.59 found that neither NBO nor HBO commenced 15 min after poisoning provided any significant protection against learning dysfunction or hippocampal pyknosis.59 LOC was induced in this model with a 4–9 min exposure to 50,000 ppm of CO after 40 min at 1000 ppm. It could be argued that this produces a sufficiently severe injury to be irreversible. This experimental poisoning regimen can be compared with Thom’s well established rat model which uses 3000 ppm for up to 20 min in order to induce LOC after 40 min of exposure to 1000 ppm.60 Treatment commences after a more clinically relevant interval of 45 min and the HBO regimen used is 45 min at 2.8ATA. This model was used to demonstrate that another structural protein, myelin basic protein (MBP) is released after CO poisoning and triggers delayed neuropathology via a cell mediated immune response.61 Most recently, Thom62 has shown that intervention with HBO, but not NBO, reduces learning dysfunction and the sensitization of lymphocytes to MBP, but this effect is only partial.62 This study adds to work showing that early HBO can block leukocyte adherence to endothelium and reduces leukocyte mediated oxidative damage that compounds CO-related injury upon reoxygenation.63 Interpreted together, the mechanisms of action of HBO and the studies referred to above suggest there is potential for benefit from early HBO. It is not clear, however, to what extent reversal of acute toxicity is responsible for outcome benefit as opposed to secondary injury reduction via mechanisms such as modulation of delayed immune-response and up-regulation of antioxidant systems. It is also not clear whether these effects are useful if HBO is applied much later after termination of CO exposure. Oxygen dosing is another major variable, with various pressures between 1.5 and 3.0 ATA used in therapeutic research, for durations between 45 min and 2 h. Single dose therapy and multiple dose therapy at varying intervals, have been used. In clinical practice, therapeutic oxygen is used in variable amounts and durations before and during hospital admission and therapeutic and toxic effect in animals should ideally be studied with varying doses of HBO and both with and without NBO in the periods before and after HBO, if findings are to be translated into human clinical practice or research design. It is also important to remember that developing a satisfactory animal model is critical to laboratory research. An experimental poisoning regimen must be found that produces sufficient toxicity to be associated with measurable pathology but the

8417: “8417_c018” — 2007/9/19 — 13:22 — page 397 — #7

398

Carbon Monoxide Poisoning

animals must survive and the damage should not be so severe as to be irreversible. Such controlled exposures are not the reality for humans and even in the animal models, demonstrated benefit may only apply in certain degrees of injury severity and both the optimal HBO dose and level of poisoning amenable to therapy seem likely to be species specific. The issue of animal model generalizability to humans is particularly important in CO poisoning. Most CO poisoning studies are done in rodents, which have the advantages of being relatively cheap, easy to handle and importantly, they are now available in a wide range of genetically altered strains which can be used to test biochemical mechanism hypotheses. Although rodents are mammals, they have critical differences from larger species. Their small body mass and high metabolic rate result in much more rapid uptake and distribution of gases than is the case for humans.52,64 More importantly, rodents are relatively hypoxia and carbon dioxide resistant, presumably evolved traits associated with survival in the rodent environment but ones which may have critical impacts on the interpretability of findings regarding CO poisoning, given the importance of hypoxia in the pathology of CO poisoning. This is not to argue against rodent research; it has generated and will continue to generate significant advances in our knowledge regarding the mechanisms of both CO and HBO. Generating clinical outcome predictions, case selection criteria or dosing requirements from rodent work has significant potential for error however. Gorman argues this in reporting results from his instrumented sheep model of CO poisoning. In this model, animals lose consciousness with an exposure to 1% CO for 120 min and seem to tolerate LOC-inducing levels of CO poisoning without the same neuropathology described in other models, although some peri-ventricular white matter infarcts and gliosis do occur. In addition to species differences, tolerance almost certainly depends upon cardiovascular and cerebrovascular reflex responses being intact and sufficient to maintain ongoing cerebral oxygenation. This has required chronic instrumentation of the animals, as if CO poisoning is provided too early after cannulation of the carotid vessels, more severe neuropathology was seen, localized to the side of instrumentation possibly related to impairment in the normal compensatory increase in cerebral blood flow.45,65−68 In summary, most animal studies suggest benefit should arise from using HBO as a therapy for CO poisoning, but much caution is needed in translating findings from pure and highly controlled CO exposure in healthy animals, treated rapidly with HBO, to the messy and complicated realities of human CO poisoning.

18.2.3 HBO FOR CARBON MONOXIDE POISONING: HUMAN EVIDENCE In October 2006, the American College of Emergency Medicine released its draft “Clinical Policy: Critical Issues in the Management of Adult Patients Presenting to the Emergency Department with Acute Symptomatic CO Poisoning”.69 This document uses an evidence-based approach. In response to the question “Should HBOT be used for the treatment of patients with CO poisoning; and can clinical or laboratory criteria identify CO-poisoned patients who are most likely to benefit from this therapy?”—the best it could do was “Level C” recommendations, that is, recommendations based

8417: “8417_c018” — 2007/9/19 — 13:22 — page 398 — #8

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

399

on preliminary, inconclusive or conflicting evidence. The Level C recommendations were: 1. HBOT should not be mandated for the treatment of CO poisoning. 2. HBOT remains a therapeutic option to potentially reduce the incidence of neurological sequelae. 3. Do not use COHb levels alone to choose therapy in CO poisoning. 4. The available evidence does not identify either a subgroup of patients for whom HBOT is clearly indicated, or a subgroup of patients who clearly have no potential to benefit from HBO. Prior to this, Phin,70 in reviewing the evidence for therapy of CO poisoning, concluded that there was no evidence to support HBOT being of clear effectiveness. Evidence—based on call: Acute Medicine in 200171 recommended avoiding HBO in CO therapy, classifying this as an A class recommendation. In November 2000, The Medicare Services Advisory Committee of the Department of Health and Aged Care of the Australian Government,72 released its Report on HBOT. It concluded that there should be no support for “public funding for HBOT in either a multi-place or mono-place chamber for CO poisoning.” On February 9, 2001, the Minister for Health and Aged Care accepted this recommendation and such funding was withdrawn. This recommendation was based on a systematic review by Juurlink et al.73 in 2000 on behalf of the Cochrane Database of Systematic Reviews. “The review collected six reports of randomized controlled trials involving nonpregnant adults acutely poisoned with CO. Only three studies scored ≥ 3/5 on the Jadad quality scale (assessment based on randomization, double blinding and withdrawals and dropouts),74 and these three by Raphael,75 Thom,76 and Scheinkestel77 were analyzed. Juurlink et al.73 found that the severity of CO poisoning varied between trials and each trial employed different doses of HBO. The results for a total of 455 patients were available for analysis. Nonspecific neurological symptoms were present in 81/237 patients (34.1%) in the HBO group compared to 81/218 (37.2%) in the NBO group [odds ratio (OR) = 0.82; 95% Confidence Interval (CI) = 0.4, 1.66]. This systematic review failed to demonstrate a significant reduction in neurologic sequelae following HBOT for CO poisoning.” The German government has similarly recommended discontinuation of HBOT for CO poisoning78 stating “no studies could be identified, which could justify a continuation of HBOT for CO poisoning.” The review by Juurlink et al.73 underwent a substantive amendment in November, 2004 and was subsequently published in the Cochrane Library 2006.73 In this review, six trials were evaluated, the previous three plus a further one from Raphael,79 one from Mathieu,80 and one from Weaver.81 Of these, the latest one from Raphael is in abstract form only and that from Mathieu is an interim analysis. Using the Jadad scale again,74 Mathieu’s trial scored 2/5, the two from Raphael and the one by Thom scored 3/5 with those by Weaver and Scheinkestel scoring 5/5. Juurlink et al.73 state: “Of the six trials included, four found no benefit of HBO for the reduction of neurologic sequelae, while two did. While pooled analysis does

8417: “8417_c018” — 2007/9/19 — 13:22 — page 399 — #9

400

Carbon Monoxide Poisoning

not suggest a benefit for HBOT, (OR for neurological deficits = 0.78; 95% CI = 0.54, 1.12, p = 0.18), significant methodological and statistical heterogeneity means that this result must be interpreted with caution. Design and analysis flaws were evident in all the trials and importantly, the conclusion of one positive trial may have been influenced by failure to adjust for multiple hypothesis testing while the other positive trial is hampered by apparent changes in the primary outcome during the course of the trial.” Juurlink et al.73 again concluded that: “there was no evidence to support the use of HBO for treatment of patients with CO poisoning.” It is worth carefully reviewing all these studies, and also including for review that by Ducasse,82 which is also a prospective, randomized trial, but not usually included because of the use of surrogate outcome measures. Most of the discussion will concentrate on the two main studies: Weaver’s and Scheinkestel’s. Case series, retrospective reviews, nonrandomized studies, animal work and manuscripts based on theory have not been included in this review. They are discussed in chronological order below. 18.2.3.1 “Trial of Normobaric and Hyperbaric Oxygen for Acute Carbon Monoxide Intoxication” by Raphael Raphael’s study published in 1989,75 enrolled only patients poisoned in the 12 h prior to hospital admission. Three hundred and forty-three (343) patients with mild CO poisoning (no impairment of consciousness) were randomized to receive either NBO or HBO. Two hundred and eighty six (286) severely poisoned patients (with impairment of consciousness) were randomized to either one or two sessions of HBO, 12 h apart. Critically ill patients were excluded. Patients refusing the allocated treatment after randomization (n = 19), were still retained in the study and analyzed according to treatment intended. NBO therapy consisted of 6 h of 100% inspired oxygen by facemask or endotracheal tube. HBOT was 2 h of HBO in a mono-place chamber (0.5 h for compression, 1 h at 2.0 ATA and 0.5 h for decompression), plus 4 h of NBO. Eleven percent (11%) of patients were lost to follow-up. Thirty-nine patients (39) were intolerant of HBO and five had confirmed barotrauma. Raphael reported persistent neurological sequelae (PNS) in 32–34% of the patients with no LOC, and in 46–48% of those with LOC. The diagnosis of PNS was based on gross signs and symptoms, as neuropsychological testing was not performed. Patient assessment at 1 month consisted of a self-assessment questionnaire and a physical examination performed by the patients’ own doctor. If there was no response to the questionnaire, patients were telephoned. Raphael concluded that in patients without LOC, HBO had no advantage over NBO. At the 1-month review, recovery occurred, respectively, in 66% of 170 NBO patients and 68% of 173 HBO patients. Ninety-seven percent of patients resumed their usual occupation and social activities irrespective of treatment. In patients with transient LOC, he found no difference in outcome at 1 month between those patients having one or two HBO treatments (54% versus 52% recovery, p = 0.42).

8417: “8417_c018” — 2007/9/19 — 13:22 — page 400 — #10

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

401

The problems identified with this study include: • Individuals were excluded on the basis of COHb measurement alone. • Those patients who refused randomization were retained in the study and included in the final analysis as members of intended treatment group.83 • NBO was given by facemask only, making the exact percentage of inspired oxygen unclear. • Lack of true controls. Only less severely CO poisoned patients were randomized to HBO versus NBO groups.78 All severely poisoned patients received HBO. • The HBO regimes used are considered by some to be ineffective (2.0 ATA instead of 2.8 ATA).84 • The times from poisoning to treatment entry criteria (up to 12 h) were too long.85,86 Oxygen treatment did not begin until after more than 6 h after poisoning in approximately half the cases.75 • Patients were not stratified according to interval between exposure and therapy.83 • Neither the investigators nor the patients were blind to treatment group. • The use of insensitive outcome measures (self-assessment questionnaire by telephone or mail one month after poisoning, discussion with the patients’ personal physicians),87 with recovery being determined by a lack of symptoms and/or resumption of former activities and delayed neurologic sequelae (DNS) being diagnosed when patients reported any of a variety of complaints.88 • Not using neuropsychometric tests to assess outcome resulted in an absence of objective or quantitative evaluation of cortical function. • The physical examination was performed by the patients’ own physician, rather than a physician experienced in CO-poisoning. Inexperienced physicians, not used to assessing these patients, may miss the more subtle signs and symptoms. • The use of multiple physicians precluded consistency. • 70 of 629 (11.1%) patients were lost to follow-up at 1 month. 18.2.3.2 “Non-Comatose Patients with Acute Carbon Monoxide Poisoning: Hyperbaric or Normobaric Oxygenation?” by Ducasse Ducasse’s56 series comprised 26 noncomatose patients with a glasgow coma score (GCS) of 15 on admission; 13 patients were randomized to HBO and 13 patients were randomized to NBO. They started treatment at a mean time after CO exposure of 53 min. The HBO group received HBOT for 120 min at 2.5 ATA pressure, followed by 100% NBO for 4 h and a further 6 h of 50% NBO. The NBO group received oxygen through a face-mask at 100% for 6 h, then at 50% for 6 h. Patients were assessed on clinical signs and symptoms, electroencephalogram (EEG) and cerebral blood flow response to acetazolamide. No neuropsychological assessments were performed. The reported incidence of both persistent neurologic sequelae (PNS) and DNS was zero.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 401 — #11

Carbon Monoxide Poisoning

402

Clinical assessment at 2 and 12 h favored the HBO group. Two patients in the NBO group were changed to HBO at 12 h and were asymptomatic at completion of treatment. All 26 patients were discharged home well, an average of 28 h after presentation. There was no difference in the EEG at 24 h between HBO and NBO groups. Eight of twenty-six patients (31%) were lost to follow-up (three in the NBO group and five in the HBO group). Follow-up EEG in the remainder was worse in the NBO group (p ≤ 0.02), but all patients were clinically normal. Cerebral blood flow was assessed in 20 patients (four HBO, six NBO, and ten controls). There were no differences between HBO, NBO and controls with regard to perfusion of the basal ganglia or in cerebral blood flow values. Reactivity to acetazolamide was similar in the controls and the HBO group, and were said to be statistically significantly different from the NBO group (p ≤ 0.04). Adverse events due to HBO are not mentioned. Ducasse concluded that HBO reduced the time to initial recovery and the number of delayed functional abnormalities in noncomatose patients with acute CO poisoning. He attributed some of his success to the rapidity of treatment. The problems identified with this study include: • • • • • • • • • • • •

Small study with few patients.26 Only patients with mild CO poisoning were enrolled. The study was nonblinded. The use of surrogate outcome measures which are of questionable significance.85,87,89 The significance of an abnormal EEG in clinically normal patients is unclear. No statistical significance values were reported for the EEG results. Thirty-one percent (31%) of subjects we lost to follow-up. While controls were part of the 3-week evaluation, the authors did not define this subgroup’s characteristics. There were no measures of cognitive function, as standardized neuropsychiatric testing was not performed. The test data appear to conflict with one of the author’s conclusions: that HBO treated patients showed a better cerebral blood flow response and greater improvement in the 3 week EEG studies.83 Inadequate allocation concealment.89 Statistical significance of reactivity to acetazolamide is questionable, given the small numbers and large range of results.

18.2.3.3 “Delayed Neuropsychologic Sequelae after Carbon Monoxide Poisoning: Prevention by Treatment With Hyperbaric Oxygen” by Thom Thom’s study76 also included only mild poisonings. Patients with a history of LOC were excluded. Patients presented within 6 h of exposure and usually commenced

8417: “8417_c018” — 2007/9/19 — 13:22 — page 402 — #12

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

403

treatment in about 1 h. Thirty-two patients were allocated to NBO and thirty-three to HBO. HBO patients were treated once for 30 min at 2.8 ATA, followed by 90 min at 2.0 ATA. NBO patients received 100% oxygen through a nonrebreather face-mask until all symptoms resolved (4.2 ± 3 h). The presence or absence of five signs and symptoms together with a COHb level were used to assume the two groups were of similar severity of poisoning. Thom used a CO neuropsychological screening battery (CONSB) designed by Myers after completion of treatment. Formal neuropsychological testing was performed at 1 month. The 3-month review consisted of a telephone interview only. Twelve of sixty-five patients (18%) were lost to follow-up. Thom reported no DNS in the HBO group, while 7 of 30 in the NBO group developed problems (p < 0.05). No specific treatment was given to those who developed DNS. Three of these patients refused follow-up. In the remaining four, neuropsychometric testing was repeated at intervals of 2–3 weeks until scores returned to baseline. It would appear that all DNS resolved. No adverse events resulted from HBO treatment. Thom concluded that HBOT decreased the incidence of DNS after CO poisoning. Problems reported with this study include: • • • • • • • •

•

Neither the patients nor the investigators were blinded to treatment. The randomization process is unclear. The consent procedures are unclear. Only patients with mild poisoning were included, sick patients were excluded.90 While a COHb level was one of the measures used to assume that the two groups were of similar severity of poisoning, the delay taken in measurement of COHb is not mentioned. Baseline neuropsychometric testing was not performed to ensure that the two groups were similar. There was greater comorbidity in the NBO group at randomization.85,87 . NBO patients were slightly older and had a higher incidence of cardiovascular and respiratory disease. Neuropsychological tests were repeated at intervals of 2–3 weeks until scores returned to baseline. The effect of repetition and learning of tests on outcome was controlled by comparing to the effects achieved by practice on learning in a control group of eight patients. This is an insufficient number of control subjects to assist with the interpretation of neuropsychological tests86,91 and no details of the characteristics of the control group were provided, in particular how they matched with the HBO and NBO groups.90,91 The location and conditions for neuropsychometric testing were inconsistent. Some were performed immediately on completion of treatment, but if patients were “fatigued”, the tests were performed in the patients’ homes within the next 12 h. No details are provided as to how the neuropsychometric testing was performed on the controls.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 403 — #13

404

Carbon Monoxide Poisoning

• The paper does not state the number of clinicians involved in performing the neuropsychometric testing, nor their experience. • The neuropsychological tests used (CONSB) are said not to adequately measure memory. • The definition of DNS was a deterioration in one or more subtest scores on the neuropsychological test battery, but this “deterioration” is not defined in the paper.91 • The incidence of delayed sequelae may have been significantly altered in the normobaric group if normal psychometric testing had been a prerequisite for discharge.88 • At the 4 week follow-up, the NBO group had a worse score in one subtest only, Trail-Making, which Thom states may reflect the presence of a subtle impairment of learning ability when these patients first took the psychometric test and hence a lack of familiarity on retesting.76 • The 3-month review consisted of a telephone interview only. • 18% of patients were lost to follow-up.86 • All patients reported complete resolution of symptoms. Hampson89 states that this study was stopped early due to a treatment advantage in the HBO group, but the actual paper makes no mention of this. In fact, Juurlink73 points out that in an interim analysis describing the outcome of 58 patients, published in 1992, there was no difference in symptoms between patients in the NBO versus HBO groups (4 of 29 patients versus 0 of 29 patients, respectively). Juurlink notes that seven additional patients were recruited after this interim analysis, three to the NBO arm (all of these patients experienced neurological sequelae), and four to the HBO arm (none of these experienced neurological sequelae). He states that although the recruitment of seven additional patients with this distribution of allocation and outcomes could be due to chance (p = 0.014 by Fisher’s exact test), it may reflect premature termination of the trial after recruitment of only seven more patients, greatly exaggerating the treatment effect. A statistical penalty to adjust for inflation of the type I error rate was not introduced, and would have rendered the final result statistically insignificant.73 Had adjustment for multiple comparisons been performed, no significant difference between treatments would have been identified.92 18.2.3.4 “Randomized Prospective Study Comparing the Effect of HBO Versus 12 h NBO in Noncomatose CO Poisoned Patients: Results of the Interim Analysis” by Mathieu Mathieu80 reported on an interim analysis after 3 years of a 5-year multicenter study. Only patients noncomatose on hospital admission, and poisoned in the preceding 12 h were enrolled. Delay to treatment is not specified. Treatment was either one HBO session of 90 min at 2.5 ATA (299 patients), or 12 h of 100% NBO (276 patients). Patients were neurologically normal at the time of hospital discharge, but at 1 month approximately one quarter had sequelae with no difference between HBO and NBO (23% versus 26%). At three months, the incidence

8417: “8417_c018” — 2007/9/19 — 13:22 — page 404 — #14

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

405

fell and there was a statistically significant difference between HBO and NBO (9.5% versus 15%; p = 0.016), but this was no longer evident at 6 months (6.4% versus 9.5%; p = 0.09) or 12 months (4.3% versus 5%). The limitations of this report are: • • • • • • • • • •

It is only available in abstract form. There was no blinding. The definitions of the primary outcome are missing.92 There was a long entry time of 12 h from poisoning to treatment.85 No details of the neurological assessment are provided. Neurological manifestations are referred to as PNS, but given that the patients were neurologically normal at the time of hospital discharge, it is not clear why they are PNS and not DNS. Neuropsychological assessments were not performed either pre- or posttreatment. Mathieu provides no details of attrition rates for follow-up.92 Recovery was considered complete if the patient had no complaints.88 Had the investigators adjusted their analysis for multiple comparisons, no significant difference between treatments would have been identified at any interval.92

18.2.3.5 “Managing Carbon Monoxide Poisoning with Hyperbaric Oxygen” by Raphael Raphael79 in 2004, published in abstract form the results of a second randomized controlled trial. Patient recruitment took place between 1996 and 2000 and involved 385 victims of accidental, domestic CO poisonings presenting within 12 h of CO exposure. One hundred and seventy nine (179) patients had transient LOC and were randomized to receive either 6 h of NBO (A0 group, 86 patients) or one treatment with HBO (at a plateau of 2 ATA for 60 min), plus 4 h of NBO (A1 group, 93 patients). Two hundred and six (206) comatose patients were randomized to receive either one (B1 group-101 patients) or two (B2 group-105 patients) HBO treatments (at a plateau of 2 ATA for 60 min) plus 4 h of NBO. The primary end-point was the proportion of patients who recovered at 1 month, with recovery being defined by a normal self-assessment questionnaire and normal blinded neurological examination. The trial was stopped prematurely because the interim analysis showed that giving two HBO sessions rather than one, was associated with a poorer outcome in comatose patients. Furthermore, in patients with transient LOC, the recovery rate was not modified by addition of HBO to NBO. The percentage who had recovered at 1 month were as follows: A0 61% 45/74 A1 58% 46/79 (OR = 0.90, 95% CI = 0.47–1.71, p = 0.87) B1 68% 54/80 B2 47% 42/90 (OR = 0.42, 95% CI = 0.23–0.79, p = 0.007)

8417: “8417_c018” — 2007/9/19 — 13:22 — page 405 — #15

Carbon Monoxide Poisoning

406

There were four deaths and four survivors with severe neurological sequelae in the B2 arm, (two HBO treatments). All but one of the deaths and severe neurological sequelae were in the B group, that with comatose patients receiving HBO treatment. Severe neurological sequelae occurred only in those with coma at presentation. Seventeen (17%) percent of patients were lost to follow-up. The authors concluded “There is no evidence supporting the routine use of HBO in patients with acute CO poisoning” The limitations of this report are: • All the ones of his previous study, as the trials are essentially the same: • Lack of true controls. Only less severely poisoned patients were randomized to HBO versus NBO.73 All severely poisoned patients received HBO. • The HBO regimes used are considered by some to be ineffective (2.0 ATA instead of 2.8 ATA).84 • The times from poisoning to treatment entry criteria (up to 12 h) were too long.85,86 • Patients were not stratified according to interval between exposure and therapy.83 • Neither the investigators nor the patients were blind to treatment group. • The use of insensitive outcome measures (self-assessment questionnaire by telephone or mail 1 month after poisoning, discussion with the patients’ personal physicians),87 with recovery being determined by a lack of symptoms and/or resumption of former activities and DNS being diagnosed when patients reported any of a variety of complaints.88 • Not using neuropsychometric tests to assess outcome resulted in an absence of objective or quantitative evaluation of cortical function. • The physical examination was performed by the patients’ own physician, rather than a physician experienced in CO-poisoning and inexperienced physicians, not used to assessing these patients may miss the more subtle signs and symptoms. • The use of multiple physicians precluded consistency. In addition, • The study is only available in abstract form. • 17% of patients were lost to follow-up. 18.2.3.6 “Hyperbaric or Normobaric Oxygen for Acute Carbon Monoxide Poisoning: A Randomised Controlled Clinical Trial” by Scheinkestel We98 commenced a trial in 1993 in which we aimed to overcome the shortcomings of the previous trials. We randomized patients with all grades of CO poisoning, including severely poisoned patients, within 24 h of poisoning, used sham treatments in the multiplace chamber for the NBO group with both patients and the outcome assessor

8417: “8417_c018” — 2007/9/19 — 13:22 — page 406 — #16

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

407

blinded to treatment group. Our primary end-point was outcome at completion of treatment. Our secondary end-point was outcome at 1 month. In the absence of universally accepted recommendations for depth or duration of pressurization for HBO treatments, we used what was at the time, the best available data to determine the HBO treatment protocol, which consisted of three treatments of 100 min with 60 min at 2.8 ATA over 3 days, based on the conclusions of Gorman and Runciman93 that such protocols were associated with the lowest mortality and neurological deficits and would provide the maximum potential advantage for HBOT. In between treatments, patients received high-flow oxygen by nonocclusive facemask. In order for the treatment and control groups to be identical in every way except for the hyperbaric component, the NBO group had exactly the same treatment as the HBO group but had NBO in the chamber. Patient assessment at entry included the clinical effects of poisoning and a minimental examination. After the third treatment, patients were reassessed medically and underwent full neuropsychological testing. Patients with a poor outcome had a further three treatments (NBO or HBO as previously allocated) and continued with high flow oxygen between treatments. We used a pretreatment, baseline minimental examination, an extended series of neuropsychological tests to assess both persistent and delayed neuropsychological deficits, a clinical psychologist trained in neuropsychological assessment of brain injured patients to perform all the tests (at completion of treatment and at follow-up), and computerized testing to standardize administration and increase objectivity of these tests. Following consent, patients were stratified into four groups: suicide versus accidental, then mechanically ventilated versus non-ventilated prior to randomization to HBO (104 patients) or NBO (87 patients). In our prospective randomized controlled trial of 191 patients, in which both groups received high doses of oxygen, the HBO regimen used, did not benefit and may have worsened outcome. More patients in the HBO group required additional treatments (28% versus 15%, p = 0.01 for all patients; 35% versus 13%, p = 0.001 for severely poisoned patients). HBO patients had a worse outcome in the learning test at completion of treatment (p = 0.01 for all patients; p = 0.005 for severely poisoned patients) and a greater number of abnormal test results at completion of treatment (3.4 versus 2.7, p = 0.02 for all patients; 3.7 versus 2.6, p = 0.008 for severely poisoned patients). A greater percentage of severely poisoned patients in the HBO group had a poor outcome at the completion of treatment (85% versus 65%, p = 0.03). DNS were restricted to the HBO patients (p = 0.03). No outcome measure was worse in the NBO group. The comprehensive assessment of all patients at completion of treatment showed no benefit for HBO. Our study has attracted a number of criticisms: •

Cluster randomization

In the accompanying editorial, Moon and DeLong,99 while acknowledging that “the study design is amongst the most rigorous yet published”, expressed concern about

8417: “8417_c018” — 2007/9/19 — 13:22 — page 407 — #17

Carbon Monoxide Poisoning

408

cluster randomization. Kao87 and Weaver100 expressed similar concerns. To minimize the impact of the trial on daily practice, cluster randomization was used for simultaneously presenting patients from the same exposure: the treatment group assigned to the first patient was allocated to the other simultaneously presenting patient(s). Randomization took place only after the group was assembled. Cluster randomization accounted for the difference in numbers between the HBO and NBO groups. We used cluster randomization, allocating more than one person simultaneously to the same treatment on 22 occasions, (2 patients on 12 occasions, 3 patients on 5 occasions and 4 patients on 5 occasions). Overall 14 clusters (40 patients) were allocated to HBO and 8 clusters (19 patients) to NBO. As patients presenting simultaneously could be uniquely identified by having identical measurements for three continuous baseline severity measurements (exposure time, time to COHb measurement and time to treatment), any effects due to cluster randomisation could be controlled and adjusted for by including these variables in the generalized linear model. Continuous outcome variables were also analyzed by the mixed procedures in statistical analysis software (SAS)94 which allows a repeated measures analysis of variance, with the variable cluster being treated as a random repeated measurement, thus “adjusting for” within cluster variation. We also repeated the analysis excluding all patients who were allocated as part of a cluster. It has been shown95,96 that the effect of cluster randomization is to increase the size of standard errors and p-values. By including these three variables we found that the standard errors and the p-values were increased in comparison to models excluding these variables. To further validate results, sensitivity analysis was performed for the binomial outcome variables by comparing results with those obtained by excluding all participants who were randomized into a cluster.

Variable

Original OR (95% CI)

Results Excluding All Clusters

Required > 3 treatments 2.8 (1.3–6.2) 3.3 (1.3–8.2) Required > 3 treatments (severe) 5.4 (2.0–14.8) 5.3 (1.7–16.0) Poor outcome 1.7 (0.8–4.0) 2.2 (0.8–5.9) Poor outcome (severe) 3.6 (1.1–11.9) 3.2 (0.9–11.9) This analysis produced consistent results for all variables which suggests that our results have not been biased by cluster randomisation.97,98 • Allocation Concealment The Cochrane review73 scores our study a “B” for allocation concealment because of concern of failure of concealment related to fixed block sizes if the treating team were aware of previous assignment. We specifically addressed this by having blocks of various sizes so that you could not predict the next treatment. Randomization (HBO or NBO) was performed by a hyperbaric technician who opened progressively numbered, sealed, opaque envelopes (from random blocks of 4, 6, 8 and 10 envelopes, each block containing equal numbers of HBO and NBO selections).

8417: “8417_c018” — 2007/9/19 — 13:22 — page 408 — #18

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

•

409

Delay to Treatment

Both Moon99 and Weaver100 also express concern about our delay to treatment (geometric mean of 7.1 h (95% CI = 1.9–26.5 h). Firstly this is not dissimilar from some of the other studies. Raphael’s inclusion criteria was up to 12 h in both his studies74,79 with a mean time to treatment of 6.4 h74 . Mathieu’s80 entry criteria was also within 12 h of exposure. Weaver’s study81 included patients within 24 h of exposure and the mean times in his study were only marginally different from ours: for HBO 5.8 ± 2.9 versus 7.5 (6.6–8.6), NBO 5.7 ± 2.9 versus 6.6 (5.7–7.5), respectively, for Weaver’s and Scheinkestel’s patients. As Tighe points out,101 “these time delays are representative of most clinical practice because of late presentation and the need for stabilization and transport to a remote hyperbaric facility.” Although the geometric mean treatment delay was 7.1 h, we performed subgroup analysis of patients treated within 4 h (all patients and just severely poisoned patients). There were 44 patients treated within 4 h (22 HBO and 22 NBO), 33 of which were severely poisoned (15 HBO and 18 NBO) with no outcome measure favoring the use of HBO. We further analyzed time to treatment in quartiles (12 h) and found no difference in outcome between HBO and NBO. Further multivariable analysis did not identify delay in treatment as a predictor of poor outcome. Thus there was no evidence that delay in treatment could have explained the lack of benefit of HBO.97 •

Concomitant depression, suicide attempt and use of psycho-active drugs

Weaver,100 Shank,85 and Moon99 express concern that we included patients whose exposure was due to suicidal intent, had consumed cointoxicants and had a history of depression. They question whether this might have influenced the outcome of the neuropsychological tests and therefore the results of the study. While it is true that depression and the use of medication may have resulted in a higher incidence of poor outcome overall, this would not in any way have biased the comparison between normobaric and hyperbaric groups as patients were specifically stratified for suicide attempt prior to randomization to therapy. As we also stated, analysis of accidental poisonings (excluding suicide attempts) also showed no differences between HBO and NBO groups.77 Furthermore the incidence of self-administration of drugs and alcohol was identical in both groups (44%).77 In a subsequent letter to the British Medical Journal (BMJ)102 Weaver states “I agree that attempted suicide probably did not bias the outcome between the two arms”. • Lack of pretreatment neuropsychological assessment Our lack of pretreatment neuropsychological assessment has been considered a problem by Moon,99 Schiltz103 and Denson and Hay.104 In order to address our specific research question, whether HBOT is superior to NBO therapy in preventing residual cognitive impairment following CO poisoning,

8417: “8417_c018” — 2007/9/19 — 13:22 — page 409 — #19

Carbon Monoxide Poisoning

410

two alternative research designs were considered: a cross sectional design wherein subjects are randomly assigned to two groups (HBO versus NBO), and a longitudinal design where within group comparisons are made (pre-treatment verses post-treatment). When employing neuropsychological tests to evaluate patient outcome within a longitudinal design, both practice effects and spontaneous recovery may over-estimate treatment benefits. As Dr. Schiltz points out,103 it would be pertinent within a longitudinal design to take into account premorbid intelligence and psychiatric status. In selecting a randomized cross-sectional design (where both assessor and data analyzer were blind to treatment membership) however, statistical comparisons are not confounded by practice effects nor spontaneous recovery. Additionally, while we acknowledge that we could not clearly establish that our two groups (NBO and HBO) were equal with respect to premorbid intellectual activity and level of depression, there was no difference in initial mini-mental score between groups (27.0 (26.1–27.9) versus 26.4 (25.4–27.4), p = 0.27). There was similar improvement in mini-mental score in both groups, for all patients (p = 0.53), and for severely poisoned patients (p = 0.71). The score improved further to normal levels in both NBO and HBO groups in those attending follow-up for all patients as well as severely poisoned ones. Clinical considerations were equally important in selecting the most appropriate research design and methodology. Firstly, there is the issue of obtaining meaningful data. In an acutely ill, disoriented, agitated and distressed patient, prolonged psychometric testing is not practical or meaningful. In more cases than not, the patient would not be able to sustain concentration for a 1.5-h assessment. Further, availability of a trained psychologist 24 h a day, 7 days a week to perform pretreatment full neuropsychological assessments and delaying treatment by a further 90 min was not practical. We therefore performed the mini-mental test on admission, which is quick and easy to administer, rather than comprehensive neuropsychological assessment which was performed at completion of treatment once patient’s mental status had stabilized. • Type of tests There has been criticism that the neuropsychological tests we used were not standard ones.87 Neuropsychological assessment of the CO population is challenging in light of the confounding psychiatric variables. A further challenge to researchers is the collection of data in a clinical setting that needs to be sensitive to patient motivation, level of cooperation and imminent discharge. For these reasons, highly sensitive cognitive tests were selected that sampled the pertinent neuropsychological realms (attention, new learning, visuo-construction and executive functioning) in 90 min in order to maximize cooperation and minimize fatigue. The choice reaction time (RT) measure we used, required participants to respond to stimuli that appeared at various positions around the periphery of the screen and as such required rapid visual scanning. RT is considered sensitive to cerebral lesions of any localization. Should specific visuo-spatial deficits occur in the CO-poisoned population, it is likely that they would be reflected in the RT. In addition to being sensitive to diffuse cerebral lesions, the choice RT task can be argued to be sensitive specifically to visuospatial deficits.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 410 — #20

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

•

411

Multitude of tests

Moon99 and Denson and Hay104 raise the matter that we performed a multitude of tests and that only one showed a statistically significant result (in favor of NBO). Although we performed a multitude of tests, not one showed a benefit in favor of HBO. We were not trying to prove a statistically significant advantage for NBO, rather we could not demonstrate a benefit in favor of HBO. As multiple tests were considered, there is an increased likelihood of a type I error. But, given that an outcome measure based on combining all tests was used, in conjunction with death and > 3 treatments, the chance of a spurious result was minimized. We agree that there were two major limitations to our study: •

Low follow-up rate

While our study was inclusive with 83% of potential patients being entered (4% were excluded and 13% refused consent), only 46% of patients entered attended the followup at 1 month. Thus, longterm follow-up was only available in 38% of all potential patients. This unfortunate result occurred despite significant effort. Patients were requested to attend for review at 1 month. The review appointment was confirmed by mail and if required, patients were actively pursued by telephone. Despite repeated efforts, only 46% of patients attended follow-up. This low rate of attendance at follow-up is indeed a major problem in interpreting our patients’ outcomes. Our different patient population, with characteristics associated with suicide attempts and depression, many referrals from distant locations and lack of incentive, probably contributed to the low follow-up rate, which was however, equal in both groups, and evenly distributed across subgroups. For those attending follow-up, our assessment was rigorous (as opposed to a telephone survey) and failed to show any benefit for HBO. The previously published randomized studies have experienced lower but significant nonattendance rates at delayed review of 11.1%,75 31%,82 18%,76 17%,79 with Mathieu’s study not quoting the attrition rate. • The treatment regimens We have been criticized for using nonstandard HBO and NBO therapy46,87,99 because this is not representative of usual practice. The treatment regimens in our study were not conventional. It should be noted however, that established practice continues to vary very widely.4 In the absence of universally accepted recommendations for depth or duration of pressurization for HBO treatments, our protocol for the management of CO poisoning consisted of three treatments of 100 min with 60 min at 2.8 ATA over 3 days, based on the conclusion of Gorman and Runciman93 that this achieved the lowest mortality and neurological deficits and would provide the maximum potential advantage for HBOT. Our hyperbaric treatment regimen for CO poisoning included interval NBO. To retain blinding, we also provided the control group with NBO for 3 days. As a result, our normobaric group received more NBO than used in previous studies and this may have been a factor in the lack of outcome difference between the treatment groups.105

8417: “8417_c018” — 2007/9/19 — 13:22 — page 411 — #21

412

Carbon Monoxide Poisoning

As Tighe comments,101 “previous studies have been flawed by a failure to optimize treatment in the normobaric group, and Scheinkestel et al’s study clearly shows that such cheap, available, safe treatment is also effective.” Moon99 raised the concept that repetitive treatments at 2.8 ATA could have been neurotoxic and offset any potential benefit from HBO. This proposal of oxygen neurotoxicity is a valid one, but it is important to note that repeated daily treatments at 2.8 ATA is standard therapy for CO poisoning in many centers and also for cerebrally-impaired patients with decompression illness. We set out to demonstrate an advantage for HBO and selected the optimum treatment regime to achieve this. We have now demonstrated that there is no advantage in outcome for HBO over 3 days of high flow oxygen therapy. What we cannot now answer is whether the same outcomes can be achieved with less days of high flow oxygen therapy. Piantadosi46 states that the difference in O2 doses between the groups was negligible. We calculated that the HBO group received oxygen therapy equating to approximately 35.7 COHb dissociation half-lives, while the NBO group received the equivalent of 28.5 COHb dissociation half-lives, a difference of 7.2 COHb dissociation half-lives. Most other studies have used total oxygen doses of less than 7.0 COHb dissociation half-lives.74,75,80,82 The difference in our two groups is greater than the difference in doses used in these other studies. We just used a higher baseline oxygen dose. 18.2.3.7 “Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning” by Weaver Weaver81 in 2002 reported on the outcome of 152 patients enrolled within 24 h of exposure to CO in his double-blind randomized trial. Seventy-six patients were assigned to have three hyperbaric sessions within a 24-h period and 76 had one NBO treatment and two sessions of exposure to normobaric room air. Patients were stratified according to whether or not they had lost consciousness, the interval between end of exposure and entry into the chamber (6 h) and age (40 years). Treatments took place in a Sechrist monoplace chamber. The first HBO session was with 100% O2 for 2.5 h with 55 min at 3.0 ATA, the second and third were with 100% O2 for 2 h with 90 min at 2.0 ATA. It is not clear from the paper as to the exact NBO treatment given. In one section it states: “patients in the NBO group were exposed to air at one ATA for all three chamber sessions.” In another section it states: “100% oxygen was delivered to those in the NBO group during chamber session 1. During NBO sessions 2 and 3, patients were exposed to 1 ATA and breathed air” unless the patients were intubated and ventilated or had arterial oxygen saturations 25%, age > 50 and metabolic acidosis. In patients with none of these four criteria, HBO did not improve outcome. While proponents of HBO for CO-poisoning consider this to be a landmark paper, and Dr Weaver believes his paper to be the only one worth considering and the definitive paper on HBO in CO-poisoning,78 we believe there are some fundamental problems with the Weaver study: •

Primary Outcome analysis

Dr. Juurlink summarizes this eloquently.78 “In 1995, Weaver and colleagues published the first interim analysis of their study.106 In that report, the only test of statistical significance was applied to DNS, and the investigators indicated that enrollment would continue because the p value had not achieved the threshold required for premature termination of the trial. That same year,107 Weaver presented a description of his ongoing trial and gave an explicit definition of its primary outcome: “Our major question is, does HBO2 reduce the incidence of DNS?” He also wrote: “During the course of the trial, it became evident that operational definitions of DNS and PNS were needed . . . Our definition for DNS is: development of a new neurologic abnormality not present at day 1, and/or decrement of neuropsychologic subtest score of more than two SDs below the mean or two subtest scores more than one SD below the mean compared to standardized norms (prior normal neuropsychologic test). If the prior neuropsychologic test is abnormal, then we use a decrement of an abnormal subtest of more than one SD compared to the prior score or more than 0.5 SDs below each of at least two abnormal subtests.” In his letter to the editor,107 written with the study underway and 47 patients enrolled, neither the definition of DNS nor that of PNS includes patient symptoms. They are both clearly defined and these definitions are solely based on the outcome of neuropsychological tests.” DNS appears to be the original intended outcome of the trial. This outcome has never been subsequently reported in any forum. “In 2001, a discussion of his soon-to-be published study89 contained no mention of DNS. Indeed, a very different outcome (“cognitive sequelae”) had been defined: “Cognitive sequelae were considered present if any 6-week neuropsychological subtest score was >2 standard deviations below the mean (or if at least two

8417: “8417_c018” — 2007/9/19 — 13:22 — page 413 — #23

Carbon Monoxide Poisoning

414

subtest scores were each more than one SD below the mean) of demographically corrected standardized scores. Cognitive sequelae were present if a neuropsychological subtest score was >1 SD below the mean or if two subtest scores each were >0.5 SD below the mean and the patient complained of memory, attention, and/or concentration difficulties.” In the final publication,81 the cut-off values used to define abnormal results changed yet again. It is not clear why the statistical analysis on the T-scores for the neuropsychological subtests is performed on the aggregate of the three scores obtained after chamber session 3, at 2 weeks and at 6 weeks, given that “the primary outcome was cognitive sequelae at 6 weeks”. The other two scores (after chamber session 3 and at 2 weeks) are not relevant to this end-point.” In summary, the original intended outcome of this trial was DNS defined by rigorous clinical criteria. In contrast, the final publication reports “cognitive sequelae,” a distinctly different outcome. PNS and DNS have been bundled together. The definition of a poor outcome was changed to include a subjective component of patient symptoms and the results of neuropsychological tests after treatment, at 2 weeks and at 6 weeks have also been bundled together. Juurlink states: “Dr. Weaver and his coinvestigators have obviously collected the data necessary to examine DNS as an outcome, and we urge them to present this analysis.73 Juurlink states, “doing so would help settle the present debate. While HBO enthusiasts may argue that ‘cognitive sequelae’ is a meaningful outcome, skeptics may legitimately wonder if the revised outcome was simply that which cast the most favorable light on HBO once all the data were collected.” Buckley108 asserts that a significant difference between HBO and NBO would almost certainly not have been demonstrated if the originally intended outcome had been analyzed. Other problems identified with Weaver’s study include: •

Early termination of the study

In 1999, Weaver wrote109 “Blinded interim analysis showed no difference in outcome between the two groups after 50 and 100 patients. Yet after 52 additional patients, the results were so clear-cut, that the trial was terminated early, “after the third of four scheduled interim analyses”. We interpret this to mean that there was one more interim analysis scheduled and presumably 100 more patients still to enroll. When Weaver adjusted for differences in baseline severity (cerebellar dysfunction—see section on failure of randomization), the difference between groups only just makes statistical significance at p = 0.05, well below the predetermined “stopping rules.” •

Enrollment and follow-up

Only 33% of potential patients were entered: 28% were excluded and 39% refused consent, thus introducing the possibility of significant selection bias. The 95% followup rate of these highly selected patients resulted in only 31% of all potential patients being assessed. Weaver argues110 that he has followed up 91 of the 180 who declined to be in the trial. They had a lower suicide rate (16% versus 30%) and a lower COHb (20% versus 25%). Weaver concludes that they were similar to those enrolled in the

8417: “8417_c018” — 2007/9/19 — 13:22 — page 414 — #24

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

415

study. Further he states that the incidence of 6 week cognitive sequelae of 43% was similar to that of the NBO group in the trial (46.1%), but it is clear from his reasons for “declined enrollment in the trial” group,81 that at least some of these declined because the referring physician insisted on HBO. •

Failure of randomization

This is the only study to have a large difference in baseline severity. Despite randomization, Weaver et al.’s normobaric group appeared to have suffered more severe CO poisoning. The normobaric group had a CO exposure of 22 ± 64 h, almost double that of the hyperbaric group (13 ± 41 h), and the COHb saturation at first entry to the chamber was significantly higher (p = 0.02). Furthermore, the incidence of pretreatment cerebellar dysfunction was 15% in the normobaric group and only 4% in the hyperbaric group, an almost fourfold difference that was also statistically different (p = 0.03). The presence of cerebellar dysfunction before treatment was associated with cognitive sequelae (p = 0.005). Shorter exposure to CO, a lower level of COHb, and a lower incidence of cerebellar dysfunction would be expected to favor a better outcome in the hyperbaric group. It is also worth noting that the COHb was actually only available in 83 patients (in 36 HBO, less than half of the 76 HBO patients enrolled, and in 47 NBO). In the other 69, missing values were imputed on the basis of the data in the other 83 patients. The duration of exposure to CO has been presented as a mean with a standard deviation. From these results (13 ± 41 versus 22 ± 64) it is clear that the distribution of duration is skewed. Variables measuring duration are often well characterized by a log-normal distribution and would have been more appropriately presented as geometric means with 95% confidence intervals, or if the distribution was unknown (nonparametric), as medians with interquartile ranges. It is likely that there are significant outliers included and this may have had a detrimental effect on the significance of baseline differences and therefore on the multivariate analysis as a whole. Weaver argues that the CO exposure was not statistically significant and produces the requested statistics:110 • Median exposures: 4.2 (range 0.2–308 h) for HBO versus 5.0 (0.3–397 h) for NBO • Geometric means: 4.0 h (95% CI = 2.9–5.4) for HBO versus 5.4 (95% CI = 3.8–7.6) for NBO, p = 0.2 • Interquartile range 7.3 h (1.6–8.9) for HBO versus 9.5 (2.0–13.5) h for NBO, p = 0.2 These all confirm a trend, albeit non-significant, towards increased exposure in the NBO group, which is confirmed by the statistically, significantly higher COHb in the NBO group and is in keeping with the increased cerebellar signs in the NBO group. Weaver argues however that the difference in COHb is of no consequence as there is no relationship between COHb and outcome.110 This still leaves the fourfold increase in prechamber cerebellar signs in the NBO group compared to the HBO group. If these patients are not excluded, his intention to treat analysis finds a highly significant difference in six weeks outcome (25.0% versus 46.1%, p = 0.007) in favor of HBO.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 415 — #25

Carbon Monoxide Poisoning

416

If only patients with normal cerebellar function are analyzed, the benefit for HBO only just reaches statistical significance (23.2% versus 39.0%, p = 0.05) and the trial would not we assume have been terminated early. • Inadequate oxygen dose to normobaric group The amount of oxygen therapy given to the normobaric group may have been suboptimal. In all studies, patients received variable amounts of NBO therapy prior to hospital arrival. Current recommendations for in-hospital NBO administration most commonly involve occlusive face-mask and reservoir bag for 6–12 h or until symptoms resolve. In this study, unless the patient’s oxygen saturation was less than 90%, or they were intubated, the normobaric group only received 135 min of oxygen therapy with a reservoir and a nonocclusive face-mask while in the chamber. Thus in-hospital oxygen treatment was, by conventional criteria, short. The four other randomized studies that have been published all utilized longer duration of oxygen therapy in the normobaric group: 6–12 h80 6 h75 12 h,82 and an average of 4.2 h.76 Weaver argues110 that the sum of oxygen therapy for the NBO group was 6.9 h, 135 min in the chamber and 4.5 ± 2.2 h before chamber session 1. The range was 3.9–18.8 h. However, he is including the prehospital time, whereas all other studies provide at least 6 h once at the hospital. The median duration of oxygen therapy in Weaver’s NBO group was 6.2 h. By definition 50% of patients had less than 6.2 h of oxygen therapy. Further, Weaver states109 that referring physicians gave NBO therapy to CO patients, but there is no statement to guarantee that these patients were given high flow oxygen from a reservoir through a non-rebreathing face mask prior to arrival at LDS Hospital. Conventional teaching is that hyperoxia is required in CO poisoning to help “offload” CO. Not providing oxygen therapy to patients whose oxygen saturations were >90% would be considered inappropriate by most centers. Weaver argues110 that they treated all CO-poisoned patients with NBO until the COHb was less than 5%. This is not stated in the paper. In fact, the paper is quite clear in several sections, that supplemental oxygen was only provided if the oxygen saturation was 90%. If this was done using pulse oximetry, as is the norm, then there is considerable literature as to the inadequacy of pulse oximetry to monitor oxygen saturation as this cannot differentiate between COHb and oxyhemoglobin, and therefore over-estimates oxyhaemoglobin.111 •

Non-conventional HBO regime

Kao raised concerns that Weaver used a nonstandard HBO protocol.87 The first HBO session was with 100% O2 for 2.5 h with 55 min at 3.0 ATA, the second and third were with 100% O2 for 2 h with 90 min at 2.0 ATA. These are not the normally used treatments and are not in keeping with The HBOT Committee Report 2003,112 which states as follows: “the optimal number of hyperbaric treatments, the time following poisoning after which therapy is no longer effective and the optimal treatment pressure

8417: “8417_c018” — 2007/9/19 — 13:22 — page 416 — #26

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

417

will require additional study . . . all patients at high risk deserve a single treatment . . . subsequent treatments may be performed within 6–8 h and continued once or twice daily until there is no further improvement . . . the optimal dose cannot be clearly stated . . . between 2.5 and 3.0 ATM seems appropriate.” Weaver’s HBO regime was less conventional than Scheinkestel’s. Weaver only gave three treatments, 18% of the HBO group did not complete the three treatments and the treatments were at lower ATA than recommended in the HBO Committee Report. This compares with Scheinkestel’s treatment regime that gave a further three treatments (total of six) if abnormalities persisted and all were at the recommended ATA. •

Unjustified assumptions made in interpreting six week data

The intention to treat analyzes used by Weaver as “patients with missing data for neuropsychological tests at six weeks were assumed to have cognitive sequelae.” This is contrary to the standard intention to treat (ITT) approach of carrying the last observation forward. Further, only one patient was lost to follow-up in the HBO group whereas four were lost in the NBO group, thus this assumption favored the HBO group.92 The impact of such arbitrary definitions on outcomes can be seen in Weaver’s 2002 abstract113 where he took the opposite approach and if data were missing, the outcome was deemed to be “favorable.” Data from the 6 and 12 month follow-up are apparently combined. Weaver states “a favorable outcome was found in 62/76 (82%) of HBO patients compared with 50/76 (66% treated with NBO (p = 0.027). If data from patients with unknown 6 and 12 month outcome data were excluded, a favorable outcome was present in 49/58 (84%) treated with HBO compared to 42/60 (70%) treated with NBO (p = 0.061).” •

Unjustified assumptions made in interpreting 6 and 12 month data

In the final publication,81 the definitions change again. With respect to the outcomes at 6 and 12 months, these were performed on the basis that ‘if patients had cognitive sequelae at 6 weeks, and missing data at 6 or 12 months, they were assumed to have cognitive sequelae at those times. This is invalid, as it does not allow for the improvement with time, which was demonstrated in patients with complete data. Such a definition couples the first outcome to later events and creates a spurious outcome dependent on the first. The statistical differences reported at 6 and 12 months merely reflect the results at 6 weeks, not necessarily the true longterm outcome. If the analysis is restricted to those patients with complete data, the statistically significant difference in late outcome is lost. Dr Weaver disputes110 that statistical significance is lost at 12 months but a p = 0.08 is not statistically significant. •

Soft outcome measures

The entire positive outcome of this study is based on reported symptoms. These were the primary determinant of a statistical difference between treatments. In the final publication,81 neuropsychological testing identified no difference between HBO and NBO; indeed, the mean neuropsychological testing scores for patients treated with NBO were within the normal range.78,92

8417: “8417_c018” — 2007/9/19 — 13:22 — page 417 — #27

Carbon Monoxide Poisoning

418

HBOT patients reported fewer difficulties with memory (p = 0.004) and this was the only significant symptom difference between the two groups. However, there were no statistically significant differences between groups on any of the memory-related neuropsychology test scores. •

Interpretation of neuropsychological tests

Disproportionate numbers of patients with cerebellar problems entered one arm of the Weaver study. Neurological sequelae occurred more commonly if there were cerebellar signs at the time of enrollment into the study. This was particularly so because two of the six neuropsychiatric tests involved “Trail-Making,” and this would be affected by even minor degrees of cerebellar dysfunction.78 This imbalance alone could have accounted for half the actual observed difference between groups because the absolute difference between arms was 16 individuals, yet there were 8 more individuals with cerebellar dysfunction and neurological sequelae in the NBO group.92 Olsen is more critical114 and states: “The neuropsychological data presented by Weaver are clinically underwhelming. The raw scores show a statistically significant difference between treatment groups in only one of six subtests (Trail-Making, Part 1), and even in this subtest, the normobaric group was at the mean demographically corrected score for a normal population at 6-week follow-up.” Buckley also comments that the mean performance of patients in the NBO group was normal for five of the six neuropsychiatric tests and above normal in the sixth.92,115 He questions how a meaningful outcome could label 46% of patients in the control group as having “cognitive sequelae,” when in fact, five of six of the mean test scores in that group were actually normal or above average. Weaver 110 responds that the neuropsychological test scores of patients with dysfunction are obscured by those without sequelae and that therefore the group mean scores do not detect a difference between groups, but no data are provided to support this. It is also of interest to note that the frequency of cognitive sequelae amongst patients who completed three HBO sessions was not different from those who did not complete the three sessions, and 18.4% of the hyperbaric group did not complete the required number of chamber sessions. Weaver et al.’s outcome measures were performed by any of ten different psychologists. Weaver states that the examiners were all psychology Ph.D. candidates with proper training, with inter-rater reliability being well established for these tests at 0.9 or higher. He also states that as the trial took 6 years, this necessitated multiple examiners. • Interpretation of 6 and 12 month data Weaver claims115 that “Cognitive sequelae at six months and 12 months were less frequent in the HBO group than in the NBO group, both according to the intentionto-treat analysis (p = 0.02 at 6 months, p = 0.04 at 12 months) and according to the efficacy analysis, (p = 0.03 at 6 months, p = 0.08 at 12 months).” Interestingly, in his presentation at the Undersea and Hyperbaric Medicine Society Annual Scientific Meeting July, 2002, and presented in abstract form, Weaver states “if data from patients with unknown 6 and 12 month data were excluded, a “favorable

8417: “8417_c018” — 2007/9/19 — 13:22 — page 418 — #28

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

419

outcome” was present in 49/59 (84%) treated with HBO compared to 42/60 (70%) treated with NBO (p = 0.061).” • Lack of pretreatment neuropsychological assessment Weaver’s study performed no baseline neuropsychological assessment. As Raphael states, baseline information on abnormal cognitive tests is not provided. Given the heterogeneity of the population and the rather small sample size, one cannot rule out an imbalance between the treatment groups with respect to abnormal results of cognitive tests just as there was an imbalance with respect to cerebellar signs.116 •

Outcomes of other tests

There were no statistically significant differences between groups on the Geriatric Depression Scale, the Katz index of activities of daily living, nor in scores on the subscales of the SF 36 (social function, physical role, mental health, and energy). Weaver argues110 that these tests do not test cognition and confirm that the cognitive impairment is not due to depression. He also states that the “activities of daily Living” measures gross abilities to perform daily activities not cognition. He also states “the SF36 measures health related quality of life. It does not measure cognition but rather the patients’ perception as to whether the CO poisoning resulted in decreased quality of life for physical and psychological functioning.” However, in regard to the latter, his paper states: “We found no treatment-related differences in scores on the subscales (social function, physical role, mental health and energy) of the SF36. Clearly then, the conclusion is that the patients did not have the perception that the CO-poisoning results in decreased quality of life for physical and psychological functioning. • Adverse events Weaver reports a significant difference in the incidence of nystagmus post-treatment, with hyperbaric patients having a 12% incidence compared to 2.7% for the normobaric group (p = 0.05). The reason for this adverse effect in the hyperbaric group is not clear. It is also worth noting that 18.4% of the hyperbaric group did not complete the required number of chamber sessions for reasons including anxiety and middle ear problems. In an abstract presented to the ASM of UHMS117 in 2001, Weaver states that the NBO group tolerated chamber therapy better (96% versus 82%, p = 0.002). While these adverse events are not major, they must be taken in the context of the degree of benefit obtained from the treatment. •

Cost analysis

In his 1995 letter to the editor of the Annals of Emergency Medicine,107 Weaver states that other questions his trial may answer include differences between the two therapies (HBO and NBO) related to cost (including transport). This analysis is yet to be seen. Other criticisms include the following: • Small number of intubated patients (12) prevents interpretation of what HBO may or may not have to offer seriously injured patients.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 419 — #29

Carbon Monoxide Poisoning

420

• Having stratified for time to treatment less than or greater than 6 h, Weaver does not report if there was a difference between these two groups, yet he is so critical of those studies where there was a delay of >6 h. • Weaver’s subsequent analysis concluded that: “HBO improved outcome if any of the following: LOC, COHb >25%, age > 50 and metabolic acidosis were present. In patients with none of these four criteria, HBO did not improve outcome. This conclusion uses an absolute value of COHb at the time this was first sampled and the recommended selection criteria could therefore exclude patients with significant exposure but delay in COHb measurement. It is worth noting that the actual COHb levels were only available in 55% of his patients. This statement is in contradiction with the literature and Weaver’s previous and subsequent declaration that “the difference in COHb between the groups is of no consequence as there is no relationship between COHb and outcome.110 Further it is difficult to understand the choice of age > 50, when his study stratified for age > 40 years and did not present the data for age greater than or less than 40. • The outcome should not be derived, as Weaver’s was, from complex interpretations of pooled differences in test scores, especially when those tests are not routinely conducted in clinical practice.115 Buckley92 concludes overall that the unbalanced recruitment, changed primary outcome, and the intention to treat (ITT) assumption when considered with the very small number of patients (16) resulting in the finding in favor of HBO, could easily mean that the trial outcome would have changed from significant to nonsignificant had these factors been otherwise. Given all the above, Weaver et al.’s conclusion of benefit arising of HBOT is not convincing. The benefit of HBOT demonstrated in this study, if there is one, may not be clinically significant.

18.3 CONCLUSIONS There are significant difficulties with comparing the outcomes of these investigations. Variations in study design, HBO and NBO protocols used, outcomes measured and patient populations included, make it difficult to draw firm conclusions.87 In addition, many of the studies show bias towards use of HBO in the more severely affected patients, follow-up is incomplete and overall, the numbers of patients studied is low. No reliable method to identify patients at high risk for neurologic sequelae has been identified. The efficacy of one HBO protocol over another has not been determined. Timing of evaluation (discharge, 1 month, 6 weeks, 1 year) has also not been determined. The ongoing debate about the efficacy of HBO is driven largely by these discrepant results.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 420 — #30

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

421

The best, most sophisticated, unbiased and comprehensive analysis of published human trial work to date, without conflicts of interest, and no pre-existing financial or reputational bias either pro or against HBO, was performed by Buckley92 who concluded that “The role of HBO remains unclear and the weight of the available evidence neither confirms nor refutes a clinically meaningful net benefit.” Buckley assessed the effectiveness of HBO compared with NBO for the prevention of neurological sequelae in patients with CO poisoning. Eight randomized controlled trials were identified. Two had no evaluable data and were excluded. A pooled OR for the presence of neurological symptoms at 1-month follow-up was calculated. At 1 month follow-up after treatment, sequelae possibly related to CO poisoning were present in 242 of 761 patients (36.1%) treated with NBO compared with 259 of 718 patients (31.8%) treated with HBO. The OR for neuropsychiatric symptoms with HBO was 0.77 (95% CI = 0.51, 1.14). A further analysis was performed by removing the results of trials with a Jadad score of less than 3/5 (the two trials published in abstract form only), leaving a total of 146 of 383 patients (38%) randomized to HBO with neurological sequelae at one month compared with 152 of 368 patients (41%) randomized to NBO [OR for neurological sequelae 0.70 (95% CI = 0.34, 1.47)]. When restricting the analysis to the two studies that enrolled more severely poisoned patients (the only two with a Jadad score of 5/5) the results remained inconclusive [OR for neurological sequelae 0.73 (95% CI = 0.22, 2.48)]. Buckley states that there were methodological shortcomings in all trials and empiric evidence of bias in some, particularly those suggesting benefit from HBO. The trials enrolled patients with CO poisoning of varying severity, employed different regimens of HBO and NBO. Only two were conducted with double-blinding through sham treatment. Pooled analysis of such inconsistent studies should be interpreted with caution. In the following issue of Toxicological Reviews, four knowledgeable clinical toxicologists from different parts of the world were asked to provide editorials on the subject. Brent, the editor, comments: “Given six relevant randomized clinical trials involving 1479 patients, if the effect of HBO were real and large, it is difficult to imagine that the trials would not be more definitive.118 Even the positive ones can be interpreted as having only marginal benefit.” “There are clearly insufficient data to consider HBO as a standard of care and it should be considered to be a therapy of as yet unproven benefit. Further, any benefit deriving from HBO is likely to be small. Thus, no physician or poison center should be held liable for withholding HBO given the uncertainty and even possibility of harm associated with this treatment.” Henry119 states: “Most of us would probably choose HBO because we know there is no real evidence of harm and “because it might do some good.” However, the benefit is not likely to be great.” Olsen114 concludes: “There is no recognized standard of care mandating the use of HBO.” Bentur120 was the most positive in favor of HBO, stating “it is impossible to state that HBOT should not be offered.” He goes on to provide an algorithm for treatment of CO poisoning based on the results of Weaver’s and Scheinkestel’s studies.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 421 — #31

Carbon Monoxide Poisoning

422

Seger121 states: “believers believe that the studies that support HBO have fewer limitations than the studies that do not support HBO (the believers are very fervent, and tend to lump the atheists and agnostics together and suggest both be put to the sword). The more sceptical nonbelievers feel that the studies demonstrating no benefit have fewer limitations. The camps are divided,” with many in each camp seeming to have an investment in maintaining their belief system, regardless of the data. Lastly, there is in fact a downside to HBO. There is a cost both for providing the service and transport to the facility. This has not to-date been quantified. Transports of critically ill patients are associated with risk and while this has not specifically been quantified for CO-poisoning, there is considerable evidence for this in critically-ill patients. Further, there is documented morbidity to patients due to the treatment in a hyperbaric chamber. A report by the Department of Health and Human Services, Office of Inspector General, June Gibbs Brown, in October 2000 on HBOT: Its Use and Appropriateness,122 states: “According to our review, 18% of beneficiaries exhibit side effects (significantly greater than the literature suggests). The most common side effect is ear-related trauma, representing 63% of all observed side effects. While side effects are generally not severe, two individuals within our sample showed signs of oxygen toxicity. This relates to 1.3% of the population which also is significantly greater than the expected value cited in the literature. These statistics were based on our analysis of the 1998 National Claims History file maintained by Health Care Financing Administration (HCFA).”

18.4 WHERE TO NOW? While independent reviewers have, predictably, called for large, multi-center randomized, controlled and blinded, clinical trials, many in the hyperbaric community appear to believe this is either unnecessary, unethical or impractical, if not all three. Plans are therefore emerging for further randomized studies comparing different doses of HBO, selected from existing regimens and a pilot has been conducted.123 The optimal dose of HBO is certainly an important question for all indications for HBO, but it could be argued that at the present time we know too little about HBO for CO to select either the optimal doses or the optimal patient eligibility criteria for an optimal study. It will be apparent from the preceding that we agree that it would be absolutely necessary for further randomized trials to provide proof of benefit before HBO could be generally and unconditionally recommended for CO poisoning. We also believe that it would not in any way be unethical to randomize patients, given the lack of clear-cut evidence for the superiority of any one form of treatment over any other. It is not clear to us, however, that another randomized controlled trial is the appropriate next step. Large multicenter studies carry a large financial cost however and sometimes there can be a significant opportunity cost as well. If one study consumes the limited numbers of eligible patients, capable centers and research funding available for several years, this can be to the detriment of potentially better studies conceived after

8417: “8417_c018” — 2007/9/19 — 13:22 — page 422 — #32

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

423

new basic science or epidemiologically generated information becomes available. It is therefore most important that any future studies have an optimal design and conduct, and are designed to answer a reasonable hypothesis. Herein arises a significant issue. Simple and broad hypotheses applied to heterogeneous populations with many confounding variables require very large studies to ensure matching of active and control groups, to ensure that relatively small but significant outcome improvements are detected and to allow for the subgroup analyses necessary to identify response in important subpopulations. Another factor that mandates large sample sizes is low incidence of the outcome measure of importance and this applies to the CO poisoning outcomes of death after rescue and development of delayed onset neurological sequelae (DNS). As an example, a highly important 50% reduction in mortality or DNS rate from, say, 10% to 5% would require a two armed randomized study sample size of nearly 1000 subjects to provide 80% likelihood of detection at p = 0.05 (2-tailed T test). The alternative trial design strategy is to utilize tight enrollment criteria to identify a specific subgroup for a test of therapy thought to provide significant benefit, based upon pilot data or coherent and generalizable large animal research. This could allow for smaller study sizes if the therapeutic effect is powerful and the outcome measure is relatively close to evenly distributed in the control group. Caution is subsequently needed in generalizing the results of any such study to different populations, doses or treatment timings. Patient selection for any therapy is clearly important and this could be a critical issue in the case of CO poisoning. To date, recommendations regarding the optimal population for treatment with HBO have generally used combinations of age, COHb level, history of LOC and presenting signs and symptoms. Consideration of age and indicators of cardiovascular or cerebral disease do seem very logical given the way that cerebral or cardiac hypoxia can exacerbate the severity of clinical poisoning that results from any given COHb level. It is not clear from the studies published to date whether such criteria are valid. While outcome for patients who have suffered CO-poisoning-related cardiac arrest is universally bad,124 there may or may not be a definable larger group of patients so severely brain injured as to have no reasonable prospect of recovery. It is clear that long-term harm can occur with and without LOC and recovery can follow severe as well as mild poisoning. It is not yet clear, however, whether there is a minimum threshold for injury and most significantly, whether there is better response to HBO amongst any particular CO poisoning population. In attempting to generate potential clinical study designs, a number of relevant hypotheses can usefully be generated from the existing human and animal data and from basic principles: • The value of HBO may vary significantly with the stage of injury. • The optimal dose of HBO may vary with stage of injury. • Optimal dose, especially in very early stage treatment, may be variable with degree of poisoning and this would indicate the need for therapy tailored to some measurable variable. • HBO and even NBO might be harmful at certain stages of the secondary injury process, perhaps only in susceptible individuals, by increasing

8417: “8417_c018” — 2007/9/19 — 13:22 — page 423 — #33

Carbon Monoxide Poisoning

424

•

• • • • • • •

lipid peroxidation or other manifestations of oxidative injury rather than reducing them. Unconsciousness or gross neurological deficit on admission is correlated with poorer outcomes and this is usually taken to be an indicator of severity of brain injury but this may not mean that this is the most treatment-responsive group. If one could measure and allow for the effects of residual intracellular CO and/or cellular hypoxia this seems likely to correlate with injury severity, but again might not select for the most treatment responsive group. LOC may be possible without harm, provided that cerebral oxygen delivery is maintained by greatly increased cerebral blood flow, and this might be a group of patients who can recover without any active therapy. Overall outcomes and response to therapy may be different depending upon the presence or absence of cotoxins such as volatile hydrocarbons, cyanides and other products of combustion. There may be significant inter-individual variability in susceptibility to CO mediated secondary cerebral injury and in the degree of recovery possible. There may be a significant inter-individual variability in response to HBO. Such variabilities may have both genetically determined and acquired elements. If HBO does accelerate recovery and reduce hospital stay, this may be sufficiently valuable to be worthwhile even if there is no net change in longterm outcome. If so, therapy could be most valuable to populations selected for characteristics other than susceptibility, for instance working age patients, parents, and carers or the psychiatrically disturbed.

The size and complexity of the studies likely to be needed and the extent of the above and, no doubt, other unknowns would suggest that the time is not yet right for large and costly human randomized controlled trials. Given that the provision of HBO for CO poisoning is established practice in some centers, it is not unreasonable for these centers to undertake studies comparing different but commonly used doses of HBO provided adequately powered, well governed, collaborative studies can be achieved at moderate cost, with study designs that address the limitations of previous work. Meanwhile, a huge natural experiment is continuing, unfortunately with little analyzable data being collected; CO poisoning is common and patients currently receives a wide range of different oxygen doses in both the normobaric and hyperbaric treatment environment. A well-designed clinical registry has the potential to generate specific hypotheses for testing with clinical trials and even to answer many questions outright. Meaningful data must be collected, however, and this will require agreement on markers of poisoning severity and outcome measures. It would also be much more valuable if any registry drew data from centers that do not use HBO as well as from the hyperbaric community’s patients. In addition to enabling comparison of outcomes between patients receiving NBO and HBO, the optimal initial duration of NBO for different poisoning severity is unknown, as is the place of increased inspired oxygen fraction during the postpoisoning, secondary brain injury phase.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 424 — #34

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

425

Animal and cell based research will continue to produce information that should be taken into consideration and of particular note here is the field of CO therapeutics. In recent years, identification of the role of CO as an intracellular signalling molecule and element of the hemoxygenase stress response system has led to two important conclusions: some elements of the pathology of CO poisoning may be related to disturbances of the physiological role of CO45,125,126 and exogenous CO might be a useful therapeutic substance.9,127−139 . Both of these ideas have implications for CO poisoning therapeutics. Exogenous CO can inhibit ischemia-reperfusion injury and appears to have neuro-protective, cardio-protective and lung protective properties in some animal models and this has already led to some human clinical trials. The doses used in these trials have been as high as 500 ppm,140 an order of magnitude above current occupational health limits and a level that has been associated with pathological outcomes, a fact which has been pointed out in critical commentaries.11,12 Nevertheless, the potential for exogenous CO to modulate immune and oxidative stress-related processes raises the question as to whether residual CO may have some protective effects after the bulk of excessive CO is eliminated following CO poisoning. It is clear that both CO and HBO biochemistry are in a state of evolution that should be closely monitored by toxicologists and clinical researchers. Work in other types of neurological injury is also likely to yield important information, both with regard to the potential for HBO use in the later and delayed onset stages of CO-related injury and with respect to alternative therapies to minimize CO related brain injury. Strategies which show promise for diffuse and hypoxic brain injury, such as moderate hypothermia and various anti-oxidant drugs,141,142 could well provide benefit for CO poisoning patients. If alternative neuroprotective strategies show significant benefit for CO, it would be appropriate to compare these with HBO and to test combined therapy. CO-related cardiac injury is also be a field needing further investigation. Recent work indicates that measurable injury is far more common than previously considered143 and HBO may moderate this, although this needs investigation. Cardiac injury might also prove a useful and more easily measured outcome than neurotoxicity for research aiming to select optimal dosing of oxygen and might even prove to be a usable surrogate enabling early assessment of overall severity of injury. Given the lack of predictive power of clinical signs and current biochemical markers, new markers of neurological injury severity would be useful if they were shown to correlate with either outcome or response to therapy. Cleaved tau, nonspecific enolase, MBP and S100B57,144−146 are amongst markers being considered. Serum could be readily collected from CO poisoning patients to produce a database which could yield useful information, provided good follow up and outcome measures are available to enable correlations to be sought. Medical imaging is advancing rapidly and various markers of severity have been proposed based upon Computed Tomography (CT), Magnetic Resonance Imaging (MRI) and Tc-Hexamethylpropyleneamine Oxime (HMPAO) Single Photon Emission Computed Tomography scanning (SPECT).147−157 To date none have become established as reliable determinants of response to treatment and any such markers would need to be readily available with rapid turn-around if they are to be

8417: “8417_c018” — 2007/9/19 — 13:22 — page 425 — #35

Carbon Monoxide Poisoning

426

useful in deciding therapeutic directions. Quantitative EEG monitoring has also been tested as a marker of response, no doubt, will other measures.158 A final critical factor is that it seems likely that there are variations in innate susceptibility or resistance to injury and responsiveness or resistance to therapy for CO poisoning, as is increasingly seen as important in many other areas. Gender is almost certainly significant, as is the case in trauma and stroke.159−161 As this may be based upon antioxidant protective mechanisms of female hormones, it is conceivable that response to exogenous oxygen therapy could be different and any future research should look for potential gender differences. Recently, Hopkins162 has reported in abstract that ApoE typing is a significant predictor of outcome in patients followed-up at his institution. The ApoE e4 allele, associated with poor outcomes in many other forms of degenerative and acute brain conditions, was associated with higher incidence of cognitive sequelae after CO poisoning and benefit from HBO appeared to be confined to patients with ApoE4—in those without the e4 allele there was no difference in outcome between HBO and non HBO treated groups. At present this measure is not available as a point of care test but this and other markers, including those drawn from the fields of genomics and proteomics may hold the future of case selection and progress monitoring for specific treatments such as HBO. Despite over four decades of hope and persistence, HBO has, as yet, failed to live up to it’s promises as a valuable therapy for CO poisoning. While there are sufficient indications of potential benefit to support ongoing research, there is insufficient evidence to support promotion of HBO use outside of clinical trials or for the development of new HBO facilities specifically for CO poisoning treatment. Given an ongoing problem with CO poisoning and an increasing availability of high standard hyperbaric facilities incorporated into major hospitals, the infrastructure for next generation clinical trials exists. Hopefully lessons will be taken from trials to date in developing future trials but it should not be assumed that existing protocols are optimal and an open mind should be kept to alternatives. Meanwhile, those facilities treating CO poisoning regularly will hopefully contribute data to registries which will also gather data on patients treated with NBO, which continues to be the standard of care, albeit without any good data regarding the optimal duration of therapy.

References 1. End E, Long CW. Oxygen under pressure in carbon monoxide poisoning. J. Ind. Hyg. Toxicol. 1942; 24: 302–306. 2. Pace N, Strajman E, Walker EL. Acceleration of carbon monoxide elimination in man by high pressure oxygen. Science 1950; 111(2894): 652–654. 3. Hampson NB. Emergency department visits for carbon monoxide poisoning in the Pacific Northwest. J. Emerg. Med. 1998; 16(5): 695–698. 4. Hampson NB. Trends in the incidence of carbon monoxide poisoning in the United States. Am. J. Emerg. Med. 2005; 23(7): 838–841. 5. Brennan C, Routley V, Ozanne-Smith J. Motor vehicle exhaust gas suicide in Victoria, Australia 1998–2002. Crisis 2006; 27(3): 119–124.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 426 — #36

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

427

6. Morrell S, Page AN, Taylor RJ. The decline in Australian young male suicide. Soc. Sci. Med. 2007; 64(3): 747–754. 7. Ryter SW, Morse D, Choi AM. Carbon monoxide and bilirubin: potential therapies for pulmonary/vascular injury and disease. Am. J. Respir. Cell. Mol. Biol. 2007; 36(2): 175–182. 8. Hill-Kapturczak N, Agarwal A. Carbon monoxide: from silent killer to potential remedy. Am. J. Physiol. Renal. Physiol. 2006; 290(4): F787–788. 9. Hoetzel A, Schmidt R. [Carbon monoxide - poison or potential therapeutic]. Der Anaesthetist 2006; 55(10): 1068–1079. 10. Mannaioni PF, Vannacci A, Masini E. Carbon monoxide: the bad and the good side of the coin, from neuronal death to anti-inflammatory activity. Inflamm. Res. 2006; 55(7): 261–273. 11. Hampson NB, Weaver LK, Piantadosi CA. "Low-level" carbon monoxide administration may carry risk. Am. J. Respir. and Crit. Care Med. 2005; 172(6): 784. 12. Thom SR, Weaver LK, Hampson NB. "Therapeutic" carbon monoxide may be toxic. Am. J. Respir. Crit. Care Med. 2005; 171(11): 1318. 13. Choi AM, Dolinay T. "Therapeutic" carbon monoxide may be a reality soon. Am. J. Respir. Crit. Care Med. 2005; 171(11): 1318–1319. 14. Weaver LK. Operational use and patient care in the monoplace hyperbaric chamber. Resp. Care Clinics of N. Am. 1999; 5(1): 51–92. 15. Hunt TK, Ellison EC, Sen CK. Oxygen: at the foundation of wound healing introduction. World J. Surg. 2004; 28(3): 291–293. 16. Thom SR, Bhopale VM, Velazquez OC, Goldstein LJ, Thom LH, Buerk DG. Stem cell mobilization by hyperbaric oxygen. Am. J. Physiol. Heart. Circ. Physiol. 2006; 290(4): H1378–1386. 17. Buras J. Basic mechanisms of hyperbaric oxygen in the treatment of ischemiareperfusion injury. Internatl. Anesthesiology Clin. 2000; 38(1): 91–109. 18. Buras JA, Stahl GL, Svoboda KK, Reenstra WR. Hyperbaric oxygen downregulates ICAM-1 expression induced by hypoxia and hypoglycemia: the role of NOS. Am. J. Physiol. Cell. Physiol. 2000; 278(2): C292–302. 19. Brown SD, Piantadosi CA. Recovery of energy metabolism in rat brain after carbon monoxide hypoxia. J. Clin. Invest. 1992; 89(2): 666–672. 20. Brown SD, Piantadosi CA. Reversal of carbon monoxide-cytochrome c oxidase binding by hyperbaric oxygen in vivo. Adv. Exper. Med. Biol. 1989; 248: 747–754. 21. Sukoff MH, Ragatz RE. Hyperbaric oxygenation for the treatment of acute cerebral edema. Neurosurgery 1982; 10(1): 29–38. 22. Al-Waili NS, Butler GJ, Beale J, et al. Hyperbaric oxygen in the treatment of patients with cerebral stroke, brain trauma, and neurologic disease. Adv. In. Therapy. 2005; 22(6): 659–678. 23. Veltkamp R, Siebing DA, Sun L, et al. Hyperbaric oxygen reduces blood-brain barrier damage and edema after transient focal cerebral ischemia. Stroke 2005; 36(8): 1679–1683. 24. Yang ZJ, Camporesi C, Yang X, et al. Hyperbaric oxygenation mitigates focal cerebral injury and reduces striatal dopamine release in a rat model of transient middle cerebral artery occlusion. Eur. J. Appl. Physiol. 2002; 87(2): 101–107. 25. Hills BA. A role for oxygen-induced osmosis in hyperbaric oxygen therapy. Med. Hypotheses. 1999; 52(3): 259–263. 26. Freiberger J, Coulombe K, Suliman H, Carraway M, Piantadosi C. Superoxide dismutase responds to hyperoxia in rat hippocampus. Undersea Hyperb. Med. 2004; 31(2): 227–232.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 427 — #37

428

Carbon Monoxide Poisoning 27. Speit G, Dennog C, Eichhorn U, Rothfuss A, Kaina B. Induction of heme oxygenase-1 and adaptive protection against the induction of DNA damage after hyperbaric oxygen treatment. Carcinogenesis 2000; 21(10): 1795–1799. 28. Thom SR, Mendiguren I, Hardy K, et al. Inhibition of human neutrophil beta2integrin-dependent adherence by hyperbaric O2 . Am. J. Physiol. 1997; 272(3 Pt 1): C770–777. 29. Wilson HD, Wilson JR, Fuchs PN. Hyperbaric oxygen treatment decreases inflammation and mechanical hypersensitivity in an animal model of inflammatory pain. Brain Res. 2006; 1098(1): 126–128. 30. Sumen G, Cimsit M, Eroglu L. Hyperbaric oxygen treatment reduces carrageenaninduced acute inflammation in rats. Eur. J. Pharmacol. 2001; 431(2): 265–268. 31. Lou M, Chen Y, Ding M, Eschenfelder CC, Deuschl G. Involvement of the mitochondrial ATP-sensitive potassium channel in the neuroprotective effect of hyperbaric oxygenation after cerebral ischemia. Brain Res. Bull. 2006; 69(2): 109–116. 32. Vlodavsky E, Palzur E, Soustiel JF. Hyperbaric oxygen therapy reduces neuroinflammation and expression of matrix metalloproteinase-9 in the rat model of traumatic brain injury. Neuropath. and Appl. Neurobiol. 2006; 32(1): 40–50. 33. Vlodavsky E, Palzur E, Feinsod M, Soustiel JF. Evaluation of the apoptosisrelated proteins of the BCL-2 family in the traumatic penumbra area of the rat model of cerebral contusion, treated by hyperbaric oxygen therapy: a quantitative immunohistochemical study. Acta. Neuropathologica. 2005; 110(2): 120–126. 34. Ostrowski RP, Colohan AR, Zhang JH. Mechanisms of hyperbaric oxygen-induced neuroprotection in a rat model of subarachnoid hemorrhage. J. Cereb. Blood Flow Metab. 2005; 25(5): 554–571. 35. Yu Y, Matsuyama Y, Yanase M, et al. Effects of hyperbaric oxygen on GDNF expression and apoptosis in spinal cord injury. Neuroreport 2004; 15(15): 2369–2373. 36. Li Y, Zhou C, Calvert JW, Colohan AR, Zhang JH. Multiple effects of hyperbaric oxygen on the expression of HIF-1 alpha and apoptotic genes in a global ischemiahypotension rat model. Exper. Neurol. 2005; 191(1): 198–210. 37. Yin D, Zhang JH. Delayed and multiple hyperbaric oxygen treatments expand therapeutic window in rat focal cerebral ischemic model. Neurocritical care 2005; 2(2): 206–11. 38. Sloan EP, Murphy DG, Hart R, et al. Complications and protocol considerations in carbon monoxide-poisoned patients who require hyperbaric oxygen therapy: report from a ten-year experience. Annals Emerg. Med. 1989; 18(6): 629–634. 39. Plafki C, Peters P, Almeling M, Welslau W, Busch R. Complications and side effects of hyperbaric oxygen therapy. Aviation, Space, Environ. Med. 2000; 71(2): 119–124. 40. Yildiz S, Aktas S, Cimsit M, Ay H, Togrol E. Seizure incidence in 80,000 patient treatments with hyperbaric oxygen. Aviation, Space, Environ. Med. 2004; 75(11): 992–994. 41. Hampson N,Atik D. Central nervous system oxygen toxicity during routine hyperbaric oxygen therapy. Undersea Hyperb. Med. 2003; 30(2): 147–153. 42. Smerz RW. Incidence of oxygen toxicity during the treatment of dysbarism. Undersea Hyperb. Med. 2004; 31(2): 199–202. 43. Hampson NB, Simonson SG, Kramer CC, Piantadosi CA. Central nervous system oxygen toxicity during hyperbaric treatment of patients with carbon monoxide poisoning. Undersea Hyperb. Med. 1996; 23(4): 215–219. 44. Haldane J. The Relation of the Action of Carbonic Oxide to Oxygen Tension. J. Physiol. 1895; 18(3): 201–217.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 428 — #38

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

429

45. Gorman D, Drewry A, Huang YL, Sames C. The clinical toxicology of carbon monoxide. Toxicology 2003; 187(1): 25–38. 46. Piantadosi CA. Carbon monoxide poisoning. Undersea Hyperb. Med. 2004; 31(1): 167–177. 47. Weaver L. Carbon monoxide (CO) poisoning in 12 young, college educated skiiers. Undersea Hyperb. Med. 2006; 33(5): 338. 48. McGrath JJ. Effects of altitude on endogenous carboxyhemoglobin levels. J. Toxicol. Environ. Health 1992; 35(2): 127–133. 49. Kleinman MT, Leaf DA, Kelly E, Caiozzo V, Osann K, O’Niell T. Urban angina in the mountains: effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Archives. Environ. Health 1998; 53(6): 388–397. 50. Foutch RG, Henrichs W. Carbon monoxide poisoning at high altitudes. Am. J. Emerg. Med. 1988; 6(6): 596–598. 51. Lipman GS. Carbon monoxide toxicity at high altitude. Wilderness Environ. Med. 2006; 17(2): 144–145. 52. Purser D. Interactions among carbon monoxide, hydrogen cyanide, low oxygen hypoxia, carbon dioxide, and inhaled irritant gases. In: Penney DG, Ed. Carbon Monoxide Toxicity. Boca Raton: CRC Press; 2000: 157–191. 53. White S. Update on the clinical treatment of carbon monoxide poisoning. In: Penney DG, Ed. Carbon Monoxide Toxicity. Boca Raton: CRC Press; 2000: 264. 54. Shimada H, Morita T, Kunimoto F, Saito S. Immediate application of hyperbaric oxygen therapy using a newly devised transportable chamber. Ame. J Emerg. Med. 1996; 14: 412–415. 55. Rogatsky GG, Mayevsky A. Mobile hyperbaric chamber: an essential need in emergency medicine? Ann. Emerg. Med. 2003; 42(6): 849–850. 56. Ducasse JL, Celsis P, Marc-Vergnes JP. Non-comatose patients with acute carbon monoxide poisoning: hyperbaric or normobaric oxygenation? Undersea Hyperb. Med. 1995; 22(1): 9–15. 57. Brvar M, Finderle Z, Suput D, Bunc M. S100B protein in conscious carbon monoxidepoisoned rats treated with normobaric or hyperbaric oxygen. Crit. Care Med. 2006; 34(8): 2228–30. 58. Bunc M, Luzar B, Finderle Z, Suput D, Brvar M. Immediate oxygen therapy prevents brain cell injury in carbon monoxide poisoned rats without loss of consciousness. Toxicology 2006; 225(2–3): 138–141. 59. Gilmer B, Kilkenny J, Tomaszewski C, Watts JA. Hyperbaric oxygen does not prevent neurologic sequelae after carbon monoxide poisoning. Acad. Emerg. Med. 2002; 9: 1–8. 60. Thom SR. Carbon monoxide-mediated brain lipid peroxidation in the rat. J. Appl. Physiol. 1990; 68(3): 997–1003. 61. Thom SR, Bhopale VM, Fisher D, Zhang J, Gimotty P. Delayed neuropathology after carbon monoxide poisoning is immune-mediated. PNAS 2004; 101(37): 13660– 13665. 62. Thom SR, Bhopale VM, Fisher D. Hyperbaric oxygen reduces delayed immunemediated neuropathology in experimental carbon monoxide toxicity. Toxicol. Appl. Pharm. 2006; 213(2): 152–159. 63. Thom SR. Functional inhibition of leukocyte B2 integrins by hyperbaric oxygen in carbon monoxide-mediated brain injury in rats. Toxicol. Appl. Pharm. 1993; 123(2): 248–256. 64. Penney DG. Acute carbon monoxide poisoning: animal models: a review. Toxicology 1990; 62(2): 123–160.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 429 — #39

430

Carbon Monoxide Poisoning 65. Gorman D, Lin HY, Williams C. Blockade of haem oxygenase and nitric oxide synthetase causes cortical dysfunction in sheep exposed to carbon monoxide. Toxicology 2005; 209(3): 237–243. 66. Gorman D, Huang YL, Williams C. A narcotic dose of carbon monoxide induces neuronal haeme oxygenase and nitric oxide synthetase in sheep. Toxicology 2002; 179(1–2): 79–84. 67. Gorman DF, Huang YL, Williams C. Prolonged exposure to one percent carbon monoxide causes a leucoencephalopathy in un-anaesthetised sheep. Toxicology 2001; 165(2–3): 97–107. 68. Langston P, Gorman D, Runciman W, Upton R. The effect of carbon monoxide on oxygen metabolism in the brains of awake sheep. Toxicology 1996; 114(3): 223–232. 69. American College of Emergency Physicians. Clinical policy: critical issues in the management of adult patients presenting to the emergency department with acute symptomatic carbon monoxide poisoning. 2006. 70. Phin N. Carbon monoxide poisoning (acute). Clin. Evid. 2005; (13): 1732–43. 71. Ray J, Ball C. Carbon monoxide poisoning. In: Ball CM and Phillips RS, Ed. Evidence-Based on Call: Acute Medicine: Churchill Livingstone; 2001: 144–148. 72. Medicare Services Advisory Committee (MSAC). Hyperbaric Oxygen TherapyMSAC Applications 1018–1020 Assessment report. http://wwwhealyhgovau/ haf/msac 2001. 73. Juurlink DN, Buckley NA, Stanbrook MB, Isbister GK, Bennett M, McGuigan MA, Hyperbaric oxygen for carbon monoxide poisoning (Review). The Cochrane Library 2006(2). 74. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control. Clin. Trials. 1996; 17(1): 1–12. 75. Raphael JC, Elkharrat D, Jars-Guincestre MC, et al. Trial of normobaric and hyperbaric oxygen for acute carbon monoxide intoxication. Lancet 1989; 2(8660): 414–419. 76. Thom SR, Taber RL, Mendiguren, II, Clark JM, Hardy KR, Fisher AB. Delayed neuropsychologic sequelae after carbon monoxide poisoning: prevention by treatment with hyperbaric oxygen. Ann. Emerg. Med. 1995; 25(4): 474–480. 77. Scheinkestel CD, Bailey M, Myles PS, et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial. Med. J. Aust. 1999; 170(5): 203–210. 78. Federal committee of the physicians and health insurance companies. Hyperbare Sauerstofftherapie (HBO) Zusammenfassender Bericht des Arbeitsausschusses "Ärztliche Behandlung" des Bundesausschusses der Ärzte und Krankenkassen über die Beratungen der Jahre 1999 und 2000 zur Bewertung der Hyperbaren Sauerstofftherapie gemäß §135 Abs.1 SGB V [Hyperbaric Oxygen Therapy (HBO) Summary Report of the Working Committee on “Medical Treatment” of the Federal Committee of Physicians and Health Insurance Companies Covering Years 1999 and 2000 for the Evaluation of Hyperbaric Oxygen Therapy in Accordance with §135 Abs.1 SGB V] [German]. Köln, Germany:. In: Geschäftsführung des Arbeitsausschusses; 2006. 79. Raphael JC, Chevret, S, Driheme, A, Annane, D,. managing carbon monoxide poisoning with hyperbaric oxygen (Abstract). J. Toxicol.-Clin. Toxicol. 2004; 42: 455–456. 80. Mathieu D, Wattel F, Mathieu-Nolf M, Durak C, Tempe JP, Bouachour G, Sainty JM. Randomised prospective study comparing the effect of HBO versus 12 hours NBO

8417: “8417_c018” — 2007/9/19 — 13:22 — page 430 — #40

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

81. 82.

83.

84. 85.

86. 87. 88. 89.

90. 91. 92.

93. 94. 95. 96.

97.

98.

99. 100.

431

in nin-comatose CO poisoned patients: Results of the interim analysis. Undersea Hyperb. Med. 1996; 23(Suppl.): 7. Weaver LK, Hopkins RO, Chan KJ, et al. Hyperbaric oxygen for acute carbon monoxide poisoning. NEJM 2002; 347: 1057–1067. Ducasse JL. Celsis P, Marc-Vergnes JP. Non-Comatose patients with acute carbon monoxide poisoning: hyperbaric or normobaric oxygenation? Undersea Hyperb. Med. 1995; 22(1): 9–15. Tibbles PM, Perrotta PL. Treatment of carbon monoxide poisoning: a critical review of human outcome studies comparing normobaric oxygen with hyperbaric oxygen. Ann. Emerg. Med. 1994; 24(2): 269–276. Mark P. Carbon Monoxide poisoning: a review. South Pacific Underwater Med. Soc. 1992; 22(3): 127–135. Shank ES, Muth CM. Decompression illness, iatrogenic gas embolism, and carbon monoxide poisoning: the role of hyperbaric oxygen therapy. Internatl. Anesthesiol. Clin. 2000; 38(1): 111–138. Weaver LK. Carbon monoxide poisoning. Crit. Care Clin. 1999; 15(2): 297–317, viii. Kao LW, Nanagas KA. Carbon monoxide poisoning. Emerg. Med. Clin. of N. Am. 2004; 22(4): 985–1018. Seger D, Welch L. Carbon monoxide controversies: neuropsychologic testing, mechanism of toxicity, and hyperbaric oxygen. Ann. Emerg. Med. 1994; 24(2): 242–248. Hampson NB, Mathieu D, Piantadosi CA, Thom SR, Weaver LK. Carbon monoxide poisoning: interpretation of randomized clinical trials and unresolved treatment issues. Undersea Hyperb. Med. 2001; 28(3): 157–164. Olson KR, Seger D. Hyperbaric oxygen for carbon monoxide poisoning: does it really work? Ann. Emerg. Med. 1995; 25(4): 535–537. Weaver LK, Hopkins RO, Larson-Lohr V. Hyperbaric oxygen and carbon monoxide poisoning. Ann. Emerg. Med. 1995; 26(3): 390–392. Buckley N.A. Isbister,GK, Stokes B. Hyperbaric oxygen for carbon monoxide poisoning: asystematic review and critical analysis of the evidence. Toxicol. Rev. 2005; 24: 75–92. Gorman DF, Runciman WB. Carbon monoxide poisoning. Anaesth. Intens. Care 1991; 19(4): 506–511. Littell RC Milliken GA, Stroup W.W. Wolfinger R.D. SAS (computer program). In: Cary, NC: SAS Institute Inc., 1996. Bland JMK, Sally M. Statistical Notes: Trials randomised in clusters. BMJ 1997; 315(6 September): 7108. Simpson JM, Klar N, Donnor A. Accounting for cluster randomization: a review of primary prevention trials, 1990 through 1993. Am. J. Public Health 1995; 85(10): 1378–1383. Scheinkestel CD, Tuxen DV, Bailey M, et al. Hyperbaric oxygen in carbon monoxide poisoning. Authors of study clarify points that they made. BMJ (Clinical research ed.) 2000; 321(7253): 109–111. Scheinkestel CD, Bailey M, Myles PS, et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial. Med. J. Aust. 1999; 170(11): 563–565. Moon RE, DeLong E. Hyperbaric oxygen for carbon monoxide poisoning. Med. J. Aust. 1999; 170(5): 197–199. Weaver LK. Hyperbaric oxygen in carbon monoxide poisoning. BMJ (Clinical Research ed.) 1999; 319(7217): 1083–1084.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 431 — #41

432

Carbon Monoxide Poisoning

101. Tighe SQ. Hyperbaric oxygen in carbon monoxide poisoning. 100% oxygen is best option. BMJ (Clinical Research ed.) 2000; 321(7253): 110–111. 102. Weaver LK. Hyperbaric oxygen in carbon monoxide poisoning. BMJ (Clinical research ed.) 2000; 321(8 July): 109. 103. Schiltz KL. Failure to assess motivation, need to consider psychiatric variables, and absence of comprehensive examination: a skeptical review of neuropsychologic assessment in carbon monoxide research. Undersea Hyperb. Med. 2000; 27(1): 48–50. 104. Denson LA, Hay PJ. Was the neuropsychological assessment appropriate? MJA 1999; 170(7 June): 563. 105. Scheinkestel CD, Bailey M, Myles PS, Jones K, Cooper DJ, Millar I, Tuxen DV, Hyperbaric oxygen for carbon monoxide poisoning. Med. J. Aust. 2000; 172: 141–142. 106. Weaver LK. Double-blind, controlled, Prospective, Randomized Clinical trial (RCT) in patients With Acute Carbon Monoxide poisoning: Outcome of patients treated With Normobaric Oxygen or hyperbaric Oxygen-An Interim Report. Undersea Hyperb. Med. 1995; 22(suppl.): 14. 107. Weaver LK, Hopkins RO, Larson-Lohr V. Carbon monoxide poisoning: a review of human outcome studies comparing normobaric oxygen with hyperbaric oxygen. Ann. Emerg. Med. 1995; 25(2): 271–272. 108. Buckley NA. Isbister,GK, Juurlink DN, Hyperbaric oxygen for carbon monoxide poisoning. Toxicol. Rev. 2005; 24(3): 159–160. 109. Weaver LK, Hopkins RO, Elliott G. Carbon monoxide poisoning. NEJM 1999; 340(16): 1290. 110. Weaver LK, Hopkins RO, Chan KJ, et al. Carbon Monoxide Research Group, LDS Hospital, Utah in reply to Scheinkestel et al. and Emerson: the role of hyperbaric oxygen in carbon monoxide poisoning. Emerg. Med. Aust. 2004; 16(5–6): 394–399; discussion 481–482. 111. Blumenthal I. Carbon monoxide poisoning. J. Royal Soc. Med. 2001; 94(6): 270–272. 112. The Hyperbaric Oxygen Therapy Committee Report. Hyperbaric Oxygen 2003: Indications and Results. Kensington, MD: Undersea Hyperb. Med. Soc. 2003. 113. Weaver LK. 6 & 12 month outcome of acute carbon monoxide poisoning treated with hyperbaric or normobaric oxygen. UHMS, Annual Scientific Meeting 2002 (July 2002, published in UHM, 2002, Vol. 29, No. 2, p. 139). 114. Olson KR. Hyperbaric oxygen or normobaric oxygen? Toxicol. Rev. 2005; 24(3): 151; discussion 9–60. 115. Juurlink DN. The Cochrane Library — Feedback. http://wwwcochranefeedbackcom/ cf/cda/citationdo?id=9531# 9531 2006. 116. Raphael JC, Annane D, Chevret S. Hyperbaric oxygen for acute carbon monoxide poisoning. NEJM 2003; 348(6): 557–60. 117. Weaver LK. Outcome of carbon monoxide poisoning treated with hyperbaric or normobaric oxygen. UHMS 2001 (Abstracts of the UHMS Annual Scientific Meeting June 14–16, 2001). 118. Brent J. What does the present state of knowledge tell us about the potential role of hyperbaric oxygen therapy for the treatment of carbon monoxide poisoning? Toxicol. Rev. 2005; 24(3): 145–147. 119. Henry JA. hyperbaric therapy for carbon monoxide poisoning. Toxicol. Rev. 2005; 24(3): 149–150. 120. Bentur Y. Hyperbaric oxygen for carbon monoxide poisoning. Toxicol. Rev. 2005; 24(3): 153–154; discussion 9–60.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 432 — #42

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

433

121. Seger D. WL. The Myth. Toxicol. Rev. 2005; 24(3): 155–156. 122. Gibbs Brown J, Inspector General. Hyperbaric oxygen therapy: its use and appropriateness. In: Services Department of Health and Human Services, ed.: Office Of Inspector General; 2000. 123. Hampson NB, Dunford RG, Ross DE, Wreford-Brown CE. A prospective, randomized clinical trial comparing two hyperbaric treatment protocols for carbon monoxide poisoning. Undersea Hyperb. Med. 2006; 33(1): 27–32. 124. Hampson NB, Zmaeff JL. Outcome of patients experiencing cardiac arrest with carbon monoxide poisoning treated with hyperbaric oxygen. Ann. Emerg. Med. 2001; 38(1): 36–41. 125. Piantadosi CA, Carraway MS, Suliman HB. Carbon monoxide, oxidative stress, and mitochondrial permeability pore transition. Free Rad. Biol. Med. 2006; 40(8): 1332–1339. 126. Cronje FJ, Carraway MS, Freiberger JJ, Suliman HB, Piantadosi CA. Carbon monoxide actuates O(2)-limited heme degradation in the rat brain. Free Rad. Biol. Med. 2004; 37(11): 1802–1812. 127. Wu L, Wang R. Carbon monoxide: endogenous production, physiological functions, and pharmacological applications. Pharmacol. Rev. 2005; 57(4): 585–630. 128. Kirkby KA, Adin CA. Products of heme oxygenase and their potential therapeutic applications. Am. J. Physiol. Renal. Physiol. 2006; 290(3): F563–571. 129. Nakao A, Choi AM, Murase N. Protective effect of carbon monoxide in transplantation. J. Cell. Mol. Med. 2006; 10(3): 650–671. 130. Zuckerbraun BS, Chin BY, Wegiel B, et al. Carbon monoxide reverses established pulmonary hypertension. J. Exper. Med. 2006; 203(9): 2109–2119. 131. Kohmoto J, Nakao A, Kaizu T, et al. Low-dose carbon monoxide inhalation prevents ischemia/reperfusion injury of transplanted rat lung grafts. Surgery 2006; 140(2): 179–185. 132. Li Volti G, Rodella LF, Di Giacomo C, et al. Role of Carbon Monoxide and Biliverdin in Renal Ischemia/Reperfusion Injury. Nephron Exp. Nephrol. 2006; 104(4): e135–e139. 133. Botros FT, Navar LG. Interaction between endogenously produced carbon monoxide and nitric oxide in regulation of renal afferent arterioles. Am. J. Physiol. Heart Circ. Physiol. 2006; 291(6): H2772–2778. 134. Rodella L, Lamon BD, Rezzani R, et al. Carbon monoxide and biliverdin prevent endothelial cell sloughing in rats with type I diabetes. Free Rad. Biol. Med. 2006; 40(12): 2198–2205. 135. Chen YH, Tsai HL, Chiang MT, Chau LY. Carbon monoxide-induced early thrombolysis contributes to heme oxygenase-1-mediated inhibition of neointimal growth after vascular injury in hypercholesterolemic mice. J. Biomed. Sci. 2006; 13(5): 721–730. 136. Bilban M, Bach FH, Otterbein SL, et al. Carbon monoxide orchestrates a protective response through PPARgamma. Immunity 2006; 24(5): 601–610. 137. Li MH, Cha YN, Surh YJ. Carbon monoxide protects PC12 cells from peroxynitriteinduced apoptotic death by preventing the depolarization of mitochondrial transmembrane potential. BBRC 2006; 342(3): 984–990. 138. Lee BS, Heo J, Kim YM, et al. Carbon monoxide mediates heme oxygenase 1 induction via Nrf2 activation in hepatoma cells. BBRC 2006; 343(3): 965–972. 139. Schober P, Koch A, Zacharowski K, Loer SA. [Carbon monoxide: toxic molecule with antiinflammatory and cytoprotective properties]. Anasthesiol. Intensivmed. Notfallmed. Schmerzther 2006; 41(3): 140–149.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 433 — #43

434

Carbon Monoxide Poisoning

140. Mayr FB, Spiel A, Leitner J, et al. Effects of carbon monoxide inhalation during experimental endotoxemia in humans. Am. J. Resp. and Crit. Care Med. 2005; 171(4): 354–360. 141. Giffard RG, Jaffe RA. Advances in understanding protection from cerebral ischemia. Curr. Opin. Anaesthesiol. 2002; 15(5): 495–500. 142. Wang KK, Larner SF, Robinson G, Hayes RL. Neuroprotection targets after traumatic brain injury. Curr. Opin. Neurol. 2006; 19(6): 514–519. 143. Henry CR, Satran D, Lindgren B, Adkinson C, Nicholson CI, Henry TD. Myocardial injury and long-term mortality following moderate to severe carbon monoxide poisoning. JAMA 2006; 295(4): 398–402. 144. Rasmussen LS, Poulsen MG, Christiansen M, Jansen EC. Biochemical markers for brain damage after carbon monoxide poisoning. Acta Anaesthesiologica Scand 2004; 48(4): 469–73. 145. Berger RP, Adelson PD, Richichi R, Kochanek PM. Serum biomarkers after traumatic and hypoxemic brain injuries: insight into the biochemical response of the pediatric brain to inflicted brain injury. Develop. Neurosci. 2006; 28(4–5): 327–335. 146. Brvar M, Mozina H, Osredkar J, et al. S100B protein in carbon monoxide poisoning: a pilot study. Resuscitation 2004; 61(3): 357–360. 147. Gale SD, Hopkins RO. Effects of hypoxia on the brain: neuroimaging and neuropsychological findings following carbon monoxide poisoning and obstructive sleep apnea. J. Int. Neuropsychol. Soc. 2004; 10(1): 60–71. 148. Wu CI, Changlai SP, Huang WS, Tsai CH, Lee CC, Kao CH. Usefulness of 99mTc ethyl cysteinate dimer brain SPECT to detect abnormal regional cerebral blood flow in patients with acute carbon monoxide poisoning. Nucl. Med. Comm. 2003; 24(11): 1185–1188. 149. Sener RN. Acute carbon monoxide poisoning: diffusion MR imaging findings. AJNR 2003; 24(7): 1475–7. 150. Teksam M, Casey SO, Michel E, Liu H, Truwit CL. Diffusion-weighted MR imaging findings in carbon monoxide poisoning. Neuroradiology 2002; 44(2): 109–113. 151. Kesler SR, Hopkins RO, Blatter DD, Edge-Booth H, Bigler ED. Verbal memory deficits associated with fornix atrophy in carbon monoxide poisoning. J Int Neuropsychol. Soc 2001; 7(5): 640–646. 152. Hurley RA, Hopkins RO, Bigler ED, Taber KH. Applications of functional imaging to carbon monoxide poisoning. J. Neuropsychiatry and Clin. Neurosci. 2001; 13(2): 157–160. 153. Gale SD, Hopkins RO, Weaver LK, Bigler ED, Booth EJ, Blatter DD. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning. Brain Inj. 1999; 13(4): 229–243. 154. Silver DA, Cross M, Fox B, Paxton RM. Computed tomography of the brain in acute carbon monoxide poisoning. Clin. Radiology 1996; 51(7): 480–483. 155. Pracyk JB, Stolp BW, Fife CE, Gray L, Piantadosi CA. Brain computerized tomography after hyperbaric oxygen therapy for carbon monoxide poisoning. Undersea Hyperb. Med. 1995; 22(1): 1–7. 156. Murata T, Itoh S, Koshino Y, et al. Serial cerebral MRI with FLAIR sequences in acute carbon monoxide poisoning. J. Computer-Assisted Tomogr. 1995; 19(4): 631–634. 157. Gotoh M, Kuyama H, Asari S, Ohmoto T, Akioka T, Lai MY. Sequential changes in MR images of the brain in acute carbon monoxide poisoning. Comput. Med. Imaging Graph. 1993; 17(1): 55–59.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 434 — #44

Hyperbaric Oxygen for Acute Carbon Monoxide Poisoning

435

158. Murata M, Suzuki M, Hasegawa Y, Nohara S, Kurachi M. Improvement of occipital alpha activity by repetitive hyperbaric oxygen therapy in patients with carbon monoxide poisoning: a possible indicator for treatment efficacy. J. neurol. Sci. 2005; 235(1–2): 69–74. 159. Vink R, Nimmo AJ, Cernak I. An overview of new and novel pharmacotherapies for use in traumatic brain injury. Clin. Exper. Pharm. Physiol. 2001; 28(11): 919–921. 160. Roof RL, Hall ED. Gender differences in acute CNS trauma and stroke: neuroprotective effects of estrogen and progesterone. J. Neurotrauma 2000; 17(5): 367–388. 161. Choudhry MA, Bland KI, Chaudry IH. Gender and susceptibility to sepsis following trauma. Endocrine, Metab. Immune Disorders Drug Targets 2006; 6(2): 127–135. 162. Hopkins RO. Effects of the ApoE4 allele on cognitive outcome in acute CO poisoning. Undersea Hyperb. Med. 2006; 33(5): 337–338.

8417: “8417_c018” — 2007/9/19 — 13:22 — page 435 — #45

8417: “8417_c018” — 2007/9/19 — 13:22 — page 436 — #46

19

A Challenge to the Healthcare Community: The Diagnosis of Carbon Monoxide Poisoning David G. Penney

CONTENTS 19.1 19.2 19.3 19.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems in the Diagnosis of CO Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chronic Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Addendum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.4.1 The RAD-57 cm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

437 439 443 446 446 448

19.1 INTRODUCTION Overview Statement: If you were able to chose the kind of brain injury you were to incur, it would be better in terms of the potential for recovery to have a stroke, a concussion in a motor vehicle accident, etc. than carbon monoxide poisoning.

There are many problems with CO poisoning generally (Table 19.1). Although CO is a very simple molecule, its mechanisms of action are complex and multiple. There is more to its pathophysiology than simply the tying up of hemoglobin such that oxygen cannot be transported. It traverses the blood-barrier and dissolves in the protoplasm of cells, attaching there to a variety of molecules. As stated elsewhere in this book, CO is a very “smart poison.” It is colorless, tasteless, odorless, and is completely nonirritating to respiratory and mucous membranes. Thus it is undetectable by humans with unaided senses. Unlike other “dumb” poisons, it leaves the body quickly when the victim again breathes unpolluted air, leaving behind just the damage it has done. CO doesn’t hang around to be detected and identified weeks, months, or years later, like lead and mercury. 437

8417: “8417_c019” — 2007/9/11 — 12:14 — page 437 — #1

Carbon Monoxide Poisoning

438

TABLE 19.1 Some of the Major Problems in Dealing with Carbon Monoxide Poisoning1 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Mechanisms of toxicity involve more than CO attaching to hemoglobin. Lack of public awareness of the dangers of accidental CO poisoning. Insufficient professional awareness in the diagnosis of CO poisoning. Extent of suicidal use of motor exhaust and other fumes. Assessment of the severity of the poisoning. Criteria to be used in the treatment of poisoned patients. Pathogenesis and prediction of the development of delayed sequelae. Effectiveness of hyperbaric oxygen therapy is largely unproven. Are there any effective alternative treatments to HBO? Unknown morbidity and mortality of undiagnosed cases.

The public and indeed most members of the healthcare community are ignorant of many of the risks involved in working and playing around equipment that emits CO (see Chapter 15). While automobiles have enjoyed the reputation of being extremely dangerous in terms of CO emissions, this is no longer the case when they are in good repair and are used properly. Because of catalytic converters, virtually all of the CO produced by the automobile engine is now converted to harmless carbon dioxide and water vapor. Other vehicles such as high-powered, watercraft fueled by gasoline and using engines similar to those used in cars, now pose the greatest risk of deadly CO poisoning to individuals. Some of these boats, such as a single inboard, twin-engine ski boat may emit as much CO as 250–300 automobiles. CO concentrations immediately behind such boats may range to over 20,000 ppm depending on boat movement, wind direction, and speed, CO concentrations at least 40 times the lethal dose for humans. People have difficulty recognizing and appreciating the danger posed by these boats, since it usually occurs in the open air, not inside a structure. Medical personnel have traditionally underappreciated the frequency of CO poisonings, both the acute and the chronic types, but especially the latter, and have been notorious in misdiagnosing CO poisoning. Because the symptoms produced are mostly general and nonspecific, they are usually associated with other diseases that lie more central to internal medicine, conditions that result from bacteria, viruses, hormonal changes, physical trauma, and so forth. In fact, the very nature of the multiplicity of symptoms that CO poisoning produces, tends to confuse physicians and nurses. If the flu, hypothyroidism, stroke, and so forth are not suggested, then psychosomatic or psychiatric conditions often are. Since few symptoms or signs are produced that occur only with CO-poisoning (i.e., pathognomonic), it is often difficult to immediately identify a cause. Identification of CO poisoning as the cause is made easier by the taking of a full situational history. When possible, this consists of determining where the patient was during the past few hours. Who was with him and were the others also sick? Was it a house, apartment, recreational vehicle, and so forth? Were pets also sick or died? How was the space heated? Did the patient notice

8417: “8417_c019” — 2007/9/11 — 12:14 — page 438 — #2

A Challenge to the Healthcare Community

439

a pattern in that he/she was less sick when he/she left the site, and became sicker again upon return? Had there been reports of heating devices malfunctioning, or the use of internal combustion engines near or within the structure? Assessment of the severity of CO poisoning has been a thorny problem. Far too much reliance has been placed on carboxyhemoglobin (COHb) measured at the emergency center, and too little on evaluation of the conscious state and behavior of the victim. Recent studies suggest that gait and balance may be among the most important prognostic signs, and should be used far more often in deciding the course of treatment. Long or “soaking” exposures to CO invite more rapid and aggressive therapy than shorter CO exposures. The early use of the new generation of pulse CO-oximeters (see addendum and Chapter 33), possibly even at the site of discovery, will aid in assessing poisoning severity and treatment regimen. As long as the use of hyperbaric oxygen therapy (HBOT) has been available to treat CO-poisoned patients, it remains unclear whether it is really effective in reduced long-term health damage. Two other chapters in this book discuss in detail the pros and cons of the half a dozen or so clinical studies that have tested the HBO hypothesis. Little progress appears to have been made in the past decade in developing new approaches to treating CO poisoning, or in understanding HBOT.

19.2 PROBLEMS IN THE DIAGNOSIS OF CO POISONING CO poisoning is difficult to diagnose for a number of reasons, and the rate of misdiagnosis has been in some cases, shockingly high. Part of the problem has to do with its special properties, and some has to do with healthcare workers ignorance of the symptoms by which it presents (Table 19.2). There has been overreliance on COHb level, and also the fact that COHb is not routinely measured in patients

TABLE 19.2 Major Reasons for Failure to Diagnose Carbon Monoxide Poisoning in the Emergency Room 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Usually a low (or zero) index of suspicion for CO poisoning by physicians. Lack of training in toxicology and myopic thinking by physicians. Too many disparate, seemingly unrelated, multisystems symptoms involved. Nonspecific symptoms confused physicians. Incorrect clue given by patient or companion. Failure to obtain complete situational histories. Presentation in ER that appears not to require emergency measures. Mistaken belief that COHb must always be greatly elevated. Distraction by other traumatic or metabolic condition/disease. Dependence on old style, two wavelength pulse-oximeter for measurement of O2 saturation. Use of less than the best, even irrelevant clinical tests for CO, unless goal is R/O (i.e., rule out). COHb is NOT routinely measured in the ER.

8417: “8417_c019” — 2007/9/11 — 12:14 — page 439 — #3

Carbon Monoxide Poisoning

440

TABLE 19.3 Carbon Monoxide Exposure Types •

Acute, brief exposure

•

Acute, “soaking”

•

Chronic (>24 h)

Immediate presentation Delayed presentation + 100% oxygen Immediate presentation Delayed presentation +100% oxygen Immediate presentation Delayed presentation +100% oxygen

high COHb COHb may be low COHb may be low - high COHb COHb may be low COHb may be low COHb high to low COHb may be low COHb may be low

TABLE 19.4 A Situational History • • • • • • • • • •

Living abode - house, apartment, etc. Heating systems Other people sick there Pets sick/dead Feel better when away, worse when return Periodicity with work, weekends, vacations/season Car driven - year, engine Use of small gasoline engine/attached garage Prolonged illness/seen other docs previously/recently New vs. old building

entering a physician’s office or an emergency room (ER). Recent studies show that there is virtually no relationship between reported COHb saturation and long-term outcome. In some cases COHb was barely elevated or even normal, yet the patient sustained brain damage. The grid of the various types of CO poisonings shown in Table 19.3 indicates why COHb may be normal or well below levels considered toxic in individuals who sustained serious CO poisonings. We see oxygen saturation obtained by two-wavelength pulse oximeters continuing to be printed on the charts of CO-poisoned patients although it has no reality in such cases. The falsely high number could even lead to tragic action by a healthcare worker who does not understand that the number is not accurate. There are numerous examples of healthcare workers failing to avail themselves of all of the clues available to them when a patient presents with symptoms consistent with CO poisoning. Many of these cases end tragically. A high index of suspicion must be maintained about possible CO poisoning when seeing patients (Table 19.4). One very important tool in diagnosing CO poisoning, and indeed other poisons, is the “situational history”. Table 19.5 shows some of the components, that is, questions asked when and if possible, when building a situational history.

8417: “8417_c019” — 2007/9/11 — 12:14 — page 440 — #4

A Challenge to the Healthcare Community

441

TABLE 19.5 Index of Suspicion (for CO Poisoning) Should be Raised by: • • • • • • • • • • •

Presentation during winter Several/many people presenting “ill” from same household Elevated plasma glucose c/o Memory problems c/o Furnace/heat problem at home Presentation after electrical power outage c/o Pets sick/dead c/o Headache, Nz, Dz, fatigue/weakness c/o “Illness” when going to/staying at, one place, not at others Flushed pink/cherry red skin Unexplained LOC/syncope

Very recently in Florida, a man and his adult son were staying at a motel resort. Overnight both were poisoned by CO. The son died, while the father survived. It was later learned that other motel patrons had been sick the week before, had gone to the hospital, where they were eventually diagnosed with CO poisoning. With an appropriate index of suspicion of CO, the use of proper diagnostic techniques by emergency personnel, and effective communication, tragedies like this can be averted. Clearly, getting the right diagnosis can be critical with possible CO poisoning. Table 19.6 lists possible misdiagnoses of CO poisoning, and Table 19.7 lists the frequency of reported incidences of major misdiagnoses of CO poisoning. In another instance several years ago, a prominent urologist was attending a meeting in the west with his wife, a surgical nurse. After the first night in a posh new ski hotel, both became violently sick. They went to the local ER. Upon entering the facility the urologist said he believed they had eaten some “bad” chicken sandwiches on the plane coming from the east. The ER personnel, hearing this from a fellow physician, accepted his diagnosis and treated him and his wife with antiemetics, and so forth. The doctor and his wife then went back to the hotel to continue the urology meeting, skiing, and dining. The next morning the urologist was dead, and his wife was in a coma that lasted several days. She suffered permanent, irreversible brain damage, involving changes in cognitive ability, personality, physical pain, and so forth. If only COHb had been measured at the ER. Later it was learned through testimony and examination of internal memos that employees and management of the hotel had been aware of a “CO problem” there for over 6 months before the doctor died. In another case in a large eastern US city, an elderly woman was taken to a very prominent local hospital known for its expertise and training programs in emergency and trauma medicine. ER personnel there failed to diagnose the patient with CO poisoning, apparently because the emergency medical technicians who went to her house and transported the woman saw a bottle near the chair she was found in. She was treated as an alcohol overdose case. Shortly afterward, it was discovered that there

8417: “8417_c019” — 2007/9/11 — 12:14 — page 441 — #5

Carbon Monoxide Poisoning

442

TABLE 19.6 Some Possible Mistaken Diagnoses in Patients with Carbon Monoxide Poisoning1 Misdiagnosis

Cause

Neurological Cerebrovascular accident Migraine, tension headache Epilepsy Meningitis, encephalitis Parkinsonism

Cerebral ischemic accident due to CO poisoning Headache Anoxic convulsions Vomiting, headache, bizarre neurological symptoms Late-onset parkinsonian symptoms

Psychiatric Depression Anxiety state Hyperventilation syndrome Acute confusional state

Lethargy, somatic symptoms Hyperventilation, headache, malaise Hyperventilation Confusion, hallucinations

Cardiac Myocardial infarction Cardiac arrhythmias

A critical coronary artery lesion decompensated through hypoxia Conduction system hypoxia

Pharmacological and toxicological Drug overdose Ethylene glycol poisoning Ethanol intoxication Drug abuse

Hypoxic coma, nontraumatic rhabdomyolysis Coma and renal failure Vomiting, ataxia, slurred speech, coma Agitation, confusion, hallucinations

Infections Influenza and other viral infections Post viral syndrome Gastroenteritis and food poisoning Pneumonia Sinusitis

Muscle aches, tachypnea, headache, exhaustion Lethargy, myalgia Nausea, vomiting Dyspnea, delirium Headache, malaise

Others Cholecystitis and other acute abdominal conditions

Abdominal pain, nausea, vomiting

was no liquor of any kind in the house. The woman had been sitting, then slumping, in the same chair for as long as 3 days. At the hospital the usual “tox screen” found no alcohol or other drugs in her body. Shortly after that, substantial concentrations of CO were discovered in her house by the woman’s daughter. The woman remained comatose for many days in hospital, had a long drawn out period of recovery, and sustained severe brain damage. This is the second example of comments by less qualified personnel being allowed to influence the decision about the proper medical work-up to achieve an accurate diagnosis. In this case the patient failed to receive the proper treatment, which in this case would almost certainly have been HBOT.

8417: “8417_c019” — 2007/9/11 — 12:14 — page 442 — #6

A Challenge to the Healthcare Community

443

TABLE 19.7 Common Misdiagnosis for Carbon Monoxide Poisoning by Reported Frequency2 Food poisoning Psychiatric disease, hysteria, confusion, depression Heart disease presenting as angina or syncope Alcohol intoxication or delirium tremens Acute solvent intoxication Migraine headache Ischemic cerebral disease Cerebral hemorrhage Cerebral tumor (convulsions)

38% 18% 13% 7% 7% 6% 4% 4% 3%

TABLE 19.8 Characteristics of Chronic Carbon Monoxide Poisoning6 1 2 3 4 5 6 7 8 9

Lasts more than 24 h Often goes long undetected Masquerades as the flu, fatigue, and so forth. Often many people “sick” simultaneously The “sickness” may show a periodicity synchronous with season May go away upon leaving poisoning site (to work, on vacation, etc.) Nearly always misdiagnosed by physicians May involve pets “sick,” dead at same time Rarely involves sinus congestion, cough [when present, it may be due to other compounds (e.g., NOx , SO2 ) in exhaust gases]

19.3 CHRONIC CARBON MONOXIDE POISONING Working Assumption: For every single case of chronic carbon monoxide poisoning reported / successfully diagnosed, there are 10 cases that go unreported/undiscovered/undiagnosed.

Chronic CO poisoning as I define it is an exposure to CO of more than 24 h, whether continuous or discontinuous.3 Beyond this, the characteristics of such CO poisonings (Table 19.8) often go long undetected, sometimes for years. It may masquerade as the flu, chronic fatigue, fibromyalgia, chemical hypersensitivity (MCS), and so forth. Often several or many people are “sick” simultaneously, and for extended periods uncharacteristic of viral or bacterial infections. The condition often subsides or disappears when the victim leaves the poisoning site, but reappears when the victim returns. It may show a seasonal periodicity if it is the primary space heating appliance that is malfunctioning. This type of CO poisoning has the highest misdiagnosis

8417: “8417_c019” — 2007/9/11 — 12:14 — page 443 — #7

Carbon Monoxide Poisoning

444 Cycle of misdiagnosis

CO ignored

CO regarded as rare

Patient ill and stays at home

Chronic exposure to CO

Misdiagnosis

Effects not understood

FIGURE 19.1 Cycle of misunderstanding and misdiagnosis.

rate by healthcare professionals. In the CO Support Study in the UK it was the heating professionals who were most likely to diagnose the problem.4 This type of CO poisoning may result in pets sick with symptoms not unlike those of the humans. It rarely involves sinus congestion, cough or sneezing, although leaking exhaust gases containing irritants could do this—CO alone will not.5 Figure 19.1 illustrates the cycle of misdiagnosis often encountered with chronic CO poisoning. The presence of CO in a breathing space usually goes long undetected, and even when it is detected may be denied by those in authority. The presence of the CO and its possible health damaging effects are often misunderstood by those being poisoned, and worse, by the medical personnel consulted. Finally, people feeling ill from what eventually turns out to be CO, almost invariably spend increased time in the contaminated breathing space attempting to recover, but only become worse or possibly even die with severe poisoning. Some of the comments that I hear from the public regarding such poisonings are found in Table 19.9. I have condensed these comments into what I call “commonalities.” The symptoms consist of headache, fatigue, nausea, dizziness, and so forth.3 Nearly everyone who contacts me is certain he/she was exposed to CO in one way or another. There has to be a source, since CO doesn’t come out of thin air. A dominant theme I’ve heard hundreds of times is that “my doctor won’t take my complaints seriously,” that “CO comes, goes (once you are in fresh air), and then you are ok,” and finally, “there are no medical people in my area who understand CO’s effects!”. The public wants to know how it is diagnosed, when he/she will get better, and what the treatments are. Diagnosis of CO poisoning is a problem as discussed above, particularly if the victim has left the site of exposure for some days or weeks, since CO quickly leaves the body. Making pronouncements about when people will

8417: “8417_c019” — 2007/9/11 — 12:14 — page 444 — #8

A Challenge to the Healthcare Community

445

TABLE 19.9 Inquiries About Carbon Monoxide Poisoning—Commonalities in my Experience6 • • • • • •

Complaints of continuous headache, fatigue, nausea, dizziness, and so forth. Certainty by complainant that CO exposure occurred. “My doctor won’t take my complaints seriously!”—“CO comes, goes, and . . . (he/she says) you are ok!” “There are no medical people are in my area who understand CO’s effects!” How is it diagnosed? When will I get better? What are the treatments? What long-term damage might the CO have caused to my child/children?

TABLE 19.10 Clues to the Discovery of Chronic Carbon Monoxide Poisoning6 • • • • • • • • •

Lethargy, headache, and so forth of long duration Long-standing “illness” intractable to medical solutions Multiple cases of similar illness at one location “Illness” that may suddenly improve when leaving site “Illness” that improves when combustion device(s) is turned off or taken away Morbidity/mortality of pets CO alarm sounding, once or repeatedly Presence of malfunctioning furnace, water heater, and so forth. Measurement of CO by fireman, service personnel, and so forth at the presumed site of poisoning.

be well again is very risky even under circumstances of thorough knowledge of a case, and telling people it is unlikely they will get well or that their health damage is permanent is very hard to do. I do make recommendations about treatment, but as we all know, brain damage currently remains irreversible. Finally, many people ask what long-term damage might the CO have caused their children, born and unborn. Many effects of CO on both children and fetuses have been described by Penney,7 but identifying damage caused by CO in any one case is difficult. There are many clues to discovery of chronic CO poisoning (Table 19.10). Headache and lethargy of long duration should raise the suspicion of CO exposure. Similar to the comments above, a long-standing “illness” intractable to medical solutions should have CO poisoning placed high on the differential diagnosis list. Likewise, the incidence of multiple cases of similar “illness” at one location is an important clue, and an “illness” that suddenly improves when the victim leaves the site, or when a key combustion device is rendered nonfunctional or is removed, should cause the investigator to suspect a site-specific poisoning. As above, morbidity or even mortality of pets is important. “CO is an equal opportunity poison,” not discriminating on the basis of skin color, gender, religion, or even species. If the organism is warm-blooded and uses hemoglobin as its circulating oxygen carrier, it is subject to injury or death from CO poisoning. The degradation of a heating appliance

8417: “8417_c019” — 2007/9/11 — 12:14 — page 445 — #9

Carbon Monoxide Poisoning

446

TABLE 19.11 Common Misdiagnoses for Chronic Carbon Monoxide Poisoning6 • • • • • • • • • • • • • • •

The “flu” Other viral or bacterial pulmonary or gastrointestinal infections Bad/tainted/poisoned food Psychosomatic problem, malingering General “run-down” condition A “female” problem Allergy/asthma Psychiatric condition, for example, depression Chronic fatigue syndrome Chemical hypersensitivity (i.e., MCS) Fibromyalgia Multiple sclerosis (MS) Lymes disease Endocrine problem (e.g., hyper- or hypo-thyroid condition) Immune deficiency

such as furnace through corrosion of one or another part, often produces chronic CO poisoning, while discovery of the problem usually occurs when the CO problem is particularly severe. A CO detector/alarm that sounds intermittently over a period of days or weeks is probably identifying a site of chronic CO poisoning. Table 19.11 lists common misdiagnoses made in cases of chronic CO poisoning. The “flu” is probably the most common misdiagnosis. Lack of a fever does not guarantee that the condition is not being caused by a “bug.” “Bad or tainted” food may be pronounced the cause, but this curse is more commonly encountered as the misdiagnosis in cases of acute CO poisoning, as revealed in the tragic story above. A third large area of false cause identification centers around accusations of mental illness, depression, psychosomatic condition, malingering/faking it, and exaggerating minor irritations. This usually stems from the confusion that CO poisoning causes in medical personnel owing to the large number of symptoms generated. Other classic misdiagnoses for chronic CO poisoning include multiple sclerosis, lyme’s disease, fibromyalgia, chronic fatigue syndrome, hypothyroidism, chemical hypersensitivity, allergic reaction, immune deficiency, a “female” problem, general run-down condition, and endocrine problem. Many of these can be easily ruled out by doing the appropriate clinical test.

19.4 ADDENDUM 19.4.1 THE RAD-57 CM Masimo is the innovator of Signal Extraction Technology (SET)® pulse oximetry and the inventor of pulse CO-oximetry. This technology is capable of continuously and noninvasively measuring COHb and methemoglobin (MetHb) in the blood.

8417: “8417_c019” — 2007/9/11 — 12:14 — page 446 — #10

A Challenge to the Healthcare Community

447

FIGURE 19.2 Masimo Rad-57, hand-held pulse CO-oximeter. It allows clinicians and others to continuously and noninvasively measure carbon monoxide levels in the blood, reducing the need for invasive and expensive arterial blood gas analysis. (From Masimo Corporation website. Rad-57 Pulse CO-oximeter. Available at: http://www.masimo.comlrad-57/index.htm. Accessed September 26, 2005.)

Masimo Rainbow SET™ was developed out of Masimo SET® . It informs the operator about the vascular oxygen status of patients, while continuously and noninvasively monitoring other species of hemoglobin, such as COHb (SpCO%) and MetHb (SpMet%). Rainbow extends Masimo SET® through the addition of signal analysis algorhithms that run in parallel with it, to reveal the presence and levels of these hemoglobin species (see Figure 19.2) Because the Masimo parallel engines and adaptive filters receive more than seven discrete wavelengths of light (note incoming signal λ1 . . . λn ), multiple constituents of hemoglobin are detectable and can be quantitated, as compared to conventional pulse-oximeters that employ only two wavelengths. The wavelengths used span those necessary to recognize oxygenated hemoglobin as well as COHb and MetHb. For example, when CO is bound to hemoglobin, a conventional red/infrared oximeter misreads COHb essentially as oxygenated hemoglobin, producing a falsely high SpO2 value that may have disastrous immediate, delayed, or even chronic effects on brain and cardiac function. Pulse CO-oximetry relies on the same principles of spectrophotometry used to determine blood oxygen saturation in the laboratory. The underlying physical principle in pulse oximetry is the absorption of specific light wavelengths while passing

8417: “8417_c019” — 2007/9/11 — 12:14 — page 447 — #11

Carbon Monoxide Poisoning

448

through tissues. In pulse CO-oximetry the percent of total hemoglobin found as either COHb (SpCO%) and/or MetHb (SpMet%) is measured. The Masimo Rainbow sensor emits multiple wavelengths of light. Output of the adjoining photo detector is sent to the instrument, where Rainbow SET™ technology employs parallel algorhithms and adaptive digital filters to process the data. The Rainbow device displays as percentages: oxyhemoglobin (SpO2 %), COHb (SpCO%), and MetHb (SpMet%), plus the core parameters derived from the Masimo SET technology platform, pulse rate (PR), perfusion index (PI), and signal IQ® (SIQ). A number of single-use, latex-free, adhesive sensors are available for patients . . . from adults to neonates. Reusable adult and slender finger clip sensors are also available for short-term monitoring and spot checks. Note that a Rainbow-empowered device provides a choice of two wavelength sensors for SpO2 , pulse rate, and perfusion index measurements, or Rainbow SET® multiple-wavelength sensors to add COHb (SpCO%), MetHb (SpMet%), or both to the pulse-oximeter. Editors note: This section has been included in this book for informational purposes only. The editor is not an employee of Masimo Corporation and has taken no money, gifts, whatever, and will not do so in the future, for providing this inclusion.

References 1. Lowe-Ponsford, F.L., Henry, J.A. Clinical aspects of carbon monoxide poisoning. Adverse Drug Reactions and Acute Poisoning Reviews, 8, 217–240, 1989. 2. Bartlett, R. Carbon monoxide poisoning. In: Clinical Management of Poisoning and Drug Overdose, 3rd ed., Haddad, L.M., Shannon, M.W., Winchester, J.F., ed., W.B. Saunders Co., Philadelphia, Chapt.70, pp. 885–898. 1998. 3. Penney, D.G. Chronic carbon monoxide poisoning. In: Carbon Monoxide Toxicity, D.G. Penney, Ed., CRC Press, NY, 2000, Chapt. 18, pp. 393–418. 4. Hay, A.W.H., Jaffer, S., Davis, D. Chronic carbon monoxide exposure: The CO Support study. In: Carbon Monoxide Toxicity, D.G. Penney, ed., CRC Press, NY, 2000, Chapt. 19, pp. 419–438. 5. Dwyer, B., Leatherman, Manclark, Kimball, K., Rasmussen, R. Carbon Monoxide: A Clear and Present Danger, ESCO Press, USA, 3rd ed., 2003 (www.escoinst.com). 6. Available at: www.coheadquarters.com/CO1.htm 7. Penney, D.G. Effects of carbon monoxide exposure on developing animals and humans. In: Carbon Monoxide, D.G. Penney, Ed., CRC Press, NY, 1996, pp. 109–144.

8417: “8417_c019” — 2007/9/11 — 12:14 — page 448 — #12

20

Neuroimaging after Carbon Monoxide Exposure Gunnar Heuser

CONTENTS 20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.2 Sequential Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.3 Location and Symmetry of Lesions, and Hypofrontality and Disinhibition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.4 Review of the Recent Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

449 450 450 451 453 454 454

20.1 INTRODUCTION Neuroimaging was previously reviewed by I.S.S. Choi in the book Carbon Monoxide Toxicity, by Penney, published in 2000.1 His review included representative Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Single Photon Emission Computed Tomography (SPECT), and Magnetic Resonance Spectroscopy (MRS) scans. My review is meant to update and complement that of Choi’s. Neuroimaging helps to understand the clinical picture and to correlate findings with other evaluations. The field has advanced in recent years in that the techniques are in some respects vastly improved. The reader is invited to compare the SPECT scans presented in Choi’s chapter with the ones presented here. Present-day imaging of anatomical (MRI, CT) and functional [(SPECT, Positron Emission Tomography (PET)] impairment provides a significant visual aid to the assessment of brain anatomy and function, especially when presented in color. Brain scans now also lend themselves to statistical analysis, comparing a given scan to a control population or analyzing regions of interest (ROI) with other regions in the same brain scan.2

449

8417: “8417_c020” — 2007/9/19 — 13:23 — page 449 — #1

Carbon Monoxide Poisoning

450

20.2 SEQUENTIAL STUDIES CO exposure, acute and chronic, can result in long lasting (i.e., years) impairment of neurological and other functions. The need for sequential follow-up exams is obvious in view of the potential long-term effects of CO poisoning. Follow-up neuroimaging and other testing is often mandatory for assessment of a given patient as the clinical picture evolves. It is also important if one wants to assess the efficacy of the course of treatment. Sequential neuropsychological testing is a very useful assessment tool. However, it is expensive. Also, some tests administered suffer from potential learning effects, thus giving the false impression of improvement at the time of follow-up, because the patient has learned to perform better in the follow-up test since he or she is repeating the test. Examinations by a neurologist are not always comparable, especially if the patient is not followed by the same physician. In view of the above limitation, sequential neuroimaging studies appears to be the procedure of choice in assessing a patient over time, provided the patient returns to the same imaging facility which uses the same imaging protocol.

20.3 LOCATION AND SYMMETRY OF LESIONS, AND HYPOFRONTALITY AND DISINHIBITION From a diagnostic point of view, abnormalities after CO exposure would be expected to be present in similar locations and to be symmetrical. This is not the case. One would also expect clinical observations and neuroimaging abnormalities to show a correlation. Often, this is actually not the case. Finally, one would naively assume that damage in different brain regions should be viewed as a focal abnormality which is not functionally connected to other areas of the brain. This is also not true since frontal hypofunction (hypofrontality) may result in hyperfunction in the posterior areas of the brain and also in subcortical areas. Since the frontal brain exerts an inhibitory function, frontal lobe damage may result in disinhibition.3,4 While neurotoxic exposure (e.g., solvents, pesticides, mold toxins) in general may not be obvious on the MRI at all, or result in nonspecific small high intensity (demyelinating/vascular) lesions (foci), MRI studies in CO-exposed patients very frequently show abnormal MRIs with multiple and significant abnormalities, especially in the white matter and the globus pallidus. This is in contrast to other neurotoxic insults and therefore, allows for a tentative diagnosis of CO poisoning on the basis of an anatomical MRI study. Functional SPECT, PET, and MRS studies are frequently abnormal, but not diagnostic of CO poisoning. In other words, the abnormalities seen are compatible with but not diagnostic of CO-induced impairment. Heuser and Mena in 19982 studied more than 70 patients after neurotoxic (pesticides, fumes, perfumes, etc.) exposure and found hypoperfusion in frontal, temporal, and parietal lobes using SPECT. The abnormalities were often asymmetrical, but showed no specific signature after exposure to a given neurotoxin. In other words,

8417: “8417_c020” — 2007/9/19 — 13:23 — page 450 — #2

Neuroimaging after Carbon Monoxide Exposure SY +4

451

CK-BRAIN

03/27/06

Baseline data versus adult normals I Right lateral view

Anterior view

Superior view

Left lateral view

Posterior view

Interactive view-no cerebelhm

+3

+2

-2

-3

-4

-5

FIGURE 20.1 (See Color insert following page 422) A SPECT scan of cortical function after carbon monoxide poisoning. The color scale (left side) displays normal perfusion in gray, subnormal perfusion in green and blue, and hyperperfusion in red. In other cases hypoperfusion is found in the frontal areas. Thus, the abnormalities found vary from one patient to another. (Credit to J. Michael Uszler)

unrelated neurotoxins resulted in similar, if not identical abnormalities on SPECT. This is why we cannot show a typical, that is, diagnostic SPECT or PET brain scan after CO exposure. Figure 20.1 shows a SPECT brain scan of a patient who was chronically exposed to CO and became symptomatic on a long-term basis. This figure is meant to illustrate the advance in technology which now allows us to display SPECT brain scans in an understandable and almost three-dimensional fashion. The technology and display of SPECT was developed by Ismael Mena at UCLA (now in Santiago, Chile).

20.4 REVIEW OF THE RECENT LITERATURE More than 100 publications on neuroimaging after CO exposure were reviewed for this chapter and the results of significant contributions are summarized below.

8417: “8417_c020” — 2007/9/19 — 13:23 — page 451 — #3

452

Carbon Monoxide Poisoning

Hopkins and Woon provide an excellent review in 20065 of structural and functional neuroimaging and neuropsychological evaluation after CO poisoning. They also discuss in some detail the mechanism of CO-induced injury. Globus pallidus lesions are described more frequently than all others.6−22 White matter lesions are described by many authors.9,13,15,18,21−26 Lesions can occur in many other locations such as centrum semiovale, hippocampus, fornix, frontal, temporal, parietal, occipital lobes, cerebellum, and of course, the basal ganglia. These locations are discussed in some detail by Hopkins and Woon.5 While lesions can occur in any part of the brain, they are predominant in the globus pallidus and also in the white matter. Wu et al.27 studied ten patients within 2–5 h after CO exposure and found SPECT to be more sensitive than CT scanning. SPECT scans showed hypoperfusion in seven patients. Denays et al.28 studied 12 patients on admission day and found abnormal SPECT scans in the temporo-parietal-occipital areas in eight patients. Abnormalities were unilateral in some patients, bilateral in others. These findings confirm my earlier comment that lesions after neurotoxic exposure are not always symmetrical. Parkinson et al.29 studied 73 patients during the acute exposure phase and then followed these patients and retested them 6 months later. White matter lesions could not be correlated to the clinical severity. Centrum semiovale lesions were correlated with cognitive impairment—periventricular lesions were not. Devine et al. in 200230 compared MRI and neuropsychological evaluations in one patient who had multiple bilateral lesions in the basal ganglia 15 months after exposure. Porter et al.31 studied MRIs and neuropsychological evaluations on the day of exposure and 6 months thereafter in 62 patients. They found corpus callosum atrophy and described correlation in detail. The neuropsychological test results did not correlate with the level of corpus callosum atrophy. Kim et al.32 studied the diffusivity of white matter lesions in five patients, describing periventricular and centrum semiovale demyelinating hyperintense bilateral lesions. Ku et al. in 20064 describes a case of mania and discusses the possible disinhibition of the frontal lobes as a possible mechanism after CO poisoning (this case had frontal white matter lesions on MRI). This discussion supports my earlier comments on disinhibition from frontal lobe lesions. MRI and/or CT scans were done in conjunction with neuropsychological evaluation by many authors: n = 16,7 n = 5,13 n = 156,14 n = 9,16 n = 21,33 n = 69.34 Sequential studies were performed and discussed: n = 5,13 n = 69,34 n = 2135 and others. Reports from various studies show patients being followed for 80 days,7 3 months,37 6 months,38 3 years,36 4 years,22 6 years,39 9 years,24 10 years,40 and 33 years.23 In one study, MRI and CT scans were performed in 107 patients.18 Parkinsonism was described, but does not necessarily correlate with the location of the anatomical or functional lesions.14,16,21,41 Cerebellar lesions were also described.7,15,23,24,30 Gale et al.33 used MRI, quantitative MRI (QMRI), SPECT, and neuropsychological evaluation in 21 patients and found SPECT and QMRI to be the most sensitive neuroimaging tools for the evaluation of CO poisoning.

8417: “8417_c020” — 2007/9/19 — 13:23 — page 452 — #4

Neuroimaging after Carbon Monoxide Exposure

453

Durak et al. in 20057 describe cortical atrophy, cerebellar atrophy, vermian atrophy, corpus callosum atrophy, having examined eight patients. Neuroimaging changes may be present even if patients are asymptomatic. Kesler et al.34 showed that cognitive impairment can occur in the absence of abnormal imaging studies. However, they specifically describe fornix atrophy associated with verbal memory deficits. Uchino et al.22 examined thirteen patients, 25 years after a coal mine explosion. Globus pallidus and white matter lesions were found and the parieto-occipital region was the most affected cortical area. Seven patients had definite asymmetrical cortical and subcortical lesions. Dunham, et al.42 showed that almost identical exposure in some of their patients, did not lead to the same neuropsychological impairment. Silver et al.18 studied 107 patients. Eight had CT scans. Seven of those had typical globus pallidus and white matter lesions. Bianco and Floris in 19966 described hemorrhagic infarction of the globus pallidus in two patients after CO exposure. Silverman et al.19 also described hemorrhage in the globus pallidus in a patient three years after exposure. Pinkston et al. in a remarkable study in 2000,38 describe abnormal test studies in two adults 3 years after chronic CO exposure and correlate these abnormalities with the neuropsychological evaluation of these same two adults. Neuroimaging studies showed decreased functionality in the pre-frontal cortex and the temporal lobe. Tom et al.20 studied eighteen patients and found globus pallidus lesions in seven patients and white matter lesions in five of those patients. Pach et al.43 studied SPECT, MRS and neuropsychological test results in 10 patients. They found only partial correlation in acute and 6 months postexposure studies. Hurley et al.44 showed an increase in choline and a decrease in N-acetyl-aspartate using MRS in one of their patients. Prockop in a 2005 study16 described abnormal MRS studies in nine patients, some of whom showed decreased N-acetyl-aspartate, especially in the basal ganglia. In some patients the changes were asymmetrical, if not unilateral. MRI and neuropsychological studies were also performed in these patients. Sohn et al.45 studied a husband and wife pair like Pinkston et al.,38 with apparently identical exposure to CO, but different clinical outcomes—only the husband developed parkinsonism. Scanning revealed more severe white matter damage in the husband. Only the wife had bilateral pallidal necrosis. This study illustrates the fact that the same exposure can have different results in different people. No review of this type would be complete without mentioning the important contributions of Lapresle and Fardeau,46 who published anatomical–pathological studies of 22 cases after severe CO poisoning.

20.5 CONCLUSIONS A review of the literature and my own experience show that no single neuroimaging or other test can be used as the one and only diagnostic or prognostic indicator. In other words, a good clinical or otherwise pertinent history, a review of medical records, a physical examination, neuropsychological testing, and neuroimaging

8417: “8417_c020” — 2007/9/19 — 13:23 — page 453 — #5

Carbon Monoxide Poisoning

454

studies (structural and functional) are essential for a correct diagnosis, appropriate treatment, and for thorough follow-up evaluation.

ACKNOWLEDGMENTS Dr Luke Curtis and Sylvia Heuser contributed valuable ideas to the subject discussed in this manuscript. Virginia Salisbury accomplished the difficult and meticulous task of typing the manuscript. Dr. J.M. Uszler of Santa Monica Imaging contributed the SPECT scan in Figure 20.1.

References 1. Choi, I.S.S., Use of scanning techniques in the diagnosis of damage from carbon monoxide, In: Carbon Monoxide Toxicity, David G. Penney, Ed., CRC Press, 2000, pp. 363–380. 2. Heuser, G., and Mena, I. Neurospect in neurotoxic chemical exposure. Demonstration of long-term functional abnormalities, Toxicol. Industr. Health, 14, 813–827, 1998. 3. Heuser, G., and Wu, J.C. Deep subcortical (including limbic) hypermetabolism in patients with chemical intolerance: Human PET studies, In: The role of neural plasticity in chemical intolerance, B.A. Song and I.R. Bell, Ed., Ann. NY Acad. Sci., 933, 319–323, 2001. 4. Ku, B.D., Shin, H.Y., Kim, E.J., Park, K.C., Seo, S.W., and Na, D.L. Secondary mania in a patient with delayed anoxic encephalopathy after carbon monoxide intoxication, J. Clin. Neuroscience, 13, 860–862, 2006 (Epub, 2006, Aug. 28). 5. Hopkins, R.O., and Woon, F.L.M. Neuroimaging, cognitive, and neurobehavioral outcomes following carbon monoxide poisoning, Behav. Cogn. Neurosci. Rev., 5, 141–155, 2006. 6. Bianco, F., and Floris, R. MRI appearances consistent with hemorrhagic infarction as an early manifestation of carbon monoxide poisoning, Neuroradiology, 38 (Suppl. 1), 870–872, 1996. 7. Durak, A.C., Coskun, A., Yikilmaz, A., Erdogan, F., Mavili, E., and Guven, M. Magnetic resonance imaging findings in chronic carbon monoxide intoxication, Acta Radiol. 46, 322–327, 2005. 8. Fine, R.D., and Parker, G.D. Disturbance of central vision after carbon monoxide poisoning, Aust. NZ. J. Ophthalmol., 24, 137–141, 1996. 9. Gotoh, M., Kuyama, H., Asari, S., Ohmoto, T., Akioka, T., and Lai, M.Y. Sequential changes in MR images of the brain in acute carbon monoxide poisoning, Comput. Med. Imaging Graph, 17, 55–59, 1993. 10. Horowitz, A.L., Kaplan, R., and Sarpel, G. Carbon monoxide toxicity: MR imaging in the brain, Radiology, 162, 787–788, 1987. 11. Kanaya, N., Imaizumi, H., Nakayama, M., Nagai, H., Yamaya, K., and Namiki, A. The utility of MRI in acute stage of carbon monoxide poisoning, Intensive Care Med., 18, 371–372, 1992. 12. Kleinert, A., Sinczuk-Walczak, H., and Goraj, B. Acute poisoning by carbon monoxide affecting the extrapyramidal system, Med. Pr., 49, 573–577, 1998. 13. Klostermann, W., Vieregge, P., and Bruckmann, H. Carbon monoxide poisoning; the importance of computed and magnetic resonance tomographic cranial findings for the clinical picture and follow-up, ROFO, 159, 361–367, 1993.

8417: “8417_c020” — 2007/9/19 — 13:23 — page 454 — #6

Neuroimaging after Carbon Monoxide Exposure

455

14. Mimura, K., Harada, M., Sumiyoshi, S., Tohya, G., Takagi, M., Fujita, E., Takata, A., and Tatetsu, S. Longterm follow-up study on sequelae of carbon monoxide poisoning, Seishin Shinkeigaku Zasshi, 101, 592–618, 1999. 15. O’Donnell, P., Buxton, P.J., Pitkin, A., and Jarvis, L.J. The magnetic resonance imaging appearances of the brain in acute carbon monoxide poisoning, Clin. Radiol., 55, 273–280, 2000. 16. Prockop, L.D. Carbon monoxide brain toxicity: clinical, magnetic resonance imaging, magnetic resonance spectroscopy, and neuropsychological effects in 9 people, J. Neuroimaging, 15, 144–149, 2005. 17. Sawa, G.M., Watson, C.P., Terbrugge, K., and Chiu, M. Delayed encephalopathy following carbon monoxide intoxication, Can. J. Neurol. Sci., 8, 77–79, 1981. 18. Silver, D.A., Cross, M., Fox, B., and Paxton, R.M. Computed tomography of the brain in acute carbon monoxide poisoning, Clin. Radiol., 51, 480–483, 1996. 19. Silverman, C.S., Brenner, J., and Murtagh, F.R. Hemorrhagic necrosis and vascular injury in carbon monoxide poisoning; MRI demonstration, Przegl. Lek., 52, 267–270, 1995. 20. Tom, T., Abedon, S., Clark, R.I., and Wong, W. Neuroimaging characteristics in carbon monoxide toxicity, J. Neuroimaging, 6, 161–166, 1996. 21. Tvedt, B., Krogstad, J.M., and Berstad, J. Hypoxic brain damage after carbon monoxide poisoning, Tidsskr Nore Laegeforen, 116, 3005–3008, 1996. 22. Uchino, A., Hasuo, K., Shida, K., Matsumoto, S., Yasumori, K., and Masuda, K. MRI of the brain in chronic carbon monoxide poisoning, Neuroradiology, 36, 399–401, 1994. 23. Coskun, A., Yikilmaz, A., Guven, M., and Erdogan, F. Cranial MR imaging findings of carbon monoxide poisoning in asymptomatic patients, the chronic stage, Tani Girisim Radvol 9, 146–151, 2003. 24. Mascalchi, M., Petruzzi, P., and Zampa, V. MRI of cerebellar white matter damage due to carbon monoxide poisoning: case report, Neuroradiology, 38 (Suppl. 1), 873–874, 1996. 25. Murata, T., Itoh, S., Koshino, Y., Sakamoto, K., Nishio, M., Maeda, M., Yamada, H., Ishil, Y., and Isaki, F. Serial cerebral MRI with FLAIR sequences in acute carbon monoxide poisoning, Comput. Assist. Tomogr., 19, 631–634, 1995. 26. Prockop, L.D., and Naidu, K.A., Brain CT and MRI findings after carbon monoxide toxicity, J. Neuroimaging, 9, 175–181, 1999. 27. Wu, C.I., Changlai, S.P., Huang, W.S., Tsai, C., Lee, C.C., and Kao, C.H., Usefulness of 99mTc ethyl cysteinate dimer brain SPECT to detect abnormal regional cerebral blood flow in patients with acute carbon monoxide poisoning, Nuclear Med. Comm., 24, 1185–1188, 2003. 28. Denays, R., Makhoul, E., Dachy, B., Tondeur, M., Noel, P., Ham, H.R., and Mois, P. Electroencephalographic mapping and 99mTc HMPAO single photon emission computed tomography in carbon monoxide poisoning, Ann. Emerg. Med., 24, 947–952, 1994. 29. Parkinson, R.B., Hopkins, R.O., Cleavinger, H.B., Weaver, L.K., Victoroff, J., Foley, J.F., and Bigler, E.D. White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning, Neurology, 58, 1525–1532, 2002. 30. Devine, S.A., Kirkley, S.M., Palumbo, C.L., and White, R.F. MRI and neuropsychological correlates of carbon monoxide exposure: A case report, Environ. Health Persp., 110, 1051–1055, 2002.

8417: “8417_c020” — 2007/9/19 — 13:23 — page 455 — #7

456

Carbon Monoxide Poisoning 31. Porter, S.S., Hopkins, R.O., Weaver, L.K., Bigler, E.D., and Blatter, D.O. Corpus callosum atrophy and neuropsychological outcome following carbon monoxide poisoning, Arch. Clin. Neuropsychology, 17, 195–204, 2002. 32. Kim, J. Chang, K., Song, I.C., Kim, K.H., Kwon, J., Kim, H.C., Kim, J.H., and Han, M.H. Delayed encephalopathy of acute carbon monoxide intoxication: diffusivity of cerebral white matter lesions, Am. J. Neuroradiol., 24, 1592–1597, 2003. 33. Gale, S.D., Hopkins, R.O., Weaver, L.K., Bigler, E.D., Booth, E.J., and Blatter, D.D. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning, Brain Inj. 13, 229–243, 1999. 34. Kesler, S.R., Hopkins, R.O., Blatter, D.D., Edge-Booth, H., and Bigler, E.D. Verbal memory deficits associated with fornix atrophy in carbon monoxide poisoning, J. Int. Neuropsychol. Soc., 7, 640–646, 2001. 35. Hayashi, R., Hayashi, K., Inouse, K., and Yanagisawa, N. A serial computerized tomographic study of the interval form of CO poisoning, Eur. Neurol., 33, 27–29, 1993. 36. Quattrocolo, G., Leotta, D., Appendino, L., Tarenzi, L., and Duca, S. A case of cortical blindness due to carbon monoxide poisoning, Ital. J. Neurol. Sci., 8, 57–58, 1987. 37. Sawada, Y., Takahashi, M., Ohashi, N., Fusamoto, H, Maemura, K., Kobayashi, H., and Yoshioka, T. Computerised tomography as an indication of long-term outcome after acute carbon monoxide poisoning, Lancet, 1, 783–784, 1980. 38. Pinkston, J.B., Wu, J.C., Gouvier, W.D., and Varney, N.R. Quantitative PET scan findings in carbon monoxide poisoning: deficits seen in a matched pair, Arch. Clin. Neuropsychol., 15, 545–553, 2000. 39. Bruno, A., Wagner, W., and Orrison, W.W. Clinical outcome and brain MRI four years after carbon monoxide intoxication, Acta. Neurol. Scand., 87, 205–209, 1993. 40. Pasquier, F., DePoorter, M.D., Jacquemotte, N. Adnet-Bonte, C. and Petit, H. Cerebellar syndrome after carbon monoxide poisoning, Magnetic Resonance imaging single photon emission tomography, Rev. Neurol. (Paris), 149, 805–806, 1993. 41. Ikeda, K., Sasaki, S., Ichijo, S., Matsuoka, Y., and Irimajiri, S. Atlas of cranial and spinal MRI--magnetic resonance imaging in carbon monoxide poisoning and Parkinsonian syndrome, No To Shinkei, 50, 86–87, 1998. 42. Dunham, M.D., and Johnstone, B. Variability of neuropsychological deficits associated with carbon monoxide poisoning, four case reports, Brain Inj., 13, 917–925, 1999. 43. Pach, D., Urbanik, A., Szczepanska, L., Hubalewska, A., Huszno, B., Groszek, B., and Jenner, B. (99m)Tc-HmPAO single photon emission tomography, magnetic resonance proton spectroscopy and neuropsychological testing in evaluation of carbon monoxide neurotoxicity, Przegl. Lek., 62, 441–445, 2005. 44. Hurley, R.A., Hayman, L.A., and Taber, K.H. Applications of functional imaging to carbon monoxide poisoning, J. Neuropsychiatry Clin. Neurosci., 13, 157–160, 2001. 45. Sohn, Y.H., Jeong, Y., Kim, H.S., Im, J.H., and Kim, J.S. The brain lesion responsible for parkinsonism after carbon monoxide poisoning, Arch. Neurol., 57, 1214–1218, 2000. 46. Lapresle, J., and Fardeau, M. The central nervous system and carbon monoxide poisoning. II. Anatomical study lesions following intoxication with carbon monoxide (22 cases), Prog. Brain Res., 24, 31–74, 1967.

8417: “8417_c020” — 2007/9/19 — 13:23 — page 456 — #8

21

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning S. Gregory Hipskind

CONTENTS 21.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Neuroimaging Modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Anatomical Imaging Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3.1 Computerized Tomography (CT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3.2 Magnetic Resonance Imaging (MRI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.4 Functional Imaging Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.4.1 Positron Emission Tomography (PET) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.4.2 Functional Magnetic Resonance Imaging (FMRI) . . . . . . . . . . . . . . . . 21.4.3 Magnetic Resonance Spectroscopy (MRS) . . . . . . . . . . . . . . . . . . . . . . . 21.4.4 SPECT Findings in Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . 21.5 Neuroimaging Findings in Chronic, Lower-Level CO Poisoning . . . . . . . . 21.6 High Resolution SPECT Findings in Ten Cases of Delayed Carbon Monoxide-Induced Encephalopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

457 458 458 458 459 460 460 461 461 462 464 464 471 473 474

21.1 INTRODUCTION It has been established in this book and earlier books in this series, that carbon monoxide (CO) poisoning is an insidious cause of death and disability in the United States and throughout the world.1 Tissue anoxia is most commonly implicated as the underlying pathophysiologic mechanism of toxic CO exposure as a result of its displacement of oxygen from hemoglobin forming carboxyhemoglobin (COHb). In addition, CO has been shown to have a direct effect on several key biochemical

457

8417: “8417_c021” — 2007/9/19 — 13:23 — page 457 — #1

Carbon Monoxide Poisoning

458

processes by binding directly to hydroxyperoxidase, cytochrome oxidase, and cytochrome p 450 and other key enzymes in the oxidative metabolic process.2,3 Furthermore, it has been shown that CO binds directly to the heme iron in the globus pallidus in substantia nigra.4 Among the various types of neurotoxins, CO produces a unique clinical syndrome in which, after survival of an acute intoxication, a lucid period of variable duration can ensue followed by the onset of delayed neurological symptoms (DNS) in sequelae or a fraction of patients.4 The mechanism for the delayed a fraction effects of CO poisoning is not well understood, but involves elements of reperfusion injury, vascular oxidative stress by generation of reactive oxygen species, lipid peroxidation,5 neuronal exitotoxicity, and apoptosis.6,7 When these molecular changes affect large enough areas of neural tissue, they may be visualized by various neuroimaging techniques. We will see later how the pattern, distribution, and nature of these induced neuropathalogic changes may provide characteristic neuroimaging clues associated with CO poisoning. The role of specific neuroimaging modalities during the various stages of CO poisoning will be reviewed.

21.2 NEUROIMAGING MODALITIES Various neuroimaging modalities have been employed over the years to evaluate the neuropathological changes associated with both acute and chronic CO poisoning. Also see Chapter 20 in this book. Typically, these modalities are divided into two general categories—anatomical or structural studies and functional modalities. Each modality, anatomic and functional, has made its own unique contribution to our understanding of the neuropathological changes associated with CO poisoning. For purposes of this discussion, functional brain imaging will refer to those imaging modalities that assess the level of regional differences in the metabolic activity of brain tissue. Current prominent examples of functional brain imaging include Single Photon Emission Computed Tomography (SPECT), Positron Emission Tomography (PET), Functional Magnetic Resonance Imaging (fMRI), and Magnetic Resonance (MR) Spectroscopy. The primary anatomical modalities include Computerized Tomography (CT) and Magnetic Resonance Imaging (MRI) with its various perturbations—Diffusion Weighted Imaging (DWI) and Fluid Attenuation Inversion Recovery (FLAIR). A summary of the recent functional and anatomical imaging findings in CO poisoning will be presented.

21.3 ANATOMICAL IMAGING FINDINGS 21.3.1 COMPUTERIZED TOMOGRAPHY (CT) The most common finding in acute CO poisoning on CT imaging is symmetrical, diffuse frontal lobe white matter damage (low density findings). In two relatively large series, one by Muira (n = 60) and one by Choi (n = 129), the sensitivity of CT imaging in acute CO poisoning in detecting these changes was shown to be 38.5% and 48.0%, respectively.8,9 The second most common finding was that of low density areas seen in the globus pallidus, occurring in 30% of patients in

8417: “8417_c021” — 2007/9/19 — 13:23 — page 458 — #2

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

459

the Muira study8 and in 25.6% of the patients in the Choi study.9 It was suggested that the low density lesions in the cerebral white matter and globus pallidus were related to necrosis. Although patients with white matter lesions seemed to have a worse prognosis than those with lesions in the globus pallidus, no consistent or reliable prediction of outcome could be established with this imaging modality because of its relatively low sensitivity, particularly for cortical lesions. These suggested neuropathological changes are similar in location and pathology to those found in the pathology studies of Lapresle and Fardeau who performed autopsies on 22 patients who died of acute CO poisoning.10 Decreased CT sensitivity was confirmed in a 2003 study by Wu et al.,11 using 99 m Tc -ECD SPECT in ten patients with acute CO poisoning. All ten patients had negative CT scans. However, brain SPECT imaging showed areas of abnormal hypoperfusion in the basal ganglia of five patients and in the cortical areas of seven patients. The ongoing role and utility of the use of standard CT imaging in the evaluation of CO poisoning continues to be an issue.

21.3.2 MAGNETIC RESONANCE IMAGING (MRI) Although anatomical MRI findings have a slightly superior sensitivity relative to CT, the findings are essentially the same as CT with the exception that, in the 1992 study by Chang,7 additional abnormalities were seen in the subcortical gray matter of the thalamus, putamen, and caudate nucleus. In addition, 9 of 15 patients demonstrated abnormal findings in the globus pallidus bilaterally. Abnormal cortical findings were minimal. Subsequent studies by Gale and Bigler, et al. in 1999,12 using a 1.5 Tesla magnet found cerebral abnormalities in only 38% of 21 patients by MRI who had sustained moderate to severe acute CO poisoning. The majority of the abnormalities detected were subcortical. By contrast, they identified significant abnormalities in 67% of these same patients by using either SPECT or Quantitative Magnetic Resonance Imaging (QMRI). The low sensitivity of MRI for the detection of cortical lesions in CO poisoning somewhat limits it’s utility as either a diagnostic aide or in guiding interventional strategies. However, in a study by Murata et al. in 199513 using serial MRI with FLAIR pulse imaging to follow a single patient with acute CO poisoning, abnormal findings in cerebral white matter were noted which suggested progressive demyelination in spite of normal SPECT and neuropsychological testing. This anecdotal observation seemed to suggest a possible prognostic role for MRI with FLAIR for detecting early neuropathological changes which might require more aggressive interventional strategies. However, in 2005 Durak et al.14 studied 16 patients between 1 and 10 years after their acute episode of CO poisoning using MRI with FLAIR enhancement. All patients initially presented in an unconscious state in the acute setting. At the time of follow-up, eight patients were asymptomatic and eight continued to experience chronic neuropsychiatric sequelae from their acute episode. MRI with FLAIR detected white matter abnormalities in all 16 patients, with lesions noted predominantly in the centrum semiovale area. It thus appears that early evaluation of acute CO poisoning using MRI with FLAIR enhancement carries little prognostic value.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 459 — #3

Carbon Monoxide Poisoning

460

21.4 FUNCTIONAL IMAGING FINDINGS 21.4.1 POSITRON EMISSION TOMOGRAPHY (PET)

993TC

derived rCBF (ml/min/100g)

PET gets its name from positron emitting metabolic substrate analogs, typically [18 F]—Fluorodeoxyglucose (FDG) or [15 O]H2 O. These isotopes are typically shortacting and are generated in cyclotrons. With their capability to directly measure either glucose metabolism or oxygen consumption, regional differences in cerebral metabolic activity can be expressed quantitatively in units of mL/min/100 g brain tissue. Traditionally, PET imaging has been considered superior to SPECT in that it has been shown to achieve intrinsic spatial resolutions of 4–5 mm versus 6–10 mm for SPECT. In addition, PET’s ability to quantify regional differences in cerebral metabolism directly has been considered a more accurate measurement of cerebral function in comparison to SPECT’s relative measurements of regional cerebral blood flow (rCBF). However, because of PET’s increased cost, decreased availability, and the need for cyclotron-generated pharmaceuticals, it is a less practical neuroimaging modality. In addition, with the advent of cheaper, more readily accessible SPECT cameras which can now achieve intrinsic spatial resolutions of 2.3 mm and less (NeuroQuad, NC Systems), PET’s advantage of superior spatial resolution no longer exists. Furthermore, it has been shown that a very close relationship exists between oxygen consumption, glucose metabolism, and CBF. As seen in Figure 21.1, a very tight correlation between CBF and metabolism is seen when inhaled Xenon 133, which measures CBF quantitatively, is compared with 99mTc-HMPAO, the most common radiopharmaceutical used in SPECT studies. Therefore, the differences between these modalities now become somewhat academic. In a 1993 PET study, DeReuck et al.15 demonstrated a global decrease in cerebral metabolic activity, localized primarily in the frontal and temporal cortices. A PET study by Yoshii et al.16 of a patient two months following acute CO poisoning demonstrated findings primarily in the basal ganglia, but with some involvement of the caudate and amputamen. Pinkston et al.17 studied two patients, man and wife, 3 years following their acute exposure. Using statistical parametric mapping to a

120

120

ECD

100

100

80

80

60

60

40

m = 0.85 b=6 r = 0.86

20 0

0

20

40

133Xe

60

80

100 120

rCBF (ml/min/100g)

HMPAO

40

m = 0.83 b = 11 r = 0.93

20 0

0

20

40 133Xe

60

80

100 120

rCBF (ml/min/100g)

FIGURE 21.1 Comparison of the “derived” regional cerebral blood flow (rCBF) for 99m TcECD (left) or 99m Tc HMPAO (right) scanning techniques relative to rCBF (With permission of Ismael Mena).

8417: “8417_c021” — 2007/9/19 — 13:23 — page 460 — #4

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

461

control group they demonstrated decreased metabolic activity in the bilateral orbit frontal, dorsolateral prefrontal, and temporal cortices. Subsequently, Tengvar et al.18 evaluated a patient five months following acute CO poisoning with PET and noted decreased metabolic activity in the frontal lobes, anterior singular gyri, and sub-cortical white matter. In a study of three patients with acute CO poisoning, Ellis-Hon et al.19 noted significant decreases in basal ganglia metabolism in two of three patients using PET imaging. As discussed earlier, it is apparent from the review of the four prior articles which evaluated a total of seven patients only, that PET is not that widely used as a neuroimaging modality for the evaluation of CO poisoning. It remains to be determined whether or not this modality will continue to be utilized in the ongoing neuroimaging evaluations of CO poisoning patients.

21.4.2 FUNCTIONAL MAGNETIC RESONANCE IMAGING (FMRI) FMRI, when employed with blood oxygen level determination (BOLD) enhancement, is able to quantitatively assess regional differences in cerebral metabolic activity. Essentially, by measuring the oxygen extraction between the arterial and venous circulation in various brain regions, with the increased deoxyhemoglobin levels in the venous circulation being detected on T2 images, quantitative measurements of oxygen consumption can be expressed in terms of mL/min/100 g brain tissue. A study by Kondo in 200620 on one patient poisoned with CO, revealed abnormal findings in the deep cortex, hippocampus, and globus pallidus. Hantson21 found symmetrical distribution of deep gray and basal ganglia abnormalities in a study of five patients with delayed effects of CO poisoning. Again, the limited availability of this technology, along with the lack of studies with adequate sample size limit fMRI’s ability to appropriately study the neuropathological changes that occur following recovery from acute CO poisoning.

21.4.3 MAGNETIC RESONANCE SPECTROSCOPY (MRS) This relatively new functional neuroimaging modality uses the response of metabolic protons to an electromagnetic field to evaluate areas of abnormal cerebral activity. It has been shown that the metabolic byproducts of such neuropathological processes as demyelination, neuronal degeneration or ongoing anaerobic glycolysis can result in regional differences in NAA (N-acetyl-aspartate). MRS detected decreases in NAA are associated with neuronal loss or degeneration. Increased choline concentration has been associated with active demyelination. In 1995, Murata et al.13 performed serial MRS, MRI, and SPECT on a patient after the onset of symptoms associated with a delayed CO encephalopathic process. Initially, increased choline was detected which suggested axonal demyelination. The patient was followed with serial MR spectroscopy scans which later revealed increased levels of lactate and NAA which suggested neuronal death. These findings were detected before changes were noted on either MRI or SPECT. In 1998, Sakamoto et al.22 studied three patients with the chronic encephalopathic form of CO poisoning using MRI, electroencephalogram (EEG),

8417: “8417_c021” — 2007/9/19 — 13:23 — page 461 — #5

462

Carbon Monoxide Poisoning

N-isopropyl-p-[1231]iodoarmphetamine (IMP) SPECT, and HMRS. A white matter lesion detected by MRI persisted long after the improvement of the patient’s clinical symptoms. SPECT findings did not always correlate with clinical symptomatology. MRS studies in the frontal lobe area revealed increases in choline-containing compounds and reductions of NAA in all cases. There was close correlation between clinical symptomatology and levels of lactate suggesting a possible role for MRS as a predictor of clinical outcome in patients with the “interval form” of CO poisoning. In 2005, Pach et al.23 studied ten patients following recovery from acute CO poisoning. Neuropsychological testing and 99m Tc –HMPAO-SPECT was performed at 6 months following the acute injury and MRS was performed at 8 months following the acute episode. SPECT revealed significant abnormalities in the frontal cortex, basal ganglia and parietal cortex. All ten patients evaluated by MRS at 8 months post acute CO exposure exhibited elevations in either mobile lipid and/or lactate concentration in the frontal lobes and basal ganglia.

21.4.4 SPECT FINDINGS IN CARBON MONOXIDE POISONING A growing body of scientific information has evolved in the area of the SPECT evaluation of acute CO poisoning. In a study of 12 patients with acute CO poisoning in 1994, Denays et al.24 demonstrated abnormal CBF in the temporal, partial, and occipital lobes in eight of them. In a study of ten acutely poisoned CO patients in 1998, Kao and colleagues25 identified abnormal cortical findings in seven out of ten patients and abnormal basal ganglia findings in six out of ten patients. In 2003, Wu11 found abnormalities in the cortex of seven acutely poisoned patients and abnormalities in the basal ganglia of five out of ten patients. See the chapter on scanning techniques by Choi in Penney, 2000.26 In a 2004 study of 20 acutely poisoned CO patients by Pach et al.,27 brain SPECT imaging with 99 mTc-HMPAO revealed abnormalities in 17 of the 20 patients (85%). Predominant findings in this study included decreases in perfusion in the frontal and parietal cortices and basal ganglia. Their conclusion was that the use of 99 mTcHMPAO brain imaging seemed to be useful in the demonstration of early central nervous system (CNS) dysfunction. A follow-up study by the same investigators of ten patients revealed the same findings. Another study in 2004 of five patients with the chronic effects of CO poisoning by Kon Chu et al.28 showed decreased bilateral white matter, globus pallidus, and frontal cortex. A summary of the SPECT findings noted above is given in Table 21.1. In 1995, Choi29 studied 13 patients with delayed neurological sequelae of CO poisoning and found a diffuse, patchy, decreased pattern which was more prominent in the frontal lobes. Seven of these patients were seen in follow-up six months later at which time their SPECT scans showed increased frontal blood flow in six out of seven patients which was also associated with improved neuropsychological performance. In 1999, Gale and Bigler et al.12 evaluated 21 patients with delayed sequelae of CO poisoning with the multiple modalities of MRI, QMRI, SPECT, and comprehensive neuropsychological testing. In their study, 14/21 patients were noted to have abnormalities on both SPECT and QMRI imaging, but only 38% of patients were noted to have abnormal findings on routine MRI imaging. In the patients with

8417: “8417_c021” — 2007/9/19 — 13:23 — page 462 — #6

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

463

TABLE 21.1 SPECT Areas of Decreased Perfusion Seen in Studies of Carbon Monoxide Poisoning Year

Author

1994

Denays24

1995

Choi26

1998

Kao25

1999

Gale12

2002

Watanabe30

2003

Wu11

2004

Pach23

2004

Chu28

Brain Area Temporal lobe Parietal lobe Occipital lobe Diffuse patchy cortical Bilateral frontal Cortex Basal ganglia Frontal lobes Parietal lobes Bilateral frontal lobes Bilateral insula Right temporal lobe Cortex Basal ganglia Parietal cortex Frontal lobes Basal ganglia Frontal cortex Globus pallidus

Perfusion Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased Decreased

abnormal SPECT findings, the perfusion abnormalities were noted in the frontal lobes (71%), parietal lobes (57%), and temporal lobes (36%). The QMRI findings demonstrated decreased hippocampal size, increased ventricle to brain ratio (VBR) and evidence of cortical atrophy. In 2002, Watanabe et al.30 studied eight patients with delayed neurosequelae of acute CO poisoning and ten patients with no neuropsychiatric sequelae. Using statistical parametric mapping and comparing them to a control group of 44 normal patients, patients with delayed neuropsychiatric sequelae had significantly reduced perfusion in the bilateral frontal lobes, bilateral insula, and right temporal lobe. Interestingly, patients with no neuropsychiatric sequelae had significantly reduced bilateral frontal perfusion as well, worse on the left, compared to controls. This study suggests that relative preservation of left frontal lobe function following acute CO poisoning may have some prognostic value and that right frontal lobe dysfunction may result in more significant neurosequelae. This observation deserves further investigation. In the 2004 study by Chu et al.,28 five patients who manifested delayed hypoxic encephalopathy following acute CO poisoning showed the typical T2 -weighted MRI findings of abnormalities in bilateral periventricular white matter and white matter in the frontal and parieto-occipital areas. In addition, symmetrical abnormalities were found in the globus pallidus. Simultaneous brain SPECT imaging using 99 mTcHMPAO revealed decreased profusion in the bilateral frontal cortex, globus pallidus, and bilateral white matter.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 463 — #7

464

Carbon Monoxide Poisoning

21.5 NEUROIMAGING FINDINGS IN CHRONIC, LOWER-LEVEL CO POISONING Beck’s 1936 axiom is that “No noxious gas so potent when inhaled in atmospheric dilutions of 1% or even less as to cause almost instantaneous death, can be incapable of producing symptoms if inhaled at a lesser concentration for a longer period of time”.31 Population studies have shown that even small increases in the ambient CO concentration can lead to significant increases in mortality,32 low birth weight,33 and psychiatric hospitalizations.34 In a 1998 study by Bayer,35 it was reported that chronic COHb levels between 0.4% and 5.8% were associated with persistent neurological symptoms (PNS). Furthermore, in a 2006 study of chronic CO exposure Tellez and colleagues36 described lack of memory, attention-concentration, and Parkinson type movements as the main neuropsychological and neurological symptoms occurring most frequently. If one assumes that the basic neuropathologic damage of PNS is the same in chronic, lower-level, “occult” CO poisoning as it is in the “interval form” of CO poisoning, it might be anticipated that the neuroimaging findings in these two entities would also be similar. A review of the literature reveals very few neuroimaging studies performed for this entity. A single case of chronic CO exposure in a 2 12 month old infant in 1990 revealed bilateral basal ganglia hypodensities on MRI.37

21.6 HIGH RESOLUTION SPECT FINDINGS IN TEN CASES OF DELAYED CARBON MONOXIDE-INDUCED ENCEPHALOPATHY As much as it is useful to describe the imaging findings of the neuropathological changes which occur in the acute CO poisoning setting from a public health standpoint, the ability to identify those cases of chronic lower-level CO exposure would seem relevant. As has been shown, chronic lower-level CO poisoning often presents clinically as a confusing constellation of seemingly unrelated symptoms which have been variously described as neurasthenia, chronic fatigue syndrome, fibromyalgia, hypochondriasis, and other psychiatric conditions. As such, it often goes undiagnosed and therefore is improperly treated. To the extent that brain function imaging might reveal a characteristic perfusion pattern, it might prove a useful diagnostic tool when the clinical presentation is unclear. This author has had the opportunity to perform high resolution brain SPECT imaging on a series of ten patients over the last 18 months, who were experiencing the long-term effects of acute CO poisoning. A brain dedicated four headed gamma camera (NeuroQuad SPECT, NC Systems Inc.) with an intrinsic spatial resolution of two mm was used. In addition, Mirage software from Segami Corporation which used a normative database allowed comparisons of individual patients on a Brodman area by Brodman area basis for the cortical structures and the subcortical structures of the thalamus, lentiform nucleus, and caudate. A representation of the type of data obtained are seen in Table 21.2.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 464 — #8

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

465

TABLE 21.2 Baseline Data versus Adult Normals

ROI # 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 22

ROI Label

# ELTS

Volume (ml)

Area 1,2,3-Both Area 1,2,3-Left Area 1,2,3-Right Area 4-Both Area 4-Left Area 4-Right Area 5-Both Area 5-Left Area 5-Right Area 6-Both Area 6-Left Area 6-Right Area 7-Both Area 7-Left Area 7-Right Area 8-Both Area 8-Left Area 8-Right Area 9-Both Area 9-Left Area 9-Right Area 10-Both Area 10-Left Area 10-Right

5756 2878 2878 9944 4972 4972 2610 1305 1305 15932 7966 7966 15052 7526 7526 9226 4613 4613 8432 4216 4216 11378 5689 5689

21 10 10 37 18 18 9 4 4 59 29 29 56 28 28 34 17 17 31 15 15 42 21 21

Maximum (sd)

Minimum (sd)

Mean (sd)

Standard deviation (sd)

1.3 1.3 1.3 3.0 3.0 3.0 0.3 0.3 0.5 1.8 1.8 1.4 1.9 1.9 1.7 2.3 2.2 2.3 3.6 3.6 3.3 5.0 5.0 5.0

−5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0 −5.0

−2.6 −3.1 −2.1 −2.6 −2.7 −2.5 −2.7 −2.6 −2.8 −2.4 −2.7 −2.2 −1.3 −1.3 −1.1 −2.2 −2.3 −2.1 −2.4 −2.4 −2.3 −1.7 −1.7 −1.7

1.9 1.8 1.8 1.7 1.7 1.7 1.5 1.8 1.1 1.6 1.6 1.6 1.7 1.7 1.6 2.1 2.1 2.1 2.3 2.2 2.3 2.4 2.3 2.5

Four adults and six children underwent high-resolution brain SPECT imaging with 99mTc-HMPAO. The average age of the adults was 34.8 years (range 31.3–44.6 years) and the average age of the children was 10.3 years (range 7.4–12.8 years). In nine of the patients the average time from acute exposure to imaging was 2.7 years. These nine individuals, six children and three adults, were all apparently exposed to the same level of CO and all were found in an unconscious state. One adult was seen 4.1 years following acute exposure in which unconsciousness was also a factor in the acute setting. All patients were experiencing various delayed neuropsychological sequelae from their acute poisoning and most had been found to have significant abnormalities on neuropsychological testing which primarily involved problems with executive function and memory, as well as such affective disturbances as depression and anxiety. Tables 21.3 and 21.4 represent a summary of their SPECT findings. The following data (Table 21.5) are a summary of the number of statistically abnormal findings in the Brodman areas of the nine patients experiencing the same level of acute CO poisoning.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 465 — #9

Bilateral frontal Bilateral temporal Bilateral parietal Patchy cortical

1, L—2.5 2, L—2.5 3, L—2.5 4, L—2.6, R—2.6 5, L—2.0, R—2.3 6, L—2.2 11, L—4.6, R—4.5 25, L—2.1 28, L—2.14, R—2.1 31, R—1.9

D

E

Bilateral frontal Bilateral occipital Patchy cortical

Bilateral frontal Bilateral temporal L > R Left medial temporal Patchy cortical

21, L—1.7 23, L—1.9, R—2.1 24, L—2.6, R—3.0 28, L—2.3, R—2.3 32, L—2.5, R—2.3

C

21, L—2.2, R—2.4 24, L—3.5, R—3.2 25, L—4.5, R—4.1 28, L—2.7, R—2.4 32, R—2.1 38, L—2.5, R—2.6 47, L—2.9, R—3.1

Bilateral medial temporal Patchy cortical

Bilateral frontal L > R Bilateral temporal L > R Bilateral occipital L > R Patchy cortical

24, L—1.8, R—2.6 32, R—2.3

(Area), (Left value), (Right value) 17, L—1.9, R—2.0 18, L—1.9

A

Cortical Findings

B

Broadman Areas (S.D.)

Patient

TABLE 21.3 Summary of SPECT Findings in 10 individuals designated as A–J

Left globus pallidus Right posterior thalamus Left caudate

Bilateral lentiform Right posterior thalamic Bilateral caudate

Left caudate body

R Globus pallidus Bilateral posterior thalamic R > L

R Globus pallidus

Sub-Cortical Findings

466 Carbon Monoxide Poisoning

8417: “8417_c021” — 2007/9/19 — 13:23 — page 466 — #10

32, L—2.8 36, L—2.7, R—2.9 38, L—2.4, R—2.9 47, R—2.1

20, L—2.3 21, L—2.4 24, L—3.8, R—3.2 25, L—2.6, R—2.6 28, L—3.4, R—2.4 32, L—3.2, R—2.7 20, L—2.1 R—2.0 21, L—2.1 R—0

G

H

Right cerebellar Patchy cortical Bilateral frontal Left medial temporal Right occipital Patchy cortical Bilateral frontal Bilateral cerebellar Bilateral posterior temporal Bilateral occipital

17, L—1.9 24, R—1.9

5, L—2.1 17, L—2.6, R—2.5 28, L—2.0 31, L—2.0, R—2.0

J

Left occipital

Bilateral temporal

Patchy cortical Bilateral frontal

Bilateral Frontal Bilateral Temporal L>R Bilateral Parietal Left Occipital Patchy Cortical

Bilateral orbital frontal Bilateral temporal Bilateral occipital L > R

I

25, L—3.1 R—3.5 28, L—3.6 R—3.7 31, L—2.0

24, L—3.9 R—2.4

36, L—2.6 38, R—2.3 44, L—2.9 R—2.3 45, L—2.1 R—2.5

17, L—2.2, R—1.9 21, L—1.9 24, R—2.0 25, R—2.1 28, L—2.4, R—2.1 38, L—2.5, R—2.0

F

Bilateral globus pallidus Right putamen Left posterior thalamus Left caudate

Bilateral putamen Right thalamus Bilateral globus pallidus

—

Bilateral globus pallidus Bilateral posterior thalamus Left caudate

Bilateral lentiform Left globus pallidus

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

8417: “8417_c021” — 2007/9/19 — 13:23 — page 467 — #11

467

Carbon Monoxide Poisoning

468

TABLE 21.4 Summary of Regional Abnormal Areas in Nine Patients Using High Resolution SPECT Cortical Diffuse, decreased patchiness 8/9

Cerebellar 2/9

Frontal 8/9

Temporal 8/9

Parietal 1/9

Occipital 5/9

Subcortical Globus pallidus 7/9

Caudate 4

Putamen 4/9

Thalamus 5/9

TABLE 21.5 Abnormal Cortical Findings by Age after Carbon Monoxide Exposure Patient

Gender

Age (years)

Number of Abnormal Brodman Areas

C H D Adult average A B E F I J Child average

Male Female Male

31 31 33 31.7 12 11 12 11 7 9 10.3

11 17 15 14.3 3 1 4 9 2 6 4.2

Male Male Male Male Male Male

As can be seen, given the same level of acute CO poisoning, the general cortical and subcortical pattern of involvement was similar in both children and adults, while the number of statistically abnormal Brodman areas in the children were less than those seen in the adults. This finding suggests that children sustain milder neuropathological effects from the same level of acute CO poisoning. Various explanations have been offered for this previously observed phenomenon which are consistent with the findings of other observers that children are more resistant to the acute effects of CO poisoning.38 In addition, the sole adult female in the group experienced a smaller number of statistically abnormal areas than the two adult males. This, too, is consistent with the epidemiological evidence suggesting that women are more resistant than men to the effects of acute CO poisoning.39,40 (see Figures 21.2–21.6).

8417: “8417_c021” — 2007/9/19 — 13:23 — page 468 — #12

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

469

Finding No. 1(cont.): Toxic injury from carbon monoxide poisoning

"CO poisoned"

Normal

FIGURE 21.2 (See color insert following page 422). Example of diffuse neuronal injury two years after acute carbon monoxide poisoning.

Mild

Medium

Severe

FIGURE 21.3 (See color insert following page 422). SPECT scans of three patients with mild, medium and severe cognitive defects two years after acute carbon monoxide poisoning.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 469 — #13

Carbon Monoxide Poisoning

470

Adult male

Adult female

FIGURE 21.4 (See color insert following page 422). SPECT scans of male and female patients two years following acute carbon monoxide poisoning with identical carboxyhemoglobin levels (34.5% vs. 34.9%)

LC

RL

IG

TM

BAW

SP

BRW

FIGURE 21.5 (See color insert following page 422). Superior, transverse views of SPECT perfusion findings in isolated lentiform nuclei of seven patients, two years following acute carbon monoxide poisoning. Yellow areas in the color plates are areas of abnormally decreased perfusion.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 470 — #14

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

ROI label

# Elts

Volume

Maximum

Minimum

471 Standard deviation

Mean

Caudate nucleus—Left

975

0.3 %

71.1 %

26.9 %

42.6 %

9.3 %

Caudate nucleus—Right

975

0.3 %

74.1 %

37.9 %

52.9 %

7.7 %

Right lateral view

Anterior view

Superior view

Left lateral view

Posterior view

Interactive view

FIGURE 21.6 (See color insert following page 422). Six SPECT isolation views of caudate nuclei of a patient two years following acute carbon monoxide poisoning. All areas other than red in the color plates represent areas of abnormally decreased perfusion.

21.7 DISCUSSION CO poisoning, for the purposes of discussion, can be divided into three distinct categories: 1. Acute CO poisoning. This may involve immediate deficits, that is, PNS. 2. Acute CO poisoning which displays delayed effects, that is, DNS (the so-called “interval form” of CO poisoning). 3. Chronic, lower-level or “occult” CO poisoning, which can also result in long-term health effects. In acute exposures, multiple organ systems can be affected with life-threatening results. Suicide by voluntary CO inhalation remains the number one cause of toxic death in the United States.1 Severe acute exposures can result in the development of cardiac arrhythmias, pulmonary edema, renal failure, and metabolic acidosis which must be managed aggressively. High concentrations of CO can cause neuropathologic changes that may manifest themselves immediately or at some interval following the acute, possibly life-threatening episode. Unfortunately, it has been shown that the degree of neuropathologic damage caused by CO does not always correlate with longterm prognosis.41 Although other organ systems can be permanently affected, it is the neurological, neuropsychological, and neuropsychiatric sequelae of CO poisoning that accounts for the bulk of the chronic morbidity often seen in this disorder.42 To the extent that the neuropathological changes that occur in the acute setting are significant and sufficiently widespread, they create “pathological

8417: “8417_c021” — 2007/9/19 — 13:23 — page 471 — #15

472

Carbon Monoxide Poisoning

footprints” that can be identified by various neuroimaging techniques. In the acute setting, from a neuroimaging standpoint, it is important to establish, if possible, the location and extent of the brain injury. This not only aids in the prediction of clinical outcomes, but can assist in the determination of the need for ongoing interventional measures. Perhaps the greatest potential contribution that neuroimaging might make is the ability to aid in the identification of cases of lower-level, chronic CO poisoning. As has been pointed out, chronic CO poisoning has been associated with various vague neuropsychiatric conditions such as neurasthenia, chronic fatigue syndrome, fibromyalgia, and other nonspecific, yet debilitating illnesses which often go undiagnosed and improperly treated.43,44 To the extent that neuroimaging might be capable of identifying certain characteristic perfusion patterns associated with the chronic effects of CO poisoning, improved diagnosis, and better treatment might be possible for these vague syndromes with their protean manifestations. The focus of this chapter, therefore, is the review of current neuroimaging findings in acute CO poisoning, delayed neurosequelae of acute CO poisoning and chronic CO poisoning. In addition, an attempt has been made to assess its current role in the evaluation of the various forms of this insidious illness. As we have seen, CO poisoning, either acute and life threatening, with delayed neuropsychological sequelae or of the lower-level, insidious chronic type, can cause serious neurological damage to various areas of the brain. The damage, from frank necrosis of gray matter to demyelination of white matter, to impaired mitochondrial, and aerobic metabolic processes, results in impaired neuronal activity. This often results in significant neuropsychological disability and morbidity and represents a significant disease burden for our society. Evidence from the anatomic neuroimaging studies, CT and MRI, suggest a regional predilection for CO to damage frontal lobe white matter including corpus callosum and the centrum semiovale as well as subcortical gray areas, particularly the globus pallidus. These changes are often seen earlier in the course of CO poisoning relative to the changes seen with PET or SPECT. CO’s affinity for the high iron content contained within the globus pallidus and substantia nigra is the proposed mechanism whereby the globus pallidus and the substantia nigra are selectively damaged. Subcortical damage has also been seen in the caudate, putamen and thalamus. Regarding CO’s affect on cortical brain matter, numerous studies have shown PET and SPECT’s superiority to CT and MRI. Since PET and SPECT primarily measure rCBF in the more metabolically active gray matter, white matter abnormalities are only occasionally detected. The preponderance of the functional neuroimaging data suggest a predilection of CO to damage the frontal and temporal lobes, often the more medial aspects of the temporal lobes. However, almost every other cortical lobe has been implicated as well including the parietal, occipital, and cerebellar lobes of the brain. MR Spectroscopy may serve as an alarm modality because of its apparent ability to detect active pathologic changes sooner after the onset of the delayed sequelae of CO poisoning as compared to MRI, PET, or SPECT. Further studies are needed to confirm this initial impression. The apparent earlier ability of MRS to detect the onset of neuropsychological symptoms related to CO poisoning may allow for earlier intervention with modalities such as hyperbaric oxygen (HBO). HBO therapy has

8417: “8417_c021” — 2007/9/19 — 13:23 — page 472 — #16

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

473

been shown to be effective in treating many aspects of brain injury, including CO poisoning (see chapter 17 in this book). The data from the various anatomic and functional neuroimaging modalities suggest that a prototypical pattern of CO’s affects on the brain is evolving. This picture appears to be one which demonstrates abnormalities of function in the globus pallidus, bilateral frontal, and bilateral temporal cortices with a tendency toward medial versus lateral temporal involvement. In addition, a patchy cortical pattern of decreased activity as seen in various toxic-anoxic insults to the brain is also observed often but is fairly nonspecific. The proper neuroimaging evaluation of these findings would suggest that a combination of CT and MRI with SPECT or PET is required on the basis of the relative sensitivities and intrinsic spatial resolution anatomical studies for the globus pallidus and the superior sensitivity of functional studies for cortical structures. However, newer generation, high resolution SPECT cameras with intrinsic spatial resolutions of 2.0 mm (NeuroQuad SPECT, NC Systems) may prove useful as a single imaging modality. When one considers this from a radiation exposure standpoint, the ability to do a functional neuroimaging scan with 0.26 rems of radiation exposure versus a CT involving 6.0 rems of radiation exposure, the choice seems obvious. Although there is zero radiation exposure associated with MRI, the ability to avoid two separate procedures and the associated increased cost would also seem preferable. The series of nine patients imaged with the brain-dedicated NeuroQuad SPECT system at Brain Matters, Inc. in Denver, CO confirms its ability to not only assess cortical neuroactivity, but also functional activity in smaller subcortical structures, particularly the globus pallidus. From a public health perspective, the ability of neuroimaging studies to detect CO poisoning of the lower-level chronic type would seem very important. As has been discussed, this insidious malady with its many manifestations often goes undiagnosed and inappropriately treated. Increased awareness and education, particularly in the primary care setting will be important in assisting clinicians in detecting this occult illness in many of their patients who may present with a myriad of seemingly unrelated symptoms. In this setting, clinicians should consider a neuroimaging evaluation to include either a combination of CT or MRI with standard resolution SPECT, or where available, newer generations of high resolution, brain-dedicated SPECT cameras. With the improved diagnostic capabilities of the newer neuroimaging modalities and increased knowledge of CO’s evolving “fingerprint,” earlier intervention and better outcomes after CO poisoning can be achieved.

21.8 CONCLUSIONS 1. Anatomical neuroimaging studies such as CT and MRI are capable of detecting pathological changes in frontal lobe white matter and subcortical gray matter, particularly the globus pallidus, following the development of the neuropsychological sequelae of CO poisoning. 2. Functional neuroanatomical studies such as PET and SPECT appear more sensitive to changes in the regional metabolic activity and rCBF in cortical grey matter, particularly in the frontal and temporal lobes.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 473 — #17

Carbon Monoxide Poisoning

474

3. A prototypical, but not necessarily pathognomonic neuroimaging profile for CO poisoning is evolving which suggests preferential involvement of the globus pallidus subcortically and the frontal and temporal lobes cortically. 4. MR Spectroscopy may serve as an early indicator for evaluating neuropathological changes which develop following the onset of the neuropsychological sequelae of CO poisoning. 5. Neuroimaging studies, either CT or MRI coupled with SPECT or PET, or alternatively high resolution SPECT should be considered as part of the evaluation process for patients presenting subacutely in the primary care setting with various vague, ill-defined symptoms which are often confused with chronic fatigue syndrome, fibromyalgia, flu, Lyme’s disease, multiple sclerosis, psychiatric condition, and so forth.

References 1. Cobb, N., and Etzel, RA. Unintentional carbon monoxide-related deaths in the United States, 1979 through 1988, J. Am. Med. Assoc., 266, 659, 1995. 2. Brown, S., and Piantadosi, C. In vivo binding of CO to cytochrome oxidase in rat brain, J. Appl. Physiol., 68, 604,1990. 3. Goldbaum, L.R., et al. Mechanism of the toxic action of carbon monoxide, Ann. Clin. Lab. Sci. 1976; 6: 372–376. 4. Auer, R.N., Benveniste H. Hypoxia and related conditions. In: Graham DI, Langtos PL (eds): Greenfield’s Neuropathology, London, Arnold, 1997, pp. 275–276. 5. Thom, S.R. Dehydrogenase conversion to oxidase and lipid peroxidation in brain after carbon monoxide poisoning. J. Appl. Physiol., 73: 1584–1589, 1992. 6. Thom, S.R. Carbon monoxide-mediated brain lipid peroxidation in the rat. J. Appl. Physiol., 68, 997, 1990. 7. Chang, K.H. Delayed encephalopathy after acute carbon monoxide intoxication: MR imaging features and distribution of cerebral white matter lesions, Radiology. 1992; 184: 117–122. 8. Miura, T., Mitomo, M., Kawai, R., and Harada, K. CT of the brain in acute carbon monoxide intoxication: characteristic features and prognosis, AJNR, 6, 739, 1985. 9. Choi, I.S., Kim, S.K., Choi, Y.D., Lee, S.S., and Lee, M.S. Evaluation of outcome after acute carbon monoxide poisoning by brain CT, J. Kor. Neurol. Assoc., 8, 78, 1993. 10. Lapresle, J., and Fardeau, M. The central nervous system and carbon monoxide poisoning. II. Anatomical study of brain lesions following intoxication with carbon monixide (22 cases), Prog. Brain Res. 1967; 24: 31–74. 11. Wu, C.I., et al. Usefulness of 99mTc ethyl cysteinate dimer brain SPECT to detect abnormal regional cerebral blood flow in patients with acute carbon monoxide poisoning. Nucl. Med. Commun. 2003; 24: 1185–1188. 12. Gale, S.D., Bigler, E.D., et al. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning, Brain Inj. 1999; 13: 229–243. 13. Murata, T., et al. Serial proton magnetic resonance spectroscopy in a patient with the interval form of carbon monoxide poisoning, J. Neurol. Neurosurg. Psychiatry. 1995; 58: 100–103.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 474 — #18

Recent Advances in Brain SPECT Imaging after Carbon Monoxide Poisoning

475

14. Durak, A.C., et al. Magnetic resonance imaging findings in chronic carbon monoxide intoxication, Acta. Radiol. 2005; 46: 322–327. 15. De Reuck, J., Decoo, D., Lemanhieu, I., Stijckmans, K., Boon, P., Man Maele, G., Buylaert, W., Leys, D., and Petit. H. A positron emission tomography study of patients with acute carbon monoxide poisoning treated by hyperbaric oxygen, J. Neurol., 240, 430, 1993. 16. Yoshii, R., Kozuma, R., Takahashi, W., Haida, M., Takagi, S., and Shinohara, Y. Magnetic resonance imagin and 11c-n-methylspiperone/positron emission tomography studies in a patient with the interval form of carbon monoxide poisoning. Neurol. Sci. 160, 1, 1998. 17. Pinkston, J.B., Wu J.C., and Gouvier W.D. Varney NR. Quantitative PET scan findings in carbon monoxide poisoning: deficits seen in a matched pair. Arch. Clin. Neuropsychol. 2000; 15: 545–553. 18. Tengvar, C., Johansson, B., and Sorensen, J. Frontal lobe and cingulate cortical metabolic dysfunction in acquired akinetic mutism: a PET study of the interval form of carbon monoxide poisoning. Brain Inj. 2004; 18: 615–625. 19. Ellis-Hon, K.L., Yeung W.L., Ho C.H., Leung W.K., Li AM., Chiu-Wing Chu W., and Chan Y.L. Neurologic and radiologic manifestations of three girls surviving acute carbon monoxide poisoning. J Child Neurol. 2006; 21: 737–741. 20. Kondo, A., Saito, Y., Seki, A., Sugiura, C., Maegaki, Y., Nakyama, Y., Yagi, K., and Ohno, K. Delayed neuropsychiatric syndrome in a child following carbon monoxide poisoning. Brain Dev. 2007; 29(3): 174–177. 21. Hanston, P., and Duprez, T. The value of morphological neuroimaging after acute exposure to toxic substances. Toxicol. Rev. 25, 2, 1993. 22. Sakamoto, K., et al. Clinical studies on three cases of the interval form of carbon monoxide poisoning: serial proton magnetic resonance spectroscopy as a prognostic predictor, Psychiatry Res. 1998; 83: 179–192. 23. Pach, D., et al. Evaluation of regional cerebral perfusion using 99mTc-HmPAO single photon emission computed tomography (SPECT) in carbon monoxide acutely poisoned patients, Przegl Lek. 62, 6, 2005. 24. Denays, R., et al. Electroencephalographic mapping and 99mTc HMPAO singlephoton emission computed tomography in carbon monoxide poisoning, Ann. Emerg. Med. 1994; 24: 947–952. 25. Kao, C.H., et al. HMPAO brain SPECT in acute carbon monoxide poisoning, J. Nucl. Med. 1998; 39: 769–772. 26. Choi, I.S.S. Use of scanning techniques in the diagnosis of damage from carbon monoxide. In: Carbon Monoxide Toxicity, D.G. Penney, Ed., CRC Press, 2000, Chapt. 16, pp. 363–380. 27. Pach, D., et al. Evaluation of regional cerebral perfusion using 99mTc-HmPAO single photon emission computed tomography (SPECT) in carbon monoxide acutely poisoned patients, Przegl Lek. 61, 4, 2004. 28. Chu, K., et al. Diffusion-weighted MRI and 99mTc-HMPAO SPECT in delayed relapsing type of carbon monoxide poisoning: evidence of delayed cytotoxic edema, Eur. Neurol. 2004; 51: 98–103. Epub Jan. 28, 2004. 29. Choi, IS., et al. Evaluation of outcome of delayed neurologic sequelae after carbon monoxide poisoning by technetium-99m hexamethylpropylene amine oxime brain single photon emission computed tomography, Eur. Neurol. 1995; 35: 137–142. 30. Watanabe, N., et al. Statistical parametric mapping in brain single photon computed emission tomography after carbon monoxide intoxication, Nucl. Med. Commun. 2002; 23: 355–366.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 475 — #19

476

Carbon Monoxide Poisoning 31. Beck, H.G. Carbon monoxide asphyxiation: a neglected clinical problem, J. Am. Med. Assoc., 107, 1025–1029, 1936. 32. Hexter, A.D., and Goldsmith, J.R. Carbon monoxide: association of community air pollution with mortality, Science, 172, 265–276, 1971. 33. Ritz, B., and Yu, F. The effect of ambient carbon monoxide on low birth weight among children born in southern California between 1989 and 1993, Environ. Health Perspect., 107, 17–25, 1999. 34. Strathilevitz, M., Strahilevitz, A., and Miller, J.E. Air pollutants and the admission rate of psychiatric patients, Am. J. Psychiatr., 136, 205–207, 1979. 35. Bayer, M.J., Orlando, J., McCormick, M.A., Weiner, A., and Deckel, A.W. Persistent neurological sequelae following chronic exposure to carbon monoxide, in Carbon Monoxide: The Unnoticed Poison of the 21st Century, Satellite Meeting, IUTOX VIIIth International Congress of Toxicology, Dijon, France, July 3–4, 1998. 36. Tellez, J., Rodriguez, A., and Fajardo A. Carbon monoxide contamination; an environmental health problem, Rev. Salud Publica, 8, 1, 108–117, 2006. 37. Piatt, J.P., Kaplan, A.M. Bond, G.R., and Berg, R.A. Occult carbon monoxide poisoning in an infant, Pediatr. Emerg. Care, 6, 21, 1990. 38. White, S. Pediatric carbon monoxide poisoning. In: Carbon Monoxide Toxicity, Penney, D.G., Ed., 2000, Chapt. 21, 463–491. 39. Howe, S., Hopkins, R.O., and Weaver, L.K. A retrospective demographic analysis of carbon monoxide poisoned patients [abstract]. Undersea Hyperb. Med., 23 (Suppl.), 84, 1996. 40. Dodson, W.W., Santamaria, J.P., Etzel, R.A., Desautels, D.A., and Bushnell, J.D. Epidemiologic study of carbon monoxide poisoning cases receiving hyperbaric oxygen treatment [abstract], Undersea Hyperb. Med., 24 (Suppl.), 38, 1997. 41. Vieregga, P., Klostermann, W., Blumm, R.G., and Borgis, K.J. Carbon monoxide poisoning: clinical, neurophysiological, and brain imaging observations in acute disease and follow up, J. Neurol. , 236, 478, 1989. 42. Choi, I.S. Peripheral neuropathy following acute carbon monoxide poisoning. Muscle Nerve, 9, 965, 1986. 43. Roy, B. Crawford, R. Pitfalls in diagnosis and management of carbon monoxide poisoning, J. Accid. Emerg. Med. 1996; 13: 62–63. 44. Grace, T.W., and Platt, F.W. Subacute carbon monoxide poisoning; another great imitator, J. Am. Med. Assoc., 246, 1698–1700, 1981.

8417: “8417_c021” — 2007/9/19 — 13:23 — page 476 — #20

22

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning Ramona O. Hopkins

CONTENTS 22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Mechanisms of Brain Injury. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3 Cognitive Sequelae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.1 Persistent and Delayed Neuropsychological Sequelae. . . . . . . . . . . . 22.3.2 Cognitive Impairments in Lower Level Carbon Monoxide Poisoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.3 Markers of Carbon Monoxide Poisoning Severity and Outcome 22.3.4 Effect of HBO on Cognitive Impairments. . . . . . . . . . . . . . . . . . . . . . . . . 22.3.5 Relationship between Cognitive Sequelae and Neuroimaging Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.6 Functional Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4 Affective Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4.1 Depression and Anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4.2 Obsessive Compulsive Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4.3 Kluver–Bucy Syndrome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

477 478 478 481 483 484 484 485 486 487 487 488 488 488 489

22.1 INTRODUCTION Carbon monoxide (CO) is a colorless, odorless, tasteless, and nonirritating gas produced as a by-product of combustion of carbon-containing compounds. CO is the leading cause of poisoning injury and death worldwide,1 and the most common cause of accidental and intentional poisoning in the United States. CO results in approximately 40,000 emergency department visits2 and 470 unintentional deaths per year in the United States.3 The brain and heart are particularly vulnerable 477

8417: “8417_c022” — 2007/9/11 — 12:14 — page 477 — #1

Carbon Monoxide Poisoning

478

to the pathological effects of CO.1 Neurologic morbidity following CO poisoning includes neurologic sequelae,4 abnormalities on brain imaging,4 affective changes,5 and cognitive impairments.4

22.2 MECHANISMS OF BRAIN INJURY Exposure to CO may damage multiple organ systems with high oxygen utilization, especially the cardiovascular and central nervous systems. For more detail on this topic, see the Raub et al. review published in 2000.1 The mechanisms of CO-induced neural damage are complex and multifactorial. Although the neuropathological injury associated with CO poisoning are related to CO-induced hypoxia (CO binds to hemoglobin),6 other biochemical mechanisms appear to be involved. Mechanisms of brain injury following CO poisoning include: binding to intracellular proteins cytochrome c-oxidase, myoglobin, or cytochrome c (P-450) reductase, leading to mitochondrial dysfunction and disruption of cellular metabolism.7 Other mechanisms include hypoxia,8 release of excitatory amino acids (e.g., glutamate) resulting in calcium influx and cell damage or death,9 interference with intracellular enzyme function,10 lipid peroxidation leading to oxidative injury,11 deposition of peroxynitrate and subsequent blood vessel endothelium damage,12 oxidative stress from intracellular iron deposition,13 and apoptosis.14 CO-mediated oxidative stress alters myelin basic protein, resulting in an immune response and inflammation in the central nervous system.15 Consistent with the multifactorial neuropathologic mechanisms, the resultant neuropathology, cognitive, affective, and neurobehavioral sequelae are heterogeneous.

22.3 COGNITIVE SEQUELAE Cognitive impairments frequently occur following CO poisoning in healthy individuals.16 It is estimated that between 15% and 49% of individuals with diagnosed CO poisoning will develop cognitive sequelae.17 Some CO-poisoned people will develop persistent neuropsychological sequelae (i.e., initial impairments that persist over time);16 however, others who have intact initial cognitive performance may present with delayed neurologic or cognitive sequelae (i.e., normal initial cognitive scores with impairment developing from 2 to 40 days post-CO poisoning; see discussion below). CO-related cognitive sequelae are heterogeneous regarding onset, severity, and cognitive domain affected.4 A list of some CO-related cognitive impairments are shown in Table 22.1. Common CO-related cognitive sequelae include impaired memory,18 executive function,19 slow mental-processing speed, decreased intellectual function,4 apraxia, aphasia, and agnosia.20 Cognitive sequelae lasting 1 month21,22 or more16,23 occurs in 25–50% of participants with loss of consciousness (LOC) or carboxyhemoglobin (COHb) levels greater than 25%.22,24 Even people with less severe CO poisoning may develop cognitive impairments.25 Recent studies utilizing comprehensive standardized neuropsychological tests find significant cognitive impairments both immediately and several years after recovery from the

8417: “8417_c022” — 2007/9/11 — 12:14 — page 478 — #2

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

TABLE 22.1 Carbon Monoxide-Related Cognitive Impairments Shown by Cognitive Domain Cognitive Domain Arithmetic Attention

Executive function

Intelligence Memory

Motor

Processing speed Spatial

Verbal

Visual

Impairments Acalculia Distractibility Divided attention Preservative errors Sustained attention Decision making Disorganization Impulsivity Planning Working memory Verbal intelligence Performance intelligence Anterograde memory Delayed memory Recall Recognition memory Retrograde memory Short term or working memory Apraxia Athetosis Ballism Bradykenesia Chorea Dyskenesia Dystonia Incoordination Myoclonus Parkinsonism Rigidity Tremor Mental-processing speed Visuoconstruction Visuoperception Visuospatial Aphasia Dysarthria Hypophonia Mutism Achromotopsia Apperceptive agnosia Cortical blindness Homonymous hemianopsia Prosopagnosia Scotomas Visual form agnosia

8417: “8417_c022” — 2007/9/11 — 12:14 — page 479 — #3

479

Carbon Monoxide Poisoning

480

initial CO-poisoning symptoms.4,26 The cognitive impairments range in severity from mild, to moderate, and severe. While impairments in memory, attention, and executive function occur most frequently, a consistent pattern of neuropsychological deficits, or a CO “syndrome,” has not been observed. A recent review of the literature from 1995 to 2005 assessed cognitive impairments following CO poisoning identified by standardized neuropsychological tests.27 Hopkins and Woon27 identified 18 group studies that included 979 CO-poisoned patients and 16 case studies that included 35 CO-poisoned patients, for a total of 1014 patients. The mean age of the patients was 38.1 years and 40.2 years for the group and case studies, respectively. While some studies included elderly individuals (age >65 years), the majority of studies contained predominately young to middle age adults. CO-poisoning severity was moderate to severe, with a mean COHb level of 23% (normal ≤2.0%) for both the group and case studies. LOC occurred in 30.1% (295/979) of patients in the group studies and 42.8% (15/35) of patients in the case studies, with a range of 10.8–100% for the group studies. All of the studies reviewed demonstrated the adverse effect of CO poisoning on cognition. The cognitive impairments occurred in multiple cognitive domains and were heterogeneous regarding onset, severity, and cognitive domain affected. Of the group studies, impaired memory was the most frequently reported impairment, followed by impaired attention, motor impairments, executive dysfunction, slow mental-processing speed, and impaired visual spatial abilities. Similarly, in the case studies, memory impairments occurred most frequently, followed by executive dysfunction, impaired attention, motor impairments, visual spatial deficits, and slow mental-processing speed (Figure 22.1). These data indicate that memory is the most common cognitive domain affected by CO poisoning.27 Memory impairments following CO poisoning can be mild, moderate, or severe. For example, a 48-year-old male with LOC and a COHb of 9.1%

Case studies

Processing speed

Cognitive impairments

Group studies Visual spatial Executive function Motor Attention Memory 0

20

40 60 Percent of studies reporting

80

100

FIGURE 22.1 Percent of group and case studies reporting cognitive impairments following carbon monoxide poisoning.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 480 — #4

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

481

had severe memory impairments associated with hippocampal atrophy (structure in the temporal lobe associated with the formation and recall of memories) 6 months postCO poisoning.28 Memory impairments following CO poisoning include impaired short-term memory, anterograde memory, retrograde memory, recall, and recognition memory.28−31 While memory impairments usually include the inability to learn or remember new information, global amnesia (both anterograde and retrograde memory impairments) are reported.31 There was the case of a patient with CO-induced amnesia and bilateral hippocampal atrophy who recovered recognition memory, but remained impaired for verbal and spatial memory over time.32 This suggests that some types of memory may improve, whereas other memory impairments may not improve over time. Other CO-induced cognitive impairments included impaired attention, executive function, motor, visual spatial, and slowed mental-processing speed. Significant deficits in visual tracking, visuomotor skills, visuospatial planning, and abstract thinking occur following CO poisoning.25 Impaired executive function, attention and concentration, visual–perceptual abilities, and information processing speed are also reported.33 Although the common wisdom has been that most people who survive initial CO poisoning will recover, recent studies suggest that the cognitive impairments they incur may last years and sometimes become permanent. The identification of CO-related neurocognitive deficits may be instrumental for future studies in determining if they are lessened through therapy, such as cognitive rehabilitation or medications. In addition to the more common cognitive impairments such as impaired memory and executive function, CO poisoning can produce other impairments such as visual agnosia.34 As an example, a 34-year-old female (patient DF) who sustained severe CO poisoning developed visual form agnosia (e.g., inability to visually recognize objects), owing to bilateral diffuse damage to the ventral portion of the lateral occipital regions, including the ventral visual pathway.34 DF was unable to recognize objects, especially line drawings of common objects. She could not discriminate between vertical and horizontal line gratings or between simple geometric shapes.35 Her color vision was intact and she was able to draw objects from memory, but could not recognize the objects she drew. Despite the problem with object recognition, she had no problem adjusting her finger-thumb grip to the width of objects.36 She was adept at interacting manually with objects and she used the structural features of objects to control visually guided grasping movements, showing the capacity to act upon visual information that she was unable to report at a conscious level.36 Further, she was able to negotiate obstacles when walking though a room.35 The ability to interact with objects or avoid obstacles when walking was due to preservation of the dorsal visual pathway. Thus, she had impaired ability to visually recognize objects but, preserved visual processing for object orientation and location of objects in a room, even though she denied “seeing” the objects.

22.3.1 PERSISTENT AND DELAYED NEUROPSYCHOLOGICAL SEQUELAE Neurological or cognitive sequelae can occur immediately and persist over time (“persistent neurocognitive sequelae,” PNS), or the onset can be delayed in its

8417: “8417_c022” — 2007/9/11 — 12:14 — page 481 — #5

482

Carbon Monoxide Poisoning

onset (“delayed neurocognitive sequelae” or the so-called interval form, DNS).16,22,23 Delayed neurologic or neurocognitive sequelae presentation includes a lucid period of between 2 and 40 days after CO poisoning onset.37 A number of studies have assessed CO-associated DNS,17,22,23,38 however few studies recognize or assess PNS.16 PNS and DNS are common following acute CO poisoning. Delayed neuropsychological sequelae occur in 0.06–40% of CO poisonings.17,38 Symptoms of DNS include mental deterioration, urinary and/or fecal incontinence, gait disturbance and other neurologic problems such as cognitive impairments.39 There was the case of a 50-year-old male, who sustained LOC owing to CO poisoning and developed DNS 1 month later with behavioral changes, disorientation, impaired memory, masked face, hypophonia, muscle rigidity, and bradykinesia manifested by slow movements and gait disturbance.40 At 4 months all symptoms had resolved. Other studies show that DNS persists over time.23 There are a paucity of data regarding risk factors for DNS, however, COHb levels are not associated with the severity of the symptoms.41 The etiology(s) of PNS and DNS are unknown, although it has been hypothesized that DNS is due to extensive myelin and neuronal loss.42 Other theories include immuopathological damage due to activation of polymorphonuclear leukocytes leading to demyelination and dopaminergic and serotonergic disturbances.43 One method that appears to be sensitive to the underlying pathophysiology of DNS is magnetic resonance spectroscopy (MRS). In proton MRS of the brain, signals are present from N-acetylaspartate (located primarily in neurons, a marker for neurons and axons), choline (principally phosphotidyl choline, a membrane constituent), and creatine (used as an internal standard because its level is usually stable). For comparison purposes, N-acetylaspartate and choline are expressed relative to creatine. MRS appears to be sensitive to CO-poisoning related changes in white matter.44 Data show that N-acetylaspartate/creatine were below normal and choline/creatine were elevated at the time of DNS symptom onset, yet both structural magnetic resonance (MR) and cerebral blood flow measures were normal.45 Increased severity of the spectral changes at the onset of DNS was related to more profound clinical symptoms. Clinical recovery correlated with normalization of the spectral changes.44 Thus, proton MRS may provide a much-needed marker of the development of DNS following CO poisoning. Weaver and colleagues46 assessed both PNS and DNS in a prospective outcome study. Participants (N = 238) with acute CO poisoning were followed prospectively. Persistent neuropsychological sequelae were defined as cognitive dysfunction initially, which persisted at 2 weeks and 6 weeks after CO poisoning. Delayed neuropsychological sequelae are defined as a decline of at least one standard deviation on a neuropsychological subtest score from a prior score, and meeting the definition for cognitive sequelae at 6 weeks.46 Thirty-seven percent of participants had cognitive sequelae at 6 weeks, of which 59% had PNS and 28% had DNS, a ratio of 2:1. Hyperbaric oxygen (HBO) reduced the incidence of PNS but not DNS.46 One possible explanation for the effect of HBO on PNS, but not on DNS is that HBO favorably modulates mechanisms of early brain injury but does not influence mechanisms that may be involved in the development of DNS.15

8417: “8417_c022” — 2007/9/11 — 12:14 — page 482 — #6

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

483

22.3.2 COGNITIVE IMPAIRMENTS IN LOWER LEVEL CARBON MONOXIDE POISONING Many clinicians believe that only patients who experience LOC37 or have moderate to severe CO poisoning (COHb > 25%)47 will develop sequelae. A survey of hyperbaric medical centers in North America indicated that nearly all treating facilities would use HBO for a CO-poisoned patient with a COHb level of 40% and a COHb level of 25% was identified most often as an indication for HBO therapy.48 While most studies of cognitive impairments following CO poisoning include patients with moderate to severe poisoning, information is accumulating regarding cognitive impairments in patients with lower-level or less severe CO poisoning. Less severe CO poisoning has been defined using a variety of criteria. It has been defined as COHb level of ≤10%,49 a COHb level of 10%,50 a COHb of 5–15%,47 or COHb of 0.01–11%.25 Most studies agree that less severe CO poisoning occurs at COHb levels of 15%. About 55 patients had less severe and 201 had more severe CO poisoning. Cognitive sequelae occurred in 35% versus 39% for patients with less severe versus more severe CO poisoning, respectively. There was no difference in the prevalence of cognitive sequelae (P = 0.91) in patients

8417: “8417_c022” — 2007/9/11 — 12:14 — page 483 — #7

484

Carbon Monoxide Poisoning

with less severe compared to the more severe CO-poisoning at 6 weeks, 6 and 12 months. Regardless of less severe or more severe CO poisoning, CO-poisoned patients had significant cognitive sequelae.54 These data suggest that CO-related cognitive outcomes may be independent of poisoning severity.

22.3.3 MARKERS OF CARBON MONOXIDE POISONING SEVERITY AND OUTCOME One common belief is that markers of CO-poisoning severity, such as LOC and COHb levels are good predictors of patient outcome. One study found the length of LOC was related to outcome55 and development of DNS.37 Alternatively, COHb level is not a reliable predictor of CO-poisoning severity, symptoms, or neurologic outcome.55−57 Similarly, COHb levels do not correlate with severity of poisoning or cognitive outcomes.58−60 For example, the rate of cognitive sequelae in patients with severe CO poisoning with mean COHb levels of 25.2 ± 9.2% did not differ from that of patients with less severe CO poisoning with mean COHb levels of 6.8 ± 4.7%.54 Other studies have shown that COHb levels are not associated with cognitive deficits.61 Alternatively markers of poisoning severity such as LOC, duration of coma, elevated COHb, or duration of exposure do not predict cognitive sequelae.23 Several studies have assessed the relationship between markers of CO poisoning severity (LOC and COHb levels) with brain imaging findings and cognitive impairments. Hopkins et al.62 found that LOC was not required for the development of cognitive sequelae following CO poisoning. Another study found no association between corpus callosum atrophy and COHb and or LOC. Even with significantly elevated COHb levels and LOC in approximately 50% of the patients, these markers (e.g., COHb level and LOC) did not correlate with the presence of corpus callosum atrophy or development of cognitive impairments.62 Similarly, neither fornix atrophy nor verbal memory impairments correlate with COHb or LOC following CO poisoning.61 Findings in primates indicate that neither severity nor duration of CO-exposure is related to the severity of white matter damage.63 In summary, neither symptoms of poisoning, cognitive impairment4,23 white matter hyperintensities,64 fornix atrophy,61 or corpus callosum atrophy65 are related to COHb levels or LOC. The lack of association between COHb levels and cognitive and neuropathological outcomes raises the question as to why this may be the case. One possible explanation is that the lack of association between COHb and outcome measures is due part in to the variability in measured COHb levels in CO-poisoned individuals. The variability in measured COHb levels is directly related to the delay in removal from the CO environment to medical treatment and the amount and duration of supplemental oxygen given prior to COHb measurement.66,67 Thus, the time to COHb measurement and supplemental oxygen impacts the result in decreased COHb levels.

22.3.4 EFFECT OF HBO ON COGNITIVE IMPAIRMENTS The usual treatment for acute CO poisoning is 100% normobaric oxygen, commonly delivered by a reservoir nonrebreathing face-mask, or by HBO.41,68 HBO

8417: “8417_c022” — 2007/9/11 — 12:14 — page 484 — #8

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

485

therapy is recommended for patients with acute CO poisoning, especially if they have experienced LOC, or have severe poisoning.2,41,69,70 Comparisons of the several randomized clinical trials in the treatment of acute CO poisoning is difficult owing to methodological differences.22,23,71,72 A recent prospective double-blind, randomized treatment trial found that HBO therapy reduced cognitive sequelae by 46% at 6 weeks compared to normobaric oxygen.16 Both groups of participants improved over time, but the difference in cognitive sequelae was maintained at 12 months. Those patients with cognitive sequelae had moderate to severe cognitive impairments, falling below the 16th percentile of the normal distribution.29 They communicate and perform activities of daily living normally, but find activities that require executive function, memory, and attention/concentration skills as challenging or impossible.16 Similar findings of benefit for HBO have been reported.23 Alternatively, Scheinkestel et al.71 reported that HBO might worsen outcome in CO-poisoned patients. The study by Weaver et al.16 differed substantially from that of Scheinkestel et al. regarding the number of intubated patients, CO exposure duration, time from the end of the CO exposure to HBO therapy, randomization methods (equal proportions versus cluster), follow-up rate (97% versus 46%), suicide rate (31% versus 69%), statistical analyses, and oxygen treatment protocols.71

22.3.5 RELATIONSHIP BETWEEN COGNITIVE SEQUELAE AND NEUROIMAGING FINDINGS CO poisoning may result in focal and generalized neuroanatomical abnormalities observed on MR and Computed Tomography (CT) imaging. Brain lesions following CO poisoning occur in cortex,73 cerebellum,74 thalamus,75 and substantia nigra.76 Subcortical lesions are found in the white matter77 and basal ganglia including the globus pallidus,78 caudate and putamen.79,80 White matter hyperintensities are common in the periventricular and centrum semiovale or deep white matter regions.64 Generalized atrophy of white matter structures like the corpus callosum65 and white matter degeneration in the temporal, parietal and occipital regions are reported post-CO poisoning.73 A prospective study in consecutive CO-poisoned participants demonstrated that white matter lesions occur more frequently than basal ganglia lesions.64 While some studies indicate that white matter lesions are the most common lesion following CO poisoning,77,81 others indicate basal ganglia lesions are the most common.82,83 Alternative evidence supporting basal ganglia lesions as the most frequent CO-related lesion comes from a recent review that found globus pallidus lesions occurred in 32–86% of patients.84 Early (within 24 h post-CO poisoning) brain imaging can be normal with basal ganglia lesions observed on subsequent scans.64 Alternatively, basal ganglia lesions can occur within the first day following CO poisoning.85 Basal ganglia lesions have been reported at 1 month,86 6 months,31 1-year,87 2 years,86 4 years,74 and 5 years88 post-CO poisoning. It appears that COHb levels are not associated with the development of basal ganglia lesions, as the lesions occurred with COHb levels as low of 9.1%28 and as high as 54%.89

8417: “8417_c022” — 2007/9/11 — 12:14 — page 485 — #9

Carbon Monoxide Poisoning

486

In addition to neural lesions, CO poisoning may cause neuronal cell loss and concomitant structural atrophy. Atrophy has been reported in the fornix,61 hippocampus,4 and corpus callosum.65 Generalized atrophy can occur with brain volume reduction manifested by reduced gyral volume, increased sulcal space, and passive ventricular enlargement.4 A recent study found CO-poisoned patients had atrophy in the putamen, caudate, and globus pallidus in the absence of basal ganglia lesions.90 The relationship between neuropathologic findings and cognitive impairments has only recently been assessed. Slow mental-processing speed and impaired memory were associated with smaller putamen and globus pallidus volumes in CO-poisoned patients. Impaired verbal memory was associated with fornix atrophy,61 while slow mental-processing speed was associated with white matter hyperintensities64 following CO poisoning. Thus, basal ganglia atrophy, fornix atrophy, and white matter hyperintensities likely all contribute to the observed cognitive impairments in CO victims. There are significant correlations between neuropsychological impairments and abnormalities in cerebral perfusion, clinical MR, and/or brain volumetric measures in CO poisoned patients.4 A study of consecutively CO-poisoned patients using quantitative MR (brain volumetric measures) compared neural volumes from the initial MR scan (e.g., day of CO poisoning) with MR 6 months post CO exposure. The patients had atrophic changes of the fornix, corpus callosum, and basal ganglia 6 months postexposure compared to their initial MR scans that correlated with cognitive impairments.61,65 Alternatively, MR and CT structural imaging carried out at the onset of the symptoms of DNS found no association between the symptoms and the imaging abnormalities. Pavese et al.91 found 50% of patients (11/22) had abnormalities on MR imaging, whereas 27% of the patients had adverse symptoms 1 month after the CO poisoning. Other authors report many patients with delayed symptoms (30–42%) have normal neuroimaging examinations.92,93

22.3.6 FUNCTIONAL IMAGING Measures of cerebral blood flow may be more sensitive to CO poisoning-related neural changes than standard structural imaging. Decreased glucose metabolism on positron emission tomography (PET),94 hypoperfusion on single photon emission computed tomography (SPECT),4,95 and abnormal electroencephalography (EEG)20,21 parallel the focal and diffuse changes observed with structural imaging following CO poisoning. Regional cerebral blood flow (using PET) abnormalities may also be present in the absence of abnormalities on CT or MR.96 Neuropsychological impairments are associated with abnormalities in cerebral perfusion, clinical MR, and/or brain volumetric measures.4 A SPECT study found patchy hypoperfusion in patients that developed delayed neurological sequelae.95 Parkinsonian symptoms are associated with decreased perfusion of the basal ganglia and cognitive deficits are associated with decreased cerebral blood flow in cortical areas.96 CO poisoningrelated cerebral blood flow abnormalities predict poor outcome (death, remote memory impairment).97 In contrast, cerebral blood flow abnormalities did not predict outcome 3–5 days after CO poisoning.98

8417: “8417_c022” — 2007/9/11 — 12:14 — page 486 — #10

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

487

TABLE 22.2 Carbon Monoxide-Related Affective and Neurobehavioral Changes Affective Sequelae Depression Anxiety Apathy Irritability Mood swing Elated mood Neurobehavioral Changes Obsessive and compulsive behaviors Delusions Hallucinations Violent outbursts Fear Disinhibition Anger

22.4 AFFECTIVE DISORDERS 22.4.1 DEPRESSION AND ANXIETY Some affective and behavioral changes associated with CO poisoning are shown in Table 22.2. Affective and personality changes following CO poisoning appear to be heterogeneous regarding time of onset, severity, and duration.26,89,99,100 While depression is frequently reported following CO poisoning, fewer studies have assessed anxiety.4,26,92,101 Other psychological and personality changes following CO poisoning include obsessive and compulsive behavior,26,92,102,103 delusions and hallucinations,20,21,92 violent outbursts,20 fear,26 and elated mood.21 Jasper and colleagues found depression and anxiety in 45% of CO-poisoned patients at 6 weeks, 44% at 6 months, and 43% at 12 months following CO poisoning.5 The prevalence of depression and anxiety following CO poisoning varied by study from a low of 33% to a high of 100%.4,99,104 Differences in patient populations, patient selection, affective measures, and length of follow-up may account for the between study differences. The consistent between study findings include the high rate of depression and anxiety in CO-poisoned people. Similar prevalence rates of depression and anxiety occur in other pulmonary disorders with depression reported in 25–28% of patients with cardiac and pulmonary disorders105,106 and anxiety in 10–40% of patients with pulmonary disorders.107,108 Accidentally CO-poisoned patients are as likely as those with intentional CO poisoning to have depression and anxiety at 6 and 12 months.5 Other studies have reported significant depression and anxiety in participants who attempt suicide with CO.99,104 Hay et al.99 found that depression in CO-poisoned participants was

8417: “8417_c022” — 2007/9/11 — 12:14 — page 487 — #11

Carbon Monoxide Poisoning

488

similar to control patients who had a psychiatric disorder. Smith and Brandon109 found a higher prevalence of depression and personality changes in patients with intentional CO poisoning (35%) compared with accidentally poisoned patients (21%) and psychiatric controls (9%). Thus both intentional and accidentally CO-poisoned patients are at risk to develop depression and anxiety. CO-poisoned patients with cognitive impairments may develop CO morbid depression and anxiety.4,74 Alternatively, depression following CO poisoning can occur in the absence of cognitive impairments.78,109 Mori et al.78 described a patient with no prior history of psychiatric disorders who developed dramatic personality changes in the absence of cognitive deficits following accidental CO poisoning. Smith and Brandon109 found that 33% of CO-poisoned patients developed personality and affective morbidity, but only 11% developed cognitive impairments. CO poisoning appears to result in a high rate of depression and anxiety.

22.4.2 OBSESSIVE COMPULSIVE DISORDER Symptoms of obsessive-compulsive disorder include behavior mannerisms associated with Tourettes syndrome and obsessive thoughts. Compulsive stereotypic routines may develop in patients with basal ganglia lesions.102 A case of obsessive-compulsive disorder secondary to CO poisoning was reported in the 34-year-old male.102 Shortly after hospital discharge, the patient reported an irresistible urge to spit and pull the hair on his legs. He developed obsessive thoughts of harming his friends and family, fear of germ contamination, and sexual fantasies. His compulsive behaviors included repetitive hand-washing, checking behaviors, and counting rituals. Neuroimaging showed bilateral globus pallidus lesions. Thus, CO poisoning can lead to acquired obsessive-compulsive disorder with concomitant bilateral globus pallidus lesions.102

22.4.3 KLUVER–BUCY SYNDROME Muller110 reported the case of an 18-year-old female with LOC due to CO poisoning. Her COHb was 14%. During rehabilitation she had oral tendencies (i.e., put everything into her mouth), decreased social distance, and object agnosia. She had flattened affect and appeared placid and indifferent toward people and events. She was diagnosed with a Kluver–Bucy-like syndrome. Neuropsychological testing showed cognitive deficits similar to dementia. Lesions on brain imagining were located in the lateral temporal lobes sparing the hippocampus. The Kluver-Bucy like symptoms resolved 6 months later, but the cognitive deficits including impaired attention, distractibility, and amnesia persisted.110

22.5 CONCLUSION CO poisoning is common, often goes unrecognized and may result in significant morbidity. Morbidity following CO poisoning includes neurologic sequelae, neuropathologic abnormalities on brain imaging, affective sequelae, and cognitive impairments. Morbidity appears to be independent of poisoning severity as measured

8417: “8417_c022” — 2007/9/11 — 12:14 — page 488 — #12

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

489

by COHb level and LOC. Most CO poisoning is preventable, therefore the associated morbidity is also preventable. Given the high rate of brain related morbidity and the fact that the majority of CO poisoning is avoidable, awareness and prevention of CO exposure is warranted. With increased awareness of the dangers and causes of CO exposure, and with the availability of CO alarms, CO poisoning and its adverse effects may be significantly reduced.

References 1. Raub, J.A. et al. Carbon monoxide poisoning: a public health perspective, Toxicology, 145, 1, 2000. 2. Hampson, N.B., Ed. Hyperbaric Oxygen Therapy: 1999 Committee Report Revised, Undersea and Hyperbaric Medical Society, Kensington, MD, 1999. 3. Piantadosi, C.A. Carbon monoxide poisoning, N. Engl. J. Med., 347, 1054, 2002. 4. Gale, S.D. et al. MRI, quantitative MRI, SPECT, and neuropsychological findings following carbon monoxide poisoning, Brain Inj., 13, 229, 1999. 5. Jasper, B.W. et al. Affective outcome following carbon monoxide poisoning: A prospective longitudinal study, Cogn. Behav. Neurol., 18, 127, 2005. 6. Okeda, R. et al. Comparative study on pathogenesis of selective cerebral lesions in carbon monoxide poisoning and nitrogen hypoxia in cats, Acta. Neuropathol. (Berl). 56, 265, 1982. 7. Piantadosi, C.A. Carbon monoxide, oxygen transport, and oxygen metabolism, Undersea Hyperb. Med., 2, 27, 1987. 8. Caine, D. and Watson, J.D. Neuropsychological and neuropathological sequelae of cerebral anoxia: A critical review, J. Int. Neuropsychol. Soc., 6, 86, 2000. 9. Jarrard, L.E. and Meldrum, B.S. Selective excitotoxic pathology in the rat hippocampus, Neuropathol. Appl. Neurobiol., 19, 381, 1993. 10. Coburn, R.F. Mechanisms of carbon monoxide toxicity, Prev. Med., 8, 310, 1979. 11. Thom, S.R. Antagonism of carbon monoxide-mediated brain lipid peroxidation by hyperbaric oxygen, Toxicol. Appl. Pharmacol., 105, 340, 1990. 12. Thom, S.R. et al. Vascular nitrosative stress from carbon monoxide (CO) exposure, Undersea Hyperb. Med., 25 (Suppl.), 47, 1998. 13. Piantadosi, C.A. Carbon monoxide poisoning, Undersea Hyperb. Med., 31, 167, 2004. 14. Piantadosi, C.A. et al. Apoptosis and delayed neuronal damage after carbon monoxide poisoning in the rat, Exp. Neurol., 147, 103, 1997. 15. Thom, S.R. et al. Delayed neuropathology after carbon monoxide poisoning is immune-mediated, Proc. Natl. Acad. Sci. U. S. A., 101, 13660, 2004. 16. Weaver, L.K. et al. Hyperbaric oxygen for acute carbon monoxide poisoning, N. Engl. J. Med., 347, 1057, 2002. 17. Myers, R.A., DeFazio, A. and Kelly, M.P. Chronic carbon monoxide exposure: A clinical syndrome detected by neuropsychological tests, J. Clin. Psychol., 54, 555, 1998. 18. Hopkins, R.O., Weaver, L.K. and Kesner, R.P. Long-term memory impairments and hippocampal magnetic resonance imaging in carbon monoxide poisoned subjects., Undersea Hyperb. Med., 20, 15, 1993. 19. Gale, S.D. and Hopkins, R.O. Effects of hypoxia on the brain: Neuroimaging and neuropsychological findings following carbon monoxide poisoning and obstructive sleep apnea, J. Int. Neuropsychol. Soc., 10, 60, 2004.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 489 — #13

490

Carbon Monoxide Poisoning 20. Garland, H. and Pearce, J. Neurological complications of carbon monoxide poisoning, Q. J. Med., 36, 445, 1967. 21. Min, S.K. A brain syndrome associated with delayed neuropsychiatric sequelae following acute carbon monoxide intoxication, Acta. Psychiatr. Scand., 73, 80, 1986. 22. Raphael, J.C. et al. Trial of normobaric and hyperbaric oxygen for acute carbon monoxide intoxication, Lancet, 2 (8660), 414, 1989. 23. Thom, S.R. et al. Delayed neuropsychologic sequelae after carbon monoxide poisoning: Prevention by treatment with hyperbaric oxygen, Ann. Emerg. Med., 25, 474, 1995. 24. Gorman, D.F. et al. A longitudinal study of 100 consecutive admissions for carbon monoxide poisoning to the Royal Adelaide Hospital, Anaesth. Intensive Care, 20, 311, 1992. 25. Amitai, Y. et al. Neuropsychological impairment from acute low-level exposure to carbon monoxide, Arch. Neurol., 55, 845, 1998. 26. Dunham, M.D. and Johnstone, B. Variability of neuropsychological deficits associated with carbon monoxide poisoning: Four case reports, Brain Inj., 13, 917, 1999. 27. Hopkins, R.O. and Woon, F.L. Neuroimaging, cognitive, and neurobehavioral outcomes following carbon monoxide poisoning, Behav. Cogn. Neurosci. Rev., 5, 141, 2006. 28. Gottfried, J.A. and Chatterjee, A. Carbon monoxide-mediated hippocampal injury, Neurology, 57, 17, 2001. 29. Bastin, C. et al. Dissociation between recall and recognition memory performance in an amnesic patient with hippocampal damage following carbon monoxide poisoning, Neurocase, 10, 330, 2004. 30. Bourgeois, J.A. Amnesia after carbon monoxide poisoning, Am. J. Psychiatry, 157, 1884, 2000. 31. Carlesimo, G.A. et al. Selective sparing of face learning in a global amnesic patient, J. Neurol. Neurosurg. Psychiatr., 71, 340, 2001. 32. Henke, K. et al. Memory lost and regained following bilateral hippocampal damage, J. Cogn. Neurosci., 11, 682, 1999. 33. Helffenstein, D. Neuropsychological evaluation of the carbon monoxide-poisoned patient, In: Carbon Monoxide Toxicity, Penney, D.G., Ed., CRC Press, Boca Raton, FL, 2000, Chapt. 20, pp. 439–461. 34. Farah, M. Visual Agnosia, MIT Press, Cambridge, MA, 1990. 35. Patla, A.E. and Goodale, M.A. Obstacle avoidance during locomotion is unaffected in a patient with visual form agnosia, Neuroreport, 8, 165, 1996. 36. Goodale, M.A. et al. A neurological dissociation between perceiving objects and grasping them, Nature, 349 (6305), 154, 1991. 37. Choi, I.S. Delayed neurologic sequelae in carbon monoxide intoxication, Arch. Neurol., 40, 433, 1983. 38. Shillito, F., Drinker, C. and Chaughness, T. The problem of nervous and mental sequelae in carbon monoxide poisoning, JAMA, 106, 669, 1936. 39. Choi, I.S. Carbon monoxide poisoning: systemic manifestations and complications, J. Korean Med. Sci., 16, 253, 2001. 40. Yoshii, F. et al. Magnetic resonance imaging and C-N-methylspiperone/positron emission tomography studies in a patient with the interval form of carbon monoxide poisoning, J. Neurol. Sci., 160, 87, 1998. 41. Ernst, A. and Zibrak, J.D. Carbon monoxide poisoning, N. Engl. J. Med., 339, 1603, 1998. 42. Yant, W. et al. Studies in asphyxia, US Public Health Bull., 211, 1, 1934.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 490 — #14

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

491

43. Hurley, R.A. et al. Applications of functional imaging to carbon monoxide poisoning, J. Neuropsychiat. Clin. Neurosci., 13, 157, 2001. 44. Sohn, Y.H. et al. The brain lesion responsible for parkinsonism after carbon monoxide poisoning, Arch. Neurol., 57, 1214, 2000. 45. Kamada, K. et al. Cerebral metabolic changes in delayed carbon monoxide sequelae studied by proton MR spectroscopy, Neuroradiology, 36, 104, 1994. 46. Weaver, L.K. et al. Persistent and delayed carbon monoxide cognitive sequelae, Undersea Hyperb. Med., Vol. 32(4), No. 144, 289, 2005. 47. Bleecker, M.L. and Lindgren, K.N. The mere presence of low levels of carboxyhemoglobin is not causal proof for altered neuropsychological performance, Arch. Neurol., 56, 1299, 1999. 48. Hampson, N.B. et al. Selection criteria utilized for hyperbaric oxygen treatment of carbon monoxide poisoning, J. Emerg. Med., 13, 227, 1995. 49. Sadovnikoff, N., Varon, J. and Sternbach, G.L. Carbon monoxide poisoning. An occult epidemic, Postgrad. Med., 92, 86, 1992. 50. Crawford, R., Campbell, D.G. and Ross, J. Carbon monoxide poisoning in the home: Recognition and treatment, Br. Med. J., 301 (6758), 977, 1990. 51. Horvath, S.M., Dahms, T.E. and O’Hanlon, J.F. Carbon monoxide and human vigilance. A deleterious effect of present urban concentrations, Arch. Environ. Health, 23, 343, 1971. 52. Wright, J. Chronic and occult carbon monoxide poisoning: We don’t know what we’re missing, Emerg. Med. J., 19, 386, 2002. 53. Guarnieri, M. Cave painting hazard?, Science, 283 (5410), 2019, 1999. 54. Chambers, C., Hopkins, R.O. and Weaver, L.K. Cognitive and affective outcomes compared dichotomously in patients with acute carbon monoxide poisoning, Undersea Hyperb. Med., Vol. 33(5), 338–339, 2006. 55. Jain, K.K. Carbon Monoxide Poisoning, Warren H. Green, Inc., St. Louis, 1990. 56. Levy, D.E. et al. Predicting outcome from hypoxic-ischemic coma, JAMA, 253, 1420, 1985. 57. Martindale, L.G. Carbon monoxide poisoning: The rest of the story, J. Emerg. Med., 15 (2 Pt 1), 101, 1989. 58. Camporesi, E.M. Hyperbaric Oxygen Therapy: A Committee Report, Undersea and Hyperbaric Medical Society, Bethesda, MD, 1996. 59. Weaver, L.K. et al. Double-blind, controlled, prospective randomized clinical trial (RCT) in patients with acute carbon monoxide poisoning: Outcome of patients treated with normobaric oxygen or hyperbaric oxygen (HBO2) - an interim report., Undersea Hyperb. Med., 22 (Suppl.), 14, 1995. 60. Winter, P.M. and Miller, J.N. Carbon monoxide poisoning, JAMA, 236, 1502, 1976. 61. Kesler, S.R. et al. Verbal memory deficits associated with fornix atrophy in carbon monoxide poisoning, J. Int. Neuropsychol. Soc., 7, 640, 2001. 62. Hopkins, R.O. et al. Severe anoxia with and without concomitant brain atrophy and neuropsychological impairments, J. Int. Neuropsychol. Soc., 1, 501, 1995. 63. Ginsberg, M.D., Myers, R.E. and McDonagh, B.F. Experimental carbon monoxide encephalopathy in the primate: II. Clinical aspects, neuropathology, and physiologic correlation, Arch. Neurol., 30, 209, 1974. 64. Parkinson, R.B. et al. White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning, Neurology, 58, 1525, 2002. 65. Porter, S.S. et al. Corpus callosum atrophy and neuropsychological outcome following carbon monoxide poisoning, Arch. Clin. Neuropsychol., 17, 195, 2002.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 491 — #15

492

Carbon Monoxide Poisoning 66. Sokal, J.A. and Kralkowska, E. The relationship between exposure duration, carboxyhemoglobin, blood glucose, pyruvate and lactate and the severity of intoxication in 39 cases of acute carbon monoxide poisoning in man, Arch. Toxicol., 57, 196, 1985. 67. Sokal, J.A. The effect of exposure duration on the blood level of glucose, pyruvate and lactate in acute carbon monoxide intoxication in man, J. Appl. Toxicol., 5, 395, 1985. 68. Weaver, L.K. Carbon monoxide poisoning, Crit. Care Clin., 15, 297, 1999. 69. Tibbles, P.M. and Edelsberg, J.S. Hyperbaric-oxygen therapy, N. Engl. J. Med., 334, 1642, 1996. 70. Piantadosi, C.A. Diagnosis and treatment of carbon monoxide poisoning, Respir. Care Clin. N. Am., 5, 183, 1999. 71. Scheinkestel, C.D. et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: A randomized controlled clinical trial, Med. J. Aust., 170, 203, 1999. 72. Ducasse, J.L., Celsis, P. and Marc-Vergnes, J.P. Non-comatose patients with acute carbon monoxide poisoning: Hyperbaric or normobaric oxygenation?, Undersea Hyperb. Med., 22, 9, 1995. 73. Uchino, A. et al. MRI of the brain in chronic carbon monoxide poisoning, Neuroradiology, 36, 399, 1994. 74. Bruno, A., Wagner, W. and Orrison, W.W. Clinical outcome and brain MRI four years after carbon monoxide intoxication, Acta Neurol. Scand., 87, 205, 1993. 75. Chang, K.H. et al. Delayed encephalopathy after acute carbon monoxide intoxication: MR imaging features and distribution of cerebral white matter lesions, Radiology, 184, 117, 1992. 76. Kawanami, T. et al. The pallidoreticular pattern of brain damage on MRI in a patient with carbon monoxide poisoning, J. Neurol. Neurosurg. Psychiatr., 64, 282, 1998. 77. Watanabe, N. et al. Statistical parametric mapping in brain single photon computed emission tomography after carbon monoxide intoxication, Nucl. Med. Commun., 23, 355, 2002. 78. Mori, E. et al. Isolated athymhormia following hypoxic bilateral pallidal lesions, Behav. Neurol., 9, 17, 1996. 79. Hsiao, C.L., Kuo, H.C. and Huang, C.C. Delayed encephalopathy after carbon monoxide intoxication—long-term prognosis and correlation of clinical manifestations and neuroimages, Acta Neurol. Taiwan, 13, 64, 2004. 80. Martinez Bermejo, A. et al. [Bilateral hypodensity of the basal ganglia. Clinicoevolutionary correlation in children], Rev. Neurol., 33, 101, 2001. 81. Smallwood, P. and Murray, G.B. Neuropsychiatric aspects of carbon monoxide poisoning: A review and single case report suggesting a role for amphetamines, Ann. Clin. Psychiatry, 11, 21, 1999. 82. Pracyk, J.B. et al. Brain computerized tomography after hyperbaric oxygen therapy for carbon monoxide poisoning., Undersea Hyperb. Med., 22, 1, 1995. 83. Roohi, F., Kula, R.W. and Mehta, N. Twenty-nine years after carbon monoxide intoxication, Clin. Neurol. Neurosurg., 103, 92, 2001. 84. Hopkins, R.O. et al. Basal ganglia lesions following carbon monoxide poisoning, Brain Inj., 20, 273, 2006. 85. Gottfried, A.W. Intellectual consequences of perinatal anoxia, Psychol. Bull., 80, 231, 1973. 86. Stuppaeck, C.H. et al. Akathisia induced by necrosis of the basal ganglia after carbon monoxide intoxication, Mov. Disord., 10, 229, 1995.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 492 — #16

Neurocognitive and Affective Sequelae of Carbon Monoxide Poisoning

493

87. Shimosegawa, E. et al. Cerebral blood flow and glucose metabolism measurements in a patient surviving one year after carbon monoxide intoxication, J. Nucl. Med., 33, 1696, 1992. 88. Vion-Dury, J. et al. Sequelae of carbon monoxide poisoning: An MRI study of two cases, J. Neuroradiol., 14, 60, 1987. 89. Jaeckle, R.S. and Nasrallah, H.A. Major depression and carbon monoxide-induced parkinsonism: Diagnosis, computerized axial tomography, and response to L-dopa, J. Nerv. Ment. Dis., 173, 503, 1985. 90. Pulsipher, D.T., Hopkins, R.O. and Weaver, L.K. Basal ganglia volumes following carbon monoxide poisoning: A prospective longitudinal study, Undersea Hyperb. Med., 33, 245, 2006. 91. Pavese, N. et al. Clinical outcome and magnetic resonance imaging of carbon monoxide intoxication. A long-term follow-up study, Ital. J. Neurol. Sci., 20, 171, 1999. 92. Lee, M.S. and Marsden, C.D. Neurological sequelae following carbon monoxide poisoning clinical course and outcome according to the clinical types and brain computed tomography scan findings, Mov. Disord., 9, 550, 1994. 93. Choi, I.S. and Cheon, H.Y. Delayed movement disorders after carbon monoxide poisoning, Eur. Neurol., 42, 141, 1999. 94. de Reuck, J. et al. A positron emission tomography study of patients with acute carbon monoxide poisoning treated by hyperbaric oxygen, J. Neurol., 240, 430, 1993. 95. Choi, I.S. et al. Evaluation of outcome of delayed neurologic sequelae after carbon monoxide poisoning by technetium-99m hexamethylprophylene amine oxime brain single photon emission computed tomography, Eur. Neurol., 35, 137, 1995. 96. Kao, C.H. et al. HMPAO brain SPECT in acute carbon monoxide poisoning, J. Nucl. Med., 39, 769, 1998. 97. Turner, M. and Kemp, P.M. Isotope brain scanning with Tc-HMPAO: A predictor of outcome in carbon monoxide poisoning?, J. Accid. Emerg. Med., 14, 139, 1997. 98. Sesay, M. et al. Regional cerebral blood flow measurements with Xenon-CT in the prediction of delayed encephalopathy after carbon monoxide intoxication, Acta Neurol. Scand. Suppl., 166, 22, 1996. 99. Hay, P.J. et al. The neuropsychiatry of carbon monoxide poisoning in attempted suicide: A prospective controlled study, J. Psychosom. Res., 53, 699, 2002. 100. Vieregge, P. et al. Carbon monoxide poisoning: Clinical, neurophysiological, and brain imaging observations in acute disease and follow-up, J. Neurol., 236, 478, 1989. 101. Jefferson, J. Subtle neuropsychiatric sequelae of carbon monoxide intoxication: Two case reports, Am. J. Psychiatry, 133, 961, 1976. 102. Escalona, P.R. et al. Obsessive-compulsive disorder following bilateral globus pallidus infarction, Biol. Psychiatry, 42, 410, 1997. 103. Lugaresi, A. et al. "Psychic akinesia" following carbon monoxide poisoning, Eur. Neurol., 30, 167, 1990. 104. Skopek, M.A. and Perkins, R. Deliberate exposure to motor vehicle exhaust gas: The psychosocial profile of attempted suicide, Aust. N. Z. J. Psychiatry, 32, 830, 1998. 105. Silverstone, P.H. Prevalence of psychiatric disorders in medical inpatients, J. Nerv. Ment. Dis., 184, 43, 1996. 106. Silverstone, P.H. et al. The prevalence of major depressive disorder and low selfesteem in medical inpatients, Can. J. Psychiatry., 41, 67, 1996. 107. Karajgi, B. et al., The prevalence of anxiety disorders in patients with chronic obstructive pulmonary disease, Am. J. Psychiatry, 147, 200, 1990.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 493 — #17

494

Carbon Monoxide Poisoning

108. Pollack, M.H. et al., Prevalence of panic in patients referred for pulmonary function testing at a major medical center, Am. J. Psychiatry, 153, 110, 1996. 109. Smith, J.S. and Brandon, S. Morbidity from acute carbon monoxide poisoning at three-year follow-up, Br. Med. J., 1, 318, 1973. 110. Muller, N.G. and Gruber, O. High-resolution magnetic resonance imaging reveals symmetric bitemporal cortical necrosis after carbon monoxide intoxication, J. Neuroimaging, 11, 322, 2001.

8417: “8417_c022” — 2007/9/11 — 12:14 — page 494 — #18

23

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide Poisoning: A Retrospective Study and Case Presentation Dennis A. Helffenstein

CONTENTS 23.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.1.2 Incidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.1.3 Physical, Cognitive and Emotional/Affective Sequelae of Chronic Carbon Monoxide Poisoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.1.4 Published Chronic Carbon Monoxide Studies Utilizing Neuropsychological Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2 Helffenstein Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.1 Sample and Exposure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.1.1 Admission Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.1.2 Sample Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.1.3 Source and Location of Exposure . . . . . . . . . . . . . . . . . . . . . . . 23.2.1.4 Level of Carbon Monoxide Exposure . . . . . . . . . . . . . . . . . . . 23.2.1.5 Duration and Frequency of Exposure . . . . . . . . . . . . . . . . . . . 23.2.1.6 Time Since Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.2 Battery Administered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.3 Norms Utilized . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.4 Symptoms Experienced by Participants During Carbon Monoxide Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.5 Persisting Symptoms Reported by Participants . . . . . . . . . . . . . . . . . . .

496 496 497 498 502 506 506 506 507 507 507 507 508 508 509 509 511 495

8417: “8417_c023” — 2007/9/11 — 12:14 — page 495 — #1

Carbon Monoxide Poisoning

496

23.2.5.1 Physical Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.5.2 Visual Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.5.3 Cognitive Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.5.4 Psychological / Behavioral Symptoms . . . . . . . . . . . . . . . . 23.2.6 Neuropsychological Testing Outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.1 Index Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.2 IQ Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.3 Halstead–Reitan Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.4 Memory Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.5 Academic Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.6 Visual–Visual Perceptual Testing . . . . . . . . . . . . . . . . . . . . . . 23.2.6.7 Speed of Information Processing . . . . . . . . . . . . . . . . . . . . . . 23.2.6.8 Other Motor Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.6.9 Miscellaneous Tests of Executive Function . . . . . . . . . . . 23.2.6.10 Language Comprehension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.7 Minnesota Multiphasic Personality Inventory-2 . . . . . . . . . . . . . . . . . . 23.2.8 Vocational Outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.9 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1 Patient Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2 Exposure Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.3 Educational History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.4 Persisting Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.5 Results from Initial Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.6 Neuropsychological Re-evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.6.1 Circumstances of Re-evaulation . . . . . . . . . . . . . . . . . . . . . . . 23.3.6.2 Self-reported Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.6.3 Comparison of Test Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.6.4 Results from Re-evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.7 Summary of Case Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.8 Takeaway Messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4 Addendum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

513 513 515 515 516 516 518 522 522 523 525 527 528 529 530 531 536 538 540 540 540 540 541 541 542 542 542 543 543 545 545 546 547 548

23.1 BACKGROUND 23.1.1 Introduction Carbon monoxide (CO) is the most common cause of poisoning in the United States and may result in neuropathologic changes which in turn result in a wide range of cognitive, visual, affective, and neurologic sequelae. The effects of moderate to severe acute CO poisoning have been well documented in the literature. For example, Gale et al.1 utilized neuropsychological testing in addition to brain imaging techniques [i.e., Single Photon Emission Computed Tomography (SPECT),

8417: “8417_c023” — 2007/9/11 — 12:14 — page 496 — #2

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

497

Magnetic Resonance Imaging (MRI), and quantitative Magnetic Resonance Imaging (qMRI)] to identify the patient’s residual neuropsychological and neuropathological impairments. Ninety-three percent of the sample group showed cognitive impairments on neuropsychological testing, including difficulties with attention, memory, executive function, and mental processing speed. Ninety-five percent of the patients were experiencing ongoing problems with depression and anxiety. The results of the imaging studies revealed that 38% of the patients had abnormal MRI scans, 67% had abnormal SPECT scans, and 67% had abnormal qMRI findings. The qMRI technique identified hippocampal atrophy and/or diffuse cortical atrophy. The SPECT identified cerebral profusion deficits, most notably in the frontal and temporal lobes. Significant correlation was identified between the neuroimaging techniques and deficits noted on neuropsychological testing. The effects of chronic CO poisoning are less well researched and documented. Most CO poisoning studies to date have focused on short-term effects of a one time, lower-level exposure to CO in experimental settings or on the long-term effects of accidental acute CO poisoning. As with any toxin, there are three components in determining the severity of exposure. These include the level or amount of toxin the individual is exposed to, the frequency of exposure, and the duration of each exposure. One difficulty faced in evaluating the effects of chronic CO poisoning is that it is often difficult to calculate the above variables with any degree of certainty. However, the health risks of exposure to lower levels of CO repeatedly or for an extended duration should not be minimized. Wright2 states, “There is a strong possibility that low level exposure to CO is responsible for widespread and significant morbidity. However, the clinical syndrome produced is often overlooked because of a range of presentation, obscure symptoms, and a lack of awareness of the problem” (p. 387). There is now a growing body of evidence, which clearly shows that chronic exposure to CO can, and often does, result in permanent neurological, cognitive, and visual dysfunction. Regarding this issue, Penney3 states, “Clearly, prolonged exposure to this poison even at what were previously thought to be ultra-low levels is capable of producing many and varied residual health effects. Furthermore, the incidence of such unpleasant and often debilitating effects is far higher than was previously believed by the medical and public health community and can continue for a very long period of time” (pp. 414–415). Indeed, there is some evidence to suggest that because of the repeated exposures, which typically occur in cases of chronic CO poisoning, the pathophysiological changes and damage to the brain may actually be more significant than in cases of acute poisoning.4

23.1.2 INCIDENCE CO exposure is the most common cause of death by poisoning in the US and results in an estimated 40,000 emergency room (ER) visits per year.5 Typically, these are acute poisoning incidents which have resulted in severe medical problems requiring emergency intervention. Townsend and Maynard6 note, “The Institute for Environment and Health commented in their 1998 publication on CO that, ‘It is likely that many more sub-acute CO intoxications occur than are brought to the attention of medical practitioners.’ Because of the difficulty in recognizing the effects of exposure

8417: “8417_c023” — 2007/9/11 — 12:14 — page 497 — #3

498

Carbon Monoxide Poisoning

to low concentrations of CO, there is currently limited knowledge on the size of the problem” (p. 708). These authors go on to note that there are two key questions, which must be addressed. One, the number of people potentially affected by low levels of CO in their homes, workplaces or other settings, and, two, the likelihood of long-term negative health effects. With regard to the second question, there is some evidence to suggest that a significant proportion (43%) of individuals who experience a chronic exposure to CO have permanent neurological sequelae at 3-year follow-up.7 The number of individuals potentially affected by low levels of CO is more difficult to estimate, but is certainly more common than acute poisoning events. Hampson8 points out that the signs and symptoms of CO poisoning are nonspecific and that under-diagnosis in ERs is well described in the literature. He goes on to point out that not all patients chronically exposed to CO are treated in ERs and are often seen in medical offices or clinics and, therefore, their statistics would not appear in the national database. It is also noted that these patients will often attribute their nonspecific symptoms to alternate causes (e.g., viral illness) and not seek medical attention. This latter event is typically referred to as “occult CO poisoning.”2 The problem with accurately diagnosing chronic CO poisoning has been well documented by other authors. Regarding this issue, Penney3 states, “For every single case of chronic CO poisoning reported/successfully diagnosed, there are ten cases that go unreported/undiscovered/undiagnosed” (pp. 396–397). The answer to the question regarding the frequency of chronic CO poisoning remains elusive. Heckerling et al.9 estimated that 3–5% of individuals who present to urban ERs for headaches and dizziness might well be victims of chronic CO poisoning. Whatever the number of chronic occult CO poisonings, it is clearly a significant health risk in the United States. Regarding this issue, Halpern10 states, “Chronic occult CO poisoning is a diagnosis that is not frequently recognized in patients seen initially in an ER or by a primary care provider. It is not readily recognized because of a limited history, vague and variable clinical presentation and a failure of emergency care providers to suspect the cause of symptoms. It is a serious and potentially lethal condition and should be suspected whenever a person is seen at a health care facility for multiple ‘flu-like’ complaints, especially during the winter months when homes are heated, or in the late fall when furnaces are started” (p. 107).

23.1.3 PHYSICAL, COGNITIVE AND EMOTIONAL/AFFECTIVE SEQUELAE OF CHRONIC CARBON MONOXIDE POISONING To date, there has been several retrospective survey studies conducted concerning the residual or persisting symptoms associated with chronic CO poisoning. These include two studies conducted by Penney3 and a comprehensive questionnaire study of individuals chronically exposed to CO in the United Kingdom conducted by CO Support in 1996, a registered charity, headed by Ms. Debbie Davis. The CO Support Study was originally published as a technical paper in October 1997,11 and a detailed summary of that study was later published by Hay et al.12 Penney3 conducted two retrospective studies of chronic CO poisoning. Study A included data from 66 individuals who had sustained chronic CO poisoning, defined as CO exposure lasting more than 24 h. Data were obtained through the Internet.

8417: “8417_c023” — 2007/9/11 — 12:14 — page 498 — #4

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

499

Duration of exposure, as reported by 29 participants ranged from 0.18 to 120 months. The mean duration of exposure was 30.7 months plus or minus 6.0 months (SEM). Air CO concentrations reported in 15 instances were 427.8 ppm, plus or minus 115.2 ppm. Carboxyhemoglobin (COHb) level, as reported in 11 instances was 9.65%, plus or minus 2.46%. Study A assessed symptoms that respondents indicated they experienced during their chronic CO exposure, as well as their residual or persisting symptoms. Respondents reported 103 different symptoms in total during their exposure. The typical respondent reported multiple symptoms in multiple systems. A partial list of these symptoms appears in Table 23.1. Penney notes that the misdiagnosis rate in cases of chronic CO poisoning is very high. He believes chronic CO poisoning is not better recognized because “It almost invariably presents with too many disparate, seemingly unrelated and for the most part, nonspecific symptoms. This tends to confuse physicians who act mainly on pattern recognition of one or a few symptoms to come up with a probable diagnosis, or at least a ‘short list.’ The result of being presented with 5, 10, or 15 or more symptoms is likely to yield a diagnosis of hypochondriasis (faking), psychiatric condition, or both” (p. 395). When an ER or family physician is presented with a history of multiple symptoms in multiple systems, it is easy to understand how they may misdiagnose chronic CO poisoning. Halpern10 emphasizes that, “A high index of suspicion is necessary to recognize this condition.” In addition, accurate and detailed history taking is important in identifying chronic CO poisoning.

TABLE 23.1 Symptoms Reported During Chronic Carbon Monoxide Poisoning Exposure—Penney Study A, 20003 Physical

Cognitive

Emotional/affective

Headaches Nausea Vomiting Fatigue/lethargy Dizziness/vertigo Shortness of breath Muscle/joint aches Balance problems Hearing problems Tremors Short-term memory Mental confusion Attention/concentration Word finding Depression Anxiety Irritability

Weakness Tinnitus Syncope, partial/total Sleep disturbance Vision problems Heart palpitations Paresthesias Muscle cramps Chest pain/tightness Spelling Speech Disorientation Personality changes Mood swings Apathy

a Summarized and reprinted with permission of author.

8417: “8417_c023” — 2007/9/11 — 12:14 — page 499 — #5

Carbon Monoxide Poisoning

500

TABLE 23.2 Persisting Symptoms Following Chronic Carbon Monoxide Exposure— Penney Study A, 20003 Physical

Cognitive

Emotional/affective

Ataxia Balance Muscle/joint pain Temperature deregulation Chest pain/tightness Choking Motor incoordination Muscle cramps Dizziness Dysarthria Fatigue Shortness of breath Tinnitus G. I. problems Attention/concentration Short-term memory Mental confusion Disorientation Executive dysfunction Slow speed of information processing Math Paraphasias (literal and verbal) Verbal fluency (word finding) Depression Anxiety Panic attacks Irritability Personality change

Headaches Hearing problems Multiple chemical sensitivity Nausea Paresthesias Peripheral neuropathy Heart palpitation Motor tremors Photophobia Phonophobia Sleep disturbance Vision problems

Reading Speech Spelling Writing Vocabulary (reduced)

Aggression Libido (reduced) Motivation (reduced)

a Summarized and re-printed with permission of author.

Study A also assessed residual or persisting symptoms experienced by the 66 respondents. Ninety-five different persisting symptoms were identified. A select sample of those symptoms is presented in Table 23.2. Penney’s Study B3 involved the analysis of questionnaires completed by 82 individuals who claimed to have suffered chronic CO poisoning. The mean duration of exposure was 28.4 months, plus or minus 4.4 months, with a range of 3 weeks to 120 months. The mean period after termination of the CO exposure, to the time their responses were given was 21.4 months, plus or minus 2.2 months, or nearly 2 years after the exposure stopped. This study reviewed only residual or permanent symptoms that the respondents were experiencing at the time they completed their questionnaires. Regarding physical symptoms, 100% reported persisting problems with fatigue. Greater than 80% reported residual problems with headaches, muscle

8417: “8417_c023” — 2007/9/11 — 12:14 — page 500 — #6

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

501

and joint pain, dizziness/balance problems, and sleep disturbance. Greater than 50% reported ongoing difficulties with chest pain, tingling and numbness, and vision problems. More than 45% reported ongoing change in their perception of or sensitivity to smell or taste. From a cognitive standpoint, greater than 70% reported ongoing problems with decision-making, following directions, short-term memory, and attention and concentration. More than 40% reported ongoing difficulties with spatial disorientation and organization. With regard to emotional and affective symptoms, more than 80% of the sample reported ongoing problems with mood change/swings, temper problems/irritability, and personality changes. More than 40% reported ongoing social and family problems. Thirty percent reported school problems. Regarding these results, Penney3 states, “This study suggests that a multitude of physical, cognitive, and emotional symptoms persist for very long periods of time following chronic exposure to CO. The CO exposure need not produce altered consciousness at any time for this to occur. In fact, the CO concentrations and COHb saturations are quite low and in the range previously thought incapable of producing lasting health harm in humans” (p. 413). The CO Support study11 involved the analysis of questionnaires completed by 65 individuals who were chronically exposed to CO. None of those individuals lost consciousness (LOC) as part of their exposure. The results of 12 questionnaires completed by individuals who did experience LOC were also reviewed. Ten of those individuals were involved in chronic CO poisonings and two experienced acute poisonings. This study is unique in that it used controls matched for gender, age, and income. For the chronically exposed patients, data regarding symptoms experienced during exposure were summarized as well as persisting symptoms. A detailed summary of this study will not be presented as part of this chapter. For a detailed review of the results of this study, the reader is referred to Hay et al.12 It is important to note, however, that as in the Penney studies; respondents experienced multiple symptoms in multiple systems during their exposure. It is important to note that results of this study also suggest that, while some symptoms do abate to some degree once the exposure stops, in every case the symptoms may persist long-term. The CO Support study also gathered data regarding employment outcome. Although this was not discussed in detail in the Hay et al.12 summaries, a review of the original technical paper provide more detail regarding employment outcome. Those results are presented in Table 23.3. Of note, 32% of the patients chronically exposed to CO were unable to return to work following the exposure. When LOC occurred because of the CO poisoning, 75% were unable to return to competitive employment. This would suggest that when LOC occurs in conjunction with CO poisoning, the possibility of total and permanent vocational disability increases dramatically. Both of these figures stand in contrast to the control group where none of those individuals was disabled from employment. Hay and his colleagues12 conclude their summary of this study by stating, “The results of this survey indicate that there is a continuing and unrecognized problem associated with chronic exposure to CO. Most physicians do not recognize the symptoms of CO poisoning, and, therefore, do not diagnose it. Many individuals

8417: “8417_c023” — 2007/9/11 — 12:14 — page 501 — #7

Carbon Monoxide Poisoning

502

TABLE 23.3 Employment Outcome—CO Support Study, 199711

Able to work full time Able to work part time Unable to work Not applicable Working full time Working part time Unable to work Not applicable

Chronic CO Poisoning (N = 65)

CO Poisoning with LOC* (N = 12)

35% 8% 32% 22%

17% 8% 75% 0%

Matched Nondisabled Control Group (N = 65)

44.6% 24.6% 0% 30.8%

∗ This group included ten individuals chronically exposed to CO and two individuals who experienced acute exposures.

suffer for many years because of their exposure to this gas, and as the survey indicates, many people continue to suffer symptoms years after the exposure has stopped. Respondents of the questionnaire indicate that they have experienced a wide range of symptoms up to 2 years after the exposure ended” (p. 434).

23.1.4 PUBLISHED CHRONIC CARBON MONOXIDE STUDIES UTILIZING NEUROPSYCHOLOGICAL ASSESSMENT In 1990, Ryan13 presented the case study of a 48-year-old, right-handed, married woman who reported a 3-year history of constant headaches, lethargy and memory problems. She indicated that she did not have any difficulty recalling events from the distant past but had difficulty recalling information that is more recent. She was reporting ongoing episodes of mental confusion, periods of depression, and anxiety. She reported that on one occasion she nearly LOC in her basement and, therefore, had the gas company check her furnace. The furnace was found to be releasing 180 ppm CO. The patient was running a typing service out of her basement and the possibility existed that the exposure may have persisted for up to 3 years. No COHb level was obtained and it is noted that the patient never LOC. While the woman’s headaches stopped once her furnace was replaced, her memory problems persisted. Her history was negative for alcohol or drug abuse, head trauma, and psychiatric problems. She had never previously been exposed to any other toxic substances. The results of her neuropsychological testing revealed deficits in the area of incidental memory, as measured by the Digit Symbol Incidental Memory Test, as well as clear deficits in her ability to both learn and recall verbal and visual information. Dr. Ryan summarizes the results by stating, “There is no doubt that this patient has developed a clinically significant memory disorder. Prior to her exposure, she worked in positions that placed demands on concentration and memory skills; following

8417: “8417_c023” — 2007/9/11 — 12:14 — page 502 — #8

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

503

this exposure, she was unable to keep track of verbal information that was presented seconds earlier, and had difficulty accurately retrieving both verbal and visual information that she had learned within the past 30 min” (p. 64). Based on his overall assessment of the case, Dr. Ryan attributed her neuropsychological disturbance, affective disorder, and somatic complaints to her 3-year history of low level CO exposure. Dr. Ryan further notes that memory disturbances are one of the most frequently reported cognitive problems following CO poisoning. He also notes that this case supports the conclusion that a LOC is not necessary for the development of neuropsychiatric symptoms following a period of CO exposure. Roy Myers et al.14 state that “Chronic exposure to CO produces a clinical syndrome that is often overlooked because of obscure symptomatology, a range of presentations, and a lack of awareness of the problems” (p. 555). They go on to note that neurological exams will often not identify subtle changes in functioning and that neuropsychological testing is often more sensitive to the neurotoxic effects of chronic CO poisoning. Seven patients were included in this study. Each individual had been exposed to CO intermittently or constantly over periods ranging from 3 weeks to 3 years. Once the exposure was identified, each was sent for hyperbaric oxygen (HBO) treatment. Each of the individuals was exposed to a minimum of 200 ppm CO. Each individual was administered the CO Neuropsychometric Screening Battery, as well as other specific neuropsychological tests including the WAIS-R, Trails A and B, Finger Tapping Test, Logical Reasoning and Visual Reproduction from the Wechsler Memory Scale, and Minnesota Multiphasic Personality Inventory (MMPI) or MMPI2. Six of the seven individuals received HBO therapy, ranging from 5 to 59 treatments. The individual who received only five treatments was intolerant to the chamber and that individual’s treatment was considered incomplete. One individual was included in the study who did not receive any HBO treatment. Individuals who received HBO treatment were tested every 2 weeks “until the psychometric tests reached a plateau or returned to normal” (p. 557). The individual who did not receive HBO treatment was tested at two months after the exposure stopped, and then after 10 months of rehabilitation, was tested again at 1 year after the exposure stopped. In addition to neuropsychological testing, a questionnaire was also completed by each participant regarding his or her symptoms. The most common symptoms experienced during the exposure (acknowledged by 50% or more of the group) included problems with headaches, dizziness, motor tremors, difficulties with shortterm memory, sleep disturbance, cognitive set loss, anxiety, reading comprehension, vision, gait/balance, muscle tremors, paresthesias, altered sense of smell, body aches, tinnitus, and spatial disorientation. One participant who received ten HBO treatment sessions reported significant resolution of their symptoms following treatment. One individual who received 50 HBO treatments reported moderate resolution of their symptoms. Three of the participants who received 19, 59, and 29 HBO treatments, respectively, reported minimal or no functional improvement of their symptoms. The individual who received five incomplete treatments reported only minimal improvements in his symptoms. The individual, who received no HBO treatment, but ten months of rehabilitation, reported that his condition improved significantly, but that patient continued to report ongoing and multiple symptoms. In summary, five of the seven (71% of the sample) reported minimal or no resolution of their symptoms

8417: “8417_c023” — 2007/9/11 — 12:14 — page 503 — #9

504

Carbon Monoxide Poisoning

after a period of recovery from the exposure, HBO treatment, and/or rehabilitation services. With regard to the neuropsychological testing, the authors found that the CO Neuropsycho-metric Screening Battery was of little value in identifying residual cognitive deficits and recommended neuropsychological evaluation that is more detailed. By the completion of testing, four of the seven participants continued to demonstrate a significant split in their Verbal and Performance IQs (PIQs). Three of those individuals had Verbal IQs (VIQs) significantly greater than PIQs. One had PIQ significantly greater than VIQ. Three of the seven continued to demonstrate impairments on the Trails B Test, a test of alternating attention and logical sequencing. Three of the seven also demonstrated residual deficits in fine motor speed. One weakness of this study is that there is no discussion of the impact of practice effects. It appears that most individuals in this study were tested every 2 weeks and repeated exposure to these tests can result in significant gains due to practice. The authors note that ongoing problems with emotional lability, irritability, depression, and anxiety are common sequelae of chronic CO poisoning. Pinkston et al.15 conducted Positron Emission Tomography (PET) scans and neuropsychological testing of two adult patients 3 years following a chronic CO poisoning. The patients were both right-handed, middle-aged individuals who had been married for many years. Both worked in professional occupations and they had no history of prior psychiatric or neurologic conditions. They suffered exposure to CO for a 3-year period due to faulty furnace exhaust/ducting. Neuropsychological testing was conducted on both subjects four times over a 3-year period. The results of the testing indicated a significant anterior frontal lobe syndrome. In addition, both individuals demonstrated frontal symptoms in their activities of daily living, such as indecisiveness, mental passivity, and disorganization. Both individuals experienced a significant vocational disability because of their persistent and ongoing symptoms. Indeed, both individuals were rendered vocationally disabled because of their residual deficits. In addition, both experienced losses in their level of independence and both experienced difficulties with various activities of daily living. Both subjects demonstrated a similar pattern of hypometabolism on PET imaging. Substantially reduced metabolism was evident in the orbital frontal and dorsal lateral prefrontal cortex, as well as areas of the temporal lobe for both individuals. It was determined that the individual scans were consistent with the patients’ presenting symptoms and reduced level of functioning. While hypometabolism was evident in various regions of the temporal lobe, both individuals were performing within normal limits on memory tests by the time of their final evaluation. However, both individuals were reporting and experiencing significant memory dysfunction in their day-to-day activities and activities of daily living. It is noted that one of the subjects received a more significant exposure, being in the home for significant more prolonged periods of time. That subject demonstrated more problems on neuropsychological testing and his PET scan demonstrated more areas of significant hypometabolism. That individual also demonstrated more severe behavioral/affective problems. Both patients ultimately developed epilepsy and were begun on Depakote. The authors note that the development of a seizure disorder was

8417: “8417_c023” — 2007/9/11 — 12:14 — page 504 — #10

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

505

not surprising given the temporal and hippocampal dysfunction noted on the PET imaging. Hartman,16 presents a case study of a 65-year-old woman chronically exposed to CO over a six-to-seven-year period. The exposure occurred because of improper installation of a water heater. The woman had 2 years of college and a professional degree. Her symptoms began within a year of the new water heater being installed. As is typical for cases of chronic CO poisoning, her symptoms gradually emerged and worsened over time. Physical symptoms that she developed during the exposure were fatigue, muscle spasms, loss of muscle tone in her face, paresthesias, muscle fatigue, migrating neuritis (sharp pains), problems with balance, her fingernails, and toenails turned black, sleep disturbance, and one near-blackout episode. Other symptoms that she experienced included depression, panic attacks, sleep disturbance, problems with short-term memory, spatial disorientation, and difficulties with vision. By the end of the exposure, she had severe body pain, even at rest. Also during the exposure period, she experienced frequent urinary tract infections, other chronic infections, and developed severe allergies, suggesting a possible compromise of her immune system. It was also noted during the exposure that house plants died and silverware turned black quickly. It was noted that her symptoms did abate to some degree when she would leave the house for several days at a time, but worsened upon returning home. Once the exposure stopped, some symptoms resolved but some persisted, most notably difficulty with her vision, short-term memory, allergies, and sleep disturbance. It was also noted that she had become sensitive to a variety of chemicals and substances (e.g., perfume). The patient underwent serial neuropsychological testing. By the time she was evaluated several years after the exposure stopped, she was functionally reporting ongoing problems with language comprehension, verbal short-term memory, and occasional episodes of spatial disorientation, mild emotional lability, and intermittent sleep disturbance. She was continuing to note ongoing sensitivity to various chemicals and substances (e.g., pesticides). The patient had also developed Crohn’s disease. On neuropsychological testing, she was demonstrating persistent sensory/motor and spatial integration deficits. There was a bilateral loss of her ability to discriminate one versus two-point touch on her fingertips and some errors in fingertip localization. Fine motor coordination on the Grooved Pegboard Test was severely impaired. The patient’s ultimate DSM-III-R diagnosis was 294.80 Organic Mental Disorder, not otherwise specified/probable CO exposure etiology. Divine et al.17 present a case study of a 45-year-old woman chronically exposed to CO for approximately 1 year. The exposure occurred because of a faulty furnace at her place of employment where she worked as a cook. For approximately 1 year, she experienced the following symptoms: “severe flu,” inability to walk in a straight line, bumping into things, problems with balance, severe headache, fatigue, verbal fluency, hearing problems, paresthesias, irritability, and facial pain. Her condition was misdiagnosed as a sinus infection. She had been off work for a period of 5 days and upon returning to work immediately became ill and contacted the gas company. “Extremely high” levels of CO were identified in her work area, at which point she left the premises. The exact CO concentration in her work space was not given.

8417: “8417_c023” — 2007/9/11 — 12:14 — page 505 — #11

Carbon Monoxide Poisoning

506

Seventeen months following cessation of the exposure, many of her symptoms had resolved. However, she was reporting ongoing problems with reading, writing, speaking, verbal fluency, and dysarticulation. An MRI performed at 15-months after the exposure ended was read as abnormal. The scan revealed multiple small lesions bilaterally in the basal ganglia. The lesions were more severe in the globitus palidus than in the putamen. The radiologist concluded that the lesions were consistent with chronic CO poisoning. The patient had a Bachelor’s degree and no prior neuropsychological history. Neuropsychological testing conducted at 17-months after the end of exposure revealed deficits in attention and concentration, learning and memory retrieval. Testing also suggested problems with depression at that time. Behaviorally, during the course of the evaluation, lapses in attention, perseverations, sequencing problems, slight concreteness, and verbal fluency difficulties were noted. The authors concluded that the testing was consistent with subtle frontal lobe dysfunction. Retesting was performed 12-months later and similar deficits were found, suggesting persistent frontal lobe dysfunction.

23.2 HELFFENSTEIN STUDY 23.2.1 SAMPLE AND EXPOSURE DATA 23.2.1.1 Admission Criteria Participants in this study were 21 consecutively evaluated patients who had been chronically exposed to CO. There were a number of criteria established for admission to the study: 1. The full WAIS-III was administered as part of the evaluation 2. No prior psychological or psychiatric history requiring treatment 3. No prior neuropsychological history (e.g., no prior head injuries, toxic exposures, or neurological illness) 4. No history of substance abuse As all participants were involved in some type of active litigation, no patient was admitted to the study if there was any indication of symptom magnification or exaggeration. All 21 participants of this study performed satisfactorily on three symptom validity tests. Eighty-one percent of the participants were administered and passed the Tombaugh Test of Memory Malingering, Hiscock Digit Recognition Test, and Rey II 15 Item Memory Malingering Test. The remaining 19% of the cohort were administered and passed the Computerized Assessment of Response Bias, Word Memory Test, and Tombaugh Test of Memory Malingering. In addition, the MMPI-2 profiles of all 21 participants were devoid of any indications of symptom magnification or exaggeration. In addition, all 21 participants demonstrated behavioral indicators of good effort during the neuropsychological testing process. Participants were admitted to the study only if there was some documented evidence of dangerous levels of CO in their living or work environments. In 14 cases, measurements of CO in ppm had been made in the living or work space. For two of the

8417: “8417_c023” — 2007/9/11 — 12:14 — page 506 — #12

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

507

participants, COHb levels were obtained shortly after the exposures were identified. For five of the participants high levels of CO were identified as being emitted from a furnace or other appliance and there was a reasonable mechanism by which CO entered the living or workspace of the individual. 23.2.1.2 Sample Demographics The mean age of participants in this study was 39.6 years (range 16–78). The mean level of education was 13.8 years (range 9–20 years of formal education). All participants in this study were Caucasian. Eighteen were female and three were male. All participants were right-handed. Nineteen of the participants were working full time at the time of the exposure and two were full-time students. 23.2.1.3 Source and Location of Exposure Fifteen of the 21 participants experienced chronic CO poisoning in their homes. Six of the participants experienced CO exposure at work. Eleven of the participants experienced CO exposure because of a faulty furnace or boiler. Five were exposed as a result of a faulty furnace and water heater combination. One experienced CO exposure because of a faulty gas range and water heater combination. Two participants experienced CO exposure as a result of faulty gas fireplace installation. One participant experienced CO exposure because of a faulty water heater installation. One participant experienced CO exposure because of auto exhaust fumes from a parking garage entering his apartment. 23.2.1.4 Level of Carbon Monoxide Exposure As noted above, in 14 of the cases included in this study, actual measurements of CO in parts per million were made in the living or work space of the individual. The average concentration of CO was 123 ppm, with a range of 13–467 ppm. It is important to note that 100-ppm CO exposure over many hours will result in a COHb saturation of 14%. Two of the participants had their COHb levels measured shortly after the presence of CO was discovered. Regression (i.e., back) calculations were performed to determine the COHb levels at the time the participant left the toxic environment. On average, these two individuals had COHb saturations of 14.5% at that time. For four of the participants in the study, a furnace and water heater at their place of employment were housed in a confined space, which limited combustion air. This created a situation where both the furnace and water heater had a tendency to back draft. The furnace was producing 200+ ppm CO and the water heater was producing 2000+ ppm CO. These readings were taken at the exhaust vents. When back drafting occurred, CO was entering the work space of these individuals. 23.2.1.5 Duration and Frequency of Exposure For each of the 21 participants, it was possible to estimate with some certainty the duration of exposure. The mean duration was 28.9 months (2.4 years) with a range of

8417: “8417_c023” — 2007/9/11 — 12:14 — page 507 — #13

Carbon Monoxide Poisoning

508

0.5–120 months. It was also possible to estimate with some certainty the frequency of exposure. The average frequency of exposure was 6.3 days per week with a range of 3–7 days per week. 23.2.1.6 Time Since Exposure Each of the 21 participants was evaluated because they were experiencing persisting symptoms, which were great enough to negatively impact their day-to-day, academic, and/or work activities. Clearly, this sample represents the sub-set of chronically exposed CO patients who do not make a full and complete recovery. Testing was conducted an average of 46.8 months (3.9 years) postexposure. The range was 16–111 months postexposure (1.3–9.25 years).

23.2.2 BATTERY ADMINISTERED Each participant was administered a comprehensive battery of neuropsychological tests. Each individual received the Expanded Halstead–Reitan Neuropsychological Test Battery, as well as most of the tests renormed for age, gender and education by Heaton, Grant and Matthews.19 Each individual was administered the complete WAIS-III, as well as the Peabody Individual Achievement Test (PIAT). In addition to the above, the following tests were administered: 1. Complex ideation subtest of the Boston Diagnostic Aphasia Examination 2. Nonverbal agility and verbal agility subtest of the Boston Diagnostic Aphasia Examination 3. Ruff Figural Fluency Test (Ruff FFT) 4. Hooper visual organization test 5. Padula visual midline screening test20 6. Line bisection test 7. Behavioral dyscontrol scale 8. Buschke verbal selective reminding test 9. Rey-Osterreith complex figure test 10. Paced Auditory Serial Addition Test (PASAT) (Levin version) In addition to the above, each of the participants completed a MMPI-2 or Minnesota Multiphasic Personality Inventory-Adolescent (MMPI-A). For each of the participants, the following index scores were generated: 1. 2. 3. 4.

Halstead Impairment Index (HII) Average Impairment Rating (AIR) Global Deficit Score (GDS) General Neuropsychological Deficit Scale (GNDS)

For an earlier discussion of neuropsychological testing in a case study of CO-poisoning by this author, see Helffenstein, 2000.18

8417: “8417_c023” — 2007/9/11 — 12:14 — page 508 — #14

Neurocognitive and Neurobehavioral Sequelae of Chronic Carbon Monoxide

509

23.2.3 NORMS UTILIZED For the purpose of this study, the Heaton et al.19 demographic norms were used. These norms are demographically corrected for age, gender, and education. In addition, for the purpose of this study, the performance levels proposed by Heaton et al. were utilized. Table 23.4 shows the ranges of performance proposed by Heaton et al. as well as the accompanying T-scores and percentile scores. Where nondemographically corrected scores (i.e., non-Heaton norms) are utilized, those norms will be individually identified by level of performance and will be discussed using the Heaton guidelines.

23.2.4 SYMPTOMS EXPERIENCED BY PARTICIPANTS DURING CARBON MONOXIDE EXPOSURE Table 23.5 is a summary of the most common symptoms reported by the participants of this study during chronic CO poisoning. Consistent with prior studies of chronic CO poisoning, people experienced multiple symptoms in multiple systems. Table 23.5 clearly shows that the symptoms involve a wide variety of physical, sensorymotor, visual, cognitive, and affective/mood conditions. Penney3 discussed in some detail misdiagnosis of chronic CO poisoning, and notes that common misdiagnoses include chronic fatigue syndrome, viral/bacterial/pulmonary or gastrointestinal infection, “rundown condition,” endocrine problem, immune deficiency disorders, psychiatric/psychosomatic problems, allergies, and food poisoning. As noted earlier in this chapter, prior studies have found that during chronic CO poisoning, individuals typically experience multiple symptoms in multiple organ systems. On average, the participants in the current study experienced 25 symptoms during the chronic CO poisoning event, and, as noted by Dr. Penney, most of these individuals received a variety of misdiagnoses during their exposure. Often, when

TABLE 23.4 Heaton Range of Performance on Neuropsychological Testing (Heaton, et al.19 ) Level of Performance Above average Average Below average* Mild impairment Mild/Moderate impairment Moderate impairment Moderate/severe impairment Severe impairment

T-Scores

Percentile Scores

55+ 45–54 40–44 35–39 30–34 25–29 20–24 1–19

68+ 30–67 12–29 6–13 3–5 1–2 15%. Two-hundred and one patients were included in the more severe CO-poisoning group and 55 in the less severe CO poisoning group. Cognitive sequelae occurred in 39% of the first group and in 35% of the second group. The difference was not statistically significant at any time. Thus, both the more and the less severe groups had significant cognitive sequelae. Therefore, CO-induced cognitive outcomes appear to be unrelated to measures of apparent CO poisoning severity. One could ask why it has taken so long for the effects of less severe, acute, and chronic CO poisoning to be recognized. With respect to chronic CO poisoning, this has come about primarily through the work of neuropsychologists such as Bronstein et al., Ryan, Hartman, Pinkston, Devine, Helffenstein, Hopkins, and others,2,3,5−9 toxicologists such as Hay and Penney1,10,11 and epidemiologists such as Ritz, Morris, and others,12,13 not largely through the work of physicians. This is probably owing to the fact that most of the symptoms produced by chronic CO poisoning are not considered seriously by those in mainline medicine and the practice of neuropsychological evaluation is a discipline that has grown up outside of internal medicine. In my case as a specialized toxicologist, I may have recognized them because they were thrust into my face by victims of chronic CO poisoning, and moreover since I entered the field from pure science and had no preconceived ideas about what to expect.

24.5 CONCLUSION This study of a series of 61 patients demonstrates that chronic CO poisoning is not without long-term health consequences. The study results show not only the frequency of the reporting of symptoms known to be part of the Carbon Monoxide Poisoning Syndrome1 but now also the intensity with which each of these symptoms is reported. Again, we see a multiplicity of symptoms and signs in the five arenas, many more than most of the other usual or more common diseases seen by health care professionals. And again, these symptoms and signs appear to involve a number of organ systems, but all with the common connection of the central nervous system and the well recognized damage that CO can do to this organ system.

References 1. Penney, D.G. Chronic carbon monoxide poisoning. In: Carbon Monoxide Toxicity, D.G. Penney, Ed.,CRC Press, NY, 2000, Chapt. 18, pp. 393–418.

8417: “8417_c024” — 2007/9/11 — 12:14 — page 565 — #15

566

Carbon Monoxide Poisoning 2. Helffenstein, D.A. Cognitive and behavioral sequelae of chronic CO poisoning: Data on large case series. In: Carbon Monoxide Poisoning, D.G. Penney, Ed., CRC-Taylor and Francis Press, NY, 2007, Chapt. 23. 3. Hopkins, R.O. Neurocognitive and affective sequelae of carbon monoxide poisoning In: Carbon Monoxide Poisoning, D.G. Penney, Ed., CRC-Taylor and Francis Press, NY, 2007, Chapt. 23. 4. Chambers, C., Hopkins, R.O., and Weaver, L.K. Cognitive and affective outcomes compared dichotomously in patients with acute carbon monoxide poisoning. Undersea Hyperbar. Med., 48, 2006. 5. Bronstein, A.C., Kadushin, F.S., and Teitelbaum, D.T. Neurobehavioral findings in two cases of chronic low level carbon monoxide poisoning Veter. Human Toxicol., 29, 479, 1987 (abstract only). 6. Ryan, C.M. Memory disturbances following chronic, low-level carbon monoxide exposure. Arch. Clin. Neuropsychol., 5, 59–67, 1990. 7. Hartman, D.E. Neuropsychological Toxicology Identification and Assessment of Human Neurotoxic Syndromes (2nd ed.), Plenum Press, New York and London, 1995. 8. Pinkston, J.B., Wu, J.C., Gouvier, W.D., and Varney, N.R. Quantitative PET scan findings in carbon monoxide poisoning: Deficits seen in a matched pair. Arch. Clin. Neuropsychol., 15, 545–553, 2000. 9. Devine, S.A., Kirkley, S.M., Palumbo, C.L., and White, R.F. MRI and neuropsychological correlates of carbon monoxide exposure: A case report, Environ. Health Perspect., 2002; 110: 1051–1055. 10. Hay, A.W.H., Jaffer, S., and Davis, D. Carbon Monoxide Support. Effects of chronic exposure to CO:Aresearch study conducted by CO Support. Technical Paper. October, 1997. 47 pp, appendices. 11. Hay, A.W.H., Jaffer, S., and Davis, D. Chronic carbon monoxide exposure: The CO Support study. In: Carbon Monoxide Toxicity, D.G. Penney, Ed., CRC Press, NY, 2000, Chapt. 19, pp. 419–437. 12. Ritz, B., and Yu, F. The effect of ambient carbon monoxide on low birth weight among children born in Southern California between 1989 and 1993. Environ. Health Perspect., 107, 17–25, 1999. 13. Morris, R.D. Low-level carbon monoxide and human health. In: Carbon Monoxide Toxicity, D.G. Penney, Ed., CRC Press, NY, 2000, Chapt. 17, pp. 381–391.

Editor’s note: On March 7, 2007 I presented a 50 minute talk in a room at the House of Lords, London, UK, along with others. Several lords and members of parliament were in attendance at various times. The meeting was organized by “CO Awareness”, an English “charity”, whose mission is the reduction of injuries and deaths from unintentional CO poisoning in that country. I spoke about the dangers of CO poisoning, its proper diagnosis, this book, and particularly about the patient (case) series described in this chapter. The impression I gained at the meeting and in talking to a number of participants, live and via E-mail, was that some British are suspicious of statistics released by the government on the numbers of deaths from CO poisoning. Many believe the numbers are far higher than official values. They are also frustrated by what they perceive as inaction on the part of governmental agencies that could address the problems of CO exposure. For more on this, see the comments by Rob Aiers

8417: “8417_c024” — 2007/9/11 — 12:14 — page 566 — #16

Chronic Carbon Monoxide Poisoning: A Case Series

567

in Chapter 11, who manages the number one website for CO poisoning matters, “carbonmonoxidekills.com”. Another impression I gained at the House of Lords meeting is that in Britain there are inadequate numbers of people with special training and experience in CO poisoning. Victims see, for the most part, only physicians, who are principally general practitioners, and who are unschooled and inexperienced in toxicology and in the outcomes of CO poisoning. The same can be said for the neurologists, internists, cardiologists, etc. that they see. As discussed elsewhere (Chapters 14 and 19) these are not generally the health professionals who can be most helpful in the diagnosis, testing and management of people with CO poisoning. In some respects this is similar to the situation in the USA. Finally, it was my impression that high quality neuropsychological evaluation for CO victims such as we have in the USA (see Chapters 22, 23 and 25) may not be encouraged by the medical establishment and/or is unavailable to most victims of CO poisoning in Britain, even though it is often needed. Since the majority of the lasting health effects of CO poisoning generally occur in the cognitive-memory and affective-emotional arenas, and neuropsychological evaluation is considered the gold-standard for assessing CO-induced brain damage, this problem is extremely serious with regard to the proper testing and long-term management of CO poisoned patients.

8417: “8417_c024” — 2007/9/11 — 12:14 — page 567 — #17

8417: “8417_c024” — 2007/9/11 — 12:14 — page 568 — #18

25

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children Carol L. Armstrong and Jacqueline Cunningham

CONTENTS 25.1 25.2 25.3 25.4 25.5 25.6

Prevalence and Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanisms of Toxicity Effects on the Brain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evidence of Structural Brain Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neurobehavioral Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbon Monoxide Toxicity as a Model of White Matter Injury . . . . . . . . . . Developmental Effects of Carbon Monoxide Toxicity . . . . . . . . . . . . . . . . . . . . 25.6.1 Pediatric Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6.2 Age as a Critical Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6.3 Developmental Neuropsychological Approach to Assessing Toxicity Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.1 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.2 Case Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

569 570 571 571 572 573 573 574 575 577 577 585 586 587

25.1 PREVALENCE AND SYMPTOMS Carbon monoxide (CO) poisoning is a relatively common environmental toxic exposure risk that can masquerade as other illnesses. Attribution to CO toxicity can be made based on clinical findings and knowledge of a malfunctioning heater, without evidence of elevated carboxyhemoglobin (COHb) levels or acute mental status changes. In the United States, CO toxicity is implicated in more than 40,000 emergency department visits made annually.1,2 CO is found wherever there is incomplete combustion of carbonaceous material, and, after carbon dioxide, it is the most

569

8417: “8417_c025” — 2007/9/11 — 12:14 — page 569 — #1

570

Carbon Monoxide Poisoning

abundant air pollutant.3 Common sources of CO are fires, faulty combustion heating systems, exhaust from internal combustion engines, and heating gases other than natural gas. Auto exhaust and exhaust gases from oil heat were most commonly associated with elevation of COHb in a study of serially admitted children to an urban emergency department who had elevated COHb levels.2 Normal COHb saturation is 0.4–0.7% at rest, while the ambient level is 0.5–1.5% in the general population due to added environmental exposure. Tobacco smokers have COHb levels ranging from 4% to 20%; the mean for one-pack/day smokers is 5–6%. COHb levels may be higher during pregnancy. Infants of mothers who smoke may have COHb elevations up to 4.3%. The presenting symptoms of a chronic CO exposure are occult, different and less specific than those that are more readily (though not easily) diagnosed in an acute/peak exposure. These differences in presenting symptoms, neuroimaging, COHb measurements, and other clinical characteristics often lead to misdiagnosis.4 Symptoms of chronic exposure may be mistaken as flu-like symptoms or other etiology, especially if there was no loss of consciousness (LOC). Early symptoms of an acute or a peak CO exposure are more obvious. These may include headache (experienced at COHb levels of ≥ 10%), fatigue and lethargy, dizziness, paresthesias, chest pain, palpitations, and nausea. Severe exposures result in obtunded consciousness, reducing the victim’s ability to recognize danger. Peak exposure may be followed by an acute encephalopathy, abnormal hyperintensity signal changes in the brain on magnetic resonance imaging (MRI), and long-lasting neurobehavioral and cognitive changes. Delayed sequelae, or encephalopathy, follows a period of apparent recovery after acute encephalopathy. Problems in functioning can have a delayed onset, and appear suddenly after days or weeks of apparent normal functioning. Symptoms may progress to severe neurocognitive impairments and neuropsychiatric (personality) changes, indicating evidence of hypoxic/anoxic brain damage.

25.2 MECHANISMS OF TOXICITY EFFECTS ON THE BRAIN The harm created by CO exposure is mainly due to reduction in oxygen transport rather than to any significant poisoning effect on respiratory enzymes.5 The major determining factors of toxicity are CO concentration, duration of exposure, and alveolar ventilation. The disruption to cerebral blood flow created by reduction in oxygen transport produces brain injury through hypoxia. The basic mechanisms of behavioral changes are hypoxic brain damage resulting from: (1) the displacement of O2 when CO binds with hemoglobin and interferes with O2 delivery to tissues, (2) the inhibition of mitochondrial oxidative respiration and cardiomyopathy due to displacement of O2 from myoglobin by CO, with associated hypotension and systemic acidosis, and (3) the increased permeability of the vascular endothelium caused by CO. Neural inflammatory processes may also be involved in delayed neurobehavioral problems because of brain iron extravasion into neural tissues and release of myelin basic protein following CO poisoning.6

8417: “8417_c025” — 2007/9/11 — 12:14 — page 570 — #2

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children

571

Minimal effects on cognitive performance are reliably observed at 7% COHb level, and a 5% COHb level causes decrements in vigilance tests and in difficult dual task tests. A 2–4% COHb level has been associated with reduced video game performance.7−10

25.3 EVIDENCE OF STRUCTURAL BRAIN INJURY Although initial COHb levels do not predict later impairment well, radiographic neural changes are predictive of functional impairments. Peak exposures are more likely to produce obvious structural changes on clinical MRI scans of the brain when caused by hypoxia and by cardioembolic or hypotensive strokes related to myocardial ischemia. The most common MRI finding on T2-weighted images and Fluid Attenuation Inversion Recovery (FLAIR) is often, but not always, bilateral symmetric hyperintensity of the white matter, greater in the centrum semiovale, with relatively less involvement of the temporal lobes and anterior frontal lobes.11,12 The incidence of white matter hyperintensities varies in natural history group studies. In one large prospective controlled group study (N = 73), the incidence was 12%, significantly more in the periventricular area than in the centrum semiovale.13 Cognitive impairment was greater in those patients with hyperintensities in the centrum semiovale, and signal changes persisted over the 6-month study period. Atrophy is common in chronic CO exposure. Most reports of structural abnormalities are based on case reports, but one group report cited cerebral cortical atrophy in 60% of adults, mild atrophy of cerebellar hemispheres in 50%; vermian atrophy in almost 70%, with less common atrophy found in the globus pallidus and corpus callosum.6,12 Atrophy of the hippocampus,11 basal ganglia,14 thalamus and medial temporal lobe15 can also occur. Structures with high-energy demands, including the hippocampus and the cerebellum, are particularly vulnerable to hypoxia,11 as are those high in iron content, including the basal ganglia.16 Initial Computed Tomography (CT) or MRI often fails to detect low-density lesions in the globus pallidus, a frequent site of injury, and has poor clinical correlation as abnormalities seen on the scan may be found in both wellrecovering and comatose patients. At follow-up, lesions may disappear, diminish, or remain unchanged, and resolution of white matter lesions, ventricular enlargement, and cortical atrophy later on may be represented by an improved clinical picture with persistent neurocognitive and neuropsychiatric impairment at a less severe level.6,17,18 However, neurocognitive impairments are likely even when neurological symptoms resolve. Visual evoked potentials are often done, but may miss effects, if white matter lesions result in demyelination in other parts of the brain.19 The most robust findings may be found at long-term follow-up.

25.4 NEUROBEHAVIORAL FINDINGS Neurobehavioral symptoms vary and correspond with damage to cortex, subcortical nuclei, and white matter. The following symptoms and signs have been reported: amnesia,11,20 apraxia,21 visual object agnosia,22 akinetic mutism,23,24 parkinsonism,25 choreoathetosis,26 peripheral neuropathy,27,28 cortical blindness,29,30 depression and

8417: “8417_c025” — 2007/9/11 — 12:14 — page 571 — #3

572

Carbon Monoxide Poisoning

anxiety,31 and impaired attention and executive functioning.32 Varied memory processes have been implicated in CO toxicity, including incidental memory, recognition, paired associate learning, delayed recall, and visual short- and long-term memory, without effects on intelligence, visuospatial functions, speed, dexterity, nor letter fluency.33

25.5 CARBON MONOXIDE TOXICITY AS A MODEL OF WHITE MATTER INJURY Chang et al.34 identified three types of white matter injury that occur following CO poisoning: (1) multiple small necrotic foci in the centrum semiovale and interhemispheric commissures; (2) areas of necrosis in the deep periventricular white matter that is associated with axonal destruction and lipid-laden macrophages; and (3) demyelination in the deep white matter.13,32,35,36 The multifocal and diffuse injuries associated with CO toxicity link it with traumatic brain injury (TBI) in terms of its commonality with the latter as a white matter disease.37 Whereas the acute clinical course of CO toxicity is differentiated from that of deceleration or rotational brain injury, the chronic state of either condition is associated with events, including the production of free radicals and excitatory amino acids and the disruption of normal calcium homeostasis, which act in concert to exacerbate the hypoxic-ischemic insult that can occur in association with diffuse axonal injury (DAI).38 Given the commonality of neurochemical mechanisms underlying axonal injuries, TBI may be a relevant comparison when examining the neurobehavioral effects of CO toxicity, as well as in speculating on the responsiveness of either condition to similar approaches to rehabilitation.39 In the cognitive domain, a pattern of relative strengths and weaknesses has long been known to characterize DAI: the relative preservation of “crystallized” and prior-learned knowledge, and the severe effects on “fluid” processing,40 which is resource-limited. Resource-limited processes require higher levels of processing resources for optimal performance and are affected by inter-and intra-task competition.41 Fluid intelligence is largely nonverbal and a culture-reduced form of mental efficiency that is used when a task requires adaptation to a new situation. Crystallized intelligence is content-related, and represents the corpus of what has been learned. Whereas no narrowly defined cognitive pattern has been identified as characterizing CO toxicity,33 a frequent cognitive profile obtained in the neuropsychological evaluation of the CO-poisoned patient is one that mirrors that found in TBI in that the profiles demonstrate relative sparing of crystallized abilities, but serious involvement of fluid processes.42 Most common deficits in the CO-poisoned patient are observed in executive functioning (i.e., planning, monitoring, resisting distraction, and maintaining flexibility in thinking), attention and concentration, memory, visual perceptual abilities, and speed of information-processing.42 In both adults and children, deficits in these areas are also expected to occur, singly or in combination, as sequelae to closed head injury.43

8417: “8417_c025” — 2007/9/11 — 12:14 — page 572 — #4

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children

573

25.6 DEVELOPMENTAL EFFECTS OF CARBON MONOXIDE TOXICITY 25.6.1 PEDIATRIC VULNERABILITY It is important to differentiate the typical CO exposures in adults as compared to children. Adults often are exposed to more lethal doses, but children have heightened vulnerability to CO toxicity based on constitutional factors.44 Low birth weight is a well-known sequelae of CO exposure.45,46 Across all pediatric age groups, roughly 50% of fatalities are unintentional and of those, 50% are related to motor vehicle exhaust inhalation.47 Parameters that are relevant to neural development are the duration of exposure (chronic or peak exposure); acute, subacute or delayed symptoms, timing of exposure (prenatal vs postnatal), and dose (moderate levels analogous to cigarette smoking or higher levels). Symptoms in children appear to manifest more variably than in adults, with even poorer clinical correlation with a wider range of COHb levels,16 and with variable outcomes occurring even in related individuals with similar exposure levels.48 This variability can lead to the conclusion that no clear relationship can be predicted between severity of CO poisoning (COHb levels at admission) and residual neuropsychological deficits.13,48 Even in cases of known elevated COHb levels, some individuals demonstrate little chronic cognitive impairment, although a degree of impairment remains likely. There are very few reported case outcomes for CO-exposed children. Severe memory impairment was reported in four pediatric cases with chronic exposure over several years; in three of the four children, exposure was lower-level and without LOC.33 Physiological factors putting children at greater risk than adults include children’s rapid uptake of CO into the bloodstream, high metabolic needs coupled with small blood volume, reduced pulmonary transport of CO while sleeping (occurring for longer periods in children), and reduced pulmonary transport of CO due to young age.16 There is also heightened risk for children, as compared with adults, from exposure to CO in utero. CO is a known potent fetotoxin and teratogen with prenatal toxicity retarding growth and development.49 Direct effects on cognition are shown by translational studies. Long-lasting learning and memory deficits have been seen to result from gestational exposure of rats to levels of COHb in the maternal blood supply equivalent to cigarette smoking in humans. Postnatal exposure to this dose of CO from 1 to 10 days postbirth has resulted in no memory or learning deficits at 3 and 18 months of age in rats.50 However, exposure to the same level of CO from day 0 to day 20 of gestation in rats has led to impairments in habituation, working memory, and the ability to explore novel objects, but no alteration in spontaneous motor activity.51 The behavioral effect of prenatal exposure may be caused in part by changes in mesolimbic dopaminergic transmission52 and in cholinergic and catecholaminergic pathways.53 In one study, guinea pigs were exposed to CO for 10 h a day from Day 23 to 25 of gestation until term at 68 days. The fetus brains were examined 5–7 days prior to birth. Prenatal exposure to CO affected cholinergic and catecholaminergic pathways in the medulla, particularly in cardio-respiratory centers, a finding which may be related to smoking and increased rate of sudden infant death

8417: “8417_c025” — 2007/9/11 — 12:14 — page 573 — #5

Carbon Monoxide Poisoning

574

syndrome.53 Prenatal exposure can also cause developmental alterations in the areas of the brainstem responsible for respiratory control, causing increased sensitivity to CO after birth.54 Other factors mediate the impact of CO as a toxicant on child development, some potentiating and some moderating. The level of CO poisoning, age at the time of exposure, and the intellectual level of the child mediate long-term functional capacity.55

25.6.2 AGE AS A CRITICAL VARIABLE Because the growth of children is incomplete at any one point in time, the developmental stage reached at the time of a brain insult holds particular importance in assessing the injury’s significance. The primary challenge to successful coping imposed by pediatric brain injury, is that an acquired brain injury generates “rippling effects” that have long-term consequences over the whole course of a child’s future development.56 The most enduring effects are thus expected to occur at the youngest ages when the child has not yet had the opportunity to master fundamental developmental tasks. In the period from birth to 2 years of age, the child is expected to achieve an understanding of cause-and-effect relationships. By understanding that certain events are routinely paired together, he/she becomes capable of basic self-regulation, an essential achievement that paves the way for the ability to integrate thinking, emotion, and behavior—the task of the developmental period from 3 to 5 years. In the middle-childhood years, the achievement of classroom self-organizational skills forms the precursor for the development of the planning and goal-direction capacities necessary for movement and problem-solving through the middle school years. Failure to develop age-appropriate skills at any stage in the developmental trajectory thwarts the development of the judgment skills eventually necessary for successful progress toward adolescent autonomy.56 The commonly reported finding that permanent deficits in mood regulation and executive functioning, in areas including decision-making and attention-monitoring57 frequently follow CO poisoning, warrants looking at developmental factors which are influential in the expression of impairment. Disruptions in achieving major milestones in the first year of life create lock-step constitutional changes that affect overarching consequences as the child matures. For example, the infant who fails to understand that certain events are routinely paired together becomes despondent when the mother does not immediately attend to his/her cry, resulting in the infant’s difficulty in regulating the sleep-wake cycle.58 Taking a developmental perspective on pediatric CO toxicity extends the focus from the physiological mechanisms causing dysfunction to psychosocial considerations. In the absence of such a perspective, valuable information may be lost because subtle signs, such as irritability, are not detected as markers of organicity. The paradoxical results obtained in a study59 that rated infants with relatively high levels of COHb as asymptomatic have in hindsight been attributed to failure to consider the significance of feeding difficulty and irritability in interpreting the findings.16

8417: “8417_c025” — 2007/9/11 — 12:14 — page 574 — #6

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children

575

Empirical findings of CO toxicity in children, both in utero and in early childhood, have shown wide-ranging degrees of neurobehavioral injury. Faulty heater installation resulted in a family being exposed to high levels of accumulated CO in an apartment in which they had lived for 1 year. A 4-year-old boy and a 5-year-old girl were hospitalized with COHb levels of 20% and 13% respectively, and were treated with oxygen through masks. Over the course of a year following exposure, behavioral symptoms were recurrent: headaches, lethargy, sleepiness, hyperactivity, violent tantrums, and extreme mood swings in the boy, and irritability, intermittent fatigue, and falling asleep in school in the girl. However, testing of cognitive ability (receptive language, vocabulary, adaptive behavior) was normal in both children 5 months after admission for treatment. The mechanism of the effect was thought to be hypoxemia (rather than hypoxia) partly because psychometric tests were normal.60 In another study, a 22-year-old pregnant woman exposed to elevated CO levels, who developed neurologic symptoms as well as tachycardia, tachypnea, signs of preterm labor, and a mildly elevated COHb level, delivered her baby at full term, without sequelae for the infant, following treatment with hyperbaric O2 .61

25.6.3 DEVELOPMENTAL NEUROPSYCHOLOGICAL APPROACH TO ASSESSING TOXICITY EFFECTS The particular imprint CO toxicity bears on a child’s functioning may represent the toxicant’s direct impact on specific cell groups or brain structures. Impact on neural structures which are vulnerable to CO toxicity and which are essential to the expression of certain functions will be represented as focal effects of injury. As examples, injury to the globus pallidus and the basal ganglia will affect the development of motor repertoires, and injury to structures critically involved in memory, such as the hippocampus, will disrupt this domain of cognitive functioning.11 In contrast to the specific brain-behavior relationships which can, at least to a significant degree, be ascribed to injury in the well-differentiated adult brain, the assessment of injury in the less-differentiated brain must consider the injury’s potential for interfering with complex processes that interact with one another and that represent consequences of both biological and psychosocial developmental forces. A multiplicity of neural and nonneural factors interact to determine a toxin’s impact in children’s functioning.62 In the child, whose behavioral repertoire is in the process of developing, toxins do not simply hit given neural structures—the “where”; they hit a “what” at both a “where” and a “when”.63 Because knowledge of a toxin’s particular signature in behavior demands a consideration of how injury to specific structures (the “where”) disrupts a developing system over time (the “what” and the “when”), the evaluation of the functional effects of toxicity in a child suspected of CO poisoning is necessarily multifaceted. A developmental neuropsychological approach to assessing effects of toxicity proposed by Bernstein64 stipulates that two levels of data-gathering are needed to account for the interactions possible between exposure to a toxicant and the developmental events occurring at the time of exposure. At a first level of assessment, data are obtained in broad domains of functioning to determine whether or not

8417: “8417_c025” — 2007/9/11 — 12:14 — page 575 — #7

576

Carbon Monoxide Poisoning

deficits actually exist. Domains of interest include behavioral regulation, attentional capacities, learning skills, memory capacities, problem-solving skills, motor skills (gross, fine, graphomotor), sensory capacities, general cognitive ability, communicative competence/language abilities, visuospatial processing, social cognition, socio-emotional status/adjustment, and academic achievement. These domains are tapped by the direct physical and neurological examination of the child, administration of standardized tests, and observation of the child’s behavior in both naturalistic and structured situations. Naturalistic observations complemented by information from behavioral questionnaires and history-taking interviews, contributed by caretakers including parents, teachers, and health-care professionals, form relevant components of assessment. Their importance resides in the fact that the well-structured condition of formal testing may not detect the derailing of self-regulatory functions in the child’s familiar and habituated settings. Often, the injured child’s behavior is much worse at home than in school in part because behavioral disruption may actually be leading to a child’s inability to organize moment-to-moment behavior when the situation is not strongly structuring environmental stimuli. That such derailing is possible as a result of toxicity requires documenting the child’s ability to adapt to the demands of ordinary situations as an important aspect of evaluation, one usually based on family report. As proposed by Bernstein,64 the second level of a developmental neuropsychological assessment moves from identifying deficits to determining their linkages to a particular neurotoxicant. Knowledge of the toxicant’s mode of action and predilection for specific brain systems and/or processes, as well as of the developmental significance of the timing, amount, and duration of exposure, determines the ability to address the questions of specificity, sensitivity, and causation posed at this level. Information on the where, when, and what of CO toxicity comes from animal studies on early brain injury. However, the “where” of CO toxicity (decreased oxygenation of blood) and the “when” (last trimester) in the association between developmental age and CO’s power to produce injury, leave open the knowledge of “what” is disrupted as a result of damage. In work with pre- and postgestational rats and kittens, different functional outcomes have been robustly associated with the precise age at injury.65 The most optimal outcomes have been associated with embryonic stages characterized by maximal astrocyte generation and synapse development, when the developing brain carries the potential for recovery due to the redundancy in neural generation occurring at that point. Conversely, the developing brain is at its worst time for compensation when it is in a stage of neural migration and of the initiation of synaptic formation. In rats, this is the first week of life, a period that is likened to the third trimester in the human fetus.65 During a period when the fetus is most vulnerable to CO poisoning owing to pulmonary development, it is also at a stage when genetically-programmed developmental processes work against its ability to withstand the effects of neural toxicity. Thus, the timing of a CO insult can cause cumulative biological risks66 to a developing system, which can be diverse and lifelong in their aversive impacts).67 Four case studies are next presented to describe the wide-ranging variability in outcomes possible in cases of pediatric CO toxicity.

8417: “8417_c025” — 2007/9/11 — 12:14 — page 576 — #8

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children

577

25.7 CASE STUDIES The following case studies are shown to illustrate the complexity of consequences possible in young children following CO exposures. The first three cases, CB, LB and DB, are siblings with chronic lower-level exposures lasting from 4.5 years in two children (ages 18 months and 3 years at the beginning of exposure), to lower level in utero exposure plus 2.5 postnatal years in the youngest. A fourth case is presented, HK, who had chronic exposure in utero and postnatally, with a probable peak exposure at almost two years of age. Table 25.1 summarizes psychometric test data as well as information obtained clinically. The children’s neuropsychological findings demonstrate a range of outcome morbidity. Test batteries for the children vary largely because of the applicability of measures at different age groups. For example, there was no neuropsychological battery for the 4-year-old at the time of the assessment, and the Wechsler Intelligence Battery for Children68 was a primary instrument, whereas by age six, a theoretically based neuropsychological battery 69 could be used. However, the constructs that the different tests measured are given, and all tests are converted to a single standard (i.e., percentile), in order to facilitate comparisons.

25.7.1 OBSERVATIONS The two cases exposed in utero (DB and HK) showed respiratory symptoms (postnatal plethora, chronic respiratory infections, asthma, premature lung development). Respiratory problems were not observed in the other two children whose exposure onset was at 4.5 and 2.5 years of age. Skin conditions were observed in three of the four children, all from the same household and thus the same exposure conditions. Systemic problems that were observed in all cases were nervous habits, difficulty concentrating, emotional lability and irritability, and attention deficit disorders. Partial autistic behaviors were observed in the child with the most severe exposure (HK), which included regression in speech at age two years, poor eye contact, difficulty playing rule-guided games, lack of imaginative play, and little social cognition. Furthermore, the longitudinal observations in this child with the most severe exposure indicated that he changed from a pleasant-appearing but poorly concentrating child at age four years, to a rigid, fearful, and easily frustrated child at age six and a half. Motor development was delayed in two of the four children (LB and HK). Some dyspraxia was measured in all four children. In the child with longitudinal data, his ability to copy figures (constructional or graphomotor praxis) declined significantly in relation to developmental expectations. By 6.5 years of age, he was significantly impaired in all motor skills tested. Tactile discrimination (identifying two fingers touched simultaneously) was impaired in the two oldest children in whom this could be tested. Developmental language impairment was often observed by the parents as poor comprehension or impaired articulation, and slow or regressed speech development in two children (LB and HK). The child with the most severe exposure as well as a history of familial reading disability (HK) had receptive and expressive language

8417: “8417_c025” — 2007/9/11 — 12:14 — page 577 — #9

8417: “8417_c025” — 2007/9/11 — 12:14 — page 578 — #10 Faulty furnace emissions and ventilation into house 4 years, 4.0 months

Faulty furnace emissions and ventilation into house 7 years, 0 months

Faulty furnace emissions and ventilation into house 8 years, 7.5 months

Age at test

Chronic

Chronic

Chronic

Peak or chronic CO exposure Source of CO exposure

2.5 years + in utero

4.5 years

4.5 years

From conception

DB

Years of CO exposure

LB 1.5 years

CB

Cases

Clinical characteristics Age at initial CO 3 years exposure

Case Initials

TABLE 25.1 Summary of Psychometric Test and Clinical Data for Four Case Studies

4 years, 1.5 months

Faulty furnace

Chronic low-level in utero, and peak exposure age 1 year, 11 months Possible chronic exposure in utero, one documented peak exposure with unresponsiveness, vomiting, diarrhea, pallor, dilated, equal, slowed pupil responses; other less severe peak exposures possible. Total exposure 2 years + in utero Chronic + peak

HK Time 1

6 years, 6.0 months

Faulty furnace

HK Time 2

578 Carbon Monoxide Poisoning

4 years college 2 years college None, though doesn’t enjoy school

Right None known

Hand dominance Other risk factors including family history

Parental educational level Mother Father Depression

Chronic cough, fatigue, skin bumps, itching, restlessness, difficulty concentrating, difficulty learning, sleep disturbance, chronic diarrhea, frequent muscle cramps, nausea/vomiting, irritable, hyperactive, episode of lethargy and hallucination, nervous habits, messy habits/smearing of feces, temper tantrums

Symptoms observed by family without other known cause

4 years college 2 years college None, happy, even-tempered

Right None known

Skin bumps, flushing of cheeks, speech problems, loss of appetite, nausea, irritability, severe constipation, incident of acute flu-like symptoms and cyanotic hands, fine motor incoordination and dressing difficulty at age of 6 years, problems of speech articulation

4 years college 2 years college None, happy, outgoing, verbally quiet

Right None known

Red skin at birth that normalized, became plethoric 2 days after birth and required suctioning, chronic cough and respiratory problems that onset postnatally, often tired, low energy, frequent knee pain, impaired articulation and reduced intelligibility of speech, dysphagia, asthma, borderline anemia

12th grade 10th grade None reported, pleasant presentation

Right Paternal reading disability (mild); mother smoked 15 cigarettes/day during pregnancy

Asthma onset postnatally, premature lung development, failed to crawl, hyperactive, falls often, impaired concentration, emotional lability, bites fingernails, temper tantrums, fatigued

8417: “8417_c025” — 2007/9/11 — 12:14 — page 579 — #11

(Continued)

None, fearful, easily frustrated, rigid

Same

Symptoms persist with exception of hyperactivity but Attention Deficit Disorder persists, poor attention/concentration, poor comprehension, overly dependent on parents, partial autistic-like behaviors (regression in speech at age 2 years following peak exposure) with recovery of speech, lacks imaginary play, poor eye contact, difficulty playing rule-guided games, little interest in others No change Same

Functional and Developmental Effects of Carbon Monoxide Toxicity in Children 579

Slow, first noticed after 2 years of exposure

Timing of onset of symptoms

Neuropsychological findings Sensory and motor function Copying hand 9th percentile movements Copying figures 25th percentile (NEPSY) Grooved Pegboard test 63rd percentile on right 25th percentile on left

Normal

CB

Early motor and speech development

Case Initials

TABLE 25.1 (Continued)

25th percentile 63rd percentile (NEPSY)

9th percentile (NEPSY) 32nd percentile on right 19th percentile on left

Lethargy and viral-like symptoms postnatally

Normal

DB

9th percentile

Slow

Motor development normal prior to CO exposure; speech development slowed following 6 months of CO exposure

LB

Cases

66th percentile (Beery)

Developmental and neurological evaluations at age 2 years 7–8 months revealed clumsiness, impaired balance, lack of speech or language progression after 3 months of therapy; delay in behaviors: cognitive 8 months, gross motor 10 months, fine motor 7 months, personal/social 9 months Respiratory symptoms postnatally; speech/motor symptoms after peak exposure

HK Time 1

5th percentile (NEPSY)

5th percentile

Globally impaired, with greater expressive speech impairment—first sentence structure at age 5.5 years, communicates primarily by gesture; genetic syndrome and autism ruled out

HK Time 2

580 Carbon Monoxide Poisoning

8417: “8417_c025” — 2007/9/11 — 12:14 — page 580 — #12

Impaired: age seven equivalent

Body part naming Attention and information processing speed Symbol digit 47th percentile modalities test—oral version (coding) Coding (digit symbol subtest—written version) Trail making test—A 31st percentile (tracking of numbers) Trail making test—B 19th percentile (tracking of letters and numbers) Visual cancellation 63rd percentile tests Language Word fluency 99th percentile (phonemic) Semantic fluency 84th percentile Token test (linguistic 81st percentile comprehension) Comprehension of instructions Peabody picture 77th percentile vocabulary test-III

Visuomotor precision Oral praxis Finger localization test

77th percentile

>80th percentile 19th percentile

5th percentile

63rd percentile

1st percentile

1st percentile

n/a

Impaired: age six equivalent

73rd percentile

84th percentile

75th percentile

75th percentile

84th percentile

50th percentile

75th percentile 1st percentile