United States Sentemher 2009
Environmental Protection aeptemoer ZUUV
Agency EPA/600/R-09/019B
Integrated Science Assessment for
Carbon Monoxide -
Second External Review Draft
National Center for Environmental Assessment-RTF Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document is the second external review draft for review purposes only and does not constitute
U.S. Environmental Protection Agency policy. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
September 2009 \\ DRAFT-DO NOT QUOTE OR CITE
-------
Authors and Contributors
Authors
Dr. Thomas Long (CO Team Leader)—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Craig Hansen— Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens— National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Matthew Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
September 2009 iii DRAFT - DO NOT CITE OR QUOTE
-------
Dr. Kazuhiko Ito—Department of Environmental Medicine, New York University School of
Medicine, Tuxedo, NY
Dr. Jennifer Peel—Department of Environmental and Radiological Health Sciences, Colorado State
University, Fort Collins, CO
Contributors
Dr. Richard Baldauf—National Risk Management Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Vernon Benignus—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Lance McCluney—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Kris Novak—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Adam Reff—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Mark Schmidt—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Rhonda Thompson—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Reviewers
Dr. Richard Baldauf—National Risk Management Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Vernon Benignus—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Souad Benromdhane—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Philip Bromberg—School of Medicine, University of North Carolina, Chapel Hill, NC
Dr. Matthew Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Daniel Costa—National Program Director for Air, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Andrew Ohio—National Health and Environmental Effects Research Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Kazuhiko Ito—Department of Environmental Medicine, New York University School of
Medicine, Tuxedo, NY
September 2009 iv DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Petros Koutrakis—Harvard School of Public Health, Harvard University, Cambridge, MA
Mr. John Langstaff—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barry Lefer—Department of Geosciences, University of Houston, Houston, TX
Dr. Karen Martin—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Dave McKee—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Connie Meacham—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ines Pagan—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Parker—National Center for Health Statistics, Centers for Disease Control, Atlanta, GA
Dr. Jennifer Peel—Department of Environmental and Radiological Health Sciences, Colorado State
University, Fort Collins, CO
Dr. Pradeep Raj an—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Harvey Richmond—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph Somers—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
Dr. John Vandenberg—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Alan Vette—National Exposure Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. William Vizuete—Department of Environmental Sciences and Engineering, University of North
Carolina, Chapel Hill, NC
Ms. Debra Walsh—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Lin Weaver—Department of Internal Medicine, LDS Hospital, Salt Lake City, UT
Dr. Lewis Weinstock—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Ron Williams—National Exposure Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
September 2009 v DRAFT-DO NOT QUOTE OR CITE
-------
CO Project Team
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTF Division,
Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division,
Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Scientific Staff
Dr. Thomas Long (CO Team Leader)—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Craig Hansen— Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens— National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
September 2009 vi DRAFT-DO NOT QUOTE OR CITE
-------
Dr. Joseph Pinto—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Technical Support Staff
Ms. Laeda Baston— National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Ellen Lorang— National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Deborah Wales—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Barbara Wright— National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
September 2009 vii DRAFT-DO NOT QUOTE OR CITE
-------
Clean Air Scientific Advisory Committee
- CO NAAQS Review Panel
Chairperson
Dr. Joseph Brain*, Department of Environmnetal Health, Harvard School of Public Health, Harvard
University, Boston, MA
Members
Dr. Thomas Dahms, Department of Anesthesiology Research and Critical Care, St. Louis University
School of Medicine, St. Louis, MO
Dr. Russell R. Dickerson, Department of Meteorology, University of Maryland, College Park, MD
Dr. Laurence Fechter, Research Service, Department of Veterans Affairs, Loma Linda VA Medical
Center, Loma Linda, CA
Dr. H. Christopher Frey*, College of Engineering, Department of Civil, Construction, and
Environmental Engineering, North Carolina State University, Raleigh, NC
Dr. Milan Hazucha, Department of Medicine, Center for Environmental Medicine, Asthma and Lung
Biology, University of North Carolina, Chapel Hill, NC
Dr. Michael T. Kleinman, Department of Community & Environmental Medicine, University of
California-Irvine, Irvine, CA
Dr. Arthur Penn, Department of Comparative Biomedical Sciences, Louisiana State University
School of Veterinary Medicine, Baton Rouge, LA
Dr. Beate Ritz, School of Public Health, Epidemiology, University of California at Los Angeles, Los
Angeles, CA
Dr. Paul Roberts, Sonoma Technology, Inc., Petaluma, CA
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering, Georgia Institute
of Technology, Atlanta, GA
Dr. Stephen R. Thorn, Institute for Environmental Medicine, University of Pennsylvania,
Philadelphia, PA
* Members of the statutory Clean Air Scientific Advisory Committee (CAS AC) appointed by the
EPA Administrator
September 2009 viii DRAFT-DO NOT QUOTE OR CITE
-------
Science Advisory Board Staff
1 Dr. Ellen Rubin, Designated Federal Officer, 1200 Pennsylvania Avenue, N.W., Washington, DC,
2 20460, Phone: 202-343-9975, Fax: 202-233-0643, Email: rubin.ellen@epa.gov
Physical/Courier/FedEx Address:
Dr. Ellen Rubin, U.S. EPA Science Advisory Board Staff Office, Mail Code 1400F, Woodies
Building, Room 3610E, 1025 F Street, N.W., Washington, DC 20004
September 2009 ix DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
AUTHORS AND CONTRIBUTORS
TABLE OF CONTENTS
LIST OF TABLES xv
LIST OF FIGURES xix
ACRONYMS AND ABBREVIATIONS xxv
CHAPTER 1. INTRODUCTION 1-1
1.1. Legislative Requirements 1-2
1 .2. History of the NAAQS for CO
1.3. ISA Development
1.4. Document Organization
1.5. Document Scope
1 .6. EPA Framework for Causal Determination
1.6.1. Scientific Evidence Used in Establishing Causality
1.6.2. Association and Causation
1 .6.3. Evaluating Evidence for Inferring Causation
1 .6.4. Application of Framework for Causal Determination
1 .6.5. Determination of Causality
1.6.5.1. Effects on Human Populations
1.6.5.2. Effects on Ecosystems or Public Welfare
1 .6.6. Concepts in Evaluating Adversity of Health Effects
1.7. Summary
1-4
1-5
1-9
1-10
1-11
1-12
1-12
1-13
1-17
1-19
1-21
1-22
1-23
1-24
References 1-25
CHAPTER 2. INTEGRATIVE HEALTH EFFECTS OVERVIEW 2-1
2.1. Ambient CO Sources and Concentrations 2-2
2.2. Climate Forcing Effects 2-3
2.3. Exposure to Ambient CO 2-4
2.4. Dosimetry, Pharmacokinetics, and Mode of Action 2-5
2.4.1. Dosimetry and Pharmacokinetics 2-5
2.4.2. Mode of Action 2-6
2.5. Health Effects 2-7
2.5.1. Cardiovascular Morbidity 2-8
2.5.2. Central Nervous System Effects 2-9
2.5.3. Birth Outcomes and Developmental Effects 2-10
2.5.4. Respiratory Morbidity 2-12
2.5.5. Mortality 2-13
2.6. Policy-Relevant Considerations 2-15
2.6.1. Susceptible Populations 2-15
September 2009 x DRAFT-DO NOT QUOTE OR CITE
-------
2.6.2. Concentration-Response Relationship 2-18
2.7. Integration of CO Health Effects 2-19
References 2-26
CHAPTER 3. SOURCE TO EXPOSURE 3-1
3.1. Introduction 3-1
3.2. Sources and Emissions of CO 3-1
3.3. Physics and Chemistry of Atmospheric CO 3-10
3.3.1. CO Climate Forcing Effects 3-13
3.4. Ambient Measurements 3-18
3.4.1. Ambient Measurement Instruments 3-18
3.4.2. Ambient Sampling Network Design 3-22
3.4.2.1. Monitor Siting Requirements 3-22
3.4.2.2. Spatial and Temporal Coverage 3-23
3.5. Environmental Concentrations 3-32
3.5.1. Spatial Variability 3-32
3.5.1.1. National Scale 3-32
3.5.1.2. Urban Scale 3-42
3.5.1.3. Micro-to- Neighborhood Scale and the Near-Road Environment 3-57
3.5.2. Temporal Variability 3-70
3.5.2.1. Multiyear Trends 3-70
3.5.2.2. Hourly Variation 3-72
3.5.3. Associations with Copollutants 3-76
3.5.4. Policy-Relevant Background 3-81
3.5.4.1. Surface-based Determinations 3-82
3.5.4.2. Limitations of Other Possible Methods 3-84
3.6. Issues in Exposure Assessment 3-86
3.6.1. Summary of Findings from 2000 CO AQCD 3-86
3.6.2. General Exposure Concepts 3-87
3.6.3. Exposure Modeling 3-89
3.6.3.1. Stochastic Population-Based Time-Weighted Microenvironmental
Exposure Models 3-89
3.6.3.2. Using Spatial Models to Estimate Exposure 3-92
3.6.4. Personal Exposure Monitors for CO 3-94
3.6.5. Indoor Exposure to CO 3-95
3.6.5.1. Infiltration of Ambient CO 3-95
3.6.5.2. Exposure to Nonambient CO 3-96
3.6.6. Exposure Assessment Studies at Different Spatial Scales 3-98
3.6.6.1. Neighborhood- to Urban-Scale Studies of Ambient CO Exposure 3-98
3.6.6.2. Microscale Studies of Ambient CO Exposure: Near-Road and On-Road
Exposures 3-100
3.6.7. Association between Personal CO Exposure and Copollutants 3-105
3.6.8. Implications for Epidemiology 3-106
3.6.8.1. Measurement Error 3-107
3.6.8.2. Exposure Issues Related to Nonambient CO 3-108
3.6.8.3. Spatial Variability 3-109
3.6.8.4. Temporal Variability 3-110
3.6.8.5. CO Exposure in Copollutant Mixtures 3-111
3.6.8.6. Conclusions 3-112
3.7. Summary and Conclusions 3-113
3.7.1. Sources of CO 3-113
3.7.2. Physics and Chemistry of Atmospheric CO and Related Climate Forcing Effects 3-113
3.7.3. Ambient CO Measurements 3-114
September 2009 xi DRAFT-DO NOT QUOTE OR CITE
-------
3.7.4. Environmental CO Concentrations 3-115
3.7.5. Exposure Assessment and Implications for Epidemiology 3-116
References 3-118
CHAPTER 4. DOSIMETRY AND PHARMACOKINETICS OF CARBON MONOXIDE 4-1
4.1. Introduction 4-1
4.2. Carboxyhemoglobin Modeling 4-2
4.2.1. The Coburn-Forster-Kane and Other Models 4-2
4.2.2. Multicompartment Models 4-6
4.2.3. Model Comparison 4-8
4.2.4. Mathematical Model Usage 4-9
4.3. Absorption, Distribution, and Elimination 4-13
4.3.1. Pulmonary Absorption 4-13
4.3.1.1. Mass Transfer of Carbon Monoxide 4-14
4.3.1.2. Lung Diffusion of Carbon Monoxide 4-15
4.3.2. Tissue Uptake 4-16
4.3.2.1. The Respiratory Tract 4-16
4.3.2.2. The Blood 4-16
4.3.2.3. Heart and Skeletal Muscle 4-18
4.3.2.4. Other Tissues 4-20
4.3.3. Pulmonary and Tissue Elimination 4-21
4.3.4. COHb Analysis Methods 4-23
4.4. Conditions Affecting Uptake and Elimination 4-23
4.4.1. Environment and Activity 4-23
4.4.2. Altitude 4-24
4.4.3. Physical Characteristics 4-25
4.4.3.1. Fetal Pharmacokinetics 4-26
4.4.4. Health Status 4-27
4.5. Endogenous CO Production and Metabolism 4-27
4.6. Summary and Conclusions 4-33
References 4-35
CHAPTERS. INTEGRATED HEALTH EFFECT 5-1
5.1. Mode of Action of CO Toxicity 5-1
5.1.1. Introduction 5-1
5.1.2. Hypoxic Mechanisms 5-1
5.1.3. Non-Hypoxic Mechanisms 5-2
5.1.3.1. Non-Hypoxic Mechanisms Reviewed in the 2000 CO AQCD 5-3
5.1.3.2. Recent Studies of Non-Hypoxic Mechanisms 5-5
5.1.3.3. Implications of Non-Hypoxic Mechamisms 5-13
5.1.3.4. Summary 5-17
5.2. Cardiovascular Effects 5-18
5.2.1. Epidemiologic Studies with Short-Term Exposure 5-18
5.2.1.1. Heart Rate and Heart Rate Variability 5-19
5.2.1.2. ECG Abnormalities Indicating Ischemia 5-24
5.2.1.3. Arrhythmia 5-24
5.2.1.4. Cardiac Arrest 5-26
5.2.1.5. Myocardial Infarction 5-27
5.2.1.6. Blood Pressure 5-27
5.2.1.7. Vasomotor Function 5-28
5.2.1.8. Blood Markers of Coagulation and Inflammation 5-28
5.2.1.9. Hospital Admissions and Emergency Department Visits 5-32
September 2009 xii DRAFT-DO NOT QUOTE OR CITE
-------
5.3.
5.4.
5.5.
5.6.
5.7.
5.2.2. Epidemiologic Studies with Long-Term Exposure
5.2.3. Summary of Epidemiologic Studies of Exposure to CO and Cardiovascular
Effects
5.2.4. Controlled Human Exposure Studies
5.2.5. Toxicological Studies
5.2.5.1. Endothelial Dysfunction
5.2.5.2. Cardiac Remodeling Effects
5.2.5.3. Electrocardiographic Effects
5.2.5.4. Summary of Cardiovascular Toxicology
5.2.6. Summary of Cardiovascular Effects
Central Nervous System Effects
5.3.1 . Controlled Human Exposure Studies
5.3.2. Toxicological Studies
5.3.3. Summary of Central Nervous System Effects
Birth Outcomes and Developmental Effects
5.4.1. Epidemiologic Studies
5.4.1.1. Preterm Birth
5.4.1.2. Birth Weight, Low Birth Weight, and Intrauterine Growth
Restriction/Small for Gestational Age
5.4.1.3. Congenital Anomalies
5.4.1.4. Neonatal and Post-Neonatal Mortality
5.4.1.5. Summary of Epidemiologic Studies of Birth Outcomes and
Developmental Effects
5.4.2. Toxicological Studies of Birth Outcomes and Developmental Effects
5.4.2.1. Birth Outcomes
5.4.2.2. Developmental Effects
5.4.3. Summary of Birth Outcomes and Developmental Effects
Respiratory Effects
5.5.1 . Epidemiologic Studies with Short-Term Exposure
5.5.1.1. Pulmonary Function, Respiratory Symptoms, and Medication Use
5.5.1.2. Respiratory Hospital Admissions, ED Visits and Physician Visits
5.5.2. Epidemiologic Studies with Long-Term Exposure
5.5.2.1. Pulmonary Function
5.5.2.2. Asthma and Asthma Symptoms
5.5.2.3. Allerav
5.5.2.4. Summary of Associations between Long-Term Exposure to CO and
Respiratory Morbidity
5.5.3. Controlled Human Exposure Studies
5.5.4. Toxicological Studies
5.5.5. Summary of Respiratory Health Effects
5.5.5.1. Short-Term Exposure to CO
5.5.5.2. Long-Term Exposure to CO
Mortality
5.6.1 . Epidemiologic Studies with Short-Term Exposure to CO
5.6. 1 . 1 . Summary of Findings from 2000 CO AQCD
5.6.1.2. Multicitv Studies
5.6.1.3. Meta-Analysis of All Criteria Pollutants
5.6.1.4. Sinqle-Citv Studies
5.6.1.5. Summary of Mortality and Short-Term Exposure to CO
5.6.2. Epidemiologic Studies with Long-Term Exposure to CO
5.6.2.1. U.S. Cohort Studies
5.6.2.2. U.S. Cross-Sectional Analysis
5.6.2.3. Summary of Mortality and Long-Term Exposure to CO
Susceptible Populations
5.7.1. Pre-Existing Disease
5.7.1.1. Cardiovascular Disease
5-56
5-57
5-57
5-61
5-62
5-63
5-65
5-66
5-66
5-68
5-68
5-68
5-69
5-70
5-70
5-70
5-74
5-84
5-86
5-88
5-89
5-90
5-96
5-113
5-115
5-115
5-115
5-126
5-136
5-137
5-138
5-139
5-140
5-141
5-141
5-143
5-143
5-144
5-144
5-144
5-145
5-145
5-153
5-153
5-157
5-158
5-160
5-164
5-165
5-166
5-167
5-168
September 2009 xiii DRAFT-DO NOT QUOTE OR CITE
-------
5.8.
5.7.1.2. Obstructive Lung Disease
5.7.1.3. Anemia
5.7.1.4. Diabetes
5.7.2. Lifestaqe
5.7.2.1. Older Adults
5.7.2.2. Gestational Development
5.7.3. Gender
5.7.4. Altitude
5.7.5. Exercise
5.7.6. Proximity to Roadways
5.7.7. Medications and Other Substances
5.7.8. Summary of Susceptible Populations
Summary
References
5-169
5-170
5-171
5-172
5-172
5-173
5-174
5-175
5-176
5-176
5-177
5-177
5-179
5-180
ANNEX A. ATMOSPHERIC SCIENCE A-1
ANNEXE. DOSIMETRY STUDIES B-1
References B-6
ANNEX C. EPIDEMIOLOGY STUDIES C-1
References C-96
ANNEX D. CONTROLLED HUMAN EXPOSURE STUDIES D-1
References D-5
ANNEX E. TOXICOLOGICAL STUDIES E-1
References E-24
September 2009 xiv DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Table 1-1
Table 1-2
Table 2-1
Table 2-2
Table 3-1
Table 3-2
Table 3-3
Table 3-4
Table 3-5
Table 3-6
Table 3-7
Table 3-8
Table 3-9
Table 3-10
Table 3-1 1
Table 3-1 2
Table 3-1 3
Table 4-1
Table 4-2
Table 4-3
Table 5-1
Table 5-2
Table 5-3
Table 5-4
Aspects to aid in iudqinq causality.
Weiqht of evidence for causal determination.
Causal determinations for health effects cateqories.
Range of mean and 99th percentile concentrations (ppm) in US and Canadian studies of
short-term CO exposure and CVD hospitalizations.
Literature values for CO yields from hydrocarbons in per carbon units except as noted.
Specific hydrocarbons are noted in parentheses.
Performance specifications for analytical detection of CO, based on 40 CFR Part 53.
Counts of CO monitors by sampling scale meeting 75% completeness criteria for use in the
U.S. durinq 2005-2007.
Proximity to CO monitors for the total population by city.
Proximity to CO monitors for adults aqed 65 and older by city.
Distribution of 1-h avq CO concentration (ppm) derived from AQS data.
Distribution of 24-h avq CO concentration (ppm) derived from AQS data.
Distribution of 1-h daily max CO concentration (ppm) derived from AQS data.
Distribution of 8-h daily max CO concentration (ppm) derived from AQS data.
Table of inter-sampler comparison statistics, as defined in the text, including Pearson r,
P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to AQS for
2005-2007 in Denver, CO.
Table of inter-sampler comparison statistics, as defined in the text, including Pearson r,
P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to AQS for
2005-2007 in Los Anqeles, CA.
National distribution of all hourly observations, 1-h daily max, 1-h daily average, and 8-h
daily max concentration (ppm) derived from AQS data, based on monitor scale
designations, 2005-2007.
Percentage of time exposed to ambient CO (adjusted to reflect the absence of non-ambient
CO from ETS and gas cooking), average CO exposures, and percentage of exposure
estimated for the population.
Predicted COHb levels resulting from 1 , 8, and 24 h CO exposures in a modeled human at
rest
CO concentration in pmol/mg wet weight tissue and fold tissue CO concentration changes
[normalized to background tissue concentrations! - human.
CO concentration in pmol/mg fresh weight tissue and fold tissue CO concentration changes
[normalized to background tissue concentrationsl - adult mouse.
Responses to low and moderate CO exposures.
Tissue concentration of CO following inhalation exposure.
Tissue concentration of CO following increased endogenous production.
Summary of studies investigating the effect of CO exposure on HRV parameters.
1-18
1-20
2-8
2-22
3-12
3-20
3-26
3-31
3-31
3-35
3-37
3-38
3-40
3-46
3-51
3-59
3-99
4-10
4-20
4-20
5-12
5-15
5-16
5-23
September 2009 xv DRAFT-DO NOT QUOTE OR CITE
-------
Table 5-5
Table 5-6
Table 5-7
Table 5-8
Table 5-9
Table 5-10
Table 5-1 1
Table 5-1 2
Table 5-1 3
Table 5-1 4
Table 5-1 5
Table 5-1 6
Table 5-1 7
Table 5-1 8
Table 5-1 9
Table 5-20
Table 5-21
Table 5-22
Table 5-23
Table 5-24
Table 5-25
Table 5-26
Table A-1
Table A-2
Table A-3
Table A-4
Table A-5
Summary of studies investiqatinq the effect of CO exposure on cardiac arrhythmias.
Summary of studies investigating the effect of CO exposure on blood markers of
coaqulation and inflammation.
Summary of CHD hospital admission studies.1
Summary of stroke hospital admission studies.1
Summary of CHF hospital admission studies.
Association of ambient air pollution levels and cardiovascular morbidity in visits with and
without specific secondary conditions.
Summary of non-specific CVD hospital admission studies.
Brief summary of PTB studies.
Brief summary of birth weiqht studies.
Behavioral responses to low and moderate CO exposure
Neuronal responses to low and moderate CO exposure
Neurotransmitter chanqes from low and moderate CO exposure
Developinq auditory system responses to low and moderate CO exposure
Cardiovascular and systemic developmental responses to low and moderate CO exposure
Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with short-term exposure to CO.
Range of CO concentrations reported in key respiratory hospital admission and ED visit
studies that examine effects associated with short-term exposure to CO.
Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with lonq-term exposure to CO.
Range of CO concentrations reported in multicity studies that examine mortality effects
associated with short-term exposure to CO.
Range of CO concentrations reported in single-city studies that examine mortality effects
associated with short-term exposure to CO.
Range of CO concentrations reported in U.S.-based studies that examine mortality effects
associated with lonq-term exposure to CO.
Definitions of susceptible and vulnerable in the CO literature.
Percent of the U.S. population in 2007 with respiratory diseases and cardiovascular
diseases.
Listinq of all carbon monoxide monitors currently in use, alonq with their limits of detection.
Microscale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes that the
value was not reported.
Middle scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes that
the value was not reported.
Neighborhood scale monitors meeting 75% completeness criteria, 2005-2007. "NR"
denotes that the value was not reported.
Urban scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes that
5-26
5-32
5-38
5-43
5-47
5-49
5-53
5-74
5-83
5-98
5-102
5-104
5-108
5-111
5-117
5-127
5-137
5-146
5-154
5-159
5-167
5-169
A-8
A-9
A-11
A-1 2
the value was not reported.
A-15
September 2009
XVI
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-6
Table A-7
Table A-8
Table A-9
Table A-10
Table A-11
TableA-12
Table A-13
Table A-14
Table A-15
Table A-16
Table A-17
Table A-18
Table A-19
Table A-20
Table A-21
Table A-22
Table A-23
Regional scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes
that the value was not reported.
A-15
Monitors meeting 75% completeness criteria, 2005-2007 with no scale delared. "NR"
denotes that the value was not reported.
Numbers of high LOD and trace-level monitors in each state that met completeness criteria
for 2005-2007.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Anchorage, AK.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in Atlanta,
GA.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in Boston,
MA.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Houston, TX.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in New
York City, NY.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Phoenix, AZ.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Pittsburgh, PA.
Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in St.
Louis, MO.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Atlanta, GA.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Boston, MA.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Denver, CO.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Houston, TX.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Los Angeles, CA.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for New York, NY.
A-16
A-18
A-31
A-34
A-37
A-40
A-43
A-46
A-49
A-54
A-56
A-56
A-57
A-57
A-58
A-59
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Phoenix, AZ.
A-60
September 2009
XVII
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-24
Table A-25
Table A-26
Table B-1
Table C-1
Table C-2
Table C-3
Table C-4
Table C-5
Table C-6
Table C-7
Table C-8
Table D-1
Table E-1.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Pittsburgh, PA.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for Seattle, WA.
Comparison of distributional data at different monitoring scales for hourly, 1-h daily max,
24-h avg, and 8-h daily max data for St. Louis, MO.
Recent studies related to CO dosimetry and pharmacokinetics.
Studies of CO exposure and cardiovascular morbidity.
Studies of CO exposure and cardiovascular hospital admissions and ED visits._
Studies of CO exposure and neonatal and postneonatal outcomes.
Studies of short-term CO exposure and respiratory morbidity
Studies of short-term CO exposure and respiratory hospital admissions and ED visits..
Studies of long-term CO exposure and respiratory morbidity.
Studies of short-term CO exposure and mortality.
Studies of long-term CO exposure and mortality.
Controlled human exposure studies.
Human and animal studies.
. A-61
. A-61
. A-62
_ B-1
C-1
.C-1 3
. C-23
. C-32
. C-41
. C-64
. C-69
. C-91
E-1
September 2009
XVIII
DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Figure 1-1
Figure 2-1
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Figure 3-5
Figure 3-6
Figure 3-7
Figure 3-8
Figure 3-9
Figure 3-10
Figure 3-1 1
Figure 3-1 2
Figure 3-1 3
Figure 3-1 4
Figure 3-1 5
Figure 3-1 6
Figure 3-1 7
Figure 3-1 8
Figure 3-1 9
Figure 3-20
Identification of studies for inclusion in the ISA.
Excess risk estimates from epidemiologic studies of short-term CO exposure and CVD
hospitalizations alonq with mean and 99th percentile CO concentrations.
CO emissions (tons) in the U.S. bv source sector in 2002.
Trends in anthropogenic CO emissions (MT) in the U.S. by source category for 1990 and
1996-2002.
Mean CO concentrations in the boundary layer (0-1 .5 km altitude) during the ICARTT
campaign (July 1 -August 15, 2004) (left).
Surface air CO concentrations at Cheboque Point during the ICARTT campaign.
CO concentrations measured by satellite at the 700 hectoPascal level (~1 0,000 feet above
sea level) from MOPITT for the period 15-23 July 2004 during intense wildfires in Alaska
and Yukon.
Trends in sub-national CO emissions in the 1 0 U.S. EPA Regions for 1 990 and 1 996-2002.
CO emissions density map and distributions for the state of Colorado, and for selected
counties in Colorado.
Data from collocated monitors in Charlotte, NC.
Map of 376 CO monitor locations in the U.S. in 2007.
Map of CO monitor locations with respect to population density in the Denver, CO CBSA,
total population.
Map of CO monitor locations with respect to population density in the Denver, CO CBSA,
age 65 and older.
Map of CO monitor locations with respect to population density in the Los Angeles, CA
CSA, total population.
Map of CO monitor locations with respect to population density in the Los Angeles, CA
CSA, age 65 and older.
County-level map of second-highest 1-h avq CO concentrations in the U.S. in 2007.
County-level map of second-highest 8-h avq CO concentrations in the U.S. in 2007.
Seasonal plots showing the variability in correlations between 24-h avg CO concentration
with 1-h daily max and 8-h daily max CO concentrations and between 1-h daily max and
8-h daily max CO concentrations.
Map of CO monitor locations and maior highways for Denver, CO.
Box plots illustrating the distribution of 2005-2007 hourly CO concentrations in Denver, CO.
Map of CO monitor locations and maior highways for Los Angeles, CA.
Box plots illustrating the distribution of 2005-2007 hourly CO concentrations in Los
Angeles, CA.
1-7
2-21
3-3
3-4
3-6
3-6
3-8
3-8
3-9
3-21
3-25
3-27
3-28
3-29
3-30
3-33
3-34
3-42
3-44
3-45
3-47
3-48
Figure 3-21 Aerial view of the location of CO monitors A and B (marked by the red pins) in Denver, CO,
depicting their proximity to the urban core. 3-54
September 2009 xix DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-22
Figure 3-23
Figure 3-24
Figure 3-25
Figure 3-26
Figure 3-27
Figure 3-28
Figure 3-29
Figure 3-30
Figure 3-31
Figure 3-32
Figure 3-33
Figure 3-34
Figure 3-35
Figure 3-36
Figure 3-37
Figure 3-38
Figure 3-39
Figure 3-40
Figure 3-41
Figure 3-42
Figure 3-43
Figure 3-44
Aerial view of the location of CO monitor I (marked by the red pin) in Azusa, CA (Los
Anqeles CSA), depicting its proximity to mixed use land.
Aerial view of the location of CO monitor Q (marked by the red pin) in Pasadena, CA (Los
Anqeles CSA), depicting its proximity to a residential neighborhood.
Distribution of hourly CO concentration data by city and monitoring scale.
Distribution of 1-h daily max CO concentration data by city and monitoring scale.
Relative concentrations of CO and copollutants at various distances from the 1-71 0 freeway
in Los Angeles.
CO concentration time series 20 m and 300 m from the I-440 highway in Raleigh, NC.
CO concentration profile 10 m from I-440 in Raleigh, NC behind a noise barrier and in open
terrain.
Dimensionless tracer gas concentration on the windward and leeward sides of the canyon
plotted against the elevation of the measurement
Normalized difference between CO measurements taken at ground level and from the 39th
floor of a building in a Phoenix, AZ street canyon as a function of bulk Richardson number
(Ri).
(Top) Trends in ambient CO in the U.S., 1 980-2006, reported as the annual second daily
highest 8-h concentrations (ppm) for the mean, median, 10% and 90% values.
Trends in ambient CO in the U.S., 1 980-2005, reported as the annual second highest daily
8-h concentrations (ppm) for the EPA Regions 1 through 10, along with a depiction of the
geographic extent of those Regions
Diel plot generated from weekday hourly CO data (ppm) for the eleven CSAs and CBSAs
2005-2007.
Diel plot generated from weekend hourly CO data (ppm) for the eleven CSAs and CBSAs
2005-2007.
Seasonal plots showing the variability in correlations between hourly CO concentration and
co-located hourly SO?, NO?, d, PMmand PM?<; concentrations.
Seasonal plots showing the variability in correlations between hourly CO concentration and
co-located hourly SO?, NO?, d, PMm and PM?^ concentrations for Denver, CO.
Seasonal plots showing the variability in correlations between hourly CO concentration and
co-located hourly SO?, NO?, d, PMm and PM?^ concentrations for Los Angeles, CA.
Linear regression of n-butane and isopentane concentration as a function of CO
concentration, Riverside, CA.
Map of the baseline monitor sites used in this assessment to compute policy-relevant
background concentrations.
Monthly (circles) and annual (squares) average CO concentrations (ppb), 2005-2007.
Distribution of time sample population spends in various environments, from the National
Human Activity Pattern Survey.
Hourly personal versus ambient CO concentrations obtained in Baltimore, MD
Box plots of the ratio of personal to ambient concentrations obtained in Baltimore, MD
Comparison of in-vehicle (solid line) and outside the vehicle (dotted line) results for (left)
driving with windows closed and air conditioner in recirculating air mode, and (right) driving
with windows closed and air conditioner in fresh air mode.
3-56
3-57
3-61
3-62
3-64
3-64
3-66
3-68
3-69
3-71
3-72
3-74
3-75
3-76
3-78
3-79
3-81
3-83
3-84
3-91
3-97
3-98
3-104
September 2009 xx DRAFT-DO NOT QUOTE OR CITE
-------
Figure 4-1. Plot of fractional sensitivities of selected variables versus time of exposure. 4-4
Figure 4-2 Simulated COHb formation for two 5 day workweeks 4-6
Figure 4-3 Overall structure of the Bruce and Bruce (2008,193977) multicompartment model of
storage and transport of CO. 4-8
Figure 4-4 Predicted COHb levels in healthy commuters exposed to various CO concentrations over a
60-min commute twice a day. 4-11
Figure 4-5 Predicted COHb levels due to various endogenous CO production rates. 4-12
Figure 4-6 Predicted COHb levels in an active or sedentary individual. CO concentration was constant
at 1 ppm. 4-13
Figure 4-7 Diagrammatic presentation of CO uptake and elimination pathways and CO body stores. 4-15
Figure 4-8 02Hb dissociation curve of normal human blood, of blood containing 50% COHb, and of
blood with only 50% Hb because of anemia. 4-18
Figure 4-9 Changes in blood COHb after short-term and long-term exposure to CO, representing the
biphasic nature of CO elimination. Note: y-axis is log-scale. 4-22
Figure 4-10 Predicted maternal and fetal COHb during prolonged exposure to CO (30-300 ppm) and
washout from equilibrium values with no CO. 4-26
Figure 4-11 Representative estimates of endogenous CO production rates resulting from various
conditions and diseases. 4-29
Figure 4-12 Representative COHb saturation resulting from various diseases and conditions. SBP:
Spontaneous bacterial peritonitis 4-30
Figure 4-13. Representative exhaled CO concentrations (ppm) resulting from various conditions plotted
as fold increases over healthy human controls from each study. SBP: Spontaneous
bacterial peritonitis; URTI: Upper respiratory tract infection 4-31
Figure 5-1 Direct Effects of CO. The dashed line refers to uptake of inhaled CO by respiratory
epithelial cells and resident macrophages in the lung. 5-18
Figure 5-2 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for various froms of CHD. 5-37
Figure 5-3 Summary of effect estimates (95% confidence intervals) associated with ED visits and
hospital admissions for stroke. 5-42
Figure 5-4 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CHF. 5-46
Figure 5-5 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CVD. 5-52
Figure 5-6 Effect estimates from studies of ED visits and hospital admissions for CVD outcomes other
than stroke from single pollutant (CO only, closed circles) and particulate copollutant (CO
plus PM, open circles) models. 5-54
Figure 5-7 Effect estimates from studies of ED visits and HAs for CVD outcomes other than stroke
from single pollutant. 5-55
Figure 5-8 Summary of effect estimates (95% confidence intervals) for PTB associated with maternal
exposure to ambient CO. 5-73
Figure 5-9 Summary of change in birth weight (95% confidence intervals) associated with maternal
exposure to ambient CO. 5-80
September 2009 xxi DRAFT-DO NOT QUOTE OR CITE
-------
Figure 5-10
Figure 5-1 1
Figure 5-1 2
Figure 5-1 3
Figure 5-1 4
Figure 5-1 5
Figure 5-1 6
Figure 5-1 7
Figure 5-1 8
Figure A-1
Figure A-2
Figure A-3
Figure A-4
Figure A-5
Figure A-6
Figure A-7
Figure A-8
Figure A-9
Figure A-1 0
Figure A-1 1
Figure A-1 2
Figure A-1 3
Figure A-1 4
Summary of effect estimates (95% confidence intervals) for LBW associated with maternal
exposure to ambient CO.
Summary of effect estimates (95% confidence intervals) for SGA associated with maternal
exposure to ambient CO.
Estimated effect (95% Cl) on pulmonary function due to a 10th to 90th percentile increment
chanqe in pollutant concentration in sinqle-pollutant models.
Asthma symptoms, respiratory symptoms and medication use in asthmatic individuals
associated with short-term exposure to CO.
Summary of associations between short-term exposure to CO and respiratory hospital
admissions:
Summary of associations between short-term exposure to CO and respiratory ED visits.
Posterior means and 95% posterior intervals of national average estimates for CO effects
on total (non-accidental) mortality at lags 0, 1 , and 2 within sets of the 90 U.S. cities with
available pollutant data.
Summary of mortality risk estimates for short-term exposure to CO from multicity studies.
Summary of mortality risk estimates for lonq-term exposure to CO.
CO emissions density map and distribution for the state of Alaska and for Yukon-Koyukuk
County in Alaska.
CO emissions density map and distribution for the state of Utah and for selected counties in
Utah.
CO emissions density map and distribution for the state of Massachusetts and for selected
counties in Massachusetts.
CO emissions density map and distribution for the state of Georgia and for selected
counties in Georgia (1 of 2).
CO emissions distribution for selected counties in Georgia (2 of 2).
CO emissions density map and distribution for the state of California and for selected
counties in California.
CO emissions density map and distribution for the state of Alabama and for Jefferson
County in Alabama.
Map of CO monitor locations with respect to population density in the Anchorage CBSA,
total population.
Map of CO monitor locations with respect to population density in the Anchorage CBSA,
aqes 65 and older.
Map of CO monitor locations with respect to population density in the Atlanta CSA, total
population.
Map of CO monitor locations with respect to population density in the Atlanta CSA, ages 65
and older.
Map of CO monitor locations with respect to population density in the Boston CSA, total
population.
Map of CO monitor locations with respect to population density in the Boston CSA, ages 65
and older.
Map of CO monitor locations with respect to population density in the Houston CSA, total
population.
5-81
5-82
5-119
5-124
5-132
5-135
5-147
5-152
5-164
A-1
A-2
A-3
A-4
A-5
A-6
A-7
A-20
A-20
A-21
A-21
A-22
A-22
A-23
September 2009 xxii DRAFT-DO NOT QUOTE OR CITE
-------
Figure A-1 5
Figure A-1 6
Figure A-1 7
Figure A-1 8
Figure A-1 9
Figure A-20
Figure A-21
Figure A-22
Figure A-23
Figure A-24
Figure A-25
Figure A-26
Figure A-27
Figure A-28
Figure A-29
Figure A-30
Figure A-31
Figure A-32
Figure A-33
Figure A-34
Figure A-35
Figure A-36
Figure A-37
Figure A-38
Map of CO monitor locations with respect to population density in the Houston CSA, ages
65 and older.
Map of CO monitor locations with respect to population density in the New York City CSA,
total population.
Map of CO monitor locations with respect to population density in the New York City CSA,
aqes 65 and older.
Map of CO monitor locations with respect to population density in the Phoenix CSA, total
population.
Map of CO monitor locations with respect to population density in the Phoenix CSA, ages
65 and older.
Map of CO monitor locations with respect to population density in the Pittsburgh CSA, total
population.
Map of CO monitor locations with respect to population density in the Pittsburgh CSA, ages
65 and older.
Map of CO monitor locations with respect to population density in the Seattle CSA, total
population.
Map of CO monitor locations with respect to population density in the Seattle CSA, ages 65
and older.
Map of CO monitor locations with respect to population density in the St. Louis CSA, total
population.
Map of CO monitor locations with respect to population density in the St. Louis CSA, ages
65 and older.
Map of CO monitor locations with AQS Site IDs for Anchoraqe, AK.
Box plots illustrating the seasonal distribution of hourly CO concentrations in Anchorage,
AK. Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for Atlanta, GA.
Box plots illustrating the seasonal distribution of hourly CO concentrations in Atlanta, GA.
Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for Boston, MA.
Box plots illustrating the seasonal distribution of hourly CO concentrations in Boston, MA.
Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for Houston, TX.
Box plots illustrating the seasonal distribution of hourly CO concentrations in Houston, TX.
Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for New York City, NY.
Box plots illustrating the seasonal distribution of hourly CO concentrations in New York
City, NY. Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for Phoenix, AZ.
Box plots illustrating the seasonal distribution of hourly CO concentrations in Phoenix, AZ.
Note: 1 = winter, 2, = sprinq, 3 = summer, and 4 = fall on the x-axis.
Map of CO monitor locations with AQS Site IDs for Pittsburqh, PA.
A-24
A-24
A-25
A-25
A-26
A-27
A-27
A-28
A-28
A-29
A-29
A-30
A-32
A-33
A-35
A-36
A-38
A-39
A-41
A-42
A-44
A-45
A-47
A-48
September 2009 xxiii DRAFT-DO NOT QUOTE OR CITE
-------
Figure A-39 Box plots illustrating the seasonal distribution of hourly CO concentrations in Pittsburgh,
PA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis. A-50
Figure A-40 Map of CO monitor locations with AQS Site IDs for Seattle, WA. A-51
Figure A-41 Box plots illustrating the seasonal distribution of hourly CO concentrations in Seattle, WA.
Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis. A-52
Figure A-42 Map of CO monitor locations with AQS Site IDs for St. Louis, MO. A-53
Figure A-43 Box plots illustrating the seasonal distribution of hourly CO concentrations in St. Louis, MO.
Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis. A-55
Figure A-44 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5 concentrations for Anchorage, AK. A-63
Figure A-45 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5 concentrations for Atlanta, GA. A-64
Figure A-46 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5concentrations for Boston, MA. A-65
Figure A-47 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5 concentrations for New York City, NY. A-66
Figure A-48 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5 concentrations for Phoenix, AZ. A-67
Figure A-49 Seasonal plots of correlations between hourly CO concentration with hourly (1) S02, (2)
N02, (3) 03, (4) PM10, and (5) PM2.5 concentrations for Seattle, WA. A-68
September 2009 xxiv DRAFT-DO NOT QUOTE OR CITE
-------
Acronyms and Abbreviations
a alpha, ambient exposure factor
a air exchange rate of the microenvironment
AA abdominal aorta(s)
ABR auditory brainstem response
ACS American Cancer Society
ACS-CPS-II ACS Cancer Prevention Study II
ADP adenosine diphosphate
AEFV area under the expiratory flow-volume curve
AGL above ground level
Akt Akt cell signaling pathway
AMI acute myocardial infarction
AMP adenosine monophosphate
ANOVA analysis of variance
APO E apolipoprotein E
ARI acute respiratory infection
AP action potential
APD action potential duration
APEX Air Pollution Exposure
APHEA Air Pollution and Health: A European Approach
APTT activated partial thromboplastin time
AQ air quality
AQCD Air Quality Criteria Document
AQS Air Quality System
AR gastronomy reared
ARCO gastronomy reared + CO exposure
ARIC Atherosclerosis Risk in Communities
ARID gastronomy reared with iron deficient diet
ARIDCO gastronomy reared with iron deficient diet + CO exposure
ATP adenosine triphosphate
ATS American Thoracic Society
AVP aortic valve prosthesis
P beta, beta coefficient, slope
September 2009
xxv
DRAFT - DO NOT CITE OR QUOTE
-------
B lymphocytes bursa-dependent lymphocytes
BALF bronchoalveolar lavage fluid
BC black carbon
BEAS-2B human bronchial epithelial cell line
BEIS Biogenic Emissions Inventory System
BELD Biogenic Emissions Landcover Database
BHR bronchial hyper-responsiveness
BKCa voltage and Ca2+-activated K+ channel(s)
BP blood pressure
BQ-123 endothelin A (ETA) receptor antagonist
BS black smoke
BSP black smoke particles
Ca ambient concentration
CA cardiac arrhythmia
Ca2+ calcium ion
CAA Clean Air Act
CAD coronary artery disease
CALINE California Line Source Dispersion Model
CAMP Childhood Asthma Management Program
cAMP cyclic AMP
CAP(s) concentrated ambient particles, compound action potential(s)
CASAC Clean Air Scientific Advisory Committee
CASN Cooperative Air Sampling Network
CAth cardiac atherosclerosis
CBS A Core-Based Statistical Area
CCGG Carbon Cycle Greenhouse Gases Group
CD cardiac dysrhthmias
CD-I mouse strain
CDC Centers for Disease Control and Prevention
CdCl2 cadmium chloride
CFK Coburn-Forster-Kane
CFR Code of Federal Regulations
cGMP cyclic GMP
CH2O formaldehyde
CH2O2 formic acid
September 2009
XXVI
DRAFT - DO NOT CITE OR QUOTE
-------
CH3 methyl groups
CH3CHO acetaldehyde
CH3CO acetyl radical(s)
CH3CO3NO2 PAN, peroxyacetyl nitrate
CH3O2 methyl peroxy radical
CH3OOH methyl hydroperoxide
CH4 methane
ChAT choline acetyl-transferase
CHD coronary heart disease
CHF congestive heart failure
CI confidence interval(s)
CIS cerebral ischemic stroke
Cj airborne concentration at location j
CL/P cleft lip with or without palate
CNS central nervous system
CO carbon monoxide
CO2 carbon dioxide
COD coefficient of divergence
CoH, COH coefficient of haze
COHb carboxyhemoglobin (% concentration measured in (mL CO/mL
blood))
COMb carboxymyoglobin
CONUS contiguous U.S.
COPD chronic obstructive pulmonary disease
CPS II Cancer Prevention Study II
C-R concentration-response
CRC Coordinating Research Council
CrMP collapsin response mediator protein
CRP C-reactive protein
CSA Combined Statistical Area
CVD cardiovascular disease
d straight-line distance between monitor pairs
df degrees of freedom
DL lung diffusing capacity
DLCO lung diffusing capacity of CO
DmCO capacity for diffusion of CO into the muscle
September 2009
XXVII
DRAFT - DO NOT CITE OR QUOTE
-------
DMT-1
DMV
DNA
DOCA
dP/dtLV
dP/dtRV
DSA
E
Ea
EC
ED
EKG, EGG
Ena
eNOS
EPA
EPO
EPR
EPRI
ESRL
ET-1
ETA
ETS
EXPOLIS
FAS
FC
FEF
FEF25.75
FEM
FEVj
f;
FjCO
Finf
fo
divalent metal transporter-1
dorsal motor nucleus of the vagus nerve
deoxyribonucleic acid
Deoxycorticosterone acetate
left ventricular maximal and minimal first derived pressure
(+dP/dtLV, -dP/dtLV)
right ventricular maximal and minimal first derived pressure
(+dP/dtRV, -dP/dtRv)
deletion/sub stitution/addition
exposure over some duration
exposure to pollutant of ambient origin
elemental carbon
emergency department
electrocardiogram
exposure to pollutant of non-ambient origin
endothelial nitric oxide synthase
U.S. Environmental Protection Agency
erythropoietin
Electron Paramagnetic Resonance
Electric Power Research Institute
Earth System Research Laboratory
endothelin-1
endothelin A (ETA) receptor
environmental tobacco smoke
six-city European air pollution study
apoptosis stimulating fragment
interference filter
forced expiratory flow (L/s)
forced expiratory flow between the times at which 25% and 75% of
the vital capacity is reached
Federal equivalent method
forced expiratory volume in 1 second
fraction of time spent indoors
fractional concentration of CO in ambient air
infiltration factor
fraction of time spent outdoors
September 2009
XXVIII
DRAFT - DO NOT CITE OR QUOTE
-------
FR
FOR
FRM
FSH
FVC
FVII
FW
GAM
GD
GEE
GEM
GFAP
GFC
GLM
GLMM
GMD
GMP
GSH
GSSG
GTP
GWP(s)
H
h
H202
H9c2
Hb
HC(s)
HCFC(s)
HCO
HEAPSS
HEK293
Hep3B
HF
HFLFR
HH
Federal Register
fetal growth restriction(s)
Federal reference method
follicle stimulating hormone
forced vital capacity
Factor VII
fresh weight
generalized additive model(s)
gestational day
generalized estimating equations
gas extraction monitor
glial fibrillary acidic protein
gas filter correlation
generalized linear models
generalized linear mixed models
Global Monitoring Division
guanosine monophosphate
glutathione
oxidized glutathione
guanosine triphosphate
global warming potential(s)
atomic hydrogen, hydrogen radical, height
hour
hydrogen peroxide
rat embryonic cardiomyocytes
hemoglobin
hydrocarbon(s)
hydrochlorofluorocarbon(s)
formyl radical
Health Effects of Air Pollution among Susceptible Subpopulations
human embryonic kidney cells
Human hepatocarcinoma cell line
heart failure, high frequency (HRV parameter)
high frequency to low frequency ratio (HRV parameter)
hypobaric hypoxia
September 2009
XXIX
DRAFT - DO NOT CITE OR QUOTE
-------
HIF-la
HO
H02
HO-1
HO-2
HO/CO
HR
H/R
HRV
HUVEC(s)
hv
IARC
1C
ICAM-1
ICD
ICR
IDW
IHD
IL-x
INDAIR
IOM
IQR
IR
IS
ISA
ITA
Ito
IUGR
K+
k
kco
Km
k02
LEW
LCA+
hypoxia-inducible factor
heme oxygenase
hydroperoxy radical
inducible isoform of heme oxygenase
constitutively expressed isoform of heme-oxygenase
heme oxygenase/carbon monoxide system
heart rate, hazard ratio
hypoxia followed by reoxygenation
heart rate variability
human umbilical vein endothelial cell(s)
photon
International Agency for Research on Cancer
inferior colliculus
intercellular adhesion molecule
implantable cardioverter defibrillator(s)
Institute for Cancer Research
inverse-distance-weighted
ischemic heart disease
interleukin-6, 8, etc.
Indoor Air Model
Institute of Medicine
interquartile range
immunoreactivity
ischemic stroke
Integrated Science Assessment
internal thoracic artery of the heart
transient outward current
intrauterine growth restriction
potassium ion
dissociation rate
dissociation rate of carbon monoxide from hemoglobin
Michaelis Constant in Michaelis-Menten equation of enzyme kinetics
Dissociation rate of oxygen from hemoglobin
low birth weight
leucocyte common antigen cells
September 2009
xxx
DRAFT - DO NOT CITE OR QUOTE
-------
LD
LDH
LDL
LF
LH
LOAEL
LOD
LOESS
LPS
LTP
LUR
LV
LV+S
LVDP
LVESP
LVSF
LVW
M
MAPK
MAO-A
Mb
MC
METs
MHC
MI
min
MIP-2
mitral E to A ratio
MMEF
MMP
MOA(s)
MOBILE6
MODIS
MONICA
MOPITT
lactational day
lactate dehydrogenase
low-density lipoprotein
low frequency (HRV parameter)
luetenizing hormone
lowest observed adverse effect level
limit of detection
locally weighted scatterplot smoothing
lipopolysaccharide
long-term potentiation
land use regression
left ventricle
left ventricular plus septum
left ventricular developed pressure
left ventricular end diastolic pressure
left ventricular shortening fraction
left ventricular work
Haldane coefficient representing CO chemical affinity
mitogen-activated protein kinase
monoamine oxidase A
myoglobin
ultrafine particle mass concentration
metabolic equivalent unit(s)
major histocompatibility complex
myocardial infarction, "heart attack"
minute(s)
macrophage inflammatory protein-2
mitral ratio of peak early to late diastolic filling velocity
maximal midexpiratory flow
matrix metalloproteinase
mode(s) of Action
Mobile source emission factor model
Moderate Resolution Imaging Spectroradiometer
Monitoring of Trends and Determinants in Cardiovascular Disease
Measurement of Pollution in the Troposphere
September 2009
XXXI
DRAFT - DO NOT CITE OR QUOTE
-------
MPO
MPT
MR
mRNA
MSA
MSNA
MT
MV02
NAAQS
NADPH
NADH-TR
NAPAP
NARSTO
NAS
NASA
Nb
NC
NDIR
NE
NEI
NF-KB
NIHL
NMDA
NMHC(s)
NMMAPS
NN
nNOS
NO
NO'
N02
NOAA
NOAEL
NO'-Hb
NO'-Mb
NOX
myeloperoxidase
mitochondrial permeability transition
maternally reared
messenger RNA
Metropolitan Statistical Area
muscle sympathetic nerve activity
million tons
myocardial oxygen consumption
National Ambient Air Quality Standards
nicotinamide adenine dinucleotide phosphate
nicotinamide adenine dinucleotide - tetrazolium reductase
National Acid Precipitation Assessment Program
North American Research Strategy for Tropospheric Ozone
National Academy of Sciences
National Aeronautics and Space Administration
neuroglobin
ultrafine particle number concentration
nondispersive infrared
norepinephrine
National Emissions Inventory
nuclear factor kappa B
noise-induced hearing loss
N-methyl-D-aspartate
nonmethane hydrocarbon(s)
National Morbidity, Mortality, and Air Pollution Study
normal-to-normal (NN or RR) time interval between each QRS
complex in the EKG
neuronal nitric oxide synthase (NOS)
nitric oxide
nitric oxide free radical
nitrogen dioxide
National Oceanic and Atmospheric Administration
no observed adverse effect level
nitrosyl bound Hb
nitrosyl bound Mb
nitrogen oxides, oxides of nitrogen
September 2009
XXXII
DRAFT - DO NOT CITE OR QUOTE
-------
NRC
NTS
03
02Hb
O2Mb
OAE
OAQPS
OC
OH, OH'
OR
OS
OSPM
P
P,P
P90
PA
PA
PACF
PACO
PAF
PAH
PAHT
PAN
PA02
Pa02
PARP
PB
PEN
PC
pCO
p-co2
PDGF
PEE
PEF
PEFD(s)
National Research Council
nucleus of the solitary tract (in brainstem)
ozone
oxyhemoglobin
oxymyoglobin
otoacoustic emissions
Office of Air Quality Planning and Standards
organic carbon
hydroxyl group, hydroxyl radical
odds ratio
occlusive stroke
Operational Street Pollution Model
penetration factor
probability
90th percentile of the absolute difference in concentrations
alveolar pressure
pulmonary artery (myocytes)
partial auto-correlation functions
alveolar pressure for carbon monoxide
platelet activating factor
polycyclic aromatic hydrocarbon
pulmonary artery hypertension
peroxyacetyl nitrate
alveolar pressure for oxygen
arterial oxygen pressure
poly(ADP-ribose) polymerase
barometric pressure (in mmHg)
N-tert-butyl-alpha-phenylnitrone
average partial pressure in lung capillaries
partial pressure of CO
average partial pressure of O2 in lung capillaries
platelet derived growth factor
prediction equation estimates
peak expiratory flow
Personal Exposures Frequency Distributions
September 2009
XXXIII
DRAFT - DO NOT CITE OR QUOTE
-------
PEM(s)
Pmo
PHD
PI
Pi
PI3K
P,CO
PIH
PKB
PM
PM2.5
PM10
PM10.2.5
PMN
PNC
PND
pNEM/CO
PNN
PNN50
PNS
pO2
pPRB
PT
PTB
PVCD
Pv02
PVO2
g
QCP
personal exposure monitor(s)
saturation pressure of water vapor
pulmonary heart disease
partial pressure of inhaled air
inorganic phosphate
phosphoinositide 3-kinase
CO partial pressure in inhaled air
primary intracerebral hemorrhage
protein kinases B
particulate matter
particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 um (referred to as fine PM)
particulate matter with a nominal mean aerodynamic diameter less
than or equal to 10 um
particulate matter with a nominal mean aerodynamic diameter greater
than 2.5 um and less than or equal to 10 um (referred to as thoracic
coarse particulate matter or the course fraction of PM10).
Concentration may be measured or calculated as the difference
between measured PMi0 and measured PM2 5 concentrations.
polymorphonuclear leukocytes
particle number concentration / count
post natal day
probabilistic NAAQS Exposure Model for CO
proportion of interval differences of successive normal-beat intervals
inEKG
proportion of interval differences of successive normal-beat intervals
greater than 50 ms in EKG
peripheral nervous system
partial pressure of oxygen in lung capillaries
policy-relevant background
prothrombin time
preterm birth
peripheral vascular and cerebrovascular disease
venous oxygen tension
peak oxygen consumption
cardiac output
Quantitative Circulatory Physiology
blood flow to muscle
blood flow to other tissues
September 2009
xxxiv
DRAFT - DO NOT CITE OR QUOTE
-------
RA radial artery of the heart
RAW 264.7 mouse macrophage cell line
RBC red blood cell
rho(O) rho(O) cells (cells lacking mitochondrial DNA)
Ri Richardson number
rMSSD mean squared differences of successive difference normal-beat to
normal-beat (NN or RR) time intervals between each QRS complex in
the EKG
RNA ribonucleic acid
ROE Report on the Environment
ROFA residual oil fly ash (particles)
ROS reactive oxygen species
RR normal-to-normal (NN or RR) time interval between each QRS
complex in the EKG
RR risk ratio(s)
RUPERT Reducing Urban Pollution Exposure from Road Transport
RV right ventricle (of heart)
RVEDP right ventricular end diastolic pressure
RVESP right ventricular end-systolic pressure
RVSF right ventricular shortening fraction
RVW right ventricular work
SA sphinganine
SAA serum amyloid A
SAB Science Advisory Board
SBP systolic blood pressure, spontaneous bacterial peritonitis
SDNN standard deviation normal-to-normal (NN or RR) time interval
between each QRS complex in the EKG
sEng soluble endoglin
SES socioeconomic status
SF6 sulfur hexafluoride (tracer gas)
sFlt soluble Fms-like tyrosine kinase-1
SGA small for gestational age
sGC soluble guanylate cyclase
SHEDS Stochastic Human Exposure and Dose Simulation
SHR Spontaneously hypertensive rat strain
SIDS sudden infant death syndrome
SIPs State Implementation Plan(s)
September 2009
xxxv
DRAFT - DO NOT CITE OR QUOTE
-------
siRNA
SLAMS
SMC
SnMP
SNP
SnPP-IX
SO
S02
SO42'
SOD
SOPHIA
STEMS
STN
STPD
SV
SVEB
T
T lymphocytes
TEARS
TC
TFAM
Tg
TH
THP-1
TIA
TNF-a
TPM
TSP
UFP
ULTRA
URI
URTI
use
V A
Vb
small inhibitory RNA
State and Local Air Monitoring Stations
smooth muscle cell(s)
tin-(IV)-mesoporphyrin
single-nucleotide polymorphism
tin protoporphyrin IX
sphingosine
sulfur dioxide
sulfate
superoxide dismutase
Study of Particles and Health in Atlanta
Space-Time Exposure Modeling System
Speciation Trends Network
standard temperature and pressure, dry
stroke volume
supraventricular ectopic beats
tau, photochemical lifetime
thymus-dependent lymphocytes
thiobarbituric acid reactive substances
total carbon
mitochondrial transcription factor A
teragram(s)
tyrosine hydroxylase
human monocyte-derived cell line
transient ischemic attack
tissue necrosis factor alpha
total particulate matter
total suspended particles
ultrafine particle(s)
Ultrafine Particles in Ambient Air
upper respiratory infection
upper respiratory tract infection
U.S. Code
alveolar ventilation
blood volume
September 2009
xxxvi
DRAFT - DO NOT CITE OR QUOTE
-------
Vco
VD
VE
VEGF
VLF
V
v max
VO2 max
VOC(s)
VPB
vWF
W
WBC
WHI
WKY
ZnPPIX
CO uptake rate
endogenous CO production rate
Dead space volume
ventilation rate
vascular endothelial growth factor
very low energy frequency (HRV parameter)
maximum velocity
maximum volume per time, of oxygen
volatile organic compound(s)
ventricular premature beat
von Willebrand factor
width
white blood cell
Women's Health Initiative
Wistar-Kyoto rat strain
Zn protoporphyrin IX
September 2009
XXXVII
DRAFT - DO NOT CITE OR QUOTE
-------
Chapter 1. Introduction
1 The Integrated Science Assessment (ISA) is a concise evaluation and synthesis of the most
2 policy-relevant science for reviewing the national ambient air quality standards (NAAQS). Because
3 the ISA communicates critical science judgments relevant to the NAAQS review, it forms the
4 scientific foundation for the review of the NAAQS for carbon monoxide (CO). The existing primary
5 CO standards include a 1-hour (h) average (avg) standard set at 35 parts per million (ppm), and an
6 8-h avg standard set at 9 ppm, neither to be exceeded more than once per year. There is currently no
7 secondary standard for CO.
8 The ISA accurately reflects "the latest scientific knowledge useful in indicating the kind and
9 extent of identifiable effects on public health which may be expected from the presence of [a]
10 pollutant in ambient air" (42 U.S.C. 7408). Key information and judgments formerly contained in
11 the Air Quality Criteria Document (AQCD) for CO are incorporated in this assessment. Additional
12 details of the pertinent scientific literature published since the last review, as well as selected older
13 studies of particular interest, are included in a series of annexes. This second external draft IS A thus
14 serves to update and revise the evaluation of the scientific evidence available at the time of the
15 previous review of the NAAQS for CO that was completed in 2000.
16 The integrated Plan for Review of the NAAQS for CO (U.S. EPA, 2008, 193995) identifies
17 key policy-relevant questions that provide a framework for this assessment of the scientific evidence.
18 These questions frame the entire review of the NAAQS for CO and thus are informed by both
19 science and policy considerations. The ISA organizes, presents, and integrates the scientific evidence
20 which is considered along with findings from risk analyses and policy considerations to help the U.S.
21 Environmental Protection Agency (EPA) address these questions during the NAAQS review. In
22 evaluating the health evidence, the focus of this assessment is on scientific evidence that is most
23 relevant to the following questions taken directly from the Integrated Review Plan:
24 • Has new information altered the scientific support for the occurrence of health effects
25 following short- and/or long-term exposure to levels of CO found in the ambient air?
26 • To what extent is key evidence becoming available that could inform our understanding
27 of human subpopulations that are particularly sensitive to CO exposures? Is there new or
28 emerging evidence on health effects beyond cardiovascular and respiratory endpoints
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 1-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 (e.g., systemic effects, developmental effects, birth outcomes) that suggest additional
2 sensitive subpopulations should be given increased focus in this review (e.g., neonates)?
3 • What do recent studies focused on the near-roadway environment, including bus stops
4 and intersections, tell us about high-exposure human subpopulations and the health
5 effects of CO? What information is available on elevated exposures due to other
6 transportation sources, such as shipping, port operations, and recreational vehicles? What
7 is the effect of altitude on CO sources and health effects?
8 "At what levels of CO exposure do health effects of concern occur?
9 "To what extent is key scientific evidence becoming available to improve our
10 understanding of the health effects associated with various time periods of CO exposures,
11 including not only daily, but also chronic (months to years) exposures? To what extent is
12 critical research becoming available that could improve our understanding of the
13 relationship between various health endpoints and different lag periods (e.g., single day,
14 multiday distributed lags)?
15 "To what extent does the evidence suggest that alternate dose indicators other than
16 carboxyhemoglobin (COHb) levels (e.g., tissue oxygenation) should be evaluated to
17 characterize the biological effect?
18 • Has new information altered conclusions from previous reviews regarding the
19 plausibility of adverse health effects caused by CO exposure?
20 • To what extent have important uncertainties identified in the last review been reduced
21 and/or have new uncertainties emerged?
22 • Have new information or scientific insights altered the scientific conclusions regarding
23 the occurrence of direct (or indirect) welfare effects associated with levels of CO found
24 in the ambient air?
1.1. Legislative Requirements
25 Two sections of the Clean Air Act (CAA, the Act) govern the establishment and revision of the
26 NAAQS. Section 108 of the Act (42 U.S.C. 7408) directs the Administrator to identify and list "air
27 pollutants" that "in [her] judgment, may reasonably be anticipated to endanger public health and
28 welfare" and whose "presence ... in the ambient air results from numerous or diverse mobile or
September 2009 1-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 stationary sources" and to issue air quality criteria for those that are listed (42 U.S.C. 7408). Air
2 quality criteria are intended to "accurately reflect the latest scientific knowledge useful in indicating
3 the kind and extent of identifiable effects on public health or welfare which may be expected from
4 the presence of [a] pollutant in ambient air..." 42 U.S.C. 7408(b).
5 Section 109 of the Act (42 U.S.C. 7409) directs the EPA Administrator to propose and
6 promulgate "primary" and "secondary" National Ambient Air Quality Standards (NAAQS) for
7 pollutants listed under Section 108. Section 109(b)(l) defines a primary standard as one "the
8 attainment and maintenance of which in the judgment of the Administrator, based on such criteria
9 and allowing an adequate margin of safety, are requisite to protect the public health."1 A secondary
10 standard, as defined in Section 109(b)(2), must "specify a level of air quality the attainment and
11 maintenance of which, in the judgment of the U.S. EPA Administrator, based on such criteria, is
12 required to protect the public welfare from any known or anticipated adverse effects associated with
13 the presence of [the] pollutant in the ambient air."2 The requirement that primary standards include
14 an adequate margin of safety was intended to address uncertainties associated with inconclusive
15 scientific and technical information available at the time of standard setting. It was also intended to
16 provide a reasonable degree of protection against hazards that research has not yet identified. See
17 Lead Industries Association v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied, 449 U.S.
18 1042 (1980); American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981) cert.
19 denied, 455 U.S. 1034 (1982). The aforementioned uncertainties are components of the risk
20 associated with pollution at levels below those at which human health effects can be said to occur
21 with reasonable scientific certainty. Thus, in selecting primary standards that include an adequate
22 margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
23 demonstrated to be harmful but also to prevent lower pollutant levels that may pose an unacceptable
24 risk of harm, even if the risk is not precisely identified as to nature or degree.
25 In selecting a margin of safety, the EPA considers such factors as the nature and severity of the
26 health effects involved, the size of sensitive population(s) at risk, and the kind and degree of the
27 uncertainties that must be addressed. The selection of any particular approach to providing an
28 adequate margin of safety is a policy choice left specifically to the Administrator's judgment. See
29 Lead Industries Association v. EPA, supra, 647 F.2d at 1161-62.
30 In setting standards that are "requisite" to protect public health and welfare, as provided in
31 Section 109(b), EPA's task is to establish standards that are neither more nor less stringent than
1 The legislative history of section 109 of the Clean Air Act indicates that a primary standard is to be set at "the maximum permissible
ambient air level . . . which will protect the health of any [sensitive] group of the population," and that for this purpose "reference should
be made to a representative sample of persons comprising the sensitive group rather than to a single person in such a group" [S. Rep. No.
91-1196, 91st Cong., 2d Sess. 10 (1970)].
2 Welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects on soils, water, crops, vegetation,
man-made materials, animals, wildlife, weather, visibility and climate, damage to and deterioration of property, and hazards to
transportation, as well as effects on economic values and on personal comfort and well-being."
September 2009 1-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 necessary for these purposes. In so doing, EPA may not consider the costs of implementing the
2 standards. See Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-76 (D.C.
3 Cir. 2001).
4 Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-year intervals
5 thereafter, the Administrator shall complete a thorough review of the criteria published under Section
6 108 and the national ambient air quality standards... and shall make such revisions in such criteria
7 and standards and promulgate such new standards as may be appropriate..." Section 109(d)(2)
8 requires that an independent scientific review committee "shall complete a review of the
9 criteria... and the national primary and secondary ambient air quality standards... and shall
10 recommend to the Administrator any new... standards and revisions of existing criteria and standards
11 as may be appropriate..." Since the early 1980s, this independent review function has been
12 performed by the Clean Air Scientific Advisory Committee (CASAC) of EPA's Science Advisory
13 Board (SAB).
1.2. History of the NAAQS for CO
14 On April 30, 1971, EPA promulgated identical primary and secondary NAAQS for CO, under
15 Section 109 of the Clean Air Act, set at 9 ppm, 8-h avg and 35 ppm, 1-h avg, neither to be exceeded
16 more than once per year (36 FR 8186). In 1979, EPA published the Air Quality Criteria Document
17 for Carbon Monoxide (1979, 017687), which updated the scientific criteria upon which the initial
18 CO standards were based. A Staff Paper (U.S. EPA, 1979, 194665) was prepared and, along with the
19 AQCD, served as the basis for development of proposed rulemaking (45 FR 55066) published on
20 August 18, 1980. Delays due to uncertainties regarding the scientific basis for the final decision
21 resulted in EPA announcing a second public comment period (47 FR 26407). Following substantial
22 reexamination of the scientific data, EPA prepared an Addendum to the 1979 AQCD (1984, 012690)
23 and an updated Staff Paper (1984, 012691). Following review by CASAC, EPA announced its final
24 decision (50 FR 37484) not to revise the existing primary standard and to revoke the secondary
25 standard for CO on September 13, 1985, due to a lack of evidence of direct effects on public welfare
26 at ambient concentrations.
27 In 1987, EPA initiated action to revise the criteria for CO and released a revised AQCD for
28 CASAC and public review. In a "closure letter" (McClellan, 1991, 194666) sent to the
29 Administrator, the CASAC concluded that the AQCD (U.S. EPA, 1991, 017643) ". . . provides a
30 scientifically balanced and defensible summary of current knowledge of the effects of this pollutant
31 and provides an adequate basis for the EPA to make a decision as to the appropriate primary NAAQS
32 for CO." A revised Staff Paper subsequently was reviewed by CASAC and the public, and in a
33 "closure letter" (McClellan, 1992, 194667) sent to the Administrator, CASAC stated ". . . that a
September 2009 1-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 standard of the present form and with a numerical value similar to that of the present standard would
2 be supported by the present scientific data on health effects of exposure to carbon monoxide." Based
3 on the revised AQCD (U.S. EPA, 1991, 017643) and staff conclusions and recommendations
4 contained in the revised Staff Paper (U.S. EPA, 1992, 084191). the Administrator announced the
5 final decision (59 FR 38906) on August 1, 1994, that revision of the primary NAAQS for CO was
6 not appropriate at that time.
7 In 1997, revisions to the 1991 AQCD were initiated. A workshop was held in September 1998
8 to review and discuss material contained in the revised AQCD. On June 9, 1999, CAS AC held a
9 public meeting to review the draft AQCD and a draft exposure analysis methodology document.
10 Comments from CAS AC and the public were considered in a second draft AQCD, which was
11 reviewed at a CAS AC meeting, held on November 18, 1999. After revision of the second draft
12 AQCD, the final AQCD (U.S. EPA, 2000, 000907) was released in August 2000. EPA put the review
13 on hold when Congress called on the National Research Council (NRC) to conduct a review of the
14 impact of meteorology and topography on ambient CO concentrations in high altitude and extreme
15 cold regions of the U.S. In response, the NRC convened the committee on Carbon Monoxide
16 Episodes in Meteorological and Topographical Problem Areas, which focused on Fairbanks, Alaska
17 as a case study in an interim report, which was completed in 2002. A final report, Managing Carbon
18 Monoxide Pollution in Meteorological and Topographical Problem Areas, was published in 2003
19 (NRC, 2003, 042550) and offered a wide range of recommendations on management of CO air
20 pollution, cold start emissions standards, oxygenated fuels, and CO monitoring. EPA did not
21 complete the NAAQS review which started in 1997.
1.3. ISA Development
22 EPA initiated the current review of the NAAQS for CO on September 13, 2007 with a call for
23 information from the public (72 FR 52369). In addition to the call for information, publications were
24 identified through an ongoing literature search process that includes extensive computer database
25 mining on specific topics. Literature searches were conducted routinely to identify studies published
26 since the last review, focusing on publications from 1999 to May 2009. Search strategies were
27 iteratively modified to optimize identification of pertinent publications. Additional papers were
28 identified for inclusion in several ways: review of pre-publication tables of contents for journals in
29 which relevant papers may be published; independent identification of relevant literature by expert
30 authors; and identification by the public and CAS AC during the external review process.
31 Publications considered for inclusion in the ISA were added to the Health and Environmental
32 Research Online (HERO) database recently developed by EPA (http://cfpub.epa.gov/ncea/hero/);
33 note that all references in the ISA include a HERO ID that provides a link to the database. Typically,
September 2009 1-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 only information that had undergone scientific peer review and had been published or accepted for
2 publication was considered, along with analyses conducted by EPA using publicly available data.
3 This review has attempted to evaluate all relevant data published since the last review pertaining to
4 the atmospheric science of CO, human exposure to ambient CO, and epidemiologic, controlled
5 human exposure, and animal toxicological studies on CO, including those related to exposure-
6 response relationships, mode(s) of action (MOA), or susceptible subpopulations. Added to the body
7 of research on CO effects were EPA's analyses of air quality and emissions data, studies on
8 atmospheric chemistry, transport, and fate of these emissions, as well as issues related to exposure to
9 CO. An extensive literature search for data on the ecological effects of ambient CO did not identify
10 any relevant information.
11 In general, in assessing the scientific quality and relevance of health and environmental effects
12 studies, the following considerations have been taken into account when selecting studies for
13 inclusion in the ISA or its annexes. The selection process for studies included in this ISA is shown in
14 Figure 1-1.
15 • Are the study populations, subjects, or animal models adequately selected and are they
16 sufficiently well defined to allow for meaningful comparisons between study or exposure
17 groups?
18 • Are the statistical analyses appropriate, properly performed, and properly interpreted?
19 Are likely covariates adequately controlled or taken into account in the study design and
20 statistical analysis?
21 • Are the air quality data, exposure, or dose metrics of adequate quality and sufficiently
22 representative of information regarding ambient CO?
23 • Are the health or welfare effect measurements meaningful and reliable?
September 2009 1-6 DRAFT - DO NOT CITE OR QUOTE
-------
Informative
studies
are identified
KEY DEFINITIONS
INFORMATIVE studies are well-designed,
properly implemented, thoroughly described.
HIGHLY INFORMATIVE studies reduce
uncertainty on critical issues, may include
analyses of confounding or effect modification
by copollutants or other variables, analyses of
concentration-response or dose-response
relationships, analyses related to time
between exposure and response, and offer
innovation in method or design.
POLICY-RELEVANT studies may include
those conducted at or near ambient concen-
trations and studies conducted in U.S. and
Canadian airsheds.
Studies are
evaluated for inclusion
in the ISA and^
or Annexes.
ISA
Policy relevant and highly informative studies discussed in the ISA text include
those that provide a basis for or describe the association between the criteria
pollutant and effects. Studies summarized in tables and figures are included
because they are sufficiently comparable to be displayed together, A study
highlighted in the ISA text does not necessarily appear in a summary table or
figure.
ANNEXES
All newly identified informative studies are included in the Annexes. Older, key
studies included in previous assessments may be included as well.
Figure 1 -1 Identification of studies for inclusion in the ISA.
1 In selecting epidemiologic studies, EPA considered whether a given study presented
2 information on associations with short- or long-term CO exposures at or near ambient levels of CO;
3 considered approaches to evaluate issues related to potential confounding and modification of effects
4 by other pollutants; addressed health endpoints and populations not previously extensively
5 researched; and evaluated important methodologic issues (e.g., lag or time period between exposure
6 and effects, model specifications, thresholds, mortality displacement) related to interpretation of the
7 health evidence. Among the epidemiologic studies selected, particular emphasis was placed on those
8 studies most relevant to the review of the NAAQS. Specifically, studies conducted in the United
9 States (U.S.) or Canada were discussed in more detail than those from other geographical regions.
September 2009
1-7
DRAFT - DO NOT CITE OR QUOTE
-------
1 Particular emphasis was placed on: (1) recent multicity studies that employ standardized analysis
2 methods for evaluating effects of CO and that provide overall estimates for effects based on
3 combined analyses of information pooled across multiple cities, (2) studies that help understand
4 quantitative relationships between exposure concentrations and effects, (3) new studies that provide
5 evidence on effects in susceptible or vulnerable populations, and (4) studies that consider and report
6 CO as a component of a complex mixture of air pollutants.
7 Criteria for the selection of research evaluating controlled human exposure or animal
8 toxicological studies included a focus on studies conducted using relevant pollutant exposures. For
9 both types of studies, relevant pollutant exposures are considered to be those generally within one or
10 two orders of magnitude of ambient CO concentrations. Studies in which higher doses were used
11 may also be considered if they provide information relevant to understanding MO As or mechanisms,
12 as noted below.
13 Evaluation of controlled human exposure studies focused on those that approximated expected
14 human exposure conditions in terms of concentration and duration. In the selection of controlled
15 human exposure studies, emphasis is placed on studies that (1) investigate potentially susceptible
16 populations such as people with cardiovascular diseases; (2) address issues such as concentration-
17 response or time-course of responses; (3) include control exposures to filtered air; and (4) have
18 sufficient statistical power to assess findings.
19 Review of the animal toxicological evidence focused on studies that approximate expected
20 human dose conditions, which will vary depending on the toxicokinetics and biological sensitivity of
21 the particular laboratory animal species or strains studied. Due to resource constraints on exposure
22 duration and numbers of animals tested, animal studies typically utilize high-concentration
23 exposures to acquire data relating to mechanisms and assure a measureable response. Such studies
24 were considered to the extent that they provided useful information to inform our understanding of
25 interspecies differences and potential sensitivity differences between healthy and susceptible human
26 populations.
27 These criteria provide benchmarks for evaluating various studies and for focusing on the
28 policy-relevant studies in assessing the body of health and welfare effects evidence. Detailed critical
29 analysis of all CO health and welfare effects studies, especially in relation to the above
30 considerations, is beyond the scope of this document. Of most relevance for evaluation of studies is
31 whether they provide useful qualitative or quantitative information on exposure-effect or
32 exposure-response relationships for effects associated with current ambient air concentrations of CO
33 that can inform decisions on whether to retain or revise the standards.
34 In developing the CO ISA, EPA began by reviewing and summarizing the evidence on
35 atmospheric sciences and exposure and the health effects evidence from in vivo and in vitro
36 toxicological studies, controlled human exposure studies, and epidemiologic studies. In November
September 2009 1-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 2008, EPA invited EPA staff and other researchers with expertise in CO to a teleconference meeting
2 to review the scientific content of preliminary draft materials for the draft ISA and the annexes. The
3 purpose of the initial peer review teleconference was to ensure that the ISA is up-to-date and focused
4 on the most policy-relevant findings, and to assist EPA with integration of evidence within and
5 across disciplines. Subsequently, EPA addressed comments and completed the initial integration and
6 synthesis of the evidence.
7 The integration of evidence on health or welfare effects involves collaboration between
8 scientists from various disciplines. As described in the section below, the ISA organization is based
9 on health effect categories. As an example, an evaluation of health effects evidence would include
10 summaries of findings from epidemiologic, controlled human exposure, and toxicological studies,
11 and integration of the results to draw conclusions based on the causal framework described below.
12 Using the causal framework described in Section 1.6, EPA scientists consider aspects such as
13 strength, consistency, coherence and biological plausibility of the evidence, and develop draft
14 judgments on the whether the relationships are causal. The draft integrative synthesis sections and
15 conclusions are reviewed by EPA internal experts and, as appropriate, by outside expert authors. In
16 practice, causality determinations often entail an iterative process of review and evaluation of the
17 evidence. This draft ISA is released for review by the CASAC and the public. Comments on the
18 characterization of the science as well as the implementation of the causal framework are carefully
19 considered in revising and completing the ISA.
1.4. Document Organization
20 The IS A is composed of five chapters. This introductory chapter presents background
21 information, and provides an overview of EPA's framework for making causal judgments. Chapter 2
22 is an integrated summary of key findings and conclusions regarding the source to dose paradigm,
23 MOA, and important health effects of CO, including cardiovascular, nervous system,
24 perinatal/developmental, respiratory, and mortality outcomes. Chapter 3 highlights key concepts and
25 evidence relevant to understanding the sources, ambient concentrations, atmospheric behavior, and
26 exposure to ambient CO. Chapter 4 describes the dosimetry and pharmacokinetics of CO, including
27 formation and fate of carboxyhemoglobin (COHb). Chapter 5 presents a discussion of the MOA of
28 CO and evaluates and integrates epidemiologic, human clinical, and animal toxicological
29 information on the health effects of CO, including cardiovascular and systemic effects, central
30 nervous system (CNS) effects, birth outcomes and developmental effects, respiratory effects, and
31 mortality.
September 2009 1-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 A series of annexes supplement this ISA. The annexes provide tables summarizing additional
2 details of the pertinent literature published since the last review, as well as selected older studies of
3 particular interest. These annexes contain information on:
4 • atmospheric chemistry of CO, sampling and analytic methods for measurement of CO
5 concentrations, emissions, sources and human exposure to CO (Annex A)
6 • studies on the dosimetry and pharmacokinetics of CO (Annex B)
7 • epidemiologic studies of health effects from short- and long-term exposure to CO
8 (Annex C)
9 • controlled human exposure studies of health effects related to exposure to CO (Annex
10 D); and
11 • toxicological studies of health effects in laboratory animals (Annex E)
12 Within the annexes, detailed information about methods and results of health studies is
13 summarized in tabular format, and generally includes information about concentrations of CO and
14 averaging times, study methods employed, results and comments, and quantitative results for
15 relationships between effects and exposure to CO. As noted in the section above, the most pertinent
16 results of this body of studies are brought into the ISA.
1.5. Document Scope
17 For the current review of the primary CO standards, relevant scientific information on human
18 exposures and health effects associated with exposure to ambient CO has been assessed. Health
19 effects resulting from accidental exposures to very high concentrations of non-ambient CO (i.e., CO
20 poisoning) are not directly relevant to ambient exposures, and as such, a discussion of these effects
21 has deliberately been excluded from this document. For a detailed review of the effects of high level
22 exposures to CO, the reader is referred to the extensive body of literature related to CO poisoning
23 (Ernst and Zibrak, 1998, 049822; Penney, 2007, 194668; Raub et al, 2000, 002180). The possible
24 influence of other atmospheric pollutants on the interpretation of the role of CO in health effects
25 studies is considered. This includes other pollutants with the potential to co-occur in the environment
26 (e.g., nitrogen dioxide [NO2], sulfur dioxide [SO2], ozone [O3], and particulate matter [PM]). The
27 review also assesses relevant scientific information associated with known or anticipated public
28 welfare effects that may be identified. As discussed in Section 1.3, a critical review of the ecological
29 effects literature identified no information pertinent to ambient CO exposures; hence, no section on
September 2009 1-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 ecological effects appears in this assessment. The definition of public welfare for the NAAQS
2 includes considerations of climate; thus, the climate forcing effects of CO are summarized in
3 Chapter 2 and discussed in detail in the physics and chemistry section of Chapter 3 where
4 distinctions are drawn between the necessarily global-scale conclusions related to climate and the
5 strongly variable continental and regional climate forcing effects from CO.
1.6. EPA Framework for Causal Determination
6 The EPA has developed a consistent and transparent basis to evaluate the causal nature of air
7 pollution-induced health or environmental effects. The framework described below establishes
8 uniform language concerning causality and brings more specificity to the findings. This standardized
9 language was drawn from across the federal government and wider scientific community, especially
10 from the recent National Academy of Sciences (NAS) Institute of Medicine (IOM) document,
11 Improving the Presumptive Disability Decision-Making Process for Veterans, (2008, 156586) the
12 most recent comprehensive work on evaluating causality.
13 This introductory section focuses on the evaluation of health effects evidence. While focusing
14 on human health outcomes, the concepts are also generally relevant to causality determination for
15 welfare effects. This section:
16 • describes the kinds of scientific evidence used in establishing a general causal
17 relationship between exposure and health effects;
18 • defines cause, in contrast to statistical association;
19 • discusses the sources of evidence necessary to reach a conclusion about the existence of
20 a causal relationship;
21 • highlights the issue of multifactorial causation;
22 • identifies issues and approaches related to uncertainty; and
23 • provides a framework for classifying and characterizing the weight of evidence in
24 support of a general causal relationship.
25 Approaches to assessing the separate and combined lines of evidence (e.g., epidemiologic,
26 human clinical, and animal toxicological studies) have been formulated by a number of regulatory
27 and science agencies, including the IOM of the NAS (2008, 156586). International Agency for
28 Research on Cancer (2006, 093206). EPA Guidelines for Carcinogen Risk Assessment (2005,
29 086237). Centers for Disease Control and Prevention (2004, 056384). and National Acid
September 2009 1-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 Precipitation Assessment Program (1991, 095894). These formalized approaches offer guidance for
2 assessing causality. The frameworks are similar in nature, although adapted to different purposes,
3 and have proven effective in providing a uniform structure and language for causal determinations.
4 Moreover, these frameworks have supported decision-making under conditions of uncertainty.
1.6.1. Scientific Evidence Used in Establishing Causality
5 Causality determinations are based on the evaluation and synthesis of evidence from across
6 scientific disciplines; the type of evidence that is most important for such determinations will vary
7 by pollutant or assessment. The most compelling evidence of a causal relationship between pollutant
8 exposures and human health effects comes from human clinical studies. This type of study
9 experimentally evaluates the health effects of administered exposures in human volunteers under
10 highly-controlled laboratory conditions.
11 In epidemiologic or observational studies of humans, the investigator does not control
12 exposures or intervene with the study population. Broadly, observational studies can describe
13 associations between exposures and effects. These studies fall into several categories:
14 cross-sectional, prospective cohort, and time-series studies. "Natural experiments" offer the
15 opportunity to investigate changes in health with a change in exposure; these include comparisons of
16 health effects before and after a change in population exposures, such as closure of a pollution
17 source.
18 Experimental animal data complement the clinical and observational data; these studies can
19 help characterize effects of concern, exposure-response relationships, susceptible subpopulations and
20 MOAs. In the absence of clinical or epidemiologic data, animal data alone may be sufficient to
21 support a likely causal determination, assuming that humans respond similarly to the experimental
22 species.
1.6.2. Association and Causation
23 "Cause" is a significant, effectual relationship between an agent and an effect on health or
24 public welfare. "Association" is the statistical dependence among events, characteristics, or other
25 variables. An association is prima facie evidence for causation; alone, however, it is insufficient
26 proof of a causal relationship between exposure and disease. Unlike an association, a causal claim
27 supports the creation of counterfactual claims; that is, a claim about what the world would have been
28 like under different or changed circumstances (IOM, 2008, 156586). Much of the newly available
29 health information evaluated in this ISA comes from epidemiologic studies that report a statistical
30 association between ambient exposure and health outcome.
September 2009 1-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 Many of the health and environmental outcomes reported in these studies have complex
2 etiologies. Diseases such as asthma, coronary heart disease (CHD) or cancer are typically initiated
3 by multiple agents. Outcomes depend on a variety of factors, such as age, genetic susceptibility,
4 nutritional status, immune competence, and social factors (Gee and Payne-Sturges, 2004, 093070;
5 IOM, 2008, 156586). Effects on ecosystems are often also multifactorial with a complex web of
6 causation. Further, exposure to a combination of agents could cause synergistic or antagonistic
7 effects. Thus, the observed risk represents the net effect of many actions and counteractions.
1.6.3. Evaluating Evidence for Inferring Causation
8 Moving from association to causation involves elimination of alternative explanations for the
9 association. In estimating the causal influence of an exposure on health or environmental effects, it is
10 recognized that scientific findings incorporate uncertainty. Uncertainty can be defined as a state of
11 having limited knowledge where it is impossible to exactly describe an existing state or future
12 outcome; e.g., the lack of knowledge about the correct value for a specific measure or estimate.
13 Uncertainty characterization and uncertainty assessment are two activities that lead to different
14 degrees of sophistication in describing uncertainty. Uncertainty characterization generally involves a
15 qualitative discussion of the thought processes that lead to the selection and rejection of specific
16 data, estimates, scenarios, etc. The uncertainty assessment is more quantitative. The process begins
17 with simpler measures (e.g., ranges) and simpler analytical techniques and progresses, to the extent
18 needed to support the decision for which the assessment is conducted, to more complex measures
19 and techniques. Data will not be available for all aspects of an assessment and those data that are
20 available may be of questionable or unknown quality. In these situations, evaluation of uncertainty
21 can include professional judgment or inferences based on analogy with similar situations. The net
22 result is that the assessments will be based on a number of assumptions with varying degrees of
23 uncertainty. Uncertainties commonly encountered in evaluating health evidence for the criteria air
24 pollutants are outlined below for epidemiologic and experimental studies. Various approaches to
25 characterizing uncertainty include classical statistical methods, sensitivity analysis, or probabilistic
26 uncertainty analysis, in order of increasing complexity and data requirements. The ISA generally
27 evaluates uncertainties qualitatively in assessing the evidence from across studies; in some situations
28 quantitative analysis approaches, such as metaregression may be used.
29 Controlled human exposure studies evaluate the effects of exposures to a variety of pollutants
30 in a highly controlled laboratory setting. Also referred to as human clinical studies, these
31 experiments allow investigators to expose subjects to known concentrations of air pollutants under
32 carefully regulated environmental conditions and activity levels. In some instances, controlled
33 human exposure studies can also be used to characterize concentration-response relationships at
34 pollutant concentrations relevant to ambient conditions. Controlled human exposures are typically
September 2009 1-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 conducted using a randomized crossover design with subjects exposed both to CO and a clean air
2 control. In this way, subjects serve as their own controls, effectively controlling for many potential
3 confounders. However, human clinical studies are limited by a number of factors including a small
4 sample size and short exposure time. The repetitive nature of ambient CO exposures at levels that
5 can vary widely may lead to cumulative health effects, but this type of exposure is not practical to
6 replicate in a laboratory setting. In addition, although subjects do serve as their own controls,
7 personal exposure to pollutants in the hours and days preceding the controlled exposures may vary
8 significantly between and within individuals. Endogenous production of CO creates a body burden
9 of CO that, together with personal exposure from nonambient sources, contributes to baseline COHb
10 levels. Endogenous production rates vary within and among individuals, particularly for individuals
11 with diseases such as hemolytic anemia or chronic inflammation. This body burden of CO and
12 COHb limits the lower range of exposures that can be practically covered in controlled human
13 exposure studies. Finally, human clinical studies require investigators to adhere to stringent health
14 criteria for a subject to be included in the study, and therefore the results cannot necessarily be
15 generalized to an entire population. Although some human clinical studies have included health-
16 compromised individuals such as those with coronary artery disease (CAD), these individuals must
17 also be relatively healthy and do not represent the most sensitive individuals in the population. Thus,
18 a lack of observation of effects from human clinical studies does not necessarily mean that a causal
19 relationship does not exist. While human clinical studies provide important information on the
20 biological plausibility of associations observed between air pollutant exposure and health outcomes
21 in epidemiologic studies, observed effects in these studies may underestimate the response in certain
22 subpopulations.
23 Epidemiologic studies provide important information on the associations between health
24 effects and exposure of human populations to ambient air pollution. In the evaluation of
25 epidemiologic evidence, one important consideration is potential confounding. Confounding is ". . .
26 a confusion of effects. Specifically, the apparent effect of the exposure of interest is distorted because
27 the effect of an extraneous factor is mistaken for or mixed with the actual exposure effect (which
28 may be null)" (Rothman and Greenland, 1998, 086599). One approach to remove spurious
29 associations due to possible confounders is to control for characteristics that may differ between
30 exposed and unexposed persons; this is frequently termed "adjustment." Scientific judgment is
31 needed regarding likely sources and magnitude of confounding, together with consideration of how
32 well the existing constellation of study designs, results, and analyses address this potential threat to
33 inferential validity. One key consideration in this review is evaluation of the potential contribution of
34 CO to health effects when it is a component of a complex air pollutant mixture. Reported CO effect
35 estimates in epidemiologic studies may reflect independent CO effects on health outcomes. Ambient
36 CO may also be serving as an indicator of complex ambient air pollution mixtures that share the
September 2009 1-14 DRAFT - DO NOT CITE OR QUOTE
-------
1 same source as CO (e.g., motor vehicle emissions). Alternatively, copollutants may mediate the
2 effects of CO or CO may influence the toxicity of copollutants.
3 Multivariable regression models constitute one tool for estimating the association between
4 exposure and outcome after adjusting for characteristics of participants that might confound the
5 results. The use of multipollutant regression models has been the prevailing approach for controlling
6 potential confounding by copollutants in air pollution health effects studies. Finding the likely causal
7 pollutant from multipollutant regression models is made difficult by the possibility that one or more
8 air pollutants may be acting as a surrogate for an unmeasured or poorly-measured pollutant or for a
9 particular mixture of pollutants. In addition, more than one pollutant may exert similar health effects,
10 resulting in independently observed associations for multiple pollutants. For example, PM2.5 and
11 NO2 have each been linked to cardiovascular effects in epidemiologic studies. Correlation between
12 CO concentrations and various copollutants, such as PM2 5 and NO2, makes it difficult to
13 quantitatively interpret associations between different pollutant exposures and health effects. Thus,
14 results of models that attempt to distinguish CO effects from those of copollutants must be
15 interpreted with caution. The number and degree of diversity of covariates, as well as their relevance
16 to the potential confounders, remain matters of scientific judgment. Despite these limitations, the use
17 of multipollutant models is still the prevailing approach employed in most air pollution
18 epidemiologic studies, and provides some insight into the potential for confounding or interaction
19 among pollutants.
20 Another way to adjust for potential confounding is through stratified analysis, i.e., examining
21 the association within homogeneous groups with respect to the confounding variable. The use of
22 stratified analyses has an additional benefit: it allows examination of effect modification through
23 comparison of the effect estimates across different groups. If investigators successfully measured
24 characteristics that distort the results, adjustment of these factors help separate a spurious from a true
25 causal association. Appropriate statistical adjustment for confounders requires identifying and
26 measuring all reasonably expected confounders. Deciding which variables to control for in a
27 statistical analysis of the association between exposure and disease or health outcome depends on
28 knowledge about possible mechanisms and the distributions of these factors in the population under
29 study. Identifying these mechanisms makes it possible to control for potential sources that may result
30 in a spurious association.
31 Adjustment for potential confounders can be influenced by differential exposure measurement
32 error. There are several components that contribute to exposure measurement error in epidemiologic
33 studies, including the difference between true and measured ambient concentrations, the difference
34 between average personal exposure to ambient pollutants and ambient concentrations at central
35 monitoring sites, and the use of average population exposure rather than individual exposure
36 estimates. Consideration of issues important for evaluation of exposure to ambient CO include
September 2009 1-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 spatial variability of CO concentrations across urban areas, particularly with respect to highly
2 traveled roadways; location of CO monitors at varying distances from roads; and the detection limit
3 of instruments in the CO monitoring network. Previous AQCDs have examined the role of
4 measurement error for non-reactive pollutants in time-series epidemiologic studies using simulated
5 data and mathematical analyses and suggested that "transfer of effects" would only occur under
6 unusual circumstances (i.e., "true" predictors having high positive or negative correlation;
7 substantial measurement error; or extremely negatively correlated measurement errors) (U.S. EPA,
8 2004, 0569051
9 Confidence that unmeasured confounders are not producing the findings is increased when
10 multiple studies are conducted in various settings using different subjects or exposures; each of
11 which might eliminate another source of confounding from consideration. Thus, multicity studies
12 which use a consistent method to analyze data from across locations with different levels of
13 covariates can provide insight on potential confounding in associations. Intervention studies, because
14 of their quasi-experimental nature, can be particularly useful in characterizing causation.
15 In addition to clinical and epidemiologic studies, the tools of experimental biology have been
16 valuable for developing insights into human physiology and pathology. Laboratory tools have been
17 extended to explore the effects of putative toxicants on human health, especially through the study of
18 model systems in other species. These studies evaluate the effects of exposures to a variety of
19 pollutants in a highly-controlled laboratory setting, and allow exploration of MOAs or mechanisms
20 by which a pollutant may cause effects. Background knowledge of the biological mechanisms by
21 which an exposure might or might not cause disease can prove crucial in establishing, or negating, a
22 causal claim. Consideration of evidence on the non-hypoxic effects of CO via cell signaling and
23 alteration of heme protein function along with evidence on COHb-mediated hypoxic stress provides
24 a more complete understanding of the biological response to CO. There are, however, uncertainties
25 associated with quantitative extrapolations between laboratory animals and humans on the
26 pathophysiological effects of any pollutant. Animal species can differ from each other in
27 fundamental aspects of physiology and anatomy (e.g., metabolism, airway branching, hormonal
28 regulation) that may limit extrapolation.
29 Interpretations of experimental studies of air pollution effects in laboratory animals, as in the
30 case of environmental comparative toxicology studies, are affected by limitations associated with
31 extrapolation models. The differences between humans and rodents with regard to pollutant
32 absorption and distribution profiles based on metabolism, hormonal regulation, breathing pattern,
33 exposure dose, and differences in lung structure and anatomy all have to be taken into consideration.
34 Also, in spite of a high degree of homology and the existence of a high percentage of orthologous
35 genes across humans and rodents (particularly mice), extrapolation of molecular alterations at the
36 gene level is complicated by species-specific differences in transcriptional regulation. Given these
September 2009 1-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 molecular differences, there are uncertainties associated with quantitative extrapolations at this time
2 between laboratory animals and humans of observed pollutant-induced pathophysiological
3 alterations under the control of widely varying biochemical, endocrine, and neuronal factors.
1.6.4. Application of Framework for Causal Determination
4 EPA uses a two-step approach to evaluate the scientific evidence on health or environmental
5 effects of criteria pollutants. The first step determines the weight of evidence in support of causation
6 and characterizes the strength of any resulting causal classification. The second step includes further
7 evaluation of the quantitative evidence regarding the concentration-response relationships and the
8 loads or levels, duration and pattern of exposures at which effects are observed.
9 To aid judgment, various "aspects"1 of causality have been discussed by many philosophers
10 and scientists. The most widely cited aspects of causality in epidemiology, and public health, in
11 general, were articulated by Sir Austin Bradford Hill in 1965 and have been widely used (CDC,
12 2004, 056384; IARC, 2006, 093206; IOM, 2008, 156586; U.S. EPA, 2005, 086237).These aspects
13 (Hill, 1965, 071664) have been modified (Table 1-2) for use in causal determinations specific to
14 health and welfare effects or pollutant exposures (U.S. EPA, 2009, 179916).2 Some aspects are more
15 likely than others to be relevant for evaluating evidence on the health or environmental effects of
16 criteria air pollutants. For example, the analogy aspect does not always apply, especially for the
17 gaseous criteria pollutants, and specificity would not be expected for multi-etiologic health
18 outcomes, such as asthma or cardiovascular disease, or ecological effects related to acidification.
19 Aspects that usually play a larger role in determination of causality are consistency of results across
20 studies, coherence of effects observed in different study types or disciplines, biological plausibility,
21 exposure-response relationship, and evidence from "natural" experiments.
1 The "aspects" described by Hill (1965, 071664) have become, in the subsequent literature, more commonly described as "criteria." The
original term "aspects" is used here to avoid confusion with 'criteria' as it is used, with different meaning, in the Clean Air Act.
2 The Hill aspects were developed for interpretation of epidemiologic results. They have been modified here for use with a broader array of
data, i.e., epidemiologic, controlled human exposure, and animal toxicological studies, as well as in vitro data, and to be more consistent
with EPA's Guidelines for Carcinogen Risk Assessment.
September 2009 1-17 DRAFT - DO NOT CITE OR QUOTE
-------
Table 1-1 Aspects to aid in judging causality.
An inference of causality is strengthened when a pattern of elevated risks is observed across several independent studies. The
Consistency of the reproducibility of findings constitutes one of the strongest arguments for causality. If there are discordant results among
observed association investigations, possible reasons such as differences in exposure, confounding factors, and the power of the study are
considered.
An inference of causality from epidemiologic associations maybe strengthened by other lines of evidence (e.g., clinical and
animal studies) that support a cause-and-effect interpretation of the association. Evidence on ecological or welfare effects may
be drawn from a variety of experimental approaches (e.g., greenhouse, laboratory, and field) and subdisciplines of ecology
(e.g., community ecology, biogeochemistry and paleological/historical reconstructions). The coherence of evidence from
various fields greatly adds to the strength of an inference of causality. The absence of other lines of evidence, however, is not a
reason to reject causality.
Coherence
An inference of causality tends to be strengthened by consistency with data from experimental studies or other sources
demonstrating plausible biological mechanisms. A proposed mechanistic linking between an effect and exposure to the agent is
an important source of support for causality, especially when data establishing the existence and functioning of those
mechanistic links are available. A lack of biologic understanding, however, is not a reason to reject causality.
Biological plausibility.
A well characterized exposure-response relationship (e.g., increasing effects associated with greater exposure) strongly
Biological gradient suggests cause and effect, especially when such relationships are also observed for duration of exposure (e.g., increasing
(exposure-response effects observed following longer exposure times). There are, however, many possible reasons that a study may fail to detect
relationship) an exposure-response relationship. Thus, although the presence of a biologic gradient may support causality, the absence of
an exposure-response relationship does not exclude a causal relationship.
The finding of large, precise risks increases confidence that the association is not likely due to chance, bias, or other factors.
Strength of the observed However, given a truly causal agent, a small magnitude in the effect could follow from a lower level of exposure, a lower
association potency, or the prevalence of other agents causing similar effects. While large effects support causality, modest effects
therefore do not preclude it.
Experimental evidence.
The strongest evidence for causality can be provided when a change in exposure brings about a change in occurrence or
frequency of health or welfare effects.
Temporal relationship of Evidence of a temporal sequence between the introduction of an agent, and appearance of the effect, constitutes another
the observed association argument in favor of causality.
As originally intended, this refers to increased inference of causality if one cause is associated with a single effect or disease
(Hill, 1965,071664). Based on our current understanding this is now considered one of the weaker guidelines for causality; for
Specificity of the observed example, many agents cause respiratory disease and respiratory disease has multiple causes. At the scale of ecosystems, as
association in epidemiology, complexity is such that single agents causing single effects, and single effects following single causes, are
extremely unlikely. The ability to demonstrate specificity under certain conditions remains, however, a powerful attribute of
experimental studies. Thus, although the presence of specificity may support causality, its absence does not exclude it.
Structure activity relationships and information on the agent's structural analogs can provide insight into whether an association
Analogy is causal. Similarly, information on mode of action for a chemical, as one of many structural analogs, can inform decisions
regarding likely causality.
1 Although these aspects provide a framework for assessing the evidence, they do not lend
2 themselves to being considered in terms of simple formulas or fixed rules of evidence leading to
3 conclusions about causality (Hill, 1965, 071664). For example, one cannot simply count the number
4 of studies reporting statistically significant results or statistically nonsignificant results and reach
5 credible conclusions about the relative weight of the evidence and the likelihood of causality. Rather,
6 these important considerations are taken into account with the goal of producing an objective
7 appraisal of the evidence, informed by peer and public comment and advice, which includes
8 weighing alternative views on controversial issues. In addition, it is important to note that the aspects
9 in Table 1-1 cannot be used as a strict checklist, but rather to determine the weight of the evidence
10 for inferring causality. In particular, not meeting one or more of the principles does not automatically
11 preclude a determination of causality (e.g., see discussion in CDC (2004, 056384)).
September 2009
1-18
DRAFT - DO NOT CITE OR QUOTE
-------
1.6.5. Determination of Causality
1 In the ISA, EPA assesses the results of recent relevant publications, building upon evidence
2 available during the previous NAAQS review, to draw conclusions on the causal relationships
3 between relevant pollutant exposures and health or environmental effects. This ISA uses a five-level
4 hierarchy that classifies the weight of evidence for causation, not just association1; that is, whether
5 the weight of scientific evidence makes causation at least as likely as not, in the judgment of the
6 reviewing group. In developing this hierarchy, EPA has drawn on the work of previous evaluations,
7 most prominently the lOM's Improving the Presumptive Disability Decision-Making Process for
8 Veterans (2008, 156586). EPAs Guidelines for Carcinogen Risk Assessment (2005, 086237). and the
9 U.S. Surgeon General's smoking reports (CDC, 2004, 056384). In the ISA, EPA uses a series of five
10 descriptors to characterize the weight of evidence for causality. This weight of evidence evaluation is
11 based on various lines of evidence from across the health and environmental effects disciplines.
12 These separate judgments are integrated into a qualitative statement about the overall weight of the
13 evidence and causality. The five descriptors for causal determination are described in Table 1-2.
1 It should be noted that the CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A five-level
hierarchy is used here to be consistent with the EPA Guidelines for Carcinogen Risk Assessment and to provide a more nuanced set of
categories.
September 2009 1-19 DRAFT - DO NOT CITE OR QUOTE
-------
Table 1-2 Weight of evidence for causal determination.
Health Effects Ecological and Welfare Effects
Evidence is sufficient to conclude that there is a causal relationship
Evidence is sufficient to conclude that there is a causal relationship with relevant pollutant exposures. That is, the pollutant has been
with relevant pollutant exposures. That is, the pollutant has been shown to result in effects in studies in which chance, bias, and
shown to result in health effects in studies in which chance, bias, and confounding could be ruled out with reasonable confidence.
Causal confounding could be ruled out with reasonable confidence. For Controlled exposure studies (laboratory or small- to medium-scale
relationshirj example: a) controlled human exposure studies that demonstrate field studies) provide the strongest evidence for causality, but the
" consistent effects; or b) observational studies that cannot be explained scope of inference may be limited. Generally, determination is based
by plausible alternatives or are supported by other lines of evidence on multiple studies conducted by multiple research groups, and
(e.g., animal studies or mode of action information). Evidence includes evidence that is considered sufficient to infer a causal relationship is
replicated and consistent high-quality studies by multiple investigators, usually obtained from the joint consideration of many lines of
evidence that reinforce each other.
Evidence is sufficient to conclude that a causal relationship is likely to
±KWpr,wz-i
in studies in which chance and bias can be ruled out with reasonable assoclallon wlln re'eyam pollutant exposures , mat is an association
nas beetl obsetved between the pollutant and the Outcome in
animal toxicological evidence from multiple studies from different
laboratories that demonstrate effects, but limited or no human data are
available. Evidence generally includes replicated and high-quality
studies by multiple investigators.
Suggestiveof b Evidence is suggestive of a caus, relationship with relevant
a causal canot be ruled out. For example, at least one high-quality
relationship epidemiologic study shows an association with a given health outcome ,t th nf n h.r ap
effect' but the results of other studles are lnconslstent-
but the results of other studies are inconsistent.
Inadequate to {jjff^j^ exists The available studies are of insufficient quality, consistency or
±S quS £ a^isS^'a^'^ to permit a -™ J^Mo-m* a conclusion Jegalg the presence or
absence of an effect.
conciusion arding the presence or absence of an effect .
Evidence is suggestive of no causal relationship with relevant pollutant
Not likely to exposures. Several adequate studies, covering the full range of levels Several adequate studies, examining relationships with relevant
be a causal of exposure that human beings are known to encounter and exposures, are consistent in failing to show an effect at any level of
relationship considering susceptible or vulnerable subpopulations, are mutually
exposure.
consistent in not showing an effect at any level of exposure.
1 For the CO ISA, determination of causality involved the evaluation of evidence for different
2 types of health effects associated with short- and long-term exposure periods. In making
3 determinations of causality for CO, evidence was evaluated for health outcome categories, such as
4 cardiovascular effects, and then conclusions were drawn based upon the integration of evidence from
5 across disciplines (e.g., epidemiology, clinical studies and toxicology) and also across the suite of
6 related individual health outcomes. To accomplish this integration, evidence from multiple and
7 various types of studies was considered. Response was evaluated over a range of observations which
8 was determined by the type of study and methods of exposure or dose and response measurements.
9 Results from different protocols were compared and contrasted.
10 In drawing judgments regarding causality for the criteria air pollutants, EPA focuses on
1 1 evidence of effects at relevant pollutant exposures. To best inform reviews of the NAAQS, these
12 evaluations go beyond a determination of causality at any dose or concentration to emphasize the
13 relationship apparent at relevant pollutant exposures. Concentrations generally within an order of
14 magnitude or two of ambient pollutant measurements are considered to be relevant for this
September 2009 1-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 determination. Building upon the determination of causality are questions relevant to quantifying
2 health or environmental risks based on our understanding of the quantitative relationships between
3 pollutant exposures and health or welfare effects. While the causality determination is based
4 primarily on evaluation of health or environmental effects evidence, EPA also evaluates evidence
5 related to the doses or levels at which effects are observed. Considerations relevant to evaluation of
6 quantitative relationships for health and environmental effects are summarized below.
1.6.5.1. Effects on Human Populations
7 Once a determination is made regarding the causal relationship between the pollutant and
8 outcome category, important questions regarding quantitative relationships include:
9 • What is the concentration-response or dose-response relationship in the human
10 population?
11 • What is the interrelationship between incidence and severity of effect?
12 • What exposure conditions (dose or exposure, duration and pattern) are important?
13 • What subpopulations appear to be differentially affected i.e., more susceptible or
14 vulnerable to effects?
15 To address these questions, the entirety of policy-relevant quantitative evidence is evaluated to
16 best quantify those concentration-response relationships that exist. This requires evaluation of
17 pollutant concentrations and exposure durations at which effects were observed for exposed
18 populations, including potentially susceptible subpopulations. This integration of evidence resulted
19 in identification of a study or set of studies that best approximated the concentration-response
20 relationships between health outcomes and CO, given the current state of knowledge and the
21 uncertainties that surrounded these estimates. To accomplish this, evidence is considered from
22 multiple and diverse types of studies. To the extent available, the ISA evaluates results from across
23 epidemiologic studies that use various methods to evaluate the form of relationships between CO
24 and health outcomes, and draws conclusions on the most well-supported shape of these relationships.
25 Animal data may also inform evaluation of concentration-response relationships, particularly relative
26 to MOAs, and characteristics of susceptible subpopulations. Chapter 2 presents the integrated
27 findings informative for evaluation of population risks.
28 An important consideration in characterizing the public health impacts associated with
29 exposure to a pollutant is whether the concentration-response relationship is linear across the full
30 concentration range encountered, or if nonlinear relationships exist along any part of this range. Of
31 particular interest is the shape of the concentration-response curve at and below the level of the
September 2009 1-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 current standards. The shape of the concentration-response curve varies, depending on the type of
2 health outcome, underlying biological mechanisms and dose. At the human population level,
3 however, various sources of variability and uncertainty (such as the low data density in the lower
4 concentration range, possible influence of measurement error, and individual differences in
5 susceptibility to air pollution health effects) tend to smooth and "linearize" the
6 concentration-response function. In addition, many chemicals and agents may act by perturbing
7 naturally occurring background processes that lead to disease, which also linearizes population
8 concentration-response relationships (Clewell and Crump, 2005, 156359; Crump et al., 1976,
9 003192; Hoel, 1980, 156555). These attributes of population dose-response may explain why the
10 available human data at ambient concentrations for some environmental pollutants (e.g., PM, O3,
11 lead [Pb], environmental tobacco smoke [ETS], radiation) do not exhibit evident thresholds for
12 cancer or noncancer health effects, even though likely mechanisms include nonlinear processes for
13 some key events. These attributes of human population dose-response relationships have been
14 extensively discussed in the broader epidemiologic literature (Rothman and Greenland, 1998,
15 086599).
16 Publication bias is a source of uncertainty regarding the magnitude of health risk estimates. It
17 is well understood that studies reporting non-null findings are more likely to be published than
18 reports of null findings, and publication bias can also result in overestimation of effect estimate sizes
19 (loannidis, 2008, 188317). For example, effect estimates from single-city epidemiologic studies have
20 been found to be generally larger than those from multicity studies (Anderson et al., 2005, 087916)
21 Although publication bias commonly exists for many research areas, it may be present to a lesser
22 degree for epidemiologic studies on CO. In general, epidemiologic studies have focused on the
23 effects of PM, and CO was largely considered as a potentially confounding copollutant of PM; thus,
24 CO effect estimates may have been presented in these studies regardless of the statistical significance
25 of the results.
26 Finally, identification of the susceptible or vulnerable population groups contributes to an
27 understanding of the public health impact of pollutant exposures. Epidemiologic studies can help
28 identify susceptible subpopulations by evaluating health responses in the study population. Examples
29 include stratified analyses for subsets of the population under study, or testing for interactions or
30 effect modification by factors such as gender, age group, or health status. Experimental studies using
31 animal models of susceptibility or disease can also inform the extent to which health risks are likely
32 greater in specific population subgroups.
1.6.5.2. Effects on Ecosystems or Public Welfare
33 Key questions for understanding the quantitative relationships between exposure (or
34 concentration or deposition) to a pollutant and risk to ecosystems or the public welfare include:
September 2009 1-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 • What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
2 functions, etc.) appear to be affected, or are more sensitive to effects?
3 • Under what exposure conditions (amount deposited or concentration, duration and
4 pattern) are effects seen?
5 • What is the shape of the concentration-response or exposure-response relationship?
6 Evaluations of causality generally consider the probability of quantitative changes in
7 ecological and welfare effects in response to exposure. A challenge to the quantification of exposure-
8 response relationships for ecological effects is the great regional and local variability in ecosystems.
9 Thus, exposure-response relationships are often determined for a specific ecological system and
10 scale, rather than at the national or even regional scale. Quantitative relationships therefore are
11 available site by site. For example, an ecological response to deposition of a given pollutant can
12 differ greatly between ecosystems. Where results from greenhouse or animal ecotoxicological
13 studies are available, they may be used to aid in characterizing exposure-response relations,
14 particularly relative to mechanisms of action, and characteristics of sensitive biota.
1.6.6. Concepts in Evaluating Adversity of Health Effects
15 In evaluating the health evidence, a number of factors can be considered in determining the
16 extent to which health effects are "adverse" for health outcomes such as changes in lung function or
17 in cardiovascular health measures. Some health outcome events, such as hospitalization for
18 respiratory or cardiovascular diseases, are clearly considered adverse; what is more difficult is
19 determining the extent of change in the more subtle health measures that is adverse. What constitutes
20 an adverse health effect may vary between populations. Some changes in healthy individuals may
21 not be considered adverse while those of a similar type and magnitude are potentially adverse in
22 more susceptible individuals.
23 For example, the extent to which changes in lung function are adverse has been discussed by
24 the American Thoracic Society (ATS) in an official statement titled What Constitutes an Adverse
25 Health Effect of Air Pollution? (2000, 011738). This statement updated the guidance for defining
26 adverse respiratory health effects that had been published 15 years earlier (ATS, 1985, 006522).
27 taking into account new investigative approaches used to identify the effects of air pollution and
28 reflecting concern for impacts of air pollution on specific susceptible groups. In the 2000 update,
29 there was an increased focus on quality of life measures as indicators of adversity and a more
30 specific consideration of population risk. Exposure to air pollution that increases the risk of an
31 adverse effect to the entire population is viewed as adverse, even though it may not increase the risk
32 of any identifiable individual to an unacceptable level. For example, a population of asthmatics
September 2009 1-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 could have a distribution of lung function such that no identifiable individual has a level associated
2 with significant impairment. Exposure to air pollution could shift the distribution such that no
3 identifiable individual experiences clinically relevant effects; this shift toward decreased lung
4 function, however, would be considered adverse because individuals within the population would
5 have diminished reserve function and, therefore, would be at increased risk to further environmental
6 insult.
7 It is important to recognize that the more subtle health outcomes may be linked to health
8 events that are clearly adverse. For example, air pollution has been shown to affect markers of
9 transient myocardial ischemia such as ST-segment abnormalities and onset of exertional angina. In
10 some cases, these effects are silent yet may still increase the risk of a number of cardiac events,
11 including myocardial infarction and sudden death.
1.7. Summary
12 This second external review draft ISA is a concise evaluation and synthesis of the most
13 policy-relevant science for reviewing the NAAQS for CO, and it is the chief means for
14 communicating the critical science judgments relevant to that NAAQS review. It reviews the most
15 policy-relevant evidence from atmospheric science, exposure, and health and environmental effects
16 studies and includes mechanistic evidence from basic biological science. This draft ISA incorporates
17 clarification and revisions based on public comments and advice and comments provided by EPA's
18 CAS AC (Brain and Samet, 2009, 194669). Annexes to the ISA provide additional details of the
19 literature published since the last review. A framework for making critical judgments concerning
20 causality was presented in this chapter. It relies on a widely accepted set of principles and
21 standardized language to express evaluation of the evidence. This approach can bring rigor and
22 clarity to the current and future assessments. This ISA should assist EPA and others, now and in the
23 future, to accurately represent what is presently known—and what remains unknown—concerning
24 the effects of CO on human health and public welfare.
September 2009 1-24 DRAFT - DO NOT CITE OR QUOTE
-------
References
ATS. (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special reference to epidemiologic
studies of air pollution. Am Rev Respir Dis, 131: 666-668. 006522
ATS. (2000). What constitutes an adverse health effect of air pollution? Official statement of the American Thoracic
Society. Am J Respir Grit Care Med, 161: 665-673. 011738
Anderson HR; Atkinson RW; Peacock JL; Sweeting MJ; Marston L. (2005). Ambient particulate matter and health effects:
Publication bias in studies of short-term associations. Epidemiology, 16: 155-163. 087916
Brain JD; Samet JM. (2009). Review of EPA's Integrated Science Assessment for Carbon Monoxide (First External Review
Draft). CleanAir Scientific Advisory Committee, Science Advisory Board. Washington, DC. EPA-CASAC-09-011.
194669
CDC. (2004). The health consequences of smoking: A report of the Surgeon General. Centers for Disease Control and
Prevention, U.S. Department of Health and Human Services. Washington, DC. 056384
Clewell HJ; Crump KS. (2005). Quantitative estimates of risk for noncancer endpoints. Risk Anal, 25: 285-289. 156359
Crump KS; Hoel DG; Langley CH; Peto R. (1976). Fundamental carcinogenic processes and their implications for low
dose risk assessment. Cancer Res, 36: 2973-2979. 003192
Ernst A; Zibrak JD. (1998). Carbon monoxide poisoning. N Engl J Med, 339: 1603-1608. 049822
Gee GC; Payne-Sturges DC. (2004). Environmental health disparities: A framework integrating psychosocial and
environmental concepts. Environ Health Perspect, 112: 1645-1653. 093070
Hill AB. (1965). The environment and disease: Association or causation. J R Soc Med, 58: 295-300. 071664
Hoel DG. (1980). Incorporation of background in dose-response models. Fed Proc, 39: 73-75. 156555
IARC. (2006). IARC monographs on the evaluation of carcinogenic risks to humans: Preamble. International Agency for
Research on Cancer. Lyon,France . http://monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdf. 093206
IOM. (2008). Improving the Presumptive Disability Decision-Making Process for Veterans; Committee on Evaluation of
the Presumptive Disability Decision-Making Process for Veterans, Board on Military and Veterans Health.
Washington, DC: Institue of Medicine of the National Academies, National Academies Press. 156586
loannidis JPA. (2008). Why most discovered true associations are inflated. Epidemiology, 19: 640-648. 188317
McClellanRO. (1991). CASAC Closure on the 1991 AQCD for Carbon Monoxide. CleanAir Scientific Advisory
Committee, Science Advisory Board, U.S. Environmental Protection Agency. Washington, DC. US EPA-SAB-
CASAC-LTR-92-015. 194666
McClellanRO. (1992). CASAC Closure on the OAQPS Staff Paper for Carbon Monoxide. CleanAir Scientific Advisory
Committee, Science Advisory Board, U.S. Environmenal Protection Agency. Washington, DC. US EPA-SAB-
CASACLTR-2-016. 194667
NAPAP. (1991). The experience and legacy of NAPAP report of the Oversight Review Board. National Acid Precipitation
Assessment Program. Washington, DC. 095894
NRC. (2003). Managing carbon monoxide pollution in meteorological and topographical problem areas. Washington, DC:
National Research Council of the National Academies Press. 042550
Penney DG. (2007). Carbon Monoxide Poisoning. Boca Raton, FL: CRC Press. 194668
Raub JA; Mathieu-Nolf M; Hampson NB; Thorn SR. (2000). Carbon monoxide poisoning—a public health perspective.
Toxicology, 145: 1-14. 002180
RothmanKJ; Greenland S. (1998). Modern epidemiology. Philadelphia, PA: Lippincott-Raven Publishers. 086599
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 1-25 DRAFT - DO NOT CITE OR QUOTE
-------
U.S. EPA U.S. EPA. (1979). Assessment of adverse health effects from carbon monoxide and implications for possible
modifications of the standard. U.S. EPA. Research Triangle Park, NC. 194665
U. S. EPA. (1979). Air quality criteria for carbon monoxide. U. S. Environmental Protection Agency. Washington, DC.
017687
U. S. EPA. (1984). Review of the NAAQS for carbon monoxide: Reassessment of scientific and technical information.
Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park, NC.
012691
U. S. EPA. (1984). Revised evaluation of health effects associated with carbon monoxide exposure: an addendum to the
1979 EPA air quality criteria document for carbon monoxide. 012690
U.S. EPA. (1991). Air quality criteria for carbon monoxide. 017643
U.S. EPA. (1992). Review of the national ambient air quality standards for carbon monoxide: 1992 reassessment of
scientific and technical information OAQPS staff paper. 084191
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 600/P-
99/001F. 000907
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U. S. EPA. (2005). Guidelines for carcinogen risk assessment, Risk Assessment Forum Report. U. S. Environmental
Protection Agency. Washington, DC. EPA/630/P-03/001F. http://cfpub.epa.gov/ncea/index.cfm. 086237
U.S. EPA. (2008). Plan for review of the National Ambient Air Quality Standards for carbon monoxide. U.S.
Environmental Protection Agency. Research Triangle Park, NC. EPA?HQ?OAR?2008?0015; FRL?8792?1. 193995
U.S. EPA. (2009). Integrated Science Assessment for Particulate Matter. U.S. Environmental Protection Agency.
Washington, DC. EPA/600/R-08/139. 179916
September 2009 1-26 DRAFT - DO NOT CITE OR QUOTE
-------
Chapter 2. Integratiye Health Effects
Overview
1 The subsequent chapters of this ISA present the most policy-relevant information related to
2 this review of the NAAQS for CO, including a synthesis of the evidence presented in the 2000 CO
3 AQCD along with the results of more recent studies. This chapter integrates important findings from
4 the disciplines evaluated in this current assessment of the CO scientific literature, which includes the
5 atmospheric sciences, ambient air data analyses, climate forcing effects, exposure assessment,
6 dosimetry, and health effects research (animal toxicological studies, controlled human exposure
7 studies, and epidemiologic studies). The EPA framework for causal determinations described in
8 Chapter 1 has been applied to the body of evidence evaluated in this assessment in order to
9 characterize the relationship between exposure to CO at relevant concentrations and health effects.
10 The EPA framework applied here employs a five-level hierarchy that classifies the weight of
11 evidence for causation:
12 • Causal relationship
13 • Likely to be a causal relationship
14 • Suggestive of a causal relationship
15 • Inadequate to infer a causal relationship
16 • Not likely to be a causal relationship
17 This evaluation led to causal determinations for several health outcome categories and
18 characterization of the magnitude of the response, including responses in susceptible populations,
19 over a range of relevant concentrations. This integration of evidence also provides a basis for
20 characterizing the concentration-response relationships of CO and adverse health outcomes for the
21 U.S. population, given the current state of knowledge.
22 This chapter summarizes and integrates the newly available scientific evidence that best
23 informs consideration of the policy-relevant questions that frame this assessment, which are
24 presented in Chapter 1. Section 2.1 discusses the trends in ambient concentrations and sources of
25 CO. Section 2.2 provides an overview of climate forcing related directly and indirectly to CO.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 2-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 Section 2.3 provides a brief summary of factors influencing personal exposure to ambient CO.
2 Section 2.4 summarizes CO dosimetry and pharmacokinetics and describes what is known regarding
3 the modes of action of CO. Section 2.5 integrates the evidence from studies that examined health
4 effects related to short- and long-term exposure to CO and discusses important uncertainties
5 identified in the interpretation of the scientific evidence. Section 2.6 summarizes policy-relevant
6 considerations associated with exposure to CO including evidence of effects in potentially
7 susceptible populations and information on the shape of the concentration-response function. Finally,
8 Section 2.7 presents an integrated summary of the health effects of CO, reports the levels at which
9 effects are observed, and discusses important uncertainties to consider in the interpretation of the
10 scientific evidence.
2.1. Ambient CO Sources and Concentrations
11 CO is formed by incomplete combustion of carbon-containing fuels and by photochemical
12 reactions in the atmosphere. Nationally, on-road mobile sources constituted more than half of total
13 CO emissions in 2002, or ~63 of-109 million tons (MT) of total CO emissions, based on the most
14 recent publicly available data meeting data quality objectives from EPA's National Emissions
15 Inventory (NEI). In metropolitan areas in the U.S., as much as 75% of all CO emissions result from
16 on-road vehicle exhaust. The majority of these on-road CO emissions are derived from gasoline-
17 powered vehicles. When emissions from incomplete combustion of fuels powering non-road mobile
18 sources, such as farm and construction equipment, lawnmowers, boats, ships, snowmobiles, and
19 aircraft are included, all mobile sources accounted for -80% of total CO emissions in the U.S. in
20 2002. Other primary sources of CO include wildfires, controlled vegetation burning, residential
21 biomass combustion, and industrial processes. While CO emissions from non-road mobile sources,
22 fire, and industry have remained fairly constant, on-road mobile source CO emissions have
23 decreased by roughly 5% per year since the early 1990s. Secondary sources of CO, which can be
24 large in some areas, include the oxidation of both anthropogenic and biogenic hydrocarbons such as
25 methane and isoprene and other carbon containing species including aldehydes and alcohols.
26 Significant reductions in ambient CO concentrations and in the number of NAAQS
27 exceedances have been observed over the past 25 yr, a continuation of trends documented in the
28 2000 CO AQCD. Nationwide ambient CO data from the EPA Air Quality System (AQS), for the
29 years 2005-2007, show that the median 1-h daily maximum (max) concentration across the U.S. was
30 0.7 ppm; the mean was 0.9 ppm; the 95th percentile was 2.4 ppm; and the 99th percentile was
31 3.8 ppm. The median 8-h daily max ambient CO concentration for the years 2005-2007 was
32 0.5 ppm; the mean was 0.7 ppm; the 95th percentile was 1.7 ppm; and the 99th percentile was
33 2.6 ppm. The current CO NAAQS are 35 ppm (1-h avg) and 9 ppm (8-h avg), not to be exceeded
September 2009 2-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 more than once per year. During the years 2005-2007, 1-h and 8-h CO concentrations did not exceed
2 the NAAQS level more than once per year at any monitoring site. Moreover, in these 3 yr, a 1-h avg
3 concentration in excess of 35 ppm was reported only once (39 ppm), and there were only 7 reported
4 8-h avg values nationwide in excess of 9 ppm in all 3 yr. Seasonally divided box plots of data from
5 2005-2007 compiled for spatially diverse urban metropolitan areas illustrate the tendency for higher
6 median CO concentrations and wider variations in concentrations in the winter and fall compared
7 with the spring and summer (see Section 3.5).
8 Policy-relevant background (PRB) concentrations include contributions from natural sources
9 everywhere in the world and from anthropogenic sources outside the U.S., Canada, and Mexico.
10 PRB concentrations of CO were estimated for this assessment using data for the years 2005-2007
11 collected at 12 remote sites in the U.S. which are part of the National Oceanic and Atmospheric
12 Administration's (NOAA) Global Monitoring Division (GMD) and are not part of the EPA national
13 regulatory network. The 3-yr avg CO PRB averaged ~0.13 ppm in Alaska, ~0.10 ppm in Hawaii, and
14 ~0.13 ppm over the contiguous U.S. (CONUS). (Note that the analysis for North American PRB in
15 this assessment was made by segregating the three Alaska sites based on their high latitude and the
16 two Hawaii sites based on their distance from the continent and then treating the remaining seven
17 sites as representative of the CONUS PRB.)
2.2. Climate Forcing Effects
18 Recent data do not alter the current well-established understanding of the role of urban and
19 regional CO in continental and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
20 2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
21 Intergovernmental Panel on Climate Change (2001, 156587; 2007, 092765). CO is a weak direct
22 contributor to greenhouse warming because its fundamental absorption band near 4.63 (im is far
23 from the spectral maximum of Earth's longwave radiation at ~10 (im. Sinha and Toumi (1996,
24 193747) estimated the direct radiative forcing (RF) of CO computed for all-sky conditions at the
25 tropopause - IPCC's preferred form for the calculation (2007, 092765) - to be 0.024 W/m2 from the
26 change in CO mean global concentration since pre-industrial times. The RF value similarly
27 computed by Sinha and Toumi (1996, 193747) for a more than two-fold increase in the current mean
28 global background concentration to 0.290 ppm was 0.025 W/m2.
29 However, because reaction with CO is the major sink for OH on a global scale, increased
30 concentrations of CO can lead to increased concentrations of other trace gases whose loss processes
31 also involve OH chemistry. Some of those trace gases, CH4 and O3 for example, absorb infrared
32 radiation from the Earth's surface and contribute to the greenhouse effect directly; others, including
33 the hydrochlorofluorocarbons (HCFCs), methyl chloride, and methyl bromide, can deplete
September 2009 2-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 stratospheric O3, increasing the surface-incident UV flux. Because of these chemical
2 interdependencies, calculations of an indirect RF for any of these short-lived O3 precursor species
3 are most often made for all of the most important ones together. So, for example, the combined effect
4 of increased CH4, CO, NMVOC, and NOX emissions since 1750 has produced tropospheric O3
5 concentrations associated with a net RF of-0.35 W/m2 (IPCC, 2001, 156899). The integrated 20-yr
6 and 100-yr time horizon RFs were computed by IPCC (2007, 092936) for year 2000 emissions of
7 CO, NMVOC, and NOX to be ~0.19 W/m2, just slightly lower than the RF of year 2000 black carbon
8 emissions from fossil fuel and biomass burning on the same horizons.
9 Overall, the evidence reviewed in this assessment is sufficient to conclude that a Causal
10 relationship exists between current atmospheric concentrations of CO and effects on
11 Climate. The most significant of these effects do not arise directly from the CO molecules; rather
12 they result indirectly from CO's role in the CO-CH4-O3-NOx-OH chemical system in the
13 atmosphere, and are mediated by the greenhouse gas species CH4, O3, and CO2 produced by
14 reactions with CO. The combined RF computed for all emissions and changes in CO in the years
15 1750-2005 for all indirect effects of CO through O3, CH4, and CO2 was determined by IPCC (2007,
16 092936) to be ~0.2 W/m2. Of the three indirect effects from CO emissions, the O3-related component
17 was the largest, accounting for approximately one-half of this radiative forcing (IPCC, 2007,
18 092936).
2.3. Exposure to Ambient CO
19 Very few recent exposure assessment studies involve ambient CO concentration data. The
20 studies of personal exposure to ambient CO presented here generally found that the largest
21 percentage of time in which an individual is exposed to ambient CO occurs indoors, but the highest
22 ambient CO exposure levels occur in transit. In-vehicle CO concentrations are typically reported to
23 be between 2 and 5 times higher than ambient concentrations measured at the roadside, but have
24 been reported to be as much as 25 times higher. Among commuters, exposures were higher for those
25 traveling in automobiles in comparison with those traveling on buses and motorbikes and with those
26 cycling or walking. Ambient CO exposure in automobiles has been demonstrated to vary with
27 vehicle ventilation settings, and a very small portion of that exposure is thought to come from the
28 vehicle in which the exposed person travels. High near-road CO concentrations can be important for
29 those living in the near-road environment because virtually all of ambient CO infiltrates indoors.
30 Hence, indoor exposure to ambient CO is determined by the CO concentration outside the building.
31 CO concentration in the near-road environment has been shown to decrease sharply with downwind
32 distance from a highway; wind direction, emission source strength (e.g., number of vehicles on a
33 highway), and natural and urban topography also influence localized ambient CO concentrations.
September 2009 2-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 Recent exposure assessment studies support one of the main conclusions of the 2000 CO
2 AQCD that central site ambient CO monitors may overestimate or underestimate individuals'
3 personal exposure to ambient CO because ambient CO concentration is spatially variable,
4 particularly when analyzing exposures in the near-road environment. Exposure error may occur
5 when the ambient CO concentration measured at the central site monitor is used as an ambient
6 exposure surrogate and differs from the actual ambient CO concentration outside a subject's
7 residence and/or worksite. For example, measurement at a "hot spot" could skew community
8 exposure estimates upwards, and likewise measurement at a location with few CO sources could
9 skew exposure estimates downwards. Correlations across CO monitors can vary widely within and
10 between cities across the U.S. as a function of natural and urban topography, meteorology, source
11 strength and proximity to sources. Typically, intersampler correlation ranges from 0.35 to 0.65 for
12 monitors sited at different scales within a metropolitan area, although it can be greater than 0.8 in
13 some areas. Health effects estimates from time-series epidemiologic studies are not biased by spatial
14 variability in CO concentrations if concentrations at different locations are correlated in time.
15 Exposure assessment is also complicated by the existence of CO in multipollutant mixtures emitted
16 by combustion processes, making it difficult to quantify the health effects related specifically to CO
17 exposure compared with those related to another combustion-related pollutant or mix of pollutants.
18 In most circumstances, exposure error tends to bias a health effect estimate downward (Sheppard et
19 al., 2005, 079176: Zeger et al, 2000, 001949). Spatial and temporal variability not fully captured by
20 ambient monitors and correlation of CO with copollutants are examples of sources of uncertainty
21 that could widen confidence intervals of health effects estimates.
2.4. Dosimetry, Pharmacokinetics, and Mode of Action
2.4.1.Dosimetry and Pharmacokinetics
22 Upon inhalation, CO elicits various health effects by binding to and altering the function of a
23 number of heme-containing molecules, mainly hemoglobin (Hb). The formation of COHb reduces
24 the oxygen (O2)-carrying capacity of blood and impairs the release of O2 from oxyhemoglobin
25 (O2Hb) to the tissues. The 2000 CO AQCD has a detailed description of the well-established
26 Coburn-Forster-Kane (CFK) equation, which has been used for many years to model COHb
27 formation. Since then, models have been developed that include myoglobin (Mb) and extravascular
28 storage compartments, as well as other dynamics of physiology relevant to CO uptake and
29 elimination. These models have indicated that CO has a biphasic elimination curve, due to initial
30 washout from the blood followed by a slower flux from the tissues. The flow of CO between the
31 blood and alveolar air or tissues is controlled by diffusion down the pCO gradient. The uptake of CO
September 2009 2-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 is governed not only by this CO pressure differential, but also by physiological parameters, such as
2 minute ventilation and lung diffusing capacity, that can, in turn, be affected by factors such as
3 exercise, age, and medical conditions (e.g., obstructive lung disease). Susceptible populations, such
4 as health-compromised individuals, are at a greater risk from COHb induced health effects due to
5 altered CO kinetics, compromised cardiopulmonary processes, and increased baseline hypoxia
6 levels. Altitude also may have a substantial effect on the kinetics of COHb formation, especially for
7 visitors to high altitude areas. Compensatory mechanisms, such as increased cardiac output, combat
8 the decrease in barometric pressure. Altitude also increases the endogenous production of CO
9 through upregulation of heme oxygenase (HO). CO is considered a second messenger and is
10 endogenously produced from the catabolism of heme proteins by enzymes such as HO-1 (the
11 inducible form of heme oxygenase) and through endogenous lipid peroxidation. Finally, CO is
12 removed from the body by expiration and oxidation to CO2.
2.4.2.Mode of Action
13 The diverse effects of CO are dependent upon concentration, duration of exposure, and the cell
14 types and tissues involved. Responses to CO are not necessarily due to a single process and may
15 instead be mediated by a combination of effects including COHb-mediated hypoxic stress and other
16 mechanisms such as free radical production and the initiation of cell signaling. However, binding of
17 CO to reduced iron in heme proteins with subsequent alteration of heme protein function is the
18 common mechanism underlying the biological responses to CO (see Section 5.1).
19 As discussed in the 2000 CO AQCD, the most well-known pathophysiological effect of CO is
20 tissue hypoxia caused by binding of CO to Hb. Not only does the formation of COHb reduce the O2-
21 carrying capacity of blood, but it also impairs the release of O2 from O2Hb. Compensatory
22 alterations in hemodynamics, such as vasodilation and increased cardiac output, protect against
23 tissue hypoxia. Depending on the extent of CO exposure, these compensatory changes may be
24 effective in people with a healthy cardiovascular system. However, hemodynamic responses
25 following CO exposure may be insufficient in people with decrements in cardiovascular function,
26 resulting in health effects as described in Section 5.2. Binding of CO to Mb, as discussed in the 2000
27 CO AQCD and in Section 4.3.2.1, can also impair the delivery of O2 to tissues. Mb has a high
28 affinity for CO, about 25 times that of O2; however, pathophysiologic effects are seen only after high
29 dose exposures to CO, resulting in COMb concentrations far above baseline levels.
30 Non-hypoxic mechanisms underlying the biological effects of CO have been the subject of
31 recent research since the 2000 CO AQCD. Most of these mechanisms are related to CO's ability to
32 bind heme-containing proteins other than Hb and Mb. These mechanisms, which may be interrelated,
33 include alteration in nitric oxide (NO) signaling, inhibition of cytochrome c oxidase, heme loss from
34 proteins, disruption of iron homeostasis, alteration in cellular redox status, alteration in ion channel
September 2009 2-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 activity and modulation of protein kinase pathways. CO is a ubiquitous cell signaling molecule with
2 numerous physiological functions. The endogenous generation and release of CO from heme by HO-
3 1 and HO-2 is tightly controlled, as is any homeostatic process. However, exogenously-applied CO
4 has the capacity to disrupt multiple heme-based signaling pathways due to its nonspecific nature.
5 Only a limited amount of information is available regarding the impact of exogenous CO on tissue
6 and cellular levels of CO and on signaling pathways. However recent animal studies demonstrated
7 increased tissue CO levels and biological responses following exposure to 50 ppm CO. Whether or
8 not environmentally-relevant exposures to CO lead to adverse health effects through altered cell
9 signaling is an open question for which there are no definitive answers at this time. However,
10 experiments demonstrating oxidative/nitrosative stress, inflammation, mitochondrial alterations and
11 endothelial dysfunction at concentrations of CO within 1 or 2 orders of magnitude higher than
12 ambient concentrations suggest a potential role for such mechanisms in pathophysiologic responses.
13 Furthermore, prolonged increases in endogenous CO resulting from chronic diseases may provide a
14 basis for the enhanced sensitivity of susceptible populations to CO-mediated health effects such as is
15 seen in individuals with coronary artery disease.
2.5. Health Effects
16 This assessment reviewed health effects evidence regarding the effect of CO on several
17 categories of health outcomes. Table 2-1 presents the overall conclusions of the ISA regarding the
18 presence of a causal relationship between exposure to relevant CO concentrations and health
19 outcome categories. Summaries of the evidence supporting each causal determination and
20 considerations relevant to application of the causal framework are provided in the following
21 subsections.
September 2009 2-7 DRAFT - DO NOT CITE OR QUOTE
-------
Table 2-1 Causal determinations for health effects categories.
Outcome Category
Cardiovascular morbidity
Central nervous system effects
Birth outcomes and Developmental effects
Respiratory morbidity
Mortality
Exposure Period
Short-term
Short- and long-term
Long-term
Short-term
Long-term
Short-term
Long-term
Causality Determination
Likely to be a causal relationship
Suggestive of a causal relationship
Suggestive of a causal relationship
Suggestive of a causal relationship
Inadequate to infer a causal relationship
Suggestive of a causal relationship
Not likely to be a causal relationship
2.5.1.Cardiovascular Morbidity
1 The most compelling evidence of a CO-induced effect on the cardiovascular system at COHb
2 levels relevant to the current NAAQS comes from a series of controlled human exposure studies
3 among individuals with coronary artery disease (CAD) (see Section 5.2). These studies, described in
4 the 1991 and 2000 CO AQCDs, demonstrate consistent decreases in the time to onset of exercise-
5 induced angina and ST-segment changes following CO exposures resulting in COHb levels of 3-6%,
6 with one multicenter study reporting similar effects at COHb levels as low as 2.0-2.4% (see
7 Section 5.2.4). No human clinical studies have evaluated the effect of controlled exposures to CO
8 resulting in COHb levels lower than 2%. Human clinical studies published since the 2000 CO
9 AQCD have reported no association between CO and ST-segment changes or arrhythmia; however,
10 none of these studies included individuals with diagnosed heart disease.
11 While the exact physiological significance of the observed ST-segment changes among
12 individuals with CAD is unclear, ST-segment depression is a known indicator of myocardial
13 ischemia. It is also important to note that the individuals with CAD who participated in these
14 controlled exposure studies may not be representative of the most sensitive individuals in the
15 population. It is conceivable that the most sensitive individuals respond to levels of COHb lower
16 than 2%. Variability in activity patterns and severity of disease among individuals with CAD is
17 likely to influence the critical level of COHb which leads to adverse cardiovascular effects.
18 The degree of ambient CO exposure which leads to attainment of critical levels of COHb will
19 also vary between individuals. Although endogenous COHb is generally less than 1% in healthy
20 individuals, higher endogenous COHb levels are observed in individuals with certain medical
21 conditions. Nonambient exposures to CO, such as exposure to environmental tobacco smoke (ETS),
22 may increase COHb above endogenous levels, depending on the gradient of pCO. Ambient
23 exposures may cause a further increase in COHb. Modeling results described in Chapter 4 indicate
September 2009
2-8
DRAFT - DO NOT CITE OR QUOTE
-------
1 that increases of-1% COHb are possible with exposures of several ppm CO depending on exposure
2 duration and exercise level.
3 Findings of epidemiologic studies conducted since the 2000 CO AQCD are coherent with
4 results of the controlled human exposure studies. These recent studies observed associations between
5 ambient CO concentration and emergency department (ED) visits and hospital admissions for
6 ischemic heart disease (IHD), congestive heart failure (CHF) and cardiovascular diseases (CVD) as a
7 whole and were conducted in locations where the mean 24-h avg CO concentrations ranged from
8 0.5 ppm to 9.4 ppm (Table 5-7). All of these studies that evaluated CAD outcomes (IHD, MI,
9 angina) reported positive associations (Figure 5-2). Although CO is often considered a marker for the
10 effects of another traffic-related pollutant or mix of pollutants, evidence indicates that CO
11 associations generally remain robust in copollutant models and supports a direct effect of short-term
12 ambient CO exposure on CVD morbidity. These studies add to findings reported in the 2000 CO
13 AQCD that demonstrated associations between short-term variations in ambient CO concentrations
14 and exacerbation of heart disease.
15 The known role of CO in limiting O2 availability lends biological plausibility to ischemia-
16 related health outcomes following CO exposure. However, it is not clear whether the small changes
17 in COHb associated with ambient CO exposures results in substantially reduced O2 delivery to
18 tissues. Recent toxicological studies suggest that CO may also act through other mechanisms by
19 initiating or disrupting cellular signaling. Studies in healthy animals demonstrated oxidative injury
20 and inflammation in response to 50-100 ppm CO while studies in animal models of disease
21 demonstrated exacerbation of cardiomyopathy and increased vascular remodeling in response to
22 50 ppm CO. Further investigations will be useful in determining whether altered cell signaling
23 contributes to adverse health effects following ambient CO exposure.
24 Given the consistent and coherent evidence from epidemiologic and human clinical studies,
25 along with biological plausibility provided by CO's role in limiting O2 availability, it is concluded
26 that a causal relationship is likely to exist between relevant short-term CO exposures and
27 cardiovascular morbidity.
2.5.2.Central Nervous System Effects
28 Exposure to high levels of CO has long been known to adversely affect central nervous system
29 (CNS) function, with symptoms following acute CO poisoning including headache, dizziness,
30 cognitive difficulties, disorientation, and coma. However, the relationship between ambient levels of
31 CO and neurological function is less clear and has not been evaluated in epidemiologic studies.
32 Studies of controlled human exposures to CO discussed in the 2000 CO AQCD reported inconsistent
33 neural and behavioral effects following exposures resulting in COHb concentrations of 5-20%. No
34 new human clinical studies have evaluated central nervous system or behavioral effects of exposure
September 2009 2-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 to CO. At ambient-level exposures, healthy adults may be protected against CO-induced
2 neurological impairment owing to compensatory responses including increased cardiac output and
3 cerebral blood flow. However, these compensatory mechanisms are likely impaired among certain
4 potentially susceptible groups including individuals with reduced cardiovascular function.
5 Toxicological studies that were not discussed in the 2000 CO AQCD employed rodent models
6 to show that CO exposure during the in utero or perinatal period can adversely affect adult outcomes
7 including behavior, neuronal myelination, neurotransmitter levels or function, and the auditory
8 system (discussed in Section 5.3). In utero CO exposure, including both intermittent and continuous
9 exposure, has been shown to impair multiple behavioral outcomes in offspring (75-150 ppm). In
10 utero CO exposure (75 and 150 ppm) was associated with significant myelination decrements and
11 neurotransmitter effects (up to 200 ppm). Finally, perinatal CO exposure has been shown to affect
12 the developing auditory system of rodents, inducing permanent changes into adulthood
13 (12.5-100 ppm), some of which appear to be reactive oxygen species mediated. Considering the
14 combined evidence from controlled human exposure and toxicological studies, the evidence is
15 suggestive of a causal relationship between relevant short- and long-term CO exposures
16 and central nervous system effects.
2.5.3.Birth Outcomes and Developmental Effects
17 The most compelling evidence for a CO-induced effect on birth and developmental outcomes
18 is for PTB and cardiac birth defects. These outcomes were not addressed in the 2000 CO AQCD,
19 which included only two studies that examined the effect of ambient CO on LEW. Since then, a
20 number of studies have been conducted looking at varied outcomes, including PTB, birth defects,
21 fetal growth (including LEW), and infant mortality.
22 There is limited epidemiologic evidence that CO during early pregnancy (e.g., first month and
23 first trimester) is associated with an increased risk of PTB. The only U.S. studies to investigate the
24 PTB outcome were conducted in California, and these reported consistent positive associations with
25 CO exposure during early pregnancy when exposures were assigned from monitors within close
26 proximity of the mother's residential address. Additional studies conducted outside of the U.S.
27 provide supportive, though less consistent, evidence of an association between CO concentration and
28 PTB.
29 Very few epidemiologic studies have examined the effects of CO on birth defects. Two of
30 these studies found maternal exposure to CO to be associated with an increased risk of cardiac birth
31 defects. This insult to the heart is coherent with results of human clinical studies demonstrating the
32 heart as a target for CO effects (Section 5.2). Animal toxicological studies provide additional
33 evidence for such an insult to the heart, and reported transient cardiomegaly at birth after continuous
34 in utero CO exposure (60, 125, 250 and 500 ppm CO) and delayed myocardial electrophysiological
September 2009 2-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 maturation (150 ppm CO). Toxicological studies have also shown that continuous in utero CO
2 exposure (250 ppm) induced teratogenicity in rodent offspring in a dose-dependent manner that was
3 further exacerbated by dietary protein (65 ppm CO) or zinc manipulation (500 ppm CO).
4 Toxicological studies of CO exposure over the duration of gestation have shown skeletal alterations
5 (7 h/day, CO 250 ppm) or limb deformities (24 h/day, CO 180 ppm) in prenatally exposed offspring.
6 There is evidence of ambient CO exposure during pregnancy having a negative effect on fetal
7 growth in epidemiologic studies. In general, the reviewed studies, summarized in Figures 5-7
8 through 5-9, reported small reductions in birth weight (ranging -5-20 g). Several studies examined
9 various combinations of birth weight, LEW, and SGA/IUGR and inconsistent results are reported
10 across these metrics. It should be noted that having a measurable, even if small, change in a
11 population is different than having an effect on a subset of susceptible births and increasing the risk
12 of IUGR/LBW/SGA. It is difficult to conclude if CO is related to a small change in birth weight in
13 all births across the population, or a marked effect in some subset of births. Toxicology studies have
14 found associations between CO exposure in laboratory animals and decrements in birth weight
15 (90-600 ppm), as well as reduced prenatal growth (65-500 ppm CO).
16 In general, there is limited epidemiologic evidence that CO is associated with an increased risk
17 of infant mortality during the neonatal or post-neonatal periods. In support of this limited evidence,
18 animal toxicological studies provide some evidence that exogenous CO exposure to pups in utero
19 significantly increased postnatal mortality (7 h/day and 24 h/day, 250 ppm CO; 24 h/day, 90 or
20 180 ppm CO) and prenatal mortality (7 h/day, 250 ppm CO).
21 Evidence exists for additional developmental outcomes which have been examined in
22 toxicological studies, but not epidemiologic or human clinical studies, including behavioral
23 abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes, and
24 changes in the auditory system. Structural aberrations of the cochlea involving neuronal activation
25 (12.5, 25 and 50 ppm CO) and auditory related nerves (25 ppm CO) were seen in pups after neonatal
26 CO exposure. Auditory functional testing using otoacoustic emissions testing (OAE at 50 ppm CO)
27 and 8th cranial nerve action potential (AP) amplitude measurements (12, 25, 50, 100 ppm CO) on
28 rodents exposed perinatally to CO showed that CO-exposed nenonates had auditory decrements at
29 PND22 (OAE and AP) and permanent changes in AP into adulthood (50 ppm CO). Furthermore,
30 exogenous CO may interact with or disrupt the normal physiological roles that endogenous CO plays
31 in the body. There is evidence that CO plays a role in maintaining pregnancy, controlling vascular
32 tone, regulating hormone balance, and sustaining normal follicular maturation.
33 Overall, there is limited, though positive, epidemiologic evidence for a CO-induced effect on
34 PTB and birth defects, and weak evidence for a decrease in birth weight, other measures of fetal
35 growth, and infant mortality. Animal toxicological studies provide support and coherence for these
36 effects. Both hypoxic and non-hypoxic mechanisms have been proposed in the toxicological
September 2009 2-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 literature (Section 5.1), though a clear understanding of the mechanisms underlying reproductive and
2 developmental effects is still lacking. Taking into consideration the positive evidence for some birth
3 and developmental outcomes from epidemiologic studies and the resulting coherence for these
4 associations in animal toxicological studies, the evidence is Suggestive Of 3 Causal relationship
5 between long-term exposures to relevant CO concentrations and developmental effects
6 and birth outcomes.
2.5.4.Respiratory Morbidity
7 New epidemiologic studies, supported by the body of literature summarized in the 2000 CO
8 AQCD (U.S. EPA, 2000, 000907). provide evidence of positive associations between short-term
9 exposure to CO and respiratory-related outcomes including pulmonary function, respiratory
10 symptoms, medication use, hospital admissions, and ED visits. The majority of this literature does
11 not report results of extended analyses to examine the potential influence of model selection, effect
12 modifiers, or confounders on the association between CO and respiratory morbidity. The lack of
13 copollutant models, specifically, has contributed to the inability to disentangle the effects attributed
14 to CO from the larger complex air pollution mix (particularly motor vehicle emissions), and this
15 creates uncertainty in interpreting the results observed in the epidemiologic studies evaluated. As
16 discussed in previous sections, authors often attributed associations reported with CO to the broader
17 mixture of combustion-related pollutants, citing a lack of understanding of the biological
18 mechanisms for CO-related effects. However, animal toxicological studies do provide some evidence
19 that short-term exposure to CO (50-100 ppm) can cause oxidative injury and inflammation and alter
20 pulmonary vascular remodeling. Controlled human exposure studies have not extensively examined
21 the effect of short-term exposure to CO on respiratory morbidity, but a few studies have found
22 inconsistent evidence for CO-induced effects on pulmonary function. Overall, the limited number of
23 controlled human exposure studies that have been conducted prior to and since the 2000 CO AQCD
24 provide very little evidence of any adverse effect of CO on the respiratory system at COHb
25 concentrations relevant to the NAAQS. Although controlled human exposure studies have not
26 provided evidence to support CO-related respiratory health effects, epidemiologic studies show
27 positive associations for CO-induced lung-related outcomes and animal toxicological studies
28 demonstrate the potential for an underlying biological mechanism, which together provide evidence
29 that is suggestive of a causal relationship between short-term exposure to relevant CO
30 concentrations and respiratory morbidity.
31 Currently, only a few studies have been conducted that examine the association between long-
32 term exposure to CO and respiratory morbidity including allergy. Although some studies did observe
33 associations between long-term exposure to CO and respiratory health outcomes, key uncertainties
34 still exist. These uncertainties include: the lack of replication and validation studies to evaluate new
September 2009 2-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 methodologies (i.e., Deletion/Substitution/Addition (DSA) algorithm) that have been used to
2 examine the association between long-term exposure to CO and respiratory health effects; whether
3 the respiratory health effects observed in response to long-term exposure to CO can be explained by
4 the proposed biological mechanisms; and the lack of copollutant analyses to disentangle the
5 respiratory effects associated with CO due to its high correlation with NO2 and other combust'on-
6 related pollutants. Overall, the evidence available is inadequate to Conclude that 3 C3USal
7 relationship exists between long-term exposure to relevant CO concentrations and
8 respiratory morbidity.
2.5.5.Mortality
9 The recently available multicity studies, which consist of larger sample sizes, along with the
10 single-city studies evaluated reported associations that are generally consistent with the results of the
11 studies evaluated in the 2000 CO AQCD (U.S. EPA, 2000, 000907). However, to date the majority
12 of the literature has not conducted extensive analyses to examine the potential influence of model
13 selection, effect modifiers, or confounders on the association between CO and mortality.
14 The multicity studies reported comparable CO mortality risk estimates for total (non-
15 accidental) mortality with the APHEA2 European multicity study showing slightly higher estimates
16 for cardiovascular mortality in single-pollutant models. However, when examining potential
17 confounding by copollutants these studies consistently showed that CO mortality risk estimates were
18 reduced when NO2 was included in the model, but this observation may not be "confounding" in the
19 usual sense in that NO2 may also be an indicator of other pollutants or pollution sources
20 (e.g., traffic).
21 Of the studies evaluated only the APHEA2 study focused specifically on the CO-mortality
22 association, and in the process examined: (1) model sensitivity; (2) the CO-mortality C-R
23 relationship; and (3) potential effect modifiers of CO mortality risk estimates. The sensitivity
24 analysis indicated an approximate 50 - 80% difference in CO risk estimates from a reasonable range
25 of alternative models, which suggests that some model uncertainty likely influences the range of CO
26 mortality risk estimates obtained in the studies evaluated. The examination of the CO-mortality
27 concentration-response relationship found only weak evidence for a CO threshold at 0.5 mg/m3
28 (0.43 ppm). Finally, when examining a variety of city-specific variables to identify potential effect
29 modifiers of the CO-mortality relationship the APHEA2 study found that geographic region
30 explained most of the heterogeneity in CO mortality risk estimates.
31 The results from the single-city studies are generally consistent with the multicity studies in
32 that some evidence of a positive association was found for mortality upon short-term exposure to
33 CO. However, the CO-mortality associations were often, but not always, attenuated when
34 copollutants were included in the regression models. In addition, limited evidence was available to
September 2009 2-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 identify cause-specific mortality outcomes (e.g., cardiovascular causes of death) associated with
2 short-term exposure to CO.
3 The evidence from the recent multi- and single-city studies suggests that an association
4 between short-term exposure to CO and mortality exists, but limited evidence is available to evaluate
5 cause-specific mortality outcomes associated with CO exposure. In addition, the attenuation of CO
6 risk estimates which was often observed in copollutant models contributes to the uncertainty as to
7 whether CO is acting alone or as an indicator for other combustion-related pollutants. Overall, the
8 epidemioiogic evidence is suggestive of a causal relationship between short-term exposure
9 to relevant CO concentrations and mortality.
10 The evaluation of new epidemioiogic studies conducted since the 2000 CO AQCD (U.S. EPA,
11 2000, 000907) that investigated the association between long-term exposure to CO and mortality
12 consistently found null or negative mortality risk estimates. No such studies were discussed in the
13 2000 CO AQCD. The re-analysis of the ACS data by Jerrett et al. (2003, 087380) found no
14 association between long-term exposure to CO and mortality. Similar results were obtained in an
15 updated analysis of the ACS data when using earlier (1980) CO data, but negative associations were
16 found when using more recent (1982-1998) data. These results were further confirmed in an
17 extended analysis of the ACS data. The Women's Health Initiative (WHI) Study also found no
18 association between CO and CVD events (including mortality) using the mortality data from recent
19 years (1994-1998), while the series of Veterans Cohort studies found no association or a negative
20 association between mean annual 95th percentile of hourly CO values and mortality. An additional
21 study was identified that used a cross-sectional study design, which reported results for a study of
22 U.S. counties that are generally consistent with the cohort studies: positive associations between
23 long-term exposure to PM2.5 and SO42" and mortality, and generally negative associations with CO.
24 Overall, the consistent null and negative associations observed across epidemioiogic studies which
25 included cohort populations encompassing potentially susceptible populations (i.e., post-menopausal
26 women and hypertensive men) combined with the lack of evidence for respiratory and
27 cardiovascular morbidity outcomes following long-term exposure to CO; and the absence of a
28 proposed mechanism to explain the progression to mortality following long-term exposure to CO
29 provide supportive evidence that there is not likely to be a causal relationship between long-
30 term exposure to CO and mortality.
September 2009 2-14 DRAFT - DO NOT CITE OR QUOTE
-------
2.6. Policy-Relevant Considerations
2.6.1.Susceptible Populations
1 The examination of populations potentially at greater risk for health effects due to CO
2 exposure is an important consideration in setting NAAQS to provide an adequate margin of safety
3 for both the general population and sensitive populations (see Section 5.7 for a more detailed
4 discussion). During the evaluation of the CO literature, numerous studies were identified that
5 examined whether underlying factors increased the susceptibility of an individual to CO-related
6 health effects. These types of studies were those that included stratified analyses, examined
7 individuals with an underlying health condition, or used animal models of disease.
8 The most important susceptibility characteristic for increased risk due to CO exposure is CAD,
9 also known as coronary heart disease (CHD). As discussed in Section 5.7, there were approximately
10 13.7 million individuals with CHD in the US in 2007. Persons with a normal cardiovascular system
11 can tolerate substantial concentrations of CO, if they vasodilate or increase cardiac output in
12 response to the hypoxia produced by CO. In contrast, individuals unable to vasodilate in response to
13 CO exposure may show evidence of ischemia at low concentrations of COHb. Many of the
14 controlled human exposure studies have focused on individuals with CAD, and several studies have
15 found that controlled exposures to CO resulting in COHb concentrations of 2-6% result in significant
16 decreases in time to onset of exercise-induced angina or ST segment changes in patients with stable
17 angina. Epidemiologic studies found limited evidence for increased hospital admissions for ischemic
18 heart disease (IHD) in individuals with secondary diagnoses of dysrhythmias or congestive heart
19 failure (CHF). This combined evidence from controlled human exposure and epidemiologic studies
20 indicates that individuals with underlying cardiovascular disease, particularly CAD, are a large
21 population that is susceptible to increased health effects in response to exposure to ambient CO.
22 Additional evidence for increased CO-induced cardiovascular effects is provided by toxicological
23 studies that observed altered cardiac outcomes in animal models of cardiovascular disease.
24 Other medical conditions that have been linked to increased susceptibility to CO-induced
25 health effects include COPD, anemia, and diabetes. Individuals with hypoxia resulting from COPD
26 may be particularly sensitive to CO during submaximal exercise typical of normal daily activity. The
27 results available from epidemiologic, controlled human exposure, and toxicological studies provide
28 preliminary evidence that individuals with obstructive lung disease (e.g., COPD patients with
29 underlying hypoxia, asthmatics) may be a potentially susceptible population for increased health
30 effects due to ambient CO exposure. Individuals with various forms of anemia experience lowered
31 hematocrit which decreases blood O2 content; in addition, individuals with hemolytic anemia exhibit
32 increased endogenous CO production rates and COHb levels. Both make individuals with anemia a
September 2009 2-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 potentially susceptible population for ambient CO effects. Diabetics are known to have elevated
2 exhaled CO concentrations indicative of increased endogenous CO production rates. In addition,
3 some recent epidemiologic studies provide preliminary evidence for increased associations between
4 short-term CO exposure and ED visits and hospital admissions for cardiovascular disease (CVD)
5 among diabetics compared to non-diabetics. Increased endogenous CO production in diabetics
6 combined with the limited epidemiologic evidence suggests that diabetics are potentially susceptible
7 to health effects induced by short-term exposure to CO.
8 Aging alters physiological parameters that influence the uptake, distribution, and elimination
9 of CO. The general impact of these changes over an individual's lifetime increases the time required
10 for both loading and elimination of CO from the blood. As noted in the 2000 CO AQCD, changes in
11 metabolism that occur with age, particularly declining maximal oxygen uptake, may make the aging
12 population susceptible to the effects of CO via impaired oxygen delivery to the tissues. Some
13 epidemiologic studies reported increases in IHD or myocardial infarction (MI) hospital admissions
14 among older adults as compared to all age groups or younger adults in response to short-term
15 exposure to CO. Older adults represent a large and growing fraction of the U.S. population, and
16 have a higher prevalence of CAD than the general population; combined with the limited evidence
17 available from epidemiologic studies, this indicates that older adults are a potentially susceptible
18 population for increased health effects due to CO.
19 During gestational exposure, fetal CO pharmacokinetics differ from maternal kinetics, in part
20 because human fetal Hb has a higher CO affinity than adult Hb. At steady-state conditions, fetal
21 COHb is up to 10-15% higher than maternal COHb levels, and these levels are maintained over a
22 longer period since the half-life for fetal CO Hb is approximately twice that of maternal COHb
23 (7.5 h versus 4 h). Some epidemiologic studies reported higher associations between short-term CO
24 exposure and IHD or myocardial infarction (MI) hospital admissions among older adults as
25 compared to all age groups or younger adults. Epidemiologic studies provide some evidence that CO
26 exposure during pregnancy is associated with changes in birth outcomes, including PTB, cardiac
27 birth defects, reductions in birth weight, and infant mortality in the post-neonatal period.
28 Toxicological studies report effects in laboratory animals that lend biological plausibility to
29 outcomes observed in epidemiologic studies, including decrements in birth weight, reduced prenatal
30 growth, and effects on the heart. Toxicological evidence also exists for additional developmental
31 outcomes which have not been examined in epidemiologic or human clinical studies, including
32 behavioral abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes,
33 and changes in the auditory system. This evidence suggests that critical developmental phases may
34 be characterized by enhanced sensitivity to CO exposure.
35 COHb concentrations are generally higher in males than in females, and the COHb half-life is
36 longer in healthy men than in women of the same age. However, women experience fluctuating
September 2009 2-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 COHb levels through the menstrual cycle due to variations in the endogenous CO production rate.
2 Only a limited number of epidemiologic studies have examined gender differences, and found some
3 evidence for larger effects in males compared to females when examining the association between
4 short-term CO exposure and IHD hospital admissions. The limited epidemiologic evidence,
5 combined with the gender-related differences in endogenous CO production, contributes to the
6 inability to conclude whether CO disproportionately affects males or females.
7 Increased altitude induces a number of physiological changes as compensatory mechanisms to
8 counteract the effects of decreased barometric pressure and the resulting altitude-induced hypobaric
9 hypoxia (HH). These changes generally increase both CO uptake and elimination, with increased
10 COHb levels observed in subjects at rest and decreased COHb observed in individuals exposed to
11 CO during exercise. In addition, baseline COHb levels increase due to increased endogenous CO
12 production. A controlled human exposure study observed an additive effect of CO exposure and
13 simulated high altitude on the reduction in time to onset of angina among a group of individuals with
14 CAD. Acclimatization occurs as the length of stay at high altitude increases, indicating that visitors
15 to high altitude locations may have an increased risk of health effects due to CO exposure and
16 represent a potentially susceptible population.
17 Physiological changes associated with exercise tend to increase both uptake and elimination of
18 CO. In a controlled human exposure study, healthy subjects exposed to CO and achieving COHb
19 levels of approximately 5% observed a significant decrement in exercise duration and maximal effort
20 capability. Due to the counterbalancing effects of increased COHb formation and elimination rates, it
21 is unclear whether individuals engaging in light to moderate exercise represent a population
22 potentially susceptible to ambient CO exposure.
23 CO concentrations on and adjacent to heavily traveled roadways are several times higher than
24 concentrations measured at fixed-site monitors not located adjacent to roadways. In addition, studies
25 of commuters have shown that commuting time is an important determinant of CO exposure for
26 those traveling by car, bicycle, public transportation, and walking. Census data indicate that 17.9
27 million occupied homes nationwide (16.1%) are located within approximately 90 m of a freeway,
28 railroad, or airport, and that 5.5 million U.S. workers (5%) commute 60 minutes or more to work in
29 automobiles. This evidence for elevated on-road and near-road CO concentrations combined with
30 residential and commuting data indicates that the large numbers of individuals who spend a
31 substantial amount of time on or near heavily traveled roadways are an important potentially
32 susceptible population for increased health risks due to ambient CO exposure.
33 Endogenous CO production can be altered by medications or other substances, including
34 nicotinic acid, allyl-containing compounds (acetamids and barbiturates), diphenylhydantoin,
35 progesterone, contraceptives, and statins. One epidemiologic study observed an association between
36 short-term CO exposure and an increase in SDNN for CAD patients taking beta blockers; however,
September 2009 2-17 DRAFT - DO NOT CITE OR QUOTE
-------
1 this association did not persist in CAD patients taking beta blockers. Other compounds such as
2 carbon disulfide and sulfur-containing chemicals (parathion and phenylthiourea) increase CO
3 following metabolism by cytochrome p450s. The P450 system may also cause large increases in CO
4 produced from the metabolic degradation of dihalomethanes. Minor sources of endogenous CO
5 include the auto-oxidation of phenols, photo-oxidation of organic compounds, and lipid peroxidation
6 of cell membrane lipids. Taken together, this evidence indicates that individuals ingesting
7 medications and other substances that enhance endogenous or metabolic CO production are a
8 potentially susceptible population for increased health effects due to additional exposure to ambient
9 CO.
10 Overall, the controlled human exposure, epidemiologic, and toxicological studies evaluated in
11 this assessment provide evidence for increased susceptibility among various populations. Medical
12 conditions that increase endogenous CO production rates may also contribute to increased
13 susceptibility to health effects from ambient CO exposure. The level and type of evidence varies
14 depending on the factor being evaluated, with the strongest evidence indicating that individuals with
15 CAD are most susceptible to an increase in CO-induced health effects.
2.6.2.Concentration-Response Relationship
16 Currently, very limited information is available in the human clinical and epidemiologic
17 literature regarding the CO concentration-response (C-R) relationship and the potential existence of
18 a CO threshold. Two human clinical studies described in the 1991 and 2000 CO AQCDs have
19 evaluated the C-R relationship between CO and onset of exercise-induced angina among individuals
20 with CAD. Anderson et al. (1973, 023134) exposed 10 adult men with stable angina (5 smokers and
21 5 non-smokers) for 4 h to CO concentrations of 50 and 100 ppm, which resulted in average COHb
22 levels of 2.9% and 4.5%, respectively. Both exposures significantly decreased the time to onset of
23 exercise-induced angina relative to room air control (1.6% COHb). However, there was no
24 difference in response between the two exposure concentrations of CO. In a much larger study, 63
25 adults with stable angina were exposed for 1 h to 2 concentrations of CO (average exposure
26 concentrations of 117 and 253 ppm) resulting in average COHb concentrations in the range of 2.0-
27 2.4% and 3.9-4.7% (Allred et al., 1989, 013018; Allred et al., 1989, 012697; Allred et al., 1991,
28 011871). Relative to control (average COHb 0.6-0.7%), COHb levels of 2.0-2.4% and 3.9-4.7%
29 were observed to decrease the time required to induce ST-segment changes indicative of myocardial
30 ischemia by 5.1% (p = 0.01) and 12.1% (p < 0.001), respectively. Increasing COHb concentration
31 was similarly shown to decrease the time to onset of exercise-induced angina. As described in Allred
32 et al. (1989), the apparent dose-response relationship observed was further evaluated by regressing
33 the percent change in time to ST-segment change or time to angina on actual COHb concentration
34 (0.2% -5.1%) using the three exposures (air control and two CO exposures) for each subject. This
September 2009 2-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 analysis demonstrated statistically significant decreases in time to angina and ST-segment change of
2 approximately 1.9% and 3.9%, respectively, per 1% increase in COHb concentration. Although the
3 C-R relationship has not been explicitly evaluated in human clinical studies with exposures resulting
4 in COHb concentrations < 2.0%, the findings of Allred et al. provide some evidence of a significant
5 C-R relationship over a range of COHb concentrations relevant to the NAAQS.
6 Two studies in the epidemiologic literature attempted to examine the C-R relationship at the
7 low end of CO concentrations through a threshold analysis. Samoli et al. (2007, 098420) in their
8 examination of the association between short-term exposure to CO and mortality conducted an
9 ancillary analysis to examine the potential presence of a CO threshold. In this analysis the authors
10 compared city-specific models to the threshold model, which consisted of thresholds at 0.5 mg/m3
11 (0.43 ppm) increments. Samoli et al. (2007, 098420) then computed the deviance between the two
12 models and summed the deviances for a given threshold over all cities. While the minimum deviance
13 suggested a potential threshold of 0.43 ppm (the lowest threshold examined), the comparison with
14 the linear no-threshold model indicated weak evidence (p-value > 0.9) for a threshold. However,
15 determining the presence of a threshold at the very low range of CO concentrations (i.e., at
16 0.43 ppm) in this data set is challenging, because, in seven of the 19 European cities examined, the
17 lowest 10% of the CO distribution was at or above 2 mg/m3 (1.74 ppm). By only using the 12 cities
18 in the analysis that had minimum CO concentrations approaching 0.5 mg/m3 (0.43 ppm), a limited
19 number of observations were examined around the threshold of interest, which subsequently
20 contributed to the inability to draw conclusions regarding the potential presence of a threshold with
21 any certainty. In addition to the time-series analyses investigating the association of CO
22 concentrations with hospital admissions due to CVD among Medicare enrollees, Bell et al. (2009,
23 193780) performed subset analyses using datasets that included only days with CO levels below
24 certain specified values, ranging from 1 to 10 ppm (in 1 ppm increments). When these various CO
25 limit values were evaluated, there were positive associations between cardiovascular health effects
26 and CO concentrations at each level investigated in this study, thus providing no evidence for the
27 existence of a threshold. The investigators also estimated an exposure-response curve allowing a
28 non-linear relationship between CO concentration and risk of CVD hospital admissions, and reported
29 no evidence of departure from a linear exposure-response curve.
2.7. Integration of CO Health Effects
30 This section summarizes the main conclusions of this assessment regarding the health effects
31 of CO and the concentrations at which those effects are observed. It also discusses important
32 uncertainties that were considered in interpreting the health effects evidence. The clearest evidence
33 for health effects associated with short-term exposure to CO is provided by studies of cardiovascular
September 2009 2-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 morbidity. The combined health effects evidence supports a likely causal relationship for this
2 outcome. Controlled human exposure studies provide strong evidence of independent effects of CO
3 on cardiac function, with effects being observed in patients with CAD following short-term CO
4 exposures resulting in 2.0-2.4% COHb, the lowest levels tested. Epidemiologic studies of ED visits
5 and hospital admissions for ischemic heart disease report consistent positive associations with
6 additional preliminary evidence for an increase in cardiovascular-related mortality provided by a
7 multicity study. This epidemiologic evidence is coherent with ischemia-related effects observed in
8 controlled human exposure studies. Recent toxicological evidence suggests that other mechanisms
9 involving altered cellular signaling may play a role in cardiovascular disease outcomes following CO
10 exposure.
11 Consistent decreases in time to onset of exercise-induced angina, along with ST-segment
12 changes indicative of myocardial ischemia, were observed in individuals with CAD following
13 controlled CO exposures resulting in COHb concentrations of 2-6%, with no evidence of a threshold
14 at the lowest levels tested. Modeling results described in Chapter 4indicated that increases of ~1%
15 COHb are possible with exposures of several ppm CO, depending on exposure duration and exercise
16 level. Baseline COHb levels are <1% in healthy individuals, with higher endogenous CO production
17 observed in individuals with certain medical conditions. The volunteers who participated in these
18 studies were diagnosed with moderate to severe CAD, although they may not be representative of
19 the most sensitive individuals in the population. Variability in activity patterns and severity of
20 disease combined with daily fluctuations in baseline COHb levels may influence the critical level of
21 increased COHb which leads to adverse cardiovascular effects in a particular individual. In addition,
22 arterial COHb is transiently higher than venous COHb for several minutes following a rapid increase
23 in inhaled CO concentration. Transient increases in ambient CO have the potential to elevate COHb
24 to higher levels in the coronary arteries than in other vascular beds, possibly increasing heart CO
25 levels and cardiovascular symptoms in diseased individuals. Quantification of the magnitude of
26 effects at ambient concentrations from the results of controlled human exposure studies is difficult
27 due to the gap between ambient concentrations and the higher concentrations used in these studies
28 (i.e., experimental studies have not been conducted at levels within the range of current maximum
29 ambient concentrations).
30 Epidemiologic studies consistently show associations between ambient CO concentrations and
31 cardiovascular endpoints other than stroke, particularly hospitalizations and emergency department
32 visits for ischemic heart disease, myocardial infarction, and angina. These effects are robust to
33 adjustment for copollutants. Figure 2-1 presents health effect estimates from U.S. and Canadian
34 studies of short-term CO exposure and CVD hospitalizations, along with mean and 99th percentile
35 concentrations during the study periods. Table 2-2 summarizes the range of mean and 99th percentile
36 concentrations observed in the studies presented in Figure 2-1. This evidence for ischemia-related
September 2009 2-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 outcomes is coherent with effects observed in controlled human exposure studies, although
2 uncertainty regarding the plausibility of reduced O2 delivery to tissues following exposure to
3 ambient CO concentrations contributes to the uncertainty in quantitative interpretation of effect
4 estimates.
Study, Location
Mean
(99th %*) CO
Concentratio
in ppm
Avg Tim
,u,
Effect Estimate and 95% Confidence Interval
Non-stroke CVD Endpoints
Symonsetal. (2006, 0912581: Baltimore, MD
Szyszkowicz (2007, 1937931: Montreal Canada
Szyszkowicz (2007, 1937931: Montreal Canada
Wellenius et al. (2005, 0874831: Pittsburgh, PA
Bell et al. (2009, 1937801: 126 U.S. Counties
Fung et al. (2005, 0743221: Wndsor, Canada
Fung et al. (2005, 0743221: Wndsor, Canada
Metzger et al. (2004, 0442221: Atlanta, GA
Metzger et al. (2004, 0442221: Atlanta, GA
Tolbert et al. (2007, 0903161: Atlanta, GA
Peel et al. (2007, 0904421: Atlanta, GA
Peel et al. (2007, 0904421: Atlanta, GA
Koken et al. (2003, 0494661: Denver, CO
Mann et al. (2002, 0367231: California, US
Mann et al. (2002, 0367231: California, US
Mann et al. (2002, 0367231: California, US
Linn et al. (2000, 0028391: Los Angeles, CA
Linn et al. (2000, 0028391: Los Angeles, CA
Linn et al. (2000, 0028391: Los Angeles, CA
0.4 (2.3)
0.5
0.5
1.03(3.3-8.9)
1.3(1.2-22.1)
1.3
1.3
1.5(5.5-5.9)
1.5(5.5-5.9)
1.6(5.5-5.9)
1.8(5.5-5.9)
1.8(5.5-5.9)
0.9 (2.5-3.9)
2.07(1.3-15.9)
2.07(1.3-15.9)
2.07(1.3-15.9)
1.5(1.1-8.3)
1.5(1.1-8.3)
1.5(1.1-8.3)
8h max
24 h avg
24 h avg
1 hmax
1 hmax
1 h max
1 hmax
1 hmax
1 hmax
1 hmax
1 hmax
1 hmax
24 h avg
8 hmax
8 hmax
8 hmax
24 h avg
24 h avg
24 h avg
0-3
0
0
0
0
0-2
0-2
0-2
0-2
0-2
0-2
0-2
3
0-3
0-3
0-3
0
0
0
_^
unr
IHD
CVD
CVD-
IHD
CHF-
CVD
IHD
CHF
CHF
IHD
IHD
IHD
Ml
CHF
CVD
Villeneuve et al. (2006, 0901911: Edmonton, Canada
Villeneuve et al (2006 090191V Edmonton Canada
Villeneuve et al. (2006, 0901911: Edmonton, Canada
Villeneuve et al. (2006, 0901911: Edmonton, Canada
Wellenius et al. (2005, 088685); Multicity, US
Linn et al. (2000, 0028391: Los Angeles, US
0.8
0.8
0.8
0.8
1.02(1.2-7.1)
1.5(1.1-8.3)
24 h avg
24 h avg
24 h avg
24 h avg
24 h avg
24 h avg
0-2
0-2
0-2
0-2
0-2
0
* Range of 99th percentile concentrations during study period across all moniotrs presented when multiple monitors were
— •-
IHD *
• 65+ yrs
CHF— •
• 65+ yrs
-• 65+ yrs
-*—
•* —
-
-• —
-—
-
-•- sCHF
-•-sARR
—
-
-*-
Stroke
Cer Isc Stroke 65+ yrs
Hern Stroke 65+ "r^
-m- Isch Stroke, 65+ yrs
- Hem Stroke
-*- Isch Stroke
1 ill
0.9 1.0 1.1 12 1.3
available in the study location
Figure 2-1 Excess risk estimates from epidemiologic studies of short-term CO exposure and
CVD hospitalizations along with mean and 99th percentile CO concentrations.
September 2009
2-21
DRAFT - DO NOT CITE OR QUOTE
-------
Table 2-2 Range of mean and 99th percentile concentrations (ppm) in US and Canadian studies of
short-term CO exposure and CVD hospitalizations.
Metric 1-h daily max 8-h daily max 24-havg
Mean 1.03-1.8 0.4-2.07 0.5-1.5
99th percentile 1.2-22.1 1.3-15.9 1.1-8.3
1 Additional studies provide evidence for associations between CO exposure and other health
2 outcomes, including central nervous system effects, birth outcomes and developmental effects,
3 respiratory effects, and mortality. Although inconsistent results were reported in controlled human
4 exposure studies on neural and behavioral effects, toxicological studies in rodents found that
5 perinatal exposure to CO can have a range of effects on the adult nervous system. This combined
6 evidence is suggestive of a causal relationship between both short- and long-term CO exposure and
7 central nervous system effects. Differences in fetal pharmacokinetics from those of the mother result
8 in fetal COHb levels that are up to 10-15% higher than maternal COHb levels. Epidemiologic
9 studies provide some evidence that CO exposure during pregnancy is associated with changes in
10 birth outcomes, including increased risk of PTB, cardiac birth defects, small reductions in birth
11 weight, and infant mortality in the post-neonatal period. This evidence, in conjunction with
12 developmental effects observed in toxicological studies, is suggestive of a causal relationship
13 between long-term exposure to CO and birth and developmental effects.
14 Evidence regarding the effect of short-term exposure to CO on respiratory morbidity is
15 suggestive of a causal relationship, based on associations observed in epidemiologic studies and
16 animal toxicological studies which indicate the potential for an underlying biological mechanism,
17 while the evidence on long-term exposure and respiratory morbidity is inadequate to infer the
18 presence of a causal relationship.
19 An evaluation of epidemiologic studies that examined the effect of short-term exposure to CO
20 on mortality provides evidence that is suggestive of a causal relationship. Epidemiologic studies that
21 examined mortality and long-term exposure to CO reported consistent null associations, which,
22 combined with the lack of respiratory and cardiovascular morbidity or a proposed biological
23 mechanism for mortality following long-term exposure, indicate that there is not likely to be a causal
24 relationship between long-term exposure to CO and mortality.
25 Issues such as exposure error and isolation of the independent effect of CO as a component of
26 a complex air pollutant mixture contribute to uncertainty in interpreting the results of epidemiologic
27 studies. Studies published since the 2000 CO AQCD have provided insight regarding the nature and
28 magnitude of these uncertainties. Exposures in near-road and on-road microenvironments are likely
29 to be higher than concentrations measured at community-oriented regulatory monitors, which may
September 2009 2-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 result in over- or under-estimation of the magnitude of ambient exposure for some individuals.
2 Individuals who are susceptible to CO-induced health effects, such as those with coronary artery
3 disease, may be at additional risk when experiencing elevated on-road CO concentrations. However,
4 as discussed in Section 2.3 and in more detail in Section 3.6, spatial variability in absolute
5 concentration will not introduce error into time-series epidemiologic studies if the concentrations are
6 correlated in time. A recent study by Sarnat et al. (2009, 180084) found that associations between
7 CO and cardiovascular ED visits were similar when based on different monitors within an urban
8 center, regardless of monitor location or distance to population, while an association was not
9 observed when using a rural monitor outside the urban area. This may have been related to the
10 similarity of driving patterns and peak rush hour times in the urban center as compared to the area
11 around the rural monitor, where the temporal driving patterns were different. Simulations of ambient
12 and nonambient exposures to a non-reactive pollutant indicated that nonambient exposure has no
13 effect on the association between ambient exposure and health outcomes for the case where ambient
14 and nonambient concentrations are independent, although variability is introduced. Nonambient
15 exposure to CO is not expected to be temporally correlated with ambient CO concentrations, and
16 therefore nonambient CO will not act as a confounder in epidemiologic associations with ambient
17 CO. Exposure error is not likely to affect the magnitude of the population-averaged effect estimates
18 observed in epidemiologic studies, although it would tend to widen the confidence intervals.
19 Epidemiologic studies consider the effects of CO as a component of a complex mixture of air
20 pollutants that varies across space and time, with moderate to high correlations observed between
21 CO concentrations and those of other combustion-related pollutants. On-road vehicle exhaust
22 emissions are a nearly ubiquitous source of combustion pollutant mixtures that include CO, NO2,
23 and PM2.5, and these emissions are the most important contributor to ambient CO in near-road
24 locations. Correlations between CO and NO2 reported in epidemiologic studies of short-term
25 exposure to CO generally ranged from 0.3 to 0.86, with correlations reported in US studies ranging
26 from 0.55-0.86. Correlations between CO and PM2.5 reported in all studies ranged from 0.17 to 0.74,
27 with correlations in US studies ranging from 0.43-0.62. This complicates the quantitative
28 interpretation of effect estimates in these studies to apportion the relative extent to which CO at
29 ambient concentrations is independently associated with cardiovascular or other effects, and the
30 extent to which CO acts as a marker for the effects of another combustion-related pollutant or mix of
31 pollutants.
32 As summarized in Tolbert et al. (2007, 090316). when toxicological or controlled human
33 exposure studies of two correlated pollutants provide evidence that each exerts an independent health
34 effect, two-pollutant models may be appropriate to adjust the effect estimate for each pollutant for
35 confounding by the other pollutant. PM25 and NO2 have each been linked to cardiovascular health
36 effects in epidemiologic studies. In two-pollutant models in which one of the pollutants is linked to
September 2009 2-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 the measured outcome, and the other is a surrogate for the first pollutant, the copollutant model can
2 help identify which is the better predictor of the effect, particularly if the etiologically linked
3 pollutant is measured with more error than the second pollutant. Uncertainty is introduced in the size
4 of the effect estimate and the portion of the effect size represented by each of the coefficients in the
5 model by correlation between the two pollutants and by differential exposure measurement error.
6 Since the spatial variability of CO is a larger contributor to measurement error than for other more
7 homogenously distributed pollutants such as PM2.5, robustness of CO effect estimates indicates that
8 CO is the better predictor of effects in copollutant models. Although this complicates quantitative
9 interpretation of the effect estimates reported in epidemiologic studies, the epidemiologic evidence
10 for cardiovascular morbidity summarized in this assessment indicates that CO associations generally
11 remain robust in copollutant models (see Figure 5-6 and Figure 5-7), which, combined with the
12 consistency of effects observed across studies, the coherence of epidemiologic health outcomes with
13 effects observed in controlled human exposure studies, and the emerging evidence on the potential
14 role for cell signaling effects at low tissue CO concentrations, supports an independent effect of
15 short-term CO exposure on cardiovascular morbidity. This combined evidence supports a
16 determination that the relationship between CO and cardiovascular morbidity is likely causal, while
17 still recognizing that CO is a component of a mixture of combustion-related pollutants.
18 Evidence from controlled human exposure and epidemiologic studies indicates that individuals
19 with underlying cardiovascular disease, specifically CAD, are an important susceptible population at
20 increased risk of health effects due to ambient CO. Potentially susceptible populations include those
21 with other underlying diseases, including anemia, obstructive lung disease, or diabetes; older adults
22 and fetuses during critical phases of development; commuters and those living near heavily traveled
23 roadways; visitors to high-altitude locations; and individuals ingesting medications and other
24 substances that enhance endogenous or metabolic CO production. Limited evidence is available from
25 controlled human exposure studies of CAD patients indicating a statistically significant inverse
26 relationship between COHb concentration and time to ST segment change or time to exercise-
27 induced angina, although the C-R relationship has not been explicitly evaluated with controlled
28 exposures resulting in COHb concentrations below 2.0%. Epidemiologic analyses investigating the
29 exposure-response relationship for mortality and cardiovascular morbidity did not find evidence for
30 a departure from linearity or a threshold for CO effects.
31 The new evidence reviewed in this ISA builds upon the health effects evidence summarized in
32 the 2000 CO AQCD, with many new epidemiologic studies adding to the body of evidence showing
33 associations between acute cardiovascular effects and CO measured at ambient monitors. Controlled
34 human exposure studies reviewed both in this ISA and the 2000 CO AQCD show definitive evidence
35 of cardiovascular effects among individuals with CAD following short-term CO exposure resulting
36 in COHb concentrations as low as 2.0-2.4%. Emerging toxicological evidence points to the potential
September 2009 2-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 role for CO in modes of action not directly related to COHb's role in O2 delivery. In evaluating the
2 several epidemiologic studies available at the time that reported associations between ambient CO
3 and cardiovascular effects, the 2000 CO AQCD considered those findings to be inconclusive for
4 multiple reasons, including questions regarding the consistency of the results among studies; the
5 ability of community fixed-site monitors to represent spatially variable ambient CO concentrations
6 and personal exposures; the small expected increase in COHb due to ambient CO concentrations; the
7 lack of biological plausibility for health effects to occur at such COHb levels, even in diseased
8 individuals; the potentially greater impact of non-ambient exposure on COHb; and the possibility
9 that ambient CO is serving as a surrogate for a mixture of combustion-related pollutants. Some of
10 these uncertainties remain and complicate the quantitative interpretation of the epidemiologic
11 findings, particularly regarding the biological plausibility of health effects occurring at COHb levels
12 resulting from exposures to ambient CO concentrations. New research summarized in this
13 assessment reduces several of the other uncertainties noted in the 2000 CO AQCD, and demonstrates
14 the lack of influence of nonambient exposure on effect estimates in epidemiologic studies, the
15 consistency of epidemiologic study results, their robustness in copollutant models, and the coherence
16 of ischemia-related outcomes with evidence from controlled human exposure studies. This consistent
17 and coherent evidence from epidemiologic and human clinical studies, along with biological
18 plausibility provided by CO's role in limiting O2 availability, is sufficient to conclude that a causal
19 relationship is likely to exist between relevant short-term CO exposures and cardiovascular
20 morbidity.
September 2009 2-25 DRAFT - DO NOT CITE OR QUOTE
-------
References
Allied EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Hayes D; Pagano M; Selvester RH; Walden
SM; Warren J. (1989). Acute effects of carbon monoxide exposure on individuals with coronary artery disease.
Health Effects Institute. Boston, MA. 012697
Allred EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; Selvester RH; Walden SM;
Warren J. (1989). Short-term effects of carbon monoxide exposure on the exercise performance of subjects with
coronary artery disease. N Engl J Med, 321: 1426-1432. 013018
Allred EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; Selvester RH; Walden SM;
Warren J. (1991). Effects of carbon monoxide onmyocardial ischemia. Environ Health Perspect, 91: 89-132.
011871
Anderson EW; Andelman RJ; Strauch JM; Fortuin NJ; Knelson JH. (1973). Effect of low-level carbon monoxide exposure
on onset and duration of angina pectoris: a study in ten patients with ischemic heart disease. Ann Intern Med, 79:
46-50. 023134
Bell ML; Peng RD; Dominici F; Samet JM. (2009). Emergency admissions for cardiovascular disease and ambient levels
of carbon monoxide: Results for 126 U.S. urban counties, 1999-2005. Circulation, xx: xxx-xxx. 193780
Fung KY; Luginaah I; Gorey KM; Webster G (2005). Air pollution and daily hospital admissions for cardiovascular
diseases in Windsor, Ontario. Can J Public Health, 96: 29-33. 074322
IPCC. (2001). IPCC Third Assessment Report (TAR): Climate Change 2001: Working Group I: The Scientific Basis.
Cambridge, U.K, and New York, NY: Intergovernmental Panel on Climate Change, Cambridge University Press.
156587
IPCC. (2007). IPCC Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical
Science Basis. Cambridge, UK and New York, NY: Intergovernmental Panel on Climate Change, Cambridge
University Press. 092765
IPCC; Forster P; Ramaswamy V; Artaxo P; Berntsen T; Betts R; Fahey DW; Haywood J; Lean J; Lowe DC; Myhre G;
Nganga J; Prinn R; Raga G; Schultz M; Van Dorland R. (2007). Changes in atmospheric constituents and in
radiative forcing, Chapter 2. In Solomon S, Qin D; Manning M; Chen Z; Marquis M; Averyt KB; Tignor M; Miller
HL (Ed.),IPCC Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical
Science Basis, (pp. 129-234). Cambridge, U.K. and New York, NY: Intergovernmental Panel on Climate Change;
Cambridge University Press. 092936
IPCC; Ramaswamy V; Boucher O; Haigh J; Hauglustaine D; Haywood J; Myhre G; Nakajima T; Shi G; Solomon S.
(2001). Radiative forcing of climate change, Chapter 6. In Houghton JT; Ding Y; Griggs DJ; Noguer M; van der
Linden PJ; Da X; Maskell K; Johnson CA (Ed.),IPCC Third Assessment Report (TAR): Climate Change 2001:
Working Group I: The Scientific Basis (pp. 349-416). Cambridge, U.K. and New York, NY: Intergovernmental
Panel on Climate Change; Cambridge University Press. 156899
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of the air
pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health A, 66: 1735-1777.
087380
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver. Environ Health Perspect, 111: 1312-1317. 049466
Linn WS; Szlachcic Y; Gong H Jr; Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions in
metropolitan Los Angeles. Environ Health Perspect, 108: 427-434. 002839
Mann JK; Tager IB; Lurmann F; Segal M; Quesenberry CP Jr; Lugg MM; Shan J; Van den Eeden SK. (2002). Air pollution
and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia. Environ
Health Perspect, 110: 1247-1252. 036723
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 2-26 DRAFT - DO NOT CITE OR QUOTE
-------
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Samoli E; Touloumi G; Schwartz J; Anderson HR; Schindler C; Forsberg B; Vigotti MA; Vonk J; Kosnik M; Skorkovsky J;
Katsouyanni K. (2007). Short-term effects of carbon monoxide on mortality: An analysis within the APHEA
project. Environ Health Perspect, 115: 1578-1583. 098420
Sarnat SE; Klein M; Sarnat JA; Flanders WD; Waller LA; Mulholland JA; Russell AG; Tolbert PE. (2009). An examination
of exposure measurement error from air pollutant spatialvariability in time-series studies. J Expo Sci Environ
Epidemiol, In Press: 1-12. 180084
Sheppard L; Slaughter JC; Schildcrout J; Liu L-JS; Lumley T. (2005). Exposure and measurement contributions to
estimates of acute air pollution effects. J Expo Sci Environ Epidemiol, 15: 366-376. 079176
Sinha A; Toumi R. (1996). A comparison of climate forcings due to chloroflurocarbons and carbon monoxide. Geophys Res
Lett, 23:65-68. J_93747
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Szyszkowicz M. (2007). Air pollution and emergency department visits for ischemic heart disease in Montreal, Canada. Int
J Occup Med Environ Health, 20: 167-173. 193793
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 600/P-
99/001F. 000907
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and hemorrhagic
stroke among medicare beneficiaries. Stroke, 36: 2549-2553. 088685
Zeger SL; Thomas D; Dominici F; Samet JM; Schwartz J; Dockery D; Cohen A. (2000). Exposure measurement error in
time-series studies of air pollution: concepts and consequences. Environ Health Perspect, 108: 419-426. 001949
September 2009 2-27 DRAFT - DO NOT CITE OR QUOTE
-------
Chapter 3. Source to Exposure
3.1. Introduction
1 This chapter reviews concepts and findings in atmospheric sciences and exposure assessment
2 that provide a foundation for the detailed presentation of evidence of CO-related health effects in
3 subsequent chapters. Section 3.2 provides an overview of the sources of CO and examples of their
4 spatial distribution. Atmospheric chemistry involved in the production and removal of CO by
5 oxidation processes is discussed in Section 3.3 along with a description of climate forcing caused
6 directly and indirectly by CO. Descriptions of CO measurement methods, monitor siting
7 requirements, and monitor locations are presented in Section 3.4. Ambient CO concentrations and
8 their spatial and temporal variability are characterized in Section 3.5. The background concentrations
9 of CO useful for risk and policy assessments informing decisions about the NAAQS, referred to as
10 policy-relevant background (PRB) concentrations, are also presented in Section 3.5. For this
11 document, PRB concentrations include contributions from natural sources everywhere in the world
12 and from anthropogenic sources outside the U.S., Canada, and Mexico. Factors related to human
13 exposure to ambient CO, and their implications for epidemiologic studies, are discussed in
14 Section 3.6. Finally, a summary and conclusions of the chapter are presented in Section 3.7.
3.2. Sources and Emissions of CO
15 CO is a colorless, odorless, tasteless gas consisting of one carbon atom covalentry bonded to
16 one oxygen atom; its molar mass is 28.0101 g/mol. CO is formed primarily by incomplete
17 combustion of carbon-containing fuels and photochemical reactions in the atmosphere. In general,
18 any increase in fuel O2 content, burn temperature, or mixing time in the combustion zone will tend
19 to decrease production of CO relative to CO2.
20 CO emissions from large fossil-fueled power plants are typically very low since the boilers at
21 these plants are tuned for highly efficient combustion with the lowest possible fuel consumption.
22 Additionally, by allowing time for the furnace flue gases to mix with air and be oxidized by OH to
23 CO2 in the hot gas stream before the OH concentrations drop as the flue gases cool, the CO-to-CO2
24 ratio in these emission is shifted toward CO2.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 3-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 Internal combustion engines used in mobile sources, by contrast, have more widely varying
2 operating conditions and thus higher and more varying rates of CO formation. Moreover, the
3 gasoline-powered spark ignition engines that predominate in light-duty on-road vehicles have higher
4 uncontrolled CO emission rates than other combustion sources because they typically operate closer
5 to the stoichiometric air-to-fuel ratio, have relatively short residence times at peak combustion
6 temperatures, and have very rapid cooling of cylinder exhaust gases. By contrast, the diesel-powered
7 engines which predominate in heavy-duty on-road vehicles and in off-road and non-road fixed
8 combustion sources have much lower engine-out CO emission than do the spark-ignition engines
9 because the diesels typically operate at very high air-to-fuel ratios which promotes mixing oxygen
10 and the fuel, thus improving carbon burn.
11 Figure 3-1 lists CO emissions totals in tons segregated by individual source sectors in the U.S.
12 for 2002, which is the most recent publicly available CO emissions data meeting EPA's data quality
13 assurance objectives. In the U.S., CO emissions data are tracked in the National Emissions Inventory
14 (U.S. EPA, 2006, 157070). a composite of data from various sources including industries and state,
15 tribal, and local air agencies. NEI data are collected for all states, the District of Columbia, the U.S.
16 territories of Puerto Rico and Virgin Islands, and some of the territories of federally recognized
17 American Indian nations. Different data sources use different data collection methods, most of which
18 are based on empirical estimates and engineering calculations rather than measurements. Most fuel
19 combustion and industrial sources, for example, estimate their CO emissions using EPA-approved
20 emission factors, as do on-road and non-road mobile source emitters where models (MOBILE6,
21 MOVES, NONROAD) are available to calculate inventories (U.S. EPA, 2006, 157070). Although
22 these estimates are generated using well-established approaches, uncertainties inhere in the emission
23 factors and models used to represent sources for which emissions have not been directly measured.
24 Nationally, on-road mobile sources in the NEI constituted more than half of total CO
25 emissions in 2002, or ~63 MT of-109 MT total. For this reason, high concentrations of CO can
26 often occur in areas of heavy traffic. In metropolitan areas in the U.S., for example, as much as 75%
27 of all CO emissions came from on-road vehicle exhaust in the 2002 NEI (U. S. EPA, 2006, 157070).
28 The majority of these on-road CO emissions derive from gasoline-powered vehicles since the O2
29 content, pressure, and temperature required for diesel fuel ignition result in much less CO
30 production. When the emissions from incomplete combustion of fuels powering non-road mobile
31 sources were included, all mobile sources accounted for -80% of total CO emissions in the U.S. in
32 2002; see Figure 3-1.
September 2009 3-2 DRAFT - DO NOT CITE OR QUOTE
-------
Nationwide Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
1 62,957,908
14,520,530
22,414,896
^2,704,197
] 1,499,367
] 652,314
^| 2,414,055
^2,018,496
33,786
3,294
20,000,000
40,000,000
Emissions (tons)
60,000,000
80,000,000
Source: U.S. EPA (2006, 157070)
Figure 3-1 CO emissions (tons) in the U.S. by source sector in 2002.
1 Figure 3-2 shows present and historical CO emissions from the traditionally inventoried
2 anthropogenic source categories: (1) fuel combustion, which includes emissions from coal-, gas-,
3 and oil-fired power plants and industrial, commercial, and institutional sources, as well as residential
4 heaters (e.g., wood-burning stoves) and boilers; (2) industrial processes, which includes chemical
5 production, petroleum refining, metals production, and industrial processes other than fuel
6 combustion; (3) on-road vehicles, which includes cars, trucks, buses, and motorcycles; and (4) non-
7 road vehicles and engines, such as farm and construction equipment, lawnmowers, chainsaws, boats,
8 ships, snowmobiles, aircraft, locomotive, and others. Using these NEI data, trends in the national CO
9 emissions can be computed and compared over time. So, for example, the national-scale estimated
10 anthropogenic CO emissions decreased 35% between 1990 and 2002; see Figure 3-2. The trend plot
11 in Figure 3-2 demonstrates that controls in the on-road vehicle sector have produced nearly all the
12 national-level CO reductions since 1990. (Data are presented here for 1990 and from 1996-2002
13 because only 1990 data have been updated to be comparable to the more recent inventories made
14 since 1996.)
September 2009
3-3
DRAFT - DO NOT CITE OR QUOTE
-------
160
^140
"c/r
g 120
••—•
1 100
1, 80
CO
I 60
'c/3
•| 40
m 20
0
Fuel combustion
Other industrial processes
On-road vehicles
Nonroad vehicles and engines
•go
'96 '97 '98 '99 '00 '01 '02
Year
Source: U.S. EPA (2008, 1570761
Figure 3-2 Trends in anthropogenic CO emissions (MT) in the U.S. by source category for
1990 and 1996-2002.
1 With the exception of this downward trend resulting from emissions controls, anthropogenic
2 CO emissions demonstrate less interannual variability than biogenic emissions (Bergamaschi et al.,
3 2000, 192377). Several recent reports using both ambient concentrations and fuel-based emissions
4 estimates have explored this annual-to-decadal emissions decrease in anthropogenic CO in finer
5 detail; they include, Harley et al. (2001, 193922: 2005, 088154). Parrish et al. (2002, 052472).
6 Parrish (2006, 090352). Pollack et al. (2004, 184461). and Mobley et al. (2005, 194008). The
7 consistent conclusion from those investigations has been that annual average U.S. on-road vehicle
8 CO emissions have decreased at a rate of ~5% per year since the early 1990s. Additional analyses by
9 Harley et al. (2005, 088154) and Parrish (2006, 090352) were also consistent with the suggestion in
10 Pollack et al. (2004, 184461) that the EPA MOBILE6 vehicle emissions model
11 (http://www.epa. gov/otaq/m6.htm) now overestimates vehicle CO emissions by a factor of ~2.
12 Parrish's (2006, 090352) findings that the measured trends of CO and NOX concentrations from
13 mobile sources in the U.S. indicated that modeled CO emission estimates were substantially too high
14 were subsequently confirmed by field measurements by Bishop and Stedman (2008, 194670).
15 Improvements in emissions technologies not correctly represented in MOBILE emission
16 models have been suggested as one cause for this discrepancy. For example, Pokharel et al. (2002,
17 052473; 2003, 053740) demonstrated substantial decrements in the CO fraction of tailpipe exhaust in
18 several U.S. cities and Burgard et al. (2006, 193222) documented improvements in emission from
19 heavy-duty on-road diesel engines. It appears likely that some of the largest errors in the MOBILE
20 models may be addressed when the successor model, MOVES, is released in final form; see
September 2009
3-4
DRAFT - DO NOT CITE OR QUOTE
-------
1 http://www.epa.gov/oms/models/moves/420b09008.pdf. The public schedule lists a release of
2 MOVES in final form by the end of calendar year 2009.
3 Estimates of non-anthropogenic CO emissions are made using the Biogenic Emissions
4 Inventory System (BEIS) model with data from the Biogenic Emissions Landcover Database
5 (BELD) and annual meteorological data; see http://www.epa.gov/ttnchiel/emch/biogenic. National
6 biogenic emissions, excluding fires, were estimated to contribute ~5% of total CO emissions from all
7 sources in 2002; fires in 2002 added another 13%, or -14.5 MT, to the national CO emissions total.
8 Geogenic emissions of CO, also included in this inventory, include volcanic gases released from
9 molten rock in the earth's mantle. Mixing ratios of dissolved CO in this rock vary in a range from
10 0.01 to 2% as a function of the rock stratum surrounding the volcano and other geologic conditions.
11 This high variability and infrequent though often violent release mean geogenic CO measurements
12 are very difficult to make with precision, though on non-local scales the magnitude of their
13 contribution is small relative to anthropogenic sources. Photodecomposition of organic matter in
14 oceans, rivers, lakes, and other surface waters, and from soil surfaces also releases CO (Goldstein
15 and Galbally, 2007, 193247). However, soils can act as a CO source or a sink depending on soil
16 moisture, UV flux reaching the soil surface, and soil temperature (Conrad and Seiler, 1985, 029520).
17 Soil uptake of CO is driven by anaerobic bacteria (Inman et al, 1971, 010972). Emissions of CO
18 from soils appear to occur by abiotic processes, such as thermodecomposition or
19 photodecomposition of organic matter. In general, warm and moist conditions found in most soils
20 favor CO uptake, whereas hot and dry conditions found in deserts and some savannas favor the
21 release of CO (King, 1999, 002828). An extensive measurement and modeling study by Hudman
22 et al. (2008, 191253) established that the NEI CO emissions estimate for the eastern third of the
23 CONUS could be overestimated by 60% in summer. Using aircraft measurements from the ICARTT
24 campaign (Fehsenfeld et al., 2006, 190531) and the GEOS-Chem model (Bey et al., 2001, 051218)
25 (configured as described by Hudman et al.(2007, 089474)). Hudman et al. (2008, 191253)
26 determined that anthropogenic CO emissions over eastern North America between July and August
27 2004 were 6.4 Tg CO including 4.6 Tg from direct emissions and 1.8 from oxidation of
28 anthropogenic VOCs, and that the biogenic CO from oxidation of isoprene and other biogenic VOCs
29 was 8.3 Tg; see Figure 3-3 and Figure 3-4 taken from Hudman et al. (2008, 191253).
September 2009 3-5 DRAFT - DO NOT CITE OR QUOTE
-------
Observed
Simulated (NEI99 emissions)
Simulated
i NEI99 emissions reduced by 60%)
50°N
40°N
30°N
125 150 175 200
Source: Hudman et al. (2008,191253)
Figure 3-3 Mean CO concentrations in the boundary layer (0-1.5 km altitude) during the
ICARTT campaign (July 1-August 15, 2004) (left). Observations averaged over the
2° x 2.5° GEOS-Chem model grid are compared to model results using the
(middle) U.S. EPA NEI emissions estimates from 1999 and (right) anthropogenic
CO emissions reduced by 60%. Model results are samples along the flight
tracksat the time of the flights.
350
300
250
200
15°
100
50
Observed
Simulated (NEI 99 emissions)
Simulated (NBI 99 emissions reduced by 60%)
180
190
200 210
Day of the Year
220
230
Source: Hudman et al. (2008,1912531
Figure 3-4 Surface air CO concentrations at Chebogue Point during the ICARTT campaign.
Observations (black) are compared to model results using the 1999 NEI
anthropogenic emissions (green) and with these CO emissions reduced by 60%
(blue). Yellow bands are periods of U.S. outflow diagnosed by Millet et al. (2006,
195106). Overestimate near day 200 is due to model misplacement of a large
Alaskan/Canadian biomass burning plume.
1 Biomass burning consists of wildfires and the intentional burning of vegetation to clear new
2 land for agriculture and population resettlement; to control the growth of unwanted plants on pasture
September 2009
3-6
DRAFT - DO NOT CITE OR QUOTE
-------
1 land; to manage forest resources with prescribed burning; to dispose of agricultural and domestic
2 waste; and as fuel for cooking, heating, and water sterilization. Globally, most wildfires may be
3 ignited directly as the result of human activities leaving only 10-30% initiated by lightning (Andreae,
4 1991, 078147). However, because fire management practices suppress natural wildfires, the buildup
5 of fire fuels increases the susceptibility of forests to more severe but less frequent fires in the future.
6 Thus there is considerable uncertainty in attributing the fraction of wildfire emissions to human
7 activities because the emissions from naturally occurring fires that would have been present in the
8 absence of fire suppression practices are not known.
9 Biomass burning also exhibits strong seasonality and interannual variability (van der Werf et
10 al., 2006, 157084). with most biomass burned during the local dry season. This is true for both
11 prescribed burns and wildfire. The unusually warm and dry weather in central Alaska and western
12 Yukon in the summer of 2004, for example, contributed to the burning of 11 million acres there.
13 These fires, the largest on record for this region, produced CO emissions easily tracked by the
14 Measurement of Pollution in the Troposphere (MOPITT) instrument on NASA's Terra satellite; see
15 Figure 3-5. The high CO concentration measured by MOPITT coincided with the surface location of
16 fires tracked using aerosol plumes identified by the Moderate Resolution Imaging Spectroradiometer
17 (MODIS) also on Terra. Subsequent modeling by Pfister et al. (2005, 093009) showed that the CO
18 contribution from these fires in July 2004 was 30 (± 5) teragrams (Tg) that summer, or in the range
19 of the total U.S. anthropogenic CO emissions during the same time.
20 The smoldering phase of combustion yields higher CO emissions than the flaming phase.
21 Using controlled combustion chamber experiments Lobert et al. (1991, 029473) found that with a
22 wide variety of vegetation types, on average, 84% of the CO from biomass fires was produced
23 during the smoldering phase and 16% during the flaming phase of combustion.
24 CO emissions data for EPA's ten administrative Regions in the U.S. depicted in Figure 3-6
25 show a more nuanced view of the national concentrations and trends described just above. Net
26 anthropogenic CO emissions were estimated to have declined in all EPA Regions between 1990 and
27 2002 with the largest decrease (10.8 MT) occurring in Region 9 and the smallest (1.3 MT) in
28 Region 10.
29 On still finer scales, CO emissions from on-road mobile sources or from fires can dominate in
30 different places across the U.S. Figure 3-7 illustrates this variability with CO state-level emissions
31 totals and selected county totals in 2002 for Colorado. (Annex A includes analogous data for Alaska,
32 Utah, Massachusetts, Georgia, California, and Alabama.) In Colorado, emissions from fires and on-
33 road vehicles were nearly equal: ~0.9 MT from fires and ~1.1 MT from on-road vehicles; emissions
34 sources varied strongly across counties, however, with urban Denver County dominated by on-road
35 vehicle emissions at 71% and rural Garfield County dominated by fire emissions at 67%.
September 2009 3-7 DRAFT - DO NOT CITE OR QUOTE
-------
o
co
0)
•o
o
CD
LSi&x^i!
"&P *>..
••S-**" - .•»
-150
-120
-90 -60
Longitude
-30
0
100 110 120 130
150 160 170 180 190 200 ppbv
Source: Fishman et al. (2008,1939271
Figure 3-5 CO concentrations measured by satellite at the 700 hectoPascal level (~10,000
feet above sea level) from MOPITT for the period 15-23 July 2004 during intense
wildfires in Alaska and Yukon.
30
_ 25
20
E. 15
10
-R1
-R2
-R3
R4
-R5
R6
R7
R8
-R9
-R10
90
'96 '97 '98 '99 '00 '01 '02
Year
Data are presented for 1990
and 1996-2002, as datasets
from these inventory years are
all fully up to date. Data are
available for inventory years
1991-1995, but these data have
not been updated to allow
comparison with data from
1990 and 1996-2002.
EPA Regions
© o°
o ® o
Source: U.S. EPA (2008, 157076'
Figure 3-6 Trends in sub-national CO emissions in the 10 U.S. EPA Regions for 1990 and
1996-2002.
September 2009
3-8
DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Monoxide ErniK-Jotv:- i< ..::CO.:: ;->;p^ 7:.q ;;•--
108 - 5.03
5.45 - 16.26
-E.96 - TB7.70
Colorado State Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
H 1,110,980
H 966,816
H 385,460
HI 78,308
| 25,743
7,290
6,838
527
334
53
Emissions (Tons)
Denver County Emissions 2002
Garfield County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
1,864
H 129,554
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
] 5,607
] 3,121
] 2,239
] 20,309
II64,816
Emissions (Tons)
Emissions (Tons)
Source:: U.S. EPA (2006,1570701
Figure 3-7 CO emissions density map and distributions for the state of Colorado, and for
selected counties in Colorado.
September 2009
3-9
DRAFT - DO NOT CITE OR QUOTE
-------
3.3. Physics and Chemistry of Atmospheric CO
1 In addition to being emitted directly by combustion sources, CO is produced by
2 photooxidation of methane (CH4) and other VOCs including nonmethane hydrocarbons (NMHCs) in
3 the atmosphere, and of organic molecules in surface waters and soils. CH4 oxidation is summarized
4 in this reaction sequence:
CH+OH -
4
CH3 + O2 (+M) -» CH3O2
CH3O2 + NO^ CH3O + NO2
CH3O + O2^ CH2O + HO2
or
or CH2O + OH -» HCO + H2O
HCO+O2 -^CO+HO2
Reaction 3-1
5 where M is a reaction mediator stabilizing the reaction product that is neither created nor destroyed.
6 Photolysis of formaldehyde (CH2O) proceeds by two pathways. The first produces molecular
7 hydrogen (H2) and CO with a reaction yield of 55% in conditions of clear skies and low zenith
8 angles; the second yields a hydrogen radical (H) and the formyl radical (HCO). HCO then reacts
9 with O2 to form hydroperoxy radical (HO2; OH and HO2 together are termed HOX) and CO.
10 Reaction of methyl peroxy radical (CH3O2) with HO2 radicals (reaction not shown) to form methyl
1 1 hydroperoxide (CH3OOH) is also operative, especially in low oxides of nitrogen (NO+NO2=NOX)
12 conditions. Heterogeneous removal of the water-soluble intermediate products CH3OOH, CH2O,
13 and radicals will decrease CO yields from CH4 oxidation.
14 While oxidation of CH2O nearly always produces CO and some small quantities of formic
1 5 acid (CH2O2) in the reaction of CH2O with HO2 (not shown here), oxidation of acetaldehyde
16 (CH3CHO) does not always yield two CO molecules. Reaction of CH3CHO with OH can yield
17 acetyl radicals (CH3CO) which then will participate with O2 in a termolecular recombination
18 reaction to form peroxyacyl radicals, which then can react with nitric oxide (NO) to form CH3 and
19 CO2; or the peroxyacyl radicals can react with NO2 to form peroxy acetyl nitrate (PAN),
20 CH3CO3NO2. In this way, one carbon atom is oxidized directly to CO2 without passing through CO.
21 The yield of CO from these pathways depends on the OH concentration and the photolysis rate of
22 CH3CHO, as well as on the abundance of NO, since peroxyacyl radicals also will react with other
23 odd hydrogen radicals like HO2.
September 2009 3-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 Estimating the CO yield from oxidation of hydrocarbons (HCs) larger than CH4 requires
2 computing the yields of CH2O, CH3CHO, CH3CO, and analogous radicals from oxidation of the
3 parent molecules. Moreover, the extent of heterogeneous removal of soluble intermediate products
4 also affects oxidation of more complex HCs. However, the detailed gas-phase kinetics for many HCs
5 with more than a few carbons is still unknown, and this is especially the case for several important
6 classes of VOCs including the aromatics, biogenic HCs including isoprene, and their intermediate
7 oxidation products like epoxides, nitrates, and carbonyls. It has long been known that as much as
8 30% of the carbon in HCs in many urban areas is in the form of aromatics largely from mobile
9 sources since gasoline contains significant quantities of aromatics (Grosjean and Fung, 1984,
10 040120; Seila et al., 1989, 043362). Yet mass balance analyses performed on irradiated smog
11 chamber mixtures of aromatic HCs indicate that only about one-half of the carbon is in the form of
12 compounds that can be identified. In addition, reactions like the oxidation of terpenes that produce
13 condensable products are also significant because these reactions produce secondary organic
14 aerosols, thereby reducing the potential yield of CO. The CO yield from oxidation of CH4, for
15 example, is -0.9 on a per carbon basis (Kanakidou and Crutzen, 1999, 011760). Yields from other
16 compounds range from less than 0.1 for anthropogenic alkanes (Altshuller, 1991, 192375) to ~0.9 for
17 ethane; yields from other compounds are given in Table 3-1 taken from Kanakidou and Crutzen
18 (1999.011760).
September 2009 3-11 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-1 Literature values for CO yields from hydrocarbons in per carbon units except as noted.
Specific hydrocarbons are noted in parentheses.
Reference
CO Yields
Zimmerman et al. (1978,010758)
0.3 (hydrocarbons)
Brewer etal. (1984,194402)
0.22-0.27 (isoprene)
Hanstetal. (1980.011988)
According to chamber experiments, CO and C02 yield:
-0.85 (ethylene)
-0.90 (ethane)
-0.80 (propane)
-0.58 (n-butane)
-0.73 (isoprene)
-0.30 (alpha-pinene)
Crutzen (1987, 002848)
0.9ofCH4
Kanakidou et al. (1991, 029701)
0.39 (C2H6 and C3H8)
Jacob and Wofsy (1990,029668)
@ low NOX: 0.2 (isoprene)
@ high NOX: 0.6 (isoprene)
Crutzen etal. (1985.194403)
=0.8 (isoprene + OH)
Kirchhoff and Marinho (1990,194406)
Isoprene oxidation may form 10 ppbv CO/d over the Amazon (3 km deep boundary layer)
Altshuller (1991,192375)
Conversion factors of 19 (C2-C6) anthropogenic alkenes vary between 0.010 and 0.075
Manning etal. (1997.194401)
CH4intheSH:0.7
Kanakidou and Crutzen (1999, 011760)
Annual tropospheric mean conversion factors:
CH4:0.9
Isoprene: 0.4
1
2
3
4
5
6
7
8
9
10
11
12
13
Other nonmethane hydrocarbons: 0.7
Source: adapted from Kanakidou and Crutzen (1999, 0117601
The major pathway for removal of CO from the atmosphere is reaction with OH to produce
CO2 and H radicals that rapidly combine with O2 to form HO2 radicals with a rate constant at 1 atm
in air of ~2.4xlO~13 cm3/molecule/s (Finlayson-Pitts and Pitts, 2000, 055565). The mean tropospheric
photochemical lifetime (T) of CO in the northern hemisphere is ~57 days (Khalil and Rasmussen,
1990, 012352; Thompson and Cicerone, 1986, 019374). Owing to variation in atmospheric water
vapor, OH concentration, and insolation, shorter T are found nearer the tropics and longer ones at
higher latitudes. During winter at high latitudes CO has nearly no photochemical reactivity on urban
and regional scales. Because the CO T is shorter than the characteristic time scale for mixing between
the hemispheres of ~1 year a large gradient in concentrations can exist between the hemispheres. In
addition, the CO T at high latitudes is long enough to result in much smaller gradients between 30°
latitude and the pole of either hemisphere. The typical residence times of CO in urban areas when
assuming a diel-average OH concentration of 3 xio6/cm3 in urban areas is ~16 days, so CO will not
typically be destroyed in urban areas where it is emitted and will likely be mixed on continental and
September 2009
3-12
DRAFT - DO NOT CITE OR QUOTE
-------
1 larger scales. OH concentrations are orders of magnitude lower in indoor environments and so CO
2 will generally not be destroyed by indoor air reactions.
3.3.1.CO Climate Forcing Effects
3 Recent data do not alter the current well-established understanding of the role of urban and
4 regional CO in continental and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
5 2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
6 Intergovernmental Panel on Climate Change (IPCC, 2001, 156587; IPCC, 2007, 092765). CO is a
7 weak direct contributor to greenhouse warming because its fundamental absorption band near
8 4.63 (im is far from the spectral maximum of Earth's longwave radiation at ~10 (im. Sinha and
9 Toumin (1996, 193747) estimates the direct radiative forcing (RF) of CO computed for all-sky
10 conditions at the tropopause - IPCC's preferred form for the calculation (IPCC, 2007, 092765) - to
11 be 0.024 W/m2 from the change in CO mean global concentration since pre-industrial times. The RF
12 value similarly computed by Sinha and Toumin for more than doubling the current mean global
13 background concentration to 290 ppb was 0.025 W/m2.
14 However, because reaction with CO is the major sink for OH on a global scale, increased
15 concentrations of CO can lead to increased concentrations of other trace gases whose loss processes
16 also involve OH chemistry. Some of those trace gases, CH4 and O3 for example, absorb infrared
17 radiation from the Earth's surface and contribute to the greenhouse effect directly; others, including
18 the chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), methyl chloride, and methyl
19 bromide, can deplete stratospheric O3, increasing the surface-incident UV flux.
20 This indirect effect of CO on stratospheric O3 concentrations is opposite in sign to the effect of
21 CO on O3 in the troposphere where CO reacts in a manner similar to other VOCs in the presence of
22 NOX and UV to create O3. (See the detailed description of O3 formation from VOCs and NOX in the
23 2008 NOX ISA (U.S. EPA, 2008, 157073). Because the chemical lifetime of CO is longer than the
24 VOCs most prominent on urban and regional scales and because of the one-to-one stoichiometry of
25 CO oxidation (whereby one molecule of CO converts only one molecule of NO to NO2), CO has a
26 significantly lower O3 forming potential than other VOCs in the troposphere. Carter (1998, 192380)
27 computed a maximum incremental reactivity for CO of 0.07 g O3 for 1 g CO, as compared to
28 reactivities of total on-road vehicle exhaust emissions in the range of 3 to 4. However, because the
29 total mass of CO emissions is substantially greater than those of the other VOCs with higher carbon
30 numbers and faster reactivities, CO can contribute significantly to O3 formation even though its
31 photochemical processing is slow. Using data from instrumented models including that of Jeffries
32 (1995, 003055). the NRC (1999, 010614) estimated, for example, that CO can contribute 15-25% of
33 the total O3 forming potential of gasoline exhaust emissions though this estimate shows strong
September 2009 3-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 regionality. The contribution of CO to urban and regional O3 concentration is often less than 10%
2 owing to its very slow reactivity on these scales and to locally variable radical concentration ratios.
3 Emissions of CO and the other O3 precursors, nonmethane volatile organic compounds
4 (NMVOCs) and NOX, affect the oxidizing capacity of the atmosphere largely by perturbing HOX
5 concentrations. From a climate perspective, this HOX perturbation chiefly affects the CH4 T and
6 production of O3 in the troposphere. Changes in the concentration of O3 and hence in its RF occur
7 mainly in the time of a few months. However, Prather (1996, 193195) showed that changes in CH4
8 concentration and its RF extend to the 'primary mode' timescale of troposphere chemistry of about
9 14 yr; see also Wild et al. (2001, 193196): Derwent et al. (2001, 047912). The primary mode time-
10 scale of CH4 is in part determined by the positive feedbacks in the CH4-OH-CO system in which
11 even low concentration additions of CH4 produce additional CO through oxidation by OH. That
12 additional CO then further decreases atmospheric OH concentrations when OH oxidized it to CO2.
13 The resulting decreased OH concentration then further increases the CH4 T (Daniel and Solomon,
14 1998, 193235: Isaksen and Hov, 1987, 019490). Atmospheric CH4 concentrations since 1750 have
15 increased by more than a factor of 2, giving an RF of-0.5 W/m2 (IPCC, 2001, 156587). Roughly
16 25% of the global mean tropospheric CO is produced by CH4 oxidation (Wuebbles and Hayhoe,
17 2002, 044159). Using a 2-D global model on a coarse grid Wang and Prinn (1999, 011758) showed
18 that increasing CO and CH4 concentrations leading to decreased OH concentrations can extend the
19 CO T as well as the CH4 T. Wang and Prinn varied the CO emissions and other model inputs and
20 parameters in a matrix of simulations that showed with increased or even constant 20th century CO
21 concentrations the CO T was increased by more than 50% in 100 yr.
22 CH4 is long-lived and in general well-mixed in the atmosphere; but the reaction of CH4 and
23 OH, and hence the CH4 T, is governed by the behavior and location of emissions of the short-lived
24 gases including CO, VOCs, and NOX. This produces high regional variability and uncertainty in the
25 concentrations and RFs from CO and its related climate forcing gases; see Fuglestvedt et al. (1999,
26 047431): Berntsen et al. (2006, 193244). NOX, for example, can produce effects on the combined
27 indirect RF opposite in direction to those of CH4 since under most global background conditions an
28 increase in NOX increases the global average OH concentration and decreases CH4 T and RF
29 (Berntsen et al., 2005, 193241: Wild et al., 2001, 193196) showed that emissions changes in CO and
30 NOX in Southeast Asia were more influential on the global O3 concentration and its RF (and hence
31 for the indirect O3 RF from CO) than were emissions changes in CO and NOX in Europe.
32 Using the 3-D global chemistry model MOZART-2 (Horowitz et al., 2003, 057770) Naik et al.
33 (2005, 193194) simulated changes in global tropospheric O3 concentrations and RF resulting from
34 differing reductions in emissions of NOX alone, or a combination of NOX, CO, and NMHCs in nine
35 regions of the Earth. For the reductions in Europe, North America, and Southeast Asia, reducing CO
September 2009 3-14 DRAFT - DO NOT CITE OR QUOTE
-------
1 and NMHCs in addition to reducing NOX lowered the spatial inhomogeneity of the O3 concentration
2 and RF because of the longer lifetime of CO.
3 Wild et al. (2001, 193196) used the University of California Irvine chemical transport model
4 (Wild and Prather, 2000, 052402) driven by the NASA GISS IF general circulation model (Rind and
5 Lerner, 1996, 193750) to compute changes in O3 concentrations and RF from regional emissions of
6 NOX and CO. Changes in O3 and CH4 resulting from increases in global surface NOX emissions
7 alone and run for 10 yr produced negative net RFs ranging from -0.2 in East Asia to -0.5 W/m2 in the
8 Tropics owing to the long-term interdependencies in the CO-CH4-NOX system described above.
9 When global CO emissions were increased by a 10 Tg pulse for one year together with the same one-
10 year pulsed NOX surface emissions and run again for 10 yr, the global net RF reversed in sign to
11 1.7 W/m2 (Wild et al., 2001, 193196).
12 Determining effects on several species T and RF from pulses or continuing (so-called step-
13 wise) emissions of the short-lived O3 precursor species NMVOC, CO, and NOX will increase or
14 decrease is additionally complicated by where on the Earth a particular region is on the O3
15 production response surface; see the description of the O3 production response surface and its
16 dependence on NOX and radical concentrations in the 2008 NOX ISA (U.S. EPA, 2008, 157073).
17 Fiore et al. (2002, 051221) and Fiore et al. (2008, 193749) have described the closely coupled
18 system of CH4 and O3 and its regional variation with NOX concentrations. Using the weighted
19 average results from 12 3-D global chemistry models exercised for the IPCC Third Assessment
20 Report (2001, 156587). Wigley et al. (2002, 047883) confirmed that increases in CO and VOC
21 emissions increased the O3 RF both directly and indirectly through the CH4 effects described above,
22 and that NOX emissions produced a mix of direct and indirect increases in RF mostly dominated by
23 the direct effects for all modeled scenarios. Wigley et al. (2002, 047883) concluded that tropospheric
24 O3 RF influences were larger than CH4 influences and that the short-lived reactive gases produced
25 60% to 80% of that forcing, with the remainder coming from CH4.
26 Because of these chemical interdependencies, calculations of an indirect RF for any of these
27 short-lived O3 precursor species are most often made for all of the most important ones together. So,
28 for example, the combined effect of increased CH4, CO, NMVOC, and NOX emissions since 1750
29 has produced tropospheric O3 concentrations associated with a net RF of-0.35 W/m2 (IPCC, 2001,
30 156587). The integrated 20-yr and 100-yr time horizon RFs were computed by IPCC (2007, 092765)
31 for year 2000 emissions of CO, NMVOC, and NOX to be ~0.19 W/m2, just slightly lower than the
32 RF of year 2000 black carbon emissions from fossil fuel and biomass burning on the same horizons.
33 The combined RF computed for all emissions and changes in CO in the years 1750-2005 for all
34 indirect effects of CO through O3, CH4, and CO2 was also ~0.2 W/m2, more than a factor of 3 larger
35 than the indirect effect of the shorter-lived NMVOCs on the same three GHGs, 0.06 W/m2. Of the
September 2009 3-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 three indirect effects from CO emissions, the O3-related component was the largest, accounting for
2 approximately one-half of the forcing (IPCC, 2007, 092765).
3 It is also possible to compute individual contributions to the integral RF from CO from
4 separate emissions sectors. Unger et al. (2009, 193238) used the NASA GISS model for Physical
5 Understanding of Composition-Climate Interactions and Impacts (G-PUCCINI) (Shindell et al.,
6 2006, 193751) and divided the 1995 global anthropogenic CO emissions total of 846.7 Tg/yr into
7 sectors for on-road transport (ORT) and power generation (PG), and then separated contributions
8 from each of these sectors for the U.S. and other large geographic regions of the Earth. ORT CO
9 emissions in the U.S. were 76.3 Tg/yr; PG CO emissions were 0.5 Tg/yrout of the total U.S.
10 anthropogenic CO emissions of 102.1 Tg/yr. Unger et al. concluded from analysis of 7 yr of runs that
11 the CO indirect CH4 effects (that is, the CO effects through CH4 changes as described above) in the
12 1995 emissions run were -0.004 W/m2 for the global ORT and -0.022 W/m2 for the global PG. In the
13 U.S., the indirect CH4 RF was positive at +0.009 W/m2 because the positive effects on CH4 T from
14 the CO emissions dominated over the negative effects from NOX through OH. This RF fraction from
15 indirect CH4 is approximately the same as the direct O3 RF from ORT in the U.S., 0.010 W/m2.
16 Because the PG sector emits NOX but less CO relative to the ORT, the indirect CH4 RF from the
17 U.S. PG was not dominated by the positive CO effects and remained a net negative at -0.006 W/m2
18 (Unger et al., 2009, 193238).
19 These gross emissions sectors can also be subdivided to demonstrate more clearly the
20 localized chemical interdependencies of the CO-CH4-NOX system. Fuglestvedt et al. (2008, 193242)
21 used the Oslo CTM2 model to simulate effects from all emissions and changes in all transportation
22 subsectors from 1850-2000. Fuglestvedt et al. found that global transport has been responsible for
23 -15% of the total anthropogenic CO2 RF and -15% of the total anthropogenic O3 RF. Of the total
24 O3 RF, the largest contributor was the shipping sector, because its high NOx-to-CO and NOx-to-
25 VOC ratios produced OH increases and hence CH4 decreases in regions of naturally low NOX. For
26 the shipping segment of the transport sector, the high NOX emissions there reduced the CH4 T but
27 increased O3. The global mean effect from these two was small and still smaller than the direct
28 negative effect from SO4 aerosols. In the on-road segment of global transportation, emissions of CO
29 and VOCs together with NOX produce an O3 RF larger than the negative RF from CH4.
30 Caution is warranted before using any of these these results too freely. RF values are global
31 model calculations using the assumption that global climate sensitivities are equal for all forcing
32 mechanisms, whether CO2, sulfates and other aerosols, or the short-lived gases like CO (Berntsen et
33 al., 2005, 193241; Berntsen et al., 2006, 193244). That assumption is under challenge now by CGM
34 results using regionalized RF values separately for different forcing mechanisms and with CO2, O3,
35 and solar input changes (Joshi et al., 2003, 193752). Joshi et al. found that global climate system
36 sensitivities from non-CO2 RF varied by ±30% compared to CO2 RF. Other GCM experiments by
September 2009 3-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 Lelieveld et al. (2002, 190361). Rotstayn and Penner (2001, 193754). Menon et al. (2002, 155978).
2 and Kristjansson (2002, 045282) have indicated that regionally changing RF can induce changes in
3 large-scale circulation patterns that control the regionalized cycles of flooding and drought through
4 disruptions in regional temperature and hydrologic cycles. Using the U.K. Meteorological Office
5 3-D Lagrangian CTM STOCHEM (Collins et al., 1997, 193193). Derwent et al. (2008, 193245) have
6 shown the scale of RF differences from changing surface-level NOX emissions to be large and
7 variable in affecting O3 T and RF, but that the counter-effects on CH4 - increased oxidation to CO
8 from increased OH concentrations from NOX - are larger still. However, such regionalized patterns
9 resulting from GCM experiments are so uncertain and so widely variable across models that even the
10 sign of these regionalized changes can vary with model type and any of the models' unconstrained
11 assumptions (Berntsen et al., 2006, 193244). Even with such uncertainty and variability, though, the
12 consensus of the climate community is that the climate effects of changes to emissions of the long-
13 and especially the short-lived pollutants including CO are very likely not independent of location.
14 Because the greenhouse warming effects from CO are nearly completely indirect, and because
15 CO concentrations are spatially heterogeneous, neither the IPCC nor EPA computes direct global
16 warming potentials (GWPs) for CO, just as they do not for tropospheric O3, NO, NO2, or VOCs
17 (U.S. EPA, 2008, 184463). GWP is a widely used relative measure of the potential effect of different
18 emissions on climate usually defined as the time integrated commitment to climate forcing from an
19 instantaneous pulsed release of 1 kg of a trace gas relative to the effects from a pulsed release of 1 kg
20 of CO2. The GWP values evaluated and summarized by IPCC are global and cannot reflect effects of
21 localized emissions or emissions changes, making the values for the short-lived species NMVOC,
22 CO, and NOX more uncertain than the values for the long-lived well mixed species because of the
23 OH chemistry described above. Moreover, urban and regional-scale oxidation of CO to CO2 under
24 current atmospheric conditions proceeds very slowly and IPCC considers production of CO2 through
25 this pathway to be double counting of CO effects (IPCC, 2007, 092765).
26 However, some groups of atmospheric scientists have made estimates of CO GWP and those
27 have been reviewed by IPCC though without a final conclusive statement. The unusually large
28 heterogeneity in model type and form, pulsed or stepped emissions increase, time horizon unit, and
29 integral or differential indirect effects in several combinations - with or without NOX emissions
30 changes, including or excluding CO2 effects - imparts variation to the CO GWP range of estimates.
31 Even with such variability in methods and tools, when carefully considered, the CO GWPs have
32 been largely in agreement for approximately 10 yr. For example, Daniel and Solomon (1998,
33 193235) used a global box model for changes through CH4 and O3 effects from pulsed CO
34 emissions and estimated a CO GWP exclusive of the effect through CO2 to be between 1 and 4.4.
35 Using the STOCHEM CTM, Derwent et al. (2001, 047912) estimated a pulsed emissions CO GWP,
36 again exclusive of effects through CO2, to be 1.5. Johnson and Derwent (1996, 193192) had
September 2009 3-17 DRAFT - DO NOT CITE OR QUOTE
-------
1 previously computed and integrated GWP of 2.1 for the CH4 and O3 effect from a step-wise
2 emissions change using a 2-D and a 100-y time horizon. Derwent et al. (2001, 047912) and Collins
3 et al. (2002, 044156) subsequently differentiated that integral for each effect and reported GWP for
4 step-wise CO emissions changes on a 100-year time horizon of 1.0, 0.6, and 1.6 through the effects
5 on CH4, O3, and CO2, respectively. Most recently, Berntsen et al. (2005, 193241) used the model
6 LMDz v3.3 (Hauglustaine et al., 2004, 193191) to compute 100-year GWP values for pulsed CO
7 emissions through all indirect effects to be 1.9 as resolved for Europe and 2.4 for Asia,
8 demonstrating the strong regionality in the indirect effects from these short-lived precursors.
3.4. Ambient Measurements
3.4.1.Ambient Measurement Instruments
9 For enforcement of the air quality standards set forth under the Clean Air Act, EPA has
10 established provisions in the Code of Federal Regulations (CFR) under which analytical methods can
11 be designated as federal reference methods or federal equivalent methods (FRM or FEM,
12 respectively). Measurements for determinations of NAAQS compliance must be made with FRMs or
13 FEMs. As of August 2009, 20 automated FRMs and no FEMs had been approved for CO
14 (http://www.epa.gov/ttn/amtic/criteria.html).
15 All EPA FRMs for CO operate on the principle of nondispersive infrared (NDIR) detection
16 and can include the gas filter correlation (GFC) methodology. NDIR is an automated and continuous
17 method based on the specific absorption of infrared radiation by the CO molecule. Most
18 commercially available analyzers incorporate a gas filter to minimize interferences from other gases
19 and operate near atmospheric pressure. NDIR is based on the physics of CO's characteristic infrared
20 absorption near 4.63 um. NDIR methods have several practical advantages over other techniques for
21 CO detection in that they are not sensitive to flow rate changes, require no wet chemicals, are
22 reasonably independent of ambient air temperature changes, are sensitive over wide concentration
23 ranges, and have fast response times. An extensive and comprehensive review of NDIR, GFC, and
24 alternative, non-FRM techniques for CO detection including tunable diode laser spectroscopy, gas
25 chromatography, mercury liberation, and resonance fluorescence was made for the 2000 CO AQCD
26 (U.S. EPA, 2000, 000907). and the reader is directed there for additional information. The
27 description here is limited to a brief outline of the FRM NDIR and GFC techniques.
28 GFC spectroscopy analyzers are used most frequently now in documenting compliance with
29 ambient air standards. A GFC monitor has all of the advantages of an NDIR instrument and the
30 additional advantages of smaller size, no interference from CO2, and very small interference from
31 water vapor. During operation, air flows continuously through a sample cell. Radiation from the
September 2009 3-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 infrared source is directed by optical transfer elements through two main optical subsystems: (1) the
2 rotating gas filter and (2) the optical multipass (sample) cell. The beam exits the sample cell through
3 an interference filter (FC), which limits the spectral passband to a few of the strongest CO absorption
4 lines. Detection of the transmitted radiation occurs at the infrared detector. The gas correlation cell is
5 constructed with two compartments, one filled with 0.5 atm CO, and a second with pure nitrogen gas
6 (N2). Radiation transmitted through the CO is completely attenuated at the wavelengths where CO
7 absorbs strongly. The radiation transmitted through the N2 is reduced by coating the exit window of
8 the cell with a neutral attenuator so that the amounts of radiation transmitted by the two cells are
9 made approximately equal in the passband that reaches the detector. In operation, radiation passes
10 alternately through the two cells as they are rotated to establish a signal modulation frequency. If CO
11 is present in the sample, the radiation transmitted through the CO is not appreciably changed,
12 whereas that through the N2 cell is changed. This imbalance is linearly related to CO concentrations
13 in ambient air.
14 Specifications for CO monitoring are designed to help states demonstrate whether they have
15 met compliance criteria; operational parameters required under 40 CFR 53 are provided in Table 3-2.
16 Given the 1-h level of the NAAQS of 35 ppm and the 8-h level of the NAAQS of 9 ppm, a 1.0 ppm
17 LOD is sufficient for demonstration of compliance. However, with ambient CO levels now routinely
18 at or below 1 ppm, there is greater uncertainty in the monitoring data because a large percentage is
19 below the LOD. For this reason, a new generation of ambient CO monitors has been designed for
20 trace-level measurements. Additionally, trace-level CO measurements are needed to support
21 additional objectives such as validating the inputs to chemical transport models (CTMs) and
22 assessing differences between CO levels in urban and rural areas, because background CO
23 concentrations are on the order of 0.1 ppm. Effective LOD is influenced by instrumental noise and
24 drift and by the amount of water vapor in the air. Recent improvements in the instruments' optical
25 components and dehumidification of the air stream help to reduce the amount of noise and drift in
26 the CO measurements. Newer GFC instruments have been designed for automatic zeroing to
27 minimize drift (U.S. EPA, 2000, 000907).
September 2009 3-19 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-2 Performance specifications for analytical detection of CO, based on 40 CFR Part 53.
Range 0-50 ppm
Noise 0.5 ppm
LOD 1.0 ppm
Interference equivalent
Each interfering substance ±1.0 ppm
Total interfering substances 1.5 ppm
Zero drift
12 h ±1.0 ppm
24 h ±1.0 ppm
Span drift, 24-h
20% of upper range limit ±10.0%
80% of upper range limit ±2.5%
Lag time 10min
Rise time 5 min
Fall time 5 min
Precision
20% of upper range limit 0.5 ppm
80% of upper range limit 0.5 ppm
1 Currently, 24 types of CO monitors are in use; the models are listed in Annex Table A-l.
2 Among them, 20 are older NDIR instruments listed to have a limit of detection (LOD) of 0.5 ppm,
3 and 4 are trace-level GFC instruments listed to have an LOD of 0.04 ppm. States do not routinely
4 report the operational limit of detection, precision, and accuracy of the monitors to the U.S. EPA's
5 Air Quality System (AQS). Some states report the raw monitored data, while others report the
6 concentration as 50% of the LOD (0.25 ppm for high-LOD instruments and 0.02 for low-LOD
7 instruments) when reported data are below the LOD. Among several of the older instruments still in
8 use (Federal Reference Method codes 008, 012, 018, 033, 041, 050, 051, and 054), performance
9 testing has shown effective LODs of 0.62-1.05 ppm, with 24-h drift ranging from 0.044-0.25 ppm
10 and precision ranging from 0.022-0.067 ppm at 20% of the upper range limit of the instrument
11 (Michie RM et al., 1983, 194043). Among newer GFC trace-level instruments, manufacturer-
12 declared LODs range from 0.02-0.04 ppm, with 24-h zero drift varying between 0.5% within 1 ppm
13 and 0.1 ppm, and precision varying from 0.5% to 0.1 ppm.
14 Comparison of older and newer, trace-level monitors calls attention to several data quality
15 issues with the older monitors; Figure 3-8 shows data from collocated older and trace-level monitors
16 in Charlotte, NC to illustrate this point. First, the data appearing below the LOD of 0.5 ppm for the
17 older monitor comprise 58% of the data obtained by that monitor. In contrast, no data from the trace-
18 level monitor are reported below the LOD of 0.04 ppm. Second, the data from the older monitor are
September 2009 3-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 reported in units of 0.1 ppm, as seen in the lower resolution of the data points. Last, it is possible
2 from the data that the older monitor exhibits some upward drift, since newer models have automatic
3 zeroing functions. The median data are 0.4 ppm for the older monitor and 0.24 ppm for the trace-
4 level monitor. However the mean from the older monitor is 0.4 ppm, in contrast with 0.330 ppm for
5 the trace-level monitor. The 99th percentile is 1.8 ppm for the older monitor, in contrast with the
6 newer monitor, whose 99th percentile level is 1.485 ppm. However, because both the older and the
7 trace-level CO monitors require calibration, it is not possible to state with certainty that drift exists
8 for the older monitor.
3.5
• Method 593
• Method D54above LOD
A Method D54belouu LOD
D 2DDD 4000 6000 80DO 10DDO 12000 14000 16DOO 18000 20000
Figure 3-8 Data from collocated monitors in Charlotte, NC. Data from method 054 are from
an older (Thermo Electron Model 48C, Waltham, MA) model, while data from
method 593 are from a new trace-level instrument (Teledyne API Model 300EU,
San Diego, CA).
September 2009
3-21
DRAFT - DO NOT CITE OR QUOTE
-------
3.4.2.Ambient Sampling Network Design
3.4.2.1. Monitor Siting Requirements
1 Minimum monitoring requirements for CO were revoked in the 2006 revisions to ambient
2 monitoring requirements (see 71 FR 61236, October 17, 2006). This action was made to allow for
3 reductions in measurements of CO and some other pollutants (SO2, NO2, and Pb) where measured
4 levels were well below the applicable NAAQS and air quality problems were not expected. CO
5 monitoring activities have been maintained at some State and Local Air Monitoring Stations
6 (SLAMS), and these measurements of CO using FRM are required to continue until discontinuation
7 is approved by the EPA Regional Administrator. CO monitors are typically sited at the following
8 spatial scales (40 CFR Part 58 Appendix D):
9 Microscale: Data represents concentrations within a 100 m radius of the monitor. For CO,
10 microscale monitors are sited 2-10 m from a roadway. Measurements are intended to represent the
11 near-road or street canyon environment.
12 Middle scale: Data represents concentrations averaged over areas defined by 100-500 m radii.
13 Measurements are intended to represent several city blocks.
14 Neighborhood scale: Data represents concentrations averaged over areas defined by 0.5-4.0
15 km radii. Measurements are intended to represent extended portions of a city.
16 In 2007, there were 376 CO monitors reporting values to the EPA Air Quality System (AQS)
17 database. Where CO monitoring is ongoing, 40 CFR Part 58 requires at least one CO monitor to
18 capture maximum levels in a given region. This requirement is met with a monitor situated at the
19 CFR-defmed microscale distance from the side of a roadway for CO. Microscale monitor locations
20 also have sample inlets mounted at 3 ± 0.5 m above ground level, unlike the monitors sampling for
21 larger scales, whose inlet heights can vary between 2 and 15m. For the CFR-defmed neighborhood
22 scale monitoring, the minimum monitor distance from a major roadway is directly related to the
23 average daily traffic counts on that roadway to ensure that measurements are not substantially
24 influenced by any one roadway. For example, the minimum distance of a neighborhood scale CO
25 monitor from a roadway with an average daily traffic count of 15,000 vehicles per day is 25 m, while
26 the minimum distance is 135 m for a roadway with an average daily traffic of 50,000 vehicles per
27 day. Occasionally, CO monitors are sited at urban (covering areas of 4-50 km) or regional (covering
28 areas of tens to hundreds of km) scale. More detail on siting requirements can be found in 40 CFR
29 Part 58 Appendices D and E.
30 In addition to monitoring for determining compliance with the NAAQS, the U.S. EPA is
31 currently in the process of implementing plans for a new network of multipollutant stations called
32 National Core (NCore) that is intended to meet multiple monitoring objectives. A subset of the
September 2009 3-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 SLAMS network, NCore stations are intended to address integrated air quality management needs to
2 support long-term trends analysis, model evaluation, health and ecosystem studies, as well as the
3 more traditional objectives of NAAQS compliance and Air Quality Index reporting. States were
4 required to submit Annual Monitoring Network Plans describing their candidate NCore stations by
5 July 1, 2009. EPA is reviewing these plans and intends to provide station approvals later in 2009.
6 The complete NCore network, required to be fully implemented by January 1, 2011, will consist of
7 approximately 60 urban and 20 rural stations and will include some existing SLAMS sites that have
8 been modified for the additional measurements. Each state will contain at least one NCore station,
9 and 46 of the states plus Washington, D.C. will have at least one urban station. CO will be measured
10 using trace-level monitors at all sites, as will SO2, NO, and NOv1; surface meteorology will also be
11 measured at NCore sites. The advantage to the NCore strategy is that time-resolved, simultaneous
12 measurements of multiple pollutants will be obtained at each site. The disadvantage is that the NCore
13 network will be sparse, and so spatial variability will be difficult to ascertain from the data obtained.
3.4.2.2. Spatial and Temporal Coverage
14 Figure 3-9 depicts the distribution of the 376 regulatory CO monitors operating in the U.S. in
15 2007. Data from 291 of the 376 CO monitors operating year-round at 290 sites in the years
16 2005-2007 met the data completeness criteria for inclusion in the multiyear ambient data analyses
17 for this assessment. Completeness criteria require that data be collected for 75% of the hours in a
18 day, 75% of the days in a quarter, and three complete quarters in a year for all 3 yr; criteria for
19 Region 10 were relaxed to two complete quarters a year because it contains Alaska. The greatest
20 density of monitors is in the CSAs for Los Angeles, CA and San Francisco, CA, and along the Mid-
21 Atlantic sea board. Monitors are also located in regions where biomass burning is more prevalent,
22 such as Anchorage, AK, but not all of these monitors report values from all seasons of all years. The
23 number of monitors per sampling scale is provided in Table 3-3, and locations of monitors with
24 nearby roadway types and traffic counts are provided in Annex Tables A-2 through A-7 for each
25 monitoring scale.
26 Figure 3-9 also shows the locations of trace-level CO monitors throughout the U.S in 2007.
27 The trace-level monitors included in the analysis are located in Baton Rouge, LA; Boston, MA;
28 Charlotte, NC; Dallas, TX; Decatur, GA; Houston, TX; Portland, OR; Presque Isle, ME; San Jose,
29 CA; and rural locations within Georgia and South Carolina. Other trace-level monitors not meeting
30 completeness criteria for the 2005-2007 analysis were located in Beltsville, MD; Cedar Rapids, IA;
31 Davenport, IA; Des Moines, IA; Nederland, TX; Northbrook, IL; Plant City, FL; Seattle, WA;
•' NCore sites must measure, at a minimum, PM2.5 particle mass using continuous and integrated/filter-based samplers, speciated PM2.5,
PMio-2.5 particle mass, speciated PM 10-2.5, O3, SO2, CO, NO/NOY, wind speed, wind direction, relative humidity, and ambient
temperature.
September 2009 3-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 Thomaston, CT; Tulsa, OK; Westport, CT; and rural locations in Maryland and Wisconsin. A listing
2 of trace-level and high-LOD monitors meeting completeness criteria by state for 2005-2007 is
3 provided in Annex Table A-8.
4 Eleven metropolitan regions were chosen for closer investigation of monitor siting based on
5 their relevance to the health studies assessed in subsequent chapters of this ISA and to demonstrate
6 specific points about geospatial distributions of CO emissions and concentrations. These regions
7 were: Anchorage, AK; Atlanta, GA; Boston, MA; Denver, CO; Houston, TX; Los Angeles, CA; New
8 York City, NY; Phoenix, AZ; Pittsburgh, PA; Seattle, WA; and St. Louis, MO. Core-Based Statistical
9 Areas (CBSAs) and Combined Statistical Areas (CSAs), as defined by the U.S. Census Bureau
10 (http://www.census/gov/). were used to determine which counties, and hence which monitors, to
11 include for each metropolitan region.1 As an example, Figure 3-10 through Figure 3-13 display CO
12 monitor density with respect to population density (for total population and elderly adults aged 65
13 and over) for the Denver and Los Angeles CSAs. (Annex A, Figures A-7 through A-22 show
14 analogous plots for the other nine metropolitan regions.) Figure 3-17 and Figure 3-19 in Section 3.5
15 and additional figures in Annex A show the locations of CO monitors for the 11 CSAs/CBSAs in
16 relation to major roadways, including Interstate highways, U.S. highways, state highways, and other
17 major roadways required for traffic network connectivity. In the examples shown for Denver and Los
18 Angeles, the monitors were typically located near high population density neighborhoods within the
19 CS A/CBS A. The Los Angeles CSA monitors appear to be distributed fairly evenly across the city of
20 Los Angeles, while the Denver CSA had three monitors in the city center and two in the suburbs of
21 the Denver CSA. Regional background sites were not included on the maps unless they lay within
22 the CSA/CBSA.
23 Ambient monitors for CO and other criteria pollutants are located to monitor compliance
24 rather than population exposures. However, CO monitors submitting data to the AQS are often used
25 for exposure assessment. For this reason, data are presented here to assess population density in the
26 vicinity of CO monitors. Table 3-4 and Table 3-5 show the population density around CO monitors
27 for the total population and for elderly adults aged 65 and over for each CSA/CBSA. The percentage
28 of population within specific radii of the monitors for each city was, for the most part, similar
29 between the total and elderly populations. In the cases of Anchorage, Denver, Phoenix, and St. Louis
30 however, the percentage of the elderly population within given radii of the monitors was
31 considerably different compared with the total population. Between-city disparities in population
32 density were larger. Los Angeles, with 85%, and Denver, with 68%, had the largest proportion of the
•' A CBSA represents a county-based region surrounding an urban center of at least 10,000 people determined using 2000 census data and
replaces the older Metropolitan Statistical Area (MSA) definition from 1990. The CSA represents an aggregate of adjacent CBSAs tied
by specific commuting behaviors. The broader CSA definition was used when selecting monitors for the cities listed above with the
exception of Anchorage and Phoenix, which are not contained within a CSA. Therefore, the smaller CBSA definition was used for these
metropolitan areas.
September 2009 3-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 total population within 15 km of a monitor. Seattle, with 18%, had the lowest population coverage in
2 large part because ambient CO concentrations there require only a single CO monitor. For the elderly
3 population, Los Angeles, at 83%, Anchorage, at 73%, and Denver, at 70%, had the greatest
4 population coverage within 15 km of a monitor, whereas Seattle, at 18%, again had the lowest
5 coverage. Proximity to monitoring stations is considered further in Sections 3.5 and 3.6 regarding
6 spatial variability within cities. In combination, these data illustrate that population coverage varies
7 by monitor and across cities.
CO Monitor Locations in United States in 2007
~
o
* CO Monitors - included in analysis, low LOD, 2007
* CO Monitors - included in analysis,high LOD, 2007
» CO Monitors - not included in analysis, low LOD. 2007
CO Monitors - not included in analysis, high LOD, 2007
CSA/CBSA
Figure 3-9 Map of 376 CO monitor locations in the U.S. in 2007. Locations are indicated with
triangles: filled triangles show locations of the 290 sites used in data analysis for
this assessment; open triangles are at locations with monitors which did not
meet the data completeness requirements for analysis; blue lines mark the
boundaries of the 11 CSAs/CBSAs used in the data analysis for this assessment.
September 2009
3-25
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-3 Counts of CO monitors by sampling scale meeting 75% completeness criteria for use in
the U.S. during 2005-2007.
Monitoring Scale Count
Microscale 57
Middle Scale 31
Neighborhood Scale 119
Urban Scale 11
Regional Scale 2
Null 71
September 2009 3-26 DRAFT - DO NOT CITE OR QUOTE
-------
Denver Combined Statisical Area
0 S 10
30 40
I Kilometers
01530 60 90 120
2005 Population Density
^ Denver CO Monitors (5 km buffer)
Population per Sq Km
^B °-67
^B 68-135
136 - 673
674 -1347
| 1348 - 3364
^H 336S -13456
Figure 3-10 Map of CO monitor locations with respect to population density in the Denver, CO
CBSA, total population.
September 2009
3-27
DRAFT - DO NOT CITE OR QUOTE
-------
Denver Combined Statisical Area
0 S 10
30 40
2005 Population Density
| | Denver CO Monitors (5 km buffer)
Population > 65 per Sq Km
^B o-26
^B 27-53
54-263
264 - 525
^B 526-1313
^H 1314-5251
01530 60 90
I Kilometers
120
Figure 3-11 Map of CO monitor locations with respect to population density in the Denver, CO
CBSA, age 65 and older.
September 2009
3-28
DRAFT - DO NOT CITE OR QUOTE
-------
Los Angeles Combined Statisical Area
I Kilometers
0 5 10 20 30 40
2006 Population Density
| | Los Angeles CO Monitors (5 km buffer)
Population per Sq Km
^H 0 - 271
^H 272 - 542
543 - 2711
2712 - 6422
| 5423 -13556
• 13557 - 54222
0 3060 120
I Kilometers
180 240
Figure 3-12 Map of CO monitor locations with respect to population density in the Los
Angeles, CA CSA, total population.
September 2009
3-29
DRAFT - DO NOT CITE OR QUOTE
-------
Los Angeles Combined Statisical Area
I Kilometers
0 5 10 20 30 40
2006 Population Density
| | Los Angeles CO Monitors (5 km buffer)
Population > 65 per Sq Km
^H 0-33
77-382
383 - 767
768-1911
• 1912 - 7645
I Kilometers
0 3060 120 180 240
Figure 3-13 Map of CO monitor locations with respect to population density in the Los
Angeles, CA CSA, age 65 and older.
September 2009
3-30
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-4 Proximity to CO monitors for the total population by city.
Region
Anchorage, AK
Atlanta, GA
Boston, MA
Denver, CO
Houston, TX
Los Angeles, CA
New York, NY
Phoenix, AZ
Pittsburgh, PA
Seattle, WA
St. Louis, MO
Total CSA/
CBSA
N
352,225
5,316,742
7,502,707
2,952,039
5,503,320
17,655,319
22,050,940
3,818,147
2,515,383
3,962,434
2,869,955
<1km
N %
5,391 1.53
5,480 0.10
95,732 1.28
26,096 0.88
2,9068 0.53
202,340 1.15
201,350 0.91
47,478 1.24
29,136 1.16
4,814 0.12
16,638 0.58
<5km
N %
131,608 37.36
149,772 2.82
1,180,054 15.73
497,598 16.86
599,796 10.90
4,064,309 23.02
3,711,369 16.83
503,433 13.19
369,965 14.71
94,649 2.39
255,499 8.90
<10km
N %
212,834 60.43
672,701 12.65
2,432,846 32.43
1,091,444 36.97
1,669,117 30.33
11,928,427 67.56
8,385,801 38.03
1,033,102 27.06
895,252 35.59
279,976 7.07
886,412 30.89
<15km
N %
239,842 68.09
1,444,986 27.18
3,418,353 45.56
1,720,360 58.28
2,506,830 45.55
15,074,972 85.38
12,454,837 56.48
1,581,887 41.43
1,359,596 54.05
699,490 17.65
1,303,636 45.42
Table 3-5 Proximity to CO monitors for adults aged 65 and older by city.
Region
Anchorage, AK
Atlanta, GA
Boston, MA
Denver, CO
Houston, TX
Los Angeles, CA
New York, NY
Phoenix, AZ
Pittsburgh, PA
Seattle, WA
St. Louis, MO
Total CSA/
CBSA
N
17,742
362,201
945,790
232,974
377,586
1,626,663
2,710,675
388,150
449,544
390,372
358,747
<1km
N %
361 2.03
423 0.12
8,272 0.87
2,541 1.09
1 ,703 0.45
17,974 1.10
29,534 1.09
2,877 0.74
5,383 1.20
556 0.14
3,203 0.89
<5km
N %
8,986 50.65
12,758 3.52
131,198 13.87
42,760 18.35
42,312 11.21
380,079 23.37
427,601 15.77
35,839 9.23
66,967 14.90
12,142 3.11
42,890 11.96
<10km
N %
12,038 67.85
54,148 14.95
297,392 31.44
102,783 44.12
130,567 34.58
1,069,188 65.73
940,121 34.68
77,244 19.90
166,440 37.02
3,1036 7.95
127,274 35.48
<15km
N %
12,990 73.22
111,232 30.71
430,502 45.52
163,682 70.26
182,049 48.21
1,355,461 83.33
1,429,215 52.73
125,300 32.28
255,220 56.77
69,858 17.90
184,491 51.43
September 2009
3-31
DRAFT - DO NOT CITE OR QUOTE
-------
3.5. Environmental Concentrations
3.5.1.Spatial Variability
3.5.1.1. National Scale
1 The current NAAQS designates that the level of the NAAQS is not to be exceeded more than
2 once per year at a given location. Figure 3-14 and Figure 3-15 show the second-highest 1-h and
3 second-highest 8-h county-average CO concentrations, respectively, over the U.S. along with
4 estimates of the fraction of U.S. total population exposed to those concentrations. Although 93% of
5 the U.S. counties are not represented in AQS reporting, based on their population densities and
6 proximity to sources, those counties are not expected to have higher concentrations than the ones
7 analyzed here in the absence of extreme events such as wildfires. Continuous hourly averages are
8 reported from U.S. monitoring stations. 1-h and 8-h CO data were available for 243 counties and
9 autonomous cities or municipalities (e.g., Anchorage, AK, Washington, DC) where CO monitors met
10 the 75% data completeness criteria used in this analysis for the years 2005-2007. In 2007, no
11 monitored location reported a second-highest 1-h CO concentration above 35 ppm; see Figure 3-14.
12 Moreover, only two monitored locations, one in Weber Co., UT and the other in Jefferson Co., AL
13 (including Birmingham, AL), reported second-highest 1-h CO concentrations between 15.1 and
14 35.0 ppm. Figure 3-15 shows that only 5 counties reported second-highest 8-h CO concentrations
15 above 5.0 ppm: Jefferson Co., AL; Imperial Co., CA; Weber Co., UT; Philadelphia Co., PA; and
16 Anchorage Municipality, AK.
September 2009 3-32 DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Monoxide - Second Highest 1-hour Average, 2007
300 -,
250 -
200 -
150 -
100 -
50 -
Population
(millions)
Concentration:
• > 35.1 ppm
• 15.1-35.0 ppm
10.1-15.0 ppm
• 5.1 - 10.0 ppm
< 5.0 ppm
Q No data
Figure 3-14 County-level map of second-highest 1-h avg CO concentrations in the U.S. in
2007. The bar on the left shows the total U.S. population living in counties with
CO concentrations in the range indicated. Note that approximately 150 million
people live in counties with no CO monitors.
September 2009
3-33
DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Monoxide - Second Highest 8-hour Average, 2007
300 -,
250 -
200 -
Concentration:
n >9.1 ppm
1 7.6 - 9.0 ppm
5.1 -7.5 ppm
1 2.6-5.0 ppm
< 2.5 ppm
] No data
Figure 3-15 County-level map of second-highest 8-h avg CO concentrations in the U.S. in
2007. The bar on the left shows the total U.S. population living in counties with
CO concentrations in the range indicated. Note that approximately 150 million
people live in counties with no CO monitors.
September 2009
3-34
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-6 Distribution of 1-h avg CO concentration (ppm) derived from AQS data.
Percentiles
n
Mean
Min 1
5
10
25
50
75
90
95
99
Max
NATIONWIDE STATISTICS (N = NUMBER OF OBSERVATIONS)
2005-2007 7,180,700
2005 2,391,962
2006 2,402,153
2007 2,386,585
Winter (December - February) 1 ,752,340
Spring (March - May) 1 ,826,1 67
Summer (June -August) 1,811,082
Fall (September - November) 1,791,111
0.5
0.5
0.5
0.4
0.6
0.4
0.4
0.5
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.1
0.2
0.2
0.2
0.2
0.3
0.2
0.2
0.2
0.4
0.4
0.4
0.3
0.4
0.3
0.3
0.4
0.6
0.6
0.6
0.5
0.7
0.5
0.5
0.6
0.9
1.0
0.9
0.8
1.2
0.8
0.7
1.0
1.2
1.3
1.2
1.1
1.6
1.0
0.9
1.3
2.1
2.3
2.1
1.9
2.7
1.7
1.5
2.2
39.0
22.3
35.3
39.0
20.0
35.3
39.0
24.1
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007 285
2005 285
2006 285
2007 285
Winter (December - February) 285
Spring (March - May) 285
Summer (June - August) 285
Fall (September - November) 285
0.5
0.5
0.5
0.4
0.6
0.4
0.4
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.2
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.4
0.3
0.3
0.4
0.3
0.3
0.4
0.4
0.5
0.4
0.4
0.5
0.4
0.4
0.4
0.6
0.6
0.6
0.5
0.7
0.5
0.5
0.6
0.7
0.8
0.7
0.7
0.9
0.7
0.6
0.8
0.8
0.9
0.8
0.7
1.1
0.7
0.7
0.9
1.0
1.3
1.2
1.1
1.5
1.0
1.1
1.1
1.5
1.6
1.4
1.5
1.6
1.6
1.5
1.5
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007) (N = NUMBER OF OBSERVATIONS)
Anchorage3
Atlanta
Boston
Denver
Houston
Los Angeles
New York
Phoenix
Pittsburgh
Seattle
St. Louis
Not in the 11 cities
25,672
76,683
171,975
129,038
123,925
592,960
226,673
127,477
179,758
25,818
77,142
5,449,251
1.1
0.5
0.4
0.5
0.3
0.5
0.5
0.8
0.3
0.8
0.4
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.2
0.2
0.0
0.1
0.0
0.0
0.1
0.1
0.0
0.2
0.1
0.0
0.3
0.2
0.1
0.2
0.0
0.1
0.1
0.2
0.0
0.3
0.2
0.1
0.5
0.3
0.2
0.3
0.2
0.2
0.3
0.3
0.1
0.4
0.3
0.2
0.7
0.4
0.4
0.4
0.3
0.3
0.5
0.5
0.2
0.6
0.4
0.4
1.3
0.6
0.5
0.6
0.4
0.6
0.6
1.0
0.4
0.9
0.5
0.6
2.3
0.8
0.7
1.0
0.6
1.0
0.9
1.9
0.6
1.3
0.7
0.9
3.1
1.1
0.9
1.3
0.8
1.4
1.1
2.5
0.8
1.6
0.9
1.2
5.0
1.6
1.4
2.2
1.4
2.3
1.6
3.6
1.2
2.5
1.4
2.1
13.1
10.8
10.0
9.3
4.6
8.4
5.8
7.8
6.7
5.9
5.7
39.0
3CO monitoring is only available for quarters 1 and 4; since monitoring data are not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-6 contains the distribution of hourly CO measurements reported to AQS for
2 2005-2007. All monitoring locations meeting the 75% data completeness criteria have been included
3 in this table. Several monitors in EPA Region 10 including four in Alaska did not meet the data
4 completeness criteria since CO reporting was only required during the first and fourth quarters of
5 each year at these sites. Anchorage was included in the table, however, for an approximate
September 2009
3-35
DRAFT - DO NOT CITE OR QUOTE
-------
1 comparison with the other CSAs and CBSAs reporting year-round measurements to AQS.
2 Anchorage and other partial-year monitors were not, however, included in the national statistics
3 shown in the table. AQS site number 371190041 located in Charlotte, NC was the only site with
4 collocated monitors both meeting the data completeness criteria and, therefore, the nationwide data
5 in the table was derived from 286 monitors located at 285 sites. In Section 3.5.1.3 below, the
6 nationwide 1-h avg statistics shown in Table 3-6 (along with the nationwide 24-h avg, 1-h daily max
7 and 8-h daily max statistics) are further divided by monitoring scale (microscale, middle scale, etc.)
8 to address issues relating to the near-road environment.
9 The nationwide mean, median, and interquartile range for 1-h measurements reported for
10 2005-2007 were 0.5, 0.4 and 0.4 ppm, respectively, and these statistics did not change by more than
11 0.1 ppm over the 3-year period. The largest recorded second-highest 1-h concentration, 26.3 ppm,
12 for this period was reported in 2006 in Birmingham, AL (AQS site ID: 010736004). The highest 1-h
13 concentration, 39 ppm, between 2005 and 2007, was reported in Ogden, UT (AQS site ID:
14 490570006) on August 28, 2007. An annual outdoor barbeque festival held in Ogden on that day
15 resulted in a period of elevated CO concentrations. The seasonally stratified concentrations in Table
16 3-6 are generally highest in the winter (December-February) and fall (September-November) and
17 decrease on average during the spring (March-May) and summer (June-August).
18 Nationwide statistics pooled by site are listed in the center of Table 3-6 and illustrate the
19 distribution of the site average CO concentrations recorded at the 285 monitoring sites for
20 2005-2007 (see Figure 3-9 for these sites). The site reporting the highest 3-year pooled 1-h avg CO
21 concentration, 1.5 ppm, was located in San Juan, Puerto Rico (AQS site ID: 721270003). The eleven
22 individual CSAs/CBSAs discussed earlier are included in the table, none of which reported
23 concentrations above the value of the 1-h NAAQS. Four of the eleven cities (Boston, Houston,
24 Pittsburgh and St. Louis) had 95th percentile 1-h CO concentrations below 1 ppm; the 95th
25 percentile concentrations for the remaining cities were below 3.1 ppm. Lack of year-round
26 monitoring in Anchorage prevented a direct comparison with the other metropolitan regions.
27 However, Anchorage exhibited a 1-h CO distribution shifted higher in concentration when compared
28 to the U.S. average during fall or winter. The 99th percentile 1-h avg concentration in Anchorage
29 was 5.0 ppm; the other selected cities with year-round monitoring had 99th percentile concentrations
30 ranging from 0.9 ppm to 2.5 ppm.
September 2009 3-36 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-7 Distribution of 24-h avg CO concentration (ppm) derived from AQS data.
Percentiles
n
Mean
Min 1
5
10
25
50
75
90
95
99
Max
NATIONWIDE STATISTICS (N = NUMBER OF OBSERVATIONS)
2005-2007 303,843
2005 101,184
2006 101,652
2007 101,007
Winter (December - 7/11/1/1
February) M'w
Spring (March - May) 77,317
Summer (June - August) 76,562
Fall (September - November) 75,820
0.5
0.5
0.5
0.4
0.6
0.4
0.4
0.5
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.2
0.1
0.1
0.1
0.3
0.3
0.3
0.2
0.3
0.2
0.2
0.3
0.4
0.4
0.4
0.4
0.5
0.4
0.3
0.4
0.6
0.6
0.6
0.5
0.7
0.5
0.5
0.6
0.9
0.9
0.9
0.8
1.1
0.7
0.7
0.9
1.1
1.1
1.1
1.0
1.3
0.9
0.8
1.1
1.7
1.8
1.6
1.6
2.0
1.4
1.3
1.7
7.0
5.8
7.0
6.9
7.0
6.4
6.9
5.8
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007 285
2005 285
2006 285
2007 285
Winter (December - ,Rc
February) "s
Spring (March - May) 285
Summer (June - August) 285
Fall (September - November) 285
0.5
0.5
0.5
0.4
0.6
0.4
0.4
0.5
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.1
0.1
0.1
0.1
0.2
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.4
0.3
0.3
0.4
0.3
0.3
0.4
0.4
0.5
0.4
0.4
0.5
0.4
0.4
0.4
0.6
0.6
0.6
0.5
0.7
0.5
0.5
0.6
0.7
0.8
0.7
0.7
0.9
0.7
0.6
0.8
0.8
0.9
0.8
0.7
1.1
0.7
0.7
0.9
1.0
1.3
1.2
1.1
1.5
1.0
1.1
1.1
1.5
1.6
1.4
1.5
1.6
1.6
1.5
1.5
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007) (N = NUMBER OF OBSERVATIONS)
Anchorage3 1 ,074
Atlanta 3,229
Boston 7,446
Denver 5,363
Houston 5,188
Los Angeles 25,803
New York 9,513
Phoenix 5,348
Pittsburgh 7,497
Seattle 1 ,079
St. Louis 3,216
Not in the 11 cities 230,161
1.1
0.5
0.4
0.5
0.3
0.5
0.8
0.8
0.3
0.8
0.4
0.5
0.0 0.2
0.0 0.1
0.0 0.0
0.0 0.1
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.1
0.0 0.0
0.1 0.2
0.0 0.0
0.0 0.0
0.2
0.2
0.1
0.2
0.0
0.1
0.1
0.2
0.0
0.3
0.1
0.0
0.4
0.2
0.1
0.2
0.1
0.1
0.2
0.3
0.0
0.4
0.2
0.1
0.6
0.3
0.3
0.3
0.2
0.2
0.4
0.4
0.1
0.5
0.3
0.2
0.9
0.4
0.4
0.5
0.3
0.4
0.5
0.6
0.2
0.7
0.4
0.4
1.4
0.6
0.5
0.6
0.4
0.6
0.6
1.1
0.4
0.9
0.5
0.6
1.9
0.8
0.7
0.9
0.5
1.0
0.8
1.6
0.6
1.2
0.7
0.8
2.4
0.9
0.8
1.1
0.6
1.2
1.0
1.9
0.7
1.4
0.8
1.1
3.3
1.2
1.1
1.5
0.9
1.7
1.3
2.5
1.0
1.8
1.0
1.6
4.6
1.6
2.2
2.3
1.9
3.8
2.5
3.4
1.9
2.4
1.9
7.0
3CO monitoring is only available for quarters 1 and 4; since monitoring data are not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-7 contains the distribution of 24-h avg CO concentrations derived from the 1-h
2 concentrations reported to AQS and summarized in Table 3-6. The nationwide mean, median, and
3 interquartile range for 24-h avg values during 2005-2007 were 0.5, 0.4 and 0.3 ppm, respectively.
4 These were similar to those for the 1-h values. The maximum 24-h avg concentration in these years,
September 2009
3-37
DRAFT - DO NOT CITE OR QUOTE
-------
1 7 ppm, was reported in Birmingham, AL (AQS site ID: 010736004). The 99th percentile 24-h avg
2 concentrations ranged from 0.9 ppm to 2.5 ppm in the selected cities with year-round monitoring;
3 Anchorage had a 99th percentile concentration of 3.3 ppm.
Table 3-8 Distribution of 1 -h
daily max CO concentration (ppm) derived from AQS
data.
Percentiles
n
Mean
Min 1
5
10
25
50
75
90
95
99 Max
NATIONWIDE STATISTICS (N = NUMBER OF OBSERVATIONS)
2005-2007
2005
2006
2007
Winter (December - February)
Spring (March - May)
Summer (June -August)
Fall (September - November)
303,843
101,184
101,652
101,007
74,144
77,317
76,562
75,820
0.9
1.0
0.9
0.8
1.2
0.8
0.7
1.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.1
0.2
0.1
0.1
0.2
0.1
0.1
0.2
0.3
0.3
0.3
0.2
0.3
0.3
0.2
0.3
0.4
0.5
0.4
0.4
0.5
0.4
0.4
0.5
0.7
0.8
0.7
0.7
0.9
0.7
0.6
0.8
1.2
1.3
1.2
1.1
1.6
1.0
0.9
1.3
1.8
2.0
1.9
1.7
2.5
1.6
1.3
2.0
2.4
2.6
2.4
2.1
3.1
2.0
1.6
2.5
3.8 39.0
4.1 22.3
3.9 35.3
3.4 39.0
4.7 20.0
3.0 35.3
2.5 39.0
3.8 24.1
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007
2005
2006
2007
Winter (December - February)
Spring (March - May)
Summer (June -August)
285
285
285
285
285
285
285
0.9
1.0
0.9
0.8
1.2
0.8
0.7
0.1
0.1
0.1
0.1
0.0
0.1
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.3
0.4
0.3
0.3
0.4
0.3
0.2
0.5
0.5
0.5
0.4
0.6
0.4
0.3
0.6
0.7
0.6
0.6
0.8
0.6
0.5
0.8
0.9
0.9
0.8
1.0
0.8
0.6
1.1
1.2
1.1
1.0
1.5
1.0
0.8
1.5
1.6
1.6
1.4
2.1
1.3
1.1
1.7
2.0
1.8
1.6
2.5
1.5
1.3
2.3
2.5
2.3
2.0
3.4
2.1
2.2
3.9
3.7
4.8
3.1
4.1
4.0
3.3
Fall (September - November)
285
1.0
0.1
0.1
0.3
0.5
0.7
0.9
1.2
1.7
2.0
2.4 4.1
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007) (N = NUMBER OF OBSERVATIONS)
Anchorage3
Atlanta
Boston
Denver
Houston
Los Angeles
New York
Phoenix
Pittsburgh
Seattle
St. Louis
Not in the 11 cities
1,074
3,229
7,446
5,363
5,188
25,803
9,513
5,348
7,497
1,079
3,216
230,161
2.6
0.8
0.7
1.2
0.7
1.0
0.9
1.9
0.6
1.5
0.8
0.9
0.0
0.0
0.0
0.1
0.0
0.0
0.0
0.0
0.0
0.2
0.0
0.0
0.3
0.2
0.1
0.2
0.0
0.1
0.1
0.3
0.0
0.4
0.1
0.0
0.6
0.3
0.2
0.4
0.1
0.2
0.2
0.5
0.0
0.5
0.3
0.1
0.8
0.3
0.3
0.5
0.2
0.3
0.4
0.6
0.1
0.7
0.4
0.2
1.3
0.4
0.4
0.7
0.4
0.5
0.6
0.9
0.2
0.9
0.5
0.4
2.2
0.7
0.6
1.0
0.6
0.8
0.8
1.6
0.5
1.3
0.6
0.7
3.5
1.1
0.9
1.5
0.8
1.3
1.1
2.5
0.8
1.8
0.9
1.2
5.0
1.4
1.2
2.2
1.3
2.0
1.5
3.5
1.1
2.4
1.3
1.8
6.1
1.7
1.6
2.7
1.7
2.6
1.8
4.1
1.4
2.9
1.7
2.4
7.6
2.2
2.6
3.9
2.6
4.0
2.5
5.3
2.0
4.3
2.7
3.8
13.1
10.8
10.0
9.3
4.6
8.4
5.8
7.8
6.7
5.9
5.7
39.0
3CO monitoring is only available for quarters 1 and 4; since monitoring data are not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
September 2009
3-38
DRAFT - DO NOT CITE OR QUOTE
-------
1 Table 3-8 contains the distribution of 1-h daily max CO concentrations derived from 1-h
2 values reported to AQS for all monitors meeting the inclusion criteria described earlier. The
3 nationwide mean, median, and interquartile range for 1-h daily max concentrations reported for
4 2005-2007 were 0.9, 0.7 and 0.8 ppm, respectively. The 99th percentile 1-h daily max concentrations
5 ranged from 2.0 ppm to 5.3 ppm in the selected cities with year-round monitoring; Anchorage had a
6 99th percentile concentration of 7.6 ppm.
September 2009 3-39 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-9 Distribution of 8-h daily max CO concentration (ppm) derived from AQS data.
Percentiles
Mean Min 1 5 10 25 50 75 90 95 99 Max
NATIONWIDE STATISTICS (N = NUMBER OF OBSERVATIONS)
2005-2007
2005
2006
2007
Winter (December - February)
Spring (March - May)
Summer (June - August)
Fall (September - November)
303,843
101,184
101,652
101,007
74,144
77,317
76,562
75,820
0.7
0.7
0.7
0.6
0.9
0.6
0.5
0.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.3
0.3
0.5
0.6
0.5
0.5
0.7
0.5
0.4
0.6
0.8
0.9
0.8
0.8
1.1
0.7
0.6
0.9
1.3
1.4
1.3
1.2
1.7
1.1
0.9
1.4
1.7
1.8
1.7
1.5
2.1
1.3
1.1
1.8
2.6
2.8
2.6
2.3
3.2
2.0
1.7
2.7
10.9
9.7
9.8
10.9
9.8
9.6
10.9
9.0
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007
2005
2006
2007
Winter (December - February)
285
285
285
285
285
0.7
0.7
0.7
0.6
0.9
0.2
0.3
0.2
0.2
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.4
0.4
0.4
0.4
0.4
0.5
0.5
0.5
0.5
0.6
0.6
0.6
0.6
0.6
0.8
0.8
0.9
0.8
0.7
1.1
1.0
1.1
1.1
1.0
1.4
1.2
1.4
1.2
1.1
1.7
1.7
1.9
1.8
1.6
2.4
2.1
2.2
2.4
2.0
2.6
Spring (March - May)
Summer (June -August)
Fall (September - November)
STATISTICS FOR INDIVIDUAL
Anchorage3
Atlanta
Boston
Denver
Houston
Los Angeles
New York
Phoenix
Pittsburgh
Seattle
St. Louis
Not in the 11 cities
285
285
285
0.6
0.5
0.7
CSAS/CBSAS (2005-2007)
1,074
3,229
7,446
5,363
5,188
25,803
9,513
5,348
7,497
1,079
3,216
230,161
1.7
0.6
0.6
0.8
0.5
0.7
0.7
1.3
0.5
1.1
0.6
0.7
0.2
0.2
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.4
0.4
0.4
0.5
0.5 0.7
0.5 0.6
0.6 0.9
0.9
0.8
1.2
1.1
0.9
1.3
1.6 2.2
1.5 2.0
1.8 2.2
(N = NUMBER OF OBSERVATIONS)
0.3
0.0
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.0
0.3
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.3
0.6
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.5
0.3
0.3
0.9
0.4
0.3
0.5
0.3
0.3
0.4
0.6
0.3
0.7
0.3
0.3
1.5 2.3
0.5 0.8
0.5 0.7
0.7 1.0
0.4 0.6
0.6 0.9
0.6 0.9
1 .0 1 .8
0.3 0.6
1 .0 1 .4
0.5 0.7
0.5 0.8
3.3
1.1
0.9
1.4
0.9
1.5
1.2
2.5
0.9
1.8
0.9
1.3
3.9
1.3
1.1
1.8
1.1
1.8
1.4
3.0
1.0
2.2
1.2
1.6
5.0 6.5
1.7 2.5
1.8 5.8
2.4 3.4
1.7 3.3
2.7 6.2
1.8 3.0
3.8 5.8
1.5 3.7
3.2 4.0
1 .9 4.2
2.5 10.9
3CO monitoring is only available for quarters 1 and 4; since monitoring data is not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-9 contains the distribution of 8-h daily max concentrations derived from the 1-h CO
2 concentrations reported to AQS. This was done by first calculating the average concentration for
3 each successive 8-h period, thereby producing 24 8-h avg per day. The maximum of these values for
4 a given monitor within a given day (midnight-to-midnight) was used as the 8-h daily max statistic
5 for that monitor and day. The nationwide mean, median, and interquartile range for 8-h daily max
September 2009 3-40 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations reported for 2005-2007 were 0.7, 0.5, and 0.5 ppm, respectively. The highest 8-h
2 daily max concentration, 10.9 ppm, was recorded at a monitor located 5 mi north of Newkirk, OK
3 (AQS site ID: 400719010). The 99th percentile 8-h daily max concentrations ranged from 1.5 ppm to
4 3.8 ppm in the selected cities with year-round monitoring; Anchorage had a 99th percentile 8-h daily
5 max concentration of 5.0 ppm.
6 Table 3-7 through Table 3-9 show distributions of CO data based on the 24-h avg, 1-h daily
7 max and 8-h daily max concentration. The current standards are based on 1-h and 8-h calculations.
8 While the nationwide concentrations vary in absolute magnitude based on these three statistics, the
9 shape of the distributions are quite similar up to the 99th percentile. The relative increase from the
10 99th percentile to the max for the 1-h daily max is larger than for the 24-h or 8-h daily max. This is
11 to be expected since this statistic is more sensitive to short term (less than 8 h) increases in CO
12 concentration. Box plots showing the range in Pearson correlation coefficients (r) between the
13 different statistics are shown in Figure 3-16. Included are the correlation of the 24-h avg with the 1-h
14 daily max and 8-h daily max as well as the correlation between the 1-h daily max and 8-h daily max,
15 all calculated using the same 2005 2007 data set stratified by season. Correlations are generally quite
16 high across all seasons and all comparisons with medians above 0.8. Correlations are higher on
17 average in the wintertime compared to the summertime for the two comparisons involving the 1-h
18 daily max statistic. The correlations between the 24-h avg and the 8-h daily max are the highest in all
19 seasons, which is in agreement with the distributional similarities shown in the preceding tables.
September 2009 3-41 DRAFT - DO NOT CITE OR QUOTE
-------
Winter
Spring
daily avg vs. daily max 1h -
daily avg vs. daily max 8h -
daily max 1 h vs. daily max 8h -
— r-Hrr-
-r-SH-
If
—HHH-
0.0 0.2
0.4 0.6 0.8 1.0
Summer
0.0 0.2 0.4 0.6 0.8 1.0
Fall
daily avg vs. daily max 1h -
daily avg vs. daily max 8h -
daily max 1 h vs. daily max 8h -
—r-ffl-r-
• --I-HH-
H-
--H1H-
•-HH-
—h-HH-
0.0 0.2 0.4 0.6 0.8
correlation (r)
1.0
0.0 0.2 0.4 0.6 0.8
correlation (r)
1.0
Figure 3-16 Seasonal plots showing the variability in correlations between 24-h avg CO
concentration with 1-h daily max and 8-h daily max CO concentrations and
between 1-h daily max and 8-h daily max CO concentrations. Red bars denote the
median, green stars denote the arithmetic mean, the box incorporates the IQR
and the whiskers extend to the 5th and 95th percentiles. Correlations outside the
5th and 95th percentiles are shown as individual points.
3.5.1.2. Urban Scale
1 This section describes urban variability in CO concentrations reported to AQS at the individual
2 CSA/CBSA level. Denver, CO and Los Angeles, CA were selected for this assessment to illustrate
3 the variability in CO concentrations measured across contrasting metropolitan regions. Information
4 on the other nine cities evaluated for this assessment is included in Appendix A. Maps of the Denver
5 CSA and Los Angeles CSA shown in Figure 3-17 and Figure 3-19, respectively, illustrate the
6 location of all CO monitors meeting the inclusion criteria described earlier. Letters on the maps
7 identify the individual monitor locations and correspond with the letters provided in the
8 accompanying concentration box plots (Figure 3-18 and Figure 3-20) and pair-wise monitor
9 comparison tables (Table 3-10 and Table 3-11). The box plots for each monitor include the hourly
10 CO concentration median and interquartile range with whiskers extending from the 5th to the 95th
September 2009
3-42
DRAFT - DO NOT CITE OR QUOTE
-------
1 percentile. Data from 2005-2007 were used to generate the box plots, which are stratified by season
2 as follows: 1 = winter (December-February), 2 = spring (March-May), 3 = summer (June-August),
3 and 4 = fall (September-November). The comparison tables include the Pearson correlation
4 coefficient (r), the 90th percentile of the absolute difference in concentrations (P90) in ppm, the
5 coefficient of divergence (COD) and the straight-line distance between monitor pairs (d) in km. The
6 COD provides an indication of the variability across the monitoring sites within each CSA/CBSA
7 and is defined as follows:
Equation 3-1
8 where X]_j andXik represent the observed hourly concentrations for time period / at sites j and k, and/?
9 is the number of paired hourly observations. A COD of 0 indicates there are no differences between
10 concentrations at paired sites (spatial homogeneity), while a COD approaching 1 indicates extreme
11 spatial heterogeneity. Similar maps, box plots, and comparison tables for the nine remaining
12 CSAs/CBSAs are included in Annex A.
13 The information contained in these figures and tables should be used with some caution since
14 many of the reported concentrations for the years 2005-2007 are near or below the monitors' stated
15 lowest detection limits. Because ambient concentrations are now in large part very near the detection
16 limit for the majority of FRMs of 0.5 ppm and the coarsely reported measurement resolution is
17 0.1 ppm, the comparison statistics shown in these tables might be biased to exhibit specious
18 heterogeneity in the box plots.
September 2009 3-43 DRAFT - DO NOT CITE OR QUOTE
-------
Denver Combined Statistical Area
01
• Denver CO Monitors
Denver Major Highways
Denver
0 15 30 60
90
120
• Kilometers
Figure 3-17 Map of CO monitor locations and major highways for Denver, CO.
September 2009
3-44
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
08-001-3001
08-013-0009 08-031-0002
08-031-0019
08-123-0010
Scale
Neighborhood Micro
Micro
Micro
Neighborhood
Mean
0.52
Obs
25920
SD
0.36
0.42
0.65
0.52
0.55
25559
25959
25552
26048
0.38
0.42
0.46
0.46
3-
1f 2"
o
O
o
"(5
concenti
0-
I
I
1 1
*1
,
i i
1234 1 234 1 234 1234 1 234
season
Figure 3-18 Box plots illustrating the distribution of 2005-2007 hourly CO concentrations in
Denver, CO. The data are stratified by season along the x-axis where 1 = winter, 2
= spring, 3 = summer, and 4 = fall. The box plots show the median and
interquartile range with whiskers extending from the 5th to the 95th percentile.
Identifiers and statistics for each site are shown at the top of the figure.
September 2009
3-45
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-10 Table of inter-sampler comparison statistics, as defined in the text, including Pearson r,
P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to AQS for
2005-2007 in Denver, CO. The table is grouped and identified by monitoring scale.
A
A 1.00
0.0
0.00
0
B
C
o
.2
D
Micro
B
0.76
0.5
0.34
1.3
1.00
0.0
0.00
0
Neighborhood
C
0.46
0.7
0.44
46.9
0.49
0.7
0.47
47.0
1.00
0.0
0.00
0
D
0.45
0.7
0.36
78.3
0.46
0.7
0.42
79.0
0.54
0.6
0.43
44.6
1.00
0.0
0.00
E
0.59
0.6
0.29
10.1
0.64
0.5
0.37
10.9
0.53
0.6
0.43
38.5
0.52
0.6
0.34
* E
o
.£
.Q
.£
'o
z
Legend
R
P90
COD
d
0 68.2
1.00
0.0
0.00
0
September 2009
3-46
DRAFT - DO NOT CITE OR QUOTE
-------
Los Angeles Combined Statistical Area
0 15 30
• Los Ange les CO Mon Itore
Los Angeles Msjor Highways
Los Angeles
60 90 120
• ~^^^^™ Kilometers
Figure 3-19 Map of CO monitor locations and major highways for Los Angeles, CA.
September 2009
3-47
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
06-037-
0002
06-037-
0113
06-037-
1002
06-037-
1103
06-037-
1201
06-037-
1301
06-037-
1701
06-037-
2005
Scale
Null
Null
Null
Null
Null
Middle
Null
Null
Mean
0.42
0.41
0.66
0.56
0.57
0.98
0.69
0.72
Obs
2,5001 24916
24892
24645
24281
24825
24912
24804
SD
0.27
0.36
0.59
0.50
0.54
0.89
0.45
0.48
E
CL
Q.
O
4.5-
4,0-
3.5-
3.0-
2.5-
2.0-
1.5-
1.0-
0.5-
o.o-
1
n
I I
ll
i
I
'1
i i
1
I i
1
ll
I i
I i
I
1
I
234 1234 1234 1234 1234 1234 1234 1234
season
Figure 3-20 Box plots illustrating the distribution of 2005-2007 hourly CO concentrations in
Los Angeles, CA. The data are stratified by season along the x-axis where 1 =
winter, 2 = spring, 3 = summer, and 4 = fall. The box plots show the median and
interquartile range with whiskers extending from the 5th to the 95th percentile.
Identifiers and statistics for each site are shown at the top of the figure (monitors
without scale designations in AQS are labeled Null). Part 1 of 3 of Figure 3-20.
See the next two pages for parts 2 and 3 of figure 3-20.
September 2009
3-48
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
M
06-037-
4002
06-037-
5005
06-037-
6012
06-037-
9033
06-059-
0007
06-059-
1003
06-059-
2022
06-059-
5001
Scale
Null
Neighbor- Nu||
Middle Urban
Middle Null
Null
Mean
0.69
0.24
0.30
0.23
0.42
0.31
0.26
0.62
Obs
24259
24965
24860
24135
24264
24760
24831
24705
SD
0.56
0.37
0.25
0.29
0.46
0.47
0.25
0.55
4.5-
4.0-
3.5-
-p 3.0-
E
Q.
Q.
^T 2.5-
g
15
•g 2.0-
8
c
o 1.5-
1.0-
0.5-
o.o-
L
i i
III
i i
ll
1
1234 1234 1234 1
hi
i i i
i
I
n i i
ul
i i i i
li
n i
234 1234 1234 1234 1234
season
Part 2 of 3 for Figure 3-20
September 2009
3-49
DRAFT - DO NOT CITE OR QUOTE
-------
w
Site ID
Scale
Mean
Obs
SD
06-065-
1003
Micro
0.67
24885
0.42
06-065-
5001
Null
0.25
24938
0.14
06-065-
8001
Null
0.60
24778
0.46
06-065-
9001
Neighbor-
hood
0.29
24792
0.20
06-071-
0001
Null
0.17
24105
0.17
06-071-
0306
Null
0.30
24796
0.28
06-071-
1004
Null
0.59
24767
0.32
06-071-
9004
Middle
0.53
24844
0.38
Q.
Q.
.Q
"S
•g
4.5-
4.0-
3.5-
3.0-
2.5-
2.0-
1.5-
1.0-
0.5-
o.o-
I,
II
I I
--
I I I I
I
i
I
i i
1234 1234 1234 1234 1
•--
i i i
234 1
l!
i
I
234 1
II
II
I I
I ,
H
I l
234 1234
season
Part 3 of 3 for Figure 3-20
September 2009
3-50
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-11 Table of inter-sampler comparison statistics, as defined in the text, including Pearson r,
P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to AQS for
2005-2007 in Los Angeles, CA. The table is grouped and identified by monitoring
scale (monitors without scale designations in AQS are labeled Null).
Mic-
ro
A B
A 1.00 0.56
2 0.0 0.8
jg 0.00 0.66
0 57.1
B 1.00
0.0
0.00
0
C
€ D
E
F
•a
0
0
£ G
Z Lea
H P£
!i ^
i
j
K
L
z M
N
0
P
Middle NJ0
C D E F
0.56 0.54 0.73 0.72
0.9 1.1 0.5 0.7
0.67 0.30 0.30 0.46
104.6 74.8 21.3 30.5
0.55 0.67 0.50 0.46
0.6 1.3 0.7 0.5
0.66 0.70 0.64 0.59
112.0 38.6 76.9 55.1
1.00 0.55 0.50 0.50
0.0 1.6 0.7 0.4
0.00 0.72 0.64 0.57
0 82.5 100.4 132.5
1.00 0.44 0.39
0.0 1.3 1.6
0.00 0.42 0.56
0 88.6 86.0
1.00 0.69
0.0 0.6
0.00 0.42
0 48.0
1.00
0.0
0.00
0
0
u
od
G
0.45
0.9
0.73
95.0
0.60
0.5
0.69
55.8
0.39
0.6
0.74
84.4
0.63
1.5
0.76
20.4
0.43
0.8
0.72
108.0
0.43
0.5
0.70
106.1
1.00
0.0
0.00
0
Ur-
ban
H
0.62
0.7
0.46
51.3
0.75
0.5
0.55
17.3
0.56
0.7
0.60
94.7
0.71
1.1
0.51
27.4
0.55
0.6
0.46
68.5
0.53
0.6
0.42
58.7
0.58
0.6
0.71
47.4
1.00
0.0
0.00
0
No Scale Identified
I
0.54
0.7
0.33
52.6
0.14
0.7
0.64
51.2
0.27
0.6
0.62
62.2
0.21
1.5
0.44
35.0
0.48
0.6
0.35
59.9
0.56
0.4
0.38
74.8
0.19
0.7
0.72
51.0
0.29
0.6
0.43
33.9
1.00
0.0
0.00
0
J
0.35
1.0
0.68
110.6
0.28
0.7
0.69
158.5
0.45
0.4
0.62
104.0
0.33
1.7
0.73
152.6
0.31
0.8
0.65
90.2
0.30
0.4
0.58
137.8
0.18
0.6
0.75
166.0
0.31
0.8
0.62
144.7
0.17
0.6
0.62
117.7
1.00
0.0
0.00
0
K
0.70
0.6
0.29
88.2
0.70
0.9
0.63
66.3
0.59
1.1
0.65
57.4
0.70
0.9
0.35
29.1
0.65
0.7
0.35
96.3
0.58
1.0
0.46
106.5
0.64
1.0
0.72
27.0
0.72
0.7
0.41
51.8
0.43
1.0
0.35
36.5
0.35
1.2
0.67
142.7
1.00
0.0
0.00
0
L
0.66
0.6
0.24
51.0
0.72
0.7
0.62
27.9
0.63
0.9
0.64
84.2
0.78
0.9
0.30
23.8
0.55
0.7
0.33
65.7
0.55
0.8
0.43
63.7
0.59
0.8
0.72
44.2
0.81
0.5
0.38
10.5
0.34
0.8
0.33
23.6
0.34
1.0
0.66
137.1
0.75
0.6
0.26
43.6
1.00
0.0
0.00
0
M
0.46
0.8
0.39
74.1
0.64
0.9
0.64
29.5
0.43
1.1
0.69
94.0
0.70
1.0
0.39
11.8
0.39
0.9
0.46
90.1
0.36
1.1
0.54
81.1
0.59
1.0
0.75
26.4
0.63
0.8
0.48
23.2
0.17
1.1
0.47
42.4
0.24
1.3
0.70
159.7
0.62
0.8
0.41
40.8
0.62
0.7
0.37
24.6
1.00
0.0
0.00
0
N
0.62
0.6
0.37
77.4
0.58
0.8
0.63
51.6
0.53
0.9
0.63
67.5
0.74
1.0
0.41
15.3
0.58
0.6
0.39
87.9
0.53
0.8
0.46
93.4
0.59
0.8
0.73
22.7
0.70
0.6
0.40
37.3
0.46
0.8
0.38
29.0
0.34
1.0
0.66
143.4
0.84
0.5
0.29
14.7
0.74
0.5
0.31
29.8
0.60
0.8
0.45
27.1
1.00
0.0
0.00
0
0
0.61
0.8
0.58
43.3
0.62
0.5
0.56
23.7
0.51
0.4
0.61
122.7
0.57
1.5
0.65
59.5
0.51
0.6
0.56
64.6
0.51
0.3
0.50
32.4
0.42
0.5
0.73
78.3
0.67
0.5
0.46
32.9
0.42
0.5
0.52
60.6
0.33
0.4
0.63
152.3
0.69
1.0
0.56
84.6
0.67
0.8
0.54
41.5
0.48
1.1
0.58
52.1
0.63
0.8
0.55
70.2
1.00
0.0
0.00
0
P Q
0.48 0.53
0.9 0.7
0.47 0.35
80.1 70.1
0.34 0.50
0.6 0.9
0.59 0.68
129.6 54.0
0.41 0.45
0.4 1.0
0.56 0.70
171.9 59.6
0.28 0.53
1.7 1.0
0.56 0.39
154.5 23.8
0.46 0.52
0.7 0.7
0.43 0.41
73.3 78.6
0.49 0.47
0.3 0.9
0.32 0.53
75.7 89.2
0.26 0.40
0.5 1.0
0.70 0.75
174.8 34.4
0.37 0.54
0.7 0.8
0.41 0.52
129.2 37.7
0.33 0.41
0.5 0.8
0.36 0.43
131.4 18.7
0.29 0.31
0.3 1.1
0.55 0.71
123.6 131.7
0.40 0.67
1.2 0.7
0.46 0.39
167.7 18.1
0.40 0.58
1.0 0.7
0.42 0.37
130.6 28.1
0.24 0.41
1.2 0.9
0.53 0.46
152.4 34.7
0.32 0.67
1.0 0.7
0.46 0.44
157.4 11.7
0.44 0.55
0.3 0.9
0.47 0.62
107.9 69.5
1.00 0.39
0.0 1.0
0.00 0.54
0 149.6
R
0.78
0.4
0.19
35.0
0.60
0.9
0.66
46.3
0.61
0.9
0.66
75.4
0.65
1.0
0.29
45.0
0.68
0.6
0.31
44.2
0.69
0.8
0.46
58.1
0.51
1.0
0.73
63.9
0.69
0.7
0.44
31.4
0.57
0.8
0.33
17.7
0.40
1.1
0.68
113.3
0.78
0.5
0.26
53.5
0.77
0.5
0.20
24.3
0.52
0.8
0.38
48.6
0.77
0.5
0.33
43.8
0.70
0.9
0.58
48.9
0.47
1.0
0.47
114.2
S T
0.73 0.67
0.6 0.6
0.29 0.37
108.0 6.1
0.57 0.40
0.8 0.9
0.62 0.67
80.7 59.3
0.53 0.41
0.9 0.9
0.62 0.67
64.0 99.2
0.49 0.35
1.2 1.2
0.41 0.45
42.1 73.7
0.71 0.64
0.6 0.6
0.32 0.41
116.3 17.7
0.66 0.66
0.8 0.8
0.40 0.49
125.1 36.6
0.52 0.43
0.9 0.9
0.72 0.74
29.1 93.7
0.61 0.49
0.7 0.8
0.41 0.49
68.3 51.7
0.44 0.47
0.7 0.8
0.33 0.43
56.5 49.2
0.31 0.19
1.0 1.1
0.64 0.69
158.2 105.4
0.74 0.52
0.6 0.8
0.28 0.42
20.0 85.3
0.61 0.50
0.7 0.8
0.29 0.38
61.5 50.2
0.44 0.31
0.9 1.0
0.44 0.48
52.3 74.0
0.64 0.49
0.7 0.8
0.34 0.45
31.8 75.1
0.55 0.43
0.8 0.9
0.54 0.59
101.2 47.5
0.42 0.40
0.9 0.9
0.40 0.50
187.6 82.4
U
0.54
0.8
0.53
114.5
0.25
0.7
0.67
96.2
0.40
0.5
0.62
48.4
0.23
1.6
0.60
58.2
0.50
0.6
0.51
119.3
0.56
0.3
0.47
135.3
0.24
0.6
0.73
48.7
0.29
0.7
0.53
81.8
0.43
0.5
0.47
61.9
0.25
0.5
0.62
148.8
0.39
1.1
0.52
30.1
0.34
0.9
0.52
73.4
0.15
1.2
0.60
69.4
0.34
0.9
0.52
44.7
0.28
0.5
0.59
114.6
0.40
0.4
0.45
192.2
V
0.70
0.5
0.20
27.4
0.36
0.8
0.65
54.9
0.49
0.7
0.65
77.9
0.39
1.3
0.35
57.0
0.64
0.5
0.29
32.7
0.68
0.6
0.43
54.8
0.27
0.9
0.73
75.8
0.47
0.7
0.45
41.6
0.61
0.5
0.30
27.4
0.36
0.8
0.66
103.7
0.59
0.7
0.29
63.8
0.54
0.6
0.25
35.8
0.28
0.9
0.41
60.3
0.57
0.6
0.36
55.2
0.57
0.7
0.56
52.6
0.47
0.7
0.43
104.2
W X
0.55 0.57
0.8 0.7
0.54 0.42
62.8 98.1
0.55 0.47
0.6 0.7
0.62 0.59
107.5 64.4
0.67 0.42
0.4 0.7
0.57 0.60
75.4 74.9
0.51 0.50
1.5 1.3
0.61 0.50
103.4 26.4
0.51 0.53
0.6 0.6
0.52 0.41
44.9 109.1
0.49 0.55
0.4 0.5
0.47 0.39
92.3 112.0
0.41 0.59
0.6 0.6
0.72 0.69
118.6 11.4
0.54 0.59
0.6 0.5
0.50 0.39
93.7 53.7
0.24 0.45
0.6 0.5
0.50 0.38
68.4 50.0
0.43 0.25
0.4 0.7
0.59 0.62
51.0 161.1
0.58 0.69
1.0 0.8
0.53 0.38
97.8 18.9
0.58 0.59
0.8 0.6
0.51 0.37
86.4 48.5
0.43 0.43
1.1 0.9
0.59 0.48
109.6 35.2
0.51 0.71
0.8 0.6
0.53 0.38
95.9 21.2
0.45 0.57
0.4 0.6
0.56 0.50
102.5 85.9
0.38 0.35
0.4 0.6
0.44 0.38
102.8 178.1
September 2009
3-51
DRAFT - DO NOT CITE OR QUOTE
-------
"JJ- Middle NJ0h0bor JJ- No Scale Identified
ABCDEFGH 1 JKLMNOPQRS
Q 1.00 0.65 0.54
0.0 0.6 0.8
0.00 0.34 0.42
0 35.4 38.0
R 1.00 0.70
0.0 0.6
0.00 0.30
0 73.4
S 1.00
0.0
0.00
0
T
U
V
w
X
T
0.39
0.8
0.46
67.2
0.60
0.7
0.38
31.8
0.62
0.7
0.40
105.2
1.00
0.0
0.00
0
U
0.32
1.0
0.58
46.2
0.47
0.9
0.53
79.6
0.53
0.8
0.49
20.4
0.46
0.9
0.56
110.8
1.00
0.0
0.00
0
V
0.53
0.7
0.35
46.0
0.78
0.5
0.18
12.0
0.58
0.6
0.30
83.8
0.54
0.6
0.37
22.8
0.53
0.6
0.51
88.3
1.00
0.0
0.00
0
W
0.46
0.9
0.59
84.3
0.58
0.9
0.54
62.4
0.55
0.8
0.50
115.6
0.38
0.9
0.58
57.0
0.39
0.5
0.54
110.7
0.47
0.7
0.52
52.7
1.00
0.0
0.00
0
X
0.49
0.8
0.49
31.6
0.63
0.7
0.41
65.0
0.64
0.7
0.34
17.8
0.55
0.7
0.46
96.1
0.36
0.6
0.50
37.4
0.50
0.6
0.40
76.4
0.41
0.6
0.50
115.3
1.00
0.0
0.00
0
1 The Denver CSA in Figure 3-17 incorporates an area of 33,723 km2 with a maximum straight-
2 line distance between CO monitors of 79 km. Of the five CO monitors meeting the inclusion criteria,
3 three were sited for microscale monitoring and two were sited for neighborhood scale monitoring.
4 Sites A and B are located in downtown Denver while Site E is located in an industrial region north of
5 town and surrounded on three sides by three heavily-traveled interstate highways. Sites C and D are
6 located in two smaller towns (Longmont and Greeley, respectively) north of Denver. The means and
7 seasonal patterns shown in Figure 3-18 are similar for all five monitors within this CSA. The highest
8 annual mean concentration (0.7 ppm) was observed at Site A, a downtown microscale monitor, while
9 the lowest annual mean concentration (0.4 ppm) was observed at Site C, a microscale monitor in
10 Longmont.
11 The Los Angeles CSA in Figure 3-19 incorporates an area of 88,054 km2 and a maximum
12 straight-line distance between monitors of 192 km, making it more than twice the size of the Denver
13 CSA. Of the eleven CS As/CBS As investigated, Los Angeles had the largest number of CO monitors
14 (N = 24) meeting the inclusion criteria. One monitor was sited for microscale, four for middle scale,
15 two for neighborhood scale and one for urban scale. The remaining 16 monitors did not contain a
16 siting classification in AQS. The monitors were evenly distributed around the Los Angeles and
17 Riverside areas with outlying monitors in Santa Clarita (Site U), Lancaster (Site C), Victorville (Site
18 W), Barstow (Site J) and Palm Springs (Site P). A large amount of variability is present in the means
September 2009 3-52 DRAFT - DO NOT CITE OR QUOTE
-------
1 and seasonal patterns displayed in Figure 3-20. Generally speaking, lower annual mean
2 concentrations (< 0.3 ppm) were measured in the outlying towns including those listed above as well
3 as Lake Elsinore (Site F) and Mission Viejo (Site O). In addition, a neighborhood scale upwind
4 background site (Site G) located on the grounds of the Los Angeles International Airport and 1.5 km
5 from the Pacific Ocean reported a relatively low mean annual concentration of 0.2 ppm. The highest
6 annual mean concentration (1.0 ppm) was observed at Site D, a middle scale maximum
7 concentration site located 25 m from a busy surface street and adjacent to the Imperial Shopping
8 Mall. This site is also 180 m from a major highway intersection and 350 m from Interstate 105.
9 The pair-wise comparisons for measurements at the monitors in each of the eleven
10 CS As/CBS As included in this analysis reveal a wide range of response between monitors in each
11 city and among the cities judged against each other (see Table 3-10, Table 3-11 and Annex Tables A-
12 9 through A-16). While this wide range is produced by the interactions of many physical and
13 chemical elements, the location of each monitor and the uniqueness of its immediate surroundings
14 can often explain much of the agreement or lack thereof.
15 For the monitor comparisons within the Denver CBSA (Table 3-10, the correlations tend to be
16 inversely related to the monitor separation distance, with the highest correlation (r = 0.76) for the
17 two downtown Denver monitors (Sites A and B) separated by 1.3 km and the lowest correlations (r <
18 0.46) between the downtown Denver monitors and the Greeley monitor (Site D) located roughly 80
19 km north. While Sites A and B have a high correlation, the comparative magnitudes of the
20 concentrations measured at these two sites—as determined by the P90 and COD—is comparable to
21 comparisons with much less proximal monitors. This is likely caused by the location of these two
22 monitors on opposite sides of downtown Denver, as illustrated by the aerial view of monitors A and
23 B in Figure 3-21. While there is no prevailing wind direction in Denver, the wind comes from the
24 south-southwest with a slightly higher frequency than other directions, making Site A downwind of
25 the urban core more frequently than Site B. Assuming traffic within the urban core is a major source
26 of CO, this would explain the higher mean concentrations measured at Site A relative to Site B
27 despite their close proximity.
September 2009 3-53 DRAFT - DO NOT CITE OR QUOTE
-------
KS^ySv* • ••
^P^lp1
^i^l^'^X r<
-<$%AV 0.62) with Sites A, E,
September 2009
3-54
DRAFT - DO NOT CITE OR QUOTE
-------
1 F and T, all east of Los Angeles and all over 100 km away. Site S is located in a densely populated
2 urban area with a mixture of commercial and residential land whereas the other four sites are located
3 in less densely populated regions with commercial, residential and undeveloped land. Sites S and T
4 contain no monitoring scale information in AQS, but Sites A, E and F are classified as microscale,
5 middle scale and neighborhood scale, respectively. In contrast to the above example, Sites I and Q
6 are located only 19 km apart in Azusa and Pasadena, respectively, and they correlate less well (r =
7 0.41). While these two locations are relatively close in proximity with similar topography, the siting
8 of the two monitors is quite different. Site I in Azusa is located 700 m from 1-210 in a mixed use
9 community containing warehouses, small industry, housing and a gravel operation (see Figure 3-22)
10 while Site Q in Pasadena is located between a large residential neighborhood and the California
11 Institute of Technology campus (see Figure 3-23). Neither of these sites has monitoring scale
12 designations reported in AQS. The contrasting CO emission sources surrounding these two monitors
13 result in disparate concentrations with poor correlations despite their close proximity. Topography
14 and micrometeorology can also play an important role in the correlation between monitors. For
15 example, Sites C and P are isolated from the other sites in the Los Angeles CSA by the San Gabriel
16 Mountains and the San Bernardino Mountains, respectively, resulting in lower than average
17 concentrations (Figure 3-20) and relatively low pair-wise correlations (Table 3-11) for these two
18 sites. This analysis demonstrates that agreement between monitors on an urban scale is a complex
19 function of monitor siting, location relative to sources, geography, and micrometeorology.
September 2009 3-55 DRAFT - DO NOT CITE OR QUOTE
-------
Figure 3-22 Aerial view of the location of CO monitor I (marked by the red pin) in Azusa, CA
(Los Angeles CSA), depicting its proximity to mixed use land.
September 2009
3-56
DRAFT - DO NOT CITE OR QUOTE
-------
Figure 3-23 Aerial view of the location of CO monitor Q (marked by the red pin) in Pasadena,
CA (Los Angeles CSA), depicting its proximity to a residential neighborhood.
3.5.1.3. Micro- to- Neighborhood Scale and the Near-Road Environment
1 Table 3-12 shows the 2005-2007 nationwide distributional data for all hourly, 1-h daily max,
2 1-h daily avg, and 8-h daily max CO concentrations broken down by spatial sampling scale. The
3 different sampling scales included in the table—microscale, middle scale, neighborhood scale and
4 urban scale—were defined in Section 3.4.2.1. While monitors classified under all four scales are
September 2009
3-57
DRAFT - DO NOT CITE OR QUOTE
-------
1 used for highest concentration monitoring and regulatory compliance, individual monitors are
2 classified by spatial scale to be used for addressing more particular monitoring objectives.
3 Microscale, middle scale and neighborhood scale monitors are used to quantify source impacts while
4 neighborhood scale and urban scale monitors are used for population oriented monitoring (40 CFR
5 Part 58 Appendix D). For CO, traffic is the major source in an urban setting and therefore microscale
6 data are sited "to represent distributions within street canyons, over sidewalks, and near major
7 roadways" while middle scale monitors are sited to represent "air quality along a commercially
8 developed street or shopping plaza, freeway corridors, parking lots and feeder streets" (40 CFR Part
9 58 Appendix D). The data used to create Table 3-12 were subject to the same 75% completeness
10 criteria described in Section 3.5.1.1. More than 50% of the reported hourly data fell below the
11 reported LOD (reported as 0.5 ppm for the majority of monitors reporting to AQS).
September 2009 3-58 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-12 National distribution of all hourly observations, 1-h daily max, 1-h daily average, and 8-h
daily max concentration (ppm) derived from AQS data, based on monitor scale
designations, 2005-2007.
PERCENTILES
Time scale
n
mean
min
1
5
10
25
50
75
90
95
99
max
ALL HOURLY
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
1428745
771941
2878993
279311
0.6
0.5
0.4
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.0
0.2
0.1
0.0
0.0
0.3
0.2
0.2
0.1
0.5
0.4
0.3
0.3
0.8
0.6
0.5
0.5
1.1
1.0
0.8
0.7
1.4
1.3
1.1
0.9
2.2
2.3
2.1
1.6
19.6
18.9
35.3
10.8
1-H DAILY MAX
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
59905
32659
121328
11784
1.2
1.0
0.9
0.7
0.0
0.0
0.0
0.0
0.2
0.1
0.0
0.0
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.7
0.5
0.4
0.3
1.0
0.8
0.6
0.5
1.5
1.2
1.1
0.9
2.1
2.0
1.8
1.3
2.5
2.5
2.4
1.8
3.9
4.0
4.0
3.1
19.6
18.9
35.3
10.8
1-H DAILY AVERAGE
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
59905
32659
121328
11784
0.6
0.5
0.4
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.0
0.0
0.2
0.1
0.1
0.0
0.4
0.3
0.2
0.2
0.5
0.4
0.3
0.3
0.8
0.6
0.5
0.5
1.0
0.9
0.8
0.7
1.2
1.2
1.0
0.8
1.7
1.9
1.6
1.2
4.0
5.5
7.0
2.5
8-H DAILY MAX
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
59905
32659
121328
11784
0.8
0.7
0.6
0.5
0.3
0.1
0.0
0.0
0.3
0.3
0.3
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.5
0.3
0.3
0.3
0.7
0.6
0.4
0.4
1.1
0.9
0.8
0.7
1.5
1.4
1.2
1.0
1.8
1.9
1.6
1.3
2.6
2.8
2.7
2.1
5.8
6.2
10.9
4.0
1 The median hourly CO concentration across the U.S. obtained at microscale monitors was
2 25% higher than at middle scale and 67% higher than at neighborhood scale. However,
3 measurements at or below the median hourly concentration were almost entirely below the LOD for
4 all scales, thereby limiting the usefulness of hourly median comparisons. The upper percentiles (90%
5 and above), however, were all above the LOD and reveal consistently lower hourly concentrations
6 for the urban scale monitors relative to the other monitors. For example, the 99th percentile of
7 reported hourly values was 2.2, 2.3, and 2.1 ppm for microscale, middle scale and neighborhood
8 scale, respectively, compared to 1.6 ppm for urban scale. Similar patterns were present in the 1-h
9 daily max, 1-h daily average, and 8-h daily max distributions. Overall, the urban scale nationwide
10 distributions tended to have lower concentrations relative to neighborhood scale, middle scale and
11 microscale distributions in Table 3-12.
September 2009
3-59
DRAFT - DO NOT CITE OR QUOTE
-------
1 Distributions categorized by spatial scale and CSA/CBSA are provided in Figure 3-24 for
2 hourly data and in Figure 3-25 for 1-h daily max data for the select CSAs/CBSAs where data were
3 available at multiple scales (not all scales were reported by each CSA/CBSA studied). Tables A-17
4 through A-26 of Annex A contain tabular distributions for all CSAs/CBSAs except Anchorage. On a
5 city-by-city basis, there was considerable variability when comparing distributions at the available
6 spatial scales. With a few exceptions, however, the distribution of microscale and middle scale
7 monitors tended to be higher than those obtained from neighborhood and urban scale monitors. For
8 example, in CSAs/CBSAs containing both microscale and neighborhood scale monitors (Boston,
9 Denver, Houston, Los Angeles, New York and Phoenix), median hourly concentrations at monitors
10 sited for microscale were 20-40% higher than for middle scale and 0-150% greater than those sited
11 for neighborhood scales. At the 99th percentile, microscale concentrations ranged from 31% less
12 than to 59% greater than middle scale concentrations and from 14% less than to 67% greater than
13 neighborhood scale. For most cities, the median hourly data are near or below the 0.5 ppm LOD
14 reported for most monitors in use. In general, these data suggest that near road CO concentrations
15 measured with monitors designated at microscale and middle scale locations were somewhat
16 elevated compared with neighborhood and urban scale monitor locations, but the magnitude of these
17 differences varies by city and is difficult to discern given the predominance of CO concentrations
18 near or below the LOD.
19 Despite differences in concentrations observed at different scales in Figure 3-24 and Figure
20 3-25, intersampler correlations do not follow a distinct trend with respect to spatial monitoring scale
21 (see Table 3-10 and Table 3-11). For instance, intersampler correlation in Denver ranged from 0.46
22 to 0.76 among microscale monitors and was 0.52 for the correlation between the two neighborhood
23 scale monitors (no monitors in Denver reporting to the AQS are sited at middle scale). Intersampler
24 correlation in Los Angeles ranged from 0.44 to 0.73 for middle scale and the one pair of
25 neighborhood scale monitors had a correlation of 0.43. Only one monitor was sited each at
26 microscale and urban scale, and 16 of the 24 CO monitors in Los Angeles are not declared to sample
27 at any spatial scale (scale designation = "null"). In Denver, the distribution of hourly CO data
28 obtained at microscale was nearly identical to that obtained at neighborhood scale. In Los Angeles,
29 the microscale data was typically higher than middle, neighborhood, or urban scale data except at the
30 upper end of the distribution, where middle scale data were higher for both hourly and 1-h daily max
31 data (See Figure 3-24 and Figure 3-25).
September 2009 3-60 DRAFT - DO NOT CITE OR QUOTE
-------
Atlanta
Boston
20
40 60
Percentile
80
100
1.0
0.8
0.6
0.4
20
40 60
Percentile
80
100
Denver
Houston
0.0
20
40 60
Percentile
Los Angeles
20
40 60
Percentile
Phoenix
80
100
80
100
20
40 60
Percentile
New York City
20
40 60
Percentile
Pittsburgh
80
100
80
100
0 8
JB 0 6
0 0
//
2
//
^^
. *r**^^~^
0 20 40 60 80 1C
Percentile
Figure 3-24 Distribution of hourly CO co
scale. For comparison purp<
city-specific 99th percentile
Seattle, and St. Louis CSAs
cities do not have monitors
0 8
J
D n 4
o.o -
)0 C
ncentration i
>ses, the y-a,
concentratio
are not inclu
sited at diffe
/
h
^7
~. — -^
^^=^^
) 20 40 60
Percentile
data by city and monitoring
Kis has been scaled to the
n. Note that Anchorage,
ded here because these
rent scales.
^
80 100
— microscale
—middle scale
—neighborhood scale
—urban scale
September 2009
3-61
DRAFT - DO NOT CITE OR QUOTE
-------
Atlanta
Boston
20
40 60
Percentile
80
100
0.0
20
40 60
Percentile
80
100
Denver
Houston
0.0
0.0
20
40 60
Percentile
Los Angeles
20
40 60
Percentile
Phoenix
20
40 60
Percentile
80
100
80
100
80
100
20
40 60
Percentile
New York City
20
40 60
Percentile
Pittsburgh
40 60
Percentile
80
100
80
100
80
100
Figure 3-25
Distribution of 1-h daily max CO concentration data by
city and monitoring scale. For comparison purposes, the
y-axis has been scaled to the city-specific 99th
percentile concentration. Note that Anchorage, Seattle,
and St. Louis CSAs are not included here because these
cities do not have monitors sited at different scales.
— microscale
— middle scale
— neighborhood scale
— urban scale
September 2009
3-62
DRAFT - DO NOT CITE OR QUOTE
-------
1 The microscale and middle scale CO data reported here are consistent with hourly
2 concentrations reported in the literature for the near road environment within the United States.
3 Baldauf et al. (2008, 190239) reported CO concentrations obtained by an open-path Fourier
4 transform infrared spectrometer 20 m from an interstate highway in Raleigh, NC to have a median
5 around 0.25 ppm and with maximum concentration less than 2.0 ppm. Zhu et al. (2002, 041553)
6 reported CO concentration of 1.9-2.6 ppm at a distance of 17 m from an interstate highway in Los
7 Angeles, with concentration decreasing exponentially with distance from the highway. Zhu et al.
8 (2002, 041553) observed on-road CO concentrations to be approximately 10 times higher than at an
9 upwind monitoring site, as shown in Figure 3-26. Concentrations continued to decrease and were
10 still two times higher than upwind levels at a monitoring site 300 m away. Baldauf et al. (2008,
11 190239) also reported a drop in concentration at a monitoring site 300 m from the road compared
12 with the 20 m site. Figure 3-27 illustrates the distribution of measurements taken throughout a day.
13 In this plot, the near-road (20 m distance) CO concentrations tend to be significantly higher than
14 those obtained at 300 m, and the daily variability in the CO concentration time series is greater at the
15 20 m site than at the 300 m site. The ratio of 20 m to 200 m concentrations is higher for the Zhu
16 et al. (2002, 041553) paper. This is likely due to the fact that the 300 m site was always downwind in
17 Zhu et al. (2002, 041553). whereas winds were more variable in Baldauf et al. (2008, 190239).
18 Chang et al. (2000, 001276) reported near-road ambient CO measurements obtained in downtown
19 Baltimore (distance to road not specified) in the range of 0.5-1.3 ppm. Riediker et al. (2003, 043761)
20 reported measurements of CO concentration obtained near one of four heavily-trafficked roads in
21 Wake County, NC to average 1.1 ppm (range: 0.4-1.7 ppm). Neighborhood scale measurements
22 reported in the literature were also consistent with if not slightly lower than those reported by AQS.
23 Gentner et al. (2009, 194034) reported CO concentrations ranging from roughly 0.4-0.9 ppm in
24 Riverside, CA. Singh et al. (2006, 190136) reported 24-h avg CO concentrations obtained with a
25 trace-level CO monitor in Long Beach, CA within 0.5 km of 1-405 and 1.5 km of 1-710 to range from
26 0.2-1.4 ppm.
September 2009 3-63 DRAFT - DO NOT CITE OR QUOTE
-------
1.0
0.0
Upwind -200
-100 0 100 200
Distance to the 710 Freeway (m)
300 Downwind
Source: Zhu et al. (2002, 041553) (Zhu et al., 2002, 041553)
Figure 3-26 Relative concentrations of CO and copollutants at various distances from the I-
710 freeway in Los Angeles.
2.0
I
(/)
d
o
"ro
"c
0)
o
o
O
O
O
1.5
1.0 -
0.5
0.0
• 20 m site
a 300 m site
4:00 6:00 8:00 10:00 12:00 14:00
Time of day (hrmin)
16:00 18:00
Source: Baldauf et al. (2008,190239)
Figure 3-27 CO concentration time series 20 m and 300 m from the I-440 highway in Raleigh,
NC.
1 Determinants of spatial variability in ambient CO concentration include roadway density.
2 traffic counts, meteorology, and natural and urban topography. Mobile sources are the largest single
3 source of CO, and their abundance and density affect the magnitude of CO production. Rodes et al.
4 (1998, 010611) compared traffic volume, roadway type, and concentrations of CO and several
September 2009
3-64
DRAFT - DO NOT CITE OR QUOTE
-------
1 copollutants in Los Angeles and Sacramento, CA in a study of on-road traffic emissions. They noted
2 that there was little difference in CO concentration between arterial roads and freeways for Los
3 Angeles. Rodes et al. (1998, 010611) found that traffic was also much more congested throughout
4 Los Angeles, not surprisingly given that Los Angeles is a much larger city with substantially higher
5 traffic volumes than Sacramento. Under similar wind conditions, morning concentrations were much
6 higher in Los Angeles than Sacramento. Rodes et al. (1998, 010611) observed that high afternoon
7 winds ventilate Los Angeles, but Sacramento is not as well ventilated. As a result, Sacramento has
8 nearly the same concentrations as Los Angeles in the afternoon. This observation is consistent with
9 measurements by Gentner et al. (2009, 194034) showing that CO concentrations varied inversely
10 with wind speed.
11 The size of the gradient between on-road or road-side CO concentrations and what is
12 measured outside a home in the near-road environment may relate to the traffic volume. Among the
13 291 active sites where monitors met completeness criteria during 2005-2007, 57 were declared by
14 state agencies as microscale with average annual daily traffic (AADT) counts on the nearby roads
15 ranging from 500 vehicles per day at one site in Denver, CO to 133,855 vehicles per day in Tampa,
16 FL with a geometric mean of 17,462 vehicles per day and a geometric standard deviation of 2.5; see
17 Table A-2 of Annex A. Within a geometric standard deviation, the data range from 6,576-40,000
18 vehicles per day. Only two monitors were sited at roads with 100,000 vehicles per day or more. In
19 contrast, the site where Zhu et al. (2002, 041553) collected data had 160,000-178,000 vehicles per
20 day in 2001 (CalTrans, 2009, 194036). Microscale sites near roads in the mid-range of the traffic
21 count data may record data that are not substantially different from those obtained from
22 neighborhood scale measurements, as indicated in Table 3-12. Likewise, with little microscale data
23 at roads with AADT of more than 100,000 vehicles per day, there is still much uncertainty regarding
24 the size of concentration gradients in the near-road environment.
25 Field measurements, computational modeling, and wind tunnel experiments have shown that
26 roadway design, roadside structures and vegetation, and on-road traffic levels can affect
27 concentrations of CO and other pollutant concentrations near roadways. Field measurements
28 reported by Baldauf et al. (2008, 191017) indicated that noise barriers could reduce near-road
29 pollutant concentrations by as much as 50 percent, although this effect was highly dependent on
30 meteorological conditions; these results are illustrated in Figure 3-28. This study also showed that
31 the presence of mature vegetation further reduced concentrations and flattened the concentration
32 gradient away from the road. Urban dispersion and wind-field modeling by Bowker et al. (2007,
33 149997) also demonstrated the influence of noise barriers and vegetation on the concentrations and
34 spatial variability of nonreactive pollutants emitted from traffic sources. Heist et al (2009, 194037)
35 ran wind tunnel experiments using a model of a road with different roadside features and a tracer gas
36 line source emitted from the simulated road to study how concentrations of gaseous traffic emissions
September 2009 3-65 DRAFT - DO NOT CITE OR QUOTE
-------
1 vary spatially in the near-road environment. They demonstrated that noise barriers and roadway
2 design characteristics, such as the presence of embankments and elevated roadway segments can
3 alter airflow and contaminant dispersion patterns in the near road environment. For example, their
4 results indicated that roadway design having below-grade sections of road and embankments
5 reduced concentrations away from the road. These results showed similar concentrations as Zhu
6 et al. (2002, 041553) both for roadway segments at-grade with no obstructions to air flow and for
7 elevated roadway segments with different road fill conditions. Additionally, Khare et al. (2005,
8 194016) illustrated in a wind tunnel study that vertical dispersion of a nonreactive gas increased with
9 increasing simulated traffic volume; this effect was also sensitive to changes in approaching wind
10 direction. These studies taken together suggest that localized turbulence induced by roadside
11 structures, roadway design, and traffic provide some mixing and resulting dilution of the CO
12 concentration in the near-road environment; the extent of mixing effects varies by meteorological
13 conditions and the specific roadway design and traffic loading.
2.0
1.5-
^ 1.0-
0 0.5
E
Q.
0.0
Open Terrain CO
Behind Barrier CO
5.00 6.30 8.00 9.3011.0012.3014.0015.3017.0018.30
Time (hrmin)
Source: Baldauf et al. (2008,191017)
Figure 3-28 CO concentration profile 10 m from I-440 in Raleigh, NC behind a noise barrier
and in open terrain.
14 The geometry of urban street canyons has a profound effect on the distribution of CO
15 concentrations on a micro-scale. A number of studies have performed computational and wind tunnel
16 modeling of street canyons using nonreactive tracers and demonstrated the potential variability in
17 concentration within a canyon (e.g., Borrego et al., 2006, 155697; Chang and Meroney, 2003,
18 090298: Kastner-Klein and Plate, 1999, 001961: So et al., 2005, 110746: Xiaomin et al., 2006,
19 156165). Because CO is a pollutant with very low reactivity on urban and regional scales, results
20 from these models are directly relevant to CO concentration distributions in street canyons.
21 Influential parameters include canyon height to width ratio (HAY), source positioning, wind speed
22 and direction, building shape, and upstream configuration of buildings. Figure 3-29 shows
September 2009
3-66
DRAFT - DO NOT CITE OR QUOTE
-------
1 dimensionless concentrations obtained from wind tunnel and computational fluid dynamics
2 simulations of tracer gas transport and dispersion in an infinitely long street canyon with a line
3 source centered at the bottom of the canyon (Xiaomin et al, 2006, 156165). When the canyon height
4 was equal to the street width (typical of moderate density suburban or urban fringe residential
5 neighborhoods) and lower background wind speed existed, concentrations on the leeward
6 (downwind) canyon wall were four times those of the windward (upwind) wall near ground level.
7 When the canyon height was twice the street width (typical of higher-density cities) and background
8 winds were somewhat higher, near ground-level concentrations on the windward canyon wall were
9 roughly three times higher than those measured at the leeward wall. These results suggest that the
10 magnitude of microscale CO concentrations may vary by factors of three or four times at different
11 locations within a street canyon and are heavily influenced by wind speed and street canyon
12 topography. The relationship between in-cany on concentration and wind speed and turbulence is
13 well established with concentration varying inversely with the magnitude of wind speed and
14 turbulence (Britter and Hanna, 2003, 090295). When studying the effect of wind direction on street
15 canyon concentration levels for a continuous "line source" of traffic exhaust, concentration levels
16 were at local maxima under two conditions: wind perpendicular to or parallel to the street canyon.
17 Wind gusts at the turbulence interface at the top of the canyon or traffic-based turbulence can also
18 cause dilution of the exhaust concentration within the canyon (Kastner-Klein et al., 2000, 194035).
September 2009 3-67 DRAFT - DO NOT CITE OR QUOTE
-------
A
windward,
measured
windward
simulated
m leeward,
measured
- - - - leeward,
simulated
1.2
0.8
0.6
0.4
0.2
H/W=1
wind speed = 3 m/s
1.2
0.8
0.6
0.4
0.2
H/W = 2
wind speed = 5 m/s
0 20 40 60 80
Dimensionless concentration
100 0 100 200 300 400 500 600
lb> Dimensionless concentration
Source: Xiaomin et al. (2006,156165)
Figure 3-29 Dimensionless tracer gas concentration on the windward and leeward sides of
the canyon plotted against the elevation of the measurement (Z) scaled by
building height (H) under two different H/Wand wind speed conditions. Shown
are measurements obtained in a wind tunnel (symbols) and model simulations
using computational fluid dynamics (lines).
1 Street canyon field studies support the computational and wind tunnel modeling results
2 described above. In a multisite survey of curbside CO concentration in London, U.K., Croxford and
3 Penn (1998, 087176) observed up to three-fold differences in concentration related to the side of the
4 street on which the monitor was positioned relative to the wind direction with H/W varying between
5 0.7 and 1.7 depending on position within the canyon. Bogo et al. (2001, 192378) measured CO
6 concentrations in a street canyon with H/W of 1 in Buenos Aires, Argentina using a continuous CO
7 monitor. Similar to the Xiaomin et al. (2006, 156165) simulation results for H/W of 1, Bogo et al.
8 (2001, 192378) observed slightly higher leeward concentrations than windward concentrations
9 within the canyon, where recirculating airflow inside the canyon causes pollutants to collect in
10 higher concentration on one side. However, for the case of a deep street canyon (H/W of 5.7) in
September 2009
3-68
DRAFT - DO NOT CITE OR QUOTE
-------
1 Naples, Italy, (Murena et al., 2008, 194038) observed that the concentrations on two sides of the
2 canyon differed by less than 15% with wind direction varying between 10° and 80° from the street
3 axis. Doran et al. (2003, 143352) measured CO concentration in a street canyon in Phoenix, AZ
4 during the morning hours and observed that CO concentration decreases with elevation above the
5 ground if turbulent mixing is small, but that the difference between ground level and 39th-floor (50
6 m AGL) measurements of CO concentration decreases when turbulent mixing increases (with
7 maximum measurements at any elevation not exceeding 2 ppm). As shown in Figure 3-30, the larger
8 difference in concentration as a function of turbulent mixing can occur when there are
9 meteorologically stable conditions in the lower boundary layer. These results support findings from
10 the modeling studies that CO concentration can vary by several times within a street canyon and are
11 greatly influenced by local meteorology and building topography.
oj 1-00
£ °-75
v 0.50
i^__
M_
'•n 0.25
o
o 0.00
"S -0.25
N
'"5 -O.bO
fc -0.75
c -1.00
1
-
o
o
o
0
Do°
O o
° <£
0 1
-
'
a 0200 and 0500 LSI -
« 0600 LSI
0 0700 AND 0800 LST -
-
^
n *
v . . - -.
*n an Dn n
no r^D
(a)
-10 -50 5 10 15
bulk Richardson number
20
Source: Doran et al. (2003,1433521
Figure 3-30 Normalized difference between CO measurements taken at ground level and from
the 39th floor of a building in a Phoenix, AZ street canyon as a function of bulk
Richardson number (Ri). Bulk Ri is a dimensionless number that describes the
ratio of potential to kinetic energy, and it is used here as a measure of stability
within the street canyon, with greater Ri corresponding to greater stability and
values near or less than zero indicating greater mixing.
12 Research by Kaur and Nieuwenhuijsen (2009, 194014) and Carslaw et al. (2007, 148210)
13 suggests that CO exposures are related to traffic volume and fleet mix in the street-canyon
14 environment. Kaur and Nieuwenhuijsen (2009, 194014) used multiple linear regression to model CO
15 concentration data from central London as a function of mode of transport (broken down by vehicle
16 type), traffic count, wind speed, and temperature. They added each variable successively and found
17 traffic count, temperature, wind speed, and walking to be significant parameters in the model, with
September 2009
3-69
DRAFT - DO NOT CITE OR QUOTE
-------
1 traffic count being the strongest determinant. Analysis of variance showed variability in traffic count
2 to explain 78% of the variability in CO levels for these data, and variability in mode of transport
3 explained 6% of the variability. Likewise, Carslaw et al. (2007, 148210) used a generalized additive
4 model to determine how CO concentration data (log-transformed) obtained in central London varied
5 as a function of light- and heavy-duty traffic counts, along-street and cross-street components of
6 wind, temperature, year, and Julian day. Light-duty vehicle count was a more important determinant
7 of CO concentration than heavy-duty (i.e., diesel) vehicle count in this study. They found that the CO
8 declined steadily with year and that wind was the most significant covariate. In addition to showing
9 meteorology to be an important determinant of concentration, these modeling exercises also suggest
10 a linear or log-linear relationship between concentration and traffic.
3.5.2.Temporal Variability
3.5.2.1. Multiyear Trends
11 Figure 3-31 (top) shows ambient CO concentrations in ppm from 1980-2006 based on
12 continuous measurements averaged over 8-h time segments. Figure 3-31 (bottom) depicts trends in
13 the annual second-highest 8-h CO concentrations for 144 sites in 102 counties nationwide having
14 data either in the State and Local Air Monitoring Stations (SLAMS) network or from other special
15 purpose monitors.
September 2009 3-70 DRAFT - DO NOT CITE OR QUOTE
-------
16
= 14
I_ 12
S I 10
90% of sites have concentrations below this line
NAAQS = 9 ppm
11
Average
10% of sites have
concentrations below this line
'80 '82 '84 '86 '88 '90 '92 '9
-------
1 from 68% in Region 7 to 85% in Region 1; see Figure 3-32. This is also consistent with declining
2 emissions seen in many regions of the U.S., shown in Figure 3-32.
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
Coverage: 141 monitoring sites
in the EPA Regions (out of a total
of 375 sites measuring CO in
2006) that have sufficient data to
assess CO trends since 1980.
EPA Regions
.
Source: U.S. EPA (2008, 1570761
Figure 3-32 Trends in ambient CO in the U.S., 1980-2005, reported as the annual second
highest daily 8-h concentrations (ppm) for the EPA Regions 1 through 10, along
with a depiction of the geographic extent of those Regions
3.5.2.2. Hourly Variation
3 Weekday and weekend diel variation for the mean, median, 5th, 10th, 90th, and 95th
4 percentiles of hourly CO concentration over 2005-2007 are shown in Figure 3-33 and Figure 3-34,
5 respectively, for the eleven CSAs and CBSAs examined in this assessment. Since these figures
6 represent the distribution of hourly observations over a 3-yr period, any fluctuations or changes in
7 the timing of the daily peaks would result in a broadening of the curves shown in the diel plot
8 compared to the actual daily temporal behavior on any specific day measured by an individual
9 monitor. However, these figures are useful for comparing the general hourly variation in CO
10 concentrations across cities and by day of the week (i.e., weekday versus weekend). The weekday
11 data showed that the Anchorage mean, median, 5th and 10th percentile CO concentration curves
12 exhibit pronounced morning and evening rush hour peak CO levels. Boston, Denver, Houston, Los
September 2009
3-72
DRAFT - DO NOT CITE OR QUOTE
-------
1 Angeles, Phoenix, Pittsburgh, and St. Louis all exhibited similar trends, although the magnitude of
2 the concentrations shown was roughly twice as high for Anchorage as the other cities. The curves
3 had less overall variability for Boston, Pittsburgh, and St. Louis. The Atlanta plot shows that the
4 median concentration was fairly constant throughout the 24-h period, with a slightly elevated mean
5 during the morning hours. The 90th and 95th percentile curves exhibit stronger morning and evening
6 CO concentration peaks. New York City shows fairly constant CO mean and median concentration
7 throughout the day with slight elevations throughout the morning rush hour and a slight trough
8 between 1:00 and 5:00 a.m. The Seattle plot shows a daytime plateau beginning around 5:00 a.m.
9 and lasting until roughly 10:00 p.m., with higher concentrations during morning and afternoon rush
10 hour. Differences in hourly variation among the eleven CSAs and CBS As reflect city-to-city
11 variation in source characteristics and meteorology. For instance, the rush hour peaks in many cities
12 likely correspond to increased mobile source emissions during those periods. Local meteorology and
13 topography, which influence mixing heights, can also affect hourly variation in CO concentration.
14 Figure 3-34 illustrates weekend diel trends for the eleven CSAs and CBSAs considered in this
15 assessment. For Anchorage during the period 2005-2007, the mean and median concentration curves
16 peaked during the morning and evening hours. A daytime concentration trough is evident. The 90th
17 and 95th percentiles of concentration were similar but more pronounced. The shape of this plot is
18 also characteristic of Atlanta, Boston, Denver, Houston, Los Angeles, Phoenix, Pittsburgh, Seattle,
19 and St. Louis, although the Anchorage CO concentrations are nearly 100% higher than
20 concentrations in the other cities. The weekend diel plot for New York shows that the mean and
21 median CO concentrations remain fairly constant throughout the day, with a slight reduction between
22 2:00 and 7:00 a.m. The 90th and 95th percentile curves illustrate more diel variation.
September 2009 3-73 DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage
Atlanta
Boston
Weekday(N =771)
Weekday (N = 2317)
Weekday (N-5216)
58
^- 4.4 •
Q.
3 29
O
° 15
0 •
1
3
?22
O.
S 1.5
O
O 08
0
3 -
E "'
S.
S 1.5 -
0 08-
0 •
3 •
-p. 2.2 •
3 1.5
0 0.8
0 -
1
3
/*», 2.2-
---. '*
*>.---' ,' - , \V ,.' 0.8-
6 12 18 24
Denver
Weekday (N = 3806)
3-
2.2
/' - ''•- "' '" 08
^vxr^^-'- ""-"-••'' ^
r , ^i^: , , o -
6 12 IB 24
New York
Weekday (N- 6821 )
5.8
4.4
29
'-, "" "1.5
-
6 12 18 24
St. Louis
Weekday (N = 2303)
3
22
1.5
--'"--- ---- «
6 12 18 24 0
3 -
22-
1.5 -
'•'»i-fi ..--'"' ' 08-
, 1 1 1 0 -
6 12 18 24
Houston
Weekday (N = 3734)
3-
22 -
r.s-
,---,
6 12 18 24 ° -
Phoenix
Weekday (N - 3836)
3 -
2.2-
,'\ 1 5 -
0.8-
I-----' '.-I-.--.-.-*-.-;'.-?-.- 7-.-, 0 •
6 12 IS 24
Seattle
Weekday (N = 776)
,-' "-, ,--'
• -''^.~"-——'''^:~~.i~-\
S 12 18 24
Los Angeles
WoekdayfN- 15437)
--., / V; ,,--"'"
'V^<^~""~-'"'.-
6 12 18 24
Pittsburgh
Weekday (N = 5352)
' "'" ""---— •;.---'. .'.'.'— -"-'~
6 12 18 24
Median
Mean
90th&101h
95th & 5th
Figure 3-33 Diel plot generated from weekday hourly CO data (ppm) for the eleven CSAs and
CBSAs 2005-2007. Included are the number of monitor days (N) and the median,
mean, 5th, 10th, 90th and 95th percentiles of composite CO concentrations
plotted by time of day. Note that the y-axis of the Anchorage and Phoenix plots
are scaled to 5.8 ppm while the other plots are scaled to 3.0 ppm.
September 2009
3-74
DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage
Atlanta
Boston
Weekend(N = 309)
Weekend (N - 932)
Weekend (N = 2155)
5.8-
^ 4.4
I
0
0 15
0
3
E "
Q-
3 1.5
O
0 0,8
0
3
I"
3: 15
0
O 08
3 -
E "'
CL
Q. ! 5
O
0 o,&-
o 4-
1
3 -
2.2 -
w-
~~ -'' '-^ ,-'' 0.8-
~ 0-
6 12 Ifl 24
Denver
Weekend (N = 1564)
3 •
2.2 •
1.5
""•-.. .-----.. -'"' 08
I '."' --. , : — , 0 J
6 12 18 24
New York
Weekend (N = 2746)
5.8 -
4.4 -
2.9-
"" . .„ . . . _^. 1 5 -
6 12 18 24 0
Seattle
Weekend (N = 311)
3 -
2.2-
<^i— ^-^ U.
6 12 18 24 ° '
3-
2.2-
1.6 -
— , . . . . . , , , o -
6 12 18 24
Houston
Weekend (N = 151 2)
3 -
2,2-
15 -
0.8
, . , , 0 -
6 12 18 24
Phoenix
Weekend (N = 1550)
3-
2.2-
.--. 1-S '
""'-- ''"
. ,'
0.8 -
6 12 18 24
St. Louis
Weekend (N- 931)
•-- .---"--•
^-^- ^, " ' "-
6 12 18 24
Los Angeles
Weekend (N = 621 6)
'*», ' ---- .-'"""
6 12 18 24
Pittsburgh
Weekend (N = 21 61)
. ,
6 12 18 24
Median
Mean
90th & 10th
95'th&5'th
1 6 12
18 24
Figure 3-34 Diel plot generated from weekend hourly CO data (ppm) for the eleven CSAs and
CBSAs 2005-2007. Included are the number of monitor days (N) and the median,
mean, 5th, 10th, 90th and 95th percentiles of composite CO concentrations
plotted by time of day. Note that the y-axis of the Anchorage and Phoenix plots
are scaled to 5.8 ppm while the other plots are scaled to 3.0 ppm.
September 2009
3-75
DRAFT - DO NOT CITE OR QUOTE
-------
3.5.3.Associations with Copollutants
1 Associations between hourly CO and other copollutants, including SO2, NO2, O3, PMi0, and
2 PM2.s are provided in box plots in Figure 3-35 using AQS data across the U.S. AQS data were
3 obtained from all available collocated monitors across the U.S. after application of the 75%
4 completeness criteria described earlier in Section 3.5.1.1. Pearson correlation coefficients (r) were
5 calculated using 2005-2007 data stratified by season. Correlation plots analogous to Figure 3-35 for
6 select individual cities are provided in Annex A, Figures A.43 to A.48.
Winter
Spring
PM,
••ml—
-KB— 1-
-1.0 -0.8 -0.6 -0.4 -0.2 0
-Q*D — f-
1 \ — LJ r*
•
t^§^J
0 0.2 0.4 0.6 0.8 1.0
4
"i ULJ
ontm
-1.0 -0.8 -0.6 -0.4 -0.2 0
-CH — h--
[AA\-
H * I-H--
T T* *
0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
correlation (r) with CO
-4
• *
• mm
~
-1.0 -0.8 -0.6 -0.4 -0.2 0
fin — f-«
— I [±LH-
n fM»*
I— I T I— h
0 0.2 0.4 0.6 0.8 1.0
correlation (r) with CO
Figure 3-35 Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02,03, PMi0 and PM2.s
concentrations. Red bars denote the median, green stars denote the arithmetic
mean, the box incorporates the IQR and the whiskers extend to the 5th and 95th
percentiles. Correlations outside the 5th and 95th percentiles are shown as
individual points.
7 In all cases, a wide range of correlations existed between CO and copollutants as illustrated in
8 Figure 3-35. The mean and median correlation between CO and copollutants were positive for NO2,
9 PM10, and PM25; near zero for SO2; and negative for O3. These findings reflect common
September 2009
3-76
DRAFT - DO NOT CITE OR QUOTE
-------
1 combustion sources for CO, NO2, and PM; CO is highly correlated with NO2 and PM2 5 because
2 they are both emitted directly during incomplete combustion and because secondary nitrate PM
3 comes from NOX, which is largely produced from mobile sources. Among those copollutants with
4 positive associations, NO2 had the highest mean and median correlations, followed by PM2 5 and
5 PMio (correlations vary by season). The IQR of correlations with SO2 spanned from positive to
6 negative for all seasons; SO2 would not be expected to correlate well with CO because SO2
7 emanates primarily from industrial sources. Correlations between CO and O3 were almost entirely
8 negative for winter, when CO emissions tend to be high and O3 formation is low. During the other
9 three seasons, most of the CO-O3 correlations were also negative. The wide range of correlations
10 displayed in the nationwide plots reflects the large pool of data in addition to the
11 micrometeorological factors in each city.
12 Within and between individual metropolitan areas, the distribution of copollutant correlations
13 varied substantially. Figure 3-36 and Figure 3-37 illustrate the correlations between CO and
14 copollutants for Denver, CO and Los Angeles, CA to exemplify these differences. For instance,
15 correlations between CO and copollutants are all positive for SO2, NO2, PMi0, and PM25 and are all
16 negative for O3 in Denver. In contrast, the correlations in Los Angeles span from negative to positive
17 for O3, PMio, and PM25, in various seasons. The larger span of correlations for Los Angeles in
18 comparison with Denver could result from several factors. For example, more variation in
19 meteorology, topography, or source distribution with respect to monitor placement in Los Angeles
20 may cause the distribution of copollutant correlations to be wider. In addition, fewer collocated
21 monitors in Denver compared with Los Angeles may be causing some of the observed differences.
September 2009 3-77 DRAFT - DO NOT CITE OR QUOTE
-------
Winter
Spring
so?-
NOi^
0.3-
PMio -
PM2.S '
oo
GD
O
00
O
SQ2-
N02-
OaJ
PMw-
PM2.S -
am
©
O
O
O
-1.0 -0,8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.8 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0..4 0.6 0.8 1.0
Summer Fall
SO-
NCfe-
Or
PMto-
PM2.5 •
GD
O
oo
©
O
SOr
NO?'
03'
PMio-
PM« •
€D
oo
O
©
O
•1.0 -0.8 -0,6 -0.4 -0.2 0.0 0,2 0.4 0.6 0.8 1.1
r (correlation coefficient)
-1.0 -0,8 -0.6 -0.4 -0,2 0.0 0,2 0.4 0.6
r (correlation coefficient)
Figure 3-36 Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02,03, PMio and PM2.s
concentrations for Denver, CO.
September 2009
3-78
DRAFT - DO NOT CITE OR QUOTE
-------
W i n te r
Spring
SCfe-
NOz-
DJ-
PM» •
PM2.5 -
OOQDOCS)
O
dBPQD
SCte
N02-
OB-
PMio-
PM25-
(BOD
(ID (08DO
C
KPBOIDS
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0,4 0.6 0.8 1.0
Fall
SCfe-
N02-
DJ-
PMto
PM25
CD
00 CD
KK"»Mft<'0-)))
OQTOOO
SCte-
NO?-
0.1-
PMio-
PM2.5 -
OODO
O
O
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0,8 1,(
r (correlation coefficient)
-1.0 -0.8 -0.6 -0.4 -0.2 8.0 0.2 0.4 0.6 0.8 1.0
r (correlation coefficient)
Figure 3-37 Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02,03, PMi0 and PM2.s
concentrations for Los Angeles, CA.
1 Several recent studies reported correlations between ambient CO and other pollutants.
2 Reported relationships were generally consistent with the correlations shown above using AQS data.
3 Sarnat et al. (2001, 019401) reported significant positive Spearman's correlations between CO and
4 NO2 (r = 0.76) and PM2.5 (r = 0.69) and significant negative correlations between CO and O3
5 (r = -0.67) in Baltimore (concentration averaging periods not specified). Correlation of CO with SO2
6 was insignificant (r = -0.12). The Sarnat et al. (2001, 019401) study focused on correlations of
7 ambient and personal PM2.5 with gaseous copollutants, so seasonal information is only available for
8 the correlation between PM25 and CO. High correlation of ambient CO with NO2 is expected given
9 that both are closely related to mobile source combustion emissions. Sarnat et al (2005, 087531) also
10 reported significant year-round association between CO and PM25 and significant associations
11 between CO and SO42" aerosols. Kim et al. (2006, 089820) measured CO, NO2, and PM2.5 at
12 ambient fixed sites in Toronto, Canada and found associations, averaged over monitoring stations, of
13 CO with PM25 (Spearman's r = 0.38, non-significant) and of CO with NO2 (r = 0.72, significant).
14 Tolbert et al. (2007, 090316) reported correlations between multiple pollutants in Atlanta and also
15 showed the highest Spearman's correlation for CO with NO2 (r = 0.70). CO was also reported to
16 have fairly high correlation with PM25 elemental carbon (EC) (r = 0.66), PM25 organic carbon (OC)
17 (r = 0.59), and PM25 total carbon (TC) (r = 0.63). Correlations were reported to be much lower for
18 CO with O3 (r = 0.27) and PM2.5 SO42" (r = 0.14). The higher correlations of CO with EC, OC, and
September 2009
3-79
DRAFT - DO NOT CITE OR QUOTE
-------
1 TC are likely related to the fact that CO and carbonaceous PM are both emitted by mobile sources.
2 Gentner et al. (2009, 194034) analyzed the relationship between ambient CO and VOC
3 concentrations, serving as markers of gasoline vehicle emissions in Riverside, CA. Correlations of
4 CO with two compounds, n-butane and isopentane, are shown in Figure 3-38 for summer and fall.
5 Higher concentrations of n-butane per unit of CO were observed for fall, as well as higher
6 correlation (fall: r = 0.88; summer: r = 0.65). For isopentane, the slopes of regression are much
7 closer for fall and summer, with higher correlations between isopentane and CO (fall: r = 0.93;
8 summer: r = 0.86). Gentner et al. (2009, 194034) noted that isopentane vapor fraction was higher in
9 summer than winter and that the n-butane vapor fraction increases in winter. This reflects the higher
10 volatility of n-butane compared with isopentane. In this work, Gentner et al. (2009, 194034) used
11 emissions modeling to estimate that overall VOC emissions from gasoline varies with CO emissions
12 with a ratio of 0.086 with a correlation of r = 0.80 in summer. Gentner et al. (2009, 194034) suggest
13 that the near-road slope of ambient VOC to CO concentration might be influenced by upwind CO
14 concentration and secondary CO production by oxidation of VOCs.
September 2009 3-80 DRAFT - DO NOT CITE OR QUOTE
-------
(a)
6-
.Q
.0
Q.
TO
13
.£>
4H
2-
(b)
_Q
-Q
Q.
CO
4->
c
Q.
O
2-
•^
D Fall (r = 0.88)
Summer (r = 0.65)
i !
400 800 1200 1600
Carbon Monoxide [ppbv]
D
D Fall (r = 0.93)
O Summer (r = 0.86)
I
400 800 1200 1600
Carbon Monoxide [ppbv]
Source: Centner et al. (2009,194034)
Figure 3-38 Linear regression of n-butane and isopentane concentration as a function of CO
concentration, Riverside, CA.
3.5.4.Policy-Relevant Background
1 Background concentrations of pollutants used for informing policy decisions about national
2 standards in the U.S. are commonly referred to at EPA as policy-relevant background (PRB)
3 concentrations. In this assessment, PRB concentrations exclude anthropogenic emissions in the U.S.
September 2009
3-81
DRAFT - DO NOT CITE OR QUOTE
-------
1 Canada, and Mexico and include to the extent possible world-wide biogenic emissions including
2 from the U.S., Canada, and Mexico, and all anthropogenic emissions elsewhere in the world.
3.5.4.1. Surface-based Determinations
3 For this assessment, PRB concentrations of CO were determined from the extensive and long-
4 running network of remote-site baseline CO measurements conducted by NOAA's Earth System
5 Research Laboratory (ESRL), Global Monitoring Division (GMD), as part of their Carbon Cycle
6 Greenhouse Gases Group (CCGG) Cooperative Air Sampling Network (CASN); see
7 http://www.esrl.noaa.gov/gmd/ccgg/iadv. Unique among the EPA Criteria Pollutants, surface-based
8 CO measurements have been made for more than 10 yr with exceptionally high sensitivity and
9 selectivity at locations significantly away from local sources. In this assessment, for example, CO
10 data through December 2007 are available with extensive quality assurance and control information
11 from the worldwide network of 72 nodes active in December 2008. ESRL GMD uses the highly
12 sensitive gas chromatography-mercury liberation photometric detection technique with precision to 1
13 part per billion (ppb) in 50 ppb or 2 ppb in 200 ppb and accuracy to 1.5 ppb in 500 ppb or 2 ppb in
14 200 ppb.
September 2009 3-82 DRAFT - DO NOT CITE OR QUOTE
-------
Pcmt Arena\C alifcm
Shemya tslana.
* CO monitors
Monitor Location
Figure 3-39 Map of the baseline monitor sites used in this assessment to compute policy-
relevant background concentrations.
1 In order to smooth interannually changing meteorological and emissions effects, data from
2 2005—2007 at 12 remote sites in the U.S. were used to determine PRB. A map of these sites is
3 shown in Figure 3-39; they are: Cold Bay, AK; Barrow, AK; Shemya Island, AK; Cape Kumukahi,
4 HI; Mauna Loa, HI; Trinidad Head, CA; Point Arena, CA; Wendover, UT; Niwot Ridge, CO; Park
5 Falls, WI; Southern Great Plains, OK; and Key Biscayne, FL. Average concentrations for each
6 month and for each of the 3 yr are shown for each site in Figure 3-40. All sites demonstrate the well-
7 known seasonality in background CO with minima in the summer and fall and maxima in the winter
8 and spring in the Northern Hemisphere (NH). NH summer-time minima are related in large measure
9 to the enhanced photochemical reaction of CO with OH, as described in Section 3.3. Analysis for
10 North American PRB is made here by segregating the three Alaska sites (owing to their high
11 latitude) and the two Hawaii sites (owing to their distance from the continent) and treating the
12 remaining seven sites as representative of the CONUS surface-level background concentrations.
13 Outside the defined CONUS domain used here, the 3-y avg CO PRB in Alaska ranged from 127 to
14 135 ppb with an average of 130 ppb, and from 95.3 to 103.1 ppb with an average of 99.2 ppb in
September 2009
3-83
DRAFT - DO NOT CITE OR QUOTE
-------
1 Hawaii. Over the CONUS domain the 3-y avg CO PRB concentration ranged from 118 to 146 ppb
2 with an average of 132 ppb.
zoo
ISO
Stnmya Island. AK
1U
1BO
&4Q
5m
100
M
60
200
ISO
ISO
JOO
HO-
160 •
100'
80.
K*y fttelyn*. Fl.
J FMAMJJASONDJFHAHJJASONDJFMAMJJA$ONO
200! SOOt JOOf
JFMAMJJASONDJFMAMJJASONDJFMAMJJASONO JFMAMJJASONDJFMaMJJASOHOJFMAHJJOSOND
SMS IO«6 !007 200S 2008 ZM7
Figure 3-40 Monthly (circles) and annual (squares) average CO concentrations (ppb),
2005-2007. Cold Bay, AK; Barrow, AK; Shemya Island, AK; Cape Kumukahi, HI;
Wendover, UT; Niwot Ridge, CO; Mauna Loa, HI; Trinidad Head, CA;
Point Arena, CA; Park Falls, Wl; Southern Great Plains, OK; and
Key Biscayne, FL.
3.5.4.2. Limitations of Other Possible Methods
3 The significance of CO for surface-level air quality and for its indirect climate forcing effects
4 through CH4, O3, and CO2 as described previously in this chapter and its long T relative to that of
5 other primarily urban and regional pollutants make it an important species for measurement and
6 evaluation on multiple spatial, temporal, and chemical scales.
September 2009
3-84
DRAFT - DO NOT CITE OR QUOTE
-------
1 In additional to the ESRL GMD surface network used in this assessment's determination of
2 CO PRB, CO concentrations away from local sources can be measured from space. So, for example,
3 CO has been observed from space by the Measurement of Air Pollution from Satellites (MAPS)
4 instrument on Space Shuttle orbiter flights for three 10-d missions in 1984 and 1994 (Connors et al.,
5 1994, 193755) and by the Measurement of Pollution in the Troposphere (MOPITT) on the Tera
6 satellite since 2000 (Emmons et al., 2004, 193756). Surface spatial coverage with both space-based
7 instruments was limited by the common problems of cloud cover, high surface albedo and
8 emissivities, and image swath pattern and timing with the result that much of the CONUS, for
9 example, was missed some of the time. In addition, all these satellite measurements were limited
10 though somewhat differently in the vertical resolution of their total column CO concentration values.
11 For a determination of a PRB-equivalent background concentration for 2008, the MAPS data
12 would be of no use, excepting for comparisons on temporal tends and even that is limited by the very
13 few observations from MAPS. MOPITT data might seem more useful were it not for MOPITT's
14 very low precision and accuracy in the lowest few kilometers above the Earth's surface of its
15 integrated total column CO measurement by thermal infrared radiances (Shindell et al., 2005,
16 193746). MOPITT CO profile sensitivities are so very low at the surface that retrievals at the
17 850 hPa level - the lowest reported - do not capture the surface concentration with fidelity but
18 actually stand for a broad and deep vertical slice of the lower troposphere with an integral
19 concentration that often peaks well above 850 hPa (Shindell et al., 2006, 091028). Error analysis by
20 Emmons et al. (2004, 193756) reported in Shindell et al. (2006, 091028) revealed that MOPITT
21 concentration error in the lower troposphere was 7% and had greater bias over cleaner sites. The
22 cleaner sites are the ones most of interest, of course, when estimating a CONUS PRB.
23 Since the integrated total column measurements of CO from space-borne instruments are
24 dominated by CO in the mid and upper troposphere comparisons to surface measurements are highly
25 fraught. Using a subset of 7-9 of the ESRL GMD network nodes in North America, for example, to
26 compare to the MAPS and MOPITT data, Shindell et al. (2005, 193746) found that the satellite data
27 showed an increase of between 3 and 13 ppb CO while the surface data at these locations showed a
28 decrease of 20 ppb in the years 2000-2002 relative to 1994. Mean global concentrations of CO were
29 apparently decreasing before 2000 but that trend has now mostly ended (Duncan and Logan, 2008,
30 194042). so that the integrated column CO total measured from space may have indicated a false
31 trend.
32 CO concentrations can also be predicted with numerical CTMs on regional, continental, and
33 global scales. Hence it would, in principle, be possible to predict CO PRB concentrations for the
34 CONUS. The chief limitation to this method comes from the highly uncertain emissions of CO
35 worldwide, most particularly from biomass burning the Southern Hemisphere (SH) and east Asia
36 (EA) needed to drive the global CTMs which in turn set the boundary conditions and chemical flow
September 2009 3-85 DRAFT - DO NOT CITE OR QUOTE
-------
1 fields for the finer-scale models which might be used to compute PRB. Interannual variability in CO
2 emissions from global biomass burning is very high and the emissions source strength of this signal
3 is, of course, a very strong component of the CONUS PRB given the CO T of-57 d. The long T
4 means that CO can mix from SH to NH, requiring even more fidelity in global biomass burning
5 emissions to predict NH background levels (Shindell et al, 2008, 193748) citing Arellano et al.
6 (2004, 193757): Petron et al. (2004, 193758): and Begamaschi et al. (2000, 192377).Thus. for
7 example, in the years just following the intensive El Nino Southern Oscillation in 1997-1998, large-
8 scale Indonesian biomass fires and forest fires in Canada and Siberia were responsible for increases
9 even in the NH in 1998 of, on average, 10-20 ppb compared to other years ((Dentener et al., 2004,
10 194040) citing (Duncan et al., 2003, 193760)). Estimates of total global CO emissions used in recent
11 forward and inverse model experiments range from <1,000 Tg/yr to more than 3000 Tg/yr (Shindell
12 etal, 2005, 193746).
13 A comprehensive evaluation of 26 state-of-the-science atmospheric chemistry models
14 exercised for present-day and future CO simulations was performed and reported by Shindell et al.
15 (2006, 091028). They found substantial under-prediction of CO in the extra-tropical NH compared to
16 satellite and local surface observations and large variability among the models as well even when
17 using identical CH4 abundances and CO emissions. In North America, for example, the multimodel
18 avg underestimated the observations of lower troposphere CO by 60 ppb or more, or by -50% or
19 more of the measured background concentration at many of the ESRL GMD sites. The Pearson r
20 values for the multimodel average against MOPITT data globally for 2000-2001 was 0.84 ± 0.08 for
21 April at 850 hPa (as near to the surface as tested) but only 0.55 ± 0.11 in October (Shindell et al.,
22 2006, 091028). Shindell et al. (2006, 091028) proposed several reasons for this pervasive under-
23 prediction in addition to the widely acknowledged underrepresentation of CO emissions from SH
24 biomass burning: 1) the models do not adequately simulate CO build-up during the wintertime
25 periods of lower OH flux; 2) the models have no seasonal CH4 cycle with build-up in the NH winter;
26 and 3) variability in the models' OH concentrations which accounted for -80% of the CO intermodal
27 variance (Shindell et al., 2006, 091028).
3.6. Issues in Exposure Assessment
3.6.1 .Summary of Findings from 2000 CO AQCD
28 The 2000 CO AQCD (U.S. EPA, 2000, 000907) describes the results of studies completed
29 prior to 1999 on personal exposures and microenvironmental concentrations of CO. Although these
30 studies may no longer be representative of current exposure levels due to declining ambient CO
31 concentrations, the personal-microenvironmental-ambient relationships are still instructive. Time
September 2009 3-86 DRAFT - DO NOT CITE OR QUOTE
-------
1 spent commuting, particularly in cars, was a major contributor to personal CO exposures. Many
2 studies measured in-vehicle concentrations of CO and found elevated concentrations compared to
3 fixed-site monitors. Roadside CO monitors were elevated compared to ambient levels, and equal to
4 or lower than in-vehicle levels (Ott et al., 1994, 076546: Rodes et al, 1998, 010611). A small portion
5 of the CO concentrations inside a vehicle cabin comes from the vehicle itself, while a substantial
6 fraction comes from roadway mobile source emissions entering the cabin via air exchange. Studies
7 summarized in the 2000 CO AQCD found that in-vehicle CO concentrations were generally two to
8 five times higher than ambient CO concentrations obtained at fixed-site monitors within the cities
9 studied. High traffic volumes contributed to increased in-vehicle concentrations.
10 Prior to the 2000 CO AQCD, it was well-known that CO levels in residences may be elevated
1 1 above ambient due to nonambient indoor sources, such as cooking, space heating, and smoking.
12 Separation of indoor CO into ambient and nonambient components was found to be important for
13 determining the effect of ambient CO concentrations, although this had not been done successfully in
14 studies conducted to date. Two large studies performed in Denver, CO and Washington, DC in the
15 early 1980s found that fixed-site monitor concentrations were higher than personal exposures for
16 those with low-level exposures, while fixed site monitor concentrations were lower than exposures
17 for those with high-level exposures (Akland et al., 1985, 011618: Johnson, 1984, 024652).
18 Nonambient sources contributing to high total exposures likely obscured this relationship. In Denver,
19 gas stove operation, passive smoking, and attached garages increased residential indoor exposure by
20 2.6, 1.6, and 0.4 ppm respectively compared to individuals without those sources present.
21 Categorical analyses found significantly higher personal exposures on high ambient concentration
22 days than on low ambient concentration days, suggesting that personal exposures are related to
23 ambient levels. Nonambient exposures tend to obscure the relationship between ambient CO
24 concentrations, as measured at ambient monitors, and total personal CO exposure.
3. 6. 2. General Exposure Concepts
25 A theoretical model of personal exposure is presented to highlight measurable quantities and
26 the uncertainties that exist in this framework. An individual's time-integrated total exposure to CO
27 can be described based on a compartmentalization of the person's activities throughout a given time
28 period:
ET =
Equation 3-2
29 where £T = total (T) exposure over a time-period of interest, Cj = airborne CO concentration at
30 microenvironmentj, and dt = portion of the time-period spent in microenvironmentj. Equation 3-2
September 2009 3-87 DRAFT - DO NOT CITE OR QUOTE
-------
1 can be decomposed into a model that accounts for exposure to CO of ambient (Ea) and nonambient
2 (Ena) origin of the form:
ET=Ea+Ena
Equation 3-3
3 Examples of ambient CO sources include industrial and mobile source emissions, biomass
4 combustion, and agricultural processes. Examples of nonambient sources include environmental
5 tobacco smoke (ETS), cooking, and home heating. CO concentrations generated by ambient and
6 nonambient sources are subject to spatial and temporal variability that can affect estimates of
7 exposure and resulting health effects. Exposure factors affecting interpretation of epidemiologic
8 studies are discussed in detail in Section 3.6.8.
9 This assessment focuses on the ambient component of exposure because this is more relevant
10 to the NAAQS review. Ea can be expressed in terms of the fraction of time spent in various outdoor
11 and indoor microenvironments (Wallace et al., 2006, 089190: Wilson et al., 2000, 010288):
Equation 3-4
12 where/= fraction of the relevant time period (equivalent to dt in Equation 3-2), subscript o = index
13 of outdoor microenvironments, subscript / = index of indoor microenvironments, subscript o,i =
14 index of outdoor microenvironments adjacent to a given indoor microenvironment /', and F^z =
15 infiltration factor for indoor microenvironment i. Equation 3-4 is subject to the constraint S/0 +
16 S/j = 1 to reflect the total exposure over a specified time period, and each term on the right hand side
17 of the equation has a summation because it reflects various microenvironmental exposures. Here,
18 "indoors" refers to being inside any aspect of the built environment, e.g., home, office buildings,
19 enclosed vehicles (automobiles, trains, buses), and/or recreational facilities (movies, restaurants,
20 bars). "Outdoor" exposure can occur in parks or yards, on sidewalks, and on bicycles or motorcycles.
21 Finf is a function of the building air exchange characteristics. Assuming steady state ventilation
22 conditions, the infiltration factor is a function of the penetration (P) of CO, the air exchange rate (a)
23 of the microenvironment, and the rate of CO loss (K) in the microenvironment; F^f = Pa/(a+k).
24 Given that k -> 0 for CO, Finf reduces to P. Studies of CO infiltration are reviewed in
25 Section 3.6.5.1.
26 In epidemiologic studies, Ca is often used in lieu of outdoor microenvironmental data to represent
27 these exposures based on the availability of data. Thus it is often assumed that C0 = Ca and that the
28 fraction of time spent outdoors can be expressed cumulatively as f0; the indoor terms still retain a
29 summation because infiltration differs among different microenvironments. If an epidemiologic
September 2009 3-88 DRAFT - DO NOT CITE OR QUOTE
-------
1 study employs only Ca, then the assumed model of an individual's exposure to ambient CO, first
2 given in Equation 3-4, is re-expressed solely as a function of Ca:
Equation 3-5
3 Meteorology, strength of CO sources, spatial variability of CO concentration, proximity of the
4 study population to sources of CO, design of the epidemiologic study, and other factors determine
5 whether or not Equation 3-5 is a reasonable approximation for Equation 3-4. Errors and uncertainties
6 inherent in use of Equation 3-5 in lieu of Equation 3-4 are described in Section 3.6.8 with respect to
7 implications for interpreting epidemiologic studies. Epidemiologic studies often use concentration
8 measured at a central site monitor to represent ambient concentration; thus a, the ratio between
9 personal exposure to ambient CO and the ambient concentration of CO, is defined as:
Equation 3-6
10 Combination of Equations 3-5 and 3-6 yield:
Equation 3-7
11 a varies between 0 and 1. If a person's exposure occurs in a single microenvironment, the ambient
12 component of a microenvironmental CO concentration can be represented as the product of the
13 ambient concentration and P. Wallace et al. (2006, 089190) note that time-activity data and
14 corresponding estimates of P for each microenvironmental exposure are needed to compute an
15 individual's a with accuracy. If local sources and sinks exist and are significant but not captured by
16 central site monitors, then the ambient component of the local outdoor concentration may be
17 estimated using dispersion models, land use regression models, receptor models, fine scale
18 chemistry-transport models or some combination of these techniques. These techniques are described
19 in Section 3.6.3.
3.6.3.Exposure Modeling
3.6.3.1. Stochastic Population-Based Time-Weighted Microenvironmental Exposure
Models
20 Population-based methods, such as the Air Pollution Exposure (APEX) and Stochastic Human
21 Exposure and Dose Simulation (SHEDS) models, involve stochastic treatment of the model input
22 factors (Burke et al., 2001, 014050: U.S. EPA, 2009, 194009). These are described in detail in
September 2009 3-89 DRAFT - DO NOT CITE OR QUOTE
-------
1 Annex 3.7 of the 2008 NOX ISA (U.S. EPA, 2008, 157073). Stochastic models utilize distributions
2 of pollutant-related and individual-level variables, such as ambient and local CO concentration
3 source contributions and breathing rate respectively, to compute the distribution of individual
4 exposures across the modeled population. The models also have the capability to estimate received
5 dose through a dosimetry model. Using distributions of input parameters in the model framework
6 rather than point estimates allows the models to explicitly incorporate uncertainty and variability into
7 exposure estimates (Zidek et al., 2007, 190076). These models estimate time-weighted exposure for
8 modeled individuals by summing exposure in each microenvironment visited during the exposure
9 period. For example, Bruinen de Bruin et al. (2004, 190943) utilized the EXPOLIS (exposure in
10 polis, or cities) model to predict CO population exposures in Milan, Italy based on subjects' time-
11 activity data broken into 15-min intervals. The simulation results showed that the U.S. 8-h NAAQS
12 level was exceeded in one case out of 1,000. The model also showed that exposures exceeded
13 20 ppm in one case out of 100,000. The results were not shown to be very sensitive to the number of
14 microenvironments (e.g., outdoors, indoors, in vehicle) included in the model.
15 The initial set of input data for population exposure models is ambient air quality data, which
16 may come from a monitoring network or model estimates. Estimates of concentrations in a set of
17 microenvironments are generated either by mass balance methods or microenvironmental factors.
18 Microenvironments modeled include residential indoor microenvironments; other indoor
19 microenvironments, such as schools, offices, and public buildings; and vehicles. The sequence of
20 microenvironments and exertion levels during the exposure period is determined from characteristics
21 of each modeled individual. The APEX model does this by generating a profile for each simulated
22 individual by sampling from distributions of demographic variables such as age, gender, and
23 employment; physiological variables such as height and weight; and situational variables such as
24 living in a house with a gas stove or air conditioning. Activity patterns from a database such as
25 Consolidated Human Activity Database (CHAD) are assigned to the simulated individual using age,
26 gender, and biometric characteristics (U.S. EPA, 2009, 194010). Breathing rates are calculated for
27 each activity based on exertion level, and the corresponding received dose is then computed. For
28 APEX, the CO dosimetry algorithm calculates venous COHb levels using the nonlinear CFK model
29 as described in Chapter 4. (U.S. EPA, 2008, 191775). Summaries of individual- and population-level
30 metrics are produced, such as maximum exposure or dose, number of individuals exceeding a
31 specified exposure/dose threshold, and number of person-days at or above benchmark exposure
32 levels. The models also consider the non-ambient contribution to total exposure. Nonambient source
33 terms are added to the infiltration of ambient pollutants to calculate the total concentration in the
34 microenvironment. Output from model runs with and without nonambient sources can be compared
35 to estimate the ambient contribution to total exposure and dose.
September 2009 3-90 DRAFT - DO NOT CITE OR QUOTE
-------
1 Recent larger-scale human activity databases, such as those developed for the CHAD or the
2 National Human Activity Pattern Survey (NHAPS), have been designed to characterize exposure
3 patterns among much larger population subsets than can be examined during individual panel studies
4 (Klepeis et al, 2001, 002437: McCurdy et al, 2000, 000782). CHAD consists of a consolidation of
5 human activity data obtained during several panel studies in which diary or retrospective activity
6 data were obtained, while NHAPS acquired sample population time activity data through surveys
7 about human activity (Klepeis et al., 2001, 002437). The complex human activity patterns across the
8 population (all ages) are illustrated in Figure 3-41 (Klepeis et al., 2001, 002437). This figure is
9 presented to illustrate the diversity of daily activities among the entire population as well as the
10 proportion of time spent in each microenvironment. Different patterns would be anticipated when
11 breaking down activity patterns for subgroups such as children or the elderly. Population exposures
12 can be estimated using CO concentration data in each microenvironment.
B
O
s*
C3
O
i-l
CO
O -—i
Time of Day
O i—i (N
Source: Klepeis et al. (2001, 0024371.
Figure 3-41
Distribution of time sample population spends in various environments, from the
National Human Activity Pattern Survey.
13 Compartmental models, such as the Indoor Air Model (INDAIR), can be used to assess
14 exposure to infiltrated ambient air pollutants in a deterministic or probabilistic framework
15 (Dimitroulopoulou et al., 2001, 014737). To examine indoor concentrations of ambient CO,
September 2009
3-91
DRAFT - DO NOT CITE OR QUOTE
-------
1 Dimitroulopoulou et al. (2006, 090302) used the probabilistic formulation of the INDAIR model to
2 examine indoor exposure to ambient CO, along with NOX and PM for a given distribution of
3 background CO levels, meteorology, residential air exchange rate, and residential room dimensions.
4 They found that 24-h avg CO concentration increased from 1.86 ppm outdoors to 1.90-1.93 ppm
5 indoors in the absence of non-ambient sources, and that indoor 24-h avg CO concentration could
6 increase to 1.93-2.00 ppm in the presence of smoking and to 1.98-2.32 ppm in the presence of gas
7 cooking. Similarity between the outdoor and non-source indoor concentrations was attributed to the
8 lack of CO loss mechanisms. In the Reducing Urban Pollution Exposure from Road Transport
9 (RUPERT) study, Bell et al. (2004, 192376) presented methodology to use the probabilistic form of
10 INDAIR for development of personal exposure frequency distributions of CO, NOX, and PM based
11 on time spent in residential, transportation, school, office, and recreational environments with inputs
12 from transportation source categories (Chen et al., 2008, 193986).
3.6.3.2. Using Spatial Models to Estimate Exposure
13 Another set of approaches to improve exposure estimates in urban areas involves construction
14 of a concentration surface over the geographic area. This does not estimate exposure directly because
15 it does not account for activity patterns or concentrations in different microenvironments. It provides
16 an improved estimate of the expected local outdoor concentration near residences, schools or
17 workplaces, and roadways across the area. There are two main types of approaches: spatial
18 interpolation of measured concentrations, and regression models using land use, roadway
19 characteristics, and other variables to predict concentrations at receptors in the domain. Rigorous
20 first-principles models, such as dispersion models and chemical transport models, can also be used
21 for this type of application, but are less suitable because they have intensive resource requirements
22 and are typically applied over larger domains.
23 The STEMS model provides an example of an integrated exposure modeling approach using a
24 range of spatial inputs. STEMS maps exposures based on inputs for traffic levels, atmospheric
25 dispersion, background concentrations, and geography. Gulliver and Briggs (2005, 191079) tested
26 the STEMS model for CO and observed some correlation between modeled and measured CO
27 concentrations (R2 = 0.41), which was consistent with results for PMi0 and NOX. Exposures were
28 estimated from the predicted ambient CO concentration using a term similar to a that varied
29 depending on whether the individual was walking or in a vehicle. Gulliver and Briggs (2005,
30 191079) noted that a limitation to modeling CO is the scarcity of background CO data obtained at
31 rural sites. For this reason, they assumed a constant value obtained from estimates made over the
32 North Atlantic Ocean. Although the authors only presented detailed results for a model of PMi0
33 based on traffic and meteorology in Northampton, U.K., they found that the majority of variation on
34 a given day in modeled exposure among school children was due to differences in travel routes.
September 2009 3-92 DRAFT - DO NOT CITE OR QUOTE
-------
1 Variation across days was also influenced by background and meteorological conditions. Similar
2 results can be expected for CO based on the tendency for variation of the CO concentration profile
3 on the neighborhood and micro-scales (Jerrett et al, 2005, 092864) . Flachsbart (1999, 015857)
4 tested numerous meteorological, traffic, and background CO input variables in a regression approach
5 to predicting CO exposure among individuals while traveling in a vehicle. This work showed travel
6 time and average speed of on-road vehicles to be important determinants of CO exposure in a
7 vehicle. Results from individual models of this nature can be pooled to develop a distribution for
8 examination of population effects or for comparison with population exposure models.
Dispersion Models
9 Dispersion models have been used both for direct estimation of exposure and as inputs for
10 stochastic modeling systems, as described above. Location-based exposures have been predicted
11 using a model such as California Line Source Dispersion Model (CALINE), the American
12 Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD),
13 CALPUFF (long-range plume transport model created by the California Air Resources Board), or the
14 Operational Street Pollution Model (OSPM) for estimation of street-level ambient CO exposure
15 (e.g., Abdul-Wahab, 2004, 194011; Delfmo et al., 2009, 190254; Zhou and Levy, 2008, 190091).
16 CALINE, CALPUFF, and AERMOD utilize Gaussian dispersion models to describe pollutant
17 transport, while OSPM is a semi-empirical model of airflow and pollutant transport within an infinite
18 street canyon that assumes the street canyon airflow to be similar to a driven cavity. Delfmo et al.
19 (2009, 190254) used CALINE (version 4) to model exposure in the near-road environment for
20 estimation of relative risks of asthma hospitalizations as a function of increases in ambient CO and
21 NOX concentrations. The concentration at each subject's home was computed with the dispersion
22 model, and then the data were aggregated to estimate a population risk. Zhou and Levy (2008,
23 190091) used results from an OSPM simulation to compute intake fraction, defined as the fraction of
24 emissions that are inhaled or ingested, for ambient CO and other copollutants. Daytime activity
25 patterns were modeled using both CHAD and the American Community Survey to model
26 commuting behaviors that would affect both mobile source emissions and population-based
27 exposures. With an individualized exposure approach, the model is deterministic. However,
28 population exposures were estimated by performing repeated simulations using various housing
29 characteristics and then computing a posterior probability distribution function for exposure. When
30 comparing street canyon exposure computed by OSPM with near-road exposure computed simply
31 with a Gaussian dispersion model, Zhou and Levy (2008, 190091) estimated that the street canyon
32 exposures would be three times greater than those in the general community. Isakov et al. (2009,
33 191192) developed a methodology to link a chemical transport model, used to compute regional
September 2009 3-93 DRAFT - DO NOT CITE OR QUOTE
-------
1 scale spatiotemporally-varying concentration in an urban area, with stochastic population exposure
2 models to predict annual and seasonal variation in urban population exposure within urban
3 microenvironments. Although this approach was demonstrated for PM2.5, it is similar to the one used
4 by Zhou and Levy (2008, 190091) for linking ambient CO concentrations with population activity
5 pattern data to link the spatial concentration field to personal exposure to ambient CO.
Land Use Regression Models
6 Marshall et al. (2008, 193983) compared four spatial interpolation techniques for estimation of
7 CO concentrations in Vancouver, BC. The investigators assigned a daily average CO concentration
8 to each of the 51,560 postal code centroids using one of the following techniques: (1) the
9 concentration from the nearest monitor within 10 km, (2) the average of all monitors within 10 km,
10 (3) the inverse-distance-weighted (IDW) average of all monitors in the area, and (4) the IDW
11 average of the three closest monitors within 50 km. Method 1 (the nearest-monitor approach) and
12 Method 4 (IDW-50 km) had similar mean and median estimated annual average concentrations,
13 although the 10th-90th percentile range was smaller for IDW-50. This is consistent with the
14 averaging of extreme values inherent in IDW methods. The Pearson correlation coefficient between
15 the two methods was 0.88. Methods 2 and 3 were considered sub-optimal and were excluded from
16 further analysis. In the case of Method 2, a single downtown high-concentration monitor skewed the
17 results in the vicinity, partially as a result of the asymmetric layout of the coastal city of Vancouver.
18 Method 3 was too spatially homogenous, because it assigned most locations a concentration near the
19 regional average, except for locations immediately adjacent to a monitoring site. LUR results were
20 also reported in this study for NO and NO2, and indicated that LUR's higher spatial precision
21 reflects neighborhood-scale effects from nearby land use, but may not account for urban-scale
22 variation. These results highlight the variation in local concentration estimates with choice of
23 estimation technique.
3.6.4.Personal Exposure Monitors for CO
24 Portable monitors for measuring personal CO exposure include the Langan and Draeger
25 monitors, both of which use electrochemical oxidation-reduction techniques (Langan, 1992,
26 046120). These monitors continuously log CO concentrations, making them suitable for use in
s JO " O
27 personal monitoring studies. Electrochemical CO sensors typically have a limit of detection of 1 ppm
28 and a 90% sensor response time (or the time required for the sensor to register 90% of a step change
29 in CO concentration, of 20-60 s. The 2000 CO AQCD (U.S. EPA, 2000, 000907) provided detail on
30 design updates of electrochemical CO sensors made during the 1990s. Commercially available CO
31 personal exposure monitors are not designed to detect concentrations below 1 ppm. Electrochemical
September 2009 3-94 DRAFT - DO NOT CITE OR QUOTE
-------
1 personal CO monitors are also typically sensitive to temperature changes, so that data correction is
2 normally required.
3.6.5.Indoor Exposure to CO
3.6.5.1. Infiltration of Ambient CO
3 CO is a relatively inert gas, making the indoor decay rate negligible compared to typical air
4 exchange rates (~l/h). In the absence of indoor sources, this would lead to an indoor-outdoor
5 concentration ratio (I/O) of approximately 1. For this reason, few studies have calculated I/O for CO.
6 Polidori et al. (2007, 156877) calculated I/O of 0.94-1.21 for two retirement communities in the Los
7 Angeles area. The authors suggested that similarity between I/O for CO and NOX can be attributed
8 to lack of indoor sources of either gas. Chaloulakou and Mavroidis (2002, 026050) reported I/O
9 using CO measurements in the absence of indoor sources in a school building in Athens, Greece and
10 found that I/O varied with season. During the summer, median I/O was reported to be 0.57 on
11 weekdays, 0.91 on Saturdays, and 0.81 on Sundays. In winter, median I/O was reported to be 0.82
12 during weekdays, 0.90 on Saturdays, and 0.74 on Sundays. In a related study, Chaloulakou et al.
13 (2003, 190945) reported the median I/O over all days as 0.8 for the same school and 0.9 for an
14 Athens office building with no ETS (the presence of other sources was not clearly stated but
15 assumed zero). However, observed indoor values are often greater than outdoor concentrations in the
16 presence of indoor sources. A recent study in the U.K. reported I/O of 3.9-4.3 in homes with gas
17 cookers (Dimitroulopoulou et al., 2006, 090302). which is consistent with previous studies. A
18 multipollutant study conducted in 2000-2001 attempted to measure I/O for CO and calculated
19 residential infiltration factors, but low CO concentrations resulted in a large number of
20 measurements below the limit of detection (Williams et al., 2003, 053335). Ni Riain et al. (2003,
21 053792) examined the effects of mechanical ventilation and wind speed on I/O. In this study, the
22 authors measured indoor and outdoor concentrations at two buildings located on a six-lane highway
23 in central London with natural and mechanical ventilation. Ni Riain et al. (2003, 053792) (2003
24 Atmos Environ 37: 4121-432) found that outdoor concentrations for each building and ventilation
25 condition ranged from 1.5 ± 0.1 ppm to 1.9 ± 0.1 ppm. Ni Riain et al. (2003, 053792) reported
26 cumulative I/O approaching 0.9 within 30 min of sampling for the mechanical ventilation case and
27 cumulative I/O varying between 0.65 and 0.8 for more than 70 h of sampling for the natural
28 ventilation case. Ni Riain et al. (2003, 053792) found that wind speed and direction influenced the
29 variation in I/O.
30 Indoor air flow may affect CO exposure in the absence of indoor sources. Milner et al. (2006,
31 123100) compared hourly CO concentration time series from different parts of a building (with a
32 mix of natural and mechanical ventilation) located near a busy road and intersection in central
September 2009 3-95 DRAFT - DO NOT CITE OR QUOTE
-------
1 London, U.K. They found that, within a given floor, CO concentration is greater in rooms that are
2 closer to busy roads or an intersection. They noted that the correlation coefficient between indoor
3 and outdoor CO concentrations also decreased within the building with distance from the road; the
4 correlation coefficients were reported to be 0.80 for two time series obtained in rooms near the road,
5 while they were reported to range between 0.46 and 0.55 on the sides of the building furthest from
6 the road. The magnitude of the difference between CO concentrations in different rooms located
7 nearer or further from the roads also depended on wind direction. Milner et al. (2006, 123100) noted
8 that I/O tended to decrease with increasing wind speed, but Chaloulakou et al. (2003, 190945) also
9 noted that indoor CO concentration varied inversely with wind speed. Chaloulakou et al. (2003,
10 190945) attributed their observation to reduced concentrations related to dilution effects. Milner
11 et al. (2006, 123100) stated that this relationship could be due to dilution of CO or to the tendency of
12 people to keep windows closed on windy days. Additionally, CO concentrations were higher on
13 lower floors of the building and varied over a given day throughout the building. These findings
14 suggest that differences in exposure can occur within the same building as a result of differences in
15 air exchange related to access to windows, mechanical ventilation, and outdoor meteorological
16 conditions.
3.6.5.2. Exposure to Nonambient CO
17 Several papers have investigated the microenvironmental sources of total personal CO
18 exposure. The CDC conducted a survey of emergency department (ED) visits for non-fatal CO
19 poisoning, CO exposure, or potential CO exposure and found that home heating was the largest
20 known source of CO exposure, prompting 16.4% of CO-related ED visits, followed by motor vehicle
21 exhaust exposure accounting for 8.1% of ED visits (Annest et al., 2008, 190236). Aim et al. (2000,
22 192374; 2001, 020237) studied factors that contributed to elevated CO exposures among pre-school
23 children and found that presence of a gas stove at home, ETS, natural ventilation, and living in a
24 high rise building all contributed to increased CO exposures. Time-activity diaries were linked to
25 personal CO exposures in the EXPOLIS study. Here, Georgoulis et al. (2002, 025563) observed that
26 geometric mean exposure among smokers ranged from 0.33 ppm in Helsinki, Finland to 3.2 ppm in
27 Athens, Greece, while among nonsmokers it ranged from 0.36 ppm in Helsinki to 1.7 ppm in Milan
28 and ambient CO concentration ranged from 0.42 ppm in Helsinki to 3.2 ppm in Athens. Bruinen de
29 Bruin (2004, 190943) found, for a panel of 46 subjects in Milan, that indoor CO concentrations were
30 3.4 ppm in the presence of gas cooking and ETS, compared with 2.9 ppm only in the presence of
31 ETS, 2.4 ppm only in the presence of gas cooking, and 1.8 ppm in the absence of indoor CO sources.
32 Scotto di Marco et al. (2005, 144054) reported that average indoor CO increased in the presence of
33 ETS from 0.96-1.2 ppm for the home indoor environment and from 1.0-1.4 ppm for the work indoor
34 environment. CO concentrations were measured to decrease from 1.5 to 1.3 ppm in other (not home
September 2009 3-96 DRAFT - DO NOT CITE OR QUOTE
-------
1 or work) indoor environments, but those locations included garages, restaurants, and bars and could
2 have been differently influenced by CO from cooking, indoor automobiles, or other sources.
3 Personal CO concentrations can also be much more variable than ambient measurements.
4 Figure 3-42 shows hourly versus personal CO concentration data obtained by Chang et al. (2000,
5 001276) for a 1998-1999 multipollutant sampling campaign in Baltimore, MD. Personal exposures
6 were obtained in five separate microenvironments in this study. A high degree of scatter is evident in
7 this figure, which suggests that these personal exposures are influenced by both ambient and non-
8 ambient sources of CO. Figure 3-43 is a box plot of the personal-to-ambient CO concentration ratio
9 for the same five microenvironments. Wide variability is seen in these plots, particularly during the
10 summer. Much of that variability could be due to the influence of non-ambient sources, which would
11 then result in poor correlation between total personal exposure and ambient concentration.
I
o
o
o
3 •
I -
O Indoor Renidence
A Indoor (Mi«»
O Outdoor near Roadway
V Outdoor «way Irom Road
+ In Vehicle
**>
Hourly Ambient CO (ppm)
Source: Chang et al. (2000, 001276'
Figure 3-42 Hourly personal versus ambient CO concentrations obtained in Baltimore, MD
during summer of 1998 in five settings: indoor residence, indoor other, outdoor
near road, outdoor away from road, in vehicle.
September 2009
3-97
DRAFT - DO NOT CITE OR QUOTE
-------
swswswswsw
Indoor Indoor Outdoor Outdoor In Vehicle
Residence Other Near away from
Road Road
Source: Adapted from Chang et al. (2000, 001276'
Figure 3-43 Box plots of the ratio of personal to ambient concentrations obtained in
Baltimore, MD during summer of 1998 and winter of 1999 in five settings: indoor
residence, indoor other, outdoor near road, outdoor away from road, in vehicle.
The grey line shows the mean, and the black mid-line shows the median. S =
summer; W = winter.
3.6.6.Exposure Assessment Studies at Different Spatial Scales
3.6.6.1. Neighborhood-to Urban-Scale Studies of Ambient CO Exposure
1 Although several multipollutant exposure studies have been conducted recently in the U.S.,
2 (e.g., Sarnat et al., 2006, 089784). most have not included CO in the suite of pollutants, possibly due
3 to high detection limits in personal monitors. A few studies conducted in Europe and Canada
4 measured personal-ambient relationships for CO. This section summarizes CO exposure assessment
5 studies that compare personal exposure measurements with ambient concentration measurements for
6 the purpose of examining how well these measures correspond.
7 The EXPOLIS study (Georgoulis et al., 2002, 025563) found that 48-h personal exposures
8 were significantly correlated with ambient concentrations in each of five European cities (Athens,
9 Basel, Helsinki, Milan, and Prague). Controlling for source terms, including ETS, traffic, and natural
10 gas appliances, regression coefficients between personal exposure and ambient concentration ranged
11 from 0.28 in Milan to 1.99 in Helsinki. The ambient concentration was the only variable that was
12 statistically significantly associated with 48-h personal exposure for all five cities in this study, with
13 correlations between personal CO exposure and ambient CO concentration ranging from 0.33 to
14 0.77. Georgoulis et al. (2002, 025563) reported that CO exposure in traffic ranged from 0.99 ppm in
September 2009
3-98
DRAFT - DO NOT CITE OR QUOTE
-------
1 Helsinki to 4.2 ppm in Athens, while ambient CO concentration ranged from 0.42 ppm in Helsinki to
2 3.2 ppm in Athens. As part of this study, personal CO exposure was measured for a panel of 50 office
3 workers in Milan (Bruinen de Bruin et al., 2004, 190943). Average measured 1-h personal exposures
4 were 7.3 ppm in comparison with 5.0 ppm for fixed site 1-h measurements. Average 8-h (3.3 ppm)
5 and 24-h (2.1 ppm) CO concentrations were the same for personal and fixed site measurements.
6 Percentage of time exposed, exposures, and percentage of exposure from the Bruinen de Bruin et al.
7 (2004, 190943) study, in the absence of non-ambient CO from ETS and gas cooking, are shown in
8 Table 3-13. The largest percentage of time-weighted CO exposure was attributed to home indoor
9 exposure in the absence of indoor sources, while the highest exposure levels were observed during
10 transit. Scotto di Marco et al. (2005, 144054) found similar results. Bruinen de Bruin et al. (2004,
11 190943) and Scotto di Marco et al. (2005, 144054) found that mobile source emissions were
12 important contributors to personal exposure, as described in the following subsection.
Table 3-13 Percentage of time exposed to ambient CO (adjusted to reflect the absence of
non-ambient CO from ETS and gas cooking), average CO exposures, and percentage of
exposure estimated for the population.
Percent of time exposed (%)
INDOORS 89.6
Home 56.5
Work 29.1
Other 4.1
OUTDOORS 1.8
Home 0.2
Work 0.6
Other 1.0
IN-TRANSIT 8.5
Walking 3.0
Train/metro 0.7
Bus/tram 2.0
Motorbike 0.2
Car/taxi 2.6
13 EXPOLIS also looked at the special cas
14 generally do not produce CO in their daily act:
15 (2000. 192374; 2001. 020237) reported higher
16 children aged 3-6 yr old in Helsinki. Their me;
17 1.4 ppm measured at a fixed-site monitor. For
18 the corresponding values were 2.9 ppm and 2.
19 respectively, for fixed site measurements. The
Exposure (ppm) Percent of exposure (%)
81.1
1.8 49.4
1.9 26.8
2.5 4.9
2.1
2.3 0.2
2.1 0.6
2.6 1.2
16.8
3.0 4.4
3.0 1.0
3.8 3.7
4.5 0.4
5.7 7.2
Source: Bruinen de Bruin et al. (2004, 1909431.
e of children's exposure to CO because children
ivities and have no occupational exposures. Aim et al.
• personal exposures than ambient concentrations for
an 1-h daily max exposure was 5.2 ppm, compared to
the average of 8-h and 24-h daily max concentrations,
1 ppm for personal exposure and 0.8 and 0.6 ppm,
Spearman rank correlation, although statistically
September 2009
3-99
DRAFT - DO NOT CITE OR QUOTE
-------
1 significant, was relatively low (r = 0.15) between individual 24-h avg exposure and the ambient
2 monitor. The correlation improved when the average exposure of children measured on the same day
3 (r = 0.33, 3-6 children) or the same week (r = 0.55, 10-23 children) was compared to the monitor
4 data. A regression model using questionnaire data found that parental smoking status, parental
5 education, and presence of a gas stove explained only 12% of the variability in the 8-h max
6 exposures, indicating that other factors, such as time spent outdoors and proximity to roadways are
7 likely to be important in determining personal exposure.
8 Kim et al. (2006, 089820) reported mean CO concentrations of 1.4 ppm for a panel of 28
9 cardiac-compromised individuals in Toronto, Canada. Corresponding fixed-site monitor mean
10 concentrations ranged from 0.5 to 1.4 ppm, with an overall mean of 1.0 ppm. The observed higher
11 personal exposures may have been due to both indoor sources and proximity to roadways when
12 outdoors. Personal-ambient Spearman correlations ranged from -0.65 to 0.93, with a median of
13 r = 0.31, indicating that while moderate correlations are observed overall, inter-individual
14 differences based on time spent in different microenvironments have a strong influence on the
15 observed correlation. Lai et al. (2004, 056811) measured relationships between personal CO
16 exposure and microenvironmental (home indoor, home outdoor, and work indoor) concentrations in
17 Oxford, U.K.. The highest personal exposures were associated with smoking, cooking, and
18 transportation while low correlations were observed between personal and indoor residential
19 concentrations, further indicating the importance of indoor sources and the need to separate ambient
20 contributions to personal exposure from total personal exposure.
21 The studies presented above present mixed results regarding the association between ambient
22 CO concentration measurements and personal CO exposures. Some personal CO measurements have
23 been reported to be higher than ambient concentrations, while others are similar. Additionally,
24 correlation between ambient CO concentration and personal exposure has varied in the literature.
25 Nonambient (described in Section 3.6.5) and in-transit sources (described in Section 3.6.6.2) have
26 been identified as important contributors to personal exposure. These observations raise questions
27 about where and when ambient CO concentration can be used as a surrogate for personal CO
28 exposure; these concepts are explored further in Section 3.6.8 Implications for Epidemiology
3.6.6.2. Microscale Studies of Ambient CO Exposure: Near-Road and On-Road
Exposures
29 The 2007 American Housing Survey (AHS) (U.S. Census Bureau, 2008, 194013) reports that
30 17.9 million occupied homes nationwide (16.1%) are within 91.4 m (300 ft) of a "4-or-more-lane
31 highway, railroad, or airport" and so are exposed to the near-road environment. Within city centers,
32 6.2 million occupied homes (19.7% of those living in city centers) are within 91.4 m of ahighway,
33 railroad, or airport; whereas in rural areas outside designated Metropolitan Statistical Areas (MSA),
September 2009 3-100 DRAFT - DO NOT CITE OR QUOTE
-------
1 1.4 million occupied homes (9.2% of those in rural areas outside MS As) are near a highway,
2 railroad, or airport. Those data can be put into context for exposure assessment in the near-road
3 environment. Section 3.5.1.3 describes near-road studies in which ambient CO was measured within
4 the vicinity of a road and microscale AQS data obtained in the near-road environment. The AQS data
5 suggest some spatial variability (20-40% difference between microscale and middle scale monitors,
6 with the hourly microscale concentration having a median of 0.5 ppm and a 99th percentile value of
7 2.2 ppm), which was much lower than that reported by Zhu et al. (2002, 041553) for the near-road
8 environment, in which the average concentration at 17 m from the road was 2.3 ppm (range
9 1.9-2.6 ppm) and a factor of about 12.5 lower for the monitoring site located 300 m from the road.
10 The larger discrepancy observed between the Zhu et al. (2002, 041553) data and the AQS data might
11 be attributed to the fact that the sampling equipment used by Zhu et al. (2002, 041553) were
12 downwind of the freeway for the entire sampling period, while the hourly AQS data represents a
13 range of wind speeds and directions that vary across different monitoring sites. For those living in
14 the 16.1% of occupied homes situated in the near-road environment (within approximately 90 m),
15 median hourly CO concentrations are typically higher than those further from the road, but the
16 magnitude of the outdoor concentration is still in most circumstances measured to be below 2.2 ppm.
17 Kaur and Nieuwenhuij sen (2009, 194014) and Carslaw et al. (2007, 148210) suggest that CO
18 exposures are related to traffic volume and fleet mix in the street-canyon environment. In this
19 research, Kaur and Nieuwenhuij sen (2009, 194014) developed a multiple linear regression of CO as
20 a function of mode of traffic, broken down by vehicle type, wind speed, temperature, and traffic
21 count for data obtained in central London as part of the DAPPLE study of traffic-related pollution.
22 They added each variable successively and found traffic count, temperature, wind speed, and
23 walking to be significant parameters in the model, with traffic count being the strongest determinant.
24 Analysis of variance showed variability in traffic count to explain 78% of the variability in CO levels
25 for these data, and variability in mode of transport explained 6% of the variability. Likewise,
26 Carslaw et al. (2007, 148210) used a generalized additive model to determine how CO concentration
27 (log-transformed) varies as a function of year, the along-street and cross-street components of wind,
28 temperature, Julian day, light and heavy traffic counts, and temperature for data obtained in central
29 London. Light duty vehicle count was a more important determinant of CO concentration than was
30 heavy duty (i.e., diesel) vehicle count in this study, which is not surprising because gasoline powered
31 vehicles are known to emit more CO than diesel engines. They found that the CO concentration
32 declined steadily with year and that wind was the most significant covariate. The decline in CO
33 concentration with year, adjusted for all other covariates, was usually significantly different than the
34 simple relationship between concentration and year, but the adjusted and unadjusted trends were
35 similar. In addition to showing meteorology to be an important determinant of concentration, these
September 2009 3-101 DRAFT - DO NOT CITE OR QUOTE
-------
1 modeling exercises also suggest a linear or log-linear relationship between concentration and traffic
2 count.
3 Findings regarding meteorology are consistent with in-vehicle CO concentration studies.
4 Gomez-Perales et al. (2007, 138816) also noted that meteorology can impact in-vehicle exposures,
5 with evening increases in wind speed causing a 50% reduction in CO exposures among bus and
6 minibus commuters. Aim et al. (1999, 047196) made a similar observation in a study of urban
7 commuters' exposure within a vehicle. These observations are sensible given the influence of
8 meteorology on near-road concentrations shown by Baldauf et al. (2008, 190239) and Gokhale and
9 Khare et al (2007, 194015).
10 A number of studies have focused on transit-time CO exposure, which can occur while in a
11 vehicle or cycling (on-road) or while walking (near-road). Chang et al. (2000, 001276) showed that
12 personal exposures in vehicles were on average 2.8 times higher than ambient during the summer
13 and 4.1 times higher than ambient in the winter (see Figure 3-43). For the other four
14 microenvironments tested, the average ratio was around 1. Kaur et al. (2005, 086504) found that
15 transit time exposures in London, U.K. were significantly higher than measurements made at a fixed
16 site background monitor away from traffic (0.3 ±0.1 ppm) for car riders (1.3 ± 0.2 ppm), taxi riders
17 (1.1 ± 0.1 ppm), bicyclers (1.1 ± 0.2 ppm), walkers (0.9 ± 0.2 ppm), and bus riders (0.8 ± 0.1 ppm).
18 Curbside measurements (1.5 ± 0.7 ppm) in this study were slightly higher than car riders' exposures.
19 Duci et al. (2003, 044199) found that average in-transit concentrations in Athens, Greece were
20 highest for cars (winter: 21.4 ± 4 ppm), followed by pedestrians (winter: 11.5 ± 2.6 ppm; summer:
21 10.1 ± 1.7 ppm), buses (winter: 10.4 ± 2.9 ppm; summer: 9.4 ± 3.6 ppm), trolleys (winter: 9.6 ±
22 1.9 ppm; summer: 8.2 ± 3 ppm), and rail transit (winter: 4 ± 0.6 ppm; summer: 3.4 ± 0.7 ppm). Duci
23 et al. (2003, 044199) did not provide fixed site CO concentrations but stated that in-transit exposures
24 were higher in each case. Gomez-Perales et al. (2004, 054418) measured CO exposures on buses,
25 mini-buses, and metro cars in Mexico City, Mexico to be 12 ppm, 15 ppm, and 7 ppm, respectively.
26 These values are much higher than CONUS measurements and those presented by Kaur et al. (2005,
27 086504). but the relative difference between the minibus and bus exposures in the Gomez-Perales
28 et al. study are similar to those seen for the taxi-to-bus or car-to-bus comparisons in Kaur et al.
29 (2005, 086504). These studies indicate that on-road exposures might be influenced by vehicle type,
30 but that city-to-city differences are likely larger than differences between different modes of
31 transport.
32 Additional analyses from the EXPOLIS study indicated that on-road mobile source emissions
33 were the most important source of CO exposure for non-ETS-exposed subjects (Bruinen de Bruin et
34 al., 2004, 190943: Scotto Di Marco et al., 2005, 144054). Scotto di Marco et al. (2005, 144054)
35 found that, for a panel of 201 adult Helsinki, Finland residents (aged 25-55 yr), subjects spent 8.1%
36 (1.9 h) of their time in transit, which accounted for 12.6% of their total exposure (range of means =
September 2009 3-102 DRAFT - DO NOT CITE OR QUOTE
-------
1 0.96 ppm on a train - 2.8 ppm in a car). Similarly, in a panel study of 50 office workers, Bruinen de
2 Bruin et al. (2004, 190943) found that, in the absence of non-ambient sources, the subjects spent
3 8.5% (2 h) of their time in transit, which accounted for 16.8% of their total exposure, with 2.6% of
4 time spent in a car or taxi accounting for 7.2% of exposure (mean = 5.7 ppm). Commuting time was
5 an important predictor of exposure, such that subjects living in low CO concentration suburban areas
6 and commuting to work experienced higher levels than urban residents with short commute times.
7 According to the 2007 AHS (U.S. Census Bureau, 2008, 194013). 110.1 million U.S. workers
8 (87.8% of those working) commute to work in automobiles. 32.8% of U.S. workers work at home or
9 commute less than 15 min to work, 32.1% commute 15-29 min to work, 15.1% commute 30-44 min
10 to work, 5.7% commute 45-59 min to work, and 5.0% commute 1 h or longer to work.
11 Vehicle ventilation can be an important determinant of in-vehicle concentrations. A study from
12 Abi Esber et al. (2007, 190941) is presented because they observed in-vehicle CO concentration
13 time-series under a range of ventilation conditions, although the in-vehicle CO concentrations
14 measured are substantially higher than those observed in the U.S. Abi Esber et al. (2007, 190941)
15 report results from CO concentration measurements taken within an automobile in Beirut, Lebanon
16 during the morning commute period of 7:30 - 9:30 a.m. Weekday trip CO levels ranged from
17 10.8 ppm with the windows open and vents closed to 37.4 ppm when driving with windows and
18 vents closed. Mean and standard deviation for ambient CO concentrations, obtained using a roadside
19 monitor in Beirut during the periods September-December 2003, August-September 2004, and
20 May-August 2005 were 1.4 ± 0.7, 1.6 ± 0.4, and 1.1 ± 0.7 ppm, respectively. Abi Esber and El-Fadel
21 (2008, 190939) compared the amount of CO produced by an automobile, driving the same route of
22 Beirut described in Abi Esber et al. (2007, 190941) above, by sampling CO directly outside the
23 vehicle and separately from the cabin of the car under three different ventilation conditions. Cabin
24 CO concentration of 2 ppm was reported at the beginning of the experiments, and average ambient
25 CO levels were reported by Abi Esber et al. (2007, 190941) to be 1.1-1.6 ppm for measurement
26 periods in 2003-2005. For the case when one window was half-open and vents were closed, outside
27 CO concentrations averaged 12.6 ppm while in-vehicle concentrations averaged 17.7 ppm, which
28 was 40.5% higher. With windows closed and the air conditioner operating on "recirculating air"
29 mode, CO concentrations averaged 13 ppm from outside and 30.2 ppm in the vehicle cabin, a 132%
30 increase. With windows closed and the air conditioner on "fresh air" mode, outdoor CO
31 concentrations averaged 18.3 ppm while in-vehicle concentrations were 20.5 ppm, which was only a
32 12% increase. Figure 3-44 shows that the time series for the cabin and outdoor CO samples are very
33 similar for the fresh air scenario, but for the recirculating air ventilation the concentration increases
34 then reaches a plateau as CO builds up in the cabin of the vehicle. These values are substantially
35 higher than in-vehicle concentrations reported above for other studies but illustrate the role in a
36 vehicle's ventilation system on CO build-up within the cabin.
September 2009 3-103 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
£
cL
o.
75
50 -
825H
0
10 20 30
Time {inin)
40 50
0 10 20 30 40 50
Time (min)
Source: Abi Esber and El Fadel (2008,190939)
Figure 3-44
Comparison of in-vehicle (solid line) and outside the vehicle (dotted line) results
for (left) driving with windows closed and air conditioner in recirculating air
mode, and (right) driving with windows closed and air conditioner in fresh air
mode.
Riediker et al. (2003, 043761) measured CO concentrations inside patrol cars during shifts.
Troopers recorded in a time-activity diary the ventilation settings of their cars and exit/entry from the
vehicle, and the air conditioning was typically set to recirculation mode during the shifts. Riediker
et al. (2003, 043761) found that CO concentrations (mean, SD: 2.6 ±1.1 ppm) were higher than
ambient monitor concentrations (0.8 ±0.3 ppm). They were also higher than roadside CO
concentrations (1.1 ± 0.3 ppm), indicating that either the vehicle itself contributes to in-cabin CO, or
on-road concentrations are higher than roadside concentrations, or both. Riediker et al. (2003,
043761) noted that within-shift variability was higher than between-shift variability, which
underscores the variability in police officers' activities during a given shift. Data were not segregated
by ventilation settings, although the police officers typically operated the air conditioning
continually because the study was performed during the summer. Aim et al. (1999, 047196) reported
in-vehicle CO concentrations of 5.7 ppm in the morning and 3.1 ppm in the afternoon commute for
Kuopio, Finland. These data indicate that within-shift variability observed by Riediker et al. (2003,
043761) might, consistent with Aim et al. (1999, 047196). be related to time of day. Chang et al.
(2000, 001276) measured CO concentrations during a scripted activity study in Baltimore, MD in
1998 and 1999. Mean 1-h CO concentrations were near the 1 ppm detection limit of the Langan CO
monitor. Microenvironmental CO concentrations were significantly correlated with concentrations
measured at a fixed-site ambient monitor for residential, other indoor, in-vehicle, and outdoor near-
road microenvironments during the winter. Significant correlations were observed only for
residential microenvironments during the summer. The location of the ambient monitor near a
roadway may have contributed to the lack of correlation with concentrations measured at outdoor
September 2009
3-104
DRAFT - DO NOT CITE OR QUOTE
-------
1 locations away from roadways. Microenvironmental concentrations inside vehicles were
2 significantly higher than those for other microenvironments.
3 Vehicle self-pollution, defined by Behrentz et al. (2004, 155682) as the fraction of a vehicle's
4 own exhaust entering the vehicle microenvironment, is another potential source of CO exposure.
5 This has been studied using inert tracer gases to evaluate exposures of children riding school buses.
6 Behrentz et al. (2004, 155682) used sulfur hexafluoride (SF6) tracer gas emitted from school bus
7 engines to determine the proportion of in-vehicle pollution related to self-pollution. Based on the
8 SF6 concentration, they calculated that 0.04-0.29% of the bus cabin air contained exhaust for high
9 emitting diesel engines, 0.01-0.03% for "regular" diesel buses, 0.02-0.04% for buses fitted with a
10 particle trap, and 0.03-0.04% for buses running on compressed natural gas. SF6 concentrations were
11 higher when bus windows were closed. In addition to demonstrating that some portion of the in-
12 vehicle concentration is due to self-pollution, results from Behrentz et al. (2004, 155682) support the
13 Abi Esber and El Fadel (2008, 190939) and Riediker et al. (2003, 043761) studies cited above that
14 vehicle ventilation is an important determinant of in-vehicle CO concentration.
15 In their review of roadway exposures to CO and PM, Kaur et al. (2007, 190070) listed a
16 number of factors that may influence near-road or on-road exposure. Vertical CO concentration
17 gradients have been documented in which concentrations decreased with height; lower breathing
18 zone height among children may make them more likely to be exposed to higher CO tailpipe
19 emissions. With respect to transportation, Kaur et al. (2007, 190070) suggested that vehicle
20 ventilation, speed, position in traffic, and start/stop activity influence in-vehicle exposures. Abi Esber
21 and El Fadel (2008, 190939) and Riediker et al. (2003, 043761) illustrated the effect of vehicle
22 ventilation on in-vehicle concentrations. The influence of vehicle speed and start/stop activity is
23 consistent with the turbulence research of Khare et al. (2005, 194016) and Gokhale and Khare (2007,
24 194015) that suggested an increase in traffic volume and vehicle movement acts to dilute the on-road
25 concentration of CO discussed in Section 3.5.1.3.
3.6./.Association between Personal CO Exposure and Copollutants
26 Since incomplete combustion is the primary source of ambient CO in urban areas, exposure to
27 ambient CO is accompanied by exposure to other combustion-related pollutants, such as NOX, PM,
28 and VOCs. Thus, ambient CO is often considered a surrogate for exposure to traffic-generated
29 pollutants. However, the specific mix of CO with NOX and PM depends on the source; for example,
30 the mixture generated by gasoline engines differs from that produced by natural gas combustion.
31 Correlations between ambient CO and ambient PM2.5, PM10, NO2, SO2, and O3 from AQS data and
32 the peer-reviewed literature were presented in Section 3.5.3. Nationwide, ambient CO was most
33 highly correlated with ambient NO2 followed by PM25 and PM10. Correlations between CO and
34 PM2 5 were not consistently positive on a national basis; correlations spanned from negative to
September 2009 3-105 DRAFT - DO NOT CITE OR QUOTE
-------
1 positive for ambient CO with ambient SO2 and ambient PMi0, and ambient CO was negatively
2 correlated with ambient O3. The correlation between ambient CO and specific ambient VOCs
3 depends on parameters such as ambient temperature and the volatility of a specific compound.
4 Relationships between personal CO exposures and copollutants were reported less frequently
5 in the literature, but results from these studies were consistent with the findings cited above. In a
6 study of personal exposures to CO, PM25, and ultrafme PM in a street canyon, Kaur et al. (2005,
7 086504) found low Pearson's correlation of total personal CO exposure with personal PM25
8 exposure (r = 0.23). Personal CO exposure had much better correlation with personal ultrafme PM
9 exposure (r = 0.68). Chang et al. (2000, 001276) reported correlations of personal CO exposure with
10 personal PM2.5, personal toluene, and personal benzene exposures in Baltimore, MD at five
11 locations, labeled indoor residential, indoor nonresidential, outdoors near roadway, outdoors away
12 from road, and in vehicle. Much variability was observed in the correlations for different locations
13 and seasons (winter versus summer). In general, the correlations of personal CO with personal VOCs
14 tended to be stronger in the winter. Chang et al. (2000, 001276) suggested that lower wintertime
15 indoor air exchange rates could increase exposure to nonambient sources of both CO and VOCs,
16 such as ETS and hence increase correlations between personal exposure of CO to VOCs. Significant
17 associations of CO with benzene and toluene were also observed in vehicle microenvironments.
3.6.8.Implications for Epidemiology
18 Exposure error can be an important contributor to variability in epidemiologic study results.
19 Community time-series studies may involve thousands or millions of people whose exposure and
20 health status is estimated over the course of a few years using a short monitoring interval (hours to
21 days). Community-averaged concentration is typically used as a surrogate for ambient exposure in
22 community time-series studies. Exposures and health effects are spatially aggregated over the time
23 intervals of interest because they are designed to examine health effects and their potential causes at
24 the community level (e.g., Bell et al., 2009, 194033). A longitudinal cohort epidemiology study
25 typically involves hundreds or thousands of subjects followed over several years or decades.
26 Concentrations are generally aggregated over time and by community to estimate exposures (e.g.,
27 Rosenlund et al., 2006, 089796). In addition, panel studies, which consist of a relatively small
28 sample (typically tens) of study participants followed over a period of days to months, have been
29 used to examine the health effects associated with exposure to ambient concentrations of air
30 pollutants. An example of panel studies include time-activity diary studies (Akland et al., 1985,
31 011618; e.g., Bruinen de Bruin et al., 2004, 190942; Scotto Di Marco et al., 2005, 144054). These
32 studies may apply a microenvironmental model to represent exposure to an air pollutant.
33 The importance of exposure misclassification varies with study design and is dependent on the
34 spatial and temporal aspects of the design. For example, the use of a community-averaged CO
September 2009 3-106 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentration in a community time-series epidemiologic study may not allow for adequate
2 examination of the role of spatial variability. Other factors that could influence exposure estimates
3 include spatial and temporal variability related to source strength, topography of the natural and built
4 environment, and meteorology; measurement errors; use of ambient CO concentration as a surrogate
5 for ambient CO exposure; and the presence of CO in a mixture of combustion-related pollutants. The
6 following sections will consider various sources of error and how they affect the interpretation of
7 results from epidemiologic studies of different designs.
3.6.8.1. Measurement Error
Measurement Error at Community-Based Ambient Monitors and Exposure
Assessment
8 Because CO concentrations measured with community-based ambient monitors are often used
9 as surrogates for ambient CO exposure in epidemiology studies, the limitations of the
10 instrumentation are important to consider. As stated in Section 3.4.2, among the 291 monitors
11 meeting completeness criteria for 2005-2007, only 8 were trace-level monitors; the other monitors
12 have limits of detection of 0.5 ppm. Among the nationwide AQS data for 2005-2007 from these 291
13 monitors, more than 50% of the hourly CO concentration data were below the LOD of the
14 instrumentation. Data below the LOD adds uncertainty to the association between CO exposure and
15 health effects estimates.
16 Instrumental measurement error, other than that related to high LOD, is not expected to bias
17 health effect estimates substantially in most circumstances. Because there will be some random
18 component to instrumental measurement error, the correlation of the measured CO concentration
19 with the true CO concentration will likely be less than 1. When analyzing the effect of instrument
20 error for measuring nonreactive ambient pollutants, Zeger et al. (2000, 001949) stated that the
21 instrument error for ambient measurements "is close to the Berkson type". In the Berkson error
22 model, the measured exposure estimate is used instead of the true exposure based on the assumption
23 that the average measurement is the average of the true exposure. It is generally expected that the
24 health effects estimate will not be biased by using measured values with error but may have more
25 uncertainty than would an estimate based on the true average exposure. In order for instrument error
26 to cause substantial bias in health effects estimates, the error term (the difference between the true
27 concentrations and the measured concentrations) must be strongly correlated with the measured
28 concentrations.
September 2009 3-107 DRAFT - DO NOT CITE OR QUOTE
-------
Measurement Error for Personal Exposure Monitors
1 Personal electrochemical CO monitors are subject to interference and drift, and have a
2 relatively high detection limit (approximately 1 ppm) relative to current ambient concentrations.
3 Previous studies in the 1980's and 1990's, when ambient levels were higher, were able to
4 successfully deploy these monitors, but more recent exposure studies have avoided personal CO
5 measurements due to the high percentage of non-detects. The lack of a suitable personal monitor for
6 measuring low-level exposures (<1 ppm) has hampered field studies assessing personal exposure to
7 ambient CO. Chang et al. (2001, 019216) evaluated the Langan CO monitor as part of an air quality
8 sampling manifold. At high (0.4-3.0 ppm) CO concentrations, the instrument correlated well (R2 =
9 0.93) with a reference NDIR CO monitor, with the Langan underestimating the CO concentration by
10 41%. When ambient levels fell consistently below that level, coefficient of determination (R2)
11 between the Langan and reference monitor fell to R2 = 0.4 in summer and R2 = 0.59 in winter with
12 the arithmetic average concentration underestimated by 47% in summer and by 63% in winter.
13 Chang et al. (2001, 019216) pointed out the need for frequent instrument zeroing to minimize
14 instrument drift. Abi Esber et al. (2007, 190940) evaluated a similar personal electrochemical CO
15 sensor, the GEM™ 2000, by comparing measured concentrations with those obtained through
16 co-located grab bag sampling in a vehicle cabin. Differences between the GEM™ 2000 and the
17 reference samples were fairly low during weekday driving (differences = 2.1-10.6%). Differences on
18 Sundays, when traffic was significantly lower than during weekdays, were dependent on vehicle
19 ventilation conditions, with better agreement when vehicle ventilation allowed for higher cabin CO
20 concentrations (differences = 3.4-5.6%), but the electrochemical sensor did not compare well with
21 reference values when concentrations were low (differences = 20-71%). In general, it is difficult to
22 separate the large instrumental measurement error seen at concentrations below instrument LOD
23 from variation related to non-ambient CO sources. This large variation in personal measurements can
24 result in high levels of classical measurement error (Sheppard et al., 2005, 079176).
3.6.8.2. Exposure Issues Related to Nonambient CO
25 The focus of the ISA is on ambient CO because that is relevant to the NAAQS. Uncertainty
26 related to nonambient CO exposure may make it difficult to distinguish the effect of ambient CO on
27 health effects. Wallace and Ziegenfus (1985, 011656) used NHANES II (1976-1980) data to evaluate
28 the relationship between COHb levels and ambient CO concentration in 20 U.S. cities. They found a
29 significant slope of 0.066% per 1 ppm increase of CO concentration. However, there was much
30 scatter in the data, and variability in ambient CO concentration only accounted for 3% of the
31 variation in COHb. The authors attributed this scatter to variability in nonambient sources such as
September 2009 3-108 DRAFT - DO NOT CITE OR QUOTE
-------
1 gas cooking and ETS. This finding illustrates the importance of considering the relative role of
2 ambient and nonambient CO in total personal exposure.
3 Ambient and nonambient CO are chemically identical and so exert the same health effects. At
4 the same time, ambient and nonambient sources are distinct and not correlated with each other
5 (Wilson and Suh, 1997, 077408) and so would not confound the association between ambient CO
6 exposure and the health effect (see also (Sheppard et al., 2005, 079176)). Zeger and Diggle (2001,
7 026017) noted that, because ambient and nonambient CO exposures are uncorrelated, in a health
8 effects model the regression coefficient of ambient concentration should be equal to the product of a
9 (the ratio of ambient exposure to ambient concentration) and the regression coefficient obtained
10 when average personal exposure is used. The confidence intervals around the estimate obtained
11 using total personal exposure would be wider because nonambient CO concentrations add variability.
12 This is true even for the case when the chemical compound is the same for the ambient and
13 nonambient pollutants, as in the case of CO. Likewise, Sheppard et al. (2005, 079176) simulated
14 ambient and nonambient exposures to a non-reactive pollutant and observed that nonambient
15 exposure has no effect on the association between ambient exposure and health outcomes for the
16 case where ambient and nonambient concentrations were independent. Hence, the bias that will be
17 introduced to epidemiologic models by using ambient CO concentration instead of ambient CO
18 exposure or personal CO exposure is given by the average a. Random variations in daily values of a
19 would not change the health effects estimate but would also widen the confidence intervals around
20 the health effect.
3.6.8.3. Spatial Variability
21 CO concentration is known to be spatially heterogeneous, as evidenced by the near-road and
22 in-vehicle studies cited in Section 3.5.1.3 and 3.6.6.2 as well as the intraurban correlations provided
23 in Section 3.5.1.2 and Tables A-9 through A-16 of Annex A. Results from Zhu et al. (2002, 041553).
24 which showed a large CO concentration gradient in the near-road environment, support the
25 contention that CO exposures for those living in the near-road environment but far from a monitor
26 might be underestimated. Conversely, exposures for those living away from roads might be
27 overestimated by near-road CO concentration measurements. Exposure error may occur if the
28 ambient CO concentration measured at the central site monitor is used as an ambient exposure
29 surrogate and differs from the actual ambient CO concentration outside a subject's residence and/or
30 worksite (in the absence of considering indoor CO sources). Averaging data from a large number of
31 samplers will dampen inter-sampler variability, and use of multiple monitors over smaller land areas
32 may allow for more variability to be incorporated into an epidemiologic analysis. This is consistent
33 with conclusions presented in the 2000 AQCD (U.S. EPA, 2000, 000907).
September 2009 3-109 DRAFT - DO NOT CITE OR QUOTE
-------
1 Community exposure may not be well represented when monitors cover large areas with
2 several subcommunities having different sources and topographies. The intersampler correlations of
3 AQS data from monitors, presented in Section 3.5.1.2, reflect how well the time-series of
4 concentration data correspond across metropolitan areas. Overall, the data show moderate site-to-site
5 correlation; for example, in the Los Angeles CSAthe mean of the correlation was 0.50, and within
6 one standard deviation of the mean, the range of correlations was 0.36-0.65. Bell et al. (2009,
7 194033) tested the association between monitor density and 1-h max CO effect estimates for CVD
8 hospitalizations for 126 U.S. counties and found an 8% increase in effect estimate size (95% CI: -7%
9 to - 24%) with an IQR decrease in area covered by the monitor. This difference was not statistically
10 significant but suggested that the magnitude of the effect estimate might be related to monitor
11 coverage. Sarnat et al. (2009, 180084) studied the spatial variability of CO, along with NO2, O3, and
12 PM2.5, in the Atlanta, GA metropolitan area and how spatial variability affects interpretation of
13 epidemiologic results, using time-series data for circulatory disease emergency department visits.
14 Sensitivity to spatial variability was examined at slightly greater than neighborhood scale (8 km) in
15 this study. Interestingly, Sarnat et al. (2009, 180084) found that relative risk varied with distance
16 between the monitor and study population when comparing urban to rural locations, but distance of
17 the study population to the monitor was not an important factor when comparing urban population
18 groups. This suggests that, even for spatially heterogeneous CO, urban scale measures may produce
19 results comparable to neighborhood-scale exposures in some circumstances. This may be due to
20 comparability of sites throughout a city, for example as a result of similar traffic patterns. However,
21 Sarnat et al. (2009, 180084) caution that, because their study was limited to 8 km radii, it is not
22 possible to interpret this work with respect to near-road and on-road microscale exposures.
3.6.8.4. Temporal Variability
Temporal Correlation
23 Within a city, lack of correlation of relevant time series at various sites results in smoothing
24 the exposure/surrogate concentration function over time and resulting loss of peak structure from the
25 data series. At the same time if monitors are well correlated across a metropolitan area, even if the
26 magnitude of concentration varies over space, time series analyses should provide comparable
27 results across larger spatial areas. Such temporal correlation resulted in the small variation in relative
28 risk estimates within the metropolitan region in Sarnat et al. (2009, 180084). where peak rush hour
29 times were similar throughout the city, in comparison with the rural area where temporal driving
30 patterns were different. Burnett and Goldberg (2003, 042798) found that community time-series
31 epidemiologic study results reflect actual population dynamics only when five conditions are met:
September 2009 3-110 DRAFT - DO NOT CITE OR QUOTE
-------
1 environmental covariates are fixed spatially but vary temporally; the probability of the health effect
2 estimate is small at any given time; each member of the population has the same probability of the
3 health effect estimate at any given time after adjusting for risk factors; each member of the
4 population is equally affected by environmental covariates; and, if risk factors are averaged across
5 members of the population, they will exhibit smooth temporal variation. Note that for this study,
6 Burnett and Goldberg (2003, 042798) analyzed mortality related to PM exposure, but the results are
7 not specific to a given pollutant or health effect and thus are generalized here for time-series
8 analysis. Dominici et al. (2000, 005828) note that ensuring correlation between ambient and
9 community average exposure time series air pollutant data is made difficult by limitations in
10 availability and duration of detailed ambient concentration and exposure time series data, resulting in
11 a source of uncertainty. If sufficient data are available and the time-series of concentration data
12 adequately represent population dynamics, then high temporal correlation between sampling sites
13 should limit bias in health effects estimates, even if the magnitude of the concentrations differ.
Seasonality
14 Community time-series epidemiologic studies can be designed to investigate seasonal effects
15 by incorporating seasonal interaction terms for the exposure surrogate and/or meteorology (e.g.,
16 Dominici et al., 2000, 005828). Sheppard et al. (2005, 079176) examined the role of seasonality on
17 epidemiologic models. They found that a for the population will vary seasonally. This makes sense
18 because a is a function of the amount of time spent indoors and outdoors and of indoor ventilation.
19 Given that use of ambient CO concentration instead of ambient CO exposure biases the coefficient
20 used in epidemiologic models by a, Sheppard et al. (2005, 079176) found that seasonal trends
21 causing a change in a would contribute additional positive or negative bias, depending on the season
22 and region of the country. However, several studies discussed in Chapter 5 investigated seasonal
23 effects. No consistent seasonal pattern across health outcomes in these studies.
3.6.8.5. CO Exposure in Copollutant Mixtures
24 Because CO exposures most often occur together with exposure to other combustion-related
25 pollutants, especially in traffic, interpretation of health studies using ambient CO data can be a
26 challenge, as discussed further in Chapter 5. Ambient CO concentrations from AQS data (see Section
27 3.5.3) have been shown to be correlated with ambient concentrations of NO2 and VOCs, and
28 personal CO exposures have been correlated with personal PM and VOC exposures (see Section
29 3.6.7). Correlation between factors is one condition for confounding, so it is possible that NO2 or
30 VOCs could confound estimates of the health effects of ambient CO concentrations, and CO
31 concentration could potentially confound estimates of the health effects of NO2 or VOCs. For this to
September 2009 3-111 DRAFT - DO NOT CITE OR QUOTE
-------
1 be true, both CO and the copollutant would have to be correlated with the health outcome of interest.
2 The moderately high correlations between ambient CO and copollutants make it difficult to discern
3 the extent to which CO and other compounds are associated with a given health effect.
4 It is also possible that the factor of interest may be the multipollutant mixture emitted from on-
5 road or other combustion processes. The HEI Report on Traffic Related Pollutants (HEI, 2009,
6 191009) suggests that ambient CO, NO2, and benzene could all be considered as surrogates for
7 mobile source-related pollution, but none are ideal surrogates for mobile source pollution because
8 ambient CO concentration tends to decrease rapidly with distance from the source (Baldauf et al.,
9 2008, 190239: e.g., Zhu et al., 2002, 041553). NO2 is reactive, and benzene is volatile. Additionally,
10 PM components of mobile source emissions change rapidly in size and composition from secondary
11 formation and other atmospheric processing. Given that the mixture of mobile source-related
12 emissions changes rapidly as a result of these factors, the ratio of CO to other components of mobile
13 source emissions also changes. Hence, even if CO is itself stable within the mixture of copollutants,
14 the dynamic evolution of the mixture may change the representativeness of CO as an indicator of
15 that mixture over time. Additionally, reductions in CO emissions over the past 30 yr have brought
16 ambient CO concentrations down substantially, with more than half of hourly measurements below
17 the LOD for most instruments (see Section 3.5.1.1). Furthermore, CO and other copollutants found
18 in mobile-source emissions have multiple anthropogenic and biogenic sources and, as a result, are
19 difficult to attribute solely to mobile source pollution. For all of these reasons, the representativeness
20 of CO as an indicator of the multipollutant mixture of mobile source emissions has not been clearly
21 determined.
3.6.8.6. Conclusions
22 This section presents considerations for exposure assessment and the exposure
23 misclassification issues that can potentially affect health effects estimates. These issues can be
24 categorized into the following areas: measurement, nonambient sources, spatial variability, temporal
25 variability, and CO in copollutant mixtures. Potential influences of each of these sources on health
26 effect estimates derived from panel, time-series, and longitudinal epidemiologic studies are
27 described above. Additionally, error sources have the potential to interact with each other. For
28 example, CO concentrations have been shown to decrease rapidly with distance from a highway, and
29 so spatial variability is an important issue in assessing CO exposure. Exposure error may occur if the
30 ambient CO concentration measured at the central site monitor is used as an ambient exposure
31 surrogate and differs from the actual ambient CO concentration outside a subject's residence and/or
32 worksite. However in time-series epidemiologic studies, spatial variability will only be an important
33 source of error if the time-series of CO concentration at different locations are not well correlated in
34 time. The spatial variability of CO, in mixture with the dynamically changing group of mobile
September 2009 3-112 DRAFT - DO NOT CITE OR QUOTE
-------
1 source pollutants, adds to the difficulty of quantifying the health effects related specifically to CO
2 compared with those related to other combustion-related copollutants. In most circumstances,
3 exposure error tends to bias a health effect estimate downward (Sheppard et al., 2005, 079176; Zeger
4 et al., 2000, 001949). Insufficient spatial or temporal resolution to capture true variability and
5 correlation of CO with copollutants are examples of sources of uncertainty that could widen
6 confidence intervals and so reduce the statistical significance of health effects estimates.
3.7. Summary and Conclusions
3.7.1.Sources of CO
7 In the U.S., on-road mobile sources constituted more than half, or ~63 MT of-109 MT total,
8 of total CO emissions in 2002, which is the most recent publicly available CO emission dataset
9 meeting EPA's data quality assurance objectives. In metropolitan areas in the U.S., for example, as
10 much as 75% of all CO emissions can come from on-road vehicle exhaust (U.S. EPA, 2006,
11 157070). The majority of these on-road CO emissions derive from gasoline-powered vehicles since
12 the O2 content, pressure, and temperature required for diesel fuel ignition do not produce large
13 quantities of CO. Anthropogenic CO emissions are estimated to have decreased 35% between 1990
14 and 2002. On-road vehicle sector emissions controls have produced nearly all these national-level
15 CO reductions. Nationally, biogenic emissions, excluding fires, were estimated to contribute -5% of
16 total CO emissions from all sources in 2002, and fires in 2002 added another 13%, or -14.5 MT, to
17 the national CO emissions total.
3.7.2.Physics and Chemistry of Atmospheric CO and Related Climate
Forcing Effects
18 In addition to being emitted directly by incomplete combustion, CO is produced by
19 photooxidation of CH4 and other VOCs in the atmosphere, including NMHCs. Estimating the CO
20 yield from oxidation of HCs larger than CH4 requires computing the yields of several intermediate
21 products and reactants from oxidation of the parent molecules. The major pathway for removal of
22 CO from the atmosphere is reaction with OH to produce CO2 and HO2. The mean photochemical
23 lifetime (T) of CO in the northern hemisphere is -57 days. During winter at high latitudes, CO has
24 nearly no photochemical reactivity on urban and regional scales.
25 Recent data do not alter the current well-established understanding of the role of urban and
26 regional CO in continental and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
27 2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
September 2009 3-113 DRAFT - DO NOT CITE OR QUOTE
-------
1 Intergovernmental Panel on Climate Change (IPCC, 2001, 156587). CO is a weak direct contributor
2 to greenhouse warming because its fundamental absorption band near 4.63 (im is far from the
3 spectral maximum of earth's longwave radiation at -10 (im. Sinha and Toumin (1996, 193747)
4 estimated the direct radiative forcing (RF) of CO computed for all-sky conditions at the tropopause
5 to be 0.024 W/m2 from the change in CO mean global concentrations since pre-industrial times. The
6 RF value similarly computed by Sinha and Toumin for more than doubling the current mean global
7 background concentration to 290 ppb was 0.025 W/m2. However, because reaction with CO is the
8 major sink for OH on a global scale, increased concentrations of CO can lead to increased
9 concentrations of other trace gases whose loss processes also involve OH chemistry. Some of those
10 trace gases, CH4 and O3 for example, absorb infrared radiation from the Earth's surface and
11 contribute to the greenhouse effect directly; others, including the chlorofluorocarbons (CFCs),
12 hydrochlorofluorocarbons (HCFCs), methyl chloride, and methyl bromide, can deplete stratospheric
13 O3, increasing the surface-incident UV flux. Because of these chemical interdependencies,
14 calculations of an indirect RF for any of these short-lived O3 precursor species are most often made
15 for all of the most important ones together. So, for example, the combined effect of increased CH4,
16 CO, NMVOC, and NOX emissions since 1750 has produced tropospheric O3 concentrations
17 associated with a net RF of-0.35 W/m2. The integrated 20-year and 100-year time horizon RFs were
18 determined by IPCC (2007, 092765) for year 2000 emissions of CO, NMVOC, and NOX to be -0.19
19 W/m2, or just slightly lower than the RF of year 2000 black carbon emissions from fossil fuel and
20 biomass burning on the same time horizons.
3.7.3.Ambient CO Measurements
21 As of August 2009, 24 automated FRMs and no FEMs had been approved for CO. All EPA
22 FRMs for CO operate on the principle of nondispersive infrared (NDIR) detection and can include
23 gas filter correlation (GFC). Current specifications for CO monitoring are designed to help states
24 demonstrate whether they have met compliance criteria, with requirements for an LOD of 1 ppm.
25 The reported LOD for 20 of the 24 FRMs is 0.5 ppm, and four trace-level FRMs are in operation
26 with an LOD of 0.04 ppm. FRMs with higher LOD also are limited to a precision of 0.1 ppm and are
27 more subject to drift compared with newer trace-level monitors with automatic drift correction
28 options.
29 For 2005-2007, there were 291 CO monitors meeting the 75% completeness requirements and
30 reporting values year-round to the AQS in the 50 states, plus the District of Columbia, Puerto Rico,
31 and the Virgin Islands. 57 monitors across the U.S. have been sited at microscale to capture
32 near-road concentrations, 31 have been sited at middle scale, and 119 are sited for neighborhood-
33 scale monitoring; among the remaining 84 monitors, states did not declare the spatial scale of
34 monitoring for 71 monitors, and 13 are sited for monitoring urban or regional scale. For CO, traffic
September 2009 3-114 DRAFT - DO NOT CITE OR QUOTE
-------
1 is the major source in an urban setting and therefore microscale data are sited "to represent
2 distributions within street canyons, over sidewalks, and near major roadways" while middle scale
3 monitors are sited to represent "air quality along a commercially developed street or shopping plaza,
4 freeway corridors, parking lots and feeder streets" (40 CFR Part 58 Appendix D). At middle and
5 neighborhood scales, monitor distance from a road is directly related to the road's average daily
6 traffic count to capture community averages. Ambient monitors for CO and other criteria pollutants
7 are located to monitor compliance rather than population exposures. However, AQS monitors are
8 often used for exposure assessment. When comparing CO monitor location with population density,
9 it was observed that population coverage varies both within and between cities.
3.7.4.Environmental CO Concentrations
10 CO concentration data for 1-h and 8-h intervals were available for 243 counties and
11 autonomous cities or municipalities that maintained active CO monitoring stations meeting the 75%
12 completeness criteria for the years 2005-2007. There were no violations of the 1-h or 8-h NAAQS in
13 those years. The nationwide mean, median, and interquartile range for 1-h measurements reported
14 between 2005 and 2007 were 0.5, 0.4, and 0.4 ppm, respectively, and these statistics did not change
15 by more than 0.1 ppm for each year of the 3-year period. The nationwide mean, median, and
16 interquartile range for 8-h daily max concentrations, reported between 2005-2007, were 0.7, 0.5, and
17 0.5 ppm, respectively. The 2006 annual second highest 8-h CO concentration, averaged across 144
18 monitoring sites nationwide, was 75% below that for 1980 and is the lowest concentration recorded
19 during the past 27 yr. The mean annual second highest 8-h ambient CO concentration has been
20 below 5 ppm since 2004. The downward trend in CO concentrations in the 1990s parallels the
21 downward trend observed in CO emissions and can be attributed largely to decreased mobile source
22 emissions.
23 The correlation structures for measurements at the monitors in each of the 11 CS As/CBS As
24 examined for this assessment reveal a wide range of response between monitors in each city and
25 among the cities. While this wide range is produced by the interactions of many physical and
26 chemical elements, the location of each monitor and the uniqueness of its immediate surroundings
27 can often explain much of the agreement or lack thereof. CO concentrations can be elevated near
28 roadways and decrease with increasing distance from the road. Anchorage, AK had concentrations
29 roughly twice those of the other metropolitan areas. Most of the CS As/CBS As examined here had
30 diel concentration curves with pronounced morning and evening rush hour peak CO levels, although
31 diel CO concentrations had less variability for New York City, Atlanta, and Seattle than for the other
32 eight cities. For most metropolitan areas examined here, concentrations were generally highest in the
33 winter (December-February) and fall (September-November) and decreased, on average, during the
34 spring (March-May) and summer (June-August). Measurements near or below the LOD of most
September 2009 3-115 DRAFT - DO NOT CITE OR QUOTE
-------
1 instruments of 0.5 ppm, coupled with the coarsely reported measurement resolution of 0.1 ppm, can
2 artificially influence the comparison statistics shown in the tables and result in apparent
3 heterogeneity in the box plots (Figure 3-18 through Figure 3-20).
4 CO measurements obtained at different monitoring scales were compared to assess spatial
5 variability of CO concentration. The median hourly CO concentration across the U.S. obtained at
6 microscale monitors was 25% higher than at middle scale and 67% higher than at neighborhood
7 scale. The microscale and middle scale CO data reported here are consistent with hourly
8 concentrations reported in the literature for the near road environment within the United States, with
9 CO concentration decaying with downwind distance from the road. Determinants of spatial
10 variability of ambient CO concentration within the near-road environment include roadway density,
11 traffic counts, meteorology, and natural and urban topography.
12 In all cases, a wide range of correlations existed between CO and copollutants computed from
13 AQS data. The mean and median correlation between CO and copollutants were positive for NO2,
14 PMio, and PM2.5; near zero for SO2; and negative for O3. These findings might reflect common
15 combustion sources for CO, NO2, and PM. Among those copollutants with positive associations,
16 NO2 had the highest mean and median correlations, followed by PM25 and PMi0. Within and
17 between individual metropolitan areas, the distribution of copollutant correlations varied
18 substantially. Studies in the literature also found fairly high correlations of CO with EC and certain
19 VOCs.
20 This assessment has used data from 2005-2007 at 12 remote sites as part of the international
21 CCGG CASN in the CONUS, Alaska, and Hawaii to determine PRB. All sites demonstrate the well-
22 known seasonality in background CO with minima in the summer and fall and maxima in the winter
23 and spring. The 3-yr avg CO PRB in Alaska was 130 ppb; in Hawaii it was 99 ppb; and over the
24 CONUS it was 132 ppb.
3.7.5.Exposure Assessment and Implications for Epidemiology
25 Very few recent exposure assessment studies involve ambient CO concentration data. The
26 studies of personal exposure to ambient CO presented here generally found that the largest
27 percentage of time in which an individual is exposed to ambient CO occurs indoors, but the highest
28 ambient CO exposure levels occur in transit. In-vehicle CO concentrations are typically reported to
29 be between 2 and 5 times higher than ambient concentrations measured at the roadside, but have
30 been reported to be as much as 25 times higher. Among commuters, exposures were higher for those
31 traveling in automobiles in comparison with those traveling on buses and motorbikes and with those
32 cycling or walking. Ambient CO exposure in automobiles has been demonstrated to vary with
33 vehicle ventilation settings, and a very small portion of that exposure is thought to come from the
34 vehicle in which the exposed person travels. High near-road CO concentrations can be important for
September 2009 3-116 DRAFT - DO NOT CITE OR QUOTE
-------
1 those living in the near-road environment because virtually all of ambient CO infiltrates indoors.
2 Hence, indoor exposure to ambient CO is determined by the CO concentration outside the building.
3 Residents of the 17.9 million occupied homes located within approximately 90 m of a highway,
4 railroad, or airport may be exposed to elevated ambient CO levels. However, CO concentration in
5 the near-road environment has been shown to decrease sharply with downwind distance from a
6 highway; wind direction, emission source strength (e.g., number of vehicles on a highway), and
7 natural and urban topography also influence localized ambient CO levels.
8 Recent exposure assessment studies support one of the main conclusions of the 2000 CO
9 AQCD that central site ambient CO monitors may overestimate or underestimate individuals'
10 personal exposure to ambient CO because ambient CO concentration is spatially variable,
11 particularly when analyzing exposures in the near-road environment. Exposure error may occur if the
12 ambient CO concentration measured at the central site monitor is used as an ambient exposure
13 surrogate and differs from the actual ambient CO concentration outside a subject's residence and/or
14 worksite. For example, measurement at a "hot spot" could skew community exposure estimates
15 upwards, and likewise measurement at a location with few CO sources could skew exposure
16 estimates downwards. Correlations across CO monitors can vary widely from within and between
17 cities across the U.S. as a function of natural and urban topography, meteorology, and strength and
18 proximity to sources. Typically, intersampler correlation ranges from 0.35 to 0.65 for monitors sited
19 at different scales within a metropolitan area, although it can be greater than 0.8 in some areas.
20 Health effects estimates from time-series epidemiologic studies are not biased by spatial variability
21 in CO concentrations if concentrations at different locations are correlated in time. Additionally,
22 exposure assessment is complicated by the existence of CO in multipollutant mixtures emitted by
23 combustion processes. Because ambient CO exists in a mixture with volatile and reactive pollutants,
24 the correlation between exposure to ambient CO and copollutants can vary substantially over time
25 and across locations. For this reason, it is difficult to quantify the effects related specifically to CO
26 exposure compared with those related to another combustion-related pollutant or mix of pollutants.
27 In most circumstances, exposure error tends to bias a health effect estimate downward. Spatial and
28 temporal variability not fully captured by ambient monitors and correlation of CO with copollutants
29 are examples of sources of uncertainty that could widen confidence intervals of health effects
30 estimates.
September 2009 3-117 DRAFT - DO NOT CITE OR QUOTE
-------
References
Abdul-Wahab SA. (2004). An application and evaluation of the CAL3QHC model for predicting carbon monoxide
concentrations from motor vehicles near a roadway intersection in Muscat, Oman. Environ Manage, 34: 372-382.
194011
Abi-Esber L; El-Fadel M. (2008). In-vehicle CO ingression: Validation through field measurements and mass balance
simulations. Sci Total Environ, 394: 75-89. 190939
Abi-Esber L; El-Fadel M; Nuwayhid I; Saliba N. (2007). The effect of different ventilation modes on in-vehicle carbon
monoxide exposure. Atmos Environ, 41: 3644-3657. 190941
Abi-Esber L; El-Fadel M; Shihadeh A. (2007). Comparison of trip average in-vehicle and exterior CO determinations by
continuous and grab sampling using an electrochemical sensing method. Atmos Environ, 41: 6087-6094. 190940
Akland GG; Hartwell TD; Johnson TR; Whitmore RW. (1985). Measuring human exposure to carbon monoxide in
Washington, DC, and Denver, Colorado, during the winter of 1982-1983. Environ Sci Technol, 19: 911-918.
011618
Aim S; Jantunen MJ; Vartiainen M. (1999). Urban commuter exposure to particle matter and carbon monoxide inside an
automobile. JExpo Sci Environ Epidemiol, 9: 237-244. 047196
Aim S; Mukala K; Jantunen MJ. (2000). Personal carbon monoxide exposures of preschool children in Helsinki, Finland:
Levels and determinants. Atmos Environ, 34: 277-285. 192374
Aim S; Mukala K; Tiittanen P; Jantunen MJ. (2001). Personal carbon monoxide exposures of preschool children in
Helsinki, Finland-comparison to ambient air concentrations. Atmos Environ, 35: 6259-6266. 020237
Altshuller AP. (1991). Estimating product yields of carbon-containing products from the atmospheric photooxidation of
ambient air alkenes. J Atmos Chem, 13: 131-154. 192375
Andreae MO. (1991). Biomass burning: its history, use, and distribution and its impact on environmental quality and global
climate. In Levine JS (Ed.),Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications (pp. 1-
21). Cambridge, MA: MIT Press. 078147
Annest JH; Glower J; Yip F; Stock A; Lucas M; Iqbal S. (2008). Nonfatal, unintentional, non—fire-related carbon monoxide
exposures-United States, 2004-2006. JAMA, 57: 896-9. 190236
Arellano AFJ; Kasibhatla PS; Giglio L; Van der Werf GR; Randerson JT. (2004). Top-down estimates of global CO sources
using MOPITT measurements. Geophys Res Lett, 34: L01104. 193757
Baldauf R; Thoma E; Hays M; Shores R; Kinsey J; Gullett B; Kimbrough S; Isakov V; Long T; Snow R; Khlystov A;
Weinstein J; Chen FL; Seila R; Olson D; Gilmour I; Cho SH; Watkins N; Rowley P; Bang J. (2008). Traffic and
meteorological impacts on near-road air quality: Summary of methods and trends from the raleigh near-road study.
J Air Waste Manag Assoc, 58: 865-878. 190239
Baldauf R; Thoma E; Khlystov A; Isakov V; Bowker G; Long T; Snow R. (2008). Impacts of noise barriers on near-road air
quality. Atmos Environ, 42: 7502-7507. 191017
Behrentz E; Fitz DR; Pankratz DV; Sabin LD; Colome SD; Fruin SA; Winer AM. (2004). Measuring self-pollution in
school buses using a tracer gas technique. Atmos Environ, 38: 3735-3746. 155682
Bell M; Namdeo A; Chen H; Ashmore M; Terry A; Dimitroulopoulou S. (2004). Reducing Urban Pollution Exposure from
Road Transport (RUPERT). In Urban Transport (pp. 829-838). Southampton, UK: WIT Press. 192376
Bell ML; Peng RD; Dominici F; Samet JM. (2009). Emergency hospital admissions for cardiovascular diseases and
ambient levels of carbon monoxide: results for 126 U.S. urban counties, 1999-2005. Circulation, 120: 949-955.
194033
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS)
September 2009 3-118 DRAFT - DO NOT CITE OR QUOTE
-------
Bergamaschi P; Hein R; Heimann M; Crutzen PJ. (2000). Inverse modeling of the global CO cycle 1. Inversion of CO
mixing ratios. J Geophys Res, 105: 1909-1927. 192377
Berntsen T; Fuglestvedt JS; Joshi M; Shine KP; Stuber N; Sausen R; Li L; Hauglustaine DA; Ponater M. (2005). Response
of climate to regional emissions of ozone precursors: Sensitivities and warming potentials. Tellus B, 57: 283-304.
193241
Berntsen T; Fuglestvedt JS; Myhre G; Stordal F; Berglen TF. (2006). Abatement of greenhouse gases: Does location
matter?. Clim Change, 74: 377-411. 193244
Bey I; Jacob DJ; Yantosca RM; Logan JA; Field B; Fiore AM; Li Q; Liu H; Mickley LJ; Schultz MG (2001). Global
modeling of tropospheric chemistry with assimilated meteorology: model description and evaluation. J Geophys
Res, 106: 23073-23095. 051218
Bishop GA; StedmanDH. (2008). A Decade of On-road Emissions Measurements. Environ Sci Tech, 42: 1651-1656.
194670
BogoH; Gomez DR; Reich SL;NegriRM; Roman ES. (2001). Traffic pollution in a downtown site of Buenos Aires City.
Atmos Environ, 35: 1717-1727. 192378
Borrego C; Tchepel O; Costa AM; Martins H; Ferreira J; Miranda AI. (2006). Traffic-related particulate air pollution
exposure in urban areas. Atmos Environ, 40: 7205-7214. 155697
Bowker GE; Baldauf R; Isakov V; Khlystov A; Petersen W. (2007). The effects of roadside structures on the transport and
dispersion of ultrafine particles from highways. Atmos Environ, 41: 8128-8139. 149997
Brewer DA; Ogliaruso MA; Augustsson TR; Levine JS. (1984). The oxidation of isoprene in the troposphere: Mechanism
and model calculations. Atmos Environ, 18: 2723-2744 . 194402
Britter RE; Hanna SR. (2003). Flow and dispersion in urban areas. Annu Rev Fluid Mech, 35: 469-496. 090295
Bruinen de Bruin Y; Carrer P; Jantunen M; Hanninen O; Di Marco GS; Kephalopoulos S; Cavallo D; Maroni M. (2004).
Personal carbon monoxide exposure levels: contribution of local sources to exposures and microenvironment
concentrations in Milan. JExpo Sci Environ Epidemiol, 14: 312-322. 190943
Bruinen de Bruin Y; Hanninen O; Carrer P; Maroni M; Kephalopoulos S; Scotto di Marco G; Jantunen M. (2004).
Simulation of working population exposures to carbon monoxide using EXPOLIS-Milan microenvironment
concentration and time-activity data. J Expo Anal Environ Epidemiol, 14: 154-163. 190942
Burgard DA; Bishop GA; Stedman DH; Gessner VH; Daeschlein C. (2006). Remote sensing of in-use heavy-duty diesel
trucks. Environ Sci Technol, 40: 6938-6942. 193222
Burke JM; Zufall MJ; Ozkaynak H. (2001). A population exposure model for particulate matter: case study results for
PM25 in Philadelphia, PA. J Expo Sci Environ Epidemiol, 11: 470-489. 014050
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. Health
Effects Institute. Boston, MA. 042798
CalTrans. (2009). CalTrans. Retrieved 02-AUG-09, from http://traffic-counts.dot.ca.gov/. 194036
Carslaw DC; Beevers SD; Tate JE. (2007). Modelling and assessing trends in traffic-related emissions using a generalised
additive modelling approach. Atmos Environ, 41: 5289-5299. 148210
Carter WPL. (1998). Updated maximum incremental reactivity scale for regulatory applications. Air Pollution Research
Center and College of Engineering, Center for Environmental Research and Technology. University of California,
Riverside CA. 192380
Chaloulakou A; Mavroidis I. (2002). Comparison of indoor and outdoor concentrations of CO at a public school Evaluation
of an indoor air quality model. Atmos Environ, 36: 1769-1781. 026050
Chaloulakou A; Mavroidis I; Duci A. (2003). Indoor and outdoor carbon monoxide concentration relationships at different
microenvironments in the Athens area. Chemosphere, 52: 1007-1019. 190945
Chang C-H; Meroney RN. (2003). Concentration and flow distributions in urban street canyons: wind tunnel and
computational data. J WindEng Ind Aerod, 91: 1141-1154. 090298
Chang L-T; Koutrakis P; Catalano PJ; Suh HH. (2000). Hourly personal exposures to fine particles and gaseous pollutants-
results from Baltimore, Maryland. J Air Waste Manag Assoc, 50: 1223-1235. 001276
September 2009 3-119 DRAFT - DO NOT CITE OR QUOTE
-------
Chang L-T; Suh HH; Wolfson JM; Misra K; Allen GA; Catalano PJ; Koutrakis P. (2001). Laboratory and field evaluation
of measurement methods for one-hour exposures to O3, PM25, and CO. J Air Waste Manag Assoc, 51: 1414-1422.
019216
Chen H; Namdeo A; Bell M. (2008). Classification of road traffic and roadside pollution concentrations for assessment of
personal exposure. Environ Modell Softw, 23: 282-287. 193986
Collins WJ; Derwent RG; Johnson CE; Stevenson DS. (2002). The oxidation of organic compounds in the troposphere and
their global warming potentials. Clim Change, 52: 453-479. 044156
Collins WJ; Stevenson DS; Johnson CE; Derwent RG. (1997). Tropospheric ozone in a global-scale three-dimensional
Lagrangian model and its response to NOX emissions controls. J Atmos Chem, 26: 223-274. 193193
Connors VS; Gormsen BB; Nolf S; Reichle HG (1994). Space-borne observations of the global distribution of carbon
monoxide in the middle troposphere during April and October 1994. J Geophys Res, 104: 21455-21470. 193755
Conrad R; Seiler W. (1985). Influence of temperature, moisture, and organic carbon on the flux of H2 and CO between soil
and atmosphere: field studies in subtropical regions. J Geophys Res, 90: 5699-5709. 029520
CroxfordB; Perm A. (1998). Siting considerations for urban pollution monitors. Atmos Environ, 32: 1049-1057. 087176
Crutzen PJ. (1987). Role of the tropics in atmospheric chemistry. In Dickinson RE (Ed.),Geophysiology of amazonia:
vegetation and climate interactionsNew York, NY: John Wiley and Sons. 002848
Crutzen PJ; Delany AC; Greenberg J; Haagenson P; Heidt L; Lueb R; Pollock W; Seiler W; Wartburg A; Zimmerman P.
(1985). Tropospheric chemical composition measurements in Brazil during the dry season. J Atmos Chem, 2: 233-
256. 194403
Daniel JS; Solomon S. (1998). On the climate forcing of carbon monoxide. J Geophys Res, 103: 13249-13260. 193235
Delfino RJ; Chang J; Wu J; Ren C; Tjoa T; Nickerson B; Cooper D; Gillen DL. (2009). Repeated hospital encounters for
asthma in children and exposure to traffic-related air pollution near the home. Ann Allergy Asthma Immunol, 102:
138-44. 190254
Dentener F; Stevenson D; Cofala J; Mechler R; Amann M; Bergamaschi P; Raes F; Derwent R. (2004). The impact of air
pollutant and methane emission controls on tropospheric ozone and radiative forcing: CTM calculations for the
period 1990-2030. Atmos Chem Phys Discuss, 4: 8471-8538. 194040
Derwent RG; Collins WJ; Johnson CE; Stevenson DS. (2001). Transient behaviour of tropospheric ozone precursors in a
global 3-D CTM and their indirect greenhouse effects. Clim Change, 49: 463-487. 047912
Derwent RG; Stevenson DS; Doherty RM; Collins WJ; Sanderson MG; Johnson CE. (2008). Radiative forcing from
surface NOX emissions: Spatial and seasonal variations. Clim Change, 88: 385-401. 193245
Dimitroulopoulou C; Ashmore MR; Byrne MA; Kinnersley RP. (2001). Modelling of indoor exposure to nitrogen dioxide
in the UK. Atmos Environ, 35: 269-279. 014737
Dimitroulopoulou C; Ashmore MR; Hill MTR; Byrne MA; Kinnersley R. (2006). INDAIR: a probabilistic model of indoor
air pollution in UK homes. Atmos Environ, 40: 6362-6379. 090302
Dominici F; Zeger SL; Samet JM. (2000). A measurement error model for time-series studies of air pollution and mortality.
Biostatistics, 1: 157-175. 005828
Doran JC; Berkowitz CM; Coulter RL; Shaw WJ; Spicer CW (2003). The 2001 Phoenix Sunrise experiment: vertical
mixing and chemistry during the morning transition in Phoenix. Atmos Environ, 37: 2365-2377. 143352
Duci A; Chaloulakou A; Spyrellis N. (2003). Exposure to carbon monoxide in the Athens urban area during commuting.
Sci Total Environ, 309: 47-58. 044199
Duncan BN; Logan JA. (2008). Model analysis of the factors regulating the trends and variability of carbon monoxide
between 1988 and 1997. Atmos Chem Phys, 8: 7389-7403. 194042
Duncan BN; Martin RV; Staudt AC; Yevich R; Logan JA. (2003). Interannual and seasonal variability of biomass burning
emissions constrained by satellite observations. J Geophys Res, 108: 4100. 193760
September 2009 3-120 DRAFT - DO NOT CITE OR QUOTE
-------
Emmons LK; Deeter MN; Gille JC; Edwards DP; Attie JL; Warner J; Ziskin D; Francis G; Khattatov B; Yudin V;
Lamarque IF; Ho SP; Mao D; Chen JS; Drummond J; Novell! P; Sachse G; Coffey MT; Hannigan JW; Gerbig C;
Kawakami S; Kondo Y; Takegawa N; Schlager H; Baehr J; Ziereis H. (2004). Validation of Measurements of
Pollution In The Troposphere (MOPITT) CO retrievals with aircraft in situ profiles. J Geophys Res, 109: D03309.
193756
Fehsenfeld FC; Ancellet G; Bates TS; Goldstein AH; Hardesty RM; Honrath R; Law KS; Lewis AC; Leaitch R; McKeen S;
Meagher J; Parrish DD; Pszenny AAP; Russell PB; Schlager H; Seinfeld J; Talbot R; Zbinden R. (2006).
International consortium for atmospheric research on transport and transformation (ICARTT): North America to
Europe -overview of the 2004 summer field study . J Geophys Res, 111: D23S01.1-D23S01.36. 190531
Finlayson-Pitts BJ; Pitts JN Jr. (2000). Chemistry of the upper and lower atmosphere: theory, experiments and applications.
San Diego, CA: Academic Press. 055565
Fiore AM; Jacob DJ; Field BD; Streets DG; Fernandes SD; Jang C. (2002). Linking ozone pollution and climate change:
the case for controlling methane. Geophys Res Lett, 2919: 1919. 051221
Fiore AM; West JJ; Horowitz LW; Naik V; Schwartzkopf MD. (2008). Characterizing the tropospheric ozone response to
methane emission controls and the benefits to climate and air quality. J Geophys Res, 113: D08307. 193749
Fishman J; Bowman KW; Burrows JP; Richter A; Chance KV; Edwards DP; Martin RV; Morris GA; Pierce RB; Ziemke
JR; Al-Saadi JA; Creilson JK; Schaack TK; Thompson AM. (2008). Remote sensing of tropospheric pollution from
space. Bull Am Meteorol Soc, 89: 805-821. 193927
Flachsbart PG (1999). Models of exposure to carbon monoxide inside a vehicle on a Honolulu highway. J Expo Sci
Environ Epidemiol, 9: 245-260. 015857
Fuglestvedt J; Berntsen T; Myhre G; Rypdal K; Skeie RB. (2008). Climate forcing from the transport sectors. Proc Natl
Acad Sci USA, 105: 454-458. 193242
Fuglestvedt JS; Berntsen TK; Isaksen ISA; Mao H; Liang X-Z; Wang W-C. (1999). Climatic forcing of nitrogen oxides
through changes in tropospheric ozone and methane; global 3D model studies. Atmos Environ, 33: 961-978.
047431
Gentner DR; Harley RA; Miller AM; Goldstein AH. (2009). Diurnal and seasonal variability of gasoline-related volatile
organic compound emissions in Riverside, California. Environ Sci Technol, 43: 4247-4252. 194034
Georgoulis LB; Hanninen O; Samoli E; Katsouyanni K; Kunzli N; Polanska L; De Bruin YB; Aim S; Jantunen M. (2002).
Personal carbon monoxide exposure in five European cities and its determinants. Atmos Environ, 36: 963-974.
025563
Gokhale S; Khare M. (2007). Statistical behavior of carbon monoxide from vehicular exhausts in urban environments.
Environ Modell Softw, 22: 526-535. 194015
Goldstein A; Galbally I. (2007). Known and unexplored organic constituents in the earth's atmosphere. Environ Sci
Technol, 41: 1514-1521. 193247
GrosjeanD; Fung K. (1984). Hydrocarbons and carbonyls in Los Angeles air. J Air Waste Manag Assoc, 34: 537-543.
040120
Gulliver J; Briggs DJ. (2005). Time-space modeling of journey-time exposure to traffic-related air pollution using GIS.
Environ Res, 97: 10-25. 191079
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez-Avedoy V; Paramo-Figueroa VH; Blanco-Jimenez S;
Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo-Navarro M; Nieuwenhuijsen MJ. (2007). Bus,
minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City. Atmos Environ, 41: 890-
901. 138816
Gomez-Perales JE; Colvile RN; NieuwenhuijsenMJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-Figueroa
VH; Blanco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R. (2004). Commuters' exposure to
PM25, CO, and benzene in public transport in the metropolitan area of Mexico City. Atmos Environ, 38: 1219-
1229.054418
HEI. (2009). Traffic-related air pollution: A critical review of the literature on emissions, exposure, and health effects.
Health Effects Institute. Boston. Special Report 17. 191009
September 2009 3-121 DRAFT - DO NOT CITE OR QUOTE
-------
Hanst PL; Spence JW; Edney EO. (1980). Carbon monoxide production in photooxidation of organic molecules in the air.
Atmos Environ, 14: 1077-1088. 011988
Harley RA; Marr LC; Lehner JK; Giddings SN. (2005). Changes in motor vehicle emissions on diurnal to decadal time
scales and effects on atmospheric composition. Environ Sci Technol, 39: 5356-5362. 088154
Harley RA; McKeen SA; Pearson J; Rodgers MO; Lonneman WA. (2001). Analysis of motor vehicle emissions during the
Nashville/Middle Tennessee Ozone Study. J Geophys Res, 106: 355973567. 193922
Hauglustaine DA; Hourdin F; Jourdain L; Filiberti MA; Walters S; Lamarque JF; Holland EA. (2004). Interactive
chemistry in the Laboratoire de Meteorologie Dynamique general circulation model: Description and background
tropospheric chemistry evaluation. J Geophys Res, 109: D04314. 193191
Heist DK; Perry SG; Brixey LA. (2009). A wind tunnel study of the effect of roadway configurations on the dispersion of
traffic-related pollution. Atmos Environ, In Press: In Press. 194037
Horowitz LW; Walters S; Mauzerall DL; Emmons LK; Rasch PJ; Granier C; Tie X; Lamarque J-F; Schultz MG; Tyndall
GS; Orlando JJ; Brasseur GP (2003). A global simulation of tropospheric ozone and related tracers: Description
and evaluation of MOZART, version 2. J Geophys Res, 108: D24. 057770
Hudman RC; Jacob DJ; Turquety S; Leibensperger EM; Murray LT; Wu S; Gilliland AB; Avery M; Bertram TH; Brune W;
Cohen RC; Dibb JE; Flocke FM; Fried A; Holloway J; Neuman JA; Orville R; Perring A; Ren X; Sachse GW;
Singh HB; Swanson A; Wooldridge PJ. (2007). Surface and lightning sources of nitrogen oxides over the United
States: Magnitudes, chemical evolution, and outflow. J Geophys Res, 112: D12S05. 089474
Hudman RC; Murray LT; Jacob DJ; Millet DB; Turquety S; Wu S; Blake DR; Goldstein AH; Holloway J; Sachse GW.
(2008). Biogenic versus anthropogenic sources of CO in the United States. Geophys Res Lett, 35. 191253
IPCC. (2001). IPCC Third Assessment Report (TAR): Climate Change 2001: Working Group I: The Scientific Basis.
Cambridge, U.K, and New York, NY: Intergovernmental Panel on Climate Change, Cambridge University Press.
156587
IPCC. (2007). IPCC Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical
Science Basis. Cambridge, UK and New York, NY: Intergovernmental Panel on Climate Change, Cambridge
University Press. 092765
InmanRE; Ingersoll RB; Levy EA. (1971). Soil: a natural sink for carbon monoxide. Science, 172: 1229-1231.010972
Isakov V; Touma JS; Burke J; Lobdell DT; Palma T; Rosenbaum A; Ozkaynak H. (2009). Combining regional- and local-
scale air quality models with exposure models for use in environmental health studies. J Air Waste Manag Assoc,
59: 461-472. 191192
Isaksen ISA; Hov O. (1987). Calculation of trends in the tropospheric concentration of O3, OH, CO, CH4 and NOx. Tellus
B Chem Phys Meteorol, 39: 271-285. 019490
Jacob DJ; Wofsy SC. (1990). Budgets of reactive nitrogen, hydrocarbons, and ozone over the Amazon forest during the wet
season. J Geophys Res, 95: 16,737-16,754. 029668
Jeffries HE. (1995). Photochemical air pollution. In Singh, H. B. (Ed.),Composition, chemistry, and climate of the
atmosphereNew York, NY: Van Nostrand Reinhold. 003055
Jerrett M; Arain A; Kanaroglou P; Beckerman B; Potoglou D; Sahsuvaroglu T; Morrison J; Giovis C. (2005). A review and
evaluation of intraurban air pollution exposure models. J Expo Sci Environ Epidemiol, 15:185-204. 092864
Johnson CE; Derwent RG (1996). Relative radiative forcing consequences of global emissions of hydrocarbons, carbon
monoxide, and NOX from human activities estimated with a zonally averaged two-dimensional model. Clim
Change, 34: 439-462. 193192
Johnson TR. (1984). A study of personal exposure to carbon monoxide in Denver, Colorado. Presented at Presented at: 77th
annual meeting of the Air Pollution Control Association; June; San Francisco, CA. Pittsburgh, PA: Air Pollution
Control Association; paper no. 84-121.3., California. 024652
Joshi M; Shine KP; Ponater M; Stuber N; Sausen R; Li. (2003). A comparison of climate response to different radiative
forcings in three general circulation models: Towards an improved metric of climate change. Clim Dynam, 20: 843-
854. 193752
September 2009 3-122 DRAFT - DO NOT CITE OR QUOTE
-------
KanakidouM; CrutzenPJ. (1999). The photochemical source of carbon monoxide: importance, uncertainties and
feedbacks. Chemosphere Global Change Sci, 1: 91-109. 011760
Kanakidou M; Singh HB; Valentin KM; Crutzen PJ. (1991). A two-dimensional study of ethane and propane oxidation in
the troposphere. J Geophys Res, 96: 15,395-15,413. 029701
Kastner-Klein P; Fedorovich E; Ois Sini J-F; Mestayer PG (2000). Experimental and numerical verification of similarity
concept for dispersion of car exhaust gases in urban street canyons . Environ Monit Assess, 65: 353-361. 194035
Kastner-Klein P; Plate EJ. (1999). Wind-tunnel study of concentration fields in street canyons. Atmos Environ, 33: 3973-
3979. 001961
Kaur S; Nieuwenhuijsen M; Colvile R. (2005). Personal exposure of street canyon intersection users to PM2.5, ultrafine
particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641. 086504
Kaur S; Nieuwenhuijsen MJ. (2009). Determinants of personal exposure to PM2.5, ultrafine particle counts, and CO in a
transport microenvironment. Environ Sci Technol, 43: 4737-4743. 194014
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2007). Fine particulate matter and carbon monoxide exposure concentrations in
urban street transport microenvironments. Atmos Environ, 41: 4781-4810. 190070
Khalil MAK; Rasmussen RA. (1990). The global cycle of carbon monoxide: trends and mass balance. Chemosphere, 20:
227-242. 012352
Khare M; Chaudhry KK; Gowda RMM; Ahmad K. (2005). Effects of the homogeneous traffic on vertical dispersion
parameter in the near field of roadways -A wind tunnel study. Environ Monit Assess, 10: 55-62. 194016
Kim D; Sass-Kortsak A; Purdham JT; Dales RE; Brook JR. (2006). Associations between personal exposures and fixed-site
ambient measurements of fine particulate matter, nitrogen dioxide, and carbon monoxide in Toronto, Canada. J
Expo Sci Environ Epidemiol, 16: 172-183. 089820
King GM. (1999). Characteristics and significance of atmospheric carbon monoxide consumption by soils. Chemosphere,
1:53-63.002828
Kirchhoff VWJH; Marinho EVA. (1990). Surface carbon monoxide measurements in Amazonia. J Geophys Res, 95:
16933-16943. 194406
Klepeis NE; Nelson WC; OttWR; Robinson JPTsang AM; Switzer P; Behar JV; Hern SC; Engelmann WH. (2001). The
National Human Activity Pattern Survey (NHAPS): a resource for assessing exposure to environmental pollutants. J
Expo Sci Environ Epidemiol, 11: 231-252. 002437
Kristjansson JE. (2002). Studies of the aerosol indirect effect from sulfate and black carbon aerosols. J Geophys Res, 107:
4246. 045282
Lai HK; Kendall M; Ferrier H; Lindup I; Aim S; Hanninen O; Jantunen M; Mathys P; Colvile R; Ashmore MR; Cullinan P;
NieuwenhuijsenMJ. (2004). Personal exposures and microenvironment concentrations of PM25, VOC, NO2 and
CO in Oxford, UK. Atmos Environ, 38: 6399-6410. 056811
Langan L. (1992). Portability in measuring exposure to carbon monoxide. Presented at In: Measuring, understanding, and
predicting exposures in the 21st century: proceedings of the conference; November 1991; Atlanta, GA. J. Exposure
Anal. Environ. Epidemiol. l(suppl. 1): 223-239. 046120
Lelieveld J; Berresheim H; Borrmann S; Crutzen PJ; Dentener FJ; Fischer H; et al. (2002). Global air pollution crossroads
over the Mediterranean . Science, 298: 794-799. 190361
Lobert JM; Scharffe DH; Hao W-M; Kuhlbusch TA; Seuwen R; Warneck P; Crutzen PJ. (1991). Experimental evaluation
of biomass burning emissions: nitrogen and carbon containing compounds. In Levine, J. S. (Ed.),Global biomass
burning: atmospheric, climatic, and biospheric implicationsCambridge, MA: MIT Press. 029473
Manning MR; Brenninkmeijer CAM; Allan W. (1997). Atmospheric carbon monoxide budget of the southern hemisphere:
Implications of 13C/12C measurements. J Geophys Res, 102: 10673-10682 . 194401
Marshall JD; Nethery E; Brauer M. (2008). Within-urban variability in ambient air pollution: Comparison of estimation
methods. Atmos Environ, 42: 1359-1369. 193983
McCurdy T; Glen G; Smith L; Lakkadi Y. (2000). The National Exposure Research Laboratory's Consolidated Human
Activity Database. J Expo Sci Environ Epidemiol, 10: 566-578. 000782
September 2009 3-123 DRAFT - DO NOT CITE OR QUOTE
-------
Menon S; Hansen J; Nazarenko L; Luo Y. (2002). Climate effects of black carbon aerosols in China and India. Science,
297: 2250-2253. 155978
Michie RM Jr; McElroy FF; Sokash JA; Thompson VL; Dayton DP; Sutcliffe CR. (1983). Performance test results and
comparative data for designated reference methods for carbon monoxide. U.S. Environmental Protection Agency.
Research Triangle Park. 194043
Millet DB; Jacob DJ; Turquety S; Hudman RC; Wu S; Fried A; Walega J; Heikes BG; Blake DR; Singh HB; Anderson BE;
Clarke AD. (2006). Formaldehyde distribution over North America: Implications for satellite retrievals of
formaldehyde columns and isoprene emission. J Geophys Res, 111: D24S02. 195106
Milner JT; ApSimon HM; Croxford B. (2006). Spatial variation of CO concentrations within an office building and outdoor
influences. Atmos Environ, 40: 6338-6348. 123100
Mobley D; Deslauriers M; Rojas-Brachos L. (2005). Improving Emission Inventories for Effective Air-Quality
Management across North America?a NARSTO Assessment. NARSTO. Pasco, WA . NARSTO-05-001. 194008
Murena F; Garofalo N; Favale F. (2008). Monitoring CO concentration at leeward and windward sides in a deep street
canyon. Atmos Environ, 42: 8204-8210. 194038
NRC. (1999). Ozone-forming potential of reformulated gasoline. Washington, DC: National Research Council, National
Academy of Sciences. 010614
Naik V; Mauzerall D; Horowitz L; Schwarzkopf MD; Ramaswamy V; Oppenheimer M. (2005). Net radiative forcing due
to changes in regional emissions of tropospheric ozone precursors. J Geophys Res, 110: D24306. 193194
Ni Riain CM; Mark D; Davies M; Harrison RM; Byrne MA. (2003). Averaging periods for indoor-outdoor ratios of
pollution in naturally ventilated non-domestic buildings near a busy road. Atmos Environ, 37: 4121-4132. 053792
Ott W; Switzer P; Willits N. (1994). Carbon monoxide exposures inside an automobile traveling on an urban arterial
highway. J Air Waste Manag Assoc, 44: 1010-1018. 076546
Parrish DD. (2006). Critical evaluation of US on-road vehicle emission inventories. Atmos Environ, 40: 2288-2300.
090352
Parrish DD; Trainer M; Hereid D; Williams EJ; Olszyna KJ; Harley RA; Meagher JF; Fehsenfeld FC. (2002). Decadal
change in carbon monoxide to nitrogen oxide ratio in US vehicular emissions. J Geophys Res, 107: 4140. 052472
Petron G; Granier C; Khattatov B; Yudin V; Lamarque JF; Emmons L; Gille J; Edwards DP. (2004). Monthly CO surface
sources inventory based on the 2000-2001 MOPITT satellite data. Geophys Res Lett, 31: L21107. 193758
Pfister G; Hess PG; Emmons LK; Lamarque JF; Wiedinmyer C; Edwards DP; Petron G; Gille JC; Sachese GW (2005).
Quantifying CO emissions from the 2004 Alaskan wildfires using MOPITT CO data. Geophys Res Lett, 32:
LI 1809. 093009
Pokharel SS; Bishop GA; Stedman DH. (2002). An on-road motor vehicle emissions inventory for Denver: An efficient
alternative to modeling. Atmos Environ, 36: 5177-5184. 052473
Pokharel SS; Bishop GA; Stedman DH. (2003). Emissions reductions as a result of automobile improvement. Environ Sci
Technol, 37: 5097-5101.053740
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and carbonaceous
content of fine particulate matter in retirement homes of the Los Angeles Basin. J Air Waste Manag Assoc, 57: 366-
379. 156877
Pollack AK; Lindhjem C; Stoeckenius TE; Tran C; Mansell G; Jimenez M; Wilson G; Coulter-Burke S. (2004). Final
Report: Evaluation of the US EPAMobile6 highway vehicle emission factor model. ENVIRON International
Corporation. Novato, CA. CRC Project E-64. 184461
Prather MJ. (1996). Natural modes and timescales in atmospheric chemistry: Theory, GWPs for CH4 and CO, and runaway
growth. Geophys Res Lett, 23: 2597-2600. 193195
Riediker M; Williams R; Devlin R; Griggs T; Bromberg P. (2003). Exposure to particulate matter, volatile organic
compounds, and other air pollutants inside patrol cars. Environ Sci Technol, 37: 2084-2093. 043761
RindD; Lerner J. (1996). Use of on-line tracers as a diagnostic tool in general circulation model development: 1.
Horizontal and vertical transport in the troposphere. J Geophys Res, 101: 12667-12683. 193750
September 2009 3-124 DRAFT - DO NOT CITE OR QUOTE
-------
Rodes C; Sheldon L; Whitaker D; Clayton A; Fitzgerald K; Flanagan J; DiGenova F; Hering S; Frazier C. (1998).
Measuring concentrations of selected air pollutants inside California vehicles [final report]. Sacramento, CA:
California Environmental Protection Agency, Air Resources Board; South Coast Air Quality Management District.
010611
Rosenlund M; Berglind N; Pershagen G; Hallqvist J; Jonson T; Bellander T. (2006). Long-term exposure to urban air
pollution and myocardial infarction. Epidemiology, 17: 383-390. 089796
Rotstayn LD; Penner JE. (2001). Indirect aerosol forcing, quasi-forcing, and climate response. J Clim, 14: 2960-2975.
193754
Sarnat JA; Brown KW; Schwartz J; Coull BA; Koutrakis P. (2005). Ambient gas concentrations and personal particulate
matter exposures: implications for studying the health effects of particles. Epidemiology, 16: 385-395. 087531
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology: Confounders
or surrogates?. Environ Health Perspect, 109: 1053-1061. 019401
Sarnat SE; Coull BA; Schwartz J; Gold DR; Suh HH. (2006). Factors affecting the association between ambient
concentrations and personal exposures to particles and gases. Environ Health Perspect, 114: 649-654. 089784
Sarnat SE; Klein M; Sarnat JA; Flanders WD; Waller LA; Mulholland JA; Russell AG; Tolbert PE. (2009). An examination
of exposure measurement error from air pollutant spatialvariability in time-series studies. J Expo Sci Environ
Epidemiol, In Press: 1-12. 180084
Scotto Di Marco G; Kephalopoulos S; Ruuskanen J; Jantunen M. (2005). Personal carbon monoxide exposure in Helsinki,
Finland. Atmos Environ, 39: 2697-2707. 144054
Seila RL; Lonneman WA; Meeks SA. (1989). Determination of C2 to C12 ambient air hydrocarbons in 39 US cities from
1984 through 1986. U. S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory. RTF, NC. EPA/600/3-89/058. 043362
Sheppard L; Slaughter JC; Schildcrout J; Liu L-JS; Lumley T. (2005). Exposure and measurement contributions to
estimates of acute air pollution effects. J Expo Sci Environ Epidemiol, 15: 366-376. 079176
Shindell DT; Faluvegi G; Emmons LK. (2005). Inferring carbon monoxide pollution changes from space-based
observations. J Geophys Res, 110: D23303. 193746
Shindell DT; Faluvegi G; Stevenson DS; Krol MC; Emmons LK; Lamarque J-F; Petron G; Dentener FJ; Ellingsen K;
Schultz MG; Wild O; Amann M; Atherton CS; Bergmann DJ; Bey I; Butler T; Cofala J; Collins WJ. (2006).
Multimodel simulations of carbon monoxide: comparison with observations and projected near-future changes. J
Geophys Res, 111: D19306. 091028
Shindell DT; Faluvegi G; Unger N; Aguilar E; Shmidt GA; Koch DM; Bauer SE; Miller RL. (2006). Simulations of pre-
industrial, present-day, and 2100 conditions in the NASA GISS composition and climate model G-PUCCINI.
Atmos Chem Phys, 6: 4427-4459. 193751
Shindell DT; Lamarque JF; Unger N; Koch D; Faluveg G; Bauer S; Teich H. (2008). Climate forcing and air quality change
due to regional emissions reductions by economic sector. Atmos Chem Phys, 8: 7101-7113. 193748
Singh M; Phuleria HC; Bowers K; Sioutas C. (2006). Seasonal and spatial trends in particle number concentrations and
size distributions at the children's health study sites in Southern California. JExpo Sci Environ Epidemiol, 16: 3-18.
190136
Sinha A; Toumi R. (1996). A comparison of climate forcings due to chloroflurocarbons and carbon monoxide. Geophys Res
Lett, 23:65-68.193747
So ESP; Chan AT Y; Wong AYT. (2005). Large-eddy simulations of wind flow and pollutant dispersion in a street canyon.
Atmos Environ, 39: 3573-3582. 110746
ThompsonAM; Cicerone RJ. (1986). Possible perturbations to atmospheric CO, CH4, and OH. J Geophys Res, 91: 10,853-
10,864. 019374
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
U.S. Census Bureau. (2008). American Housing Survey for the United States: 2007. U.S. Government Printing Office.
Washington, DC. 194013
September 2009 3-125 DRAFT - DO NOT CITE OR QUOTE
-------
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 600/P-
99/001F. 000907
U.S. EPA. (2006). 2002 National Emissions Inventory Data and Documentation. 157070
U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental Protection
Agency. Research Triangle Park, NC. 157073
U.S. EPA. (2008). Inventory of U.S. greenhouse gas emissions and sinks: 1990 - 2006. US Environmental Protection
Agency. Washington, DC. 184463
U.S. EPA. (2008). Total Risk Integrated Methodology (TRIM) Air Pollutants Exposure Model Documentation
(TRIM.Expo/APEX, Version 4.3). Volume 1: User's Guide. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. EPA-452/B-08-001b. 191775
U.S. EPA. (2008). U.S. EPA's 2008 Report on the Environment (Final Report). EPA. DC. 157076
U.S. EPA. (2009). Consolidated Human Activity Database. Retrieved 27-AUG-09, from http://www.epa.gov/chadnetl/.
194010
U.S. EPA. (2009). Human exposure modeling - Air pollutants exposure model(APEX/ TRIM.Expo Inhalation). Retrieved
27-AUG-09, from http://www.epa.gov/ttn/fera/human_apex.html. 194009
Unger N; Shindell DT; Wang TS. (2009). Climate forcing by the on-road transportation and power generation sectors.
Atmos Environ, 43: 3077-3085 . 193238
Wallace L; Williams R; Suggs J; Jones P. (2006). Estimating contributions of outdoor fine particles to indoor
concentrations and personal exposures: effects of household characteristics and personal activities. U.S.
Environmental Protection Agency; National Exposure Research Laboratory. Research Triangle Park, NC.
EPA/600/R-06/023. 089190
Wallace LA; Ziegenfus RC. (1985). Comparison of carboxyhemoglobin concentrations in adult nonsmokers with ambient
carbon monoxide levels. J Air Waste Manag Assoc, 35: 944-949. 011656
Wang C; Prinn RG (1999). Impact of emissions, chemistry and climate on atmospheric carbon monoxide: 100-yr
predictions from a global chemistry-climate model. Chemosphere Global Change Sci, 1: 73-81. 011758
Wigley TML; Smith SJ; Prather MJ. (2002). Radiative forcing due to reactive gas emissions. J Clim, 15: 2690-2696.
047883
Wild O; Prather MJ. (2000). Excitation of the primary tropospheric chemical mode in a global three-dimensional model. J
Geophys Res, 105: 24,647-24,660. 052402
Wild O; Prather MJ; Akimoto H. (2001). Indirect long-term global radiative cooling from NOX emissions. Geophys Res
Lett, 28: 1719-1722. 193196
Williams R; Suggs J; Rea A; Leovic K; Vette A; Croghan C; Sheldon L; Rodes C; Thornburg J; Ejire A; Herbst M; Sanders
W Jr. (2003). The Research Triangle Park particulate matter panel study: PM mass concentration relationships.
Atmos Environ, 37: 5349-5363. 053335
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient particulate
matter for epidemiology and risk assessment: why and how. J Air Waste Manag Assoc, 50: 1167-1183. 010288
Wilson WE; Suh HH. (1997). Fine particles and coarse particles: concentration relationships relevant to epidemiologic
studies. J Air Waste Manag Assoc, 47: 1238-1249. 077408
Wuebbles DJ; Hayhoe K. (2002). Atmospheric methane and global change. Earth Sci Rev, 57: 177-210. 044159
Xiaomin X; Zhen H; Jiasong W. (2006). The impact of urban street layout on local atmospheric environment. Build
Environ, 41: 1352-1363. 156165
Zeger SL; Diggle PJ. (2001). Correction: exposure measurement error in time-series air pollution studies. Environ Health
Perspect, 109: A517. 026017
Zeger SL; Thomas D; Dominici F; Samet JM; Schwartz J; Dockery D; Cohen A. (2000). Exposure measurement error in
time-series studies of air pollution: concepts and consequences. Environ Health Perspect, 108: 419-426. 001949
September 2009 3-126 DRAFT - DO NOT CITE OR QUOTE
-------
Zhou Y; Levy JI. (2008). The impact of urban street canyons on population exposure to traffic-related primary pollutants.
Atmos Environ, 42: 3087-3098. 190091
Zhu Y; Hinds WC; Kim S; Shen S; Sioutas C. (2002). Study of ultrafine particles near a major highway with heavy-duty
diesel traffic. Atmos Environ, 36: 4323-4335. 041553
Zidek JV; Shaddick G; Meloche J; Chatfield C; White R. (2007). A framework for predicting personal exposures to
environmental hazards. Environ Ecol Stat, 14: 411-431. 190076
Zimmerman PR; Chatfield RB; Fishman J; Crutzen PJ; Hanst PL. (1978). Estimates on the production of CO and H2 from
the oxidation of hydrocarbon emissions from vegetation. Geophys Res Lett, 5: 679-682. 010758
van der Werf GR; Randerson JT; Giglio L; Collate GJ; Kasibhatla PS; Arellano Jr AF. (2006). Interannual variability in
global biomass burning emissions from 1997 to 2004. Atmos Chem Phys, 6: 3423-3441. 157084
September 2009 3-127 DRAFT - DO NOT CITE OR QUOTE
-------
Chapter 4. Dosimetry and
Pharmacokinetics of Carbon Monoxide
4.1. Introduction
1 Inhaled ambient CO elicits various health effects by binding with and altering the function of a
2 number of heme-containing molecules, mainly Hb. Traditional concepts for CO pathophysiology
3 have been based on the high affinity of CO for deoxyhemoglobin, resulting in COHb formation and
4 consequent reduction in O2-carrying capacity of blood and impaired O2 delivery to tissues. Research
5 on CO pharmacokinetics dates back to the 1890s, but since the late 1970s has become limited.
6 Current literature primarily focuses on endogenous CO produced by the metabolic degradation of
7 heme by heme oxygenase (HO) and its role as a gaseous messenger. This chapter reviews the
8 physiology and pharmacokinetics of CO. The chapter draws heavily from Chapter 5 of the previous
9 AQCD (U.S. EPA, 2000, 000907). Relevant new data are included when available. Recent models of
10 Hb binding are characterized, as well as measurements of tissue CO concentrations using new
11 methods of extraction.
12 CO binds with a number of heme-containing molecules including Mb and cytochromes, but
13 none have been studied as extensively as Hb. The primary focus of this chapter is placed on the
14 models and kinetics of such binding and the factors influencing this event. The chapter discusses
15 effects at ambient or near ambient levels of CO leading to low COHb levels (< 5%); however few
16 studies are available at ambient CO concentrations. Both human and animal studies using higher CO
17 exposure concentrations, resulting in moderate to high COHb levels (< 20%), are discussed where
18 needed to understand CO kinetics, pathophysiologic processes, and mechanisms of cytotoxicity.
19 Where human studies could not experimentally test certain hypotheses or were unavailable, animal
20 experiments were used as surrogates. CO uptake and elimination has been shown to be inversely
21 proportional to body mass over environmentally relevant exposure levels, meaning the smaller the
22 animal, the faster the rate of absorption and elimination (Klimisch et al, 1975, 010762; Tyuma et al.,
23 1981, 011226). However, the basic mechanisms of CO toxicity between experimental animals and
24 humans are similar and are thus extrapolated from animals to humans in this chapter, keeping in
25 mind a number of interspecies differences.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 4-1 DRAFT - DO NOT CITE OR QUOTE
-------
4.2. Carboxyhemoglobin Modeling
4.2.1.The Coburn-Forster-Kane and Other Models
1 Investigators have modeled the effect of CO binding to Hb in a number of ways. Empirical
2 and mechanistic models are two distinct approaches that have been taken to model in vivo COHb
3 formation after CO exposure. First, empirical models were used to predict COHb by regressing
4 concentration and duration of exogenous CO exposure with observed COHb, with or without the
•
5 inclusion of physiological predictors such as initial COHb levels and alveolar ventilation (\k ).
6 These methods were reviewed in depth in the previous AQCD (U.S. EPA, 2000, 000907). It is
7 important to note that CO empirical regression models are limited to estimating COHb in the exact
8 conditions on which the models were based. These simple models include those by Peterson and
9 Stewart (1970, 012416) and Ott and Mage (1978, 011124). as well as various others (Chung, 1988,
10 012749: Forbes et al., 1945, 012850: Selvakumar et al., 1992, 013750: Sharan et al., 1990, 003798:
11 Singh et al., 1991, 013583). Using a linear differential equation where ambient CO concentrations
12 varied, it was shown that the presence of brief ambient CO concentration spikes averaged over
13 hourly intervals may lead to underestimating the COHb concentration by as much as 21% of the true
14 value. To avoid this problem, it was suggested that ambient CO measurements be monitored and
15 averaged over 10-15 min periods (Ott and Mage, 1978, 011124). Other empirical models predict
16 COHb as a function of exposure time (Sharan et al., 1990, 003798: Singh et al., 1991, 013583) or
17 altitude (Selvakumar et al., 1992, 013750). A comparison of empirical model predictions showed a
18 wide disparity in predicted COHb values, highlighting the inaccuracy of these models outside of the
19 conditions on which they were presented (Tikuisis, 1996, 080960).
20 Secondly, mechanistic models use physical and physiological processes and an understanding
21 of biological processes to predict COHb production. The most commonly used mechanistic method
22 for predicting levels of blood COHb after CO inhalation is the Coburn-Forster-Kane equation or
23 CFK model developed in 1965 (Coburn et al., 1965, 011145). This differential equation was
24 developed to examine endogenous CO production, using the major physiological and physical
25 variables influencing this value. Since then, it has been shown to provide a good approximation to
26 the COHb level at a steady level of inhaled exogenous CO (Peterson and Stewart, 1975, 010696:
27 Stewart et al., 1973, 012428). The CFK model describes a four-element, physical system containing
28 an exogenous CO source, a transfer interface, an endogenous CO source, and a storage compartment.
29 The linear CFK model assumes O2Hb concentration is constant and is as follows in Equation 4-1:
September 2009 4-2 DRAFT - DO NOT CITE OR QUOTE
-------
d[COHb]t
dt
= Vco -
[COHb]0P-02
[02Hb]M
1
1 P -P
1 i 1 B 1 H2O
DLCO '
+
P,CO
1 P -P
1 i 1 B 1 H2O
DLCO V,
Equation 4-1
1 where V\, is blood volume in milliliters (mL); [COHb]t is the COHb concentration at time t in mL
•
2 CO/mL blood, at standard temperature and pressure, dry (STPD); Vco is the endogenous CO
3 production rate in mL/min, STPD; [COHb]0 is the COHb concentration at time zero in mL CO/mL
4 blood, STPD; [O2Hb] is the O2Hb concentration in mL O2/mL blood, STPD; M is the Haldane
5 coefficient representing the CO chemical affinity for Hb; 5 2 js me average partial pressure of O2 in
•
6 lung capillaries in mmHg; VA is the alveolar ventilation in mL/min, STPD; DLCO is the lung
7 diffusing capacity of CO in mL/min/mmHg, STPD; PB is the barometric pressure in mmHg; PH2o is
8 the saturation pressure of water vapor at body temperature in mmHg (47 mmHg); and P:CO is the
9 CO partial pressure in inhaled air in mmHg.
10 The linear CFK model assumes instant equilibration of COHb concentration between venous
11 and arterial blood, gases in the lung, and COHb concentrations between blood and extravascular
12 tissues, which is not physiologically representative. The nonlinear CFK equation incorporates the
13 interdependence of COHb and O2Hb levels since they are derived from the same pool of blood Hb.
14 The nonlinear equation is more physiologically accurate; however the linear CFK equation gives a
15 good approximation to the nonlinear solution over a large range of values during CO uptake and
16 during low levels of CO elimination (Smith, 1990, 013164). The linear equation prediction of COHb
17 concentration at or below 6% will only differ ±0.5% from the nonlinear equation prediction.
18 Sensitivity analysis of the CFK equations has shown that alterations in each variable of the equation
19 will affect the outcome variably at different times of exposure, so that the relative importance of the
20 CFK variables will change with the experimental conditions (McCartney, 1990, 013162). Figure 4-1
21 illustrates the temporal changes in fractional sensitivities of the principal physiological determinants
22 of CO uptake for the linear form of the CFK equation, where THb is the total blood concentration of
23 Hb in g Hb/mL blood and F:CO is the fractional concentration of CO in ambient air in ppm. The
24 fractional sensitivity of unity means that, for example, a 5% error in the selected variable induces a
25 5% error in the predicted COHb value by the nonlinear model. As Figure 4-1 demonstrates, a
26 constant or given percent error in one variable of the model does not generally produce the same
27 error in the calculated blood COHb, and the error is time dependent. Thus, each variable influencing
28 CO uptake and elimination will exert its maximal influence at different times of exposure. This
29 analysis found that only F:CO and VCo will not affect the rate at which equilibrium is reached.
September 2009
4-3
DRAFT - DO NOT CITE OR QUOTE
-------
1.00
1.00
- -0.75
-1.00
0.6s
-2
6s
-1
i min 10 min
0 1
Log Time
1.7h
2
17h
3
-1.00
Source: modified fromMcCartney (1990, 013162'
Figure 4-1. Plot of fractional sensitivities of selected variables versus time of exposure.
1 The mechanistic CFK model contains a number of assumptions under which the model is
2 solely applicable, including 1) ventilation is a continuous process, 2) equilibrium between plasma
3 CO concentration and COHb concentration is obtained in the pulmonary system, 3) percent COHb
4 can exceed 100% saturation in the linear model, and 4) it does not account for the shape of the O2 or
5 CO saturation versus pO2 or pCO relation (McCartney, 1990, 013162). Estimations outside of these
6 assumptions have been attempted but with less predictive agreement. For example, transient
7 exposures such as those that would simulate everyday conditions would violate the assumption of a
8 single, well-mixed vascular compartment. COHb levels during exposure of subjects exposed to
9 frequent but brief high CO exposures (667-7,500 ppm for 75 s to 5 min) were not accurately
10 predicted by CFK modeling (Benignus et al., 1994, 013908: Tikuisis et al., 1987, 012219: Tikuisis et
11 al., 1987, 012138). Consistently, the predicted COHb value overpredicted venous COHb (0.8-6%)
12 and underpredicted arterial COHb (1.5-6.1%) and this disparity increased after exercise. Individual
13 differences between arterial and venous COHb varied from 2.3-12.1% COHb (Benignus et al., 1994,
14 013908). These inaccuracies between measured and predicted COHb values disappeared after
September 2009
4-4
DRAFT - DO NOT CITE OR QUOTE
-------
1 simulated mixing of arterial and venous blood and thus are likely due to delays in mixing of arterial
2 and venous blood and differences in cardiac output and lung wash-in. A modified CFK was created
3 to adjust for these issues and produce a more accurate COHb prediction (Smith et al, 1994, 076564).
4 This expanded CFK model used multiple compartments to model the lung, arm circulation, and the
5 rest of the body (quickly and slowly perfused tissues). This model was more accurate than the
6 nonlinear CFK in predicting the individual peak or maximal values of arterial and venous COHb
7 during CO uptake in the first 10 min after exposure. However, both the nonlinear CFK and this
8 expansion produced accurate predictions several minutes after the 5 min exposure ended. The
9 expanded model required the use of two parameters that were not measured individually or derived
10 from the literature, and instead were estimated by adjustments between the simulations and
11 experimental subject data.
12 In addition to the limitations discussed above, the CFK model does not account for
13 extravascular storage sites for CO, such as muscle Mb. CO will undergo reversible muscle Mb
14 binding, similar to Hb, as well as uptake into other extravascular tissues (Vreman et al., 2006,
15 098272). The most recent adaptation to the CFK equation incorporates alveoli-blood and blood-
16 tissue CO exchanges and mass conservation of CO at all times (Gosselin et al., 2009, 190946). This
17 model has a single free parameter whose value is estimated from one data set, however it better
18 predicted COHb formation over a wide range of CO levels and several temporal scenarios (Stewart
19 et al., 1970, 013972: Tikuisis et al., 1987, 012138: Tikuisis et al., 1987, 012219: Tikuisis et al., 1992,
20 013592) compared to the linear CFK model. Like the linear CFK model, this modified model
21 assumes a constant level of oxyhemoglobin. Sensitivity analysis of the model showed that the most
P-O
22 important parameter influencing the level of COHb in this model is M, followed by c 2 and VA .
23 Ambient exposure scenarios were simulated with this model to determine the CO concentrations
24 needed to reach certain COHb levels in humans from 3 months of age to 40 year old adults. The CO
25 concentrations needed to achieve 2% COHb vary from 24.4-48.1 ppm for a 1 h exposure, from
26 11.1-13.1 ppm for an 8 h exposure, and from 9.8-10.1 ppm for a daily exposure. Children (1 yr old)
27 were most sensitive to CO concentrations, whereas babies (3 months old) required the highest CO
28 concentration to reach 2% COHb. The model was also used to simulate time profiles of COHb
29 formation for two workweek exposure scenarios in a healthy 40-year-old man. Figure 4-2A
30 represents a high exposure scenario where the work period is spent at 35 ppm and the rest of the time
31 at 3 ppm. Figure 4-2B represents a lower exposure scenario where there is a constant 3 ppm
32 exposure. Both figures consist of 5 days where 24 h are broken up into three consecutive 8-h
33 periods: sleeping from 12 a.m. to 8 a.m., working with light exercise from 8 a.m. to 4 p.m., and
34 sitting from 4 p.m. to 12 a.m..
September 2009 4-5 DRAFT - DO NOT CITE OR QUOTE
-------
A 7
B 1
0.8
y
I
'0.6
I 0.4
0.2
Light
exercise
Sleeping | Sitting
'h4***
Monday Tuesday | Wednesday [ Thursday Friday
Time (days)
Light
exercise
Sleeping I Sitting
Monday Tuesday | Wednesday | Thursday
Time (days)
Friday
48
40
32
24
16
O
0)
3-
o
3
o
o
A -i S>
42 3.
3.6 £
9
8
§
"I
I
T>
"O
1.2 3.
Source: Gosselin et al. (2009,190946)
Figure 4-2 Simulated COHb formation for two 5 day workweeks "The 24-h day consists of
three consecutive 8-h periods: sleeping from 12 a.m. to 8 a.m., working (light
exercise) from 8 a.m. to 4 p.m., and sitting from 4 p.m. to 12 a.m.. (A) High
exposure: work period at 35 ppm and the rest of the time at 3 ppm. (B) Low daily
exposure at 3 ppm. The CO exposure periods are represented by dotted lines
(—) and the COHb simulations by solid lines (—)."
4.2.2.Multicompartment Models
1 A third approach applied more recently to model COHb formation is the use of
2 multicompartment or physiologically based pharmacokinetic (PBPK) models. Cronenberger et al.
3 (2007, 194671) described a two-compartment population-based model to describe and predict COHb
September 2009
4-6
DRAFT - DO NOT CITE OR QUOTE
-------
1 pharmacokinetics from smoking. This model required a compartment for extravascular binding of
2 CO to accurately predict COHb formation during multiple short and rapid inhalations followed by a
3 period of no exposure, as occurs in smoking.
4 A five compartment PBPK model has been proposed to predict CO uptake and distribution
5 from acute inhalation exposure and contains components for lung, arterial blood, venous blood,
6 muscle tissue, and nonmuscle tissue (Bruce and Bruce, 2003, 193975; Bruce and Bruce, 2006,
7 193980; Bruce et al., 2008, 193977). This model structure is illustrated in Figure 4-3. This model
8 includes the dynamics of CO storage in the lung and its dependence on ventilation and CO pressure
9 of mixed venous blood, relaxes the assumption that Hb is saturated by including the role of CO in
10 altering the O2 dissociation curve, includes a subcompartmentalized muscle tissue compartment,
11 accounts for dissolved CO in blood and tissue, and predicts COHb based on age and body
12 dimensions. This multicompartment model is limited by its exclusion of cellular metabolism or Mb
13 diffusion, simplification of within tissue bed spatial variability, and assumption that ventilation and
14 PAO2 are constant. Another limitation of this model is that some of the physiological parameters used
15 in simulations are estimated through visual fits to the COHb profile and not from experimental or
16 published data. This model better predicts COHb levels when inspired CO levels change rapidly or
17 when incomplete blood mixing has occurred, and better predicts the CO washout time course
18 compared to the CFK equation. Bruce and Bruce (2003, 193975) compared the two models and
19 found similar results for long term exposure settings (1,000 min), however, the multicompartment
20 model predicted somewhat lower COHb levels compared to the CFK model during transient CO
21 uptake conditions when using data taken from Peterson and Stewart (1970, 012416).
September 2009 4-7 DRAFT - DO NOT CITE OR QUOTE
-------
LUNGS
PACO
1
2
3
4
5
6
7
10
11
12
1
tXTlf I
MIXED
ARTERIAL
NONMUSCLE
TISSUE
v
1 "-"
V,:.,
i
Qot
MUSCLE TISSUE
i V
Mb,O, Vjrn. M
Mb;.O, Vtnv M
b^O
b2CO
^
Qm
Source: Modified from Bruce and Bruce (2008,193977)
Figure 4-3 Overall structure of the Bruce and Bruce (2008,193977) multicompartment model
of storage and transport of CO. Includes compartments for lung, arterial blood,
venous blood, muscle tissue, and nonmuscle tissue. The muscle compartment is
divided into two subcompartments for diffusion of gases within the tissue.
A multicompartment model of the human respiratory system was developed using
characteristics of the lung representation described in Selvakumar et al. (1992, 013750) and Sharan
(1999, 194673). which considered the exchanges of CO, O2, and CO2, and the tissue representation
of Bruce and Bruce (2003, 193975) and Neto et al. (2008, 194672). The model contains six
compartments including: alveolar, pulmonary capillaries, arterial, venous, tissue capillary, and
tissues (muscular and non-muscular). The model was applied to four simulated physical activity
levels, resting, sitting, standing, and walking, in a healthy subject exposed to the urban atmosphere
of a metropolitan area of Brazil. The highest and lowest COHb levels were simulated in the walking
individual, suggesting that greater variability in COHb occurs at higher physical activity levels.
4.2.3.Model Comparison
A number of models have been presented which predict COHb formation over numerous
exposure scenarios. These models are often compared to determine the most accurate predictive
model under certain exposure conditions. As was mentioned in Section 4.2.1, Tikuisis (1996,
September 2009
4-8
DRAFT - DO NOT CITE OR QUOTE
-------
1 080960) conducted a comparison of empirical model predictions that showed a wide disparity in
2 predicted COHb values, highlighting the inaccuracy of these models outside of the conditions on
3 which they were presented. Smith et al. (1990, 013164) compared the linear and nonlinear CFK
4 equations and concluded that the linear CFK equation gives a good approximation (within 1%) to the
5 nonlinear solution over a large range of values during CO uptake and over a somewhat smaller range
6 during CO elimination. The linear equation prediction of COHb concentration at or below 6% will
7 only differ ± 0.5% from the nonlinear equation prediction. Additionally, the most recently modified
8 CFK model (Gosselin et al., 2009, 190946) better predicted COHb formation over a wide range of
9 CO levels (50-4,000 ppm) and several temporal scenarios (Stewart et al., 1970, 013972; Tikuisis et
10 al., 1987, 012138: Tikuisis et al., 1987, 012219: Tikuisis et al., 1992, 013592) compared to the linear
11 CFK model. Linear regression slopes between the simulated COHb values and the observed
12 experimental values were closer to 1 in all experimental scenarios, indicating a better fit to the
13 observed data. When evaluating all validation studies the modified model had an estimated slope of
14 0.996 (95% CI: 0.986-1.001) compared to 0.917 (95% CI: 0.906-0.927) using the CFK model. Bruce
15 and Bruce (2003, 193975) compared their model to the CFK and found similar results for long term
16 exposure settings (1,000 min [16.5 h]), however, their multicompartment model predicted somewhat
17 lower COHb levels over transient CO uptake conditions when using data taken from Peterson and
18 Stewart (1970, 012416). The Bruce and Bruce model better predicts COHb levels when inspired CO
19 levels change rapidly or when incomplete blood mixing has occurred, and better predicts the CO
20 washout time course compared to the CFK equation.
4.2.4.Mathematical Model Usage
21 Since measurements of COHb in the population are not readily available, mathematical
22 models are used to predict the resulting COHb levels from various CO exposure scenarios. Table 4-1
23 illustrates the predictions of venous COHb after 1, 8, or 24 h of CO exposure at a range of
•
24 concentrations in a healthy adult human at rest (\A =6 L/min; DLCO = 20 (mL/min)/mmHg),
•
25 during light exercise (\A =15 L/min; DLCO = 34 [mL/min]/mmHg), and during moderate exercise
•
26 ( VA = 22 L/min; DLCO = 43 [mL/min]/mmHg). The contribution of alveolar ventilation and lung
27 diffusion to the changes in COHb levels is discussed in Section 4.3.1.2. The Quantitative Circulatory
28 Physiology (QCP) model, which integrates human physiology using over 4,000 variables and
29 equations based on published biological interactions, was used to predict these values (Abram et al.,
30 2007, 193859; Benignus et al., 2006, 151344). This dynamic whole body model uses the nonlinear
31 CFK equation with modifications presented in Smith et al. (1994, 076564). Endogenous CO
32 production varies as described in Section 4.5 but generally results in less than 1% COHb, with a
33 QCP modeled value of 0.39% at time zero. The rate of endogenous CO production was set at
34 0.007 mL/min for this simulation, whereas both higher and lower values have been reported (Coburn
September 2009 4-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 et al, 1966, 010984) (see Section 4.5). Table 4-1 illustrates that 35 ppm CO for 1-h results in
2 between 0.9-2.0% COHb and 9 ppm CO for 8 h results in between 1.2-1.3% COHb, depending upon
3 activity level. Also, this table shows that low concentration CO exposure over several hours can
4 result in equivalent COHb levels compared to higher concentration, acute exposure. For example, in
5 a resting condition without additional baseline COHb, COHb resulting from 35 ppm for 1 h (0.9%)
6 is approximately equivalent to 6 ppm for 8 h (0.9%) or 4 ppm for 24 h (0.9%).
Table 4-1 Predicted COHb levels resulting from 1, 8, and 24 h CO exposures in a modeled human
at rest (VA = 6 L/min; DLCO = 20 (mL/min)/mmHg; VCo = 0.007 mL/min;
initial COHb = 0.38%; Hb = 0.15 g/mL), during light exercise (\, =15 L/min;
DLCO = 34 (mL/min)/mmHg), and during moderate exercise (\ = 22 L/min; DLCO =
43 (mL/min)/mmHg). The QCP model used a dynamic nonlinear CFK with affinity
constant M = 230.
CO (ppm)
2
3
6
9
15
24
35
6 L/min
0.30
0.32
0.38
0.43
0.54
0.71
0.91
1h
15 L/min
0.30
0.34
0.45
0.56
0.78
1.12
1.52
22 L/min
0.30
0.35
0.50
0.65
0.95
1.40
1.95
6 L/min
0.46
0.56
0.86
1.15
1.74
2.61
3.67
8h
1 5 L/min
0.40
0.54
0.94
1.34
2.13
3.32
4.74
22 L/min
0.37
0.51
0.92
1.34
2.16
3.37
4.84
6 L/min
0.57
0.71
1.15
1.59
2.46
3.74
5.26
24 h
15 L/min
0.42
0.57
1.02
1.47
2.34
3.63
5.18
22 L/min
0.37
0.52
0.95
1.38
2.24
3.49
4.98
7 The QCP model was also used to simulate several population exposure scenarios including
8 various commuting concentrations (Figure 4-4), endogenous production rates (Figure 4-5), and
9 activity levels (Figure 4-6). Commuting concentrations were modeled since the highest ambient CO
10 exposure levels are generally observed during transit (Section 3.6.6.2). Figure 4-4 presents simulated
11 COHb levels in a healthy adult throughout the second of five modeled days containing a 60 min
12 commute at various CO concentrations. The U.S. Census Bureau estimates that 5% of the population
13 commutes in automobiles for 60 or more minutes to work daily (U.S. Census Bureau, 2008, 194013)
14 and exposure studies have reported in-vehicle transit concentrations up to 50 ppm (Abi-Esber and
15 El-Fadel, 2008, 190939: Duci et al., 2003, 044199). However, U.S. studies have reported in-vehicle
16 concentrations of less than 5 ppm (Riediker et al., 2003, 043761). CO concentrations during
17 commuting lead to spikes in COHb in this model scenario with a 1% COHb increase over the initial
18 COHb (0.4%) after 50 ppm exposure. Figure 4-4 also illustrates that the COHb saturation after CO
19 exposure from commuting is not fully eliminated by the next commuting period. Modeling
20 successive days results in the same pattern and degree of COHb formation, indicating no
21 accumulation of COHb over time.
September 2009 4-10 DRAFT - DO NOT CITE OR QUOTE
-------
1.6
1.4
o
u
0.2
2 ppm 10 ppm 20 ppm 50 ppm
360
720
Time (min)
1080
1440
1
2
3
4
5
6
Figure 4-4 Predicted COHb levels in healthy commuters exposed to various CO
concentrations over a 60-min commute twice a day. Ambient CO concentration
not during commuting time was 1 ppm. The activity pattern simulated 1) sleeping
for 8 h, 2) standing and light exercise for 30 min, 3) sitting during a 60-min
commute, 4) light exercise for 8.5 h, 5) sitting during a second 60-min commute,
6) moderate exercise for 60 min, 7) sitting for 4 h. The graph illustrates the
second day simulated under these conditions.1
Figure 4-5 presents simulated COHb levels in adults with various endogenous CO production
rates throughout the second of five modeled days containing a 60-min commute at 20 ppm CO. The
normal endogenous rate of CO production in young adult males with an average COHb of 0.88%
averages 0.007 mL/min (18.7 ± 0.8 (imol/h) (Coburn et al., 1963, 013971). However, a number of
diseases and conditions described in Section 4.5 can affect this production rate. Patients with
hemolytic anemia have endogenous CO production rates ranging from 0.012 to 0.053 mL/min (31 to
-1 Sleeping/lying human parameters: VA- 3.8 L/min, VT- 467 mL, VD- 147 mL, Vco- 0.007 mL/min, DLCO- 17.9 mL/min/mmHg, M- 230,
initial COHb-0.38%.
Sitting human parameters: VA- 5.2 L/min, VT- 560 mL, VD- 155 mL, VCo- 0.007 mL/min, DLCO- 18 mL/min/mmHg.
Standing human parameters: VA- 6.4 L/min, VT- 636 mL, VD- 161 mL, VCo- 0.007 mL/min, DLCO- 19.3 mL/min/mmHg.
Light exercise (1 MPH, 32 W) human parameters: VA- 13.4 L/min, VT- 994 mL, VD- 218 mL, Vco- 0.007 mL/min, DLCO- 30.4
mL/min/mmHg.
Heavy exercise (3 MPH, 96 W) human parameters: VA- 31.4 L/min, VT- 1642 mL, VD- 241 mL, Vco- 0.007 mL/min, DLCO- 49.6
mL/min/mmHg.
September 2009
4-11
DRAFT - DO NOT CITE OR QUOTE
-------
1 143 (imol/h) (Coburn et al, 1966, 010984). The venous COHb levels in these same patients ranged
2 from 0.77 to 2.62%.
— 0.007 mL/min 0.02 mL/min 0.04 mL/min 0.06 mL/min
0
360
720
Time (min)
1080
1440
Figure 4-5 Predicted COHb levels due to various endogenous CO production rates. The
activity pattern presented in Figure 4-4 was used. Ambient CO concentration not
during commuting time was 1 ppm and commuting CO concentration was
20 ppm. The graph illustrates the second day simulated under these conditions.
3 Figure 4-6 presents simulated COHb levels throughout the second of five modeled days in a
4 healthy adult performing two activity patterns at a constant 1 ppm CO exposure. The sedentary
5 individual maintains a higher COHb saturation compared to the active individual due to increased
6 gas exchange during physical exertion.
September 2009
4-12
DRAFT - DO NOT CITE OR QUOTE
-------
0.8
— 0.6
I
o
o
0.4 -
0.2
—Active Sedentary
360
720
Time (min)
1080
1440
Figure 4-6 Predicted COHb levels in an active or sedentary individual. CO concentration was
constant at 1 ppm. The activity pattern presented in Figure 4-4 was used for the
active individual. The 24 h period of the sedentary individual included 1) sleeping
for 8 h, 2) sitting for 4 h, 3) standing for 1 h, 4) sitting for 4 h, 5) lying down for
7 h. The graph illustrates the second day simulated under these conditions.
4.3. Absorption, Distribution, and Elimination
4.3.1 Pulmonary Absorption
1 Pulmonary uptake of CO accounts for all environmental CO absorption and occurs at the
2 respiratory bronchioles and alveolar ducts and sacs. CO and O2 share various physico-chemical
3 properties, thus allowing for the extension of the knowledge about O2 kinetics to those of CO despite
4 the differences in the reactivity of the gases. The exchange of CO between the air and the body
5 depends on a number of physical (e.g., mass transfer and diffusion), as well as physiological factors
6 (e.g., alveolar ventilation and cardiac output), which are controlled by environmental conditions,
7 physical exertion, and other processes discussed in Section 4.4. The ability of the lung to take up
September 2009
4-13
DRAFT - DO NOT CITE OR QUOTE
-------
1 inhaled CO is measured by DLCO, and CO uptake (VCo) representing the product of DLCO and the
2 mean alveolar pressure (PACO). The importance of dead space volume, gas mixing and
3 homogeneity, and ventilation/perfusion matching were discussed in depth in the 2000 CO AQCD
4 (U.S. EPA, 2000, 000907).
4.3.1.1. Mass Transfer of Carbon Monoxide
5 Mass transfer refers to the molecular and convective transport of CO molecules within the
6 body stores, driven by random molecular motion from high to low concentrations. CO enters through
7 the airway opening (mouth and nose) and transfers in a gas phase to the alveoli. CO transport is due
8 to convective flow, the mechanical action of the respiratory system, and diffusion in the acinar zone
9 of the lung (Engel et al., 1973, 014336). Then, CO diffuses across the air-blood interface, binding
10 red blood cell (RBC) Hb. At environmental CO levels, CO uptake into RBC is limited by the
11 reaction rate of binding of CO to O2Hb forming COHb. Pulmonary capillary RBC CO diffusion is
12 rapidly achieved (Chakraborty et al., 2004, 193759; Gibson and Roughton, 1955, 193941; Reeves
13 and Park, 1992, 193847; Roughton and Forster, 1957, 193862). The formation rate and level of
14 COHb depends upon pCO, pO2 in the air, time of exposure, and the ventilation rate (Roughton and
15 Forster, 1957, 193862). Most of the body CO is bound to Hb; however, 10-15% of the total body CO
16 is located in extravascular tissues primarily bound to other heme proteins (Coburn, 1970, 013916).
17 Considerable concentrations of CO have been measured in spleen, lung, kidney, liver, muscle, and
18 heart (Vreman et al., 2005, 193786; Vreman et al., 2006, 098272). whereas less CO is localized to
19 fatty tissues, such as adipose and brain. The transfer of CO occurs by a partitioning of CO between
20 Hb and tissue. Less than 1% of the total body CO stores appear as dissolved in body fluids, due to
21 the insolubility and small tissue partial pressure of CO (Coburn, 1970, 013916). Transport pathways
22 and body stores of CO are shown in Figure 4-7.
September 2009 4-14 DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Monoxide in the Ambient Air
Endogenous ;
production
of CO
i Metabolism
j ofCOtoCO! !
Myoglobin
Intracellular
enzymes
'•-. _.....'
Extravascular compartment
Figure 4-7
Source: Adapted from Co-burn (1967, 0111441
Found in U.S. EPA (2000, 000907)
Diagrammatic presentation of CO uptake and elimination pathways and CO body
stores.
4.3.1.2. Lung Diffusion of Carbon Monoxide
1 Lung diffusion of CO is an entirely passive process of gas diffusion across the alveolo-
2 capillary membrane, through the plasma, across the RBC membrane and into the RBC stroma, where
3 CO binding to Hb rapidly occurs. Membrane and blood phase transfer are governed by physico-
4 chemical laws, including Pick's first law of diffusion. The diffusing capacity of the lung for CO,
5 represented as DLCO, is a measurement of the partial pressure difference between inspired and
6 expired CO. Due to the rapid binding of CO to Hb, a high pressure differential between air and blood
7 exists when CO air levels are increased. Inhalation of CO-free air reverses the pressure differential
8 (higher CO pressure on the blood side than the alveolar side), and then CO is released into the
9 alveolar air. Since CO is also produced endogenously, CO release will also be affected by this
10 production pressure. However, the air-blood gradient for CO is usually higher than the blood-air
11 gradient; therefore, CO uptake will be a proportionately faster process than CO elimination.
12 A number of factors have been found to affect DLCO including Hb concentration, cardiac
13 output (Q), erythrocyte flow, COHb concentration, PACO2, body position, exercise, time of day, age,
14 etc. (Forster, 1966, 180430; Hsia, 2002, 193857). DLCO consistently decreases after intense bouts of
15 exercise, likely due to the redistribution of blood volume to the periphery (Hanel et al., 1997,
16 193918; Manier et al., 1991, 193979). However, in going from rest to exercise DLCO can increase
September 2009
4-15
DRAFT - DO NOT CITE OR QUOTE
-------
1 linearly from: lung expansion leading to unfolding and distension of alveolar septa, opening and/or
2 distension of capillaries as Q increases, increased capillary hematocrit, and more homogeneous
3 distribution of capillary erythrocytes (Hsia, 2002, 193857). DLCO is less dependent upon lung
4 volume at mid-range vital capacity, but at extreme volumes the diffusion rate is varied, higher than
5 average at total lung capacity and lower at residual volume (McClean et al, 1981, 012411).
6 DLCO is also altered by a number of diseases. Decreased DLCO is evident in patients with
7 restrictive lung disease (i.e., decreased lung volumes) since a loss of lung tissue leads to a loss of
8 functional lung units. DLCO also shows a good correlation with the severity of restrictive lung
9 disease (Arora et al., 2001, 186713). Conditions affecting DLCO vary and include chronic
10 obstructive pulmonary disease (Terzano et al., 2009, 108046). ulcerative colitis (Marvisi et al., 2000,
11 186703: Marvisi et al., 2007, 186702). severe gastroesophageal reflux (Schachter L. et al., 2003,
12 186707). beta thalassemia (Arora et al., 2001, 186713). thoracic or abdominal aortic aneurysm
13 (Sakamaki et al., 2002, 186706). pulmonary arterial hypertension (Proudman et al., 2007, 186705).
14 and chemotherapy for breast cancer (Yerushalmi et al., 2009, 186711). Diseases affecting CO
15 kinetics and DLCO are also discussed in section 4.4.4.
4.3.2.Tissue Uptake
4.3.2.1. The Respiratory Tract
16 The upper respiratory tract contributes little to the overall COHb uptake. The lung has nearly
17 constant exposure to CO; however, relatively little CO diffuses into the tissue except at the alveolar
18 region en route to the circulation. No detectable uptake of CO was observed in the human nasal
19 cavity or upper airway (Guyatt et al., 1981, 011196) or in the monkey oronasal cavity after high CO
20 exposure (Schoenfisch et al., 1980, 011404).
4.3.2.2. The Blood
21 The blood is the largest reservoir for CO, where it reversibly binds to Hb. The chemical
22 affinity of CO for adult human Hb is approximately 218 times greater than that of O2, meaning one
23 part CO and 218 (210-250) parts O2 would form equal parts of O2Hb and COHb (Engel et al., 1969,
24 193914: Rodkey et al., 1969, 008151: Roughton, 1970, 013931). This would happen when breathing
25 air containing 21% O2 and 960 ppm CO. This concept was presented by Haldane and Smith
26 (Haldane, 1895, 010538) and later represented as the Haldane constant M (210-250) in the Haldane
27 equation by Douglas, Haldane, and Haldane (Douglas et al., 1912, 013965). M is relatively
28 unaffected by changes in physiological pH, CO2, temperature, or 2,3-diphosphoglycerate:
September 2009 4-16 DRAFT - DO NOT CITE OR QUOTE
-------
COHb -T- O2Hb =Mx (pCO + pO2)
Equation 4-2
1 The Hb association rate for CO is 10% slower than O2 and occurs in a cooperative manner
2 (Chakraborty et al., 2004, 193759; Sharma VS, Schmidt and Ranney, 1976, 193766). Hb is
3 composed of four globin chains each containing a heme group capable of binding CO or O2. The
4 associative reaction rates become faster with successive heme binding, attributed to interactions
5 within the protein and to strains imposed on the heme and its ligands (Alcantara et al., 2007,
6 193867). More simply, the greater the number of heme sites bound to CO, the greater the affinity of
7 free heme sites for O2, thus causing Hb to bind and retain O2 that would normally be released to
8 tissues. Cooperativity is greatly reduced in CO dissociation, but the rate of dissociation of CO from
9 Hb is orders of magnitude slower than O2 (kco = 4 x 10-4 k02), which accounts for the high affinity
10 values (Chakraborty et al., 2004, 193759). The half-time of dissociation reaction is about 11 s at
11 37°C (Holland, 1970, 193856). In general, CO uptake to COHb equilibrium is slower in humans and
12 large animals, requiring 8-24 h, than in smaller species such as rats, which will equilibrate in 1-2 h
13 (Penney, 1988, 012519). Also, COHb equilibrium within the blood stream is not instantaneous. Men
14 exposed to brief (~5 min) high dose CO had an initial delay of 1-2 min in the appearance of venous
15 COHb after the start of CO inhalation (Benignus et al., 1994, 013908; Smith et al., 1994, 076564).
16 Additionally, arterial COHb concentrations were considerably higher than venous concentrations
17 during CO exposure; however, they quickly converged after the end of exposure, as venous and
18 arterial blood mixed.
19 CO binding to Hb also has effects on the O2 dissociation curve of the remaining Hb by shifting
20 the curve progressively to the left and altering the normal S-shaped curve to become more
21 hyperbolic due to increased cooperative O2 binding (Roughton, 1970, 013931). This is referred to as
22 the "Haldane effect" and causes tissues to have more trouble obtaining O2 from the blood, even
23 compared to the same extent of reduced Hb resulting from anemia. For example, Figure 4-8 (as
24 explained in the 2000 CO AQCD) illustrates that in an acute anemia patient (50% of Hb) at a venous
25 pO2 of 26 mmHg (v'l), 5 vol % of O2 (50% saturation) was extracted from the blood. In contrast, for
26 a CO poisoned person with 50% COHb, the venous pO2 will have to drop to 16 mmHg (v'2) to
27 release the same 5 vol % O2. This more severe effect on O2 pressure may lead to brain O2 depletion
28 and loss of consciousness if any higher demand of O2 is needed (e.g., exercise).
September 2009 4-17 DRAFT - DO NOT CITE OR QUOTE
-------
20
100
O
O
.Q
o
O
0
6
•4-t
0
o
Q.
0)
E
15 -
10 -
50% Anemia
V' (02Hb Capacity = lOmL/IOOmL)
20
40 60
PO2 (mmHg)
80
100
Source: U.S. EPA (1991, 017643)
Figure 4-8 02Hb dissociation curve of normal human blood, of blood containing 50% COHb,
and of blood with only 50% Hb because of anemia.
4.3.2.3. Heart and Skeletal Muscle
1 Mb is a globular heme protein that facilitates O2 diffusion from the muscle sarcoplasm to
2 mitochondria, acting as an O2 supply buffer to maintain adequate pO2 for mitochondria when the O2
3 supply changes, as in exercise. O2 has a greater affinity for Mb than Hb, which allows small changes
4 in tissue pO2 to release large amounts of O2 from O2Mb (Wittenberg et al., 1975, 012436). Small
5 reductions in O2 storage capacity of Mb, due to CO binding, may have a profound effect on the
6 supply of O2 to the tissue.
September 2009
4-18
DRAFT - DO NOT CITE OR QUOTE
-------
1 Like Hb, Mb will undergo reversible CO binding, however the affinity constant is
2 approximately eight-times lower than Hb (M = 20-40 versus 218, respectively) (Haab, 1990,
3 013359). The association rate constant of CO and Mb is approximately 27 times lower than O2,
4 however the dissociation rate constant is approximately 630 times lower than O2 (Gibson et al.,
5 1986, 016289) causing CO to be retained and possibly stored in the muscle. CO levels have been
6 measured in human muscle and heart tissues with less than 2% COHb concentrations at background
7 levels (15 and 31 picomole (pmol) CO/mg ww, respectively) (Vreman et al., 2006, 098272) (Table
8 4-2). Under conditions of CO asphyxiation, tissue concentrations increased 17-18 fold (265 and 527
9 pmol CO/mg ww muscle and heart tissue, respectively); however, heart tissue concentrations varied
10 widely between individuals. Mouse muscle did not show this increase after exogenous CO exposure
11 (Table 4-3). This may be due to the fact that human muscle has a 15-fold higher concentration of
12 myoglobin protein than mouse muscle (Weller et al., 1986, 187298). The capacity for diffusion of
13 CO into the muscle is represented by the coefficient DmCO and is generally larger in males than in
14 females, likely due to the differences in muscle mass and capillary density (Bruce and Bruce, 2003,
15 193975). COMb concentrations in the heart and skeletal muscle increase with work load, due to a
16 higher relative rate of CO binding to Mb relative to Hb. This causes an increase in COMb/COHb
17 that is not seen at rest (Sokal et al., 1984, 011591). Subjects with 2% COHb, but not those with 20%
18 COHb levels, showed a significant uptake of CO from the blood to the muscle with increasing work
19 intensity of the quadriceps muscle (Richardson et al., 2002, 037513).
September 2009 4-19 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-2 CO concentration in pmol/mg wet weight tissue and fold tissue CO concentration
changes [normalized to background tissue concentrations] - human.
Exposure
Background
Fire
Fire + CO
CO asphyxiation
Adipose
3±1
5±4
[1.7]
18 ±29
[6.0]
25 ±27
[8.3]
Brain
3±3
7±5
[2.3]
17±14
[5.7]
72 ±38
[24.0]
Muscle
15±9
24 ±16
[1.6]
168 ±172
[11.2]
265 ±157
[17.7]
Heart
31 ±23
54 ±33
[1.7]
128 ±63
[4.1]
527 ± 249
[17.0]
Kidney
23 ±18
27 ±11
[1.2]
721 ± 427
[31.3]
885 ± 271
[38.5]
Lung
57 ±59
131 ±127
[2.3]
1097 ±697
[19.2]
2694 ±1730
[47.3]
Spleen
79 ±75
95 ±69
[1.2]
2290 ±1409
[29.0]
3455 ±1347
[43.7]
Blood
165 ±143
286 ±127
[1.7]
3623 ±1975
[22.0]
51 96 ±2625
[31.5]
Source: Vreman et al
% COHb
1 .5 ± 1 .2
3.8 ±3.2
[2.5]
40.7 ± 28.8
[27.1]
56.4 ±28.9
[37.6]
. (2006, 0982721
Table 4-3 CO concentration in pmol/mg fresh weight tissue and fold tissue CO concentration
changes [normalized to background tissue concentrations] - adult mouse.
Exposure
Background
500 ppm CO
30 pM heme
Testes
2 + 1
6 + 3
[3.0]
2 + 0
[1.0]
Intestine
4 + 2
9 + 7
[2.3]
3+1
[0.8]
Muscle
10 + 1
14 + 1
[1.4]
7 + 1
[0.7]
Brain
2 + 0
18 + 4
[9.0]
2 + 0
[1.0]
Heart
6+1
100+18
[16.7]
14 + 3
[2.3]
Liver
5+1
115 + 31
[23.0]
8 + 3
[1.6]
Kidney
7 + 2
120+12
[17.1]
7 + 2
[1.0]
Spleen
6+1
229 + 55
[38.2]
11 + 1
[1.8]
Lung
3+1
250 + 2
[83.3]
8 + 3
[2.7]
Blood
45 + 5
2648 + 400
[58.8]
88 + 10
[2.0]
% COHb
0.5
28
[56.0]
0.9
[1.8]
Source: Vreman et al. (2005,193786)
4.3.2.4. Other Tissues
1 CO binds with other hemoproteins, such as cytochrome P450, cytochrome c oxidase, catalase,
2 and peroxidase, but the possibility of this binding influencing CO-O2 kinetics has not been
3 established. CO transfers between COHb and tissue, the extent of which varies between organs.
4 Blood to tissue flux causes less CO to be expired following CO exposure than what is lost from the
5 blood in terms of COHb (Roughton and Root, 1945, 180418). This value is estimated to be 0.3-0.4%
6 min-1 or 0.24 mL/min (Bruce and Bruce, 2003, 193975: Prommer and, 2007, 180421). The
7 equilibration rate from blood to tissue is uncertain. Newly modeled CO trafficking kinetics shows
8 that CO continues to be taken up by the muscle and extravascular tissues well beyond the end of
9 exposure because of a less than instant equilibration (Bruce and Bruce, 2006, 193980). Table 4-2 and
10 Table 4-3 contain tissue CO concentrations from human and mouse under different CO exposure
11 conditions. The distribution of CO between the different human organs was shown to follow the
12 same pattern versus percent of the blood CO concentration, irrespective of the level of blood CO
13 (Vreman et al., 2006, 098272). Consistently, the spleen, lung, and kidney had the highest measured
14 CO concentration and the most dramatic increases over basal levels. The brain and adipose had the
September 2009
4-20
DRAFT - DO NOT CITE OR QUOTE
-------
1 lowest CO concentrations. In addition to these fatty tissues, the muscular tissues including the heart
2 and skeletal muscle had similarly low increases over background CO levels. This pattern was also
3 found in rodents exposed to exogenous CO (Table 4-3), however increased endogenous CO
4 produced after heme administration did not follow this pattern of uptake. Increased endogenous CO
5 production led to moderately increased CO present in the lung, heart, liver, and spleen and no change
6 in CO concentration in the testes, intestine, muscle, brain, and kidney. The spleen and liver have an
7 abundance of HO-1 expression and are involved in the catabolism of heme, thus it is expected to
8 have elevated CO concentrations in these organs after heme treatment. Also, elevated CO in the lung
9 is not surprising since it is the site of CO excretion. The tissues analyzed in these studies were
10 blanched before analysis, however contamination of the tissue sonicates with blood from the vessels
11 within each organ is a possible source of error. The measurements were presented by the authors as
12 minimum tissue CO concentrations, due to the possibility of rapid loss of CO from blood and tissue
13 exposed to the atmosphere, light, and elevated temperature (Chace et al., 1986, 012020; Ocak et al,
14 1985, 011641). These results are not consistent with older papers suggesting that negligible retention
15 of CO occurs in the liver or brain (Sokal et al., 1984, 011591: Topping, 1975, 193784).
4.3.3.Pulmonary and Tissue Elimination
16 Blood COHb concentrations are generally considered to have a monotonically decreasing,
17 second-order (logarithmic or exponential) elimination rate from equilibrium. However, more recent
18 reports have presented evidence for a biphasic washout curve, especially after short-term CO
19 exposure (Figure 4-9) (Bruce and Bruce, 2006, 193980; Shimazu et al., 2000, 016420; Wagner et al.,
20 1975, 010989). This event is modeled by a two-compartment system where the initial rapid decrease
21 is the washout rate from the blood, followed by a slower phase due to CO flux from the muscle and
22 extravascular compartments back to the blood. Tissue elimination rates have been reported as slower
23 than those for blood (Landaw, 1973, 010803). The biphasic curve is more obvious after short-term
24 CO exposure (less than 1 h), whereas long-term CO exposure (5 h or more) results in a virtually
25 monoexponential elimination, which could account for the historical findings. However, this
26 elimination curve also follows a biphasic curve with a slightly higher rate of elimination initially
27 (Shimazu et al., 2000, 016420). Differences in elimination kinetics could also be a result of the
28 variation in CO exposure duration (Weaver et al., 2000, 016421).
29 The elimination of COHb is affected by a number of factors, including duration of exposure,
30 PaO2, minute ventilation, the time post-exposure for analysis due to extravascular stores, as well as
31 inter-individual variability (Bruce and Bruce, 2006, 193980; Landaw, 1973, 010803; Shimazu, 2001,
32 016331). The elimination rate does not seem to be dependent upon the CO exposure source
33 (e.g., fire, non-fire CO exposure) (Levasseur et al., 1996, 080895). In addition, in a series of
34 poisoning cases, the COHb elimination half-life was not influenced by gender, age, smoke
September 2009 4-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 inhalation, history of loss of consciousness, concurrent tobacco smoking, degree of initial metabolic
2 acidosis (base excess), or the initial COHb level (Weaver et al., 2000, 016421). On the contrary, in
3 modeling the nonlinear kinetics of CO, a subject with a higher initial COHb will detoxify and
4 eliminate CO more rapidly (Gosselin et al., 2009, 190946). Similarly, it has been shown that the
5 absolute elimination rates are associated positively with the initial concentration of COHb, however
6 the relative rate of elimination, expressed as a percentage decline in COHb% after a measured time,
7 is independent of the initial COHb concentration (Wagner et al., 1975, 010989). COHb elimination
8 half-life falls as the fractional inspired O2 concentration increases. While breathing air at sea level
9 pressure, the expected half-life in adult males is approximately 285 min, but may be shorter in adult
10 females. With inhalation of normobaric 40% O2, the half-life falls to 75 min and further to 21 min
11 when breathing 100% O2 because of greater competition for Hb by O2 (Landaw, 1973, 010803).
12 Another study reports the half-life falls to 74 min (mean) after breathing 100% O2, although the
13 range in this particular study was 26-148 min (Weaver et al., 2000, 016421). In addition, COHb half-
14 life will fall further after normocapnic hyperoxic hyperpnea (i.e., hyperventilation while maintaining
15 normal CO2 pressure in high O2) (Takeuchi et al., 2000, 005675).
100
.a
I
O
O
30 60 90 120
Time (min)
150
180
Source: Adapted from Shimazu et al. (2000, 0164201
Figure 4-9 Changes in blood COHb after short-term and long-term exposure to CO,
representing the biphasic nature of CO elimination. Note: y-axis is log-scale.
September 2009
4-22
DRAFT - DO NOT CITE OR QUOTE
-------
4.3.4.COHb Analysis Methods
1 Blood COHb saturation can be analyzed using numerous methods with various benefits and
2 limitations. The most popular current techniques include gas chromatography (GC) and
3 spectrophotometry, specifically using CO-oximeters. CO-oximeters are commonly used because they
4 require little sample preparation and simultaneously measure COHb, O2Hb, methemoglobin, and
5 total hemoglobin concentration. However, at low concentrations of COHb relevant to ambient
6 exposure (< 5%), CO-oximeters overestimate COHb levels determined by GC (Mahoney et al.,
7 1993, 013859: Widdop, 2002, 030493). Conversely, at higher COHb levels (> 5%), CO-oximeters
8 will underestimate COHb concentrations. In addition to the inaccuracy of the CO-oximeters, some
9 studies report considerable imprecision in the results. Also, numerous substances or conditions can
10 interfere with CO-oximeter measurements (i.e., temperature, bilirubin, fetal hemoglobin).
11 Alternatively, GC is an accurate, precise, highly specific analysis method and is generally used as the
12 reference method for COHb analysis. GC requires the CO incorporated into blood or tissue samples
13 to first be released using a liberating agent such as potassium ferricyanide or sulfosalicylic acid
14 (Vreman et al., 2005, 193786; Vreman et al., 2006, 098272) and then measured directly or indirectly.
15 This methodology is more complex and time-consuming than spectrophotometry. In either analysis
16 method, it is important to remember that COHb measured at one site in the body does not necessarily
17 represent whole body CO uptake.
18 CO can also be measured directly in air or breath samples by using an electrochemical sensor
19 that depends on the electrical signal generated by the oxidation of CO. There are conflicting reports
20 on the correlation of exhaled CO (COex) with COHb. Multiple reports present positive correlation
21 coefficients (r) ranging from 0.92 and 0.98 in smoking subjects (Jarvis et al., 1980, 011813; Jarvis et
22 al., 1986, 012043; Landaw, 1973, 010803). Positive linear correlations have also been shown in
23 diseased patients with increased COHb (De las Heras et al., 2003, 194087). Others have reported no
24 correlation between low level COHb and COex and have suggested less correlation exists at the
25 lower levels of COex relevant to ambient exposures (Horvath et al., 1998, 087191; Scharte et al.,
26 2000, 194112). Finally, CO is endogenously produced in the nose and paranasal sinus which may
27 contribute to COex concentrations (Andersson et al., 2000, 011836).
4.4. Conditions Affecting Uptake and Elimination
4.4.1.Environment and Activity
28 Elevated CO exposure and COHb levels are dependent upon the changes in CO concentration
29 in the local environment. Pedestrians are exposed to high levels of CO for short time periods from
September 2009 4-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 vehicle exhaust at busy intersections (see also Chapter 3, Section 3.6). Higher exposure can also
2 result from riding in an automobile or stopping at busy intersections (Ott et al., 1994, 076546).
3 Indoor exposure occurs from ETS and unvented combustion appliances, such as natural gas cooking
4 stoves, attached garages, and gas fireplaces, the latter of which can result in CO concentrations of
5 over 100 ppm (Dutton et al., 2001, 021307). Recreational exposure at levels exceeding 200 ppm and
6 peaks of 1,600 ppm could occur in indoor ice rinks using fossil fuel powered ice resurfacers and
7 coliseums housing malfunctioning equipment or poor ventilation (Levesque et al., 2000, 011886;
8 Pelham et al., 2002, 025716). Certain occupations provide instances and conditions for transient
9 moderate-to-high CO levels, including fire fighters and machinery operators. Such transient
10 exposures have the ability to increase COHb levels. For example, exposure for 5 min or less of a
11 resting individual to 6,600 ppm CO will result in up to 20% COHb (Benignus et al., 1994, 013908).
12 Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase
13 in gas exchange. O2 consumption can increase more than 10 fold during exercise. Similarly,
14 ventilation, membrane and lung diffusing capacity, pulmonary capillary blood volume, and cardiac
15 output increase proportional to work load. The majority of these changes facilitate CO uptake and
16 transport, by increasing gas exchange efficiency. Likewise, the COHb elimination rate increases with
17 physical activity, causing a decrease in COHb half-life (Joumard et al., 1981, 011330).
4.4.2.Altitude
18 Increased altitude changes a number of factors that contribute to the uptake and elimination of
19 CO. The relationship between altitude and CO exposure has been discussed in depth in the 2000 CO
20 AQCD and other documents (U.S. EPA, 1978, 086321). In an effort to maintain proper O2 transport
21 and supply, physiological changes occur as compensatory mechanisms to combat the decreased
22 barometric pressure and resulting altitude induced hypobaric hypoxia (HH). HH, unlike CO hypoxia,
23 causes humans to hyperventilate, which reduces arterial blood CO2 (hypocapnia) and increases
24 alveolar partial pressure of O2. Hypocapnia will lead to difficulty of O2 dissociation and decreased
25 blood flow, thus reducing tissue O2 supply. HH increases blood pressure (BP) and cardiac output and
26 leads to redistribution of blood from skin to organs and from blood vessels to extravascular
27 compartments. Generally these changes will favor increased CO uptake and COHb formation, as
28 well as CO elimination. In hypoxic conditions both CO and O2 bind reduced Hb through a
29 competitive-parallel reaction (Chakraborty et al., 2004, 193759). Breathing CO (9 ppm) at rest at
30 altitude produced higher COHb compared to sea level (McGrath et al., 1993, 013865). whereas high
31 altitude exposure with exercise caused a decrease in COHb levels versus similar exposure at sea
32 level (Horvath et al., 1988, 012725). This decrease could be a shift in CO storage or suppression of
33 COHb formation, or both. Altitude also increases the baseline COHb levels by inducing endogenous
September 2009 4-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 CO production. Initial HH increased lung HO-1 protein and activity, whereas chronic HH induced
2 endogenous CO production in nonpulmonary sites (see Section 4.5) (Carraway et al., 2000, 021096).
3 As the length of stay increases at high altitude, acclimatization occurs, inducing
4 hyperventilation, polycythemia or increased red blood cell count, and increased tissue capillarity and
5 Mb content in skeletal muscle, which could also favor increased CO uptake. Most of the early
6 adaptive changes gradually revert to sea level values. However, differences in people raised at high
7 altitude persist even after reacclimatization to sea level (Hsia, 2002, 193857).
4.4.3.Physical Characteristics
8 Certain physical characteristics (e.g., age, sex, pregnancy) can alter the variables that influence
9 the uptake, distribution, and elimination of CO. Values of CO uptake and elimination change with
10 age. Young children eliminate COHb more rapidly than adults after CO exposure (Joumard et al.,
11 1981, 011330; Klasner et al., 1998, 087196). After infancy, the COHb half-life increases with age,
12 nearly doubling between 2 and 70 yr (Joumard et al., 1981, 011330). The rate of this increase in CO
13 elimination is very rapid in the growing years (2-16 yr of age), but slows beyond adolescence.
14 Alveolar volume and DLCO increase with increasing body length of infants and toddlers (Castillo et
15 al., 2006, 193234). suggesting a further degree of lung development and faster CO uptake. After
16 infancy, increasing age decreases DLCO and increases VA/Q mismatch, causing it to take longer to
17 both load and eliminate CO from the blood (Neas and Schwartz, 1996, 079363).
18 COHb concentrations are generally lower in female subjects than in male subjects (Horvath et
19 al., 1988, 012725) and the COHb half-life may be longer in healthy men than in women of the same
20 age, which may be partially explained by differences in muscle mass or the slight correlation
21 between COHb half-life and increased height (Joumard et al., 1981, 011330). However, women do
22 have a higher rate of endogenous production while in the progesterone phase of the menstrual cycle
23 and during pregnancy (see Section 4.5). The rate of decline of DLCO with age is lower in middle-
24 aged women than in men; however, it evens out towards older age (Neas and Schwartz, 1996,
25 079363). Women also tended to be more resistant to altitude hypoxia (Horvath et al., 1988, 012725).
26 Ethnicity does alter physiological variables that determine CO uptake and kinetics. Lung
27 volumes are 10-15% less in both Asian and African-American populations when compared to
28 Caucasians. This causes a reduced alveolar surface area (20% less than estimated values) for gas
29 exchange, leading to a 13% difference in diffusion capacity, DLCO (Pesola et al., 2004, 193842;
30 Pesola et al., 2006, 193855). Certain factors such as socioeconomic status (SES) were not controlled
31 for in these studies. SES has been shown to affect pulmonary function, including decreasing DLCO
32 (Hegewald and Crapo, 2007, 193923).
September 2009 4-25 DRAFT - DO NOT CITE OR QUOTE
-------
4.4.3.1. Fetal Pharmacokinetics
1 Inhaled CO by pregnant animals quickly passes the placental barriers and enters the fetal
2 circulation (Longo, 1977, 012599). Fetal CO pharmacokinetics do not follow the same kinetics as
3 maternal CO exposure, making it difficult to estimate fetal COHb based on maternal levels. Human
4 fetal Hb has a higher affinity for CO than adult Hb (Di Cera et al., 1989, 193998). Maternal and fetal
5 COHb concentrations have been modeled as a function of time using a modified CFK equation
6 (Figure 4-10) (Hill et al., 1977, 011315). At steady-state conditions, the fetal COHb is up to 10-15%
7 higher than the maternal COHb levels, for example, exposure to 30 ppm CO results in a maternal
8 COHb of 5% and a fetal COHb of 5.75%. The fetal CO uptake lags behind the maternal for the first
9 few hours but later may overtake the maternal values. Fetal COHb equilibrium may not be reached
10 for 36-48 h after exposure. Similarly, during washout, the fetal COHb levels are maintained for
11 longer, with a half-life of around 7.5 h versus the maternal half-life of around 4 h (Longo and Hill,
12 1977. 010802).
300
\ Maternal <
F.tol
12 16 20 24 eoQ 4
TIME (Hourt)
12
16 20 24
Source: Hill et al. (1977, 0113151
Figure 4-10 Predicted maternal and fetal COHb during prolonged exposure to CO
(30-300 ppm) and washout from equilibrium values with no CO.
September 2009
4-26
DRAFT - DO NOT CITE OR QUOTE
-------
4.4.4.Health Status
1 Health status can influence the toxicity involved with CO exposure by influencing the severity
2 of hypoxia resulting from CO exposure. Any condition that would alter the blood O2 carrying
3 capacity or content will result in a greater risk from COHb induced hypoxia and decreased tissue O2
4 delivery. The severity of this effect depends upon the initial level of hypoxia.
5 Anemias are a group of diseases that result in insufficient blood O2 or hypoxia due to Hb
6 deficiency through hemolysis, hemorrhage, or reduced hematopoiesis. Anemia may result from
7 pathologic conditions characterized by chronic inflammation such as malignant tumors or chronic
8 infections (Cavallin-Stahl et al, 1976, 086306: Cavallin-Stahl et al., 1976, 193239). The bodies of
9 people with anemia compensate causing cardiac output to increase as both heart rate and stroke
10 volume increase. The endogenous production of CO, thus COHb, is increased in patients with
11 hemolytic anemia due to increased heme catabolism, causing an increased baseline COHb
12 concentration. One of the most prevalent anemias arises from a single-point mutation of Hb, causing
13 sickle cell diseases. The Hb affinity for O2 and O2 carrying capacity is reduced causing a shift to the
14 right in the O2 dissociation curve. It is well documented that African-American populations have a
15 higher incidence of sickle cell anemia, which may be a risk factor for CO hypoxia.
16 Chronic obstructive pulmonary disease (COPD) is often accompanied by a number of changes
17 in gas exchange, including increased deadspace volume (VD) and ventilation-perfusion ratio (VA/Q)
18 inequality (Marthan et al., 1985, 086334). which could slow both CO uptake and elimination.
19 Patients with pulmonary sarcoidosis, a restrictive lung disease, may also have a decrease in lung
20 volumes, a loss of DLCO, and gas exchange abnormalities during exercise, including decreased
21 arterial oxygen pressure (PaO2) and increased alveolar-arterial oxygen pressure difference (Lamberto
22 etal.,2004,193_845).
23 Individuals with heart disease may be at a greater risk from CO exposure since they may
24 already have compromised O2 delivery. Time to onset of angina was reduced after exposure to
25 100 ppm carbon monoxide, compared to clean air (Kleinman et al., 1998, 047186). Hyperlipidemic
26 patients may have decreased CO diffusion capacity, a loss of V/Q gradient, and a decrease in PaO2
27 (Enzi et al., 1976) (see section 5.2 discussing cardiovascular effects).
4.5. Endogenous CO Production and Metabolism
28 Humans breathing air containing no environmental sources of CO will still have a low
29 measurable level of circulating COHb. This is due to endogenous CO production from heme protein
30 catabolism among other sources. In the normal degradation of RBC Hb, the porphyrin ring of heme
31 is broken at the a-methene bridge by HO. HO is colocalized with NADPH-flavoprotein reductase
September 2009 4-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 and biliverdin reductase on the endoplasmic reticulum, where it catabolizes heme in an O2 and
2 NADPH-dependent manner to biliverdin, ferrous iron, and CO. Biliverdin is then further broken
3 down by biliverdin reductase into bilirubin, a powerful endogenous antioxidant. Two main HO
4 isoforms exist, HO-1 and HO-2. Expression of HO-1 is inducible, whereas HO-2 is constitutively
5 expressed. The major site of heme catabolism, and thus the major organ of CO production, is the
6 liver, followed by the spleen, brain, and erythropoietic system (Berk et al., 1976, 012603). These
7 rates of CO formation may be due to higher levels of HO activity in these tissues. The whole body
8 production rate of CO is approximately 18.8 umol/h (0.42 mL/h or 0.007 mL/min) and produces
9 between 400-500 umol CO per day (Coburn et al., 1963, 013971: Coburn et al., 1964, 013956:
10 Coburn et al., 1966, 010984) (Figure 4-11). The endogenous rate of production varied somewhat
11 within individuals measured on multiple days (±4.5 umol/h and ±0.35% COHb) (Coburn et al.,
12 1966, 010984). However, these measurements of day-to-day CO production variability were
13 comparable to the equipment measurement error reported (±3.1 umol/h). The endogenous rate of CO
14 formation has been shown to vary between different tissues, ranging from 0.029 nmol/mg protein/h
15 in chorionic villi of term human placenta to 0.28 nmol/mg protein/h in rat olfactory receptor neurons
16 in culture and in rat liver perfusate (Marks et al., 2002, 030616). however these estimations are
17 uncertain since CO is quickly scavenged in the cytosol of living cells. CO is endogenously produced
18 in the nose and paranasal sinus which may contribute to exhaled CO concentrations (Andersson et
19 al., 2000, 011836). It is also important to note that increased endogenous CO production does not
20 universally lead to an increase in COHb saturation.
21 HO mediated metabolism functions as the rate-limiting enzyme step in heme degradation and
22 endogenous CO production (Wu and Wang, 2005, 180411). Three isoforms of HO exist, but HO-1 is
23 the only inducible form (Maines and Kappas, 1974, 193976: Maines et al., 1986, 193978:
24 McCoubrey WK et al., 1997, 016715). Endogenous CO production can be increased by the up-
25 regulation of HO-1 expression and activity by inducers such as oxidative stress, hypoxia, heavy
26 metals, sodium arsenite, heme and heme derivatives, various cytokines, and also exogenous CO (Wu
27 and Wang, 2005, 180411). High levels of CO (2,500 ppm) have been shown to increase HO-1
28 activity in the brain of rats, as well as liberate intracellular heme to further stimulate endogenous CO
29 production (Cronj e et al., 2004, 180440).
September 2009 4-28 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Coburnetal. (1963, 013971)
Delivoria-Papadopoulos et al. (1974, 086316)
Delivoria-Papadopoulos et al. (1974, 0863161
Lonao (1970, 013922)
Longo (1970, 013922)
Lonao (1970, 013922)
Longo (1970, 013922)
Coburnetal. (1966, 010984)
Coburnetal. (1966, 010984)
Zegdi et al. (2002, 037461)
Scharteetal. (2006, 194115)
Scharteetal. (2006, 194115)
Condition
1 Adult male
1 Female - Estrogen phase
I Pregnancy
] Fetal
I Rnstpartnm - 74 h
I Postpartum - day 4
i Hpmnlyfr inpmia min
1 Hemnlytic anemia may
1 Sepsis si irvivnrs
1 Cardiac disease patients may
1 Critically ill on dialysis may
0.01 0.02 0.03 0.04 0.05
Endogenous CO production (mL/min)
0.06
Figure 4-11
Representative estimates of endogenous CO production rates resulting from
various conditions and diseases.
1 Not all endogenous CO production is derived from Hb breakdown. Other hemoproteins, such
2 as Mb, cytochromes, peroxidases, and catalase, contribute 20-25% to the total amount of endogenous
3 CO (Berk et al., 1976, 012603). All of these sources result in a normal blood COHb concentration
4 between 0.4 and 1% (Coburn et al., 1965, 011145). The level of endogenous production can be
5 altered by drugs or a number of physiological conditions that alter RBC destruction, other
6 hemoprotein breakdown, or HO-1 expression and activity (Figure 4-11). Nicotinic acid (Lundh et al.,
7 1975, 086332). allyl-containing compounds (acetamids and barbiturates) (Mercke et al., 1975,
8 086303). diphenylhydantoin (Coburn, 1970, 010625). progesterone (Delivoria-Papadopoulos et al.,
9 1974, 086316). contraceptives (Mercke et al., 1975, 086308). and statins (Muchova et al., 2007,
10 194098) will increase CO production. Compounds such as carbon disulfide and sulfur-containing
11 chemicals (parathion and phenyltiourea) will increase CO by acting on P450 system moieties
12 (Landaw et al., 1970, 012605). The P450 system may also cause large increases in CO produced
13 from the metabolic degradation of dihalomethanes leading to very high (>10%) COHb levels (Bos et
14 al., 2006, 194084; Manno et al., 1992, 013707). which can be further enhanced by prior exposure to
15 hydrocarbons or ethanol (Pankow et al., 1991, 013551; Wirkner et al., 1997, 082642). Minor sources
16 of endogenous CO include auto-oxidation of phenols, flavenoids, and halomethanes, photo-oxidation
17 of organic compounds, and lipid peroxidation of cell membrane lipids (Rodgers et al., 1994,
18 076440).
19 Women experience fluctuating COHb levels through the menstrual cycle when endogenous
20 CO production doubles in the progesterone phase (0.62 mL/h versus 0.32 mL/h in estrogen phase)
September 2009
4-29
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
(Delivoria-Papadopoulos et al, 1974, 086316: Mercke and Lundh, 1976, 086309). Similarly,
endogenous CO production increases during pregnancy (0.92 mL/h) due to contributions from fetal
endogenous CO production (0.036 mL/h) and altered hemoglobin metabolism (Hill et al., 1977,
011315: Longo, 1970, 013922).
Any disturbance in RBC hemostasis by acceleration of destruction of hemoproteins will lead
to increased production of CO (Figure 4-12 and Figure 4-13). Pathologic conditions such as anemias,
hematomas, thalassemia, Gilbert's syndrome with hemolysis, and other hematological diseases and
illness will accelerate CO production (Berk et al., 1974, 012386: Hampson and Weaver, 2007,
190272: Meyer et al., 1998, 047530: Solanki et al., 1988, 012426: Sylvester et al., 2005, 191954).
Patients with hemolytic anemia exhibit COHb levels at least 2- to 3-fold higher than healthy
individuals and CO production rates 2- to 8-fold higher (Coburn et al., 1966, 010984). A recent study
reports COHb levels elevated to levels between 4.6% and 9.7% due to drug-induced hemolytic
anemia (Hampson and Weaver, 2007, 190272) and between 3.9% and 6.7% due to an unstable
hemoglobin disorder (Hb Zurich) (Zinkham et al., 1980, 011435). Endogenous CO production rate
varied from 0.70 to 3.18 mL/h in anemic patients (Coburn et al., 1966, 010984).
Study
Hampson and Wfeaver (2007, 190272)
Hampson and Wfeaver (2007, 190272)
Zinkham etal. (1980, 011435)
Zinkham etal. (1980, 011435)
Coburn etal. (1966, 010984)
Coburn etal. (1966, 010984)
Iran et al. (2007, 090752)
De Las Heras et al. (2003, 194087)
De Las Heras et al. (2003, 1940871
Yasuda et al. (2005, 191953)
Morimatsu et al. (2006, 194097)
De Las Heras et al. (2003, 194087)
Condition
I Hemolytic anemia mav
I Hemolytic anemia min
I Hh Zurich min
I Hh Zurich min
I Hemnlytic anemia may
i Hemolytic anemia min
I Liver transplant
I Peritonitis (SBP)
I Cirrhosis
I Exacerbated COPD
I Critically ill
• Healthy
246
COHb (%)
10
Figure 4-12
Representative COHb saturation resulting from various diseases and conditions.
SBP: Spontaneous bacterial peritonitis
Critically ill patients exhale more CO and have higher endogenous CO production than
healthy controls, likely due to both increased heme turnover as well as upregulation of the
expression and activity of HO-1 (Morimatsu et al., 2006, 194097: Scharte et al., 2000, 194112:
Scharte et al., 2006, 194115) (Figure 4-13). CO production weakly correlates with the multiple organ
dysfunction score (MODS), which estimates severity of organ dysfunction; however, it did not
September 2009
4-30
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
correlate with Acute Physiology and Chronic Health Evaluation II score (APACHE II) (Scharte et
al., 2006, 194115) or the sequential organ failure assessment score (SOFA) (Morimatsu et al., 2006,
194097). Critically ill patients that survived had a higher exhaled CO (COex) concentration than
nonsurvivors (median 3.9 ppm versus 2.4 ppm) (Morimatsu et al., 2006, 194097). Similarly, patients
that survived severe sepsis had a higher CO production than those that did not survive (14.7 ± 5.3
versus 8.5 ± 3.3 (il/kg/h) (Zegdi et al., 2002, 0374611
Study
Yamava et al. (1998, 047525), Zavasu et al. (1997, 084291)
De Las Heras et al. (2003, 194087)
De Las Heras et al. (2003, 194087)
Sylvester et al. (2005, 191954)
Yamava etal. (1998, 047525)
Zavasu etal. (1997, 0842911
Zavasu etal. (1997, 084291)
Pared! etal. (1999, 118798)
Horvath et al. (1998, 087190)
Pared! etal. (1999, 1187981
Pared! etal. (1999, 1187981
Morimatsu et al. (2006, 1940971
Morimatsu et al. (2006, 194097)
Condition
I Peritonitis (SBP)
I Sickle Coll Anemia
I IIRTI
I Asthma w/ steroids
I Asthma
| Cystic fihrnsis
I Bronchiectasis
i Type 2 diabetes
I Type 1 diabetes
I Critically ill
H Healthy
3 6 9 12
Exhaled CO (average fold change from healthy)
15
Figure 4-13.
Representative exhaled CO concentrations (ppm) resulting from various
conditions plotted as fold increases over healthy human controls from each
study. SBP: Spontaneous bacterial peritonitis; URTI: Upper respiratory tract
infection
Diseases involving inflammation and infection tend to have increased endogenous CO
production. Patients with severe sepsis or septic shock had a higher COex and CO endogenous
production compared to control patients, and the CO production decreased with treatment of the
disease (i.e., antibiotics, surgery) (Zegdi et al., 2002, 037461). Similarly, patients with pre-existing
cardiac disease, as well as patients with renal failure, who undergo dialysis, produced higher
amounts of endogenous CO compared to other critically ill patients (Scharte et al., 2006, 194115).
High plasma COHb levels were found in nonsmoking patients evaluated for liver transplantation
(mean, 2.1%), however this increase was not correlated with the Model for End Stage Liver Disease
(MELD) score or Child Turcotte Pugh score, used to assess the degree of liver impairment (Tran et
al., 2007, 090752). Further investigation, in cirrhotic patients, with and without ascites, provided
September 2009
4-31
DRAFT - DO NOT CITE OR QUOTE
-------
1 evidence for increased plasma CO concentrations, HO-1 activity in polymorphonuclear cells,
2 exhaled CO, and blood COHb (De las Heras et al, 2003, 194087: Tarquini et al, 2009, 1941171
3 COex, plasma CO, and COHb levels were correlated with the Child-Pugh score, and thus the
4 severity of disease. These parameters were significantly higher in patients with ascites or with
5 spontaneous bacterial peritonitis (SBP) (COHb, healthy: 0.6 ± 0.1%; cirrhosis: 1.0 ± 0.1%; with
6 ascites: 1.6 ± 0.2%; with SBP: 1.9 ± 0.2%). Both COex and COHb levels decreased after resolution
7 of the infection in patients with SBP, reaching values similar to noninfected patients within 1 month
8 (De las Heras et al., 2003, 194087). Endotoxin concentration was correlated with plasma CO levels,
9 suggesting a link between systemic endotoxemia and increased activity or expression of the HO/CO
10 system (Tarquini et al., 2009, 194117). COex concentrations are also elevated in patients with
11 diabetes (Type 1: 4.0 ± 0.7 ppm; Type 2: 5.0 ± 0.4 ppm; healthy: 2.9 ± 0.2 ppm), and correlated with
12 blood glucose levels and duration of disease (Paredi et al., 1999, 194102). Likewise, obese Zucker
13 rats, a model of metabolic syndrome with insulin resistance, have increased respiratory CO excretion
14 and COHb levels compared to lean Zucker rats (3.9 ± 0.1% versus 3.0 ± 0.1% COHb), which is
15 decreased by HO inhibition (Johnson et al., 2006, 193874).
16 Endogenous CO is also increased in airway inflammatory diseases. Patients with upper
17 respiratory tract infections exhaled higher CO concentrations than normal controls and this increase
18 was attenuated after recovery (Yamaya et al., 1998, 047525). Arterial COHb levels have been related
19 to disease severity in COPD patients (Yasuda et al., 2005, 191953). Bronchiectasis patients had
20 higher COex, however anti-inflammatory treatment did not decrease the CO levels (Horvath et al.,
21 1998, 087191). Patients with cystic fibrosis had higher COex than normal controls (6.7 ±0.6 ppm
22 versus 2.4 ± 0.4 ppm) and patients treated with steroids had a decrease in CO levels (8.4 ±1.0 ppm
23 versus 5.1 ± 0.5 ppm) (Paredi et al., 1999, 118798). Increased arterial COHb was reported in patients
24 with bronchial asthma, pneumonia, idiopathic pulmonary fibrosis, pyelonephritis, and active
25 rheumatoid arthritis (Yasuda et al., 2002, 035206: Yasuda et al., 2004, 191955). Similarly, asthmatic
26 patients exhibit an elevation of COex that decreases with corticosteroid therapy (nonsmoking
27 controls: 1.5 ±0.1 ppm; asthmatics without corticosteroids: 5.6 ± 0.6 ppm; with corticosteroids: 1.7
28 ±0.1 ppm; smoking controls: 21.6 ±2.8 ppm) (Zayasu et al., 1997, 084291). These results were
29 confirmed and associated with increased expression of HO-1 in airway macrophages (Horvath et al.,
30 1998, 087190). Similarly, COex was increased in patients with allergic rhinitis during the pollen
31 season; however, their COex was similar to control subject levels out of season (Monma et al., 1999,
32 180426). Similarly, endogenous CO production and HO-1 expression in nasal mucosa was correlated
33 with allergic rhinitis in guinea pigs as described in Section 5.1 (Yu et al., 2008, 192384).
34 Altitude has been shown to be positively associated with baseline COHb concentrations
35 (McGrath, 1992, 013528: McGrath et al., 1993, 013865). This increase in COHb with altitude
36 induced hypoxia has also been associated with increases in the mRNA, protein, and activity of HO-1
September 2009 4-32 DRAFT - DO NOT CITE OR QUOTE
-------
1 in rats and cells leading to enhanced endogenous CO production (Carraway et al., 2002, 026018; Lee
2 et al., 1997, 082641). Whether other variables such as an accelerated metabolism or a greater pool of
3 Hb, transient shifts in body stores, or a change in the elimination rate of CO play a role has not been
4 explored.
5 Because of the sensitivity of COHb to changes in the metabolic state, ranges of endogenous
6 COHb levels in the population are uncertain. However, baseline levels of COHb, which include
7 ambient, non-ambient, and endogenous production of CO, have been measured in the population.
8 COHb levels in packed red blood cell units reserved for use between 2004-2005 averaged 0.78 ±
9 1.48%, with 10.3% of samples having COHb levels of 1.5% or greater and a maximum measurement
10 of 12% (Ehlers et al., 2009, 194089). This study reported a decrease from a study conducted in!982-
11 83 in the number of units with elevated COHb; at that time, 49% of units had COHb levels >1.5%
12 ((Aronow et al., 1984, 194083) versus 10.3% in 2004-05). Another study calculated that 23% of
13 donated blood units had COHb levels exceeding 1.5%, with the highest measurement being 7.2%
14 (Aberg et al., 2009, 194082). Smoking is the main factor causing increased blood concentrations of
15 CO. A dose response relationship existed with the number of cigarettes smoked a day (nonsmoker:
16 1.59 ± 1.72%; 1-5 cig/day: 2.31 ± 1.94%; 6-14 cig/day: 4.39 ± 2.48%; 15-24 cig/day: 5.68 ± 2.64%;
17 > 25 cig/day: 6.02 ± 2.86% COHb). The mean baseline COHb value for former smokers was higher
18 than that of never smokers in this prospective cohort study (1.96 ± 1.87 versus 1.59 ± 1.72%) (Hart
19 etal, 2006, 194092).
20 Endogenous CO is removed from the body mainly by expiration and oxidation. CO will
21 diffuse across the alveolar-capillary membrane and then is exhaled. This event has been used as a
22 noninvasive measurement of endogenous CO and CO body load (Stevenson et al., 1979, 193767).
23 CO can also be oxidized to CO2 by cytochrome c oxidase in the mitochondria (Fenn, 1970, 010821;
24 Young and Caughey, 1986, 012091). However, the rates of CO metabolism are much slower than the
25 rates of endogenous CO production, with the rate of consumption representing only 10% of the rate
26 of CO production in dogs (Luomanmaki and Coburn, 1969, 012319).
4.6. Summary and Conclusions
27 CO elicits various health effects by binding with and altering the function of a number of
28 heme-containing molecules, mainly Hb. The formation of COHb reduces the O2-carrying capacity of
29 blood and impairs the release of O2 from O2Hb to the tissues. Venous COHb levels have been
30 modeled mainly by the CFK equation, but more recent models have included venous and arterial
31 blood mixing and Mb and extravascular storage compartments, as well as other dynamics of CO
32 physiology. These models have indicated that CO has a biphasic elimination curve, due to initial
33 washout from the blood followed by a slower flux from the tissues. The flow of CO between the
September 2009 4-33 DRAFT - DO NOT CITE OR QUOTE
-------
1 blood and alveolar air or tissues is controlled by diffusion down the pCO gradient. The uptake of CO
2 is governed not only by this CO pressure differential, but also by physiological factors, such as
3 minute ventilation and lung diffusing capacity, that can, in turn, be affected by conditions such as
4 exercise, age, and health. Susceptible populations, including health compromised individuals and
5 developing fetuses, are at a greater risk from COHb induced health effects due to altered CO
6 kinetics, compromised cardiopulmonary processes, and increased baseline hypoxia levels. Altitude
7 may also significantly affect the kinetics of COHb formation. Compensatory mechanisms, such as
8 increased cardiac output, combat the decrease in barometric pressure. Altitude also increases the
9 endogenous production of CO through upregulation of HO-1. CO is considered a second messenger
10 and is endogenously produced from the catabolism of heme proteins by enzymes such as HO-1. A
11 number of diseases and conditions affect endogenous CO production, possibly causing a higher
12 endogenous COHb level. Finally, CO is removed from the body by expiration or oxidation to CO2.
September 2009 4-34 DRAFT - DO NOT CITE OR QUOTE
-------
References
Aberg AM; Sojka BN; Winso O; Abrahamsson P; Johansson G; Larsson JE. (2009). Carbon monoxide concentration in
donated blood: Relation to cigarette smoking and other sources. Transfusion, 49: 347-353. 194082
Abi-Esber L; El-Fadel M. (2008). In-vehicle CO ingression: Validation through field measurements and mass balance
simulations. Sci Total Environ, 394: 75-89. 190939
Abram SR; Hodnett BL; Summers RL; Coleman TG; Hester RL. (2007). Quantitative circulatory physiology: An
integrative mathematical model of human physiology for medical education. Adv Physiol Educ, 31: 202-210.
193859
Alcantara RE; Xu C; Spiro TG; Guallar V. (2007). A quantum-chemical picture of hemoglobin affinity. Proc Natl Acad Sci
USA, 104: 18451-18455. 193867
Andersson JA; Uddman R; Cardell L-O. (2000). Carbon monoxide is endogenously produced in the human nose and
paranasal sinuses. J Allergy Clinlmmunol, 105: 269-273. 011836
Aronow WS; O'Donohue WJ Jr; Freygang J; Sketch MH. (1984). Carboxyhemoglobin levels in banked blood. Chest, 85:
694-695. 194083
Arora M; Chandra J; Suri JC; Narayan S; Dutta AK. (2001). Pulmonary function tests in beta thalassemia. Indian J Pediatr,
68: 239-242. 186713
Benignus VA; Coleman T; Eklund CR; Kenyon EM. (2006). A general physiological and toxicokinetic (GPAT) model for
simulating complex toluene exposure scenarios in humans. Toxicol Mech Meth, 16: 27-36. 151344
Benignus VA; Hazucha MJ; Smith MV; Bromberg PA. (1994). Prediction of Carboxyhemoglobin formation due to transient
exposure to carbon monoxide. J Appl Physiol, 76: 1739-1745. 013908
Berk PD; Blaschke TF; Scharschmidt BF; Waggoner JG; Berlin NI. (1976). A new approach to quantitation of the various
sources of bilirubin in man. J Lab Clin Med, 87: 767-780. 012603
Berk PD; Rodkey FL; Blaschke TF; Collison HA; Waggoner JG. (1974). Comparison of plasma bilirubin turnover and
carbon monoxide production in man. J Lab Clin Med, 83: 29-37. 012386
Bos PM; Zeilmaker MJ; van Eijkeren JC. (2006). Application of physiologically based pharmacokinetic modeling in
setting acute exposure guideline levels for methylene chloride. Toxicol Sci, 91: 576-585. 194084
Bruce EN; Bruce MC. (2003). A multicompartment model of Carboxyhemoglobin and carboxymyoglobin responses to
inhalation of carbon monoxide. J Appl Physiol, 95: 1235-1247. 193975
Bruce EN; Bruce MC; Erupaka K. (2008). Prediction of the rate of uptake of carbon monoxide from blood by extravascular
tissues. Respir Physiol Neurobiol, 161: 142-159. 193977
Bruce MC; Bruce EN. (2006). Analysis of factors that influence rates of carbon monoxide uptake, distribution, and
washout from blood and extravascular tissues using a multicompartment model. J Appl Physiol, 100: 1171-1180.
193980
Carraway MS; Ghio AJ; Carter JD; Piantadosi CA. (2000). Expression of heme oxygenase-1 in the lung in chronic
hypoxia. Am J Physiol, 278: L806-L812. 021096
Carraway MS; Ghio AJ; Suliman HB; Carter JD; Whorton AR; Piantadosi CA. (2002). Carbon monoxide promotes
hypoxic pulmonary vascular remodeling. Am J Physiol, 282: L693-L702. 026018
Castillo A; Llapur CJ; Martinez T; Kisling J; Williams-Nkomo T; Coates C; Tepper RS. (2006). Measurement of single
breath-hold carbon monoxide diffusing capacity in healthy infants and toddlers. Pediatr Pulmonol, 41: 544-550 .
193234
Cavallin-Stahl E; Mercke C; Lundh B. (1976). Carbon monoxide production in patients with breast carcinoma. Br J
Haematol, 32: 177-182. 086306
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS)
September 2009 4-35 DRAFT - DO NOT CITE OR QUOTE
-------
Cavallin-Stahl E; Mercke C; Lundh B. (1976). Erythropoiesis and carbon monoxide production in Hodgkin's disease. Br J
Haematol, 32: 167-175. 193239
Chace DH; Goldbaum LR; Lappas NT. (1986). Factors affecting the loss of carbon monoxide from stored blood samples. J
AnalToxicol, 10: 181-189. 012020
Chakraborty S; Balakotaiah V; Bidani A. (2004). Diffusing capacity reexamined: relative roles of diffusion and chemical
reaction in red cell uptake of O2, CO, CO2, and NO . J Appl Physiol, 97: 2284-2302. 193759
Chung SJ. (1988). Formulas predicting carboxyhemoglobin resulting from carbon monoxide exposure. Vet Hum Toxicol,
30: 528-532. 012749
Coburn RF. (1967). Endogenous carbon monoxide production and body CO stores. J Intern Med, 472: 269-282. 011144
Coburn RF. (1970). Enhancement by phenobarbital and diphenylhydantoin of carbon monoxide production in normal man.
N EnglJ Med, 283: 512-515. 010625
Coburn RF. (1970). The carbon monoxide body stores. Ann N YAcad Sci, 174: 11-22. 013916
Coburn RF; Blakemore WS; Forster RE. (1963). Endogenous carbon monoxide production in man. J Clin Invest, 42: 1172-
1178.013971
Coburn RF; Forster RE; Kane PB. (1965). Considerations of the physiological variables that determine the blood
carboxyhemoglobin concentration in man. J Clin Invest, 44: 1899-1910. 011145
Coburn RF; Williams WJ; Forster RE. (1964). Effect of erythrocyte destruction on carbon monoxide production in man. J
Clin Invest, 43: 1098-1103. 013956
Coburn RF; Williams WJ; Kahn SB. (1966). Endogenous carbon monoxide production in patients with hemolytic anemia. J
Clin Invest, 45: 460-468. 010984
Cronenberger C; Mould DR; Roethig H-J; Sarkar M. (2007). Population pharmacokinetic analysis of carboxyhaemoglobin
concentrations in adult cigarette smokers. Br J Clin Pharmacol, 65: 30-39. 194671
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA. (2004). Carbon monoxide actuates O2-limited heme
degredations in the rat brain. Free Radic Biol Med, 3 7: 1802-1812. 180440
Delivoria-Papadopoulos M; Coburn RF; Forster RE. (1974). Cyclic variation of rate of carbon monoxide production in
normal women. J Appl Physiol, 36: 49-51. 086316
De las Heras D; Fernandez J; Gines P; Cardenas A; Ortega R; Navasa M; Barbera JA; Calahorra B; Guevara M; Bataller R;
Jimenez W; Arroyo V; Rodes J. (2003). Increased carbon monoxide production in patients with cirrhosis with and
without spontaneous bacterial peritonitis. Hepatology, 38: 452-459. 194087
Di Cera E; Doyle ML; Morgan MS; De Cristofaro R; Landolfi R; Bizzi B; Castagnola M; Gill SJ. (1989). Carbon
monoxide and oxygen binding to human hemoglobin FO. Biochemistry, 28: 2631-2638. 193998
Douglas CG; Haldane JS; Haldane JBS. (1912). The laws of combination of haemoglobin with carbon monoxide and
oxygen. J Physiol, 44: 275-304. 013965
Duci A; Chaloulakou A; Spyrellis N. (2003). Exposure to carbon monoxide in the Athens urban area during commuting.
Sci Total Environ, 309: 47-58. 044199
Dutton SJ; Hannigan MP; Miller SL. (2001). Indoor pollutant levels from the use of unvented natural gas fireplaces in
Boulder, Colorado. J Air Waste ManagAssoc, 51: 1654-1661. 021307
Ehlers M; Labaze G; Hanakova M; McCloskey D; Wilner G (2009). Alarming levels of carboxyhemoglobin in banked
blood. J Cardiothorac Vase Anesth, 23: 336-338 . 194089
Engel LA; Wood LDH; Utz G; Macklem PT. (1973). Gas mixing during inspiration. J Appl Physiol, 35: 18-24. 014336
Engel RR; Rodkey FL; O'Neal JD; CollisonHA. (1969). Relative affinity of human fetal hemoglobin for carbon monoxide
and oxygen. Blood, 33: 37-45. 193914
FennWO. (1970). The burning of CO in tissues. Ann N YAcad Sci, 174: 64-71. 010821
Forbes WH; Sargent F; Roughton FJW (1945). The rate of carbon monoxide uptake by normal men. Am J Physiol, 143:
594-608. 012850
Forster RE. (1966). Diffusion of gases in the lungs. Physiologist, 9: 110-122. 180430
September 2009 4-36 DRAFT - DO NOT CITE OR QUOTE
-------
Gibson QH; Olson JS; McKinnie RE; Rohlfs RJ. (1986). Akinetic description of ligand binding to sperm whale
myoglobin. J Biol Chem, 261: 10228-10239. 016289
Gibson QH; Roughton FJ. (1955). The kinetics of dissociation of the first oxygen molecule from fully saturated
oxyhaemoglobin in sheep blood solutions. Proc Biol Sci, 912: 143. 193941
Gosselin NH; Brunei RC; Carrier G (2009). Determination of carboxyhaemoglobin in humans following low-level
exposures to carbon monoxide. Inhal Toxicol, 0: 00-00. 190946
Guyatt AR; Holmes MA; Gumming G. (1981). Can carbon monoxide be absorbed from the upper respiratory tract in man?.
Eur Respir J, 62: 383-390. 011196
HaabP. (1990). The effect of carbon monoxide on respiration. Cell Mol Life Sci, 46: 1202-1206. 013359
Haldane J. (1895). The action of carbonic oxide on man. J Physiol, 18: 430-462. 010538
Hampson NB; Weaver LK. (2007). Carbon monoxide poisoning: a new incidence for an old disease. Undersea Hyperb
Med, 34: 163-8. 190272
Hanel B; Teunissen I; Rabol A; Warberg J; Secher NH. (1997). Restricted postexercise pulmonary diffusion capacity and
central blood volume depletion. J Appl Physiol, 83: 11-17. 193918
Hart CL; Smith GD; Hole DJ; Hawthorne VM. (2006). Carboxyhaemoglobin concentration, smoking habit, and mortality
in 25 years in the Renfrew/Paisley prospective cohort study. Heart, 92: 321-324. 194092
Hegewald MJ; Crapo RO. (2007). Socioeconomic status and lung function. Chest, 132: 1608-1614. 193923
Hill EP; Hill JR; Power GG; Longo LD. (1977). Carbon monoxide exchanges between the human fetus and mother: A
mathematical model. Am J Physiol, 232: H311-H323. 011315
Holland RA. (1970). Reaction rates of carbon monoxide and hemoglobin. Ann N Y Acad Sci, 174: 154-171. 193856
Horvath I; Donnelly LE; Kiss A; Paredi P; Kharitonov SA; Barnes PJ. (1998). Raised levels of exhaled carbon monoxide
are associated with an increased expression of heme oxygenase-1 in airway macrophages in asthma: a new marker
of oxidative stress. Thorax, 53: 668-672. 087190
Horvath I; Loukides S; Wodehouse T; Kharitonov SA; Cole PJ; Barnes PJ. (1998). Increased levels of exhaled carbon
monoxide in bronchiectasis: a new marker of oxidative stress. Thorax, 53: 867-870. 087191
Horvath SM; Agnew JW; Wagner JA; Bedi JR (1988). Maximal aerobic capacity at several ambient concentrations of
carbon monoxide at several altitudes. 012725
Hsia CC. (2002). Recruitment of lung diffusing capacity: Update of concept and application. Chest, 122: 1774-1783.
193857
Jarvis MJ; Belcher M; Vesey C; Hutchison DCS. (1986). Low cost carbon monoxide monitors in smoking assessment.
Thorax, 41: 886-887. 012043
Jarvis MJ; Russell MAH; Saloojee Y. (1980). Expired air carbon monoxide: a simple breath test of tobacco smoke intake.
Br Med J, 281: 484-485. 011813
Johnson FK; Johnson RA; Durante W; Jackson KE; Stevenson BK; Peyton KJ. (2006). Metabolic syndrome increases
endogenous carbon monoxide production to promote hypertension and endothelial dysfunction in obese Zucker
rats. Am J Physiol Regul Integr Comp Physiol, 290: 601-608. 193874
JoumardR; Chiron M; Vidon R; Maurin M; Rouzioux J-M. (1981). Mathematical models of the uptake of carbon
monoxide on hemoglobin at low carbon monoxide levels. Environ Health Perspect, 41: 277-289. 011330
Klasner AE; Smith SR; Thompson MW; Scalzo AJ. (1998). Carbon monoxide mass exposure in a pediatric population.
Acad Emerg Med, 5: 992-996. 087196
KleinmanMT; Leaf DA; Kelly E; Caiozzo V; OsannK; O'Niell T. (1998). Urban angina in the mountains: effects of carbon
monoxide and mild hypoxemia on subjects with chronic stable angina. Arch Environ Occup Health, 53: 388-397.
047186
Klimisch HJ; Chevalier HJ; Harke HP; Dontenwill W. (1975). Uptake of carbon monoxide in blood of miniature pigs and
other mammals. Toxicology, 3: 301-310. 010762
September 2009 4-37 DRAFT - DO NOT CITE OR QUOTE
-------
Lamberto C; Nunes H; Le Toumelin P; Duperron F; Valeyre D; Clerici C. (2004). Membrane and capillary blood
components of diffusion capacity of the lung for carbon monoxide in pulmonary sarcoidosis: relation to exercise
gas exchange. Chest, 125: 2061-2068. 193845
Landaw SA. (1973). The effects of cigarette smoking on total body burden and excretion rates of carbon monoxide. J
Occup Environ Med, 15: 231-235. 010803
Landaw SA; Callahan EW Jr; Schmid R. (1970). Catabolism of heme in vivo: comparison of the simultaneous production
of bilirubin and carbon monoxide. J Clin Invest, 49: 914-925. 012605
Lee PJ; Jiang B-H; Chin BY; Iyer NV; Alam J; Semenza GL; Choi AMK. (1997). Hypoxia-inducible factor-1 mediates
transcriptional activation of the heme oxygenase-1 gene in response to hypoxia. J Biol Chem, 272: 5375-5381.
082641
LevasseurL; Galliot-Guilley M; Richter F; Scherrmann JM; Baud FJ. (1996). 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. 080895
Levesque B; Allaire S; Prud'homme H; Dupuis K; Bellemare D. (2000). Air quality monitoring during indoor monster
truck and car demolition shows. J Expo Sci Environ Epidemiol, 10: 58-65. 011886
Longo LD. (1970). Carbon monoxide in the pregnant mother and fetus and its exchange across the placenta. Ann N Y Acad
Sci. 174:312-341.013922
Longo LD. (1977). The biological effects of carbon monoxide on the pregnant woman, fetus, and newborn infant. Am J
Obstet Gynecol, 129: 69-103. 012599
Longo LD; Hill EP (1977). Carbon monoxide uptake and elimination in fetal and maternal sheep. Am J Physiol, 232:
H324-H330. 010802
LundhB; Cavallin-Stahl E; Mercke C. (1975). Nicotinic acid and the endogenous production of carbon monoxide. J Intern
Med, 197: 173-176. 086332
Luomanmaki K; Coburn RF. (1969). Effects of metabolism and distribution of carbon monoxide on blood and body stores.
Am J Physiol, 217: 354-363. 012319
Mahoney JJ; Vreman HJ; Stevenson DK; Van Kessel AL. (1993). Measurement of carboxyhemoglobin and total
hemoglobin by five specialized spectrophotometers (CO-oximeters) in comparison with reference methods. Clin
Chem, 39: 1693-1700. 013859
Maines MD; Kappas A. (1974). Cobalt induction of hepatic heme oxygenase; with evidence that cytochrome P-450 is not
essential for this enzyme activity. Proc Natl Acad Sci U S A, 71: 4293-4297. 193976
Maines MD; Trakshel GM; Kutty RK. (1986). Characterization of two constitutive forms of rat liver microsomal heme
oxygenase. Only one molecular species of the enzyme is inducible. J Biol Chem, 261: 411-419. 193978
Manier G; Moinard J; Techoueyres P; Varene N; Guenard H. (1991). Pulmonary diffusion limitation after prolonged
strenuous exercise. Respir Physiol, 83: 143-153. 193979
Manno M; Rugge M; Cocheo V. (1992). Double fatal inhalation of dichloromethane. Hum Exp Toxicol, 11: 540-545.
013707
Marks GS; Vreman HJ; McLaughlin BE; Brien JF; Nakatsu K. (2002). Measurement of endogenous carbon monoxide
formation in biological systems. Antioxid Redox Signal, 4: 271-277. 030616
Marthan R; Castaing Y; Manier G; Guenard H. (1985). Gas exchange alterations in patients with chronic obstructive lung
disease. Chest, 87: 470-475. 086334
Marvisi M; Bassi E; Bonassi R; Civardi G; Delsignore R. (2007). DLCO correlates with intestinal inflammation in
ulcerative colitis, but albuminuria does not. Minerva Gastroenterol Dietol, 53: 321-7. 186702
Marvisi M; Borrello PD; Brianti M; Fornarsari G; Marani G; Guariglia A. (2000). Changes in the carbon monoxide
diffusing capacity of the lung in ulcerative colitis. Eur Respir J, 16: 965-8. 186703
McCartney ML. (1990). Sensitivity analysis applied to Coburn-Forster-Kane models of carboxyhemoglobin formation.
AIHA J, 51: 169-177. 013162
September 2009 4-38 DRAFT - DO NOT CITE OR QUOTE
-------
McClean PA; Duguid NJ; Griffin PM; Newth CJL; Zamel N. (1981). Changes in exhaled pulmonary diffusing capacity at
rest and exercise in individuals with impaired positional diffusion. EurRespir J, 17: 179-186. 012411
McCoubrey WK Jr; Huang TJ; Maines MD. (1997). Isolation and characterization of a cDNA from the rat brain that
encodes hemoprotein heme oxygenase-3. FEES J, 247: 725-732. 016715
McGrath JJ. (1992). Effects of altitude on endogenous carboxyhemoglobin levels. J Toxicol Environ Health, 35: 127-133.
013528
McGrath JJ; Schreck RM; Lee PS. (1993). Carboxyhemoglobin levels in humans: Effects of altitude. Inhal Toxicol, 5: 241-
249. 013865
Mercke C; Cavallin-Stahl E; Lundh B. (1975). Heme catabolism during short-term treatment with phenobarbital, diazepam
and oxazepam. Acta Med Scand, 198: 149-154. 086303
Mercke C; Cavallin-Stahl E; Lundh B. (1975). Carbon monoxide production and reticulocyte count in normal women:
effect of contraceptive drugs and smoking. Acta Med Scand, 198: 155-160. 086308
Mercke C; Lundh B. (1976). Erythrocyte filterability and heme catabolism during the menstrual cycle. Ann Intern Med, 85:
322-324. 086309
Meyer J; Prien T; Van Aken H; Bone H-G; Waurick R; Theilmeier G; Booke M. (1998). Arterio-venous
Carboxyhemoglobin difference suggests carbon monoxide production by human lungs. Biochem Biophys Res
Commun, 244: 230-232. 047530
Monma M; Yamaya M; Sekizawa K; Ikeda K; Suzuki N; Kikuchi T; Takasaka T; Sasaki H. (1999). Increased carbon
monoxide in exhaled air of patients with seasonal allergic rhinitis. Clin Exp Allergy, 29: 1537-1541. 180426
Morimatsu H; Takahashi T; Maeshima K; Inoue K; Kawakami T; Shimizu H; Takeuchi M; Yokoyama M; Katayama H;
Morita K. (2006). Increased heme catabolism in critically ill patients: Correlation among exhaled carbon monoxide,
arterial Carboxyhemoglobin, and serum bilirubin IXalpha concentrations. Am J Physiol Lung Cell Mol Physiol,
290: L114-L119.194097
Muchova L; Wong RJ; Hsu M; Morioka I; Vitek L; Zelenka J; Schroder H; Stevenson DK. (2007). Statin treatment
increases formation of carbon monoxide and bilirubin in mice: A novel mechanism of in vivo antioxidant
protection. Can J Physiol Pharmacol, 85: 800-810. 194098
Neas LM; Schwartz J. (1996). The determinants of pulmonary diffusing capacity in a national sample of US adults. Am J
Respir Grit Care Med, 153: 656-664. 079363
Neto CA; Yanagihara JI; Turri F. (2008). A carbon monoxide transport model of the human respiratory system applied to
urban atmosphere exposure analysis. J Braz Chem Soc, 30: 253-260. 194672
Ocak A; Valentour JC; Blanke RV. (1985). The effects of storage conditions on the stability of carbon monoxide in
postmortem blood. J Anal Toxicol, 9: 202-206. 011641
Ott W; Switzer P; Willits N. (1994). Carbon monoxide exposures inside an automobile traveling on an urban arterial
highway. J Air Waste Manag Assoc, 44: 1010-1018. 076546
Ott WR; Mage DT. (1978). Interpreting urban carbon monoxide concentrations by means of a computerized blood COHb
model. J Air Waste Manag Assoc, 28: 911-916. 011124
PankowD; MatschinerF; WeigmannH-J. (1991). Influence of aromatic hydrocarbons on the metabolism of
dichloromethane to carbon monoxide in rats. Toxicology, 68: 89-100. 013551
Paredi P; Biernacki W; Invernizzi G; Kharitonov SA; Barnes PJ. (1999). Exhaled carbon monoxide levels elevated in
diabetes and correlated with glucose concentration in blood: A new test for monitoring the disease?. Chest, 116:
1007-1011. 194102
Paredi P; Shah PL; Montuschi P; Sullivan P; Hodson ME; Kharitonov SA; Barnes PJ. (1999). Increased carbon monoxide
in exhaled air of patients with cystic fibrosis. Thorax, 54: 917-20. 118798
Pelham TW; Holt LE; Moss MA. (2002). Exposure to carbon monoxide and nitrogen dioxide in enclosed ice arenas. Occup
Environ Med, 59: 224-233. 025716
Penney DG (1988). Areview: Hemodynamic response to carbon monoxide. Environ Health Perspect, 77: 121-130. 012519
September 2009 4-39 DRAFT - DO NOT CITE OR QUOTE
-------
Pesola GR; Huggins G; Sherpa TY. (2006). Abnormal predicted diffusion capacities in healthy Asians: an inequality with a
solution. Respiration, 73: 799-807. 193855
Pesola GR; Sunmonu Y; Huggins G; Ford JG (2004). Measured diffusion capacity versus prediction equation estimates in
blacks without lung disease. Respiration, 71: 484-492. 193842
Peterson JE; Stewart RD. (1970). Absorption and elimination of carbon monoxide by inactive young men. Arch Environ
Occup Health, 21: 165-171. 012416
Peterson JE; Stewart RD. (1975). Predicting the carboxyhemoglobin levels resulting from carbon monoxide exposures. J
ApplPhysiol, 39: 633-638. 010696
Prommer N; Schmidt. (2007). Loss of CO from the intravascular bed and its impact on the optimised CO-rebreathing
method. Eur J Appl Physiol, 100: 383-391. 180421
Proudman SM; Stevens WM; Sahhar J; Celermajer D. (2007). Pulmonary arterial hypertension in systemic sclerosis: the
need for early detection and treatment. J Intern Med, 37: 485-94. 186705
Reeves RB; Park HK. (1992). CO uptake kinetics of red cells and CO diffusing capacity. Respir Physiol, 88: 1-21. 193847
Richardson RS; Noyszewski EA; Saltin B; Gonzalez-Alonso J; et al. (2002). Effect of mild carboxy-hemoglobin on
exercising skeletal muscle: Intravascular and intracellular evidence. Am J Physiol, 283: R1131-R1139. 037513
Riediker M; Williams R; Devlin R; Griggs T; Bromberg P. (2003). Exposure to particulate matter, volatile organic
compounds, and other air pollutants inside patrol cars. Environ Sci Technol, 37: 2084-2093. 043761
Rodgers PA; Vreman HJ; Dennery PA; Stevenson DK. (1994). Sources of carbon monoxide (CO) in biological systems and
applications of CO detection technologies. Semin Perinatol, 18: 2-10. 076440
Rodkey FL; O'Neal JD; Collison HA. (1969). Oxygen and carbon monoxide equilibria of human adult hemoglobin at
atmospheric and elevated pressure. Blood, 33: 57-65. 008151
Roughton FJ; Forster RE. (1957). Relative importance of diffusion and chemical reaction rates in determining rate of
exchnage of gases in the human lung, with special reference to true diffusing capacity of pulmonary membrane and
volume of blood in the lung capillaries. J Appl Physiol, 11: 290-302. 193862
Roughton FJW. (1970). The equilibrium of carbon monoxide with human hemoglobin in whole blood. Ann N Y Acad Sci,
174: 177-188.013931
Roughton FJW; Root WS. (1945). The fate of CO in the body during recovery from mild carbon monoxide poisoning in
man. Am J Physiol, 145: 239-252. 180418
Sakamaki F; Oya H; Nagaya N; Kyotani S; Satoh T; Nakanishi N. (2002). Higher prevalence of obstructive airway disease
in patients with thoracic or abdominal aortic aneurysm. J Vase Surg, 36: 35-40. 186706
Schachter L. M; Dixon J; Pierce R. J; O'Brien P. (2003). Severe gastroesophageal reflux is associated with reduced carbon
monoxide diffusing capacity. Chest, 123: 1932-8. 186707
Scharte M; Bone HG; Van Aken H; Meyer J. (2000). Increased carbon monoxide in exhaled air of critically ill patients.
Biochem Biophys Res Commun, 267: 423-426. 194112
Scharte M; von Ostrowski TA; Daudel F; Freise H; Van Aken H; Bone HG. (2006). Endogenous carbon monoxide
production correlates weakly with severity of acute illness. Eur J Anaesthesiol, 23: 117-122. 194115
Schoenfisch WH; Hoop KA; Struelens BS. (1980). Carbon monoxide absorption through the oral and nasal mucosae of
cynomolgus monkeys. Arch Environ Health, 35: 152-154. 011404
Selvakumar S; Sharan M; Singh MP. (1992). Mathematical model for the exchange of gases in the lungs with special
reference to carbon monoxide. Med Biol Eng Comput, 30: 525-532. 013750
Sharan M; Selvakumar S. (1999). A mathematical model for the simultaneous transport of gases to compute blood
carboxyhaemoglobin build-up due to CO exposures: application to the end-expired breath technique. Environ
Pollut, 105: 231-242. 194673
Sharan M; Selvakumar S; Singh MP. (1990). Mathematical model for the computation of alveolar partial pressure of carbon
monoxide. Int J Bio Med Comput, 26: 135-147. 003798
Sharma VS, Schmidt MR; Ranney HM. (1976). Dissociation of CO from carboxyhemoglobin. J Biol Chem, 251: 4267-
4272. 193766
September 2009 4-40 DRAFT - DO NOT CITE OR QUOTE
-------
ShimazuT. (2001). Half-life of blood carboxyhemoglobin. Chest, 119: 661-663. 016331
Shimazu T; Ikeuchi H; Sugimoto H; Goodwin CW; Mason AD Jr; Pruitt BA Jr. (2000). Half-life of blood
carboxyhemoglobin after short-term and long-term exposure to carbon monoxide. J Trauma, 49: 126-131. 016420
Singh MP; Sharan M; Selvakumar S. (1991). A mathematical model for the computation of carboxyhaemoglobin in human
blood as a function of exposure time. Proc Biol Sci, 334: 135-147. 013583
Smith MV. (1990). Comparing solutions to the linear and nonlinear CFK equations for predicting COHb formation. Math
Biosci, 99: 251-263. 013164
Smith MV; Hazucha MJ; Benignus VA; Bromberg PA. (1994). Effect of regional circulation patterns on observed HbCO
levels. J Appl Physiol, 77: 1659-1665. 076564
Sokal JA; Majka J; Palus J. (1984). The content of carbon monoxide in the tissues of rats intoxicated with carbon monoxide
in various conditions of acute exposure. ArchToxicol, 56: 106-108. 011591
Solanki DL; McCurdy PR; Cuttitta FF; Schechter GP (1988). Hemolysis in sickle cell disease as measured by endogenous
carbon monoxide production: A preliminary report. Am J Clin Pathol, 89: 221-225. 012426
Stevenson DK; Ostrander CE; Johnson JD. (1979). Effect of erythrocyte destruction on the pulmonary excretion rate of
carbon monoxide in adult male Wistar rats. J Lab Clin Med, 94: 649-654. 193767
Stewart RD; Peterson JE; Baretta ED; Bachand RT; Hosko MJ; Herrmann AA. (1970). Experimental human exposure to
carbon monoxide. Arch Environ Health, 21: 154-164. 013972
Stewart RD; Peterson JE; Fisher TN; Hosko MJ; Baretta ED; Dodd HC; Herrmann AA. (1973). Experimental human
exposure to high concentrations of carbon monoxide. Arch Environ Occup Health, 26: 1-7. 012428
Sylvester KP; Patey RA; Rafferty GF; Rees D; Thein SL; Greenough A. (2005). Exhaled carbon monoxide levels in
children with sickle cell disease. Eur J Pediatr, 164: 162-165. 191954
Takeuchi A; Vesely A; Rucker J; Sommer LZ; Tesler J; Lavine E; Slutsky AS; Maleck WA; Volgyesi G; Fedorko L; Iscoe
S; Fisher JA. (2000). A simple "new" method to accelerate clearance of carbon monoxide. Am J Respir Crit Care
Med, 161: 1816-1819. 005675
Tarquini R; Masini E; La Villa G; Barletta G; Novelli M; Mastroianni R; Romanelli RG; Vizzutti F; Santosuosso U; Laffi
G. (2009). Increased plasma carbon monoxide in patients with viral cirrhosis and hyperdynamic circulation. Am J
Gastroenterol, 104: 891-897. 194117
Terzano C; Conti V; Petroianni A; Ceccarelli D; De Vito C; Villari P. (2009). Effect of Postural Variations on Carbon
Monoxide Diffusing Capacity in Healthy Subjects and Patients with Chronic Obstructive Pulmonary Disease.
Respiration, 77: 51-57. 108046
Tikuisis P. (1996). Modeling the uptake and elimination of carbon monoxide. 080960
Tikuisis P; Buick F; Kane DM. (1987). Percent carboxyhemoglobin in resting humans exposed repeatedly to 1,500 and
7,500 ppm CO. J Appl Physiol, 63: 820-827. 012219
Tikuisis P; Kane DM; McLellan TM; Buick F; Fairburn SM. (1992). Rate of formation of carboxyhemoglobin in exercising
humans exposed to carbon monoxide. J Appl Physiol, 72: 1311-1319. 013592
Tikuisis P; Madill HD; Gill BJ; Lewis WF; Cox KM; Kane DM. (1987). A critical analysis of the use of the CFK equation
in predicting COHb formation. Am Ind Hyg Assoc J, 48: 208-213. 012138
Topping DL. (1975). Acute effects of carbon monoxide on the metabolism of perfused rat liver. Biochem J, 152: 425-427.
193784
Tran DN; Ota LC; Jacobson JD; Patton WC; Chan PJ. (2007). Influence of nanoparticles on morphological differentiation
of mouse embryonic stem cells. Fertil Steril, 87: 965-970. 090752
Tyuma I; Ueda Y; Imaizumi K; Kosaka H. (1981). Prediction of the carbonmonoxyhemoglobin levels during and after
carbon monoxide exposures in various animal species. J Physiol Sci, 31: 131-143. 011226
U.S. Census Bureau. (2008). American Housing Survey for the United States: 2007. U.S. Government Printing Office.
Washington, DC. 194013
U.S. EPA. (1978). Altitude as a factor in air pollution. 086321
September 2009 4-41 DRAFT - DO NOT CITE OR QUOTE
-------
U.S. EPA. (1991). Air quality criteria for carbon monoxide. 017643
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 600/P-
99/001F. 000907
Vreman HJ; Wong RJ; Kadotani T; Stevenson DK. (2005). Determination of carbon monoxide (CO) in rodent tissue: effect
of heme administration and environmental CO exposure. Anal Biochem, 341: 280-289. 193786
Vreman HJ; Wong RJ; Stevenson DK; Smialek JE; Fowler DR; Li L; Vigorito RD; Zielke HR. (2006). Concentration of
Carbon Monoxide (CO) in Postmortem Human Tissues: Effect of Environmental CO Exposure. J Forensic Sci, 51:
1182-1190.098272
Wagner JA; Horvath SM; Dahms TE. (1975). Carbon monoxide elimination. Respir Physiol Neurobiol, 23: 41-47. 010989
Weaver LK; Howe S; Hopkins R; Chan K. (2000). Carboxyhemoglobin half-life in carbon monoxide-poisoned patients
treated with 100% oxygen at atmospheric pressure. Chest, 117: 801-808. 016421
Weller PA; Price M; Isenberg H; Edwards YH; Jeffreys AJ. (1986). Myoglobin expression: early induction and subsequent
modulation of myoglobin and myoglobin mRNA during myogenesis. Mol Cell Biol, 6: 4539-4547. 187298
Widdop B. (2002). Analysis of carbon monoxide. Ann Clin Biochem, 39: 378-391. 030493
Wirkner K; Damme B; Poelchen W; Pankow D. (1997). Effect of long-term ethanol pretreatment on the metabolism of
dichloromethane to carbon monoxide in rats. Toxicol Appl Pharmacol, 143: 83-88. 082642
Wittenberg BA; Wittenberg JB; Caldwell PRB. (1975). Role of myoglobin in the oxygen supply to red skeletal muscle. J
Biol Chem, 250: 9038-9043. 012436
WuL; Wang R. (2005). Carbon monoxide: endogenous production, physiological functions, and pharmacological
applications. Pharmacol Rev, 57: 585-630. 180411
Yamaya M; Sekizawa K; Ishizuka S; Monma M; Mizuta K; Sasaki H. (1998). Increased carbon monoxide in exhaled air of
subjects with upper respiratory tract infections. Am J Respir Crit Care Med, 158: 311-314. 047525
Yasuda H; Sasaki T; Yamaya M; Ebihara S; Maruyama M; Kanda A; Sasaki H. (2004). Increased arteriovenous
Carboxyhemoglobin differences in patients with inflammatory pulmonary diseases. Chest, 125: 2160-2168. 191955
Yasuda H; Yamaya M; Nakayama K; Ebihara S; Sasaki T; Okinaga S; Inoue D; Asada M; Nemoto M; Sasaki H. (2005).
Increased arterial Carboxyhemoglobin concentrations in chronic obstructive pulmonary disease. Am J Respir Crit
Care Med, 171: 1246-1251. 191953
Yasuda H; Yamaya M; Yanai M; Ohrui T; Sasaki H. (2002). Increased blood carboxyhaemoglobin concentrations in
inflammatory pulmonary diseases. Thorax, 57: 779-783. 035206
Yerushalmi R; Kramer MR; Rizel S; Sulkes A; Gelmon K; Granot T; Neiman V; Stemmer SM. (2009). Decline in
pulmonary function in patients with breast cancer receiving dose-dense chemotherapy: a prospective study. Ann
Oncol, 20: 437-440. 186711
Young LJ; Caughey WS. (1986). Mitochondrial oxygenation of carbon monoxide. Biochemistry, 239: 225-227. 012091
Yu S; Zhang R; Chen Y; Chen J; Zhu C; Tang J; Li G. (2008). Possible contribution of endogenous carbon monoxide to the
development of allergic rhinitis in guinea pigs. J Inflamm, 5: -. 192384
Zayasu K; Sekizawa K; Okinaga S; Yamaya M; Ohrui T; Sasaki H. (1997). Increased carbon monoxide in exhaled air of
asthmatic patients. Am J Respir Crit Care Med, 156: 1140-1143. 084291
Zegdi R; Perrin D; Burdin M; Boiteau R; Tenaillon A. (2002). Increased endogenous carbon monoxide production in
severe sepsis. Intensive Care Med, 28: 793-796. 037461
Zinkham WH; Houtchens RA; Caughey WS. (1980). Carboxyhemoglobin levels in an unstable hemoglobin disorder (Hb
Zurich): effect onphenotypic expression. Science, 209: 406-408. 011435
September 2009 4-42 DRAFT - DO NOT CITE OR QUOTE
-------
Chapters. Integrated Health Effects
5.1. Mode of Action of CO Toxicity
5.1.1.Introduction
1 The diverse effects of CO are dependent upon concentration and duration of exposure as well
2 as on the cell types and tissues involved. Responses to CO are not necessarily due to a single process
3 and may instead be mediated by a combination of effects including COHb-mediated hypoxic stress
4 and other mechanisms such as free radical production and the initiation of cell signaling. However,
5 binding of CO to reduced iron in heme proteins with subsequent alteration of heme protein function
6 is the common mechanism underlying the biological responses to CO.
5.1.2.Hypoxic Mechanisms
7 As discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907). the most well-known
8 pathophysiologic effect of CO is tissue hypoxia caused by binding of CO to Hb. Not only does the
9 formation of COHb reduce the O2-carrying capacity of blood, but it also impairs the release of O2
10 from O2Hb. Compensatory alterations in hemodynamics, such as vasodilation and increased cardiac
11 output, protect against tissue hypoxia. Depending on the extent of CO exposure, these compensatory
12 changes may be effective in people with a healthy cardiovascular system. However, hemodynamic
13 responses following CO exposure may be insufficient in people with decrements in cardiovascular
14 function, resulting in health effects as described in Section 5.2.
15 The 2000 CO AQCD (U.S. EPA, 2000, 000907) reported changes in vasodilation due to CO
16 levels between 500-2,000 ppm (Kanten et al., 1983, 011333: MacMillan, 1975, 012909). In one
17 study, the vasodilatory response to CO in cerebral blood vessels was attributed to decreased O2
18 availability (Koehler et al., 1982, 011341). In another study, exposure of rats to 1000 ppm CO
19 resulted in increased cerebral blood flow which was not triggered by tissue hypoxia since no changes
20 in intramitochondrial NADH levels preceded vasodilation (Meilin et al., 1996, 079919). However,
21 the response was blocked by the inhibition of nitric oxide synthase (NOS) indicating a role for the
22 free radical species nitric oxide (NO) in CO-mediated vasodilation (Meilin et al., 1996, 079919).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 5-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 Increased cardiac ouput was also discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907)
2 as a compensatory response to CO-mediated tissue hypoxia. Findings of studies which measured
3 hemodynamic alterations following CO exposure were equivocal and sometimes contradictory
4 (Penney, 1988, 012519). While most studies reported a positive correlation between COHb and
5 cardiac output at COHb levels above 20%, one study demonstrated increased cardiac output in
6 humans following acute exposure to 5% CO which resulted in the rapid rise in COHb levels to
7 about 9% (Ayres et al., 1973, 193943). However, there was no increase in cardiac output following a
8 more gradual increase in COHb levels to about 9% achieved by exposure to 0.1% CO over a longer
9 period of time (Ayres et al., 1973, 193943). Increased heart rate and stroke volume (SV) were
10 observed in response to CO exposure in one study (Stewart et al., 1973, 012428): however, some
11 experiments found no change in SV in humans with 18-20% COHb (Vogel and Gleser, 1972,
12 010898) or 12.5% COHb (Klausen et al., 1968, 193936). The 2000 CO AQCD (U.S. EPA, 2000,
13 000907) reported that blood pressure was generally unchanged in human CO exposure studies, while
14 a number of animal studies demonstrated CO-induced hypotension (Penney, 1988, 012519). No
15 changes in forearm blood flow, blood pressure, or heart rate were reported in humans with
16 approximately 8% COHb (Hausberg and Somers, 1997, 083450). However, high concentration
17 animal exposures (3,000-10,000 ppm) showed diminished organ blood flow (Brown and Piantadosi,
18 1992, 013441). In depth discussion of hemodynamic changes resulting from CO exposure in recent
19 human clinical studies can be found in Section 5.2.2.
20 Binding of CO to Mb, as discussed in the 2000 CO AQCD (U. S. EPA, 2000, 000907) and in
21 Section 4.3.2.3, can also impair the delivery of O2 to tissues. Mb has a high affinity for CO, about 25
22 times that of O2; however, pathophysiologic effects are seen only after high dose exposures to CO,
23 resulting in COMb concentrations far above baseline levels. High energy phosphate production in
24 cardiac myocytes was inhibited when COMb concentrations exceeded 40%, corresponding to an
25 estimated COHb level between 20-40% (Wittenberg and Wittenberg, 1993, 013909). Conversely, rat
26 hearts perfused with solutions containing 6% CO (60,000 ppm) exhibited no change in high energy
27 phosphate production, respiration rate, or contractile function (Chung et al., 2006, 193987; Glabe et
28 al.. 1998. 086704).
5.1.3.Non-Hypoxic Mechanisms
29 Non-hypoxic mechanisms underlying the biological effects of CO were discussed in the 2000
30 CO AQCD (U.S. EPA, 2000, 000907) and are summarized below. Most of these mechanisms are
31 related to CO's ability to bind heme-containing proteins other than Hb and Mb (Raub and Benignus,
32 2002, 041616). Since then, additional experiments have confirmed and extended these findings.
33 While the majority of the older studies utilized concentrations of CO far higher than ambient levels,
34 many of the newer studies have employed more environmentally-relevant concentrations of CO.
September 2009 5-2 DRAFT - DO NOT CITE OR QUOTE
-------
5.1.3.1. Non-Hypoxic Mechanisms Reviewed in the 2000 CO AQCD
1 Inhibition of heme-containing proteins such as cytochrome c oxidase and cytochrome P450
2 reductases may alter cellular function. CO interacts with the ferrous heme a3 of the terminal enzyme
3 of the electron transport chain, cytochrome c oxidase (Petersen, 1977, 193764). Cytochrome c
4 oxidase inhibition not only interrupts cellular respiration and energy production, but can also
5 enhance reactive oxygen species (ROS) production. In vivo studies observed CO binding to
6 cytochrome c oxidase under conditions where COHb concentrations were above 50% (Brown and
7 Piantadosi, 1992, 013441). It is unlikely that this could arise under physiologic conditions or under
8 conditions relevant to ambient exposures.
9 A series of studies from the laboratory of Thorn, Ischiropoulos and colleagues indicated that
10 CO exposure produced a pro-oxidant cellular environment by liberation of NO. Exposure to CO
11 concentrations of 10-20 ppm and above caused isolated rat platelets, as well as cultured bovine
12 pulmonary endothelial cells, to release NO (Thorn and Ischiropoulos, 1997, 085644). This response
13 was blocked by treatment with a NOS inhibitor indicating that the NO released was dependent on
14 NOS activity. An increase in available NO was also seen in the lung and brain of CO-exposed rats
15 (Ischiropoulos et al, 1996, 079491; Thorn et al, 1999, 016757). Reaction of NO with superoxide to
16 form the highly active oxidant species, peroxynitrite (Thorn et al., 1997, 084337). was thought to
17 lead to the activation and sequestration of leukocytes in brain vessels (Thorn et al., 2001, 193779)
18 and aorta, oxidation of plasma lipoproteins (Thorn et al., 1999, 016753). and the formation of protein
19 nitrotyrosine (Ischiropoulos et al., 1996, 079491; Thorn et al., 1999, 016757; Thorn et al., 1999,
20 016753). NO release by CO was attributed to the displacement of NO from nitrosyl-bound heme
21 proteins. The rate of this event was slow; however it occurred at environmentally-relevant
22 concentrations of CO (Thorn et al., 1997, 084337).
23 CO exposure also increased the production of other pro-oxidant species, including hydrogen
24 peroxide (H2O2) and hydroxyl radical (OH'). High level CO exposure (2,500 ppm) increased OH* in
25 rat brain and this response was distinct from tissue hypoxia (Piantadosi et al., 1997, 081326). The
26 mechanism for enhanced H2O2 production was unclear. The release of H2O2 in the lung of
27 CO-exposed rats was dependent upon the production of NO, as it was inhibited by the pretreatment
28 with a NOS inhibitor (Thorn et al., 1999, 016757). It is possible that peroxynitrite formed after CO
29 exposure inhibited electron transport at complexes I through III, or that cytochrome c oxidase
30 inhibition led to mitochondrial dysfunction and ROS production.
31 Evidence was presented for CO-mediated vasorelaxation by three different mechanisms. First,
32 CO may inhibit the synthesis of vasoconstrictors by P450 heme proteins (Wang, 1998, 086074).
33 Vasodilation in isolated vessels was demonstrated via this P450-dependent mechanism using high
34 concentrations of CO (approximately 90,000 ppm) (Coceani et al., 1988, 040493). In the case of
September 2009 5-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 cytochrome P450 enzymes, tissue CO levels may need to be abnormally high to elicit a response
2 since the Warburg binding coefficients (the ratio of CO to O2 at which half the reactive sites are
3 occupied by CO) for cytochrome P450s range from 0.1-12 (Piantadosi, 2002, 037463). P450
4 inhibition may reduce the hypoxia-induced expression of mitogens such as erythropoietin (EPO),
5 vascular endothelial growth factor (VEGF), endothelin-1 (ET-1), and platelet derived growth factor
6 (PDGF) which may decrease smooth muscle proliferation in response to hypoxia (Wang, 1998,
7 086074). CO also interfered with the metabolism of barbiturates and other drugs; however, this was
8 probably due to the hypoxic actions of CO rather than to P450 inhibition (Roth RA and Rubin, 1976,
9 012703: Roth RA and Rubin, 1976, 012420).
10 Secondly, CO has been shown to play a physiological role in vasomotor control and in signal
11 transduction by activation of soluble guanylate cyclase (sGC), causing a conversion of GTP to cyclic
12 GMP (cGMP). CO reversibly ligates the heme core of sGC and the resulting protoporphyrin IX
13 intermediate triggers cGMP production (Ndisang et al, 2004, 180425). CO caused vascular
14 relaxation, independent of the endothelium, in human arterial rings (Achouh et al., 2008, 179918).
15 rat tail artery (Wang et al., 1997, 084341). and rat thoracic aorta (Lin and McGrath, 1988, 012773).
16 but not in cerebral vessels (Andresen et al., 2006, 180449; Brian et al., 1994, 076283). Activation of
17 sGC by CO has been linked to neurotransmission, vasodilation, bronchodilation, inhibition of
18 platelet aggregation, and inhibiton of smooth muscle proliferation (Briine and Ullrich, 1987, 016535;
19 Cardell et al., 1998, 086700; Cardell et al., 1998, 011534; Morita et al., 1997, 085345; Verma et al.,
20 1993. 193999).
21 CO-mediated vasorelaxation can also be caused by activation of voltage- or Ca2+ -activated
22 potassium (K+) channels in smooth muscle cells, which leads to membrane hyperpolarization,
23 voltage-dependent Ca2+ channel closing, reduction of resting Ca2+ concentration and vascular tissue
24 relaxation (Farrugia et al., 1993, 013826; Wang et al., 1997, 084341). This effect may be linked to
25 sGC activity; however it has also been reported to occur independently (Dubuis et al., 2003, 180439;
26 Naik and Walker, 2003, 193852). Developmental stage and tissue type will determine whether K+
27 channels or the sGC/cGMP pathway plays more of a role in vasorelaxation (Ndisang et al., 2004,
28 180425).
29 Collectively, these older studies demonstrated that exposures to high concentrations of CO
30 resulted in altered functions of heme proteins other than Hb and Mb. Decreased cellular respiration
31 and energy production and increased ROS following cytochrome c oxidase inhibition would likely
32 predispose towards cellular injury and death. The release of NO from sequestered stores could
33 contribute to the pro-oxidant status if superoxide levels are simultaneously increased. Furthermore,
34 increased ROS and reactive nitrogen species are known to promote cell signaling events leading to
35 inflammation and endothelial dysfunction. An inappropriate increase in vasorelaxation due to
36 inhibition of vasoconstrictor production or to activation of vasodilatory pathways (sGC and ion
September 2009 5-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 channels) could potentially limit compensatory alterations in hemodynamics. Alternatively,
2 CO-binding to sGC could result in decreased vasorelaxation by interfering with the binding of NO to
3 sGC. NO can also activate sGC, and with a 30-fold greater affinity than CO is 1,000-fold more
4 potent with respect to vasodilation and sGC activation (Stone and Marietta, 1994, 076455). CO
5 could further contribute to endothelial dysfunction by this mechanism. Although the 2000 CO
6 AQCD (U.S. EPA, 2000, 000907) made no definitive links between these non-hypoxic mechanisms
7 of CO and CO-mediated health effects, it did document the potential for CO to interfere with basic
8 cellular and molecular processes that could lead to dysfunction and/or disease.
5.1.3.2. Recent Studies of Non-Hypoxic Mechanisms
9 Since the 2000 CO AQCD (U.S. EPA, 2000, 000907). new studies have provided additional
10 evidence for non-hypoxic mechanisms of CO which involve the binding of CO to reduced iron in
11 heme proteins. These mechanisms, which may be inter-related, are described below and include:
12 • Alteration in NO signaling
13 • Inhibition of cytochrome c oxidase
14 • Heme loss from protein
15 • Disruption of iron homeostasis
16 • Alteration in cellular redox status
17 Recent studies have also demonstrated non-hypoxic mechanisms of CO which are either
18 indirectly linked to heme protein interactions or not yet understood. These mechanisms are described
19 below and include:
20 • Alteration in ion channel activity
21 • Modulation of protein kinase pathways
22 This assessment evaluates these non-hypoxic mechanisms in terms of their potential to
23 contribute to health effects associated with environmentally-relevant CO exposures. As discussed
24 above, CO at high concentrations may promote oxidative stress, cell injury and death, inflammation
25 and endothelial dysfunction. Whether lower CO concentrations trigger these same processes is of
26 key interest since they may potentially contribute to adverse health effects following ambient
27 exposures.
28 In addition, a large number of studies published since the 2000 CO AQCD (U.S. EPA, 2000,
29 000907) has focused on the role of CO derived from HO-catalyzed heme metabolism as an
September 2009 5-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 endogenous signaling molecule and on the potential therapeutic effects of exogenous CO
2 administered at high concentrations. This assessment addresses aspects of these topics pertaining to
3 the evaluation of health effects associated with environmentally-relevant CO exposures.
Alteration in NO Signaling
4 Work by Thorup et al. (1999, 193782) demonstrated altered NO signaling in isolated rat renal
5 resistance arteries. In one set of experiments, rapid release of NO was observed in response to
6 exogenous CO. This response was biphasic, peaking at 100 nM CO in the perfusate and declining at
7 higher concentrations. It was also NOS-dependent as it required L-arginine and was blocked by a
8 NOS inhibitor. The authors attributed the effects of CO on NO release to either stimulated eNOS or
9 to displacement of preformed NO from intracellular binding sites. These findings are similar to those
10 of Thorn and colleagues (Ischiropoulos et al., 1996, 079491; Thorn and Ischiropoulos, 1997, 085644;
11 Thorn et al., 1994, 076459; Thorn et al., 1997, 084337; Thorn et al., 1999, 016753; Thorn et al.,
12 1999, 016757; Thorn et al., 2000, 011574; Thorn et al., 2006, 098418) who demonstrated NO
13 release, presumably from sequestered stores in platelets, endothelial cells, aorta and lung in response
14 to CO (see above). Furthermore in a second set of experiments, Thorup et al., (Thorup et al., 1999,
15 193782) demonstrated inhibition of agonist-stimulated NOS activity in isolated rat renal resistance
16 arteries. Here rapid NOS-dependent release of NO following carbachol stimulation was blocked by
17 pretreatment with 100 nM CO or by upregulation of intracellular HO-1. Additional experiments
18 using blood-perfused isolated juxtamedullary afferent arterioles demonstrated a biphasic response to
19 CO with rapid vasodilation observed at lower, but not higher, concentrations of CO. These same
20 higher concentrations of CO inhibited agonist-stimulated vasodilation in the arterioles. In order to
21 determine whether CO had a direct effect on the activity of NOS, which is a heme protein, purified
22 recombinant eNOS was exposed in vitro to CO in the presence of the necessary substrates and
23 cofactors. A dose-dependent inhibition of NOS by CO was observed, suggesting that CO-mediated
24 NO release in the isolated vessels was not due to stimulated NOS activity. The authors concluded
25 that CO effects on vascular tone were due to the liberation of NO from intracellular binding sites at
26 lower concentrations and to the inhibition of NOS at higher concentrations.
27 These findings illustrate the potential of CO to alter processes dependent on endogenous NO
28 either by enhancing intracellular concentrations of free NO (increased vasodilatory influence) or by
29 inhibiting agonist-induced NO production by NOS (decreased vasodilatory influence). In addition,
30 CO may compete with NO for binding to sGC as discussed above. Since NO activates sGC to a
31 greater extent than CO, NO-dependent vasodilation may be significantly impaired in the presence of
32 CO. In fact, a recent study in transgenic mice demonstrated that chronic overexpression of HO-1 in
33 vascular smooth muscle resulted in attenuated NO-mediated vasodilation and elevated blood
September 2009 5-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 pressure (Imai et al., 2001, 193864). Results of this study suggested that decreased sensitivity of sGC
2 to NO contributed to the changes in vascular function. The considerations mentioned above,
3 however, do not preclude an important role for CO in maintaining vasomotor tone in vessels where
4 CO and NO do not compete for available heme sites on sGC. This could occur when both mediators
5 are present at low concentrations compared with sGC or in situations where NOS does not co-
6 localize with sGC, as discussed by Thorup et al., (Thorup et al., 1999, 193782).
Inhibition of Cytochrome c Oxidase
7 High concentrations of CO are known to inhibit cytochrome c oxidase, the terminal enzyme in
8 the mitochondrial electron transport chain, resulting in inhibition of mitochondrial respiration and
9 the formation of superoxide from mitochondrial substrates. Several recent studies demonstrated
10 CO-mediated decreases in cytochrome c oxidase activity in model systems ranging from isolated
11 mitochondria to whole animals. In a study by Alonso et al. (2003, 193882). exposure of isolated
12 mitochondria from human skeletal muscle to 50-500 ppm CO for 5 min decreased cytochrome c
13 oxidase activity. Similarly, exposure of cultured macrophages to 250 ppm CO for 1 h inhibited
14 cytochrome c oxidase (Zuckerbraun et al., 2007, 193884). In this latter study, increased ROS were
15 observed following exposure to 250 ppm CO, as well as to CO concentrations as low as 50 ppm, for
16 1 h. Animal studies demonstrated that exposure of rats to 250 ppm CO for 90 min inhibited
17 cytochrome c oxidase activity in myocardial fibers (Favory et al., 2006, 184462). Exposure of mice
18 to 1,000 ppm CO for 3 h resulting in COHb levels of 61% decreased cytochrome c oxidase activity
19 in heart mitochondria (Iheagwara et al., 2007, 193861).
Heme Content Loss from Proteins
20 In addition to decreasing the activity of cytochrome c oxidase, exposure of mice to 1,000 ppm
21 CO for 3 h resulted in decreased protein levels and heme content of cytochrome c oxidase in heart
22 mitochondria (Iheagwara et al., 2007, 193861). CO-mediated heme release was also seen in a study
23 by Cronje et al., (Cronje et al., 2004, 180440). and was followed by increased endogenous CO
24 production through the activation of HO-2 and the induction of HO-1. Loss of heme from proteins
25 leads to loss of protein function and often to protein degradation.
Disruption of Iron Homeostasis
26 Exposure of rats to 50 ppm CO for 24 h increased levels of iron and ferritin in the
27 bronchoalveolar lavage fluid (BALF), decreased lung non-heme iron and increased liver non-heme
28 iron (Ghio et al., 2008, 096321). Furthermore in this same study, exposure of cultured human
29 respiratory epithelial cells to 10-100 ppm CO for 24 h caused a dose-dependent decrease in cellular
September 2009 5-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 non-heme iron and ferritin. Heme loss, which was observed in other studies (Cronje et al, 2004,
2 180440; Iheagwara et al., 2007, 193861). may also contribute to disruption of iron homeostasis. Iron
3 homeostasis is critical for the sequestration of free iron and the prevention of iron-mediated redox
4 cycling which leads to ROS generation and lipid peroxidation.
Alteration in Cellular Redox Status
5 Recent studies demonstrated that exposure to low, moderate and high levels of CO increased
6 cellular oxidative stress in cultured cells (Kim et al., 2008, 193961; Zuckerbraun et al., 2007,
7 193884). A dose-dependent increase in dichlorofluorescein (DCF) fluorescence (an indicator of
8 ROS) occurred following 1-h exposure to 50-500 ppm CO in macrophages and following 1-h
9 exposure to 250 ppm CO in hepatocytes. NOS inhibition had no effect on the increase in DCF
10 fluorescence in CO-treated macrophages indicating that the effects were not due to an interaction of
11 CO and NO (Zuckerbraun et al., 2007, 193884). Mitochondria were identified as the source of the
12 increased ROS since mitochondria-impaired cells (rho zero cells and treatment with antimycin A)
13 did not respond to CO with an increase in DCF fluorescence. Furthermore, 1-h exposure to 250 ppm
14 CO inhibited mitochondrial cytochrome c oxidase enzymatic activity in macrophages (Zuckerbraun
15 et al., 2007, 193884). Recently, inhibition of cytochrome c oxidase was demonstrated in HEK-293
16 cells transfected with HO-1 and in macrophages with induced HO-1, and this effect was attributed to
17 endogenously produced CO (D'Amico et al., 2006, 193992). In hepatocytes, exposure to 250 ppm
18 CO for 1 h resulted in Akt phosphorylation and nuclear translocation of nuclear factor kappa B
19 (NF-KB), effects which were blocked by antioxidants (Kim et al., 2008, 193961). Significant
20 increases in apoptosis were also observed in this model. Thus in this study, CO exposure led to
21 uncoupled mitochondrial respiration and ROS-induced programmed cell death.
22 Further evidence for cellular redox stress is provided by studies in which glutathione stores
23 were altered following CO exposure in vitro (Kim et al., 2008, 193961; Patel et al., 2003, 043155).
24 In addition, mitochondrial redox stress was observed in livers of rats exposed to 50 ppm CO
25 (Piantadosi et al., 2006, 180424). Furthermore, an adaptive increase in intracellular antioxidant
26 defenses (i.e., superoxide dismutase) was observed in endothelial cells exposed to 10 ppm CO for 40
27 min (Thorn et al., 2000, 011574) and mitochondrial biogenesis was observed in hearts of mice
28 exposed to 250 ppm CO for 1 h (Suliman et al., 2007, 193768).
29 Several mechanisms could contribute to the cellular redox stress elicited by CO exposure.
30 First, inhibition of cytochrome c oxidase could result in increased mitochondrial superoxide
31 generation. Secondly, interactions of CO with heme proteins could lead to the release of heme and
32 free iron and subsequently to the generation of ROS. As mentioned above, increased ROS generation
33 has been linked to cellular injury and death, inflammation, and endothelial dysfunction.
September 2009 5-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 Two of the above-mentioned studies demonstrated that CO-mediated mechanisms were
2 unrelated to hypoxia by showing that hypoxic conditions failed to mimic the results obtained with
3 CO. Hence the mitochondrial redox stress and mitochondrial pore transition observed in livers from
4 rats exposed to CO (Piantadosi et al., 2006, 180424) and the cardiac mitochondrial biogenesis
5 observed in mice exposed to CO (Suliman et al., 2007, 193768) were attributed specifically to non-
6 hypoxic mechanisms of CO.
Alteration in Ion Channel Activity
7 Work by Dubuis et al., (Dubuis et al., 2002, 193911) demonstrated increased current through
8 Ca2+-activated K+ channels in smooth muscle cells from pulmonary arteries of rats exposed to
9 530 ppm CO for 3 wk. These findings provide further evidence for non cGMP-dependent
10 vasodilatory actions of CO.
Modulation of Protein Kinase Pathways
11 Endogenously produced CO is a gaseous second messenger molecule in the cell. Work from
12 numerous laboratories has demonstrated the potential for CO to be used as a therapeutic gas with
13 numerous possible clinical applications, since it can produce anti-inflammatory, anti-apoptotic, and
14 anti-proliferative effects (Durante et al., 2006, 193778; Ryter et al., 2006, 193765). These studies
15 generally involved pretreatment with CO followed by exposure to another agent 12-24 h later. There
16 is extensive literature on this topic as reviewed by Ryter et al.(Ryter et al., 2006, 193765). Durante et
17 al., (Durante et al., 2006, 193778) and others. A number of these processes are mediated through
18 cGMP while others involve redox-sensitive kinase pathways, possibly secondary to CO-dependent
19 generation of ROS. For example, 250 ppm CO inhibited growth of airway smooth muscle cells by
20 attenuating the activation of the extracellular signal-regulated kinase 1/2 (ERK 1/2) pathway,
21 independent of sGC and other MAP kinases (Song et al., 2002, 037531). A second example is
22 provided by the study of Kim et al., (Kim et al., 2005, 193959) where 250 ppm CO inhibited PDGF-
23 induced smooth muscle cell proliferation by upregulating p2lWafl/Clpl and caveolin-1, and down-
24 regulating cyclin A expression. In this case, effects were dependent upon cGMP and the p38 MAPK
25 pathway (Kim et al., 2005, 193959). Thirdly, rat endothelial cells exposed to 15 ppm CO escaped
26 anoxia/reoxygenation-induced apoptosis via modulation of the signaling pathways involving
27 phosphoinositide 3-kinase (PI3K), Akt, p38 MAP kinase, Signal Transducers and Activators of
28 Transcription (STAT-1) and STAT-3 (Zhang et al., 2005, 184460). In a fourth study, Akt was found
29 to be responsible for the CO-induced activation of NF-KB, protecting against hepatocyte cell death
30 (Kim et al., 2008, 193961). While research focusing on therapeutic applications of CO generally
31 involves high level, short-term exposure to CO (i.e., 250-1,000 ppm for up to 24 h), some studies
September 2009 5-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 found effects below 20 ppm (Zhang et al, 2005, 184460). Few if any studies on the therapeutic
2 effects of CO have explored the dose-response relationship between CO and pathway
3 activation/deactivation, so it remains unclear how these effects may be related to environmentally -
4 relevant exposures.
Concentration-Response Relationships
5 In many cases the concentrations of exogenous CO required for these non-hypoxic effects was
6 much higher (Alonso et al., 2003, 193882; Favory et al., 2006, 184462; Iheagwara et al., 2007,
7 193861; Thorup et al., 1999, 193782) than concentrations of CO in ambient air. However in some
8 studies the effects were mimicked by upregulation of HO-1 which would result in increased local
9 production of CO as well as of iron and biliverdin (D'Amico et al., 2006, 193992; Imai et al., 2001,
10 193864; Thorup et al., 1999, 193782). For example, HO-1 upregulation or overexpression attenuated
11 carbachol-mediated NO release and NO-mediated vasodilation, similar to the effects of exogenous
12 CO in these same models (Imai et al., 2001, 193864; Thorup et al., 1999, 193782). In the study by
13 D'Amico et al., (D'Amico et al., 2006, 193992). overexpression of HO-1 in cells inhibited cellular
14 respiration by 12% and decreased cytochrome c oxidase activity by 23%. It is not clear how
15 comparable these conditions involving increased intracellular concentrations of endogenous CO are
16 to increased intracellular concentrations of CO resulting from exogenous CO exposures. Neither is it
17 clear what concentrations of intracellular CO are generated locally within cells as a result of HO-
IS catalyzed heme metabolism. However, a small amount of a relatively high local concentration of
19 endogenous CO produced in a regulated manner by HO-1 and HO-2 may be sufficient to react with
20 local targets (e.g., heme proteins) while a larger amount of exogenous CO may be required to reach
21 the same targets. This may be due to indiscriminate reactions of exogenous CO with other target
22 proteins or to other issues related to compartmentalization. It is conceivable that acute or chronic
23 exposures to ambient CO could "sensitize (or "desensitize") targets of endogenous cellular CO
24 production, but there is no experimental evidence to support this mechanism.
25 There is a growing appreciation that non-hypoxic mechanisms may contribute to the effects
26 associated with CO toxicity and poisoning (Ischiropoulos et al., 1996, 079491; Thorn et al., 1994,
27 076459; Weaver et al., 2007, 193939). On the other hand, recent studies suggest that exogenous CO
28 at lower concentrations may have beneficial anti-inflammatory, anti-proliferative and cytoprotective
29 effects under certain circumstances (Durante et al., 2006, 193778; Ryter et al., 2006, 193765). Since
30 the focus of this assessment is on mechanisms which are relevant to ambient exposures, it is
31 important to understand which mechanisms may occur at "low" (50 ppm and less) and "moderate"
32 (50-250 ppm CO) concentrations of CO. Hence, both recent animal studies and relevant older ones
33 which add to the understanding of mechanisms in this range of CO concentrations are briefly
September 2009 5-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 summarized in Table 5-1. It should be noted that most of the above-mentioned non-hypoxic
2 mechanisms were demonstrated at CO concentrations of 50 ppm and less.
September 2009 5-11 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-1 Responses to low and moderate CO exposures.
Reference
Model System
CO Exposure
Response
Notes
IN VITRO
Alonso et al. (2003,
1938821
Thorn and Ischiropoulos
(1997, 0856441
Thorn etal. (1997,
0843371
Thorn et al. (2000,
0115741
Song et al. (2002,
0375311
Kim et al. (2005,
1939591
Kim et al. (2008,
1939611
Zhang et al. (2005,
1844601
Zuckerbraun et al.
(2007, 1938841
Ghio et al. (2008,
0963211
Human muscle
mitochondria
Rat
platelets
Bovine pulmonary artery
endothelial cells
Bovine pulmonary artery
endothelial cells
Human aortic smooth
muscle cells
Rat pulmonary artery
smooth muscle cells
Rat hepatocytes
Rat pulmonary artery
endothelial cells
Mouse macrophages
Human bronchial epithelial
cells
50, 100, 500 ppm
5 min
10 ppm
20 ppm
30-60 min
10 ppm
40 min
50-500 ppm
24 h
250 ppm
1h
250 ppm
1h
2x per day
250 ppm
1 h
15 ppm
0.5-24 h
50 and 250 ppm
1h
10-1 00 ppm
24 h
Decreased cytochrome c oxidase activity
Increased free NO
Increased free NO and peroxynitrite
Increased MnSOD and protection against
toxic effects of 100 ppm CO
Inhibition of cellular proliferation
Inhibited PDGF- induced smooth muscle cell
proliferation
Blocked spontaneous apoptosis
Increased mitochondrial ROS generation,
increased mitochondrial glutathione
oxidation, and decreased cellular ascorbic
acid
Blocked anoxia-reoxygenation mediated
apoptosis
Increased ROS generation (dose dependent
response for 50-500 ppm CO)
Dose-dependent decrease in cellular non-
heme iron (effect at 10 ppm was signficant,
effect at 50 ppm maximal)
Dose-dependent decrease in cellular ferritin
at 50-100 ppm
50 ppm blocked iron uptake by cells
50 ppm increased iron release from cells
Reported to correspond to 7% COHb
Adaptive responses
Blocked activation of ERK1/2 pathway,
independent of sGC and other MAP
kinases
Upregulated P2iwaf1/cip1 and caveolin-
1, and down-regulated cyclin A
expression.
CO induced Akt phosphorylation via
ROS production
CO activated NFKB
Modulation of PI3K/Akt/p38 MAP
kinase and STAT-1 and STAT-3
Mitochondrial derived ROS and
cytochrome c oxidase inhibition
demonstrated for 250 ppm
Compare with in vivo experiments in
same paper
IN VIVO
Ghio et al. (2008,
0963211
Thorn etal. (1999,
0167531
Rats
Rats
50 ppm
24 h
50 ppm
1 h
100 ppm
1h
Mild neutrophil accumulation in BALF
Increased lavage MIP-2, protein, LDH
Lavage iron and ferritin were increased by
CO
Lung non-heme iron was decreased by CO
Liver non-heme iron was increased by CO
Increased nitrotyrosine in aorta
Leukocyte sequestration in aorta after 18 h
Albumin efflux from skeletal muscle
microvasculature 3 h after CO
LDL oxidation
Compare with in vitro experiments in
same paper
Effects blocked by NOS inhibitor
September 2009
5-12
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Thorn etal. (1999,
0167571
Sorhaug et al. (2006,
1804141
Leonnechen et al.
(1999, 0115491
Favory et al. (2006,
1844621
Piantadosi et al. (2006,
1804241
Suliman et al. (2007,
1937681
Wellenius et al. (2004,
0878741
Wellenius et al. (2006,
1561521
Carraway et al. (2002,
0260181
Gautier et al. (2007,
0964711
Melin et al. (2005,
1938331
Melin et al. (2002,
0375021
Model System
Rats
Rats
Rats
Rats
Rats
Mice
Rats
Model of Ml
Rats
Model of Ml
Rats
Model of hypoxic
pulmonary vascular
remodeling
Rats
Model of right ventricle
hypertrophy secondary to
chronic hypoxia
Rats
Model of right ventricle
hypertrophy secondary to
chronic hypoxia
Rats
Model of right ventricle
hypertrophy secondary to
chronic hypoxia
CO Exposure
100ppm
1 h
SOppm
1h
200 ppm
72 wk
100 and 200 ppm
1-2wk
250 ppm
90min
50 ppm CO or
hypobaric hypoxia for
1, 3, or 7 days
250 ppm
1 h
35 ppm
1h
35 ppm
1 h
Hypobaric hypoxia
+ 50 ppm CO
3wk
3 wk of hypobaric
hypoxia with 50 ppm
CO during last week
SOppm
10 wk
50 ppm
10 wk
Response
Elevated free NO during CO exposure
(EPR)
Elevated nitrotyrosine in lung homogenates
Lung capillary leakage 18 h after exposure
No changes in lung morphology
No pulmonary hypertension
No atherosclerotic lesions in systemic
vessels
Ventricular hypertrophy
Increased ET-1 mRNA in the heart
ventricles, increased right and left
ventricular weight
Complex IV inhibition in myocardial fibers
Inhibition of vasodilatory response to
acetylcholine and SNP, Increased coronary
perfusion pressure and contractility
Liver mitochondrial oxidative and nitrosative
stress, altered mitochondrial permeability
pore transition sensitivity
Cardiac mitochondrial biogenesis
Decreased delayed ventricular beat
frequency
Decreased supraventricular ectopic beats
CO promoted remodeling and increased
pulmonary vascular resistance
Rats with pulmonary hypertension were
more sensitive to CO which altered the right
ventricular adaptive response to pulmonary
hypertension leading to ischemic lesions
CO increased cardiac dilation and
decreased left ventricular function
CO increased right ventricular hypertrophy,
decreased right ventricular diastolic function
and increased left ventricular weights
Notes
Inhibition of NOS abrogated CO
effects
12and23%COHb
11%COHb
CO effects not mimicked by hypobaric
hypoxia
Activation of GC involved. No role for
NOS. Increased mitochondrial H202
and activation of Akt proposed
Altered arrhythmogenesis
Altered arrhythmogenesis
5.1.3.3. Implications of Non-Hypoxic Mechanisms
1 A key issue in understanding the biological effects of environmentally-relevant exposures to
2 CO is whether the resulting partial pressures of CO (pCO) in cells and tissues can initiate cell
3 signaling which is normally mediated by endogenously generated CO or perturb signaling which is
4 normally mediated by other signaling molecules such as NO.
September 2009
5-13
DRAFT - DO NOT CITE OR QUOTE
-------
1 Several aspects need to be considered. First of all, during a period of exogenous CO uptake,
2 Hb acts as a buffer for most cells and tissues by limiting the availability of free CO. Nevertheless,
3 COHb delivers CO to cells and tissues. This delivery involves CO's dissociation from Hb followed
4 by its diffusion down a pCO gradient. Hence, greater release of CO from COHb will occur under
5 conditions of low cell/tissue pCO. Conversely, higher cell/tissue pCO in cells/tissues than in the
6 blood will lead to the egress of CO from cells/tissues.
7 A second consideration is the role played by O2 in competing with CO for binding to
8 intracellular heme protein targets. In general, heme proteins (e.g., cytochrome c oxidase) are more
9 sensitive to CO when O2 is limited. Hence hypoxic conditions would be expected to enhance the
10 effects of CO. This concept is demonstrated in the study by D'Amico et al., (D'Amico et al., 2006,
11 193992). NO, which also competes with O2 and CO for binding to heme proteins may have a similar
12 impact.
13 A third consideration is whether certain cell types serve as primary targets for the effects of
14 CO. Besides the blood cells (including leukocytes and platelets), the first cells encountering CO
15 following its dissociation from Hb are the endothelial cells which line blood vessels. An exception to
16 this situation is in the lungs where epithelial and inflammatory cells found in airways and alveoli are
17 exposed to free CO prior to CO binding to Hb. These lung cells may also serve as unique targets for
18 CO. Processes such as pulmonary microvascular endothelial dysfunction, inflammatory cell
19 activation and respiratory epithelial injury may ensue as a result of preferential targeting of these cell
20 types.
21 Since there is potential for exogenous CO to affect endogenous pools of CO, the
22 concentrations of CO in cells and tissues before and after exogenous exposures are of great interest.
23 Table 5-2 summarizes findings from 4 recent studies relevant to this issue. It should be noted that
24 exposure to 50 ppm CO resulted in a three- to fivefold increase in tissue CO concentration.
September 2009 5-14 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-2 Tissue concentration of CO following inhalation
Reference
Cronje et al. (2004,
1804401
Vreman et al. (2005,
1937861
Piantadosi et al. (2006,
1804241
Vreman et al. (2005,
1937861
CO Exposure
Rat
2,500 ppm
45min
Mice
500 ppm
SOmin
Rats
50 ppm
1-7 days
Mice
50, 250 and
1250 ppm
1h
Tissue CO Concentrations COHb
Blood: 27,500 (800) pmol/mg
Heart: 800 (300) pmol/mg
Muscle: 90 (80) pmol/mg fifi 79o/
Brain: 60 (40) pmol/mg
Control levels in parentheses
Blood: 2648 + 400 (45) pmol/mg
Heart: 100 + 18(6) pmol/mg
Muscle: 14+1 (10) pmol/mg
Brain: 18 + 4(2) pmol/mg
Kidney: 120 + 12 (7) pmol/mg 9RO/
Spleen: 229 + 55 (6) pmol/mg M'°
Liver: 115 + 31 (5) pmol/mg
Lung: 250 + 2 (3) pmol/mg
Intestine: 9 + (4) pmol/mg
Testes: 6 + 3 (2) pmol/mg
Control levels in parentheses
Liver: 30-40 pmol/mg 4-5%
Control liver 10 pmol/mg Control 1 %
Heart (left ventricle)
50 ppm: 50 pmol/mg
250 ppm: 95 pmol/mg
1250 ppm: 160 pmol/mg
Control heart: 9 pmol/mg
exposure.
Notes
CO concentration increased in the heart but not in brain or
skeletal muscle after CO exposure
A later report stated that these tissue CO values were too
high due to a computational error (Piantadosi et al., 2006,
1804241
CO concentration relative to 100% blood:
Lung: 9.4%
Spleen: 8.6%
Kidney: 4.5%
Liver: 4.3%
Heart: 3.8%
Brain: 0.7%
Muscle: 0.5%
Intestine: 0.3%,
Testes: 0.2%
CO concentration reached a plateau after 1 day
No mention of COHb% but exposures were similar to those
in Cronje et al. (2004, 1804401
Data is expressed as pmol CO/mg tissue wet weight
1 Furthermore, endogenous CO production is known to be increased during inflammation,
2 hypoxia, increased heme availability and other conditions of cellular stress where HO-1 or HO-2
3 activity is increased. A few studies reported cell and tissue concentrations of CO along with
4 accompanying COHb levels resulting from enhanced endogenous CO production. Table 5-3
5 summarizes these findings. Additional measurements of CO levels in cells and tissues following
6 increased endogenous production and following inhalation of exogenous CO may provide further
7 insight into the relationship between the CO tissue concentration and biological responses.
September 2009
5-15
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-3 Tissue concentration of CO following increased endogenous production.
Reference
Carraway et al. (2000,
0210961
Piantadosi et al. (2006,
1804241
Vreman et al. (2005,
1937861
Exposure
Rats
Hypobaric hypoxia for 21 days
Rats
Hypobaric hypoxia
1-7 days
Mice
30 pM heme
Tissue CO COHb
1.5-2.8%
Control 0.5%
Liver:5-12pmol/mg 1-1.25%
Control liver 10 pmol/mg Control 1%
Blood: 88 + 10 (45) pmol/mg
Heart: 14 + 3 (6) pmol/mg
Muscle: 7 + 1 (10) pmol/mg
Brain: 2 + 0 (2) pmol/mg
Kidney: 7 + 2 (7) pmol/mg n q%
Spleen: 11 + 1 (6) pmol/mg u's/0
Liver: 8 + 3 (5) pmol/mg
Lung: 8 + 3 (3) pmol/mg
Intestine: 3 + 1 (4) pmol/mg
Testes: 2 + 0 (2) pmol/mg
Control levels in parentheses
Notes
COHb highest after days 1 and 21
at 3-4 fold higher than controls
CO concentration reached a plateau after 1 day
CO concentration relative to 100% blood:
Heart: 16%
Spleen: 13%
Lung: 9%
Liver: 9%
Kidney: 8%
Muscle: 8%
Intestine: 3%
Brain: 2%
Testes: 2%
Data is expressed as pmol CO/mg tissue wet weight
1 It should be noted that increased cellular and tissue concentrations of biliverdin and iron
2 accompany the increased endogenous production of CO by HO-1 and HO-2. Biliverdin and iron
3 have known biological effects, with biliverdin exhibiting antioxidant properties and iron exhibiting
4 pro-oxidant properties (Piantadosi et al., 2006, 180424). which could complicate interpretation of
5 results from studies in which HO-1 and HO-2 activities are increased. In addition, indiscriminate
6 reactions occurring in the case of exogenous CO would likely lead to less specific responses than
7 those mediated by reactions of endogenously-produced CO with local targets. Hence the situations of
8 increased endogenous CO production and of exogenous CO exposure are not equivalent.
9 A further consideration is that in the numerous conditions and disease states where HO-1 is
10 induced, increased levels of endogenously produced CO may represent an adaptive response to stress
11 (Durante et al., 2006, 193778; Piantadosi, 2008, 180423). These increases and the accompanying
12 increases in COHb generally fall in the range of 1.5-4 fold, with the exception of some situations of
13 hemolytic anemia and hemoglobin disorders (see Figure 4-12 for results in humans). The resulting
14 excess endogenous CO may react intracellularly with heme proteins or diffuse into the blood
15 according to the gradient of pCO in the cell/tissue and blood compartments. In many cases,
16 beneficial effects or compensatory mechanisms may result as a result of short-term induction of
17 HO-1, as reviewed by Ryter et al., (Ryter et al., 2006, 193765) and Durante et al., (Durante et al.,
18 2006, 193778). Longer term increases in HO-1 are sometimes associated with protective responses
19 as in the case of atherosclerosis (Cheng et al., 2009, 193775; Durante et al., 2006, 193778) and
20 sometimes with pathophysiologic responses as demonstrated in hypoxic pulmonary vascular
21 remodeling (Carraway et al., 2002, 026018) and models of salt-sensitive hypertension (Johnson et
September 2009
5-16
DRAFT - DO NOT CITE OR QUOTE
-------
1 al., 2003, 193868: Johnson et al., 2004, 193870) and metabolic syndrome (Johnson et al., 2006,
2 193874). Increased endogenous CO in hearts of individuals with ischemic heart disease and in lungs
3 of individuals with various forms of inflammatory lung disease, might also be expected (Scharte et
4 al., 2006, 194115: Yamada et al., 2008, 193232: Yasuda et al., 2005, 191953) (see Figure 4-12). It is
5 conceivable that prolonged increases in endogenous CO production in chronic disease states may
6 result in less of a reserve capacity to handle additional intracellular CO resulting from exogenous
7 exposures but there is no experimental evidence to support this mechanism. Perhaps these
8 circumstances lead to dysregulated functions or toxicity. Thus CO may be responsible for a
9 continuum of effects from cell signaling to adaptive responses to cellular injury (Piantadosi, 2008,
10 180423) depending on intracellular concentrations of CO, heme proteins and molecules which
11 modulate CO binding to heme proteins.
5.1.3.4. Summary
12 CO is a ubiquitous cell signaling molecule with numerous physiological functions. The
13 endogenous generation and release of CO from heme by HO-1 and HO-2 is tightly controlled, as is
14 any homeostatic process. However, exogenously-applied CO has the capacity to disrupt multiple
15 heme-based signaling pathways due to its nonspecific nature. Only a limited amount of information
16 is available regarding the impact of exogenous CO on tissue and cellular levels of CO and on
17 signaling pathways. However recent animal studies demonstrated increased tissue CO levels and
18 biological responses following exposure to 50 ppm CO. Whether or not environmentally-relevant
19 exposures to CO lead to adverse health effects through altered cell signaling is an open question for
20 which there are no definitive answers at this time. However, experiments demonstrating
21 oxidative/nitrosative stress, inflammation, mitochondrial alterations and endothelial dysfunction at
22 concentrations of CO within one or two orders of magnitude higher than ambient concentrations
23 suggest a potential role for such mechanisms in pathophysiologic responses. Furthermore, prolonged
24 increases in endogenous CO resulting from chronic diseases may provide a basis for the enhanced
25 sensitivity of susceptible populations to CO-mediated health effects such as is seen in individuals
26 with coronary artery disease.
September 2009 5-17 DRAFT - DO NOT CITE OR QUOTE
-------
Intracellular
Heme Oxygenases
HO-1 and HO-2
Ambient and
Non-ambient CO
Cell Signaling
Pathways
Physiologic or
Pathophysiologic
Effects
Figure 5-1 Direct Effects of CO. The dashed line refers to uptake of inhaled CO by
respiratory epithelial cells and resident macrophages in the lung. The uptake of
CO by all other cells and tissues is dependent on COHb.
5.2. Cardiovascular Effects
5.2.1.Epidemiologic Studies with Short-Term Exposure
1 The 2000 CO AQCD (U.S. EPA, 2000, 000907) examined the association between short-term
2 variations in ambient CO concentrations and cardiovascular morbidity. While the results presented
3 by these studies did provide suggestive evidence of ambient CO levels being associated with
4 exacerbation of heart disease, the AQCD determined that the evidence was inconclusive. The reasons
5 for this conclusion, which are shared with those studies that examined the effect of short-term
6 exposure to CO on mortality and other types of morbidity, were given as: internal inconsistencies
7 and lack of coherence of the reported results within and across studies; the degree to which average
8 ambient CO levels derived from fixed-site monitors are representative of spatially heterogeneous
9 ambient CO values or of personal exposures that often include nonambient CO; and the lack of
10 biological plausibility for any harmful effects occurring with the very small changes in COHb levels
11 (from near 0 up to 1.0%) over typical baseline levels (about 0.5%) that would be expected with the
12 low average ambient CO concentrations reported in the epidemiologic studies (generally <5.0 ppm,
13 1-h daily max) (U.S. EPA, 2000, 000907). The AQCD also posed the possibility that the ambient CO
14 levels used as exposure indices in the epidemiology studies may be surrogates for ambient air mixes
September 2009
5-18
DRAFT - DO NOT CITE OR QUOTE
-------
1 impacted by combustion sources and/or other constituent toxic components of such mixes. Overall,
2 the AQCD observed that the epidemiologic evidence was stimulating increased scientific interest
3 regarding ambient CO exposures as a potential risk factor for exacerbation of heart disease and other
4 health effects although the epidemiologic studies were subject to considerable biological and
5 statistical uncertainty. Furthermore, the AQCD called for additional research on the health effects of
6 ambient CO exposure alone and CO as a component of the overall ambient air pollution mixture.
7 The following section reviews the literature since the 2000 CO AQCD, including numerous
8 new studies on relevant cardiac endpoints and biomarkers and additional studies of daily hospital
9 admissions for heart disease. New epidemiologic evidence addresses some of the aforementioned
10 uncertainties, including consistency and coherence of results and the possibility that CO may be
11 acting as a surrogate for other combustion-derived air pollutants.
5.2.1.1. Heart Rate and Heart Rate Variability
12 Heart rate variability (HRV) refers to the beat-to-beat alterations in the heart and is generally
13 determined by analyses of time and frequency domains measured by electrocardiograms (ECG). The
14 time domains often analyzed are (a) normal-to-normal (NN or RR) time interval between each QRS
15 complex, (b) standard deviation of the normal-to-normal interval (SDNN), and (c) mean squared
16 differences of successive difference normal-beat to normal-beat intervals (rMSSD), shorter time
17 domain variables results in lower HRV. The frequency domains often analyzed are a) the ratio of low
18 energy frequency (LF) to high energy frequency (HF) and b) the proportion of interval differences of
19 successive normal-beat intervals greater than 50 ms (PNN50), reflecting autonomic balance.
20 Decreased HRV is associated with a variety of adverse cardiac outcomes such as arrhythmia,
21 myocardial infarction (MI), and heart failure (De Jong and Randall, 2005, 193996; Deedwania et al.,
22 2005, 195134; Huikuri et al., 1999, 184464; Rajendra Acharya et al., 2006, 193787).
23 Two studies investigated the association between ambient air pollution, including CO, and
24 HRV in Boston, MA and reported inconsistent results. The earlier of these studies recruited
25 twenty-one 53- to 87-yr old active residents and performed up to 12 ECG assessments on each
26 subject over a period of 4 months (during summer 1997). Particles (PMi0, PM2.5) and several
27 gaseous pollutants (O3, NO2, and SO2) were monitored at fixed sites (up to 4.8 mi from the study
28 site) while CO was monitored 0.25 mi from each participants' residence. Lag periods for the
29 preceding 1 h, 4 h, and 24 h before each subject's HRV assessment were analyzed and results
30 showed that only PM2.5 and O3 were associated with HRV parameters (Gold et al., 2000, 011432).
31 A similar study by the same group of researchers 2 yr later involved 28 older subjects (aged
32 61-89 yr) who were living at or near an apartment complex located on the same street as the Harvard
33 School of Public Health. The subjects were seen once a week for up to 12 wk and HRV parameters
34 (SDNN, r-MSSD, PNN50, LF/HF ratio) were measured for 30 min each session. Data for PM2.5,
September 2009 5-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 black carbon (BC), and CO were recorded at the Harvard School of Public Health (<1 km from the
2 residence) while data for NO2, O3, and SO2 were collected from government regulatory monitoring
3 sites. There were moderate correlations between CO and PM2.5 (r = 0.61) and NO2 (r = 0.55), but not
4 with SO2 (r = 0.18) or O3 (r = 0.21). Similarly PM25 was associated with HRV, whereas in contrast
5 to the previous study, CO was associated1 with a negative change in SDNN (% change: -13
6 [95% CI: -24.06 to -1.88]), r-MSSD (% change: -31.88 [95% CI: -38 to -7.5]), and PNN50
7 (% change: -46.25 [95% CI -103.95 to -9.38] per 0.5 ppm increase in 24-h avg CO concentration)
8 (Schwartz et al, 2005, 074317).
9 A later Boston, MA study examined HRV parameters (SDNN, LF, HF, LF/HF ratio) among
10 603 persons from the Normative Aging Study, a longitudinal study that originally recruited 2,280
11 men in the greater Boston area during 1963. The cohort members were examined (November
12 2000-October 2003) and the ECG data were linked to air pollution data for PM2 5, particle number
13 concentration, BC, O3, NO2, SO2, and CO. Lagged pollutant effects for a 4-h, 24-h, and 48-h
14 moving avg were used. Since previous studies established variable CO results, the main pollutant
15 effects were with PM2.5 and O3 while CO was not associated with HRV (Park et al., 2005, 057331).
16 A study in Mexico City selected 30 subjects from the outpatient clinic at the National Institute
17 of Cardiology and followed them for -10 h (starting at 9:00 a.m.) (Riojas-Rodriguez et al., 2006,
18 156913). Each subject was connected to a Holter ECG monitor (e.g., a portable ECG monitor) and
19 also given personal PM25 and CO monitors. The subjects went about their usual daily activities and
20 the personal PM25 and CO data were linked to various ECG parameters (heart rate [HR], R-R, LF,
21 HF) at various lags. In copollutant models (PM25 and CO) personal CO exposure for the same 5-min
22 period was significantly associated with a decrease in LF and very low energy frequency (VLF)
23 parameters with coefficients equal to -0.024 (95% CI: -0.041 to -0.007) and -0.034 (95% CI: -0.061
24 to -0.007) respectively for a 1 ppm increase in 1-h CO concentration.
25 In Mexico City, 34 residents from a nursing home underwent HRV analysis every other day
26 for 3 months (Holguin et al., 2003, 057326). Exposure assessment for ambient PM25 was based on
27 data recorded at a monitor on the roof of the nursing home while exposures to ambient O3, NO2,
28 SO2, and CO were derived from data recorded at a fixed site 3 km from the nursing home.
29 Exposures for the same day and 1-day lags were analyzed and only O3 and PM25 were positively
30 associated with HRV.
31 Wheeler et al. (2006, 088453) examined 18 individuals with COPD and 12 individuals with
32 recent MI living in Atlanta, GA. Morning ECG readings were collected by a Holter system by a field
33 technician in the subjects' homes. Ambient air pollution exposures for PM25, O3, NO2, SO2 and CO
•' The effect estimates from epidemiologic studies have been standardized to a 1 ppm increase in ambient CO for 1-h max CO
concentrations, 0.75 ppm for 8-h max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations throughout this section (text,
tables, and figures).
September 2009 5-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 were derived from data recorded at fixed sites throughout metropolitan Atlanta. Three exposure
2 periods were analyzed: the hour of the ECG reading, 4-h mean and 24-h mean before the reading.
3 While positive effects were reported for NO2 and PM2.5, no quantitative results were reported for
4 CO.
5 After reviewing 2,000 patient charts, Dales (2004, 099036) recruited 36 subjects with coronary
6 artery disease (CAD) from the Toronto Western Hospital's noninvasive cardiac diagnostic unit. HR
7 and HRV (SDNN, N-N, HF, LF, HF/LH ratio) were assessed 1 day each week for up to 10 wk by a
8 Holter monitoring system. Personal air sampling for PM2.5 and CO was carried out for the same 24-h
9 period whereby subjects went about their usual daily activities for that period. Stratified results
10 showed that among those not on beta-receptor-blockers, personal CO exposure was positively
11 associated with SDNN (p = 0.02). However, in the group taking beta blockers there was a negative
12 association (p = 0.06). Personal exposure to PM2.5 was not associated with HRV.
13 HR was examined among a sub-sample of the Monitoring of Trends and Determinants in
14 Cardiovascular Disease (MONICA) study (n = 2,681) in Augsburg, Germany by Peters and
15 colleagues (1999, 011554). Total suspended particles (TSP), SO2, and CO data were collected from a
16 single monitoring station located in the center of the city and linked to each subject to estimate
17 exposures on the same day and 5 days prior. A 0.5 ppm change in 24-h CO concentration was
18 associated with an increase in HR of approximately 1 beat per minute, whereas CO based on a 5-day
19 exposure had no effect on HR.
20 Thirty-one subjects with CHF had their pulse rate recorded daily over a 2-mo period and the
21 correlation between pulse rate and air pollutants was examined (Goldberg et al., 2008, 180380).
22 There was weak evidence for a decrease in pulse rate associated with the lag 1 SO2 concentration
23 after adjustment for personal and meteorological factors, and no evidence for an effect associated
24 with any of the other air pollutants.
25 Liao et al (2004, 056590) investigated men and women aged 45-64 yr from the Atherosclerosis
26 Risk in Communities (ARIC) study (Washington County, MD; Forsyth County, NC; and selected
27 suburbs of Minneapolis, MN). The sample sizes were 4,899, 5,431, 6,232, 4,390 and 6,784 for
28 analyses involving PMi0, O3, CO, NO2, and SO2 respectively. County level exposure estimates for
29 24 h CO were calculated for 1, 2, and 3 days prior to clinical examination. A 0.5 ppm increase in
30 24-h CO concentration (at lag 1) was associated with an increase in HR (beats/minute) ((3 = 0.357,
31 p < 0.05). CO was not significantly associated with changes in SDNN.
32 The Exposure and Risk Assessment for Fine and Ultrafine Particles in Ambient Air (ULTRA)
33 study was carried out in three European cities: Amsterdam, the Netherlands, Erfurt, Germany, and
34 Helsinki, Finland, whereby a panel of subjects with CAD was followed for 6 mo with biweekly
35 clinical visits, which included an ECG reading to assess HRV (Timonen et al., 2006, 088747). The
36 time domain measures of HRV (SDNN and rMSSD) were analyzed along with frequency domain
September 2009 5-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 measures, which included power spectrum densities for LF and HF. Exposures to ambient air
2 pollution (PM2.5, PMi0, NO2, CO) were derived from data recorded at fixed monitoring site
3 networks within each city. Correlation coefficients for NO2 and CO ranged from 0.32 to 0.86 in the 3
4 cities. CO was moderately correlated with PMi0 in Helsinki (r = 0.40) and with PM2.5 in Amsterdam
5 (r = 0.58) and more highly correlelated with PMi0 in Erfurt (r = 0.77). Various lag periods were
6 examined including lag 0 (24 h prior to the clinical visit) through a 0-2-day avg lag and a 0-4-day
7 avg lag. In total there were 1,266 ECG recordings used in the final analyses. In the pooled analyses
8 (e.g., across cities) a 0.5 ppm increase in 24-h CO concentration was associated with a decrease in
9 LF/HF ratio at lag 1-day ((3 -16.4 [95% CI: -29.9 to -0.3]), and a decrease in SDNN and HF at lag
10 2-day (P -3.4 [95% CI: -6.1 to -0.4]; (3= -17.6 [95% CI: -34.4 to -0.9], respectively). However, the
11 same study reported no effect for CO on BP and HR (Ibald-Mulli et al., 2004, 087415).
12 A small panel study in Kuopio, Finland, which was designed as the pilot study for the ULTRA
13 study examined simultaneous ambulatory ECG and personally monitored CO readings among
14 6 male patients with CAD (Tarkiainen et al., 2003, 053625). The patients were asked to follow their
15 usual daily activities, but data were recorded only three times with 1-week intervals. The CO
16 exposures were divided into low (< 2.7 ppm) and high (>2.7 ppm) and during the high CO exposure
17 r-MSSD increased on average by 2.4 ms. However, there was no effect on RR or SDNN.
18 A study in Taiwan recruited 83 patients (aged 40-75 yr) from the National Taiwan University
19 Hospital, Taipei and conducted ambulatory ECG readings using a Holter system (Chan et al., 2005,
20 088988). Ambient air pollution exposures for PMi0, NO2, SO2, and CO were derived from 12 fixed
21 site monitoring stations across Taipei. Lag periods of 1 h to 8 h prior to the ECG reading were
22 analyzed and only NO2 was associated with HRV parameters (SDNN and LF). CO was not
23 associated with HRV.
24 In summary, few studies have examined the effect of CO on HR and while two of the three
25 studies reported a positive association, further research is warranted to corroborate the current
26 results. Similarly, while a larger number of studies have examined the effect of CO on various HRV
27 parameters, mixed results have been reported throughout these studies. Furthermore, with several
28 HRV parameters often examined, there are mixed results across the studies as to the HRV parameters
29 that are positively associated with CO exposure. Table 5-4 shows a summary of the reviewed studies.
September 2009 5-22 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-4 Summary of studies investigating the effect of CO exposure on HRV parameters.
Study
Gold et al. (2000, 0114321
Schwartz et al. (2005, 0743171
Parketal. (2005, 0573311
Riojas-Rodriguez et al. (2006,
1569131
Holguin et al. (2003, 0573261
Wheeler et al. (2006, 0884531
Dales(2004, 0990361
Peters etal. (1999, 0115541
Goldberg et al. (2008, 1803801
Liao et al. (2004, 0565901
Timonen et al. (2006, 0887471
Ibald-Mulli etal. (2004,
0874151
Tarkiainen et al. (2003,
0536251
Chan et al. (2005, 0889881
Location (Sample
Size)
Boston, MA
(n = 21)
Boston, MA
(n = 28)
Boston, MA
(n = 4 97)
Mexico City, Mexico
(n = 30)
Mexico City, Mexico
(n = 34)
Atlanta, GA
(n = 30)
Toronto, Canada
(n = 36)
Augsburg, Germany
(n = 2681)
Montreal, Canada
(n-31)
Maryland,
North Carolina,
Minnesota,
(n = 4899-6784)
Amsterdam,
the Netherlands;
Erfurt, Germany;
Helsinki, Finland
(n = 131)
Amsterdam,
the Netherlands;
Erfurt, Germany;
Helsinki, Finland
(n = 131)
Kuopio, Finland
(n = 6)
Taipei, Taiwan
(n = 83)
Cardiac Endpoint
HR, SDNN,
r-MSSD
SDNN,
r-MSSD, PNN, LF/HF
SDNN, LF, HF, LF/HF
HF, LF, VLF, HR, R-R
HF.LF, LF/HF
SDNN, r-MSSD, PNN,
LF, HF, LF/HF
SDNN, HF, LF, LF/HF,
N-N
HR
Pulse rate
HR, SDNN, LF, HF
SDNN, HF, LF/HF
BP, HR
PNN, SDNN, r-MSSD
SDNN,
r-MSSD, LF
Upper CO
Concentrations
from AQS* in
ppm
98th%: 0.80-2.48
99th%: 0.89-2.57
(24 h)t
98th%: 0.95-2.14
99th%: 0.96-2.60
(24 h)
98th%: 0.92-1. 45
99th%: 0.99-1. 66
(24 h)
NA
NA
98th%: 2.8-3.1
99th%: 2.9-3.8
(8h)
NA
NA
NA
98th%: 0.39-2.29
99th%: 0.43-2.66
(24 h)
NA
NA
NA
NA
CO Concentrations
Reported by Study
Authors in ppm
Mean: 0.47(24 h)
Range: 0.12-0.82
25th, 50th, 75th
percentiles:
0.38, 0.45, 0.54
Mean: 0.50 (24 h)
Range: 0.13-1. 8
Mean: 2.9 (11 h)
Range: 0.1-18
Mean: 3.3(24 h)
Range: 1.8-4.8
Mean: 362 ppb (4h)
25th, 50th, 75th
percentiles:
221.5,304.3,398.1
Mean: 2.4**
Range: 0.4-16.5
Mean: 3.6
Range: 1.5-7.1
NR;IQR:1.8ppm
Mean: 0.65 (24 h)
Mean: 0.35-0.52
Range: 0.09-2. 17
Mean: 0.35-0.52
Range: 0.09-2. 17
Mean: 4.6
Range: 0.5-27.4
Mean: 1.1
Range: 0.1-7.7
Copollutants
PM,o, PM25, 03, N02,
S02
PM25, BC, N02, 03
PM25, BC, 03, N02,
S02
PM25
PM25, 03, N02, S02
PM25, 03, N02, S02
PM25
TSP, S02
N02, 03, S02, PM25
PM10, 03, N02, S02
PM25, PM10, N02
UFP, PM,o, PM25,
N02, S02
None
PM10, N02,S02
NA: Not Available
Includes range across individual monitors in study site; AQS data available for U.S. studies only
*95th percentile of 24-h levels
September 2009
5-23
DRAFT - DO NOT CITE OR QUOTE
-------
5.2.1.2. ECG Abnormalities Indicating Ischemia
1 The ST-segment of an ECG represents the period of slow repolarization of the ventricles and
2 ST-segment depression can be associated with adverse cardiac outcomes. Gold et al. (2005, 087558)
3 recruited a panel of 28 older adults living at or near an apartment complex located within 1 km of a
4 monitoring site in Boston, MA. Each subject underwent weekly ECGs for 12 wk in summer 1999
5 with the main outcome of interest being the ST-segment. Air pollution data in the form of PM2.5,
6 black carbon (BC), and CO were collected from a central site within 0.5 km of the residences of the
7 subjects and averaged over various lag periods (1-24 h, 12 h and 24 h moving average [ma]) before
8 the ECG. The final analyses included 24 subjects with 269 observations and results showed
9 consistent negative associations of ST-segment level with increased BC with the strongest
10 association with the 5-h lag. CO during the same lag period also showed a negative association with
11 ST-segment depression, however only BC remained significant in multipollutant models.
5.2.1.3. Arrhythmia
12 Cardiac arrhythmia refers to a broad group of conditions where there is irregular electrical
13 activity in the heart. The main types of arrhythmias are fibrillation, tachycardia, and bradycardia, all
14 of which can be associated with the upper (atria) and lower (ventricle) chambers of the heart. Briefly,
15 fibrillation refers to when a chamber of the heart quivers chaotically rather than pumps in an orderly
16 fashion, tachycardia refers to a rapid heart beat (e.g., >100 beats/min) while bradycardia refers to a
17 slow heart beat (e.g., <60 beats/min). A few air pollution panel studies have examined the occurrence
18 of cardiac arrhythmias by analyzing data recorded by implantable cardioverter defibrillators (ICDs)
19 among cardiac patients. The majority of these studies were conducted in North America with the
20 main outcome investigated being tachycardia. Results of these studies provide little evidence for an
21 association between cardiac arrhythmia and ambient CO.
22 For example, Dockery and colleagues (2005, 078995) analyzed the relationship between
23 ambient air pollution and the daily incidence of ventricular tachyarrhythmia among 203 patients with
24 ICDs in Boston, MA. An hourly city average for the Boston metropolitan area was calculated for
25 CO, O3, NO2, SO2, SO42", BC, and PM25. Although positive associations between ventricular
26 arrhythmic episode days were found for all mean pollutant levels on the same day and previous days,
27 none of these associations approached statistical significance. However, when the analyses were
28 stratified by patients who had a previous incidence of ventricular arrhythmia within 3 days, or
29 greater than 3 days to the day of interest, a 0.5 ppm increase in 24-h CO concentration was positively
30 associated with incidence of ventricular arrhythmia (OR: 1.68 [95% CI: 1.18-2.41) among those who
31 had a ventricular arrhythmia within the last 3 days.
September 2009 5-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 A similar study in eastern Massachusetts examined cardiac arrhythmia by analyzing
2 defibrillator discharges precipitated by either ventricular tachycardia or fibrillation among 100
3 cardiac patients (Peters et al, 2000, 011347). Exposure to ambient CO was estimated for the same
4 day, 1-day, 2-day, 3-day, and a 5-day mean lag period. CO was moderately correlated with PMi0
5 (r = 0.51) and PM2.5 (r = 0.56) and more highly correlated with NO2 (r = 0.71). When analyzing
6 patients who had at least one defibrillator discharge (n = 33) there was no association with CO.
7 However, when analyzing patients who had at least 10 discharges (n = 6), a 0.5 ppm increase in 24-h
8 CO concentration (lag 0-4) was associated with an increased odds of a defibrillator discharge (OR:
9 1.66 [95% CI: 1.01-2.76]).
10 In contrast, other air pollution panel studies conducted in St Louis, MO (among 56 subjects)
11 (Rich et al., 2006, 089814). Atlanta, GA (among 518 subjects) (Metzger et al., 2007, 092856).
12 Boston, MA (among 203 subjects) (Rich et al., 2005, 079620). and Vancouver, Canada (Rich et al.,
13 2004, 055631): (Vedal et al., 2004, 055630) (among 34 and 50 subjects respectively) did not find an
14 association between short term changes in ambient CO and occurrence of cardiac arrhythmia in
15 patients with implantable defibrillators. The study in Boston also examined atrial fibrillation
16 episodes among the same group of subjects and also did not find an association with ambient CO
17 (Rich et al., 2005, 079620).
18 An alternative method used to assess the relationship between cardiac arrhythmia and ambient
19 air pollution is to analyze cardiac data recorded via ECG. Two studies have employed this method
20 and reported inconsistent results. A study in Steubenville, OH, which is located in an industrial area,
21 examined weekly ECG data among 32 non-smoking older adults for 24 wk during summer and fall
22 (Sarnat et al., 2006, 090489). Ambient exposures for up to 5 days prior to the health assessment
23 (based on a 5-day moving average) were calculated for PM2.5, SO42", elemental carbon (EC), O3,
24 NO2, SO2, and CO from data recorded at one central monitoring site. Increases in ambient CO were
25 not associated with increased odds of having at least one arrhythmia during the study period.
26 In contrast, a study in Germany examined the relationship between ambient air pollution and
27 the occurrence of supraventricular (atria) and ventricular tachycardia recorded via monthly 24-h
28 ECGs among 57 subjects over a 6 month period (Berger et al., 2006, 098702). Exposure estimates
29 were calculated for ambient ultrafine particles, PM25, CO, NO, NO2, and SO2 for various lag
30 periods (0-23 h, 24-47 h, 48-71 h, 72-95 h, and 5-day avg) prior to the ECG. Results showed that a
31 0.5 ppm increase in ambient 24-h CO concentration (lag 0-4 days prior to ECG) was positively
32 associated with the occurrence of supraventricular tachycardia (OR: 1.36 [95% CI: 1.08-1.74]).
33 However, ambient CO was not associated with ventricular tachycardia.
34 In summary, the studies that have examined associations between CO and the occurrence of
35 cardiac arrhythmias provided little evidence of a CO effect on cardiac arrhythmias. While most
36 studies analyzed data from ICDs, very few reported significant associations. This was similar for the
September 2009 5-25 DRAFT - DO NOT CITE OR QUOTE
-------
1 mixed results from the two studies that analyzed ECG data to evaluate cardiac arrhythmias in
2 association with CO exposures. Table 5-5 summarizes the reviewed studies.
Table 5-5 Summary of studies investigating the effect of CO exposure on cardiac arrhythmias.
Ventricular
arrhythmia
98th %' 0892 33
99th%: 0.99-2.55
25th, 50th, 75th percentiles:
0.4,0.5,0.6 (24h)
PM2.5,EC,03,N02,S02
Atlanta, GA
Metzger etal. (2007, . 51ffl
092856)
Ventricular
Tachycardia
: 5.0
99tho/0.56
99th/0'5'6
(1 h)
Mean:1.7(1h)
Range'01-77
PM10,PM2.5,03,N02,S02
Rich et al (2005
079620)
-
(n = 203)
Atrial fibrillation
99tho/0-o 99-255
(24 h)
25th' 50th. 75th' 95th'
percentiles: 0.53, 0.80, 1.02,
PM2.5,BC,03,N02,S02
Rieta,,2004,
^discharge
arrhythmia
Mean: 0.55 (24 h)
IQR:0.16
PM25,PM10,EC,03,N02,S02,
S04?~
Vedal et al. (2004,
(n = 50)
ARRHYTHMIAS (VIA ECG)
arrhythmia
NA
Mean: 0.6 (24 h)
Range: 0.3-1 .6
PM10,03,N02,S02
Sarnat etal (2006 Steubenville, OH Atrial or
' ' (n = 32) JESS.
98th%: 1.42
994thJ1-81
Mean: 0.2 (24 h)
Range: 0.1, 1.5
PM2.5,03,N02,S02,S042", EC
Be,gereta,.(2006, Erfurt, Germany Atrialor^
"^ =
Mean: 0.45 (24 h)
0,0,0.38,1.68
PM10,PM2.5,N02,NO, S02,UF
NA: Not Available
* includes range across individual monitors in study site; AQS data available for U.S. studies only
5.2.1.4. Cardiac Arrest
3 Cardiac arrest refers to the abrupt loss of heart function due to failure of the heart to contract
4 effectively during systole, which can lead to sudden cardiac death if not treated immediately. Very
5 few studies have investigated the association between ambient CO exposure and the risk of cardiac
September 2009
5-26
DRAFT - DO NOT CITE OR QUOTE
-------
1 arrest and none reported a significant link between increased CO exposure and the occurrence of
2 cardiac arrest.
3 Two similar studies were conducted in Seattle, WA, and both did not report an association
4 between ambient CO and cardiac arrest. Both studies employed a case-crossover design and
5 examined air pollution exposures for black smoke particles (BSP), PMi0, SO2, and CO. The
6 correlation coefficient for PMi0 and CO was 0.8 in both studies. The first of these studies examined
7 paramedic-attended out-of-hospital primary cardiac arrests among 362 cases (between 1998-1994) in
8 Seattle and King County, WA whereby lags of 0-5 days were analyzed (Levy et al., 2001, 017171).
9 The second of these studies examined out-of-hospital primary cardiac arrest for a 10-yr period
10 (1985-1994) among subjects within a health organization database (the Group Health Cooperative of
11 Puget Sound) whereby 0-day through 2-day lags were analyzed (Sullivan et al., 2003, 043156).
5.2.1.5. Myocardial Infarction
12 As previously stated, MI is commonly referred to as 'heart attack' and is another cardiac
13 outcome that has received limited attention within the area of air pollution research. Only one study
14 has investigated the association between short-term changes in ambient CO and the onset of MI.
15 Peters and colleagues (2001, 016546) employed a case-crossover study design to analyze short term
16 exposures (0-5 h and 0-5 days before the onset of MI) to particles (PMi0, PM2.5, PMi0_2.5, BC) and
17 gases (CO, O3, NO2, SO2) among 772 patients with MI in the greater Boston area. While all
18 pollutants showed positive associations with the onset of MI, only PM2.5 reached statistical
19 significance with the main exposure period being 2 h before the onset.
5.2.1.6. Blood Pressure
20 Only two studies have investigated whether short-term exposure to CO influences BP. The
21 earlier of these two studies examined BP among 2607 men and women aged 25-64 yr who
22 participated in the Augsburg, Germany MONICA study (Ibald-Mulli et al., 2001, 016030).
23 Exposures to ambient TSP, SO2 and CO (from one monitor in the center of the city) during the same
24 day as the BP reading and an average over the 5 days prior were examined. Results showed that
25 ambient CO had no association with BP.
26 Similarly, the second of these studies extracted baseline and repeated-measures of cardiac
27 rehabilitation data from a Boston, MA hospital for 62 subjects with 631 visits and analyzed ambient
28 air pollution exposures (with particular focus on PM2 5) averaged over various periods up to 5 days
29 before the visit (Zanobetti et al., 2004, 087489). While results showed significant associations
30 between increased BP and ambient PM25, SO2, O3, and BC, there was no significant effect for CO.
September 2009 5-27 DRAFT - DO NOT CITE OR QUOTE
-------
5.2.1.7. Vasomotor Function
1 Gaseous pollutants, including SO2, NO and CO, were found to affect large artery endothelial
2 function among 40 healthy white male nonsmokers in Paris, France, whereas particulate matter was
3 found to exaggerate the dilatory response of small arteries to ischemia (Briet et al., 2007, 093049).
4 Changes in amplitude of flow-mediated dilatation were highly dependent on changes in 5-day lag
5 concentrations of SO2, NO and CO, but not NO2, PM2.5 or PMi0. The effect attributed to CO was the
6 smallest in magnitude when compared to those for SO2 and NO, but overall the effect estimates were
7 similar and all were statistically significant. Similarly, PM25, PMi0, NO2 and CO were positively
8 correlated with small artery reactive hyperemia, and the effect attributed to CO was the smallest in
9 magnitude when compared to those for PM25, PMi0, and NO2, but overall the effect estimates were
10 similar and all were statistically significant.
5.2.1.8. Blood Markers of Coagulation and Inflammation
11 Several studies have investigated the association between ambient CO and various blood
12 markers related to coagulation and inflammation. The main endpoints analyzed have been plasma
13 fibrinogen, B-type natriuretic peptide (BNP), endothelial function, Factor VII, C-reactive protein
14 (CRP), prothrombin, intercellular adhesion molecule (ICAM-1), and white blood cell count (WBC).
15 Delfino et al. (2008, 156390) measured blood plasma biomarkers in a panel of 29 nonsmoking,
16 elderly subjects with a history of coronary artery disease living in retirement communities in the Los
17 Angeles, CA air basin in order to identify associations with systemic inflammation. The blood
18 plasma biomarkers included CRP, fibrinogen, tumor necrosis factor-a (TNF- a) and its soluble
19 receptor-II (sTNF-RII), interleukin-6 (IL-6) and its soluble receptor (IL-6sR), fibrin D-dimer, soluble
20 platelet selectin (sP-selectin), soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble ICAM-
21 1, and myeloperoxidase (MPO). Overall, there were statistically significant associations for many of
22 the biomarker and pollutant combinations, with some of the strongest effects for CRP, IL-6 and
23 sTNF-RII with indoor and outdoor concentrations of NO2 and CO. Only the outdoor concentrations
24 indicated an effect of PM for these three biomarkers of inflammation. There was weaker evidence
25 for an effect of outdoor and indoor CO on the biomarker of platelet activation (sP-selectin), and
26 suggestive evidence for an effect of many of the air pollutants examined on fibrinogen, TNF- a,
27 sVCAM-1, sICAM-1, and MPO. Parameter estimates for fibrin D-dimer were close to zero for most
28 models. Overall, the results suggest that traffic related pollutants, including PM2 5, UFPs, OC and
29 CO lead to increases in systemic inflammation and platelet activation in elderly people with a history
30 of coronary artery disease.
31 Circulating levels of BNP are directly associated with cardiac hemodynamics and symptom
32 severity in patients with heart failure and serve as a marker of functional status. Wellenius et al.
September 2009 5-28 DRAFT - DO NOT CITE OR QUOTE
-------
1 (2007, 092830) examined the association of BNP levels with short-term changes in ambient
2 pollution levels among 28 patients with chronic stable heart failure and impaired systolic function.
3 The authors reported no association between any pollutant and measures of BNP at any lag.
4 Pekkanen et al. (2000, 013250) examined the association between daily concentrations of air
5 pollution and concentrations of plasma fibrinogen measured among 4,982 male and 2,223 female
6 office workers in Whitehall, London, U.K. between September 1991 and May 1993. Plasma
7 fibrinogen data were linked to ambient exposure to BS, PMi0, O3, NO2, SO2, and CO, where the
8 exposures were derived from data recorded at 5 fixed sites across London. There was a high
9 correlation between levels of CO and NO2 (r = 0.81) and more moderate correlations with PMi0
10 (r = 0.57) and SO2 (r = 0.61). The pollution data on the same day when the blood sampling was done
11 (lag 0) and on the 3 previous days (lags 1-3) were analyzed. Results showed that ambient CO at all
12 lags was significantly associated with an increase in plasma fibrinogen. Results were similar for NO2
13 while all other pollutants were not associated with an increase in plasma fibrinogen.
14 Liao et al. (2005, 088677) examined associations between various air pollutants and
15 hemostatic and inflammatory markers (fibrinogen, factor VIII-C, von Willebrand factor, serum
16 albumin, WBC) among 10,208 middle-aged males and females from the ARIC study. Exposure
17 estimates for ambient PMi0, NO2, SO2, O3 and CO were calculated for days 1-3 prior to the blood
18 sampling. A 0.5 ppm increment in 24-h CO concentration was significantly associated with
19 0.015 g/dL decrease in serum albumin among persons with a history of CVD. CO was not associated
20 with other hemostatic or inflammatory factors.
21 In Israel, Steinvil et al. (2008, 188893) examined WBC, fibrinogen, and CRP among 3,659
22 study subjects enrolled in the Tel-Aviv Sourasky Medical Center inflammation survey, in which
23 subjects lived <11 km from an ambient air pollution monitor. Air pollution data in the form of PMi0,
24 NO2, SO2, O3, and CO were derived from data recorded at fixed sites. The correlations coefficients
25 were high between CO and NO2 (r = 0.86) and PMi0 (r = 0.75). Exposures for lag days 1-7 were
26 analyzed and ambient CO had a significant negative effect on fibrinogen only among males.
27 Significant associations were reported for lag 0 (e.g., same day) and lags 2-5 with the decrease in
28 fibrinogen ranging from -5.5 mg/dL to -9.8 mg/dL per 0.5 ppm increase in 24-h CO concentration. A
29 similar negative effect for CO was observed on WBC among males only. The average CO exposure
30 over the week prior to the sampling yielded the largest reduction in WBC (-263 cells/(iL).
31 In a German study, Riickerl and colleagues (Ruckerl et al., 2006, 088754) recruited 57 non-
32 smoking male patients with coronary heart disease (CHD) who were scheduled for 12 subsequent
33 clinical visits where samples of blood were collected. The authors tested the primary hypothesis that
34 CRP would increase in association with a rise in air pollution levels. CRP is an acute phase protein
35 that increases during inflammatory processes in the body. Other markers of inflammation (serum
36 amyloid A [SAA]), cell adhesion (E-selectin, von Willebrand factor antigen [vWF], ICAM-1), and
September 2009 5-29 DRAFT - DO NOT CITE OR QUOTE
-------
1 coagulation (fibrinogen, factor VII [FVII], prothrombin fragment 1+2) were also examined. Ambient
2 air pollution in the form of PMi0, ultrafine particles (UFP), EC, NO2, and CO was monitored at one
3 central site and a 24-h avg immediately preceding the clinic visit (lag 0) and up to 5 days (lags 1-4)
4 was calculated for each patient. For CRP, the odds of observing concentrations above the 90th
5 percentile were 2.41 (95% CI: 1.23-5.02) in association with a 0.5 ppm increase in 24-h CO
6 concentration (lag 2). CO concentration during lags 1 and 2 was associated with observing ICAM-1
7 concentrations above the 90th percentile (OR: 2.41 [95% CI: 1.49-4.04]; OR: 3.17
8 [95% CI: 1.77-6.11], respectively). CO concentration during lags 0-3 was associated with a decrease
9 in FVII.
10 A similar study by Ruckerl and colleagues (2007, 156931) was conducted among 1,003 MI
11 survivors across six European cities (Athens, Greece; Augsburg, Germany; Barcelona, Spain;
12 Helsinki, Finland; Rome, Italy; Stockholm, Sweden). The study compared repeated measurements of
13 interleukin-6 (IL-6), CRP and fibrinogen with concurrent ambient levels of air pollution (particle
14 number count [PNC], PMi0, PM2.5, NO, NO2, O3, SO2, CO) from fixed sites across each city. Lags
15 0-1 and the 5-day mean prior to the blood sampling were analyzed and ambient CO was not
16 associated with any of the inflammatory endpoints.
17 Baccarelli et al. (2007, 090733) recruited 1,218 healthy individuals from the Lombardia region
18 in Italy and assessed whether blood coagulability is associated with ambient air pollution. The main
19 blood coagulability endpoints of interest were prothrombin time (PT) and activated partial
20 thromboplastin time (APTT), which are measures of the quality of the coagulation pathways,
21 assuming that, if shortened these measures would reflect hypercoagulability. Air pollution data
22 (PMio, O3, NO2, and CO) were obtained from 53 fixed stations across the Lombardia region, which
23 was divided into nine different study areas and a network average for each pollutant was calculated
24 across the available monitors within each of the nine study areas. The analyses examined air
25 pollution at the time of the blood sampling as well as averages for the 7 days prior and 30 days prior.
26 Results showed that ambient CO at the time of blood sampling was associated with a decrease in PT
27 (coefficient = -0.11 [95% CI: -0.18 to -0.05, p O.001), indicating hypercoagulability. However,
28 PMio and NO2 at the time of blood sampling were also associated with a decrease in PT and results
29 from multipollutant models were not reported. Acute phase reactants such as fibrinogen, and
30 naturally occurring anticoagulants such as antithrombin, protein C and protein S were examined and
31 none were associated with ambient air pollution.
32 Rudez et al. (2009, 193783) collected 13 consecutive blood samples within a 1-yr period and
33 measured light-transmittance platelet aggregometry, thrombin generation, fibrinogen and CRP in
34 40 healthy individuals in Rotterdam, the Netherlands. In general, air pollution increased platelet
35 aggregation as well as coagulation activity but had no clear effect on systemic inflammation.
36 Specifically, there were notable associations between maximal aggregation and CO, NO and NO2
September 2009 5-30 DRAFT - DO NOT CITE OR QUOTE
-------
1 and between late aggregation and CO. The effects for CO were the highest in magnitude and
2 persisted over most of the lag times investigated. There also was evidence of an increase in
3 endogenous thrombin potential and peak thrombin generation associated with CO, NO, NO2 and O3,
4 but no clear associations between PMi0 and peak height or lag time of thrombin generation. There
5 was no evidence for an effect of any of the air pollutants examined on CRP or fibrinogen levels.
6 These prothrombotic effects may partly explain the relationship between air pollution and the risk of
7 ischemic cardiovascular disease.
8 Ljungman et al. (2009, 191983) investigated the effect of CO and NO2 on inflammation in
9 certain genetic subpopulations of MI survivors. Specifically they examined whether IL-6 and
10 fibrinogen gene variants could affect plasma IL-6 response to CO or NO2. The study included
11 955 MI survivors from six European cities. This study provides evidence of gene-environment
12 interaction where IL-6 and fibrinogen gene polymorphisms modified the effects of CO and NO2 on
13 IL-6 levels in this panel of subjects with existing cardiovascular disease. Subjects with the
14 homozygous major allele genotypes for all 3 IL-6 polymorphisms examined showed larger IL-6
15 responses to increased CO, and there was evidence of a genetic interaction with NO2 for one of the
16 polymorphisms. Subjects with the homozygote minor allele genotype for 1 fibrinogen polymorphism
17 showed both a larger and clearer effect modification for the IL-6 response to increased CO compared
18 to the IL-6 polymorphisms. Similar magnitudes of effect modification were seen for NO2, but the
19 effect modification pattern was not statistically significant. A second fibrinogen polymorphism did
20 not modify the response to air pollution. Overall, this study provides evidence for the influence of
21 CO on IL-6 levels in subjects with genetic polymorphisms of the IL-6 and fibrinogen genes. In this
22 study, 16% of the subjects had a polymorphism combination that resulted in a statistically significant
23 gene-gene-environment interaction potentially implicating a higher risk of health effects from air
24 pollution in these patients with ischemic heart disease.
25 In summary, a growing number of studies provides some evidence of a link between CO
26 exposure and blood markers of coagulation and inflammation. The prothrombotic effects
27 characterized by many of the blood markers may partly explain the relationship between air
28 pollution and the risk of ischemic cardiovascular disease. The results of a recent gene-gene-
29 environment interaction study are particularly interesting. Table 5-6 summarizes the reviewed
30 studies.
September 2009 5-31 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-6 Summary of studies investigating the effect of CO exposure on blood markers of
coagulation and inflammation.
1
Study
Delfino et al. (2008, 1563901
Wellenius et al. (2007,
0928301
Pekkanen et al (2000, 0132501
Liao et al (2005, 0886771
Steinvil et al (2008, 1888931
Ruckerl et al (2006, 0887541
Rukerl et al (2007, 1569311
Baccarelli et al (2007, 0907331
Rudez et al. (2009, 1937831
Ljungman et al. (2009,
1919831
Location,
Sample Size
Los Angeles, CA
(n=29)
Boston, MA
(n=28)
London, U.K.
(n = 7205)
USA
(n = 10.208)
Tel-Aviv, Israel
(n = 3659)
Erfurt, Germany
(n = 57)
Six European cities
(n = 1003)
Lombardia region,
Italy
(n = 1218)
Rotterdam, the
Netherlands
(n=40)
Six European cities
(n=955)
Cardiac Endpoint
CRP, fibrinogen, TNF-
a, IL-6, fibrin D-dimer,
sP-selectin, sVCAM-1,
slCAM-1, MPO
BNP
Plasma fibrinogen
Fibrinogen, VI I-C,
WBC, albumin, vWF
CRP, fibrinogen,
WBC
CRP, SAA, cell
adhesions and
coagulation
IL-6, CRP, fibrinogen
PT, APTT, fibrinogen,
anticoagulants
Platelet aggregation,
thrombin generation,
fibrinogen, CRP
IL-6 and fibrinogen
polymorphisms
Upper CO
Concentrations
from AQS* in
ppm
98th%: 2.9
99th%:3.1
98th%: 0.75-2.22
99th%: 0.92-2.48
(24 h)
NA
98th%: 0.39-2.29
99th%: 0.43-2.66
(24 h)
NA
NA
NA
NA
NA
NA
CO Concentrations
Reported by Study Copollutants
Authors in ppm
Outdoor Mean: 0.71 (1 h) 03, N02, EC, OC, BC,
PM025,PMo25-25,
Indoor Mean: 0.78 (1 h) PM2s-io
Mean: 0.44 (24 h) PM25, S02, N02, 03, BC
Mean: 1.22 (24 h)
10th, 50th, 90th, Max: PM10, BS, 03, N02, S02
0.61,1.04,2.0,8.61
Mean: 1.4 (24 h)
PMio,03, N02,S02
Mean: 0.8
25th, 50th, 75th percentiles: PM10, 03, N02, S02
0.7, 0.8, 1.0
Mean: 0.45 (24 h)
PM10,PM25,UFP, EC, N02
Range: 0.10, 1.68
Mean: 0.29-1.48 (24 h) PM10, PM25, 03, N02, S02
Mean: 1.14-3.11
PMio,03, N02,S02
Max: 5.52-11.43
Median: 0.29 (24 h) PM10, NO, N02, 03
Mean: 0.25-1. 29 (24 h) N02, PM10, PM25
NA: Not Available
* includes range across individual monitors in study site; AQS data available for U.S. studies only
5.2.1.9. Hospital Admissions and Emergency Department Visits
Since the 2000 CO AQCD there have been a number of studies investigating the effect of
2 ambient CO on hospital admissions and ED visits for cardiovascular diseases. Some of these studies
3 have focused solely on one specific CVD outcome, and these studies are discussed first. This is
4 followed by a discussion of studies that investigated admissions for all CVD outcomes (e.g., non-
5 specific) or a variety of specific CVD outcomes.
September 2009
5-32
DRAFT - DO NOT CITE OR QUOTE
-------
Coronary Heart Disease
1 Ischemic heart disease (IHD), also known as CHD, is caused by inadequate circulation of the
2 blood to the heart muscle, which is a result of the coronary arteries being blocked by cholesterol
3 deposits. IHD can lead to sudden episodes such as MI ("heart attack") or death, as well as chronic
4 conditions such as angina pectoris (chest pain).
Ischemic Heart Disease
5 A number of studies have focused directly on hospitalizations for IHD. There is a lot of
6 variation among these studies with regard to methods employed and results reported. It should be
7 noted that within these studies IHD included MI and angina pectoris (ICD-9 codes 410-414; ICD-10
8 codes 120, 121-123, 124). Mann and colleagues (2002, 036723) investigated the modifying effect of
9 secondary diagnosis of arrhythmia and congestive heart failure (CHF) on the relationship between
10 hospital admissions for IHD (ICD-9: 410-414) and ambient air pollutants for the period of
11 1988-1995 in southern California. There were 54,863 visits analyzed and a 0.75 ppm increase in 8-h
12 max CO concentration was associated with a 2.69% (95% CI: 1.21-4.19) increase in same-day IHD
13 admissions among persons with a secondary diagnosis of CHF, a 2.23% (95% CI: 1.35-3.13)
14 increase among persons with a secondary diagnosis of arrhythmia, and a 1.21% (95% CI: 0.49-1.94)
15 increase among persons without either secondary diagnosis. Of all pollutants examined (PMi0, NO2,
16 O3, CO), only NO2 showed similar positive effects to CO and no multipollutant models were
17 analyzed. The correlation coefficients between CO and NO2 ranged from 0.64 to 0.86 across the
18 seven regions. This study indicated that people with IHD and accompanying CHF and /or arrhythmia
19 are a sensitive group in relation to the effects of ambient air pollution.
20 By using a time-series approach, ED visits for IHD (ICD-9: 410-414) in Montreal, Canada
21 (1997-2002) were examined in relation to ambient CO concentrations (lags 0 and 1) (Szyszkowicz,
22 2007, 193793). A total of 4,979 visits were analyzed and results showed significant positive effects
23 with a 0.5 ppm increase in 24-h CO concentration (lag 0) attributing to a 14.1% (95% CI: 5.8-20.6)
24 increase in daily ED visits among all patients. Stratified analyses showed that this effect was mostly
25 among male patients (19.8% [95% CI: 9.2-31.6]). NO2 was the only other pollutant examined, and it
26 too showed significant positive associations with ED visits for IHD for same-day exposure; however,
27 no multipollutant models were examined.
28 Lee and colleagues (2003, 095552) examined daily counts of hospital admissions for IHD in
29 Seoul, Korea for the period from December 1997 to December 1999. Single-day lags 0-5 were
30 analyzed and the lag period with the strongest association for each pollutant was chosen. For CO, lag
31 5 showed the strongest effect with a 1 ppm increase in 1-h maximum (max) CO concentration
32 associated with a daily increase in the number of hospital admissions for IHD; however, this was
33 only among patients 64+ yr of age (RR: 1.07 [95% CI: 1.01-1.13]). All other pollutants (PMi0, O3,
September 2009 5-33 DRAFT - DO NOT CITE OR QUOTE
-------
1 NO2) except SO2 showed similar significant effects and in a two-pollutant model with PMi0 the CO
2 effect attenuated toward the null.
3 Other studies have examined hospital admissions for IHD while investigating a broad group of
4 CVD outcomes. A study was conducted in Atlanta, GA, where over 4 million ED visits from
5 31 hospitals for the period 1993-2000 were analyzed (Study of Particles and Health in Atlanta
6 [SOPHIA]). Several articles have been published from this research with two examining
7 cardiovascular admissions in relation to CO concentrations. The first of these (Metzger et al., 2004,
8 044222) used a time-series design and analyzed a 3-day moving average over single-day lags 0-2 as
9 the a priori lag structure. Although of borderline statistical significance, CO was positively
10 associated with an increase in ED visits for IHD (RR 1.016 [95% CI: 0.999-1.034] per 1 ppm
11 increase in 1-h max CO concentration).
12 The second of these reports (Peel et al., 2007, 090442) examined the association of ambient air
13 pollution levels and cardiovascular morbidity in visits with and without specific secondary
14 conditions (e.g., comorbidity). Within a time-stratified case-crossover design using the same lag
15 structure already mentioned, the main results showed that a 1 ppm increase in 1-h max CO
16 concentration was associated with an increase in IHD among those without diabetes (OR: 1.023
17 [95% CI: 1.004-1.042]), and without CHF (OR: 1.024 [95% CI: 1.006-1.042]).
18 Two Australian studies have also examined associations between ambient CO concentrations
19 and increased hospital admissions for various CVD outcomes. The first of these studies (Barnett et
20 al., 2006, 089770) analyzed data from 5 of the largest cities in Australia (Brisbane, Canberra,
21 Melbourne, Perth, Sydney) and two New Zealand cities (Auckland, Christchurch) for the period
22 1998-2001. A time-stratified case-crossover design was employed and the age groups of 15-64 yr
23 and > 65 yr were analyzed for the 0-1 lag period. The pooled estimates across all cities showed that a
24 0.75 ppm increase in 8-h max CO concentration was associated with a 1.9% (95% CI: 0.7-3.2)
25 increase in admissions for IHD, but only among the elderly group (> 65 yr).
26 The second of the Australian studies (Jalaludin et al., 2006, 189416) examined ED visits for
27 CVD outcomes in the elderly (65+ yr) in Sydney for the period 1997-2001. Using a time-series
28 approach, single-day lags of 0, 1, 2, 3 and an average over lags 0 and 1 were examined. A 0.75 ppm
29 increase in 8-h max CO concentration (lag 0) was associated with increases in IHD emergency
30 department visits of 3.1% (95% CI: 1.3-4.9).
Angina Pectoris
31 In the current literature, only one study was identified that focused solely on angina pectoris
32 as an endpoint. Admissions data for angina pectoris were collected from 25 academic hospitals in
33 Tehran, Iran, and linked to ambient air pollution for the period of 19962001 (Hosseinpoor et al.,
34 2005, 087413). Using a time-series approach, single-day lags of 0-3 were analyzed and a 0.5 ppm
September 2009 5-34 DRAFT - DO NOT CITE OR QUOTE
-------
1 increase in 24-h avg CO concentration at lag 1 was associated with increased hospital admissions for
2 angina (OR: 1.005 [95% CI: 1.003-1.007). This result persisted in a multipollutant model that also
3 included NO2, PMi0, and O3 with CO being the only significant pollutant (OR: 1.005
4 [95% CI: 1.001-1.008]).
Myocardial Infarction
5 Linn et al. (2000, 002839) examined the association between ambient air pollution and
6 hospital admissions for cardiopulmonary illnesses in metropolitan Los Angeles for the years
7 1992-1995. Using a time-series approach, a 0.5 ppm increase in same-day 24-h avg CO
8 concentration was associated with a 2.0% increase in MI hospital admissions among people aged
9 >30 yr. When the analyses were stratified by season, no significant effects were observed (No
10 quantitative seasonal effects reported).
11 A time-series study in Denver, Colorado, investigated daily hospital admissions for various
12 CVD outcomes among older adults (>65 yr) across 11 hospitals (Koken et al., 2003, 049466). Data
13 between July and August for the period 1993-1997 were analyzed. Single-day lags 0-4 were
14 examined and CO showed no association with hospital admissions for MI (quantitative results were
15 not reported).
16 As part of the HEAPSS (Health Effects of Air Pollution among Susceptible Subpopulations)
17 study, Lanki et al. (2006, 089788) investigated the association between traffic-related exposure to air
18 pollutants and hospitalization for first acute myocardial infarction (AMI). Data were collected from
19 five European cities with either AMI registers (Augsburg, Barcelona), or hospital discharge registers
20 (Helsinki, Rome, Stockholm). Correlation coefficients between CO and NO2 ranged from 0.43 to
21 0.75 across the five cities, and for PMi0 the range was 0.21 to 0.56. Atotal of 26,854 hospitalizations
22 were analyzed and pooled estimates from all 5 cities showed that there was a weak positive
23 association with AMI hospitalizations and 24-h avg CO concentrations at lag 0 (RR: 1.014
24 [95% CI: 1.000-1.029] per 0.5 ppm increase), but more so when only using data from the three cities
25 (Helsinki, Rome, Stockholm) with hospital discharge registers (RR: 1.020 [95% CI: 1.003-1.035] per
26 0.5 ppm increase). When analyses were stratified by fatality and age, results showed that the CO
27 effect was significantly associated with fatal AMI among the <75-yr age group (RR: 1.080
28 [95% CI: 1.017-1.144), and with non-fatal AMI in the > 75-yr age group (RR: 1.044
29 [95% CI: 1.011-1.076).
30 Further analyses within the HEAPSS cohort were conducted using the event of cardiac
31 readmission among the first MI survivors (n = 22, 006) (Von Klot et al., 2005, 088070). The
32 readmissions of interest were those with primary diagnosis of AMI, angina pectoris, dysrhythmia,
33 and heart failure that occurred at least 29 days after the index event. Single-day lags 0-3 were
34 examined and pooled estimates from all 5 cities showed that a 0.5 ppm increase in same-day (lag 0)
September 2009 5-35 DRAFT - DO NOT CITE OR QUOTE
-------
1 CO was associated with an increase in cardiac (e.g., any of the diagnoses) readmissions (RR: 1.041
2 [95% CI: 1.003-1.076]) and this persisted in two-pollutant models that included either PMi0 or O3.
3 Correlation coefficients with CO ranged from 0.21 to 0.57 for PMi0 and 0.44 to 0.75 for NO2.
4 A study in Rome, Italy, also found an association between ambient CO and hospitalizations for
5 first episode MI among 6,531 subjects (January 1995-June 1997) (D'Ippoliti et al., 2003, 0743111 A
6 case-crossover design with stratification of time into separate months was used to select referent
7 days as the days falling on the same day of the week within the same month as the index day. CO
8 concentration was positively associated for lag 2 (OR: 1.019 [95% CI: 1.001-1.037]). The other
9 pollutants analyzed were NO2 and TSP, both of which exhibited a significant positive effect at lag 0.
10 TSP also showed a significant positive effect at lag 0-2 and when entered into a model with CO, the
11 CO effect did not persist.
12 The previously mentioned Australian and New Zealand study that analyzed data from seven
13 cities (Brisbane, Canberra, Melbourne, Perth, Sydney, Auckland, and Christchurch) for the period
14 1998-2001 also reported an association between CO and MI hospitalization (Barnett et al., 2006,
15 089770). The pooled estimates across all cities showed that a 0.75 ppm increase in 8-h max CO
16 concentration was associated with a 2.4% (95% CI: 0.6-4.1) increase in admissions for MI, but only
17 among older adults (> 65 yr). Table 5-7 shows a summary of the IHD hospital admission studies that
18 examined CO exposures.
19 In summary, the majority of studies reported significant increases in the daily number of
20 admissions for IHD, angina and MI in relation to CO exposures. In studies that stratified by age
21 groups and/or sex, the effects were larger among the elderly and males. Among the different lag
22 periods being examined, the associations were more commonly observed with same day CO (lag 0)
23 or an average over the same day and previous day (lag 0-1). Figure 5-2 shows the effect estimates
24 associated with daily admissions for various forms of IHD from selected studies.
September 2009 5-36 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location
Lag Effect Estimate
IHD
Metzger et al. (2004, 044222)
Peel et al. (2007. 090442)
Mann et al.(2002, 036723)
Mann et al.(2002, 036723)
Mann et al.(2002, 036723)
Barnett etal. (2006, 089770)
Barnett etal. (2006. 089770)
Jalaludin et al. (2006, 189416)
Szyszkowicz (2007, 193793)
Szyszkowicz (2007, 193793)
Szyszkowicz (2007, 193793)
Szyszkowicz (2007, 193793)
Szyszkowicz (2007, 193793)
Szyszkowicz (2007, 193793)
Lee etal. (2003. 095552)
Lee etal. (2003. 095552)
Atlanta, GA
Atlanta, GA
California, US
California, US
California, US
Australia, New Zealand
Australia, New Zealand
Sydney, Australia
Montreal, Canada
Montreal, Canada
Montreal, Canada
Montreal, Canada
Montreal, Canada
Montreal, Canada
Seoul, Korea
Seoul, Korea
0-2
0-2
0-3
0-3
0-3
0-1
0-1
0-1
0
o
n
0 >64yr
0
0
5 •
5
von Klot etal. (2005.088070)
Hosseinpoor et al. (2005, 087413)
Multicity, Europe
Tehran, Iran
0 —>
1
Linn et al. (Linn et al., 2000, 002839)
Barnett etal. (2006. 089770)
Barnett etal. (2006. 089770)
Lanki et al. (2006. 089788)
Lanki etal. (2006, 089788)
Lanki etal. (2006, 089788)
Lanki etal. (2006, 089788)
Lanki et al. (2006. 089788)
von Klot etal. (2005.088070)
D'lppoliti etal. (2003.074311)
D'lppoliti etal. (2003.074311)
D'lppoliti etal. (2003.074311)
D'lppoliti etal. (2003.074311)
Los Angeles, CA
Australia, New Zealand
Australia, New Zealand
Multicity, Europe
Multicity, Europe
Multicity, Europe
Multicity, Europe
Multicity, Europe
Multicity, Europe
Rome, Italy
Rome, Italy
Rome, Italy
Rome, Italy
0
0-1
0-1
0
0
0
0
0 -
n
0-2
0-2
0-2 —
0-2 —
•*-
-»-
•*
-*- sCHF
.*. sARR
*-15-64yr
•*- 65+ yr
-*- 65+ yr
^61 yr * M
All ages
s
males
ales
Angina
<-m 35+ yr
i
Ml
-»• All year
•*- 15-64yr
-»- 65+ yr
•*- >34yr, All cities
h- <75 yr, Non-fatal
— • — 75+ yr, Non-fatal
• 75+ yr, Fatal
-»— 18+ yr
— » — 18-64yr
•*— 65-74 yr
•*— 75+ yr
1
0.8 1.0 1.2
1.4
Figure 5-2 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for various froms of CHD. Effect estimates have been standardized to
a 1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-37
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-7 Summary of CHD hospital admission studies.1
Study
*""•" Ixa±S
Copollutants Ex^ed
Upper CO
Concentrations from
AQS* in ppm
CO Concentrations
Reported by Study
Authors in ppm
STUDIES THAT FOCUSED SOLELY ON CHD
Mann et al. (2002, 0367231
Szyszkowicz (2007,
1937931
Lee etal. (2003, 0955521
Lanki et al. (2006, 08978812
von Klot et al. (2005,
08807012
D'lppoliti etal. (2003,
074311 12
Hosseinpoor et al. (2005,
08741 312
Southern
California IHD
(1988-1995)
Montreal,
Canada |HD
(1997-2002)
Seoul, Korea
IHD
(1997-1999)
5 European cities M|(first
(1992-2000) acute)
5 European cities M|]Angina]
(1992-2001) Cardiac*
Rome, Italy
Ml
(1995-1997)
Tehran, Iran
Angina
(1996-2001)
PM10, N02, 03 0,1,2, 2-4ma
N02 0,1
PM10,N02,S02, Q12345
U3
PM,o, N02, 03, nlo,
PNC U'U'J
PM,o, N02, 03, nlo,
PNC U'U'J
TSP, N02, S02 0,1,2,3,4,0-2
PM,o, N02, S02, n, -,
03 0'V'J
98th%: 1.0-13.8
99th%: 1.3-15.9 (8 h)
NA
NA
NA
NA
NA
NA
Mean: 2.07 (8h)
Mean: 0.5 (24 h)
Mean: 1.8
Highest city was Rome.
25th = 1.5
75th = 2.9 mg/m3
Mean: highest city was
Rome: 1.9 (24 h)
Mean: 3.8 (24 h)
Mean: 9.4 (24 h)
STUDIES THAT EXAMINED CHD OTHER CVDS
Metzger et al.(2004,
044222)
Peel etal. (2007,090442)
Barnett etal. (2006,
089770)
Jalaludin et al. (2006,
189416)
Linn etal. (2000, 002839)
Kokenetal.(2003,
049466)
Atlanta, GA IHD, All
CVD, CD,
(1993-2000) CHF, PVCD
Atlanta, GA IHD, All
CVD, CD,
(1993-2000) CHF, PVCD
Australia and ^p ^ /^
New Zealand CVD, CA,
(1998-2001) strok'e
Sydney, ^p ^|
Australia CVD' ^ote,
(1997-2001) Cardiac
Los Angeles,
CA Ml, All CVD,
CHF, CA, OS
(1992-1995)
Denver, CO MI.CAth,
PHD, CD,
(1993-1997) CHF
™<°.0N°2. 0-2ma
OW2, Us
™io,N02, 0-2ma
OU2, Vs
PM10,N02,03 Lag 0-1
PM10 N02, 0,1,2,3,0-1
o(J2, U3
PM10,N02,03 0
PM10,N02, 01234
S02,03 0,1,2,3,4
98th%: 5.0-5.1
99th%: 5.5-5.9(1 h)
98th%: 5.0-5.1
99th%: 5.5-5.9(1 h)
NA
NA
98th%: 1.0-7.8
99th%:1.1-8.3(24h)
98th%: 1.2-2.0
99th%: 1.3-2.0 (24 h)
Mean: 1.5(1 h)
Mean: 1.5(1 h)
Mean: (8 h)
0.5-2.1
Mean: 0.82 (8 h)
Mean: (24 h)
Winter 1.7, Spring 1.0,
Summer 1.2, Fall 2.1
Mean: 0.9 ppm (24 h)
1Cardiac =AMI, angina, dysrhythmia, orHF; CA = Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhythmias; CHF = Congestive heart failure; PHD = Pulmonary heart disease;
OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average.
2These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which assumes standard atmosphere
and temperature. NA: Not Available; * includes range across individual monitors in study site; AQS data available for U.S. studies only
September 2009
5-38
DRAFT - DO NOT CITE OR QUOTE
-------
Stroke
1 A stroke is the result of either the blood supply to the brain being blocked (e.g., embolism),
2 which refers to an ischemic stroke (80% of strokes), or the occurrence of a burst blood vessel or
3 hemorrhaging, referred to as a hemorrhagic stroke. Hemorrhagic stroke has two main groupings;
4 intracerebral hemorrhagic stroke (10% of strokes), which is when a blood vessel in the brain leaks,
5 and subarachnoid hemorrhage (3% of strokes), which is bleeding under the outer membranes of the
6 brain. The third type of stroke is a transient ischemic attack (TIA), or mini-stroke, which has the
7 same early symptoms as a normal stroke but the symptoms disappear within 24 h, leaving no
8 apparent deficits.
9 A small number of air pollution studies have investigated hospital admissions for the three
10 main forms of stroke with the majority reporting positive associations with ambient CO and lag
11 periods between 0 and 3 days.
12 A U.S. study across 9 cities investigated hospital admissions for ischemic and hemorrhagic
13 stroke among Medicare beneficiaries aged 65+ yr of age (155,503 ischemic and 19,314 hemorrhagic
14 admissions from the ED) (Wellenius et al., 2005, 088685). Single-day lags 0-2 were examined and
15 based on a pooled estimate, same-day CO (lag 0) was associated with an increase in admissions of
16 1.98% (95% CI: 0.86-3.12) per 0.5 ppm increase in 24-h CO concentration) for ischemic stroke
17 admissions but not hemorrhagic stroke admissions (-1.14%, 95% CI: -3.40 to 1.18). All other
18 pollutants examined (PMi0, NO2, SO2) were associated with an increase in ischemic stroke
19 admissions, but not hemorrhagic stroke admissions.
20 Villeneuve and colleagues (2006, 090191) studied ED visits for hemorrhagic strokes, acute
21 ischemic strokes and transient ischemic attacks among individuals 65+ yr of age at 5 hospitals within
22 the Edmonton area in Canada between April 1992 and March 2002 (12,422 visits). Within a time-
23 stratified case-crossover design the analyses were stratified by two seasonal groups (October-March
24 and April-September) and CO only had an effect on ischemic stroke during April-September. A
25 0.5 ppm increase the CO concentration for a 3-day avg across lags 0-2 was associated with a 32%
26 increase in risk (OR: 1.32 [95% CI 1.09-1.60]). CO had no effect on any other stroke subtype. In
27 two-pollutant models the CO effect on ischemic stroke persisted after controlling for PMi0, PM2.5,
28 SO2, and O3. When all seasons and all strokes were combined there was no statistically significant
29 association between all the pollutants examined and increased admissions for stroke.
30 In Kaohsiung City, Taiwan, CO averaged over lags 0-2 was associated with increased
31 admissions for stroke across 63 hospitals (Tsai et al., 2003, 080133). From 1997-2000 a total of
32 23,179 admissions were analyzed and on warm days (> 20°C) the odds ratios for primary
33 intracerebral hemorrhage and ischemic stroke were 1.39 (95% CI: 1.16-1.66) and 1.39
34 (95% CI: 1.25-1.53) respectively for a 0.5 ppm increase in 24-h CO concentration. For the same
September 2009 5-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 increase in CO on cool days (<20°C) the odds ratios were 1.33 (95% CI: 0.38-2.55) for intracerebral
2 hemorrhage and 2.68 (95% CI: 1.59-4.49) for ischemic stroke. These results persisted in two-
3 pollutant models that included PMi0, SO2, and O3, but did not persist when controlling for NO2.
4 Earlier research conducted in metropolitan Los Angeles examined hospital admissions for
5 cardiopulmonary illnesses from 1992-1995 (Linn et al., 2000, 002839). Using a time-series
6 approach, a 0.5 ppm increase in 24-h CO concentration (lag 0) was associated with a 2.18%
7 (95% CI: 1.73-2.62) increase in occlusive (ischemic) stroke hospital admissions among people aged
8 >30 yr. When the analyses were stratified by season there was a 1.8% increase during winter, a
9 4.55% increase during summer, and a 1.6% increase during fall (results for spring were not
10 reported).
11 A study in Taipei, Taiwan analyzed 8,582 emergency admissions for cerebrovascular diseases,
12 hemorrhagic stroke, ischemic stroke, and all strokes during 1997-2002 (Chan et al., 2006, 090193).
13 Single-day lags 0-3 were analyzed and a 0.75 ppm increase in 8-h max CO concentration (lag 2) was
14 associated with an increase in cerebrovascular diseases (OR: 1.03 [95% CI: 1.01-1.05]) and all
15 strokes (OR: 1.03 [95% CI: 1.01-1.05]). These results persisted in two- and three-pollutant models
16 that included O3 and PMi0. There was no association with individual ischemic or hemorrhagic
17 stroke. CO was moderately correlated with PMi0 (r = 0.47) and PM2.5 (r = 0.44), and the correlation
18 was higher with NO2 (r = 0.77).
19 The only time-series study that focused specifically on stroke hospital admissions that did not
20 report a significant association with ambient CO was conducted in Dijon, France (Henrotin et al.,
21 2007, 093270). Hospital admissions for different types of first-ever stroke (e.g., ischemic,
22 hemorrhagic) among subjects over 40 yr of age were analyzed for the period of 1994-2004. A bi-
23 directional case-crossover study design was employed where single-day lags of 0-3 were examined
24 and CO had no significant association across all lags. This was also the case when the analyses were
25 stratified by gender and types of ischemic stroke (large arteries, lacunar, cardioembolic, transient).
26 Of all pollutants examined (PMi0, NOX, O3, SO2, CO) only O3 showed a significant effect.
27 Two Australian studies examined associations between ambient CO and hospital admissions
28 for various CVDs. The first of these studies analyzed data from five of the largest cities in Australia
29 (Brisbane, Canberra, Melbourne, Perth, Sydney) and two New Zealand cities (Auckland,
30 Christchurch) for the period 1998-2001 (Barnett et al., 2006, 089770). A time-stratified case-
31 crossover design was employed and the age groups of 15-64 yr and > 65 yr were analyzed for the 0-
32 1 lag period (average over lag 0 and 1). The pooled estimates across all cities showed that CO had no
33 effect on stroke admissions (quantitative results not reported).
34 The second of the Australian studies examined ED visits for CVDs in older adults (65+ yr) in
35 Sydney for the period from 1997-2001 (Jalaludin et al., 2006, 189416). Using a time-series
36 approach, single-day lags of 0-3 and an average over lags 0 and 1 (e.g., lag 0-1) were examined and
September 2009 5-40 DRAFT - DO NOT CITE OR QUOTE
-------
1 CO showed no effect on stroke ED visits. When the analyses were stratified by cool and warm
2 periods a 0.75 ppm increase in 8-h max CO concentration during the cool period was associated with
3 a 3.8% (95% CI: 0.76-6.94) increase in stroke ED visits.
4 Figure 5-3 shows the effect estimates associated with daily admissions for stroke from selected
5 studies. Table 5-8 shows a summary of the stroke hospital admission studies that examined CO
6 exposures.
7 In summary, there was some evidence that increased ambient CO concentrations were
8 associated with an increase in the number of hospital admissions for stroke. The largest positive
9 effects came from the Taiwan study in Kaohsiung (Tsai et al, 2003, 080133) with slightly larger
10 effects during the warmer period (>20°C). Similarly, in the Canadian study by Villeneuve and
11 colleagues (2006, 090191) there was a stronger effect during the warmer period (April-September).
September 2009 5-41 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location
Lag
Effect Estimate
Ischemic Stroke
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Tsaietal. (2003. 0801 33)
Tsaietal. (2003. 0801 33)
Linn etal. (2000, 002839)
Chan etal. (2006. 0901 93)
Henrotin etal. (2007.093270)
Multicity, US
Edmonton, Canada
Edmonton, Canada
Edmonton, Canada
Edmonton, Canada
Edmonton, Canada
Edmonton, Canada
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Los Angeles, CA
Taipei, Taiwan
Dijon, France
0-2 ^ 65+ yr
0-2 IS -f- 65+ yr, All seasons
0-2 IS
0-2 IS -.
0-2 ClSn
0-2 CIS-i
0-2 CIS i
0-2
0-2
0
1
3
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Te^i at •ai /onm non*! T3\
isai et al. (^uuo, Uouioo)
Tsaietal. (2003. 0801 33)
Chan etal. (2006. 0901 93)
Henrotin etal. (2007.093270)
Multicity, US
Edmonton, Canada
Edmonton, Canada
Edmonton, Canada
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Taipei, Taiwan
Dijon, France
0-2
0-2 •<
0-2
0-2 -*i
n o
U-£
0-2
1 -<
1
Chan etal. (2006. 0901 93)
Jalaludin et al. (2006, 189416)
Taipei, Taiwan
Sydney, Australia
2
— * — 65+ yr, April-September
• 65+yr, October-March
65+ yr, All seasons
- 65+ yr, April-September
65+yr, October-March
-•- >20°CTemp
'
»-
Hemorrhagic Stroke
65+yr
• 65+ yr, All seasons
— • 65+ yr, April-September
65+ yr, October-March
• -"n"O T r- r
20°CTemp
-
Stroke (non-specific)
i
0-1 k
i
05 1.0 15 ZO 25
15 4,0 4.5
Figure 5-3
Summary of effect estimates (95% confidence intervals) associated with ED visits
and hospital admissions for stroke. Effect estimates have been standardized to a
1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations
IS=ischemic stroke, CIS=cerebral ischemic stroke.
September 2009
5-42
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-8 Summary of stroke hospital admission studies.1
Type Of . Upper CO CO Concentrations
Study Location Stroke Copollutants examined Concentrations Reported by Study
Examined examinee from AQS* jn ppm Authors in ppm
STUDIES THAT FOCUSED SOLELY ON STROKE
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)
Tsaietal. (2003. 0801 33)
Chan etal. (2006. 0901 93)
Henrotin et al. (2007. 093270)2
9 cities, USA
(1993-1999)
Edmonton,
Canada
(1992-2002)
Kaohsiung, Taiwan
(1997-2000)
Taipei, Taiwan
(1997-2002)
Dijon, France
(1994-2004)
STUDIES THAT EXAMINED STROKE AMONG
Linn etal. (2000, 002839)
Barnett etal. (2006. 089770)
Jalaludin et al. (2006, 189416)
Los Angeles, CA
(1992-1995)
Australia and New
Zealand
(1998-2001)
Sydney, Australia
(1997-2001)
Isch, Hem
Isch.Hem.TIA
Isch, Hem
All, Isch, Hem
Isch, Hem
OTHER CVDS
Isch
All
All
PM10,N02,S02
N02,S02,03
PM10,N02,S02,
03
PM10,N02,S02,
03
PM10,NOX,S02,
03
PM10,N02,03
PM10,N02,03
PM10,N02,S02,
03
0,1,2
0,1,0-2
0-2
0,1,2,3
0,1,2,3
LagO
Lag 0-1
0,1,2,3,0-1
98th%: 0.9-5.9
99th%: 1.2-7.1 (24 h)
NA
NA
NA
NA
98th%: 1.0-7.8
99th%: 1.1-8.3 (24 h)
NA
NA
25th, 50th, 75th
percentiles:
0.73,1.02,1.44
Mean: 0.8 (24 h)
Mean: 0.79 (24 h)
Mean:1.7(8h)
Mean: 0.59 (24 h)
Mean: (24 h)
Winter 1.7, Spring 1.0,
Summer 1.2, Fall 2.1
Mean:(8h)
0.5-2.1
Mean: 0.82 (8h)
1lsch = Ischemic; Hem = Hemorrhagic; TIA= transient ischemic attack
2These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which assumes standard atmosphere and
temperature. NA: Not Available; * includes range across individual monitors in study site; AQS data available for U.S. studies only
Congestive Heart Failure
1 Heart failure (HF) is a condition in which the heart is unable to adequately pump blood to the
2 rest of the body. It does not refer to the cessation of the heart, but more to the inability of the heart to
3 operate at an optimal capacity. HF is often called congestive heart failure (CHF), which refers to
4 when the inadequate pumping leads to a buildup of fluid in the tissues. The underlying causes of
5 CHF are hypertension, CAD, MI, and diabetes.
6 Wellenius and colleagues (2005, 087483) examined the rate of hospitalization for CHF among
7 55,019 Medicare recipients (aged > 65 yr) residing in Allegheny County, PA, during 1987-1999. A
8 time-stratified case-crossover design was employed and single-day lags of 0-3 were analyzed and a
9 1 ppm increase in 1-h max CO concentration on the same-day (lag 0) was associated with a 9.31%
10 (95% CI: 6.77-11.92) increase in the rate of hospitalization for CHF. This result persisted in two-
11 pollutant models that included PM10, NO2, O3, and SO2. CO was moderately correlated with SO2
12 (r = 0.54) and PM10 (r = 0.57) and more highly correlated with NO2 (r = 0.70).
September 2009
5-43
DRAFT - DO NOT CITE OR QUOTE
-------
1 Another U.S. study recruited 125 patients diagnosed with CHF who were admitted to Johns
2 Hopkins Bayview Medical Center in Baltimore, MD (Symons et al., 2006, 091258). The patients
3 were interviewed after admission through the ED during their stays in overnight wards. The
4 interview was designed to collect information about symptom onset, health conditions, and factors
5 related to air pollution exposure. Various lag periods (single day and cumulative days 0-3) prior to
6 the onset of symptoms were analyzed and although the focus of this study was exposure to PM25, of
7 all the pollutants examined (PM25, CO, NO2, O3) only 8-h max CO concentration at lag 2 was
8 significantly associated with the onset of CHF symptoms (OR: 1.68 [95% CI: 1.28- 2.80]).
9 Earlier research conducted in metropolitan Los Angeles, CA examined hospital admissions for
10 cardiopulmonary illnesses 1992-1995 (Linn et al., 2000, 002839). Using a time-series approach, a
11 0.5 ppm increase in same-day 24-h avg CO concentration was associated with a 1.25% increase in
12 CHF hospital admissions among people aged >30 yr. When the analyses were stratified by seasons
13 only summer showed a significant increase (3.7%); however, the study did not report the results for
14 the other seasons.
15 A time-series study in Denver, Colorado, investigated daily admissions for various CVDs
16 among older adults (>65 yr) across 11 hospitals (Koken et al., 2003, 049466). Single-day lags 0-4
17 were examined and an increase of 0.5 ppm in 24-h avg CO concentration for lag 3 was associated
18 with an 18% (95% CI: 0.2-39.3) increase in risk of hospitalization for CHF.
19 As stated earlier, a study was conducted in Atlanta, GA, where over 4 million ED visits from
20 31 hospitals for the period 1993-2000 were analyzed (Metzger et al., 2004, 044222). A time-series
21 design was used and a 3-day moving average over single-day lags 0-2 as the a priori lag structure
22 was analyzed. Results showed that 1-h max CO concentration was not associated with an increase in
23 ED visits for CHF (RR: 1.010 [95% CI: 0.988-1.032] per 1 ppm increase). When the analyses
24 examined the same CVDs among those with and without specific secondary conditions
25 (e.g., comorbidity) 1-h max CO concentration was associated with an increase in ED visits for CHF
26 only among those with COPD (OR: 1.058 [95% CI: 1.003-1.115] per 1 ppm increase) (Peel et al.,
27 2007, 090442).
28 In Kaohsiung city, Taiwan, a study analyzed 13,475 admissions for CHF across 63 hosptials
29 for the period 1996 through 2004 (Lee et al., 2007, 093271). A 0.5 ppm increase in 24-h avg CO
30 concentration averaged over lag days 0-2 was positively associated with CHF hospital admissions on
31 cool days (<25°C) (OR: 1.70 [95% CI: 1.43-2.01) with a slightly weaker effect on warm days
32 (>25oC) (OR: 1.32 [95% CI: 1.15-1.55]). These results persisted in two-pollutant models that
33 included PMi0, SO2, O3, and models with NO2 only on warmer days, not with NO2 on cooler days.
34 A case-crossover analysis was undertaken to examine the association between levels of
35 ambient air pollutants and hospital admissions for CHF among individuals residing in Taipei, Taiwan
36 from 1996-2004 (Yang, 2008, 157160). During the 9 yr of the study, there were 24.240 CHF hospital
September 2009 5-44 DRAFT - DO NOT CITE OR QUOTE
-------
1 admissions for the 47 hospitals in Taipei. The analyses were stratified by temperature, either warm
2 days (>20 C; n = 2325 d) or cool days (<20 C; n = 963 d). The number of CHF admissions was
3 associated with concentrations of PMi0, NO2, CO and O3 on warm days, however on cool days, the
4 positive effects on increased CHF admissions remained positive, though were diminished for NO2
5 and CO, and disappeared completely for PMi0 and O3 concentrations. In two-pollutant models, CO
6 remained statistically significant after the inclusion of PMi0, SO2 or O3 on warm days. On cool
7 days, the effects associated with CO remained positive, but were no longer statistically significant
8 after the inclusion of PMi0, SO2, or NO2, but became statistically significant and negative after the
9 inclusion of O3 in the model (see Figure 5-6).
10 Figure 5-4 shows the effect estimates for associations between CO and daily admissions for
11 CHF from selected studies. Table 5-9 summarizes the CHF hospital admission studies that examined
12 CO exposures.
13 In summary, many of the studies that examined associations between ambient CO
14 concentrations and daily hospital admissions for CHF reported positive associations at lags of
15 0-3 days.
September 2009 5-45 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location Lag
Effect Estimate
Metzger et al. (2004, 044222) Atlanta, GA
Peel et al. (2007, 090442) Atlanta, GA
Yang (2008,157160)
Yang (2008,157160)
0-2
0-2
Wellenius et al.(2005,087483) Pittsburgh, PA 0
Linn etal. (2000, 002839) Los Angeles, CA 0
Kokenetal. (2003.049466) Denver, CO
Symons et al. (2006,091258) Baltimore, MD 0-3
Lee etal. (2007.093271) Kaohsiung,Taiwan 0-2
Lee et al. (2007,093271) Kaohsiung, Taiwan 0-2
Taipei, Taiwan 0-2
Taipei, Taiwan 0-2
65+yr
65+yr
<25°C Temp
>25°CTemp
<20°C Temp
-*- >20°CTemp
I 1 1 1 1 1 1 1 1—
0.4 0.6 12 1.3 2.0
Figure 5-4 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CHF. Effect estimates have been standardized to a 1 ppm
increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h max CO
concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-46
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-9 Summary of CHF hospital
Study
Location
Endpoints
Examined
admission studies.
Laas Upper CO
Copollutants c " . Concentrations
bxammea fr0m AQS* in ppm
CO Concentrations
Reported by Study
Authors in ppm
STUDIES THAT FOCUSED SOLELY ON HF
Wellenius et al. (2005,
087483)
Symonsetal.(2006,
091258)
Lee etal. (2007,
093271)
Yang (2008, 157160)
Pittsburgh, PA
(1987-1999)
Baltimore, MD
(2002)
Kaohsiung,
Taiwan
(1996-2004)
Taipei, Taiwan
(1996-2004)
CHF
CHF
CHF
CHF
PM
SO
PM
PM
SO
PM
SO
10uN°2' 0,1,2,3
2, O3
2.5,N02,03 0,1,2,3
i.,oN02, Q.2
,o,N02, 0.2
2, U3
98th%:
99th%:
98th%:
99th%:
NA
NA
0.9-5.9
1.2-7.1 (24 h)
1 .2-1 .3
1 .51 (24 h)
Mean : 1.03 (24 h)
Mean: 0.4 (24 h)
Mean: 0.76 (24 h)
Mean: 1.26 (24 h)
STUDIES THAT EXAMINED HF AMONG OTHER CVDS
Linn et al. (2000,
002839)
Koken etal. (2003,
049466)
Metzger et al. (2004,
044222)
Peel et al. (2007,
090442)
Los Angeles, CA
(1992-1995)
Denver, CO
(1993-1997)
Atlanta, GA
(1993-2000)
Atlanta, GA
(1993-2000)
CHF, Ml, All
CVD, CA, OS
CHF, Ml, CAth,
PHD, CD
CHF, IHD.AII
CVD, CD,
PVCD
CHF, IHD.AII
CVD, CD,
PVCD
PM
PM
SO
PM
SO
PM
SO
10,N02,03 0
1°'nN°2' 0,1,2,3
2, O3
<°.nN02< 0-2ma
2, O3
<°'0N3°2' 0-2ma
98th%:
99th%:
98th%:
99th%:
98th%:
99th%:
98th%:
99th%:
1.0-7.8
1. 1-8.3 (24 h)
1.2-2.0
1. 3-2.0 (24 h)
5.0-5.1
5.5-5.9 (1 h)
5.0-5.1
5.5-5.9 (1 h)
Mean: (24 h)
Winter 1.7; Spring 1.0
Summer 1.2; Fall 2.1
Mean: 0.9 (24 h)
Mean 1 .5 (1 h)
Mean 1 .5 (1 h)
*Cardiac = AMI, angina, dysrhythmia, or HF; CA = Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhythmias; CHF = Congestive heart failure; PHD = Pulmonary heart disease;
OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average. NA: Not Available; * includes range across individual monitors in study site; AQS data
available for U.S. studies only
Cardiovascular Diseases
1 The following section reviews studies that have investigated the effect of CO on ED visits and
2 hospital admissions for all CVD outcomes (e.g., non-specific). Several of these studies also
3 examined specific CVDs and were briefly discussed in previous sections.
4 A multicity time-series studies was conducted to estimate the risk of CVD hospitalization
5 associated with short-term CO exposure in 126 U.S. urban counties from 1999-2005 for over 9
6 million Medicare enrollees 65 yr old and older (Bell et al., 2009, 193780). The analyses yielded
7 positive associations between same day CO concentration and increased risk of hospitalization for
8 total CVD outcomes, which remained positive and statistically significant, but were attenuated, with
9 copollutant adjustment, especially NO2 (see Figure 5-6). Overall, a 1 ppm increase in same day 1-h
10 max CO was associated with a 1.010 (95% PI: 1.008-1.011) increase in risk of CVD admissions.
11 After adjustment for NO2, the estimate was attenuated to 1.005 (95% PI: 1.004-1.007). For most
12 cause-specific CVD hospitalizations (IHD, heart rhythm, CHF, cerebrovascular) associations were
September 2009
5-47
DRAFT - DO NOT CITE OR QUOTE
-------
1 positive and statistically significant for same day CO concentration adjusted for same day NO2.
2 Cause-specific effect estimates were not presented for CO alone (without adjustment for NO2).
3 As discussed earlier, a study was conducted in Atlanta, GA where over 4 million ED visits
4 from 31 hospitals for the period 1993-2000 were analyzed (SOPHIA). Several articles have been
5 published from this research with three examining cardiovascular admissions in relation to CO
6 exposures. The first of these used a time-series design and analyzed a 3-day moving average over
7 single-day lags 0-2 as the a priori lag structure (Metzger et al, 2004, 044222). Results showed that a
8 1 ppm increase in 1-h max CO concentration was associated with an increase in daily ED visits for
9 all CVDs (RR: 1.017 [95% CI: 1.008-1.027]). This persisted in two-pollutant models that included
10 NO2andPM2.5.
11 The second of these publications examined the association of ambient air pollution levels and
12 cardiovascular morbidity in visits with and without specific secondary conditions (Peel et al., 2007,
13 090442). Within a time-stratified case-crossover design, a 3-day ma over single-day lags 0-2 was
14 used as the a priori lag structure. Results from the case-crossover analyses on all cardiovascular and
15 peripheral vascular and cerebrovascular disease were similar to the time-series results presented
16 earlier. Results from the various comorbidity analyses are presented in Table 5-10. Similar to the
17 results from the earlier publication, CO was mostly associated with peripheral vascular and
18 cerebrovascular disease (PVCD) among those with and without the comorbidities, except among
19 those with CHF. Overall, there is limited, if any, evidence of susceptibility to the effects of CO
20 concentration for those with comorbid conditions.
September 2009 5-48 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-10 Association of ambient air pollution levels and cardiovascular morbidity in visits with
and without specific secondary conditions.
Co-morbidity
IHD
Dysrhythmias
PVCD
CHF
HYPERTENSION
-With
- Without
1.007(0.978-1.037)
1 .022 (1 .000-1 .043)
1.065(1.015-1.118)
1.008(0.988-1.029)
1.038(1.004-1.074)
1.027(1.002-1.054)
1.037(0.997-1.079)
1.010(0.985-1.037)
DIABETES
-With
- Without
0.985(0.945-1.027)
1 .023 (1 .004-1 .042)
1.058(0.976-1.146)
1.014(0.995-1.034)
1.065(1.012-1.121)
1.025(1.003-1.048)
1.020(0.975-1.067)
1.018(0.993-1.044)
COPD
-With
- Without
0.996(0.938-1.057)
1.018(1.000-1.036)
0.972 (0.878-1 .077)
1.018(0.999-1.038)
1.113(1.027-1.205)
1.026(1.004-1.047)
1.058(1.003-1.115)
1.011(0.987-1.036)
CHF
-With
- Without
0.956(0.907-1.007)
1 .024 (1 .006-1 .042)
1.065(0.968-1.173)
1.015(0.996-1.034)
1.072(0.981-1.172)
1.029(1.008-1.051)
-
-
DYSRHYTHMIAS
-With
- Without
1.028(0.985-1.072)
1.014(0.995-1.033)
-
-
PVCD - peripheral vascular and cerebrovascular disease, IHD = ischemic heart disease, CHF = congestive heart failure.
The third study utilizing the SOPHIA data extended
2004 (Tolbert et al., 2007, 090316) and focused on two lar;
1.072(1.011-1.138)
1 .026 (1 .004-1 .048)
1.004(0.960-1.051)
1.023(0.998-1.049)
Source: Peel et al. (2007, 0904421
the time period to include 1993 through
ge outcome groups: a respiratory diseases
2
3 group and a cardiovascular diseases group. The combined cardiovascular case group included the
4 following groups of primary ICD-9 diagnostic codes: IHD (410-414), cardiac dysrhythmias (427),
5 CHF (428), and peripheral vascular and cerebrovascular disease (433-437, 440, 443-445, 451-453).
6 Results showed that a 1 ppm increase in 1-h max CO concentration was associated with an increase
7 in daily ED visits for all CVDs (RR: 1.016 [95% CI: 1.008-1.024]). CO was the strongest predictor
8 of CVD effects in models with two-pollutant combinations of NO2, CO and total carbon, as well as
9 in a model including all three pollutants.
10 Earlier research conducted in Los Angeles, CA, showed that a 0.5 ppm increase in same-day
11 24-h avg CO concentration was associated with a 1.6% increase in CVD hospital admissions among
12 people aged >30 yr (Linn et al., 2000, 002839). When the analyses were stratified by season the
13 significant CO effect was strongest during winter (1.9% increase) followed by summer (1.8%) and
14 fall (1.4%) with no effect in spring.
15 In contrast to other North American studies, a study in Spokane, WA, did not find an
16 association between CO (lags of 1-3 days) and an increase in the number of daily cardiac hospital
17 admissions (quantitative results not reported) (Slaughter et al., 2005, 073854). Similarly, a time-
September 2009
5-49
DRAFT - DO NOT CITE OR QUOTE
-------
1 series study in Windsor, Ontario, did not find an association between ambient CO and daily hospital
2 admissions for CVDs (defined as HF, IHD, or dysrhythmias) (Fung et al, 2005, 074322). A total of
3 11,632 cardiac admissions were analyzed for the period of 1995-2000. The lag periods analyzed in
4 this study were lag 0 (same-day), a 2-day avg (lag 0-1), and a 3-day avg (lag 0-2). For a 1 ppm
5 increase in 1-h max CO concentration the mean percent change in daily admissions for the <65 age
6 group (lag 0) was -2.6 (95% CI: -6.2 to 3.3); and for the 65+ age group, 0.4 (95% CI: -1.9 to 2.7).
7 The authors reported moderate to low correlations with NO2 (r = 0.38), PMi0 (r = 0.21) and SO2
8 (r = 0.16).
9 Two case-crossover studies in Taiwan reported an association between ambient CO and
10 hospital admissions for CVDs. In Taipei, a total of 74,509 CVD admissions from 47 hospitals for the
11 period of 1997-2001 were analyzed (Chang et al., 2005, 080086). An increase of 0.5 ppm in 24-h
12 avg CO concentration (average over lags 0-2) during warmer periods (> 20°C) was associated with
13 an increase in daily hospital admissions (OR: 1.09 [95% CI: 1.065-1.121) but not cooler periods
14 (<20°C) (OR: 0.98 [95% CI: 0.93-1.004]). These results persisted after controlling for PMi0, SO2, or
15 O3 in two-pollutant models. An identical study in Kaohsiung analyzed 29,661 CVD admissions for
16 the period 1997-2000 (Yang et al., 2004, 094376). Results showed that a 0.5 ppm increase in 24-h
17 avg CO concentration was associated with an increase in CVD hospital admissions during both the
18 warmer periods (OR: 1.50 [95% CI: 1.38-1.63) and cooler periods (OR: 1.89 [95% CI: 1.69-2.12]).
19 Similarly, two Australian studies also reported associations between ambient CO
20 concentrations and increased hospital admissions among older adults. The first of these studies
21 analyzed data from five of the largest cities in Australia (Brisbane, Canberra, Melbourne, Perth,
22 Sydney) and two New Zealand cities (Auckland, Christchurch) for the period 1998-2001 (Barnett et
23 al., 2006, 089770). The combined estimates showed that an increase of 0.75 ppm in the average 8-h
24 max CO concentration over the current and previous day (lag 0-1) was associated with a 1.8%
25 (95% CI: 0.7-2.8) increase in all CVD admissions among those aged 65+ yr. Among those aged 15-
26 64 yr there was a smaller increase in CVD admissions (1.0% [95% CI: 0.2-1.7]). The second of the
27 Australian studies examined ED visits for CVDs in older adults (65+ yr) in Sydney for the period
28 1997-2001 (Jalaludin et al., 2006, 189416). A 0.75 ppm increase in 8-h max CO concentration for
29 single-day lags 0 and 1 was associated with increases in admissions of 2.5% (95% CI: 1.6-3.5) and
30 1.4% (95% CI: 0.5-2.4) respectively. Based on an average over lags 0 and 1 (e.g., lag 0-1) there was
31 an increase of 2.6% (95% CI: 1.5-3.6). There were positive increases of approximately 3% in CVD
32 ED visits during the cool (May-October) period, but not the warm period (November-April).
33 Very few studies investigating the association between CO and cardiovascular hospital
34 admissions have been conducted in European cities. Ballester et al.(2001, 013257) analyzed
35 emergency hospital admissions in Valencia, Spain for the period 1994 - 1996. The mean daily
36 number of CVD admissions was 7 and when using a time-series approach there was no association
September 2009 5-50 DRAFT - DO NOT CITE OR QUOTE
-------
1 between CO and admissions for all CVDs (RR: 1.009 [95% CI: 0.99-1.016] per 1 ppm increase in
2 1-h max CO concentration), heart diseases (RR: 1.010 [95% CI: 0.993-1.028] per 1 ppm increase),
3 and cerebrovascular diseases (RR: 0.985 [95% CI: 0.959-1.012] per 1 ppm increase). When the
4 analyses were stratified by hot and cold seasons, only CO concentrations during the hot season were
5 associated with an increase in all cardiovascular admissions (RR: 1.033 [95% CI: 1.006-1.064] per
6 1 ppm increase), heart disease admissions (RR: 1.033 [95% CI: 1.000-1.067] per 1 ppm increase),
7 and cerebrovascular admissions (RR: 1.074 [95% CI: 1.007-1.113] per 1 ppm increase).
8 Ballester et al. (2006, 088746) extended this research to include data from 14 Spanish cities
9 for the period of 1995-1999. An average exposure period over lags 0-1 was analyzed and for the
10 combined estimates a 0.75 ppm increase in 8-h max CO concentration was associated with a 1.77%
11 (95% CI: 0.56-2.99) increase in all cardiovascular emergency hospital admissions and a larger
12 increase of 3.57% (95% CI: 1.12-6.08) for heart disease admissions. These results persisted in two-
13 pollutant models that included NO2, O3 and SO2.
14 A study was carried out to evaluate the association between air pollution cardiovascular ED
15 visits in subjects with and without diabetes in Sao Paulo, Brazil (Filho et al., 2008, 190260). From
16 January 2001 to July 2003 45,000 ED visits were registered due to cardiovascular diseases, of which
17 700 were registered due to cardiovascular diseases in diabetic patients. SO2 and NO2 were positively
18 and statistically significantly associated with CVD ED visits among diabetics and non-diabetics,
19 while CO was only positive and statistically significant among non-diabetic patients. PMi0 and O3
20 were not positively associated with ED admissions among either group.
21 Table 5-11 summarizes the non-specific CVD hospital admission studies that examined CO
22 exposures. Figure 5-5 shows the effect estimates associated with daily admissions for non-specific
23 CVD hospital admissions from selected studies.
24 In summary, many of the studies that examined associations between ambient CO
25 concentrations and ED visits and daily hospital admissions for CVD reported small yet precise
26 positive associations at short (0-1 day) lags. Among studies that conducted stratified analyses, there
27 were slightly stronger effects among older adults and possibly during warmer periods.
September 2009 5-51 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Bell etal. (2009,193780)
Tolbertetal. (2007,090316)
Linn etal. (2000, 002839)
Linn etal. (2000, 002839)
Linn etal. (2000, 002839)
Linn etal. (2000, 002839)
Linn etal. (2000, 002839)
Fung etal. (2005, 074322)
Fung etal. (2005, 074322)
Barnett etal. (2006, 089770)
Barnett etal. (2006. 089770)
Jalaludin et al. (2006, 189416)
Ballester et al. (2006, 088746)
Pereira Filho et al. (2008, 190260)
Pereira Filho et al. (2008, 190260)
Chang etal. (2005, 080086)
Chang etal. (2005, 080086)
Yang et al. (2004, 094376)
Yang et al. (2004, 094376)
Location
126 U.S. Counties
Atlanta, GA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Windsor, Canada
Windsor, Canada
Australia, New Zealand
Australia, New Zealand
Sydney, Australia
Multicity, Spain
Sao Paulo, Brazil
Sao Paulo, Brazil
Taipei, Taiwan
Taipei, Taiwan
Kaohsiuno Taiwan
Kaohsiuno Taiwan
Lag
0
0-2
0
0
0
0
0
0-2 —i
0-2
0-1
0-1
0-1
0-1
0
o
0-2
0-2 — •
0-2
0-2
Effect Estimate
1 65+ yr
•
» All year
- Spring
• Summer
i Fall
* Winter
— <65yr
•- 65+ yr
» 15-64yr
» 65+ yr
« 65+ yr
•
-•— Diabetic
» Non-diabetic
•+- >20°CTemp
<20°C Temp
^°0°CTemp •
0.9 1.1 O 1.5
1.7
1.9 2.1
Figure 5-5 Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CVD. Effect estimates have been standardized to a 1 ppm
increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h max CO
concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-52
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-11 Summary of non-specific CVD hospital admission studies.
Study
Bell etal. (2009,193780)
Metzger et al. (2004, 044222)
Peel etal. (2007,090442)
Tolbert etal. (2007,090316)
Linn etal. (2000, 002839)
Slaughter et al. (2005, 073854)
Fung etal. (2005, 074322)
Chang etal. (2005, 080086)
Yang et al. (2004, 094376)
Barnett etal. (2006, 089770)
Jalaludin et al. (2006, 189416)
Ballester etal. (2001,013257)1
Ballester et al. (2006, 088746) 1
Pereira Filho et al. (2008,
190260)
Location
126 urban U.S.
counties
(1999-2005)
Atlanta, GA
(1993-2000)
Atlanta, GA
(1993-2000)
Atlanta, GA
(1993-2004)
Los Angeles, CA
(1992-1995)
Spokane, WA
(1995-2001)
Windsor, Canada
(1995-2000)
Taipei, Taiwan
(1997-2001)
Kaohsiung, Taiwan
(1997-2000)
Australia and New
Zealand
(1998-2001)
Sydney, Australia
(1997-2001)
Valencia, Spain
(1994-1996)
Multicity, Spain
(1995-1999)
Sao Paulo, Brazil
(2001-2003)
CVD Codes
Total CVD
All CVD
All CVD
All CVD
All CVD
All CVD
(ICD9: 390-
459)
All CVD (HF,
IMF, or
Dysrhythmia)
All CVD
(ICD9:410-
429)
All CVD
(ICD9:410-
429)
All CVD
(ICD9: 390-
459)
All CVD
(ICD9: 390-
459)
All CVD
(ICD9: 390-
459)
All CVD
(ICD9: 390-
459)
All CVD
Copollutants
PM25,N02,
EC
PM10,N02,
S02,03
PMio,N02,
S02,03
PM10,N02,
S02,03
PM10,N02,03
PMio,PM25,
CO
PM10,N02,
S02,03
PM10,N02,
S02,03
PM10,N02,
S02,03
PM10,N02,03
PM10,N02,
S02,03
BS, N02,S02,
03
BS,PM10,
TSP, N02,
S02,03
PM10,N02,
S02,03
Lags
Examined
0,1,2
0-2ma
0-2ma
0-2ma
0
1,2,3
0,0-1,0-2
0-2
0-2
0-1
0,1,2,3,0-1
1,2,3,4,5
0-1
0,1,2,0-1,
0-2, 0-3
Upper CO
Concentrations from
AQS* in ppm
98th%: 1.1-19.1
99th%: 1.2-22.1 (1 h)
98th%: 5.0-5.1
99th%: 5.5-5.9(1 h)
98th%: 5.0-5.1
99th%: 5.5-5.9(1 h)
98th%: 1.0-7.8
99th%:1.1-8.3(24h)
98th%: 1.5-4.6
99th%: 1.7-5.0 (24 h)
NA
NA
NA
NA
NA
NA
NA
NA
CO Concentrations
Reported by Study
Authors in ppm
Median: 1.3(1 h)
Median: 0.5 (24 h)
Mean: 1.5(1 h)
Mean 1 .5 (1 h)
Mean 1.6(1 h)
Mean: (24 h)
Winter 1.7; Spring 1.0;
Summer 1.2; Fall 2.1
Mean: range across 5
monitors 0.42-1 .82 (24
h)
Mean : 1.3 (24 h)
Mean : 1.37 (24 h)
Mean: 0.79 (24 h)
Mean: (8h)
0.5-2.1
Mean: 0.82 (8h)
Mean: 0.54 (24 h)
Mean: range across 14
cities
0.1 2-0.24 (8h)
Mean: 2.7 (8 h)
1These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which assumes standard atmosphere and
temperature.
1 Figure 5-6 and Figure 5-7 summarizes the effects of CO concentration on ED visits and
2 hospital admissions for all CVD outcomes other than stroke from studies that presented the results
3 from two-pollutant models. Generally, the CO effect estimates from these studies are robust to the
4 inclusion of copollutants, including PMi0, PM2.5, NO2, SO2, and O3. In all but two instances (Lee
September 2009
5-53
DRAFT - DO NOT CITE OR QUOTE
-------
1 al., 2007, 093271): <25°C adjusted for NO2 and (Yang, 2008, 157160): <20°C adjusted for O3) when
2 the single pollutant effect estimate was positive for CO, it remained positive after the addition of any
3 of the copollutants investigated.
Study
Outcome
Lag
Effect Estimate
PM10
Wellenius et al. (2005, 087483)
Wellenius et al. (2005, 087483)
Chan etal. (2006. 0901 93)
Chan etal. (2006. 0901 93)
Lee et al. (2007, 093271)
Lee etal. (2007. 093271)
D'lppoliti etal. (2003. 074311)
D'lppoliti etal. (2003. 074311)
Chang etal. (2005, 080086)
Chang etal. (2005, 080086)
Chang etal. (2005, 080086)
Chang etal. (2005, 080086)
Lee etal. (2003. 095552)
Lee etal. (2003. 095552)
Ballester et al. (2006, 088746)
Ballester et al. (2006, 088746)
von Klot etal. (2005.088070)
von Klot etal. (2005.088070)
*
Lee et al. (2007, 093271)
Lee etal. (2007. 093271)
Yang (2008, 157160)
Yang (2008, 157160)
Yang (2008, 157160)
Yang (2008, 157160)
CHF
CHF
CD
CD
CHF
CHF
IHD
IHD
CVD
CVD
CVD
CVD
IHD
IHD
CVD
CVD
CVD
CVD
CHF
CHF
CHF
CHF
CHF
CHF
0
0
2
2
0
0-2
0-2
0
0
0
1
0
0
0-1
0-1
0
0
0
0-2
0-2
0-2
0-2
CO Alone ! •
i
CO+PM10 | •*
CO Alone J«-
CO+PM10 X
i
i
i
CO Alone r*-
i
CO + TSP *•
CO Alone | -»->20°C
i
i
CO Alone — <*- <20°C
CO+PM10 — •— i
i
CO Alone r«- 64+ yr
i
CO+PM10 -*-
CO Alone r*~
i
CO+PM10 -i*-
CO Alone [-*-
CO+PM10 -r*-
i
i
0} Alone — • — >20°C
i
CO+PM10 — *
CO Alone — j-» <20°C
i
PM2.5
Bell etal. (2009. 193780)
Bell etal. (2009. 193780)
Tolbert etal. (2007.090316)
Tolbert etal. (2007.090316)
Chan etal. (2006. 0901 93)
Chan etal. (2006. 0901 93)
CVD
CVD
CVD
CVD
CD
CD
0
0
0-2
0-2
2
2
CO Alone •*•
CO + PM2.5 •*•
CO Alone <*
CO+PM25 ,*•
CO Alone K-
CO+PM25 "*•
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Figure 5-6
Effect estimates from studies of ED visits and hospital admissions for CVD
outcomes other than stroke from single pollutant (CO only, closed circles) and
particulate copollutant (CO plus PM, open circles) models. Effect estimates have
been standardized to a 1 ppm increase in ambient CO for 1-h max CO
concentrations, 0.75 ppm for 8-h max CO concentrations, and 0.5 ppm for 24-h
avg CO concentrations
September 2009
5-54
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Outcome
Lag
Effect Estimate
N02
Bell et al. (2009, 193780)
Bell et al. (2009, 1937801
Wellenius et al. (2005, 087483)
Wellenius et al. (2005, 087483)
Tolbert etal. (2007, 090316)
Tolbert etal. (2007, 090316)
Ballesteretal. (2006, 0887461
Ballesteretal. (2006, 0887461
Chang et al. (2005, 0800861
Chang et al. (2005, 0800861
Chang et al. (2005, 0800861
Chang et al. (2005, 0800861
Leeetal. (2007, 093271)
Leeetal. (2007, 093271)
Leeetal. (2007, 093271)
Lee et al. (2007, 0932711
Yang (2008, 157160)
Yang (2008, 1571601
Yang (2008, 157160)
Yang (2008, 157160)
CVD
CVD
CHF
CHF
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
0
0
0
0
0-2
0-2
0-1
0-1
0
0
0
3
0
0
0
0 CO
0-2
0-2
0-2
0-2
CO Alone «.
CO + N02 A
CO Alone *
CO + N02 A-
CO Alone +
CO + N02 4.
CO Alone -«_
C0+N02 i —
CO Alone1 +->2Q°C
CO + N02 [
CO Alone _«- <20°C
C0+N02 1_
rn i |\p.
4. up- _^_i
CQAIone _,-_ >20 °C
CO + N02 — 1
CO Alone -^ <20°C
rn + |\p. [
; so2
Wellenius et al. (2005, 087483) CHF 0
Wellenius et al. (2005, 087483) CHF 0
Ballesteretal. (2006, 0887461 CVD 0-1
Ballesteretal. (2006, 0887461 CVD 0-1
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 2
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Yang (2008, 157160) CHF 0-2
Yang (2008, 157160) CHF 0-2
Yang (2008, 157160) CHF 0-2
Yang (2008, 1571601 CHF 0-2
Wellenius et al. (2005, 0874831 CHF 0
Wellenius et al. (2005, 087483) CHF 0
Chan et al. (2006, 0901931 CD 2
Chan et al. (2006, 0901931 CD 2
Ballesteretal. (2006. 0887461 CVD 0-1
Ballesteretal. (2006, 0887461 CVD 0-1
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 0
Chang et al. (2005, 0800861 CVD 4
von Klot et al. (2005, 0880701 Cardiac 0
von Klot et al. (2005, 088070) Cardiac 0
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Leeetal. (2007, 093271) CHF 0
Yang (2008, 157160) CHF 0-2
Yang (2008, 157160) CHF 0-2
Yang (2008, 157160) CHF 0-2
Yang (2008, 1571601 CHF 0-2
Effect estimates have been standardized to a 1 ppm increase in ambient CO for 1-h max CO
concentrations, 0.75 ppm for 8-h max CO concentrations, and 0.5 ppm for 24-h avg CO concentra
CO Alone, •
CO + S02( A
COAIonS-*-
CO + S02 "-A-
COAIyne-*- >20°C
CQ+ S02 -A-
CO Alone -20°C
qo + so2 — A —
CO Alone -MP <20°C
C0 + S02 — LA
03
CO Alone *
CO + 03 -
CO Alone -«-
CO + 03 —
CO Alone r-»-
COAItme*. >20°C
C50 + 03
CO Alone -*- <20°C
CO + 03 , — A—
CO Alone L-»_
rn Jvjnnr «. ->°5 °r
rr\+ n
, rn Ainpp ( ^05 or
COJMone — « — >20°C
CO Alone -(-• <20°C
i
(CO only=closed circles) and gaseous copollutant (CO plus N02, S02 and 03= triangles) models.
0.8 1.0 12 1.4 1.6 1.8 ZO
Figure 5-7 Effect estimates from studies of ED visits and HAs for CVD outcomes other than
stroke from single pollutant.
September 2009
5-55
DRAFT - DO NOT CITE OR QUOTE
-------
5.2.2.Epidemiologic Studies with Long-Term Exposure
1 Two studies examined CVD outcomes in association with long-term exposure to CO.
2 Rosenlund et al. (2006, 089796) investigated long-term exposure (30 yr) to urban air pollution and
3 the risk of MI in Sweden. The study included 2,246 cases and 3,206 controls aged between 45-70 yr
4 and residing in Stockholm County during 1992-1993. A detailed postal questionnaire was completed
5 by 4067 subjects and all addresses inhabited during more than 2 yr since 1960 were geocoded. The
6 exposures were then derived from dispersion calculations based on emissions data for each decade
7 since 1960. These calculations were estimates of annual mean levels of traffic-generated NOX, NO2,
8 CO, PMio, and PM2 5, with the addition of SO2 from heating sources. The analyses were stratified by
9 all cases, nonfatal cases, fatal cases, in-hospital death, and out-of-hospital death. Based on a 30-yr
10 avg exposure all pollutants were not associated with overall MI incidence. However, increased CO
11 was associated with out-of-hospital death from MI (OR: 1.81 [95% CI: 1.02-3.23] per 0.5 ppm
12 increase in 30-yr avg CO concentration). Similar results were reported for NO2. The correlation
13 between the 30-yr NO2 and CO exposures was reasonably strong (r = 0.74) and multipollutant
14 models with both these pollutants included (NO2, CO) were not examined. No other pollutants were
15 significantly associated with all other MI outcomes. The study period was extended to include
16 43,275 cases of MI during 1985-1996 and 507,000 controls (Rosenlund et al., 2009, 190309). Five-
17 year average exposures to NO2, PMi0 and CO were associated with incidence of MI, especially with
18 fatal disease; when examining only nonfatal disease no association was observed. The effect estimate
19 for CO (OR: 1.03; 95%C CI: 1.02-1.04 per 0.5 ppm increase in 5-yr average) was similar in
20 magnitude to those for NO2 and PMi0. When the analysis was restricted to the group that did not
21 move between population censuses (the least expected misclassification of true individual exposure),
22 the effect estimate for CO increased to 1.17 (95% CI: 1.11-1.24) per 0.5 ppm increase in 5-yr
23 average, and although the effect estimates for NO2 and PMi0 remained similar to the estimate for
24 CO, in this analysis the effect estimate for CO was slightly greater in magnitude than the effect
25 estimate for PM10.
26 A small-area ecologic study analyzed mortality and hospital admissions for stroke across 1,030
27 census districts in Sheffield, U.K. (Maheswaran et al., 2005, 088683). Stroke counts within each
28 census district were linked to modeled air pollution data which was then grouped into quintiles of
29 exposure. For stroke hospital admissions, when the analyses were adjusted for only sex and age
30 demographics there was an exposure-response pattern exhibited across the quintiles of CO exposure
31 with all levels reaching significance (RR: 1.37 [95% CI: 1.24-1.52] for the highest exposure group
September 2009 5-56 DRAFT - DO NOT CITE OR QUOTE
-------
1 compared to the lowest group). However, this result did not persist when also adjusting for a
2 deprivation index and smoking rates across the districts (RR: 1.11 [95% CI: 0.99-1.25]).
5.2.3.Summary of Epidemiologic Studies of Exposure to CO and
Cardiovascular Effects
3 A substantial number of epidemiologic studies have examined the potential association
4 between exposure to CO and various relevant cardiac endpoints or biomarkers. Overall, despite some
5 mixed results reported among panel and retrospective cohort studies, there was evidence that
6 exposure to CO has an effect on HR, various HRV parameters, and blood markers of coagulation and
7 inflammation. Conversely, based on results from panel studies there was little evidence of a link
8 between CO and cardiac arrhythmia, cardiac arrest, the occurrence of myocardial infarction, and
9 increased BP.
10 Studies of ED visits and hospital admissions provide evidence that CO is associated with
11 various forms of CVD with lag periods ranging from 0 to 3 days. There is little evidence that
12 ambient CO is associated with an increase in hospital admissions for ischemic stroke. Studies of
13 hospital admissions and ED visits for IHD and CHF provide the strongest evidence of ambient CO
14 being associated with adverse CVD outcomes. It is difficult to determine from this group of studies
15 the extent to which CO is independently associated with CVD outcomes or if CO is a marker for the
16 effects of another traffic-related pollutant or mix of pollutants. On-road vehicle exhaust emissions
17 are a nearly ubiquitous source of combustion pollutant mixtures that include CO and can be an
18 important contributor to CO in near-road locations. Although this complicates the efforts to
19 disentangle specific CO-related health effects, the evidence indicates that CO associations generally
20 remain robust in copollutant models and supports a direct effect of short-term ambient CO exposure
21 on CVD morbidity.
5.2.4.Controlled Human Exposure Studies
22 Controlled human exposure studies provide valuable information related to the health effects
23 of short-term exposure to air pollutants. Results of controlled human exposure studies can be used to
24 provide coherence with the evidence from epidemiologic studies by expanding the understanding of
25 potential mechanisms for the observed health outcomes. However, they may also provide
26 information that can be used directly in quantitatively characterizing the exposure concentration-
27 health response relationships at ambient or near-ambient concentrations.
28 Several human clinical studies cited in the 2000 CO AQCD observed changes in measures of
29 cardiovascular function among individuals with coronary artery disease (CAD) following short term
30 exposures to CO. Principal among these is a large multilaboratory study of men with stable angina
31 (n = 63) designed to evaluate the effect of CO exposure resulting in COHb concentrations of 2% and
September 2009 5-57 DRAFT - DO NOT CITE OR QUOTE
-------
1 4% on exercise-induced angina and ST-segment changes indicative of myocardial ischemia Allred
2 et al. (1989, 013018: 1989, 012697: 1991, 011871). The majority of subjects were following an
3 antiischemic medication regimen (e.g., beta blockers, nitrates, or calcium channel antagonists) which
4 was maintained throughout the study. On two separate occasions, subjects underwent graded
5 exercise treadmill tests following 50-70 min exposures to average CO concentrations of 117 ppm
6 (range 42-202 ppm) and 253 ppm (range 143-357 ppm). The post-exposure target COHb
7 concentrations were set at values 10% greater than the post-exercise targets (i.e., 2.2% and 4.4%) to
8 compensate for the elimination of CO during exercise testing in clean air following exposure. CO
9 uptake constants were determined for each subject individually during a qualifying visit and were
10 used to compute the inhaled concentration required to attain the target COHb concentrations.
11 Although CO-oximetry was used at each center to rapidly provide approximate concentrations of
12 COHb during the actual exposure, COHb concentrations determined by a gas chromatographic
13 technique were used in the statistical analyses as this method is known to be more accurate than
14 spectrophotometric measurements, particularly for samples containing COHb concentrations < 5%.
15 For the two CO exposures, the average post-exposure COHb concentrations were reported as 2.4%
16 and 4.7% (3.2% and 5.6% using CO-oximetry), and the average post-exercise COHb concentrations
17 were reported as 2.0% and 3.9% (2.7% and 4.7% using CO-oximetry). While the average COHb
18 concentrations during the exercise tests were clearly between the concentrations measured in post-
19 exposure and post-exercise blood samples, the study authors noted that the samples at the end of the
20 exercise test represent the COHb concentrations at the approximate time of onset of myocardial
21 ischemia as indicated by angina and ST segment changes. Relative to clean air exposure (COHb
22 a 0.6-0.7%), exposures to CO resulting in post-exercise COHb concentrations of 2.0% and 3.9%
23 were shown to decrease the time required to induce ST-segment changes by 5.1% (p = 0.01) and
24 12.1% (p <0.001), respectively. These changes were well correlated with the onset of exercise-
25 induced angina. The apparent dose-response relationship observed was further evaluated by
26 regressing the percent change in time to ST-segment change or time to angina on actual post-exercise
27 COHb concentration (0.2% - 5.1%) using the three exposures (air control and two CO exposures) for
28 each subject. This analysis demonstrated significant decreases in time to angina and ST-segment
29 change of approximately 1.9% and 3.9%, respectively, per 1% increase in COHb concentration.
30 In addition to work of Allred et al., a number of other studies involving individuals with stable
31 angina have also demonstrated a CO-induced decrease in time to onset of angina as well as reduction
32 in duration of exercise at COHb concentrations between 3 and 6%, measured using
33 spectrophotometric methods (Adams et al., 1988, 012692: Anderson et al., 1973, 023134: Kleinman
34 et al., 1989, 012696: Kleinman et al., 1998, 047186). However, Sheps et al. (1987, 012212) observed
35 no change in time to onset of angina or maximal exercise time following a 1-h exposure to 100 ppm
36 CO (targeted COHb of 4%) among a group of 30 patients with CAD. In a subsequent study
September 2009 5-58 DRAFT - DO NOT CITE OR QUOTE
-------
1 conducted by the same laboratory, a significant increase in number of ventricular arrhythmias during
2 exercise was observed relative to room air among individuals with CAD following a 1-h exposure to
3 200 ppm CO (targeted COHb of 6%), but not following a 1-h exposure to 100 ppm CO (targeted
4 COHb of 4%) (Sheps et al., 1990, 013286). It should be noted that although the subjects evaluated in
5 the studies described above are not necessarily representative of the most sensitive population, the
6 level of disease in these individuals was relatively severe, with the majority either having a history of
7 MI or having > 70% occlusion of one or more of the coronary arteries.
8 The 2000 CO AQCD presented very little evidence of CO-induced changes in cardiovascular
9 function in healthy adults. Davies and Smith (1980, 011288) exposed healthy young adults
10 continuously for 7 days to CO concentrations of 0, 15, or 50 ppm. In this study, a marked
11 ST-segment depression was demonstrated in only 1 out of 16 subjects following exposure to 15 ppm
12 CO (2.4% COHb) or 50 ppm CO (7.2% COHb).
13 Since the publication of the 2000 CO AQCD, no new human clinical studies have been
14 published involving controlled CO exposures among subjects with CAD. However, a number of new
15 studies have evaluated changes in various measures of cardiovascular and systemic responses
16 following controlled exposures to CO in healthy adults. Adir et al. (1999, 001026) exposed 15 young
17 healthy adult males to room air or CO for approximately 4 min, using a CO exposure concentration
18 which had been shown to produce the targeted COHb level of 4-6%. Following each exposure,
19 subjects performed an exercise treadmill test at their maximal capacity. Exposure to CO was not
20 observed to cause arrhythmias, ST-segment changes, or changes in myocardial perfusion (thallium
21 scintigraphy) during post-exposure exercise. However, CO was demonstrated to decrease the post-
22 exposure duration of exercise by approximately 10% (p = 0.0012). In addition, the authors reported
23 significant CO-induced decreases in metabolic equivalent units (p <0.001), which is a relative
24 measure of O2 consumption. These results support the findings of several studies cited in the 2000
25 CO AQCD which observed decreases in exercise duration and maximal aerobic capacity among
26 healthy adults at COHb levels > 3% (Drinkwater et al., 1974, 041332: Ekblom and Huot, 1972,
27 010886: Horvath et al., 1975, 010887: Raven et al., 1974, 041340). While these decreases in exercise
28 duration were relatively small and only likely to be noticed by competing athletes, the findings are
29 nonetheless important in providing coherence with the observed effects of CO on exercise-induced
30 myocardial ischemia among patients with CAD.
31 Kizakevich et al.(2000, 052691) evaluated the cardiovascular effects of increasing CO
32 concentration in healthy adults engaged in upper and lower body exercise. Subjects were initially
33 exposed for 4-6 min to CO concentrations between 1,000 and 3,000 ppm, followed by continued
34 exposure to 27, 55, 83, and 100 ppm to maintain COHb levels of 5, 10, 15, and 20%, respectively.
35 Relative to room air control, CO exposure was not observed to cause ST-segment changes or affect
36 cardiac rhythm at any concentration during either upper or lower body exercise. Compensation
September 2009 5-59 DRAFT - DO NOT CITE OR QUOTE
-------
1 mechanisms for reduced O2 carrying capacity during CO exposure were demonstrated, with
2 statistically significant increases in heart rate occurring at COHb levels > 5%, and statistically
3 significant increases in cardiac output and cardiac contractility observed at COHb levels > 10%. In a
4 human clinical study designed to evaluate the contribution of CO to cardiovascular morbidity
5 associated with cigarette smoking, Zevin et al. (2001, 021120) exposed 12 healthy male smokers for
6 7 consecutive days to clean air, CO, or cigarette smoke, with each subject serving as his own control.
7 The COHb levels were similar between the exposures to cigarette smoke and CO, with average
8 concentrations of 6% and 5%, respectively. Cigarette smoke, but not CO, was observed to
9 significantly increase plasma levels of CRP and plasma platelet factor 4 relative to the air control
10 arm of the study. Neither cigarette smoke nor CO was shown to affect BP. Hanada et al. (2003,
11 193915) observed an increase in leg muscle sympathetic nerve activity (MSNA) following
12 controlled exposures to CO (COHb ~ 20%) under normoxic or hyperoxic conditions. Although an
13 increase in the magnitude of sympathetic activation is typically associated with regional
14 vasoconstriction, no CO-induced changes in femoral venous blood flow were observed in this study.
15 These findings are in agreement with those of Hausberg and Somers (1997, 083450) who observed
16 no change in forearm blood flow or BP in a study of 10 healthy men and women following a
17 controlled exposure to CO (COHb ~ 8%). Interestingly, one recent study did observe an increase in
18 retinal blood flow, retinal vessel diameter, and choroidal blood flow following controlled exposures
19 to CO at a concentration of 500 ppm (Resch et al., 2005, 193853). This protocol resulted in COHb
20 concentrations of 5.6% and 9.4% following exposures of 30 and 60 min, respectively, with
21 statistically significant increases in retinal and choroidal blood flow observed at both time points
22 relative to synthetic air control. This CO-induced change in ocular hemodynamics may have been
23 due to local tissue hypoxia; however, the clinical significance of this finding is unclear. Exposures to
24 CO have also been shown to affect skeletal muscle function, with one recent human clinical study
25 reporting a decrease in muscle fatigue resistance in healthy adult males using both voluntary and
26 electrically-induced contraction protocols following controlled exposures to CO resulting in an
27 average COHb level of 6% (Morse et al., 2008, 097980).
28 In summary, controlled human exposures to CO among individuals with CAD have been
29 shown to consistently increase markers of myocardial ischemia at COHb levels between 3 and 6%,
30 with one study reporting similar effects following CO exposures resulting in COHb concentrations
31 of 2.0-2.4%. No such effects have been observed in healthy adults following controlled exposures to
32 CO. Although some studies have reported CO-induced hemodynamic changes among healthy adults
33 at COHb concentrations as low as 5%, this effect has not been observed consistently across studies.
September 2009 5-60 DRAFT - DO NOT CITE OR QUOTE
-------
5.2.5.lexicological Studies
1 While there was no toxicological research reported in the 2000 CO AQCD that involved CO
2 exposures at or below the NAAQS levels, adverse cardiovascular effects were reported for higher
3 CO concentrations. The lowest observed effect levels for cardiovascular effects in experimental
4 animals included 50 ppm (6-wk exposure, 2.6% COHb) for cardiac rhythm effects, 100 ppm
5 (46 days, 9.3% COHb) for hematology effects, 150 ppm (30 min, 7.5% COHb) for hemodynamic
6 effects, 200 ppm (30 days, 15.8% COHb) for cardiomegaly and 250 ppm (10 wk, 20% COHb) for
7 atherosclerosis and thrombosis (Table 5-11) (U.S. EPA, 2000, 000907). Conflicting experimental
8 data relating to the role of CO in promoting atherosclerotic vessel disease was discussed. While
9 some animal studies have linked chronic CO exposure with atherosclerosis development resulting
10 from increased fatty streaking and cellular lipid loading (Davies et al., 1976, 010660; Thomsen,
11 1974, 193781; Turner et al., 1979, 012328). other studies have failed to see this association (Penn et
12 al., 1992, 013728; Stupfel and Bouley, 1970, 010557). Vascular insults due to acute exposure to CO
13 concentrations of 50 ppm and higher were also reported (Ischiropoulos et al., 1996, 079491; Thorn,
14 1993, 013895; Thorn et al., 1998, 016750; Thorn et al., 1999, 016757; Thorn et al., 1999, 016753). In
15 addition, chronic CO exposure has been shown to result in ventricular hypertrophy (Penney et al.,
16 1984, 011567; Penney et al., 1988, 012521).
17 The following sections describe recent studies dealing with toxicity of low to moderate
18 concentrations of CO. There has been little new research with the overt purpose of examining
19 environmentally-relevant levels of CO. For the most part, studies were designed to mimic exposures
20 related to cigarette smoke, either side-stream or mainstream, accidental CO poisoning, or for the
21 purposes of therapeutic application. Thus, few studies examined levels of CO within the current 1 h
22 (35 ppm) or 8 h (9 ppm) NAAQS levels, and fewer still examined concentration response curves to
23 delineate no effects levels. However, it is apparent that CO, at low to moderate levels (35-250 ppm),
24 has pathophysiologic effects on the cardiovascular system and on relatively ubiquitous cellular
25 pathways. In evaluating these studies, it should be kept in mind that the traditional concept of CO
26 pathophysiology resulting from reduced O2 delivery is likely to be more relevant for higher
27 concentrations of CO than are currently found in the ambient environment.
28 CO exposure at environmentally-relevant levels is unlikely to cause overt toxicity in a healthy
29 cell; however, susceptibility may be rendered by disease or early development. A common theme
30 appears to be the vulnerability of vascular cells, especially the endothelium, which could be
31 considered the first organ of contact once CO is taken up into the circulation. While relatively little
32 research has been conducted since the 2000 CO AQCD, several key studies conducted at
33 environmentally-relevant CO levels provide important clues to the potential public health
34 implications of ambient CO exposure.
September 2009 5-61 DRAFT - DO NOT CITE OR QUOTE
-------
5.2.5.1. Endothelial Dysfunction
1 While the preferential binding to heme and effective displacement of O2 by CO has been well
2 established for over a century, new information from various fields of study are beginning to
3 elucidate non-hypoxic mechanisms that may lead to cardiovascular abnormalities associated with
4 CO exposure. Research by Thorn, Ischiropoulos, and colleagues (Ischiropoulos et al., 1996, 079491)
5 (Thorn and Ischiropoulos, 1997, 085644; Thorn et al., 1994, 076459; Thorn et al., 1997, 084337)
6 (Thorn et al., 1999, 016753; Thorn et al., 1999, 016757). some of which was reported in the 2000 CO
7 AQCD, has focused on CO-mediated displacement of NO from heme-binding sites. Some of this
8 work demonstrates a specific pathway by which severe CO poisoning can lead to the release of NO
9 from platelets with subsequent neutrophil activation and vascular injury (Ischiropoulos et al., 1996,
10 079491; Thorn et al., 2006, 098418). The steps include (1) peroxynitrite generation from the reaction
11 of NO from platelets with neutrophil-derived superoxide followed by (2) stimulation of intravascular
12 neutrophil degranulation that can result in (3) myeloperoxidase deposition along the vascular lining.
13 Products from myeloperoxidase-mediated reactions can cause endothelial cell activation (Thorn et
14 al., 2006, 098418) and can lead to endothelial dysfunction. The concentrations used in these studies
15 are greatly in excess of the NAAQS levels, but certainly within the range of accidental or
16 occupational exposures. Research by these same investigators at more environmentally-relevant CO
17 levels was partially reviewed in the 2000 CO AQCD. The release of free NO was noted in isolated
18 rat platelets exposed to 10-20 ppm CO (Thorn and Ischiropoulos, 1997, 085644). Increased
19 nitrotyrosine content of the aorta was observed in rats exposed to 50 ppm CO for 1 h (Thorn et al.,
20 1999, 016757; Thorn et al., 1999, 016753). Furthermore in this same study, a 1-h exposure to
21 100 ppm CO led to albumin efflux from skeletal muscle microvasculature at 3 h and leukocyte
22 sequestration in the aorta at 18 h. LDL oxidation was also reported. These effects were dependent on
23 NOS but not on neutrophils or platelets. A second study demonstrated NO-dependent effects of
24 50-100 ppm CO in lungs and is described in Section 5.5.4 (Thorn et al., 1999, 016757). Studies in
25 cultured endothelial cells were also conducted using buffer saturated with 10-100 ppm CO (Thorn et
26 al., 1997, 084337). These experiments were designed to mimic conditions where blood COHb levels
27 were between 3.8 and 28% resulting in exposure of endothelial cells to 11-110 nM CO.
28 CO-stimulated release of NO from endothelial cells along with peroxynitrite formation; delayed cell
29 death was observed at CO concentrations of 22 nM and higher (Thorn et al., 1997, 084337). A more
30 recent study demonstrated adaptive responses in endothelial cells exposed to this same range of CO
31 concentrations (Thorn et al., 2000, 011574). Specifically, 1-h exposure to 11 nM CO resulted in
32 MnSOD and HO-1 induction and resistance to the apoptotic effects of 110 nM CO. These protective
33 effects of CO were mediated by NO, as demonstrated using an inhibitor of NOS and a scavenger of
34 peroxynitrite. Collectively, these experiments demonstrated oxidative and nitrosative stress, the
September 2009 5-62 DRAFT - DO NOT CITE OR QUOTE
-------
1 initiation of inflammation, increased microvascular permeability and altered cell signaling in animals
2 and isolated cells following exposure to 10-100 ppm CO.
3 CO is an endogenous regulator of vasomotor tone through vasodilatory effects mediated by
4 activation of soluble guanylate cyclase and activation of large conductance Ca2+ activated K+
5 channels. However, CO does not cause vasodilation in every vascular bed. For example, 5, 100, 500
6 and 2,500 ppm CO administered by inhalation to near-term fetal lambs did not induce pulmonary
7 vasodilation and the HO inhibitor zinc protoporphyrin IX failed to affect baseline vascular tone
8 (Grover et al., 2000, 097088). In some cases CO promotes vasoconstriction, which is thought to be
9 mediated by inhibition of endothelial NOS (Johnson and Johnson, 2003, 053611; Thorup et al., 1999,
10 193782) or decreased NO bioavailability. An interesting series of studies has also suggested that
11 endogenous CO derived from HO-1 which is induced in a variety of disease models (salt-sensitive
12 forms of hypertension, metabolic syndrome in obese rats) is responsible for skeletal muscle arterial
13 endothelial dysfunction (Johnson and Johnson, 2003, 053611; Johnson et al., 2006, 193874; Teran et
14 al., 2005, 193770). Additional studies will be useful in determining whether environmentally-
15 relevant concentrations of CO have detrimental effect on pre-existing conditions such as
16 hypertension, metabolic syndrome or pregnancy.
17 Several recent animal studies examined the vascular effects of controlled exposures to
18 complex combustion mixtures containing CO. Vascular dilatation was decreased following exposure
19 to diesel (4 h at 4 ppm) (Knuckles et al., 2008, 191987) and gasoline engine emissions (6 h/day x 1,
20 3, and 7 day at 80 ppm) (Lund et al., 2009, 180257). Furthermore, evidence of vascular ROS
21 following gasoline emissions has been shown in certain animal models (6 h/day x 50 day at
22 8-80 ppm) (Lund et al., 2009, 180257). While none of these studies examined the potential
23 independent role of CO, it is clearly a common factor in the various combustion atmospheres and
24 future work will be needed to reveal its importance on vascular health.
5.2.5.2. Cardiac Remodeling Effects
25 Cardiomyopathy, or abnormal growth of the cardiac muscle, can manifest in different ways,
26 depending on the nature of the insult. The adverse effects of cardiac hypertrophy are due to reduction
27 of ventricular chamber volume and a diminishing efficiency of the heart. Such concentric
28 hypertrophy typically occurs in response to chronic increases in load, as occurs with hypertension.
29 Ischemia of the cardiac tissue can also lead to cardiac remodeling and myopathy. During and after an
30 acute infarction or obstruction of major coronary vessels, downstream tissues can suffer severe
31 regional ischemia that leads to significant necrosis. Such regions will lose the ability to contract, and
32 surrounding tissue will show deficits in contractility. Decreased contractility is often a result of
33 structural thinning of the ventricular wall, as well as metabolic impairments. Chronic ischemia, such
34 as may result from CAD, may similarly impair cardiomyocyte function and cause decreased
September 2009 5-63 DRAFT - DO NOT CITE OR QUOTE
-------
1 contractility and remodeling. However, ultimately cardiomyopathies are of a complex origin
2 involving mismanagement of fluid balance, abnormal hormonal influences (epinephrine,
3 angiotensin), and insufficient perfusion/nutrition. Assessing the role of exogenous CO in altering
4 pathways leading to cardiomyopathy is a relatively new endeavor and several new findings are of
5 great interest.
6 The heart is a known target for CO toxicity, potentially due to its high rate of O2 consumption.
7 Effects of CO on the healthy heart have only been observed at relatively high concentrations. For
8 example, a recent study by Sorhaug et al. (2006, 180414) demonstrated cardiac hypertrophy in rats
9 exposed for 72 wk to 200 ppm CO. COHb levels were reported to be 14.7%. Neither structural signs
10 of hypertension in the pulmonary arteries or atherosclerotic lesions in the systemic arteries were
11 observed. A follow-up study by the same investigators (Bye et al., 2008, 193777) found reduced
12 aerobic capacity and contractile function leading to pathologic cardiac hypertrophy in rats exposed
13 for 18 mo to 200 ppm CO. Cardiac hypertrophy was also demonstrated in rats exposed to 100-
14 200 ppm CO for 1-2 wk (Loennechen et al., 1999, 011549). This response was accompanied by an
15 increase in endothelin-1 expression. COHb levels were reported to be 12-23% in this latter study.
16 Effects of CO on the healthy heart have also been demonstrated following short-term
17 exposures. In a study by Favory et al. (2006, 184462) rats were exposed to 90 min of 250 ppm CO,
18 which led to peak COHb values of roughly 11%; recovery of 96 h was needed for COHb levels to
19 return to baseline. The authors noted that within the first 24 h of recovery, while COHb values
20 decreased from 11% to 5%, the coronary vascular perfusion pressure and the left ventricular
21 developed pressure were significantly increased compared to baseline. Concomitantly, the ratio of
22 cGMP to cAMP decreased and the sensitivity of the coronary vascular bed to both acetylcholine and
23 a NO donor were reduced by CO exposure. The authors concluded that the discordant alterations in
24 contractility (increased) and perfusion (decreased) may place the heart at risk of O2 limitations
25 following this exposure to CO.
26 Several studies examined the impact of lower levels (50 ppm) on pre-existing or concurrent
27 cardiac pathologies. In one such study, CO exacerbated the effects of a hypoxia-based model of right
28 ventricular remodeling and failure (Gautier et al., 2007, 096471). In controlled laboratory settings,
29 chronic hypobaric hypoxia (HH) caused right ventricular hypertrophy as a result of pulmonary
30 arterial vasoconstriction and increased pulmonary resistance. Using such a model (Wistar rats
31 exposed for 3 wk to hypoxia), CO (50 ppm during the last week of hypoxia, continuous) only
32 increased COHb from 0.5% to 2.4% in the hypoxia model, yet had significant effects on blocking
33 compensatory functional responses to hypoxia, such as increased fractional shortening and
34 contractility. Also, while right ventricular weight was increased by hypoxia alone, significant
35 pathology related to necrosis was observed in the hypoxia + CO-exposed rats. The reduced coronary
36 perfusion of the right ventricle in hypoxia + CO-exposed rats may help explain the histopathologic
September 2009 5-64 DRAFT - DO NOT CITE OR QUOTE
-------
1 findings. The authors cited previous work demonstrating that exogenous CO can inhibit NOS
2 (Thorup et al., 1999, 193782). which is essential for coronary dilation and angiogenesis. Thus, this
3 study provided evidence that exogenous CO may interrupt or downregulate pathways that
4 endogenous CO may activate.
5 In 2 studies by Melin et al. (2002, 037502; 2005, 193833). Dark Agouti rats were exposed for
6 10 wk to either HH, 50 ppm CO or HH plus 50 ppm CO. CO exposure amplified the right ventricular
7 cardiac hypertrophy and decreased the right ventricular diastolic function which occurred in
8 response to HH. In addition, the combined exposure led to effects on left ventricular morphology and
9 function which were not seen with either exposure alone. Changes in HRV were also reported.
10 Results from both of these studies combined with results of Gautier and colleagues (Gautier et al.,
11 2007, 096471) indicated that CO may interfere with normal homeostatic responses to hypoxia. This
12 could occur by blocking HIF-la-responsive elements (vascular endothelial growth factor,
13 erythropoietin) or other cell signaling pathways.
14 In a similar study, Carraway et al. (2002, 026018) exposed rats to HH (380 torr) with or
15 without co-exposure to CO (50 ppm). These exposures were continuous for up to 21 days and
16 focused on pulmonary vascular remodeling. While the addition of CO to HH did not alter the
17 thickness or diameter of vessels in the lung, there was a significant increase in the number of small
18 (<50 um) diameter vessels compared to control, HH only, and CO-only exposures. Despite the
19 greater number of vessels, the overall pulmonary vascular resistance was increased in the combined
20 CO + hypoxic exposure, which the authors attributed to enhancement of muscular arterioles and p-
21 actin. Results of this study taken together with results from the studies of Gautier et al. (2007,
22 096471) and Melin et al. (2002, 037502: 2005, 193833) suggested that the combined effect of low
23 levels of CO with hypoxia is an enhanced right ventricle workload and an exacerbated
24 cardiomyopathy related to pulmonary hypertension. The population at risk of primary pulmonary
25 hypertension is low, but secondary pulmonary hypertension is a frequent complication of COPD and
26 certain forms of heart failure.
5.2.5.3. Electrocardiograph^ Effects
27 In two related studies, Wellenius et al. (2004, 087874; 2006, 156152) examined the effect of
28 CO on a rat model of arrhythmia that was previously shown to produce significant results with
29 exposures to PM (Wellenius et al., 2002, 025405). ECG changes were observed during exposure to
30 residual oil fly ash (ROFA) particles in a rat model of MI. Thus, using an anesthetized model of post-
31 infarction myocardial sensitivity, Wellenius and colleagues tested the effects of 35 ppm CO (1-h
32 exposure) on the induction of spontaneous arrhythmias in Sprague Dawley rats (Wellenius et al.,
33 2004, 087874). CO exposure caused a statistically significant decrease (60.4%) in ventricular
34 premature beat (VPB) frequency during the exposure period in rats with a high number of pre-
September2009 5-65 DRAFT - DO NOT CITE OR QUOTE
-------
1 exposure VPB. No interaction was observed with co-exposure to carbon concentrated particles,
2 which independently reduced VPB frequency during the post-exposure period when administered
3 alone. In a follow-up publication, results from the analysis of supraventricular ectopic beats (SVEB)
4 were provided (Wellenius et al., 2006, 156152). A decrease in the number of SVEB was observed
5 with CO (average concentration 37.9 ppm) compared to filtered air. While the authors concluded that
6 CO exposure did not increase risk of SVEB in this particular rodent model of coronary occlusion, the
7 fact that cardiac electrophysiological dynamics are significantly altered by short-term exposure to
8 low level CO may be of concern for other models of susceptibility.
5.2.5.4. Summary of Cardiovascular Toxicology
9 Experimental studies demonstrated that short-term exposure to 50-100 ppm CO resulted in
10 aortic injury as measured by increased nitrotyrosine and the sequestration of activated leukocytes in
11 healthy rats. In addition, skeletal muscle microvascular permeability was increased. Short term-
12 exposure to 35 ppm CO altered cardiac electrophysiology in a rat model of arrhythmia. Furthermore,
13 short-term exposure to 50 ppm CO exacerbated cardiac pathology and impaired function in an
14 animal model of hypertrophic cardiomyopathy and enhanced vascular remodeling and increased
15 pulmonary vascular resistance in an animal model of pulmonary hypertension. Ventricular
16 hypertrophy was observed in healthy rats in response to chronic exposures of 100-200 ppm CO.
17 These studies provide some support for the development of adverse health effects resulting from
18 exposures to CO at environmentally-relevant concentrations.
5.2.6.Summary of Cardiovascular Effects
19 The most compelling evidence of a CO-induced effect on the cardiovascular system at COHb
20 levels relevant to the current NAAQS comes from a series of controlled human exposure studies
21 among individuals with CAD. These studies, described in the 1991 and 2000 CO AQCDs,
22 demonstrate consistent decreases in the time to onset of exercise-induced angina and ST-segment
23 changes following CO exposures resulting in COHb levels of 3-6%, with one multicenter study
24 reporting similar effects at COHb levels as low as 2.0-2.4% (see Section 5.2.4). No human clinical
25 studies have evaluated the effect of controlled exposures to CO resulting in COHb levels lower than
26 2%. Human clinical studies published since the 2000 CO AQCD have reported no association
27 between CO and ST-segment changes or arrhythmia; however, none of these studies included
28 individuals with diagnosed heart disease.
29 While the exact physiological significance of the observed ST-segment changes among
30 individuals with CAD is unclear, ST-segment depression is a known indicator of myocardial
31 ischemia. It is also important to note that the individuals with CAD who participated in these
32 controlled exposure studies may not be representative of the most sensitive individuals in the
September 2009 5-66 DRAFT - DO NOT CITE OR QUOTE
-------
1 population. It is conceivable that the most sensitive individuals respond to levels of COHb lower
2 than 2%. Variability in activity patterns and severity of disease among individuals with CAD is
3 likely to influence the critical level of COHb which leads to adverse cardiovascular effects.
4 The degree of ambient CO exposure which leads to attainment of critical levels of COHb will
5 also vary between individuals. Although endogenous COHb is generally less than 1% in healthy
6 individuals, higher endogenous COHb levels are observed in individuals with certain medical
7 conditions. Nonambient exposures to CO, such as exposure to ETS, may increase COHb above
8 endogenous levels, depending on the gradient of pCO. Ambient exposures may cause a further
9 increase in COHb. Modeling results described in Chapter 4 indicate that increases of-1% COHb are
10 possible with exposures of several ppm CO depending on exposure duration and exercise level.
11 Findings of epidemiologic studies conducted since the 2000 CO AQCD (U.S. EPA, 2000,
12 000907) are coherent with results of the controlled human exposure studies. These recent studies
13 observed associations between ambient CO concentration and ED visits and hospital admissions for
14 IHD, CHF and cardiovascular disease as a whole and were conducted in locations where the mean
15 24-h avg CO concentrations ranged from 0.5 ppm to 9.4 ppm (Table 5-7). All but one of these
16 studies that evaluated CAD outcomes (IHD, MI, angina) reported positive associations (Figure 5-2).
17 Although CO is often considered a marker for the effects of another traffic-related pollutant or mix
18 of pollutants, evidence indicates that CO associations generally remain robust in copollutant models
19 and supports a direct effect of short-term ambient CO exposure on CVD morbidity. These studies
20 add to findings reported in the 2000 CO AQCD (U.S. EPA, 2000, 000907) that demonstrated
21 associations between short-term variations in ambient CO concentrations and exacerbation of heart
22 disease.
23 The known role of CO in limiting O2 availability lends biological plausibility to ischemia-
24 related health outcomes following CO exposure. However it is not clear whether the small changes
25 in COHb associated with ambient CO exposures results in substantially reduced O2 delivery to
26 tissues. Recent toxicological studies suggest that CO may also act through other mechanisms by
27 initiating or disrupting cellular signaling. Studies in healthy animals demonstrated oxidative injury
28 and inflammation in response to 50-100 ppm CO while studies in animal models of disease
29 demonstrated exacerbation of cardiomyopathy and increased vascular remodeling in response to
30 50 ppm CO. Further investigations will be useful in determining whether altered cell signaling
31 contributes to adverse health effects following ambient CO exposure.
32 Given the consistent and coherent evidence from epidemiologic and human clinical studies,
33 along with biological plausibility provided by CO's role in limiting O2 availability, it is concluded
34 that a causal relationship is likely to exist between relevant short-term CO exposures and
35 cardiovascular morbidity.
September 2009 5-67 DRAFT - DO NOT CITE OR QUOTE
-------
5.3. Central Nervous System Effects
5.3.1.Controlled Human Exposure Studies
1 The behavioral effects of controlled human exposures to CO have been examined by several
2 laboratories, and these studies were summarized in the 2000 CO AQCD. Briefly, decreases in visual
3 tracking as well as visual and auditory vigilance were observed following exposures to CO resulting
4 in COHb levels between 5% and 20% (Benignus et al., 1987, 012250: Fodor and Winneke, 1972,
5 011041: Horvath et al., 1971, 011075: Putz et al., 1979, 023137). One study reported similar
6 behavioral effects (time discrimination) among a group of healthy volunteers with COHb levels <3%
7 (Beard and Wertheim, 1967, 011015). though subsequent studies were unable to replicate these
8 findings at such low exposure concentrations (Otto et al., 1979, 010863: Stewart et al., 1973,
9 093412). These outcomes represent a potentially important adverse effect of CO exposure resulting
10 in COHb levels > 5%, although it is important to note that these findings have not been consistent
11 across studies. Similarly, some studies demonstrated decreases in reaction time as well as decrements
12 in cognitive function and fine motor skills following controlled exposures to CO; however, these
13 studies were not typically conducted using double-blind procedures, which may significantly affect
14 the outcome of behavioral studies (Benignus, 1993, 013645). It should be noted that all behavioral
15 studies of controlled CO exposure were conducted in normal, healthy adults. No new human clinical
16 studies have evaluated CNS or behavioral effects of exposure to CO.
5.3.2.lexicological Studies
17 The evidence for toxicological effects of CO exposure in laboratory animal models comes
18 from in utero or perinatal exposure involving relatively low to relatively high concentrations of CO
19 (25-750 ppm). Affected endpoints from this early, developmental CO exposure include behavior,
20 memory, learning, locomotor ability, peripheral nervous system myelination, auditory decrements,
21 and neurotransmitter changes. These data are addressed in detail in the Birth Outcomes and
22 Developmental Effects section of the ISA (Section 5.4.2). Further, a group of studies have found that
23 high dose CO (500-1,200 ppm) can result in CO-dependent ototoxicity, specifically loss of threshold
24 of cochlear compound action potentials (CAP) and potentiation of noise-induced hearing loss
25 (NIHL) (Chen et al., 2001, 193985: Fechter et al., 1997, 081322: Fechter et al., 2002, 193926: Liu
26 and Fechter, 1995, 076524). Proposed mechanisms for these effects include ROS generation and
27 glutamate release.
September 2009 5-68 DRAFT - DO NOT CITE OR QUOTE
-------
5.3.3.Summary of Central Nervous System Effects
1 Exposure to high levels of CO has long been known to adversely affect CNS function, with
2 symptoms following acute CO poisoning including headache, dizziness, cognitive difficulties,
3 disorientation, and coma. However, the relationship between ambient levels of CO and neurological
4 function is less clear and has not been evaluated in epidemiologic studies. Studies of controlled
5 human exposures to CO discussed in the 2000 CO AQCD reported inconsistent neural and
6 behavioral effects following exposures resulting in COHb levels of 5-20%. No new human clinical
7 studies have evaluated central nervous system or behavioral effects of exposure to CO. At ambient-
8 level exposures, healthy adults may be protected against CO-induced neurological impairment owing
9 to compensatory responses including increased cardiac output and cerebral blood flow. However,
10 these compensatory mechanisms are likely impaired among certain potentially susceptible groups,
11 including individuals with reduced cardiovascular function.
12 Toxicological studies that were not discussed in the 2000 CO AQCD employed rodent models
13 to show that low to moderate CO exposure during the in utero or perinatal period can adversely
14 affect adult outcomes including behavior, neuronal myelination, neurotransmitter levels or function,
15 and the auditory system (discussed in Section 5.4). In utero CO exposure, including both intermittent
16 and continuous exposure, has been shown to impair multiple behavioral outcomes in offspring
17 including active avoidance behavior (150 ppm CO), non-spatial memory (75 and 150 ppm CO),
18 spatial learning (endogenous CO inhibition), homing behavior (150 ppm CO), locomotor movement
19 (150 ppm CO), and negative geotaxis (125 and 150 ppm). In two separate studies, in utero CO
20 exposure (75 and 150 ppm) was associated with significant myelination decrements without
21 associated changes in motor activity in adult animals. Multiple studies demonstrated that in utero CO
22 exposure affected glutamatergic, cholinergic, catecholaminergic, and dopaminergic neurotransmitter
23 levels or transmission. Possible or demonstrated adverse outcomes from the CO-mediated aberrant
24 neurotransmitter levels or transmission include respiratory dysfunction (200 ppm CO), impaired
25 sexual behavior (150 ppm CO), and an adverse response to hyperthermic insults resulting in
26 neuronal damage (200 ppm). Finally, perinatal CO exposure has been shown to affect the developing
27 auditory system of rodents, inducing permanent changes into adulthood. This is manifested by
28 atrophy of cochlear cells innervating the inner hair cells (25 ppm CO), decreased immunostaining
29 associated with impaired neuronal activation (12.5 ppm CO), impaired myelination of auditory
30 associated nerves (25 ppm CO), decreased energy production in the sensory cell organ of the inner
31 ear or the organ of corti (25 ppm CO). Some of these changes have been proposed to be mediated by
32 ROS. Functional tests of the auditory system of rodents exposed neonatally to CO using OAE testing
33 (50 ppm) and amplitude measurements of the 8th cranial nerve action potential (12, 25, 50,
34 100 ppm), revealed decrements in auditory function at PND22 and permanent changes into
September 2009 5-69 DRAFT - DO NOT CITE OR QUOTE
-------
1 adulthood using action potential (AP) testing (50 ppm). Additionally, high dose CO has been shown
2 to result in CO-dependent ototoxicity in adult animals, possibly through glutamate and ROS
3 dependent mechanisms. Together, these animal studies demonstrated that in utero or perinatal
4 exposure to CO can adversely affect adult behavior, neuronal myelination, neurotransmission, and
5 the auditory system in adult rodents. Considering the combined evidence from controlled human
6 exposure and toxicoiogicai studies, the evidence is suggestive of a causal relationship
7 between relevant short- and long-term CO exposures and central nervous system
8 effects.
5.4. Birth Outcomes and Developmental Effects
5.4.1. Epidemiologic Studies
9 Although the body of literature is growing, the research focusing on adverse birth outcomes is
10 limited when compared to the numerous studies that have examined the more well-established health
11 effects of air pollution. Among this small number of studies, various dichotomized measures of birth
12 weight, such as LEW, SGA, and IUGR, have received more attention in air pollution research while
13 preterm birth (<37 wk gestation; [PTB]), congenital malformations, and infant mortality are less
14 studied.
15 In the 2000 CO AQCD only two studies were cited that examined the effect of ambient air
16 pollution on adverse birth outcomes and both of these studies investigated LEW as an endpoint
17 (Alderman et al., 1987, 012243): (Ritz and Yu, 1999, 086976). At that time this area of research was
18 in its infancy and since then there has been increasing interest.
5.4.1.1. Preterm Birth
19 A small number of air pollution-birth outcome studies have investigated the possible
20 association between PTB and maternal exposure to CO with the majority of U.S. studies conducted
21 in southern California. The earliest of these studies examined exposures to ambient CO during the
22 first month of pregnancy and the last 6 wk prior to birth among a cohort of 97,158 births in southern
23 California between 1989 and 1993 (Ritz et al., 2000, 012068). The exposure assessment within this
24 study was based on data from fixed site monitors that fell within a 2-mi radius of the mother's ZIP
25 code area. The crude relative risks for PTB associated with a 1 ppm increase in 3-h avg CO
26 concentration (6:00 to 9:00 a.m.) during the last 6 wk prior to birth and the first month of pregnancy
27 were 1.04 (95% CI: 1.03-1.5) and 1.01 (95% CI: 1.00-1.03) respectively. However, when the authors
28 controlled for other risk factors, only the effect associated with CO during the last 6 wk prior to birth
September 2009 5-70 DRAFT - DO NOT CITE OR QUOTE
-------
1 persisted (RR: 1.02 [95% CI: 1.01-1.03]). Furthermore, when the analyses included variables for
2 either season or other pollutants the CO effect estimates generally were reduced such that they
3 remained positive, but were no longer statistically significant.
4 Expanding on this research, Wilhelm and Ritz (2005, 088668) examined PTB among a cohort
5 of 106,483 births in Los Angeles County, CA between 1994 and 2000. Based on data recorded at
6 monitoring stations of varying proximities to the mother's residence, the main exposure windows
7 examined were the first trimester and the last 6 wk prior to birth. Among women living within a 1-mi
8 radius of a CO monitoring station, a 0.5 ppm increase in 24-h avg CO concentration during the first
9 trimester was associated with a 3% (RR: 1.03 [95% CI: 1.00-1.06]) increased risk of PTB. This
10 result persisted after simultaneously adjusting for NO2 and O3 (RR: 1.05 [95% CI: 1.00-1.10]), but
11 not with the inclusion of PMi0 into the regression model (RR: 0.99 [95% CI: 0.91-1.09]). The result
12 from the single pollutant model for CO exposures averaged over the 6 wk prior to birth was similar
13 in magnitude but failed to reach statistical significance (RR: 1.02 [95% CI: 0.99-1.04]).
14 A limitation of many air pollution-birth outcome studies is the limited availability of detailed
15 information on maternal lifestyle factors and time-activity patterns during pregnancy. To assess
16 possible residual confounding due to these factors, Ritz and colleagues (2007, 096146) were able to
17 analyze detailed maternal information from a survey of 2,543 from a cohort of 58,316 eligible births
18 in 2003 in Los Angeles County. Based on data from the closest monitor to the mother's ZIP code
19 area, exposures to CO, NO2, O3, and PM2.s during the first trimester and last 6 wk prior to delivery
20 were examined and results from the overall cohort (n = 58,316) with limited maternal information
21 were compared to the more detailed nested case-control cohort (n = 2,543). Within the overall
22 cohort, 24-h avg CO during the first trimester was associated with an increased risk of 25% (OR:
23 1.25 [95% CI: 1.12-1.38]; highest exposure group >1.25 ppm versus lowest < 0.58 ppm). This result
24 persisted within the nested case-control cohort (OR: 1.21 [95% CI: 0.88-1.65]) where factors such as
25 passive smoking and alcohol use during pregnancy were included in the model; however, the
26 confidence intervals were wider due to the smaller sample. Any possible association between CO
27 and PTB was less evident during the last 6 wk prior to birth. A strength of this study was that it also
28 highlighted how there was little change in the air pollution effect estimates when controlling for
29 more detailed maternal information (e.g., smoking, alcohol use), as opposed to only controlling for
30 more limited maternal information that is routinely collected on birth registry forms.
31 In contrast to the Los Angeles studies, a case-control study of PTB across California for the
32 period 1999 through 2000 found a positive association with 24-h CO concentration during the entire
33 pregnancy (OR: 1.03 [95% CI: 0.98-1.09] per 0.5 ppm increase), the first month of gestation (OR:
34 1.05 [95% CI: 0.99-1.10] per 0.5 ppm increase), and the last 2 wk of gestation (OR: 1.00
35 [95% CI: 0.96-1.04] per 0.5 ppm increase) (Huynh et al., 2006, 091240). Although there was an
36 indication of an effect during early pregnancy, the small sample size (when compared to other
September 2009 5-71 DRAFT - DO NOT CITE OR QUOTE
-------
1 studies) may not have provided sufficient power to detect statistical significance. Furthermore,
2 exposures within this study were assigned based on a county-level average which may explain the
3 lack of effect, given the poor level of exposure assessment.
4 Studies outside of the U.S. have been conducted in Canada, Australia, and Korea with mixed
5 results reported. In Vancouver, Canada, based on a city-wide average across available monitoring
6 sites, 24-h avg CO concentration during the last month of pregnancy was associated with a 4% (OR:
7 1.04 [95% CI: 1.00-1.07]) increased risk of PTB per 0.5 ppm increase while there was no association
8 found during the first month of pregnancy (OR: 0.98 [95% CI: 0.94-1.00]) (Liu et al., 2003, 089548).
9 This study investigated maternal exposures to ambient gaseous pollutants (CO, NO2, SO2, O3)
10 averaged over the first and last month of pregnancy among a cohort of 229,085 births between 1985
11 and 1998.
12 In a cohort of 52,113 births in Incheon, Korea between 2001-2002, a kriging technique was
13 used to assign the maternal exposures to CO, which is a statistical mapping technique that allows the
14 prediction of an average concentration over a spatial region from data collected at specific points.
15 The spatial average CO concentrations were then linked to each study subject's residential address.
16 CO concentrations during the first trimester were associated with a 26% (RR: 1.26
17 [95% CI: 1.11-1.44]) increased risk of PTB for the highest quartile of exposure when compared to
18 the lowest quartile (Leem et al., 2006, 089828). There was also a strong significant trend exhibited
19 across the quartiles. A similar result was found for 24-h avg CO concentration during the last
20 trimester although the effect was less pronounced (RR: 1.16 [95% CI: 1.01-1.24]).
21 Conversely, a study in Sydney, Australia, examined maternal exposure to ambient air pollution
22 during the first and last month, and the first and last trimester of pregnancy among a cohort of
23 123,840 births between 1998-2000 and found no association between PTB and CO (Jalaludin et al.,
24 2007, 156601). Maternal exposure estimates in this study were based on a city-wide average of
25 available monitoring sites and also based on data from fixed sites within 5 km of the mother's
26 postcode area. The odds ratios for PTB associated with 8-h avg CO concentrations during the first
27 trimester and last three months of gestation were 1.18 (95% CI: 0.85-1.63) and 1.08
28 (95% CI: 0.95-1.22), respectively, when including births within 5 km of a monitor. Interestingly,
29 when all births were included in the analyses and the exposure was based on a city-wide average,
30 these effects had become protective for the first trimester (OR: 0.82 [95% CI: 0.77-0.87]) and null
31 for the last 3 mo of gestation (OR: 0.99 [95% CI: 0.92-1.07]). This suggests that exposures based on
32 data from fixed sites closer to the mother's address are more likely to detect an effect than a city-
33 wide average.
34 Figure 5-8 shows the risk ratios for the risk of delivering a preterm infant from the reviewed
35 studies. Table 5-12 provides a brief overview of the PTB studies. In summary there are mixed results
36 across the studies. Although these studies are difficult to compare directly due to the different
September 2009 5-72 DRAFT - DO NOT CITE OR QUOTE
-------
1 exposure assessment methods employed, there is some evidence that CO during early pregnancy
2 (e.g., first month and trimester) is associated with an increased risk of PTB. The most consistency is
3 exhibited within the studies conducted around Los Angeles, CA and surrounding areas whereby all
4 studies reported a significant association with CO exposure during early pregnancy, and exposures
5 were assigned from monitors within close proximity of the mother's residential address (Ritz et al.,
6 2000, 012068): (Ritz et al., 2007, 096146): (Wilhelm and Ritz, 2005, 088668). It should also be
7 noted that the mixed results when analyzing different cohorts that resided within varying proximities
8 to a monitor may be attributable to analyzing different populations.
9
Study
Ritz et si (2000, 012068)
Ritz et si (2000, 012068)
Wilhelm & Ritz (2005, 088668)
Wilhelm & Ritz (2005, 088668)
Huynh et al (2006, 091240)
Huynh et a\ (2006, 091240)
Huynh et a\ (2006, 091240)
Liu et a\ (2003, 089548)
Liu et a\ (2003, 089548)
Jslsludin et a\ (2007, 156601)
Jslsludin et a\ (2007, 156601)
Jslsludin et a\ (2007, 156601)
Location
California, USA
California, USA
Los Angeles, CAUSA
Los Angeles, CAUSA
California, USA
California, USA
California, USA
Vancouver, Canada
Vancouver, Canada
Sydney, Australia
Sydney, Australia
Sydney, Australia
Period Effect Estimate
Mo1
Last 6 wk
First trimester
Last 6 wk
First mo
Last 2 wk __
Entire pregnancy
First month ,
Last mo
First mo , Cit
First trimester , Citv wide
Last mo §
•*- <2 mi of monitor
*• <2 mi of monitor
-•- ZIP code level
•*- ZIP code level
— •— — . County level
— , County level
t rnnntw lowol
, City wide
, City wide
/wide
City wide
a?S 1.00 1.2H
Figure 5-8 Summary of effect estimates (95% confidence intervals) for PTB associated with
maternal exposure to ambient CO. Effect estimates have been standardized to a
1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-73
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-12 Brief summary of PTB studies.
Study
Ritzetal. (2000,012068)
Wilhelm and Ritz (2005, 088668)
Ritzetal.(2007, 096146)
Huynh et al. (2006, 091240)
Liu etal. (2003,089548)
Leem et al. (2006, 089828)
Jalaludin et al. (2007, 156601)
Location (Sample
Size)
California
(n = 97,158)
Los Angeles, CA
(n = 106,483)
Los Angeles, CA
(n = 58,31 6)
California
(n = 42,692)
Vancouver, Canada
(n = 229,085)
Incheon, Korea
(n = 52,11 3)
Sydney, Australia
(n = 123,840)
Mean CO
(ppm)
2.7 (6-9 a.m.)
1 .4 (24 h)
0.87 (24 h)
0.8 (24 h)
1 .0 (24 h)
0.9 (24 h)
0.9 (8 h)
Exposure Assessment
<2 mi of monitor
Varying distances to monitor
Nearest monitor to ZIP code
County level
City wide avg
Residential address within Dong-based on
Kriging
City wide avg and
<5 km from monitor
Exposure Window
Mo1
Last 6 wk
Last 6 wk
Entire pregnancy
Trimester 1
Last 6 wk
Entire pregnancy
Mo1
Last 2 wk
Mo1
Last mo
First trimester
Last trimester
First mo
First trimester
Last trimester
Last mo
5.4.1.2. Birth Weight, Low Birth Weight, and Intrauterine Growth Restriction/Small for
Gestational Age
1 With birth weight routinely collected in vital statistics and being a powerful predictor of infant
2 mortality, it is the most studied outcome within air pollution-birth outcome research. Air pollution
3 researchers have analyzed birth weight as a continuous variable, and/or as a dichotomized variable in
4 the forms of low birth weight (LEW) (<2,500g [5 Ibs, 8 oz]) and small for gestational age (SGA).
5 It should be noted that the terms SGA, which is defined as a birth weight <10th percentile for
6 gestational age (and often sex), and intrauterine growth restriction (IUGR) are used interchangeably.
7 However, this definition of SGA does have limitations. For example, using this definition of IUGR
8 may overestimate the percentage of 'growth-restricted' neonates as it is unlikely that 10% of
9 neonates have growth restriction (Wollmann, 1998, 193812). On the other hand, when the 10th
10 percentile is based on the distribution of live births at a population level the percentage of SGA
11 among preterm births is most likely underestimated (Hutcheon and Platt, 2008, 193795).
12 Nevertheless, the terms SGA and IUGR are often used interchangeably and it therefore should be
13 noted that SGA represents a statistical description of a small neonate, whereas the term IUGR is
14 reserved for those with clinical evidence of abnormal growth. Thus, all IUGR neonates will be SGA,
September 2009
5-74
DRAFT - DO NOT CITE OR QUOTE
-------
1 but not all SGA neonates will be IUGR (Wollmann, 1998, 193812). In the following sections the
2 terms SGA and IUGR are referred to as each cited study used the terms.
3 Over the past decade a number of studies examined various metrics of birth weight in relation
4 to maternal exposure to CO with the majority conducted in the U.S. Given that most studies
5 examined multiple birth weight metrics, in order to avoid overlap of the studies the following section
6 focuses on each study only once and presents results for each metric within that study.
7 Most of the U.S. studies have been conducted in southern California with inconsistent results
8 reported with regard to gestational timing of the CO effects. The first of these studies was reviewed
9 in the 2000 CO AQCD and is briefly summarized here. Ritz and Yu (1999, 086976) examined the
10 effect of ambient CO during the last trimester on LEW among 125,573 births in Los Angeles
11 between 1989 and 1993. When compared to neonates born to women in the lowest CO exposure
12 group (<2.2 ppm), neonates born to women in the highest exposure group (5.5 ppm-95th percentile)
13 had a 22% (OR: 1.22 [95% CI: 1.03-1.44]) increased risk of being born as LEW.
14 Building upon this research, Wilhelm and Ritz (2005, 088668) reported similar results when
15 extending this study to include 136,134 births for the period of 1994-2000. Exposure to ambient CO
16 during each trimester was based on data recorded at monitoring stations of varying proximities to the
17 mother's residence. For women residing within 1 mi of a station, there was 36% (OR: 1.36
18 [95% CI: 1.04-1.76]) increased risk of having a term LEW baby for women with third-trimester
19 exposure above the 75th percentile when compared to women below the 75th percentile. There was
20 also an increased risk of term LEW (OR: 1.28 [95% CI: 1.12-1.47]) among women in the highest
21 exposure group when the analyses included women within a 5-mi radius of a station. However, when
22 the analyses included women within a 1-2 mi or 2-4 mi radius of a station, the CO effects failed to
23 reach statistical significance and there was no evidence of an exposure-response pattern exhibited
24 across the varying distances to a station. Furthermore, none of the significant CO results persisted
25 after controlling for other pollutants. Although standard errors were certainly increased after
26 controlling for the other pollutants leading to non-significant results, some of the effect sizes were
27 similar, providing some consistency. It is interesting to note, however, that maternal exposure to CO
28 during trimesters one and two was not associated with LEW (quantitative results not reported by
29 authors).
30 Further validation in association with exposure times was observed in an analysis using a
31 susbset of participants in the Children's Health Study. Salam and colleagues (2005, 087885) found
32 that CO only during the first trimester was associated with reduced fetal growth. Their research
33 examined birth weight, LEW, and IUGR among a subset of participants in the Children's Health
34 Study (Peters et al, 1999, 087243) who were born in California between 1975-1987 (n = 3,901). The
35 study examined term births with a gestational age between 37-44 wk. Exposures in this study were
36 based on CO data from up to the three nearest monitoring sites within 50 km of the centroid of the
September 2009 5-75 DRAFT - DO NOT CITE OR QUOTE
-------
1 mother's ZIP code. Exposures for the entire pregnancy and each trimester were analyzed and a
2 0.5 ppm increase in 24-h CO concentration during the first trimester was associated with a 7.8 g
3 (95% CI: 15.1-0.4) decrease in birth weight, which also translated to a 6.7% (OR: 1.07 [95% CI:
4 1.00-1.13]) increased risk of IUGR; however, there was no association with LEW (OR: 1.00
5 [95% CI: 0.88-1.16]).
6 In contrast to the previous studies, another California study of 18,247 singleton births born at
7 40-wk gestation during 2000 found no association between ambient 24-h CO concentration and
8 reduced birth weight or SGA where the highest quartile of exposure was 0.98 ppm. Based on data
9 from fixed sites within 5 mi of the mother's residence, exposures to CO and PM2.5 during the entire
10 pregnancy and each trimester were analyzed. Although CO during the entire pregnancy was
11 associated with a 20 g (95% CI: 40.1-0.8) reduction in birth weight, this did not persist after
12 controlling for PM2.5. PM2.5 was found to have a strong effect on birth weight within each trimester
13 (Parker et al, 2005, 087462).
14 Two similar studies were conducted in the northeastern U.S. with inconsistent results. A study
15 of 89,557 singleton term births in Boston, MA, Hartford, CT, Philadelphia, PA, Pittsburgh, PA, and
16 Washington, DC between 1994-1996 found that exposure to ambient 24-h avg CO during the third
17 trimester was associated with an increased risk of LEW (OR: 1.14 [95% CI: 1.03-1.27] per 0.5 ppm
18 increase) (Maisonet et al., 2001, 016624). When stratified by race this effect was only significant
19 among African Americans for the first and third trimesters (first OR: 1.32 [95% CI: 1.22-1.43]; third
20 OR: 1.20 [95% CI: 1.09-1.32]). Exposures to PMi0 and SO2 were examined and there was no strong
21 evidence that these pollutants were associated with LEW Exposures for this study were based on a
22 city-wide average of monitors within the mother's city of residence. The second study examined
23 358,504 births at 32-44-wk gestation between 1999-2002 in Connecticut and Massachusetts (Bell et
24 al., 2007, 091059). 24-h CO exposures were estimated from fixed sites within each mother's county
25 of residence (e.g., county level). CO averaged over the entire pregnancy was associated with a
26 reduction in birth weight of 27.0 g (95% CI: 21.0-32.8). This result persisted after controlling for
27 each additional pollutant (PMi0, PM2.5, NO2, and SO2) in two-pollutant models. However, this
28 reduction in birth weight did not translate to an increased risk of LEW (OR: 1.05
29 [95% CI: 0.97-1.12] per 0.5 ppm increase in CO). When controlling for exposure during each
30 trimester, the reduction in birth weight associated with a 0.5 ppm increase in 24-h CO concentration
31 during the first trimester ranged from 18.8 to 16.5 g while the reductions associated with third
32 trimester exposure ranged between 27.2 and 23.3 g. It is interesting to note that, although the
33 exposures were based on data averaged at the county level, CO was associated with a reduction in
34 birth weight. Whereas, in a previously cited California study by Huynh and colleagues (2006,
35 091240) exposures were also at the county level yet there was no association with PTB. This
September 2009 5-76 DRAFT - DO NOT CITE OR QUOTE
-------
1 difference may be due to the counties being smaller in New England than in California, resulting in
2 more precise exposure estimates.
3 Two studies in Canada investigated the effects of ambient air pollution on fetal growth with
4 exposures derived from a city-wide average across the available monitoring sites. The first of these
5 studies was among a cohort of 229,085 singleton term births (37-42-wk gestation) in Vancouver, BC
6 with monthly and trimester exposures to CO investigated in relation to LEW and IUGR (Liu et al.,
7 2003, 089548). For a 0.5 ppm increase in 24-h CO concentration during the first month of pregnancy
8 there was an increased risk of IUGR (OR: 1.03 [95% CI: 1.00-1.05]) and this was of borderline
9 significance when CO was averaged over the first trimester (OR: 1.02 [95% CI: 1.00-1.05]). This
10 result persisted after controlling for other gaseous pollutants. Conversely, maternal exposure to CO
11 was not associated with LEW. The more recent of these 2 studies examined 386,202 singleton term
12 births (37-42-wk gestation) in Calgary, Edmonton and Montreal between 1986 and 2000 (Liu et al.,
13 2007, 090429). The study examined monthly and trimester exposures to CO with IUGR being the
14 only endpoint. A 0.5 ppm increase in 24-h CO concentration was associated with an increased risk of
15 IUGR in the first (OR: 1.09 [95% CI: 1.07-1.11]), second (OR: 1.07 [95% CI: 1.05-1.09]), and third
16 trimesters (OR: 1.09 [95% CI: 1.07-1.11]) of pregnancy. This result translated to CO exposure
17 having a positive effect on IUGR within each individual month of pregnancy with the highest effect
18 during the first and last months. This result persisted after controlling for concurrent NO2 and PM2.5.
19 Two studies in Sao Paulo, Brazil, a city with notably high levels of air pollution (mean CO
20 3.7 ppm) investigated associations between maternal exposures to CO in relation to reduced birth
21 weight and LEW within two consecutive time periods and found similar results. In both studies the
22 exposures were derived from a city-wide average across the available monitoring sites. The first
23 study examined 179,460 singleton term births during 1997 and found that a 0.75 ppm increase in 8-h
24 CO concentration averaged over the first trimester was associated with a 17.3 g (95% CI: 31.0-3.7)
25 reduction in birth weight (Gouveia et al., 2004, 055613). The second of these studies examined
26 311,735 singleton births (37-41-wk gestation) between 1998 and 2000 and reported a 6.0 g (95 % CI:
27 7.75-4.1) reduction in birth weight associated with a 0.5 ppm increase in 24-h CO concentration
28 averaged over the first trimester (Medeiros and Gouveia, 2005, 156750). It is important to note that
29 neither of these studies found an association between CO exposure and an increased risk of LEW.
30 Therefore, despite CO during the first trimester being associated with reduced birth weight, it was
31 not associated with LEW.
32 Similar to the two studies in Sao Paulo, Brazil, researchers in Seoul, South Korea conducted
33 two studies using data from two consecutive time periods. Both of these studies based the exposure
34 estimates on a city-wide average from all available fixed sites and as would be expected, the results
35 pertaining to CO were similar for both studies. Ha and colleagues (2001, 019390) examined
36 maternal exposures to CO during the first and third trimesters among 276,763 singleton term births
September 2009 5-77 DRAFT - DO NOT CITE OR QUOTE
-------
1 in Seoul between 1996 and 1997. Exposure to CO during the first trimester was associated with a
2 decrease in birth weight of 13.3 g, which also translated into an increased risk of LEW (RR: 1.10
3 [95% CI: 1.05-1.14] per 0.5 ppm increase in 24-h CO concentration). When Lee and colleagues
4 (2003, 043202) extended this study to include singleton term births for the period of 1996-1998 with
5 24-h CO concentrations averaged over each month of pregnancy and trimester, CO exposure during
6 the first trimester was associated with an increased risk of LEW (OR: 1.04 [95% CI: 1.01-1.07] per
7 0.5 ppm increase). No associations were found in the third trimester for any of the pollutants.
8 Monthly-specific exposures showed that the risk of LEW tended to increase with CO exposure
9 between months 2-5 of pregnancy.
10 In contrast to other studies reporting that early and late pregnancy are the critical periods for
11 CO exposure, a Sydney, Australia study of 138,056 singleton births between 1998-2000 reported a
12 reduction in birth weight of 21.7 g (95% CI: 38.2-5.1) and 17.2 g (95% CI: 33.4-0.9) associated with
13 a 0.75 ppm increase in maternal exposure to 8-h CO averaged over the second and third trimesters
14 respectively (Mannes et al., 2005, 087895). However, this result did not persist after controlling for
15 other pollutants (PMi0, NO2) and was only statistically significant when including births where the
16 mother resided within 5 km of a monitor. Furthermore, this result did not translate to an increased
17 risk of SGA, which was defined as a birth weight two standard deviations below the mean. The odds
18 ratios for SGA for CO exposures during the first, second and third trimesters were 0.96 (95% CI:
19 0.91-1.03), 0.99 (95% CI: 0.92-1.07), and 1.01 (95% CI: 0.93-1.08) per 0.75 ppm increase in 8-h
20 CO, respectively. While the majority of studies restrict the analyses to term births as a method of
21 controlling for gestational age, it is important to note that the Sydney study used all births and
22 controlled for gestational age in the birth weight analyses and SGA was derived for each gestational
23 age group.
24 Of all studies reviewed, only two did not find an association between maternal exposure to CO
25 and birthweight variables. In northern Nevada, Chen and colleagues (2002, 024945) examined CO,
26 PMio, and O3 exposures among a cohort of 39,338 term births (37-44-wk gestation) between 1991
27 and 1999 and found no association between CO exposure during the entire pregnancy (and each
28 trimester) and a reduction in birth weigh or an increased risk of LEW. For a 0.75 ppm increase in 8-h
29 CO concentration averaged over the entire pregnancy there was a reduction in birth weight of 6 g,
30 however it failed to reach statistical significance. Exposures for this study were based on data from
31 all monitoring sites across Washoe County, Nevada.
32 In a retrospective cohort study among 92,288 singleton term births (37-44-wk gestation) in
33 Taipei and Kaoshiung, Taiwan between 1995-1997, maternal exposures to CO, SO2, O3, NO2, and
34 PMio in each trimester of pregnancy were examined and only SO2 during the third trimester showed
35 evidence of contributing to LEW. Exposure assessment was based on data from the monitor closest
36 to the centroid of the mother's residential district and the final analyses only included mothers whose
September 2009 5-78 DRAFT - DO NOT CITE OR QUOTE
-------
1 district centroid was within 3 km of a monitor. CO exposures were grouped into low (-1.1 ppm),
2 medium (-1.2-15.0 ppm), and high (>15.0 ppm) and when compared to the lowest exposure group,
3 the odds ratio for LEW in the highest exposure group was 0.90 (95% CI: 0.75-1.09) for the first
4 trimester, 1.00 (95% CI: 0.82-1.22) for the second trimester, and 0.86 (95% CI: 0.71-1.03) for the
5 third trimester (Lin et al., 2004, 089827).
6 Table 5-13 provides a brief overview of the birth weight studies. In summary, there is evidence
7 of ambient CO during pregnancy having a negative effect on fetal growth. From the reviewed studies
8 Figure 5-9 shows the change in birth weight (grams), Figure 5-10 shows the effect estimates for
9 LEW, and Figure 5-11 shows the effect estimates for SGA. In general the reported reductions in birth
10 weight are small (ranging ~10-20g). It is difficult to conclude whether CO is related to a small
11 change in birth weight in all births across the population, or a marked effect in some subset of births.
12 Furthermore, there is a large degree of inconsistency across these studies. This may be due to several
13 factors such as inconsistent exposure assessment and statistical methods employed, different CO
14 concentrations, and/or different demographics of the birth cohorts analyzed. The main inconsistency
15 among these findings is the gestational timing of the CO effect. Although the majority of studies
16 reported significant effects during either the first or third trimester, other studies failed to find a
17 significant effect during these periods. Several studies found an association with exposure during the
18 entire pregnancy, providing evidence for a possible accumulative effect; however, these results are
19 inconclusive and this may be the result of correlated exposure periods.
20 Several studies examined various combinations of birth weight, LEW, and SGA/IUGR and
21 inconsistent results are reported across these metrics. For example, several studies reported an
22 association between maternal exposure to CO and decreased birth weight yet the decrease in birth
23 weight did not translate to an increased risk of LEW or SGA. However, it needs to be noted that a
24 measureable change, even if only a small one, on a population is different than an effect on a subset
25 of susceptible births which may increase the risk of IUGR/LBW/SGA.
26 The possibility exists that the small reductions in birth weight associated with maternal CO
27 exposures are the result of residual confounding associated with other factors (e.g., other pollutants,
28 temperature, and spatial/temporal variation in maternal factors) or other correlated pollutants. For
29 example, in some studies the CO effect did not persist after controlling for other pollutants (Mannes
30 et al., 2005, 087895): (Parker et al., 2005, 087462): (Wilhelm and Ritz, 2005, 088668) while in some
31 studies it did persist (Bell et al., 2007, 091059): (Gouveia et al., 2004, 055613): (Liu et al., 2003,
32 089548). and other studies did not report results from multipollutant models (Ha et al., 2001,
33 019390): (Lee et al., 2003, 043202): (Maisonet et al., 2001, 016624): (Medeiros and Gouveia, 2005,
34 156750). In addition, various methods have been employed to control for seasonality and trends
35 (e.g., month of birth, season of birth, year of birth, smoothed function of time), which may explain
36 some of the mixed results.
September 2009 5-79 DRAFT - DO NOT CITE OR QUOTE
-------
1 The two U.S. studies conducted in the Northeast compared results from analyses stratified by
2 race. The earlier of these studies found an association between CO and LEW among African
3 Americans but not among whites and Hispanics (Maisonet et al., 2001, 016624). In contrast, despite
4 reporting an llg reduction in birth weight among African-Americans and a 17 g reduction among
5 whites, the more recent of the two studies found no significant difference between these reductions
6 by race (Bell et al., 2007, 091059). Parker and colleagues (2005, 087462) also tested for interactions
7 between race and found no significant association.
Study
Location
Exposure
Period
Change in Birthweight (gms)
Bell etal (2007,091059)
Salamet al (2005. 087885)
Salamet al (2005.087885)
Salamet al (2005.087885)
Salamet al (2005.087885)
Chen et al (2002, 024945)
Chen et al (2002, 024945)
Chen et al (2002, 024945)
Chen et al (2002, 024945)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Gouveia et al (2004, 055613)
Gouveia et al (2004, 055613)
Gouveia et al (2004, 055613)
Medeiros and Gouveia (2005,1567501
Medeiros and Gouveia (2005,156750)
Medeiros and Gouveia (2005,156750)
CT&MA, USA
California, USA
California, USA
California, USA
California, USA
Entire pregnancy
First trimester
Second trimester
Third trimester
Entire pregnancy
Northern NV, USA First trimester
Northern NV, USA Second trimester
Northern NV, USA Third trimester
Northern NV, USA Entire pregnancy
Sydney, Australia First trimester
Sydney, Australia Second trimester
Sydney, Australia Third trimester
Sydney, Australia Last month
Sao Paulo, Brazil First trimester
Sao Paulo, Brazil Second trimester
Sao Paulo, Brazil Third trimester
Sao Paulo, Brazil First trimester
Sao Paulo, Brazil Second trimester
Sao Paulo, Brazil Third trimester
County wide
ZIP code level
ZIP code level
- ZIP code level
ZIP code level
County level
. County level
County level
County level
City level
City level
_ City level
City level
City level
City level
City level
Cityi level
City level
_^ City level
-35 -25 -15 -505 10 15 20
Figure 5-9 Summary of change in birth weight (95% confidence intervals) associated with
maternal exposure to ambient CO. Effect estimates have been standardized to a
1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-80
DRAFT - DO NOT CITE OR QUOTE
-------
Author
Belief al(2007,091059)
Maisonet et al (2001 , 016624)
Maisonet et al (2001 , 016624)
Maisonet et al (2001 , 016624)
Wilhelm & Ritz (2005, 088668)
Salametal (2005, 087885)
Salametal (2005, 087885)
Salametal (2005, 087885)
Salametal (2005. 087885)
Liu et al (2003, 089548)
Liu et al (2003, 089548)
Ha et al (2001 , 019390)
Ha et al (2001 , 019390)
Lee etal (2003. 043202)
Lee etal (2003. 043202)
Lee etal (2003, 043202)
Lee etal (2003, 043202)
Location
CT&MA, USA
Northeastern USA, USA
Northeastern USA, USA
Northeastern USA, USA
Los Angeles County, CA USA
California, USA
California, USA
California, USA
California, USA
Vancouver, Canada
Vancouver, Canada
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Exposure Effect Estjmate
Period
Entire preanancv _ o.. .„,.,....,
Third trimester
Third trimester
First trimester
Rer-nnri trimester f
Thirri trimester ,
Fntire prennannv m
First mo
Second month ,.
First trimester
Third trimester t C
First trimester
Second trimester
Third trimester m
Entire pregnancy
• nt" i°v°i
, Pity loi/ol
, ZIP code level
7IPr-nrielewel
7IPr-nHe lewel
ZIP code level
ZIP code level
, City level
,_ City level
t City level
ty level
9 City level
a City level
City level
— • — City level
1
0.75 1.00 1.2S
Figure 5-10 Summary of effect estimates (95% confidence intervals) for LEW associated with
maternal exposure to ambient CO. Effect estimates have been standardized to a
1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-81
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Salam et al (2005, 087885)
Salamet al (2005. 087885)
Salamet al (2005. 087885)
Salamet al (2005. 087885)
Liu et al (2003, 089548)
Liu et al (2003, 089548)
Liu et al (2003, 089548)
Liu et al (2003, 089548)
Liu et al (2003, 089548)
Liu et al (2007, 090429)
Liu et al (2007, 090429)
Liu et al (2007, 090429)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Mannes et al (2005, 087895)
Location
California USA
California, USA
California, USA
California, USA
Vancouver, Canada
Vancouver, Canada
Vancouver, Canada
Vancouver, Canada
Vancouver, Canada
3 Cities, Canada
3 Cities, Canada
3 Cities, Canada
Sydney, Australia
Sydney, Australia
Sydney, Australia
Sydney, Australia
Period Effect Estimate
First trimester
Second trimester __
ThirH trimester
Fntirp prpgnanm/
First mo
First trimester
Second trimester — (_
Third trimester t
Last mo _».
First trimester
Second trimester
Third trimester
First trimpstpr
Sprnnri trimpstpr §
Third trimester
Last mo
f 7IP code level
— ZIP code level
7iD™Hp level
» City level
_•_ City level
P City level
. City level
_ City level
, City level
, City level
, City level
City level
nit\; IPVPI
, City level
g City level
1
0.78 1.00 1.2S
Figure 5-11 Summary of effect estimates (95% confidence intervals) for SGA associated with
maternal exposure to ambient CO. Effect estimates have been standardized to a
1 ppm increase in ambient CO for 1-h max CO concentrations, 0.75 ppm for 8-h
max CO concentrations, and 0.5 ppm for 24-h avg CO concentrations.
September 2009
5-82
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-13 Brief summary of
Study
Outcomes
Examined
birth weight studies.
Location (Sample
Size)
Mean CO (ppm)
Exposure Assessment
Exposure Windows
UNITED STATES
Ritz and Yu (1999. 086976)
Wilhelm and Ritz (2005,
088668)
Salametal. (2005. 087885)
Parker et al. (2005, 087462)
Maisonet et al. (2001 , 016624)
Bell etal. (2007.091059)
Chen etal. (2002, 024945)
LBW
LBW
Birth weight
LBW
IUGR
Birth weight
SGA
LBW
Birth weight
LBW
Birth weight
LBW
Los Angeles, CA
(n = 125, 573)
Los Angeles County, CA
(n = 136, 134)
California
(n = 3901)
California
(n = 18,247)
Boston, MA; Hartford,
CT; Philadelphia &
Pittsburg, PA;
Washington DC
(n = 103, 465)
Connecticut and
Massachusetts,
(n = 358,504)
Northern Nevada,
(n = 36,305)
2.6 (6-9 a.m.)
1 .4 (24 h)
1 .8 (24 h)
0.75 (8 h)
1 .1 (24 h)
0.6 (24 h)
0.9 (8 h)
<2 mi of monitor
Varying distances from monitor
ZIP code level
<5 mi from monitor
City wide avg
County level avg
County level
Trimester 3
Trimesters 1,2,3
Entire pregnancy
Trimesters 1,2,3
Entire pregnancy
Trimesters 1,2,3
Trimesters 1,2, 3
Entire pregnancy
Trimesters 1 , 3
Trimesters 1,2,3
CANADA
Liu et al. (2003. 089548)
Liu et al. (2007. 090429)
LBW
IUGR
IUGR
Vancouver, Canada
(n = 229,085)
Calgary, Edmonton,
Montreal, Canada
(n = 386,202)
1 .0 (24 h)
1 .1 (24 h)
City wide avg
City wide avg
Trimester 1
Trimesters 1,2, 3
SOUTH AMERICA
Gouveia etal. (2004. 05561 3)
Medeiros and Gouveia (2005,
156750)
Birth weight
LBW
Birth weight
LBW
Sao Paulo, Brazil
(n = 179, 460)
Sao Paulo, Brazil
(n = 31 1,735)
3.7 (8 h)
3.0 (24 h)
(Presented in
graph)
City wide avg
City wide avg
Trimesters 1,2,3
Trimesters 1,2, 3
AUSTRALIA/ASIA
Ha et al. (2001 , 019390)
Lee etal. (2003. 043202)
Mannes et al. (2005. 087895)
Lin et al. (2004, 089827)
Birth weight
LBW
LBW
Birth weight
SGA
LBW
Seoul, Korea
(n = 276,763)
Seoul, Korea
(n = ?)
Sydney, Australia
(n = 138,056)
Taipei, Kaoshiung,
Taiwan
(n = 92,288)
1 .2 (24 h)
1 .2 (24 h)
0.8 (8 h)
Taipei 1 .1 ,
Kaoshiung 8.1
City wide avg
City wide avg
City wide avg and
<5 km from monitor
<3 km of monitor
Trimesters 1 and 3
Entire pregnancy
Trimesters 1,2,3
Trimesters 1,2,3
Last 30 days
Entire pregnancy
Trimesters 1,2,3
September 2009
5-83
DRAFT - DO NOT CITE OR QUOTE
-------
5.4.1.3. Congenital Anomalies
1 Despite the growing evidence of an association between ambient air pollution and various
2 adverse birth outcomes, fewer studies have investigated the effect of temporal variations in ambient
3 air pollution on congenital anomalies. Given the higher prevalence and associated mortality, heart
4 defects have been the main focus of the majority of these recent air pollution studies. The other
5 study's focus was cleft lip/palate.
6 The first of these studies was conducted in southern California (Ritz et al, 2002, 023227).
7 Exposure to ambient CO, NO2, O3 and PMi0 during each of the first three months of pregnancy was
8 examined among births during 1987-1993. Maternal exposure estimates were based on data from the
9 fixed site closest to the mother's ZIP code area and when using a case-control design where cases
10 were matched to 10 randomly selected controls, results showed that CO during the second month of
11 pregnancy was associated with cardiac ventricular septal defects. The CO exposures were grouped
12 by quartiles (25th = 1.14, 50th = 1.57, 75th = 2.39 ppm) and when compared to those in the lowest
13 quartile exposure group (<1.14 ppm), the odds ratios for ventricular septal defects across the 3
14 exposure groups were 1.62 (95% CI: 1.05-2.48), 2.09 (95% CI: 1.19-3.67), and 2.95 (95% CI: 1.44-
15 6.05) respectively. In a multipollutant model a similar exposure-response pattern was exhibited
16 across the quartiles with the highest quartile of exposure reaching statistical significance (OR: 2.84
17 [95% CI: 1.15-6.99]). The only other pollutant associated with a defect was O3 during the second
18 month of pregnancy, which was associated with aortic artery and valve defects.
19 Another study was conducted in Texas (Gilboa et al., 2005, 087892), where exposure to
20 ambient CO, NO2, SO2, O3 and PMi0 during the 3rd to 8th week of gestation was examined among
21 births between 1997-2000. Maternal exposure estimates were calculated by assigning the data from
22 the closest monitor to the mother's residential address. If data were missing on a particular day then
23 data from the next closest site were used. The median distances from a monitor ranged from 8.6-14.2
24 km with maximum distances ranging from 35.5-54.5 km. The main results showed that CO was
25 associated with multiple conotruncal defects and Tetralogy of Fallot. CO exposures were grouped
26 into quartiles of much lower concentrations (25th = 0.4, 50th = 0.5, 75th = 0.7 ppm) than the
27 California study (Ritz et al., 2002, 023227) and when compared to the lowest quartile, the odds
28 ratios for conotruncal defects across the 3 CO exposure groups were 1.38 (95% CI: 0.97-1.97), 1.17
29 (95% CI: 0.81-1.70), and 1.46 (1.03-2.08) respectively without a significant test for trend (p for trend
30 = 0.0870). A strong exposure-response pattern was exhibited across the quartiles of CO exposure for
31 Tetralogy of Fallot (25th OR: 0.82 [95% CI: 0.52-1.62]; 50th OR: 1.27 [95% CI: 0.75-2.14]; 75th
32 OR: 2.04 [95% CI: 1.26-3.29]; p for trend = 0.0017). The only significant associations found with
33 other pollutants were between PMi0 and isolated atrial septal defects, and SO2 and ventricular septal
34 defects.
September 2009 5-84 DRAFT - DO NOT CITE OR QUOTE
-------
1 A study conducted in Atlanta, GA investigated the associations between ambient air pollution
2 concentrations during weeks 3-7 of pregnancy and risks of cardiovascular malformations among a
3 cohort of pregnancies conceived during 1986-2003 (Strickland et al., 2009, 190324). The mean 24-h
4 CO concentration during this period was 0.75 ppm. The authors did not report any statistically
5 significant associations with ambient CO concentrations and cardiac malformations, though there
6 were elevated risk ratios for ambient CO concentration and patent ductus arteriosus, Tetralogy of
7 Fallot, and right ventricular outflow tract defect. These results remained consistently positive in five
8 sensitivity analyses conducted, and were closer to achieving statistical significance in these
9 sensitivity analyses. The only statistically significant results were for the association between PMi0
10 and patent ductus arteriosus.
11 The last of these studies was a case-control study that examined maternal exposure to various
12 air pollutants during the first three months of pregnancy and the risk of delivering an infant with an
13 oral cleft, namely cleft lip with or without palate (CL/P). Birth data from the Taiwanese birth registry
14 from 2001-2003 was linked to air pollutant data that were spatially interpolated from all fixed
15 monitoring sites across Taiwan. Based on data at the center of the townships or districts, exposure
16 estimates for PMi0, SO2, NOX, O3, and CO were averaged over each of the first three months of
17 pregnancy. The mean 8-h avg CO concentration was 0.69 ppm. Interestingly, of all the pollutants
18 examined, only O3 during the first two months of pregnancy was significantly associated with an
19 increased risk of CL/P. In multipollutant models CO was not associated with CL/P (Hwang and
20 Jaakkola, 2008, 193794).
21 The main results from the southern California study showed that CO was associated with an
22 increased risk of ventricular septal defects and this was exhibited by an exposure-response pattern
23 across the quartiles of exposure, yet there was no indication that ambient CO concentration in Texas
24 was associated with ventricular septal defects. Conversely, ambient CO concentration in Texas was
25 associated with an increased risk of conotruncal defects, yet there was no indication that CO in
26 southern California was associated with conotruncal defects. The Atlanta study (Strickland et al.,
27 2009, 190324) found positive, though not statistically significant associations for patent ductus
28 arteriosus, Tetralogy of Fallot, and right ventricular outflow tract defect. The elevated risk ratio for
29 Tetrology of Fallot is consistent with the result observered in Texas (Gilboa et al., 2005, 087892).
30 Interestingly, similar inconsistencies were also found for PMi0 between these studies. For
31 example, PMi0 in Texas was associated with an increased risk of atrial septal defects and with patent
32 ductus arteriosus in Atlanta, GA, yet there was no indication of such an effect in southern California
33 where PMi0 concentrations were markedly higher.
34 The authors of the Texas study (Gilboa et al., 2005, 087892) provide little discussion toward
35 the inconsistent results with the southern California study. One suggestion is the different CO
36 concentrations across the studies with the 75th quartile in southern California being 2.39 ppm while
September 2009 5-85 DRAFT - DO NOT CITE OR QUOTE
-------
1 in Texas it was much lower at 0.7 ppm. However, this suggests that different defects are associated
2 with different concentrations of CO, yet it still does not explain why particular associations were
3 reported in Texas and not southern California where concentrations were higher. Similarly, the
4 authors of the Texas study (Gilboa et al., 2005, 087892) also suggested the inconsistency was due to
5 different exposure periods. In Texas the exposures were averaged over the 3rd to 8th week while in
6 southern California the exposures were averaged over the second month of pregnancy. However,
7 there was no reason provided as to why this small difference in the examined exposure period would
8 explain the inconsistent results.
9 Overall, there is some evidence that maternal exposure to CO is associated with an increased
10 risk of congenital anomalies, namely heart defects and cleft lip and palate. Further research is
11 required to corroborate these findings.
5.4.1.4. Neonatal and Post-Neonatal Mortality
12 A handful of studies examined the effect of ambient air pollution on neonatal and post-
13 neonatal mortality with the former the least studied. These studies varied somewhat with regard to
14 the outcomes and exposure periods examined, and study designs employed.
Neonatal
15 In Sao Paulo, Brazil, a time-series study examined daily counts of neonatal (up to 28 days
16 after birth) deaths for the period of 1998-2000 in association with concurrent day exposure to SO2,
17 CO, O3, and PM10. Moving averages from 27 days were examined. The mean city-wide CO
18 concentration was 2.8 ppm and there was no association between daily ambient CO and neonatal
19 deaths. Despite CO being correlated with PM10 (r = 0.71) and SO2 (r = 0.55), only PM10 and SO2
20 were associated with an increase in the daily rate of neonatal deaths (Lin et al., 2004, 095787).
21 In another study of neonatal death, Hajat et al. (2007, 093276) created a daily time-series of air
22 pollution and all infant deaths between 1990 and 2000 in 10 major cities in England. The mean daily
23 CO concentration across the ten cites was 0.57 ppm. This study provided no evidence for an
24 association between ambient CO concentration and neonatal deaths.
Post-Neonatal
25 Two studies in the U.S. examined the potential association between ambient CO and post-
26 neonatal (from 28 days to 1 yr after birth) mortality and inconsistent results were reported. These
27 studies, however, varied somewhat in study design.
28 The first of these studies employed a case-control design and examined all infant deaths
29 during the first year of life among infants born alive during 1989-2000 within 16 km from a
September 2009 5-86 DRAFT - DO NOT CITE OR QUOTE
-------
1 monitoring site within the South Coast Air Basin of California. Exposures for 2-wk, 1-mo, 2-mo, and
2 6-mo periods before death were linked to each individual death. Extensive analyses were conducted
3 for all-cause infant deaths, respiratory causes of death, and sudden infant death syndrome (SIDS).
4 Given the long time period of the data analyzed, in order to alleviate the confounding trends in infant
5 mortality and CO levels this study was able to match by year (Ritz et al, 2006, 089819). Ambient
6 1-h max CO concentrations averaged over the 2 mo before death were associated with an 11% (OR:
7 1.11 [95% CI: 1.06-1.16]) increase in risk of all-cause post-neonatal death (per 1 ppm increase) and
8 a 19% (OR: 1.19 [95% CI: 1.10-1.28]) increase in risk of SIDS. In the multipollutant models
9 (including PMi0, NO2, O3) the positive CO mortality effect decreased by around 50% and was not
10 statistically significant. Based on exposure from 2 wk before death, CO was associated with an
11 increased risk of respiratory related post-neonatal deaths occurring 28 days to 1 yr after birth (OR:
12 1.14 [95% CI: 1.03-1.25] per 1 ppm increase) and 28 days to 3 mo after birth (OR: 1.20
13 [95% CI: 1.02-1.40]), but no effect was observed for respiratory related deaths occurring 4-12 mo
14 after birth. These results persisted in the multipollutant models and exposure-response patterns were
15 exhibited across the exposures groupings of 1.02 to <2.08, and > 2.08 ppm. To control for gestational
16 age and birth weight the analyses were stratified by 'term/normal-weight infants' and 'preterm
17 and/or LEW infants.' When these two strata were analyzed, CO was associated with an increased
18 risk of all-cause death and SIDS within both strata (ORs ranged from 1.12 to 1.46). However, these
19 effects did not persist in multipollutant models (Ritz et al., 2006, 089819).
20 Another study examined 3,583,495 births, including 6,639 post-neonatal deaths occurring in
21 96 counties throughout the U.S. (in counties with more than 250,000 residents) between 1989 and
22 2000 (Woodruff et al., 2008, 098386). Only exposure during the first two months of life was
23 examined and this was based on an average of CO concentrations recorded across all available
24 monitors within the mother's county of residence. In contrast to the other postnatal mortality study in
25 California, CO averaged over the first two months of life was not associated with all-cause death
26 (OR: 1.01 [95% CI: 0.94-1.09] per 0.5 ppm increase in 24-h CO concentration), or with respiratory
27 related deaths (OR: 1.08 [95% CI: 0.91-1.54] per 0.5 ppm increase in 24-h CO concentration), SIDS
28 (OR 0.85 [95% CI: 0.70-1.04] per 0.5 ppm increase in 24-h CO concentration), or other causes of
29 post-neonatal mortality (OR: 1.03 [95% CI: 0.96-1.09] per 0.5 ppm increase in 24-h CO
30 concentration). These null findings may be due to higher error of the exposure assessment at the
31 county-level as opposed to using data from monitors within close proximity to the residence.
32 In a study that included 10 major cities in England, Hajat et al. (2007, 093276) created a daily
33 time-series of air pollution and all infant deaths between 1990 and 2000. While there was no
34 evidence for an association with neonatal deaths and ambient CO concentrations, there was a strong
35 adverse effect of CO in post-neonatal deaths, although the confidence intervals were wide due to a
36 small sample size (RR 1.09, 95% CI: 0.94-1.25).
September 2009 5-87 DRAFT - DO NOT CITE OR QUOTE
-------
1 The only other postnatal mortality studies have been conducted throughout Asia. Two identical
2 studies in Taiwan failed to find an association between daily counts of post-neonatal deaths and
3 ambient air pollutants, including CO. The data analyzed were from the cities of Taipei (Yang et al.,
4 2006, 090760) and Kaohsiung (Tsai et al., 2006, 090709) with ambient CO concentrations being
5 1.6 ppm and 0.8 ppm respectively. Both studies examined deaths for the period of 1994-2000 and
6 employed a case-crossover design that compared air pollution levels 1 wk before and after each
7 infant's death.
8 Similarly, another study in South Korea examined post-neonatal mortality for the period of
9 1995-1999 using a time-series design. Same-day CO was not associated with all-cause death (RR:
10 1.02 [95% CI: 0.97-1.06] per 0.5 ppm increase). However, same-day CO was associated with post-
11 neonatal mortality when the analyses were restricted to respiratory mortality (RR: 1.33
12 [95% CI: 1.01-1.76] per 0.5 ppm increase) (Ha et al., 2003, 042552). An additional study examined
13 the relationship between air pollution and postneonatal mortality for all causes in Seoul, Korea. This
14 study used both case-crossover and time- series analyses for all firstborn infants during 1999-2003.
15 The mean 8-h max CO concentration during this time period was 1.01 ppm. The association between
16 ambient CO concentration and postneonatal mortality was the strongest in magnitude for CO when
17 compared to the other criteria pollutants, though the confidence intervals were wide (RR: 1.02 [95%
18 CI: 0.87-1.20] for case-crossover analysis; RR: 1.23 [1.06-1.44] for time-series analysis per
19 0.75 ppm increase in 8-h max CO concentration).
20 In general, the inconsistent exposure periods examined among these studies allows for limited
21 direct comparison and interpretation. Nevertheless, there is limited evidence that CO is associated
22 with an increased risk of infant mortality during the post-neonatal period. The exposure periods
23 examined varied from the same-day CO to lag periods up to a 6-mo period prior to birth with one
24 study alternatively exploring exposures averaged over the first two months of life. Furthermore,
25 given that birth weight and gestational age are strong predictors of infant mortality, in all of the
26 reviewed studies these factors have not been considered at either the design or analysis stage. Hence,
27 the link between fetal exposures, neonatal exposures, and post-neonatal exposures, and the possible
28 interaction that birth weight and gestational age may have on the results yielded from these
29 examined exposure periods, needs further attention within this field of research.
5.4.1.5. Summary of Epidemiologic Studies of Birth Outcomes and Developmental
Effects
30 There is some evidence that CO during early pregnancy (e.g., first month and first trimester) is
31 associated with an increased risk of PTB. Additionally, there is evidence of ambient CO during
32 pregnancy having a negative effect on fetal growth. In general, the reviewed studies (Figure 5-9
33 through Figure 5-11) reported small reductions in birth weight (ranging -10-20 g). Although the
September 2009 5-88 DRAFT - DO NOT CITE OR QUOTE
-------
1 majority of studies reported significant effects during either the first or third trimester, other studies
2 failed to find a significant effect during these periods. Several studies examined various
3 combinations of birth weight, LEW, and SGA/IUGR and inconsistent results are reported across
4 these metrics. For example, six studies reported an association between maternal exposure to CO and
5 decreased birth weight yet the decrease in birth weight did not translate to an increased risk of LEW
6 or SGA. It should be noted that having a measurable, even if small, change in a population is
7 different than having an effect on a subset of susceptible births, which may increase the risk of
8 IUGR/LBW/SGA. It is difficult to conclude if CO is related to a small change in birth weight in all
9 births across the population, or a marked effect in some subset of births.
10 Three studies examined the effects of CO on cardiac birth defects and found maternal
11 exposure to CO to be associated with an increased risk of cardiac birth defects. While there was
12 limited coherence for the specific cardiac malformations associated with CO exposure in these
13 studies, this insult to the heart is coherent with the CO effects on the heart characterized in
14 Section 5.2. In general, there is limited evidence that CO is associated with an increased risk of
15 infant mortality during the post-neonatal period.
5.4.2.lexicological Studies of Birth Outcomes and Developmental
Effects
16 The brief overview of the reproductive and development toxicology of CO that follows is not
17 limited to the past 10 yr as are other areas discussed in this document. This is because reproductive
18 and developmental toxicology endpoints have not been covered in previous CO AQCDs. Effects of
19 both exogenous CO exposure and endogenous production of CO are discussed since exposure to
20 exogenous CO could possibly alter pathways normally regulated by endogenous CO production.
21 This document details how in utero or perinatal CO exposure in pregnant dams or pups affects
22 outcomes in the offspring including post-natal mortality, skeletal development, the ability of the
23 developing fetus to tolerate maternal dietary manipulation, behavioral outcomes, neurotransmitters,
24 brain development, the auditory system, myocardial development, and immune system development.
25 Similarly, endogenous CO is discussed in relation to pregnancy maintenance, vascular tone during
26 gestation, the placenta, the ovaries, the anterior pituitary axis, and lactation. Together, this
27 toxicological summary documents the importance of CO in reproductive and developmental
28 toxicology in laboratory animal models.
September 2009 5-89 DRAFT - DO NOT CITE OR QUOTE
-------
5.4.2.1. Birth Outcomes
Decreased Birth Weight
1 Multiple reports have been published associating CO exposure in laboratory animals and
2 decrements in birth weight (90-600 ppm); some of these studies also noted reduced growth evident
3 in the prenatal period (65-500 ppm CO).
4 Multiple studies have reported decreased body weights in pups collected near term. Significant
5 decreases in fetal body weight at GD21 after 21 days of continuous CO exposure (125, 250, or
6 500 ppm) in pregnant Wistar rats have been reported (Prigge and Hochrainer, 1977, 012326). This
7 decrease was not found in rats exposed to 60 ppm CO. Penney et al. (1983, 011385) exposed
8 pregnant rats to CO (200 ppm) for the final 17 days of prenatal development and also found
9 significant decreases in near-term fetal rat weight at GD20-21; gestation in rats is ~ 22 days. Penney
10 et al. continued to find decreased body weight to PND210 after postnatal CO exposure (500 ppm,
11 PND1-32), and to a larger extent in male pups when compared to female pups (Penney et al., 1982,
12 011387). Singh et al. (1984, 011409: 1993, 013892) found significant decreases in fetal weight in
13 gestationally CO-exposed mouse pups (65, 125, 250 or 500 ppm) in two studies. Near-term fetal
14 body weight was decreased at GD18 in mice exposed from GD7-18 to 125, 250, and 500 ppm CO,
15 but not 65 ppm CO (Singh and Scott, 1984, 011409). However, a second study found decreased fetal
16 weight at GD18 with all CO exposures (65-500 ppm) from GD8-18 (Singh et al., 1993, 013892).
17 A number of studies have found decreases birth weight after CO exposure. Fechter and Annau
18 (1977, 010688) exposed pregnant rats to 150 ppm CO (dam COHb 15%) continuously during
19 gestation via inhalation and found a 5% decrease in birth weight in PND1 pups versus control
20 animals with weight decrements measurable to weaning (PND4: 16% decrease; PND21: 13%
21 decrease); in this study, lactational cross fostering did not ameliorate these reduced growth rates,
22 indicating that maternal postnatal contributions from CO exposure did not affect these growth rates.
23 Decreased birth weight and pre-weaning weight were seen in CO-exposed pups despite a lack of
24 weight decrement in CO-exposed dams. A decrease in body weight at birth was also seen in neonates
25 of pregnant rats exposed to 157, 166, and 200 ppm CO over GD6-GD19 (Penney et al., 1983,
26 011385). Singh et al. (2006, 190512) showed decreases in birth weight of mouse pups gestationally
27 exposed for 6 h/day for the first 2 wk of pregnancy to 125 ppm, but not 65 ppm. Carmines et al.
28 (2008, 188440) exposed Sprague-Dawley rats to -600 ppm CO (dam COHb 30%) via nose-only
29 inhalation (levels similar to those seen in cigarette smoke) during GD6-GD19 of gestation for 2
30 h/day and found significant decreases in birth weight (0.5 g or 13%) of exposed pups versus
31 controls. Maternal body weight was unchanged through gestation, but corrected terminal body
32 weight (body weight minus uterine weight) was significantly elevated in CO-exposed dams at term,
September 2009 5-90 DRAFT - DO NOT CITE OR QUOTE
-------
1 indicating a decrease in uterine weight. Other studies have not found decreases in birth weight after
2 gestational CO exposure (Carratu et al., 2000, 015839: Mereu et al., 2000, 193838).
3 Other animal models have been used to examine decreased birth weight resulting from CO
4 exposure. Astrup et al. (1972, 011121) found significant decreases (11 and 20%, respectivelty) in
5 birth weight of rabbits exposed to either 90 or 180 ppm CO continuously over the duration of
6 gestation. Tolcos et al. (2000, 015997) found significant decreases in body, brain, and liver weights,
7 and crown to rump length in guinea pig fetuses after exposure to 200 ppm CO for 1 Oh/day from
8 GD23-GD25 until GD61-GD63, at which time the fetuses were collected (term ranges from GD68 to
9 GD72). In other studies, there was no significant differences in birth weight of guinea pig pups after
10 a similar exposure (200 ppm from GD23-GD25 to term, fetal and maternal COHb levels of 13% and
11 8.5%, respectively) (McGregor et al., 1998, 085342: Tolcos et al., 2000, 010468) or in Long Evans
12 rats (150 ppm CO continuous exposure over all of gestation) (Fechter and Annau, 1977, 010688).
13 Fetal mouse weight was significantly greater than control in the 7 h/day exposures and significantly
14 less than control animals in the 24 h/day (250 ppm CO, GD6-GD15) exposure groups with
15 corresponding significant differences in crown to rump length in the two groups (Schwetz et al.,
16 1979, 011855). However, as neonates animals that showed no decrement in birth weight were
17 significantly smaller at PND4 compared to control guinea pigs (McGregor et al., 1998, 085342) with
18 dam and fetal COHb levels were 13% and 8.5%, respectively during pregnancy.
Pregnancy Loss and Perinatal Death
19 Two studies have provided evidence for CO-induced pregnancy loss and perinatal death at CO
20 concentrations between 90-250 ppm. Schwetz et al. (1979, 011855) exposed CF-1 mice and New
21 Zealand rabbits to 250 ppm CO over GD6-GD15 (mice) or GD6-GD18 (rabbits) for either 7 h/day or
22 24 h/day, yielding 4 exposure paradigms. The fetuses were then collected at the termination of
23 exposure, near term. Maternal COHb in the 7 h/day exposure groups was approximately 10-15%
24 COHb in rabbits and mice; COHb was not followed in the 24 h exposure groups. The mice exposed
25 to CO for 7 h/day, but not 24 h/day, had a significant increase in the number of resorbed pups.
26 Rabbits were less affected by CO exposure manifesting no significant perinatal death or pregnancy
27 loss. Astrup et al. (1972, 011121) studied the effect of CO on fetal development after continuous CO
28 exposure (90 or 180 ppm CO) over the duration of gestation in rabbits. COHb was 8-9% and 16-18%
29 in the 90 and 180 ppm exposure groups, respectively. In the immediate neonatal period, 24 h
30 postpartum, 35% (180 ppm) and 9.9% (90 ppm) of CO-exposed animals died. In the postpartum
31 period after the first 24 h and extending out to PND21, 90 ppm CO-exposed pups experienced 25%
32 mortality versus 13% in controls; there was no difference from control at the 180 ppm CO exposure
33 level. Gestation length was unchanged with CO exposure. Conversely, Fechter and Annau (1977,
September 2009 5-91 DRAFT - DO NOT CITE OR QUOTE
-------
1 010688) exposed Long-Evans rats in utero to 150 ppm CO continuously through gestation (dam
2 COHb 15%) and saw no effects of CO on litter mortality at PND1.
Effect of Maternal Diet
3 As mentioned above, CO induced offspring mortality after prenatal exposure. Alterations in
4 maternal dietary protein and zinc further exacerbated offspring mortality and teratogenicity caused
5 by CO (65-500 ppm).
Maternal Protein Intake and Neonatal Mouse Mortality and Teratogenicity
6 Pregnant CD-I mice were exposed intermittently (6 h/day for first 2 wk of pregnancy) to CO
7 (0, 65, or 125 ppm) in combination with protein modified diets [27% (supplemental protein), 16%
8 (control), 8% (low), or 4% (very low protein)] to assess the role of dietary protein in modulating CO
9 effects on neonatal mortality at 1 wk of age (Singh, 2006, 190512). Litter size was not affected by
10 CO exposure. Pup weight was inversely related to CO exposure and directly related to dam diet
11 protein content during pregnancy. Pup mortality at birth was directly related to CO exposure in
12 certain protein groups (supplemental, and 4% protein) and inversely related to the dam's dietary
13 protein content. At 1 wk of age, pup mortality was significantly increased by CO-exposure as well as
14 dietary protein restriction; all pups in the 4% protein diet died by 1 wk of age. CO exposure (65 ppm
15 only) combined with a normal protein diet (16%) and CO exposure (65 and 125 ppm) with a
16 supplemental protein diet (27%) significantly increased pup mortality at 1 wk versus control air pups
17 (0 ppm CO). Contrary to other findings, low protein diet (8%) combined with CO (125 ppm) led to a
18 slight yet significant decrease in pup mortality at 1 wk of age versus control (0 ppm CO). In
19 summary, these data show that in utero CO exposure induced increased neonatal mouse deaths at
20 1 wk in supplemental protein and normal protein diet exposure groups and increased perinatal
21 mortality when combined with supplemental or restricted protein.
22 The role of diet as a contributor to teratogenicity of CO (0, 65, 125, or 250 ppm CO) in CD-I
23 mice given various protein diets (4, 8, 16, or 27% protein) during pregnancy was explored by Singh
24 et al. (1993, 013892). Timed pregnant CD-I mice were exposed continuously to CO from GD8-
25 GDI8 at which point animals were sacrificed and fetuses collected. Work by this group has shown
26 that low protein diets plus CO exposure act in an additive fashion to increase placental COHb in
27 mice (Singh, 2003, 053624; Singh et al., 1992, 013759). As expected, all levels of CO and the lowest
28 protein diet (4 or 8% protein) given to the dams during gestation resulted in significantly decreased
29 near-term weight of normal fetuses at GDI 8. CO exposure did not produce maternal toxicity except
30 for a significant decrease in maternal weight at GDI8 with 4 and 8% protein diets versus control diet
31 in non-CO-exposed animals. Dam dietary protein levels were inversely related to gross fetal
32 malformations including jaw changes. All concentrations of CO exposure within each maternal
September 2009 5-92 DRAFT - DO NOT CITE OR QUOTE
-------
1 dietary protein level significantly increased the percentage of litters with malformations in a dose-
2 dependent manner. Skeletal malformations were present in offspring with the percent of litters
3 affected inversely related to dietary protein levels. CO exposure concomitant with a low protein diet
4 exacerbated the percent of skeletal malformations in offspring. The percent of dead, resorbed, or
5 grossly malformed fetuses was directly related to CO concentration and inversely related to maternal
6 dietary protein levels. CO and maternal dietary protein restriction had a synergistic effect on mouse
7 offspring mortality and an additive effect on malformations.
Maternal Zinc and Protein Intake and Neonatal Mortality and Teratogenicity
8 Singh et al. (2003, 053624) explored how teratogenicity and fetal mortality were affected by
9 zinc (Zn) modulation in CO-exposed (500 ppm from GD8-GD18) pregnant dams (CD-I mouse)
10 given protein insufficient diets. CO exposure in low protein conditions (9% protein) decreased the
11 mean implants per litter as compared to air exposure. CO exposure also increased the near-term fetal
12 mortality over all groups, and to a larger extent in the low protein groups, both Zn normal (57%
13 versus 6% mortality) and Zn deficient groups (86.6% versus 70.9% mortality). Under low protein
14 conditions, CO exposure increased the incidence of malformations (9.4% versus 0%) when Zn levels
15 were normal and increased the incidence of gastroschisis (5% versus 0%) when Zn levels were low.
16 Joint protein and Zn deficiency led to 60% of litters with gastroschisis. Conversely, CO exposure
17 under Zn deficiency decreased the incidence of other malformations such as exencephaly, jaw,
18 syndactyly, and tail malformations.
19 Further studies by Neggers and Singh (2006, 193964) only partially confirmed these findings.
20 As before, diets deficient in both Zn and protein had significant detrimental influence on both fetal
21 malformations and mortality. Exposure to 500 ppm CO increased fetal mortality and malformation
22 rates under deficient protein (9%) and supplemental Zn (3.3 g/kg diet) conditions; however CO had a
23 negligible effect on these endpoints under deficient protein and deficient or normal Zn conditions.
Role of Endogenous CO
24 CO is produced endogenously from heme protein catabolism by heme oxygenases, HO-1, HO-
25 2, and HO-3. CO has recently been recognized as a second messenger signaling molecule, similar to
26 NO, with a number of normal physiological roles in the body. Some of these roles are played in
27 maintaining pregnancy, controlling vascular tone, regulating hormone balance, and sustaining
28 normal follicular maturation. These areas could be potential areas of interaction of exogenous CO.
Pregnancy Maintenance
29 HO-1 is known to protect organs from rejection (Kotsch et al., 2006, 193899) and thus, HO
30 may also protect the developing fetus from rejection by the non-self maternal immune system.
September 2009 5-93 DRAFT - DO NOT CITE OR QUOTE
-------
1 Idiopathic spontaneous abortions are more frequent in women with HO-1 polymorphisms (GT)n
2 microsatellite polymorphisms associated with altered HO-1 transcription) in their genome
3 (Denschlag et al., 2004, 193894). Similarly, administering HO-inhibitors to pregnant rodents induced
4 total litter loss, possibly due to vasoconstriction and associated ischemia of the placental vascular
5 bed (Alexandreanu and Lawson, 2002, 192373). Also, mice over-expressing HO-1 had a
6 significantly decreased rate of spontaneous abortion (Zenclussen et al., 2006, 193873). Various
7 pathologies of pregnancy, including intrauterine growth restriction and pre-eclampsia, are associated
8 with significant decreases in placental HO activity (Denschlag et al., 2004, 193894; McLaughlin et
9 al., 2003, 193904). Oxygenation is important in early pregnancy and triggers trophoblast invasion of
10 the spiral arteries (Kingdom and Kaufmann, 1997, 193897). Women living at high altitude have an
11 increased risk of adverse pregnancy outcomes versus women living at lower altitudes (Zamudio et
12 al., 1995, 193908). Also, women living at high altitude, women with pre-eclampsia, or women who
13 had pregnancies with fetal growth restrictions (FOR) produced term placenta with significant
14 decreases in HO-2 versus women living at lower altitude with uncomplicated pregnancies (Barber et
15 al., 2001, 193891: Lyall et al., 2000, 193902). Thus, the HO/CO system is crucial for the developing
16 fetus, helps in maintaining pregnancy, and plays a role in spontaneous abortions.
Vascular Control
17 During pregnancy, there is increased blood volume without a concurrent increase in systemic
18 BP, which is accomplished by a decrease in total peripheral vascular resistance (Zhao et al., 2008,
19 193883). CO through the production of soluble guanylate cyclase is able to stimulate the relaxation
20 of vascular smooth muscle (Villamor et al., 2000, 015838) and relaxation of pregnant rat tail artery
21 and aortic rings (Longo, et al., 1999, 011548). Further, the administration of the HO inhibitor SnMP
22 increased maternal BP (systolic, diastolic, and mean arterial pressure) and significantly increased
23 uterine artery blood flow velocity during pregnancy in mice (Zhao et al., 2008, 193883). Zhao et al.
24 also showed pregnancy induced increased total body CO exhalation, and that this increased CO
25 production could be significantly decreased by SnMP administration. Abdominal aortas (AA) of
26 pregnant dams are significantly dilated with pregnancy and SnMP treatment leads to AA
27 vasoconstriction to levels similar to non-pregnant mice. Isolated human placenta exposed to
28 solutions containing CO demonstrated a concentration-dependent decrease in perfusion pressure
29 (Bainbridge et al., 2002, 043161) further demonstrating the role of CO in maintaining basal
30 vasculature tone. However, the addition of exogenous CO to isolated human and rat uterine tissue
31 during pregnancy failed to induce relaxation and quiet the spontaneous contractility of rat or human
32 myometrium (uterine smooth muscle)(Longo, et al., 1999, 011548). CO is not able to relax all types
33 of vascular smooth muscle (Brian et al., 1994, 076283). and pregnancy appears to modulate the
34 response of tissues to CO (Katoue et al., 2005, 193896). Thus, it appears that the increased CO
September 2009 5-94 DRAFT - DO NOT CITE OR QUOTE
-------
1 production during pregnancy may partially account for the decreased peripheral vascular resistance
2 seen in pregnancy that prevents the increased blood volume of pregnancy from affecting BP.
Hormone Regulation
3 Endogenous CO has been shown to regulate neuroendocrine functions. Disruption of normal
4 CO signaling causes changes in the cycles of a number of hormones involved in pregnancy. HO
5 inhibition in rats significantly decreased ovarian production of gonadotrophin-induced
6 androstenedione and progesterone without affecting estradiol levels (Alexandreanu and Lawson,
7 2002, 192373). However, treatment with the HO inducer, hemin, caused androstenedione and
8 estradiol production from rat ovaries in vitro. CO also has been shown to have a stimulatory effect
9 on gonadotropin-releasing hormone (GnRH) release from rat hypothalamic explants in vitro (Lamar
10 et al, 1996, 190997). while in vivo CO appears not to influence GnRH secretion (Kohsaka et al,
11 1999, 191000). HO-1 induction and HO concentration have been shown to be regulated by estrogen
12 in the rat uterus (Cella et al., 2006, 193240) during pregnancy and in non-gravid rats. This agrees
13 with work by (Tschugguel et al., 2001, 193785) in which CO was generated by primary endothelial
14 cells from human umbilical veins and uterine arteries after exogenous 17-(3 estradiol administration.
15 HO inhibition by CrMP decreased time in estrous in a dose-dependent manner (Alexandreanu and
16 Lawson. 2002. 192373).
17 HO-1 and HO-2 are expressed in rat anterior pituitary and the secretion of gonadotropins and
18 prolactin is affected by HO inhibitor and HO substrate administration (Alexandreanu and Lawson,
19 2003, 193871). The estrogen-induced afternoon surge of luteinizing hormone (LH) was advanced
20 forward in time by HO inhibition and this advance could be reversed by concomitant administration
21 of hemin. The serum follicle stimulating hormone (FSH) surge was unaffected by HO inhibition or
22 hemin but in vitro treatment of GnRH-stimulated pituitaries with hemin led to a significant increase
23 in FSH release. The estrogen-dependent afternoon prolactin surge was inhibited or delayed by HO
24 inhibition and significantly decreased prolactin release. In vitro studies using pituitary explants
25 showed that LH release was significantly increased by HO inhibition. HO inhibition also decreased
26 litter weight gain during lactation, which the authors attributed to decreased maternal milk
27 production or milk ejection problems as cross-fostered pups regained weight that was lost during
28 nursing on HO inhibited dams (Alexandreanu and Lawson, 2002, 192373). The lactational effects
29 seen in this model may be explained by changes in prolactin (Alexandreanu and Lawson, 2003,
30 193871). It is possible that HO inhibition by CrMP may also inhibit NO production, a mechanism
31 that is distinct from CO-dependent effects.
September 2009 5-95 DRAFT - DO NOT CITE OR QUOTE
-------
Ovarian Follicular Atresia
1 As a part of normal follicular maturation in the ovaries, the majority of follicles undergo
2 artresia via apoptosis prior to ovulation. Harada et al. (2004, 193920) harvested porcine granulosa
3 cells from ovaries and found that cells naturally undergoing atresia or cell death more strongly
4 expressed HO-1 than did successful follicles. Addition of the HO substrate hemin or the HO
5 inhibitor Zn protoporphyrin IX (ZnPP IX) significantly induced or inhibited granulosa cell apoptosis,
6 respectively. In this porcine model, HO was able to augment granulosa cell apoptosis allowing for
7 proper follicular maturation.
Summary of lexicological Studies on Birth Outcomes
8 There is some evidence that CO exposure leads to altered birth outcomes, including decreased
9 birth and near term body weight, increased pregnancy loss and perinatal death, and increased
10 malformations. These events occurred at levels as low as 65 ppm for fetal body weight decrements
11 and 90 ppm for changes in birth weight and perinatal death. Pregnancy loss was seen after exposure
12 to 250 ppm CO, whereas skeletal malformations were present after 180 ppm CO. Dietary protein and
13 zinc modifications exacerbated these CO induced effects on birth outcomes. Maternal protein
14 restriction and CO had a synergistic effect on peri- and postnatal mortality and an additive effect on
15 malformations. Dietary zinc alterations resulted in inconsistent changes to CO-induced
16 malformations and fetal mortality.
17 Endogenous CO is recognized as a second messenger signaling molecule with normal
18 physiological roles in maintaining pregnancy and for proper fetal and postnatal development. The
19 endogenous HO/CO system is also involved in controlling vascular tone, follicular maturation,
20 ovarian steroidogenesis, secretion of gonadotropin and prolactin by the anterior pituitary, lactation,
21 and estrous cyclicity in rodent studies. These areas could be potential points of interaction of
22 exogenous CO with endogenous HO/CO.
5.4.2.2. Developmental Effects
Congenital Abnormalities
23 Studies by Schwetz et al. (1979, 011855) found that gestational CO exposure (250 ppm) in
24 CF-1 mice for 7 or 24 h/day over GD6-GD15 resulted in minor fetal skeletal alterations in the form
25 of extra lumbar ribs and spurs (dam gestational COHb 10-15% for 7h/day exposure, 24 h/day dam
26 COHb not measured). Similarly exposed rabbits did not exhibit these changes.
September 2009 5-96 DRAFT - DO NOT CITE OR QUOTE
-------
1 Astrup et al. (1972, 011121) studied the effect of CO exposure on fetal rabbit development via
2 continuous CO exposure (90 or 180 ppm with gestational dam COHb of 9 and 17%, respectively)
3 over the duration of gestation. Three pups in the 180 ppm CO group (n = 123) had deformities in
4 their extremities at birth, whereas no control and no 90 ppm CO-exposed animals manifested with
5 this malformation.
6 Further skeletal malformations were seen after gestational CO exposure in mice as described
7 above ("Effect of Maternal Diet") (Singh et al., 1993, 013892). Briefly, pregnant CD-I mice were
8 exposed intermittently to CO (65-250 ppm; GD8-18) in combination with protein modified diets
9 [27% (supplemental protein), 16% (control), 8% (low), or 4% (very low protein)] to assess the role
10 of dietary protein in modulating CO effects on neonates at 1 wk of age. Maternal dietary protein
11 restriction additively compounded the CO induced skeletal malformations. Further, dietary
12 restriction in Zn and protein led to increased teratogenicity, specifically increased incidence of
13 gastroschisis (Singh, 2003, 053624). Conversely, Carmines et al. (2008, 188440) did not find
14 evidence of external malformations (teratogenicity) in rats after exposure to -600 ppm CO from
15 GD6-GD19.
CMS Developmental Effects
Behavioral
16 Investigators have used animal models to study the effects of moderate CO exposure
17 (65-150 ppm) during gestation on behavioral outcomes after birth, including active avoidance,
18 learning and memory, homing, and motor activity. These studies generally found decrements in
19 behavior in early life after in utero exposure to CO concentrations greater than 125 ppm and in some
20 cases as low as 65 ppm. Table 5-14 shows results of behavioral response studies with CO exposure
21 <150ppm.
September 2009 5-97 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-14 Behavioral responses to low and moderate CO exposure
Reference Model System CO Exposure Response
BEHAVIORAL
75 and 150 ppm
Deetal. (1995, p. mn«nn™ic Impaired acquisition (Sand 18 months) and reacquisition (18
0794411 Kals connnuous months) of avoidance behavior at 150 ppm , not 75 ppm
GDO-GD20
Notes
Mactutus and Fechter 150PPm
(1985, 0115361 Rats continuous Delayed acquisition of active avoidance (PND1 20) and disrupted CQHb156+11%
GDO-GD20
75 and 150 ppm CO (150 ppm) reduced the minimum frequency of ultrasonic calls as
Di et al. (1 993, p t ., well as decreased responsiveness to a challenge dose of diazepam.
0138221 Kals connnuous There was no change in locomotion however CO impaired learning
GDO-GD20 in a two-way active avoidance task.
Mactutus and Fechter
(1984 011355) p_te -icn nnm Acquisition did not improve with age/maturation, failure to learn; rnl,h ^0,
Kals lsuppm impaired reqacquisition (PND31), failure to retain UJHD ISA,
75 and 150 ppm
011538) Rats continuous Decreased exploration, habituation, non-spatial working memory
GDO-GD20
Mouse ZnPPIX (HO
Zhuo et al. (1993, : ""rLmn,i hr,in inhibitor) and HO inhibition blocked long term potentiation and CO evoked
013905) nippocampai cram synaptic potentials and long-term enhancement
' sectlons 0.1-1.0 pM CO f H H a
Stevens and V\feng Mouse and rat -, DDIV K ^K • •> HO inhibition blocked long term potentiation but not long-term
(1993. 1884581 hippocampal brain ZnPPIX (5-15 pM) depression a v a
Mereu (2000, 193838) pat hinnnrammi 150 ppm
reai nippocampai impaired long term potentiation maintenance
brain sections QDO-GD20
150 ppm
(ie980eoai295inaU Ra's continuous Delayed homing behavior and poor reflexive response
GDO-GD20
COHb: 1.6 ±0.1%
0 ppm); 7.36 ±0.2%
75 ppm); 16.1 ±0.9%
150 ppm)
150 ppm
("wTmoeSS)™ Rats continuous Decreased locomotor activity at PND1, 4, and 14, but not PND21 COHb 15%
1
2
3
4
5
6
7
GDO-GD20
65 and 125 ppm
cinnhMoaR f-Moa™ u:^ ™n«n,,™,,. Impaired aerial righting score at PND14 (65 and 125 ppm), impaired
Singh (1986, 012827) Mice continuous negative geotaxis at PND10 and righting reflex on PND1 (125 ppm)
GD7-GD18
Active AvoidanCG BGhavior. To assess behavioral changes after in utero exposure, pregnant
Wistar rats were exposed to CO (0, 75, or 150 ppm) continuously over GDO-GD20 (De Salvia et al.,
1995, 079441). Male pups from exposed dams were evaluated for active avoidance behavior (mild
shock avoidance) during acquisition and reacquisition. This work was designed to expand on the
studies of Mactutus and Fechter (1985, 011536). who showed delayed acquisition (120 days of age)
of an active avoidance task and disruption of retention at a later test date (360 days) after continuous
in utero CO exposure (150 ppm CO, dam COHb concentrations of 15.6 ± 1.1%), and to determine if
September 2009
5-98
DRAFT - DO NOT CITE OR QUOTE
-------
1 these behavioral changes were permanent. De Salvia et al. (1995, 079441) found there were no
2 significant behavioral impairments in moderate dose animals (75 ppm). However, animals exposed
3 to the 150 ppm in utero had significantly impaired acquisition (at 3 and 18 months of age) and
4 reacquisition (at 18 months of age) of conditioned avoidance behavior. This impaired learning was
5 also seen in gestationally CO (150 ppm, trend seen at 75 ppm) exposed rats at PND90 (Di Giovanni
6 et al., 1993, 013822). The authors speculated that this CO-dependent behavioral change may be
7 mediated through neurotransmitter signaling, specifically changes in dopamine in the neostriatum or
8 nucleus accumbens. These studies demonstrate that moderate CO exposure in utero can lead to
9 permanent behavioral changes in male offspring.
10 Mactutus and Fechter (1984, 011355) also found that acquisition in a two-way conditioned
11 avoidance test (flashing light warnings followed by mild footshock) failed to improve with age of in
12 utero CO-exposed (150 ppm, dam COHb 15%) Long-Evans rats (male and female offspring) in
13 contrast to air-exposed controls who improved with age/maturation, indicating a failure in the
14 associative process of learning. They also found impairments in reacquisition performance, an index
15 of retention, in PND31 rats that had received continuous in utero CO exposure. Overall, prenatal CO
16 exposure (150 ppm, not 75 ppm) induced learning and memory deficits in male and female
17 offspring.
18 HabituatJOn, Memory, and Learning. Giustino et al. (1999, 011538) exposed primiparious
19 pregnant Wistar rats to CO (0, 75 or 150 ppm) by inhalation from GDO-GD20. Blood COHb
20 concentrations (mean % ± SEM) on GD20 were reported (0 ppm: 1.6 ± 0.1; CO 75 ppm: 7.36 ± 0.2;
21 CO 150 ppm: 16.1 ± 0.9). Male offspring at age 40 days were given two habituation trials. In the
22 first trial (Tl), two similar objects were presented. In the second trial (T2), one object from the first
23 trial was presented as well as one novel object. Results were quantified three ways. Exploration
24 activity was defined as the time exploring both objects during each trial. Global habituation was
25 quantified as a comparison of the time spent exploring the two objects in Tl to the time spent
26 exploring objects in T2. Discrimination between new and familiar objects was measured in T2 by
27 contrasting the time spent exploring the familiar object to the time spent exploring the new object.
28 These recognition sessions test for the preference that rats have for investigating novel objects in lieu
29 of familiar objects and are a measurement of non-spatial working memory. The results of this study
30 showed 40 day old animals that were gestationally exposed to CO (both 75 and 150 ppm) spent less
31 time exploring novel objects when compared to control animals. Control rabbits habituated or
32 learned after a second exposure to a previously explored object (T2
-------
1 animals (75 or 150 ppm). The authors speculated that the mesolimbic dopaminergic system may be
2 responsible for these changes, possibly involving the nucleus accumbens. The human literature
3 shows a possible connection with these CO-dependent rodent effects; infants whose mothers smoked
4 during pregnancy manifest with habituation defects (Fried et al., 1998, 190210; Fried et al., 2003,
5 190209). Nonetheless, CO is just one of many constituents of cigarette smoke. The results from
6 these animal toxicology studies showed that in utero exposure to CO affects non-spatial working
7 memory in young adult male rats.
8 Studies have shown that endogenous and exogenous CO may be involved in the generation of
9 the hippocampal long-term potentiation (LTP), which is believed to correlate with learning and
10 memory (Hawkins et al., 1994, 076503: Mereu et al., 2000, 193838: Stevens and Wang, 1993,
11 188458: Zhuo et al., 1993, 013905). It is possible that CO can act as a retrograde synaptic signaling
12 messenger, allowing a signal to travel from a postsynaptic to presynaptic neuron. Treatment of
13 mouse or rat hippocampal brain sections with ZnPPIX, a HO inhibitor, blocked induction of the LTP,
14 but not long-term depression (Stevens and Wang, 1993, 188458: Zhuo et al., 1993, 013905).
15 Exogenous CO exposure (0.1-1.0 (iM) also evoked long-term enhancement and evoked synaptic
16 potentials (Zhuo et al., 1993, 013905). Similarly, hippocampal slices from gestationally CO exposed
17 (150 ppm from GDO-20) Wistar rats exhibited an impaired ability to maintain LTP over time and a
18 modest reduction in post-tetanic potentiation (Mereu et al., 2000, 193838). Conversely, other studies
19 have found no correlation between CO and LTP using step through, step down, and water maze tests
20 (Bing et al., 1995, 079418: Toyoda et al., 1996, 079945). Thus, distinct types of learning may be
21 differentially regulated by CO exposure; and endogenous CO, as modulated by HO inhibitors, may
22 manifest with different outcomes when compared to outcomes seen for exogenous CO.
23 Homing and LOCOmotor Effects. Fechter and Annau (1977, 010688: 1980, 011295)
24 exposed Long-Evans rats in utero to 150 ppm CO continuously through gestation (dam COHb 15%)
25 and saw significant effects of CO on pup locomotor activity measured across 10-minute intervals for
26 a 1-h period. CO-exposed pups showed consistently less activity than air-exposed controls through
27 the pre-weaning window, with significantly reduced activity seen at PND1 and PND4 (both after
28 subcutaneous L-DOPA administration to induce movement) and at PND14, but not at PND21.
29 However, the PND14 rats only showed decreased activity after 30 min of testing. Di Giovanni et al.
30 (1993, 013822) found that prenatal CO (75 and 150 ppm) did not significantly affect locomotor
31 activity or D-amphetamine induced hyperactivity at PND14 or PND21, but the rats were only
32 subjected to a 30-min session. This study may have overlooked the later window of decreased
33 activity.
34 Under analogous exposure conditions, Fechter and Annau (1980, 011295) found that the
35 development of homing behavior, orientation by the rat toward its home cage, was significantly
36 delayed in rats prenatally exposed to 150 ppm. Also, exposed offspring manifested with poorer than
September 2009 5-100 DRAFT - DO NOT CITE OR QUOTE
-------
1 delayed in rats prenatally exposed to 150 ppm. Also, exposed offspring manifested with poorer than
2 normal performance on the negative geotaxis test, a reflexive response that results in a directional
3 movement with or against gravity. Similarly, continuous prenatal CO exposure (125 ppm, GD7-
4 GDIS) in CD-I mice impaired negative geotaxis at PND10 (Singh, 1986, 012827). The
5 standardization and use of geotaxis as a vestibular, motor, or postural metric in infant rodents has
6 been debated in the literature (Kreider and Blumberg, 2005, 193944).
7 Prenatal exposure to CO (125 ppm, GD7-GD18) significantly affected the righting reflex (the
8 turning of an animal from its supine position to its feet) in exposed CD-I mice on PND1. Also, the
9 aerial righting score, or turning 180° and landing on the feet when dropped from the supine position
10 at a height, was significantly decreased in pups exposed to CO in utero (65 and 125 ppm) at PND14
11 (Singh, 1986, 012827). The same trend of impaired righting reflex was seen in gestationally CO
12 (150 ppm) exposed rats (Fechter and Annau, 1980, 011295). These behavioral tests indicated
13 neuromuscular, vestibular, or postural effects in the CO-exposed neonate.
14 Conversely, no gross impairment of motor activity measured by infrared movement
15 monitoring in Wistar rats treated in utero (GDO-GD20) to moderate levels of CO (0, 75 or 100 ppm)
16 was found (Carratu et al, 2000, 015839). Monitoring was done at PND40 and PND90 and may have
17 been too late to detect CO-dependent changes. Earlier studies by Fechter and Annau (Fechter and
18 Annau, 1977, 010688) identified an early window of sensitivity for CO-dependent motor activity
19 deficits of PND1-PND14, with recovery by PND21.
20 Emotionality. In utero CO exposure caused subtle alterations in the ontogeny of emotionality
21 measured by the ultrasonic vocalization emitted by rat pups removed from their nest. Prenatal CO
22 exposure (150 ppm) caused a reduction in the minimum frequency of ultrasonic calls emitted by
23 PND5 pups (Di Giovanni et al., 1993, 013822). The rate of calling, maximum frequency, and
24 duration and sound pressure level were not affected by prenatal CO. However, the rate of calling and
25 responsiveness to a challenge dose of diazepam was decreased by prenatal CO exposure (150 ppm).
26 Pup vocalization is mediated by the GABAergic neuron function which is altered by CO exposure
27 (see below).
Neuronal
28 Since behavioral changes have been caused by CO exposure, studies have investigated
29 whether CO exposure results in changes to neuronal structures and electrical excitability. Moderate
30 levels of CO (75 -150 ppm) decrease peripheral nervous system (PNS) myelination due to impaired
31 sphingomyelin homeostasis and can reversibly delay the rate of ion channel development after
32 gestational exposure. In utero CO exposure also results in irreversible changes in sodium equilibrium
33 potential. Further details of these studies are given below in Table 5-15.
September 2009 5-101 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-15 Neuronal responses to low and moderate CO exposure
Reference
SMysfem
C0
Response
Notes
NEURONAL
Carratu et al. (2000,
0158391
Rats
75 and 150ppm
continuous
GDO-GD20
Decreased peripheral nerve fiber myelin sheath thickness
COHb:0 ppm (GD10: 0.97 +
0.02; GD20: 1.62 ±0.1),
75 ppm (GD10: 7.20 ±0.12;
GD20: 7.43 ±0.62), and
150 ppm (GD10: 14.42 ±0.52;
GD20: 16.08 ±0.88)
Carratu et al. (2000,
0159351
Rats
150ppm
continuous
GDO-GD20
Impaired sphingomyelin homeostasis by increasing sphingosine
Carratu etal. (1993,
0138121
Rats
75 and 150ppm
continuous
GDO-GD20
Produced partly reversible changes in membrane excitability through
delayed inward current inactivation and decreased inward current COHb: 15% at 150 ppm
reversal potential
DeLuca 1996 (1996,
0809111
Rats
75 and 150ppm
continuous
GDO-GD20
Delayed development of the ion channels responsible for passive and
active membrane electrical properties of skeletal muscle
Montagnani et al. (1996,
0809021
Rats
75 or 150 ppm
GDO-GD20
CO (150 ppm) increased the tetrodotoxin-inhibition of PNS-evoked
vasoconstriction at PND5-7. CO exposure caused the relaxant effect by
ACh to appear earlier and the contractile response to disappear earlier
(vasodilator effects).
Dyer etal. (1979,
1909941
Rats
150 ppm
GDO-GD21
Increased early components (P1-N1 and N1-P1) of the cortical flash .. t , rnuh *r0,
evoked potential peak-to-peak amplitudes at PND65 in female rats Maternal LUMtx lb/0
1 Peripheral Nerve MyeNnation. In utero exposure (GDO-GD20) to moderate levels of CO (0,
2 75 or 150 ppm) and its effect on sciatic nerve myelination in male offspring was studied in Wistar
3 rats (Carratu et al., 2000, 015839). The dam CO blood concentration expressed as %COHb was
4 determined for 0 ppm (GD10: 0.97 ± 0.02; GD20: 1.62 ± 0.1.), 75 ppm (GD10: 7.20 ± 0.12; GD20:
5 7.43 ± 0.62), and 150 ppm (GD10: 14.42 ± 0.52; GD20: 16.08 ± 0.88). The myelin sheath thickness
6 of the peripheral nerve fibers was significantly decreased in CO-exposed animals (75 and 150 ppm),
7 however axon diameter was not affected. As mentioned above, even though CO affected
8 myelination, it did not significantly affect motor activity of CO-exposed mice at 40 and 90 days. It is
9 possible that these deficits in PNS myelination are due to impaired sphingomyelin homeostasis. In
10 utero exposure (GDO-GD20) of Wistar rats to CO (150 ppm) caused a 2-fold increase in sphingosine
11 (SO), but not sphinganine (SA) in the sciatic nerve at 90 days of age (Carratu et al., 2000, 015935).
12 SO is an intermediate in sphingolipid turnover and SA is an intermediate of de novo sphingolipid
13 biosynthesis. Hypoxia has been shown to induce sphingomyelin changes which could lead to
14 impaired myelination and motor activity decrements (Ueda et al., 1998, 195136; Yoshimura et al.,
15 1999, 195135). Prenatal CO exposure had no effect on brain SA or SO levels in male offspring at
16 90 days of age. These results demonstrate prenatal CO exposure could interrupt sphingolipid
September 2009
5-102
DRAFT - DO NOT CITE OR QUOTE
-------
1 homeostasis in the PNS but not CNS, causing a decrease in nerve myelination without changes in
2 motor activity.
Electrophysiological Changes.
3 Gestational exposure of Wistar rats to continuous CO (75 or 150 ppm (15% COHb at
4 150 ppm) yielded electrophysiological changes in the PNS (Carratu et al., 1993, 013812). Changes
5 were noticeable in voltage- and time-dependent properties of sodium channels in the sciatic nerve
6 after in utero CO exposure. Sodium channel inactivation kinetics were reversible (present at PND40
7 and absent at PND270), but changes in the sodium equilibrium potential were irreversible. In utero
8 CO exposure (150 ppm) also delayed the development of the resting chloride conductance (GC1) and
9 resting potassium conductance (GK), with levels matching the control by PND80 and PND60,
10 respectively (De Luca et al., 1996, 080911). CO exposure (75 and 150 ppm) also altered the
11 pharmacological properties of the chloride channel and excitability parameters of skeletal muscle
12 fibers. These changes in the nerve electrophysiological properties could account for increased
13 tetrodotoxin-inhibition of the vasoconstriction evoked by the PNS in 5-7 day old prenatally exposed
14 pups (Montagnani et al., 1996, 080902). Finally, gestational CO exposure increased early
15 components (P1-N1 and N1-P1) of the cortical flash evoked potential peak-to-peak amplitudes at 65
16 days post exposure (PND65) in female, not male, rats (Dyer et al., 1979, 190994). The early waves
17 of the cortical evoked potential, an indicator of visual cortical functioning, generally indicate activity
18 in the retinogeniculostriate system. These studies showed that in utero CO exposure had both
19 reversible and irreversible effects on sodium and potassium channels, which are essential for proper
20 electrophysiological function of the muscles and PNS.
Neurotransmitter Changes
21 The developing nervous system is extremely sensitive to decreased oxygen availability.
22 Virtually all neurotransmitter systems are present at birth but require further maturation. The studies
23 listed below in Table 5-16 have shown that prenatal exposure to CO alters a number of
24 neurotransmitters and their pathways at levels from 75-300 ppm, both transiently and permanently.
September 2009 5-103 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-16 Neurotransmitter changes from low and moderate CO exposure
-^™ sMy£e:
CO Exposure
Response
Notes
NEUROTRANSMITTER CHANGES
StaL(20°°< Guinea pigs
5StaL(2000' Gui™
|||fetaL<1998< Guinea pigs
Cagianoetal. (1998, D t
0871701 Rats
Hermans etal. (1993, p t
1905101 rals
Fechter etal. (1987, p.
0122591 Kals
Storm and Fechter D .
(1985, 0116531 Kals
Storm and Fechter p t
(1985, 0116521 rals
Storm etal. (1986, p.
0121361 Kals
Benagiano et al. (2005, p .
1804451 Kals
Benagiano (2007, p t
1938921 Kats
Antonelli (2006, 1938851 Rats
200 ppm
1 0h/day
GD23-25 to
GD61-63
200 ppm
1 0h/day
GD23-25 to birth
Hyperthermia on
PND4
200 ppm
1 0h/day
GD23-25 to birth
75 and 150 ppm
GDO-GD20
Hypoxia(10.5%
02)
GD15-GD21
75, 150, and
300 ppm
GDO-GD20 or
PND10
150 ppm
GDO-GD20
75, 150, and
300 ppm
GDO-GD20
75, 150, and
300 ppm
GDO-PND10
75 ppm
GDO-GD20
75 ppm
GD5-GD20
75 ppm
GD5-GD20
CO affected catecholaminergic system in brainstem by reducing tyrosine
hydroxylase. Affected cholinergic system by increasing choline acetyl-
transferase.
CO sensitizes the brain to the effects of a short period of hyperthermia on
PND4. The exposure combination resulted in lesions in the brain, as well
as increased serotonin and glial fibrillary acidic protein. The exposure also
caused reactive astrogliosis.
CO increased tidal volume during steady state hypercapnia and
progressive asphyxia, due to increased ventilation.
In utero CO (150 ppm) exposure increased mount/intromission latency ,
decreased mount/intromission frequency, and induced ejaculatory
abnormalities. CO also blunted the amphetamine-induced increase in
dopamine.
Hypoxia caused delayed initiation latencies of male sexual behavior and
decreased number of ejaculations.
Prenatal CO exposure continuing to PND10 leads to increased
concentrations of dopamine but not dopamine metabolites in striatal tissue.
Prenatal CO exposure increased mean and total cerebellar norepinephrine
concentration from PND14 to PND42, but not in the cortex.
CO transiently decreased 5HT and NE in the pons/medulla. CO increased
NE in the cortex and hippocampus at PND42. CO dose-dependently
reduced cerebellum wet weight.
CO decreased cerebellar weight (150-300 ppm at PND10, 75-300 ppm at
PND21) and decreased total cerebellar GABA (150-300 ppm at PND10
and PND21). CO (300 ppm) exposed cerebella has fewer fissures.
CO reduced the number of GABA and GAD 65/67 positive neuronal bodies
and axon terminals in the cerebellar cortex.
Adult offspring exposed prenatally to CO exhibited decreased GABA and
GAD in the molecular layer and Purkinje neuron layers of the cerebellar
cortex
CO decreased cortical glutamatergic transmission both at rest and after a
chemical depolarizing stimulus.
Fetal COHb: 13%
Maternal COHb: 8.5%
Fetal COHb: 13%
Maternal COHb: 8.5%
Fetal COHb: 13%
Maternal COHb: 8.5%
Maternal COHb: GD10-1,
7, and15%;GD20-1.5, 7,
and16%(0, 75, and
150 ppm CO, respectively)
Maternal COHb: 2.5+0.1%,
11.4+0.3%, 18.5+0.5%,
26.8+1.1% (0,75, 150,
and 300 ppm, respectively)
Maternal COHb: 2.5%,
11.5%, 18.5%, and 26.8%
(0, 75, 150, and 300 ppm,
respectively)
Maternal COHb: 2.5%,
11.5%, 18.5%, and 26.8%
(0, 75, 150, and 300 ppm,
respectively)
1 Medullar NeurOtransmitterS. Maternal smoking during pregnancy is associated with
2 Sudden Infant Death Syndrome (SIDS) which involves the aberrant development of brainstem nuclei
3 controlling respiratory, cardiovascular, and arousal activity. To investigate changes in the structure
4 and neurochemistry of the brainstem, Tolcos et al. (2000, 015997) exposed pregnant guinea pigs to
September 2009
5-104
DRAFT - DO NOT CITE OR QUOTE
-------
1 CO (200 ppm) over the last 60% of gestation. Guinea pigs and humans both have the majority of
2 CNS development in utero. CO-exposed pups were found to have significant decrements in body,
3 brain, and liver weights, crown to rump length, and medullar volume. Neurotransmitter systems were
4 also affected after CO exposure. Specifically, the brainstem displayed significant decreases in protein
5 and immunoreactivity for tyrosine hydroxylase (TH), an enzyme necessary for catecholamine
6 production, which is likely due to decreased cell number in specific medullar regions responsible for
7 cardiorespiratory control. This was consistent with earlier work showing that prenatal CO exposure
8 leads to aberrant respiratory responses to asphyxia and CO2 (McGregor et al., 1998, 085342). The
9 cholinergic system was also affected by prenatal CO exposure with significant increases in choline
10 acetyl-transferase (ChAT) immunoreactivity of the medulla, however no changes in muscarinic
11 acetylcholine receptor. This is in contrast to human infants with SIDS who show decreased
12 brainstem muscarinic receptor binding (Kinney et al., 1995, 193898). ChAT changes in this study
13 (Tolcos et al., 2000, 015997) were from areas of the medulla associated with tongue innervation,
14 which is crucial to swallowing, possibly in relation to breathing.
15 A second risk factor for SIDS is hyperthermia. To explore the interaction of hyperthemia and
16 CO-induced hypoxia, pregnant guinea pigs were exposed to CO (0 or 200 ppm) for 10 h/day for the
17 last 60% of gestation (Tolcos et al., 2000, 010468). At PND4 male pups were exposed to
18 hyperthermia or ambient temperature as a control. Brains were then collected at 1 and 8 wk of age.
19 In utero CO exposure sensitized some areas of the brain to future hyperthermic insults. Specifically,
20 CO plus hyperthermia induced significant increases in serotonin in multiple brain regions (NTS,
21 DMV, and hypoglossal nucleus) at 1 wk of age; this change was no longer evident at 8 wk of age.
22 Hyperthermia exposure alone induced decreased met-enkephalin neurotransmitter immunoreactivity
23 at 1 wk of age that was absent at 8 wk and absent in CO plus hyperthermia exposed animals. Brain
24 stem neurotransmitter (met-enkephalin, serotonin, TH, substance P) immunohistochemical
25 differences were not apparent with CO treatment alone. At 8 wk of age, CO plus hyperthermia
26 exposure induced glial aggregations and gliosis surrounding infarct or necrotic areas in the brain and
27 the medulla lesions stained positive for glial fibrillary acidic protein (GFAP). GFAP upregulation is
28 classically seen with neuronal diseases or following neurodegeneration. Gross structural
29 observations revealed no differences in the medulla or cerebellum following in utero CO exposure
30 alone. Together, these data showed that CO exposure in utero sensitizes the brain to future
31 hyperthermic insults leading to generation of necrotic lesions in the brain and changes in
32 neurotransmitter levels.
33 DopaiTlinGrgiC EffGCtS. Dopamine is a catecholamine neurotransmitter that plays an
34 important role in the regulation of male rat sexual behavior. Experiments assessing sexual behavior
35 and mesolimbic dopaminergic function were conducted on adult (5 and 10 months of age) male
36 offspring gestationally exposed to CO (0, 75 or 150 ppm) (Cagiano et al., 1998, 087170). Maternal
September 2009 5-105 DRAFT - DO NOT CITE OR QUOTE
-------
1 COHb at GD10 was 1, 7, and 15% and 1.5, 7, and 16% at GD20 (0, 75, and 150 ppm CO,
2 respectively). At 5 months of age, CO-exposed male offspring showed decrements in sexual
3 behavior including an increase in mount to intromission latency, a decrease in mount to intromission
4 frequency, and a decrease in ejaculation frequency. Further, administration of amphetamine, which
5 stimulates copulatory activity, did not alter CO-induced changes in mount to intromission latency or
6 frequency. Basal extracellular dopamine concentration in the nucleus accumbens was unchanged
7 after CO exposure. However, when stimulated with amphetamine administration, control rats had
8 increased release of dopamine that was absent with CO-exposed rats. Rats followed to ten months of
9 age showed no significant changes in copulatory activity or neurochemical parameters after CO
10 exposure, indicating recovery from earlier decrements. This altered male sexual behavior in
11 CO-exposed offspring paralleled earlier studies of mice exposed gestationally to hypoxia (Hermans
12 et al., 1993, 190510). In summary, in utero exposure to CO delayed copulatory sexual behavior in
13 male offspring with accompanying changes in the mesolimbic dopaminergic system.
14 A second study also found no change in dopamine metabolite levels after prenatal exposure to
15 CO, however it did find an elevation in dopamine concentration in rats exposed both pre- and
16 postnatally to CO. Exposure of Long Evans rat dams and pups continuously to CO (75, 150, or
17 300 ppm with maternal COHb of 11, 19, and 27%, respectively) from conception to PND10 induced
18 significant elevations in dopamine in the striatum at PND21 in CO-exposed offspring versus air
19 exposed controls (Fechter et al., 1987, 012259).
20 NoradrenergiC and SerotonergiC Changes. Other monoamine neurotransmitters,
21 norepinephrine (NE) and serotonin (5HT), were tested for sensitivity to CO during development.
22 Long Evans rats exposed to CO (75, 150, or 300 ppm) over the duration of gestation yielded a dose-
23 dependent reduction in cerebellum wet weight (significant at 150 and 300 ppm) at PND21 with
24 increases in NE concentration found in the cortex and hippocampus at PND42 but not PND21
25 (Storm and Fechter, 1985, 011652). In a separate experiment, CO-exposed (150 ppm) animals
26 presented with increased mean and total NE concentrations in the cerebellum, but not cortex when
27 monitored from PND14 to PND42 (Storm and Fechter, 1985, 011653). Also, NE concentration in the
28 pons/medulla decreased linearly with increasing CO exposure at PND21 but not at PND42. A
29 transitory decrease in 5HT concentration was also shown in the pons/medulla after gestational CO
30 exposure (Storm and Fechter, 1985, 011652). Thus, in these studies, it appeared that CO both
31 transiently and permanently altered the pattern of postnatal neurotransmitter development in a
32 region-specific manner and postnatal growth of the cerebellum.
33 GlutamatergJC System. Glutamate is an abundant excitatory neurotransmitter that serves as
34 a precursor for the synthesis of the inhibitory neurotransmitter y-aminobutyric acid (GAB A)
35 catalyzed by glutamic acid decarboxylase (GAD). Primary cell cultures obtained from the cerebral
36 cortex of offspring (PND1) gestationally (GD5-GD20) exposed to CO (75 ppm) had decreased
September 2009 5-106 DRAFT - DO NOT CITE OR QUOTE
-------
1 extracellular glutamate (basal and K+-evoked) levels versus air exposed controls (Antonelli et al,
2 2006, 193885). Similarly, CO-exposed (300 ppm only) pups at PND21 had significant decreases in
3 cerebellar GABA content, decreased uptake of exogenous radio-labeled GAB A, decreased fissures in
4 the cerebellum, and decreased cerebellum size (Storm et al., 1986, 012136). It is possible this
5 decrease in GABA content is due to a diminished activity of GAD. Rats exposed to CO (75 ppm) in
6 utero (GDO-20) exhibited decreased GABA and GAD in the molecular layer and Purkinje neuron
7 layer of the vermian cerebellar cortex (Benagiano et al., 2005, 180445; Benagiano et al., 2007,
8 193892). This alteration may functionally impair cortical glutamatergic transmission in CO-exposed
9 offspring, possibly affecting learning and memory.
The Developing Auditory System
10 Prenatal exposure to tobacco smoke can cause auditory system deficits as seen in animal tests
11 for auditory responsiveness, habituation, and auditory arousal. Similarly, term human infants born to
12 smoking mothers have impaired cochlear development, albeit mild, with decreased amplitudes of
13 transient evoked otoacoustic emissions (OAE) at the highest test frequency (4 kHz) versus newborns
14 born to non-smokers (Korres et al., 2007, 190908); CO is one of many potential affective
15 components of cigarette smoke. The developing auditory system of rodents has recently been
16 investigated as a target of CO exposure at levels as low as 12 ppm. The rat brain and auditory system
17 goes through extensive cell division and multicellular organization during a major growth spurt in
18 the postnatal period (PND7-PND20), making it a probable target for CO induced effects. These
19 studies showed exposure to low concentrations of CO during development can lead to permanent
20 changes in the auditory system that persist into adulthood.
September 2009 5-107 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-17 Developing auditory system responses to low and moderate CO exposure
Reference
Model
System
CO Exposure
Response Notes
DEVELOPING AUDITORY SYSTEM
Stockard-Sullivan et al.
(2003, 1909471
Lopez (2003, 1939011
Webber etal. (2003,
1905151
Webber etal. (2005,
1905141
Lopez et al. (2008,
0973431
Rats
Rats
Rats
Rats
Rats
12-100 ppm
22 h/day
PND6-PND21-23
12 and 25 ppm
PND8-PND22
12.5, 25, 50 ppm
PND8-PND20-22
25 and 100 ppm
PND9-PND24
25 ppm
10-18 h/day
GD5-20orGD5-
GD20 and PND5-
PND20
CO (50 ppm) reduced otoacoustic emissions (preneural cochlear function)
at 7.13 and 8.01 kHz. CO persistently attenuated the amplitude of the mwh m w nnn i
action potential of the eighth cranial nerve (12-50 ppm), persisting to ^ VJ? ,.™. z, °A! ,M m)l
PND73. No functional impairment in the Morris Water Maze after CO a3/0 (AK)' ^ ' /0 I|VIK)
exposure.
CO (25 ppm) led to swelling and mild vacuolization of nerve terminals
innervating inner hair cells and the fibers of the 8th cranial nerve. CO
(25 ppm) decreased expression of neurofilament and myelin basic
proteins, cytochrome oxidase, NADH-TR, and calcium ATPase.
CO decreased c-Fos immunoreactivity in the central inferior colliculus at
both PND27 and PND75-PND77 over all dose groups (12.5, 25, or 50 ppm
CO)
CO exposure (25 and 100 ppm) decreased neurofilament proteins,
decreased c-Fos expression in the central 1C, and increased CuZnSOD in
the spiral ganglion neurons. Iron deficiency ablated these responses.
Prenatal CO exposure led to increased oxidative stress in the cochlear
vasculature (high HO-1, SOD-1, iNOS, and nitrotyrosine) and decreased
neurofilament proteins and synapsin-1 . CO caused morphological
deterioration of putative afferent terminals and mild deterioration in the
inner hair cells at the basal region of the cochlea.
1 Studies on the developing auditory system have used an artificial feeding system where pups
2 were removed from their respective dams and fed a milk substitute comparable to natural rat milk
3 via intragastric cannulation. This allowed nursing pups to be exposed to CO without possible
4 confounding by lactational and maternal CO co-exposure. However, this invasive rat model does
5 cause decreased brain, cerebellum, and lung weight at PND16. A summary of these studies and
6 others are presented in the above table (Table 5-17).
7 Using this model, Stockard-Sullivan et al. (2003, 190947) examined Sprague-Dawley rat pups
8 receiving low dose CO (12, 25, or 50 ppm) to determine how perinatal CO exposure (PND6-PND21-
9 23) functionally affected hearing in the developing rat. Rodent pups were either maternally reared
10 (MR), nutritionally supported with the artificial feeding system (AR), or received AR plus CO
11 exposure (ARCO). CO (50 ppm, not 25 ppm) exposure caused significant reductions in distortion
12 product otoacoustic emissions (DPOAE) levels at certain frequencies (7.13 and 8.01 kHz), a measure
13 of preneural cochlear function and thus not affected by eighth cranial nerve function. However, the
14 frequency range where significant CO results were seen is very narrow and low compared to the
15 normal rat audiogram. The eighth cranial nerve or vestibulocochlear nerve is responsible for
16 transmitting sound from the inner ear to the brain. This study also found significant attenuation of
17 the action potential (AP) of the eighth cranial nerve with ARCO exposure (12, 25, and 50 ppm CO)
18 versus AR controls at PND22. This is complicated by the finding that AR control animals had
19 significant attenuation of the eighth cranial nerve AP versus MR control animals, implying that
September 2009
5-108
DRAFT - DO NOT CITE OR QUOTE
-------
1 artificial rearing contributes to AP changes before CO was introduced. Nonetheless, the
2 CO-dependent attenuation of the eighth cranial nerve AP (versus AR control) was permanent,
3 persisting until adulthood in the 50 ppm CO exposure group (the only CO group monitored).
4 Auditory brainstem response (ABR) conduction time was not affected in CO-exposed animals (12,
5 25, 50, 100 ppm).These functional tests reported that neonatal exposure to low concentrations of CO
6 can induce auditory functional changes in rodents.
7 Further studies have investigated physiological changes in cochlear development during mild
8 chronic CO exposure. Sprague Dawley rats exposed to low levels of CO (12 or 25 ppm, ARCO)
9 from PND6-PND27 had no evidence of damage to the inner or outer hair cells (Lopez et al., 2003,
10 193901). However, CO (25 ppm) caused atrophy or vacuolization of the nerve cells that innervate
11 the inner (not outer) hair cells. Also, fibers of the eighth cranial nerve at the level of the internal
12 auditory canal had distorted myelination and vacuolization of the axoplasm after 25 ppm CO
13 exposure. Energy production markers in the organ of corti and spiral ganglion neurons including
14 cytochrome oxidase (electron transport chain complex IV) and NADH-TR (marker of complex I
15 reductase activity) were significantly decreased after 25 ppm (not 12 ppm) CO exposure versus
16 control (AR and MR). Reduced energy production likely led to the decreased expression of the
17 calcium-mediated myosin ATPase and neurofilament proteins in the organ of corti and spiral
18 ganglion neurons (25 ppm CO). Since no changes in body weight were found after CO exposure in
19 these experiments (Stockard-Sullivan et al., 2003, 190947). it is likely that the decreased electron
20 transport chain enzymes are specific to vulnerable areas such as the cochlea.
21 Further analysis focused attention on CO-induced changes in the inferior colliculus (1C), an
22 auditory integrative section of the midbrain. Low concentrations of CO (12.5, 25, or 50 ppm) inhaled
23 over PND8-PND22 decreased c-Fos immunoreactivity in the central 1C at both PND27 and
24 PND75-PND77; immnunostaining of other subregions of the 1C were not affected by CO (Webber et
25 al., 2003, 190515). c-Fos is an immediate early gene whose tonotopic expression corresponds to
26 neuronal activation in the auditory system. The same decrease in c-Fos expression was seen in rats
27 exposed to 25 or 100 ppm CO from PND9-PND24 (Webber et al., 2005, 190514). These CO-
28 exposed rats also exhibited decreased neurofilament proteins and increased Cu-Zn superoxide
29 dismutase (SOD1) in the spiral ganglion neurons. This response could be ablated by dietary iron
30 restriction, suggesting an ROS-dependent contribution to the auditory changes seen after CO
31 exposure. These authors postulated that CO creates a persistent oxidative stress condition where
32 ROS generated via the interaction of peroxide and iron (via the Fenton reaction or Haber Weiss
33 chemistry) leads to impaired cochlear development; decreasing the available iron decreases the total
34 pool available for ROS generation. Further, the attenuation of the elevated SOD levels with iron
35 restriction post CO-exposure gives credence to this model.
September 2009 5-109 DRAFT - DO NOT CITE OR QUOTE
-------
1 A recent study has found comparable auditory system responses after prenatal (GD5-GD20)
2 exposure to CO with postnatal exposure (GD5-PND20) similar to the studies described above
3 (Lopez et al, 2008, 097343). Prenatal CO (25 ppm) exposure led to high levels of the oxidative
4 stress markers HO-1, SOD-1, iNOS, and nitrotyrosine in cochlea vasculature and stria vascularis at
5 PND12, however unlike postnatally exposed pups, HO-1 and SOD1 levels returned to normal at
6 PND20. Both groups of CO exposed rats exhibited spiral ganglion cytoplasmic vacuolization, a
7 decrease in type I spiral ganglion neuron neurofilament proteins, thinning and damage in the cells of
8 the stria vascularis, and mild deterioration of the innervation of the inner hair cells. These nerve
9 terminals also had a persistent decrease in synapsin-1, a regulatory neuronal phosphoprotein. These
10 studies suggest that mild chronic CO exposure disrupts the developing auditory system, more often
11 at the IHC innervation and the eighth cranial nerve of the spiral ganglion, by creating an oxidative
12 stress that may be reflected as hearing impairment.
Summary of Toxicological Studies on Developmental Central Nervous System Effects
13 Toxicological studies employing rodent models have shown that low level CO exposure
14 during the in utero period can adversely affect adult outcomes including behavior, neuronal
15 myelination, neurotransmitter levels or function, and the auditory system. In utero CO exposure has
16 been shown to impair active avoidance behavior (150 ppm), habituation (75 and 150 ppm), non-
17 spatial memory (75 and 150 ppm), and emotionality (150 ppm). These behavioral changes could be
18 due to neuronal changes or altered neurotransmitter signaling. In utero CO exposure (75 and
19 150 ppm) was associated with PNS myelination decrements from impaired sphingolipid homeostasis
20 (150 ppm CO). These neuronal changes were also accompanied by electrophysiological changes
21 such as reversible delays in ion channel development and irreversible changes in sodium equilibrium
22 potential (150 ppm). Also, multiple studies demonstrated that in utero CO exposure affected
23 cholinergic (200 ppm), catecholaminergic (200 ppm), noradrenergic (150 ppm), serotonergic
24 (75 ppm), dopaminergic (75 ppm) and glutamatergic (75 ppm), neurotransmitter levels or
25 transmission in exposed rodents. Possible or demonstrated adverse outcomes from the CO-mediated
26 aberrant neurotransmitter levels or transmission include respiratory dysfunction (150 ppm), impaired
27 sexual behavior (150 ppm), and an adverse response to hyperthermic insults resulting in neuronal
28 damage (200 ppm). Finally, in utero CO exposure has been shown to affect the developing auditory
29 system of rodents, inducing permanent changes into adulthood at concentrations as low as 12 ppm.
30 Together, these animal studies demonstrate that in utero or perinatal exposure to CO can adversely
31 affect adult behavior, neuronal function, neurotransmission, and the auditory system in rodents.
September 2009 5-110 DRAFT - DO NOT CITE OR QUOTE
-------
Cardiovascular and Systemic Developmental Effects
1 In utero exposure to moderate to high concentrations of CO (60, 125, 150, 250, or 500 ppm) is
2 able to induce transient changes in cardiac morphology, cardiac action potentials, and systemic
3 immunity that may make a CO-exposed animal more susceptible to other outside stressors during the
4 immediate neonatal period. Studies of cardiovascular and systemic developmental responses to CO
5 levels of 500 ppm and less are presented below in Table 5-18.
Table 5-18 Cardiovascular and systemic developmental responses to low and moderate CO
exposure
Reference
Model
System
CO Exposure
Response
Notes
CARDIOVASCULAR AND SYSTEMIC DEVELOPMENT
Sartiani et al. (2004,
1908981
Prigge and Hochrainer
(1977, 0123261
Fechteretal. (1980,
0112941
Penney etal. (1982,
0113871
Styka and Penney
(1978, 0111661
Giustino et al. (1993,
0138331
Giustino et al. (1994,
0763431
Rats
Rats
Rats
Rats
Rats
Rats
Rats
150 ppm
GDO-GD20
60, 125, 250, and
500 ppm
GDO-GD21
150 ppm
GDO-GD20
500 ppm
PND1-PND32
400 or 500 ppm
increased to
1,100 ppm
Adult
6wk
75 and 150 ppm
GDO-GD20
75 and 150 ppm
GDO-GD20
CO delayed action potential duration shortening, decreased the density of
lto channels and increased the density of ICa,L channels.
CO depressed fetal hemoglobin (250 and 500 ppm), reduced fetal weight
(125, 250, and 500 ppm), decreased hematocrit (250 and 500 ppm), and
increased heart weight (60-500 ppm).
CO transiently increased wet heart weight. There was no increase in dry
heart weight.
CO increased heart weight to body weight ratio, which remained high to
PND107. Right ventricular weight was high through PND217.
Hydroxyproline and cardiac cytochrome c was depressed but only during
CO exposure. Neither lactate dehydrogenase nor myoglobin were altered
by CO.
CO caused increased heart weight to body weight that regressed within a
couple of months after CO exposure.
CO decreased splenic macrophage killing (75 and 150 ppm), phagocytosis
(150 ppm), and superoxide release (150 ppm). These alterations were
reversible, not seen at PND60.
CO (150 ppm) decreased the frequency of splenic leukocyte common
antigen (LCA+) cells at PND21, but not PND15 or PND540
COHb: 15%
COHb: 400 ppm-35%;
1,100ppm-58%
COHb: 150 ppm-15%
6
7
8
9
10
11
12
Myocardial Electrophysiological Maturation
A rat model of in utero exposure was employed to study CO effects on the development of
cardiac myocytes. Results demonstrated that in utero CO exposure (150 ppm) alters postnatal
cellular electrophysiological maturation in the rat heart (Sartiani et al., 2004, 190898). Specifically,
at 4 wk of age, the action potential duration (APD) of isolated cardiac myocytes from CO-exposed
animals failed to shorten or mature as the APD of control animals did. Further, the two ion
conduction channels Ito (transient outward current, K+-mediated) and ICa,L (L-type Ca2+ current),
which largely control the rat APD, were significantly different from control animals after in utero
September 2009
5-111
DRAFT - DO NOT CITE OR QUOTE
-------
1 CO exposure at 4 wk of age. These CO-dependent changes were resolved by 8 wk of age, reflecting
2 a delayed maturation. Further, these authors postulated that a CO-dependent delay in
3 electrophy siological maturation of the cardiac myocyte (lack of APD shortening) could lead to
4 arrhythmias and thus could be associated with SIDS deaths. However, no SIDS-like cardiac
5 aberrations were followed in intact Holter-monitored rats in this study.
Heart Morphological Changes After In Utero or Perinatal CO Exposure
6 Multiple authors have reported cardiomegaly following in utero low level CO exposure.
7 Prigge and Hochrainer (1977, 012326) reported increased fetal Wistar rat heart wet weight or
8 cardiomegaly following continuous in utero CO (60, 125, 250, and 500 ppm) exposure with no
9 decreases in near term fetal hematocrit or Hb levels seen at exposures below 250 ppm. Fechter et al.
10 (1980, 011294) found that prenatal exposure to CO affected cardiac development in exposed
11 offspring. Long Evans rats that were exposed to CO continuously (150 ppm) during gestation
12 manifested with significant elevations in wet heart weight, as well as heart weight in relation to body
13 weight at PND1, but not PND4, PND14, or PND21. Dry to wet weight ratios revealed that the
14 increased heart weight of CO-exposed pups at birth was due to edema or water content. Penney et al.
15 (1982, 011387) studied CO-dependent (500 ppm) cardiomegaly in neonates (continuous CO
16 exposure for 32 days starting at PND1). Other studies of adult male Charles River derived rats
17 exposed to CO for 6 wk (at 400 or 500 to 1,100 ppm CO) as adults only developed CO-dependent
18 cardiomegaly during exposure that significantly regressed within a couple of months after
19 termination of CO exposure (Styka and Penney, 1978, 011166).
Systemic Immune Toxicology After In Utero CO Exposure
20 In utero exposure (GDO-GD20) of male Wistar rats to moderate CO (0, 75, or 150 ppm)
21 concentrations induced reversible changes in macrophage function (Giustino et al., 1993, 013833).
22 The killing of Candida albicans (yeast) by splenic macrophages was significantly decreased at
23 PND15 in gestationally CO-exposed male offspring (75 and 150 ppm) but recovered function by
24 PND21. Macrophage phagocytosis of C. albicans was significantly reduced at PND15 and PND21 in
25 CO-exposed males (150 ppm only) and recovery was seen at PND60. Superoxide production by the
26 splenic macrophage respiratory burst was significantly decreased at PND15 and PND21 after in
27 utero CO exposure (150 ppm only) with recovery to control levels at PND60. In summary, CO
28 exposure in utero leads to a reversible and dose dependent loss of function of splenic macrophages
29 with decreased killing ability, decreased phagocytosis, and decreased ROS production during the
30 macrophage respiratory burst.
31 Further studies by the same laboratory showed that in utero exposure of male rats to CO
32 (150 ppm) induced a subtle decrease in the frequency of splenic immunocompotent cells (leukocyte
33 common antigen (LCA+) cells) in a population of splenic immune cells at PND21, but not PND15 or
September 2009 5-112 DRAFT - DO NOT CITE OR QUOTE
-------
1 PND540 (Giustino et al, 1994, 076343). Specific LCA+ cell subpopulations including macrophages,
2 Major Histocompatibility (MHC) II cells, T and B lymphocytes showed a decreasing trend but were
3 not significant with CO exposure.
Summary of Toxicological Studies of Cardiovascular and Systemic Development
4 In utero CO exposure is associated with various adverse, albeit non-persistent, cardiac
5 aberrations. Exposure to 150 ppm induced a delayed maturation of the cardiac action potential in
6 CO-exposed offspring. In other studies, continuous in utero CO exposure (60-500 ppm) induced
7 cardiomegaly at PND1 which was transient and regressed by PND4. CO (75 and 150 ppm) also
8 affects nonspecific immunity, shown through a reversible and dose dependent loss of function of
9 splenic macrophages with decreased killing ability, decreased phagocytosis, and decreased
10 macrophage ROS production (150 ppm). Also, the distribution of splenic immunocompotent cells
11 was slightly skewed because of a decrease in the number of LCA+ cells in PND21 male rats exposed
12 during gestation to 150 ppm CO. In conclusion, in utero exposure to moderate doses of CO
13 (60-500 ppm) is able to induce transient changes in cardiac morphology, cardiac action potentials,
14 and systemic nonspecific immunity.
5.4.3.Summary of Birth Outcomes and Developmental Effects
15 The most compelling evidence for a CO-induced effect on birth and developmental outcomes
16 is for PTB and cardiac birth defects. These outcomes were not addressed in the 2000 CO AQCD,
17 which included only two studies that examined the effect of ambient CO on LEW. Since then, a
18 number of studies have been conducted looking at varied outcomes, including PTB, birth defects,
19 fetal growth (including LEW), and infant mortality.
20 There is limited epidemiologic evidence that CO during early pregnancy (e.g., first month and
21 first trimester) is associated with an increased risk of PTB. The only U.S. studies to investigate the
22 PTB outcome were conducted in California, and these reported consistent positive associations with
23 CO exposure during early pregnancy when exposures were assigned from monitors within close
24 proximity of the mother's residential address. Additional studies conducted outside of the U.S.
25 provide supportive, though less consistent, evidence of an association between CO concentration and
26 PTB.
27 Very few epidemiologic studies have examined the effects of CO on birth defects. Two of
28 these studies found maternal exposure to CO to be associated with an increased risk of cardiac birth
29 defects. This insult to the heart is coherent with results of human clinical studies demonstrating the
30 heart as a target for CO effects (Section 5.2). Animal toxicological studies provide additional
31 evidence for such an insult to the heart, and reported transient cardiomegaly at birth after continuous
32 in utero CO exposure (60, 125, 250 and 500 ppm CO) and delayed myocardial electrophysiological
September 2009 5-113 DRAFT - DO NOT CITE OR QUOTE
-------
1 maturation (150 ppm CO). Toxicological studies have also shown that continuous in utero CO
2 exposure (250 ppm) induced teratogenicity in rodent offspring in a dose-dependent manner that was
3 further exacerbated by dietary protein (65 ppm CO) or zinc manipulation (500 ppm CO).
4 Toxicological studies of CO exposure over the duration of gestation have shown skeletal alterations
5 (7 h/day, CO 250 ppm) or limb deformities (24 h/day, CO 180 ppm) in prenatally exposed offspring.
6 There is evidence of ambient CO exposure during pregnancy having a negative effect on fetal
7 growth in epidemiologic studies. In general, the reviewed studies, summarized in Figure 5-9 through
8 Figure 5-11, reported small reductions in birth weight (ranging -5-20 g). Several studies examined
9 various combinations of birth weight, LEW, and SGA/IUGR and inconsistent results are reported
10 across these metrics. It should be noted that having a measurable, even if small, change in a
11 population is different than having an effect on a subset of susceptible births and increasing the risk
12 of IUGR/LBW/SGA. It is difficult to conclude if CO is related to a small change in birth weight in
13 all births across the population, or a marked effect in some subset of births. Toxicology studies have
14 found associations between CO exposure in laboratory animals and decrements in birth weight
15 (90-600 ppm), as well as reduced prenatal growth (65-500 ppm CO).
16 In general, there is limited epidemiologic evidence that CO is associated with an increased risk
17 of infant mortality during the neonatal or post-neonatal periods. In support of this limited evidence,
18 animal toxicological studies provide some evidence that exogenous CO exposure to pups in utero
19 significantly increased postnatal mortality (7 h/day and 24 h/day, 250 ppm CO; 24 h/day, 90 or
20 180 ppm CO) and prenatal mortality (7 h/day, 250 ppm CO).
21 Evidence exists for additional developmental outcomes which have been examined in
22 toxicological studies, but not epidemiologic or human clinical studies, including behavioral
23 abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes, and
24 changes in the auditory system. Structural aberrations of the cochlea involving neuronal activation
25 (12.5, 25 and 50 ppm CO) and auditory related nerves (25 ppm CO) were seen in pups after neonatal
26 CO exposure. Auditory functional testing using otoacoustic emissions testing (OAE at 50 ppm CO)
27 and 8th cranial nerve action potential (AP) amplitude measurements (12, 25, 50, 100 ppm CO) on
28 rodents exposed perinatally to CO showed that CO-exposed nenonates had auditory decrements at
29 PND22 (OAE and AP) and permanent changes in AP into adulthood (50 ppm CO). Furthermore,
30 exogenous CO may interact or disrupt the normal physiological roles that endogenous CO plays in
31 the body. There is evidence that CO plays a role in maintaining pregnancy, controlling vascular tone,
32 regulating hormone balance, and sustaining normal follicular maturation.
33 Overall, there is limited, though positive, epidemiologic evidence for a CO-induced effect on
34 PTB and birth defects, and weak evidence for a decrease in birth weight, other measures of fetal
35 growth, and infant mortality. Animal toxicological studies provide support and coherence for these
36 effects. Both hypoxic and non-hypoxic mechanisms have been proposed in the toxicological
September 2009 5-114 DRAFT - DO NOT CITE OR QUOTE
-------
1 literature (Section 5.1), though a clear understanding of the mechanisms underlying reproductive and
2 developmental effects is still lacking. Taking into consideration the positive evidence for some birth
3 and developmental outcomes from epidemiologic studies and the resulting coherence for these
4 associations in animal toxicological studies, the evidence IS Suggestive Of 3 Causal
5 relationship between long-term exposures to relevant CO concentrations and
6 developmental effects and birth outcomes
5.5. Respiratory Effects
5.5.1.Epidemiologic Studies with Short-Term Exposure
7 This section evaluates the key epidemiologic studies published since the 2000 CO AQCD
8 (U.S. EPA, 2000, 000907) that further examine the association between short-term exposure to CO
9 and respiratory morbidity. Although the number of studies that have specifically examined the CO-
10 respiratory health relationship has increased, it is still considerably less than that for the other criteria
11 air pollutants (e.g., PM and O3). The epidemiologic studies discussed below represent those studies:
12 conducted in locations with ambient CO concentrations similar to those in the U.S.; determined to
13 use a reasonable study design and analytical methods; and adequately adjusted for confounding
14 using accepted methods. If limitations in the design or analytical methods used in a study were
15 identified they were noted. It is recognized that each of the studies evaluated have a varying degree
16 of exposure measurement error due to: the number of monitors used within the study, the geographic
17 size of the study area, the spatial variability of CO, and differences in personal exposure distributions
18 in the population (see Section 3.6.8), all of which could influence the associations observed. As a
19 result, in some instances specific details of a study are mentioned to address any potential bias in the
20 reported CO associations. Finally, the issue of confounding by measured or unmeasured copollutants
21 was evaluated, if possible, for each study through the interpretation of multipollutant models. The
22 results from multipollutant models were used as an attempt to disentangle the effect of CO from
23 other pollutants while recognizing the high correlation between CO and other combustion-related
24 pollutants.
5.5.1.1. Pulmonary Function, Respiratory Symptoms, and Medication Use
25 The 2000 CO AQCD (U.S. EPA, 2000, 000907) briefly discussed the potential acute
26 respiratory health effects associated with short-term exposure to CO. An evaluation of the
27 epidemiologic literature at the time did not find any evidence of an association between short-term
28 exposure to CO and lung function, respiratory symptoms, or respiratory disease. As a result, the 2000
September 2009 5-115 DRAFT - DO NOT CITE OR QUOTE
-------
1 CO AQCD (U.S. EPA, 2000, 000907) did not conclude that a causal association exists between
2 short-term exposure to CO and respiratory health effects. Multiple uncertainties were identified in
3 the epidemiologic literature that contributed to this conclusion, which were discussed in
4 Section 5.2.1. The following section evaluates the current literature that examines the potential
5 association between short-term exposure to CO and respiratory health effects. Table 5-19 lists the
6 studies evaluated in this section along with the respiratory health outcomes examined and CO
7 concentrations reported.
September 2009 5-116 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-19
Author
O'Connor etal.
(2008, 156818)2
Rabinovitch et al.
(2004, 096753)
Silkoff etal. (2005,
087471)
Fischer et al.
(2002, 025731)1
Ranzi etal. (2004,
089500)1
Lagorio et al.
(2006, 089800)1
Penttinen et al.
(2001.030335)1
Timonen et al.
(2002, 025653)1
Chen etal. (1999,
011149)
Delfino et al.
(2003, 050460)
Slaughter et al.
(2003, 086294)
Yu et al. (2000,
013254)
Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with short-term exposure to CO.
Location
7 U.S. cities
Denver, CO
Year 1 : n = 41
Year 2: n = 63
Year 3: n = 43
Denver, CO
Year1:n=16
Year 2: n= 18
The Netherlands
n = 68
Emilia-Romagna
Region, Italy
n = 120
Rome, Italy
(n = 29)
Helsinki, Finland
n = 57
Kuopio, Finland
n = 33
Taiwan
n = 941
Los Angeles, CA
n = 22
Seattle, WA
n = 133
Seattle, WA
n = 133
Years
8/1998-7/2001
11/1999-3/2000;
11/2000-3/2001;
11/2001-3/2002
1999-2000
(winter);
2000-2001 (winter)
March -Aprils
2/1999-5/1999
5/1999-6/1999;
11/1999-12/1999
11/1996-4/1997
2/1994-4/1994
5/1995-1/1996
11/1999-1/2000
12/1993-8/1995
11/1993-8/1995
Health
Outcome
Pulmonary
function;
Respiratory
symptoms
Pulmonary
function;
Medication use
Pulmonary
function;
Medication use
Pulmonary
function
Pulmonary
function;
Respiratory
symptoms;
Medication use
Pulmonary
Function
Pulmonary
function
Pulmonary
function
Pulmonary
function
Asthma symptoms
Asthma symptoms;
Medication use
Asthma symptoms
Metric
8-hmax;24-h
avg
24- h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
1-hmax;
24-h avg
1-hmax;
8-h max
24-h avg
24-h avg
Mean
Concentration
(ppm)
NR
1.0
1999-2000:1.2
2000-2001 : 1 .1
0.80
Urban: 1.34
Rural: 1.06
Spring: 1.83
Winter: 10.7
Overall: 6.4
NR
0.52
NR
1-hmax: 7.7
8-h max: 5.0
NR
1.6
Middle/Upper Percentile
Concentrations (ppm)
8-h max:
50th: 1.2
75th 1 .8
99th 3.8
24-h avg:
50th: 0.7
75th 0.9
99th: 1.8
50th: 0.9
75th: 1.2
Maximum: 3.5
1999-2000
50th: 1.10
75th: 1.43
Maximum: 3.79
2000-2001
50th: 0.975
75th: 1.34
Maximum: 2.81
Max: 1.34
NR
Overall
Max: 25.1
50th: 0.35
75th: 0.43
Maximum: 0.96
Maximum: 2.43
1 -h max
Maximum: 3.6
1 -h max
90th: 12.0
Maximum: 17
8-h max
90th: 7. 9
Maximum: 10
50th: 1.47
75th: 1.87
50th: 1.47
Maximum: 4. 18
September 2009
5-117
DRAFT - DO NOT CITE OR QUOTE
-------
Author Location Years Ouf^ome Metric
Schildcrout et al.
(2006,089812)
von Klot et al.
(2002, 034706)1
Park etal. (2005,
088673)
Rodriguez et al.
(2007, 092842)
de Hartog et al.
(2003, 001061)1
8 North
American cities 11/1993.9/1995 Asthma symptoms; 24-havq
n = 990
10/1996-3/1997 i\/i H~ t ' 24-h 3V9
n = 53 meaicaiion use
Incheon, Korea Acthma cumntnmc-
•Wfin? fi/9nf)9 Mbmimbymyiumb, ~, .
O/£UU£ U/£UU£ MaHi^otinn nca ""9
n = 64 meaicaiion use
Perth, Australia Symptoms
6/1996-7/1998 associated with 8-h max
n = 263 respiratory illness
Amsterdam, the
Netherlands
n = 37
Erfurt, Germany 1998-1999 (winter) sySf^ 24-h avg
Helsinki, Finland
n = 47
Concentration
(ppm)
NR
0.78
Control days: 0.64
Dust days: 0.65
1.41
Amsterdam: 0.52
Erfurt: 0.35
Helsinki: 0.35
Middle/Upper Percentile
Concentrations (ppm)
50th: 0.63-1 .49
75th: 0.77-1 .90
90th: 0.95-2 .40
50th: 0.70
75th: 1.04
Maximum: 2.60
NR
Maximum: 8.03
Amsterdam: 1 .39
Erfurt: 2.17
Helsinki: 0.87
1These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which
assumes standard atmosphere and ambient temperature.
2This study did not present air quality statistics quantitatively, as a result, the air quality statistics presented were estimated from a figure.
3This study did not provide the year(s) in which air quality data was collected.
Pulmonary Function
I As part of the Inner-City Asthma Study (1CAS), O'Connor et al. (2008, 156818) examined the
2 effect of air pollutants (i.e., PM2.5, O3, NO2, CO, and SO2) on lung function in a population of 861
3 children (5-12) with persistent asthma in 7 urban U.S. communities. Throughout the study, %
4 predicted forced expiratory volume in 1 s (FEVi) and peak expiratory flow (PEF) were examined for
5 each subject during 2-week periods twice daily every 6 months for 2 yr. Lung function was
6 examined in single pollutant models using both same-day (lag 0) and 5-day (lag 0-4) moving average
7 pollutant concentrations (see Figure 5-12). CO was not found to be associated with % predicted
8 FEVi at lag 0, but there was evidence for a reduction in % predicted FEVi when using the 5-day
9 moving average (-0.32 [95% CI: -0.75, 0.11] per 0.5 ppm increase in 24-h avg CO concentrations).
10 When examining % predicted PEF, a reduction was observed at lag 0 (not reported quantitatively),
11 but the effect was found to be larger at lag 0-4 (-0.28 [95% CI: -0.71, 0.15]). In this study, CO was
12 found to be moderately correlated with other combustion related pollutants (e.g., PM2.5 [r = 0.44]
13 and NO2 [r = 0.54]), but CO was not included in the multipollutant models examined, limiting the
14 interpretation of the small reductions in lung function observed. Although the observed reductions in
15 lung function did not reach significance, the results do suggest a potential effect of CO on lung
16 function at relatively low CO concentrations (99th percentile max 8-h avg concentrations: ~
17 3.8 ppm).
September 2009
5-118
DRAFT - DO NOT CITE OR QUOTE
-------
Ozone PM 2.5 SO2
NO2
O/one PM 2.5
SO2 NO2 CO
Source: O'connor et al. (2008,156818)
Figure 5-12 Estimated effect (95% Cl) on pulmonary function due to a 10th to 90th percentile
increment change in pollutant concentration in single-pollutant models. The
estimates shown are from models that included either a 1-day or 5-day average of
pollutant concentration. Effect estimates were adjusted for site, month, site-by-
month interaction, temperature, and intervention group in mixed models. Figure
A, Percent predicted FEV1 as outcome variable. Figure B, Percent predicted
PEFR as outcome variable.
1 The remaining U.S.-based studies evaluated consisted of single-city studies conducted in
2 Denver, CO. Rabinovitch et al. (2004, 096753) examined the association between exposure to
3 ambient air pollutants and asthma exacerbation in a panel of urban minority children, 6-12 yr old,
4 with moderate to severe asthma over three winters. The investigators examined pulmonary function
5 by measuring FEVi and peak expiratory flow (PEF) in the morning on school days, and also at night
6 on weekends or other nonschool days. Using a 3-day moving average (lag 0-2) for all pollutants,
7 Rabinovitch et al. (2004, 096753) did not find an association between CO and either lung function
8 parameter during the morning or at night. Silkoff et al. (2005, 087471). also examined lung function
9 during the winter months, but in a panel of former smokers that were at least 40 yr old and had been
10 diagnosed with COPD. In this study, CO concentrations were similar to those reported in
11 Rabinovitch et al. (2004, 096753). The authors examined the association between exposure to air
12 pollutants and lung function (i.e., FEV i and PEF) in both the morning and the evening. Silkoff et al.
13 (2005, 087471) found contradictory results when examining the effects of CO for each of the winter
14 periods separately, 1999-2000 and 2000-2001. During the analysis of the first winter (i.e., 1999-
15 2000), CO was not found to be associated with lung function decrements in the morning at any lag,
16 but there was some evidence for lung function decrements during the evening at lag 0. Of note is the
17 increase in FEVi during the morning that was observed at lag 1 during this time period. For the
18 second winter (i.e., 2000-2001) the authors found a significant negative association between CO
19 exposure and FEVi in the evening at lag 2, and a moderate negative association with PEF at lag 0 in
20 the morning and lag 2 in the evening. Silkoff et al. (2005, 087471) postulated that the difference in
21 the FEVi results for the two study periods could be due to higher pollution concentrations along
September 2009
5-119
DRAFT - DO NOT CITE OR QUOTE
-------
1 with somewhat lower temperatures and higher humidity in 2000-2001. However, mean CO levels
2 remained relatively constant between the first and second winters, whereas, PMi0, PM2.5, and NO2
3 concentrations all increased. The decrements in FEV i observed in the second winter, therefore, may
4 have been due to the slightly worse, although not significantly different, baseline lung function of the
5 panel of subjects used during the second winter (Silkoff et al, 2005, 087471).
6 In the recent literature, the majority of studies that examined the association between short-
7 term exposure to CO and lung function have been conducted in Europe and the results provide
8 stronger evidence for CO-induced decrements in lung function parameters than studies conducted in
9 the U.S. Negative associations between short-term exposure to CO and lung function were observed
10 primarily in individuals with underlying respiratory conditions; however, some evidence also exists
11 for effects in children that live in urban environments. Penttinen et al. (2001, 030335) examined the
12 association between CO and lung function in a panel consisting of 57 non-smoking adult asthmatics
13 during the winter and spring in Helsinki, Finland. The authors observed negative associations with
14 PEF (L/min) for a 0.5 ppm increase in 24-h avg CO concentrations in the morning at lag 1 (|3 = -
15 0.54, SE = 0.084), and in the afternoon ((3 = -1.52, SE = 0.29) and evening ((3 = -1.81, SE = 0.27) for
16 a 5-day average. In two-pollutant models with daily mean particle number concentration (PNC), CO
17 effects on PEF in the morning were attenuated at lag 1, but remained negative. In addition, negative
18 associations with PEF persisted in the afternoon and evening in a two-pollutant model at lag 0. In
19 this study, moderate correlations between UFP and other traffic generated pollutants (e.g., CO
20 [r=0.44], NO [r=0,60], and NO2 [r=0.44]) make it difficult to attribute the observed respiratory
21 effects to a specific pollutant.
22 Lagorio et al. (2006, 089800) also conducted a study that examined the association between
23 CO and lung function in adults. In this study, 3 panels of subjects with underlying asthma, COPD, or
24 IHD that resided in Rome, Italy were selected. The ages of the subjects varied depending on the
25 panel, but overall the subjects ranged from 18-80 yr old. In single-pollutant models with CO, a
26 reduction in FVC (forced vital capacity) and FEVi was observed at most of the lags examined
27 (i.e., 0, 0-1, and 0-2) for both the COPD and asthma panels. No association was observed between
28 CO and FVC or FEV i in the IHD panel. Lagorio et al. (2006, 089800) did observe a relatively high
29 correlation between CO and PM2.5, but not NO2 (r=0.05). Copollutant models were not conducted in
30 this analysis to identify whether the CO associations observed are potentially confounded by other
31 pollutants.
32 Studies that focused on alterations in lung function in asthmatic children reported results
33 consistent with those observed in adult asthmatics. Timonen et al. (2002, 025653) examined the
34 effect of CO on bronchial responsiveness and pulmonary function (i.e., FVC, FEVi, MMEF, and
35 AEFV) at rest and after exercise in a panel of children 7-12 yr old with chronic respiratory
36 symptoms during the winter in Kuopio, Finland. The authors found that CO was significantly
September 2009 5-120 DRAFT - DO NOT CITE OR QUOTE
-------
1 associated with decrements in baseline lung function (i.e., lung function measured prior to exercise)
2 for FVC (mL) at lags 2 (-17.5 mL), 3 (-24.8 mL), and 4-day avg (-52.5 mL), and for FEVi (mL) at
3 lag 3 (-20.9 mL) for a 0.5 ppm increase in 24-h avg CO concentration. CO was not found to be
4 associated with exercise induced changes in lung function or bronchial responsiveness. Overall,
5 Timonen et al. (2002, 025653) found that increased concentrations of combustion-related byproducts
6 (i.e., BS, PMio, particle numbers, NO2, and CO) was associated with impairment in baseline lung
7 function. These associations, along with the high correlation between CO and combustion-related
8 pollutants (e.g., PMio [r=0.64]; NO2 [r=0.88]) contributed to the inability of the authors to conclude
9 that the lung function effects observed were due to biological changes in lung pathology specific to
10 CO exposure.
11 Chen et al. (1999, 011149) examined the effect of CO on lung function in 941 8-13 yr old
12 asthmatic children in Taiwan. The authors observed an association between short-term exposure to
13 CO and decrements in FVC (mL) at a 2-day lag when using daytime average CO concentrations
14 (from 8:00 a.m. to 6:00 p.m.) in a single-pollutant model. However, the authors found a high
15 correlation between CO and NO2 concentrations (r = 0.86-0.98), and did not conduct multipollutant
16 analyses.
17 One additional study, Fischer et al. (2002, 025731). examined the association between CO and
18 respiratory health, specifically lung function in anon-selected cohort study of 68 children ages 10-11
19 that live in an urban environment (Utrecht, the Netherlands). In this study, the authors examined
20 whether eNO was a more sensitive measure of lung damage than the traditional pulmonary function
21 measurements (i.e., FVC, FEVi, PEF, and MMEF). Fischer et al. (2002, 025731) found negative
22 associations between CO and FEVi, PEF, and MMEF at both lags 1 and 2, as well as, an association
23 between CO and an increase in eNO at lag 1. However, the lack of pollutant correlations and the
24 examination of copollutant models limit the interpretation of these results.
Respiratory Symptoms in Asthmatic Individuals
25 Upon evaluating the literature that examined the association between short-term exposure to
26 CO and respiratory symptoms in asthmatic individuals, consistent, positive associations were
27 observed across studies. The studies evaluated that included children enrolled in the Childhood
28 Asthma Management Program (CAMP) study found that CO was positively associated with asthma
29 symptoms. Yu et al. (2000, 013254) found an increase of 1.14-fold in asthma symptoms
30 ([95% CI: 1.05-1.23] per 0.5 ppm increase in 24-h avg CO concentrations at lag 1) in a population of
31 5-13 yr old asthmatic children (n = 133) in Seattle, WA. Similar effects were observed at lag 0 and
32 lag 2. These effects persisted when controlling for previous day's asthma symptoms at all lags, with
33 the largest effect at lag 1 (1.12 [95% CI: 1.05-1.19]), and in multipollutant models with PM 1.0 and
September 2009 5-121 DRAFT - DO NOT CITE OR QUOTE
-------
1 SO2. Using the same population of children, Slaughter et al. (2003, 086294) found an association
2 between short-term exposure to CO at lag 1 and asthma severity both with and without controlling
3 for the previous day's asthma severity, (RR = 1.04 [95% CI: 1.01-1.08]) and (RR = 1.03 [95% CI:
4 1.00-1.05]), respectively. However, this study only examined the effect of copollutant models on PM
5 risk estimates, not CO. Schildcrout et al. (2006, 089812) examined the association between air
6 pollutants and asthma symptoms in 990 children ages 5-12 in 8 North American cities. The authors
7 found a positive association between short-term exposure to CO and asthma symptoms at lag 0 (OR
8 =1.04 [95% CI: 1.00-1.07] per 0.5 ppm increase in 24-h avg CO), but similar effects were also
9 observed at lag 1, 2, and the 3-day moving sum. The CO effects observed persisted when NO2,
10 PMio, and SO2 where included in joint pollutant models.
11 As previously mentioned, O'Connor et al. (2008, 156818) conducted an additional multicity
12 study to examine the effect of air pollutants (i.e., PM2.5, O3, NO2, CO, and SO2) on respiratory
13 health in a population of 861 children (5-12) with persistent asthma in 7 U.S. urban communities.
14 The authors collected information on asthma symptoms every 2 months and examined the
15 association between a 2-week recall of the asthma symptoms and each air pollutant. O'Connor et al.
16 (2008, 156818)used a 19-day lag, which encompassed the 14 days of the symptom recall period and
17 the 5-day lag period proceeding the symptom recall period. In a single-pollutant model, CO was
18 significantly associated with number of days with a wheeze-cough (14% [95% CI: 2-29%]), number
19 of nights with asthma symptoms (i.e., nighttime asthma) (19%[95% CI: 4-36%]), and number of
20 days a child slowed down or stopped play (15% [95% CI: 2-30%]) per 0.5 ppm increase in 24-h avg
21 CO concentrations during the 2-week recall period. In this study, CO effects were not examined in a
22 multipollutant model.
23 U.S.-based single-city studies also found positive associations between CO and asthma
24 symptoms (Delfino et al., 2003, 050460: Rabinovitch et al., 2004, 096753). Rabinovitch et al. (2004,
25 096753) found evidence for an increase in asthma exacerbations in response to 24-h avg CO
26 concentrations for a 3-day moving average (lag 0-2) (OR = 1.02 [95% CI: 0.89-1.16] per 0.5 ppm
27 increase in 24-h avg CO) in a population of urban poor children with moderate to severe asthma in
28 Denver, CO. Delfino et al. (2003, 050460) also reported evidence of a positive association between
29 CO and asthma symptoms (based on symptoms that interfere with daily activities) using a population
30 of Hispanic children with asthma in a Los Angeles, CA, community. However, Delfino et al. (2003,
31 050460) only found positive associations at 1-day lags when using either the 1-hr maximum
32 (OR=1.05 [95% CI: 0.88-1.26] per 1 ppm increase in 1-hr max CO concentrations) or maximum 8-h
33 avg (OR=1.09 [95% CI: 0.80-1.50] per 0.75 ppm increase in max 8-hr avg CO concentrations) CO
34 concentration as the exposure metric. It should be noted that in comparison to Rabinovitch et al.
35 (2004, 096753) and the other respiratory symptoms studies discussed above, the mean ambient
36 concentrations for 1-h max and maximum 8-h avg reported by Delfino et al. (2003, 050460) were
September 2009 5-122 DRAFT - DO NOT CITE OR QUOTE
-------
1 7.7 ppm and 5.0 ppm, respectively, both of which are approximately 3.5 times higher than the
2 corresponding 24-h avg concentrations reported in the other studies.
3 In contrast to the U.S.-based studies presented above, international studies were evaluated that
4 examined the association between short-term exposure to CO and asthma symptoms in study
5 populations that included adults. Figure 5-13 summarizes the results from studies that provided
6 usable quantitative results and examined the association between short-term exposure to CO and
7 asthma or respiratory symptoms in asthmatic individuals. A panel study consisting of 53 adults with
8 asthma or asthma symptoms in Germany (Von Klot et al., 2002, 034706) observed a marginal
9 association between CO concentration and the prevalence of wheezing at lag 0 (OR = 1.03
10 [95% CI: 0.97-1.08] per 0.5 ppm increase in 24-h avg CO), and a positive association for a 5-day
11 mean concentration (OR = 1.12 [95% CI: 1.05-1.21] per 0.5 ppm increase in 24-h avg CO).
12 However, the authors found CO to be highly correlated with UFPs (r=0.66), complicating the
13 interpretation of the associations observed. Additionally, Park et al. (2005, 088673) in a panel study
14 of individuals 16-75 yr old in Incheon, Korea with bronchial asthma did not find an association
15 between CO and nighttime asthma symptoms or cough.
16 To further examine the effect of CO on asthma and asthma symptoms some studies also
17 analyzed medication use in asthmatic individuals in response to an increase in air pollutant
18 concentrations. The majority of U.S.-based studies (i.e., (Rabinovitch et al., 2004, 096753;
19 Schildcrout et al., 2006, 089812: Slaughter et al., 2003, 086294) focused on rescue inhaler use in
20 children with ages ranging from 5-13 yr old. Rabinovitch et al. (2004, 096753) found a weak
21 association (OR = 1.08 [95% CI: 1.00-1.17] per 0.5 ppm increase in 24-h avg CO) between rescue
22 inhaler use in a population of 6-12 yr old urban minority children with moderate to severe asthma in
23 the winter in Denver, CO. In a population of 5-12 yr old children with asthma in Seattle, WA,
24 Slaughter et al. (2003, 086294) found a stronger association with rescue inhaler use both with and
25 without taking into consideration the previous day's asthma severity, (RR: 1.04 [95% CI: 1.01-1.08]
26 per 0.5 ppm increase in 24-h avg CO) and (RR: 1.03 [95% CI: 1.00-1.05] per 0.5 ppm increase in
27 24-h avg CO), respectively. Similar results were observed in a multicity study conducted by
28 Schildcrout et al. (2006, 089812). which analyzed rescue inhaler use in 990 children ages 5-13 with
29 asthma in eight North American cities. Schildcrout et al. (2006, 089812) found that short-term
30 exposure to CO was positively associated with rescue inhaler use at lags of 0, 2, and a 3-day moving
31 sum, and that the association was fairly robust to a simultaneous increase in CO and other pollutants
32 (i.e., NO2, PMi0, and SO2) in joint models. Overall, Slaughter et al. (2003, 086294) and Schildcrout
33 et al. (2006, 089812) question the associations observed due to the lack of biological plausibility for
34 CO-induced respiratory effects, and the high correlation between CO and NO2 (which suggests that
35 other pollutants from mobile sources are driving the associations observed), respectively. Additional
36 studies (Park et al., 2005, 088673: Silkoff et al., 2005, 087471: Von Klot et al., 2002, 034706)
September 2009 5-123 DRAFT - DO NOT CITE OR QUOTE
-------
1 conducted in Denver, CO; Erfurt, Germany; and Incheon, Korea, respectively, found associations
2 between CO and medication use that are consistent with those previously reported, but in
3 populations with combined ages ranging from 16-77. Figure 5-13 presents the risk estimates from
4 studies that examined the association between short-term exposure to CO and medication use in
5 asthmatic individuals.
6
Reference
Location
Population
Lag
Respiratory Symptoms
Schildcrout et al. (2006,
0898121
O'Connor et al. (2008,
1568181
O'Connor et al. (2008,
1568181
Yu et al. (2000, 0132541
Rabinovitch et al. (2004,
0967531
Delfino et al. (2003, 0504601
Delfino et al. (2003, 0504601
von Klot et al. (2002,
034706)
8 N. American
cities
7 U.S. cities
7 U.S. cities
Seattle, WA
Denver, CO
Los Angeles, CA
Los Angeles, CA
Erfurt, Germany
Asthmatics (n=990)
Asthmatics (n=867)
Asthmatics (n=867)
Mild-to-moderate
asthmatics (n=133)
Asthmatics (n=1 47)
Hispanic children with
asthma (n=22)
Hispanic children with
asthma (n=22)
Asthma/asthma
symptoms (n=53)
0-2a Children
0-18
0-18
1
n °
1
•]
0-5 Adults
Schildcrout et al. (2006,
0898121
Slaughter etal. (2003,
0862941
Slaughter etal. (2003,
0862941
Rabinovitch et al. (2004,
0967531
von Klot et al. (2002,
0347061
von Klot et al. (2002,
0347061
8 N. American
cities
Seattle, WA
Seattle, WA
Denver, CO
Erfurt,
Germany
Erfurt,
Germany
Asthmatics (n=990)
Mild-to-moderate
asthmatics (n=133)
Mild-to-moderate
asthmatics (n=133)
Asthmatics (n=147)
Asthma/asthma
symptoms (n=53)
Asthma/asthma
symptoms (n=53)
0~23 Children
1
1
0-2
0-5 Adults
0-5
, ,
'-•- Asthma Symptoms, 5-12 yr
i/i/u i. r H-,
• Wheeze-cough, 5-12 yr
M- i-u.' 11- r ^
« Asthma Symptoms, 5-13 yr
A u. r- u. !• r- H-,
0 1 0 ^1 /H I.
Symptom Score >2 (
10-1 6 yr
— • Wheeze, 37-77 yr
0-1 6 yr
max 8-h avg),
Medication Use
» Inhaler Use, 5-12 yr
-~- Inhaler Use", 5-1 Syr
-•- Inhaler Use0, 5-1 Syr
— • Inhaler Use, 6-12 yr
— — Inhaler Use (|32-agonist), 37-77 yr
— • — Inhaler Use (Corticosteroid)
37-77 yr
0.75
1.00
125
1.50 1.75 2.00
Figure 5-13 Asthma symptoms, respiratory symptoms and medication use in asthmatic
individuals associated with short-term exposure to CO.1 Effect estimates were
standardized depending on the averaging time used in the study: 0.5 ppm for
24-h avg, 0.75 ppm for max 8-h avg, and 1.0 ppm for 1-h max.
•' Effect estimates from Park et al. (2005, 088673) were not included in this figure because the study did not provide the increment at
which the effect estimates were calculated. Additionally, estimates for Silkoff et al. (2005, 087471) were not included in the figure
because results were not presented quantitatively.
September 2009
5-124
DRAFT - DO NOT CITE OR QUOTE
-------
Respiratory Symptoms in Non-Asthmatic Individuals
1 In addition to examining the association between short-term exposure to CO and respiratory
2 symptoms (e.g., cough, wheeze, shortness of breath, etc.) in asthmatic populations some studies
3 examined respiratory effects in individuals classified as non-asthmatics. Rodriguez et al. (2007,
4 092842) examined the effect of CO on respiratory symptoms in a panel of 263 children 0-5 yr old at
5 high risk for developing asthma in Perth, Australia. Rodriguez et al. (2007, 092842) found CO
6 concentrations to be positively associated with wheeze/rattle chest and runny/blocked nose at both a
7 5-day lag and a 0-5-day lag. It is unclear which pollutant is driving the effect observed by Rodriguez
8 et al. (2007, 092842) because multipollutant models were not examined, CO correlations with other
9 pollutants were not presented, and additional analyses were not conducted to further characterize the
10 associations observed.
11 In a panel of individuals > 50 yr of age with CHD in three European locations (Amsterdam,
12 the Netherlands, Erfurt, Germany, and Helsinki, Finland) during the winter, de Hartog et al. (2003,
13 001061) observed some marginal associations, specifically between CO concentration and the
14 incidence of the respiratory symptoms shortness of breath and phlegm at lag 3, OR=1.17 (95% CI:
15 0.96, 1.40) and OR=1.22 (95% CI: 0.93, 1.57), respectively per 0.5 ppm increase in 24-h avg CO
16 concentrations. However, the authors found that the associations between air pollution exposure and
17 respiratory symptoms were stronger for PM2.s than for gaseous air pollutants. Overall, the
18 associations observed in this study should be viewed with caution because they are for a panel of
19 medicated individuals with CHD.
Summary of Associations between Short-Term Exposure to CO and Pulmonary
Function, Respiratory Symptoms, and Medication Use
20 A limited body of evidence is available that examined the effect of short-term exposure to CO
21 on various respiratory health endpoints. Among asthmatics, the studies reviewed generally found
22 positive associations between short-term exposure to CO and respiratory-related health effects
23 (i.e., decrements in lung function/lung function growth, respiratory symptoms, and medication use).
24 However, it can be observed that study authors often concluded that observed associations were due
25 to CO acting as an indicator for other traffic-related pollutants, primarily referring to the lack of an
26 understood biological mechanism for CO-induced respiratory effects. On-road vehicle exhaust
27 emissions are a nearly ubiquitous source of combustion pollutant mixtures that include CO and can
28 be an important contributor to CO-related health effects in near-road locations, which is evident by
29 the high correlations reported between CO and other combustion-related pollutants (i.e., NO2 and
30 PM). A lack of copollutant analyses among this group of studies complicates the efforts to
31 disentangle the health effects attributed to CO from the larger traffic-related pollutant mix.
September 2009 5-125 DRAFT - DO NOT CITE OR QUOTE
-------
1 Additional uncertainty exists as to a biologically plausible mechanism that could explain the effect of
2 CO on respiratory health.
5.5.1.2. Respiratory Hospital Admissions, ED Visits and Physician Visits
3 The 2000 CO AQCD (U.S. EPA, 2000, 000907) evaluated a limited amount of literature that
4 examined the association between short-term exposure to CO and respiratory hospital admissions
5 (HAs), ED visits, and physician visits in the U.S. (i.e., Seattle, WA, Reno, NV, and Anchorage, AK)
6 and Europe (i.e., Barcelona, Spain). From these studies, the 2000 CO AQCD (U.S. EPA, 2000,
7 000907) concluded that positive associations were observed for short-term exposure to CO with
8 several respiratory outcomes, including asthma and COPD. However, the lack of a biologically
9 plausible mechanism for CO-induced respiratory morbidity at that time brought into question
10 whether the results observed could be attributed to CO independently of other pollutants in the air
11 pollutant mixture. Additional uncertainties were identified in the epidemiologic literature that
12 contributed to this conclusion, which were discussed in Section 5.2.1.
13 This section evaluates those studies published since the 2000 CO AQCD (U.S. EPA, 2000,
14 000907) that examined the association between short-term exposure to CO at ambient concentrations
15 similar to those found in the U.S. and respiratory-related HAs (Figure 5-14), ED visits (Figure 5-15),
16 and physician visits. Unlike previous sections, which also evaluated studies conducted outside of
17 North America, the expansive number of studies conducted in the U.S. and Canada provides
18 adequate evidence to examine the association between short-term exposure to CO and respiratory
19 HAs and ED visits. Although not discussed in this section, collectively, the studies conducted outside
20 of the U.S. observed associations that are consistent with those observed in the U.S.- and Canadian-
21 based studies evaluated below (see Annex C for results from the international studies evaluated).
22 Overall, this section focuses on respiratory-related HAs because the majority of the literature
23 examines HAs as opposed to ED visits or physician visits (Table 5-20 presents the studies evaluated
24 in this section along with the range of CO concentrations measured in each study). It must be noted
25 that when examining the association between short-term exposure to CO and health outcomes that
26 require medical attention, it is important to distinguish between hospital admissions, ED visits, and
27 physician visits for respiratory outcomes (more so than for cardiovascular outcomes). This is because
28 it is likely that a small percentage of respiratory ED visits will be admitted to the hospital and,
29 therefore, may represent potentially less serious, but more common outcomes. To adequately
30 distinguish between the results presented in hospital admission, ED visit, and physician visit studies,
31 each outcome is evaluated in individual sections. In addition, each section presents results separately
32 for respiratory health outcomes which includes all respiratory diagnoses (ICD-9: 460-519) or
33 selected diseases (e.g., asthma, COPD, pneumonia and other respiratory infections) in order to
34 evaluate the potential effect of short-term exposure to CO on each outcome.
September 2009 5-126 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-20 Range of CO concentrations reported in key respiratory hospital admission and ED visit
studies that examine effects associated with short-term exposure to CO.
Author
Location
Type of Visit (ICD9)
Metric Mean Concentration (ppm)
Cakmaketal. 10 Canadian
(2006,093272) cities
HospitalAdmissions: Respiratory
disease (i.e., Acute bronchitis and
bronchiolitis; Pneumonia; Chronic
and unspecific bronchitis;
Emphysema;Asthma;
Bronchiectasis; Chronic airway
obstruction)
24-h avg 0.8
Maximum: 6.5
Linn et al.
(2000, 002839)
Slaughter et al.
(2005, 073854)
Burnett etal.
(2001.093439)
Yang et al.
(2003, 055621)
Lin et al. (2003,
042549)
Lin et al. (2004,
055600)
Moolgavkar
(2003, 042864)
Yang et al.
(2005, 090184)
Karretal.
(2006, 088751)
Los Angeles,
CA
Spokane, WA
Toronto, ON,
Canada
Vancouver, BC,
Canada
Toronto, ON,
Canada
Vancouver, BC,
Canada
Cook County,
IL; Los Angeles
County, CA
Vancouver, BC,
Canada
South Coast Air
Basin, CA
HospitalAdmissions: Pulmonary;
Asthma; COPD
ED Visits and HospitalAdmissions:
Respiratory; Asthma; COPD;
Pneumonia; Acute Respiratory
Infection
HospitalAdmissions: Respiratory
disease (i.e., Asthma;Acute
bronchitis/bronchiolitis; Croup;
Pneumonia)
HospitalAdmissions: Respiratory
diseases
Hospital Admissions : Asthma
Hospital Admissions : Asthma
HospitalAdmissions: COPD
HospitalAdmissions: COPD
Hospital Admissions : Acute
bronchiolitis
24-h avg
24-h avg
1-hmax
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
24-h avg
Winter: 1.7; Spring: 1.0;
Summer: 1.2; Fall: 2.1
Hamilton St.: 1.73
Backdoor Tavern: 1.29
Spokane Club: 1.41
Third and Washington: 1 .82
Rockwood:0.42
1.9
0.98
1.18
0.96
NR
0.71
Lag1:
Index: 1.730
Referrent: 1.750
Lag 4:
Index: 1.760
Referrent: 1.790
Maximum:
Winter: 5.3; Spring: 2.2;
Summer: 2.7; Fall: 4.3;
95th: 3. 05
50th: 1.8; 75th: 2.3;
95th: 3.3; 99th: 4.0
Maximum: 6.0
50th: 0.82; 75th: 1.1 6
Maximum: 4.90
50th: 1.10; 75th: 1.40
Maximum: 6. 10
50th: 0.80; 75th: 1.12
Maximum: 4.90
Cook:
50th: .99; 75th: 1.25
Maximum: 3.91
Los Angeles:
50th:1.35;75th:2.16
Maximum: 5.96
50th: 0.64
Maximum: 2.48
Lag1:
Index:
50th: 1.52; 75th: 2.26;
90th: 3. 16
Maximum: 9.60
Referrent:
50th: 1.51; 75th: 2.29;
90th: 3. 23
Maximum: 9.60
Lag 4:
Index:
50th: 1.54; 75th: 2.31;
90th: 3. 23
Maximum: 8.71
Referrent:
50th: 1.55; 75th: 2.35;
90th: 3. 30
Maximum: 9.60
September 2009
5-127
DRAFT - DO NOT CITE OR QUOTE
-------
Author Location
Karretal.
(2007, 090719)
Zanobetti and
Schwartz (2006,
090195)
Lin et al. (2005,
087828)
Peel et al.
(2005, 056305)
Tolbert et al.
(2007, 090316)
Ito et al. (2007,
156594)
Villeneuve et al.
(2006, 091179)
Sinclair etal.
(2004, 088696)
South Coast Air
Basin, CA
Boston, MA
Toronto, ON,
Canada
Atlanta, GA
Atlanta, GA
New York, NY
Toronto, ON,
Canada
Atlanta, GA
Type of Visit (ICD9)
Hospital Admissions : Acute
bronchiolitis
HospitalAdmissions: Pneumonia
HospitalAdmissions: Respiratory
infections
ED Visits: All respiratory; Asthma;
COPD;URI; Pneumonia
ED Visits: Respiratory diseases
(i.e., Asthma; COPD;URI;
Pneumonia; Bronchiolitis)
ED Visits: Asthma
Physicians Visits:
Allergic rhinitis
Urgent Care Visits:
Asthma; Respiratory infections
Metric Mean Concentration
24-havg: 1.720
M4onthTavg Monthly: 1 .770
24-h avg NR
24-havg 1.16
1-hmax 1.8
1-hmax 1.6
8-h max 1 .31
24-h avg 1 .1
1-hmax 1.3
, . Middle/Upper Percentile
(PPm) Concentrations (ppm)
24-havg:
50th: 1.61; 75th: 2.08;
90th: 2. 75
Maximum: 5.07
Monthly avg:
50th: 1.63; 75th: 2.13;
90th: 2. 88
Maximum: 8.30
50th: 0.48; 75th: 0.60;
95th: 0.88
50th: 1.05; 75th: 1.37
Maximum: 2.45
90th: 3. 4
50th: 1.3; 75th: 2.0; 90th: 3.0
Maximum: 7.7
50th: 1.23; 75th: 1.52;
95th: 2.11
Maximum: 2.2
NR
Hospital Admissions
Respiratory Disease
1 The majority of studies from North America that examined the association between short-term
2 exposure to CO and HAs for all respiratory diseases were conducted in Canada, and only one of
3 these studies presented results from a combined analysis of multiple cities (Cakmak et al., 2006,
4 093272). In a study of 10 of the largest Canadian cities, Cakmak et al. (2006, 093272) examined
5 respiratory HAs (ICD-9: 466, 480-486, 490-494, 496) in relation to ambient gaseous pollutant
6 concentrations for the time period 1993-2000. This study reported a 0.37% (95% CI: 0.12-0.50)
7 increase in respiratory hospital admissions for all ages for a 0.5 ppm increase in 24-h avg CO (lag
8 2.8 days averaged over the 10 cities1). However, Cakmak et al. (2006, 093272) found that this effect
9 was eliminated when including CO in a multipollutant model with other gaseous pollutants (i.e.,
10 NO2, SO2, and O3). U.S.-based studies (Los Angeles and Spokane) that examined HAs for all
11 respiratory diseases reported similarly weak or null associations with CO (Linn et al., 2000,
12 002839)(Slaughter et al., 2005, 073854). However, two single-city studies conducted in Canada
13 reported stronger associations, primarily through evidence from copollutant models, between short-
•' To determine the lag for the combined estimate across all 10 cities, Cakmak et al. averaged the strongest associations from lags 0-5 days
from each city.
September 2009
5-128
DRAFT - DO NOT CITE OR QUOTE
-------
1 term exposure to CO and respiratory disease HAs (Burnett et al., 2001, 093439; Yang et al, 2003,
2 055621). In a study conducted in Toronto, Canada for the time period 1980-1994, Burnett et al.
3 (2001, 093439) reported a relatively strong association between 1-h max CO and respiratory disease
4 HAs in children less than 2 yr of age, for the diagnoses of asthma (493), acute
5 bronchitis/bronchiolitis (466), croup (464.4), and pneumonia (480-486). The authors found a 9.7%
6 (95% CI: 4.1-15.5) increase in HAs for a 2-day avg (lag 0-1) per 1 ppm increase in 1-h max CO. In
7 the two-pollutant model analysis, the estimates for both CO and O3 remained elevated, but CO was
8 not found to be highly correlated with O3 (r=0.24). Yang et al. (2003, 055621) reported similar
9 results (OR = 1.04 [95% CI: 1.01-1.06] at lag 1 per 0.5 ppm increase in 24-h avg CO) for pediatric
10 (<3 yr of age) respiratory disease (ICD-9: 460-519) HAs in Vancouver for the time period
11 1986-1998. Yang et al. (2003, 055621) also reported elevated associations with 24-h avg CO and
12 respiratory HAs (ICD-9: codes 460-519) for ages 65 and over in Vancouver, Canada (OR = 1.02
13 [95% CI: 1.00-1.04]) at lag 1 for a 0.5 ppm increase in 24-h avg CO. Similar to Burnett et al. (2001,
14 093439). the authors found that the CO risk estimates remained elevated when O3 was included in
15 the model, which could be attributed to the negative correlation between CO and O3 (r=-0.52).
Asthma
16 Some studies that examined the effect of short-term exposure to CO on asthma HAs conducted
17 all age and age-stratified analyses, specifically to examine effects in children. In a few studies
18 conducted in Canada, evidence was observed for increased pediatric (ages 6-12) asthma hospital
19 admissions (ICD-9: 493) in boys, but not girls (Lin et al., 2003, 042549: Lin et al., 2004, 055600):
20 however, a biological explanation was not provided which could explain this difference. Lin et al.
21 (2003, 042549) used a bi-directional case-crossover analysis in Toronto, Canada for the years 1981-
22 1993. The authors reported an OR of 1.05 (95% CI: 1.00-1.11) per 0.5 ppm increase in 24-h avg CO
23 for a 1-day lag for boys with similar results being reported when averaging CO concentrations up to
24 7 days prior to a HA. Risk estimates for girls did not provide evidence of an association using the
25 same lag structure that was used in the boys' analysis (OR = 1.00 [95% CI: 0.93-1.06]); lag 1). In
26 this study, CO levels were moderately correlated with NO2 (r=0.55) and PM2.5 (r= 0.45), and weakly
27 correlated with SO2 (r=0.37). Lin et al. (2003, 042549) further examined the CO association in a
28 multipollutant analysis, and found that the estimates for boys were essentially unchanged when
29 adjusting for PMi0_2.5 and PM2.5; however, the study did not adjust for gaseous pollutants. It should
30 be noted that this study used a bi-directional case-crossover analysis, which may be biased (Levy et
31 al., 2001, 017172). Studies that examined the various referent selection strategies for the case-
32 crossover study design have concluded that the preferred control selection strategy is the time-
33 stratified framework (Levy et al., 2001, 017172). Lin et al. (2004, 055600) also examined the
34 association between air pollutants and asthma HAs (Lin et al., 2003, 042549) in children, but using a
September 2009 5-129 DRAFT - DO NOT CITE OR QUOTE
-------
1 time-series study design in Vacouver during the years 1987-1998. In this study the authors stratified
2 results by socioeconomic status (SES) and found some evidence for an association between CO and
3 asthma HAs for both girls and boys of both high and low SES at lag 1 (RR=1.01-1.06 per 0.5 ppm
4 increase in 24-h avg CO), but overall the evidence was less consistent for a greater effect in boys
5 versus girls compared to Lin et al. (2003, 042549). In a study that examined asthma HAs for all ages
6 and genders combined, Slaughter et al. (2005, 073854) observed some evidence for an increase in
7 asthma HAs (ICD-9 493) in Spokane (1995-2000) for CO at lag 2 (RR = 1.03 [ 95% CI: 0.98-1.08])
8 for a 0.5 ppm increase in 24-h avg CO,, but not for the other two lags examined (lag 1 and lag 3).
Chronic Obstructive Pulmonary Disease
9 A few of the studies examined the effect of short-term exposure to CO on COPD, or
10 obstructive lung disease, and HAs. Moolgavkar (2003, 042864) (a reanalysis of (Moolgavkar, 2000,
11 010274) examined HAs for COPD plus "allied diseases" (ICD-9 490-496) in two U.S. counties
12 (Cook County, IL and Los Angeles County, CA) for the years 1987-1995 using Poisson generalized
13 linear models (GLMs) or generalized additive models (GAM) with the more stringent convergence
14 criteria. Overall, the results from both models were similar. Using the GAM models the study
15 reported percent increases in HAs of 0.53-1.20% for all ages in Los Angeles County, and 0.17-1.41%
16 for ages > 65 in Cook County, for a 0.5 ppm increase in 24-h avg CO and lags ranging from 0 to
17 5 days. However, CO was found to be highly correlated with NO2 in both Cook County (r=0.63) and
18 Los Angeles County (r=0.80), but Moolgavkar (2003, 042864) did not examine the influence of
19 copollutants on CO risk estimates. Yang et al. (2005, 090184) reported similar results for COPD HAs
20 (ICD-9 490-492, 494, 496) in Vancouver for ages > 65 for the years 1994-1998 for a moving
21 average of 0-6 day lags (RR =1.14 [95% CI: 1.03-1.23] per 0.5 ppm increase in 24-h avg CO). In
22 this study, CO concentrations were moderately correlated with NO2, SO2, and PMi0 and moderately
23 negatively correlated with O3. In copollutant models, Yang et al. (2005, 090184) found that risk
24 estimates for CO and COPD HAs remained elevated with O3 or SO2, but were attenuated when
25 adjusting for NO2 or PM. Contradictory to Moolgavkar (2003, 042864) and Yang et al. (2005,
26 090184). Slaughter et al. (2005, 073854) found no association between short-term exposure to CO
27 and COPD HAs (ICD-9 491, 492, 494, 496) in Spokane, WA at lag 1-day (RR = 0.97
28 [95% CI: 0.93-1.01] per 0.5 ppm increase in 24-h avg CO) with similar results being reported for 2-
29 and 3-day lags. However, this study did not examine correlations between CO and other gaseous
30 pollutants or conduct copollutant analyses.
Acute Bronchiolitis in Infants
31 Karr et al. (2006, 088751: 2007, 090719) examined both short-term (lag 0 or 1) and longer
32 term levels of CO in relation to acute bronchiolitis (ICD-9: 466) hospital admissions during the first
33 year of life from 1995-2000 in the South Coast Air Basin in California. Karr et al. (2006, 088751)
September 2009 5-130 DRAFT - DO NOT CITE OR QUOTE
-------
1 found no evidence of a short-term association between ambient CO concentrations and HAs for
2 acute bronchiolitis at lag 1 day (OR= 0.99 [95%CI: 0.98-1.01] per 0.5 ppm increase in 24-h avg
3 CO). In addition, Karr et al. (2007, 090719). which examined longer term exposures (average in the
4 month prior to a HA and lifetime average) in a matched case-control study, did not provide any
5 evidence of an association with CO. Neither of these studies examined the correlation between CO
6 and other pollutants nor conducted copollutant analyses.
Pneumonia and Other Respiratory Infections
1 In addition to examining the effect of short-term exposure to CO on health outcomes that can
8 limit the function of the respiratory system, some studies examined the effect of CO on individuals
9 with pneumonia (ICD-9: 480-486) separately or in combination with other respiratory infections.
10 Zanobetti and Schwartz (2006, 090195) examined pneumonia HAs (ICD-9 480-487) in Boston, MA,
11 for the years 1995-1999 for individuals ages 65 and older using a time-stratified case-crossover
12 analysis. The authors reported an increase in pneumonia HAs at lag 0 of 5.4% (95% CI: 1.2-10.0)
13 per 0.5 ppm increase in 24-h avg CO. While Zanobetti and Schwartz (2006, 090195) did not report
14 multipollutant results, they suggested that CO was most likely acting as a marker for traffic-related
15 pollutants because CO was highly correlated with both BC (r = 0.80) and NO2 (r = 0.67), and
16 moderately correlated with PM2.5 (r = 0.52). Instead of examining the effect of CO on pneumonia
17 HAs separately, as was done by Zanobetti and Schwartz (2006, 090195). Lin et al. (2005, 087828)
18 presented results for the overall effect of CO on respiratory infection HAs (ICD-9: 464, 466, 480-
19 487). In this analysis, Lin et al. (2005, 087828) examined the potential increase in respiratory HAs in
20 children less than 15 yr of age in Toronto, Canada for 1998-2001 using a bi-directional case-
21 crossover approach. The authors reported elevated estimates for boys (OR=1.17 [95% CI: 1.03-1.32]
22 per 0.5 ppm increase in 24-h avg CO for a 6-day ma) while the estimate for girls was weaker and
23 with wider confidence intervals (OR=1.06 [95%CI: 0.91-1.23]). In multipollutant models with both
24 PM2.s and PMi0_2.5 the CO risk estimates were slightly attenuated, but remained positive (boys:
25 OR=1.10 [95% CI: 0.96-1.26]; girls: OR=1.03 [95% CI: 0.88-1.06]). Lin et al. (2005, 087828) did
26 not provide an explanation as to why the estimates are stronger for boys than for girls. It should be
27 noted that this study used a bi-directional case-crossover analysis, which, as discussed previously,
28 may bias the results (Levy et al., 2001, 017172).
September 2009 5-131 DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Location
Lag
Effect Estimates
n u. if\nnA nno ,11m
Burnett et al. (2001 , 093439)
Yang et al. (2003, 055621)
Linn etal. (2000, 002839)
Yang et al. (2003, 055621)
Cakmak etal. (2006,093272)
Slaughter et al. (2005, 073854)
I in at a I f9fim 049^40^
I in pt al I7t\m ndP'idm
Slaughter et al (2005 073854)
Slaughter et al. (2005, 073854)
Yang et al (2005 090184)
Zanobetti & Schwartz (2006, 090195)
Karr etal. (2006.088751)
Lin et al. (2005, 087828)
Lin et al. (2005, 087828)
loronto, LAN
Vancouver, CAN
Los Angeles, CA
Vancouver, CAN
10 Canadian Cities
Spokane, WA
Spokane WA
Spokane, WA
Vancouver CAN
Boston, MA
SQAB, CA
Toronto, CAN
Toronto, CAN
U-1
1
o
1
2.8
2 -c
•1
•1 .
2
2 — ,
0-6
0
1 -«i
0-5 <15yr
0-0
.
All Respiratory
- <2yr
• <3yr
• 30+ yr
— 65+ yr
> All ages
•— All ages
Asthma
COPD
-*. 15+ yr
R^J. \/r
Pneumonia
m 65+ yr
Acute Bronchiolitis
- <1 yr
Respiratory Infection
* Girls
.
O-9 1.O 1.1
1_2 1.3
1.4
Figure 5-14 Summary of associations between short-term exposure to CO and respiratory
hospital admissions.1'2 Effect estimates were standardized depending on the
averaging time used in the study: 0.5 ppm for 24-h avg, 0.75 ppm for max 8-h avg,
and 1.0 ppmfor1-h max.
•' Risk estimates from Moolgavkar (2003, 042864) were not included in this figure because the study presented a range of effect estimates
using different statistical models. The results from this study were more adequately highlighted in the evaluation of the study in the
COPD section.
• Risk estimates from Lin et al. (2004, 055600) were not included in the figure because the results were stratified by SES and therefore
could not be readily compared to effect estimates from Lin et al. (2003, 042549).
September 2009
5-132
DRAFT - DO NOT CITE OR QUOTE
-------
Emergency Department Visits
Respiratory Disease
1 Peel et al. (2005, 056305) conducted a large single-city respiratory disease ED visit study in
2 Atlanta, GA, which included data from 31 hospitals for the time period 1993-2000. In this study,
3 results were reported for various respiratory-related visits (ICD-9 460-466, 477, 480-486, 491-493,
4 496, 786.09). In an all ages analysis, the authors found a RR=1.01 (95% CI: 1.00-1.02) for all
5 respiratory disease ED visits for a 3-day avg (lag 0-2) per 1 ppm increase in 1-h max CO
6 concentration. Tolbert et al. (2007, 090316) expanded the time period used in the Peel et al. (2005,
7 056305) study to include ED visits through 2004, and reported similar results for all respiratory
8 disease ED visits (RR=1.013 [95% CI: 1.007-1.018] per 1 ppm increase in 1-h max CO). The CO
9 risk estimates from the Atlanta, GA, ED visits studies were attenuated when O3, NO2, or PM were
10 added to the model, which could potentially be explained by the high correlations between CO and
11 NO2 (r=0.70) and EC (r=0.66); and the moderate correlation with PM2.5 (r=0.51) reported in Tolbert
12 et al. (2007, 090316). One additional ED visits study that also examined respiratory disease
13 (Slaughter et al., 2005, 073854) presented essentially null results at lag 1 and 2, but found similar
14 results to Peel et al. (2005, 056305) and Tolbert et al. (2007, 090316) at lag 3 (RR=1.02
15 [95% CI: 1.00-1.03] per 0.5 ppm increase in 24-h avg CO). Slaughter et al. (2005, 073854) reported
16 a weak to moderate correlation between CO and various PM size fractions, but did not report the
17 correlation between CO and gaseous pollutants, limiting the comparison of this study with Peel et al.
18 (2005, 056305) and Tolbert et al. (2007, 090316).
Asthma
19 The association between short-term exposure to CO and asthma ED visits (ICD-9 493, 786.09)
20 was also examined in Atlanta, GA by Peel et al. (2005, 056305). In this study the authors reported
21 results from distributed lag models including lags 0-13 in addition to a moving average of lags 0, 1,
22 and 2 (lag 0-2) for specific respiratory outcomes (e.g., asthma). Effect estimates from the distributed
23 lag models were stronger than those produced from models that used 3-day moving average CO
24 concentrations (RR = 1.01 [95% CI: 0.99-1.02] for lags 0-2 compared to RR=1.08
25 [95% CI: 1.05-1.11] for an unconstrained distributed lag of 0-13 for a 1 ppm increase in 1-h max
26 CO). These results demonstrated the potential effect of CO exposures up to 13 days prior to an
27 asthma ED visit. Estimates were stronger for pediatric ED visits (ages 2-18 yr) (RR=1.02
28 [95% CI: 1.00-1.04] per 1 ppm increase in 1-h max CO) for a 3-day avg (lag 0-2) compared to all
29 ages (Peel et al., 2005, 056305). Although Peel et al. did not examine copollutant models, an
30 examination of pollutant correlations from a different publication from the same group (Metzger et
31 al., 2004, 044222). found that CO concentrations were moderately correlated with NO2, and PM and
September 2009 5-133 DRAFT - DO NOT CITE OR QUOTE
-------
1 weakly correlated with O3 and SO2. Slaughter et al. (2005, 073854). which also examined ED visits
2 for Spokane (1995-2001), reported an increase in asthma ED visits for all ages for CO at lag 3
3 (RR=1.03 [95% CI: 1.00-1.05] per 0.5 ppm increase in 24-h avg CO), but not for the other two lags
4 examined (lags 1 and 2). The results from Ito et al. (2007, 156594) also provide evidence of
5 increased ED visits for asthma (ICD-9 493) for all ages in New York City for 1999-2002. Using
6 three different models that adjusted for weather variables via either different degrees of smoothing
7 and/or a different number of weather variables, the authors found that CO effect estimates remained
8 elevated in both an all year analysis and in analyses stratified by warm and cold months. In addition,
9 Ito et al. (2007, 156594) examined copollutant models and found that CO effect estimates were
10 robust to the inclusion of PM25> O3 and SO2 in the model, but the CO risk estimate was attenuated,
11 resulting in a negative effect estimate when including NO2 in the model.
Chronic Obstructive Pulmonary Disease
12 In the examination of the effect of short-term exposure to CO on COPD ED visits (ICD-9 491,
13 492, 496), Peel et al. (2005, 056305) reported elevated estimates for Atlanta, GA for 1993-2000
14 (RR=1.03 [95%CI: 1.00-1.05] per 1 ppm increase in 1-h max CO for a moving average of lag 0-2)
15 with similar results for the distributed lag model (RR=1.03 [95% CI: 0.98-1.09). However, results
16 from Slaughter et al. (2005, 073854) from Spokane, WA were consistent with a null or slightly
17 protective association at lag 1 (RR=0.96 [95% CI: 0.92-1.00] per 0.5 ppm increase in 24-h avg CO at
18 lag 1) with similar results for lags 2 and 3.
Pneumonia and Other Respiratory Infections
19 Similar to the HA analysis conducted by Zanobetti and Schwartz (2006, 090195). discussed
20 above, Peel et al. (2005, 056305) examined the effect of CO on pneumonia separately (ICD-9: 480-
21 486), but also included an analysis of upper respiratory infection (ICD-9: 460-466, 477) ED visits for
22 all ages in Atlanta, GA during the years 1993-2000. The authors reported a weak estimate for
23 pneumonia for the three-day moving average (lag 0-2) (RR=1.01 [95% CI: 0.996-1.021] per 1 ppm
24 increase in 1-h max CO). However, when using an unconstrained distributed lag model (days 0-13),
25 Peel et al. (2005, 056305) observed evidence of an association (RR=1.045 [95% CI: 1.01-1.08]). An
26 examination of upper respiratory infection (URI) ED visits, the largest of the respiratory ED groups,
27 found slightly increased risk estimates for both the three-day moving average (lag 0-2) (RR=1.01
28 [95% CI: 1.00-1.02]) and the unconstrained distributed lag for days 0-13 (RR=1.07 [
29 95% CI: 1.05-1.09]) per 1 ppm increase in 1-h max CO. In copollutant models, CO risk estimates
30 were largely attenuated when PMi0, O3, or NO2 were included in the model, which could potentially
31 be explained by the correlation between CO and NO2, PM indices, and SO2 reported in Metzger et
32 al. (2004, 044222). Upon conducting an age-stratified analysis, Peel et al. (2005, 056305) also found
September 2009 5-134 DRAFT - DO NOT CITE OR QUOTE
-------
1 that infant (< 1 yr of age) and pediatric (ages 2-18) URI ED visit CO risk estimates were
2 substantially stronger than the all age risk estimates.
3
Reference
Location
Lag Effect Estimates
All Respiratory
Slaughter et al. (2005, 073854)
Peel etal. (2005,056305)
Tolbertetal. (2007.090316)
Spokane, WA
Atlanta, GA
Atlanta, GA
3
0-2
0-2
Peel et al. (2005. 056305)
Peel et al. (2005. 056305)
Peel et al. (2005. 056305)
Slaughter et al. (2005, 073854)
Atlanta, GA
Atlanta, GA
Atlanta, GA
Spokane, WA
0-2
0-2
0-13*
3
Slaughter et al. (2005, 073854)
Peel et al. (2005. 056305)
Peel et al. (2005. 056305)
Spokane, WA
Atlanta, GA
Atlanta, GA
1 .
0-2
0-13*
Peel et al. (2005. 056305)
Peel et al. (2005. 056305)
Atlanta, GA
Atlanta, GA
0-2
0-13*
Peel et al. (2005. 056305)
Peel et al. (2005. 056305)
Atlanta, GA
Atlanta, GA
0-2
0-13*
. . 1
— — — All ages
— • — All ages
— • — All ages
Asthma
1 1 8 yr
— • — All ages
COPD
15+yr
Pneumonia
— • — All ages
Respiratory Infection
—All ages
O.9O
O.95
1 .DO 1.05
1.1O 1.15
* Unconstrained distributed lag
Figure 5-15 Summary of associations between short-term exposure to CO and respiratory ED
visits. Effect estimates were standardized depending on the averaging time used
in the study: 0.5 ppm for 24-h avg, 0.75 ppm for max 8-h avg, and 1.0 ppm for 1-h
max.
Physician Visits
4 Although hospital admissions and ED visits are the two most well studied measures of
5 morbidity, a few studies also examined the effect of CO on unscheduled physician visits. In a time-
6 series study, Villeneuve et al. (2006, 091179) examined the effect of CO on physician visits for
7 allergic rhinitis in individuals 65 and older in Toronto, Canada. Although quantitative results were
8 only presented in figures, upon observation it was evident that estimates were consistent with a null
9 association for lags 0-6 (Villeneuve et al., 2006, 091179). In an additional study, Sinclair et al. (2004,
September 2009
5-135
DRAFT - DO NOT CITE OR QUOTE
-------
1 088696) reported results for urgent care visits for asthma and respiratory infections in a health
2 maintenance organization in Atlanta, GA; however, the study only reported statistically significant
3 results, of which none were for CO.
Summary of Associations between Short-Term Exposure to CO and Respiratory
Hospital Admissions, ED Visits, and Physicians Visits
4 Compared to other criteria air pollutants (e.g., O3 and PM), relatively few studies evaluated
5 the association between short-term exposure to ambient CO and hospital admissions and ED visits
6 for various respiratory outcomes. Although evidence for consistent positive associations (See Figure
7 5-14 and Figure 5-15) has been found across the studies evaluated, there remains uncertainty as to a
8 biologically plausible mechanism which could explain the association between CO exposure and
9 respiratory-related health effects. As observed in the preceding section, several authors suggest that
10 the observed associations are due to CO acting as an indicator of combustion-related pollution
11 (e.g., traffic). The interpretation of the associations observed in the studies evaluated is further
12 complicated by the moderate to high correlations reported between CO and other traffic-related
13 pollutants such as NO2, PM2.s, EC, or BC. Only a few studies examined potential confounding of
14 CO risk estimates by other pollutants through copollutant models, and these studies found that CO
15 risk estimates were robust or slightly attenuated, but remained positive in two-pollutant models with
16 O3, NO2, or PM indices.
5.5.2.Epidemiologic Studies with Long-Term Exposure
17 The 2000 CO AQCD (U.S. EPA, 2000, 000907) did not evaluate any studies that examined the
18 effect of long-term exposure to CO on respiratory health. The following section discusses those
19 studies that analyze the effect of long-term exposure to CO on pulmonary function, asthma/asthma
20 symptoms, and allergic rhinitis. Table 5-20 lists the studies evaluated in this section along with the
21 respiratory health outcomes examined and CO concentrations reported.
September 2009 5-136 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-21 Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with long-term exposure to CO.
Author1
Mortimer etal. (2008,
122163)
Meng et al. (2007, 093275)
Wilhelm etal. (2008,
191912)
Goss et al. (2004, 055624)
Hirsch etal. (1999.003537)
Guo etal. (1999.010937)
Wang etal. (1999.008105)
Hwang et al. (2005, 089454)
Hwang et al. (2006, 088971)
Lee etal. (2003. 049201)
Arnedo-Pena et al. (2009,
190238)
Mortimer etal. (2008,
187280)
Location
San Joaquin
Valley, CA
n=232
Los Angeles and
San Diego
counties, CA
Los Angeles and
San Diego
counties, CA
n=612
U.S.
Dresden, Germany
Taiwan
Kaohsiung and
Pintong, Taiwan
Taiwan
Taiwan
Taiwan
7 Spanish cities
Fresno, CA
n=170
Y-r(s) oEl
1989.2000 Pulmonary
11/2000-9/2001 Asthma symptoms
1999-2001 Asthma symptoms
Pulmonary
2000 function; Asthma
symptoms
9/1995-6/1996 Jg^
Igg4 Asthma; Asthma
symptoms
1996 Asthma
2000 Asthma
2000 Allergic rhinitis
1994 Allergic rhinitis
Asthma, allergic
2000 rhinitis, atopic
eczema
11/2000-4/2005 ^a(ion
Metric
Monthly mean
of max8-h
avg
Annual mean
of 1 -h avg
Annual mean
of 1 -h avg
Annual mean
of 1 -h avg
Annual mean
of 0.5-havg
Annual mean
of monthly avg
Annual avg
Annual mean
of monthly avg
Annual mean
of monthly avg
Annual avg
Annual avg
Monthly mean
of 24-h avg
Mean Middle/Upper
Concentration Percentile
(ppm) Concentrations (ppm)
NR
NR
1.0
0.69
0.60
0.85
NR
0.66
0.66
0.85
NR2
NR
NR
Maximum:
25th: 0.48
50th: 0.59
75th: 0.83
75th: 0.76
Maximum:
50th: 0.84
75th: 1.00
Maximum:
50th: 0.80
50th: 0.65
75th: 0.75
Maximum:
50th: 0.65
75th: 0.75
Maximum:
50th: 0.84
75th: 1.00
Maximum:
50th: 0.61
75th: 0.78
Maximum:
NR2
1.8
1.34
1.61
0.96
0.96
1.61
1.04
1The number of individuals included in the study population was only provided for those studies that included less than 1,000 participants.
2This study only presented air quality data graphically.
5.5.2.1. Pulmonary Function
1 Mortimer et al. (2008, 122163) examined the effect of prenatal and lifetime exposures to air
2 pollutants on pulmonary function in 232 asthmatic children that resided in the San Joaquin Valley of
September 2009
5-137
DRAFT - DO NOT CITE OR QUOTE
-------
1 California. The strong temporal correlation between pollutants and pollutant metrics for different
2 time periods in the study area contributed to the inability to draw conclusions about the effect of
3 individual pollutant metrics on pulmonary function (Mortimer et al., 2008, 122163). The authors
4 used a newly developed Deletion/Substitution/Addition (DSA) algorithm "to identify which
5 pollutant metrics were most predictive of pulmonary function" (Mortimer et al., 2008, 122163). This
6 methodology uses an exploratory process to identify the best predictive model for each outcome of
7 interest. Focusing specifically on the exposure durations after birth, using this approach, Mortimer
8 et al. (2008, 122163) found that exposure to CO early in life, ages 0-3, was negatively associated
9 with FEVi/FVC, resulting in an effect size of-2.5% per IQR increase in CO.1 Additional negative
10 associations were observed between exposure to CO during the first 6 yr of life and FEF25 (-6.7%)
11 and FEF25-75/FVC (-4.8%) in children diagnosed with asthma prior to 2 yr of age. Overall, Mortimer
12 et al. (2008, 122163) found that these effects were limited to subgroups, including African
13 Americans and individuals diagnosed with asthma before the age of 2 yr. It must be noted that
14 research still needs to be conducted to validate the aforementioned results obtained using the DSA
15 algorithm and the subsequent calculation of effect estimates using GEE because the current model
16 could underestimate the uncertainty surrounding the associations reported (Mortimer et al., 2008,
17 122163). Although the authors did find associations between long-term exposure to CO and
18 decrements in pulmonary function, they also observed high correlations between CO and NO2,
19 which together are markers for pollutants generated by urban combustion sources (e.g., mobile
20 sources) (Mortimer et al., 2008, 122163).
21 Goss et al. (2004, 055624) also examined the effect of long-term exposure to CO on
22 pulmonary function in a cohort of cystic fibrosis patients > 6 yr of age enrolled in the Cystic Fibrosis
23 National Patient Registry in 1999 and 2000. When examined cross-sectionally in 2000, using a
24 multiple linear regression model, the authors found no association between CO and a reduction in
25 FEVi. However, Goss et al. recognize that the CO results could be influenced by measurement error
26 and subsequently exposure misclassification.
5.5.2.2. Asthma and Asthma Symptoms
27 U.S.-based studies consistently reported no association between long-term exposure to CO and
28 asthma and asthma symptoms. Wilhelm et al. (2008, 191912) and Meng et al. (2007, 093275) both
29 examined the association between long-term exposure to air pollutants and asthma symptoms in
30 respondents to the 2001 California Health Interview Survey (CHIS) in populations consisting of
31 children (0-17) and adults (> 18), respectively, that resided in Los Angeles and San Diego counties.
32 Using a cross-sectional study design Meng et al. (2007, 093275) found no association between long-
•' The study did not present the IQR for CO; therefore, the effect estimates presented were not standardized using the approach mentioned
previously in this ISA.
September 2009 5-138 DRAFT - DO NOT CITE OR QUOTE
-------
1 term exposure to CO and poorly controlled asthma in adults, while Wilhelm et al. (2008, 191912)
2 reported no associations between long-term exposure CO and asthma symptoms or asthma HA and
3 ED visits in children during the study period (i.e., 2000-2001). In an additional U.S.-based study,
4 Goss et al. (2004, 055624) found no association (OR=1.01 [95% CI: 0.92-1.10] per 0.5 ppm increase
5 in annual average CO concentrations) between long-term exposure to CO and pulmonary
6 exacerbations in a national cohort of individuals with cystic fibrosis > 6 yr of age.
7 Among studies conducted in other countries, a study conducted in Germany (Hirsch et al.
8 (1999, 003537). and studies conducted in Taiwan (Guo et al. (1999, 010937). Wang et al. (1999,
9 008105). and Hwang et al. (2005, 089454). all found positive associations between long-term
10 exposure to CO and asthma or asthma symptoms in populations ranging from 6-16 yr old. In these
11 studies, the authors addressed the observed associations differently. Guo et al. (1999, 010937) and
12 Hwang et al. (2005, 089454) both concluded that it is unlikely CO directly affects the respiratory
13 system; Hirsch et al. (1999, 003537) attributed the increase in the prevalence of cough and bronchitis
14 to exposure to traffic-related air pollutants (i.e., NO2, CO, and benzene); and Wang et al. (1999,
15 008105) did not interpret the association observed between long-term exposure to CO and adolescent
16 asthma. Only Hwang et al. (2005, 089454) conducted a copollutant analysis and found that the
17 asthma effects observed were robust to the inclusion of PMi0, SO2 and O3 in the model. However,
18 this study did not include NOX in a copollutant model, which is notable because NOX was found to
19 be highly correlated with CO (r=0.88).
5.5.2.3. Allergy
20 Allergy is a major contributor to asthma and upper respiratory symptoms; as a result, studies
21 have examined the effect of air pollutants on allergic outcomes. The studies evaluated that examined
22 the association between long-term exposure to CO and allergic outcomes were primarily conducted
23 outside of the U.S. and Canada. A multicity study conducted in 7 Spanish cities, found that the
24 annual average concentration of CO was associated with a higher prevalence of allergic rhinitis,
25 rhinoconjunctivitis, and atopic eczema in 6-7 year-old children (Arnedo-Pena et al., 2009, 190238).
26 NO2 was also examined and found to be positively associated with allergic rhinitis, but, unlike CO,
27 was negatively associated with eczema and rhinoconjunctivitis. It should be noted that in this data
28 set CO and NO2 concentrations were negatively correlated (r=-0.55). Additionally, sulfur dioxide
29 (SO2) was positively associated with all allergic outcomes, while total suspended particulate (TSP)
30 matter was inversely associated with rhinitis and rhinoconjunctivitis. Hwang et al. (2006, 088971)
31 and Lee et al. (2003, 049201) both examined the effect of long-term exposure to air pollutants on the
32 prevalence of allergic rhinitis in a population of schoolchildren in Taiwan. Both studies found an
33 association between allergic rhinitis prevalence and CO, but they also observed an association with
34 NOX. As a result, although Hwang et al. (2006, 088971) and Lee et al. (2003, 049201) observed an
September 2009 5-139 DRAFT - DO NOT CITE OR QUOTE
-------
1 increase in the prevalence of allergic rhinitis in response to an increase in long-term CO levels, they
2 concluded that the combination of an association being observed for both CO and NOX can be
3 attributed to the complex mixture of traffic-related pollutants and not necessarily CO alone.
4 Although questions surround the associations observed between long-term exposure to CO and
5 allergic outcomes, the results are consistent with those presented in a multicity study that examined
6 the association between short-term exposure to CO and allergic symptoms. Moon et al. (2009,
7 190297) observed associations between short-term CO exposure and allergic symptoms in children
8 in South Korea. However, allergic symptoms were also associated with other pollutants, including
9 PMio, SO2, and NO2, and the study did not present correlation coefficients to allow for further
10 analysis of the results. It should be noted, toxicological experiments suggest that endogenously
11 produced CO may play an integral part in the pathogenesis of allergic rhinitis resulting in an
12 additional potential pathway for CO-induced allergic outcomes (Yu et al., 2008, 192384).
13 Allergic symptoms such as rhinitis are a direct result of allergic sensitization, which is
14 commonly measured by skin prick testing or IgE antibody measurement. Hirsch et al. (1999,
15 003537) in a single-city study conducted in Dresden, Germany observed no associations between
16 annual average concentrations of CO, NO2, SO2, or O3 and allergy assessed by skin prick testing or
17 serum IgE measurement in schoolchildren. However, prenatal exposure to CO was associated with
18 allergic sensitization in a cohort of 6-11 year-old asthmatic children in California (Mortimer et al.,
19 2008, 187280). Skin prick tests indicated higher levels of sensitization to indoor and outdoor
20 allergens with an increase in CO exposure during the prenatal period; the association with
21 sensitization to outdoor allergens remained after adjustment for effect modifiers, copollutants, and
22 other potential confounders. Mortimer et al. (2008, 187280) also found that PMi0 exposure was
23 associated with sensitization to indoor allergens, but was not significant after adjustment.
24 Additionally, despite strong correlations between CO and NO2, no associations were reported with
25 NO2. It should be noted, these results were produced using the DSA algorithm and as discussed
26 previously additional research is still needed to evaluate the use of this method in air pollution
27 epidemiology (Mortimer et al., 2008, 122163).
5.5.2.4. Summary of Associations between Long-Term Exposure to CO and
Respiratory Morbidity
28 To date, a limited number of studies have examined the potential association between long-
29 term exposure to CO and respiratory morbidity. Although studies have reported positive associations
30 for various respiratory outcomes, the limited evidence available, the new analytical methods
31 employed, and the lack of studies that examined potential confounders of the CO-respiratory
32 morbidity relationship, especially due to the high correlation between CO and other traffic-related
September 2009 5-140 DRAFT - DO NOT CITE OR QUOTE
-------
1 pollutants, makes it difficult to attribute the associations observed to CO independent of other air
2 pollutants.
5.5.3.Controlled Human Exposure Studies
3 Human clinical studies provide very little and inconsistent evidence of changes in pulmonary
4 function following exposure to CO. In one older study, Chevalier et al. (1966, 010641) observed a
5 significant decrease in total lung capacity following a short term exposure to 5,000 ppm resulting in
6 a COHb level of 4%. However, a similar study conducted at a higher CO concentration resulting in
7 COHb levels of 17-19% found no CO-induced changes in lung volume or mechanics (Fisher et al.,
8 1969, 012381). The 2000 CO AQCD (U.S. EPA, 2000, 000907) reported no evidence of CO-induced
9 changes in exercise ventilation at COHb levels <15% during submaximal exercise (Koike et al.,
10 1991, 013500). In two recent human clinical studies, exposure to CO (COHb ~ 10%) was not found
11 to significantly affect resting pulmonary ventilation compared with exposure to clean air under either
12 hypoxic or hyperoxic exposure conditions (Ren et al., 2001, 193850; Vesely et al., 2004, 194000).
13 The results of these studies demonstrate that the hypoxia- and CO2-induced increases in pulmonary
14 ventilation are not affected by CO. One recent study evaluated the potential anti-inflammatory
15 effects of controlled exposures to CO in the airways of 19 individuals with COPD (Bathoorn et al.,
16 2007, 193963). Subjects were exposed to both CO at concentrations of 100-125 ppm as well as room
17 air for 2 h on each of four consecutive days. The authors reported a small decrease in sputum
18 eosinophils, as well as a slight increase in the provocative concentration of methacholine required to
19 cause a 20% reduction in FEVi following exposure to CO. Although this study appears to
20 demonstrate some evidence of an anti-inflammatory effect of CO among subjects with COPD, it
21 must be noted that two of these patients experienced exacerbations of COPD during or following CO
22 exposure. A similar study found no evidence of systemic anti-inflammatory effects following
23 exposure to higher CO concentrations (500 ppm for 1 h) in a group of healthy adults (Mayr et al.,
24 2005, 193984).
5.5.4.Toxicological Studies
25 As discussed in Section 5.2.3., the work of Thorn, Ischiropoulos and colleagues (Ischiropoulos
26 et al., 1996, 079491; Thorn and Ischiropoulos, 1997, 085644; Thorn et al., 1997, 084337; Thorn et
27 al., 1999, 016753; Thorn et al., 1999, 016757) focused on CO-mediated displacement of NO from
28 heme-binding sites. Although the concentrations of CO used in many of their studies were far higher
29 than ambient levels, some of this research involved more environmentally-relevant CO levels. In one
30 study, 1-h exposure of rats to 50 ppm CO resulted in increased lung capillary leakage 18 h later
31 (Thorn et al., 1999, 016757). Increased NO was observed in the lungs by electron paramagnetic
September 2009 5-141 DRAFT - DO NOT CITE OR QUOTE
-------
1 resonance during 1-h exposure to 100 ppm CO and was accompanied by increases in H2O2 and
2 nitrotyrosine. All of these effects were blocked by inhibition of NOS. These results, which were
3 partially discussed in the 2000 CO AQCD, demonstrate the potential for exogenous CO to interact
4 with NO-mediated pathways and to lead to pathophysiological effects in the lung.
5 Recent work by Ohio et al, (Ohio et al, 2008, 096321) showed a disruption of cellular iron
6 homeostasis following exposure to a low level of CO (50 ppm x 24 h) in rats. In lungs of inhalation-
7 exposed rats, non-heme iron was significantly reduced, while lavagable iron was increased
8 dramatically, suggesting an active removal of cellular iron. Lavagable ferritin was also increased
9 following the CO exposure. Concurrently, liver iron levels increased, implying that the anatomical
10 distribution of iron stores may significantly shift during/after CO exposures. These investigators
11 were able to replicate the effect of loss of cellular iron in an in vitro model of cultured BEAS-2B
12 cells and reported statistically significant effects at 10 ppm CO and an apparent maximal effect at
13 50 ppm CO (concentrations up to 500 ppm did not significantly enhance the iron loss beyond
14 50 ppm). Similar responses were observed for cellular ferritin. Both enhancement of iron removal
15 and diminished iron uptake were noted in CO-exposed cells. Furthermore, decreased oxidative
16 stress, mediator release and proliferation were noted in respiratory cells. These effects were
17 reversible with a recovery period in fresh air. Interestingly, the in vivo exposure to CO induced mild,
18 but significant neutrophilia in the lungs compared to air-exposed rats. This finding is contrary to the
19 concept that CO acts as an anti-inflammatory agent; however, with alterations in iron handling
20 several potential pathways could be initiated to recruit inflammatory cells into airways. The authors
21 pointed out that while CO derived from HO activity may have an important role in iron regulation,
22 the non-specific application of exogenous CO will have little capacity to discriminate between
23 excessive and/or inappropriate iron which catalyzes oxidative stress and iron which may be required
24 for normal homeostasis.
25 A chronic inhalation study by Sorhaug et al. (2006, 180414) demonstrated no alterations in
26 lung morphology in Wistar rats exposed to 200 ppm CO for 72 wk. COHb levels were reported to be
27 14.7% and morphological changes were noted in the heart as described in Section 5.2.3.
28 A recent study by Carraway et al. (2002, 026018) involved continuous exposure of rats to HH
29 (380 torr) with or without co-exposure to CO (50 ppm) for up to 21 days. The focus of this study was
30 on remodeling of the pulmonary vasculature. While the addition of CO to HH did not alter the
31 thickness or diameter of vessels in the lung, there was a significant increase in the number of small
32 (<50 um) diameter vessels compared to control, HH only, and CO-only exposures. Despite the
33 greater number of vessels, the overall pulmonary vascular resistance was increased in the combined
34 CO + HH exposure, which the authors attribute to enhancement of muscular arterioles and p-actin.
35 One new study found an association between increased endogenous CO and the development
36 of allergic rhinitis (Yu et al., 2008, 192384). In this model, guinea pigs which were sensitized and
September 2009 5-142 DRAFT - DO NOT CITE OR QUOTE
-------
1 challenged with ovalbumin exhibited high immunoreactivity of HO-1 in the nasal mucosa and a
2 more than doubling of blood COHb levels (measured by gas chromatography). It is not known
3 whether the observed increase in endogenous CO resulting from ovalbumin-mediated
4 inflammation/oxidative stress plays a role in the development of allergic rhinitis but suggests a
5 potential mechanism by which exogenous CO could impact an allergic phenotype.
6 In summary, one older study (Thorn et al., 1999, 016757) and two new studies (Carraway et
7 al., 2002, 026018: Ohio et al., 2008, 096321) demonstrated effects of 50-100 ppm CO on the lung.
8 Responses included an increase in alveolar capillary permeability, disrupted iron homeostasis, mild
9 pulmonary inflammation and an exacerbation of pulmonary vascular remodeling elicited by HH.
10 These results should be considered in view of the potential for inhaled CO to interact directly with
11 lung epithelial cells and resident macrophages. However, a chronic study involving 200 ppm CO
12 demonstrated no changes in pulmonary morphology (Sorhaug et al., 2006, 180414).
5.5.5.Summary of Respiratory Health Effects
5.5.5.1. Short-Term Exposure to CO
13 New epidemiologic studies, supported by the body of literature summarized in the 2000 CO
14 AQCD, provide evidence of positive associations between short-term exposure to CO and
15 respiratory-related outcomes including pulmonary function, respiratory symptoms, medication use,
16 hospital admissions, and ED visits. The majority of this literature does not report results of extended
17 analyses to examine the potential influence of model selection, effect modifiers, or confounders on
18 the association between CO and respiratory morbidity. The lack of copollutant models, specifically,
19 has contributed to the inability to disentangle the effects attributed to CO from the larger complex air
20 pollution mix (particularly motor vehicle emissions), and this creates uncertainty in interpreting the
21 results observed in the epidemiologic studies evaluated. As discussed in previous sections, authors
22 often attributed associations reported with CO to the broader mixture of combustion-related
23 pollutants, citing a lack of understanding of the biological mechanisms for CO-related effects.
24 However, animal toxicological studies do provide some evidence that short-term exposure to CO
25 (50-100 ppm) can cause oxidative injury and inflammation and alter pulmonary vascular remodeling.
26 Controlled human exposure studies have not extensively examined the effect of short-term exposure
27 to CO on respiratory morbidity, but a few studies have found inconsistent evidence for CO-induced
28 effects on pulmonary function. Overall, the limited number of controlled human exposure studies
29 that have been conducted prior to and since the 2000 CO AQCD provide very little evidence of any
30 adverse effect of CO on the respiratory system at COHb concentrations relevant to the NAAQS.
31 Although controlled human exposure studies have not provided evidence to support CO-related
32 respiratory health effects, epidemiologic studies show positive associations for CO-induced lung-
September 2009 5-143 DRAFT - DO NOT CITE OR QUOTE
-------
1 related outcomes and animal toxicological studies demonstrate the potential for an underlying
2 biological mechanism, which together provide evidence that JS Suggestive Of 3 C3USal
3 relationship between short-term exposure to relevant CO concentrations and respiratory
4 morbidity.
5.5.5.2. Long-Term Exposure to CO
5 Currently, only a few studies have been conducted that examine the association between long-
6 term exposure to CO and respiratory morbidity including allergy. Although some studies did observe
7 associations between long-term exposure to CO and respiratory health outcomes key uncertainties
8 still exist. These uncertainties include: the lack of replication and validation studies to evaluate new
9 methodologies (i.e., Deletion/Substitution/Addition (DSA) algorithm) that have been used to
10 examine the association between long-term exposure to CO and respiratory health effects; whether
11 the respiratory health effects observed in response to long-term exposure to CO can be explained by
12 the proposed biological mechanisms; and the lack of copollutant analyses to disentangle the
13 respiratory effects associated with CO due to its high correlation with NO2 and other combust'on-
14 related pollutants. Overall, the evidence available is inadequate to conclude that a causal
15 relationship exists between long-term exposure to relevant CO concentrations and
16 respiratory morbidity.
5.6. Mortality
5.6.1.Epidemiologic Studies with Short-Term Exposure to CO
17 Epidemiologic studies have traditionally focused on mortality effects associated with exposure
18 to PM and O3, resulting in a limited number of studies that have conducted extended analysis to
19 examine the potential influence of model selection, effect modifiers, or confounders on the
20 association between CO and mortality. This has contributed to the inability to formulate a clear
21 understanding of the association between short-term exposure to CO and mortality. This section
22 summarizes the main findings of the 2000 CO AQCD (U.S. EPA, 2000, 000907). and evaluates the
23 newly available information on the relationship between short-term exposure to CO and daily
24 mortality in an effort to disentangle the CO-mortality effect from those effects attributed to other
25 criteria air pollutants.
September 2009 5-144 DRAFT - DO NOT CITE OR QUOTE
-------
5.6.1.1. Summary of Findings from 2000 CO AQCD
1 The 2000 CO AQCD (U.S. EPA, 2000, 000907) examined the association between short-term
2 exposure to CO and mortality through the analysis of primarily single-city time-series studies, with
3 additional evidence from one multicity study, which included 11 Canadian cities. While the results
4 presented by these studies did provide suggestive evidence that an association exists between CO
5 and mortality the AQCD concluded that inadequate evidence existed to infer a causal association
6 between mortality and short-term exposure to ambient concentrations of CO. Multiple uncertainties
7 were identified in the epidemiologic literature that contributed to this conclusion, which were
8 discussed in Section 5.2.1.
9 The majority of the recent time-series mortality studies, as mentioned previously, have not
10 extensively examined the CO-mortality relationship. As such, CO has usually been considered as one
11 of the potential confounding copollutants in air pollution epidemiologic studies. Given the limitation
12 that most of these studies were not conducted to examine CO, the goal of this review is to evaluate
13 the CO-mortality association, and specifically the: magnitude of associations; evidence of
14 confounding; and evidence of effect modification.
5.6.1.2. Multicity Studies
15 The following sections evaluate the recent literature that examined the association between
16 short-term exposure to CO and mortality, and in addition discuss newly available information with
17 regard to the issues specific to CO mentioned above. This evaluation focuses primarily on multicity
18 studies because they provide: a more representative sample of potential CO-related mortality effects;
19 and especially useful information by analyzing data from multiple cities using a consistent method,
20 and thus avoiding potential publication bias.1 Table 5-22 the multicity studies evaluated along with
21 the mean CO concentrations reported in each study.
•' To compare studies in this section that used different averaging times, effects estimates were standardized to the following: 0.5 ppm
studies that used 24-h avg concentrations and 0.75 ppm for studies that used max 8-h avg concentrations. These standardized values
represent the range of current mean ambient concentrations across the U.S.
for
September 2009 5-1 45 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-22 Range of CO concentrations reported in multicity studies that examine mortality effects
associated with short-term exposure to CO.
Author
Dominici et al. (2003, 056116; 2005,
087912)
Reanalysis of Samet et al. (2000, 156939)
Burnett etal. (2004, 086247)
Samoli et al. (2007. 098420)2
Location Years
82 U.S. cities1
1987-1994
(NMMAPS)
12 Canadian cities 1981-1999
19 European cities 100*1007-,
(APHEA2) 1990-19973
Averaging
Time
24-h avg
24-h avg
8-h max
Mean Range of Mean
Concentration Concentrations
(ppm) Across Cities (ppm)
Baton Rouge = 0.43
1.02
Spokane = 2. 19
Winnipeg = 0.58
1.02
Toronto = 1 .31
Basel = 0.52
2.12
Athens = 5.3
The study actually consisted of 90 U.S. cities, but only 82 had CO data.
2 This study presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm= 1.15 mg/m3, which assumes
standard atmosphere and temperature.
3 The study period varied from city to city. These years represent the total years in which data was collected across all cities.
National Morbidity, Mortality, and Air Pollution Study of 90 U.S. Cities
1 The time-series analysis of the 90 largest U.S. cities (82 cities for CO) in the National
2 Morbidity, Mortality, and Air Pollution Study (NMMAPS) (Dominici et al., 2003, 056116; Dominici
3 et al., 2005, 087912) (reanalysis of Samet et al., 2000, 156939) is by far the largest multicity study
4 conducted to date to investigate the mortality effects of air pollution, but the study primarily focused
5 on PMio. The range in 24-h avg CO concentrations in a subset of the largest 20 cities (by population
6 size) was 0.66 ppm (Detroit, MI) to 2.04 ppm (New York City). The analysis in the original report
7 used GAM with default convergence criteria. In response to the bias observed in the estimates
8 generated using GAM models with default convergence criteria (Dominici et al., 2002, 030458).
9 Dominici et al. (2003, 056116; 2005, 087912) (reanalysis of Samet et al. (2000, 156939) conducted a
10 reanalysis of the original data using GAM with stringent convergence criteria as well as GLM.
11 Focusing on the results obtained using GLM, PMi0 and O3 (in summer) appeared to be more
12 strongly associated with mortality than the other gaseous pollutants. The authors stated that the
13 results did not indicate associations between CO, SO2, or NO2, and total (nonaccidental) mortality.
14 However, as with PMi0, the gaseous pollutants CO, SO2, and NO2 each showed the strongest
15 association at a 1-day lag (for O3, a 0-day lag). Figure 5-16 presents the total mortality risk estimates
16 for CO from Dominici et al. (2003, 056116). The authors found a mortality risk estimate of 0.23%
17 (95% PI: 0.09, 0.36) per 0.5 ppm increase in 24-h avg CO for a 1-day lag in a single-pollutant
18 model. The inclusion of PMi0 or PMi0 and O3 in the model did not reduce CO risk estimates.
19 However, the confidence intervals were wider in the multipollutant models, but this could be
20 attributed to: (1) PMi0 data in many of the cities being collected every 6th day, as opposed to daily
21 data for gaseous pollutants; and (2) O3 being collected in some cities only during warm months. The
September 2009
5-146
DRAFT - DO NOT CITE OR QUOTE
-------
1 addition of NO2 (along with PMi0) to the model resulted in a reduced CO risk estimate. Some
2 caution is required when interpreting this apparent reduction because a smaller number of cities
3 (57 cities1) were available for the CO multipollutant analysis with PMi0 and NO2 compared to the
4 single-pollutant CO analysis (82 cities). However, most of the cities that did not have NO2 data (26
5 out of 32), and subsequently were not included in the multipollutant analysis, were some of the least
6 populated cities. Thus, the difference in the number of cities in the multi- and single-pollutant
7 analyses is unlikely to be the underlying cause for the reduction in the CO risk estimate in the CO
8 multipollutant analysis with PMi0 and NO2. In comparison to the PMi0 risk estimates, which were
9 not reduced in multipollutant models, the CO risk estimates from multipollutant models indicate less
10 consistent associations with mortality.
Figure 5-16
Q. 1.5; LagO Log1 Lag 2
T! 1.0:
a^;
C '
IT" °-5 - i .
]§ o * . " i
OO 0—T_: - - : _. _:_ -
S.E ' : • . : '" ' •
.£ CD -O.s: t
D5 CC ';
i 2 -1.0. . • :
§- -1.5-;
° ABCDEABCDEABCDE
Models
Source: Dominici et al.(2003, 0561161
Posterior means and 95% posterior intervals of national average estimates for
CO effects on total (non-accidental) mortality at lags 0,1, and 2 within sets of the
90 U.S. cities with available pollutant data. Models A = CO alone; B = CO + PMi0;
C = CO + PM10 + 03; D = CO + PM10 + N02; E = CO + PM10 + S02.
Canadian Multicity Studies
11 Since the 2000 CO AQCD (U. S. EPA, 2000, 000907) two Canadian multi city studies have
12 been published that examined the association between mortality and short-term exposure to air
13 pollutants: (1) an analysis of PMi0, PM25, PMi0_2.5, and gaseous pollutants in 8 cities from
14 1986-1996 (Burnett et al, 2000, 010273); and (2) an analysis of PM10, PM2.5, PM10.2.5, and gaseous
15 pollutants in 12 cities from 1981-1999 (Burnett et al., 2004, 086247). The 2000 study utilized GAM
•' One city was excluded from the multipollutant analysis because it contained NO2 data, but did not contain CO data.
September 2009
5-147
DRAFT - DO NOT CITE OR QUOTE
-------
1 with default convergence criteria, and upon reanalysis only examined PM indices (Burnett and
2 Goldberg, 2003, 042798).
3 Burnett et al. (2004, 086247) is the most extensive Canadian multicity study conducted to
4 date, both in terms of the length of the study and the number of cities covered. This study focused
5 primarily on NO2-mortality associations because it was found to be the best predictor of fluctuations
6 in mortality among the air pollutants examined (NO2, O3, SO2, CO, PM2.5, and PMi0_2.5); however,
7 the study did present single- and copollutant results for all pollutants included in the analysis. The
8 mean CO concentrations reported by Burnett et al. (2004, 086247) are similar to those reported in
9 NMMAPS (see Table 5-19).
10 Burnett et al. (2004, 086247) examined the effect of short-term exposure to CO on total
11 (nonaccidental) mortality. The authors found the strongest mortality association at lag 1-day for CO,
12 SO2, PM2.5, PMio_2.5, PMio (arithmetic addition of PM2.5 and PMi0_2.5), and CoH, whereas for NO2,
13 the strongest association was for the 3-day moving average (i.e., average of 0-, 1-, and 2-day lags),
14 and for O3, it was the 2-day moving average. In this study, Burnett et al. (2004, 086247) used 24-h
15 avg pollutant concentrations because these values showed stronger associations with mortality than
16 the daily 1-h max values for all of the gaseous pollutants and CoH, but not for O3. In a single-
17 pollutant model the CO risk estimate for total (nonaccidental) mortality was 0.33%
18 (95% CI: 0.12-0.54) per 0.5 ppm increase in 24-h avg CO at lag 1. After adjusting for NO2, the CO
19 risk estimate was reduced to 0.04% (95% CI: -0.19 to 0.26), while the NO2 risk estimate was only
20 slightly affected (increased from 2.25% to 2.35%) when including CO in the model. In this analysis,
21 a copollutant model including both CO and PM was not presented. The results presented in this
22 Canadian multicity study and NMMAPS are similar in that the CO risk estimates appeared to be
23 sensitive to the addition of NO2 in the regression model. However, interpretation of these results
24 requires some caution because: (1) NO2 tends to have a more spatially uniform distribution within a
25 city compared to CO; (2) CO and NO2 share common sources (e.g., traffic); and (3) CO and NO2 are
26 often moderately to highly correlated.
Air Pollution and Health: A European Approach
27 Most of the Air Pollution and Health: A European Approach (APHEA) analyses have focused
28 on the mortality effects of PM (PMi0 and BS), SO2, NO2, and O3, but not CO. In addition, some of
29 the analyses have not even considered CO as a potential confounder, such as the extended analysis
30 (APHEA2) of PM (Katsouyanni et al., 2001, 019008). and NO2. Gryparis et al. (2004, 057276) did
31 consider CO as a potential confounder in an analysis of O3, and found that the addition of CO
32 increased O3 mortality risk estimates both in the summer and winter although the number of cities
33 included in the copollutant model were reduced from 21 to 19. However, the study did not present
September 2009 5-148 DRAFT - DO NOT CITE OR QUOTE
-------
1 CO risk estimates. Unlike other APHEA studies (or the NMMAPS and Canadian multicity studies),
2 the Samoli et al. (2007, 098420) analysis focused specifically on CO.
3 Samoli et al. (2007, 098420) investigated the effect of short-term exposure to CO on total
4 (nonaccidental) and cardiovascular mortality in 19 European cities participating in the APHEA2
5 project by using a two-stage analysis to examine city-specific effects and potential sources of
6 heterogeneity in CO-mortality risk estimates. The mean levels of the max 8-h avg CO concentration
7 in this study ranged from 0.52 ppm (Basel, Switzerland, and the Netherlands) to 5.3 ppm (Athens,
8 Greece). The max 8-h avg CO concentration across all cities in the APHEA2 study of 2.12 ppm is
9 higher than the estimated max 8-h avg CO concentrations reported for the U.S. cities examined in
10 Dominici et al. (2003, 056116; 2005, 087912) and the Canadian cities examined in Burnett et al.
11 (2004, 086247) of 1.53 ppm.1 In APHEA cities, the correlation between CO and BS (r = 0.67-0.82)
12 was higher than the correlation between CO and PMi0 (r = 0.16-0.70) or CO and 1-h max NO2
13 (r = 0.03-0.68).
14 To examine the CO-mortality relationship, Samoli et al. (2007, 098420) conducted a time-
15 series analysis of individual cities following the revised APHEA2 protocol.2 The primary results
16 presented by the authors are from a sensitivity analysis that used two alternative methods to select
17 the extent of adjustment for temporal confounding. These methods consisted of: (1) confining the
18 extent of smoothing to 8 degrees of freedom per year (df/yr); and (2) selecting the appropriate extent
19 of smoothing through minimization of the absolute value of the sum of partial auto-correlation
20 functions (PACF) of the residuals, which resulted in the analysis using on average 5 df/yr for total
21 (nonaccidental) mortality and 4 df/yr for cardiovascular mortality. The authors also conducted
22 copollutant analyses using PMi0, BS, SO2, NO2, or O3 (1 h). In the second stage model Samoli et al.
23 (2007, 098420) examined heterogeneity in CO risk estimates between cities by regressing risk
24 estimates from individual cities on potential effect modifiers including: a) the air pollution level and
25 mix in each city (i.e., mean levels of pollutants, ratio PMi0/NO2); b) the exposure (number of CO
26 monitors, correlation between monitors' measurements); c) variables describing the health status of
27 the population (e.g., crude mortality rate); d) the geographic area (northern, western, and central-
28 eastern European cities); and e) climatic conditions (mean temperature and relative humidity levels).
29 Samoli et al. (2007, 098420) found that CO was associated with total (nonaccidental) and
30 cardiovascular mortality. The primary results represent the combined random effects estimate for a
31 0.75 ppm increase in max 8-h avg CO concentrations for the average of 0- and 1-day lag for total
32 (nonaccidental) mortality (1.03% [95% CI: 0.55-1.53]) and for cardiovascular mortality (1.08%
33 [95% CI: 0.25-1.90]). These results were obtained using PACF to choose the extent of adjustment for
•' The max 8-h avg concentration for the Dominici et al. (2003, 056116) and Burnett et al. (2004, 086247) studies was calculated using the
conversion factor of 2:3 to convert 24-h avg concentrations to max 8-h avg concentrations.
•2 The APHEA2 protocol used a Poisson GAM model with penalized splines as implemented in the statistical package R.
September 2009 5-149 DRAFT - DO NOT CITE OR QUOTE
-------
1 temporal trends. Although the results obtained using PACF are insightful, the use of 8 df/yr would
2 have been more consistent with the NMMAPS model (7 df/yr), and would have allowed for a more
3 accurate comparison of the results between APHEA2 and NMMAPS. The corresponding risk
4 estimates obtained using the 8 df/yr model are: 0.57% (95% CI: 0.23-0.91) for total (nonaccidental)
5 mortality and 0.70% (95% CI: 0.31-1.09) for cardiovascular mortality. In the sensitivity analysis,
6 Samoli et al. (2007, 098420) used 8 or 12 df/yr to adjust for temporal confounding. Both approaches
7 resulted in similar risk estimates, but using PACF to choose the extent of smoothing separately in
8 each city generally resulted in larger CO risk estimates (by -50-80%). This can be attributed to the
9 smaller number of df/yr used in the model (on average 5 df/yr for total (nonaccidental) mortality and
10 4 df/yr for cardiovascular mortality), which increases the magnitude of the effect and the amount of
11 observed heterogeneity (Samoli et al., 2007, 098420).
12 During the examination of the results obtained from the copollutant models, the authors noted
13 that there was indication of confounding of CO risk estimates by BS and NO2, but not PMi0. These
14 results are consistent with CO, BS, and NO2 being part of the traffic pollution mixture and PMi0
15 likely including secondary aerosols that do not correlate well with traffic-derived pollution. The risk
16 estimates from the model using 8 df/yr that included NO2 were: 0.26% (-0.09 to 0.61) for total
17 (nonaccidental) mortality and 0.37% (-0.05 to 0.80) for cardiovascular mortality. Thus, the inclusion
18 of NO2 in the model nearly halved the CO risk estimates (whereas the NO2 risk estimate was not
19 sensitive to the inclusion of CO in the model). CO risk estimates were reduced by a similar
20 magnitude when including BS in the model. Overall, the sensitivity of CO risk estimates to the
21 inclusion of NO2 in the model is consistent with the results presented in NMMAPS (Dominici et al.,
22 2003, 056116) and the Canadian multicity study (Burnett et al., 2004, 086247).
23 In the second stage model, Samoli et al. (2007, 098420) found that geographic region was the
24 most significant effect modifier, while the other effect modifiers (mentioned above) did not result in
25 strong associations. Effects were primarily found in western and southern European cities, and were
26 larger in cities where the standardized mortality rate was lower. Earlier APHEA studies also reported
27 a regional pattern of air pollution associations for BS and SO2, and found that western cities showed
28 stronger associations than eastern cities. However, the heterogeneity in CO risk estimates by
29 geographic region does not provide specific information to evaluate the CO-mortality association.
30 An ancillary analysis conducted by Samoli et al. (2007, 098420) examined the possible
31 presence of a CO threshold. The authors compared city-specific models to the threshold model,
32 which consisted of thresholds at 0.5 mg/m3 (0.43 ppm) increments. Samoli et al. (2007, 098420) then
33 computed the deviance between the two models and summed the deviances for a given threshold
34 over all cities. While the minimum deviance suggested a potential threshold of 0.43 ppm (the lowest
35 threshold examined), the comparison with the linear no-threshold model indicated weak evidence
36 (p-value >0.9) for a threshold. However, determining the presence of a threshold at the very low
September 2009 5-150 DRAFT - DO NOT CITE OR QUOTE
-------
1 range of CO concentrations (i.e., at 0.43 ppm) in this data set is challenging because in seven of the
2 19 European cities examined, the lowest 10% of the CO distribution was at or above 2 mg/m3
3 (1.74 ppm). Thus, the interpretation of the suggestive indication of athreshold is limited.
4 In summary, the APHEA2 analysis of CO in 19 cities found an association between CO and
5 total (nonaccidental) and cardiovascular mortality in single-pollutant models, but the associations
6 were substantially reduced when NO2 or BS was included in copollutant models. The evidence for
7 potential confounding of CO risk estimates by NO2 is consistent with the findings from NMMAPS
8 and Canadian 12 cities study. In addition, Samoli et al. (2007, 098420) found that geographic region
9 was a potential effect modifier, but such geographic heterogeneity is not specific to CO, based on
10 previously conducted APHEA studies. Finally, examination of the CO concentration-response
11 relationship found weak evidence of a CO threshold, which requires further investigation.
Other European Multicity Studies
12 An additional European multicity study was conducted by Biggeri et al. (2005, 087395) in
13 eight Italian cities. The authors examined the effect of short-term exposure to CO on mortality in
14 single-pollutant models using a time-series approach. In this analysis, all of the pollutants showed
15 positive associations with the mortality endpoints examined and the correlations among the
16 pollutants were not presented, therefore, it is unclear if the observed associations are shared or
17 confounded.
Summary of Multicity Studies
18 In summary, the mortality risk estimates from single-pollutant models are comparable for the
19 NMMAPS and Canadian 12-city studies, 0.23 and 0.33, respectively; with the estimate from the
20 APHEA2 study being slightly larger (0.57%) (Figure 5-17). In both the NMMAPS and Canadian
21 studies, a 1-day lag showed the strongest association; but the APHEA2 study used an a priori
22 exposure window (i.e., average of 0- and 1-day lags), which has been found to be the exposure
23 window most strongly associated with mortality in PM analyses.
24 The APHEA2 risk estimates presented in Figure 5-17 are from a model that used a fixed
25 amount of smoothing to adjust for temporal confounding (8 df/yr), which is similar to that used in
26 the NMMAPS study (7 df/yr). However, the APHEA2 sensitivity analysis suggested an approximate
27 50-80% difference in CO risk estimates between the models that used 8 or 12 df/yr, and the models
28 that used minimization of the absolute value of the sum of PACF of the residuals as a criterion to
29 choose the smoothing parameters. Thus, some model uncertainty likely influences the range of CO
30 risk estimates obtained from the studies evaluated.
September 2009 5-151 DRAFT - DO NOT CITE OR QUOTE
-------
1 The CO risk estimates from the aforementioned studies are also consistently sensitive to the
2 inclusion of NO2 in a copollutant model (0.11, 0.03, and 0.26%, for the NMMAPS, Canadian
3 12-city study, and APHEA2, respectively). Thus, these results suggest confounding by NO2.
4 However, this interpretation is further complicated because as with CO, NO2 itself may be an
5 indicator of combustion sources, such as traffic.
% Increase
Dominici et al. (2003,066116:2006,087912)
(reanalysis Samet et al. (2000,166939) NMMAPS, lag 1
82 cities COAIone
82 cities CO + PMio i
57 cities CO •*• PM 10 and NO-, ,
Burnett etal. (1998, 086427)
CO A
CO + NOna i
Samoli etal. (2007, 098420)
CO+ PMio
CO + BS H
CO + NO^ i
12 Canadian cities, lag 1
APHEA2 (19 cities), lag 0-1
CO Alone •
-0.4 -0.2 0.0 0.2 0.4 0.6 0.6 1.0
Figure 5-17 Summary of mortality risk estimates for short-term exposure to CO from multicity
studies. Estimates were standardized to 0.5 ppm and 0.75 ppm for studies that
used 24-h avg CO and max 8-h avg CO exposure metrics, respectively.
September 2009
5-152
DRAFT - DO NOT CITE OR QUOTE
-------
5.6.1.3. Meta-Analysis of All Criteria Pollutants
1 Stieb et al. (2002, 025205) reviewed the time-series mortality studies published between 1985
2 and 2000, and conducted a meta-analysis to estimate combined effects for PMi0, CO, NO2, O3, and
3 SO2. Because many of the studies reviewed in the 2000 analysis used GAM with default
4 convergence criteria, Stieb et al. (2003, 056908) updated the estimates from the meta-analysis by
5 separating the GAM versus non-GAM studies. In this meta-analysis the authors also presented
6 separate combined estimates for single- and multipollutant models. Overall, there were more GAM
7 estimates than non-GAM estimates for all of the pollutants except SO2. For CO, 4 single-pollutant
8 model risk estimates were identified, resulting in a combined estimate of 3.18% (95% CI: 0.76-5.66)
9 per 0.5 ppm increase in 24-h avg CO, and only 1 multipollutant model risk estimate (0.00%
10 [95% CI: -1.71 to 1.74]) from the non-GAM studies. Thus, for CO, this study did not provide useful
11 meta-estimates because the number of studies that contributed to the combined estimates for CO was
12 rather small.
5.6.1.4. Single-City Studies
13 In addition to the multicity studies discussed above, there have also been several single-city
14 U.S.- and Canadian-based time-series mortality studies that examined CO. The single-city studies,
15 similar to the multicity studies, often focused on the PM-mortality association, but also provided
16 additional information that is not available in the multicity studies. Because the sample size used in
17 each single-city study is small, and subsequently results in wide confidence intervals, a quantitative
18 comparison of the results from single- and multicity studies is difficult. In addition, some studies do
19 not present CO results quantitatively adding to the inability to adequately compare studies. Table
20 5-23 lists the single-city studies evaluated along with the mean CO concentrations reported in each
21 study.
September 2009 5-153 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-23 Range of CO concentrations reported in single-city studies that examine mortality
effects associated with short-term exposure to CO.
Study Location
De Leon et al. (2003, 055688) New York, NY
Klemm et al. (2004, 056585) Atlanta, GA
Vedal et al. (2003, 039044)1 Vancouver, BC, Canada
Villeneuve et al. (2003, 055051) Vancouver, BC, Canada
Goldberg et al. (2003, 035202) Montreal, Quebec, Canada
Hoeketal. (2000, 01 0350; 2001,
016550): Reanalyzed by Hoek (2003, The Netherlands
042818)
Years
1985-1994
1998-2000
1994-1996
1986-1999
1984-1993
1986-1994
Averaging
Time
24-h avg
1 -h max
24-h avg
24-h avg
24-h avg
24-h avg
Mean Concentration
(ppm)
2.45
1.31
0.5
1.0
0.8
Entire Country: 0.46
Four Major Cities: 0.59
Upper Percentile
Concentrations
(ppm)
95th: 4.04
Max: 7.40
75th: 1.66
Max: 1.9
90th: 0.9
Max: 4.9
90th: 1.6
Max: 5.1
75th: 1.0
Max.
Entire Country: 2.6
Four Major Cities: 4.6
1Study reported median CO concentrations.
Single-City Studies Conducted in the United States
1 De Leon et al. (2003, 055688) focused on the role of contributing respiratory diseases on the
2 association between air pollution (i.e., PMi0, O3, NO2, SO2, and CO) and primary non-respiratory
3 mortality (circulatory and cancer) in New York City, NY during the period 1985-1994. This study
4 only presented risk estimates graphically for each of the pollutants analyzed, except PMi0. In single-
5 pollutant models, PMi0, CO, SO2, and NO2 all showed the same pattern of association with
6 circulatory mortality for individuals > 75, indicating a larger risk of death in individuals with
7 contributing respiratory diseases compared to those without. In two-pollutant models, PMi0 and CO
8 risk estimates were reduced, but each remained significantly positive.
9 Klemm et al. (2004, 056585) analyzed 15 air pollutants for their associations with mortality in
10 Atlanta, GA, for a 2-yr period starting in August 1998. These pollutants included PM25, PMi0_2.5,
11 UFP surface area and counts, aerosol acidity, EC, OC, SO42", O3, CO, SO2, and NO2. This study
12 presented risk estimates using three levels of smoothing (quarterly, monthly, and biweekly knots) for
13 temporal trend adjustment, and suggested that the risk estimates were rather sensitive to the extent of
14 smoothing. It should be noted that temporal smoothing using biweekly knots is a more aggressive
15 modeling approach than the degrees of freedom approach used by most studies. In the single-
16 pollutant models for nonaccidental mortality, the strongest association, which was also statistically
17 significant, was found for PM25. CO, SO42", and PMi0_2.5 also showed positive associations with
18 nonaccidental mortality (CO: Quarterly knots and Monthly Knots (3 = 0.00002 [SE = 0.00001];
September 2009
5-154
DRAFT - DO NOT CITE OR QUOTE
-------
1 Biweekly knots (3 = 0.00001 [SE = 0.00002]). However, CO was significantly associated with
2 circulatory mortality in older adults (> 65), and these associations remained when PM2.5 was
3 included in the model (results were presented graphically).
Single-City Studies Conducted in Canada
4 Vedal et al. (2003, 039044) examined the association between short-term exposure to "low
5 levels" of air pollution (i.e., PMi0, O3, NO2, SO2, and CO) and daily morality in Vancouver, British
6 Columbia, Canada for the years 1994-1996. In this analysis, all of the risk estimates were presented
7 graphically; however, the results suggested that O3 in the summer and NO2 in the winter showed the
8 strongest associations with mortality. Vedal et al. (2003, 039044) found that CO was positively, but
9 not significantly associated with mortality. Additionally, the association between short-term exposure
10 to NO2 and mortality was found to be consistent with the results from the Canadian multicity study
11 conducted by Burnett et al. (2004, 188612).
12 Villeneuve et al. (2003, 055051) also conducted an analysis using data from Vancouver,
13 Canada, using a cohort of 550,000 individuals whose vital status was ascertained between 1986 and
14 1999. In this study, PM2.5, PM10.2.5, TSP, CoH, PM10, SO42", O3, CO, SO2, and NO2 were examined
15 for their associations with all-cause (nonaccidental), cardiovascular, and respiratory mortality. When
16 examining the association between gaseous pollutants and all-cause (nonaccidental) mortality in this
17 data set, NO2 and SO2 showed the strongest associations, while the association between CO and all-
18 cause mortality were generally weaker than those for NO2 and SO2. For cardiovascular mortality,
19 SO2 risk estimates were smaller than those for NO2 or CO, while for respiratory mortality, SO2
20 showed the strongest associations. However, the wider confidence intervals for these categories and
21 the smaller daily counts make it difficult to assess CO associations with cause-specific mortality
22 outcomes.
23 Goldberg et al. (2003, 035202) contrasted associations between air pollution and mortality in
24 individuals with underlying CHF versus mortality in individuals who were identified as having CHF
25 one year prior to death based on information from the universal health insurance plan in Montreal,
26 Quebec, Canada, during the period 1984-1993. In this study, Goldberg et al. (2003, 035202)
27 examined associations between PM2 5, CoH, SO42", O3, CO, SO2, and NO2, and mortality. The
28 authors found no association between any of the air pollutants and mortality with underlying CHF.
29 However, Goldberg et al. (2003, 035202) found positive associations between air pollution and
30 mortality in individuals diagnosed with CHF one year prior to death. Of the air pollutants examined,
31 CoH, NO2, and SO2 were most consistently associated with mortality for ages 65 and older, while
32 CO showed positive but weaker associations compared to these three pollutants.
September 2009 5-155 DRAFT - DO NOT CITE OR QUOTE
-------
Single-City Studies Conducted in Other Countries
1 Of the epidemiologic studies conducted in other countries that examine the association
2 between short-term exposure to CO and mortality only those studies conducted in European
3 countries that have CO levels comparable to the U.S. were evaluated. However, because Samoli et
4 al. (2007, 098420) conducted a multicity study of European cities that focused on short-term
5 exposure to CO, there are only a few single-city studies that provide additional information,
6 specifically those studies conducted in the Netherlands. The Netherland studies were evaluated
7 because they provide risk estimates for multiple pollutants and cause-specific mortality, and
8 consisted of relatively large sample sizes (i.e., the mortality time-series of the entire country was
9 analyzed).
10 Hoek et al. (2000, 010350) re-analyzed by Hoek (2003, 042818) examined associations
11 between air pollution and all-cause (nonaccidental), cardiovascular, COPD, and pneumonia deaths in
12 the entire Netherlands, the four major cities combined, and the entire country minus the four major
13 cities for the period 1986-1994. The air pollutants analyzed included BS, PMi0, O3, NO2, SO2, CO,
14 SO42~ and NO3~. In the single-pollutant models, all of the pollutants were significantly associated
15 with nonaccidental mortality at lag 1-day and 0-6 days when using the entire Netherlands data set. In
16 the two-pollutant model, CO risk estimates were reduced to null when PMi0, BS, SO42" and NO3~
17 were included in the model while the risk estimates for these copollutants remained significantly
18 positive. BS, CO, and NO2 were highly correlated (r >0.85) in this data set, and the authors noted
19 "all these pollutants should be interpreted as indicators for motorized traffic emissions" (Hoek et al.,
20 2000, 010350). The authors found that O3 showed the most consistent and independent associations
21 with mortality and that the risk estimates for all of the pollutants were substantially higher in the
22 summer months than in the winter months. Pneumonia deaths showed the largest risk estimates for
23 most pollutants including CO. The result from the Hoek et al. (2000, 010350) study is somewhat in
24 contrast to the result from the Samoli et al. (2007, 098420) multicity study in that, in the Hoek et al.
25 (2000, 010350) analysis, CO was more sensitive to the addition of PM indices in copollutant models.
26 This may be due to the high correlation between CO and PM indices in the Netherlands.
27 Hoek et al. (2001, 016550) reanalysis by Hoek (2003, 042818) analyzed the Netherlands data
28 using more specific cardiovascular causes of death: MI and other IHD, arrhythmia, heart failure,
29 cerebrovascular mortality, and embolism/thrombosis. In this analysis, the authors analyzed O3, BS,
30 PMio, CO, SO2, and NO2 in only single-pollutant models. For all of the pollutants, risk estimates
31 were larger for arrhythmia, heart failure, and cerebrovascular mortality than for the combined
32 cardiovascular mortality outcome. Thus, the results suggested larger impacts of air pollution on more
33 specific cardiovascular causes, but it is difficult to distinguish the effects of each pollutant from the
34 larger air pollution mixture.
September 2009 5-156 DRAFT - DO NOT CITE OR QUOTE
-------
5.6.1.5. Summary of Mortality and Short-Term Exposure to CO
1 The recently available multicity studies, which consist of larger sample sizes, along with the
2 single-city studies evaluated reported associations that are generally consistent with the results of the
3 studies evaluated in the 2000 CO AQCD (U.S. EPA, 2000, 000907). However, to date the majority
4 of the literature has not conducted extensive analyses to examine the potential influence of model
5 selection, effect modifiers, or confounders on the association between CO and mortality.
6 The multicity studies reported comparable CO mortality risk estimates for total (non-
7 accidental) mortality with the APHEA2 European multicity study (Samoli et al., 2007, 098420)
8 showing slightly higher estimates for cardiovascular mortality in single-pollutant models. However,
9 when examining potential confounding by copollutants these studies consistently showed that CO
10 mortality risk estimates were reduced when NO2 was included in the model, but this observation
11 may not be "confounding" in the usual sense in that NO2 may also be an indicator of other pollutants
12 or pollution sources (e.g., traffic).
13 Of the studies evaluated only the APHEA2 study focused specifically on the CO-mortality
14 association (Samoli et al., 2007, 098420). and in the process examined: (1) model sensitivity; (2) the
15 CO-mortality C-R relationship; and (3) potential effect modifiers of CO mortality risk estimates. The
16 sensitivity analysis indicated an approximate 50 - 80% difference in CO risk estimates from a
17 reasonable range of alternative models, which suggests that some model uncertainty likely influences
18 the range of CO mortality risk estimates obtained in the studies evaluated. The examination of the
19 CO-mortality concentration-response relationship found only weak evidence for a CO threshold at
20 0.5 mg/m3 (0.43 ppm). Finally, when examining a variety of city-specific variables to identify
21 potential effect modifiers of the CO-mortality relationship the APHEA2 study found that geographic
22 region explained most of the heterogeneity in CO mortality risk estimates.
23 The results from the single-city studies are generally consistent with the multicity studies in
24 that some evidence of a positive association was found for mortality upon short-term exposure to
25 CO. However, the CO-mortality associations were often, but not always, attenuated when
26 copollutants were included in the regression models. In addition, limited evidence was available to
27 identify cause-specific mortality outcomes (e.g., cardiovascular causes of death) associated with
28 short-term exposure to CO.
29 The evidence from the recent multi- and single-city studies suggests that an association
30 between short-term exposure to CO and mortality exists, but limited evidence is available to evaluate
31 cause-specific mortality outcomes associated with CO exposure. In addition, the attenuation of CO
32 risk estimates which was often observed in copollutant models contributes to the uncertainty as to
33 whether CO is acting alone or as an indicator for other combustion-related pollutants. Overall, the
September 2009 5-157 DRAFT - DO NOT CITE OR QUOTE
-------
i epidemioiogic evidence is suggestive of a causal relationship between short-term exposure
2 to relevant CO concentrations and mortality.
5.6.2.Epidemioiogic Studies with Long-Term Exposure to CO
3 The 2000 CO AQCD (U.S. EPA, 2000, 000907) did not evaluate the association between long-
4 term exposure to CO and mortality because there were no studies at the time that examined this
5 relationship. Since then there have been several new studies that examined the association between
6 long-term exposure to CO and mortality, but it should be noted that these studies primarily focused
7 on PM, and CO was only considered in these studies as a potential confounder. Therefore, the
8 information available from these new long-term exposure studies is somewhat limited, especially in
9 comparison to that for PM. Table 5-24 lists the U.S.-based studies evaluated that examined the
10 association between long-term exposure to CO and morality along with the mean CO concentrations
11 reported in each study.
September 2009 5-158 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-24 Range of CO concentrations reported in U.S.-based studies that examine mortality
effects associated with long-term exposure to CO.
Study
Jerrett et al. (2003, 087380)
Pope et al. (2002, 024689)
Krewski et al. (2009, 191193)
Miller etal. (2007.090130)
Lipfert et al. (2000,004087)
Lipfert etal. (2006.088756)
Lipfert etal. (2006.088218)
Lipfert and Morris (2002, 019217)
Location
107 U.S. cities
1980: 113 U.S. cities
1982-1998: 122 U.S.
cities
108 U.S. cities
36 U.S. cities
U.S.
U.S.
U.S.
1960-1969: 44 U.S.
counties
1970-1974: 206 U.S.
counties
1979-1981: 272 U.S.
counties
1989-1991: 246 U.S.
counties
1995-1997:261 U.S.
counties
Years
1980
1980
1982-1998
1980
2000
1960-1974
1975-1981
1982-1988
1989-1996
1999-2001
1976-1981
1982-1988
1989-1996
1997-2001
1960-1969
1970-1974
1979-1981
1989-1991
1995-1997
Averaging
Time
Annual avg
Annual avg
Annual avg
Annual avg
Mean annual 95th
percentile of
hourly CO values
Mean annual 95th
percentile of
hourly CO values
Mean annual 95th
percentile of
hourly CO values
Mean annual 95th
percentile of
hourly CO values
Mean Concentration ^SSSS
(Ppm) (ppm)
1.56
1980:1.7
1982-1998:
1.68
NR
1960-1974:
1975-1981:
1982-1988:
1989-1996:
1.63
1976-1981:
1982-1988:
1989-1996:
1997-2001:
1960-1969:
1970-1974:
1979-1981:
1989-1991:
1995-1997:
1.1
10.82
7.64
3.42
2.36
7.6
3.4
2.4
1.6
13.8
9.64
5.90
2.69
1.72
Maximum: 3.95
NR
75th: 2.13
90th: 2.58
95th: 3.05
Maximum: 3.95
NR
1960-1974
50th: 9.31
Maximum: 35.3
1975-1981
50th: 7.04
Maximum: 22.4
1982-1988
50th: 3.33
Maximum: 15.20
1989-1996
50th: 2.30
Maximum: 7. 10
Maximum: 6.7
NR
NR
September 2009
5-159
DRAFT - DO NOT CITE OR QUOTE
-------
5.6.2.1. U.S. Cohort Studies
American Cancer Society Cohort Studies
1 Pope et al. (1995, 045159) investigated associations between long-term exposure to PM and
2 mortality outcomes in the ACS cohort. In this study, ambient air pollution data from 151 U.S.
3 metropolitan areas in 1981 were linked with individual risk factors in 552,138 adults who resided in
4 these areas when enrolled in the prospective study in 1982. Death outcomes were ascertained
5 through 1989. PM2.5 and SO42~ were associated with total (nonaccidental), cardiopulmonary, and
6 lung cancer mortality, but not with mortality for all other causes (i.e., nonaccidental minus
7 cardiopulmonary and lung cancer). Gaseous pollutants were not analyzed in Pope et al. (1995,
8 045159). Jerrett et al. (2003, 087380). using data from Krewski et al. (2000, 012281). conducted an
9 extensive sensitivity analysis of the Pope et al. (1995, 045159) ACS data, augmented with additional
10 gaseous pollutants data. Due to the smaller number of CO monitors available compared to SO42", the
11 number of metropolitan statistical areas (MS As) included in the CO analysis were reduced (from 151
12 with SO42~) to 107. The mean annual CO concentrations in these MSAs ranged from 0.19 to
13 3.95 ppm. CO was weakly negatively correlated with SO42" (r = -0.07). Among the gaseous
14 pollutants examined (CO, NO2, O3, SO2), only SO2 showed positive associations with mortality, and
15 in addition was the only copollutant that reduced SO42" risk estimates. For CO, the relative risk
16 estimates for total (nonaccidental) mortality in single- and copollutant models with SO42" was 0.99
17 (95% CI: 0.96-1.01) and 0.98 (95% CI: 0.96-1.01), respectively, per 0.5 ppm increase in mean
18 annual average CO concentrations.
19 Pope et al. (2002, 024689) conducted an extended analysis of the ACS cohort with double the
20 follow-up time (to 1998) and triple the number of deaths compared to the original Pope et al. (2002,
21 024689) study. In addition to PM25, data for all of the gaseous pollutants were retrieved for the
22 extended period and analyzed for their associations with mortality-specific outcomes. As in the 1995
23 analysis, the air pollution exposure estimates were based on the MSA-level averages. The authors
24 found that PM2 5 and SO42~ were both associated with all-cause, cardiopulmonary, and lung cancer
25 mortality.1 In this study, the CO analysis used two different data sets. The first data set consisted of
26 1980 data and 113 MSAs; while the second data set used averages of the years 1982-1998 and
27 122 MSAs. The authors found, when using the 1980 data, that CO was not associated (risk estimates
28 ~ 1) (See Figure 5-18) with all-cause, cardiopulmonary, lung cancer, or mortality for all other causes.
29 However, the analysis of the 1982-1998 data found that CO was negatively (and significantly)
30 associated with all-cause, cardio-pulmonary, and lung cancer mortality. It is unclear why significant
31 negative associations were observed when analyzing the 1982-1998 data, but evidence from other
•' These results were presented graphically in Pope et al. (2002, 024689) and were estimated for Figure 5-18.
September 2009 5-160 DRAFT - DO NOT CITE OR QUOTE
-------
1 mortality studies that examined the association between long-term exposure to CO and mortality
2 does not suggest that CO elicits a protective effect.
3 Krewski et al. (2009, 191193) further analyzed the ACS cohort by adding two additional years
4 of mortality data (total period: 1982-2000). This study extended the range of the analysis to
5 incorporate sophisticated adjustment for bias and confounding as well as intra-urban analyses.
6 However, the CO analysis was limited to using: nationwide data, only 1980 CO concentrations, and
7 the standard Cox proportional hazards model. In addition to the death categories examined in Pope et
8 al. (2002, 024689). this analysis also examined ischemic heat disease (IHD) mortality. As was the
9 case with the Pope et al. (2002, 024689) analysis, Krewski et al. (2009, 191193) found that 1980 CO
10 data was not associated with any of the mortality categories examined: all-cause mortality HR=1.00
11 (95%CI: 0.99-1.01); cardio-pulmonary mortality, HR=1.00 (95% CI: 0.99-1.00); and IHD mortality,
12 HR=1.00 (95% CI: 0.99-1.01) per 0.5 ppm increase in CO.
Women's Health Initiative Cohort Study
13 Miller et al. (2007, 090130) studied 65,893 postmenopausal women between the ages of 50
14 and 79 yr without previous CVD in 36 U.S. metropolitan areas from 1994-1998. The authors
15 examined the association between one or more fatal or nonfatal cardiovascular events and air
16 pollutant concentrations. Exposures to air pollution were estimated by assigning the year 2000 mean
17 concentration of air pollutants measured at the nearest monitor to the location of residence of each
18 subject on the basis of its five-digit ZIP code centroid, which allowed estimation of effects due to
19 both within-city and between-city variation of air pollution. The investigators excluded monitors
20 whose measurement objective focused on a single point source or those with "small measurement
21 scale (0-100 m)." Thus, presumably these criteria reduced some of the exposure measurement error
22 associated with monitors that are highly impacted by local sources.
23 During the course of the study, a total of 1,816 women had one or more fatal or nonfatal
24 cardiovascular events, including 261 cardiovascular-related deaths. Hazard ratios for the initial
25 cardiovascular event were estimated. The following results are for models that only included subjects
26 with non-missing exposure data for all pollutants (n = 28,402 subjects, resulting in 879 CVD events).
27 In the single-pollutant models, PM2.5 showed the strongest associations with CVD events among all
28 pollutants (HR = 1.24 [95% CI: 1.04-1.48] per 10-(ig/m3 increase in annual average), followed by
29 SO2 (HR = 1.07 [95% CI: 0.95-1.20] per 5-ppb increase in the annual average). For CO the single-
30 pollutant risk estimate was slightly (but not significantly) negative (HR = 0.96 [95%CI: 0.84-1.10]).
31 In the multipollutant model, which included all pollutants (i.e., PM2.5, PMi0_2.5, SO2, NO2, and O3),
32 the CO risk estimate was similar to the one presented in the single-pollutant model (HR = 0.96
33 [95% CI: 0.82-1.14]). In addition, CO was not associated with CVD events in a single pollutant
September 2009 5-161 DRAFT - DO NOT CITE OR QUOTE
-------
1 model (HR = 1.00 [95%CI: 0.90-1.10] per 0.5 ppm increase in mean annual average CO
2 concentration) that used all available observations. Overall this study found that PM2.5 was clearly
3 the best predictor of cardiovascular events.
The Washington University-EPRI Veterans' Cohort Mortality Studies
4 Lipfert et al. (2000, 004087) conducted an analysis of a national cohort of -70,000 male U. S.
5 military veterans who were diagnosed as hypertensive in the mid 1970s and were followed for
6 approximately 21 yr (up to 1996). Demographically, 35% of the cohort consisted of African
7 American men and 57% of the cohort was defined as current smokers; however, 81% of the cohort
8 had been smokers at one time in their life. The study examined mortality effects in response to long-
9 term exposure to multiple pollutants including, PM2.5, PMio, PMi0_2.5, TSP, SO42", CO, O3, NO2,
10 SO2, and Pb. Lipfert et al. (2000, 004087) estimated exposures by indentifying the county of
11 residence at the time of entry to the study. Four exposure periods (1960-1974, 1975-1981, 1982-
12 1988, and 1989-1996) were defined, and deaths during each of the three most recent exposure
13 periods were considered. The mean annual 95th percentile of hourly CO values during these periods
14 declined from 10.8 ppm to 2.4 ppm. The authors noted that the pollution risk estimates were
15 sensitive to the regression model specification, exposure periods, and the inclusion of ecological and
16 individual variables. Lipfert et al. (2000, 004087) reported that indications of concurrent mortality
17 risks (i.e., associations between mortality and air quality for the same period) were found for NO2
18 and peak O3. The estimated CO mortality risks were all negative, but not significant.
19 Lipfert et al. (2006, 088756) examined associations between traffic density and mortality in
20 the same Veterans' Cohort, but in this analysis the follow-up period was extended to 2001. As in
21 their 2000 study, four exposure periods were considered but more recent years were included in the
22 2006 analysis. The authors used the mean annual average of the 95th percentile of 24-h avg CO in
23 each of the exposure periods as the averaging metric. The traffic density variable was the most
24 significant predictor of mortality in their analysis, remaining so in two- and three pollutant models
25 with other air pollutants (i.e., CO, NO2, O3. PM2.5, SO42~, non-SO42~ PM2.5, and PM 10-2.5). In the
26 multipollutant models, mortality risk estimates were not statistically significant for any of the other
27 pollutants, except O3. The natural log of the traffic density variable (VKTA = vehicle-km traveled
28 per year) was not correlated with CO (r = -0.06), but moderately correlated with PM2 5 (r = 0.50) in
29 this data set. For the 1989-1996 data period, the estimated mortality relative risk was 1.02
30 (95% CI: 0.98-1.06) per 1 ppm increase in the mean annual 95th percentile of hourly CO
31 concentration in a single-pollutant model. The two-pollutant model, which included the traffic
32 density variable, resulted in a relative risk of 1.00 (95% CI: 0.96-1.04). Lipfert et al. (2006, 088218)
33 note that the low risk estimates for CO in this study were consistent with those observed in other
September 2009 5-162 DRAFT - DO NOT CITE OR QUOTE
-------
1 long-term exposure studies, and may have been due to the localized nature of CO, which can lead to
2 exposure errors when data from centralized monitors is used to represent an entire county.
3 Interestingly, as Lipfert et al. (2006, 088756) pointed out, the risk estimates due to traffic density did
4 not vary appreciably across these four periods even though regulated tailpipe emissions declined
5 during the study period. The authors speculated that some combination of other environmental
6 factors such as road dust, psychological stress, and noise (all of which constitute the environmental
7 effects of vehicular traffic) along with spatial gradients in SES might contribute to the non-negative
8 effects observed.
9 Lipfert et al. (2006, 088218) extended the analysis of the Veterans Cohort data to include the
10 EPA's Speciation Trends Network (STN) data, which collected chemical components of PM2.5. The
11 authors analyzed the STN data for the year 2002, and again used county-level averages. In addition,
12 they analyzed PM2.s and gaseous pollutants data for 1999 through 2001. As in the other Lipfert et al.
13 (2006, 088218) study, traffic density was the most important predictor of mortality, but associations
14 were also observed for EC, vanadium (V), nickel (Ni), and NO3 -. Ozone, NO2, and PMi0 also
15 showed positive, but weaker associations. The authors found no association between the mean
16 annual 95th percentile of hourly CO values and mortality (RR = 0.995 [95% CI: 0.988-1.001] per
17 1 ppm increase in CO concentration) in a single-pollutant model. The study did not present
18 multipollutant model results for CO.
September 2009 5-163 DRAFT - DO NOT CITE OR QUOTE
-------
Endpoint
All-cause
Cardio-pulmonary
Cardiovascular
IHD
Lung Cancer
Reference
Pope et al. (2002,
024689)"
Krewski et al. (2009,
191193)
Pope et al. (2002,
024689)3
Lipfert et al. (2006,
088218)
Lipfert etal. (2006,
088218)
Lipfert etal. (2006,
088756)
Jerrett et al. (2003,
087380)
Jerrett et al. (2003,
087380)
Pope et al. (2002,
024689)"
Krewski et al. (2009,
191193)
Pope et al. (2002,
024689)"
Miller etal. (2007,
090130)"
Miller etal. (2007,
090130)"
Krewski et al. (2009,
191193)
Krewski et al. (2009,
191193)
Cohort
ACS
ACS
ACS
Veterans
Veterans
Veterans
ACS
ACS
ACS
ACS
ACS
WHI
WHI
ACS
ACS
MoSyData *?Sta Effect Estimates
1982-1998
1982-2000
1982-1998
1976-2001
1976-2001
1997-2001
1982-1998
1982-1998
1982-1998
1982-2000
1982-1998
1994-1998
1994-1998
1982-2000
1982-2000
1980
1980
1982-1998 -«-
iqaq iqqfi
1 080 1 0Ofi •
1999-2001 -•
1982C *
1982C — •-
1980
1980
1982-1998 — •—
>-
i.
. + |nA/KTA^d
_
- + S042'
1-
h
+ PM 2.5, PM 10-2.5, S02,N02,
1980 >-
1980 -»j-
0.80 0.90 1.00 1.10 1.20
'The study does not present CO results quantitatively. This effect estimate and 95% confidence interval were estimated from Figure 5 in Pope et al. (2002, 024689).
bEffect estimate is only for subjects with non-missing exposure data for all pollutants.
'The study did not report the range of years of CO data used; however, it does specify that air quality data was obtained from pollution monitoring stations operating in 1982.
dNatural log of Vehicle-km Traveled variable.
Figure 5-18 Summary of mortality risk estimates for long-term exposure to CO. Estimates
were standardized to 0.5 ppm and 1.0 ppm for studies that used mean annual
average CO and the mean annual 95th percentile of hourly CO values exposure
metrics, respectively.
5.6.2.2. U.S. Cross-Sectional Analysis
1 An ecological cross-sectional analysis involves regressing county- (or city) average health
2 outcome values on county-average explanatory variables such as air pollution and census statistics.
September 2009
5-164
DRAFT - DO NOT CITE OR QUOTE
-------
1 Unlike the cohort studies described above, to the extent that individual level confounders are not
2 adjusted for, the cross-sectional study design is considered to be subject to ecologic confounding.
3 However, all of the cohort studies described above are also semi-ecologic in that the air pollution
4 exposure variables are ecologic (Kunzli and Tager, 1997, 086180). In this sense, cross-sectional
5 studies may be useful in evaluating the correlation among exposure variables.
6 Lipfert and Morris (2002, 019217) conducted ecological cross-sectional regressions for U.S.
7 counties (except Alaska) during five periods: 1960-1969, 1970-1974, 1979-1981, 1989-1991, and
8 1995-1997. They regressed age-specific (15-44, 45-64, 65-74, 76-84, and 85+) all-cause (excluding
9 AIDS and trauma) mortality on air pollution, demography, climate, SES, lifestyle, and diet. The
10 authors analyzed TSP, PMi0, PM2.5, SO42", SO2, CO, NO2, and O3. However, air pollution data was
11 only available for limited periods of time depending on the pollutant: TSP up to 1991; PMi0 between
12 1995 and 1999; and PM2.5 between 1979-1984 and 1999. In response to the varying number of
13 counties with valid air pollution data by pollutant and time, Lipfert and Morris (2002, 019217)
14 employed a staged regression approach. In the first stage, a national model was developed for each
15 dependent variable, excluding air pollution variables. In the second stage, regressions were
16 performed with the residuals on concurrent and previous periods' air pollution variables to identify
17 the pollutants of interest. Many results were presented because of the large number of age groups,
18 lagged exposure time windows, and mortality study periods examined in the study; overall, the
19 results were similar to those presented in the ACS cohort studies (i.e., PM25 and SO42" were found
20 to be consistently and positively associated with mortality). Lipfert and Morris (2002, 019217)
21 generally found the strongest associations in the earlier time periods, and when mortality and air
22 quality were measured in different periods (e.g., mortality data 1995-1997 and CO data 1970-1974).
23 Also, consistent with the Lipfert et al. (2000, 012281) and the Pope et al. (2002, 024689) cohort
24 studies, CO was frequently negatively (and often significantly) associated with mortality in older age
25 groups, especially when mortality was matched with CO levels in more recent time periods. The
26 younger age group (15-44) often showed a positive association with CO, but considering the small
27 number of deaths attributed to this age group (less than 1% of total deaths), the association was not
28 informative. Overall, this study highlighted that the CO-mortality associations presented in purely
29 ecologic study designs are generally consistent with those found in semi-individual cohort studies.
5.6.2.3. Summary of Mortality and Long-Term Exposure to CO
30 The evaluation of new epidemiologic studies conducted since the 2000 CO AQCD (U.S. EPA,
31 2000, 000907) that investigated the association between long-term exposure to CO and mortality
32 consistently found null or negative mortality risk estimates. No such studies were discussed in the
33 2000 CO AQCD (U.S. EPA, 2000, 000907). The re-analysis of the ACS data (Pope et al., 1995,
34 045159) by Jerrett et al. (2003, 087380) found no association between long-term exposure to CO and
September 2009 5-165 DRAFT - DO NOT CITE OR QUOTE
-------
1 mortality. Similar results were obtained in an updated analysis of the ACS data (Pope et al, 2002,
2 024689) when using earlier (1980) CO data, but negative associations were found when using more
3 recent (1982-1998) data. These results were further confirmed in an extended analysis of the ACS
4 data (Krewski et al., 2009, 191193). The Women's Health Initiative (WHI) Study also found no
5 association between CO and CVD events (including mortality) using the mortality data from recent
6 years (1994-1998) (Miller et al., 2007, 090130). while the series of Veterans Cohort studies found no
7 association or a negative association between mean annual 95th percentile of hourly CO values and
8 mortality (Lipfert et al., 2006, 088218: Lipfert et al., 2006, 088756). An additional study was
9 identified that used a cross-sectional study design, Lipfert and Morris (2002, 019217). which
10 reported results for a study of U.S. counties that are generally consistent with the cohort studies:
11 positive associations between long-term exposure to PM2.5 and SO42" and mortality, and generally
12 negative associations with CO. Overall, the consistent null and negative associations observed across
13 epidemiologic studies which included cohort populations encompassing potentially susceptible
14 subpopulations (i.e., post-menopausal women and hypertensive men) combined with the lack of
15 evidence for respiratory and cardiovascular morbidity outcomes following long-term exposure to
16 CO; and the absence of a proposed mechanism to explain the progression to mortality following
17 long-term exposure to CO provide supportive evidence that there is not likely to be 3 Causal
is relationship between long-term exposure to CO and mortality.
5.7. Susceptible Populations
19 Interindividual variation in human responses to air pollutants indicates that some
20 subpopulations are at increased risk for adverse health effects resulting from ambient CO exposure.
21 The NAAQS are intended to provide an adequate margin of safety for both the general population
22 and populations potentially at increased risk for health effects due to ambient air pollution (See
23 Section 1.1). To facilitate the identification of populations at the greatest risk for CO-related health
24 effects, studies have evaluated factors that contribute to the susceptibility and/or vulnerability of an
25 individual to CO. These terms have sometimes been used interchangeably in the literature, and in
26 other cases have been defined to represent two different categories that could contribute to a
27 population experiencing increased risk to CO-related health effects, resulting in the lack of a clear
28 and consistent definition (see Table 5-25). Additionally, in some cases, "at-risk" has been used as a
29 term encompassing these concepts more generally. In this ISA, the term 'susceptibility' will be used
30 to represent populations that have a greater likelihood of experiencing effects related to ambient CO
31 exposure. This increased likelihood of response to CO can result from a multitude of factors,
32 including pre-existing disease states, gender, age, or lifestyle (e.g., visiting high-altitude location,
33 medication use).
September 2009 5-166 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-25 Definitions of susceptible and vulnerable in the CO literature.
Definition Reference
Susceptible: predisposed to develop a noninfectious disease
Merriam-Webster (2009,1921461
Vulnerable: capable of being hurt: susceptible to injury or disease
Susceptible: greater likelihood of an adverse outcome given a specific exposure, in comparison with the general
population. Includes both host and environmental factors (e.g., genetics, diet, physiologic state, age, gender, social,
economic, and geographic attributes). American Lung Association (2001, 0166261
Vulnerable: periods during an individual's life when they are more susceptible to environmental exposures.
Susceptible: those who are more likely to experience adverse effects of CO exposure than normal healthy adults ,,o ppA ,-..- 1CmQc\
(e.g., persons with cardiovascular disease, COPD, reduced or abnormal hemoglobin, older adults, neonates). ( ' lajjaa)
Susceptible: greater or lesser biological response to exposure.
U.S. EPA (2009, 1921491
Vulnerable: more or less exposed.
Vulnerable: to be susceptible to harm or neglect, that is, acts of commission or omission on the part of others that . , , . ,onnl 1Q01Cm
can wound. Aaay' LA'(zum' m-m>
Susceptible:may be those who are significantly more liable than the general population to be affected by a stressor
due to life stage (e.g., children, the elderly, or pregnant women), genetic polymorphisms (e.g., the small but
significant percentage of the population who have genetic susceptibilities), prior immune reactions (e.g., individuals
who have been "sensitized" to a particular chemical), disease state (e.g., asthmatics), or prior damage to cells or u s EPA /2go3 192145)
systems (e.g., individuals with damaged ear structures due to prior exposure to toluene, making them more sensitive
to damage by high noise levels).
Vulnerable: differential exposure and differential preparedness (e.g., immunization).
Susceptible: intrinsic (e.g., age, gender, pre-existing disease (e.g., asthma) and genetics) and extrinsic (previous K|pphprnpr and Ohteuka mm 1T>48Q1
exposure and nutritional status) factors. weeoerger ana untsura (^uus, jdUfiBy)
1 To examine whether CO differentially affects certain subpopulations, epidemiologic studies
2 conduct stratified analyses to identify the presence or absence of effect modification. These analyses
3 require the proper identification of confounders and their subsequent adjustment in statistical
4 models, which helps separate a spurious association from a true causal association. In experimental
5 research, the study of individuals with underlying disease and the use of animal models of disease
6 allow for comparisons between subgroups. Therefore, the results from these studies, combined with
7 results obtained through stratified analyses of comparison groups in epidemiologic studies,
8 contribute to the overall weight of evidence for the increased susceptibility of specific populations to
9 CO. The following section discusses the epidemiologic, controlled human exposure, and
10 toxicological studies evaluated in previous sections of Chapter 5 that provide information on
11 potentially susceptible populations.
5.7.1.Pre-Existing Disease
12 The 2000 CO AQCD identified certain subpopulations within the general population that may
13 be more susceptible to the effects of CO exposure, including individuals (particularly older adults)
14 with CHD and other vascular diseases, anemia, or COPD. As discussed in the 2000 CO AQCD and
15 reviewed in Section 4.5 of this assessment, diseases which cause inflammation and systemic stress
16 are known to increase endogenous CO production, potentially putting individuals with such
September 2009 5-167 DRAFT - DO NOT CITE OR QUOTE
-------
1 conditions at additional risk from ambient CO exposure. The critical level of COHb leading to
2 adverse health effects varies depending on health outcome and disease state of individuals. The
3 following sections summarize the evidence presented in the 2000 CO AQCD along with the new
4 evidence for the potential increased susceptibility of individuals with various pre-existing diseases to
5 CO-induced health effects.
5.7.1.1. Cardiovascular Disease
6 Controlled exposures to CO resulting in COHb concentrations of 2-6% have been shown to
7 affect cardiovascular function among individuals with coronary artery disease (CAD). Several
8 studies have reported significant decreases in the time to onset of exercise-induced angina or ST-
9 segment changes following CO exposure in patients with stable angina. In the largest such study,
10 COHb concentrations as low as 2.0-2.4% were observed to significantly decrease the time required
11 to induce ST-segment changes indicating myocardial ischemia (p = 0.01) (see 5.2.4). In addition to
12 the effects of CO on myocardial ischemia, there is some evidence to suggest that CO may provoke
13 cardiac arrhythmia in patients with CAD; however, this has not been observed at COHb
14 concentrations below 6%. While healthy adults have been shown to experience a decrease in
15 exercise performance following or during exposure to CO, no changes in cardiac rhythm or ECG
16 parameters have been demonstrated.
17 Evidence of CO-induced health effects in individuals with CAD is coherent with results of
18 epidemiologic studies that examined the effect of pre-existing cardiovascular conditions through
19 either secondary diagnoses or underlying comorbidities on associations between CO and emergency
20 department (ED) visits and hospital admissions (HAs). Mann et al. (2002, 036723) found increased
21 associations between CO and HAs for IHD in individuals with secondary diagnoses of either CHF or
22 arrhythmia in southern California. Peel et al. (2007, 090442) also examined the effect of underlying
23 cardiovascular conditions on cardiovascular-related HAs in response to short-term exposure to air
24 pollutants including CO in Atlanta, GA. Individuals with underlying dysrhythmia were found to have
25 increased HAs for IHD, but unlike Mann et al. (2002, 036723) underlying CHF was not found to
26 increase IHD HAs. Peel et al. (2007, 090442) also examined other underlying conditions and found
27 increased HAs for a variety of cardiovascular effects including: dysrhythmia, PVCD, and CHF in
28 individuals with underlying hypertension; dysrhythmia and PVCD in individuals with underlying
29 CHF, and PVCD in indiviudals with underlying dysrhythmia. Although a clear pattern of
30 associations is not evident across the epidemiologic studies evaluated, the available evidence
31 suggests that underlying dysrhythmia increases IHD HAs in response to short-term exposure to CO.
32 The combined evidence from controlled human exposure and epidemiologic studies provides
33 coherence and biological plausibility for the association between CO and cardiovascular morbidity
34 in individuals with CAD, particularly those with IHD. Approximately 13.7 million people in the U.S.
September 2009 5-168 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
have been diagnosed with CAD (also known as CHD), some fraction of whom have IHD (see Table
5-26). These individuals therefore represent a large population that may be more susceptible to
ambient CO exposure than the general population. Additional evidence for increased CO-induced
cardiovascular effects is provided by toxicological studies that used animal models of cardiovascular
disease. Short-term exposure to 50 ppm CO exacerbated cardiomyopathy and vascular remodeling
related to pulmonary hypertension (Carraway et al., 2002, 026018; Gautier et al., 2007, 096471;
Melin et al., 2002, 037502; 2005, 193833). Although the population at risk of primary pulmonary
hypertension is low, secondary pulmonary hypertension is a frequent complication of COPD and
certain forms of heart failure. These studies demonstrate the potential for short-term exposure to CO
to adversely affect indivduals with underlying cardiovascular conditions.
Table 5-26 Percent of the U.S. population in 2007 with respiratory diseases and cardiovascular
diseases.
Chronic Condition/ Disease
Age
Adults (18+) 18-44 45-64
Number (x106) % % %
Regional
65-74 75+ NE MW S W
% % % % % %
COPD1
Chronic bronchitis
Emphysema
7.6 3.4 2.3 4.2
3.7 1.6 0.2 2.3
5.5 4.8 2.8 3.2 4.0 2.9
4.5 5.2 1.1 1.8 1.8 1.6
CARDIOVASCULAR DISEASES2
All heart disease3
Coronary heart disease4
Hypertension
Stroke
25.1 11.2 4.1 12.2
13.7 6.1 0.9 6.7
52.9 23.2 8.2 32.1
5.4 2.4 0.3 2.8
27.1 35.8 10.6 12.3 11.3 10.2
18.6 23.6 5.3 6.7 6.4 5.5
50.9 57.4 21.3 23.4 25.1 21.0
6.3 10.6 2.2 2.3 2.7 2.2
1 Respondents were asked if they had ever been told by a doctor or other health professional that they had emphysema. In a separate question, respondents were asked if they had been told by a doctor or
other health professional in the last 12 months that they had bronchitis. A person maybe represented in more than one row.
2 In separate questions, respondents were asked if they had ever been told by a doctor or other health professional that they had: hypertension for high blood pressure), coronary heart disease, angina (or
angina pectoris), heart attack (or myocardial infarction), any other heart condition or disease not already mentioned, or a stroke. A person may be represented in more than one row.
3 Heart disease includes coronary heart disease, angina pectoris, heart attack, or any other heart condition or disease.
" Coronary heart disease includes coronary heart disease, angina pectoris, or heart attack.
Source: National Center for Health Statistics, Summary Health Statistics for U.S. Adults: National Health Interview Survey, 2007, Tables 1 &2.
5.7.1.2. Obstructive Lung Disease
COPD is a progressive disease resulting in decreased air flow to the lungs, and is especially
prevalent among smokers. The national prevalence of chronic bronchitis and emphysema, the two
main forms of COPD, was estimated to be 7.6 million and 3.7 million people in 2007, respectively
(see Table 5-26), although there could be overlap among these two subpopulations. The 2000 CO
AQCD identified individuals with obstructive lung diseases, such as COPD, as a susceptible
population due to a majority of COPD patients having exercise limitations as demonstrated by a
decrease in O2 saturation during mild to moderate exercise. This may heighten the sensitivity of
September 2009
5-169
DRAFT - DO NOT CITE OR QUOTE
-------
1 these individuals to CO during submaximal exercise typical of normal daily activity. COPD patients
2 who are smokers may have baseline COHb levels of 4-8% (U.S. EPA, 2000, 000907). increasing
3 their susceptibility to additional increases in COHb resulting from ambient exposure. COPD is often
4 accompanied by a number of changes in gas exchange, including increased VD and VA/Q inequality
5 (Marthan et al, 1985, 086334). which could slow both CO uptake and elimination.
6 A controlled human exposure study, which consisted of individuals with COPD (Bathoorn et
7 al., 2007, 193963). found that two of the patients experienced COPD exacerbation during or
8 following CO exposure at 100-125 ppm for 2 h, although a slight anti-inflammatory effect was also
9 observed. The few epidemiologic studies that evaluated the relationship between ambient CO and
10 increased hospital admissions or ED visits for COPD show weak positive associations. For example,
11 Peel et al. (2007, 090442) found that associations between short-term CO exposure and hospital
12 admissions for PVCD or CHF were increased in individuals with secondary diagnoses of COPD.
13 However, underlying COPD was not associated wth increased IHD and dysrhythmia HAs. Although
14 the majority of the evidence for CO-induced effects comes from studies that focus on individuals
15 with COPD, epidemiologic studies also report weak positive associations for asthmatics, who can
16 also experience exercise-induced airflow limitation.
17 As described in Section 5.7.1.3, evidence from animal toxicological studies indicates CO-
18 induced exacerbation of vascular remodeling related to pulmonary hypertension; secondary
19 pulmonary hypertension is a frequent complication of COPD. Preliminary evidence is also available
20 for CO-induced pulmonary inflammation, which is important for exacerbation of COPD and asthma,
21 from a recent animal toxicological study that indicated mild pulmonary inflammation in response to
22 50 ppm CO (Ohio et al., 2008, 096321).
23 Taken together, the results from epidemiologic, controlled human exposure, and toxicological
24 studies provide preliminary evidence which suggests that individuals with obstructive lung disease
25 (e.g., COPD patients with underlying hypoxia, asthmatics) may be susceptible to CO exposure.
26 Overall individuals with obstructive lung disease represent approximately 5% of the U.S. population,
27 and, therefore, represent a rather large population that is potentially susceptible to increased health
28 effects due to ambient CO exposure.
5.7.1.3. Anemia
29 As discussed in the 2000 CO AQCD, conditions such as anemia that alter the blood O2
30 carrying capacity or content will result in a greater risk from COHb induced hypoxia. Anemias are a
31 group of diseases that lower hematocrit and result in insufficient blood O2 or hypoxia due to Hb
32 deficiency through hemolysis, hemorrhage, or reduced hematopoiesis. Anemia may result from
33 pathologic conditions characterized by chronic inflammation such as malignant tumors or chronic
34 infections (Cavallin-Stahl et al., 1976, 086306; Cavallin-Stahl et al., 1976, 193239). The
September 2009 5-170 DRAFT - DO NOT CITE OR QUOTE
-------
1 cardiovascular system of people with anemia compensate for the reduction in O2 carrying capacity
2 by increasing cardiac output as both heart rate and stroke volume increase. One of the most prevalent
3 forms of anemia arises from a single-point mutation in the Hb gene, resulting in sickle cell diseases.
4 The affinity of Hb for O2 and its O2 carrying capacity is reduced causing a shift to the right in the O2
5 dissociation curve. It is well documented that African-American populations have a higher incidence
6 of sickle cell anemia, which may be a risk factor for CO hypoxia. Overall, lowered hematocrit due to
7 anemia will result in increased susceptibility and a greater response to inhalation of ambient CO. No
8 controlled human exposure or epidemiologic studies were identified that specifically investigated the
9 effect of anemia on health effects due to CO exposure.
10 Anemia may also increase the susceptibility of an individual to CO exposure through the
11 increased production of endogenous CO as a result of the disturbance of RBC hemostasis by
12 accelerated destruction of hemoproteins. Pathologic conditions such as hemolytic anemias,
13 hematomas, thalassemia, Gilbert's syndrome with hemolysis, and other hematological diseases and
14 illness will accelerate endogenous CO production (Berk et al., 1974, 012386; Hampson and Weaver,
15 2007, 190272: Meyer et al., 1998, 047530: Solanki et al., 1988, 012426: Sylvester et al., 2005,
16 191954). Patients with hemolytic anemia exhibit COHb at least levels 2- to 3-fold higher than
17 healthy individuals and CO production rates 2- to 8-fold higher (Coburn et al., 1966, 010984).
18 Recent studies report elevated COHb levels of 4.6-9.7% due to drug-induced hemolytic anemia
19 (Hampson and Weaver, 2007, 190272) and between 3.9% and 6.7% due to an unstable hemoglobin
20 disorder (Hb Zurich) (Zinkham et al., 1980, 011435). Taken together, this evidence indicates that
21 individuals with anemia are a potentially susceptible population for increased health effects due to
22 ambient CO as a result of their diminished O2-carrying capacity or high baseline COHb levels.
5.7.1.4. Diabetes
23 Exhaled CO concentrations are elevated in individuals with diabetes and are correlated with
24 blood glucose levels and duration of disease, indicating increased endogenous CO production (see
25 Section 4.5). Diabetics have been observed to be at increased risk for ED visits and hospital
26 admissions for heart diseases compared to non-diabetics in response to short-term exposure to CO
27 (Filho et al., 2008, 190260: Zanobetti and Schwartz, 2001, 016710). Peel et al. (2007, 090442) also
28 observed an increase in cardiovascular-related ED visits in individuals with diabetes but only for
29 dysrhythmias or PVCD, not IHD or CHF ED visits. Although no evidence was identified from
30 controlled human exposure or toxicological studies regarding CO exposure and diabetes, vascular
31 dysfunction was demonstrated in an animal model of metabolic syndrome and was attributed to
32 increased endogenous CO production (Johnson et al., 2006, 193874). Thus, increased endogenous
33 CO production in diabetics combined with the limited epidemiologic evidence suggests that
34 diabetics are potentially susceptible to short-term exposure to CO.
September 2009 5-171 DRAFT - DO NOT CITE OR QUOTE
-------
5.7.2. Lifestage
1 Age alters the variables that influence the uptake, distribution, and elimination of CO (see
2 Section 4.4.3). COHb levels decline more rapidly in young children than adults after CO exposure
3 (Joumard et al, 1981, 011330: Klasner et al., 1998, 087196). After infancy, the COHb half-life
4 increases with age, practically doubling between the ages of 2 and 70 (Joumard et al., 1981, 011330).
5 However, it should be noted that the rate of this reduction in CO elimination is very rapid in the
6 growing years (2-16 yr of age), but slows beyond adolescence. Increases in alveolar volume and
7 DLCO were observed with increasing body length of infants and toddlers (Castillo et al., 2006,
8 193234); these changes suggest faster CO uptake due to more advanced lung development. After
9 infancy, increasing age decreases DLCO and increases VA/Q mismatch, resulting in a longer duration
10 for both loading and elimination of CO from the blood (Neas and Schwartz, 1996, 079363).
5.7.2.1. OlderAdults
11 The 2000 CO AQCD noted that changes in metabolism that occur with age, particularly
12 declining maximal oxygen uptake, may make the aging population susceptible to the effects of CO
13 via impaired oxygen delivery to the tissues. Several epidemiologic studies compared cardiovascular
14 outcomes in older and younger adults, although no such studies were conducted in the U.S. In a
15 study in Australia and New Zealand, Barnett et al. (2006, 089770) found an increase in IHD and MI
16 HAs among individuals >65 yr of age compared with individuals aged 15-64 in response to short-
17 term exposure to CO. Lee et al. (2003, 095552) also found an association with IHD hospital
18 admissions in Seoul, Korea for individuals >65 yr of age, but not when all individuals were included
19 in the analysis. Lanki et al. (2006, 089788) found an association with hospital admissions for non-
20 fatal MI in a multicity European study among those aged >75 yr, but not for those <75 yr old. In
21 contrast, D'Ippoliti et al. (2003, 074311) observed higher associations for MI hospital admissions in
22 Rome among 18-64 year olds than among either 65-74 year olds or those 75 yr and over.
23 Szyszkowicz (2007, 193793) found slightly lower associations for IHD hospital admissions among
24 those >64 yr of age than for the all-age group. No controlled-human exposure studies or
25 toxicological studies were identified that compared CO effects among older and younger adults or
26 animal models of senescence, respectively. Overall, the epidemiologic studies evaluated provide
27 limited evidence that older adults may be susceptible to CO exposure. It should be noted that this
28 population also has a much higher prevalence of CAD than the general population; 18.6% of adults
29 aged 65-74 and 23.6% of adults age 75 and over reported having CHD, as compared with 6.1% of
30 the population as a whole, which may also contribute to any increase observed in CO-induced
31 cardiovascular effects. Both the higher prevalence of CAD and the gradual decline in physiological
32 processes associated with aging (U.S. EPA, 2006, 192082) may contribute to increased health effects
September 2009 5-172 DRAFT - DO NOT CITE OR QUOTE
-------
1 in response to CO in this population. Older adults represent a large and growing fraction of the U.S.
2 population, from 12.4% or 35 million people in 2000 to a projected 19.3% or 72 million people in
3 2030 (U.S. Census, 2000, 157064). and, as a result, are a large, potentially susceptible population for
4 CO-induced health effects.
5.7.2.2. Gestational Development
5 CO inhaled by pregnant animals quickly crosses the placental barriers and enters fetal
6 circulation. Effects of ambient CO may be increased during gestation because fetal CO
7 pharmacokinetics do not follow the same kinetics as maternal CO exposure; which contributes to the
8 difficulty in estimating fetal COHb based on maternal levels. Human fetal Hb has a higher affinity
9 for CO than adult Hb (Di Cera et al, 1989, 193998). Maternal and fetal COHb concentrations have
10 been modeled as a function of time using a modified CFK equation (Hill et al., 1977, 011315). At
11 steady-state conditions, fetal COHb has been found to be 10-15% higher than maternal COHb levels.
12 For example, exposure to 30 ppm CO results in a steady-state maternal COHb of 5% and a fetal
13 COHb of 5.75%. Fetal CO uptake lags behind maternal uptake for the first few hours, but later may
14 overtake the maternal values. Similarly, during washout, fetal COHb levels are maintained for
15 longer, with a half-life of around 7.5 h versus the maternal half-life of around 4 h (Longo and Hill,
16 1977, 010802). In addition, maternal endogenous CO production increases during pregnancy (0.92
17 mL/h) due to contributions from fetal endogenous CO production (0.036 mL/h) and altered
18 hemoglobin metabolism (Hill et al., 1977, 011315; Longo, 1970, 013922).
19 Epidemiologic studies provide limited evidence that in utero CO exposure is associated with
20 changes in various birth outcomes (see Section 5.4.1). CO exposure during early pregnancy was
21 associated with an increased risk of PTB. In the two studies that examined associations between CO
22 and birth defects, maternal CO exposure was associated with an increased risk of cardiac birth
23 defects, which is also coherent with evidence in Section 5.2 identifying the heart as a target organ for
24 CO. There is evidence for small reductions in birth weight (10-20 g) associated with CO exposure,
25 generally in the first or third trimester, although the decrease does not generally translate to an
26 increased risk of LEW or SGA. It is therefore difficult to conclude if CO is related to a small change
27 in birth weight across all births or a marked effect in some subset of births. There is limited evidence
28 that prenatal CO exposure is associated with an increased risk of infant mortality in the post-neonatal
29 period.
30 Toxicological studies lend biological plausibility to outcomes observed in epidemiologic
31 studies (see Section 5.4.2). Associations have been observed between CO exposure in laboratory
32 animals and decrements in birth weight as well as reduced prenatal growth. Animal toxicological
33 studies also provide evidence for effects on the heart, including transient cardiomegaly at birth after
34 continuous in utero CO exposure and delayed myocardial electrophysiological maturation. Evidence
September 2009 5-173 DRAFT - DO NOT CITE OR QUOTE
-------
1 exists for additional developmental outcomes which have been examined in toxicological studies,
2 but not epidemiologic or human clinical studies, including behavioral abnormalities, learning and
3 memory deficits, locomotor effects, neurotransmitter changes, and changes in the auditory system.
4 Furthermore, exogenous CO may interact or disrupt the normal physiological roles that endogenous
5 CO plays in the body. There is evidence that CO plays a role in maintaining pregnancy, controlling
6 vascular tone, regulating hormone balance, and sustaining normal follicular maturation.
7 Outcomes evaluated in epidemiologic studies affect a substantial portion of the U.S.
8 population. PTB and LEW have been established as strong predictors of infant mortality and
9 morbidity (Barker et al., 2002, 193960: Berkowitz and Papiernik, 1993, 055466: Li et al., 2003,
10 193965: Mclntire et al., 1999, 015310). In 2004, 36.5 percent of all infant deaths in the U.S. were
11 preterm-related (MacDorman et al., 2007, 193973). Vital statistics for the year 2005 in the U.S.
12 showed that the rate for PTB was 12.7%, which has risen 20% since 1990, and the rate for LBW was
13 8.2%, which has risen 17% since 1990 (Martin et al., 2007, 193982). Data from the Metropolitan
14 Atlanta Congenital Defects Program (MACDP), which is one of the most comprehensive birth defect
15 registries in the U.S., showed that the prevalence of congenital heart defects had increased between
16 1968 and 1997. During 1995-1997 the rate was 90.2 per 10,000 births (0.9%) and this had increased
17 from 58.7 per 10,000 births since 1986-1972 (Botto et al., 2001, 192379). Cardiovascular defects are
18 the single largest contributor to infant mortality attributable to birth defects (CDC, 1998, 193243).
19 Between 1979 and 1997, 1 in 10 infant deaths (9.8%) was associated with a congenital heart defect,
20 and 1 in 13 infant deaths (7.4%) was due to a congenital heart defect (Boneva et al., 2001, 193972).
21 The combined evidence from epidemiologic and toxicological studies, along with the increasing
22 prevalence of PTB, LBW, and cardiac birth defects in the U.S. population, indicates that critical
23 developmental phases may be characterized by enhanced sensitivity to CO exposure.
5.7.3.Gender
24 COHb concentrations are generally higher in male subjects than in female subjects (Horvath et
25 al., 1988, 012725). In addition, the COHb half-life is longer in healthy men than in women of the
26 same age, which may be partially explained by differences in muscle mass or the slight correlation
27 between COHb half-life and increased height (Joumard et al., 1981, 011330). The rate of decline of
28 DLCO with age is lower in middle-aged women than in men; however, it is similar in older adults
29 (Neas and Schwartz, 1996, 079363). This is supported by the fact that women tend to be more
30 resistant to altitude hypoxia (Horvath et al., 1988, 012725). Women also experience fluctuating
31 COHb levels through the menstrual cycle when endogenous CO production doubles in the
32 progesterone phase (0.62 mL/h versus 0.32 mL/h in estrogen phase) (Delivoria-Papadopoulos et al.,
33 1974, 086316: Mercke and Lundh, 1976, 086309). Similarly, endogenous CO production increases
34 during pregnancy due to contributions from fetal CO production and altered hemoglobin metabolism
September 2009 5-174 DRAFT - DO NOT CITE OR QUOTE
-------
1 as described above. In an epidemiologic study investigating the association between short-term CO
2 exposure and IHD hospital admissions (Szyszkowicz, 2007, 193793). males had higher associations
3 than females in both the all-ages group and in those >64 yr of age. The limited epidemiologic
4 evidence, combined with the gender-related differences in endogenous CO production, contributes to
5 the inability to conclude whether CO disproportionately affects males or females.
5.7.4.Altitude
6 Higher altitude results in changes in a number of factors that contribute to the uptake and
7 elimination of CO. The relationship between altitude and CO exposure has been discussed in depth
8 in the 2000 CO AQCD and other documents (U.S. EPA, 1978, 086321) and is reviewed in
9 Section 4.4.2 of this ISA. In an effort to maintain proper O2 transport and supply, physiological
10 changes occur as compensatory mechanisms to combat the decreased barometric pressure and
11 resulting altitude-induced hypobaric hypoxia (HH). These changes, which include increases in BP
12 and cardiac output and redistribution of blood from skin to organs and from blood to extravascular
13 compartments, generally will favor increased CO uptake and COHb formation, as well as CO
14 elimination. It has been demonstrated that breathing CO (9 ppm) at rest at altitude produces higher
15 COHb compared to sea level (McGrath et al, 1993, 013865). whereas high altitude exposure in
16 combination with exercise causes a decrease in COHb levels versus similar exposure at sea level
17 (Horvath et al., 1988, 012725). This decrease could be a shift in CO storage or suppression of COHb
18 formation, or both. In a controlled human exposure study on the health effects of CO at altitude,
19 Kleinman et al. (1998, 047186) observed an additive effect of CO exposure and simulated high
20 altitude on the reduction in time to onset of angina among a group of individuals with CAD. No
21 epidemiologic studies were identified that specifically examined the effect of altitude on health
22 effects due to CO exposure.
23 Altitude also increases the baseline COHb levels by inducing endogenous CO production and
24 has been shown to be positively associated with baseline COHb concentrations (McGrath, 1992,
25 001005; McGrath et al., 1993, 013865). This increase in COHb with altitude-induced hypoxia has
26 also been associated with increases in mRNA, protein, and activity of HO-1 in rats and cells leading
27 to enhanced endogenous CO production (Carraway et al., 2002, 026018; Chin et al., 2007, 190601).
28 Early HH increased lung HO-1 protein and activity, whereas chronic HH induced endogenous CO
29 production in nonpulmonary sites (see Section 4.5) (Carraway et al., 2000, 021096). Whether other
30 variables (such as an accelerated metabolism or a greater pool of Hb, transient shifts in body stores,
31 or a change in the elimination rate of CO) play a role has not been explored.
32 As the length of stay increases at high altitude, acclimatization occurs, inducing
33 hyperventilation, polycythemia or increased red blood cell count, and increased tissue capillarity and
34 Mb content in skeletal muscle, which could also favor increased CO uptake. Most of the initial
September 2009 5-175 DRAFT - DO NOT CITE OR QUOTE
-------
1 adaptive changes gradually revert to sea level values. However, these adaptive changes persist in
2 people raised at high altitude even after reacclimatization to sea level (Hsia, 2002, 193857). This
3 evidence indicates that visitors to high altitude locations may represent a potentially susceptible
4 population for increased risk of health effects due to CO exposure.
5.7.5.Exercise
5 Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase
6 in gas exchange. O2 consumption can increase more than 10 fold during exercise. Similarly,
7 ventilation, membrane and lung diffusing capacity, pulmonary capillary blood volume, and cardiac
8 output increase proportional to work load. The majority of these changes facilitate CO uptake and
9 transport, by increasing gas exchange efficiency. Likewise, the COHb elimination rate increases with
10 physical activity, causing a decrease in COHb half-life (Joumard et al., 1981, 011330). In a
11 controlled human exposure study, healthy subjects exposed to CO and achieving COHb levels of
12 approximately 5% observed a significant decrement in exercise duration and maximal effort
13 capability (measured by metabolic equivalent units) (Adir et al., 1999, 001026). It is possible that
14 CO lowers the anaerobic threshold, allowing earlier fatigue of the skeletal muscles and decreased
15 maximal effort capability. Due to the counterbalancing effects of increased rates of COHb formation
16 and elimination, it is unclear whether individuals engaging in light to moderate exercise are a
17 potentially susceptible population for increased health effects due to ambient CO exposure.
5.7.6.Proximity to Roadways
18 Individuals that spend a substantial amount of time on or near heavily traveled roadways, such
19 as commuters and those living or working near freeways, are likely to be exposed to elevated CO
20 concentrations, as discussed in Chapter 1. CO concentrations measured at the roadside in research
21 studies are several-fold higher than concentrations measured a few hundred meters downwind
22 (Baldauf et al., 2008, 191017; Zhu et al., 2002, 041553). with the shape of the concentration profile
23 dependent on wind speed and direction. AQS monitoring data aggregated across multiple sites with
24 no adjustment for wind conditions show somewhat higher concentrations for microscale (near-road)
25 monitors relative to middle-scale monitors, although the ratio is lower than that observed in the
26 roadside studies. Elevated near-road concentrations are important for occupants of the estimated 17.9
27 million occupied homes nationwide (16.1%) that are within approximately 90 m of a freeway,
28 railroad, or airport, according to the 2007 American Housing Survey (2008, 194013)
29 Studies of commuters have shown that commuting time is an important determinant of CO
30 exposure for those traveling by car, bicycle, public transportation, and walking (Bruinen de Bruin et
31 al., 2004, 190943; Kaur et al., 2005, 086504; Scotto Di Marco et al., 2005, 144054). In-vehicle
September 2009 5-176 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations have been measured to be several times higher than concentrations measured at fixed-
2 site monitors not located adjacent to roadways (Bruinen de Bruin et al., 2004, 190943; Chang et al.,
3 2000, 001276: Kaur et al., 2005, 086504: Riediker et al., 2003, 043761: Scotto Di Marco et al., 2005,
4 144054). Commuting is likely to be an important contributor to CO exposure for the 5.5 million U.S.
5 worker (5%) who drive 60 min or more to work (U.S. Census Bureau, 2008, 194013).This evidence
6 for elevated on-road and near-road CO concentrations combined with residential and commuting
7 data indicates that the large numbers of individuals who spend a substantial amount of time on or
8 near heavily traveled roadways are an important potentially susceptible population for increased
9 health risks due to ambient CO exposure.
5.7.7.Medications and Other Substances
10 Endogenous CO production can be altered by medications or a number of physiological
11 conditions that increase RBC destruction, the breakdown of hemoproteins other than Hb, and the
12 production of bilirubin (see Section 4.5). Nicotinic acid, allyl-containing compounds (acetamids and
13 barbiturates), diphenylhydantoin, progesterone, contraceptives, and statins increase CO production.
14 One epidemiologic study (Dales, 2004, 099036) investigated the effect of medication use on the
15 relationship between ambient CO and HRV in individuals with CAD. The authors observed an
16 association between short-term CO exposure and an increase in SDNN for CAD patients not taking
17 beta blockers; however, this association did not persist in CAD patients taking beta blockers.
18 Compounds such as carbon disulfide and sulfur-containing chemicals (parathion and
19 phenylthiourea) increase CO following metabolism by cytochrome p450s. The P450 system may
20 also cause large increases in CO produced from the metabolic degradation of dihalomethanes leading
21 to very high (>10%) COHb levels which can be further enhanced by prior exposure to HCs or
22 ethanol. Minor sources of endogenous CO include the auto-oxidation of phenols, photo-oxidation of
23 organic compounds, and lipid peroxidation of cell membrane lipids. Taken together, this evidence
24 indicates that individuals ingesting medications and other substances that enhance endogenous or
25 metabolic CO production are a potentially susceptible population for increased health effects due to
26 additional exposure to ambient CO.
5.7.8.Summary of Susceptible Populations
27 Individuals with CAD represent the population most susceptible to increased risk of CO-
28 induced health effects, based on evidence of significant decreases in the time to onset of exercise-
29 induced angina or ST-segment changes observed in controlled human exposure studies of individuals
30 with CAD, along with coherent results from epidemiologic studies that observed associations
31 between short-term CO exposure and ED visits and HAs for IHD and related outcomes. Limited
September 2009 5-177 DRAFT - DO NOT CITE OR QUOTE
-------
1 evidence from stratified analyses in epidemiologic studies indicated that secondary diagnoses of
2 CHF or arrhythmia increased associations between short-term CO exposure and IHD HAs.
3 Additional evidence is provided by toxicological studies that demonstrated exacerbation of
4 cardiomyopathy and increased vascular remodeling in animal models of cardiovascular disease.
5 Although it is not clear whether the small changes in COHb associated with ambient CO exposures
6 result in substantially diminished O2 delivery to tissues, the known role of CO in limiting O2
7 availability lends a degree of biological plausibility to ischemia-related health outcomes following
8 CO exposure.
9 Potentially susceptible populations also include individuals with other pre-existing diseases,
10 such as COPD, anemia, or diabetes. Although the limited evidence available from controlled human
11 exposure, epidemiologic, and toxicological studies relating to respiratory and pulmonary health
12 effects contributes to uncertainty regarding the specific nature of CO-induced health effects in
13 individuals with COPD, those with underlying hypoxia may be a potentially susceptible population
14 for increased health effects due to ambient CO exposure. Individuals with various types of anemia
15 are a potentially susceptible population for increased health effects due to ambient CO as a result of
16 their diminished O2-carrying capacity or high baseline COHb levels. Increased endogenous CO
17 production in diabetics combined with limited epidemiologic evidence suggests that diabetics may
18 be potentially susceptible to health effects induced by short-term exposure to CO.
19 There is also evidence that older adults and the developing young represent potentially
20 susceptible population to CO-induced health effects. Epidemiologic studies provide limited evidence
21 from stratified analyses indicating that associations between short-term CO exposure and hospital
22 admissions for CAD are higher among those > 65 yr old than for those <65. The older adult
23 population also has a much higher prevalence of CAD than the population as a whole, which may
24 contribute to increased susceptibility. Recent studies on birth outcomes have provided limited
25 evidence of associations between in utero CO exposure and PTB, LEW and cardiac birth defects.
26 Toxicological studies provide evidence of effects on birth weight and growth as well as development
27 of the cardiovascular and nervous systems following prenatal exposure to CO. This evidence,
28 combined with differences between fetal and maternal CO pharmacokinetics, indicates that critical
29 developmental phases may be characterized by enhanced sensitivity to CO exposure.
30 Visitors to high altitude locations may represent a potentially susceptible population due to
31 changes in factors which affect the uptake and elimination of CO, although acclimatization occurs as
32 length of stay increases. Individuals with substantial exposure to mobile source emissions, such as
33 commuters and those living near heavily traveled roadways, represent an important subpopulation
34 potentially susceptible to increased risk of CO-induced health effects due to elevated on-road and
35 roadside CO concentrations.
September 2009 5-178 DRAFT - DO NOT CITE OR QUOTE
-------
1 Overall, the controlled human exposure, epidemiologic, and toxicological studies evaluated in
2 this assessment provide evidence for increased susceptibility among various populations. Medical
3 conditions that increase endogenous CO production rates may also contribute to increased
4 susceptibility to health effects from ambient CO exposure. The level and type of evidence varies
5 depending on the factor being evaluated, with the strongest evidence indicating that individuals with
6 CAD are most susceptible to an increase in CO-induced health effects.
5.8. Summary
7 The evidence reviewed in this chapter describes recent findings regarding the health effects of
8 ambient CO. Section 5.1 presents evidence on the mode of action of CO, including its role in
9 limiting O2 availability as well as its role in altered cell signaling. Evidence is presented in
10 subsequent sections on the effect of short- and long-term exposure to CO on cardiovascular
11 morbidity (Section 5.2), the central nervous system (Section 5.3), birth outcomes and developmental
12 effects (Section 5.4), respiratory morbidity (Section 5.5), and mortality (Section 5.6). Potentially
13 susceptible populations at increased risk of CO-induced health effects are discussed in Section 5.7.
14 Table 5-1 summarizes causal determinations for the health outcome categories reviewed in this
15 assessment. An integrative overview of the health effects of ambient CO and uncertainties associated
16 with interpretation of the evidence is provided in Chapter 2. The strongest evidence regarding CO-
17 induced health effects relates to cardiovascular morbidity, and the combined evidence from
18 controlled human exposure studies and epidemiologic studies indicates that a causal relationship is
19 likely to exist between relevant short-term CO exposures and cardiovascular morbidity, particularly
20 in individuals with CAD. The evidence is suggestive of a causal relationship between short-term
21 exposure to CO and respiratory morbidity as well as between short-term CO exposure and mortality.
22 The evidence is also suggestive of a causal relationship for birth outcomes and developmental effects
23 following long-term exposure to CO, and for central nervous system effects linked to short- and
24 long-term exposure to CO. The evidence indicates that there is not likely to be a causal relationship
25 between long-term exposure to CO and mortality. For respiratory morbidity following long-term
26 exposure to CO, the evidence was inadequate to infer a causal relationship.
September 2009 5-179 DRAFT - DO NOT CITE OR QUOTE
-------
References
Achouh PE; Simonet S; Fabiani JN; Verbeuren TJ. (2008). Carbon monoxide induces relaxation of human internal thoracic
and radial arterial grafts. Interact Cardiovasc Thorac Surg, 7: 959-963. 179918
Adams KF; Koch G; Chatterjee B; Goldstein GM; O'Neil J J; Bromberg PA; Sheps DS; McAllister S; Price CJ; Bissette J.
(1988). Acute elevation of blood carboxyhemoglobin to 6% impairs exercise performance and aggravates
symptoms in patients with ischemic heart disease. J Am Coll Cardiol, 12: 900-909. 012692
Aday LA. (2001). At risk in America: The health and health care needs of vulnerable populations in the United States. San
Francisco, CA: Jossey-Bass, Inc. 192150
Adir Y; Merdler A; Haim SB; Front A; Harduf R; Bitterman H. (1999). Effects of exposure to low concentrations of carbon
monoxide on exercise performance and myocardial perfusion in young healthy men. Occup Environ Med, 56: 535-
538. 001026
Alderman BW; Baron AE; Savitz DA. (1987). Maternal exposure to neighborhood carbon monoxide and risk of low infant
birth weight. Public Health Rep, 102: 410-414. 012243
Alexandreanu 1C; Lawson DM. (2002). Effects of chronic administration of a heme oxygenase substrate or inhibitor on
progression of the estrous cycle, pregnancy and lactation of Sprague-Dawley rats. Life Sci, 72: 153-162. 192373
Alexandreanu 1C; Lawson DM. (2003). Heme oxygenase in the rat anterior pituitary: Immunohistochemical localization
and possible role in gonadotropin and prolactin secretion. Exp Biol Med, 228: 64-69. 193871
Allred EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Hayes D; Pagano M; Selvester RH; Walden
SM; Warren J. (1989). Acute effects of carbon monoxide exposure on individuals with coronary artery disease.
Health Effects Institute. Boston, MA. 012697
Allred EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; Selvester RH; Walden SM;
Warren J. (1989). Short-term effects of carbon monoxide exposure on the exercise performance of subjects with
coronary artery disease. N Engl J Med, 321: 1426-1432. 013018
Allred EN; Bleecker ER; Chaitman BR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; Selvester RH; Walden SM;
Warren J. (1991). Effects of carbon monoxide on myocardial ischemia. Environ Health Perspect, 91: 89-132.
011871
Alonso JR; CardellachF; Lopez S; Casademont J; Miro O. (2003). Carbon monoxide specifically inhibits cytochrome c
oxidase of human mitochondrial respiratory chain. Pharmacol Toxicol, 93: 142-146. 193882
American Lung Association. (2001). Urban air pollution and health inequities: A workshop report. Environ Health
Perspect, 3: 357-374. 016626
Anderson EW; Andelman RJ; Strauch JM; Fortuin NJ; Knelson JH. (1973). Effect of low-level carbon monoxide exposure
on onset and duration of angina pectoris: a study in ten patients with ischemic heart disease. Ann Intern Med, 79:
46-50. 023134
Andresen JJ; Shafi NI; Durante W; Bryan RM Jr. (2006). Effects of carbon monoxide and heme oxygenase inhibitors in
cerebral vessels of rats and mice. Am J Physiol Heart Circ Physiol, 291: H223-H230. 180449
Antonelli T; Tomasini MC; Tattoli M; Cassano T; Finetti S; Mazzoni E; Trabace L; Carratu MR; Cuomo V; Tanganelli S;
Ferraro L. (2006). Prenatal exposure to the cannabinoid receptor agonist WIN 55,212-2 and carbon monoxide
reduces extracellular glutamate levels in primary rat cerebral cortex cell cultures. Neurochem Int, 49: 568-576.
193885
Arnedo-Pena A; Garcia-Marcos L; Carvajal Uruena I; Busquets Monge R; Morales Suarez-Varela M; Miner Canflanca I;
Batlles Garrido J; Blanco Quiros A; Lopez-Silvarrey Varela A; Garcia Hernandez G; Aguinaga Ontoso I; Gonzalez
Diaz C. (2009). Air pollution and recent symptoms of asthma, allergic rhinitis, and atopic eczema in schoolchildren
aged between 6 and 7 years. Arch Bronconeumol, 45: 224-229. 190238
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 5-180 DRAFT - DO NOT CITE OR QUOTE
-------
Astrup P; Olsen HM; Trolle D; Kjeldsen K. (1972). Effect of moderate carbon-monoxide exposure on fetal development.
Lancet, 7789: 1220-1222. 011121
Ayres SM; Giannelli S, Jr; Mueller H; Criscitiello A. (1973). Myocardial and systemic vascular responses to low
concentration of carboxyhemoglobin. AnnClinLab Sci, 3: 440-447. 193943
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM; Bertazzi PA;
Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb Haemost, 5: 252-260.
090733
Bainbridge SA; Farley AE; McLaughlin BE; Graham CH; Marks GS; Nakatsu K; Brien JF; Smith GN. (2002). Carbon
monoxide decreases perfusion pressure in isolated human placenta. Placenta, 23: 563-569. 043161
Baldauf R; Thoma E; Khly stov A; Isakov V; Bowker G; Long T; Snow R. (2008). Impacts of noise barriers on near-road air
quality. Atmos Environ, 42: 7502-7507. 191017
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB; Canada A;
Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air pollution and
cardiovascular admisisons association in Spain: results within the EMECAS project. J Epidemiol Community
Health, 60: 328-336. 088746
Ballester F; Tenias JM; Perez-Hoyos S. (2001). Air pollution and emergency hospital admissions for cardiovascular
diseases in Valencia, Spain. J Epidemiol Community Health, 55: 57-65. 013257
Barber A; Robson SC; Myatt L; Bulmer JN; Lyall F. (2001). Heme oxygenase expression in human placenta and placental
bed: reduced expression of placenta endothelial HO-2 in preeclampsia and fetal growth restriction. ,15: 1158-68.
193891
Barker DJ; Eriksson JG; Forsen T; Osmond C. (2002). Fetal origins of adult disease: Strength of effects and biological
basis. Int J Epidemiol, 31: 1235-1239. 193960
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW. (2006). The effects of air
pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities.
Environ Health Perspect, 114: 1018-1023. 089770
Bathoorn E; Slebos DJ; Postma DS; Koeter GH; van Oosterhout AJ; van der Toorn M; Boezen HM; Kerstjens HA. (2007).
Anti-inflammatory effects of inhaled carbon monoxide in patients with COPD: A pilot study. Eur Respir J, 30:
1131-1137. 193963
Beard RR; Wertheim GA. (1967). Behavioral impairment associated with small doses of carbon monoxide. Am J Public
Health, 57: 2012-2022. 011015
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and Massachusetts.
Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Peng RD; Dominici F; Samet JM. (2009). Emergency admissions for cardiovascular disease and ambient levels
of carbon monoxide: Results for 126 U.S. urban counties, 1999-2005. Circulation, xx: xxx-xxx. 193780
Benagiano V; Lorusso L; Coluccia A; Tarullo A; Place P; Girolamo F; Bosco L; Cagiano R; Ambrosi G. (2005). Glutamic
acid decarboxylase and GABA immunoreactivities in the cerebellar cortex of adult rat after prenatal exposure to a
low concentration of carbon monoxide. Neuroscience, 135: 897-905. 180445
Benagiano V; Lorusso L; Place P; Girolamo F; Rizzi A; Sabatini R; Auteri P; Bosco L; Cagiano R; Ambrosi G. (2007).
Effects of prenatal exposure to the CB-1 receptor agonist WIN 55212-2 or CO on the GABAergic neuronal systems
of rat cerebellar cortex. Neuroscience, 149: 592-601. 193892
Benignus VA. (1993). Importance of experimenter-blind procedure in neurotoxicology. Neurotoxicol Teratol, 15: 45-49.
013645
Benignus VA; Muller KE; Barton CN; Prah JD. (1987). Effect of low level carbon monoxide on compensatory tracking and
event monitoring. Neurotoxicol Teratol, 9: 227-234. 012250
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; WichmannHE; Peters A. (2006). Runs of ventricular
and supraventricular tachycardia triggered by air pollution in patients with coronary heart disease. J Occup Environ
Med, 48: 1149-58. 098702
September 2009 5-181 DRAFT - DO NOT CITE OR QUOTE
-------
Berk PD; Rodkey FL; Blaschke TF; Collison HA; Waggoner JG. (1974). Comparison of plasma bilirubin turnover and
carbon monoxide production in man. J Lab Clin Med, 83: 29-37. 012386
Berkowitz GS; Papiernik E. (1993). Epidemiology of preterm birth. Epidemiol Rev, 15: 414-443. 055466
Bernard N; Saintot M; Astre C; Gerber M. (1998). Personal exposure to nitrogen dioxide pollution and effect on plasma
antioxidants. J Lab Clin Med, 53: 122-128. 086427
Biggeri A; Baccini M; Bellini P; Terracini B. (2005). Meta-analy sis of the Italian studies of short-term effects of air
pollution (MISA), 1990-1999. Int J Occup Environ Health, 11: 107-122. 087395
Bing O; Grundemar L; Ny L; Moller C; Heilig M. (1995). Modulation of carbon monoxide production and enhanced
spatial learning by tin protoporphyrin. Neuroreport, 6: 1369-1372. 079418
Boneva RS; Botto LD; Moore CA; Yang QH; Correa A; Erickson JD. (2001). Mortality associated with congenital heart
defects in the United States: Trends and racial disparities, 1979-1997. Circulation, 103: 2376-2381. 193972
Botto LD; Correa A; Erickson JD. (2001). Racial and temporal variations in the prevalence of heart defects. Pediatrics,
107: E32. 192379
Brian JEJr; HeistadDD; Faraci FM. (1994). Effect of carbon monoxide on rabbit cerebral arteries. Stroke, 25: 639-644.
076283
Briet M; Collin C; Laurent S; Tan A; Azizi M; Agharazii M; Jeunemaitre X; Alhenc-Gelas F; Boutouyrie P. (2007).
Endothelial function and chronic exposure to air pollution in normal male subjects. Hypertension, 50: 970-976.
093049
Brown SD; Piantadosi CA. (1992). Recovery of energy metabolism in rat brain after carbon monoxide hypoxia. J Clin
Invest, 89: 666-672. 013441
Bruinen de Bruin Y; Carrer P; Jantunen M; Hanninen O; Di Marco GS; Kephalopoulos S; Cavallo D; Maroni M. (2004).
Personal carbon monoxide exposure levels: contribution of local sources to exposures and microenvironment
concentrations in Milan. JExpo Sci Environ Epidemiol, 14: 312-322. 190943
Brilne B; Ullrich V. (1987). Inhibition of platelet aggregation by carbon monoxide is mediated by activation of guanylate
cyclase. Mol Pharmacol, 32: 497-504. 016535
Burnett RT; Brook J; Dann T; Delocla C; Philips O; Cakmak S; Vincent R; Goldberg MS; Krewski D. (2000). Association
between particulate- and gas-phase components of urban air pollution and daily mortality in eight Canadian cities.
Inhal Toxicol, 12: 15-39. 010273
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. Health
Effects Institute. Boston, MA. 042798
Burnett RT; Smith-Doiron M; Stieb D; Raizenne ME; Brook JR; Dales RE; Leech JA; Cakmak S; Krewski D. (2001).
Association between ozone and hospitalization for acute respiratory diseases in children less than 2 years of age.
Am J Epidemiol, 153: 444-452. 093439
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Health, 59: 228-36. 188612
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59: 228-236.
086247
Bye A; S0rhaug S; Ceci M; H0ydal MA; St01en T; Heinrich G; Tj0nna AE; Najjar SM; Nilsen OG; Catalucci D; Grimaldi
S; Contu R; Steinshamn S; Condorelli G; Smith GL; Ellingsen O; Waldum H; WisMf U. (2008). Carbon monoxide
levels experienced by heavy smokers impair aerobic capacity and cardiac contractility and induce pathological
hypertrophy. Inhal Toxicol, 20: 635-646. 193777
CDC. (1998). Trends in infant mortality attributable to birth defects-United States, 1980-1995. MMWR Morb Mortal
Wkly Rep, 47: 773-778. 193243
Cagiano R; Ancona D; Cassano T; Tattoli M; Trabace L; Cuomo V (1998). Effects of prenatal exposure to low
concentrations of carbon monoxide on sexual behaviour and mesolimbic dopaminergic function in rat offspring. Br
J Pharmacol, 125: 909-915. 087170
September 2009 5-182 DRAFT - DO NOT CITE OR QUOTE
-------
Cakmak S; Dales RE; Judek S. (2006). Respiratory health effects of air pollution gases: modification by education and
income. Arch Environ Occup Health, 61: 5-10. 093272
Cardell L-O; Ueki IF; Stjarne P; Agusti C; Takeyama K; Linden A; Nadel JA. (1998). Bronchodilatation in vivo by carbon
monoxide, a cyclic GMP related messenger. Br J Pharmacol, 124: 1065-1068. 086700
Cardell LO; Lou YP; Takeyama K; Ueki IF; Lausier J; Nadel JA. (1998). Carbon monoxide, a cyclic GMP-related
messenger, involved in hypoxic bronchodilation in vivo. Pulm Pharmacol Ther, 11: 309-315. 011534
Carmines E; RajendranN. (2008). Evidence for carbon monoxide as the major factor contributing to lower fetal weights in
rats exposed to cigarette smoke. Toxicol Sci, 102: 383. 188440
Carratu MR; Cagiano R; Desantis S; Labate M; Tattoli M; Trabace L; Cuomo V. (2000). Prenatal exposure to low levels of
carbon monoxide alters sciatic nerve myelination in rat offspring. Life Sci, 67: 1759-1772. 015839
Carratu MR; Cagiano R; Tattoli M; Trabace L; Borracci P; Cuomo V. (2000). Prenatal exposure model simulating CO
inhalation in human cigarette smokers: sphingomyelin alterations in the rat sciatic nerve. Toxicol Lett, 117: 101-
106.015935
Carratu MR; Renna G; Giustino A; De Salvia MA; Cuomo V. (1993). Changes in peripheral nervous system activity
produced in rats by prenatal exposure to carbon monoxide. Arch Toxicol, 67: 297-301. 013812
Carraway MS; Ghio AJ; Carter JD; Piantadosi CA. (2000). Expression of heme oxygenase-1 in the lung in chronic
hypoxia. Am J Physiol, 278: L806-L812. 021096
Carraway MS; Ghio AJ; Suliman HB; Carter JD; Whorton AR; Piantadosi CA. (2002). Carbon monoxide promotes
hypoxic pulmonary vascular remodeling. Am J Physiol, 282: L693-L702. 026018
Castillo A; Llapur CJ; Martinez T; Kisling J; Williams-Nkomo T; Coates C; Tepper RS. (2006). Measurement of single
breath-hold carbon monoxide diffusing capacity in healthy infants and toddlers. Pediatr Pulmonol, 41: 544-550 .
193234
Cavallin-Stahl E; Mercke C; Lundh B. (1976). Carbon monoxide production in patients with breast carcinoma. Br J
Haematol, 32: 177-182. 086306
Cavallin-Stahl E; Mercke C; Lundh B. (1976). Erythropoiesis and carbon monoxide production in Hodgkin's disease. Br J
Haematol, 32: 167-175. 193239
Cella M; Farina MG; Sarmiento MIK; Chianelli M; Rosenstein RE; Franchi AM. (2006). Heme oxygenase-carbon
monoxide (HO-CO) system in rat uterus : Effect of sexual steroids and prostaglandins. J Steroid Biochem Mol Biol,
99: 59-66 . 193240
Chan C-C; Chuang K-J; Chien L-C; Chen W-J; Chang W-T. (2006). Urban air pollution and emergency admissions for
cerebrovascular diseases in Taipei, Taiwan. Eur Heart J, 27: 1238-1244. 090193
Chan C-C; Chuang K-J; Su T-C; Lin L-Y. (2005). Association between nitrogen dioxide and heart rate variability in a
susceptible population. Eur J Cardiovasc Prev Rehabil, 12: 580-586. 088988
Chang C-C; Tsai S-S; Ho S-C; Yang C-Y (2005). Air pollution and hospital admissions for cardiovascular disease in
Taipei, Taiwan. EnvironRes, 98: 114-119. 080086
Chang L-T; Koutrakis P; Catalano PJ; Suh HH. (2000). Hourly personal exposures to fine particles and gaseous pollutants-
results from Baltimore, Maryland. J Air Waste Manag Assoc, 50: 1223-1235. 001276
Chen GH; Kong J; Reinhard K; Fechter LD. (2001). NMDA receptor blockage protects against permanent noise-induced
hearing loss but not its potentiation by carbon monoxide . Hear Res, 154: 108-115. 193985
ChenL; Yang W; JennisonBL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada, 1991-
1999. Inhal Toxicol, 14: 141-157. 024945
Chen P-C; Lai Y-M; Chan C-C; Hwang J-S; Yang C-Y; Wang J-D. (1999). Short-term effect of ozone on the pulmonary
function of children in primary school. Environ Health Perspect, 107: 921-925. 011149
Cheng C; Noordeloos AM; Jeney V; Scares MP; Moll F; Pasterkamp G; Serruys PW; Duckers HJ. (2009). Heme
oxygenase 1 determines atherosclerotic lesion progression into a vulnerable plaque. Circulation, 119: 3017-3027.
193775
September 2009 5-183 DRAFT - DO NOT CITE OR QUOTE
-------
Chevalier RB; Krumholz RA; Ross JC. (1966). Reaction of nonsmokers to carbon monoxide inhalation: cardiopulmonary
responses at rest and during exercise. JAMA, 198: 1061-1064. 010641
Chin BY; Jiang G; Wegiel B; Wang HJ; Macdonald T; Zhang XC; Gallo D; Cszimadia E; Bach FH; Lee PJ; Otterbein LE.
(2007). Hypoxia-inducible factor 1 alpha stabilization by carbon monoxide results in cytoprotective preconditioning.
ProcNatlAcadSciUSA, 104: 5109-5114. 190601
Chung Y; Huang SJ; Glabe A; Jue T. (2006). Implication of CO inactivation on myoglobin function. Am J Physiol Lung
Cell Mol Physiol, 290: 1616-1624. 193987
Coburn RF; Williams WJ; Kahn SB. (1966). Endogenous carbon monoxide production in patients with hemolytic anemia. J
Clin Invest, 45: 460-468. 010984
Coceani F; Breen CA; Lees JG; Falck JR; Olley PM. (1988). Further evidence implicating a cytochrome P-450-mediated
reaction in the contractile tension of the lamb ductus arteriosus. Circ Res, 62: 471-477. 040493
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA. (2004). Carbon monoxide actuates O2-limited heme
degredations in the rat brain. Free Radic Biol Med, 3 7: 1802-1812. 180440
D'Amico G; Lam F; Hagen T; Moncada S. (2006). Inhibition of cellular respiration by endogenously produced carbon
monoxide. J Cell Sci, 119: 2291-2298. 193992
D'Ippoliti D; Forastiere F; Ancona C; Agabiti N; Fusco D; Michelozzi P; Perucci CA. (2003). Air pollution and myocardial
infarction in Rome: a case-crossover analysis. Epidemiology, 14: 528-535. 074311
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery disease. J
Occup Environ Med, 46: 1217-1221.099036
Davies DM; Smith DJ. (1980). Electrocardiographic changes in healthy men during continuous low-level carbon monoxide
exposure. Environ Res, 21: 197-206. 011288
Davies RF; Topping DL; Turner DM. (1976). The effect of intermittent carbon monoxide exposure on experimental
atherosclerosis in the rabbit. Atherosclerosis, 24: 527-536. 010660
De Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; Van Wrjnen JH;
Kreyling W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on cardiorespiratory symptoms
in elderly subjects with coronary heart disease: the ULTRA study. Am J Epidemiol, 157: 613-623. 001061
De Jong MJ; Randall DC. (2005). Heart rate variability analysis in the assessment of autonomic function in heart failure. J
Cardiovasc Nurs, 20: 186-195; quiz 196-197. 193996
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality associations with
air pollution. Am J Respir Grit Care Med, 167: 1117-1123. 055688
De Luca A; Pierno S; Tricarico D; Carratu MR; Cagiano R; Cuomo V; Camerino DC. (1996). Developmental changes of
membrane electrical properties of rat skeletal muscle fibers produced by prenatal exposure to carbon monoxide.
Environ Toxicol Pharmacol, 2: 213-221. 080911
De Salvia MA; Cagiano R; Carratu MR; Di Giovanni V; Trabace L; Cuomo V. (1995). Irreversible impairment of active
avoidance behavior in rats prenatally exposed to mild concentrations of carbon monoxide. Psychopharmacology,
122:66-71.079441
Deedwania PC; Stein PK; Koren A; Mukherjee R; Pitt B. (2005). Decreased heart rate variability is an independent
predictor of sudden cardiac death in post-ML heart failure: Results of the EPHESUS arrhythmia and heart-rate
variability analysis. In JT Dipiro (Ed.),Pharmacotherapy: A Pathophysiological Approach (pp. 3029). New York
City: McGraw-Hill. 195134
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily ambient
exposures to toxic and criteria air pollutants. Environ Health Perspect, 111: 647-656. 050460
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J; Zaldivar F;
Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are
associated with primary combustion aerosols in subjects with coronary artery disease. Environ Health Perspect,
116:898-906. 156390
Delivoria-Papadopoulos M; Coburn RF; Forster RE. (1974). Cyclic variation of rate of carbon monoxide production in
normal women. J Appl Physiol, 36: 49-51. 086316
September 2009 5-184 DRAFT - DO NOT CITE OR QUOTE
-------
Denschlag D; Marculescu R; Unfried G; Hefler LA; Exner M; Hashemi A; Riener EK; Keck C; Tempfer CB; Wagner O.
(2004). The size of a microsatellite polymorphism of the haem oxygenase 1 gene is associated with idiopathic
recurrent miscarriage. , 10: 211-4. 193894
Di Cera E; Doyle ML; Morgan MS; De Cristofaro R; Landolfi R; Bizzi B; Castagnola M; Gill SJ. (1989). Carbon
monoxide and oxygen binding to human hemoglobin FO. Biochemistry, 28: 2631-2638. 193998
Di Giovanni V; Cagiano R; De Salvia MA; Giustino A; Lacomba C; Renna G; Cuomo V. (1993). Neurobehavioral changes
produced in rats by prenatal exposure to carbon monoxide. Brain Res, 616: 126-131. 013822
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD; Verrier RL.
(2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias recorded by implanted
cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dominici F; McDermott A; Daniels M; Zeger SL; Samet JM. (2003). Mortality among residents of 90 cities. In: Revised
analyses of time-series studies of air pollution and health. Health Effects Institute. Boston, MA. HEI Special
Report. http://www.healtheffects.org/Pubs/TimeSeries.pdf. 056116
Dominici F; McDermott A; Daniels M; Zeger SL; Samet JM. (2005). Revised analyses of the national morbidity, mortality,
and air pollution study: mortality among residents of 90 cities. J Toxicol Environ Health A, 68: 1071-1092. 087912
Dominici F; McDermott A; Zeger SL; Samet JM. (2002). On the use of generalized additive models in time-series studies
of air pollution and health. Am J Epidemiol, 156: 193-203. 030458
Drinkwater BL; Raven PB; Horvath SM; Gliner JA; Ruhling RO; Bolduan NW; Taguchi S. (1974). Air pollution, exercise,
and heat stress. Arch Environ Occup Health, 28: 177-181. 041332
Dubuis E; Gautier M; Melin A; Rebocho M; Girardin C; Bonnet P; Vandier C. (2002). Chronic carbon monoxide enhanced
IbTx-sensitive currents in rat resistance pulmonary artery smooth muscle cells. Am J Physiol Lung Cell Mol
Physiol, 283: L120-L129. 193911
Dubuis E; Gautier M; Melin A; Rebocho M; Girardin C; Bonnet P; Vandier C. (2003). Chronic carbon monoxide exposure
of hypoxic rats increases in vitro sensitivity of pulmonary artery smooth muscle. Can J Physiol Pharmacol, 81: 711-
719. 180439
Durante W; Johnson FK; Johnson RA. (2006). Role of carbon monoxide in cardiovascular function. J Cell Mol Med, 10:
672-686. 193778
Dyer R; Eccles U; Swartzwelder H; Fechter L; Annau Z. (1979). Prenatal carbon monoxide and adult evoked potentials in
rats. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev, 13: 107-120. 190994
Ekblom B; Huot R. (1972). Response to submaximal and maximal exercise at different levels of carboxyhemoglobin. Acta
Physiol, 86: 474-482. 010886
Farrugia G; Irons WA; Rae JL; Sarr MG; Szurszewski JH. (1993). Activation of whole cell currents in isolated human
jejunal circular smooth muscle cells by carbon monoxide. Am J Physiol, 264: Gil84-G1189. 013826
Favory R; Lancel S; Tissier S; Mathieu D; Decoster B; Neviere R. (2006). Myocardial dysfunction and potential cardiac
hypoxia in rats induced by carbon monoxide inhalation. Am J Respir Crit Care Med, 174: 320-325. 184462
Fechter LD; Annau Z. (1977). Toxicity of mild prenatal carbon monoxide exposure. Science, 197: 680-682. 010688
Fechter LD; Annau Z. (1980). Prenatal carbon monoxide exposure alters behavioral development. Neurotoxicol Teratol, 2:
7-11.011295
Fechter LD; Chen GD; Rao D. (2002). Chemical Asphyxiants and Noise. Noise Health, 4: 49-61. 193926
Fechter LD; Karpa MD; Proctor B; Lee AG; Storm JE. (1987). Disruption of neostriatal development in rats following
perinatal exposure to mild, but chronic carbon monoxide. Neurotoxicol Teratol, 9: 277-281. 012259
Fechter LD; Liu Y; Pearce TA. (1997). Cochlear protection from carbon monoxide exposure by free radical blockers in the
guinea pig. Toxicol Appl Pharmacol, 142: 47-55. 081322
Fechter LD; Thakur M; Miller B; Annau Z; Srivastava U. (1980). Effects of prenatal carbon monoxide exposure on cardiac
development. Toxicol Appl Pharmacol, 56: 370-375. 011294
September 2009 5-185 DRAFT - DO NOT CITE OR QUOTE
-------
Filho MAP; Pereira LAA; Arbex FF; Arbex M; Conceicao GM; Santos UP; Lopes AC; Saldiva PHN; Braga ALF; Cendon
S. (2008). Effect of air pollution on diabetes and cardiovascular diseases in Sao Paulo, Brazil. Braz J Med Biol Res,
41:526-532. 190260
Fischer PH; Steerenberg PA; Snelder JD; Van Loveren H; Van Amsterdam JGC. (2002). Association between exhaled nitric
oxide, ambient air pollution and respiratory health in school children. Int Arch Occup Environ Health, 75: 348-353.
025731
Fisher AB; Hyde RW; Baue AE; Reif JS; Kelly DF. (1969). Effect of carbon monoxide on function and structure of the
lung. J Appl Physiol, 26: 4-12. 012381
Fodor GG; Winneke G (1972). Effect of low CO concentrations on resistance to monotony and on psychomotor capacity.
Gefahrst Reinhalt Luft, 32: 46-54. 011041
Fried PA; Watkinson B; Gray R. (1998). Differential effects on cognitive functioning in 9- to 12-year olds prenatally
exposed to cigarettes and marihuana. Neurotoxicol Teratol, 20: 293-306. 190210
Fried PA; Watkinson B; Gray R. (2003). Differential effects on cognitive functioning in 13- to 16-year-olds prenatally
exposed to cigarettes and marihuana. Neurotoxicol Teratol, 25: 427-436. 190209
Fung KY; Luginaah I; Gorey KM; Webster G. (2005). Air pollution and daily hospital admissions for cardiovascular
diseases in Windsor, Ontario. Can J Public Health, 96: 29-33. 074322
Gautier M; Antier D; Bonnet P; Le Net JL; Hanton G; Eder V (2007). Continuous inhalation of carbon monoxide induces
right ventricle ischemia and dysfunction in rats with hypoxic pulmonary hypertension. Am J Physiol Heart Circ
Physiol, 293: H1046-52. 096471
Gerritzen-Bruning, M. J.; van denlngh, T. S.; Rothuizen, J. (2006). Diagnostic value of fasting plasma ammonia and bile
acid concentrations in the identification of portosystemic shunting in dogs. J Vet Intern Med, 20: 13-19. 193395
Ohio AJ; Stonehuerner JG; Dailey LA; Richards JH; Madden MD; Deng Z; Nguyen NB; Callaghan KD; Yang F;
Piantadosi CA. (2008). Carbon Monoxide Reversibly Alters Iron Homeostasis and Respiratory Epithelial Cell
Function. Am J Respir Cell Mol Biol, 38: 715-723. 096321
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation between
ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J Epidemiol, 162: 238-
252. 087892
Giustino A; Cagiano R; Carratu MR; Cassano T; Tattoli M; Cuomo V (1999). Prenatal exposure to low concentrations of
carbon monoxide alters habituation and non-spatial working memory in rat offspring. Brain Res, 844: 201-205.
011538
Giustino A; Cagiano R; Carratu MR; De Salvia MA; Panaro MA; Jirillo E; Cuomo V (1993). Immunological changes
produced in rats by prenatal exposure to carbon monoxide. Basic Clin Pharmacol Toxicol, 73: 274-278. 013833
Giustino A; Carratu MR; Brigiani GS; De Salvia MA; Pellegrino NM; Steardo L; Jirillo E; Cuomo V (1994). Changes in
the frequency of splenic immunocompetent cells in rats exposed to carbon monoxide during gestation.
Immunopharmacol Immunotoxicol, 16: 281-292. 076343
Glabe A; Chung Y; Xu D; Jue T. (1998). Carbon monoxide inhibition of regulatory pathways in myocardium. Am J
Physiol, 274: H2143-2151. 086704
Gold DR; Litonjua A; Schwartz J; Lovett E; Larson A; Nearing B; Allen G; Verrier M; Cherry R; Verrier R. (2000).
Ambient pollution and heart rate variability. Chest, 101: 1267-1273. 011432
Gold DR; Litonjua AA; Zanobetti A; Coull BA; Schwartz J; MacCallum G; Verrier RL; Nearing BD; Canner MJ; Suh H;
Stone PH. (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health Perspect, 113: 883-
887. 087558
Goldberg MS; Burnett RT; Valois M-F; Flegel K; Bailar JC III; Brook J; Vincent R; Radon K. (2003). Associations
between ambient air pollution and daily mortality among persons with congestive heart failure. Environ Res, 91:8-
20. 035202
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive heart
failure to estimate the short-term effects from personal factors and environmental conditions on oxygen saturation
and pulse rate. Occup Environ Med, 65: 659-666. 180380
September 2009 5-186 DRAFT - DO NOT CITE OR QUOTE
-------
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. Am J Respir Grit Care Med, 169: 816-821. 055624
Gouveia N; Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in Sao Paulo,
Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Grover TR; Rairigh RL; Zenge JP; Abman SH; Kinsella JP. (2000). Inhaled carbon monoxide does not cause pulmonary
vasodilation in the late-gestation fetal lamb. Am J Physiol Lung Cell Mol Physiol, 278: L779-84. 097088
Gryparis A; Forsberg B; Katsouyanni K; Analitis A; Touloumi G; Schwartz J; Samoli E; Medina S; Anderson HR; Niciu
EM; Wichmann H-E; Kriz B; Kosnik M; Skorkovsky J; Vonk JM; Dortbudak Z. (2004). Acute effects of ozone on
mortality from the "Air Pollution and Health: A European Approach" project. Am J Respir Grit Care Med, 170:
1080-1087. 057276
Guo YL; Lin Y-C; Sung F-C; Huang S-L; Ko Y-C; Lai J-S; Su H-J; Shaw C-K; Lin R-S; Dockery DW. (1999). Climate,
traffic-related air pollutants, and asthma prevalence in middle-school children in Taiwan. Environ Health Perspect,
107: 1001-1006. 010937
Ha E-H; Hong Y-C; Lee B-E; Woo B-H; Schwartz J; Christiani DC. (2001). Is air pollution a risk factor for low birth
weight in Seoul?. Epidemiology, 12: 643-648. 019390
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul, South Korea.
Pediatrics, 111: 284-290. 042552
Hajat S; Armstrong B; Wilkinson P; Busby A; Dolk H. (2007). Outdoor air pollution and infant mortality: Analysis of daily
time-series data in 10 English cities. J Epidemiol Community Health, 61: 719-722. 093276
Hampson NB; Weaver LK. (2007). Carbon monoxide poisoning: a new incidence for an old disease. Undersea Hyperb
Med, 34: 163-8. 190272
Hanada A; Sander M; Gonzalez-Alonso J. (2003). Human skeletal muscle sympathetic nerve activity, heart rate and limb
haemodynamics with reduced blood oxygenation and exercise. J Physiol, 551: 635-647. 193915
Harada T; Koi H; Kubota T; Aso T. (2004). Haem oxygenase augments porcine granulosa cell apoptosis in vitro. J
Endocrinol, 181: 191-205. 193920
Hausberg M; Somers VK. (1997). Neural circulatory responses to carbon monoxide in healthy humans. Hypertension, 29:
1114-1118.083450
Hawkins RD; Zhuo M; Arancio O. (1994). Nitric oxide and carbon monoxide as possible retrograde messengers in
hippocampal long-term potentiation. Dev Neurobiol, 25: 652-665. 076503
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic stroke
occurrence: A case-crossover analysis from a 10-year population-based study in Dijon, France. Occup Environ
Med, 64: 439-445. 093270
Hermans RH; McGivern RF; Chen W; Longo LD. (1993). Altered adult sexual behavior in the male rat following chronic
prenatal hypoxia. Neurotoxicol Teratol, 15: 353-63. 190510
Hill EP; Hill JR; Power GG; Longo LD. (1977). Carbon monoxide exchanges between the human fetus and mother: A
mathematical model. Am J Physiol, 232: H311-H323. 011315
Hirsch T; Weiland SK; Von Mutius E; Safeca AF; Grafe H; Csaplovics E; Duhme H; Keil U; Leupold W. (1999). Inner city
air pollution and respiratory health and atopy in children. Eur Respir J, 14: 669-677. 003537
Hoek G. (2003). Daily mortality and air pollution in The Netherlands. 042818
Hoek G; Brunekreef B; Fischer P; Van Wrjnen J. (2001). The association between air pollution and heart failure,
arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study. Epidemiology,
12: 355-357. 016550
Hoek G; Brunekreef B; Verhoeff A; Van Wijnen J; Fischer P. (2000). Daily mortality and air pollution in the Netherlands. J
Air Waste Manag Assoc, 50: 1380-1389. 010350
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air pollution
and heart rate variability among the elderly in Mexico City. Epidemiology, 14: 521-527. 057326
September 2009 5-187 DRAFT - DO NOT CITE OR QUOTE
-------
Horvath SM; Agnew JW; Wagner JA; Bedi JF. (1988). Maximal aerobic capacity at several ambient concentrations of
carbon monoxide at several altitudes. 012725
Horvath SM; Dahms TE; O'Hanlon JF. (1971). Carbon monoxide and human vigilance: A deleterious effect of present
urban concentrations. Arch Environ Occup Health, 23: 343-347. 011075
Horvath SM; Raven PB; Dahms TE; Gray DJ. (1975). Maximal aerobic capacity at different levels of carboxyhemoglobin.
JApplPhysiol, 38: 300-303. 010887
Hosseinpoor AR; Forouzanfar MH; Yunesian M; Asghari F; Naieni KH; Farhood D. (2005). Air pollution and
hospitalization due to angina pectoris in Tehran, Iran: A time-series study. Environ Res, 99: 126-131. 087413
Hsia CC. (2002). Recruitment of lung diffusing capacity: Update of concept and application. Chest, 122: 1774-1783.
193857
Huikuri HV; Jokinen V; Syvanne M; Nieminen MS; Airaksinen KEJ; Ikaheimo MJ; Koistinen JM; Kauma H; Kesaniemi
AY; Majahalme S; Niemela KO; Frick MH. (1999). Heart rate variability and progression of coronary
atheroscerosis. Arterioscler Thromb Vase Biol, 19: 1979-1985. 184464
Hutcheon JA; Platt RW. (2008). The missing data problem in birth weight percentiles and thresholds for "small-for-
gestational-age". Am J Epidemiol, 167: 786-792. 193795
Huynh M; Woodruff TJ; Parker JD; Schoendorf KC. (2006). Relationships between air pollution and preterm birth in
California. Paediatr Perinat Epidemiol, 20: 454-461. 091240
Hwang B-F; Jaakkola JJK; Lee Y-L; Lin Y-C; Guo Y-LL. (2006). Relation between air pollution and allergic rhinitis in
Taiwanese schoolchildren. Respir Res, 7: 23. 088971
Hwang B-F; Lee Y-L; Lin Y-C; Jaakkola JJK; Guo YL. (2005). Traffic related air pollution as a determinant of asthma
among Taiwanese school children. Thorax, 60: 467-473. 089454
Hwang BF; Jaakkola JJ. (2008). Ozone and other air pollutants and the risk of oral clefts. Environ Health Perspect, 116:
1411-1415. 193794
Ibald-Mulli A; Stieber J; Wichmann H-E; Koenig W; Peters A. (2001). Effects of air pollution on blood pressure: A
population-based approach. Am J Public Health, 91: 571-577. 016030
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J; Hoek G;
Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and heart rate in subjects
with cardiovascular disease: a multicenter approach. Environ Health Perspect, 112: 369-377. 087415
Iheagwara KN; Thorn SR; Deutschman CS; Levy RJ. (2007). Myocardial cytochrome oxidase activity is decreased
following carbon monoxide exposure. Biochim Biophys Acta, 1772: 1112-1116. 193861
Imai T; Morita T; Shindo T; Nagai R; Yazaki Y; Kurihara H; Suematsu M; Katayama S. (2001). Vascular smooth muscle
cell-directed overexpression of heme oxygenase-1 elevates blood pressure through attenuation of nitric oxide-
induced vasodilation in mice. Circ Res, 89: 55-62. 193864
Ischiropoulos H; Beers MF; Ohnishi ST; Fisher D; Garner SE; Thorn SR. (1996). Nitric oxide production and perivascular
tyrosine nitration in brain after carbon monoxide poisoning in the rat. J Clin Invest, 97: 2260-2267. 079491
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions
in the context of time-series health effects models. J Expo Sci Environ Epidemiol, 17 Suppl 2: S45-S60. 156594
Jalaludin B; Mannes T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution on
gestational age is modified by season in Sydney, Australia. Environ Health, 6:16. 156601
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient air
pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+ years),
Sydney, Australia. JExpo Sci Environ Epidemiol, 16: 225-37. 189416
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of the air
pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health A, 66: 1735-1777.
087380
Johnson FK; Durante W; Peyton KJ; Johnson RA. (2003). Heme oxygenase inhibitor restores arteriolar nitric oxide
function in Dahl rats. Hypertension, 41: 149-155. 193868
September 2009 5-188 DRAFT - DO NOT CITE OR QUOTE
-------
Johnson FK; Durante W; Peyton KJ; Johnson RA. (2004). Heme oxygenase-mediated endothelial dysfunction in DOCA-
salt, but not in spontaneously hypertensive, rat arterioles. Am J Physiol Heart Circ Physiol, 286: 1681-1687. 193870
Johnson FK; Johnson RA. (2003). Carbon monoxide promotes endothelium-dependent constriction of isolated gracilis
muscle arterioles. Am J Physiol, 285: R536-R541. 053611
Johnson FK; Johnson RA; Durante W; Jackson KE; Stevenson BK; Peyton KJ. (2006). Metabolic syndrome increases
endogenous carbon monoxide production to promote hypertension and endothelial dysfunction in obese Zucker
rats. Am J Physiol Regul Integr Comp Physiol, 290: 601-608. 193874
JoumardR; Chiron M; Vidon R; Maurin M; Rouzioux J-M. (1981). Mathematical models of the uptake of carbon
monoxide on hemoglobin at low carbon monoxide levels. Environ Health Perspect, 41: 277-289. 011330
Kanten WE; Penney DG; Francisco K; Thill JE. (1983). Hemodynamic responses to acute carboxyhemoglobinemia in the
rat. Am J Physiol, 244: H320-H327. 011333
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Karr C; Lumley T; Shepherd K; Davis R; Larson T; Ritz B; Kaufman J. (2006). A case-crossover study of wintertime
ambient air pollution and infant bronchiolitis. Environ Health Perspect, 114: 277-281. 088751
Katoue MG; Khan I; Oriowo MA. (2005). Increased expression and activity of heme oxygenase-2 in pregnant rat aorta is
not involved in attenuated vasopressin-induced contraction. Naunyn Schmiedebergs Arch Pharmacol, 372: 220-7.
193896
Katsouyanni K; Touloumi G; Samoli E; Gryparis A; Le Tertre A; Monopolis Y; Rossi G; Zmirou D; Ballester F; Boumghar
A; Anderson HR; Wojtyniak B; Paldy A; Braunstein R; Pekkanen J; Schindler C; Schwartz J. (2001). Confounding
and effect modification in the short-term effects of ambient particles on total mortality: results from 29 European
cities within the APHEA2 project. Epidemiology, 12: 521-531. 019008
Kaur S; Nieuwenhuijsen M; Colvile R. (2005). Personal exposure of street canyon intersection users to PM2.5, ultrafine
particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641. 086504
Kim HP; Wang X; Nakao A; Kim SI; Murase N; Choi ME; Ryter SW; Choi AM. (2005). Caveolin-1 expression by means
of mitogen-activated protein kinase mediates the antiprofiferative effect of carbon monoxide . Proc Natl Acad Sci U
S A, 102: 11319-11324. 193959
Kim HS; Loughran PA; Rao J; Billar TR; Zuckerbraun BS. (2008). Carbon monoxide activates NF-kappaB via ROS
generation and Akt pathways to protect against cell death of hepatocytes. Am J Physiol Gastrointest Liver Physiol,
295: G146-G152. 193961
Kingdom JC; KaufmannP. (1997). Oxygen and placental villous development: origins of fetal hypoxia. Placenta, 18: 613-
21; discussion 623-6. 193897
Kinney HC; Filiano JJ; Sleeper LA; Mandell F; Valdes-Dapena M; White WF. (1995). Decreased muscarinic receptor
binding in the arcuate nucleus in sudden infant death syndrome. Science, 269: 1446-1450. 193898
Kizakevich PN; McCartney ML; Hazucha MJ; Sleet LH; Jochem WJ; Hackney AC; Bolick K. (2000). Noninvasive
ambulatory assessment of cardiac function in healthy men exposed to carbon monoxide during upper and lower
body exercise. Eur J Appl Physiol, 83: 7-16. 052691
Klasner AE; Smith SR; Thompson MW; Scalzo AJ. (1998). Carbon monoxide mass exposure in a pediatric population.
Acad Emerg Med, 5: 992-996. 087196
Klausen K; Rasmussen B; Gjellerod H; Madsen H; Petersen E. (1968). Circulation, metabolism and ventilation during
prolonged exposure to carbon monoxide and to high altitude. Scand J Clin Lab Invest, 103: 26-38. 193936
Kleeberger SR; Ohtsuka Y. (2005). Gene-particulate matter-health interactions. Toxicol Appl Pharmacol, 207: S276-S281.
130489
Kleinman MT; Davidson DM; Vandagriff RB; Caiozzo VJ; Whittenberger JL. (1989). Effects of short-term exposure to
carbon monoxide in subjects with coronary artery disease. Arch Environ Occup Health, 44: 361-369. 012696
Kleinman MT; Leaf DA; Kelly E; Caiozzo V; OsannK; O'Niell T. (1998). Urban angina in the mountains: effects of carbon
monoxide and mild hypoxemia on subjects with chronic stable angina. Arch Environ Occup Health, 53: 388-397.
047186
September 2009 5-189 DRAFT - DO NOT CITE OR QUOTE
-------
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from
ARIES. Inhal Toxicol, 16 Suppl 1: 131-141. 056585
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling of
eNOS. Toxicol Appl Pharmacol, 230: 346. 191987
Koehler RC; Jones MD Jr; Traystman RJ. (1982). Cerebral circulatory response to carbon monoxide and hypoxic hypoxia
in the lamb. Am J Physiol, 243: H27-H32. 011341
Kohsaka A; Watanobe H; Kakizaki Y; Suda T. (1999). A comparative study of the effects of nitric oxide and carbon
monoxide on the in vivo release of gonadotropin-releasing hormone and neuropeptide Y from rat hypothalamus
during the estradiol-induced luteinizing hormone surge: estimation by push-pull perfusion. Neuroendocrinology,
69: 245-253. 191000
Koike A; Wasserman K; Armon Y; Weiler-Ravell D. (1991). The work-rate-dependent effect of carbon monoxide on
ventilatory control during exercise. Respir Physiol Neurobiol, 85: 169-183. 013500
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver. Environ Health Perspect, 111: 1312-1317. 049466
Korres S; Riga M; Balatsouras D; Papadakis C; Kanellos P; Ferekidis E. (2007). Influence of smoking on developing
cochlea: Does smoking during pregnancy affect the amplitudes of transient evoked otoacoustic emissions in
newborns?. Int J Pediatr Otorhinolaryngol, 71: 781-786. 190908
Kotsch K; Francuski M; Pascher A; Klemz R; Seifert M; Mittler J; Schumacher G; Buelow R; Volk HD; Tullius SG;
Neuhaus P; Pratschke J. (2006). Improved long-term graft survival after HO-1 induction in brain-dead donors !. , 6:
477-86. 193899
Kreider JC; Blumberg MS. (2005). Geotaxis and beyond: Commentary on Motz and Alberts. Neurotoxicol Teratol, 27:
529-533. 193944
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett M; Abrahamowicz M; White WH. (2000).
Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate air pollution and
mortality: a special report of the Institute's Particle Epidemiology Reanalysis Project. Health Effects Institute.
Cambridge, MA. 012281
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Kunzli N; Tager IB. (1997). The semi-individual study in air pollution epidemiology: a valid design as compared to
ecologic studies. Environ Health Perspect, 105: 1078-1083. 086180
Lagorio S; Forastiere F; Pistelli R; lavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air
pollution and lung function among susceptible adult subjects: a panel study. Environ Health, 5:11. 089800
Lamar C; Mahesh V; Brann D. (1996). Regulation of gonadotrophin-releasing hormone (GnRH) secretion by heme
molecules: a regulatory role for carbon monoxide?. Endocrinology, 137: 790-793. 190997
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; DTppoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S; Salomaa V;
Sunyer J; Tiittanen P; Von Klot S; Forastiere F; for the HEAPSS Study Group. (2006). Associations of traffic-
related air pollutants with hospitalisation for first acute myocardial infarction: the HEAPSS study. Occup Environ
Med, 63: 844-851.089788
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different gestational
phases contributes to risks of low birth weight. Hum Reprod, 18: 638-643. 043202
Lee IM; Tsai SS; Ho CK; Chiu HF; Yang CY (2007). Air pollution and hospital admissions for congestive heart failure in a
tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 19: 899-904. 093271
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic heart
diseases among individuals 64+years of age residing in Seoul, Korea. Arch Environ Health, 58: 617-623. 095552
Lee Y-L; Shaw C-K; Su H-J; Lai J-S; Ko Y-C; Huang S-L; Sung F-C; Guo Y-L. (2003). Climate, traffic-related air
pollutants and allergic rhinitis prevalence in middle-school children in Taiwan. Eur Respir J, 21: 964-970. 049201
September 2009 5-190 DRAFT - DO NOT CITE OR QUOTE
-------
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers IF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy andpreterm delivery. Environ Health Perspect, 114: 905-910. 089828
Levy D; Lumley T; Sheppard L; Kaufman J; Checkoway H. (2001). Referent selection in case-crossover analyses of acute
health effects of air pollution. Epidemiology, 12: 186-192. 017172
Levy D; Sheppard L; Checkoway H; Kaufman J; Lumley T; Koenig J; Siscovick D. (2001). A case-crossover analysis of
particulate matter air pollution and out-of-hospital primary cardiac arrest. Epidemiology, 12: 193-199. 017171
Li CI; Daling JR; Emanuel I. (2003). Birthweight and risk of overall and cause-specific childhood mortality. Paediatr
Perinat Epidemiol, 17: 164-170. 193965
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J Epidemiol,
159:768-777.056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ Epidemiol, 15: 319-328.
088677
Lin C-M; Li C-Y; Mao I-F. (2004). Increased risks of term low-birth-weight infants in a petrochemical industrial city with
high air pollution levels. Arch Environ Occup Health, 59: 663-668. 089827
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and neonatal deaths
in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
LinH;McGrath JJ. (1988). Carbon monoxide effects on calcium levels in vascular smooth muscle. Life Sci, 43: 1813-
1816.012773
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2003). Effect of short-term exposure to gaseous pollution on
asthma hospitalisation in children: a bi-directional case-crossover analysis. J Epidemiol Community Health, 57: 50-
55. 042549
Lin M; Chen Y; Villeneuve PJ; Burnett RT; Lemyre L; Hertzman C; McGrail KM; Krewski D. (2004). Gaseous air
pollutants and asthma hospitalization of children with low household income in Vancouver, British Columbia,
Canada. Am J Epidemiol, 159: 294-303. 055600
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children
younger than 15 years in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240. 087828
Linn WS; Szlachcic Y; Gong H Jr; Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions in
metropolitan Los Angeles. Environ Health Perspect, 108: 427-434. 002839
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Morris SC. (2002). Temporal and spacial relationships between age-specific mortality and ambient air quality
in the United States: preliminary results for counties, 1960-97. Occup Environ Med, 59: 156-174. 019217
Lipfert FW; Perry HM Jr; Miller JP; Baty JD; Wyzga RE; Carmody SE. (2000). The Washington University-EPRI veterans'
cohort mortality study: preliminary results. Inhal Toxicol, 4: 41-73. 004087
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006). Traffic density as a surrogate measure of environmental exposures in
studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos Environ, 40: 154-169.
088218
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air pollutants
during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432. 090429
Liu S; Krewski D; Shi Y; Chen Y; Burnett RT. (2003). Association between gaseous ambient air pollutants and adverse
pregnancy outcomes in Vancouver, Canada. Environ Health Perspect, 111: 1773-1778. 089548
Liu Y; Fechter LD. (1995). MK-801 protects against carbon monoxide-induced hearing loss. Toxicol Appl Pharmacol, 132:
196-202.076524
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni K; von Klot
S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and fibrinogen
polymorphisms. Environ Health Perspect, 117: 1373-1379. 191983
September 2009 5-191 DRAFT - DO NOT CITE OR QUOTE
-------
Loennechen JP; Beisvag V; Arbo I; Waldum HL; Sandvik AK; Knardahl S; Ellingsen O. (1999). Chronic carbon monoxide
exposure in vivo induces myocardial endothelin-1 expression and hypertrophy in rat. Basic Clin Pharmacol Toxicol,
85: 192-197.011549
Longo LD. (1970). Carbon monoxide in the pregnant mother and fetus and its exchange across the placenta. AnnN YAcad
Sci, 174: 312-341.013922
Longo LD; Hill EP. (1977). Carbon monoxide uptake and elimination in fetal and maternal sheep. Am J Physiol, 232:
H324-H330. 010802
Longo, M; Jain V; Vedernikov YP; Saade GR; Goodrum L; Facchinetti F; Garfield RE. (1999). Effect of nitric oxide and
carbon monoxide on uterine contractility during human and rat pregnancy. Am J Obstet Gynecol, 181: 981-988.
011548
Lopez I; Acuna D; Webber DS; Korsak RA; Edmond J. (2003). Mild carbon monoxide exposure diminishes selectively the
integrity of the cochlea of the developing rat. , 74: 666-75. 193901
Lopez IA; Acuna D; Beltran-Parrazal L; Espinosa-Jeffrey A; Edmond J. (2008). Oxidative stress and the deleterious
consequences to the rat cochlea after prenatal chronic mild exposure to carbon monoxide in air. Neuroscience, 151:
854-867. 097343
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009). Vehicular
emissions induce vascular MMP-9 expression and activity associated with endothelin-l?mediated pathways.
Arterioscler Thromb Vase Biol, 29: 511-517. 180257
Lyall F; Barber A; Myatt L; Bulmer JN; Robson SC. (2000). Hemeoxygenase expression in human placenta and placental
bed implies a role in regulation of trophoblast invasion and placental function. , 14: 208-19. 193902
MacDorman MF; Callaghan WM; Mathews TJ; Hoyert DL; Kochanek KD. (2007). Trends in preterm-related infant
mortality by race and ethnicity, United States 1999-2004. Int J Health Serv, 37: 635-641. 193973
MacMillan V (1975). Regional cerebral blood flow of the rat in acute carbon monoxide intoxication. Can J Physiol
Pharmacol, 53: 644-650. 012909
Mactutus CF; Fechter LD. (1984). Prenatal exposure to carbon monoxide: learning and memory deficits. Science, 223:
409-411.011355
Mactutus CF; Fechter LD. (1985). Moderate prenatal carbon monoxide exposure produces persistent, and apparently
permanent, memory deficits in rats. Teratology, 31: 1-12. 011536
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air pollution
and stroke in Sheffield, United Kingdom: A small-area level geographical study. Stroke, 36: 239-243. 088683
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth weight in the
northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
Mann JK; Tager IB; Lurmann F; Segal M; Quesenberry CP Jr; Lugg MM; Shan J; Van den Eeden SK. (2002). Air pollution
and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia. Environ
Health Perspect, 110: 1247-1252. 036723
Marines T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on birth
weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Marthan R; Castaing Y; Manier G; Guenard H. (1985). Gas exchange alterations in patients with chronic obstructive lung
disease. Chest, 87: 470-475. 086334
Martin JA; Hamilton BE; Sutton PD; Ventura SJ; Menacker F; Kirmeyer S; Munson ML. (2007). Births: Final data for
2005. National Center for Health Statistics. Hyattsville, MD. 193982
Mayr FB; Spiel A; Leitner J; Marsik C; Germann P; Ullrich R; Wagner O; Jilma B. (2005). Effects of carbon monoxide
inhilation during experimental endotoxemia in humans. Am J Respir Crit Care Med, 171: 354-360. 193984
McGrath JJ. (1992). Effects of altitude on endogenous carboxyhemoglobin levels. J Air Waste Manag Assoc, 35: 127-133.
001005
McGrath JJ; Schreck RM; Lee PS. (1993). Carboxyhemoglobin levels in humans: Effects of altitude. Inhal Toxicol, 5: 241-
249.013865
September 2009 5-192 DRAFT - DO NOT CITE OR QUOTE
-------
McGregor HP; Westcott K; Walker DW. (1998). The effect of prenatal exposure to carbon monoxide on breathing and
growth of the newborn guinea pig. Pediatr Res, 43: 126-131. 085342
Mclntire DD; Bloom SL; Casey BM; Leveno KJ. (1999). Birth weight in relation to morbidity and mortality among
newborn infants. N Engl J Med, 340: 1234-1238. 015310
McLaughlin BE; Lash GE; Smith GN; Marks GS; Nakatsu K; Graham CH; Brien JF. (2003). Heme oxygenase expression
in selected regions of term human placenta. Exp Biol Med, 228: 564-7. 193904
Medeiros A; Gouveia N. (2005). [Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil].
Rev Saude Publica, 39: 965-972. 156750
Meilin S; Rogatsky GG; Thorn SR; ZarchinN; Guggenheimer-Furman E; Mayevsky A. (1996). Effects of carbon
monoxide on the brain may be mediated by nitric acid. J Appl Physiol, 81: 1078-1083. 079919
Melin A; Bonnet P; Eder V; Antier D; Obert P; Fauchier L. (2005). Direct implication of carbon monoxide in the
development of heart failure in rats with cardiac hypertrophy subjected to air pollution. Cardiovasc Toxicol, 5: 311-
320. 193833
Melin A; Obert P; Rebocho M; Bonnet P. (2002). Cardiac morphology and function following long-term exposure to
carbon monoxide at high altitude in rats. J Toxicol Environ Health A, 65:1981-1998. 037502
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007). Traffic and outdoor air pollution levels near residences and
poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-463. 093275
Mercke C; Lundh B. (1976). Erythrocyte filterability and heme catabolism during the menstrual cycle. Ann Intern Med, 85:
322-324. 086309
Mereu G; Cammalleri M; Fa M; Francesconi W; Saba P; Tattoli M; Trabace L; Vaccari A; Cuomo V. (2000). Prenatal
exposure to a low concentration of carbon monoxide disrupts hippocampal long-term potentiation in rat offspring. J
Pharmacol Exp Ther, 294: 728-734. 193838
Merriam-Webster. (2009). Merriam-Webster on-line: http://www.merriam-webster.com/medical/sussceptible,
http://merriam-webster.com/medical/vulerable. Retrieved 15-JUN-09, from http://www.merrian-
webster.com/medical/. 192146
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592. 092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Meyer J; Prien T; Van Aken H; Bone H-G; Waurick R; Theilmeier G; Booke M. (1998). Arterio-venous
carboxyhemoglobin difference suggests carbon monoxide production by human lungs. Biochem Biophys Res
Commun, 244: 230-232. 047530
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term exposure
to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458. 090130
Montagnani M; Serio M; Potenza MA; Mansi G; De Salvia MA; Cagiano R; Cuomo V; Mitolo-Chieppa D. (1996).
Prenatal exposure to carbon monoxide and vascular responsiveness of rat resistance vessels. Life Sci, 59: 1553-
1561.080902
Moolgavkar SH. (2000). Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
metropolitan areas in the United States. Inhal Toxicol, 12: 75-90. 010274
Moolgavkar SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook counties. Health
Effects Institute. Boston, MA. 042864
Moon JS; Kim YS; Kim JH; Son BS; Kim DS; Yang W. (2009). Respiratory health effects among schoolchildren and their
relationship to air pollutants in Korea. Int J Environ Health Res, 19: 31-48. 190297
Morita T; Mitsialis SA; Koike H; Liu Y; Kourembanas S. (1997). Carbon monoxide controls the proliferation of hypoxic
vascular smooth muscle cells. J Biol Chem, 272: 32,804-32,809. 085345
Morse CI; Pritchard LJ; Wust RC; Jones DA; Degens H. (2008). Carbon monoxide inhalation reduces skeletal muscle
fatigue resistance. Eur Phys J A, 192: 397-401. 097980
September 2009 5-193 DRAFT - DO NOT CITE OR QUOTE
-------
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Air pollution and pulmonary function in
asthmatic children: effects of prenatal and lifetime exposures. Epidemiology, 19: 550-557. 122163
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air pollution and
allergic sensitization in children with asthma. J Asthma, 45: 874-881. 187280
Naik JS; Walker BR. (2003). Heme oxygenase-mediated vasodilation involves vascular smooth muscle cell
hyperpolarization. Am J Physiol Heart Circ Physiol, 285: 220-228. 193852
Ndisang JF; Tabien HEN; Wang R. (2004). Carbon monoxide and hypertension. Am J Hypertens, 22: 1057-1074. 180425
Neas LM; Schwartz J. (1996). The determinants of pulmonary diffusing capacity in a national sample of US adults. Am J
Respir Grit Care Med, 153: 656-664. 079363
Neggers YH; Singh J. (2006). Zinc supplementation to protein-deficient diet in CO-exposed mice decreased fetal mortality
and malformation. Biol Trace Elem Res, 114: 269-279. 193964
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Grain EF; Evans R, 3rd; Gruchalla R; Morgan W; Stout J; Adams
GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with asthma in US inner
cities. J Allergy Clinlmmunol, 121: 1133-1139 el!31. 156818
Otto DA; Benignus VA; Prah JD. (1979). Carbon monoxide and human time discrimination: failure to replicate Beard-
Wertheim experiments. Aviat Space Environ Med, 50: 40-43. 010863
Park JW; Lim YH; Kyung S Y; An CH; Lee SP; Jeong SH; Ju S-Y (2005). Effects of ambient particulate matter on peak
expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in Korea. Respirology, 10:
470-476. 088673
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate variability: The VA
normative aging study. Environ Health Perspect, 113: 304-309. 057331
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in California.
Pediatrics, 115: 121-128. 087462
Patel AP; Moody JA; Handy RD; Sneyd JR. (2003). Carbon monoxide exposure in rat heart: glutathione depletion is
prevented by antioxidants. Biochem Biophys Res Commun, 302: 392-396. 043155
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Pekkanen J; Brunner EJ; Anderson HR; Tiittanen P; Atkinson RW. (2000). Daily concentrations of air pollution and plasma
fibrinogen in London. Occup Environ Med, 57: 818-822. 013250
Perm A; Currie J; Snyder C. (1992). Inhalation of carbon monoxide does not accelerate arteriosclerosis in cockerels. Eur J
Clin Pharmacol, 228: 155-164. 013728
Penney DG. (1988). Areview: Hemodynamic response to carbon monoxide. Environ Health Perspect, 77: 121-130. 012519
Penney DG; BaylerianMS; Thill JE; Fanning CM; Yedavally S. (1982). Postnatal carbon monoxide exposure: immediate
and lasting effects in the rat. Am J Physiol, 243: H328-H339. 011387
Penney DG; Baylerian MS; Thill JE; Yedavally S; Fanning CM. (1983). Cardiac response of the fetal rat to carbon
monoxide exposure. Am J Physiol, 244: H289-H297. 011385
Penney DG; Davidson SB; Gargulinski RB; Caldwell-Ayre TM. (1988). Heart and lung hypertrophy, changes in blood
volume, hematocrit and plasma renin activity in rats chronically exposed to increasing carbon monoxide
concentrations. J Appl Toxicol, 8: 171-178. 012521
Penney DG; Stryker AE; BaylerianMS. (1984). Persistent cardiomegaly induced by carbon monoxide and associated
tachycardia. J Appl Physiol, 56: 1045-1052. 011567
Penttinen P; Timonen KL; Tiittanen P; Mirme A; Ruuskanen J; Pekkanen J. (2001). Ultrafine particles in urban air and
respiratory health among adult asthmatics. Eur Respir J, 17: 428-435. 030335
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of
myocardial infarction. Circulation, 103: 2810-2815. 016546
September 2009 5-194 DRAFT - DO NOT CITE OR QUOTE
-------
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery DW.
(2000). Air pollution and incidence of cardiac arrhythmia. Epidemiology, 11: 11-17. 011347
Peters A; Perz S; Doring A; Stieber J; Koenig W; Wichmann H-E. (1999). Increases in heart rate during an air pollution
episode. Am J Epidemiol, 150: 1094-1098. 011554
Peters JM; Avol E; Navidi W; London SJ; Gauderman WJ; Lurmann F; Linn WS; Margolis H; Rappaport E; Gong H Jr;
Thomas DC. (1999). A study of twelve southern California communities with differing levels and types of air
pollution I Prevalence of respiratory morbidity. Am J Respir Crit Care Med, 159: 760-767. 087243
PetersenLC. (1977). The effect of inhibitors on the oxygen kinetics of cytochrome c oxidase. Biochim Biophys Acta, 460:
299-307. 193764
Piantadosi CA. (2002). Biological chemistry of carbon monoxide. Antioxid Redox Signal, 4: 259-270. 037463
Piantadosi CA. (2008). Carbon monoxide, reactive oxygen signaling, and oxidative stress. Free Radic Biol Med, 45: 562-
569. 180423
Piantadosi CA; Carraway MS; Suliman HB. (2006). Carbon monoxide, oxidative stress, and mitochondrial permeability
pore transition. Free Radic Biol Med, 40: 1332-1339. 180424
Piantadosi CA; Zhang J; Demchenko IT. (1997). Production of hydroxyl radical in the hippocampus after CO hypoxia or
hypoxic hypoxia in the rat. Free Radic Biol Med, 22: 725-732. 081326
Pope CAIII; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-1141. 024689
Pope CAIII; Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW Jr. (1995). Particulate air
pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Crit Care Med, 151: 669-674.
045159
Prigge E; Hochrainer D. (1977). Effects of carbon monoxide inhalation on erythropoiesis and cardiac hypertrophy in fetal
rats. Toxicol Appl Pharmacol, 42: 225-228. 012326
Putz VR; Johnson BL; Setzer JV. (1979). A comparative study of the effects of carbon monoxide and methylene chloride on
human performance. J Environ Pathol Toxicol Oncol, 2: 97-112. 023137
RabinovitchN; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient air
pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin
Immunol, 114: 1131-1137.096753
Rajendra Acharya U; Paul Joseph K; Kannathal N; Lim CM; Suri JS. (2006). Heart rate variability: A review. Med Biol
Eng Comput, 44: 1031-1051. 193787
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P. (2004). Air
pollution and respiratory status in asthmatic children: Hints for a locally based preventive strategy AIRE study. Eur
J Epidemiol, 19: 567-576. 089500
Raub JA; Benignus VA. (2002). Carbon monoxide and the nervous system. Neurosci Biobehav Rev, 26: 925-940. 041616
Raven PB; Drinkwater BL; Ruhling RO; Bolduan N; Taguchi S; Gliner J; Horvath SM. (1974). Effect of carbon monoxide
and peroxyacetyl nitrate on man's maximal aerobic capacity. J Appl Physiol, 36: 288-293. 041340
RenX; DorringtonKL; Robbins PA. (2001). Respiratory control in humans after 8 h of lowered arterial PO2,
hemodilution, or carboxyhemoglobinemia. J Appl Physiol, 90: 1189-1195. 193850
Resch H; Zawinka C; Weigert G; Schmetterer L; Garhofer G. (2005). Inhaled carbon monoxide increases retinal and
choroidal blood flow in healthy humans. Invest Ophthalmol Vis Sci, 46: 4275-4280. 193853
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol, 161:
1123-1132.079620
September 2009 5-195 DRAFT - DO NOT CITE OR QUOTE
-------
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of participate air pollution and cardiac arrhythmia
in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riediker M; Williams R; Devlin R; Griggs T; Bromberg P. (2003). Exposure to particulate matter, volatile organic
compounds, and other air pollutants inside patrol cars. Environ Sci Technol, 37: 2084-2093. 043761
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa C;
Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with known
ischemic heart disease in Mexico City. JExpo Sci Environ Epidemiol, 16: 131-137. 156913
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment and
pregnancy outcomes study at the University of California, Los Angeles. Am JEpidemiol, 166: 1045-1052. 096146
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics, 118: 493-
502. 089819
Ritz B; Yu F. (1999). 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. 086976
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern California
between 1989 and 1993. Epidemiology, 11: 502-511. 012068
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in Southern
California. Am J Epidemiol, 155: 17-25. 023227
Rodriguez C; Tonkin R; Heyworth J; Kusel M; De Klerk N; Sly PD; Franklin P; Runnion T; Blockley A; Landau L;
Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young children. Int
J Environ Health Res, 17: 351-360 . 092842
Rosenlund M; Bellander T; Nordquist T; Alfredsson L. (2009). Traffic-generated air pollution and myocardial infarction.
Epidemiology, 20: 265-71. 190309
Rosenlund M; Berglind N; Pershagen G; Hallqvist J; Jonson T; Bellander T. (2006). Long-term exposure to urban air
pollution and myocardial infarction. Epidemiology, 17: 383-390. 089796
Roth RA Jr; Rubin RJ. (1976). Comparison of the effect of carbon monoxide and of hypoxic hypoxia I In vivo metabolism,
distribution and action of hexobarbital. J Pharmacol Exp Ther, 199: 53-60. 012703
Roth RA Jr; Rubin RJ. (1976). Comparison of the effect of carbon monoxide and of hypoxic hypoxia. II. Hexobarbital
metabolism in the isolated, perfused rat liver. J Pharmacol Exp Ther, 199: 61-66. 012420
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
JacqueminB; vonKlot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J;
Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial
infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl R; Ibald-Mulli A; Koenig WSchneider A; Woelke G; Cyrys J; Heinrich J; Marder V; Frampton M; Wichmann HE;
Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients with coronary heart
disease. Environ Health Perspect, 173: 432-441. 088754
Rudez G; JanssenNA; Kilinc E; Leebeek FW; Gerlofs-Nijland ME; Spronk HM; ten Gate H; Cassee FR; de Maat MP.
(2009). Effects of ambient air pollution on hemostasis and inflammation. Environ Health Perspect, 117: 995-1001.
193783
Ryter SW; Alam J; Choi AMK. (2006). Heme oxygenase-1/carbon monoxide: From basic science to therapeutic
applications . Physiol Rev, 86: 583-650. 193765
Salam MT; Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal exposure to
ozone, carbon monoxide, and particulate matter: results from the Children's Health Study. Environ Health Perspect,
113: 1638-1644. 087885
Samet JM; Zeger SL; Dominici F; Curriero F; Coursac I; Dockery DW; Schwartz J; Zanobetti A. (2000). The National
Morbidity, Mortality, and Air Pollution Study. Part II: Morbidity and mortality from air pollution in the United
States. 156939
September 2009 5-196 DRAFT - DO NOT CITE OR QUOTE
-------
Samoli E; Touloumi G; Schwartz J; Anderson HR; Schindler C; Forsberg B; Vigotti MA; Vonk J; Kosnik M; Skorkovsky J;
Katsouyanni K. (2007). Short-term effects of carbon monoxide on mortality: An analysis within the APHEA
project. Environ Health Perspect, 115: 1578-1583. 098420
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Sartiani L; Cerbai E; Lonardo G; DePaoli P; Tattoli M; Cagiano R; Carratu M; Cuomo V; Mugelli A. (2004). Prenatal
exposure to carbon monoxide affects postnatal cellular electrophysiological maturation of the rat heart: a potential
substrate for arrhythmogenesis in infancy. Circulation, 109-: 419-423. 190898
Scharte M; von Ostrowski TA; Daudel F; Freise H; Van Aken H; Bone HG (2006). Endogenous carbon monoxide
production correlates weakly with severity of acute illness. Eur J Anaesthesiol, 23: 117-122. 194115
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG (2006). Ambient air pollution and asthma
exacerbations in children: An eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G; Speizer
FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly subjects. Thorax, 60:
455-461.074317
Schwetz BA; Smith FA; Leong BKJ; Staples RE. (1979). Teratogenic potential of inhaled carbon monoxide in mice and
rabbits. Teratology, 19: 385-391. 011855
Scotto Di Marco G; Kephalopoulos S; Ruuskanen J; Jantunen M. (2005). Personal carbon monoxide exposure in Helsinki,
Finland. Atmos Environ, 39: 2697-2707. 144054
Sheps DS; Adams KF Jr; Bromberg PA; Goldstein GM; O'Neil JJ; Horstman D; Koch G. (1987). Lack of effect of low
levels of carboxyhemoglobin on cardiovascular function in patients with ischemic heart disease. Arch Environ
Occup Health, 42: 108-116. 012212
Sheps DS; Herbst MC; Hinderliter AL; Adams KF; Ekelund LG; O'Neil JJ; Goldstein GM; Bromberg PA; Dalton JL;
Ballenger MN; Davis SM; Koch GG. (1990). Production of arrhythmias by elevated carboxyhemoglobin in patients
with coronary artery disease. Ann Intern Med, 113: 343-351. 013286
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and disease
parameters in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J Allergy Clin
Immunol, 115: 337-344. 087471
Sinclair AH; Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an ambulatory
care setting: 25-month results from the aerosol research and inhalation epidemiological study. J Air Waste Manag
Assoc, 54: 1212-1218. 088696
Singh J. (1986). Early behavioral alterations in mice following prenatal carbon monoxide exposure. Neurotoxicology, 7:
475-481.012827
Singh J. (2003). Gastroschisis is caused by the combination of carbon monoxide and protein-zinc deficiencies in mice.
Birth Defects Res B Dev Reprod Toxicol, 68: 355-362. 053624
Singh J. (2006). Interaction of maternal protein and carbon monoxide on pup mortality in mice: implications for global
infant mortality. Birth Defects Res B Dev Reprod Toxicol, 77: 216-26. 190512
Singh J; Aggison L Jr; Moore-Cheatum L. (1993). Teratogenicity and developmental toxicity of carbon monoxide in
protein-deficient mice. Teratology, 48: 149-159. 013892
Singh J; Scott LH. (1984). Threshold for carbon monoxide induced fetotoxicity. Teratology, 30: 253-257. 011409
Singh J; Smith CB; Moore-Cheatum L. (1992). Additivity of protein deficiency and carbon monoxide on placental
carboxyhemoglobin in mice. Am J Obstet Gynecol, 167: 843-846. 013759
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci Environ Epidemiol,
15: 153-159.073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on symptom
severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346-353. 086294
September 2009 5-197 DRAFT - DO NOT CITE OR QUOTE
-------
Solanki DL; McCurdy PR; Cuttitta FF; Schechter GP. (1988). Hemolysis in sickle cell disease as measured by endogenous
carbon monoxide production: A preliminary report. Am J Clin Pathol, 89: 221-225. 012426
Song R; Mahidhara RS; Liu F; Ning W; Otterbein LE; Choi AMK. (2002). Carbon monoxide inhibits human airway
smooth muscle cell proliferation via mitogen-activated protein kinase pathway. Am J Respir Cell Mol Biol, 27: 603-
610.037531
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution and
inflammation-sensitive biomarkers. Environ Res, 106: 51-61. 188893
Stevens CF; Wang Y. (1993). Reversal of long-term potentiation by inhibitors of haem oxygenase. Nature, 364: 147-149.
188458
Stewart RD; Newton PE; Hosko MJ; Peterson JE. (1973). Effect of carbon monoxide on time perception. Arch Environ
Occup Health, 27: 155-160. 093412
Stewart RD; Peterson JE; Fisher TN; Hosko MJ; Baretta ED; Dodd HC; Herrmann AA. (1973). Experimental human
exposure to high concentrations of carbon monoxide. Arch Environ Occup Health, 26: 1-7. 012428
Stieb DM; Judek S; Burnett RT. (2002). Meta-analysis of time-series studies of air pollution and mortality: effects of gases
and particles and the influence of cause of death, age, and season. J Air Waste Manag Assoc, 52: 470-484. 025205
Stieb DM; Judek S; Burnett RT. (2003). Meta-analysis of time-series studies of air pollution and mortality: Update in
relation to the use of generalized additive models. J Air Waste Manag Assoc, 53: 258-261. 056908
Stockard-Sullivan JE; Korsak RA; Webber DS; Edmond J. (2003). Mild carbon monoxide exposure and auditory function
in the developing rat. JNeurosci Res, 74: 644-654. 190947
Stone JR; Marietta MA. (1994). Soluble guanylate cyclase from bovine lung: Activation with nitric oxide and carbon
monoxide and spectral characterization of the ferrous and ferric states. Biochemistry, 33: 5636-5640. 076455
Storm JE; Fechter LD. (1985). Alteration in the postnatal ontogeny of cerebellar norepinephrine content following chronic
prenatal carbon monoxide. JNeurochem, 45: 965-969. 011653
Storm JE; Fechter LD. (1985). Prenatal carbon monoxide exposure differentially affects postnatal weight and monoamine
concentration of rat brain regions. Toxicol Appl Pharmacol, 81: 139-146. 011652
Storm JE; Valdes JJ; Fechter LD. (1986). Postnatal alterations in cerebellar GABA content, GABA uptake and morphology
following exposure to carbon monoxide early in development. DevNeurosci, 8: 251-261. 012136
Strickland MJ; Klein M; Correa A; Reller MD; Mahle WT; Riehle-Colarusso TJ; Botto LD; Flanders WD; Mulholland JA;
Siffel C; Marcus M; Tolbert PE. (2009). Ambient air pollution and cardiovascular malformations in Atlanta,
Georgia, 1986-2003. Am JEpidemiol, 169: 1004-14. 190324
Stupfel M; Bouley G (1970). Physiological and biochemical effects on rats and mice exposed to small concentrations of
carbon monoxide for long periods. Ann N Y Acad Sci, 174: 342-368. 010557
StykaPE; Penney DG (1978). Regression of carbon monoxide-induced cardiomegaly. Am J Physiol, 235: H516-H522.
011166
Suliman HB; Carraway MS; Tatro LG; Piantadosi CA. (2007). A new activating role for CO in cardiac mitochondrial
biogenesis. J Cell Sci, 120: 299-308. 193768
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine particulate
matter and primary cardiac arrest among persons with and without clinically recognized heart disease. Am J
Epidemiol, 157: 501-509. 043156
Sylvester KP; Patey RA; Rafferty GF; Rees D; Thein SL; Greenough A. (2005). Exhaled carbon monoxide levels in
children with sickle cell disease. Eur J Pediatr, 164: 162-165. 191954
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Szyszkowicz M. (2007). Air pollution and emergency department visits for ischemic heart disease in Montreal, Canada. Int
J Occup Med Environ Health, 20: 167-173. 193793
September 2009 5-198 DRAFT - DO NOT CITE OR QUOTE
-------
S0rhaug S; Steinshamn S; Nilsen OG; Waldum HL. (2006). Chronic inhalation of carbon monoxide: Effects on the
respiratory and cardiovascular system at doses corresponding to tobacco smoking. Toxicology, 228: 280-290.
180414
Tarkiainen TH; Timonen KL; Vanninen EJ; Aim S; Hartikainen JEK; Pekkanen J. (2003). Effect of acute carbon monoxide
exposure on heart rate variability in patients with coronary artery disease. Clin Physiol Funct Imaging, 23: 98-102.
053625
Teran FJ; Johnson RA; Stevenson BK; Peyton KJ; Jackson KE; Appleton SD; Durante W; Johnson FK. (2005). Heme
oxygenase-derived carbon monoxide promotes arteriolar endothelial dysfunction and contributes to salt-induced
hypertension in Dahl salt-sensitive rats. Am J Physiol Regul Integr Comp Physiol, 288: 615-622. 193770
Thorn SR. (1993). Functional inhibition of leukocyte B2 integrins by hyperbaric oxygen in carbon monoxide-mediated
brain injury in rats. Toxicol Appl Pharmacol, 123: 248-256. 013895
Thorn SR; Bhopale VM; Han ST; Clark JM; Hardy KR. (2006). Intravascular neutrophil activation due to carbon monoxide
poisoning. Am J Respir Grit Care Med, 174: 1239-1248. 098418
Thorn SR; Fisher D; Manevich Y. (2001). Roles for platelet-activating factor and *NO-derived oxidants causing neutrophil
adherence after CO poisoning. Am J Physiol Heart Circ Physiol, 281: H923-H930. 193779
Thorn SR; Fisher D; Xu YA; Garner S; Ischiropoulos H. (1999). Role of nitric oxide-derived oxidants in vascular injury
from carbon monoxide in the rat. Am J Physiol, 276: H984-H992. 016753
Thorn SR; Fisher D; Xu YA; Notarfrancesco K; Ischiropoulos H. (2000). Adaptive responses and apoptosis in endothelial
cells exposed to carbon monoxide. ProcNatlAcad SciUSA, 97: 1305-1310. 011574
Thorn SR; Garner S; Fisher D; Ischiropoulos H. (1998). Vascular nitrosative stress from CO exposure. Presented at.
016750
Thorn SR; Ischiropoulos H. (1997). Mechanism of oxidative stress from low levels of carbon monoxide. 085644
Thorn SR; Ohnishi ST; Fisher D; Xu YA; Ischiropoulos H. (1999). Pulmonary vascular stress from carbon monoxide.
Toxicol Appl Pharmacol, 154: 12-19. 016757
Thorn SR; Ohnishi ST; Ischiropoulos H. (1994). Nitric oxide released by platelets inhibits neutrophil B2 integrin function
following acute carbon monoxide poisoning. Toxicol Appl Pharmacol, 128: 105-110. 076459
Thorn SR; Xu YA; Ischiropoulos H. (1997). Vascular endothelial cells generate peroxynitrite in response to carbon
monoxide exposure. Chem Res Toxicol, 10: 1023-1031. 084337
ThomsenHK. (1974). Carbon monoxide-induced atherosclerosis in primates: an electron-microscopic study on the
coronary arteries of Macaca trus monkeys. Atherosclerosis, 20: 233-240. 193781
Thorup C; Jones CL; Gross SS; Moore LC; Goligorsky MS. (1999). Carbon monoxide induces vasodilation and nitric
oxide release but suppresses endothelial NOS. Am J Physiol, 277: F882-F889. 193782
Timonen KL; Pekkanen J; Tiittanen P; Salonen RO. (2002). Effects of air pollution on changes in lung function induced by
exercise in children with chronic respiratory symptoms. Occup Environ Med, 59: 129-134. 025653
Timonen KL; Vanninen E; De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G; Lanki T; Peters A;
Tarkiainen T; Tiittanen P; Kreyling W; Pekkanen J. (2006). Effects of ultrafine and fine particulate and gaseous air
pollution on cardiac autonomic control in subjects with coronary artery disease: The ULTRA study. J Expo Sci
Environ Epidemiol, 16: 332-341.088747
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
Tolcos M; Mallard C; McGregor H; Walker D; Rees S. (2000). Exposure to prenatal carbon monoxide and postnatal
hyperthermia: short and long-term effects on neurochemicals and neuroglia in the developing brain. Exp Neurol,
162: 235-246. 010468
Tolcos M; McGregor H; Walker D; Rees S. (2000). Chronic prenatal exposure to carbon monoxide results in a reduction in
tyrosine hydroxylase-immunoreactivity and an increase in choline acetyltransferase-immunoreactivity in the fetal
medulla: implications for sudden infant death syndrome. J Neuropathol Exp Neurol, 59: 218-228. 015997
Toyoda M; Saito H; Matsuki N. (1996). Nitric oxide but not carbon monoxide is involved in spatial learning of mice. J
Pharmacol Sci, 71: 205-211. 079945
September 2009 5-199 DRAFT - DO NOT CITE OR QUOTE
-------
Tsai S-S; Chen C-C; Hsieh H-J; Chang C-C; Yang C-Y. (2006). Air pollution and postneonatal mortality in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol, 18: 185-189. 090709
Tsai S-S; Goggins WB; Chiu H-F; Yang C-Y. (2003). Evidence for an association between air pollution and daily stroke
admissions in Kaohsiung, Taiwan. Stroke, 34: 2612-2616. 080133
Tschugguel W; Stonek F; Zhegu Z; Dietrich W; Schneeberger C; Stimpfl T; Waldhoer T; Vycudilik W; Huber JC. (2001).
Estrogen increases endothelial carbon monoxide, heme oxygenase 2, and carbon monoxide-derived cGMP by a
receptor-mediated system. J Clin Endocrinol Metab, 86: 3833-3839. 193785
Turner DM; Lee PN; Roe FJC; Gough KJ. (1979). Atherogenesis in the White Carneau pigeon: further studies of the role of
carbon monoxide and dietary cholesterol. Atherosclerosis, 34: 407-417. 012328
U.S. Census Bureau. (2000). Census Bureau projects doubling of nation's population by 2100. Retrieved , from . 157064
U.S. EPA. (2003). Framework for cumulative risk assessment. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Washington, DC. EPA/630/P-02/001F. 192145
U.S. EPA. (2009). Susceptible subpopulations. Retrieved 15-MAY-09, from
http://www.epa.gov/nerl/goals/health/populations.html. 192149
U.S. Census Bureau. (2008). American Housing Survey for the United States: 2007. U.S. Government Printing Office.
Washington, DC. 194013
U.S. EPA. (1978). Altitude as a factor in air pollution. 086321
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 600/P-
99/001F. 000907
U.S. EPA. (2006). Aging and Toxic Response: Issues Relevant to Risk Assessment (Final). U.S. Environmental Protection
Agency. Washington, DC.http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 56648. 192082
Ueda N; Kaushal GP; Hong X; Shah SV. (1998). Role of enhanced ceramide generation in DNA damage and cell death in
chemical hypoxic injury to LLC-PK1 cells. Kidney Int, 54: 399-406. 195136
Vedal S; Brauer M; White R; Petkau J. (2003). Air pollution and daily mortality in a city with low levels of pollution.
Environ HealthPerspect, 111: 45-51. 039044
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with implantable
cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Verma A; Hirsch DJ; Glatt CE; Ronnett GV; Snyder SH. (1993). Carbon monoxide: a putative neural messenger. Science,
259: 381-384. 193999
Vesely AE; Somogyi RB; Sasano H; Sasano N; Fisher JA; Duffin J. (2004). The effects of carbon monoxide on respiratory
chemoreflexes in humans. Environ Res, 94: 227-233. 194000
Villamor E; Perez-Vizcaino F; Cogolludo AL; Conde-Oviedo J; Zaragoza-Arnaez F; Lopez-Lopez JG; Tamargo J. (2000).
Relaxant effects of carbon monoxide compared with nitric oxide in pulmonary and systemic vessels of newborn
piglets. PediatrRes, 48: 546-553. 015838
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of
air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci Environ Epidemiol, 13: 427-
435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Villeneuve PJ; Doiron M-S; Stieb D; Dales R; Burnett RT; Dugandzic R. (2006). Is outdoor air pollution associated with
physician visits for allergic rhinitis among the elderly in Toronto, Canada?. Allergy, 61: 750-758. 091179
Vogel JA; GleserMA. (1972). Effect of carbon monoxide on oxygen transport during exercise. J Appl Physiol, 32: 234-
239. 010898
September 2009 5-200 DRAFT - DO NOT CITE OR QUOTE
-------
Von Klot S; Peters A; Aalto P; Bellander T; Berglind N; D'Ippoliti D; Elosua R; Hermann A; Kulmala M; Lanki T; Lowel
H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible Subpopulations
(HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk of hospital cardiac
readmissions of myocardial infarction survivors in five European cities. Circulation, 112: 3073-3079. 088070
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A. (2002).
Increased asthma medication use in association with ambient fine and ultrafine particles. Eur Respir J, 20: 691-702.
034706
Vreman HJ; Wong RJ; Kadotani T; Stevenson DK. (2005). Determination of carbon monoxide (CO) in rodent tissue: effect
of heme administration and environmental CO exposure. Anal Biochem, 341: 280-289. 193786
Wang R. (1998). Resurgence of carbon monoxide: an endogenous gaseous vasorelaxing factor. Can J Physiol Pharmacol,
76: 1-15.086074
Wang R; Wu L; Wang Z. (1997). The direct effect of carbon monoxide on KCa channels in vascular smooth muscle cells.
Pflugers Arch, 434: 285-291. 084341
Wang T-N; Ko Y-C; Chao Y-Y; Huang C-C; Lin R-S. (1999). Association between indoor and outdoor air pollution and
adolescent asthma from 1995 to 1996 in Taiwan. Environ Res, 81: 239-247. 008105
Weaver LK; Valentine KJ; Hopkins RO. (2007). Carbon monoxide poisoning: risk factors for cognitive sequelae and the
role of hyperbaric oxygen. Am J Respir Grit Care Med, 176: 491 - 497. 193939
Webber D; Lopez I; Korsak R; Hirota S; Acuna D; Edmond J. (2005). Limiting iron availability confers neuroprotection
from chronic mild carbon monoxide exposure in the developing auditory system of the rat. J Neurosci Res, 80: 620-
633.190514
Webber DS; Korsak RA; Sininger LK; Sampogna SL; Edmond J. (2003). Mild carbon monoxide exposure impairs the
developing auditory system of the rat. J Neurosci Res, 74: 655-665. 190515
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac effects of
carbon monoxide and ambient particles in a rat model of myocardial infarction. Toxicol Sci, 80: 367-376. 087874
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal Toxicol, 18: 1077-1082.
156152
Wellenius GA; Saldiva PHN; Batalha JRF; Murthy GGK; Coull BA; Verrier RL; Godleski JJ. (2002). Electrocardiographic
changes during exposure to residual oil fly ash (ROFA) particles in a rat model of myocardial infarction. Toxicol
Sci, 66: 327-335. 025405
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and hemorrhagic
stroke among medicare beneficiaries. Stroke, 36: 2549-2553. 088685
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution on
functional status in patients with chronic congestive heart failure: a repeated-measures study. Environ Health, 6: 1-
7. 092830
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air pollution
and heart rate variability differs for individuals with heart and pulmonary disease. Environ Health Perspect, 114:
560-566. 088453
Wilhelm M; Meng Y; Rull RP. (2008). Environmental Public Health Tracking of Childhood Asthma Using California
Health Interview Survey, Traffic, and Outdoor Air Pollution Data. Environ Health Perspect, 116: 1254-1260.
191912
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in Los Angeles
County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Wittenberg BA; Wittenberg JB. (1993). Effects of carbon monoxide on isolated heart muscle cells. Health Effects Institute.
Cambridge, MA. 013909
September 2009 5-201 DRAFT - DO NOT CITE OR QUOTE
-------
WollmannHA. (1998). Intrauterine growth restriction: definition and etiology. , 49: 1-6. 193812
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-
2002. Environ Health Perspect, 116: 110-5. 098386
Yamada K; Sorimachi A; Wang Q; Yi J; Cheng S; Zhou Y; Sakamoto K. (2008). Abatement of indoor air pollution achieved
with coal-biomass household briquettes. Atmos Environ, 42: 7924-7930. 193232
Yang C-Y; Chen Y-S; Yang C-H; Ho S-C. (2004). Relationship between ambient air pollution and hospital admissions for
cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A, 67: 483-493. 094376
Yang C-Y; Hsieh H-J; Tsai S-S; Wu T-N; Chiu H-F. (2006). Correlation between air pollution and postneonatal mortality in
a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 69: 2033-2040. 090760
Yang CY (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei, Taiwan. J
Toxicol Environ Health A, 71: 1085-1090. 157160
Yang Q; Chen Y; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). Effect of short-term exposure to low levels of
gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environ Res, 99: 99-105. 090184
Yang Q; Chen Y; Shi Y; Burnett RT; McGrail KM; Krewski D. (2003). Association between ozone and respiratory
admissions among children and the elderly in Vancouver, Canada. Inhal Toxicol, 15: 1297-1308. 055621
Yasuda H; Yamaya M; Nakayama K; Ebihara S; Sasaki T; Okinaga S; Inoue D; Asada M; Nemoto M; Sasaki H. (2005).
Increased arterial carboxyhemoglobin concentrations in chronic obstructive pulmonary disease. Am J Respir Grit
CareMed, 171: 1246-1251. 191953
Yoshimura S; Banno Y; Nakashima S; Hayashi K; Yamakawa H; Sawada M; Sakai N; Nozawa Y. (1999). Inhibition of
neutral sphingomyelinase activation and ceramide formation by glutathione in hypoxic PC 12 cell death. J
Neurochem, 73: 675-683. 195135
Yu O; Sheppard L; Lumley T; Koenig JQ; S. (2000). Effects of ambient air pollution on symptoms of asthma in Seattle-
area children enrolled in the CAMP study. Environ Health Perspect, 108: 1209-1214. 013254
Yu S; Zhang R; Chen Y; Chen J; Zhu C; Tang J; Li G (2008). Possible contribution of endogenous carbon monoxide to the
development of allergic rhinitis in guinea pigs. J Inflamm, 5: -. 192384
Zamudio S; Palmer SK; Dahms TE; Berman JC; Young DA; Moore LG (1995). Alterations in uteroplacental blood flow
precede hypertension in preeclampsia at high altitude. J Appl Physiol, 79: 15-22. 193908
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold DR.
(2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. , 110: 2184-2189. 087489
Zanobetti A; Schwartz J. (2001). Are diabetics more susceptible to the health effects of airborne particles?. Am J Respir
Grit Care Med, 164: 831-833. 016710
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health,
60: 890-895. 090195
Zenclussen ML; Anegon I; Bertoja AZ; Chauveau C; Vogt K; Gerlof K; Sollwedel A; Volk HD; Ritter T; Zenclussen AC.
(2006). Over-expression of heme oxygenase-1 by adenoviral gene transfer improves pregnancy outcome in a
murine model of abortion. Am J Reprod Immunol, 69: 35-52. 193873
Zevin S; Saunders S; Gourlay SG; Jacob PHI; BenowitzNL. (2001). Cardiovascular effects of carbon monoxide and
cigarette smoking. J Am Coll Cardiol, 38: 1633-1638. 021120
Zhang X; Shan P; Alam J; Fu XY; Lee PJ. (2005). Carbon monoxide differentially modulates STAT1 and STAT3 and
inhibits apoptosis via a phosphatidylinositol 3-kinase/akt and p38 kinase-dependent STAT3 pathway during anoxia-
reoxygenation injury. J Biol Chem, 280: 8714-8721. 184460
Zhao H; Wong RJ; Doyle TC; Nayak N; Vreman HJ; Contag CH; Stevenson DK. (2008). Regulation of maternal and fetal
hemodynamics by heme oxygenase in mice. Biol Reprod, 78: 744-751. 193883
Zhu Y; Hinds WC; Kim S; Shen S; Sioutas C. (2002). Study of ultrafine particles near a major highway with heavy-duty
diesel traffic. Atmos Environ, 36: 4323-4335. 041553
Zhuo M; Small SA; Kandel ER; Hawkins RD. (1993). Nitric oxide and carbon monoxide produce activity-dependent long-
term synaptic enhancement in hippocampus. , 260: 1946-1950. 013905
September 2009 5-202 DRAFT - DO NOT CITE OR QUOTE
-------
Zinkham WH; Houtchens RA; Caughey WS. (1980). Carboxyhemoglobin levels in an unstable hemoglobin disorder (Hb
Zurich): effect onphenotypic expression. Science, 209: 406-408. 011435
Zuckerbraun BS; Chin BY; Bilban M; d'Avila JC; Rao J; Billiar TR; Otterbein LE. (2007). Carbon monoxide signals via
inhibition of cytochrome c oxidase and generation of mitochondrial reactive oxygen species. , 21: 1099-1106.
193884
September 2009 5-203 DRAFT - DO NOT CITE OR QUOTE
-------
Annex A. Atmospheric Science
Cerbco Monocode
Alaska State Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
Emissions (Tons)
Yukon-Koyukuk County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
Figure A-1 CO emissions density map and distribution for the state of Alaska and for Yukon-
Koyukuk County in Alaska.
September 2009
A-1
DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Morx^ide Emissicns in 2QQ(? Cfons per Squere Mile)
Utah State Emissions 2002
] 159 - 5.90
16.97 - B.34
I "B.56 - 396.50
Weber County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
H 32,365
H 42,466
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
12,079
Emissions (Tons)
Emissions (Tons)
Utah County Emissions 2002
Grand County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
|2,439
j 2,048
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Emissions (Tons)
Emissions (Tons)
Figure A-2 CO emissions density map and distribution for the state of Utah and for selected
counties in Utah.
September 2009
A-2
DRAFT - DO NOT CITE OR QUOTE
-------
Cwbon M<*XMdde Emissions *i 2002 (Tons per Square Mile)
Massachusetts State Emissions 2002
49.37 - 117.27
139,75 - 198,59
252.78 - 225154
Middlesex County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
H 926,079
H 490,378
] 28,111
|11,569
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
II 7,063
H 101,247
Emissions (Tons)
Norfolk County Emissions 2002
Emissions (Tons)
Suffolk County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
H 8,803
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
Emissions (Tons)
Emissions (Tons)
Figure A-3 CO emissions density map and distribution for the state of Massachusetts and
for selected counties in Massachusetts.
September 2009
A-3
DRAFT - DO NOT CITE OR QUOTE
-------
Carbon Monoride Emissions in 2002 (Tens per Squ^e Mte)
14.62 - 22.32
122.43-49.92
15111- 603.87
Georgia State Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
Fulton County Emissions 2002
Dekalb County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Emissions (Tons)
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Emissions (Tons)
Figure A-4 CO emissions density map and distribution for the state of Georgia and for
selected counties in Georgia (1 of 2).
September 2009
A-4
DRAFT - DO NOT CITE OR QUOTE
-------
Chatham County Emissions 2002
Liberty County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
HI 25,9(
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
]4,832
Emissions (Tons)
Emissions (Tons)
Glynn County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
lesidential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
] 6,619
Emissions (Tons)
Figure A-5 CO emissions distribution for selected counties in Georgia (2 of 2).
September 2009
A-5
DRAFT - DO NOT CITE OR QUOTE
-------
San Diego
Carb
-------
Carbon Monoxide &nrsacns in 2002 (Tons per Square M*e)
1907 - 2115
12164 - 3742
I40L58-3Q122
Alabama State Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Sol vent Use
J 277,465
1375.49
1106.5:
Jefferson County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Sol vent Use
Figure A-7 CO emissions density map and distribution for the state of Alabama and for
Jefferson County in Alabama.
September 2009
A-7
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-1 Listing of all carbon monoxide monitors currently in use, along with their limits of
detection.
Method Code
008
012
018
033
041
048
050
051
054
055
066
067
088
093
106
108
147
158
167
172
554
588
593
Method Description
BENDIX8501-5CA
BECKMAN 866
MSA 202S
HORIBAAQM-10-11-12
MONITOR LABS 8310
HORIBA300E/300SE
MASS-CO 1 (MASSACHUSETTS)
DASIBI 3003
THERMO ELECTRON 48, 48C
Gas Filter Correlation Thermo Electron 48C-TL
MONITOR LABS 8830
DASIBI 3008
LEAR SIEGLER MODEL ML 9830
API MODEL 300 GAS FILTER
HORIBAINSTR. MODEL APMA-360
ENVIRONMENT SA MODEL C011 M
Environnement S.A. Model C012M Co Analyzer
HORIBAINSTR. MODEL APMA-370
DKK-TOA Cork Mode GFC-311E
SIR S.A> Model S5006
Gas Filter Correlation Thermo Electron 48C-TLE
Ecotech EC9830T
API Model 300 EU
Reference Method Id
RFCA-0276-008
RFCA-0876-012
RFCA-0177-018
RFCA-1 278-033
RFCA-0979-041
RFCA-1 180-048
RFCA-1 280-050
RFCA-0381-051
RFCA-0981-054
RFCA-0388-066
RFCA-0488-067
RFCA-0992-088
RFCA-1 093-093
RFCA-0895-106
RFCA-0995-108
RFCA-0206-147
RFCA-0506-158
RFCA-0907-167
RFCA-0708-172
RFCA-0992-088
RFCA-1 093-093
Fed MDL (ppm)
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.02000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.02000
0.02000
0.02000
September 2009
A-8
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-2 Microscale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes that
the value was not reported.
Monitor Code
2-90-2-42101-1
4-13-16-42101-1
4-19-1014-42101-1
6-65-1003-42101-1
6-73-7-42101-1
8-13-9-42101-1
8-31-2-42101-2
8-31-19-42101-1
8-41-15-42101-1
8-77-18-42101-1
9-3-17-42101-1
11-1-23-42101-1
12-57-1070-42101-1
12-86-4002-42101-1
12-95-1005-42101-1
12-103-24-42101-1
12-103-2008-42101-1
12-115-1004-42101-1
13-121-99-42101-1
17-31-63-42101-1
17-31-6004-42101-1
17-143-36-42101-1
17-167-8-42101-1
17-201-11-42101-1
18-3-11-42101-1
18-89-15-42101-1
18-97-72-42101-1
18-163-19-42101-1
21-111-1019-42101-1
27-53-954-42101-1
27-123-50-42101-1
27-137-18-42101-1
27-145-3048-42101-1
30-29-10-42101-1
30-31-13-42101-1
33-11-1009-42101-1
34-5-1001-42101-1
34-17-1002-42101-1
37-67-23-42101-1
39-35-48-42101-1
State_Name
Alaska
Arizona
Arizona
California
California
Colorado
Colorado
Colorado
Colorado
Colorado
Connecticut
District Of Columbia
Florida
Florida
Florida
Florida
Florida
Florida
Georgia
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Kentucky
Minnesota
Minnesota
Minnesota
Minnesota
Montana
Montana
New Hampshire
New Jersey
New Jersey
North Carolina
Ohio
City_Name
Fairbanks
Phoenix
Tucson
Riverside
San Diego
Longmont
Denver
Denver
Colorado Springs
Grand Junction
Hartford
Washington
Tampa
Miami
Orlando
Saint Petersburg
Clearwater
Sarasota
Atlanta
Chicago
Maywood
Peoria
Springfield
Rockford
Fort Wayne
East Chicago
Indianapolis
Evansville
Louisville
Minneapolis
St. Paul
Duluth
St. Cloud
Kalispell
Not in a city
Nashua
Burlington
Jersey City
Winston-Salem
Cleveland
Traffic_Count
NR
50000
41200
40000
6000
20000
17200
500
44200
13525
10000
30000
133855
5000
30000
35000
67751
31000
44000
5000
NR
18500
16400
11400
30430
NR
21237
24498
22000
29352
NR
12000
NR
NR
2000
40000
8000
25000
22000
24300
Type_Road
NR
ARTERIAL
MAJSTORHY
FREEWAY
THRUST OR HY
MAJSTORHY
MAJSTORHY
MAJSTORHY
MAJSTORHY
THRUST OR HY
THRUST OR HY
THRUST OR HY
ARTERIAL
LOCAL ST OR HY
MAJSTORHY
MAJSTORHY
MAJSTORHY
MAJSTORHY
MAJSTORHY
LOCAL ST OR HY
NR
ARTERIAL
MAJSTORHY
ARTERIAL
MAJSTORHY
NR
MAJSTORHY
LOCAL ST OR HY
MAJ ST OR HY
MAJSTORHY
NR
MAJSTORHY
NR
THRUST OR HY
THRUST OR HY
MAJSTORHY
THRUST OR HY
THRUST OR HY
MAJSTORHY
THRUST OR HY
September 2009
A-9
DRAFT - DO NOT CITE OR QUOTE
-------
Monitor Code
State Name
City_Name
Traffic Count
Type_Road
39-35-51-42101-1
Ohio
Cleveland
16150
MAJ STORMY
39-35-53-42101-1
Ohio
Cleveland
19550
MAJ STORMY
39-49-36-42101-1
Ohio
Columbus
16800
MAJ STORMY
39-61-21-42101-1
Ohio
Cincinnati
17250
LOCAL STORMY
39-85-6-42101-1
Ohio
Mentor
25240
MAJ ST OR MY
39-113-34-42101-1
Ohio
Dayton
7100
THRU STORMY
39-153-22-42101-1
Ohio
Akron
13150
MAJ STORMY
41-29-18-42101-1
Oregon
Medford
NR
NR
41-39-13-42101-1
Oregon
Eugene
17500
MAJ STORMY
41-51-87-42101-1
Oregon
Portland
4150
LOCAL STORMY
45-79-20-42101-1
South Carolina
Columbia
31500
MAJ ST OR MY
47-37-21-42101-1
Tennessee
Nashville
15000
MAJ ST OR MY
47-157-36-42101-1
Tennessee
Memphis
25000
THRU STORMY
48-29-46-42101-1
Texas
San Antonio
5820
MAJ STORMY
48-201-75-42101-1
Texas
Houston
6576
LOCAL ST OR HY
53-33-19-42101-1
Washington
Bellevue
100000
MAJSTORHY
53-63-49-42101-1
Washington
Spokane
10000
MAJSTORHY
September 2009
A-10
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-3 Middle scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes
that the value was not reported.
Monitor Code
4-13-3010-42101-1
6-29-10-42101-1
6-37-1301-42101-1
6-37-9033-42101-1
6-59-1003-42101-1
6-71-9004-42101-1
6-85-5-42101-1
12-11-10-42101-1
12-31-80-42101-1
12-31-84-42101-1
12-99-1004-42101-1
12-103-2006-42101-1
17-31-3103-42101-1
20-209-21-42101-1
24-510-40-42101-1
32-31-22-42101-1
34-3-4-42101-1
36-61-56-42101-1
39-49-5-42101-1
39-81-1001-42101-1
39-151-20-42101-1
40-143-191-42101-1
42-3-38-42101-1
42-101-47-42101-1
45-19-46-42101-1
45-45-8-42101-1
45-45-9-42101-1
47-163-7-42101-1
48-439-1002-42101-1
50-7-14-42101-1
72-127-3-42101-1
State_Name
Arizona
California
California
California
California
California
California
Florida
Florida
Florida
Florida
Florida
Illinois
Kansas
Maryland
Nevada
New Jersey
New York
Ohio
Ohio
Ohio
Oklahoma
Pennsylvania
Pennsylvania
South Carolina
South Carolina
South Carolina
Tennessee
Texas
Vermont
Puerto Rico
City_Name
Phoenix
Bakersfield
Lynwood
Lancaster
Costa Mesa
San Bernardino
San Jose
Fort Lauderdale
Jacksonville
Jacksonville
Palm Beach
Clearwater
Schiller Park
Kansas City
Baltimore
Reno
Fort Lee
New York
Columbus
Mingo Junction
Canton
Tulsa
Pittsburgh
Philadelphia
Not in a city
Greenville
Taylors
Kingsport
Fort Worth
Burlington
San Juan
Traffic_Count
18500
30300
35000
2320
1000
21900
NR
1000
1000
500
30000
23400
47900
7720
15300
NR
250000
45000
36600
2500
11000
50800
15000
NR
NR
NR
9500
NR
100
NR
64000
Type_Road
ARTERIAL
ARTERIAL
ARTERIAL
LOCAL ST OR HY
LOCAL ST OR HY
THRU STORMY
LOCAL STORMY
LOCAL STORMY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
ARTERIAL
MAJ ST OR HY
THRUSTORHY
NR
ARTERIAL
MAJ ST OR HY
FREEWAY
LOCAL ST OR HY
MAJ ST OR HY
FREEWAY
MAJ ST OR HY
NR
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
NR
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
September 2009
A-11
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-4 Neighborhood scale monitors meeting 75% completeness criteria, 2005-2007. "NR"
denotes that the value was not reported.
Monitor Code
1-73-1003-42101-1
1-73-6004-42101-1
2-20-18-42101-1
2-20-48-42101-1
2-90-20-42101-1
4-13-19-42101-1
4-13-3002-42101-1
4-19-2-42101-1
4-19-1011-42101-1
4-19-1028-42101-1
6-1-1001-42101-1
6-13-2-42101-1
6-37-5005-42101-1
6-53-1003-42101-1
6-65-9001-42101-1
6-67-7-42101-1
6-73-1-42101-1
6-73-1002-42101-1
6-73-2007-42101-1
6-83-1025-42101-1
6-83-2004-42101-1
6-83-2011-42101-1
6-83-4003-42101-1
8-1-3001-42101-1
8-67-7001-42101-1
8-69-1004-42101-1
8-123-10-42101-1
11-1-41-42101-1
12-11-2004-42101-1
12-11-3002-42101-1
12-31-83-42101-1
12-86-31-42101-1
12-86-1019-42101-1
12-95-2002-42101-1
12-103-18-42101-1
17-31-4002-42101-1
17-163-10-42101-1
18-97-73-42101-1
20-173-10-42101-1
21-111-46-42101-1
State_Name
Alabama
Alabama
Alaska
Alaska
Alaska
Arizona
Arizona
Arizona
Arizona
Arizona
California
California
California
California
California
California
California
California
California
California
California
California
California
Colorado
Colorado
Colorado
Colorado
District Of Columbia
Florida
Florida
Florida
Florida
Florida
Florida
Florida
Illinois
Illinois
Indiana
Kansas
Kentucky
City_Name
Fairfield
Birmingham
Anchorage
Anchorage
Fairbanks
Phoenix
Phoenix
Tucson
Tucson
Tucson
Fremont (Centerville)
Concord
Los Angeles
Salinas
Lake Elsinore
Sacramento
Chula Vista
Escondido
Otay Mesa
Capitan
Lompoc
Goleta
Vandenberg Air Force Base
Welby
Not in a city
Fort Collins
Greeley
Washington
Pompano Beach
Hollywood
Jacksonville
Miami
Miami
Winter Park
Saint Petersburg
Cicero
East Saint Louis
Indianapolis (Remainder)
Wichita
Louisville
Traffic_Count
5000
NR
NR
5000
NR
NR
24000
37400
47000
52900
500
41218
1252
33193
NR
20000
5000
NR
18000
NR
NR
5000
NR
500
2436
5000
6650
540
1000
1000
10000
62000
8000
7000
2000
NR
8900
11261
6884
6500
Type_Road
LOCAL STORMY
NR
NR
LOCAL STORMY
NR
LOCAL STORMY
ARTERIAL
MAJ ST OR MY
MAJ ST OR MY
MAJ ST OR MY
LOCAL STORMY
MAJ ST OR MY
LOCAL ST OR MY
THRU STORMY
NR
THRU STORMY
LOCAL STORMY
NR
LOCAL ST OR HY
NR
NR
THRUSTORHY
NR
EXPRESSWAY
LOCAL ST OR HY
THRUSTORHY
THRUSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
NR
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
THRUSTORHY
September 2009
A-12
DRAFT - DO NOT CITE OR QUOTE
-------
Monitor Code
22-33-9-42101-1
25-13-16-42101-1
25-17-7-42101-1
25-25-42-42101-1
27-3-600-42101-1
27-37-20-42101-1
27-37-423-42101-1
29-510-86-42101-1
30-111-85-42101-1
31-55-35-42101-1
32-3-538-42101-1
32-3-539-42101-1
32-3-561-42101-1
32-3-1021-42101-1
32-3-2002-42101-1
32-31-16-42101-1
32-31-20-42101-1
32-31-25-42101-1
32-31-1005-42101-1
32-31-2009-42101-1
32-510-4-42101-1
33-11-20-42101-1
34-3-5001-42101-1
34-7-3-42101-1
35-1-19-42101-1
35-1-23-42101-1
35-1-24-42101-1
35-1-28-42101-1
35-1-1014-42101-1
35-43-9004-42101-1
36-63-2008-42101-1
37-119-41-42101-1
37-119-41-42101-3
39-35-70-42101-1
39-113-28-42101-1
39-153-20-42101-1
40-21-9002-42101-1
40-71-9010-42101-1
40-109-47-42101-1
41-51-80-42101-1
42-3-31-42101-1
42-13-801-42101-1
42-17-12-42101-1
State_Name
Louisiana
Massachusetts
Massachusetts
Massachusetts
Minnesota
Minnesota
Minnesota
Missouri
Montana
Nebraska
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
New Hampshire
New Jersey
New Jersey
New Mexico
New Mexico
New Mexico
New Mexico
New Mexico
New Mexico
New York
North Carolina
North Carolina
Ohio
Ohio
Ohio
Oklahoma
Oklahoma
Oklahoma
Oregon
Pennsylvania
Pennsylvania
Pennsylvania
City_Name
Baton Rouge
Springfield
Lowell
Boston
Fridley
Rosemount
Inver Grove Heights (RR name Inver Grove)
St. Louis
Billings
Omaha
Las Vegas
Las Vegas
Las Vegas
Las Vegas
Las Vegas
Reno
Reno
Reno
Sparks
Lemmon Valley-Golden Valley
Carson City
Manchester
Hackensack
Camden
Albuquerque
Albuquerque
Albuquerque
Albuquerque
Albuquerque
Not in a city
Niagara Falls
Charlotte
Charlotte
Cleveland
Dayton
Akron
Park Hill
Not in a city
Oklahoma City
Portland
Pittsburgh
Altoona
Bristol
Traffic_Count
5000
5000
15000
12785
1400
NR
NR
81850
5700
2900
20000
21000
28400
NR
6750
22700
NR
NR
2600
NR
1
500
15000
45000
1
41200
15500
20600
8000
100
5000
16400
16400
100
5100
200
10300
300
27000
5000
4562
100
500
Type_Road
LOCALS! OR HY
LOCAL ST OR HY
THRUST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
NR
NR
MAJ ST OR HY
THRUST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
NR
THRUSTORHY
LOCAL ST OR HY
NR
NR
LOCAL ST OR HY
NR
LOCAL ST OR HY
LOCAL ST OR HY
THRUSTORHY
MAJ ST OR HY
ARTERIAL
MAJ ST OR HY
MAJ ST OR HY
THRUSTORHY
THRUSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
September 2009
A-13
DRAFT - DO NOT CITE OR QUOTE
-------
Monitor Code
42-21-11-42101-1
42-49-3-42101-1
42-71-7-42101-1
42-73-15-42101-1
42-91-13-42101-1
42-95-25-42101-1
42-101-4-42101-1
42-101-27-42101-1
42-107-3-42101-1
42-125-5-42101-1
44-7-1010-42101-1
48-61-6-42101-1
48-113-69-42101-2
48-141-2-42101-1
48-141-29-42101-1
48-141-37-42101-1
48-141-44-42101-1
48-141-53-42101-1
48-141-57-42101-1
48-141-58-42101-1
48-201-24-42101-1
48-201-47-42101-1
48-201-1035-42101-1
48-201-1039-42101-1
48-439-3011-42101-1
48-453-14-42101-1
48-479-17-42101-1
49-35-3-42101-1
50-21-2-42101-1
51-59-5-42101-1
51-650-4-42101-2
51-760-24-42101-1
51-770-15-42101-1
54-9-11-42101-1
54-29-9-42101-1
54-29-1004-42101-1
State_Name
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Rhode Island
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Utah
Vermont
Virginia
Virginia
Virginia
Virginia
West Virginia
West Virginia
West Virginia
City_Name
Johnstown
Erie
Lancaster
New Castle
Norristown
Freemansburg
Philadelphia
Philadelphia
Shenandoah
Charleroi
East Providence
Brownsville
Dallas
El Paso
El Paso
El Paso
El Paso
El Paso
Socorro
El Paso
Not in a city
Houston
Houston
Deer Park
Arlington
Austin
Laredo
Not in a city
Rutland
Not in a city
Hampton
Richmond
Roanoke
Weirton
Weirton
Weirton
Traffic_Count
6000
1000
2000
4500
8500
100
13800
46000
100
NR
100000
30
1000
7270
2790
5000
15200
1992
500
1080
5300
5860
13440
16010
10573
3420
30380
16500
NR
25
2000
7591
NR
NR
NR
50
Type_Road
LOCAL ST OR HY
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
NR
FREEWAY
LOCAL ST OR HY
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
ARTERIAL
FREEWAY
LOCAL ST OR HY
LOCAL ST OR HY
MAJSTORHY
MAJ ST OR HY
MAJ ST OR HY
MAJSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
ARTERIAL
THRUSTORHY
NR
LOCAL ST OR HY
LOCAL ST OR HY
THRUSTORHY
NR
NR
NR
LOCAL ST OR HY
September 2009
A-14
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-5 Urban scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes that
the value was not reported.
Monitor Code
6-59-7-42101-1
13-89-2-42101-1
13-223-3-42101-1
25-27-23-42101-1
34-7-1001-42101-1
42-3-10-42101-1
42-7-14-42101-1
42-129-8-42101-1
42-133-8-42101-1
48-141-55-42101-1
51-59-30-42101-1
State_Name
California
Georgia
Georgia
Massachusetts
New Jersey
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Texas
Virginia
City_Name
Anaheim
Decatur
Not in a city
Worcester
Not in a city
Pittsburgh
Beaver Falls
Greensburg
York
El Paso
Franconia
Traffic_Count
1000
9250
6
NR
4000
1000
NR
100
8400
2450
200
Type_Road
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
THRU ST OR MY
MAJ STORMY
NR
THRU ST OR MY
THRUST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
Table A-6 Regional scale monitors meeting 75% completeness criteria, 2005-2007. "NR" denotes
that the value was not reported.
Monitor Code
State Name
City_Name
Traffic Count
Type_Road
23-9-103-42101-1
Maine
Not in a city
3500
LOCAL ST OR HY
35-1-29-42101-1
New Mexico
South Valley
8800
LOCAL ST OR HY
September 2009
A-15
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-7 Monitors meeting 75% completeness criteria, 2005-2007 with no scale delared. "NR"
denotes that the value was not reported.
Monitor Code
4-13-9997-42101-1
6-1-7-42101-1
6-7-2-42101-1
6-13-1002-42101-1
6-13-1004-42101-1
6-13-3001-42101-1
6-19-7-42101-1
6-19-8-42101-1
6-19-242-42101-1
6-19-5001-42101-1
6-25-5-42101-1
6-25-6-42101-1
6-25-1003-42101-1
6-37-2-42101-1
6-37-113-42101-1
6-37-1002-42101-1
6-37-1103-42101-1
6-37-1201-42101-1
6-37-1701-42101-1
6-37-2005-42101-1
6-37-4002-42101-1
6-37-6012-42101-1
6-41-1-42101-1
6-45-8-42101-1
6-45-9-42101-1
6-55-3-42101-1
6-59-2022-42101-1
6-59-5001-42101-1
6-65-5001-42101-1
6-65-8001-42101-1
6-67-2-42101-1
6-67-6-42101-1
6-67-13-42101-1
6-71-1-42101-1
6-71-306-42101-1
6-71-1004-42101-1
6-75-5-42101-1
6-77-1002-42101-1
6-81-1001-42101-1
6-87-3-42101-1
State_Name
Arizona
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
City_Name
Phoenix
Livermore
Chico
Bethel Island
San Pablo
Pittsburg
Fresno
Fresno
Fresno
Clovis
Calexico
Calexico
El Centre
Azusa
West Los Angeles
Burbank
Los Angeles
Reseda
Pomona
Pasadena
Long Beach
Santa Clarita
San Rafael
Ukiah
Willits
Napa
Mission Viejo
La Habra
Palm Springs
Rubidoux (West Riverside)
North Highlands
Sacramento
Sacramento
Barstow
Victorville
Upland
San Francisco
Stockton
Redwood City
Davenport
Traffic_Count
250
2400
44000
NR
NR
9600
500
20000
500
16461
7000
10
NR
600
NR
2400
9000
NR
NR
18000
24000
4395
15000
12000
18000
NR
42400
NR
NR
18000
NR
10000
100
NR
454
15000
240700
6000
1000
NR
Type_Road
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
NR
THRU STORMY
THRU STORMY
LOCAL STORMY
MAJSTORHY
LOCAL ST OR HY
THRUST OR HY
LOCAL ST OR HY
THRUSTORHY
NR
THRUSTORHY
NR
LOCAL ST OR HY
THRUSTORHY
NR
NR
THRUSTORHY
LOCAL ST OR HY
LOCAL ST OR HY
MAJSTORHY
LOCAL ST OR HY
MAJSTORHY
NR
MAJSTORHY
NR
NR
THRUSTORHY
NR
LOCAL ST OR HY
LOCAL ST OR HY
NR
LOCAL ST OR HY
THRUSTORHY
FREEWAY
LOCAL ST OR HY
LOCAL ST OR HY
NR
September 2009
A-16
DRAFT - DO NOT CITE OR QUOTE
-------
Monitor Code
6-95-4-42101-1
6-97-3-42101-1
6-99-5-42101-1
6-99-6-42101-1
9-3-1003-42101-1
10-3-1008-42101-1
10-3-2004-42101-1
15-3-10-42101-1
18-63-2-42101-1
25-25-2-42101-1
29-77-32-42101-1
29-189-4-42101-1
30-13-1-42101-1
31-109-18-42101-1
34-23-2003-42101-1
34-25-2001-42101-1
34-27-3-42101-1
36-1-12-42101-1
36-29-5-42101-1
36-55-1007-42101-1
36-67-17-42101-1
36-81-124-42101-1
36-93-3-42101-1
36-103-9-42101-2
48-479-16-42101-1
49-57-6-42101-1
51-13-20-42101-1
51-59-1005-42101-1
51-59-5001-42101-1
51-510-9-42101-1
56-39-1012-42101-1
State_Name
California
California
California
California
Connecticut
Delaware
Delaware
Hawaii
Indiana
Massachusetts
Missouri
Missouri
Montana
Nebraska
New Jersey
New Jersey
New Jersey
New York
New York
New York
New York
New York
New York
New York
Texas
Utah
Virginia
Virginia
Virginia
Virginia
Wyoming
City_Name
Vallejo
Santa Rosa
Modesto
Turlock
East Hartford
Not in a city
Wilmington
Ewa Beach
Pittsboro
Boston
Springfield
Sunset Hills
Great Falls
Lincoln
Perth Amboy
Freehold
Morristown
Albany
Buffalo
Rochester
Syracuse
New York
Schenectady
Holtsville
Laredo
Ogden
Not in a city
Annandale
McLean
Alexandria
Not in a city
Traffic_Count
9350
2608
NR
500
800
NR
28046
NR
500
35000
1000
33300
26155
NR
14000
NR
NR
12000
26000
NR
NR
10000
37000
10000
16180
38000
6000
24000
36845
3974
NR
Type_Road
THRUST OR HY
THRUSTORHY
NR
LOCAL ST OR HY
LOCAL ST OR HY
NR
MAJSTORHY
NR
LOCAL ST OR HY
MAJSTORHY
LOCAL ST OR HY
MAJSTORHY
MAJSTORHY
NR
LOCAL ST OR HY
NR
NR
MAJSTORHY
ARTERIAL
NR
NR
EXPRESSWAY
EXPRESSWAY
THRUSTORHY
MAJSTORHY
ARTERIAL
MAJSTORHY
MAJSTORHY
MAJSTORHY
LOCAL ST OR HY
NR
September 2009
A-17
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-8
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Numbers of high LOD and trace-level monitors in each state
criteria for 2005-2007.
State Number of high LOD monitors
2
4
9
0
65
9
2
2
2
18
3
1
0
8
6
0
2
2
0
0
1
4
0
7
0
3
4
2
12
2
9
7
9
2
0
14
4
3
19
1
that met completeness
Number of trace-level monitors
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
September 2009
A-18
DRAFT - DO NOT CITE OR QUOTE
-------
State
Number of high LOD monitors Number of trace-level monitors
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
19
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
September 2009
A-19
DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage Core Based Statisical Area
Figure A-8 Map of CO monitor locations with respect to population density in the Anchorage
CBSA, total population.
Anchorage Core Based Statisical Area
MOS Populjhoo DMHMg
Population 2 *S p*r 26 $q Km
Figure A-9 Map of CO monitor locations with respect to population density in the Anchorage
CBSA, ages 65 and older.
September 2009
A-20
DRAFT - DO NOT CITE OR QUOTE
-------
Atlanta Combined StatisicalArea
Figure A-10 Map of CO monitor locations with respect to population density in the Atlanta
CSA, total population.
Atlanta Combined StatisicalArea
0 5 10 X 30
29H PopuUbon Den**y
| | MhM» CO U»n*Mi IS km button
Population >ft5 per 2.fl Sq Km
••
Figure A-11 Map of CO monitor locations with respect to population density in the Atlanta
CSA, ages 65 and older.
September 2009
A-21
DRAFT - DO NOT CITE OR QUOTE
-------
Boston Combined StatisicalArea
0 1S30 90 90 120
Figure A-12 Map of CO monitor locations with respect to population density in the Boston
CSA, total population.
Boston Combined StatisicalArea
0 1S» 60 90 120
Figure A-13 Map of CO monitor locations with respect to population density in the Boston
CSA, ages 65 and older.
September 2009
A-22
DRAFT - DO NOT CITE OR QUOTE
-------
Houston Combined Statisical Area
zoos popubtton D*R» «y
[ I ™tan CO !**-**.(!»«
np«f IS SqKm
Figure A-14 Map of CO monitor locations with respect to population density in the Houston
CSA, total population.
September 2009
A-23
DRAFT - DO NOT CITE OR QUOTE
-------
Houston Combined Statisical Area
0 3 50 ICQ 1$0 200
Figure A-15 Map of CO monitor locations with respect to population density in the Houston
CSA, ages 65 and older.
New York Combined Statisical Area
Figure A-16 Map of CO monitor locations with respect to population density in the New York
City CSA, total population.
September 2009
A-24
DRAFT - DO NOT CITE OR QUOTE
-------
New York Combined Statisical Area
200* PopuUEion Denx.tr
| | '« rnrtCO W(nioi [S *m 6U
P*(mUt«>n 2 SS firt 38 Sn Km
Figure A-17 Map of CO monitor locations with respect to population density in the New York
City CSA, ages 65 and older.
Phoenix Core Based Statisical Area
ISW-3W7
3679 • IMO*
o »« » no m
Figure A-18 Map of CO monitor locations with respect to population density in the Phoenix
CSA, total population.
September 2009
A-25
DRAFT - DO NOT CITE OR QUOTE
-------
Phoenix Core Based Statisical Area
0 30 40 BO 120 160
Figure A-19 Map of CO monitor locations with respect to population density in the Phoenix
CSA, ages 65 and older.
September 2009
A-26
DRAFT - DO NOT CITE OR QUOTE
-------
Pittsburgh Combined Statisical Area
0 5 ID X 30 -«
2005 Population Density
( j PHi6g(*>C9Moulin6kmMhrJ
Populaltcn ptr Sq Km
Figure A-20 Map of CO monitor locations with respect to population density in the Pittsburgh
CSA, total population.
Pittsburgh Combined Statisical Area
0 15 30 tO SO 1»
Figure A-21 Map of CO monitor locations with respect to population density in the Pittsburgh
CSA, ages 65 and older.
September 2009
A-27
DRAFT - DO NOT CITE OR QUOTE
-------
Seattle Combined Statisical Area
0 1530 flO 90 130
Figure A-22 Map of CO monitor locations with respect to population density in the Seattle
CSA, total population.
Seattle Combined Statisical Area
Figure A-23 Map of CO monitor locations with respect to population density in the Seattle
CSA, ages 65 and older.
September 2009
A-28
DRAFT - DO NOT CITE OR QUOTE
-------
St. Louis Combined Statisical Area
Figure A-24 Map of CO monitor locations with respect to population density in the St. Louis
CSA, total population.
St. Louis Combined Statisical Area
Figure A-25 Map of CO monitor locations with respect to population density in the St. Louis
CSA, ages 65 and older.
September 2009
A-29
DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage Core Based Statistical Area
Anchorage CO Monitors
Anchorage Major Highways
Anchorage
0 15 30 60 90 120
i Kilometers
Figure A-26 Map of CO monitor locations with AQS Site IDs for Anchorage, AK.
September 2009
A-30
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-9 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Anchorage, AK.
Neighborhood
1.00 0.73
0.0 1.1
o
' 0.00 0.32
9.0
1.00
0.0
0.00
September 2009 A-31 DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
Mean
Obs
SD
A
020200018
1.04
12969
0.94
B
020200048
1.10
12703
1.04
4-
O -
"E"
a.
a
ro 2 "
concent
1 -
o-
i
I
1234 12
season
3 4
Figure A-27 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Anchorage, AK. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the
x-axis.
September 2009
A-32
DRAFT - DO NOT CITE OR QUOTE
-------
Atlanta Combined Statistical Area
Atlanta CO Monitors
Atlanta Major Highways
Atlanta
0 10 20 40 60 80
i Kilometers
Figure A-28 Map of CO monitor locations with AQS Site IDs for Atlanta, GA.
September 2009
A-33
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-10 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Atlanta, GA.
Micro Urban
1.00 0.60 0.10
0.0 0.5 0.7
0.00 0.27 0.38
~0 22~561J~
1.00 0.12
00 bT~
0.00 0.37
0 74.7
ilxT
0.0
ootT
September 2009 A-34 DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
Mean
Obs
SD
130890002
0.53
25531
0.35
131210099
0.58
25440
0.30
132230003
0.30
25712
0.13
1.7-
1.6-
1.5-
1.4-
1.3-
"E i'i -
Q, 1.1
Q.
^-1.0-
c
•B 0.9-
(0
i 0.8-
c
Q^ ri T
S u./ -
c
8 0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
1234
1234
.
!
j I
'
1234
season
Figure A-29 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Atlanta, GA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-35
DRAFT - DO NOT CITE OR QUOTE
-------
Boston Combined Statistical Area
Boston CO Monitors
Boston Major Highways
Boston
0 12.5 25 50 75
100
• Kilometers
Figure A-30 Map of CO monitor locations with AQS Site IDs for Boston, MA.
September 2009
A-36
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-11 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Boston, MA.
Micro
A B C
A 1.00 0.50 0.38
0.0 0.6 0.6
0 0.00 0.44 0.46
1 0 18.3 57.5
B 1.00 0.50
0.0 0.4
0.00 0.48
0 39.7
C 1.00
0.0
0.00
0
D
E
.Q
.C
0
z
F
C
.Q
5
G
Neighborhood
D
0.49
0.5
0.30
26.1
0.41
0.4
0.41
41.3
0.26
0.5
0.45
80.7
1.00
0.0
0.00
0
E
0.43
0.6
0.39
102.6
0.40
0.4
0.40
89.1
0.36
0.4
0.47
58.7
0.29
0.4
0.37
128.6
1.00
0.0
0.00
0
Urban
F
0.46
0.5
0.25
61.5
0.49
0.5
0.42
57.9
0.37
0.5
0.45
58.9
0.40
0.4
0.28
85.8
0.34
0.5
0.39
58.9
1.00
0.0
0.00
0
Null
G
0.35
0.7
0.60
55.1
0.35
0.4
0.58
37.2
0.52
0.4
0.56
2.5
0.27
0.5
0.58
78.2
0.34
0.4
0.55
60.2
0.34
0.6
0.59
58.0
1.00
0.0
0.00
1 °
September 2009
A-37
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID 25017007 2502500° f025004 25027002 f11002 330111009 f07101
Mean
Obs
SD
0.33
24362
0.22
0.26
24134
0.24
0.36
24260
0.26
0.53
24446
0.23
0.45
25197
0.27
0.60
25869
0.37
0.34
23707
0.22
Q.
a.
c
o
ts
I
o
o
1.9-
1.8-
1.7-
1.6-
1.5-
1.4-
1.3-
1.2-
1.1 -
1.0-
0.9-
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
I
|,
1
1
I
1234 1234 1234 1234 1234 1234 1234
season
Figure A-31 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Boston, MA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-38
DRAFT - DO NOT CITE OR QUOTE
-------
Houston Combined Statistical Area
• Houston CO Monitors
— Houston Major Highways
I Houston
0 10 20 40 60 80
Kilometers
Figure A-32 Map of CO monitor locations with AQS Site IDs for Houston, TX.
September 2009
A-39
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-12 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Houston, TX.
Micro Neighborhood
1.00 0.45 0.56 0.53 0.43
0.0 0.4 0.4 0.5 0.4
0.00 0.47 0.47 0.74 0.47
~00167 1039~323T"
1.00 0.72 0.56 0.68
00 03 05 oT~
0.00 0.29 0.73 0.24
0.0 17.5 19.8 32.2
100 065 oeT
0.0 0.5 0.4
000 073 029~
0.0 25.2 39.7
1.00 0.57
00 oT~
0.00 0.72
00 145~
1.00
0.0
000~
0.0
September 2009 A-40 DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
Mean
Obs
SD
482010024
0.42
23997
0.27
482010047
0.39
25241
0.33
482010075
0.35
24922
0.26
482011035
0.14
25285
0.22
482011039
0.33
24480
0.16
1.6-
1.5-
1.4-
1.3-
1.2-
S 1-0-
.1 °'9"
2 0.8-
I 0.7-
|0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
1 234
season
Figure A-33 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Houston, TX. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-41
DRAFT - DO NOT CITE OR QUOTE
-------
New York Combined Statistical Area
New York CC Monitors
New York Major Highways
Mew York
0 15 30 60 90
120
M Kilometers
Figure A-34 Map of CO monitor locations with AQS Site IDs for New York City, NY.
September 2009
A-42
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-13 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in New
York City, NY.
Micro Middle Neighborhood
A BCD
A 1.00 0.65 0.52 0.64
0.0 0.7 0.7 0.8
0 0.00 0.28 0.24 0.29
1 0 15.9 8.9 16.8
B 1.00 0.56 0.58
0.0 0.4 0.4
0.00 0.23 0.22
0 10.5 7.0
C 1.00 0.54
0.0 0.4
o 0.00 0.23
- 0 15.0
-o D 1.00
£ 0.0
£ 0.00
'o 0
E
F
G
H
I
E
0.54
0.9
0.35
29.9
0.55
0.4
0.25
45.8
0.41
0.4
0.28
37.5
0.55
0.4
0.23
45.4
1.00
0.0
0.00
0
F
0.32
0.9
0.34
55.0
0.40
0.4
0.25
70.6
0.33
0.4
0.25
61.0
0.35
0.5
0.26
71.5
0.50
0.4
0.24
27.5
1.00
0.0
0.00
0
Null
G
0.48
0.9
0.34
35.7
0.56
0.4
0.24
43.7
0.41
0.4
0.26
43.6
0.54
0.4
0.23
38.1
0.57
0.4
0.23
36.7
0.47
0.4
0.23
61.9
1.00
0.0
0.00
0
H
0.43
0.9
0.35
20.5
0.41
0.5
0.28
17.8
0.46
0.4
0.26
12.3
0.59
0.4
0.23
24.5
0.46
0.4
0.27
45.1
0.33
0.4
0.27
65.0
0.34
0.4
0.27
55.8
1.00
0.0
0.00
0
I
0.31
1.3
0.81
85.5
0.30
0.8
0.75
76.5
0.29
0.7
0.77
76.8
0.49
0.7
0.74
82.9
0.33
0.7
0.72
107.8
0.32
0.6
0.73
120.3
0.31
0.7
0.72
119.7
0.43
0.6
0.73
65.1
1.00
0.0
0.00
1 °
September 2009
A-43
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
34003000 34003500 34017100 34023200 34025200 34027000 36061005 36081012 36103000
1
1
Mean
0.55
0.52
0.85
0.48
0.50
0.49
0.62
0.47
0.12
Obs
23113
25150
25646
25028
25727
25691
25547
25022
25749
SD
0.27
0.30
0.43
0.27
0.24
0.25
0.21
0.23
0.17
Q.
Q.
.-
co
"c
CD
o
c
2.0-
1.9-
1.8-
1.7-
1.6-
1,5-
1.4-
1.3-
1.2-
1.1 -
1.0-
0.9-
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
i
i
nri
1234 1234 1234 1234 1234 1234 1234 1234 1234
season
Figure A-35 Box plots illustrating the seasonal distribution of hourly CO concentrations in
New York City, NY. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-
axis.
September 2009
A-44
DRAFT - DO NOT CITE OR QUOTE
-------
Phoenix Core Based Statistical Area
f
Phoenix CO Monitors
Phoenix Pdajor Highways
Phoenix
0 15 30 60 90
120
• Kilometers
Figure A-36 Map of CO monitor locations with AQS Site IDs for Phoenix, AZ.
September 2009
A-45
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-14 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Phoenix, AZ.
Micro Middle
A B
A 1.00 0.86
0.0 0.8
0 0.00 0.39
1 0.0 3.9
B 1.00
0.0
o 0.00
1 o.o
c
•o D
o
1
.5>
o
•z.
E
c
0.89
0.7
0.37
1.6
0.88
0.6
0.34
3.4
1.00
0.0
0.00
0.0
Neighborhood
D
0.80
1.1
0.43
8.9
0.81
0.7
0.41
6.6
0.81
0.9
0.38
9.4
1.00
0.0
0.00
0.0
Null
E
0.84
0.9
0.37
3.5
0.83
0.6
0.33
5.2
0.89
0.7
0.24
4.9
0.85
0.6
0.36
6.8
1.00
0.0
0.00
1 o.o
September 2009 A-46 DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
Mean
Obs
SD
040130016
0.93
25382
0.95
040130019
0.84
25589
0.88
040133002
0.58
25657
0.64
040133010
0.76
25414
0.72
040139997
0.79
25435
0.64
ncen
5-
4-
3-
2-
1 -
o-
I
I
I
,l
H
1
!
T*
I I
1234 1234 1234 1234 1234
season
Figure A-37 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Phoenix, AZ. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-47
DRAFT - DO NOT CITE OR QUOTE
-------
Pittsburgh Combined Statistical Area
• Pittsburgh CO Monitors
Pittsburgh Major Highways
_| Pittsburgh
10 20 40 60
80
• Kiiometers
Figure A-38 Map of CO monitor locations with AQS Site IDs for Pittsburgh, PA.
September 2009
A-48
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-15 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in
Pittsburgh, PA.
Middle
A
A 1.00
0.0
o 0.00
— 0
^
B
C
•o D
o
.£
.£
'o
z
E
F
G
Neighborhood
B C
0.25 0.39
0.7 0.6
0.65 0.51
33.3 68.2
1.00 0.26
0.0 0.5
0.00 0.68
0 101.0
1.00
0.0
0.00
0
D
0.73
0.4
0.39
0.7
0.29
0.5
0.62
33.6
0.42
0.4
0.51
68.0
1.00
0.0
0.00
0
E
0.20
0.7
0.56
43.4
0.09
0.6
0.69
75.0
0.16
0.6
0.57
27.5
0.30
0.5
0.54
43.4
1.00
0.0
0.00
0
Urban
F
0.30
0.8
0.88
44.1
0.09
0.5
0.90
37.8
0.21
0.6
0.87
104.1
0.35
0.5
0.86
43.7
0.02
0.7
0.87
84.1
1.00
0.0
0.00
0
G
0.43
0.6
0.68
1.8
0.42
0.5
0.73
34.4
0.11
0.6
0.72
66.8
0.52
0.5
0.69
2.2
0.05
0.7
0.74
41.9
0.18
0.7
0.88
45.8
1.00
0.0
c 0.00
§ 0
September 2009
A-49
DRAFT - DO NOT CITE OR QUOTE
-------
SitelD
Mean
Obs
SD
420030010
0.28
25655
0.32
420030031
0.32
25936
0.26
420030038
0.47
25818
0.33
420070014
0.28
25500
0.27
420730015
0.32
25745
0.26
421250005
0.21
25319
0.23
421290008
0.07
25785
0.15
1,4-
1,3-
1.2-
1.1 -
1.0-
"§0.9-
Q.
-So.8-
c
•Bo.7-
ra
"E 0.6-
0)
o
§0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
I
n i
hi
_ —
t i
1234 1234 1234 1234 1234 1234 1234
season
Figure A-39 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Pittsburgh, PA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-
axis.
September 2009
A-50
DRAFT - DO NOT CITE OR QUOTE
-------
Seattle Combined Statistical Area
• Seattle CO Monitors
- Seattle Major Highways
~| Seattle
0 15 30 60 90
120
• Kilometers
Figure A-40 Map of CO monitor locations with AQS Site IDs for Seattle, WA.
September 2009
A-51
DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
530330019
Mean
0.75
Obs
25818
SD
0.49
3-
E 2
Q.
•S
c
.2
"ro
k—
I
8 1
o-
i i r
234
season
Figure A-41 Box plots illustrating the seasonal distribution of hourly CO concentrations in
Seattle, WA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-52
DRAFT - DO NOT CITE OR QUOTE
-------
St Louis Combined Statistical Area
St Louis CO Monitors
St Louis Ma;or Kghways
St Louis
0 10 20 40 60 80
Kilometers
Figure A-42 Map of CO monitor locations with AQS Site IDs for St. Louis, MO.
September 2009
A-53
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-16 Table of inter-sampler comparison statistics, including Pearson r, P90 (ppm), COD, and d
(km), as defined in the text, for each pair of hourly CO monitors reporting to AQS in St.
Louis, MO.
Neighborhood Null
A 1.00 0.60 0.19
0.0 0.3 0.5
•§ 0.00 0.24 0.40
I 0 9~52TT
o
I, B 1.00 0.19
1 blb~5~~
0.00 0.42
0 19.8
c ilxT
= 0.0
^ b~otT
September 2009 A-54 DRAFT - DO NOT CITE OR QUOTE
-------
Site ID
171630010 291890004 295100086
Mean
0.44
0.43
0.42
Obs
25325
25879
25938
SD
0.25
0.25
0.29
1.3-
1.2-
1.1 -
1.0-
^ 0.9-
1. 0.8-
CL
c 0.7-
.9
ro 0.6 -
i 0.5-
o
| 0.4-
0.3-
0.2-
0.1 -
o.o-
-0.1 -
I
1
i i
i f
i
1234 1234 1234
season
Figure A-43 Box plots illustrating the seasonal distribution of hourly CO concentrations in St.
Louis, MO. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
September 2009
A-55
DRAFT - DO NOT CITE OR QUOTE
-------
TableA-17
Comparison of
max, 24-h avg,
distributional data at different monitoring
and 8-h daily max data for Atlanta, GA.
scales
for hourly,
1-h daily
PERCENTILES
Time scale
n
mean
min
1 5
10
25
50
75
90 95
99
max
ALL HOURLY
Microscale
Urban Scale
25440
51243
0.6
0.4
0.0
0.0
0.2 0.2
0.0 0.1
0.3
0.2
0.4
0.3
0.4
0.3
0.5
0.3
0.7 0.7
0.4 0.5
1.0
0.7
1.2
1.0
1-H DAILY MAX
Microscale
Urban Scale
1075
2154
1.0
0.7
0.2
0.0
0.3 0.4
0.2 0.2
0.6
0.3
0.7
0.3
0.8
0.4
1.0
0.5
1.2 1.2
0.8 0.9
1.6
1.3
1.9
1.5
1-H DAILY AVG
Microscale
Urban Scale
1075
2154
0.6
0.4
0.2
0.0
0.2 0.3
0.1 0.2
0.3
0.2
0.4
0.3
0.5
0.3
0.5
0.4
0.6 0.7
0.5 0.5
0.8
0.7
1.0
0.9
8-H DAILY MAX
Microscale
Urban Scale
1075
2154
0.8
0.5
0.3
0.0
0.3 0.3
0.1 0.2
0.4
0.3
0.5
0.3
0.6
0.3
0.7
0.4
0.9 0.9
0.6 0.7
1.2
1.0
1.3
1.3
TableA-18
Comparison of
max, 24-h avg,
distributional data at different monitoring
and 8-h daily max data for Boston, MA.
scales
for hourly,
1-h daily
PERCENTILES
Time scale
n
mean
min
1 5
10
25
50
75
90 95
99
max
ALL HOURLY
Microscale
Neighborhood Scale
Urban Scale
25869
97526
24446
0.6
0.4
0.5
0.0
0.0
0.0
0.1 0.2
0.0 0.0
0.1 0.3
0.3
0.1
0.3
0.4
0.2
0.4
0.4
0.3
0.4
0.5
0.3
0.5
0.7 0.7
0.4 0.5
0.6 0.6
1.0
0.6
0.8
1.2
0.8
0.9
1-H DAILY MAX
Microscale
Neighborhood Scale
Urban Scale
1080
4212
1086
1.2
0.6
0.8
0.2
0.0
0.0
0.4 0.5
0.1 0.2
0.3 0.4
0.6
0.3
0.5
0.7
0.4
0.6
0.8
0.4
0.6
0.9
0.5
0.8
1 .2 1 .4
0.7 0.7
0.9 1.0
2.0
1.1
1.2
2.5
1.4
1.4
1-H DAILY AVG
Microscale
Neighborhood Scale
Urban Scale
1080
4212
1086
0.6
0.4
0.5
0.1
0.0
0.0
0.2 0.3
0.0 0.1
0.1 0.3
0.3
0.1
0.4
0.4
0.2
0.4
0.5
0.3
0.5
0.6
0.4
0.5
0.7 0.7
0.4 0.5
0.6 0.6
0.9
0.6
0.7
1.1
0.7
0.8
8-H DAILY MAX
Microscale
Neighborhood Scale
Urban Scale
1080
4212
1086
0.8
0.5
0.7
0.3
0.3
0.3
0.3 0.3
0.3 0.3
0.3 0.3
0.4
0.3
0.3
0.6
0.3
0.5
0.6
0.3
0.5
0.7
0.4
0.6
0.9 1.0
0.6 0.6
0.8 0.8
1.4
0.8
1.0
1.7
1.0
1.1
September 2009
A-56
DRAFT - DO NOT CITE OR QUOTE
-------
TableA-19
Comparison of
max, 24-h avg,
distributional data at different monitoring scales for hourly,
and 8-h daily max data for Denver, CO.
1-h daily
PERCENTILES
Time scale
n
mean
min
1
5
10
25
50 75 90 95
99 max
ALL HOURLY
Microscale
Neighborhood Scale
77070
51968
0.5
0.5
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.2
0.3
0.3
0.3 0.4 0.6 0.7
0.3 0.4 0.6 0.6
1.0 1.3
1.0 1.3
1-H DAILY MAX
Microscale
Neighborhood Scale
3190
2173
1.2
1.1
0.1
0.1
0.3
0.2
0.4
0.3
0.5
0.4
0.7
0.6
0.8 1.0 1.4 1.5
0.6 0.9 1.3 1.5
2.2 2.7
2.1 2.6
1-H DAILY AVG
Microscale
Neighborhood Scale
3190
2173
0.5
0.5
0.0
0.0
0.1
0.1
0.2
0.2
0.2
0.3
0.3
0.3
0.4 0.5 0.6 0.6
0.4 0.5 0.6 0.6
0.9 1.0
0.9 1.1
8-H DAILY MAX
Microscale
Neighborhood Scale
3190
2173
0.8
0.8
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.5
0.4
0.5 0.7 0.9 1.0
0.5 0.7 0.9 1.0
1.4 1.8
1.5 1.8
Table A-20
Comparison of
max, 24-h avg,
distributional data at different monitoring scales for hourly,
and 8-h daily max data for Houston, TX.
1-h daily
PERCENTILES
Time scale
n
mean
min
1
5
10
25
50 75 90 95
99 max
ALL HOURLY
Microscale
Neighborhood Scale
24922
99003
0.3
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
0.2
0.2 0.3 0.4 0.5
0.2 0.3 0.4 0.4
0.6 0.8
0.6 0.8
1-H DAILY MAX
Microscale
Neighborhood Scale
1043
4145
0.7
0.7
0.0
0.0
0.0
0.0
0.2
0.1
0.3
0.2
0.4
0.4
0.5 0.6 0.8 0.9
0.4 0.5 0.8 0.8
1.2 1.4
1.3 1.7
1-H DAILY AVG
Microscale
Neighborhood Scale
1043
4145
0.3
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.2
0.2
0.3 0.4 0.5 0.5
0.2 0.3 0.4 0.4
0.6 0.6
0.5 0.6
8-H DAILY MAX
Microscale
Neighborhood Scale
1043
4145
0.5
0.5
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3 0.4 0.6 0.6
0.3 0.4 0.5 0.6
0.8 1.0
0.9 1.1
September 2009
A-57
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-21 Comparison of distributional data at different monitoring scales for hourly, 1-h daily
max, 24-h avg, and 8-h daily max data for Los Angeles, CA.
PERCENTILES
Time scale
n
mean
min
1
5
10
25
50
75
90
95
99
max
ALL HOURLY
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
24885
98564
49757
24264
0.7
0.5
0.3
0.4
0.0
0.0
0.0
0.0
0.2
0.0
0.0
0.0
0.3
0.0
0.0
0.1
0.3
0.0
0.0
0.1
0.4
0.1
0.1
0.1
0.4
0.2
0.1
0.2
0.5
0.4
0.2
0.3
0.7
0.6
0.3
0.4
0.8
0.7
0.3
0.5
1.2
1.1
0.6
1.0
1.6
1.6
0.8
1.4
1-H DAILY MAX
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
1080
4299
2164
1053
1.3
1.2
0.7
1.0
0.2
0.0
0.0
0.0
0.4
0.1
0.0
0.1
0.5
0.1
0.0
0.2
0.6
0.2
0.1
0.3
0.8
0.5
0.3
0.4
0.8
0.6
0.3
0.4
1.1
0.9
0.5
0.7
1.6
1.3
0.8
1.3
1.7
1.5
0.9
1.5
2.3
2.5
1.3
2.2
2.7
3.7
1.7
2.6
1-H DAILY AVG
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
1080
4299
2164
1053
0.7
0.5
0.3
0.4
0.2
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.3
0.0
0.0
0.1
0.4
0.1
0.0
0.1
0.5
0.2
0.1
0.2
0.5
0.2
0.2
0.2
0.6
0.4
0.2
0.3
0.7
0.6
0.3
0.5
0.8
0.7
0.4
0.6
1.1
1.1
0.5
0.9
1.2
1.5
0.6
1.1
8-H DAILY MAX
Microscale
Middle Scale
Neighborhood Scale
Urban Scale
1080
4299
2164
1053
0.9
0.8
0.5
0.7
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.3
0.3
0.4
0.3
0.3
0.3
0.6
0.3
0.3
0.3
0.6
0.3
0.3
0.3
0.8
0.6
0.3
0.5
1.1
0.9
0.5
0.8
1.2
1.0
0.6
0.9
1.6
1.8
0.9
1.5
1.8
2.4
1.2
1.8
September 2009
A-58
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-22 Comparison of distributional data at different monitoring scales for hourly, 1-h daily
max, 24-h avg, and 8-h daily max data for New York, NY.
PERCENTILES
Time scale
mean mm
10
25
50
75
90
95
99
ALL HOURLY
Microscale
25646
0.8
0.0
0.2
0.3
0.4
0.5
0.6
0.8
1.0
1.1
1.4
Middle Scale
48660
0.6
0.0
0.1
0.3
0.3
0.4
0.5
0.6
0.7
0.7
0.9
1.0
Neighborhood Scale
25150
0.5
0.0
0.2
0.2
0.3
0.3
0.4
0.4
0.6
0.6
0.9
1.1
1-H DAILY MAX
Microscale
1077
1.4
0.3
0.4
0.6
0.8
1.0
1.1
1.4
1.7
1.8
2.1
2.4
Middle Scale
2053
0.9
0.2
0.4
0.5
0.6
0.7
0.7
0.8
1.0
1.1
1.3
1.5
Neighborhood Scale
1053
0.9
0.2
0.3
0.4
0.4
0.6
0.6
0.8
1.0
1.1
1.5
1.9
1-H DAILY AVG
Microscale
1077
0.8
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1.0
1.0
1.3
1.4
Middle Scale
2053
0.6
0.0
0.2
0.3
0.4
0.5
0.5
0.6
0.7
0.7
0.8
0.9
Neighborhood Scale
1053
0.5
0.1
0.2
0.3
0.3
0.4
0.4
0.5
0.6
0.6
0.8
1.0
8-H DAILY MAX
Microscale
1077
1.2
0.3
0.4
0.6
0.7
0.9
0.9
1.1
1.4
1.4
1.7
1.9
Middle Scale
2053
0.7
0.3
0.3
0.4
0.4
0.6
0.6
0.7
0.8
0.9
1.0
1.2
Neighborhood Scale
1053
0.7
0.3
0.3
0.3
0.3
0.4
0.5
0.6
0.8
0.8
1.2
1.5
September 2009
A-59
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-23 Comparison of distributional data at different monitoring scales for hourly, 1-h daily
max, 24-h avg, and 8-h daily max data for Phoenix, AZ.
PERCENTILES
Time scale
mean mm
10
25
50
75
90
95
99
ALL HOURLY
Microscale
25382
0.9
0.0
0.0
0.1
0.1
0.3
0.3
0.6
1.1
1.3
2.3
3.0
Middle Scale
25414
0.8
0.0
0.0
0.1
0.1
0.3
0.3
0.5
0.9
1.0
2.3
Neighborhood Scale
51246
0.7
0.0
0.0
0.1
0.1
0.2
0.3
0.4
0.7
0.8
2.4
1-H DAILY MAX
Microscale
1063
2.2
0.0
0.2
0.5
0.7
1.1
1.2
1.9
2.8
3.1
4.2
4.7
Middle Scale
1066
1.8
0.1
0.3
0.5
0.7
1.0
1.1
1.6
2.2
2.4
3.2
3.8
Neighborhood Scale
2156
0.1
0.2
0.4
0.5
0.8
0.9
1.5
2.3
2.6
3.6
4.2
1-H DAILY AVG
Microscale
1063
0.9
0.0
0.0
0.2
0.2
0.4
0.4
0.7
1.2
1.3
2.0
2.3
Middle Scale
1066
0.8
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.9
1.0
1.5
1.7
Neighborhood Scale
2156
0.7
0.0
0.1
0.2
0.2
0.3
0.4
0.5
0.9
0.9
1.5
8-H DAILY MAX
Microscale
1063
1.5
0.3
0.3
0.3
0.4
0.6
0.7
1.2
2.0
2.2
3.1
3.5
Middle Scale
1066
1.2
0.3
0.3
0.3
0.4
0.7
0.7
1.0
1.5
1.7
2.3
2.7
Neighborhood Scale
2156
1.2
0.3
0.3
0.3
0.3
0.5
0.6
0.9
1.5
1.7
2.5
3.0
September 2009
A-60
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-24
Comparison of
max, 24-h avg,
distributional data at different monitoring
and 8-h daily max data for Pittsburgh, PA.
scales
for hourly,
1-h daily
PERCENTILES
Time scale
n
mean
min 1
5
10
25
50
75
90
95
99 max
ALL HOURLY
Middle Scale
Neighborhood Scale
Urban Scale
25818
77000
76940
0.5
0.3
0.2
0.0 0.0
0.0 0.0
0.0 0.0
0.1
0.0
0.0
0.1
0.0
0.0
0.3
0.1
0.0
0.3
0.1
0.0
0.4
0.2
0.1
0.5
0.3
0.3
0.6
0.4
0.3
0.8 1.1
0.6 0.8
0.6 0.8
1-H DAILY MAX
Middle Scale
Neighborhood Scale
Urban Scale
1079
3210
3208
0.9
0.6
0.4
0.0 0.2
0.0 0.0
0.0 0.0
0.4
0.1
0.0
0.4
0.2
0.0
0.6
0.3
0.1
0.6
0.3
0.2
0.8
0.5
0.4
1.1
0.7
0.6
1.1
0.7
0.7
1.6 1.9
1.1 1.3
1.0 1.2
1-H DAILY AVG
Middle Scale
Neighborhood Scale
Urban Scale
1079
3210
3208
0.5
0.3
0.2
0.0 0.1
0.0 0.0
0.0 0.0
0.2
0.0
0.0
0.2
0.0
0.0
0.3
0.1
0.0
0.3
0.2
0.0
0.4
0.3
0.1
0.5
0.3
0.3
0.6
0.4
0.3
0.8 0.9
0.6 0.7
0.6 0.7
8-H DAILY MAX
Middle Scale
Neighborhood Scale
Urban Scale
1079
3210
3208
0.7
0.5
0.4
0.3 0.3
0.3 0.3
0.3 0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.3
0.3
0.4
0.3
0.3
0.6
0.3
0.3
0.7
0.5
0.4
0.8
0.5
0.5
1.1 1.3
0.8 1 .0
0.8 1.0
Table A-25
Comparison of
max, 24-h avg,
distributional data at different monitoring
and 8-h daily max data for Seattle, WA.
scales
for hourly,
1-h daily
PERCENTILES
Time scale
n
mean
min 1
5
10
25
50
75
90
95
99 max
ALL HOURLY
Microscale
25818
0.8
0.0 0.1
0.2
0.3
0.4
0.5
0.6
0.9
0.9
1.3 1.6
1-H DAILY MAX
Microscale
1079
1.5
0.2 0.4
0.5
0.7
0.9
1.0
1.3
1.7
1.8
2.4 2.9
1-H DAILY AVG
Microscale
1079
0.8
0.1 0.2
0.3
0.4
0.5
0.6
0.7
0.9
0.9
1.2 1.4
8-H DAILY MAX
Microscale
1079
1.1
0.3 0.3
0.4
0.5
0.7
0.8
1.0
1.3
1.4
1.8 2.2
September 2009
A-61
DRAFT - DO NOT CITE OR QUOTE
-------
Table A-26 Comparison of distributional data at different monitoring scales for hourly, 1-h daily
max, 24-h avg, and 8-h daily max data for St. Louis, MO.
PERCENTILES
Time scale
mean min 1 5 10 25 50 75 90 95 99 max
ALL HOURLY
Neighborhood Scale 51263 0.4 0.0 0.1 0.2 0.2 0.3 0.3 0.4 0.5 0.5 0.6 0.8
1-H DAILY MAX
Neighborhood Scale 2138 0.8 0.1 0.2 0.3 0.4 0.5 0.5 0.6 0.9 1.0 1.5 2.0
1-H DAILY AVG
Neighborhood Scale 2138 0.4 0.0 0.1 0.2 0.3 0.3 0.3 0.4 0.5 0.5 0.6 0.7
8-H DAILY MAX
Neighborhood Scale 2138 0.6 0.3 0.3 0.3 0.3 0.3 0.3 0.5 0.6 0.7 1.0 1.3
September 2009
A-62
DRAFT - DO NOT CITE OR QUOTE
-------
Winter
Spring
1 •
2 -
3 •
4 •
5 •
6 •
7 •
8 -
O
o
o
(3D
(3D
®
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
O
o
(QE)
O
-------
Winter
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
oo
o o
o
CD
OD
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
O O
o o
o o
GOO
CD
OS)
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
1.0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
0 C
c
D O
) O
GDO
OOO
-1.0 -0,8 -0.6 -0.4 -0.2 0,0 0.2 0.4 0,6 0,8 1,0
r (correlation coefficient)
Figure A-45 Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) Os, (4) PMio, and (5) PM2.6 concentrations for Atlanta, GA. Also
shown are correlations between 24-h avg CO concentration with (6) daily max 1-h
and (7) daily max 8-h CO concentrations and (8) between daily max 1-h and daily
max 8-h CO concentrations. (Refer the numbers in this caption to those on the y-
axis of each seasonal plot.)
September 2009
A-64
DRAFT - DO NOT CITE OR QUOTE
-------
Winter
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
o
OGDO
OGD
O O O
oo
dD
m
0®
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
CD
OCM)
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
mo
(TOO
CSBDO
dSQ O
(ISJH5)
Qm
(Mi
1 •
2 •
3 •
4 '
5 •
6 •
7 •
a •
QDS2)
O OlED
CKODD
O
OC3D
(OBi
(0
(SBD
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
1.0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
ooc
Q9QD
CO
)
O O
O O CgO
fituKlBP 9M I
KMBB
(CHID
-1.0 -0,8 -0,6 -0.4 -0.2 0.0 02 0.4 0.6 0,8 1,0
r (correlation coefficient)
1
2 -
3 •
4 •
5 •
6
7 •
B
€000
-1.0 -0,8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0,6 0,8 1,0
r (correlation coefficient)
Figure A-47 Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) 03, (4) PMio, and (5) PM2.5 concentrations for New York City, NY.
Also shown are correlations between 24-h avg CO concentration with (6) daily
max 1-h and (7) daily max 8-h CO concentrations and (8) between daily max 1-h
and daily max 8-h CO concentrations. (Refer the numbers in this caption to those
on the y-axis of each seasonal plot.)
September 2009
A-66
DRAFT - DO NOT CITE OR QUOTE
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
DUD
O OO
O (SO)
O CTO
GO
GO®
4B1
dm
\ •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
dSD
(
O 00
O €SD
XO OGD)
OO O
(M(f&i)
(QBE)
GOB)
-0.8 -0.5 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 04 0.6 0.8 1.0
Summer
Fall
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
OO 0
c
©
OO
OO
OO
(GD>
OO
O €D
-1.0 -0.8 -0.6 -04 -0,2 0.0 0,2 0.4 0,6 0.8 1.0
r (correlation coefficient)
1
2 •
3 •
4 •
5 -
6
7 •
GD
000
-1.0 -0,8 -0.6 -0.4 -02 0.0 0.2 0.4 0,6 0,8 10
r (correlation coefficient)
Figure A-48 Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) 03, (4) PMio, and (5) PM2.6 concentrations for Phoenix, AZ. Also
shown are correlations between 24-h avg CO concentration with (6) daily max 1-h
and (7) daily max 8-h CO concentrations and (8) between daily max 1-h and daily
max 8-h CO concentrations. (Refer the numbers in this caption to those on the y-
axis of each seasonal plot.)
September 2009
A-67
DRAFT - DO NOT CITE OR QUOTE
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
o
o
o
(O
^^p)
CD
1 •
2 •
3 -
4 -
5 •
6 •
7 -
8 •
O
o
o
GSD
®2)
CD
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
1.0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 02 04 0.6 0.8 1.0
Summer
Fall
1 -
2 •
3 •
4 •
5 -
6 •
7 •
8 -
O
o
o
OK©
O
(QED
-1.0 -0.8 -0.6 -0.4 -0.2 0,0 0.2 0.4 0,6 0.8 1.0
r (correlation coefficient)
1 •
2 -
3 •
4 •
5 •
6 -
7 •
8 •
O
o
o
(25>
O
(TO
-1,0 -0.8 -0,6 -0.4 -0,2 0,0 02 0.4 0,6 0,8 1.0
r (correlation coefficient)
Figure A-49 Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) Os, (4) PMio, and (5) PM2.6 concentrations for Seattle, WA. Also
shown are correlations between 24-h avg CO concentration with (6) daily max 1-h
and (7) daily max 8-h CO concentrations and (8) between daily max 1-h and daily
max 8-h CO concentrations. (Refer the numbers in this caption to those on the y-
axis of each seasonal plot.)
September 2009
A-68
DRAFT - DO NOT CITE OR QUOTE
-------
Annex B. Dos i me try Studies
Table B-1 Recent studies related to CO dosimetry and pharmacokinetics.
Reference
Purpose
Findings
Ahorr,ot,i
Aberg et ai.
To investigate CO concentrations in blood donors in
Sweden.
The mean CO concentration is blood donors was 84.5 pmol/L
Concentrations over 130 pmol/L were found in 6% of blood and the
highest concentration was 561 pmol/L. By using a calculation, 23% of
banked blood bags could exceed 1.5% COHb with a highest fraction of
7.2% COHb.
Abram et al. (2007,1938591
To present the Quantitative Circulatory Physiology
model as a teaching module in the practice of
medicine.
QCP is a dynamic mathematical model based on published models and
parameters of biological interactions.
To use a quantum mechanics/molecular mechanics
Alcantara et al (2007 1938671 approach to understand the cooperativity of Hb ligand
Alcantara et al. (2007,193867) binding gnd differences in energy between T and R Hb
functional states.
The ligand binding energies between R and T states differ due to strain
induced in the heme and its ligands and in protein contacts in the a and
|3 chains.
Adiretal. (1999, 001026)
To determine if low concentrations of CO would affect
exercise performance and myocardial perfusion in
young healthy men.
Men with COHb levels between 4-6% had decreased exercise
performance measured by decreased mean duration of exercise
(1.52 min) and maximal effort described by metabolic equivalent units
(2.04). No changes were seen in lactate/pyruvate ratio, arrhythmias, or
myocardial perfusion.
Anderson et al. (2000, 0118361
To investigate if CO could be endogenously produced
in the nose and paranasal sinuses.
Both nose and paranasal sinuses contained HO-like immunoreactivity,
mostly in the respiratory epithelium, indicating local CO production in
the upper respiratory airways.
Aroraetal. (2001,1867131
To evaluate the effect of multiple transfusion recipient
thalassemics on pulmonary function.
DLCO was decreased in all the patients with restrictive lung disease and
fall in DLCO showed a good correlation with the severity of restrictive
disease. Thalassemics had a decrease in lung volume and a
proportional decrease in flow rate.
Benignusetal. (2006,1513441
To adapt and use a human model for toluene uptake
and elimination including a brain compartment.
The Quantitative Circulatory Physiology 2004 (QCP 2004) model was
used to construct simulations of scenarios of toxicant exposure and
human activities. QCP accurately predicted toluene blood
concentrations from inhaled exposure.
Bos et al. (2006, 1940841
To use a PBPK model to set AEGL for methylene
chloride.
This model adequately predicted COHb levels formed by various
methylene chloride concentrations, specifically in nonconjugators
lacking the GSTT-1 enzyme, and proposed AEGL values.
Bruce and Bruce (2003,1939751
To create a mathematical model to predict uptake and
distribution of CO in both vascular and tissue
compartments during constant or variable inhalation
levels of CO.
This model contains 5 compartments: lung, arterial blood, venous blood,
muscle tissue, and nonmuscle tissue. It was constructed to include
tissue compartment flux and difference between venous and arterial
COHb for short exposures which is not possible with the CFK model.
Bruce and Bruce (2006,1939801
To use their mathematical multicompartment model
along with experimental data to predict the factors that
influence the washout rates of CO, along with
predicting the rates of CO uptake, distribution in
vascular and extravascular (muscle and non-muscle
tissue) compartments, and washout over a range of
exposure and conditions.
Rates of CO washout follow a biphasic elimination where washout was
faster immediately post exposure. The difference in rates is likely due to
slow equilibration between vascular and extravascular compartments.
Important factors contributing to washout kinetics include: peak COHb
level, exposure duration and concentration, time after exposure samples
were obtained, and individual variability.
Bruce and Bruce (2008,1939771
To develop a mathematical model able to integrate a
large body of indirect experimental findings on the
uptake and distribution of CO by accounting for
arteriole to venule shunting via intra-tissue pathways
and diffusion of blood gases into tissues from pre-
capillary vessels like arterioles.
The former model of Bruce and Bruce (2006,1939801 was altered by
adding a mass balance equation for 02 so P02 is directly calculated in
the compartments and the muscle compartment is divided into two sub-
compartments of muscle and non-muscle tissue. CO uptake from blood
by muscle is much slower than 02, thus COHb% will fall rapidly while
COMb% could remain high.
Carraway et al. (2000, 0210961
To test the hypothesis that HO-1 gene expression and
protein are upregulated in the lungs of rats during
chronic hypoxia.
Rats were exposed to HH (17,000 ft) for 1-21 days. COHb increased
after 1 day and progressively after 14 days. HO-1 protein and activity
were upregulated during early chronic hypoxia. This HO-1 was localized
to inflammatory cells and then to newly muscularized arterioles.
September 2009
B-1
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Purpose
Findings
To describe a new method for measurement of CO
Castillo et al. (2006,1932341 DLCO and VA in sleeping infants (6-22 mo old), using a
single 4-s breath-hold technique.
VA30 and DLCO increased with increasing body length and the method
could be used as a measurement of lung development and growth.
Chakraborty et al. (2004,1937591
To present an analytical expression for diffusing
capacity of CO, NO, C02, and 02 to the red blood cell
in terms of optimum size and shape of the RBC,
thickness of the unstirred plasma layer surrounding the
RBC, diffusivities and solubilities of the gas in RBC and
boundary layer, hematocrit, and the slope of the
dissociation curve.
Results indicate the discoidal shape of the RBC is optimal for 02 uptake
and reaction velocity is limited by mass transfer resistance in
surrounding stagnant plasma layer. The paper overviews rate constants
and reaction kinetics for CO binding to Hb. CO diffusing capacity is
shown to be reaction rate limited at low POO under normoxic and
hyperoxic conditions, but diffusion rate limited under hypoxic and high
PCO conditions.
Cronenberger et al. (2008,
1940851
To develop a population-based model to describe and
predict the pharmacokinetics of COHb in adult
smokers.
This two compartment model included zero-order input and first-order
elimination and required a compartment for extravascular binding of CO
to accurately predict COHb formation during multiple short and rapid
inhalations followed by a period of no exposure, as occurs in smoking.
Smokers COHb ranged from 0.8 to 11.1%.
Cronje et al. (2004,1804401
To analyze CO uptake and elimination in the brain,
muscle, heart, and blood of rats, with the intent of
testing the Warburg hypothesis that CO partitioning is
directly proportional to the C0/02 ratio.
Results indicate that tissue and blood [CO] dissociate during CO
inhalation, but [CO] does not follow blood [CO] or 1/P02 as in the
Warburg theory during intake or elimination. Tissue [CO] increases later
during the resolution period and varies significantly among animals and
tissues. The deviation from the predicted values in the brain is likely due
to the release of heme and increase in NADPH stimulating endogenous
CO production by HO.
De las Heras et al. (2003,
1940871
To assess production of CO (venous COHb measured
by CO-oximeter and exhaled CO) in patients with
cirrhosis with and without spontaneous bacterial
peritonitis.
Patients with SBP had higher CO production than noninfected cirrhotic
patients and both groups of patients had higher CO production
compared to healthy controls. CO production decreased slowly after
resolution of the disease.
Duttonetal. (2001,021307)
To monitor CO, N02, and PAH emissions during the
operation of unvented natural gas fireplaces in two
residences in Boulder, CO, at various times between
1997 and 2000.
Results showed significant accumulation of CO, N02, and PAH indoors
when the fireplaces were used. CO concentrations could exceed
100 ppm. N02 concentrations avg 0.36 ppm over 4 h. PAH 4-h time avg
reached 35 ng/m .
Ehlers et al. (2009,1940891
To determine the level of COHb found in banked blood
in the Albany, NY region.
The avg COHb level was 0.78%. The highest recorded COHb level was
12% and 10.3% of packed red blood cell units had levels of 1.5% COHb
or higher.
Gosselin et al. (2009,1909461
To develop a variant of the CFK model that links COHb
levels in humans to ambient CO levels under various
environmental or occupational exposure conditions.
The model adds alveoli-blood and blood-tissue CO exchanges and
mass conservation of CO at all times to the CFK equation. The model
better predicted COHb formation over a wide range of CO levels and
scenarios with linear regression analysis of predicted versus observed
values generating a slope of 0.996 (95% Cl: 0.986-1.001) compared to
0.917 (95% Cl: 0.906-0.927) using the CFK model
Hampson and Wfeaver (2007,
1902721
To present a case study of a man with drug-induced
hemolytic anemia and hepatic failure.
The man had elevated endogenous CO production resulting in levels of
COHb as high as 9.7%.
Hart et al. (2006,1940921
To investigate the relationship between COHb and
smoking habit and mortality.
COHb was related to self reported smoking in a dose dependent
manner. COHb was positively associated with all causes of mortality
analyzed including CHD, COPD, stroke, and lung cancer. Mean COHb
levels ranged from 1.59% in never smokers to 6.02% in the most often
smoking group.
To review the current concepts and practical relevance
of the diffusing capacity/cardiac output interaction, in
hopes of aiding in the interpretation of diffusing
capacity, membrane diffusing capacity, and capillary
blood volume.
This review helped to understand the determinants of changes in
diffusing capacity, including hematocrit, erythrocyte distribution, blood
volume, lung volume, cardiac output, etc.
Johnson et al. (2006,1938741
To test that heme-derived CO formation is increased
and contributes to hypertension and arteriolar
endothelial dysfunction in obese Zucker rats.
Obese Zucker rats showed increased respiratory CO excretion that was
lowered by HO inhibition. Skeletal muscle arterioles of obese rats have
attenuated ACh and flow responses that was abolished by HO inhibition
(HO inhibition enhanced dilation).
Lambertoetal. (2004,1938451
To evaluate which component, alveolar membrane
diffusing capacity (Dm) and pulmonary capillary blood
volume (Vc), is responsible for decreased resting
DLCO in sarcoidosis patients and which component is
the best predictor of gas exchange abnormalities.
Patients with pulmonary sarcoidosis had decreased lung volumes, a
loss in DLCO, and gas exchange abnormalities during exercise
including decreased Pa02 and increased alveolar-arterial oxygen
pressure difference. Dm accounted for the majority of the decrease in
DLCO and was predictive for gas exchange abnormalities.
Levesque et al. (2000, 0118861
To describe the results of air quality monitoring in an
indoor ice skating rink during Monster Truck and car
demolition exhibitions.
Maximum time-weighted avg levels of CO were 100 ppm with several
peaks exceeding 200 ppm (max: 1,600 ppm).
Lim et al. (2000,1269691
To investigate the expression of HO-1 and HO-2 in
bronchial biopsies obtained from patients with mild
asthma compared with that of subjects without asthma.
HO-1 and HO-2 expression is widely distributed equally in healthy
subjects and subjects with asthma and is not modulated by inhaled
corticosteroid therapy.
September 2009
B-2
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Purpose
Findings
Mahonevetal. (1993, 0138591
To compare CO-oximeter measurements of COHb
against a gas chromatography reference method.
In general, the 5 CO-oximeters that were tested underestimated COHb
concentrations for COHb >2.5% and overestimated COHb concentration
for COHb < 2.5%, when compared to reference gas chromatography
method.
Marks et al. (2002, 0306161
To review the analytical methods for measurement of
endogenous formation of CO in a variety of tissues.
A variety of methods have been used to measure endogenous CO. The
rate of formation varies over a narrow range from 0.029 nmol/mg
protein/h to 0.28 nmol/mg protein/h depending on tissue. Brain and liver
regions tend to have the highest rates of CO formation likely due to high
levels of HO activity in these tissues.
To evaluate DLCO impairment and microalbuminuria in
patients with active ulcerative colitis (UC) and to assess Reduced DLCO was present in 67% of patients. Microalbuminuria was
whether these tests correlate with intestinal present in 63% of patients with ulcerative colitis.
inflammation.
Marvisi et al. (2007,1867021
Merxetal. (2001,0020061
To investigate the effect of CO inactivation of Mb in
wild-type and myo-/- mice on hemodynamics and
oxygen dynamics.
Fully oxygenated Mb treated with 20% CO had no change in left
ventricular developed pressure or coronary venous P02. Partially
02-saturated Mb (87% 02Mb) exposed to 20% CO had significantly
decreased LVDP (12%) and Pv02 (30%) in wild-type but not myo-/-
hearts.
Monmaetal. (1999,1804261.
To study whether exhaled CO levels were increased in
seasonal allergic rhinitis.
Exhaled CO concentrations were higher in allergic rhinitis patients
during cedar pollen season (3.6 ppm; SD 0.3 ppm) that out (1.2 ppm;
SD0.1 ppm).
Morimatsu et al. (2006,1940971
To examine exhaled CO, arterial COHb, and bilirubinn
IXa levels in critically ill patients.
Exhaled CO concentrations were significantly higher in critically ill
patients compared to controls. There was a significant correlation
between exhaled CO and COHb or bilirubin. There was no correlation
between exhaled CO and disease severity or degree of inflammation.
There was higher exhaled CO in survivors compared to nonsurvivors.
Muchova et al. (2007,1940981
To determine if long-term use of statins affects HO
activity and blood and organ CO and bilirubin in FvB
mice (6-8 wks).
Rosuvastatin and atorvastatin treatment increased COHb, plasma
bilirubin, and heart tissue CO content. Both statins caused an increase
in HO activity in heart tissue, whereas no changes were seen in brain or
lung. Liver HO activity was inconsistent over time and between statins.
Both statins decreased the heart antioxidant capacity; and changes in
HO activity and antioxidant capacity can be reversed by HO inhibitor
treatment.
Neto et al. (2008, 1946721
To develop a model of the respiratory system to
analyze CO transport in the human body submitted to
several physical activity levels.
The model contains six compartments including: alveolar, pulmonary
capillaries, arterial, venous, tissue capillary, and tissues (muscular and
non-muscular). The highest and lowest COHb levels were simulated in
the walking individual, suggesting that greater variability in COHb
occurs in higher physical activity levels.
Pelham et al. (2002, 0257161
To review the literature on exposure and effects of
mainly CO and N02 in enclosed ice rinks.
CO levels as high as 300 ppm were recorded after episodes of
malfunctioning ice resurfacing equipment or inadequate ventilation.
Pared! etal. (1999,194102)
To investigate the level of exhaled CO produced by
diabetic patients.
Diabetic patients (type 1 and 2) had higher levels of exhaled CO than
healthy subjects. Exhaled CO levels correlated with the incidence of
glycemia and the duration of diabetes.
Pared! etal. (1999,118798)
To investigate whether cystic fibrosis patients have
higher exhaled levels of CO and if this is reduced by
corticosteroid therapy.
Cystic fibrosis patients had higher exhaled CO concentrations compared
to healthy controls. Patients receiving corticosteroid therapy had lower
exhaled CO concentrations.
Pesola et al. (2004,1938421
To determine if healthy African Americans may be
misdiagnosed has having respiratory deficient due to
comparison using Caucasian-derived prediction
equation estimates of DLCO.
The lung volume of African American individuals is 10-15% lower than
Caucasians. The measured DLCO was consistently significantly lower in
African Americans than what would be predicted, thus the authors
suggest a race correction reduction of the Miller PEE for diffusion of
12%.
Pesola et al. (2006,1938551
To determine if healthy Asians may be misdiagnosed
has having respiratory deficient due to comparison
using Caucasian-derived prediction equation estimates
of DLCO.
The lung volume of Asian individuals is 10-15% lower than Caucasians,
thus a Chinese derived prediction for DLCO should be used.
Prommer and Schmidt (2007,
1804211
To determine the error in total Hb mass measurements
using the optimized CO-rebreathing method due to loss
of CO to Mb
Optimal blood mixing (when venous and arterial blood COHb% are
equivalent) was determined to be after 6 min. A small volume of
administered CO leaves the vascular space (0.32% per min). 2.3%
increase in total Hb mass would be found if CO diffusion was not
included.
Proudman et al. (2007,1867051
To review the signs of pulmonary arterial hypertension,
including a drop in DLCO, in patients with systemic
sclerosis.
September 2009
B-3
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Purpose
Findings
Richardson et al. (2002, 0375131
To combine invasive vascular measures of arterial and
venous blood and muscle blood flow with noninvasive
magnetic spectroscopy of deoxy-myoglobin and high
energy phosphates to determine the effects of mild CO
poisoning (20% COHb) in humans during muscular
work.
Five humans were analyzed under normoxia, hypoxia, normoxia + CO
(20% COHb), and 100% 02 + CO. Maximum works rates and maximal
oxygen uptake were reduced in H, COnorm, and COhyper. CO and H
caused elevated blood flow. Net muscle CO uptake from blood was less
during 20% COHb trials than during normoxia and hypoxia (1-2%) trials.
Sakamaki et al. (2002,1867061
To evaluate the association of patients with aortic
aneurysm to the prevalence obstructive airway disease, obstruction
Patients with AA had lower FEV1 and DLCO than controls. Presence of
AA and male gender were associated with a higher risk of airway
Scharte et al. (2000, 1941121
To investigate whether exhaled CO concentrations are
increased in critically ill patients.
Critically ill patients had higher exhaled CO concentrations and higher
total CO production rates compared to healthy controls. No correlation
was found between exhaled CO concentration and venous or arterial
COHb.
Scharte et al. (2006, 1941151
To investigate the relationship between the severity of
illness and endogenous CO production in critically ill
patients.
CO production rates weakly correlated with the multiple organ
dysfunction score (R=0.27). Cardiac disease patients and patients
undergoing dialysis produced higher amounts of CO compared to
critically ill control patients.
Schachteretal. (2003,1867071
To evaluate the association between severe
gastroesophageal reflux and lung function.
Patients with severe gastroesophageal reflux had reduced DLCO,
remaining significant after adjusting for age, gender, BMI, and smoking.
Shimazu et al. (2000, 0164201
To study the effects of short-term (min) or long-term
(several h) CO exposure on COHb elimination and
developing a mathematical model to simulate this
event.
COHb exhibited an initial rapid decrease followed by a slower phase
which is compatible with a 2-compartment model and biphasic
elimination. Both exposures fit the 2-compartment, single central outlet
mathematical model.
Shimazu (2001, 0163311
To discuss the findings of Weaver et al. (2000, 0164211
onCOHbt1/2.
The authors discuss that CO elimination is biphasic and is heavily
affected by duration of exposure which was not taken into account in the
Weaver et al. (2000, 0164211 paper.
Sylvester et al. (2005,1919541
To assess the usage of end tidal CO levels in children
with sickle cell disease for measurement of hemolysis.
Children with sickle cell disease had higher exhaled CO levels (4.9 ppm;
SD 1.7 ppm) compared to healthy controls (1.3 ppm; SD 0.4 ppm). A
positive correlation existed between end tidal CO levels and COHb and
bilirubin.
Takeuchi et al. (2000, 0056751
To examine the relationship between min ventilation
and rate of COHb reduction during breathing 100% 02
and during normocapnic hyperoxic hyperpnea.
Patients were exposed to 400-1,000 ppm CO, resulting in 10-12%
COHb. The half-time of COHb reduction was 78 + 24 min during 100%
02 treatment and 31+6 min during normocapnic hyperpnea with 02
treatment.
Tarquini et al. (2009,1941171
To measure plasma CO levels in patients with liver
cirrhosis and portal hypertension.
Plasma CO was higher in ascetic patients than non-ascitic patients and
both were higher than healthy controls. HO activity was higher in
cirrhotic patients that healthy subjects and highest in patients with
ascites.
Terzano et al. (2009, 1080461
To investigate the effect of postural changes on gas
exchange in patients with COPD and healthy subjects.
DLCO increased in healthy individuals from upright to supine position
and upright to prone position. DLCO did not significantly change in
COPD patients from upright to prone position. This is explained by
homogeneous perfusion in healthy individuals and increased rigidity of
lung capillaries due to COPD.
Tran et al. (2007,1941201
To assess the correlation of COHb to severity of liver
disease.
No correlation was found with the Model for End Stage Liver Disease
score, Child Turcotte Pugh score, or other biochemical or clinical
measures of disease severity, such as spleen size, bilirubin, disease
duration, or AST/ALT The mean COHb was 2.1%.
Vreman et al. (2005,1937861
To develop a sensitive and reproducible method of CO
quantification in rodent (mouse and rat) tissue pre- and
post-exposure in hopes of understanding endogenous
CO production.
Tissues were sonicated mixed with sulfosalicylic acid for 30 min at 0°C
and then liberated CO was analyzed by gas chromatograph. Blood
contained the highest CO concentration. Lowest concentrations were
found in brain, testes, intestine, and lung (endogenously).
Vreman et al. (2006, 0982721
To test a method of CO quantification in frozen
postmortem human tissues from 3 determined
categories of fatalities: trauma with no suspected CO
exposure (controls), fire-related, and CO asphyxiation.
CO levels were analyzed in adipose, brain, muscle, heart, kidney, lung,
spleen, and blood (ordered from approximate low to high tissue
concentration). It was suggested that blood, muscle, brain, lung, and
kidney are suitable for diagnosing death due to lethal CO exposure due
to regression analysis against COHb values.
Weaver et al. (2000, 0164211
To determine in COHb half-life is influenced by CO
poisoning vs. experimental CO exposure, loss of
consciousness, concurrent tobacco smoking, or Pa02.
COHb t1/2 determined was 74 ± 25 min with a range from 26 to 148 min
by a single exponential decrease function. This is shorter than most
clinical studies and was inversely proportionate to Pa02, however not
influenced by age, gender, smoke inhalation, loss of consciousness,
tobacco smoking, or method of 02 treatment.
September 2009
B-4
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Purpose
Findings
Whincup et al. (2006, 1951291
To report COHb levels from a population-based study in
men aged 60-79 yr during the 20-yr follow-up of the
British Regional Heart Study cohort.
Mean COHb: 0.46%; Median COHb: 0.5%
9.2% of men had COHb levels of 2.5% or greater (93% were smokers)
0.1 % of men had COHb levels of 7.5% or greater
Smoking is the highest influence on COHb levels however other factors
independently related were season, region, gas cooking and central
heating, and active smoking
Widdop (2002, 0304931
To review carbon monoxide analysis methods,
including CO-oximeters and gas chromatography
Wu and Wang (2005,1804111
To review the endogenous production of CO through
HO, as well as discuss physiological roles for CO both
toxic and therapeutic.
CO is produced endogenously by HO-1 and -2 and acts as a
gasotransmitter, inducing cell signaling cascades. The review discusses
possible roles for CO in the various organ systems. Also, it discusses
the potential pharmacological and therapeutic applications for CO.
Yamavaetal. (1998, 0475251
To determine whether upper respiratory tract infections
increase exhaled CO concentrations.
Exhaled CO increased in patients at the time of upper respiratory tract
infection symptoms but decreased to nonsmoking healthy control levels
during recovery.
Yamavaetal. (2001,1801301
To determine whether the level of CO is related to the
severity of asthma.
Severe asthmatics exhaled more CO than non-smoking controls.
Exhaled CO concentrations in unstable severe asthmatics were higher
than in stable severe asthmatics. Mild and moderate asthmatics did not
differ from controls. Exhaled CO was correlated with FEV, in all
asthmatics.
Yasuda et al. (2002, 0352061
To determine whether arterial COHb is increased in
patients with inflammatory pulmonary diseases.
Arterial COHb concentrations are increased in patients with
inflammatory pulmonary diseases including exacerbated bronchial
asthma (1.05%), pneumonia (1.08%), and idiopathic pulmonary fibrosis
(1.03%) over controls (0.6%).
Yasuda et al. (2004,1919551
To determine if COHb levels in the venous blood and
arteriovenous COHb (a-vCOHb) differences are
increased in patients with inflammatory pulmonary
diseases compared to patients with extrapulmonary
inflammation and control subjects.
Patients with inflammatory pulmonary diseases including bronchial
asthma and pneumonia had a large a-vCOHb difference. Both arterial
and venous blood COHb increased in patients with inflammatory
pulmonary disease such as bronchial asthma, pneumonia,
pyelonephritis and active rheumatoid arthritis.
Yasuda et al. (2005,1021831
To study the relationship between COHb and disease
severity in patients with COPD.
COHb concentrations increased in patients with COPD at a stable
condition over controls and patients with COPD with exacerbations were
further increased.
Yerushalmi et al. (2009,1867111
Zegdi et al. (2002, 0374611
To compare endogenous CO production in
mechanically ventilated critically ill adult patients with
and without severe sepsis.
CO production was higher in septic patients during the first 3 days of
treatment compared to controls. Survivors of sepsis had a significantly
higher CO production compared to non-survivors.
September 2009
B-5
DRAFT - DO NOT CITE OR QUOTE
-------
References
Aberg AM; Sojka BN; Winso O; Abrahamsson P; Johansson G; Larsson JE. (2009). Carbon monoxide concentration in
donated blood: Relation to cigarette smoking and other sources. Transfusion, 49: 347-353. 194082
Abram SR; Hodnett BL; Summers RL; Coleman TG; Hester RL. (2007). Quantitative circulatory physiology: An
integrative mathematical model of human physiology for medical education. Adv Physiol Educ, 31: 202-210.
193859
Adir Y; Merdler A; Haim SB; Front A; Harduf R; Bitterman H. (1999). Effects of exposure to low concentrations of carbon
monoxide on exercise performance and myocardial perfusion in young healthy men. Occup Environ Med, 56: 535-
538. 001026
Alcantara RE; Xu C; Spiro TG; Guallar V. (2007). A quantum-chemical picture of hemoglobin affinity. Proc Natl Acad Sci
USA, 104: 18451-18455. 193867
Andersson JA; Uddman R; Cardell L-O. (2000). Carbon monoxide is endogenously produced in the human nose and
paranasal sinuses. J Allergy Clinlmmunol, 105: 269-273. 011836
AroraM; Chandra J; Suri JC;Narayan S; DuttaAK. (2001). Pulmonary function tests in beta thalassemia. Indian J Pediatr,
68: 239-242. 186713
Benignus VA; Coleman T; Eklund CR; Kenyon EM. (2006). A general physiological and toxicokinetic (GPAT) model for
simulating complex toluene exposure scenarios in humans. Toxicol Mech Meth, 16: 27-36. 151344
Bos PM; Zeilmaker MJ; van Eijkeren JC. (2006). Application of physiologically based pharmacokinetic modeling in
setting acute exposure guideline levels for methylene chloride. Toxicol Sci, 91: 576-585. 194084
Bruce EN; Bruce MC. (2003). A multicompartment model of carboxyhemoglobin and carboxymyoglobin responses to
inhalation of carbon monoxide. J Appl Physiol, 95: 1235-1247. 193975
Bruce EN; Bruce MC; Erupaka K. (2008). Prediction of the rate of uptake of carbon monoxide from blood by extravascular
tissues. Respir Physiol Neurobiol, 161: 142-159. 193977
Bruce MC; Bruce EN. (2006). Analysis of factors that influence rates of carbon monoxide uptake, distribution, and
washout from blood and extravascular tissues using a multicompartment model. J Appl Physiol, 100: 1171-1180.
193980
Carraway MS; Ghio AJ; Carter JD; Piantadosi CA. (2000). Expression of heme oxygenase-1 in the lung in chronic
hypoxia. Am J Physiol, 278: L806-L812. 021096
Castillo A; Llapur CJ; Martinez T; Kisling J; Williams-Nkomo T; Coates C; Tepper RS. (2006). Measurement of single
breath-hold carbon monoxide diffusing capacity in healthy infants and toddlers. Pediatr Pulmonol, 41: 544-550 .
193234
Chakraborty S; Balakotaiah V; Bidani A. (2004). Diffusing capacity reexamined: relative roles of diffusion and chemical
reaction in red cell uptake of O2, CO, CO2, and NO . J Appl Physiol, 97: 2284-2302. 193759
Cronenberger C; Mould DR; Roethig HJ; Sarkar M. (2008). Population pharmacokinetic analysis of carboxyhaemoglobin
concentrations in adult cigarette smokers. Br J Clin Pharmacol, 65: 30-39. 194085
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA. (2004). Carbon monoxide actuates O2-limited heme
degredations in the rat brain. Free Radic Biol Med, 3 7: 1802-1812. 180440
De las Heras D; Fernandez J; Gines P; Cardenas A; Ortega R; Navasa M; Barbera JA; Calahorra B; Guevara M; Bataller R;
Jimenez W; Arroyo V; Rodes J. (2003). Increased carbon monoxide production in patients with cirrhosis with and
without spontaneous bacterial peritonitis. Hepatology, 38: 452-459. 194087
Dutton SJ; Hannigan MP; Miller SL. (2001). Indoor pollutant levels from the use of unvented natural gas fireplaces in
Boulder, Colorado. J Air Waste ManagAssoc, 51: 1654-1661. 021307
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 B-6 DRAFT - DO NOT CITE OR QUOTE
-------
Ehlers M; Labaze G; Hanakova M; McCloskey D; Wilner G. (2009). Alarming levels of carboxyhemoglobin in banked
blood. J Cardiothorac Vase Anesth, 23: 336-338 . 194089
Gosselin NH; Brunei RC; Carrier G. (2009). Determination of carboxyhaemoglobin in humans following low-level
exposures to carbon monoxide. Inhal Toxicol, 0: 00-00. 190946
Hampson NB; Weaver LK. (2007). Carbon monoxide poisoning: a new incidence for an old disease. Undersea Hyperb
Med, 34: 163-8. 190272
Hart CL; Smith GD; Hole DJ; Hawthorne VM. (2006). Carboxyhaemoglobin concentration, smoking habit, and mortality
in 25 years in the Renfrew/Paisley prospective cohort study. Heart, 92: 321-324. 194092
Hsia CC. (2002). Recruitment of lung diffusing capacity: Update of concept and application. Chest, 122: 1774-1783.
193857
Johnson FK; Johnson RA; Durante W; Jackson KE; Stevenson BK; Peyton KJ. (2006). Metabolic syndrome increases
endogenous carbon monoxide production to promote hypertension and endothelial dysfunction in obese Zucker
rats. Am J Physiol Regul Integr Comp Physiol, 290: 601-608. 193874
Lamberto C; Nunes H; Le Toumelin P; Duperron F; Valeyre D; Clerici C. (2004). Membrane and capillary blood
components of diffusion capacity of the lung for carbon monoxide in pulmonary sarcoidosis: relation to exercise
gas exchange. Chest, 125: 2061-2068. 193845
Levesque B; Allaire S; Prud'homme H; Dupuis K; Bellemare D. (2000). Air quality monitoring during indoor monster
truck and car demolition shows. J Expo Sci Environ Epidemiol, 10: 58-65. 011886
Lim S; Groneberg D; Fischer A; Gates T; Caramori G; Mattos W; Adcock I; Barnes PJ; Chung KF. (2000). Expression of
heme oxygenase isoenzymes 1 and 2 in normal and asthmatic airways: Effect of inhaled corticosteroids. Am J
Respir Grit Care Med, 162: 1912-1918. 126969
Mahoney JJ; Vreman HJ; Stevenson DK; Van Kessel AL. (1993). Measurement of carboxyhemoglobin and total
hemoglobin by five specialized spectrophotometers (CO-oximeters) in comparison with reference methods. Clin
Chem, 39: 1693-1700. 013859
Marks GS; Vreman HJ; McLaughlin BE; Brien JF; Nakatsu K. (2002). Measurement of endogenous carbon monoxide
formation in biological systems. Antioxid Redox Signal, 4: 271-277. 030616
Marvisi M; Bassi E; Bonassi R; Civardi G; Delsignore R. (2007). DLCO correlates with intestinal inflammation in
ulcerative colitis, but albuminuria does not. Minerva Gastroenterol Dietol, 53: 321-7. 186702
Merx MW; Flogel U; Stumpe T; Godecke A; Decking UK; Schrader J. (2001). Myoglobin facilitates oxygen diffusion.
FASEB J, 15: 1077-1079. 002006
Monma M; Yamaya M; Sekizawa K; Ikeda K; Suzuki N; Kikuchi T; Takasaka T; Sasaki H. (1999). Increased carbon
monoxide in exhaled air of patients with seasonal allergic rhinitis. Clin Exp Allergy, 29: 1537-1541. 180426
Morimatsu H; Takahashi T; Maeshima K; Inoue K; Kawakami T; Shimizu H; Takeuchi M; Yokoyama M; Katayama H;
Morita K. (2006). Increased heme catabolism in critically ill patients: Correlation among exhaled carbon monoxide,
arterial carboxyhemoglobin, and serum bilirubin IXalpha concentrations. Am J Physiol Lung Cell Mol Physiol,
290: L114-L119.194097
Muchova L; Wong RJ; Hsu M; Morioka I; Vitek L; Zelenka J; Schroder H; Stevenson DK. (2007). Statin treatment
increases formation of carbon monoxide and bilirubin in mice: A novel mechanism of in vivo antioxidant
protection. Can J Physiol Pharmacol, 85: 800-810. 194098
Neto CA; Yanagihara JI; Turri F. (2008). A carbon monoxide transport model of the human respiratory system applied to
urban atmosphere exposure analysis. J Braz Chem Soc, 30: 253-260. 194672
Paredi P; Biernacki W; Invernizzi G; Kharitonov SA; Barnes PJ. (1999). Exhaled carbon monoxide levels elevated in
diabetes and correlated with glucose concentration in blood: A new test for monitoring the disease?. Chest, 116:
1007-1011. 194102
Paredi P; Shah PL; Montuschi P; Sullivan P; Hodson ME; Kharitonov SA; Barnes PJ. (1999). Increased carbon monoxide
in exhaled air of patients with cystic fibrosis. Thorax, 54: 917-20. 118798
Pelham TW; Holt LE; Moss MA. (2002). Exposure to carbon monoxide and nitrogen dioxide in enclosed ice arenas. Occup
Environ Med, 59: 224-233. 025716
September 2009 B-7 DRAFT - DO NOT CITE OR QUOTE
-------
Pesola GR; Huggins G; Sherpa TY. (2006). Abnormal predicted diffusion capacities in healthy Asians: an inequality with a
solution. Respiration, 73: 799-807. 193855
Pesola GR; Sunmonu Y; Huggins G; Ford JG (2004). Measured diffusion capacity versus prediction equation estimates in
blacks without lung disease. Respiration, 71: 484-492. 193842
Prommer N; Schmidt. (2007). Loss of CO from the intravascular bed and its impact on the optimised CO-rebreathing
method. Eur J Appl Physiol, 100: 383-391. 180421
Proudman SM; Stevens WM; Sahhar J; Celermajer D. (2007). Pulmonary arterial hypertension in systemic sclerosis: the
need for early detection and treatment. J Intern Med, 37: 485-94. 186705
Richardson RS; Noyszewski EA; Saltin B; Gonzalez-Alonso J; et al. (2002). Effect of mild carboxy-hemoglobin on
exercising skeletal muscle: Intravascular and intracellular evidence. Am J Physiol, 283: R1131-R1139. 037513
Sakamaki F; Oya H; Nagaya N; Kyotani S; Satoh T; Nakanishi N. (2002). Higher prevalence of obstructive airway disease
in patients with thoracic or abdominal aortic aneurysm. J Vase Surg, 36: 35-40. 186706
Schachter L. M; Dixon J; Pierce R. J; O'Brien P. (2003). Severe gastroesophageal reflux is associated with reduced carbon
monoxide diffusing capacity. Chest, 123: 1932-8. 186707
Scharte M; Bone HG; Van Aken H; Meyer J. (2000). Increased carbon monoxide in exhaled air of critically ill patients.
Biochem Biophys Res Commun, 267: 423-426. 194112
Scharte M; von Ostrowski TA; Daudel F; Freise H; Van Aken H; Bone HG. (2006). Endogenous carbon monoxide
production correlates weakly with severity of acute illness. Eur J Anaesthesiol, 23: 117-122. 194115
ShimazuT. (2001). Half-life of blood carboxyhemoglobin. Chest, 119: 661-663. 016331
Shimazu T; Ikeuchi H; Sugimoto H; Goodwin CW; Mason AD Jr; Pruitt BA Jr. (2000). Half-life of blood
carboxyhemoglobin after short-term and long-term exposure to carbon monoxide. J Trauma, 49: 126-131. 016420
Sylvester KP; Patey RA; Rafferty GF; Rees D; Thein SL; Greenough A. (2005). Exhaled carbon monoxide levels in
children with sickle cell disease. Eur J Pediatr, 164: 162-165. 191954
Takeuchi A; Vesely A; Rucker J; Sommer LZ; Tesler J; Lavine E; Slutsky AS; Maleck WA; Volgyesi G; Fedorko L; Iscoe
S; Fisher JA. (2000). A simple "new" method to accelerate clearance of carbon monoxide. Am J Respir Crit Care
Med, 161: 1816-1819. 005675
Tarquini R; Masini E; La Villa G; Barletta G; Novelli M; Mastroianni R; Romanelli RG; Vizzutti F; Santosuosso U; Laffi
G. (2009). Increased plasma carbon monoxide in patients with viral cirrhosis and hyperdynamic circulation. Am J
Gastroenterol, 104: 891-897. 194117
Terzano C; Conti V; Petroianni A; Ceccarelli D; De Vito C; Villari P. (2009). Effect of Postural Variations on Carbon
Monoxide Diffusing Capacity in Healthy Subjects and Patients with Chronic Obstructive Pulmonary Disease.
Respiration, 77: 51-57. 108046
Iran TT; Martin P; Ly H; Balfe D; Mosenifar Z. (2007). Carboxyhemoglobin and its correlation to disease severity in
cirrhotics. , 41: 211-215. 194120
Vreman HJ; Wong RJ; Kadotani T; Stevenson DK. (2005). Determination of carbon monoxide (CO) in rodent tissue: effect
of heme administration and environmental CO exposure. Anal Biochem, 341: 280-289. 193786
Vreman HJ; Wong RJ; Stevenson DK; Smialek JE; Fowler DR; Li L; Vigorito RD; Zielke HR. (2006). Concentration of
Carbon Monoxide (CO) in Postmortem Human Tissues: Effect of Environmental CO Exposure. J Forensic Sci, 51:
1182-1190.098272
Weaver LK; Howe S; Hopkins R; Chan K. (2000). Carboxyhemoglobin half-life in carbon monoxide-poisoned patients
treated with 100% oxygen at atmospheric pressure. Chest, 117: 801-808. 016421
Whincup P; Papacosta O; Lennon L; Haines A. (2006). Carboxyhaemoglobin levels and their determinants in older British
men. BMC Public Health, 6: 189. 195129
Widdop B. (2002). Analysis of carbon monoxide. Ann Clin Biochem, 39: 378-391. 030493
WuL; Wang R. (2005). Carbon monoxide: endogenous production, physiological functions, and pharmacological
applications. Pharmacol Rev, 57: 585-630. 180411
September 2009 B-8 DRAFT - DO NOT CITE OR QUOTE
-------
Yamaya M; Hosoda M; Ishizuka S; Monma M; Matsui T; Suzuki T; Sekizawa K; Sasaki H. (2001). Relation between
exhaled carbon monoxide levels and clinical severity of asthma. Clin Exp Allergy, 31: 417-422. 180130
Yamaya M; Sekizawa K; Ishizuka S; Monma M; Mizuta K; Sasaki H. (1998). Increased carbon monoxide in exhaled air of
subjects with upper respiratory tract infections. Am J Respir Crit Care Med, 158: 311-314. 047525
Yasuda H; Sasaki T; Yamaya M; Ebihara S; Maruyama M; Kanda A; Sasaki H. (2004). Increased arteriovenous
carboxyhemoglobin differences in patients with inflammatory pulmonary diseases. Chest, 125: 2160-2168. 191955
Yasuda H; Yamaya M; Nakayama K; Sasaki H. (2005). Blood carbon monoxide will increase from a decline in pulmonary
function alone 7]. , 172: 1231-1232. 102183
Yasuda H; Yamaya M; Yanai M; Ohrui T; Sasaki H. (2002). Increased blood carboxyhaemoglobin concentrations in
inflammatory pulmonary diseases. Thorax, 57: 779-783. 035206
Yerushalmi R; Kramer MR; Rizel S; Sulkes A; Gelmon K; Granot T; Neiman V; Stemmer SM. (2009). Decline in
pulmonary function in patients with breast cancer receiving dose-dense chemotherapy: a prospective study. Ann
Oncol, 20: 437-440. 186711
Zegdi R; Perrin D; Burdin M; Boiteau R; Tenaillon A. (2002). Increased endogenous carbon monoxide production in
severe sepsis. Intensive Care Med, 28: 793-796. 037461
September 2009 B-9 DRAFT - DO NOT CITE OR QUOTE
-------
Annex C. Epidemiology Studies
Table C-1 Studies of CO exposure and cardiovascular morbidity.
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CHANGES IN HEART RATE AND HEART RATE VARIABILITY
Author: Chan et al. (2005,
088988)
Period of Study: December
2001-February 2002
Location:
Taipei, Taiwan
Health Outcome: Various measures of HRV
via ambulatory ECG (Holler system)
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects)
Age Groups Analyzed:
40-75 yr
Sample Description:
83 patients from the National Taiwan
University Hospital
Averaging Time:
1-h ma
Mean (SD) unit:
1.1 ppm
Range (Min, Max):
0.1,7.7
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined (-h ma): 1,2,3,4, 5,6, 7,8
CO had no statistically significant effect on SDNN,
rMSSD, LF, HF.
Author: Dales et al. (2004,
099036)
Period of Study: NR
Location:
Toronto, Canada.
Health Outcome: Various measures of HRV
via Holler system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects)
Age Groups Analyzed:
51-88yr (mean 65 yr)
Sample Description: 36 subjects with pre-
existing CAD
Averaging Time: 24-h Increment: NR
Regression co-efficient [Lower Cl, Upper Cl]
Mean (SD) unit:
2.40 ppm (95th
percentile) Personal
monitoring
Range (Min, Max):0.4,
16.5
Copollutant:
correlation PM25:
r = 0.17
Lags examined : NR
CO had no statistically significant effect on LF, HF,
HFLFR, SDNN among those taking Beta-blockers.
Whereas CO had a positive effect on SDNN among
those not taking Beta-blockers. Slope = 0.0111
(0.002-0.020, p= 0.02)
Author: Gold et al. (2000,
011432)
Period of Study:
June-September 1997
Location:
Boston, MA
Health Outcome (ICD9 or ICD10): Heart
Rate and various measures of HRV via
Holler system
Study Design: Panel/Cohort
Statistical Analyses: Linear regression
(fixed effects/random effects)
Age Groups Analyzed:
53-87 yr
Sample Description: 21 active Boston
residents observed up to 12 times.
Averaging Time: 24-h Increment: 0.6 ppm
% Change [Lower Cl, Upper Cl]
Lags examined : 24-h
No significant effect with CO (no results recorded)
Mean (SD) unit:
0.47 ppm
Range (Min, Max):
0.12,0.82
Copollutant: NR
Author: Gold et al. (2005,
087558)
Period of Study:
June-September 1999
Location:
Boston, MA
Health Outcome: ST- segment.
Study Design: Panel
Statistical Analyses: Linear regression
(mixed models)
Age Groups Analyzed:
61-88yr
Sample Description: 24 Active Boston
residents-each observed up to 12 times.
Averaging Time:
1 Oh, 24-h
Mean (SD) unit: NR
Range (Min, Max):
(ppm) (personal
monitoring)
10th = 0.20
90th = 1.08
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined : 1 24-h
Although CO was associated with ST-segment
depression in single pollutant models, this result did
not persist in multiple pollutant models.
September 2009
C-1
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Goldberg et al. (2008,
180380)
Period of Study:
July 2002-October 2003
Location:
Montreal, Quebec
Health Outcome: Oxygen Saturation and
Heart Rate
Study Design: Panel
Statistical Analyses: Mixed regression
models
Age Groups Analyzed: 50-85 yr
Sample Description: 31 subjects with CHF
and limits in physical functioning in the Heart
Failure and Heart Transplant Center at the
McGill University Health Center
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
PM25:r = 0.72
N02:r = 0.84
S02 and N02:r = 0.43
Increment: NR
Adjusted Mean Difference [Lower Cl, Upper Cl]
Lags examined: 0,1,2
Oxygen Saturation:
Lag 0:0.004 ppm (-0.060,0.067)
Lag 1:-0.001 ppm (-0.066, 0.065)
3-day: -0.005 ppm (-0.098.0.088)
Pulse Rate:
Lag 0:0.011 ppm (-0.290, 0.312)
Lag 1:0.227 ppm (-0.080,0.535)
3-day: 0.245 ppm (-0.209, 0.700)
Author: Holguin et al. (2003,
057326)
Period of Study:
February-April 2000
Location:
Mexico City, Mexico
Author: Ibald-Mullietal.
(2004, 087415)
Period of Study:
1998-1999
Location: Helsinki,
Finland
Erfurt, Germany
Amsterdam, Netherlands
Health Outcome: Various measures of HRV
via ECG
Study Design: Panel
Statistical Analyses: GEE
Age Groups Analyzed:
60-96 yr (mean age 79 yr)
Sample Description:
34 patients who were permanent residents of
a nursing home in the Northeast metropolitan
area.
Health Outcome: BP and HR via ECG
Study Design: Panel
Statistical Analyses: Linear regression
Age Groups Analyzed: > 50 yr
Sample Description: 131 nonsmokers with
coronary heart disease
Averaging Time: 24-h
Mean (SD) unit:
3.3 ppm
Range (Min, Max): 1.8,
4.8
Copollutant: NR
Averaging Time: 24-h
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfurt: 0.4 mg/m3
Helsinki: 0.4 mg/m3
Range (Min, Max):
Amsterdam: 0.4, 1.6
Erfurt: 0.1, 2.5
Helsinki: 0.1, 1.0
Copollutant:
Amsterdam
PM25:r = 0.58 ug/m3
N02:r = 0.76 ug/m3
S02:r= 0.50 mg/m3
UFP:r=0.22n/cm3
ACP:r = 0.60n/cm3
Erfurt
PM25:r = 0.77 ug/m3
N02:r = 0.86 ug/m3
S02:r= 0.68 mg/m3
UFP:r=0.72n/cm3
ACP:r = 0.78n/cm3
Helsinki
PM25:r = 0.40 ug/m3
N02:r = 0.32 ug/m3
S02:r=0.19mg/m3
UFP:r=0.35n/cm3
ACP:r = 0.51 n/cm3
Increment: 10 ppm
Regression Coefficients [Lower Cl, Upper Cl]
Lags examined : 0
Lag 0 :
HF: 0.003 (-0.004 to 0.001)
LF: 0.001 (-0.006 to 0.008)
LF/HF: 0.001 (-0.005 to 0.002)
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined: 0,1,2,3
Results presented graphically
September 2009
C-2
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Liao et al. (2004,
056590)
Period of Study:
1996-1998
Location:
Forsyth County, NC; Selected
suburbs of Minneapolis, MN;
Jackson, Ml
Health Outcome: Heart Rate & various
rates of HRV.
Study Design: Cohort
Statistical Analyses: Linear regression
Age Groups Analyzed:
45-64 yr (mean 62 yr)
Sample Description:
6784 study subjects from the atherosclerosis
risk in communities study
Averaging Time: 24-h
Mean (SD) unit:
0.65 ppm (0.44)
Range (Min, Max): NR
Copollutant: NR
Increment: 0.44 ppm
Regression coefficients Lags examined : 1
Lagt:
HF (log transformed) : -0.033
LF (log transformed): 0.006
SDNN :-0.274
Heart Rate (bpm): 0.404*
Confidence Intervals not recorded
*p < 0.05
Author: Park etal. (2005,
057331)
Period of Study:
2000-2003
Location:
Boston, MA
Health Outcome: Various measures of HRV Averaging Time: 24-h
via ECG
Mean (SD) unit:
0.50 ppm
Study Design: Panel/Cohort
Statistical Analyses: Linear regression
Age Groups Analyzed:
21-81yr
Sample Description:
497 men from the Normative aging study in
Greater Boston
Range (Min, Max):
0.13,1.8
Copollutant: NR
Increment: 0.24 ppm
% Change in HRV [Lower Cl, Upper Cl]
Lags examined: 4-h ma, 24-h ma, 48-h ma
Lag 4-h ma:
SDNN (LoglO): 2.0 (-2.9to 7.3)
HF(Log10):8.8(-4.6to24.1)
LF(Log10) : 3.2 (-7.0 to 14.6)
LF :HF(Log10) :-5.1 (-13.5 to 4.1)
Lag 24-h ma:
SDNN (LoglO):-2.2 (-7.7 to 3.6)
HF(Log10):-13.2 (-25.4to 1.0)
LF(Log10) :-0.6 (-11.9 to 12.1)
LF :HF(Log10) : 14.5 (2.9-27.5)
Lag 48-h ma:
SDNN(Log10):-3.4(-10.2to3.9)
HF (LoglO):-13.8 (-.28.9 to 4.4)
LF (LoglO):-2.4 (-16.2 to 13.6)
LF:HF (LoglO): 13.2 (-1.1 to 29.6)
Author: Peters etal. (1999,
011554)
Period of Study:
1984-1985
Location:
Augsburg, Germany
Health Outcome: Heart Rate
Study Design: Cohort
Statistical Analyses: Linear regression
(GEE)
Age Groups Analyzed:
25-64 yr
Sample Description:
2681 men & women who participated in the
MONICA study
Averaging Time: 24-h Increment: 6.6 mg/m3
Mean (SD) unit:
During air pollution
episode: 4.54 mg/m
Outside air pollution
episode: 4.51 mg/m3
Range (Min, Max):
During air pollution
episode: 2.39,6.85
Outside air pollution
episode: 0.91,11.51
Respectively
Copollutant: NR
Mean Change in Heart Rate (beats/min) [Lower
Cl, Upper Cl]
Lags examined: 0, 5-day avg
All
Lag 0:0.97 (0.02-1.91)
Lag 5-day avg : 0.70 (-0.09 to 1.48)
Men
Lag 0 :0.95 (-0.37 to 2.27)
Lag 5-day avg : 0.91 (-0.25 to 2.07)
Women
Lag 0 :0.98 (-0.37 to 2.34)
Lag 5-day avg : 0.52 (-0.55 to 1.59)
Author: Riojas-Rodriguez et
al. (2006.156913)
Period of Study: December
2001-April 2002
Location:
Mexico City, Mexico
Health Outcome: Various measures of HRV
via Holler system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects models)
Age Groups Analyzed:
25-76 yr (mean 55 yr)
Sample Description:
30 patients from the Outpatient clinic of the
National Institute of Cardiology of Mexico
Averaging Time: 24-h Increment: 1 ppm
Regression Coefficients [Lower Cl, Upper Cl]
Lags examined (per min): 5,10
Lag 5 min :
HF:-0.006 (-0.023 to 0.010)
LF :-0.024 (-0.041 to-0.007)
VLF : -0.034 (-0.061 to -0.007)
Notes: VLF = Very low frequency
Mean (SD) unit:
2.9 ppm
(personal monitor)
Range (Min, Max): 0.1
18.0
Copollutant: NR
September 2009
C-3
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Schwartz et al. (2005, Health Outcome: Measures of HRVvia
Holler system
074317'
Period of Study: 1999
Location:
Boston, MA
Study Design: Panel
Statistical Analyses: Linear regression
(hierarchical model)
Age Groups Analyzed:
61-89yr
Sample Description:
28 subjects living at or near an apartment
complex located on the same street at the
Harvard School of Public Health
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min,
Max): ppm
25th = 0.38; 75th =
0.54
Copollutant:
correlation
PM25:r = 0.61
N02:r = 0.55
S02:r=-0.18
03:r = 0.21
Increment: 0.16 ppm
% Change in HRV [Lower Cl, Upper Cl]
Lags examined :24-h, 1 h
Lag 1 h:
SDNN : -2.6 (-5.6 to 0.5); rMSSD : -3.9 (-10.6 to 3.3);
PNN50 : -3.5 (-13.7 to 8.0); LF :HF : 4.5 (-1.2 to
10.5)
Lag 24-h:
SDNN : -4.2 (-0.6 to -7.7); rMSSD : -10.2 (-2.4 to -
17.4);
PNN50 : -14.8 (-3.0 to -25.2); LF :HF : 6.2 (-0.6 to
13.4)
Author: Tarkiainen et al.
(2003, 053625)
Period of Study:
October 1997-May 1998
Location:
Kuopio, Finland
Health Outcome: Various measures of HRV
via Ambulatory ECG (Holler system)
Study Design: Panel
Statistical Analyses: ANOVAfor repeated
errors (GLM)
Age Groups Analyzed:
Age 55-68 yr
Sample Description: 6 male patients with
angiographically verified CAD
Averaging Time: 24-h
Mean (SD) unit:
4.6 ppm (max of CO
episode) (personal
monitoring)
Range (Min, Max): 0.5,
27.4 (max of CO
episode)
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined : 5 min prior to CO episode, 5 min
during CO episode
CO had no statically significant effect on NN, SDNN
or rMSSD. However, during high CO exposure
(>2.7 ppm) CO was associated with an increase in
rMSSD of 2.4ms (p=0.034).
Author: Timonen et al. (2006,
088747)
Period of Study:
1998-1999
Health Outcome:
Stable CAD: Various measures of HRVvia
ambulatory ECG (Holler system)
Study Design: Panel
Averaging Time: 24-h Increment: 1 mg/m3
Regression co-efficient [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2, 3, 5-day avg
Location: Statistical Analyses: Linear regression
3 Cities in Europe: Amsterdam, (mixed model)
Netherlands; Erfert, Germany;
Helsinki Finland A9e Groups Analyzed: Mean age across 3
cities; 64-71 yr.
Sample Description:
131 subjects with Stable CAD followed for 6
mo with bi-weekly clinical visits.
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfert: 0.4 mg/m3
Helsinki: 0.4 mg/m3
Range (Min, Max):
Amsterdam: 0.4,1.6
Erfert: 0.1, 2.5
Helsinki: 0.1,1.0
Copollutant:
correlation
Amsterdam:
PM25:r = 0.58
N02:r = 0.76
Erfert:
PM10:r = 0.77
N02:r = 0.86
Helsinki:
PM10:r = 0.40
N02:r = 0.32
SDNN:
Lag 0 : -1.21 (-4.44 to 2.03); Lag 1 : -1.71 (-6.05 to
2.63); Lag 2 : -5.69 (-10.7 to -0.72); Lag 3 :0.66 (-
3.83 to 5.15);
5-day avg:-3.60 (-9.88 to 2.68)
HF:
Lag 0:5.0 (-15.1 to 25.1); Lag 1 :-2.0 (-37.1 to
33.1);
Lag 2 : -30.7 (-59.8 to -1.5); Lag 3 : -9.3 (-35.8 to -
17.3);
5-day avg:-15.2 (-53.0 to 22.6)
LF/HF:
Lag 0 : -3.6 (-21.8 to 14.5); Lag 1 : -28.6 (-52.0 to -
5.3);
Lag 2 : -10.1 (-36.9 to 16.7); Lag 3 :7.7 (-16.5 to
31.9);
5-day avg:-16.9 (-51.2 to 17.3)
Author: Wheeler etal. (2006,
088453)
Period of Study:
1999-2000
Location:
Atlanta, GA
Health Outcome: Various measures of HRV Averaging Time: 1 h
via Holler system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects models)
Age Groups Analyzed:
Mean65yr-IQR55-73yr.
Sample Description:
18 subjects with COPD and 12 subjects with
recent Ml.
Mean (SD) unit:
362.0 ppb
Range (Min, Max):
25th = 221.5; 75th =
398.1
Copollutant:
correlation
PM2.5: r = 0.43
Increment: NR
RR Estimate [Lower Cl, Upper Cl]; lag :
Lags examined (h ma): 1, 4,24
No CO results reported.
September 2009
C-4
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
ONSET OF CARDIAC ARRHYTHMIA
Author: Bergeretal. (2006,
098702)
Period of Study:
October 2000-April 2001
Location:
Erfurt, Germany
Health Outcome:
Runs of supraventricular and ventricular
tachycardia recorded via 24-h ECG.
Study Design: Panel
Statistical Analyses:
Poisson regression (GAM) Linear regression
Age Groups Analyzed:
52-76 yr (mean 76years)
Sample Description:
57 men with CHD
Averaging Time: 24-h
Mean (SD) unit: 0.52
mg/m3
Range (Min, Max):
0.11,1.93
Copollutant:
correlation NR
Increment:
All: 0.27 mg/m3
5-day avg :0.22 mg/m3
RR Estimate [Lower Cl, Upper Cl]
Lags examined (h): 0, 0-23,24-47,48-71, 72-95,
5-day avg
Supraventricular extrasystoles:
Lag 0 :1.18 (1.00-1.38) Lag 0-23 :1.16 (1.02-1.31);
Lag 24-47 :1.13 (1.00-1.28); Lag 48-71 :1.18
(1.03-1.36);
Lag 72-95 :1.08 (0.98-1.20); 5-day avg: 1.18
(1.04-1.35)
Mean % Change [Lower Cl, Upper Cl]
Hourly Lags examined:
0, 0-23,24-47,48-71, 72-95, 5-day avg
Ventricular extrasystoles:
Lag 0 :0.0 (-4.1 to 4.4); Lag 0-23 :1.1 (-3.3 to 5.7);
Lag 24-47 :1.9 (-2.6 to 6.6); Lag 48-71 :4.2 (-0.3 to
8.9);
Lag 72-95 :2.7 (-1.3 to 6.9); 5-day avg: 3.0 (-1.8 to
8.0)
Author: Dockery et al. (2005,
078995)
Period of Study:
1995-2002
Location:
Boston, MA
Health Outcome:
Tachyarrhythmias:
Study Design: Panel
Statistical Analyses: Logistic regression
(GEE)
Age Groups Analyzed:
19-90 yr; mean age 64 yr
Sample Description:
203 cardiac patients with ICDs within 40km
of air monitoring site at Harvard School of
Public Health, Boston
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
25th = 0.53; 75th =
1.02
Copollutant: NR
Increment: 0.48 ppm
OR for Ventricular Arrhythmia [Lower Cl, Upper
Cl]
Lags examined (days): 0,1, 2, 3
Lag 2day ma: 1.14 (0.95-1.29)
Among those who had an Arrhythmia -
within 3 days: 1.65 (1.17-2.33)
later than 3 days : 1.04 (0.83-1.29)
Author: Metzgeretal. (2007,
092856)
Period of Study:
1993-2002
Location:
Atlanta, GA
Health Outcome:
Cardiac Arrhythmia, ICD,
Ventricular tachyarrhythmia
Study Design: Panel
Statistical Analyses:
Logistic regression (GEE)
Averaging Time: 1 h
Mean (SD) unit:
1 .7 ppm
Range (Min, Max): 0.1,
7.7
ftnnnllnfeint1 MR
Increment: 1 ppm
OR for Tachyarrhythmic event [Lower Cl, Upper
Cl]
Lags examined (days) : 0
Results for All events
I an n • n QQQ in Q7H.1 n?R\
Age Groups Analyzed:
15-88yr
Sample Description:
518 patients with ICDs with at least one
ventricular tachyarrhythmic event
Events resulting in cardiac pacing or defibrillation
Lag 0:1.008 (0.964-1.054)
Events resulting defibrillation
Lag 0:1.012 (0.925-1.10.7)
September 2009
C-5
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Peters et al. (2000,
011347)
Period of Study:
1995-1997
Location:
Eastern Massachusetts
Author: Rich et al. (2004,
055631)
Period of Study:
February-December 2000
Location:
Vancouver, Canada
Author: Rich et al. (2005,
079620)
Period of Study:
1995-1999
Location:
Boston, MA
Author: Rich et al. (2006,
089814)
Period of Study:
2001 & 2002
Location:
St. Louis, MO
Design
Health Outcome:
Defibrillated discharges for ventricular
tachycardia or fibrillation
Study Design: Panel
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Mean age of 62 yr
Sample Description:
100 patients with ICDs
Health Outcome: Cardiac arrhythmia via
patients ICD
Study Design:
Case-crossover
Statistical Analyses:
Conditional Logistic regression
Age Groups Analyzed:
15-85yr
Sample Description:
34 patients who experienced at least 1 ICD
discharge (8201 person days)
Health Outcome: Ventricular arrhythmias
via ICD
Study Design: Panel/Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
203 patients with implanted ICD at the New
England Medical Center
Health Outcome: Ventricular arrhythmia
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
60 subjects with at least 1 ICD recorded
arrhythmia who lived within 40 km of
St. Louis - Midwest supersite.
Concentrations
Averaging Time: 24-h
Mean (SD) unit:
0.58 ppm
Range (Min, Max):
25th = 0.43; 75th =
0.66
Copollutant:
correlation
PM10:r = 0.51
PM2.5: r = 0.56
N02: r = 0.71
S02:r= 0.41
03:r= -0.40
Averaging Time: 24-h
Mean (SD) unit:
553.8 ppb
Range (Min, Max):
IQR: 162.7
Copollutant:
correlation
PM10:r = 0.40
S02:r=0.75
N02'r=068
03:r=-0.56
Averaging Time:
1-h&24-h
Mean (SD) unit: NR
Range (percentiles):
1h:
25th = 0.46
75th = 1 .04
24-h:
25th = 0 52
75th = 1 .03
Copollutant: NR
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
25th = 0.4; 75th = 0.6
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment:
0.65 ppm (Lags 0, 1 , 2, 3); 0.42 ppm (Lag 5-day
mean)
OR for Defibrillated Discharge [Lower Cl, Upper
Cl]
J
Lags examined (days): 0, 1 , 2 ,3, 5-day mean
At least one discharge:
Lag 0: 1 .07 (0.62-1 .86); Lag 1 : 1 .06 (0.61-1 .85);
Lag 2 : 1 .05 (0 .62-1 .77) ; Lag 3 : 0 .09 (0 .65-1 .83) ;
Lag 5-day mean : 1.23 (0.71-2.12)
At least 10 discharges:
Lag 0: 1.12 (0.54-2. 32); Lag 1 : 1.13 (0.54-2.33);
Lag 2 : 1 .62 (0 .85-3 .09) ; Lag 3 : 1 .98 (1 .05-3 .72) ;
Lag 5-day mean : 1.94 (1 .01-75)
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0, 1 , 2, 3
No significant effect (results not reported in table).
Increment: 0.56 ppm; 0.54; 0.51 ; 0.49 respectively
for results shown below
OR Estimate [Lower Cl, Upper Cl]
Ventricular Arrythmia
Hours prior to event :
0-2 : 1.01 (0.87-1.18)
0-6:1.00(0.85-1.17)
0-23:1.03(0.84-1.25)
0-47:1.11 (0.88-1.40)
Increment: 0.2 ppm
OR for Ventricular Arrhythmia [Lower Cl, Upper
Cl]
Lags examined : 0-23 h-ma
0-23h-ma : 0.99 (0.80-1.21)
September 2009
C-6
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Rich et al. (2006,
088427)
Period of Study:
1995-1999
Location:
Boston, MA
Design
Health Outcome: ICD Episode of Atrial
fibrillation
Study Design: Panel/case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
203 patients with ICDs at the New England
Medical Center
Concentrations
Averaging Time: 1-h &
24-h
Mean (SD) unit: NR
Range (Min, Max):
1h:
25th = 0.46; 75th =
1.04
24-h:
25th = 0.52; 75th =
1.03
CO Effect Estimates (95% Cl)
Increment:
Lag (hrs) 0 : 0.58 ppm
Lag (hrs) 0-23 : 0.51 ppm
OR for Episode of Atrial Fibrillation [Lower Cl,
Upper Cl]
Lags (h) : 0, 0-23
Lag 0:0. 87 (0.56-1 .37)
Lag 0-23: 0.71 (0.39-1.28)
Copollutant: NR
Author: Sari etal. (2008,
190315)
Period of Study: June 2007
Location:
Gaziantep, Turkey
Health Outcome: P-wave dispersion
(predictors of atrial fibrillation, ventricular
arrhythmias and sudden deathjvia ECG
Study Design: Case-control
Statistical Analyses: Pearson correlation
analysis
Age Groups Analyzed:
Barbecue workers mean age: 33.66 ± 9.43
Control group mean age: 35.15 ± 6.78
Sample Description: 48 healthy males
working at various indoor barbecue
restaurants for at least 3 yr.(avg:15.6 ± 7.1
yr), 51 age-matched healthy men for control
group
Averaging Time: NR
Mean (SD) unit:
COHb%
Indoor barbecue
workers: 6.48% ±1.43
Control Group:
2.19% ±1.30
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Correlation Coefficient for COHb [p-value]
Lags examined : NR
Pmin:-0.132 (0.245)
Pmax: 0.215 (0.057)
Pd: 0.315 (0.005)
QTmin: 0.080 (0.454)
QTmax: 0.402 (<0.001)
QTd: 0.573 (<0.001)
cQTd: 0.615 (<0.001)
Author: Sarnat etal. (2006,
090489)
Period of Study:
24 wk during the Summer and
Fall of 2000
Location:
Steubenville, OH
Health Outcome: Arrhythmia via ECG
measurements
Study Design: Panel
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
53-90 yr (mean age 71)
Sample Description:
32 non-smoking older adults
Averaging Time: 24-h Increment: 0.2 ppm
RR Estimate [Lower Cl, Upper Cl]; lag :
Lags examined (days): 1,2,3, 4, 5, 5-day ma
Lag 5-day ma :
Mean (SD) unit:
0.02 ppm
Range (Min, Max): -
0.1,1.5
Copollutant:
correlation
PM25:r = 0.45
S02:r=0.62
N02:r = 0.66
03:r=-0.37
Supraventricular Ectopy
SVE: 0.99 (0.76-1.29)
Ventricular Ectopy
VE: 1.05 (0.75-1.46)
Author: Vedal et al. (2004,
055630)
Period of Study:
1997-2000
Location:
Vancouver, Canada
Health Outcome: Cardiac arrythmia via
patients with ICD
Study Design: Panel
Statistical Analyses:
Logistic regression (GEE)
Age Groups Analyzed: Range from
12-77 (mean age 53)
Sample Description:
50 patients who experienced 1 or more
arrhythmia event days during the four yr
Averaging Time: 24-h Increment: 0.2 ppm
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0,1, 2,3
No significant effect for CO (results shown in plots)
Mean (SD) unit:
0.6 ppm
Range (Min, Max):
0.3,1.6
Copollutant:
correlation PM10:
r = 0.43
S02:r=0.62
N02:r = 0.74
03:r=-0.52
September 2009
C-7
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CARDIAC ARREST
Author: Levy etal. (2001,
017171)
Period of Study:
1988-1994
Location:
Seattle, WA
Author: Sullivan et al. (2003,
043156)
Period of Study:
1985-1994
Location:
Washington State
Health Outcome: Out of hospital primary
cardiac arrest
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
25-75 yr
Sample Description:
362 cases
Health Outcome:
Out of Hospital Cardiac Arrest.
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
1 ,542 members of a large health
maintenance organization
Averaging Time: 24-h
Mean (SD) unit:
1.79ppm
Range (Min, Max):
0.52,5.92
Copollutant:
correlation PMi0:
r = 081
S02:r=0.29
Averaging Time: 24-h
Mean (SD) unit:
1.92 ppm
Range (Min, Max):
0.52,7.21
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0, 1
Lag 1 :0.99 (0.83-1.18)
Increment: 1 .02 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags examined (days) : 0, 1 , 2
Lag 0:0. 95 (0.85-1 .05)
Lag1 : 0.97 (0.87-1 .08)
Lag 2: 0.99 (0.89-1 .11)
MYOCARDIAL INFARCTION
Author: Peters et al. (2001 ,
016546)
Period of Study:
1995-1996
Location:
Boston, MA
Author: Rosenlund et al.
(2006, 089796)
Period of Study:
1992-1994
Location:
Stockholm, Sweden
Author: Rosenlund et al.
(2009, 190309)
Period of Study: NR
Location: Stockholm County,
Sweden
Health Outcome: Onset of Ml:
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
772 participants
Health Outcome: Ml
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
45-70 yr
Sample Description:
1,397 cases, 1,870 controls
Health Outcome: Fatal and nonfatal Ml
Study Design: Case-control
Statistical Analyses: Various multiple
regression models
Age Groups Analyzed: 15-79 yr
Sample Description: 43,275 Ml cases
during 1985-1996, 511,065 controls
Averaging Time: 24-h
Mean (SD) unit: 1.09
Range
(percentiles): ppm
5th = 0.49
95th = 1 .78
Copollutant: NR
Averaging Time:
Mean (SD) unit:
66.8 ug/m
(est. 30yr residential
exposure)
Range (percentiles):
5th = 13.9; 95th =
295.7
Copollutant: NR
Averaging Time: 1-yr
Mean (SD) unit:
Cases: 64.2 ug/m3
Controls: 55.8 ug/m3
Range (percentiles):
Cases: 5th = 7.3; 95th
= 267.4
Controls: 5th = 6.1;
95th = 261 .8
Copollutant: PM10,
N02
Increment: 2 H-1 ppm; 24-h - 0.6 ppm
OR Estimate [Lower Cl, Upper Cl]
Onset of Ml:
2-h prior: 1.22 (0.89-1 .67)
24-h prior: 0.98 (0.70-1 .36)
Increment: 300 ug/m3
OR Estimate [Lower Cl, Upper Cl] ; lag :
Estimated 30 yr avg. exposure
All cases: 1.04 (0.89-1 .21)
Non-fatal cases : 0.98 (0.82-1.16)
Fatal cases : 1.22 (0.98-1.52)
In-hospital death: 1.16 (0.89-1 .51)
Out-of-hospital death : 1 .36 (1 .01-1 .84)
Increment: NR
OR Estimate [Lower Cl, Upper Cl]
Syr. avg. exposure
All subjects (n = 301 ,273)
All cases :1.01 (0.97-1.05)
Non-fatal cases : 0.94 (0.89-1 .00)
Fatal cases : 1.1 4 (1.07-1 .21)
In-hospital death: 1.00 (0.91-1 .10)
Out-of-hospital death : 1 .23 (1 .14-1 .32)
Restriction to subjects who did not move between
population census ( n = 80,155)
All cases : 1.04 (0.94-1 .14)
Non-fatal cases : 0.96 (0.87-1 .06)
Fatal cases : 2.03 (1.59-2.60)
In-hospital death : 2.04 (1 .35-3.08)
Out-of-hospital death : 2.03 (1 .50-2.74)
September 2009
C-8
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CHANGES IN BLOOD PRESSURE
Author:
Ibalde-Mulli et al (2001
016030)
Period of Study:
1984-1985
Location:
Augsburg, Germany
Health Outcome: BP-SPB
Study Design: Cohort
Statistical Analyses:
Gaussian regression for repeated measures
Age Groups Analyzed:
25-64 yr
Averaging Time: 24-h
Mean (SD) unit:
4.1 mg/m3
Range (Min, Max):
1.7,8.2
Copollutant: NR
Increment: Lag 0 : 5.6 mg/m3
5-day prior avg.
Mean Change [Lower Cl, Upper Cl]
SPB mmHg
Lag 0 (days):
All : 0.53 (-0.66 to 1 .72): Men : 0.68 (-0.94 to 2.31):
Sample Description:
2,607 men & women aged 25-64 yr
Women :0.51 (-1.31 to 2.19)
5-day prior avg:
All: 1.06 (-0.17 to 2.29); Men : 0.92 (-0.87 to 2.70);
Women :0.91 (-0.87 to 2.70)
Author: Zanobetti et al. (2004, Health Outcome: BP
087489)
Study Design: Cohort/Panel
Period of Study:
1999-2001 Statistical Analyses: Random effects
Location:
Boston, MA
Age Groups Analyzed: 39-90 yr
Sample Description:
62 subjects with 631 total visits
Averaging Time:
1-h&120-havg
Mean (SD) unit:
Same Hr: 0.81 ppm
120 Hrav: 0.66 ppm
Range (Min, Max):
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
CO had no significant effect on BP
10th = 0.48; 90th =
1.22
120-hav:
10th = 0.48; 90th =
0.86
Copollutant: NR
CHANGES IN BLOOD MARKERS OF COAGULATION AND INFLAMMATION
Author: Baccarelli et al. (2007,
090733)
Period of Study:
1995-2005
Location:
Milan, Italy
Health Outcome: Prothrombin time (PT)
and Activated partial thromboplastin time
(APTT)
Study Design: Panel
Statistical Analyses: GAMS
Age Groups Analyzed:
11-84yr (mean 43years)
Sample Description:
1 ,218 healthy individuals who were partners
or friends of patients with thrombosis who
attended the thrombosis center of the
University of Milan.
Averaging Time: 1-h
Mean (SD) unit: NR
Range (percentiles):
Sept-Nov:
25th = 1.36; 75th =
3.52
Dec-Feb:
25th = 2. 00; 75th =
A -3-1
f.O 1
Mar- May:
25th = 1.03; 75th =
914
£.. I *T
h in Ai in*
JUI I r\uy .
25th = 0.73; 75th =
1.58
Increment: NR
Regression co-efficient [Lower Cl, Upper Cl]
Lags examined (time of blood sampling - avg): 0,
7,30
PT:
Lag 0 : -0.11 (-0.18 to -0.05); Lag 7 : -0.07 (-0.14to
0.01); Lag 30: -0.05 (-0.13 to 0.02)
APTT:
Lag 0 : 0.03 (-0.04 to 0.10); Lag 7 : 0.04 (-0.04 to
0.11); Lag 30: 0.06 (-0.01 to 0.14)
Notes: CO had no effect on fibrinogen, functional
antithrombin, functional protein C, protein C antigen,
functional protein S, free protein S for all lag periods.
Copollutant: NR
September 2009
C-9
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Delfinoetal. (2008,
156390)
Period of Study:
2005-2006
Location: Los Angeles, CA
Author: Liao et al. (2005,
088677)
Period of Study:
1996-1998
Location:
Forsyth County, NC; Selected
suburbs of Minneapolis, MN;
Jackson, Ml
Author: Ljungman et al.
(2009, 191983)
Period of Study:
May 2003-July 2004
Location: Athens, Greece;
Augsberg, Germany;
Barcelona, Spain; Helsinki,
Finland; Rome, Italy;
Stockholm, Sweden
Design
Health Outcome: Biomarkers of systemic
inflammation
Study Design: Panel
Statistical Analyses:
Linear mixed-effects models
Age Groups Analyzed:
>65yr(mean85.7yr)
Sample Description: 29 nonsmoking
subjects with history of CAD living in
retirement communities
Health Outcome: Various measures of
hemostasis/ inflammation
Study Design: Cohort
Statistical Analyses:
Linear regression
Age Groups Analyzed:
45-64 yr
Sample Description:
10,208 subjects from the Atherosclerosis
Risk in Communities Study
Health Outcome: Plasma lnterleukin-6 (IL-
6), Fibrinogen
Study Design: Panel/Field
Statistical Analyses: Linear Mixed Effects
Model
Age Groups Analyzed: 35-80 yr (mean =
62.2 yr)
Sample Description: 955 subjects who had
experienced myocardial infarction between 4
mo and 6 yr before start of the study
Concentrations
Averaging Time: 24-h
Mean (SD) unit:
0.78 ± 0.30 ppb
Range (Min, Max):
0.22, 1.97
Copollutant
(Outdoor):
EC: r = 0.84
OC:r = 0.69
OCprimary:r= 0.73
N02:r = 0.78
03:r=-0.35
PM0.25:r=0.84
PM0.25-2.5:r = 0.14
PM2.5-io:r = 0.51
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Averaging Time: 24-h
Mean (SD) unit:
Individual cities: 0.29-
1 .48 mg/m3
Mean for all cities:
0.78 mg/m
Range (percentiles):
25th = 0.56; 75th =
0.90 (for mean of all
cities)
Copollutant:
(mean for all cities)
N02:r= 0.69
PM10:r= 0.47
PM2.5: r = 0.55
PNC: r= 0.67
CO Effect Estimates (95% Cl)
Increment: NR
Estimated co-efficient
Relationship to Outdoor Air Pollutants:
CRP (ng/mL): Lag 0: 847.52; 3-day avg: 728.79; 9-
dayavg: 236.51
IL-6 (pg/mL): Lag 0: 0.52; 3-day avg: 0.51 ; 9-day
avg: 0.50 sTNF-RII (pg/mL): Lag 0: 154.05; 3-day
avg : 1 39 .45 ; 9-day avg : 225 .60
Relationship to Indoor Air Pollutants:
CRP (ng/mL): Lag 0: 695.39; 3-day avg: 527.37; 9-
dayavg:760.15
IL-6 (pg/mL): Lag 0: 0.54; 3-day avg: 0.47; 9-day
avg: 0.77 sTNF-RII (pg/mL): Lag 0: 114.22; 3-day
avg : 1 07 .95 ; 9-day avg : 273 .38
Relationship of sP-selction (ng/mL) to:
Indoor Air Pollutants: Lag 0: 0.77; 5-day avg: 1 .40; 9-
day avg: 2. 19
Outdoor Air Pollutants: Lag 0: 0.84; 5-day avg: 1.23;
9-day avg: 4.29
Relationship of Cu, Zn-SOD (U/g Hb) to:
Indoor Air Pollutants: Lag 0: -145.54; 5-day avg: -
238.72; 9-day avg: -70.10
Outdoor Air Pollutants: Lag 0: -105.73; 5-day avg: -
176. 72; 9-day avg: -41 .92
Increment: 0.6 ppm
Regression coefficients [SE]
Lags examined (days): 1
Lag1:
Fibrinogen (mg/dL) :-0.16 (0.67)
Factor VIII -C (%) : 0.45 (0.42)
vWF%: -0.29 (0.50)
WBC(x103/mm3): 0.003 (0.017)
Albumin (g/dL) : -0.01 8 (0.003)**
** p < 0.01
Increment: 0.34 mg/m3
Change of IL-6
% of overall mean per IQ range increase
Genotypes:! 1,1 2,22
IL6 rs2069832
1 1:2.0(0.3,3.6);! 2: -0.2 (-1.7, 1.3); 2 2: -2.0 (-4.7,
0.8); p-value: 0.03
IL6 rs2069840
1 1:2.0(0.3,3.8);! 2: 0.4 (-0.9, 1.7); 2 2: -1.2 (-3.4,
1.1); p-value: 0.04
IL6 rs2069845
1 1:1.9(0.2,3.5);! 2: -0.1 (-1.5, 1.4); 2 2: -1.6 (-4.3,
1.2); p-value: 0.31
FGArs2070011
1 1:1.0 (-0.7,2.7);! 2:0.7 (0.6, 2.0); 2 2:0.4 (-1.9,
2.7); p-value: 0.64
FGBrs! 800790
1 1: -0.2 (-1.8, 1.3); 1 2:2.1 (0.4, 3.8); 2 2:4.5 (1.1,
8.0); p-value: 0.02
September 2009
C-10
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Pekkanen et al.
(2000, 013250)
Period of Study:
1991-1993
Location:
London, England
Author: Ruckerl et al. (2006,
088754)
Period of Study:
2000-2001
Location:
Erfert, Germany
Design
Health Outcome: Fibrinogen
Study Design: Cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
35-55 yr
Sample Description:
7,205 office workers
Health Outcome: Blood markers of
inflammation and coagulation
Study Design: Panel
Statistical Analyses:
Linear and logistic regression (fixed effects)
Age Groups Analyzed:
51 -76 yr (mean age 66 yr)
Sample Description:
57 male patients with CHD
Concentrations
Averaging Time: 8 h
Mean (SD) unit:
1 .4 mg/m3
Range (Min, Max):
Min= NR, Max = 9.9
Copollutant
correlation:
PM10:r = 0.57
N02:r = 0.81
on • r - n fii
OW2- 1 ~ U.D 1
03:r=-0.45
Averaging Time: 24-h
Mean (SD) unit:
0.52 mg/m
Range (Min, Max):
0.11,1.93
Copollutant
correlation: N02:
r = 0.82
CO Effect Estimates (95% Cl)
Increment: 1 .6 mg/m3
% Change in Fibrinogen Concentration [p value] ;
Lags examined : 0,1,2,3
Lag 0 : 1 .43 (<0.01); Lag 1 : 1 .49 (<0.01);
Lag 2 : 1 .59 (<0.01); Lag 3 : 1 .26 (<0.01)
OR for having Fibrinogen above 3.1 9 g/l [p value]
Lags examined : 0,1,2,3
Lag 0 : 1.1 7 (0.05); Lag 1 : 1.09 (0.31);
Lag 2: 1.14 (0.11); Lag 3:1. 22 (<0.01)
Increment: 0.27 mg/m3
OR Estimate for blood marker >90th percentile
[Lower Cl, Upper Cl]
Lags examined (h): 0-23, 24-47, 48-71 ,
5-day avg.
CRP (C-reactive protein)
0-23: 0.9 (0.7-1 .2); 24-47: 1.0 (0.7-1 .5);
48-71 : 1.5 (1.1-2.1); 5-day avg 1.1 (0.8-1.6)
ICAM-1 (Intercellular adhesion molecule 1)
0-23: 0.8 (0.6-1.0); 24-47: 1.5 (1.2-1 .9);
48-71 : 1 .7 (1 .3-2.3) ; 5-day avg 1 .2 (1 .0-1 .6)
% of change from the mean of blood marker
vWF (von Willebrand factor antigen)
0-23 :4.4 (1.4- 7.5); 24-47 :2.7 (-0.8 to 6.1);
48-71 :2.0 (-1.7 to 5.8); 5-day avg : 4.9 (1.0-8.8)
FVII (Factor VI I)
0-23 : -1.4 (-3.8to 1.1);24-47 :-2.6 (-4.8to 0.3);
48-71 : -2.8 (-5.1 to -0.4); 5-day avg : -3.0 (-5.5 to -
0.4)
Author: Ruckerl et al. (2007,
156931)
Period of Study:
May 2003-July 2004
Location:
6 cities across Europe:
Athens, Greece;
Augsburg, Germany;
Barcelona, Spain;
Helsinki, Finland;
Rome, Italy;
Stockholm, Sweden
Health Outcome: lnterleukin-6,
C-reactive protein, Fibrinogen
Study Design: Panel/Cohort
Statistical Analyses:
Linear regression (mixed effects)
Age Groups Analyzed:
37-81 yr
Sample Description:
1,003 Ml survivors who had at least 2 valid
repeated blood samples
Averaging Time: 24-h
Mean (SD) unit:
Athens: 1.48 mg/m3
Augsburg: 0.58 mg/m
Barcelona: 0.59 mg/m3
Helsinki: 0.31 mg/m3
Rome: 1.40 mg/m
Stockholm: 0.29 mg/m3
Range (Min, Max): NR
Copollutant: NR
Increment: 0.34 mg/m
% Change in mean [Lower Cl, Upper Cl]
Lags examined: 0,1,2, 5-day avg
(Pooled estimates)
lnterleukin-6
Lag 0:0.57 (-0.63 to 1.79); Lag 1 :0.44 (-0.79 to
1.68);
Lag 2: -2.36 (-4.82 to 0.17); 5-day avg: -0.28 (-2.53
to 2.02)
C-reactive protein
Lag 0:-0.01 (-1.72 to 1.73); Lag 1 :-1.51 (-3.30 to
0.32); Lag 2:-2.35 (-6.84to 2.36);
5-day avg :-0.85 (-.5.37 to 3.90)
Fibrinogen
Lag 0 :0.24 (-0.54 to 0.92); Lag 1 :0.32 (-0.35 to
1.00);
Lag 2:-0.44 (-1.11 to 0.23); 5-day avg : 0.12 (-0.81
to 1.05)
September 2009
C-11
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Rudez et al. (2009,
193783)
Period of Study:
January 2005-December 2006
Location: Rotterdam, the
Netherlands
Health Outcome: Platelet aggregation,
thrombin generation, Fibrinogen, C-reactive
protein
Study Design: Panel
Statistical Analyses:
Linear regression
Age Groups Analyzed:
mean age 41 yr
Sample Description:
40 healthy individuals
Averaging Time: 24-h Increment: NR
Median (SD) unit: 333
ug/m3
Range (percentiles):
25th = 276; 75th = 412
Copollutant:
PM10:r >0.6
N0:r>0.6
N02:r>0.6
03:-0.4>r>-0.6
Estimated Changes [Lower Cl, Upper Cl]
Platelet Aggregation Parameters
Maximal Platelat Aggregation:
DO-6: -3.6 (-9.3, 2.1); DO-12: -4.7 (-11.0,1.5); DO-24:
-2.6 (-7.9, 2.7); 124-48: -1.1 (-7.2,4.9); 148-72:8.4
(2.5,14.3); !72-96:-0.1 (-5.1, 5.0); 0+10-96:9.5 (1.6,
17.4)
Late Aggregation:
DO-6:10.5 (0.8,20.3); DO-12:11.6 (1.2, 21.9); DO-
24: 11.2 (1.4,21.0); 124-48:7.5 (-2.2,17.1); 148-72:
18.1 (8.4, 27.8); 172-96:4.2 (-5.5,13.9); D+IO-96:
20.4 (8.4,32.4)
Thrombin Generation
ETP
DO-6:-1.51 (-3.7, 0.80); DO-12:-1.1 (-3.4,1.1); DO-
24: -1.5 (-3.9,0.9); 124-48: -0.7 (-3.4,2.0); 148-72:
0.8 (-1.9, 3.4); 172-96:3.5 (0.8, 6.2); D+IO-96:0.8 (-
2.7, 4.3)
Peak
DO-6: -2.5 (-6.3,1.3) DO-12: -1.9, (-5.7,1.9); DO-24: -
3.3 (-7.3, 0.7); !24-48;-1.3 (-6.1, 3.6); !48-72:-0.5 (-
5.0, 4.0) 172-96:3.8 (-0.8,8.4) D+IO-96: -1.7 (-7.5,
4.2)
Lag Time
DO-6:1.0 (-0.5, 2.5); DO-12:1.0 (-0.5,2.5); DO-24:
1.6 (0.1,3.1); 124-48; 0.4 (-1.3, 2.2); 148-72: -1.0 (-
2.7, 0.7); 172-96: -1.5 (-3.2, 0.2); D+IO-96:0.1 (-2.1,
2.2)
Inflammatory Markers
Fibrinogen
124-48; 0.0 (-1.7,1.8); 148-72:0.0 (-1.8,1.9) 172-96:
-0.1 (-1.9,1.7)
CRP
124-48; 3.2 (-6.4,12.8); 148-72: -1.9 (-12.5, 8.7); I72-
96:-4.5 (-15.3, 6.3)
Author: Steinvil et al. (2008,
188893)
Period of Study:
2003-2006
Location: Tel Aviv, Israel
Health Outcome: Various measures of
inflammation sensitive biomarkers
Study Design: Cohort
Statistical Analyses:
Linear regression
Age Groups Analyzed:
mean age 46 yr
Sample Description:
3,659 subjects living within 11 km of
monitoring site
Averaging Time: 24-h Increment: 0.3 ppm
Mean (SD) unit:
0.8 ppm
Range (percentiles):
25th = 0.7; 75th = 1.0
Copollutant:
correlation
PM10:r = 0.75
N02:r = 0.857
S02:r= 0.671
03: r = -0.656
Regression co-efficient [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2,3,4,5,6,7, last wk
avg
Fibrinogen-Men
Lag 0 :-3.3 (-6.1 to-0.6); Lag 1 :-2.6 (-5.5 to 0.4);
Lag 2 : -3.4 (-6.6 to -0.3); Lag 3 : -3.4 (-6.5 to -0.2);
Lag 4 : -5.9 (-8.9 to -2.9); Lag 5 : -4.7 (-7.8 to -1.6);
Lag 6 : -2.0 (-5.1 to 1.0); Lag 7 : -2.7 (-5.7 to 0.2);
Last wk avg:-7.7 (-12.1 to-3.3)
Notes: No effect on fibrinogen among women. CO
had no effect on CRP among men and no effect on
CRP and WBC among women for all Lag times
examined.
September 2009
C-12
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
VARIOUS MEASURES OF CARDIOVASCULAR HEALTH
Author: Brietetal. (2007,
093049)
Period of Study: NR
Location: Paris, France
Health Outcome: Endothelial function,
Reactive Hyperemia
Study Design: Case-crossover
Statistical Analyses: Multiple regression
models
Age Groups Analyzed: 18-35 yr
Sample Description: 40 healthy white male
nonsmokers
Averaging Time: 24-h Increment: NR
Mean (SD) unit: NR p-Coefficient [Lower Cl, Upper Cl]
Range (Min, Max): NR Flow-mediated Brachial Artery Dilatation:
Copollutant: -0.68 (-1.22,-0.15)
Small Artery Reactive Hyperemia:
10.46(1.73,19.31)
PM25, PM10, NO, N02,
S02
Author: Nautiyal et al. (2007,
190301)
Period of Study:
August 1999-May 2000
Location:
Mandi Gobindgarh, India
Morinda, India
Health Outcome: Various measures of
cardiovascular health via ECG (Minnesota
Code)
Study Design: Cross-sectional
Statistical Analyses: NR
Age Groups Analyzed: +15yr
Sample Description:
200 total survey participants (100/town)
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max):
Morinda
Pure residential Site: 0-
1 ppm GT Road
Site: 2-3 ppm
Mandi Gobindgarh
Mixed Habitat Site: 0-3
ppm GT Road Site:
1-3 ppm
Copollutant:
PM2.5, PM10, NOx, SOx
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined : NR
No quantitative results presented
Author: Wellenius et al. (2007, Health Outcome: Congestive heart failure
092830)
Study Design: Cohort (retrospective)
Period of Study: February
2002-March 2003
Location: Boston, MA
Statistical Analyses: Linear mixed models
Age Groups Analyzed: 33-88 yr.
Tai Chi Group mean age (n=14): 66 ±
13yr.
Control Group mean age (n=14): 63 ±
14yr.
Sample Description: 28 patients with CHF
and impaired systolic function
Averaging Time: 24-h
Mean (SD) unit: 0.44
ppm
Range (IQ): 0.20 ppm
Copollutant:
PM2.5: r = 0.35
N02, S02,03, BC
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined : 0,1,2,3
Results presented graphically
Table C-2 Studies of CO exposure and cardiovascular hospital admissions and ED visits.
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
STROKE
Author: Chan et al. (2006,
090193)
Period of Study: 1997-2002
Location:
Taipei, Taiwan
ED Visits
Health Outcome (ICD9):
Cerebrovascular disease
(430-437); Strokes (430-434);
Hemorrhagic stroke (430-432); Ischemic
stroke (433-434)
Study Design: Time-series
Statistical Analyses: GAM
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 8 h
Mean (SD) unit:
1.7 ppm
Range (Min, Max): 0.6, 4.4
Copollutant: correlation
03:r = 0.30
S02:r - 0.63
N02: r- 0.77
PM2.5: r = 0.44
PMi0:r- 0.47
Increment: 0.8 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags (days) examined 0,1,2,3
Cerebrovascular disease: Lag 2, 1 .03 (1 .01 ,
1 06)
Stroke : Lag 2, 1.03 (1.01, 1.05)
Ischemic and Hemorrhagic stroke : not
significant.
Cerebrovascular 2 pollutant model:
CO + 03: Lag 2, 1.03(1.01-1.05)
CO + PM25: Lag 2, 1.02 (1.00-1 .04)
CO + PM10: Lag 2, 1.03(1.01-1.05)
September 2009
C-13
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Henrotin et al. (2007,
093270)
Period of Study: 1994-2004
Location:
Dijon, France
Health Outcome (ICD9 or ICD10):
Stroke (Ischemic & Hemorrhagic)
Study Design: Bi-directional Case-
crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: > 40 yr
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 683 |jg/m3
Range (Min, Max): 0,4014
Copollutant: NR
Increment: 10 |jg/m3
OR Estimate [Lower Cl, Upper Cl]
Lags (days) examined: 0,1,2, 3.
Ischemic:
Lag 0:0.999 (0.997-1.001)
Lag1 :0.998 (0.997-1.001)
Lag 2:0.999 (0.998-1.001)
Lag 3:1.000 (0.998-1.001)
Hemorrhagic:
Lag 0:1.000 (0.996-1.004)
Lag1 :1.001 (0.997-1.005)
Lag 2 :0.999 (0.995-1.004)
Lag 3:0.998 (0.994-1.002)
Also not significant when stratified by sex.
Author: Maheswaran et al.
(2005, 090769)
Period of Study: 1994-1998
Location:
Sheffield, UK
Health Outcome (ICD9 or ICD10):
Stroke deaths (ICD9:430-438); Stroke
Hospital admissions (ICD10:160-I69)
Study Design: Ecological
Statistical Analyses:
Poisson regression
Age Groups Analyzed: > 45 yr
Sample Description:
1,030 census districts
Averaging Time: NR
Mean (SD) unit: Quintiles
Range (Min, Max): NR
Copollutant: NR
Increment:
NR-Quintiles of exposure
RR Estimate [Lower Cl, Upper Cl]
Adjusted for sex, age, deprevation,
smoking.
Quintiles:
2nd: 1.04 (0.94-1.16)
3rd: 1.01 (0.91-1.13)
4th : 1.10(0.99-1.23)
5th: 1.11 (0.99-1.25)
Adjusted for sex, age:
2nd: 1.11 (1.01-1.22)
3rd: 1.15 (1.04-1.27)
4th: 1.29 (1.17-1.42)
5th : 1.37 (1.24-1.52)
Author: Tsai et al. (2003,
080133)
Period of Study:
1997-2000
Location:
Kaohsiung, Taiwan
Study Design: Case-crossover
Health Outcome (ICD9 or ICD10):
Cerebrovascular diseases: ICD9: 430 to
438 (Subarachnoid hemorrhagic stroke
430, Primary intracerebral hemorrhage
(PIH): 431-432, Ischemic stroke (IS):
433-435).
Statistical Analyses: NR
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 0.79 ppm
Range (Min, Max): 0.24, 1.72
Copollutant: NR
Increment: 0.8 ppm (IQR)
RR Estimate [Lower Cl, Upper Cl]
Lag (days): 0-2
>20°C
PIH : OR 1.21 (1.09-1.34)
IS : OR 1.21 (1.14-1.28)
<20°C
PIH : OR 1.1 8 (0.80-0.72)
IS : OR 1.77 (1.31-2.39)
Notes:
2-pollutant models:
PIH results nersisterl when arliustinn for
S02 and 03
IS results persisted when controlling for
PM10, S02 and 03
Author: Villeneuve et al. (2006, ED Visits (within 5 hospitals)
Period of Study: 1992-2002
Location:
Edmonton, Canada
Health Outcome (ICD9): Stroke (430-
438); Ischemic (434-436) Hemorrhagic
(430-432); Transient Ischemic Attack
(435)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: 65+ yr
Sample Description:
12,422 visits
Averaging Time: 24-h
Mean (SD) unit: 0.8 ppm
Range (percentiles):
25th = 0.5; 75th = 1.0
Copollutant correlation :
03:r=-0.54
PM25:r = 0.43
PM10:r = 0.30
Increment: 0.5 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags (days) examined : 0,1 & 0-2
Ischemic (April-Sept)
LagO : 1.16 (1.00,1.33)
Lag1 :1.17 (1.01,1.36)
Lag 0-2 :1.32 (1.09,1.60)
Notes:
- Not significant for all seasons or Oct-Mar.
- Hemorrhagic : Not significant for all
seasons or Oct-Mar, Apr-Sept.
- Transient Ischemic Attack : Not significant
for all seasons or Oct-Mar, Apr-Sept.
September 2009
C-14
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Wellenius et al. (2005, ED Visits
088685)
Period of Study: NR
Location:
9 U.S. cities: Chicago, Detroit,
Pittsburgh, Cleveland,
Birmingham, New Haven,
Seattle, Minneapolis, Salt Lake
City
Health Outcome:
Stroke among Medicare beneficiaries:
(Ischemic, hemorrhagic)
Study Design: Time-series
Statistical Analyses:
Logistic regression
Age Groups Analyzed: > 65 yr
Sample Description: 155,503 visits
Averaging Time: NR
Mean (SD) unit: NR
Range (percentiles):
25th = 0.73; 50th = 1.02;
75th = 1.44 (ppm)
Copollutant: correlation
PM10:r = 0.43
Increment: 0.71 ppm
% Change [Lower Cl, Upper Cl]
Lag:0
Ischemic: 2.83 (1.23-4.46)
Hemorrhagic : -1.61 (-4.79 to 1.68)
ISCHEMIC HEART DISEASE
Author: D'lppoliti et al. (2003,
074311)
Period of Study: 1995-1997
Location:
Rome, Italy
Hospital Admissions
Health Outcome (ICD9): Ml (410)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: 18+yr
Sample Description:
6,531 patients.
Averaging Time: 24-h
Mean (SD) unit: 4.4 mg/m3
Range (percentiles):
25th = 2.8; 75th = 4.3
Copollutant: correlation
TSP: r = 0.35
S02:r=0.56
N02:r = 0.31
Increment: 1 mg/m3
OR Estimate [Lower Cl, Upper Cl]; lag :
Lags examined (days): 0,1,2, 3,4, 0-2
Acute Ml
Lag 0:1.021 (0.988-1.054)
Lag1 :1.020 (0.988-1.054)
Lag 2:1.033 (1.001-1.066)
Lag 3:1.010(0.982-1.040)
Lag 4:1.025 (0.996-1.055)
Lag 0-2:1.044 (1.000-.089
Author: Hosseinpoor et al.
(2005, 087413)
Period of Study:
1996-2001
Location:
Tehran, Iran
Health Outcome:
Angina Pectoris (ICD9:413; ICD10:120)
Study Design: Time-series
Statistical Analyses:
Poisson regression
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 10.8 mg/m3
Range (Min, Max): 1.6, 57.8
Copollutant: NR
Increment: 1 mg/m3
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2, 3
Lag1 :1.00957 (1.00600-1.01315)
September 2009
C-15
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Lanki et al. (2006,
089788)
Period of Study: 1994-2000
Location:
5 European cities:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Author: Lee etal. (2003,
095552)
Period of Study: 1997-1 999
Location:
Seoul, Korea
Author: Maheswaran et al.
(2005, 090769)
Period of Study:
1994-1998
Location:
Sheffield, UK
Design
Health Outcome:
First AMI (ICD9: 410; ICD10: 121,122)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed:
35+ yr
Sample Description:
26,854 Hospital Admissions
Study Design: Time-series
Health Outcome (ICD9 or ICD10):
Angina: ICD1 0:120
AMI: ICD10: 121-123
Other Acute IHDs:ICD1 0:124
Statistical Analyses:
Poisson regression, GAM
Age Groups Analyzed: 64+ yr
Sample Description: 822 days
Emergency Hospital Admission
Health Outcome (ICD9):
CH 0(410-41 4)
Study Design: Ecological
Statistical Analyses: Poisson
regression
Age Groups Analyzed: 45+ yr
Sample Description:
11, 407 Emergency Hospital Admissions
for CHD in patients 45+ yr (within 1 ,030
census districts)
Concentrations
Averaging Time: 24-h
Mean (SD) unit: NR
Unit: mg/m3
Range (percentiles):
Augsburg, Germany
25th = 0.7; 75th = 1.1
Barcelona, Spain
25th = 0.6; 75th = 1.4
Helsinki, Finland
25th = 0.3; 75th = 0.5
Rome Italy
25th = 1.7; 75th = 2. 9
Stockholm, Sweden
25th = 0.3; 75th = 0.5
Copollutant: correlation
PM10:r = 0.21 -0.56
N02:r = 0.43 -0.75
03 : r = -.023 - 020
Averaging Time:
Daily max
Mean (SD) unit:
1 .8 ppm
Range (percentiles):
25th = 1 .2
75th = 2.2
Copollutant: correlation
PM20:0.60
S02:0.81
N02: 0.79
03:-0.39
Averaging Time: NR
Mean (SD) unit: Quintiles
Range (Min, Max): NR
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: 0.2 mg/m3
RR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined : 0,1,2,3
All 5 cities:
LagO : 1.005 (1.000-1 .010)
Lag1 : 1.002 (0.996-1 .007)
Lag 2 : 1.002 (0.997-1 .007)
Lag3 : 0.998 (0.992-1 .003)
3 cities with Hospital Discharge
Register(HDR):
Lag 0:1. 007 (1.001 -1.01 2)
Lag1 : 1.002 (0.996-1 .008)
Lag 2: 1.003 (0.998-1 .009)
Lag 3: 1.004 (0.988-1 .020)
3 cities with HDR-<75years
Fatal:
LagO : 1.027 (1.006-1 .048)
Lag1 : 1.021 (1.000-1 .042)
Lag 2 : 1.01 8 (0.997-1 .039)
Lag3 : 1.01 5 (0.994-1 .037)
Non-Fatal:
Lag 0:1. 001 (0.995-1 .008)
Lag1 : 1.000 (0.994-1 .007)
Lag 2 : 1.004 (0.998-1 .011)
Lag 3: 0.999 (0.992-1 .006)
3 cities with HDR - > 75years
Fatal:
LagO : 1.009 (0.992-1 .006)
Lag1 : 1.001 (0.985-1.018)
Lag 2: 1.006 (0.990-1 .023)
Lag3 : 1.000 (0.983-1 .017)
Non-Fatal:
LagO : 1.01 5 (1.004-1 .086)
Lag1 : 1.006 (0.995-1 .017)
Lag 2: 0.995 (0.983-1 .006)
LagS : 0.998 (0.987-1 .009)
Increment: 1 ppm (IQR)
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days) : 0, 1 , 2, 3, 4, 5, 6
Allyr:
Lag 5: All ages: 0.94 (0.91 0.98)
Lag 5: 64+ age : 1.07 (1.01-1.13)
Summer:
Lag 5: All ages : 1.1 9 (1.02-1.38)
Lag 5: 64+ age : 1.60 (1.27-2.03)
2-pollutant model:
Lag 5 : 64+ age :
CO + PM10: 1.04 (0.98-1 .11)
Increment: NA
RR Estimate [Lower Cl, Upper Cl]
Lowest quintile reference category
Adjusted for sex, age, deprivation, smoking:
2nd: 0.97 (0.89-1 .07)
3rd: 0.94 (0.86-1 .04)
4th : 0.96 (0.97-1 .06)
5th : 0.88 (0.79- 0.98)
Adjusted for sex, age:
2nd : 1.09 (1.00-1 .19)
3rd: 1.1 5 (1.05-1 .26)
4th : 1.1 9 (1.09-1 .30)
5th : 1 .20 (1 .09-1 .32)
September 2009
C-16
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Mannetal. (2002,
036723)
Period of Study: 1988-1995
Location:
Southern California
Health Outcome (ICD9): IHD (IHD)
(410-414); Myocardial Infarction (Ml)
(410)
Study Design: Time-series
Statistical Analyses:
Poisson regression, GAM
Age Groups Analyzed: All
Sample Description:
54,863 IHD admissions among Southern
California Kaiser- Permanente members
(within 20km of monitor)
Averaging Time: 8 h
Mean (SD) unit: 2.07 ppm
Range (Min, Max): 0.30,11.;
Copollutant: correlation
Ranging across 7 regions:
N02:r = 0.64,0.86
03:r =-0.37,0.28
PM10:r = 0.15, 0.40
Increment: 1 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2, 2 ma,
3 ma, 4 ma
With arrythmia:
Lag 0 :2.99 (1.80-4.99)
Lag1 :1.51 (0.37-2.66)
Lag 2:1.26 (0.15-2.38)
2 ma: 2.66 (1.40-3.94)
3 ma: 2.59 (1.27-3.92)
4ma : 2.25 (0.90-3.63)
With CHF:
Lag 0 :3.60 (1.620-5.63)
Lag1 :3.34 (1.48-5.22)
Lag 2:1.90 (0.11-3.72)
2 ma: 4.23 (2.13-6.37)
3 ma: 4.14 (1.96-6.37)
4 ma : 4.07 (1.81 -6.38)
Without secondary diagnosis:
Lag 0:1.62 (0.65-2.59)
Lag1 :1.45 (0.54-2.37)
Lag 2:0.92 (0.04-1.82)
2 ma: 1.83 (0.80-2.86)
3 ma: 1.79 (0.72-2.87)
4 ma: 1.82 (0.71-2.94)
Author: Szyszkowicz (2007,
193793)
Period of Study: 1997-2003
Location: Montreal, Canada
Study Design: Time-series
Health Outcome (ICD9 or ICD10):
ED Visits.
IHD:ICD9:410-414
Statistical Analyses:
Poisson regression (GLMM)
Age Groups Analyzed: All
Sample Description: 4,979 ED Visits
Averaging Time: 24-h
Mean (SD) unit: 0.5 ppm
Range (Min, Max): 0.1, 3.1
Copollutant: NR
Increment: 0.2 ppm
% Change [Lower Cl, Upper Cl]; lag :
Lags examined (days): 0,1
All Patients: Lag 0:5.4 (2.3-8.5)
Males: Lag 0:7.5 (3.6-11.6)
Females: Lag 0:2.7 (-2.0 to 7.6)
Ages > 64
All Patients: Lag 0:4.9 (1.3-8.7)
Males: Lag 0:7.5 (2.6-12.6)
Females: Lag 0 :2.4 (-3.0 to.O)
Lag 1 not significant for all results
Author: von Klot et al. (2005,
088070)
Period of Study: 1992-2001
Location:
5 European cities:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Health Outcome: Hospital Cardiac
(Myocardial Infacrtion (Ml), Angina,
Dysrythmia, Heart Failure) Re-
admissions
Study Design:
Prospective Cohort
Statistical Analyses: Poisson
regression
Age Groups Analyzed: All
Sample Description:22 006 survivors
of first Ml
Averaging Time: 24-h
Unit: mg/m3
Mean (SD) unit:
Augsburg, Germany: 0.93
Barcelona, Spain: 1.00
Helsinki, Finland: 0.42
Rome, Italy: 2.21
Stockholm, Sweden: 0.43
Range (Min, Max): NR
Copollutant: correlation
PM10:r = 0.21-0.57
N02:r = 0.44-0.75
03: r = -.027 - 0.47
Increment: 0.2 mg/m (0.172 ppm)
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2, 3
LagO:
Ml:1.022(0.998-.047)
Angina :1.009 (0.992-.02)
Cardiac: 1.014 (1.001-.026)
September 2009
C-17
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
HEART FAILURE
Author: Lee etal. (2007,
093271)
Period of Study: 1996-2004
Location:
Kaohsiung City, Taiwan
Hospital Admissions
Health Outcome (ICD9): CHF (428)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: All
Sample Description:
13,475 Hospital Admissions
(63 Hospitals)
Averaging Time: 24-h
Mean (SD) unit: 0.76 ppm
Range (Min, Max): 0.14,1.72
Copollutant: NR
Increment: 0.31 ppm
OR Estimate [Lower Cl, Upper Cl]
Lag examined (days): 0-2
a 25°C: 1.19 (1.09-1.31)
<25°C: 1.39 (1.24-1.54)
Adjusted for PM10:
a 25°C; 1.15 (1.04-1.27)
<25°C;1.21 (1.206-1.38)
Adjusted for S02:
>25°C:1.23 (1.11-1.36)
<25°C: 1.39 (1.24-1.55)
Adjusted for N02:
a 25°C: 1.22 (1.08-1.39)
<25°C: 0.94 (0.81-1.10)
Adjusted for 03:
a 25°C: 1.17 (1.07-1.28)
<25°C: 1.36 (1.22-1.51)
Author: Symons et al. (2006,
091258)
Period of Study: 2002
(April-November)
Location:
Hospital Admissions
Health Outcome: NR
Study Design: Case-crossover
Statistical Analyses:
Averaging Time: 24-h
Mean (SD) unit: 0.4 ppm
Range (Min, Max): 0.1, 1.0
Copollutant: NR
Increment: 0.2 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags examined (days):
0, 1,2,3, cum 1, cum 2, cum 3
Johns Hopkins Bayview
Medical Center, Baltimore, MD
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
398 Hospital Admissions for CHF
Lag 0:0.86 (0.67-1.11)
Lag1 :0.90 (0.70-1.17)
Lag 2:0.96 (0.73-1.26)
Lag 3:0.88 (0.67-1.16)
Cum. Lag1 :0.82 (0.60-1.13)
CumLag3:o!27(o!46-1/l4)
Author: Wellenius et al. (2005,
087483)
Period of Study: 1987-1 999
Location:
Pittsburgh, PA
Hospital Ad missions
Health Outcome (ICD9): CHF
(428,428.1)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: 65+ yr
Sample Description:
54,019 Hospital Admissions among
Medicare beneficiaries
Averaging Time: 24-h
Mean(SD) unit: 1.03 ppm
Range (percentiles):
25th = 0.68; 75th = 1.23
Copollutant: correlation
PM10:r = 0.57
N02: r = 0.70
03:r=-0.25
S02:r= 0.54
Increment: 0.55 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0, 1 , 2, 3
LagO:
Single pollutant model: 4.55 (3.33-5.79)
Adjusted for PM10 : 5.18 (3.49-6.89)
Adjusted for N02: 4.84 (3.06-6.66)
Adjusted for 03 : 4.35 (3.08-5.64)
Adjusted for S02 : 4.51 (3.15-5.90)
Author: Yang (2008,157160)
Period of Study: 1996-2004
Location: Taipei, Taiwan
Hospital Admissions
Health Outcome: CHF
Study Design: Case-crossover
Statistical Analyses: NR
Age Groups Analyzed: NR
Sample Description: 24,240 CHF HA
from 47 hospitals
Averaging Time: 24-h
Mean (SD) unit: 1.26 ppm
Range (Min, Max): 0.12, 3.66
Copollutant: PM10, N02,03, S02
Increment: NR
OR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2
Single Pollutant Model
Warm days (>20o C): 1.24 (1.16,1.33)
Cool days (<20oC): 1.05 (0.96,1.15)
Two Pollutant Models
Warm days (>20o C)
Adjusted for PM10:1.16 (1.08,1.26)
Adjusted for N02:1.02 (0.92,1.13)
Adjusted for 03:1.25 (1.17,1.34)
Adjusted for S02:1.32 (1.22,1.42)
Cool days (<20o C)
Adjusted for PM10:1.09 (0.97,1.21)
Adjusted for N02:1.07 (0.92,1.25)
Adjusted for 03:0.89 (0.80, 0.99)
Adjusted for S02:1.03 (0.92,1.16)
September 2009
C-18
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CARDIOVASCULAR DISEASES - NON-SPECIFIC
Author: Ballester et al. (2001,
013257)
Period of Study: 1994-1 996
Location: Valencia, Spain
ED Visits
Health Outcome (ICD9: CVD (390-
459); Heart Diseases (410-414, 427,
428); Cerebrovascular Disease
(430-438)
Study Design: Time-series
Statistical Analyses: Poisson
regression
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 6.2 mg/m3
Range (Min, Max): 0.6, 17.8
Copollutant: correlation
BS:r= 0.64
N02:r = 0.03
S02:r=0.74
03:r=-0.26
Increment: 1 mg/m3
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0, 1 , 2, 3,
All cardiovascular:
Lag 2: 1.0077 (0.991 2-1 .01 38)
Heart Disease:
Lag1 : 1.0092 (0.9945-1 .0242)
Cerebrovascular Disease:
Lag1 : 0.9874 (0.9646-1 .01 07)
;lag:
4,5
Author: Ballester et al. (2006,
088746)
Period of Study: 1995-1999
Location: 14 Cities in Spain
Health Outcome (ICD9:
All CVD (390-459) ;Heart Diseases (410-
414,427,428)
Study Design: Time-series
Statistical Analyses: GAM
Age Groups Analyzed: All
Averaging Time: 8 h
Mean (SD) unit:
Range across 14 cities,
1.4-2.8 mg/m3
Range (percentiles):
10th = 0.4-1.7; 90th = 2.0-3.9
Increment: 1 mg/m3
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0-1
All CVD: Lag 0-1 :2.06 (0.65-3.48)
Heart Disease: Lag 0-1 :4.15 (1.31-7.08)
Author: Barnett et al. (2006,
089770)
Period of Study: 1998-2001
Location:
Brisbane, Canberra,
Melbourne, Perth, Sydney
Australia
Auckland & Christchurch, New
Zealand
Sample Description: NR
Hospital Admissions with
Cardiovascular Diseases
Health Outcome (ICD9:Arrythmia
(247); Cardiac Disease (390-429);
Cardiac Failure (428); IHD (410-413); M
(410); Total CVD (390-459)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
15-64yr&>65yr
Sample Description: NR
Copollutant: NR
Averaging Time: 8 h
Mean (SD) unit: ppm
Brisbane: 1.7
Canberra: 0.9
I Melbourne: 1.0
Perth: 1.0
Sydney: 0.8
Auckland: 2.1
Christchurch: 0.5
Range (Min, Max): ppm
Brisbane: 0.0, 7.0
Canberra: 0.0, 5.8
Melbourne: 0.1, 8.0
Perth: 0.1, 4.0
Sydney: 0.0, 4.5
Auckland: 0.2, 7.9
Christchurch: 0.0, 5.4
Increment: 0.9 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0-1
15-64yr
Arrythmia: 2.5 (0.1-4.9)
Cardiac: 1.7 (0.5-2.9)
Cardiac Failure: 4.2 (0.6-7.8)
IHD: 1.6 (-0.6 to 3.9)
MM 8 (-07 to 4 3)
Total CVD: 1.2 (0.3-2.1)
>65yr
Arrythmia: 0.1 (-1.8 to 2.1)
Cardiac: 2.8 (1.3-4.4)
Cardiac Failure: 6.0 (3.5-8.5)
IHD: 2.3 (0.9-3.8)
Ml: 2.9 (0.8-4.9)
Total CVD: 2.2 (0.9-3.4)
Copollutant NR
September 2009
C-19
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Bell etal. (2009,
193780)
Period of Study: 1999-2005
Location:
126 U.S. urban counties
Hospital Admissions with
Cardiovascular Diseases
Health Outcome (ICD9): Cardiac
Failure (428); Cerebrovascular Events
(430-438); Heart Rhythm Disturbances
(426-427); IHD (410-414,429);
Peripheral Vascular Disease (440-448)
Study Design: Time-series
Statistical Analyses: Log-linear over-
dispersed Poisson regression
Averaging Time: 1 h
Mean (SD) unit: 1.6 ppm
Median (SD) unit: 1.3 ppm
Median Range (Min, Max):
0.2,9.7
Copollutant:
PM25:r = 0.26
N02:r = 0.56
Increment: 1 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0-2
LagO:
Single pollutant model: 0.96 (0.79-1.12)
Adjusted for PM25:0.76 (0.57-0.96)
Adjusted for N02:0.55 (0.36-0.74)
Adjusted for EC : 0.97 (0.38-1.57)
Author: Chang et al. (2005,
080086)
Period of Study: 1997-2001
Location:
Taipei, Taiwan
Author: Filhol. (2008, 190260)
Period of Study:
January 2001 -July 2003
Location:
Sao Paulo, Brazil
Age Groups Analyzed: > 65 yr
Sample Description:
>9.3 million Medicare subjects
Health Outcome (ICD9: CVD Hospital
Admissions
(410-429)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
74,509 CVD hospital admissions
(47 Hospitals)
ED Visits
Health Outcome (ICD10): Hypertension
and Cardiac Ischemic Disease (110-125)
Study Design: Time-series
Statistical Analyses: Linear Poisson
regression models
Age Groups Analyzed: >18yr
Sample Description: 45,000
Cardiovascular emergency room visits
from diabetic and non-diabetic patients
(tertiary referral teaching hospital)
tu: r = UAQ
Averaging Time: 24-h
Mean(SD) unit: 1.37 ppm
Range (Min, Max): 0.37, 3.66
Copollutant: NR
Averaging Time: 8 h
Mean (SD) unit: 2.7 ppm
Range (Min, Max): 0.7, 12.1
Copollutant: correlation
PM10:r = 0.69
N02:r = 0.58
S02:r=0.52
03:r = 0.07
Increment: 0.49 ppm
OR Estimate [Lower Cl, Upper Cl]
Lag examined (days) : 0-2
>20°C: 1.090 (1.064-1.118)
<20°C: 0.984 (0.927-1 .044)
Adjusted for PM10:
>20°C'1 171 (1 132-1 211)
<20°C: 0.946 (0.892-1 .003)
Adjusted for S02:
>20°C: 1.232 (1.1 94-1 .272)
<20°C: 1.098 (1.034-1 .165)
Adjusted for N02:
>20°C: 1.048 (1.003-1 .095
<20°C: 0.983 (0.91 4-1 .058)
Adjusted for 03:
>20°C: 1.1 96 (1.1 61 -1.232)
<20°C: 1.092 (1.031 -1.1 57)
Increment: 1 .2 ppm
Regression Coefficients [SEM]
Lags examined (days): 0,1,2
CVD Visits/Diabetes:
Lag 0:0. 0575 (0.0410)
Lag 1:-0. 0056 (0.041 8)
Lag 2: -0.0324 (0.0426)
2-day moving avg: 0.0324 (0.0470)
3-day moving avg: 0.0074 J0.0528)
4-day moving avg: -0.0025 (0.0582)
CVD Visits/Non-Diabetes:
Lag 0:0.0286 (0.0095)
Lag 1:0.0098 (0.0091)
Lag 2:0.0102 (0.0089)
2-day moving avg: 0.0271 (0.0108)
3-day moving avg: 0.0281 J0.0120)
4-day moving avg: 0.0306 (0.0131)
Author: Fung et al. (2005,
074322)
Period of Study: 1995-2000
Location:
Windsor, Ontario,
Canada
Hospital Admissions of
Cardiovascular Diseases
Health Outcome (ICD9): CHF (428);
IHD (410-414); Dysrythmias (427)
Study Design: Time-series
Statistical Analyses :GLM
Age Groups Analyzed: All
Sample Description:
11,632 Cardiac hospital admissions
Averaging Time: 24-h
Mean (SD) unit: 1.3 ppm
Range (Min, Max): 0.0,11.8
Copollutant: correlation
PM10:r = 0.21
N02:r = 0.38
S02:r=0.16
03:r = 0.10
Increment: 1.2 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0, 0-1, 0-2
<65yr
LagO:-3.1 (-7.4 to 1.4)
Lag 0-1 : -2.7 (-8.1 to 3.0)
Lag 0-2 : -0.5 (-6.7 to 6.0)
>65yr
Lag 0:0.5 (-2.2 to 3.3)
Lag 0-1 :2.3 (-1.1 to 5.9)
Lag 0-2:2.8 (-1.1 to 7.0)
September 2009
C-20
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Jalaludin et al. (2006,
189416)
Period of Study:
1997-2001
Location:
Sydney, Australia
ED Visits
Health Outcome (ICD9):
All Cardiovascular (390-459); Cardiac
Disease (390-429); IHD (410-413);
Cerebrovascular or Stroke (430-438)
Study Design: Time-series
Statistical Analyses:
GLM & GAM
Age Groups Analyzed:
65+yr
Sample Description: NR
Averaging Time: 8 h
Mean (SD) unit: 0.82 ppm
Range (Min, Max): 0.02, 4.63
Copollutant: correlation
PM10:r = 0.31
N02:r = 0.71
S02:r=0.51
03:r = 0.19
Increment: 0.69 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2, 3,0-1
All Cardiovascular:
Lag 0:2.32 (1.45-3.19)
Lag1 :1.33 (0.47-2.20)
Lag 0-1 :2.35 (1.39-3.32)
Cardiac Disease:
Lag 0:2.52 (1.50-3.54)
Lag1 :1.85 (0.83-2.88)
Lag 2 :1.11 (0.0-2.15)
Lag 0-1 :2.85 (1.71-4.01)
IHD:
Lag 0 :2.83 (1.22-4.48)
Lag1 :1.58 (0.01-3.19)
Lag 0-1 :2.86 (1.07-4.68)
Stroke: No results were significant for
Stroke.
All Cardiovascular Disease:
Cool period: Lag 0 :3.26 (2.00-4.53)
Cardiac Disease:
Cool period: Lag 0 :3.43 (1.95-4.93)
IHD:
Cool period: Lag 0 :3.64 (1.28-6.06)
Warm period: Lag 0 :2.29 (0.01-4.62)
Stroke:
Cool period: Lag 0 :3.54 (0.78-6.37)
Notes:
Cool: May to October
Warm : November to April
Author: Kokenetal. (2003,
049466)
Period of Study: 1993-1997
Location:
Denver, CO
Hospital Admissions for Cardiovascular
Disease
Health Outcome (ICD9: Ml (410-
410.92); Coronary Atherosclerosis
(414-414.05); Pulmonary Heart Disease
(416-416.9); Cardiac Dysrythmia
(427-427.9);
CHF (428)
Study Design: Time-series
Statistical Analyses: GLM
Age Groups Analyzed: >65 yr
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 0.9 ppm
Range (Min, Max): 0.3,1.6
Copollutant: correlation
PM10:r = 0.25
N02:r = 0.73
S02:r=0.21
03:r=-0.40
Increment: 0.3 ppm
% Change [Lower Cl, Upper Cl]
Lags examined (days): 1,2,3,4
CHF: Lag 3:10.5 (0.1-22.0)
CO not significantly associated with other
Lag periods.
Author: Linn et al. (2000,
002839)
Period of Study: 1992-1995
Location:
Los Angeles, CA
Health Outcome:
Hospital Admissions for Cardiovascular,
Cerebrovascular, Pulmonary.
Study Design: Time-series
Statistical Analyses:
Ordinary least squares regression;
Poisson regression
Age Groups Analyzed: >30 yr
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit:
Winter: 1.7; Spring: 1.0;
Summer: 1.2; Fall: 2.1
Range (Min, Max):
Winter: 0.5,5.3; Spring: 0.4, 2.2;
Summer: 0.3,2.7; Fall: 0.2, 4.3
Copollutant: correlation
Winter:
PM10: r = 0.78; N02:r = 0.89;
03:r = -0.43;
Spring:
PM10: r = 0.54; N02:r = 0.92;
03:0.29
Summer:
PM10: r = 0.72; N02:r = 0.94;
03:0.03
Fall:
PM10: r = 0.58; N02:r = 0.84;
03:r=-0.36
Increment: 1 ppm
Co-efficient [SE]
Lags examined (lags): 0,1
LagO:
Cardiovascular
All: 0.032 (0.003)* (e.g. 3.2% increase)
Winter: 0.038 (0.006)*
Spring: 0.010(0.015)
Summer: 0.035 (0.014)*
Fall: 0.027 (0.006)*
Cerebrovascular
All: 0.009 (0.007)
Winter:-0.008 (0.014)
Spring: 0.107 (0.033)*
Summer: 0.030 (0.033)
Fall: 0.008 (0.012)
Myocardial Infarction
All: 0.040 (0.009)*
CHF
All: 0.025 (0.009)*
CardiacArrythmia
All: 0.023 (0.009)*
Stroke
All: 0.044 (0.009)*
Notes:*p< 0.05
September 2009
C-21
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Metzger et al. (2004,
044222)
Period of Study: 1993-2000
Location:
Atlanta, GA
ED Visits (from 31 hospitals)
Health Outcome (ICD9:
Cardiovascular:
IHD(410-414);AcuteMI(410);
Dysrythmia (427); Cardiac Arrest
(427.5); CHF (428); Peripheral Vascular
& Cereberovascular Disease (PVCD)
(433-437,440,443,444,451-453);
Atherosclerosis (440); Stroke (436)
Study Design: Case-crossover
Statistical Analyses:
Poisson regression (GLM)
Age Groups Analyzed: All
Sample Description: 4,407,535 visits
Averaging Time: 1 h
Median (SD) unit: 1.5 ppm
Range (percentiles):
10th = 0.5; 90th = 3.4
Copollutant: correlation
PM10:r = 0.47
N02:r=0.68
S02:r=0.26
03: r =0.20
Increment: 1 ppm
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0-2ma
All CVD: 1.017 (1.008-1.027)
Dysrythmia : 1.012 (0.993-1.031)
CHF: 1.010(0.988-1.032)
IHD: 1.016 (0.999-1.034)
PVCD: 1.031 (1.010-1.052)
Author: Peel et al. (2007,
090442)
Period of Study: 1993-2000
Location:
Atlanta, GA
ED Visits (from 31 hospitals)
Health Outcome (ICD9:
Cardiovascular:
IHD (410-414); Dysrythmia (427); CHF
(428); PVCD (433-437, 440, 443, 444,
451-453)
Averaging Time: 1-h
Mean(SD) unit: 1.8 ppm
Range (SD):SD: 1.2
Copollutant: NR
Increment: 1 .2 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0-2ma
IHD:
Without Diabetes : 1 .023 (1 .004-1 .420)
\Afithniit rut- 1 r\OA 11 nnR_i r\AO\
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: All
Sample Description: 4,407,535 visits
Dysrythmias:
With Hypertension : 1.065 (1.015-1.118)
PVCD:
With Hypertension : 1.038 (1.004-1.074)
Without Hypertension: 1.027 (1.002-1.054)
With Diabetes: 1.065 (1.012-1.121)
Without Diabetes: 1.025 (1.003-1.048)
With COPD: 1.113 (1.027-1.205)
Without COPD: 1.026 (1.004-1.047)
Without CHF: 1.029 (1.008-1.051)
With Dysrythmias: 1.072 (1.011-1.138)
Without Dysrythmias : 1.026 (1.004-1.048)
CHF: With COPD : 1.058 (1.003-1.115)
Author: Slaughter et al. (2005,
073854)
Period of Study: 1995-2001
Location:
Spokane, WA
Author: Tolbertetal. (2007,
090316)
Period of Study: 1993-2004
Location:
Atlanta, GA
Health Outcome (ICD9: Cardiac Hospital
Admissions: (390-459)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GLM & GAM)
Age Groups Analyzed: All
Sample Description: NR
ED Visits (from 41 hospitals)
Health Outcome (ICD9): IHD (410-
414), cardiac dysrhythmias (427), CHF
(428), peripheral vascular and
cerebrovascular diseases (433-437,
440, 443-445 and 451 -453)
Study Design: Time-series
Statistical Analyses: Poisson
generalized linear model
Averaging Time: 24-h
Mean (SD) unit: 0.42-1 .82
Range (Min, Max): NR
Copollutant correlation :
PM10:r = 0.32
PM2.5: r = 0.62
Averaging Time: 1 h
Mean (SD) unit: 1 .6 ppm
Range (Min, Max): 0.1, 7.7
Copollutant:
PM10:r = 0.51
N02:r=0.70
S02:r=0.28
03:r=0.27 PM25:
r = 0.47
Increment: NR
RR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined (days): 1,2,3
No significant association. Results not
reported.
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 1,2,3
Single Pollutant Model
3-day moving avg: 1 .020 (1 .010, 1 .030)
Results for multi-pollutant models presented
graphically
Age Groups Analyzed: NR
Sample Description: 10,234,490 ED
Visits (238,360 CVD group)
September 2009
C-22
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Yang etal. (2004,
094376)
Health Outcome (ICD9:
Cardiovascular diseases (410-429)
Period of Study: 1997-2000 Study Design: Case-crossover
Location:
Kaohsiung City, Taiwan
Statistical Analyses: Conditional
logistic regression
Age Groups Analyzed: All
Sample Description:
29,661 Cardiovascular hospital
admissions (63 hospitals)
Averaging Time: 24-h
Mean (SD) unit: 0.79 ppm
Range (Min, Max): 0.24,1.72
Copollutant: NR
Increment: 0.28 ppm
OR Estimate [Lower Cl, Upper Cl]
Lag examined (days): 0-2
>25°C: 1.264 (1.205-1.326)
<25°C: 1.448 (1.357-1.545)
Adjusted for PM10:
>25°C: 1.206 (1.146-1.270)
<25°C: 1.314 (1.213-1.423)
Adjusted for S02:
>25°C: 1.406 (1.327-1.489)
<25°C: 1.3450 (1.352-1.555)
Adjusted for N02:
>25°C: 1.246 (1.166-1.332)
<25°C: 0.905 (0.819-0.999)
Adjusted for 03:
>25°C:1.250 (1.191-1.311)
<25°C: 1.447 (1.356-1.545)
Table C-3 Studies of CO exposure and neonatal and postneonatal outcomes.
Study
Author: Bell etal. (2007,
091059)
Period of Study: 1999-2002
Location:
Connecticut and
Massachusetts
Author: Brauer etal. (2008,
156292)
Period of Study: 1999-2004
Location:
Vancouver, Canada
Author: Chen et al. (2002,
024945)
Period of Study: 1991 -1999
Location:
Northern Nevada
Design
Health Outcome:
Birth weight and LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
358,504 full term live
singleton births (32-44 wk)
Health Outcome: LBW,
PTB and SGA
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
70,249 live singleton births
Health Outcome:
Birth weight & LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
39,338 full term live
singleton births (37-44 wk)
Concentrations
Averaging Time: 24-h
Mean (SD) unit:
0.65 ppm J0.18)
Range (Min, Max): NR
Copollutant: NR
Averaging Time:
Land use regression model
Mean (SD) unit: 633 ug/m3
Range (Min, Max):
124, 1409
Copollutant: correlation:
PM10:r = 0.73
N02:r = 0.75
S02:r=0.82
03:r=-0.39
Averaging Time: 8 h
Mean (SD) unit: 0.98 ppm
Range (Min, Max):
0.25,4.87
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: Interquartile range -0.30 ppm
Regression co-efficient for birth weight (g) [Lower
Cl, Upper Cl]
Entire pregnancy: -16.2 (-19.7 to -12.6)
Stratified by race.
Black mother: -10.9 (-20.2 to -1.6)
White mother: -17.5 (-21 .3 to -13.7)
OR for LBW [Lower Cl, Upper Cl]
Entire pregnancy : 1 .028 (0.983-1 .074)
Increment: 100 ug/m3
OR for SGA [Lower Cl, Upper Cl];
Entire pregnancy : 1 .06 (1 .03-1 .08)
OR for term LBW [Lower Cl, Upper Cl] ;
Entire pregnancy : 1 .02 (0.96-1 .09)
OR PTB [Lower Cl, Upper Cl];
Entire pregnancy : 1 .1 6 (1 .01 -1 .33)
Increment: NR
Regression co-efficient for birth weight (g) [SE]
Trimesters:
First: -1.02 (6.68)
Second : -0.07 (6.58)
Third : -3.95 (6.76)
Entire pregnancy :-8.28 (14.9)
Notes: CO not associated with LBW
September 2009
C-23
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Conceicao et al.
(2001.016628)
Health Outcome: Child
mortality, under 5 yr of age
Period of Study: 1994-1997 Study Design: Time-series
Location:
Sao Paulo, Brazil
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit:
4.4 ppm (2.2)
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Regression co-efficient for Child mortality - under 5
yr of age [SE];
Lags examined : 0,1,2,3
Lag 2 :0.0306 (0.0076) (p < 0.01)
Lag chosen for best fitting model
Author: Gilboaetal. (2005,
087892)
Period of Study: 1997-2000
Location:
Texas
Health Outcome:
Birth defects (heart defects &
orofacial clefts)
Study Design: Case-control
Statistical Analyses:
Conditional Logistic
regression
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: Exposure categories (ppm):<0.4; 0.4-
0.5;0.5-0.7;>0.7
OR for Birth Defects [Lower Cl, Upper Cl];
Exposure period : wk 3 - 8 of pregnancy
Conotruncal defects:
1.00; 1.38 (0.97-1.97); 1.17 (0.81-1.70); 1.46
(1.03-2.08)
Tetralogy of Fallot:
1.00; 0.92 (0.52-1.62); 1.27 (0.75-2.14);2.04
(1.26-3.29)
Notes: CO was not associated with the following
defects: Aortic artery & valve, atrial septal,
pulmonary artery & valve, ventricular septal,
endocardia! cushion & mitral valve , cleft lip, cleft
palate, aortic valve stenosis, coarctation of the
aorta, ostium secundum.
Author: Gouveia et al. (2004,
055613)
Period of Study: 1997
Location:
Sao Paulo, Brazil
Author: Ha etal. (2001,
019390)
Period of Study: 1996-1 997
Location:
Seoul, South Korea
Author: Ha et al. (2003,
042552)
Period of Study: 1995-1 999
Location:
Seoul, South Korea
Health Outcome:
Birth weight & LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
179,460 live singleton term
births (>37 wk)
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression (GAM)
Age Groups Analyzed: NA
Sample Description:
276 763 full term live
singleton births (>37 wk)
Health Outcome:
Post-neonatal mortality
(1 mo-1 yr) (also looked at
older age groups)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: 8 h
Mean (SD) unit: 3.7 ppm
Range (Min, Max):
1.1,11.4
Copollutant: NR
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
Percentiles:
25th: 0.99 ppm
75th: 1.41 ppm
Copollutant correlation: TSP: r = 0.73
N02:r = 0.75
S02:r=0.82
03:r=-0.39
Averaging Time: 24-h
Mean (SD) unit: 1 .2 ppm
Range (Min, Max):
0.39,3.38
Copollutant correlation:
PM10:r = 0.63
N02:r = 0.72
S02:r=0.75
03:r=-0.46
Increment: 1 ppm
Regression co-efficient for birth weight (g) [Lower
Cl, Upper Cl]
Trimesters:
First : -23.1 (-41 .3 to -4.9); Second : 3.2 (-18.2 to
24.5);
Third: 1.9 (-18.2 to 22.0)
OR for LBW) [Lower Cl, Upper Cl]
4th quartile exposure (compared to lowest
quartile):
First : 1 .02 (0.82-1 .27); Second : 1 .07 (0.88-1 .30);
Third: 0.93 (0.76-1 .12)
Increment: 0.42 ppm
RR for LBW [Lower Cl, Upper Cl]
Trimesters:
First: 1.08 (1.04, 1.12)
Third : 0.91 (0.87, 0.96)
Increment: 0.57 ppm
RR for Post-neonatal mortality (1 mo-1 yr)
[Lower Cl, Upper Cl]
Lags examined : 0
Total mortality:
Lag 0:1. 020 (0.976-1 .067)
Respiratory mortality:
Lag 0:1. 388 (1.009-1 .911)
September 2009
C-24
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Hajat et al. (2007,
093276)
Period of Study: NR
Location:
Birmingham, Bristol, Leeds,
Liverpool, London,
Manchester, Middlesbrough,
Newcastle, Nottingham,
Sheffield
England
Health Outcome: Neonatal
and post-neonatal mortality
Study Design: Time-series
Statistical Analyses:
Poisson regression (GLM)
Age Groups Analyzed: NA
Sample Description:
22,288 total infant deaths
between 1990 and 2000
Averaging Time: 3 days
Mean (SD) unit: (mg/m3)
Increment: 1 mg/m3
RR Estimate [Lower Cl, Upper Cl]
Birmingham: 0.64; Bristol: 1.01; Leeds: Lags examined (days): 0,1, 2
0.73; Liverpool: 0.51; London: 0.77;
Manchester: 0.63; Middlesbrough: 0.37; All infant deaths: 1.02 (0.96,1.09)
Newcastle: 0.67; Nottingham: 0.62;
Sheffield: 0.60
Neonatal deaths: 0.99 (0.92,1.07)
Range (Min, Max):
Birmingham: 0.4,0.8; Bristol: 0.6,1.2;
Leeds: 0.5,0.9; Liverpool: 0.3,0.6;
London: 0.5, 0.9; Manchester: 0.4,0.7;
Middlesbrough: 0.2, 0.4; Newcastle: 0.5,
0.8; Nottingham: 0.4,0.7; Sheffield: 0.3,
0.7
Copollutant: S02, N02, NO, 03, PM10
Post-neonatal deaths: 1.09 (0.94,1.25)
City specific results of all infant mortality displayed
graphically
Author: Huynh et al. (2006,
091240)
Health Outcome:
PTB (24-36 wk gestation)
Period of Study: 1999-2000 Study Design: Case-control
Location:
California
Statistical Analyses:
Conditional Logistic
regression
Age Groups Analyzed:
Cases = 24-36 wk gestation;
Controls = 39-44 wk
Sample Description:
10,673 PTBs (cases); 32,119
term births (controls)
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: 1 ppm
Exposure level - Quartiles of exposure for first mo
and last two wk of gestation (mg/m3)
First: <0.61; Second : 0.61 - 0.82;Third : 0.82 -
1.07;
Fourth :>1.07
Quartiles for entire pregnancy and last two wk of
pregnancy were similar.
OR for PTB [Lower Cl, Upper Cl]
First mo of gestation:
Per 1 ppm increase : 1.10 (0.99-1.20)
Second quartile : 0.94 (0.88-1.01)
Third quartile: 1.04 (0.97-1.11)
Fourth quartile : 1.05 (0.96-1.14)
Last two wk of gestation:
Per 1 ppm increase : 1.00 (0.93-1.09)
Second quartile : 1.03 (0.97-1.10)
Third quartile: 1.04 (0.97-1.12)
Fourth quartile : 0.99 (0.91-1.08)
Entire pregnancy:
Per 1 ppm increase : 1.06 (0.95-1.18)
Second quartile :0.97 (0.91-1.04)
Third quartile : 0.99 (0.92-1.05)
Fourth quartile : 1.02 (0.94-1.09)
Lowest quartile used as reference group
Author: Hwang and Jaakkola
(2008,193794)
Period of Study: 2001-2003
Location: Taiwan
Health Outcome: Oral clefts Averaging Time: 8 h
(with or without palate)
v H ' Mean (SD) unit: 0.69 (0.4)
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
6,530 cases from 721,289
newborns
Range (Min, Max): 0.25, 2.7
Copollutant correlation: PM10: r =
0.19
N0x:r = 0.82
S02:r=0.24
03:r=-0.19
Increment: 100ppb
RR for oral cleft [Lower Cl, Upper Cl]
Month! : 1.00 (0.96-1.04)
Month 2 :1.00 (0.96-1.03)
Months : 1.00 (0.96-1.03)
Author: Jalaludin et al. (2007, Health Outcome: PTB
Study Design:
Period of Study: 1998-2000 Retrospective cohort
Averaging Time: 8 h
Increment: 1 ppm
RR for PTB [Lower Cl, Upper Cl]
Location:
Sydney, Australia
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
123,840 full term live
singleton births (<42 wk)
Range (Min, Max): NR
Mean (SD) unit:
0.9 ppm (0.68)
™ ^ ' First mo:
All of Sydney : 0.89 (0.84-0.95)
Within 5km of site : 1.03 (0.68-1.54)
Copollutant correlation: PM10: r = 0.28 pjrs( trimester'
N02: r = 0.60 An of Sydney : 0.77 (0.71-0.83)
S02:r=0.24 WithinSkmof site : 1.24 (0.81-1.91)
03:r--0.21 1 mo prior to birth:
All of Sydney : 0.96 (0.88-1.04)
Within 5km of site: 1.00 (0.86-1.15)
3 mo prior to birth:
All of Sydney : 0.99 (0.90-1.09)
Within 5km of site :1.11 (0.94-1.31)
September 2009
C-25
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Lee etal. (2003,
043202)
Period of Study:
1996-1998
Location:
Seoul, South Korea
Author: Leem etal. (2006,
089828)
Period of Study: 2001 -2002
Location:
Incheon, Korea
Author: Lin et al. (2004,
095787)
Period of Study: 1998-2000
Location:
Sao Paulo, Brazil
Author: Lin et al. (2004,
089503)
Period of Study: 1995-1 997
Location:
Taipei & Kaoshiung, Taiwan
Design
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
388,105 full term live
singleton births (37-44 wk)
Health Outcome:
PTB
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
52,113 live singleton births
Health Outcome:
Neonatal death (within first
28 days of life)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed: NA
Sample Description: NR
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
92,288 full term live
singleton births (>37 wk)
within 3km of monitoring
site.
Concentrations
Averaging Time: 24-h
Mean (SD) unit: 1 .2 ppm
Range (Min, Max): 0.4, 3.4
Copollutant correlation: PM10: r = 0.47
N02:r = 0.77
S02:r=0.79
Averaging Time:
Kriging was used to estimate exposure
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant correlation: PM10: r = 0.27
N02: r = 0.63
S02:r=0.31
Averaging Time: 24-h
Mean (SD) unit: 2.83 ppm
Range (Min, Max):
0.54,10.25
Copollutant correlation:
PM10:r = 0.71
N02:r = 0.67
S02'r=055
03:r = 0.03
Averaging Time: 24-h
Mean (SD) unit:
Taipei (avg over 5 sites)
0.84-1.31
Kaohsiung (avg over 5 sites)
5.56-10.05
Range (Min, Max): NR
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: 0.5 ppm
OR for LBW [Lower Cl, Upper Cl]
First: 1.04 (1.01 -1.07)
Second: 1.03 (1.00-1 .06)
Third : 0.96 (0.93-0.99)
Entire pregnancy : 1 .05 (1 .01 -1 .09)
Increment:
Exposure level - Quartiles of exposure for first
trimester (mg/m3)
First: 0.47-0.63; Second: 0.6 -0.77;
Third: 0.78-0.90; Fourth: 0.91-1 .27
- exposure groups for third trimester was similar
OR for PTB [Lower Cl, Upper Cl]
First Trimester:
Second quartile : 0.92 (0.81-1.05)
Third quartile: 1.1 4 (1.01 -1.29)
Fourth quartile : 1.26 (1.11-1.44)
Third Trimester:
Second quartile : 1.07 (0.95-1.21)
Third quartile: 1.07 (0.94-1 .22)
Fourth quartile : 1.16 (1.01-1.34)
Lowest quartile used as reference group.
Increment: NR
Regression coefficent for neonatal death [SE]
Lags examined : 0
Lag 0:0.0061 (0.0110)
Increment: Exposure groups
M = Median exposure 1.1-14.2 ppm
H = High exposure >14.2 ppm
OR for LBW [Lower Cl, Upper Cl]
Trimesters:
First : M 1 .01 (0.89, 1 .16), H 0.90 (0.75, 1 .09)
Second :M 1.02 (0.90, 1.16), H 1.00 (0.82, 1.22)
Third : M 0.88 (0.77, 1 .00), H 0.86 (0.71 , 1 .03)
Entire pregnancy :
M 0.89 (0.77, 1.01), H 0.77 (0.63, 0.94)
Notes: Cut off for exposures groups for second and
third trimester were similar to those presented
above.
September 2009
C-26
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Liu et al. (2003,
089548)
Period of Study: 1985-1 998
Location:
Vancouver, BC, Canada
Design
Health Outcome:
PTB, IUGR, LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Concentrations
Averaging Time: 24-h
Mean (SD) unit: 1 .0 ppm
Range (Min, Max):
25th: 0.7; 75th: 1.2
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: 1 .0 ppm
OR for LBW [Lower Cl, Upper Cl]
Month of pregnancy:
First mo: 1.01 (0.93-1.09)
Last mo: 0.96 (0.88-1 .04)
Age Groups Analyzed: NA
Sample Description:
229,085 live singleton births
First mo : 0.95 (0.89-1.01)
Last mo: 1.08 (1.01-1.15)
OR for IUGR [Lower Cl- Upper Cl]
First mo : 1.06 (1.01-1.10)
Last mo: 0.98 (0.94-1.03)
Trimester! : 1.05 (1.00-1.10)
Trimester 2 :0.97 (0.92-1.01)
Trimesters :0.97 (0.93-1.02)
Author: Liu et al. (2007,
090429)
Period of Study: 1995-2000
Location:
Calgary, Edmonton,
and Montreal, Canada
Author: Maisonet et al.
(2001.016624)
Period of Study: 1994-1 996
Location:
Northeastern USA
Author: Mannes
etal. (2005, 087895)
Period of Study: 1998-2000
Location:
Sydney, Australia
Health Outcome: IUGR
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
386,202 live singleton births
Health Outcome:
Live birth weight
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
89,557 live singleton term
births (37-44 wk)
Health Outcome:
Birth weight and SGA
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
138,056 full term all
singleton births (including
stillbirths'! fat least ?0 wk
Averaging Time: 24-h
Mean (SD) unit: 1.1 ppm
Range (Min, Max):
25th: 0.6; 75th: 1.3
Copollutant correlation: PM:5: r = 0.31
N02:r = 0.71
S02:r=0.21
03:r=-0.42
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
Percentiles:
25th: 0.93 ppm; 75th: 1.23 ppm
Copollutant: NR
Averaging Time: 8 h
Mean (SD) unit: 0.8 ppm
Range (Min, Max): 0.0, 4.6
Copollutant: correlation
PMi0:r = 0.26
N02: r = 0.57
03:r= -0.20
Increment: 1 ppm
RR for LBW [Lower Cl, Upper Cl]
Notes: CO was associated with an increased risk
of IUGR of approximately 16% and 23% in the first
and nine mo of pregnancy.
(All results presented in Figures)
Increment: 1 ppm
OR for LBW [Lower Cl, Upper Cl]
Trimester:
First : 1 .08 (0.91 -1 .28); Second : 1 .1 4 (0.83-1 .58) ;
Third : 1 .31 (1 .06-1 .62)
Stratified results among African-Americans:
First : 1 .43 (1 .18-1 .74); Second : 1 .27 (0.87-1 .86);
Third: 1.75 (1.50-2.04)
Notes: CO had no effect on whites or Hispanics
Increment: 1 ppm
Regression coefficients for birth weight (g) [Lower
Cl, Upper Cl]
All births:
First trimester : 1 .86 (-8.31 to 12.03)
Second trimester : -10.72 (-23.09 to 1 .65)
Third trimester : -6.63 (-18.57 to 5.31)
One mo prior to birth : -15.28 (-25.59 to -4.97)
Births within 5 km of monitor:
First trimester : -8.56 (-28.60 to 10.68)
Second trimester : -28.87 (-50.98 to -6.76)
Third trimester : -22.88 (-44.58 to -1 .18)
One mo prior to birth : -10.41 (-30.03 to 9.21)
gestation)
OR for SGA [Lower Cl, Upper Cl]
All births:
First trimester: 0.95 (0.88-1.04)
Second trimester: 0.99 (0.90-1.10)
Third trimester: 1.01 (0.91-1.11)
One mo prior to birth : 1.06 (0.98-1.16)
Births within 5km of monitor:
First trimester: 0.99 (0.86-1.14)
Second trimester: 1.06 (0.90-1.25)
Third trimester: 1.05 (0.90-1.23)
One mo prior to birth : 1.10 (0.96-1.27)
September 2009
C-27
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Medeiros et al.
(2005, 156750)
Period of Study: 1998-2000
Health Outcome: Birth
weight and LBW
Study Design:
Retrospective cohort
Averaging Time: 24-h
Mean (SD) unit:
Daily mean shown in Figure (see paper)
Ranna IMin MflYV MR
Increment: 1 ppm
Regression co-efficient for birth weight (g) [Lower
Cl, Upper Cl]
Sao Paulo, Brazil
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
311,735 full term live
singleton births (37-41 wk)
Copollutant: NR
First: -11.9 (-15.5 to -8.2); Second : 4.9 (0.5-9.3);
Third: 12.1 (7.6-16.6)
OR for LBW [Lower Cl, Upper Cl]
4th quartile exposure (compared to lowest quartile)
First: 0.98 (0.91 -1.06); Second : 0.97 (0.90-1.05);
Third: 1.03 (0.96-1.11)
Author: Mortimer et al. (2008,
187280)
Period of Study:
November 2000-April 2005
Location: Central Valley of
Californinia
Health Outcome: Allergic
sensitization
Study Design: Cohort
Statistical Analyses: Chi-
square tests
Age Groups Analyzed:
6-11 yrs.
Sample Description: 170
children with asthma from
the FACES-LITE study
Averaging Time: 8 h
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
Entire Prenatal:
PM10:r = 0.32
N02:r = 0.74
03:r=-0.40
Trimester 2:
PM10:r = 0.32
N02:r = 0.68
03:r=-0.26
Increment: NR
Trimester specific results presented graphically
Single-pollutant Model for "sensitized to at least
one outdoor allergen"
OR adjusted foryr of birth and sex [Lower Cl,
Upper Cl]
Entire Pregnancy
24-h avg: 1.45 (1.02,2.07)
Daily max: 1.53 (1.01, 2.33)
8-h max: 1.55 (1.01, 2.37)
2nd Trimester
24-h avg: 1.52 (0.93,2.47)
Daily max: 1.50 (0.92, 2.45)
8-h max: 1.45 (0.90, 2.35)
Coefficient adjusted for yr of birth and sex [SE]
Entire Pregnancy
24-h avg: 1.33 (0.68)
Daily max:0.54 (0.27)
8-h max: 0.84 (0.42)
2nd Trimester
24-h avg: 0.57 (0.34)
Daily max: 0.21 (0.13)
8-h max: 0.32 (0.21)
Author: Parker etal. (2005,
087462)
Period of Study: 2000
Location:
California
Health Outcome:
Birth weight & SGA
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Sample Description:
18,247 full term live
singleton births (40 wk)
within 5 miles of a monitor
Averaging Time: 24-h
Mean (SD) unit: 0.78 ppm
Range (Min, Max): NR
Copollutant: NR
Increment: Quartiles of exposure for first trimester
First: <0.57; Second : 0.57-0.76 ;
Third : 0.76- 0.93; Fourth : >0.93
- exposure groups for other trimesters were similar
Regression co-efficient for birth weight (g) [Lower
Cl, Upper Cl]
Trimesters:
4th quartile exposure (compared to lowest quartile)
First: -7.3 (-29.7 to 15.0); Second : 14.2 (-8.9 to
37.3);
Third:-8.4 (-32.2 to 15.3);
Entire pregnancy: -20.5 (-40.1 to -0.8)
OR for SGA [Lower Cl, Upper Cl]
4th quartile exposure (compared to lowest quartile)
First: 0.91 (0.76-1.09); Second: 0.80 (0.66-0.97);
Third: 0.90 (0.75-1.10);
Entire pregnancy: 0.95 (0.81 -1.12)
September 2009
C-28
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Ritzetal. (2000,
012068)
Period of Study: 1989-1 993
Location:
Southern California
Health Outcome: PTB
Study Design:
Retrospective Cohort
Statistical Analyses:
Logistic regression
Averaging Time:
6-9 a.m.
Mean (SD) unit: 2.70 ppm
Range (Min, Max):
0.36,9.12
Increment: 3 ppm
RR for PTB [Lower Cl, Upper Cl]
Adjusted for various risk factors and season of
birth and conception
6 wk prior to birth : 1.04(0.99-1.10)
1 st mo of pregnancy : 1 .04 (0.99-1 .09)
Eligible study subjects were
singletons born at 26-44 wk
gestation
Sample Description:
97,518 neonates born in
Southern California
N0'2:r = 0.60
03:r=-0.44
Adjusted for various risk factors
6 wk prior to birth : 1.06 (1.02-1.10)
1st mo of pregnancy :1.01 (0.97-1.04)
Author: Ritzetal. (2002,
023227)
Period of Study: 1987-1993
Location:
Southern California
Health Outcome:
Birth defects (heart defects &
orofacial clefts)
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: Exposure categories: ppm
<1.14;1.14-1.57;1.57-2.39;>2.39
OR for Birth defects [Lower Cl, Upper Cl]:
Period of exposure - Second mo of pregnancy.
Aortic artery & valve defects:
1.00 (ref group); 1.10 (0.73-1.66); 1.25 (0.74-2.13);
0.93(0.47-1.85)
Pulmonary artery & valve anomalies:
1.00 (ref group); 1.09 (0.69-1.73); 0.92 (0.50-1.70);
1.00(0.46-2.17)
Ventricular septal defects:
1.00 (ref group); 1.62 (1.05-2.48); 2.09 (1.19-3.67);
2.95(1.44-6.05)
Conotruncal defects:
1.00 (ref group); 0.79 (0.47-1.32); 0.73 (0.36-1.47);
0.95 (0.38-2.38)
Notes: Results also presented for more specific
defects, however CO showed no association (see
paper Table 3.). CO not associated with orofacial
clefts)
Author: Ritzetal. (2006,
089819)
Period of Study: 1989-2000
Location:
Southern California
Health Outcome: Post-
neonatal mortality (28 days
to 1yr); All causes; SIDS
Averaging Time: 24-h
Mean(SD) unit: 1.63 ppm
Study Design: Case-control Ran9e (Min, Max):
0.38,3.44
Statistical Analyses:
Conditional Logistic
regression
Sample Description:
Mothers residing within
16 km of monitoring site
Copollutant: correlation
PM10:r = 0.33
N02:r = 0.72
03:r=-0.57
Increment: 1 ppm
OR for Post-neonatal death [Lower Cl, Upper Cl]
Exposure period : 2 wk prior to death, 1 mo prior to
death, 2 mo prior to death, 6 mo prior to death
All causes:
2 wk prior to death : 1.14 (1.03-1.25)
2 mo prior to death :1.11 (1.06-1.16)
SIDS:
2 mo prior to death : 1.19 (1.10-1.28)
Term/normal weight births
2 mo prior to death:
All causes: 1.12 (1.05-1.19)
SIDS: 1.17 (1.07-1.29)
Respiratory: 1.14 (0.95-1.36)
Preterm &/or LBW births
2 mo prior to death:
All causes: 1.12 (1.01-1.25)
SIDS: 1.46 (1.09-1.94)
Respiratory : 1.03 (0.83-1.27)
Notes: These results did not persist in
multipollutant models (CO, N02, PM10, 03)
September 2009
C-29
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Ritzetal. (2007,
096146)
Period of Study:
January-December 2003
Location:
Los Angeles, CA
Author: Salam et al. (2005,
087885)
Period of Study: 1975-1 987
Location:
California
Design
Health Outcome: PTB
Study Design:
Nested case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
A survey of 2,543 of 6,374
women sampled from a
cohort of 58,316 eligible
births in Los Angeles county.
Health Outcome: Birth
weight, LBW , IUGR
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Age Groups Analyzed: NA
Concentrations
Averaging Time: 24-h
Mean (SD) unit: NR
Copollutant correlation:
TSP: r = 0.73
N02:r = 0.75
S02:r=0.82
03:r=-0.39
Averaging Time: 24-h
Mean (SD) unit:
1.8ppm(0.9)
(Entire pregnancy)
Range: NR
Copollutant: correlation
PM10:r = 0.41
N02:r = 0.69
03:r=-0.27
CO Effect Estimates (95% Cl)
Increment: Exposure categories (ppm):
Less than 0.58: 0.59-0.91 ; 0.92-1 .25; >1 .25
RR for LBW [Lower Cl, Upper Cl]
First trimester:
1 .00 (Ref group); 1.17 (1 .08-1 .26); 1 .15
(1.05-1 .26); 1.25 (1.1 2-1 .38)
6 wk prior to birth
1 .00 (Ref group); 1 .00 (0.93-1 .08) ; 1 .08
(0.98-1 .20); 1.03 (0.93-1 .14)
Entire pregnancy:
1.00 (Ref group); 0.76 (0.70-0.82); 0.84
(0.77-0.91); 1.03 (0.91-1 .17)
Increment: Entire pregnancy 1 .2 ppm
Trimesters: First : 1 .4 ppm; Second : 1 .4 ppm;
Third : 1 .3 ppm
Regression co-efficient for birth weight (g) [Lower
Cl, Upper Cl]
Trimesters:
First: -21.7 (-42.3 to -1.1); Second: 11.3 (-9.7 to
32.3);
Third: 11. 8 (-8.4 to 32.1);
Sample Description:
3,901 infants from the
California Children's Health
Study
Entire pregnancy: 2.2 (-20.1 to 24.4)
OR for LBW [Lower Cl, Upper Cl]
Trimesters:
First: 1.0 (0.7-1.5); Second: 0.9 (0.6-1.3);
Third: 0.7 (0.5-1.1); Entire pregnancy: 0.8 (0.6-1.3)
OR for IUGR [Lower Cl, Upper Cl]
Trimesters:
First: 1.2 (1.0-1.4); Second: 1.0 (0.9-1.1);
Third: 1.0 (0.8-1.1); Entire pregnancy: 1.0 (0.9-1.2)
Author: Son etal. (2008,
190323)
Period of Study: NR
Location: Seoul, Korea
Health Outcome: Post-
neonatal mortality from all
causes
Study Design: Case-
crossover and Time-series
Statistical Analyses:
Conditional logistic
regression
Age Groups Analyzed: NA
Sample Description: 1,286
firstborn birth and infant
death records from
1999-2003 (only post-
neonatal deaths)
Averaging Time: 8 h
Mean (SD) unit: 1.01 ppm
Range (Min, Max): 0.29, 3.54
Copollutant:
PM10, N02, 03, S02
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0-7
Time Series: 1.323 (1.077,1.625)
Case-crossover(1:6): 1.029 (0.833,1.271)
CLR Analyses using different control selection
schemes
1:2:1.076(0.839,1.379)
1:4:0.981 (0.784,1.228)
1:6:1.029(0.833,1.271)
September 2009
C-30
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Strickland et al.
(2009, 190324)
Period of Study: NR
Location: Atlanta, GA
Author: Tsai et al. (2006,
090709)
Period of Study: 1994-2000
Location:
Kaoshiung, Taiwan
Author: Wilhelm et al. (2005,
088668)
Period of Study: 1994-2000
Location:
Los Angeles, CA
Design
Health Outcome:
Cardiovascular
malformations
Study Design:
Retrospective cohort
Statistical Analyses:
Poisson GLM
Age Groups Analyzed: NA
Sample Description:
Pregnancies reaching at
least 20 wk' gestation that
were conceived during
January 1, 1986-March 12,
2003
Health Outcome:
Postneonatal death
(27 days-1 yr old)
Study Design: Case-
crossover
Statistical Analyses:
Poisson regression
Age Groups Analyzed: NA
Sample Description: NR
Health Outcome:
Term LBW and PTB
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
51 8,254 births within 4 mi of
a monitoring station. Varied
according to analyses.
Concentrations
Averaging Time: 24-h
Mean (SD) unit:
By season of conception:
March-May: 0.9 ppm
June-August: 0.8 ppm
Sept.-Nov.: 0.9 ppm
Dec.-Feb.:0.7ppm
By yr of conception:
1986-1991:0.7 ppm
1992-1 997: 0.8 ppm
1998-2003: 0.7 ppm
Range (IQR): 0.3
Copollutant:
PM10 (24-h): r = 0.32
N02 (24-h): r = 0.41
03(8h):r=0.07
S02 (24-h): r = 0.23
Averaging Time: 24-h
Mean (SD) unit:
8.27 ppm x10
Range (Min, Max):
2.26,17.7
Copollutant: NR
Averaging Time: 24-h
Mean (SD) unit:
Trimester 1 : 1 .42 ppm
Results for third trimester and 6 wk prior
to birth were similar to first trimester
Range (Min, Max):
0.26,2.82
Copollutant correlation:
First Trimester:
PM10:r = 0.12
PM25:r = 0.57
N02:r = 0.81
S02:r=-0.31
CO Effect Estimates (95% Cl)
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Atrial septal defect, secundum: 1 .16 (0.67, 2.00)
Coarctation of the aorta: 1.15 (0.65, 2.06)
Hypoplastic left heart syndrome: 0.82 (0.37, 1 .84)
Patent ductus arteriosus: 1 .39 (0.72, 2.68)
Pulmonary stenosis, valvar: 0.97 (0.53, 1 .75)
Tetralogy of Fallot: 1 .09 (0.59, 2.00)
Transposition of the great arteries: 1 .29 (0.58,
2.85)
Ventricular septal defect, muscular: 1.08 (0.77,
1.50)
Ventricular septal defect, perimembranous: 1 .06
(0.67,1.68)
Conotruncal defect: 1 .22 (0.81 , 1 .85)
Left ventricular outflow tract defect: 1 .09 (0.70,
1.68)
Right ventricular outflow tract defects: 0.73 (0.44,
1.22)
Increment: Interquartile range : 0.31 ppm
OR for Post-neonatal mortality [Lower Cl, Upper
Cl]
Lag examined : 0-2
Lag 0-2: 1.051 (0.304-3.630)
Increment: 1 ppm
RR for PTB [Lower Cl, Upper Cl]
First trimester:
Less than 1 mile: 1.06 (1.00-1.12)
1-2 miles: 1.06 (1.03-1 .10)
2-4 miles: 1.08 (1.06-1 .09)
ZIP code level: 1.04 (1.01 -1.07)
6 wk prior to birth:
Less than 1 mile: 1.04 (0.98-1.09)
1-2 miles: .04 (1.01 -1.08)
2-4 miles: 1.01 (0.99-1.02)
ZIP code level: 1.03 (1.00-1 .06)
Notes: All results above did not persist in
multipollutant model (CO, N02, 03, PMm)
OR for term LBW [Lower Cl, Upper Cl]
Third trimester:
Less than 1 mile: 1.10 (0.98-1.23)
1-2 miles: 1.05 (0.99-1.13)
2-4 miles: 1.06 (1.02-1.10)
ZIP code level: 1.12 (1.05-1.19)
Notes: All results above did not persist in
multipollutant model (CO, N02, 03, PM10)
See paper for results based on exposure category
groupings.
September 2009
C-31
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Woodruff et al.
(2008, 098386)
Period of Study: 1999-2002
Location:
U.S. counties with >250,000
residents
Health Outcome:
Post-neonatal deaths
All causes; respiratory;
SIDS; ill-defined + SIDS;
other causes.
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression (GEE)
Averaging Time: 24-h
Mean (SD) unit:
All causes: 0.70 ppm
Range (Min, Max):
Percentiles:
25th: 0.48; 75th: 0.87
Copollutant correlation: PM10: r = 0.18
S02:r= 0.27
03:r=-0.46
Increment: 0.39 ppm
OR for Post-neonatal mortality [Lower Cl, Upper
Cl]
Avg exposure over the first 2 mo of life:
All causes: 1.01 (0.95-1.07)
Respiratory: 1.1 4 (0.93-1 .40)
SIDS: 0.88 (0.76-1 .03)
Ill-defined + SIDS: 0.93 (0.84-1.02)
Other causes: 1.02 (0.97-1 .07)
Age Groups Analyzed: NA
Sample Description: NR
Author: Yang etal. (2004,
094376)
Period of Study: 1994-2000
Location:
Taipei, Taiwan
Health Outcome:
Post-neonatal mortality
(27 days-1 yr old)
Study Design: Case-
crossover
Statistical Analyses:
Poisson regression
Averaging Time: 24-h
Mean (SD) unit:
15.8 ppmxIO
Range (Min, Max):
3.20,48.4
Copollutant: NR
Increment: Interquartile range: 0.56 ppm
OR for Post-neonatal mortality [Lower Cl,
Cl]
Lag examined : 0-2
Lag 0-2: 1.038 (0.663-1 .624)
Upper
Age Groups Analyzed: NA
Sample Description: NR
Table C-4 Studies of short-term CO exposure and respiratory morbidity
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Andersen et al.
(2008, 096150)
Period of Study:
Dec1998-Dec2004
Location:
Copenhagen, Denmark
Health Outcome: Wheezing
symptoms
Study Design: Panel
Statistical Analyses: Logistic
regression (GEE)
Age Groups Analyzed: 0-3 yrs
Sample Description:
205 children of mothers with asthma
Averaging Time: 24h
Mean (SD) unit:
0.29 (0.10) ppm
Range (percentiles):
25th = 0.22; 75th = 0.34
Copollutant: correlation
PM10:r=0.45
PM2.5: r = 0.45
UFPNC:r = 0.52
N02: r = 0.75
NOX: r = 0.74
03:r=-0.63
Increment: NR
OR Estimate [Lower Cl, Upper Cl]; lag :
Lags examined: 0,1, 2, 3, 4,2-4
Lag 0:0.96 (0.80,1.15)
Lag 1:0.92 (0.77,1.10)
Lag 2:1.08 (0.92,1.28)
Lag 3:1.07 (0.90,1.26)
Lag 4:1.02 (0.84,1.23)
3d mean: 1.07 (0.87,1.32)
Author: Bhattacharyya et
al. (2009.180154)
Period of Study: 1997-
2006
Location: NR (National
Health Interview Survey as
aggregated in the
Integrated Health Interview
Series served as data
Health Outcome: Respiratory
morbidity
Study Design: Cross-sectional study
Statistical Analyses: SPSS version
14.0, univariate linear regression
analysis
Age Groups Analyzed: 18+ yr (avg:
45.2yr)
Sample Description: Hay fever,
weak/failing kidneys, sinusitis all in
past 12 m
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): 2.209-4.157ppm
(decreased with increasing yr)
Copollutant: NR
Increment: NR
Linear regression analysis for disease
condition prevalence: Hayfever: Standardized
B- 0.012, p-value- <0.001; Sinusitis:
Standardized B- 0.027, p-value- <0.001;
Kidney Weak/Failin: Standardized B- -0.001, p-
value- <0.001
Lags examined: NR
September 2009
C-32
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Chen etal. (1999,
011149)
Period of Study:
5/1995-1/1996
Location:
3 Taiwan communities
Author: Chen et al. (2000,
011931)
Period of Study:
8/1996-6/1998
Location:
Washoe County, NV
Design
Health Outcome: Lung function (FVC,
FEV1, FEV1/FVC, FEF25-75%, PEF)
Study Design: Cross-sectional survey
Statistical Analyses:
Multivariate linear model
Population:
941 children (Boys: 453; Girls: 488)
Age Groups Analyzed: 8-13 yr
Health Outcome: School absenteeism
Study Design: Time-series
Statistical Analyses: Maximum
likelihood
Population: 1st to 6th grade children:
27,793
Concentrations
Pollutant: CO
Averaging Time:
1-h max;24-h avg
Mean (SD) unit: NR
Range (Min, Max):
1-h max: (0.4, 3.6)
Copollutant correlation:
N02:r = 0.86 -0.98
Note: To represent the
schoolchildren's exposure the
daytime avg and peak concentrations
were measured from 0800 to 1800.
Pollutant: CO
Averaging Time: 1-h max
Mean (SD) unit:
2.73 (1.1 54) ppm
Range (Min, Max): (0.65, 2.73)
f^nnnlliitant rnrratatinn1
CO Effect Estimates (95% Cl)
Increment: NR
p Coefficient (SE); lag:
FVC (mL)
24-h avg
-66.6(40.73);!
-147.71 (64.48); 2
2.2 (48.13); 7
1 -h max
-33.25(20.74);!
-16.48 (19.67); 2
-5.18 (16.48); 7
FFV1 lm\ \
re v i ^IIILJ
24-h avg
20.55(38.24);!
-82. 42 (60. 95); 2
48. 23 (45. 58); 7
1 -h max
1.2(19.48);!
-1.44 (18.57); 2
20. 96 (15. 67); 7
Increment: 1.0 ppm
% Increase (Lower Cl, Upper Cl); lag:
3. 79% (1.04-6. 55) ;0
Age Groups Analyzed: 1st to 6th PM10:r= 0.721
grade children 03:r=-0.204
Author: de Hartog et al.
(2003, 001061)
Period of Study:
1998-1999
Location:
Amsterdam, Netherlands;
Erfurt, Germany;
Helsinki, Finland
Health Outcome: Respiratory
symptoms (shortness of breath, being
awakened by breathing problems,
phlegm, wheezing, tripping heart)
Study Design: Time-series
Statistical Analyses:
Logistic regression
Population:
Non-smoking individuals with CHD:
Amsterdam: 37
Erfurt: 47
Helsinki: 47
Age Groups Analyzed: 2 50 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfurt:0.4mg/m3
Helsinki: 0.4 mg/m3
Range (Min, Max):
Amsterdam: (0.4,1.6)
Erfurt: (0.1, 2.5)
Helsinki: (0.1,1.0)
Copollutant:
PM25;
NO,
Increment: 0.25 mg/m3
Odds Ratio (Lower Cl, Upper Cl); lag:
Incidence of symptoms
Shortness of breath
1 (0.92-1.1);0
0.96(0.88-1.05);!
1 (0.92-1.09); 2
1.07 (0.98-1.16); 3
1.03 (0.9-1.18); 0-4
Being awakened by breathing problems
1.02(0.92-1.14);!
1.03 (0.93-1.15); 2
1.11 (1-1.22);3
1.16 (0.98-1.37); 0-4
Phlegm
1.05 (0.93-1.19) ;0
1.02 (0.91-1.14) ;1
1.08 (0.96-1.22); 2
1.09 (0.97-1.22); 3
1.13 (0.94-1.35); 0-4
Prevalence of symptoms
Shortness of breath
1 (0.94-1.06); 0
0.99(0.94-1.05);!
0.99 (0.93-1.05); 2
1.01 (0.95-1.07); 3
0.98 (0.9-1.07); 0-4
Being awakened by breathing problems
1.01 (0.93-1.1);!
0.99 (0.91-1.08); 2
1.1 (1.02-1.19); 3
1.13 (1-1.29); 0-4
September 2009
C-33
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Delfino etal.
(2003, 050460)
Period of Study:
11/1999-1/2000
Location:
Los Angeles, CA
Author: Estrella et al.
(2005, 099124)
Period of Study:
1/2000-4/2000
Location:
Quito, Ecuador
Author: Fischer et al.
(2002, 025731)
Period of Study: NR
Location:
Utrecht, Netherlands
Author: Ho et al. (2007,
093265)
Period of Study:
Oct1995-Mar1996
Location:
Taipei, Taiwan
Design
Health Outcome:
Asthma symptoms (Cough, wheeze,
sputum production, shortness of
breath, chest tightness) (symptom
scores >1 , symptoms scores >2); Lung
function (PEF)
Study Design: Panel study
Statistical Analyses:
Asthma symptoms: GEE
Lung function: Generalized linear
IIIIACU IIIUUcI
Population:
22 asthmatic Hispanic children
Age Groups Analyzed: 1 0-1 Syr
Health Outcome:
Acute respiratory infection
Study Design: Prospective study
Statistical Analyses:
Logistic regression; Poisson
Population: 960 children
Age Groups Analyzed: 6-11 yr
Health Outcome:
Lung function (FVC, FEV,, PEF,
MMEF)
Study Design: Panel study
Statistical Analyses:
Restricted max likelihood linear model
Population: 68 children
Age Groups Analyzed: 10-11
Health Outcome: asthma
Study Design: panel
Statistical Analyses: Logistic
regression (GEE)
Age Groups Analyzed: 10-17yr
Sample Description:
a stratified cluster random sample of
students (n=69,367) from 1 ,139,452
students sampled nationwide
Concentrations
Pollutant: CO
Averaging Time:
1-h max;8-h max
Mean (SD) unit:
1-h max: 7.7 (3.1) ppb
8-h max: 5.0 (2.0) ppb
Range (Min, Max):
1-h max: (2, 17)
8-h max:(1, 10)
Copollutant correlation:
N02: r= 0.65; 03:r = -0.17;
Acetaldehyde:r= 0.51;
Acetone: r= 0.28;
Formaldehyde: r= 0.41;
Benzene: r = 0.50;
Ethylbenzene:r = 0.62;
Tetrachloroethylene: r = 0.63;
Toluene: r= 0.71;
m,p-Xylene:r = 0.72;
PM10:r=0.50;
EC: r = 0.60;
OC:r = 0.55;
S02:r=0.69
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 921 ug/m3
Range (Min, Max): (31 9, 1540)
Copoiiutcint!
PM10;BS;N02;NO
Averaging Time: 8h
Mean (SD) unit:
NR
Range (min, max):
NR
Copollutant:
NO, N02, NOX, 03, S02, PM10, PSI
CO Effect Estimates (95% Cl)
Increment: 5.0 ppb & 3.0 ppb
Odds Ratio (Lower Cl, Upper Cl); lag:
1-max
Increment: 5.0 ppb
Symptom scores >1
0.95 (0.52-1. 75) ;0
1.11 (0.75-1.65); 1
Symptom scores >2
0.48 (0.07-3.53) ;0
.28(0.53-3.12);!
8-h max
Increment: 3.0 ppb
Symptom scores >1
0.95 (0.55-1. 62) ;0
1.2(0.77-1.86);!
Symptom scores >2
0.53 (0.1 0-2.92) ;0
1.43 (0.41 -5.00) ;1
Increment: NR
Odds Ratio (Lower Cl, Upper Cl); lag:
Acute respiratory infection
ARI in children COHb >2.5% vs. COHb <2.5%:
Adjusted Logistic Regression Model
3.25(1.65-6.38)
ARI in children COHb >2.5% vs. COHb <2.5%:
Crude Logistic Regression Model
2.06(1.30-3.20)
Log-Linear Model (Each Percent Increase in
COHb above 2.5%)
1.15(1.03-1.28)
Increment: 100 ug/m3
mL(SE);lag:
FVC: 0.5 (0.4); 1; 0.1 (0.2); 2
FEV1:-0.4(0.5);1;-0.2(0.2);2
m/s (SE);lag:
PEF: -1.1 (2.8); 1; -0.6 (1.1); 2
MMEF: -0.5 (1. 4); 1; -0.3 (0.6); 2
Increment: very high, high, med, low, very low
OR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined: NR
Females: 1.984 (1.536, 2.561)
Males: 1.780 (1.377, 2.302)
Monthly attack rate vs. Single Air Pollutant
cone.
Estimate (p-value):
0.0750 (0.3336)
September 2009
C-34
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Lagorio et al.
(2006, 089800)
Period of Study:
5/1999-6/1999;
11/1999-12/1999
Location:
Rome, Italy
Author: Moonetal. (2009,
190297)
Period of Study:
Apr 2003-May 2003
Location:
Seoul, lncheon,Busan,&
Jeju, Korea
Author: Mortimer et al.
(2008, 187280)
Period of Study:
Nov 2000-Apr 2005
Location:
Fresno, California
Design
Health Outcome:
Lung function (FVC.FEV1)
Study Design:
Time-series panel study
Statistical Analyses:
Generalized estimating equations
(GEE)
Population:
COPD panel: 11
Asthma panel: 11
IHD panel: 7
Age Groups Analyzed:
COPD panel: 50-80 yr
Asthma panel :1 8-64 yr
IHDpanel: 40-64 yr
Notes: Asthma panel was restricted to
never smokers, while COPD and IHD
panels include former smokers if
smoking cessation occurred at least
1 yr prior to enrollment.
Health Outcome: respiratory
symptoms
Study Design: panel
Statistical Analyses: Logistic
regression (GEE)
Age Groups Analyzed: < 13yr
Sample Description:
696 children
Health Outcome: allergic sensitization
Study Design: panel
Statistical Analyses: Multi-step
modeling
Age Groups Analyzed: 6-11 yr
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Overall: 7.4 (6.2) mg/m Spring: 2.1
(0.3) mg/m3 Winter: 12.3 (4.9) mg/m3
Range (Min, Max):
Overall: (1.6, 28.9)
Copollutant correlation:
PM25:r = 0.67
PM10-2.5:r=-0.09
PM • r - n ^
r IVI10- ' u.vjo
N02: r = 0.05
03:r=-0.87
S02:r=0.65
Averaging Time: 24h
Mean (SD) unit:
NR
IQ Range:
0.12ppm
Copollutant:
PM10, S02, N02, 03
Averaging Time: 24h avg, 24h max,
8h max
Mean (SD) unit:
NR
IQ Range (24h avg, 24h max, 8h
max):
CO Effect Estimates (95% Cl)
Increment: 1 mg/m3
p Coefficient (SE); lag:
COPD panel
FVC (% of predicted)
-0.14 (0.15) ;0
-0.13 (0.18); 0-1
0.15 (0.23); 0-2
FEV1 (% of predicted)
-0.05 (0.13); 0
-0.12 (0.16); 0-1
-0.03 (0.2); 0-2
Asthma panel
FVC (% predicted)
0.02 (0.12); 0
-0.001 (0.13); 0-1
-0.06 (0.16); 0-2
FEV1 (% predicted)
-0.05 (0.14); 0
-0.16 (0.15); 0-1
-0.28 (0.18); 0-2
IHD panel
FVC (% of predicted)
0.176 (0.101);0
0.132(0.120);0-1/
0.1 32 (0.1 65); 0-2
FEV1 (% of predicted)
0.204 (0.1 20) ;0
0.114 (0.142); 0-1
0.1 59 (0.1 94); 0-2
Increment: 0.12 ppm (IQR)
OR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined: lag days 0-3
Lower resp. symptoms: 1.005 (1.003, 1.008),
lagO
Upper resp. symptoms: 1.006 (1.003, 1.008),
lag 0-2
Irritation symptoms : 1 .004 (1 .001 , 1 .006) , lag
1-3
Increment: IQR
OR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined: NR
Fntire Prennancv
Sample Description: 0.28,0.79,0.52
170 children with physician diagnosed Copollutant: entire prenatal
asthma correlation
N02:r = 0.74
03:r=-0.40
PM10:r=0.32
CO 24h avg: 1.45 (1.02, 2.07)
CO 24h max: 1.53 (1.01, 2.33)
CO 24h avg: 1.55 (1.01, 2.37)
September 2009
C-35
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Nkwocha etal.
(2008, 190304)
Period of Study:
Feb 2005 -Jul 2006
Location:
Port Harcourt, Nigeria
Health Outcome: respiratory
symptoms
Study Design: panel
Statistical Analyses: Mixed Effects
models
Ana f*rnnnc An^lx^pH1 0-fi \/r
Averaging Time: 8h
Mean (SD) unit:
NR
Range (min, max):
1 .3 |jg/m3, 1 .83 |jg/m3
ftnnnllnfeint1 NO- SO- PMJn
Increment: NR
Lags examined: NR
R Estimate:
Dry season: 0.13
Author: O'Connor etal.
(2008, 156818)
Period of Study:
Aug1998-Jul2001
Location:
Boston, MA; the Bronx,
NY; Chicago, IL; Dallas,
TX; New York, NY; Seattle,
WA;Tuscon,AZ
Sample Description:
250 children
Health Outcome: respiratory
symptoms
Study Design: panel
Statistical Analyses: Mixed Effects
Models
Age Groups Analyzed: 5-12 yr
Sample Description:
Rfi1 rhilHrpn with nprcictpnt acthma
Averaging Time: 8h
Mean (SD) unit:
NR
Range (10th-90th):
872.1 ppb
Copollutant:
PM10, S02, N02, 03
Wet season: 0.25
Increment: 872.1 ppb
Lags examined: NR
Change Estimate [Lower Cl, Upper Cl]:
FEV1: -0.56 (-1.31, 0.20)
PEFR: -0.49 (-1.24, 0.27)
and atopy living in low-income census
tracts
Pollution lmpact*[Lower Cl, Upper Cl]:
Wheeze-cough: 1.26 (1.03,1.55)
Nighttime asthma: 1.35 (1.07,1.71)
Slow play: 1.28 (1.04,1.59)
OR [Lower Cl, Upper Cl]:
Missed School: 1.08 (0.76,1.53)
'coefficients from the negative binomial model
and indicate the multiplicative effect per unit
change
Author: Park etal. (2002,
093798)
Period of Study:
3/1996-12/1999
Location:
Seoul, Korea
Health Outcome: School absenteeism
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Population:
~ 1,264 children (671 Boys, 593 girls)
Age Groups Analyzed:
1st through 6th grade students
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.11 (0.40) ppm
Range (Min, Max): (0.39, 2.97)
Copollutant correlation:
PM10:r=0.56;
N02:r=0.70;
S02:r = 0.67;
03:r=-0.46
Increment: 0.52 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Total Absences:
0.95 (0.94-0.97);0
Non-Illness Related Absences:
0.99 (0.96-1.02);0
Illness-Related Absences:
0.96 (0.94-0.98);0
Author: Park etal. (2005,
088673)
Period of Study:
3/2002-6/2002
Location:
Incheon, Korea
Health Outcome:
Lung function (PEF variability (>20%),
Mean PEF); Respiratory symptoms
(night respiratory symptoms, cough,
inhaler use)
Study Design: Panel study
Statistical Analyses:
GEE; Poisson GAM
Population: 64 bronchial asthmatics
Age Groups Analyzed: 16-75 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Control days: 0.6368 (0.1522) ppm
Dust days: 0.6462 (0.0945) ppm
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Relative Risk (Lower Cl, Upper Cl); lag:
PEF variability (>20%): 1.43 (0.54-3.75)
Night respiratory symptoms:
0.98(0.51-1.86)
P Coefficient (SE); lag:
PEF variability (>20%): 0.9737 (0.3187)
Mean PEF (L/min):-10.103 (2.7146)
Night respiratory symptoms:
-0.018(0.3654)
Cough: 0.0855 (0.1826)
Inhaler Use: 0.0796 (0.1733)
September 2009
C-36
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Penttinen et al. Health Outcome:
Period of Study:
11/1996-4/1997
Location:
Helsinki, Finland
Lung function (PEF)
Study Design: Panel study
Statistical Analyses:
First order autoregressive linear model
Population:
57 non-smoking adult asthmatics
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time: 24-h avg
Median unit: 0.4 mg/m3
Range (Min, Max):
(0.1,1.1) mg/m3
Copollutant correlation:
PM10:r=-0.03
PM10-2.5:r = -0.30
PM25:r=0.32
PM1:r=0.39
PNC: r = 0.44
NC0.01-0.1:r = 0.43
NC0.1-1:r = 0.47
NO: r = 0.60
N02: r = 0.44
Increment: 0.2 mg/m3
p Coefficient (SE); lag:
PEF Deviations (L/min)
Morning
0.27 (0.38); 0
-1.08(0.36);!
0.23 (0.38); 2
-1.11 (1.19); 5-day avg
Afternoon
-0.4(0.43);0
-0.13(0.41);!
-0.71 (0.41);2
-3.03 J1.06); 5-day avg
Evening
-0.7 (0.45);0;
-0.31 (0.44); 1
0.3 (0.44); 2
-3.62 (1.19); 5-day avg
Co-pollutant models with PNC
Morning:-0.67(0.64);!
Afternoon:-0.46 (0.69); 0
Evening:-0.46 (0.73); 0
Author: Rabinovitch et al.
(2004, 096753)
Period of Study:
11/1999-3/2000;
11/2000-3/2001;
11/2001-3/2002
Location:
Denver, CO
Health Outcome:
Lung function (FEV1); asthma
exacerbation; bronchodilator use
Study Design: Panel study
Statistical Analyses:
Pulmonary function: Mixed effects
model; Asthma exacerbation and
medication use:GLM
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .0 (0.4) ppm
Range (Min, Max): (0.3, 3.5)
Copollutant:
PM2.5;PM10;N02;S02;03
Increment: 0.4 ppm
P Coefficient (SE); lag: FEV1
AM: -0.001 (0.008); 3-day ma
PM: 0.015 (0.01); 3-day ma
Odds Ratio (Lower Cl, Upper Cl); lag:
Asthma exacerbation:
1.012 (0.91 3-1 .123); 3-day ma
Population:
Urban poor asthmatic children:
1999-2000:41
2000-2001:63
2001-2002:43
Age Groups Analyzed: 6-12 yr
Bronchodilator use:
1.065 (1.001-1.133); 3-day ma
Author: Ranzi et al. (2004,
089500)
Period of Study:
2/1999-5/1999
Location:
Emilia-Romagna, Italy
Health Outcome:
Lung function; respiratory symptoms,
medication use
Study Design: Panel study
Statistical Analyses: GLM
Population:
120 "asthma-like" school children
Age Groups Analyzed: 6-11 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Urban area: 1.54 mg
Rural area: 1.22 mg
Range (Min, Max): NR
Copollutant: N02; TSP; PM25
The study did not present quantitative results
for CO.
Author: Rodriguez et al.
(2007, 092842)
Period of Study:
1996-2003
Location:
Perth, Australia
Health Outcome:
Respiratory symptoms (body
temperature, cough, wheeze/rattle
chest, runny/blocked nose)
Study Design: Panel study
Statistical Analyses:
Logistic regression, GEE
Population:
263 children at high risk of developing
asthma
Age Groups Analyzed: 0-5 yr
Pollutant: CO
Averaging Time: 8-h avg
Mean (SD) unit: 1 .408 ppm
Range (Min, Max): (0.012, 8.031)
Copollutant: NR
Increment: NR
Odds Ratio (Lower Cl, Upper Cl); lag:
Body Temperature
1.024 (0.911-1.151); 0
1.056 (0.943-1 .184); 5
0.991 (0.962-1 .021); 0-5
Cough
1.001 (0. 996-1. 005) ;0
1.064 (0.941 -1.02); 5
1.028 (0.996-1 .061); 0-5
Wheeze/Rattle Chest
1.089 (0.968-1. 226) ;0
1.1 36 (1.01 6-1 .26); 5
1.035 (1.005-1 .066); 0-5
Runny/Blocked Nose
1. 094 (0.824-1. 453) ;0
1.38 (1.028-1 .853); 5
1.101 (1.025-1 .183); 0-5
September 2009
C-37
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Schildcrout et al.
(2006, 089812)
Period of Study:
11/1993-9/1995
Location:
8 North American cities:
Albuquerque, NM
Baltimore, MD
Boston, MA
Denver, CO
San Diego, CA
Seattle, WA
St. Louis, MO
Toronto, ON, Canada
Health Outcome:
Asthma symptoms; rescue inhaler use
Study Design: Panel study
Statistical Analyses:
Asthma symptoms: Logistic
regression; Rescue Inhaler Use:
Poisson regression
Population:
990 asthmatic children
Age Groups Analyzed: 5-12 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
N02;03;PM10;S02
Increment: 1.0 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Asthma Symptoms
1.08(1.01-1.14);0
1.07(0.99-1.16);!
1.08 (1.02-1.15); 2
1.05 (1.01-1.09); 0-2
Asthma Symptoms
+ 20 ppb increase in N02
1.07(1-1.14);0
1.04(0.96-1.11);!
1.09 (1.02-1.16); 2
1.04 (1-1.08); 0-2
+ 25 ug/m3 increase in PM10
1.08(1.01-1.15);0
1.06(0.99-1.14);!
1.08 (1.02-1.14); 2
1.05 (1.011-1.08); 0-2
+ 10 ppb increase in S02
1.07 (0.99-1.16) ;0
1.06 (0.96-1.19) ;1
1.1 (1.02-1.18); 2
1.05 (1-1.09); 0-2
Rescue Inhaler Use
1.07(1.01-1.13);0
1.05(0.99-1.1);!
1.06(1.01-1.1);2
1.04 (1.01-1.07); 0-2
Rescue Inhaler Use
+ 20 ppb increase in N02
1.05 (0.99-1.12) ;0
1.04(0.98-1.11);!
1.07 (1.02-1.12); 2
1.04 (1-1.07); 0-2
+ 25 ug/m3 increase in PM10
1.06 (0.99-1.13) ;0
1.05(0.99-1.11);!
1.05 (1.01-1.09); 2
1.03 (1-1.07); 0-2
+ 10 ppb increase in S02
1.04 (0.96-1.12) ;0
1.04(0.97-1.1);!
1.08 (1.03-1.13); 2
1.04 (1-1.08); 0-2
Author: Silkoffetal.
(2005, 087471)
Period of Study:
11/11/1999-3/31/2000;
11/1/2000-3/16/2001
Location:
Denver, CO
Health Outcome:
Lung function (FEV1 , PEF); recorded
symptoms; rescue medication use
Study Design: Panel study
Statistical Analyses:
Rescue medication use and total
symptom score: GEE;
Lung function: Mixed effects model
Population:
1 st winter: 1 6 with a history of more
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
1999-2000: 1.2 (0.555) ppm
2000-2001:1.1 (0.5) ppm
Range (Min, Max):
1999-2000: (0.340, 3.790)
2000-2001: (0.360, 2.810)
Copollutant: NR
The study did not present quantitative results
for CO.
than 10 pack yr of tobacco use, airflow
limitation with FEV1 of less than 70%
of predicted value, and FEV1/ FVC
ratio of less than 60%
2nd winter: 18 with a history of more
than 10 pack yr of tobacco use, airflow
limitation with FEV1 of less than 70%
of predicted value, and FEV1/ FVC
ratio of less than 60%
Age Groups Analyzed: > 40 yr
September 2009
C-38
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Slaughter etal.
(2003, 086294)
Period of Study:
12/1994-8/1995
Location:
Seattle, WA
Author: Steerenberg et al.
(2001.017157)
Period of Study: NR
Location:
Bilthoven and Utrecht,
the Netherlands
Author: Timonen et al.
(2002, 025653)
Period of Study:
2/1994-4/1994
Location:
Kuopio, Finland
Design
Health Outcome:
Asthma severity; medication use
Study Design: Panel study
Statistical Analyses:
Asthma severity: Ordinal logistic
regression; Medication use: Poisson
Population: 133 mild-to-moderate
asthmatic children
Age Groups Analyzed: 5-13
Health Outcome:
Lung function (PEF); exhaled nitric
oxide; inflammatory nasal markers
Study Design: Panel study
Statistical Analyses:
Restricted max likelihood linear model
Population: 126 children
Age Groups Analyzed: 8-13 yr
Notes : The study was only conducted
for a two mo period: February and
March.
Health Outcome:
Exercise induced bronchial
responsiveness; Lung function (FVC,
FEV1,MMEF,AEFV)
Study Design: Panel study
Statistical Analyses: Linear
regression
Population:
33 children with chronic respiratory
symptoms
Age Groups Analyzed: 7-12 yr
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Median unit: 1 .47 ppm
IQR (25th, 75th): (0.23, 1.87)
Copollutant: NR
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Utrecht :0. 8 mg/m3
Bilthoven: 0.5 mg/m3
Range (Min, Max):
Utrecht: (0.3, 2.3)
Bilthoven: (0.3, 0.9)
Copollutant: NR
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.6 mg/m3
Range (Min, Max): (0.1, 2.8)
Copollutant correlation:
PM10:r=0.52
BS:r = 0.80
PNC0.01-0.03:r = 0.81
PNC0.03-0.1:r=0.87
PNC0.1-0.3:r=0.71
PNC0.3-1.0:r=0.60
PNC1. 0-3.2: r= 0.84
PNC3.2-10:r=0.79
N02: r = 0.85
CO Effect Estimates (95% Cl)
Increment:
Increased asthma attack severity: 0.67 ppm
Increased rescue inhaler use: 1 .0 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Increased asthma attack severity:
Without transition :1. 21 ;1
With transition:!. 17; 1
Increased rescue inhaler use:
Without transition : 1 .09 (1 .03-1 .1 6) ; 1
With transition: 1.06 (1.01-1.1);!
The study did not present quantitative results
for CO.
Increment: 0.32 mg/m3
P Coefficient (SE); lag:
Exercise induced responsiveness
AFEV1 (%) FEV1 (mL)
-0.081 (0.647) ;0 19.2 (13.2); 0
0.03(0.262);! -9.04(5.45);!
0.087 (0.26); 2 -9.15 (5.21); 2
-0.091 (0.275); 3 -11.7 (5.77); 3
0.19 (0.599); 0-3 -17.5 (12.5); 0-3
AMMEF (%) MMEF (mL/s)
0. 442(1. 79) ;0 22. 2 (36. 9) ;0
0.52(0.723);! -23(15.2);!
0.313 (0.719); 2 -4.63 (14.7); 2
-0.616 (0.75); 3 -30.9 (16); 3
0.096 (1.64); 0-3 -24.9 (34.8); 0-3
AAEFV (%) AEFV (L2/s)
0.287 (1. 19); 0 -0.093 (0.088) ;0
0.281(0.482);! -0.068(0.036);!
0.904 (0.474); 2 -0.06 (0.035); 2
0.15 (0.483); 3 -0.05 (0.039); 3
1.6 (1.05); 0-3 -0.076 (0.083); 0-3
FVC (mL)
0.064(10.9);0
-4.79(4.51);!
-9.78 (4 .24); 2
-13.9 (4.7); 3
-29.4 (10.1); 0-3
September 2009
C-39
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: von Klotetal.
(2002, 034706)
Period of Study:
9/1996-3/1997
Location:
Erfurt, Germany
Design
Health Outcome:
Asthma symptoms; medication use
Study Design: Panel study
Statistical Analyses:
Logistic regression
Population:
53 adults with asthma or asthma
symptoms
Age Groups Analyzed: 37-77 yr
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.9 mg/m3
Range (Min, Max): (0.3, 3.0)
Copollutant correlation:
NC0.01-0.1:r = 0.66
NC0.1-0.5:r=0.79
NC0.5-2.5:r=0.46
MC0.1-0.5:r=0.66
MC0.01-2.5:r = 0.65
PM ' r ~ fi 49
rlVl2.5-lo- 1 U.4£
PM • r- n fin
r IVI1Q. I U.UU
NOi' r — 0 82_
2' ~ '
CO Effect Estimates (95% Cl)
Increment:
0 and 5-day avg lag: 0.6 mg/m3
14-day avg lag: 0.54 mg/m
Odds Ratio (Lower Cl, Upper Cl); lag:
Prevalence: Inhaled p2-agonist use
0.98 (0.93-1. 03) ;0
1.04 (0.97-1 .12); 0-4
0.93 (0.86-1 .01); 0-1 3
Prevalence: Inhaled corticosteroid use
125 (1.1 7-1 .34); 0-4
1.06 (0.97-1 .15); 0-1 3
Prevalence: Wheezing
1. 03 (0.97-1. 08) ;0
1.1 3 (1.05-1 .22); 0-4
1.1 4 (1.05-1 .25); 0-1 3
Co-pollutant models
Inhaled (32-agonist use
CO+MCO.01-2.5:
1 (0.91-1.11); 0-4
CO+NCOO.01-0.1:
1.01 (0.91-1 .11); 0-4
Inhaled corticosteroid use
CO+MCO.01-2.5:
0.89 (0.81 -0.98); 0-1 3
CO+NC: 0.01-0.
1:0.81 (0.72-0.91); 0-13
Wheezing
CO+MCO.01-2.5:
1.1 5 (1.04-1 .27); 0-4
CO+NCO.01-0.1:
1.09 (0.98-1 .22); 0-4
Author: Yuetal. (2000,
013254)
Period of Study:
11/1993-8/1995
Location:
Seattle, Washington
Health Outcome: Pollutant: CO
Asthma symptoms (Wheezing,
coughing, chest tightness, shortness of Averaging Time: 24-h avg
breath)
Study Design: Panel study
Mean (SD) unit: 1.6 ppm
Range (Min, Max): (0.65, 4.18)
R^a'ted'meTs'ure^iogistic regression ™Tnlutann'correlation:
mnrlpk (Gff\ PM1.U. r-U.8«>
moaeis (btt) PM10:r=0.86
Population: S02:r=0.31
133 mild-to-moderate asthmatics
Age Groups Analyzed: 5-13 yr
Increment: 1.0 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Marginal GEE
1.22(1.03-1.45);0
1.3(1.11-1.52);!
1.26 (1.09-1.46); 2
Transition GEE
1.18(1.02-1.37);0
1.25(1.1-1.42);!
1.18 (1.04-1.33); 2
September 2009
C-40
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-5 Studies of short-term CO exposure and respiratory hospital admissions and ED visits.
Study
Author: Abe etal.
(2009, 190536)
Period of Study:
January 1 -December
31,2005
Location: Tokyo,
Japan
Design
ED Visits
Health Outcome: Asthma
Study Design: Time-series
Statistical Analyses: Bivariate
Pearson correlation
coefficitnes, ARIMA model
Concentrations
Averaging Time: NR
Mean (SD) unit: 11 .5ppm
Range (Min, Max): 3-44ppm
Copollutant: NR
Effect Estimates (95% Cl)
lncrement:0.1ppm
ARIMA model for ambulance transports to ED for asthma
exacerbation among adults: p coefficient: 0.1 51, SE:
0.098, t statistic: 1 .537, P value: .125
ARIMA model for ambulance transports to ED for asthma
exacerbation among children: p coefficient: 0.019, SE:
0.034, t statistic: 0.549, P value: 0.583
Age Groups Analyzed:
Children: <14yr, Adults: < 15 yr
Sample Description: Data
from daily number of
ambulance transports to ED for
asthma
Lags examined :0
On the day with the highest CO the number of transports
was 25. The number of transports for adults and CO had
significant bivariate correlations. The fitted ARIMA model
had no significant associations.
Author: Anderson et
al. (2001 , 017033)
Period of Study:
10/1994-12/1996
Location:
West Midlands; U.K.
Author: Andersen et
al. (2007.093201)
Period of Study:
1/1999-12/2004
Location:
Copenhagen,
Denmark
Hospital Admission
Health Outcome (ICD9):
Respiratory Diseases
Asthma (493)
COPD (490-492, 494-496)
Study Design: Time-series
Statistical Analyses:
Regression with quasi-
likelihood approach and GAM
Age Groups Analyzed:
All ages
0-1 4 yr
15-64yr
>65yr
Hospital Admission
Health Outcome (ICD10):
Respiratory diseases: Chronic
bronchitis (J41 -42),
Emphysema (J43), COPD
(J44), Asthma (J45), Status
asthmaticus (J46), Pediatric
asthma (J45), Pediatric
asthmaticus (J46)
Study Design: Time-series
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 0.8 (0.7) ppm
Range (Min, Max): (0.2, 10)
CoPollutant: correlation
PM10:r = 0.55;
PM2.5: r = 0.54;
PM25-io: r = 0.10;
BS:r = 0.77;
S042-:r=0.17;
N02:r = 0.73;
03:r = -0.29;
S02: r = 0.49
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.3 (0.1) ppm
IQR (25th, 75th): (0.22, 0.34)
Copollutant; correlation:
PM10:r = 0.45
Increment: 1.0 ppm
% Increase (Lower Cl, Upper Cl); lag:
Respiratory Diseases
Age Group
All ages: 0.3% (-1.10 to 1.70); 0-1
0-14: 1.50% (-0.60 to 3.60); 0-1
15-64: -0.70% (-3.60 to 2.30); 0-1
>65: 0.00% (-2.10 to 2.10); 0-1
Asthma
Age Group
0-14: 3.90% (-0.50 to 8.50); 0-1
15-64: -4.90% (-10.60 to 1.10); 0-1
COPD
Age Group
>65: 1.00% (-2.50 to 4.60); 0-1
Increment: 0.1 2 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Respiratory Disease
Age Group: > 65
CO: 1.024 (0.997-1 .053); 0-4
CO, PM10: 1.001 (0.961 -1.042); 0-4
Asinms
Age Group: 5-1 8
CO: 1.1 04 (1.01 8-1 .198); 0-5
CO, PM10: 1.023 (0.911-1 .149); 0-5
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: 5-18
yr;>65yr
September 2009
C-41
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Period of Study:
1/1992-12/1994
Location:
London,
U.K.
Author: Atkinson etal. ED Visits
(1999 007882)
v Health Outcome (ICD9):
Respiratory complaints:
wheezing, inhaler request,
chest infection, chronic
obstructive lung disease
(COLD), difficulty breathing,
cough, other respiratory
complaints, e.g., croup,
pleurisy, noisy breathing;
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
0-14 yr
15-64yr
>65yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.4) ppm
Range (Min, Max): (0.2, 5.6)
Copollutant; correlation:
N02
03
S02
PM10
BS
Increment: 0.8 ppm
% Increase (Lower Cl, Upper Cl); lag:
Respiratory complaints
Age Group
All ages: 0.76% (-0.83, 2.38); 1
0-14:2.92% (0.60,5.30);!
15-64:2.15% (-0.27,4.63);!
> 65:4.29% (1.15, 7.54); 0
Asthma visits:
Single-pollutant model
Age Group:
All ages: 3.32% (0.56,6.16);!
0-14:4.13% (-0.11, 8.54);0
15-64:4.41% (0.46,8.52);!
Multi-pollutant model
Age Group:
0-14
CO, N02:2.05% (-2.25, 6.54); 0
CO, 03:4.48% (0, 9.16); 0
CO, S02:2.34% (-1.94,6.81); 0
CO, PM10:2.93% (-1.53, 7.58); 0
CO, BS: 4.19% (-0.04,8.60); 0
Author: Bedeschi et
al. (2007,090712)
Period of Study:
1/2001-3/2002
Location:
Reggio Emilia,
Italy
ED Visits
Health Outcome (ICD9):
Asthma (493); Asthma-like
disorders, i.e., asthma,
bronchiolitis,
dyspnea/shortness of breath;
Other respiratory disorders
(i.e..upperand lower
respiratory illness including
sinusitis, bronchitis, and
pneumonia)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, penalized
splines
Age Groups Analyzed: <15yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.4 (0.7) mg/m3
Range (Min, Max): (0.4, 4.6)
Copollutant; correlation:
PM10:r = 0.61
TSP:r = 0.61
S02:r = 0.71
N02:r=0.77
The study did not provide quantitative results for CO.
September 2009
C-42
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Bell etal.
(2008, 091268)
Period of Study:
1/1995-12/2002
Location:
Taipei, Taiwan
Design
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (486); Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SE) unit: 0.9 ppm
Range (Min, Max): (0.3, 3.6)
CoPollutant: NR
Effect Estimates (95% Cl)
Increment: 0.5 ppm
% Increase (Lower Cl, Upper Cl); lag
Asthma (avg correlation between monitor pairs
monitors))
3. 29% (-0.74 to 7. 49) ;0
.49% (-4.25 to 3.41) ;1
-0.84% (-4.43 to 2.88); 2
0.48% (-4.02 to 3.18); 3
0.74% (-4.62 to 6.4): 0-3
= 0.75(13
Pneumonia (avg correlation between monitor pairs = 0.75
(13 monitors))
1.91% (-1.97 to 5.95); 0
0.03% (-3.65 to 3.85);1
0.36% (-3.2 to 4.04); 2
-1.29% (-4.77 to 2.32); 3
0.21% (-5.03 to 5.73); 0-3
Asthma (avg correlation between monitor pairs = 0.88 (5
monitors))
1.68% (-1.68 to 5.15); 0
-1.19% (-4.29 to 2.01); 1
-0.83% (-3.83 to 2.26); 2
-0.35% (-3.32 to 2.71); 3
-0.31% (-4.9 to 4.5); 0-3
Pneumonia (avg correlation between monitor pairs = 0.88
(5 monitors))
1.24% (-2.02 to 4.6);0
-0.01% (-3.06 to 3.13); 1
0.57% (-2.4 to 3.62); 2
-0.85% (-3.78 to 2.16); 3
0.31% (-4.23 to 5.06); 0-3
Asthma (monitors with 2 0.75 between monitor
correlations (11 monitors), avg correlation between
monitor pairs = 0.81)
2.87% (-0.91 to 6.79); 0
-0.71% (-4.2 to 2.91); 1
-0.73% (-4.08 to 2.73); 2
-0.41% (-3.72 to 3.01); 3
0.51% (-4.6 to 5.89); 0-3
Pneumonia (monitors with > 0.75 between monitor
correlations (11 monitors) to avg correlation between
monitor pairs = 0.81)
0.98% (-1.68 to 5.76);0
-0.12% (-3.54to 3.42); 1
0.37% (-2.95 to 3.8); 2
-1.08% (-4.34 to 2.3); 3
0.3% (-4.71 to 5.57); 0-3
Author: Bellini et al.
(2007, 097787)
Period of Study:
1996-2002
Location:
15 Italian cities
Hospital Admissions
Health Outcome:
Respiratory Conditions
Study Design:
Time-series; Meta-analysis
Statistical Analyses:
1. GLM for city-specific
estimates
2. Bayesian random-effects for
meta analysis
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
CoPollutant: correlation NR
Increment: 1 mg/m
% Increase (Lower Cl, Upper Cl); Lag
Respiratory conditions
All ages:
Season:
Winter: 0.58%; 0-1
Summer: 3.47%; 0-1
All Season: 1.25%; 0-3
Note: Estimates from Biggeri et al. (2004)
September 2009
C-43
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Braga et al.
(2001.016275)
Period of Study:
1/1993-11/1997
Location:
Sao Paulo, Brazil
Author: Burnett et al.
(1999.017269)
Period of Study:
1/1980-12/1994
Location:
Toronto, ON,
Canada
Author: Burnett et al.
(2001.093439)
Period of Study:
1/1980-12/1994
Location:
Toronto, ON,
Canada
Author: Cakmak et al.
(2006, 093272)
Period of Study:
4/1993-3/2000
Location:
10 Canadian cities
Design
Hospital Admissions
Health Outcome (ICD9):
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
£ 2yr
3-5 yr
6-1 Syr
14-1 9 yr
0-1 9 yr
Hospital Admissions
Health Outcome (ICD9):
Asthma (493); COPD (490-492,
496);
Respiratory infection
(464, 466, 480-487, 494)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, LOESS
Age Groups Analyzed: All
ages
Hospital Admissions
Health Outcome (ICD9):
Asthma (493); Acute bronchitis/
bronchiolitis (466); Croup
(464.4) ; Pneumonia (480-486)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed: <2 yr
Hospital Admissions
Health Outcome (ICD9):
Actue bronchitis/bronchiolitis
(466); Pneumonia (480-486);
Chronic/ unspecified bronchitis
(490, 491); Emphysema (492);
Asthma (493); Bronchiectasis
(494); Chronic airway
obstruction (496)
Study Design: Time-series
Concentrations
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 4.8 (2.3) ppm
Range (Min, Max): (0.6, 19.1)
CoPollutant:correlationCopollutant:
correlation
PM10:r = 0.60
03: r = -0.07
S02: r = 0.47
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.18 ppm
IQR (25th, 75th): (0.9, 1.4)
CoPollutant: correlation
PM • r - 0 4Q
r IVI2 5- 1 — U.'rc?
PMio-2.5:r = 0.20
PMi0:r = 0.43
MO • r - 0 55
INW2- 1 ~ U.vJvJ
S02: r = 0.37
03: r = -0.23
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 1.9 ppm
IQR (25th, 75th): (1.3, 2.3)
CoPollutant: correlation
03:r = 0.24
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 ppm
Range (Min, Max): (0.0, 6.5)
CoPollutant: correlation
S02
N02
03
Effect Estimates (95% Cl)
Increment: 3 ppm
% Increase (Lower Cl, Upper Cl); lag:
Respiratory
Age Group:
< 2: 5.00% (3.30-6.80); 0-6
3-5: 4.90% (1.40-8.50); 0-6
6-13: 1.00% (-2.50 to4.60); 0-6
14-19: 11. 30% (5.90-16.80); 0-6
0-19: 4.90% (3.50-6.40); 0-6
Increment: 1.18 ppm
% Increase (t-value); lag:
Asthma: 5.35% (3.92); 0
COPD: 2.93% (1. 48) ;0
Respiratory Infection: 5.00% (4.25); 0
Asthma:
Multi-pollutant model
CO, S02, 03: 5.15%
CO PM25 S02 03' 463%
CO, PM10-2.5, S02, 03: 5.25%
CO, PM10, S02, 03: 4.80%
CO, PM10-2.5, 03: 4.00%
COPD:
Multi-pollutant model
CO, S02, 03: 3.02%
CO,PM25,S02,03: 2.46%
CO, PM10-2.5, S02, 03: 3.00%
CO, PM10, S02, 03: 2.75%
CO, PM10-2.5, 03: 3.00%
Increment: 1.9 ppm
% Increase (Lower Cl, Upper Cl); lag
Respiratory problems
CO: 19. 20%; 0-1
CO, 03: 14.30%; 0-1
Increment: 0.8 ppm
% Increase (Lower Cl, Upper Cl); lag:
Respiratory disease
CO: 0.60% (0.20,1); 2.8
CO, S02, N02, 03: -0.20% (-0.70- 0.30); 2.8
Statistical Analyses:
1. Poisson
2. Restricted Maximum
Likelihood Method
Age Groups Analyzed: All
ages
September 2009
C-44
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Cheng etal.
(2007, 093034)
Period of Study:
1996-2004
Location:
Kaohsiung, Taiwan
Design
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-486)
Study Design: Bi-directional
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.76 ppm
Range (Min, Max): (0.1 4, 1.72)
CoPollutant: correlation
PM10
S02
N02
03
Effect Estimates (95% Cl)
Increment: 0.31 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
OR for pneumonia and exposure to various pollutants for
all ages in areas > 25°C or <25°C
Pollutant and Temperature
CO, > 25 °C: 1.1 8 (1.1 4-1 .23); 0-2
CO, <25°C: 1.47 (1.41-1 .53); 0-2
CO,PM10,>25°C:1.15(1.11-1.2);0-2
CO,PM10,<25°C:1.28(1.21-1.35);0-2
CO, S02, > 25 °C: 1.22 (1.17-1.27); 0-2
CO, S02, <25 °C: 1.49 (1.42-1.56); 0-2
CO, N02, a 25 °C: 1.2 (1.15-1.27); 0-2
CO, N02, <25 °C: 1.01 (0.95-1.08); 0-2
C0,03, a 25 °C: 1.16 (1.12-1.2); 0-2
C0,03, <25°C: 1.44 (1.38-1.5); 0-2
Author: Chiu etal.
(2009, 190249)
Period of Study:
1996-2004
Location:
Taipei, Taiwan
Hospital Admissions Averaging Time: 24h
Health Outcome: pneumonia Mean (SD) unit:
HA 1 .26 ppm
Study Design: case-crossover Range (min, max):
012 366
Statistical Analyses:
Increment: 0.57 ppm (IQR)
OR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined: one wk before to one wk after
CO:
Age Groups Analyzed: All
ages
Sample Description:
152,594 HA for 47 hospitals in
Taipei city
PM10:r = 0.34
S02: r = 0.57
N02:r=0.69
03:r = -0.31
>23°C: 1.25 (1.21,1.29)
<23°C: 1.12 (1.09,1.15)
CO + PM10:
>23°C:1.23(1.19,1.27)
<23°C: 1.05 (1.02,1.09)
CO + S02:
>23°C: 1.25 (1.21,1.30)
<23°C: 1.27 (1.22,1.31)
C0+N02:
>23°C: 0.97 (0.93,1.02)
<23°C:1.14(1.09,1.20)
C0 + 03:
>23°C: 1.24 (1.20,1.28)
<23°C:1.21 (1.17,1.24)
September 2009
C-45
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Cho et al.
(2000, 099051)
Period of Study:
1/1996-12/1996
Location:
3 South Korea cities:
Author: Delfino etal.
(2008, 156390)
Period of Study:
January 1,
2000-December 31 ,
2003
Location: Orange
County, California
Design
Hospital Admissions
Health Outcome (ICD9):
Bronchial asthma ;COPD;
Bronchitis
Study Design: Time-series
Statistical Analyses: Poisson
GAM, LOESS
Age Groups Analyzed: All
Ages
ED Visits
Health Outcome: Asthma
Study Design: Longitudinal,
Cohort
Statistical Analyses:
Proportional hazards models in
SAS version 9.2
Age Groups Analyzed: 0-1 8 yr
Sample Description: Various
gender, race, insurance status,
income, poverty level,
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Daejeon: 1 .424 (0.611) ppm
Ulsan: 0.950 (0.211) ppm
Suwon: 1.270 (0.549) ppm
Range (Min, Max):
Daejeon: (.364, 3.504)
Ulsan: (.380, 1.675)
Suwon: (.250, 3.616)
CoPollutant: correlation
Daejeon
S02: r = 0.280; N02:r = 0.041;
TSP: r = 0.193;03: r = -0.101;
03 Max: r = -0.069
Ulsan
S02: r = 0.108; N02:r = 0.446;
TSP: r = 0.286; 03:r = -0.195;
03 Max: r = -0.107
Suwon
S02: r = 0.556; N02:r = 0.291;
TSP: r = 0.496; 03:r = -0.371;
03 Max: r = -0.365
Averaging Time: NR
Mean (SD) unit: Cool season:
0.114 (0.052), Warm season: 0.103
(0.048)
Range (Min, Max): Cool season:
0.014-0.378, Warm season: 0.013
-0.482
f^nPnlliifeint1 MO
wUr UllUOni. INWx
Effect Estimates (95% Cl)
Increment: 1,000 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Estimates obtained using dummy variables to apply
environmental indicators to the model
Daejeon
CO: 1.26(1.08-1.47)
TSP, S02,N02,03: 1.21 (1.02-1 .44)
Ulsan
CO: 3.55(1.65-7.63)
TSP, S02,N02,03: 2.51 (1.06-5.93)
Suwon
CO: 1.24(0.97-1.59)
TSP,S02, N02,03:1.19(0.88-1.61)
Estimates obtained using actual measured integrated
environmental pollution indicator values
Daejeon
CO: 1.34 (1.1 4-1 .58)
Ulsan
CO: 1.27 (0.94-1 .71)
Suwon
CO: 3.55 (1.27-9.93)
Increment: 0.056 ppm
HR (95% Cl): Unadjusted: 1 .072 (1 .016-1 .131),
Adjusted : 1 .073 (1 .01 3 - 1 .1 37) , Male : 1 .054 (0.978 -
1 .1 37), Female: 1 .1 00 (1 .011-1 .1 97), 0 yr: 1 .1 58 (1 .041 -
1 .289), 1-5 yr: 1 .021 (0.933 - 1 .117), 6-18 yr: 1 .076 (0.972
- 1 .191), Median or less poverty: 1 .054 (0.979 - 1 .134),
Greater than the median poverty: 1 .094 (1 .006 - 1 .190),
Greater than the median income: 1.120 (1.034-1.213),
Median or less income: 1.041 (0.959-1.129), Private
insurance: 1 .1 02 (1 .006 - 1 .206), Government sponsored
or self-pay insurance: 1.061 (0.989-1.138), Unknown
insurance : 0.91 3 (0.591 - 1 .41 2) , White : 1 .1 1 3 (1 .027 -
1 .205), Hispanic: 1 .081 (0.996 - 1 .173), Non-Hispanic
nonwhite: 0.804 (0.601 -1.074)
residence distance to treating
hospital
Lags examined: NR
The point estimates for CO are stronger in girls than in boy
and in infants than in older children. There is little
difference in coefficients between adjusted and unadjusted
CO models. There were significant increased risks of
repeated hospital encounters of 7% to 10% per IQR
increase in traffic-related CO exposure.
Author: Farhat et al.
(2005, 089461)
Period of Study:
8/1996-8/1997
Location:
Sao Paulo, Brazil
Hospital Visits & ED Visits
Health Outcome (ICD9):
Pneumonia/bronchopneumonia
(480-486); Asthma (493);
Bronchiolitis (466)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 3.8 (1.6) ppm
Range (Min, Max): (1.1, 11. 4)
CoPollutant: correlation
PM10:r = 0.72;
S02:r=0.49;
N02:r = 0.59;
03:r = -0.8
Increment: 1.8 ppm
% Increase (Lower Cl, Upper Cl); lag:
Lower Respiratory Tract Disease ED Visits
CO, PM10: -0.10% (-5.60 to 5.30); 0-2
CO,N02: -1.20% (-6.70 to 4.20); 0-2
CO, S02: 3.70% (-1 .00 to 8.40); 0-2
C0,03: 4.80% (0.50-9.10); 0-2
CO, PM10, N02, S02, 03:- 0.64% (-6.90 to 5.60); 0-2
Pneumonia/ Bronchopneumonia Hospital Admissions
CO, PM10: 4.40% (-7.90 to 16.70); 0-2
CO,N02: 4.40% (-88.70 to 17.50); 0-2
CO,S02: 7.80% (-2.50 to 18.20); 0-2
CO, 03: 9.60% (-0.50 to 19.70); 0-2
CO, PM10 to N02, S02, 03:5.1 0% (-9.60 to 19.70) ; 0-2
Asthma/ Bronchiolitis Hospital Admissions
CO, PM10: 6.10% (-14.90 to 27.10); 0-2
CO,N02: 2.40% (-16.90 to 21.70); 0-2
CO, S02: 10.60% (-6.60 to 27.80); 0-2
CO, 03: 12.40% (-3.60 to 28.40); 0-2
CO, PM10 to N02, S02, 03: 8.80% (-15.60 to 33.30); 0-2
September 2009
C-46
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Fung et al.
(2006, 089789)
Period of Study:
6/1995-3/1999
Location:
Vancouver,
Canada
Author: Fusco etal.
(2001.020631)
Period of Study:
1/199510/1997
Location:
Rome,
Italy
Hospital Admissions
Health Outcome (ICD9):
Respiratory Illness
Study Design:
1 . Dewanji and Moolgavkar
2. Time-series
3. Bi-directional case-crossover
Statistical Analyses:
1. Dewanji and Moolgavkar
2. Poisson
3. Conditional logistic
regression
Age Groups Analyzed: 2 65 yr
Hospital Admissions
Health Outcome (ICD9):
Respiratory conditions (460-
519,
excluding 470-478); Acute
respiratory infections plus
pneumonia (460-466, 480-486);
COPD (490-492, 494-496)
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
0-1 4 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.69 0.25) ppm
Range (Min, Max): (0.28, 2.03)
CoPollutant: correlation
CoH: r= 0.85; 03:r = -0.53;
N02: r = 0.74; S02:r = 0.61;
PM10: r = 0.46; PM25:r= 0.23;
PM10-2.5:r = 0.51
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
3.6(1.2)mg/m3
IQR (25th, 75th): (2.8, 4.3)
CoPollutant: correlation
All Year
9(V r - n "ifi
OW2. ' U.UU
N02:r=0.31
03: r = -0.57
Cold Season
cri • r - n ?7
OW2. 1 — U.O(
N02:r = 0.41
03: r = -0.44
Warm Season
S02: r = 0.44
N02:r=0.59
0- r - 0 ?8
3. 1 ~ ~U.OO
Increment: 0.24 ppm
Relative Risk (Lower Cl, Upper Cl); lag
Dewanji and Moolgavkar
1. 008 (0.997-1. 02) ;0
1.012 (0.999-1 .025); 0-2
1.010 (0.995-1 .025); 0-4
1.009 (0.991 -1.026); 0-6
Tim6-S6ri6s
1. 01 2 (1.000-1. 023) ;0
1.01 7 (1.003-1 .032); 0-2
1 017(1 001-1 035) '0-4
1.016 (0.996-1 .036); 0-6
Bi-directional case-crossover
1.010 (0.006-1. 023); 0
1.012 (0.996-1 .027); 0-2
1.012 (0.995-1 .03); 0-4
1.010(0.991 1.031; 0-6
Increment: 1 .5 mg/m3
% Increase (Lower Cl, Upper Cl); lag:
Age Group: All Ages
Respiratory conditions
2.80% (1. 30-4.30) ;0
1.80% (0.20-3.30);!
0.20% (-1.30 to 1.80); 2
0.50% (-2.00 to 1.10); 3
0.70% (-0.80 to 2. 20); 4
CO, N02: 2.30% (0.60-4.00); 0
Acute Respiratory Infections plus Pneumonia
2.20% (0.00-4.40) ;0
2.10% (-0.10 to 4.40); 0
1.70% (-0.50 to 4.00); 2
-0.90% (-3.00 to 1.30); 3
1.50% (-0.70 to 3. 70); 4
CO, N02: 0.00% (-2.30 to 2.40); 0
Asthma
5.50% (0.90-1 0.40) ;0
0. 80% (-3 .80 to 5. 70) ;1
-1.30% (-5.90 to 3.50); 2
-3.00% (-7.40 to 1.60); 3
0.60% (-4.00 to 5.30); 4
CO, N02: 4.80% (0.30-9.50); 0
COPD
4.30% (1. 60-7.1 0);0
-0.20% (-2.90 to 2.50); 1
-0.20% (-2.90 to 2.60); 2
-0.30% (-3.00 to 2.40); 3
-0.10% (-2.80 to 2.60); 4
CO, N02: 4.80% (0.90-7.90); 0
Warm Season
Respiratory Conditions:
10.80% (6.70-14.80); 0
Acute respiratory infections plus pneumonia:
8.60% (2.90-1 4.60) ;0
COPD:
13.90% (6.80-21. 50); 0
Age Group: 0-1 4
Respiratory conditions
2. 50 (-0.30 to 5. 50) ;0
0.80 (-2.10 to 3.80); 1;
0.20 (-2.70 to 3.10); 2
-1.00 (-3.70 to 1.90); 3
3. 20 (0.40- 6. 20); 4
CO, N02: 4.10 (-1.20 to 9.80); 1
Acute Respiratory Infections plus Pneumonia
2. 50 (-0.80 to 5. 80) ;0
-0.10 (-3.40 to 3.20); 1
0.90 (-2.30 to 4.30); 2
-2.00 (-5.10 to 1.20); 3
3.20 (0.00-6.60); 4
CO, N02: 6.90 (0.80-13.40);!
Asthma
6. 30 (-0.50 to 13. 50) ;0
8.20(1.10-15.70);!;
-0.70 (-7.30 to 6.30); 2
September 2009
C-47
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
3.50 (-3.20 to 10.60); 3;
4.80 (-1.90 to 12.00); 4
CO, N02:3.30 (-4.20 to 11.30);1
Author: Gouveia and
Fletcher (2000
010436)
Period of Study:
11/1992-9/1994
Location:
Sao Paulo,
Brazil
Hospital Admissions
Health Outcome (ICD9): All
respiratory diseases
Pneumonia (480-486); Asthma
(493); Bronchitis (466, 490,
491)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 5.8 (2.4) ppm
Range (Min, Max): (1.3, 22.8)
CoPollutant: correlation
PM10:r = 0.63
S02: r = 0.65
N02:r=0.35
Increment: 6.9 ppm
Relative Risk (Lower Cl, Upper Cl);
All respiratory diseases
Age Group:
<5: 1.017 (0.971-1. 065); 0
Pneumonia
Age Group:
<5: 1.015 (0.961-1. 071); 0;
<1: 1.035 (0.975-1 .099); 2
lag:
Asthma
Age Group:
<5:1.081 (0.98-1.192); 0
Author: Hajat et al.
(1999.000924)
Period of Study:
1/1992-12/1994
Location:
London,
U.K.
General Practitioner Visits
Health Outcome (ICD9):
Asthma (493); Lower
Respiratory Diseases (464,
466,476,480-483, 485-487,
490-492,494-496,500,501,
503-505,510-515,518,519,
786)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
0-14 yr
15-64yr
>65yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
All yr: 0.8 (0.4) ppm
Warm Season
(April-September): 0.7 (0.3) ppm
Cool Season
(October-March): 1.0 (0.5) ppm
Range (10th, 90th):
All Year: (0.5,1.3)
Warm Season: (0.4,1.0)
Cool Season: (0.5,1.6)
CoPollutant: correlation
All Year
N02:r = 0.72;
S02:r=0.51;
BS:r = 0.85;
03:r = -0.40;
PM10:r = 0.56
Warm Season
N02:r = 0.70;
S02:r=0.32;
BS:r = 0.65;
03:r = -0.12;
PM10:r = 0.58
Cool Season
N02:r = 0.84;
S02:r=0.58;
BS:r = 0.87
Increment: 0.8 & 0.7 ppm
% Increase (Lower Cl, Upper Cl); Lag
All Year:
Asthma - Single Day Lags
Increment: 0.8 ppm
Age Group
0-14:4.10% (-0.10 to 8.40); 2
15-64:0.90% (-2.10 to 4.10);0
> 65:7.50% (0.50-14.90); 2
All ages: 1.60% (-1.20 to 4.60); 2
Asthma - Cumulative exposure
Increment: 0.7 ppm
Age Group
0-14:6.90% (1.30-12.90); 0-3
15-64:1.00% (-3.20 to 5.40); 0-2
> 65:8.20% (0.40-16.60);0-2
All ages: 1.80% (-1.50 to5.20);0-2
Lower Respiratory Diseases - Single Day Lags
Increment: 0.8 ppm
Age Group
0-14:4.40 (1.70-7.10); 2
15-64:1.10 (-0.70 to 3.00); 2
> 65:-2.60 (-4.80 to-0.30); 3
All ages: 2.00 (0.50-3.40); 2
Lower Respiratory Diseases - Cumulative exposure
Increment:
0.7 ppm for 0-2 and 0-3; 0.8 for 0-1
Age Group
0-14:3.00% (-1.00 to 7.20); 0-3
15-64:-0.70% (-2.90 to 1.50); 0-1
> 65:-1.60% (-5.10 to 2.00);0-3
All ages: 1.80% (0.10-3.60); 0-2
Warm or Cold Seasons:
Asthma, Increment: 0.8 ppm
Age Groups Season
0-14 & Warm Season: 11.40% (3.30-20.00); 2
0-14 & Cold Season: 2.90% (-3.20 to 9.40); 2
15-64 & Warm Season: 4.80% (-0.60 to 10.60); 0
15-64 & Cold Season: -0.30% (-4.80 to 4.50); 0
> 65 & Warm Season: 15.60% (3.10-29.60); 2
> 65 & Cold Season: 4.20% (-6.00 to 15.60); 2
Lower Respiratory Diseases, Increment: 0.8 ppm
Age Group& Season
0-14 & Warm Season: 2.70% (-2.90 to 8.60); 2
0-14 & Cold Season: 6.20% (2.30-10.20); 2
15-64 & Warm Season: 6.20% (2.30-10.20); 2
15-64 & Cold Season: 2.40% (-1.20 to 6.10); 2
> 65 & Warm Season: 1.00% (-1.60 to 3.80); 2
> 65 & Cold Season: -2.20% (-6.50 to 2.40); 3
September 2009
C-48
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author:
Hajat et al. (2002,
030358)
Period of Study:
1/1992-12/1994
Location:
London, U.K.
Author: Hapcioglu et
al. (2006. 093263)
Period of Study:
1/1997-12/2001
Location:
Istanbul,
Turkey
Author: Hinwood et
al. (2006. 088976)
Period of Study:
1/1992-12/1998
Location:
Perth,
Australia
Author: Hwang and
Chan (2002, 023222)
Period of Study:
1998
Location:
50 communities in
Taiwan
Design Concentrations
General Practitioner Visits Pollutant: CO
Health Outcome (ICD9): Averaging Time: 24-h avg
Upper Respiratory Diseases
(IJ£D) Mean (SD) unit:
Allyr:0.8(0.4)ppm
Study Design: Time-series Warm Season (April-September): 0.7
(0 3) ppm
Statistical Analyses: cool Season (October-March) :
Poisson, GAM, LOESS ^ 0 (0.5) ppm
Age Groups Analyzed: Range (10th, 90th):
0-1 4 yr All Year: (0.5, 1.3)
15-64yr Warm Season: (0.4, 1.0)
265Yr Cool Season: (0.5, 1.6)
CoPollutant: NR
Hospital Admissions Pollutant: CO
Health Outcome (ICD9): Averaging Time: Monthly
COPD (490-492, 494-496) .. „.„, .t llpl
Mean (SD) unit: NR
Study Design: Cross-sectional
Range (Min, Max): NR
Statistical Analyses:
Pearson Correlation Coefficient CoPollutant: NR
Age Groups Analyzed: All
ages
Hospital Admissions Pollutant: CO
Health Outcome (ICD9): Averaging Time:
COPD (490.00-496.99 Maximum 8-h avg
excluding asthma)
Pneumonia/influenza (480 00- Mean (SD) umt:
489.99); Asthma (493) All Year: 2.3 (1 .3) ppm;
November-April: 2.2 (1.3) ppm;
Study Design: Case-crossover May-October: 2.4 (1 .2) ppm
Statistical Analyses: Range (10th, 90th):
Conditional logistic regression All Year: (0.9, 4.2)
November-April: 0.8, 4.2)
Age Groups Analyzed: All May-October' (1 1 42)
ages
CoPollutant: correlation
All Year:
N02:r = 0.57
03:r = 0.00
November-April:
N02:r = 0.55
03:r = 0.00
May-October:
N02:r=0.57
03:r = 0.16
Clinic Visits Pollutant: CO
Health Outcome (ICD9): Averaging Time:
Lower respiratory tract Maximum 8-h avg
infections (466, 480-486)
Mean (SD) unit:
Study Design: Time-series 1 .00 (0.30) ppm
Statistical Analyses: Range (Min, Max): (0.51 , 1 .71)
1. General linear regression
2. Bayesian hierarchical CoPollutant: NR
modeling
Age Groups Analyzed:
All Anac
r\\\ rtyco
n 1 4 \/r
u 1 1 yi
1 ^ RA \/r
1 0-04 yi
> R^i \/r
^ oo yi
Effect Estimates (95% Cl)
Increment: 0.6 ppm, 0.8 ppm, & 1 .1 ppm
% Increase (Lower Cl, Upper Cl); lag:
Warm Season, Increment: 0.6 ppm
Age Group
0-14: 2.90% (-0.60 to 6.40); 1
14-64: 7.90% (4.80-11.10);!
> 65: 4.90% (-1.80 to 12.10); 3
Cold Season, Increment: 1 .1 ppm
Age Group
0-1 4: -2.50% (-4.90 to 0.1 0);1
1 4-64: 0.60% (-1.60 to 2.90) ;1
> 65: 5.60% (0.90-10.60); 3
All Year, Increment: 0.8 ppm
Age Group
0-14: -2.20% (-4.00 to -0.30); 1
14-64: 2.70% (0.10-5.50);!
> 65: 5.80% (2.40 to 9.30); 3
Correlation Coefficient:
Between CO exposure and COPD: 0.57
Between CO exposure and COPD when controlling for
temperature: 0.25
Increment: 2.3 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
Pneumonia
0.99999 (0.9737-1. 0268) ;0
1.00650(0.9806-1.0331);!
1.00351 (0.9779-1 .0298); 2
1.00424 (0.9790-1 .0301); 3
1.00581 (0.9752-1 .0374); 0-1
1.01 005 (0.9755-1 .0458); 0-2
1.00805 (0.9701 -1.0474); 0-3
COPD
0.9991 5 (0.9693-1. 0297) ;0
1.00205(0.9727-1.0323);!
0.98630 (0.9577-1 .01 58); 2
0.98970 (0.9619-1.01 82);3
0.99960 (0.9647-1 .0357); 0-1
0.99260 (0.9538-1 .0329); 0-2
0.991 60 (0.9493-1 .0357); 0-3
Increment: 0.1 ppm
% Increase (Lower Cl, Upper Cl); Lag
Age Group: All Ages
0.80% (0.60-1. 00) ;0
0.10% (-0.10 to 0.30); 1
0.10% (-0.10 to 0.30); 2
Age Group: 0-1 4
0.70% (0.50-1. 00) ;0
0.10% (-0.20 to 0.30); 1
0.20% (-0.10 to 0.40); 2
Age Group: 15-64
0.90% (0.60-1 .10);0
0.20% (0.00-0.50);!
0.20% (-0.10 to 0.40); 2
Age Group: > 65
1. 10% (0.80-1. 50) ;0
0.60% (0.30-1. 00) ;1
0.40% (0.1 0-0.80); 2
September 2009
C-49
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Ito et al.
(2007, 091262)
Period of Study:
1999-2002
Location:
New York City, NY
Design
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
GLM
Age Groups Analyzed: All
ages
Concentrations
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
All Season: 1.31 (0.43) ppm
Warm Months (April-September):
1.22 (0.32) ppm
Cold Months (October-March): 1 .41
(0.5) ppm
Effect Estimates (95% Cl)
Increment: 1.3 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
Warm mo: 1.1 5 (1.07-1 .25); 0-1
Range (5th, 95th):
All season: (0.77,2.11)
Warm Months (April-September):
(0.75,1.82)
Cold Months (October-March): (0.78,
2.33)
CoPollutant: NR
Author: Jayaraman et
al. (2008. 180352)
Period of Study:
2004-2005
Location:
New Delhi, India
Author: Karretal.
(2007, 090719)
Period of Study:
1995-2000
Location:
South Coast Air Basin,
CA
Hospital Admissions
Health Outcome: respiratory
Study Design: time series
Statistical Analyses: Poisson
regression (GAM)
Age Groups Analyzed:
all
Sample Description:
daily HA for respiratory unit of
Safdarjung hospital
Hospital Admissions
Health Outcome (ICD9):
Acute bronchiolitis (466.1)
Study Design: Matched case-
control
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Infants : 3 wk-yr
Averaging Time: 24-h
Mean (SD) unit:
2,379.14 (1,289.18) ug/m3
Range (min, max):
588, 8458
CoPollutant:
S02:r = 0.217*
N02:r= 0.204*
SPM:r = 0.071
RSPM: r = 0.120
03: r = 0.063
*p<0.05
Pollutant:CO
Averaging Time: 24-h avg
Mean (SD) unit:
Chronic: 1,770 ppb
Subchronic:1,720ppb
Range (Min, Max):
Chronic: (120, 8300)
Subchronic: (130, 5070)
CoPollutan :NR
Increment: 10 ug/m3
RR Estimate [Lower Cl, Upper Cl] ;
Lags examined: lag days 0-3
lag:
Single Pollutant: 0.9989 (0.985, 2.715), 2
Multi-pollutant: 0.998 (0.993, 1.004),
Winter, all ages: 1 .027 (1 .004, 1 .051)
Winter, males 50-69: 2.625 (1.048,1
Increment: 910 ppb, 960 ppb
Odds Ratio (Lower Cl, Upper Cl); lag
Increment: 910 ppb
2
,2
.158)
Subchronic broncholitis: 1 (0.97-1.03)
Increment: 960 ppb
Chronic broncholitis: 1 (0.97-1 .03)
September 2009
C-50
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Karretal.
(2006, 088751)
Period of Study:
1995-2000
Location:
South Coast Air Basin,
CA
Hospital Admissions
Health Outcome (ICD9):
Acute bronchiolitis (466.1)
Study Design: Case-
Crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Infants: 3 wk-1 yr
Pollutant:CO
Averaging Time: 24-h avg
Mean (SD) unit:
1 -day lag:
Index*: 1 ,730 ppb
Referent*: 1 ,750 ppb
4-day lag:
Index*: 1 ,760 ppb
Referent*: 1 ,790 ppb
Increment: 1361, 1400 ppb
Odds Ratio (Lower Cl, Upper Cl);
Increment: 1361 ppb
Age Group:
Overall: 0.99(0.96-1.02);!
25-29 wk: 0.86(0.68-1.1);!
291/7-34 wk: 1 (0.86-1.15);!
34 1/7- 37 wk: 0.95(0.87-1.04);!
37 1/7- 44 wk: 1 (0.97-1.03);!
Increment: 1400 ppb
Lag
Range (Min, Max):
Lag1:
Index*: (4, 9600) Referent*: (4, 9600)
Lag 4:
Index* (4, 8710) Referent* (4,9600)
Copollutant: NR
* Index days: days lagged in
reference to date of hospitalization of
a case.
Referent days: are for each case and
includes all days that are the same
day of wk and in the same mo as the
index day for that case for CO.
Age Group:
Overall: 0.97 (0.94-1); 4
25-29 wk: 0.93 (0.72-1.2); 4
29 1/7-34 wk: 0.89 (0.77-1.03); 4
34 1/7-37 wk: 0.98 (0.90-1.08); 4
37 1/7-44 wk: 0.97 (0.94-1); 4
Author: Kimetal.
(2007, 092837)
Period of Study:
2002
Location:
Seoul,
Korea
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45 and J46)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All Ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
Daily Concentration: 8.6 (4.6) ppm
Relevant Concentration:
2.8 (2.8) ppm
Range (Min, Max):
Daily Concentration: (0.8, 44.0)
Relevant Concentration: (0.0, 30.4)
Copollutant: NR
Relative Risk (Lower Cl, Upper Cl); lag:
Individual Level SEP
Quintile1:1.06(1.02-1.09);1-3ma
Quintile 2:1.05(1.02-1.09); 1-3 ma
Quintile 3:1.05 (1.01-1.08); 1-3 ma
Quintile 4:1.07 (1.03-1.11); 1-3 ma
Quintile 5:1.05 (1.00-1.09); 1-3 ma
Regional Level SEP
Quintile 1:0.99 (0.92-1.07); 1-3 ma
Quintile 2:1.06 (1.02-1.11); 1-3 ma
Quintile 3:1.04 (1.02-1.07); 1-3 ma
Quintile4:1.10(1.06-1.15); 1-3 ma
Quintile 5:1.06 (1.03-1.09); 1-3 ma
Overall: 1.06 (1.04-1.07); 1-3 ma
Relative Effect Modification for SES
Individual Level SEP
Quintile 1:1
Quintile 2:1 (0.95.1.04); 1-3 ma
Quintile 3:0.99 (0.94-1.03); 1-3 ma
Quintile 4:1.02 (0.97-1.06); 1-3 ma
Quintile 5:0.99 (0.94-1.04); 1-3 ma
Regional Level SEP
Quintile 1:1
Quintile 2:1.05 (0.97-1.14); 1-3 ma
Quintile 3:1.03 (0.96-1.11); 1-3 ma
Quintile 4:1.08 (1-1.16); 1-3 ma
Quintile5:1.05 (0.97-1.13); 1-3 ma
Author: Kontosetal.
(1999, 011326) Period
of Study:
1/1987-12/1992
Location:
Piraeus, Greece
Hospital Admissions
Health Outcome (ICD9):
Respiratory conditions
(laryngitis, bronchiolitis,
tonsillitis, acute
rhinopharyngitis, otitis,
bronchopneumonia,
pneumonia, asthma)
Pollutant: CO
Averaging Time: 24-h avg
Mean Range (SD) unit:
1987: 4.2 mg/m3
1992:3.6mg/m3
Range (Min, Max): NR
This study did not present quantitative results for CO.
Study Design: Time-series
Statistical Analyses:
Stochastic dynamical system
approach
Age Groups Analyzed: 0-14 yr
Copollutant: correlation
1987-1989
Smoke: r = 0.2979;
S02:r=0.2166;
N02:r= 0.1913
1990-1992
Smoke: r = 0.5383;
S02:r= 0.43283;
N02:0.5223
September 2009
C-51
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Lee et al.
(2002, 034826)
Period of Study:
12/1997-12/1999
Location:
Seoul, Korea
Author: Lee et al.
(2006, 098248)
Period of Study:
1/2002-12/2002
Location:
Seoul,
Korea
Author: Lee et al.
(2007, 090707)
Period of Study:
1996-2003
Location:
Kaohsiung, Taiwan
Author: Lin et al.
(1999.040437)
Period of Study:
5/1991-4/1993
Location:
Sao Paulo, Brazil
Design
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45, J46)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: <15yr
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45-46)
Study Design: Time-series
Statistical Analyses:
GAM with stringent parameters
Age Groups Analyzed: <15yr
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Bi-directional
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
ED Visits
Health Outcome (ICD9):
Respiratory illness (lower
respiratory illness, upper
respiratory illness, wheezing)
Study Design: Time-series
Concentrations
Pollutant: CO
Averaging Time: 1-h max
Mean Range (SD) unit:
1.8(0.7) ppm
IQR (25th, 75th): (1.2, 2.2)
Copollutant: correlation
PM10:r = 0.598
S02:r = 0.812
N02:r= 0.785
03: r = -0.388
Pollutant: CO
Averaging Time:
Maximum 2-h avg
Mean (SD) unit:
High SES: 6.08 (2.10) ppb
Moderate SES: 6.35 (2.44) ppb
Low SES: 6.67 (2.59) ppb
Range (Min, Max): NR
Copollutant: correlation
N02:r=0.55
S02: r = 0.72
PM10:r = 0.28
03: r = -0.36
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.77 ppm
Range (Min, Max): (0.23, 1.72)
Copollutant:
PM10
en
OW2
N02
3
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 5 ppm
Range (Min, Max): (1,1 2)
CoDollutant: correlation
Effect Estimates (95% Cl)
Increment: 1.0 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
RR for asthma and exposure to various pollutants for
children under 1 Syr old
Pollutant:
CO: 1.1 6 (1.1 0-1 .22); 2-3 avg
CO, PM10: 1.1 3 (1.07-1 .20); 2-3 avg
CO,S02: 1.1 7 (1.08-1 .27); 2-3 avg
CO, N02: 1.04 (0.95-1 .14); 2-3 avg
C0,03: 1.1 6 (1.1 1-1 .22); 2-3 avg
CO, 03, PM10: 1 .148 (1 .084-1 .217); 2-3 avg
CO, 03, PM10, S02: 1 .1 68 (1 .075-1 .269); 2-3 avg
CO, 03, PM10, S02, N02: 1 .098 (0.994-1 .214); 2-3 avg
Increment: 3.01 ppb, 0.26 ppb, 4.52 ppb, 3.68 ppb
Relative Risk (Lower Cl, Upper Cl); lag:
Increment: 3.01 ppb
Overall: 1.07 (0.96-1 .20); 0
Increment: 0.26 ppb
High SES: 1.06 (0.96-1 .17); 0
Increment: 4.52 ppb
Moderate SES: 0.96 (0.84-1. 10); 0
Increment: 3.68 ppb
Low SES: 1. 02 (0.85-1. 24) ;0
Increment: 0.29 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
CO
<25°C: 1.398 (1.306-1 .496); 0-2
>25°C: 1.1 89 (1.1 23-1 .259); 0-2
CO. PM10
<25°C '1 257(1 152-1 371) '0-2
>25°C: 1.1 49 (1.079-1 .224); 0-2
CO, S02
<25°C: 1.396 (1.295-1 .504); 0-2
>25°C: 1.241 (1.1 61 -1.326); 0-2
CO, N02
<25°C: 0.973 (0.877-1 .080); 0-2
>25°C: 1.1 96 (1.1 04-1 .297); 0-2
C0,03
<25°C: 1.378 (1.286-1 .477); 0-2
>25°C: 1.1 70 (1.1 05-1 .239); 0-2
Increment: NR
Relative Risk (Lower Cl, Upper Cl); lag:
Overall Respiratory Illnesses
CO: 1.206 (1.066-1 .364); 0-5
CO, PM10, 03, S02, N02: 0.945 (0.808-1 .105); 0-5
Statistical Analyses: Poisson
Age Groups Analyzed: <13yr
PM10:r = 0.50
N02:r=0.35
S02: r = 0.56
03:r = 0.04
CO: 1.203 (0.867-1.669); 0-5
CO, PM10,03, S02, N02:0.971 (0.641-1.472); 0-5
Upper Respiratory Illness
CO: 1.237 (1.072-1.428); 0-5
CO, PM10,03, S02, N02:0.944 (0.785-1.135); 0-5
Wheezing
CO: 0.813 (0.606-1.091); 0-5
CO, PM10, N02, S02, 03:0.74 (0.505-1.085); 0-5
September 2009
C-52
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lin et al.
(2003, 042549)
Period of Study:
1/1981-12/1993
Location:
Toronto, ON,
Canada
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.18 (0.50) ppm
Range (Min, Max): (0,6.10)
Copollutant: correlation
Age Groups Analyzed: 6-12 yr N02: r = 0.55
03:r = -0.16
PM25:r = 0.45
PM10-2.5:r = 0.17
PM10:r = 0.38
Increment: 0.5 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Boys:
Adjusting for Daily Weather Variables
1.05 (1-1.11); 1 /1.07 (1.01-1.14); 2
l!o7(o!99-l!l6);5/l!o7(0.98-1.17);6
1.07 (0.98-1.17); 7
Adjusting for PM and Daily Weather Variables
1.05(0.99-1.11);! /1.08 (1.01-1.16);2
1.09 (1.01-1.18); 3/1.10 (1.02-1.20); 4
1.09(1.00-1.18);5/1.09(0.99-1.19);6
1.09 (0.99-1.20); 7
Girls:
Adjusting for Daily Weather Variables
1.00(0.93-1.06);! /1.01 (0.94-1.10);2
1.00 (0.91 -1.09); 3 / 0.98 (0.89-1.09); 4
1.01(0.91-1.13);5/1.03(0.92-1.16);6
1.04 (0.93-1.17); 7
Adjusting for PM and Daily Weather Variables
1.00 (0.93-1.07); 1/1.01 (0.92-1.10);2
0.99 (0.90-1.09); 3 / 0.97 (0.87-1.08); 4
0.99 (0.89-1.11); 5/1.02 (0.90-1.15); 6
1.05 (0.93-1.20); 7
Author: Lin et al.
(2004, 055600)
Period of Study:
1/1987-12/1998
Location:
Vancouver, BC
Canada
Author: Lin et al.
(2005, 087828)
Period of Study:
1998-2001
Location:
Toronto,
Canada
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
GAM, LOESS
Age Groups Analyzed: 6-1 2 yr
Hospital Admissions
Health Outcome (ICD9):
Respiratory Infections (464,
466, and 480-487)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: <15yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.96 (0.52) ppm
Range (Min, Max): (0.23, 4.90)
Copollutant: correlation
S02: r = 0.67
N02:r=0.73
03: r = -0.35
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.1 6 (0.38) ppm
Range (Min, Max): (0.38, 2.45)
Copollutant: correlation
ply] • r = o 1 0
PM 10-2.5: r = 0.06
PM10:r = 0.10
S02' r = 0 12
N02:r=0.20
03:r = -0.11
Increment: 0.5 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Boys
HighSES:
1.06(0.98-1.14);1/1.06(0.97-1.15);2
1.07 (0.97-1 .17); 3/1. 03 (0.93-1 .14); 4
1.01 (0.91-1 .12); 5/1. 01 (0.91-1.13);6
1.06 (0.94-1 .18); 7
LowSES:
1.06(0.99-1.14);1/1.03(0.95-1.12);2
1.01(0.93-1.11);3/0.99(0.90-1.09);4
0.96 (0.87-1 .06); 5 / 0.98 (0.88-1 .08); 6
0.98 (0.88-1 .09); 7
Girls
HighSES:
1.05(0.94-1.16);1/1.02(0.90-1.15);2
0.97(0.85-1.11);3/0.95(0.83-1.10);4
0.93 (0.80-1 .08); 5 / 0.95 (0.82-1 .11); 6
1.01 (0.87-1 .19); 7
LowSES:
1.01 (0.92-1.11);! /0.98(0.89-1.10);2
0.99(0.88-1.11);3/1.05(0.93-1.19);4
1.07(0.94-1.21);5/1.07(0.94-1.23);6
1.04 (0.91 -1.20); 7
Increment: 0.44 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
Boys
No adjustment:
1 .1 1 (1 .01 -1 .22) ; 0-3 / 1 .1 0 (1 .00-1 .22) ; 0-5
Adjustment for weather variables :
1 .1 3 (1 .03-1 .24) ; 0-3 / 1 .1 3 (1 .02-1 .25) ; 0-5
Adjustment for weather variables and PM :
1 .08 (0.98-1 .20) ; 0-3 / 1 .08 (0.97-1 .20) ; 0-5
Girls
No adjustment:
0.99 (0.89-1 .1 0) ; 0-3 / 1 .00 (0.89-1 .1 3) ; 0-5
Adjustment for weather variables :
1. 02 (0.92-1 .14); 0-3: / 1.05 (0.93-1 .18); 0-5
Adjustment for weather variables and PM :
1 .01 (0.90-1 .1 3) ; 0-3 / 1 .02 (0.90-1 .1 5) ; 0-5
Total
No adjustment:
1 .06 (0.98-1 .1 4) ; 0-3 / 1 .06 (0.98-1 .1 5) ; 0-5
Adjustment for weather variables :
1.09 (1.01 -1.1 7); 0-3/1. 10 (1.01 -1.1 9); 0-5
Adjustment for weather variables and PM :
1 .05 (0.97-1 .1 4) ; 0-3 / 1 .06 (0.97-1 .1 5) ; 0-5
September 2009
C-53
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Linn et al.
(2000, 002839)
Period of Study:
1992-1995
Location:
Los Angeles, CA
Hospital Admissions
Health Outcome (ICD9):
APR-DRG Codes: Pulmonary
(75-101);COPD(88)
ICD9 Codes: Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
0-29 yr; > 30 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Winter 1.7(0.8) ppm
Spring 1.0 (0.3) ppm
Summer 1.2 (0.4) ppm
Fall 2.1 (0.8) ppm
Range (Min, Max):
Winter: (0.5, 5.3)
Spring: (0.4,2.2)
Summer: (0.3,2.7)
Fall: (0.6, 4.3)
Copollutant: correlation
Winter
N02: r= 0.89; PM10:r= 0.78;
03: r = -0.43
Spring
N02: r= 0.92; PM10:r= 0.54;
03:r = 0.29
Summer
N02: r= 0.94; PM10:r= 0.72;
03:r = 0.03
Fall
N02: r= 0.84; PM10:r= 0.58;
03: r = -0.36
Increment: 1.0 ppm
P(SE);lag:
Pulmonary
Age Group: > 30
All Year: 0.007
Winter: 0.016
Spring: 0.014
Summer: 0.020
Fall: 0.020
Asthma
Age Group 0-29
All Year: 0.036
Asthma
Age Group: > 30;
All Year: 0.028
Winter: 0.045
Fall: 0.039
COPD
Age Group: > 30
All Year: 0.019
Winter: 0.035
Fall: 0.029
September 2009
C-54
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Luginaah et
al. (2005. 057327)
Period of Study:
4/1995-12/2000
Location:
Hospital Admissions
Health Outcome (ICD9):
Respiratory illness (460-519)
Study Design:
Time-series and Case-
Windsor, ON, Canada crossover
Statistical Analyses:
1. Time-series: Poisson
2. Case-crossover: conditional
logistic regression
Age Groups Analyzed:
All ages
0-14 yr
15-64yr
>65yr
Pollutant: CO
Averaging Time: 1-h max
Mean(SD)unit:1.3(1.0)ppm
Range (Min, Max): (0,11.82)
Copollutant: correlation
N02:r=0.38
S02:r = 0.16
03:r = 0.10
CoH:r=0.31
PM10:r = 0.21
Increment: 1.17 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
Females and Case-crossover study design
Age Group: All ages:
1.037(0.968-1.111);!
1.063 (0.976-1.158); 2
1.087 (0.982-1.203); 3
Age Group: 0-14:
1.147(1.006-1.307);!
1.186 (1.020-1.379); 2
1.221 (1.022-1.459); 3
Age Group: 15-64:
1.005(0.884-1.141);!
1.007 (0.859-1.181);2
1.032 (0.858-1.240); 3
Age Group: a 65:
1.014(0.922-1.116);!
1.024 (0.907-1.156); 2
1.035 (0.893-1.200); 3
Males and Case-crossover study design
Age Group:AII Ages:
0.950(0.884-1.020);!
0.945 (0.862-1.036); 2
0.965 (0.866-1.075); 3
Age Group: 0-14:
1.003(0.904-1.113);!
0.997 (0.871-1.141);2
0.970 (0.824-1.141);3
Age Group: 15-64:
1.036(0.870-1.233);!
1.033 (0.821-1.299); 2
0.991 (0.760-1.293); 3
Age Group: > 65:
0.867(0.775-0.970);!
0.865 (0.752-0.994); 2
0.946 (0.807-1.109); 3
Female and Time-series study design
Age Group:AII Ages:
1.049(0.993-1.108);!
1.032 (0.993-1.188); 2
1.051 (0.993-1.112); 3
Age Group: 0-14:
1.077(0.979-1.184);!
1.068 (1.001-1.139); 2
1.100 (0.997-1.213);3
Age Group: 15-64:
1.072(0.962-1.195);!
1.025 (0.944-1.112);2
1.081 (0.963-1.213); 3
Age Group: a 65:
1.029(0.957-1.118);!
1.030 (0.928-1.144); 2
1.013 (0.899-1.142); 3
Male and Time-series study design
Age Group:AII Ages:
0.989(0.932-1.049);!
0.986 (0.946-1.029); 2
0.987 (0.929-1.048); 3
Age Group: 0-14:
1.034(0.949-1.126);!
0.996 (0.933-1.062); 2
0.968 (0.881-1.064); 3
Age Group: 15-64:
0.994(0.854-1.157);!
0.988 (0.884-1.104); 2
0.951 (0.806-1.121); 3
Age Group: a 65:
0.901 (0.817-0.994);!
0.904 (0.803-1.019); 2
0.963 (0.845-1.098); 3
September 2009
C-55
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Martins et al.
(2002, 035059)
Period of Study:
5/1996-9/1998
Location:
Sao Paulo, Brazil
Author: Masjedi et al.
(2003, 052100)
Period of Study:
9/1997-2/1998
Location:
Tehran, Iran
Design
ED Visits
Health Outcome (ICD10):
Chronic Lower Respiratory
Disease (CLRD: J40-47) for
chronic bronchitis, emphysema,
other COPD, asthma, and
bronchiectasia
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: >64 yr
ED Visits
Health Outcome (ICD9):
Total acute respiratory
conditions; Asthma (493);
COPD (490-492, 494, 496)
Study Design: Time-series
Statistical Analyses:
Multiple stepwise regression
Age Groups Analyzed: Adults
Concentrations
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 3.7(1. 7) ppm
Range (Min, Max): (1.0, 12.6)
Copollutant: correlation
N02:r = 0.62;
S02:r=0.51;
PM10:r = 0.73;
03:r = 0.07
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 8.85 ppm
Range (Min, Max): (2.1 5, 23.8)
Copollutant: NR
Effect Estimates
Increment: 1.63 ppm
P(SE);lag:
(95% Cl)
Chronic Lower Respiratory Diseases
Age Group
>64: 0.0489 (0.0274); 2
Increment: NR
p(p-value); lag;
Asthma: -0.779 (0.1 2)
COPD: 0.012 (0.71)
Acute Respiratory conditions: -0.086 (0.400)
Correlation coefficients:
Mean 3-day CO levels and asthma
Mean weekly CO level and asthma
: -0.300 (0.1 49)
: -0.1 4 (0.2)
Mean 10-day CO levels and asthma: -0.05 (0.43)
Author: McGowan et
al. (2002. 030325)
Period of Study:
6/1988-12/1998
Location:
Christchurch,
New Zealand
Author: Migliaretti et
al. (2007. 193772)
Period of Study:
1/1997-12/1999
Location:
Turin, Italy
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-487); Acute
respiratory infections (460-466);
Chronic lung Diseases (491-
492, 494-496); Asthma (493)
Study Design: Time-series
Statistical Analyses:
Generalized Additive Model
Age Groups Analyzed:
<15yr;>64yr
Hospital Admissions
Health Outcome (ICD9):
Respiratory Illness (chronic
bronchitis, emphysema, and
other COPD)
(490-496)
Study Design: Case-control
Statistical Analyses:
Multiple logistic regression
Age Groups Analyzed:
>15yr
15-64yr
>64yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.16 (1.51) mg/m3
Range (Min, Max): (0, 15.7)
Copollutant: NR
Pollutant: CO
Averaging Time: 8-h median
Median (SD) unit:
3. 36 (1.57) mg/m3
Range (Min, Max): NR
Copollutant: correlation TSP
This study did not provide quantitative results for CO.
Increment: 1 mg/m3
Odds Ratio (Lower Cl, Upper Cl)
CO
Age Group
> 15: 1.053 (1.030-1 .070)
15-64:1.040(0.987-1.085)
>64: 1 .054 (1 .027-1 .083)
CO , TSP
Age Group
> 15: 1.058 (1.024-1 .096)
15-64:1.062(0.993-1.135)
>64: 1.054 (1.01 1-1 .099)
Hag:
September 2009
C-56
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Moolgavkar
(2000, 010274)
Period of Study:
1987-1995
Location:
3 U.S. counties:
Los Angeles
County.CA
Cook County, IL
Maricopa County, AZ
Design
Hospital Admissions
Health Outcome (ICD9):
COPD plus asthma (490-496)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All Ages
0-1 9 yr
20-64 yr
>65yr
Concentrations
Pollutant: CO
Averaging Time: 24-h median
Median unit:
Cook: 993 ppb
LA: 1347 ppb
Maricopa: 1240 ppb
Range (Min, Max):
Cook: (224, 391 2)
LA:(237,5955)
Maricopa: (269, 4777)
Copollutant: correlation
Cook County:
N02: r= 0.63; S02:r = 0.35;
0- r — fl OQ
3. 1 U.£O
LA County:
N02: r= 0.80; S02:r = 0.78;
03: r = -0.52
Maricopa County:
Effect Estimates (95% Cl)
Increment: 1.0 ppm
% Increase (t-statistic); lag:
Age Group: > 65
Cook County
CO:
2.60 (1 . 9); O;/ 3.00 (2.2); 1;/ 1.30 (1.0); 2;
1 40(1 1 ) ' 3 ' / 1 1 0 (0 8) ' 4 ' / 2 30 (1 8) ' 5
Los Angeles County
CO:
5.40 (11. 3); O;/ 4.90 (10.1); 1;/5.00 (10.2);2;
4.90 (10.1); 3; / 4.00 (8.3); 4; / 4.30 (8.6); 5;
pn PM •
L-U, rlvl-io-
4.30 (3.3); 0; / 5.30 (4.2); 1 ; / 5.10 (4.0); 2;
6.80 (5.6); 3; / 6.90 (5.4); 4; / 6.30 (4.7); 5;
CO, PM25:
3.00 (1 .9); 0; / 3.90 (2.5); 1 ; / 4.20 (2.6); 2;
6.50 (4.4); 3; / 5.80 (3.8); 4; / 5.10 (3.1); 5
Maricopa County
CO:
1 .40 (1 .0); 0; / 0.80 (0.6); 1 ; / 1 .20 (0.9); 2;
1. 20 (0.9); 3; / 1. 50 (1.1); 4; / 4.90 (3.8); 5
N02: r= 0.66; S02:r = 0.53;
03:r = -0.61
Age Group: 0-19
Los Angeles County
CO:
8.20 (14.4); 0;/9.00 (15.9); 1;/9.20(16.4);2;
8.50 (15.0); 3;/7.00 (12.1); A;14.80 (8.1); 5;
CO, PM10:
7.50 (14.4); 0; /: 7.40 (5.2); 1; / 6.40 (4.3); 2;
8.00 (5.5); 3;/ 6.30 (4.0); 4;/ 5.30 (3.5); 5;
CO, PM10-2.5:
5.70 (3.4); 0; / 7.50 (4.9); 1;/ 5.60 (3.3); 2;
5.40 (3.5); 3; / 4.40 (2.7); 4;/1.80 (1.1); 5
Age Group: 20-64
Los Angeles County
CO:
Author: Moolgavkar
(2003, 042864)
Period of Study:
1987-1995
Location:
2 U.S. counties:
Los Angeles County,
CA, and Cook County,
IL
Hospital Admissions
Health Outcome (ICD9):
COPD plus asthma (490-496)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, Poisson GLM
with natural splines
Age Groups Analyzed:
All Ages;
>65yr
Pollutant: CO
Averaging Time: 24-h median
Median unit:
Cook: 993 ppb
LA: 1347 ppb
Maricopa: 1240 ppb
Range (Min, Max):
Cook: (224, 3912)
LA: (237,5955)
Copollutant: correlation
Cook County:
N02: r= 0.63; S02:r = 0.35;
03: r = -0.28
Los Angeles County:
N02: r= 0.80; S02:r = 0.78;
03: r = -0.52
3.70 (8.6); 0; / 3.90 (9.1); 1 ; / 4.50 (10.6); 2;
3.50 (8.3); 3; / 3.40 (7.9); 4; / 3.50 (7.9); 5;
CO,PM10:
5.00 (4.6); O;/ 3.00 (2.7); 1;/ 3.10 (2.8); 2;
5.20 (4.7); 3; / 5.90 (5.1); 4; / 4.90 (4.4); 5;
CO,PM25:
3.50 (2.5); O;/ 0.60 (0.4); 1 ;/ 1 .10 (0.8); 2;
5.70 (4.1); 3; / 4.70 (3.3); 4; / 3.90 (2.8); 5;
CO, PM10-2.5:
2.80 (2.2); 0; / 2.50 (2.0); 1 ;/ 0.60 (0.5); 2;
3.90 (3.2); 3; / 3.40 (2.8); 4; / 4.00 (3.4); 5
Increment: 1 ppm
% Increase (t-statistic); lag:
COPD-Los Angeles County
CO-
GAM-30(10-8):
5.48 (17.67); 0;/5.67 (18.22); 1;/ 5.90 (19.01); 2;
5.28 (16.94); 3; / 4.59 (14.50); 4; / 4.10 (12.80); 5
GAM-100(10-8):
2.37 (8.67); O;/ 2.41 (8.73); 1;/ 2.41 (8.76);2;
1 .81 6.58 ; 3; / 1 .38 4.94 ; 4; / 1 .07 3.82 ; 5
NS-100'
2.28 (5.65); O;/ 2.29 (5.50); 1;/ 2.32 (5.33); 2;
1 .74 (4.10); 3; / 1 .30 (3.16); 4; / 1 .00 (2.46); 5
COPD-Cook County
CO-
GAM-100(10-8):
2.11 (1.62);0;/2.85(2.16);1;/1.14(0.86);2;
1 .05 (0.79); 3; / 0.43 (0.33); 4; / 0.34 (0.26); 5
September 2009
C-57
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Neidelletal.
(2004, 057330)
Period of Study:
1992-1998
Location:
California
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
Linear Regression
Age Groups Analyzed:
0-1 yr
1-3yr
3-6 yr
6-12 yr
12-1 Syr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.777 (1.037) ppm
Range (Min, Max): NR
Copollutant: correlation
03
PM10
N02
Increment: NR
p(SE);lag;
Single-pollutant model
Age Group
0-1:-0.007 (0.009);
1-3:0.027(0.009);
3-6:0.053(0.010);
6-12:0.047(0.009);
12-18:0.025(0.008)
Fixed effect controlling for 03, PM10, and N02
Age Group
0-1:-0.01 (0.01);
1-3:0.024(0.011);
3-6:0.049(0.011);
6-12:0.023(0.011);
12-18:0.021 (0.009)
Fixed effect controlling for 03, PM10, N02 and Avoidance
Behavior
Age Group
0-1:-0.010(0.010);
1-3:0.027(0.011);
3-6:0.051 (0.011);
6-12:0.025(0.011);
12-18:0.021 (0.009)
Author: Morris et al.
(1999.040774)
Period of Study:
9/1995-12/1996
Location:
Seattle, WA
Author: Peel et al.
(2005, 056305)
Period of Study:
1/1993-8/2000
Location:
Atlanta, GA
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
Semiparametric Poisson GAM
Age Groups Analyzed: <18yr
ED Visits
Health Outcome (ICD9):
Asthma (493, 786.09) ;COPD
(491, 492, 496); URI (460-466,
477); Pneumonia (480-486)
Study Design: Time-series
Statistical Analyses:
1. Poisson GEE or asthma,
URI, all respiratory
2. Poisson GLM for pneumonia
and COPD
Age Groups Analyzed:
Primary Analysis : All Ages
Secondary Analvsis : 2-1 8 vr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.6 (0.5) ppm
Range (Min, Max): (0.6, 4.1)
Copollutant: correlation
PM10:r = 0.74
N02 (1-h max): r = 0.47
N02(24-h avg.): r = 0.66
S02 (1-h max): r = 0.15
S02 (24-h avg.): r= 0.32
Pollutant: CO
Averaging Time: 1-h max
Mean (SD) unit: 1.8 (1.2) ppm
Range (10th, 90th): (0.5, 3.4)
Copollutant: NR
Increment: 0.6 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
High Utilization: 1.04 (0.93-1.16);!
Low Utilization: 1.1 5 (1.05-1.28);!
All: 1.1 0(1 .02-1.19);!
Increment: 1.0 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
Health Condition
All respiratory illnesses : 1 .01 1 (1 .004-1 .019); 0-2
URI:
1 .01 2 (1 .003-1 .021 ) ; 0-2 / 1 .066 (1 .045-1 .087) ; 0-1 3
Asthma:
1 010(0999-1 022) '0-2
1.076 (1.047-1 .105); 0-1 3
Pneumonia:
1.009 (0.996-1 .021); 0-2
1.045 (1.011-1.080);0-13
COPD:
1.026 (1.004-1 .048); 0-2
1.032 (0.975-1 .092); 0-13
RR for asthma and exposure to CO for children age 2-18:
1.019 (1.004-1.035); 0-2
RR for all respiratory illnesses and CO exposure for all
ages
AQS (1 /1 /93- 8/31 /OO): 1.011 (1.004-1.019); 0-2
AQS (8/1 /98- 8/31 /OO): 1.010 (1.000-1.021); 0-2
ARIES (8/1/98- 8/31/00): 1.018 (1.003-1.033); 0-2
September 2009
C-58
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Sauerzapf et
al. (2009,180082)
Period of Study:
Jan 2006-Feb 2007
Location:
Norfolk county,
England
Design
Hospital Admissions
Health Outcome: COPD
Study Design: case-crossover
Statistical Analyses: Logistic
Regression
Aae Grouns Analyzed:
Concentrations
Averaging Time: 24h
Mean (SD) unit:
Control days: 194.46 (80.93)
Case days: 204.73 (11 9.97)
Range (min, max):
Control days: 105.20, 408.10
Effect Estimates (95% Cl)
Increment: 10 ug/m3
Lags examined: 0-8
OR Estimate [Lower Cl, Upper Cl]; lag:
Unadjusted: 1 .01 0 (1 .001 , 1 .01 9); lag 0-7
18+ yr (90% of patients 60+ yr) Case days: 108.70, 432.20
Sample Description:
1050 COPD admissions
CoPollutant:
NO, N02, NOX, 03
Adjusted: 1.015 (1.005,1.025); lag 0-7
Unadjusted: 1.013 (1.001,1.025); lag 1-8
Adjusted: 1.018 (1.005,1.031); lag 1-8
* Control days= 7 days prior to
admission; Case days= day of
admission
Author: Sheppard et
al. (1999,086921)
Period of Study:
1987-1994
Location:
Seattle, WA
Author: Slaughter et
al. (2005. 073854)
Period of Study:
1/1995-6/2001
Location:
Spokane, WA
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed: <65 yr
Hospital Admissions & ED Visits
Health Outcome (ICD9):
Respiratory causes (460-519)
Asthma (493); COPD (491,
492, 494, 496) Acute
respiratory tract infections not
including colds and sinusitis
(464-466, 490)
Study Design: Time-series
Statistical Analvses:
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1831 ppb
IQR (25th, 75th): (1277, 2201)
CoPollutant: correlation
PM10: r = 0.83; PM25:r= 0.78;
PM10-2.5: r = 0.56; 03:r = -0.18;
S02: r = 0.24
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (5th, 95th): (1.25, 3.05)
Copollutant: correlation
PM1:r = 0.63
PM2.5:r = 0.62
PM10:r = 0.32
PM10-2.5:r=0.32
Increment: 924 ppb
% Increase (Lower Cl, Upper Cl); Lag
CO: 6% (3, 9); 3
CO,PM2.5: 5% (1,8); 3
Increment: 1.0 ppm
Relative Risk (Lower Cl, Upper Cl);
ED Visits
All Respiratory Illnesses
Age Group:AII Ages:
0.99 (0.96-1 .02); 1/1. 01 (0.98-1.04);
1.03 (1.00-1 .06); 3
Asthma
Age Group:AII Ages:
1.00 (0.95-1 .06); 1/1. 01 (0.96-1.07);
1 nfi 11 nn 1 11\-1
lag:
2
2
Poisson GLM, Natural Splines
Age Groups Analyzed:
All ages,
Adults
COPD
Age Group: Adults:
0.92(0.85-1.00);1/0.99(0.91-1.08);
1.01 (0.93-1.10); 3
HospitalAd missions:
All Respiratory Illnesses
Age Group:AII Ages:
0.99 (0.95-1.02); 1/1.00 (0.96-1.04);
0.99 (0.96-1.03); 3
Asthma
Age Group:AII Ages:
1.02(0.92-1.13);1/1.06(0.96-1.17);
1.00 (0.91-1.11); 3
COPD
Age Group: Adults:
0.94(0.86-1.03);1/1.04(0.95-1.13);
0.97 (0.88-1.06); 3
September 2009
C-59
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Steibetal.
(2000, 011675)
Period of Study:
7/1992-3/1996
Location:
Saint John,
Canada
Author: Sunetal.
(2006, 090768)
Period of Study:
1/2004-12/2004
Location:
Taiwan
Author: Tenias etal.
(2002, 026077)
Period of Study:
1/1994-12/1995
Location:
Valencia, Spain
Design
ED Visits
Health Outcome (ICD9):
Asthma; COPD; Respiratory
infections; All respiratory
illnesses
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: All
ages
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Cross-sectional
Statistical Analyses:
Pearson correlation analysis
Age Groups Analyzed:
<16; 16-55
ED Visits
Health Outcome (ICD9):
COPD (491 , 492, 494, 496)
Study Design: Time-series
Statistical Analyses:
Concentrations
Pollutant: CO
Averaging Time:
24-h avg
1 -h max
Mean (SD) unit:
All yr: 0.5 (0.3) ppm
May-September: 0.6 (0.3) ppm
Allyr:1.6(1.1)ppm,
May-September: 1 .7 (0.9) ppm
Range (Min, Max): NR
Copollutant: correlation
H2S: r = -0.10; N02:r = 0.68;
03: r = -0.05; S02:r = 0.31;
TRS: r = 0.07; PM10:r = 0.28;
PM25: r = 0.27; H+:r = 0.23;
S042-: r = 0.27; CoH:r= 0.55
Pollutant: CO
Averaging Time: Monthly
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Pollutant: CO
Averaging Time:
24-h avg
1 -h max
Mean (SD) unit:
24-h avg
Effect Estimates (95% Cl)
Increment: 0.5 & 1 .7 ppm
Al% Increase (Lower Cl, Upper Cl); lag:
I Respiratory Illnesses
Increment: 0.5 ppm
All Ypar 1 4(V 7
Mil ICdl. O.tU, /
Increment: 1.7 ppm
May- September: -5.70
Increment: NR
Correlation Coefficient:
Asthma
Age Group:
<16'0653
16-55:0.425
Increment: 1 mg/m3
Relative Risk (Lower Cl, Upper Cl); Lag
24-h avg
All Year: 1.074 (0.998-1156);!
Cold Months: 1.070 (0.991-1.156); 1
Warm Months: 1 .129 (0.960-1 .329); 1
2. Sensitivity: GAM, LOESS
Age Groups Analyzed: >14yr
All yr: 3.1 mg/m3
Warm Months: 2.5 mg/m3
Cold Months:3.7 mg/m3
1-havg
Allyr:6.7 mg/m
Warm Months: 5.4 mg/m3
Cold Months:8.0 mg/m
Range (Min, Max):
24-h avg: (0.9,7.1)
1-h max: (1.6,17.2)
Copollutant: correlation
S02: r = 0.734; N02:r = 0.180;
03:r = -0.517
1 -h max
All Year: 1.039 (1.014-1.066);!
Cold Months: 1.037 (1.010-1.064); 1
Warm Months: 1.058 (0.994-1.127); 1
All Year: sinusoidal terms:
1.039 (1.010-1.066) ;1
All Year: humidity and temperature variables:
1.040(1.014-1.067);!
All Year: GAM, LOESS:
1.042(1.019-1.066);!
Author: Thompson et
al. (2001,073513)
Period of Study:
1/1993-12/1995
Location:
Belfast,
Northern Ireland
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
Children
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Warm Season: 0.57 (0.41) ppm
Cold Season: 0.74 (0.73) ppm
IQR (25th, 75th):
Warm Season: (0.3,0.7)
Cold Season: (0.4, 0.8)
Copollutant: correlation
S02 (log): r = 0.64;
PM10 (log): r= 0.57;
03: r =-0.52; NOX (log): r = 0.74;
NO (log): r = 0.71 ;N02:r = 0.69
Increment: NR
Relative Risk (Lower Cl, Upper Cl); lag:
Temperature included in the model:
1.04 (1.00-1.09); 0 /1.07 (1.02-1.12); 0-1
1.06 (1.00-1.12); 0-2 /1.07 (1.00-1.14); 0-3
Warm Season: 1.06 (0.98-1.16); NR
Cold Season:1.07(1.01-1.14); NR
Adjusted for benzene level:
0.92 (0.83-.02); 0-1 avg.
Note: The increment the study uses to calculate effect
estimates is a doubling in CO levels, but The study did not
provide this value.
September 2009
C-60
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Tolbertetal.
(2007, 090316)
Period of Study:
1/1993-12/2004
Location:
Atlanta, GA
Author: Trapasso and
Keith (1999, 180127)
Period of Study:
1/1994-12/1994
Location:
Bowling Green, KY
Author: Tsai et al.
(2006, 089768)
Period of Study:
1996-2003
Location:
Kaohsiung, Taiwan
Author: Vigottietal.
(2007, 090711)
Period of Study:
1/2000-12/2000
Location:
Pisa, Italy
Design
ED Visits
Health Outcome (ICD9):
Respiratory Disease: Asthma
(493,786.07,786.09);COPD
(491, 492, 496); URI (460-465,
460.0, 477); Pneumonia (480-
496);Bronchiolitis
(466.1,466.11,466.19))
Study Design: Time-series
Statistical Analyses:
Poisson GLM
Age Groups Analyzed: All
ages
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
Spearman Rank Correlation
Coefficient
Age Groups Analyzed: All
ages
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
ED Visits
Health Outcome (ICD9):
Respiratory Disease: Asthma
(493); Dry cough (468); Acute
bronchitis (466)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: <10yr;
>65yr
Concentrations
Pollutant: CO
Averaging Time: 1-h max
Mean (SD) unit: 1.6 ppm
Range (Min, Max): (0.1, 7.7)
Copollutant: correlation
PM10: r = 0.51 ;03:r = 0.27;
N02: r= 0.70; S02:r = 0.28;
Coarse PM:r= 0.38; PM25:r = 0.47;
S04:r=0.14;EC:r=0.66;
OC: r= 0.59;TC:r = 0.63;
OHC:r = 0.29
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.77 ppm
Range (Min, Max): (0.23, 1.72)
Copollutant:
PM10
S02
N02
03
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.5 (0.7) ug/m3
Range (Min, Max): (0.3, 3.5)
Copollutant: correlation
N02:r=0.62
PM10:r = 0.70
Effect Estimates (95% Cl)
Increment: 1.22 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Respiratory Diseases: 1 .01 6 (1 .009-1 .022); 3
Note: The study only provides results of the multi-pollutant
models in figures, not quantitatively.
Increment: NR
Correlation Coefficient (lag)
COMean:r = 0.19;0
CO Mean: r = 0.27; 1
CO Mean: r = 0.21; 2
CO Max: r = 0.26; 0
CO Max: r = 0.36; 1
CO Max: r = 0.24; 2
Increment: 0.29 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
OR for getting asthma and exposure to various pollutants
for all ages at either <25°C or > 25°C
CO
<25°C: 1.41 4 (1.300-1 .537); 0-2
a 25°C: 1.222 (1.1 38-1 .31 2); 0-2
CO, PM10
<25°C: 1.251 (1.1 25-1 .393); 0-2
>25°C: 1.1 78 (1.088-1 .274); 0-2
CO, S02
<25°C: 1.207 (1.076-1 .354); 0-2
>25°C: 1.290 (1.1 88-1 .400); 0-2
CO, N02
<25°C: 0.916 (0.807-1 .039); 0-2
>25°C: 1.249 (1.1 27-1 .384); 0-2
C0,03
<25°C: 1.396 (1.282-1 .520); 0-2
>25°C: 1.1 95 (1.1 13-1 .284); 0-2
Increment: 1mg/m3
% Increase (Lower Cl, Upper Cl); Lag
Age Group
<10: 18.60% (-6.90 to 51. 10); 1
>65: 26.50% (3.40-54.80); 4
September 2009
C-61
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Villeneuve et
al. (2006,091179)
Period of Study:
1995-2000
Location:
Toronto, ON,
Canada
Author: Xirasagar et
al. (2006. 093267)
Period of Study:
1998-2001
Location:
Taiwan
Author: Yang etal.
(2007, 092848)
Period of Study:
1996-2003
Location:
Taipei,
Taiwan
Author: Yang etal.
(2007, 092847)
Period of Study:
1996-2003
Location:
Taipei, Taiwan
Design Concentrations
Physician Visits Pollutant: CO
Health Outcome (ICD9): Averaging Time: 24-h avg
Allergic rhinitis (177)
a v ' Mean (SD) unit: 1.1 (0.4) ppm
Study Design: Time-series
Range (Mm, Max): (0.0, 2.2)
Statistical Analyses: Poisson
Ql_[¥] Copollutant:
PM2.5
Age Groups Analyzed: >65 yr PM 10
PM10-2.5
S02
N02
03
Hospital Admissions Pollutant: CO
Health Outcome (ICD9): Averaging Time: Monthly
Asthma (493) Mean (SD) unit: NR
Study Design: Cross-sectional Rgnge (Mi|% Max); NR
Statistical Analyses:
Spearman Rank Correlations Copollutant: NR
Age Groups Analyzed:
0-14 yr;<2yr; 2-5 yr;>5yr
Hospital Admissions Pollutant: CO
Health Outcome (ICD9): Averaging Time: 24-h avg
Asthma (493)
Mean (SD) unit: 1.33 ppm
Study Design: Case-crossover
Range (Mm, Max): (0.32, 3.62)
Statistical Analyses:
Conditional logistic regression Copollutant:
r M-|o
Age Groups Analyzed: S02
All ages N02
0.
3
Hospital Admissions Pollutant: CO
Health Outcome (ICD9): Averaging Time: 24-h avg
COPD: (490-492, 494, 496)
Mean (SD) unit: 1.33 ppm
Study Design: Case-crossover
Range (Min, Max):
Statistical Analyses: (0.32, 3.66) ppm
Conditional logistic regression
Copollutant:
Age Groups Analyzed: All PM10
ages S02
Effect Estimates (95% Cl)
Increment: 0.4 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
The study did not present quantitative results for CO.
Increment: NR
Correlation Coefficient (Lag)
Age Group:
<2:r= -0.208
2-5: r = -0.281
>5:r= -0.134
Increment: 0.53 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
CO
<25°C: 1.076 (1.01 9-1 .136); 0-2
>25°C : 1.277 (1.1 79-1 .383); 0-2
CO, PM10
<25°C: 1.050 (0.983-1 .122); 0-2
>25°C : 1.332 (1.21 6-1 .459); 0-2
CO S02
<25°C:1.131 (1.059-1 .207); 0-2
>25°C : 1.278 (1.1 74-1 .392); 0-2
CO, N02
<25°C: 0.915 (0.839-0.997); 0-2
>25°C : 1.1 77 (1.049-1 .320); 0-2
C0,03
<25°C: 1.1 69 (1.1 02-1 .240); 0-2
>25°C : 1.275 (1.1 77-1 .382); 0-2
Increment: 0.53 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
CO
<20°C: 0.975 (0.921 ,1.033); 0-2
>20°C: 1.227 (1.1 78-1 .277); 0-2
CO, PM10
<20°C: 0.925 (0.863-0.992); 0-2
>20°C: 1.1 77 (1.1 23-1 .235); 0-2
CO, S02
<20°C: 0.895 (0.832-0.962): 0-2
>20°C: 1.274 (1.219-1.331); 0-2
CO, N02
<20°C: 1.000 (0.910-1.099); 0-2
>20°C: 1.061 (0.998-1.129); 0-2
C0,03
<20°C: 0.935 (0.875-0.999); 0-2
>20°C: 1.234 (1.185-1.285); 0-2
September 2009
C-62
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Yang etal.
(2005, 090184)
Period of Study:
1/1994-12/1998
Location:
Vancouver,
Canada
Design
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65yr
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: .71 (0.28) ppm
Range (Min, Max): (0.30, 2.48)
Copollutant: correlation
03: r = -0.56
N02:r=0.73
S02: r = 0.67
PM10:r = 0.50
Effect Estimates (95% Cl)
Increment: 0.3 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
CO
1 .03 (1 .00-1 .06) ; 0 / 1 .04 (1 .01 -1 .08) ; 0-1
1 .05 (1 .01 -1 .09) ; 0-2 / 1 .05 (1 .00-1 .10); 0-3
1 .06 (1 .01 -1 .11) ; 0-4 / 1 .07 (1 .02-1 .1 2); 0-5
1.08 (1.02-1 .13); 0-6
MultiPollutant:
CO, 03: 1.11 (1.04-1 .18); 0-6
CO, N02: 1.04 (0.95-1 .14); 0-6
CO,S02:1.11 (1.01-1.22); 0-6
CO, PM10:1.02 (0.93-1.12); 0-6
CO, PM10,03, N02, S02:1.08 (0.96-1.22); O-l
CO, 03, N02, S02:1.10 (0.98-1.23); 0-6
Author: Yang etal.
(2003, 055621)
Period of Study:
1/1986-12/1998
Location:
Vancouver, BC,
Canada
Author: Yang etal.
(2004, 087488)
Period of Study:
6/1/1995-3/31/1999
Location:
Vancouver, Canada
Author: Zanobetti and
Schwartz (2006
090195)
Period of Study:
1995-1999
Location:
Boston, MA
Hospital Admissions
Health Outcome (ICD9):
Respiratory diseases (460-519)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
<3yr;>65yr
Hospital Admissions
Health Outcome (ICD9):
Respiratory diseases (460-
51 9); Pneumonia (480-486);
Asthma (493)
Study Design: Case-control
Statistical Analyses:
Pearson's correlation coefficient
Age Groups Analyzed: <3 yr
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-487)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.98 (0.54) ppm
IQR (25th, 75th): (0.62, 1.1 6)
Copollutant: correlation
03: r = -0.52
CoH
N02
S02
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.70 (0.30) ppm
IQR (25th, 75th): (0.50, 0.80)
Copollutant: correlation
PM10: r = 0.46; PM15:r= 0.24;
PMio-2.5: r = 0.33; 03: r = -0.53;
N02: r= 0.74; S02:r= 0.61
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
IQR (25th, 75th): (0.39, 0.60)
Copollutant: correlation
PM25: r = 0.52; BC:r = 0.82;
N02: r = 0.67; 03:r= -0.30
Increment: 0.54 ppm
Odds Ratio (Lower Cl, Upper Cl); Lag
OR for respiratory diseases and exposure to various
pollutants for people <3 and > 65
Age Group: <3
CO alone: 1.04(1.01-1.07);!
C0,03: 1.04(1.01-1.07);!
CO, 03, CoH, N02, S02: 1 .02 (0.96-1 .08);
Age Group: > 65
CO alone: 1.02(1.00-1.04);!
C0,03: 1.02(1.00-1.04);!
CO, 03, CoH, N02, S02: 0.96 (0.93-1 .00);
1
1
This study did not present quantitative results for CO.
Increment: 0.475 ppm
% Increase (Lower Cl, Upper Cl); lag:
5.45 (1.10, 9.51); 0
5.12 (0.83, 9.16); 0-1
September 2009
C-63
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-6 Studies of long-term CO exposure and respiratory morbidity.
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Goss etal.
(2004, 055624)
Period of Study:
1999-2000
Location: U.S.
Health Outcome: Lung function
(FEV1 , Cystic fibrosis pulmonary
exacerbation)
Study Design: Cohort
Statistical Analyses:
Logistic regression
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit:
0.692 (0.295) ppm
IOR (9Kth 7KthV lt\ 4R fl R?
Increment: 1 .0 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Two or More Pulmonary Exacerbations During 2000
1.02(0.85-1.22)
\
Population:
11,484 cystic fibrosis patients
Age Groups Analyzed: >6 yr
Copollutant: NR
Author: Guo et al.
(1999.010937)
Period of Study:
10/1995-5/1996
Location: Taiwan
Health Outcome: Asthma
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population:
331,686 non-smoking children
Age Groups Analyzed:
Middle-school children
(mean age: 13.8 yr)
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 853 (277) ppb
Range (Min, Max): (381,1610)
Copollutant: NR
Increment: 326 ppb
% Increase (Lower Cl, Upper Cl);
Boys
Physician-diagnosed asthma:
1.17% (0.63-1.72)
Questionnaire-diagnosed asthma:
1.10% (0.45-1.75)
Girls
Physician-diagnosed asthma:
0.84% (0.45-1.22)
Questionnaire-diagnosed asthma:
1% (0.44-1.56)
September 2009
C-64
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Hirschetal.
(1999.003537)
Period of Study:
Population:
9/1995-6/1996
Air:
4/1994-4/1995
Location:
Dresden, Germany
Health Outcome:
Asthma symptoms in the past 12
mo (wheeze, morning cough);
Doctor's diagnosis (asthma,
bronchitis); Lung function
(bronchial hyperresponsiveness
(BHR), FEV1 <85% pred., FEF25-
75% <70% pred.)
Study Design: Cross-sectional
Statistical Analyses:
Multiple logistic regression
Population:
5-7:2,796; 9-11:2,625
Age Groups Analyzed:
5-7 and 9-11 yr
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit: 0.69 mg/m3
Range (Min, Max): (0.32,1.54)
Copollutant: NR
Increment: 0.2 ug/m3
Prevalence Odds Ratio (Lower Cl, Upper Cl); lag:
Symptoms in the past 12 mo: Wheeze
Home Exposure
Age Groups: 5-7; 9-11:1.05 (0.93-1.18)
Home/School Exposure
Age Groups: 9-11:1.02 (0.85-1.22)
Morning Cough
Home Exposure
Age Groups: 5-7; 9-11:1.12 (1.01-1.23)
Age Group: 9-11:1.13 (0.98-1.3)
Doctor's diagnosis: Asthma
Home Exposure
Age Groups: 5-7; 9-11:1.07 (0.94-1.21)
Age Groups: 9-11:1.16 (0.97-1.38)
Doctor's diagnosis: Bronchitis
AgeGroups:5-7;9-11:1.19(1.11-1.27)
Age Group:9-11:1.24(1.12-1.38)
Lung function: BHR
Age Groups: 5-7; 9-11:0.79 (0.63-0.99)
Age Group: 9-11:0.77 (0.6-0.99)
Lung function: FEV1 <85% pred.
Age Groups: 5-7; 9-11:1.09 (0.81-1.47)
Age Group: 9-11:1.01 (0.73-1.41)
Lung function: FEV25-75% <70% pred.
Age Groups: 5-7; 9-11:1.15 (0.94-1.39)
Age Group: 9-11:1.07 (0.86-1.34)
Symptoms in the past 12 mo: Wheeze
Age Groups: 5-7; 9-11
Atopicchildren:! (0.81-1.24)
Nonatopic children: 1.05 (0.83-1.31)
Morning cough
Age Groups: 5-7; 9-11
Atopic children: 1.03 (0.82-1.29)
Nonatopic children: 1.22 (1.05-1.41)
Doctor's diagnosis: Asthma
Atopic children: 1.05 (0.83-1.32)
Nonatopic children: 1.29 (1.05-1.59)
Doctor's diagnosis: Bronchitis
Age Groups: 5-7; 9-11
Atopic children:! (0.86-1.16)
Nonatopic children: 1.21 (1.1-1.33)
Notes: Atopic Children were defined as those children
with specific IgE to aeroallergens >0.7 kU-L-1;
Nonatopic Children were defined as those children with
specific IgE to aeroallergens £ 0.7 kU-L-1.
Author: Hwang et al.
(2006, 088971)
Period of Study:
2001
Location: Taiwan
Health Outcome:
Allergic rhinitis
Study Design: Cross-sectional
Statistical Analyses:
Two-stage hierarchical model
(logistic and linear regression)
Population:
32,143 Taiwanese school children
Age Groups Analyzed: 6-15 yr
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit: 664(153) ppb
Range (Min, Max): (416,964)
Copollutant: correlation
N0x:r = 0.88
03:r=-0.37
PM10:r = 0.27
S02:r=0.40
Increment: 100 ppb
Adjusted Odds Ratio (Lower Cl, Upper Cl); lag:
Physician-diagnosed allergic rhinitis
1.05(1.04-1.07)
CO, S02:1.04 (1.02-1.06)
CO, PM10:1.05 (1.03-1.07)
CO, 03:1.07 (1.05-1.09)
Male: 1.06 (1.03-1.08); Female: 1.05 (1.02-1.08)
Parental atopy: Yes: 1.05 (1.02-1.08)
Parental atopy: No: 1.06 (1.03-1.08)
Parental Education: <6:1 (0.91-1.09)
Parental Education: 6-8:1.07 (1.02-.12)
Parental Education: 9-11:1.05 (1.02-1.08)
Parental Education: > 12:1.06 (1.03-1.09)
ETS: Yes: 1.06 (1.03-1.08); ETS: No: 1.05 (1.02-1.08)
Visible Mold: Yes: 1.07 (1.03-1.11)
Visible Mold: No: 1.05 (1.03-1.07)
September 2009
C-65
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Hwang et al.
(2005, 089454)
Period of Study:
2001
Location: Taiwan
Health Outcome: Asthma
Study Design: Cross-sectional
Statistical Analyses:
Two-stage hierarchical model
(logistic and linear regression)
Population:
32,672 Taiwanese school children
Age Groups Analyzed: 6-15 yr
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 664(153) ppb
Range (Min, Max): (416,964)
Copollutant: correlation
N0x:r = 0.88
03:r=-0.37
PM10:r = 0.27
S02:r=0.40
Increment: 100 ppb
Adjusted Odds Ratio (Lower Cl, Upper Cl); lag:
Physician-diagnosed asthma: 1.045 (1.017-1.074)
CO, S02:1.066 (1.034-1.099)
CO, PM10:1.079 (1.047-1.112)
CO, 03:1.063 (1.1-1.474)
CO, S02,03:1.111 (1.074-1.15)
CO, PM10,03:1.119 (1.084-1.155)
Male: 1.49 (1.37-1.63); Female:!
Parental atopy: Yes :1
Parental atopy: No: 2.72 (2.5-2.97)
Parental Education :<6:1
Parental Education: 6-8:1.17 (0.9-1.52)
Parental Education: 9-11:1.61 (1.26-2.05)
Parental Education: > 12:2.43 J1.9-3.09)
ETS: Yes: 0.85 (0.78-0.92); ETS: No: 1
Visible Mold: Yes: 1.27 (1.16-1.4); Visible Mold: No: 1
Maternal smoking during pregnancy:
Yes: 1.18 (0.89-1.56)
Maternal smoking during pregnancy: No: 1
Cockroaches noted monthly:
Yes: 1.15 (1.03-1.29)
Cockroaches noted monthly: No: 1
Water damage: Yes: 0.96 (0.81-1.12)
Water damage: No:1
Author: Lee et al.
(2003, 049201)
Period of Study:
10/1995-5/1996
Location: Taiwan
Author: Mengetal.
(2007, 093275)
Period of Study:
11/2000-9/2001
Location:
Los Angeles County
and San Diego
County, California
Health Outcome:
Allergic rhinitis
Study Design: Cohort
Statistical Analyses:
Multiple logistic regression
Population:
331 ,686 non-smoking children
Age Groups Analyzed: 12-1 4 yr
Health Outcome: Asthma
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population:
1 ,609 physician-diagnosed
asthmatics
Age Groups Analyzed: > 18yr
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit: 853 (277) ppb
Range (Min, Max): (381, 1610)
Copollutant: NR
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: correlation
Traffic: r = -0.04; 03:r= -0.55;
PM10: r = 0.42; PM25:r = 0.52;
N02:r = 0.55
The study did not present quantitative results for CO.
The study did not present quantitative results for CO.
September 2009
C-66
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Mortimer et
al. (2008,122163)
Period of Study:
1989-2000
Location:
San Joaquin Valley,
CA
Health Outcome: Lung function
(FVC,FEV1,PEF, FEF25-75,
FEV1/FVC, FEF25-75/FVC,
FEF25, FEF75)
Study Design: Cohort
Statistical Analyses:
1 . DSA algorithm
2. GEE
Pollutant: CO
Averaging Time:
8-h max monthly mean
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant; correlation:
Increment: NR
Effect Size per IQR Increase in Pollutant (SE):
FEF25-75:
24-h avg CO exposure during 1st trimester
0.90% (0.0113)
FEV1/FVC
Daily max CO exposure during ages 0 to 3
-2.50% (0.001 6)
Population: 232 asthmatic children
Age Groups Analyzed: 6-11 yr
Lifetime
N02 (24-h avg): r= 0.68
03(8-h max):r= -0.40
PM10 (24-h avg): r= 0.05
Prenatal
CO (8-h max): r= 0.52
N02 (24-h avg): r= 0.37
03 (8-h max): r=-0.16
PM10 (24-h avg): r = -0.05
FEF25-75/FVC
24-h avg CO exposure during ages 0 to 6 and
diagnosed with asthma <2 yr old
-4.80% (0.0446)
FEF25
24-h avg CO exposure during ages 0 to 6 and
diagnosed with asthma <2 yr old plus 24-h avg PM10
exposure during 2nd trimester and mother smoked
when pregnant
-6.70% (0.015)
Coefficient (SE):
FVC
24-h avg CO exposure during 2nd trimester
-0.0878(0.0415)
FEF25-75
Lifetime 24-h avg CO exposure
-0.94454 (0.3975)
FEF25-75/FVC
-0.1090(0.0303)
FEV1/FVC
Prenatal 8-h max CO exposure: 0.1711 (0.0653)
Lifetime 1-h max CO exposure: -0.3242 (0.0919)
24-h avg CO exposure during ages 0-3 and diagnosed
with asthma <2yr old: -0.1814 (0.0599)
FEF25
24-h avg CO exposure during ages 0-6 and diagnosed
with asthma <2 yr old: -1.0460 (0.1953)
FEF75
Lifetime 8-h max CO exposure: -0.4214 (0.1423)
Author: Singh et al.
(2003, 052686)
Period of Study: NR
Location:
Jaipur, India
Health Outcome: Lung function
Study Design: Panel study
Statistical Analyses: Parametric
statistical methods
Population:
Campus panel: 142
Commuter panel: 158
Age Groups Analyzed: -20 yr
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit:
Roadside: 3,175 ug/m3
Campus: 2,150 ug/m
Range (Min, Max): NR
CoPollutant: NR
The study did not present quantitative results for CO.
September 2009
C-67
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Sole et al.
(2007, 090706)
Period of Study:
Location:
Sao Paulo West, Sao
Paulo South, Santo
Andre, Curitba, &
Porto Alegre, Brazil
Author: Wang et al.
(1999.008105)
Period of Study:
10/1995-6/1996
Location:
Kaohsiung and
Pintong, Taiwan
Author: Wilhelm et al.
(2008, 191912)
Period of Study:
2000-2001
Location:
Los Angeles County
or San Diego County,
California
Design
Health Outcome: symptoms of
asthma, rhinitis, & eczema
Study Design: panel
Statistical Analyses: Logistic
Regression
Age Groups Analyzed:
13-1 4 yr
Sample Description:
Health Outcome: Asthma
Study Design: Cross-sectional
Statistical Analyses:
Multiple logistic regression
Population:
165,173 high school students
Age Groups Analyzed: 11-16yr
Health Outcome: asthma
symptoms/ED visit/HA
Study Design: panel
Statistical Analyses: Logistic
regression
Age Groups Analyzed:
0-1 7 yr
Concentrations
Averaging Time: annual
Mean (SD) unit:
Sao Paulo West: 7.70 ppm
Sao Paulo South: 7.50 ppm
Santo Andre: 9. 80 ppm
Curitba: 7.90 ppm
Porto Alegre: 1.51 ppm
Range (min, max):
MD
NK
Copollutant:
N02, S02, 03
Pollutant: CO
Averaging Time: Annual median
Median (SD) unit: 0.80 ppm
Range (Min, Max): NR
Copollutant: NR
Averaging Time: annual
Mean (SD) unit:
1 .0 ppm
Range (min, max):
0.34,1.8
CoPollutant: correlation
03: r= -0.67
Effect Estimates (95% Cl)
Increment: Risk in relation to center w/ lowest annual
mean (Porto Alegre = ref)
OR Estimate [Lower Cl, Upper Cl]:
Lags examined: NR
Current Wheezing:
Sao Paulo West: 1.26 (1.11, 1.42)
Sao Paulo South: 1.03 (0.91, 1.18)
Santo Andre: 1.36 (1.20, 1.56)
Curitba:1.05 (0.93, 1.19)
Severe Asthma:
Sao Paulo West: 1.20 (0.95, 1.50)
Sao Paulo South: 0.59 (0.45, 0.78)
Santo Andre: 0.62 (0.48, 0.81)
Curitba: 0.64 (0.50, 0.82)
Nighttime Coughing:
Sao Paulo West: 1.06 (0.95, 1.17)
Sao Paulo South: 0.93 (0.84, 1 .03)
Santo Andre: 0.91 (0.82,1.02)
Curitba:0.99 (0.89, 1.10)
Rhinoconjunctivitis:
Sao Paulo West: 1.31 (1.15,1.15)
Sao Paulo South: 0.73 (0.64, 0.85)
Santo Andre: 0.85 (0.74, 0.97)
Curitba:1.10(0.96,1.25)
Severe Rhinits:
Sao Paulo West: 1.01 (0.91,1.49)
Sao Paulo South: 0.68 (0.59, 0.77)
Santo Andre: 0.73 (0.64, 0.83)
Curitba: 1.03 (0.91, 1.1 6)
Eczema:
Sao Paulo West: 1.45 (1.20, 1.74)
Sao Paulo South: 1.03 (0.85, 1.25)
Santo Andre: 1.03 (0.85, 1.25)
Curitba:0.90 (0.75, 1.10)
Flexural Eczema:
Sao Paulo West: 1.42 (1.15, 1.76)
Sao Paulo South: 0.71 (0.56,0.91)
Santo Andre: 0.68 (0.53, 0.87)
Curitba: 0.73 (0.57, 0.92)
Severe Eczema:
Sao Paulo West: 1.08 (0.86, 1.35)
Sao Paulo South: 0.42 (0.31 , 0.56)
Santo Andre: 0.38 (0.28, 0.51)
Curitba: 0.30 (0.22, 0.41)
Increment: NR
Adjusted Odds Ratio (Lower Cl, Upper Cl); lag:
CO Concentrations: <0.80 ppm: 1 .0
CO Concentrations > 0.80 ppm: 1 .23 (1 .19-1 .28)
Multivariate analysis with variables for exercise,
smoking, alcohol, incense use, ETS: 1.15 (1.1-1.2)
Increment: NR
OR Estimate [Lower Cl, Upper Cl] ; lag :
Lags examined: NR
No associations observed between asthma symptom
outcome measures (no results shown)
Sample Description:
612 children who reported a
physician diagnosis of asthma at
some point in their lives
PM10:r=0.41
PM2.5: r= 0.60
N02: r= 0.57
traffic density: r= 0.02
September 2009
C-68
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-7 Studies of short-term CO exposure and mortality.
Study
Author: Anderson et al.
(2001.017033)
Period of Study:
10/1994-12/1996
Location:
West Midlands,
United Kingdom
Author: Bellini et al. (2007,
097787)
Period of Study: 1996-2002
Location:
15 Italian cities
Author: Berglind et al.
(2009, 190068)
Period of Study: 1992-2002
Location: Augsburg,
Germany; Barcelona, Spain;
Helsinki, Finland; Rome,
Italy; Stockholm, Sweden
Design
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses:
PniQQnn HAM
rUlooUII UMIVI
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Meta-analysis
Statistical Analyses:
Pniccnn f^l M
rUlooUII OLIVI
Age Groups Analyzed: All
ages
Health Outcome: Mortality
Study Design: Cohort
Statistical Analyses: Poisson
regression analysis
Age Groups Analyzed: £35 yr
Sample Description: First-
time Ml patients
Concentrations
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit: 0.8 (0.7) ppm
Range (Min, Max): (0.2, 10.0)
Copollutant correlation:
PM10: r= 0.55; PM25:r = 0.54;
PM10-2.5: r = 0.10; BS:r= 0.77;
S042-: r = 0.17; N02:r = 0.73;
03: r= -0.29; S02:r = 0.49
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
so.
O \J2
N02
03
PM10
Averaging Time: 24-h
Mean (SD) unit: Median calculated
from daily 24-h means:
Augsburg: 0.85
Barcelona: 0.75
Helsinki: 0.36
Rome: 1.66
Stockholm: 0.38
Range (IQR): Augsburg: 0.43
Barcelona: 0.75
Helsinki: 0.36
Rome: 1.11
Stockholm: 0.38
Copollutant: NR
Effect Estimates (95% Cl)
Increment: 1 .0 ppm
% Increase (Lower Cl, Upper Cl); lag:
All-cause
0.8% (-0.6 to 2.2); 0-1
Cardiovascular
2.5% (0.4-4.6); 0-1
Respiratory
1.2% (-2.1 to 4.6); 0-1
Increment: 1 mg/m3
% Increase (Lower Cl, Upper Cl); lag:
All-cause
1.19% (0.61-1 .72); 0-1
Respiratory
0.66% (-1.46 to 2.88); 0-1
Cardiovascular
0.93% (-0.10 to 1.77); 0-1
Increment: 0.2 mg/m3
% Change in Daily Nontrauma Deaths [Lower Cl,
Upper Cl]: Mean of Lag 0 and 1 : 2.61 (-0.26-5.56)
Mean of Lag 0-4: 3.82 (1.00-6.72)
Mean of Lag 0-14: 4.92 (2.11-7.81)
Lags examined: 0, 1,4, 14
CO had a trend towards or positive associations with
all cities for 2-day mean effects on daily mortality.
CO was associated with risk for the 5-day avg. The
strongest association was observed for the 1 5-day
avg.
September 2009
C-69
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Biggeri et al. (2005,
087395)
Period of Study: 1990-1999
Location:
8 Italian Cities (Turin, Milan,
Verona, Bologna, Ravenna,
Florence, Rome, and
Palermo)
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519); Cardio-
respiratory
Study Design: Meta-analysis
Statistical Analyses: Poisson
GLM, cubic splines
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit:
Turin, 1991-1994:5.8 mg/m3
Turin, 1995-1998:4.0 mg/m3
Milan, 1990-1994:5.9 mg/m3
Milan, 1995-1997:4.0 mg/m3
Verona, 1995-1999:2.5 mg/m3
Ravenna, 1991 -1995:1.8 mg/m
Bologna, 1996-1998:2.4 mg/m3
Florence, 1996-1998:2.7 mg/m3
Rome, 1992-1994:6.5 mg/m3
Rome, 1995-1997:5.4 mg/m3
Palermo, 1997-1999:2.1 mg/m3
Range (Min, Max):
Turin, 1991-1994: (NR, 24.7)
Turin, 1995-1998: (NR, 19.8)
Milan, 1990-1994: (NR, 26.5)
Milan, 1995-1997: (NR, 12.3)
Verona, 1995-1999: (NR, 10.2)
Ravenna, 1991-1995: (NR, 7.0)
Bologna, 1996-1998: (NR, 11.1)
Florence, 1996-1998: (NR, 8.7)
Rome, 1992-1994: (NR, 22.3)
Rome, 1995-1997: (NR, 18.5)
Palermo, 1997-1999: (NR, 8.0)
Copollutant: NR
Increment: 1.0 mg/m
% Increase (Lower Cl, Upper Cl); lag:
Non-accidental
Fixed: 0.93 (0.50-1.36); 0-1
Random: 0.93 (0.50-1.36); 0-1
Cardiovascular
Fixed: 1.29 (0.62-1.96); 0-1
Random: 1.29 (0.62-1.96); 0-1
Respiratory
Fixed: 2.44 (0.74-4.17); 0-1
Random: 2.47 (0.14-4.85); 0-1
Author: Bolter etal. (2002,
011922)
Health Outcome (ICD9):
Mortality
Period of Study: 1991-1993 Study Design:
Longitudinal study
Location:
Sao Paulo, Brazil
Statistical Analyses:
State space model
Age Groups Analyzed: > 65 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: TSP; N02; 03; S02
Increment: NR
p(SE):
Model 1:0.0053 (0.0036)
Model 2:0.0046 (0.0028)
Model 3:0.0040 (0.0028)
Model 4:0.0032 (0.0028)
Author: Bremner et al.
(1999.007601)
Period of Study:
1/1992-12/1994
Location:
London, U.K.
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Poisson,
cubic splines
Age Groups Analyzed:
All ages
0-64 yr
>65yr
65-74 yr
>75yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.4) ppm
Range (Min, Max): (0.2,5.6)
Copollutant:
N02;
03;
S02;
PM10;
BS
Increment: 0.8 ppm
% Increase (Lower Cl, Upper Cl); lag:
All-cause
Age Group:
All ages: 0.9% (-0.2 to 2.0); 1
0-64:1.2% (-1.0 to 3.5) ;1
> 65:0.8% (-0.4 to 1.9); 2
65-74:0.8% (-1.2 to 2.8); 3
> 75:0.9% (-0.4 to 2.2); 2
Respiratory
Age Group:
All ages: 2.0% (-0.3 to 4.5); 3
0-64:7.8% (0.2-15.9); 3
> 65:0.7% (-1.7 to 3.2); 3
65-74:7.5% (2.1-13.2); 3
>75:2.3%(-0.5to5.3);0
Multipollutant:
CO, S02:1.90% (0.18-3.64); 3
CO, PM10:1.25% (0.04-2.47); 3
CO, BS: 2.41% (-0.65 to 5.57); 3
Cardiovascular
Age Group:
All ages: 1.4% (-0.1 to 3.0); 1
0-64:2.1% (-1.7 to 6.0); 2
> 65:1.1% (-0.4 to 2.8); 2
65-74:2.4% (-0.6 to 5.5); 2
> 75:1.9% (0.0-3.9); 2
Multipollutant:
CO, N02:2.55% (0.40-4.75);!
CO, 03:3.98% (0.85-7.21);!
CO, PM10:0.62% (-0.59 to 1.85); 1
CO, BS: 1.29% (-1.53 to 4.19); 1
September 2009
C-70
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Burnett et al. (2000,
010273)
Period of Study: 1986-1 996
Location:
8 Canadian cities
Design
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800)
Study Design: Time-series
Statistical Analyses:
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.9 ppm
Range (Max): 7.2 ppm
Effect Estimates (95% Cl)
Increment: 0.9 ppm
% Increase (t-value); lag:
Temporally filtered daily non-accidental mortality
(days in which PMio data available)
CO: 0.4 (0.4); 0; 2.0(2.3);1
1. Single-pollutant models:
Poisson GAM, LOESS
2. Multi-pollutant models:
Principal component regression
analysis
Age Groups Analyzed:
All ages
Copollutant: correlation
03:r=-0.05
S02: r = 0.42
PM25:r = 0.44
PM10-2.5:r = 0.29
PM10:r=0.45
CO, PM25:-0.7(-0.7);0;1.1(1.1);1
CO, PM10-2.5:0.1 (0.2);0; 1.8 (2.1); 1
CO, PM™:-0.5 (-0.6); 0; 1.2(1.3); 1
Daily filtered non-accidental mortality
Single-pollutant model: 2.1 (2.1)
Multi-pollutant models:
Model 1: CO, PM25, PM10-2.5, 03, N02, S02:0.7
(1.9)
Model 2: CO, S04, Ni, Fe, In, 03, N02:0.7 (1.7)
Author: Burnett et al. (2004,
086247)
Period of Study: 1981 -1999
Location:
12 Canadian cities
Author: Cakmak et al.
(2007, 091170)
Period of Study:
1/1997-12/2003
Location:
Chile-7 cities
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800)
Study Design: Time-series
Statistical Analyses:
1 . Poisson, natural splines
2. Random effects regression
model
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular diseases (390-
459); Respiratory diseases
(460-519)
Studv Desicin: Time-series
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .02 ppm
Range (Min, Max): NR
Copollutant:
N02;
03;
S02;
PM2.5;
PM10-2.5
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .29 ppm
Range (Min, Max): NR
f^rtnrtlh it<3nt ^nrmlatii'in •
Increment: 1 .02 ppm
% Increase (t-value); lag:
0.68% (3.12); 1
CO, N02: 0.07% (0.30); 1
Increment: 1 .29 ppm
% Increase (t-value); lag:
Non-accidental:
5.88% (6.42); 1; 9.39% (6.89); 0-5
CO+PM10+03+S02:6.13% (4.34); 1
Age Group: £64
Statistical Analyses: Poisson;
Random effects regression
model
Age Groups Analyzed:
All ages
<64yr
65-74 yr
75-84 yr
>85yr
03:r=-0.55 to-0.01
S02:r = 0.31 to 0.67
PM10:r = 0.49 to 0.82
Note: Correlations are between
pollutants for seven monitoring
stations.
4.10% (2.52); 1;/4.76% (2.19);0-5
Age Group: 65-74
6.24%(3.17);1;/8.12%(3.88);0-5
Age Group: 75-84
8.64% (4.82); 1;/13.12% (5.12); 0-5
Age Group: > 85
8.58% (4.45); 1;/13.20% (4.82);0-5
April-September
7.09% (4.02); 1; /9.65% (4.50); 0-5
October-March
5.45%(1.14);1;/7.80%(1.89);0-5
Cardiac
7.79% (4.56); 1;/11.22% (4.8);0-5
Respiratory
12.93% (5.78); 1;/21.31 % (6.34); 0-5
Author: Chock etal. (2000,
010407)
Period of Study:
1989-1991
Location:
Pittsburgh, PA
Health Outcome (ICD9):
Mortality: Respiratory (480-486,
490-496, 507); Cardiovascular
(390-448); Influenza (487)
Study Design: Time-series
Statistical Analyses:
Poisson GAM; Cubic B-spline
basis functions
Age Groups Analyzed:
All ages
<75yr
>75yr
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
PM •
r IVho,
PM2.5;
03;
9(V
OW2,
MO
\\\J2
Increment: NR
P(SE);lag:
Age Group: <75
CO alone: 0.0080(1. 56) ;0
PM10, CO: 0.0030 (0.48); 0
PM10,N02, CO: 0.0079 (1.14); 0
PM10, 03, S02, N02, CO: 0.072 (1 .02) ; 0
CO
-0.00738 (-1 .42); -3; / 0.00133 (0.23); -2;
-0.0021 9 (-0.38) ; -1 ; / 0.00809 (1 .48) ; 0;
-0.00129 (-0.22); 1;/0.00512 (0.90); 2;
-0.00974 (-1.87); 3
CO, PM10, 03, S02, N02
-0.01103 (-1 .48); -3; / -0.00097 (-0.13); -2;
0.00514 (0.67);-1;/0.00853(1.15);0;
-0 .00404 (-0 .52) ; 1 ;/ -0 .00296 (-0 .39) ; 2 ;
-0.00346 (-0.46); 3
Season
CO
Winter: 0.00539 (0.78); 0
Spring: 0.01 655(1. 90); 0
Summer: 0.00155 (0.14); 0
September 2009
C-71
DRAFT - DO NOT CITE OR QUOTE
-------
Study Design Concentrations Effect Estimates (95% Cl)
Fall: 0.00797(1.14); 0
CO, PM10
Winter:-0.00563 (-0.50);0
Spring: 0.01233 (0.99); 0
Summer:-0.00712 (-0.48); 0
Fall: 0.00661 (0.73) ;0
CO, PM10, 03, S02, N02
Winter:-0.01326 (-0.95); 0
Spring: 0.02501 (1.54);0
Summer: 0.01874 (0.92) ;0
Fall: 0.01011 (0.88); 0
Age Group:>75
CO Alone:-0.0035 (-0.67); 0
CO, PM10:-0.0104 (-1.67); 0
CO, PM10,N02:-0.0128 (-1.80); 0
CO, PM10, 03, S02, N02: -0.0144 (-1.99); 0
CO
-0.00025 (-0.05); -3;/ -0.00242 (-0.42); -2;
-0.00238 (-0.41); -1; / -0.00302 (-0.54); 0;
-0.00116 (-0.20); 1;/ -0.00508 (-0.88); 2;
-0.00251 (-0.48); 3
CO, PM10, 03, S02, N02
-0.00123 (-0.17); -3; / -0.00876 (-1.13); -2;
-0.00682 (-0.88); -1;/ -0.01248 (-1.66); 0;
-0.00672 (-0.86); 1;/-0.00181 (-0.23); 2;
-0.00515 (-0.69); 3
Season
CO
Winter:-0.00304 (-0.43);0
Spring: 0.00482(0.54) ;0
Summer: 0.01178(1.07); 0
Fall:-0.01011 (-1.43);0
CO, PM10
Winter:-0.02303 (-2.03);0
Spring:-0.00517 (-0.40);0
Summer: 0.00735 (0.50); 0
Fall:-0.01042 (-1.14); 0
CO, PM10, 03, S02, N02
Winter:-0.03370 (-2.41);0
Spring:-0.00652 (-0.39);0
Summer: 0.01258 (0.61); 0
Fall:-0.01250 (-1.07); 0
September 2009 C-72 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Cifuentes et al.
(2000, 010351)
Period of Study:
1988-1996
Location:
Santiago, Chile
Health Outcome (ICD9):
Mortality: All causes (non-
accidental) (<800)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, GAM with filtered
variables & GLM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 2.5 ppb
Range (5th, 95th): (0.6, 6.2)
Copollutant correlation:
PM25:r = 0.80
PM10-2.5:r = 0.47
S02:r=0.62
N02: r = 0.65
03:r=-0.01
Increment:
All yr: 2.5 ppm
Winter: 3.6 ppm
Summer: 1.3 ppm
Relative Risk (t-ratio); Lag
All Year
CO: 1.041 (7.2); 0-1
CO, PM25:1.025 (3.5); 0-1
CO, PMm-2.5:1.035 (4.9); 0-1
CO, S02:1.038 (6.0); 0-1
CO, N02:1.026 (3.9); 0-1
CO, 03:1.036 (4.8); 0-1
Winter
CO: 1.052 (5.9); 0-1
CO, PM25:1.025 (2.1); 0-1
CO, PM10-2.5:1.049 (4.3); 0-1
CO, S02:1.049 (5.0); 0-1
CO, N02:1.027 (2.6); 0-1
CO, 03:1.051 (4.4); 0-1
Summer
CO: 1.053 (6.0); 0-1
CO, PM25:1.053 (5.3); 0-1
CO, PM10-2.5:1.053 (5.3); 0-1
CO, S02:1.050 (5.2); 0-1
CO, N02:1.047 (5.2); 0-1
CO, 03:1.042 (3.6); 0-1
All Year
GAM model
CO: 1.041 (7.2); 0-1
CO, PM25, PM10-2.5,S02, N02,03:
1.032 (4.6); 0-1
GAM Filtered Variables
CO: 1.030 (4.3); 0-1
CO, PM25, PM10-2.5, S02, N02, 03:
1.022 (2.4); 0-1
GLM
CO: 1.023 (2.4); 0-1
CO, PM25, PM10-2.5, S02, N02, 03:
1.013 (1.1); 0-1
Author: Conceicao et al.
(2001.016628)
Period of Study:
1994-1997
Location:
Sao Paulo, Brazil
Health Outcome (ICD9):
Mortality: Respiratory diseases
(460-519)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: <5 yr
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit:
Total: 4.4 (2.2) ppm
1994:5.1 (2.4) ppm
1995:5.1 (2.4) ppm
1996:3.9 (2.0) ppm
1997:3.7 (1.6) ppm
Range (Min, Max): NR
Copollutant:
PM10;S02;03
Increment: NR
p(SE);lag:
CO: 0.0306 (0.0076); 2
CO, S02, PM10,03:0.0259 (0.0116); 2
Model 1: Pollutant concentration:
0.0827 (0.0077); 2
Model2:1+loess(time):
0.0285 (0.0074); 2
Model 3:2+loess(temperature)+humidity:
0.0309 (0.0076); 2
Model 4:3+nonrespiratory counts:
0.0306 (0.0076); 2
Model 5:4+autoregressive parameters:
0.0292 (0.0118); 2
Author: De Leon et al.
(2003, 055688)
Period of Study:
1/1985-12/1994
Location:
New York, NY
Health Outcome (ICD9):
Mortality: Circulatory (390-459);
Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
<75yr
>75yr
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 2.45 ppm
IQR (25th, 75th): (1.80, 2.97)
Copollutant:
PM10;
03;
S02;
N02
The study did not present quantitative results for CO.
September 2009
C-73
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Dominici et al.
(2003, 056116)
Period of Study:
1987-1994
Design
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental); Cardiovascular;
Respiratory
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Effect Estimates (95% Cl)
Increment: 1 ppm
% Increase (Lower Cl, Upper Cl); Lag
CO
Location:
90 U.S. cities (NMMAPS)
Study Design: Time-series
Statistical Analyses:
1. GAM with S-PLUS default
convergence criteria
2. GAM with more stringent
convergence criteria
3. Poisson GLM with natural
cubic splines
Age Groups Analyzed:
All ages
Range (Min, Max): NR
Copollutant:
03;N02;S02;CO
0.08% (-0.18 to 0.34); 0
0.46% (0.18-0.73); 1
0.16% (-0.12 to 0.45); 2
Author: Fairleyetal. (1999,
000896)
Period of Study:
1989-1996
Location:
Santa Clara, CA
Author: Fischer et al. (2003,
043739)
Period of Study:
1986-1994
Location: The Netherlands
Health Outcome (ICD9):
Mortality: Respiratory;
Cardiovascular
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Pneumonia
(480-486) ;COPD (490-496);
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Pollutant: CO
Averaging Time:
24-h avg; Maximum 8-h avg
Median (SD) unit:
24-h avg: 1.4 (1.0) ppm
Maximum 8-h avg: 2.1 (1 .6) ppm
Range (Min, Max):
24-h avg: (0.0, 7.6)
Maximum 8-h avg: (0.2, 2.5)
Copollutant: correlation
PM10:r= 0.609;
PM25:r= 0.435;
PM10-2.5:r= 0.326;
COM: r= 0.736;
N03:r= 0.270;
S04: r = 0.146; 03:r = -0.215
Pollutant: CO
Averaging Time: 24-h avg
Median (SD) unit: 406 ug/m3
Range (Min, Max): (174, 2620)
Copollutant:
PM10;BS;03;N02;S02
Increment: 2.2 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
1980-1986
C0:1.04;0;
C0:1.05;1;
CO, COM: 1.00; 1;
CO, N03:1.03;
CO, N03, 03, COM: 1.00
1989-1996
C0'1 02' 0-
C0:1.04;1;
CO, PM25:0.98;
CO N03' 1 01 '
CO, N02,03,N03:1.06
Respiratory mortality: CO: 1 .08; 1
Cardiovascular mortality: CO: 1 .04; 1
Increment: 1 ,200 ug/m3
Relative Risk (Lower Cl, Upper Cl); lag:
Cardiovascular
Age Group:
<45: 0.965 (0.750-1 .240); 0-6
45-64:1.029 (0.941-1 .125); 0-6
65-74:1.038 (0.972-1 .108); 0-6
> 75: 1.024 (0.984-1 .065); 0-6
Age Groups Analyzed:
<45yr
45-64 yr
65-74 yr
>75yr
COPD
Age Group:
<45:1.710 (0.852-3.435); 0-6
45-64:1.181 (0.850-1.640); 0-6
65-74:1.377 (1.147-1.654); 0-6
> 75:1.072 (0.963-1.193); 0-6
Pneumonia
Age Group:
<45:0.927 (0.463-1.856); 0-6
45-64:2.691 (1.509-4.800); 0-6
65-74:1.118(0.743-1.683);0-6
> 75:1.230 (1.090-1.389); 0-6
Author: Forastiere et al.
(2005, 086323)
Period of Study:
1998-2000
Location:
Rome, Italy
Health Outcome (ICD9): Pollutant: CO
Mortality : IHD (41 0-41 4)
Averaging Time: 24-h avg
Timeystra^r!ed case-crossover Mean (SD) unit: 2.4 (1 .0) mg/m3
Statistical Analyses: IQR (25th, 75th): (1 .7, 2.9)
Conditional logistic regression CopoNutant corre|atjon:
Age Groups Analyzed: >35 yr PNC: r = 0.89; PM10: r = 0.34;
N02: r = 0.54; S02: r = 0.52;
03:r=0.01
Increment: 1.2 mg/m3
% Increase (Lower Cl, Upper Cl); lag:
6.5% (1.0-1 2.3) ;0
4 7% (-09 to 10 7)' 1
2.6% (-3.0 to 8.5); 2
-0.1% (-5.5 to 5.5); 3
7.0% (0.8-1 3.7); 0-1
September 2009
C-74
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Forastiere et al.
(2007, 090720)
Period of Study:
1998-2001
Location: Rome, Italy
Health Outcome (ICD9):
Mortality: Malignant Neoplasms
(140-208); Diabetes Mellitus
(250); Hypertensive (401-405);
Previous AM I (410,412); IHD
(410-414); Conduction
disorders of the heart (426);
Dysrhythmia (427); Heart
Failure (428);Cerebrovascular
(430-438); Peripherical Artery
disease (440-448) ;COPD
(490-496)
Study Design: Time-stratified
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: >35 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
IQR (25th, 75th): NR
Copollutant:
PM10; PM2.5; NOX; Benzene
This study did not present quantitative results for
CO.
Author: Goldberg et al.
(2001.016548)
Period of Study:
1984-1993
Location: Montreal,
Quebec, Canada
Author: Goldberg et al.
(2003, 035202)
Period of Study: 1984-1 993
Location:
Montreal, Quebec, Canada
Author: Goldberg et al.
(2006, 088641)
Period of Study:
1984-1993
Location:
Montreal, Quebec, Canada
Author: Gouveia et al.
(2000, 012132)
Period of Study:
1991-1993
Location:
Sao Paulo, Brazil
Health Outcome (ICD9):
Mortality: Upper respiratory
diseases (472-478); Acute
Upper respiratory diseases
(460-465); Acute Lower
Respiratory (466, 480-487,
512,513,518,519)
Study Design: Time-series
Statistical Analyses: Poisson
GAM; LOESS
Age Groups Analyzed:
<65yr;>65yr
Health Outcome (ICD9):
Mortality: CHF (428)
Study Design: Time-series
Statistical Analyses: Poisson
GLM, natural splines
Age Groups Analyzed:
>65yr
Health Outcome (ICD9):
Mortality: Diabetes (250)
Study Design: Time-series
Statistical Analyses: Poisson,
natural splines
Age Groups Analyzed: > 65 yr
Health Outcome (ICD9):
Mortality: Respiratory;
Cardiovascular; All other
causes
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
>65yr
<5yr
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (0.1, 5.1)
Copollutant:
TSP; PM10; PM25; Sulfates; COH;
S02; N02; NO; 03
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (0.1, 5.1)
Copollutant:
PM2.5;Sulfate;S02;N02;03
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (0.1, 5.1)
Copollutant:
PM2.5;
Sulfate;
S02;
N02;
03
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit: 5.8 (2.1) ppm
Range (Min, Max): (1.3, 16.2)
Copollutant:
PMi0;S02; N02;03
The study did not present quantitative results for CO.
Increment: 0.50 ppm
% Increase (Lower Cl, Upper Cl); lag:
Daily mortality from CHF
-0.99% (-6.31 to 4.63) ;0
0.12% (-5.29 to 5.84); 1
-1.38% (-8.81 to 6.66); 0-2
Daily mortality among persons classified as having
CHF before death
2.10% (-0.24 to 4.49); 0
2.28% (-0.09 to 4.72); 1
2.86% (-0.46 to 6.29); 0-2
Increment: 0.50 ppm
% Increase (Lower Cl, Upper Cl); lag:
Daily mortality from diabetes
2.64% (-2.56 to 8.12); 0
6.54% (1.31-12.03);!
8.08% (1.03-15.62); 0-2
Daily mortality among persons classified as having
diabetes before death
1.15%(-1.69to4.07);0
1. 30% (-1.58 to 4.27) ;1
2.63% (-1.42 to 6.85) ;0-2
Increment: 5.1 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Age Group: All ages:
All-causes 1.012 (0.994-1. 031); 0
Age Group: >65
All-causes: 1 .020 (0.996-1 .046); 0
Respiratory: 0.981 (0.927-1 .037); 2
CVD: 1.041 (1 .007-1. 076) ;0
Age Group: <5
Respiratory: 1 .086 (0.950-1 .238); 0
Pneumonia: 1.141 (0.962-1 .321); 2
September 2009
C-75
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Gwynn et al. (2000,
074109)
Period of Study:
5/1988-10/1990
Location:
Buffalo, NY
Author: Hoeketal. (2001,
016550)
Period of Study:
1986-1994
Location: The Netherlands
Author: Hoeketal. (2000,
010350)
Period of Study:
1986-1994
Location: The Netherlands
Design
Health Outcome (ICD9):
Mortality: Respiratory (466,
480-486); Circulatory (401 -405,
410-414, 41 5-41 7); All non-
accidental causes (<800)
Study Design: Time-series
Statistical Analyses:
Poisson GLM
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Heart Failure (428);
Arrhythmia (427);
Cerebrovascular (430-436);
Thrombocytic (433, 434, 444,
452, 453); Cardiovascular
(390-448)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Pneumonia
(480-486) ;COPD (490-496);
Cardiovascular diseases (CVD)
(390-448)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant correlation:
H+: r =0.15; S042-:r= 0.24;
03' r= -0 23 S02'r = 0 11'
N02:r = 0.65
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
03;BS;PM10;S02;N02
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Netherlands: 457 ug/m3
Four Major Cities: 589 ug/m
Range (Min, Max):
Netherlands: (174, 2620)
Crtnr Uoirtr Pitiac- /OflO ylfiO-n
Effect Estimates (95% Cl)
Increment: NR
P(SE);lag:
Respiratory mortality: 0.032466 (0.053802); 0
Circulatory mortality: 0.039216 (0.026544); 3
Total mortality: 0.040214 (0.015205); 3
Increment: 120 ug/m3
Relative Risk (Lower Cl, Upper Cl); Lag
Total CVD mortality: 1 .026 (0.993-1 .060) ; 0-6
Ml and other IHD mortality:
1.050 (1.004-1 .099); 0-6
Arrhythmia: 1.062 (0.937-1 .203); 0-6
Heart failure mortality: 1 .1 09 (1 .01 2-1 .216); 0-6
Cerebrovascular mortality:
1.066 (1.029-1 .104); 0-6
Embolism, thrombosis: 1 .065 (0.926-1 .224); 0-6
Increment:
Single-day lag (1): 1 ,500 ug/m
Weekly avg (0-6): 1200 ug/m3
Relative Risk (Lower Cl, Upper Cl); Lag
CO
Four Major Cities: 1 .022 (0.995-1 .050) ; 1
Four Major Cities : 1 .044 (1 .008-1 .082) ; 0-6
Netherlands w/o Major Cities : 1 .040 (1 .020-1 .060) ; 1
Age Groups Analyzed: All
ages
Four Major Cities: (202, 4621)
Copollutant correlation:
PM10: r= 0.64; BS:r = 0.89;
03: r=-0.48; N02:r = 0.89;
S02: r= 0.65; S042-:r = 0.55;
N03-:r = 0.58
Netherlands w/o Major Cities:
1.051 (1.026-1.076); 0-6 avg
Entire Netherlands: 1.035 (1.018-1.052); 1
Entire Netherlands: 1.046 (1.025-1.068); 0-6
CVD: 1.044 (1.012-1.077); 0-6
COPD: 1.194 (1.099-1.298); 0-6
Pneumonia: 1.276 (1.143-1.426); 0-6
Winter: 1.038 (1.013-1.063); 0-6
Summer: 1.199 (1.108-1.296); 0-6
Multi-pollutant model
CO, PM10
Total mortality: 0.969 (0.914-1.028); 0-6
CVD: 1.005 (0.918-1.101); 0-6
BS.CO
Total mortality: 0.980 (0.933-1.030); 0-6
CVD: 0.927 (0.860-0.999); 0-6
CO, S042-
Total mortality: 0.990 (0.951 -1.030); 0-6
CVD: 0.999 (0.939-1.063); 0-6
Author: Honda etal. (2003,
193774)
Period of Study:
1976-1990
Location:
Tokyo, Japan
Health Outcome (ICD9):
Mortality:
Total (non-accidental) (<800)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24- h avg
Median (SD) unit: 1 .6 ppm
Range (Min, Max): (0,6.8)
Copollutant correlation:
NO: r = 0.403; N02:r = 0.415;
Oxidant:r = 0.396; S02:r = 0.675
Increment: NR
Rate Ratio (Lower Cl, Upper Cl); lag:
CO concentration
<1 .1 ppm: 1 .00 (reference category)
1.1-1.6 ppm:1.017 (1.009,1.026)
1.6-2.2 ppm: 1.031 (1.020,1.041)
>2.2 ppm: 1.051 (1.039,1.063)
September 2009
C-76
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Hong et al. (2002,
035060)
Period of Study:
1/1991-12/1997
Location:
Seoul, Korea
Author: Hong etal. (1999,
011195)
Period of Study:
1/1995-12/1995
Location:
Inchon, Korea
Author: Hong et al. (2002,
024690)
Period of Study:
1/1995-12/1998
Location:
Seoul, Korea
Author: Hong etal. (1999,
008087)
Period of Study:
1/1995-8/1996
Location:
Inchon, South Korea
Author: Keatinge et al.
(2001.017063)
Period of Study:
1976-1995
Location:
London, England
Design
Health Outcome (ICD9):
Mortality: Hemorrhagic and
ischemic stroke (431 -434)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Cardiovascular (400-
440); Respiratory
(460-51 9); Non-accidental
causes (<800)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Stroke (160-1 69)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory;
Cardiovascular
Study Design: Time-series
Statistical Analyses: Poisson
GAM' LOESS
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Non-accidental
causes (<800)
Study Design: Time-series
Statistical Analyses: Single
and multiple delay regression
Age Groups Analyzed:
All ages
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1.44 (0.70) ppm
Range (Min, Max): (0.430, 5.14)
Copollutant:
TSP;S02;N02;03
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .7 (0.8) ppm
Range (Min, Max): (0.3, 5.1)
Copollutant:
en • MO • O
GW2, l*1*-^, ^3
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .2 (0.5) ppm
Range (Min, Max): (0.4, 3.4)
Copollutant: correlation
PM10: r= 0.22; N02:r = 0.64;
S02: r = 0.90; 03:r = -0.35
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 15.2 (7.1) ppb
Range (Min, Max): (2.9, 51 .2)
Copollutant:
PM10;N02;S02;03
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
S02; PM10
Effect Estimates (95% Cl)
Increment: 0.76 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
1.06(1.02,1.09);!
Multipollutant:
CO, TSP: 1.07(1.03,1.11);!
CO, N02: 1.06 (1.00,1. 11); 1
CO, S02: 1.05 (1.01,1.10);!
CO, 03:1.09 (1.05, 1.13);!
Increment: 1 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Total mortality:
0.993 (0.950, 1.037); 0-4
Cardiovascular mortality:
0.965 (0.892, 1.044); 0-4
Increment: 0.3 ppm
% Increase (Lower Cl, Upper Cl); lag:
CO: 2.2% (0.4, 4.1); 2
CO (stratified by PM10 concentration):
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Kettunen et al.
(2007, 091242)
Period of Study: 1998-2004
Location:
Helsinki, Finland
Health Outcome (ICD10):
Mortality: Stroke (160-161,
I63-I64)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, penalized thin-
plate splines
Age Groups Analyzed:
>65yr
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Median (SD) unit:
Cold Season: 0.5 mg/m3
Warm Season: 0.4 mg/m3
Range (Min, Max):
Cold Season: (0.1, 2.4)
Warm Season: (0.1,1.1)
Copollutant: correlation
Cold Season:
PM2.5: r = 0.32; UFP:r= 0.47
Warm Season:
PM2.5: r = 0.24; UFP:r= 0.39
Increment: 0.2 mg/m
% Increase (Lower Cl, Upper Cl); lag:
Cold Season
0.47 (-3.29 to 4.39); 0; /-0.63 (-4.39 to 3.28); 1;
-2.69 (-6.46 to 1.24); 2; / -0.19 (-3.93 to 3.69); 3
Warm Season
3.95 (-3.78 to 12.30); 0; / 8.33 (0.63 to 16.63); 1;
6.97 (-0.59 to 15.11); 2; / 7.54 (-0.05 to 15.71); 3
Author: Klemm et al. (2004,
056585)
Period of Study:
8/1998-7/2000
Location:
Fulton County and DeKalb
County, GA (ARIES)
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(390-459); Respiratory
(460-51 9); Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson GLM, natural cubic
splines
Age Groups Analyzed:
<65yr;>65yr
Pollutant: CO
Averaging Time: 1-h max
Median (SD) unit:
1,310 (939.13) ppb
Range (Min, Max):
(303.58, 7400)
Copollutant:
PM2s' PMio-2 5' 03' N02' S02'
EC;OC;S04;
Oxygenated HCs;
NMHCs;N03
Increment: NR
P(SE);lag:
Quarterly Knots: 0.00002 (0.00001); 0-1
Monthly Knots: 0.00002 (0.00001); 0-1
Biweekly Knots: 0.00001 (0.00002); 0-1
Acid-
Author: Knoxetal. (2008,
193776)
Period of Study: 1996-2004
Location: 352 English local
authorities
Health Outcome: Mortality
Study Design: Cross-sectional
Statistical Analyses: Linear
regression
Age Groups Analyzed: NR
Sample Description: Data
from Oxford Cancer
Intelligence Unit
Averaging Time: NR
Meuan (SD) nit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Significant (p<0.01) correlations (r) between CO and
diseases: Lung cancer: 0.28, Stomach cancer: 0.20,
Oesophagus cancer: -0.20, Prostate cancer: -0.25,
Brain cancer:-0.24, Melanoma:-0.24, Hodgkin's:-
0.19, Peripheral vascular disease: 0.15, Stroke:
0.16, Rheumatic heart disease: 0.27, Peptic ulcer:
0.28, Diabetes: 0.17, COPD: 0.25, Asthma: 0.14,
Pneumonia: 0.44, Multiple sclerosis: -0.16,
Motorneurone disease: -0.24, Parkinsons disease: -
0.15
Significant (p<0.01) socially standardized
correlations between diseases and exposures: Lung
cancer: 0.25, Stomach cancer: 0.18, RHD: 0.19,
Pneumonia: 0.37, COPD: 0.17, Peptic ulcer: 0.16
Lags examined: NR
September 2009
C-78
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Kwon et al. (2001 ,
016699)
Period of Study:
1994-1998
Location:
OCUUI, rVUIcd
Design
Health Outcome (ICD9):
Mortality :CHF (428);
Cardiovascular (390-459)
Study Design:
1 . Time-series
2. Bi-directional case-crossover
Statistical Analyses:
LPoissonGLM, LOESS
2. Conditional logistic
regression
Age Groups Analyzed:
<55yr
55-64 yr
65-74 yr
75-84 yr
>85yr
Concentrations
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 1 2.4 ppb
Range (Min, Max): (4.1, 38.0)
Copollutant correlation:
PM10: r= 0.713; N02:r = 0.744;
S02: r = 0.843; 03:r = -0.367
Effect Estimates (95% Cl)
Increment: 0.59 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
From GAM approach
CHF patients: 1 .054 (0.991-1 .121); 0; 0
General Population : 1 .022 (1 .017-1 .029) ; 0
From case-crossover design
CHF patients: 1.033 (0.946-1. 127); 0
General Population: 1 .007 (0.997- .016); 0
Modifiers and CHF patients (case-crossover design)
Gender
Male: 1.025 (0.890-1. 180); 0
Female: 1.035 (0.925-1. 157); 0
Age Group:
<75:0.948(0.890-1.180);0
> 75: 1.11 6 (0.989-1. 258) ;0
Time from admission to death
4 or less wk: 1.088 (0.907-1. 306); 0
>4 wk: 1.01 7 (0.920-1.1 24) ;0
Total mortality: 1 .033 (0.946-1.1 27) ;0
Cardiovascular mortality: 1 .033 (0.920-1 .160);
Cardiac death : 1 .052 (0.91 9-1 .204) ; 0
0
Two-pollutant model in CHF patients (case-
crossover design)
CO alone: 1.054 (0.991-1.121); 0
CO, PM10:1.096 (0.981-1.224); 0
CO, N02:1.022 (0.932-1.122); 0
CO, S02:1.014 (0.909-1.131); 0
CO, 03:1.056 (0.992-1.124); 0
Author: Lee etal. (2007,
093042)
Period of Study:
1/2000-12/2004
Location:
Seoul, Korea
Author: Lipfert etal. (2000,
004088)
Period of Study:
5/1992-9/1995
Location:
Philadelphia, PA, three
nearby suburban
Pennsylvania counties, and
three nearby New Jersey
counties
Author: Lippmann et al.
(2000, 011938)
Period of Study:
Health Outcome (ICD10):
Mortality: Non-accidental (AOO-
R99)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Respiratory
(460-51 9); Cardiac (390-448);
Cancer; Other causes (<800)
Study Design: Time-series
Statistical Analyses:
Stepwise regression
Age Groups Analyzed:
<65yr
>65yr
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Circulatory (390-459);
Respiratory (460-51 9)
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit:
w/ Asian dust days: 0.92 (0.42) ppm
w/o Asian dust days:0.92 (0.41) ppm
Asian dust days only: 1.00
(0.47) ppm
Range (Min, Max): NR
Copollutant: PM10; N02; S02; 03
Pollutant: CO
Averaging Time:
24-h avg; 1-h max
Mean (SD) unit:
Camden:
24-h avg: 0.75 (0.40) ppm
Philadelphia:
24-h avg: 0.63 (0.40) ppm
1-h max: 1.44 (1.04)
Range (Min, Max):
Camden: (0.1 0,3.8)
Philadelphia:
24-h avg: (0.1 0,3.3)
1-h max: (0.0, 7.8)
Copollutant:
NO;N02;03;S02;S042-;
PM10;PM2.5
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Increment: 0.54 ppm
% Increase (Lower Cl, Upper Cl); lag:
Model with Asian Dust Days:3.3% (2.5-4.1); 1
Model without Asian dust days: 3.3% (2.5-4.2); 1
Increment: NR
Attributable Risk; lag:
Peak CO
All-cause
Philadelphia: 0.0054; 0-1
4 Pennsylvania Counties: 0.0081 ; 0-1
Pennsylvania + NJ: 0.0085; 0-1
CO
All seven counties in Pennsylvania and New Jersey
All ages
Respiratory: -0.0067; Cardiac: 0.0131 ;
Other: 0.0078
All-cause:
<65: 0.0148; 0-1 ;> 65: 0.0054; 0-1
Joint model with CO
Philadelphia: 0.0059; 0-1
4 Pennsylvania Counties: 0.0089; 0-1
Pennsylvania + NJ: 0.0096; 0-1
Cardiac: 0.01 35; 0-1;
Other causes: 0.0084
<65: 0.0154; 0-1;
> 65: 0.0060; 0-1
Increment:
1985-1990: 11 .5 ppm; 1992-1994: 8.4 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
September 2009
C-79
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
1985-1990
1992-1994
Location:
Detroit, Ml and Windsor, ON
Study Design: Time-series
Statistical Analyses:
Poisson GLM
Age Groups Analyzed:
>65yr
1985-1990:0.9 ppm
1992-1994:0.72 ppm
Range (5th, 95th):
1985-1990: (.46,1.61)
1992-1994: (0.36,1.2)
Copollutant correlation:
1985-1990
PM10: r= 0.35; TSP:r = 0.28;
TSP-PM10:r = 0.02;
TSP-S042-:r = 0.18;
03: r= -0.22; S02:r = 0.36;
N02:r = 0.58
1992-1994
PM10: r= 0.38; PM25:r = 0.38;
PM10-2.5:r=0.24;
H+: r= 0.16; S042-:r= 0.32;
03: r = 0.16; S02:r = 0.42;
N02: r = 0.68
1985-1990
Total Mortality:
0.9842 (0.9667-1
1.0103(0.9926-1
1.0075(0.9898-1
1.0145(0.9967-1
0.9968 (0.9789-1
1.0105(0.9925-1
1.0134(0.9954-1
1.0003(0.9823-1
1.0152(0.9971-1
1.0053(0.9873-1
.002) ;0
.0284);!
.0254); 2
.0326); 3
.0151); 0-1
.0288); 1-2
.0317); 2-3
.0187); 0-2
.0336); 1-3
.0236); 0-3
Circulatory Mortality:
0.9818 (0.9574-1.0068);0
0.9991 (0.9745-1.0243) ;1
0.9980 (0.9735-1.0232); 2
1.0088 (0.9841-1.0341); 3
0.9888 (0.9640-1.0144); 0-1
0.9981 (0.9732-1.0237); 1-2
1.0042 (0.9792-1.0298); 2-3
0.9900 (0.9650-1.0157); 0-2
1.0029 (0.9777-1.0287); 1-3
0.9944 (0.9692-1.0202); 0-3
Respiratory Mortality;
0.9644 (0.9042-1.0287) ;0
1.0142 (0.9518-1.0808);1
1.0483 (0.9845-1.1164); 2
1.0468 (0.9828-1.1149); 3
0.9868 (0.9248-1.053); 0-1
1.0372 (0.9730-1.1056); 1-2
1.0554 (0.9904-1.1246); 2-3
1.0088 (0.9457-1.0762); 0-2
1.0466 (0.9817-1.1158); 1-3
1.0205 (0.9569-1.0884); 0-3
Total minus respiratory and circulatory mortality:
0.9939 (0.9668-1.0217); 0
1.0278(1.0001-1.0562);!
1.0178 (0.9902-1.0461); 2
1.0227 (0.9948-1.0514); 3
1.0127 (0.9860-1.0412); 0-1
1.0269 (0.9989-1.0556); 1-2
1.0249 (0.9968-1.0538); 2-3
1.0172 (0.9893-1.0458); 0-2
1.0322 (1.0041-1.0612); 1-3
1.0229 (0.9950-1.0516); 0-3
1992-1994
Total Mortality
0.9933 (0.9636-1.024);0
1.0162(0.9860-1.0473);!
1.0116 (0.9816-1.0426); 2
0.9947 (0.9648-1.0254); 3
1.0056 (0.9756-1.0366); 0-1
1.0165 (0.9864-1.0476); 1-2
1.0038 (0.9739-1.0476); 2-3
1.0098 (0.9796-1.0409); 0-2
1.0104 (0.9862-1.0414); 1-3
1.0064 (0.9755-1.0382); 0-3
Circulatory Mortality
1.0076 (0.9640-1.0531); 0
1.0307(0.9865-1.0768);!
1.0142 (0.9705-1.0598); 2
0.9523 (0.9102-0.9964); 3
1.0229 (0.9788-1.0688); 0-1
1.0267 (0.9827-1.0727); 1-2
0.9802 (0.9375-1.0248); 2-3
1.0243 (0.9801-1.0726); 0-2
0.9987 (0.9553-1.0441); 1-3
1.0019 (0.9573-1.0487); 0-3
Respiratory Mortality
0.9894 (0.8912-1.0984) ;0
0.9474(0.8521-1.0533);!
0.9652 (0.8682-1.0732); 2
0.9931 (0.8934-1.1040); 3
September 2009
C-80
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
0.9626 (0.8668-1.0691); 0-1
0.9485 (0.8535-1.0541); 1-2
0.9752 (0.8775-1.0838); 2-3
0.9555 (0.8802-1.0615); 0-2
0.9567 (0.8607-1.0635); 1-3
0.9584 (0.9604-1.0675); 0-3
Total minus respiratory and circulatory mortality:
0.9769 (0.9332-1.0227);0
1.0135(0.9682-1.0609);!
1.0195 (0.9747-1.0664); 2
1.0429 (0.9974-1.0905); 3
0.9940 (0.9494-1.0406); 0-1
1.0197 (0.9746-1.0670); 1-2
1.0371 (0.9918-1.0845); 2-3
1.0045 (0.9596-1.0515); 0-2
1.0353 (0.9896-1.0831); 1-3
1.0215 (0.9749-1.0702); 0-3
Author: Maheswaran et al.
(2005, 090769)
Period of Study:
1994-1998
Location:
Sheffield, United Kingdom
Health Outcome (ICD9):
Mortality:CHD (410-414)
Study Design: Ecological
Statistical Analyses: Poisson
Age Groups Analyzed:
>45yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
NOX;
PM10
Notes: Quintiles represent the
following mean CO concentrations
and category limits:
5:482 ug/m3 (> 455)
4:443 ug/m3 (> 433 to <455)
3:426ug/m3(>419to<433)
2:405ug/m3(>387to<419)
1:360 ug/m3 (<387)
Increment: NR
Rate Ratios (Lower Cl, Upper Cl):
CO
Adjusted for sex and age
Quintile:
5 (highest): 1.24 (1.14,1.36)
4:1.30(1.19,1.41)
3:1.15(1.05,1.25)
2:1.08(0.99,1.17)
1: (lowest): 1.00
CO
Adjusted for sex, age, deprivation, and smoking
Quintile:
5 (highest): 1.05 (0.95,1.16);
4:1.16(1.06,1.28);
3:1.04(0.95,1.14);
2:1.03(0.94,1.13);
1 (lowest): 1.00
CO
Adjusted for sex, age, deprivation, and smoking
(spatially smoothed using a 1 km radius)
Quintile:
5 (highest): 1.07 (0.96,1.18);
4:1.13(1.03,1.24);
3:1.04(0.95,1.14);
2:1.01 (0.92,1.10);
1 (lowest): 1.00
September 2009
C-81
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Maheswaran et al.
(2005, 088683)
Period of Study:
1994-1998
Location:
Sheffield, United Kingdom
Health Outcome (ICD9):
Mortality: Stroke deaths
(430-438)
Study Design: Ecological
Statistical Analyses: Poisson
Age Groups Analyzed:
>45yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Quintile:
5:482 |jg/m3;
4:443 |jg/m3;
3:426 |jg/m3;
2:405 |jg/m3;
1:360 |jg/m3
Range (Min, Max): NR
Copollutant correlation :
PM10: r= 0.88; N0x:r = 0.87
Notes: Quintiles represent the
following mean CO concentrations
and category limits:
5:482 ug/m5 (> 455)
4:443 |jg/m3 (> 433 to <455)
3:426ug/m3(>419to<433)
2:405ug/m3(>387to<419)
1:360 |jg/m3 (<387)
Increment: NR
Rate Ratios (Lower Cl, Upper Cl); lag:
RR for mortality and CO modeled outdoor air
pollution
Adjusted for sex and age
Quintile:
5 (highest): 1.35 (1.19,1.53);
4:1.40(1.24,1.58);
3:1.08(0.95,1.23);
2:1.10(0.97,1.24);
1 (lowest): 1.00
Adjusted for sex, age, deprivation, and smoking
Quintile:
5 (highest): 1.26 (1.10,1.46);
4:1.32(1.15,1.50);
3:1.07(0.93,1.22);
2:1.12(0.99,1.28);
1 (lowest): 1.00
Not spatially smoothed CO outdoor air pollution
Quintile:
5 (highest): 1.26 (1.10,1.46);
4:1.32(1.15,1.50);
3:1.07(0.93,1.22);
2:1.12(0.99,1.28);
1 (lowest): 1.00
Spatially smoothed using a 1-km radius
Quintile:
5 (highest): 1.16 (1.01,1.34);
4:1.22(1.07,1.39);
3:0.95(0.83,1.09);
2:0.97(0.85,1.11);
1 (lowest): 1.00
Author: Mar et al. (2000,
001760)
Period of Study:
1995-1997
Location:
Phoenix, AZ
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Cardiovascular (390-
449)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.5 (0.8) ppm
Range (Min, Max):
1995: (0.5, 4.0) ppm
1996: (0.3, 4.0) ppm
1997: (0.3, 3.7) ppm
Copollutant correlation:
PM25:r=0.85;
PM10:r=0.53;
PM10-2.5:r=0.34;
N02:r=0.87;
03:r=-0.40;
S02:r=0.53
Increment: 1.19 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Total Mortality (CO exposure):
1.06(1.02,1.09); 0;
1.05(1.01,1.09);!
Cardiovascular Mortality (CO exposure):
1.05(1.00,1.11);0;
1.10(1.04,1.15);!;
1.07(1.02,1.12);2;
1.07(1.02,1.12);3;
1.08(1.03,1.13);4
September 2009
C-82
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Moolgavkar et al.
(2000, 012054)
Period of Study:
1987-1995
Location:
Cook County, IL
Los Angeles County, CA
Maricopa County, AZ
Health Outcome (ICD9):
Mortality:
Circulatory (390-448);
Cardiovascular (390-429);
Cerebrovascular (430-448);
COPD (490-496); Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, spline smoother
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Median unit:
Cook county: 993 ppb
Los Angeles: 1347 ppb
Maricopa: 1240 ppb
Range (Min, Max):
Cook county: (224, 3912)
Los Angeles: (237, 5955)
Maricopa: (269,4777)
Copollutant correlation :
PM10:
Cook: r= 0.30;
LA: r = 0.45;
Maricopa: r= 0.20
N02:
Cook:r= 0.63;
LA: r = 0.80;
Maricopa: r= 0.66
S02:
Cook:r= 0.35;
LA: r = 0.78;
Maricopa: r= 0.53
03:
Cook:r= -0.28;
LA: r = -0.52;
Maricopa: r= -0.61
Increment: 1 ppm
% Change (Lower Cl, Upper Cl); lag:
CVD Mortality
Cook County
CO
-1.07 (-2.67,0.54); 0;/1.25 (-0.36,2.87); 1;
1.49 (-0.09,3.07); 2;/1.90 (0.32, 3.48); 3;
1.44 (-0.16,3.03); 4; / 0.72 (-0.89,2.32); 5
Los Angeles County
CO
3.47 (2.94, 4.00); 0;/ 3.93 (3.41, 4.46); 1;
4.08 (3.56, 4.60); 2;/ 3.76 (3.24, 4.28); 3;
2.91 (2.37, 3.44);4;/2.63 (2.09, 3.17);5
CO, PM10
2.27 (0.88, 3.66); O;/4.33 (2.96, 5.69); 1;
4.72 (3.38, 6.05); 2;/ 4.26 (2.90, 5.63); 3;
2.49 (1.10, 3.88); 4;/ 5.93 (4.60, 7.27); 5
COandPM25
0.43 (-1.35,2.20);0;/2.88 (1.16,4.60);!;
4.65 (2.93, 6.37); 2;/ 5.93 (4.20, 7.65); 3;
3.88 (2.13, 5.63); 4;/ 5.85 (4.12, 7.58); 5
Maricopa County
CO
0.81 (-0.79,2.39);0;/2.20 (0.61,3.79);!;
3.05 (1.49, 4.61); 2;/ 3.78 (2.27, 5.28); 3;
3.73 (2.27, 5.19); 4;/ 2.25 (0.76, 3.72); 5
COPD Mortality
Cook County
CO
-2.65 (-7.05,1.75); O;/2.80 (-1.60,7.19);!;
0.98 (-3.34,5.31); 2;/2.20 (-2.12,6.53); 3;
1.31 (-3.06,5.68); 4; /1.59 (-2.78,5.97); 5
Los Angeles County
CO
3.78 (2.31, 5.25); 0;/ 5.23 (3.78, 6.69); 1;
5.71 (4.26, 7.17);2;/5.42 (3.95, 6.89);3;
4.01 (2.51, 5.50); 4;/ 3.82 (2.31, 5.33); 5
Maricopa County
CO
1.29 (-2.19,4.76);0;/4.63 (1.17,8.09);!;
0.07 (-3.36,3.50); 2;/3.00 (-0.30,6.30); 3;
6.21 (3.02, 9.40); 4;/ 3.27 (0.04, 6.50); 5
Cerebrovascular Disease Mortality
Cook County
-0.41 (-3.30,2.47); O;/3.13 (0.23, 6.02); 1;
2.12 (-0.73,4.97); 2;/1.00 (-1.85,3.86); 3;
2.50 (-0.36,5.37); 4; /1.88 (-1.00,4.76); 5
Los Angeles County
3.31 (2.32, 4.31);0;/3.88 (2.89,4.87);!;
3.23 (2.25, 4.22); 2;/ 2.65 (1.66, 3.65); 3;
2.11 (1.11,3.12);4;/2.04(1.02,3.06);5
Maricopa County
0.26 (-2.65,3.16); O;/3.50 (0.60, 6.41); 1;
3.52 (0.66, 6.38); 2;/ 4.61 (1.85, 7.37); 3;
4.78 (2.10, 7.46); 4;/ 5.15 (2.45, 7.84); 5
Notes: Total Mortality effect estimates were not
presented quantitatively.
Author: Moolgavkar et al.
(2003, 051316)
Period of Study: 1987-1995
Location:
Cook County, Illinois & Los
Angeles County, California
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Circulatory (390-448)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All Ages
Pollutant: CO
Averaging Time: 24-h avg
Median unit:
Cook County: 993 ppb
LA County: 1347 ppb
Range (Min, Max):
Cook County: (224, 3912) ppb
LA County: (237, 5955) ppb
Increment: 1 ppm
% Increase (t-statistic); lag
Total Mortality Cook County
CO:
0.6% (1.2); 0;/2.5% (5.4); 1;/1.2% (2.6); 2;
1.5% (3.2); 3;/1.1 % (2.5); 4; / 0.6% (1.3); 5
CO, PM10:
-0.5% (-1.0);0;/2.2% (4.3); 1;/1.1% (2.2);2;
September 2009
C-83
DRAFT - DO NOT CITE OR QUOTE
-------
Study Design Concentrations Effect Estimates (95% Cl)
Copollutant correlation: 1.0% (1.9); 3;/1.1% (2.1); 4; /1.4% (2.7); 5
Cook County: Tota| Morta|ity Los Angeles County
NO2: r = u.bo; QQ.
?£' =-°n2-32c 1-3% <7-4); °'n•9% (10-5); 11/1.6% (8.9);2;
S°2-r - °--». 1.4% (8.1); 3;/1.0% (5.9); 4; / 0.7% (4.1); 5
rMio. r - U.oU
LA County: CO, PM10:
N02: r = 0.80; o% (0); 0;/2.2% (4.8); 1;/1.4% (3.1);2;
03: r = -0.52; 0.8% (1.8); 3;/ 0.7% (1.6); 4; /1.3% (3.0); 5
S02:r = 0.78;
PM10:r=0.45; CO, PM2.5:
PM25- r = 0 58 -0.1% (-1.5); O;/1.5% (2.5); 1;/2.4% (3.8); 2;
0.3% (0.5); 3;/1.6% (2.8); 4; /1.5% (2.6); 5
Total Mortality (Season-specific) Cook County
Spring (CO):
0.8% (0.9); O;/2.4% (2.9); 1;/0% (0); 2;
1.2% (1.5); 3; /0.8% J1.0); 4; / -0.1 % (-0.2); 5
Summer (CO):
1.2% (1.0); 0;/3.6% (3.0); 1;/4.2% (3.6); 2;
-0.3% (-0.2); 3;/-1.1 % (-1.0);4;/-0.7% (-0.6); 5
Fall (CO):
0% (0);3;/-o!s% (-0.6);4;'/-0.7% (-0.9);5
Winter (CO):
-0.7% (-1.0); 0;/1.8% (2.3); 1;/-0.2% (-0.3); 2;
0.5% (0.6); 3;/1.2% (1.5); 4; /1.0% (1.3); 5
Los Angeles County
Total Mortality (Season-specific)
Spring (CO):
3.6% (6.3); O;/3.5% (6.2); 1;/1.9% (3.4); 2;
0.6% (1.0);3;/-0.5% (-0.8);4;/-0.7% (-1.2); 5
Summer (CO):
3.0% (3.0); 0;/4.7% (4.6); 1;/5.2% (5.1); 2;
4.1 % J3.8); 3;/1.9% J1.8); 4; /1.4% J1.3); 5
Fall (CO):
1.8% (4.6);0;/2.0% (5.1); 1;/1.0% (2.6);2;
0.6% (1.5); 3;/ 0.4% (1.2); 4; / 0.2% (0.6); 5
Winter (CO):
0% (0); 0;/0.8% (2.5); 1;/0.9% (3.1);2;
1.0% (3.4); 3;/ 0.5% (1.7); 4; / 0.5% (1.6); 5
CVD Mortality Cook County
CO:
-1.1% (-1.5);O;/1.8% (2.5); 1;/1.5% (2.2);2;
1.6% (2.4); 3;/1.4% (2.1); 4; / 0.7% (1.0); 5
CO, PM10:
-2.1 % (-2.6); 0;/1.5% (1.8); 1;/1.4% (1.7); 2;
CVD Mortality Los Angeles County
CO:
1.6% (6.3);O;/1.9% (7.6); 1;/1.6% (6.6);2;
1.9% (8.2); 3;/1.6% J7.1); 4; /1.4% J6.1); 5
CO, PM10:
-0.8% (-1.2); 0;/1.9% (3.0); 1;/2.7% (4.3); 2;
1.3% (2.2); 3;/ 0.5% (0.9); 4; / 2.8% (4.7); 5
CO, PM25:
-2.2% (-2.7); 0;/1.5% (1.8); 1;/1.9% (2.0); 2;
1.9% (2.2); 3;/2.1 % (2.6); 4;/3.7%(4.5);5
CVD Mortality (Season Specifid) Cook County
Spring (CO):
0.7% (0.5); 0;/1.4% (1.1); 1;/0.3% (0.3); 2;
1.1 % (0.9); 3;/ 0.4% (3.1); 4; / 0.1% (0.6); 5
Summer (CO):
-2.6% (-1.4); 0;/2.5% (1.4); 1;/6.5% (3.7); 2;
0.9% (0.5); 3;/-1.9% (-1.1);4;/-1.0% (-0.6); 5
Fall (CO):
0% O); 3; /-0.8% (-0.7) ; / 4; 0% (0) ; 5
September 2009 C-84 DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Winter (CO):
-2.5% (-2.2); 0;/0.7% (0.6); 1;/0% (0); 2;
1.3% (1.1); 3;/ 0.8% (0.7); 4; / 0.4% (0.4); 5
Los Angeles County
CVD Mortality (Season-specific)
Spring (CO):
3.0% (3.7); 0;/3.3% (4.1); 1;/2.3% (2.9); 2;
0.7% (0.9);3;/-1.2% (-1.6);4;/0% (0);5
Summer (CO):
4.0% (2.8); 0;/5.2% (3.5); 1;/6.3% (4.3); 2;
5.0% (3.3); 3;/ 3.1% (2.0); 4; / 3.6% (2.3); 5
Fall (CO):
2.3% (4.2);0;/2.1% (3.7); 1;/1.1% (1.9);2;
1.2% (2.2); 3;/1.5% (2.9); 4; /1.0% (1.8); 5
Winter (CO):
0.3% (0.8);/0; 0.7% (1.7); 1;/0.8% (2.0); 2;
1.4% (3.4); 3;/1.0% (2.3); 4; /1.1 % (2.5); 5
Author: Ostroetal. (1999,
006610)
Period of Study:
1989-1992
Location:
Coachella Valley, California
Author: Penttinen et al.
(2004, 087432)
Period of Study:
1988-1996
Location:
Helsinki, Finland
Author: Peters et al. (2000,
001756)
Period of Study:
1982-1994
Location:
Northern Bavaria (Rural
Germany) and the Coal
Basin of the Czech Republic
Author: Rainhametal.
(2003, 053202)
Period of Study:
1980-1996
Location:
Toronto, ON, Canada
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-51 9);
Cardiovascular (393-440)
Study Design: Time-series
Statistical Analyses: Poisson
riAiu|- 1 OF^
VJrtlvl , LWCOO
Age Groups Analyzed: >50
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-51 9);
Cardiovascular (393-440)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, LOESS
Age Groups Analyzed:
All ages
15-64yr
65-74 yr
>75
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800) ; Cardiovascular (390-
459); Respiratory (460-51 9);
Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
(1) Poisson Regression Models
by logistic regression analyses
with a cubic function;
(2) Poisson GAM, natural
splines
Age Groups Analyzed:
All Ages
Health Outcome (ICD9):
Mortality: Cardiac (390-459);
Respiratory (480-51 9); Total
(non-accidental) (<800)
Study Design: Time-series
Statistical Analyses:Poisson
GAM, natural cubic splines
Age Groups Analyzed:
<65
>65
Pollutant: CO
Averaging Time: 1-h max
Mean (SD) unit: 1 .35 ppm
Range (Min, Max): (0, 6.0)
Copollutant correlation:
PM10: r= -0.18; 03:r= -0.47; N02:r =
0.65
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Median unit: 1 .2 mg/m3
Range (Min, Max): (0,12.4)
Copollutant correlation:
03: r= -0.46; N02:r = 0.59;
S02:r= 0.55; PM10:r= 0.45; TSP:
I- — n oc •
r- 02b,
TSP Blackness: r = 0.26
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Coal Basin: 0.58 (0.39) mg/m3
Northeast Bavaria:
0.88 (0.69) mg/m3
Range (Min, Max):
Coal Basin: (-0.1, 2.88)
Northeast Bavaria: (0.1 , 6.2)
Copollutant correlation:
S02: r= 0.37; TSP: r = 0.37; N02:r =
0.32; 03:r = -0.57; PM10:r = 0.44;
PM25: r = 0.42
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .0 (0.4) ppm
Range (Min, Max): (0.0, 4.0)
Copollutant: 03; N02; S02
Increment: NR
P(SE);lag:
CO: 0.0371 (0.01 57); 2
CO, PM10: 0.0300 (0.01 94); 2
Increment: 1 mg/m3
% Increase (Lower Cl, Upper Cl); lag:
Total Mortality
-1 .50% (-2.78, -0.22) ;0
0.15% (-1.09,1.39);!
-1.00% (-2.80, 0.81); 0-3
Cardiovascular Mortality
-2.48% (-4.30, -0.66) ;0
-0.84% (-2.61, 0.93); 1
-1 .87% (-4.43, 0.69); 0-
Respiratory Morality
-0.48% (-4.84, 3.87); 0
-0.14% (-4.43,4.15);!
-1.49% (-7.73, 4.74); 0-3
Increment: 1 mg/m3
Relative Risk (Lower Cl, Upper Cl); lag:
Coal Basin of the Czech Republic
Total Mortality:
1 .016 (0.998, 1 .035); 0; / 1 .016 (0.998, 1 .034); 1 ;
1 .013 (0.996, 1 .030); 2; / 1 .012 (0.995, 1 .028); 3
Northeast Bavaria
Total Mortality:
1 .014 (0.994, 1 .034); 0; / 1 .023 (1 .005, 1 .041); 1 ;
1 .013 (0.995, 1 .031); 2; / 1 .003 (0.985, 1 .021); 3
Cardiovascular Disease Mortality:
1 .018 (0.994, 1 .044); 0; / 1 .012 (0.987, 1 .038); 1 ;
1 .01 6 (0.991 , 1 .041) ; 2; / 1 .004 (0.980, 1 .029); 3
The study did not present quantitative results for CO.
September 2009
C-85
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Roemer et al.
(2001.019391)
Period of Study:
1/1987-11/1998
Location: Amsterdam
Health Outcome (ICD9): Pollutant: CO
Mortality: Total (non-accidental)
(<8oo) Averaging Time: 24-h avg
Study Design: Time-series Mean (SD) unit:
Air pollution background:
Statistical Analyses: 836 ug/m3
Poisson GAM Air pollution traffic: 1805 ug/m3
Age Groups Analyzed: Range (10th, 90th):
All ages Air pollution background:
(448, 1315) ug/nr
Air pollution traffic:
(727, 31 92) ug/m3
C^niinlliifeint1
Increment:
Lag 1 and 2: 100 ug/m3
Lag 0-6: 50 ug/m
Relative Risk (Lower Cl, Upper Cl); lag:
Total Population using Background sites
1.002(1.000-1.004);!;
1.001 (0.999-1 .003); 2;
1.001 (1.000-1 .003); 0-6
Traffic Population using Background Sites
1.003(0.997-1.008);!;
1.008 (1.003-1 .01 3); 2;
1.003 (0.999-1 .007); 0-6
BS;PM10;S02;N02;NO;03
Total population using Traffic Sites
1.000(1.000-1.001);!;
1.000 (0.999-1.001); 2;
1.000 (1.000-1.001); 0-6
Author: Samet et al. (2000,
013132)
Period of Study:
1987-1994
Location:
20 U.S. Cities: Los Angeles,
CA; New York, NY; Chicago,
IL; Dallas, TX; Houston, TX;
San Diego, CA; Anaheim,
CA; Phoenix, AZ; Detroit, Ml;
Miami, FL; Philadelphia, PA;
Minneapolis, MN; Seattle,
WA; San Jose, CA;
Cleveland, OH;
San Bernardino, CA;
Pittsburgh, PA; Oakland,
CA; Atlanta, GA;
San Antonio, TX
Health Outcome (ICD9):
Mortality: Cardiovascular (390-
459); Respiratory
(460-519); Other (non-
accidental) (<800)
Study Design: Time-series
Statistical Analyses:
Two-stage log linear regression
model
Age Groups Analyzed:
<65
65-74
>75
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Los Angeles: 15.1 ppm
New York: 20.4 ppm
Chicago: 7.9 ppm
Dallas: 7.4 ppm
Houston: 8.9 ppm
San Diego: 11.0 ppm
Anaheim: 12.3 ppm
Phoenix: 12.6 ppm
Detroit: 6.6 ppm
Miami: 10.6 ppm
Philadelphia: 11.8 ppm
Minneapolis: 11.8 ppm
Seattle: 17.8 ppm
San Jose: 9.4 ppm
Cleveland: 8.5 ppm
San Bernardino: 10.3 ppm
Pittsburgh: 12.2 ppm
Oakland: 9.1 ppm
Atlanta: 8.0 ppm
San Antonio: 10.1 ppm
Range (10th, 90th):
Los Angeles: (5.9, 28.3)
New York: (14.8, 27.6)
Chicago: (4.5,11.9)
Dallas: (3.6,12.0)
Houston: (4.0,14.2)
San Diego: (4.5, 20.5)
Anaheim: (3.7,25.2)
Phoenix: (5.4, 22.6)
Detroit: (3.2,11.1)
Miami: (6.5,15.9)
Philadelphia: (7.0,17.2)
Minneapolis: (7.0,17.0)
Seattle: (10.5,26.4)
San Jose: (1.7,21.3)
Cleveland: (3.7,13.8)
San Bernardino: (4.0,17.5)
Pittsburgh: (6.1,19.8)
Oakland: (2.9,17.0)
Atlanta: (3.2,14.3)
San Antonio: (4.1,17.3)
Copollutant correlation:
PM10: r= 0.45; 03:r =-0.19;
N02: r = 0.64; S02:r = 0.41
This study did not provide quantitative results for
CO.
Author: Samoli et al. (2007,
098420)
Period of Study:
1990-1997
Location:
19 European Cities
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Cardiovascular (390-
459)
Study Design: Time-series
Statistical Analyses:
Pollutant: CO
Averaging Time: 24-h avg
Mean Range (unit-mg/m3):
Athens: 6.1 ; Barcelona: 0.9; Basel:
0.6; Birmingham: 1 .0; Budapest: 5.1 ;
Geneva: 1 .5; Helsinki: 1 .2; Ljubljana:
Increment: 1 mg/m3
% Increase (Lower Cl, Upper Cl); lag:
Non-accidental mortality
8 Degrees of Freedom peryr
Fixed Effects :
CO: 0.59% (0.41-0.78); 0-1
September 2009
C-86
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
(APHEA2)
Poisson and two-stage
hierarchical model
Age Groups Analyzed:
All ages
1.6;London:1.4;Lyon:3.8;
Milano: 5.4; Netherlands: 0.6;
Prague: 0.9; Rome: 4.1; Stockholm:
0.8; Teplice: 0.7; Torino: 5.5;
Valencia: 4.1; Zurich: 1.2
Range (10th, 90th):
Athens: (3.5, 9.2)
Barcelona: (0.4,1.7)
Basel: (0.4,1.1)
Birmingham: (0.5,1.6)
Budapest: (3.3,7.4)
Geneva: (0.8, 2.6)
Helsinki: (0.7,1.9)
Ljubljana: (0.6, 3.0)
London: (0.7, 2.2)
Lyon:(2.0,6.0)
Milano: (2.9, 8.7)
Netherlands: (0.4,1.2)
Prague: (0.5,1.5)
Rome: (2.5, 5.9)
Stockholm: (0.5,1.2)
Teplice: (0.3,1.2)
Torino: (2.8, 9.1)
Valencia: (2.4, 5.9)
Zurich: (0.7,2.0)
Copollutant correlation:
PM10:r = 0.16 to 0.70
BS:r = 0.67 to 0.82
S02:r = 0.35 to 0.82
N02:r= 0.03 to 0.68
03:r=-0.25 to-0.65
CO, BS: 0.35% (-0.03 to 0.72); 0-1
CO, PM10:0.48% (0.24-0.72); 0-1
CO, S02:0.44% (0.21-0.67); 0-1
CO, 03:0.66% (0.46-0.86); 0-1
CO, N02:0.27% (0.03-0.51); 0-1
Random Effects:
CO: 0.66% (0.27-1.05); 0-1
CO, BS: 0.45% (-0.01 to 0.92); 0-1
CO, PM10:0.58% (0.12-1.04); 0-1
CO, S02:0.46% (0.07-0.85); 0-1
CO, 03:0.76% (0.45-1.06); 0-1
CO, N02:0.30% (-0.11 to 0.71); 0-1
PACF: (Partial Autocorrelation Function) Plot Fixed
Effects:
CO: 1.00% (0.83-1.18); 0-1
CO, BS:0.67% (0.30-1.04);0-1
CO, PM10:0.78% (0.55-1.00); 0-1
CO, S02:0.68% (0.47-0.90); 0-1
C0,03:1.12%(0.93-1.31);0-1
CO, N02:0.72% (0.50-0.95); 0-1
Random Effects:
CO: 1.20% (0.63-1.77); 0-1
CO, BS:0.77% (0.28-1.26);0-1
CO, PM10:1.09% (0.36-1.83); 0-1
CO, S02:0.75% (0.26-1.26); 0-1
C0,03:1.37%(0.81-1.95);0-1
CO, N02:0.88% (0.22-1.55); 0-1
Cardiovascular Mortality
8 Degrees of Freedom per Year
Fixed Effects:
CO: 0.80% (0.53-1.07); 0-1
CO, BS: 0.49% (-0.04 to 1.02); 0-1
CO, PM10:0.73% (0.39-1.07); 0-1
CO, S02:0.72% (0.39-1.04); 0-1
CO, 03:0.91% (0.62-1.20); 0-1
CO, N02:0.44% (0.10-0.79); 0-1
Random Effects:
CO: 0.81% (0.36-1.26); 0-1
CO, BS: 0.49% (-0.04 to 1.02); 0-1
CO, PM10:0.73% (0.39-1.07); 0-1
CO, S02:0.68% (-0.03 to 1.40); 0-1
CO, 03:1.02% (0.58-1.46); 0-1
CO, N02:0.43% (-0.06 to 0.93); 0-1
PACF (Partial Autocorrelation Function) Fixed
Effects:
CO: 1.06% (0.80-1.32); 0-1
CO, BS:0.83% (0.31-1.35);0-1
CO, PM10:0.95% (0.62-1.27); 0-1
CO, S02:0.91% (0.59-1.22); 0-1
C0,03:1.28%(1.01-1.56);0-1
CO, N02:0.68% (0.35-1.00); 0-1
Random Effects:
CO: 1.25% (0.30-2.21); 0-1
CO, BS:0.83% (0.31-1.35);0-1
CO, PM10:1.13% (0.60-1.67); 0-1
CO, S02:0.86% (0.06-1.66); 0-1
CO, 03:1.62% (0.72-2.52); 0-1
CO, N02:0.84% (-0.03 to 1.71); 0-1
Effect Modifiers
Non-accidental Mortality
8 Degrees of Freedom per Year
Number of CO monitors:
25th Percentile: 0.71% (0.48-0.94); 0-1
75th Percentile: 0.54% (0.34-0.74); 0-1
Mean PM10 Levels:
25th Percentile: 0.37% (0.08-0.66); 0-1
75th Percentile: 0.49% (0.28-0.69); 0-1
Standardized Mortality Rate:
25th Percentile:0.79% (0.55-1.03); 0-1
75th Percentile: 0.44% (0.22-0.66); 0-1
Western cities: 0.75% (0.47-1.03); 0-1
Southern cities: 0.61% (0.32-0.91); 0-1
Eastern cities: 0.03% (-0.47 to 0.53); 0-1
PACF (Partial Autocorrelation Function)
September 2009
C-87
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Number of CO monitors:
25th Percentile: 1.18% (0.96-1.39); 0-1
75th Percentile:0.92% (0.73-1.11);0-1
Mean PM10 Levels:
25th Percentile:0.74% (0.46-1.02); 0-1
75th Percentile: 1.07% (0.87-1.27); 0-1
Standardized Mortality Rate:
25th Percentile: 1.29% (1.06-1.52); 0-1
75th Percentile: 0.77% (0.56-0.98); 0-1
Western cities: 1.15% (0.90-1.40); 0-1
Southern cities: 1.08% (0.79-1.38); 0-1
Eastern cities: 0.27% (-0.20 to 0.74); 0-1
Cardiovascular Mortality
8 Degrees of Freedom per Year
Mean03:
25th Percentile: 1.04% (0.67-1.41); 0-1
75th Percentile:0.82% (0.55-1.10); 0-1
Standardized Mortality Rate:
25th Percentile: 1.06% (0.71-1.42); 0-1
75th Percentile: 0.61% (0.30-0.93); 0-1
Population >75yrofage (%):
25th Percentile: 0.58% (0.25-0.92); 0-1
75th Percentile:0.94% (0.64-1.24); 0-1
Western cities: 1.06% (0.67-1.46); 0-1
Southern cities: 0.70% (0.26-1.14); 0-1
Eastern cities: 0.21 % (-0.48 to 0.90); 0-1
PACF (Partial Autocorrelation Function)
Mean03:
25th Percentile: 1.32% (0.96-1.68); 0-1
75th Percentile: 1.09% (0.83-1.14); 0-1
Standardized Mortality Rate:
25th Percentile: 1.40% (1.06-1.75); 0-1
75th Percentile:0.85% (0.55-1.14); 0-1
Population >75yrofage (%):
25th Percentile:0.74% (0.41-1.06); 0-1
75th Percentile: 1.25% (0.96-1.54); 0-1
Western cities: 1.38% (1.00-1.76); 0-1
Southern cities: 0.90% (0.47-1.33); 0-1
Eastern cities: 0.48% (-0.14 to 1.11); 0-1
Author: Schwartz et al.
(1999.017915)
Period of Study:
1989-1995
Location:
Spokane, WA
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit:
Dust Storm Days:
09/08/1990:6.37 ppm
09/12/1990:3.40 ppm
10/04/1990:3.15 ppm
11/09/1990:2.45 ppm
11/23/1990:2.50 ppm
09/13/1991:4.60 ppm
10/16/1991:2.10 ppm
10/21/1991:2.20 ppm
09/04/1992:3.43 ppm
09/12/1992:1.80 ppm
09/13/1992:1.65 ppm
09/25/1992:2.95 ppm
09/26/1992:4.30 ppm
10/08/1992:3.85 ppm
09/11/1993:1.88 ppm
11/3/1993:5.33 ppm
07/24/1994:2.10 ppm
08/30/1996:2.85 ppm
Range (Min, Max): NR
Copollutant: PM10
The study did not present quantitative results for CO.
September 2009
C-88
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Sharovsky et al.
(2004, 156976)
Period of Study:
1996-1998
Location:
Sao Paulo, Brazil
Author: Slaughter etal.
(2005, 073854)
Period of Study:
1/1995-6/2001
Location:
Spokane, WA
Design
Health Outcome (ICD10):
Mortality: Myocardial Infarction
(1.21)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: 35-
109
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-51 9);
Asthma (493) ;COPD (491,
492, 494, 496); Pneumonia
(480-487); Acute Upper
Respiratory Tract Infections
(464-466, 490); Cardiac
Outcomes (390-459)
Study Design: Time-series
Statistical Analyses:
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 3.7(1. 6) ppm
Range (Min, Max): (1.0, 11.8)
Copollutant: correlation
S02: r= 0.73; PM10:r= 0.51
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Areas in Spokane
Hamilton St:1.73(0.46) ppm
Backdoor Tavern:
1.29 (0.23) ppm
Spokane Club: 1.41 (0.32) ppm
Third and Washington:
1.82 (0.33) ppm
Rockwood:0.42 (0.15) ppm
Effect Estimates
Increment: NR
px100(SE);lag:
CO: 1.42 (1.01)
CO, S02,PM10: 0.97 (1.27)
(95% Cl)
Notes: The study did not present the lag used for
CO.
The study did not present quantitative results for CO.
Log-linear Poisson GLM,
natural splines for calendar
time
Age Groups Analyzed: All
ages
Range (Min, Max): NR
Copollutant correlation :
PM1:r = 0.63; PM25:r = 0.62;
PM10:r=0.32;
PM10-2.5:r = 0.32
Author: Stieb etal. (2003,
056908)
Period of Study:
1985-2000
Location: All locations
Author: Stolzel et al. (2007,
091374)
Period of Study:
9/1995-8/2001
Location:
Erfurt, Germany
Author: Sunyer et al. (2001 ,
019367)
Period of Study:
1990-1995
Location:
Barcelona, Spain
Author: Sunyer etal. (2002,
034835)
Period of Study:
1985-1995
Location:
Barcelona, Spain
Health Outcome (ICD9):
Mortality: Non-accidental
Study Design: Meta-analysis
Statistical Analyses: NR
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800);
Cardio-respiratory (390-459,
460-519,785,786)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality COPD
(491,492,494,496)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: >35
Health Outcome (ICD9):
Mortality: Respiratory mortality
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: >14
Study population:
Asthmatic individuals: 5,610
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: NR
IQR (25th, 75th) : NR
Copollutant: NR
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
0.47 (0.39) mg/m3
IQR (25th, 75th): (0.23, 0.57)
Copollutant correlation:
MC0.1-0.5:r=0.58;
MC0.01-2.5:r = 0.57;
PMi0:r= 0.50; NO: r = 0.70;
N02:r = 0.71
Pollutant: CO
Averaging Time: 8-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: PM10; N02; 03
Pollutant: CO
Averaging Time: 24- h avg
Median (SD) unit: 7.7 ug/m3
Range (Min, Max): (0.6, 66.0)
Copollutant:
PM10;BS;N02;03;S02
Increment: 1 .1 ppm
% Excess Mortality (Lower Cl, Upper Cl); lag:
Non-GAM:
Single-pollutant model (4 studies): 4.7% (1.1-8.4)
Multi-pollutant model (1 study): 0.0% (-3.8 to 3.8)
GAM'
Single-pollutant model (18 studies): 1 .6% (1 .1-2.1)
Multi-pollutant model (11 studies): 0.7% (-0.1 to 1 .5)
Increment: 0.34 mg/m3
Relative Risk (Lower Cl, Upper Cl); lag:
Total (non-accidental)
1.000 (0.977-1. 023); 0;
1.002 (0.980-1. 024); 1;
1.013(0.991-1.035);2;
1.007 (0.986-1 .029); 3;
1.012(0.990-1.034);4;
0.995 (0.974-1 .01 7); 5
Increment: 4.5 ug/m3
Odds Ratio (Lower Cl, Upper Cl); lag:
CO: 1.052 (0.990-1 .117); 0-2
CO, PM10: 1.017 (0.947-1 .091); 0-2
Increment: 7.2 ug/m3
Odds Ratio (Lower Cl, Upper Cl); lag:
Asthmatic individuals with 1 ED visit
1.1 27 (0.895-1 .41 8); 0-2
Asthmatic individuals with >1 ED visit
1.1 25 (0.773-1 .638); 0-2
Asthma/COPD individuals with >1 ED visit
0.81 5 (0.61 4-1 .082); 0-2
September 2009
C-89
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Author: Tsai et al. (2003,
050480)
Period of Study:
1994-2000
Location:
Kaohsiung, Taiwan
Author: Tsai et al. (2006,
090709)
Period of Study:
1994-2000
Location:
Kaohsiung, Taiwan
Author: Vedal et al. (2003,
039044)
Period of Study:
1/1994-12/1996
Location:
Vancouver, BC, Canada
Author: Villeneuve et al.
(2003, 055051)
Period of Study:
1986-1999
Location:
Vancouver, BC, Canada
Design
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-51 9);
Circulatory (390-459)
Study Design:
Bidirectional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
27 days old to <1 yr of age
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-51 9);
Cardiovascular (390-459)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: All
ages
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(401 -440); Respiratory
(460-51 9); Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson, natural splines
Age Groups Analyzed: > 65
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.827 ppm
Range (Min, Max): (0.226, 1.770)
Copollutant:
PU.»- ^o,- NO,- n.
TIVI10, *J^2, '"^2, ^3
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 8.27 ppm
Range (Min, Max): (2.26, 17.70)
Copollutant:
PM10;S02;03;N02
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.6 (0.2) ppm
Range (Min, Max): (0.3, 1.9)
Copollutant correlation:
Summer:
PM10:r = 0.71 ;03:r = 0.12;
N02:r = 0.81 ;S02: r = 0.67
Winter:
PM10: r= 0.76; 03:r = -0.65;
N02: r = 0.78; S02:r = 0.83
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .0 ppm
Range (Min, Max): (0.2, 4.9)
Copollutant:
PM25;PM10;PM10-2.5;TSP;
S04;CO;COH;03;N02;S02
Effect Estimates (95% Cl)
Increment: 0.313 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Total (non-accidental): 1 .003 (0.968-1 .039); 0-2
Respiratory: 1 .01 1 (0.883-1 .1 59) ; 0-2
Circulatory: 0.986 (0.914-1 .063); 0-2
Increment: 0.31 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Postneonatal Mortality
1.051 (0.304-3.630); 0-2
The study did not present quantitative results for CO.
Increment: 1.1 ppb
% Increase (Lower Cl, Upper Cl); lag:
Non-accidental
0.5% (-1 .9 to 2.9); 0-2; / -0.3% (-2.2 to 1 .7); 0;
0.6% (-1.3 to 2.6); 1;/ 0.5% (-1.4 to 2.5); 2
Cardiovascular
2.3% (-1 .6 to 6.3); 0-2; / 1 .6% (-1 .5 to 4.7); 0;
1.2% (-2.0 to 4.5); 1 ;/ 1.5% (-1.5 to 4.4); 2
Respiratory
-1 .0% (-7.3 to 5.8) ; 0-2; / 1 .3% (-4.4 to 7.3) ; 0;
-0.1% (-5.3 to 5.4); 1 ; -/2.8% (-7.8 to 2.6); 2
Cancer
-2.8% (-7.6 to 2.4); 0-2; / -3.0% (-6.9 to 1 .1); 0;
-1 .6% (-5.6 to 2.4); 1 ; / -0.5% (-4.7 to 3.8); 2
Author: Wang et al. (2008,
179974)
Period of Study: Daily CO
content: 2000-2005 (data
from Beijing Environment
Protection Bureau), Death
rate: 2000-2003
Location: Beijing, China
Health Outcome: Mortality
Study Design: Time series,
Granger causality, Back
propagation neural network
model, MIV
Statistical Analyses: Eviews
3.1.SAS9.0, Matlab7.0
Age Groups Analyzed: NR
Sample Description: Death
rate of respiratory diseases in
Beijing from China Centers for
Disease Control and
Prevention
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Granger causality: Acute respiratory diseases
probability: 0.03122
COPD probability: 0.00047
Change of death rate of acute respiratory diseases:
Increasing 10%:+0.437, Decreasing 10%: -0.386
Change of death rate of COPD: Increasing 10%:
+0.181, Decreasing 10%:-0.316
Lags examined: 10
September 2009
C-90
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Wichmann et al.
(2000, 013912)
Period of Study:
9/1995-12/1998
Location:
Erfurt, Germany
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(401-440); Respiratory
(460-519)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
<70
70-79
>80
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.6 (0.5) mg/m3
Range (Min, Max): (0.10, 2.50)
Copollutant correlation:
PM25: r = 0.62; PM10:r = 0.58;
TSP: r= 0.57; S02:r= 0.59;
N02:r = 0.71
Increment: 0.5 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
Single-Day Lag
CO: 1.055 (1.003-1.110); 4
Polynomial Distributed Lag
Multi-pollutant model: 1.076 (1.017-1.138); 4
Total Mortality
CO: 1.012 (0.977-1.049); 0
Log-transformed: 1.016 (0.962-1.073); 0
1.004(0.969-1.040);!
Log-transformed: 1.027 (0.973-1.083);!
1.020 (0.984-1.057); 2
Log-transformed: 1.024 (0.970-1.081); 2
1.019 (0.984-1.055); 3
Log-transformed: 1.037 (0.984-1.093); 3
1.029 (0.995-1.063); 4
Log-transformed: 1.055 (1.003-1.110);4
0.997 (0.965-1.031); 5
Log-transformed: 1.014 (0.966-1.065); 5
Total Mortality (Season-specific):
Log-transformed
Winter: 1.002 (0.922-1.088); 4
Spring: 1.019 (0.942-1.102); 4
Summer: 1.085 (1.018-1.156); 4
Fall: 1.111 (1.039-1.188); 4
Winter-specific: Log-transformed
10/95-3/96:1.046 (0.949-1.153); 4
10/96-3/97:1.091 (0.998-1.193); 4
10/97-3/98:1.028 (0.966-1.095); 4
One-pollutant Model: Log-transformed
CO: 1.055 (1.003-1.110); 4
Author: Yang etal.
055603)
Period of Study:
1994-1998
Location:
Taipei, Taiwan
(2004, Health Outcome (ICD9):
Mortality:
Non-accidental (<800);
Circulatory (390-459);
Respiratory (460-51 9)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .16 ppm
Range (Min, Max): (0.24, 4.42)
Copollutant:
PM10;S02;N02;03
Increment: 0.52 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Non-accidental: 1.005 (0.980-1 .031); 0-2
Respiratory: 1 .01 4 (0.925-1 .1 1 0) ; 0-2
Circulatory: 0.996 (0.948-1 .046); 0-2
Table C-8
Study
Studies of long-term CO exposure and mortality.
Design
Author: Krewski et al. Health Outcome: mortality
(2009, 191193)
Period of Study:
1983-2000
Location:
United States
Study Design: cohort
Statistical Analyses: random
effects Cox model
Age Groups Analyzed:
30+ yrs
Sample Description:
508,538 adults living in large US
cities
Concentrations
Averaging Time: 1980 annual avg
Mean (SD) unit:
1.68 (0.66) ppm
Range (min, max):
0.19,3.95
CoPollutant:
PM15, PM25,S02, S04.TSP, 03,
N02
Effect Estimates (95% Cl)
Increment: 1 ppm
HR Estimate [Lower Cl, Upper Cl] :
Lags examined: NR
All Causes: 1.00 (0.99, 1.01)
Cardiopulmonary: 1.00 (0.99, 1.01)
IHD:1.01 (0.99,1.03)
Lung Cancer: 0.99 (0.97, 1.03)
All Other Causes: 0.99 (0.98, 1.01)
September 2009
C-91
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lipfertetal.
(2000, 004087)
Period of Study:
1975-1996
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-90,000 hypertensive male U.S.
veterans
Statistical Analyses:
Staged regression
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1960-1974:10.82 (5.15) ppm
1975-1981:7.64 (2.94) ppm
1982-1988:3.42 (0.95) ppm
1989-1996:2.36 (0.67) ppm
Range (Min, Max):
1960-1974: (0.94, 35.30)
1975-1981: (0.43, 22.38)
1982-1988: (0.30,15.20)
1989-1996: (0.30,7.10)
Copollutants; correlation:
1960-1974:
03:r = 0.004;
N02:r = 0.690;
S042-:r = 0.469
1975-1981:
03:r = 0.109;
N02:r = 0.249;
S042-:r =-0.155;
IP S042-:r = 0.356;
PM25:r = 0.634;
PM 10-2.5: r = 0.498;
PM15:r = 0.626
1982-1988
03:r = 0.158; N02:r = 0.413; S042-:
r =-0.518;
IP S042-:r = 0.075;
PM25:r = 0.296;
PM10-2.5:r = 0.135
PM15:r = 0.284
1989-1996
03:r = 0.397;
N02:r = 0.492;
S042-:r =-0.551
Increment: NR
Coefficient:
Baseline Model
Exposure Period: up to 1975
Single Period:-0.000
Deaths, 1976-81:0.0043
Deaths, 1982-88:-0.0002
Deaths after 1988:-0.0041
Exposure Period: 1975-81
Single Period:-0.013
Deaths, 1976-81:-0.0170
Deaths, 1982-88:-0.0217
Deaths after 1988:-0.0240
Exposure Period: 1982-88
Single Period:-0.028
Deaths, 1976-81:-0.0294
Deaths, 1982-88:-0.0484
Deaths after 1988:-0.0424
Exposure Period: 1989-96
Single Period:-0.046
Deaths, 1976-81:-0.0590
Deaths, 1982-88:-0.0581
Deaths after 1988:-0.0536
Final Model w/ Ecological Variables
Exposure Period: up to 1975
Single Period:-0.001
Deaths, 1976-81:0.0013
Deaths, 1982-88:-0.0022
Deaths after 1988:-0.0061
Exposure Period: 1975-81
Single Period:-0.008
Deaths, 1976-81:-0.0128
Deaths, 1982-88:-0.0186
Deaths after 1988:-0.0203
Exposure Period: 1982-88
Single Period:-0.009
Deaths, 1976-81:-0.0007
Deaths, 1982-88:-0.0246
Deaths after 1988:-0.0216
Exposure Period: 1989-96
Single Period:-0.009
Deaths, 1976-81:-0.0106
Deaths, 1982-88:-0.0136
Deaths after 1988:-0.0078
Notes: Mortality risks based on mean
concentrations of pollutants less estimated
background weighted by the number of subjects in
each county, but The study did not present this
value for each pollutant.
Author: Lipfert and
Morris (2002, 019217)
Period of Study:
1960-1997
Location:
U.S. counties
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design:
Ecological/ cross-sectional
Statistical Analyses:
Staged regression
Age Groups Analyzed:
15-44
45-64
65-74
75-84
2 85
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit:
1960-1969:13.81 (8.47) ppm
1970-1 974:9. 64 (5. 63) ppm
1979-1 981:5. 90 (3. 54) ppm
1989-1 991:2.69 (1.22) ppm
1995-1997:1 .72 (0.76) ppm
Range (Min, Max): NR
Copollutant:
TSP
S042-
S02
N02
03
Increment: NR
Attributable risk (SE):
Attributable Risks of mortality (1960-4)
Peak CO 1 960-1 964, All locations
Ages 15-44: 0.1299 (0.0341)
Ages 45-64: 0.0340 (0.0280)
Ages 65-74: -0.0058 (0.0220)
Ages 75-84: 0.01 21 (0.0188)
Ages > 85: 0.0374 (0.0225)
Log Mean: 0.0365 (0.0149)
Attributable Risks of mortality (1970-4)
Peak CO 1 970-1 974, All locations
Ages 15-44: 0.0553 (0.0240)
Ages 45-64: 0.01 81 (0.0148)
Ages 65-74: -0.0146 (0.0134)
Ages 75-84: -0.01 28 (0.0098)
Ages > 85: -0.0151 (0.0093)
Log Mean: 0.0038 (0.0086)
September 2009
C-92
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Attributable Risks of mortality (1979-81)
Peak CO 1979-1981, All locations
Ages 15-44:0.0054 (0.0174)
Ages 45-64:-0.0060 (0.0141)
Ages 65-74:-0.0251 (0.0105)
Ages 75-84: -0.0331 (0.0086)
Ages > 85:-0.0123 (0.0079)
Log Mean:-0.0183 (0.0077)
Peak CO 1970-1974, All locations
Ages 15-44:0.0218 (0.0200)
Ages 45-64:0.0327 (0.0161)
Ages 65-74:-0.0136 (0.0119)
Ages 75-84:-0.0250 (0.0105)
Ages > 85:-0.0202 (0.0085)
Log Mean: -0.0048 (0.0077)
Peak CO 1960-1969, All locations
Ages 15-44:0.0506 (0.0478)
Ages 45-64:0.0704 (0.0337)
Ages 65-74:0.0100 (0.0211)
Ages 75-84:-0.0124 (0.0143)
Ages > 85:0.0187 (0.0135)
Log Mean: 0.0084 (0.0149)
Peak CO 1979-1981, CO 1970-1974
Ages 15-44:0.0244 (0.0209)
Ages 45-64:0.0016 (0.0181)
Ages 65-74:-0.0183 (0.0128)
Ages 75-84:-0.0382 (0.0108)
Ages > 85:-0.0201 (0.0089)
Log Mean:-0.0165 (0.0089)
Peak CO 1979-1981, CO 1960-1969
Ages 15-44:0.0748 (0.0679)
Ages 45-64:0.0844 (0.0496)
Ages 65-74:0.0144 (0.0259)
Ages 75-84:-0.0158 (0.0168)
Ages > 85:-0.0073 (0.0170)
Log Mean: 0.0109 (0.0218)
Peak CO 1979-1981, CO 1960-1969
Ages 15-44:0.1191 (0.0709)
Ages 45-64:0.1163 (0.0491)
Ages 65-74:0.0177 (0.0310)
Ages 75-84:-0.0120 (0.0212)
Ages > 85:-0.0040 (0.0202)
Log Mean: 0.0211 (0.0231)
Attributable Risks of mortality (1989-91)
Peak CO 1989-1991, All locations
Ages 15-44:0.0404 (0.0322)
Ages 45-64:-0.0262 (0.0162)
Ages 65-74:-0.0397 (0.0115)
Ages 75-84: -0.0464 (0.0097)
Ages > 85:-0.0209 (0.0073)
Log Mean:-0.0178 (0.0098)
Peak CO 1979-1981, All locations
Ages 15-44:0.0522 (0.0227)
Ages 45-64:-0.0047 (0.0121)
Ages 65-74:-0.0165 (0.0078)
Ages 75-84: -0.0268 (0.0068)
Ages > 85:-0.0027 (0.0055)
Log Mean: -0.0020 (0.0065)
Peak CO 1970-1974, All locations
Ages 15-44:0.0685 (0.0274)
Ages 45-64:0.0022 (0.0148)
Ages 65-74:-0.0051 (0.0091)
Ages 75-84:-0.0158 (0.0079)
Ages > 85:-0.0069 (0.0060)
Log Mean: 0.0038 (0.0077)
Peak CO 1960-1969, All locations
Ages 15-44:0.0578 (0.0713)
Ages 45-64:0.0583 (0.0347)
Ages 65-74:0.0007 (0.0174)
September 2009
C-93
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Ages 75-84:-0.0245 (0.0130)
Ages > 85:-0.0138 (0.0113)
Log Mean: 0.0041 (0.0176)
Attributable Risks of mortality (1995-97)
Peak CO 1995-1997, All locations
Ages 15-44:0.0344 (0.0256)
Ages 45-64:-0.0203 (0.0198)
Ages 65-74:-0.0346 (0.0146)
Ages 75-84:-0.0378 (0.0161)
Ages > 85:-0.0283 (0.0119)
Log Mean:-0.0188 (0.0103)
Peak CO 1989-1991, All locations
Ages 15-44:0.0289 (0.0248)
Ages 45-64:-0.0192 (0.0192)
Ages 65-74:-0.0466 (0.0140)
Ages 75-84:-0.0497 (0.0147)
Ages > 85:-0.0301 (0.0108)
Log Mean: -0.0240 (0.0096)
Peak CO 1979-1981, All locations
Ages 15-44:0.0336 (0.0176)
Ages 45-64:-0.0037 (0.0135)
Ages 65-74: -0.0298 (0.0096)
Ages 75-84:-0.0301 (0.0105)
Ages > 85:-0.0087 (0.0078)
Log Mean:-0.0094 (0.0071)
Peak CO 1970-1974, All locations
Ages 15-44:0.0464 (0.0202)
Ages 45-64:0.0202 (0.0155)
Ages 65-74:-0.0032 (0.0112)
Ages 75-84:-0.0157 (0.0122)
Ages > 85:-0.0142 (0.0084)
Log Mean: 0.0007 (0.0077)
Peak CO 1960-1969, All locations
Ages 15-44:0.0679 (0.0441)
Ages 45-64:0.0772 (0.0405)
Ages 65-74:0.0059 (0.0173)
Ages 75-84:-0.0085 (0.0213)
Ages > 85:-0.0158 (0.0162)
Log Mean: 0.0162 (0.0149)
Author: Lipfertetal.
(2006, 088218)
Period of Study:
1976-2001
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-70,000 hypertensive male U.S.
veterans
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1976-1981:7.6 (2.9) ppm
1982-1988:3.4 (9.5) ppm
1989-1996:2.4 (0.67) ppm
1997-2001:1.6 (5.6) ppm
Range (Min, Max): NR
Copollutants correlation:
ln(VKTA):r = -0.06
Avg N02:r= 0.43
Peak 03:r= 0.08
Peak S02:r =-0.05
PM25:r = 0.08
S042-:r = -0.16
Note:VKTA=annualvehicle-km
traveled/km2
Increment: 2 ppm
Relative risk (Lower Cl, Upper Cl):
CO: 1.032 (0.954-1.117)
CO, InVKTA: 0.999 (0.923-1.081)
CO, lnVKTA,N02:1.012 (0.923-1.110)
CO, InVKTA, N02+03:1.023 (0.939-1.115)
September 2009
C-94
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lipfertetal.
(2006, 088756)
Period of Study:
1997-2002
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-18,000 hypertensive male U.S.
veterans
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1999-2001:1.63 (0.84) ppm
1999-2001 (STN sites only): 1.73
(0.77)
Range (Min, Max):
1999-2001: (0.40,6.7)
1999-2001 (STN sites only): (0.47,
4.2)
Copollutants correlation:
Inftraffic density): r = -0.199
PM25:r = 0.040;As:r=0.148
Cr: r = 0.448; Cu:r = 0.177
Fe: r = -0.138; Pb:r = 0.420
Mn: r= 0.357; Ni:r = 0.090
Se: r= -0.110; V:r= 0.230
Zn: r = 0.472; OC:r = 0.470
EC: r = 0.234; S042-:r = -0.123
N03-:r=-0.088
PM2s comp.:r = 0.133
N02:r=0.418
Peak03:r=0.172
Peak S02:r = 0.405
Increment: NR
p coefficient (SE); t-statistic:
-0.00000536(0.0000324);-0.165
Author: Jerrettetal.
(2003, 087380)
Period of Study:
1982-1989
Location:
1(17 1 1 9 ritips
1 U( U.O. UllICo
Mortality
Health Outcome (ICD9):
Cardiovascular;CHD;
Cerebrovascular disease
Study Design: Cohort
Study Population:
65, 893 postmenopausal women
without previous cardiovascular
disease
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 1.56 ppm
Range (Min, Max): (0.1 9, 3.95)
Copollutants correlation:
Sulfates: r= -0.07
N02
03
S02
Increment: 1 ppm
Relative risk (Lower Cl, Upper Cl):
CO: 0.98 (0.92-1 .03)
CO, Sulfates: 0.97 (0.92-1 .03)
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: 2 30
Author: Miller etal.
(2007, 090130)
Period of Study:
1994-1998
Location:
Qfi 1 1 C pitipc
OU U .O. UllICo
Mortality
Health Outcome (ICD9):
Cardiovascular;CHD;
Cerebrovascular disease
Study Design: Cohort
Study Population:
65, 893 postmenopausal women
without previous cardiovascular
disease
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: NR
Range (Min, Max):NR
Copollutants:
PM2.5
PM10-2.5
S02
M/"l
Increment: 1 ppm
Hazard ratio (Lower Cl, Upper Cl):
All subjects
CO: 1.0 (0.81-1 .22)
Only subjects with non-missing exposure data
CO: 0.92 (0.71-1 .21)
CO, PM25, PMm-2.5, S02, N02, 03: 0.93 (0.67,
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: 50-79
03
1.30)
Author: Pope et al.
(2002, 024689)
Period of Study:
1980-1998
Location:
All 50 States,
Washington DC, and
Puerto Rico (ACS-CPS-
Mortality
Health Outcome (ICD9):
Total (non-accidental) (<800); Lung
Cancer (162);Cardiopulmonary
(401-440,460-519)
Study Design:
Prospective cohort
Statistical Analyses:
Cox proportional hazards model
Age Groups Analyzed: 2 30
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
1980:1.7 (0.7) ppm
1982-1998:1.1 (0.4) ppm
Range (Min, Max): NR
Co pollutant:
PM2.5;PM10;TSP;S02;N02;03
The study presents results for CO graphically.
September 2009
C-95
DRAFT - DO NOT CITE OR QUOTE
-------
References
Abe T; Tokuda Y; Ohde S; Ishimatsu S; Nakamura T; Birrer RB. (2009). The relationship of short-term air pollution and
weather to ED visits for asthma in Japan. Am J Emerg Med, 27: 153-159. 190536
Andersen ZJ; Loft S; Ketzel M; Stage M; Scheike T; Mette MN; Bisgaard H. (2008). Ambient Air Pollution Triggers
Wheezing Symptoms in Infants. Thorax, 63: 710-716. 096150
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Scheike T; Loft S. (2007). Ambient particle source apportionment and daily
hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ Epidemiol, 17: 625-636.
093201
Anderson HR; Bremner SA; Atkinson RW; Harrison RM; Walters S. (2001). Particulate matter and daily mortality and
hospital admissions in the west midlands conurbation of the United Kingdom: associations with fine and coarse
particles, black smoke and sulphate. Occup Environ Med, 58: 504-510. 017033
Atkinson RW; Bremner SA; Anderson HR; Strachan DP; Bland JM; Ponce de Leon A. (1999). Short-term associations
between emergency hospital admissions for respiratory and cardiovascular disease and outdoor air pollution in
London. Arch Environ Occup Health, 54: 398-411. 007882
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM; Bertazzi PA;
Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb Haemost, 5: 252-260.
090733
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB; Canada A;
Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air pollution and
cardiovascular admisisons association in Spain: results within the EMECAS project. J Epidemiol Community
Health, 60: 328-336. 088746
Ballester F; Tenias JM; Perez-Hoyos S. (2001). Air pollution and emergency hospital admissions for cardiovascular
diseases in Valencia, Spain. J Epidemiol Community Health, 55: 57-65. 013257
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW. (2006). The effects of air
pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities.
Environ Health Perspect, 114: 1018-1023. 089770
Bedeschi E; Campari C; Candela S; Collini G; Caranci N; Frasca G; Galassi C; Francesca G; Vigotti MA. (2007). Urban air
pollution and respiratory emergency visits at pediatric unit, Reggio Emilia, Italy. J Toxicol Environ Health A, 70:
261-265. 090712
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and Massachusetts.
Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions in
Taipei, Taiwan. Occup Environ Med, 65:104- 111. 091268
Bell ML; Peng RD; Dominici F; Samet JM. (2009). Emergency admissions for cardiovascular disease and ambient levels
of carbon monoxide: Results for 126 U.S. urban counties, 1999-2005. Circulation, xx: xxx-xxx. 193780
Bellini P; Baccini M; Biggeri A; Terracini B. (2007). The meta-analysis of the Italian studies on short-term effects of air
pollution (MISA): Old and new issues on the interpretation of the statistical evidences. Environmetrics, 18: 219-
229. 097787
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; WichmannHE; Peters A. (2006). Runs of ventricular
and supraventricular tachycardia triggered by air pollution in patients with coronary heart disease. J Occup Environ
Med, 48: 1149-58.098702
Berglind N; Bellander T; Forastiere F; von Klot S; Aalto P; Elosua R; Kulmala M; Lanki T; Lowel H; Peters A; Picciotto S;
Salomaa V; Stafoggia M; Sunyer J; Nyberg F. (2009). Ambient air pollution and daily mortality among survivors of
myocardial infarction. Epidemiology, 20: 110-118. 190068
Bhattacharyya N. (2009). Air quality influences the prevalence of hay fever and sinusitis. Laryngoscope, 119: 429-433.
180154
September 2009 C-96 DRAFT - DO NOT CITE OR QUOTE
-------
Biggeri A; Baccini M; Bellini P; Terracini B. (2005). Meta-analy sis of the Italian studies of short-term effects of air
pollution (MISA), 1990-1999. Int J Occup Environ Health, 11: 107-122. 087395
Better DA; Jorgensen B; Peres AAQ. (2002). A longitudinal study of mortality and air pollution for Sao Paulo, Brazil. J
Expo Sci Environ Epidemiol, 12: 335-343. 011922
Braga ALF; Saldiva PHN; Pereira LAA; Menezes JJC; Conceicao GMC; Lin CA; Zanobetti A; Schwartz J; Dockery DW.
(2001). Health effects of air pollution exposure on children and adolescents in Sao Paulo, Brazil. Pediatr Pulmonol,
31: 106-113.016275
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air pollution
impacts on birth outcomes. Environ Health Perspect, 116: 680-686. 156292
Bremner SA; Anderson HR; Atkinson RW; McMichael AJ; Strachan DP; Bland JM; Bower JS. (1999). Short term
associations between outdoor air pollution and mortality in London 1992-4. Occup Environ Med, 56: 237-244.
007601
Briet M; Collin C; Laurent S; Tan A; Azizi M; Agharazii M; Jeunemaitre X; Alhenc-Gelas F; Boutouyrie P. (2007).
Endothelial function and chronic exposure to air pollution in normal male subjects. Hypertension, 50: 970-976.
093049
Burnett RT; Brook J; Dann T; Delocla C; Philips O; Cakmak S; Vincent R; Goldberg MS; Krewski D. (2000). Association
between particulate- and gas-phase components of urban air pollution and daily mortality in eight Canadian cities.
Inhal Toxicol, 12: 15-39. 010273
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air pollution on
cardiorespiratory hospitalizations. Arch Environ Occup Health, 54: 130-139. 017269
Burnett RT; Smith-Doiron M; Stieb D; Raizenne ME; Brook JR; Dales RE; Leech JA; Cakmak S; Krewski D. (2001).
Association between ozone and hospitalization for acute respiratory diseases in children less than 2 years of age.
Am J Epidemiol, 153: 444-452. 093439
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59: 228-236.
086247
Cakmak S; Dales RE; Judek S. (2006). Respiratory health effects of air pollution gases: modification by education and
income. Arch Environ Occup Health, 61: 5-10. 093272
Cakmak S; Dales RE; Vidal CB. (2007). Air pollution and mortality in Chile: Susceptibility among the elderly. Environ
Health Perspect, 115: 524-7. 091170
Chan C-C; Chuang K-J; Chien L-C; Chen W-J; Chang W-T. (2006). Urban air pollution and emergency admissions for
cerebrovascular diseases in Taipei, Taiwan. Eur Heart J, 27: 1238-1244. 090193
Chan C-C; Chuang K-J; Su T-C; Lin L-Y (2005). Association between nitrogen dioxide and heart rate variability in a
susceptible population. Eur J Cardiovasc Prev Rehabil, 12: 580-586. 088988
Chang C-C; Tsai S-S; Ho S-C; Yang C-Y. (2005). Air pollution and hospital admissions for cardiovascular disease in
Taipei, Taiwan. EnvironRes, 98: 114-119. 080086
Chen L; Jennison BL; Yang W; Omaye ST. (2000). Elementary school absenteeism and air pollution. Inhal Toxicol, 12:
997-1016.011931
ChenL; Yang W; Jennison BL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada, 1991-
1999. Inhal Toxicol, 14: 141-157. 024945
Chen P-C; Lai Y-M; Chan C-C; Hwang J-S; Yang C-Y; Wang J-D. (1999). Short-term effect of ozone on the pulmonary
function of children in primary school. Environ Health Perspect, 107: 921-925. 011149
Cheng M-F; Tsai S-S; Wu T-N; Chen P-S; Yang C-Y. (2007). Air pollution and hospital admissions for pneumonia in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 70: 2021-6. 093034
Chiu HF; Cheng MH; Yang CY (2009). Air pollution and hospital admissions for pneumonia in a subtropical city: Taipei,
Taiwan. Inhal Toxicol, 21: 32-7. 190249
Cho B; Choi J; Yum YT. (2000). Air pollution and hospital admissions for respiratory disease in certain areas of Korea. J
Occup Health, 42: 185-191. 099051
September 2009 C-97 DRAFT - DO NOT CITE OR QUOTE
-------
Chock DP; Winkler SL; Chen C. (2000). A study of the association between daily mortality and ambient air pollutant
concentrations in Pittsburgh, Pennsylvania. J Air Waste Manag Assoc, 50: 1481-1500. 010407
Cifuentes LA; Vega J; Kopfer K; Lave LB. (2000). Effect of the fine fraction of particulate matter versus the coarse mass
and other pollutants on daily mortality in Santiago, Chile. J Air Waste Manag Assoc, 50: 1287-1298. 010351
Conceicao GMS; Miraglia SGEK; Kishi HS; Saldiva PHN; Singer JM. (2001). Air pollution and child mortality: a time-
series study in Sao Paulo, Brazil. Environ Health Perspect, 1093: 347-350. 016628
D'Ippoliti D; Forastiere F; Ancona C; Agabiti N; Fusco D; Michelozzi P; Perucci CA. (2003). Air pollution and myocardial
infarction in Rome: a case-crossover analysis. Epidemiology, 14: 528-535. 074311
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery disease. J
Occup Environ Med, 46: 1217-1221.099036
De Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; Van Wrjnen JH;
Kreyling W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on cardiorespiratory symptoms
in elderly subjects with coronary heart disease: the ULTRA study. Am J Epidemiol, 157: 613-623. 001061
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality associations with
air pollution. Am J Respir Grit Care Med, 167: 1117-1123. 055688
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily ambient
exposures to toxic and criteria air pollutants. Environ Health Perspect, 111: 647-656. 050460
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J; Zaldivar F;
Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are
associated with primary combustion aerosols in subjects with coronary artery disease. Environ Health Perspect,
116:898-906. 156390
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD; Verrier RL.
(2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias recorded by implanted
cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dominici F; McDermott A; Daniels M; Zeger SL; Samet JM. (2003). Mortality among residents of 90 cities. In: Revised
analyses of time-series studies of air pollution and health. Health Effects Institute. Boston, MA. HEI Special
Report. http://www.healtheffects.org/Pubs/TimeSeries.pdf. 056116
Estrella B; Estrella R; Oviedo J; Narvaez X; Reyes MT; Gutierrez M; Naumova EN. (2005). Acute respiratory diseases and
carboxyhemoglobin status in school children of Quito, Ecuador. Environ Health Perspect, 113: 607-11. 099124
Fairley D. (1999). Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ Health Perspect,
107:637-641.000896
Farhat SCL; Paulo RLP; Shimoda TM; Conceicao GMS; Lin CA; Braga ALF; Warth MPN; Saldiva PHN. (2005). Effect of
air pollution on pediatric respiratory emergency room visits and hospital admissions. Braz J Med Biol Res, 38: 227-
235. 089461
Filho MAP; Pereira LAA; Arbex FF; Arbex M; Conceicao GM; Santos UP; Lopes AC; Saldiva PHN; Braga ALF; Cendon
S. (2008). Effect of air pollution on diabetes and cardiovascular diseases in Sao Paulo, Brazil. Braz J Med Biol Res,
41: 526-532. 190260
Fischer P; Hoek G; Brunekreef B; Verhoeff A; van Wijnen J. (2003). Air pollution and mortality in the Netherlands: are the
elderly more at risk?. Eur Respir J, 40: 34S-38S. 043739
Fischer PH; Steerenberg PA; Snelder JD; Van Loveren H; Van Amsterdam JGC. (2002). Association between exhaled nitric
oxide, ambient air pollution and respiratory health in school children. Int Arch Occup Environ Health, 75: 348-353.
025731
Forastiere F; Stafoggia M; Picciotto S; Bellander T; D'Ippoliti D; Lanki T; Von Klot S; Nyberg F; Paatero P; Peters A;
Pekkanen J; Sunyer J; Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary deaths and air
pollution in Rome, Italy. Am J Respir Grit Care Med, 172: 1549-1555. 086323
Forastiere F; Stafoggia M; Tasco C; Picciotto S; Agabiti N; Cesaroni G; Perucci CA. (2007). Socioeconomic status,
particulate air pollution, and daily mortality: differential exposure or differential susceptibility. Am J Ind Med, 50:
208-216. 090720
September 2009 C-98 DRAFT - DO NOT CITE OR QUOTE
-------
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory hospitalizations
among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Fung KY; Luginaah I; Gorey KM; Webster G. (2005). Air pollution and daily hospital admissions for cardiovascular
diseases in Windsor, Ontario. Can J Public Health, 96: 29-33. 074322
Fusco D; Forastiere F; Michelozzi P; Spadea T; Ostro B; Area M; Perucci CA. (2001). Air pollution and hospital
admissions for respiratory conditions in Rome, Italy. EurRespir J, 17: 1143-1150. 020631
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation between
ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J Epidemiol, 162: 238-
252. 087892
Gold DR; Litonjua A; Schwartz J; Lovett E; Larson A; Nearing B; Allen G; Verrier M; Cherry R; Verrier R. (2000).
Ambient pollution and heart rate variability. Chest, 101: 1267-1273. 011432
Gold DR; Litonjua AA; Zanobetti A; Coull BA; Schwartz J; MacCallum G; Verrier RL; Nearing BD; Canner MJ; Suh H;
Stone PH. (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health Perspect, 113: 883-
887. 087558
Goldberg MS; Burnett RT; Bailar JC III; Brook J; Bonvalot Y; Tamblyn R; Singh R; Valois M-F. (2001). The association
between daily mortality and ambient air particle pollution in Montreal, Quebec 1 Nonaccidental mortality. Environ
Res, 86: 12-25. 016548
Goldberg MS; Burnett RT; Valois M-F; Flegel K; Bailar JC III; Brook J; Vincent R; Radon K. (2003). Associations
between ambient air pollution and daily mortality among persons with congestive heart failure. Environ Res, 91: 8-
20. 035202
Goldberg MS; Burnett RT; Yale JF; Valois MF; Brook JR. (2006). Associations between ambient air pollution and daily
mortality among persons with diabetes and cardiovascular disease. Environ Res, 100: 255-267. 088641
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive heart
failure to estimate the short-term effects from personal factors and environmental conditions on oxygen saturation
and pulse rate. Occup Environ Med, 65: 659-666. 180380
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. Am J Respir Grit Care Med, 169: 816-821. 055624
Gouveia N; Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in Sao Paulo,
Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Gouveia N; Fletcher T. (2000). Respiratory diseases in children and outdoor air pollution in Sao Paulo, Brazil: a time series
analysis. Occup Environ Med, 57: 477-483. 010436
Gouveia N; Fletcher T. (2000). Time series analysis of air pollution and mortality: effects by cause, age and socioeconomic
status. J Epidemiol Community Health, 54: 750-755. 012132
Guo YL; Lin Y-C; Sung F-C; Huang S-L; Ko Y-C; Lai J-S; Su H-J; Shaw C-K; Lin R-S; Dockery DW. (1999). Climate,
traffic-related air pollutants, and asthma prevalence in middle-school children in Taiwan. Environ Health Perspect,
107: 1001-1006. 010937
Gwynn RC; Burnett RT; Thurston GD. (2000). A time-series analysis of acidic particulate matter and daily mortality and
morbidity in the Buffalo, New York, region. Environ Health Perspect, 108:125-133. 074109
Ha E-H; Hong Y-C; Lee B-E; Woo B-H; Schwartz J; Christiani DC. (2001). Is air pollution a risk factor for low birth
weight in Seoul?. Epidemiology, 12: 643-648. 019390
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul, South Korea.
Pediatrics, 111: 284-290. 042552
Hajat S; Anderson HR; Atkinson RW; Haines A. (2002). Effects of air pollution on general practitioner consultations for
upper respiratory diseases in London. Occup Environ Med, 59: 294-299. 030358
Hajat S; Armstrong B; Wilkinson P; Busby A; Dolk H. (2007). Outdoor air pollution and infant mortality: Analysis of daily
time-series data in 10 English cities. J Epidemiol Community Health, 61: 719-722. 093276
Hajat S; Haines A; Goubet SA; Atkinson RW; Anderson HR. (1999). Association of air pollution with daily GP
consultations for asthma and other lower respiratory conditions in London. Thorax, 54: 597-605. 000924
September 2009 C-99 DRAFT - DO NOT CITE OR QUOTE
-------
Hapcioglu B; Issever H; Kocyigit E; Disci R; Vatansever S; Ozdilli K. (2006). The effect of air pollution and
meteorological parameters on chronic obstructive pulmonary disease at an Istanbul hospital. Indoor Built Environ,
15: 147-153.093263
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic stroke
occurrence: A case-crossover analysis from a 10-year population-based study in Dijon, France. Occup Environ
Med, 64: 439-445. 093270
Hinwood AL; De Klerk N; Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett J.
(2006). The relationship between changes in daily air pollution and hospitalizations in Perth, Australia 1992-1998:
A case-crossover study. Int J Environ Health Res, 16: 27-46. 088976
Hirsch T; Weiland SK; Von Mutius E; Safeca AF; Grafe H; Csaplovics E; Duhme H; Keil U; Leupold W. (1999). Inner city
air pollution and respiratory health and atopy in children. Eur Respir J, 14: 669-677. 003537
Ho W-C; Hartley WR; Myers L; Lin M-H;. (2007). Air pollution, weather, and associated risk factors related to asthma
prevalence and attack rate. Environ Res, 104: 402-409. 093265
Hoek G; Brunekreef B; Fischer P; Van Wijnen J. (2001). The association between air pollution and heart failure,
arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study. Epidemiology,
12: 355-357. 016550
Hoek G; Brunekreef B; Verhoeff A; Van Wijnen J; Fischer P. (2000). Daily mortality and air pollution in the Netherlands. J
Air Waste Manag Assoc, 50: 1380-1389. 010350
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air pollution
and heart rate variability among the elderly in Mexico City. Epidemiology, 14: 521-527. 057326
Honda Y; Nitta H; Ono M. (2003). Low level carbon monoxide and mortality of persons aged 65 or older in Tokyo, Japan
1976-1990. Eisei Kagaku, 49: 454-458. 193774
Hong Y-C; Lee J-T; Kim H; Ha E-H; Schwartz J; Christian! DC. (2002). Effects of air pollutants on acute stroke mortality.
Environ Health Perspect, 110: 187-191.024690
Hong Y-C; Lee J-T; Kim H; KwonH-J. (2002). Air pollution: A new risk factor in ischemic stroke mortality. Stroke, 33:
2165-2169. 035060
Hong Y-C; Leem J-H; Ha E-H; Christiani DC. (1999). PM10 exposure, gaseous pollutants, and daily mortality in Inchon,
South Korea. Environ Health Perspect, 107: 873-8. 008087
Hong YC; Leem JH; Ha EH. (1999). Air pollution and daily mortality in Inchon, Korea. J Korean Med Sci, 14: 239-244.
011195
Hosseinpoor AR; Forouzanfar MH; Yunesian M; Asghari F; Naieni KH; Farhood D. (2005). Air pollution and
hospitalization due to angina pectoris in Tehran, Iran: A time-series study. Environ Res, 99: 126-131. 087413
Huynh M; Woodruff TJ; Parker JD; Schoendorf KC. (2006). Relationships between air pollution and preterm birth in
California. Paediatr Perinat Epidemiol, 20: 454-461. 091240
Hwang B-F; Jaakkola JJK; Lee Y-L; Lin Y-C; Guo Y-LL. (2006). Relation between air pollution and allergic rhinitis in
Taiwanese schoolchildren. Respir Res, 7: 23. 088971
Hwang B-F; Lee Y-L; Lin Y-C; Jaakkola JJK; Guo YL. (2005). Traffic related air pollution as a determinant of asthma
among Taiwanese school children. Thorax, 60: 467-473. 089454
Hwang BF; Jaakkola JJ. (2008). Ozone and other air pollutants and the risk of oral clefts. Environ Health Perspect, 116:
1411-1415. 193794
Hwang J-S; Chan C-C. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am J
Epidemiol, 155: 1-10. 023222
Ibald-Mulli A; Stieber J; Wichmann H-E; Koenig W; Peters A. (2001). Effects of air pollution on blood pressure: A
population-based approach. Am J Public Health, 91: 571-577. 016030
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J; Hoek G;
Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and heart rate in subjects
with cardiovascular disease: a multicenter approach. Environ Health Perspect, 112: 369-377. 087415
September 2009 C-100 DRAFT - DO NOT CITE OR QUOTE
-------
Ito K; Thurston GD; Silverman RA;. (2007). Association between coarse particles and asthma emergency department (ED)
visits in New York City. Presented at American Thoracic Society international conference 2007, San Francisco.
091262
Jalaludin B; Mamies T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution on
gestational age is modified by season in Sydney, Australia. Environ Health, 6:16. 156601
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient air
pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+ years),
Sydney, Australia. JExpo Sci Environ Epidemiol, 16: 225-37. 189416
Jayaraman G; Nidhi. (2008). Air pollution and associated respiratory morbidity in Delhi. Health Care Manag Sci, 11: 132-
138. 180352
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of the air
pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health A, 66: 1735-1777.
087380
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Karr C; Lumley T; Shepherd K; Davis R; Larson T; Ritz B; Kaufman J. (2006). A case-crossover study of wintertime
ambient air pollution and infant bronchiolitis. Environ Health Perspect, 114: 277-281. 088751
Keatinge WR; Donaldson GC. (2001). Mortality related to cold and air pollution in London after allowance for effects of
associated weather patterns. Environ Res, 86: 209-216. 017063
Kettunen J; Lanki T; Tiittanen P; Aalto PP; Koskentalo T; Kulmala M; Salomaa V; Pekkanen J. (2007). Associations of fine
and ultrafine particulate air pollution with stroke mortality in an area of low air pollution levels. Stroke, 38: 918-
922. 091242
Kim S Y; O'Neill MS; Lee JT; Cho Y; Kim J; Kim H. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health, 80: 701-710. 092837
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from
ARIES. Inhal Toxicol, 16 Suppl 1: 131-141. 056585
Knox EG. (2008). Atmospheric pollutants and mortalities in English local authority areas. J Epidemiol Community Health,
62: 442-447. 193776
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver. Environ Health Perspect, 111: 1312-1317. 049466
Kontos AS; Fassois SD; Deli MF. (1999). Short-term effects of air pollution on childhood respiratory illness in Piraeus,
Greece, 1987-1992: nonparametric stochastic dynamic analysis. Environ Res, 81: 275-296. 011326
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Kwon H-J; Cho S-H; Nyberg F; Pershagen G. (2001). Effects of ambient air pollution on daily mortality in a cohort of
patients with congestive heart failure. Epidemiology, 12: 413-419. 016699
Lagorio S; Forastiere F; Pistelli R; lavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air
pollution and lung function among susceptible adult subjects: a panel study. Environ Health, 5:11. 089800
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; DTppoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S; Salomaa V;
Sunyer J; Tiittanen P; Von Klot S; Forastiere F; for the HEAPSS Study Group. (2006). Associations of traffic-
related air pollutants with hospitalisation for first acute myocardial infarction: the HEAPSS study. Occup Environ
Med, 63: 844-851.089788
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different gestational
phases contributes to risks of low birth weight. Hum Reprod, 18: 638-643. 043202
Lee I-M; Tsai S-S; Chang C-C; Ho C-K; Yang C-Y (2007). Air pollution and hospital admissions for chronic obstructive
pulmonary disease in a tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 19: 393-398. 090707
September 2009 C-101 DRAFT - DO NOT CITE OR QUOTE
-------
Lee IM; Tsai SS; Ho CK; Chiu HF; Yang CY. (2007). Air pollution and hospital admissions for congestive heart failure in a
tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 19: 899-904. 093271
Lee J-T; Kim H; Song H; Hong Y-C; Cho Y-S; Shin S-Y; Hyun Y-J; Kim Y-S. (2002). Air pollution and asthma among
children in Seoul, Korea. Epidemiology, 13: 481-484. 034826
Lee J-T; Son J-Y; Cho Y-S. (2007). A comparison of mortality related to urban air particles between periods with Asian dust
days and without Asian dust days in Seoul, Korea, 2000-2004. Environ Res, 105: 409-13. 093042
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic heart
diseases among individuals 64+years of age residing in Seoul, Korea. Arch Environ Health, 58: 617-623. 095552
Lee JT; Son JY; Kim H; Kim S Y (2006). Effect of air pollution on asthma-related hospital admissions for children by
socioeconomic status associated with area of residence. Arch Environ Occup Health, 61: 123-130. 098248
Lee Y-L; Shaw C-K; Su H-J; Lai J-S; Ko Y-C; Huang S-L; Sung F-C; Guo Y-L. (2003). Climate, traffic-related air
pollutants and allergic rhinitis prevalence in middle-school children in Taiwan. Eur Respir J, 21: 964-970. 049201
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect, 114: 905-910. 089828
Levy D; Sheppard L; Checkoway H; Kaufman J; Lumley T; Koenig J; Siscovick D. (2001). A case-crossover analysis of
particulate matter air pollution and out-of-hospital primary cardiac arrest. Epidemiology, 12: 193-199. 017171
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J Epidemiol,
159:768-777.056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. JExpo Sci Environ Epidemiol, 15: 319-328.
088677
Lin C-M; Li C-Y; Yang G-Y; Mao I-F. (2004). Association between maternal exposure to elevated ambient sulfur dioxide
during pregnancy and term low birth weight. Environ Res, 96: 41-50. 089503
Lin CA; Martins MA; Farhat SCL; Pope CA III; Conceicao GMS; Anastacio VM; Hatanaka M; Andrade WC; Hamaue
WR; Bohm GM; Saldiva PHN. (1999). Air pollution and respiratory illness of children in Sao Paulo, Brazil.
Paediatr Perinat Epidemiol, 13: 475-488. 040437
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and neonatal deaths
in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2003). Effect of short-term exposure to gaseous pollution on
asthma hospitalisation in children: a bi-directional case-crossover analysis. J Epidemiol Community Health, 57: 50-
55. 042549
Lin M; Chen Y; Villeneuve PJ; Burnett RT; Lemyre L; Hertzman C; McGrail KM; Krewski D. (2004). Gaseous air
pollutants and asthma hospitalization of children with low household income in Vancouver, British Columbia,
Canada. Am J Epidemiol, 159: 294-303. 055600
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children
younger than 15 years in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240. 087828
Linn WS; Szlachcic Y; Gong H Jr; Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions in
metropolitan Los Angeles. Environ Health Perspect, 108: 427-434. 002839
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Morris SC. (2002). Temporal and spacial relationships between age-specific mortality and ambient air quality
in the United States: preliminary results for counties, 1960-97. Occup Environ Med, 59: 156-174. 019217
Lipfert FW; Morris SC; Wyzga RE. (2000). Daily mortality in the Philadelphia metropolitan area and size-classified
particulate matter. J Air Waste Manag Assoc, 50: 1501-1513. 004088
Lipfert FW; Perry HM Jr; Miller JP; Baty JD; Wyzga RE; Carmody SE. (2000). The Washington University-EPRI veterans'
cohort mortality study: preliminary results. Inhal Toxicol, 4: 41-73. 004087
September 2009 C-102 DRAFT - DO NOT CITE OR QUOTE
-------
Lipfert FW; Wyzga RE; Baty JD; Miller JP (2006). Traffic density as a surrogate measure of environmental exposures in
studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos Environ, 40: 154-169.
088218
Lippmann M; Ito K; Nadas A; Burnett RT. (2000). Association of particulate matter components with daily mortality and
morbidity in urban populations. 011938
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air pollutants
during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432. 090429
Liu S; Krewski D; Shi Y; Chen Y; Burnett RT. (2003). Association between gaseous ambient air pollutants and adverse
pregnancy outcomes in Vancouver, Canada. Environ Health Perspect, 111: 1773-1778. 089548
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni K; von Klot
S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and fibrinogen
polymorphisms. Environ Health Perspect, 117: 1373-1379. 191983
Luginaah IN; Fung K Y; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with respiratory
hospitalization in a government designated "area of concern": the case of Windsor, Ontario. Environ Health
Perspect, 113: 290-296. 057327
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air pollution
and stroke in Sheffield, United Kingdom: A small-area level geographical study. Stroke, 36: 239-243. 088683
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air
pollution, mortality, and hospital admissions from coronary heart disease in Sheffield, UK: A small-area level
ecological study. Eur Heart J, 26: 2543-2549. 090769
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth weight in the
northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
Mann JK; Tager IB; Lurmann F; Segal M; Quesenberry CP Jr; Lugg MM; Shan J; Van den Eeden SK. (2002). Air pollution
and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia. Environ
Health Perspect, 110: 1247-1252. 036723
Marines T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on birth
weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Mar TF; Norris GA; Koenig JQ; Larson TV. (2000). Associations between air pollution and mortality in Phoenix, 1995-
1997. Environ Health Perspect, 108: 347-353. 001760
Martins LC; Latorre MRDO; Saldiva PHN; Braga ALF. (2002). Air pollution and emergency room visits due to chronic
lower respiratory diseases in the elderly: An ecological time-series study in Sao Paulo, Brazil. J Occup Environ
Med, 44: 622-627. 035059
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K; Ghavam
SM. (2003). The effects of air pollution on acute respiratory conditions. Respirology, 8: 213-230. 052100
McGowan JA; Hider PN; Chacko E; Town GI. (2002). Particulate air pollution and hospital admissions in Christchurch,
New Zealand. Aust N Z J Public Health, 26: 23-29. 030325
Medeiros A; Gouveia N. (2005). [Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil].
Rev Saude Publica, 39: 965-972. 156750
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007). Traffic and outdoor air pollution levels near residences and
poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-463. 093275
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592. 092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Migliaretti G; Dalmasso P; Gregori D. (2007). Air pollution effects on the respiratory health of the resident adult
population in Turin, Italy. Int J Environ Health Res, 17: 369-379. 193772
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term exposure
to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458. 090130
September 2009 C-103 DRAFT - DO NOT CITE OR QUOTE
-------
Moolgavkar SH. (2000). Air pollution and daily mortality in three US counties. Environ Health Perspect, 108: 777-784.
012054
Moolgavkar SH. (2000). Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
metropolitan areas in the United States. Inhal Toxicol, 12: 75-90. 010274
Moolgavkar SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook counties. Health
Effects Institute. Boston, MA. 042864
Moolgavkar SH. (2003). Air pollution and daily mortality in two US counties: season-specific analyses and exposure-
response relationships. Inhal Toxicol, 15: 877-907. 051316
Moon JS; Kim YS; Kim JH; Son BS; Kim DS; Yang W. (2009). Respiratory health effects among schoolchildren and their
relationship to air pollutants in Korea. Int J Environ Health Res, 19: 31-48. 190297
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Air pollution and pulmonary function in
asthmatic children: effects of prenatal and lifetime exposures. Epidemiology, 19: 550-557. 122163
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air pollution and
allergic sensitization in children with asthma. J Asthma, 45: 874-881. 187280
Nautiyal J, 3rd,Garg ML,Kumar MS,Khan AA,Thakur JS,Kumar R. (2007). Air pollution and cardiovascular health in
Mandi-Gobindgarh, Punjab, India - a pilot study. Int J Environ Health Res, 4: 268-82. 190301
Neidell MJ. (2004). Air pollution, health, and socio-economic status: The effect of outdoor air quality on childhood asthma.
J Health Econ, 23: 1209-1236. 057330
Nkwocha EE; Egejuru RO. (2008). Effects of industrial air pollution on the respiratory health of children. Int J Environ Sci
Tech, 5:509-516. 190304
Norris G; Young-Pong SN; Koenig JQ; Larson TV; Sheppard L; Stout JW. (1999). An association between fine particles
and asthma emergency department visits for children in Seattle. Environ Health Perspect, 107: 489-493. 040774
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Grain EF; Evans R, 3rd; Gruchalla R; Morgan W; Stout J; Adams
GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with asthma in US inner
cities. J Allergy Clinlmmunol, 121: 1133-1139 el!31. 156818
Ostro BD; Hurley S; Lipsett MJ. (1999). Air pollution and daily mortality in the Coachella Valley, California: A study of
PM10 dominated by coarse particles. Environ Res, 81: 231-238. 006610
Park H; Lee B; Ha E-H; Lee J-T; Kim H; Hong Y-C. (2002). Association of air pollution with school absenteeism due to
illness. Arch Pediatr Adolesc Med, 156: 1235-1239. 093798
Park JW; Lim YH; Kyung S Y; An CH; Lee SP; Jeong SH; Ju S-Y (2005). Effects of ambient particulate matter on peak
expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in Korea. Respirology, 10:
470-476. 088673
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate variability: The VA
normative aging study. Environ Health Perspect, 113: 304-309. 057331
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in California.
Pediatrics, 115: 121-128. 087462
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Pekkanen J; Brunner EJ; Anderson HR; Tiittanen P; Atkinson RW. (2000). Daily concentrations of air pollution and plasma
fibrinogen in London. Occup Environ Med, 57: 818-822. 013250
Penttinen P; Tiittanen P; Pekkanen J. (2004). Mortality and air pollution in metropolitan Helsinki, 1988-1996. Scand J
Work Environ Health, 2: 19-27. 087432
Penttinen P; Timonen KL; Tiittanen P; Mirme A; Ruuskanen J; Pekkanen J. (2001). Ultrafine particles in urban air and
respiratory health among adult asthmatics. Eur Respir J, 17: 428-435. 030335
September 2009 C-104 DRAFT - DO NOT CITE OR QUOTE
-------
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of
myocardial infarction. Circulation, 103: 2810-2815. 016546
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery DW.
(2000). Air pollution and incidence of cardiac arrhythmia. Epidemiology, 11: 11-17. 011347
Peters A; Perz S; Doring A; Stieber J; Koenig W; Wichmann H-E. (1999). Increases in heart rate during an air pollution
episode. Am J Epidemiol, 150: 1094-1098. 011554
Peters A; Skorkovsky J; Kotesovec F; Brynda J; Spix C; Wichmann HE; Heinrich J. (2000). Associations between
mortality and air pollution in central Europe. Environ Health Perspect, 108: 283-287. 001756
Pope CAIII; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-1141. 024689
RabinovitchN; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW (2004). Effects of wintertime ambient air
pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin
Immunol, 114: 1131-1137.096753
Rainham DGC; Smoyer-Tomic KE. (2003). The role of air pollution in the relationship between a heat stress index and
human mortality in Toronto. Environ Res, 93: 9-19. 053202
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P. (2004). Air
pollution and respiratory status in asthmatic children: Hints for a locally based preventive strategy AIRE study. Eur
J Epidemiol, 19: 567-576. 089500
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR; Dockery DW.
(2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in ambient air
pollution. Environ Health Perspect, 114: 120-123. 088427
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol, 161:
1123-1132.079620
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac arrhythmia
in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa C;
Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with known
ischemic heart disease in Mexico City. JExpo Sci Environ Epidemiol, 16: 131-137. 156913
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment and
pregnancy outcomes study at the University of California, Los Angeles. Am JEpidemiol, 166: 1045-1052. 096146
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics, 118: 493-
502. 089819
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern California
between 1989 and 1993. Epidemiology, 11: 502-511. 012068
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in Southern
California. Am J Epidemiol, 155: 17-25. 023227
Rodriguez C; Tonkin R; Heyworth J; Kusel M; De Klerk N; Sly PD; Franklin P; Runnion T; Blockley A; Landau L;
Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young children. Int
J Environ Health Res, 17: 351-360 . 092842
Roemer WH; Van Wijnen JH. (2001). Daily mortality and air pollution along busy streets in Amsterdam, 1987-1998.
Epidemiology, 12: 649-653. 019391
Rosenlund M; Bellander T; Nordquist T; Alfredsson L. (2009). Traffic-generated air pollution and myocardial infarction.
Epidemiology, 20: 265-71. 190309
September 2009 C-105 DRAFT - DO NOT CITE OR QUOTE
-------
Rosenlund M; Berglind N; Pershagen G; Hallqvist J; Jonson T; Bellander T. (2006). Long-term exposure to urban air
pollution and myocardial infarction. Epidemiology, 17: 383-390. 089796
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
JacqueminB; vonKlot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J;
Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial
infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl R; Ibald-Mulli A; Koenig WSchneider A; Woelke G; Cyrys J; Heinrich J; Marder V; Frampton M; Wichmann HE;
Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients with coronary heart
disease. Environ Health Perspect, 173: 432-441. 088754
Rudez G; Janssen NA; Kilinc E; Leebeek FW; Gerlofs-Nijland ME; Spronk HM; ten Gate H; Cassee FR; de Maat MR
(2009). Effects of ambient air pollution on hemostasis and inflammation. Environ Health Perspect, 117: 995-1001.
193783
Salam MT; Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal exposure to
ozone, carbon monoxide, and particulate matter: results from the Children's Health Study. Environ Health Perspect,
113: 1638-1644. 087885
Samet JM; Dominici F; Curriero FC; Coursac I; Zeger SL. (2000). Fine particulate air pollution and mortality in 20 US
cities, 1987-1994. NEnglJMed, 343: 1742-1749. 013132
Samoli E; Touloumi G; Schwartz J; Anderson HR; Schindler C; Forsberg B; Vigotti MA; Vonk J; Kosnik M; Skorkovsky J;
Katsouyanni K. (2007). Short-term effects of carbon monoxide on mortality: An analysis within the APHEA
project. Environ Health Perspect, 115: 1578-1583. 098420
Sari I; Zengin S; Ozer O; Davutoglu V; Yildirim C; Aksoy M. (2008). Chronic carbon monoxide exposure increases
electrocardiographic P-wave and QT dispersion. Inhal Toxicol, 20: 879-84. 190315
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Sauerzapf V; Jones AP; Cross J. (2009). Environmental factors and hospitalisation for chronic obstructive pulmonary
disease in a rural county of England. J Epidemiol Community Health, 63: 324-328. 180082
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG (2006). Ambient air pollution and asthma
exacerbations in children: An eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G; Speizer
FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly subjects. Thorax, 60:
455-461.074317
Schwartz J; Norris G; Larson T; Sheppard L; Claiborne C; Koenig J. (1999). Episodes of high coarse particle
concentrations are not associated with increased mortality. Environ Health Perspect, 107: 339-342. 017915
Sharovsky R; Cesar LA; Ramires JA. (2004). Temperature, air pollution, and mortality from myocardial infarction in Sao
Paulo, Brazil. Braz J Med Biol Res, 37: 1651-1657. 156976
Sheppard L; Levy D; Norris G; Larson TV; Koenig JQ. (1999). Effects of ambient air pollution on nonelderly asthma
hospital admissions in Seattle, Washington, 1987-1994. Epidemiology, 10: 23-30. 086921
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and disease
parameters in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J Allergy Clin
Immunol, 115: 337-344. 087471
Singh V; Khandelwal R; Gupta AB. (2003). Effect of air pollution on peak expiratory flow rate variability. J Asthma, 40:
81-86. 052686
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci Environ Epidemiol,
15: 153-159.073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on symptom
severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346-353. 086294
September 2009 C-106 DRAFT - DO NOT CITE OR QUOTE
-------
Sole D; Camelo-Nunes 1C; Wandalsen GF; Pastorino AC; Jacob CMA; Gonzalez C; WandalsenNF; Rosario Filho NA;
Fischer GB; Naspitz CK. (2007). Prevalence of symptoms of asthma, rhinitis, and atopic eczema in Brazilian
adolescents related to exposure to gaseous air pollutants and socioeconomic status. J Investig Allergol Clin
Immunol, 17: 6-13. 090706
Son JY; Cho YS; Lee JT. (2008). Effects of air pollution on postneonatal infant mortality among firstborn infants in Seoul,
Korea: case-crossover and time-series analyses. Arch Environ Occup Health, 63: 108-13. 190323
Steerenberg PA; Nierkens S; Fischer PH; Van Loveren H; Opperhuizen A; Vos JG; Van Amsterdam JG (2001). Traffic-
related air pollution affects peak expiratory flow, exhaled nitric oxide, and inflammatory nasal markers. Arch
Environ Occup Health, 56: 167-174. 017157
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution and
inflammation-sensitive biomarkers. Environ Res, 106: 51-61. 188893
Stieb DM; Beveridge RC; Brook JR; Smith-Doiron M; Burnett RT; Dales RE; Beaulieu S; Judek S; Mamedov A. (2000).
Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint John, Canada. J Expo Sci
Environ Epidemiol, 10: 461-477. 011675
Stieb DM; Judek S; Burnett RT. (2003). Meta-analysis of time-series studies of air pollution and mortality: Update in
relation to the use of generalized additive models. J Air Waste Manag Assoc, 53: 258-261. 056908
Strickland MJ; Klein M; Correa A; Reller MD; Mahle WT; Riehle-Colarusso TJ; Botto LD; Flanders WD; Mulholland JA;
Siffel C; Marcus M; Tolbert PE. (2009). Ambient air pollution and cardiovascular malformations in Atlanta,
Georgia, 1986-2003. Am J Epidemiol, 169: 1004-14. 190324
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Heinrich J; Wichmann H-E; Peters A. (2007). Daily
mortality and particulate matter in different size classes in Erfurt, Germany. J Expo Sci Environ Epidemiol, 17:
458-467. 091374
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine particulate
matter and primary cardiac arrest among persons with and without clinically recognized heart disease. Am J
Epidemiol, 157: 501-509. 043156
Sun HL; Chou MC; Lue KH. (2006). The relationship of air pollution to ED visits for asthma differ between children and
adults. Am J Emerg Med, 24: 709-713. 090768
Sunyer J; Basagana X. (2001). Particles, and not gases, are associated with the risk of death in patients with chronic
obstructive pulmonary disease. Int J Epidemiol, 30: 1138-1140. 019367
Sunyer J; Basagana X; Belmonte J; Anto JM. (2002). Effect of nitrogen dioxide and ozone on the risk of dying in patients
with severe asthma. Thorax, 57: 687-693. 034835
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Szyszkowicz M. (2007). Air pollution and emergency department visits for ischemic heart disease in Montreal, Canada. Int
J Occup Med Environ Health, 20: 167-173. 193793
Tarkiainen TH; Timonen KL; Vanninen EJ; Aim S; Hartikainen JEK; Pekkanen J. (2003). Effect of acute carbon monoxide
exposure on heart rate variability in patients with coronary artery disease. Clin Physiol Funct Imaging, 23: 98-102.
053625
Tenias JM; Ballester F; Perez-Hoyos S; Rivera ML. (2002). Air pollution and hospital emergency room admissions for
chronic obstructive pulmonary disease in Valencia, Spain. Arch Environ Occup Health, 57: 41-47. 026077
Thompson AJ; Shields MD; Patterson CC. (2001). Acute asthma exacerbations and air pollutants in children living in
Belfast, Northern Ireland. Arch Environ Occup Health, 56: 234-241. 073513
Timonen KL; Pekkanen J; Tiittanen P; Salonen RO. (2002). Effects of air pollution on changes in lung function induced by
exercise in children with chronic respiratory symptoms. Occup Environ Med, 59: 129-134. 025653
Timonen KL; Vanninen E; De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G; Lanki T; Peters A;
Tarkiainen T; Tiittanen P; Kreyling W; Pekkanen J. (2006). Effects of ultrafine and fine particulate and gaseous air
pollution on cardiac autonomic control in subjects with coronary artery disease: The ULTRA study. J Expo Sci
Environ Epidemiol, 16: 332-341. 088747
September 2009 C-107 DRAFT - DO NOT CITE OR QUOTE
-------
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
Trapasso LM; Keith L. (1999). Relationships between selected meteorological/pollution parameters and hospital
admissions for asthma in south central kentucky . J Ky Acad Sci, 60: 73-77. 180127
Tsai S-S; Chen C-C; Hsieh H-J; Chang C-C; Yang C-Y. (2006). Air pollution and postneonatal mortality in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol, 18: 185-189. 090709
Tsai S-S; Cheng M-H; Chiu H-F; Wu T-N; Yang C-Y. (2006). Air pollution and hospital admissions for asthma in a tropical
city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 549-554. 089768
Tsai S-S; Goggins WB; Chiu H-F; Yang C-Y. (2003). Evidence for an association between air pollution and daily stroke
admissions in Kaohsiung, Taiwan. Stroke, 34: 2612-2616. 080133
Tsai S-S; Huang C-H; Goggins WB; Wu T-N; Yang C-Y. (2003). Relationship between air pollution and daily mortality in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 66: 1341-1349. 050480
Vedal S; Brauer M; White R; Petkau J. (2003). Air pollution and daily mortality in a city with low levels of pollution.
Environ HealthPerspect, 111: 45-51. 039044
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with implantable
cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Vigotti MA; Chiaverini F; Biagiola P; Rossi G (2007). Urban air pollution and emergency visits for respiratory complaints
in Pisa, Italy. J Toxicol Environ Health A, 70: 266-269. 090711
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of
air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci Environ Epidemiol, 13: 427-
435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Villeneuve PJ; Doiron M-S; Stieb D; Dales R; Burnett RT; Dugandzic R. (2006). Is outdoor air pollution associated with
physician visits for allergic rhinitis among the elderly in Toronto, Canada?. Allergy, 61: 750-758. 091179
Von Klot S; Peters A; Aalto P; Bellander T; Berglind N; DTppoliti D; Elosua R; Hermann A; Kulmala M; Lanki T; Lowel
H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible Subpopulations
(HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk of hospital cardiac
readmissions of myocardial infarction survivors in five European cities. Circulation, 112: 3073-3079. 088070
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A. (2002).
Increased asthma medication use in association with ambient fine and ultrafine particles. Eur Respir J, 20: 691-702.
034706
Wang Q; Liu Y; Pan X. (2008). Atmosphere pollutants and mortality rate of respiratory diseases in Beijing. Sci Total
Environ, 391: 143-148. 179974
Wang T-N; Ko Y-C; Chao Y-Y; Huang C-C; Lin R-S. (1999). Association between indoor and outdoor air pollution and
adolescent asthma from 1995 to 1996 in Taiwan. Environ Res, 81: 239-247. 008105
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and hemorrhagic
stroke among medicare beneficiaries. Stroke, 36: 2549-2553. 088685
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution on
functional status in patients with chronic congestive heart failure: a repeated-measures study. Environ Health, 6: 1-
7. 092830
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air pollution
and heart rate variability differs for individuals with heart and pulmonary disease. Environ Health Perspect, 114:
560-566. 088453
September 2009 C-108 DRAFT - DO NOT CITE OR QUOTE
-------
Wichmann H-E; Spix C; Tuch T; Wolke G; Peters A; Heinrich J; Kreyling WG; Heyder J. (2000). Daily mortality and fine
and ultrafine particles in Erfurt, Germany Part I: role of particle number and particle mass. 013912
Wilhelm M; Meng Y; Rull RP (2008). Environmental Public Health Tracking of Childhood Asthma Using California
Health Interview Survey, Traffic, and Outdoor Air Pollution Data. Environ Health Perspect, 116: 1254-1260.
191912
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in Los Angeles
County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Woodruff T J; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-
2002. Environ Health Perspect, 116: 110-5. 098386
Xirasagar S; Lin HC; Liu TC. (2006). Seasonality in pediatric asthma admissions: the role of climate and environmental
factors. Eur J Pediatr, 165: 747-752. 093267
Yang C-Y; Chang C-C; Chuang H-Y; Tsai S-S; Wu T-N; Ho C-K. (2004). Relationship between air pollution and daily
mortality in a subtropical city: Taipei, Taiwan. Environ Int, 30: 519-523. 055603
Yang C-Y; Chen Y-S; Yang C-H; Ho S-C. (2004). Relationship between ambient air pollution and hospital admissions for
cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A, 67: 483-493. 094376
Yang CY (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei, Taiwan. J
Toxicol Environ Health A, 71: 1085-1090. 157160
Yang CY; Chen CC; Chen CY; Kuo HW. (2007). Air pollution and hospital admissions for asthma in a subtropical city:
Taipei, Taiwan. J Toxicol Environ Health A, 70: 111-117. 092848
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-1219. 092847
Yang Q; Chen Y; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). Effect of short-term exposure to low levels of
gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environ Res, 99: 99-105. 090184
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air pollution and
first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch Environ Occup Health, 59: 14-
21.087488
Yang Q; Chen Y; Shi Y; Burnett RT; McGrail KM; Krewski D. (2003). Association between ozone and respiratory
admissions among children and the elderly in Vancouver, Canada. Inhal Toxicol, 15: 1297-1308. 055621
Yu O; Sheppard L; Lumley T; Koenig JQ; S. (2000). Effects of ambient air pollution on symptoms of asthma in Seattle-
area children enrolled in the CAMP study. Environ Health Perspect, 108: 1209-1214. 013254
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold DR.
(2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. , 110: 2184-2189. 087489
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health,
60: 890-895. 090195
September 2009 C-109 DRAFT - DO NOT CITE OR QUOTE
-------
Annex D. Controlled Human
Exposure Studies
Table D-1 Controlled human exposure studies.
Study
Subjects
Exposure
Findings
Adiretal. (1999,
0010261
15 healthy non-
smokers
Gender: M
Age: 22-34 yr
Inhaled Concentration: Not provided
Exposure Duration: 3 min 45 s
COHb Concentration: 4-6%
COHb Analysis: CO-oximeter (IL-282)
Exposures to CO and room air were separated
by 1 mo with the order of exposure randomly
assigned.
Exposure to CO resulted in a decrease in post-exposure
exercise duration (Bruce protocol) relative to clean air
exposure in 13 out of 15 subjects (p=0.0012). Statistically
significant decreases in metabolic equivalent units (METs)
were also reported following CO exposure (p = 0.0001). No
CO-induced changes in heart rate (HR), BP, ECG parameters,
or myocardial perfusion were observed.
Bathoorn et al. (2007,
1939631
19 former smokers with
COPD
Inhaled Concentration: 100 ppm (9 subjects) or Following the fourth day of exposure, CO inhalation reduced
Gender: 18 M/1 F
Age: 66-70 yr
125ppm (10 subjects)
Exposure Duration: 2-h on each of four
consecutive days
COHb Concentration: 2.7% (following fourth
day exposure)
COHb Analysis: Not provided
Exposures to CO and room air conducted were
separated by at least 1 wk using a randomized
crossover design.
sputum eosinophils relative to room air and also increased the
provocative concentration of methacholine required to cause a
20% reduction in FEV1. Neither of these effects were shown to
reach statistical significance. No changes in sputum
neutrophils, white blood cell counts or serum C-reactive
protein (CRP) were observed. Although this study appears to
demonstrate some evidence of an anti-inflammatory effect of
CO among subjects with COPD, it must be noted that two of
these patients experienced exacerbations of COPD during or
following CO exposure, with one patient requiring
hospitalization two mo after exposure (initial symptoms first
experienced 1-wk post-exposure).
Hanada et al. (2003,
1939151
20 healthy adults Inhaled Concentration: Not provided
Exposure Duration: 20 min
Gender: M COHb Concentration: 20-24%
Age: 26 + 1 yr COHb Analysis: CO-oximeter (OSM-3)
15 subjects exposed for 20 min (10 min rest, 5
min handgrip exercise, 2 min post-exercise
ischemia, 3 min recovery) under the following
four conditions: (1) normoxia (inspiratory 02
fraction 21.4%), (2) hypoxia (inspiratory 02
fraction 10.3%), (3) CO + normoxia, (4) CO +
hyperoxia (inspiratory 02 fraction 95.9%). Trials
involving exposure to CO were conducted last in
this sequence. Each of the four conditions was
separated from the next by 20 min of rest. 5
subjects served as controls (four consecutive 20
min periods of normoxia).
Blood oxygenation, BP, HR and respiratory rate were
measured during exposure. Muscle sympathetic nerve activity
(MSNA) and leg hemodynamics were evaluated in two subsets
of the study group (n = 8 and 7, respectively). Arterial oxygen
saturation (pulse oximetry) was significantly lower, and resting
HR and ventilation significantly higher during the period of
hypoxia compared to the other periods; none of these
measures were affected by exposure to CO. MSNA was
shown to increase during hypoxia and CO exposure relative to
normoxia. Neither hypoxia nor CO was found to affect leg
blood flow or vasoconstriction.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009
D-1
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Subjects
Kizakevich et al. (2000, 16 healthy non-
052691) smokers
Gender: M
Age: 18-29 yr
Exposure
Inhaled Concentration: Initial short term (4-6
min) exposure to 1,000 or 3,000 ppm followed by
exposures to 27, 55, 83, or 100 ppm to maintain
COHb Concentration.
Exposure Duration: 4-6 min at 1,000 or 3,000
ppm followed by 20 min at 27, 55, 83, or 100
ppm.
Target COHb Concentrations: 5,10,15, and
Findings
At all levels of upper- and lower-body exercise, exposures to
CO resulted in increases in HR, cardiac output, and cardiac
contractility relative to clean air exposures. Increases in HR
reached statistical significance at COHb Concentrations 2
5%, and increases in both cardiac output and cardiac
contractility reached statistical significance at COHb
Concentrations 2 10%. CO exposure during exercise was not
observed to cause ventricular arrhythmias or affect ECG wave
shape (no evidence of ST-segment depression) at COHb
Concentrations < 20%.
COHb Analysis: CO-oximeter (IL-282)
Subjects exposed on 4 separate days to
increasing CO concentrations during either
upper-body exercise (hand-crank) or lower-body
exercise (treadmill). Targeted COHb
Concentrations were initially attained using
short term (4-6 min) exposures to CO at
concentrations of 1,000 or 3,000 ppm. Chamber
exposures were then conducted at CO
concentrations required to maintain COHb levels
of <2% (room air), 5% (27 ppm), 10% (55 ppm),
15% (83 ppm), and 20% (100 ppm).
Mayretal. (2005,
1939841
13 healthy non-
smokers
Gender: M
Inhaled Concentration: 500 ppm
Exposure Duration: 1 h
COHb Concentration: 7%
COHb Analysis: CO-oximeter (AVL 91 2)
Infusion of IPS significantly increased plasma concentrations
of TNF-a, CRP, IL-6, and IL-8, with no difference in the
inflammatory response between clean air and CO exposures.
Age: 18-38 yr
Subjects exposed to both CO and clean air with
exposures separated by a 6-wk period.
Immediately following exposure, subjects were
administered an intravenous bolus dose
(2 ng/kg) of lipopolysaccharide (IPS).
September 2009
D-2
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Subjects
Exposure
Findings
Morse et al. (2008,
0979801
12 healthy non-
smokers
Gender: M
Age: 25 ±2.9 yr
Inhaled Concentration: 3,000 ppm
Exposure Duration: 3-8 min
COHb Concentration: 6.2%
COHb Analysis: Electrochemical sensor
(Smokerlyzer) measuring CO in exhaled breath
Leg strength and muscle fatigue were evaluated immediately
following exposure. CO exposure did not affect muscle
strength (maximal voluntary isometric contraction), but did
cause a statistically significant increase in muscle fatigue (p
<0.05).
Exposures conducted on two separate occasions
to room air (6 min) and CO. Subjects were
exposed to CO until COHb reached 6% (3-8 min
exposures).
Ren etal. (2001,
1938501
12 healthy adults
(10nonsmokersand 1
smoker)
Gender: 9 M/3 F
Age: 20-32 yr
Inhaled Concentration: 0.4% (= 4,000 ppm)
A statistically significant increase in ventilation was observed
following hypoxia, but no such increase was found following
Exposure Duration: 10-30 min at 0.4% followed any Of the other 3 protocols, including exposure to CO. One
by ~ 8-h with periodic exposure to maintain subject fe|( faint during the blood withdrawal protocol and did
COHb Concentration not complete the study.
COHb Concentration: 10%
COHb Analysis: Not provided
Each subject underwent four different 8-h
experimental protocols: (1) isocapnic hypoxia
(end-tidal P02 held at 55 mmHg), (2) withdrawal
of 500 ml of venous blood at the start of an 8-h
period, (3) CO exposure at a concentration
required to maintain a COHb level of 10%, and
(4) a control exposure where subjects breathed
room air with no intervention.
Resch et al. (2005,
1938531
15 healthy non-
smokers
Gender: M
Age: 27 + 4 yr
Inhaled Concentration: 500 ppm
Exposure Duration: 1 h
COHb Concentration:-10%
COHb Analysis: CO-oximeter (AVL 912)
COHb levels averaged 5.6% after 30 min and 9.4% after 60
min of exposure. Statistically significant increases in retinal
blood flow, retinal vessel diameter, and choroidal blood flow
were observed with CO exposure relative to synthetic air at
both time points. Exposure to CO did not affect oxygen
saturation of arterial blood.
Exposures to CO and synthetic air control were
separated by a period of at least 1 wk.
Vesely et al. (2004,
1940001
10 healthy non-
smokers
Gender: M
Age: 22-52 yr
Inhaled Concentration: 1,200 ppm
Exposure Duration: 30-45 min
COHb Concentration: 10%
COHb Analysis: CO-oximeter (OSM-3)
Ventilation rate was observed to significantly increase during
hypoxic rebreathing relative to hyperoxic rebreathing.
However, exposure to CO had no effect on ventilation under
either hypoxic or hyperoxic conditions. The authors concluded
that exposure to low levels of CO does not significantly affect
chemoreflex sensitivity of the C02-induced stimulation of
ventilation.
Prior to and following exposure, subjects
performed hypoxic and hyperoxic rebreathing
tests. Four subjects were exposed to hypoxic
conditions first, while six subjects were exposed
to hyperoxic conditions first, both prior to and
following CO exposure.
September 2009
D-3
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Zevinetal. (2001,
0211201
Subjects
Exposure
12 healthy smokers Inhaled Concentration: 1,200-1,500 ppm
Gender: M
Age: 27-47 yr
Exposure Duration: 10 min each h, 16-h each
day, over 7 days
COHb Concentration: 5-6%
COHb Analysis: CO-oximeter (Ciba Corning
2500)
Findings
COHb levels were similar during smoking and exposure to CO,
with avg concentrations of 6% and 5%, respectively. Blood
was drawn on day 4 of each exposure and analyzed for CRP,
plasma platelet factor 4, and white blood cell count. Plasma
levels of CRP and platelet factor 4 were significantly elevated
with smoking, but not with CO exposure, relative to air control.
HR and BP were evaluated on day 3 of each protocol.
Cigarette smoke, but not CO, was observed to significantly
increase HR, while no difference in BP was observed between
any of the three exposures.
Exposures were conducted over 21 consecutive
days under three different protocols, with each
protocol lasting 7 days. In one protocol, subjects
smoked 20 cigarettes per day, one every 45 min.
In the other two protocols, every 45 min (20
times per day) subjects breathed either air or CO
from a 1 liter bag once per min for 10 min at a
time. Subjects completed all three protocols, with
six subjects exposed sequentially to CO,
smoking, then air, and the other six exposed
sequentially to air, smoking, then CO.
September 2009
D-4
DRAFT - DO NOT CITE OR QUOTE
-------
References
Adir Y; Merdler A; Haim SB; Front A; Harduf R; Bitterman H. (1999). Effects of exposure to low concentrations of carbon
monoxide on exercise performance and myocardial perfusion in young healthy men. Occup Environ Med, 56: 535-
538. 001026
Bathoorn E; Slebos DJ; Postma DS; Koeter GH; van Oosterhout AJ; van der Toorn M; Boezen HM; Kerstjens HA. (2007).
Anti-inflammatory effects of inhaled carbon monoxide in patients with COPD: Apilot study. Eur Respir J, 30:
1131-1137. 193963
Hanada A; Sander M; Gonzalez-Alonso J. (2003). Human skeletal muscle sympathetic nerve activity, heart rate and limb
haemodynamics with reduced blood oxygenation and exercise. J Physiol, 551: 635-647. 193915
Kizakevich PN; McCartney ML; Hazucha MJ; Sleet LH; Jochem WJ; Hackney AC; Bolick K. (2000). Noninvasive
ambulatory assessment of cardiac function in healthy men exposed to carbon monoxide during upper and lower
body exercise. Eur J Appl Physiol, 83: 7-16. 052691
Mayr FB; Spiel A; Leitner J; Marsik C; Germann P; Ullrich R; Wagner O; Jilma B. (2005). Effects of carbon monoxide
inhilation during experimental endotoxemia in humans. Am J Respir Crit Care Med, 171: 354-360. 193984
Morse CI; Pritchard LJ; Wust RC; Jones DA; Degens H. (2008). Carbon monoxide inhalation reduces skeletal muscle
fatigue resistance. Eur Phys J A, 192: 397-401. 097980
RenX; DorringtonKL; Robbins PA. (2001). Respiratory control in humans after 8 h of lowered arterial PO2,
hemodilution, or carboxyhemoglobinemia. J Appl Physiol, 90: 1189-1195. 193850
Resch H; Zawinka C; Weigert G; Schmetterer L; Garhofer G (2005). Inhaled carbon monoxide increases retinal and
choroidal blood flow in healthy humans. Invest Ophthalmol Vis Sci, 46: 4275-4280. 193853
Vesely AE; Somogyi RB; Sasano H; Sasano N; Fisher JA; Duffin J. (2004). The effects of carbon monoxide on respiratory
chemoreflexes in humans. Environ Res, 94: 227-233. 194000
Zevin S; Saunders S; Gourlay SG; Jacob PHI; Benowitz NL. (2001). Cardiovascular effects of carbon monoxide and
cigarette smoking. J Am Coll Cardiol, 38: 1633-1638. 021120
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 D-5 DRAFT - DO NOT CITE OR QUOTE
-------
Annex E. Toxicological Studies
Table E-1. Human and animal studies.
Reference Species / Model
CO Concentration
Findings
Acevedo and Ahmed Human pregnant
(1998,0160031 myometrium
HO-1 and HO-2 (mRNAand protein) were upregulated in
pregnant myometrium when compared to non-pregnant
myometrium. The HO activator hemin inhibited spontaneous and
oxytocin-induced contractility of the myometrium. Progesterone
induced HO-1 and HO-2 mRNA expression.
Humanarteries
Until equilibrium Approximately 30 pM
CO induced endothelium- and N0--independent relaxation of
precontracted human ITA and RA graft by partially stimulating
cGMP production. The mechanism and extent of relaxation
depended upon the tissue.
Si8'3'''2000' Human placenta
Placental HO-1 was significantly higher at term. HO-1
significantly attenuated TNFa-dependent cellular damage in
placental explants. HO-1 was significantly attenuated in pre-
eclampsia pregnancies vs non-pre-eclamptic pregnancies.
Placental arteries exposed to the HO activator hemin
demonstrated reduced vascular tension (i.e., placental blood
vessel relaxation).
Ahmed et al. (2005,
1938651
Human placental
cotyledons
The source of CO in term human placental chorionic villi was
found to be the catalysis of heme by HO and not endogenous
lipid peroxidation.
«reta,.(2007, S^fagueDawley
-^222) Adult female
Modulation of the HO/CO system in the anterior pituitary of the
female rat led to altered secretion of gonadotropins and prolactin.
Rat
Alexandreanu et al. o^r,,,^ rv*i,i*w
(2002,192373) Female
The role of the HO/CO system in estrous cyclicity, pregnancy and
lactation was evaluated using HO inhibitors and substrates. The
HO inhibitor CrMP decreased time in estrous. Administering HO-
inhibitors to pregnant rodents induced total litter loss. CrMP
induced decreased litter weight gain during lactation, which the
authors attribute to maternal milk production or ejection problems
as cross-fostered pups regained weight lost during nursing on
CrMP dams.
Alexandreanu and Rat
Lawson ((2003, Sprague Dawley
1938711 Adult female
Modulation of the HO/CO system in the anterior pituitary of the
female rat led to altered secretion of gonadotropins and prolactin.
Alexandreanu and Rat
Lawson (2003, Sprague Dawley
1938761 Adult female ovary
HO-1 and HO-2 were localized in the ovaries in rats and
treatment of rat ovaries in vitro with CrMP, an inhibitor of HO, or
with hemin, a substrate for HO induced steroidogenic changes in
the ovaries.
Alonso et al. (2003,
1938821
Human muscle tissue
mitochondria
5min
50-500 ppm
CO significantly reduced muscle mitochondrial cytochrome c
oxidase activity by 20%, 42%, and 55% after treatment with 50,
100, and 500 ppm CO respectively but did not change the activity
of three other electron transport proteins.
Rat
Long Evans
Male
Andersen et al. (2006, .,
1804491 Mouse
•lm^> C57BL/6J
Male
Cerebral vessels
1-100 pM
CO did not dilate rat or mouse cerebral arteries until 100 pM,
which is not a physiological concentration. Also, the HO inhibitors
constricted vessels in a nonspecific manner.
Antonelli et al. (2006, Rat
1938851 Wistar
GD5-GD20
75 ppm
Pups exposed to CO in utero had significant impairment of
cortical neuronal glutamanergic transmission at PND1 in both
neurons at rest and in neurons stimulated with depolarization.
September 2009
E-1
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Appleton and Marks H , t
(2002, 193935) numan Placenla
Endogenous CO production by HO in the human placenta was
regulated by 02 availability. Placental HO activity was directly
dependent on 02 availability; this does not vary between pre-
eclamptic and normotensive placentas.
Ashfeqetal.(2003, Human p|acenta
Placentas were collected from smokers and nonsmokers who
gave birth to male infants. Premature aging and a statistically
significant increase in apoptotic cells were seen in placentas from
smokers vs nonsmokers.
Astrupetal. (1972, Rabbit
0111211 (strain not identified)
Continuous CO
exposure over
gestation
90 or 180 ppm
Skeletal abnormalities: Three pups (from n = 123) in the 180 ppm
CO group had deformities in their extremities at birth, whereas no
control and no 90 ppm CO-exposed animals manifested with this
malformation.
72-3369 nM
Isolated human placenta exposed to solutions containing CO
demonstrated a concentration-dependent decrease in perfusion
pressure further demonstrating the role of CO in maintaining
basal vasculature tone.
Bainbridge et al.
(2006, 1939491
Human placenta
6h
Starting concentrations of
CO: 3.9 pM CO in cell
culture media (control)
and CO-exposed groups:
116 pM, 145 pM,
181 fjM.
After 3 h, the CO in the
culture media was
3.7 pM (control), and
CO-exposed cells
10.2,12, and 15.9 pM.
C-section placentas were collected from healthy term
pregnancies.Villous explants of placentas were cultured under
hypoxia followed by reoxygenation (H/R). H/R + CO-exposed
placental tissue had decreased apoptosis and decreased PARP
(a protein marker of apoptosis) vs control H/R exposed cells.
Secondary necrosis of the placental tissue post H/R was inhibited
by CO treatment.
Human placenta
The role of HO in the placenta and during pregnancy are
reviewed in this article. The conflicting data on the activity,
localization, and expression of HO in the placentas of pre-
eclamptic women are presented.
Human placenta
Expression and tissue localization of soluble guanylyl cyclase in
human placenta using antibody localization were characterized.
These tools can be used in future studies to elucidate the
NO'/CO/cGMP pathway.
?9a39e53)taL(19"' Human myometrium
HO and NOS did not maintain human uterine quiescence during
pregnancy.
Barber etal. (2001, Human placenta
Women who had pregnancies with fetal growth restrictions (FGR)
produced term placenta with significant decreases in HO-2 vs
healthy pregnancies.
End-tidal CO measurements in women with pregnancy-induced
hypertension and pre-eclampsia were significantly lower than in
normotensive pregnant women.
Benagiano et al.
(2005, 1804451
Rat
Wistar
Female
GDO-GD20
75 ppm
CO caused a significant reduction in glutamic acid decarboxylase
and GABA immunoreactivities in the cerebellar cortex of adult
rats prenatally exposed to CO (number of positive neuronal
bodies and axon terminals and the area they covered). No
difference was found in the microscopic structure of the
cerebellar cortex or distribution patterns of GAD or GABA.
Benagiano (2007,
1938921
Rat
Wistar
Female
GD5-GD20
75 ppm
Prenatal CO reduced GAD and GABA immunoreactivities. There
were no structural alterations of the cerebellar cortex.
Bergeron etal. (1998, Rat
Brain
To address the developmental changes of HO staining in the
brain, immunohistochemical staining for HO-1 was performed on
the developing rat brain at PND7, PND14, and PND21. HO-1
staining was most intense at PND7 and by PND21 reached its
adult pattern of staining localizing to the hippocampus, thalamic
and hypothalamic nuclei, with virtually no staining of endothelium,
white matter and cortex. HO-2 is the dominant HO isoform in the
brain.
Bing etal. (1995, D . t
079418) Rodent
Spatial learning in the Morris water maze was enhanced in
rodents exposed to the HO inhibitor tin protoporphyrin (Sn-PP).
September 2009
E-2
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Burmester et al.
(2000. 099998)
Bye et al. (2008,
1937771
Cagianoetal. (1998,
0871701
Carmines and
Rajendran (2008,
1884401
Carratuetal. (1993,
0138121
Species / Model
Human
and
Mouse
Rat
Wistar
Female
Rat
Wistar
Female
Rat
Sprague Dawley
Rat
Wistar
Male pups
Exposure
Duration
100h/wkfor
18 mo
GDO-GD20
GD6-GD19of
gestation for
2 h/day
GDO-GD20
CO Concentration Findings
Nb had a high oxygen affinity similar to Mb, thus may increase
the availability of 0, to brain tissue.
CO-exposed (11-14.7% COHb) rats experienced a 24% decrease
in aerobic capacity evidenced by V02 max deficits. Left
2W PPm ventricular cardiomyocytes were longer and wider, had increased
expression of growth-related proteins, and had impaired
contraction-relaxation cycles. CO increased cGMP and impaired
cardiomyocyte Ca2* handling. No change in BP was observed.
At 5 mo of age, CO-exposed male offspring showed decrements
in sexual behavior including an increase in mount to intromission
latency, a decrease in mount to intromission frequency, and a
75 or 150 Dom decrease in ejaculation frequency. Basal extracellular dopamine
ppm concentration in the nucleus accumbens was unchanged after
CO-exposure. However, when stimulated with amphetamine
administration, control rats had increased release of dopamine
that is absent with CO-exposed rats.
Significant decreases in birth weight were reported after CO
exposure. Maternal body weight was unchanged during
600 ppm gestation, but corrected terminal body weight (body weight minus
uterine weight) was significantly elevated in CO-exposed dams at
term.
Prenatal CO exposure slowed the inactivation kinetics of
transient sodium current in the sciatic nerve fibers of 40-day-old
7"- I5n m male rats. The maximum number of activatable Na channels at
ppm normg| restjng p0tentja| was jncreasec| jn co exposed rats and
the voltage-current relationship showed a negative shift of
sodium equilibrium potential.
Carratuetal. (1995, Rat
0794271 (Wstar)
150 ppm
Sphingolipid homeostasis was disrupted in male offspring of
prenatally exposed rats, without a disruption in motor function.
Carratu et al. (2000,
0159351
Rat
Wstar
GDO-GD20
150ppm
Maternal COHb (mean % ± SEM) was 1.9 ±0.04 and 16.02 ±
0.98 in control and 150 ppm CO-exposed animals, respectively.
Prenatal CO exposure had no effect on brain sphinganine (SA) or
sphingosine (SO) levels in male offspring at 90 days of age.
However, the sciatic nerve had significant increases in SO after
CO exposure, no changes in SA at 90 days of age. Motor activity,
which could be affected by changes in myelination, showed no
differences between CO and control animals at 90 days of age.
Carratu et al. (2000,
0158391
Rat
Wstar
GDO-GD20
75 or 100 ppm
The myelin sheath thickness of the nerve fibers was significantly
decreased in CO-exposed animals (75 and 150 ppm). Axon
diameter was not affected by CO exposure. Even though CO
affected myelination, it did not significantly affect motor activity of
CO-exposed rats at 40 and 90 days.
Rat model of hypoxic
Carraway et al. (2002, pulmonary vascular
0260181 remodeling
(Strain of rat not stated)
3wk
Hypobaric hypoxia +
50 ppm
CO promoted remodeling and increased pulmonary vascular
resistance in response to HH. The number of small muscular
vessels was increased compared with HH alone. Changes in cell
proliferation, apoptosis, actin and HO-1 gene and protein
expression correlated with structural changes. COHb levels were
<0.5% in controls, 1.5-2.8% in the HH treatment group and
3.5-3.9% in the HH + CO treatment group.
Cella et al. (2006,
1932401
Rat
Sprague Dawley
HO-1 production and HO concentration were shown to be
regulated by estrogen in the rat uterus.
Rat
Long Evans
Male
2 mo old
3.5 h
1201 +18 ppm
CO potentiates noise induced hearing loss. The NMDA inhibitor
(+1-MK-801 did not block the potentiation of the NIHL by CO.
Cheng et al. (2009,
1937751
Human atherectomy
biopsy (clinical carotid
artery disease)
Mouse model of
vulnerable plaque
ApoE-/- mouse
HO-1 expression correlated with features of vulnerable human
atheromatous plaque. HO-1 expression was upregulated in
vulnerable lesions in the mouse model. Induction of HO-1 in the
mouse impeded lesion progression into vulnerable plaques.
Inhibition of HO-1 augmented plaque vulnerability.
Overexpression of HO-1 resulted in plaque stabilization. It was
concluded that HO-1 induction was atheroprotective.
Rat
Sprague Dawley
Male
3-6%
CO inactivation of Mb does not induce any change in the
respiration rate, contractile function, or high-energy phosphate
levels in perfused rat hearts.
September 2009
E-3
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Cronje et al. (2004,
1804401
Rat
Sprague Dawley
Male
240-325 g
45 min
2,500 ppm
Results indicate that tissue and blood [CO] (66-72% COHb)
dissociate during CO inhalation, but tissue [CO] does not follow
blood [CO] or 1/p02 as in the Warburg theory during intake or
elimination. Tissue [CO] increases later during the resolution
period and varies significantly among animals and tissues. The
deviation from the predicted values in the brain is likely due to the
release of heme and increase in NADPH stimulating endogenous
CO production by HO. Immediately following exposure, tissue
CO concentrations were found to be:
Blood: 27,500 (800) pmol/mg
Heart: 800 (300) pmol/mg
Muscle: 90 (80) pmol/mg
Brain: 60 (40) pmol/mg
These values are estimates taken from a graph, with control
levels in parentheses
A later report stated that these tissue CO values were too high
due to a computational error (Piantadosi et al., 2006,1804241
Human placenta
Human (HUVEC)
Mouse
Cudmore et al (2007 (HO-1 deficient mouse on
1939911 ' 129/SVxC57BL/6
background)
Pig
(Porcine aortic endothelial
cells)
HUVEC cells, porcine aortic endothelial cells, HO-1 null mice and
placental villous explants (normotensive and pre-eclamptic
pregnancies) were used in this study. The HO-1/CO system
inhibited sFlt-1 and sEng release, two factors upregulated in pre-
eclampsia.
D'Amico et al. (2006, Human embryonic kidney n ,n min
1939921 (HEK293) cells u-dumm
Exogenous CO inhibited respiration in HEK293 cells under
ambient 02 concentration (21%). Inhibition was enhanced under
hypoxic conditions. Increased endogenous CO resulting from
HO-1 overexpression inhibited respiration by 12% and
cytochrome c oxidase activity by 23%. This effect was enhanced
under hypoxic conditions.
Dani et al. (2007, Human
1939941 (neonatal blood)
CO was lower at birth and 48-72 h postpartum in infants born by
elective C-section and higher in vaginally born infants.
Rat
DeLucaetal. (1996, Wistar
0809111 Female
Male pups
GDO-GD20
75 or 150 ppm
Prenatal CO (150 ppm) delayed development of the ion channels
responsible for passive and active membrane electrical
properties of skeletal muscle. CO induced lower values of resting
chloride conductance was reversed at PND80. CO induced
delayed developmental reduction of resting potassium
conductance was reversed at PND60.
De Salvia et al. (1995, Rat
0794411 Wistar
Animals exposed to the higher dose of CO (150 ppm) in utero
GDO-GD20 75 or 150 ppm had significantly impaired acquisition (at 3 and 18 mo) and
reacquisition (at 18 mo) of conditioned avoidance behavior.
Denschlag et al. ullm-n
(2004, 1938941 Muman
Genetic polymorphisms in human HO-1 are linked to idiopathic
recurrent miscarriages.
Dewildeetal. (2001,
0193181
Nb exists as a reversibly hexacoordinated Hb type with a His-
Fe2+-His binding scheme. Dissociation of the internal ligand by
02 or CO is the rate limiting step.
Di Giovanni et al.
(1993, 0138221
Rat
GDO-GD20
75 and 150 ppm
CO (150 ppm) reduced the minimum frequency of ultrasonic calls
as well as decreased responsiveness to a challenge dose of
diazepam. There was no change in locomotion however CO
impaired learning in a two-way active avoidance task.
Rat
Dubois et al. (2002, Wistar
1939111 Adult female
250 g
3wk
530 ppm
Intrapulmonary resistance artery smooth muscle cells were
isolated from control and exposed rats. Electrophysiological re-
cordings provided evidence of increased Ca2*-activated K+
current consequent to chronic CO exposure. The authors specu-
lated that this could in part explain the vasodilatory effect of CO
in the pulmonary circulation.
Dubois et al. (2005,
1804351
Rat
Wstar
Male
21 days
50 ppm
CO attenuated PAHT by activating BKCa channels in PA
myocytes and reduced hemodynamic changes of PAHT.
September 2009
E-4
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
CO Concentration
Findings
Duboisetal.(2003, ?
1804391 ™jffr
' Male
21 days
50 ppm
CO induced relaxation of pulmonary artery rings in normoxic,
hypoxic, and hypoxic-CO rats and it was not endothelium
dependent. Chronic hypoxia decreased acute CO sensitivity,
while CO-hypoxia increased it. K+ channel blocker reduced this
effect while sGC blocker did not.
Durante et al. (2006,
1937781
Reviews the role of CO in cardiovascular function.
Favory et al. (2006,
1844621
Rat
250-300 g
(Strain not stated)
90min
250 ppm
CO inhibited myocardial permeabilized fiber respiration (complex
IV), increased coronary perfusion pressure and left ventricular
developed pressure (LVDP) first derivative and decreased the
cGMP/cAMP ratio in the heart. These changes were maintained
over 24-48 h of recovery in air. Cardiac function and vasodilatory
responses were evaluated at 3-h recovery in air. (3-adrenergic
blockade had no effect on coronary perfusion pressure or LVDP
first derivative. Total inhibition of vasodilator response to
acetylcholine and partial inhibition of vasodilator response to
nitroprusside were observed. An increase in myofilament calcium
sensitivity was also observed. Thus CO promotes abnormalities
in mitochondrial respiration, coronary vascular relaxation and
myocardial contractility. The authors speculated that CO may
have a detrimental effect on heart 02 supply-to-utilization which
could potentially lead to myocardial hypoxia because of the
increased 02 demand resulting from increased contractility, the
inhibited mitochondrial respiration and the reduced coronary
blood-flow reserve resulting from the decreased vasodilatory
capacity.
COHb was found to be 11% immediately after exposure. COHb
levels gradually returned to baseline (1.5%) over the next 96 h.
Fechter and Annau Rat
(1977, 0106881 Long Evans
Garofolo et al. (2002,
Human infants
Rat
Continuous CO
exposure
throughout
pregnancy
The authors found a 5% significantly decreased birth weights at
PND1 in gestationally CO-exposed pups vs control animals with
weight decrements persisting to weaning; lactational cross
fostering did not ameliorate the CO-dependent reduced growth
150 ppm CO rates. Dams exposed to CO during gestation had COHb over
gestation of 15% with control dams having less than 1%.
Decreased birth weight and pre-weaning weight were seen in
CO-exposed pups despite a lack of weight decrement in
CO-exposed damsvs air-exposed control dams.
Fechter etal. (1980,
0112941
Fechter and Annau
(1980, 0112951
Fechter etal. (1987,
0121941
Fechter etal. (1987,
0122591
Fechter etal. (1997,
0813221
Fechter etal. (1986,
012030)
Rat
Long Evans
Rat
Long Evans
Rat
Long-Evans
Male
Rat
(Long Evans)
Male
Guinea pigs
Continuous CO
exposure
throughout
pregnancy
Continuous CO
exposure
throughout
pregnancy
Continuous CO
exposure
throughout
pregnancy or from
GDOtoPNDIO
150 ppm
150 ppm
1-4mL/100gBW(ip)
75, 150, or 300 ppm
35 ml/kg gas (ip)
40% COHb
CO-exposed animals had cardiomegaly at birth (wet heart
weight) that dissipated by PND4.
CO-exposed animals had decreased birth weight, impaired
righting reflexes, impaired negative geotaxis, and delayed
homing behavior.
High dose CO led to dose-dependent, reversible loss of the
compound action potential sensitivity for high frequency tone
bursts. Also, CO produced a dose dependent elevation in the
cochlear blood flow.
The neostriatum of PND21 rat brains were collected and showed
disrupted development following CO exposure (GDO-PND10
group, 300 ppm CO). Dopamine levels were also significantly
elevated in CO-exposed animals (GDO-PND10, 150 and
300 ppm CO).
CO impairs high-frequency auditory sensitivity shown by
increased compound action potential threshold at higher test
frequencies. Free radical inhibitors blocked this response.
Reviews the effects of carbon monoxide on brain development.
Rat: PND2-PND5
Human infants who die from SIDS show decreased brainstem
muscarinic receptor binding vs infants dying from other causes.
B-adrenergic modulation of muscarinic receptors in developing
heart was observed.
Rodent B-adrenergic agonists at PND2-PND5 induced
muscarinic receptor decrement in adenylyl cyclase.
September 2009
E-5
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Gautier et al. (2007,
0964711
Rat
Wistar
Adult male
Model of right ventricular
hypertrophy secondary to
chronic hypoxia
3wkofHH±CO
in final wk
Or1 wkofCO
SOppm
CO altered the right ventricular adaptive response to pulmonary
hypertension which occurs secondarily to chronic hypoxia. Right
ventricular end-systolic pressure (RVESP) and right ventricular
shortening fraction (RVSF) were smaller in rats treated with
CO+HH compared with rats treated with HH alone. CO alone had
no effect on these measures. Hypobaric hypoxia had no effect on
left ventricular function while 00+ HH led to an increased left
ventricular shortening fraction (LVSF). CO alone led to a
decrease in LVSF and the mitral E-to-A ratio, indicative of a LV
filling impairment. Hypobaric hypoxia decreased the relative RV
perfusion and increased the relative LV perfusion. These effects
were prevented with concomitant exposure to CO although expo-
sure to CO alone had no effects on myocardial perfusion.
Morphologic and histologic analysis demonstrated RV hyper-
trophy in both the HH group and the CO+HH group and fibrotic
lesions in the CO+HH group. The authors concluded that the
1-wk exposure to 50 ppm CO had a deleterious effect on RV
myocardial perfusion adaptation to chronic hypoxia and pressure
overload. Although the reduced RV pressure overload was
beneficial it was counterbalanced by impaired RV perfusion and
redistribution of perfusion toward the LV.
Gaworski et al. (2004, Rat
1939331 Sprague Dawley
Rat
Sprague Dawley
Adult male
2 h/day, 7 days/wk
by nose-only
inhalation
Males: 4 wk prior
to and during
mating; and
Females: 2 wk
prior to mating,
during mating,
and through
weaning to
PND21
24 h
Cigarette smoke:
150, 300, or600mg/m3
Total Particulate Matter
(TPM)
SOppm
Maternal exposure to high concentrations of cigarette smoke
during gestation and lactation reduced pup birth weight and
retarded neonatal pup growth. Developmental and
neurobehavioral testing of neonates did not show any behavioral
effects following parental smoke exposure.
Mild neutrophil accumulation was observed in BALF
accompanied by increases in BALF MIP-2, protein and LDH. Iron
status was altered since CO exposure led to an increase in BALF
iron and ferritin, a decrease in lung non-heme iron and an
increase in liver non-heme iron.
Ghio et al. (2008,
0963211
Human bronchial
epithelial cells
(BEAS-2B)
2-24 h
10-100 ppm
CO exposure for 24 h led to a dose-dependent decrease in
cellular non-heme iron, with the effect at 10 ppm statistically
significant and the effect at 50 ppm maximal. This effect was
reversible since removing the cells after 2 h of CO and incubating
them in air restored non-heme iron concentrations at 24 h. A
dose-dependent decrease in cellular ferritin was observed
following exposure for 24 h to 50-500 ppm CO. In addition,
exposure to 50 ppm CO for 20 h blocked iron uptake by cells
while exposure to 50 ppm CO for 2 h increased iron release from
cells. Increased protein expression of the iron transporter DMT-1
was also noted after 24 h exposure to 50 ppm CO. Oxidative
stress, mediator release and cell proliferation were also
decreased by exposure to 50 ppm for 24 h. This effect was also
reversible upon removal to air. Effects of CO on cell proliferation
indices were mimicked by with the iron-depleting agent
deferoxamine. The authors concluded that CO exposure altered
lung iron homeostasis possibly by initially causing heme release
from proteins.
Giustino et al. (1999,
0115381
Rat
Wistar GDO-GD20
Male and pregnant female
75 or 150 ppm
This study showed that CO (75 and 150 ppm) exposed male
animals at 40 days of age had a significantly decreased time of
exploration of novel objects. The 150 ppm CO group showed a
lack of habituation after the second exposure to a previously
viewed object. Blood COHb concentrations (mean % + SEM) on
GD20were reported (0 ppm: 1.6 + 0.1; CO 75 ppm: 7.36 + 0.2;
CO 150 ppm: 16.1+0.9).
Giustino etal. (1993, Rat
0138331 (Wistar
GDO-GD20
75 or 150 ppm
CO exposure in utero led to a reversible and dose dependent
loss of function of splenic macrophages with decreased killing
ability, decreased phagocytosis, and decreased ROS production
during the macrophage respiratory burst.
September 2009
E-6
DRAFT - DO NOT CITE OR QUOTE
-------
Reference Species / Model nlfratinn6 co Concentration
Findings
Giustinoetal. (1994,
0763431
;
Glabeetal. (1998,
0867041
Male pups
Rat
Sprague Dawley
Male,
Myocardium
CO (150 ppm) decreased the number of leukocyte common
antigen (LCA+) cells at PND21. This was reversed by PND540.
GDO-GD20 75 or 150 ppm CO (75 ppm) and other measures of immunological changes
showed trends toward reduction (macrophages, T cells, B cells,
and MHC II cells).
Increased PCO and increased COMb saturation did not alter high
energy phosphate signals (ATP, phosphocreatine, Pi). MV02
began to decline at 87.6% COMb and is likely not due to
cytochrome c oxidase inhibition.
PCO = 0-107Torr
Graver et al. (2000, Fetal lamb
0104651 (mixed breed)
10min
500 ppm
Fetal methoxyhemoglobin (COHb%) ranged from 3.8 + 0.2 to 8.1
+ 2.0 at 0 and 500 ppm CO, respectively. Inhaled 0-500 ppm CO
administered to near-term fetal lambs did not induce pulmonary
vasodilation (main pulmonary artery, left pulmonary artery, aorta
and left atrium) and the HO inhibitor zinc protoporphyrin IX failed
to affect baseline vascular tone.
CO exposure increased extracellular dopamine levels and
decreased its major metabolites in a Na -dependent pathway. CO
40 min 1,000-3,000 ppm withdrawal and reoxygenation caused levels to return to control
or overshoot which may suggest an increase in oxidative
metabolism of CO, mediated by MAO-A.
Haraetal.(2002,
Male
Harada et al. (2004, Pig
1939201 Granulosa cells
In this porcine model, HO was able to augment granulosa cell
apoptosis allowing for proper follicular maturation.
Hendler and Baum
(2004, 1939251
Human
End-tidal breath CO measurements in pregnant women with
contractions (term and pre-term) were lower than those
measurements in noncontracting women.
Hofmann and Brittain
(1998, 0520191
Human
Partitioning of 02 and CO in the human embryonic Hb is
discussed.
Iheagwaraetal.
(2007, 193861) Male
3h
1,000 ppm
CO significantly reduced cytochrome c oxidase activity and Vma>
but not Km in myocardial mitochondria. Cytochrome c oxidase
protein levels and heme content were significantly decreased.
The average COHb level was 61 % but no tissue hypoxia was
observed in the heart.
HO-1 transgenic mice
Imai et al. (2001, which specifically over-
1938641 express HO-1 in smooth
muscle
Transgenic mice had a significant increase in arterial pressure
and impaired nitrovasodilatory aortic responses. The mice had
enhanced NO' production and impaired sGC activity. The authors
speculated that the effect of HO-1 overexpression was to
suppress vasodilatory responses to NO' in vascular smooth
muscle.
CO poisoning resulted in free NO' in brains as measured by
electron paramagnetic resonance spectroscopy and in a 10-fold
increase in nitrotyrosine as measured by immunohistochemical
staining. These responses were blocked by pretreatment with a
NOS inhibitor but not by neutrophil depletion.
Brain nitrotyrosine formation was blocked by platelet depletion
60 min 1,000-3,000 ppm following 40-min but not 60-min exposure to 1,000 ppm CO.
40-60 min 1,000 ppm Following CO poisoning, myeloperoxidase activity, a measure of
leukocyte sequestration, was increased in brain microvessels.
This response was blocked by NOS inhibition but not by platelet
depletion. Similar effects were noted forxanthine oxidase
activation.
The authors concluded that perivascular reactions mediated by
peroxynitrite are key to CO poisoning effects in brain.
Rat
Ischiropoulos et al. Wistar
(1996,0794911 Male
200-290 g
Rat
Johnson and Johnson Sprague Dawley
(2003,0536111 Male
250-300 g
0-100|JM
CO produced a concentration dependent, endothelium-
dependent vasoconstriction in isolated gracilis muscle arterioles,
evident at 1 pM CO. Pre-treatment with a NOS substrate
prevented this response while pretreatment with a NOS inhibitor
converted this response to a vasodilation. The authors concluded
that exogenous CO was acting through NOS inhibition.
September 2009
E-7
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Johnson et al. (2003,
1938681
Rat
Dahl/Rapp salt-resistant
and salt-sensitive model
Male
High-salt diet increased COHb, BP, and aortic HO-1 protein
levels in salt-sensitive Dahl rats. Enhanced immunostaining was
observed for HO-1 but not HO-2 in isolated gracilis muscle
arterioles. Compared with the low-salt diet, the high-salt diet
resulted in a smaller vasoconstrictor response when NOS was
inhibited. Vasoconstriction was exacerbated in arterioles from
both low salt- and high salt-treated rats using both NOS and HO
inhibitors. Acetylcholine-induced vasodilation was diminished in
the high-salt diet group compared with the low-salt diet group.
This effect was not seen using the HO inhibitor. The high-salt diet
did not alter endothelium-independent vasodilation. The authors
concluded that HO-derived CO caused dysfunction of the NO'
system in salt-sensitive rats treated with a high-salt diet.
Johnson et al. (2004,
1938701
Rat
Sprague Dawley
Male
Deoxycorticosterone
acetate (DOCA)-salt
hypertension model
WKYrats
Spontaneously
hypertensive rats (SHR)
Salt-sensitive DOCA rats, but not SHR, had elevated aortic HO-1
expression and blood COHb levels. Both had elevated mean
arterial BP compared with controls. Acetylcholine-mediated
vasodilation of isolated gracilis muscle arterioles was attenuated
in DOCA rats but not SHR. Pretreatment with a HO inhibitor
restored the response in DOCA rats. The authors concluded that
HO-1-derived CO contributes to endothelial dysfunction in DOCA
but not SHR.
Johnson et al. (2006,
1938741
Rat
Zucker
Lean and obese
Male
lOOpMCO
The obese rats had increased CO expiration and mean arterial
pressure which was decreased by pretreatment with a HO
inhibitor. No difference was observed in HO-1 protein between
lean and obese rats. Acetylcholine- and flow-mediated vasodila-
tion of isolated gracilis muscle arterioles was attenuated in obese
but not lean rats. Pretreatment with a HO inhibitor restored the
response in obese rats. Exogenous CO prevented the restoration
of flow-induced dilation by the HO inhibitor. The authors
concluded that HO-derived CO contributes to endothelial
dysfunction in this model of metabolic syndrome.
Katoue et al. (2005,
1938961
Rat
Wistar
HO activity in the aorta is significantly increased during
pregnancy but aortic AVP-dependent Vasoconstriction appears to
be HO/CO independent.
Katoue et al. (2006,
1939541
Rat
Wistar
Pregnancy-induced modulation of calcium mobilization and
down-regulation of Rho-kinase expression contributed to
attenuated vasopressin-induced contraction of the rat aorta.
Khan et al. (2006,
1939551
Nb overexpressing BDF *
CD1 mice
Cerebral and myocardial infarcts were decreased in neuroglobin
overexpressing mice, decreasing ischemic injury.
Kim et al. (2005,
1939591
Primary rat pulmonary
artery smooth muscle
cells
Rat
Inbred LEW
Sprague Dawley
200-250 g
24hor
pretreatment for
1-2 h followed by 250 ppm
24 h
posttreatment
Exposure of cells in culture to 250 ppm CO for 24 h inhibited
serum-stimulated cell proliferation, increased expression of
p21Waf1/Cip1 and decreased expression of cyclin A. CO also
inhibited PDGF-stimulated cell proliferation and reversed the
inhibitory effect of PDGF on caveolin-1 expression. Genetic
silencing of caveolin-1 using siRNA, prevented the
antiproliferative effect of CO. Endogenous CO derived from HO-1
in an overexpression system was found to upregulate caveolin-1
expression. Effects of CO on caveolin-1 were found to be
mediated by p38 MAPK and cGMP. Experiments in fibroblasts
deficient in p38 confirmed a role for p38 in CO-mediated
inhibition of cellular proliferation via effects on p21Waf1/Cip1,
cyclin A and caveolin-1. Experiments in fibroblasts deficient in
caveolin-1 confirmed the role of caveolin-1 in the anti-proliferative
effects of CO.
In a model of neointimal injuries induced by balloon injuries in
intact animals, exposure to CO inhibited neointimal formation and
increased caveolin-1 expression in the intima and media.
September 2009
E-8
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Kim et al. (2008,
1939611
Primary rat hepatocytes
Primary mouse
hepatocytes
Respiration-deficient
human HepSB cells
10-60 min
250 ppm
Exposure of cells in culture to 250 CO for 1 h twice a day
prevented spontaneous hepatocyte death over 6 days in culture.
CO also decreased caspase-3 activity. Cell death was deter-
mined to be partly due to apoptosis. CO also increased ROS as
measured by dichlorofluorescein fluorescence in rat hepatocytes,
mouse hepatocytes and HepSB cells but not in respiration-
deficient HepSB cells indicating that ROS were mitochondrial in
origin. An increase in mitochondrial oxidized glutathione was
noted in rat hepatocytes treated with CO for 30 min. Increased
Akt phosphorylation occurred following 10-30 min CO and was
diminished by treatment with antioxidants. CO was found to
activate NFKB through a PI3K and oxidant-dependent pathway.
CO mediated spontaneous cell death was found to be dependent
on ROS and Akt phosphorylation. The authors concluded that CO
prevents hepatocyte apoptosis through redox mechanisms
leading to cytoprotection.
Sheep
Kinobe et al. (2006, Gravid and non-gravid
1884471 sheep and their near-term
fetuses
There were no significant differences in hypoxic adult and
hypoxic fetal sheep when compared to their normoxic controls.
Knuckles etal. (2008, M
1919871 Mouse
4h
Diesel emissions:
350 pg/m3
Diesel exhaust enhanced vasoconstriction in veins but not
arteries. It was suggested that this is through the uncoupling of
eNOS.
Korres et al. (2007, Ullm,n
1909081 Human
Transient evoked otoacoustic emissions response and amplitude
at 4000 Hz was lower in neonates with prenatal exposure to
cigarette smoke. There was no dose dependent change in
response depending on the amount cigarettes per day that was
smoked.
^reta.,2004,
End tidal CO concentrations were lower in pregnant women with
gestational hypertension and pre-eclampsia than normotensive
Lash et al. (2003,
1938491
Human
Term placental chroionic
villi from healthy or pre-
eclamptic placentas
Infarcted areas of placenta had decreased HO expression (in
pre-eclamptic placenta only).
Li et al. (2008,
1870031
Mouse
ICR(CD-I)
Pregnant
The effect of maternal LPS exposure on fetal liver HO was
measured. HO-1 was upregulated in fetal livers post-LPS
exposure and this HO-1 upregulation was attenuated with the
spin trap agent PBN, pointing to a ROS dependent HO-1
upregulation post maternal LPS treatment.
Liu and Fechter (1995, Guinea pig
0765241 Male
CO increased the compound action potential threshold at high
35 mL/kg (ip) frequencies. This could be blocked by inhibition of the glutamate
receptor.
Loennechen et al.
(1999, 0115491
Rat
Sprague Dawley
Female
220-240g
1wk 100 ppm
1 wk 100 ppm and
1wk 200pm 100-200 ppm
Endothelin-1 expression increased by 53% and 54% in the left
and right ventricle respectively during the 2-wk exposure and by
43% and 12% in the left and right ventricle respectively during
the 1 -wk exposure. Right ventricular to body weight ratio was
increased by 18% and 16% in the 2-wk and 1-wk exposure
groups respectively. COHb levels were 23% and 12% in the 2-wk
and 1-wk exposure groups respectively.
Longoetal. (1999,
0115481
Rat uterine tissue and tail
artery rings
Sprague Dawley
Human uterine biopsies
10-4 M
The addition of exogenous CO to isolated human and rat uterine
tissue failed to induce relaxation of uterine tissue. Isolated rat
aortic rings and tail artery rings from pregnant dams can be
relaxed by submersion in exogenous CO solutions.
September 2009
E-9
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
CO Concentration
Findings
Lopez et al. (2008,
0973431
Rat
Sprague-Dawley
Rat
Pregnant rats
exposed to CO
GD5-GD20
(Group A) or
GD5-GD20 plus
PND5-PND20 25 ppm
(Group B);
Group C (control
air exposure).
10-18h/day
CO exposure induced damage to the spiral ganglia neurons and
inner hair cells with oxidative stress seen in cochlear blood
vessels. At PND20 groups A and B showvacuolization of afferent
terminals at the base of the cochlea. At PND3, group A shows
decreased synapsin-1 staining of the efferent nerve terminals. At
PND20, groups A and B show decreased neurofilament-IR
(staining) in type I spiral ganglia neurons and afferent nerve
fibers. At PND12 and PND20, group B shows increased HO-1
and SOD-1-IR in blood vessels of the stria vasularis; group A is
similar to controls. From PND3-PND20, there is increased iNOS
and increased nitrotyrosine-IR in blood vessels of the cochlea.
Lopez et al. (2003,
1939011
Rat
Sprague-Dawley
PND6 to weaning 12 or 25 ppm
(PND19-PND20)
In the cochlea, atrophy orvacuolization of the nerve cells that
innervate the inner (not outer) hair cells was seen. Fibers of the
8th cranial nerve (internal auditory canal of the ARCO animals,
25 ppm) had distorted myelination and vacuolization of the
axoplasm. In the organ of corti and spiral ganglion neurons,
cytochrome c oxidase and NADH-TR were significantly
decreased in 25 ppm exposure group vs control. Expression of
the calcium-mediated myosin ATPase in the organ of corti and
spiral ganglion neurons was significantly decreased in the
25 ppm CO exposure group vs controls.
Lund et al. (2007,
1257411
Mouse
ApoE-/-
Male
High fat diet
6 h/day, 7days/wk,
7wk
8, 40,or60(jg/m PM
whole gasoline exhaust;
or filtered exhaust with
gases matching the 60
pg/m3 concentration. CO
concentrations were 9,
50, and 80 ppm
corresponding to the 8,
40, and 60 pg/m3 PM
whole exhaust exposures
Both whole and filtered exhaust increased aortic mRNA
expression of matrix metalloproteinase-3 (MMP-3), MMP-7, and
MMP-9, tissue inhibitor of metalloproteinases-2, endothelin-1 and
HO-1 at 60 pg/m . Aortas also showed increased immunostaining
for MMP-9 and nitrotyrosine in 60 pg/m3 PM whole exhaust and
PM-filtered exhaust exposed groups. Aortic TBARS, a measure
of lipid peroxidation, was also increased in all treatment groups.
Lund et al. (2009,
1802571
Mouse
ApoE-/-
Male
High fat diet
6 h/day, 1 or
7 days
Gasoline engine exhaust
containing 60 pg/m3 PM
and 80 ppm CO
Gasoline exhaust exposure increased aortic MMP-2/9 activity at
1 and 7 days. Protein levels of aortic MMP-9, MMP-2, TMP-2 and
plasma MMP-9 were also increased after 7 days. Lipid
peroxidation in aorta resulting from gasoline exhaust exposure
was inhibited by treatment with the antioxidant Tempol, while
increases in mRNA for ET-1 and MMP-9 in aortas were inhibited
by treatment with BQ-123, an antagonist of ETA receptor.
Treatment with BQ-123 also reduced aortic MMP-2/9 activity in
aortas following gasoline exhaust exposure. The authors
concluded that ETA receptor pathway is a key mediator of
gasoline engine exhaust effects in the vasculature.
Lyall and Myatt (2002,
1939711
Human
Women with pre-eclampsia, produced term placenta with
significant decreases in HO-2 vs women with healthy
pregnancies.
Lyall et al. (2000,
1939021
Human
(placentas from 8-19 wk
pregnancy and term
placentas)
The use of a HO inhibitor, ZnPP, increased placental perfusion
pressure. HO-1 and HO-2 were expressed in the placenta and
placental bed and vary in expression over the course of
pregnancy. HO may thus be involved in trophoblast invasion,
placental function and perfusion pressure.
Mactutus and Fechter
(1984, 0113551
Rat
Long Evans
Continuous
exposure
to CO over
gestation
150ppm
Acquisition as measured in a two-way conditioned avoidance
(flashing light warnings followed by mild footshock) test failed to
improve with age of in utero CO-exposed (150 ppm, dam COHb
15%) rats (male and female offspring) in contrast to air-exposed
controls who improved with age/maturation, indicating a failure in
the associative process of learning. The authors also found
impairments in reacquisition performance, an index of retention,
in PND31 rats that had received continuous in utero CO
exposure. Prenatal CO exposure induced learning and memory
deficits in male and female offspring.
McGregor etal. (1998,
GD23-GD25 until
term
r •
Guinea pig
10 h/day
Aberrant respiratory responses (to asphyxia and C02) of
offspring with prenatal CO exposure. The authors hypothesized
this may be related to changes in the brainstem. COHb in
maternal (8.53 ± 0.6% vsO.25 ± 0.1%) and fetal blood (13.0 ±
0.4% vs 1.6 + 0.1 %) from CO-treated vs controls.
September 2009
E-10
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Mclaughlin et al.
(2001, 1938231
Mclaughlin et al.
(2000, 0158151
Species / Model
Human placenta
Human placenta
Deration* C0 Concentration Findings
Various pathologies of pregnancy including IUGR and pre-
eclampsia are associated with significant decreases in placental
HO activity. The endogenous generation of CO in the placenta
has been demonstrated in chroionicvilli of term placenta.
Placental regional localization of HO was explored. The chorionic
plate, chorionic villi, basal plate and choorio-decidua had
significantly higher HO activity than the amnion.
Mclaughlin et al.
(2003, 1938271
Human placenta
HO expression in various regions of term placentas was
explored. Microsomal HO-2 protein content was not different
between normotensive and milk pre-eclamptic pregnancies.
There was increased expression of microsomal HO-1 protein in
chorionic villi and fetal membranes from pre-eclamptic
pregnancies vs normotensive pregnancies.
McLean et al. (2000,
0162691
Human placenta
HO activity was highest in the placenta near term.
Melin et al. (2002,
0375021
Rat
Dark Agouti
Male
Model of right ventricle
hypertrophy secondary to
chronic hypoxia (HH 10
wk)
10 wk
50 ppm alone
or concomitant with HH
Hb and hematocrit levels were increased above controls in HH
rats, CO rats and HH+CO rats, with the increase due to the
combined treatment significantly higher than the increase due to
HH. COHb levels were 1.1% in controls, 1.3% in HH rats, 4.7% in
CO rats and 9.1% in HH plus CO rats. HH treatment significantly
increased right ventricular (RV) heart weight above controls while
CO treatment had no effect on any postmortem heart weights.
Combined treatment with HH+CO resulted in a significant
increase in left ventricular plus septum (LV+S) weight and RV
weight compared with HH treatment alone. Echocardiographic
left ventricular morphology and mass also showed the greatest
changes in the HH+CO group. Hemodynamic measurements of
LV function demonstrated significant effects in the HH+CO group
for left ventricular end diastolic pressure (LVESP), left ventricular
maximal first derived pressure (+dP/dtLV), and left ventricular
work (LVW) compared with controls. Hemodynamic
measurements of RV function demonstrated significant effects in
the HH group for right ventricular end systolic and diastolic
pressure (RVESP, RVEDP), right ventricular maximal and
minimal first derived pressure (+dP/dtRV,, -dP/dtRV) and right
ventricular work (RVW). CO significantly enhanced the effects of
HH on RVEDP and significantly diminished the effects of HH on
dP/dtRV and RVW. The authors concluded that CO intensified
the HH-induce RV hypertrophy, increased LV weight and induced
severe hematological responses that could hamper adaptation.
Melin et al. (2005,
1938331
Rat
Dark Agouti
Male and female
Model of right ventricle
hypertrophy secondary to , n ,,
chronic hypoxia (HH, 10 IUWK
wk)
Half of the animals were
exercise- trained to induce
LV hypertrophy
50 ppm alone
or concomitant with HH
In untrained animals, combined treatment with HH+CO led to
increased LV+S and RV weights compared with HH treatment
alone. HH+CO led to several changes in measured
echocardiographic parameters including increased anterior and
posterior wall thickness in diastole (AVVTd, PVVTd) and to
increased fraction of shortening. These effects were not seen
with HH alone. In addition RVEDP was enhanced in HH+CO
compared with HH alone. HRV components were altered by
HH+CO but not by CO alone.
Mereu et al (2000,
1938381
Middendorffetal.
(2000, 0158421
Montagnani et al.
(1996, 0809021
R t GDO-GD20
St continuous CO 150 ppm
exposure
Human
Adult males aged 65-
75 yr.
Testicular tissue from
orchiectomy
Rat
Wistar GDO-GD20 75 or 150 ppm
Male pups
In utero exposure to CO disrupted hippocampal LTP with
concomitant HO-2 and nNOS reductions. The authors surmised
that these changes may be related to the memory deficits seen in
animals exposed to CO in utero.
Zn protoporphryin (ZnPP) and Hb both significantly reduced
seminiferous tubular cGMP generation, suggesting a role for CO
in human testicular tissue.
CO caused an increase in tetrotoxin induced inhibition of
perivascular nerve stimulation PNS-evoked vasoconstriction,
increased the time to NO-related relaxant effect by ACh, and
decreased the contractile response evoked by ACh on resting
tone.
September 2009
E-11
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
CO Concentration
Findings
Naik and Walker
(2003, 1938521
Rat
Sprague-Dawley
Male
21 OpL of CO/1 00 ml
of physiological
saline solution
Endogenous CO mediated vasorelaxation involved cGMP-
independent activation of vascular smooth muscle large-conduc-
tance Ca2*-activated K+ channels. However exogenous CO
vasodilation was cGMP dependent.
Ndisang et al. (2004,
1804251
Review of CO and hypertension. CO is a vasorelaxant due to
activation of the big conductance calcium-activated potassium
channels and soluble guanylate cyclase/cGMP pathway.
Developmental stage and tissue type will determine which of
these pathways plays more of a role in vasorelaxation.
Meggers and Singh Mouse
(2006, 1939641 CD-1
GD8-GD18
500 ppm
Developmental toxicitiy of CO was attenuated by protein
supplementation, i.e., protein supplemented animals (27%)
showed a significantly lower incidence of fetal mortality vs 8%
and 16% protein groups. Further, dietary restriction of both
protein and zinc with CO-exposure to CO during gestation
increased the incidence of pup mortality and malformations
including gastroschisis. Zinc supplementation to protein deficient
diet in CO-exposed mice decreased fetal mortality and
malformation.
in culture
Term human placental cells were grown in cell culture under
basal and hypoxic conditions to explore changes in HO
expression. HO-1 was unchanged in cytotrophoblasts under
hypoxia, but HO-1 was significantly decreased in hypoxic
syncytiotrophoblasts. HO-2 was unchanged in either cell type
with hypoxia. These cell culture data can give insight into what
cell types might be responsive to hypoxia through the HO/CO
system in the human placenta.
Odrcich et al.
1939581
Guinea pig
Immunohistochemical localization of HO in guinea pig placentae
showed that HO-1 staining was highest near term (PND62) and
lesser at term or earlier in pregnancy. HO-1 was localized in the
advential layer of fetal blood vessels.
Ozawa et al. (2002,
1938411
Rat
Wistar
Adult male
The role of HO-1 in spermatogenesis was explored. CdCI2
induced testicular HO-1 and reduced HO-2 protein in rats.
Pretreatment with ZnPPIX attenuated CdCI2-dependent
apoptosis. Leydig cells use HO-1 derived CO to tirgger apoptosis
of pre-meiotic germ cells and modulate spermatogenesis under
CdCI2 dependent oxidative stress.
Patel et al. (2003,
0431551
Rat
Sprague Dawley
Male
262 ± 30 g
Isolated hearts
SOmin
The ventricular glutathione content, both reduced and oxidized,
decreased by 76% and 84% 90 min post-exposure to 0.01 % and
Buffer saturated with 0.05% CO, respectively. Treatment with antioxidants partially
0.01 and 0.05% CO blocked the decreases in glutathione. Increased creatine kinase
activity was observed in heart perfusate during and after
treatment.
Penney etal. (1983,
0113851
Rat
(strain not reported)
In utero CO exposure induced decreased fetal body weight,
GD17-GD22 157,166 or 200 ppm decreased placental weight, increased wet heart weight at birth,
and altered cardiac enzymes at birth.
Penney etal. (1982,
0113871
Rat
COBS
GDO-GD32
350 ppm PND1-3, then
425 ppm PND4-7, then
500 ppm PND8-32
Postnatal CO exposure decreased body weight, to a greater
extent in male pups. The heart to body weight ratio and left
ventricle plus interventricular septum and right ventricle weight
increased after birth in CO exposed pups. This persistent
cardiomegaly was not explained by increasing in DMA or
hydroxyproline.
Piantadosi (2002,
0374631
Reviews the biochemical activities of CO, including various heme
protein binding. The review stresses the importance of the C0/02
ratio in determining the physiological effects of CO.
Piantadosi (2008,
1804231
Reviews the physiologic responses to exogenous and
endogenous CO and biochemical effects including the binding to
heme proteins, the generation of reactive 02 species and
activation related signaling pathways.
September 2009
E-12
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
CO Concentration
Findings
Piantadosi et al.
(2006, 1804241
Rat
Sprague Dawley
Adult male
1,3, or 7 days 50 ppm or HH
COHb produced COHb levels of 4-5% (controls approximately
1 %) and liver CO concentration of 30-40 pmol/mg wet weight
(controls approximately 10 pmol/mg wet weight). Both CO and
HH led to increased expression of hypoxia-sensitive proteins
HO-1 and HIF-1a and mitochondrial antioxidant protein SOD-2.
CO caused a greater change in mitochondrial GSH/GSSG than
HH. Only CO increased mitochondrial 3-nitrotyrosine and protein
mixed disulfides. Mitochondria isolated from CO-exposed rats,
but not from HH-exposed rats, showed an increase in the calcium
sensitivity of the mitochondrial permeability transition (MPT).
Exposure to CO or HH resulted in a loss of the ability of adenine
nucleotides to protect mitochondria from MPT. This effect was
restored in the presence of a strong reductant. The authors
conclude that CO causes mitochondrial pore stress
independently of its hypoxic effects
Prigge and Hochrainer Rat
(1977, 0123261 Wistar, SPF
GDO-GD20
60, 100, 250, 500 ppm
Fetuses were collected by C-section after 21-days exposure.
Significant increases in fetal heart weight were seen in fetuses
exposed to CO in all dose groups. Fetal body weight was signifi-
cantly decreased (NOAEL125 ppm CO).
(Raub and Benignus,
2002, 0416161
Reviews the physiology of CO and the effects on the nervous
system. It is estimated that COHb would have to rise to 15-20%
before a 10% reduction in any behavioral or visual measurement
could be observed.
Richardson et al.
(2002, 0375131
Human
Male
20% COHb
20% COHb did not influence 02Mb binding indicated by unaltered
deoxy-myoglobin signal. Resting skeletal muscle metabolic rate
was unaffected by 20% COHb. V02 max was decreased. No
decrement in intracellular P02 was found. 20% COHb altered
exercising bioenergetics, pH, PCr, and ATP levels.
Ryteretal. (2006,
1937651
Reviews the basic science of exogenous and endogenous CO
including HO-1 regulation. It also reviews some therapeutic
applications for CO.
Sartiani et al. (2004, Rat
1908981 Wistar
In utero inhalation
exposure
150 ppm
At 4 wk of age, the action potential duration APD of isolated
cardiac myocytes from CO-exposed animals failed to shorten or
mature as did the APD of control animals. Further, the two ion
conduction channels Ito (transient outward current, K*-mediated)
and ICa.L (L-type Ca2* current), which largely control the rat APD,
were significantly different from control animals after CO
exposure at 4 wk of age. All of these CO-dependent changes
were no longer different from controls at 8 wk of age, showing a
delayed maturation.
Mouse
Schwetzetal. (1979, CF'1
OJ1855) Rabbit
New Zealand
7 or 24-h/day
GD6-GD15(Mice)
GD6-GD18
(Rabbits)
In mice there was a significant increase in number of skeletal
abnormalities in CO-exposed mice. Decreased birth weight in
mice exposed to 24 h/day CO vs control. Increased birth weight
in mice exposed to 7 h/day CO vs controls. No similar effects
were seen in rabbits.
Singh etal. (1992, Mouse
0137591 CD-1
CO exposure concomitant with a low protein diet exacerbated the
percent of skeletal malformations in offspring. The percent of
dead, resorbed, or grossly malformed fetuses was directly related
GD8-GD18 65,125, or 250 ppm to CO concentration and inversely related to maternal dietary
protein levels. CO and maternal dietary protein restriction have a
synergistic effect on offspring survival and an additive effect on
malformations.
Singh (2006, 1905121 J^58
6 h/day during
the first
2ndwk
of pregnancy
65 or 125 ppm
Modulating dam protein intake during in utero CO exposure
altered pup mortality.
Singh etal. (1993, Mouse
0138921 Albino CD-1
GD8-GD18
65, 125, 250, or 500 ppm
Mice were given various protein diets (4, 8,16, or 27% protein)
during pregnancy along with CO exposure. All concentrations of
CO exposure within each maternal dietary protein level
significantly increased the percentage of litters with
malformations in a dose-dependent manner. CO exposure
concomitant with a low protein diet exacerbated the percent of
skeletal malformations in offspring. The percent of dead,
resorbed, or grossly malformed fetuses was directly related to
CO concentration and inversely related to maternal dietary
protein levels. CO and maternal dietary protein restriction had a
synergistic effect on mouse offspring mortality and an additive
effect on malformations.
September 2009
E-13
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Singh (2003, 053624)
GD8-18
500 ppm
CO decreased the mean implants per litter. CO increased the
incidence of fetal mortality. Under low protein conditions, CO
exposure increased the incidence of malformations (9.4% vs 0%)
when Zn levels were normal and increased the incidence of
gastroschisis (5% vs 0%) when Zn levels were low.
Singh and Scott
(1984, 0114091
Mouse
Albino CD-1
GD7-18
65, 125, 250, or 500 ppm
All concentration of CO decreased fetal weight mouse pups.
Near-term fetal body weight was decreased at GD18 in mice
exposed from GD7-GD18 to 125, 250, and 500 ppm CO, but not
65 ppm CO.
Singh (1986, 012827)
Impaired aerial righting score at PND14 (65 and 125 ppm),
GD7-18 65 or 125 ppm impaired negative geotaxis at PND10 and righting reflex on
PND1 (125 ppm)
Sitdikova et al. (2007, Frog neuro-muscular on . QR,,..
1804171 junctions 20 mln 96 ^M
CO induced acetylcholine release, without effects on the pre-
synaptic action potential or functional properties of post-synaptic
receptors in frog neuro-muscular preparations.
Human
Primary human airway 0-48 h
smooth muscle cells
10-250 ppm
CO inhibited SMC proliferation at concentrations from
50-500 ppm. The cell cycle arrest occurred at the GO/G1 phase
of the cell cycle. CO increased expression of the cell cycle
inhibitor p21Cip1 at 1 h and decreased expression of cyclin D1
over 24-48 h. The antiproliferative actions of CO were found to be
independent of sGC, but instead exerted through the inhibition of
ERK MAPK activation since 15 min exposure to 250 ppm CO
blocked serum-mediated ERK phosphorylation.
Sorhaug et al. (2006,
180414)
Rat
Wistar
Female
169 + 4.5 g
on h/rlav
x 5 days/wk, 200 ppm
COHb was 14.7% in CO-exposed animals and 0.3% in controls.
Total Hb was also increased in following CO exposure. CO
caused no changes in lung morphology or pulmonary
hypertenstion. No atherosclerotic lesions were found in aorta or
femoral artery. Weight increases of 20% and 14% were observed
in the right ventricle and left ventricle plus septum, respectively,
indicative of ventricular hypertrophy following chronic CO
exposure.
Mouse
C57/BI-6J
Stevens and Wang p t
(1993,188458) gfague.Daw|ey
Hippocampal brain slices
HO inhibition blocked long-term potentiation but not long-term
depression.
Stockard-Sullivan et
al. (2003, 1909471
Rat
Sprague-Dawley
22 h/day,
PND6-PND22
Using functional OAE testing and ABR showed that with perinatal
CO exposure (50 and 100 ppm CO), there were significant
decrements in OAE in CO-exposed animals. ABR showed no
12, 25, 50, or 100 ppm functional deficits with CO exposure. Using another otoacoustic
test revealed significant attenuation of the AP of the 8th cranial
nerve with CO exposure (12, 25, and 50 ppm CO) vs controls at
PND22.
Storm and Fechter Rat
(1985,0116531 Long-Evans
GDO-parturition 150 ppm
Prenatal CO exposure increased mean and total cerebellar
norepinephrine concentration from PND14-PND42, but not in the
cortex.
CO transiently decreased 5HT and NE in the pons/medulla. CO
75,150, and 300 ppm increased NE in the cortex and hippocampus at PND42. CO
GDO-GD20 ' ' dose-dependently reduced cerebellum wet weight. Maternal
COHb: 2.5%, 11.5%, 18.5%, and 26.8% (0, 75,150, and 300
ppm, respectively).
Storm and Fechter Rat
(1985,0116521 Long-Evans
Storm etal. (1986, Rat
0121361 Long-Evans
CO decreased cerebellar weight (150-300 ppm at PND10,
GDO-PND10 75, 150, and 300 ppm
cerebella had fewer fissures.
Styka and Penney
(1978, 0111661
Rat
Charles River
Male
6wk
400 ppm or gradual CO caused increased heart weight to body weight that regressed
increase from 500 to within a couple of mo after CO exposure. COHb: 400 ppm - 35%;
1,100 ppm 1,100ppm-58%
September 2009
E-14
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Suliman et al. (2007,
1937681
Mouse
C57BL/6
Wild-type and eNOS
deficient
Male
Rat
Embryonic
cardiomyocytes H9c2
cells
1h
50-1,250 ppm
OrHH
OMOOmM
dichlorom ethane
1-h exposure of mice to 1,250 ppm CO increased cardiac
mitochondrial content of all 5 respiratory complexes 24 h later.
The volume density of interfibrillar mitochondria was increased by
30% after 24 h demonstrating that CO caused cardiac
mitochondrial biogenesis. The CO concentration in heart
increased from 9 pmol/mg to 50-150 pmol/mg in mice exposed to
50-1,250 ppm CO for 1 h. These levels declined to baseline by 6
h. Peroxisome proliferator-activated receptor gamma coactivator
1 alpha (PGC-1a) expression was increased 6 h following
exposure to 50-1,250 ppm CO. Expression of DMA polymerase
and mitochondrial transcription factor A (TFAM) was increased 6
and 24 h after exposure, while mitochondrial DMA was increased
2-3 fold 24 h after exposure. CO activated gene expression of
these proteins involved in cardiac mitochondrial biogenesis
beginning at 2 h post exposure for PGC-1a, nuclear respiratory
factors 1 and 2 (NRF-1 and -2) and at 6 h postexposure for
TFAM. These effects were independent of NOS and not seen
with HH. CO exposure resulted in phosphorylation of p38 MAPK
and Akt at 2 and 6 h post-exposure to 1,250 ppm CO for 1 h.
Inhibition of p38 activation failed to inhibit the CO-mediated
increase in cardiac mitochondrial biogenesis.
In cell culture experiments, CO derived from dichloromethane
metabolism resulted in increased cGMP, protein levels of SOD2,
TFAM, NRF-1, NRF-2, PGC-1,mitochondrial ROS, Akt
phosphorylation, and mitochondrial DMA. Inhibition of GCor
PISK/Akt, but not p38, blocked the responses to CO. A role for
mitochondrial H202 in Akt regulation was demonstrated.
Mitochondrial H202 and the PISK/Akt pathway were important
mediators of TFAM expression.
The authors concluded that CO exposure increased
mitochondrial ROS which promoted mitochondrial biogenesis in
the heart.
Sun etal. (2001,
0260221
Mouse
Neuronal cultures
prepared from the
cerebral hemispheres of
16-day Charles River CD1
mouse embryos
Nb expression was increased by neuronal hypoxia in vitro and
focal cerebral ischemia in vivo. Inhibiting Nb reduced neuronal
survival after hypoxia whereas Nb overexpression enhanced
neuronal survival.
Tattolietal. (1999,
0115571
Rat
Wistar
Male and pregnant female
PND1-PND10 75 and 150 ppm
Cognitive function was assessed in rats after postnatal CO
exposure at 3 and 18 mo of age. Postnatal CO exposure did not
affect the acquisition and reacquisition of an active avoidance
task. This is different from previous findings by the same
laboratory indicating that in utero exposure to CO (75 and
150 ppm) induced long-lasting learning and memory deficits.
Telferetal. (2001,
1937691
Human
Myometrium tissue
obtained from gravid [pre-
term (25-34 wk gestation),
term not in labor or term in
labor] and non-gravid
cGMP was monitored in various myometrial tissues. cGMP was
significantly higher than that from nonpregnant tissue and
decreased at term, especially in tissue from laboring women.
Teran et al. (2005,
1937701
Rat
Dahl/Rapp
salt-sensitive rats
Male
A high salt diet for 1-4 wk resulted in increased aortic HO-1
protein expression, an increase in mean arterial pressure and
time-dependent inhibition of flow- and acetylcholine-mediated
vasodilation in isolated gracilis muscle arterioles. A smaller
degree of inhibition of acetylcholine-mediated vasodilation was
observed with a low salt diet for 1-4 wk. Pretreatment with a HO
inhibitor restored these responses but this effect was reversed in
the presence of exogenous CO. Mean arterial pressure was
decreased in intact animals fed a high salt diet for 4 wk and then
treated with a HO inhibitor. The authors concluded that the HO-
derived CO contributed to the development of hypertension and
the impairment of endothelium-dependent vasodilator responses
in this model.
September 2009
E-15
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Thorn etal. (1994,
0764591
Rat
Wistar
Male
Isolated blood cells
1 h
OR
>1 h
30 min
1,000ppm
OR
1,000-3,000 and
higher ppm
0.5 ml of pure CO
CO poisoning inhibited B2 integrin-dependent PMN adherence in
heparinized blood obtained from rats immediately after exposure.
Adherence was restored when platelet number was decreased.
Adherence was also decreased when PMN from control animals
were incubated with platelets from poisoned animals. Adherence
of activated PMN was reduced in the presence of SOD and
enhanced by NOS inhibition. Platelet production of NO' was
significantly greater, while platelet NOS activity was significantly
inhibited after poisoning.
When whole blood or platelet-rich plasma was incubated with
CO, PMN adherence was inhibited.
The authors concluded that PMN B2 integrin activity was
inhibited by CO-dependent release of NO' from the platelets into
the blood.
Platelets isolated from rats exposed to 20-1,000 ppm CO for 1-h
released NO' in a dose-dependent manner. COHb levels were
0.7% in controls, and 3.2%, 7.8% and 51.0% in 20,100 and
1,000 ppm exposure groups respectively.
Isolated platelets released NO-when incubated for 30 min with
20-100 ppm CO. NOS activity was not enhanced by 100 ppm
CO. Platelets released NO' in response to 10-100 ppm CO after
30 min pretreatment with a NOS inhibitor, suggesting that CO
displaces NO- from heme-binding sites. Longer incubations (2 h)
with the NOS inhibitor led to a diminished response to 100 ppm
CO. There appears to be a discrepancy in the results depending
on how NO' was measured (electrode vs Greiss reaction).
Endothelial cells released NO' in response to 20-100 ppm CO.
NOS inhibition blocked the response to 100 ppm CO. CO was
found not to affect arginine transport or NOS activity in
endothelial cells. Exposure to 40-100 ppm CO resulted in the
1 n 20-1 000 ppm release of short-lived oxidants. This response was blocked by
NOS inhibition. Lysates from cells exposed to 50 and 100 ppm
30 min or 2 h 10-20 ppm CO had increased nitrotyrosine content. This response was
blocked by NOS inhibition. Cellular reduced sulfhydryls were not
1 h 10-100 ppm decreased by 100 ppm CO. Dihydrorhodamine 123 oxidation, a
measure of peroxynitrite formation, was increased by exposure to
100 ppm CO. This effect was blocked by NOS inhibition.
Cytotoxicity of CO was evaluated by the release of 51 chromium.
Cytotoxicity was evident 4 h following a 2-h incubation with
100 ppm CO, but not immediately after exposure. This response
was not blocked by NOS inhibition, although NOS inhibition had
protective effects under conditions of continuous CO exposure of
4 h. Exposure to 20 and 100 ppm CO for 2 h led to the loss of
membrane integrity, measured byethidium homodimer-1
staining, 18 h later.
Results demonstrate that 10-20 ppm CO released NO- from
platelets and endothelial cells in vitro. Platelets from rats that in-
haled 20 ppm CO also released NO' in vitro. The authors
suggested that CO-mediated NO' release from platelets and
endothelial cells resulted from disrupted intracellular scavenging
for NO'. They also suggested that peroxynitrite may have been
generated in response to CO.
Thorn and
lschiropoulos(1997,
0856441
Ra
(Wistar
Male
200-290 g
Platelet-rich plasma from
rats was used as the
source of platelets
Bovine pulmonary artery
endothelial cells
September 2009
E-16
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
CO Concentration
Findings
1-h exposure to 111-110 nM CO led to a dose-dependent
increase in NO' release, as measured by nitrite+nitrate. Signifi-
cance was achieved at 22 nM (corresponding to an interstitial
partial pressure of 20 ppm and a blood COHb level of 7%). NOS
inhibition blocked the response to 110 nM CO. A dose-
dependent increase in cellular nitrotyrosine was also observed
following a 2-h exposure to CO, with significance achieved at 55
nM CO. NOS inhibition blacked the response to 110 nM. CO
exposure failed to decrease the concentration of reduced
suifhydryls, but did result in the extracellular release of a short-
lived oxidant species which was blocked by NOS inhibition.
Dihydrorhodamine oxidation, a measure of peroxynitrite
formation, occurred in response to 110 nM CO, an effect which
blocked by NOS inhibition. Cytotoxicity of CO was evaluated by
the release of 51 chromium. Cytotoxicity was evident 4 h following
30 min-4 h a 2-h incubation with 110 nM CO, but not immediately after
10-100 ppm (11-110 nM) exposure. This response was not blocked by NOS inhibition, al-
though NOS inhibition had protective effects under conditions of
continuous CO exposure of 4 h. Exposure to 110 nM CO for 2 h
led to the loss of membrane integrity, measured by ethidium
homodimer-1 staining, 18 h later. This response was blocked by
NOS inhibition. Exposure to 110 nM CO had no effect on 02 con-
sumption, production of intracellular H202 or cellular redox
activity. Exposure to 110 nM did not alter arginine transport or
NOS activity. NO- release from cells which had been pre-treated
with a NOS inhibitor and then exposed briefly to 5% CO was
measured using a NO-selective electrode suggesting that CO
competed with intracellular binding sites of NO-.
The authors concluded that endothelial cells release NO' and
NO'-derived oxidants in response to CO. A delayed cell death
occurred following exposures to 22 nM and higher concentrations
of CO.
Thorn etal. (1997,
0843371
Bovine pulmonary artery
endothelial cells
Thorn etal. (1999,
0167531
Rat
Wistar
Male
200-290 g
Some rats fed a high
cholesterol diet
1h
Nitrotyrosine immunoreactivity was found in aortic intima in rats
exposed to CO for 1 h but not in controls. Nitrotyrosine content
was quantitated and found to be increased in a dose-dependent
manner following 1-h exposure to 50-1,000 ppm CO. The effect
was significant at 50 ppm but the COHb content measured
immediately after exposure was not different than controls. Plate-
let and neutrophil depletion did not alter nitrotyrosine content
following CO exposure. Leukocyte adherence to the aorta
occurred 18 h, but not immediately, after a 1-h exposure to
100 ppm CO. This effect was blocked by NOS inhibition. The
influx of albumin from the microvasculature into skeletal muscle
50-1,000 ppm increased during the 3 h after exposure to 100 ppm CO but was
not seen 18 h later. This effect was blocked by NOS inhibition.
Rats fed a high cholesterol diet and exposed to 100 ppm CO for
1 h had increased aortic nitrotyrosine content which was not
different than that in CO-exposed rats fed the standard diet.
However, rats on the high cholesterol diet had a 6-fold increase
in LDL oxidation immediately after 1-h exposure to 100 ppm CO.
This effect was not blocked by NOS inhibition.
The authors concluded that CO can alter vascular status by
several mechanisms linked to NO'-derived oxidants.
September 2009
E-17
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Thorn etal. (1999,
0167571
Rat
Wistar
Male
200-290 g
1 h
Leakage of albumin into lung parenchyma occurred 18 h after
rats were exposed to 100 ppm CO for 1 h. This response was
not observed at earlier time points following CO exposure. It
response was also observed using 50 and 1,000 ppm but not
20 ppm CO. Leakage resolved by 48 h. Furthermore no leakage
occurred when rats which were exposed to 100 ppm CO were
pretreated with a NOS inhibitor. COHb levels were 0.9% in
controls and 4.8%, 10.6% and 53.7% following 1-h exposure to
50,100 and 1,000 ppm CO, respectively. Elevated free NO-,
determined by EPR, was observed in lungs of rats exposed to
50-1,000 ppm 100 ppm CO for 1 h. This effect was blocked when rats were
pretreated with a NOS inhibitor. Lung H202 was elevated by
exposure to 100 ppm CO for 1 h and this effect was blocked
when rats were pretreated with a NOS inhibitor. Elevated
nitrotyrosine content was observed in lung homogenates 2-4 h
following 1-h exposure of rats to 100 ppm CO. This effect was
also blocked by pretreatment with a NOS inhibitor. No leukocyte
sequestration was observed in Iungs18 h following exposure to
100 ppm CO. CO-induced lung leak was not affected by
neutrophil depletion. The authors concluded that CO causes lung
vascular injury which is dependent on NO'.
Thorn et al. (2000,
0115741
Bovine pulmonary artery
endothelial cells
40 min-2 h
11-110nM
(10-100 ppm)
Increased uptake of ethidium homodimer-1, a measure of
decreased membrane integrity and cell death, was observed in
endothelial cells 18 h after exposure to 110 nM for 60-120 min.
Exposures of 20-40 nM were ineffective in this regard. Ethidium
uptake was also increased by 2-h exposure to 88 nM CO.
Preincubation for 2 h with an inhibitor of eNOS, an antioxidant,
and an inhibitor of peroxynitrite reactions blocked the
CO-mediated cell death. Morphological changes in cells were ob-
served 2 h following a 2-h exposure to 110 nM CO. Cell death
induced by 110 nM CO was also blocked by inhibition of protein
synthesis and inhibition of caspase-1 but of caspase-3.
Caspase-1 activity was increased following 2-h exposure to
110 nM CO; this effect was blocked by inhibiting eNOS. Pre-
exposureofcellsto 11 nM CO for 40 min followed bya3-h
incubation period resulted in an increased level of MnSOD and
protection against cell death 18 h following a 2-h exposure to
110 nM CO. The authors concluded that exposure to 11 nM CO
led to an adaptive response which protected cells from injury and
apoptosis resulting from N0--derived oxidants.
Thorn etal. (2001,
1937791
Rat
Until lost
consciousness
Neutrophils sequestration was observed in the brain vessels of
rats exposed to high dose CO. CO also led to increased
1,000-3,000 ppm nitrotyrosine formation in the brain vessels. These events were
blocked by pretreatment with a peroxynitrite scavenger or a PAF
receptor antagonist.
Thorn et al. (2006,
0984181
Human
Rat
Wistar
Male
Mouse
C57B6J
MPO-deficient
Blood samples and brain
tissue
1 h
Humans:
Acute CO poisoning
Rats and mice:
1,000-3,000 ppm
In humans, COHb was 20-30.5%. Increased cell surface
expression of CD18 and PAC1 was observed in neutrophils from
people with CO poisoning. Increased surface-bound
myeloperoxidase (MPO, indicative of neutrophil degranulation),
increased plasma MPO and more numerous platelet-neutrophil
aggregates were also observed.
Similar changes were observed in blood of CO-poisoned rats.
Platelet depletion, inhibition of NOS and inhibition of platelet
integrin-dependent adhesion blocked these responses. Brains
from poisoned rats had significant elevations in MPO which could
reflect either an increase number of neutrophils or an increase in
neutrophil degranulation. Perivascular MPO and nitrotyrosine
were CO-localized in brain. CO poisoning also resulted in altered
brain myelin basic protein.
Similar changes were observed in blood of CO-poisoned mice.
MPO deficiency blocked the CO-mediated alteration in brain
myelin basic protein.
The authors concluded that exposure to CO triggers intravascular
interactions between platelets and neutrophils that lead to
neutrophil degranulation in experimental animals and people with
CO poisoning.
September 2009
E-18
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Thorupetal. (1999,
1937821
Rat
Sprague Dawley
Male
200-250 g
0.01-10 pM
Perfusion of Isolated rat renal resistance arteries with
CO-containing buffer (0.001-10 pM) led to the biphasic release of
NO', peaking at 100 nM and declining to undetectable responses
at 10 |jM. Sequential pulses of 100 nM resulted in a blunting of
NO' release with consecutive pulses, consistent with a depletion
of intracellular NO' stores. NO release was dependent on
arginine concentrations and was inhibited by pretreatment with a
NOS inhibitor. Perfusion with 100 nM CO blocked carbachol-
dependent NO- release from vessels.
Rats were treated with a HO-1 inducer and renal resistance
arteries were isolated 12 h later. Carbachol-induced NO- release
was smaller in the HO-1 induced rats compared with controls
suggesting that endogenous CO has a similar effect as 100 nM
exogenous CO. This effect was reversed in the presence of
excess arginine.
Vasodilation was measured in blood-perfused afferent arterioles
perfused with CO in solution. A biphasic vasodilatory response
was observed as well as a blunted muscarinic vasorelaxation.
CO (0.1-10 pM) suppressed the release of NO- from purified
recombinant eNOS in solution.
The authors concluded that low levels of CO may release NO'
and elicit vasorelaxation and modulate basal vascular tone while
higher levels of CO may inhibit eNOS and NO' generation.
Tolcos et al. (2000,
0159971
Guinea pig
10h/day over the
last 60% of 200 ppm
gestation
Fetal and maternal COHb were 13% and 8.5% respectively.
Neurotransmitter systems were affected after CO exposure. The
catecholaminergic system of the brainstem displayed significant
decreases in immunoreactivityfortyrosine hydroxylase (TH),
which was likely due to decreased cell number in specific
medullar regions. The cholinergic system was also affected by
prenatal CO exposure with significant increases in ChAT
immunoreactivity of the medulla and no changes in muscarinic
acetylcholine receptor.
l (2000
Guinea pig
10h/dayforthe
last 60% of 200 ppm
gestation
Brains were collected at 1 and 8 wk of age. These data showed
that CO exposure in utero sensitized the brain to hyperthermia at
PND4 leading to generation of necrotic lesions in the brain and
changes in neurotransmitter levels.
Toyodaetal. (1996,
0799451
Tschugguel et al.
(2001, 1937851
Human
HUVEC
CO was generated by primary endothelial cells from human
umbilical veins and uterine arteries after exogenous 17-B estra-
diol administration.
Villamoretal. (2000,
0158381
Mouseprotein
The authors present the X-ray structure of CO-bound ferrous
murine Nb. When CO binds, the heme group slides deeper into
the protein crevice.
Vreman et al. (2000,
0969151
Human
Umbilical cord (artery and
Rat
Aorta, vena cavae, liver
and heart
HO activity was quantified in human umbilical cord and in the rat
vasculature (aorta and vena cavae). Human umbilical artery and
vein HO activity were equal. The rat aorta and vena cavae
produced equal amounts of HO activity (wet weight/g tissue) but
generated 3x greater HO than the heart and 0.2x of the liver. HO
activity in rat vasculature was 3x that of the human cord tissues.
Use of the HO inhibitor CrMP effectively blocked HO activity in
the rat liver and heart but was less effective at blocking HO
activity in the human umbilical cord or the rat vasculature (only
50% effective). The activity of HO in the umbilical vessels may
provide a role for CO in control of vasculature tone during
pregnancy.
September 2009
E-19
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Vreman et al. (2005,
1937861
Mouse
BALB/c
500 ppm
OR
Following CO exposure, COHb levels were 28%. Tissue
concentrations of CO were as follows with control levels in
parenthesis.
Blood: 2648 ±400 (45) pmol/mg
Heart:100±18(6)pmol/mg
Muscle: 14+1 (10) pmol/mg
Brain: 18 + 4(2) pmol/mg
Kidney: 120+12 7) pmol/mg
Spleen: 229 + 55 6) pmol/mg
Liver: 115 + 31 (5) pmol/mg
Lung: 250 + 2 (3) pmol/mg
Intestine: 9 + 7 (4) pmol/mg
Testes: 6 + 3 (2) pmol/mg
CO concentration relative to 100% blood:
Lung: 9.4%,
Spleen: 8.6%
Kidney: 4.5%,
Liver: 4.3%,
Heart: 3.8%,
Brain: 0.7%,
Muscle: 0.5%,
Intestine: 0.3%,
Testes: 0.2%
SOmin
Heme arainate Injection of heme arginate resulted in a 3-fold increase in CO
arginate excretion reaching a maximum at 60 min. Animals were sacri-
30 umol/kg body weight f'080'a' 90 min. COHb levels were 0.9%. Tissue concentrations
j v of CO were as follows with control levels in parenthesis.
Blood: 88+10 (45) pmol/mg
Heart: 14 + 3(6) pmol/mg
Muscle: 7 + 1 (10) pmol/mg
Brain: 2 + 0 (2) pmol/mg
Kidney: 7 + 2 (7) pmol/mg
Spleen: 11 + 1 (6) pmol/mg
Liver: 8 + 3 (5) pmol/mg
Lung: 8 +3 (3) pmol/mg
Intestine: 3 + 1 (4) pmol/mg
Testes: 2 + 0 (2) pmol/mg
CO concentration relative to 100% blood:
Heart: 16%
Spleen: 13%
Lung: 9%
Liver: 9%
Kidney: 8%
Muscle: 8%
Intestine: 3%
Brain: 2%
Testes: 2%
Weaver et al. (2007,
1939391
Human
Acute CO poisoning
Mean COHb in humans with acute CO poisoning was 35%.
Hyperbaric 02 reduces cognitive sequelae in a randomized
clinical trial of CO-poisoned patients. Risk factors for cognitive
sequelae without hyperbaric 02 included older age and longer
CO exposures. Patients with loss of consciousness or high initial
COHb levels should also be treated with hyperbaric 02.
Webber etal. (2003,
1905151
Rat
(Strain not stated)
PND8-PND22 12.5, 25, or 50 ppm
Immunostaining of c-Fos, a marker of neuronal activation in the
nervous system was followed. C-Fos immunoreactivity in the
central 1C was significantly decreased in the CO-exposed
animals at both PND27 and PND75-PND77 over all dose groups
of CO; immnunostaining of other subregions of the 1C were not
affected by CO. These studies show exposure to CO during
development can lead to permanent changes in the auditory
system of rats that persist into adulthood.
September 2009
E-20
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Webber etal. (2005,
1905141
Rat
(Strain not stated)
PND9-PND24 25or100ppm
Neurofilament loss from the spiral ganglion neurons and somas
after ARCO treatment was rescued (no detectable neurofilament
loss) with low iron+CO (ARIDCO); ARID (low iron) treatment
induced no change in neurofilaments. CuZn superoxide
dismutase (SOD1) was significantly increased with CO exposure
(ARCO) and rescued in ARIDCO animals; SOD1 was unchanged
in low iron only animals (ARID). Low iron treatment or CO
exposure alone led to significant decreases in c-fos positive cell
numbers of the central 1C, but c-fos levels were unchanged after
low iron diet concomitant with CO exposure (ARIDCO).
Wellenius et al. (2004,
0878741
Rat
Sprague Dawley
250 g
Diazepam-sedated
Model of acute Ml induced
by thermocoagulation
1 h, 12-18 h after
surgery
35 ppm
CO exposure decreased ventricular premature beat frequency by
60.4% during the exposure period compared to controls. 1-h
exposure to CAPs (318 jjg/m3) decreased ventricular premature
beat frequency in specific subgroups. Neither CAPs nor CO had
an effect on heart rate. There were no significant interactions
between their effects when rats were exposed to both CO and
CAPs.
Wellenius et al. (2006,
1561521
Rat
Sprague Dawley
250 g
Diazepam-sedated
Model of acute Ml induced
by thermocoagulation
1 h, 12-18 h after
surgery
35 ppm
Exposure to CO failed to increase the probability of observing
supraventricular ectopic beats (SVEB). Exposure to CAPs
(646 pg/m3) for 1 h decreased the frequency of SVEB. There
were no significant effects observed when rats were exposed to
both CO and CAPs. Among a subset of rats with one or more
SVEB at baseline, a significant decrease in number of SVEB
during the exposure period was observed with either CO or CAPs
exposure compared with controls.
Yoshiki etal.(2001,
1937901
Human
HO localization in human endometrium and its changes in
expression over the menstrual cycle were explored in this study.
HO-1 was constitutively expressed throughout the menstrual
cycle and HO-2 was greater in the secretory than the proliferative
phase of the menstrual cycle. HO-1 was localized to the epithelial
cells and macrophages. HO-2 was found in endothelial cells and
smooth muscle cells of endometrial blood vessels.
Yu et al. (2008,
1923841
Guinea pig
Allergic rhinitis model
using nasal ovalbumin
sensitization
Indicators of allergic rhinitis were enhanced by treatment with a
HO-1 inducer and decreased by treatment with a HO-1 inhibitor.
Immunoreactivity for HO-1 was shown in the lamina of mucosa of
sensitized guinea pigs. Endogenous CO may play a role in the
inflammation process of allergic rhinitis.
Zamudioetal. (1995, u
1939081 Human
Women living at high altitude had an increased risk of adverse
pregnancy outcomes vs women living at lower altitudes.
Zenclussen et al.
(2006, 1938731
Mouse
CBA/J x DBA/2J
To evaluate the role of HO-1 in spontaneous abortion, a mouse
model that spontaneously undergoes abortion (CBA/J x DBA/2J
mice) was used with and without HO adenovirus treatment to see
if pregnancy outcome could be modulated by changing HO
concentration. Pregnancy outcome was significantly better
(abortion rate significantly decreased) in mice overexpressing HO
due to adenovirus transfer.
Zhang et al. (2005,
1844601
Rat
Pulmonary artery
endothelial cells
8-28 h
15 ppm
Exposure to 15 ppm CO during anoxia resulted in decreased
phosphorylation of STAT1 and increased phosphorylation of
STATS at 8-24 h. Similar responses were observed when 24-h
anoxia was followed by a period of reoxygenation (0.5-4 h). DMA
binding of STAT1 was decreased while that of STATS was
enhanced by CO treatment during anoxia/reoxygenation.
Exposure to 15 ppm during 8-24-h anoxia or 24 h anoxia followed
by 0.5-4 h reoxygenation resulted in increased phosphorylation of
Akt and p38 MAPK. Inhibitor studies demonstrated that activation
of the PI3K pathway by CO was upstream of p38 MAPK
activation during anoxia/reoxygenation. Similarly, the PISKand
p38 MAPK pathways were found to be upstream of STAT
modulation. The anti-apoptotic effects of 15 ppm CO during
anoxia-reoxygenation involved decreased FAS expression and
decreased caspase 3 acvitiry These effects were dependent on
activation of the PI3K, p38 MAPK and STATS pathways.
The authors concluded that CO blocks anoxia-reoxygenation
mediated apoptosis through modulation of PI3K/Akt/p38 MAPK
and STAT1 and STATS.
September 2009
E-21
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Zhang et al. (2007,
1938791
Mouse
A single dose of IPS administered to pregnant mice induced up-
regulation of HO-1 but not HO-2 in the mouse placenta 12-48 h
post-LPS treatment. Pre-treatment of mice with the spin trap
agent PBN or the TNF inhibitor pentoxifylline prevented the
LPS-dependent HO-1 upregulation. Thus ROS may mediate the
LPS-dependent upregulation of HO-1.
Zhao et al. (2008,
1938831
Mouse
FVB
With pregnancy, there was an increased blood volume without a
concurrent increase in systemic BP; this was accomplished by a
decrease in total vascular resistance, to which CO contributed as
determined by using HO inhibitors.
Zhuoetal. (1993,
0139051
Guinea pig
Adult male
Hippocampal LTP of brain sections is significantly affected by CO
exposure with ZnPP IX, a HO inhibitor, blocking hippocampal
LTP.
Exposure of RAW macrophages to 250 ppm CO for 10-60 min
increased ROS generation, measured as dichlorofluorescein
(DCF) fluorescence. ROS generation at 1 h was dose-dependent
with significant effects observed at 50, 250 and 500 ppm CO.
This response was not blocked with a NOS inhibitor. A 1-h
exposure to 250 ppm resulted in decreased intracellular
glutathione levels. CO treatment was found to block TNFa
production and to enhance p38 MAPK phosphorylation in LPS-
stimulated cells. These effects were diminished by pretreatment
with antioxidants. The source of CO-derived oxidants was
determined to be mitochondrial since respiration-deficient THP-1
10 min-24 h 50-500 ppm macrophages, unlike wild-type cells, failed to generate ROS in
response to 250 ppm CO. Furthermore, treatment of RAW cells
with the mitochondrial complex III inhibitor antimycinC, blocked
ROS generation in response to 250 ppm CO. Exposure of RAW
cells to 250 ppm CO for 1 h inhibited cytochrome c oxidase
activity by 50%. Exposure to 250 ppm CO for 6 h had no effect
on cellular ATP levels or mitochondrial membrane potential.
Antimycin C treatment was found to reverse the effects of CO on
LPS-mediated responses (TNFa and p38 MAPK), suggesting
that mitochondrial-derived ROS mediated the effects of CO. The
authors concluded that CO increased the generation of
mitochondrial-derived ROS.
Zuckerbraun et al.
(2007, 1938841
Macrophages
RAW 264.7
THP-1 cells, wild-type and
respiration-deficient
September 2009
E-22
DRAFT - DO NOT CITE OR QUOTE
-------
References
Acevedo CH; Ahmed A. (1998). Hemeoxygenase-1 inhibits human myometrial contractility via carbon monoxide and is
upregulated by progesterone during pregnancy. J Clin Invest, 101: 949-955. 016003
Achouh PE; Simonet S; Fabiani JN; Verbeuren TJ. (2008). Carbon monoxide induces relaxation of human internal thoracic
and radial arterial grafts. Interact Cardiovasc Thorac Surg, 7: 959-963. 179918
Ahmed A; Rahman M; Zhang X; Acevedo CH; Nijjar S; Rushton I; Bussolati B; St John J. (2000). Induction of placental
heme oxygenase-1 is protective against TNFalpha-induced cytotoxicity and promotes vessel relaxation. , 6: 391-
409. 193863
Ahmed H; McLaughlin BE; Soong J; Marks GS; Brien JF; Nakatsu K. (2005). The source of endogenous carbon monoxide
formation in human placental chorionic villi. Cell Mol Biol Incl Cyto Enzymol, 51: 447-451. 193865
Alexander PG; Chau L; Tuan RS. (2007). Role of nitric oxide in chick embryonic organogenesis and dysmorphogenesis.
Birth Defects Res AClin Mol Teratol, 79: 581-594. 193869
Alexandreanu 1C; Lawson DM. (2002). Effects of chronic administration of a heme oxygenase substrate or inhibitor on
progression of the estrous cycle, pregnancy and lactation of Sprague-Dawley rats. Life Sci, 72: 153-162. 192373
Alexandreanu 1C; Lawson DM. (2003). Heme oxygenase in the rat anterior pituitary: Immunohistochemical localization
and possible role in gonadotropin and prolactin secretion. Exp Biol Med (Maywood), 228: 64-69. 193871
Alexandreanu 1C; Lawson DM. (2003). Heme oxygenase in the rat ovary: Immunohistochemical localization and possible
role in steroidogenesis. Exp Biol Med (Maywood), 228: 59-63. 193876
Alonso JR; CardellachF; Lopez S; Casademont J; Miro O. (2003). Carbon monoxide specifically inhibits cytochrome c
oxidase of human mitochondrial respiratory chain. , 93: 142-146. 193882
Andresen JJ; Shafi NI; Durante W; Bryan RM Jr. (2006). Effects of carbon monoxide and heme oxygenase inhibitors in
cerebral vessels of rats and mice. Am J Physiol Heart Circ Physiol, 291: H223-H230. 180449
Antonelli T; Tomasini MC; Tattoli M; Cassano T; Finetti S; Mazzoni E; Trabace L; Carratu MR; Cuomo V; Tanganelli S;
Ferraro L. (2006). Prenatal exposure to the cannabinoid receptor agonist WIN 55,212-2 and carbon monoxide
reduces extracellular glutamate levels in primary rat cerebral cortex cell cultures. Neurochem Int, 49: 568-576.
193885
Appleton SD; Marks GS; Nakatsu K; Brien JF; Smith GN; Graham CH. (2002). Heme oxygenase activity in placenta:
Direct dependence on oxygen availability. Am J Physiol Heart Circ Physiol, 282: 2055-2059. 193935
AshfaqM; JanjuaMZ; Nawaz M. (2003). Effects of maternal smoking on placental morphology . JAMC, 15: 1-5. 194002
Astrup P; Olsen HM; Trolle D; Kjeldsen K. (1972). Effect of moderate carbon-monoxide exposure on fetal development.
Lancet, 7789: 1220-1222. 011121
Bainbridge SA; Belkacemi L; Dickinson M; Graham CH; Smith GN. (2006). Carbon monoxide inhibits
hypoxia/reoxygenation-induced apoptosis and secondary necrosis in syncytiotrophoblast. Am J Pathol, 169: 774-
783. 193949
Bainbridge SA; Farley AE; McLaughlin BE; Graham CH; Marks GS; Nakatsu K; Brien JF; Smith GN. (2002). Carbon
monoxide decreases perfusion pressure in isolated human placenta. Placenta, 23: 563-569. 043161
Bainbridge SA; Smith GN. (2005). HO in pregnancy. , 38: 979-988. 193946
Bamberger A-M; KoglinM; Kempfert J; Loning T; Scholz H; Behrends S. (2001). Expression and tissue localization of
soluble guanylyl cyclase in the human placenta using novel antibodies directed against the alpha2 subunit. J Clin
Endocrinol Metab, 86: 909-912. 016271
Barber A; Robson SC; Lyall F. (1999). Hemoxygenase and nitric oxide synthase do not maintain human uterine quiescence
during pregnancy. Am J Pathol, 155: 831-840. 193953
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk Information System (IRIS).
September 2009 E-23 DRAFT - DO NOT CITE OR QUOTE
-------
Barber A; Robson SC; Myatt L; Bulmer JN; Lyall F. (2001). Heme oxygenase expression in human placenta and placental
bed: Reduced expression of placenta endothelial HO-2 in preeclampsia and fetal growth restriction. , 15: 1158-
1168. 193956
Baum M; Schiff E; Kreiser D; Dennery PA; Stevenson DK; Rosenthal T; Seidman DS. (2000). End-tidal carbon monoxide
measurements in women with pregnancy-induced hypertension and preeclampsia. Am J Obstet Gynecol, 183: 900-
903. 016435
Benagiano V; Lorusso L; Coluccia A; Tarullo A; Place P; Girolamo F; Bosco L; Cagiano R; Ambrosi G. (2005). Glutamic
acid decarboxylase and GABA immunoreactivities in the cerebellar cortex of adult rat after prenatal exposure to a
low concentration of carbon monoxide. Neuroscience, 135: 897-905. 180445
Benagiano V; Lorusso L; Place P; Girolamo F; Rizzi A; Sabatini R; Auteri P; Bosco L; Cagiano R; Ambrosi G. (2007).
Effects of prenatal exposure to the CB-1 receptor agonist WIN 55212-2 or CO on the GABAergic neuronal systems
of rat cerebellar cortex. Neuroscience, 149: 592-601. 193892
Bergeron M; Ferriero DM; Sharp FR. (1998). Developmental expression of heme oxygenase-1 (HSP32) in rat brain: An
immunocytochemical study. , 105: 181-194. 193967
Bing O; Grundemar L; Ny L; Moller C; Heilig M. (1995). Modulation of carbon monoxide production and enhanced
spatial learning by tin protoporphyrin. Neuroreport, 6: 1369-1372. 079418
Burmester T; Weich B; Reinhardt S; Hankeln T. (2000). A vertebrate globin expressed in the brain. , 407: 520-523. 099998
Bye A; S0rhaug S; Ceci M; H0ydal MA; St01en T; Heinrich G; Tj0nna AE; Najjar SM; Nilsen OG; Catalucci D; Grimaldi
S; Contu R; Steinshamn S; Condorelli G; Smith GL; Ellingsen O; Waldum H; Wiskrff U. (2008). Carbon monoxide
levels experienced by heavy smokers impair aerobic capacity and cardiac contractility and induce pathological
hypertrophy. Inhal Toxicol, 20: 635-646. 193777
Cagiano R; Ancona D; Cassano T; Tattoli M; Trabace L; Cuomo V. (1998). Effects of prenatal exposure to low
concentrations of carbon monoxide on sexual behaviour and mesolimbic dopaminergic function in rat offspring. Br
JPharmacol, 125: 909-915. 087170
Carmines E; RajendranN. (2008). Evidence for carbon monoxide as the major factor contributing to lower fetal weights in
rats exposed to cigarette smoke. , 102: 383. 188440
Carratu MR; Cagiano R; De Salvia MA; Trabace L; Cuomo V. (1995). Developmental neurotoxicity of carbon monoxide.
Presented at In: Degen, G. H.; Seiler, J. P.; Bentley, P., eds. Toxicology in transition: proceedings of the 1994
EUROTOX congress; August 1994; Basel, Switzerland. Arch. Toxicol. Suppl. 17: 295-301. 079427
Carratu MR; Cagiano R; Desantis S; Labate M; Tattoli M; Trabace L; Cuomo V. (2000). Prenatal exposure to low levels of
carbon monoxide alters sciatic nerve myelination in rat offspring. Life Sci, 67: 1759-1772. 015839
Carratu MR; Cagiano R; Tattoli M; Trabace L; Borracci P; Cuomo V. (2000). Prenatal exposure model simulating CO
inhalation in human cigarette smokers: sphingomyelin alterations in the rat sciatic nerve. Toxicol Lett, 117: 101-
106.015935
Carratu MR; Renna G; Giustino A; De Salvia MA; Cuomo V. (1993). Changes in peripheral nervous system activity
produced in rats by prenatal exposure to carbon monoxide. Arch Toxicol, 67: 297-301. 013812
Carraway MS; Ghio AJ; Suliman HB; Carter JD; Whorton AR; Piantadosi CA. (2002). Carbon monoxide promotes
hypoxic pulmonary vascular remodeling. Am J Physiol, 282: L693-L702. 026018
Cella M; Farina MG; Sarmiento MIK; Chianelli M; Rosenstein RE; Franchi AM. (2006). Heme oxygenase-carbon
monoxide (HO-CO) system in rat uterus : Effect of sexual steroids and prostaglandins. J Steroid Biochem Mol Biol,
99: 59-66 . 193240
Chen GH; Kong J; Reinhard K; Fechter LD. (2001). NMDA receptor blockage protects against permanent noise-induced
hearing loss but not its potentiation by carbon monoxide . Hear Res, 154: 108-115. 193985
Cheng C; Noordeloos AM; Jeney V; Scares MP; Moll F; Pasterkamp G; Serruys PW; Duckers HJ. (2009). Heme
oxygenase 1 determines atherosclerotic lesion progression into a vulnerable plaque. Circulation, 119: 3017-3027.
193775
Chung Y; Huang SJ; Glabe A; Jue T. (2006). Implication of CO inactivation on myoglobin function. Am J Physiol Lung
Cell Mol Physiol, 290: 1616-1624. 193987
September 2009 E-24 DRAFT - DO NOT CITE OR QUOTE
-------
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA. (2004). Carbon monoxide actuates O2-limited heme
degredation in the rat brain. Free Radic Biol Med, 37: 1802-1812. 180440
Cudmore M; Ahmad S; Al-Ani B; Fujisawa T; Coxall H; Chudasama K; Devey LR; Wigmore SJ; Abbas A; Hewett PW;
Ahmed A. (2007). Negative regulation of soluble Fit-1 and soluble endoglin release by heme oxygenase-1.
Circulation, 36: 2973-2979. 193991
D'Amico G; Lam F; Hagen T; Moncada S. (2006). Inhibition of cellular respiration by endogenously produced carbon
monoxide. ,119: 2291-2298. 193992
Dani C; Giannini L; Bertini G; Pratesi S; Corsini I; Longini M; Buonocore G; Masini E; Rubaltelli FF. (2007). Changes of
nitric oxide, carbon monoxide and oxidative stress in term infants at birth. Free Radic Res, 41: 1358-1363. 193994
De Luca A; Pierno S; Tricarico D; Carratu MR; Cagiano R; Cuomo V; Camerino DC. (1996). Developmental changes of
membrane electrical properties of rat skeletal muscle fibers produced by prenatal exposure to carbon monoxide.
Environ Toxicol Pharmacol, 2: 213-221. 080911
De Salvia MA; Cagiano R; Carratu MR; Di Giovanni V; Trabace L; Cuomo V. (1995). Irreversible impairment of active
avoidance behavior in rats prenatally exposed to mild concentrations of carbon monoxide. Psychopharmacology
(Berl), 122: 66-71. 079441
Denschlag D; Marculescu R; Unfried G; Hefler LA; Exner M; Hashemi A; Riener EK; Keck C; Tempfer CB; Wagner O.
(2004). The size of a microsatellite polymorphism of the haem oxygenase 1 gene is associated with idiopathic
recurrent miscarriage. , 10: 211-4. 193894
Dewilde S; Kiger L; Burmester T; Hankeln T; Baudin-Creuza V; Aerts T; Marden MC; Caubergs R; Moens L. (2001).
Biochemical characterization and ligand binding properties of neuroglobin, a novel member of the globin family. J
Biol Chem, 276: 38949-38955. 019318
Di Giovanni V; Cagiano R; De Salvia MA; Giustino A; Lacomba C; Renna G; Cuomo V. (1993). Neurobehavioral changes
produced in rats by prenatal exposure to carbon monoxide. Brain Res, 616: 126-131. 013822
Dubuis E; Gautier M; Melin A; Rebocho M; Girardin C; Bonnet P; Vandier C. (2002). Chronic carbon monoxide enhanced
IbTx-sensitive currents in rat resistance pulmonary artery smooth muscle cells. Am J Physiol Lung Cell Mol
Physiol, 283: L120-L129. 193911
Dubuis E; Gautier M; Melin A; Rebocho M; Girardin C; Bonnet P; Vandier C. (2003). Chronic carbon monoxide exposure
of hypoxic rats increases in vitro sensitivity of pulmonary artery smooth muscle. Can J Physiol Pharmacol, 81: 711-
719. 180439
Dubuis E; Potier M; Wang R; Vandier C. (2005). Continuous inhalation of carbon monoxide attenuates hypoxic
pulmonaryhypertension development presumably through activation of BKCa channels. Cardiovasc Res, 65: 751-
761. 180435
Durante W; Johnson FK; Johnson RA. (2006). Role of carbon monoxide in cardiovascular function. ,10: 672-686. 193778
Favory R; Lancel S; Tissier S; Mathieu D; Decoster B; Neviere R. (2006). Myocardial dysfunction and potential cardiac
hypoxia in rats induced by carbon monoxide inhalation. Am J Respir Crit Care Med, 174: 320-325. 184462
Fechter LD; AnnauZ. (1977). Toxicity of mild prenatal carbon monoxide exposure. ,197: 680-682. 010688
Fechter LD; Annau Z. (1980). Prenatal carbon monoxide exposure alters behavioral development. Neurotoxicol Teratol, 2:
7-11.011295
Fechter LD; Karpa MD; Proctor B; Lee AG; Storm JE. (1987). Disruption of neostriatal development in rats following
perinatal exposure to mild, but chronic carbon monoxide. Neurotoxicol Teratol, 9: 277-281. 012259
Fechter LD; Liu Y; Pearce TA. (1997). Cochlear protection from carbon monoxide exposure by free radical blockers in the
guinea pig. Toxicol Appl Pharmacol, 142: 47-55. 081322
Fechter LD; Mactutus CF; Storm JE. (1986). Carbon monoxide and brain development. Neurotoxicology, 7: 463-473.
012030
Fechter LD; Thakur M; Miller B; Annau Z; Srivastava U. (1980). Effects of prenatal carbon monoxide exposure on cardiac
development. Toxicol Appl Pharmacol, 56: 370-375. 011294
Fechter LD; Thorne PR; Nuttall AL. (1987). Effects of carbon monoxide on cochlear electrophysiology and blood flow.
Hear Res, 27: 37-45. 012194
September 2009 E-25 DRAFT - DO NOT CITE OR QUOTE
-------
Garofolo MC; Seidler FJ; Auman JT; Slotkin TA. (2002). beta-Adrenergic modulation of muscarinic cholinergic receptor
expression and function in developing heart. Am J Physiol Regul Integr Comp Physiol, 282: R1356-1363. 193930
Gautier M; Antier D; Bonnet P; Le Net JL; Hanton G; Eder V. (2007). Continuous inhalation of carbon monoxide induces
right ventricle ischemia and dysfunction in rats with hypoxic pulmonary hypertension. Am J Physiol Heart Circ
Physiol, 293: H1046-52. 096471
Gaworski CL; Carmines EL; Faqi AS; Rajendran N. (2004). In utero and lactation exposure of rats to 1R4F reference
cigarette mainstream smoke: Effect on prenatal and postnatal development. Toxicol Sci, 79: 157-169. 193933
Ohio AJ; Stonehuerner JG; Dailey LA; Richards JH; Madden MD; Deng Z; Nguyen NB; Callaghan KD; Yang F;
Piantadosi CA. (2008). Carbon Monoxide Reversibly Alters Iron Homeostasis and Respiratory Epithelial Cell
Function. Am J Respir Cell Mol Biol, 38: 715-723. 096321
Giustino A; Cagiano R; Carratu MR; Cassano T; Tattoli M; Cuomo V. (1999). Prenatal exposure to low concentrations of
carbon monoxide alters habituation and non-spatial working memory in rat offspring. Brain Res, 844: 201-205.
011538
Giustino A; Cagiano R; Carratu MR; De Salvia MA; Panaro MA; Jirillo E; Cuomo V. (1993). Immunological changes
produced in rats by prenatal exposure to carbon monoxide. , 73: 274-278. 013833
Giustino A; Carratu MR; Brigiani GS; De Salvia MA; Pellegrino NM; Steardo L; Jirillo E; Cuomo V. (1994). Changes in
the frequency of splenic immunocompetent cells in rats exposed to carbon monoxide during gestation.
Immunopharmacol Immunotoxicol, 16: 281-292. 076343
Glabe A; Chung Y; Xu D; Jue T. (1998). Carbon monoxide inhibition of regulatory pathways in myocardium. Am J
Physiol, 274: H2143-2151. 086704
Grover TR; Rairigh RL; Zenge JP; Abman SH; Kinsella JP (2000). Inhaled carbon monoxide does not cause pulmonary
vasodilation in the late-gestation fetal limb. Am J Physiol, 278: L779-L784. 010465
Kara S; Mukai T; Kurosaki K; Kuriiwa F; Endo T. (2002). Modification of the striatal dopaminergic neuron system by
carbon monoxide exposure in free-moving rats, as determined by in vivo brain microdialysis. Arch Toxicol, 76:
596-605. 037497
Harada T; Koi H; Kubota T; Aso T. (2004). Haem oxygenase augments porcine granulosa cell apoptosis in vitro. J
Endocrinol, 181: 191-205. 193920
Hendler I; Baum M; Kreiser D; Schiff E; Druzin M; Stevenson DK; Seidman DS. (2004). End-tidal breath carbon
monoxide measurements are lower in pregnant women with uterine contractions. , 24: 275-278. 193925
Hofmann OM; Brittain T. (1998). Partitioning of oxygen and carbon monoxide in the three human embryonic hemoglobins.
Hemoglobin, 22: 313-319. 052019
Iheagwara KN; Thorn SR; Deutschman CS; Levy RJ. (2007). Myocardial cytochrome oxidase activity is decreased
following carbon monoxide exposure. Biochim Biophys Acta, 1772: 1112-1116. 193861
Imai T; Morita T; Shindo T; Nagai R; Yazaki Y; Kurihara H; Suematsu M; Katayama S. (2001). Vascular smooth muscle
cell-directed overexpression of heme oxygenase-1 elevates blood pressure through attenuation of nitric oxide-
induced vasodilation in mice. Circ Res, 89: 55-62. 193864
Ischiropoulos H; Beers MF; Ohnishi ST; Fisher D; Garner SE; Thorn SR. (1996). Nitric oxide production and perivascular
tyrosine nitration in brain after carbon monoxide poisoning in the rat. J Clin Invest, 97: 2260-2267. 079491
Johnson FK; Durante W; Peyton KJ; Johnson RA. (2003). Heme oxygenase inhibitor restores arteriolar nitric oxide
function in Dahl rats. Hypertension, 41: 149-155. 193868
Johnson FK; Durante W; Peyton KJ; Johnson RA. (2004). Heme oxygenase-mediated endothelial dysfunction in DOCA-
salt, but not in spontaneously hypertensive, rat arterioles. Am J Physiol Heart Circ Physiol, 286: 1681-1687. 193870
Johnson FK; Johnson RA. (2003). Carbon monoxide promotes endothelium-dependent constriction of isolated gracilis
muscle arterioles. Am J Physiol, 285: R536-R541. 053611
Johnson FK; Johnson RA; Durante W; Jackson KE; Stevenson BK; Peyton KJ. (2006). Metabolic syndrome increases
endogenous carbon monoxide production to promote hypertension and endothelial dysfunction in obese Zucker
rats. Am J Physiol Regul Integr Comp Physiol, 290: 601-608. 193874
September 2009 E-26 DRAFT - DO NOT CITE OR QUOTE
-------
Katoue MG; Khan I; Oriowo MA. (2005). Increased expression and activity of heme oxygenase-2 in pregnant rat aorta is
not involved in attenuated vasopressin-induced contraction. Naunyn Schmiedebergs Arch Pharmacol, 372: 220-7.
193896
Katoue MG; Khan I; Oriowo MA. (2006). Pregnancy-induced modulation of calcium mobilization and down-regulation of
Pvho-kinase expression contribute to attenuated vassopressin-induced contraction of the rat aorta. Vascul Pharmacol,
44: 170-176. 193954
Khan AA; Wang Y; Sun Y; Mao XO; Xie L; Miles E; Graboski J; Chen S; Ellerby LM; Jin K; Greenberg DA. (2006).
Neuroglobin-overexpressing transgenic mice are resistant to cerebral and myocardial ischemia. Proc Natl Acad Sci
USA, 103: 17944-17948. 193955
Kim HP; Wang X; Nakao A; Kim SI; Murase N; Choi ME; Ryter SW; Choi AM. (2005). Caveolin-1 expression by means
of mitogen-activated protein kinase mediates the antiprofiferative effect of carbon monoxide . Proc Natl Acad Sci U
S A, 102: 11319-11324. 193959
Kim HS; Loughran PA; Rao J; Billar TR; Zuckerbraun BS. (2008). Carbon monoxide activates NF-kappaB via ROS
generation and Akt pathways to protect against cell death of hepatocytes. , 295: G146-G152. 193961
Kinobe R; Vlahakis J; Soong J; Szarek W; Brien J; Longo L; Nakatsu K. (2006). Heme oxygenase activity in fetal and
adult sheep is not altered by acclimatization to high altitude hypoxia. Can J Physiol Pharmacol, 84: 893 - 901.
188447
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling of
eNOS. Toxicol Appl Pharmacol, 230: 346. 191987
Korres S; Riga M; Balatsouras D; Papadakis C; Kanellos P; Ferekidis E. (2007). Influence of smoking on developing
cochlea: Does smoking during pregnancy affect the amplitudes of transient evoked otoacoustic emissions in
newborns?. Int J Pediatr Otorhinolaryngol, 71: 781-786. 190908
Kreiser D; Baum M; Seidman DS; Fanaroff A; Shah D; Hendler I; Stevenson DK; Schiff E; Druzin ML. (2004). End tidal
carbon monoxide levels are lower in women with gestational hypertension and pre-eclampsia. ,24: 213-217.
193948
Lash GE; McLaughlin BE; MacDonald-Goodfellow SK; Smith GN; Brien JF; Marks GS; Nakatsu K; Graham CH. (2003).
Relationship between tissue damage and heme oxygenase expression in chorionic villi of term human placenta. Am
J Physiol Heart Circ Physiol, 284: H160-H167. 193849
Li X; Zhang C; Wang S; Ji Y; Wang H; Zhao L; Xu D. (2008). Maternally administered lipopolysaccharide (LPS)
upregulates the expression of heme oxygenase-1 in fetal liver: The role of reactive oxygen species. Toxicol Lett,
176: 169-177. 187003
Liu Y; Fechter LD. (1995). MK-801 protects against carbon monoxide-induced hearing loss. Toxicol Appl Pharmacol, 132:
196-202.076524
Loennechen JP; Beisvag V; Arbo I; Waldum HL; Sandvik AK; Knardahl S; Ellingsen O. (1999). Chronic carbon monoxide
exposure in vivo induces myocardial endothelin-1 expression and hypertrophy in rat. , 85: 192-197. 011549
Longo M; Jain V; Vedernikov YP; Saade GR; Goodrum L; Facchinetti F; Garfield RE. (1999). Effect of nitric oxide and
carbon monoxide on uterine contractility during human and rat pregnancy. Am J Obstet Gynecol, 181: 981-988.
011548
Lopez I; Acuna D; Webber DS; Korsak RA; Edmond J. (2003). Mild carbon monoxide exposure diminishes selectively the
integrity of the cochlea of the developing rat. , 74: 666-75. 193901
Lopez IA; Acuna D; Beltran-Parrazal L; Espinosa-Jeffrey A; Edmond J. (2008). Oxidative stress and the deleterious
consequences to the rat cochlea after prenatal chronic mild exposure to carbon monoxide in air. Neuroscience, 151:
854-867. 097343
Lund AK; Knuckles TL; Obot Akata C; ShohetR; McDonald JD; Gigliotti A; Seagrave JC; Campen MJ. (2007). Gasoline
exhaust emissions induce vascular remodeling pathways involved in atherosclerosis. Toxicol Sci, 95: 485-94.
125741
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009). Vehicular
emissions induce vascular MMP-9 expression and activity associated with endothelin-l?mediated pathways.
Arterioscler Thromb Vase Biol, 29: 511-517. 180257
September 2009 E-27 DRAFT - DO NOT CITE OR QUOTE
-------
Lyall F; Barber A; Myatt L; Buhner JN; Robson SC. (2000). Hemeoxygenase expression in human placenta and placental
bed implies a role in regulation of trophoblast invasion and placental function. , 14: 208-19. 193902
Lyall F; Myatt L. (2002). The role of the placenta in pre-eclampsia — a workshop report. Placenta, 23: S142-145. 193971
Mactutus CF; Fechter LD. (1984). Prenatal exposure to carbon monoxide: learning and memory deficits. , 223: 409-411.
011355
McGregor HP; Westcott K; Walker DW. (1998). The effect of prenatal exposure to carbon monoxide on breathing and
growth of the newborn guinea pig. Pediatr Res, 43: 126-131. 085342
McLaughlin BE; Hutchinson JM; Graham CH; Smith GN; Marks GS; Nakatsu K; Brien JF. (2000). Heme oxygenase
activity in term human placenta. Placenta, 21: 870-873. 015815
McLaughlin BE; Lash GE; Graham CH; Smith GN; Vreman HJ; Stevenson DK; Marks GS; Nakatsu K; Brien JF. (2001).
Endogenous carbon monoxide formation by chorionic villi of term human placenta. Placenta, 22: 886-888. 193823
McLaughlin BE; Lash GE; Smith GN; Marks GS; Nakatsu K; Graham CH; Brien JF. (2003). Heme oxygenase expression
in selected regions of term human placenta. Exp Biol Med (Maywood), 228: 564-567. 193827
McLean M; Bowman M; Clifton V; Smith R; Grossman AB. (2000). Expression of the heme oxygenase - carbon monoxide
signalling system in human placenta. J Clin Endocrinol Metab, 85: 2345-2349. 016269
Melin A; Bonnet P; Eder V; Antier D; Obert P; Fauchier L. (2005). Direct implication of carbon monoxide in the
development of heart failure in rats with cardiac hypertrophy subjected to air pollution. Cardiovasc Toxicol, 5: 311-
320. 193833
Melin A; Obert P; Rebocho M; Bonnet P. (2002). Cardiac morphology and function following long-term exposure to
carbon monoxide at high altitude in rats. J Toxicol Environ Health A, 65:1981-1998. 037502
Mereu G; Cammalleri M; Fa M; Francesconi W; Saba P; Tattoli M; Trabace L; Vaccari A; Cuomo V. (2000). Prenatal
exposure to a low concentration of carbon monoxide disrupts hippocampal long-term potentiation in rat offspring. J
Pharmacol Exp Ther, 294: 728-734. 193838
Middendorff R; Kumm M; Davidoff MS; Holstein AF; Muller D. (2000). Generation of cyclic guanosine monophosphate
by heme oxygenases in the human testis—a regulatory role for carbon monoxide in Sertoli cells?. Biol Reprod, 63:
651-657. 015842
Montagnani M; Serio M; Potenza MA; Mansi G; De Salvia MA; Cagiano R; Cuomo V; Mitolo-Chieppa D. (1996).
Prenatal exposure to carbon monoxide and vascular responsiveness of rat resistance vessels. Life Sci, 59: 1553-
1561.080902
Naik JS; Walker BR. (2003). Heme oxygenase-mediated vasodilation involves vascular smooth muscle cell
hyperpolarization. Am J Physiol Heart Circ Physiol, 285: 220-228. 193852
Ndisang JF; Tabien HEN; Wang R. (2004). Carbon monoxide and hypertension. Am J Hypertens, 22: 1057-1074. 180425
Neggers YH; Singh J. (2006). Zinc supplementation to protein-deficient diet in CO-exposed mice decreased fetal mortality
and malformation. Biol Trace Elem Res, 114: 269-279. 193964
Newby D; Cousins F; Myatt L; Lyall F. (2005). Heme oxygenase expression in cultured human trophoblast cells during in
vitro differentiation: Effects of hypoxia. Placenta, 26: 201-209. 193966
Odrcich MJ; Graham CH; Kimura KA; McLaughlin BE; Marks GS; Nakatsu K; Brien JF. (1998). Heme oxygenase and
nitric oxide synthase in the placenta of the guinea-pig during gestation. Placenta, 19: 509-516. 193958
Ozawa N; Goda N; Makino N; Yamaguchi T; Yoshimura Y; Suematsu M. (2002). Ley dig cell-derived heme oxygenase-1
regulates apoptosis of premeiotic germ cells in response to stress. , 109: 457-467. 193841
Patel AP; Moody JA; Handy RD; Sneyd JR. (2003). Carbon monoxide exposure in rat heart: glutathione depletion is
prevented by antioxidants. Biochem Biophys Res Commun, 302: 392-396. 043155
Penney DG; BaylerianMS; Thill JE; Fanning CM; Yedavally S. (1982). Postnatal carbon monoxide exposure: immediate
and lasting effects in the rat. Am J Physiol, 243: H328-H339. 011387
Penney DG; Baylerian MS; Thill JE; Yedavally S; Fanning CM. (1983). Cardiac response of the fetal rat to carbon
monoxide exposure. Am J Physiol, 244: H289-H297. 011385
Piantadosi CA. (2002). Biological chemistry of carbon monoxide. Antioxid Redox Signal, 4: 259-270. 037463
September 2009 E-28 DRAFT - DO NOT CITE OR QUOTE
-------
Piantadosi CA. (2008). Carbon monoxide, reactive oxygen signaling, and oxidative stress. Free Radic Biol Med, 45: 562-
569. 180423
Piantadosi CA; Carraway MS; Suliman HB. (2006). Carbon monoxide, oxidative stress, and mitochondrial permeability
pore transition. Free Radic Biol Med, 40: 1332-1339. 180424
Prigge E; Hochrainer D. (1977). Effects of carbon monoxide inhalation on erythropoiesis and cardiac hypertrophy in fetal
rats. Toxicol Appl Pharmacol, 42: 225-228. 012326
Raub JA; Benignus VA. (2002). Carbon monoxide and the nervous system. Neurosci Biobehav Rev, 26: 925-940. 041616
Richardson RS; Noyszewski EA; Saltin B; Gonzalez-Alonso J; et al. (2002). Effect of mild carboxy-hemoglobin on
exercising skeletal muscle: Intravascular and intracellular evidence. Am J Physiol, 283: R1131-R1139. 037513
Ryter SW; Alam J; Choi AMK. (2006). Heme oxygenase-1/carbon monoxide: From basic science to therapeutic
applications . Physiol Rev, 86: 583-650. 193765
Sartiani L; Cerbai E; Lonardo G; DePaoli P; Tattoli M; Cagiano R; Carratu M; Cuomo V; Mugelli A. (2004). Prenatal
Exposure to Carbon Monoxide Affects Postnatal Cellular Electrophysiological Maturation of the Rat Heart: A
Potential Substrate for Arrhythmogenesis in Infancy. Circulation, 109-: 419-423. 190898
Schwetz BA; Smith FA; Leong BKJ; Staples RE. (1979). Teratogenic potential of inhaled carbon monoxide in mice and
rabbits. , 19: 385-391.011855
Singh J. (1986). Early behavioral alterations in mice following prenatal carbon monoxide exposure. Neurotoxicology, 7:
475-481.012827
Singh J. (2003). Gastroschisis is caused by the combination of carbon monoxide and protein-zinc deficiencies in mice.
Birth Defects Res B Dev Reprod Toxicol, 68: 355-362. 053624
Singh J. (2006). Interaction of maternal protein and carbon monoxide on pup mortality in mice: implications for global
infant mortality. Birth Defects Res B Dev Reprod Toxicol, 77: 216-26. 190512
Singh J; Aggison L Jr; Moore-Cheatum L. (1993). Teratogenicity and developmental toxicity of carbon monoxide in
protein-deficient mice. , 48: 149-159. 013892
Singh J; Scott LH. (1984). Threshold for carbon monoxide induced fetotoxicity. , 30: 253-257. 011409
Singh J; Smith CB; Moore-Cheatum L. (1992). Additivity of protein deficiency and carbon monoxide on placental
carboxyhemoglobin in mice. Am J Obstet Gynecol, 167: 843-846. 013759
Sitdikova GF; Islamov RR; Mukhamedyarov MA; Permyakova VV; Zefirov AL; Palotas A. (2007). Modulation of
neurotransmitter release by carbon monoxide at the frogneuro-muscular junction. Curr Drug Metab, 8: 177-184.
180417
Song R; Mahidhara RS; Liu F; Ning W; Otterbein LE; Choi AMK. (2002). Carbon monoxide inhibits human airway
smooth muscle cell proliferation via mitogen-activated protein kinase pathway. Am J Respir Cell Mol Biol, 27: 603-
610.037531
Stevens CF; Wang Y. (1993). Reversal of long-term potentiation by inhibitors of haem oxygenase. Nature, 364: 147-149.
188458
Stockard-Sullivan JE; Korsak RA; Webber DS; Edmond J. (2003). Mild carbon monoxide exposure and auditory function
in the developing rat. J Neurosci Res, 74: 644-654. 190947
Storm JE; Fechter LD. (1985). Alteration in the postnatal ontogeny of cerebellar norepinephrine content following chronic
prenatal carbon monoxide. JNeurochem, 45: 965-969. 011653
Storm JE; Fechter LD. (1985). Prenatal carbon monoxide exposure differentially affects postnatal weight and monoamine
concentration of rat brain regions. Toxicol Appl Pharmacol, 81: 139-146. 011652
Storm JE; Valdes JJ; Fechter LD. (1986). Postnatal alterations in cerebellar GABA content, GABA uptake and morphology
following exposure to carbon monoxide early in development. Dev Neurosci, 8: 251-261. 012136
StykaPE; Penney DG (1978). Regression of carbon monoxide-induced cardiomegaly. Am J Physiol, 235: H516-H522.
011166
Suliman HB; Carraway MS; Tatro LG; Piantadosi CA. (2007). A new activating role for CO in cardiac mitochondrial
biogenesis. , 120: 299-308. 193768
September 2009 E-29 DRAFT - DO NOT CITE OR QUOTE
-------
Sun Y; Jin K; Mao XO; Zhu Y; Greenberg DA. (2001). Neuroglobin is up-regulated by and protects neurons from hypoxic-
ischemic injury. Proc Natl Acad Sci USA, 98: 15,306-15,311.026022
S0rhaug S; Steinshamn S; Nilsen OG; Waldum HL. (2006). Chronic inhalation of carbon monoxide: Effects on the
respiratory and cardiovascular system at doses corresponding to tobacco smoking. Toxicology, 228: 280-290.
180414
Tattoli M; Carratu MR; Cassano T; Cortese I; Di Giovanni V; Elia G; Renna G; Cagiano R. (1999). Effects of early
postnatal exposure to low concentrations of carbon monoxide on cognitive functions in rats. Pharmacol Res, 40:
271-274. 011557
Telfer JF; Itoh H; Thomson AJ; Norman JE; Nakao K; Campa JS; Poston L; Tribe RM; Magness RR. (2001). Activity and
expression of soluble and particulate guanylate cyclases in myometrium from nonpregnant and pregnant women:
down-regulation of soluble guanylate cyclase at term. J Clin Endocrinol Metab, 86: 5934-5943. 193769
Teran FJ; Johnson RA; Stevenson BK; Peyton KJ; Jackson KE; Appleton SD; Durante W; Johnson FK. (2005). Heme
oxygenase-derived carbon monoxide promotes arteriolar endothelial dysfunction and contributes to salt-induced
hypertension in Dahl salt-sensitive rats. Am J Physiol Regul Integr Comp Physiol, 288: 615-622. 193770
Thorn SR; Bhopale VM; Han ST; Clark JM; Hardy KR. (2006). Intravascular neutrophil activation due to carbon monoxide
poisoning. Am J Respir Grit Care Med, 174: 1239-1248. 098418
Thorn SR; Fisher D; Manevich Y. (2001). Roles for platelet-activating factor and *NO-derived oxidants causing neutrophil
adherence after CO poisoning. Am J Physiol Heart Circ Physiol, 281: H923-H930. 193779
Thorn SR; Fisher D; Xu YA; Garner S; Ischiropoulos H. (1999). Role of nitric oxide-derived oxidants in vascular injury
from carbon monoxide in the rat. Am J Physiol, 276: H984-H992. 016753
Thorn SR; Fisher D; Xu YA; Notarfrancesco K; Ischiropoulos H. (2000). Adaptive responses and apoptosis in endothelial
cells exposed to carbon monoxide. Proc Natl Acad Sci US A, 97: 1305-1310. 011574
Thorn SR; Ischiropoulos H. (1997). Mechanism of oxidative stress from low levels of carbon monoxide. 085644
Thorn SR; Ohnishi ST; Fisher D; Xu YA; Ischiropoulos H. (1999). Pulmonary vascular stress from carbon monoxide.
ToxicolAppl Pharmacol, 154: 12-19. 016757
Thorn SR; Ohnishi ST; Ischiropoulos H. (1994). Nitric oxide released by platelets inhibits neutrophil B2 integrin function
following acute carbon monoxide poisoning. ToxicolAppl Pharmacol, 128: 105-110. 076459
Thorn SR; Xu YA; Ischiropoulos H. (1997). Vascular endothelial cells generate peroxynitrite in response to carbon
monoxide exposure. Chem Res Toxicol, 10: 1023-1031. 084337
Thorup C; Jones CL; Gross SS; Moore LC; Goligorsky MS. (1999). Carbon monoxide induces vasodilation and nitric
oxide release but suppresses endothelial NOS. Am J Physiol, 277: F882-F889. 193782
Tolcos M; Mallard C; McGregor H; Walker D; Rees S. (2000). Exposure to prenatal carbon monoxide and postnatal
hyperthermia: short and long-term effects on neurochemicals and neuroglia in the developing brain. Exp Neurol,
162: 235-246. 010468
Tolcos M; McGregor H; Walker D; Rees S. (2000). Chronic prenatal exposure to carbon monoxide results in a reduction in
tyrosine hydroxylase-immunoreactivity and an increase in choline acetyltransferase-immunoreactivity in the fetal
medulla: implications for sudden infant death syndrome. J Neuropathol Exp Neurol, 59: 218-228. 015997
Toyoda M; Saito H; Matsuki N. (1996). Nitric oxide but not carbon monoxide is involved in spatial learning of mice. J
Pharmacol Sci, 71: 205-211. 079945
Tschugguel W; Stonek F; Zhegu Z; Dietrich W; Schneeberger C; Stimpfl T; Waldhoer T; Vycudilik W; Huber JC. (2001).
Estrogen increases endothelial carbon monoxide, heme oxygenase 2, and carbon monoxide-derived cGMP by a
receptor-mediated system. J Clin Endocrinol Metab, 86: 3833-3839. 193785
Vallone B; Nienhaus K; Matthes A; Brunori M; Nienhaus GU. (2004). The structure of carbonmonoxy neuroglobin reveals
a heme-sliding mechanism for control of ligand affinity. , 101: 17351-17356. 193993
Villamor E; Perez-Vizcaino F; Cogolludo AL; Conde-Oviedo J; Zaragoza-Arnaez F; Lopez-Lopez JG; Tamargo J. (2000).
Relaxant effects of carbon monoxide compared with nitric oxide in pulmonary and systemic vessels of newborn
piglets. PediatrRes, 48: 546-553. 015838
September 2009 E-30 DRAFT - DO NOT CITE OR QUOTE
-------
Vreman HJ; Wong RJ; Kadotani T; Stevenson DK. (2005). Determination of carbon monoxide (CO) in rodent tissue: effect
of heme administration and environmental CO exposure. Anal Biochem, 341: 280-289. 193786
Vreman HJ; Wong RJ; Kim EC; Nabseth DC; Marks GS; Stevenson DK. (2000). Haem oxygenase activity in human
umbilical cord and rat vascular tissues. Placenta, 21: 337-44. 096915
Weaver LK; Valentine KJ; Hopkins RO. (2007). Carbon monoxide poisoning: risk factors for cognitive sequelae and the
role of hyperbaric oxygen. Am J Respir Grit Care Med, 176: 491 - 497. 193939
Webber D; Lopez I; Korsak R; Hirota S; Acuna D; Edmond J. (2005). Limiting iron availability confers neuroprotection
from chronic mild carbon monoxide exposure in the developing auditory system of the rat. J Neurosci Res, 80: 620-
633.190514
Webber DS; Korsak RA; Sininger LK; Sampogna SL; Edmond J. (2003). Mild carbon monoxide exposure impairs the
developing auditory system of the rat. J Neurosci Res, 74: 655-665. 190515
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac effects of
carbon monoxide and ambient particles in a rat model of myocardial infarction. Toxicol Sci, 80: 367-376. 087874
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal Toxicol, 18: 1077-1082.
156152
Yoshiki N; Kubota T; Aso T. (2001). Identification of heme oxygenase in human endometrium. J Clin Endocrinol Metab,
86: 5033-5038. 193790
Yu, S; Zhang, R; Chen, Y; Chen, J; Zhu, C; Tang, J; Li, G (2008). Possible contribution of endogenous carbon monoxide to
the development of allergic rhinitis in guinea pigs. , 5: -. 192384
Zamudio S; Palmer SK; Dahms TE; Berman JC; Young DA; Moore LG (1995). Alterations in uteroplacental blood flow
precede hypertension in preeclampsia at high altitude. J Appl Physiol, 79: 15-22. 193908
Zenclussen ML; Anegon I; Bertoja AZ; Chauveau C; Vogt K; Gerlof K; Sollwedel A; Volk HD; Ritter T; Zenclussen AC.
(2006). Over-expression of heme oxygenase-1 by adenoviral gene transfer improves pregnancy outcome in a
murine model of abortion. , 69: 35-52. 193873
Zhang C; Li XY; Zhao L; Wang H; Xu DX. (2007). Lipopolysaccharide (LPS) up-regulates the expression of haem
oxygenase-1 in mouse placenta. Placenta, 28: 951-957. 193879
Zhang X; Shan P; Alam J; Fu XY; Lee PJ. (2005). Carbon monoxide differentially modulates STAT1 and STAT3
andinhibits apoptosis via a phosphatidylinositol 3-kinase/akt and p38 kinase-dependent STAT3 pathway during
anoxia-reoxygenation injury. J Biol Chem, 280: 8714-8721. 184460
Zhao H; Wong RJ; Doyle TC; Nayak N; Vreman HJ; Contag CH; Stevenson DK. (2008). Regulation of maternal and fetal
hemodynamics by heme oxygenase in mice. Biol Reprod, 78: 744-751. 193883
Zhuo M; Small SA; Kandel ER; Hawkins RD. (1993). Nitric oxide and carbon monoxide produce activity-dependent long-
term synaptic enhancement in hippocampus. , 260: 1946-1950. 013905
Zuckerbraun BS; Chin BY; Bilban M; d'Avila JC; Rao J; Billiar TR; Otterbein LE. (2007). Carbon monoxide signals via
inhibition of cytochrome c oxidase and generation of mitochondrial reactive oxygen species. , 21: 1099-1106.
193884
September 2009 E-31 DRAFT - DO NOT CITE OR QUOTE
------- |