u n ited States January 2010
Environmental Protection m \/sr\r\m nrwmrn-
Agency EPA/600/R-09/019F
Integrated Science Assessment for
Carbon Monoxide
National Center for Environmental Assessment-RTF Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
January 2010
-------
Table of Contents
TABLE OF CONTENTS
LIST OF TABLES vm
LIST OF FIGURES xn
CO PROJECT TEAM xvm
AUTHORS, CONTRIBUTORS, AND REVIEWERS xx
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE CO NAAQS REVIEW PANEL xxiv
ACRONYMS AND ABBREVIATIONS xxvi
CHAPTER 1. INTRODUCTION 1 -1
1.1. Legislative Requirements
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
References
1-2
1-3
1-4
1-6
1-7
1-8
1-8
1-9
1-9
1-12
1-13
1-15
1-16
1-16
1-17
1-18
CHAPTER 2. INTEGRATIVE OVERVIEW 2-1
2.1. Ambient CO Sources and Concentrations 2-1
2.2. Climate Forcing Effects 2-2
2.3. Exposure to Ambient CO 2-3
2.4. Dosimetry, Pharmacokinetics, and Mode of Action 2-4
2.4.1. Dosimetry and Pharmacokinetics 2-4
2.4.2. Mode of Action 2-4
2.5. Health Effects 2-5
2.5.1. Cardiovascular Morbidity 2-5
2.5.2. Central Nervous System Effects 2-6
January 2010
-------
2.5.3. Birth Outcomes and Developmental Effects 2-7
2.5.4. Respiratory Morbidity 2-8
2.5.5. Mortality 2-9
2.6. Policy-Relevant Considerations 2-10
2.6.1. Susceptible Populations 2-10
2.6.2. Concentration- and Dose-Response Relationships 2-12
2.7. Integration of CO Health Effects 2-13
References 2-18
CHAPTER 3. SOURCE TO EXPOSURE 3-1
3.1. Introduction 3-1
3.2. CO Sources, Emissions, and Chemistry 3-1
3.2.1. Direct CO Emissions 3-1
3.2.2. Secondary CO Emissions and Associated Chemistry 3-9
3.3. CO Climate Forcing Effects 3-11
3.4. Ambient Measurements 3-16
3.4.1. Ambient Measurement Instruments 3-16
3.4.2. Ambient Sampling Network Design 3-19
3.4.2.1. Monitor Siting Requirements 3-19
3.4.2.2. Spatial and Temporal Coverage 3-20
3.5. Environmental Concentrations 3-28
3.5.1. Spatial Variability 3-28
3.5.1.1. National Scale 3-28
3.5.1.2. Urban Scale 3-37
3.5.1.3. Micro- to Neighborhood Scale and the Near-Road Environment 3-53
3.5.2. Temporal Variability 3-62
3.5.2.1. Multiyear Trends 3-62
3.5.2.2. Hourly Variation 3-65
3.5.3. Associations with Copollutants 3-68
3.5.4. Policy-Relevant Background 3-72
3.5.4.1. Surface-Based Determinations 3-72
3.5.4.2. Limitations of Other Possible Methods 3-74
3.6. Issues in Exposure Assessment 3-76
3.6.1. Summary of Findings from 2000 CO AQCD 3-76
3.6.2. General Exposure Concepts 3-76
3.6.3. Exposure Modeling 3-78
3.6.3.1. Stochastic Population-Based Time-Weighted Microenvironmental
Exposure Models 3-78
3.6.3.2. Using Spatial Models to Estimate Exposure 3-79
3.6.4. Personal Exposure Monitors for CO 3-82
3.6.5. Indoor Exposure to CO 3-82
3.6.5.1. Infiltration of Ambient CO 3-82
3.6.5.2. Exposure to Nonambient CO 3-83
3.6.6. Exposure Assessment Studies at Different Spatial Scales 3-85
3.6.6.1. Neighborhood to Urban Scale Studies of Ambient CO Exposure 3-85
3.6.6.2. Microscale Studies of Ambient CO Exposure: Near-Road and On-Road
Exposures 3-87
3.6.7. Association between Personal CO Exposure and Copollutants 3-90
3.6.8. Implications for Epidemiology 3-90
3.6.8.1. Measurement Error 3-91
3.6.8.2. Exposure Issues Related to Nonambient CO 3-92
3.6.8.3. Spatial Variability 3-92
3.6.8.4. Temporal Variability 3-93
January 2010
-------
3.6.8.5. CO Exposure in Copollutant Mixtures 3-94
3.6.8.6. Conclusions 3-94
3.7. Summary and Conclusions 3-95
3.7.1. CO Sources, Emissions, and Chemistry 3-95
3.7.2. Climate Forcing Effects Related to CO 3-95
3.7.3. Ambient CO Measurements 3-96
3.7.4. Environmental CO Concentrations 3-96
3.7.5. Exposure Assessment and Implications for Epidemiology 3-97
References 3-99
CHAPTER 4. DOSIMETRY AND PHARMACOKINETICS OF CARBON MONOXIDE 4-1
4.1. Introduction 4-1
4.2. Carboxyhemoglobin Modeling 4-1
4.2.1. The Coburn-Forster-Kane and Other Models 4-1
4.2.2. Multicompartment Models 4-6
4.2.3. Model Comparison 4-8
4.2.4. Mathematical Model Usage 4-8
4.3. Absorption, Distribution, and Elimination 4-11
4.3.1. Pulmonary Absorption 4-11
4.3.1.1. Mass Transfer of Carbon Monoxide 4-12
4.3.1.2. Lung Diffusion of Carbon Monoxide 4-12
4.3.2. Tissue Uptake 4-13
4.3.2.1. Respiratory Tract 4-13
4.3.2.2. Blood 4-13
4.3.2.3. Heart and Skeletal Muscle 4-15
4.3.2.4. Other Tissues 4-16
4.3.3. Pulmonary and Tissue Elimination 4-17
4.3.4. COHb Analysis Methods 4-18
4.4. Conditions Affecting Uptake and Elimination 4-19
4.4.1. Physical Activity 4-19
4.4.2. Altitude 4-19
4.4.3. Physical Characteristics 4-20
4.4.3.1. Fetal Pharmacokinetics 4-20
4.4.4. Health Status 4-21
4.5. Endogenous CO Production and Metabolism 4-22
4.6. Summary and Conclusions 4-26
References 4-28
CHAPTERS. INTEGRATED HEALTH EFFECTS 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. Nonhypoxic Mechanisms 5-2
5.1.3.1. Nonhypoxic Mechanisms Reviewed in the 2000 CO AQCD 5-2
5.1.3.2. Recent Studies of Nonhypoxic Mechanisms 5-3
5.1.3.3. Implications of Nonhypoxic Mechanisms 5-9
5.1.3.4. Summary 5-12
5.2. Cardiovascular Effects 5-12
5.2.1. Epidemiologic Studies with Short-Term Exposure 5-12
5.2.1.1. Heart Rate and Heart Rate Variability 5-13
5.2.1.2. ECG Abnormalities Indicating Ischemia 5-16
5.2.1.3. Arrhythmia 5-17
January 2010
-------
5.3.
5.4.
5.5.
5.6.
5.2.1.4. Cardiac Arrest
5.2.1.5. Myocardial Infarction
5.2.1.6. Blood Pressure
5.2.1.7. Vasomotor Function
5.2.1.8. Blood Markers of Coagulation and Inflammation
5.2.1.9. Hospital Admissions and Emergency Department Visits
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
5.2.6. 1 . Short-Term Exposure to CO
5.2.6.2. Long-Term Exposure to CO
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 Postneonatal 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. Respiratory Allergy and Other Allergic Responses
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. Multicity Studies
5.6.1.3. Meta-Analysis of All Criteria Pollutants
5.6.1.4. Single-City Studies
5-18
5-19
5-19
5-19
5-20
5-23
5-40
5-40
5-41
5-44
5-44
5-46
5-47
5-47
5-47
5-47
5-48
5-49
5-49
5-49
5-49
5-50
5-50
5-50
5-53
5-60
5-61
5-63
5-63
5-63
5-68
5-79
5-81
5-81
5-81
5-88
5-96
5-96
5-97
5-98
5-98
5-99
5-99
5-100
5-100
5-101
5-101
5-101
5-101
5-101
5-106
5-107
January 2010
-------
5.7.
5.8.
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. Preexisting Disease
5.7.1.1. Cardiovascular Disease
5.7.1.2. Obstructive Lung Disease
5.7.1.3. Diabetes
5.7.1.4. Anemia
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
5-109
5-109
5-110
5-113
5-114
5-114
5-116
5-116
5-117
5-118
5-118
5-119
5-119
5-120
5-121
5-121
5-122
5-122
5-123
5-123
5-124
5-125
ANNEX A. ATMOSPHERIC SCIENCE A-1
ANNEXE. DOSIMETRY STUDIES B-1
References B-6
ANNEX C. EPIDEMIOLOGY STUDIES C-1
References C-99
ANNEX D. CONTROLLED HUMAN EXPOSURE STUDIES D-1
References D-4
ANNEX E. TOXICOLOGICAL STUDIES E-1
References E-23
January 2010 vii
-------
List of Tables
Table 1 -1. Aspects to aid in judging causality. 1 -13
Table 1 -2. Weight of evidence for causal determination. 1 -14
Table 2-1. Causal determinations for health effects categories. 2-5
Table 2-2. Range of mean and 99th percentile concentrations (ppm) in US and Canadian studies of
short-term CO exposure and CVD hospitalizations. 2-15
Table 3-1. Literature values for CO yields from hydrocarbons in per carbon units, except as noted.
Specific hydrocarbons are noted in parentheses. 3-10
Table 3-2. Performance specifications for analytical detection of CO, based on 40 CFR Part 53. 3-18
Table 3-3. Counts of CO monitors by sampling scale meeting 75% completeness criteria for use in the
U.S. during 2005-2007. 3-21
Table 3-4. Proximity to CO monitors for the total population by city. 3-27
Table 3-5. Proximity to CO monitors for adults aged 65 and older by city. 3-28
Table 3-6. Distribution of 1-h avg CO concentration (ppm) derived from AQS data. 3-31
Table 3-7. Distribution of 24-h avg CO concentration (ppm) derived from AQS data. 3-33
Table 3-8. Distribution of 1-h daily max CO concentration (ppm) derived from AQS data. 3-34
Table 3-9. Distribution of 8-h daily max CO concentration (ppm) derived from AQS data. 3-35
Table 3-10. Table of intersampler 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. 3-40
Table 3-11. Table of intersampler 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. 3-46
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. 3-53
Table 3-13. Percentage of time exposed to ambient CO (adjusted to reflect the absence of nonambient
CO from ETS and gas cooking), average CO exposures, and percentage of exposure
estimated for the population. 3-86
Table 4-1. Predicted COHb levels resulting from 1,8, and 24 h CO exposures in a modeled human at
rest 4-9
Table 4-2. CO concentration in pmol/mg wet weight tissue and fold tissue CO concentration changes
(normalized to background tissue concentrations) - human. 4-16
Table 5-1. Responses to CO exposures at low and moderate concentrations. 5-8
Table 5-2. Tissue concentration of CO following inhalation exposure. 5-10
Table 5-3. Tissue concentration of CO following increased endogenous production. 5-11
Table 5-4. Summary of studies investigating the effect of CO exposure on HRV parameters. 5-16
Table 5-5. Summary of studies investigating the effect of CO exposure on cardiac arrhythmias. 5-18
January 2010 viii
-------
Table 5-6.
Summary of studies investigating the effect of CO exposure on blood markers of
Table 5-7.
Table 5-8.
Table 5-9.
Table 5-10.
Table 5-11.
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.
coaqulation and inflammation.
Summary of CHD hospital admission studies.3
Summary of stroke hospital admission studies.3
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 nonspecific CVD hospital admission studies.
Brief summary of PTB studies.
Brief summary of birth weiqht studies.
Behavioral responses.
Neuronal responses.
Neurotransmitter changes.
Developing auditory system.
Cardiovascular and systemic developmental responses.
Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with short-term exposure to CO.
5-23
5-28
5-31
5-33
5-34
5-37
5-53
5-59
5-69
5-72
5-74
5-76
5-78
5-82
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.
Table A-6.
Table A-7.
Table A-8.
Range of CO concentrations reported in key respiratory HA 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 long-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 long-term exposure to CO.
Range of definitions of "susceptible" and "vulnerable" in the CO literature..
Adult U.S. population in 2007 with respiratory diseases and cardiovascular diseases..
Listing of all CO monitors currently in use, along with their limits of detection.
Microscale monitors meeting 75% completeness criteria, 2005-2007.
Middle scale monitors meeting 75% completeness criteria, 2005-2007..
Neighborhood scale monitors meeting 75% completeness criteria, 2005-2007..
Urban scale monitors meeting 75% completeness criteria, 2005-2007.
Regional scale monitors meeting 75% completeness criteria, 2005-2007.
Monitors meeting 75% completeness criteria, 2005-2007 with no scale delared.
Numbers of high LOD and trace-level monitors in each state that met completeness criteria
for 2005-2007.
5-8
5-96
_5-102
_5-107
A-8
A-9
A-11
A-12
A-15
A-15
A-16
A-18
January 2010
-------
Table A-9.
Table A-10.
Table A-11.
TableA-12.
TableA-13.
TableA-14.
TableA-15.
TableA-16.
TableA-17.
TableA-18.
TableA-19.
Table A-20.
Table A-21.
Table A-22.
Table A-23.
Table A-24.
Table A-25.
Table A-26.
Table B-1.
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 City, NY.
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.
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 for hourly, 1-h daily max, 24-h avg, and 8-h daily max
data for Seattle, WA.
Comparison of distributional data 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..
A-30
A-33
A-36
A-39
A-42
A-45
A-48
A-53
A-55
A-55
A-56
A-56
A-57
A-58
A-59
A-60
A-60
A-61
B-1
January 2010
-------
Table C-1. Studies of CO exposure and cardiovascular morbidity. C-1
Table C-2. Studies of CO exposure and cardiovascular hospital admissions and ED visits. C-17
Table C-3. Studies of CO exposure and neonatal and postneonatal outcomes. C-26
Table C-4. Studies of short-term CO exposure and respiratory morbidity C-35
Table C-5. Studies of short-term CO exposure and respiratory hospital admissions and ED visits. C-42
Table C-6. Studies of long-term CO exposure and respiratory morbidity. C-65
Table C-7. Studies of short-term CO exposure and mortality. C-70
Table C-8. Studies of long-term CO exposure and mortality. C-94
Table D-1. Controlled human exposure studies. D-1
Table E-1. Human and animal studies. E-1
January 2010
-------
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 3-11.
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.
Figure 3-21.
Figure 3-22.
Identification of studies for inclusion in the ISA.
Excess risk estimates from epidemiologic studies of short-term CO exposure and CVD
hospitalizations along with author-reported mean and AQS-derived 99th percentile CO
concentrations.
CO emissions (tons) in the U.S. bv source sector in 2002 from the NEI and the BEIS.
Trends in anthropogenic CO emissions (MT) in the U.S. by source category for 1990 and
1996-2002.
Surface air CO concentrations at Cheboque Point durinq the ICARTT campaiqn.
CO concentrations centered at ~3,000 m above sea level measured by the MOPITT sensor
on the Terra satellite for the period July 15-23, 2004, during intense wildfires in Alaska and
the Yukon.
Trends in subnational CO emissions in the 10 U.S. EPA Reqions for 1990 and 1996-2002.
CO emissions density map and distributions for the state of Colorado and for selected
counties in Colorado in 2002, from the NEI and the BEIS.
Components of RF in 2005 resulting from emissions since 1 750. (S) and (T) indicate
stratospheric and tropospheric changes, respectively.
Inteqrated RF of year 2000 emissions over 20-yr and 100-vrtime horizons.
Scatterplot comparinq data from co-located monitors in Charlotte, NC.
Map of CO monitor locations in the U.S. in 2007.
Map of CO monitor locations with respect to population density in the Denver, CO CSA,
total population.
Map of CO monitor locations with respect to population density in the Denver, CO CSA,
aqe 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, aqe 65 and older.
County-level map of second-hiqhest 1-h avq CO concentrations in the U.S. in 2007.
County-level map of second-hiqhest 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 illustratinq the distribution of 2005-2007 hourly CO concentrations in Denver, CO.
Intersampler correlation versus distance for monitors located within the Denver CSA.
Map of CO monitor locations and maior hiqhways for Los Anqeles, CA.
Box plots illustrating the distribution of 2005-2007 hourly CO concentrations in Los
Angeles, CA.
1-5
2-15
3-3
3-4
3-5
3-7
3-7
3-8
3-14
3-15
3-19
3-22
3-23
3-24
3-25
3-26
3-29
3-30
3-36
3-38
3-39
3-41
3-42
3-43
January 2010 xii
-------
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.
Figure 3-45.
Figure 3-46.
Intersampler correlation versus distance for monitors located within the Los Anqeles CSA.
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.
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. Scale: 1 cm = 145 m.
Aerial view of the location of CO monitor Q (marked by the red pin) in Pasadena, CA (Los
Anqeles CSA), depictinq its proximity to a residential neiqhborhood. Scale: 1 cm = 145 m.
Distribution of hourly CO concentration data by city and monitorinq scale.
Distribution of 1-h daily max CO concentration data by city and monitorinq scale.
Relative concentrations of CO and copollutants at various distances from the 1-71 0 freeway
in Los Anqeles.
CO concentration time series 20 m and 300 m from the I-440 hiqhway in Raleiqh, NC.
CO concentration profile 10m 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., 1980-2006, reported as the annual second 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 8-h
concentrations (ppm) for the EPA Regions 1 through 10, along with a depiction of the
qeoqraphic extent of those Reqions.
Diel plot generated from weekday hourly CO data (ppm) for the 1 1 CSAs and CBSAs,
2005-2007.
Diel plot generated from weekend hourly CO data (ppm) for the 1 1 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 Anqeles, 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 PRB concentrations.
Monthly (circles) and annual (squares) averaqe CO concentrations (ppb), 2005-2007.
Distribution of time that the sample population spends in various environments, from the
NHAPS.
Hourly personal versus ambient CO concentrations obtained in Baltimore, MD,
Box plots of the ratio of personal to ambient concentrations obtained in Baltimore, MD,
3-48
3-50
3-51
3-52
3-55
3-56
3-57
3-58
3-59
3-60
3-61
3-63
3-64
3-66
3-67
3-68
3-69
3-70
3-71
3-73
3-74
3-79
3-84
3-85
January 2010 xiii
-------
Figure 3-47. 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
Figure 4-1 .
Figure 4-2.
Figure 4-3.
Figure 4-4.
Figure 4-5.
Figure 4-6.
Figure 4-7.
Figure 4-8.
Figure 4-9.
Figure 4-10.
Figure 4-11.
Figure 4-1 2.
Figure 5-1 .
Figure 5-2.
Figure 5-3.
Figure 5-4.
Figure 5-5.
Figure 5-6.
with windows closed and air conditioner in fresh air mode.
Plot of fractional sensitivities of selected variables versus time of exposure.
Simulated COHb formation for two 5-dav workweeks.
Overall structure of the Bruce and Bruce (2008, 193977) multicompartment model of
storaqe and transport of CO.
Predicted COHb levels in healthy commuters exposed to various CO concentrations over a
60-min commute twice a day.
Predicted COHb levels due to various endoqenous CO production rates.
Diaqrammatic presentation of CO uptake and elimination pathways and CO body stores.
02Hb dissociation curve of normal human blood, of blood containing 50% COHb, and of
blood with only 50% Hb because of anemia.
Changes in blood COHb after exposure to CO for a few minutes (A) or several hours (B),
representing the biphasic nature of CO elimination.
Predicted maternal and fetal COHb during periodic exposure to CO (50 ppm for 16 h
followed by 0 ppm for 8 h ).
Representative estimates of endogenous CO production rates resulting from various
conditions and diseases.
Representative COHb saturation resulting from various diseases and conditions.
Representative exhaled CO concentrations (ppm) resulting from various conditions plotted
as fold increases over healthy human controls from each study.
Direct effects of CO.
Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for various forms of CHD.
Summary of effect estimates (95% confidence intervals) associated with ED visits and
hospital admissions for stroke.
Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CHF.
Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for CVD.
Effect estimates from studies of ED visits and hospital admissions for CVD outcomes other
3-89
4-4
4-6
4-7
4-10
4-11
4-12
4-15
4-18
4-21
4-23
4-24
4-25
5-12
5-27
5-30
5-32
5-36
than stroke from single pollutant (CO only: black circles) and particulate copollutant (CO +
PM2.5: red triangles; CO + PM10 or TSP: purple triangles) models. 5-38
Figure 5-7. Effect estimates from studies of ED visits and HAs for CVD outcomes other than stroke
from single pollutant (CO only: black circles) and gaseous copollutant models (CO + N02,
S02 and 03= green, blue, and orange triangles, respectively). 5-39
Figure 5-8. Regression of the percent change in time to ST endpoint between the pre- and
postexposure exercise tests ([postexposure-pre-exposure]/pre-exposure) and the
measured blood COHb levels at the end of exercise for the 63 subjects combined. 5-42
Figure 5-9. Summary of effect estimates (95% confidence intervals) for PTB associated with maternal
exposure to ambient CO. 5-52
January 2010 xiv
-------
Figure 5-10.
Figure 5-11.
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 5-1 9.
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.
Summary of change in birth weight (95% confidence intervals) associated with maternal
exposure to ambient CO.
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% confidence intervals) on pulmonary function due to a 10th to 90th
percentile increment chanqe in pollutant concentration in sinqle-pollutant models.
Summary of associations for short-term exposure to CO and asthma symptoms, respiratory
symptoms and medication use in asthmatic individuals.
Summary of associations for short-term exposure to CO and respiratory hospital
admissions:
Summary of associations for short-term exposure to CO and respiratory ED visits.
Posterior means and 95% posterior intervals of national average estimates for CO effects
on total (nonaccidental) mortality at lags 0, 1, and 2 within sets of the 90 U.S. cities with
available pollutant data.
Summary of percent increase in total (nonaccidental) mortality 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 (Figure 1 of 2).
CO emissions distribution for selected counties in Georgia (Figure 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,
ages 65 yr 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
yr and older.
Map of CO monitor locations with respect to population density in the Boston CSA, total
population.
5-57
5-57
5-58
5-84
5-87
5-93
5-95
5-103
5-106
5-113
A-1
A-2
A-3
A-4
A-5
A-6
A-7
A-20
A-20
A-21
A-21
A-22
January 2010 xv
-------
Figure A-1 3.
Figure A-1 4.
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 Boston CSA, ages 65
vr and older.
Map of CO monitor locations with respect to population density in the Houston CSA, total
population.
Map of CO monitor locations with respect to population density in the Houston CSA, ages
65 yr 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 yr 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 vr 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 vr 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
vr 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 vr 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.
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.
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.
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
Citv, NY.
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.
Map of CO monitor locations with AQS Site IDs for Pittsburgh, PA.
A-22
A-23
A-23
A-24
A-24
A-25
A-25
A-26
A-26
A-27
A-27
A-28
A-28
A-29
A-31
A-32
A-34
A-35
A-37
A-38
A-40
A-41
A-43
A-44
A-46
A-47
January 2010 xvi
-------
Figure A-39. Box plots illustrating the seasonal distribution of hourly CO concentrations in Pittsburgh,
PA. A-49
Figure A-40. Map of CO monitor locations with AQS Site IDs for Seattle, WA. A-50
Figure A-41. Box plots illustrating the seasonal distribution of hourly CO concentrations in Seattle, WA. A-51
Figure A-42. Map of CO monitor locations with AQS Site IDs for St. Louis, MO. A-52
Figure A-43. Box plots illustrating the seasonal distribution of hourly CO concentrations in St. Louis, MO. A-54
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-62
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-63
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-64
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-65
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-66
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-67
January 2010 xvii
-------
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
January 2010 xviii
-------
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—Senior Environmental Employment Program, 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—Senior Environmental Employment Program, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
January 2010 xix
-------
Authors, Contributors, and Reviewers
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
January 2010 xx
-------
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. Antonio Fernandez—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
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
Ms. Kathryn Sargeant—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
Mr. Mark Schmidt—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph H. Somers—Office of Transportation and Air Quality, Office of Air and Radiation,
U.S. Environmental Protection Agency, Ann Arbor, MI
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
January 2010 xxi
-------
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
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. Deirdre Murphy—Office of Air Quality Planning and Standards, Office of Air and Radiation,
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 H. 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
January 2010 xxii
-------
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
January 2010 xxiii
-------
Clean Air Scientific Advisory Committee
CO NAAQS Review Panel
Chair of the Environmental Protection Agency's Clean Air Scientific Advisory Committee
Dr. Jonathan M. Samet*, Department of Preventive Medicine at the Keck School of Medicine, and
Director of the Institute for Global Health at the University of Southern California, Los Angeles, CA
Chair of the Carbon Monoxide Review Panel
Dr. Joseph Brain*, Department of Environmental Health, Harvard School of Public Health, Harvard
University, Boston, MA
Members
Dr. Paul Blanc, Department of Occupational Medicine, University of California-San Francisco, San
Francisco, CA
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. Joel Kaufman, Department of Environmental & Occupational Health Sciences, University of
Washington, Seattle, WA
Dr. Michael T. Kleinman, Department of Community & Environmental Medicine, University of
California-Irvine, Irvine, CA
Dr. Francine Laden, Department of Environmental Health, Harvard School of Public Health,
Harvard University, Boston, MA
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
January 2010 xxiv
-------
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering, Georgia Institute
of Technology, Atlanta, GA
Dr. Anne Sweeney, Department of Epidemiology & Biostatistics, School of Rural Public Health,
Texas A&M Health Science Center, College Station, TX
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
Science Advisory Board Staff
Ms. Kyndall Barry, Designated Federal Officer, 1200 Pennsylvania Avenue, N.W., Washington, DC,
20460, Phone: 202-343-9868, Fax: 202-233-0643, Email:barry.kyndall@epa.gov
Physical/Courier/FedEx Address:
Ms. Kyndall Barry, U.S. EPA Science Advisory Board Staff Office, Mail Code 1400F, Ariel Rios
Building, Room 3610A, 1025 F Street, N.W., Washington, DC 20004
January 2010 xxv
-------
Acronyms and Abbreviations
a alpha, ambient exposure factor
a air exchange rate of the microenvironment
AA abdominal aorta(s)
AADT annual average daily traffic
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
January 2010
XXVI
-------
P beta, beta coefficient, slope
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
January 2010
XXVII
-------
CH2O2 formic acid
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
January 2010
XXVIII
-------
DmCO
DMT-1
DMV
DNA
DOCA
dP/dtLV
dP/dtRV
DSA
E
Ea
EC
ED
EKG, EGG
p
J-^na
eNOS
EPA
EPO
EPR
EPPJ
ESRL
ET-1
ETA
ETS
EXPOLIS
FAS
FC
FEF
FEF25.75
FEM
FEVj
f;
FjCO
Fmf
capacity for diffusion of CO into the muscle
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
January 2010
XXIX
-------
fo
FR
FOR
FRM
FSH
FVC
FVII
FW
GAM
GD
GEE
GEM
GFAP
GFC
GLM
GLMM
GMD
GMP
GSH
GSSG
GTP
GWP(s)
H
h
H2O2
H9c2
Hb
HC(s)
HCFC(s)
HCO
HEAPSS
HEK293
HepSB
HF
HFLFR
fraction of time spent outdoors
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 (experimentally transformed cell line)
Human hepatocarcinoma cell line
heart failure, high frequency (HRV parameter)
high frequency to low frequency ratio (HRV parameter)
January 2010
xxx
-------
HH
HIF-la
HO
H02
HO-1
HO-2
HO/CO
HR
H/R
HRV
HS
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
hypobaric hypoxia
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
hemorrhagic stroke
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; Michaelis-Menten equation of enzyme kinetics
Dissociation rate of oxygen from hemoglobin
January 2010
XXXI
-------
LEW
LCA+
LD
LDH
LDL
LF
LH
LOAEL
LOD
LOESS
EPS
LTP
EUR
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
low birth weight (<2,500 grams, (=51bs, 8 oz))
leucocyte common antigen cells
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 the CO chemical affinity for Hb [or
Mb]), Reaction mediator.
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
January 2010
XXXII
-------
MONICA
MOPITT
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
Monitoring of Trends and Determinants in Cardiovascular Disease
Measurement of Pollution in the Troposphere
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
January 2010
XXXIII
-------
NO'-Mb
NOX
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
PC02
PDGF
PEE
nitrosyl bound Mb
nitrogen oxides, oxides of nitrogen
National Research Council
nucleus of the solitary tract (in brainstem)
ozone
oxyhemoglobin (% concentration in mL O2 / mL blood)
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 (CH3CO3NO2)
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
January 2010
xxxiv
-------
PEF
PEFD(s)
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
PvO2
PV02
Q
QCP
peak expiratory flow
Personal Exposures Frequency Distributions
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 PM10 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
January 2010
xxxv
-------
<2OT blood flow to other tissues
RA radial artery of the heart
RAW 264.7 mouse macrophage cell line
RBC red blood cell
RF radiative forcing
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
January 2010
xxxvi
-------
SIDS sudden infant death syndrome
SIPs State Implementation Plan(s)
siRNA small inhibitory RNA
SLAMS State and Local Air Monitoring Stations
SMC smooth muscle cell(s)
SnMP tin-(IV)-mesoporphyrin
SNP single-nucleotide polymorphism
SnPP-IX tin protoporphyrin IX
SO sphingosine
SO2 sulfur dioxide
SO42" sulfate
SOD superoxide dismutase
SOPHIA Study of Particles and Health in Atlanta
STEMS Space-Time Exposure Modeling System
STN Speciation Trends Network
STPD standard temperature and pressure, dry
SV stroke volume
SVEB supraventricular (atrium or atrioventricular node) ectopic beats
T tau, photochemical lifetime
T lymphocytes thymus-dependent lymphocytes
TEARS thiobarbituric acid reactive substances
TC total carbon
TFAM mitochondrial transcription factor A
Tg teragram(s)
TH tyrosine hydroxylase
THP-1 human monocyte-derived cell line, (can differentiate into
macrophages)
TIA transient ischemic attack
TNF-a tissue necrosis factor alpha
TPM total particulate matter
TSP total suspended particles
UFP ultrafine particle(s)
ULTRA Exposure and Risk Assessment for Fine and Ultrafine Particles in
Ambient Air (Study)
URI upper respiratory infection
URTI upper respiratory tract infection
January 2010
XXXVII
-------
use
VA
Vb
Vco
VD
VE
VEGF
VLF
V
v max
VO2 max
VOC(s)
VPB
vWF
W
WBC
WHI
WKY
ZnPPIX
U.S. Code
alveolar ventilation
blood volume
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 (maximal oxygen
consumption, maximal oxygen uptake or aerobic capacity)
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
January 2010
XXXVIII
-------
Chapter 1. Introduction
The Integrated Science Assessment (ISA) is a concise evaluation and synthesis of the most
policy-relevant science for reviewing the national ambient air quality standards (NAAQS). Because
the ISA communicates critical science judgments relevant to the NAAQS review, it forms the
scientific foundation for the review of the NAAQS for carbon monoxide (CO). The existing primary
CO standards include a 1-hour (h) average (avg) standard set at 35 parts per million (ppm), and an
8-h avg standard set at 9 ppm, neither to be exceeded more than once per year. There is currently no
secondary standard for CO.
The ISA accurately reflects "the latest scientific knowledge useful in indicating the kind and
extent of identifiable effects on public health which may be expected from the presence of [a]
pollutant in ambient air" (42 U.S.C. 7408). Key information and judgments formerly contained in
the Air Quality Criteria Document (AQCD) for CO are incorporated in this assessment. Additional
details of the pertinent scientific literature published since the last review, as well as selected older
studies of particular interest, are included in a series of annexes. This ISA thus serves to update and
revise the evaluation of the scientific evidence available at the time of the previous review of the
NAAQS for CO that was completed in 2000.
The integrated Plan for Review of the National Ambient Air Quality Standards for Carbon
Monoxide (U.S. EPA, 2008, 193995) identifies key policy-relevant questions that provide a
framework for this assessment of the scientific evidence. These questions frame the entire review of
the NAAQS for CO and thus are informed by both science and policy considerations. The ISA
organizes, presents, and integrates the scientific evidence which is considered along with findings
from risk analyses and policy considerations to help the U.S. Environmental Protection Agency
(EPA) address these questions during the NAAQS review. In evaluating the health evidence, the
focus of this assessment is on scientific evidence that is most relevant to the following questions
taken directly from the Integrated Review Plan:
• Has new information altered the scientific support for the occurrence of health effects
following short- and/or long-term exposure to levels of CO found in the ambient air?
• To what extent is key evidence becoming available that could inform our understanding
of human subpopulations that are particularly sensitive to CO exposures? Is there new or
emerging evidence on health effects beyond cardiovascular and respiratory endpoints
(e.g., systemic effects, developmental effects, birth outcomes) that suggest additional
sensitive subpopulations should be given increased focus in this review (e.g., neonates)?
• What do recent studies focused on the near-roadway environment, including bus stops
and intersections, tell us about high-exposure human subpopulations and the health
effects of CO? What information is available on elevated exposures due to other
transportation sources, such as shipping, port operations, and recreational vehicles? What
is the effect of altitude on CO sources and health effects?
• At what levels of CO exposure do health effects of concern occur?
• To what extent is key scientific evidence becoming available to improve our
understanding of the health effects associated with various time periods of CO exposures,
including not only daily but also chronic (months to years) exposures? To what extent is
critical research becoming available that could improve our understanding of the
relationship between various health endpoints and different lag periods (e.g., single-day,
multiday distributed lags)?
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).
January 2010 1-1
-------
To what extent does the evidence suggest that alternate dose indicators other than
carboxyhemoglobin (COHb) levels (e.g., tissue oxygenation) should be evaluated to
characterize the biological effect?
Has new information altered conclusions from previous reviews regarding the
plausibility of adverse health effects caused by CO exposure?
To what extent have important uncertainties identified in the last review been reduced
and/or have new uncertainties emerged?
Have new information or scientific insights altered the scientific conclusions regarding
the occurrence of direct (or indirect) welfare effects associated with levels of CO found
in the ambient air?
1.1. Legislative Requirements
Two sections of the Clean Air Act (CAA, the Act) govern the establishment and revision of the
NAAQS. Section 108 of the Act (42 U.S.C. 7408) directs the Administrator to identify and list "air
pollutants" that "in [her] judgment, may reasonably be anticipated to endanger public health and
welfare" and whose "presence ... in the ambient air results from numerous or diverse mobile or
stationary sources" and to issue air quality criteria for those that are listed (42 U.S.C. 7408). Air
quality criteria are intended to "accurately reflect the latest scientific knowledge useful in indicating
the kind and extent of identifiable effects on public health or welfare which may be expected from
the presence of [a] pollutant in ambient air..." 42 U.S.C. 7408(b).
Section 109 of the Act (42 U.S.C. 7409) directs the EPA Administrator to propose and
promulgate "primary" and "secondary" National Ambient Air Quality Standards (NAAQS) for
pollutants listed under Section 108. Section 109(b)(l) defines a primary standard as one "the
attainment and maintenance of which in the judgment of the Administrator, based on such criteria
and allowing an adequate margin of safety, are requisite to protect the public health."1 A secondary
standard, as defined in Section 109(b)(2), must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the U.S. EPA Administrator, based on such criteria, is
required to protect the public welfare from any known or anticipated adverse effects associated with
the presence of [the] pollutant in the ambient air."2 The requirement that primary standards include
an adequate margin of safety was intended to address uncertainties associated with inconclusive
scientific and technical information available at the time of standard setting. It was also intended to
provide a reasonable degree of protection against hazards that research has not yet identified. See
Lead Industries Association v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied, 449 U.S.
1042 (\9%Q); American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981) cert.
denied, 455 U.S. 1034 (1982). The aforementioned uncertainties are components of the risk
associated with pollution at levels below those at which human health effects can be said to occur
with reasonable scientific certainty. Thus, in selecting primary standards that include an adequate
margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
demonstrated to be harmful, but also to prevent lower pollutant levels that may pose an unacceptable
risk of harm, even if the risk is not precisely identified as to nature or degree.
In selecting a margin of safety, the EPA considers such factors as the nature and severity of the
health effects involved, the size of susceptible population(s), and the kind and degree of the
uncertainties that must be addressed. The selection of any particular approach to providing an
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."
January 2010 1-2
-------
adequate margin of safety is a policy choice left specifically to the Administrator's judgment. See
Lead Industries Association v. EPA, supra, 647 F.2d at 1161-62.
In setting standards that are "requisite" to protect public health and welfare, as provided in
Section 109(b), EPA's task is to establish standards that are neither more nor less stringent than
necessary for these purposes. In so doing, EPA may not consider the costs of implementing the
standards. See Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-76 (D.C.
Cir. 2001).
Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-year intervals
thereafter, the Administrator shall complete a thorough review of the criteria published under Section
108 and the national ambient air quality standards... and shall make such revisions in such criteria
and standards and promulgate such new standards as may be appropriate..." Section 109(d)(2)
requires that an independent scientific review committee "shall complete a review of the
criteria... and the national primary and secondary ambient air quality standards... and shall
recommend to the Administrator any new... standards and revisions of existing criteria and standards
as may be appropriate..." Since the early 1980s, this independent review function has been
performed by the Clean Air Scientific Advisory Committee (CASAC) of EPA's Science Advisory
Board (SAB).
1.2. History of the NAAQS for CO
On April 30, 1971, EPA promulgated identical primary and secondary NAAQS for CO, under
Section 109 of the Clean Air Act, set at 9 ppm, 8-h avg and 35 ppm, 1-h avg, neither to be exceeded
more than once per year (36 FR 8186). In 1979, EPA published the Air Quality Criteria Document
for Carbon Monoxide (1979, 017687), which updated the scientific criteria upon which the initial
CO standards were based. A Staff Paper (U.S. EPA, 1979, 194665) was prepared and, along with the
AQCD (1979, 017687). served as the basis for development of proposed rulemaking (45 FR 55066)
published on August 18, 1980. Delays due to uncertainties regarding the scientific basis for the final
decision resulted in EPA announcing a second public comment period (47 FR 26407). Following
substantial reexamination of the scientific data, EPA prepared an Addendum to the 1979 AQCD
(1984, 012690) and an updated Staff Paper (1984, 012691). Following review by CASAC, EPA
announced its final decision (50 FR 37484) not to revise the existing primary standard and to revoke
the secondary standard for CO on September 13, 1985, due to a lack of evidence of direct effects on
public welfare at ambient concentrations.
In 1987, EPA initiated action to revise the criteria for CO and subsequently released a revised
AQCD (U.S. EPA, 1991, 017643) for CASAC and public review. In a "closure letter" (McClellan,
1991, 194666) sent to the Administrator, the CASAC concluded that the AQCD (U.S. EPA, 1991,
017643) ". . . provides a scientifically balanced and defensible summary of current knowledge of the
effects of this pollutant and provides an adequate basis for the EPA to make a decision as to the
appropriate primary NAAQS for CO." A revised Staff Paper subsequently was reviewed by CASAC
and the public, and in a "closure letter" (McClellan, 1992, 194667) sent to the Administrator,
CASAC stated ". . . that a standard of the present form and with a numerical value similar to that of
the present standard would be supported by the present scientific data on health effects of exposure
to carbon monoxide." Based on the revised AQCD (U.S. EPA, 1991, 017643) and staff conclusions
and recommendations contained in the revised Staff Paper (U.S. EPA, 1992, 084191). the
Administrator announced the final decision (59 FR 38906) on August 1, 1994, that revision of the
primary NAAQS for CO was not appropriate at that time.
In 1997, revisions to the 1991 AQCD (U.S. EPA, 1991, 017643) were initiated. A workshop
was held in September 1998 to review and discuss material contained in the revised draft AQCD. On
June 9, 1999, CASAC held a public meeting to review the draft AQCD and a draft exposure analysis
methodology document. Comments from CASAC and the public were considered in a second draft
AQCD, which was reviewed at a CASAC meeting, held on November 18, 1999. After revision of the
second draft AQCD, the final AQCD (U.S. EPA, 2000, 000907) was released in August 2000. EPA
put the review on hold when Congress called on the National Research Council (NRC) to conduct a
review of the impact of meteorology and topography on ambient CO concentrations in high altitude
and extreme cold regions of the U.S. In response, the NRC convened the committee on Carbon
Monoxide Episodes in Meteorological and Topographical Problem Areas, which focused on
January 2010 1-3
-------
Fairbanks, Alaska as a case study in an interim report, which was completed in 2002. A final report,
Managing Carbon Monoxide Pollution in Meteorological and Topographical Problem Areas, was
published in 2003 (National Research Council, 2003, 042550) and offered a wide range of
recommendations on management of CO air pollution, cold start emissions standards, oxygenated
fuels, and CO monitoring. EPA did not complete the NAAQS review which started in 1997.
1.3. ISA Development
EPA initiated the current review of the NAAQS for CO on September 13, 2007 with a call for
information from the public (72 FR 52369). In addition to the call for information, publications were
identified through an ongoing literature search process that includes extensive computer database
mining on specific topics. Literature searches were conducted routinely to identify studies published
since the last review, focusing on publications from 1999 to May 2009. Search strategies were
iteratively modified to optimize identification of pertinent publications. Additional papers were
identified for inclusion in several ways: review of pre-publication tables of contents for journals in
which relevant papers may be published; independent identification of relevant literature by expert
authors; and identification by the public and CASAC during the external review process.
Publications considered for inclusion in the ISA were added to the Health and Environmental
Research Online (HERO) database recently developed by EPA (http://cfpub.epa.gov/ncea/hero/);
note that all references in the ISA include a HERO ID that provides a link to the database. Typically,
only information that had undergone scientific peer review and had been published or accepted for
publication was considered, along with analyses conducted by EPA using publicly available data.
This review has attempted to evaluate all relevant data published since the last review pertaining to
the atmospheric science of CO, human exposure to ambient CO, and epidemiologic, controlled
human exposure, and animal toxicological studies on CO, including those related to exposure-
response relationships, mode(s) of action (MOA), or susceptible populations. Added to the body of
research on CO effects were EPA's analyses of air quality and emissions data, studies on atmospheric
chemistry, transport, and fate of these emissions, as well as issues related to exposure to CO. An
extensive literature search for data on the ecological effects of ambient CO did not identify any
relevant information published since the review of the ecological effects evidence in the 1979 CO
AQCD (U.S. EPA, 1979, 017687).
In general, in assessing the scientific quality and relevance of health and environmental effects
studies, the following considerations have been taken into account when selecting studies for
inclusion in the ISA or its annexes. The selection process for studies included in this ISA is shown in
Figure 1-1.
• Are the study populations, subjects, or animal models adequately selected and are they
sufficiently well defined to allow for meaningful comparisons between study or exposure
groups?
• Are the statistical analyses appropriate, properly performed, and properly interpreted?
Are likely covariates adequately controlled or taken into account in the study design and
statistical analysis?
• Are the air quality data, exposure, or dose metrics of adequate quality and sufficiently
representative of information regarding ambient CO?
• Are the health or welfare effect measurements meaningful and reliable?
January 2010 1-4
-------
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.
V
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.
In selecting epidemiologic studies, EPA considered whether a given study presented
information on associations with short- or long-term CO exposures at or near ambient levels of CO;
considered approaches to evaluate issues related to potential confounding by other pollutants;
assessed potential effect modifiers; addressed health endpoints and populations not previously
extensively researched; and evaluated important methodologic issues (e.g., lag or time period
between exposure and effects, model specifications, thresholds, mortality displacement) related to
interpretation of the health evidence. Among the epidemiologic studies selected, particular emphasis
was placed on those studies most relevant to the review of the NAAQS. Specifically, studies
conducted in the United States (U.S.) or Canada were discussed in more detail than those from other
geographical regions. Particular emphasis was placed on: (1) recent multicity studies that employ
standardized analysis methods for evaluating effects of CO and that provide overall estimates for
effects based on combined analyses of information pooled across multiple cities; (2) studies that help
understand quantitative relationships between exposure concentrations and effects; (3) new studies
that provide evidence on effects in susceptible populations; and (4) studies that consider and report
CO as a component of a complex mixture of air pollutants.
January 2010
1-5
-------
Criteria for the selection of research evaluating controlled human exposure or animal
toxicological studies included a focus on studies conducted using relevant pollutant exposures. For
both types of studies, relevant pollutant exposures are considered to be those generally within one or
two orders of magnitude of ambient CO concentrations. Studies in which higher doses were used
may also be considered if they provide information relevant to understanding MO As or mechanisms,
as noted below.
Evaluation of controlled human exposure studies focused on those that approximated expected
human exposure conditions in terms of concentration and duration. In the selection of controlled
human exposure studies, emphasis is placed on studies that (1) investigate potentially susceptible
populations such as people with cardiovascular diseases; (2) address issues such as concentration-
response or time-course of responses; (3) include control exposures to filtered air; and (4) have
sufficient statistical power to assess findings.
Review of the animal toxicological evidence focused on studies that approximate expected
human dose conditions, which will vary depending on the toxicokinetics and biological sensitivity of
the particular laboratory animal species or strains studied. Due to resource constraints on exposure
duration and numbers of animals tested, animal studies typically utilize high-concentration
exposures to acquire data relating to mechanisms and assure a measureable response. Such studies
were considered to the extent that they provided useful information to inform our understanding of
interspecies differences and potential sensitivity differences between healthy and susceptible human
populations.
These criteria provide benchmarks for evaluating various studies and for focusing on the
policy-relevant studies in assessing the body of health and welfare effects evidence. Detailed critical
analysis of all CO health and welfare effects studies, especially in relation to the above
considerations, is beyond the scope of this document. Of most relevance for evaluation of studies is
whether they provide useful qualitative or quantitative information on exposure-effect or
exposure-response relationships for effects associated with current ambient air concentrations of CO
that can inform decisions on whether to retain or revise the standards.
In developing the CO ISA, EPA began by reviewing and summarizing the evidence on
atmospheric sciences and exposure and the health effects evidence from in vivo and in vitro
toxicological studies, controlled human exposure studies, and epidemiologic studies. In November
2008, EPA invited EPA staff and other researchers with expertise in CO to a teleconference to review
the scientific content of preliminary draft materials for the draft ISA and the annexes. The purpose of
the initial peer review teleconference was to ensure that the ISA is up to date and focused on the
most policy-relevant findings, and to assist EPA with integration of evidence within and across
disciplines. Subsequently, EPA addressed comments and completed the initial integration and
synthesis of the evidence.
The integration of evidence on health or welfare effects involves collaboration between
scientists from various disciplines. As described in the section below, the ISA organization is based
on health effect categories. As an example, an evaluation of health effects evidence would include
summaries of findings from epidemiologic, controlled human exposure, and toxicological studies,
and integration of the results to draw conclusions based on the causal framework described below.
Using the causal framework described in Section 1.6, EPA scientists consider aspects such as
strength, consistency, coherence and biological plausibility of the evidence, and develop draft
causality judgments on the nature of the relationships. The draft integrative synthesis sections and
conclusions are reviewed by EPA internal experts and, as appropriate, by outside expert authors. In
practice, causality determinations often entail an iterative process of review and evaluation of the
evidence. The draft ISA is released for review by the CASAC and the public, and comments received
on the characterization of the science as well as the implementation of the causal framework are
carefully considered in revising and completing the ISA.
1.4. Document Organization
The ISA is composed of five chapters. This introductory chapter presents background
information and provides an overview of EPA's framework for making causal judgments. Chapter 2
is an integrated summary of key findings and conclusions regarding the source to dose paradigm,
MOA, and important health effects of CO, including cardiovascular, nervous system,
January 2010 1-6
-------
perinatal/developmental, respiratory, and mortality outcomes. Chapter 3 highlights key concepts and
evidence relevant to understanding the sources, ambient concentrations, atmospheric behavior, and
exposure to ambient CO. Chapter 4 describes the dosimetry and pharmacokinetics of CO, including
formation and fate of carboxy hemoglobin (COHb). Chapter 5 presents a discussion of the MO A of
CO and evaluates and integrates epidemiologic, human clinical, and animal toxicological
information on health effects related to short-term exposures (i.e., hours, days, or weeks) and long-
term exposures (i.e., months or years) to CO, including cardiovascular and systemic effects, central
nervous system (CNS) effects, birth outcomes and developmental effects, respiratory effects, and
mortality.
A series of annexes supplement this ISA. The annexes provide tables summarizing additional
details of the pertinent literature published since the last review, as well as selected older studies of
particular interest. These annexes contain information on:
• atmospheric chemistry of CO, sampling and analytic methods for measurement of CO
concentrations, emissions, sources and human exposure to CO (Annex A);
• studies on the dosimetry and pharmacokinetics of CO (Annex B);
• epidemiologic studies of health effects from short- and long-term exposure to CO
(Annex C);
• controlled human exposure studies of health effects related to exposure to CO (Annex
D);and
• toxicological studies of health effects in laboratory animals (Annex E)
Within Annexes B through E, detailed information about methods and results of health studies
is summarized in tabular format, and generally includes information about concentrations of CO and
averaging times, study methods employed, results and comments, and quantitative results for
relationships between effects and exposure to CO. As noted in the section above, the most pertinent
results of this body of studies are brought into the ISA.
1.5. Document Scope
For the current review of the primary CO standards, relevant scientific information on human
exposures and health effects associated with exposure to ambient CO has been assessed. Health
effects resulting from accidental exposures to very high concentrations of non-ambient CO (i.e., CO
poisoning) are not directly relevant to ambient exposures, and as such, a discussion of these effects
has deliberately been excluded from this document. For a detailed review of the effects of high-level
exposures to CO, the reader is referred to the extensive body of literature related to CO poisoning
(Ernst and Zibrak, 1998, 049822: Penney, 2007, 194668: Raub et al, 2000, 002180). In addition,
results of studies investigating the relationship between blood COHb concentrations and health
effects (e.g., Hedblad et al., 2006, 199512) may be informative regarding the biological plausibility
of health effects associated with changes in COHb concentrations. However, the lack of data on
ambient concentrations and the likely contribution of non-ambient CO to COHb in these studies
complicates the interpretation of the results with respect to ambient CO exposure, and therefore these
studies will not be discussed in this review. The possible influence of other atmospheric pollutants on
the interpretation of the role of CO in health effects studies is considered in this assessment. This
includes other pollutants with the potential to co-occur in the environment (e.g., nitrogen dioxide
[NO2], sulfur dioxide [SO2], ozone [O3], and particulate matter [PM]).
The review also assesses relevant scientific information associated with known or anticipated
public welfare effects that may be identified. The 1979 CO AQCD (U.S. EPA, 1979, 017687)
reviewed research on the effects of CO on vegetation and soil microflora, which showed that visible
symptoms and effects on growth, yield, and reproduction were observed in some studies at very high
CO concentrations (1,000-10,000 ppm or greater), while biochemical and physiological responses,
including reduced nitrogen fixation, were observed at lower concentrations (1,000 ppm and below).
As discussed in Section 1.3, a critical review of the ecological effects literature identified no
January 2010 1-7
-------
information published since the 1979 CO AQCD (U.S. EPA, 1979, 017687) pertinent to ambient CO
exposures; hence, no section on ecological effects appears in this assessment. The reader is referred
to the 1979 CO AQCD (U.S. EPA, 1979, 017687) for a detailed discussion of the effects of high CO
concentrations on plants and microorganisms. The definition of public welfare for the NAAQS
includes considerations of climate. Thus, the climate forcing effects of CO are summarized in
Chapter 2 and are discussed in detail in Chapter 3, where distinctions are drawn between global-
scale conclusions related to climate and the strongly variable continental and regional climate
forcing effects from CO.
1.6. EPA Framework for Causal Determination
The EPA has developed a consistent and transparent basis to evaluate the causal nature of air
pollution-induced health or environmental effects. The framework described below establishes
uniform language concerning causality and brings more specificity to the findings. This standardized
language was drawn from across the federal government and wider scientific community, especially
from the recent National Academy of Sciences (NAS) Institute of Medicine (IOM) document,
Improving the Presumptive Disability Decision-Making Process for Veterans, (2008, 156586) the
most recent comprehensive work on evaluating causality.
This introductory section focuses on the evaluation of health effects evidence. While focusing
on human health outcomes, the concepts are also generally relevant to causality determination for
welfare effects. This section:
• describes the kinds of scientific evidence used in establishing a general causal
relationship between exposure and health effects;
• defines cause, in contrast to statistical association;
• discusses the sources of evidence necessary to reach a conclusion about the existence of
a causal relationship;
• highlights the issue of multifactorial causation;
• identifies issues and approaches related to uncertainty; and
• provides a framework for classifying and characterizing the weight of evidence in
support of a general causal relationship.
Approaches to assessing the separate and combined lines of evidence (e.g., epidemiologic,
human clinical, and animal toxicological studies) have been formulated by a number of regulatory
and science agencies, including the IOM of the NAS (2008, 156586). International Agency for
Research on Cancer (2006, 093206). EPA Guidelines for Carcinogen Risk Assessment (2005,
086237). Centers for Disease Control and Prevention (2004, 056384). and National Acid
Precipitation Assessment Program (1991, 095894). These formalized approaches offer guidance for
assessing causality. The frameworks are similar in nature, although adapted to different purposes,
and have proven effective in providing a uniform structure and language for causal determinations.
Moreover, these frameworks have supported decision-making under conditions of uncertainty.
1.6.1. Scientific Evidence Used in Establishing Causality
Causality determinations are based on the evaluation and synthesis of evidence from across
scientific disciplines; the type of evidence that is most important for such determinations will vary
by pollutant or assessment. The most compelling evidence of a causal relationship between pollutant
exposures and human health effects comes from human clinical studies. This type of study
experimentally evaluates the health effects of administered exposures in human volunteers under
highly controlled laboratory conditions.
January 2010 1-8
-------
In epidemiologic or observational studies of humans, the investigator does not control
exposures or intervene with the study population. Broadly, observational studies can describe
associations between exposures and effects. These studies fall into several categories:
cross-sectional, prospective cohort, and time-series studies. "Natural experiments" offer the
opportunity to investigate changes in health with a change in exposure; these include comparisons of
health effects before and after a change in population exposures, such as closure of a pollution
source.
Experimental animal data can help characterize effects of concern, exposure-response
relationships, susceptible populations and MOAs. In the absence of controlled human exposure or
epidemiologic data, animal data alone may be sufficient to support a likely causal determination,
assuming that humans respond similarly to the experimental species.
1.6.2. Association and Causation
"Cause" is a significant, effectual relationship between an agent and an effect on health or
public welfare. "Association" is the statistical dependence among events, characteristics, or other
variables. An association is prima facie evidence for causation; alone, however, it is insufficient
proof of a causal relationship between exposure and disease. Unlike an association, a causal claim
supports the creation of counterfactual claims; that is, a claim about what the world would have been
like under different or changed circumstances (IOM, 2008, 156586). Much of the newly available
health information evaluated in this ISA comes from epidemiologic studies that report a statistical
association between ambient exposure and health outcome.
Many of the health and environmental outcomes reported in these studies have complex
etiologies. Diseases such as asthma, coronary heart disease (CHD) or cancer are typically initiated
by multiple agents. Outcomes depend on a variety of factors, such as age, genetic susceptibility,
nutritional status, immune competence, and social factors (Gee and Payne-Sturges, 2004, 093070;
IOM, 2008, 156586). Effects on ecosystems are often also multifactorial with a complex web of
causation. Further, exposure to a combination of agents could cause synergistic or antagonistic
effects. Thus, the observed risk represents the net effect of many actions and counteractions.
1.6.3. Evaluating Evidence for Inferring Causation
Moving from association to causation involves the elimination of alternative explanations for
the association. In estimating the causal influence of an exposure on health or environmental effects,
it is recognized that scientific findings incorporate uncertainty. "Uncertainty" can be defined as a
state of having limited knowledge where it is impossible to exactly describe an existing state or
future outcome, e.g., the lack of knowledge about the correct value for a specific measure or
estimate. Uncertainty characterization and uncertainty assessment are two activities that lead to
different degrees of sophistication in describing uncertainty. Uncertainty characterization generally
involves a qualitative discussion of the thought processes that lead to the selection and rejection of
specific data, estimates, scenarios, etc. Uncertainty assessment is more quantitative. The process
begins with simpler measures (e.g., ranges) and simpler analytical techniques and progresses, to the
extent needed to support the decision for which the assessment is conducted, to more complex
measures and techniques. Data will not be available for all aspects of an assessment and those data
that are available may be of questionable or unknown quality. In these situations, evaluation of
uncertainty can include professional judgment or inferences based on analogy with similar situations.
The net result is that the assessment will be based on a number of assumptions with varying degrees
of uncertainty. Uncertainties commonly encountered in evaluating health evidence for the criteria air
pollutants are outlined below for epidemiologic and experimental studies. Various approaches to
evaluating uncertainty include classical statistical methods, sensitivity analysis, or probabilistic
uncertainty analysis, in order of increasing complexity and data requirements. The ISA generally
evaluates uncertainties qualitatively in assessing the evidence from across studies; in some situations
quantitative analysis approaches, such as metaregression, may be used.
Meta-analysis may be a valuable tool for evaluating evidence by combining results from a
body of studies. Blair et al. (1995, 079190) observe that meta-analysis can enhance understanding of
associations between exposures and effects that are not readily apparent in examination of individual
study results and can be particularly useful for formally examining sources of heterogeneity.
January 2010 1-9
-------
However, these authors note that meta-analysis may not be useful when the relationship between the
exposure and outcome is obvious, when only a few studies are available for a particular exposure-
outcome relationship, where there is limited access to data of sufficient quality, or where there is
substantial variation in study design or population. In addition, important differences in effect
estimates, exposure metrics, or other factors may limit or even preclude quantitative statistical
combination of multiple studies.
Controlled human exposure studies evaluate the effects of exposures to a variety of pollutants
in a highly controlled laboratory setting. Also referred to as human clinical studies, these
experiments allow investigators to expose subjects to known concentrations of air pollutants under
carefully regulated environmental conditions and activity levels. In some instances, controlled
human exposure studies can also be used to characterize concentration-response relationships at
pollutant concentrations relevant to ambient conditions. Controlled human exposures are typically
conducted using a randomized crossover design, with subjects exposed both to CO and a clean air
control. In this way, subjects serve as their own controls, effectively controlling for many potential
confounders. However, human clinical studies are limited by a number of factors, including a small
sample size and short exposure times. The repetitive nature of ambient CO exposures at levels that
can vary widely may lead to cumulative health effects, but this type of exposure is not practical to
replicate in a laboratory setting. In addition, although subjects do serve as their own controls,
personal exposure to pollutants in the hours and days preceding the controlled exposures may vary
significantly between and within individuals. Endogenous production of CO creates a body burden
of CO that, together with personal exposure from nonambient sources, contributes to baseline COHb
levels. Endogenous production rates vary within and among individuals, particularly for individuals
with diseases such as hemolytic anemia or chronic inflammation. This body burden of CO and
COHb limits the lower range of exposures that can be practically covered in controlled human
exposure studies. Finally, human clinical studies require investigators to adhere to stringent health
criteria for a subject to be included in the study, and therefore the results cannot necessarily be
generalized to an entire population. Although some human clinical studies have included health-
compromised individuals such as those with coronary artery disease (CAD), these individuals must
also be relatively healthy and do not represent the most sensitive individuals in the population. Thus,
a lack of observation of effects from human clinical studies does not necessarily mean that a causal
relationship does not exist. While human clinical studies provide important information on the
biological plausibility of associations observed between air pollutant exposure and health outcomes
in epidemiologic studies, observed effects in these studies may underestimate the response in certain
populations.
Epidemiologic studies provide important information on the associations between health
effects and exposure of human populations to ambient air pollution. In the evaluation of
epidemiologic evidence, one important consideration is potential confounding. Confounding is ". . .
a confusion of effects. Specifically, the apparent effect of the exposure of interest is distorted because
the effect of an extraneous factor is mistaken for or mixed with the actual exposure effect (which
may be null)" (Rothman and Greenland, 1998, 086599). One approach to remove spurious
associations due to possible confounders is to control for characteristics that may differ between
exposed and unexposed persons; this is frequently termed "adjustment." Scientific judgment is
needed regarding likely sources and magnitude of confounding, together with consideration of how
well the existing constellation of study designs, results, and analyses address this potential threat to
inferential validity. One key consideration in this review is evaluation of the potential contribution of
CO to health effects when it is a component of a complex air pollutant mixture. Reported CO effect
estimates in epidemiologic studies may reflect independent CO effects on health outcomes. Ambient
CO may also be serving as an indicator of complex ambient air pollution mixtures that share the
same source as CO (e.g., motor vehicle emissions). Alternatively, copollutants may mediate the
effects of CO or CO may influence the toxicity of copollutants.
Another important consideration in the evaluation of epidemiologic evidence is effect
modification. "Effect-measure modification differs from confounding in several ways. The main
difference is that, whereas confounding is a bias that the investigator hopes to prevent or remove
from the effect estimate, effect-measure modification is a property of the effect under study ... In
epidemiologic analysis one tries to eliminate confounding but one tries to detect and estimate effect-
measure modification" (Rothman and Greenland, 1998, 086599). Examples of effect modifiers in
some of the studies evaluated in this ISA include environmental variables, such as temperature or
humidity, individual risk factors, such as education, cigarette smoking status, age in a prospective
January 2010 1-10
-------
cohort study, and community factors, such as percent of population > 65 yr old. It is often possible to
stratify the relationship between health outcome and exposure by one or more of these risk factor
variables. For variables that modify the association, effect estimates in each stratum will be different
from one another and different from the overall estimate, indicating a different exposure-response
relationship may exist in populations represented by these variables. Effect modifiers may be
encountered (a) within single-city time-series studies or (b) across cities in a two-stage hierarchical
model or meta-analysis.
Several statistical methods are available to detect and control for potential confounders, with
none of them being completely satisfactory. Multivariable regression models constitute one tool for
estimating the association between exposure and outcome after adjusting for characteristics of
participants that might confound the results. The use of multipollutant regression models has been
the prevailing approach for controlling potential confounding by copollutants in air pollution health
effects studies. Finding the likely causal pollutant from multipollutant regression models is made
difficult by the possibility that one or more air pollutants may be acting as a surrogate for an
unmeasured or poorly measured pollutant or for a particular mixture of pollutants. In addition, more
than one pollutant may exert similar health effects, resulting in independently observed associations
for multiple pollutants. For example, PM2 5 and NO2 have each been linked to cardiovascular effects
in epidemiologic studies. Correlation between CO concentrations and various copollutants, such as
PM2.5 and NO2, makes it difficult to quantitatively interpret associations between different pollutant
exposures and health effects. Thus, results of models that attempt to distinguish CO effects from
those of copollutants must be interpreted with caution. The number and degree of diversity of
covariates, as well as their relevance to the potential confounders, remain matters of scientific
judgment. Despite these limitations, the use of multipollutant models is still the prevailing approach
employed in most air pollution epidemiologic studies and provides some insight into the potential for
confounding or interaction among pollutants.
Another way to adjust for potential confounding is through stratified analysis, i.e., examining
the association within homogeneous groups with respect to the confounding variable. The use of
stratified analyses has an additional benefit: it allows examination of effect modification through
comparison of the effect estimates across different groups. If investigators successfully measured
characteristics that distort the results, adjustment of these factors help separate a spurious from a true
causal association. Appropriate statistical adjustment for confounders requires identifying and
measuring all reasonably expected confounders. Deciding which variables to control for in a
statistical analysis of the association between exposure and disease or health outcome depends on
knowledge about possible mechanisms and the distributions of these factors in the population under
study. Identifying these mechanisms makes it possible to control for potential sources that may result
in a spurious association.
Adjustment for potential confounders can be influenced by differential exposure measurement
error. There are several components that contribute to exposure measurement error in epidemiologic
studies, including the difference between true and measured ambient concentrations, the difference
between average personal exposure to ambient pollutants and ambient concentrations at central
monitoring sites, and the use of average population exposure rather than individual exposure
estimates. Consideration of issues important for evaluation of exposure to ambient CO include: (1)
spatial variability of CO concentrations across urban areas, particularly with respect to highly
traveled roadways; (2) location of CO monitors at varying distances from roads; and (3) the
detection limit of instruments in the CO monitoring network. Previous AQCDs have examined the
role of measurement error for non-reactive pollutants in time-series epidemiologic studies using
simulated data and mathematical analyses and suggested that transfer of effects from the "causal"
variable to the confounder would only occur under unusual circumstances (i.e., "true" predictors
having high positive or negative correlation; substantial measurement error; or extremely negatively
correlated measurement errors) (U.S. EPA, 2004, 056905).
Confidence that unmeasured confounders are not producing the findings is increased when
multiple studies are conducted in various settings using different subjects or exposures, each of
which might eliminate another source of confounding from consideration. Thus, multicity studies
which use a consistent method to analyze data from across locations with different levels of
covariates can provide insight on potential confounding in associations. Intervention studies, because
of their quasi-experimental nature, can be particularly useful in characterizing causation.
In addition to clinical and epidemiologic studies, the tools of experimental biology have been
valuable for developing insights into human physiology and pathology. Laboratory tools have been
January 2010 1-11
-------
extended to explore the effects of putative toxicants on human health, especially through the study of
model systems in other species. These studies evaluate the effects of exposures to a variety of
pollutants in a highly controlled laboratory setting and allow exploration of MO As or mechanisms
by which a pollutant may cause effects. Background knowledge of the biological mechanisms by
which an exposure might or might not cause disease can prove crucial in establishing or negating a
causal claim. Consideration of evidence on the non-hypoxic effects of CO via cell signaling and
alteration of heme protein function along with evidence on COHb-mediated hypoxic stress, provides
a more complete understanding of the biological response to CO. There are, however, uncertainties
associated with quantitative extrapolations between laboratory animals and humans on the
pathophysiological effects of any pollutant. Animal species can differ from each other in
fundamental aspects of physiology and anatomy (e.g., metabolism, airway branching, hormonal
regulation) that may limit extrapolation.
Interpretations of experimental studies of air pollution effects in laboratory animals, as in the
case of environmental comparative toxicology studies, are affected by limitations associated with
extrapolation models. The differences between humans and rodents with regard to pollutant
absorption and distribution profiles based on metabolism, hormonal regulation, breathing pattern,
exposure dose, and differences in lung structure and anatomy, all have to be taken into consideration.
Also, in spite of a high degree of homology and the existence of a high percentage of orthologous
genes across humans and rodents (particularly mice), extrapolation of molecular alterations at the
gene level is complicated by species-specific differences in transcriptional regulation. Given these
molecular differences, at this time there are uncertainties associated with quantitative extrapolations
between laboratory animals and humans of observed pollutant-induced pathophysiological
alterations under the control of widely varying biochemical, endocrine, and neuronal factors.
1.6.4. Application of Framework for Causal Determination
EPA uses a two-step approach to evaluate the scientific evidence on health or environmental
effects of criteria pollutants. The first step determines the weight of evidence in support of causation
and characterizes the strength of any resulting causal classification. The second step includes further
evaluation of the quantitative evidence regarding the concentration-response relationships and the
loads or levels, duration and pattern of exposures at which effects are observed.
To aid judgment, various "aspects" of causality have been discussed by many philosophers
and scientists. The most widely cited aspects of causality in epidemiology, and public health, in
general, were articulated by Sir Austin Bradford Hill (1965, 071664) and have been widely used
(CDC, 2004, 056384: IARC, 2006, 093206: IOM, 2008, 156586: U.S. EPA, 2005, 086237).These
aspects (Hill, 1965, 071664) have been modified (Table 1-2) for use in causal determinations
specific to health and welfare effects or pollutant exposures (U.S. EPA, 2009, 179916).2 Some
aspects are more likely than others to be relevant for evaluating evidence on the health or
environmental effects of criteria air pollutants. For example, the analogy aspect does not always
apply, especially for the gaseous criteria pollutants, and specificity would not be expected for multi-
etiologic health outcomes, such as asthma or cardiovascular disease, or ecological effects related to
acidification. Aspects that usually play a larger role in determination of causality are consistency of
results across studies, coherence of effects observed in different study types or disciplines, biological
plausibility, 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.
January 2010 1-12
-------
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 may be 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
Coherence ^e c'rawn ^om a vanety 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.
An inference of causality tends to be strengthened by consistency with data from experimental studies or other sources
Bioloaical olausibilitv demonstrating plausible biological mechanisms. A proposed mechanistic linking between an effect and exposure to the agent is
9 " "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.
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.
Strength of the observed
association
Experimental evidence.
The finding of large, precise risks increases confidence that the association is not likely due to chance, bias, or other factors.
However, given a truly causal agent, a small magnitude in the effect could follow from a lower level of exposure, a lower
potency, or the prevalence of other agents causing similar effects. While large effects support causality, modest effects
therefore do not preclude it.
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
the observed association
Evidence of a temporal sequence between the introduction of an agent, and appearance of the effect, constitutes another
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
example, many agents cause respiratory disease and respiratory disease has multiple causes. At the scale of ecosystems, as
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.
Specificity of the observed
association
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.
Although these aspects provide a framework for assessing the evidence, they do not lend
themselves to being considered in terms of simple formulas or fixed rules of evidence leading to
conclusions about causality (Hill, 1965, 071664). For example, one cannot simply count the number
of studies reporting statistically significant results or statistically nonsignificant results and reach
credible conclusions about the relative weight of the evidence and the likelihood of causality. Rather,
these important considerations are taken into account with the goal of producing an objective
appraisal of the evidence, informed by peer and public comment and advice, which includes
weighing alternative views on controversial issues. In addition, it is important to note that the aspects
in Table 1-1 cannot be used as a strict checklist, but rather to determine the weight of the evidence
for inferring causality. In particular, not meeting one or more of the principles does not automatically
preclude a determination of causality (See discussion in CDC, 2004, 056384).
1.6.5. Determination of Causality
In the ISA, EPA assesses the results of recent relevant publications, building upon evidence
available during the previous NAAQS review, to draw conclusions on the causal relationships
between relevant pollutant exposures and health or environmental effects. This ISA uses a five-level
hierarchy that classifies the weight of evidence for causation, not just association1; that is, whether
the weight of scientific evidence makes causation at least as likely as not, in the judgment of the
reviewing group. In developing this hierarchy, EPA has drawn on the work of previous evaluations,
most prominently the lOM's Improving the Presumptive Disability Decision-Making Process for
Veterans (2008, 156586). EPAs Guidelines for Carcinogen Risk Assessment (2005, 086237). and the
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.
January 2010
1-13
-------
U.S. Surgeon General's smoking reports (CDC, 2004, 056384). In the ISA, EPA uses a series of five
descriptors to characterize the weight of evidence for causality. This weight of evidence evaluation is
based on various lines of evidence from across the health and environmental effects disciplines.
These separate judgments are integrated into a qualitative statement about the overall weight of the
evidence and causality. The five descriptors for causal determination are described in Table 1-2.
Table 1-2. Weight of evidence for causal determination.
Health Effects
Evidence is sufficient to conclude that there is a causal relationship
with relevant pollutant exposures. That is, the pollutant has been
shown to result in health effects in studies in which chance, bias, and
confounding could be ruled out with reasonable confidence. For
Causal example: a) controlled human exposure studies that demonstrate
relationship consistent effects ; or b) observational studies that cannot be
explained by plausible alternatives or are supported by other lines of
evidence (e.g., animal studies or mode of action information).
Evidence includes replicated and consistent high-quality studies by
multiple investigators.
Evidence is sufficient to conclude that a causal relationship is likely to
exist with relevant pollutant exposures, but important uncertainties
remain. That is, the pollutant has been shown to result in health
effects in studies in which chance and bias can be ruled out with
Likelvto be a reasonable confidence but potential issues remain. For example: a)
causal observational studies show an association, but copollutant exposures
relationshirj are difficult to address and/or other lines of evidence (controlled
" human exposure, animal, or mode of action information) are limited
or inconsistent; or b) 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.
Evidence is suggestive of a causal relationship with relevant pollutant
Suggestive of exposures, but is limited because chance, bias and confounding
a causal cannot be ruled out. For example, at least one high-quality
relationship epidemiologic study shows an association with a given health
outcome but the results of other studies are inconsistent.
Inade ate , Evidence is inadequate to determine that a causal relationship exists
infer a causal with relevant pollutant exposures. The available studies are of
rplatinnshin insufficient quantity, quality, consistency or statistical power to permit
re nsmp a conclusion regarding the presence or absence of an effect.
Evidence is suggestive of no causal relationship with relevant
Not likely to pollutant exposures. Several adequate studies, covering the full
be a causal range of levels of exposure that human beings are known to
relationship encounter and considering susceptible populations, are mutually
consistent in not showing an effect at any level of exposure.
Ecological and Welfare Effects
Evidence is sufficient to conclude that there is a causal relationship
with relevant pollutant exposures. That is, the pollutant has been
shown to result in effects in studies in which chance, bias, and
confounding could be ruled out with reasonable confidence.
Controlled exposure studies (laboratory or small- to medium-scale
field studies) provide the strongest evidence for causality, but the
scope of inference may be limited. Generally, determination is based
on multiple studies conducted by multiple research groups, and
evidence that is considered sufficient to infer a causal relationship is
usually obtained from the joint consideration of many lines of
evidence that reinforce each other.
Evidence is sufficient to conclude that there is a likely causal
association with relevant pollutant exposures. That is, an association
has been observed between the pollutant and the outcome in studies
in which chance, bias and confounding are minimized, but
uncertainties remain. For example, field studies show a relationship,
but suspected interacting factors cannot be controlled, and other
lines of evidence are limited or inconsistent. Generally, determination
is based on multiple studies in multiple research groups.
Evidence is suggestive of a causal relationship with relevant pollutant
exposures, but chance, bias and confounding cannot be ruled out.
For example, at least one high-quality study shows an effect, but the
results of other studies are inconsistent.
The available studies are of insufficient quality, consistency or
statistical power to permit a conclusion regarding the presence or
absence of an effect.
Several adequate studies, examining relationships with relevant
exposures, are consistent in failing to show an effect at any level of
exposure.
For the CO ISA, determination of causality involved the evaluation of evidence for different
types of health effects associated with short- and long-term exposure periods. In making
determinations of causality for CO, evidence was evaluated for health outcome categories, such as
cardiovascular effects, and then conclusions were drawn based upon the integration of evidence from
across disciplines (e.g., epidemiology, clinical studies and toxicology) and also across the suite of
related individual health outcomes. To accomplish this integration, evidence from multiple and
various types of studies was considered. Response was evaluated over a range of observations which
was determined by the type of study and methods of exposure or dose and response measurements.
Results from different protocols were compared and contrasted.
In drawing judgments regarding causality for the criteria air pollutants, EPA focuses on
evidence of effects at relevant pollutant exposures. To best inform reviews of the NAAQS, these
evaluations go beyond a determination of causality at any dose or concentration to emphasize the
relationship apparent at relevant pollutant exposures. Concentrations generally within an order of
magnitude or two of ambient pollutant measurements are considered to be relevant for this
determination. Building upon the determination of causality are questions relevant to quantifying
health or environmental risks based on our understanding of the quantitative relationships between
pollutant exposures and health or welfare effects. While the causality determination is based
primarily on evaluation of health or environmental effects evidence, EPA also evaluates evidence
related to the doses or levels at which effects are observed. Considerations relevant to evaluation of
quantitative relationships for health and environmental effects are summarized below.
January 2010
1-14
-------
1.6.5.1. Effects on Human Populations
Once a determination is made regarding the causal relationship between the pollutant and
outcome category, important questions regarding quantitative relationships include:
• What is the concentration-response, exposure-response, or dose-response relationship in
the human population?
• What is the interrelationship between incidence and severity of effect?
• What exposure conditions (dose or exposure, duration and pattern) are important?
• What populations appear to be differentially affected (i.e., more susceptible to effects)?
To address these questions, the entirety of policy-relevant quantitative evidence is evaluated to
best quantify those concentration-response relationships that exist. This requires evaluation of
pollutant concentrations and exposure durations at which effects were observed for exposed
populations, including potentially susceptible populations. This integration of evidence resulted in
identification of a study or set of studies that best approximated the concentration-response
relationships between health outcomes and CO, given the current state of knowledge and the
uncertainties that surrounded these estimates. To accomplish this, evidence is considered from
multiple and diverse types of studies. To the extent available, the ISA evaluates results from across
epidemiologic studies that use various methods to evaluate the form of relationships between CO
and health outcomes and draws conclusions on the most well-supported shape of these relationships.
Animal data may also inform evaluation of concentration-response relationships, particularly relative
to MOAs and characteristics of susceptible populations. Chapter 2 presents the integrated findings
informative for evaluation of population risks.
An important consideration in characterizing the public health impacts associated with
exposure to a pollutant is whether the concentration-response relationship is linear across the full
concentration range encountered or if nonlinear relationships exist along any part of this range. Of
particular interest is the shape of the concentration-response curve at and below the level of the
current standards. The shape of the concentration-response curve varies, depending on the type of
health outcome, underlying biological mechanisms and dose. At the human population level,
however, various sources of variability and uncertainty, such as the low data density in the lower
concentration range, possible influence of exposure measurement error, and individual differences in
susceptibility to air pollution health effects, tend to smooth and "linearize" the
concentration-response function. In addition, many chemicals and agents may act by perturbing
naturally occurring background processes that lead to disease, which also linearizes population
concentration-response relationships (Clewell and Crump, 2005, 156359; Crump et al., 1976,
003192; Hoel, 1980, 156555). These attributes of population dose-response may explain why the
available human data at ambient concentrations for some environmental pollutants (e.g., PM, O3,
lead [Pb], environmental tobacco smoke [ETS], radiation) do not exhibit evident thresholds for
cancer or noncancer health effects, even though likely mechanisms include nonlinear processes for
some key events. These attributes of human population dose-response relationships have been
extensively discussed in the broader epidemiologic literature (Rothman and Greenland, 1998,
086599).
Publication bias is a source of uncertainty regarding the magnitude of health risk estimates. It
is well understood that studies reporting non-null findings are more likely to be published than
reports of null findings, and publication bias can also result in overestimation of effect estimate sizes
(loannidis, 2008, 188317). For example, effect estimates from single-city epidemiologic studies have
been found to be generally larger than those from multicity studies (Anderson et al., 2005, 087916)
Although publication bias commonly exists for many research areas, it may be present to a lesser
degree for epidemiologic studies on CO. In general, epidemiologic studies have focused on the
effects of PM, and CO was largely considered as a potentially confounding copollutant of PM. Thus,
CO effect estimates may have been presented in these studies regardless of the statistical significance
of the results.
Finally, identification of the susceptible population groups contributes to an understanding of
the public health impact of pollutant exposures. In this ISA, the term "susceptible population" will
be used as an overarching concept to encompass populations variously described as susceptible,
January 2010 1-15
-------
vulnerable, or sensitive. "Susceptible populations" is defined here as those populations that have a
greater likelihood of experiencing health effects related to exposure to an air pollutant (e.g., CO) due
to a variety of factors including but not limited to: genetic or developmental factors, race, gender,
lifestage, lifestyle (e.g., smoking status and nutrition) or preexisting disease; as well as population-
level factors that can increase an individual's exposure to an air pollutant (e.g., CO) such as
socioeconomic status [SES], which encompasses reduced access to health care, low educational
attainment, residential location, and other factors. Epidemiologic studies can help identify
susceptible populations by evaluating health responses in the study population. Examples include
stratified analyses for subsets of the population under study or testing for interactions or effect
modification by factors such as gender, age group, or health status. Experimental studies using
animal models of susceptibility or disease can also inform the extent to which health risks are likely
greater in specific population groups. Further discussion of these groups is presented in Section 5.7.
1.6.5.2. Effects on Ecosystems or Public Welfare
Key questions for understanding the quantitative relationships between exposure (or
concentration or deposition) to a pollutant and risk to ecosystems or the public welfare include:
• What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
functions, etc.) appear to be affected, or are more sensitive to effects?
• Under what exposure conditions (amount deposited or concentration, duration and
pattern) are effects seen?
• What is the shape of the concentration-response or exposure-response relationship?
Evaluations of causality generally consider the probability of quantitative changes in
ecological and welfare effects in response to exposure. A challenge to the quantification of exposure-
response relationships for ecological effects is the great regional and local variability in ecosystems.
Thus, exposure-response relationships are often determined for a specific ecological system and
scale, rather than at the national or even regional scale. Quantitative relationships therefore are
available site by site. For example, an ecological response to deposition of a given pollutant can
differ greatly between ecosystems. Where results from greenhouse or animal ecotoxicological
studies are available, they may be used to aid in characterizing exposure-response relations,
particularly relative to mechanisms of action, and characteristics of sensitive biota.
1.6.6. Concepts in Evaluating Adversity of Health Effects
In evaluating the health evidence, a number of factors can be considered in determining the
extent to which health effects are "adverse" for health outcomes such as changes in lung function or
in cardiovascular health measures. Some health outcome events, such as hospitalization for
respiratory or cardiovascular diseases, are clearly considered adverse; what is more difficult is
determining the extent of change in the more subtle health measures that is adverse. What constitutes
an adverse health effect may vary between populations. Some changes in healthy individuals may
not be considered adverse while those of a similar type and magnitude are potentially adverse in
more susceptible individuals.
For example, the extent to which changes in lung function are adverse has been discussed by
the American Thoracic Society (ATS) in an official statement titled What Constitutes an Adverse
Health Effect of Air Pollution? (2000, 011738). This statement updated the guidance for defining
adverse respiratory health effects that had been published 15 years earlier (ATS, 1985, 006522).
taking into account new investigative approaches used to identify the effects of air pollution and
reflecting concern for impacts of air pollution on specific susceptible groups. In the 2000 update,
there was an increased focus on quality of life measures as indicators of adversity and a more
specific consideration of population risk. Exposure to air pollution that increases the risk of an
adverse effect to the entire population is viewed as adverse, even though it may not increase the risk
of any identifiable individual to an unacceptable level. For example, a population of asthmatics
could have a distribution of lung function such that no identifiable individual has a level associated
with significant impairment. Exposure to air pollution could shift the distribution such that no
January 2010 1-16
-------
identifiable individual experiences clinically relevant effects. This shift toward decreased lung
function, however, would be considered adverse because individuals within the population would
have diminished reserve function and therefore would be at increased risk to further environmental
insult.
It is important to recognize that the more subtle health outcomes may be linked to health
events that are clearly adverse. For example, air pollution has been shown to affect markers of
transient myocardial ischemia such as ST-segment abnormalities and onset of exertional angina. In
some cases, these effects are silent yet may still increase the risk of a number of cardiac events,
including MI and sudden death.
1.7. Summary
This ISA is a concise evaluation and synthesis of the most policy-relevant science for
reviewing the NAAQS for CO, and it is the chief means for communicating the critical science
judgments relevant to that NAAQS review. It reviews the most policy-relevant evidence from
atmospheric science, exposure, and health and environmental effects studies and includes
mechanistic evidence from basic biological science. This final IS A incorporates clarification and
revisions based on public comments and advice and comments provided by EPA's CASAC on the
first and second draft ISAs (Brain and Samet, 2009, 194669; Brain and Samet, 2010, 202840).
Annexes to the ISA provide additional details of the literature published since the last review. A
framework for making critical judgments concerning causality was presented in this chapter. It relies
on a widely accepted set of principles and standardized language to express evaluation of the
evidence. This approach can bring rigor and clarity to current and future assessments. This ISA
should assist EPA and others, now and in the future, to accurately represent what is presently known
and what remains unknown concerning the effects of CO on human health and public welfare.
January 2010 1-17
-------
References
Anderson HR; Atkinson RW; Peacock JL; Sweeting MJ; Marston L (2005). Ambient participate matter and health effects:
Publication bias in studies of short-term associations. Epidemiology, 16: 155-163. 087916
ATS (1985). American Thoracic Society: 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
Blair A; Burg J; Foran J; Gibb H; Greenland S; Morris R; Raabe G; Savitz D; Teta J; Wartenberg D; Wong O; Zimmerman
R (1995). Guidelines for application of meta-analysis in environmental epidemiology. Regul Toxicol Pharmacol,
22: 189-197.079190
Brain JD; Samet JM (2009). Review of EPA's Integrated Science Assessment for carbon monoxide (first external review
draft). Clean Air Scientific Advisory Committee, Science Advisory Board. Washington, DC. EPA-CASAC-09-011.
194669
Brain JD; Samet JM.. (2010). Review of Integrated Science Assessment for Carbon Monoxide (Second External Review
Draft). Clean Air Scientific Advisory Committee, Science Advisory Board. Washington, DC. EPA-CASAC-10-005.
202840
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
Hedblad B; Engstrom G; Janzon E; Berglund G; Janzon L (2006). COHb% as a marker of cardiovascular risk in never
smokers: results from a population-based cohort study. Scand J Public Health, 34: 609-615. 199512
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, World Health Organization. Lyon,France .
http://monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdf. 093206
loannidis JPA (2008). Why most discovered true associations are inflated. Epidemiology, 19: 640-648. 188317
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
McClellanRO (1991). CASAC Closure on the 1991 AQCD for Carbon Monoxide. Clean Air 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. Clean Air Scientific Advisory
Committee, Science Advisory Board, U.S. Environmenal Protection Agency. Washington, DC. US EPA-SAB-
CASACLTR-2-016. 194667
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).
January 2010 1-18
-------
NAPAP(1991). The experience and legacy of NAPAP report of the Oversight Review Board. National Acid Precipitation
Assessment Program. Washington, DC. 095894
National Research Council (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
U.S. EPA (1979). Air quality criteria for carbon monoxide. U.S. Environmental Protection Agency. Washington, DC. EPA-
600/8-79-022. 017687
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 (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.
EPA-450/5-84-004. 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. U. S. Environmental Protection Agency. Washington,
DC. EPA-600/8-83-033F. 012690
U. S. EPA (1991). Air quality criteria for carbon monoxide. U. S. Environmental Protection Agency. Research Triangle Park,
NC. EPA/600/8-90/045F. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000554R.txt. 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. U.S. Environmental Protection Agency. Washington, DC.
EPA-452/R-92-004. 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-03/001B.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 16283. 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
January 2010 1-19
-------
Chapter 2. Integrative Overview
The subsequent chapters of this ISA present the most policy-relevant information related to
this review of the NAAQS for CO, including a synthesis of the evidence presented in the 2000 CO
AQCD (U.S. EPA, 2000, 000907). along with the assessment of more recent studies. This chapter
integrates important findings from the disciplines evaluated in this current assessment of the CO
scientific literature, which includes the atmospheric sciences, ambient air data analyses, climate
forcing effects, exposure assessment, dosimetry, and health effects research (animal toxicological
studies, controlled human exposure studies, and epidemiologic studies). The EPA framework for
causal determinations described in Chapter 1 has been applied to the body of evidence evaluated in
this assessment in order to characterize the relationship between exposure to CO at relevant
concentrations and health effects. The EPA framework applied here employs a five-level hierarchy
that classifies the weight of evidence for causation:
• Causal relationship
• Likely to be a causal relationship
• Suggestive of a causal relationship
• Inadequate to infer a causal relationship
• Not likely to be a causal relationship
This evaluation led to causal determinations for several health outcome categories and
characterization of the magnitude of the response, including responses in susceptible populations,
over a range of relevant concentrations. This integration of evidence also provides a basis for
characterizing the concentration-response relationships of CO and adverse health outcomes for the
U.S. population, given the current state of knowledge.
This chapter summarizes and integrates the newly available scientific evidence that best
informs consideration of the policy-relevant questions that frame this assessment, which are
presented in Chapter 1. Section 2.1 discusses the trends in ambient concentrations and sources of
CO. Section 2.2 provides an overview of climate forcing related directly and indirectly to CO.
Section 2.3 provides a brief summary of factors influencing personal exposure to ambient CO.
Section 2.4 summarizes CO dosimetry and pharmacokinetics and describes what is known regarding
the modes of action of CO. Section 2.5 integrates the evidence from studies that examined health
effects related to short- and long-term exposure to CO and discusses important uncertainties
identified in the interpretation of the scientific evidence. Section 2.6 summarizes policy-relevant
considerations associated with exposure to CO including evidence of effects in potentially
susceptible populations and information on the shape of the concentration-response function. Finally,
Section 2.7 presents an integrated summary of the health effects of CO, reports the levels at which
effects are observed and discusses important uncertainties to consider in the interpretation of the
scientific evidence.
2.1. Ambient CO Sources and Concentrations
CO is formed by incomplete combustion of carbon-containing fuels and by photochemical
reactions in the atmosphere. Nationally, on-road mobile sources constituted more than half of total
CO emissions in 2002, or ~61 of-117 million tons (MT) of total CO emissions, based on the most
recent publicly available data meeting data quality objectives from EPA's National Emissions
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).
January 2010 2-1
-------
Inventory (NEI). In metropolitan areas in the U.S., as much as 75% of all CO emissions result from
on-road vehicle exhaust. The majority of these on-road CO emissions are derived from gasoline-
powered vehicles. When emissions from incomplete combustion of fuels powering nonroad mobile
sources, such as farm and construction equipment, lawnmowers, boats, ships, snowmobiles, and
aircraft, are included, all mobile sources accounted for -80% of total CO emissions in the U.S. in
2002. Other primary sources of CO include wildfires, controlled vegetation burning, residential
biomass combustion, and industrial processes. While CO emissions from nonroad mobile sources,
wild fires, and industry have remained fairly constant, on-road mobile source CO emissions have
decreased by roughly 5% per year since the early 1990s. Secondary sources of CO include the
oxidation of both anthropogenic and biogenic hydrocarbons, such as methane and isoprene and other
carbon containing species including aldehydes and alcohols. During summer when biogenic
emissions are at their peak, secondary sources of CO are estimated to be a significant fraction of
total U.S. sources; however, secondary sources are dispersed over the entire country, while direct
emissions are concentrated near primary sources, such as on-road mobile sources, which are mainly
in urban areas. Although these estimates are generated using well-established approaches,
uncertainties are inherent in the emission factors and models used to represent sources for which
emissions have not been directly measured, and these uncertainties vary by source category, season,
and region.
Significant reductions in ambient CO concentrations and in the number of NAAQS
exceedances have been observed over the past 25 yr, a continuation of trends documented in the
2000 CO AQCD (U.S. EPA, 2000, 000907). Nationwide ambient CO data from the EPA Air Quality
System (AQS) for the years 2005-2007 show that the median 1-h daily maximum (max)
concentration across the U.S. was 0.7 ppm; the mean was 0.9 ppm; the 95th percentile was 2.4 ppm;
and the 99th percentile was 3.8 ppm. Roughly one-third of the 1-h daily max data fell below the limit
of detection (LOD) for the majority of CO monitors reporting to AQS. The median 8-h daily max
ambient CO concentration for the years 2005-2007 was 0.5 ppm; the mean was 0.7 ppm; the 95th
percentile was 1.7 ppm; and the 99th percentile was 2.6 ppm. Half of the 8-h daily max
concentrations fell below the LOD for the majority of CO monitors in the field. The current CO
NAAQS are 35 ppm (1-h avg) and 9 ppm (8-h avg), not to be exceeded more than once per year.
During the years 2005-2007, 1-h and 8-h CO concentrations did not exceed the NAAQS level more
than once per year at any monitoring site. Moreover, in these 3 yr, a 1-h avg concentration in excess
of 35 ppm was reported only once (39 ppm), and there were only 7 reported 8-h avg values
nationwide in excess of 9 ppm in all 3 yr. Seasonally divided box plots of data from 2005-2007
compiled for spatially diverse urban metropolitan areas illustrate the tendency for higher median CO
concentrations and wider variations in concentrations in the winter and fall compared with the spring
and summer (Section 3.5).
Policy-relevant background (PRB) concentrations include contributions from natural sources
everywhere in the world and from anthropogenic sources outside the U.S., Canada, and Mexico.
PRB concentrations of CO were estimated for this assessment using data for the years 2005-2007
collected at 12 remote sites in the U.S. which are part of the National Oceanic and Atmospheric
Administration's (NOAA) Global Monitoring Division (GMD) and are not part of the EPA national
regulatory network. The 3-yr avg CO PRB averaged ~0.13 ppm in Alaska, ~0.10 ppm in Hawaii, and
-0.13 ppm over the contiguous U.S. (CONUS). The analysis for North American PRB in this
assessment was made by segregating the three Alaska sites based on their high latitude and the two
Hawaii sites based on their distance from the continent, and then treating the remaining seven sites
as being more representative of the CONUS PRB. Note that these seven sites are affected by
anthropogenic emissions in North America to varying degrees.
2.2. Climate Forcing Effects
Recent data do not alter the current well-established understanding of the role of urban and
regional CO in continental- and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
Intergovernmental Panel on Climate Change (IPCC, 2001, 156587; IPCC, 2007, 092765). CO is a
weak direct contributor to radiative forcing (RF) and greenhouse warming. Sinha and Toumi (1996,
193747) estimated the direct RF of CO computed for all-sky conditions at the tropopause to be
January 2010 2-2
-------
0.024 W/m2 based on an assumed change in CO mean global concentrations from 25 to 100 ppb
since preindustrial times. The direct RF attributed to CO over this time frame is -1.5% of the direct
RF for CO2 estimated by the IPCC (Forster et al., 2007, 092936).
More importantly, CO can indirectly cause increased RF because it reacts with tropospheric
OH and thus can increase the lifetime of trace gases in the atmosphere including the GHGs CH4 and
O3. Additionally, the major pathway for removal of CO from the atmosphere is reaction with OH to
produce CO2. CH4, O3, and CO2 absorb infrared radiation from the Earth's surface and contribute to
the greenhouse effect. Indirect RF attributed to 1750-2005 emissions of CO through changes in
concentration of the GHGs O3, CH4, and CO2 was estimated by Forster et al. (2007, 092936) to be
-0.2 W/m2, or -12% of the direct RF of CO2 (Figure 3-7). The future direct and indirect integrated
RF for year 2000 emissions of CO was estimated to be -0.2 W/m2-yr with -50% uncertainty over
both 20-yr and 100-yr time horizons (Figure 3-8). The RF related to short-lived CO is -25% of that
for CO2 for a 20-yr time horizon but only -7% of that for longer-lived CO2 over a 100-yr horizon.
Overall, the evidence reviewed in this assessment is sufficient to conclude that a Causal
relationship exists between current atmospheric concentrations of CO and effects on
climate
2.3. Exposure to Ambient CO
Very few recent exposure assessment studies involve ambient CO concentration data. The
studies of personal exposure to ambient CO presented here generally found that the largest
percentage of time in which an individual is exposed to ambient CO occurs indoors but that the
highest ambient CO exposure levels occur in transit. In-vehicle CO concentrations are typically
reported to be between 2 and 5 times higher than ambient concentrations, although peak in-vehicle
concentrations more than an order of magnitude higher than corresponding ambient monitor
concentrations have also been reported. Among commuters, exposures were higher for those
traveling in automobiles in comparison with those traveling on buses and motorbikes and with those
cycling or walking. Ambient CO exposure in automobiles has been demonstrated to vary with
vehicle ventilation settings, and a very small portion of that exposure is thought to come from the
vehicle in which the exposed person travels. High near-road CO concentrations can be important for
those living in the near-road environment because virtually all of ambient CO infiltrates indoors.
Hence, indoor exposure to ambient CO is determined by the CO concentration outside the building.
CO concentration in the near-road environment has been shown to decrease sharply with downwind
distance from a highway, wind direction, and emission source strength (e.g., number of vehicles on a
highway); natural and urban topography also influence localized ambient CO concentrations.
Recent exposure assessment studies support one of the main conclusions of the 2000 CO
AQCD (U.S. EPA, 2000, 000907). that central site ambient CO monitors may overestimate or
underestimate individuals' personal exposure to ambient CO because ambient CO concentration is
spatially variable, particularly when analyzing exposures in the near-road environment. Exposure
error may occur when the ambient CO concentration measured at the central site monitor is used as
an ambient exposure surrogate and differs from the actual ambient CO concentration outside a
subject's residence and/or worksite. For example, measurement at a "hot spot" could skew
community exposure estimates upwards, and likewise measurement at a location with few CO
sources could skew exposure estimates downwards. Correlations across CO monitors can vary
widely within and between cities across the U.S. as a function of natural and urban topography,
meteorology, source strength and proximity to sources. Typically, intersampler correlation ranges
from 0.35 to 0.65 for monitors sited at different scales within a metropolitan area, although it can be
greater than 0.8 in some areas.
Health effects estimates from time-series epidemiologic studies are not biased by spatial
variability in CO concentrations if concentrations at different locations are correlated in time.
Exposure assessment in epidemiologic studies is also complicated by the existence of CO in
multipollutant mixtures emitted by combustion processes, making it difficult to quantify the health
effects related specifically to CO exposure compared with those related to another combustion-
related pollutant or mix of pollutants. In most circumstances, exposure error tends to bias a health
effect estimate downward (Sheppard et al., 2005, 079176: Zeger et al., 2000, 001949). Spatial and
temporal variability not fully captured by ambient monitors and correlation of CO with copollutants
January 2010 2-3
-------
are examples of sources of uncertainty that could widen confidence intervals of health effects
estimates.
2.4. Dosimetry, Pharmacokinetics, and Mode of Action
2.4.1. Dosimetry and Pharmacokinetics
Upon inhalation, CO elicits various health effects by binding to and altering the function of a
number of heme-containing molecules, mainly hemoglobin (Hb). The formation of COHb reduces
the oxygen (O2)-carrying capacity of blood and impairs the release of O2 from oxyhemoglobin
(O2Hb) to the tissues. The 2000 CO AQCD (U.S. EPA, 2000, 000907) has a detailed description of
the well-established Coburn-Forster-Kane (CFK) equation, which has been used for many years to
model COHb formation. Since then, models have been developed that include myoglobin (Mb) and
extravascular storage compartments, as well as other dynamics of physiology relevant to CO uptake
and elimination. These models have indicated that CO has a biphasic elimination curve, due to initial
washout from the blood followed by a slower flux from the tissues. The flow of CO between the
blood and alveolar air or tissues is controlled by diffusion down the pCO gradient. The uptake of CO
is governed not only by this CO pressure differential but also by physiological parameters, such as
minute ventilation and lung diffusing capacity that can, in turn, be affected by factors such as
exercise, age, and medical conditions (e.g., obstructive lung disease). Susceptible populations, such
as health-compromised individuals, are at a greater risk from COHb-induced health effects due to
altered CO kinetics, compromised cardiopulmonary processes, and increased baseline hypoxia
levels. Altitude also may have a substantial effect on the kinetics of COHb formation, especially for
visitors to high-altitude areas. Compensatory mechanisms, such as increased cardiac output, combat
the decrease in barometric pressure. Altitude also increases the endogenous production of CO
through upregulation of heme oxygenase (HO). CO is considered a second messenger and is
endogenously produced from the catabolism of heme proteins by enzymes such as HO-1 (the
inducible form of heme oxygenase) and through endogenous lipid peroxidation. Finally, CO is
removed from the body by expiration and oxidation to CO2.
2.4.2. Mode of Action
The diverse effects of CO are dependent upon concentration, duration of exposure, and the cell
types and tissues involved. Responses to CO are not necessarily due to a single process and may
instead be mediated by a combination of effects including COHb-mediated hypoxic stress and other
mechanisms such as free radical production and the initiation of cell signaling. However, binding of
CO to reduced iron in heme proteins with subsequent alteration of heme protein function is the
common mechanism underlying the biological responses to CO (see Section 5.1).
As discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907). the most well-known
pathophysiological effect of CO is tissue hypoxia caused by binding of CO to Hb. Not only does the
formation of COHb reduce the O2-carrying capacity of blood, but it also impairs the release of O2
from O2Hb. Compensatory alterations in hemodynamics, such as vasodilation and increased cardiac
output, protect against tissue hypoxia. Depending on the extent of CO exposure, these compensatory
changes may be effective in people with a healthy cardiovascular system. However, hemodynamic
responses following CO exposure may be insufficient in people with decrements in cardiovascular
function, resulting in health effects, as described in Section 5.2. Binding of CO to Mb, as discussed
in the 2000 CO AQCD (U.S. EPA, 2000, 000907) and in Section 4.3.2.3, can also impair the delivery
of O2 to tissues. Mb has a high affinity for CO, about 25 times that of O2; however, pathophysiologic
effects are seen only after high-dose exposures to CO, resulting in COMb concentrations far above
baseline levels.
Nonhypoxic mechanisms underlying the biological effects of CO have been the subject of
recent research since the 2000 CO AQCD (U.S. EPA, 2000, 000907). Most of these mechanisms are
related to CO's ability to bind heme-containing proteins other than Hb and Mb. These mechanisms,
which may be interrelated, include alteration in nitric oxide (NO) signaling, inhibition of
January 2010 2-4
-------
cytochrome c oxidase, heme loss from proteins, disruption of iron homeostasis, alteration in cellular
redox status, alteration in ion channel activity and modulation of protein kinase pathways. CO is a
ubiquitous cell signaling molecule with numerous physiological functions. The endogenous
generation and release of CO from heme by HO-1 and HO-2 is tightly controlled, as is any
homeostatic process. However, exogenously-applied CO has the capacity to disrupt multiple heme-
based signaling pathways due to its nonspecific nature. Only a limited amount of information is
available regarding the impact of exogenous CO on tissue and cellular levels of CO and on signaling
pathways. However, recent animal studies demonstrated increased tissue CO levels and biological
responses following exposure to 50 ppm CO. Whether or not environmentally-relevant exposures to
CO lead to adverse health effects through altered cell signaling is an open question for which there
are no definitive answers at this time. However, experiments demonstrating oxidative/nitrosative
stress, inflammation, mitochondrial alterations and endothelial dysfunction at concentrations of CO
within one or two orders of magnitude higher than ambient concentrations suggest a potential role
for such mechanisms in pathophysiologic responses. Furthermore, prolonged increases in
endogenous CO resulting from chronic diseases may provide a basis for the enhanced sensitivity of
susceptible populations to CO-mediated health effects such as is seen in individuals with coronary
artery disease.
2.5. Health Effects
This assessment reviewed health effects evidence regarding the effect of CO on several
categories of health outcomes. Table 2-1 presents the overall conclusions of the ISA regarding the
presence of a causal relationship between short-term (i.e., hours, days, or weeks) or long-term (i.e.,
months or years) exposure to relevant CO concentrations (defined in Chapter 1 as generally within
one or two orders of magnitude of ambient CO concentrations) and health outcome categories.
Summaries of the evidence supporting each causal determination and considerations relevant to
application of the causal framework are provided in the following subsections.
Table 2-1. Causal determinations for health effects categories.
Outcome Category Exposure Period Causality Determination
. , . ,.... Short-term Likely to be a causal relationship
Cardiovascular morbidity
Long-term Inadequate to infer a causal relationship
Central nervous system effects Short- and long-term Suggestive of a causal relationship
Birth outcomes and Developmental effects Long-term Suggestive of a causal relationship
Short-term Suggestive of a causal relationship
Respiratory morbidity
Long-term Inadequate to infer a causal relationship
Short-term Suggestive of a causal relationship
Mortality
Long-term Not likely to be a causal relationship
2.5.1. Cardiovascular Morbidity
The most compelling evidence of a CO-induced effect on the cardiovascular system at COHb
levels relevant to the current NAAQS comes from a series of controlled human exposure studies
among individuals with coronary artery disease (CAD) (Section 5.2). These studies, described in the
1991 (U.S. EPA, 1991, 017643) and 2000 (U.S. EPA, 2000, 000907) CO AQCDs, demonstrate
consistent decreases in the time to onset of exercise-induced angina and ST-segment changes
following CO exposures resulting in COHb levels of 2-6% (Section 5.2.4). No human clinical
studies have been designed to evaluate the effect of controlled exposures to CO resulting in COHb
concentrations lower than 2%. Human clinical studies published since the 2000 CO AQCD
January 2010 2-5
-------
(U.S. EPA, 2000, 000907) have reported no association between CO and ST-segment changes or
arrhythmia; however, none of these studies included individuals with diagnosed heart disease.
While the exact physiological significance of the observed ST-segment changes among
individuals with CAD is unclear, ST-segment depression is a known indicator of myocardial
ischemia. It is also important to note that the individuals with CAD who participated in these
controlled exposure studies may not be representative of the most sensitive individuals in the
population. It is conceivable that the most sensitive individuals respond to levels of COHb lower
than those evaluated in controlled human exposure studies. Variability in activity patterns and
severity of disease among individuals with CAD is likely to influence the critical level of COHb
which leads to adverse cardiovascular effects.
The degree of ambient CO exposure which leads to attainment of critical levels of COHb will
also vary between individuals. Although endogenous COHb is generally <1% in healthy individuals,
higher endogenous COHb levels are observed in individuals with certain medical conditions.
Nonambient exposures to CO, such as exposure to environmental tobacco smoke (ETS), may
increase COHb above endogenous levels, depending on the gradient of pCO. Ambient exposures
may cause a further increase in COHb. Modeling results described in Chapter 4 indicate that
increases of ~1% COHb are possible with exposures of several ppm CO depending on exposure
duration and exercise level.
Findings of epidemiologic studies conducted since the 2000 CO AQCD (U.S. EPA, 2000,
000907) are coherent with results of the controlled human exposure studies. These recent studies
observed associations between ambient CO concentration and emergency department (ED) visits and
hospital admissions (HAs) for ischemic heart disease (IHD), congestive heart failure (CHF) and
cardiovascular diseases (CVD) as a whole and were conducted in locations where the mean 24-h avg
CO concentrations ranged from 0.5 ppm to 9.4 ppm (Table 5-7). All but one of these studies that
evaluated CAD outcomes (IHD, MI, angina) reported positive associations (Figure 5-2). Although
CO is often considered a marker for the effects of another traffic-related pollutant or mix of
pollutants, evidence indicates that CO associations generally remain robust in copollutant models
and supports a direct effect of short-term ambient CO exposure on CVD morbidity. These studies
add to findings reported in the 2000 CO AQCD (U.S. EPA, 2000, 000907) that demonstrated
associations between short-term variations in ambient CO concentrations and exacerbation of heart
disease.
The known role of CO in limiting O2 availability lends biological plausibility to ischemia-
related health outcomes following CO exposure. However, it is not clear whether the small changes
in COHb associated with ambient CO exposures result in substantially reduced O2 delivery to
tissues. Recent toxicological studies suggest that CO may also act through other mechanisms by
initiating or disrupting cellular signaling. Studies in healthy animals demonstrated oxidative injury
and inflammation in response to 50-100 ppm CO, while studies in animal models of disease
demonstrated exacerbation of cardiomyopathy and increased vascular remodeling in response to
50 ppm CO. Further investigations will be useful in determining whether altered cell signaling
contributes to adverse health effects following ambient CO exposure.
Given the consistent and coherent evidence from epidemiologic and human clinical studies,
along with biological plausibility provided by CO's role in limiting O2 availability, it is concluded
that a causal relationship is likely to exist between relevant short-term exposures to CO
and cardiovascular morbidity.
Only two epidemiologic studies were identified that investigated the relationship between
long-term exposure to CO and cardiovascular effects, and the results of these studies provide very
limited evidence of an association (Section 5.2.2). Considering the lack of evidence from controlled
human exposure studies and the very limited evidence from toxicological studies on cardiovascular
effects following long-term exposure to CO, the available evidence is inadequate to Conclude
that a causal relationship exists between relevant long-term exposures to CO and
cardiovascular morbidity.
2.5.2. Central Nervous System Effects
Exposure to high levels of CO has long been known to adversely affect central nervous system
(CNS) function, with symptoms following acute CO poisoning including headache, dizziness,
January 2010 2-6
-------
cognitive difficulties, disorientation, and coma. However, the relationship between ambient levels of
CO and neurological function is less clear and has not been evaluated in epidemiologic studies.
Studies of controlled human exposures to CO discussed in the 2000 CO AQCD (U.S. EPA, 2000,
000907) reported inconsistent neural and behavioral effects following exposures resulting in COHb
concentrations of 5-20%. No new human clinical studies have evaluated central nervous system or
behavioral effects of exposure to CO. At ambient-level exposures, healthy adults may be protected
against CO-induced neurological impairment owing to compensatory responses including increased
cardiac output and cerebral blood flow. However, these compensatory mechanisms are likely
impaired among certain potentially susceptible groups including individuals with reduced
cardiovascular function.
Toxicological studies that were not discussed in the 2000 CO AQCD (U.S. EPA, 2000,
000907) employed rodent models to show that CO exposure during the in utero or perinatal period
can adversely affect adult outcomes, including behavior, neuronal myelination, neurotransmitter
levels or function, and the auditory system (discussed in Section 5.3). In utero CO exposure,
including both intermittent and continuous exposure, has been shown to impair multiple behavioral
outcomes in offspring (75-150 ppm). In utero CO exposure (75 and 150 ppm) was associated with
significant myelination decrements and neurotransmitter effects (up to 200 ppm). Finally, perinatal
CO exposure has been shown to affect the developing auditory system of rodents, inducing
permanent changes into adulthood (12.5-100 ppm), some of which appear to be reactive oxygen
species mediated. Considering the combined evidence from controlled human exposure and
toxicoiogicai studies, the evidence is suggestive of a causal relationship between relevant
short- and long-term exposures to CO and central nervous system effects.
2.5.3. Birth Outcomes and Developmental Effects
The most compelling evidence for a CO-induced effect on birth and developmental outcomes
is for preterm birth (PTB) and cardiac birth defects. These outcomes were not addressed in the 2000
CO AQCD (U.S. EPA, 2000, 000907). which included only two studies that examined the effect of
ambient CO on low birth weight (LEW). Since then, a number of studies have been conducted
looking at varied outcomes, including PTB, birth defects, fetal growth (including LEW), and infant
mortality.
There is limited epidemiologic evidence that CO during early pregnancy (e.g., first month and
first trimester) is associated with an increased risk of PTB. The only U.S. studies to investigate the
PTB outcome were conducted in California, and these reported consistent positive associations with
CO exposure during early pregnancy when exposures were assigned from monitors within close
proximity of the mother's residential address. Additional studies conducted outside of the U.S.
provide supportive, though less consistent, evidence of an association between CO concentration and
PTB.
Very few epidemiologic studies have examined the effects of CO on birth defects. Two of
these studies found maternal exposure to CO to be associated with an increased risk of cardiac birth
defects. Human clinical studies also demonstrated the heart as a target for CO effects (Section 5.2).
Animal toxicoiogicai studies provide additional evidence for cardiac effects with reported transient
cardiomegaly at birth after continuous in utero CO exposure (60, 125, 250 and 500 ppm CO) and
delayed myocardial electrophysiological maturation (150 ppm CO). Toxicoiogicai studies have also
shown that continuous in utero CO exposure (250 ppm) induced teratogenicity in rodent offspring in
a dose-dependent manner that was further affected by dietary protein (65 ppm CO) or zinc
manipulation (500 ppm CO). Toxicoiogicai studies of CO exposure over the duration of gestation
have shown skeletal alterations (7 h/day, CO 250 ppm) or limb deformities (24 h/day, CO 180 ppm)
in prenatally exposed offspring.
There is evidence of ambient CO exposure during pregnancy having a negative effect on fetal
growth in epidemiologic studies. In general, the reviewed studies, summarized in Figures 5-7
through 5-9, reported small reductions in birth weight (ranging -5-20 g). Several studies examined
various combinations of birth weight, LBW, and small for gestational age (SGA)/intrauterine growth
restriction (IUGR) and inconsistent results are reported across these metrics. It should be noted that
having a measurable, even if small, change in a population is different than having an effect on a
subset of susceptible births and increasing the risk of IUGR/LBW/SGA. It is difficult to conclude if
CO is related to a small change in birth weight in all births across the population, or a marked effect
January 2010 2-7
-------
in some subset of births. Toxicology studies have found associations between CO exposure in
laboratory animals and decrements in birth weight (90-600 ppm), as well as reduced prenatal growth
(65-500 ppm CO).
In general, there is limited epidemiologic evidence that CO is associated with an increased risk
of infant mortality during the neonatal or post-neonatal periods. In support of this limited evidence,
animal toxicological studies provide some evidence that exogenous CO exposure to pups in utero
significantly increased postnatal mortality (7 h/day and 24 h/day, 250 ppm CO; 24 h/day, 90 or
180 ppm CO) and prenatal mortality (7 h/day, 250 ppm CO).
Evidence exists for additional developmental outcomes which have been examined in
toxicological studies but not epidemiologic or human clinical studies, including behavioral
abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes, and
changes in the auditory system. Structural aberrations of the cochlea involving neuronal activation
(12.5, 25 and 50 ppm CO) and auditory related nerves (25 ppm CO) were seen in pups after neonatal
CO exposure. Auditory functional testing using otoacoustic emissions testing (OAE at 50 ppm CO)
and 8th cranial nerve action potential (AP) amplitude measurements (12, 25, 50, 100 ppm CO) in
rodents exposed perinatally to CO showed auditory decrements at postnatal day (PND) 22 (OAE and
AP) and permanent changes in AP into adulthood (50 ppm CO). Furthermore, exogenous CO may
interact with or disrupt the normal physiological roles that endogenous CO plays in the body. There
is evidence that CO plays a role in maintaining pregnancy, controlling vascular tone, regulating
hormone balance, and sustaining normal ovarian follicular maturation.
Overall, there is limited, though positive, epidemiologic evidence for a CO-induced effect on
PTB and birth defects, and weak evidence for a decrease in birth weight, other measures of fetal
growth, and infant mortality. Animal toxicological studies provide support and coherence for these
effects. Both hypoxic and nonhypoxic mechanisms have been proposed in the toxicological literature
(Section 5.1), though a clear understanding of the mechanisms underlying reproductive and
developmental effects is still lacking. Taking into consideration the positive evidence for some birth
and developmental outcomes from epidemiologic studies and the resulting coherence for these
associations in animal toxicological studies, the evidence is Suggestive Of 3 Causal relationship
between relevant long-term exposures to CO and developmental effects and birth
outcomes
2.5.4. Respiratory Morbidity
New epidemiologic studies, supported by the body of literature summarized in the 2000 CO
AQCD (U.S. EPA, 2000, 000907). provide evidence of positive associations between short-term
exposure to CO and respiratory-related outcomes including pulmonary function, respiratory
symptoms, medication use, hospital admissions, and ED visits. The majority of the studies evaluated
have not conducted extensive analyses to examine the potential influence of model selection or effect
modifiers on the association between CO and respiratory morbidity. A limited number of studies
have examined the potential confounding effects of copollutants on CO risk estimates, and found
that CO risk estimates were generally robust to the inclusion of O3, SO2, and PM in two-pollutant
models, but were slightly attenuated in models with NO2. However, the limited amount of evidence
from studies that have examined the effect of gaseous pollutants on CO-respiratory morbidity risk
estimates in two-pollutant models, specifically NO2, has contributed to the inability to disentangle
the effects attributed to CO from the larger complex air pollution mix (particularly motor vehicle
emissions), and this limits interpretation of the results observed in the epidemiologic studies
evaluated. A key uncertainty in interpreting the epidemiologic studies evaluated is the biological
mechanism(s) that could explain the effect of CO on respiratory health. Animal toxicological studies,
however, provide some evidence that short-term exposure to CO (50-100 ppm) can cause oxidative
injury and inflammation and alter pulmonary vascular remodeling. Controlled human exposure
studies have not extensively examined the effect of short-term exposure to CO on respiratory
morbidity, with a very limited number of studies reporting inconsistent effects of CO on pulmonary
function. Although these controlled human exposure studies do not provide evidence to support
CO-related respiratory health effects, epidemiologic studies show positive associations for CO-
induced lung-related outcomes and animal toxicological studies demonstrate the potential for an
underlying biological mechanism, which together provide evidence that is Suggestive Of 3 CdUSdl
relationship between relevant short-term exposures to CO and respiratory morbidity.
January 2010 2-8
-------
Currently, only a few studies have been conducted that examine the association between long-
term exposure to CO and respiratory morbidity, including allergy. Although some studies did observe
associations between long-term exposure to CO and respiratory health outcomes, key uncertainties
still exist. These uncertainties include: the lack of replication and validation studies to evaluate new
methodologies (i.e., Deletion/Substitution/Addition (DSA) algorithm) that have been used to
examine the association between long-term exposure to CO and respiratory health effects; whether
the respiratory health effects observed in response to long-term exposure to CO can be explained by
the proposed biological mechanisms; and the lack of copollutant analyses to disentangle the
respiratory effects associated with CO due to its high correlation with NO2 and other combustion-
related pollutants. Overall, the evidence available is inadequate to Conclude that 3 C3USal
relationship exists between relevant long-term exposures to CO and respiratory
morbidity.
2.5.5. Mortality
The recently available multicity studies, which consist of larger sample sizes, along with the
single-city studies, evaluated reported associations that are generally consistent with the results of
the studies evaluated in the 2000 CO AQCD (U.S. EPA, 2000, 000907). However, to date the
majority of the literature has not conducted extensive analyses to examine the potential influence of
model selection, effect modifiers, or confounders on the association between CO and mortality.
The multicity studies reported comparable CO mortality risk estimates for total
(nonaccidental) mortality, with the APHEA2 European multicity study showing slightly higher
estimates for cardiovascular mortality in single-pollutant models. However, when examining
potential confounding by copollutants, these studies consistently showed that although CO mortality
risk estimates remained positive, they were reduced when NO2 was included in the model. But this
observation may not be "confounding" in the usual sense in that NO2 may also be an indicator of
other pollutants or pollution sources (e.g., traffic).
Of the studies evaluated, only the APHEA2 study focused specifically on the CO-mortality
association and in the process examined: (1) model sensitivity; (2) the CO-mortality C-R
relationship; and (3) potential effect modifiers of CO mortality risk estimates. The sensitivity
analysis indicated an approximate 50-80% difference in CO risk estimates from a reasonable range
of alternative models, which suggests that some model uncertainty likely influences the range of CO
mortality risk estimates obtained in the studies evaluated. The examination of the CO-mortality
concentration-response relationship found very weak evidence for a CO threshold at 0.5 mg/m3
(0.43 ppm). Finally, when examining a variety of city-specific variables to identify potential effect
modifiers of the CO-mortality relationship, the APHEA2 study found that geographic region
explained most of the heterogeneity in CO mortality risk estimates.
The results from the single-city studies are generally consistent with the multicity studies in
that some evidence of a positive association was found for mortality upon short-term exposure to
CO. However, the CO-mortality associations were often but not always attenuated when copollutants
were included in the regression models. In addition, limited evidence was available to identify cause-
specific mortality outcomes (e.g., cardiovascular causes of death) associated with short-term
exposure to CO.
The evidence from the recent multi- and single-city studies suggests that an association
between short-term exposure to CO and mortality exists, but limited evidence is available to evaluate
cause-specific mortality outcomes associated with CO exposure. In addition, the attenuation of CO
risk estimates which was often observed in copollutant models contributes to the uncertainty as to
whether CO is acting alone or as an indicator for other combustion-related pollutants. Overall, the
epidemioiogic evidence is suggestive of a causal relationship between relevant short-term
exposures to CO and mortality.
The evaluation of new epidemioiogic studies conducted since the 2000 CO AQCD (U.S. EPA,
2000, 000907) that investigated the association between long-term exposure to CO and mortality
consistently found null or negative mortality risk estimates. No such studies were discussed in the
2000 CO AQCD (U.S. EPA, 2000, 000907). The reanalysis of the American Cancer Society (ACS)
data by Jerrett et al. (2003, 087380) found no association between long-term exposure to CO and
mortality. Similar results were obtained in an updated analysis of the ACS data when using earlier
(1980) CO data, but negative associations were found when using more recent (1982-1998) data.
January 2010 2-9
-------
These results were further confirmed in an extended analysis of the ACS data. The Women's Health
Initiative (WHI) Study also found no association between CO and CVD events (including mortality)
using the mortality data from recent years (1994-1998), while the series of Veterans Cohort studies
found no association or a negative association between mean annual 95th percentile of hourly CO
values and mortality. An additional study was identified that used a cross-sectional study design,
which reported results for a study of U.S. counties that are generally consistent with the cohort
studies: positive associations between long-term exposure to PM2.5 and SO42" and mortality, and
generally negative associations with CO. Overall, the consistent null and negative associations
observed across epidemiologic studies which included cohort populations encompassing potentially
susceptible populations (i.e., post-menopausal women and hypertensive men) combined with the
lack of evidence for respiratory and cardiovascular morbidity outcomes following long-term
exposure to CO; and the absence of a proposed mechanism to explain the progression to mortality
following long-term exposure to CO provide supportive evidence that there is DOt likely to be 3
causal relationship between relevant long-term exposures to CO and mortality.
2.6. Policy-Relevant Considerations
2.6.1. Susceptible Populations
The examination of populations potentially at greater risk for health effects due to CO
exposure is an important consideration in setting NAAQS to provide an adequate margin of safety
for both the general population and sensitive populations (see Section 5.7 for a more detailed
discussion). During the evaluation of the CO literature, numerous studies were identified that
examined whether underlying factors increased the susceptibility of an individual to CO-related
health effects. These types of studies were those that included stratified analyses, examined
individuals with an underlying health condition, or used animal models of disease.
The most important susceptibility characteristic for increased risk due to CO exposure is CAD,
also known as coronary heart disease (CHD). As discussed in Section 5.7, there were approximately
13.7 million individuals with CHD in the U.S. in 2007. Persons with a normal cardiovascular system
can tolerate substantial concentrations of CO, if they vasodilate or increase cardiac output in
response to the hypoxia produced by CO. In contrast, individuals unable to vasodilate in response to
CO exposure may show evidence of ischemia at low concentrations of COHb. Many of the
controlled human exposure studies have focused on individuals with CAD, and several studies have
found that controlled exposures to CO resulting in COHb concentrations of 2-6% result in significant
decreases in time to onset of exercise-induced angina or ST-segment changes in patients with stable
angina. Epidemiologic studies found limited evidence for increased hospital admissions for ischemic
heart disease (IHD) in individuals with secondary diagnoses of dysrhythmias or congestive heart
failure (CHF). This combined evidence from controlled human exposure and epidemiologic studies
indicates that individuals with underlying cardiovascular disease, particularly CAD, are a large
population that is susceptible to increased health effects in response to exposure to ambient CO.
Additional evidence for increased CO-induced cardiovascular effects is provided by toxicological
studies that observed altered cardiac outcomes in animal models of cardiovascular disease.
Other medical conditions that have been linked to increased susceptibility to CO-induced
health effects include COPD, diabetes, and anemia. Individuals with hypoxia resulting from COPD
may be particularly sensitive to CO during submaximal exercise typical of normal daily activity. The
results available from epidemiologic and controlled human exposure studies provide preliminary
evidence that individuals with obstructive lung disease (e.g., COPD patients with underlying
hypoxia, asthmatics) may be susceptible to cardiovascular or respiratory effects due to CO exposure.
Diabetics are known to have elevated exhaled CO concentrations indicative of increased endogenous
CO production rates. In addition, some recent epidemiologic studies provide preliminary evidence
for increased associations between short-term CO exposure and ED visits and hospital admissions
for cardiovascular disease (CVD) among diabetics compared to non-diabetics, as well as associations
between short-term CO exposure and changes in HRV parameters among subjects with metabolic
syndrome, but not among healthy subjects. Increased endogenous CO production and the potential
January 2010 2-10
-------
for higher baseline COHb concentrations in individuals with diabetes, combined with the limited
epidemiologic evidence showing cardiovascular effects, suggests that diabetics are potentially
susceptible to short-term exposure to CO. Individuals with various forms of anemia experience
lowered hematocrit or produce altered forms of hemoglobin, resulting in decreased arterial O2
content; in addition, individuals with hemolytic anemia exhibit increased endogenous CO production
rates and COHb levels. This suggests that individuals with anemia who have diminished O2-carrying
capacity and/or high baseline COHb levels may be more susceptible to health effects due to ambient
CO exposure, although no studies were identified that evaluated specific CO-related health effects in
anemic individuals.
Aging alters physiological parameters that influence the uptake, distribution, and elimination
of CO. The general impact of these changes over an individual's lifetime increases the time required
for both loading and elimination of CO from the blood. As noted in the 2000 CO AQCD (U.S. EPA,
2000, 000907). changes in metabolism that occur with age, particularly declining maximal oxygen
uptake, may make the aging population susceptible to the effects of CO via impaired oxygen
delivery to the tissues. Some epidemiologic studies reported increases in IHD or myocardial
infarction (MI) HAs among older adults as compared to all-age groups or younger adults in response
to short-term exposure to CO. Older adults represent a large and growing fraction of the U.S.
population and have a higher prevalence of CAD and other cardiovascular conditions than the
general population; combined with the limited evidence available from epidemiologic studies, this
indicates that older adults are a potentially susceptible population for increased health effects due to
CO.
During gestational exposure, fetal CO pharmacokinetics differ from maternal kinetics, in part
because human fetal Hb has a higher CO affinity than adult Hb. At steady-state conditions, fetal
COHb concentrations are up to 10-15% higher on a relative basis than maternal COHb levels, and
these levels are maintained over a longer period since the half-life for fetal CO Hb is approximately
twice that of maternal COHb (7.5 h versus 4 h). Some epidemiologic studies reported higher
associations between short-term CO exposure and IHD or MI HAs among older adults as compared
to all-age groups or younger adults. Epidemiologic studies provide some evidence that CO exposure
during pregnancy is associated with changes in birth outcomes, including PTB, cardiac birth defects,
reductions in birth weight, and infant mortality in the postneonatal period. Toxicological studies
report effects in laboratory animals that lend biological plausibility to outcomes observed in
epidemiologic studies, including decrements in birth weight, reduced prenatal growth, and effects on
the heart. Toxicological evidence also exists for additional developmental outcomes which have not
been examined in epidemiologic or human clinical studies, including behavioral abnormalities,
learning and memory deficits, locomotor effects, neurotransmitter changes, and changes in the
auditory system. This evidence suggests that critical developmental phases may be characterized by
enhanced sensitivity to CO exposure.
COHb concentrations are generally higher in males than in females, and the COHb half-life is
longer in healthy men than in women of the same age. However, women experience fluctuating
COHb levels through the menstrual cycle due to variations in the endogenous CO-production rate.
Only a limited number of epidemiologic studies have examined gender differences, and found some
evidence for larger effects in males compared to females when examining the association between
short-term CO exposure and IHD HAs. The limited epidemiologic evidence combined with known
gender-related differences in endogenous CO production do not provide sufficient basis for
determining whether CO disproportionately affects males or females.
Increased altitude induces a number of physiological changes as compensatory mechanisms to
counteract the effects of decreased barometric pressure and the resulting altitude-induced hypobaric
hypoxia (HH). These changes generally increase both CO uptake and elimination, with increased
COHb levels observed in subjects at rest and decreased COHb observed in individuals exposed to
CO during exercise. In addition, baseline COHb levels increase due to increased endogenous CO
production. A controlled human exposure study observed an additive effect of CO exposure and
simulated high altitude on the reduction in time to onset of angina among a group of individuals with
CAD. Acclimatization occurs as the length of stay at high altitude increases, indicating that visitors
to high-altitude locations may have an increased risk of health effects due to CO exposure and
represent a potentially susceptible population.
Physiological changes associated with exercise tend to increase both uptake and elimination of
CO. In a controlled human exposure study, healthy subjects exposed to CO and achieving COHb
levels of-5% observed a significant decrement in exercise duration and maximal effort capability
January 2010 2-11
-------
during heavy exercise. Due to the counterbalancing effects of increased COHb formation and
elimination rates, it is unclear whether individuals engaging in light to moderate exercise represent a
population potentially susceptible to ambient CO exposure.
CO concentrations on and adjacent to heavily traveled roadways are several times higher than
concentrations measured at fixed-site monitors not located adjacent to roadways. In addition, studies
of commuters have shown that commuting time is an important determinant of CO exposure for
those traveling by car, bicycle, public transportation, and walking. Census data indicate that 17.9
million occupied homes nationwide (16.1%) are located within approximately 90 m of a freeway,
railroad, or airport, and that 5.5 million U.S. workers (5%) commute 60 min or more to work in
automobiles. This evidence for elevated on-road and near-road CO concentrations combined with
residential and commuting data indicates that the large numbers of individuals who spend a
substantial amount of time on or near heavily traveled roadways are an important population that is
potentially susceptible to increased health risks due to ambient CO exposure.
Endogenous CO production can be altered by medications or other substances, including
nicotinic acid, allyl-containing compounds (acetamids and barbiturates), diphenylhydantoin,
progesterone, contraceptives, and statins. One epidemiologic study observed an association between
short-term CO exposure and an increase in SDNN for CAD patients not taking beta blockers;
however, this association did not persist in CAD patients taking beta blockers. Other compounds
such as carbon disulfide and sulfur-containing chemicals (parathion and phenylthiourea) increase CO
following metabolism by cytochrome p450s. The p450 system may also cause large increases in CO
produced from the metabolic degradation of dihalomethanes such as methylene chloride. Minor
sources of endogenous CO include the auto-oxidation of phenols, photo-oxidation of organic
compounds, and lipid peroxidation of cell membrane lipids. Taken together, this evidence indicates
that individuals ingesting medications and other substances that enhance endogenous or metabolic
CO production represent a population that is potentially susceptible to increased health effects due to
additional exposure to ambient CO.
Overall, the controlled human exposure, epidemiologic, and toxicological studies evaluated in
this assessment provide evidence for increased susceptibility among multiple populations. Medical
conditions that increase endogenous CO production rates may also contribute to increased
susceptibility to health effects from ambient CO exposure. Although the weight of evidence varies
depending on the factor being evaluated, the clearest evidence indicates that individuals with CAD
are most susceptible to an increase in CO-induced health effects.
2.6.2. Concentration- and Dose-Response Relationships
Currently, very limited information is available in the human clinical and epidemiologic
literature regarding the CO concentration- or dose-response (C-R, D-R) relationships and the
potential existence of a CO threshold. Two human clinical studies described in the 1991 (U.S. EPA,
1991, 017643) and 2000 (U.S. EPA, 2000, 000907) CO AQCDs have evaluated the D-R relationship
between percent COHb (a measure of internal dose of CO) and onset of exercise-induced angina
among individuals with CAD. Anderson et al. (1973, 023134) exposed 10 adult men with stable
angina (5 smokers and 5 nonsmokers) for 4 h to CO concentrations of 50 and 100 ppm, which
resulted in average COHb concentrations of 2.9% and 4.5%, respectively. Both exposures
significantly decreased the time to onset of exercise-induced angina relative to room air control
(1.6% COHb). However, there was no difference in response between the two exposure
concentrations of CO. In a much larger study, 63 adults with stable angina were exposed for 1 h to
2 concentrations of CO (average exposure concentrations of 117 and 253 ppm) resulting in average
COHb concentrations in the range of 2.0-2.4% and 3.9-4.7% (Allred et al., 1989, 013018: Allred et
al., 1989, 012697: Allred et al., 1991, 011871). Relative to control (average COHb 0.6-0.7%), COHb
concentrations of 2.0-2.4% and 3.9-4.7% were observed to decrease the time required to induce ST-
segment changes indicative of myocardial ischemia by 5.1% (p = 0.01) and 12.1% (p < 0.001),
respectively. Increasing COHb concentration was similarly shown to decrease the time to onset of
exercise-induced angina. As described in Allred et al. (1989, 013018: 1989, 012697: 1991, 011871).
the observed dose-response relationship was further evaluated by regressing the percent change in
time to ST-segment change or time to angina on actual COHb concentration (0.2% -5.1%) using the
three exposures (air control and two CO exposures) for each subject. Regression analyses were
conducted separately for each individual and the averages of the intercepts and slopes across subjects
were reported. This analysis demonstrated statistically significant decreases in time to angina and
January 2010 2-12
-------
ST-segment change of approximately 1.9% and 3.9%, respectively, per 1% increase in COHb
concentration, with no evidence of a measurable threshold. The findings of Allred et al. (1989,
013018; 1989, 012697; 1991, 011871) provide evidence of a significant D-R relationship over a
range of COHb concentrations relevant to the NAAQS. While several other laboratory studies have
evaluated cardiovascular effects of CO exposure among adults with CAD, differences in study
protocols and analytical methods do not allow for an informative pooled or quantitative meta-
analysis of the D-R relationship across studies (Section 5.2.4).
Two studies in the epidemiologic literature attempted to examine the C-R relationship at the
low end of CO concentrations through a threshold analysis. Samoli et al. (2007, 098420) in their
examination of the association between short-term exposure to CO and mortality conducted an
ancillary analysis to examine the potential presence of a CO threshold. In this analysis the authors
compared city-specific models to the threshold model, which consisted of thresholds at 0.5 mg/m3
(0.43 ppm) increments. Samoli et al. (2007, 098420) then computed the deviance between the two
models and summed the deviances for a given threshold over all cities. While the minimum deviance
suggested a potential threshold of 0.43 ppm (the lowest threshold examined), the comparison with
the linear no-threshold model indicated weak evidence (p-value > 0.9) for a threshold. However,
determining the presence of a threshold at the very low range of CO concentrations (i.e., at
0.43 ppm) in this data set is challenging, because, in 7 of the 19 European cities examined, the
lowest 10% of the CO distribution was at or above 2 mg/m3 (1.74 ppm). By only using the 12 cities
in the analysis that had minimum CO concentrations approaching 0.5 mg/m3 (0.43 ppm), a limited
number of observations were examined around the threshold of interest, which subsequently
contributed to the inability to draw conclusions regarding the potential presence of a threshold with
any certainty. In addition to the time-series analyses investigating the association of CO
concentrations with hospital admissions due to CVD among Medicare enrollees, Bell et al. (2009,
193780) performed subset analyses using datasets that included only days with CO levels below
certain specified values, ranging from 1 to 10 ppm (in 1 ppm increments). When these various CO-
limit values were evaluated, there were positive associations between cardiovascular health effects
and CO concentrations at each level investigated in this study, thus providing no evidence for the
existence of a threshold. The investigators also estimated an exposure-response curve allowing a
nonlinear relationship between CO concentration and risk of CVD hospital admissions, and reported
no evidence of departure from a linear exposure-response curve.
2.7. Integration of CO Health Effects
This section summarizes the main conclusions of this assessment regarding the health effects
of CO and the concentrations at which those effects are observed. It also discusses important
uncertainties that were considered in interpreting the health effects evidence. The clearest evidence
for health effects associated with short-term exposure to CO is provided by studies of cardiovascular
morbidity. The combined health effects evidence supports a likely causal relationship for this
outcome. Controlled human exposure studies provide strong evidence of independent effects of CO
on cardiac function, with effects being observed in patients with CAD following short-term CO
exposures resulting in 2.0-2.4% COHb. Epidemiologic studies of ED visits and hospital admissions
for ischemic heart disease report consistent positive associations with additional preliminary
evidence for an increase in cardiovascular-related mortality provided by a multicity study. This
epidemiologic evidence is coherent with ischemia-related effects observed in controlled human
exposure studies. Recent toxicological evidence suggests that other mechanisms involving altered
cellular signaling may play a role in cardiovascular disease outcomes following CO exposure.
Consistent decreases in time to onset of exercise-induced angina, along with ST-segment
changes indicative of myocardial ischemia, were observed in individuals with CAD following
controlled CO exposures resulting in COHb concentrations of 2-6%, with no evidence of a threshold
at the lowest levels tested. Modeling results described in Chapter 4 indicated that increases of ~1%
COHb are possible with exposures of several ppm CO, depending on exposure duration and exercise
level. Baseline COHb levels are <1% in healthy individuals, with higher endogenous CO production
observed in individuals with certain medical conditions. The volunteers who participated in these
studies were diagnosed with moderate to severe CAD, although they may not be representative of
the most sensitive individuals in the population. Variability in activity patterns and severity of
January 2010 2-13
-------
disease combined with daily fluctuations in baseline COHb levels may influence the critical level of
increased COHb which leads to adverse cardiovascular effects in a particular individual. In addition,
arterial COHb is transiently higher than venous COHb for several minutes following a rapid increase
in inhaled CO concentration. Transient increases in ambient CO have the potential to elevate COHb
to higher levels in the coronary arteries than in other vascular beds, possibly increasing heart CO
levels and cardiovascular symptoms in diseased individuals. Quantification of the magnitude of
effects at ambient concentrations from the results of controlled human exposure studies is difficult
due to the gap between ambient concentrations and the higher concentrations used in these studies
(i.e., experimental studies have not been conducted at levels within the range of current maximum
ambient concentrations).
Epidemiologic studies consistently show associations between ambient CO concentrations and
cardiovascular endpoints other than stroke, particularly hospitalizations and ED visits for ischemic
heart disease, MI, and angina. These effects are robust to adjustment for copollutants. Since the
heterogeneity of endpoints in these studies does not lend itself to a quantitative meta-analysis, a
forest plot was used to summarize the results. Figure 2-1 presents unadjusted health effect estimates
from U.S. and Canadian studies of short-term CO exposure and CVD hospitalizations, along with
mean and 99th percentile concentrations during the study periods. Table 2-2 summarizes the range of
mean and 99th percentile concentrations observed in the studies presented in Figure 2-1. This
evidence for ischemia-related outcomes is coherent with effects observed in controlled human
exposure studies, although uncertainty regarding the extent of reduced O2 delivery to tissues
following exposure to ambient CO concentrations contributes to the uncertainty in quantitative
interpretation of effect estimates.
Study Location Mean*^ Avg Time Lag Outcome/
Group
Svmonsetal. (2006, 091258) Baltimore, MD 0.4(2.3) 8-h max 0-3 CHF
Szvszkowicz (2007. 193793) Montreal, Can 0.5 24-havg 0 IHD
0 IHD, 65+
Wellenius et al. (2005. 087483) Pittsburgh. PA 1.03(1.6-3.9) 24-havg 0 CHF
Bell etal. (2009. 193780) 126 U.S. Counties 1.3(1. 2-22.1)" 1-hmax 0 CVD, 65+
Effect Estimate (95% Cl)
Non-Stroke CVD Endpoint
^ 1 • ^>
i •
i -•-
i.
Fung etal. (2005, 074322) Windsor, Can 1.3 1-hmax 0-2 CVD, <65 *
0-2 CVD, 65+ -i-. —
Metzger etal. (2004,044222) Atlanta.GA 1.8(5.5-5.9) 1-hmax 0-2 IHD L^-
0-2 CHF
Tolbert etal. (2007,090316) Atlanta.GA 1.6(5.3-5.4) 1-hmax 0-2 CVD
_^_
Peel etal. (2007,090442) Atlanta.GA 1.8(5.5-5.9) 1-hmax 0-2 IHD i-.-
0-2 CHF ,-•-
Koken etal. (2003,049466) Denver, CO 0.9(1.3-2.0) 24-havg 3 CHF
Mann etal. (2002,036723) California, US 2.07(1.3-15.9) 8-h max 0-3 IHD
i •
i»
0-3 IHD.sCHF i-.-
0-3 IHD, sARR i*
Linn etal. (2000,002839) LosAnaeles.CA 1.5(1.1-8.3) 24-havg 0 Ml '-•-
0 CHF U-
0 CVD ' *
Villeneuve etal. (2006, 090191) Edmonton, Can 0.8 24-hava 0-2 IS, 65+
0-2 CIS, 65+
0-2 HS, 65+
1< Stroke
• — ,—
» i
Wellenius etal. (2005, 088685) Multicity, US 1.02** (1.2-7.1) 24-havg 0-2 IS, 65+ ,-*•
0-2 HS, 65+ -w-
Linn etal. (2000,002839) Los Angeles, CA 1.5(1.1-8.3) 24-havg 0 IS
*CO Mean and (99th percentile) concentrations in ppm. ** Median f
a Mean concentrations (ppm) are presented as reported in the references. Studies are arranged in approximate order of
increasing mean concentration, with adjustment for different averaging times using a 4:3:2 ratio of 1 -h max, 8-h max, 0-
and 24-h avg values based on relationships derived from the national AQS CO distribution.
b 99th percentile concentrations (ppm) were obtained from the AQS database for durations and locations chosen to match
those of the U.S. studies. When multiple monitors were available at the study location, the range of monitor specific 99th
percentile concentrations during the study period is presented. No 99th percentile data are presented for Canadian studies.
c For the Bell etal. (2009, 193780) study, the concentration statistics represent the 1999-2005 average of daily county-specific
values. The central estimate is the median county-average across the U.S. The 99th percentile values represent the
counties with the lowest and highest 99th percentile concentrations. Additional cause-specific effect estimates adjusted for
N02are presented in Section 5.2.1.
i •
i
i 1 1 1
9 1.0 1.1 1.2 1.3
Excess Risk
Figure 2-1. Excess risk estimates from epidemiologic studies of short-term CO exposure and
CVD hospitalizations along with author-reported mean and AQS-derived 99th
percentile CO concentrations. See the footnotes related to concentration data.
January 2010
2-14
-------
Table 2-2. Range of mean and 99th percentile concentrations (ppm) in US and Canadian studies of
short-term CO exposure and CVD hospitalizations. See the notes in Figure 2-1 for
sources of concentration data.
Metric
Mean
99th percentile
1-h daily max
1.3-1.8
1.2-22.1
8-h daily max
0.4-2.07
1.3-15.9
24-h avg
0.5-1.5
1.1-8.3
Additional studies provide evidence for associations between CO exposure and other health
outcomes, including CNS effects, birth outcomes and developmental effects, respiratory effects, and
mortality. Although inconsistent results were reported in controlled human exposure studies on
neural and behavioral effects, toxicological studies in rodents found that perinatal exposure to CO
can have a range of effects on the adult nervous system. This combined evidence is suggestive of a
causal relationship between both short- and long-term CO exposure and CNS effects. Differences in
fetal pharmacokinetics from those of the mother result in fetal COHb levels that are up to 10-15%
higher than maternal COHb levels. Epidemiologic studies provide some evidence that CO exposure
during pregnancy is associated with changes in birth outcomes, including increased risk of PTB,
cardiac birth defects, small reductions in birth weight, and infant mortality in the postneonatal
period. This evidence, in conjunction with developmental effects observed in toxicological studies, is
suggestive of a causal relationship between long-term exposure to CO and birth and developmental
effects.
Evidence regarding the effect of short-term exposure to CO on respiratory morbidity is
suggestive of a causal relationship, based on associations observed in epidemiologic studies and
animal toxicological studies which indicate the potential for an underlying biological mechanism,
while the evidence on long-term exposure and respiratory morbidity is inadequate to infer the
presence of a causal relationship.
An evaluation of epidemiologic studies that examined the effect of short-term exposure to CO
on mortality provides evidence that is suggestive of a causal relationship. Epidemiologic studies that
examined mortality and long-term exposure to CO reported consistent null associations, which,
combined with the lack of respiratory and cardiovascular morbidity or a proposed biological
mechanism for mortality following long-term exposure, indicate that there is not likely to be a causal
relationship between long-term exposure to CO and mortality.
Issues such as exposure error and isolation of the independent effect of CO as a component of
a complex air-pollutant mixture contribute to uncertainty in interpreting the results of epidemiologic
studies. Studies published since the 2000 CO AQCD (U.S. EPA, 2000, 000907) have provided
insight regarding the nature and magnitude of these uncertainties. Exposures in near-road and on-
road microenvironments are likely to be higher than concentrations measured at community-oriented
regulatory monitors, which may result in over- or underestimation of the magnitude of ambient
exposure for some individuals. Individuals who are susceptible to CO-induced health effects, such as
those with CAD, may be at additional risk when experiencing elevated on-road CO concentrations.
However, as discussed in Section 2.3 and in more detail in Section 3.6, spatial variability in absolute
concentration will not introduce error into time-series epidemiologic studies if the concentrations are
correlated in time. A recent study by Sarnat et al. (2009, 180084) found that associations between
CO and cardiovascular ED visits were similar when based on different monitors within an urban
center, regardless of monitor location or distance to population, while an association was not
observed when using a rural monitor outside the urban area. This may have been related to the
similarity of driving patterns and peak rush-hour times in the urban center as compared to the area
around the rural monitor, where the temporal driving patterns were different. Simulations of ambient
and nonambient exposures to a nonreactive pollutant indicated that nonambient exposure has no
effect on the association between ambient exposure and health outcomes for the case where ambient
and nonambient concentrations are independent, although variability is introduced. Nonambient
exposure to CO is not expected to be temporally correlated with ambient CO concentrations, and
therefore nonambient CO will not act as a confounder in epidemiologic associations with ambient
CO. Exposure error is not likely to affect the magnitude of the population-averaged effect estimates
observed in epidemiologic studies, although it would tend to widen the confidence intervals.
January 2010 2-15
-------
Epidemiologic studies consider the effects of CO as a component of a complex mixture of air
pollutants that varies across space and time, with moderate to high correlations observed between
CO concentrations and those of other combustion-related pollutants. On-road vehicle exhaust
emissions are a nearly ubiquitous source of combustion pollutant mixtures that include CO, NO2,
and PM2.5, and these emissions are the most important contributor to ambient CO in near-road
locations. Correlations between CO and NO2 reported in epidemiologic studies of short-term
exposure to CO generally ranged from 0.3 to 0.86, with correlations reported in US studies ranging
from 0.55 to 0.86. Correlations between CO and PM2.5 reported in all studies ranged from 0.17 to
0.74, with correlations in US studies ranging from 0.43 to 0.62. This complicates the quantitative
interpretation of effect estimates in these studies to apportion the relative extent to which CO at
ambient concentrations is independently associated with cardiovascular or other effects, and the
extent to which CO acts as a marker for the effects of another combustion-related pollutant or mix of
pollutants.
As summarized in Tolbert et al. (2007, 090316), when toxicological or controlled human
exposure studies of two correlated pollutants provide evidence that each exerts an independent health
effect, two-pollutant models may be appropriate to adjust the effect estimate for each pollutant for
confounding by the other pollutant. PM2 5 and NO2 have each been linked to cardiovascular health
effects in epidemiologic studies. In two-pollutant models in which one of the pollutants is linked to
the measured outcome and the other is a surrogate for the first pollutant, the copollutant model can
help identify which is the better predictor of the effect, particularly if the etiologically linked
pollutant is measured with more error than the second pollutant. Uncertainty is introduced in the size
of the effect estimate and the portion of the effect size represented by each of the coefficients in the
model by correlation between the two pollutants and by differential exposure measurement error.
Since the spatial variability of CO is a larger contributor to measurement error than for other more
homogenously distributed pollutants such as PM2 5, robustness of CO effect estimates indicates that
CO is the better predictor of effects in copollutant models. Although this complicates quantitative
interpretation of the effect estimates reported in epidemiologic studies, the epidemiologic evidence
for cardiovascular morbidity summarized in this assessment indicates that CO associations generally
remain robust in copollutant models (Figure 5-6 and Figure 5-7), which, combined with the
consistency of effects observed across studies, the coherence of epidemiologic health outcomes with
effects observed in controlled human exposure studies, and the emerging evidence on the potential
role for cell signaling effects at low tissue CO concentrations, supports an independent effect of
short-term CO exposure on cardiovascular morbidity. This combined evidence supports a
determination that the relationship between CO and cardiovascular morbidity is likely causal, while
still recognizing that CO is a component of a mixture of combustion-related pollutants.
Evidence from controlled human exposure and epidemiologic studies indicates that individuals
with underlying CVD, specifically CAD, are an important susceptible population at increased risk of
health effects due to ambient CO. Potentially susceptible populations include those with other
underlying diseases, including anemia, obstructive lung disease, or diabetes; older adults and fetuses
during critical phases of development; commuters and those living near heavily traveled roadways;
visitors to high-altitude locations; and individuals ingesting medications and other substances that
enhance endogenous or metabolic CO production. Limited evidence is available from controlled
human exposure studies of CAD patients indicating a statistically significant inverse relationship
between COHb concentration and time to ST segment change or time to exercise-induced angina.
Epidemiologic analyses investigating the exposure-response relationship for mortality and
cardiovascular morbidity did not find evidence for a departure from linearity or a threshold for CO
effects.
The new evidence reviewed in this ISA builds upon the health-effects evidence summarized in
the 2000 CO AQCD (U.S. EPA, 2000, 000907). with many new epidemiologic studies adding to the
body of evidence showing associations between acute cardiovascular effects and CO measured at
ambient monitors. Controlled human exposure studies reviewed both in this ISA and the 2000 CO
AQCD (U.S. EPA, 2000, 000907) show definitive evidence of cardiovascular effects among
individuals with CAD following short-term CO exposure, resulting in COHb concentrations as low
as 2.0-2.4%. Emerging toxicological evidence points to the potential role for CO in modes of action
not directly related to COHb's role in O2 delivery. In evaluating the several epidemiologic studies
available at the time that reported associations between ambient CO and cardiovascular effects, the
2000 CO AQCD (U.S. EPA, 2000, 000907) considered those findings to be inconclusive for multiple
reasons, including: questions regarding the consistency of the results among studies; the ability of
January 2010 2-16
-------
community fixed-site monitors to represent spatially variable ambient CO concentrations and
personal exposures; the small expected increase in COHb due to ambient CO concentrations; the
lack of biological plausibility for health effects to occur at such COHb levels, even in diseased
individuals; the potentially greater impact of non-ambient exposure on COHb; and the possibility
that ambient CO is serving as a surrogate for a mixture of combustion-related pollutants. Some of
these uncertainties remain and complicate the quantitative interpretation of the epidemiologic
findings, particularly regarding the biological plausibility of health effects occurring at COHb levels
resulting from exposures to ambient CO concentrations measured at AQS monitors. New research
summarized in this assessment reduces several of the other uncertainties noted in the 2000 CO
AQCD (U.S. EPA, 2000, 000907) and demonstrates the lack of influence of nonambient exposure on
effect estimates in epidemiologic studies, the consistency of epidemiologic study results, their
robustness in copollutant models, and the coherence of ischemia-related outcomes with evidence
from controlled human exposure studies. This consistent and coherent evidence from epidemiologic
and human clinical studies, along with biological plausibility provided by the role of CO in limiting
O2 availability, is sufficient to conclude that a causal relationship is likely to exist between relevant
short-term CO exposures and cardiovascular morbidity.
January 2010 2-17
-------
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, 120: 949-955. 193780
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
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). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment
Report (TAR) of the Intergovernmental Panel on Climate Change [Houghton, J.T.,Y Ding, D.J. Griggs, M. Noguer,
PJ. van der Linden, X. Dai, K. Maskell, and C.A. Johnson (eds.)]. Cambridge, UK and New York, NY, USA:
Intergovernmental Panel on Climate Change, Cambridge University Press. 156587
IPCC (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report (AR4) of the Intergovernmental Panel on Climate Change . Cambridge, UK and New York, NY:
Intergovernmental Panel on Climate Change, Cambridge University Press. 092765
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 Curr Iss, 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
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 JEpidemiol, 165: 625-633. 090442
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).
January 2010 2-18
-------
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 spatial variability 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. 193747
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. JExpo Sci Environ Epidemiol, 17: S29-S35. 090316
U. S. EPA (1991). Air quality criteria for carbon monoxide. U. S. Environmental Protection Agency. Research Triangle Park,
NC. EPA/600/8-90/045F. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000554R.txt. 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
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
January 2010 2-19
-------
Chapter 3. Source to Exposure
3.1. Introduction
This chapter reviews concepts and findings in atmospheric sciences and exposure assessment
that provide a foundation for the detailed presentation of evidence of CO-related health effects in
subsequent chapters and for a causality finding regarding climate forcing effects of CO. Section 3.2
provides an overview of the primary and secondary sources of CO as well as the atmospheric
chemistry involved in the production and removal of CO by oxidation processes. Section 3.3
provides a description of climate forcing caused directly and indirectly by CO. Descriptions of CO
measurement methods, monitor siting requirements, and monitor locations are presented in Section
3.4. Ambient CO concentrations and their spatial and temporal variability are characterized in
Section 3.5. The background concentrations of CO useful for risk and policy assessments informing
decisions about the NAAQS, referred to as policy-relevant background (PRB) concentrations, are
also presented in Section 3.5. Factors related to human exposure to ambient CO, and their
implications for epidemiologic studies, are discussed in Section 3.6. Finally, a summary and
conclusions of the chapter are presented in Section 3.7.
3.2. CO Sources, Emissions, and Chemistry
3.2.1. Direct CO Emissions
CO is formed primarily by incomplete combustion of carbon-containing fuels and
photochemical reactions in the atmosphere. In general, any increase in fuel O2 content, burn
temperature, or mixing time in the combustion zone will tend to decrease production of CO relative
to CO2. CO emissions from large fossil-fueled power plants are typically very low since the boilers
at these plants are tuned for highly efficient combustion with the lowest possible fuel consumption.
Additionally, by allowing time for the furnace flue gases to mix with air and be oxidized by OH to
CO2 in the hot gas stream before the OH concentrations drop as the flue gases cool, the CO-to-CO2
ratio in these emissions is shifted toward CO2.
Figure 3-1 lists CO emissions totals in tons segregated by individual source sectors in the U.S.
for 2002, which is the most recent publicly available CO emissions data meeting EPA's data quality
assurance objectives. In the U.S., direct CO emissions data are tracked in the National Emissions
Inventory (U.S. EPA, 2006, 157070). a composite of data from various sources including industries
and state, tribal, and local air agencies, and from the Biogenic Emissions Inventory System (BEIS).
NEI data are collected for all states, the District of Columbia, the U.S. territories of Puerto Rico and
Virgin Islands, and some of the territories of federally recognized American Indian nations. Different
data sources use different data collection methods, most of which are based on empirical estimates
and engineering calculations rather than measurements. Most fuel combustion and industrial sources,
for example, estimate their CO emissions using EPA-approved emission factors, as do on-road and
non-road mobile source emitters where models (MOBILE6, MOVES, NONROAD) are available to
calculate inventories (U.S. EPA, 2006, 157070). The NEI includes fires of anthropogenic and natural
origin. Anthropogenic fires include structural fires, agricultural fires, prescribed burning, and slash
burning; forest wildfires are considered to be of natural origin. Estimates of direct CO emission from
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).
January 2010 3-1
-------
soil are calculated by the EPA using the BEIS model. Although these estimates are generated using
well-established approaches, uncertainties are inherent in the emission factors and models used to
represent sources for which emissions have not been directly measured. These uncertainties vary by
source category, season, and region. Discussion of uncertainties is provided in subsequent
paragraphs related to mobile sources, the largest source category.
Nationally, on-road mobile sources in the NEI constituted more than half of total CO
emissions in 2002, or -60.6 MT of-116.8 MT total, which includes anthropogenic and biogenic
emissions reported in the NEI and the BEIS (http://www.epa.gov/ttnchie 1 /emch/biogenic). High
concentrations of CO can often occur in areas of heavy traffic. In metropolitan areas in the U.S., for
example, as much as 75% of all CO emissions came from on-road vehicle exhaust in the 2002 NEI
(U.S. EPA, 2006, 157070). When the emissions from incomplete combustion of fuels powering
non-road mobile sources were included, all mobile sources accounted for -80% of total CO
emissions in the U.S. in 2002 (Figure 3-1).
CO emissions from internal combustion engines vary substantially with ambient temperature
and operating conditions. Substantial light-duty gasoline vehicle CO emissions occur during the cold
start before the catalyst is warmed up. Most emission tests such as the Federal Test Procedure (FTP)
which includes a cold start portion and is used to certify that vehicles meet EPA emission standards,
are conducted at about 75°F. Lower ambient temperatures result in increased CO emissions because
spark ignition engines are required to run richer air: fuel ratios for longer periods of time, and also
because the time before the catalyst is warmed up increases compared to the time for catalyst warm-
up occurring at 75°F (U.S. EPA, 2006, 199897). Thus, in addition to the vehicle CO emissions
standards EPA implemented starting with the 1968 model year, EPA has also implemented a cold
temperature CO emission standard for light-duty gasoline vehicles and trucks at 20°F that phased in
for 40% of the new fleet in the 1994 model year, 80% for the 1995 model year, and 100% with the
1996 and succeeding model years. The emission standard of 10 g/mile results in a reduction of about
20-30% in CO emissions at 20° F (57 FR 3188-31923 July 17, 1992). Increased vehicle CO
emissions can also occur under conditions such as high rates of acceleration, rapid speed
fluctuations, heavy-vehicle load demands (such as occur while pulling a trailer or going up a steep
hill), and use of air-conditioning. Such driving conditions were not originally fully reflected in the
FTP. EPA has issued a Supplemental Federal Test Procedure (SFTP) to control excess CO emissions
under these conditions. These regulations were phased in for the 1998-2000 model years (61 FR
54852-54906 October 22, 1996). Moreover, the gasoline-powered spark ignition engines that
predominate in light-duty on-road vehicles have higher uncontrolled CO emission rates than other
combustion sources because they typically operate closer to the stoichiometric air-to-fuel ratio, have
relatively short residence times at peak combustion temperatures, and have very rapid cooling of
cylinder exhaust gases. By contrast, the diesel-powered engines that predominate in heavy-duty on-
road vehicles and in off-road and non-road fixed combustion sources have much lower engine-out
CO emissions than do the spark-ignition engines because the diesels typically operate at very high
air-to-fuel ratios, which promote mixing oxygen and fuel, thus improving carbon burn.
January 2010 3-2
-------
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
SOILS
FUEL COMB. OTHER
WASTE DISPOSAL & RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC. UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM & RELATED INDUSTRIES
CHEMICAL & ALLIED PRODUCT MFC
STORAGE & TRANSPORT
MISCELLANEOUS
SOLVENT UTILIZATION
60,600,000
22,700,000
5,820,000
3,550,000
| 1,590,000
| 1,270,000
] 987,000
657,000
490,000
357,000
284,000
118,000
25,600
1,660
18,500,000
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 from the NEI and the
BEIS. The "fires" category has been extracted from Tier 3 miscellaneous
categories of the NEI, and biogenic fires are those attributed to forest wildfires.
The "soils" category comprises the BEIS data. The "roadway vehicles" and "non-
road vehicles" categories have been renamed here for clarity.
Figure 3-2 shows present and historical CO emissions from the traditionally inventoried
anthropogenic source categories: (1) fuel combustion, which includes emissions from coal-, gas-,
and oil-fired power plants and industrial, commercial, and institutional sources, as well as residential
heaters (e.g., wood-burning stoves) and boilers; (2) industrial processes, which include chemical
production, petroleum refining, metals production, and industrial processes other than fuel
combustion; (3) on-road vehicles, which include cars, trucks, buses, and motorcycles; and (4) non-
road vehicles and engines, such as farm and construction equipment, boats, ships, snowmobiles,
aircraft, locomotive, and the two-stroke engines found in lawnmowers, chainsaws, and other small
gasoline-powered equipment. Using these NEI data, trends in the national CO emissions can be
computed and compared over time. So, for example, the national-scale estimated anthropogenic CO
emissions decreased 35% between 1990 and 2002. The trend in Figure 3-2 demonstrates that
controls in the on-road vehicle sector have produced nearly all the national-level CO reductions
since 1990. (Data are presented here for 1990 and from 1996-2002 because only 1990 data have
been updated to be comparable to the more recent inventories made since 1996.)
January 2010
3-3
-------
160
1 120
H—I
1100
E 80
I 60
•| 40
m 20
0
Fuel combustion
Other industrial processes
On-road vehicles
Nonroad vehicles and engines
'90
'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.
With the exception of this downward trend resulting from emissions controls, anthropogenic
CO emissions demonstrate less interannual variability than biogenic emissions (Bergamaschi et al.,
2000, 192377). Several recent reports using both ambient concentrations and fuel-based emissions
estimates have explored this annual-to-decadal emissions decrease in anthropogenic CO in finer
detail; they include Harley et al. (2001, 193922: 2005, 088154). Parrish et al. (2002, 052472).
Parrish (2006, 090352). Pollack et al. (2004, 184461). and Mobley et al. (2005, 194008). The
consistent conclusion from those investigations has been that annual average U.S. on-road vehicle
CO emissions have decreased at a rate of-5% per year since the early 1990s. This can be seen from
Figure 3-2 as well. Additional analyses by Harley et al. (2005, 088154) and Parrish (2006, 090352)
were also consistent with the suggestion in Pollack et al. (2004, 184461) that the EPA MOBILE6
vehicle emissions model (http://www.epa. gov/otaq/m6.htm) now overestimates vehicle CO
emissions by a factor of ~2. Field measurements by Bishop and Stedman (2008, 194670) were in
accord with Parrish's (2006, 090352) findings that the measured trends of CO and NOX
concentrations from mobile sources in the U.S. indicated that modeled CO emission estimates were
substantially too high. Hudman et al. (2008, 191253) found that the NEI overestimated
anthropogenic CO emissions by 60% for the eastern U.S. during the period July 1-August 15, 2004
using aircraft observations of CO from the International Consortium for Atmospheric Research on
Transport and Transformation (ICARTT) campaign (Fehsenfeld et al., 2006, 190531) and results
from a tropospheric chemistry model (GEOS-Chem)(Figure 3-3).
January 2010
3-4
-------
350
300
250
200
15°
100
50
0
Observed
Simulated (NE199 emissions)
Simulated (NEI99 emissions reduced by 60%)
ISO
190
200 210
Day of the Year
220
230
Source: Reprinted with Permission of the Amercian Geophysical Union from Hudman et al. (2008,1912531
Figure 3-3. 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). Overestimation near day 200 is due to model misplacement of a large
Alaskan/Canadian biomass burning plume.
Improvements in emissions technologies not correctly represented in MOBILE emissions
models have been suggested as one cause for this discrepancy. For example, Pokharel et al. (2002,
052473; 2003, 053740) demonstrated substantial decrements in the CO fraction of tailpipe exhaust in
several U.S. cities, and Burgard et al. (2006, 193222) documented improvements in emissions from
heavy-duty on-road diesel engines. The Motor Vehicle Emission Simulator (MOVES) model has
been designed to address some of the largest errors in the MOBILE model. It was released in final
form in December 2009 (http://www.epa.gov/otaq/models/moves/index.htm).
Estimates of non-anthropogenic CO emissions are made using the BEIS model with data from
the Biogenic Emissions Landcover Database (BELD) and annual meteorological data. National
biogenic emissions, excluding fires, were estimated to contribute 5%, or -5.8 MT, of total CO
emissions from all sources in 2002. Biogenic wildfires in 2002 added another 12%, or -14.1 MT, to
the national CO emissions total and were responsible for 76.1% of all CO emissions estimates from
fires. This is shown in Figure 3-1 using the NEI and BEIS data. Geogenic emissions of CO, also
included in this inventory, include volcanic gases released from molten rock in the Earth's mantle.
Mixing ratios of dissolved CO in this rock vary in a range from 0.01 to 2% as a function of the rock
stratum surrounding the volcano and other geologic conditions. This high variability and infrequent
though often violent release mean geogenic CO measurements are very difficult to make with
precision, though on non-local scales the magnitude of their contribution is small relative to
anthropogenic sources. Photodecomposition of organic matter in oceans, rivers, lakes, and other
surface waters, and from soil surfaces also releases CO (Goldstein and Galbally, 2007, 193247).
However, soils can act as a CO source or a sink depending on soil moisture, UV flux reaching the
soil surface, and soil temperature (Conrad and Seiler, 1985, 029520). Soil uptake of CO is driven by
anaerobic bacteria (Inman et al., 1971, 010972). Emissions of CO from soils appear to occur by
abiotic processes, such as thermodecomposition or photodecomposition of organic matter. In
general, warm and moist conditions found in most soils favor CO uptake, whereas hot and dry
conditions found in deserts and some savannas favor the release of CO (King, 1999, 002828).
Biomass burning consists of wildfires and the intentional burning of vegetation to clear new
land for agriculture and population resettlement; to control the growth of unwanted plants on pasture
land; to manage forest resources with prescribed burning; to dispose of agricultural and domestic
waste; and as fuel for cooking, heating, and water sterilization. Globally, most wildfires may be
ignited directly as the result of human activities, leaving only 10-30% initiated by lightning
January 2010
3-5
-------
(Andreae, 1991, 078147). However, because fire management practices suppress natural wildfires,
the buildup of fire fuels increases the susceptibility of forests to more severe but less frequent fires in
the future. Thus there is considerable uncertainty in attributing the fraction of wildfire emissions to
human activities because the emissions from naturally occurring fires that would have been present
in the absence of fire suppression practices are not known.
Biomass burning also exhibits strong seasonality and interannual variability (van der Werf et
al., 2006, 157084). with most biomass burned during the local dry season. This is true for both
prescribed burns and wildfire. The unusually warm and dry weather in central Alaska and western
Yukon in the summer of 2004, for example, contributed to the burning of 11 million acres there.
These fires, the largest on record for this region, produced CO emissions easily tracked by the
Measurement of Pollution in the Troposphere (MOPITT) instrument on NASA's Terra satellite
(Figure 3-4). The high CO concentration measured by MOPITT coincided with the surface location
of fires tracked using aerosol plumes identified by the Moderate Resolution Imaging
Spectroradiometer (MODIS) also on Terra. Subsequent modeling by Pfister et al. (2005, 093009)
snowed that the CO contribution from these fires in July 2004 was 33.1 (± 5.5) MT that summer, or
in the range of the total U.S. anthropogenic CO emissions during the same time. The smoldering
phase of combustion yields higher CO emissions than the flaming phase. Using controlled
combustion chamber experiments, Lobert et al. (1991, 029473) found that with a wide variety of
vegetation types, on average, 84% of the CO from biomass fires was produced during the smoldering
phase and 16% during the flaming phase of combustion.
CO emissions data for EPA's 10 administrative Regions in the U.S., depicted in Figure 3-5,
show a more nuanced view of the national concentrations and trends described above. Net
anthropogenic CO emissions were estimated to have declined in all EPA Regions between 1990 and
2002, with the largest decrease (10.8 MT) occurring in Region 9 and the smallest (1.3 MT) in
Region 10.
At state and local levels, CO emissions from on-road mobile sources or from fires can
dominate in different locations across the U.S. Figure 3-6 illustrates this variability with CO state-
level emissions totals and selected county totals in 2002 for Colorado (Annex A includes analogous
data for Alaska, Utah, Massachusetts, Georgia, California, and Alabama). In Colorado, emissions
from fires and on-road vehicles were nearly equal: ~0.9 MT from fires and ~1.1 MT from on-road
vehicles. Emissions sources varied strongly across counties, however, with urban Denver County
dominated by on-road vehicle emissions at 71% and rural Garfield County dominated by fire
emissions at 67%.
January 2010 3-6
-------
-150
-120
-90 -60
Longitude
-30
100
120 130 140 150 160 170 180 190 200 ppbv
Source: Reprinted with Permission of the American Meteorological Society from Fishman et al. (2008,1939271
Figure 3-4. CO concentrations centered at ~3,000 m above sea level measured by the
MOPITT sensor on the Terra satellite for the period July 15-23, 2004, during
intense wildfires in Alaska and the Yukon.
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
Source: U.S. EPA (2008, 157076)
Figure 3-5. Trends in subnational CO emissions in the 10 U.S. EPA Regions for 1990 and
1996-2002.
January 2010
3-7
-------
Carbon Monoxide Ernisaons in 2002 (Tons per Square Mile)
] 108 - 5.09
] 5.45 - 16.26
1-G.96 - TB7.7Q
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
SOILS
FUEL COMB. OTHER
WASTE DISPOSAL S RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC. UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM S RELATED INDUSTRIES
CHEMICAL S ALLIED PRODUCT MFC
STORAGE STRANSPORT
MISCELLANEOUS
SOLVENT UTILIZATION
] 21,OX
NON-ROAD EQUIPMENT
WASTE DISPOSAL & RECYCLING
METALS PROCESSING
OTHER INDUSTRIAL PROCESSES
CHEMICAL & ALLIED PRODUCT MFC
SOLVENT UTILIZATION
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FUEL COMB. OTHER
FUEL COMB. ELEC
OTHER INDUSTRIAL PROC
PETROLEUM & RELATED INDUS
CHEMICAL SALLIED PRODUC
STORAGE STRANS
SOLVENT UTILIZATION
] 3,700
| 3,200
Source:: U.S. EPA (2006,1570701
Figure 3-6. CO emissions density map and distributions for the state of Colorado and for
selected counties in Colorado in 2002, from the NEI and the BEIS. The "fires"
category has been extracted from Tier 3 miscellaneous categories of the NEI, and
biogenic fires are those attributed to wildfires. The "soils" category comprises
the BEIS data. The "roadway vehicles" and "non-road vehicles" categories have
been renamed here for clarity.
January 2010
3-8
-------
3.2.2. Secondary CO Emissions and Associated Chemistry
Oxidation of anthropogenic and biogenic VOCs constitute important secondary sources of CO.
For example, Hudman et al. (2008, 191253) determined that oxidation of isoprene and other biogenic
VOCs contributed 9.1 MT of atmospheric CO (with isoprenes contributing 7.1 MT), and oxidation
of anthropogenic VOCs contributed another 2.0 MT of CO emissions during the period July 1-
August 15, 2004, for the eastern U.S. In contrast, direct anthropogenic CO emissions were estimated
to be 5.1 MT for this time period and location. Hence, secondary biogenic formation was found to be
a more important source of CO emissions than direct anthropogenic activities for the study period.
Hudman et al. (2008, 191253) noted that biogenic CO emissions were highest in the southeastern
U.S., where isoprene emissions are also greatest. These estimates were obtained using aircraft
measurements from the ICARTT campaign (Fehsenfeld et al., 2006, 190531) and estimates from the
GEOS-Chem model (Bey et al., 2001, 051218). configured as described by Hudman et al.(2007,
089474).
Secondary CO production occurs by photooxidation of methane (CH4) and other VOCs,
including nonmethane hydrocarbons (NMHCs) in the atmosphere and organic molecules in surface
waters and soils. CH4 oxidation is summarized in this reaction sequence:
CH4 + OH^ CH3 + H2O
CH3 + O2(+M)^> CH3O2 (+ M)
CH3O2 + 7V<9 -» CH3O + NO 2
CH3O2 + HO 2 -» CH3OOH + O2
CH3O + <92 -» CH2O + HO 2
or
or CH2O + OH^> HCO + H2O
HCO + O2^CO + HO 2
Reaction 3-1
where M is a reaction mediator that is neither created nor destroyed and stabilizes the reaction
product.
Photolysis of formaldehyde (CH2O) proceeds by two pathways. The first produces molecular
hydrogen (H2) and CO with a reaction yield of 55% in conditions of clear skies and low zenith
angles; the second yields a hydrogen radical (H) and the formyl radical (HCO). HCO then reacts
with O2 to form hydroperoxy radical (HO2; OH and HO2 together are termed HOX) and CO.
Reaction of methyl peroxy radical (CH3O2) with HO2 radicals to form methyl hydroperoxide
(CH3OOH) is also operative, especially in low oxides of nitrogen (NO+NO2=NOX) conditions.
Heterogeneous removal of the partially water-soluble intermediate products, such as CH3OOH and
CH2O, will decrease CO yields from CH4 oxidation.
While oxidation of CH2O nearly always produces CO and some small quantities of formic acid
(CH2O2) in the reaction of CH2O with HO2 (not shown here), oxidation of acetaldehyde (CH3CHO)
does not always yield two CO molecules. Reaction of CH3CHO with OH can yield acetyl radicals
(CH3CO) which then will participate with O2in a termolecular recombination reaction to form
peroxy acyl radicals, which then can react with nitric oxide (NO) to form CH3 and CO2; or the
peroxy acyl radicals can react with NO2 to form peroxy acetyl nitrate (PAN), CH3CO3NO2. In this
way, one carbon atom is oxidized directly to CO2 without passing through CO. The yield of CO from
these pathways depends on the OH concentration and the photolysis rate of CH3CHO, as well as on
the abundance of NO, since peroxy acyl radicals also will react with other odd hydrogen radicals like
HO2.
Estimating the CO yield from oxidation of hydrocarbons (HCs) larger than CH4 requires
computing the yields of CH2O, CH3CHO, CH3CO, and analogous radicals from oxidation of the
parent molecules. Moreover, the extent of heterogeneous removal of soluble intermediate products
also affects oxidation of more complex HCs. However, the detailed gas-phase kinetics for many HCs
January 2010 3-9
-------
with more than a few carbons is still unknown. This is especially the case for several important
classes of VOCs, including the aromatics, biogenic HCs including isoprene, and their intermediate
oxidation products like epoxides, nitrates, and carbonyls. Mass-balance analyses performed on
irradiated smog chamber mixtures of aromatic HCs indicate that only about one-half of the carbon is
in the form of compounds that can be identified. In addition, reactions like the oxidation of terpenes
that produce condensable products are also significant because these reactions produce secondary
organic aerosols, thereby reducing the potential yield of CO. The CO yield from oxidation of CH4,
for example, is -0.9 on a per carbon basis (Kanakidou and Crutzen, 1999, 011760). Yields from
other compounds range from <0.1 for anthropogenic alkanes (Altshuller, 1991, 192375) to ~0.9 for
ethane; yields from other compounds are given in Table 3-1 taken from Kanakidou and Crutzen
(1999.011760).
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 C02yield:
-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
Other nonmethane hydrocarbons: 0.7
Source: Adapted with Permission of Elsevier Ltd. from Kanakidou and Crutzen (1999, 011760)
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 ~2.4><10~13 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 -1 yr characteristic time scale for mixing
between the hemispheres and because northern hemisphere CO emissions are higher due to
anthropogenic activity (Khalil and Rasmussen, 1990, 012443). a large gradient in concentrations
January 2010
3-10
-------
exists between the hemispheres (Edwards et al., 2004, 199889). 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 time of CO in urban areas when assuming a diel-average
OH concentration of 3><106/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 larger scales. OH
concentrations are orders of magnitude lower in indoor environments, and so CO will generally not
be affected by indoor air reactions.
3.3. CO Climate Forcing Effects
Recent data do not alter the current well-established understanding of the role of urban and
regional CO in continental and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
Intergovernmental Panel on Climate Change (IPCC) (2001, 156587; 2007, 092765). CO is a weak
direct contributor to greenhouse warming because its fundamental absorption band near 4.63 (im is
far from the spectral maximum of Earth's longwave radiation at ~10 (im. Sinha and Toumi (1996,
193747) estimated the direct RF of CO computed for all-sky conditions at the tropopause, which is
the IPCC's preferred form for the calculation (Forster et al., 2007, 092936). to be 0.024 W/m2 based
on an assumed change in CO mean global concentration from 25 to 100 ppb since preindustrial
times. The direct RF value similarly projected by Sinha and Toumi (1996, 193747) if the mean
global background concentration were to increase from 25 to 290 ppb was 0.057 W/m2.
However, because reaction with CO is the major sink for OH on a global scale, increased
concentrations of CO can lead to increased concentrations of other trace gases whose loss processes
also involve OH chemistry. Some of those trace gases, CH4 and O3 for example, absorb infrared
radiation from the Earth's surface and contribute to the greenhouse effect directly. Others, including
hydrochlorofluorocarbons (HCFCs), methyl chloride, and methyl bromide, can deplete stratospheric
O3, increasing the surface-incident UV flux.
This indirect effect of CO on stratospheric O3 concentrations is opposite in sign to the effect of
CO on O3 in the troposphere where CO reacts in a manner similar to other VOCs in the presence of
NOX and UV to create O3; see the detailed description of O3 formation from VOCs and NOX in the
2006 O3 ISA (U.S. EPA, 2006, 088089). Because CO's chemical lifetime is longer than those of the
VOCs most important for O3 formation on urban and regional scales and because CO oxidation has
one-to-one stoichiometry (whereby one molecule of CO converts only one molecule of NO to NO2),
CO has a significantly lower O3 forming potential than other VOCs in the troposphere. Carter (1998,
192380) computed a maximum incremental reactivity for CO of 0.06 g O3 for 1 g CO, as compared
to reactivities of total on-road vehicle exhaust emissions typically in the range of 3 to 4 g O3 per
g VOC. However, because the total mass of CO emissions is substantially greater than those of the
other VOCs with higher carbon numbers and faster reactivities, CO can contribute significantly to O3
formation even though its photochemical processing is slow. Using data from instrumented models,
including that of Jeffries (1995, 003055). the NRC (1999, 010614) estimated, for example, that CO
can contribute 15-25% of the total O3 forming potential of gasoline exhaust emissions, although this
estimate shows strong regionality. The contribution of CO to urban and regional O3 concentration is
often <10% owing to its very slow reactivity on these scales and to locally variable radical
concentration ratios.
Emissions of CO and the other O3 precursors, nonmethane VOCs (NMVOCs) and NOX, affect
the oxidizing capacity of the atmosphere largely by perturbing HOX concentrations. From a climate
perspective, this HOX perturbation chiefly affects the CH4 T and production of O3 in the troposphere.
Changes in the concentration of O3 and hence in its RF occur mainly in the time of a few months.
However, Prather (1996, 193195) showed that changes in CH4 concentration and its RF extend to the
"primary mode" timescale of troposphere chemistry of about 14 yr (Derwent et al., 2001, 047912;
Wild et al., 2001, 193196). The primary mode timescale of CH4 is in part determined by the positive
feedbacks in the CH4-OH-CO system in which even low concentration additions of CH4 produce
additional CO through oxidation by OH. That additional CO then further decreases atmospheric OH
concentrations when OH oxidizes it to CO2. The resulting decreased OH concentration then further
increases the CH4 T (Daniel and Solomon, 1998, 193235; Isaksen and Hov, 1987, 019490).
Atmospheric CH4 concentrations since 1750 have increased by more than a factor of 2, giving an RF
January 2010 3-11
-------
of-0.6 W/m2 (Forster et al., 2007, 092936). Roughly 25% of the global mean tropospheric CO is
produced by CH4 oxidation (Wuebbles and Hayhoe, 2002, 044159). Using a 2-D global model on a
coarse grid, Wang and Prinn (1999, 011758) showed that increasing CO and CH4 concentrations
leading to decreased OH concentrations can extend the CO T as well as the CH4 T. Wang and Prinn
(1999, 011758) varied the CO emissions and other model inputs and parameters in a matrix of
simulations that showed, with increased or even constant 20th century CO concentrations, the CO T
was increased by more than 50% in 100 yr. However, Wang and Prinn (1999, 011758) stated that
their simulation omitted NMHCs and therefore likely underestimated CO concentrations while
under- or overestimating hydroxyl radical concentrations. Likewise, low spatial resolution of the
model likely incurred additional error in the solution.
CH4 is long-lived and, in general, well mixed in the atmosphere. However, the reaction of CH4
and OH, and hence the CH4 T, is governed by the behavior and location of emissions of the short-
lived gases, including CO, VOCs, and NOX. This produces high regional variability and uncertainty
in the concentrations and RFs from CO and its related climate forcing gases (Berntsen et al., 2006,
193244; Fuglestvedt et al., 1999, 047431). NOX, for example, can produce effects on the combined
indirect RF opposite in direction to those of CH4 since under most global background conditions an
increase in NOX increases the global average OH concentration and decreases CH4 T and RF
(Berntsen et al., 2005, 193241: Wild et al., 2001, 193196). Wild et al. (2001, 193196) also showed
that emissions changes in CO have effects opposite in sign to those of NOX because increases in CO
act to depress OH concentrations and that the combined effect of CO and NOX emissions yields a
positive RF. The results of this study underscore the need to consider the combined effects of
pollutants emitted from similar sources.
Using the 3-D global chemistry model MOZART-2 (Horowitz et al., 2003, 057770). Naik
et al. (2005, 193194) simulated changes in global tropospheric O3 concentrations and RF resulting
from differing reductions in emissions of NOX alone or a combination of NOX, CO, and NMHCs in
nine regions of the Earth. For the reductions in Europe, North America, and Southeast Asia, reducing
CO and NMHCs in addition to reducing NOX lowered the spatial inhomogeneity of the O3
concentration and RF because CO has a longer lifetime than NOX.
Wild et al. (2001, 193196) used the University of California-Irvine chemical transport model
(CTM) (Wild and Prather, 2000, 052402) driven by the NASA GISS II general circulation model
(Rind and Lerner, 1996, 193750) to compute changes in O3 concentrations and RF from regional
emissions of NOX and CO. Changes in O3 and CH4 result from increases in global surface NOX
emissions alone and run for 10-yr periods produced negative net RFs, ranging from -0.2 W/m in
East Asia to -0.5 W/m2 in the Tropics owing to the long-term interdependencies in the CO-CH4-NOX
system described above. When global CO emissions were increased by an 11 MT pulse for 1 yr
together with the same 1-yr pulsed NOX surface emissions and run again for a 10-yr period, the
global net RF rose to 1.7 W/m2 with an estimated 20% uncertainty based on the spatial variability
and short-term reactivity of O3 (Wild et al., 2001, 193196).
Determining whether several species' T and RF will increase or decrease in response to pulsed
or step-wise emissions of the short-lived O3 precursor species (NMVOC, CO, and NOX) is
complicated by its global location with respect to the O3 production response surface. See the
description of the O3 production response surface and its dependence on NOX and radical
concentrations in the 2008 NOX ISA (U.S. EPA, 2008, 157073) for additional details. Fiore et al.
(2002, 051221; 2008, 193749) found that O3 is closely coupled with CH4 and that their relationship
is influenced by regional variation in NOX concentrations. Using the weighted average results from
12 3-D global chemistry models exercised for the IPCC Third Assessment Report (2001, 156587).
Wigley et al. (2002, 047883) confirmed that increases in CO and VOC emissions increased the O3
RF both directly and indirectly through the CH4 effects described above. Furthermore, Wigley et al.
(2002, 047883) demonstrated that NOX emissions produced a mix of direct and indirect increases in
RF, mostly dominated by the direct effects for all modeled scenarios. Wigley et al. (2002, 047883)
concluded that tropospheric O3 RF influences were larger than CH4 influences and that the short-
lived reactive gases produced 60-80% of that forcing, with the remainder coming from CH4. Given
these chemical interdependencies, calculations of an indirect RF for any of these short-lived O3
precursor species are most often made for all of the most important ones together.
Figure 3-7 illustrates model estimates of the combined RF of increased short-lived O3
precursor species relative to long-lived GHGs, aerosols, and other changes (Forster et al., 2007,
092936). The combined effect of increased CH4, CO, NMVOC, and NOX emissions for the period
1750-2005 has produced tropospheric O3 concentrations associated with a net RF of-0.4 W/m2 with
January 2010 3-12
-------
±50% uncertainty based on Shindell et al. (2005, 080129). Indirect effects of CO through the GHGs
O3, CH4, and CO2 were estimated to contribute RF ~0.2 W/m2, which is more than a factor of 2
larger than the indirect effect of the shorter-lived NMVOCs on the same three GHGs (Forster et al.,
2007, 092936). Of the indirect effects on these three GHGs from CO emissions, the O3-related
component was the largest, accounting for approximately one-half of the forcing (Forster et al.,
2007, 092936). In comparison, CO2 contributed a direct forcing of 1.6 ± 0.2 W/m over this time
period.
Integrated RF estimates over longer time horizons may indicate the future climate effects of
present-day emissions. Modeled integrated 20-yr and 100-yr time horizon RFs are presented by
Forster et al. (2007, 092936) in Figure 3-8 for year 2000 emissions of short-lived and long-lived
GHGs. The integrated RF for CO was estimated to be -0.2 W/m2-yr with -50% uncertainty. It can be
seen that the integrated RF of CO2 is much smaller for the 20-yr horizon because the lifetime of CO2
perturbations is roughly 150 yr. As a result, the RF related to short-lived CO is -25% of that for CO2
for the 20-yr horizon (-0.7 W/m2-yr) but only -7% of that for longer-lived CO2 over a 100-yr time
horizon (-2.4 W/m2-yr). This indirect forcing is just slightly lower than the RF of year 2000 black
carbon emissions from fossil fuel and biomass burning on the same horizons according to this
assessment.
It is also possible to compute individual contributions to the integral RF from CO based on
separate emissions sectors. Unger et al. (2009, 193238) used the NASA GISS model for Physical
Understanding of Composition-Climate Interactions and Impacts (G-PUCCINI) (Shindell et al.,
2006, 193751). Unger et al. (2009, 193238) divided the 1995 global anthropogenic CO emissions
total of 933.3 MT/yr into sectors for on-road transport (ORT) and power generation (PG), and then
separated contributions from each of these sectors for the U.S. and other large geographic regions of
the Earth. ORT CO emissions in the U.S. were 84.1 MT/yr; PG CO emissions were 0.55 MT/yr out
of the total U.S. anthropogenic CO emissions of 112.5 MT/yr. Unger et al. (2009, 193238) concluded
from analysis of 7-yr runs that the CO indirect CH4 effects (that is, the CO effects through CH4
changes as described above) in the 1995 emissions run were -0.004 W/m2 for the global ORT and
-0.022 W/m2 for the global PG. In the U.S., the indirect CH4 RF was positive at +0.009 W/m2
because the positive effects on CH4 T from the CO emissions dominated over the negative effects
from NOX through OH. This RF fraction from indirect CH4 is approximately the same as the direct
O3 RF from ORT in the U.S., 0.010 W/m2. Because the PG sector emits NOX but less CO relative to
the ORT, the indirect CH4 RF from the U.S. PG was not dominated by the positive CO effects and
remained a net negative at -0.006 W/m2 (Unger et al., 2009, 193238). The authors acknowledged
some uncertainty in this relationship related to the influence of NOX emissions on CO
transformation, but no quantification of uncertainty in these modeled estimates was provided.
These gross emissions sectors can also be subdivided to demonstrate more clearly the
localized chemical interdependencies of the CO-CH4-NOX system. Fuglestvedt et al. (2008, 193242)
used the Oslo CTM2 model to simulate effects from all emissions and changes in all transportation
subsectors from 1850-2000. Fuglestvedt et al. (2008, 193242) found that global transport has been
responsible for -15% of the total anthropogenic CO2 RF and -15% of the total anthropogenic O3 RF.
Of the total O3 RF, the largest contributor was the shipping sector at 0.03 W/m2, because its high
NOx-to-CO and NOx-to-VOC ratios produced OH increases and hence CH4 decreases in regions of
naturally low NOX. For the shipping segment of the transport sector, the high NOX emissions there
reduced the CH4 T but increased O3. The global mean effect from these two was small and still
smaller than the direct negative effect from SO42" aerosols. In the on-road segment of global
transportation, emissions of CO and VOCs together with NOX produce an O3 RF larger than the
negative RF from CH4.
January 2010 3-13
-------
Black carbon
SO2
Organic carbon
Mineral dust
Aerosols
Aircraft
Land use
Solar irradiance
-0.5 0 0.5
Radiative Forcing (Wnrr2)
1 1.5
Source: Reprinted with Permission of Cambridge University Press from Forster et al. (2007, 092936)
Figure 3-7. Components of RF in 2005 resulting from emissions since 1750. (S) and (T)
indicate stratospheric and tropospheric changes, respectively.
January 2010
3-14
-------
Long-lived
greenhouse gases
Nitrate
Organic
carbon (FF)
S02(FF)-I
>x V Short-lived gases
Aerosols and aerosol
precursors
Cloud albedo —l
Long-lived J
greenhouse gases
Nitrate
Organic
carbon (FF)
Cloud albedo —I
CFCs —| |— SF6 + PFCs + MFCs
-HCFCs
SO, (FF)-I
co, NMVOC + NOX Short-lived gases
Aerosols and aerosol
precursors
-2
-1012
Integrated radiative forcing (W nrr2yr1)
Source: Reprinted with Permission of Cambridge University Press from Forster (2007, 0929
Figure 3-8. Integrated RF of year 2000 emissions over 20-yr and 100-yr time horizons. The
values provided refer to global annual emissions, but effects are expected to vary
regionally for short-lived gases.
Caution is warranted in interpreting RF estimates. RF values are global model calculations
using the assumption that global climate sensitivities are equal for all forcing mechanisms, whether
CO2, SO42" and other aerosols, or the short-lived gases like CO (Berntsen et al., 2005, 193241;
Berntsen et al., 2006, 193244). That assumption is under challenge now by GCM results using
regionalized RF values separately for different forcing mechanisms and with CO2, O3, and solar
input changes (Joshi et al., 2003, 193752). Joshi et al. (2003, 193752) found that global climate
system sensitivities from non-CO2 RF varied by ±30% compared to CO2 RF. Other GCM
experiments by Lelieveld et al. (2002, 190361). Rotstayn and Penner (2001, 193754). Menon et al.
(2002, 155978). and Kristjansson (2002, 045282) have indicated that regionally changing RF can
induce changes in large-scale circulation patterns that control the regionalized cycles of flooding and
drought through disruptions in regional temperature and hydrologic cycles. However, such
regionalized patterns resulting from GCM experiments are so uncertain and so widely variable
across models that even the sign of these regionalized changes can vary with model type and any of
January 2010
3-15
-------
the models' unconstrained assumptions (Berntsen et al., 2006, 193244). Even with such uncertainty
and variability, though, the consensus of the climate community is that the climate effects of changes
to emissions of the long- and especially the short-lived pollutants, including CO, very likely depend
on location.
Global warming potential (GWP) is a widely used relative measure of the potential effect of
different emissions on climate, usually defined as the time-integrated RF from an instantaneous
pulsed release of 1 kg of a trace gas relative to the effects from a pulsed release of 1 kg of CO2.
Because the greenhouse warming effects from CO are nearly completely indirect and because CO
concentrations are spatially heterogeneous, neither the IPCC nor EPA computes direct GWPs for
CO, just as they do not for tropospheric O3, NO, NO2, or VOCs (U.S. EPA, 2008, 184463). IPCC
does estimate indirect GWP for CO using the integrated indirect RF (Forster et al., 2007, 092936).
However, the indirect GWP values evaluated and summarized by IPCC are global and cannot reflect
effects of localized emissions or emissions changes, making the values for the short-lived species
NMVOC, CO, and NOX more uncertain than the values for the long-lived, well-mixed species as a
result of the OH chemistry described above. Moreover, urban- and regional-scale oxidation of CO to
CO2 under current atmospheric conditions proceeds very slowly, and IPCC considers production of
CO2 through this pathway to be double counting of CO effects (Forster et al., 2007, 092936).
Variation in the CO GWP range of estimates can be attributed to the unusually large heterogeneity in
model type and form, pulsed or stepped emissions increase, time-horizon unit, and integral or
differential indirect effects in several combinations (with or without NOX emissions changes,
including or excluding CO2 effects).
Even with such variability in methods and tools, the CO GWPs have been largely in agreement
for approximately 10 yr. The IPCC estimated the indirect GWP of CO to be 1.9 (Forster et al., 2007,
092936). Daniel and Solomon (1998, 193235) used a global box model for changes through CH4 and
O3 effects from pulsed CO emissions and estimated a CO GWP exclusive of the effect through CO2
to be between 1 and 4.4. Using the STOCHEM CTM, Derwent et al. (2001, 047912) estimated a
pulsed emissions CO GWP, again exclusive of effects through CO2, to be 1.5. Johnson and Derwent
(1996, 193192) had previously computed and integrated GWP of 2.1 for the CH4 and O3 effect from
a step-wise emissions change using a 2-D and a 100-yr time horizon. Derwent et al. (2001, 047912)
and Collins et al. (2002, 044156) subsequently differentiated that integral for each effect and
reported GWP for step-wise CO emissions changes on a 100-yr time horizon of 1.0, 0.6, and 1.6
through the effects on CH4, O3, and CO2, respectively. Berntsen et al. (2005, 193241) used the model
LMDz v3.3 (Hauglustaine et al., 2004, 193191) to compute 100-yr GWP values for pulsed CO
emissions through all indirect effects to be 1.9 as resolved for Europe and 2.4 for Asia,
demonstrating the strong regionality in the indirect effects from these short-lived precursors. Most
recently, Shindell et al. (2009, 201599) compared GWPs separately for CO, CH4, and NOX with and
without interactions with aerosols. When including direct and indirect radiative effects related to
interaction of CO with aerosols, the GWP for CO was estimated to rise from a range of 1-3 to a
range of 3-8.
3.4. Ambient Measurements
3.4.1. Ambient Measurement Instruments
For enforcement of the air quality standards set forth under the Clean Air Act, EPA has
established provisions in the Code of Federal Regulations (CFR) under which analytical methods can
be designated as federal reference methods or federal equivalent methods (FRM or FEM,
respectively). Measurements for determinations of NAAQS compliance must be made with FRMs or
FEMs. As of August 2009, 20 automated FRMs and no FEMs had been approved for CO
(http://www.epa.gov/ttn/amtic/criteria.html).
All EPA FRMs for CO operate on the principle of nondispersive infrared (NDIR) detection
and can include the gas filter correlation (GFC) methodology. NDIR is an automated and continuous
method based on the specific absorption of infrared radiation by the CO molecule. Most
commercially available analyzers incorporate a gas filter to minimize interferences from other gases
and operate near atmospheric pressure. NDIR is based on the physics of CO's characteristic infrared
January 2010 3-16
-------
absorption near 4.63 um. NDIR methods have several practical advantages over other techniques for
CO detection in that they are not sensitive to flow rate changes, require no wet chemicals, are
reasonably independent of ambient air temperature changes, are sensitive over wide concentration
ranges, and have fast response times. An extensive and comprehensive review of NDIR, GFC, and
alternative, non-FRM techniques for CO detection, including tunable diode laser spectroscopy, gas
chromatography, mercury liberation, and resonance fluorescence, was made for the 2000 CO AQCD
(U.S. EPA, 2000, 000907). and the reader is directed there for additional information. The
description here is limited to a brief outline of the FRM NDIR and GFC techniques.
GFC spectroscopy analyzers are used most frequently now in documenting compliance with
ambient air standards. A GFC monitor has all of the advantages of an NDIR instrument and the
additional advantages of smaller size, no interference from CO2, and very small interference from
water vapor. During operation, air flows continuously through a sample cell. Radiation from the
infrared source is directed by optical transfer elements through two main optical subsystems: (1) the
rotating gas filter; and (2) the optical multipass (sample) cell. The beam exits the sample cell through
an interference filter, which limits the spectral passband to a few of the strongest CO absorption
lines. Detection of the transmitted radiation occurs at the infrared detector. The gas correlation cell is
constructed with two compartments, one filled with 0.5 atm CO, and a second with pure nitrogen gas
(N2). Radiation transmitted through the CO is completely attenuated at the wavelengths where CO
absorbs strongly. The radiation transmitted through the N2 is reduced by coating the exit window of
the cell with a neutral attenuator so that the amounts of radiation transmitted by the two cells are
made approximately equal in the passband that reaches the detector. In operation, radiation passes
alternately through the two cells as they are rotated to establish a signal modulation frequency. If CO
is present in the sample, the radiation transmitted through the CO is not appreciably changed,
whereas that through the N2 cell is changed. This imbalance is linearly related to CO concentrations
in ambient air.
Specifications for CO monitoring are designed to help states demonstrate whether they have
met compliance criteria; operational parameters required under 40 CFR 53 are provided in Table 3-2.
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
limit of detection (LOD) is sufficient for demonstration of compliance, where the LOD is set at three
times an instrument's noise level when analyzing a zero air sample to ensure that reported signals are
in response to actual ambient CO concentrations. However, with ambient CO levels now routinely at
or below 1 ppm, there is greater uncertainty in the monitoring data because a large percentage is
below the LOD. For this reason, a new generation of ambient CO monitors has been designed for
measurements below 0.5 ppm, with LOD = 0.04 ppm. Additionally, CO measurements at
concentrations below 0.5 ppm are needed to support additional objectives, such as validating the
inputs to CTMs, improving estimates of low-concentration CO exposure, and assessing differences
between CO levels in urban and rural areas, because background CO concentrations are on the order
of 0.1 ppm. Effective LOD is influenced by instrumental noise and drift and by the amount of water
vapor in the air. Recent improvements in the instruments' optical components and dehumidification
of the air stream help to reduce the amount of noise and drift in the CO measurements. Newer GFC
instruments have been designed for automatic zeroing to minimize drift (U.S. EPA, 2000, 000907).
January 2010 3-17
-------
Table 3-2. Performance specifications for analytical detection of CO, based on 40 CFR Part 53.
Parameter Specification
Range 0-50 ppm
Noise 0.5 ppm
LOD 1.0 ppm
Interference equivalent
Each interfering substance
Total interfering substances
±1 .0 ppm
1 .5 ppm
Zero drift
12h
24 h
±1 .0 ppm
±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
Currently, 24 models of CO monitors are in use; the models are listed in Annex A, Table A-l.
Among them, 20 are older NDIR instruments listed to have an LOD of 0.5 ppm, and 4 are GFC
instruments listed to have an LOD of 0.04 ppm. States do not routinely report the operational LOD,
precision, and accuracy of the monitors to EPA's Air Quality System (AQS). When the monitored
value is below the LOD, some states report the raw monitored data, while others report the
concentration as 50% of the LOD (0.25 ppm for high-LOD instruments and 0.02 ppm for low-LOD
instruments) when reported data are below the LOD. Among several of the older instruments still in
use (Federal Reference Method codes 008, 012, 018, 033, 041, 050, 051, and 054), performance
testing has shown effective LODs of 0.62-1.05 ppm, with 24-h drift ranging from 0.044-0.25 ppm
and precision ranging from 0.022-0.067 ppm at 20% of the upper range limit of the instrument
(Michie et al., 1983, 194043). Among newer GFC instruments, manufacturer-declared LODs range
from 0.02-0.04 ppm, with 24-h zero drift varying between 0.5% within 1 ppm and 0.1 ppm and
precision varying from 0.5% to 0.1 ppm.
Comparison of older and newer monitors with LOD = 0.5 ppm and 0.04 ppm, respectively,
calls attention to several data quality issues with the older monitors; Figure 3-9 illustrates this point
with data from two co-located monitors with LODs of 0.5 ppm and 0.04 ppm, in Charlotte, NC.
First, the data appearing below the LOD of 0.5 ppm for the older monitor comprise 58% of the data
obtained by that monitor. In contrast, no data from the 0.04 ppm LOD monitor are reported below
0.04 ppm. Below 0.5 ppm, observations obtained with the older monitor are on average more than
five times higher than those from the newer monitor. Second, the data from the older monitor are
reported in units of 0.1 ppm, as seen in the lower resolution of the data with respect to the x-axis.
Last, it is possible from the data that the older monitor exhibits some upward drift, since newer
models have automatic zeroing functions. Above 0.5 ppm, the slope of the scatterplot is 0.95,
suggesting that readings from the older monitor are on average 5% higher than those from the newer
monitor. The median data are 0.4 ppm for the older monitor and 0.24 ppm for the newer monitor.
However the mean from the older monitor is 0.5 ppm, in contrast with 0.330 ppm for the newer
monitor. The 99th percentile is 1.8 ppm for the older monitor, in contrast with the newer monitor,
whose 99th percentile level is 1.485 ppm.
January 2010 3-18
-------
3 -
o
I 2f
o
O
JJ1.5H
•O
O
.C
1 1
0.5 -
O Below Method 054 LOD
• Above Method 054 LOD
1.5 2 2.5 3
Method 054 CO Concentration (ppm)
3.5
Figure 3-9. Scatterplot comparing data from co-located monitors in Charlotte, NC. Data from
Method 054 are from an older model (Thermo Electron Model 48C, Waltham, MA)
with LOD = 0.5 ppm, while data from Method 593 are from a newer instrument
(Teledyne API Model 300EU, San Diego, CA) with LOD = 0.04 ppm. Above the
Method 054 LOD, the methods vary linearly as: [Method 593] = 0.95[Method 054] -
0.20 (R2 = 0.88, n = 6990); below the Method 054 LOD, the regression changes to
[Method 593] = 0.19[Method 054] + 0.16 (R2 = 0.07, n = 9856).
3.4.2. Ambient Sampling Network Design
3.4.2.1. Monitor Siting Requirements
Minimum monitoring requirements for CO were revoked in the 2006 revisions to ambient
monitoring requirements (71 FR 61236, October 17, 2006). This action was made to allow for
reductions in measurements of CO and some other pollutants (SO2, NO2, and Pb) where measured
levels were well below the applicable NAAQS and air quality problems were not expected. CO
monitoring activities have been maintained at some State and Local Air Monitoring Stations
(SLAMS), and these measurements of CO using FRM are required to continue until discontinuation
is approved by the EPA Regional Administrator. CO monitors are typically sited at the following
spatial scales (40 CFR Part 58 Appendix D):
• Microscale: Data represent concentrations within a 100 m radius of the monitor. For CO,
microscale monitors are sited 2-10 m from a roadway. Measurements are intended to
represent the near-road or street canyon environment.
January 2010
3-19
-------
• Middle scale: Data represent concentrations averaged over areas defined by 100-500 m
radii. Measurements are intended to represent several city blocks.
• Neighborhood scale: Data represent concentrations averaged over areas defined by
0.5-4.0 km radii. Measurements are intended to represent extended portions of a city.
In 2007, there were 376 CO monitors reporting values to the EPA Air Quality System (AQS)
database. Where CO monitoring is ongoing, 40 CFR Part 58 requires at least one CO monitor to
capture maximum levels in a given region. This requirement is met with a monitor situated at the
CFR-defined microscale distance from the side of a roadway for CO. Microscale monitor locations
also have sample inlets mounted at 3 ± 0.5 m above ground level, unlike the monitors sampling for
larger scales, whose inlet heights can vary between 2 and 15m. For the CFR-defmed neighborhood
scale monitoring, the minimum monitor distance from a major roadway is directly related to the
average daily traffic counts on that roadway to ensure that measurements are not substantially
influenced by any one roadway. For example, the minimum distance of a neighborhood scale CO
monitor from a roadway with an average daily traffic count of 15,000 vehicles per day is 25 m, while
the minimum distance is 135 m for a roadway with an average daily traffic count of 50,000 vehicles
per day. Occasionally, CO monitors are sited at urban (covering areas of 4-50 km) or regional
(covering areas of tens to hundreds of km) scale. More detail on siting requirements can be found in
40 CFR Part 58 Appendices D and E.
In addition to monitoring for determining compliance with the NAAQS, EPA is currently in
the process of implementing plans for a new network of multipollutant stations called National Core
(NCore) that is intended to meet multiple monitoring objectives. A subset of the SLAMS network,
NCore stations are intended to address integrated air quality management needs to support long-term
trends analysis, model evaluation, health and ecosystem studies, as well as the more traditional
objectives of NAAQS compliance and Air Quality Index reporting. The complete NCore network,
required to be fully implemented by January 1, 2011, will consist of approximately 60 urban and 20
rural stations and will include some existing SLAMS sites that have been modified for the additional
measurements. Each state will contain at least one NCore station, and 46 of the states plus
Washington, DC, will have at least one urban station. CO will be measured using 0.04 ppm LOD
monitors at all sites, as will SO2, NO, and NOy1; surface meteorology will also be measured at
NCore sites. The advantage to the NCore strategy is that time-resolved, simultaneous measurements
of multiple pollutants will be obtained at each site. The disadvantage is that the NCore 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
Figure 3-10 depicts the distribution of the 376 regulatory CO monitors operating in the U.S. in
2007. Data from 291 of the 376 CO monitors operating year-round at 290 sites in the years
2005-2007 met the data completeness criteria for inclusion in the multiyear ambient data analyses
for this assessment. Completeness criteria require that data be collected for 75% of the hours in a
day, 75% of the days in a quarter, and 3 complete quarters for all 3 yr; criteria for Region 10 were
relaxed to 2 complete quarters because it contains Alaska. The greatest density of monitors is in the
CSAs for Los Angeles, and San Francisco, CA; and along the Mid-Atlantic seaboard. Monitors are
also located in regions where biomass burning is more prevalent, such as Anchorage, AK, but not all
of these monitors report values from all seasons of all years. The number of monitors per sampling
scale is provided in Table 3-3, and locations of monitors with nearby roadway types and traffic
counts are provided in Annex A, Tables A-2 through A-7, for each monitoring scale. Twenty-four
percent of the monitors meeting completeness criteria are categorized as "Null", meaning that no
scale has been identified for those monitors. Furthermore, given the overlap between scales
regarding the type of road at which the monitor is sited, it is possible that scale has been
misclassified for some of the monitors.
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 PM10-2.5, O3, SO2, CO, NO/NOY, wind speed, wind direction, relative humidity, and ambient temperature.
January 2010 3-20
-------
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
Figure 3-10 also shows the locations of the newer 0.04 ppm LOD CO monitors throughout the
U.S in 2007, indicated by blue and purple triangles. The newer monitors included in the analysis are
located in: Baton Rouge, LA; Boston, MA; Charlotte, NC; Dallas, TX; Decatur, GA; Houston, TX;
Portland, OR; Presque Isle, ME; San Jose, CA; and rural locations within Georgia and South
Carolina. Other 0.04 ppm LOD monitors not meeting completeness criteria for the 2005-2007
analysis were located in: Beltsville, MD; Cedar Rapids, IA; Davenport, IA; Des Moines, IA;
Nederland, TX; Northbrook, IL; Plant City, FL; Seattle, WA; Thomaston, CT; Tulsa, OK; Westport,
CT; and rural locations in Maryland and Wisconsin. A listing of 0.04 ppm and 0.5 ppm LOD
monitors meeting completeness criteria by state for 2005-2007 is provided in Annex A, Table A-8.
Eleven metropolitan regions were chosen for closer investigation of monitor siting based on
their relevance to the health studies assessed in subsequent chapters of this ISA and to demonstrate
specific points about geospatial distributions of CO emissions and concentrations. These regions
were: Anchorage, AK; Atlanta, GA; Boston, MA; Denver, CO; Houston, TX; Los Angeles, CA; New
York City, NY; Phoenix, AZ; Pittsburgh, PA; Seattle, WA; and St. Louis, MO. Core-Based Statistical
Areas (CBSAs) and Combined Statistical Areas (CSAs), as defined by the U.S. Census Bureau
(http://www.census/gov/). were used to determine which counties, and hence, which monitors, to
include for each metropolitan region.1 As an example, Figure 3-11 through Figure 3-14 display CO
monitor density with respect to population density (for total population and elderly adults aged 65
and over) for the Denver and Los Angeles CSAs (Annex A, Figures A-7 through A-22 show
analogous plots for the other nine metropolitan regions). Figure 3-18 and Figure 3-21 and additional
figures in Annex A show the locations of CO monitors for the 11 CSAs/CBSAs in relation to major
roadways, including interstate highways, U.S. highways, state highways, and other major roadways
required for traffic network connectivity. In the examples shown for Denver and Los Angeles, the
monitors were typically located near high population density neighborhoods within the CSA. The
Los Angeles CSA monitors appear to be distributed fairly evenly across the city of Los Angeles,
while the Denver CSA had three monitors in the city center and two in the suburbs of the Denver
CSA. Regional background sites were not included on the maps unless they lay within the
CSA/CBSA.
1 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
January 2010 3-21
-------
CO Monitor Locations in United States in 2007
}
A CO Monitors - included in analysis, low LOD, 2007
•» CO Monitors - included in analysis,high LOD, 2007
A CO Monitors - not included in analysis, low LOD, 2007
A CO Monitors - not included in analysis, high LOD, 2007
~ CSWCBSA
Figure 3-10. Map of CO monitor locations in the U.S. in 2007. Locations are indicated with
triangles: blue and red triangles show locations of the sites used in data analysis
for this assessment; purple and green 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.
January 2010
3-22
-------
I Kilometers
0 5 10 20 30 40
2005 Population Density
| ^J Denver CO Monitors (5 km buffer)
Population per Sq Km
^^| 0-67
| 68-135
136-673
674-1347
| 1348 - 3364
^H 3365-13456
0 1530 60 90
I Kilometers
120
Figure 3-11. Map of CO monitor locations with respect to population density in the Denver, CO
CSA, total population. The circles indicate 5 km buffers around the monitors.
January 2010
3-23
-------
I Kilometers
0 5 10 20 30 40
2005 Population Density
I ^J Denver CO Monitors (5 km butter)
Population > 65 per Sq Km
^B 0-26
| 27 - 53
54 - 263
264-525
526-1313
^^| 1314-5251
01530 60 90
I Kilometers
120
Figure 3-12. Map of CO monitor locations with respect to population density in the Denver, CO
CSA, age 65 and older. The circles indicate 5 km buffers around the monitors.
January 2010
3-24
-------
I Kilometers
0510 20 30 40
2005 Population Density
I | Los Angeles CO Monitors (5 km buffer)
Population per Sq Km
^H 272 -542
543-2711
2712-5422
5423- 13556
• 13557-54222
0 3060 120 180
I Kilometers
240
Figure 3-13. Map of CO monitor locations with respect to population density in the Los
Angeles, CA CSA, total population. The circles indicate 5 km buffers around the
monitors.
January 2010
3-25
-------
I Kilometers
0510 20 30 40
2005 Population Density
I | Los Angeles CO Monitors (5 km buffer)
Population > 65 per Sq Km
^H 39 -76
77 - 382
383 - 767
768-1911
• 1912-7645
0 3060 120 180
I Kilometers
240
Figure 3-14. Map of CO monitor locations with respect to population density in the Los
Angeles, CA CSA, age 65 and older. The circles indicate 5 km buffers around the
monitors.
January 2010
3-26
-------
Ambient monitors for CO and other criteria pollutants are located to monitor compliance
rather than population exposures. However, CO monitors submitting data to the AQS are often used
for exposure assessment. For this reason, data are presented here to assess population density in the
vicinity of CO monitors. Table 3-4 and Table 3-5 show the population density around CO monitors
for the total population and for elderly adults aged 65 and over for each CSA/CBSA. The percentage
of population within specific radii of the monitors for each city was, for the most part, similar
between the total and elderly populations. In the cases of Anchorage, Denver, Phoenix, and St. Louis
however, the percentage of the elderly population within given radii of the monitors was
considerably different compared with the total population. Between-city disparities in population
density were larger. Los Angeles, with 85%, and Denver, with 68%, had the largest proportion of the
total population within 15 km of a monitor. Seattle, with 18%, had the lowest population coverage in
large part because ambient CO concentrations there require only a single CO monitor. For the elderly
population, Los Angeles, at 83%, Anchorage, at 73%, and Denver, at 70%, had the greatest
population coverage within 15 km of a monitor; Seattle, at 18%, again had the lowest coverage.
Proximity to monitoring stations is considered further in Sections 3.5 and 3.6 regarding spatial
variability within cities. In combination, these data illustrate that population coverage varies by
monitor and across cities.
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
<1 km
N %
5,391 1.53
5,480 0.10
95,732 1.28
26,096 0.88
29,068 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
January 2010
3-27
-------
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
31,036 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
3.5. Environmental Concentrations
3.5.1. Spatial Variability
3.5.1.1. National Scale
The current NAAQS designates that the level of the NAAQS is not to be exceeded more than
once per year at a given monitoring site. Figure 3-15 and Figure 3-16 show the second-highest 1-h
and second-highest 8-h county-average CO concentrations, respectively, over the U.S. along with
estimates of the fraction of the U.S. total population exposed to those concentrations. Although 93%
of the U.S. counties are not represented in AQS reporting, based on their population densities and
proximity to sources, those counties are not expected to have higher concentrations than the ones
analyzed here in the absence of extreme events such as wildfires. Continuous hourly averages are
reported from U.S. monitoring stations. One-hour (1-h) and 8-h CO data were available for 243
counties and autonomous cities or municipalities (e.g., Anchorage, AK, Washington, DC) where CO
monitors met the 75% data completeness criteria used in this analysis for the years 2005-2007. In
2007, no monitored location reported a second-highest 1-h CO concentration above 35 ppm (Figure
3-15). Moreover, only two monitored locations, one in Weber County, UT and the other in Jefferson
County, AL (including Birmingham, AL), reported second-highest 1-h CO concentrations between
15.1 and 35.0 ppm. Figure 3-16 shows that only 5 counties reported second-highest 8-h CO
concentrations above 5.0 ppm: Jefferson County, AL; Imperial County, CA; Weber County, UT;
Philadelphia County., PA; and Anchorage Municipality, AK.
January 2010
3-28
-------
Carbon Monoxide - Second Highest 1-hour Average, 2007
300
250 -
200
150
100
^MilSlfcai
*^&gj^£^^&&*ssfi
*^m:f\m$
Sf
ii*
:^/-Xy^^&$ tf
> 35.1 ppm
15.1-35.0 ppm
10.1 - 15.0 ppm
5.1-10.0 ppm
< 5.0 ppm
1 No data
f
Figure 3-15. 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.
January 2010
3-29
-------
Carbon Monoxide - Second Highest 8-hour Average, 2007
300
250
200
> 9.1 ppm
7.6 - 9.0 ppm
5.1-7.5 ppm
2.6-5.0 ppm
< 2.5 ppm
1 No data
Figure 3-16. 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.
January 2010
3-30
-------
Table 3-6. Distribution of 1-h avg CO concentration (ppm) derived from AQS data.
Percentiles
N
Mean Min
1
10 25 50 75 90 95 99 Max
NATIONWIDE STATISTICS (N = NUMBER OF OBSERVATIONS)
2005-2007
7,180,700
0.5
0.0
0.0 0.0
0.1
0.2
0.4
0.6
0.9
1.2
Spring (March - May)
1,826,167
0.4
0.0
0.0 0.0
0.1
0.2
0.3
0.5
0.8
1.0
2.1
39.0
2005 2,391,962
2006 2,402,153
2007 2,386,585
Winter (December - February) 1 ,752,340
0.5
0.5
0.4
0.6
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
0.2
0.2
0.3
0.4
0.4
0.3
0.4
0.6
0.6
0.5
0.7
1.0
0.9
0.8
1.2
1.3
1.2
1.1
1.6
2.3
2.1
1.9
2.7
22.3
35.3
39.0
20.0
1.7 35.3
Summer (June - August)
1,811,082
0.4
0.0
0.0 0.0
0.0
0.2
0.3
0.5
0.7
0.9
1.5
39.0
Fall (September - November) 1,791,111
0.5
0.0
0.0 0.0
0.1
0.2 0.4 0.6 1.0
1.3
2.2 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)
Fall (September - November)
285
285
285
285
285
285
285
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.
January 2010
3-31
-------
Table 3-6 contains the distribution of hourly CO measurements reported to AQS for
2005-2007. All monitoring locations meeting the 75% data completeness criteria have been included
in this table. Several monitors in EPA Region 10, including four in Alaska, did not meet the data
completeness criteria since CO reporting was only required during the first and fourth quarters of
each year at these sites. Anchorage was included in the table, however, for an approximate
comparison with the other CSAs and CBSAs reporting year-round measurements to AQS.
Anchorage and other partial-year monitors were not, however, included in the national statistics
shown in the table. AQS site number 371190041, located in Charlotte, NC, was the only site with co-
located monitors both meeting the data completeness criteria and, therefore, the nationwide data in
the table was derived from 286 monitors located at 285 sites. In Section 3.5.1.3, the nationwide 1-h
avg statistics shown in Table 3-6 (along with the nationwide 24-h avg, 1-h daily max and 8-h daily
max statistics) are further divided by monitoring scale (microscale, middle scale, etc.) to address
issues relating to the near-road environment.
The nationwide mean, median, and interquartile range for 1 -h measurements reported for
2005-2007 were 0.5, 0.4 and 0.4 ppm, respectively, and these statistics did not change by more than
0.1 ppm over the 3-yr period. More than 50% of the data nationwide were below the LOD for the
majority of monitors in use. The largest recorded second-highest 1-h concentration, 26.3 ppm, for
this period was reported in 2006 in Birmingham, AL (AQS site ID: 010736004). The highest 1-h
concentration, 39 ppm, between 2005 and 2007, was reported in Ogden, UT (AQS site ID:
490570006) on August 28, 2007. An annual outdoor barbeque festival held in Ogden on that day
resulted in a period of elevated CO concentrations. The seasonally stratified concentrations in Table
3-6 are generally highest in the winter (December-February) and fall (September-November) and
decrease on average during the spring (March-May) and summer (June-August).
Nationwide statistics pooled by site are listed in the center rows of Table 3-6 and illustrate the
distribution of the site average CO concentrations recorded at the 285 monitoring sites for
2005-2007 (Figure 3-10). The site reporting the highest 3-yr pooled 1-h avg CO concentration,
1.5 ppm, was located in San Juan, Puerto Rico (AQS site ID: 721270003). The 11 individual
CS As/CBS As discussed earlier are included in the table, none of which reported concentrations
above the value of the 1-h NAAQS. Four of the 11 cities (Boston, Houston, Pittsburgh and St. Louis)
had 95th percentile 1-h CO concentrations below 1 ppm; the 95th percentile concentrations for the
remaining cities were below 3.1 ppm. Lack of year-round monitoring in Anchorage prevented a
direct comparison with the other metropolitan regions. However, Anchorage exhibited a 1-h CO
distribution shifted higher in concentration when compared to the U.S. average during fall or winter.
The 99th percentile 1-h avg concentration in Anchorage was 5.0 ppm; the other selected cities with
year-round monitoring had 99th percentile concentrations ranging from 0.9 ppm to 2.5 ppm.
January 2010 3-32
-------
Table 3-7. Distribution of 24-h avg CO concentration (ppm) derived from AQS data.
Percentiles
N
Mean Min
1
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)
303,843
101,184
101,652
101,007
74,144
77,317
76,562
0.5
0.5
0.5
0.4
0.6
0.4
0.4
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
0.1
0.1
0.1
0.1
0.2
0.1
0.1
0.3
0.3
0.3
0.2
0.3
0.2
0.2
0.4
0.4
0.4
0.4
0.5
0.4
0.3
0.6
0.6
0.6
0.5
0.7
0.5
0.5
0.9
0.9
0.9
0.8
1.1
0.7
0.7
1.1
1.1
1.1
1.0
1.3
0.9
0.8
1.7
1.8
1.6
1.6
2.0
1.4
1.3
7.0
5.8
7.0
6.9
7.0
6.4
6.9
Fall (September - November)
75,820
0.5
0.0
0.0 0.0 0.1 0.3
0.4
0.6 0.9
1.1
1.7
5.8
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007
2005
2006
2007
Winter (December -
February)
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
285
285
285
285
285
285
285
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
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
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
0.0
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
0.0
0.0
0.1 0.2
0.1 0.2
0.1 0.2
0.1 0.2
0.2 0.2
0.1 0.2
0.1 0.2
0.1 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 0.7
0.6 0.8
0.6 0.7
0.5 0.7
0.7 0.9
0.5 0.7
0.5 0.6
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
(N = NUMBER OF OBSERVATIONS)
0.2
0.1
0.0
0.1
0.0
0.0
0.0
0.1
0.0
0.2
0.0
0.0
0.2 0.4
0.2 0.2
0.1 0.1
0.2 0.2
0.0 0.1
0.1 0.1
0.1 0.2
0.2 0.3
0.0 0.0
0.3 0.4
0.1 0.2
0.0 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 1 .9
0.6 0.8
0.5 0.7
0.6 0.9
0.4 0.5
0.6 1.0
0.6 0.8
1.1 1.6
0.4 0.6
0.9 1.2
0.5 0.7
0.6 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.
Table 3-7 contains the distribution of 24-h avg CO concentrations derived from the 1-h
concentrations reported to AQS and summarized in Table 3-6. The nationwide mean, median, and
interquartile range for 24-h avg values during 2005-2007 were 0.5, 0.4 and 0.3 ppm, respectively.
These were similar to those for the 1-h values and showed more than half the data falling below the
LOD for the majority of monitors in the field. The maximum 24-h avg concentration in these years,
7 ppm, was reported in Birmingham, AL (AQS site ID: 010736004). The 99th percentile 24-h avg
January 2010
3-33
-------
concentrations ranged from 0.9 ppm to 2.5 ppm in the selected cities with year-round monitoring;
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)
NATIONWIDE STATISTICS,
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 0.0
1.0 0.0
0.9 0.0
0.8 0.0
1.2 0.0
0.8 0.0
0.7 0.0
1.0 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
POOLED BY SITE (N = NUMBER OF SITES)
285
285
285
285
285
285
285
285
0.9 0.1
1.0 0.1
0.9 0.1
0.8 0.1
1.2 0.0
0.8 0.1
0.7 0.0
1.0 0.1
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007) (N
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 0.0
0.8 0.0
0.7 0.0
1.2 0.1
0.7 0.0
1.0 0.0
0.9 0.0
1.9 0.0
0.6 0.0
1.5 0.2
0.8 0.0
0.9 0.0
0.1
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
0.3
0.5
0.5
0.5
0.4
0.6
0.4
0.3
0.5
0.6
0.7
0.6
0.6
0.8
0.6
0.5
0.7
0.8
0.9
0.9
0.8
1.0
0.8
0.6
0.9
1.1
1.2
1.1
1.0
1.5
1.0
0.8
1.2
1.5
1.6
1.6
1.4
2.1
1.3
1.1
1.7
1.7
2.0
1.8
1.6
2.5
1.5
1.3
2.0
2.3 3.9
2.5 3.7
2.3 4.8
2.0 3.1
3.4 4.1
2.1 4.0
2.2 3.3
2.4 4.1
= NUMBER OF OBSERVATIONS)
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 13.1
2.2 10.8
2.6 10.0
3.9 9.3
2.6 4.6
4.0 8.4
2.5 5.8
5.3 7.8
2.0 6.7
4.3 5.9
2.7 5.7
3.8 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.
Table 3-8 contains the distribution of 1-h daily max CO concentrations derived from 1-h
values reported to AQS for all monitors meeting the inclusion criteria described earlier. The
nationwide mean, median, and interquartile range for 1-h daily max concentrations reported for
2005-2007 were 0.9, 0.7 and 0.8 ppm, respectively. Roughly one-third of the 1-h daily max data fall
below the LOD for the majority of CO monitors reporting to AQS. The 99th percentile 1-h daily max
January 2010
3-34
-------
concentrations ranged from 2.0 ppm to 5.3 ppm in the selected cities with year-round monitoring;
Anchorage had a 99th percentile concentration of 7.6 ppm.
Table 3-9. Distribution of 8-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)
NATIONWIDE STATISTICS,
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
00
00
0.0
0.0
0.0
0.0
00
0.3 0.3
03 03
03 03
0.3 0.3
0.3 0.3
0.3 0.3
0.3 0.3
03 03
0.3
03
03
0.3
03
03
0.3
03
0.3 0.5
03 06
03 05
0.3 0.5
0.4 0.7
0.3 0.5
0.3 0.4
03 06
0.8
09
08
0.8
1.1
07
0.6
09
1.3
1 4
1 3
1.2
1 7
1.1
0.9
1 4
1.7
1 8
1 7
1.5
?1
1 3
1.1
1 8
2.6
28
26
2.3
3?
?0
1.7
27
10.9
97
98
10.9
9.8
9.6
10.9
90
POOLED BY SITE (N = NUMBER OF SITES)
285
285
285
285
285
285
285
285
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007)
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
0.7
0.7
0.7
0.6
0.9
0.6
0.5
0.7
02
03
0.2
0.2
0.2
02
02
0.2
03 03
03 03
0.3 0.3
0.3 0.3
0.3 0.4
03 03
03 03
0.3 0.3
04
04
04
04
04
04
03
04
05 06
05 06
0.5 0.6
0.5 0.6
0.6 0.8
04 05
04 05
0.5 0.6
08
09
08
07
1.1
07
06
OP
1 0
1 1
1.1
1 0
1 4
09
08
1 ?
1 2
1 4
1 ?
1 1
1 7
1 1
09
1 3
1 7
1 9
1 8
1 fi
?4
1 6
1 5
1 8
21
22
2.4
2.0
2.6
22
20
2.2
(N = NUMBER OF OBSERVATIONS)
1.7
0.6
0.6
0.8
0.5
0.7
0.7
1.3
0.5
1.1
0.6
0.7
03
0.0
0.3
0.3
03
0.3
0.3
0.3
0.3
0.3
0.3
0.0
03 04
0.2 0.2
0.3 0.3
0.3 0.3
03 03
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 0.3
06
03
03
03
03
0.3
0.3
0.4
0.3
0.5
0.3
0.3
09 15
0.4 0.5
0.3 0.5
0.5 0.7
03 04
0.3 0.6
0.4 0.6
0.6 1.0
0.3 0.3
0.7 1.0
0.3 0.5
0.3 0.5
23
08
07
1 0
06
0.9
0.9
1.8
0.6
1.4
0.7
0.8
33
1.1
OP
1 4
09
1.5
1.2
2.5
0.9
1.8
0.9
1.3
39
1 3
1 1
1 8
1 1
1.8
1.4
3.0
1.0
2.2
1.2
1.6
50
1 7
1 8
?4
1 7
2.7
1.8
3.8
1.5
3.2
1.9
2.5
65
2.5
5.8
3.4
33
6.2
3.0
5.8
3.7
4.0
4.2
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.
Table 3-9 contains the distribution of 8-h daily max concentrations derived from the 1-h CO
concentrations reported to AQS. This was done by first calculating the average concentration for
each successive 8-h period, thereby producing 24 8-h avg per day. The maximum of these values for
a given monitor within a given day (midnight-to-midnight) was used as the 8-h daily max statistic
for that monitor and day. The nationwide mean, median, and interquartile range for 8-h daily max
January 2010
3-35
-------
concentrations reported for 2005-2007 were 0.7, 0.5, and 0.5 ppm, respectively. Half of the 8-h daily
max concentrations fell below the LOD for the majority of CO monitors in the field. The highest 8-h
daily max concentration, 10.9 ppm, was recorded at a monitor located 5 mi north of Newkirk, OK
(AQS site ID: 400719010). The 99th percentile 8-h daily max concentrations ranged from 1.5 ppm to
3.8 ppm in the selected cities with year-round monitoring; Anchorage had a 99th percentile 8-h daily
max concentration of 5.0 ppm.
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-
Erf-
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+
• --(-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 3-17. 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.
Table 3-7 through Table 3-9 show distributions of CO data based on the 24-h avg, 1-h daily
max and 8-h daily max concentration. The current standards are based on 1-h and 8-h calculations.
While the nationwide concentrations vary in absolute magnitude based on these three statistics, the
shape of the distributions are quite similar up to the 99th percentile. The relative increase from the
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
to be expected since this statistic is more sensitive to short-term (less than 8 h) increases in CO
concentration. Box plots showing the range in Pearson correlation coefficients (r) between the
different statistics are shown in Figure 3-17. Included are the correlation of the 24-h avg with the 1-h
daily max and 8-h daily max, as well as the correlation between the 1-h daily max and 8-h daily max,
all calculated using the same 2005-2007 data set stratified by season. Correlations are generally quite
high across all seasons and all comparisons, with median r > 0.8. Correlations are higher on average
in the wintertime compared to the summertime for the two comparisons involving the 1-h daily max
January 2010
3-36
-------
statistic. The correlations between the 24-h avg and the 8-h daily max are the highest in all seasons,
which is in agreement with the distributional similarities shown in the preceding tables.
3.5.1.2. Urban Scale
This section describes urban variability in CO concentrations reported to AQS at the individual
CSA/CBSA level. Denver, CO, and Los Angeles, CA, were selected for this assessment to illustrate
the variability in CO concentrations measured across contrasting metropolitan regions. Information
on the other nine cities evaluated for this assessment is included in Appendix A. Maps of the Denver
CSA and the Los Angeles CSA shown in Figure 3-18 and Figure 3-21, respectively, illustrate the
location of all CO monitors meeting the inclusion criteria described earlier. Letters on the maps
identify the individual monitor locations and correspond with the letters provided in the
accompanying concentration box plots (Figure 3-19 and Figure 3-22) and pair-wise monitor
comparison tables (Table 3-10 and Table 3-11). The box plots for each monitor include the hourly
CO concentration median and interquartile range with whiskers extending from the 5th to the 95th
percentile. Data from 2005-2007 were used to generate the box plots, which are stratified by season
as follows: 1 = winter (December-February), 2 = spring (March-May), 3 = summer (June-August),
and 4 = fall (September-November). The comparison tables include the Pearson correlation
coefficient (r), the 90th percentile of the absolute difference in concentrations (P90) in ppm, the
coefficient of divergence (COD) and the straight-line distance between monitor pairs (d) in km. The
COD provides an indication of the variability across the monitoring sites within each CSA/CBSA
and is defined as follows:
Equation 3-1
whereXtj andXik represent the observed hourly concentrations for time period / at sites j and k, and/?
is the number of paired hourly observations. A COD of 0 indicates there are no differences between
concentrations at paired sites (spatial homogeneity), while a COD approaching 1 indicates extreme
spatial heterogeneity. Pearson correlation is also plotted as a function of distance for Denver and Los
Angeles in Figure 3-20 and Figure 3-23, respectively. Similar maps, box plots, and comparison
tables for the nine remaining CSAs/CBSAs are included in Annex A.
The information contained in these figures and tables should be used with some caution since
many of the reported concentrations for the years 2005-2007 are near or below the monitors' stated
LOD. Because ambient concentrations are now in large part very near or below the 0.5 ppm LOD for
the majority of FRMs and the coarsely reported measurement resolution is 0.1 ppm, the comparison
statistics shown in these tables might be biased to exhibit specious heterogeneity in the box plots.
January 2010 3-37
-------
Denver Combined Statistical Area
Ql
r
Denver CSA
CO Monitors
Interstate Highways
Major Highways
0 15 30 60 90
120
• Kilometers
Figure 3-18. Map of CO monitor locations and major highways for Denver, CO.
January 2010
3-38
-------
ABC
D E
sitoin nftnnnnn? 08-°31- 08-°13- °8-123- 08-°01-
SitelD 08-031-0002 oolg ooog 0010 3001
Mean 0.65 0.52 0.42 0.55 0.52
SD 0.42 0.46 0.38 0.46 0.36
Obs 25959 25552 25559 26048 25920
Scale Micro Micro Micro Negjbor- ™**«-
3-
- — v
E
Q_
£ 2-
C.
O
•t—>
05
concent
0-
ll
I I I I
I I I
I I I
I i
1
r
I I I
1234 1234 1234 1234 1234
season
Figure 3-19. 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.
January 2010
3-39
-------
Table 3-10. Table of intersampler 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.
Micro
A B
A 1.00 0.76
0.0 0.5
0.00 0.34
0 1.3
B 1.00
o 0.0
ii o.oo
0
c
Neighborhood
C D
0.46
0.7
0.44
46.9
0.49
0.7
0.47
47.0
1.00
0.0
0.00
0
D
•a
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
0
o
!> E
z
Legend
r
P90
COD
d
0 68.2
1.00
0.0
0.00
0
January 2010
3-40
-------
0.9
0.8
0.7 -
0.6
o
I 0.5-1
o
O
0.4
0.3
0.2
0.1 -
10
20
30
40 50
Distance (km)
60
70
80
90
Figure 3-20. Intersampler correlation versus distance for monitors located within the Denver
CSA.
January 2010
3-41
-------
Los Angeles Combined Statistical Area
Ql
r
Los Angeles CSA
CO Monitors
Interstate Highways
Major Highways
0 15 30 60 90
120
• Kilometers
Figure 3-21. Map of CO monitor locations and major highways for Los Angeles, CA.
January 2010
3-42
-------
Site ID
Mean
SD
Obs
Scale
06-065-
1003
0.67
0.42
24885
Micro
06-059-
1003
0.31
0.47
24760
Middle
06-037-
9033
0.23
0.29
24135
Middle
06-037-
1301
0.98
0.89
24825
Middle
06-071-
9004
0.53
0.38
24844
Middle
06-065-
9001
0.29
0.20
24792
Neighbor-
06-037-
5005
0.24
0.37
24965
Neighbor-
06-059-
0007
0.42
0.46
24264
Urban
4.5-
4.0-
- — v
E 3.5-
Q_
S 3.0-
C
O 2.5-
,|_^
£ 2.0-
^^j
^^^
C
0) 1.5-
O
c
o 1.0-
o
0.5-
0.0-
i
.
ll
I:
hi
\\
!"!
Li
i
li
1234 1234 1234 1234 1234 1234 1234 1234
season
Figure 3-22. 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 3-22. See the next two pages for parts 2 and 3 of Figure 3-22.
January 2010
3-43
-------
Site ID
06-037-
0002
06-071-
0001
06-037-
1002
06-059-
5001
M
06-037-
4002
06-037-
1103
06-059-
2022
06-065-
5001
Mean 0.42
0.17
0.66
0.62
0.69
0.56
0.26
0.25
SD
0.27
0.17
0.59
0.55
0.56
0.50
0.25
0.14
Obs
2,5001
24105 24892
24705
24259
24645
24831
24938
Scale
Null
Null
Null
Null
Null
4.5-
4.0-
•— V
E 3.5-
Q_
B 3.0-
C
O 2.5-
•!_>
E 2.0-
^^«
^•^
C
0) 1.5-
0
o 1.0-
o
0.5-
0.0-
I
III
,..
I
I
'I
i
I
I
i
hi
•Ht
1234 1234 1234 1234 1234 1234 1234 1234
season
Part 2 of 3 for Figure 3-22
January 2010
3-44
-------
Site ID
06-037-
2005
06-037-
1701
06-037-
1201
06-065-
8001
06-037-
6012
06-071-
1004
W
06-071-
0306
06-037-0113
Mean 0.72
0.69
0.57
0.60
0.30
0.59
0.30
0.41
SD
0.48
0.45
0.54
0.46
0.25
0.32
0.28
0.36
Obs 24804
24912
24281
24778
24860
24767
24796
24916
Scale Null Null
4.5-
4.0-
-— N
E 3.5-
Q_
S 3.0-
g 2.5-
2 2.0-
4— •
0 1.5-
0
^™
o 1.0-
o
0.5-
0.0-
ll
i
i
Null Null Null Null Null
1
ll
1
,,
T T 1 T
1
III
k
r T
Null
||
T
•\
1234 1234 1234 1234 1234 1234 1234 1234
season
Part 3 of 3 for Figure 3-22
January 2010
3-45
-------
Table 3-11. Table of intersampler 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).
cro Middle
ABC
A 1.00 0.56 0.56
K 0.0 0.8 0.9
jg 0.00 0.66 0.67
0 57.1 104.6
B 1.00 0.55
0.0 0.6
0.00 0.66
0 112.0
C 1.00
0.0
0.00
i o
€ D
E
D
0.54
1.1
0.30
74.8
0.67
1.3
0.70
38.6
0.55
1.6
0.72
82.5
1.00
0.0
0.00
0
E
0.73
0.5
0.30
21.3
0.50
0.7
0.64
76.9
0.50
0.7
0.64
100.4
0.44
1.3
0.42
88.6
1.00
0.0
0.00
0
F
•a
8
Neighbor
-hood
F G
0.72 0.45
0.7 0.9
0.46 0.73
30.5 95.0
0.46 0.60
0.5 0.5
0.59 0.69
55.1 55.8
0.50 0.39
0.4 0.6
0.57 0.74
132.5 84.4
0.39 0.63
1.6 1.5
0.56 0.76
86.0 20.4
0.69 0.43
0.6 0.8
0.42 0.72
48.0 108.0
1.00 0.43
0.0 0.5
0.00 0.70
•g 0 106.1
£ G
r
H P90
s d
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
Null
I JKLMNOPQRSTUVWX
0.54 0.35 0.70 0.66 0.46 0.62 0.61 0.48 0.53 0.78 0.73 0.67 0.54 0.70 0.55 0.57
0.7 1.0 0.6 0.6 0.8 0.6 0.8 0.9 0.7 0.4 0.6 0.6 0.8 0.5 0.8 0.7
0.33 0.68 0.29 0.24 0.39 0.37 0.58 0.47 0.35 0.19 0.29 0.37 0.53 0.20 0.54 0.42
52.6 110.6 88.2 51.0 74.1 77.4 43.3 80.1 70.1 35.0 108.0 6.1 114.5 27.4 62.8 98.1
0.14 0.28 0.70 0.72 0.64 0.58 0.62 0.34 0.50 0.60 0.57 0.40 0.25 0.36 0.55 0.47
0.7 0.7 0.9 0.7 0.9 0.8 0.5 0.6 0.9 0.9 0.8 0.9 0.7 0.8 0.6 0.7
0.64 0.69 0.63 0.62 0.64 0.63 0.56 0.59 0.68 0.66 0.62 0.67 0.67 0.65 0.62 0.59
51.2 158.5 66.3 27.9 29.5 51.6 23.7 129.6 54.0 46.3 80.7 59.3 96.2 54.9 107.5 64.4
0.27 0.45 0.59 0.63 0.43 0.53 0.51 0.41 0.45 0.61 0.53 0.41 0.40 0.49 0.67 0.42
0.6 0.4 1.1 0.9 1.1 0.9 0.4 0.4 1.0 0.9 0.9 0.9 0.5 0.7 0.4 0.7
0.62 0.62 0.65 0.64 0.69 0.63 0.61 0.56 0.70 0.66 0.62 0.67 0.62 0.65 0.57 0.60
62.2 104.0 57.4 84.2 94.0 67.5 122.7 171.9 59.6 75.4 64.0 99.2 48.4 77.9 75.4 74.9
0.21 0.33 0.70 0.78 0.70 0.74 0.57 0.28 0.53 0.65 0.49 0.35 0.23 0.39 0.51 0.50
1.5 1.7 0.9 0.9 1.0 1.0 1.5 1.7 1.0 1.0 1.2 1.2 1.6 1.3 1.5 1.3
0.44 0.73 0.35 0.30 0.39 0.41 0.65 0.56 0.39 0.29 0.41 0.45 0.60 0.35 0.61 0.50
35.0 152.6 29.1 23.8 11.8 15.3 59.5 154.5 23.8 45.0 42.1 73.7 58.2 57.0 103.4 26.4
0.48 0.31 0.65 0.55 0.39 0.58 0.51 0.46 0.52 0.68 0.71 0.64 0.50 0.64 0.51 0.53
0.6 0.8 0.7 0.7 0.9 0.6 0.6 0.7 0.7 0.6 0.6 0.6 0.6 0.5 0.6 0.6
0.35 0.65 0.35 0.33 0.46 0.39 0.56 0.43 0.41 0.31 0.32 0.41 0.51 0.29 0.52 0.41
59.9 90.2 96.3 65.7 90.1 87.9 64.6 73.3 78.6 44.2 116.3 17.7 119.3 32.7 44.9 109.1
0.56 0.30 0.58 0.55 0.36 0.53 0.51 0.49 0.47 0.69 0.66 0.66 0.56 0.68 0.49 0.55
0.4 0.4 1.0 0.8 1.1 0.8 0.3 0.3 0.9 0.8 0.8 0.8 0.3 0.6 0.4 0.5
0.38 0.58 0.46 0.43 0.54 0.46 0.50 0.32 0.53 0.46 0.40 0.49 0.47 0.43 0.47 0.39
74.8 137.8 106.5 63.7 81.1 93.4 32.4 75.7 89.2 58.1 125.1 36.6 135.3 54.8 92.3 112.0
0.19 0.18 0.64 0.59 0.59 0.59 0.42 0.26 0.40 0.51 0.52 0.43 0.24 0.27 0.41 0.59
0.7 0.6 1.0 0.8 1.0 0.8 0.5 0.5 1.0 1.0 0.9 0.9 0.6 0.9 0.6 0.6
0.72 0.75 0.72 0.72 0.75 0.73 0.73 0.70 0.75 0.73 0.72 0.74 0.73 0.73 0.72 0.69
51.0 166.0 27.0 44.2 26.4 22.7 78.3 174.8 34.4 63.9 29.1 93.7 48.7 75.8 118.6 11.4
0.29 0.31 0.72 0.81 0.63 0.70 0.67 0.37 0.54 0.69 0.61 0.49 0.29 0.47 0.54 0.59
0.6 0.8 0.7 0.5 0.8 0.6 0.5 0.7 0.8 0.7 0.7 0.8 0.7 0.7 0.6 0.5
0.43 0.62 0.41 0.38 0.48 0.40 0.46 0.41 0.52 0.44 0.41 0.49 0.53 0.45 0.50 0.39
33.9 144.7 51.8 10.5 23.2 37.3 32.9 129.2 37.7 31.4 68.3 51.7 81.8 41.6 93.7 53.7
I 1.00 0.17 0.43
0.0 0.6 1.0
0.00 0.62 0.35
0 117.7 36.5
J 1.00 0.35
0.0 1.2
0.00 0.67
0 142.7
K 1.00
0.0
0.00
0
L
M
N
0
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
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
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
0.42
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
0.33
0.5
0.36
131.4
0.29
0.3
0.55
123.6
0.40
1.2
0.46
167.7
0.40
1.0
0.42
130.6
0.24
1.2
0.53
152.4
0.32
1.0
0.46
157.4
0.44
0.3
0.47
107.9
0.41
0.8
0.43
18.7
0.31
1.1
0.71
131.7
0.67
0.7
0.39
18.1
0.58
0.7
0.37
28.1
0.41
0.9
0.46
34.7
0.67
0.7
0.44
11.7
0.55
0.9
0.62
69.5
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
0.44
0.7
0.33
56.5
0.31
1.0
0.64
158.2
0.74
0.6
0.28
20.0
0.61
0.7
0.29
61.5
0.44
0.9
0.44
52.3
0.64
0.7
0.34
31.8
0.55
0.8
0.54
101.2
0.47
0.8
0.43
49.2
0.19
1.1
0.69
105.4
0.52
0.8
0.42
85.3
0.50
0.8
0.38
50.2
0.31
1.0
0.48
74.0
0.49
0.8
0.45
75.1
0.43
0.9
0.59
47.5
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
0.61
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
0.24
0.6
0.50
68.4
0.43
0.4
0.59
51.0
0.58
1.0
0.53
97.8
0.58
0.8
0.51
86.4
0.43
1.1
0.59
109.6
0.51
0.8
0.53
95.9
0.45
0.4
0.56
102.5
0.45
0.5
0.38
50.0
0.25
0.7
0.62
161.1
0.69
0.8
0.38
18.9
0.59
0.6
0.37
48.5
0.43
0.9
0.48
35.2
0.71
0.6
0.38
21.2
0.57
0.6
0.50
85.9
January 2010
3-46
-------
Mi- Mi,Mi. Neighbor Ur- K
cro Mlddle -hood ban NuN
P 1.00 0.39 0.47 0.42
0.0 1.0 1.0 0.9
0.00 0.54 0.47 0.40
0 149.6 114.2 187.6
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
z
U
V
W
X
0.40
0.9
0.50
82.4
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
0.40
0.4
0.45
192.2
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
0.47
0.7
0.43
104.2
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
0.38
0.4
0.44
102.8
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
0.35
0.6
0.38
178.1
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
January 2010
3-47
-------
0.9 -
0.8 -
0.7 -
0.6 -
o
I 0.5 H
o
O
0.4 -
0.3 -
0.2 -
0.1 -
•. •• • ••• • • A • •
•\ m ™ • ••••••-•!
• • • V« • ^ • ••
• • V • *• •
• •v»r*»*»*.j
• ..' :"/."*.".••.''
• • _ • •
• • §
• • •
•w . '
• . • •
v.*.-»
50
100 150
Distance (km)
200
250
Figure 3-23. Intersampler correlation versus distance for monitors located within the Los
Angeles CSA.
The Denver CSA in Figure 3-18 incorporates an area of 33,723 km2 with a maximum straight-
line distance between CO monitors of 79 km. Of the five CO monitors meeting the inclusion criteria,
three were sited for microscale monitoring and two were sited for neighborhood scale monitoring.
Sites A and B are located in downtown Denver while Site E is located in an industrial region north of
town and surrounded on three sides by three heavily-traveled interstate highways. Sites C and D are
located in two smaller towns (Longmont and Greeley, respectively) north of Denver. The means and
seasonal patterns shown in Figure 3-19 are similar for all five monitors within this CSA. The highest
annual mean concentration (0.7 ppm) was observed at Site A, a downtown microscale monitor, while
the lowest annual mean concentration (0.4 ppm) was observed at Site C, a microscale monitor in
Longmont. The step-wise nature of the box plots is attributed to the 0.1 ppm resolution of the CO
monitors used in the Denver CSA. Because these monitors have LOD of 0.5 ppm, it is also likely
that the means and statistical distributions are biased as well.
The Los Angeles CSA in Figure 3-21 incorporates an area of 88,054 km2 and a maximum
straight-line distance between monitors of 192 km, making it more than twice the size of the Denver
CSA. Of the 11 CSAs/CBSAs investigated, Los Angeles had the largest number of CO monitors (N
= 24) meeting the inclusion criteria. One monitor was sited for microscale, four for middle scale, two
for neighborhood scale, and one for urban scale. The remaining 16 monitors did not contain a siting
classification in AQS. The monitors were evenly distributed around the Los Angeles and Riverside
areas, with outlying monitors in Santa Clarita (Site U), Lancaster (Site C), Victorville (Site W),
Barstow (Site J) and Palm Springs (Site P). A large amount of variability is present in the means and
seasonal patterns displayed in Figure 3-22. Generally speaking, lower annual mean concentrations
(<0.3 ppm) were measured in the outlying towns including those listed above as well as Lake
Elsinore (Site F) and Mission Viejo (Site O). In addition, a neighborhood scale upwind background
site (Site G) located on the grounds of the Los Angeles International Airport and 1.5 km from the
January 2010
3-48
-------
Pacific Ocean reported a relatively low mean annual concentration of 0.2 ppm. The highest annual
mean concentration (1.0 ppm) was observed at Site D, a middle scale maximum concentration site
located 25 m from a busy surface street and adjacent to the Imperial Shopping Mall. This site is also
180 m from a major highway intersection and 350 m from Interstate 105. The step-wise nature of the
box plots is attributed to the 0.1 ppm resolution of the CO monitors used in the Los Angeles CSA.
Because these monitors have LOD of 0.5 ppm, it is also likely that the means and statistical
distributions are biased as well.
The pair-wise comparisons for measurements at the monitors in each of the 11 CSAs/CBSAs
included in this analysis reveal a wide range of response between monitors in each city and among
the cities judged against each other (Table 3-10, Table 3-11 and Annex Tables A-9 through A-16).
While this wide range is produced by the interactions of many physical and chemical elements, the
location of each monitor and the uniqueness of its immediate surroundings can often explain much
of the agreement or lack thereof.
For the monitor comparisons within the Denver CSA (Table 3-10 and Figure 3-20), the
correlations tend to be inversely related to the monitor separation distance, with the highest
correlation (r = 0.76) for the two downtown Denver monitors (Sites A and B) separated by 1.3 km
and the lowest correlations (r < 0.46) between the downtown Denver monitors and the Greeley
monitor (Site D) located roughly 80 km north. While Sites A and B have a high correlation, the
comparative magnitudes of the concentrations measured at these two sites, as determined by the P90
and COD, is comparable to comparisons with much less proximal monitors. This is likely caused by
the location of these two monitors on opposite sides of downtown Denver, as illustrated by the aerial
view of monitors A and B in Figure 3-24. While there is no prevailing wind direction in Denver, the
wind comes from the south-southwest with a slightly higher frequency than other directions, making
Site A downwind of the urban core more frequently than Site B. Assuming traffic within the urban
core is a major source of CO, this would explain the higher mean concentrations measured at Site A
relative to Site B despite their close proximity.
Greater variability in the pair-wise comparison statistics is observed in the Los Angeles CSA
compared to the Denver CSA, partially due to the greater number of monitors spread over a larger
area. Factors other than the distance between monitors, however, can contribute substantially to
concentration disparities observed between monitors. To illustrate this point, Site S (located in
Reseda, a suburb in the Simi Valley northwest of Los Angeles) correlates well (r = 0.73) with Site A
(located 108 km to the southeast in Riverside). In fact, Site S correlates well (r > 0.62) with Sites A,
E, F and T, all east of Los Angeles and all over 100 km away. Site S is located in a densely populated
urban area with a mixture of commercial and residential land whereas the other four sites are located
in less densely populated regions with commercial, residential and undeveloped land. Sites S and T
contain no monitoring scale information in AQS, but Sites A, E and F are classified as microscale,
middle scale and neighborhood scale, respectively. In contrast to the above example, Sites I and Q
are located only 19 km apart in Azusa and Pasadena, respectively, and they correlate less well
(r = 0.41). While these two locations are relatively close in proximity with similar topography, the
siting of the two monitors is quite different. Site I in Azusa is located 700 m from 1-210 in a mixed
use community containing warehouses, small industry, housing and a gravel operation (Figure 3-25)
while Site Q in Pasadena is located between a large residential neighborhood and the California
Institute of Technology campus (Figure 3-26). Neither of these sites has monitoring scale
designations reported in AQS. The contrasting CO emission sources surrounding these two monitors
result in disparate concentrations with poor correlations despite their close proximity. Topography
and micrometeorology can also play an important role in the correlation between monitors. For
example, Sites C and P are isolated from the other sites in the Los Angeles CSA by the San Gabriel
Mountains and the San Bernardino Mountains, respectively, resulting in lower than average
concentrations (Figure 3-22) and relatively low pair-wise correlations (Table 3-11) for these two
sites. This analysis demonstrates that agreement between monitors on an urban scale is a complex
function of monitor siting, location relative to sources, geography, and micrometeorology.
January 2010 3-49
-------
$3W
-------
Figure 3-25. 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.
Scale: 1 cm = 145m.
January 2010
3-51
-------
Figure 3-26. 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.
Scale: 1 cm = 145m.
January 2010
3-52
-------
3.5.1.3. Micro- to Neighborhood Scale and the Near-Road Environment
Table 3-12 shows the 2005-2007 nationwide distributional data for all hourly, 1-h daily max,
1-h daily avg, and 8-h daily max CO concentrations broken down by spatial sampling scale. The
different sampling scales included in the table (microscale, middle scale, neighborhood scale and
urban scale) were defined in Section 3.4.2.1. While monitors classified under all four scales are used
for highest concentration monitoring and regulatory compliance, individual monitors are classified
by spatial scale to be used for addressing more particular monitoring objectives. Microscale, middle
scale, and neighborhood scale monitors are used to quantify source impacts while neighborhood
scale and urban scale monitors are used for population-oriented monitoring (40 CFR Part 58
Appendix D). For CO, traffic is the major source in an urban setting and therefore microscale data
are sited "to represent distributions within street canyons, over sidewalks, and near major roadways"
with at least one monitor sited to capture maximum concentrations, while middle scale monitors are
sited to represent "air quality along a commercially developed street or shopping plaza, freeway
corridors, parking lots and feeder streets" (40 CFR Part 58 Appendix D). The data used to create
Table 3-12 were subject to the same 75% completeness criteria described in Section 3.5.1.1. More
than 50% of the reported hourly data fell below the reported LOD (reported as 0.5 ppm for the
majority of monitors reporting to AQS).
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
1,428,745
771 ,941
2,878,993
279,311
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
59,905
32,659
121,328
11,784
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
59,905
32,659
121,328
11,784
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
59,905
32,659
121,328
11,784
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
January 2010
3-53
-------
The median hourly CO concentration across the U.S. obtained at microscale monitors was
25% higher than at middle scale and 67% higher than at neighborhood scale. However,
measurements at or below the median hourly concentration were almost entirely below the LOD for
all scales, thereby limiting the usefulness of hourly median comparisons. The upper percentiles (90%
and above), however, were all above the LOD and reveal consistently lower hourly concentrations
for the urban scale monitors relative to the other monitors. For example, the 99th percentile of
reported hourly values was 2.2, 2.3, and 2.1 ppm for microscale, middle scale and neighborhood
scale, respectively, compared to 1.6 ppm for urban scale. Similar patterns were present in the 1-h
daily max, 1-h daily average, and 8-h daily max distributions. Overall, the urban scale nationwide
distributions tended to have lower concentrations relative to neighborhood scale, middle scale and
microscale distributions (Table 3-12).
Distributions categorized by spatial scale and CSA/CBSA are provided in Figure 3-27 for
hourly data and in Figure 3-28 for 1-h daily max data for the select CSAs/CBSAs where data were
available at multiple scales (not all scales were reported by each CSA/CBSA studied). Tables A-17
through A-26 of Annex A contain tabular distributions for all CSAs/CBSAs except Anchorage. On a
city-by-city basis, there was considerable variability when comparing distributions at the available
spatial scales. With a few exceptions, however, the distribution of microscale and middle scale
monitors tended to be higher than those obtained from neighborhood and urban scale monitors. For
example, in CSAs/CBSAs containing both microscale and neighborhood scale monitors (Boston,
Denver, Houston, Los Angeles, New York and Phoenix), median hourly concentrations at monitors
sited for microscale were 20-40% higher than for middle scale and 0-150% greater than those sited
for neighborhood scales. At the 99th percentile, microscale concentrations ranged from 31% less
than to 59% greater than middle scale concentrations and from 14% less than to 67% greater than
neighborhood scale. For most cities, the median hourly data are near or below the 0.5 ppm LOD
reported for most monitors in use. In general, these data suggest that CO concentrations measured
with monitors sited at micro- and middle scales, typically near roads, were somewhat elevated
compared with neighborhood and urban scale monitor locations. However, the magnitude of these
differences varies by city and is difficult to discern given the predominance of CO concentrations
near or below the LOD.
Despite differences in concentrations observed at different scales (Figure 3-27 and Figure
3-28), intersampler correlations do not follow a distinct trend with respect to spatial monitoring scale
(Table 3-10 and Table 3-11). For instance, intersampler correlation in Denver ranged from 0.46 to
0.76 among microscale monitors and was 0.52 for the correlation between the two neighborhood
scale monitors (no monitors in Denver reporting to the AQS are sited at middle scale). Intersampler
correlation in Los Angeles ranged from 0.44 to 0.73 for middle scale, and the one pair of
neighborhood scale monitors had a correlation of 0.43. Only one monitor was sited each at
microscale and urban scale, and 16 of the 24 CO monitors in Los Angeles are not declared to sample
at any spatial scale (scale designation = "null"). In Denver, the distribution of hourly CO data
obtained at microscale was nearly identical to that obtained at neighborhood scale. In Los Angeles,
the microscale data was typically higher than middle, neighborhood, or urban scale data except at the
upper end of the distribution, where middle scale data were higher for both hourly and 1-h daily max
data (Figure 3-27 and Figure 3-28).
January 2010 3-54
-------
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
1.0
0.8
0.6
0.4
0.2
0.0
20
40 60
Percentile
Los Angeles
20
40 60
Percentile
Phoenix
80
100
80
100
20
40 60
Percentile
80
100
20
40 60
Percentile
New York City
20
40 60
Percentile
Pittsburgh
20
40 60
Percentile
80
100
80
100
80
100
Figure 3-27.
Distribution of hourly 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
January 2010
3-55
-------
Atlanta
Boston
0.0
20
40 60
Percentile
80
100
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
20
40 60
Percentile
80
100
80
100
80
100
Figure 3-28.
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
January 2010
3-56
-------
The microscale and middle scale CO data reported here are consistent with hourly
concentrations reported in the literature for the near-road environment within the U.S. Baldauf et al.
(2008, 190239) reported CO concentrations obtained 20 m from an interstate highway in Raleigh,
NC, to have a median around 0.25 ppm and with maximum concentration <4.0 ppm. Zhu et al.
(2002, 041553) reported CO concentration of 1.9-2.6 ppm at a distance of 17 m from an interstate
highway in Los Angeles, with concentration decreasing exponentially with distance from the
highway. Zhu et al. (2002, 041553) observed on-road CO concentrations to be approximately 10
times higher than at an upwind monitoring site, as shown in Figure 3-29. Concentrations continued
to decrease and were still two times higher than upwind levels at a monitoring site 300 m away.
Baldauf et al. (2008, 190239) also reported a drop in concentration at a monitoring site 300 m from
the road compared with the 20 m site. Figure 3-30 illustrates the distribution of measurements taken
throughout a day. In this plot, the near-road (20 m distance) CO concentrations tended to be
significantly higher than those obtained at 300 m, and the daily variability in the CO concentration
time series was greater at the 20 m site than at the 300 m site. The ratio of 20 m to 200 m
concentrations was higher for the Zhu et al. (2002, 041553) paper. This was likely due to the fact that
the 300 m site was always downwind in Zhu et al. (2002, 041553). whereas winds were more
variable in Baldauf et al. (2008, 190239). Other near-road measurements reported in the literature are
similar to those from the Zhu et al. (2002, 041553) and Baldauf et al. (2008, 190239) studies. Chang
et al. (2000, 001276) reported near-road ambient CO measurements obtained in downtown Baltimore
(distance to road not specified) in the range of 0.5-1.3 ppm. Riediker et al. (2003, 043761) reported
measurements of CO concentration obtained near one of four heavily-trafficked roads in Wake
County, NC, to average 1.1 ppm (range: 0.4-1.7 ppm). Neighborhood scale measurements reported
in the literature were also consistent with if not slightly lower than those reported by AQS. Gentner
et al. (2009, 194034) reported CO concentrations ranging from roughly 0.4-0.9 ppm in Riverside,
CA, 1 km east of an interstate highway. Singh et al. (2006, 190136) reported 24-h avg CO
concentrations, obtained with a 0.04 ppm LOD CO monitor in Long Beach, CA, within 0.5 km and
1.5 km of two interstate highways, to range from 0.2-1.4 ppm.
1.0
0.0
Upwind -200
-100 0 100 200 300 Downwind
Distance to the 710 Freeway (m)
Source: Reprinted with Permission of Elsevier Ltd. from Zhu et al. (2002, 0415531
Figure 3-29. Relative concentrations of CO and copollutants at various distances from the
1-710 freeway in Los Angeles.
January 2010
3-57
-------
z.u
IT
Q.
~ 1,5
CO
o
•" ^™
^rf
(TJ
- 1.0 -
c=
0
f —
8 0.5
O
0
o.o -
i
i
a c
i i
• 20 m site
a 300 m site
i
•
^
H-hiiihiii
i
4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00
Figure 3-30.
Time of day (hrmin)
Source: Reprinted with Permission of Air and Waste Management Association from Baldauf et al. (2008, 1 90239)
CO concentration time series 20 m and 300 m from the I-440 highway in Raleigh,
NC. Symbols denote the mean concentration, and whiskers denote standard
deviations.
Determinants of spatial variability in ambient CO concentration include roadway density,
traffic counts, meteorology, and natural and urban topography. Mobile sources are the largest single
source of CO, and their abundance and density affect the magnitude of CO production. Rodes et al.
(1998, 010611) compared traffic volume, roadway type, and concentrations of CO and several
copollutants in Los Angeles and Sacramento, CA, in a study of on-road traffic emissions. They noted
that there was little difference in CO concentration between arterial roads and freeways for Los
Angeles. Rodes et al. (1998, 010611) found that traffic was also much more congested throughout
Los Angeles. This finding was not surprising given that Los Angeles is a much larger city with
substantially higher traffic volumes than Sacramento. Under similar wind conditions, morning
concentrations were much higher in Los Angeles than Sacramento. Rodes et al. (1998, 010611)
observed that high afternoon winds ventilate Los Angeles, but Sacramento is not as well ventilated.
As a result, Sacramento has nearly the same concentrations as Los Angeles in the afternoon. This
observation is consistent with measurements by Gentner et al. (2009, 194034). showing that CO
concentrations varied inversely with wind speed.
Measured on-road and road-side CO concentrations may also relate to the traffic volume.
Among the 291 active sites where monitors met completeness criteria during 2005-2007, 57 were
declared by state agencies as microscale with average annual daily traffic (AADT) counts on the
nearby roads ranging from 500 vehicles per day at one site in Denver, CO to 133,855 vehicles per
day in Tampa, FL with a geometric mean of 17,462 vehicles per day and a geometric standard
deviation of 2.5 (Table A-2 of Annex A). Within a geometric standard deviation, the data range from
6,576-40,000 vehicles per day. Only two monitors were sited at roads with 100,000 vehicles per day
or more. In contrast, the site where Zhu et al. (2002, 041553) collected data had 160,000-178,000
vehicles per day in 2001 (Cal Trans, 2009, 194036). Microscale sites near roads in the mid-range of
the traffic count data may record data that are not substantially different from those obtained from
neighborhood scale measurements, as indicated in Table 3-12. Likewise, with little microscale data
at roads with AADT of more than 100,000 vehicles per day, there is still much uncertainty regarding
the magnitude of concentrations in the near-road environment.
Field measurements, computational modeling, and wind tunnel experiments have shown that
roadway design, roadside structures and vegetation, and on-road traffic levels can affect
concentrations of CO and other pollutant concentrations near roadways. Field measurements
reported by Baldauf et al. (2008, 191017) indicated that noise barriers could reduce near-road
pollutant concentrations by as much as 50%, although this effect was highly dependent on
January 2010
3-58
-------
meteorological conditions; these results are illustrated in Figure 3-31. This study also showed that
the presence of mature vegetation further reduced concentrations and flattened the concentration
gradient away from the road. Urban dispersion and wind-field modeling by Bowker et al. (2007,
149997) also demonstrated the influence of noise barriers and vegetation on the concentrations and
spatial variability of nonreactive pollutants emitted from traffic sources. Heist et al. (2009, 194037)
ran wind tunnel experiments using a model of a road with different roadside features and a tracer gas
line source emitted from the simulated road to study how concentrations of gaseous traffic emissions
vary spatially in the near-road environment. They demonstrated that noise barriers and roadway
design characteristics, such as the presence of embankments and elevated roadway segments, can
alter airflow and contaminant dispersion patterns in the near-road environment. For example, their
results indicated that roadway design having below-grade sections of road and embankments
reduced concentrations away from the road. These results showed similar concentrations as those of
Zhu et al. (2002, 041553). both for roadway segments at-grade with no obstructions to air flow and
for elevated roadway segments with different road fill conditions. Additionally, Khare et al. (2005,
194016) illustrated in a wind tunnel study that vertical dispersion of a nonreactive gas increased with
increasing simulated traffic volume; this effect was also sensitive to changes in approaching wind
direction. These studies taken together suggest that localized turbulence induced by roadside
structures, roadway design, and traffic provide some mixing and resulting dilution of the CO
concentration in the near-road environment; the extent of mixing effects varies by meteorological
conditions and the specific roadway design and traffic loading.
Q.
2.0
1.5
3 1.QH
8 0-5
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: Reprinted with Permission of Elsevier Ltd. from Baldauf et al. (2008,191017
Figure 3-31.
CO concentration profile 10 m from I-440 in Raleigh, NC, behind a noise barrier
and in open terrain.
The geometry of urban street canyons has a profound effect on the distribution of CO
concentrations on a microscale. A number of studies have performed computational and wind tunnel
modeling of street canyons using nonreactive tracers and demonstrated the potential variability in
concentration within a canyon (e.g., Borrego et al., 2006, 155697; Chang and Meroney, 2003,
090298: Kastner-Klein and Plate, 1999, 001961: So et al., 2005, 110746: Xiaomin et al., 2006,
156165). Because CO is a pollutant with very low reactivity on urban and regional scales, results
from these models are directly relevant to CO concentration distributions in street canyons.
Parameters influencing street canyon dispersion include canyon height to width ratio (H/W), source
positioning, wind speed and direction, building shape, and upstream configuration of buildings.
Figure 3-32 shows dimensionless concentrations obtained from wind tunnel and computational fluid
dynamics simulations of tracer gas transport and dispersion in an infinitely long street canyon with a
line source centered at the bottom of the canyon (Xiaomin et al., 2006, 156165). When the canyon
height was equal to the street width (typical of moderate density suburban or urban fringe residential
neighborhoods) and lower background wind speed existed, concentrations on the leeward
(downwind) canyon wall were four times those of the windward (upwind) wall near ground level.
January 2010
3-59
-------
When the canyon height was twice the street width (typical of higher-density cities) and background
winds were somewhat higher, near ground-level concentrations on the windward canyon wall were
roughly three times higher than those measured at the leeward wall. These results suggest that the
magnitude of microscale CO concentrations may vary by factors of three or four times at different
locations within a street canyon and are heavily influenced by wind speed and street canyon
topography. The relationship between in-cany on concentration and wind speed and turbulence is
well established with concentration varying inversely with the magnitude of wind speed and
turbulence (Britter and Hanna, 2003, 090295). When studying the effect of wind direction on street
canyon concentration levels for a continuous "line source" of traffic exhaust, concentration levels
were at local maxima under two conditions: wind perpendicular to or parallel to the street canyon.
Wind gusts at the turbulence interface at the top of the canyon or traffic-based turbulence can also
cause dilution of the exhaust concentration within the canyon (Kastner-Klein et al., 2000, 194035).
A
windward,
measured
simulated
H 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
Source: Reprinted with Permission of Elsevier Ltd. from Xiaomin et al. (2006, 1 561 65)
Figure 3-32. 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).
Street canyon field studies support the computational and wind tunnel modeling results
described above. In a multisite survey of curbside CO concentration in London, U.K., Croxford and
Penn (1998, 087176) observed up to threefold differences in concentration related to the side of the
January 2010
3-60
-------
street on which the monitor was positioned relative to the wind direction, with H/W varying between
0.7 and 1.7 depending on position within the canyon. Bogo et al. (2001, 192378) measured CO
concentrations in a street canyon with H/W of 1 in Buenos Aires, Argentina, using a continuous CO
monitor. Similar to the Xiaomin et al. (2006, 156165) simulation results for H/W of 1, Bogo et al.
(2001, 192378) observed slightly higher leeward concentrations than windward concentrations
within the canyon, where recirculating airflow inside the canyon causes pollutants to collect in
higher concentration on one side. However, for the case of a deep street canyon (H/W of 5.7) in
Naples, Italy, Murena et al. (2008, 194038) observed that the concentrations on two sides of the
canyon differed by <15%, with wind direction varying between 10° and 80° from the street axis.
Doran et al. (2003, 143352) measured CO concentration in a street canyon in Phoenix, AZ, during
the morning hours and observed that CO concentration decreases with elevation above the ground if
turbulent mixing is small, but that the difference between ground level and 39th-floor (50 m AGL)
measurements of CO concentration decreases when turbulent mixing increases (with maximum
measurements at any elevation not exceeding 2 ppm). As shown in Figure 3-33, the larger difference
in concentration as a function of turbulent mixing can occur when there are meteorologically stable
conditions in the lower boundary layer. These results support findings from the modeling studies that
CO concentration can vary by several times within a street canyon and is greatly influenced by local
meteorology and building topography.
CD
U
OJ
i_
~-
0 ° * D -
O °
*n Qn DD n
no gin
-10
-50 5 10 15
bulk Richardson number
20
Source: Reprinted with Permission of Elsevier Ltd. From Doran et al. (2003,1433521
Figure 3-33. 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.
Research by Kaur and Nieuwenhuijsen (2009, 194014) and Carslaw et al. (2007, 148210)
suggests that CO exposures are related to traffic volume and fleet mix in the street-canyon
environment. Kaur and Nieuwenhuijsen (2009, 194014) used multiple linear regression to model CO
concentration data from central London as a function of mode of transport (broken down by vehicle
type), traffic count, wind speed, and temperature. They added each variable successively and found
traffic count, temperature, wind speed, and walking to be significant parameters in the model, with
traffic count being the strongest determinant. Analysis of variance showed variability in traffic count
to explain 78% of the variability in CO levels for these data, and variability in mode of transport
explained 6% of the variability. Likewise, Carslaw et al. (2007, 148210) used a generalized additive
model to determine how CO concentration data (log-transformed) obtained in central London varied
as a function of light- and heavy-duty traffic counts, along-street and cross-street components of
January 2010
3-61
-------
wind, temperature, year, and Julian day. Light-duty vehicle count was a more important determinant
of CO concentration than heavy-duty (i.e., diesel) vehicle count in this study. They found that the CO
declined steadily with year and that wind was the most significant covariate. In addition to showing
meteorology to be an important determinant of concentration, these modeling exercises also suggest
a linear or log-linear relationship between concentration and traffic.
3.5.2. Temporal Variability
3.5.2.1. Multiyear Trends
Figure 3-34 (top) shows ambient CO concentrations in ppm from 1980 to 2006 based on
continuous measurements averaged over 8-h time segments. Figure 3-34 (bottom) depicts trends in
the annual second-highest 8-h CO concentrations for 144 sites in 102 counties nationwide having
data either in the SLAMS network or from other special purpose monitors.
The 2006 annual second highest 8-h CO concentration averaged across 144 monitoring sites
nationwide was 75% below that for 1980 and is the lowest recorded during the past 27 yr (Figure
3-34 [top]). Since 1992, more than 90% of these sites have reported second highest CO
concentrations below the 8-h NAAQS of 9 ppm. The mean annual second highest 8-h ambient CO
concentration has been below 5 ppm since 2004. The downward trend in CO concentrations in the
1990s parallels the downward trend observed in CO emissions, attributed largely to decreased
mobile source emissions. In addition, of the 144 sites used to determine this trend, from a total of
375 monitoring sites operating in 2006, the number reporting second-highest 8-h CO concentrations
above the level of the NAAQS declined to zero over the same period (Figure 3-34 [bottom]).
Consistent with the nationwide trends in emissions and concentrations, CO concentrations in
all 10 EPA Regions have steadily decreased since 1980, with reductions over this period ranging
from 68% in Region 7 to 85% in Region 1 (Figure 3-35). This is also consistent with declining
emissions seen in many regions of the U.S., shown in Figure 3-5.
January 2010 3-62
-------
-a o
c -tj
o «
- -
cu =
co a>
8
16
14
12
10
8
6
4
2
0
1
90% of sites have concentrations below this line
NAAQS = 9 ppm
10% of sites have
concentrations below this line
'80 '82 '84 '86 '88 '90 '92 '94 '96 '98 '00 '02 '04 '06
Year
E OD
'80 '82 '84 '86
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
Coverage: 144 monitoring sites in 102 counties nationwide (out of
a total of 375 sites measuring CO in 2006) that have sufficient
data to assess CO trends since 1 980.
Source: U.S. EPA (2008, 1570761
Figure 3-34. (Top) Trends in ambient CO in the U.S., 1980-2006, reported as the annual second
highest 8-h concentrations (ppm) for the mean, median, 10% and 90% values.
(Bottom) Trends in ambient CO in the U.S., 1980-2006, reported as the number of
trend sites (y-axis) with annual second highest 8-h concentrations
above the level of the NAAQS (9 ppm).
January 2010
3-63
-------
l\l A l\
no
n
•R1
R2
•R3
R4
•R5
R6
R7
R8
R9
•R10
•Nat'I
'80 '82 '84 '86 '88 '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, 157076'
Figure 3-35. Trends in ambient CO in the U.S., 1980-2005, reported as the annual second
highest 8-h concentrations (ppm) for the EPA Regions 1 through 10, along with a
depiction of the geographic extent of those Regions.
January 2010
3-64
-------
3.5.2.2. Hourly Variation
Weekday and weekend diel variation for the mean, median, 5th, 10th, 90th, and 95th
percentiles of hourly CO concentration over 2005-2007 are shown in Figure 3-36 and Figure 3-37,
respectively, for the 11 CSAs and CBSAs examined in this assessment. Since these figures represent
the distribution of hourly observations over a 3-yr period, any fluctuations or changes in the timing
of the daily peaks would result in a broadening of the curves shown in the diel plot compared to the
actual daily temporal behavior on any specific day measured by an individual monitor. However,
these figures are useful for comparing the general hourly variation in CO concentrations across cities
and by day of the week (i.e., weekday versus weekend). The weekday data showed that the
Anchorage mean, median, 5th and 10th percentile CO concentration curves exhibit pronounced
morning and evening rush hour peak CO levels. Boston, Denver, Houston, Los Angeles, Phoenix,
Pittsburgh, and St. Louis all exhibited similar trends, although the magnitude of the concentrations
shown was roughly twice as high for Anchorage as the other cities. The curves had less overall
variability for Boston, Pittsburgh, and St. Louis. The Atlanta plot shows that the median
concentration was fairly constant throughout the 24-h period, with a slightly elevated mean during
the morning hours. The 90th and 95th percentile curves exhibit stronger morning and evening CO
concentration peaks. New York City shows fairly constant CO mean and median concentration
throughout the day, with slight elevations throughout the morning rush hour and a slight trough
between 1:00 and 5:00 a.m. The Seattle plot shows a daytime plateau beginning around 5:00 a.m.
and lasting until roughly 10:00 p.m., with higher concentrations during morning and afternoon rush
hour. Differences in hourly variation among the 11 CSAs and CBSAs reflect city-to-city variation in
source characteristics and meteorology. For instance, the rush hour peaks in many cities likely
correspond to increased mobile source emissions during those periods. Local meteorology and
topography, which influence mixing heights, can also affect hourly variation in CO concentration.
Figure 3-37 illustrates weekend diel trends for the 11 CSAs and CBSAs considered in this
assessment. For Anchorage during the period 2005-2007, the mean and median concentration curves
peaked during the morning and evening hours. A daytime concentration trough is evident. The 90th
and 95th percentiles of concentration were similar but more pronounced. The shape of this plot is
also characteristic of Atlanta, Boston, Denver, Houston, Los Angeles, Phoenix, Pittsburgh, Seattle,
and St. Louis, although the Anchorage CO concentrations are nearly 100% higher than
concentrations in the other cities. The weekend diel plot for New York City shows that the mean and
median CO concentrations remain fairly constant throughout the day, with a slight reduction between
2:00 and 7:00 a.m. The 90th and 95th percentile curves illustrate more diel variation.
January 2010 3-65
-------
CO (ppm)
0 +
1
3
?22
Q.
S 1.5
O
0 0.8
0
Anchorage
Weekday (N = 771)
3 -
•"/V\ 2.2
"-----'' /"' '•-, \ •••.-•''-•' 0.8 -
_ _, , , , | — o 4
6 12 18 24 1
Denver
Weekday (N = 3806)
3 -
2.2 •
/-•'~,'\ / ""••-... 1.5-
.."_-• •'• '-^f- X-'"-. -••'-•'' "'' 0.8-
j- ,-.--.-r";' — , ••,•••••••••- /•;••;•-•• , ^i^, o-
16 12 18 24
3 -
Q.
5 1.5 -
O
0 0.8-
0-
3 -
E "'
Q.
3 1.5-
o
0 0.8-
0 -
New York
Weekday (N- 6821 1
5.8 •
...... 2.9 -
irmTr.-----7 -------
6 12 18 24 ° 1
St. Louis
Weekday (N = 2303)
3 •
2.2
1.5 •
6 12 18 24 OH
Atlanta
Weekday (N = 231 7)
6 12 18 24
Houston
Weekday (N = 3734)
6 12 18 24
Phoenix
Weekday (N = 3836)
6 12 18 24
Seattle
Weekday (N = 776)
6 12 18 24
3 -
2.2 -
1.5 -
0.8 -
0 -
3 -
2.2 -
1.5 -
0.8 -
0 -
3 -
2.2 -
1.5 -
0.8 -
0 -
Boston
Weekday (N = 521 6)
6 12 18 24
Los Angeles
Weekday (N = 15437)
6 12 18 24
Pittsburgh
Weekday (N = 5352)
6 12 18 24
Median
Mean
90th & 10th
95'th&5'th
Figure 3-36. Diel plot generated from weekday hourly CO data (ppm) for the 11 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.
January 2010
3-66
-------
5.8-
^~. 4.4
Q.
3 2.9
O
0 1.5-
0-
3
Q.
S 1.5
O
0 0.8
0
3
Q.
S 1.5
o
0 0.8
0
3
Q.
S 1.5-
o
0 0.8-
*
Anchorage
Weekend (N = 309)
3 •
2.2
/•"••, •--... 1.5-
-~ 0 •
6 12 18 24
Denver
Weekend (N = 1564)
3 -
..-••--.. 1.5-
Atlanta
Weekend (N - 932)
3 •
2.2 -
1.5 -
**-> ...•••.•-'-•"' 0.8-
Boston
Weekend (N = 2155)
, , , , o !•••-•---- i- T "
6 12 18 24 1 6 12
Houston
Weekend (N = 151 2)
3
1.5 •
hr-"-n--" — :-:--r..-------"--. .-.•••.-, o -I , , -, , OH
16 12 18 24 1 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 •
X-^---'"~_~.I.'.. _.-^^*<^ 08.
^-^^^
i 6 12 18 24 0 •
Phoenix
Weekend (N = 1550)
3 -
2.2 -
~~-~"X'\. .;'-' _ 0.8-
i " ~~i -"• *•! — " i 0 4
6 12 18 24
St. Louis
Weekend (N- 931)
^^ ^^m
6 12 18 24
18 24
Los Angeles
Weekend (N = 621 6)
^^S^^<^-^^.>^^^:-
6 12
Pittsburgh
Weekend (N = 21 61)
18 24
6 12 18 24
Median
Mean
90'th&10'th
I
Figure 3-37. Diel plot generated from weekend hourly CO data (ppm) for the 11 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.
January 2010
3-67
-------
3.5.3. Associations with Copollutants
Associations between hourly CO and other copollutants, including SO2, NO2, O3, PMi0, and
PM2.5 are provided in box plots in Figure 3-38 using AQS data across the U.S. AQS data were
obtained from all available co-located monitors across the U.S. after application of the 75%
completeness criteria described earlier in Section 3.5.1.1. Pearson correlation coefficients (r) were
calculated using 2005-2007 data stratified by season. Correlation plots analogous to Figure 3-38 for
select individual cities are provided in Annex A, Figures A-43 to A-48.
-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
PM,
Xh-h
•-KB-
-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
correlation (r) with CO correlation (r) with CO
Figure 3-38. Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02, Os, PM™ 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.
In all cases, a wide range of correlations existed between CO and copollutants as illustrated in
Figure 3-38. The mean and median correlation between CO and copollutants were positive for NO2,
PMio, and PM2 5; near zero for SO2; and negative for O3. These findings reflect common combustion
sources for CO, NO2, and PM. CO is highly correlated with NO2 and PM2 5 because they are both
emitted directly during incomplete combustion and because secondary nitrate PM comes from NOX,
which is largely produced from mobile sources. Among those copollutants with positive
associations, NO2 had the highest mean and median correlations, followed by PM2 5 and PM10
(correlations vary by season). The IQR of correlations with SO2 spanned from positive to negative
for all seasons; SO2 would not be expected to correlate well with CO because SO2 emanates
primarily from industrial sources. Correlations between CO and O3 were almost entirely negative for
January 2010
3-68
-------
winter, when CO emissions tend to be high and O3 formation is low. During the other three seasons,
most of the CO-O3 correlations were also negative. Given the role of CO in O3 chemistry, cross-
correlation functions were also computed by season for the CO-O3 relationship (Annex A, Figure
A-50). The data showed negative correlations at all lags with minima at zero lag for winter, spring,
and fall. During the summer, a small positive peak correlation (r = 0.071) was centered at a lag
of -8 h and a minima occurred at a lag of 1 h, r = -0.272. It is not known whether the positive lagged
correlations in summertime are related to interaction of CO with O3 through chemistry, coinciding
independent effects such as peak times for rush hour CO emissions and afternoon O3 formation, or
some combination of interactive and independent effects. However, given the very low magnitude of
these correlations, it is clear that many other factors influence the O3 and CO time series.
Within and between individual metropolitan areas, the distribution of copollutant correlations
varied substantially. Figure 3-39 and Figure 3-40 illustrate the correlations between CO and
copollutants for Denver, CO, and Los Angeles, CA, to exemplify these differences. Copollutant
correlation plots are also shown for other cities in Annex A, Figures A-44 through A-49. For
instance, correlations between CO and copollutants are all positive for SO2, NO2, PMi0, and PM2.5
and are all negative for O3 in Denver. In contrast, the correlations in Los Angeles span from negative
to positive for O3, PMi0, and PM2.5, in various seasons. The larger span of correlations for Los
Angeles in comparison with Denver could result from several factors. For example, more variation
in meteorology, topography, or source distribution with respect to monitor placement in Los Angeles
may cause the distribution of copollutant correlations to be wider. In addition, fewer co-located
monitors in Denver compared with Los Angeles may be causing some of the observed differences.
Winter
S p ri n g
S02-
N02-
Os-
PMio-
PM2.5-
(DO
GD
O
00
0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4
0.8 1.0
S02-
N02-
03-
PMio-
PM2.5-
OJD
©
O
0
0
-1.0
-0.4 -0.2 0.0 0.2 0.4
0.8 1.0
Summer
Fall
S02-
N02-
Os-
PMio-
PM2.5-
(Q) GD
O
oo
O
O
S02
NO
03
PMio
PM2.5'
(O£>
OO
O
(0)
0
-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)
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
r (correlation coefficient)
Figure 3-39. Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02, 03, PM10 and PM2.s concentrations
for Denver, CO.
January 2010
3-69
-------
Winter
S02
N02
Ob-
PMio-
PM2.5 •
OOGDOCD)
O (QX381Q
QBE) CD
S02
N02
03
PMio
PM2.5
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
S p ri n g
<2B
GOOD
CXOHHffiJ)
(08DO
O
-1.0 -0.8
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Sum mer
PMio-
PM2.5 •
GO
GOOD
M»I;»MI;:>»
ooosaooo
S02
N02
03
PMio
PM2.5
Fall
(DOOGDO
-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)
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
r (correlation coefficient)
Figure 3-40. Seasonal plots showing the variability in correlations between hourly CO
concentration and co-located hourly S02, N02, Os, PMio and PM2.s concentrations
for Los Angeles, CA.
Several recent studies reported correlations between ambient CO and other pollutants.
Reported relationships were generally consistent with the correlations shown above using AQS data.
Sarnat et al. (2001, 019401) reported significant positive Spearman's correlations between CO and
NO2 (r = 0.76) and PM2.5 (r = 0.69) and significant negative correlations between CO and O3
(r = -0.67) in Baltimore (concentration averaging periods not specified). Correlation of CO with SO2
was insignificant (r = -0.12). The Sarnat et al. (2001, 019401) study focused on correlations of
ambient and personal PM2.5 with gaseous copollutants, so seasonal information is only available for
the correlation between PM2 5 and CO. High correlation of ambient CO with NO2 is expected given
that both are closely related to mobile source combustion emissions. Sarnat et al (2005, 087531) also
reported significant year-round association between CO and PM2 5 and significant associations
between CO and SO42" aerosols. Kim et al. (2006, 089820) measured CO, NO2, and PM2.5 at ambient
fixed sites in Toronto, Canada, and found associations, averaged over monitoring stations, of CO
with PM2 5 (Spearman's r = 0.38, nonsignificant) and of CO with NO2 (r = 0.72, significant). Tolbert
et al. (2007, 090316) reported correlations between multiple pollutants in Atlanta and also showed
the highest Spearman's correlation for CO with NO2 (r = 0.70). CO was also reported to have fairly
high correlation with PM25 elemental carbon (EC) (r = 0.66), PM25 organic carbon (OC) (r = 0.59),
and PM25 total carbon (TC) (r = 0.63). Correlations were reported to be much lower for CO with O3
(r = 0.27) and PM2.5 SO42" (r = 0.14). The higher correlations of CO with EC, OC, and TC are likely
related to the fact that CO and carbonaceous PM are both emitted by mobile sources. Gentner et al.
(2009, 194034) analyzed the relationship between ambient CO and VOC concentrations, serving as
markers of gasoline vehicle emissions in Riverside, CA. Correlations of CO with two compounds,
n-butane and isopentane, are shown in Figure 3-41 for summer and fall. Higher concentrations of
n-butane per unit of CO were observed for fall, as well as higher correlation (fall: r = 0.88; summer:
r = 0.65). For isopentane, the slopes of regression are much closer for fall and summer, with higher
correlations between isopentane and CO (fall: r = 0.93; summer: r = 0.86). Gentner et al. (2009,
194034) noted that isopentane vapor fraction was higher in summer than winter and that the n-butane
vapor fraction increases in winter. This reflects the higher volatility of n-butane compared with
isopentane. In this work, Gentner et al. (2009, 194034) used emissions modeling to estimate that
overall VOC emissions from gasoline varies with CO emissions with a ratio of 0.086 with a
correlation of r = 0.80 in summer. Gentner et al. (2009, 194034) suggest that the near-road slope of
January 2010
3-70
-------
ambient VOC to CO concentration might be influenced by upwind CO concentration and secondary
CO production by oxidation of VOCs.
(a)
6-
-Q
Q.
4H
-D
2-
(b)
6-
_
-Q
Q.
CD
4->
c
a.
o
in
4H
2-
D Fall (r = 0.88)
Summer (r = 0.65)
.
400 800 1200 1600
Carbon Monoxide [ppbv]
D Fall (r = 0.93)
Summer (r = 0.86)
I i
400 800 1200 1600
Carbon Monoxide [ppbv]
Source: Reprinted with Permission of ACS from Centner et al. (2009,1940341
Figure 3-41. Linear regression of n-butane and isopentane concentration as a function of CO
concentration, Riverside, CA.
January 2010
3-71
-------
3.5.4. Policy-Relevant Background
Background concentrations of pollutants used for informing policy decisions about national
standards in the U.S. are commonly referred to at EPA as PRB concentrations. In this assessment,
PRB concentrations include contributions from natural sources everywhere in the world and from
anthropogenic sources outside the U.S., Canada, and Mexico.
3.5.4.1. Surface-Based Determinations
For this assessment, PRB concentrations of CO were determined from the extensive and long-
running network of remote-site baseline CO measurements conducted by NOAA's Earth System
Research Laboratory (ESRL), Global Monitoring Division (GMD), as part of their Carbon Cycle
Greenhouse Gases Group (CCGG) Cooperative Air Sampling Network (CASN); see
http://www.esrl.noaa.gov/gmd/ccgg/iadv. Surface-based CO measurements have been made for more
than 10 yr with exceptionally high sensitivity and selectivity at locations significantly away from
local sources. In this assessment, for example, CO data through December 2007 are available with
extensive quality assurance and control information from the worldwide network of 72 stations
active in December 2008. ESRL GMD uses the highly sensitive gas chromatography-mercury
liberation photometric detection technique with precision to 1 ppb in 50 ppb or 2 ppb in 200 ppb and
accuracy to 1.5 ppb in 500 ppb or 2 ppb in 200 ppb.
In order to smooth interannually changing meteorological and emissions effects, data from
2005-2007 at 12 remote sites in the U.S. were used to determine PRB. A map of these sites is shown
in Figure 3-42; they are: Cold Bay, AK; Barrow, AK; Shemya Island, AK; Cape Kumukahi, HI;
Mauna Loa, HI; Trinidad Head, CA; Point Arena, CA; Wendover, UT; Niwot Ridge, CO; Park Falls,
WI; Southern Great Plains, OK; and Key Biscayne, FL. These sites are affected by anthropogenic
emissions in North America to varying degrees. Average concentrations for each month and for each
of the 3 yr are shown for each site in Figure 3-43. All sites demonstrate the well-known seasonality
in background CO with minima in the summer and fall and maxima in the winter and spring in the
Northern Hemisphere. Northern Hemisphere summer-time minima are related in large measure to
the enhanced photochemical reaction of CO with OH, as described in Section 3.2.2. Analysis for
North American PRB is made here by segregating the three Alaska sites (owing to their high
latitude) and the two Hawaii sites (owing to their distance from the continent) and treating the
remaining seven sites as being more representative of the CONUS surface-level background
concentrations. Outside the defined CONUS domain used here, the 3-yr avg CO PRB in Alaska
ranged from 127 to 135 ppb with an average of 130 ppb, and from 95.3 to 103.1 ppb with an average
of 99.2 ppb in Hawaii. Over the CONUS domain the 3-yr avg CO PRB concentration ranged from
118 to 146 ppb with an average of 132 ppb.
January 2010 3-72
-------
Pcint Arena.iCalifcm
* Shemya Island, Alaska
* CO monitors
Monitor Location
Figure 3-42. Map of the baseline monitor sites used in this assessment to compute PRB
concentrations.
January 2010
3-73
-------
200
180
60
200
180
160
1J140
0120
Shemya Island, AK
Cape Kumukahi, HI
Niwot Ridge, CO
•g.140
^
0,20
IVIaunaLoa, HI
Trinidad Head, CA
Point Arena, CA
200 -•
180 -•
160 --
•g.140 -•
0120--
100 --
Southern Great Plains, OK
JFMAMJJASONDJ FMAMJJASONDJFMAMJJASOND JFMAMJJASONDJ FMAMJJASONDJFMAMJJASOND JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
2005 2006 2007 2005 2006 2007 2005 2006 2007
Figure 3-43. 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
The significance of CO for surface-level air quality and for its indirect climate forcing effects
through CH4, O3, and CO2 as described previously in this chapter, and its long T relative to that of
other primarily urban and regional pollutants make it an important species for measurement and
evaluation on multiple spatial, temporal, and chemical scales.
In addition to the ESRL GMD surface network used in this assessment's determination of CO
PRB, CO concentrations away from local sources can be measured from space. So, for example, CO
has been observed from space by the Measurement of Air Pollution from Satellites (MAPS)
instrument on Space Shuttle orbiter flights for three 10-day missions in 1984 and 1994 (Connors et
al., 1994, 193755) and by the Measurement of Pollution in the Troposphere (MOPITT) on the Terra
satellite since 2000 (Emmons et al., 2004, 193756). Surface spatial coverage with both space-based
instruments was limited by the common problems of cloud cover, high surface albedo and
emissivities, and image swath pattern and timing, with the result that much of the CONUS, for
example, was missed some of the time. In addition, all of these satellite measurements were limited
though somewhat differently in the vertical resolution of their total column CO concentration values.
January 2010
3-74
-------
For a determination of a PRB-equivalent background concentration for 2008, the MAPS data
would be of no use, except for comparisons on temporal trends, and even that is limited by the very
few observations from MAPS. MOPITT data might seem more useful were it not for MOPITT's
very low precision and accuracy in the lowest few kilometers above the Earth's surface of its
integrated total column CO measurement by thermal infrared radiances (Shindell et al., 2005,
193746). MOPITT CO profile sensitivities are so low at the surface that retrievals at the 850 hPa
level, the lowest reported, do not capture the surface concentration accurately but actually represent
a broad and deep vertical slice of the lower troposphere with an integral concentration that often
peaks well above 850 hPa (Shindell et al., 2006, 091028). Error analysis by Emmons et al. (2004,
193756). reported in Shindell et al. (2006, 091028) revealed that MOPITT concentration error in the
lower troposphere was 7% and had greater bias over cleaner sites, which are of most interest when
estimating a CONUS PRB.
Since the integrated total column measurements of CO from space-borne instruments are
dominated by CO in the mid- and upper troposphere, comparisons to surface measurements are
subject to appreciable error. Using a subset of seven to nine of the ESRL GMD network stations in
North America, for example, to compare to the MAPS and MOPITT data, Shindell et al. (2005,
193746) found that the satellite data showed an increase of between 3 and 13 ppb CO while the
surface data at these locations showed a decrease of 20 ppb in the years 2000-2002 relative to 1994.
Mean global concentrations of CO were apparently decreasing before 2000, but that trend has now
mostly ended (Duncan and Logan, 2008, 194042). so that the integrated column CO total measured
from space may have indicated a false trend.
CO concentrations can also be predicted with numerical CTMs on regional, continental, and
global scales. Hence it would, in principle, be possible to predict CO PRB concentrations for the
CONUS. The chief limitation to this method comes from the highly uncertain emissions of CO
worldwide needed to drive the global CTMs, which in turn set the boundary conditions and chemical
flow fields for the finer-scale models that might be used to compute PRB. Interannual variability in
CO emissions from global biomass burning is very high, and the emissions source strength of this
signal is a very strong component of the CONUS PRB given the CO T of ~57 d. The long T means
that PRB monitoring sites are subject to contamination by regional pollution. Estimates of total
global CO emissions used in recent forward and inverse model experiments range from
<1,100 MT/yr to >3,300 MT/yr (Shindell et al., 2005, 193746).
A comprehensive evaluation of 26 state-of-the-science atmospheric chemistry models
exercised for present-day and future CO simulations was performed and reported by Shindell et al.
(2006, 091028). They found substantial under-prediction of CO in the extra-tropical Northern
Hemisphere compared to satellite and local surface observations and large variability among the
models as well, even when using identical CH4 abundances and CO emissions. In North America, for
example, the multimodel average underestimated the observations of lower troposphere CO by 60
ppb or more, or by -50% or more of the measured background concentration at many of the ESRL
GMD sites. The Pearson r values for the multimodel average against MOPITT data globally for
2000-2001 was 0.84 ± 0.08 for April at 850 hPa (as near to the surface as tested) but only 0.55 ±0.11
in October (Shindell et al., 2006, 091028). Shindell et al. (2006, 091028) proposed several reasons
why this could contribute to pervasive underprediction: (1) the models do not adequately simulate
CO build-up during the wintertime periods of lower OH flux; (2) the models have no seasonal CH4
cycle with build-up in the Northern Hemisphere winter; and (3) variability in the models' OH
concentrations, which accounted for -80% of the CO intermodel variance (Shindell et al., 2006,
091028).
All of the above methods have their own advantages and disadvantages. The levels determined
by the ESRL/GMD network show the temporal and spatial variability of CO levels. Although these
sites are subject to North American pollution sources to varying degrees, these data could be used
provided this caveat is borne in mind. Resulting errors in estimating excess risks will be very small
because the concentrations are only a small fraction of the CO NAAQS.
January 2010 3-75
-------
3.6. Issues in Exposure Assessment
3.6.1. Summary of Findings from 2000 CO AQCD
The 2000 CO AQCD (U.S. EPA, 2000, 000907) describes the results of studies completed
prior to 1999 on personal exposures and microenvironmental concentrations of CO. Although these
studies may no longer be representative of current exposure levels due to declining ambient CO
concentrations, the personal-microenvironmental-ambient relationships are still instructive. Time
spent commuting, particularly in cars, was a major contributor to personal CO exposures. Many
studies measured in-vehicle concentrations of CO and found elevated concentrations compared to
fixed-site monitors. Roadside CO monitors were elevated compared to ambient levels and equal to or
lower than in-vehicle levels (Ott et al., 1994, 076546; Rodes et al., 1998, 010611). A small portion of
the CO concentrations inside a vehicle cabin comes from the vehicle itself, while a substantial
fraction comes from roadway mobile source emissions entering the cabin via air exchange. Studies
summarized in the 2000 CO AQCD found that in-vehicle CO concentrations were generally two to
five times higher than ambient CO concentrations obtained at fixed-site monitors within the cities
studied. High-traffic volumes contributed to increased in-vehicle concentrations.
Prior to the 2000 CO AQCD, it was well known that CO levels in residences may be elevated
above ambient due to nonambient indoor sources, such as cooking, space heating, and smoking.
Separation of indoor CO into ambient and nonambient components was found to be important for
determining the effect of ambient CO concentrations, although this had not been done successfully in
studies conducted to date. Two large studies performed in Denver, CO, and Washington, DC, in the
early 1980s found that fixed-site monitor concentrations were higher than personal exposures for
those with low-level exposures, while fixed-site monitor concentrations were lower than exposures
for those with high-level exposures (Akland et al., 1985, 011618; Johnson, 1984, 024652).
Nonambient sources contributing to high-total exposures likely obscured this relationship. In
Denver, gas stove operation, passive smoking, and attached garages increased residential indoor
exposure by 2.6, 1.6, and 0.4 ppm, respectively, compared to individuals without those sources
present. Categorical analyses found significantly higher personal exposures on high-ambient
concentration days than on low-ambient concentration days, suggesting that personal exposures are
related to ambient levels. Nonambient exposures tend to obscure the relationship between ambient
CO concentrations, as measured at ambient monitors, and total personal CO exposure.
3.6.2. General Exposure Concepts
A theoretical model of personal exposure is presented to highlight measurable quantities and
the uncertainties that exist in this framework. An individual's time-integrated total exposure to CO
can be described based on a compartmentalization of the person's activities throughout a given time
period:
ET=\C}dt
Equation 3-2
where EI = total (T) exposure over a time-period of interest, C, = airborne CO concentration at
microenvironmentj, and dt = portion of the time-period spent in microenvironmentj. Equation 3-2
can be decomposed into a model that accounts for exposure to CO of ambient (Ea) and nonambient
(£na) origin of the form:
ET = Ea + Ena
Equation 3-3
Examples of ambient CO sources include industrial and mobile source emissions, biomass
combustion, and agricultural processes. Examples of nonambient sources include environmental
tobacco smoke (ETS), cooking, and home heating. CO concentrations generated by ambient and
January 2010 3-76
-------
nonambient sources are subject to spatial and temporal variability that can affect estimates of
exposure and resulting health effects. Exposure factors affecting interpretation of epidemiologic
studies are discussed in detail in Section 3.6.8.
This assessment focuses on the ambient component of exposure because this is more relevant
to the NAAQS review. Ea can be expressed in terms of the fraction of time spent in various outdoor
and indoor microenvironments (Wallace et al, 2006, 089190: Wilson et al, 2000, 010288):
Equation 3-4
where/= fraction of the relevant time period (equivalent to dt in Equation 3-2), subscript o = index
of outdoor microenvironments, subscript / = index of indoor microenvironments, subscript o,i =
index of outdoor microenvironments adjacent to a given indoor microenvironment i, and Finf i =
infiltration factor for indoor microenvironment z. Equation 3-4 is subject to the constraint S/0 +
E/i = 1 to reflect the total exposure over a specified time period, and each term on the right hand side
of the equation has a summation because it reflects various microenvironmental exposures. Here,
"indoors" refers to being inside any aspect of the built environment, e.g., home, office buildings,
enclosed vehicles (automobiles, trains, buses), and/or recreational facilities (movies, restaurants,
bars). "Outdoor" exposure can occur in parks or yards, on sidewalks, and on bicycles or motorcycles.
Finf is a function of the building air exchange characteristics. Assuming steady state ventilation
conditions, the infiltration factor is a function of the penetration (P) of CO, the air exchange rate (a)
of the microenvironment, and the rate of CO loss (k) in the microenvironment; F^ = Pa/(a+k).
Given that k -> 0 for CO, Finf reduces to P. Studies of CO infiltration are reviewed in Section 3.6.5.1.
In epidemiologic studies, Ca is often used in lieu of outdoor microenvironmental data to represent
these exposures based on the availability of data. Thus it is often assumed that C0 = Ca and that the
fraction of time spent outdoors can be expressed cumulatively as/,; the indoor terms still retain a
summation because infiltration differs among different microenvironments. If an epidemiologic
study employs only Ca, then the assumed model of an individual's exposure to ambient CO, first
given in Equation 3-4, is re-expressed solely as a function of Ca:
Equation 3-5
Meteorology, strength of CO sources, spatial variability of CO concentration, proximity of the
study population to sources of CO, design of the epidemiologic study, and other factors determine
whether or not Equation 3-5 is a reasonable approximation for Equation 3-4. Errors and uncertainties
inherent in use of Equation 3-5 in lieu of Equation 3-4 are described in Section 3.6.8 with respect to
implications for interpreting epidemiologic studies. Epidemiologic studies often use concentration
measured at a central site monitor to represent ambient concentration; thus a, the ratio between
personal exposure to ambient CO and the ambient concentration of CO, is defined as:
a = —
ca
Equation 3-6
Combination of Equation 3-5 and Equation 3-6 yield:
Equation 3-7
a varies between 0 and 1. If a person's exposure occurs in a single microenvironment, the ambient
component of a microenvironmental CO concentration can be represented as the product of the
ambient concentration and P. Wallace et al. (2006, 089190) note that time-activity data and
corresponding estimates of P for each microenvironmental exposure are needed to compute an
individual's a with accuracy. If local sources and sinks exist and are significant but not captured by
central site monitors, then the ambient component of the local outdoor concentration may be
January 2010 3-77
-------
estimated using dispersion models, land use regression models, receptor models, fine scale CTMs or
some combination of these techniques. These techniques are described in Section 3.6.3.
3.6.3. Exposure Modeling
3.6.3.1. Stochastic Population-Based Time-Weighted Microenvironmental Exposure
Models
Population-based methods, such as the Air Pollution Exposure (APEX) and Stochastic Human
Exposure and Dose Simulation (SHEDS) models, involve stochastic treatment of the model inputs
(Burke et al., 2001, 014050; U.S. EPA, 2009, 194009). These are described in detail in the 2008 NOX
ISA (U.S. EPA, 2008, 157073). in Annex AX 3.6.1. Stochastic models utilize distributions of
pollutant-related and individual-level variables, such as ambient and local CO concentration source
contributions and breathing rate respectively, to compute the distribution of individual exposures
across the modeled population. The models also have the capability to estimate received dose
through a dosimetry model. Using distributions of input parameters in the model framework rather
than point estimates allows the models to incorporate uncertainty and variability explicitly into
exposure estimates (Zidek et al., 2007, 190076). These models estimate time-weighted exposure for
modeled individuals by summing exposure in each microenvironment visited during the exposure
period. For example, Bruinen de Bruin et al. (2004, 190943) utilized the EXPOLIS (exposure in
polis, or cities) model to predict CO population exposures in Milan, Italy, based on subjects' time-
activity data broken into 15-min intervals. The simulation results showed that the U.S. 8-h NAAQS
level was exceeded in 1 case out of 1,000. The model also showed that exposures exceeded 20 ppm
in 1 case out of 100,000. The results were not shown to be very sensitive to the number of
microenvironments (e.g., outdoors, indoors, in-vehicle) included in the model.
The initial set of input data for population exposure models is ambient air quality data, which
may come from a monitoring network or model estimates. Estimates of concentrations in a set of
microenvironments are generated either by mass balance methods or microenvironmental factors.
Microenvironments modeled include indoor residences; other indoor locations, such as schools,
offices, and public buildings; and vehicles. The sequence of microenvironments and exertion levels
during the exposure period is determined from characteristics of each modeled individual. The
APEX model does this by generating a profile for each simulated individual by sampling from
distributions of demographic variables such as age, gender, and employment; physiological variables
such as height and weight; and situational variables such as living in a house with a gas stove or air
conditioning. Activity patterns from a database such as Consolidated Human Activity Database
(CHAD) are assigned to the simulated individual using age, gender, and biometric characteristics
(U.S. EPA, 2009, 194010). Breathing rates are calculated for each activity based on exertion level,
and the corresponding received dose is then computed. For APEX, the CO dosimetry algorithm
calculates venous COHb levels using the nonlinear CFK model, as described in Chapter 4.
(U.S. EPA, 2008, 191775). Summaries of individual- and population-level metrics are produced,
such as maximum exposure or dose, number of individuals exceeding a specified exposure/dose
threshold, and number of person-days at or above benchmark exposure levels. The models also
consider the nonambient contribution to total exposure. Nonambient source terms are added to the
infiltration of ambient pollutants to calculate the total concentration in the microenvironment. Output
from model runs with and without nonambient sources can be compared to estimate the ambient
contribution to total exposure and dose.
Recent larger-scale human activity databases, such as those developed for the CHAD or the
National Human Activity Pattern Survey (NHAPS), have been designed to characterize exposure
patterns among much larger population subsets than can be examined during individual panel studies
(Klepeis et al., 2001, 002437: McCurdy et al., 2000, 000782). CHAD consists of a consolidation of
human activity data obtained during several panel studies in which diary or retrospective activity
data were obtained, while NHAPS acquired sample population time-activity data through surveys
about human activity (Klepeis et al., 2001, 002437). The complex human activity patterns across the
population (all ages) are illustrated in Figure 3-44 (Klepeis et al., 2001, 002437). which is presented
to illustrate the diversity of daily activities among the entire population as well as the proportion of
time spent in each microenvironment. Different patterns would be anticipated when breaking down
January 2010 3-78
-------
activity patterns only for subgroups such as children or the elderly. Population exposures can be
estimated using CO concentration data in each microenvironment.
01
gp
GG
f= a fi P
ra n c3 TO
ooooooooooooooooooooooooo
ooooooooooooooooooooooooo
CN T—i 01 rn TT m vo r--~ oo ci o
c- - r-
o •— '
Figure 3-44.
Time of Day
Source: Reprinted with Permission of Nature from Klepeis et al. (2001, 002437).
Distribution of time that the sample population spends in various environments,
from the NHAPS.
Compartmental models, such as the Indoor Air Model (INDAIR), can be used to assess
exposure to infiltrated ambient air pollutants in a deterministic or probabilistic framework
(Dimitroulopoulou et al., 2001, 014737). To examine indoor concentrations of ambient CO,
Dimitroulopoulou et al. (2006, 090302) used the probabilistic formulation of the INDAIR model to
examine indoor exposure to ambient CO, along with NOX and PM for a given distribution of
background CO levels, meteorology, residential air exchange rate, and residential room dimensions.
They found that 24-h avg CO concentration increased from 1.86 ppm outdoors to 1.90-1.93 ppm
indoors in the absence of nonambient sources, and that indoor 24-h avg CO concentration could
increase to 1.93-2.00 ppm in the presence of smoking and to 1.98-2.32 ppm in the presence of gas
cooking. Similarity between the outdoor and nonsource indoor concentrations was attributed to the
lack of CO loss mechanisms. In the Reducing Urban Pollution Exposure from Road Transport
(RUPERT) study, Bell et al. (2004, 192376) presented methodology to use the probabilistic form of
INDAIR for development of personal exposure frequency distributions of CO, NOX, and PM, based
on time spent in residential, transportation, school, office, and recreational environments, with inputs
from transportation source categories (Chen et al., 2008, 193986).
3.6.3.2. Using Spatial Models to Estimate Exposure
Another set of approaches to improve exposure estimates in urban areas involves construction
of a concentration surface over a geographic area, with concentration between monitors estimated
using a model to compensate for missing data. The calculated CO concentration surface can then be
used to estimate exposures outside residences, schools, workplaces, roadways, or other locations of
January 2010
3-79
-------
interest. This technique does not estimate exposure directly because it does not account for activity
patterns or concentrations in different microenvironments. There are two main types of approaches:
spatial interpolation of measured concentrations, and regression models using land use, roadway
characteristics, and other variables to predict concentrations at receptors in the domain. Rigorous
first-principle models, such as dispersion models and CTMs, can also be used for this type of
application, but are less suitable because they have intensive resource requirements and are typically
applied over larger domains.
The STEMS model provides an example of an integrated-exposure modeling approach using a
range of spatial inputs. STEMS maps exposures based on inputs for traffic levels, atmospheric
dispersion, background concentrations, and geography. Gulliver and Briggs (2005, 191079) tested
the STEMS model for CO and observed some correlation between modeled and measured CO
concentrations (R2 = 0.41), which was consistent with results for PMi0 and NOX. Exposures were
estimated from the predicted ambient CO concentration using a term similar to a that varied
depending on whether the individual was walking or in a vehicle. Gulliver and Briggs (2005,
191079) noted that a limitation to modeling CO is the scarcity of background CO data obtained at
rural sites. For this reason, they assumed a constant value obtained from estimates made over the
North Atlantic Ocean. Although the authors only presented detailed results for a model of PMi0
based on traffic and meteorology in Northampton, U.K., they found that the majority of variation on
a given day in modeled exposure among school children was due to differences in travel routes.
Variation across days was also influenced by background and meteorological conditions. Similar
results can be expected for CO based on the tendency for variation of the CO concentration profile
on the neighborhood and microscales (Jerrett et al, 2005, 092864). Flachsbart (1999, 015857) tested
numerous meteorological, traffic, and background CO input variables in a regression approach to
predicting CO exposure among individuals while traveling in a vehicle. This work showed travel
time and average speed of on-road vehicles to be important determinants of CO exposure in a
vehicle. Results from individual models of this nature can be pooled to develop a distribution for
examination of population effects or for comparison with population exposure models.
Dispersion Models
Dispersion models have been used both for direct estimation of exposure and as inputs for
stochastic modeling systems, as described above. Location-based exposures have been predicted
using a model such as California Line Source Dispersion Model (CALINE), the American
Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD),
CALPUFF (long-range plume transport model created by the California Air Resources Board), or the
Operational Street Pollution Model (OSPM) for estimation of street-level ambient CO exposure
(e.g., Abdul-Wahab, 2004, 194011; Delfmo et al., 2009, 190254; Zhou and Levy, 2008, 190091).
CALINE, CALPUFF, and AERMOD utilize Gaussian dispersion models to describe pollutant
transport, while OSPM is a semi-empirical model of airflow and pollutant transport within an infinite
street canyon. Delfino et al. (2009, 190254) used CALINE (version 4) to model exposure in the near-
road environment for estimation of relative risks of asthma hospitalizations as a function of increases
in ambient CO and NOX concentrations. The concentration at each subject's home was computed
with the dispersion model, and then the data were aggregated to estimate a population risk. Zhou and
Levy (2008, 190091) used results from an OSPM simulation to compute intake fraction, defined as
the fraction of emissions that are inhaled or ingested, for ambient CO and other copollutants.
Daytime activity patterns were modeled using both CHAD and the American Community Survey
(http://www.census. gov/acs/www) to model commuting behaviors that would affect both mobile-
source emissions and population-based exposures. With an individualized exposure approach, the
model is deterministic. However, population exposures were estimated by performing repeated
simulations using various housing characteristics and then computing the probability distribution
function for exposure. When comparing street-canyon exposure computed by OSPM with near-road
exposure computed simply with a Gaussian dispersion model, Zhou and Levy (2008, 190091)
estimated that the street-canyon exposures would be three times greater than those in the general
community. Isakov et al. (2009, 191192) developed a methodology to link a chemical transport
model, used to compute regional scale spatiotemporally-varying concentration in an urban area, with
stochastic population-exposure models to predict annual and seasonal variation in population
exposure within urban microenvironments. Although this approach was demonstrated for PM2.5, it is
January 2010 3-80
-------
similar to the one used by Zhou and Levy (2008, 190091) for linking ambient CO concentrations
with population activity pattern data to associate the spatial-concentration field with personal
exposure to ambient CO.
Land Use Regression Models
Land use regression (LUR) models have also been developed to estimate pollution levels as a
function of several land use factors, such as land use designation, traffic counts, home heating usage,
point source strength, and population density (Briggs et al, 1997, 025950; Gilliland et al, 2005,
098820; Ryan and LeMasters, 2007, 156063). LUR is a regression derived from monitored
concentration values as a function of data from a combination of the land use factors. The regression
is then used for predicting concentration at multiple locations based on the independent variables at
those particular locations without monitors. At the census-tract level, a LUR is a multivariate
description of pollution as a function of traffic, land use, and topographic variables (Briggs et al.,
1997, 025950). Like deterministic-dispersion models, LUR can be performed over wide areas to
develop a probability density function of exposure at the urban scale. However, Hoek et al. (2008,
195851) warn of several limitations of LUR, including distinguishing real associations between
pollutants and covariates from those of correlated copollutants, limitations in spatial resolution from
monitor data, applicability of the LUR model under changing temporal conditions, and introduction
of confounding factors when LUR is used in epidemiologic studies.
A GIS platform is typically used to organize the independent variable data and map the results
from an LUR simulation. The GIS software creates numerous lattice points for the regression of
concentration as a function of the covariates. For instance, Krewski et al. (2009, 191193) computed
PM2.5 concentrations for the New York City and Los Angeles metropolitan areas. For the Los
Angeles analysis, the LUR was estimated at 23 monitors and then applied to simulate PM2.5
concentration at 18,000 points in the simulation domain, and an inverse distance weighting (IDW)
kriging method was applied to interpolate the predicted concentration. In New York City, the LUR
was estimated at 49 monitors for a 3-yr model and at 36 monitors for a model of winter 2000 and
then applied to simulate PM2 5 concentration at 5,600 locations in the 28-county domain; kriging was
employed only for the purpose of visualizing the concentration between monitors. The models
explained 69% and 66% of the variation in PM2.5 in Los Angeles and New York City, respectively.
GIS-based spatial-smoothing models can be used to estimate concentrations where monitors
are not located. Yanosky et al. (2008, 099467) described an approach to estimate PM concentrations,
using a combination of reported AQS data and GIS-based and meteorological covariates. Temporally
stationary covariates included distance to nearest road for different PM size fractions, urban land use,
population density, point-source emissions within 1 and 10 km buffers, and elevation above sea
level. Time-varying covariates included area-source emissions, precipitation, and wind speed. In this
analysis, the GIS-based covariates were temporally stationary, while the meteorological and PM
monitored concentration inputs were time varying. This approach was applied to estimate PM2 5>
PMio_2.5, and PMi0 exposures for the Nurse's Health Study and provided estimates of concentration at
approximately 70,000 nodes with PM2 5 and/or PMi0 data input from more than 900 AQS sites, with
good validation of the PM25 and PM10 models (Paciorek et al., 2009, 190090; Yanosky et al., 2008,
099467; Yanosky et al., 2009, 190114).
Marshall et al. (2008, 193983) compared four spatial interpolation techniques for estimation of
CO concentrations in Vancouver, BC. The investigators assigned a daily average CO concentration
to each of the 51,560 postal-code centroids using one of the following techniques: (1) the
concentration from the nearest monitor within 10 km; (2) the average of all monitors within 10 km;
(3) the IDW average of all monitors in the area; and (4) the IDW average of the three closest
monitors within 50 km. Method 1 (the nearest-monitor approach) and Method 4 (IDW-50 km) had
similar mean and median estimated annual average concentrations, although the 10th-90th percentile
range was smaller for IDW-50. This is consistent with the averaging of extreme values inherent in
IDW methods. The Pearson correlation coefficient between the two methods was 0.88. Methods 2
and 3 were considered sub-optimal and were excluded from further analysis. In the case of Method
2, a single downtown high-concentration monitor skewed the results in the vicinity, partially as a
result of the asymmetric layout of the coastal city of Vancouver. Method 3 was too spatially
homogenous because it assigned most locations a concentration near the regional average, except for
locations immediately adjacent to a monitoring site. LUR results were also reported in this study for
January 2010 3-81
-------
NO and NO2 and indicated that LUR's higher spatial precision reflects neighborhood-scale effects
from nearby land use but may not account for urban-scale variation. Brauer et al. (2008, 156292)
also evaluated LUR and IDW-based spatial-interpolation models and suggested that LUR is
appropriate for directly-emitted pollutants with high spatial variability; Brauer et al. (2008, 156292)
used NO and BC as examples, but CO emitted from mobile sources would also fall in that category.
These results highlight the variation in local concentration estimates with choice of estimation
technique.
Originally, LUR was used to model NO2 dispersion. It has been adapted for modeling
exposures to other pollutants, although application of LUR to CO exposures has been performed in
only a few studies. Findings related to other pollutants are provided because they are instructive in
how LUR can be used to predict CO concentrations. Carslaw et al. (2007, 148210) used multiple-
regression modeling to examine the effects of traffic volumes, wind components, temperature, and
time on concentrations of CO, NOX, NO2 (O3 was also a predictive variable for NO2), benzene, and
butadiene at a single site. These results were used for forecasting concentrations at that site, but the
study lacked the spatial resolution to predict concentrations at alternate sites. Cassidy et al. (2007,
199975) applied LUR to analyze the effect of wind, temperature, traffic volume, roadway size,
number of stories of surrounding buildings, other sources of pollution, terrain, and time of day on
concentrations of CO, PM25, and PMi0 at 30 street-level sites within Baguio City, Philippines. In this
work, they found traffic volume was the only significant predictor of CO during rush-hour periods,
while winds significantly predicted early morning concentrations of CO, PM2 5, and PMi0. The
model was not used for spatial interpolation in this case. Brauer et al. (2003, 155702) used LUR to
analyze PM2 5 exposure at 40-42 sites each within Stockholm, Sweden, Munich, Germany, and
throughout The Netherlands. This study found a measure of traffic density to be the most significant
variable predicting PM2 5 exposure and used GIS to apply the model at home addresses of asthmatic
subjects to estimate exposures. Ryan et al. (2008, 156064) reported on an LUR model developed
from monitor and land-use data and then applied at the locations of children to assess their exposure
to traffic-derived EC for the Cincinnati Allergy and Air Pollution Study. Ryan et al. (2008, 156064)
found traffic to be the most important determinant of diesel exhaust particle exposure. In this case,
wind direction was also factored into the model as a determinant of EC mixing. Although these
studies differed in the number of sites and in the pollutants of focus, they are instructive in
considering how LUR can be employed for estimating CO exposure.
3.6.4. Personal Exposure Monitors for CO
Portable monitors for measuring personal CO exposure include the Langan and Draeger
monitors, both of which use electrochemical oxidation-reduction techniques (Langan, 1992,
046120). These monitors continuously log CO concentrations, making them suitable for use in
personal monitoring studies. Electrochemical CO sensors typically have an LOD of 1 ppm and a
90% sensor response time (or the time required for the sensor to register 90% of a step change in CO
concentration) of 20-60 s. The 2000 CO AQCD (U.S. EPA, 2000, 000907) provided detail on design
updates of electrochemical CO sensors made during the 1990s. Commercially available personal CO
exposure monitors are not designed to detect concentrations below 1 ppm. Electrochemical personal
CO monitors are also typically sensitive to temperature changes, so that data correction is normally
required.
3.6.5. Indoor Exposure to CO
3.6.5.1. Infiltration of Ambient CO
CO is a relatively inert gas, making the indoor decay rate negligible compared to typical air
exchange rates (~l/h). In the absence of indoor sources, this would lead to an indoor-outdoor
concentration ratio (I/O) of approximately 1. For this reason, few studies have calculated I/O for CO.
Polidori et al. (2007, 156877) calculated I/O of 0.94-1.21 for two retirement communities in the Los
Angeles area. The authors suggested that similarity between I/O for CO and NOX can be attributed to
lack of indoor sources of either gas. Chaloulakou and Mavroidis (2002, 026050) reported I/O using
CO measurements in the absence of indoor sources in a school building in Athens, Greece, and
January 2010 3-82
-------
found that I/O varied with season. During the summer, median I/O was reported to be 0.57 on
weekdays, 0.91 on Saturdays, and 0.81 on Sundays. In winter, median I/O was reported to be 0.82
during weekdays, 0.90 on Saturdays, and 0.74 on Sundays. In a related study, Chaloulakou et al.
(2003, 190945) reported the median I/O over all days as 0.8 for the same school and 0.9 for an
Athens office building with no ETS (the presence of other sources was not clearly stated but
assumed zero). However, observed indoor values are often greater than outdoor concentrations in the
presence of indoor sources. A recent study in the U.K. reported I/O of 3.9-4.3 in homes with gas
cookers (Dimitroulopoulou et al., 2006, 090302). which is consistent with previous studies. A
multipollutant study conducted in 2000-2001 attempted to measure I/O for CO and calculated
residential infiltration factors, but low CO concentrations resulted in a large number of
measurements below the LOD (Williams et al., 2003, 053335). Ni Riain et al. (2003, 053792)
examined the effects of mechanical ventilation and wind speed on I/O. In this study, the authors
measured indoor and outdoor concentrations at two buildings located on a six-lane highway in
central London with natural and mechanical ventilation. Ni Riain et al. (2003, 053792) found that
outdoor concentrations for each building and ventilation condition ranged from 1.5 ± 0.1 ppm to 1.9
±0.1 ppm. Ni Riain et al. (2003, 053792) reported cumulative I/O approaching 0.9 within 30 min of
sampling for the mechanical ventilation case and cumulative I/O varying between 0.65 and 0.8 for
more than 70 h of sampling for the natural ventilation case. Ni Riain et al. (2003, 053792) found that
wind speed and direction influenced the variation in I/O.
Indoor air flow may affect CO exposure in the absence of indoor sources. Milner et al. (2006,
123100) compared hourly CO concentration time series from different parts of a building (with a
mix of natural and mechanical ventilation) located near a busy road and intersection in central
London, U.K. They found that, within a given floor, CO concentration is greater in rooms that are
closer to busy roads or an intersection. They noted that the correlation coefficient between indoor
and outdoor CO concentrations also decreased within the building with distance from the road; the
correlation coefficients were reported to be 0.80 for two time series obtained in rooms near the road,
while they were reported to range between 0.46 and 0.55 on the sides of the building furthest from
the road. The magnitude of the difference between CO concentrations in different rooms located
nearer or further from the roads also depended on wind direction. Milner et al. (2006, 123100) noted
that I/O tended to decrease with increasing wind speed, but Chaloulakou et al. (2003, 190945) also
noted that indoor CO concentration varied inversely with wind speed. Chaloulakou et al. (2003,
190945) attributed their observation to reduced concentrations related to dilution effects. Milner
et al. (2006, 123100) stated that this relationship could be due to dilution of CO or to the tendency of
people to keep windows closed on windy days. Additionally, CO concentrations were higher on
lower floors of the building and varied over a given day throughout the building. These findings
suggest that differences in exposure can occur within the same building as a result of differences in
air exchange related to access to windows, mechanical ventilation, and outdoor meteorological
conditions.
3.6.5.2. Exposure to Nonambient CO
Several papers have investigated the microenvironmental sources of total personal CO
exposure. The CDC conducted a survey of ED visits for nonfatal CO poisoning, CO exposure, or
potential CO exposure, and found that home heating was the largest known source of CO exposure,
prompting 16.4% of CO-related ED visits, followed by motor vehicle exhaust exposure accounting
for 8.1% of ED visits (Annest et al., 2008, 190236). Aim et al. (2000, 192374: 2001, 020237) studied
factors that contributed to elevated CO exposures among preschool children and found that presence
of a gas stove at home, ETS, natural ventilation, and living in a high-rise building all contributed to
increased CO exposures. Time-activity diaries were linked to personal CO exposures in the
EXPOLIS study. Here, Georgoulis et al. (2002, 025563) observed that geometric mean exposure
among smokers ranged from 0.33 ppm in Helsinki, Finland, to 3.2 ppm in Athens, Greece, while
among nonsmokers it ranged from 0.36 ppm in Helsinki to 1.7 ppm in Milan and ambient CO
concentration ranged from 0.42 ppm in Helsinki to 3.2 ppm in Athens. Bruinen de Bruin (2004,
190943) found, for a panel of 46 subjects in Milan, that indoor CO concentrations were 3.4 ppm in
the presence of gas cooking and ETS, compared with 2.9 ppm in the presence of ETS only, 2.4 ppm
in the presence of gas cooking only, and 1.8 ppm in the absence of indoor CO sources. Scotto di
Marco et al. (2005, 144054) reported that average indoor CO increased in the presence of ETS from
0.96-1.2 ppm for the home indoor environment and from 1.0-1.4 ppm for the work indoor
January 2010 3-83
-------
environment. CO concentrations were measured to decrease from 1.5 to 1.3 ppm in other (not home
or work) indoor environments, but those locations included garages, restaurants, and bars and could
have been differently influenced by CO from cooking, indoor automobiles, or other sources.
Personal CO concentrations can also be much more variable than ambient measurements.
Figure 3-45 shows hourly versus personal CO concentration data obtained by Chang et al. (2000,
001276) for a 1998-1999 multipollutant sampling campaign in Baltimore, MD. Personal exposures
were obtained in five separate microenvironments in this study. A high degree of scatter is evident in
this figure, which suggests that these personal exposures are influenced by both ambient and
nonambient sources of CO. Figure 3-46 is a box plot of the personal-to-ambient CO concentration
ratio for the same five microenvironments. Wide variability is seen in these plots, particularly during
the summer. Much of that variability could be due to the influence of nonambient sources, which
would then result in poor correlation between total personal exposure and ambient concentration.
u
_
O
X
3 -
O Indoor Residence
A Indoor Others
O Outdoor near Roadway
V Outdoor away tH>m Road
+ In Vehicle
Hourly Ambient CO (ppm)
Source: Reprinted with Permission of the Air and Waste Management Association from Chang et al. (2000, 0012761
Figure 3-45. 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, and in vehicle.
January 2010
3-84
-------
swswswswsw
Indoor Indoor Outdoor Outdoor In Vehicle
Residence Other Near away from
Road Road
Source: Adapted with Permission of the Air and Waste Management Association from Chang et al. (2000, 0012761
Figure 3-46. 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, and in
vehicle. The grey line shows the mean, and the black mid-line shows the median.
S = summer; W = winter.
Vehicle self-pollution, defined by Behrentz et al. (2004, 155682) as the fraction of a vehicle's
own exhaust entering the vehicle microenvironment, is another potential nonambient source of CO
exposure. This has been studied using inert tracer gases to evaluate exposures of children riding
school buses. Behrentz et al. (2004, 155682) used sulfur hexafluoride (SF6) tracer gas emitted from
school bus engines to determine the proportion of in-vehicle pollution related to self-pollution.
Based on the SF6 concentration, they calculated that 0.04-0.29% of the bus cabin air contained
exhaust for high-emitting diesel engines, 0.01-0.03% for "regular" diesel buses, 0.02-0.04% for
buses fitted with a particle trap, and 0.03-0.04% for buses running on compressed natural gas. SF6
concentrations were higher when bus windows were closed.
3.6.6. Exposure Assessment Studies at Different Spatial Scales
3.6.6.1. Neighborhood to Urban Scale Studies of Ambient CO Exposure
Although several multipollutant exposure studies have been conducted recently in the U.S.,
(e.g., Sarnat et al., 2006, 089784). most have not included CO in the suite of pollutants, possibly due
to high detection limits in personal monitors. A few studies conducted in Europe and Canada
measured personal-ambient relationships for CO. This section summarizes CO exposure assessment
studies that compare personal exposure measurements with ambient concentration measurements for
the purpose of examining how well these measures correspond.
The EXPOLIS study (Georgoulis et al., 2002, 025563) found that 48-h personal exposures
were significantly correlated with ambient concentrations in each of five European cities (Athens,
Basel, Helsinki, Milan, and Prague). Controlling for source terms, including ETS, traffic, and natural
gas appliances, regression coefficients between personal exposure and ambient concentration ranged
from 0.28 in Milan to 1.99 in Helsinki and were all statistically significant (p < 0.01 for all cities
except Prague, where p = 0.05). The regression coefficient for Helsinki (>1) likely reflects
nonambient sources that were not controlled in the study. The ambient concentration was the only
variable that was statistically significantly associated with 48-h personal exposure for all five cities
January 2010
3-85
-------
in this study, with correlations between personal CO exposure and ambient CO concentration
ranging from 0.33 to 0.77. Georgoulis et al. (2002, 025563) reported that CO exposure in traffic
ranged from 0.99 ppm in Helsinki to 4.2 ppm in Athens, while ambient CO concentration ranged
from 0.42 ppm in Helsinki to 3.2 ppm in Athens. As part of this study, personal CO exposure was
measured for a panel of 50 office workers in Milan (Bruinen de Bruin et al., 2004, 190943). Average
measured 1-h personal exposures were 7.3 ppm in comparison with 5.0 ppm for fixed site 1-h
measurements. Average 8-h (3.3 ppm) and 24-h (2.1 ppm) CO concentrations were the same for
personal and fixed-site measurements. Percentage of time exposed, exposures, and percentage of
exposure from the Bruinen de Bruin et al. (2004, 190943) study, in the absence of nonambient CO
from ETS and gas cooking, are shown in Table 3-13. The largest percentage of time-weighted CO
exposure was attributed to home indoor exposure in the absence of indoor sources, while the highest
exposure levels were observed during transit; Scotto di Marco et al. (2005, 144054) found similar
results. Bruinen de Bruin et al. (2004, 190943) and Scotto di Marco et al. (2005, 144054) found that
mobile source emissions were important contributors to personal exposure, as described in Section
3.6.6.2.
Table 3-13. Percentage of time exposed to ambient CO (adjusted to reflect the absence of
nonambient CO from ETS and gas cooking), average CO exposures, and percentage of
exposure estimated for the population.
Percent of time exposed
(%)
INDOORS
Home
Work
Other
OUTDOORS
Home
Work
Other
IN-TRANSIT
Walking
Train/metro
Bus/tram
Motorbike
Car/taxi
89.6
56.5
29.1
4.1
1.8
0.2
0.6
1.0
8.5
3.0
0.7
2.0
0.2
2.6
Exposure (ppm) Percent of exposure (%)
1.8
1.9
2.5
2.3
2.1
2.6
3.0
3.0
3.8
4.5
5.7
81.1
49.4
26.8
4.9
2.1
0.2
0.6
1.2
16.8
4.4
1.0
3.7
0.4
7.2
Source: Reprinted with Permission of Nature from Bruinen de Bruin et al. (2004,190943)
EXPOLIS also looked at the special case of children's exposure to CO because children
generally do not produce CO in their daily activities and have no occupational exposures. Aim et al.
(2000, 192374; 2001, 020237) reported higher personal exposures than ambient concentrations for
children aged 3-6 yr in Helsinki. Their mean 1-h daily max exposure was 5.2 ppm, compared to
1.4 ppm measured at a fixed-site monitor. For the average of 8-h and 24-h daily max concentrations,
the corresponding values were 2.9 ppm and 2.1 ppm for personal exposure and 0.8 and 0.6 ppm,
respectively, for fixed site measurements. The Spearman rank correlation, although statistically
significant, was relatively low (r = 0.15) between individual 24-h avg exposure and the ambient
monitor. The correlation improved when the average exposure of children measured on the same day
(r = 0.33, 3-6 children) or the same week (r = 0.55, 10-23 children) was compared to the monitor
data. A regression model using questionnaire data found that parental smoking status, parental
education, and presence of a gas stove explained only 12% of the variability in the 8-h max
January 2010 3-86
-------
exposures, indicating that other factors, such as time spent outdoors and proximity to roadways, are
likely to be important in determining personal exposure.
Kim et al. (2006, 089820) reported mean CO concentrations of 1.4 ppm for a panel of 28
cardiac-compromised individuals in Toronto, Canada. Corresponding fixed-site monitor mean
concentrations ranged from 0.5 to 1.4 ppm, with an overall mean of 1.0 ppm. The observed higher
personal exposures may have been due to both indoor sources and proximity to roadways when
outdoors. Personal-ambient Spearman correlations ranged from -0.65 to 0.93, with a median of
r = 0.31, indicating that while moderate correlations are observed overall, inter-individual
differences based on time spent in different microenvironments have a strong influence on the
observed correlation. Lai et al. (2004, 056811) measured relationships between personal CO
exposure and microenvironmental (home indoor, home outdoor, and work indoor) concentrations in
Oxford, U.K. The highest personal exposures were associated with smoking, cooking, and
transportation, while low correlations were observed between personal and indoor residential
concentrations, further indicating the importance of indoor sources and the need to separate ambient
contributions to personal exposure from total personal exposure.
The studies presented above present mixed results regarding the association between ambient
CO concentration measurements and personal CO exposures. Some personal CO measurements have
been reported to be higher than ambient concentrations, while others are similar. Additionally,
correlation between ambient CO concentration and personal exposure has varied in the literature.
Nonambient (described in Section 3.6.5) and in-transit sources (described in Section 3.6.6.2) have
been identified as important contributors to personal exposure. These observations raise questions
about where and when ambient CO concentration can be used as a surrogate for personal CO
exposure; these concepts are explored further in Section 3.6.8.
3.6.6.2. Microscale Studies of Ambient CO Exposure: Near-Road and On-Road
Exposures
The 2007 American Housing Survey (AHS) (U.S. Census Bureau, 2008, 194013) reports that
17.9 million occupied homes nationwide (16.1%) are within ~90 m (300 ft) of a "4-or-more-lane
highway, railroad, or airport" and so are exposed to the near-road environment. Within city centers,
6.2 million occupied homes (19.7% of those living in city centers) are within approximately 90 m of
a highway, railroad, or airport; whereas in rural areas outside designated Metropolitan Statistical
Areas (MSA), 1.4 million occupied homes (9.2% of those in rural areas outside MS As) are near a
highway, railroad, or airport. Those data can be put into context for exposure assessment in the near-
road environment; Section 3.5.1.3 describes near-road studies in which ambient CO was measured
within the vicinity of a road and microscale AQS data obtained in the near-road environment. The
AQS data suggest some spatial variability (20-40% difference between microscale and middle scale
monitors, with the hourly microscale concentration having a median of 0.5 ppm and a 99th percentile
value of 2.2 ppm), which was much lower than that reported by Zhu et al. (2002, 041553) for the
near-road environment, in which the average concentration at 17 m from the road was 2.3 ppm
(range 1.9-2.6 ppm) and a factor of about 12.5 lower for the monitoring site located 300 m from the
road. The larger discrepancy observed between the Zhu et al. (2002, 041553) data and the AQS data
might be attributed to the fact that the sampling equipment used by Zhu et al. (2002, 041553) were
downwind of the freeway for the entire sampling period, while the hourly AQS data represents a
range of wind speeds and directions that vary across different monitoring sites. For those living in
the 16.1% of occupied homes situated in the near-road environment (within ~90 m), median hourly
CO concentrations are typically higher than those further from the road, but the magnitude of the
outdoor concentration is still in most circumstances measured to be below 2.2 ppm.
Kaur and Nieuwenhuij sen (2009, 194014) and Carslaw et al. (2007, 148210) suggest that CO
exposures are related to traffic volume and fleet mix in the street-canyon environment. In this
research, Kaur and Nieuwenhuij sen (2009, 194014) developed a multiple linear regression of CO as
a function of mode of traffic, broken down by vehicle type, wind speed, temperature, and traffic
count for data obtained in central London as part of the DAPPLE study of traffic-related pollution.
They added each variable successively and found traffic count, temperature, wind speed, and
walking to be significant parameters in the model, with traffic count being the strongest determinant.
Analysis of variance showed variability in traffic count to explain 78% of the variability in CO levels
for these data, and variability in mode of transport explained 6% of the variability. Likewise,
Carslaw et al. (2007, 148210) used a generalized additive model to determine how CO concentration
January 2010 3-87
-------
(log-transformed) varies as a function of year, the along-street and cross-street components of wind,
temperature, Julian day, light and heavy traffic counts, and temperature for data obtained in central
London. Light-duty vehicle count was a more important determinant of CO concentration than was
heavy-duty (i.e., diesel) vehicle count in this study, which is not surprising because gasoline-
powered vehicles are known to emit more CO than diesel engines. They found that the CO
concentration declined steadily with year and that wind was the most significant covariate. The
decline in CO concentration with year, adjusted for all other covariates, was usually significantly
different than the simple relationship between concentration and year, but the adjusted and
unadjusted trends were similar. In addition to showing meteorology to be an important determinant
of concentration, these modeling exercises also suggest a linear or log-linear relationship between
concentration and traffic count.
Findings regarding meteorology are consistent with in-vehicle CO concentration studies.
Gomez-Perales et al. (2007, 138816) also noted that meteorology can impact in-vehicle exposures,
with evening increases in wind speed causing a 50% reduction in CO exposures among bus and
minibus commuters. Aim et al. (1999, 047196) made a similar observation in a study of urban
commuters' exposure within a vehicle. These observations are sensible given the influence of
meteorology on near-road concentrations shown by Baldauf et al. (2008, 190239) and Gokhale and
Khare et al (2007, 194015).
A number of studies have focused on transit-time CO exposure, which can occur while in a
vehicle or cycling (on-road) or while walking (near-road). Chang et al. (2000, 001276) showed that
personal exposures in vehicles were on average 2.8 times higher than ambient concentrations during
the summer and 4.1 times higher than ambient concentrations in the winter (Figure 3-46). For the
other four microenvironments tested, the average ratio of personal exposure to ambient concentration
was ~1. Kaur et al. (2005, 086504) found that transit time exposures in London, U.K., were
significantly higher than measurements made at a fixed-site background monitor away from traffic
(0.3 ±0.1 ppm) for car riders (1.3 ± 0.2 ppm), taxi riders (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). Curbside measurements (1.5 ± 0.7 ppm) in
this study were slightly higher than car riders' exposures. Duci et al. (2003, 044199) found that
average in-transit exposures in Athens, Greece, were highest for cars (winter: 21.4 ± 4 ppm),
followed by pedestrians (winter: 11.5 ± 2.6 ppm; summer: 10.1 ± 1.7 ppm), buses (winter:
10.4 ± 2.9 ppm; summer: 9.4 ± 3.6 ppm), trolleys (winter: 9.6 ±1.9 ppm; summer: 8.2 ± 3 ppm), and
rail transit (winter: 4 ± 0.6 ppm; summer: 3.4 ± 0.7 ppm). Duci et al. (2003, 044199) did not provide
fixed-site CO concentrations but stated that in-transit exposures were higher in each case. Gomez-
Perales et al. (2004, 054418) measured CO exposures on buses, minibuses, and metro cars in Mexico
City, Mexico, to be 12 ppm, 15 ppm, and 7 ppm, respectively. These values are much higher than
CONUS measurements and those presented by Kaur et al. (2005, 086504). but the relative difference
between the minibus and bus exposures in the Gomez-Perales et al. study are similar to those seen
for the taxi-to-bus or car-to-bus comparisons in Kaur et al. (2005, 086504). These studies indicate
that on-road exposures might be influenced by vehicle type, but that city-to-city differences are
likely larger than differences between different modes of transport.
Additional analyses from the EXPOLIS study indicated that on-road mobile source emissions
were the most important source of CO exposure for non-ETS-exposed subjects (Bruinen de Bruin et
al., 2004, 190943: Scotto Di Marco et al., 2005, 144054). Scotto di Marco et al. (2005, 144054)
found that, for a panel of 201 adult Helsinki, Finland, residents (aged 25-55 yr), subjects spent 8.1%
(1.9 h) of their time in transit, which accounted for 12.6% of their total exposure (range of means =
0.96 ppm on a train - 2.8 ppm in a car). Similarly, in a panel study of 50 office workers, Bruinen de
Bruin et al. (2004, 190943) found that, in the absence of nonambient sources, the subjects spent
8.5% (2 h) of their time in transit, which accounted for 16.8% of their total exposure, with 2.6% of
time spent in a car or taxi accounting for 7.2% of exposure (mean = 5.7 ppm). Commuting time was
an important predictor of exposure, such that subjects living in low CO concentration suburban areas
and commuting to work experienced higher exposures than urban residents with short commute
times. According to the 2007 AHS (U.S. Census Bureau, 2008, 194013). 110.1 million U.S. workers
(87.8% of those working) commute to work in automobiles. 32.8% of U.S. workers work at home or
commute less than 15 min to work, 32.1% commute 15-29 min to work, 15.1% commute 30-44 min
to work, 5.7% commute 45-59 min to work, and 5.0% commute 1 h or longer to work.
Vehicle ventilation can be an important determinant of in-vehicle concentrations. A study from
Abi Esber and El Fadel (2008, 190939) in Beirut, Lebanon, is presented because the authors
observed in-vehicle CO concentration time series under a range of ventilation conditions, although
January 2010 3-88
-------
the in-vehicle CO concentrations measured are substantially higher than those typically observed in
the U.S. Abi Esber and El Fadel (2008, 190939) reported results from CO concentration
measurements taken directly outside and within an automobile in Beirut, Lebanon, during the
morning commute period of 7:30-9:30 a.m under three different ventilation conditions. Figure 3-47
shows that the time series for the cabin and outdoor CO samples are very similar for the fresh air
scenario. However, for the recirculating air ventilation scenario, the in-vehicle concentration
increases and then reaches a plateau at a higher level. Abi Esber et al. (2007, 190941) stated that
unaccounted sources of CO cause the build-up of in-cabin CO concentrations when the ventilation is
set to recirculation mode. The correspondence between in-vehicle and outside-vehicle concentrations
for the fresh air ventilation experiments, and lack thereof for the recirculation mode ventilation
experiments, observed by Abi Esber and El Fadel (2008, 190939) suggests that in-vehicle
concentrations of ambient CO are affected by mode of ventilation.
75
50 -
ex
CL
o
u
0
Figure 3-47.
10 20 30 40
Time (inin)
50 0 10 20 30 40 50
Time (min)
Source: Reprinted with Permission of Elsevier Ltd. from Abi Esber and El Fadel (2008,190939)
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.
Substantial variability can occur over time within a vehicle. Riediker et al. (2003, 043761)
measured CO concentrations inside highway 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 be related to time of day. Likewise, Rodes et al. (1998, 010611) reported CO concentrations in
vehicles in Sacramento and Los Angeles under different driving conditions (arterial, freeway, high-
occupancy-vehicle freeway lane, and "maximum" conditions at rush-hour and nonrush-hour times).
They measured peak in-vehicle concentrations spanning 7-67 ppm on a freeway during rush hour,
although the mean for each scenario was <6 ppm. In comparison, the peak roadside concentration
ranged from 3 to 11 ppm and the peak ambient CO concentration was 1.3 ppm at the time of the
measurements. The Rodes et al. (1998, 010611) data agree with results from the Riediker et al.
January 2010
3-89
-------
(2003, 043761) and Aim et al. (1999, 047196) studies showing that substantial variability in CO
concentration inside the cabin of a vehicle can occur during the course of a commute.
In their review of roadway exposures to CO and PM, Kaur et al. (2007, 190070) listed a
number of factors that may influence near-road or on-road exposure. Vertical CO concentration
gradients have been documented in which concentrations decreased with height; lower breathing
zone height among children may make them more likely to be exposed to higher CO tailpipe
emissions. With respect to transportation, Kaur et al. (2007, 190070) suggested that vehicle
ventilation, speed, position in traffic, and start/stop activity influence in-vehicle exposures. Abi Esber
and El Fadel (2008, 190939) and Riediker et al. (2003, 043761) illustrated the effect of vehicle
ventilation on in-vehicle concentrations. The influence of vehicle speed and start/stop activity is
consistent with the turbulence research of Khare et al. (2005, 194016) and Gokhale and Khare (2007,
194015) that suggested an increase in traffic volume and vehicle movement acts to dilute the on-road
concentration of CO discussed in Section 3.5.1.3.
3.6.7. Association between Personal CO Exposure and Copollutants
Since incomplete combustion is the primary source of ambient CO in urban areas, exposure to
ambient CO is accompanied by exposure to other combustion-related pollutants, such as NOX, PM,
and VOCs. Thus, ambient CO is often considered a surrogate for exposure to traffic-generated
pollutants. However, the specific mix of CO with NOX and PM depends on the source; for example,
the mixture generated by gasoline engines differs from that produced by natural gas combustion.
Correlations between ambient CO and ambient PM2.5, PMi0, NO2, SO2, and O3 from AQS data and
the peer-reviewed literature were presented in Section 3.5.3. Nationwide, ambient CO was most
highly correlated with ambient NO2, followed by PM25 and PMi0. Correlations between CO and
PM2 5 were not consistently positive on a national basis; correlations spanned from negative to
positive for ambient CO with ambient SO2 and ambient PM10, and ambient CO was negatively
correlated with ambient O3. The correlation between ambient CO and specific ambient VOCs
depends on parameters such as ambient temperature and the volatility of a specific compound.
Relationships between personal CO exposures and copollutants were reported less frequently
in the literature, but results from these studies were consistent with the findings cited above. In a
study of personal exposures to CO, PM25, and ultrafine PM in a street canyon, Kaur et al. (2005,
086504) found low Pearson's correlation of total personal CO exposure with personal PM25
exposure (r = 0.23). Personal CO exposure had much better correlation with personal ultrafine
particle (UFP) exposure (r = 0.68). Chang et al. (2000, 001276) reported correlations of personal CO
exposure with personal PM2 5, personal toluene, and personal benzene exposures in Baltimore, MD,
at five locations, labeled indoor residential, indoor nonresidential, outdoors near roadway, outdoors
away from road, and in vehicle. Much variability was observed in the correlations for different
locations and seasons (winter versus summer). In general, the correlations of personal CO with
personal VOCs tended to be stronger in the winter. Chang et al. (2000, 001276) suggested that lower
wintertime indoor air exchange rates could increase exposure to nonambient CO and VOC sources,
such as ETS, and hence increase correlations between personal exposure of CO to VOCs. Significant
associations of CO with benzene and toluene were also observed in vehicle microenvironments.
3.6.8. Implications for Epidemiology
Exposure error can be an important contributor to variability in epidemiologic study results.
Community time-series studies may involve thousands or millions of people whose exposure and
health status is estimated over the course of a few years using a short monitoring interval (hours to
days). Community-averaged concentration is typically used as a surrogate for ambient exposure in
community time-series studies. Exposures and health effects are spatially aggregated over the time
intervals of interest because community time-series studies are designed to examine health effects
and their potential causes at the community level (e.g., Bell et al., 2009, 194033). A longitudinal
cohort epidemiology study typically involves hundreds or thousands of subjects followed over
several years or decades. Concentrations are generally aggregated over time and by community to
estimate exposures (e.g., Rosenlund et al., 2006, 089796). In addition, panel studies, which consist of
a relatively small sample (typically tens) of study participants followed over a period of days to
months, have been used to examine the health effects associated with exposure to ambient
January 2010 3-90
-------
concentrations of air pollutants. Panel studies include time-activity diary studies (Akland et al., 1985,
011618; Bruinen de Bruin et al., 2004, 190942; Scotto Di Marco et al., 2005, 144054). These studies
may apply a microenvironmental model to represent exposure to an air pollutant.
The importance of exposure misclassification varies with study design and is dependent on the
spatial and temporal aspects of the design. For example, the use of a community-averaged CO
concentration in a community time-series epidemiologic study may not allow for adequate
examination of the role of spatial variability. Other factors that could influence exposure estimates
include spatial and temporal variability related to source strength, topography of the natural and built
environment, and meteorology; measurement errors; use of ambient CO concentration as a surrogate
for ambient CO exposure; and the presence of CO in a mixture of combustion-related pollutants. The
following sections will consider various sources of error and how they affect the interpretation of
results from epidemiologic studies of different designs.
3.6.8.1. Measurement Error
Measurement Error at Community-Based Ambient Monitors and Exposure
Assessment
Because CO concentrations measured with community-based ambient monitors are often used
as surrogates for ambient CO exposure in epidemiology studies, the limitations of the
instrumentation are important to consider. As stated in Section 3.4.2, among the 291 monitors
meeting completeness criteria for 2005-2007, only 8 were monitors with LOD = 0.04 ppm; the other
monitors had LOD of 0.5 ppm. Among the nationwide AQS data for 2005-2007 from these 291
monitors, more than 50% of the hourly CO concentration data were below the LOD of the
instrumentation. Data below the LOD adds uncertainty to the association between CO exposure and
health effects estimates. Additionally, many of the monitors are not sited for a specific measurement
scale, and a given scale classification can represent a range of CO source conditions, as described in
Section 3.5.1.3. These factors also contribute uncertainty in interpretation of measurements.
Instrumental measurement error, other than that related to high LOD, is not expected to bias
health effect estimates substantially in most circumstances. Because there will be some random
component to instrumental measurement error, the correlation of the measured CO concentration
with the true CO concentration will likely be <1. When analyzing the effect of instrument error for
measuring nonreactive ambient pollutants, Zeger et al. (2000, 001949) stated that the instrument
error for ambient measurements "is close to the Berkson type." In the Berkson error model, the
measured-exposure estimate is used instead of the true exposure, based on the assumption that the
average measurement is the average of the true exposure. It is generally expected that the health
effects estimate will not be biased by using measured values with error but may have more
uncertainty than would an estimate based on the true-average exposure. In order for instrument error
to cause substantial bias in health effects estimates, the error term (the difference between the true
concentrations and the measured concentrations) must be strongly correlated with the measured
concentrations.
Measurement Error for Personal Exposure Monitors
Personal electrochemical CO monitors are subject to interference and drift and have a
relatively high LOD (~1 ppm) relative to current ambient concentrations. Previous studies in the
1980s and 1990s, when ambient levels were higher, involved successful deployment of these
monitors, but more recent exposure studies have avoided personal CO measurements because there
are now a high percentage of nondetects. The lack of a suitable personal monitor for measuring low-
level exposures (<1 ppm) has hampered field studies assessing personal exposure to ambient CO.
Chang et al. (2001, 019216) evaluated the Langan CO monitor as part of an air quality sampling
manifold. At relatively high (0.4-3.0 ppm) CO concentrations, the instrument correlated well (R =
0.93) with a reference NDIR CO monitor, with the Langan underestimating the CO concentration by
41%. When ambient levels fell consistently below that level, coefficient of determination (R2)
between the Langan and reference monitor fell to R2 = 0.40 in summer and R2 = 0.59 in winter, with
January 2010 3-91
-------
the arithmetic average concentration underestimated by 47% in summer and by 63% in winter.
Chang et al. (2001, 019216) pointed out the need for frequent instrument zeroing to minimize
instrument drift. Abi Esber and El Fadel (2007, 190940) evaluated a similar personal electrochemical
CO sensor, the GEM™ 2000, by comparing measured concentrations with those obtained through
co-located grab-bag sampling in a vehicle cabin. Differences between the GEM™ 2000 and the
reference samples were fairly low during weekday driving (differences = 2.1-10.6%). Differences on
Sundays, when traffic was significantly lower than during weekdays, were dependent on vehicle
ventilation conditions, with better agreement when vehicle ventilation allowed for higher cabin CO
concentrations (differences = 3.4-5.6%). But the electrochemical sensor did not compare well with
reference values when concentrations were low (differences = 20-71%). In general, it is difficult to
separate the large instrumental measurement error seen at concentrations below instrument LOD
from variation related to nonambient CO sources. This large variation in personal measurements can
result in high levels of classical measurement error (Sheppard et al., 2005, 079176).
3.6.8.2. Exposure Issues Related to Nonambient CO
The focus of the ISA is on ambient CO because that is relevant to the NAAQS. Uncertainty
related to nonambient CO exposure may make it difficult to distinguish the effect of ambient CO on
health effects. Wallace and Ziegenfus (1985, 011656) used NHANES II (1976-1980) data to evaluate
the relationship between COHb levels and ambient CO concentration in 20 U.S. cities. They found a
significant slope of 0.066% per 1 ppm increase of CO concentration. However, there was much
scatter in the data, and variability in ambient CO concentration only accounted for 3% of the
variation in COHb. The authors attributed this scatter to variability in nonambient sources such as
gas cooking and ETS. This finding illustrates the importance of considering the relative role of
ambient and nonambient CO in total personal exposure.
Ambient and nonambient CO are chemically identical and so exert the same health effects. At
the same time, ambient and nonambient sources are distinct and not correlated with each other
(Wilson and Suh, 1997, 077408) and so would not confound the association between ambient CO
exposure and the health effect (see also Sheppard et al., 2005, 079176). Zeger and Diggle (2001,
026017) noted that, because ambient and nonambient CO exposures are uncorrelated, in a health
effects model the regression coefficient of ambient concentration should be equal to the product of a
(the ratio of ambient exposure to ambient concentration) and the regression coefficient obtained
when average personal exposure is used. The confidence intervals around the estimate obtained
using total personal exposure would be wider because nonambient CO concentrations add variability.
This is true even for the case when the chemical compound is the same for the ambient and
nonambient pollutants, as in the case of CO. Likewise, Sheppard et al. (2005, 079176) simulated
ambient and nonambient exposures to a nonreactive pollutant and observed that nonambient
exposure has no effect on the association between ambient exposure and health outcomes for the
case where ambient and nonambient concentrations were independent. Hence, the bias that will be
introduced to epidemiologic models by using ambient CO concentration instead of personal
exposure to ambient CO is given by the average a. Random variations in daily values of a would not
change the health effects estimate but would also widen the confidence intervals around the health
effect estimate.
3.6.8.3. Spatial Variability
CO concentration is known to be spatially heterogeneous, as evidenced by the near-road and
in-vehicle studies cited in Sections 3.5.1.3 and 3.6.6.2, as well as the intraurban correlations
provided in Section 3.5.1.2 and Tables A-9 through A-16 of Annex A. Results from Zhu et al. (2002,
041553). which showed a large CO concentration gradient in the near-road environment, support the
contention that CO exposures for those living in the near-road environment but far from a monitor
might be underestimated. Conversely, exposures for those living away from roads might be
overestimated by near-road CO concentration measurements. Exposure error may occur if the
ambient CO concentration measured at the central site monitor is used as an ambient exposure
surrogate and differs from the actual ambient CO concentration outside a subject's residence and/or
worksite (in the absence of indoor CO sources). Averaging data from a large number of samplers will
dampen intersampler variability, and use of multiple monitors over smaller land areas may allow for
January 2010 3-92
-------
more variability to be incorporated into an epidemiologic analysis. This is consistent with
conclusions presented in the 2000 AQCD (U.S. EPA, 2000, 000907).
Community exposure may not be well represented when monitors cover large areas with
several subcommunities having different sources and topographies. The intersampler correlations of
AQS data from monitors, presented in Section 3.5.1.2, reflect how well the time series of
concentration data correspond across metropolitan areas. Overall, the data show moderate site-to-site
correlation; for example, in the Los Angeles CSAthe mean of the correlation was 0.50, and within
one standard deviation of the mean, the range of correlations was 0.36-0.65. Bell et al. (2009,
194033) tested the association between monitor density and 1-h max CO effect estimates for CVD
hospitalizations for 126 U.S. counties and found an 8% increase in effect estimate size (95% CI: -7%
to - 24%) with an IQR decrease in area covered by the monitor. This difference was not statistically
significant but suggested that the magnitude of the effect estimate might be related to monitor
coverage. Sarnat et al. (2009, 180084) studied the spatial variability of CO, along with NO2, O3, and
PM2.5, in the Atlanta, GA, metropolitan area and how spatial variability affects interpretation of
epidemiologic results, using time-series data for circulatory disease ED visits. Sensitivity to spatial
variability was examined at slightly greater than neighborhood scale (8 km) in this study.
Interestingly, Sarnat et al. (2009, 180084) found that relative risk varied with distance between the
monitor and study population when comparing urban to rural locations, but distance of the study
population to the monitor was not an important factor when comparing urban population groups.
This suggests that, even for spatially heterogeneous CO, urban scale measures may produce results
comparable to neighborhood-scale exposures in some circumstances. This may be due to
comparability of sites throughout a city, for example, as a result of similar traffic patterns. However,
Sarnat et al. (2009, 180084) caution that, because their study was limited to 8 km radii, it is not
possible to interpret this work with respect to near-road and on-road microscale exposures.
3.6.8.4. Temporal Variability
Temporal Correlation
Within a city, lack of correlation of relevant time series at various sites results in smoothing
the exposure/surrogate concentration function over time and resulting loss of peak structure from the
data series. At the same time if monitors are well correlated across a metropolitan area, even if the
magnitude of concentration varies over space, time series analyses should provide comparable
results across larger spatial areas. Such temporal correlation resulted in the small variation in relative
risk estimates within the metropolitan region in Sarnat et al. (2009, 180084). where peak rush-hour
times were similar throughout the city, in comparison with the rural area where temporal driving
patterns were different. Burnett and Goldberg (2003, 042798) found that community time-series
epidemiologic study results reflect actual population dynamics only when five conditions are met:
environmental covariates are fixed spatially but vary temporally; the probability of the health effect
estimate is small at any given time; each member of the population has the same probability of the
health effect estimate at any given time after adjusting for risk factors; each member of the
population is equally affected by environmental covariates; and, if risk factors are averaged across
members of the population, they will exhibit smooth temporal variation. Note that for this study,
Burnett and Goldberg (2003, 042798) analyzed mortality related to PM exposure, but the results are
not specific to a given pollutant or health effect and thus are generalized here for time-series
analysis. Dominici et al. (2000, 005828) noted that ensuring correlation between ambient and
community average exposure time-series air pollutant data is made difficult by limitations in
availability and duration of detailed ambient concentration and exposure time-series data, resulting
in a source of uncertainty. If sufficient data are available and the time series of concentration data
adequately represent population dynamics, then high temporal correlation between sampling sites
should limit bias in health effects estimates, even if the magnitude of the concentrations differ.
January 2010 3-93
-------
Seasonality
Community time-series epidemiologic studies can be designed to investigate seasonal effects
by incorporating seasonal interaction terms for the exposure surrogate and/or meteorology (e.g.,
Dominici et al, 2000, 005828). Sheppard et al. (2005, 079176) examined the role of seasonality on
epidemiologic models. They found that a for the population will vary seasonally. This makes sense
because a is a function of the amount of time spent indoors and outdoors and of indoor ventilation.
Given that use of ambient CO concentration instead of ambient CO exposure biases the coefficient
used in epidemiologic models by a, Sheppard et al. (2005, 079176) found that seasonal trends
causing a change in a would contribute additional positive or negative bias, depending on the season
and region of the country. However, several studies discussed in Chapter 5 investigated seasonal
effects. No consistent seasonal pattern across health outcomes was observed in these studies.
3.6.8.5. CO Exposure in Copollutant Mixtures
Because CO exposures most often occur together with exposure to other combustion-related
pollutants, especially in traffic, interpretation of health studies using ambient CO data can be a
challenge, as discussed further in Chapter 5. Ambient CO concentrations from AQS data (Section
3.5.3) have been shown to be correlated with ambient concentrations of NO2 and VOCs, and
personal CO exposures have been correlated with personal PM and VOC exposures (Section 3.6.7).
Correlation between factors is one condition for confounding, so it is possible that coexistence of CO
with NO2 or VOCs could confound estimates of the health effects of ambient CO concentrations, and
CO concentration could potentially confound estimates of the health effects of NO2 or VOCs. For
this to be true, both CO and the copollutant would have to be correlated with the health outcome of
interest. The moderately high correlations between ambient CO and copollutants make it difficult to
discern the extent to which CO and other compounds are associated with a given health effect.
It is also possible that the factor of interest may be the multipollutant mixture emitted from on-
road or other combustion processes. The HEI Report on Traffic Related Pollutants (HEI, 2009,
191009) suggests that ambient CO, NO2, and benzene could all be considered as surrogates for
mobile source-related pollution, but none are ideal surrogates for mobile-source pollution because
ambient CO concentration tends to decrease rapidly with distance from the source (e.g., Baldauf et
al., 2008, 190239: Zhu et al., 2002, 041553). NO2 is reactive and benzene is volatile. Additionally,
PM components of mobile source emissions change rapidly in size and composition from secondary
formation and other atmospheric processing. Given that the mixture of mobile source-related
emissions changes rapidly as a result of these factors, the ratio of CO to other components of mobile-
source emissions also changes. Hence, even if CO is itself stable within the mixture of copollutants,
the dynamic evolution of the mixture may change the representativeness of CO as an indicator of
that mixture over time. Additionally, reductions in CO emissions over the past 30 yr have brought
ambient CO concentrations down substantially, with more than half of hourly measurements below
the LOD for most instruments (Section 3.5.1.1). Furthermore, CO and other copollutants found in
mobile-source emissions have multiple anthropogenic and biogenic sources and, as a result, are
difficult to attribute solely to mobile source pollution (Section 3.2). For all of these reasons, the
representativeness of CO as an indicator of the multipollutant mixture of mobile-source emissions
has not been clearly determined.
3.6.8.6. Conclusions
This section presents considerations for exposure assessment and the exposure errors and
uncertainties that can potentially affect health effects estimates. These issues can be categorized into
the following areas: measurement, nonambient sources, spatial variability, temporal variability, and
CO in copollutant mixtures. Potential influences of each of these sources on health effect estimates
derived from panel, time-series, and longitudinal epidemiologic studies are described above.
Additionally, error sources have the potential to interact with each other. For example, CO
concentrations have been shown to decrease rapidly with distance from a highway, and so spatial
variability is an important issue in assessing CO exposure. Exposure error may occur if the ambient
CO concentration measured at the central site monitor is used as an ambient exposure surrogate and
differs from the actual ambient CO concentration outside a subject's residence and/or worksite.
January 2010 3-94
-------
However in time-series epidemiologic studies, spatial variability will only be an important source of
error if the time series of CO concentration at different locations are not well correlated in time. The
spatial variability of CO, in mixture with the dynamically changing group of mobile source
pollutants, adds to the difficulty of quantifying the health effects related specifically to CO compared
with those related to other combustion-related copollutants. In most circumstances, exposure error
tends to bias a health effect estimate downward (Sheppard et al., 2005, 079176; Zeger et al., 2000,
001949). Insufficient spatial or temporal resolution to capture true variability and correlation of CO
with copollutants are examples of sources of uncertainty that could widen confidence intervals and
so reduce the statistical significance of health effects estimates.
3.7. Summary and Conclusions
3.7.1. CO Sources, Emissions, and Chemistry
In the U.S., on-road mobile sources constituted more than half, or ~61 MT out of-117 MT, of
total CO emissions in the 2002 NEI and BEIS, which are the most recent publicly available CO
emission datasets meeting EPA's data quality assurance objectives. In metropolitan areas in the U.S.,
for example, as much as 75% of all CO emissions can come from on-road vehicle exhaust
(U.S. EPA, 2006, 157070). The majority of these on-road CO emissions derive from gasoline-
powered vehicles since the O2 content, pressure, and temperature required for diesel fuel ignition do
not produce large quantities of CO. Anthropogenic CO emissions are estimated to have decreased
35% between 1990 and 2002. On-road vehicle sector emissions controls have produced nearly all
these national-level CO reductions. Nationally, biogenic emissions, excluding fires, were estimated
to contribute ~5%, or ~5.8 MT, of total CO emissions from all sources in 2002, and fires in 2002
added another 16%, or -18.5 MT, to the national CO emissions total. Although these estimates are
generated using well-established approaches, uncertainties inherent in the emission factors and
models used to represent sources for which emissions have not been directly measured and vary by
source category, season, and region.
In addition to being emitted directly by incomplete combustion, CO is produced by
photooxidation of CH4 and other VOCs in the atmosphere, including NMHCs. Estimating the CO
yield from oxidation of HCs larger than CH4 requires computing the yields of several intermediate
products and reactants from oxidation of the parent molecules. The major pathway for removal of
CO from the atmosphere is reaction with OH to produce CO2 and HO2. The mean photochemical
lifetime (T) of CO in the northern hemisphere is ~57 days. During winter at high latitudes, CO has
nearly no photochemical reactivity on urban and regional scales.
3.7.2. Climate Forcing Effects Related to CO
Recent data do not alter the current well-established understanding of the role of urban and
regional CO in continental- and global-scale chemistry outlined in the 2000 CO AQCD (U.S. EPA,
2000, 000907) and subsequently confirmed in the recent global assessments of climate change by the
Intergovernmental Panel on Climate Change (IPCC, 2001, 156587: IPCC, 2007, 092765). CO is a
weak direct contributor to RF and greenhouse warming. Sinha and Toumi (1996, 193747) estimated
the direct RF of CO computed for all-sky conditions at the tropopause to be 0.024 W/mz based on an
assumed change in CO mean global concentrations from 25 to 100 ppb since preindustrial times. The
direct RF attributed to CO over this time-frame is -1.5% of the direct RF for CO2 estimated by the
IPCC (Forster et al., 2007, 092936).
More importantly, CO can indirectly cause increased RF because it reacts with tropospheric
OH and thus can increase the lifetime of trace gases in the atmosphere including the GHGs CH4 and
O3. Additionally, the major pathway for removal of CO from the atmosphere is reaction with OH to
produce CO2. CH4, O3, and CO2 absorb infrared radiation from the Earth's surface and contribute to
the greenhouse effect. Indirect RF attributed to 1750-2005 emissions of CO through changes in
concentration of the GHGs O3, CH4, and CO2 was estimated by Forster et al. (2007, 092936) to be
~0.2 W/m2 or -12% of the direct RF of CO2 (Figure 3-7). The future direct and indirect integrated
January 2010 3-95
-------
RF for year 2000 emissions of CO was estimated to be ~0.2 W/m2-yr with -50% uncertainty over
both 20-yr and 100-yr time horizons (Figure 3-8). The RF related to short-lived CO is -25% of that
for CO2 for a 20-yr time horizon, but only -7% of that for longer-lived CO2 over a 100-yr time
horizon. Overall, the evidence reviewed in this assessment is sufficient to conclude that a Causal
relationship exists between current atmospheric concentrations of CO and effects on
climate
3.7.3. Ambient CO Measurements
As of August 2009, 24 automated FRMs and no FEMs had been approved for monitoring CO.
All EPA FRMs for CO operate on the principle of nondispersive infrared (NDIR) detection and can
include gas filter correlation (GFC). Current specifications for CO monitoring are designed to help
states demonstrate whether they have met compliance criteria, with requirements for an LOD of
1 ppm. The reported LOD for 20 of the 24 FRMs is 0.5 ppm, and four models of FRMs are in
operation with an LOD of 0.04 ppm. FRMs with higher LOD also are limited to a precision of
0.1 ppm and are more subject to drift compared with newer monitors with automatic drift correction
options.
For 2005-2007, there were 291 CO monitors meeting the 75% completeness requirements and
reporting values year-round to the AQS in the 50 states, plus the District of Columbia, Puerto Rico,
and the Virgin Islands. 57 monitors across the U.S. have been sited at microscale to capture
near-road concentrations, 31 have been sited at middle scale, and 119 are sited for neighborhood
scale monitoring; among the remaining 84 monitors, states did not declare the spatial scale of
monitoring for 71 monitors, and 13 are sited for monitoring urban or regional scale. For CO, traffic
is the major source in an urban setting and therefore microscale data are sited "to represent
distributions within street canyons, over sidewalks, and near major roadways" while middle scale
monitors are sited to represent "air quality along a commercially developed street or shopping plaza,
freeway corridors, parking lots and feeder streets" (40 CFR Part 58 Appendix D). At middle and
neighborhood scales, required minimum monitor distance from a road is directly related to the road's
average daily traffic count to capture community averages. Ambient monitors for CO and other
criteria pollutants are located to monitor compliance rather than population exposures. However,
AQS monitors are often used for exposure assessment. When comparing CO monitor location with
population density, it was observed that population coverage varies both within and between cities.
3.7.4. Environmental CO Concentrations
CO concentration data for 1-h and 8-h intervals were available for 243 counties and
autonomous cities or municipalities that maintained active CO monitoring stations meeting the 75%
completeness criteria for the years 2005-2007. There were no violations of the 1-h or 8-h NAAQS in
those years. The nationwide mean, median, and interquartile range for 1-h measurements reported
between 2005 and 2007 were 0.5, 0.4, and 0.4 ppm, respectively, and these statistics did not change
by more than 0.1 ppm for each year of the 3-yr period. More than 50% of the data nationwide were
below the LOD for the majority of monitors in use. The nationwide mean, median, and interquartile
range for 8-h daily max concentrations, reported between 2005 and 2007, were 0.7, 0.5, and
0.5 ppm, respectively. Half of the 8-h daily max concentrations fell below the LOD for the majority
of CO monitors in the field. The 2006 annual second highest 8-h CO concentration, averaged across
144 monitoring sites nationwide, was 75% below that for 1980 and is the lowest concentration
recorded during the past 27 yr. The mean annual second highest 8-h ambient CO concentration has
been below 5 ppm since 2004. The downward trend in CO concentrations in the 1990s parallels the
downward trend observed in CO emissions and can be attributed largely to decreased mobile source
emissions.
The correlation structures for measurements at the monitors in each of the 11 CS As/CBS As
examined for this assessment reveal a wide range of responses between monitors in each city and
among the cities. While this wide range is produced by the interactions of many physical and
chemical elements, the location of each monitor and the uniqueness of its immediate surroundings
can often explain much of the agreement or lack thereof. CO concentrations can be elevated near
roadways and decrease with increasing distance from the road. Anchorage, AK, had concentrations
roughly twice those of the other metropolitan areas. Most of the CS As/CBS As examined here had
January 2010 3-96
-------
diel concentration curves with pronounced morning and evening rush hour peak CO levels, although
diel CO concentrations had less variability for New York City, Atlanta, and Seattle than for the other
eight cities. For most metropolitan areas examined here, concentrations were generally highest in the
winter (December-February) and fall (September-November) and decreased, on average, during the
spring (March-May) and summer (June-August). Measurements near or below the 0.5 ppm LOD for
most instruments, coupled with the coarsely reported measurement resolution of 0.1 ppm, can
artificially influence the comparison statistics shown in the tables and result in apparent
heterogeneity in the box plots (Figure 3-19 and Figure 3-22).
CO measurements obtained at different monitoring scales were compared to assess spatial
variability of CO concentration. The median hourly CO concentration across the U.S. obtained at
microscale monitors was 25% higher than at middle scale and 67% higher than at neighborhood
scale. The microscale and middle scale CO data reported here are consistent with hourly
concentrations reported in the literature for the near-road environment within the United States, with
CO concentration decaying with downwind distance from the road. Determinants of spatial
variability of ambient CO concentration within the near-road environment include roadway density,
traffic counts, meteorology, and natural and urban topography.
In all cases, a wide range of correlations existed between CO and copollutants computed from
AQS data. The mean and median correlations between CO and copollutants were positive for NO2,
PM10, and PM2.5; near zero for SO2; and negative for O3. These findings might reflect common
combustion sources for CO, NO2, and PM. Among those copollutants with positive associations,
NO2 had the highest mean and median correlations, followed by PM2 5 and PMi0. Within and
between individual metropolitan areas, the distribution of copollutant correlations varied
substantially. Studies in the literature also found fairly high correlations of CO with EC and certain
VOCs.
This assessment has used data from 2005-2007 at 12 remote sites as part of the international
CCGG CASN in the CONUS, Alaska, and Hawaii to determine PRB. All sites demonstrate the well-
known seasonality in background CO, with minima in the summer and fall and maxima in the winter
and spring. The 3-yr avg CO PRB in Alaska was 130 ppb; in Hawaii it was 99 ppb; and over the
CONUS it was 132 ppb.
3.7.5. Exposure Assessment and Implications for Epidemiology
Very few recent exposure assessment studies involve ambient CO concentration data. The
studies of personal exposure to ambient CO presented here generally found that the largest
percentage of time in which an individual is exposed to ambient CO occurs indoors, but the highest
ambient CO exposure levels occur in transit. In-vehicle CO concentrations are typically reported to
be between 2 and 5 times higher than ambient concentrations, although peak in-vehicle
concentrations more than an order of magnitude higher than corresponding ambient monitor
concentrations have also been reported. Among commuters, exposures were higher for those
traveling in automobiles in comparison with those traveling on buses and motorbikes and with those
cycling or walking. Ambient CO exposure in automobiles has been demonstrated to vary with
vehicle ventilation settings, and a very small portion of that exposure is thought to come from the
vehicle in which the exposed person travels. High near-road CO concentrations can be important for
those living in the near-road environment because virtually all of ambient CO infiltrates indoors.
Hence, indoor exposure to ambient CO is determined by the CO concentration outside the building.
Residents of the 17.9 million occupied homes located within approximately 90 m of a highway,
railroad, or airport may be exposed to elevated ambient CO levels. However, CO concentration in
the near-road environment has been shown to decrease sharply with downwind distance from a
highway, wind direction, emission source strength (e.g., number of vehicles on a highway). Natural
and urban topography also influence localized ambient CO levels.
Recent exposure assessment studies support one of the main conclusions of the 2000 CO
AQCD, that central-site ambient CO monitors may overestimate or underestimate individuals'
personal exposure to ambient CO because ambient CO concentration is spatially variable,
particularly when analyzing exposures in the near-road environment. Exposure error may occur if the
ambient CO concentration measured at the central-site monitor is used as an ambient exposure
surrogate and differs from the actual ambient CO concentration outside a subject's residence and/or
worksite. For example, measurement at a "hot spot" could skew community exposure estimates
upwards, and likewise measurement at a location with few nearby CO sources could skew exposure
January 2010 3-97
-------
estimates downwards. Correlations across CO monitors can vary widely from within and between
cities across the U.S. as a function of natural and urban topography, meteorology, and strength and
proximity to sources. Typically, intersampler correlation ranges from 0.35 to 0.65 for monitors sited
at different scales within a metropolitan area, although it can be greater than 0.8 in some areas.
Health effects estimates from time-series epidemiologic studies are not biased by spatial variability
in CO concentrations if concentrations at different locations are correlated in time. Additionally,
exposure assessment is complicated by the existence of CO in multipollutant mixtures emitted by
combustion processes. Because ambient CO exists in a mixture with volatile and reactive pollutants,
the correlation between exposure to ambient CO and copollutants can vary substantially over time
and across locations. For this reason, it is difficult to quantify the effects related specifically to CO
exposure compared with those related to another combustion-related pollutant or mix of pollutants.
In most circumstances, exposure error tends to bias a health effect estimate downward. Spatial and
temporal variability not fully captured by ambient monitors and correlation of CO with copollutants
are examples of sources of uncertainty that could widen confidence intervals of health effects
estimates.
January 2010 3-98
-------
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
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
Bergamaschi P; Hein R; Hermann M; Crutzen PJ (2000). Inverse modeling of the global CO cycle 1. Inversion of CO
mixing ratios. J Geophys Res, 105: 1909-1927. 192377
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)
January 2010 3-99
-------
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 Chem Phys
Meteorol, 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
Bogo H; Gomez DR; Reich SL; Negri RM; 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
Brauer M; Hoek G; van Vliet P; Meliefste K; Fischer P; Gehring U; Heinrich J; Cyrys J; Bellander T; Lewne M;
Brunekreef B (2003). Estimating long-term average particulate air pollution concentrations: application of traffic
indicators and geographic information systems. Epidemiology, 14: 228-239. 155702
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
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
Briggs DJ; Collins S; Elliott P; Fischer P; Kingham S; Lebret E; Pryl K; Van Reeuwijk H; Smallbone K; Van Der Veen A
(1997). Mapping urban air pollution using GIS: a regression-based approach. Int J Geogr Inform Sci, 11: 699-718.
025950
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; Scotto di Marco G; Kephalopoulos S; Cavallo D; Maroni M
(2004). Personal carbon monoxide exposure levels: Contribution of local sources to exposures and
microenvironment concentrations in Milan. J Expo Anal 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
PM2.5 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
Cassidy BE, Alabanza-Akers MA, Akers TA, Hall DB, Ryan PB, Bayer CW, Naeher LP (2007). Particulate matter and
carbon monoxide multiple regression models using environmental characteristics in a high diesel-use area of
Baguio City, Philippines. Sci Total Environ, 381: 47-58. 199975
January 2010 3-100
-------
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
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
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 interactions (pp. -). New 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
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
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). IND AIR: 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
Edwards DP; Emmons LK; Hauglustaine DA; Chu DA; Gille JC; Kaufman YJ; Petron G; Yurganov LN; Giglio L; Deeter
MN; Yudin V; Ziskin DC; Warner J; Lamarque J-F; Francis GL; Ho SP; Mao D; Chen J; Grechko El; Drummond
JR (2004). Observations of carbon monoxide and aerosols from the Terra satellite: Northern Hemisphere variability.
J Geophys Res, 109: D24202. 199889
January 2010 3-101
-------
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
F inlay son-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
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
Fuglestvedt J; Berntsen T; Myhre G; Rypdal K; Skeie RB (2008). Climate forcing from the transport sectors. PNAS, 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
Gilliland F; Avol E; Kinney P; Jerrett M; Dvonch T; Lurmann F; Buckley T; Breysse P; Keeler G; de Villiers T; McConnell
R (2005). Air pollution exposure assessment for epidemiologic studies of pregnant women and children: lessons
learned from the Centers for Children's Environmental Health and Disease Prevention Research. Environ Health
Perspect, 113: 1447-54. 098820
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
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
January 2010 3-102
-------
Gomez-Perales JE; Colvile RN; Nieuwenhuijsen MJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-Figueroa
VH; Blanco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R (2004). Commuters' exposure to
PM2.5, CO, and benzene in public transport in the metropolitan area of Mexico City. Atmos Environ, 38: 1219-
1229.054418
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
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: 3559-3567. 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
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
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
Hoek G; Beelen R; de Hoogh K; Vienneau D; Gulliver J; Fischer P; Briggs D (2008). A review of land-use regression
models to assess spatial variation of outdoor air pollution. Atmos Environ, 42: 7561-7578. 195851
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: L04801. 191253
Inman RE; IngersollRB; Levy EA (1971). Soil: a natural sink for carbon monoxide. Science, 172: 1229-1231.010972
IPCC (2001). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment
Report (TAR) of the Intergovernmental Panel on Climate Change [Houghton, J.T.,Y. Ding, DJ. Griggs, M. Noguer,
PJ. van der Linden, X. Dai, K. Maskell, and C.A. Johnson (eds.)]. Cambridge, UK and New York, NY, USA:
Intergovernmental Panel on Climate Change, Cambridge University Press. 156587
IPCC (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report (AR4) of the Intergovernmental Panel on Climate Change . Cambridge, UK and New York, NY:
Intergovernmental Panel on Climate Change, Cambridge University Press. 092765
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
January 2010 3-103
-------
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. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EPA-600/4-84-014. 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
Kanakidou M; 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 JF; 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). Atmospheric carbon monoxide: latitudinal distribution of sources. Geophys Res Lett,
17: 1913-1916.012443
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 inAmazonia. 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
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. Report Nr. 140. 191193
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;
Nieuwenhuijsen MJ (2004). Personal exposures and microenvironment concentrations of PM25, VOC, NO2 and
CO in Oxford, UK. Atmos Environ, 38: 6399-6410. 056811
January 2010 3-104
-------
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; Feichter J; Flatau PJ; Heland J; Holzinger R;
Korrmann R; Lawrence MG; Levin Z; Markowicz KM; Mihalopoulos N; Minikin A; Ramanathan V; De Reus M;
Roelofs GJ; Scheeren HA; Sciare J; Schlager H; Schultz M; Siegmund P; Steil B; Stephanou EG; Stier P; Traub M;
Warneke C; Williams J; Ziereis H (2002). Global air pollution crossroads over the Mediterranean. Science, 298:
794-799. 190361
Lobert JM; Scharffe DH; Hao WM; Kuhlbusch TA; Seuwen R; Warneck P; Crutzen PJ (1991). Experimental evaluation of
biomass burning emissions: Nitrogen and carbon containing compounds. In Levine JS (Ed.),Global biomass
burning: Atmospheric, climatic, and biospheric implications (pp. 289-304). Cambridge, 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. JExpo Sci Environ Epidemiol, 10: 566-578. 000782
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 RN; 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. EPA-600/S4-83-013. 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 Am erica: 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: ANARSTO 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
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
NRC (1999). Ozone-forming potential of reformulated gasoline. Washington, DC: National Research Council, National
Academy of Sciences. 010614
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
Paciorek CJ; Yanosky JD; Puett RC; Laden F; Suh HH (2009). Practical large-scale spatio-temporal modeling of particulate
matter concentrations. Ann Appl Stat, 3: 370-397. 190090
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
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
January 2010 3-105
-------
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; Iran 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
Rind D; 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
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]. California
Environmental Protection Agency, Air Resources Board; South Coast Air Quality Management District.
Sacramento, CA.http://www.arb.ca.gov/research/abstracts/95-339.htm. 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
Ryan PH; LeMasters GK (2007). A review of land-use regression models for characterizing intraurban air pollution
exposure. Inhal Toxicol, 19: 127. 156063
Ryan PH; LeMasters GK; Levin L; Burkle J; Biswas P; Hu S; Grinshpun S; Reponen T (2008). A land-use regression
model for estimating microenvironmental diesel exposure given multiple addresses from birth through childhood.
Sci Total Environ, 404: 139-147. 156064
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 spatial variability 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
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; Bell N; Schmidt GA (2005). An emissions-based view of climate forcing by methane and
tropospheric ozone. Geophys Res Lett, 32: L04803. 080129
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; Koch DM; Schmidt GA; Unger N; Bauer SE (2009). Improved attribution of climate forcing to
emissions. Science, 326: 716-718. 201599
January 2010 3-106
-------
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
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. J Expo 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. JExpo 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
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. Retrieved 02-DEC-09, from
http ://www. epa. gov/ttnchie 1 /net/2002inventory.html. 157070
U.S. EPA (2006). Air quality criteria for ozone and related photochemical oxidants. EPA. DC. 088089
U. S. EPA (2006). Cold temperature effects on vehicle HC emissions: draft report. U. S. Environmental Protection Agency.
Washington, DC. EPA420-D-06-001. http://www.epa.gov/oms/regs/toxics/420d06001.pdf. 199897
U.S. EPA (2008). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EPA/600/R-08/071. 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). U.S. Environmental Protection Agency. 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
van der Werf GR; Randerson JT; Giglio L; Collatz 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
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
January 2010 3-107
-------
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
Yanosky JD; Paciorek CJ; Schwartz J; Laden F; Puett R; Suh HH (2008). Spatio-temporal modeling of chronic PM10
exposure for the Nurses' Health Study. Atmos Environ, 42(18): 4047-4062. 099467
Yanosky JD; Paciorek CJ; Suh HH (2009). Predicting chronic fine and coarse particulate exposures using spatiotemporal
models for the northeastern and midwestern United States. Environ Health Perspect, 117: 522-529. 190114
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
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
January 2010 3-108
-------
Chapter 4. Dosimetry and
Pharmacokinetics of Carbon Monoxide
4.1. Introduction
Inhaled ambient CO elicits various health effects by binding with and altering the function of a
number of heme-containing molecules, mainly Hb. Traditional concepts for CO pathophysiology
have been based on the high affinity of CO for hemoglobin, resulting in COHb formation and
consequent reduction in O2-carrying capacity of blood and impaired O2 delivery to tissues. Research
on CO pharmacokinetics dates back to the 1890s, but since the late 1970s has become limited.
Current literature primarily focuses on endogenous CO produced by the metabolic degradation of
heme by heme oxygenase (HO) and its role as a gaseous messenger. This chapter reviews the
physiology and pharmacokinetics of CO. The chapter draws heavily from Chapter 5 of the previous
AQCD (U.S. EPA, 2000, 000907). Relevant new data are included when available. Recent models of
Hb binding are characterized, as well as measurements of tissue CO concentrations using new
methods of extraction.
CO binds with a number of heme-containing molecules including Mb and cytochromes, but
none have been studied as extensively as Hb. The primary focus of this chapter is placed on the
models and kinetics of such binding and the factors influencing this event. The chapter discusses
effects at ambient or near ambient levels of CO leading to low COHb levels (< 5%); however few
studies are available at ambient CO concentrations. Both human and animal studies using higher CO
exposure concentrations, resulting in moderate to high COHb levels (<20%), are discussed where
needed to understand CO kinetics, pathophysiologic processes, and mechanisms of cytotoxicity.
Where human studies could not experimentally test certain hypotheses or were unavailable, animal
experiments were used as surrogates. CO uptake and elimination has been shown to be inversely
proportional to body mass over environmentally relevant exposure levels, meaning the smaller the
animal, the faster the rate of absorption and elimination (Klimisch et al., 1975, 010762; Tyuma et al.,
1981, 011226). However, the basic mechanisms of CO toxicity between experimental animals and
humans are similar and are thus extrapolated from animals to humans in this chapter, keeping in
mind a number of interspecies differences.
4.2. Carboxyhemoglobin Modeling
4.2.1. The Coburn-Forster-Kane and Other Models
Investigators have modeled the effect of CO binding to Hb in a number of ways. Empirical
and mechanistic models are two distinct approaches that have been taken to model in vivo COHb
formation after CO exposure. First, empirical models were used to predict COHb by regressing
concentration and duration of exogenous CO exposure with observed COHb, with or without the
inclusion of physiological predictors such as initial COHb levels and alveolar ventilation (VA). These
methods were reviewed in depth in the previous AQCD (U.S. EPA, 2000, 000907). It is important to
note that CO empirical regression models are limited to estimating COHb in the exact conditions on
which the models were based. These simple models include those by Peterson and Stewart (1970,
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).
January 2010 4-1
-------
012416) and Ott and Mage (1978, 011124). as well as various others (Chung, 1988, 012749; Forbes
et al, 1945, 012850; Selvakumar et al, 1992, 013750; Sharan et al., 1990, 003798; Singh et al.,
1991, 013583). Using a linear differential equation where ambient CO concentrations varied, it was
shown that the presence of brief ambient CO concentration spikes averaged over hourly intervals
may lead to underestimating the COHb concentration by as much as 21% of the true value. To avoid
this problem, it was suggested that ambient CO measurements be monitored and averaged over 10-
to 15-min periods (Ott and Mage, 1978, 011124). Other empirical models predict COHb as a
function of exposure time (Sharan et al., 1990, 003798; Singh et al., 1991, 013583) or exposure time
and altitude (Selvakumar et al., 1992, 013750). A comparison of empirical model predictions showed
a wide disparity in predicted COHb values, highlighting the inaccuracy of these models outside of
the conditions on which they were presented (Tikuisis, 1996, 080960).
Secondly, mechanistic models use physical and physiological processes and an understanding
of biological processes to predict COHb production. The most commonly used mechanistic method
for predicting levels of blood COHb after CO inhalation is the Coburn-Forster-Kane equation or
CFK model developed in 1965 (Coburn et al., 1965, 011145). This differential equation was
developed to examine endogenous CO production, using the major physiological and physical
variables influencing this value. Since then, it has been shown to provide a good approximation to
the COHb level at a steady level of inhaled exogenous CO (Peterson and Stewart, 1975, 010696;
Stewart et al., 1973, 012428). The CFK model describes a four-element, physical system containing
an exogenous CO source, a transfer interface, an endogenous CO source, and a storage compartment.
The linear CFK model assumes O2Hb concentration is constant and is as follows in Equation 4-1:
V,
d[COHb]t
dt
= Vco -
[COHb]0P-02
[02HbJvI
1
1
D,CO
V
p -P
1 B L H20
VA
P,CO
p -P
B 1 H20
DLCO A
V
Equation 4-1
Vb
[COHb],
Vco
[COHb]0
[02Hb]
Pc02
M
DiCO
PB
PH20
VA
PiCO
blood volume in milliliters (ml)
COHb concentration at time t in ml CO/mL blood, at standard temperature and pressure, dry (STPD)
endogenous CO production rate in mL/min, STPD
COHb concentration at time zero in ml CO/mL blood, STPD
02Hb concentration in mL02/mL blood, STPD
average partial pressure of 02 in lung capillaries in mmHg
Haldane coefficient representing the CO chemical affinity for Hb
lung diffusing capacity of CO in mL/min/mmHg, STPD
barometric pressure in mmHg
saturation pressure of water vapor at body temperature in mmHg (47 mmHg)
alveolar ventilation in mL/min, STPD
CO partial pressure in inhaled air in mmHg
The linear CFK model assumes instant equilibration of COHb concentration between venous
and arterial blood, gases in the lung, and COHb concentrations between blood and extravascular
tissues, which is not physiologically representative. The nonlinear CFK equation extends the linear
CFK equation to incorporate the interdependence of COHb and O2Hb levels since they are derived
from the same pool of blood Hb. This interdependence can be modeled by substituting (1.38 Hb
[COHb]) for O2Hb, where THb refers to the number of grams of Hb per mililiter of blood (Peterson
and Stewart, 1975, 010696). The nonlinear CFK differential equation is as follows in Equation 4-2:
January 2010
4-2
-------
d[COHb]t •
Vh = Vco-
dt
[COHb]0P-02
(1.38[THb]-[COHb]t)M
1
1 P — P
1 , rB rH2O
DLCO ,%
V VA J
DLCO
VA
Equation 4-2
The nonlinear equation is more physiologically accurate; however the linear CFK equation
gives a good approximation to the nonlinear solution over a large range of values during CO uptake
and during low levels of CO elimination (Smith, 1990, 013164). The linear equation prediction of
COHb concentration at or below 6% will deviate by no more than ± 0.5% COHb from the nonlinear
equation prediction. Sensitivity analysis of the CFK equations has shown that alterations in each
variable of the equation will affect the outcome variably at different times of exposure, so that the
relative importance of the CFK variables will change with the experimental conditions (McCartney,
1990, 013162). Figure 4-1 illustrates the temporal changes in fractional sensitivities of the principal
physiological determinants of CO uptake for the linear form of the CFK equation, where THb is the
total blood concentration of Hb in g Hb/mL blood and F:CO is the fractional concentration of CO in
ambient air in ppm. The fractional sensitivity of unity means that, for example, a 5% error in the
selected variable induces a 5% error in the predicted COHb value by the nonlinear model. As Figure
4-1 demonstrates, a constant or given percent error in one variable of the model does not generally
produce the same error in the calculated blood COHb, and the error is time dependent. Thus, each
variable influencing CO uptake and elimination will exert its maximal influence at different times of
exposure. This analysis found that only F:CO (shown in Figure 4-1) and VCo will not affect the rate
at which equilibrium is reached (McCartney, 1990, 013162).
January 2010
4-3
-------
1.00
1.00
CD
- -0.75
-1.00
0.6s
-2
6s
-1
1 min 10 min 1.7 h
0 1 2
Log Time
17h
3
-1.00
Source: Adapted with Permission of the American Industrial Hygiene Association from McCartney (1990, 013162)
Figure 4-1. Plot of fractional sensitivities of selected variables versus time of exposure.
The mechanistic CFK model contains a number of assumptions under which the model is
solely applicable, including: (1) ventilation is a continuous process; (2) equilibrium between plasma
CO concentration and COHb concentration is obtained in the pulmonary system; (3) percent COHb
can exceed 100% saturation in the linear model; and (4) it does not account for the shape of the O2 or
CO saturation versus pO2 or pCO relation (McCartney, 1990, 013162). Estimations outside of these
assumptions have been attempted but with less predictive agreement. For example, transient
exposures such as those that would simulate everyday conditions would violate the assumption of a
single, well-mixed vascular compartment. COHb levels during exposure of subjects exposed to
frequent but brief high CO exposures (667-7,500 ppm for 75 s to 5 min) were not accurately
predicted by CFK modeling (Benignus et al., 1994, 013908: Tikuisis et al., 1987, 012219; Tikuisis et
al., 1987, 012138). Consistently, the COHb value predicted by the nonlinear CFK overpredicted
observed venous COHb (0.8-6%) and underpredicted arterial COHb (1.5-6.1%) and this disparity
increased after exercise. Individual differences between arterial and venous COHb varied from
2.3-12.1% COHb (mean, 6.2 ± 2.7% COHb), where the observed steady state COHb averaged -14%
and the observed arterial peak COHb averaged -17.5% (Smith et al., 1994, 076564)(Benignus et al.,
1994, 013908). These inaccuracies between measured and predicted COHb values disappeared after
simulated mixing of arterial and venous blood and thus are likely due to delays in mixing of arterial
and venous blood and differences in cardiac output and lung wash-in. This discrepancy in predicted
and observed COHb suggests that over a short period (<10 min) the arterial COHb levels that are
delivered to tissues could be higher than what is predicted by the CFK equation. A modified CFK
was created to adjust for these issues and produce a more accurate COHb prediction (Smith et al.,
1994, 076564). This expanded CFK model used multiple compartments to model the lung, arm
circulation, and the rest of the body (quickly and slowly perfused tissues). This model was more
January 2010
4-4
-------
accurate than the nonlinear CFK in predicting the individual peak or maximal values of arterial and
venous COHb during CO uptake in the first 10 min after exposure. However, both the nonlinear
CFK and this expansion produced accurate predictions several minutes after the 5-min exposure
ended. The expanded model required the use of two parameters, VA and Vb, that were not measured
individually or derived from the literature, and instead were estimated by adjustments between the
simulations and experimental subject data.
In addition to the limitations discussed above, the CFK model does not account for
extravascular storage sites for CO, such as muscle Mb. CO will undergo reversible muscle Mb
binding, similar to Hb, as well as uptake into other extravascular tissues (Vreman et al., 2006,
098272). The most recent adaptation to the CFK equation incorporates alveoli-blood and blood-
tissue CO exchanges and mass conservation of CO at all times (Gosselin et al., 2009, 190946). This
model has a single free parameter whose value is estimated from one data set; however, it better
predicted COHb formation over a wide range of CO levels and several temporal scenarios (Stewart
et al., 1970, 013972; Tikuisis et al., 1987, 012138; Tikuisis et al., 1987, 012219; Tikuisis et al., 1992,
013592) compared to the linear CFK model. Like the linear CFK model, this modified model
assumes a constant level of oxyhemoglobin. Sensitivity analysis of the model showed that the most
important parameter influencing the level of COHb in this model is M, followed by PcO2 and VA.
Ambient exposure scenarios were simulated with this model to determine the CO concentrations
needed to reach certain COHb levels in humans from 3 months of age to 40-yr-old adults (Gosselin
et al., 2009, 190946). The CO concentrations needed to achieve 2% COHb vary from 24.4-48.1 ppm
for a 1-h exposure, from 11.1-13.1 ppm for an 8-h exposure, and from 9.8-10.1 ppm for a daily
exposure. Infants (1 yr old) were most sensitive to CO concentrations, whereas newborns (3 mo old)
required the highest CO concentration to reach 2% COHb. Newborns required a higher CO exposure
partially because the values used in the model for the newborn blood Hb concentration (170 gnb/
Lbiood) is higher than at infancy (115 gnb/Lwood) or adulthood (150 gnb/Lwood)- The model was also
used to simulate time profiles of COHb formation for two work week exposure scenarios in a
healthy 40-yr-old man. Figure 4-2A represents a high exposure scenario where the work period is
spent at 35 ppm and the rest of the time at 3 ppm. Figure 4-2B represents a lower exposure scenario
where there is a constant 3 ppm exposure. Both figures consist of 5 days where 24 h are broken up
into 3 consecutive 8-h periods: sleeping from 12 a.m. to 8 a.m.; working with light exercise from 8
a.m. to 4 p.m.; and sitting from 4 p.m. to 12 a.m..
January 2010 4-5
-------
A 7
2 -
1 -
Monday | Tuesday | Wednesday | Thursday
Time (days)
Friday
B 1
0.8
'0.6
$0.4
0.2
n i I i r
Light
exercise
Sleeping I Sitting
Monday
Tuesday
Wednesday | Thursday
Time (days)
Friday
- 24 g
16
4.2
3.6 §•
2.4
1.2
Source: Reprinted with Permission of Informa Healthcare from Gosselin et al. (2009,190946)
Figure 4-2. Simulated COHb formation for two 5-day workweeks. "The 24-h day consists of 3
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
A third approach applied more recently to model COHb formation is the use of
multicompartment or physiologically-based pharmacokinetic (PBPK) models. Cronenberger et al.
(2008, 194085) described a two-compartment population-based model to describe and predict COHb
pharmacokinetics from smoking. This model required a compartment for extravascular binding of
January 2010
4-6
-------
CO to accurately predict COHb formation during multiple short and rapid inhalations followed by a
period of no exposure, as occurs in smoking.
A five-compartment PBPK model has been proposed to predict CO uptake and distribution
from acute inhalation exposure and contains components for lung, arterial blood, venous blood,
muscle tissue, and nonmuscle tissue (Bruce and Bruce, 2003, 193975; Bruce and Bruce, 2006,
193980; Bruce et al., 2008, 193977). This model structure is illustrated in Figure 4-3 and includes
the dynamics of CO storage in the lung and its dependence on ventilation and CO pressure of mixed
venous blood, relaxes the assumption that Hb is saturated by including the role of CO in altering the
O2 dissociation curve, includes a subcompartmentalized muscle tissue compartment, accounts for
dissolved CO in blood and tissue, and predicts COHb based on age and body dimensions. This
multicompartment model is limited by its exclusion of cellular metabolism or Mb diffusion,
simplification of within tissue bed spatial variability, and assumption that ventilation and average
partial pressure of alveolar O2 (PAO2) are constant. Another limitation of this model is that some of
the physiological parameters used in simulations are estimated through visual fits to the COHb
profile and not from experimental or published data. This model better predicts COHb levels when
inspired CO levels change rapidly or when incomplete blood mixing has occurred, and better
predicts the CO washout time course compared to the CFK equation. Bruce and Bruce (2003,
193975) compared the two models and found similar results for long-duration exposure settings
(1,000 min); however, the multicompartment model predicted somewhat lower COHb levels
compared to the CFK model during transient CO uptake conditions when using data taken from
Peterson and Stewart (1970, 012416).
LUNGS
P,co
Source: Adapted with Permission of Elsevier Science from Bruce and Bruce (2008,1939771
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 subcom part merits for diffusion of gases within the tissue.
January 2010
4-7
-------
A multicompartment model of the human respiratory system was developed using
characteristics of the tissue representation of Bruce and Bruce (2003, 193975), and the lung
representation described in Selvakumar et al. (1992, 013750) and Sharan (1999, 194673), which
considered the exchanges of CO, O2, and CO2 (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 the CFK equation to determine the most
accurate predictive model under certain exposure conditions. As was mentioned in Section 4.2.1,
Tikuisis (1996, 080960) conducted a comparison of empirical model predictions that showed a wide
disparity in predicted COHb values, highlighting the inaccuracy of these models outside of the
conditions on which they were presented. Smith et al. (1990, 013164) compared the linear and
nonlinear CFK equations and concluded that the linear CFK equation gives a good approximation
(within 1%) to the nonlinear solution over a large range of values during CO uptake and over a
somewhat smaller range during CO elimination. The linear equation prediction of COHb
concentration at or below 6% will only differ ± 0.5% COHb from the nonlinear equation prediction.
Additionally, the most recently modified CFK model (Gosselin et al., 2009, 190946) better predicted
COHb formation over a wide range of CO levels (50-4,000 ppm) and several temporal scenarios
(Stewart et al., 1970, 013972: Tikuisis et al., 1987, 012138: Tikuisis et al., 1987, 012219: Tikuisis et
al., 1992, 013592) compared to the linear CFK model. Linear regression slopes between the
simulated COHb values from Gosselin et al. (2009, 190946) and the observed experimental values
were closer to 1 in all experimental scenarios, indicating a better fit to the observed data. When
evaluating all validation studies the modified model had an estimated slope of 0.996 (95% CI:
0.986-1.001) compared to 0.917 (95% CI: 0.906-0.927) using the linear CFK model. Bruce and
Bruce (2003, 193975) compared their model to the CFK and found similar results for long-duration
exposure settings (1,000 min [16.5 h]), however, their multicompartment model predicted somewhat
lower COHb levels over transient CO uptake conditions when using data taken from Peterson and
Stewart (1970, 012416). The Bruce and Bruce model better predicts However, there has not been a
quantitative comparison of the recent multicompartment models (Bruce and Bruce, 2003, 193975:
Neto et al., 2008, 194672) and the improved CFK equation models (Gosselin et al., 2009, 190946:
Smith et al., 1994, 076564) to determine which is most accurate in predicting COHb levels under
exposure scenarios that include occasional peak concentrations. The nonlinear and linear CFK
equations remain the most extensively validated and applied models for COHb prediction. COHb
levels when inspired CO levels change rapidly or when incomplete blood mixing has occurred, and
better predicts the CO washout time course compared to the CFK equation. However, there has not
been a quantitative comparison of the recent multicompartment models (Bruce and Bruce, 2003,
193975)(Neto et al., 2008, 194672) and the improved CFK equation models (Smith et al., 1994,
076564)(Gosselin et al., 2009, 190946) to determine which is most accurate in predicting COHb
levels under exposure scenarios that include occasional peak concentrations. The nonlinear and
linear CFK equations remain the most extensively validated and applied models for COHb
prediction.
4.2.4. Mathematical Model Usage
As no new data have become available on the distribution of COHb levels in the U.S.
population since large-scale nationwide surveys - e.g., National Health and Nutrition Examination
Survey II (NCHS; et al., 1982, 011442) - and human exposure field studies - e.g., Denver, CO, and
Washington, DC (Akland et al., 1985, 011618) - were conducted in the 1970s and 1980s,
mathematical models are used to predict the resulting COHb levels from various CO exposure
scenarios. Table 4-1 illustrates the predictions of venous COHb after 1, 8, or 24 h of CO exposure at
a range of concentrations in a healthy adult human at rest (VA = 6 L/min; DLCO = 20
January 2010 4-8
-------
[mL/min]/mmHg), during light exercise (VA =15 L/min; DLCO = 34 [mL/min]/mmHg), and during
moderate exercise (VA = 22 L/min; DLCO = 43 [mL/min]/mmHg). The Quantitative Circulatory
Physiology (QCP) model, which integrates human physiology using over 4,000 variables and
equations based on published biological interactions, was used to predict these values (Abram et al.,
2007, 193859; Benignus et al., 2006, 151344). This dynamic whole body model uses the nonlinear
CFK equation with modifications presented in Smith et al. (1994, 076564). The contribution of
alveolar ventilation and lung diffusion to the changes in COHb levels is discussed in Section 4.3.1.2.
Increased ventilation leads to an increased rate of CO uptake, causing COHb levels to reach
equilibrium earlier. Also, increased ventilation leads to a decrease in steady state COHb levels due to
increased CO expiration. For example, 35 ppm CO exposure at moderate exercise (22 L/min) results
in a lower 24-h COHb saturation (4.73%), compared to COHb saturation from 35 ppm CO at rest
(5.03%) (Table 4-1). Whereas, after 1 h, COHb levels are still increasing following exposure at all
levels of exercise and have not reached steady state, thus the greater uptake from increased
ventilation leads to initially elevated COHb in higher ventilation situations.
Endogenous CO production varies as described in Section 4.5 but generally results in <1%
COHb, with a QCP modeled value of 0.27% at time zero. The rate of endogenous CO production
was set at 0.007 mL/min for this simulation, whereas both higher and lower values have been
reported (Coburn et al., 1966, 010984) (Section 4.5). Table 4-1 illustrates that 35 ppm CO for 1-h
results in between 0.9-1.9% COHb and 9 ppm CO for 8-h results in between 1.1-1.3% COHb,
depending upon activity level. Also, this table shows that low concentration CO exposure over
several hours can result in equivalent COHb levels compared to higher concentration, acute
exposure. For example, in a resting condition without additional baseline COHb, COHb resulting
from 35 ppm for 1 h (0.89%) is approximately equivalent to 6 ppm for 8 h (0.83%) or 4 ppm for 24 h
(O.i
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 = 0.27%; Hb = 0.15 g/mL), during light exercise (VA = 15 L/min;
DLCO = 34 (mL/min)/mmHg), and during moderate exercise (VA = 22 L/min; DLCO =
43 (mL/min)/mmHg). The QCP model used a dynamic nonlinear CFK with the Smith
et al. (1994,076564) COHb algorithm and affinity constant M = 218.
CO (ppm)
2
3
4
6
9
15
24
35
6 L/min
0.30
0.31
0.33
0.36
0.42
0.53
0.70
0.89
1h
15 L/min
0.30
0.33
0.36
0.44
0.55
0.77
1.10
1.50
22 L/min
0.29
0.34
0.62
0.48
0.63
0.92
1.35
1.89
6 L/min
0.45
0.54
0.64
0.83
1.12
1.69
2.55
3.58
8h
15 L/min
0.38
0.51
0.64
0.90
1.29
2.05
3.19
4.55
22 L/min
0.35
0.48
0.62
0.88
1.27
2.06
3.22
4.60
6 L/min
0.54
0.68
0.82
1.10
1.52
2.35
3.57
5.03
24 h
15 L/min
0.40
0.54
0.69
0.97
1.39
2.22
3.45
4.91
22 L/min
0.36
0.49
0.63
0.91
1.31
2.12
3.31
4.73
The QCP model incorporating the Smith et al. (1994, 076564) COHb algorithm was also used
to simulate population exposure scenarios including various commuting concentrations (Figure 4-4)
and endogenous production rates (Figure 4-5). Commuting concentrations were modeled since the
highest ambient CO exposure levels are generally observed during transit (Section 3.6.6.2). Figure
4-4 presents simulated COHb levels in a healthy adult throughout the second of 5 modeled days
containing a 60-min commute at various CO concentrations. The U.S. Census Bureau estimates that
5% of the population commutes in automobiles for 60 or more minutes to work daily
(U.S. Census Bureau, 2008, 194013) and exposure studies have reported in-vehicle transit
concentrations up to 50 ppm (Abi-Esber and El-Fadel, 2008, 190939; Duci et al., 2003, 044199).
However, U.S. studies have reported in-vehicle concentrations of <6 ppm, although peak
January 2010 4-9
-------
concentrations in congested urban areas have been reported to be higher than 50 ppm (Rodes et al.,
1998, 010611)(Riediker et al., 2003, 043761). CO concentrations during commuting lead to spikes in
COHb in this model scenario with a 1% COHb increase over the initial COHb (0.3%) after 50 ppm
exposure. Figure 4-4 also illustrates that the COHb saturation after CO exposure from commuting is
not fully eliminated by the next commuting period. Modeling successive days results in the same
pattern and degree of COHb formation, indicating no accumulation of COHb over time.
1.6
1.4
— 2 ppm 10 ppm 20 ppm —50 ppm
360
720
Time (min)
1080
1440
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; and (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 5 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
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- 218,
initial. COHb- 0.27%. 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.
January 2010
4-10
-------
hemolytic anemia have endogenous CO production rates ranging from 0.012 to 0.053 mL/min
(31-143 (imol/h) (Coburn et al, 1966, 010984). The venous COHb levels in these same patients
ranged from 0.77 to 2.62%.
— 0.007 mL/min 0.02 mL/min 0.04 mL/min —0.06 mL/min
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.
4.3. Absorption, Distribution, and Elimination
4.3.1. Pulmonary Absorption
Pulmonary uptake of CO accounts for all environmental CO absorption and occurs at the
respiratory bronchioles and alveolar ducts and sacs. CO and O2 share various physico-chemical
properties, thus allowing for the extension of the knowledge about O2 kinetics to those of CO despite
the differences in the reactivity of the gases. The exchange of CO between the air and the body
depends on a number of physical (e.g., mass transfer and diffusion), as well as physiological factors
(e.g., alveolar ventilation and cardiac output), which are controlled by environmental conditions,
physical exertion, and other processes discussed in Section 4.4. The ability of the lung to take up
inhaled CO is measured by DLCO, and CO uptake representing the product of DLCO and the mean
alveolar pressure (PACO). The importance of dead space volume, gas mixing and homogeneity, and
January 2010
4-11
-------
ventilation/perfusion matching were discussed in depth in the 2000 CO AQCD (U.S. EPA, 2000,
000907).
4.3.1.1. Mass Transfer of Carbon Monoxide
Mass transfer refers to the molecular and convective transport of CO molecules within the
body stores, driven by random molecular motion from high to low concentrations. CO enters through
the airway opening (mouth and nose) and transfers in a gas phase to the alveoli. CO transport is due
to convective flow, the mechanical action of the respiratory system, and diffusion in the acinar zone
of the lung (Engel et al., 1973, 014336). Then, CO diffuses across the air-blood interface into plasma
and subsequently into red blood cells (RBC), binding RBC Hb. At environmental CO levels, CO
uptake into RBC is limited by the reaction rate of binding of CO to O2Hb forming COHb.
Pulmonary capillary RBC CO diffusion is rapidly achieved (Chakraborty et al., 2004, 193759;
Gibson and Roughton, 1955, 193941; Reeves and Park, 1992, 193847; Roughton and Forster, 1957,
193862). The formation rate and level of COHb depends upon pCO, pO2 in the air, time of exposure,
and the ventilation rate (Roughton and Forster, 1957, 193862). Most of the body CO is bound to Hb;
however, 10-15% of the total body CO is located in extravascular tissues primarily bound to other
heme proteins (Coburn, 1970, 013916). Considerable concentrations of CO have been measured in
spleen, lung, kidney, liver, muscle, and heart (Vreman et al., 2005, 193786; Vreman et al., 2006,
098272). whereas less CO is localized to fatty tissues, such as adipose and brain. The transfer of CO
occurs by a partitioning of CO between Hb and tissue. Less than 1% of the total body CO stores
appear as dissolved in body fluids, due to the insolubility and small tissue partial pressure of CO
(Coburn, 1970, 013916). Transport pathways and body stores of CO are shown in Figure 4-6.
Carbon Monoxide in the Ambient Air
i Endogenous j
i production ^^M
I of CO j
/* **,
! Metabolism ^te
| ofCOtoCO, i
Alveolar Air
'A M •
[ * Plasma * |
RBC
Hemoglobin
• -^- -
^^
Myoglobin
v__
•"
^^^Intracellular
^^^ enzymes
V
Extravascular compartment
Source: Adapted with Permission of Wiley-Blackwell from Coburn (1967, 011144)
Figure 4-6. Diagrammatic presentation of CO uptake and elimination pathways and CO body
stores.
4.3.1.2. Lung Diffusion of Carbon Monoxide
Lung diffusion of CO is an entirely passive process of gas diffusion across the alveolo-
capillary membrane, through the plasma, across the RBC membrane and into the RBC stroma, where
January 2010
4-12
-------
CO binding to Hb rapidly occurs. Membrane and blood phase transfer are governed by physico-
chemical laws, including Pick's first law of diffusion. The diffusing capacity of the lung for CO,
represented as DLCO, is a measurement of the partial pressure difference between inspired and
expired CO. Due to the rapid binding of CO to Hb, a high pressure differential between air and blood
exists when CO air levels are increased. Inhalation of CO-free air reverses the pressure differential
(higher CO pressure on the blood side than the alveolar side), and then CO is released into the
alveolar air. Since CO is also produced endogenously, CO release will also be affected by this
production pressure. However, the air-blood gradient for CO is usually higher than the blood-air
gradient; therefore, CO uptake will be a proportionately faster process than CO elimination.
A number of factors have been found to affect DLCO including Hb concentration, cardiac
output (Q), erythrocyte flow, COHb concentration, PACO2, body position, exercise, time of day, age,
etc. (Forster, 1966, 180430; Hsia, 2002, 193857). DLCO consistently decreases after intense bouts of
exercise, likely due to the redistribution of blood volume to the periphery (Hanel et al., 1997,
193918; Manier et al., 1991, 193979). However, in going from rest to exercise, DLCO can increase
linearly from: lung expansion leading to unfolding and distension of alveolar septa, opening and/or
distension of capillaries as Q increases, increased capillary hematocrit, and more homogeneous
distribution of capillary erythrocytes (Hsia, 2002, 193857). DLCO is less dependent upon lung
volume at mid-range vital capacity, but at extreme volumes the diffusion rate is varied, higher than
average at total lung capacity and lower at residual volume (McClean et al., 1981, 012411).
DLCO is also altered by a number of diseases. Decreased DLCO is evident in patients with
restrictive lung disease (i.e., decreased lung volumes) since a loss of lung tissue leads to a loss of
functional lung units. DLCO also shows a good correlation with the severity of restrictive lung
disease (Arora et al., 2001, 186713). Conditions affecting DLCO vary and include chronic
obstructive pulmonary disease (Terzano et al., 2009, 108046). ulcerative colitis (Marvisi et al., 2000,
186703; Marvisi et al., 2007, 186702), severe gastroesophageal reflux (Schachter et al., 2003,
186707), beta thalassemia (Arora et al., 2001, 186713), thoracic or abdominal aortic aneurysm
(Sakamaki et al., 2002, 186706). pulmonary arterial hypertension (Proudman et al., 2007, 186705).
and chemotherapy for breast cancer (Yerushalmi et al., 2009, 186711). Diseases affecting CO
kinetics and DLCO are also discussed in Section 4.4.4.
4.3.2. Tissue Uptake
4.3.2.1. Respiratory Tract
The upper respiratory tract contributes little to the overall CO uptake. The lung has nearly
constant exposure to CO; however, relatively little CO diffuses into the tissue except at the alveolar
region en route to the circulation. No detectable uptake of CO was observed in the human nasal
cavity or upper airway (Guyatt et al., 1981, 011196) or in the monkey oronasal cavity after high CO
exposure (Schoenfisch et al., 1980, 011404).
4.3.2.2. Blood
The blood is the largest reservoir for CO, where it reversibly binds to Hb. The chemical
affinity of CO for adult human Hb is approximately 218 times greater than that of O2, meaning one
part CO and 218 (210-250) parts O2 would form equal parts of O2Hb and COHb (Engel et al., 1969,
193914; Rodkey et al., 1969, 008151; Roughton, 1970, 013931). This would happen when breathing
air containing 21% O2 and 960 ppm CO. This concept was presented by Haldane and Smith (1895,
010538) and later represented as the Haldane constant M (210-250) in the Haldane equation by
Douglas, Haldane, and Haldane (1912, 013965). M is relatively unaffected by changes in
physiological pH, CO2, temperature, or 2,3-diphosphoglycerate:
COHb -T-O2Hb =Mx (pCO + pO2)
Equation 4-3
The Hb association rate for CO is 10% slower than O2 and occurs in a cooperative manner
(Chakraborty et al., 2004, 193759; Sharma et al., 1976, 193766). Hb is composed of four globin
January 2010 4-13
-------
chains, each containing a heme group capable of binding CO or O2. The associative reaction rates
become faster with successive heme binding, attributed to interactions within the protein and to
strains imposed on the heme and its ligands (Alcantara et al., 2007, 193867). More simply, the
greater the number of heme sites bound to CO, the greater the affinity of free heme sites for O2, thus
causing Hb to bind and retain O2 that would normally be released to tissues. Cooperativity is greatly
reduced in CO dissociation, but the rate of dissociation of CO from Hb is orders of magnitude slower
than O2 (kco = 4 * 10-4 k02), which accounts for the high affinity values (Chakraborty et al., 2004,
193759). The half-time of dissociation reaction is about 11 s at 37°C (Holland, 1970, 193856). In
general, CO uptake to COHb equilibrium is slower in humans and large animals, requiring 8-24 h,
than in smaller species such as rats, which will equilibrate in 1-2 h (Penney, 1988, 012519). Also,
COHb equilibrium within the blood stream is not instantaneous. Men exposed to brief (~5 min)
high-dose CO had an initial delay of 1-2 min in the appearance of venous COHb after the start of CO
inhalation (Benignus et al., 1994, 013908: Smith et al., 1994, 076564). Additionally, arterial COHb
concentrations were considerably higher than venous concentrations during CO exposure; however,
they converged within 2-10 min after the end of exposure, as venous and arterial blood mixed.
CO binding to Hb also has effects on the O2 dissociation curve of the remaining Hb by shifting
the curve progressively to the left and altering the normal S-shaped curve to become more
hyperbolic due to increased cooperative O2 binding (Roughton, 1970, 013931). This is referred to as
the "Haldane effect" and causes tissues to have more trouble obtaining O2 from the blood, even
compared to the same extent of reduced Hb resulting from anemia. For example, Figure 4-7 (as
explained in the 2000 CO AQCD) illustrates that in an acute anemia patient (50% of Hb) at a venous
pO2 of 26 mmHg (v'i), 5 vol % of O2 (50% saturation) was extracted from the blood. In contrast, for
a CO poisoned person with 50% COHb, the venous pO2 will have to drop to 16 mmHg (v'2) to
release the same 5 vol % O2. This more severe effect on O2 pressure may lead to brain O2 depletion
and loss of consciousness if any higher demand of O2 is needed (e.g., exercise).
January 2010 4-14
-------
100
50% Anemia
(02Hb Capacity = lOmL/IOOmL)
20
40 60
PO2 (mmHg)
80
100
Source: U.S. EPA (1991, 017643)
Figure 4-7. 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
Mb is a globular heme protein that facilitates O2 diffusion from the muscle sarcoplasm to
mitochondria, acting as an O2 supply buffer to maintain adequate pO2 for mitochondria when the O2
supply changes, as in exercise. O2 has a greater affinity for Mb than Hb, which allows small changes
in tissue pO2 to release large amounts of O2 from O2Mb (Wittenberg et al, 1975, 012436). Small
reductions in O2 storage capacity of Mb, due to CO binding, may have a profound effect on the
supply of O2 to the tissue.
Like Hb, Mb will undergo reversible CO binding, however the affinity constant is
approximately eight-times lower than Hb (M = 20-40 versus 218, respectively) (Haab, 1990,
013359). The association rate constant of CO and Mb is approximately 27 times lower than O2;
however, the dissociation rate constant is approximately 630 times lower than O2 (Gibson et al.,
1986, 016289). causing CO to be retained and possibly stored in the muscle. CO levels have been
measured in human muscle and heart tissues with <2% COHb concentrations at background levels
January 2010
4-15
-------
(15 and 31 picomole [pmol] CO/mg ww, respectively) (Vreman et al., 2006, 098272) (Table 4-2).
Under conditions of CO asphyxiation, tissue concentrations increased 17-18 fold (265 and 527 pmol
CO/mg ww muscle and heart tissue, respectively); however, heart tissue concentrations varied
widely between individuals. Mouse muscle did not show this increase after exogenous CO exposure
(Vreman et al., 2005, 193786). This may be due to the fact that human muscle has a 15-fold higher
concentration of myoglobin protein than mouse muscle (Weller et al., 1986, 187298). The capacity
for diffusion of CO into the muscle is represented by the coefficient DmCO and is generally larger in
males than in females, likely due to the differences in muscle mass and capillary density (Bruce and
Bruce, 2003, 193975). COMb concentrations in the heart and skeletal muscle increase with work
load, due to a higher relative rate of CO binding to Mb relative to Hb. This causes an increase in
COMb/COHb that is not seen at rest (Sokal et al., 1984, 011591). Subjects with 2% COHb but not
those with 20% COHb levels showed a significant uptake of CO from the blood to the muscle with
increasing work intensity of the quadriceps muscle (Richardson et al., 2002, 037513).
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]
% COHb
1 .5 ± 1 .2
3.8 ±3.2
[2.5]
40.7 ± 28.8
[27.1]
56.4 ±28.9
[37.6]
Source: Reprinted with Permission of Wiley-Blackwell from Vreman et al. (2006, 0982721
4.3.2.4. Other Tissues
CO binds with other hemoproteins, such as cytochrome P450, cytochrome c oxidase, catalase,
and peroxidase, but the possibility of this binding influencing CO-O2 kinetics has not been
established. CO transfers between COHb and tissue, the extent of which varies between organs.
Blood-to-tissue flux causes less CO to be expired following CO exposure than what is lost from the
blood in terms of COHb (Roughton and Root, 1945, 180418). This value is estimated to be 0.3-0.4%
min"1 or 0.24 mL/min (Bruce and Bruce, 2003, 193975: Prommer and Schmidt, 2007, 180421). The
equilibration rate from blood to tissue is uncertain. Newly modeled CO trafficking kinetics shows
that CO continues to be taken up by the muscle and extravascular tissues well beyond the end of
exposure because of a less than instant equilibration (Bruce and Bruce, 2006, 193980). Table 4-2
contains tissue CO concentrations from humans under different CO exposure conditions. The
distribution of CO between the different human organs was shown to follow the same pattern versus
percent of the blood CO concentration, irrespective of the level of blood CO (Vreman et al., 2006,
098272). Consistently, the spleen, lung, and kidney had the highest measured CO concentration and
the most dramatic increases over basal levels; the brain and adipose had the lowest CO
concentrations. In addition to these fatty tissues, the muscular tissues including the heart and skeletal
muscle had similarly low increases over background CO levels. This pattern was also found in
rodents exposed to exogenous CO; however, increased endogenous CO produced after heme
administration did not follow this pattern of uptake (Vreman et al., 2005, 193786). Increased
endogenous CO production led to moderately increased CO present in the lung, heart, liver, and
spleen and no change in CO concentration in the testes, intestine, muscle, brain, and kidney. The
spleen and liver have an abundance of HO-1 expression and are involved in the catabolism of heme,
thus it is expected to have elevated CO concentrations in these organs after heme treatment. Also,
elevated CO in the lung is not surprising since it is the site of CO excretion. The tissues analyzed in
these studies were blanched before analysis; however, contamination of the tissue sonicates with
blood from the vessels within each organ is a possible source of error. The measurements were
presented by the authors as minimum tissue CO concentrations, due to the possibility of rapid loss of
CO from blood and tissue exposed to the atmosphere, light, and elevated temperature (Chace et al.,
1986, 012020; Ocak et al., 1985, 011641). These results are not consistent with older papers,
January 2010
4-16
-------
suggesting that negligible retention of CO occurs in the liver or brain (Sokal et al., 1984, 011591;
Topping, 1975. 193784).
4.3.3. Pulmonary and Tissue Elimination
Blood COHb concentrations are generally considered to have a monotonically decreasing,
second-order (logarithmic or exponential) elimination rate from equilibrium. However, more recent
reports have presented evidence for a biphasic washout curve, especially after brief CO exposure
(Figure 4-8) (Bruce and Bruce, 2006, 193980; Shimazu et al., 2000, 016420; Wagner et al., 1975,
010989). This event is modeled by a two-compartment system where the initial rapid decrease is the
washout rate from the blood, followed by a slower phase due to CO flux from the muscle and
extravascular compartments back to the blood. Tissue elimination rates have been reported as slower
than those for blood (Landaw, 1973, 010803). The biphasic curve is more obvious after short-
duration CO exposure (<1 h), whereas longer CO exposure (> 5 h) results in a virtually
monoexponential elimination, which could account for the historical findings. However, this
elimination curve also follows a biphasic curve with a slightly higher rate of elimination initially
(Shimazu et al., 2000, 016420). Differences in elimination kinetics could also be a result of the
variation in CO exposure duration (Weaver et al., 2000, 016421).
The elimination of COHb is affected by a number of factors, including duration of exposure,
PaO2, minute ventilation, the time post-exposure for analysis due to extravascular stores, as well as
inter-individual variability (Bruce and Bruce, 2006, 193980; Landaw, 1973, 010803; Shimazu, 2001,
016331). The elimination rate does not seem to be dependent upon the CO exposure source
(e.g., fire, non-fire CO exposure) (Levasseur et al., 1996, 080895). In addition, in a series of
poisoning cases, the COHb elimination half-life was not influenced by gender, age, smoke
inhalation, history of loss of consciousness, concurrent tobacco smoking, degree of initial metabolic
acidosis (base excess), or the initial COHb level (Weaver et al., 2000, 016421). On the contrary, in
modeling the nonlinear kinetics of CO, a subject with a higher initial COHb will detoxify and
eliminate CO more rapidly (Gosselin et al., 2009, 190946). Similarly, it has been shown that the
absolute elimination rates are associated positively with the initial concentration of COHb, however
the relative rate of elimination, expressed as a percentage decline in COHb% after a measured time,
is independent of the initial COHb concentration (Wagner et al., 1975, 010989). COHb elimination
half-life falls as the fractional inspired O2 concentration increases. While breathing air at sea level
pressure, the expected half-life in adult males is approximately 285 min, but may be shorter in adult
females. With inhalation of normobaric 40% O2, the half-life falls to 75 min and further to 21 min
when breathing 100% O2 because of greater competition for Hb by O2 (Landaw, 1973, 010803).
Another study reports the half-life falls to 74 min (mean) after breathing 100% O2, although the
range in this particular study was 26-148 min (Weaver et al., 2000, 016421). In addition, COHb half-
life will fall further after normocapnic hyperoxic hyperpnea (i.e., hyperventilation while maintaining
normal CO2 pressure in high O2) (Takeuchi et al., 2000, 005675).
January 2010 4-17
-------
100
.a
I
O
0
30 60 90 120
Time (min)
150
180
Source: Adapted with Permission of Lippincott Williams & Wilkins from Shimazu et al. (2000, 016420)
Figure 4-8. Changes in blood COHb after exposure to CO for a few minutes (A) or several
hours (B), representing the biphasic nature of CO elimination. Note: y-axis is log-
scale.
4.3.4. COHb Analysis Methods
Blood COHb saturation can be analyzed using numerous methods with various benefits and
limitations. The most popular current techniques include gas chromatography (GC) and
spectrophotometry, specifically using CO-oximeters. CO-oximeters are commonly used because they
require little sample preparation and simultaneously measure COHb, O2Hb, methemoglobin, and
total hemoglobin concentration. However, at low concentrations of COHb relevant to ambient
exposure (<5%), CO-oximeters overestimate COHb levels determined by GC (Mahoney et al., 1993,
013859: Widdop, 2002, 030493). Conversely, at higher COHb levels (>5%), CO-oximeters will
underestimate COHb concentrations. In addition to the inaccuracy of the CO-oximeters, some
studies report considerable imprecision in the results. Also, numerous substances or conditions can
interfere with CO-oximeter measurements (i.e., temperature, bilirubin, fetal hemoglobin).
Alternatively, GC is an accurate, precise, highly specific analysis method and is generally used as the
reference method for COHb analysis. GC requires the CO incorporated into blood or tissue samples
to first be released using a liberating agent such as potassium ferricyanide or sulfosalicylic acid
(Vreman et al., 2005, 193786; Vreman et al., 2006, 098272). and then measured directly or
indirectly. This methodology is more complex and time-consuming than spectrophotometry. In either
analysis method, it is important to remember that COHb measured at one site in the body does not
necessarily represent whole body CO distribution.
CO can also be measured directly in air or breath samples by using an electrochemical sensor
that depends on the electrical signal generated by the oxidation of CO. There are conflicting reports
on the correlation of exhaled CO (COex) with COHb. Multiple reports present positive correlation
coefficients (r) ranging from 0.92 and 0.98 in smoking subjects (Jarvis et al., 1980, 011813; Jarvis et
al., 1986, 012043; Landaw, 1973, 010803). Positive linear correlations have also been shown in
diseased patients with increased COHb (De las Heras et al., 2003, 194087). Others have reported no
correlation between low level COHb and COex and have suggested less correlation exists at the
lower levels of COex relevant to ambient exposures (Horvath et al., 1998, 087191; Scharte et al.,
2000, 194112). Finally, CO is endogenously produced in the nose and paranasal sinus which may
contribute to COex concentrations (Andersson et al., 2000, 011836).
January 2010
4-18
-------
4.4. Conditions Affecting Uptake and Elimination
4.4.1. Physical Activity
Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase
in gas exchange. O2 consumption can increase more than 10 fold during exercise. Similarly,
ventilation, membrane and lung diffusing capacity, pulmonary capillary blood volume, and cardiac
output increase proportional to work load. Also, exercise will improve the ventilation/perfusion ratio
in the lung and mobilize RBC reserves from the spleen. The majority of these changes facilitate CO
uptake and transport, by increasing gas exchange efficiency. Likewise, the COHb elimination rate
increases with physical activity, causing a decrease in COHb half-life (Joumard et al., 1981, 011330).
During a transition period from rest to exercise while exposed to CO (500 ppm/10 min), the
diffusing capacity and CO uptake were reported to rise faster than O2 consumption for each exercise
intensity (Kinker et al., 1992, 086328). The two physiological variables that are most influential in
the formation of COHb are alveolar ventilation and cardiac output. However, exercise did not affect
the ability of the CFK equation to predict COHb saturation as long as appropriate variables were
used for model analysis (Tikuisis et al., 1992, 013592).
4.4.2. Altitude
Increased altitude changes a number of factors that contribute to the uptake and elimination of
CO. The relationship between altitude and CO exposure has been discussed in depth in the 2000 CO
AQCD and other documents (U.S. EPA, 1978, 086321). In an effort to maintain proper O2 transport
and supply, physiological changes occur as compensatory mechanisms to combat the decreased
barometric pressure and resulting altitude induced hypobaric hypoxia (HH). HH, unlike CO hypoxia,
causes humans to hyperventilate, which reduces arterial blood CO2 (hypocapnia) and increases
alveolar partial pressure of O2. Hypocapnia will lead to difficulty of O2 dissociation and decreased
blood flow, thus reducing tissue O2 supply. HH increases blood pressure (BP) and cardiac output and
leads to redistribution of blood from skin to organs and from blood vessels to extravascular
compartments. Generally these changes will favor increased CO uptake and COHb formation, as
well as CO elimination. In hypoxic conditions both CO and O2 bind reduced Hb through a
competitive-parallel reaction (Chakraborty et al., 2004, 193759). Sea level residents exposed to high
altitude (3,658-5,800 m) for short or long visits (<1 year) experience negligible or minor changes in
DLCO, although these changes in DLCO can be accounted for by polycythemia or increased red
blood cell count and by the increased rate of reaction of carbon monoxide with hemoglobin due to
hypoxia (West, 1962, 199513)(Guleria et al., 1971, 199518). Breathing CO (9 ppm) at rest at altitude
produced higher COHb compared to sea level (McGrath et al., 1993, 013865), whereas high altitude
exposure with exercise caused a decrease in COHb levels versus similar exposure at sea level
(Horvath et al., 1988, 012725). This decrease could be a shift in CO storage or suppression of COHb
formation, or both. Altitude also increases the baseline COHb levels by inducing endogenous CO
production. Initial HH increased lung HO-1 protein and activity, whereas chronic HH induced
endogenous CO production in nonpulmonary sites (see Section 4.5) (Carraway et al., 2000, 021096).
As the length of stay increases at high altitude, acclimatization occurs, inducing
hyperventilation, polycythemia, and increased tissue capillarity and Mb content in skeletal muscle,
which could also favor increased CO uptake. The DLCO of sea level natives who are long-term
residents at altitude (3,100 m) increases from sea level values (Cerny et al., 1973, 199736).
Additionally, natives of high altitude (3,100-3,658 m) have increased DLCO compared to natives of
sea level or sea level natives that stay at high altitude (DeGraff et al., 1970, 199737) (Guleria et al.,
1971, 199518). This has been attributed to high pulmonary capillary blood volume and membrane
diffusing capacity, and altered lung structure. Most of the early adaptive changes gradually revert to
sea level values after individuals return to sea level. However, differences in people raised at high
altitude persist even after reacclimatization to sea level (Hsia, 2002, 193857). For example, altitude
natives (3,658 m) staying at sea-level still have increased DLCO compared to sea level natives which
suggests a permanent change in the lung structure resulting in a larger diffusing surface area (Guleria
etal, 1971. 199518).
January 2010 4-19
-------
4.4.3. Physical Characteristics
Certain physical characteristics (e.g., age, sex, pregnancy) can alter the variables that influence
the uptake, distribution, and elimination of CO. Values of CO uptake and elimination change with
age. Young children eliminate COHb more rapidly than adults after CO exposure (Joumard et al.,
1981, 011330: Klasner et al., 1998, 087196). After infancy, the COHb half-life increases with age,
nearly doubling between 2 and 70 yr (Joumard et al., 1981, 011330). The rate of this increase in CO
elimination is very rapid in the growing years (2-16 yr of age), but slows beyond adolescence.
Alveolar volume and DLCO increase with increasing body length of infants and toddlers (Castillo et
al., 2006, 193234). suggesting a further degree of lung development and faster CO uptake. After
infancy, increasing age decreases DLCO and increases VA/Q mismatch, causing it to take longer to
both load and eliminate CO from the blood (Neas and Schwartz, 1996, 079363).
COHb concentrations are generally lower in female subjects than in male subjects (Horvath et
al., 1988, 012725). and the COHb half-life may be longer in healthy men than in women of the same
age, which may be partially explained by differences in muscle mass or the slight correlation
between COHb half-life and increased height (Joumard et al., 1981, 011330). However, women do
have a higher rate of endogenous production while in the progesterone phase of the menstrual cycle
and during pregnancy (Section 4.5). The rate of decline of DLCO with age is lower in middle-aged
women than in men; however, it evens out towards older age (Neas and Schwartz, 1996, 079363).
Women also tended to be more resistant to altitude hypoxia (Horvath et al., 1988, 012725).
Ethnicity does alter physiological variables that determine CO uptake and kinetics. Lung
volumes are 10-15% less in both Asian and African-American populations when compared to
Caucasians. This causes a reduced alveolar surface area (20% less than estimated values) for gas
exchange, leading to a 13% difference in DLCO (Pesola et al., 2004, 193842: Pesola et al., 2006,
193855). Certain factors, such as socioeconomic status (SES), were not controlled for in these
studies. SES has been shown to affect pulmonary function, including decreasing DLCO (Hegewald
and Crapo, 2007, 193923).
4.4.3.1. Fetal Pharmacokinetics
Inhaled CO by pregnant animals quickly passes the placental barriers and enters the fetal
circulation (Longo, 1977, 012599). Fetal CO pharmacokinetics do not follow the same kinetics as
maternal CO exposure, making it difficult to estimate fetal COHb based on maternal levels. Fetal
COHb will vary as a function of maternal exposure but will also depend upon the rate of endogenous
fetal CO production (Section 4.5), placental diffusing capacity of CO, the relative affinity of fetal Hb
for CO compared to O2, and the affinity of fetal blood for O? (Longo, 1970, 013922). Human fetal
Hb has a higher affinity for CO than adult Hb, where the ratio of fetal COHb to maternal COHb at
steady state in humans is approximately 1.11 (Longo, 1970, 013922)(Di Cera et al., 1989,
193998)(Hayde et al., 2000, 201602). Maternal and fetal COHb concentrations have been modeled
as a function of time using a modified CFK equation (Hill et al., 1977, 011315). At steady-state
conditions, the fetal COHb is up to 10-15% higher than the maternal COHb levels. For example,
exposure to 30 ppm CO results in a maternal COHb of 5% and a fetal COHb of 5.75%. The fetal CO
uptake lags behind the maternal for the first few hours but later may overtake the maternal values
(Figure 4-9). Fetal COHb equilibrium may not be reached for 36-48 h after exposure. Similarly,
during washout, the fetal COHb levels are maintained for longer, with a half-life of around 7.5 h
versus the maternal half-life of around 4 h (Longo and Hill, 1977, 010802).
January 2010 4-20
-------
I-
o
(J
a
10
'—' C
8 5
MATERNAL
X--
,'* FETAL
12 18 24 30
TIME (HOURS)
36
42
48
Source: Reprinted with Permission of the American Physiological Society from Hill et al. (1977, 011315)
Figure 4-9. Predicted maternal and fetal COHb during periodic exposure to CO (50 ppm for
16 h followed by 0 ppm for 8 h ).
4.4.4. Health Status
Health status can influence the toxicity involved with CO exposure by influencing the severity
of hypoxia resulting from CO exposure. Any condition that would alter the blood O2 carrying
capacity or content will result in a greater risk from COHb induced hypoxia and decreased tissue O2
delivery. The severity of this effect depends upon the initial level of hypoxia.
Anemias are a group of diseases that result in insufficient blood O2 or hypoxia due to Hb
deficiency through hemolysis, hemorrhage, or reduced hematopoiesis. Anemia may result from
pathologic conditions characterized by chronic inflammation, such as malignant tumors or chronic
infections (Cavallin-Stahl et al., 1976, 086306: Cavallin-Stahl et al., 1976, 193239). The bodies of
people with anemia compensate, causing cardiac output to increase as both heart rate and stroke
volume increase. The endogenous production of CO, thus COHb, is increased in patients with
hemolytic anemia due to increased heme catabolism, causing an increased baseline COHb
concentration. One of the most prevalent anemias arises from a single-point mutation of Hb, causing
sickle cell diseases. The Hb affinity for O2 and O2 carrying capacity is reduced causing a shift to the
right in the O2 dissociation curve. It is well documented that African-American populations have a
higher incidence of sickle cell anemia, which may be a risk factor for CO hypoxia.
Chronic obstructive pulmonary disease (COPD) is often accompanied by a number of changes
in gas exchange, including increased deadspace volume (VD) and ventilation-perfusion ratio (VA/Q)
inequality (Marthan et al., 1985, 086334). which could slow both CO uptake and elimination.
Patients with pulmonary sarcoidosis, a restrictive lung disease, may also have a decrease in lung
volumes, a loss of DLCO, and gas exchange abnormalities during exercise, including decreased
arterial oxygen pressure (PaO2) and increased alveolar-arterial oxygen pressure difference (Lamberto
et al., 2004, 193845).
Individuals with heart disease may be at a greater risk from CO exposure since they may
already have compromised O2 delivery. Time to onset of angina was reduced after exposure to
January 2010
4-21
-------
100 ppm CO, compared to clean air (Kleinman et al., 1998, 047186). Hyperlipidemic patients may
have decreased CO diffusion capacity, a loss of VA/Q gradient, and a decrease in PaO2 (Enzi, 1976,
195794) (Section 5.2).
4.5. Endogenous CO Production and Metabolism
Humans breathing air containing no environmental sources of CO will still have a low
measurable level of circulating COHb due to endogenous CO production from heme protein
catabolism. In the normal degradation of RBC Hb, the porphyrin ring of heme is broken at the
a-methene bridge by HO. HO is co-localized with NADPH-flavoprotein reductase and biliverdin
reductase on the endoplasmic reticulum, where it catabolizes heme in an O2 and NADPH-dependent
manner to biliverdin, ferrous iron, and CO. Biliverdin is then further broken down by biliverdin
reductase into bilirubin, a powerful endogenous antioxidant. HO mediated metabolism functions as
the rate-limiting enzyme step in heme degradation and endogenous CO production (Wu and Wang,
2005, 180411). Three isoforms of HO exist, but HO-1 is the only inducible form (Maines and
Kappas, 1974, 193976: Maines et al., 1986, 193978: McCoubrey et al., 1997, 016715). Endogenous
CO production can be increased by the up-regulation of HO-1 expression and activity by inducers
such as oxidative stress, hypoxia, heavy metals, sodium arsenite, heme and heme derivatives, various
cytokines, and also exogenous CO (Wu and Wang, 2005, 180411).
The major site of heme catabolism, and thus the major organ of CO production, is the liver,
followed by the spleen, brain, and erythropoietic system (Berk et al., 1976, 012603). These rates of
CO formation may be due to higher levels of HO activity in these tissues. The whole body
production rate of CO is approximately 18.8 umol/h (0.42 mL/h or 0.007 mL/min) and produces
between 400-500 umol CO per day (Coburn et al., 1963, 013971: Coburn et al., 1964, 013956:
Coburn et al., 1966, 010984) (Figure 4-10). The endogenous rate of production varied somewhat
within individuals measured on multiple days (±4.5 umol/h and ±0.35% COHb) (Coburn et al.,
1966, 010984). However, these measurements of day-to-day CO production variability were
comparable to the equipment measurement error reported (±3.1 umol/h). The endogenous rate of CO
formation varies between different tissues, ranging from 0.029 nmol/mg protein/h in chorionic villi
of term human placentas to 0.28 nmol/mg protein/h in cultured rat olfactory receptor neurons and rat
liver perfusate (Marks et al., 2002, 030616). However, these estimations are uncertain since CO is
quickly scavenged in the cytosol of living cells. CO is endogenously produced in the nose and
paranasal sinus which may contribute to exhaled CO concentrations (Andersson et al., 2000,
011836). It is also important to note that increased endogenous CO production does not universally
lead to an increase in COHb saturation.
January 2010 4-22
-------
Study
Condition
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)
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 4-10.
Representative estimates of endogenous CO production rates resulting from
various conditions and diseases.
Not all endogenous CO production is derived from Hb breakdown. Other hemoproteins, such
as Mb, cytochromes, peroxidases, and catalase, contribute 20-25% to the total amount of endogenous
CO (Berk et al., 1976, 012603). All of these sources result in a normal blood COHb concentration
between 0.3 and 1% (Coburn et al., 1965, 011145). The level of endogenous production can be
changed by drugs or a number of physiological conditions that alter RBC destruction, other
hemoprotein breakdown, or HO-1 expression and activity (Figure 4-10). Nicotinic acid (Lundh et al.,
1975, 086332). allyl-containing compounds (acetamids and barbiturates) (Mercke et al., 1975,
086303). diphenylhydantoin (Coburn, 1970, 010625). progesterone (Delivoria-Papadopoulos et al.,
1974, 086316). contraceptives (Mercke et al., 1975, 086308). and statins (Muchova et al., 2007,
194098) can increase CO production. Compounds such as carbon disulfide and sulfur-containing
chemicals (parathion and phenylthiourea) will increase CO by acting on P450 system moieties
(Landaw et al., 1970, 012605). The P450 system may also cause large increases in CO produced
from the metabolic degradation of dihalomethanes, leading to very high (>10%) COHb levels (Bos
et al., 2006, 194084; Manno et al., 1992, 013707) that can be further enhanced by prior exposure to
hydrocarbons or ethanol (Pankow et al., 1991, 013551; Wirkner et al., 1997, 082642). Minor sources
of endogenous CO include auto-oxidation of phenols, flavonoids, and halomethanes, photo-oxidation
of organic compounds, and lipid peroxidation of cell membrane lipids (Rodgers et al., 1994,
076440).
Women experience fluctuating COHb levels throughout menstruation when endogenous CO
production doubles in the progesterone phase (0.62 mL/h versus 0.32 mL/h in estrogen phase)
(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 accelerated destruction of hemoproteins will lead to an
increased production of CO (Figure 4-11 and Figure 4-12). 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 two- to threefold higher than healthy
individuals and CO production rates two- to eightfold higher (Coburn et al., 1966, 010984). Recent
studies report COHb levels measured by CO-oximeter that are 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).
January 2010
4-23
-------
Endogenous CO production rate varied from 0.70 to 3.18 mL/h in anemic patients (Coburn et al.,
1966, 010984).
Study
Condition
Hampson and Wfeaver (2007, 190272)
Hampson and Wfeaver (2007, 190272)
Zinkhametal. (1980, 011435)
Zinkhametal. (1980, 011435)
Coburn etal. (1966, 010984)
Coburn etal. (1966, 0109841
Iran et al. (2007, 194120)
De Las Heras et al. (2003, 194087)
De Las Heras et al. (2003, 1940871
Yasuda et al. (2005, 1919531
Morimatsu et al. (2006, 194097)
De Las Heras et al. (2003, 194087)
I Hernolytic anemia rnin
i Hh Zurich rnin
1 Hh Zurich min
i H^mnlytic qnprriR rrw
1 Hemolytic anemia max
I Hemolytic anemia min
1 1 iver transplant
1 Peritonitis (SBP)
1 Cirrhosis
1 Exacerbated COPD
i Critically ill
• Healthy
02468
COHb (%)
10
Figure 4-11. Representative COHb saturation resulting from various diseases and conditions.
Measurements of COHb taken using CO-oximeter, except in Coburn et al. (1966,
010984), where COHb was measured using GC. 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 4-12). CO production weakly correlates with the multiple organ
dysfunction score (MODS), which estimates severity of organ dysfunction; however, it did not
correlate with the 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 ul/kg/h) (Zegdi et al., 2002, 037461).
January 2010
4-24
-------
Study
Condition
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, 084291)
Zavasu etal. (1997, 084291)
Pared! etal. (1999, 118798)
Horvath et al. (1998, 0871901
Pared! etal. (1999, 118798)
Pared! etal. (1999, 118798)
Morimatsu et al. (2006, 194097)
Morimatsu et al. (2006, 1940971
1 Peritonitis (SBP)
1 Cirrhosis
1 Sjp|<-|e re|| Anemja
1 IIRTI
i Asthma w/ steroids
1 Asthma
1 Cystic fibrosis
I Bronchiectasis
1 Type 2 diabetes
i Type 1 diabetes
1 Critically ill
H Healthy
3 6 9 12
Exhaled CO (average fold change from healthy)
15
Figure 4-12. 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 result in increased endogenous CO production.
For example, patients with severe sepsis or septic shock have a higher COex and endogenous CO
production compared to control patients, which was reduced 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 measured by CO-oximeter 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 the Child Turcotte Pugh score, used to assess the degree of liver
impairment (Tran et al., 2007, 194120). Further investigation in cirrhotic patients, with and without
ascites, provided evidence for increased plasma CO concentrations, HO-1 activity in
polymorphonuclear cells, exhaled CO, and blood COHb (De las Heras et al., 2003, 194087; Tarquini
et al., 2009, 194117). COex, plasma CO, and COHb levels were correlated with the Child-Pugh
score, and thus the severity of disease. These parameters were significantly higher in patients with
ascites or with spontaneous bacterial peritonitis (SBP) (COHb, healthy: 0.6 ± 0.1%; cirrhosis: 1.0 ±
0.1%; with ascites: 1.6 ± 0.2%; with SBP: 1.9 ± 0.2%; measured by CO-oximeter). Both COex and
COHb levels decreased after resolution of the infection in patients with SBP, reaching values similar
to noninfected patients within 1 mo (De las Heras et al., 2003, 194087). Endotoxin concentration
was correlated with plasma CO levels, suggesting a link between systemic endotoxemia and
increased activity or expression of the HO/CO system (Tarquini et al., 2009, 194117). COex
concentrations are also elevated in patients with diabetes (Type 1: 4.0 ± 0.7 ppm; Type 2: 5.0 ±
0.4 ppm; healthy: 2.9 ± 0.2 ppm), and correlated with blood glucose levels and duration of disease
(Paredi et al., 1999, 194102). Likewise, obese Zucker rats, a model of metabolic syndrome with
insulin resistance, have increased respiratory CO excretion and COHb levels compared to lean
Zucker rats (3.9 ±0.1% versus 3.0 ± 0.1% COHb), which is decreased by HO inhibition (Johnson et
al., 2006, 1938741
Endogenous CO is also increased in airway inflammatory diseases. Patients with upper
respiratory tract infections exhaled higher CO concentrations than normal controls and this increase
was attenuated after recovery (Yamaya et al., 1998, 047525). Arterial COHb levels have been related
January 2010
4-25
-------
to disease severity in COPD patients (Yasuda et al., 2005, 191953). Bronchiectasis patients had
higher COex; however, anti-inflammatory treatment did not decrease the CO levels (Horvath et al.,
1998, 087191). Patients with cystic fibrosis had higher COex than normal controls (6.7 ±0.6 ppm
versus 2.4 ± 0.4 ppm), and patients treated with steroids had a decrease in CO levels (8.4 ±1.0 ppm
versus 5.1 ± 0.5 ppm) (Paredi et al., 1999, 118798). Increased arterial COHb measured by CO-
oximeter was reported in patients with bronchial asthma, pneumonia, idiopathic pulmonary fibrosis,
pyelonephritis, and active rheumatoid arthritis (Yasuda et al., 2002, 035206; Yasuda et al., 2004,
191955). Similarly, asthmatic patients exhibited an elevation of COex that decreased with
corticosteroid therapy (nonsmoking controls: 1.5 ± 0.1 ppm; asthmatics without corticosteroids: 5.6
± 0.6 ppm; with corticosteroids: 1.7 ±0.1 ppm; smoking controls: 21.6 ± 2.8 ppm) (Zayasu et al.,
1997, 084291). These results were confirmed and associated with increased expression of HO-1 in
airway macrophages (Horvath et al., 1998, 087190). Also, COex was increased in patients with
allergic rhinitis during the pollen season; however, their COex was similar to control subject levels
out of season (Monma et al., 1999, 180426). Similarly, endogenous CO production and HO-1
expression in nasal mucosa was correlated with allergic rhinitis in guinea pigs as described in
Section 5.1 (Shaoqing et al., 2008, 192384).
Altitude is also positively associated with baseline COHb concentrations (McGrath, 1992,
001005)(McGrath et al., 1993, 013865). This increase in COHb with altitude-induced hypoxia is
associated in rats and cells with increases in the mRNA, protein, and activity of HO-1 leading to
enhanced endogenous CO production (Carraway et al., 2002, 026018; Lee et al., 1997, 082641).
Whether other variables such as an accelerated metabolism or a greater pool of Hb, transient shifts in
body stores, or a change in the elimination rate of CO, play a role has not been explored.
Because of the sensitivity of COHb to changes in the metabolic state, ranges of endogenous
COHb levels in the population are uncertain. However, baseline levels of COHb, which reflect
exposure to ambient and non-ambient CO and endogenous production of CO, have been measured in
the population. COHb levels measured by CO-oximeter in packed red blood cell units reserved for
use between 2004 and 2005 averaged 0.78 ± 1.48%, with 10.3% of samples having COHb levels of
1.5% or greater and a maximum measurement of 12% (Ehlers et al., 2009, 194089). This study
reported a decrease from a study conducted in 1982-1983 in the number of units with elevated
COHb; at that time, 49% of units had COHb levels >1.5% (Aronow et al., 1984, 194083) versus
10.3% in 2004-2005. Another study calculated that 23% of donated blood units had COHb levels
exceeding 1.5%, with the highest measurement being 7.2% (Aberg et al., 2009, 194082). Smoking is
the main factor causing increased blood concentrations of CO. A dose-response relationship was
shown to exist between COHb concentration and the number of cigarettes smoked a day
(nonsmoker: 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%; > 25 cig/day: 6.02 ± 2.86% COHb). The mean baseline COHb value for former
smokers was higher than that of never smokers in this prospective cohort study (1.96 ± 1.87 versus
1.59 ± 1.72%) (Hart et al., 2006, 194092).
Endogenous CO is removed from the body mainly by expiration and oxidation. CO diffuses
across the alveolar-capillary membrane and is exhaled. This event has been used as a noninvasive
measurement of both endogenous and body load CO (Stevenson et al., 1979, 193767). CO can also
be oxidized to CO2 by cytochrome c oxidase in the mitochondria (Fenn, 1970, 010821; Young and
Caughey, 1986, 012091). However, the rates of CO metabolism are much slower than the rates of
endogenous CO production, with the rate of consumption representing only 10% of the rate of CO
production in dogs (Luomanmaki and Coburn, 1969, 012319).
4.6. Summary and Conclusions
CO elicits various health effects by binding with and altering the function of a number of
heme-containing molecules, mainly Hb. The formation of COHb reduces the O2-carrying capacity of
blood and impairs the release of O2 from O2Hb to the tissues. Venous COHb levels have been
modeled mainly by the CFK equation, but more recent models have included venous and arterial
blood mixing and Mb and extravascular storage compartments, as well as other dynamics of CO
physiology. The CFK equation remains the most extensively validated and applied model for COHb
prediction. Recent models have indicated that CO has a biphasic elimination curve, due to initial
washout from the blood followed by a slower flux from the tissues. The flow of CO between the
January 2010 4-26
-------
blood and alveolar air or tissues is controlled by diffusion down the pCO gradient. The uptake of CO
is governed not only by this CO pressure differential, but also by physiological factors, such as
minute ventilation and lung diffusing capacity, that can, in turn, be affected by conditions such as
exercise, age, and health. Susceptible populations, including health compromised individuals and
developing fetuses, are at a greater risk from COHb induced health effects due to altered CO
kinetics, compromised cardiopulmonary function, and increased baseline hypoxia levels. Altitude
may also significantly affect the kinetics of COHb formation. Compensatory mechanisms, such as
increased cardiac output, compensate for the decrease in barometric pressure. Altitude also increases
the endogenous production of CO through upregulation of HO-1. CO is considered a second
messenger and is endogenously produced from the catabolism of heme proteins by enzymes such as
HO-1. A number of diseases and conditions affect endogenous CO production, possibly causing a
higher endogenous COHb level. Finally, CO is removed from the body by expiration or oxidation to
CO2.
January 2010 4-27
-------
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
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
Alcantara RE; Xu C; Spiro TG; Guallar V (2007). A quantum-chemical picture of hemoglobin affinity. PNAS, 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
BerkPD; 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
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)
January 2010 4-28
-------
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
Cerny FC; Dempsey JA; Reddan WG (1973). Pulmonary gas exchange in normative residents of high altitude. J Clin
Invest, 52: 2993-2999. 199736
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. Acta Med Scand Suppl, 472: 269-282.
011144
Coburn RF (1970). Enhancement by phenobarbital and diphenylhydantoin of carbon monoxide production in normal man.
N Engl J Med, 283: 512-515. 010625
Coburn RF (1970). The carbon monoxide body stores. Ann N Y Acad 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 HJ; Sarkar M (2008). Population pharmacokinetic analysis of carboxyhaemoglobin
concentrations in adult cigarette smokers. Br J Clin Pharmacol, 65: 30-39. 194085
DeGraff AC Jr; Grover RF; Johnson RL Jr; Hammond JW Jr; Miller JM (1970). Diffusing capacity of the lung in
Caucasians native to 3,100 m. J Appl Physiol, 29: 71-76. 199737
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
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; Collison HA (1969). Relative affinity of human fetal hemoglobin for carbon monoxide
and oxygen. Blood, 33: 37-45. 193914
January 2010 4-29
-------
Enzi G, Bevilacqua M, Crepaldi G (1976). Disturbances in pulmonary gaseous exchange in primary hyperlipoproteinemias.
Bull Europ Physiol Resp, 12: 433-442. 195794
FennWO(1970). The burning of CO in tissues. Ann NY Acad 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
ForsterRE (1966). Diffusion of gases in the lungs. Physiologist, 9: 110-122. 180430
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, 21: 1077-1091. 190946
Guleria JS; Pande JN; Sethi PK; Roy SB (1971). Pulmonary diffusing capacity at high altitude. J Appl Physiol, 31:536-
543. 199518
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
Haab P(1990). The effect of carbon monoxide on respiration. Experientia, 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: 163-168. 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
Hayde M; Pollak A; Bernaschek G; Weiner CP; Vreman HJ; Stevenson DK; Widness JA (2000). Association of fetal and
maternal carboxyhemoglobin levels in normal and Rh-alloimmune pregnancies. Early Hum Dev, 58: 205-212.
201602
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; Bedi JF; Wagner JA; Agnew JW (1988). Maximal aerobic capacity at several ambient concentrations of
carbon monoxide at several altitudes. J Appl Physiol, 65: 2696-2708. 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
January 2010 4-30
-------
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
Joumard R; 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
Kinker JR; Haffor A-S; Stephan M; Clanton TL (1992). Kinetics of CO uptake and diffusing capacity in transition from rest
to steady-state exercise. J Appl Physiol, 72: 1764-1772. 086328
Klasner AE; Smith SR; Thompson MW; Scalzo AJ (1998). Carbon monoxide mass exposure in a pediatric population.
Acad Emerg Med, 5: 992-996. 087196
Kleinman MT; Leaf DA; Kelly E; Caiozzo V; Osann K; 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
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; lyerNV; Alam J; Semenza GL; Choi AMK (1997). Hypoxia-inducible factor-1 mediates
transcriptional activation of the heme oxygenase-1 gene in response to hypoxia. JBiol Chem, 272: 5375-5381.
082641
Levasseur L; 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
Longo LD (1970). Carbon monoxide in the pregnant mother and fetus and its exchange across the placenta. Ann N Y Acad
Sci, 174: 313-341.013922
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
Lundh B; 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; VremanHJ; Stevenson DK; VanKessel 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. PNAS, 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
January 2010 4-31
-------
Marks GS; Vreman HJ; McLaughlin BE; Brien IF; 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
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. Eur Respir 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.
001005
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). Carbon monoxide production and reticulocyte count in normal women:
effect of contraceptive drugs and smoking. Acta Med Scand, 198: 155-160. 086308
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; 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. J_94097
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
NCHS; Radford EP; Drizd TA (1982). Blood carbon monoxide levels in persons 3-74 years of age: United States, 1976-80;
Advance Data from Vital and Health Statistics, No 76. National Center for Health Statistics; Office of Health
Research, Statistics, and Technology; Public Health Service; U.S. Department of Health and Human Services.
Hyattsville, MD. DHHS Pub. No. (PHS) 82-1250. 011442
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 Soc Mech Sci & Eng, 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 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
January 2010 4-32
-------
Pankow D; Matschiner F; Weigmann H-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
Penney DG (1988). A review: Hemodynamic response to carbon monoxide. Environ Health Perspect, 77: 121-130. 012519
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 (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
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]. California
Environmental Protection Agency, Air Resources Board; South Coast Air Quality Management District.
Sacramento, CA.http://www.arb.ca.gov/research/abstracts/95-339.htm. 010611
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 LM; Dixon J; Pierce RJ; 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
January 2010 4-33
-------
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
Shaoqing Y; Ruxin Z; Yinjian C; Jianqiu C; Chunsheng Z; Jiangfeng T; Genhong L (2008). Possible contribution of
endogenous carbon monoxide to the development of allergic rhinitis in guinea pigs. J Inflamm, 5: 23. 192384
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
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 Grit 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. In Penney DG (Ed.),Carbon Monoxide (pp.
45-67). Boca Raton, FL: CRC Press, Inc / Lewis Publishers. 080960
January 2010 4-34
-------
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 B J; 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 TT; Martin P; Ly H; Balfe D; Mosenifar Z (2007). Carboxyhemoglobin and its correlation to disease severity in
cirrhotics. J Clin Gastroenterol, 41: 211-215. 194120
Tyuma I; Ueda Y; Imaizumi K; Kosaka H (1981). Prediction of the carbonmonoxyhemoglobin levels during and after
carbon monoxide exposures in various animal species. JpnJ Physiol, 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. U.S. Envrionmental Protection Agency. Research Triangle Park, NC.
EPA-600/9-78-015. 086321
U. S. EPA (1991). Air quality criteria for carbon monoxide. U. S. Environmental Protection Agency. Research Triangle Park,
NC. EPA/600/8-90/045F. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000554R.txt. 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
West JB (1962). Diffusing capacity of the lung for carbon monoxide at high altitude. J Appl Physiol, 17: 421-426. 199513
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
Wu L; 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
January 2010 4-35
-------
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
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 Grit 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
January 2010 4-36
-------
Chapter 5. Integrated Health Effects
5.1. Mode of Action of CO Toxicity
5.1.1. Introduction
The diverse effects of CO are dependent upon concentration and duration of exposure as well
as on the cell types and tissues involved. Responses to CO are not necessarily due to a single process
and may instead be mediated by a combination of effects including COHb-mediated hypoxic stress
and other mechanisms such as free radical production and the initiation of cell signaling. However,
binding of CO to reduced iron in heme proteins with subsequent alteration of heme protein function
is the common mechanism underlying the biological responses to CO.
5.1.2. Hypoxic Mechanisms
As discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907). the most well-known
pathophysiologic effect of CO is tissue hypoxia caused by binding of CO to Hb. Not only does the
formation of COHb reduce the O2-carrying capacity of blood, but it also impairs the release of O2
from O2Hb. Compensatory alterations in hemodynamics, such as vasodilation and increased cardiac
output, protect against tissue hypoxia. Depending on the extent of CO exposure, these compensatory
changes may be effective in people with a healthy cardiovascular system. However, hemodynamic
responses following CO exposure may be insufficient in people with decrements in cardiovascular
function, resulting in health effects as described in Section 5.2.
The 2000 CO AQCD (U.S. EPA, 2000, 000907) reported changes in vasodilation due to CO
levels between 500-2,000 ppm (Kanten et al., 1983, 011333: MacMillan, 1975, 012909). In one
study, the vasodilatory response to CO in cerebral blood vessels was attributed to decreased O2
availability (Koehler et al., 1982, 011341). In another study, exposure of rats to 1,000 ppm CO
resulted in increased cerebral blood flow which was not triggered by tissue hypoxia since no changes
in intramitochondrial NADH levels preceded vasodilation (Meilin et al., 1996, 079919). However,
the response was blocked by the inhibition of NOS indicating a role for the free radical species NO
in CO-mediated vasodilation (Meilin et al., 1996, 079919).
Increased cardiac output was also discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907)
as a compensatory response to CO-mediated tissue hypoxia. Findings of studies which measured
hemodynamic alterations following CO exposure were equivocal and sometimes contradictory
(Penney, 1988, 012519). While most studies reported a positive correlation between COHb and
cardiac output at COHb levels above 20%, one study demonstrated increased cardiac output in
humans following acute exposure to 5% CO which resulted in the rapid rise in COHb levels to ~9%
(Ayres et al., 1973, 193943). However, there was no increase in cardiac output following a more
gradual increase in COHb levels to ~9% achieved by exposure to 0.1% CO over a longer period of
time (Ayres et al., 1973, 193943). Increased heart rate and stroke volume (SV) were observed in
response to CO exposure in one study (Stewart et al., 1973, 012428); however, some experiments
found no change in SV in humans with 18-20% COHb (Vogel and Gleser, 1972, 010898) or 12.5%
COHb (Klausen et al., 1968, 193936). The 2000 CO AQCD (U.S. EPA, 2000, 000907) reported that
blood pressure was generally unchanged in human CO exposure studies, while a number of animal
studies demonstrated CO-induced hypotension (Penney, 1988, 012519). No changes in forearm
blood flow, blood pressure, or heart rate were reported in humans with approximately 8% COHb
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).
January 2010 5-1
-------
(Hausberg and Somers, 1997, 083450). However, high-concentration exposures (3,000-10,000 ppm)
in animals resulted in diminished organ blood flow (Brown and Piantadosi, 1992, 013441). In-depth
discussion of hemodynamic changes resulting from CO exposure in recent human clinical studies
can be found in Section 5.2.4.
Binding of CO to Mb, as discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907) and in
Section 4.3.2.3, can also impair the delivery of O2 to tissues. Mb has a high affinity for CO, about
25 times that of O2; however, pathophysiologic effects are seen only after high-dose exposures to
CO, resulting in COMb concentrations far above baseline levels. High-energy phosphate production
in cardiac myocytes was inhibited when COMb concentrations exceeded 40%, corresponding to an
estimated COHb level between 20-40% (Wittenberg and Wittenberg, 1993, 013909). Conversely, rat
hearts perfused with solutions containing 6% CO (60,000 ppm) exhibited no change in high-energy
phosphate production, respiration rate, or contractile function (Chung et al., 2006, 193987; Glabe et
al., 1998. 086704).
5.1.3. Nonhypoxic Mechanisms
Nonhypoxic mechanisms underlying the biological effects of CO were discussed in the 2000
CO AQCD (U.S. EPA, 2000, 000907) and are summarized below. Most of these mechanisms are
related to CO's ability to bind heme-containing proteins other than Hb and Mb (Raub and Benignus,
2002, 041616). Since then, additional experiments have confirmed and extended these findings.
While the majority of the older studies utilized concentrations of CO far higher than ambient levels,
many of the newer studies have employed more environmentally-relevant concentrations of CO.
5.1.3.1. Nonhypoxic Mechanisms Reviewed in the 2000 CO AQCD
Inhibition of heme-containing proteins such as cytochrome c oxidase and cytochrome P450
reductases may alter cellular function. CO interacts with the ferrous heme a3 of the terminal enzyme
of the electron transport chain, cytochrome c oxidase (Petersen, 1977, 193764). Cytochrome c
oxidase inhibition not only interrupts cellular respiration and energy production but can also enhance
reactive oxygen species (ROS) production. In vivo studies observed CO binding to cytochrome c
oxidase under conditions where COHb concentrations were above 50% (Brown and Piantadosi,
1992, 013441). It is unlikely that this could arise under physiologic conditions or under conditions
relevant to ambient exposures.
A series of studies from the laboratory of Thorn, Ischiropoulos and colleagues indicated that
CO exposure produced a pro-oxidant cellular environment by liberation of NO. Exposure to CO
concentrations of 10-20 ppm and above caused isolated rat platelets, as well as cultured bovine
pulmonary endothelial cells, to release NO (Thorn and Ischiropoulos, 1997, 085644). This response
was blocked by treatment with an NOS inhibitor, indicating that the NO released was dependent on
NOS activity. An increase in available NO was also seen in the lung and brain of CO-exposed rats
(Ischiropoulos et al., 1996, 079491: Thorn et al., 1999, 016757). Reaction of NO with superoxide to
form the highly active oxidant species, peroxynitrite (Thorn et al., 1997, 084337). was thought to
lead to the activation and sequestration of leukocytes in brain vessels (Thorn et al., 2001, 193779)
and aorta (Thorn et al., 1999, 016753). oxidation of plasma lipoproteins (Thorn et al., 1999, 016753).
and the formation of protein nitrotyrosine (Ischiropoulos et al., 1996, 079491; Thorn et al., 1999,
016757; Thorn et al., 1999, 016753). NO release by CO was attributed to the displacement of NO
from nitrosyl-bound heme proteins. The rate of this event was slow; however, it occurred at
environmentally-relevant concentrations of CO (Thorn et al., 1997, 084337).
CO exposure also increased the production of other pro-oxidant species, including hydrogen
peroxide (H2O2) and hydroxyl radical (OH). High-level CO exposure (2,500 ppm) increased OH in
rat brain, and this response was distinct from tissue hypoxia (Piantadosi et al., 1997, 081326). The
mechanism for enhanced H2O2 production was unclear. The release of H2O2 in the lung of
CO-exposed rats was dependent upon the production of NO, as it was inhibited by pretreatment with
an NOS inhibitor (Thorn et al., 1999, 016757). It is possible that peroxynitrite formed after CO
exposure inhibited electron transport at complexes I through III, or that cytochrome c oxidase
inhibition led to mitochondrial dysfunction and ROS production.
Evidence was presented for CO-mediated vasorelaxation by three different mechanisms. First,
CO may inhibit the synthesis of vasoconstrictors by P450 heme proteins (Wang, 1998, 086074).
January 2010 5-2
-------
Vasodilation in isolated vessels was demonstrated via this P450-dependent mechanism using high
concentrations of CO (approximately 90,000 ppm) (Coceani et al., 1988, 040493). In the case of
cytochrome P450 enzymes, tissue CO levels may need to be abnormally high to elicit a response
since the Warburg binding coefficients (the ratio of CO to O2 at which half the reactive sites are
occupied by CO) for cytochrome P450s range from 0.1-12 (Piantadosi, 2002, 037463). P450
inhibition may reduce the hypoxia-induced expression of mitogens such as erythropoietin (EPO),
vascular endothelial growth factor (VEGF), endothelin-1 (ET-1), and platelet derived growth factor
(PDGF), which may decrease smooth muscle proliferation in response to hypoxia (Wang, 1998,
086074). CO also interfered with the metabolism of barbiturates and other drugs; however, this was
probably due to the hypoxic actions of CO rather than to P450 inhibition (Roth and Rubin, 1976,
012703: Roth and Rubin, 1976, 012420).
Secondly, CO has been shown to play a physiological role in vasomotor control and in signal
transduction by activation of soluble guanylate cyclase (sGC), causing a conversion of GTP to cyclic
GMP (cGMP). CO reversibly ligates the heme core of sGC, and the resulting protoporphyrin IX
intermediate triggers cGMP production (Ndisang et al., 2004, 180425). CO caused vascular
relaxation, independent of the endothelium, in human arterial rings (Achouh et al., 2008, 179918).
rat tail artery (Wang et al., 1997, 084341). and rat thoracic aorta (Lin and McGrath, 1988, 012773).
but not in cerebral vessels (Andresen et al., 2006, 180449; Brian et al., 1994, 076283). Activation of
sGC by CO has been linked to neurotransmission, vasodilation, bronchodilation, inhibition of
platelet aggregation, and inhibition of smooth muscle proliferation (Briine and Ullrich, 1987,
016535; Cardell et al., 1998, 086700; Cardell et al., 1998, 011534; Morita et al., 1997, 085345;
Verma et al., 1993, 193999).
CO-mediated vasorelaxation can also be caused by activation of voltage- or Ca2+-activated
potassium (K+) channels in smooth muscle cells, which leads to membrane hyperpolarization,
voltage-dependent Ca2+ channel closing, reduction of resting Ca2+ concentration and vascular tissue
relaxation (Farrugia et al., 1993, 013826; Wang et al., 1997, 084341). This effect may be linked to
sGC activity; however, it has also been reported to occur independently (Dubuis et al., 2003,
180439; Naik and Walker, 2003, 193852). Developmental stage and tissue type will determine
whether K+ channels or the sGC/cGMP pathway play more of a role in vasorelaxation (Ndisang et
al., 2004, 180425).
Collectively, these older studies demonstrated that exposures to high concentrations of CO
resulted in altered functions of heme proteins other than Hb and Mb. Decreased cellular respiration
and energy production and increased ROS following cytochrome c oxidase inhibition would likely
predispose towards cellular injury and death. The release of NO from sequestered stores could
contribute to the pro-oxidant status if superoxide levels are simultaneously increased. Furthermore,
increased ROS and reactive nitrogen species are known to promote cell signaling events leading to
inflammation and endothelial dysfunction. An inappropriate increase in vasorelaxation due to
inhibition of vasoconstrictor production or to activation of vasodilatory pathways (sGC and ion
channels) could potentially limit compensatory alterations in hemodynamics. Alternatively,
CO-binding to sGC could result in decreased vasorelaxation by interfering with the binding of NO to
sGC. NO can also activate sGC, and with a 30-fold greater affinity than CO, is 1,000-fold more
potent with respect to vasodilation and sGC activation (Stone and Marietta, 1994, 076455). CO
could further contribute to endothelial dysfunction by this mechanism. Although the 2000 CO
AQCD (U.S. EPA, 2000, 000907) made no definitive links between these nonhypoxic mechanisms
of CO and CO-mediated health effects, it did document the potential for CO to interfere with basic
cellular and molecular processes that could lead to dysfunction and/or disease.
5.1.3.2. Recent Studies of Nonhypoxic Mechanisms
Since the 2000 CO AQCD (U.S. EPA, 2000, 000907). new studies have provided additional
evidence for nonhypoxic mechanisms of CO which involve the binding of CO to reduced iron in
heme proteins. These mechanisms, which may be interrelated, are described below and include:
• Alteration in NO signaling
• Inhibition of cytochrome c oxidase
• Heme loss from protein
January 2010 5-3
-------
• Disruption of iron homeostasis
• Alteration in cellular redox status
Recent studies have also demonstrated nonhypoxic mechanisms of CO which are either
indirectly linked to heme protein interactions or not yet understood. These mechanisms are described
below and include:
• Alteration in ion channel activity
• Modulation of protein kinase pathways
This assessment evaluates these nonhypoxic mechanisms in terms of their potential to
contribute to health effects associated with environmentally-relevant CO exposures. As discussed
above, CO at high concentrations may promote oxidative stress, cell injury and death, inflammation
and endothelial dysfunction. Whether lower CO concentrations trigger these same processes is of
key interest since they may potentially contribute to adverse health effects following ambient
exposures.
In addition, a large number of studies published since the 2000 CO AQCD (U.S. EPA, 2000,
000907) have focused on the role of CO derived from HO-catalyzed heme metabolism as an
endogenous signaling molecule and on the potential therapeutic effects of exogenous CO
administered at high concentrations. This assessment addresses aspects of these topics pertaining to
the evaluation of health effects associated with environmentally-relevant CO exposures.
Alteration in NO Signaling
Work by Thorup et al. (1999, 193782) demonstrated altered NO signaling in isolated rat renal
resistance arteries. In one set of experiments, rapid release of NO was observed in response to
exogenous CO. This response was biphasic, peaking at 100 nM CO in the perfusate and declining at
higher concentrations. It was also NOS-dependent as it required L-arginine and was blocked by a
NOS inhibitor. The authors attributed the effects of CO on NO release to either stimulated eNOS or
to displacement of preformed NO from intracellular binding sites. These findings are similar to those
of Thorn and colleagues (Ischiropoulos et al., 1996, 079491; Thorn and Ischiropoulos, 1997, 085644;
Thorn et al., 1994, 076459; Thorn et al., 1997, 084337; Thorn et al., 1999, 016753; Thorn et al.,
1999, 016757; Thorn et al., 2000, 011574; Thorn et al., 2006, 098418) who demonstrated NO
release, presumably from sequestered stores in platelets, endothelial cells, aorta and lung in response
to CO (see above). Furthermore in a second set of experiments, Thorup et al. (1999, 193782)
demonstrated inhibition of agonist-stimulated NOS activity in isolated rat renal resistance arteries.
Here rapid NOS-dependent release of NO following carbachol stimulation was blocked by
pretreatment with 100 nM CO or by upregulation of intracellular HO-1. Additional experiments
using blood-perfused isolated juxtamedullary afferent arterioles demonstrated a biphasic response to
CO with rapid vasodilation observed at lower, but not higher, concentrations of CO. These same
higher concentrations of CO inhibited agonist-stimulated vasodilation in the arterioles. In order to
determine whether CO had a direct effect on the activity of NOS, which is a heme protein, purified
recombinant eNOS was exposed in vitro to CO in the presence of the necessary substrates and
cofactors. A dose-dependent inhibition of NOS by CO was observed, suggesting that CO-mediated
NO release in the isolated vessels was not due to stimulated NOS activity. The authors concluded
that CO effects on vascular tone were due to the liberation of NO from intracellular binding sites at
lower concentrations and to the inhibition of NOS at higher concentrations.
These findings illustrate the potential of CO to alter processes dependent on endogenous NO
either by enhancing intracellular concentrations of free NO (increased vasodilatory influence) or by
inhibiting agonist-induced NO production by NOS (decreased vasodilatory influence). In addition,
CO may compete with NO for binding to sGC as discussed above. Since NO activates sGC to a
greater extent than CO, NO-dependent vasodilation may be significantly impaired in the presence of
CO. In fact, a recent study in transgenic mice demonstrated that chronic overexpression of HO-1 in
vascular smooth muscle resulted in attenuated NO-mediated vasodilation and elevated blood
pressure (Imai et al., 2001, 193864). Results of this study suggested that decreased sensitivity of sGC
to NO contributed to the changes in vascular function. The considerations mentioned above,
however, do not preclude an important role for CO in maintaining vasomotor tone in vessels where
January 2010 5-4
-------
CO and NO do not compete for available heme sites on sGC. This could occur when both mediators
are present at low concentrations compared with sGC or in situations where NOS does not co-
localize with sGC, as discussed by Thorup et al. (1999, 193782).
Inhibition of Cytochrome c Oxidase
High concentrations of CO are known to inhibit cytochrome c oxidase, the terminal enzyme in
the mitochondrial electron transport chain, resulting in inhibition of mitochondrial respiration and
the formation of superoxide from mitochondrial substrates. Several recent studies demonstrated
CO-mediated decreases in cytochrome c oxidase activity in model systems ranging from isolated
mitochondria to whole animals. In a study by Alonso et al. (2003, 193882). exposure of isolated
mitochondria from human skeletal muscle to 50-500 ppm CO for 5 min decreased cytochrome c
oxidase activity. Similarly, exposure of cultured macrophages to 250 ppm CO for 1 h inhibited
cytochrome c oxidase (Zuckerbraun et al., 2007, 193884). In this latter study, increased ROS were
observed following exposure to 250 ppm CO, as well as to CO concentrations as low as 50 ppm, for
1 h. Animal studies demonstrated that exposure of rats to 250 ppm CO for 90 min inhibited
cytochrome c oxidase activity in myocardial fibers (Favory et al., 2006, 184462). Exposure of mice
to 1,000 ppm CO for 3 h, resulting in COHb levels of 61%, decreased cytochrome c oxidase activity
in heart mitochondria (Iheagwara et al., 2007, 193861).
Heme Content Loss from Proteins
In addition to decreasing the activity of cytochrome c oxidase, exposure of mice to 1,000 ppm
CO for 3 h resulted in decreased protein levels and heme content of cytochrome c oxidase in heart
mitochondria (Iheagwara et al., 2007, 193861). CO-mediated heme release was also seen in a study
by Cronje et al. (2004, 180440) and was followed by increased endogenous CO production through
the activation of HO-2 and the induction of HO-1. Loss of heme from proteins leads to loss of
protein function and often to protein degradation.
Disruption of Iron Homeostasis
Exposure of rats to 50 ppm CO for 24 h increased levels of iron and ferritin in the
bronchoalveolar lavage fluid (BALF), decreased lung non-heme iron and increased liver non-heme
iron (Ohio et al., 2008, 096321). Furthermore in this same study, exposure of cultured human
respiratory epithelial cells to 10-100 ppm CO for 24 h caused a dose-dependent decrease in cellular
non-heme iron and ferritin. Heme loss, which was observed in other studies (Cronje et al., 2004,
180440; Iheagwara et al., 2007, 193861). may also contribute to disruption of iron homeostasis. Iron
homeostasis is critical for the sequestration of free iron and the prevention of iron-mediated redox
cycling which leads to ROS generation and lipid peroxidation.
Alteration in Cellular Redox Status
Recent studies demonstrated that exposure to low, moderate, and high levels of CO increased
cellular oxidative stress in cultured cells (Kim et al., 2008, 193961; Zuckerbraun et al., 2007,
193884). A dose-dependent increase in dichlorofluorescein (DCF) fluorescence (an indicator of
ROS) occurred following 1-h exposure to 50-500 ppm CO in macrophages and following 1-h
exposure to 250 ppm CO in hepatocytes. NOS inhibition had no effect on the increase in DCF
fluorescence in CO-treated macrophages, indicating that the effects were not due to an interaction of
CO and NO (Zuckerbraun et al., 2007, 193884). Mitochondria were identified as the source of the
increased ROS since mitochondria-impaired cells (rho zero cells and treatment with antimycin A)
did not respond to CO with an increase in DCF fluorescence. Furthermore, 1-h exposure to 250 ppm
CO inhibited mitochondrial cytochrome c oxidase enzymatic activity in macrophages (Zuckerbraun
et al., 2007, 193884). Recently, inhibition of cytochrome c oxidase was demonstrated in HEK-293
cells transfected with HO-1 and in macrophages with induced HO-1; this effect was attributed to
January 2010 5-5
-------
endogenously produced CO (D'Amico et al., 2006, 193992). In hepatocytes, exposure to 250 ppm
CO for 1 h resulted in Akt phosphorylation and nuclear translocation of nuclear factor kappa B
(NF-KB), effects which were blocked by antioxidants (Kim et al., 2008, 193961). Significant
increases in apoptosis were also observed in this model. Thus in this study, CO exposure led to
uncoupled mitochondrial respiration and ROS-induced programmed cell death.
Further evidence for cellular redox stress is provided by studies in which glutathione stores
were altered following CO exposure in vitro (Kim et al., 2008, 193961: Patel et al., 2003, 043155).
In addition, mitochondrial redox stress was observed in livers of rats exposed to 50 ppm CO
(Piantadosi et al., 2006, 180424). Furthermore, an adaptive increase in intracellular antioxidant
defenses (i.e., superoxide dismutase) was observed in endothelial cells exposed to 10 ppm CO for
40 min (Thorn et al., 2000, 011574). and mitochondrial biogenesis was observed in hearts of mice
exposed to 250 ppm CO for 1 h (Suliman et al., 2007, 193768).
Several mechanisms could contribute to the cellular redox stress elicited by CO exposure.
First, inhibition of cytochrome c oxidase could result in increased mitochondrial superoxide
generation. Secondly, interactions of CO with heme proteins could lead to the release of heme and
free iron and subsequently to the generation of ROS. As mentioned above, increased ROS generation
has been linked to cellular injury and death, inflammation, and endothelial dysfunction.
Two of the above-mentioned studies demonstrated that CO-mediated mechanisms were
unrelated to hypoxia by showing that hypoxic conditions failed to mimic the results obtained with
CO. Hence, the mitochondrial redox stress and mitochondrial pore transition observed in livers from
rats exposed to CO (Piantadosi et al., 2006, 180424) and the cardiac mitochondrial biogenesis
observed in mice exposed to CO (Suliman et al., 2007, 193768) were attributed specifically to
nonhypoxic mechanisms of CO.
Alteration in Ion Channel Activity
Work by Dubuis et al. (Dubuis et al., 2002, 193911) demonstrated increased current through
Ca2+-activated K+ channels in smooth muscle cells from pulmonary arteries of rats exposed to
530 ppm CO for 3 wk. These findings provide further evidence for non-cGMP-dependent
vasodilatory actions of CO.
Modulation of Protein Kinase Pathways
Endogenously produced CO is a gaseous second messenger molecule in the cell. Work from
numerous laboratories has demonstrated the potential for CO to be used as a therapeutic gas with
numerous possible clinical applications since it can produce anti-inflammatory, anti-apoptotic, and
anti-proliferative effects (Durante et al., 2006, 193778; Ryter et al., 2006, 193765). These studies
generally involved pretreatment with CO followed by exposure to another agent 12-24 h later. There
is extensive literature on this topic as reviewed by Ryter et al. (2006, 193765). Durante et al. (2006,
193778) and others. A number of these processes are mediated through cGMP while others involve
redox-sensitive kinase pathways, possibly secondary to CO-dependent generation of ROS. For
example, 250 ppm CO inhibited growth of airway smooth muscle cells by attenuating the activation
of the extracellular signal-regulated kinase 1/2 (ERK 1/2) pathway, independent of sGC and other
MAP kinases (Song et al., 2002, 037531). A second example is provided by the study of Kim et al.
(2005, 193959) where 250 ppm CO inhibited PDGF- induced smooth muscle cell proliferation by
upregulating p21Wafl/Clpl and caveolin-1, and down-regulating cyclin A expression. In this case,
effects were dependent upon cGMP and the p38 MAPK pathway (Kim et al., 2005, 193959). Thirdly,
rat endothelial cells exposed to 15 ppm CO escaped anoxia/reoxygenation-induced apoptosis via
modulation of the signaling pathways involving phosphoinositide 3-kinase (PI3K), Akt, p38 MAP
kinase, Signal Transducers and Activators of Transcription (STAT-1) and STAT-3 (Zhang et al.,
2005, 184460). In a fourth study, Akt was found to be responsible for the CO-induced activation of
NF-KB, protecting against hepatocyte cell death (Kim et al., 2008, 193961). While research focusing
on therapeutic applications of CO generally involves high-level short-term exposure to CO (i.e., 250-
1,000 ppm for up to 24 h), some studies found effects below 20 ppm (Zhang et al., 2005, 184460).
Few if any studies on the therapeutic effects of CO have explored the dose-response relationship
January 2010 5-6
-------
between CO and pathway activation/deactivation, so it remains unclear how these effects may be
related to environmentally-relevant exposures.
Concentration-Response Relationships
In many cases, the concentrations of exogenous CO required for these nonhypoxic effects
were much higher (Alonso et al, 2003, 193882; Favory et al, 2006, 184462; Iheagwara et al, 2007,
193861; Thorup et al., 1999, 193782) than concentrations of CO in ambient air. However, in some
studies the effects were mimicked by upregulation of HO-1 which would result in increased local
production of CO as well as of iron and biliverdin (D'Amico et al., 2006, 193992; Imai et al., 2001,
193864; Thorup et al., 1999, 193782). For example, HO-1 upregulation or overexpression attenuated
carbachol-mediated NO release and NO-mediated vasodilation, similar to the effects of exogenous
CO in these same models (Imai et al., 2001, 193864; Thorup et al., 1999, 193782). In the study by
D'Amico et al. (2006, 193992), overexpression of HO-1 in cells inhibited cellular respiration by
12% and decreased cytochrome c oxidase activity by 23%. It is not clear how comparable these
conditions involving increased intracellular concentrations of endogenous CO are to increased
intracellular concentrations of CO resulting from exogenous CO exposures. Neither is it clear what
concentrations of intracellular CO are generated locally within cells as a result of HO-catalyzed
heme metabolism. However, a small amount of a relatively high local concentration of endogenous
CO produced in a regulated manner by HO-1 and HO-2 may be sufficient to react with local targets
(e.g., heme proteins), while a larger amount of exogenous CO may be required to reach the same
targets. This may be due to indiscriminate reactions of exogenous CO with other target proteins or to
other issues related to compartmentalization. It is conceivable that acute or chronic exposures to
ambient CO could "sensitize" (or "desensitize") targets of endogenous cellular CO production, but
there is no experimental evidence to support this mechanism.
There is a growing appreciation that nonhypoxic mechanisms may contribute to the effects
associated with CO toxicity and poisoning (Ischiropoulos et al., 1996, 079491; Thorn et al., 1994,
076459; Weaver et al., 2007, 193939). On the other hand, recent studies suggest that exogenous CO
at lower concentrations may have beneficial anti-inflammatory, anti-proliferative and cytoprotective
effects under certain circumstances (Durante et al., 2006, 193778; Ryter et al., 2006, 193765). Since
the focus of this assessment is on mechanisms which are relevant to ambient exposures, it is
important to understand which mechanisms may occur at "low" (50 ppm and less) and "moderate"
(50-250 ppm CO) concentrations of CO. Hence, both recent animal studies and relevant older ones
which add to the understanding of mechanisms in this range of CO concentrations are briefly
summarized in Table 5-1. It should be noted that most of the above-mentioned nonhypoxic
mechanisms were demonstrated at CO concentrations of 50 ppm and less.
January 2010 5-7
-------
Table 5-1. Responses to CO exposures at low and moderate concentrations.
Study
Model System
CO Exposure
Response
Notes
IN VITRO
Alonso et al. (2003,
193882)
Thorn and Ischiropoulos
(1997, 085644)
Thorn etal. (1997,
084337)
Thorn et al. (2000,
011574)
Song et al. (2002,
0375311
Kim et al. (2005,
1939591
Kim et al. (2008,
1939611
Zhang et al. (2005,
184460)
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
1 h
250 ppm
1 h
2x per day
250 ppm
1 h
15 ppm
0.5-24 h
50 and 250 ppm
1 h
10-100 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 significant,
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 p2i"™'w'h'1 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
Thorn etal. (1999,
0167571
Sorhaug et al. (2006,
1804141
Loennechen et al.
(1999, 0115491
Favory et al. (2006,
1844621
Piantadosi et al. (2006,
1804241
Suliman et al. (2007,
1937681
Wellenius et al. (2004,
087874)
Wellenius et al. (2006,
156152)
Rats
Rats
Rats
Rats
Rats
Rats
Rats
Mice
Rats
Model of Ml
Rats
Model of Ml
50 ppm
24 h
50 ppm
1 h
100 ppm
1 h
100 ppm
1 h
50 ppm
1 h
200 ppm
72 wk
100 and 200 ppm
1-2wk
250 ppm
90 min
50 ppm CO or
hypobaric hypoxia for
1, 3, or 7 days
250 ppm
1 h
35 ppm
1 h
35 ppm
1 h
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
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
Compare with in vitro experiments in
same paper
Effects blocked by NOS inhibitor
Inhibition of NOS abrogated CO
effects
12 and 23% 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
January 2010
5-8
-------
Study
Carraway et al. (2002,
0260181
Gautier et al. (2007,
0964711
Melinetal.
1938331
Melinetal.
0375021
(2005,
(2002,
Model System
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
Hypobaric hypoxia
+ 50 ppm CO
3wk
3 wk of hypobaric
hypoxia with 50 ppm
CO during last week
50 ppm
10 wk
50 ppm
10 wk
Response Notes
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
5.1.3.3. Implications of Nonhypoxic Mechanisms
A key issue in understanding the biological effects of environmentally-relevant exposures to
CO is whether the resulting partial pressures of CO (pCO) in cells and tissues can initiate cell
signaling which is normally mediated by endogenously generated CO or perturb signaling which is
normally mediated by other signaling molecules such as NO.
Several aspects need to be considered. First of all, during a period of exogenous CO uptake,
Hb acts as a buffer for most cells and tissues by limiting the availability of free CO. Nevertheless,
COHb delivers CO to cells and tissues. This delivery involves CO's dissociation from Hb followed
by its diffusion down a pCO gradient. Hence, greater release of CO from COHb will occur under
conditions of low cell/tissue pCO. Conversely, higher cell/tissue pCO in cells/tissues than in the
blood will lead to the egress of CO from cells/tissues.
A second consideration is the role played by O2 in competing with CO for binding to
intracellular heme protein targets. In general, heme proteins (e.g., cytochrome c oxidase) are more
sensitive to CO when O2 is limited. Hence, hypoxic conditions would be expected to enhance the
effects of CO. This concept is demonstrated in the study by D'Amico et al. (2006, 193992). NO,
which also competes with O2 and CO for binding to heme proteins, may have a similar impact.
A third consideration is whether certain cell types serve as primary targets for the effects of
CO. Besides the blood cells (including leukocytes and platelets), the first cells encountering CO
following its dissociation from Hb are the endothelial cells which line blood vessels. An exception to
this situation is in the lungs where epithelial and inflammatory cells found in airways and alveoli are
exposed to free CO prior to CO binding to Hb. These lung cells may also serve as unique targets for
CO. Processes such as pulmonary microvascular endothelial dysfunction, inflammatory cell
activation, and respiratory epithelial injury may ensue as a result of preferential targeting of these
cell types.
Since there is potential for exogenous CO to affect endogenous pools of CO, the
concentrations of CO in cells and tissues before and after exogenous exposures are of great interest.
Table 5-2 summarizes findings from four recent studies relevant to this issue. It should be noted that
exposure to 50 ppm CO resulted in a three- to fivefold increase in tissue CO concentration.
January 2010
5-9
-------
Table 5-2. Tissue concentration of CO following inhalation
Study
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
1,250 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
Furthermore, endogenous CO production is known to be increased during inflammation,
hypoxia, increased heme availability and other conditions of cellular stress where HO-1 or HO-2
activity is increased. A few studies reported cell and tissue concentrations of CO along with
accompanying COHb levels resulting from enhanced endogenous CO production; Table 5-3
summarizes these findings. Additional measurements of CO levels in cells and tissues following
increased endogenous production and following inhalation of exogenous CO may provide further
insight into the relationship between the CO tissue concentration and biological responses.
January 2010
5-10
-------
Table 5-3. Tissue concentration of CO following increased endogenous production.
Study
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 three- to fourfold 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
It should be noted that increased cellular and tissue concentrations of biliverdin and iron
accompany the increased endogenous production of CO by HO-1 and HO-2. Biliverdin and iron
have known biological effects, with biliverdin exhibiting antioxidant properties and iron exhibiting
pro-oxidant properties (Piantadosi et al., 2006, 180424). which could complicate interpretation of
results from studies in which HO-1 and HO-2 activities are increased. In addition, indiscriminate
reactions occurring in the case of exogenous CO would likely lead to less specific responses than
those mediated by reactions of endogenously-produced CO with local targets. Hence, the situations
of increased endogenous CO production and of exogenous CO exposure are not equivalent.
A further consideration is that in the numerous conditions and disease states where HO-1 is
induced, increased levels of endogenously produced CO may represent an adaptive response to stress
(Durante et al., 2006, 193778; Piantadosi, 2008, 180423). These increases and the accompanying
increases in COHb generally fall in the range of 1.5- to 4-fold, with the exception of some situations
of hemolytic anemia and hemoglobin disorders (see Figure 4-11 for results in humans). The resulting
excess endogenous CO may react intracellularly with heme proteins or diffuse into the blood
according to the gradient of pCO in the cell/tissue and blood compartments. In many cases,
beneficial effects or compensatory mechanisms may result as a result of short-term induction of
HO-1, as reviewed by Ryter et al. (2006, 193765) and Durante et al. (2006, 193778). Longer term
increases in HO-1 are sometimes associated with protective responses, as in the case of
atherosclerosis (Cheng et al., 2009, 193775; Durante et al., 2006, 193778). and sometimes with
pathophysiologic responses as demonstrated in hypoxic pulmonary vascular remodeling (Carraway
et al., 2002, 026018) and models of salt-sensitive hypertension (Johnson et al., 2003, 193868;
Johnson et al., 2004, 193870) and metabolic syndrome (Johnson et al., 2006, 193874). Increased
endogenous CO in hearts of individuals with ischemic heart disease and in lungs of individuals with
various forms of inflammatory lung disease might also be expected (Scharte et al., 2006, 194115;
Yamaya et al., 1998, 047525; Yasuda et al., 2005, 191953) (Figure 4-12). It is conceivable that
prolonged increases in endogenous CO production in chronic disease states may result in less of a
reserve capacity to handle additional intracellular CO resulting from exogenous exposures, but there
is no experimental evidence to support this mechanism. Perhaps these circumstances lead to
dysregulated functions or toxicity. Thus, CO may be responsible for a continuum of effects from cell
signaling to adaptive responses to cellular injury (Piantadosi, 2008, 180423). depending on
intracellular concentrations of CO, heme proteins and molecules which modulate CO binding to
heme proteins.
January 2010
5-11
-------
5.1.3.4. Summary
CO is a ubiquitous cell signaling molecule with numerous physiological functions. The
endogenous generation and release of CO from heme by HO-1 and HO-2 is tightly controlled, as is
any homeostatic process. However, exogenously-applied CO has the capacity to disrupt multiple
heme-based signaling pathways due to its nonspecific nature. Only a limited amount of information
is available regarding the impact of exogenous CO on tissue and cellular levels of CO and on
signaling pathways. However recent animal studies demonstrated increased tissue CO levels and
biological responses following exposure to 50 ppm CO. Whether or not environmentally-relevant
exposures to CO lead to adverse health effects through altered cell signaling is an open question for
which there are no definitive answers at this time. However, experiments demonstrating
oxidative/nitrosative stress, inflammation, mitochondrial alterations and endothelial dysfunction at
concentrations of CO within one or two orders of magnitude higher than ambient concentrations
suggest a potential role for such mechanisms in pathophysiologic responses. Furthermore, prolonged
increases in endogenous CO resulting from chronic diseases may provide a basis for the enhanced
sensitivity of susceptible populations to CO-mediated health effects such as is seen in individuals
with coronary artery disease.
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
The 2000 CO AQCD (U.S. EPA, 2000, 000907) examined the association between short-term
variations in ambient CO concentrations and cardiovascular morbidity. While the results presented
by these studies did provide suggestive evidence of ambient CO levels being associated with
exacerbation of heart disease, the AQCD determined that the evidence was inconclusive. The reasons
January 2010
5-12
-------
for this conclusion were given as: internal inconsistencies and lack of coherence of the reported
results within and across studies; the degree to which average ambient CO levels derived from fixed-
site monitors are representative of spatially heterogeneous ambient CO values or of personal
exposures that often include nonambient CO; and the lack of biological plausibility for any harmful
effects occurring with the very small changes in COHb levels (from near 0 up to 1.0%) over typical
baseline levels (about 0.5%) that would be expected with the low average ambient CO
concentrations reported in the epidemiologic studies (generally <5.0 ppm, 1-h daily max) (U.S. EPA,
2000, 000907). These reasons were also cited in the discussion of the effects of short-term exposure
to CO on mortality and other types of morbidity. The AQCD posited that ambient CO concentrations
used as exposure indices in epidemiologic studies may be surrogates for ambient air mixtures
produced by combustion sources and/or other constituents of such mixtures. In addition, the AQCD
noted that the epidemiologic evidence was stimulating increased scientific interest regarding ambient
CO exposures as a potential risk factor for exacerbation of heart disease and other health effects,
although the epidemiologic studies were subject to considerable biological and statistical uncertainty.
The following section reviews the literature since the 2000 CO AQCD, including numerous
new studies on relevant cardiac endpoints and biomarkers and additional studies of daily hospital
admissions for heart disease. New epidemiologic evidence addresses some of the aforementioned
uncertainties, including consistency and coherence of results and the possibility that CO may be
acting as a surrogate for other combustion-derived air pollutants.
5.2.1.1. Heart Rate and Heart Rate Variability
Heart rate variability (HRV) refers to the beat-to-beat alterations in the heart rate (HR) and is
generally determined by analyses of time and frequency domains measured by electrocardiograms
(ECG). The time domains often analyzed are (a) normal-to-normal (NN or RR) time interval
between each QRS complex, (b) standard deviation of the normal-to-normal interval (SDNN), and
(c) mean squared differences of successive difference normal-beat to normal-beat intervals (rMSSD).
Shorter time domain variables results in lower HRV. The frequency domains often analyzed are (a)
the ratio of low energy frequency (LF) to high energy frequency (HF), and (b) the proportion of
interval differences of successive normal-beat intervals greater than 50 ms (PNN50), reflecting
autonomic balance. Decreased HRV is associated with a variety of adverse cardiac outcomes such as
arrhythmia, myocardial infarction (MI), and heart failure (Deedwania et al., 2005, 195134; De Jong
and Randall, 2005, 193996: Huikuri et al., 1999, 184464: Rajendra et al., 2006, 193787).
Three studies investigated the association between ambient air pollution, including CO, and
HRV in Boston, MA and reported inconsistent results. The earlier of these studies recruited 21 active
residents aged 53-87 yr and performed up to 12 ECG assessments on each subject over a period of 4
mo (summer 1997). Particles (PMi0, PM2.5) and several gaseous pollutants (O3, NO2, and SO2) were
monitored at fixed sites (up to 4.8 mi from the study site), while CO was monitored 0.25 mi from
each participant's residence. Lag periods for the preceding 1 h, 4 h, and 24 h before each subject's
HRV assessment were analyzed, and results showed that only PM2 5 and O3 were associated with
HRV parameters (Gold et al., 2000, 011432).
A similar study by the same group of researchers 2 yr later involved 28 older subjects (aged
61-89 yr) who were living at or near an apartment complex located on the same street as the Harvard
School of Public Health. The subjects were seen once a week for up to 12 wk, and HRV parameters
(SDNN, r-MSSD, PNN50, LF/HF ratio) were measured for 30 min each session. Data for PM2.5, BC,
and CO were recorded at the Harvard School of Public Health (<1 km from the residence) while data
for NO2, O3, and SO2 were collected from government regulatory monitoring sites. There were
moderate correlations between CO and PM25 (r = 0.61) and between CO and NO2 (r = 0.55) but not
with SO2 (r = 0.18) or O3 (r = 0.21). Similarly, PM25 was associated with HRV, whereas in contrast
to the previous study, CO was associated1 with a negative change in SDNN (% change: -13
[95% CI: -24.06 to -1.88]), r-MSSD (% change: -31.88 [95% CI: -38 to -7.5]), and PNN50
(% change: -46.25 [95% CI -103.95 to -9.38] per 0.5 ppm increase in 24-h avg CO concentration)
(Schwartz et al., 2005, 074317).
1 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).
January 2010 5-13
-------
An additional Boston, MA study examined HRV parameters (SDNN, LF, HF, LF/HF ratio)
among 603 persons from the Normative Aging Study, a longitudinal study that originally recruited
2,280 men in the greater Boston area during 1963. The cohort members were examined (November
2000-October 2003) and the ECG data were linked to air pollution data for PM2 5, particle number
concentration, BC, O3, NO2, SO2, and CO. Lagged pollutant effects for a 4-h, 24-h, and 48-h ma
were examined. The main pollutant effects were with PM2 5 and O3, while CO was not associated
with HRV (Park et al., 2005, 057331).
A study in Mexico City selected 30 subjects from the outpatient clinic at the National Institute
of Cardiology and followed them for -10 h (starting at 9:00 a.m.) (Riojas-Rodriguez et al., 2006,
156913). Each subject was connected to a Holter ECG monitor (e.g., a portable ECG monitor) and
also given personal PM25 and CO monitors. The subjects went about their usual daily activities, and
the personal PM2 5 and CO data were linked to various ECG parameters (HR, R-R, LF, HF) at
various lags. In copollutant models with PM2 5, personal CO exposure for the same 5-min period was
significantly associated with a decrease in LF and very low energy frequency (VLF) parameters with
coefficients equal to -0.024 (95% CI: -0.041 to -0.007) and -0.034 (95% CI: -0.061 to -0.007),
respectively, for a 1 ppm increase in 1-h CO concentration.
In an additional study conducted in Mexico City, 34 residents from a nursing home underwent
HRV analysis every other day for 3 mo (Holguin et al., 2003, 057326). Exposure assessment for
ambient PM25 was based on data recorded at a monitor on the roof of the nursing home, while
exposures to ambient O3, NO2, SO2, and CO were derived from data recorded at a fixed site 3 km
from the nursing home. Exposures for the same day and 1-day lags were analyzed, and only O3 and
PM2 5 were positively associated with HRV.
Wheeler et al. (2006, 088453) examined 18 individuals with COPD and 12 individuals with
recent MI living in Atlanta, GA. Morning ECG readings were collected by a Holter system by a field
technician in the subjects' homes. Ambient air pollution exposures for PM25, O3, NO2, SO2 and CO
were derived from data recorded at fixed sites throughout metropolitan Atlanta. Three exposure
periods were analyzed: the hour of the ECG reading, 4-h mean, and 24-h mean before the reading.
While positive effects were reported for NO2 and PM2 5, no quantitative results were reported for
CO.
After reviewing 2,000 patient charts, Dales (2004, 099036) recruited 36 subjects with CAD
from the Toronto Western Hospital's noninvasive cardiac diagnostic unit. HR and HRV (SDNN, N-
N, HF, LF, HF/LH ratio) were assessed 1 day each week for up to 10 wk by a Holter monitoring
system. Personal air sampling for PM2 5 and CO was carried out for the same 24-h period while
subjects went about their usual daily activities for that period. Stratified results showed that among
those not on beta-receptor-blockers, personal CO exposure was positively associated with SDNN (p
= 0.02). However, in the group taking beta blockers, there was a negative association (p = 0.06).
Personal exposure to PM2 5 was not associated with HRV.
Peters et al. (1999, 011554) examined HR among a sub-sample of the Monitoring of Trends
and Determinants in Cardiovascular Disease (MONICA) study (n = 2,681) in Augsburg, Germany.
Total suspended particles (TSP), SO2, and CO data were collected from a single monitoring station
located in the center of the city and linked to each subject to estimate exposures on the same day and
5 days prior. A 0.5 ppm change in 24-h CO concentration was associated with an increase in HR of
approximately 1 beat per minute, whereas CO based on a 5-day exposure had no effect on HR.
Thirty-one subjects with CHF had their pulse rate recorded daily over a 2-mo period, and the
correlation between pulse rate and air pollutants was examined (Goldberg et al., 2008, 180380).
There was weak evidence for a decrease in pulse rate associated with the lag 1 SO2 concentration
after adjustment for personal and meteorological factors and no evidence for an effect associated
with any of the other air pollutants (adjusted mean difference for CO: 0.245 [95% CI -0.209 to
0.700] lag 0-2).
Liao et al (2004, 056590) investigated men and women aged 45-64 yr from the Atherosclerosis
Risk in Communities (ARIC) study (Washington County, MD; Forsyth County, NC; and selected
suburbs of Minneapolis, MN). The sample sizes were 4,899, 5,431, 6,232, 4,390 and 6,784 for
analyses involving PMi0, O3, CO, NO2, and SO2, respectively. County-level exposure estimates for
24-h CO were calculated for 1, 2, and 3 days prior to clinical examination. A 0.5 ppm increase in
24-h CO concentration (at lag 1) was associated with an increase in HR (beats/minute) ((3 = 0.357,
p < 0.05). CO was not significantly associated with changes in SDNN.
The Exposure and Risk Assessment for Fine and Ultrafine Particles in Ambient Air (ULTRA)
study was carried out in three European cities: Amsterdam, The Netherlands; Erfurt, Germany; and
January 2010 5-14
-------
Helsinki, Finland; a panel of subjects with CAD was followed for 6 mo with biweekly clinical visits,
which included an ECG reading to assess HRV (Timonen et al., 2006, 088747). The time-domain
measures of HRV (SDNN and rMSSD) were analyzed along with frequency-domain measures,
which included power spectrum densities for LF and HF. Exposures to ambient air pollution (PM2 5,
PMio, NO2, CO) were derived from data recorded at fixed-monitoring site networks within each city.
Correlation coefficients for NO2 and CO ranged from 0.32 to 0.86 in the three cities. CO was
moderately correlated with PM10 in Helsinki (r = 0.40), with PM2.5 in Amsterdam (r = 0.58), and
more highly correlated with PMi0 in Erfurt (r = 0.77). Various lag periods were examined, including
lag 0 (24 h prior to the clinical visit) through a 0- to 2-day avg lag and a 0- to 4-day avg lag. In total
there were 1,266 ECG recordings used in the final analyses. In the pooled analyses (e.g., across
cities) a 0.5 ppm increase in 24-h CO concentration was associated with a decrease in 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 2-day ((3 -3.4
[95% CI: -6.1 to -0.4]; (3= -17.6 [95% CI: -34.4 to -0.9], respectively). However, the same study
reported no effect for CO on BP and HR (Ibald-Mulli et al., 2004, 087415).
A small panel study in Kuopio, Finland, which was designed as the pilot study for the ULTRA
study, examined simultaneous ambulatory ECG and personally monitored CO readings among
6 male patients with CAD (Tarkiainen et al., 2003, 053625). The patients were asked to follow their
usual daily activities, but data were recorded only three times with 1-wk intervals. The CO exposures
were divided into low (< 2.7 ppm) and high (>2.7 ppm) and during the high CO exposure r-MSSD
increased on average by 2.4 ms; however, there was no effect on RR or SDNN.
A study in Taiwan recruited 83 patients (aged 40-75 yr) from the National Taiwan University
Hospital, Taipei, and conducted ambulatory ECG readings using aHolter system (Chan et al., 2005,
088988). Ambient air pollution exposures for PMi0, NO2, SO2, and CO were derived from 12 fixed-
site monitoring stations across Taipei. Lag periods of 1 h to 8 h prior to the ECG reading were
analyzed, and only NO2 was associated with HRV parameters (SDNN and LF); CO was not
associated with HRV.
Min et al. (2009, 199514) investigated the effects of CO on cardiac autonomic function by
measuring HRV in patients with and without metabolic syndrome. Several criteria were used to
classify metabolic syndrome, including waist circumference, triglycerides and cholesterol levels,
blood pressure, and fasting glucose level. The group classified as having metabolic syndrome
showed significant decreases in SDNN and HF, and those declines were significantly associated with
CO exposure with a 1- to 2-day lag. Copollutant models with PMi0 and NO2 gave similar results.
In summary, few studies have examined the effect of CO on HR, and while two of the three
studies reported a positive association, further research is warranted to corroborate the current
results. Similarly, while a larger number of studies have examined the effect of CO on various HRV
parameters, mixed results have been reported throughout these studies. Furthermore, with several
HRV parameters often examined, there are mixed results across the studies as to the HRV parameters
that are positively associated with CO exposure. Table 5-4 presents a summary of the reviewed
studies. Due to the heterogeneity of endpoints (see column "cardiac endpoint" in Table 5-4), these
studies do not lend themselves to a quantitative meta-analysis or inclusion in a summary figure.
January 2010 5-15
-------
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
Park etal. (2005, 0573311
Riojas-Rodriguez et al. (2006,
156913)
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,
053625)
Chan et al. (2005, 0889881
Min et al. (2009, 1995141
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)
South Korea
(n=986)
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
SDNN, HF, LF
Upper CO
Concentrations
from AQSa 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
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: 2.4L
Range: 0.4-16.5
Mean: 3.6
Range: 1.5-7.1
NR; IQR: 1.8 ppm
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
Mean: 0.45
Range: 0.10-7. 20
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, PM10, PM25, N02,
S02
None
PM10, N02, S02
PM,o, N02, S02
NA: Not Available
3 Includes range across individual monitors in study site; AQS data available for U.S. studies only
95th percentile of 24-h levels
5.2.1.2. ECG Abnormalities Indicating Ischemia
The ST-segment of an ECG represents the period of slow repolarization of the ventricles and
ST-segment depression can be associated with adverse cardiac outcomes, including ischemia. Gold
et al. (2005, 087558) recruited a panel of 28 older adults living at or near an apartment complex
located within 0.5 km of a monitoring site in Boston, MA. Each subject underwent weekly ECGs for
12 wk in summer 1999 with the main outcome of interest being the ST-segment. Air pollution data in
the form of PM2.5, BC, and CO were collected from a central site within 0.5 km of the residences of
the subjects and averaged over various lag periods (1- to 24-h, 12-h, and 24-h ma) before the ECG.
The final analyses included 24 subjects with 269 observations, and results showed consistent
negative associations of ST-segment change with increased BC with the strongest association with
the 5-h lag. CO during the same lag period also showed a negative association with ST-segment
change; however, only BC remained significant in multipollutant models.
The most recent study by this group of researchers utilized a repeated-measures study to
investigate the associations between ambient air pollution and ST-segment level changes averaged
over 30-min periods in patients with coronary artery disease (CAD) (Chuang et al., 2008, 155731).
January 2010
5-16
-------
The authors reported that increases in mean PM2.5, BC, NO2 and SO2 concentrations predicted
depression of 30-min averaged ST-segment levels. No association of ST-segment depression was
observed with CO or O3.
5.2.1.3. Arrhythmia
Cardiac arrhythmia refers to a broad group of conditions where there is irregular electrical
activity in the heart. The main types of arrhythmias are fibrillation, tachycardia, and bradycardia, all
resulting from dysfunction of the upper (atria) and lower (ventricle) chambers of the heart. Briefly,
fibrillation refers to when a chamber of the heart quivers chaotically rather than pumps in an orderly
fashion, tachycardia refers to a rapid heart beat (e.g., >100 beats/min), while bradycardia refers to a
slow heart beat (e.g., <60 beats/min). A few air pollution panel studies have examined the occurrence
of cardiac arrhythmias by analyzing data recorded by implantable cardioverter defibrillators (ICDs)
among cardiac patients. The majority of these studies were conducted in North America, with the
main outcome investigated being tachycardia. Results of these studies provide little evidence for an
association between cardiac arrhythmia and ambient CO.
For example, Dockery and colleagues (2005, 078995) analyzed the relationship between
ambient air pollution and the daily incidence of ventricular tachyarrhythmia among 203 patients with
ICDs in Boston, MA. An hourly city average for the Boston metropolitan area was calculated for
CO, O3, NO2, SO2, SO42", BC, and PM2 5. Although positive associations between ventricular
arrhythmic episode days were found for all mean pollutant levels on the same day and previous days,
none of these associations approached statistical significance. However, when the analyses were
stratified by patients who had a previous incidence of ventricular arrhythmia within 3 days or greater
than 3 days to the day of interest, a 0.5 ppm increase in 24-h CO concentration was positively
associated with incidence of ventricular arrhythmia (OR: 1.68 [95% CI: 1.18-2.41]) among those
who had a ventricular arrhythmia within the last 3 days.
A similar study in eastern Massachusetts examined cardiac arrhythmia by analyzing
defibrillator discharges precipitated by either ventricular tachycardia or fibrillation among 100
cardiac patients (Peters et al, 2000, 011347). Exposure to ambient CO was estimated for the same
day, 1-day, 2-day, 3-day, and a 5-day mean lag period. CO was moderately correlated with PM10
(r = 0.51) and PM2 5 (r = 0.56) and more highly correlated with NO2 (r = 0.71). When analyzing
patients who had at least one defibrillator discharge (n = 33), there was no association with CO.
However, when analyzing patients who had at least 10 discharges (n = 6), a 0.5 ppm increase in 24-h
CO concentration (lag 0-4) was associated with an increased odds of a defibrillator discharge (OR:
1.66[95%CI: 1.01-2.76]).
In contrast, other air pollution panel studies conducted in St Louis, MO (among 56 subjects)
(Rich et al., 2006, 089814). Atlanta, GA (among 518 subjects) (Metzger et al., 2007, 092856).
Boston, MA (among 203 subjects) (Rich et al., 2005, 079620). and Vancouver, Canada (Rich et al.,
2004, 055631: Vedal et al., 2004, 055630) (among 34 and 50 subjects respectively) did not find an
association between short-term changes in ambient CO and occurrence of cardiac arrhythmia in
patients with implantable defibrillators. The study in Boston also examined atrial fibrillation
episodes among the same group of subjects and did not find an association with ambient CO (Rich et
al., 2005, 079620).
An alternative method used to assess the relationship between cardiac arrhythmia and ambient
air pollution is to analyze cardiac data recorded via ECG. Two studies have employed this method
and reported inconsistent results. A study in Steubenville, OH, which is located in an industrial area,
examined weekly ECG data among 32 nonsmoking older adults for 24 wk during summer and fall
(Sarnat et al., 2006, 090489). Ambient exposures for up to 5 days prior to the health assessment
(based on a 5-day ma) were calculated for PM2 5, SO42", elemental carbon (EC), O3, NO2, SO2, and
CO from data recorded at one central monitoring site. Increases in ambient CO were not associated
with increased odds of having at least one arrhythmia during the study period.
In contrast, a study in Germany examined the relationship between ambient air pollution and
the occurrence of supraventricular (atria) and ventricular tachycardia recorded via monthly 24-h
ECGs among 57 subjects over a 6-mo period (Berger et al., 2006, 098702). Exposure estimates were
calculated for ambient ultrafine particles, PM25, CO, NO, NO2, and SO2 for various lag 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 0.5 ppm
increase in ambient 24-h CO concentration (lag 0-4 days prior to ECG) was positively associated
January 2010 5-17
-------
with the occurrence of supraventricular tachycardia (OR: 1.36 [95% CI: 1.08-1.74]). However,
ambient CO was not associated with ventricular tachycardia.
In summary, the studies that have examined associations between CO and the occurrence of
cardiac arrhythmias provided little evidence of a CO effect on cardiac arrhythmias. While most
studies analyzed data from ICDs, very few reported significant associations, which is similar to the
studies that analyzed ECG data to evaluate cardiac arrhythmias in 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.
Study
Location,
Sample Size
ARRHYTHMIAS (AMONG PATIENTS
Dockeryetal. (2005,
078995)
Peters et al. (2000,
011347)
Rich et al. (2006,
089814)
Metzger et al. (2007,
092856)
Rich et al. (2005,
079620)
Rich et al. (2004,
055631)
Vedal et al. (2004,
055630)
Boston, MA
(n = 203)
Massachusetts, US
(n = 100)
Boston, MA
(n = 56)
Atlanta, GA
(n = 518)
Boston, MA
(n = 203)
Vancouver,
Canada
(n = 34)
Vancouver, Can
(n = 50)
Cardiac
Endpoint
WITH ICDS)
Ventricular
tachycardia
Ventricular
fibrillation or
tachycardia
Ventricular
arrhythmia
Ventricular
tachycardia
Atrial fibrillation
ICD discharge
due to
arrhythmia
ICD discharge
due to
arrhythmia
Upper CO
Concentrations
from AQSa in ppm
98th%: 0.89-2.33
99th%: 0.99-2.55
(24 h)
98th%: 1.60-2.58
99th%: 1.75-2.71
(24 h)
98th%: 0.89-2.33
99th%: 0.99-2.55
(24 h)
98th%: 5.0
99th%: 5.6
(1h)
98th%: 0.89-2.33
99th%: 0.99-2.55
(24 h)
NA
NA
CO Concentrations
Reported by Study
Authors
in ppm
25th, 50th, 75th, 95th,
percentiles: 0.53, 0.80, 1.02,
1 .37 (2-day)
Mean: 0.58 (24 h)
Max: 1.66
25th, 50th, 75th percentiles:
0.4, 0.5, 0.6 (24 h)
Mean: 1.7(1 h)
Range: 0.1-7.7
25th, 50th, 75th, 95th,
percentiles: 0.53, 0.80, 1.02,
1 .37 (2-day)
Mean: 0.55 (24 h)
IQR:0.16
Mean: 0.6 (24 h)
Range: 0.3-1 .6
Copollutants
PM2.5, BC, 03, N02, S02, S042"
PM2J5, PM10, BC, 03, N02, S02,
S04
PM2.5, EC, 03, N02, S02
PM10, PM2.5, 03, N02, S02
PM2.5, BC, 03, N02, S02
PMzj;, PM10, EC, 03, N02, S02,
so7~
PM10, 03, N02, S02
ARRHYTHMIAS (VIA ECG)
Sarnatetal.(2006,
090489)
Berger et al. (2006,
098702)
Steubenville, OH
(n = 32)
Erfurt, Germany
(n = 57)
Atrial or
ventricular
tachycardia
Atrial or
ventricular
tachycardia
98th%: 1.42
99th%:1.81
(24 h)
NA
Mean: 0.2 (24 h)
Range: 0.1, 1.5
Mean: 0.45 (24 h)
Min, Med, Max
0.10,0.38,1.68
PM2.5, 03, N02, S02, S042", EC
PM10, PM2.5, N02, NO,S02, UF
NA: Not Available
3 Includes range across individual monitors in study site; AQS data available for U.S. studies only.
5.2.1.4. Cardiac Arrest
Cardiac arrest refers to the abrupt loss of heart function due to failure of the heart to contract
effectively during systole, which can lead to sudden cardiac death if not treated immediately. Very
January 2010
5-18
-------
few studies have investigated the association between ambient CO exposure and the risk of cardiac
arrest, and none reported a significant association between increased CO exposure and the
occurrence of cardiac arrest.
Two studies (Levy et al, 2001, 017171: Sullivan et al., 2003, 043156) were evaluated that
examined the association between ambient CO and cardiac arrest. Both studies were conducted in
Seattle, WA, using a case-crossover study design and found no association between short-term
exposure to CO and cardiac arrest.
These studies examined air pollution exposures for black smoke particles (BSP), PMi0, SO2,
and CO. The correlation coefficient for PMi0 and CO was 0.8 in both studies. The first of these
studies examined paramedic-attended out-of-hospital primary cardiac arrests among 362 cases
(1998-1994) in Seattle and King County, WA, whereby lags of 0-5 days were analyzed (Levy et al.,
2001, 017171). There was no indication of association between CO and out-of-hospital primary
cardiac arrest (RR 0.99 [95% CI: 0.83-1.18]). The second of these studies examined out-of-hospital
primary cardiac arrest for a 10-yr period (1985-1994) among subjects within a health organization
database (the Group Health Cooperative of Puget Sound), whereby 0- through 2-day lags were
analyzed (Sullivan et al., 2003, 043156). The relative risk of primary cardiac arrest was 0.95
(95% CI: 0.85-1.05; lag 0).
5.2.1.5. Myocardial Infarction
As previously stated, MI is commonly referred to as "heart attack" and is another cardiac
outcome that has received limited attention within the area of air pollution research. Only one study
has investigated the association between short-term changes in ambient CO and the onset of MI.
Peters and colleagues (2001, 016546) employed a case-crossover study design to analyze short_term
exposures (0-5 h and 0-5 days before the onset of MI) to particles (PMi0, PM2.5, PMi0_2.5, BC) and
gases (CO, O3, NO2, SO2) among 772 patients with MI in the greater Boston area. While all
pollutants showed positive associations with the onset of MI, only PM2 5 reached statistical
significance with the main exposure period (2 h before the onset) (OR for CO: 1.22 [95% CI:
0.89-1.67]).
5.2.1.6. Blood Pressure
Only two studies have investigated whether short-term exposure to CO influences BP. The
earlier of these two studies examined BP among 2,607 men and women aged 25-64 yr who
participated in the Augsburg, Germany, MONICA study (Ibald-Mulli et al., 2001, 016030).
Exposures to ambient TSP, SO2 and CO (from one monitor in the center of the city) during the same
day as the BP reading and an average over the 5 days prior were examined. Results showed that
ambient CO had no association with BP.
Similarly, the second of these studies extracted baseline and repeated measures of cardiac
rehabilitation data from a Boston, MA, hospital for 62 subjects with 631 visits and analyzed ambient
air pollution exposures (with particular focus on PM2 5) averaged over various periods up to 5 days
before the visit (Zanobetti et al., 2004, 087489). While results showed significant associations
between increased BP and ambient PM2 5, SO2, O3, and BC, there was no significant effect for CO
(results not presented quantitatively).
5.2.1.7. Vasomotor Function
Gaseous pollutants, including SO2, NO and CO, were found to affect large artery endothelial
function among 40 healthy white male nonsmokers in Paris, France, whereas PM was found to
exaggerate the dilatory response of small arteries to ischemia (Briet et al., 2007, 093049). Changes in
amplitude of flow-mediated dilatation were highly dependent on changes in 5-day lag concentrations
of SO2, NO and CO, but not NO2, PM2.5 or PM10. The effect attributed to CO ((3 coefficient: -0.68
[95% CI: -1.22 to -0.15]) was the smallest in magnitude when compared to those for SO2 and NO,
but overall the effect estimates were similar and all were statistically significant. Similarly, PM2 5,
PMio, NO2 and CO were positively correlated with small artery reactive hyperemia, and the effect
attributed to CO was the smallest in magnitude when compared to those for PM2 5, PMi0, and NO2;
but overall, the effect estimates were similar and all were statistically significant.
January 2010 5-19
-------
5.2.1.8. Blood Markers of Coagulation and Inflammation
Several studies have investigated the association between ambient CO and various blood
markers related to coagulation and inflammation. The main endpoints analyzed have been plasma
fibrinogen, B-type natriuretic peptide (BNP), endothelial function, Factor VII, C-reactive protein
(CRP), prothrombin, intercellular adhesion molecule (ICAM-1), and white blood cell count (WBC).
Delfino et al. (2008, 156390) measured blood plasma biomarkers in a panel of 29 nonsmoking,
elderly subjects with a history of CAD living in retirement communities in the Los Angeles, CA, air
basin, in order to identify associations with systemic inflammation. The blood plasma biomarkers
included CRP, fibrinogen, tumor necrosis factor-a (TNF- a) and its soluble receptor-II (sTNF-RII),
interleukin-6 (IL-6) and its soluble receptor (IL-6sR), fibrin D-dimer, soluble platelet selectin
(sP-selectin), soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble ICAM-1, and
myeloperoxidase (MPO). Overall, there were statistically significant associations for many of the
biomarker and pollutant combinations, with some of the strongest effects for CRP, IL-6 and sTNF-
RII with indoor and outdoor concentrations of NO2 and CO. Only the outdoor concentrations
indicated an effect of PM for these three biomarkers of inflammation. There was weak evidence for
an effect of outdoor and indoor CO on the biomarker of platelet activation (sP-selectin), and for an
effect of many of the air pollutants examined on fibrinogen, TNF- a, sVCAM-1, sICAM-1, and
MPO. Parameter estimates for fibrin D-dimer were close to zero for most models. Overall, the
results suggest that traffic related pollutants, including PM2.5, UFPs, OC and CO, lead to increases in
systemic inflammation and platelet activation in elderly people with a history of CAD.
Delfino et al. (2009, 200844) added a second year of data from 31 additional subjects to data
used in their previous analysis of 29 subjects (Delfino et al., 2008, 156390). This updated panel
study of 60 elderly individuals with CAD investigated the relationship of air pollutants to changes in
circulating biomarkers of inflammation, platelet activation and antioxidant capacity. The updated
analysis focused on the biomarkers that were most informative in the previous analysis (Delfino et
al., 2008, 156390) and included IL-6, TNF-a, sTNF-RII, CRP, and sP-selectin. Additionally, frozen-
thawed erythrocyte lysates were assayed spectrophotometrically for activities of two antioxidant
enzymes, glutathione peroxidase-1 (GPx-1) and copper-zinc superoxide dismutase (Cu,Zn-SOD).
Hourly outdoor home-air pollutants were measured over 9 days before each blood draw. There was
evidence for an association of CO with IL-6, P-selectin, TNF-RII and CRP, but not for TNF- a,
Cu,Zn-SOD, or GPx-1. Many positive associations were found for IL-6, sP-selectin, sTNF-RII,
TNF- a, and CRP with markers of traffic-related air pollution (EC, OC, BC, NOX, and CO),
confirming the earlier finding that traffic related pollutants may lead to increases in systemic
inflammation and platelet activation in elderly people with a history of CAD.
Circulating levels of BNP are directly associated with cardiac hemodynamics and symptom
severity in patients with heart failure and serve as a marker of functional status. Wellenius et al.
(2007, 092830) examined the association between BNP levels and short-term changes in ambient air
pollution levels among 28 patients with chronic stable heart failure and impaired systolic function.
The authors reported no association between CO along with the other pollutants examined and
measures of BNP at any lag.
Pekkanen et al. (2000, 013250) examined the association between daily concentrations of air
pollution and concentrations of plasma fibrinogen measured among 4,982 male and 2,223 female
office workers in Whitehall, London, U.K., between September 1991 and May 1993. Plasma
fibrinogen data were linked to ambient exposure to BS, PMi0, O3, NO2, SO2, and CO, where the
exposures were derived from data recorded at 5 fixed sites across London. There was a high
correlation between levels of CO andNO2 (r = 0.81) and more moderate correlations of CO with
PMio (r = 0.57) and SO2 (r = 0.61). The pollution data on the same day when the blood sampling was
done (lag 0) and on the 3 previous days (lags 1-3) were analyzed. Results showed that ambient CO at
all lags was associated with an increase in plasma fibrinogen. Results were similar for NO2, while all
other pollutants were not associated with an increase in plasma fibrinogen.
Liao et al. (2005, 088677) examined associations between various air pollutants and
hemostatic and inflammatory markers (fibrinogen, factor VIII-C, von Willebrand factor, serum
albumin, WBC) among 10,208 middle-aged males and females from the ARIC study. Exposure
estimates for ambient PMi0, NO2, SO2, O3 and CO were calculated for days 1-3 prior to the blood
sampling. A 0.5 ppm increment in 24-h CO concentration was significantly associated with
0.015 g/dL decrease in serum albumin among persons with a history of cardiovascular disease
(CVD). CO was not associated with other hemostatic or inflammatory factors.
January 2010 5-20
-------
In Israel, Steinvil et al. (2008, 188893) examined WBC, fibrinogen, and CRP among 3,659
study subjects enrolled in the Tel-Aviv Sourasky Medical Center inflammation survey, in which
subjects lived <11 km from an ambient air pollution monitor. Air pollution data in the form of PMi0,
NO2, SO2, O3, and CO were derived from data recorded at fixed sites. The correlation coefficients
were high between CO and NO2 (r = 0.86) and PMi0 (r = 0.75). Exposures for lag days 0-7 were
analyzed, and ambient CO had a negative effect on fibrinogen only among males. Negative
associations were reported for lag 0 (e.g., same day) and lags 2-5 with the decrease in fibrinogen
ranging from -5.5 mg/dL to -9.8 mg/dL per 0.5 ppm increase in 24-h CO concentration. A similar
negative effect for CO was observed on WBC among males only. The average CO exposure over the
week prior to the sampling yielded the largest reduction in WBC (-263 cells/(iL).
In a German study, Riickerl and colleagues (Ruckerl et al., 2006, 088754) recruited 57
nonsmoking male patients with CHD who were scheduled for 12 subsequent clinical visits where
samples of blood were collected. The authors tested the primary hypothesis that CRP would increase
in association with a rise in air pollution levels. CRP is an acute phase protein that increases during
inflammatory processes in the body. Other markers of inflammation (serum amyloid A [SAA]), cell
adhesion (E-selectin, von Willebrand factor antigen [vWF], ICAM-1), and coagulation (fibrinogen,
factor VII [FVII], prothrombin fragment 1+2) were also examined. Ambient air pollution in the form
of PMio, UFP, EC, NO2, and CO was monitored at one central site, and a 24-h avg immediately
preceding the clinic visit (lag 0) and up to 5 days (lags 1-4) was calculated for each patient. For CRP,
the odds of observing concentrations above the 90th percentile (8.5 mg/L) were 2.41
(95% CI: 1.23-5.02) in association with a 0.5 ppm increase in 24-h CO concentration (lag 2). CO
concentration during lags 1 and 2 was associated with observing ICAM-1 concentrations above the
90th percentile (OR: 2.41 [95% CI: 1.49-4.04]; OR: 3.17 [95% CI: 1.77-6.11], respectively). CO
concentration (lag 0-3) was associated with a decrease in FVII.
A similar study by Ruckerl and colleagues (2007, 156931) was conducted among 1,003 MI
survivors across six European cities (Athens, Greece; Augsburg, Germany; Barcelona, Spain;
Helsinki, Finland; Rome, Italy; Stockholm, Sweden). The study compared repeated measurements of
interleukin-6 (IL-6), CRP and fibrinogen with concurrent ambient levels of air pollution (particle
number count [PNC], PMi0, PM25, NO, NO2, O3, SO2, CO) from fixed sites across each city. Lags
0-1 and the 5-day mean prior to the blood sampling were analyzed and ambient CO was not
associated with any of the inflammatory endpoints.
Baccarelli et al. (2007, 090733) recruited 1,218 healthy individuals from the Lombardia region
in Italy and assessed whether blood coagulability is associated with ambient air pollution. The main
blood coagulability endpoints of interest were prothrombin time (PT) and activated partial
thromboplastin time (APTT), which are measures of the quality of the coagulation pathways,
assuming that, if shortened these measures would reflect hypercoagulability. Air pollution data
(PMio, O3, NO2, and CO) were obtained from 53 fixed stations across the Lombardia region, which
was divided into 9 different study areas, and a network average for each pollutant was calculated
across the available monitors within each of the 9 study areas. The analyses examined air pollution at
the time of the blood sampling, as well as averages for the 7 days prior and 30 days prior. Results
showed that ambient CO at the time of blood sampling was associated with a decrease in PT
(coefficient = -0.11 [95% CI: -0.18 to -0.05], p < 0.001), indicating hypercoagulability. However,
PM10 and NO2 at the time of blood sampling were also associated with a decrease in PT and results
from multipollutant models were not reported. Acute phase reactants such as fibrinogen and
naturally occurring anticoagulants such as antithrombin, protein C and protein S were examined and
none were associated with ambient air pollution.
Rudez et al. (2009, 193783) collected 13 consecutive blood samples within a 1-yr period and
measured light-transmittance platelet aggregometry, thrombin generation, fibrinogen and CRP in
40 healthy individuals in Rotterdam, The Netherlands. In general, air pollution increased platelet
aggregation as well as coagulation activity but had no clear effect on systemic inflammation.
Specifically, there were notable associations between maximal aggregation and CO, NO and NO2
and between late aggregation and CO. The effects for CO were the highest in magnitude and
persisted over most of the lag times investigated. There also was evidence of an increase in
endogenous thrombin potential and peak thrombin generation associated with CO, NO, NO2 and O3,
but no clear associations between PM10 and peak height or lag time of thrombin generation. In
addition, there was no evidence for an effect of any of the air pollutants examined on CRP or
fibrinogen levels. These prothrombotic effects may partly explain the relationship between air
pollution and the risk of ischemic cardiovascular disease.
January 2010 5-21
-------
Ljungman et al. (2009, 191983) investigated the effect of CO and NO2 on inflammation in
certain genetic subpopulations of MI survivors. Specifically, they examined whether IL-6 and
fibrinogen gene variants could affect plasma IL-6 response to CO or NO2. The study included
955 MI survivors from 6 European cities. This study provides evidence of gene-environment
interaction where IL-6 and fibrinogen gene polymorphisms modified the effects of CO and NO2 on
IL-6 levels in this panel of subjects with existing cardiovascular disease. Subjects with the
homozygous major allele genotypes for all 3 IL-6 polymorphisms examined showed larger IL-6
responses to increased CO, and there was evidence of a genetic interaction with NO2 for one of the
polymorphisms. Subjects with the homozygote minor allele genotype for one fibrinogen
polymorphism showed both a larger and clearer effect modification for the IL-6 response to
increased CO compared to the IL-6 polymorphisms. Similar magnitudes of effect modification were
seen for NO2, but the effect modification pattern was not statistically significant. A second fibrinogen
polymorphism did not modify the response to air pollution. Overall, this study provides evidence for
the influence of CO on IL-6 levels in subjects with genetic polymorphisms of the IL-6 and
fibrinogen genes. In this study, 16% of the subjects had a polymorphism combination that resulted in
a statistically significant gene-gene-environment interaction potentially implicating a higher risk of
health effects from air pollution in these patients with ischemic heart disease.
In summary, a growing number of studies provide some evidence of a link between CO
exposure and blood markers of coagulation and inflammation. Further studies are required to
determine whether the prothrombotic effects characterized by many of the blood markers may partly
explain the relationship between CO and the risk of ischemic cardiovascular disease. The results of a
recent gene-gene-environment interaction study are particularly interesting. Table 5-6 summarizes
the reviewed studies. Due to the heterogeneity of endpoints (see column "cardiac endpoint" in Table
5-6), these studies do not lend themselves to a quantitative meta-analysis or inclusion in a summary
figure.
January 2010 5-22
-------
Table 5-6. Summary of studies investigating the effect of CO exposure on blood markers of
coagulation and inflammation.
Study
De,finoeta,,2008,156390)
Location rnn™n*t\nn* co Concentrations
Cardiac Endpoint ,°"°Tn«s" in Reported by Study Copollutants
Sample Size ppm Authors in ppm
CRP, fibrinogen, TNF-
: 2.9
slCAM-1, MPO
Outdoor Mean: 0.71 (1 h) 03, N02, EC, OC, BC,
PMo.25, PM0 25-2.5,
Indoor Mean: 0.78 (1 h) PM25-io
Delfino et al. (2009, 2008441
Wellenius et al. (2007,
0928301
Pekkanen et al (2000, 0132501
Liao et al (2005, 0886771
Los Angeles CA
(n=60)
Boston, MA
(n=28)
London, U.K.
(n = 7205)
US
(n = 1 0.208)
CRP, TNF-a, IL-6, sP-
selectin, sTNF-RII,
Cu,Zn-SOD, GPx-1
BMP
Plasma fibrinogen
Fibrinogen, VII-C,
WBC, albumin, vWF
NA
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)
Outdoor mean: 0.50-0.58 (1h)
Mean: 0.44 (24 h)
Mean: 1.22 (24 h)
10th, 50th, 90th, Max:
0.61, 1.04,2.0,8.61
Mean: 1.4 (24 h)
03, N02,NOX, EC, OC, BC,
SOC, PN, PM025, PM025-25,
PM2.5-10
PM25, S02, N02, 03, BC
PM,o, BS, 03, N02, S02
PM10, 03, N02, S02
Steinvil et al (2008,1888931
Tel-Aviv, Israel
(n = 3659)
CRP, fibrinogen,
WBC
NA
Mean: 0.8
25th, 50th, 75th percentiles: PM10, 03, N02, S02
0.7, 0.8, 1.0
Ruckerl et al (2006, 0887541
Ruckerl et al (2007, 1569311
Baccarelli et al (2007, 0907331
Rudez et al. (2009, 1937831
Ljungman et al. (2009,
1919831
Erfurt, Germany
(n = 57)
Six European cities
(n = 1003)
Lombardia Region,
Italy
(n = 1218)
Rotterdam, The
Netherlands
(n=40)
Six European cities
(n=955)
CRP, SAA, cell
adhesions and NA
coagulation
IL-6, CRP, fibrinogen NA
PT, APTT, fibrinogen, NA
anticoagulants
Platelet aggregation,
thrombin generation, NA
fibrinogen, CRP
IL-6 and fibrinogen NA
polymorphisms
Mean: 0.45 (24 h)
PM,o, 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
PM10, 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
3 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 (U.S. EPA, 2000, 000907) there have been a number of studies that
investigated the effect of ambient CO on hospital admissions and ED visits for CVD. Some of these
studies have focused solely on one specific CVD outcome, and these studies are discussed first. The
subsequent sections provide a discussion of the studies that investigated hospital admissions and ED
visits for all CVD outcomes (e.g., nonspecific) or a variety of specific CVD outcomes.
January 2010
5-23
-------
Coronary Heart Disease
Ischemic heart disease (IHD), also known as CHD, is caused by inadequate circulation of the
blood to the heart muscle, which is a result of the coronary arteries being blocked by cholesterol
deposits or by vasospasm. CHD can lead to sudden episodes such as MI or death, as well as chronic
conditions such as angina pectoris (chest pain).
Ischemic Heart Disease
A number of studies have focused directly on hospitalizations for IHD. There is a lot of
variation among these studies with regard to methods employed and results reported. It should be
noted that within these studies IHD included MI and angina pectoris (ICD-9 codes 410-414; ICD-10
codes 120, 121-123, 124). Amulticity time-series study was conducted to estimate the risk of CVD
hospitalization associated with short-term CO exposure in 126 U.S. urban counties from 1999-2005
for over 9 million Medicare enrollees 65 yr old and older (Bell et al., 2009, 193780). The analyses
yielded positive associations between same-day CO concentration adjusted for NO2 concentration
and increased risk of hospitalization for IHD 1.004 (95% PI: 1.001-1.007). Cause-specific effect
estimates were not presented for CO alone (without adjustment for NO2).
Mann and colleagues (2002, 036723) investigated the modifying effect of secondary diagnosis
of arrhythmia and congestive heart failure (CHF) on the relationship between hospital admissions for
IHD (ICD-9: 410-414) and ambient air pollutants for the period of 1988-1995 in southern California.
There were 54,863 visits analyzed and a 0.75 ppm increase in 8-h max CO concentration was
associated with a 2.69% (95% CI: 1.21-4.19) increase in same-day IHD admissions among persons
with a secondary diagnosis of CHF, a 2.23% (95% CI: 1.35-3.13) increase among persons with a
secondary diagnosis of arrhythmia, and a 1.21% (95% CI: 0.49-1.94) increase among persons
without either secondary diagnosis. Of all the pollutants examined (PMi0, NO2, O3, CO), only NO2
showed positive effects estimates similar in magnitude to CO. Although no multipollutant models
were analyzed, a moderate to high correlation between CO and NO2 was found across the seven
regions ranging from 0.64 to 0.86. This study indicated that people with IHD and underlying CHF
and/or arrhythmia represent a potentially susceptible population relative to the effects of ambient air
pollution.
By using a time-series approach, ED visits for IHD (ICD-9: 410-414) in Montreal, Canada,
(1997-2002) were examined in relation to ambient CO concentrations (lags 0 and 1) (Szyszkowicz,
2007, 193793). A total of 4,979 visits were analyzed, and results showed significant positive effects
with a 0.5 ppm increase in 24-h CO concentration (lag 0), resulting in a 14.1% (95% CI: 5.8-20.6)
increase in daily ED visits among all patients. Stratified analyses showed that this effect was mostly
among male patients (19.8% [95% CI: 9.2-31.6]). NO2 was the only other pollutant examined, and it
too showed significant positive associations with ED visits for IHD for same-day exposure; however,
no multipollutant models were examined.
Lee and colleagues (2003, 095552) examined daily counts of hospital admissions for IHD in
Seoul, Korea, for the period from December 1997 to December 1999. Single-day lags 0-5 were
analyzed, and the lag period with the strongest association for each pollutant was presented by the
authors. For CO, lag 5 showed the strongest effect, with a 1 ppm increase in 1-h maximum (max)
CO concentration associated with a daily increase in the number of hospital admissions for IHD;
however, this was only among patients 64+ yr of age (RR: 1.07 [95% CI: 1.01-1.13]). All other
pollutants (PMio, O3, NO2) except SO2 showed similar significant effects and in a copollutant model
with PM10 the CO effect was somewhat attenuated (RR 1.04 [95% CI: 0.98-1.11]).
Other studies have examined hospital admissions for IHD while investigating a broad group of
CVD outcomes. A study was conducted in Atlanta, GA, where over 4 million ED visits from
31 hospitals for the period 1993-2000 were analyzed (Study of Particles and Health in Atlanta
[SOPHIA]). Several articles have been published from this research, with two examining
cardiovascular admissions in relation to CO concentrations. The first of these (Metzger et al., 2004,
044222) used a time-series design and analyzed a 3-day ma over single-day lags 0-2 as the a priori
lag structure. Although of borderline statistical significance, CO was positively associated with an
increase in ED visits for IHD (RR 1.016 [95% CI: 0.999-1.034] per 1 ppm increase in 1-h max CO
concentration).
The second of these reports (Peel et al., 2007, 090442) examined the association of ambient air
pollution levels and cardiovascular-related ED visits with and without specific secondary conditions
January 2010 5-24
-------
(e.g., comorbidity). Within a time-stratified case-crossover design using the same lag structure
previously mentioned, the main results showed that a 1 ppm increase in 1-h max CO concentration
was associated with an increase in IHD among those without diabetes (OR: 1.023
[95% CI: 1.004-1.042]), and without CHF (OR: 1.024 [95% CI: 1.006-1.042]).
Two Australian studies have also examined associations between ambient CO concentrations
and increased hospital admissions for various CVD outcomes. The first of these studies (Barnett et
al., 2006, 089770) analyzed data from five of the largest cities in Australia (Brisbane, Canberra,
Melbourne, Perth, Sydney) and two New Zealand cities (Auckland, Christchurch) for the period
1998-2001. A time-stratified case-crossover design was employed, and the age groups of 15-64 yr
and > 65 yr were analyzed for the 0-1 lag period. The pooled estimates across all cities showed that a
0.75 ppm increase in 8-h max CO concentration was associated with a 1.9% (95% CI: 0.7-3.2)
increase in admissions for IHD among the elderly group (> 65 yr). No association was observed for
the younger age group.
The second of the Australian studies (Jalaludin et al., 2006, 189416) examined ED visits for
CVD outcomes in the elderly (65+ yr) in Sydney for the period 1997-2001. Using a time-series
approach, single-day lags of 0, 1, 2, 3, and an average over lags 0 and 1 were examined. A 0.75 ppm
increase in 8-h max CO concentration (lag 0) was associated with increases in IHD emergency
department visits of 3.1% (95% CI: 1.3-4.9).
Angina Pectoris
In the current literature, only one study was identified that focused solely on angina pectoris
as an endpoint. Admissions data for angina pectoris were collected from 25 academic hospitals in
Tehran, Iran, and linked to ambient air pollution for the period of 1996-2001 (Hosseinpoor et al.,
2005, 087413). Using a time-series approach, single-day lags of 0-3 were analyzed and a 0.5 ppm
increase in 24-h avg CO concentration at lag 1 was associated with increased hospital admissions for
angina (OR: 1.005 [95% CI: 1.003-1.007]). This result persisted in a multipollutant model that also
included NO2, PM10, and O3 (OR: 1.005 [95% CI: 1.001-1.008]).
Myocardial Infarction
Linn et al. (2000, 002839) examined the association between ambient air pollution and
hospital admissions for cardiopulmonary illnesses in metropolitan Los Angeles for the years
1992-1995. Using a time-series approach, a 0.5 ppm increase in same-day 24-h avg CO
concentration was associated with a 2.0% increase in MI hospital admissions among people aged
>30 yr. When the analyses were stratified by season, no significant effects were observed (no
quantitative seasonal effects reported).
A time-series study in Denver, Colorado, investigated daily hospital admissions for various
CVD outcomes among older adults (>65 yr) across 11 hospitals (Koken et al., 2003, 049466). Data
between July and August for the period 1993-1997 were analyzed. Single-day lags 0-4 were
examined and CO showed no association with hospital admissions for MI (quantitative results were
not reported).
As part of the Health Effects of Air Pollution among Susceptible Subpopulations (HEAPSS)
study, Lanki et al. (2006, 089788) investigated the association between traffic-related exposure to air
pollutants and hospitalization for first acute myocardial infarction (AMI). Data were collected from
five European cities with either AMI registers (Augsburg, Barcelona) or hospital discharge registers
(Helsinki, Rome, Stockholm). Correlation coefficients between CO and NO2 ranged from 0.43 to
0.75 across the five cities, and between CO and PMi0 the range was 0.21 to 0.56. A total of 26,854
hospital admissions were analyzed, and pooled estimates from all five cities showed that there was a
weak positive association with AMI hospital admissions and 24-h avg CO concentrations at lag 0
(RR: 1.014 [95% CI: 1.000-1.029] per 0.5 ppm increase), but more so when only using data from the
three cities (Helsinki, Rome, Stockholm) with hospital discharge registers (RR: 1.020
[95% CI: 1.003-1.035] per 0.5 ppm increase). When analyses were stratified by fatality and age,
results showed that the CO effect was significantly associated with fatal AMI among the <75-yr age
group (RR: 1.080 [95% CI: 1.017-1.144]) and with non-fatal AMI in the > 75-yr age group (RR:
1.044 [95% CI: 1.011-1.076]).
Further analyses within the HEAPSS cohort were conducted using the event of cardiac
readmission among the first MI survivors (n = 22, 006) (Von Klot et al., 2005, 088070). The
readmissions of interest were those with primary diagnosis of AMI, angina pectoris, dysrhythmia,
January 2010 5-25
-------
and heart failure that occurred at least 29 days after the index event. Single-day lags 0-3 were
examined, and pooled estimates from all 5 cities showed that a 0.5 ppm increase in same-day (lag 0)
CO was associated with an increase in cardiac (e.g., any of the diagnoses) readmissions (RR: 1.041
[95% CI: 1.003-1.076]); this persisted in two-pollutant models that included either PMi0 or O3.
Correlation coefficients with CO ranged from 0.21 to 0.57 for PMi0 and 0.44 to 0.75 for NO2.
A study in Rome, Italy, also found an association between ambient CO and hospitalizations for
first-episode MI among 6,531 subjects (January 1995-June 1997) (D'Ippoliti et al, 2003, 074311). A
case-crossover design with stratification of time into separate months was used to select referent
days as the days falling on the same day of the week within the same month as the index day. CO
concentration was positively associated with lag 2 (OR: 1.019 [95% CI: 1.001-1.037]). The other
pollutants analyzed were NO2 and TSP, both of which exhibited a significant positive effect at lag 0.
TSP also showed a significant positive effect at lag 0-2 and, when entered into a model with CO, the
CO effect did not persist.
The previously mentioned Australian and New Zealand study that analyzed data from seven
cities (Brisbane, Canberra, Melbourne, Perth, Sydney, Auckland, and Christchurch) for the period
1998-2001 also reported an association between CO and MI hospital admissions (Barnett et al.,
2006, 089770). The pooled estimates across all cities showed that a 0.75 ppm increase in 8-h max
CO concentration was associated with a 2.4% (95% CI: 0.6-4.1) increase in admissions for MI, but
only among older adults (> 65 yr). Table 5-7 shows a summary of the CHD hospital admission
studies that examined CO exposures.
In summary, the majority of studies reported significant increases in the daily number of
admissions for IHD, angina and MI in relation to CO exposures. In studies that stratified by age
groups and/or sex, the effects were larger among the elderly and males. Among the different lag
periods being examined, the associations were more commonly observed with same day CO (lag 0)
or an average over the same day and previous day (lag 0-1). Figure 5-2 shows the effect estimates
associated with daily admissions for various forms of CHD from selected studies.
January 2010 5-26
-------
Study
Metzger et al. (2004, 044222)
Peel etal. (2007,090442)
Mann etal.(2002. 036723)
Barnett etal. (2006.089770)
Jalaludinetal. (2006, 189416)
Szyszkowicz (2007, 193793)
Lee etal. (2003.095552)
von Klot etal. (2005. 088070)
Hosseinpoor et al. (2005, 087413)
Linn et al. (Linn et al., 2000, 002839)
Barnett etal. (2006. 089770)
Lanki etal. (2006. 089788)
von Klot etal. (2005,088070)
D'lppoliti etal. (2003. 074311)
Location Lag
Atlanta, GA 0-2
Atlanta, GA 0-2
California, US 0-3
California, US 0-3
California, US 0-3
Australia, New Zealand 0-1
Australia, New Zealand 0-1
Sydney, Australia 0-1
Montreal, Canada 0
Montreal, Canada 0
Montreal, Canada 0
Montreal, Canada 0
Montreal, Canada 0
Montreal, Canada 0
Seoul, Korea 5
Seoul, Korea 5
Multicily, Europe 0
Tehran, Iran 1
Los Angeles, CA 0
Australia, New Zealand 0-1
Australia, New Zealand 0-1
Multicity, Europe 0
Multicity, Europe 0
Multicity, Europe 0
Multicity, Europe 0
Multicity, Europe 0
Multicity, Europe 0
Rome, Italy 0-2
Rome, Italy 0-2
Rome, Italy 0-2
Rome, Italy 0-2
Age
15-64yr
65+ yr
65+ yr
All ages
All ages
All ages
65+ yr
65+ yr
65+ yr
All ages
64+ yr
35+ yr
All ages
15-64yr
65+ yr
35+ yr
<75yr
<75yr
75+ yr
75+ yr
35+ yr
18+yr
18-64yr
65-74 yr
75+ yr
Group/Outcome Effect Estimate (95% Cl)
sCHF
sARR
j
Males
Females — i
Males and Females
Males
-• —
— i
All year
j
All cities
Nonfatal 1
Fatal
Nonfatal
Fatal ->
j
—
— i
«- IHD
I±
•
-»-
•
•-
«•
-•-
i •
— • —
r» — Angina
»
• Ml
•-
-•-
«-
t
•
— •—
!-•
I •
-•-
••—
[•—
1 I I
0.8 1.0 1.2 1.4
Relative Risk
Figure 5-2. Summary of effect estimates (95% confidence intervals) associated with hospital
admissions for various forms 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.
January 2010
5-27
-------
Table 5-7. Summary of CHD hospital admission studies.3
Endnoints Laos Upper CO CO Concentrations
Study Location cnapomcs copollutants c " . Concentrations from Reported by Study
Examined Examined AQS°inppm 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, 0897881b
von Klot et al. (2005,
0880701b
D'lppoliti etal. (2003,
07431 1lb
Hosseinpoor et al. (2005,
08741 31b
Southern
California IHD
(1988-1995)
Montreal, Can
IHD
(1997-2002)
Seoul, Korea
IHD
(1997-1999)
5 European cities M|(first
(1992-2000) acute)
5 European cities M|]Angina]
(1992-2001) Cardiaca
Rome, Italy
Ml
(1995-1997)
Tehran, Iran
Angina
(1996-2001)
PM10, N02, 03 0,1,2, 2-4ma
N02 0,1
PM10,N02,S02,
03
PM10, N02, 03, nl9,
PNC U'V'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 U'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 = 2.6
Mean: highest city was
Rome: 1.9 (24 h)
Mean: 3.8 (24 h)
Mean: 9.4 (24 h)
STUDIES THAT EXAMINED CHD AMONG 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 |un MI AM
. . _, . . IHU, Ml, All
New Zealand CVD CA
(1998-2001) stroke
Sydney, |[_|Q An
Australia CVD' ^n^e
(1997-2001) Cardiac
Los Angeles,
CA Ml, All CVD,
CHF, CA, OS
(1992-1995)
Denver, CO MI.CAth,
PHD, CD,
(1993-1997) CHF
PM10, N02, S02, ,,.
03 °-2ma
PM10, N02, S02, n,
Os °-2ma
PM10, N02, 03 Lag 0-1
PM10,N02,S02, 0^30.!
O3 ....
PM10, N02, 03 0
PM10, N02, S02, 0 -i 2 3 4
03 ' ' ' '
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: 1.8(1 h)
Mean: 1.8(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)
3 Cardiac =AMI, angina, dysrhythmia, orHF; CA= Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhythmias; CHF= Congestive heart failure;
OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average.
bThese 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
and temperature.
c Includes range across individual monitors in study site; AQS data available for U.S. studies only.
NA: Not Available
PHD = Pulmonary heart disease;
assumes standard atmosphere
January 2010
5-28
-------
Stroke
A stroke is the result of either the blood supply to the brain being blocked (e.g., embolism),
which refers to an ischemic stroke (80% of strokes), or the occurrence of a burst blood vessel or
hemorrhaging, referred to as a hemorrhagic stroke. Hemorrhagic stroke has two main groupings;
intracerebral hemorrhagic stroke (10% of strokes), which is when a blood vessel in the brain leaks,
and subarachnoid hemorrhage (3% of strokes), which is bleeding under the outer membranes of the
brain. The third type of stroke is a transient ischemic attack (TIA) or ministroke, which has the same
early symptoms as a normal stroke but the symptoms disappear within 24 h, leaving no apparent
deficits. A limited number of air pollution studies have investigated hospital admissions for the three
main forms of stroke and generally report small, positive associations or no association with ambient
CO concentrations measured during lag periods between 0 and 3 days.
In the multicity time-series study conducted by Bell et al. (2009, 193780). the analyses yielded
small, positive associations between same-day CO concentration adjusted for NO2 concentration and
increased risk of hospitalization for cerebrovascular outcomes 1.005 (95% PI: 1.002-1.009). Cause-
specific effect estimates were not presented for CO alone (without adjustment for NO2).
A U.S. study across 9 cities investigated hospital admissions for ischemic and hemorrhagic
stroke among Medicare beneficiaries aged 65+ yr of age (155,503 ischemic and 19,314 hemorrhagic
admissions from the ED) (Wellenius et al., 2005, 088685). Single-day lags 0-2 were examined and
based on a pooled estimate, same-day CO (lag 0) was associated with an increase in ischemic stroke
admissions of 1.98% (95% CI: 0.86-3.12) per 0.5 ppm increase in 24-h CO concentration), but not
hemorrhagic stroke admissions (-1.14%, 95% CI: -3.40 to 1.18). All other pollutants examined
(PMio, NO2, SO2) were also associated with an increase in ischemic stroke admissions but not
hemorrhagic stroke admissions.
Villeneuve and colleagues (2006, 090191) studied ED visits for hemorrhagic strokes, acute
ischemic strokes and transient ischemic attacks among individuals 65+ yr of age at 5 hospitals within
the Edmonton, Canada, area between April 1992 and March 2002 (12,422 visits). Within a time-
stratified case-crossover design, the analyses were stratified by two seasonal groups (October-March
and April-September). CO was found to only have an effect on ischemic stroke during April-
September (OR: 1.32 [95% CI 1.09-1.60] per a 0.5 ppm increase in 24-h CO concentration) for a
3-day avg across lags 0-2. CO had no effect on any other stroke subtype. In two-pollutant models the
CO effect on ischemic stroke persisted after controlling for PMi0, PM2 5, SO2, and O3.
In Kaohsiung City, Taiwan, CO averaged over lags 0-2 was associated with increased
admissions for stroke across 63 hospitals (Tsai et al., 2003, 080133). From 1997 through 2000 a total
of 23,179 admissions were analyzed, and on warm days (> 20°C) the odds ratios for primary
intracerebral hemorrhage and ischemic stroke were 1.39 (95% CI: 1.16-1.66) and 1.39
(95% CI: 1.25-1.53) respectively for a 0.5 ppm increase in 24-h CO concentration. For the same
increase in CO on cool days (<20°C) the odds ratios were 1.33 (95% CI: 0.38-2.55) for intracerebral
hemorrhage and 2.68 (95% CI: 1.59-4.49) for ischemic stroke. These results persisted in two-
pollutant models that included PMi0, SO2, and O3 but did not persist when controlling for NO2.
Earlier research conducted in metropolitan Los Angeles examined hospital admissions for
cardiopulmonary illnesses from 1992-1995 (Linn et al., 2000, 002839). Using a time-series
approach, a 0.5 ppm increase in 24-h CO concentration (lag 0) was associated with a 2.18%
(95% CI: 1.73-2.62) increase in occlusive (ischemic) stroke hospital admissions among people aged
>30 yr. When the analyses were stratified by season, there was a 1.8% (p=0.017) increase during
winter, a 4.55% (p=0.039) increase during summer, and a 1.6% (p=0.015) increase during fall
(results for spring were not reported).
A study in Taipei, Taiwan, analyzed 8,582 emergency admissions for cerebrovascular diseases,
hemorrhagic stroke, ischemic stroke, and all strokes during 1997-2002 (Chan et al., 2006, 090193).
Single-day lags 0-3 were analyzed, and a 0.75 ppm increase in 8-h max CO concentration (lag 2)
was associated with an increase in cerebrovascular diseases (OR: 1.03 [95% CI: 1.01-1.05]) and all
strokes (OR: 1.03 [95% CI: 1.01-1.05]). These results persisted in two- and three-pollutant models
that included O3 and PM10. There was no association with individual ischemic or hemorrhagic
stroke. CO was moderately correlated with PMio (r = 0.47) and PM2 5 (r = 0.44), and the correlation
was higher with NO2 (r = 0.77).
A time-series study that focused specifically on stroke hospital admissions conducted in Dijon,
France, did not report a significant association with ambient CO (Henrotin et al., 2007, 093270).
Hospital admissions for different types of first-ever stroke (e.g., ischemic, hemorrhagic) among
January 2010 5-29
-------
subjects >40 yr were analyzed for the period 1994-2004. A bidirectional case-crossover study design
was employed where single-day lags between 0-3 days were examined and CO had no significant
association for any lag. This was also the case when the analyses were stratified by gender and types
of ischemic stroke (large arteries, lacunar, cardioembolic, transient). Of all pollutants examined
(PMio, NOX, O3, SO2, CO), only O3 showed a significant effect.
Two Australian studies examined associations between ambient CO and hospital admissions
for various CVDs. The first of these studies analyzed data from five of the largest cities in Australia
(Brisbane, Canberra, Melbourne, Perth, Sydney) and two New Zealand cities (Auckland,
Christchurch) for the period 1998-2001 (Barnett et al., 2006, 089770). A time-stratified case-
crossover design was employed and the age groups of 15-64 yr and > 65 yr were analyzed for the
0-1 lag period (average over lag 0 and 1). The pooled estimates across all cities showed that CO had
no effect on stroke admissions (quantitative results not reported).
The second of the Australian studies examined ED visits for CVDs in older adults (65+ yr) in
Sydney for the period 1997-2001 (Jalaludin et al., 2006, 189416). Using a time-series approach,
single-day lags of 0-3 and an average over lags 0 and 1 (e.g., lag 0-1) were examined, and CO
showed no effect on stroke ED visits. When the analyses were stratified by cool and warm periods, a
0.75 ppm increase in 8-h max CO concentration during the cool period was associated with a 3.8%
(95% CI: 0.76-6.94) increase in stroke ED visits.
In summary, there was limited evidence that increased ambient CO concentrations might be
associated with hospital admissions for stroke. The largest positive effects came from the Taiwan
study in Kaohsiung (Tsai et al., 2003, 080133). with slightly larger effects during the warmer period
(>20°C). Similarly, in the Canadian study by Villeneuve and colleagues (2006, 090191). there was a
stronger effect during the warmer period (April-September). Studies in France and Australia reported
no association between ambient CO concentrations and increased hospital admissions or ED visits
for stroke. Figure 5-3 shows the effect estimates associated with daily admissions for stroke from
selected studies; Table 5-8 shows a summary of the stroke hospital admission studies that examined
CO exposures.
Study
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)
Villeneuve etal. (2006, 0901 91)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve etal. (2006, 0901 91)
Villeneuve et al. (2006, 090191)
Tsai etal. (2003,080133)
Tsai etal. (2003,080133)
Linn etal. (2000,002839)
Chan etal. (2006,090193)
Henrotin etal. (2007. 093270)
Wellenius etal. (2005. 088685)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Villeneuve et al. (2006, 090191)
Tsai etal. (2003. 0801 33)
Tsai etal. (2003,080133)
Chan etal. (2006,090193)
Henrotin et al. (2007, 093270)
Chan etal. (2006,090193)
Jalaludin etal. (2006, 189416)
Location
Multicity, US
Edmonton, Can
Edmonton, Can
Edmonton, Can
Edmonton, Can
Edmonton, Can
Edmonton, Can
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Los Angeles, CA
Taipei, Taiwan
Dijon, France
Multicity, US
Edmonton, Can
Edmonton, Can
Edmonton, Can
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Taipei, Taiwan
Dijon, France
Taipei, Taiwan
Sydney, Australia
Lag
0-2
0-2
0-2
0-2
0-2
0-2
0-2
0-2
0-2
0
1
3
0-2
0-2
0-2
0-2
0-2
0-2
1
1
2
0-1
Outcome/Group Effect Estimate (95% CI)
65+ yr
IS, 65+ yr, All Seasons
IS, 65+yr,Apr-Sep
IS. 65+ yr, Oct-Mar
CIS, 65+ yr, All Seasons
CIS, 65+ yr, Apr-Sep
CIS. 65+ yr, Oct-Mar
<20°C Temp
>20°C Temp
65+ yr
65+ yr, All Seasons
65+ yr, Apr-Sep
65+ yr Oct-Mar
<20°C Temp
>20°C Temp
•
•*•
1 — • —
«r
*
-•—
4
, -•-
»
i»
*
4
4}-
i— • —
-*
1 — • —
-if-
*
•
»
Ischemic Stroke
Hemorrhagic Stroke
Stroke (nonspecific)
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Relative Risk
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.
January 2010
5-30
-------
Table 5-8. Summary of stroke hospital admission studies.3
Study
Location
Type Of . Upper CO CO Concentrations
Stroke Copollutants examined Concentrations Reported by Study
Examined examinee from AQgc jn ppm Authors in ppm
STUDIES THAT FOCUSED SOLELY ON STROKE
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)
Tsai etal. (2003. 0801 33)
Chan etal. (2006. 0901 93)
Henrotin etal. (2007.093270)"
9 US cities
Isch, Hem
(1993-1999)
Edmonton, Can . . u__ T,A
(1992-2002) Isch, Hem, TIA
Kaohsiung, Taiwan
Isch, Hem
(1997-2000)
Taipei, Taiwan
All, Isch, Hem
(1997-2002)
Dijon, France
Isch, Hem
(1994-2004)
98th%'09-59 25th, 50th, 75th
PM10, N02, S02 0,1 , 2 percentiles:
99th%: 1.2-7.1 (24 h) 0.73,1.02,1.44
N02, S02, 03 0,1 , 0-2 NA
PM10,N02,S02, 0_2 NA
U3
PM10,N02,S02, Oi1_2i3 NA
PM10,NOX,S02, Oi1i2_3 NA
Mean: 0.8 (24 h)
Mean: 0.79 (24 h)
Mean:1.7(8h)
Mean: 0.59 (24 h)
STUDIES THAT EXAMINED STROKE AMONG OTHER CVDS
Linn etal. (2000, 002839)
Barnett etal. (2006. 089770)
Los Angeles, CA
(1992-1995)
Australia and New
Zealand
(1998-2001)
Isch
All
PM10,
PM,o,
N02,
N02,
03
03
LagO
Lag 0-1
98th%: 1.0-7.8
99th%: 1.1-8.3 (24 h)
NA
Mean: (24 h)
Winter 1.7, Spring 1.0,
Summer 1.2, Fall 2.1
Mean:(8h)
0.5-2.1
Jalaludin et al. (2006,189416)
Sydney, Australia
(1997-2001)
All
PM10,N02,S02, ^^ m
Mean: 0.82 (8h)
3 Isch = Ischemic; Hem = Hemorrhagic; TIA = transient ischemic attack
bThese 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.
c Includes range across individual monitors in study site; AQS data available for U.S. studies only.
NA: Not Available
Congestive Heart Failure
Heart failure (HF) is a condition in which the heart is unable to adequately pump blood to the
rest of the body. It does not refer to the cessation of the heart but more to the inability of the heart to
operate at an optimal capacity. HF is often called congestive heart failure (CHF), which refers to
when the inadequate pumping leads to a buildup of fluid in the tissues. The underlying causes of
CHF are hypertension, CAD, MI, and diabetes.
In the multicity time-series study conducted by Bell et al. (2009, 193780). the analyses yielded
positive associations between same-day CO concentration adjusted for NO2 concentration and
increased risk of hospitalization for HF (1.009 [95% PI: 1.005-1.012]). Cause-specific effect
estimates were not presented for CO alone (without adjustment for N02).
Wellenius and colleagues (2005, 087483) examined the rate of hospitalization for CHF among
55,019 Medicare recipients (> 65 yr) residing in Allegheny County, PA, during 1987-1999. A time-
stratified case-crossover design was employed and single-day lags of 0-3 were analyzed. A 0.5 ppm
increase in 24-h avg CO concentration on the same-day (lag 0) was associated with a 4.1%
(95% CI: 3.0-5.3) increase in the rate of hospitalization for CHF. This result persisted in copollutant
models that included PMi0, NO2, O3, and SO2. CO was moderately correlated with SO2 (r = 0.54)
and PMio (r = 0.57) and more highly correlated with NO2 (r = 0.70).
Another U.S. study recruited 125 patients diagnosed with CHF who were admitted to Johns
Hopkins Bayview Medical Center in Baltimore, MD (Symons et al., 2006, 091258). The patients
were interviewed after admission through the ED during their stays in overnight wards. The
January 2010
5-31
-------
interview was designed to collect information about symptom onset, health conditions, and factors
related to air pollution exposure. Various lag periods (single day and cumulative days 0-3) prior to
the onset of symptoms were analyzed and although the focus of this study was exposure to PM2.5, of
all the pollutants examined (PM2 5, CO, NO2, O3) only 8-h max CO concentration at lag 2 was
significantly associated with the onset of CHF symptoms (OR: 1.68 [95% CI: 1.28-2.80]).
Earlier research conducted in metropolitan Los Angeles, CA examined hospital admissions for
cardiopulmonary illnesses (1992-1995) (Linn et al, 2000, 002839). Using a time-series approach, a
0.5 ppm increase in same-day 24-h avg CO concentration was associated with a 1.25% increase in
CHF hospital admissions among people >30 yr. When the analyses were stratified by seasons, only
summer showed a significant increase (3.7%); however, the study did not report the results for the
other seasons.
A time-series study in Denver, CO, investigated daily admissions for various CVDs among
older adults (>65 yr) across 11 hospitals (Koken et al., 2003, 049466). Single-day lags 0-4 were
examined, and an increase of 0.5 ppm in 24-h avg CO concentration for lag 3 was associated with an
18% (95% CI: 0.2-39.3) increase in risk of hospitalization for CHF.
As stated earlier, a study was conducted in Atlanta, GA, where over 4 million ED visits from
31 hospitals for the period 1993-2000 were analyzed (Metzger et al., 2004, 044222). A time-series
design was used and a 3-day ma over single-day lags 0-2 as the a priori lag structure was analyzed.
Results showed that 1-h max CO concentration was not associated with an increase in ED visits for
CHF (RR: 1.010 [95% CI: 0.988-1.032] per 1 ppm increase). Peel et al. (2007, 090442) examined
the same cardiovascular-related effects among those with and without specific secondary conditions
(e.g., comorbidity) and found that 1-h max CO concentration was associated with an increase in ED
visits for CHF only among those with COPD (OR: 1.058 [95% CI: 1.003-1.115] per 1 ppm increase).
In Kaohsiung city, Taiwan, a study analyzed 13,475 admissions for CHF across 63 hospitals
for the period 1996 through 2004 (Lee et al., 2007, 090707). A 0.5 ppm increase in 24-h avg CO
concentration averaged over lag days 0-2 was positively associated with CHF hospital admissions on
cool days (<25°C) (OR: 1.70 [95% CI: 1.43-2.01]), with a slightly weaker effect on warm days
(>25° C) (OR: 1.32 [95% CI: 1.15-1.55]). These results persisted in two-pollutant models that
included PMi0, SO2, O3, and models with NO2 only on warmer days, not with NO2 on cooler days.
Study
Metzaer et al. (2004, 044222)
Peel etal. (2007, 090442)
Wellenius et al.(2005, 087483)
Linn etal. (2000,002839)
Koken etal. (2003, 049466)
Svmons etal. (2006,091258)
Lee etal. (2007,090707)
Lee etal. (2007, 090707)
Yana (2008, 157160)
Yana (2008, 157160)
Location
Atlanta, GA
Atlanta, GA
Pittsburgh, PA
Los Angeles, CA
Denver, CO
Baltimore, MD
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Taipei, Taiwan
Taipei, Taiwan
Lag
0-2
0-2
0
0
3
0-3
0-2
0-2
0-2
0-2
Group
65+ yr
65+ yr
<25°C Temp
>25°C Temp
<20°C Temp
>20°C Temp
Effect Estimate (95% CI)
f
»
j*
•
! — • —
* >
| _,
4»-
. . . , I -f- —
I I
1.4 1.8 2.2
0.2 0.6 1.0
Relative Risk
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.
A case-crossover analysis was undertaken to examine the association between levels of
ambient air pollutants and hospital admissions for CHF among individuals residing in Taipei,
Taiwan, from 1996 through 2004 (Yang, 2008, 157160). During the 9 yr of the study, there were
24,240 CHF hospital admissions for the 47 hospitals in Taipei. The analyses were stratified by
temperature, either warm days (>20°C; n = 2325 days) or cool days (<20°C; n = 963 days). The
number of CHF admissions was associated with concentrations of PMio, NO2, CO and O3 on warm
days, however on cool days, the positive effects on increased CHF admissions remained positive,
although the effects were diminished for NO2 and CO, and disappeared completely for PMi0 and O3
concentrations. In two-pollutant models, CO remained statistically significant after the inclusion of
o, SO2 or O3 on warm days. On cool days, the effects associated with CO remained positive, but
January 2010
5-32
-------
were no longer statistically significant after the inclusion of PM10, SO2, or NO2, but became
statistically significant and negative after the inclusion of O3 in the model (Figure 5-6).
Figure 5-4 shows the effect estimates for associations between CO and daily admissions for
CHF from selected studies; Table 5-9 summarizes the CHF hospital admission studies that examined
CO exposures.
In summary, many of the studies that examined associations between ambient CO
concentrations and daily hospital admissions for CHF reported positive associations at lags of
0-3 days.
Table 5-9. Summary of CHF hospital admission studies.
Study
Location
Endpoints
Examined
Copollutants ,- La9s .
r examined
Upper CO
Concentrations
CO Concentrations
Reported by Study
Authors in ppm
STUDIES THAT FOCUSED SOLELY ON HF
Wellenius et al. (2005,
087483)
Symons etal. (2006,
091258)
Lee etal. (2007,
090707)
Yang (2008, 157160)
Pittsburgh, PA
(1987-1999)
Baltimore, MD
(2002)
Kaohsiung,
Taiwan
(1996-2004)
Taipei, Taiwan
(1996-2004)
CHF
CHF
CHF
CHF
PM10, N02, S02, nl 0,
03 °'1'2'3
PM2.5,N02,03 0,1,2,3
PM10,N02,S02, n,
03 °~*
PM10,N02,S02, Q_2
03
98th%:1.4-3.4
99th%:1.6-3.9(24h)
98th%: 1.9-2.1
99th%:2.3(8h)
NA
NA
Mean : 1.03 (24 h)
Mean: 0.4 (8 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
PM10,N02,03 0
PM10, N02, S02, nl ,,
03 °'V"i
PM10,N02,S02, 0.2mg
O3
PM10,N02,S02, 0.2mg
Cardiac = AMI, angina, dysrhythmia, or HF; CA = Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhythm
98th%:1.0-7.8
99th%: 1.1-8.3 (24 h)
98th%: 1.2-2.0
99th% : 1.3-2.0 (24 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)
Mean: (24 h)
Winter 1.7; Spring 1.0
Summer 1.2; Fall 2.1
Mean: 0.9 (24 h)
Mean 1 .8 (1 h)
Mean 1 .8 (1 h)
ias; CHF = Congestive heart failure; PHD = Pulmonary heart disease;
OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average.
NA: Not Available
3 Includes range across individual monitors in study site; AQS data available for U.S. studies only.
Cardiovascular Diseases
The following section reviews studies that have investigated the effect of CO on ED visits and
hospital admissions for all CVD outcomes (e.g., nonspecific). Several of these studies also examined
specific CVDs and were briefly discussed in previous sections.
A multicity time-series study was conducted to estimate the risk of CVD hospitalization
associated with short-term CO exposure in 126 U.S. urban counties from 1999-2005 for over
9 million Medicare enrollees 65 yr old and older (Bell et al., 2009, 193780). The analyses yielded
positive associations between same-day CO concentration and increased risk of hospitalization for
total CVD outcomes, which remained positive and statistically significant but were attenuated with
copollutant adjustment, especially with NO2 (Figure 5-6). Overall, a 1 ppm increase in same-day 1-h
max CO was associated with a 1.010 (95% PI: 1.008-1.011) increase in risk of CVD admissions.
January 2010
5-33
-------
After adjustment for NO2, the estimate was attenuated to 1.005 (95% PI: 1.004-1.007). For most
cause-specific CVD hospitalizations, associations were positive and statistically significant for same
day CO concentration adjusted for same-day NO2(IHD 1.004 [95% PI: 1.001-1.007], heart rhythm
1.006 [95% PI: 1.001-1.011], HF 1.009 [95% PI: 1.005-1.012], and cerebrovascular 1.005 [95% PI:
1.002-1.009]). Cause-specific effect estimates were not presented for CO alone (without adjustment
for NO2).
As discussed earlier, a study was conducted in Atlanta, GA where over 4 million ED visits
from 31 hospitals for the period 1993-2000 were analyzed (SOPHIA). Several articles have been
published from this research, with three examining cardiovascular admissions in relation to CO
exposures. The first of these used a time-series design and analyzed a 3-day ma over single-day lags
0-2 as the a priori lag structure (Metzger et al, 2004, 044222). Results showed that a 1 ppm increase
in 1-h max CO concentration was associated with an increase in daily ED visits for all CVDs (RR:
1.017 [95% CI: 1.008-1.027]). This persisted in two-pollutant models that included NO2 and PM2.5.
The second of these publications examined the association of ambient air pollution levels and
cardiovascular morbidity in visits with and without specific secondary conditions (Peel et al., 2007,
090442). Within a time-stratified case-crossover design, a 3-day ma over single-day lags 0-2 was
used as the a priori lag structure. Results from the case-crossover analyses on all cardiovascular and
peripheral vascular and cerebrovascular disease were similar to the time-series results presented
earlier. Results from the various comorbidity analyses are presented in Table 5-10. Similar to the
results from the earlier publication, CO was mostly associated with peripheral vascular and
cerebrovascular disease (PVCD) among those with and without comorbidities, except among those
with CHF. Overall, there is limited, if any, evidence of susceptibility to the effects of CO
concentration for those with comorbid conditions.
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)
-
-
1.072(1.011-1.138)
1.026(1.004-1.048)
1.004(0.960-1.051)
1.023(0.998-1.049)
PVCD - peripheral vascular and cerebrovascular disease, IHD = ischemic heart disease, CHF = congestive heart failure.
Source: Reprinted with Permission of Oxford Journals from Peel et al. (2007, 090442)
The third study utilizing the SOPHIA data extended the time period to include 1993 through
2004 (Tolbert et al., 2007, 090316) and focused on two large outcome groups: a respiratory diseases
group and a cardiovascular diseases group. The combined cardiovascular case group included the
January 2010
5-34
-------
following groups of primary ICD-9 diagnostic codes: IHD (410-414), cardiac dysrhythmias (427),
CHF (428), and peripheral vascular and cerebrovascular disease (433-437, 440, 443-445, 451-453).
Results showed that a 1 ppm increase in 1-h max CO concentration was associated with an increase
in daily ED visits for all CVDs (RR: 1.016 [95% CI: 1.008-1.024]). CO was the strongest predictor
of CVD effects in models with two-pollutant combinations of NO2, CO and total carbon, as well as
in a model including all three pollutants.
Earlier research conducted in Los Angeles, CA, showed that a 0.5 ppm increase in same-day
24-h avg CO concentration was associated with a 1.6% increase in CVD hospital admissions among
people >30 yr (Linn et al., 2000, 002839). When the analyses were stratified by season, the strongest
CO effect occurred during the winter (1.9% increase) followed by the summer (1.8%) and fall
(1.4%) with no effect in spring.
In contrast to other North American studies, a study in Spokane, WA, did not find an
association between CO (lags of 1-3 days) and an increase in the number of daily cardiac hospital
admissions (quantitative results not reported) (Slaughter et al., 2005, 073854). Similarly, a time-
series study in Windsor, Ontario, did not find an association between ambient CO and daily hospital
admissions for CVDs (defined as HF, IHD, or dysrhythmias) (Fung et al., 2005, 074322). A total of
11,632 cardiac admissions were analyzed for the period 1995-2000. The lag periods analyzed in 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 increase
in 1-h max CO concentration the mean percent change in daily admissions for the <65-yr age group
(lag 0) was -2.6 (95% CI: -6.2 to 3.3); and for the 65+ yr age group, 0.4 (95% CI: -1.9 to 2.7). The
authors reported moderate to low correlations with NO2 (r = 0.38), PMi0 (r = 0.21) and SO2
(r = 0.16).
Two case-crossover studies in Taiwan reported an association between ambient CO and
hospital admissions for CVDs. In Taipei, a total of 74,509 CVD admissions from 47 hospitals for the
period 1997-2001 were analyzed (Chang et al., 2005, 080086). An increase of 0.5 ppm in 24-h avg
CO concentration (average over lags 0-2) during warmer periods (> 20°C) was associated with an
increase in daily hospital admissions (OR: 1.09 [95% CI: 1.065-1.121]) but not cooler periods
(<20°C) (OR: 0.98 [95% CI: 0.93-1.004]). These results persisted after controlling for PMi?, SO2, or
O3 in two-pollutant models. An identical study in Kaohsiung analyzed 29,661 CVD admissions for
the period 1997-2000 (Yang et al., 2004, 094376). Results showed that a 0.5 ppm increase in 24-h
avg CO concentration was associated with an increase in CVD hospital admissions during both the
warmer periods (OR: 1.50 [95% CI: 1.38-1.63]) and cooler periods (OR: 1.89 [95% CI: 1.69-2.12]).
Similarly, two Australian studies also reported associations between ambient CO
concentrations and increased CVD hospital admissions among older adults. The first of these studies
analyzed data from five of the largest cities in Australia (Brisbane, Canberra, Melbourne, Perth,
Sydney) and two New Zealand cities (Auckland, Christchurch) for the period 1998-2001 (Barnett et
al., 2006, 089770). The combined estimates showed that an increase of 0.75 ppm in the average 8-h
max CO concentration over the current and previous day (lag 0-1) was associated with a 1.8%
(95% CI: 0.7-2.8) increase in all CVD admissions among those aged 65+ yr. Among those aged
15-64 yr there was a smaller increase in CVD admissions (1.0% [95% CI: 0.2-1.7]). The second of
the Australian studies examined ED visits for CVDs in older adults (65+ yr) in Sydney for the period
1997-2001 (Jalaludin et al., 2006, 189416). A 0.75 ppm increase in 8-h max CO concentration for
single-day lags 0 and 1 was associated with increases in admissions of 2.5% (95% CI: 1.6-3.5) and
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
an increase of 2.6% (95% CI: 1.5-3.6). There were positive increases of approximately 3% in CVD
ED visits during the cool (May-October) period but not the warm period (November-April).
Very few studies investigating the association between CO and cardiovascular hospital
admissions have been conducted in European cities. Ballester et al. (2001, 013257) analyzed
emergency hospital admissions in Valencia, Spain, for the period 1994-1996. The mean daily number
of CVD admissions was 7, and there was no association between CO and admissions for all CVDs
(RR: 1.009 [95% CI: 0.99-1.016] per 1 ppm increase in 1-h max CO concentration), heart diseases
(RR: 1.010 [95% CI: 0.993-1.028] per 1 ppm increase), and cerebrovascular diseases (RR: 0.985
[95% CI: 0.959-1.012] per 1 ppm increase). When the analyses were stratified by hot and cold
seasons, only CO concentrations during the hot season were associated with an increase in all
cardiovascular admissions (RR: 1.033 [95% CI: 1.006-1.064] per 1 ppm increase), heart disease
admissions (RR: 1.033 [95% CI: 1.000-1.067] per 1 ppm increase), and cerebrovascular admissions
(RR: 1.074 [95% CI: 1.007-1.113] per 1 ppm increase).
January 2010 5-35
-------
Ballester et al. (2006, 088746) extended this research to include data from 14 Spanish cities
for the period 1995-1999. An average exposure period over lags 0-1 was analyzed and for the
combined estimates a 0.75 ppm increase in 8-h max CO concentration was associated with a 1.77%
(95% CI: 0.56-2.99) increase in all cardiovascular hospital admissions and a larger increase of 3.57%
(95% CI: 1.12-6.08) for heart disease admissions. These results persisted in two-pollutant models
that included NO2, O3 and SO2.
A study was carried out to evaluate the association between air pollution cardiovascular ED
visits in subjects with and without diabetes in Sao Paulo, Brazil (Pereira Filho et al., 2008, 190260).
From January 2001 to July 2003, 45,000 ED visits were registered due to CVDs, of which 700 were
registered due to CVDs in diabetic patients. SO2 and NO2 were positively and statistically
significantly associated with CVD ED visits among diabetics and nondiabetics, while CO was only
positive and statistically significant among non-diabetic patients. PMi0 and O3 were not positively
associated with ED admissions among either group.
Table 5-11 summarizes the non-specific CVD hospital admission studies that examined CO
exposures. Due to the heterogeneity of endpoints, these studies do not lend themselves to a
quantitative meta-analysis, and a forest plot was used to summarize the results of the studies on all
CVD outcomes. Figure 5-5 shows the effect estimates associated with daily admissions for
nonspecific CVD hospital admissions from selected studies.
In summary, many of the studies that examined associations between ambient CO
concentrations and ED visits and daily hospital admissions for CVD reported small yet precise
positive associations at short (0-1 day) lags. Among studies that conducted stratified analyses, there
were slightly stronger effects among older adults and possibly during warmer periods.
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)
Funa etal. (2005, 074322)
Funa etal. (2005,074322)
Barnett etal. (2006, 089770)
Barnett etal. (2006,089770)
Jalaludinetal. (2006, 189416)
Ballester et al. (2006, 088746)
Pereira Filho et al. (2008, 190260)
Pereira Filho et al. (2008, 190260)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Yana etal. (2004, 094376)
Yanq etal. (2004, 094376)
Location
126 US counties
Atlanta, GA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Windsor, Can
Windsor, Can
Australia, New Zealand
Australia, New Zealand
Sydney, Australia
Multicity, Spain
Sao Paulo, Brazil
Sao Paulo, Brazil
Taipei, Taiwan
Taipei, Taiwan
Kaohsiung, Taiwan
Kaohsiung, Taiwan
Lag
0
0-2
0
0
0
0
0
0-2
0-2
0-1
0-1
0-1
0-1
0
0
0-2
0-2
0-2
0-2
Group
65+ yr
All Year
Spring
Summer
Fall
Winter
<65yr
65+ yr
15-64yr
65+ yr
65+ yr
Diabetic
Nondiabetic
>20°C Temp
<20°CTemp
<20°CTemp
>20°C Temp
Effect Estimate (95% CI)
•
5
i»
*
»
5
•
-f-
*-
j»
;•
,•
*
-L»-
t
, *•
-»|
1 •
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Relative Risk
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.
January 2010
5-36
-------
Table 5-11. Summary of nonspecific 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, Q13257)a
Ballester et al. (2006, 088746)'
Pereira Filho et al. (2008,
190260)
Location
126 urban US
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, Can
(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
PM2.5, N02, EC
PM10,N02,
S02, 03
PM,o, 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
AQSb 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%: 4.7-4.9
99th%: 5.3-5.4(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: 1.8(1 h)
Mean 1 .8 (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(1 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)
3These 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.
b Includes range across individual monitors in study site; AQS data available for U.S. studies only.
January 2010
5-37
-------
Figure 5-6 and Figure 5-7 summarize the effects of CO concentration on ED visits and
hospital admissions for all CVD outcomes other than stroke from studies that presented the results
from two-pollutant models. Generally, the CO effect estimates from these studies are robust to the
inclusion of copollutants, including PMi0, PM2 5, NO2, SO2, and O3. In all but two instances - Lee et
al. (2007, 090707) (<25°C adjusted for NO2) and Yang (2008, 157160) (<20°C adjusted for O3) -
when the single pollutant effect estimate was positive for CO, it remained positive after the addition
of any of the copollutants investigated.
Study
Wellenius et al. (2005, 087483)
Wellenius etal. (2005, 087483)
Chan etal. (2006,090193)
Chan etal. (2006, 0901 93)
Lee etal. (2007, 090707)
Lee etal. (2007,090707)
D'lppoliti etal. (2003,074311)
D'lppoliti etal. (2003,074311)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana 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 etal. (2007, 090707)
Lee etal. (2007,090707)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Bell etal. (2009, 193780)
Bell etal. (2009, 193780)
Tolbert etal. (2007,090316)
Tolbert etal. (2007,090316)
Chan etal. (2006,090193)
Chan etal. (2006,090193)
Outcome
CHF
CHF
CD
CD
CHF
CHF
IHD
IHD
CVD
CVD
CVD
CVD
IHD
IHD
CVD
CVD
CVD
CVD
CHF
CHF
CHF
CHF
CHF
CHF
CVD
CVD
CVD
CVD
CD
CD
Lag
0
0
2
2
0
0
0-2
0-2
0
0
0
1
0
0
0-1
0-1
0
0
0
0
0-2
0-2
0-2
0-2
0
0
0-2
0-2
2
2
Pollutant
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + TSP
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM10
CO Alone
CO + PM25
CO Alone
CO + PM25
CO Alone
CO + PM25
Group
>25°C
>20°C
<20°C
64+ yr
<25°C
>20°C
<20°C
Effect Estimate (95% Cl)
;• PR/I™
,A
I*
lA
1 •
1 A
!«
L
, «.
1 -A-
«
-A-l
!*-
V
!^_
-A-
n-
4A-
1 •
1 t
it
, j«.
1 • PM2.5
1 A
1,
A
A
1 1 1 1 1 1 1 1
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Relative Risk
Figure 5-6. Effect estimates from studies of ED visits and hospital admissions for CVD
outcomes other than stroke from single pollutant (CO only: black circles) and
particulate copollutant (CO + PM2.s: red triangles; CO + PM™ or TSP: purple
triangles) 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.
January 2010
5-38
-------
Study
Bell etal. (2009, 193780)
Bell etal. (2009, 193780)
Wellenius etal. (2005, 087483)
Wellenius etal. (2005, 087483)
Tolbert etal. (2007,090316)
Tolbert etal. (2007,090316)
Ballester et al. (2006, 088746)
Ballester et al. (2006, 088746)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Lee etal. (2007,090707)
Lee etal. (2007, 090707)
Lee etal. (2007,090707)
Lee etal. (2007,090707)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Wellenius etal. (2005, 087483)
Wellenius et al. (2005, 087483)
Ballester et al. (2006, 088746)
Ballester et al. (2006, 088746)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Lee etal. (2007,090707)
Lee etal. (2007,090707)
Lee etal. (2007, 090707)
Lee etal. (2007,090707)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Wellenius et al. (2005, 087483)
Wellenius etal. (2005, 087483)
Chan etal. (2006,090193)
Chan etal. (2006,090193)
Ballester et al. (2006, 088746)
Ballester etal. (2006, 088746)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
Ghana etal. (2005, 080086)
von Klot etal. (2005,088070)
von Klot etal. (2005,088070)
Lee etal. (2007, 090707)
Lee etal. (2007,090707)
Lee etal. (2007, 090707)
Lee etal. (2007, 090707)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Yana (2008, 157160)
Outcome
CVD
CVD
CHF
CHF
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CVD
CVD
CVD
CVD
CVD
CVD
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CD
CD
CVD
CVD
CVD
CVD
CVD
CVD
Cardiac
Cardiac
CHF
CHF
CHF
CHF
CHF
CHF
CHF
CHF
Lag
0
0
0
0
0-2
0-2
0-1
0-1
0
0
0
3
0
0
0
0
0-2
0-2
0-2
0-2
0
0
0-1
0-1
0
0
0
2
0
0
0
0
0-2
0-2
0-2
0-2
0
0
2
2
0-1
0-1
0
0
0
4
0
0
0
0
0
0
0-2
0-2
0-2
0-2
Pollutant
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + N02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
CO + S02
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
CO Alone
C0 + 03
Label
>20°C
<20°C
>25°C
<25°C
>20°C
<20°C
>20°C
<20°C
>25°C
<25°C
>20°C
<20°C
>20°C
<20°C
>25°C
<25°C
>20°C
<20°C
Effect Estimate (95% Cl)
1 • NO2
1 A
•
A
!«
A
!*•
[A-
•-
hA-
-•-
—£i —
A
| — •—
— £ —
,»
— T-A
i* SO2
lA
•0-
IA-
«-
1 -A-
-•L-
A
, A
A
•
— A —
-t-t —
— IA
• 03
'A
'+
'A
v
A-
I •-
-A-
_«,-
l^.
JA-
g
A
1 •
1 A
' _»_
^
m
f^
\ i i i r
0.8 1.0 1.2 1.4 1.6 1.8 2.0
Relative Risk
Figure 5-7.
Effect estimates from studies of ED visits and HAs for CVD outcomes other than
stroke from single pollutant (CO only: black circles) and gaseous copollutant
models (CO + N02, S02 and 03= green, blue, and orange triangles, respectively).
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.
January 2010
5-39
-------
5.2.2. Epidemiologic Studies with Long-Term Exposure
Two studies examined CVD outcomes in association with long-term exposure to CO.
Rosenlund et al. (2006, 089796) investigated long-term exposure (30 yr) to urban air pollution and
the risk of MI in Sweden. The study included 2,246 cases and 3,206 controls aged 45-70 yr and
residing in Stockholm County during 1992-1993. A detailed postal questionnaire was completed by
4,067 subjects, and all addresses inhabited during more than 2 yr since 1960 were geocoded. The
exposures were then derived from dispersion calculations based on emissions data for each decade
since 1960. These calculations were estimates of annual mean levels of traffic-generated NOX, NO2,
CO, PMio, and PM2.5, with the addition of SO2 from heating sources. The analyses were stratified by
all cases, nonfatal cases, fatal cases, in-hospital death, and out-of-hospital death. Based on a 30-yr
avg exposure all pollutants were not associated with overall MI incidence. However, increased CO
was associated with out-of-hospital death from MI (OR: 1.81 [95% CI: 1.02-3.23] per 0.5 ppm
increase in 30-yr avg CO concentration). Similar results were reported for NO2. The correlation
between the 30-yr NO2 and CO exposures was reasonably strong (r = 0.74) and multipollutant
models with both these pollutants included (NO2, CO) were not examined. No other pollutants were
significantly associated with all other MI outcomes. The study period was extended to include
43,275 cases of MI during 1985-1996 and 507,000 controls (Rosenlund et al., 2009, 190309). Five-
year average exposures to NO2, PM10 and CO were associated with incidence of MI, especially with
fatal disease; when examining only nonfatal disease, no association was observed. The effect
estimate for CO (OR: 1.03 [95%C CI: 1.02-1.04] per 0.5 ppm increase in 5-yr avg) was similar in
magnitude to those for NO2 and PMi0. When the analysis was restricted to the group that did not
move between population censuses (the least expected misclassification of true individual exposure),
the effect estimate for CO increased to 1.17 (95% CI: 1.11-1.24) per 0.5 ppm increase in 5-yr avg,
and although the effect estimates for NO2 and PM10 remained similar to the estimate for CO, in this
analysis the effect estimate for CO was slightly greater in magnitude than the effect estimate for
A small-area ecologic study analyzed mortality and hospital admissions for stroke across 1,030
census districts in Sheffield, U.K. (Maheswaran et al., 2005, 088683). Stroke counts within each
census district were linked to modeled air pollution data which was then grouped into quintiles of
exposure. For stroke hospital admissions, when the analyses were adjusted for only sex and age
demographics, there was an exposure-response pattern exhibited across the quintiles of CO exposure
with all levels reaching significance (RR: 1.37 [95% CI: 1.24-1.52] for the highest exposure group
compared to the lowest group). However, this result did not persist when also adjusting for a
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
A substantial number of epidemiologic studies have examined the potential association
between exposure to CO and various relevant cardiac endpoints or biomarkers. Overall, despite some
mixed results reported among panel and retrospective cohort studies, there was evidence that
exposure to CO has an effect on HR, various HRV parameters, and blood markers of coagulation and
inflammation. Conversely, based on results from panel studies, there was little evidence of a link
between CO and cardiac arrhythmia, cardiac arrest, the occurrence of MI, and increased BP
Studies of ED visits and hospital admissions provide evidence that CO is associated with
various forms of CVD, with lag periods ranging from 0 to 3 days. Nearly all of the studies include
same day (lag 0) or next day (lagl) lag periods, which are consistent with the proposed mechanism
and biological plausibility of these CVD outcomes. When categorized by specific cardiovascular
outcome, the evidence is consistent. Studies of hospital admissions and ED visits for IHD provide
the strongest evidence of ambient CO being associated with adverse CVD outcomes. The effect
estimates for this outcome are nearly all positive, many are statistically significant, and the
magnitude of effect is similar among the studies. Though not as consistent as the IHD effects, the
effects for all CVD hospital admissions (which include IHD admissions) and CHF hospital
admissions also provide evidence for an association with ambient CO concentrations. There is very
limited evidence that ambient CO is associated with ischemic stroke. It is difficult to determine from
this group of studies the extent to which CO is independently associated with CVD outcomes or if
January 2010 5-40
-------
CO is a marker for the effects of another traffic-related pollutant or mix of pollutants. On-road
vehicle exhaust emissions are a nearly ubiquitous source of combustion pollutant mixtures that
include CO and can be an important contributor to CO in near-road locations. Although this
complicates the efforts to disentangle specific CO-related health effects, the evidence indicates that
CO associations generally remain robust in copollutant models and supports a direct effect of short-
term ambient CO exposure on CVD morbidity.
5.2.4. Controlled Human Exposure Studies
Controlled human exposure studies provide valuable information related to the health effects
of short-term exposure to air pollutants. Results of controlled human exposure studies can be used to
provide coherence with the evidence from epidemiologic studies by expanding the understanding of
potential mechanisms for the observed health outcomes. However, they may also provide
information that can be used directly in quantitatively characterizing the exposure concentration-
health response relationships at ambient or near-ambient concentrations.
Several human clinical studies cited in the 2000 CO AQCD (U.S. EPA, 2000, 000907)
observed changes in measures of cardiovascular function among individuals with CAD, following
short-term exposures to CO. Principal among these is a large multilaboratory study of men with
stable angina (n = 63), designed to evaluate the effect of CO exposure resulting in COHb
concentrations of 2% and 4% on exercise-induced angina and ST-segment changes indicative of
myocardial ischemia (Allred et al, 1989, 013018; Allred et al, 1989, 012697; Allred et al, 1991,
011871). The majority of subjects were following an anti-ischemic medication regimen (e.g., beta
blockers, nitrates, or calcium channel antagonists) which was maintained throughout the study. On
three separate occasions, subjects underwent an initial graded exercise treadmill test, followed by 50-
to 70-min exposures under resting conditions to average CO concentrations of 0.7 ppm (room air
concentration range 0-2 ppm), 117 ppm (range 42-202 ppm) and 253 ppm (range 143-357 ppm).
After the 50- to 70-min exposures, subjects underwent a second graded exercise treadmill test, and
the percent change in time to onset of angina and time to ST endpoint between the first and second
exercise tests was determined. The investigators conducted two exercise tests on exposure days (pre-
versus postexposure) to control for day-to-day variability in the endpoints of interest. The effect of
CO was evaluated by comparing the percent change in time to onset of angina or ST-segment change
between the CO and clean air exposure days. The order of the three exposures was randomly
determined and counterbalanced across subjects. For the CO exposure sessions, postexposure target
COHb concentrations were set at values 10% greater than the post-exercise targets (i.e., 2.2% and
4.4%) to compensate for the elimination of CO during exercise testing in clean air following
exposure. CO uptake constants were determined for each subject individually during a qualifying
visit and were used to compute the inhaled concentration required to attain the target COHb
concentrations. Although CO-oximetry was used at each center to rapidly provide approximate
concentrations of COHb during the actual exposure, COHb concentrations determined by a gas
chromatographic technique were used in the statistical analyses as this method is known to be more
accurate than CO-oximetry and other spectrophotometric methods, particularly for samples
containing COHb concentrations <5%. For the two CO exposures, the average postexposure COHb
concentrations were reported as 2.4% and 4.7% (3.2% and 5.6% using CO-oximetry), and the
average post-exercise COHb concentrations were reported as 2.0% and 3.9% (2.7% and 4.7% using
CO-oximetry). While the average COHb concentrations during the exercise tests were clearly
between the concentrations measured in postexposure and post-exercise blood samples, the study
authors noted that the samples at the end of the exercise test represented the COHb concentrations at
the approximate time of onset of myocardial ischemia as indicated by angina and ST segment
changes. Relative to clean air exposure (COHb ~ 0.6-0.7%), exposures to CO resulting in post-
exercise COHb concentrations of 2.0% and 3.9% were shown to decrease the time required to induce
ST-segment changes by 5.1% (p = 0.01) and 12.1% (p < 0.001), respectively. These changes were
well correlated with the onset of exercise-induced angina. The observed dose-response relationship
was further evaluated by regressing the percent change in time to ST-segment change or time to
angina on actual post-exercise COHb concentration (0.2-5.1%) using the three exposures (air control
and two CO exposures) for each subject. Regression analyses were conducted separately for each
individual and the averages of the intercepts and slopes across subjects were reported. This analysis
demonstrated significant decreases in time to angina and ST-segment change of approximately 1.9%
and 3.9%, respectively, per 1% increase in COHb concentration, with no evidence of a measurable
January 2010 5-41
-------
threshold. The relationship between percent change in time to ST-segment endpoint and post-
exercise COHb concentration is illustrated in Figure 5-8.
70-
35
a
1
-35
-70-
. t
I*
• I
••. •.
234
Postexercise %COHb
Source: Reprinted with Permission of HEI from Allred et al. (1 989, 0126971
Figure 5-8.
Regression of the percent change in time to ST endpoint between the pre- and
postexposure exercise tests ([postexposure-pre-exposure]/pre-exposure) and
the measured blood COHb levels at the end of exercise for the 63 subjects
combined. The line represents the average of individual regressions.
In addition to the work of Allred et al. (1989, 013018: 1989, 012697: 1991, 011871) a number
of other studies involving individuals with stable angina have also demonstrated a CO-induced
decrease in time to onset of angina, as well as reduction in duration of exercise at COHb
concentrations between 3 and 6%, measured using spectrophotometric methods (Adams et al., 1988,
012692: Anderson et al., 1973, 023134: Kleinman et al., 1989, 012696: Kleinman et al., 1998,
047186). However, Sheps et al. (1987, 012212) observed no change in time to onset of angina or
maximal exercise time following a 1-h exposure to 100 ppm CO (targeted COHb of 4%) among a
group of 30 patients with CAD. In a subsequent study conducted by the same laboratory, a
significant increase in number of ventricular arrhythmias during exercise was observed relative to
room air among individuals with CAD following a 1-h exposure to 200 ppm CO (targeted COHb of
6%) but not following a 1-h exposure to 100 ppm CO (targeted COHb of 4%) (Sheps et al., 1990,
013286).
While cardiovascular effects of CO have consistently been observed in studies of controlled
human exposure among individuals with CAD at COHb concentrations between 2 and 6%, a
quantitative meta-analysis of these studies is of limited value considering differences in the methods
used. For example, variation in exercise protocols resulted in substantial differences between studies
in total exercise time. More importantly, only Allred et al. (1989, 013018: 1989, 012697: 1991,
January 2010
5-42
-------
011871) analyzed COHb concentration using gas chromatography. Although all studies measured
COHb using spectrophotometric methods, these methods are only accurate within approximately 1%
COHb of the true value at COHb concentrations < 5% (U.S. EPA, 1991, 017643). Therefore, a
quantitative evaluation of changes in cardiovascular response with small increases in COHb
concentration (< 1%) as measured using CO-oximetry is neither appropriate nor informative,
particularly at low COHb concentrations.
It should be noted that although the subjects evaluated in the studies described above are not
necessarily representative of the most sensitive population, the level of disease in these individuals
was moderate to severe, with the majority either having a history of MI or having > 70% occlusion
of one or more of the coronary arteries. The 2000 CO AQCD (U.S. EPA, 2000, 000907) presented
very little evidence of CO-induced changes in cardiovascular function in healthy adults. Davies and
Smith (1980, 011288) exposed healthy young adults continuously for 7 days to CO concentrations of
0, 15, or 50 ppm. In this study, a marked ST-segment depression was demonstrated in only 1 out of
16 subjects following exposure to 15 ppm CO (2.4% COHb) or 50 ppm CO (7.2% COHb).
Since the publication of the 2000 CO AQCD (U.S. EPA, 2000, 000907). no new human
clinical studies have been published involving controlled CO exposures among subjects with CAD.
However, a number of new studies have evaluated changes in various measures of cardiovascular
and systemic responses following controlled exposures to CO in healthy adults. Adir et al. (1999,
001026) exposed 15 young healthy adult males to room air or CO for approximately 4 min, using a
CO exposure concentration which had been shown to produce the targeted COHb level of 4-6%.
Following each exposure, subjects performed an exercise treadmill test at their maximal capacity.
Exposure to CO was not observed to cause arrhythmias, ST-segment changes, or changes in
myocardial perfusion (thallium scintigraphy) during postexposure exercise. However, CO was
demonstrated to decrease the postexposure duration of exercise by approximately 10% (p = 0.0012).
In addition, the authors reported significant CO-induced decreases in metabolic equivalent units (p <
0.001), which is a relative measure of O2 consumption. These results support the findings of several
studies cited in the 2000 CO AQCD (U.S. EPA, 2000, 000907) which observed decreases in exercise
duration and maximal aerobic capacity among healthy adults at COHb levels > 3% (Drinkwater et
al., 1974, 041332: Ekblom and Huot, 1972, 010886: Horvath et al., 1975, 010887: Raven et al.,
1974, 041340). While these decreases in exercise duration were relatively small and only likely to be
noticed by competing athletes, the findings are nonetheless important in providing coherence with
the observed effects of CO on exercise-induced myocardial ischemia among patients with CAD.
Kizakevich et al.(2000, 052691) evaluated the cardiovascular effects of increasing CO
concentration in healthy adults engaged in upper and lower body exercise. Subjects were initially
exposed for 4-6 min to CO concentrations between 1,000 and 3,000 ppm, followed by continued
exposure to 27, 55, 83, and 100 ppm to maintain COHb levels of 5, 10, 15, and 20%, respectively.
Relative to room air control, CO exposure was not observed to cause ST-segment changes or affect
cardiac rhythm at any concentration during either upper or lower body exercise. Compensation
mechanisms for reduced O2 carrying capacity during CO exposure were demonstrated, with
statistically significant increases in heart rate occurring at COHb levels > 5%, and statistically
significant increases in cardiac output and cardiac contractility observed at COHb levels > 10%. In a
human clinical study designed to evaluate the contribution of CO to cardiovascular morbidity
associated with cigarette smoking, Zevin et al. (2001, 021120) exposed 12 healthy male smokers for
7 consecutive days to clean air, CO, or cigarette smoke, with each subject serving as his own control.
The COHb levels were similar between the exposures to cigarette smoke and CO, with average
concentrations of 6% and 5%, respectively. Cigarette smoke, but not CO, was observed to
significantly increase plasma levels of CRP and plasma platelet factor 4 relative to the air control
arm of the study. Neither cigarette smoke nor CO was shown to affect BP. Hanada et al. (2003,
193915) observed an increase in leg muscle sympathetic nerve activity (MSNA) following
controlled exposures to CO (COHb ~ 20%) under normoxic or hyperoxic conditions. Although an
increase in the magnitude of sympathetic activation is typically associated with regional
vasoconstriction, no CO-induced changes in femoral venous blood flow were observed in this study.
These findings are in agreement with those of Hausberg and Somers (1997, 083450) who observed
no change in forearm blood flow or BP in a study of 10 healthy men and women following a
controlled exposure to CO (COHb ~ 8%). Interestingly, one recent study did observe an increase in
retinal blood flow, retinal vessel diameter, and choroidal blood flow following controlled exposures
to CO at a concentration of 500 ppm (Resch et al., 2005, 193853). This protocol resulted in COHb
concentrations of 5.6% and 9.4% following exposures of 30 and 60 min, respectively, with
January 2010 5-43
-------
statistically significant increases in retinal and choroidal blood flow observed at both time points
relative to synthetic air control. This CO-induced change in ocular hemodynamics may have been
due to local tissue hypoxia; however, the clinical significance of this finding is unclear. Exposures to
CO have also been shown to affect skeletal muscle function, with one recent human clinical study
reporting a decrease in muscle fatigue resistance in healthy adult males, using both voluntary and
electrically-induced contraction protocols following controlled exposures to CO resulting in an
average COHb level of 6% (Morse et al., 2008, 097980).
In summary, controlled human exposures to CO among individuals with CAD have been
shown to consistently increase markers of myocardial ischemia at COHb concentrations between 2
and 6%. No such effects have been observed in healthy adults following controlled exposures to CO.
Although some studies have reported CO-induced hemodynamic changes among healthy adults at
COHb concentrations as low as 5%, this effect has not been observed consistently across studies.
5.2.5. lexicological Studies
While there was no toxicological research reported in the 2000 CO AQCD (U.S. EPA, 2000,
000907) that involved CO exposures at or below the NAAQS levels, adverse cardiovascular effects
were reported for higher CO concentrations. The lowest observed effect levels for cardiovascular
effects in experimental animals included 50 ppm (6-wk exposure, 2.6% COHb) for cardiac rhythm
effects, 100 ppm (46 days, 9.3% COHb) for hematology effects, 150 ppm (30 min, 7.5% COHb) for
hemodynamic effects, 200 ppm (30 days, 15.8% COHb) for cardiomegaly, and 250 ppm (10 wk,
20% COHb) for atherosclerosis and thrombosis (Table 6-11 in the 2000 CO AQCD) (U.S. EPA,
2000, 000907). Conflicting experimental data relating to the role of CO in promoting atherosclerotic
vessel disease were discussed. While some animal studies have linked chronic CO exposure with
atherosclerosis development resulting from increased fatty streaking and cellular lipid loading
(Davies et al., 1976, 010660: Thomsen, 1974, 010704; Turner et al., 1979, 012328). other studies
have failed to see this association (Penn et al., 1992, 013728; Stupfel and Bouley, 1970, 010557).
Vascular insults due to acute exposure to CO concentrations of 50 ppm and higher were also reported
(Ischiropoulos et al., 1996, 079491; Thorn, 1993, 013895; Thorn et al., 1998, 016750; Thorn et al.,
1999, 016757; Thorn et al., 1999, 016753). In addition, chronic CO exposure has been shown to
result in ventricular hypertrophy (Penney et al., 1984, 011567; Penney et al., 1988, 012521).
The following sections describe recent studies dealing with toxicity of low to moderate
concentrations (35-250 ppm) of CO. There has been little new research with the overt purpose of
examining environmentally-relevant levels of CO. For the most part, studies were designed to mimic
exposures related to cigarette smoke, either side-stream or mainstream, accidental CO poisoning, or
for the purposes of therapeutic application. Thus, few studies examined levels of CO within the
current 1-h (35 ppm) or 8-h (9 ppm) NAAQS levels, and fewer still examined concentration
response curves to delineate no-effects levels. However, it is apparent that CO, at low to moderate
concentrations, has pathophysiologic effects on the cardiovascular system and on relatively
ubiquitous cellular pathways. In evaluating these studies, it should be kept in mind that the
traditional concept of CO pathophysiology resulting from reduced O2 delivery is likely to be more
relevant for higher concentrations of CO than are currently found in the ambient environment.
CO exposure at environmentally-relevant levels is unlikely to cause overt toxicity in a healthy
cell; however, susceptibility may be rendered by disease or developmental stage. A common theme
appears to be the vulnerability of vascular cells, especially the endothelium, which could be
considered the first organ of contact once CO is taken up into the circulation. While relatively little
research has been conducted since the 2000 CO AQCD (U.S. EPA, 2000, 000907). several key
studies conducted at environmentally-relevant CO levels provide important clues to the potential
public health implications of ambient CO exposure.
5.2.5.1. Endothelial Dysfunction
While the preferential binding to heme and effective displacement of O2 by CO has been well
established for over a century, new information from various fields of study are beginning to
elucidate nonhypoxic mechanisms that may lead to cardiovascular abnormalities associated with CO
exposure. Research by Thorn, Ischiropoulos, and colleagues (Ischiropoulos et al., 1996, 079491;
Thorn and Ischiropoulos, 1997, 085644; Thorn et al., 1994, 076459; Thorn et al., 1997, 084337;
January 2010 5-44
-------
Thorn et al, 1999, 016753; Thorn et al, 1999, 016757). some of which was reported in the 2000 CO
AQCD (U.S. EPA, 2000, 000907). has focused on CO-mediated displacement of NO from heme-
binding sites. Some of this work demonstrates a specific pathway by which severe CO poisoning can
lead to the release of NO from platelets with subsequent neutrophil activation and vascular injury
(Ischiropoulos et al., 1996, 079491: Thorn et al., 2006, 098418). The steps include: (1) peroxynitrite
generation from the reaction of NO from platelets with neutrophil-derived superoxide; followed by
(2) stimulation of intravascular neutrophil degranulation; that can result in (3) myeloperoxidase
deposition along the vascular lining. Products from myeloperoxidase-mediated reactions can cause
endothelial cell activation (Thorn et al., 2006, 098418) and can lead to endothelial dysfunction. The
concentrations used in these studies are greatly in excess of the NAAQS levels but certainly within
the range of accidental or occupational exposures. Research by these same investigators at more
environmentally-relevant CO levels was partially reviewed in the 2000 CO AQCD (U.S. EPA, 2000,
000907). The release of free NO was noted in isolated rat platelets exposed to 10-20 ppm CO (Thorn
and Ischiropoulos, 1997, 085644). Increased nitrotyrosine content of the aorta was observed in rats
exposed to 50 ppm CO for 1 h (Thorn et al., 1999, 016757: Thorn et al., 1999, 016753). Furthermore,
in this same study, a 1-h exposure to 100 ppm CO led to albumin efflux from skeletal muscle
microvasculature at 3 h and leukocyte sequestration in the aorta at 18 h; LDL oxidation was also
reported. These effects were dependent on NOS but not on neutrophils or platelets. A second study
demonstrated NO-dependent effects of 50-100 ppm CO in lungs and is described in Section 5.5.4
(Thorn et al., 1999, 016757). Studies in cultured endothelial cells were also conducted using buffer
saturated with 10-100 ppm CO (Thorn et al., 1997, 084337). These experiments were designed to
mimic conditions where blood COHb levels were between 3.8 and 28%, resulting in exposure of
endothelial cells to 11-110 nM CO. CO stimulated the release of NO from endothelial cells along
with formation of peroxynitrite; delayed cell death was observed at CO concentrations of 22 nM and
higher (Thorn et al., 1997, 084337). A more recent study demonstrated adaptive responses in
endothelial cells exposed to this same range of CO concentrations (Thorn et al., 2000, 011574).
Specifically, 1-h exposure to 11 nM CO resulted in MnSOD and HO-1 induction and resistance to
the apoptotic effects of 110 nM CO. These protective effects of CO were mediated by NO, as
demonstrated using an inhibitor of NOS and a scavenger of peroxynitrite. Collectively, these
experiments demonstrated oxidative and nitrosative stress, the initiation of inflammation, increased
microvascular permeability and altered cell signaling in animals and isolated cells following
exposure to 10-100 ppm CO.
CO is an endogenous regulator of vasomotor tone through vasodilatory effects, mediated by
activation of soluble guanylate cyclase and activation of large conductance Ca2+-activated K+
channels. However, CO does not cause vasodilation in every vascular bed. For example, 5, 100, 500
and 2,500 ppm CO administered by inhalation to near-term fetal lambs did not induce pulmonary
vasodilation, and the HO-inhibitor zinc protoporphyrin IX failed to affect baseline vascular tone
(Grover et al., 2000, 097088). In some cases CO promotes vasoconstriction, which is thought to be
mediated by inhibition of endothelial NOS (Johnson and Johnson, 2003, 053611: Thorup et al., 1999,
193782) or decreased NO bioavailability. An interesting series of studies has also suggested that
endogenous CO derived from HO-1 which is induced in a variety of disease models (salt-sensitive
forms of hypertension, metabolic syndrome in obese rats) is responsible for skeletal muscle arterial
endothelial dysfunction (Johnson and Johnson, 2003, 053611: Johnson et al., 2006, 193874: Teran et
al., 2005, 193770). Additional studies will be useful in determining whether environmentally-
relevant concentrations of CO have detrimental effects on preexisting conditions such as
hypertension, metabolic syndrome or pregnancy.
Several recent animal studies examined the vascular effects of controlled exposures to
complex combustion mixtures containing CO. Vascular dilatation was decreased following exposure
to diesel (4 h at 4 ppm) (Knuckles et al., 2008, 191987) and gasoline engine emissions (6 h/day for 1,
3, and 7 days at 80 ppm) (Lund et al., 2009, 180257). Furthermore, evidence of vascular ROS
following gasoline emissions has been shown in certain animal models (6 h/day for 50 days at
8-80 ppm) (Lund et al., 2007, 125741). While none of these studies examined the potential
independent role of CO, it is clearly a common factor in the various combustion atmospheres, and
future work will be needed to reveal its importance on vascular health.
January 2010 5-45
-------
5.2.5.2. Cardiac Remodeling Effects
Cardiomyopathy, or abnormal growth of the cardiac muscle, can manifest in different ways,
depending on the nature of the insult. The adverse effects of cardiac hypertrophy are due to reduction
of ventricular chamber volume and a diminishing efficiency of the heart. Such concentric
hypertrophy typically occurs in response to chronic increases in load, as occurs with hypertension.
Ischemia of the cardiac tissue can also lead to cardiac remodeling and myopathy. During and after an
acute infarction or obstruction of major coronary vessels, downstream tissues can suffer severe
regional ischemia that leads to significant necrosis. Such regions will lose the ability to contract, and
surrounding tissue will show deficits in contractility. Decreased contractility is often a result of
structural thinning of the ventricular wall, as well as metabolic impairments. Chronic ischemia, such
as may result from CAD, may similarly impair cardiomyocyte function and cause decreased
contractility and remodeling. However, ultimately cardiomyopathies are of a complex origin
involving mismanagement of fluid balance, abnormal hormonal influences (epinephrine,
angiotensin), and insufficient perfusion/nutrition. Assessing the role of exogenous CO in altering
pathways leading to cardiomyopathy is a relatively new endeavor, and several new findings are of
great interest.
The heart is a known target for CO toxicity, potentially due to its high rate of O2 consumption.
Effects of CO on the healthy heart have only been observed at relatively high concentrations. For
example, a recent study by Sorhaug et al. (2006, 180414) demonstrated cardiac hypertrophy in rats
exposed for 72 wk to 200 ppm CO. COHb levels were reported to be 14.7%. Neither structural signs
of hypertension in the pulmonary arteries nor atherosclerotic lesions in the systemic arteries were
observed. A follow-up study by the same investigators (Bye et al., 2008, 193777) found reduced
aerobic capacity and contractile function leading to pathologic cardiac hypertrophy in rats exposed
for 18 mo to 200 ppm CO. Cardiac hypertrophy was also demonstrated in rats exposed to
100-200 ppm CO for 1-2 wk (Loennechen et al., 1999, 011549). This response was accompanied by
an increase in endothelin-1 expression. COHb levels were reported to be 12-23% in this latter study.
Effects of CO on the healthy heart have also been demonstrated following short-term
exposures. In a study by Favory et al. (2006, 184462) rats were exposed to 90 min of 250 ppm CO,
which led to peak COHb values of roughly 11%; recovery of 96 h was needed for COHb levels to
return to baseline. The authors noted that within the first 24 h of recovery, while COHb values
decreased from 11% to 5%, the coronary vascular perfusion pressure and the left ventricular
developed pressure were significantly increased compared to baseline. Concomitantly, the ratio of
cGMP to cAMP decreased, and the sensitivity of the coronary vascular bed to both acetylcholine and
a NO donor was reduced by CO exposure. The authors concluded that the discordant alterations in
contractility (increased) and perfusion (decreased) may place the heart at risk of O2 limitations
following this exposure to CO.
Several studies examined the impact of lower levels (50 ppm) on preexisting or concurrent
cardiac pathologies. In one such study, CO exacerbated the effects of a hypoxia-based model of right
ventricular remodeling and failure (Gautier et al., 2007, 096471). In controlled laboratory settings,
chronic hypobaric hypoxia (HH) caused right ventricular hypertrophy as a result of pulmonary
arterial vasoconstriction and increased pulmonary resistance. Using such a model (Wistar rats
exposed for 3 wk to hypoxia), CO (50 ppm during the last week of hypoxia, continuous) only
increased COHb from 0.5% to 2.4% in the hypoxia model, yet had significant effects on blocking
compensatory functional responses to hypoxia, such as increased fractional shortening and
contractility. Also, while right ventricular weight was increased by hypoxia alone, significant
pathology related to necrosis was observed in the hypoxia + CO-exposed rats. The reduced coronary
perfusion of the right ventricle in hypoxia + CO-exposed rats may help explain the histopathologic
findings. The authors cited previous work demonstrating that exogenous CO can inhibit NOS
(Thorup et al., 1999, 193782). which is essential for coronary dilation and angiogenesis. Thus, this
study provided evidence that exogenous CO may interrupt or downregulate pathways that
endogenous CO may activate.
In two studies by Melin et al. (2002, 037502: 2005, 193833). Dark Agouti rats were exposed
for 10 wk to either HH, 50 ppm CO or HH plus 50 ppm CO. CO exposure amplified the right
ventricular cardiac hypertrophy and decreased the right ventricular diastolic function which occurred
in response to HH. In addition, the combined exposure led to effects on left ventricular morphology
and function which were not seen with either exposure alone. Changes in HRV were also reported.
Results from both of these studies combined with results of Gautier and colleagues (Gautier et al.,
2007, 096471) indicated that CO may interfere with normal homeostatic responses to hypoxia. This
January 2010 5-46
-------
could occur by blocking HIF-la-responsive elements (vascular endothelial growth factor,
erythropoietin) or other cell signaling pathways.
In a similar study, Carraway et al. (2002, 026018) exposed rats to HH (380 torr) with or
without co-exposure to CO (50 ppm). These exposures were continuous for up to 21 days and
focused on pulmonary vascular remodeling. While the addition of CO to HH did not alter the
thickness or diameter of vessels in the lung, there was a significant increase in the number of small
(<50 um) diameter vessels compared to control, HH-only, and CO-only exposures. Despite the
greater number of vessels, the overall pulmonary vascular resistance was increased in the combined
CO + hypoxic exposure, which the authors attributed to enhancement of muscular arterioles and
p-actin. Results of this study, taken together with results from the studies of Gautier et al. (2007,
096471) and Melin et al. (2002, 037502: 2005, 193833). suggested that the combined effect of low
levels of CO with hypoxia is an enhanced right ventricle workload and an exacerbated
cardiomyopathy related to pulmonary hypertension. The population at risk of primary pulmonary
hypertension is low, but secondary pulmonary hypertension is a frequent complication of COPD and
certain forms of heart failure.
5.2.5.3. Electrocardiographs Effects
In two related studies, Wellenius et al. (2004, 087874; 2006, 156152) examined the effects of
CO in an animal model of post-infarction myocardial sensitivity (Wellenius et al., 2002, 025405). In
a previous study, ECG changes were observed during exposure to residual oil fly ash (ROFA)
particles in anesthetized post-Mi Sprague Dawley rats (Wellenius et al., 2002, 025405). Using this
model, Wellenius and colleagues tested the effects of 35 ppm CO (1-h exposure) on the induction of
spontaneous arrhythmias (Wellenius et al., 2004, 087874). CO exposure caused a statistically
significant decrease (60.4%) in ventricular premature beat (VPB) frequency during the exposure
period in rats with a high number of pre-exposure VPB. No interaction was observed with co-
exposure to carbon concentrated particles, which independently reduced VPB frequency during the
postexposure period when administered alone. In a follow-up publication, results from the analysis
of supraventricular ectopic beats (SVEB) were provided (Wellenius et al., 2006, 156152). A decrease
in the number of SVEB was observed with CO (average concentration 37.9 ppm) compared to
filtered air. While the authors concluded that CO exposure did not increase risk of SVEB in this
particular rodent model of coronary occlusion, the fact that cardiac electrophysiological dynamics
are significantly altered by short-term exposure to low-level CO may be of concern for other models
of susceptibility.
5.2.5.4. Summary of Cardiovascular Toxicology
Experimental studies demonstrated that short-term exposure to 50-100 ppm CO resulted in
aortic injury as measured by increased nitrotyrosine and the sequestration of activated leukocytes in
healthy rats. In addition, skeletal muscle microvascular permeability was increased. Short-term
exposure to 35 ppm CO altered cardiac electrophysiology in a rat model of arrhythmia. Furthermore,
short-term exposure to 50 ppm CO exacerbated cardiac pathology and impaired function in an
animal model of hypertrophic cardiomyopathy and enhanced vascular remodeling and increased
pulmonary vascular resistance in an animal model of pulmonary hypertension. Ventricular
hypertrophy was observed in healthy rats in response to chronic exposures of 100-200 ppm CO.
These studies provide some support for the development of adverse health effects resulting from
exposures to CO at environmentally-relevant concentrations.
5.2.6. Summary of Cardiovascular Effects
5.2.6.1. Short-Term Exposure to CO
The most compelling evidence of a CO-induced effect on the cardiovascular system at COHb
levels relevant to the current NAAQS comes from a series of controlled human exposure studies
among individuals with CAD. These studies, described in the 1991 (U.S. EPA, 1991, 017643) and
January 2010 5-47
-------
2000 (U.S. EPA, 2000, 000907) CO AQCDs, demonstrate consistent decreases in the time to onset of
exercise-induced angina and ST-segment changes following CO exposures resulting in COHb levels
of 2-6% (Section 5.2.4). No human clinical studies have been designed to evaluate the effect of
controlled exposures to CO resulting in COHb concentrations lower than 2%. Human clinical studies
published since the 2000 CO AQCD (U.S. EPA, 2000, 000907) have reported no association
between CO and ST-segment changes or arrhythmia; however, none of these studies included
individuals with diagnosed heart disease.
While the exact physiological significance of the observed ST-segment changes among
individuals with CAD is unclear, ST-segment depression is a known indicator of myocardial
ischemia. It is also important to note that the individuals with CAD who participated in these
controlled exposure studies may not be representative of the most sensitive individuals in the
population. It is conceivable that the most sensitive individuals respond to COHb concentrations
lower than those evaluated in studies of controlled human exposures. Variability in activity patterns
and severity of disease among individuals with CAD is likely to influence the critical level of COHb
which leads to adverse cardiovascular effects.
The degree of ambient CO exposure which leads to attainment of critical levels of COHb will
also vary between individuals. Although endogenous COHb is generally <1% in healthy individuals,
higher endogenous COHb levels are observed in individuals with certain medical conditions.
Nonambient exposures to CO, such as exposure to ETS, may increase COHb above endogenous
levels, depending on the gradient of pCO. Ambient exposures may cause a further increase in COHb.
Modeling results described in Chapter 4 indicate that increases of-1% COHb are possible with
exposures of several ppm CO, depending on exposure duration and exercise level.
Findings of epidemiologic studies conducted since the 2000 CO AQCD (U.S. EPA, 2000,
000907) are coherent with results of the controlled human exposure studies. These recent studies
observed associations between ambient CO concentration and ED visits and hospital admissions for
IHD, CHF and cardiovascular disease as a whole and were conducted in locations where the mean
24-h avg CO concentrations ranged from 0.5 ppm to 9.4 ppm (Table 5-7). All but one of these
studies that evaluated CAD outcomes (IHD, MI, angina) reported positive associations (Figure 5-2).
Although CO is often considered a marker for the effects of another traffic-related pollutant or mix
of pollutants, evidence indicates that CO associations generally remain robust in copollutant models
and supports a direct effect of short-term ambient CO exposure on CVD morbidity. These studies
add to findings reported in the 2000 CO AQCD (U.S. EPA, 2000, 000907) that demonstrated
associations between short-term variations in ambient CO concentrations and exacerbation of heart
disease.
The known role of CO in limiting O2 availability lends biological plausibility to ischemia-
related health outcomes following CO exposure. However, it is not clear whether the small changes
in COHb associated with ambient CO exposures results in substantially reduced O2 delivery to
tissues. Recent toxicological studies suggest that CO may also act through other mechanisms by
initiating or disrupting cellular signaling. Studies in healthy animals demonstrated oxidative injury
and inflammation in response to 50-100 ppm CO, while studies in animal models of disease
demonstrated exacerbation of cardiomyopathy and increased vascular remodeling in response to
50 ppm CO. Further investigations will be useful in determining whether altered cell signaling
contributes to adverse health effects following ambient CO exposure.
Given the consistent and coherent evidence from epidemiologic and human clinical studies,
along with biological plausibility provided by CO's role in limiting O2 availability, it is concluded
that a causal relationship is likely to exist between relevant short-term exposures to CO
and cardiovascular morbidity.
5.2.6.2. Long-Term Exposure to CO
Only two epidemiologic studies were identified that investigated the relationship between
long-term exposure to CO and cardiovascular effects, and the results of these studies provide very
limited evidence of an association. Considering the lack of evidence from controlled human
exposure studies and the very limited evidence from toxicological studies on cardiovascular effects
following long-term exposure to CO, the available evidence is inadequate to Conclude that 3
causal relationship exists between relevant long-term exposures to CO and
cardiovascular morbidity.
January 2010 5-48
-------
5.3. Central Nervous System Effects
5.3.1. Controlled Human Exposure Studies
The behavioral effects of controlled human exposures to CO have been examined by several
laboratories, and these studies were summarized in the 2000 CO AQCD (U.S. EPA, 2000, 000907).
Briefly, decreases in visual tracking as well as visual and auditory vigilance were observed following
exposures to CO resulting in COHb levels between 5% and 20% (Benignus et al., 1987, 012250;
Fodor and Winneke, 1972, 011041: Horvath et al., 1971, 011075: Putz et al., 1979, 023137). One
study reported similar behavioral effects (time discrimination) among a group of healthy volunteers
with COHb levels <3% (Beard and Wertheim, 1967, 011015). though subsequent studies were
unable to replicate these findings at such low exposure concentrations (Otto et al., 1979, 010863:
Stewart et al., 1973, 093412). These outcomes represent a potentially important adverse effect of CO
exposure resulting in COHb levels > 5%, although it is important to note that these findings have not
been consistent across studies. Similarly, some studies demonstrated decreases in reaction time as
well as decrements in cognitive function and fine motor skills following controlled exposures to CO;
however, these studies were not typically conducted using double-blind procedures, which may
significantly affect the outcome of behavioral studies (Benignus, 1993, 013645). It should be noted
that all behavioral studies of controlled CO exposure were conducted in normal, healthy adults. No
new human clinical studies have evaluated CNS or behavioral effects of exposure to CO.
5.3.2. lexicological Studies
The evidence for toxicological effects of CO exposure in laboratory animal models comes
from in utero or perinatal exposure involving relatively low to relatively high concentrations of CO
(12.5-750 ppm). Affected endpoints from this early, developmental CO exposure include behavior,
memory, learning, locomotor ability, peripheral nervous system myelination, auditory decrements,
and neurotransmitter changes. These data are addressed in detail in the Birth Outcomes and
Developmental Effects section of the ISA (Section 5.4.2). Further, a group of studies have found that
exposure to high concentrations of CO (500-1,200 ppm) can result in CO-dependent ototoxicity,
specifically loss of threshold of cochlear compound action potentials (CAP) and potentiation of
noise-induced hearing loss (NIHL) (Chen et al., 2001, 193985: Fechter et al., 1997, 081322: Fechter
et al., 2002, 193926: Liu and Fechter, 1995, 076524). Proposed mechanisms for these effects include
ROS generation and glutamate release.
5.3.3. Summary of Central Nervous System Effects
Exposure to high levels of CO has long been known to adversely affect CNS function, with
symptoms following acute CO poisoning including headache, dizziness, cognitive difficulties,
disorientation, and coma. However, the relationship between ambient levels of CO and neurological
function is less clear and has not been evaluated in epidemiologic studies. Studies of controlled
human exposures to CO discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907) reported
inconsistent neural and behavioral effects following exposures resulting in COHb levels of 5-20%.
No new human clinical studies have evaluated central nervous system or behavioral effects of
exposure to CO. At ambient-level exposures, healthy adults may be protected against CO-induced
neurological impairment owing to compensatory responses including increased cardiac output and
cerebral blood flow. However, these compensatory mechanisms are likely impaired among certain
potentially susceptible groups, including individuals with reduced cardiovascular function.
Toxicological studies that were not discussed in the 2000 CO AQCD (U.S. EPA, 2000,
000907) employed rodent models to show that low to moderate CO exposure during the in utero or
perinatal period can adversely affect adult outcomes, including behavior, neuronal myelination,
neurotransmitter levels or function, and the auditory system (discussed in Section 5.4). In utero CO
exposure, including both intermittent and continuous exposure, has been shown to impair multiple
behavioral outcomes in offspring including active avoidance behavior (150 ppm CO), nonspatial
memory (75 and 150 ppm CO), spatial learning (endogenous CO inhibition), homing behavior
January 2010 5-49
-------
(150 ppm CO), locomotor movement (150 ppm CO), and negative geotaxis (125 and 150 ppm). In
two separate studies, in utero CO exposure (75 and 150 ppm) was associated with significant
myelination decrements without associated changes in motor activity in adult animals. Multiple
studies demonstrated that in utero CO exposure affected glutamatergic, cholinergic,
catecholaminergic, and dopaminergic neurotransmitter levels or transmission. Possible or
demonstrated adverse outcomes from the CO-mediated aberrant neurotransmitter levels or
transmission include respiratory dysfunction (200 ppm CO), impaired sexual behavior (150 ppm
CO), and an adverse response to hyperthermic insults resulting in neuronal damage (200 ppm).
Finally, perinatal CO exposure has been shown to affect the developing auditory system of rodents,
inducing permanent changes into adulthood. This is manifested by atrophy of cochlear cells
innervating the inner hair cells (25 ppm CO), decreased immunostaining associated with impaired
neuronal activation (12.5 ppm CO), impaired myelination of auditory associated nerves (25 ppm
CO), decreased energy production in the sensory cell organ of the inner ear or the organ of corti
(25 ppm CO). Some of these changes have been proposed to be mediated by ROS. Functional tests
of the auditory system of rodents exposed neonatally to CO using OAE testing (50 ppm) and action
potential amplitude measurements of the 8th cranial nerve (12, 25, 50, 100 ppm), revealed
decrements in auditory function at PND22 and permanent changes into adulthood using action
potential (AP) testing (50 ppm). Additionally, exposure to high concentrations of CO has been
shown to result in CO-dependent ototoxicity in adult animals, possibly through glutamate and ROS-
dependent mechanisms. Together, these animal studies demonstrated that in utero or perinatal
exposure to CO can adversely affect adult behavior, neuronal myelination, neurotransmission, and
the auditory system in adult rodents. Considering the combined evidence from controlled human
exposure and toxicoiogicai studies, the evidence is suggestive of a causal relationship between
relevant short- and long-term exposures to CO and central nervous system effects.
5.4. Birth Outcomes and Developmental Effects
5.4.1. Epidemiologic Studies
Although the body of literature is growing, the research focusing on adverse birth outcomes is
limited when compared to the numerous studies that have examined the more well-established health
effects of air pollution. Among this small number of studies, various dichotomized measures of birth
weight, such as low birth weight (LEW), small for gestational age (SGA), and intrauterine growth
restriction (IUGR), have received more attention in air pollution research while preterm birth
(<37 wk gestation; [PTB]), congenital malformations, and infant mortality are less studied.
In the 2000 CO AQCD (U.S. EPA, 2000, 000907). only two studies were cited that examined
the effect of ambient air pollution on adverse birth outcomes, and both of these studies investigated
LEW as an endpoint (Alderman et al, 1987, 012243: Ritz and Yu, 1999, 086976). At that time this
area of research was in its infancy; however, there has since been increasing interest.
5.4.1.1. Preterm Birth
A small number of air pollution-birth outcome studies have investigated the possible
association between PTB and maternal exposure to CO, with the majority of U.S. studies conducted
in southern California. The earliest of these studies examined exposures to ambient CO during the
first month of pregnancy and the last 6 wk prior to birth among a cohort of 97,158 births in southern
California between 1989 and 1993 (Ritz et al., 2000, 012068). The exposure assessment within this
study was based on data from fixed-site monitors that fell within a 2-mi radius of the mother's ZIP
code area. The crude relative risks for PTB associated with a 1 ppm increase in 3-h avg CO
concentration (6:00-9:00 a.m.) during the last 6 wk prior to birth and the first month of pregnancy
were 1.04 (95% CI: 1.03-1.5) and 1.01 (95% CI: 1.00-1.03), respectively. However, when the
authors controlled for other risk factors, only the effect associated with CO during the last 6 wk prior
to birth persisted (RR: 1.02 [95% CI: 1.01-1.03]). Furthermore, when the analyses included variables
January 2010 5-50
-------
for either season or other pollutants, the CO effect estimates generally were reduced such that they
remained positive but were no longer statistically significant.
Expanding on this research, Wilhelm and Ritz (2005, 088668) examined PTB among a cohort
of 106,483 births in Los Angeles County, CA, between 1994 and 2000. Based on data recorded at
monitoring stations of varying proximities to the mother's residence, the main exposure windows
examined were the first trimester and the last 6 wk prior to birth. Among women living within a 1-mi
radius of a CO monitoring station, a 0.5 ppm increase in 24-h avg CO concentration during the first
trimester was associated with a 3% (RR: 1.03 [95% CI: 1.00-1.06]) increased risk of PTB. This
result persisted after simultaneously adjusting for NO2 and O3 (RR: 1.05 [95% CI: 1.00-1.10]) but
not with the inclusion of PMi0 into the regression model (RR: 0.99 [95% CI: 0.91-1.09]). The result
from the single pollutant model for CO exposures averaged over the 6 wk prior to birth was similar
in magnitude but failed to reach statistical significance (RR: 1.02 [95% CI: 0.99-1.04]).
A limitation of many air pollution-birth outcome studies is the limited availability of detailed
information on maternal lifestyle factors and time-activity patterns during pregnancy. To assess
possible residual confounding due to these factors, Ritz and colleagues (2007, 096146) were able to
analyze detailed maternal information from a survey of 2,543 from a cohort of 58,316 eligible births
in 2003 in Los Angeles County. Based on data from the closest monitor to the mother's ZIP code
area, exposures to CO, NO2, O3, and PM2.5 during the first trimester and last 6 wk prior to delivery
were examined, and results from the overall cohort (n = 58,316) with limited maternal information
were compared to the more detailed nested case-control cohort (n = 2,543). Within the overall
cohort, 24-h avg CO during the first trimester was associated with an increased risk of 25% (OR:
1.25 [95% CI: 1.12-1.38]; highest exposure group >1.25 ppm versus lowest < 0.58 ppm). This result
persisted within the nested case-control cohort (OR: 1.21 [95% CI: 0.88-1.65]) where factors such as
passive smoking and alcohol use during pregnancy were included in the model; however, the
confidence intervals were wider due to the smaller sample. Any possible association between CO
and PTB was less evident during the last 6 wk prior to birth. A strength of this study was that it also
highlighted how there was little change in the air pollution effect estimates when controlling for
more detailed maternal information (e.g., smoking, alcohol use), as opposed to only controlling for
more limited maternal information that is routinely collected on birth registry forms.
In contrast to the Los Angeles studies, a case-control study of PTB across California for the
period 1999 through 2000 found a positive association with 24-h CO concentration during the entire
pregnancy (OR: 1.03 [95% CI: 0.98-1.09] per 0.5 ppm increase) and the first month of gestation
(OR: 1.05 [95% CI: 0.99-1.10] per 0.5 ppm increase), but no association during the last 2 wk of
gestation (OR: 1.00 [95% CI: 0.96-1.04] per 0.5 ppm increase) (Huynh et al., 2006, 091240).
Although there was an indication of an effect during early pregnancy, the small sample size (when
compared to other studies) may not have provided sufficient power to detect statistical significance.
Furthermore, exposures within this study were assigned based on a county-level average which may
explain the lack of effect, given the poor level of exposure assessment.
Studies outside of the U.S. have been conducted in Canada, Australia, and Korea, with mixed
results reported. In Vancouver, Canada, based on a city-wide average across available monitoring
sites, 24-h avg CO concentration during the last month of pregnancy was associated with a 4% (OR:
1.04 [95% CI: 1.00-1.07]) increased risk of PTB per 0.5 ppm increase, while there was no
association found during the first month of pregnancy (OR: 0.98 [95% CI: 0.94-1.00]) (Liu et al.,
2003, 089548). This study investigated maternal exposures to ambient gaseous pollutants (CO, NO2,
SO2, O3) averaged over the first and last month of pregnancy among a cohort of 229,085 births
between 1985 and 1998.
In a cohort of 52,113 births in Incheon, Korea, between 2001 and 2002, a kriging technique
was used to assign the maternal exposures to CO. Kriging is a statistical mapping technique that
allows the prediction of an average concentration over a spatial region from data collected at specific
points. The spatial average CO concentrations were then linked to each study subject's residential
address. CO concentrations during the first trimester were associated with a 26% (RR: 1.26
[95% CI: 1.11-1.44]) increased risk of PTB for the highest quartile of exposure when compared to
the lowest quartile (Leem et al., 2006, 089828). There was also a strong significant trend exhibited
across the quartiles. A similar result was found for 24-h avg CO concentration during the last
trimester, although the effect was less pronounced (RR: 1.16 [95% CI: 1.01-1.24]).
Conversely, a study in Sydney, Australia, examined maternal exposure to ambient air pollution
during the first and last month and the first and last trimester of pregnancy among a cohort of
123,840 births between 1998 and 2000 and found no association between PTB and CO (Jalaludin et
January 2010 5-51
-------
al., 2007, 156601). Maternal exposure estimates in this study were based on a city-wide average of
available monitoring sites and also based on data from fixed sites within 5 km of the mother's
postcode area. The odds ratios for PTB associated with 8-h avg CO concentrations during the first
trimester and last 3 mo of gestation were 1.18 (95% CI: 0.85-1.63) and 1.08 (95% CI: 0.95-1.22),
respectively, when including births within 5 km of a monitor. Interestingly, when all births were
included in the analyses and the exposure was based on a city-wide average, these effects had
become protective for the first trimester (OR: 0.82 [95% CI: 0.77-0.87]) and null for the last 3 mo of
gestation (OR: 0.99 [95% CI: 0.92-1.07]). This suggests that exposures based on data from fixed
sites closer to the mother's address are more likely to detect an effect than a city-wide average.
Figure 5-9 shows the odds for the risk of delivering a preterm infant from the reviewed
studies; Table 5-12 provides a brief overview of the PTB studies. In summary, there are mixed
results across the studies. Although these studies are difficult to compare directly due to the different
exposure assessment methods employed, there is some evidence that CO during early pregnancy
(e.g., first month and trimester) is associated with an increased risk of PTB. The most consistency is
exhibited within the studies conducted around Los Angeles, CA, and surrounding areas, whereby all
studies reported a significant association with CO exposure during early pregnancy, and exposures
were assigned from monitors within close proximity of the mother's residential address (Ritz et al.,
2000, 012068: Ritz et al., 2007, 096146: Wilhelm and Ritz, 2005, 088668). It should also be noted
that the mixed results when analyzing different cohorts that resided within varying proximities to a
monitor may be attributable to analyzing different populations.
Study
Ritz etal. (2000,012068)
Ritz etal. (2000,012068)
Wilhelm & Ritz (2005, 088668)
Wilhelm & Ritz (2005, 088668)
Huvnh etal. (2006,091240)
Huvnh etal. (2006,091240)
Huvnh etal. (2006, 091240)
Liu etal. (2003, 089548)
Location
California, US
California, US
Los Angeles, CA
Los Angeles, CA
California, US
California, US
California, US
Vancouver, Can
Exposure Period
First month
Last 6 weeks
First trimester
Last 6 weeks
First month
Last 2 weeks
Entire pregnancy
First month
Exposure
Indicator
<2 mi of monitor
<2 mi of monitor
ZIP code level
ZIP code level
County level
County level
County level
City wide
Effect Estimate (95% CI)
•[«-
*
'-•-
r«-
^ •
— I
— r-»
— «-"•
Liu etal. (2003, 089548)
Jalaludinetal. (2007, 156601)
Jalaludinetal. (2007, 156601)
Vancouver, Can
Sydney, Australia
Sydney, Australia
Last month
First month
First trimester
City wide
City wide — • —
City wide — • —
•
Jalaludin et al. (2007,156601) Sydney, Australia
Last month
City wide
0.75
1.00
Odds Ratio
1.25
Figure 5-9.
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.
January 2010
5-52
-------
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 et al. (2003. 089548)
Leem et al. (2006, 089828)
Jalaludin et al. (2007, 156601)
Location
Sample Size
California, US
(n = 97,158)
Los Angeles, CA
(n = 106,483)
Los Angeles, CA
(n = 58,31 6)
California, US
(n = 42,692)
Vancouver, Can
(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
First mo
Last 6 wk
Last 6 wk
Entire pregnancy
First Trimester
Last 6 wk
Entire pregnancy
First mo
Last 2 wk
First mo
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
With birth weight routinely collected in vital statistics and being a powerful predictor of infant
mortality, it is the most studied outcome within air pollution-birth outcome research. Air pollution
researchers have analyzed birth weight as a continuous variable and/or as a dichotomized variable in
the forms of LEW (<2,500 g [5 Ibs, 8 oz]) and SGA.
It should be noted that the terms SGA, which is defined as a birth weight <10th percentile for
gestational age (and often sex), and IUGR are used interchangeably. However, this definition of SGA
does have limitations. For example, using it for IUGR may overestimate the percentage of "growth-
restricted" neonates as it is unlikely that 10% of neonates have growth restriction (Wollmann, 1998,
193812). On the other hand, when the 10th percentile is based on the distribution of live births at a
population level, the percentage of SGA among preterm births is most likely underestimated
(Hutcheon and Platt, 2008, 193795). Nevertheless, it should be noted that SGA represents a
statistical description of a small neonate, whereas the term IUGR is reserved for those with clinical
evidence of abnormal growth. Thus, all IUGR neonates will be SGA, but not all SGA neonates will
be IUGR (Wollmann, 1998, 193812). In the following sections the terms SGA and IUGR are referred
to as each cited study used the terms.
Over the past decade a number of studies examined various metrics of birth weight in relation
to maternal exposure to CO with the majority conducted in the U.S. Given that most studies
examined multiple birth weight metrics, the following section focuses on each study only once and
presents results for each metric within that study.
Most of the U.S. studies have been conducted in southern California, with inconsistent results
reported with regard to gestational timing of the CO effects. The first of these studies was reviewed
in the 2000 CO AQCD (U.S. EPA, 2000, 000907) and is briefly summarized here. Ritz and Yu
(1999, 086976) examined the effect of ambient CO during the last trimester on LEW among 125,573
births in Los Angeles between 1989 and 1993. When compared to neonates born to women in the
lowest CO exposure group (<2.2 ppm), neonates born to women in the highest exposure group
(5.5 ppm-95th percentile) had a 22% (OR: 1.22 [95% CI: 1.03-1.44]) increased risk of being born as
LEW
Building upon this research, Wilhelm and Ritz (2005, 088668) reported similar results when
extending this study to include 136,134 births for the period 1994-2000. Exposure to ambient CO
January 2010
5-53
-------
during each trimester was based on data recorded at monitoring stations of varying proximities to the
mother's residence. For women residing within 1 mi of a station, there was 36% (OR: 1.36
[95% CI: 1.04-1.76]) increased risk of having a term LEW baby for women with third-trimester
exposure above the 75th percentile when compared to women below the 75th percentile. There was
also an increased risk of term LEW (OR: 1.28 [95% CI: 1.12-1.47]) among women in the highest
exposure group when the analyses included women within a 5-mi radius of a station. However, when
the analyses included women within a 1- to 2-mi or 2- to 4-mi radius of a station, the CO effects
failed to reach statistical significance, and there was no evidence of an exposure-response pattern
exhibited across the varying distances to a station. Furthermore, none of the significant CO results
persisted after controlling for other pollutants. Although standard errors were certainly increased
after controlling for the other pollutants, leading to non-significant results, some of the effect sizes
were similar, providing some consistency. It is interesting to note, however, that maternal exposure
to CO during trimesters one and two was not associated with LEW (quantitative results not reported
by authors).
Further validation in association with exposure times was observed in an analysis using a
subset of participants in the Children's Health Study. Salam and colleagues (2005, 087885) found
that CO only during the first trimester was associated with reduced fetal growth. Their research
examined birth weight, LEW, and IUGR among a subset of participants in the Children's Health
Study (Peters et al, 1999, 087243) who were born in California between 1975 and 1987 (n = 3,901).
The study examined term births with a gestational age between 37 and 44 wk. Exposures in this
study were based on CO data from up to the 3 nearest monitoring sites within 50 km of the centroid
of the mother's ZIP code. Exposures for the entire pregnancy and each trimester were analyzed, and
a 0.5 ppm increase in 24-h CO concentration during the first trimester was associated with a 7.8 g
(95% CI: 15.1-0.4) decrease in birth weight, which also translated to a 6.7% (OR: 1.07 [95% CI:
1.00-1.13]) increased risk of IUGR; however, there was no association with LEW (OR: 1.00
[95% CI: 0.88-1.16]).
In contrast to the previous studies, another California study of 18,247 singleton births born at
40-wk gestation during 2000 found no association between ambient 24-h CO concentration and
reduced birth weight or SGA, where the highest quartile of exposure was 0.98 ppm. Based on data
from fixed sites within 5 mi of the mother's residence, exposures to CO and PM25 during the entire
pregnancy and each trimester were analyzed. Although CO during the entire pregnancy was
associated with a 20 g (95% CI: 40.1-0.8) reduction in birth weight, this did not persist after
controlling for PM2 5. PM2 5 was found to have a strong effect on birth weight within each trimester
(Parker et al., 2005, 087462).
Two similar studies were conducted in the northeastern U.S. with inconsistent results. A study
of 89,557 singleton term births in Boston, MA, Hartford, CT, Philadelphia, PA, Pittsburgh, PA, and
Washington, DC, between 1994 and 1996 found that exposure to ambient 24-h avg CO during the
third trimester was associated with an increased risk of LBW (OR: 1.14 [95% CI: 1.03-1.27] per
0.5 ppm increase) (Maisonet et al., 2001, 016624). When stratified by race this effect was only
significant among African-Americans for the first and third trimesters (first OR: 1.32
[95% CI: 1.22-1.43]; third OR: 1.20 [95% CI: 1.09-1.32]). Exposures to PM10and SO2were
examined, and there was no strong evidence that these pollutants were associated with LBW.
Exposures for this study were based on a city-wide average of monitors within the mother's city of
residence. The second study examined 358,504 births at 32- to 44-wk gestation between 1999 and
2002 in Connecticut and Massachusetts (Bell et al., 2007, 091059). and 24-h CO exposures were
estimated from fixed sites within each mother's county of residence (e.g., county level). CO
averaged over the entire pregnancy was associated with a reduction in birth weight of 27.0 g
(95% CI: 21.0-32.8). This result persisted after controlling for each additional pollutant (PMi0,
PM25, NO2, and SO2) in two-pollutant models. However, this reduction in birth weight did not
translate to an increased risk of LBW (OR: 1.05 [95% CI: 0.97-1.12] per 0.5 ppm increase in CO).
When controlling for exposure during each trimester, the reduction in birth weight associated with a
0.5 ppm increase in 24-h CO concentration during the first trimester ranged from 18.8 to 16.5 g,
while the reductions associated with third trimester exposure ranged between 27.2 and 23.3 g. It is
interesting to note that, although the exposures were based on data averaged at the county level, CO
was associated with a reduction in birth weight. In contrast, in a previously cited California study by
Huynh and colleagues (2006, 091240) exposures were also at the county level yet there was no
association with PTB. This difference may be due to the counties being smaller in New England than
in California, resulting in more precise exposure estimates.
January 2010 5-54
-------
Two studies in Canada investigated the effects of ambient air pollution on fetal growth with
exposures derived from a city-wide average across the available monitoring sites. The first of these
studies was among a cohort of 229,085 singleton term births (37- and 42-wk gestation) in Vancouver,
BC, with monthly and trimester exposures to CO investigated in relation to LEW and IUGR (Liu et
al., 2003, 089548). For a 0.5 ppm increase in 24-h CO concentration during the first month of
pregnancy, there was an increased risk of IUGR (OR: 1.03 [95% CI: 1.00-1.05]), and this was of
borderline significance when CO was averaged over the first trimester (OR: 1.02
[95% CI: 1.00-1.05]). This result persisted after controlling for other gaseous pollutants. Conversely,
maternal exposure to CO was not associated with LEW. The more recent of these two studies
examined 386,202 singleton term births (37- to 42-wk gestation) in Calgary, Edmonton, and
Montreal, between 1986 and 2000 (Liu et al., 2007, 090429). The study examined monthly and
trimester exposures to CO with IUGR being the only endpoint. A 0.5 ppm increase in 24-h CO
concentration was associated with an increased risk of 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 trimesters (OR: 1.09
[95% CI: 1.07-1.11]) of pregnancy. This result translated to CO exposure having a positive effect on
IUGR within each individual month of pregnancy with the highest effect during the first and last
months. This result persisted after controlling for concurrent NO2 and PM2.5.
Two studies in Sao Paulo, Brazil, a city with notably high levels of air pollution (mean CO
3.7 ppm) investigated associations between maternal exposures to CO in relation to reduced birth
weight and LEW within two consecutive time periods and found similar results. In both studies the
exposures were derived from a city-wide average across the available monitoring sites. The first
study examined 179,460 singleton term births during 1997 and found that a 0.75 ppm increase in 8-h
CO concentration averaged over the first trimester was associated with a 17.3 g (95% CI: 31.0-3.7)
reduction in birth weight (Gouveia et al., 2004, 055613). The second of these studies examined
311,735 singleton births (37- to 41-wk gestation) between 1998 and 2000 and reported a 6.0 g (95 %
CI: 7.75-4.1) reduction in birth weight associated with a 0.5 ppm increase in 24-h CO concentration
averaged over the first trimester (Medeiros and Gouveia, 2005, 089824). It is important to note that
neither of these studies found an association between CO exposure and an increased risk of LEW.
Therefore, despite CO during the first trimester being associated with reduced birth weight, it was
not associated with LEW.
Similar to the two studies in Sao Paulo, Brazil, researchers in Seoul, South Korea, conducted
two studies using data from two consecutive time periods. Both of these studies based the exposure
estimates on a city-wide average from all available fixed sites and as would be expected, the results
pertaining to CO were similar for both studies. Ha and colleagues (2001, 019390) examined
maternal exposures to CO during the first and third trimesters among 276,763 singleton term births
in Seoul between 1996 and 1997. Exposure to CO during the first trimester was associated with a
decrease in birth weight of 13.3 g, which also translated into an increased risk of LEW (RR: 1.10
[95% CI: 1.05-1.14] per 0.5 ppm increase in 24-h CO concentration). When Lee and colleagues
(2003, 043202) extended this study to include singleton term births for the period 1996-1998, with
24-h CO concentrations averaged over each month of pregnancy and trimester, CO exposure during
the first trimester was associated with an increased risk of LEW (OR: 1.04 [95% CI: 1.01-1.07] per
0.5 ppm increase). No associations were found in the third trimester for any of the pollutants.
Monthly-specific exposures showed that the risk of LEW tended to increase with CO exposure
between months two through five of pregnancy.
In contrast to other studies reporting that early and late pregnancy are the critical periods for
CO exposure, a Sydney, Australia study of 138,056 singleton births between 1998 and 2000 reported
a 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 a 0.75 ppm increase in maternal exposure to 8-h CO averaged over the second and third
trimesters, respectively (Mannes et al., 2005, 087895). However, this result did not persist after
controlling for other pollutants (PMi0, NO2) and was only statistically significant when including
births where the mother resided within 5 km of a monitor. Furthermore, this result did not translate to
an increased risk of SGA, which was defined as a birth weight two standard deviations below the
mean. The odds ratios for SGA for CO exposures during the first, second and third trimesters were
0.96 (95% CI: 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 CO, respectively. While the majority of studies restrict the analyses to term births as a
method of controlling for gestational age, it is important to note that the Sydney study used all births
and controlled for gestational age in the birth weight analyses and SGA was derived for each
gestational age group.
January 2010 5-55
-------
Of all studies reviewed, only two did not find an association between maternal exposure to CO
and birth weight variables. In northern Nevada, Chen and colleagues (2002, 024945) examined CO,
PMio, and O3 exposures among a cohort of 39,338 term births (37- to 44-wk gestation) between 1991
and 1999 and found no association between CO exposure during the entire pregnancy (and each
trimester) and a reduction in birth weigh or an increased risk of LEW. For a 0.75 ppm increase in 8-h
CO concentration averaged over the entire pregnancy, there was a reduction in birth weight of 6 g;
however it failed to reach statistical significance. Exposures for this study were based on data from
all monitoring sites across Washoe County, Nevada.
In a retrospective cohort study among 92,288 singleton term births (37- to 44-wk gestation) in
Taipei and Kaoshiung, Taiwan, between 1995 and 1997, maternal exposures to CO, SO2, O3, NO2,
and PMioin each trimester of pregnancy were examined, and only SO2 during the third trimester
showed evidence of contributing to LEW. Exposure assessment was based on data from the monitor
closest to the centroid of the mother's residential district, and the final analyses included only those
mothers whose district centroid was within 3 km of a monitor. CO exposures were grouped into low
(~1.1 ppm), medium (-1.2-15.0 ppm), and high (>15.0 ppm) and when compared to the lowest
exposure group, the odds ratio for LEW in the highest exposure group was 0.90 (95% CI: 0.75-1.09)
for the first trimester, 1.00 (95% CI: 0.82-1.22) for the second trimester, and 0.86 (95% CI: 0.71-
1.03) for the third trimester (Lin et al., 2004, 089827).
Table 5-13 provides a brief overview of the birth weight studies. In summary, there is evidence
of ambient CO during pregnancy having a negative effect on fetal growth. From the reviewed studies
Figure 5-10 shows the change in birth weight (grams), Figure 5-11 shows the effect estimates for
LEW, and Figure 5-12 shows the effect estimates for SGA. In general the reported reductions in
birth weight are small (~10-20g). It is difficult to conclude whether CO is related to a small change
in birth weight in all births across the population or a marked effect in some subset of births.
Furthermore, there is a large degree of inconsistency across these studies. This may be due to several
factors such as inconsistent exposure assessment and statistical methods employed, different CO
concentrations, and/or different demographics of the birth cohorts analyzed. The main inconsistency
among these findings is the gestational timing of the CO effect. Although the majority of studies
reported significant effects during either the first or third trimester, other studies failed to find a
significant effect during these periods. Several studies found an association with exposure during the
entire pregnancy, providing evidence for a possible accumulative effect; however, these results are
inconclusive and may be the result of correlated exposure periods.
Several studies examined various combinations of birth weight, LEW, and SGA/IUGR, and
inconsistent results are reported across these metrics. For example, several studies reported an
association between maternal exposure to CO and decreased birth weight, yet no increase in risk of
LEW or SGA. However, a measurable change, even if only a small one, at the population level is
different than an effect observed for a subset of susceptible births which may increase the risk of
IUGR/LBW/SGA. It is difficult to conclude if CO is related to a small change in birth weight in all
births across the population or a marked effect in some subset of births.
January 2010 5-56
-------
Study
Bell etal. (2007,091059)
Salametal. (2005,087885)
Salam etal. (2005,087885)
Salametal. (2005,087885)
Chen etal. (2002,024945)
Chen etal. (2002, 024945)
Chen etal. (2002,024945)
Mannes etal. (2005,087895)
Mannes etal. (2005,087895)
Mannes etal. (2005,087895)
Mannes etal (2005 087895)
Gouveia etal. (2004,055613)
Gouveia etal. (2004,055613)
Gouveia etal. (2004,055613)
Medeiros and Gouveia (2005, 089824)
Medeiros and Gouveia (2005, 089824)
Medeiros and Gouveia (2005, 089824)
Location
CT&MA.US
California, US
California, US
California, US
Northern NV, US
Northern NV, US
Northern NV, US
Sydney, Australia
Sydney, Australia
Sydney, Australia
Sao Paulo, Brazil
Sao Paulo, Brazil
Sao Paulo, Brazil
Sao Paulo, Brazil
Sao Paulo, Brazil
Sao Paulo, Brazil
Exposure
Period
Entire pregnancy
First trimester
Second trimester
Third trimester
Entire pregnancy
First trimester
Second trimester
Third trimester
Entire pregnancy
First trimester
Second trimester
Third trimester
First trimester
Second trimester
Third trimester
First trimester
Second trimester
Third trimester
Exposure
Indicator
County wide
ZIP code level
ZIP code level
ZIP code level
County level
County level
County level
City level
City level
City level
City level
City level
City level
City level
City level
City level
Effect Estimate (95% Cl)
• ,
• 1
— i— •
— i — •
*
•
•
>•
• 1-
• — i —
• i
• '
!-•
'-•
-9- '
'-e_
! -•-
i r
-35 -25 -15 -50 5 10 20
Change in birth weight (grams)
Figure 5-10. 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.
Study
Bell etal. (2007.091059)
Maisonet etal. (2001,016624)
Maisonet etal. (2001,016624)
Maisonet et al. (2001 , 016624)
Wilhelm & Ritz (2005, 088668)
Salam etal. (2005, 087885)
Salametal. (2005,087885)
Salametal. (2005.087885)
Salam etal. (2005, 087885)
Liu etal. (2003, 089548)
Liu etal. (2003. 089548)
Ha etal. (2001. 01 9390)
Ha etal. (2001,019390)
Lee etal. (2003,043202)
Lee etal. (2003,043202)
Lee etal. (2003, 043202)
Lee etal. (2003,043202)
Location
CT&MA.US
Northeastern US
Northeastern US
Northeastern US
LA County, CA
California, US
California, US
California, US
California, US
Vancouver, Can
Vancouver, Can
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Seoul, Korea
Exposure
Period
Entire pregnancy
First trimester
Second trimester
Third trimester
Third trimester
First trimester
Second trimester
Third trimester
Entire pregnancy
First month
Second month
First trimester
Third trimester
First trimester
Second trimester
Third trimester
Entire pregnancy
Exposure Effect Estimate (95% Cl)
Indicator
County level — ] — •
City level j — •
City level 1 •
City level ' •
ZIP code level ' — •—
ZIP code level •— i
ZIP code level • > —
City level — ; —
City level — »-r-
City level i — • —
City level — • '
City level |— •—
City level i-»—
City level — •— >
City level ' — • —
I 1 1
0.75 1.00 1.25
Odds Ratio
Figure 5-11. 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.
January 2010
5-57
-------
Study
Salametal. (2005.087885)
Salametal. (2005.087885)
Salametal. (2005. 087885)
Salametal. (2005.087885)
Liu etal. (2003, 089548)
Liu etal. (2003, 089548)
Liu etal. (2003, 089548)
Liu etal. (2003, 089548)
Liu etal. (2003. 089548)
Liu etal. (2007, 090429)
Liu etal. (2007, 090429)
Liu etal. (2007, 090429)
Mannes et al. (2005, 087895)
Mannes etal. (2005, 087895)
Mannes etal. (2005,087895)
Mannes etal. (2005,087895)
Location
California, US
California, US
California, US
California, US
Vancouver, Can
Vancouver, Can
Vancouver, Can
Vancouver, Can
Vancouver, Can
Multicity, Can
Multicity, Can
Multicity, Can
Sydney, Australia
Sydney, Australia
Sydney, Australia
Sydney, Australia
Exposure Period
First trimester
Second trimester
Third trimester
Entire pregnancy
First month
First trimester
Second trimester
Third trimester
Last month
First trimester
Second trimester
Third trimester
First trimester
Second trimester
Third trimester
Last month
InS? Effect Estimate (95% Cl)
ZIP code level — f—
ZIP code level • —
ZIP code level — *
City level |-»-
City level r*-
City level — •+
City level -*-1-
City level -»j-
City level i -•-
City level ' -•-
City level j -•-
City level • — \ —
City level *
City level *
City level -j — •
i I i
0.75 1.00 1.25
Odds Ratio
Figure 5-12.
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.
The possibility exists that the small reductions in birth weight associated with maternal CO
exposures are the result of residual confounding associated with other factors (e.g., other pollutants,
temperature, and spatial/temporal variation in maternal factors) or other correlated pollutants. For
example, in some studies the CO effect did not persist after controlling for other pollutants (Mannes
et al., 2005, 087895: Parker et al., 2005, 087462; Wilhelm and Ritz, 2005, 088668). while in some
studies it did persist (Bell et al., 2007, 091059; Gouveia et al., 2004, 055613; Liu et al., 2003,
089548). and other studies did not report results from multipollutant models (Ha et al., 2001,
019390; Lee et al., 2003, 043202; Maisonet et al., 2001, 016624; Medeiros and Gouveia, 2005,
089824). In addition, various methods have been employed to control for seasonality and trends
(e.g., month of birth, season of birth, year of birth, smoothed function of time), which may explain
some of the mixed results.
The two U.S. studies conducted in the Northeast compared results from analyses stratified by
race. The earlier of these studies found an association between CO and LEW among African-
Americans but not among whites and Hispanics (Maisonet et al., 2001, 016624). In contrast, despite
reporting an llg reduction in birth weight among African-Americans and a 17 g reduction among
whites, the more recent of the two studies found no significant difference between these reductions
by race (Bell et al., 2007, 091059). Parker and colleagues (2005, 087462) also tested for interactions
between race and found no significant association.
January 2010
5-58
-------
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, US
(n = 3901)
California, US
(n = 18,247)
Boston, MA
Hartford, CT
Philadelphia &
Pittsburg, PA
Washington DC
(n = 103, 465)
CT&MA.US
(n = 358,504)
Northern Nevada, US
(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, Can
(n = 229,085)
Calgary, Edmonton, &
Montreal, Can
(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,
089824)
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 = 388,1 05)
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
January 2010
5-59
-------
5.4.1.3. Congenital Anomalies
Despite the growing evidence of an association between ambient air pollution and various
adverse birth outcomes, few studies have investigated the effect of temporal variations in ambient air
pollution on congenital anomalies. Heart defects have been the focus of the majority of these recent
air pollution studies, given the higher prevalence than other congenital anomalies and associated
mortality. Another study's focus was cleft lip/palate.
The earliest of these studies was conducted in southern California (Ritz et al, 2002, 023227).
Exposure to ambient CO, NO2, O3 and PMi0 during each of the first 3 mo of pregnancy was
examined among births during 1987-1993. Maternal exposure estimates were based on data from the
fixed site closest to the mother's ZIP code area. When using a case-control design where cases were
matched to 10 randomly selected controls, results showed that CO during the second month of
pregnancy was associated with cardiac ventricular septal defects. The CO exposures were grouped
by quartiles (25th = 1.14, 50th = 1.57, 75th = 2.39 ppm), and when compared to those in the lowest
quartile exposure group (<1.14 ppm), the odds ratios for ventricular septal defects across the 3 higher
exposure groups were 1.62 (95% CI: 1.05-2.48), 2.09 (95% CI: 1.19-3.67), and 2.95
(95% CI: 1.44-6.05), respectively. In a multipollutant model a similar exposure-response pattern was
exhibited across the quartiles with the highest quartile of exposure reaching statistical significance
(OR: 2.84 [95% CI: 1.15-6.99]). The only other pollutant associated with a defect was O3 during the
second month of pregnancy, which was associated with aortic artery and valve defects.
Another study was conducted in Texas (Gilboa et al., 2005, 087892). where exposure to
ambient CO, NO2, SO2, O3 and PMi0 during the 3rd-8th week of gestation was examined among
births between 1997 and 2000. Maternal exposure estimates were calculated by assigning the data
from the closest monitor to the mother's residential address. If data were missing on a particular day
then data from the next closest site were used. The median distances from a monitor ranged from 8.6
to 14.2 km, with maximum distances ranging from 35.5 to 54.5 km. The main results showed that
CO was associated with multiple conotruncal defects and Tetralogy of Fallot. CO exposures were
grouped into quartiles of much lower concentrations (25th = 0.4, 50th = 0.5, 75th = 0.7 ppm) than
the California study (Ritz et al., 2002, 023227). and when compared to the lowest quartile, the odds
ratios for conotruncal defects across the 3 CO exposure groups were 1.38 (95% CI: 0.97-1.97), 1.17
(95% CI: 0.81-1.70), and 1.46 (1.03-2.08), respectively, without a significant test for trend (p for
trend = 0.0870). A strong exposure-response pattern was exhibited across the quartiles of CO
exposure for Tetralogy of Fallot (25th OR: 0.82 [95% CI: 0.52-1.62]; 50th OR: 1.27
[95% CI: 0.75-2.14]; 75th OR: 2.04 [95% CI: 1.26-3.29]; p for trend = 0.0017). The only significant
associations found with other pollutants were between PMi0 and isolated atrial septal defects, and
SO2 and ventricular septal defects.
A study conducted in Atlanta, GA, investigated the associations between ambient air pollution
concentrations during weeks 3-7 of pregnancy and risks of cardiovascular malformations among a
cohort of pregnancies conceived during 1986-2003 (Strickland et al., 2009, 190324). The mean 24-h
CO concentration during this period was 0.75 ppm. The authors did not report any statistically
significant associations with ambient CO concentrations and cardiac malformations, though there
were elevated risk ratios for ambient CO concentration and patent ductus arteriosus, Tetralogy of
Fallot, and right ventricular outflow tract defect. These results remained consistently positive in five
sensitivity analyses conducted and were closer to achieving statistical significance in these
sensitivity analyses. The only statistically significant results were for the association between PMi0
and patent ductus arteriosus.
The last of these studies was a case-control study that examined maternal exposure to various
air pollutants during the first 3 mo of pregnancy and the risk of delivering an infant with an oral
cleft, namely cleft lip with or without palate (CL/P). Birth data from the Taiwanese birth registry
from 2001 to 2003 was linked to air pollutant data that were spatially interpolated from all fixed
monitoring sites across Taiwan. Based on data at the center of the townships or districts, exposure
estimates for PMi0, SO2, NOX, O3, and CO were averaged over each of the first 3 mo of pregnancy.
The mean 8-h avg CO concentration was 0.69 ppm. Interestingly, of all the pollutants examined,
only O3 during the first 2 mo of pregnancy was significantly associated with an increased risk of
CL/P In multipollutant models CO was not associated with CL/P (Hwang and Jaakkola, 2008,
193794).
The main results from the southern California study showed that CO was associated with an
increased risk of ventricular septal defects, and this was exhibited by an exposure-response pattern
across the quartiles of exposure; yet there was no indication that ambient CO concentration in Texas
January 2010 5-60
-------
was associated with ventricular septal defects. Conversely, ambient CO concentration in Texas was
associated with an increased risk of conotruncal defects; yet there was no indication that CO in
southern California was associated with conotruncal defects. The Atlanta study (Strickland et al.,
2009, 190324) found positive, though not statistically significant associations for patent ductus
arteriosus, Tetralogy of Fallot, and right ventricular outflow tract defect. The elevated risk ratio for
Tetralogy of Fallot is consistent with the result observed in Texas (Gilboa et al., 2005, 087892).
Interestingly, similar inconsistencies were also found for PM^ between these studies. For
example, PMi0 in Texas was associated with an increased risk of atrial septal defects and with patent
ductus arteriosus in Atlanta, GA; yet there was no indication of such an effect in southern California
where PMi0 concentrations were markedly higher.
The authors of the Texas study (Gilboa et al., 2005, 087892) provided little discussion toward
the inconsistent results with the southern California study. One suggestion was the different CO
concentrations across the studies with the 75th quartile in southern California being 2.39 ppm while
in Texas it was much lower at 0.7 ppm. However, this suggests that different defects are associated
with different concentrations of CO; yet it still does not explain why particular associations were
reported in Texas and not southern California where concentrations were higher. Similarly, the
authors of the Texas study (Gilboa et al., 2005, 087892) also suggested the inconsistency was due to
different exposure periods. In Texas the exposures were averaged over the 3rd-8th wk while in
southern California the exposures were averaged over the second month of pregnancy. However,
there was no reason provided as to why this small difference in the examined exposure period would
explain the inconsistent results.
Overall, there is some evidence that maternal exposure to CO is associated with an increased
risk of congenital anomalies, namely heart defects and cleft lip and palate. Further research is
required to corroborate these findings.
5.4.1.4. Neonatal and Postneonatal Mortality
A handful of studies examined the effect of ambient air pollution on neonatal and postneonatal
mortality, with the former the least studied. These studies varied somewhat with regard to the
outcomes and exposure periods examined and study designs employed.
Neonatal
In Sao Paulo, Brazil, a time-series study examined daily counts of neonatal (up to 28 days
after birth) deaths for the period 1998-2000 in association with concurrent-day exposure to SO2, CO,
O3, and PMio. Moving averages from 27 days were examined. The mean city-wide CO concentration
was 2.8 ppm, and there was no association between daily ambient CO and neonatal deaths. Despite
CO being correlated with PMi0 (r = 0.71) and SO2 (r = 0.55), only PMi0 and SO2 were associated
with an increase in the daily rate of neonatal deaths (Lin et al., 2004, 095787).
In another study of neonatal death, Hajat et al. (2007, 093276) created a daily time-series of air
pollution and all infant deaths between 1990 and 2000 in 10 major cities in England. The mean daily
CO concentration across the 10 cites was 0.57 ppm. This study provided no evidence for an
association between ambient CO concentration and neonatal deaths.
Postneonatal
Two studies in the U.S. examined the potential association between ambient CO and
postneonatal (from 28 days to 1 yr after birth) mortality and inconsistent results were reported. These
studies, however, varied somewhat in study design.
The first of these studies employed a case-control design and examined all infant deaths
during the first year of life among infants born alive during 1989-2000 within 16 km of a monitoring
site within the South Coast Air Basin of California. Exposures for 2-wk, 1-mo, 2-mo, and 6-mo
periods before death were linked to each individual death. Extensive analyses were conducted for
all-cause infant deaths, respiratory causes of death, and sudden infant death syndrome (SIDS). Given
the long time period of the data analyzed, in order to alleviate the confounding trends in infant
mortality and CO levels, this study was able to match by year (Ritz et al., 2006, 089819). Ambient
January 2010 5-61
-------
1-h max CO concentrations averaged over the 2 mo before death were associated with an 11% (OR:
1.11 [95% CI: 1.06-1.16]) increase in risk of all-cause post-neonatal death (per 1 ppm increase) and
a 19% (OR: 1.19 [95% CI: 1.10-1.28]) increase in risk of SIDS. In the multipollutant models
(including PMi0, NO2, O3) the positive CO mortality effect decreased by around 50% and was not
statistically significant. Based on exposure from 2 wk before death, CO was associated with an
increased risk of respiratory related postneonatal deaths occurring 28 days to 1 yr after birth (OR:
1.14 [95% CI: 1.03-1.25] per 1 ppm increase) and 28 days to 3 mo after birth (OR: 1.20
[95% CI: 1.02-1.40]); but no effect was observed for respiratory related deaths occurring 4-12 mo
after birth. These results persisted in the multipollutant models, and exposure-response patterns were
exhibited across the exposures groupings of 1.02 to <2.08, and > 2.08 ppm. To control for gestational
age and birth weight the analyses were stratified by "term/normal-weight infants" and "preterm
and/or LEW infants." When these two strata were analyzed, CO was associated with an increased
risk of all-cause death and SIDS within both strata (ORs ranged from 1.12 to 1.46). However, these
effects did not persist in multipollutant models (Ritz et al, 2006, 089819).
Another study examined 3,583,495 births, including 6,639 postneonatal deaths occurring in
96 counties throughout the U.S. (in counties with >250,000 residents) between 1989 and 2000
(Woodruff et al., 2008, 098386). Only exposure during the first 2 mo of life was examined, and this
was based on an average of CO concentrations recorded across all available monitors within the
mother's county of residence. In contrast to the other postnatal mortality study in California, CO
averaged over the first 2 mo of life was not associated with all-cause death (OR: 1.01
[95% CI: 0.94-1.09] per 0.5 ppm increase in 24-h CO concentration), or with respiratory related
deaths (OR: 1.08 [95% CI: 0.91-1.54] per 0.5 ppm increase in 24-h CO concentration), SIDS (OR
0.85 [95% CI: 0.70-1.04] per 0.5 ppm increase in 24-h CO concentration), or other causes of
postneonatal mortality (OR: 1.03 [95% CI: 0.96-1.09] per 0.5 ppm increase in 24-h CO
concentration). These null findings may be due to higher error of the exposure assessment at the
county level as opposed to using data from monitors within close proximity to the residence.
In a study that included 10 major cities in England, Hajat et al. (2007, 093276) created a daily
time-series of air pollution and all infant deaths between 1990 and 2000. While there was no
evidence for an association with neonatal deaths and ambient CO concentrations, there was a strong
adverse effect of CO in postneonatal deaths, although the confidence intervals were wide due to a
small sample size (RR 1.09, 95% CI: 0.94-1.25).
The only other postnatal mortality studies have been conducted throughout Asia. Two identical
studies in Taiwan failed to find an association between daily counts of postneonatal deaths and
ambient air pollutants, including CO. The data analyzed were from the cities of Taipei (Yang et al.,
2006, 090760) and Kaohsiung (Tsai et al., 2006, 090709). with ambient CO concentrations being
1.6 ppm and 0.8 ppm, respectively. Both studies examined deaths for the period 1994-2000 and
employed a case-crossover design that compared air pollution levels 1 wk before and after each
infant's death.
Similarly, another study in South Korea examined postneonatal mortality for the period
1995-1999, using a time-series design. Same-day CO was not associated with all-cause death (RR:
1.02 [95% CI: 0.97-1.06] per 0.5 ppm increase). However, same-day CO was associated with
postneonatal mortality when the analyses were restricted to respiratory mortality (RR: 1.33
[95% CI: 1.01-1.76] per 0.5 ppm increase) (Ha et al., 2003, 042552). An additional study examined
the relationship between air pollution and postneonatal mortality for all causes in Seoul, Korea. This
study used both case-crossover and time-series analyses for all firstborn infants during 1999-2003.
The mean 8-h max CO concentration during this time period was 1.01 ppm. The association between
ambient CO concentration and postneonatal mortality was the strongest in magnitude for CO when
compared to the other criteria pollutants, though the confidence intervals were wide (RR: 1.02
[95% CI: 0.87-1.20] for case-crossover analysis; RR: 1.23 [1.06-1.44] for time-series analysis per
0.75 ppm increase in 8-h max CO concentration).
In general, the inconsistent exposure periods examined among these studies restricts direct
comparison and interpretation. Nevertheless, there is limited evidence that CO is associated with an
increased risk of infant mortality during the postneonatal period. The exposure periods examined
varied from the same-day CO to lag periods up to a 6-mo period prior to birth, with one study
alternatively exploring exposures averaged over the first 2 mo of life. Furthermore, given that birth
weight and gestational age are strong predictors of infant mortality, in all of the reviewed studies
these factors have not been considered at either the design or analysis stage. Hence, the link between
fetal, neonatal, and postneonatal exposures, and the possible interaction that birth weight and
January 2010 5-62
-------
gestational age may have on the results yielded from these 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
There is some evidence that CO during early pregnancy (e.g., first month and first trimester) is
associated with an increased risk of PTB. Additionally, there is evidence of ambient CO during
pregnancy having a negative effect on fetal growth. In general, the reviewed studies (Figure 5-10
through Figure 5-12) reported small reductions in birth weight (-10-20 g). Although the majority of
studies reported significant effects during either the first or third trimester, other studies failed to find
a significant effect during these periods. Several studies examined various combinations of birth
weight, LEW, and SGA/IUGR, and inconsistent results are reported across these metrics. For
example, six studies reported an association between maternal exposure to CO and decreased birth
weight, yet the decrease in birth weight did not translate to an increased risk of LEW or SGA. It
should be noted that having a measurable, even if small, change in a population is different than
having an effect on a subset of susceptible births, which may increase the risk of IUGR/LBW/SGA.
It is difficult to conclude if CO is related to a small change in birth weight in all births across the
population or a marked effect in some subset of births.
Three studies examined the effects of CO on cardiac birth defects and found maternal
exposure to CO to be associated with an increased risk of cardiac birth defects. Human clinical
studies also demonstrated the heart as a target for CO effects (Section 5.2). In general, there is
limited evidence that CO is associated with an increased risk of infant mortality during the
postneonatal period.
5.4.2. lexicological Studies of Birth Outcomes and Developmental
Effects
The brief overview of the reproductive and development toxicology of CO that follows is not
limited to the past 10 yr as are other areas discussed in this document. This is because reproductive
and developmental toxicology endpoints have not been covered in previous CO AQCDs. Effects of
both exogenous CO exposure and endogenous production of CO are discussed since exposure to
exogenous CO could possibly alter pathways normally regulated by endogenous CO production.
This document details how in utero or perinatal CO exposure in pregnant dams or pups affects
outcomes in the offspring, including postnatal mortality, skeletal development, the ability of the
developing fetus to tolerate maternal dietary manipulation, behavioral outcomes, neurotransmitters,
brain development, the auditory system, myocardial development, and immune system development.
Similarly, endogenous CO is discussed in relation to pregnancy maintenance, vascular tone during
gestation, the placenta, the ovaries, the anterior pituitary axis, and lactation. Together, this
toxicological summary documents the importance of CO in reproductive and developmental
toxicology in laboratory animal models.
5.4.2.1. Birth Outcomes
Decreased Birth Weight
Multiple reports have been published associating CO exposure in laboratory animals and
decrements in birth weight (90-600 ppm); some of these studies also noted reduced growth evident
in the prenatal period (65-500 ppm CO). Significant decreases in fetal body weight at GD21 after
21 days of continuous CO exposure (125, 250, or 500 ppm) in pregnant Wistar rats have been
reported (Prigge and Hochrainer, 1977, 012326). This decrease was not found in rats exposed to
60 ppm CO. Penney et al. (1983, 011385) exposed pregnant rats to CO (200 ppm) for the final
17 days of prenatal development and also found significant decreases in near-term fetal rat weight at
GD20-GD21; gestation in rats is ~ 22 days. Penney et al.(1982, 011387) continued to find decreased
January 2010 5-63
-------
body weight to PND210 after postnatal CO exposure (500 ppm, PND1-PND32) and to a larger
extent in male pups when compared to female pups. Singh et al. (1984, 011409; 1993, 013892)
found significant decreases in fetal weight in gestationally CO-exposed mouse pups (65, 125, 250 or
500 ppm) in two studies. Near-term fetal body weight was decreased at GDI 8 in mice exposed from
GD7-GD18 to 125, 250, and 500 ppm CO but not 65 ppm CO (Singh and Scott, 1984, 011409).
However, a second study found decreased fetal weight at GDI8 with all CO exposures (65-500 ppm)
from GD8-GD18 (Singh et al., 1993, 013892).
A number of studies have found decreases in birth weight after CO exposure. A decrease in
body weight at birth was seen in neonates of pregnant rats exposed to 157, 166, and 200 ppm CO
over GD6-GD19 (Penney et al., 1983, 011385). Singh (2006, 190512) showed decreases in birth
weight of mouse pups gestationally exposed for 6 h/day for the first 2 wk of pregnancy to 125 ppm
but not 65 ppm CO. Carmines and Rajendran (2008, 188440) exposed Sprague Dawley rats to
-600 ppm CO (dam COHb 30%) via nose-only inhalation (levels similar to those seen in cigarette
smoke) during GD6-GD19 of gestation for 2 h/day and found significant decreases in birth weight
(0.5 g or 13%) of exposed pups versus controls. Maternal body weight was unchanged through
gestation, but corrected terminal body weight (body weight minus uterine weight) was significantly
elevated in CO-exposed dams at term, indicating a decrease in uterine weight. Other studies have not
found decreases in birth weight after gestational CO exposure (Carratu et al., 2000, 015839; Mereu
etal..200Q. 193838).
Other animal models have been used to examine decreased birth weight resulting from CO
exposure. Astrup et al. (1972, 011121) found significant decreases (11 and 20%, respectively) in
birth weight of rabbits exposed to either 90 or 180 ppm CO continuously over the duration of
gestation. Tolcos et al. (2000, 015997) found significant decreases in body, brain, and liver weights,
and crown-to-rump length in guinea pig fetuses after exposure to 200 ppm CO for 1 Oh/day from
GD23-GD25 until GD61-GD63, at which time the fetuses were collected (term ranges from GD68 to
GD72). In other studies, there was no significant differences in birth weight of guinea pig pups after
a similar exposure (200 ppm from GD23-GD25 to term, fetal and maternal COHb levels of 13% and
8.5%, respectively) (McGregor et al., 1998, 085342; Tolcos et al., 2000, 010468) or in Long Evans
rats (150 ppm CO continuous exposure over all of gestation) (Fechter and Annau, 1977, 010688).
Fetal mouse weight was significantly greater than control in the 7 h/day exposures and significantly
less than control animals in the 24 h/day (250 ppm CO, GD6-GD15) exposure groups, with
corresponding significant differences in crown-to-rump length in the two groups (Schwetz et al.,
1979, 011855). However, animals that showed no decrement in birth weight were significantly
smaller at PND4 compared to control guinea pigs (McGregor et al., 1998, 085342). with dam and
fetal COHb levels of 13% and 8.5%, respectively, during pregnancy.
Pregnancy Loss and Perinatal Death
Two studies have provided evidence for CO-induced pregnancy loss and perinatal death at CO
concentrations between 90 and 250 ppm. Schwetz et al. (1979, 011855) exposed CF-1 mice and New
Zealand rabbits to 250 ppm CO over GD6-GD15 (mice) or GD6-GD18 (rabbits) for either 7 h/day or
24 h/day, yielding 4 exposure paradigms. The fetuses were then collected at the termination of
exposure, near term. Maternal COHb in the 7 h/day exposure groups was approximately 10-15%
COHb in rabbits and mice; COHb was not followed in the 24-h exposure groups. The mice exposed
to CO for 7 h/day but not 24 h/day had a significant increase in the number of resorbed pups. Rabbits
were less affected by CO exposure, manifesting no significant perinatal death or pregnancy loss.
Astrup et al. (1972, 011121) studied the effect of CO on fetal development after continuous CO
exposure (90 or 180 ppm CO, COHb 8-9% and 16-18%, respectively) over the duration of gestation
in rabbits. In the immediate neonatal period, 24 h postpartum, 35% (180 ppm) and 9.9% (90 ppm) of
CO-exposed animals died. In the postpartum period after the first 24 h and extending out to PND21,
90 ppm CO-exposed pups experienced 25% mortality versus 13% in controls; there was no
difference from control at the 80 ppm CO exposure level. Gestation length was unchanged with CO
exposure. Conversely, Fechter and Annau (1977, 010688) exposed Long Evans rats in utero to
150 ppm CO continuously through gestation (dam COHb 15%) and saw no effects of CO on litter
mortality atPNDl.
January 2010 5-64
-------
Effect of Maternal Diet
As mentioned above, CO induced offspring mortality after prenatal exposure. Alterations in
maternal dietary protein and zinc further altered offspring mortality and teratogenicity caused by CO
(65-500 ppm).
Maternal Protein Intake and Neonatal Mouse Mortality and Teratogenicity
Pregnant CD-I mice were exposed intermittently (6 h/day for first 2 wk of pregnancy) to CO
(0, 65, or 125 ppm) in combination with protein modified diets (27% [supplemental protein], 16%
[control], 8% [low], or 4% [very low protein]) to assess the role of dietary protein in modulating CO
effects on neonatal mortality at 1 wk of age (Singh, 2006, 190512). Litter size was not affected by
CO exposure. Pup weight was inversely related to CO exposure and directly related to dam diet
protein content during pregnancy. Pup mortality at birth was directly related to CO exposure in
certain protein groups (supplemental, and 4% protein) and inversely related to the dam's dietary
protein content. At 1 wk of age, pup mortality was significantly increased by CO exposure as well as
dietary protein restriction; all pups in the 4% protein diet died by 1 wk of age. CO exposure (65 ppm
only) combined with a normal protein diet (16%) and CO exposure (65 and 125 ppm) with a
supplemental protein diet (27%) significantly increased pup mortality at 1 wk versus control air pups
(0 ppm CO). Contrary to other findings, low protein diet (8%) combined with CO (125 ppm) led to a
slight yet significant decrease in pup mortality at 1 wk of age versus control (0 ppm CO). In
summary, these data show that in utero CO exposure induced increased neonatal mouse deaths at
1 wk in supplemental protein and normal protein diet exposure groups and increased perinatal
mortality when combined with supplemental or restricted protein.
The role of diet as a contributor to teratogenicity of CO (0, 65, 125, or 250 ppm CO) in CD-I
mice given various protein diets (4, 8, 16, or 27% protein) during pregnancy was explored by Singh
et al. (1993, 013892). Timed-pregnant CD-I mice were exposed continuously to CO from
GD8-GD18, at which point animals were sacrificed and fetuses collected. Work by this group has
shown that low protein diets plus CO exposure act in an additive fashion to increase placental COHb
in mice (Singh, 2003, 053624: Singh et al., 1992, 013759). As expected, all levels of CO and the
lowest protein diet (4 or 8% protein) given to the dams during gestation resulted in significantly
decreased near-term weight of normal fetuses at GDI 8. CO exposure did not produce maternal
toxicity except for a significant decrease in maternal weight at GDI8 with 4 and 8% protein diets
versus control diet in non-CO-exposed animals. Dam dietary protein levels were inversely related to
gross fetal malformations including jaw changes. All concentrations of CO exposure within each
maternal dietary protein level significantly increased the percentage of litters with malformations in
a dose-dependent manner. Skeletal malformations were present in offspring, with the percent of
litters affected inversely related to dietary protein levels. CO exposure concomitant with a low
protein diet increased 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.
Maternal Zinc and Protein Intake and Neonatal Mortality and Teratogenicity
Singh (2003, 053624) explored how teratogenicity and fetal mortality were affected by zinc
(Zn) modulation in CO-exposed (500 ppm from GD8-GD18) pregnant dams (CD-I mouse) given
protein-insufficient diets. CO exposure in low-protein conditions (9% protein) decreased the mean
implants per litter as compared to air exposure. CO exposure also increased the near-term fetal
mortality over all groups, and to a larger extent in the low-protein groups, both Zn normal (57%
versus 6% mortality) and Zn deficient groups (86.6% versus 70.9% mortality). Under low-protein
conditions, CO exposure increased the incidence of malformations (9.4% versus 0%) when Zn levels
were normal and increased the incidence of gastroschisis (5% versus 0%) when Zn levels were low.
Joint protein and Zn deficiency led to 60% of litters with gastroschisis. Conversely, CO exposure
under Zn deficiency decreased the incidence of other malformations such as exencephaly, jaw,
syndactyly, and tail malformations.
Further studies by Neggers and Singh (2006, 193964) only partially confirmed these findings.
As before, diets deficient in both Zn and protein had significant detrimental influence on both fetal
malformations and mortality. Exposure to 500 ppm CO increased fetal mortality and malformation
January 2010 5-65
-------
rates under deficient protein (9%) and supplemental Zn (3.3 g/kg diet) conditions; however, CO had
a negligible effect on these endpoints under deficient protein and deficient or normal Zn conditions.
Role of Endogenous CO
CO is produced endogenously from heme protein catabolism by heme oxygenases, HO-1,
HO-2, and HO-3. CO has recently been recognized as a second messenger signaling molecule,
similar to NO, with a number of normal physiological roles in the body. Some of these roles are
played in maintaining pregnancy, controlling vascular tone, regulating hormone balance, and
sustaining normal ovarian follicular maturation. These areas could be potential areas of interaction of
exogenous CO.
Pregnancy Maintenance
HO-1 is known to protect organs from rejection (Kotsch et al., 2006, 193899) and thus, HO
may also protect the developing fetus from rejection by the non-self maternal immune system.
Idiopathic spontaneous abortions are more frequent in women with HO-1 polymorphisms (GT)n
microsatellite polymorphisms associated with altered HO-1 transcription in their genome (Denschlag
et al., 2004, 193894). Similarly, administering HO-inhibitors to pregnant rodents induced total litter
loss, possibly due to vasoconstriction and associated ischemia of the placental vascular bed
(Alexandreanu and Lawson, 2002, 192373). Also, mice overexpressing HO-1 had a significantly
decreased rate of spontaneous abortion (Zenclussen et al., 2006, 193873). Various pathologies of
pregnancy, including IUGR and pre-eclampsia, are associated with significant decreases in placental
HO activity (Denschlag et al., 2004, 193894: McLaughlin et al., 2003, 193827). Oxygenation is
important in early pregnancy and triggers trophoblast invasion of the spiral arteries (Kingdom and
Kaufmann, 1997, 193897). Women living at high altitude have an increased risk of adverse
pregnancy outcomes versus women living at lower altitudes (Zamudio et al., 1995, 193908). Also,
women living at high altitude, women with pre-eclampsia, or women who had pregnancies with fetal
growth restrictions (FOR) produced term placenta with significant decreases in HO-2 versus women
living at lower altitude with uncomplicated pregnancies (Barber et al., 2001, 193891; Lyall et al.,
2000, 193902). Thus, the HO/CO system is crucial for the developing fetus, helps in maintaining
pregnancy, and plays a role in spontaneous abortions.
Vascular Control
During pregnancy, there is increased blood volume without a concurrent increase in systemic
BP, which is accomplished by a decrease in total peripheral vascular resistance (Zhao et al., 2008,
193883). CO through the production of soluble guanylate cyclase is able to stimulate the relaxation
of vascular smooth muscle (Villamor et al., 2000, 015838) and relaxation of pregnant rat tail artery
and aortic rings (Longo et al., 1999, 011548). Further, the administration of the HO inhibitor SnMP
increased maternal BP (systolic, diastolic, and mean arterial pressure) and significantly increased
uterine artery blood flow velocity during pregnancy in mice (Zhao et al., 2008, 193883). Zhao et al.
also showed pregnancy induced increased total body CO exhalation and that this increased CO
production could be significantly decreased by SnMP administration. Abdominal aortas (AA) of
pregnant dams are significantly dilated with pregnancy, and SnMP treatment leads to AA
vasoconstriction to levels similar to nonpregnant mice. Isolated human placenta exposed to solutions
containing CO demonstrated a concentration-dependent decrease in perfusion pressure (Bainbridge
et al., 2002, 043161). further demonstrating the role of CO in maintaining basal vasculature tone.
However, the addition of exogenous CO to isolated human and rat uterine tissue during pregnancy
failed to induce relaxation and quiet the spontaneous contractility of rat or human myometrium
(uterine smooth muscle)(Longo et al., 1999, 011548). CO is not able to relax all types of vascular
smooth muscle (Brian et al., 1994, 076283). and pregnancy appears to modulate the response of
tissues to CO (Katoue et al., 2005, 193896). Thus, it appears that the increased CO production during
pregnancy may partially account for the decreased peripheral vascular resistance seen in pregnancy
that prevents the increased blood volume of pregnancy from affecting BP.
January 2010 5-66
-------
Hormone Regulation
Endogenous CO has been shown to regulate neuroendocrine functions. Disruption of normal
CO signaling causes changes in the cycles of a number of hormones involved in pregnancy. HO
inhibition in rats significantly decreased ovarian production of gonadotrophin-induced
androstenedione and progesterone without affecting estradiol levels (Alexandreanu and Lawson,
2002, 192373). However, treatment with the HO-inducer, hemin, caused androstenedione and
estradiol production from rat ovaries in vitro. CO also has been shown to have a stimulatory effect
on gonadotropin-releasing hormone (GnRH) release from rat hypothalamic explants in vitro (Lamar
et al., 1996, 078819), while in vivo CO appears not to influence GnRH secretion (Kohsaka et al.,
1999, 191000). HO-1 induction and HO concentration have been shown to be regulated by estrogen
in the rat uterus (Cella et al., 2006, 193240) during pregnancy and in nongravid rats. This agrees
with work by Tschugguel et al. (2001, 193785) in which CO was generated by primary endothelial
cells from human umbilical veins and uterine arteries after exogenous 17-(3 estradiol administration.
HO inhibition by CrMP decreased time in estrous in a dose-dependent manner (Alexandreanu and
Lawson, 2002, 192373).
HO-1 and HO-2 are expressed in rat anterior pituitary, and the secretion of gonadotropins and
prolactin is affected by HO-inhibitor and HO-substrate administration (Alexandreanu and Lawson,
2003, 193871). The estrogen-induced afternoon surge of luteinizing hormone (LH) was advanced
forward in time by HO inhibition, and this advance could be reversed by concomitant administration
of hemin. The serum follicle stimulating hormone (FSH) surge was unaffected by HO inhibition or
hemin, but in vitro treatment of GnRH-stimulated pituitaries with hemin led to a significant increase
in FSH release. The estrogen-dependent afternoon prolactin surge was inhibited or delayed by HO
inhibition and significantly decreased prolactin release. In vitro studies using pituitary explants
showed that LH release was significantly increased by HO inhibition. HO inhibition also decreased
litter weight gain during lactation, which the authors attributed to decreased maternal milk
production or milk ejection problems as cross-fostered pups regained weight that was lost during
nursing on HO-inhibited dams (Alexandreanu and Lawson, 2002, 192373). The lactational effects
seen in this model may be explained by changes in prolactin (Alexandreanu and Lawson, 2003,
193871). It is possible that HO inhibition by CrMP may also inhibit NO production, a mechanism
that is distinct from CO-dependent effects.
Ovarian Follicular Atresia
As a part of normal follicular maturation in the ovaries, the majority of follicles undergo
atresia via apoptosis prior to ovulation. Harada et al. (2004, 193920) harvested porcine granulosa
cells from ovaries and found that cells naturally undergoing atresia or cell death more strongly
expressed HO-1 than did successful follicles. Addition of the HO-substrate hemin or the HO-
inhibitor Zn protoporphyrin IX (ZnPP IX) significantly induced or inhibited granulosa cell apoptosis,
respectively. In this porcine model, HO was able to augment granulosa cell apoptosis allowing for
proper follicular maturation.
Summary of lexicological Studies on Birth Outcomes
There is some evidence that CO exposure leads to altered birth outcomes, including decreased
birth and near-term body weight, increased pregnancy loss and perinatal death, and increased
malformations. These events occurred at levels as low as 65 ppm for fetal body weight decrements
and 90 ppm for changes in birth weight and perinatal death. Pregnancy loss was seen after exposure
to 250 ppm CO, whereas skeletal malformations were present after 180 ppm CO. Dietary protein and
zinc modifications exacerbated these CO-induced effects on birth outcomes. Maternal protein
restriction and CO had a synergistic effect on peri- and postnatal mortality and an additive effect on
malformations. Dietary zinc alterations resulted in inconsistent changes to CO-induced
malformations and fetal mortality.
Endogenous CO is recognized as a second messenger signaling molecule with normal
physiological roles in maintaining pregnancy and for proper fetal and postnatal development. The
endogenous HO/CO system is also involved in controlling vascular tone, follicular maturation,
ovarian steroidogenesis, secretion of gonadotropin and prolactin by the anterior pituitary, lactation,
January 2010 5-67
-------
and estrous cyclicity in rodent studies. These areas could be potential points of interaction of
exogenous CO with endogenous HO/CO.
5.4.2.2. Developmental Effects
Congenital Abnormalities
Studies by Schwetz et al. (1979, 011855) found that gestational CO exposure (250 ppm) in
CF-1 mice for 7 or 24 h/day over GD6-GD15 resulted in minor fetal skeletal alterations in the form
of extra lumbar ribs and spurs (dam gestational COHb 10-15% for 7h/day exposure, 24 h/day dam
COHb not measured). Similarly exposed rabbits did not exhibit these changes.
Astrup et al. (1972, 011121) studied the effect of CO exposure on fetal rabbit development via
continuous CO exposure (90 or 180 ppm with gestational dam COHb of 9 and 17%, respectively)
over the duration of gestation. Three pups in the 180 ppm CO group (n = 123) had deformities in
their extremities at birth, whereas no control and no 90 ppm CO-exposed animals manifested with
this malformation.
Further skeletal malformations were seen after gestational CO exposure in mice as described
above ("Effect of Maternal Diet") (Singh et al., 1993, 013892). Briefly, pregnant CD-I mice were
exposed intermittently to CO (65-250 ppm; GD8-GD18) in combination with protein modified diets
(27% [supplemental protein], 16% [control], 8% [low], or 4% [very low protein]) to assess the role
of dietary protein in modulating CO effects on neonates at 1 wk of age. Maternal dietary protein
restriction additively compounded the CO-induced skeletal malformations. Further, dietary
restriction in Zn and protein led to increased teratogenicity, specifically increased incidence of
gastroschisis (Singh, 2003, 053624). Conversely, Carmines and Rajendran (2008, 188440) did not
find evidence of external malformations (teratogenicity) in rats after exposure to -600 ppm CO from
GD6-GD19.
CMS Developmental Effects
Behavioral
Investigators have used animal models to study the effects of moderate CO exposure
(65-150 ppm) during gestation on behavioral outcomes after birth, including active avoidance,
learning and memory, homing, and motor activity. These studies generally found decrements in
behavior in early life after in utero exposure to CO concentrations >125 ppm and in some cases as
low as 65 ppm. Table 5-14 shows results of behavioral response studies with CO exposure
< 150 ppm.
January 2010 5-68
-------
Table 5-14. Behavioral responses.
Study
Model System CO Exposure
Response
Notes
BEHAVIORAL RESPONSES
DeSalviaetal. (1995,
0794411
Mactutus and Fechter
(1985, 0115361
Di Giovanni et al.
(1993, 0138221
Rats
Rats
Rats
75 and 150ppm
continuous
GDO-GD20
150ppm
continuous
GDO-GD20
75 and 150ppm
continuous
GDO-GD20
Impaired acquisition (3 and 18 mo) and reacquisition (18 mo) of
avoidance behavior at 150 ppm, not 75 ppm
Delayed acquisition of active avoidance (PND120) and disrupted ^nuk ic a j. i -io/
retention (PND360) COHb 15.6 ± 1.1%
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.
Mactutus and Fechter
(1984,0113551 Rats
-i^n nnm Acquisition did not improve with age/maturation, failure to learn; rnHh 1IW
I3uppm impaired reacquisition (PND31), failure to retain UU-ID io/o
Giustino et al. (1999,
0115381
Zhuoetal. (1993,
0139051
Stevens and V\feng
(1993, 1884581
Rats
Mouse
hippocampal brain
sections
Mouse and rat
hippocampal brain
slices
75 and 150 ppm
continuous
GDO-GD20
ZnPPIX (HO
inhibitor) and
0.1-1.0 pM CO
ZnPPIX (5-15 pM)
Decreased exploration, habituation, nonspatial working memory
HO inhibition blocked long-term potentiation and CO evoked
synaptic potentials and long-term enhancement
HO inhibition blocked long-term potentiation but not long-term
depression.
COHb: 1.6 ±0.1%
(0 ppm); 7.36 ±0.2%
(75 ppm); 16.1 ±0.9%
(150 ppm)
Rat hippocampal 150 ppm
brain sections GDO-GD20
Impaired long-term potentiation maintenance
Fechter and Annau
(1980, 0112951
Fechter and Annau
(1977, 0106881
Rats
Rats
150 ppm
continuous
GDO-GD20
150 ppm
continuous
GDO-GD20
Delayed homing behavior and poor reflexive response
Decreased locomotor activity at PND1, PND4, and PND14, but not
PND21
COHb 15%
Singh (1986, 012827) Mice
65 and 125 ppm
continuous
GD7-GD18
Impaired aerial righting score at PND14 (65 and 125 ppm), impaired
negative geotaxis at PND10 and righting reflex on PND1 (125 ppm)
Active Avoidance Behavior. 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
these behavioral changes were permanent. De Salvia et al. (1995, 079441) found there were no
significant behavioral impairments following exposure to 75 ppm CO. However, animals exposed to
the 150 ppm in utero had significantly impaired acquisition (at 3 and 18 mo of age) and reacquisition
(at 18 mo of age) of conditioned avoidance behavior. This impaired learning was also seen in
gestationally CO (150 ppm, trend seen at 75 ppm) exposed rats at PND90 (Di Giovanni et al., 1993,
013822). The authors speculated that this CO-dependent behavioral change may be mediated
through neurotransmitter signaling, specifically changes in dopamine in the neostriatum or nucleus
accumbens. These studies demonstrate that moderate CO exposure in utero can lead to permanent
behavioral changes in male offspring.
Mactutus and Fechter (1984, 011355) also found that acquisition in a two-way conditioned
avoidance test (flashing light warnings followed by mild footshock) failed to improve with age of in
January 2010
5-69
-------
utero CO-exposed (150 ppm, dam COHb 15%) Long Evans 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. They also found impairments in reacquisition performance, an index
of retention, in PND31 rats that had received continuous in utero CO exposure. Overall, prenatal CO
exposure (150 ppm, not 75 ppm) induced learning and memory deficits in male and female
offspring.
Habituation, Memory, and Learning. Giustino et al. (1999, 011538) exposed primiparious
pregnant Wistar rats to CO (0, 75 or 150 ppm) by inhalation from GDO-GD20. Blood COHb
concentrations (mean % ± SEM) on GD20 were reported (0 ppm: 1.6 ± 0.1; CO 75 ppm: 7.36 ± 0.2;
CO 150 ppm: 16.1 ± 0.9). Male offspring at age 40 days were given two habituation trials. In the
first trial (Tl), two similar objects were presented. In the second trial (T2), one object from the first
trial was presented as well as one novel object. Results were quantified three ways. Exploration
activity was defined as the time exploring both objects during each trial. Global habituation was
quantified as a comparison of the time spent exploring the two objects in Tl to the time spent
exploring objects in T2. Discrimination between new and familiar objects was measured in T2 by
contrasting the time spent exploring the familiar object to the time spent exploring the new object.
These recognition sessions test for the preference that rats have for investigating novel objects in lieu
of familiar objects and are a measurement of nonspatial working memory. The results of this study
showed 40 day old animals that were gestationally exposed to CO (both 75 and 150 ppm) spent less
time exploring novel objects when compared to control animals. Control rabbits habituated or
learned after a second exposure to a previously explored object (T2
-------
Under analogous exposure conditions, Fechter and Annau (1980, 011295) found that the
development of homing behavior, orientation by the rat toward its home cage, was significantly
delayed in rats prenatally exposed to 150 ppm CO. Also, exposed offspring manifested with poorer
than normal performance on the negative geotaxis test, a reflexive response that results in a
directional movement with or against gravity. Similarly, continuous prenatal CO exposure (125 ppm,
GD7-GD18) in CD-I mice impaired negative geotaxis at PND10 (Singh, 1986, 012827). The
standardization and use of geotaxis as a vestibular, motor, or postural metric in infant rodents has
been debated in the literature (Kreider and Blumberg, 2005, 193944).
Prenatal exposure to CO (125 ppm, GD7-GD18) significantly affected the righting reflex (the
turning of an animal from its supine position to its feet) in exposed CD-I mice on PND1. Also, the
aerial righting score, or turning 180° and landing on the feet when dropped from the supine position
at a height, was significantly decreased in pups exposed to CO in utero (65 and 125 ppm) at PND14
(Singh, 1986, 012827). The same trend of impaired righting reflex was seen in gestationally CO
(150 ppm) exposed rats (Fechter and Annau, 1980, 011295). These behavioral tests indicated
neuromuscular, vestibular, or postural effects in the CO-exposed neonate.
Conversely, no gross impairment of motor activity was found as measured by infrared
movement monitoring in Wistar rats treated in utero (GDO-GD20) with 0, 75 or 100 ppm CO
(Carratu et al., 2000, 015839). Monitoring was done at PND40 and PND90 and may have been too
late to detect CO-dependent changes. Earlier studies by Fechter and Annau (1977, 010688) identified
an early window of sensitivity for CO-dependent motor activity deficits of PND1-PND14, with
recovery by PND21.
Emotionality. In utero CO exposure caused subtle alterations in the ontogeny of emotionality
measured by the ultrasonic vocalization emitted by rat pups removed from their nest. Prenatal CO
exposure (150 ppm) caused a reduction in the minimum frequency of ultrasonic calls emitted by
PND5 pups (Di Giovanni et al., 1993, 013822). The rate of calling, maximum frequency, and
duration and sound pressure level were not affected by prenatal CO. However, the rate of calling and
responsiveness to a challenge dose of diazepam was decreased by prenatal CO exposure. Pup
vocalization is mediated by the GABAergic neuron function which is altered by CO exposure (see
below).
Neuronal
Since behavioral changes have been caused by CO exposure, studies have investigated
whether CO exposure results in changes to neuronal structures and electrical excitability. Moderate
levels of CO (75 -150 ppm) decrease peripheral nervous system (PNS) myelination due to impaired
sphingomyelin homeostasis and can reversibly delay the rate of ion channel development after
gestational exposure. In utero CO exposure also results in irreversible changes in sodium equilibrium
potential. Further details of these studies are given below in Table 5-15.
January 2010 5-71
-------
Table 5-15. Neuronal responses.
Study
Model
System
CO Exposure
Response
Notes
NEURONAL RESPONSES
Carratu et al. (2000,
0158391
Carratu et al. (2000,
0159351
Carratu etal. (1993,
0138121
DeLuca etal. (1996,
0809111
Montagnani et al. (1996,
0809021
Dyer etal. (1979,
1909941
Rats
Rats
Rats
Rats
Rats
Rats
75 and 150ppm
continuous
GDO-GD20
150ppm
continuous
GDO-GD20
75 and 150ppm
continuous
GDO-GD20
75 and 150ppm
continuous
GDO-GD20
75 or 150 ppm
GDO-GD20
150 ppm
GDO-GD21
Decreased peripheral nerve fiber myelin sheath thickness
Impaired sphingomyelin homeostasis by increasing sphingosine
Produced partly reversible changes in membrane excitability through
delayed inward current inactivation and decreased inward current
reversal potential
Delayed development of the ion channels responsible for passive and
active membrane electrical properties of skeletal muscle
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).
Increased early components (P1-N1 and N1-P1) of the cortical flash
evoked potential peak-to-peak amplitudes at PND65 in female rats
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)
COHb:15%at150ppm
Maternal COHb: 15%
Peripheral Nerve Myelination. The effect of in utero exposure (GDO-GD20) to 0, 75 or 150 ppm
CO on sciatic nerve myelination in male offspring was studied in Wistar rats (Carratu et al., 2000,
015839). The dam CO blood concentration, expressed as % COHb, was determined for 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). The myelin sheath thickness of the peripheral
nerve fibers was significantly decreased in CO-exposed animals (75 and 150 ppm); however, axon
diameter was not affected. As mentioned above, even though CO affected myelination, it did not
significantly affect motor activity of CO-exposed mice at 40 and 90 days. It is possible that these
deficits in PNS myelination are due to impaired sphingomyelin homeostasis. In utero exposure
(GDO-GD20) of Wistar rats to CO (150 ppm) caused a twofold increase in sphingosine (SO) but not
sphinganine (SA) in the sciatic nerve at 90 days of age (Carratu et al., 2000, 015935). SO is an
intermediate in sphingolipid turnover and SA is an intermediate of de novo sphingolipid
biosynthesis. Hypoxiahas been shown to induce sphingomyelin changes which could lead to
impaired myelination and motor activity decrements (Ueda et al., 1998, 195136; Yoshimura et al.,
1999, 195135). Prenatal CO exposure had no effect on brain SA or SO levels in male offspring at
90 days of age. These results demonstrate prenatal CO exposure could interrupt sphingolipid
homeostasis in the PNS but not CNS, causing a decrease in nerve myelination without changes in
motor activity.
Electrophysiological Changes.
Gestational exposure of Wistar rats to continuous CO (75 or 150 ppm (15% COHb at
150 ppm) yielded electrophysiological changes in the PNS (Carratu et al., 1993, 013812). Changes
were noticeable in voltage- and time-dependent properties of sodium channels in the sciatic nerve
after in utero CO exposure. Changes in sodium channel inactivation kinetics were reversible (present
at PND40 and absent at PND270) but changes in the sodium equilibrium potential were irreversible.
In utero CO exposure (150 ppm) also delayed the development of the resting chloride conductance
(GC1) and resting potassium conductance (GK), with levels matching the control by PND80 and
PND60, respectively (De Luca et al., 1996, 080911). CO exposure (75 and 150 ppm) also altered the
pharmacological properties of the chloride channel and excitability parameters of skeletal muscle
fibers. These changes in the nerve electrophysiological properties could account for increased
January 2010
5-72
-------
tetrodotoxin-inhibition of the vasoconstriction evoked by the PNS in 5- to 7-day-old prenatally
exposed pups (Montagnani et al., 1996, 080902). Finally, gestational CO exposure increased early
components (P1-N1 and N1-P1) of the cortical flash evoked potential peak-to-peak amplitudes at
65 days postexposure (PND65) in female, not male, rats (Dyer et al., 1979, 190994). The early
waves of the cortical evoked potential, an indicator of visual cortical functioning, generally indicate
activity in the retinogeniculostriate system. These studies showed that in utero CO exposure had both
reversible and irreversible effects on sodium and potassium channels, which are essential for proper
electrophysiological function of the muscles and PNS.
Neurotransmitter Changes
The developing nervous system is extremely sensitive to decreased oxygen availability.
Virtually all neurotransmitter systems are present at birth but require further maturation. The studies
listed below in Table 5-16 have shown that prenatal exposure to CO alters a number of
neurotransmitters and their pathways at concentrations ranging from 75 to 300 ppm, both transiently
and permanently.
Medullar Neurotransmitters. SIDS is a complex syndrome that involves the aberrant
development of brain stem nuclei controlling respiratory, cardiovascular, and arousal activity. To
investigate changes in the structure and neurochemistry of the brain stem, Tolcos et al. (2000,
015997) exposed pregnant guinea pigs to CO (200 ppm) over the last 60% of gestation. Guinea pigs
and humans both have the majority of CNS development in utero. CO-exposed pups were found to
have significant decrements in body, brain, and liver weights, crown-to-rump length, and medullar
volume when compared to control pups. Neurotransmitter systems were also affected after CO
exposure. Specifically, the brain stem displayed significant decreases in protein and
immunoreactivity for tyrosine hydroxylase (TH), an enzyme necessary for catecholamine
production, which is likely due to decreased cell number in specific medullar regions responsible for
cardiorespiratory control. This was consistent with earlier work showing that prenatal CO exposure
leads to aberrant respiratory responses to asphyxia and CO2 (McGregor et al., 1998, 085342). The
cholinergic system was also affected by prenatal CO exposure with significant increases in choline
acetyl-transferase (ChAT) immunoreactivity of the medulla; however, no changes in muscarinic
acetylcholine receptor were seen. This is in contrast to human infants with SIDS who show
decreased brain stem muscarinic receptor binding (Kinney et al., 1995, 193898). ChAT changes in
this study (Tolcos et al., 2000, 015997) were from areas of the medulla associated with tongue
innervation, which is crucial to swallowing, possibly in relation to breathing.
A second risk factor for SIDS is hyperthermia. To explore the interaction of hyperthermia and
CO-induced hypoxia, pregnant guinea pigs were exposed to CO (0 or 200 ppm) for 10 h/day for the
last 60% of gestation (Tolcos et al., 2000, 010468). At PND4 male pups were exposed to
hyperthermia or ambient temperature as a control. Brains were then collected at 1 and 8 wk of age.
In utero CO exposure sensitized some areas of the brain to future hyperthermic insults. Specifically,
CO plus hyperthermia induced significant increases in serotonin in multiple brain regions (NTS,
DMV, and hypoglossal nucleus) at 1 wk of age; this change was no longer evident at 8 wk of age.
Hyperthermia exposure alone induced decreased met-enkephalin neurotransmitter immunoreactivity
at 1 wk of age that was absent at 8 wk and absent in CO-plus-hyperthermia exposed animals. Brain
stem neurotransmitter (met-enkephalin, serotonin, TH, substance P) immunohistochemical
differences were not apparent with CO treatment alone. At 8 wk of age, CO-plus-hyperthermia
exposure induced glial aggregations and gliosis surrounding infarct or necrotic areas in the brain and
the medulla lesions stained positive for glial fibrillary acidic protein (GFAP). GFAP upregulation is
classically seen with neuronal diseases or following neurodegeneration. Gross structural
observations revealed no differences in the medulla or cerebellum following in utero CO exposure
alone. Together, these data showed that CO exposure in utero sensitizes the brain to future
hyperthermic insults, lea ding to generation of necrotic lesions in the brain and changes in
neurotransmitter levels.
January 2010 5-73
-------
Table 5-16. Neurotransmitter changes.
Study
Model
System
CO Exposure
Response
Notes
NEUROTRANSMITTER CHANGES
Tolcos et al. (2000, rllinMninc
0159971 Guinea pigs
Tolcos et al. (2000, r •
0104681 Guinea pigs
||fretaL<1998< Guinea pigs
Cagianoetal. (1998, p.
0871701 Kats
Hermans etal. (1993, p.
1905101 Kats
Fechter etal. (1987, p.
0122591 Kats
Storm and Fechter p .
(1985, 011653) Kats
Storm and Fechter p t
(1985, 0116521 Kats
Storm etal. (1986, pt
0121361 Kals
Benagiano et al. (2005, D t
1804451 Rats
Benagiano (2007, p t
1938921 Kats
Antonelli (2006, 1949601 Rats
200 ppm
1 0h/day
GD23-GD25to
GD61-GD63
200 ppm
1 0h/day
GD23-GD25to
birth
Hyperthermia on
PND4
200 ppm
1 0h/day
GD23-GD25to
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 brain stem 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:
GD1 0:1, 7, and 15%;
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)
Dopaminergic Effects. Dopamine is a catecholamine neurotransmitter that plays an important
role in the regulation of male rat sexual behavior. Experiments assessing sexual behavior and
mesolimbic dopaminergic function were conducted on adult (5 and 10 mo of age) male offspring
gestationally exposed to CO (0, 75 or 150 ppm) (Cagiano et al., 1998, 087170). Maternal COHb at
GD10 was 1, 7, and 15% and 1.5, 7, and 16% at GD20 (0, 75, and 150 ppm CO, respectively). 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
decrease in ejaculation frequency. Further, administration of amphetamine, which stimulates
copulatory activity, did not alter CO-induced changes in mount-to-intromission latency or frequency.
Basal extracellular dopamine concentration in the nucleus accumbens was unchanged after CO
exposure. However, when stimulated with amphetamine administration, control rats had increased
release of dopamine that was absent with CO-exposed rats. Rats followed to 10 mo of age showed
no significant changes in copulatory activity or neurochemical parameters after CO exposure,
January 2010
5-74
-------
indicating recovery from earlier decrements. This altered male sexual behavior in CO-exposed
offspring paralleled earlier studies of mice exposed gestationally to hypoxia (Hermans et al., 1993,
190510). In summary, in utero exposure to CO delayed copulatory sexual behavior in male offspring
with accompanying changes in the mesolimbic dopaminergic system.
A second study also found no change in dopamine metabolite levels after prenatal exposure to
CO; however, it did find an elevation in dopamine concentration in rats exposed both pre- and
postnatally to CO. Exposure of Long Evans rat dams and pups continuously to CO (75, 150, or
300 ppm) with maternal COHb of 11, 19, and 27%, respectively) from conception to PND10 induced
significant elevations in dopamine in the striatum at PND21 in CO-exposed offspring versus air
exposed controls (Fechter et al., 1987, 012259).
Noradrenergic and Serotonergic Changes. Other monoamine neurotransmitters, norepinephrine
(NE) and serotonin (5HT), were tested for sensitivity to CO during development. Long Evans rats
exposed to CO (75, 150, or 300 ppm) over the duration of gestation yielded a dose-dependent
reduction in cerebellum wet weight (significant at 150 and 300 ppm) at PND21, with increases in NE
concentration found in the cortex and hippocampus at PND42 but not PND21 (Storm and Fechter,
1985, 011652). In a separate experiment, CO-exposed (150 ppm) animals presented with increased
mean and total NE concentrations in the cerebellum but not cortex when monitored from PND14 to
PND42 (Storm and Fechter, 1985, 011653). Also, NE concentration in the pons/medulla decreased
linearly with increasing CO exposure at PND21 but not at PND42. A transitory decrease in 5HT
concentration was also shown in the pons/medulla after gestational CO exposure (Storm and Fechter,
1985, 011652). Thus, in these studies, it appeared that CO both transiently and permanently altered
the pattern of postnatal neurotransmitter development in a region-specific manner and stunted
postnatal growth of the cerebellum.
Glutamatergic System. Glutamate is an abundant excitatory neurotransmitter that serves as a
precursor for the synthesis of the inhibitory neurotransmitter y-aminobutyric acid (GAB A) catalyzed
by glutamic acid decarboxylase (GAD). Primary cell cultures obtained from the cerebral cortex of
offspring (PND1) gestationally (GD5-GD20) exposed to CO (75 ppm) had decreased extracellular
glutamate (basal and K+-evoked) levels versus air-exposed controls (Antonelli et al., 2006, 194960).
Similarly, CO-exposed (300 ppm only) pups at PND21 had significant decreases in cerebellar GABA
content, decreased uptake of exogenous radio-labeled GABA, decreased fissures in the cerebellum,
and decreased cerebellum size (Storm et al., 1986, 012136). It is possible this decrease in GABA
content is due to a diminished activity of GAD. Rats exposed to CO (75 ppm) in utero (GDO-GD20)
exhibited decreased GABA and GAD in the molecular layer and Purkinje neuron layer of the
vermian cerebellar cortex (Benagiano et al., 2005, 180445; Benagiano et al., 2007, 193892). This
alteration may functionally impair cortical glutamatergic transmission in CO-exposed offspring,
possibly affecting learning and memory.
The Developing Auditory System
The developing auditory system of rodents has recently been investigated as a target of CO
exposure at levels as low as 12 ppm. The rat brain and auditory system go through extensive cell
division and multicellular organization during a major growth spurt in the postnatal period (PND7-
PND20), making it a probable target for CO-induced effects. These studies showed that exposure to
low concentrations of CO during development can lead to permanent changes in the auditory system
that persist into adulthood. Similarly, prenatal exposure to tobacco smoke can cause auditory system
deficits as seen in animal tests for auditory responsiveness, habituation, and auditory arousal. Term
human infants born to smoking mothers have impaired cochlear development, albeit mild, with
decreased amplitudes of transient evoked otoacoustic emissions (OAE) at the highest test frequency
(4 kHz) versus newborns born to nonsmokers (Korres et al., 2007, 190908); CO is one of many
potential affective components of cigarette smoke.
January 2010 5-75
-------
Table 5-17. Developing auditory system.
Study
Exposure
Response
Notes
DEVELOPING AUDITORY SYSTEM
Stockard-Sullivan et al. D .
(2003, 1909471 Kats
Lopez etal. (2003, R.
1939011 Kats
Webber etal. (2003, p.
1905151 Kats
Webber etal. (2005, Dt
1905141 Kats
12-100ppm
22 h/day
PND6toPND21-PND23
12and25ppm
PND8-PND22
12.5, 25, SOppm
PND8 to PND20-PND22
25and100ppm
PND9-PND24
CO (50 ppm) reduced otoacoustic emissions (preneural cochlear
function) at 7.13 and 8.01 kHz. CO persistently attenuated the COHb: 10.2%
amplitude of the action potential of the eighth cranial nerve (12- (100 ppm); 5.5% (AR);
50 ppm), persisting to PND73. No functional impairment in the Morris 4.1% (MR)
Water Maze after CO 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.
Lopez etal. (2008, R.
0973431 Kats
25 ppm
Prenatal CO exposure led to increased oxidative stress in the cochlear
10 18 h/dav vasculature (high HO-1, SOD-1, iNOS, and nitrotyrosine) and
GD5 20 or GD5 GD20 and decreased neurofilament proteins and synapsin-1. CO caused
PND5 PND20 morphological deterioration of putative afferent terminals and mild
deterioration in the inner hair cells at the basal region of the cochlea.
Studies on the developing auditory system have used an artificial feeding system where pups
were removed from their respective dams and fed a milk substitute comparable to natural rat milk
via intragastric cannulation. This allowed nursing pups to be exposed to CO without possible
confounding by lactational and maternal CO co-exposure. However, this invasive rat model does
cause decreased brain, cerebellum, and lung weight at PND16 in normal air controls. A summary of
these studies and others are presented in the above table (Table 5-17).
Using this model, Stockard-Sullivan et al. (2003, 190947) examined Sprague Dawley rat pups
receiving low-dose CO (12, 25, or 50 ppm) to determine how perinatal CO exposure (PND6 to
PND21-PND23) functionally affected hearing in the developing rat. Rodent pups were either
maternally reared (MR), nutritionally supported with the artificial feeding system (AR), or received
AR plus CO exposure (ARCO). CO (50 ppm, not 25 ppm) exposure caused significant reductions in
distortion product otoacoustic emissions (DPOAE) levels at certain frequencies (7.13 and 8.01 kHz),
a measure of preneural cochlear function and thus not affected by eighth cranial nerve function.
However, the frequency range where significant CO results were seen is very narrow and low
compared to the normal rat audiogram. The eighth cranial nerve, or vestibulocochlear nerve, is
responsible for transmitting sound from the inner ear to the brain. This study also found significant
attenuation of the action potential (AP) of the eighth cranial nerve with ARCO exposure (12, 25, and
50 ppm CO) versus AR controls at PND22. This is complicated by the finding that AR control
animals had significant attenuation of the eighth cranial nerve AP versus MR control animals,
implying that artificial rearing contributes to AP changes before CO was introduced. Nonetheless,
the CO-dependent attenuation of the eighth cranial nerve AP (versus AR control) was permanent,
persisting until adulthood in the 50 ppm CO exposure group (the only CO group monitored).
Auditory brain stem response (ABR) conduction time was not affected in CO-exposed animals (12,
25, 50, 100 ppm).These functional tests reported that neonatal exposure to low concentrations of CO
can induce auditory functional changes in rodents.
Further studies have investigated physiological changes in cochlear development resulting
from chronic CO inhalation. Sprague Dawley rats exposed to low concentrations of CO (12 or
25 ppm, ARCO) from PND6 to PND27 had no evidence of damage to the inner or outer hair cells
(Lopez et al., 2003, 193901). However, CO (25 ppm) caused atrophy or vacuolization of the nerve
cells that innervate the inner (not outer) hair cells. Also, fibers of the eighth cranial nerve at the level
of the internal auditory canal had distorted myelination and vacuolization of the axoplasm after
25 ppm CO exposure. Energy production markers in the organ of corti and spiral ganglion neurons
January 2010
5-76
-------
including cytochrome oxidase (electron transport chain complex IV) and NADH-TR (marker of
complex I reductase activity) were significantly decreased after inhalation of 25 ppm (not 12 ppm)
CO versus control (AR and MR). Reduced energy production likely led to the decreased expression
of the calcium-mediated myosin ATPase and neurofilament proteins in the organ of corti and spiral
ganglion neurons (25 ppm CO). Since no changes in body weight were found after CO exposure in
these experiments (Stockard-Sullivan et al., 2003, 190947). it is likely that the decreased electron
transport chain enzymes are specific to vulnerable areas such as the cochlea.
Further analysis focused attention on CO-induced changes in the inferior colliculus (1C), an
auditory integrative section of the midbrain. Low concentrations of CO (12.5, 25, or 50 ppm) inhaled
over PND8-PND22 decreased c-Fos immunoreactivity in the central 1C at both PND27 and
PND75-PND77; immunostaining of other subregions of the 1C were not affected by CO (Webber et
al., 2003, 190515). c-Fos is an immediate early gene whose tonotopic expression corresponds to
neuronal activation in the auditory system. The same decrease in c-Fos expression was seen in rats
exposed to 25 or 100 ppm CO from PND9 to PND24 (Webber et al., 2005, 190514). These CO-
exposed rats also exhibited decreased neurofilament proteins and increased Cu-Zn superoxide
dismutase (SOD1) in the spiral ganglion neurons. This response could be ablated by dietary iron
restriction, suggesting an ROS-dependent contribution to the auditory changes seen after CO
exposure. These authors postulated that CO creates a persistent oxidative stress condition where
ROS generated via the interaction of peroxide and iron (via the Fenton reaction or Haber Weiss
chemistry) leads to impaired cochlear development; decreasing the available iron decreases the total
pool available for ROS generation. Further, the attenuation of the elevated SOD levels with iron
restriction post CO-exposure gives credence to this model.
A recent study has found comparable auditory system responses after prenatal (GD5-GD20)
exposure to CO with postnatal exposure (GD5-PND20,) similar to the studies described above
(Lopez et al., 2008, 097343). Prenatal CO (25 ppm) exposure led to high levels of the oxidative
stress markers HO-1, SOD-1, iNOS, and nitrotyrosine in cochlea vasculature and stria vascularis at
PND12; however, unlike postnatally exposed pups, HO-1 and SOD1 levels returned to normal at
PND20. Both groups of CO-exposed rats exhibited spiral ganglion cytoplasmic vacuolization, a
decrease in type I spiral ganglion neuron neurofilament proteins, thinning and damage in the cells of
the stria vascularis, and mild deterioration of the innervation of the inner hair cells. These nerve
terminals also had a persistent decrease in synapsin-1, a regulatory neuronal phosphoprotein. These
studies suggest that mild chronic CO exposure disrupts the developing auditory system, more often
at the IHC innervation and the eighth cranial nerve of the spiral ganglion, possibly by creating an
oxidative stress that may be reflected as hearing impairment.
Summary of Toxicological Studies on Developmental Central Nervous System Effects
Toxicological studies employing rodent models have shown that exposure to low
concentrations of CO during the in utero or perinatal period can adversely affect adult outcomes
including behavior, neuronal myelination, neurotransmitter levels or function, and the auditory
system. In utero CO exposure has been shown to impair active avoidance behavior (150 ppm),
habituation (75 and 150 ppm), nonspatial memory (75 and 150 ppm), and emotionality (150 ppm).
These behavioral changes could be due to neuronal changes or altered neurotransmitter signaling. In
utero CO exposure (75 and 150 ppm) was associated with PNS myelination decrements from
impaired sphingolipid homeostasis (150 ppm CO). These neuronal changes were also accompanied
by electrophysiological changes such as reversible delays in ion channel development and
irreversible changes in sodium equilibrium potential (150 ppm). Also, multiple studies demonstrated
that in utero CO exposure affected cholinergic (200 ppm), catecholaminergic (200 ppm),
noradrenergic (150 ppm), serotonergic (75 ppm), dopaminergic (75 ppm) and glutamatergic
(75 ppm), neurotransmitter levels or transmission in exposed rodents. Possible or demonstrated
adverse outcomes from the CO-mediated aberrant neurotransmitter levels or transmission include
respiratory dysfunction (150 ppm), impaired sexual behavior (150 ppm), and an adverse response to
hyperthermic insults resulting in neuronal damage (200 ppm). Finally, perinatal CO exposure has
been shown to affect the developing auditory system of rodents, inducing permanent changes into
adulthood at concentrations as low as 12 ppm. Together, these animal studies demonstrate that in
utero or perinatal exposure to CO can adversely affect adult behavior, neuronal function,
neurotransmission, and the auditory system in rodents.
January 2010 5-77
-------
Cardiovascular and Systemic Developmental Effects
In utero exposure to moderate to high concentrations of CO (60, 125, 150, 250, or 500 ppm) is
able to induce transient changes in cardiac morphology, cardiac action potentials, and systemic
immunity that may make a CO-exposed animal more susceptible to other outside stressors during the
immediate neonatal period. Studies of cardiovascular and systemic developmental responses to CO
levels of 500 ppm and less are presented below in Table 5-18.
Table 5-18. Cardiovascular and systemic developmental responses.
Study
Model
System
CO Exposure
Response
Notes
CARDIOVASCULAR AND SYSTEMIC DEVELOPMENTAL RESPONSES
Sartiani et al. (2004,
1908981
Prigge and Hochrainer
(1977, 0123261
Fechteretal. (1980,
011294)
Penney etal. (1982,
0113871
Styka and Penney
(1978, 0111661
Giustino etal. (1993,
0138331
Giustino et al. (1994,
076343)
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
was 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:
150ppm-15%
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
CO exposure at 4 wk of age. These CO-dependent changes were resolved by 8 wk of age, reflecting
a delayed maturation. Further, these authors postulated that a CO-dependent delay in
electrophysiological maturation of the cardiac myocyte (lack of APD shortening) could lead to
arrhythmias and thus could be associated with SIDS deaths.
Heart Morphological Changes after In Utero or Perinatal CO Exposure
Multiple authors have reported cardiomegaly following in utero exposure to low
concentrations of CO. Prigge and Hochrainer (1977, 012326) reported increased fetal Wistar rat
heart wet weight or cardiomegaly following continuous in utero CO (60, 125, 250, and 500 ppm)
exposure, with no decreases in near term fetal hematocrit or Hb levels seen at exposures below
250 ppm. Fechter et al. (1980, 011294) found that prenatal exposure to CO affected cardiac
development in exposed offspring. Long Evans rats that were exposed to CO continuously
January 2010
5-78
-------
(150 ppm) during gestation manifested with significant elevations in wet heart weight, as well as
heart weight in relation to body weight at PND1 but not at PND4, PND14, or PND21. Dry-to-wet
weight ratios revealed that the increased heart weight of CO-exposed pups at birth was due to edema
or water content. Penney et al. (1982, 011387) studied CO-dependent (500 ppm) cardiomegaly in
neonates (continuous CO exposure for 32 days starting at PND1). Other studies of adult male
Charles River-derived rats exposed to CO for 6 wk (at 400 or 500 to 1,100 ppm CO), as adults only,
developed CO-dependent cardiomegaly during exposure that significantly regressed within a couple
of months after termination of CO exposure (Styka and Penney, 1978, 011166).
Systemic Immune Toxicology after In Utero CO Exposure
In utero exposure (GDO-GD20) of male Wistar rats to moderate CO (0, 75, or 150 ppm)
concentrations induced reversible changes in macrophage function (Giustino et al., 1993, 013833).
The killing of Candida albicans (yeast) by splenic macrophages was significantly decreased at
PND15 in gestationally CO-exposed male offspring (75 and 150 ppm) but recovered function by
PND21. Macrophage phagocytosis of C. albicans was significantly reduced at PND15 and PND21 in
CO-exposed males (150 ppm only), and recovery was seen at PND60. Superoxide production by the
splenic macrophage respiratory burst was significantly decreased at PND15 and PND21 after in
utero CO exposure (150 ppm only), with recovery to control levels at PND60. In summary, CO
exposure in utero leads to a reversible and concentration-dependent loss of function of splenic
macrophages, with decreased killing ability, decreased phagocytosis, and decreased ROS production
during the macrophage respiratory burst.
Further studies by the same laboratory showed that in utero exposure of male rats to CO
(150 ppm) induced a subtle decrease in the frequency of splenic immunocompetent cells (leukocyte
common antigen [LCA+] cells) in a population of splenic immune cells at PND21 but not PND15 or
PND540 (Giustino et al., 1994, 076343). Specific LCA+ cell subpopulations, including
macrophages, Major Histocompatibility (MHC) II cells, T and B lymphocytes, showed a decreasing
trend but were not significant with CO exposure.
Summary of Toxicological Studies of Cardiovascular and Systemic Development
In utero CO exposure is associated with various adverse, albeit nonpersistent, cardiac
aberrations. Exposure to 150 ppm induced a delayed maturation of the cardiac action potential in
CO-exposed offspring. In other studies, continuous in utero CO exposure (60-500 ppm) induced
cardiomegaly at PND1, which was transient and regressed by PND4. CO (75 and 150 ppm) also
affects nonspecific immunity, shown through a reversible and dose-dependent loss of function of
splenic macrophages, with decreased killing ability, decreased phagocytosis, and decreased
macrophage ROS production (150 ppm). Also, the distribution of splenic immunocompetent cells
was slightly skewed because of a decrease in the number of LCA+ cells in PND21 male rats exposed
during gestation to 150 ppm CO. In conclusion, in utero exposure to moderate doses of CO
(60-500 ppm) is able to induce transient changes in cardiac morphology, cardiac action potentials,
and systemic nonspecific immunity.
5.4.3. Summary of Birth Outcomes and Developmental Effects
The most compelling evidence for a CO-induced effect on birth and developmental outcomes
is for PTB and cardiac birth defects. These outcomes were not addressed in the 2000 CO AQCD
(U.S. EPA, 2000, 000907). which included only two studies that examined the effect of ambient CO
on LEW. Since then, a number of studies have been conducted looking at varied outcomes, including
PTB, birth defects, fetal growth (including LEW), and infant mortality.
There is limited epidemiologic evidence that CO during early pregnancy (e.g., first month and
first trimester) is associated with an increased risk of PTB. The only U.S. studies to investigate the
PTB outcome were conducted in California, and these reported consistent positive associations with
CO exposure during early pregnancy when exposures were assigned from monitors within close
proximity of the mother's residential address. Additional studies conducted outside of the U.S.
provide supportive, though less consistent, evidence of an association between CO concentration and
PTB.
January 2010 5-79
-------
Very few epidemiologic studies have examined the effects of CO on birth defects. Two of
these studies found maternal exposure to CO to be associated with an increased risk of cardiac birth
defects. Human clinical studies also demonstrated the heart as a target for CO effects (Section 5.2).
Animal toxicological studies provided additional evidence for cardiac effects with reported transient
cardiomegaly at birth after continuous in utero CO exposure (60, 125, 250 and 500 ppm CO) and
delayed myocardial electrophysiological maturation (150 ppm CO). Toxicological studies have also
shown that continuous in utero CO exposure (250 ppm) induced teratogenicity in rodent offspring in
a dose-dependent manner that was further affected by dietary protein (65 ppm CO) or zinc
manipulation (500 ppm CO). Toxicological studies of CO exposure over the duration of gestation
have shown skeletal alterations (7 h/day, CO 250 ppm) or limb deformities (24 h/day, CO 180 ppm)
in prenatally exposed offspring.
There is evidence of ambient CO exposure during pregnancy having a negative effect on fetal
growth in epidemiologic studies. In general, the reviewed studies, summarized in Figure 5-10
through Figure 5-12, reported small reductions in birth weight (-5-20 g). Several studies examined
various combinations of birth weight, LEW, and SGA/IUGR, and inconsistent results were reported
across these metrics. It should be noted that having a measurable, even if small, change in a
population is different than having an effect on a subset of susceptible births and increasing the risk
of IUGR/LBW/SGA. It is difficult to conclude if CO is related to a small change in birth weight in
all births across the population or a marked effect in some subset of births. Toxicology studies have
found associations between CO exposure in laboratory animals and decrements in birth weight
(90-600 ppm), as well as reduced prenatal growth (65-500 ppm CO).
In general, there is limited epidemiologic evidence that CO is associated with an increased risk
of infant mortality during the neonatal or postneonatal periods. In support of this limited evidence,
animal toxicological studies provided some evidence that exogenous CO exposure to pups in utero
significantly increased postnatal mortality (7 h/day and 24 h/day, 250 ppm CO; 24 h/day, 90 or
180 ppm CO) and prenatal mortality (7 h/day, 250 ppm CO).
Evidence exists for additional developmental outcomes which have been examined in
toxicological studies but not epidemiologic or human clinical studies, including behavioral
abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes, and
changes in the auditory system. Structural aberrations of the cochlea involving neuronal activation
(12.5, 25 and 50 ppm CO) and auditory related nerves (25 ppm CO) were seen in pups after neonatal
CO exposure. Auditory functional testing using otoacoustic emissions testing (OAE at 50 ppm CO)
and 8th cranial nerve action potential (AP) amplitude measurements (12, 25, 50, 100 ppm CO) in
rodents exposed perinatally to CO showed auditory decrements at PND22 (OAE and AP) and
permanent changes in AP into adulthood (50 ppm CO). Furthermore, exogenous CO may interact
with or disrupt the normal physiological roles that endogenous CO plays in the body. There is
evidence that CO plays a role in maintaining pregnancy, controlling vascular tone, regulating
hormone balance, and sustaining normal ovarian follicular maturation.
Overall, there is limited though positive epidemiologic evidence for a CO-induced effect on
PTB and birth defects, and weak evidence for a decrease in birth weight, other measures of fetal
growth, and infant mortality. Animal toxicological studies provide support and coherence for these
effects. Both hypoxic and nonhypoxic mechanisms have been proposed in the toxicological literature
(Section 5.1), though a clear understanding of the mechanisms underlying reproductive and
developmental effects is still lacking. Taking into consideration the positive evidence for some birth
and developmental outcomes from epidemiologic studies and the resulting coherence for these
associations in animal toxicological studies, the evidence is Suggestive Of 3 Causal relationship
between relevant long-term exposures to CO and developmental effects and birth
outcomes
January 2010 5-80
-------
5.5. Respiratory Effects
5.5.1. Epidemiologic Studies with Short-Term Exposure
This section evaluates the key epidemiologic studies published since the 2000 CO AQCD
(U.S. EPA, 2000, 000907) that further examine the association between short-term exposure to CO
and respiratory morbidity. Although the number of studies that have specifically examined the CO-
respiratory health relationship have increased, there are still considerably less than that for the other
criteria air pollutants (e.g., PM and O3). The epidemiologic studies discussed below represent those
studies which (1) were conducted in locations with ambient CO concentrations similar to those in the
U.S.; (2) determined to use a reasonable study design and analytical methods; and (3) adequately
adjusted for confounding using accepted methods. If limitations in the design or analytical methods
used in a study were identified, they were noted. It is recognized that each of the studies evaluated
has a varying degree of exposure measurement error due to (1) the number of monitors used within
the study, the geographic size of the study area; (2) the spatial variability of CO; and (3) differences
in personal exposure distributions in the population; (Section 3.6.8) all of which could influence the
associations observed. As a result, in some instances specific details of a study are mentioned to
address any potential bias in the reported CO associations. Finally, the issue of confounding by
measured or unmeasured copollutants was evaluated, if possible, for each study, through the
interpretation of copollutant models. The results from copollutant models were used as an attempt to
disentangle the effect of CO from other pollutants while recognizing the high correlation between
CO and other combustion-related pollutants.
5.5.1.1. Pulmonary Function, Respiratory Symptoms, and Medication Use
The 2000 CO AQCD (U.S. EPA, 2000, 000907) briefly discussed the potential acute
respiratory health effects associated with short-term exposure to CO. An evaluation of the
epidemiologic literature at the time did not find any evidence of an association between short-term
exposure to CO and lung function, respiratory symptoms, or respiratory disease. As a result, the 2000
CO AQCD (U.S. EPA, 2000, 000907) did not conclude that a causal association exists between
short-term exposure to CO and respiratory health effects. Multiple uncertainties were identified in
the epidemiologic literature that contributed to this conclusion, which are discussed in Section 5.2.1.
The following section evaluates the current literature that examines the potential association between
short-term exposure to CO and respiratory health effects. Table 5-19 lists the studies evaluated in this
section along with the respiratory health outcomes examined and CO concentrations reported.
January 2010 5-81
-------
Table 5-1 9.
Study
O'Connor etal.
(2008, 156818)"
Rabinovitch et al.
(2004, 096753)
Silkoff etal. (2005,
087471)
Fischer et al.
(2002.025731)'
Ranzi etal. (2004,
089500)'
Lagorio et al.
(2006.089800)'
Penttinen et al.
(2001.030335)'
Timonen et al.
(2002.025653)'
Chen etal. (1999,
011149)
Delfino et al.
(2003, 050460)
Slaughter et al.
(2003, 086294)
Yu et al. (2000,
013254)
Schildcrout et al.
(2006, 089812)
von Klot et al.
(2002.034706)'
Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with short-term exposure to CO.
Location
Sample Size
7 U.S. cities
Denver, CO
(Year1: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)
8 North American
cities
(n = 990)
Erfurt, Germany
(n = 53)
Years
8/1998-7/2001
11/1999-3/2000;
11/2000-3/2001;
11/2001-3/2002
1999-2000
(winter);
2000-2001
(winter)
March -April"
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
11/1993-9/1995
10/1996-3/1997
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
Asthma symptoms;
Medication use
Asthma symptoms;
Medication use
Metric
8-h max
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
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
NR
0.78
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
50th: 0.63-1 .49
75th: 0.77-1 .90
90th: 0.95-2. 40
50th: 0.70
75th: 1.04
Maximum: 2.60
January 2010
5-82
-------
Study
Park etal. (2005,
088673)
Rodriguez et al.
(2007, 092842)
de Hartog et al.
(2003.001061)'
Location H6cilth
Sample Size ears Outcome e nc
3/2002-6/2002 j- i- ' 24-h 3VQ
(n = 64) Medication 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
(n = 47)
Helsinki, Finland
(n = 47)
Mean
Concentration
(ppm)
Control days: 0.64
Dust days: 0.65
1.41
Amsterdam: 0.52
Erfurt: 0.35
Helsinki: 0.35
Middle/Upper Percentile
Concentrations (ppm)
NR
Maximum: 8.03
Maximum:
Amsterdam: 1
Erfurt: 2.17
Helsinki: 0.87
.39
"These 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.
bThis study did not present air quality statistics quantitatively, as a result, the air quality statistics presented were estimated from a figure.
'This study did not provide the year(s) in which air quality data was collected.
Pulmonary Function
As part of the Inner-City Asthma Study (1CAS), O'Connor et al. (2008, 156818) examined the
effect of air pollutants (i.e., PM2 5, O3, NO2, CO, and SO2) on lung function in a population of 861
children (5-12 yr old) with persistent asthma in 7 urban U.S. communities. Throughout the study,
percent predicted forced expiratory volume in 1 s (FEVi) and peak expiratory flow (PEF) were
examined for each subject during 2-wk periods twice daily every 6 mo for 2 yr. Lung function was
examined in single pollutant models using both same-day (lag 0) and 5-day (lag 0-4) ma pollutant
concentrations (Figure 5-13). CO was not found to be associated with percent predicted FEVi at lag
0, but there was some evidence for a reduction in percent predicted FEVi when using the 5-day ma
(-0.32 [95% CI: -0.75 to 0.11] per 0.5 ppm increase in 24-h avg CO concentrations). When
examining percent predicted PEF, a small reduction was observed at lag 0 (not reported
quantitatively), but the effect was found to be slightly larger at lag 0-4 (-0.28 [95% CI: -0.71 to
0.15]). In this study, CO was found to be moderately correlated with other combustion related
pollutants (e.g., PM2.5 [r = 0.44] and NO2 [r = 0.54]); however, CO was not included in the
multipollutant models examined, limiting the interpretation of the small reductions in lung function
observed. Although the observed reductions in lung function did not reach statistical significance, the
results do provide some evidence for a potential effect of CO on lung function at relatively low CO
concentrations (99th percentile max 8-h avg concentrations: ~ 3.8 ppm).
January 2010
5-83
-------
Ozone PM2.5 SO2
NO2
Ozone PM2.5 SO2 NO2 CO
Source: Reprinted with Permission of Elsevier Ltd. from O'Connor et al. (2008,1568181
Figure 5-13. Estimated effect (95% confidence intervals) 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 avg of pollutant concentration. Effect estimates were adjusted for site,
month, site-by-month interaction, temperature, and intervention group in mixed
models. Panel A: percent predicted FEVi as outcome variable; Panel B: percent
predicted PEFR as outcome variable.
The remaining U.S.-based studies evaluated consisted of single-city studies conducted in
Denver, CO. Rabinovitch et al. (2004, 096753) examined the association between exposure to
ambient air pollutants and asthma exacerbation in a panel of urban minority children, 6-12 yr old,
with moderate to severe asthma over three winters. The investigators examined pulmonary function
by measuring FEVi and PEF in the morning on school days and also at night on weekends or other
nonschool days. Using a 3-day ma (lag 0-2) for all pollutants, Rabinovitch et al. (2004, 096753) did
not find an association between CO and either lung function parameter during the morning or at
night. Silkoff et al. (2005, 087471) also examined lung function during the winter months, but in a
panel of former smokers that were at least 40 yr old and had been diagnosed with COPD. In this
study, CO concentrations were similar to those reported in Rabinovitch et al. (2004, 096753). The
authors examined the association between exposure to air pollutants and lung function (i.e., FEVi
and PEF) in both the morning and the evening. Silkoff et al. (2005, 087471) found contradictory
results when examining the effects of CO for each of the winter periods (1999-2000 and 2000-2001)
separately. During the analysis of the first winter (i.e., 1999-2000), CO was not found to be
associated with lung function decrements in the morning at any lag, but there was some evidence for
lung function decrements during the evening at lag 0. Of note is the increase in FEVi during the
morning that was observed at lag 1 during this time period. For the second winter (i.e., 2000-2001)
the authors found a significant negative association between CO exposure and FEVi in the evening
at lag 2 and a moderate negative association with PEF at lag 0 in the morning and lag 2 in the
evening. Silkoff et al. (2005, 087471) postulated that the difference in the FEVi results for the two
study periods could be due to higher pollution concentrations along with somewhat lower
temperatures and higher humidity in 2000-2001. However, mean CO levels remained relatively
constant between the first and second winters, whereas, PMio, PM2.5, and NO2 concentrations all
increased. The decrements in FEVi observed in the second winter, therefore, may have been due to
the slightly worse, although not significantly different, baseline lung function of the panel of subjects
used during the second winter (Silkoff et al., 2005, 087471).
In the recent literature, the majority of studies that examined the association between short-
term exposure to CO and lung function have been conducted in Europe and Asia. These studies
provide some evidence for CO-induced changes in lung function. Negative associations between
short-term exposure to CO and lung function were observed primarily in individuals with underlying
respiratory conditions; however, some evidence also exists for effects in children that live in urban
environments. Penttinen et al. (2001, 030335) examined the association between CO and lung
function in a panel consisting of 57 nonsmoking adult asthmatics during the winter and spring in
Helsinki, Finland. The authors observed negative associations with PEF (L/min) for a 0.5 ppm
increase in 24-h avg CO concentrations in the morning at lag 1 ((3 = -0.54, SE = 0.084) and in the
January 2010
5-84
-------
afternoon ((3 = -1.52, SE = 0.29) and evening ((3 = -1.81, SE = 0.27) for a 5-day avg. In two-pollutant
models with daily mean particle number concentration (PNC), CO effects on PEF in the morning
were attenuated at lag 1 but remained negative. In addition, negative associations with PEF persisted
in the afternoon and evening in the two-pollutant model at lag 0. In this study, moderate correlations
between UFPs and other traffic-generated pollutants (e.g., CO [r = 0.44], NO [r = 0.60], and NO2
[r = 0.44]) make it difficult to attribute the observed respiratory effects to a specific pollutant.
Lagorio et al. (2006, 089800) also conducted a study that examined the association between
CO and lung function in adults. In this study, three panels of subjects with underlying asthma,
COPD, or IHD, who resided in Rome, Italy, were selected. The ages of the subjects varied depending
on the panel, but overall the subjects ranged from 18-80 yr old. In single-pollutant models with CO,
a reduction in forced vital capacity (FVC) and FEVi was observed at most of the lags examined
(i.e., 0, 0-1, and 0-2) for both the COPD and asthma panels. No association was observed between
CO and FVC or FEVi in the IHD panel. Lagorio et al. (2006, 089800) did observe a relatively high
correlation between CO and PM2.5 but not NO2 (r = 0.05). Copollutant models were not conducted in
this analysis to identify whether the CO associations observed are potentially confounded by other
pollutants.
Studies that focused on alterations in lung function in asthmatic children reported results
consistent with those observed in adult asthmatics. Timonen et al. (2002, 025653) examined the
effect of CO on bronchial responsiveness and pulmonary function (i.e., FVC, FEVi, MMEF, and
AEFV) at rest and after exercise in a panel of children, 7-12 yr old with chronic respiratory
symptoms, during the winter in Kuopio, Finland. The authors found that CO was significantly
associated with decrements in baseline lung function (i.e., lung function measured prior to exercise)
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
lag 3 (-20.9 mL), for a 0.5 ppm increase in 24-h avg CO concentration. CO was not found to be
associated with exercise-induced changes in lung function or bronchial responsiveness. Overall,
Timonen et al. (2002, 025653) found that increased concentrations of combustion-related byproducts
(i.e., BS, PMio, particle numbers, NO2, and CO) was associated with impairment in baseline lung
function. These associations, along with the high correlation between CO and combustion-related
pollutants (e.g., PMi0 [r = 0.64]; NO2 [r = 0.88]), contributed to the inability of the authors to
conclude that the lung function effects observed were due to biological changes in lung pathology
specific to CO exposure.
Chen et al. (1999, 011149) examined the effect of CO on lung function in 941 8- to 13-yr-old
asthmatic children in Taiwan. The authors observed an association between short-term exposure to
CO and decrements in FVC (mL) at a 2-day lag when using daytime average CO concentrations
(from 8:00 a.m. to 6:00 p.m.) in a single-pollutant model. However, the authors found a high
correlation between CO and NO2 concentrations (r = 0.86-0.98), and did not conduct copollutant
analyses.
An additional study, Fischer et al. (2002, 025731). examined the association between CO and
respiratory health, specifically lung function in a cohort study of 68 children ages 10-11 yr who lived
in an urban environment (Utrecht, The Netherlands). In this study, the authors examined whether
eNO was a more sensitive measure of lung damage than the traditional pulmonary function
measurements (i.e., FVC, FEVi, PEF, and MMEF). Fischer et al. (2002, 025731) found negative
associations between CO and FEVi, PEF, and MMEF at both lags 1 and 2, as well as an association
between CO and an increase in eNO at lag 1. However, the study did not provide pollutant
correlations or examine copollutant models, limiting the interpretation of these results.
Respiratory Symptoms in Asthmatic Individuals
Upon evaluating the literature that examined the association between short-term exposure to
CO and respiratory symptoms in asthmatic individuals, consistent, positive associations were
observed across studies. The studies evaluated that included children enrolled in the Childhood
Asthma Management Program (CAMP) study found that CO was positively associated with asthma
symptoms. Yu et al. (2000, 013254) found a 1.14-fold increase in asthma symptoms
([95% CI: 1.05-1.23] per 0.5 ppm increase in 24-h avg CO) at lag 1 in a population of 5- to 13-yr-old
asthmatic children (n = 133) in Seattle, WA. Similar effects were observed at lag 0 and lag 2. These
effects persisted when controlling for previous day's asthma symptoms at all lags, with the largest
effect at lag 1 (OR=1.12 [95% CI: 1.05-1.19]) and in multipollutant models with PMLO and SO2.
January 2010 5-85
-------
Using the same population of children, Slaughter et al. (2003, 086294) found an association between
short-term exposure to CO at lag 1 and asthma severity both with and without controlling for the
previous day's asthma severity, (RR = 1.04 [95% CI: 1.01-1.08] and RR = 1.03 [95% CI: 1.00-1.05]
per 0.5 ppm increase in 24-h avg CO, respectively). However, this study only examined the effect of
copollutant models on PM risk estimates, not CO. Schildcrout et al. (2006, 089812) examined the
association between air pollutants and asthma symptoms in 990 children ages 5-12 yr in 8 North
American cities. The authors found a positive association between short-term exposure to CO and
asthma symptoms at lag 0 (OR = 1.04 [95% CI: 1.00-1.07] per 0.5 ppm increase in 24-h avg CO),
but similar effects were also observed at lag 1, 2, and the 3-day moving sum. The CO effects
observed persisted when NO2, PMi0, and SO2 where included in joint-pollutant models.
As previously mentioned, O'Connor et al. (2008, 156818) conducted an additional multicity
study to examine the effect of air pollutants (i.e., PM2.5, O3, NO2, CO, and SO2) on respiratory health
in a population of 861 children (5-12 yr) with persistent asthma in 7 U.S. urban communities. The
authors collected information on asthma symptoms every 2 mo and examined the association
between a 2-wk recall of the asthma symptoms and each air pollutant. O'Connor et al. (2008,
156818) used a 19-day lag, which encompassed the 14 days of the symptom recall period and the
5-day lag period proceeding the symptom recall period. In a single-pollutant model, CO was
significantly associated with number of days with a wheeze-cough (14% [95% CI: 2-29]), number of
nights with asthma symptoms (i.e., nighttime asthma) (19% [95% CI: 4-36]), and number of days a
child slowed down or stopped play (15% [95% CI: 2-30]) per 0.5 ppm increase in 24-h avg CO
concentrations during the 2-wk recall period. In this study, CO effects were not examined in a
copollutant model.
U.S.-based single-city studies also found positive associations between CO and asthma
symptoms (Delfino et al., 2003, 050460: Rabinovitch et al., 2004, 096753). Rabinovitch et al. (2004,
096753) found evidence for an increase in asthma exacerbations in response to 24-h avg CO
concentrations for a 3-day ma (lag 0-2) (OR = 1.02 [95% CI: 0.89-1.16] per 0.5 ppm increase in 24-h
avg CO) in a population of urban poor children with moderate to severe asthma in Denver, CO.
Delfino et al. (2003, 050460) also reported evidence of a positive association between CO and
asthma symptoms (based on symptoms that interfere with daily activities) using a population of
Hispanic children with asthma in a Los Angeles, CA, community. However, Delfino et al. (2003,
050460) only found positive associations at 1-day lags when using either the 1-h max (OR=1.05
[95% CI: 0.88-1.26] per 1 ppm increase in 1-h max CO concentrations) or max 8-h avg (OR=1.09
[95% CI: 0.80-1.50] per 0.75 ppm increase in max 8-h avg CO concentrations) CO concentration as
the exposure metric. It should be noted that in comparison to Rabinovitch et al. (2004, 096753) and
the other respiratory symptoms studies discussed above, the mean ambient concentrations for 1-h
max and max 8-h avg reported by Delfino et al. (2003, 050460) were 7.7 ppm and 5.0 ppm,
respectively, both of which are approximately 3.5 times higher than the corresponding 24-h avg
concentrations reported in the other studies.
In addition to the U.S.-based studies presented above, international studies were evaluated
that examined the association between short-term exposure to CO and asthma symptoms in study
populations that included adults. Figure 5-14 summarizes the results from studies that provided
comparable quantitative results and examined the association between short-term exposure to CO
and asthma or respiratory symptoms in asthmatic individuals. A panel study consisting of 53 adults
with asthma or asthma symptoms in Germany (Von et al., 2002, 034706) observed a marginal
association between CO concentration and the prevalence of wheezing at lag 0 (OR = 1.03
[95% CI: 0.97-1.08] per 0.5 ppm increase in 24-h avg CO), and a positive association for a 5-day
mean concentration (OR = 1.12 [95% CI: 1.05-1.21] per 0.5 ppm increase in 24-h avg CO).
However, the authors found CO to be highly correlated with UFPs (r = 0.66), complicating the
interpretation of the associations observed. Additionally, Park et al. (2005, 088673) in a panel study
of individuals 16-75 yr old in Incheon, Korea, with bronchial asthma, did not find an association
between CO and nighttime asthma symptoms or cough.
To further examine the effect of CO on asthma and asthma symptoms, some studies also
analyzed medication use in asthmatic individuals in response to an increase in air pollutant
concentrations. The majority of U.S.-based studies (i.e., Rabinovitch et al., 2004, 096753;
Schildcrout et al., 2006, 089812; Slaughter et al., 2003, 086294) focused on rescue inhaler use in
children with ages ranging from 5 to 13 yr. Rabinovitch et al. (2004, 096753) found a weak
association (OR = 1.08 [95% CI: 1.00-1.17] per 0.5 ppm increase in 24-h avg CO) between rescue
inhaler use in a population of 6- to 12-yr old urban minority children with moderate to severe asthma
January 2010 5-86
-------
in the winter in Denver, CO. In a population of 5- to 12-yr-old children with asthma in Seattle, WA,
Slaughter et al. (2003, 086294) found a stronger association with rescue inhaler use both with and
without taking into consideration the previous day's asthma severity, (RR: 1.04 [95% CI: 1.01-1.08]
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
24-h avg CO), respectively. Similar results were observed in a multicity study conducted by
Schildcrout et al. (2006, 089812) which analyzed rescue inhaler use in 990 children ages 5-13 yr
with asthma in 8 North American cities. Schildcrout et al. (2006, 089812) found that short-term
exposure to CO was positively associated with rescue inhaler use at lags of 0, 2, and a 3-day moving
sum, and that the association was fairly robust to a simultaneous increase in CO and other pollutants
(i.e., NO2, PMio, and SO2) in joint models. Overall, Slaughter et al. (2003, 086294) and Schildcrout
et al. (2006, 089812) question the associations observed due to the lack of biological plausibility for
CO-induced respiratory effects and the high correlation between CO and NO2 (which suggests that
other pollutants from mobile sources are driving the associations observed), respectively. Additional
studies (Park et al., 2005, 088673; Silkoff et al., 2005, 087471; Von et al., 2002, 034706) conducted
in Denver, CO; Erfurt, Germany; and Incheon, Korea, respectively, found associations between CO
and medication use that are consistent with those previously reported, but in populations with
combined ages ranging from 16 to 77 yr. Figure 5-14 presents the risk estimates from studies that
examined the association between short-term exposure to CO and medication use in asthmatic
individuals.
£•
1
Q.
I
c
1
Symptoms
o
3
Study
Schildcrout et al. (2006, 089812)
O'Connor etal. (2008, 156818)
O'Connor et al. (2008, 156818)
Yu etal. (2000, 01 3254)
Rabinovitch etal. (2004, 096753)
Delfino etal. (2003,050460)
Delfino etal. (2003,050460)
von Klot etal. (2002, 034706)
Schildcrout etal. (2006, 08981 2)
Slaughter et al. (2003, 086294)
Slaughter et al. (2003, 086294)
Rabinovitch etal. (2004, 096753)
von Klot etal. (2002, 034706)
von Klot etal. (2002, 034706)
Location
8 NAmer cities
7 US cities
7 US cities
Seattle, WA
Denver, CO
LA.CA
LA.CA
Erfurt, Germany
8 N.Amer cities
Seattle, WA
Seattle, WA
Denver, CO
Erfurt, Germany
Erfurt, Germany
Laa
0-2J
0-18
0-18
1
0-2
1
1
0-5
"fJZ^
1
1
0-2
0-5
0-5
Age
5-12
5-12
5-12
5-13
6-12
10-16
10-16
5-12
5-13
5-13
6-12
37-77
37-77
Effect Estimate (95% CI)
Asthma Symptoms i*
Asthma Symptoms j — • —
Asthma Exacerbation 1»
Symp Score >2 (1 -h max) 1-»
Symp Score >2 (max 8-h avg) ! — •
Wheeze j — • —
Inhaler Use m
Inhaler Use" '-•-
InhalerUse'' [«•
Inhaler Use , — • —
InhalerUse (32-agonist) ->-» —
InhalerUse (Corticosteroid) ' — •
Children
Adults
Children
Adults
1
1 1 1 1 1
aO-2 lag represents a 3-day moving sum 0.75 1.00 1.25 1.50 1.75
'Estimate presented is from a model that did not control for previous day's asthma severity.
'Estimate presented is from a model that controlled for previous day's asthma severity. Odds Ratio
Figure 5-14. Summary of associations for short-term exposure to CO and asthma symptoms,
respiratory symptoms and medication use in asthmatic individuals.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.
Respiratory Symptoms in Nonasthmatic Individuals
In addition to examining the association between short-term exposure to CO and respiratory
symptoms (e.g., cough, wheeze, shortness of breath) in asthmatic populations, some studies
examined respiratory effects in individuals classified as nonasthmatics. Rodriguez et al. (2007,
092842) examined the effect of CO on respiratory symptoms in a panel of 263 children 0- to 5-yr old
at high risk for developing asthma in Perth, Australia. Rodriguez et al. (2007, 092842) found CO
concentrations to be positively associated with wheeze/rattle chest and runny/blocked nose at both a
5-day lag and a 0-5-day lag. In this study, copollutant models were not examined, CO correlations
with other pollutants were not presented, and additional analyses were not conducted to further
characterize the associations observed.
1 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.
January 2010
5-87
-------
In a panel of individuals > 50 yr of age with CHD in three European locations (Amsterdam,
The Netherlands; Erfurt, Germany; and Helsinki, Finland) during the winter, de Hartog et al. (2003,
001061) observed some marginal associations, specifically between CO concentration and the
incidence of the respiratory symptoms shortness of breath and phlegm at lag 3, OR=1.17 (95% CI:
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
concentrations. However, the authors found that the associations between air-pollution exposure and
respiratory symptoms were stronger for PM2.5 than for gaseous air pollutants. Overall, the
associations observed in this study should be viewed with caution because the panel consisted of
individuals on a variety of daily medications (i.e., beta blockers, ACE inhibitor + AT blocker,
calcium antagonist, aspirin, digitalis, inhaled beta-agonist, and nitroglycerin).
Summary of Associations between Short-Term Exposure to CO and Pulmonary
Function, Respiratory Symptoms, and Medication Use
A limited body of evidence is available that examined the effect of short-term exposure to CO
on various respiratory health endpoints. Specifically, among asthmatics, the studies reviewed
generally found positive associations between short-term exposure to CO and respiratory-related
health effects (i.e., decrements in lung function, respiratory symptoms, and medication use). On-road
vehicle exhaust emissions are a nearly ubiquitous source of combustion pollutant mixtures that
include CO and can be an important contributor to CO-related health effects in near-road locations,
which is evident by the high correlations reported between CO and other combustion-related
pollutants (i.e., NO2 and PM). However, the limited number of copollutant analyses among this
group of studies complicates the efforts to disentangle the health effects attributed to CO from the
larger traffic-related pollutant mix. Additional uncertainty exists as to a biologically plausible
mechanism that could explain the effect of CO on respiratory health.
5.5.1.2. Respiratory Hospital Admissions, ED Visits and Physician Visits
The 2000 CO AQCD (U.S. EPA, 2000, 000907) evaluated a limited amount of literature that
examined the association between short-term exposure to CO and respiratory hospital admissions
(HAs), ED visits, and physician visits in the U.S. (i.e., Seattle, WA; Reno, NV; and Anchorage, AK)
and Europe (i.e., Barcelona, Spain). From these studies, the 2000 CO AQCD (U.S. EPA, 2000,
000907) concluded that positive associations were observed for short-term exposure to CO with
several respiratory outcomes, including asthma and COPD. However, the lack of a biologically
plausible mechanism for CO-induced respiratory morbidity at that time brought into question
whether the results observed could be attributed to CO independently of other pollutants in the air
pollutant mixture. Additional uncertainties were identified in the epidemiologic literature that
contributed to this conclusion, which were discussed in Section 5.2.1.
This section evaluates those studies published since the 2000 CO AQCD (U.S. EPA, 2000,
000907) that examined the association between short-term exposure to CO at ambient concentrations
similar to those found in the U.S. and respiratory-related HAs (Figure 5-15), ED visits (Figure 5-16),
and physician visits. Unlike previous sections, which also evaluated studies conducted outside of
North America, the expansive number of studies conducted in the U.S. and Canada provide adequate
evidence to examine the association between short-term exposure to CO and respiratory HAs and
ED visits. Although not discussed in this section, collectively, the studies conducted outside of the
U.S. observed associations that are consistent with those observed in the U.S.- and Canadian-based
studies evaluated below (see Annex C for results from the international studies evaluated).
Overall, this section focuses on respiratory-related HAs because the majority of the literature
examines HAs as opposed to ED visits or physician visits (Table 5-20 presents the studies evaluated
in this section along with the range of CO concentrations measured in each study). It must be noted
that when examining the association between short-term exposure to CO and health outcomes that
require medical attention, it is important to distinguish between HAs, ED visits, and physician visits
for respiratory outcomes (more so than for cardiovascular outcomes). This is because it is likely that
a small percentage of respiratory ED visits will be admitted to the hospital and, therefore, may
represent potentially less serious but more common outcomes. To adequately distinguish between the
results presented in HAs, ED visit, and physician visit studies, each outcome is evaluated in
individual sections. In addition, each section presents results separately for respiratory health
January 2010 5-88
-------
outcomes which include all respiratory diagnoses (ICD-9: 460-519) or selected diseases
(e.g., asthma, COPD, pneumonia and other respiratory infections) in order to evaluate the potential
effect of short-term exposure to CO on each outcome.
Table 5-20.
Range of CO concentrations reported in key respiratory HA and ED
examine effects associated with short-term
Study
Cakmak et al.
(2006, 093272)
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)
(2003, 042864)
Yang et al.
(2005, 090184)
Karretal.
(2006, 088751)
Location
10 Canadian
cities
Los Angeles,
CA
Spokane, WA
Toronto, ON,
Can
Vancouver, BC,
Can
Toronto, ON,
Can
Vancouver, BC,
Can
Cook County,
IL; Los Angeles
County, CA
Vancouver, BC,
Can
South Coast Air
Basin, CA
Type of Visit (ICD9)
HospitalAdmissions: Respiratory
disease (i.e., Acute bronchitis and
bronchiolitis; pneumonia; chronic and
unspecific bronchitis; emphysema;
asthma; bronchiectasis; chronic
airway obstruction)
HospitalAdmissions: Pulmonary;
asthma; COPD
ED Visits and HospitalAdmissions:
Respiratory; asthma; COPD;
pneumonia; acute respiratory
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
Metric
24-h avg
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
exposure to CO.
Mean Concentration (ppm)
0.8
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
Referent: 1.750
Lag 4:
Index: 1.760
Referent: 1.790
visit studies that
Middle/Upper Percentile
Concentrations (ppm)
Maximum: 6.5
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
Referent:
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
Referent:
50th: 1.55; 75th: 2.35;
90th: 3. 30
Maximum: 9.60
January 2010
5-89
-------
Study
Location
Type of Visit (ICD9)
Metric Mean Concentration (ppm)
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
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
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
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
The majority of studies from North America that examined the association between short-term
exposure to CO and HAs for all respiratory diseases were conducted in Canada, and only one of
these studies presented results from a combined analysis of multiple cities (Cakmak et al., 2006,
093272). In a study of 10 of the largest Canadian cities, Cakmak et al. (2006, 093272) examined
respiratory HAs (ICD-9: 466, 480-486, 490-494, 496) in relation to ambient gaseous pollutant
concentrations for the time period 1993-2000. This study reported a 0.37% (95% CI: 0.12-0.50)
increase in respiratory hospital admissions for all ages for a 0.5 ppm increase in 24-h avg CO (lag
2.8 days averaged over the 10 cities1). However, Cakmak et al. (2006, 093272) only examined the
potential confounding effects of gaseous pollutants (i.e., NO2, SO2, and O3) on the CO risk estimate
in a multipollutant model and did not provide correlation coefficients, limiting the interpretation of
the effects observed in the single-pollutant model. U.S.-based studies (Los Angeles and Spokane)
that examined HAs for all respiratory diseases reported similarly weak or null associations with CO
(Linn et al., 2000, 002839; Slaughter et al., 2005, 073854). But two single-city studies conducted in
Canada reported stronger associations, primarily through evidence from copollutant models, between
short-term exposure to CO and respiratory disease HAs (Burnett et al., 2001, 093439; Yang et al.,
2003, 055621). In a study conducted in Toronto, Canada, for the time period 1980-1994, Burnett
et al. (2001, 093439) reported a relatively strong association between 1-h max CO and respiratory
disease HAs in children <2 yr of age for the diagnoses of asthma (493), acute bronchitis/bronchiolitis
(466), croup (464.4), and pneumonia (480-486). The authors found a 9.7% (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 the two-pollutant
model analysis with O3, the estimate for CO remained elevated (7.29% [95% CI: 1.75-13.1]), but CO
1 To determine the lag for the combined estimate across all 10 cities, Cakmak et al. (2006, 093272) averaged the strongest associations
from lags 0-5 days from each city.
January 2010
5-90
-------
was not found to be highly correlated with O3 (r = 0.24). Yang et al. (2003, 055621) reported similar
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
(<3 yr of age) respiratory disease (ICD-9: 460-519) HAs in Vancouver for the time period
1986-1998. Yang et al. (2003, 055621) also reported elevated associations with 24-h avg CO and
respiratory HAs (ICD-9: codes 460-519) for ages 65 yr and over in Vancouver, Canada, (OR = 1.02
[95% CI: 1.00-1.04]) at lag 1 for a 0.5 ppm increase in 24-h avg CO. The authors found that the CO
risk estimates remained the same when O3 was included in the model, which could be attributed to
the lack of collinearity between CO and O3 due to their negative correlation (r = -0.52).
Asthma
Some studies that examined the effect of short-term exposure to CO on asthma HAs conducted
all age and age-stratified analyses, specifically to examine effects in children. In a few studies
conducted in Canada, evidence was observed for increased pediatric (ages 6-12 yr) asthma HAs
(ICD-9: 493) in boys but not girls (Lin et al., 2003, 042549: Lin et al., 2004, 055600): however, a
biological explanation was not provided which could explain this difference. Lin et al. (2003,
042549) used a bidirectional case-crossover analysis in Toronto, Canada, for the years 1981-1993.
The authors reported an OR of 1.05 (95% CI: 1.00-1.11) per 0.5 ppm increase in 24-h avg CO for a
1-day lag for boys, with similar results being reported when averaging CO concentrations up to
7 days prior to an HA. Risk estimates for girls did not provide evidence of an association using the
same lag structure that was used in the boys' analysis (OR = 1.00 [95% CI: 0.93-1.06]); lag 1). In
this study, CO levels were moderately correlated with NO2 (r = 0.55) and PM2.5 (r = 0.45), and
weakly correlated with SO2 (r = 0.37). In this study, copollutant analyses were not conducted to
examine the potential confounding effect of different PM size fractions or gaseous pollutants on CO
risk estimates. It should be noted that this study used a bidirectional case-crossover analysis, which
may bias the results in either direction (Levy et al., 2001, 017172). Studies that examined the various
referent selection strategies for the case-crossover study design have concluded that the preferred
control selection strategy is the time-stratified framework (Levy et al., 2001, 017172). Lin et al.
(2004, 055600) also examined the association between air pollutants and asthma HAs in children,
but using a time-series study design in Vancouver during the years 1987-1998. In this study, the
authors stratified results by socioeconomic status (SES) and found some evidence for an association
between CO and 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 increase in 24-h avg CO); but overall, the evidence was less consistent
for a greater effect in boys versus girls compared to Lin et al. (2003, 042549). In a study that
examined asthma HAs for all ages and genders combined, Slaughter et al. (2005, 073854) observed
some evidence for an increase in asthma HAs (ICD-9 493) in Spokane (1995-2000) for CO at lag 2
(RR = 1.03 [ 95% CI: 0.98-1.08]) 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
A few of the studies examined the effect of short-term exposure to CO on COPD, or
obstructive lung disease, and HAs. Moolgavkar (2003, 042864) (a reanalysis of Moolgavkar, 2000,
010274) examined HAs for COPD plus "allied diseases" (ICD-9 490-496) in two U.S. counties
(Cook County, IL, and Los Angeles County, CA) for the years 1987-1995, using Poisson generalized
linear models (GLMs) or generalized additive models (GAM), with the more stringent convergence
criteria. Overall, the results from both models were similar. Using the GAM models, the study
reported increases in HAs of 0.53-1.20% for all ages in Los Angeles County and 0.17-1.41% for
ages > 65 yr in Cook County, for a 0.5 ppm increase in 24-h avg CO and lags ranging from 0 to
5 days. However, CO was found to be highly correlated with NO2 in both Cook County (r = 0.63)
and Los Angeles County (r = 0.80), but Moolgavkar (2003, 042864) did not examine the influence of
copollutants on CO risk estimates. Yang et al. (2005, 090184) reported similar results for COPD HAs
(ICD-9 490-492, 494, 496) in Vancouver for ages > 65 yr for the years 1994-1998 for a ma of 0- to
6-day lags (RR =1.14 [95% CI: 1.03-1.23] per 0.5 ppm increase in 24-h avg CO). In this study, CO
concentrations were moderately correlated with NO2, SO2, and PM10, and moderately negatively
correlated with O3. In copollutant models, Yang et al. (2005, 090184) found that risk estimates for
CO and COPD HAs remained elevated with O3 (RR=1.19 [95% CI: 1.07-1.32]) or SO2 (RR=1.19
[95% CI: 1.02-1.39]), but were attenuated when adjusting for NO2 (RR=1.07 [95% CI: 0.92-1.24])
or PM10 (RR=1.03 [95% CI: 0.89-1.21]). Contrary to Moolgavkar (2003, 042864) and Yang et al.
January 2010 5-91
-------
(2005, 090184). Slaughter et al. (2005, 073854) found no association between short-term exposure to
CO and COPD HAs (ICD-9 491, 492, 494, 496) in Spokane, WA, at lag 1-day (RR = 0.97
[95% CI: 0.93-1.01] per 0.5 ppm increase in 24-h avg CO) with similar results being reported for 2-
and 3-day lags. However, this study did not examine correlations between CO and other gaseous
pollutants or conduct copollutant analyses.
Acute Bronchiolitis in Infants
Karr et al. (2006, 088751: 2007, 090719) examined both short-term (lag 0 or 1) and longer
term levels of CO in relation to acute bronchiolitis (ICD-9: 466) HAs during the first year of life
from 1995-2000 in the South Coast Air Basin in California. Karr et al. (2006, 088751) found no
evidence of a short-term association between ambient CO concentrations and HAs for acute
bronchiolitis at lag 1 day (OR= 0.99 [95% CI: 0.98-1.01] per 0.5 ppm increase in 24-h avg CO). In
addition, Karr et al. (2007, 090719). which examined longer term exposures (average in the month
prior to a HA and lifetime average) in a matched case-control study, did not provide any evidence of
an association with CO. Neither of these studies examined the correlation between CO and other
pollutants nor conducted copollutant analyses.
Pneumonia and Other Respiratory Infections
In addition to examining the effect of short-term exposure to CO on health outcomes that can
limit the function of the respiratory system, some studies examined the effect of CO on individuals
with pneumonia (ICD-9: 480-486) separately or in combination with other respiratory infections.
Zanobetti and Schwartz (2006, 090195) examined pneumonia HAs (ICD-9 480-487) in Boston, MA,
for the years 1995-1999 for individuals ages 65 yr and older, using a time-stratified case-crossover
analysis. The authors reported an increase in pneumonia HAs at lag 0 of 5.4% (95% CI: 1.2-10.0)
per 0.5 ppm increase in 24-h avg CO. While Zanobetti and Schwartz (2006, 090195) did not report
multipollutant results, they suggested that CO was most likely acting as a marker for traffic-related
pollutants because CO was highly correlated with both BC (r = 0.80) and NO2 (r = 0.67) and
moderately correlated with PM25 (r = 0.52). Instead of examining the effect of CO on pneumonia
HAs separately, as was done by Zanobetti and Schwartz (2006, 090195). Lin et al. (2005, 087828)
presented results for the overall effect of CO on respiratory infection HAs (ICD-9: 464, 466, 480-
487). In this analysis, Lin et al. (2005, 087828) examined the potential increase in respiratory HAs in
children <15 yr of age in Toronto, Canada, for 1998-2001, using a bidirectional case-crossover
approach. The authors reported elevated estimates for boys (OR=1.17 [95% CI: 1.03-1.32] per
0.5 ppm increase in 24-h avg CO for a 6-day ma) while the estimate for girls was weaker and with
wider confidence intervals (OR=1.06 [95% CI: 0.91-1.23]). In multipollutant models with both PM2.5
and PMio_25 the CO risk estimates were slightly attenuated but remained positive (boys: 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 not provide
an explanation as to why the estimates were stronger for boys than for girls. It should be noted that
this study used a bidirectional case-crossover analysis, which, as discussed previously, may bias the
results in either direction (Levy et al., 2001, 017172).
January 2010 5-92
-------
Study
Burnett etal. (2001,093439)
Yanaetal. (2003, 055621)
Linn etal. (2000,002839)
Yanaetal. (2003, 055621)
Cakmak et al. (2006, 093272)
Slauahter et al. (2005, 073854)
Lin etal. (2003, 042549)
Lin etal. (2003, 042549)
Slaughter etal. (2005, 073854)
Slauahter et al. (2005, 073854)
Yanaetal. (2005, 0901 84)
Zanobetti & Schwartz (2006, 090195)
Karr etal. (2006,088751)
Lin etal. (2005, 087828)
Lin etal. (2005, 087828)
"Boys
bGirls
Location
Toronto, Can
Vancouver, Can
Los Angeles, CA
Vancouver, Can
10 Canadian Cities
Spokane, WA
Toronto, Can
Toronto, Can
Spokane, VW\
Spokane, WA
Vancouver, Can
Boston, MA
SCAB, CA
Toronto, Can
Toronto, Can
Age
<2
<3
30+
65+
All
All
6-1 2"
6-12°
All
15+
65+
65+
<1
<15"
<15b
Lag
0-1
1
0
1
2.8"
2
1
1
2
2
0-6
0
1
0-5
0-5
Effect Estimate (95% CI)
\ • All Respiratory
i — •-
•
'-•-
f
-•-
1 — • Asthma
if
i •
— i-u — COPD
1 •
i — • — Pneumonia
-*- Acute Bronchiolitis
1 • Respiratory Infection
1 .
i •
i i i i i i
0.9 1.0 1.1 1.2 1.3 1.4
°To determine the lag for the combined estimate across all 10 cities, Cakmak et al. (2006, 0932721
averaged the strongest associations from lags 0-5 days from each city.
Relative Risk / Odds Ratio
Figure 5-15. Summary of associations for 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.
Emergency Department Visits
Respiratory Disease
Peel et al. (2005, 056305) conducted a large single-city respiratory disease ED visit study in
Atlanta, GA, which included data from 31 hospitals for the time period 1993-2000. In this study,
results were reported for various respiratory-related visits (ICD-9 460-466, 477, 480-486, 491-493,
496, 786.09). In an all-ages analysis, the authors found a RR=1.01 (95% CI: 1.00-1.02) for all
respiratory disease ED visits for a 3-day avg (lag 0-2) per 1 ppm increase in 1-h max CO
concentration. Tolbert et al. (2007, 090316) expanded the time period used in the Peel et al. (2005,
056305) study to include ED visits through 2004 and reported similar results for all respiratory
disease ED visits (RR=1.013 [95% CI: 1.007-1.018] per 1 ppm increase in 1-h max CO). The CO
risk estimates from the Atlanta, GA, ED visits studies were attenuated when O3, NO2, or PM were
added to the model (results not presented quantitatively), which could potentially be explained by the
high correlations reported in Tolbert et al. (2007, 090316) between CO and NO2 (r = 0.70) and EC
(r = 0.66) and the moderate correlation with PM2.5 (r = 0.51). One additional ED-visits study that
also examined respiratory disease (Slaughter et ai., 2005, 073854) presented essentially null results
at lag 1 and 2 but found similar results to Peel et al. (2005, 056305) and Tolbert et al. (2007, 090316)
at lag 3 (RR=1.02 [95% CI: 1.00-1.03] per 0.5 ppm increase in 24-h avg CO). Slaughter et al. (2005,
073854) reported a weak to moderate correlation between CO and various PM size fractions but did
not report the correlation between CO and gaseous pollutants, limiting the comparison of this study
with Peel et al. (2005, 056305) and Tolbert et al. (2007, 090316).
Asthma
The association between short-term exposure to CO and asthma ED visits (ICD-9 493, 786.09)
was also examined in Atlanta, GA, by Peel et al. (2005, 056305). In this study, the authors reported
1 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).
January 2010
5-93
-------
results from distributed lag models including lags 0-13 in addition to a ma of lags 0, 1, and 2 (lag
0-2) for specific respiratory outcomes (e.g., asthma). Effect estimates from the distributed lag models
were stronger than those produced from models that used 3-day ma CO concentrations (RR = 1.01
[95% CI: 0.99-1.02] for lags 0-2 compared to RR=1.08 [95% CI: 1.05-1.11] for an unconstrained
distributed lag of 0-13 for a 1 ppm increase in 1-h max CO). These results demonstrated the potential
effect of CO exposures up to 13 days prior to an asthma ED visit. Estimates were stronger for
pediatric ED visits (ages 2-18 yr) (RR=1.02 [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 ages (Peel et al., 2005, 056305). Slaughter et al. (2005,
073854). which also examined ED visits for Spokane (1995-2001), reported an increase in asthma
ED visits for all ages for CO at lag 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 examined (lags 1 and 2). The results from Ito et al. (2007,
156594) also provide evidence of increased ED visits for asthma (ICD-9 493) for all ages in New
York City for 1999-2002. Using three different models that adjusted for weather variables via
different degrees of smoothing and/or a different number of weather variables, the authors found that
CO effect estimates remained elevated in both an all-year analysis and in analyses stratified by warm
(i.e., April to August) and cold (i.e., November to March) months. Using Model C, which adjusted
for temporal trends using 8 degrees of freedom (df) and included variables to adjust for weather and
day of the week, an all-year RR of 1.03 (95% CI: 1.01-1.06) per 0.75 ppm increase in maximum 8-h
avg CO concentrations was reported. Ito et al. (2007, 156594) also examined copollutant models
using Model C but only during the warm season. In this model CO effect estimates were robust to
the inclusion of PM2.5 (RR = 1.06 [95% CI: 1.00-1.11]), O3 (RR=1.10 [95% CI: 1.05-1.15]), and SO2
(RR=1.04 [95% CI: 0.99, 1.09]) in the model, but the CO risk estimate was attenuated, resulting in a
negative effect estimate when including NO2 (RR=0.97 [95% CI: 0.92-1.03]) in the model.
Chronic Obstructive Pulmonary Disease
In the examination of the effect of short-term exposure to CO on COPD ED visits (ICD-9 491,
492, 496), Peel et al. (2005, 056305) reported elevated estimates for Atlanta, GA, for 1993-2000
(RR=1.03 [95% CI: 1.00-1.05] per 1 ppm increase in 1-h max CO for lag 0-2 ma) with similar
results for the distributed lag model (RR=1.03 [95% CI: 0.98-1.09). However, results from Slaughter
et al. (2005, 073854) from Spokane, WA, were consistent with a null or slightly 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 lag 1)
with similar results for lags 2 and 3.
Pneumonia and Other Respiratory Infections
Similar to the HA analysis conducted by Zanobetti and Schwartz (2006, 090195) discussed
above, Peel et al. (2005, 056305) examined the effect of CO on pneumonia separately (ICD-9:
480-486) but also included an analysis of upper respiratory infection (ICD-9: 460-466, 477) ED
visits for all ages in Atlanta, GA, during the years 1993-2000. The authors reported a weak estimate
for pneumonia for the 3-day ma (lag 0-2) (RR=1.01 [95% CI: 0.996-1.021] per 1 ppm increase in 1-h
max CO). However, when using an unconstrained distributed lag model (days 0-13), Peel et al.
(2005, 056305) observed evidence of an association (RR=1.045 [95% CI: 1.01-1.08]). An
examination of upper respiratory infection (URI) ED visits, the largest of the respiratory ED groups,
found slightly increased risk estimates for both the 3-day ma (lag 0-2) (RR=1.01 [95% CI:
1.00-1.02]) and the unconstrained distributed lag for days 0-13 (RR=1.07 [95% CI: 1.05-1.09]) per
1 ppm increase in 1-h max CO. In copollutant models, CO risk estimates were largely attenuated
when PMio, O3, or NO2 were included in the model (not reported quantitatively). Upon conducting
an age-stratified analysis, Peel et al. (2005, 056305) also found that infant (<1 yr of age) and
pediatric (ages 2-18 yr) URI ED visit CO risk estimates were substantially stronger than the all-age
risk estimates.
January 2010 5-94
-------
Study
Slaughter etal. (2005, 073854)
Peel etal. (2005,056305)
Tolbert etal. (2007. 09031 6)
Peel etal. (2005. 056305)
Peel etal. (2005.056305)
Peel etal. (2005.056305)
Slaughter etal. (2005. 073854)
Ito etal. (2007,156594)
Slaughter et al. (2005, 073854)
Peel etal. (2005, 056305)
Peel etal. (2005. 056305)
Peel etal. (2005.056305)
Peel etal. (2005.056305)
Peel etal. (2005, 056305)
Peel etal. (2005,056305)
Location Age
Spokane, WA All
Atlanta, GA All
Atlanta, GA All
Atlanta, GA 2-18
Atlanta, GA All
Atlanta, GA All
Spokane, WA All
NYCity, NY All
Spokane, WA 15+
Atlanta, GA All
Atlanta, GA All
Atlanta, GA All
Atlanta, GA All
Atlanta, GA All
Atlanta, GA All
Lag Effect Estimate (95% Cl)
3
0-2
0-2
0-2
0-2
0-13"
3
0-1
1 •
0-2
0-1 y —
0-2
0-1 3a
0-2
0-1 y
~*~ All Respiratory
r»-
=*=
• Asthma
=*=
•
•
*—
COPD
' •
"•" Pneumonia
L~*~ Respiratory Infection
•
"Unconstrained distributed lag 0.90 0.95 1.00 1.05 1.10 1.15
Relative Risk
Figure 5-16. Summary of associations for 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
Although HAs and ED visits are the two most well-studied measures of morbidity, a few
studies also examined the effect of CO on unscheduled physician visits. In a time-series study,
Villeneuve et al. (2006, 091179) examined the effect of CO on physician visits for allergic rhinitis in
individuals 65 yr and older in Toronto, Canada. Although quantitative results were only presented in
figures, upon observation it was evident that estimates were consistent with a null association for
lags 0-6 (Villeneuve et al., 2006, 091179). In an additional study, Sinclair et al. (2004, 088696)
reported results for urgent care visits for asthma and respiratory infections in a health maintenance
organization in Atlanta, GA; however, the study only reported statistically significant 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
Compared to other criteria air pollutants (e.g., O3 and PM), relatively few studies evaluated the
association between short-term exposure to ambient CO and HAs and ED visits for various
respiratory outcomes. Although evidence for consistent positive associations (Figure 5-15 and Figure
5-16) has been found across the studies evaluated, there remains uncertainty as to a biologically
plausible mechanism which could explain the association between CO exposure and respiratory-
related health effects. As observed in the preceding section, several authors suggest that the observed
associations are due to CO acting as an indicator of combustion-related pollution (e.g., traffic). The
interpretation of the associations observed in the studies evaluated is further complicated by the
moderate to high correlations reported between CO and other traffic-related pollutants such as NO2,
PM2.5, EC, or BC. Only a few studies examined potential confounding of CO risk estimates by
copollutants, and these studies found that CO risk estimates were generally robust to the inclusion of
O3, SO2, and PM in two-pollutant models. However, those studies that examined two-pollutant
models with NO2 found that CO risk estimates, although positive, were slightly attenuated.
January 2010 5-95
-------
5.5.2. Epidemiologic Studies with Long-Term Exposure
The 2000 CO AQCD (U.S. EPA, 2000, 000907) did not evaluate any studies that examined the
effect of long-term exposure to CO on respiratory health. The following section discusses those
studies that analyze the effect of long-term exposure to CO on pulmonary function, asthma/asthma
symptoms, and allergic rhinitis. Table 5-21 lists the studies evaluated in this section along with the
respiratory health outcomes examined and CO concentrations reported.
Table 5-21. Range of CO concentrations reported in key respiratory morbidity studies that examined
effects associated with long-term exposure to CO.
Study3
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
(Sample Size)
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)
*•«•> oSL
1989.2000 Pulmonary
11/2000-9/2001 Asthma symptoms
1999-2001 Asthma symptoms
Pulmonary
2000 function; Asthma
symptoms
9/1995-6/1996 Jg^
19g4 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 max 8-h
avg
Annual mean
of 1 -h avg
Annual mean
of 1 -h avg
Annual mean
of 1 -h avg
Annual mean
of0.5-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
NRb
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:
NRb
1.8
1.34
1.61
0.96
0.96
1.61
1.04
"The number of individuals included in the study population was only provided for those studies that included less than 1,000 participants.
bThis study only presented air quality data graphically.
5.5.2.1. Pulmonary Function
Mortimer et al. (2008, 122163) examined the effect of prenatal and lifetime exposures to air
pollutants on pulmonary function in 232 asthmatic children who resided in the San Joaquin Valley of
California. The strong temporal correlation between pollutants and pollutant metrics for different
time periods in the study area contributed to the inability to draw conclusions about the effect of
January 2010
5-96
-------
individual pollutant metrics on pulmonary function (Mortimer et al., 2008, 122163). The authors
used a newly developed Deletion/Substitution/Addition (DSA) algorithm "to identify which
pollutant metrics were most predictive of pulmonary function" (Mortimer et al., 2008, 122163). This
methodology uses an exploratory process to identify the best predictive model for each outcome of
interest. Focusing specifically on the exposure durations after birth, using this approach, Mortimer
et al. (2008, 122163) found that exposure to CO early in life, ages 0-3 yr, was negatively associated
with FEVi/FVC, resulting in an effect size of-2.5% per IQR increase in CO.1 Additional negative
associations were observed between exposure to CO during the first 6 yr of life and FEF25 (-6.7%)
and FEF25-75/FVC (-4.8%) in children diagnosed with asthma prior to 2 yr of age. Overall, Mortimer
et al. (2008, 122163) found that these effects were limited to subgroups, including African-
Americans and individuals diagnosed with asthma before the age of 2 yr. It must be noted that
research still needs to be conducted to validate the aforementioned results obtained using the DSA
algorithm and the subsequent calculation of effect estimates using GEE because the current model
could underestimate the uncertainty surrounding the associations reported (Mortimer et al., 2008,
122163). Although the authors did find associations between long-term exposure to CO and
decrements in pulmonary function, they also observed high correlations between CO and NO2,
which together are markers for pollutants generated by urban combustion sources (e.g., mobile
sources) (Mortimer et al., 2008, 122163).
Goss et al. (2004, 055624) also examined the effect of long-term exposure to CO on
pulmonary function in a cohort of cystic fibrosis patients >6 yr of age enrolled in the Cystic Fibrosis
National Patient Registry in 1999 and 2000. When examined cross-sectionally in 2000 using a
multiple linear regression model, the authors found no association between CO and a reduction in
FEVi. However, Goss et al. (2004, 055624) recognize that the CO results could be influenced by
measurement error and subsequently exposure misclassification.
5.5.2.2. Asthma and Asthma Symptoms
U.S.-based studies consistently reported no association between long-term exposure to CO and
asthma and asthma symptoms. Wilhelm et al. (2008, 191912) and Meng et al. (2007, 093275) both
examined the association between long-term exposure to air pollutants and asthma symptoms in
respondents to the 2001 California Health Interview Survey (CHIS) in populations consisting of
children (0-17 yr) and adults (> 18 yr), respectively, who resided in Los Angeles and San Diego
counties. Using a cross-sectional study design, Meng et al. (2007, 093275) found no association
between long-term exposure to CO and poorly controlled asthma in adults, while Wilhelm et al.
(2008, 191912) reported no associations between long-term exposure CO and asthma symptoms or
asthma HA and ED visits in children during the study period (i.e., 2000-2001). In an additional
U.S.-based study, Goss et al. (2004, 055624) found no association (OR=1.01 [95% CI: 0.92-1.10]
per 0.5 ppm increase in annual average CO concentrations) between long-term exposure to CO and
pulmonary exacerbations in a national cohort of individuals with cystic fibrosis >6 yr of age.
Among studies conducted in other countries, a study conducted in Germany (Hirsch et al.,
1999, 003537) and studies conducted in Taiwan (Guo et al., 1999, 010937: Hwang et al., 2005,
089454; Wang et al., 1999, 008105), all found positive associations between long-term exposure to
CO and asthma or asthma symptoms in populations ranging from 6 to 16 yr old. In these studies, the
authors addressed the observed associations differently. Guo et al. (1999, 010937) and Hwang et al.
(2005, 089454) both concluded that it is unlikely CO directly affects the respiratory system; Hirsch
et al. (1999, 003537) attributed the increase in the prevalence of cough and bronchitis to exposure to
traffic-related air pollutants (i.e., NO2, CO, and benzene); and Wang et al. (1999, 008105) did not
interpret the association observed between long-term exposure to CO and adolescent asthma. Only
Hwang et al. (2005, 089454) conducted a copollutant analysis and found that the asthma effects
observed were robust to the inclusion of PMi0, SO2 and O3 in the model. However, this study did not
include NOX in a copollutant model, which is notable because NOX was found to be highly
correlated with CO (r = 0.88).
1 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.
January 2010 5-97
-------
5.5.2.3. Respiratory Allergy and Other Allergic Responses
Allergy is a major contributor to asthma and upper respiratory symptoms; as a result, studies
have examined the effect of air pollutants on allergic outcomes. The studies evaluated that examined
the association between long-term exposure to CO and allergic outcomes were primarily conducted
outside of the U.S. and Canada. A multicity study conducted in 7 Spanish cities found that the annual
average concentration of CO was associated with a higher prevalence of allergic rhinitis,
rhinoconjunctivitis, and atopic eczema in 6- to 7-yr-old children (Arnedo-Pena et al., 2009, 190238).
NO2 was also examined and found to be positively associated with allergic rhinitis, but, unlike CO,
was negatively associated with eczema and rhinoconjunctivitis. It should be noted that in this data
set CO and NO2 concentrations were negatively correlated (r = -0.55). Additionally, SO2 was
positively associated with all allergic outcomes, while TSP matter was inversely associated with
rhinitis and rhinoconjunctivitis. Hwang et al. (2006, 088971) and Lee et al. (2003, 049201) both
examined the effect of long-term exposure to air pollutants on the prevalence of allergic rhinitis in a
population of schoolchildren in Taiwan. Both studies found an association between allergic rhinitis
prevalence and CO, but they also observed an association with NOX. As a result, although Hwang
et al. (2006, 088971) and Lee et al. (2003, 049201) observed an increase in the prevalence of allergic
rhinitis in response to an increase in long-term CO levels, they concluded that the combination of an
association being observed for both CO and NOX can be attributed to the complex mixture of traffic-
related pollutants and not necessarily CO alone. Although questions surround the associations
observed between long-term exposure to CO and allergic outcomes, the results are consistent with
those presented in a multicity study that examined the association between short-term exposure to
CO and allergic symptoms. Moon et al. (2009, 190297) observed associations between short-term
CO exposure and allergic symptoms in children in South Korea. However, allergic symptoms were
also associated with other pollutants, including PMi0, SO2, and NO2, and the study did not present
correlation coefficients to allow for further analysis of the results. It should be noted that
toxicological experiments suggest that endogenously produced CO may play an integral part in the
pathogenesis of allergic rhinitis, resulting in an additional potential pathway for CO-induced allergic
outcomes (Shaoqing et al., 2008, 192384).
Allergic symptoms such as rhinitis are a direct result of allergic sensitization, which is
commonly measured by skin prick testing or IgE antibody measurement. Hirsch et al. (1999,
003537). in a single-city study conducted in Dresden, Germany, observed no associations between
annual average concentrations of CO, NO2, SO2, or O3 and allergy assessed by skin prick testing or
serum IgE measurement in schoolchildren. However, prenatal exposure to CO was associated with
allergic sensitization in a cohort of 6- to 11-year-old asthmatic children in California (Mortimer et
al., 2008, 187280). Skin prick tests indicated higher levels of sensitization to indoor and outdoor
allergens with an increase in CO exposure during the prenatal period; the association with
sensitization to outdoor allergens remained after adjustment for effect modifiers, copollutants, and
other potential confounders. Mortimer et al. (2008, 187280) also found that PMi0 exposure was
associated with sensitization to indoor allergens but was not significant after adjustment.
Additionally, despite strong correlations between CO and NO2, no associations were reported with
NO2. It should be noted that these results were produced using the DSA algorithm and, as discussed
previously, additional research is still needed to evaluate the use of this method in air pollution
epidemiology (Mortimer et al., 2008, 122163).
5.5.2.4. Summary of Associations between Long-Term Exposure to CO and
Respiratory Morbidity
To date, a limited number of studies have examined the potential association between long-
term exposure to CO and respiratory morbidity. Although studies have reported positive associations
for various respiratory outcomes, the limited evidence available, the new analytical methods
employed, and the lack of studies that examined potential confounders of the CO-respiratory
morbidity relationship, especially due to the high correlation between CO and other traffic-related
pollutants, makes it difficult to attribute the associations observed to CO independent of other air
pollutants.
January 2010 5-98
-------
5.5.3. Controlled Human Exposure Studies
Human clinical studies provide very little and inconsistent evidence of changes in pulmonary
function following exposure to CO. In one older study, Chevalier et al. (1966, 010641) observed a
significant decrease in total lung capacity following a short-term exposure to 5,000 ppm resulting in
a COHb level of 4%. However, a similar study conducted at a higher CO concentration resulting in
COHb levels of 17-19% found no CO-induced changes in lung volume or mechanics (Fisher et al.,
1969, 012381). The 2000 CO AQCD (U.S. EPA, 2000, 000907) reported no evidence of CO-induced
changes in exercise ventilation at COHb levels <15% during submaximal exercise (Koike et al.,
1991, 013500). In two recent human clinical studies, exposure to CO (COHb ~ 10%) was not found
to significantly affect resting pulmonary ventilation compared with exposure to clean air under either
hypoxic or hyperoxic exposure conditions (Ren et al., 2001, 193850; Vesely et al., 2004, 194000).
The results of these studies demonstrate that the hypoxia- and CO2-induced increases in pulmonary
ventilation are not affected by CO. One recent study evaluated the potential anti-inflammatory
effects of controlled exposures to CO in the airways of 19 individuals with COPD (Bathoorn et al.,
2007, 193963). Subjects were exposed to both CO at concentrations of 100-125 ppm as well as room
air for 2 h on each of 4 consecutive days. The authors reported a small decrease in sputum
eosinophils, as well as a slight increase in the provocative concentration of methacholine required to
cause a 20% reduction in FEVi following exposure to CO. 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. A similar study found no evidence of systemic anti-inflammatory effects following
exposure to higher CO concentrations (500 ppm for 1 h) in a group of healthy adults (Mayr et al.,
2005, 193984).
5.5.4. lexicological Studies
As discussed in Section 5.2.5, the work of Thorn, Ischiropoulos and colleagues (Ischiropoulos
et al., 1996, 079491: Thorn and Ischiropoulos, 1997, 085644: Thorn et al., 1997, 084337: Thorn et
al., 1999, 016753: Thorn et al., 1999, 016757) focused on CO-mediated displacement of NO from
heme-binding sites. Although the concentrations of CO used in many of their studies were far higher
than ambient levels, some of this research involved more environmentally-relevant CO levels. In one
study (Thorn et al., 1999, 016757), 1-h exposure of rats to 50 ppm CO resulted in increased lung
capillary leakage 18 h later. Increased NO was observed in the lungs by electron paramagnetic
resonance during 1-h exposure to 100 ppm CO and was accompanied by increases in H2O2 and
nitrotyrosine. All of these effects were blocked by inhibition of NOS. These results, which were
partially discussed in the 2000 CO AQCD (U.S. EPA, 2000, 000907). demonstrate the potential for
exogenous CO to interact with NO-mediated pathways and to lead to pathophysiological effects in
the lung.
Recent work by Ohio et al. (2008, 096321) showed a disruption of cellular iron homeostasis
following exposure to a low level of CO (50 ppm for 24 h) in rats. In lungs of inhalation-exposed
rats, non-heme iron was significantly reduced, while lavagable iron was increased dramatically,
suggesting an active removal of cellular iron. Lavagable ferritin was also increased following the CO
exposure. Concurrently, liver iron levels increased, implying that the anatomical distribution of iron
stores may significantly shift during/after CO exposures. These investigators were able to replicate
the effect of loss of cellular iron in an in vitro model of cultured BEAS-2B cells and reported
statistically significant effects at 10 ppm CO and an apparent maximal effect at 50 ppm CO
(concentrations up to 500 ppm did not significantly enhance the iron loss beyond 50 ppm). Similar
responses were observed for cellular ferritin. Both enhancement of iron removal and diminished iron
uptake were noted in CO-exposed cells. Furthermore, decreased oxidative stress, mediator release
and proliferation were noted in respiratory cells. These effects were reversible with a recovery period
in fresh air. Interestingly, the in vivo exposure to CO induced mild but significant neutrophilia in the
lungs compared to air-exposed rats. This finding is contrary to the concept that CO acts as an anti-
inflammatory agent; however, with alterations in iron handling several potential pathways could be
initiated to recruit inflammatory cells into airways. The authors pointed out that while CO derived
from HO activity may have an important role in iron regulation, the nonspecific application of
exogenous CO would have little capacity to discriminate between excessive and/or inappropriate
iron which catalyzes oxidative stress and iron which may be required for normal homeostasis.
January 2010 5-99
-------
A chronic inhalation study by Sorhaug et al. (2006, 180414) demonstrated no alterations in
lung morphology in Wistar rats exposed to 200 ppm CO for 72 wk. COHb levels were reported to be
14.7%, and morphological changes were noted in the heart as described in Section 5.2.5.2.
A recent study by Carraway et al. (2002, 026018) involved continuous exposure of rats to HH
(380 torr) with or without co-exposure to CO (50 ppm) for up to 21 days. The focus of this study was
on remodeling of the pulmonary vasculature. While the addition of CO to HH did not alter the
thickness or diameter of vessels in the lung, there was a significant increase in the number of small
(<50 um) diameter vessels compared to control, HH-only, and CO-only exposures. Despite the
greater number of vessels, the overall pulmonary vascular resistance was increased in the combined
CO + HH exposure, which the authors attribute to enhancement of muscular arterioles and p-actin.
One new study found an association between increased endogenous CO and the development
of allergic rhinitis (Shaoqing et al., 2008, 192384). In this model, guinea pigs which were sensitized
and challenged with ovalbumin exhibited high immunoreactivity of HO-1 in the nasal mucosa and a
more than doubling of blood COHb levels (measured by gas chromatography). It is not known
whether the observed increase in endogenous CO resulting from ovalbumin-mediated
inflammation/oxidative stress plays a role in the development of allergic rhinitis but suggests a
potential mechanism by which exogenous CO could impact an allergic phenotype.
In summary, one older study (Thorn et al., 1999, 016757) and two new studies (Carraway et
al., 2002, 026018: Ohio et al., 2008, 096321) demonstrated effects of 50-100 ppm CO on the lung.
Responses included an increase in alveolar capillary permeability, disrupted iron homeostasis, mild
pulmonary inflammation, and an exacerbation of pulmonary vascular remodeling elicited by HH.
These results should be considered in view of the potential for inhaled CO to interact directly with
lung epithelial cells and resident macrophages. However, a chronic study involving 200 ppm CO
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
New epidemiologic studies, supported by the body of literature summarized in the 2000 CO
AQCD (U.S. EPA, 2000, 000907). provide evidence of positive associations between short-term
exposure to CO and respiratory-related outcomes including pulmonary function, respiratory
symptoms, medication use, HAs, and ED visits. The majority of the studies evaluated did not
conduct extensive analyses to examine the potential influence of model selection or effect modifiers
on the association between CO and respiratory morbidity. A limited number of studies examined the
potential confounding effects of copollutants on CO risk estimates and found that CO risk estimates
were generally robust to the inclusion of O3, SO2, and PM in two-pollutant models but were slightly
attenuated in models with NO2. However, the limited amount of evidence from studies that examined
the effect of gaseous pollutants on CO-respiratory morbidity risk estimates in two-pollutant models,
specifically NO2, has contributed to the inability to disentangle the effects attributed to CO from the
larger complex air pollution mix (particularly motor vehicle emissions), and this limits interpretation
of the results observed in the epidemiologic studies evaluated. A key uncertainty in interpreting the
epidemiologic studies evaluated is the biological mechanism(s) that could explain the effect of CO
on respiratory health. Animal toxicological studies, however, provide some evidence that short-term
exposure to CO (50-100 ppm) can cause oxidative injury and inflammation and alter pulmonary
vascular remodeling. Controlled human exposure studies have not extensively examined the effect of
short-term exposure to CO on respiratory morbidity, with a very limited number of studies reporting
inconsistent effects of CO on pulmonary function. Although these controlled human exposure studies
do not provide evidence to support CO-related respiratory health effects, epidemiologic studies show
positive associations for CO-induced lung-related outcomes and animal toxicological studies
demonstrate the potential for an underlying biological mechanism, which together provide evidence
that is suggestive of a causal relationship between relevant short-term exposures to CO
and respiratory morbidity.
January 2010 5-100
-------
5.5.5.2. Long-Term Exposure to CO
Currently, only a few studies have been conducted that examine the association between long-
term exposure to CO and respiratory morbidity including allergy. Although some studies did observe
associations between long-term exposure to CO and respiratory health outcomes, key uncertainties
still exist. These uncertainties include: the lack of replication and validation studies to evaluate new
methodologies (i.e., Deletion/Substitution/Addition (DSA) algorithm) that have been used to
examine the association between long-term exposure to CO and respiratory health effects; whether
the respiratory health effects observed in response to long-term exposure to CO can be explained by
the proposed biological mechanisms; and the lack of copollutant analyses to disentangle the
respiratory effects associated with CO due to its high correlation with NO2 and other combustion-
related pollutants. Overall, the evidence available is inadequate to Conclude that 3 C3USal
relationship exists between relevant long-term exposures to CO and respiratory
morbidity.
5.6. Mortality
5.6.1. Epidemiologic Studies with Short-Term Exposure to CO
Epidemiologic studies have traditionally focused on mortality effects associated with exposure
to PM and O3, resulting in a limited number of studies that have conducted extended analysis to
examine the potential influence of model selection, effect modifiers, or confounders on the
association between CO and mortality. This has contributed to the inability to formulate a clear
understanding of the association between short-term exposure to CO and mortality. This section
summarizes the main findings of the 2000 CO AQCD (U.S. EPA, 2000, 000907) and evaluates the
newly available information on the relationship between short-term exposure to CO and daily
mortality in an effort to disentangle the CO-mortality effect from those effects attributed to other
criteria air pollutants.
5.6.1.1. Summary of Findings from 2000 CO AQCD
The 2000 CO AQCD (U.S. EPA, 2000, 000907) examined the association between short-term
exposure to CO and mortality through the analysis of primarily single-city time-series studies, with
additional evidence from one multicity study which included 11 Canadian cities. While the results
presented by these studies did provide suggestive evidence that an association exists between CO
and mortality, the AQCD concluded that inadequate evidence existed to infer a causal association
between mortality and short-term exposure to ambient concentrations of CO. Multiple uncertainties
were identified in the epidemiologic literature that contributed to this conclusion, which were
discussed in Section 5.2.1.
The majority of the recent time-series mortality studies, as mentioned previously, have not
extensively examined the CO-mortality relationship. As such, CO has usually been considered as one
of the potential confounding copollutants in air pollution epidemiologic studies. Given the limitation
that most of these studies were not conducted to examine CO, the goal of this review is to evaluate
the CO-mortality association and specifically the consistency of associations across studies, along
with evidence of confounding and effect modification.
5.6.1.2. Multicity Studies
The following sections evaluate the recent literature that examined the association between
short-term exposure to CO and mortality, and, in addition, discuss newly available information with
regard to the issues specific to CO mentioned above. This evaluation focuses primarily on multicity
studies because they provide a more representative sample of potential CO-related mortality effects
and especially useful information by analyzing data from multiple cities using a consistent method,
January 2010 5-101
-------
and thus avoiding potential publication bias.1 Table 5-22 details the multicity studies evaluated along
with the mean CO concentrations reported in each study.
Table 5-22. Range of CO concentrations reported in multicity studies that examine mortality effects
associated with short-term exposure to CO.
Study
Dominici et al. (2003, 056116; 2005,
087912): Reanalysis of Samet et al. (2000,
156939)
Burnett etal. (2004. 086247)
Samoli et al. (2007. 098420)"
Location
82 US cities3
(NMMAPS)
12 Canadian cities
19 European cities
(APHEA2)
Years
1987-1994
1981-1999
1 990-1 997C
Averaging
Time
24-h avg
24-h avg
8-h max
Mean
Concentration
(ppm)
1.02
1.02
2.12
Range of Mean
Concentrations
Across Cities (ppm)
Baton Rouge = 0.43
Spokane = 2. 19
Winnipeg = 0.58
Toronto = 1 .31
Basel = 0.52
Athens = 5.3
"The study actually consisted of 90 U.S. cities, but only 82 had CO data.
bThis 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.
'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
The time-series analysis of the 90 largest U.S. cities (82 cities for CO) in the National
Morbidity, Mortality, and Air Pollution Study (NMMAPS) (Dominici et al., 2003, 056116; Dominici
et al., 2005, 087912; a reanalysis of Samet et al., 2000, 156939) is by far the largest multicity study
conducted to date to investigate the mortality effects of air pollution; however, the study primarily
focused on PMi0. The range in 24-h avg CO concentrations in a subset of the largest 20 cities (by
population size) was 0.66 ppm (Detroit, MI) to 2.04 ppm (New York City). The analysis in the
original report used GAM with default convergence criteria. In response to the bias observed in the
estimates generated using GAM models with default convergence criteria (Dominici et al., 2002,
030458). Dominici et al. (2003, 056116; 2005, 087912)(reanalysis of Samet et al. (2000, 156939)
conducted a reanalysis of the original data using GAM with stringent convergence criteria as well as
GLM.
Focusing on the results obtained using GLM, PMi0 and O3 (in summer) appeared to be more
strongly associated with mortality than the other gaseous pollutants. The authors stated that the
results did not indicate associations between CO, SO2, or NO2, and total (nonaccidental) mortality.
However, as with PMi0, the gaseous pollutants CO, SO2, and NO2 each showed the strongest
association at a 1-day lag (for O3, a 0-day lag). Figure 5-17 presents the total mortality risk estimates
for CO from Dominici et al. (2003, 056116). The authors found a mortality risk estimate of 0.23%
(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 model.
The inclusion of PMi0 or PM10 and O3 in the model did not reduce CO risk estimates. However, the
confidence intervals were wider in the multipollutant models; however, this could be attributed to:
(1) PMio data in many of the cities being collected every 6th day as opposed to daily data for
gaseous pollutants; and (2) O3 being collected in some cities only during warm months. The addition
of NO2 (along with PMi0) to the model resulted in a reduced CO risk estimate. Some caution is
required when interpreting this apparent reduction because a smaller number of cities (57 cities2)
were available for the CO multipollutant analysis with PM10 and NO2 compared to the single-
pollutant CO analysis (82 cities). However, most of the 32 cities that were excluded due to the lack
of NO2 data were some of the least populated cities. Thus, the difference in the number of cities in
the multi- and single-pollutant analyses is unlikely to be the underlying cause for the reduction in the
1 To compare studies in this section that used different averaging times, effects estimates were standardized to the following: 0.5 ppm for
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.
2 One city was excluded from the multipollutant analysis because it contained NO2 data but did not contain CO data.
January 2010
5-102
-------
CO risk estimate in the CO multipollutant analysis with PM10 and NO2. In comparison to the PM10
risk estimates which were not reduced in multipollutant models, the CO risk estimates from
multipollutant models indicate less consistent associations with mortality.
ij
-------
and PM was not presented. The results presented in this Canadian multicity study and NMMAPS are
similar in that the CO risk estimates appeared to be sensitive to the addition of NO2 in the regression
model. However, interpretation of these results requires some caution because: (1) NO2 tends to have
a more spatially uniform distribution within a city compared to CO; (2) CO and NO2 share common
sources (e.g., traffic); and (3) CO and NO2 are often moderately to highly correlated.
Air Pollution and Health: A European Approach
Most of the Air Pollution and Health: A European Approach (APHEA) analyses have focused
on the mortality effects of PM (PMi0 and BS), SO2, NO2, and O3, but not CO. In addition, some of
the analyses have not even considered CO as a potential confounder, such as the extended analysis
(APHEA2) of PM (Katsouyanni et al, 2001, 019008) and NO2. Gryparis et al. (2004, 057276) did
consider CO as a potential confounder in an analysis of O3 and found that the addition of CO
increased O3 mortality risk estimates both in the summer and winter, although the number of cities
included in the copollutant model were reduced from 21 to 19. However, the study did not present
CO risk estimates. Unlike other APHEA studies (or the NMMAPS and Canadian multicity studies),
the Samoli et al. (2007, 098420) analysis focused specifically on CO.
Samoli et al. (2007, 098420) investigated the effect of short-term exposure to CO on total
(nonaccidental) and cardiovascular mortality in 19 European cities participating in the APHEA2
project by using a two-stage analysis to examine city-specific effects and potential sources of
heterogeneity in CO-mortality risk estimates. The mean levels of the max 8-h avg CO concentration
in this study ranged from 0.52 ppm (Basel, Switzerland, and The Netherlands) to 5.3 ppm (Athens,
Greece). The max 8-h avg CO concentration across all cities in the APHEA2 study of 2.12 ppm is
higher than the estimated max 8-h avg CO concentrations reported for the U.S. cities examined in
Dominici et al. (2003, 056116; 2005, 087912) and the Canadian cities examined in Burnett et al.
(2004, 086247) of 1.53 ppm.'In APHEA cities, the correlation between CO and BS (r = 0.67-0.82)
was higher than the correlation between CO and PMi0 (r = 0.16-0.70) or CO and 1-h max NO2
(r = 0.03-0.68).
To examine the CO-mortality relationship, Samoli et al. (2007, 098420) conducted a time-
series analysis of individual cities following the revised APHEA2 protocol.2 The primary results
presented by the authors are from a sensitivity analysis that used two alternative methods to select
the extent of adjustment for temporal confounding. These methods consisted of: (1) confining the
extent of smoothing to 8 df/yr; and (2) selecting the appropriate extent of smoothing through
minimization of the absolute value of the sum of partial autocorrelation functions (PACF) of the
residuals, which resulted in the analysis using on average 5 df/yr for total (nonaccidental) mortality
and 4 df/yr for cardiovascular mortality. The authors also conducted copollutant analyses using
PMi0, BS, SO2, NO2, or O3 (1 h). In the second stage model, Samoli et al. (2007, 098420) examined
heterogeneity in CO risk estimates between cities by regressing risk estimates from individual cities
on potential effect modifiers including: (1) the air pollution level and mix in each city (i.e., mean
levels of pollutants, ratio PMi0/NO2); (2) the exposure (number of CO monitors, correlation between
monitors' measurements); (3) variables describing the health status of the population (e.g., crude
mortality rate);(4) the geographic area (northern, western, and central-eastern European cities); and
(5) climatic conditions (mean temperature and relative humidity levels).
Samoli et al. (2007, 098420) found that CO was associated with total (nonaccidental) and
cardiovascular mortality. The primary results represent the combined random effects estimate for a
0.75 ppm increase in max 8-h avg CO concentrations for the average of 0- and 1-day lag for total
(nonaccidental) mortality (1.03% [95% CI: 0.55-1.53]) and for cardiovascular mortality (1.08%
[95% CI: 0.25-1.90]). These results were obtained using PACF to choose the extent of adjustment for
temporal trends. Although the results obtained using PACF are insightful, the use of 8 df/yr would
have been more consistent with the NMMAPS model (7 df/yr) and would have allowed for a more
accurate comparison of the results between APHEA2 and NMMAPS. The corresponding risk
estimates obtained using the 8 df/yr model are 0.57% (95% CI: 0.23-0.91) for total (nonaccidental)
mortality and 0.70% (95% CI: 0.31-1.09) for cardiovascular mortality. In the sensitivity analysis,
1 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.
January 2010 5-104
-------
Samoli et al. (2007, 098420) used 8 or 12 df/yr to adjust for temporal confounding. Both approaches
resulted in similar risk estimates, but using PACF to choose the extent of smoothing separately in
each city generally resulted in larger CO risk estimates (by -50-80%). This can be attributed to the
smaller number of df/yr used in the model (on average 5 df/yr for total [nonaccidental] mortality and
4 df/yr for cardiovascular mortality), which increases the magnitude of the effect and the amount of
observed heterogeneity (Samoli et al., 2007, 098420).
During the examination of the results obtained from the copollutant models, the authors noted
that there was indication of confounding of CO risk estimates by BS and NO2 but not PMi0. These
results are consistent with CO, BS, and NO2 being part of the traffic-pollution mixture, and PMi0
likely including secondary aerosols that do not correlate well with traffic-derived pollution. The risk
estimates from the model using 8 df/yr that included NO2 were 0.26% (-0.09 to 0.61) for total
(nonaccidental) mortality and 0.37% (-0.05 to 0.80) for cardiovascular mortality. Thus, the inclusion
of NO2 in the model nearly halved the CO risk estimates (whereas the NO2 risk estimate was not
sensitive to the inclusion of CO in the model). CO risk estimates were reduced by a similar
magnitude when including BS in the model. Overall, the sensitivity of CO risk estimates to the
inclusion of NO2 in the model is consistent with the results presented in NMMAPS (Dominici et al.,
2003, 056116) and the Canadian multicity study (Burnett et al., 2004, 086247).
In the second-stage model, Samoli et al. (2007, 098420) found that geographic region was the
most significant effect modifier, while the other effect modifiers (mentioned above) did not result in
strong associations. Effects were primarily found in western and southern European cities and were
larger in cities where the standardized mortality rate was lower. Earlier APHEA studies also reported
a regional pattern of air pollution associations for BS and SO2 and found that western cities showed
stronger associations than eastern cities. However, the heterogeneity in CO risk estimates by
geographic region does not provide specific information to evaluate the CO-mortality association.
An ancillary analysis conducted by Samoli et al. (2007, 098420) examined the possible
presence of a CO threshold. The authors compared city-specific models to the threshold model,
which consisted of thresholds at 0.5 mg/m3 (0.43 ppm) increments. Samoli et al. (2007, 098420) then
computed the deviance between the two models and summed the deviances for a given threshold
over all cities. While the minimum deviance suggested a potential threshold of 0.43 ppm (the lowest
threshold examined), the comparison with the linear no-threshold model indicated very weak
evidence (p-value >0.9) for a threshold. However, determining the presence of a threshold at the very
low range of CO concentrations (i.e., 0.43 ppm) in this data set is challenging because in 7 of the
19 European cities examined, the lowest 10% of the CO distribution was at or above 2 mg/m3
(1.74 ppm).
In summary, the APHEA2 analysis of CO in 19 cities found an association between CO and
total (nonaccidental) and cardiovascular mortality in single-pollutant models, but the associations
were substantially reduced when NO2 or BS was included in copollutant models. The evidence for
potential confounding of CO risk estimates by NO2 is consistent with the findings from NMMAPS
and the Canadian 12-city study. In addition, Samoli et al. (2007, 098420) found that geographic
region was a potential effect modifier, but such geographic heterogeneity is not specific to CO, based
on previously conducted APHEA studies. Finally, examination of the CO concentration-response
relationship found very weak evidence of a CO threshold, which requires further investigation.
Other European Multicity Studies
An additional European multicity study was conducted by Biggeri et al. (2005, 087395) in
eight Italian cities. The authors examined the effect of short-term exposure to CO on mortality in
single-pollutant models using a time-series approach. In this analysis, all of the pollutants showed
positive associations with the mortality endpoints examined. However, copollutant models were not
examined, and the correlations among the pollutants were not presented; therefore, it is unclear if the
observed associations are shared or confounded.
Summary of Multicity Studies
In summary, the mortality risk estimates from single-pollutant models are comparable for the
NMMAPS and Canadian 12-city studies, 0.23% and 0.33%, respectively, with the estimate from the
January 2010 5-105
-------
APHEA2 study being slightly larger (0.57%) (Figure 5-18). In both the NMMAPS and Canadian
studies, a 1-day lag showed the strongest association; however, the APHEA2 study used an a priori
exposure window (i.e., average of 0- and 1-day lags), which has been found to be the exposure
window most strongly associated with mortality in PM analyses.
The APHEA2 risk estimates presented in Figure 5-18 are from a model that used a fixed
amount of smoothing to adjust for temporal confounding (8 df/yr), which is similar to that used in
the NMMAPS study (7 df/yr). However, the APHEA2 sensitivity analysis suggested an approximate
50-80% difference in CO risk estimates between the models that used 8 or 12 df/yr and the models
that used minimization of the absolute value of the sum of PACF of the residuals as a criterion to
choose the smoothing parameters. Thus, some model uncertainty likely influences the range of CO
risk estimates obtained from the studies evaluated.
The CO risk estimates from the aforementioned studies are also consistently sensitive to the
inclusion of NO2 in a copollutant model (0.11, 0.03, and 0.26%, for the NMMAPS, Canadian 12-city
study, and APHEA2, respectively). Thus, these results suggest confounding by NO2. However, this
interpretation is further complicated because as with CO, NO2 itself may be an indicator of
combustion sources, such as traffic.
Study
Pollutant
Effect Estimate (95% Cl)
Dominici et al. (2003, 0561 1 : 2005, 08791 2)
(reanalvsisofSametetal. (2000, 156939))
NMMAPS, lag 1
Burnett etal. (1998,086427)
12 Canadian cities, lag 1
Samoli etal. (2007,098420)
APHEA2 (19 European cities, lag 0-1)
CO Alone
CO + PM10
CO + PM10 and N02
CO Alone
CO + N02a
CO Alone
CO + PM10
CO + BS
C0 + N02
-0.4 -0.2 0
^ «« •!.-
I I I I I
0 0.2 0.4 0.6 0.8 1.0
Percent Increase
N02 is the average of 0-day, 1-day, and 2-day lags
Figure 5-18. Summary of percent increase in total (nonaccidental) mortality 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.
5.6.1.3. Meta-Analysis of All Criteria Pollutants
Stieb et al. (2002, 025205) reviewed the time-series mortality studies published between 1985
and 2000 and conducted a meta-analysis to estimate combined effects for PMi0, CO, NO2, O3, and
SO2. Because many of the studies reviewed in the 2000 analysis used GAM with default
convergence criteria, Stieb et al. (2003, 056908) updated the estimates from the meta-analysis by
separating the GAM versus non-GAM studies. In this meta-analysis, the authors also presented
separate combined estimates for single- and multipollutant models. Overall, there were more GAM
estimates than non-GAM estimates for all of the pollutants except SO2. For CO, 4 single-pollutant
model risk estimates were identified, resulting in a combined estimate of 3.18% (95% CI: 0.76-5.66)
per 0.5 ppm increase in 24-h avg CO, and only 1 multipollutant model risk estimate (0.00%
[95% CI: -1.71 to 1.74]) from the non-GAM studies. Thus, for CO, this study did not provide useful
meta-estimates because the number of studies that contributed to the combined estimates for CO was
small.
January 2010
5-106
-------
5.6.1.4. Single-City Studies
In addition to the multicity studies discussed above, there have also been several single-city
U.S.- and Canadian-based time-series mortality studies that examined CO. The single-city studies,
similar to the multicity studies, often focused on the PM-mortality association but also provided
additional information that is not available in the multicity studies. Because the sample size used in
each single-city study is small and subsequently results in wide confidence intervals, a quantitative
comparison of the results from single- and multicity studies is difficult. In addition, some studies do
not present CO results quantitatively, adding to the inability to adequately compare studies. Table
5-23 lists the single-city studies evaluated along with the mean CO concentrations reported in each
study.
Table 5-23. Range of CO concentrations reported in single-city studies that examine mortality
effects associated with short-term exposure to CO.
Study
De Leon et al. (2003, 055688)
Klemm et al. (2004, 056585)
Vedal et al. (2003, 039044)'
Villeneuve et al. (2003, 055051)
Goldberg etal. (2003,035202)
Location
New York, NY
Atlanta, GA
Vancouver, BC, Can
Vancouver, BC, Can
Montreal, Quebec, Can
Years
1985-1994
1998-2000
1994-1996
1986-1999
1984-1993
Averaging
Time
24-h avg
1 -h max
24-h avg
24-h avg
24-h avg
Mean Concentration
(ppm)
2.45
1.31
0.5
1.0
0.8
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
Hoek et al. (2000, 010350:2001,
016550): Reanalyzed by Hoek (2003, The Netherlands
042818)
1986-1994 24-h avg
Entire Country: 0.46 Max.
Entire Country: 2.6
Four Major Cities: 0.59 Four Major Cities: 4.6
3Study reported median CO concentrations.
Single-City Studies Conducted in the United States
De Leon et al. (2003, 055688) focused on the role of contributing respiratory diseases on the
association between air pollution (i.e., PMi0, O3, NO2, SO2, and CO) and primary nonrespiratory
mortality (circulatory and cancer) in New York City, NY, during the period 1985-1994. This study
only presented risk estimates graphically for each of the pollutants analyzed, except PM10. In single-
pollutant models, PMi0, CO, SO2, and NO2 all showed the same pattern of association with
circulatory mortality for individuals > 75 yr, indicating a larger risk of death in individuals with
contributing respiratory diseases compared to those without. In two-pollutant models, PMi0 and CO
risk estimates were reduced but each remained significantly positive.
Klemm et al. (2004, 056585) analyzed 15 air pollutants for their associations with mortality in
Atlanta, GA, for a 2-yr period starting in August 1998. These pollutants included PM25, PM10_2.5,
UFP surface area and counts, aerosol acidity, EC, OC, SO42", O3, CO, SO2, and NO2. This study
presented risk estimates using three levels of smoothing (quarterly, monthly, and biweekly knots) for
temporal trend adjustment and suggested that the risk estimates were rather sensitive to the extent of
smoothing. It should be noted that temporal smoothing using biweekly knots is a more aggressive
modeling approach than the degrees-of-freedom approach used by most studies. In the single-
pollutant models for nonaccidental mortality, the strongest association, which was also statistically
significant, was found for PM25. CO, SO42", and PM10_25 also showed positive associations with
nonaccidental mortality (CO: Quarterly knots and Monthly Knots (3 = 0.00002 [SE = 0.00001];
Biweekly knots (3 = 0.00001 [SE = 0.00002]). However, CO was significantly associated with
circulatory mortality in older adults (> 65 yr), and these associations remained when PM2 5 was
included in the model (results were presented graphically).
January 2010
5-107
-------
Single-City Studies Conducted in Canada
Vedal et al. (2003, 039044) examined the association between short-term exposure to "low
levels" of air pollution (i.e., PMi0, O3, NO2, SO2, and CO) and daily mortality in Vancouver, British
Columbia, Canada, for the years 1994-1996. In this analysis, all of the risk estimates were presented
graphically; however, the results suggested that O3 in the summer and NO2 in the winter showed the
strongest associations with mortality. Vedal et al. (2003, 039044) found that CO was positively but
not significantly associated with mortality. Additionally, the association between short-term exposure
to NO2 and mortality was found to be consistent with the results from the Canadian multicity study
conducted by Burnett et al. (2004, 086247).
Villeneuve et al. (2003, 055051) also conducted an analysis using data from Vancouver,
Canada, using a cohort of 550,000 individuals whose vital status was ascertained between 1986 and
1999. In this study, PM2.5, PM10.2.5, TSP, CoH, PM10, SO42", O3, CO, SO2, and NO2 were examined
for their associations with all-cause (nonaccidental), cardiovascular, and respiratory mortality. When
examining the association between gaseous pollutants and all-cause (nonaccidental) mortality in this
data set, NO2 and SO2 showed the strongest associations, while the association between CO and all-
cause mortality were generally weaker than those for NO2 and SO2. For cardiovascular mortality,
SO2 risk estimates were smaller than those for NO2 or CO, while for respiratory mortality, SO2
showed the strongest associations. However, the wider confidence intervals for these categories and
the smaller daily counts make it difficult to assess CO associations with cause-specific mortality
outcomes.
Goldberg et al. (2003, 035202) contrasted associations between air pollution and mortality in
individuals with underlying CHF versus mortality in individuals who were identified as having CHF
1 yr prior to death based on information from the universal health insurance plan in Montreal,
Quebec, Canada, during the period 1984-1993. In this study, Goldberg et al. (2003, 035202)
examined associations between PM2 5, CoH, SO42", O3, CO, SO2, and NO2, and mortality. The
authors found no association between any of the air pollutants and mortality with underlying CHF.
However, Goldberg et al. (2003, 035202) found positive associations between air pollution and
mortality in individuals diagnosed with CHF 1 yr prior to death. Of the air pollutants examined,
CoH, NO2, and SO2 were most consistently associated with mortality for ages 65 yr and older, while
CO showed positive but weaker associations compared to these three pollutants.
Single-City Studies Conducted in Other Countries
Of the epidemiologic studies conducted in other countries that examine the association
between short-term exposure to CO and mortality, only those studies conducted in European
countries that have CO levels comparable to the U.S. were evaluated. However, because Samoli et
al. (2007, 098420) conducted a multicity study of European cities that focused on short-term
exposure to CO, there are only a few single-city studies that provide additional information,
specifically those studies conducted in The Netherlands. The Netherland studies were evaluated
because they provide risk estimates for multiple pollutants and cause-specific mortality and
consisted of relatively large sample sizes (i.e., the mortality time-series of the entire country was
analyzed).
Hoek et al. (2000, 010350) (reanalyzed by Hoek) (2003, 042818) examined associations
between air pollution and all-cause (nonaccidental), cardiovascular, COPD, and pneumonia deaths in
the entire Netherlands, the four major cities combined, and the entire country minus the four major
cities for the period 1986-1994. The air pollutants analyzed included BS, PMio, O3, NO2, SO2, CO,
SO42~ and NOV. In the single-pollutant models, all of the pollutants were significantly associated
with nonaccidental mortality at lag 1-day and 0-6 days when using the entire Netherlands data set. In
the two-pollutant model, CO risk estimates were reduced to null when PMi0, BS, SO42" and NO3~
were included in the model, while the risk estimates for these copollutants remained significantly
positive. BS, CO, and NO2 were highly correlated (r > 0.85) in this data set, and the authors noted
"all these pollutants should be interpreted as indicators for motorized traffic emissions" (Hoek et al.,
2000, 010350). The authors found that O3 showed the most consistent and independent associations
with mortality and that the risk estimates for all of the pollutants were substantially higher in the
summer months than in the winter months. Pneumonia deaths showed the largest risk estimates for
most pollutants including CO. The result from the Hoek et al. (2000, 010350) study is somewhat in
January 2010 5-108
-------
contrast to the result from the Samoli et al. (2007, 098420) multicity study in that in the Hoek et al.
(2000, 010350) analysis, CO was more sensitive to the addition of PM indices in copollutant models.
This may be due to the high correlation between CO and PM indices in The Netherlands.
Hoek et al. (2001, 016550) (reanalysis by Hoek) (2003, 042818) analyzed The Netherlands
data using more specific cardiovascular causes of death: MI and other IHD, arrhythmia, heart failure,
cerebrovascular mortality, and embolism/thrombosis. In this analysis, the authors analyzed O3, BS,
PM10, CO, SO2, and NO2 in only single-pollutant models. For all of the pollutants, risk estimates
were larger for arrhythmia, heart failure, and cerebrovascular mortality than for the combined
cardiovascular mortality outcome. Thus, the results suggested larger impacts of air pollution on more
specific cardiovascular causes; however, it is difficult to distinguish the effects of each pollutant
from the larger air-pollution mixture.
5.6.1.5. Summary of Mortality and Short-Term Exposure to CO
The recently available multicity studies, which consist of larger sample sizes, along with the
single-city studies evaluated reported associations that are generally consistent with the results of the
studies evaluated in the 2000 CO AQCD (U.S. EPA, 2000, 000907). However, to date the majority
of the literature has not conducted extensive analyses to examine the potential influence of model
selection, effect modifiers, or confounders on the association between CO and mortality.
The multicity studies reported comparable CO mortality risk estimates for total
(nonaccidental) mortality, with the APHEA2 European multicity study (Samoli et al., 2007, 098420)
showing slightly higher estimates for cardiovascular mortality in single-pollutant models. However,
when examining potential confounding by copollutants these studies consistently showed that,
although CO mortality risk estimates remained positive, they were reduced when NO2 was included
in the model. But this observation may not be confounding in the usual sense in that NO2 may also
be an indicator of other pollutants or pollution sources (e.g., traffic).
Of the studies evaluated, only the APHEA2 study focused specifically on the CO-mortality
association (Samoli et al., 2007, 098420). and, in the process, examined: (1) model sensitivity; (2)
the CO-mortality concentration-response (C-R) relationship; and (3) potential effect modifiers of CO
mortality risk estimates. The sensitivity analysis indicated an approximate 50-80% difference in CO
risk estimates from a reasonable range of alternative models, which suggests that some model
uncertainty likely influences the range of CO mortality risk estimates obtained in the studies
evaluated. The examination of the CO-mortality concentration-response relationship found very
weak evidence for a CO threshold at 0.5 mg/m (0.43 ppm). Finally, when examining a variety of
city-specific variables to identify potential effect modifiers of the CO-mortality relationship, the
APHEA2 study found that geographic region explained most of the heterogeneity in CO mortality
risk estimates.
The results from the single-city studies are generally consistent with the multicity studies in
that some evidence of a positive association was found for mortality upon short-term exposure to
CO. However, the CO-mortality associations were often but not always attenuated when copollutants
were included in the regression models. In addition, limited evidence was available to identify cause-
specific mortality outcomes (e.g., cardiovascular causes of death) associated with short-term
exposure to CO.
The evidence from the recent multi- and single-city studies suggests that an association
between short-term exposure to CO and mortality exists. But limited evidence is available to
evaluate cause-specific mortality outcomes associated with CO exposure. In addition, the attenuation
of CO risk estimates which was often observed in copollutant models contributes to the uncertainty
as to whether CO is acting alone or as an indicator for other combustion-related pollutants. Overall,
the epidemioiogic evidence is suggestive of a causal relationship between relevant short-
term exposures to CO and mortality.
5.6.2. Epidemioiogic Studies with Long-Term Exposure to CO
The 2000 CO AQCD (U.S. EPA, 2000, 000907) did not evaluate the association between long-
term exposure to CO and mortality because there were no studies at the time that examined this
relationship. Since then there have been several new studies that examined the association between
long-term exposure to CO and mortality. It should be noted, however, that these studies focused
January 2010 5-109
-------
primarily on PM, and that CO was only considered in these studies as a potential confounder.
Therefore, the information available from these new long-term exposure studies is somewhat
limited, especially in comparison to that for PM. Table 5-24 lists the U.S.-based studies evaluated
that examined the association between long-term exposure to CO and mortality, along with the mean
CO concentrations reported in each study.
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 etal. (2000.004087)
Lipfert etal. (2006.088756)
Lipfert etal. (2006.088218)
Lipfert and Morris (2002, 019217)
Location
107 US cities
1980: 113 US cities
1982-1 998: 122 US
cities
108 US cities
36 US cities
US
US
us
1960-1 969: 44 US
counties
1970-1 974: 206 US
counties
1979-1 981: 272 US
1989-1 991: 246 US
counties
1995-1997:261 US
counties
Years
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 EeSS?
(ppm)
1.56 Maximum: 3.95
1980.1.7 njp^
75th: 2.13
. RR 90th: 2.58
1'68 95th: 3.05
Maximum: 3.95
NR NR
1960-1974
50th: 9.31
Maximum: 35.3
1975-1981
1960-1974:10.82 50th: 7.04
1975-1981:7.64 Maximum: 22.4
1982-1988:3.42 1982-1988
1989-1996:2.36 50th: 3.33
Maximum: 15.20
1989-1996
50th: 2.30
Maximum: 7. 10
. c, Maximum: 6.7
1.00
•tnon -innc n A NR
1960-1969:13.8
1970-1974:9.64
1979-1981:5.90 NR
1989-1991:2.69
1995-1997:1.72
5.6.2.1. U.S. Cohort Studies
American Cancer Society Cohort Studies
Pope et al. (1995, 045159) investigated associations between long-term exposure to PM and
mortality outcomes in the ACS cohort. In this study, ambient air pollution data from 151 U.S.
metropolitan areas in 1981 were linked with individual risk factors in 552,138 adults who resided in
these areas when enrolled in the prospective study in 1982; death outcomes were ascertained through
1989. PM2.s and SO42~ were associated with total (nonaccidental), cardiopulmonary, and lung cancer
mortality, but not with mortality for all other causes (i.e., nonaccidental minus cardiopulmonary and
lung cancer). Gaseous pollutants were not analyzed in Pope et al. (1995, 045159). Jerrett et al. (2003,
January 2010
5-110
-------
087380). using data from Krewski et al. (2000, 012281). conducted an extensive sensitivity analysis
of the Pope et al. (1995, 045159) ACS data, augmented with additional gaseous pollutants data. Due
to the smaller number of CO monitors available compared to SO42", the number of metropolitan
statistical areas (MSAs) included in the CO analysis were reduced (from 151 with SO42~to 107). The
mean annual CO concentrations in these MSAs ranged from 0.19 to 3.95 ppm. CO was weakly
negatively correlated with SO42" (r = -0.07). Among the gaseous pollutants examined (CO, NO2, O3,
SO2), only SO2 showed positive associations with mortality, and, in addition, was the only
copollutant that reduced SO42" risk estimates. For CO, the relative risk estimates for total
(nonaccidental) mortality in single- and copollutant models with SO42" was 0.99 (95% CI: 0.96-1.01)
and 0.98 (95% CI: 0.96-1.01), respectively, per 0.5 ppm increase in mean annual average CO
concentrations.
Pope et al. (2002, 024689) conducted an extended analysis of the ACS cohort with double the
follow-up time (to 1998) and triple the number of deaths compared to the original Pope et al. (2002,
024689) study. In addition to PM25, data for all of the gaseous pollutants were retrieved for the
extended period and analyzed for their associations with mortality-specific outcomes. As in the 1995
analysis, the air-pollution exposure estimates were based on the MSA-level averages. The authors
found that PM2 5 and SO42~ were both associated with all-cause, cardiopulmonary, and lung cancer
mortality.1 In this study, the CO analysis used two different data sets: the first data set consisted of
1980 data and 113 MSAs; while the second data set used averages of the years 1982-1998 and
122 MSAs. The authors found, when using the 1980 data, that CO was not associated (risk estimates
~ 1) (Figure 5-19) with all-cause, cardiopulmonary, lung cancer, or mortality for all other causes.
However, the analysis of the 1982-1998 data found that CO was negatively (and significantly)
associated with all-cause, cardio-pulmonary, and lung cancer mortality. It is unclear why significant
negative associations were observed when analyzing the 1982-1998 data, but evidence from other
mortality studies that examined the association between long-term exposure to CO and mortality do
not suggest that CO elicits a protective effect.
Krewski et al. (2009, 191193) further analyzed the ACS cohort by adding two additional years
of mortality data (total period 1982-2000). This study extended the range of the analysis to
incorporate sophisticated adjustment for bias and confounding as well as intra-urban analyses.
However, the CO analysis was limited to using (1) nationwide data; (2) only 1980 CO
concentrations; and (3) the standard Cox proportional hazards model. In addition to the death
categories examined in Pope et al. (2002, 024689). this analysis also examined IHD mortality. As
was the case with the Pope et al. (2002, 024689) analysis, Krewski et al. (2009, 191193) found that
1980 CO data was not associated with any of the mortality categories examined: all-cause mortality
HR=1.00 (95%CI: 0.99-1.01); cardio-pulmonary mortality, HR=1.00 (95% CI: 0.99-1.00); and IHD
mortality, HR=1.00 (95% CI: 0.99-1.01) per 0.5 ppm increase in CO.
Women's Health Initiative Cohort Study
Miller et al. (2007, 090130) studied 65,893 postmenopausal women between the ages of 50
and 79 yr without previous CVD in 36 U.S. metropolitan areas from 1994 to 1998. The authors
examined the association between one or more fatal or nonfatal cardiovascular events and air-
pollutant concentrations. Exposures to air pollution were estimated by assigning the year 2000 mean
concentration of air pollutants measured at the nearest monitor to the location of residence of each
subject on the basis of its five-digit ZIP code centroid, which allowed estimation of effects due to
both within-city and between-city variation of air pollution. The investigators excluded monitors
whose measurement objective focused on a single point source or those with "small measurement
scale (0-100 m)." Thus, presumably, these criteria reduced some of the exposure measurement error
associated with monitors that are highly impacted by local sources.
During the course of the study, a total of 1,816 women had one or more fatal or nonfatal
cardiovascular event, including 261 cardiovascular-related deaths. Hazard ratios for the initial
cardiovascular event were estimated. The following results are for models that included only subjects
with nonmissing exposure data for all pollutants (n = 28,402 subjects, resulting in 879 CVD events).
In the single-pollutant models, PM2 5 showed the strongest associations with CVD events among all
pollutants (HR = 1.24 [95% CI: 1.04-1.48] per 10 (ig/m3 increase in annual average), followed by
1 These results were presented graphically in Pope et al. (2002, 024689) and were estimated for Figure 5-19.
January 2010 5-111
-------
SO2 (HR = 1.07 [95% CI: 0.95-1.20] per 5 ppb increase in the annual average). For CO the single-
pollutant risk estimate was slightly (but not significantly) negative (HR = 0.96 [95% CI: 0.84-1.10]).
In the multipollutant model, which included all pollutants (i.e., PM2.5, PMi0_2.5, SO2, NO2, and O3),
the CO risk estimate was similar to the one presented in the single-pollutant model (HR = 0.96
[95% CI: 0.82-1.14]). In addition, CO was not associated with CVD events in a single pollutant
model (HR = 1.00 [95% CI: 0.90-1.10] per 0.5 ppm increase in mean annual average CO
concentration) that used all available observations. Overall this study found that PM2 5 was clearly
the best predictor of cardiovascular events.
The Washington University-EPRI Veterans' Cohort Mortality Studies
Lipfert et al. (2000, 004087) conducted an analysis of a national cohort of-70,000 male U.S.
military veterans who were diagnosed as hypertensive in the mid-1970s and were followed for
approximately 21 yr (up to 1996). Demographically, 35% of the cohort consisted of African-
American men and 57% of the cohort was defined as current smokers; however, 81% of the cohort
had been smokers at one time in their life. The study examined mortality effects in response to long-
term exposure to multiple pollutants, including, PM2 5, PMi0, PMi0_2.5, TSP, SO42", CO, O3, NO2, SO2,
and Pb. Lipfert et al. (2000, 004087) estimated exposures by indentifying the county of residence at
the time of entry to the study. Four exposure periods (1960-1974, 1975-1981, 1982-1988, and
1989-1996) were defined, and deaths during each of the three most recent exposure periods were
considered. The mean annual 95th percentile of hourly CO values during these periods declined from
10.8 ppm to 2.4 ppm. The authors noted that the pollution risk estimates were sensitive to the
regression model specification, exposure periods, and the inclusion of ecological and individual
variables. Lipfert et al. (2000, 004087) reported that indications of concurrent mortality risks
(i.e., associations between mortality and air quality for the same period) were found for NO2 and
peak O3. The estimated CO mortality risks were all negative, but not significant.
Lipfert et al. (2006, 088756) examined associations between traffic density and mortality in
the same Veterans' Cohort; however, in this analysis, the follow-up period was extended to 2001. As
in their 2000 study, four exposure periods were considered, but more recent years were included in
the 2006 analysis. The authors used the mean annual average of the 95th percentile of 24-h avg CO
in each of the exposure periods as the averaging metric. The traffic-density variable was the most
significant predictor of mortality in their analysis, remaining so in two- and three pollutant models
with other air pollutants (i.e., CO, NO2, O3. PM2.5, SO42~, non-SO42~ PM2.5, and PMi0_2.5). In the
multipollutant models, mortality-risk estimates were not statistically significant for any of the other
pollutants, except O3. The natural log of the traffic-density variable (VKTA = vehicle-km traveled
per yr) was not correlated with CO (r = -0.06) but moderately correlated with PM2 5 (r = 0.50) in this
data set. For the 1989-1996 data period, the estimated mortality relative risk was 1.02 (95% CI:
0.98-1.06) per 1 ppm increase in the mean annual 95th percentile of hourly CO concentration in a
single-pollutant model. The two-pollutant model, which included the traffic-density variable,
resulted in a relative risk of 1.00 (95% CI: 0.96-1.04). Lipfert et al. (2006, 088218) noted that the
low risk estimates for CO in this study were consistent with those observed in other long-term
exposure studies and may have been due to the localized nature of CO, which can lead to exposure
errors when data from centralized monitors is used to represent an entire county. Interestingly, as
Lipfert et al. (2006, 088756) pointed out, the risk estimates due to traffic density did not vary
appreciably across these four periods, even though regulated tailpipe emissions declined during the
study period. The authors speculated that some combination of other environmental factors such as
road dust, psychological stress, and noise (all of which constitute the environmental effects of
vehicular traffic), along with spatial gradients in SES, might contribute to the nonnegative effects
observed.
Lipfert et al. (2006, 088218) extended the analysis of the Veterans Cohort data to include the
EPA's Speciation Trends Network (STN) data, which collected chemical components of PM25. The
authors analyzed the STN data for the year 2002 and again used county-level averages. In addition,
they analyzed PM25 and gaseous pollutants data for 1999-2001. As in the other Lipfert et al. (2006,
088218) study, traffic density was the most important predictor of mortality, but associations were
also observed for EC, vanadium (V), nickel (Ni), and NO3~. Ozone, NO2, and PMi0 also showed
positive but weaker associations. The authors found no association between the mean annual 95th
percentile of hourly CO values and mortality (RR = 0.995 [95% CI: 0.988-1.001] per 1 ppm increase
January 2010 5-112
-------
in CO concentration) in a single-pollutant model. The study did not present copollutant model results
for CO.
Cause-specific
Mortality
All-cause
Cardiopulmonary
Cardiovascular
IHD
Lung Cancer
Study
Pope etal. (2002,024689V
Krewskietal. (2009, 191193)
Pope etal. (2002,024689V
Lipfert etal. (2006,088218)
Lipfert etal. (2006, 08821 8)
Lipfert etal. (2006,088756)
Jerrettetal. (2003,087380)
Jerrettetal. (2003, 087380)
Pope etal. (2002, 024689V
Krewskietal. (2009, 191193)
Pope etal. (2002,024689)a
Miller etal. (2007, 0901 30V
Miller etal. (2007,090130V
Krewskietal. (2009, 191193)
Krewskietal. (2009, 191193)
Cohort
ACS
ACS
ACS
Veterans
Veterans
Veterans
ACS
ACS
ACS
ACS
ACS
WHI
WHI
ACS
ACS
Years of
Mortality
Data
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
"The study does not present CO results quantitatively. This effect estimate and 95% confidence interval
Year(s)
ofAQ
Data
1980
1980
1982-98
1989-96
1989-96
1999-01
19821-
1982°
1980
1980
1982-98
2000
2000
1980
1980
i
0.80
Effect Estimate (95% Cl)d
-*-
t
-•-i
— "-»-
— *—
— »j-
« '
±
•
— •-'
t — ! —
• — i —
-*-
-•!—
i i i 1
0.90 1.00
- +ln(VKTA)°
+ SO/'
Multipollutant1
i i i i
1.10 1.20
were estimated from Figure 5 in Pope et al. (2002, 0246891.
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.
dUnless otherwise specified, all results represent single pollutant models
'Natural log of Vehicle-km Traveled variable.
'Multipollutant model consisted of CO + PM25, PM1Cu25, S02, N02, 03.
Relative Risk / Hazard Ratio
Figure 5-19. 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
An ecological cross-sectional analysis involves regressing county- (or city-) average health
outcome values on county-average explanatory variables such as air pollution and census statistics.
Unlike the cohort studies described above, to the extent that individual level confounders are not
adjusted for, the cross-sectional study design is considered to be subject to ecologic confounding.
However, all of the cohort studies described above are also semi-ecologic in that the air-pollution
exposure variables are ecologic (Kunzli and Tager, 1997, 086180). In this sense, cross-sectional
studies may be useful in evaluating the correlation among exposure variables.
Lipfert and Morris (2002, 019217) conducted ecological cross-sectional regressions for U.S.
counties (except Alaska) during five periods: 1960-1969, 1970-1974, 1979-1981, 1989-1991, and
1995-1997. They regressed age-specific (15-44, 45-64, 65-74, 76-84, and 85+ yr) all-cause
(excluding AIDS and trauma) mortality on air pollution, demography, climate, SES, lifestyle, and
diet. The authors analyzed TSP, PMi0, PM2.5, SO42", SO2, CO, NO2, and O3. However, air pollution
data was only available for limited periods of time depending on the pollutant: TSP up to 1991; PMi0
between 1995 and 1999; and PM2.5 between 1979 and 1984 and for 1999. In response to the varying
number of counties with valid air pollution data by pollutant and time, Lipfert and Morris (2002,
019217) employed a staged-regression approach. In the first stage, a national model was developed
for each dependent variable, excluding air pollution variables. In the second stage, regressions were
performed with the residuals on concurrent and previous periods' air pollution variables to identify
the pollutants of interest. Many results were presented because of the large number of age groups,
lagged-exposure time windows, and mortality study periods examined in the study; overall, the
results were similar to those presented in the ACS cohort studies (i.e., PM2 5 and SO42" were found to
be consistently and positively associated with mortality). Lipfert and Morris (2002, 019217)
generally found the strongest associations in the earlier time periods and when mortality and air
quality were measured in different periods (e.g., mortality data 1995-1997 and CO data 1970-1974).
January 2010
5-113
-------
Also, consistent with the Lipfert et al. (2000, 012281) and the Pope et al. (2002, 024689) cohort
studies, CO was frequently negatively (and often significantly) associated with mortality in older age
groups, especially when mortality was matched with CO levels in more recent time periods. The
younger age group (15-44 yr) often showed a positive association with CO, but considering the small
number of deaths attributed to this age group (<1% of total deaths), the association was not
informative. Overall, this study highlighted that the CO-mortality associations presented in purely
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
The evaluation of new epidemiologic studies conducted since the 2000 CO AQCD (U.S. EPA,
2000, 000907) that investigated the association between long-term exposure to CO and mortality
consistently found null or negative mortality risk estimates. No such studies were discussed in the
2000 CO AQCD (U.S. EPA, 2000, 000907). The reanalysis of the ACS data (Pope et al., 1995,
045159) by Jerrett et al. (2003, 087380) found no association between long-term exposure to CO and
mortality. Similar results were obtained in an updated analysis of the ACS data (Pope et al., 2002,
024689) when using earlier (1980) CO data; however, negative associations were found when using
more recent (1982-1998) data. These results were further confirmed in an extended analysis of the
ACS data (Krewski et al., 2009, 191193). The Women's Health Initiative Study also found no
association between CO and CVD events (including mortality) using the mortality data from recent
years (1994-1998) (Miller et al., 2007, 090130). while the series of Veterans Cohort studies found no
association or a negative association between mean annual 95th percentile of hourly CO values and
mortality (Lipfert et al., 2006, 088218: Lipfert et al., 2006, 088756). An additional study (Lipfert and
Morris, 2002, 019217) was identified that used a cross-sectional study design, which reported results
for a study of U.S. counties that were generally consistent with the cohort studies: positive
associations between long-term exposure to PM2.5 and SO42" and mortality, and generally negative
associations with CO. Overall, the consistent null and negative associations observed across
epidemiologic studies which included cohort populations encompassing potentially susceptible
populations (i.e., postmenopausal women and hypertensive men) combined with the lack of evidence
for respiratory and cardiovascular morbidity outcomes following long-term exposure to CO; and the
absence of a proposed mechanism to explain the progression to mortality following long-term
exposure to CO provide supportive evidence that there is not likely to be 3 Causal relationship
between relevant long-term exposures to CO and mortality.
5.7. Susceptible Populations
Interindividual variation in human responses indicates that some populations are at increased
risk for the detrimental effects of ambient exposure to an air pollutant (Kleeberger and Ohtsuka,
2005, 130489). The NAAQS are intended to provide an adequate margin of safety for both general
populations and sensitive subgroups, or those individuals potentially at increased risk for health
effects in response to ambient air pollution (Section 1.1). To facilitate the identification of
populations at the greatest risk for CO-related health effects, studies have evaluated factors that
contribute to the susceptibility and/or vulnerability of an individual to CO. The definition for both of
these terms varies across studies, but in most instances "susceptibility" refers to biological or
intrinsic factors (e.g., lifestage, gender) while "vulnerability" refers to nonbiological or extrinsic
factors (e.g., visiting a high-altitude location, medication use) (Table 5-25). Additionally, in some
cases, the terms "at-risk" and sensitive populations have been used to encompass these concepts
more generally. However, in many cases a factor that increases an individual's risk for morbidity or
mortality effects from exposure to an air pollutant (e.g., CO) can not be easily categorized as a
susceptibility or vulnerability factor. For example, a population that is characterized as having low
SES, traditionally defined as a vulnerability factor, may have less access to healthcare resulting in
the manifestation of disease (i.e., a susceptibility factor). But they may also reside in a location that
results in exposure to higher concentrations of an air pollutant, increasing their vulnerability.
Therefore, the terms "susceptibility" and "vulnerability" are intertwined and at times cannot be
distinguished from one another.
January 2010 5-114
-------
As a result of the inconsistencies in the definitions of "susceptibility" and "vulnerability"
presented in the literature as well as the inability to clearly delineate whether an identified factor
increases an individual's susceptibility or vulnerability to an air pollutant, in this ISA, the term
"susceptible population" will be used as a blanket term and defined as follows:
Populations that have a greater likelihood of experiencing health effects related to
exposure to an air pollutant (e.g., CO) due to a variety of factors including, but not limited to:
genetic or developmental factors, race, gender, lifestage, lifestyle (e.g., smoking status and
nutrition) orpreexisting disease, as well as population-level factors that can increase an
individual's exposure to an air pollutant (e.g., CO) such as socioeconomic status [SES], which
encompasses reduced access to health care, low educational attainment, residential location,
and other factors.
Table 5-25. Range of 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 ,, ~ ppA nmf, -IQ^QQCX
(e.g., persons with cardiovascular disease, COPD, reduced or abnormal hemoglobin, older adults, neonates). ( ' lajaaa)
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 , m. Km
can wound. Aaay^uui, la^ibU)
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, preexisting disease (e.g., asthma) and genetics) and extrinsic (previous Kleeberaer and Ohtsuka (2005 1304891
exposure and nutritional status) factors. weeoerger ana untsura (^uus, jdUfiBy)
To examine whether air pollutants (e.g., CO) differentially affect certain populations,
epidemiologic studies conduct stratified analyses to identify the presence or absence of effect
modification. A thorough evaluation of potential effect modifiers may help identify populations that
are more susceptible to an air pollutant (e.g., CO). Although the design of toxicological and
controlled human exposure studies does not allow for an extensive examination of effect modifiers,
the use of animal models of disease and the study of individuals with underlying disease or genetic
polymorphisms do allow for comparisons between subgroups. Therefore, the results from these
studies, combined with those results obtained through stratified analyses in epidemiologic studies,
contribute to the overall weight of evidence for the increased susceptibility of specific populations to
an air pollutant (e.g., CO).
The remainder of this section discusses the epidemiologic, controlled human exposure, and
toxicological studies evaluated in previous sections of Chapter 5 that provide information on
potentially susceptible populations. The studies highlighted include only those studies that presented
stratified results (e.g., males versus females or <65 yr versus > 65 yr). This approach allows for a
direct comparison between populations exposed to similar CO concentrations and within the same
study design to determine whether a factor increases the susceptibility of a population to CO-related
health effects. In addition, numerous studies that focus on only one potentially susceptible
population can provide supporting evidence on susceptibility and are described in Sections 5.2
January 2010 5-115
-------
through 5.6; however, these studies are not discussed in detail in this section. It is recognized that by
using this approach to identify potentially susceptible populations, some individuals with underlying
medical conditions (i.e., reduced O2-carrying capacity or elevated COHb levels) or lifestyle
characteristics may not be identified due to the lack of studies focusing on these populations.
Discussion of conditions affecting CO uptake and elimination as well as endogenous CO production
is presented in Sections 4.4 and 4.5, respectively.
5.7.1. Preexisting Disease
The 2000 CO AQCD (U.S. EPA, 2000, 000907) identified certain populations within the
general population that may be more susceptible to the effects of CO exposure, including individuals
(particularly older adults) with CHD and other vascular diseases, anemia, or COPD. As discussed in
the 2000 CO AQCD (U.S. EPA, 2000, 000907) and reviewed in Section 4.5 of this assessment,
diseases that cause inflammation and systemic stress are known to increase endogenous CO
production, which could potentially increase the susceptibility of individuals with such conditions to
health effects induced by ambient CO exposure. The level of COHb that results in the manifestation
of health effects varies depending on health outcome and disease state of individuals. The following
sections summarize the evidence presented in the 2000 CO AQCD (U.S. EPA, 2000, 000907) along
with new evidence which identifies populations with various preexisting diseases that may be
susceptible to CO-induced health effects.
5.7.1.1. Cardiovascular Disease
Controlled exposures to CO resulting in COHb concentrations of 2-6% have been shown to
affect cardiovascular function among individuals with CAD. Several studies have reported
significant decreases in the time to onset of exercise-induced angina or ST-segment changes
following CO exposure in patients with stable angina. In the largest such study (Allred et al., 1989,
013018: Allred et al., 1989, 012697: Allred et al., 1991, 011871). COHb concentrations as low as
2.0-2.4% were observed to significantly decrease the time required to induce ST-segment changes
indicating myocardial ischemia (p = 0.01) (Section 5.2.4). In addition to the effects of CO on
myocardial ischemia, there is some evidence to suggest that CO may provoke cardiac arrhythmia in
patients with CAD; however, this has not been observed at COHb concentrations below 6% (Sheps
et al., 1990, 013286). While healthy adults have been shown to experience a decrease in exercise
performance following or during exposure to CO, no changes in cardiac rhythm or ECG parameters
have been demonstrated.
Evidence of CO-induced health effects in individuals with CAD is coherent with results from
epidemiologic studies that examined the effect of preexisting cardiovascular conditions, through
either secondary diagnoses or underlying comorbidities, on associations between CO and ED visits
and HAs. Mann et al. (2002, 036723) found increased associations between CO and IHD HAs in
individuals with secondary diagnoses of either CHF or dysrhythmia in southern California. Peel et
al. (2007, 090442) also examined the effect of underlying cardiovascular conditions on
cardiovascular-related HAs in response to short-term exposure to air pollutants, including CO in
Atlanta, GA. Individuals with underlying dysrhythmia were found to have increased HAs for IHD,
but unlike Mann et al. (2002, 036723). underlying CHF was not found to increase IHD HAs. Peel et
al. (2007, 090442) also examined other underlying conditions and found increased HAs for
additional cardiovascular effects not specifically related to IHD, including: dysrhythmia, PVCD, and
CHF in individuals with underlying hypertension; dysrhythmia and PVCD in individuals with
underlying CHF; and PVCD in individuals with underlying dysrhythmia. Although there is no
evidence for a clear pattern of CO-induced cardiovascular effects among individuals without CAD
across the epidemiologic studies evaluated, the available evidence suggests that underlying
dysrhythmia increases IHD HAs in response to short-term exposure to CO.
Additional evidence for increased CO-induced cardiovascular effects not specifically related to
IHD is provided by toxicological studies that used animal models of cardiovascular disease. These
studies have demonstrated that short-term exposure to 50 ppm CO in rats exacerbates
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 for primary pulmonary hypertension is low, secondary pulmonary hypertension is
January 2010 5-116
-------
afrequent complication of COPD (Section 5.7.1.2) and certain forms of heart failure. These studies
demonstrate the potential for short-term exposure to CO to adversely affect individuals with
underlying cardiovascular conditions.
The combined evidence from controlled human exposure and epidemiologic studies provides
coherence and biological plausibility for the association between CO and cardiovascular morbidity
in individuals with CAD, particularly those with IHD. Approximately 13.7 million people in the U.S.
have been diagnosed with CAD (also known as CHD), some fraction of whom have IHD (Table
5-26). These individuals, therefore, represent a large population that may be more susceptible to
ambient CO exposure than the general population. In addition, the continuous nature of the
progression of CAD and its close relationship with other forms of cardiovascular disease suggest that
a larger population than just those individuals with a prior diagnosis of CAD may be susceptible to
health effects from CO exposure.
Table 5-26. Adult U.S. population in 2007 with respiratory diseases and cardiovascular diseases.
Chronic Condition/
Disease
COPD3
Chronic bronchitis
Emphysema
Cardiovascular Diseases'1
All heart disease"
Coronary heart disease11
Hypertension
Stroke
Adults (18+)
(Millions)
7.6
3.7
(Millions)
25.1
13.7
52.9
5.4
Percentage of U.S
All (18+)
3.4
1.6
All (18+)
11.2
6.1
23.2
2.4
18-44
2.3
0.2
18-44
4.1
0.9
8.2
0.3
Adult
45-64
4.2
2.3
45-64
12.2
6.7
32.1
2.8
Population by Age
65-74
5.5
4.5
65-74
27.1
18.6
50.9
6.3
75+
4.8
5.2
75+
35.8
23.6
57.4
10.6
NE
2.8
1.1
NE
10.6
5.3
21.3
2.2
Percentage
MW
3.2
1.8
MW
12.3
6.7
23.4
2.3
by Region
S
4.0
1.8
S
11.3
6.4
25.1
2.7
W
2.9
1.6
W
10.2
5.5
21.0
2.2
3 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 mo that they had bronchitis. A person maybe represented in more than one row.
b In separate questions, respondents were asked if they had ever been told by a doctor or other health professional that they had: hypertension (or 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.
c Heart disease includes coronary heart disease, angina pectoris, heart attack, or any other heart condition or disease.
d Coronary heart disease includes coronary heart disease, angina pectoris, or heart attack.
Source: National Health Interview Survey, 2007, Tables 1-4 (Pleis and Lucas, 2009, 202833).
5.7.1.2. Obstructive Lung Disease
COPD is a progressive disease resulting in decreased air flow to the lungs and which is
especially prevalent among smokers. O2 limitation resulting from this reduction in air flow may
exacerbate CO-related O2 limitation and subsequent cardiovascular or respiratory effects in
individuals with COPD. 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 (Table
5-26), although there could be overlap among these two populations. The 2000 CO AQCD
(U.S. EPA, 2000, 000907) 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 these individuals to CO during submaximal exercise typical of normal daily activity. In
addition, COPD patients who are smokers may have elevated baseline COHb levels of 4-8%
(U.S. EPA, 2000, 000907). COPD is often accompanied by a number of changes in gas exchange,
including increased VD and VA/Q inequality (Marthan et al., 1985, 086334). which could slow both
CO uptake and elimination.
The few epidemiologic studies of cardiovascular effects in individuals with underlying COPD
show weak positive associations between ambient CO and increased CVD HAs or ED visits. For
example, Peel et al. (2007, 090442) found that associations between short-term CO exposure and
HAs for PVCD or CHF were increased in individuals with a secondary diagnosis of COPD.
However, underlying COPD was not associated with increased IHD or dysrhythmia HAs. As
January 2010
5-117
-------
described in Section 5.7.1.1, animal toxicological studies demonstrate CO-induced exacerbation of
vascular remodeling related to pulmonary hypertension, a form of which is a frequent complication
ofCOPD.
A controlled human exposure study of respiratory effects in individuals with COPD (Bathoorn
et al., 2007, 193963). found that two of the patients experienced COPD exacerbation during or
following CO exposure at 100-125 ppm for 2 h, although a slight anti-inflammatory effect was also
observed. Although the majority of the evidence for CO-induced effects comes from studies that
focus on individuals with COPD, epidemiologic studies also report weak positive associations for
asthmatics (Section 5.5.2.2) who can also experience exercise-induced airflow limitation. In
addition, preliminary evidence from a recent animal toxicological study indicates mild pulmonary
inflammation in response to 50 ppm CO (Ohio et al., 2008, 096321). Since pulmonary inflammation
plays an important role in the exacerbation of COPD and asthma, it may serve as a mechanism
underlying CO-induced respiratory effects; however, additional research is needed to confirm these
results. Taken together, the limited evidence from epidemiologic and controlled human exposure
studies and some preliminary evidence from toxicological studies suggests that individuals with
obstructive lung disease (e.g., COPD patients with underlying hypoxia, asthmatics) may be
susceptible to cardiovascular or respiratory effects due to CO exposure.
5.7.1.3. Diabetes
Exhaled CO concentrations are elevated in individuals with diabetes and are correlated with
blood glucose levels and duration of disease, indicating increased endogenous CO production
(Section 4.5). As a result, it has been speculated that elevated baseline COHb levels in diabetic
individuals could increase the susceptibility of diabetics to CO-induced health effects in response to
ambient CO exposures. Epidemiologic studies provide evidence which suggests that diabetics are at
increased risk for cardiovascular ED visits and HAs compared to nondiabetics in response to short-
term exposure to CO (Pereira Filho et al., 2008, 190260; Zanobetti and Schwartz, 2001, 016710).
This is consistent with results reported by Peel et al. (2007, 090442). who observed an increase in
cardiovascular-related ED visits for dysrhythmias and PVCD in individuals with diabetes but not for
IHD or CHF ED visits. The results from Peel et al. (2007, 090442) that indicate an increase in
dysrhythmia ED visits for individuals with diabetes are consistent with results from a panel study
conducted by Min et al. (2009, 199514) to investigate the relationship between CO and HRV in
individuals with metabolic syndrome. Metabolic syndrome is characterized by risk factors for both
diabetes and CVD, including elevations in blood pressure, fasting blood glucose, triglycerides, and
waist circumference, as well as low levels of HDL cholesterol. Min et al. (2009, 199514) observed
associations between short-term exposure to CO and changes in HRV parameters among subjects
with metabolic syndrome but not among healthy subjects. In addition, the observed associations
were robust in copollutant models with either PMi0 or NO2. In an analysis of individual risk factors,
the CO effects were stronger among subjects with higher levels of fasting blood glucose or
triglycerides. Although no evidence was identified from controlled human exposure or toxicological
studies regarding CO exposure and diabetes, vascular dysfunction was demonstrated in an animal
model of metabolic syndrome and was attributed to increased endogenous CO production (Johnson
et al., 2006, 193874). Thus, increased endogenous CO production and the potential for higher
baseline COHb concentrations, combined with the limited epidemiologic evidence showing
cardiovascular effects, suggests that diabetics are potentially susceptible to short-term exposure to
CO.
5.7.1.4. Anemia
Although no controlled human exposure or epidemiologic studies were identified that
specifically examined CO-related health effects in individuals with anemia, the 2000 CO AQCD
(U.S. EPA, 2000, 000907) suggested that conditions such as anemia that produce tissue hypoxia by
lowering the blood O2 carrying capacity or content will result in a greater risk of effects from COHb-
induced hypoxia due to the combined effects of both sources of hypoxia. As discussed in Section
4.4.4 of this ISA, anemias are a group of diseases that lower hematocrit and result in reduced arterial
O2 content due to Hb deficiency through hemolysis, hemorrhage, or reduced hematopoiesis.
Hereditary hemoglobinopathies such as sickle cell anemia and thalassemia also reduce the O2-
carrying capacity of the blood. Anemia may also result from pathologic conditions characterized by
January 2010 5-118
-------
chronic inflammation such as malignant tumors or chronic infections (Cavallin-Stahl et al., 1976,
086306; Cavallin-Stahl et al., 1976, 193239). The cardiovascular system of people with anemia
compensates for the reduction in O2 carrying capacity by increasing cardiac output as both heart rate
and stroke volume increase. One of the most prevalent forms of anemia arises from a single-point
mutation in the Hb gene, resulting in sickle cell diseases. The affinity of Hb for O2 and its O2
carrying capacity is reduced, causing a shift to the right in the O2 dissociation curve. It is well
documented that African-American populations have a higher incidence of sickle cell anemia, which
may be a risk factor for effects due to CO-mediated hypoxia. Other hereditary hemoglobinopathies,
such as thalassemia, also reduce the O2-carrying capacity of the blood due to the production of an
abnormal form of hemoglobin. Overall, lowered hematocrit due to anemia may result in increased
susceptibility and a greater response to inhalation of ambient CO.
Anemia may also increase the susceptibility of an individual to CO exposure in a different
manner through the increased production of endogenous CO as a result of the disturbance of RBC
hemostasis by accelerated destruction of hemoproteins (Section 4.5). Pathologic conditions such as
hemolytic anemias, hematomas, thalassemia, Gilbert's syndrome with hemolysis, and other
hematological diseases and illness will accelerate endogenous 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
two- to threefold higher than healthy individuals and CO production rates two- to eightfold higher
(Coburn et al., 1966, 010984). Recent studies report elevated COHb levels of 4.6-9.7% 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). Taken together, this
evidence suggests that individuals with anemia who have diminished O2-carrying capacity and/or
high baseline COHb levels may be more susceptible to health effects due to ambient CO exposure,
although no studies were identified that evaluated specific CO-related health effects in anemic
individuals.
5.7.2. Life stage
Age alters the variables that influence the uptake, distribution, and elimination of CO
(Section 4.4.3). COHb levels decline more rapidly in young children compared to adults after CO
exposure (Joumard et al., 1981, 011330: Klasner et al., 1998, 087196). After infancy, the COHb half-
life increases with age, practically doubling between the ages of 2 and 70 yr (Joumard et al., 1981,
011330). However, it should be noted that the rate of this reduction in CO elimination is very rapid
in the growing years (2-16 yr of age) but slows beyond adolescence. An increase in alveolar volume
and DLCO were observed with increasing body length of infants and toddlers (Castillo et al., 2006,
193234): these changes suggest faster CO uptake due to more advanced lung development. After
infancy, increasing age decreases DLCO and increases VA/Q mismatch, resulting in a longer duration
for both loading and elimination of CO from the blood (Neas and Schwartz, 1996, 079363).
5.7.2.1. Older Adults
The 2000 CO AQCD (U.S. EPA, 2000, 000907) noted that changes in metabolism that occur
with age, particularly declining maximal oxygen uptake, may make the aging population susceptible
to the effects of CO via impaired oxygen delivery to the tissues. Several epidemiologic studies
compared cardiovascular outcomes in older and younger adults, although no such studies were
conducted in the U.S. In a study in Australia and New Zealand, Barnett et al. (2006, 089770) found
an increase in IHD and MI HAs among individuals > 65 yr of age compared with individuals aged
15-64 yr in response to short-term exposure to CO. Lee et al. (2003, 095552) also found an
association with IHD HAs in Seoul, Korea, for individuals > 65 yr of age but not when all
individuals were included in the analysis. Lanki et al. (2006, 089788) found an association with HAs
for nonfatal MI in a multicity European study among those aged > 75 yr but not for those <75 yr old.
In contrast, D'Ippoliti et al. (2003, 074311) observed higher associations for MI hospital admissions
in Rome among 18- to 64-yr olds than among either 65- to 74-yr olds or those 75 yr and over.
Szyszkowicz (2007, 193793) found slightly lower associations for IHD hospital admissions in
Montreal, Canada among those >64 yr of age than for the all-age group. Another Canadian study
(Fung et al., 2005, 074322) conducted in Windsor, Ontario, found some evidence of increased
January 2010 5-119
-------
associations for between CO and CVDs (defined as HF, IHD, or dysrhythmia) among individuals
> 65 yr of age compared with the <65-yr age group. No controlled human exposure studies or
toxicological studies were identified that compared CO effects among older and younger adults or
animal models of senescence, respectively. Overall, the epidemiologic studies evaluated provide
limited evidence that older adults may be susceptible to CO exposure.
A combination of factors may be responsible for increased susceptibility to CO-related health
effects among older adults. One important factor which may contribute to the observed increases in
CO-induced cardiovascular effects is the much higher prevalence of CAD and other cardiovascular
conditions in older adults compared with the general population. As shown in Table 5-26,18.6% of
adults aged 65-74 yr and 23.6% of adults aged 75 yr and over reported having CHD, as compared
with 6.1% of the population as a whole. Both the higher prevalence of CAD and the gradual decline
in physiological processes associated with aging (U.S. EPA, 2006, 192082) may contribute to
increased health effects in response to CO in this population. Older adults represent a large and
growing fraction of the U.S. population, from 12.4% or 35 million people in 2000 to a projected
19.3% or 72 million people in 2030 (U.S. Census, 2000, 157064). and, as a result, are a large
population that is potentially susceptible to CO-induced health effects.
5.7.2.2. Gestational Development
CO inhaled by pregnant animals quickly crosses the placental barriers and enters fetal
circulation. Effects of ambient CO may be enhanced during gestation because fetal CO
pharmacokinetics do not follow the same kinetics as maternal CO exposure; this contributes to the
difficulty in estimating fetal COHb based on maternal levels. It has been demonstrated that human
fetal Hb has a higher affinity for CO than adult Hb (Di Cera et al., 1989, 193998). Maternal and fetal
COHb concentrations have been modeled as a function of time using a modified CFK equation (Hill
et al., 1977, 011315). At steady-state conditions, fetal COHb has been found to be 10-15% higher
than maternal COHb levels. For example, exposure to 30 ppm CO results in a steady-state maternal
COHb of 5% and a fetal COHb of 5.75%. Fetal CO uptake lags behind maternal uptake for the first
few hours but later may overtake the maternal values. Similarly, during washout, fetal COHb levels
are maintained for longer, with a half-life of around 7.5 h versus the maternal half-life of around 4 h
(Longo and Hill, 1977, 010802). In addition, maternal 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).
Epidemiologic studies provide limited evidence that in utero CO exposure is associated with
changes in various birth outcomes (Section 5.4.1). CO exposure during early pregnancy was
associated with an increased risk of PTB. In the studies that examined associations between CO and
birth defects, maternal CO exposure was associated with an increased risk of cardiac birth defects,
which is also coherent with evidence in Section 5.2 identifying the heart as a target organ for CO.
There is also evidence for small reductions in birth weight (10-20 g) associated with CO exposure,
generally in the first or third trimester, although the decrease does not generally translate to an
increased risk of LEW or SGA. It is therefore difficult to conclude if CO is related to a small change
in birth weight across all births or a marked effect in some subset of births. In addition, there is
limited evidence that prenatal CO exposure is associated with an increased risk of infant mortality in
the post-neonatal period.
Toxicological studies lend biological plausibility to the CO-related developmental outcomes
observed in epidemiologic studies (Section 5.4.2). Associations have been observed between CO
exposure in laboratory animals and decrements in birth weight as well as reduced prenatal growth.
Animal toxicological studies also provide evidence for effects on the heart, including transient
cardiomegaly at birth after continuous in utero CO exposure and delayed myocardial
electrophysiological maturation. Evidence exists for additional developmental outcomes which have
been examined in toxicological studies but not epidemiologic or human clinical studies, including
behavioral abnormalities, learning and memory deficits, locomotor effects, neurotransmitter changes,
and changes in the auditory system. Furthermore, exogenous CO may interact or disrupt the normal
physiological roles of endogenous CO in the body. There is evidence that CO plays a role in
maintaining pregnancy, controlling vascular tone, regulating hormone balance, and sustaining
normal follicular maturation.
The developmental outcomes examined in the epidemiologic studies evaluated affect a
substantial portion of the U.S. population. PTB and LEW have been established as strong predictors
January 2010 5-120
-------
of infant mortality and morbidity (Barker et al., 2002, 193960; Berkowitz and Papiernik, 1993,
055466; Li et al., 2003, 193965; Mclntire et al., 1999, 015310). In 2004, 36.5% of all infant deaths
in the U.S. were preterm-related (MacDorman et al., 2007, 193973). Vital statistics for the year 2005
in the U.S. showed that the rate for PTB was 12.7%, which has risen 20% since 1990, and the rate
for LEW was 8.2%, which has risen 17% since 1990 (Martin et al., 2007, 193982). Data from the
Metropolitan Atlanta Congenital Defects Program (MACDP), which is one of the most
comprehensive birth defect registries in the U.S., have shown that the prevalence of congenital heart
defects increased between 1968 and 1997. During 1995-1997 the rate was 90.2 per 10,000 births
(0.9%) and this was an increase of 58.7 per 10,000 births above the rate during 1968-1972 (Botto et
al., 2001, 192379). Cardiovascular defects are the single largest contributor to infant mortality
attributable to birth defects (CDC, 1998, 193243). Between 1995 and 1997, 1 in 13 infant deaths
(7.4%) was due to a congenital heart defect (Boneva et al., 2001, 193972). The combined evidence
from epidemiologic and toxicological studies, along with the increasing prevalence of PTB, LBW,
and cardiac birth defects in the U.S. population, indicates that critical developmental phases may be
characterized by enhanced sensitivity to CO exposure.
5.7.3. Gender
COHb concentrations are generally higher in male subjects than in female subjects (Horvath et
al., 1988, 012725). In addition, the COHb half-life is longer in healthy men than in women of the
same age, which may be partially explained by differences in muscle mass or the slight correlation
between COHb half-life and increased height (Joumard et al., 1981, 011330). The rate of decline of
DLCO with age is lower in middle-aged women than in men; however, it is similar in older adults
(Neas and Schwartz, 1996, 079363). Lower rates of decline in lung diffusing capacity are consistent
with the observation that women tend to be more resistant than men to altitude hypoxia (Horvath et
al., 1988, 012725). Women also experience fluctuating COHb levels through the menstrual cycle
when endogenous CO production doubles in the progesterone phase (0.62 mL/h versus 0.32 mL/h in
estrogen phase) (Delivoria-Papadopoulos et al., 1974, 086316; Mercke and Lundh, 1976, 086309).
Similarly, endogenous CO production increases during pregnancy due to contributions from fetal CO
production and altered hemoglobin metabolism as described above. In an epidemiologic study
investigating the association between short-term CO exposure and IHD hospital admissions
(Szyszkowicz, 2007, 193793), males had higher associations than females in both the all-ages group
and in those >64 yr of age. However, this limited epidemiologic evidence combined with known
gender-related differences in endogenous CO production do not provide sufficient basis for
determining whether CO disproportionately affects males or females.
5.7.4. Altitude
Higher altitude results in changes in a number of factors that contribute to the uptake and
elimination of CO. The relationship between altitude and CO exposure has been discussed in depth
in the 2000 CO AQCD (U.S. EPA, 2000, 000907) and other documents (U.S. EPA, 1978, 086321)
and is reviewed in Section 4.4.2 of this ISA. In an effort to maintain proper O2 transport and supply,
physiological changes occur as compensatory mechanisms to combat the decreased barometric
pressure and resulting altitude-induced hypobaric hypoxia. These changes, which include increases
in BP and cardiac output and redistribution of blood from skin to organs and from blood to
extravascular compartments, generally will favor increased CO uptake and COHb formation, as well
as CO elimination. It has been demonstrated that breathing CO (9 ppm) at rest at altitude produces
higher COHb compared to sea level (McGrath et al., 1993, 013865). whereas high-altitude exposure
in combination with exercise causes a decrease in COHb levels versus similar exposure at sea level
(Horvath et al., 1988, 012725). This decrease could be a shift in CO storage or suppression of COHb
formation, or both. In a controlled human exposure study on the health effects of CO at altitude,
Kleinman et al. (1998, 047186) observed that CO exposure and simulated high altitude reduced the
time to onset of angina relative to clean-air exposure at sea level by 9% and 11%, respectively,
among a group of individuals with CAD. In this study, the combined effects of altitude and CO
exposure were observed to be additive, with subjects experiencing, on average, an 18% decrease in
the time to onset of angina following exposure to CO and simulated altitude relative to clean air
January 2010 5-121
-------
exposure at sea level. No epidemiologic studies were identified that specifically examined the effect
of altitude on health effects due to CO exposure.
Altitude also increases the baseline COHb levels by inducing endogenous CO production and
has been shown to be positively associated with baseline COHb concentrations (McGrath, 1992,
001005: McGrath et al., 1993, 013865). This increase in COHb with altitude-induced hypoxiahas
also been associated with increases in mRNA, protein, and activity of HO-1 in rats and cells leading
to enhanced endogenous CO production (Carraway et al., 2002, 026018; Chin et al., 2007, 190601).
Early HH has been found to increase lung HO-1 protein and activity, whereas chronic HH induced
endogenous CO production in nonpulmonary sites (Section 4.5) (Carraway et al., 2000, 021096).
Whether other variables (such as an accelerated metabolism or a greater pool of Hb, transient shifts
in body stores, or a change in the elimination rate of CO) play a role in increasing COHb
concentrations at high altitudes has not been explored.
As the length of stay increases at high altitude, acclimatization occurs, inducing
hyperventilation, polycythemia or increased red blood cell count, and increased tissue capillarity and
Mb content in skeletal muscle, which could also favor increased CO uptake. Most of the initial
adaptive changes gradually revert to sea-level values. However, these adaptive changes persist in
people raised at high altitude even after reacclimatization to sea level (Hsia, 2002, 193857). This
evidence indicates that visitors to high altitude locations may represent a potentially susceptible
population for increased risk of health effects due to CO exposure.
5.7.5. Exercise
Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase
in gas exchange. O2 consumption can increase more than 10-fold during exercise. Similarly,
ventilation, membrane and lung diffusing capacity, pulmonary capillary blood volume, and cardiac
output increase proportional to work load. The majority of these changes facilitate CO uptake and
transport by increasing gas exchange efficiency. Likewise, the COHb elimination rate increases with
physical activity, causing a decrease in COHb half-life (Joumard et al., 1981, 011330). The potential
effects of CO on exercising individuals was demonstrated in a controlled human exposure study
where healthy subjects exposed to CO achieved COHb levels of approximately 5%, which resulted
in a significant decrement in exercise duration and maximal effort capability (measured by metabolic
equivalent units) (Adir et al., 1999, 001026). These effects could be attributed to CO lowering the
anaerobic threshold, allowing earlier fatigue of the skeletal muscles and decreasing maximal effort
capability during heavy exercise. Due to the counterbalancing effects of increased rates of COHb
formation and elimination, it is unclear whether individuals engaging in light to moderate exercise
are a potentially susceptible population for increased health effects due to ambient CO exposure.
5.7.6. Proximity to Roadways
Individuals that spend a substantial amount of time on or near heavily traveled roadways, such
as commuters and those living or working near freeways, are likely to be exposed to elevated CO
concentrations, as discussed in Chapter 3. Targeted sampling studies have found CO concentrations
measured at the roadside to be several-fold higher than concentrations measured a few hundred
meters downwind (Baldauf et al., 2008, 191017: Zhu et al., 2002, 041553). with the shape of the
concentration profile dependent on wind speed and direction. AQS monitoring data aggregated
across multiple sites with no adjustment for wind conditions show somewhat higher concentrations
for microscale (near-road) monitors relative to middle-scale monitors, although the ratio is lower
than that observed in the roadside studies. Elevated near-road concentrations are important for
residents of the estimated 17.9 million occupied homes nationwide (16.1%) that are within
approximately 90 m of a freeway, railroad, or airport, according to the 2007 American Housing
Survey (2008, 194013).
Studies of commuters have shown that commuting time is an important determinant of CO
exposure for those traveling by car, bicycle, public transportation, and walking (Bruinen de Bruin et
al., 2004, 190943: Kaur et al., 2005, 086504: Scotto Di Marco et al., 2005, 144054). In-vehicle
concentrations have been measured to be several times higher than concentrations measured at fixed-
site monitors not located adjacent to roadways (Bruinen de Bruin et al., 2004, 190943: Chang et al.,
2000, 001276: Kaur et al., 2005, 086504: Riediker et al., 2003, 043761: Scotto Di Marco et al., 2005,
January 2010 5-122
-------
144054). Commuting is likely to be an important contributor to CO exposure for the 5.5 million U.S.
worker (5%) who drive 60 min or more to work (U.S. Census Bureau, 2008, 194013). This evidence
for elevated on-road and near-road CO concentrations combined with residential and commuting
data indicates that the large numbers of individuals who spend a substantial amount of time on or
near heavily traveled roadways are an important population that is potentially susceptible to
increased health risks due to ambient CO exposure.
5.7.7. Medications and Other Substances
Endogenous CO production can be altered by medications or a number of physiological
conditions that increase RBC destruction, the breakdown of hemoproteins other than Hb, and the
production of bilirubin (Section 4.5). Nicotinic acid, allyl-containing compounds (acetamids and
barbiturates), diphenylhydantoin, progesterone, contraceptives, and statins increase CO production.
One epidemiologic study (Dales, 2004, 099036) investigated the effect of medication use on the
relationship between ambient CO and HRV in individuals with CAD. The authors observed an
association between short-term CO exposure and an increase in SDNN for CAD patients not taking
beta blockers; however, this association did not persist in CAD patients taking beta blockers.
Compounds such as carbon disulfide and sulfur-containing chemicals (parathion and
phenylthiourea) increase CO following metabolism by cytochrome p450s. The P450 system may
also cause large increases in CO produced from the metabolic degradation of dihalomethanes such as
methylene chloride leading to very high (>10%) COHb levels, which can be further enhanced by
prior exposure to HCs or ethanol. Minor sources of endogenous CO include the auto-oxidation of
phenols, photo-oxidation of organic compounds, and lipid peroxidation of cell membrane lipids.
Taken together, this evidence indicates that individuals ingesting medications and other substances
that enhance endogenous or metabolic CO production represent a population that is potentially
susceptible to increased health effects due to additional exposure to ambient CO.
5.7.8. Summary of Susceptible Populations
Individuals with CAD represent the population most susceptible to increased risk of CO-
induced health effects, based on evidence of significant decreases in the time to onset of exercise-
induced angina or ST-segment changes observed in controlled human exposure studies of individuals
with CAD. This is coherent with the results from epidemiologic studies that observed associations
between short-term CO exposure and ED visits and HAs for IHD and related outcomes. The limited
evidence from stratified analyses in epidemiologic studies, which indicates that secondary diagnoses
of CHF or dysrhythmia modify associations between short-term CO exposure and IHD HAs,
provides further support that individuals with cardiovascular disease represent a potentially
susceptible population. Additional evidence is provided by toxicological studies that demonstrated
exacerbation of cardiomyopathy and increased vascular remodeling in animal models of
cardiovascular disease. Although it is not clear whether the small changes in COHb associated with
ambient CO exposures result in substantially diminished O2 delivery to tissues, the known role of
CO in limiting O2 availability provides biological plausibility for ischemia-related health outcomes
following CO exposure. The continuous nature of the progression of CAD and its close relationship
with other forms of cardiovascular disease suggest that a larger population than just those individuals
with a prior diagnosis of CAD may be susceptible to health effects from CO exposure.
Populations potentially susceptible to CO-induced health effects also include individuals with
other preexisting diseases, such as COPD or diabetes. Preliminary evidence available from
controlled human exposure and epidemiologic studies suggests that individuals with obstructive lung
disease may be susceptible to increased cardiovascular or respiratory effects due to CO exposure.
Increased endogenous CO production and the potential for higher baseline COHb concentrations in
individuals with diabetes, combined with the limited epidemiologic evidence showing cardiovascular
effects, suggests that diabetics are potentially susceptible to short-term exposure to CO. Individuals
with various types of anemia who have diminished O2-carrying capacity and/or high baseline COHb
levels may be more susceptible to health effects due to ambient CO exposure, although no studies
were identified that evaluated specific CO-related health effects in individuals with anemia.
There is also evidence that older adults and the developing young represent populations
potentially susceptible to CO-induced health effects. Epidemiologic studies provide limited evidence
January 2010 5-123
-------
from stratified analyses indicating that associations between short-term CO exposure and hospital
admissions for CAD are higher among those > 65 yr old than for those <65 yr. The older adult
population also has a much higher prevalence of CAD than the general population as a whole, which
may contribute to their increased susceptibility. Recent studies on birth outcomes have provided
limited evidence of associations between in utero CO exposure and PTB, LEW and cardiac birth
defects. Toxicological studies provide evidence of effects on birth weight and growth as well as
development of the cardiovascular and nervous systems following prenatal exposure to CO. This
evidence, combined with differences between fetal and maternal CO pharmacokinetics, indicates that
critical developmental phases may be characterized by enhanced sensitivity to CO exposure.
Visitors to high-altitude locations may represent a potentially susceptible population due to
changes in factors which affect the uptake and elimination of CO, although acclimatization occurs as
length of stay increases. Individuals with substantial exposure to mobile source emissions, such as
commuters and those living near heavily traveled roadways, represent an important population
potentially susceptible to increased risk of CO-induced health effects due to elevated on-road and
roadside CO concentrations.
Overall, the controlled human exposure, epidemiologic, and toxicological studies evaluated in
this assessment provide evidence for increased susceptibility among multiple populations. Medical
conditions that increase endogenous CO production rates may also contribute to increased
susceptibility to health effects from ambient CO exposure. Although the weight of evidence varies
depending on the factor being evaluated, the clearest evidence indicates that individuals with CAD
are most susceptible to an increase in CO-induced health effects.
5.8. Summary
The evidence reviewed in this chapter describes recent findings regarding the health effects of
ambient CO. Section 5.1 presents evidence on the mode of action of CO, including its role in
limiting O2 availability as well as its role in altered cell signaling. Evidence is presented in
subsequent sections on the effect of short- and long-term exposure to CO on cardiovascular
morbidity (Section 5.2), the central nervous system (Section 5.3), birth outcomes and developmental
effects (Section 5.4), respiratory morbidity (Section 5.5), and mortality (Section 5.6). Potentially
susceptible populations at increased risk of CO-induced health effects are discussed in Section 5.7.
Table 2-1 summarizes causal determinations for the health outcome categories reviewed in this
assessment. An integrative overview of the health effects of ambient CO and uncertainties associated
with interpretation of the evidence is provided in Chapter 2. The strongest evidence regarding CO-
induced health effects relates to cardiovascular morbidity, and the combined evidence from
controlled human exposure studies and epidemiologic studies indicates that a causal relationship is
likely to exist between relevant short-term CO exposures and cardiovascular morbidity, particularly
in individuals with CAD. The evidence is suggestive of a causal relationship between short-term
exposure to CO and respiratory morbidity as well as between short-term CO exposure and mortality.
The evidence is also suggestive of a causal relationship for birth outcomes and developmental effects
following long-term exposure to CO, and for central nervous system effects linked to short- and
long-term exposure to CO. The evidence indicates that there is not likely to be a causal relationship
between long-term exposure to CO and mortality. For respiratory morbidity following long-term
exposure to CO, the evidence was inadequate to infer a causal relationship.
January 2010 5-124
-------
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 JJ; 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; Cardellach F; 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: 568-76.
194960
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).
January 2010 5-125
-------
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. FASEB J, 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, 3 0:
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, 120: 949-955. 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
January 2010 5-126
-------
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 JE Jr; Heistad DD; 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; Scotto di Marco G; Kephalopoulos S; Cavallo D; Maroni M
(2004). Personal carbon monoxide exposure levels: Contribution of local sources to exposures and
microenvironment concentrations in Milan. J Expo Anal 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 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; Wisl0ff 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
J Pharmacol, 125: 909-915. 087170
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
January 2010 5-127
-------
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
CDC (1998). Trends in infant mortality attributable to birth defects-United States, 1980-1995. MMWR Morb Mortal Wkly
Rep, 47: 773-778. 193243
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
Chen L; 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 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
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
January 2010 5-128
-------
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.
PNAS, 104: 5109-5114. 190601
Chuang KJ; Coull BA; Zanobetti A; Suh H; Schwartz J; Stone PH; Litonjua A; Speizer FE; Gold DR (2008). Participate
Air Pollution as a Risk Factor for ST-Segment Depression in Patients With Coronary Artery Disease. Circulation,
118: 1314-1320. 155731
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
degradation in the rat brain. Free Radic Biol Med, 37: 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
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: Staimer N; Tjoa T; Gillen DL; Polidori A; Arhami M; Kleinman MT; Vaziri ND; Longhurst J; Sioutas C
(2009). Air pollution exposures and circulating biomarkers of effect in a susceptible population: clues to potential
causal component mixtures and mechanisms. Environ Health Perspect, 117: 1232-1238. 200844
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
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. Mol Hum Reprod, 10: 211-4. 193894
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
January 2010 5-129
-------
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
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 Curr Iss, 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
January 2010 5-130
-------
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
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
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
January 2010 5-131
-------
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 HE; 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 Crit 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: 163-168. 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. Health Effects Institute. Boston, MA. 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
January 2010 5-132
-------
Horvath SM; Bedi JF; Wagner JA; Agnew JW (1988). Maximal aerobic capacity at several ambient concentrations of
carbon monoxide at several altitudes. J Appl Physiol, 65: 2696-2708. 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. J
Appl Physiol, 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. J Expo Sci Environ Epidemiol, 16: 225-237. 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 Curr Iss, 66:
1735-1777. 087380
Johnson FK; Durante W; Peyton KJ; Johnson RA (2003). Heme oxygenase inhibitor restores arteriolar nitric oxide function
inDahlrats. Hypertension, 41: 149-155. 193868
January 2010 5-133
-------
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
Joumard R; 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 p3 Sbeta mitogen-activated protein kinase mediates the antiproliferative effect of carbon monoxide. PNAS, 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; Kaufmann P (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; Osann K; 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
January 2010 5-134
-------
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 TL; Lund AK; Lucas SN; Campen MJ (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling
of eNOS. Toxicol Appl Pharmacol, 230: 346-351. 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). Acomparative 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. Am J
Transplant, 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. Health Effects Institute. Cambridge, MA.http://pubs.healtheffects.org/view.php?id=6. 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. Report Nr. 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 CA; Mahesh VB; Brann DW (1996). Regulation of gonadotropin-releasing hormone (GnRH) secretion by heme
molecules: a regulatory role for carbon monoxide? Endocrinology, 137: 790-793. 078819
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
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
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 andpreterm delivery. Environ Health Perspect, 114: 905-910. 089828
January 2010 5-135
-------
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 hemo static/inflammatory markers: a population-based study. JExpo 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
January 2010 5-136
-------
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. Pharmacol Toxicol, 85: 192-
197.011549
Longo LD (1970). Carbon monoxide in the pregnant mother and fetus and its exchange across the placenta. Ann N Y Acad
Sci, 174: 313-341.013922
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. J Neurosci Res, 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; Shohet R; 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-1 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. FASEB J, 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; FechterLD (1984). Prenatal exposure to carbon monoxide: learning andmemory 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 T J; 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
inhalation 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 Toxicol Environ Health, 35: 127-133.
001005
January 2010 5-137
-------
McGrath JJ; Schreck RM; Lee PS (1993). Carboxyhemoglobin levels in humans: Effects of altitude. Inhal Toxicol, 5: 241-
249.013865
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: 564-567. 193827
Medeiros A; Gouveia N (2005). Relacao entre baixo peso ao nascer e a poluicao do ar no Municipio de Sao Paulo
[Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil]. Rev Saude Publica, 39:
965-972. 089824
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 oxide. 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 Curr Iss,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. Retrieved 15-JUN-09, from http://www.merriam-
webster.com/medical/susceptible; http://www.merriam-webster.com/medical/vulnerable. 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
Min JY; Paek D; Cho SI; Min KB (2009). Exposure to environmental carbon monoxide may have a greater negative effect
on cardiac autonomic function in people with metabolic syndrome. Sci Total Environ, 407: 4807-4811. 199514
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
January 2010 5-138
-------
Morita T; Mitsialis SA; Koike H; Liu Y; Kourembanas S (1997). Carbon monoxide controls the proliferation of hypoxic
vascular smooth muscle cells. JBiol 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
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 HE; 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. JAllergy Clinlmmunol, 121: 1133-1139. 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 KG (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 JEpidemiol, 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). A review: 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; Baylerian MS (1984). Persistent cardiomegaly induced by carbon monoxide and associated
tachycardia. J Appl Physiol, 56: 1045-1052. 011567
January 2010 5-139
-------
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
Pereira Filho MA; 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
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 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
Pleis JR; Lucas JW (2009). Summary health statistics for US adults: National Health Interview Survey, 2007. National
Center for Health Statistics, Centers for Disease Control and Prevention, U.S. Department of Health and Human
Services . Hyattsville, Maryland. Series 10, Number 240; (PHS) 200971568.
http://www.cdc.gov/nchs/data/series/sr_l0/srl0_240.pdf. 202833
Pope CA III; 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 CA III; 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
January 2010 5-140
-------
Ren X; Dorrington KL; 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
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
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; Janssen NA; 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
January 2010 5-141
-------
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
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
Shaoqing Y; Ruxin Z; Yinjian C; Jianqiu C; Chunsheng Z; Jiangfeng T; Genhong L (2008). Possible contribution of
endogenous carbon monoxide to the development of allergic rhinitis in guinea pigs. J Inflamm, 5: 23. 192384
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
January 2010 5-142
-------
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
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
Sorhaug 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
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. J Neurosci 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. Dev Neurosci, 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 J Epidemiol, 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
January 2010 5-143
-------
Suliman HB; Carraway MS; Tatro LG; Piantadosi CA (2007). Anew 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
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. PNAS, 97: 1305-1310. 011574
Thorn SR; Garner S; Fisher D; Ischiropoulos H (1998). Vascular nitrosative stress from CO exposure. Undersea Hyperb
Med, 25:47. 016750
Thorn SR; Ischiropoulos H (1997). Mechanism of oxidative stress from low levels of carbon monoxide. Health Effects
Institute. Boston, MA. 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
Thomsen HK (1974). Carbon monoxide-induced atherosclerosis in primates: an electron-microscopic study on the
coronary arteries of Macaca trus monkeys. Atherosclerosis, 20: 233-240. 010704
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
January 2010 5-144
-------
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. JExpo 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
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 FJ; 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 13-MAY-01, from
http://www.census.gov/Press-Release/www/2000/cbOO-05.html. 157064
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. U.S. Envrionmental Protection Agency. Research Triangle Park, NC.
EPA-600/9-78-015. 086321
U. S. EPA (1991). Air quality criteria for carbon monoxide. U. S. Environmental Protection Agency. Research Triangle Park,
NC. EPA/600/8-90/045F. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000554R.txt. 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
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 (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=l56648. 192082
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). Susceptible subpopulations. Retrieved 15-MAY-09, from
http://www.epa.gov/nerl/goals/health/populations.html. 192149
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
January 2010 5-145
-------
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; Gleser MA (1972). Effect of carbon monoxide on oxygen transport during exercise. J Appl Physiol, 32: 234-239.
010898
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
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
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
January 2010 5-146
-------
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 YY; Rull RP; English P; Balmes J; Ritz B (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
Wollmann HA (1998). Intrauterine growth restriction: definition and etiology. Horm Res, 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: 110-115. 098386
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
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 Curr Iss, 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 Curr Iss, 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 Curr Iss, 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 Crit
Care Med, 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
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. Circulation, 110: 2184-2189.
087489
Zanobetti A; Schwartz J (2001). Are diabetics more susceptible to the health effects of airborne particles? Am J Respir Crit
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
January 2010 5-147
-------
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 P III; Benowitz NL (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 STATS 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. Science, 260: 1946-1950. 013905
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. FASEB J, 21: 1099-
1106. 193884
January 2010 5-148
-------
Annex A. Atmospheric Science
Cerbon Mcno^de Emissicns in 2002 (Tens per Square Mile)
] 0.06 — 0.33
I 0.39 - 101
I 131 - 55.53
Alaska
ON-ROAD VEHICLES
FIRES
FUEL COMB. OTHER
METALS PROCESSING
FUEL COMB. ELEC.UTIL.
PETROLEUM SRELATED INDUSTRIES
CHEMICAL SALLIED PRODUCT MFG
STORAGE STRANSPORT
SOLVENT UTILIZATION
| 145,340
| 88239
4,000,000 6,000,000 8,000,000 10,000,000
Emissions (tons)
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FUEL COMB. OTHER
METALS PROCESSING
OTHER INDUSTRIAL PROCESSES
PETROLEUM SRELATED INDUSTRIES
STORAGE STRANSPORT
Yukon-Koyukuk Census Area
00,000 6,000,000
Emissions (tons)
8,000,000 10,000,000
Figure A-1. CO emissions density map and distribution for the state of Alaska and for Yukon-
Koyukuk County in Alaska.
January 2010
A-1
-------
NON-ROAD EQUIPMENT
WASTE DISPOSAL SRECYCLING
METALS PROCESSING
OTHER INDUSTRIAL PROCESSES
CHEMICAL SALLIED PRODUCT MFG
MISCELLANEOUS
SOLVENT UTILIZATION
ON-ROAD VEHICLES
FUEL COMB. OTHER
FUEL COMB. INDUSTRIAL
UELCOMB.ELEC.UTIL.
RELATED INDUSTRIES
TORAGES TRANSPORT
MISCELLANEOUS
Ca/bon Mor :wdt En\ 9sic''is in 2QQI1 fTcos p-tr Square Mile)
Utah
1 159 - 5.90
I 6 97 - 16.34
I E.56 - 395.50
ON-ROAD VEHICLES
FUEL COMB. OTHER
FUEL COMB. INDUSTRIAL
UELCOMB.ELEC.UTIL.
RELATED INDUSTRIES
FORAGE STRAN SPORT
SOLVE NT UTILIZATION
Utah County
ON-ROAD VEHICLES
FUEL COMB. OTHER
FUEL COMB. INDUSTRIAL
FUEL COMB. ELEC.UTIL.
PETROLEUM & RELATED INDUSTRIES
STORAGE STRAN SPORT
SOLVE NT UTILIZATION
Weber County
25,000
Emissions (tons)
Grand County
50,000
Emissions (tons)
5,000 7,500
Emissions (tons)
Figure A-2. CO emissions density map and distribution for the state of Utah and for selected
counties in Utah.
January 2010
A-2
-------
ON-ROAD VEHICLES
FUEL COMB. OTHER
FUEL COMB. INDUSTRIAL
FUEL COMB. ELEC.UTIL.
PETROLEUM SRELATED INDUSTRIES
STORAGE STRANSPORT
SOLVENT UTILIZATION
ON-ROAD VEHICLES
FUEL COMB. OTHER
FUEL COMB. INDUSTRIAL
FUEL COMB. ELEC.UTIL.
PETROLEUM SRELATED INDUSTRIES
STORAGE STRANSPORT
MISCELLANEOUS
Carbon Mono^dcfe Ernisaons in 20D2 (Tens per Square Mile)
Massachusetts
] 4937 - 11727
9 T33.75 - 19&.S9
Middlesex County
FIRES
SOILS
FUEL COMB. OTHER
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM SRELATED INDUSTRIES
MISCELLANEOUS
SOLVENT UTILIZATION
0 25,000 50,000 75,000 100,000 125,000 150,000 175,000 200,000 225,000 250,000
Norfolk County
Suffolk County
FIRES
SOILS
FUEL COMB. OTHER
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
MISCELLANEOUS
SOLVE NT UTILIZATION
50,000 100,000 150,000
Emissions (tons)
25,000 50,000 75,000
Emissions (tons)
Figure A-3. CO emissions density map and distribution for the state of Massachusetts and
for selected counties in Massachusetts.
January 2010
A-3
-------
Carbcn Monoxide Bnissicns in 2002 (Tons per Squsre Mite)
1462 - £232
122.45 - 49.92
I SI 11- 803.87
FIRES
SOILS
FUEL COMB. OTHER
UELCOMB.ELEC.UTIL.
DUSTRIAL PROCESSES
MISCELLANEOUS
SOLVE NT UTILIZATION
Georgia
5UOJI03 75BJ100
Emissions (tons)
1,000,000 1,250,000
SOILS
FUEL COMB. OTHER
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
MISCELLANEOUS 28
SOLVENT UTILIZATION 0
Fulton County
100,000 158 pBO 200,000 250,000
Emissions (tons)
ON-ROAD VEHICLES
FUEL COMB. OTHER
WASTE DISPOSAL SRECYCLING
FUE COMB. INDUSTRIAL
PETROLEUM SR LATED INDUSTRIES
CHEMICAL SA IED PRODUCT MFG
STORAGE STRANSPORT
Dekalb County
Figure A-4. CO emissions density map and distribution for the state of Georgia and for
selected counties in Georgia (Figure 1 of 2).
January 2010
A-4
-------
Chatham County
Liberty County
SOILS
FUEL COMB. OTHER
WASTE DISPOSAL & RECYCLING
OTHEI
PETROLEUMS
CHEMICALS
STRIAL PROCESSES
ELATED INDUSTRIES
PRODUCT MFG
FUEL COMB. INDUSTRIAL
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM & RELATED INDUSTRIES
CHEMICAL & ALLIED PRODUCT MFG
STORAGE STRANSPORT
MISCELLANEOUS
SOLVE NT UTILIZATION
Glynn County
FUEL COMB. OTHER
WASTE DISPOSAL & RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM & RELATED INDUSTRIES
CHEMICAL SALLIED PRODUCT MFG
Figure A-5. CO emissions distribution for selected counties in Georgia (Figure 2 of 2).
January 2010
A-5
-------
_.__
Caibcn Monceode Emissions in 2002 (Tons per Square Mile)
0.85 — 17.20
17.29 - 54.04
34.72- -E46.52
California
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
STORAGE STRAN SPORT
MISCELLANEOUS
2,000,000
missions (tons)
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
CHEMICAL & ALLIED PRODUCT MFG
STORAGE STRAN SPORT
MISCELLANEOUS
SOLVE NT UTILIZATION
Los Angeles County
500,000
nissions (tons)
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
CHEMICAL SALLIED PRODUCT MFG
STORAGE STRAN SPORT
MISCELLANEOUS
San Diego County
200,000
Emissions (tons)
PE
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
SOILS
FUEL CO MB. OTHER
WASTE DISPOSAL & RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
ROLEUM & RELATED INDUSTRIES
CHEMICAL SALLIED PRODUCT MFG
MISCELLANEOUS
Orange County
Figure A-6. CO emissions density map and distribution for the state of California and for
selected counties in California.
January 2010
A-6
-------
Carbon Monoade Bnisacns in 2002 [Tons per Square Mite)
1 9.07 - 21.15
9 2164 - 3742
140.56 -30122
Alabama
PE
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
SOILS
FUEL CO MB. OTHER
WASTE DISPOSAL & RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
ROLEUM & RELATED INDUSTRIES
CHEMICAL SALLIED PRODUCT MFG
MISCELLANEOUS
Jefferson County
ON-ROAD VEHICLES
NON-ROAD EQUIPMENT
FIRES
SOILS
FUEL CO MB. OTHER
WASTE DISPOSAL & RECYCLING
FUEL COMB. INDUSTRIAL
METALS PROCESSING
FUEL COMB. ELEC.UTIL.
OTHER INDUSTRIAL PROCESSES
PETROLEUM & RELATED INDUSTRIES
CHEMICAL SALLIED PRODUCT MFG
MISCELLANEOUS
Figure A-7. CO emissions density map and distribution for the state of Alabama and for
Jefferson County in Alabama.
January 2010
A-7
-------
Table A-1. Listing of all CO 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
N/A
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
N/A
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
0.04000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.50000
0.04000
0.04000
0.04000
January 2010
A-8
-------
Table A-2. Microscale monitors meeting 75% completeness criteria, 2005-2007.
Monitor Code
02-090-0002-42101-1
04-013-0016-42101-1
04-019-1014-42101-1
06-065-1003-42101-1
06-073-0007-42101-1
08-013-0009-42101-1
08-031-0002-42101-2
08-031-0019-42101-1
08-041-0015-42101-1
08-077-0018-42101-1
09-003-0017-42101-1
11-001-023-42101-1
12-057-1070-42101-1
12-086-4002-42101-1
12-095-1005-42101-1
12-103-0024-42101-1
12-103-2008-42101-1
12-115-1004-42101-1
13-121-0099-42101-1
17-031-0063-42101-1
17-031-6004-42101-1
17-143-0036-42101-1
17-167-0008-42101-1
17-201-0011-42101-1
18-003-0011-42101-1
18-089-0015-42101-1
18-097-0072-42101-1
18-163-0019-42101-1
21-111-1019-42101-1
27-053-0954-42101-1
27-123-0050-42101-1
27-137-0018-42101-1
27-145-3048-42101-1
30-029-0010-42101-1
30-031-0013-42101-1
33-011-1009-42101-1
34-005-1001-42101-1
34-017-1002-42101-1
37-067-0023-42101-1
39-035-0048-42101-1
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
Kalis pell
Not in a city
Nashua
Burlington
Jersey City
Winston-Salem
Cleveland
Traffic Count
NR
50,000
41,200
40,000
6,000
20,000
17,200
500
44,200
13,525
10,000
30,000
133,855
5,000
30,000
35,000
67,751
31,000
44,000
5,000
NR
18,500
16,400
11,400
30430
NR
21 ,237
24,498
22,000
29,352
NR
12,000
NR
NR
2,000
40,000
8,000
25,000
22,000
24,300
Road Type
NR
ARTERIAL
MAJ ST OR HY
FREEWAY
THRUST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
THRUST OR HY
THRUSTORHY
THRUSTORHY
ARTERIAL
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
NR
ARTERIAL
MAJ ST OR HY
ARTERIAL
MAJ ST OR HY
NR
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
NR
MAJ ST OR HY
NR
THRUSTORHY
THRUSTORHY
MAJ ST OR HY
THRUSTORHY
THRUSTORHY
MAJ ST OR HY
THRUSTORHY
January 2010
A-9
-------
Monitor Code
State Name
City Name
Traffic Count
Road Type
39-035-0051-42101-1
39-035-0053-42101-1
39-049-0036-42101-1
39-061-0021-42101-1
39-085-0006-42101-1
39-113-0034-42101-1
39-153-0022-42101-1
41-029-0018-42101-1
41-039-0013-42101-1
41-051-0087-42101-1
45-079-0020-42101-1
47-037-0021-42101-1
47-157-0036-42101-1
48-029-0046-42101-1
48-201-0075-42101-1
53-033-0019-42101-1
53-063-0049-42101-1
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Oregon
Oregon
Oregon
South Carolina
Tennessee
Tennessee
Texas
Texas
Washington
Washington
Cleveland
Cleveland
Columbus
Cincinnati
Mentor
Dayton
Akron
Medford
Eugene
Portland
Columbia
Nashville
Memphis
San Antonio
Houston
Bellevue
Spokane
16,150
19,550
16,800
17,250
25,240
7,100
13,150
NR
17,500
4,150
31 ,500
15,000
25,000
5,820
6,576
100,000
10,000
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
THRUSTORHY
MAJ ST OR HY
NR
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
THRUSTORHY
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
MAJ ST OR HY
"NR" denotes that the value was not reported.
January 2010
A-10
-------
Table A-3. Middle scale monitors meeting 75% completeness criteria, 2005-2007.
Monitor Code
04-013-3010-42101-1
06-029-0010-42101-1
06-037-1301-42101-1
06-037-9033-42101-1
06-059-1003-42101-1
06-071-9004-42101-1
06-085-0005-42101-1
12-0011-0010-42101-1
12-031-0080-42101-1
12-031-0084-42101-1
12-099-1004-42101-1
12-103-2006-42101-1
17-031-3103-42101-1
20-209-0021-42101-1
24-510-0040-42101-1
32-031-0022-42101-1
34-003-0004-42101-1
36-061-0056-42101-1
39-049-0005-42101-1
39-081-1001-42101-1
39-151-0020-42101-1
40-143-0191-42101-1
42-003-0038-42101-1
42-101-0047-42101-1
45-019-0046-42101-1
45-045-0008-42101-1
45-045-0009-42101-1
47-163-0007-42101-1
48-439-1002-42101-1
50-007-0014-42101-1
72-127-0003-42101-1
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
Kings port
Fort Worth
Burlington
San Juan
Traffic Count
18,500
30,300
35,000
2,320
1,000
21 ,900
NR
1,000
1,000
500
30,000
23,400
47,900
7,720
15,300
NR
250,000
45,000
36,600
2,500
11,000
50,800
15,000
NR
NR
NR
9,500
NR
100
NR
64,000
Road Type
ARTERIAL
ARTERIAL
ARTERIAL
LOCAL STORMY
LOCAL STORMY
THRU STORMY
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
MAJ ST OR MY
MAJ ST OR MY
ARTERIAL
MAJ ST OR MY
THRU STORMY
NR
ARTERIAL
MAJ ST OR MY
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
MAJSTORHY
MAJSTORHY
"NR" denotes that the value was not reported.
January 2010
A-11
-------
Table A-4. Neighborhood scale monitors meeting 75% completeness criteria, 2005-2007.
Monitor Code
01-073-1003-42101-1
01-073-6004-42101-1
02-020-0018-42101-1
02-020-0048-42101-1
02-090-0020-42101-1
04-013-0019-42101-1
04-013-3002-42101-1
04-019-0002-42101-1
04-019-1011-42101-1
04-019-1028-42101-1
06-001-1001-42101-1
06-013-0002-42101-1
06-037-5005-42101-1
06-053-1003-42101-1
06-065-9001-42101-1
06-067-0007-42101-1
06-073-0001-42101-1
06-073-1002-42101-1
06-073-2007-42101-1
06-083-1025-42101-1
06-083-2004-42101-1
06-083-2011-42101-1
06-083-4003-42101-1
08-01-3001-42101-1
08-067-7001-42101-1
08-069-1004-42101-1
08-123-0010-42101-1
11-001-0041-42101-1
12-011-2004-42101-1
12-011-3002-42101-1
12-031-0083-42101-1
12-086-0031-42101-1
12-086-1019-42101-1
12-095-2002-42101-1
12-103-0018-42101-1
17-031-4002-42101-1
17-163-0010-42101-1
18-097-0073-42101-1
20-173-0010-42101-1
21-111-0046-42101-1
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
5,000
NR
NR
5,000
NR
NR
24,000
37,400
47,000
52,900
500
41,218
1,252
33,193
NR
20,000
5,000
NR
18,000
NR
NR
5,000
NR
500
2,436
5,000
6,650
540
1,000
1,000
10,000
62,000
8,000
7,000
2,000
NR
8,900
11,261
6,884
6,500
Road Type
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 STORMY
LOCAL STORMY
THRU STORMY
NR
THRU STORMY
LOCAL STORMY
NR
LOCAL ST OR HY
NR
NR
THRUST OR HY
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
MAJSTORHY
NR
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
THRUSTORHY
January 2010
A-12
-------
Monitor Code
22-033-0009-42101-1
25-013-0016-42101-1
25-017-0007-42101-1
25-025-0042-42101-1
27-03-0600-42101-1
27-037-0020-42101-1
27-037-0423-42101-1
29-510-0086-42101-1
30-111-0085-42101-1
31-055-0035-42101-1
32-003-0538-42101-1
32-003-0539-42101-1
32-003-0561-42101-1
32-003-1021-42101-1
32-003-2002-42101-1
32-031-0016-42101-1
32-031-0020-42101-1
32-031-0025-42101-1
32-031-1005-42101-1
32-031-2009-42101-1
32-510-0004-42101-1
33-011-0020-42101-1
34-003-5001-42101-1
34-007-0003-42101-1
35-001-019-42101-1
35-001-0023-42101-1
35-001-0024-42101-1
35-001-0028-42101-1
35-001-1014-42101-1
35-043-9004-42101-1
36-063-2008-42101-1
37-119-0041-42101-1
37-119-0041-42101-3
39-035-0070-42101-1
39-113-0028-42101-1
39-153-0020-42101-1
40-021-9002-42101-1
40-071-9010-42101-1
40-109-0047-42101-1
41-051-0080-42101-1
42-003-0031-42101-1
42-013-0801-42101-1
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
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
Traffic Count
5,000
5,000
15,000
12,785
1,400
NR
NR
81 ,850
5,700
2,900
20,000
21 ,000
28,400
NR
6,750
22,700
NR
NR
2,600
NR
1
500
15,000
45,000
1
41 ,200
15,500
2,0600
8,000
100
5,000
16,400
16,400
100
5,100
200
10,300
300
27,000
5,000
4,562
100
Road Type
LOCALS! OR HY
LOCALS! OR HY
!HRUS!ORHY
LOCALS! OR HY
LOCALS! OR HY
NR
NR
MAJS!ORHY
!HRUS!ORHY
LOCALS! OR HY
LOCALS! OR HY
MAJ S! OR HY
MAJ S! OR HY
NR
!HRUS!ORHY
LOCALS! OR HY
NR
NR
LOCALS! OR HY
NR
LOCALS! OR HY
LOCALS! OR HY
!HRUS!ORHY
MAJ S! OR HY
AR!ERIAL
MAJ S! OR HY
MAJ S! OR HY
!HRUS!ORHY
!HRUS!ORHY
LOCAL S! OR HY
LOCALS! OR HY
MAJ S! OR HY
MAJ S! OR HY
LOCALS! OR HY
LOCALS! OR HY
LOCALS! OR HY
LOCALS! OR HY
LOCALS! OR HY
MAJS!ORHY
LOCALS! OR HY
!HRUS!ORHY
LOCALS! OR HY
January 2010
A-13
-------
Monitor Code
42-017-0012-42101-1
42-021-0011-42101-1
42-049-0003-42101-1
42-071-0007-42101-1
42-073-0015-42101-1
42-091-0013-42101-1
42-095-0025-42101-1
42-101-0004-42101-1
42-101-0027-42101-1
42-107-0003-42101-1
42-125-0005-42101-1
44-007-1010-42101-1
48-061-0006-42101-1
48-113-0069-42101-2
48-141-0002-42101-1
48-141-0029-42101-1
48-141-0037-42101-1
48-141-0044-42101-1
48-141-0053-42101-1
48-141-0057-42101-1
48-141-0058-42101-1
48-201-0024-42101-1
48-201-0047-42101-1
48-201-1035-42101-1
48-201-1039-42101-1
48-439-3011-42101-1
48-453-0014-42101-1
48-479-0017-42101-1
49-035-0003-42101-1
50-021-0002-42101-1
51-059-0005-42101-1
51-650-0004-42101-2
51-760-0024-42101-1
51-770-0015-42101-1
54-009-0011-42101-1
54-029-0009-42101-1
54-029-1004-42101-1
State Name
Pennsylvania
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
Bristol
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
500
6,000
1,000
2,000
4,500
8,500
100
13800
46000
100
NR
100,000
30
1,000
7,270
2,790
5,000
15,200
1,992
500
1,080
5,300
5,860
13,440
16010
10,573
3,420
30,380
16,500
NR
25
2,000
7,591
NR
NR
NR
50
Road Type
LOCAL ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
THRUSTORHY
LOCAL ST OR HY
MAJSTORHY
LOCAL ST OR HY
MAJSTORHY
MAJSTORHY
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
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
MAJ ST OR HY
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
"NR" denotes that the value was not reported.
January 2010
A-14
-------
Table A-5. Urban scale monitors meeting 75% completeness criteria, 2005-2007.
Monitor Code
06-059-0007-42101-1
13-089-0002-42101-1
13-223-0003-42101-1
25-027-0023-42101-1
34-007-1001-42101-1
42-003-0010-42101-1
42-007-0014-42101-1
42-129-0008-42101-1
42-133-0008-42101-1
48-141-0055-42101-1
51-059-0030-42101-1
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
1,000
9,250
6
NR
4,000
1,000
NR
100
8,400
2450
200
Road Type
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
THRU ST OR MY
MAJ STORMY
NR
THRU ST OR MY
THRU ST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
"NR" denotes that the value was not reported.
Table A-6. Regional scale monitors meeting 75% completeness criteria, 2005-2007.
Monitor Code
State Name
City Name
Traffic Count
Road Type
23-009-0103-42101-1
Maine
Not in a city
3,500
LOCAL ST OR HY
35-001-0029-42101-1
New Mexico
South Valley
8,800
LOCAL ST OR HY
"NR" denotes that the value was not reported.
January 2010
A-15
-------
Table A-7. Monitors meeting 75% completeness criteria, 2005-2007 with no scale delared.
Monitor Code
04-013-9997-42101-1
06-001-0007-42101-1
06-007-0002-42101-1
06-013-1002-42101-1
06-013-1004-42101-1
06-013-3001-42101-1
06-019-0007-42101-1
06-019-0008-42101-1
06-019-0242-42101-1
06-019-5001-42101-1
06-025-0005-42101-1
06-025-0006-42101-1
06-025-1003-42101-1
06-037-0002-42101-1
06-037-0113-42101-1
06-037-1002-42101-1
06-037-1103-42101-1
06-037-1201-42101-1
06-037-1701-42101-1
06-037-2005-42101-1
06-037-4002-42101-1
06-037-6012-42101-1
06-041-0001-42101-1
06-045-0008-42101-1
06-045-0009-42101-1
06-055-0003-42101-1
06-059-2022-42101-1
06-059-5001-42101-1
06-065-5001-42101-1
06-065-8001-42101-1
06-067-0002-42101-1
06-067-0006-42101-1
06-067-0013-42101-1
06-071-0001-42101-1
06-071-0306-42101-1
06-071-1004-42101-1
06-075-0005-42101-1
06-077-1002-42101-1
06-081-1001-42101-1
06-087-0003-42101-1
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
2,400
44,000
NR
NR
9,600
500
20,000
500
16,461
7,000
10
NR
600
NR
2,400
9,000
NR
NR
18,000
24,000
4,395
15,000
12,000
18,000
NR
42,400
NR
NR
18,000
NR
10,000
100
NR
454
15,000
240,700
6,000
1,000
NR
Road Type
LOCAL STORMY
LOCAL STORMY
LOCAL STORMY
NR
THRU STORMY
THRU ST OR MY
LOCAL ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
THRUST OR HY
LOCAL ST OR HY
THRUST OR HY
NR
THRUST OR HY
NR
LOCAL ST OR HY
THRU ST OR HY
NR
NR
THRUST OR HY
LOCAL ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
LOCAL ST OR HY
MAJ ST OR HY
NR
MAJSTORHY
NR
NR
THRU ST OR HY
NR
LOCAL ST OR HY
LOCAL ST OR HY
NR
LOCAL ST OR HY
THRUST OR HY
FREEWAY
LOCAL ST OR HY
LOCAL ST OR HY
NR
January 2010
A-16
-------
Monitor Code
06-095-0004-42101-1
06-097-0003-42101-1
06-099-0005-42101-1
06-099-0006-42101-1
09-003-1003-42101-1
10-003-1008-42101-1
10-003-2004-42101-1
15-003-0010-42101-1
18-063-0002-42101-1
25-025-0002-42101-1
29-077-0032-42101-1
29-189-0004-42101-1
30-013-0001-42101-1
31-109-0018-42101-1
34-023-2003-42101-1
34-025-2001-42101-1
34-027-0003-42101-1
36-001-0012-42101-1
36-029-0005-42101-1
36-055-1007-42101-1
36-067-0017-42101-1
36-081-0124-42101-1
36-093-0003-42101-1
36-103-0009-42101-2
48-479-0016-42101-1
49-057-0006-42101-1
51-013-0020-42101-1
51-059-1005-42101-1
51-059-5001-42101-1
51-510-0009-42101-1
56-039-1012-42101-1
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
9,350
2,608
NR
500
800
NR
28,046
NR
500
35,000
1,000
33,300
26,155
NR
14,000
NR
NR
12,000
26,000
NR
NR
10,000
37,000
10,000
16,180
38,000
6,000
24,000
36,845
3,974
NR
Road Type
THRUST OR HY
THRUST OR HY
NR
LOCAL ST OR HY
LOCAL ST OR HY
NR
MAJSTORHY
NR
LOCAL ST OR HY
MAJSTORHY
LOCAL ST OR HY
MAJSTORHY
MAJ ST OR HY
NR
LOCAL ST OR HY
NR
NR
MAJSTORHY
ARTERIAL
NR
NR
EXPRESSWAY
EXPRESSWAY
THRUST OR HY
MAJSTORHY
ARTERIAL
MAJSTORHY
MAJSTORHY
MAJSTORHY
LOCAL ST OR HY
NR
"NR" denotes that the value was not reported.
January 2010
A-17
-------
Table A-8.
State
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
criteria for 2005-2007.
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
state 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
January 2010 A-18
-------
State Number of high LOD monitors Number of trace-level monitors
Rhode Island 1 0
South Carolina 3 1
South Dakota 0 0
Tennessee 3 0
Texas 19 2
Utah 2 0
Vermont 2 0
Virginia 9 0
Washington 2 0
West Virginia 3 0
Wisconsin 0 0
Wyoming 1 0
January 2010 A-19
-------
0 30 60 120 180 240
Figure A-8. Map of CO monitor locations with respect to population density in the Anchorage
CBSA, total population.
0 5 10 20 30 40
2005 Population Density
Population > 65 per 2.6 Sq Km
0 30 60 120 180 240
Figure A-9. Map of CO monitor locations with respect to population density in the Anchorage
CBSA, ages 65 yr and older.
January 2010
A-20
-------
0510 20 30 40
2006 Population Density
| | Atlanta CO Monitors (5 km buffer)
Population per 2.6 Sq Km
^H 0 - 230
231 - 459
4591 -11475
I 11476-45900
0 15 30 60 90 120
Figure A-10. Map of CO monitor locations with respect to population density in the Atlanta
CSA, total population.
| | Atlanta CO Monitors (5 km buffer;
Population > 65 per 2.6 Sq Km
Figure A-11. Map of CO monitor locations with respect to population density in the Atlanta
CSA, ages 65 yr and older.
January 2010
A-21
-------
2005 Population Density
I 1 Boston CO Monitors (5 km buffer)
Population per 2.6 Sq Km
Figure A-12. Map of CO monitor locations with respect to population density in the Boston
CSA, total population.
2005 Population Density
Population > 65 per 2.6 Sq Km
•• .;. -122
Figure A-13. Map of CO monitor locations with respect to population density in the Boston
CSA, ages 65 yr and older.
January 2010
A-22
-------
0 25 50 100 150 200
Figure A-14. Map of CO monitor locations with respect to population density in the Houston
CSA, total population.
2005 Population Density
Population > 65 per 2.6 Sq Km
0 25 50 100 150 200
Figure A-15. Map of CO monitor locations with respect to population density in the Houston
CSA, ages 65 yr and older.
January 2010
A-23
-------
0 20 40 80 120
Figure A-16. Map of CO monitor locations with respect to population density in the New York
City CSA, total population.
2005 Population Density
Population > 65 per 2.6 Sq Km
0 2040 BO 120 1
Figure A-17. Map of CO monitor locations with respect to population density in the New York
City CSA, ages 65 yr and older.
January 2010
A-24
-------
Figure A-18. Map of CO monitor locations with respect to population density in the Phoenix
CSA, total population.
Figure A-19. Map of CO monitor locations with respect to population density in the Phoenix
CSA, ages 65 yr and older.
January 2010
A-25
-------
Figure A-20. Map of CO monitor locations with respect to population density in the Pittsburgh
CSA, total population.
Figure A-21. Map of CO monitor locations with respect to population density in the Pittsburgh
CSA, ages 65 yr and older.
January 2010
A-26
-------
0 1530 60 90 120
Figure A-22. Map of CO monitor locations with respect to population density in the Seattle
CSA, total population.
Figure A-23. Map of CO monitor locations with respect to population density in the Seattle
CSA, ages 65 yr and older.
January 2010
A-27
-------
0 1530 60 90 120
Figure A-24. Map of CO monitor locations with respect to population density in the St. Louis
CSA, total population.
2005 Population Density
Population > 65 per 2.6 Sq Km
Figure A-25. Map of CO monitor locations with respect to population density in the St. Louis
CSA, ages 65 yr and older.
January 2010
A-28
-------
Anchorage Core Based Statistical Area
Ql
r
Anchorage CBSA
CO Monitors
Major Highways
0 15 30 60 90 120
I Kilometers
Figure A-26. Map of CO monitor locations with AQS Site IDs for Anchorage, AK.
January 2010
A-29
-------
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
o
.£
O
1.00
0.0
0.00
Legend
P90
COD
0.73
1.1
0.32
9.0
1.00
0.0
0.00
January 2010
A-30
-------
A B
Site ID 02-020-0018 02-020-0048
Mean 1.04 1.10
SD 0.94 1.04
Obs 12969 12703
4-
3-
'E
CL
Q.
0
"ro 2 -
c:
Q)
O
O
O
1 -
o-
I
I I I
1234 1234
season
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.
January 2010
A-31
-------
Atlanta Combined Statistical Area
Ql
r
Atlanta CSA
CO Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
I Kilometers
Figure A-28. Map of CO monitor locations with AQS Site IDs for Atlanta, GA.
January 2010
A-32
-------
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
A
A
o
in
1.00
0.0
0.00
0
Urban
B
0.60
0.5
0.27
22.5
C
0.10
0.7
0.38
61.7
1.00
0.0
0.00
Legend
P90
COD
0.12
0.7
0.37
74.7
1.00
0.0
0.00
January 2010
A-33
-------
Site ID
Mean
SD
Obs
13-121-
0099
0.58
0.30
25440
13-089-
0002
0.53
0.35
25531
13-223-
0003
0.30
0.13
25712
1.7-
1.6-
1.5-
P 1-4-
^ 123"
Q_ 1.2-
^ 1.1 -
C 1.0-
g 0.9-
05 0.8-
^ 0.7-
0 0.6-
0 0.5-
0 0.3-
0.2-
0.1 -
0.0-
ill
f !
1 1 i
1234 1234 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.
January 2010
A-34
-------
Boston Combined Statistical Area
01
r
Boston CSA
CO Monitors
Interstate Highways
Major Highways
0 12.5 25 50 75
100
M Kilometers
Figure A-30. Map of CO monitor locations with AQS Site IDs for Boston, MA.
January 2010
A-35
-------
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
A 1.00
o 0.0
ii o.oo
0
Neighborhood
B C
0.50 0.38
0.6 0.6
0.44 0.46
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
•a 0.00
1 o
1 D
'o
z
E
Legend
r
P90
COD
d
F
c
C5
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
5 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
January 2010
A-36
-------
Site ID
33-011-
1009
25-017- 25-025-
0007 0042
33-011-
0020
44-007-
1010
25-027- 25-025-
0023 0002
Mean
0.60
0.33
0.36
0.45
0.34
0.53
0.26
SD
0.37
0.22
0.26
0.27
0.22
0.23
0.24
Obs
25869
24362
24260
25197
23707
24446
24134
1.9-
1.8-
1.7-
^ 1.6-
E 1-5-
Q. 1.4-
S 13~
~ 10-
05 0.9-
-b 0.8-
^~
f \ O.o ~
C 0.5-
O 0.4-
0 0.3-
0.2-
0.1 -
0.0-
I
1
,i
1
li
1
?
I I
ii i ii i i i i MI MI i i i i MI
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.
January 2010
A-37
-------
Houston Combined Statistical Area
Houston CSA
CO Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
Kilometers
Figure A-32. Map of CO monitor locations with AQS Site IDs for Houston, TX.
January 2010
A-38
-------
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
A
1.00
A
o
H
0.0
0.00
0.0
Neighborhood
B
0.45
0.4
0.47
16.7
1.00
C
0.56
0.4
0.47
16.3
0.72
D
0.53
0.5
0.74
9.3
0.56
E
0.43
0.4
0.47
23.5
0.68
0.0
0.3
0.5
0.3
0.00
0.29
0.73
0.24
0.0
17.5
19.8
32.2
1.00
0.65
0.63
0.0
0.5
0.4
0.00
0.73
0.29
o
o
o.o
25.2
39.7
1.00
0.57
0.0
0.4
0.00
0.72
E
Legend
r
P90
COD
d
0.0 14.5
1.00
0.0
0.00
0.0
January 2010
A-39
-------
Site ID
Mean
SD
Obs
48-201-
0075
0.35
0.26
24922
48-201-
0024
0.42
0.27
23997
48-201-
0047
0.39
0.33
25241
48-201-
1035
0.14
0.22
25285
48-201-
1039
0.33
0.16
24480
1.6-
1.5-
1.4-
~ 1.3"
L 1.2-
1. 1.1 -
-X
5 0.9-
S °-8-
l 0.7-
= 0.6-
j 0.5-
= 0.4-
\ 0. 3 ~
0.2-
0.1 -
0.0-
1234 1234 1234 1234 1234
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.
January 2010
A-40
-------
New York Combined Statistical Area
New York CSA
CO Monitors
Interstate Highways
Major Highways
0 15 30 60 90
120
• Kilometers
Figure A-34. Map of CO monitor locations with AQS Site IDs for New York City, NY.
January 2010
A-41
-------
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
A
A 1.00
o 0.0
ii o.oo
0
B
Middle
B C
0.65 0.52
0.7 0.7
0.28 0.24
15.9 8.9
1.00 0.56
0.0 0.4
0.00 0.23
| 0 10.5
s c
1.00
0.0
0.00
0
"§ D
.£
0
JJ
.£
O
z
Neighborhood
D
0.64
0.8
0.29
16.8
0.58
0.4
0.22
7.0
0.54
0.4
0.23
15.0
1.00
0.0
0.00
0
Null
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
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
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
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
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
E 1.00 0.50 0.57
0.0 0.4 0.4
0.00 0.24 0.23
0 27.5 36.7
F 1.00 0.47
0.0 0.4
0.00 0.23
0 61.9
G Le9end 1.00
= r 0.0
z P90 0.00
COD 0
H d
I
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
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
0
January 2010
A-42
-------
Site ID
Mean
SD
Obs
34-im-
1002
0.85
0.43
25646
34-UU3-
0004
0.55
0.27
23113
3B-UB1-
0056
0.62
0.21
25547
34-UU3-
5001
0.52
0.30
25150
64-UZ6-
2003
0.48
0.27
25028
34-IE&-
2001
0.50
0.24
25727
64-U2I-
0003
0.49
0.25
25691
3B-U81-
0124
0.47
0.23
25022
3B-1U3-
0009
0.12
0.17
25749
2.0-
1.9-
1.8-
-^ 1.7-
C 1.6-
Q_ 1-5-
Q. 1-4-
^^ 1.3-
C 1.2-
O 1.1-
5 ag-
^ 0.8-
0 0.7-
O 0.6-
C 0.5-
0 0.4-
W 0.3-
0.2-
0.1-
0.0-
i i
1234
i i
1234
i i
1234
i i
1234
i i i i
1234
i i i i
1234
i i
1234
i i i
1234
i
rri
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.
January 2010
A-43
-------
Phoenix Core Based Statistical Area
Phoenix CBSA
• CO Monitors
Interstate Highways
Major Highways
0 15 30
60
90
120
M Kilometers
Figure A-36. Map of CO monitor locations with AQS Site IDs for Phoenix, AZ.
January 2010
A-44
-------
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
A
A
o
in
1.00
0.0
0.00
0.0
B
Middle
B
0.86
0.8
0.39
3.9
1.00
J2 0.0
ii o.oo
0.0
c
•g
o
.£
"§> D
z
E
Neighborhood
C
0.89
0.7
0.37
1.6
0.88
0.6
0.34
3.4
1.00
0.0
0.00
Legend
r
P90
COD
d
0.0
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
z 0.00
0.0
January 2010
A-45
-------
Site ID
Mean
SD
Obs
04-013-
0016
0.93
0.95
25382
04-013-
3010
0.76
0.72
25414
04-013-
0019
0.84
0.88
25589
04-013-
3002
0.58
0.64
25657
04-013-
9997
0.79
0.64
25435
5-
<•— s
E 4-
Q_
Q_
• — •
C 3-
O
4— •
e 2-
0
O
c
O -I.
O '
0-
1
1
1
1
'
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.
January 2010
A-46
-------
Pittsburgh Combined Statistical Area
Pittsburgh CSA
CO Monitors
Interstate Highways
Major Highways
0 10 20 40 60
80
I Kilometers
Figure A-38. Map of CO monitor locations with AQS Site IDs for Pittsburgh, PA.
January 2010
A-47
-------
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
1 °-°
ii o.oo
0
Neighborhood
B C
0.25 0.39
0.7 0.6
0.65 0.51
33.3 68.2
B 1 .00 0.26
0.0 0.5
0.00 0.68
0 101.0
•g c 1.00
€ 0.0
0
f °-°°
£ 0
D
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
F
c
£
5
G
Legend
r
P90
COD
d
Urban
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
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
0.00
0
January 2010
A-48
-------
SitelD
Mean
SD
Obs
42-003-
0038
0.47
0.33
25818
42-125-
0005
0.21
0.23
25319
42-073-
0015
0.32
0.26
25745
42-003-
0031
0.32
0.26
25936
42-007-
0014
0.28
0.27
25500
42-129-
0008
0.07
0.15
25785
42-003-
0010
0.28
0.32
25655
1.4-
1.3-
^1.2-
: 1.1 -
^1.0-
^ 0.9-
5 0.8-
•
J 07-
5
3 0.6-
5 0.5-
2 0.4-
5 0.3-
5 0.2-
0.1 -
0.0-
1 1
hi i i i
i i i i
i i i i
n i i i
i
n
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.
January 2010
A-49
-------
Seattle Combined Statistical Area
Seattle CSA
CO Monitors
Interstate Highways
Major Highways
0 15 30
60
90
120
• Kilometers
Figure A-40. Map of CO monitor locations with AQS Site IDs for Seattle, WA.
January 2010
A-50
-------
A
Site ID
Mean
SD
Obs
53-033-0019
0.75
0.49
25818
3-
? 2-
Q.
Q.
0
TO
concent
o-
1234
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.
January 2010
A-51
-------
St Louis Combined Statistical Area
St Louis CSA
CO Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
Kilometers
Figure A-42. Map of CO monitor locations with AQS Site IDs for St. Louis, MO.
January 2010
A-52
-------
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.
§
.£
O
1.00
0.0
0.00
Neighborhood
0.60
0.3
0.24
9.5
1.00
0.0
0.00
Legend
P90
COD
Null
0.19
0.5
0.40
21.2
0.19
0.5
0.42
19.8
1.00
0.0
0.00
January 2010
A-53
-------
Site ID
Mean
SD
Obs
A
17-163-
0010
0.44
0.25
25325
B
29-510-
0086
0.42
0.29
25938
c
29-189-0004
0.43
0.25
25879
1.3-
1.2-
^ 1.1 -
^ 1.0-
f^
Q_
^ 0.8-
.2 °7"
^3 0.5-
S °-4-
0 0.3-
O 0.2-
0 0.1.
0.0-
-0.1-
1 I I I
1234
i
1 234
i
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.
January 2010
A-54
-------
TableA-17.
Comparison of distributional data at different monitoring
max, 24-h avg, 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
25,440 0.6
51,243 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
0.4
0.7
0.5
1.0
0.7
1.2
1.0
1-H DAILY MAX
Microscale
Urban Scale
1,075 1.0
2,154 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
0.8
1.2
0.9
1.6
1.3
1.9
1.5
1-H DAILY AVG
Microscale
Urban Scale
1,075 0.6
2,154 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
0.5
0.7
0.5
0.8
0.7
1.0
0.9
8-H DAILY MAX
Microscale
Urban Scale
1,075 0.8
2,154 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
0.6
0.9
0.7
1.2
1.0
1.3
1.3
TableA-18.
Comparison of distributional data at different monitoring
max, 24-h avg, 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
25,869 0.6
97,526 0.4
24,446 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
0.4
0.6
0.7
0.5
0.6
1.0
0.6
0.8
1.2
0.8
0.9
1-H DAILY MAX
Microscale
Neighborhood Scale
Urban Scale
1,080 1.2
4,212 0.6
1,086 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
0.7
0.9
1.4
0.7
1.0
2.0
1.1
1.2
2.5
1.4
1.4
1-H DAILY AVG
Microscale
Neighborhood Scale
Urban Scale
1,080 0.6
4,212 0.4
1,086 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
0.4
0.6
0.7
0.5
0.6
0.9
0.6
0.7
1.1
0.7
0.8
8-H DAILY MAX
Microscale
Neighborhood Scale
Urban Scale
1,080 0.8
4,212 0.5
1,086 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
0.6
0.8
1.0
0.6
0.8
1.4
0.8
1.0
1.7
1.0
1.1
January 2010
A-55
-------
TableA-19.
Comparison of distributional data at different monitoring
max, 24-h avg, and 8-h daily max data for Denver, CO.
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
77,070 0.5
51,968 0.5
0.0
0.0
0.0 0.1
0.0 0.1
0.1
0.2
0.3
0.3
0.3
0.3
0.4
0.4
0.6
0.6
0.7
0.6
1.0
1.0
1.3
1.3
1-H DAILY MAX
Microscale
Neighborhood Scale
3,190 1.2
2,173 1.1
0.1
0.1
0.3 0.4
0.2 0.3
0.5
0.4
0.7
0.6
0.8
0.6
1.0
0.9
1.4
1.3
1.5
1.5
2.2
2.1
2.7
2.6
1-H DAILY AVG
Microscale
Neighborhood Scale
3,190 0.5
2,173 0.5
0.0
0.0
0.1 0.2
0.1 0.2
0.2
0.3
0.3
0.3
0.4
0.4
0.5
0.5
0.6
0.6
0.6
0.6
0.9
0.9
1.0
1.1
8-H DAILY MAX
Microscale
Neighborhood Scale
3,190 0.8
2,173 0.8
0.3
0.3
0.3 0.3
0.3 0.3
0.4
0.3
0.5
0.4
0.5
0.5
0.7
0.7
0.9
0.9
1.0
1.0
1.4
1.5
1.8
1.8
Table A-20.
Comparison of distributional data at different monitoring
max, 24-h avg, and 8-h daily max data for Houston, TX.
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
24,922 0.3
99,003 0.3
0.0
0.0
0.0 0.0
0.0 0.0
0.0
0.0
0.2
0.2
0.2
0.2
0.3
0.3
0.4
0.4
0.5
0.4
0.6
0.6
0.8
0.8
1-H DAILY MAX
Microscale
Neighborhood Scale
1,043 0.7
4,145 0.7
0.0
0.0
0.0 0.2
0.0 0.1
0.3
0.2
0.4
0.4
0.5
0.4
0.6
0.5
0.8
0.8
0.9
0.8
1.2
1.3
1.4
1.7
1-H DAILY AVG
Microscale
Neighborhood Scale
1,043 0.3
4,145 0.3
0.0
0.0
0.0 0.0
0.0 0.0
0.1
0.1
0.2
0.2
0.3
0.2
0.4
0.3
0.5
0.4
0.5
0.4
0.6
0.5
0.6
0.6
8-H DAILY MAX
Microscale
Neighborhood Scale
1,043 0.5
4,145 0.5
0.3
0.3
0.3 0.3
0.3 0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.4
0.6
0.5
0.6
0.6
0.8
0.9
1.0
1.1
January 2010
A-56
-------
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
24,885
98,564
49,757
24,264
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
1,080
4,299
2,164
1,053
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
1,080
4,299
2,164
1,053
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
1,080
4,299
2,164
1,053
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
January 2010
A-57
-------
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 City, NY.
PERCENTILES
Time Scale
Mean Min
10
25
50
75
90
95
99
ALL HOURLY
Microscale
25,646 0.8
0.0
0.2
0.3
0.4
0.5
0.6
0.8
1.0
1.1
1.4
Middle Scale
48,660 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
25,150 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
1,077 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
2,053 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
1,053 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
1,077 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
2,053 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
1,053 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
1,077 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
2,053 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
1,053 0.7
0.3
0.3
0.3
0.3
0.4
0.5
0.6
0.8
0.8
1.2
1.5
January 2010
A-58
-------
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 Min
10
25
50
75
90
95
99
ALL HOURLY
Microscale
25,382 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
25,414 0.8
0.0
0.0
0.1
0.1
0.3
0.3
0.5
0.9
1.0
2.3
Neighborhood Scale
51,246 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
1,063 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
1,066 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
2,156 1.8
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
1,063 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
1,066 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
2,156 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
1,063 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
1,066 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
2,156 1.2
0.3
0.3
0.3
0.3
0.5
0.6
0.9
1.5
1.7
2.5
3.0
January 2010
A-59
-------
Table A-24. 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.
PERCENTILES
Time Scale N
Mean
Min
1
5 10
25
50
75
90
95
99
max
ALL HOURLY
Middle Scale 25,818
Neighborhood Scale 77,000
Urban Scale 76,940
0.5
0.3
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.1 0.1
0.0 0.0
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
0.6
0.6
1.1
0.8
0.8
1-H DAILY MAX
Middle Scale 1,079
Neighborhood Scale 3,210
Urban Scale 3,208
0.9
0.6
0.4
0.0
0.0
0.0
0.2
0.0
0.0
0.4 0.4
0.1 0.2
0.0 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
1.1
1.0
1.9
1.3
1.2
1-H DAILY AVG
Middle Scale 1,079
Neighborhood Scale 3,210
Urban Scale 3,208
0.5
0.3
0.2
0.0
0.0
0.0
0.1
0.0
0.0
0.2 0.2
0.0 0.0
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
0.6
0.6
0.9
0.7
0.7
8-H DAILY MAX
Middle Scale 1,079
Neighborhood Scale 3,210
Urban Scale 3,208
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
0.8
0.8
1.3
1.0
1.0
Table A-25. Comparison of distributional data for hourly, 1-h daily max, 24-h avg, and 8-h daily
data for Seattle, WA. Microscale was the only scale at which monitoring was
performed in Seattle, WA.
max
PERCENTILES
Time Scale N
Mean
Min
1
5 10
25
50
75
90
95
99
max
ALL HOURLY
Microscale 25,818
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 1,079
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 1,079
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 1,079
1.1
0.3
0.3
0.4 0.5
0.7
0.8
1.0
1.3
1.4
1.8
2.2
January 2010
A-60
-------
Table A-26. Comparison of distributional data for hourly, 1-h daily max, 24-h avg, and 8-h daily max
data for St. Louis, MO. Neighborhood scale was the only scale at which monitoring
was performed in St. Louis, MO.
PERCENTILES
Time Scale N Mean Min 1 5 10 25 50 75 90 95 99 max
ALL HOURLY
Neighborhood Scale 51,263 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 2,138 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 2,138 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 2,138 0.6 0.3 0.3 0.3 0.3 0.3 0.3 0.5 0.6 0.7 1.0 1.3
January 2010 A-61
-------
Winter
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
O
o
o
(SO
(3D
(®
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Spring
1 •
2 •
3 •
4
5 •
6 •
7 •
O
o
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
1.0
Fall
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
O O
0
ooo
o
GO
-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-44. Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) 0s, (4) PMio, and (5) PM2.s concentrations for Anchorage, AK. 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. Note that the data are not obtained for Anchorage
during the summer, and so are not presented here.
January 2010
A-62
-------
Winter
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
oo
o o
o
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.
January 2010
A-63
-------
Winter
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
o
OGDO
O30
O O O
OO
(JOi)
cm
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
00
OCED
(Q)
OO
OOO
OOGD
(OS)
CUBE)
-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 •
OD
C
O
m o
o
-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-46. Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) Os, (4) PMio, and (5) PM2.6 concentrations for Boston, MA. 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.
January 2010
A-64
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
(O
(giro
-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 •
GDC
QEED
GO
)
O O
O O <3DO
(((OKHBB)')
flBBBl
-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 -
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
OSD
O OO
O (3D)
O CUD
GO
GD©
(SStS)
om
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
(
O GO
O <2®
XO CXJO)
OO O
(OBEI
-------
Winter
Spring
1 •
2 •
3 •
4 •
5 •
6 •
7 •
8 •
O
o
o
(D
<0>
<{®2)
O
dSJ>
-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-49. Seasonal plots of correlations between hourly CO concentration with hourly (1)
S02, (2) N02, (3) 03, (4) PM,o, and (5) PM2.5 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.
January 2010
A-67
-------
Winter
Spring
-20 -10
o
las (h)
10 20
0
ag(h)
[FT"""
20 -10 0
ag(
Fa
20 -10 0
ag(
flf
10 20
h)
I
10 20
h)
Figure A-50. Cross-correlation functions for each season combined across sites where CO
and 0s monitors were co-located in Atlanta, Boston, Denver, Los Angeles, New
York City, and Phoenix.
January 2010
A-68
-------
Annex B. Dos i me try Studies
Table B-1. Recent studies related to CO dosimetry and pharmacokinetics.
Reference
Purpose
Findings
Aberg et al. (2009,1940821
To investigate CO concentrations in blood donors in
Sweden.
The mean CO concentration in 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
(QCP) 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, 0010261
To determine if low concentrations of CO would affect
exercise performance and myocardial perfusion in
young healthy men.
Men with COHb levels between 4 and 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 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 nonmuscle
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 intratissue 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 nonmuscle tissue. CO uptake from blood
by muscle is much slower than 02, thus COHb% will fall rapidly while
COMb% could remain high.
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).
January 2010
B-1
-------
Reference
Purpose
Findings
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.
Castillo et al. (2006,1932341
To describe a new method for measurement of CO
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 pCO 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 O.Sto 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 concentration dissociate
during CO inhalation, but CO concentration does not follow blood CO
concentration or 1/p02 as in the Warburg theory during intake or
elimination. Tissue CO concentration 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 averaged 0.36 ppm over 4 h. PAH 4-h
time avg reached 35 ng/m3.
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 vs 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, and cardiac output.
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 had
attenuated ACh and flow responses that were 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.
January 2010
B-2
-------
Reference
Purpose
Findings
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.
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. (1999,180426).
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 bilirubin
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 wk).
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 6 compartments including: alveolar, pulmonary
capillaries, arterial, venous, tissue capillary, and tissues (muscular and
nonmuscular). 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 (types 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 as 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%.
To determine if healthy Asians may be misdiagnosed as
Pesola et al (2006 1938551 having respiratory deficient due to comparison using The lung volume of Asian individuals is 10-15% lower than Caucasians.
1' ' ' Caucasian-derived prediction equation estimates of Thus a Chinese-derived prediction for DLCO should be used.
DLCO.
January 2010
B-3
-------
Reference
Purpose
Findings
Optimal blood mixing (when venous and arterial blood COHb% are
To determine the error in total Hb mass measurements equivalent) was determined to be after 6 min. A small volume of
using the optimized CO-rebreathing method due to loss administered CO leaves the vascular space (0.32% per min). A 2.3%
of CO to Mb increase in total Hb mass would be found if CO diffusion was not
included.
Prommer and Schmidt (2007,
1804211
Proudman et al. (2007,1867051
To review the signs of pulmonary arterial hypertension,
including a drop in DLCO in patients with systemic
sclerosis.
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. Ohstructjon
Patients with AA had lower FEVi 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 Wfeaver 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
Wfeaver 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 nonascitic patients and
both were higher than healthy controls. HO activity was higher in
cirrhotic patients than 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
postexposure 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.
January 2010
B-4
-------
Reference
Purpose
Findings
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.
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 and 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 nonsmoking 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
To evaluate the association of dose-dense
chemotherapy in breast cancer patients with pulmonary
dysfunction.
Patients receiving dose-dense chemotherapy for breast cancer had a
significant reduction in DLCO.
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 nonsurvivors.
January 2010
B-5
-------
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. PNAS, 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
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
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
degradation in the rat brain. Free Radic Biol Med, 37: 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).
January 2010 B-6
-------
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, 21: 1077-1091. 190946
Hampson NB; Weaver LK (2007). Carbon monoxide poisoning: a new incidence for an old disease. Undersea Hyperb Med,
34: 163-168. 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 Soc Mech Sci & Eng, 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
January 2010 B-7
-------
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 (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 LM; Dixon J; Pierce RJ; 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. J Clin Gastroenterol, 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: 189-197. 195129
Widdop B (2002). Analysis of carbon monoxide. Ann Clin Biochem, 39: 378-391. 030493
Wu L; Wang R (2005). Carbon monoxide: endogenous production, physiological functions, and pharmacological
applications. Pharmacol Rev, 57: 585-630. 180411
January 2010 B-8
-------
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. Am J Respir Crit Care Med, 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
January 2010 B-9
-------
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 Averaging Time:
via ambulatory ECG (Holler system) 1 -h ma
Study Design: Panel Mean (SD) unit:
1 1 ppm
Statistical Analyses: Linear regression
(mixed effects) Range (Min, Max):
0.1,7.7
Age Groups Analyzed:
40-75 yr 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.
Sample Description:
83 patients from the National Taiwan
University Hospital
Author: Chuang (2008,
155731)
Period of Study: NR
Location: Boston, MA
Health Outcome: HRV (changes in ST-
segment)
Study Design: Panel
Statistical Analyses: Linear additive
models; Additive mixed logistic regression
models
Age Groups Analyzed: 43-75
Sample Description: 48 patients with
documented CAD who had undergone
percutaneous coronary intervention for acute
coronary syndrome (acute Ml or unstable
angina pectoris) or who had worsened CAD
Averaging Time: 12 h, 24 h
Mean (SD) unit: 12 h:
0.48ppm, 24 h:0.46ppm
Range (Min, Max): 12-h:
25th percentile- 0.35, 75th
percentile-0.62, Max-1.88;
24 h: 25th percentile-0.37,
75th percentile- 0.62, Max-
1.56
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined: NR
Estimated RR for ST-segment depression £0.1
mm (ppm): 12-h: 0.70 (0.58-0.84)
24 h: 0.84 (0.68-1.03)
Estimated ST-segment change, mm (ppm): 12-h
mean: 0.013 (0.003-0.024)
24 h mean: 0.007 (-0.004-0.019)
CO not significantly associated with ST-segment
depression.
Author: Dales et al.
(2004, 099036)
Health Outcome: Various measures of HRV
via Holler system
Period of Study: NR Study Design: Panel
Location: Statistical Analyses: Linear regression
Toronto, Canada. (mixed effects)
Age Groups Analyzed:
51-88yr(mean65yr)
Sample Description: 36 subjects with pre-
existing CAD
Averaging Time: 24 h
Mean (SD) unit:
2.40 ppm (95th percentile)
Personal monitoring
Range (Min, Max):0.4,16.5
Copollutant: correlation
PM2.5:r=0.17
Increment: NR
Regression co-efficient [Lower Cl, Upper Cl]
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)
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).
January 2010
C-1
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
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
Mean (SD) unit: 0.47 ppm % Change [Lower Cl, Upper Cl]
Range (Min, Max): 0.12,0.82 Lags examined: 24 h
Copollutant: NR No significant effect with CO (no results recorded)
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 h, 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.
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: Holguinetal.
(2003, 057326)
Period of Study:
February-April 2000
Location:
Mexico City, Mexico
Health Outcome: Various measures of HRV Averaging Time: 24 h
via ECG
Mean (SD) unit: 3.3 ppm
Study Design: Panel
Range (Mm, Max): 1.8,4.8
Statistical Analyses: GEE
Copollutant: NR
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.
Increment: 10 ppm
Regression Coefficients [Lower Cl, Upper Cl]
Lags examined: 0
LagO:
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)
January 2010
C-2
-------
Study
Author: Ibald-Mullietal.
(2004, 087415)
Period of Study:
1998-1999
Location: Helsinki,
Finland
Erfurt, Germany
Amsterdam, Netherlands
Author: Liao et al.
(2004, 056590)
Period of Study:
1996-1998
Location:
Forsyth County, NC;
Selected suburbs of
Minneapolis, MN;
Jackson, Ml
Design
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
Health Outcome: Heart rate & various rates
ofHRV.
Study Design: Cohort
Statistical Analyses: Linear regression
Age Groups Analyzed:
45-64 yr (mean 62 yr)
Sample Description:
6,784 study subjects from the atherosclerosis
risk in communities study
Concentrations
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
Frfnrt- 0 1 9 5
Cl IUI I. U. I , £.vj
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 = 0.22 n/cm3
ACP: r = 0.60 n/cm3
Erfurt
PM25:r= 0.77 ug/m3
N02:r= 0.86 ug/m3
S02:r = 0.68 mg/m3
UFP:r = 0.72 n/cm3
ACP: r = 0.78 n/cm3
Helsinki
PM25:r= 0.40 ug/m3
N02:r= 0.32 ug/m3
S02:r = 0.19mg/m3
UFP:r = 0.35 n/cm3
ACP: r = 0.51 n/cm3
Averaging Time: 24 h
Mean (SD) unit:
0.65 ppm (0.44)
Range (Min, Max): NR
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined: 0, 1,2, 3
Results presented graphically
Increment: 0.44 ppm
Regression coefficients Lags examined: 1
Lag1:
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
January 2010
C-3
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Min (2009,
199514)
Period of Study:
December 2003 -
January 2004
Location: Tae-in island
community in South
Korea
Health Outcome: HRV
Study Design: Panel
Statistical Analyses: Time-lag model
Age Groups Analyzed: 20-87
Sample Description: 986 subjects, 367 with
metabolic syndrome (MetS), 619 without
MetS
Averaging Time: 8 h
Mean (SD) unit: 0.454 ppm
(0.560)
Range (Min, Max): 0.100,
7.200 ppm
Copollutant: PM10
Increment: NR
Estimated % Increase in subjects with MetS
[Lower Cl, Upper Cl]
Lags examined: 0-1,1-2, 2-3, 3-4,4-5, 5-6
Single pollutant:
Lag 0-1:
Log(SDNN):-0.29 (-0.59,0.00), p < 0.1
Log(LF):-0.34 (-1.02,0.33)
Log(HF):-0.67 (-1.41,0.08), p < 0.1
Lag 1-2:
Log(SDNN):-0.45 (-0.81,-0.10), p < 0.05
LogJLF): -0.65 (-1.46,0.17)
Log(HF):-1.04 (-1.94,-0.14), p < 0.05
Lag 2-3:
Log(SDNN):-0.28 (-0.57,0.02), p < 0.1
Log(LF):-0.19 (-0.87,0.48)
Log(HF):-0.82 (-1.57,-0.07), p < 0.05
Lag 3-4:
Log(SDNN):-0.18 (-0.47, 0.10)
LogJLF): -0.14 (-0.80,0.51)
Log(HF):-0.46(-1.19,0.27)
Lag 4-5:
Log(SDNN):-0.20 (-0.49, 0.09)
LogJLF): -0.36 (-1.04,0.31)
Log(HF):-0.42(-1.17,0.33)
Lag 5-6:
Log(SDNN): 0.13 (-0.18,0.44)
Log(LF): 0.50 (-0.21,1.20)
Log(HF):-0.03 (-0.81,0.76)
Co-pollutant (with PM10):
Lag 0-1:
Log(SDNN):-0.25 (-0.56, 0.05)
LogJLF): -0.35 (-1.04,0.31)
Log(HF):-0.67 (-1.44,0.10), p<0.1
Lag 1-2:
Log(SDNN): -0.48 (-0.88, -0.09), p<0.05; Log(LF):
-0.72 (-1.63, 0.18); Log(HF): -1.09 (-2.09, -0.09),
p<0.05
Lag 2-3:
Log(SDNN): -0.35 (-0.67, -0.03), p < 0.05
Log(LF): -0.17 (-0.90, 0.56)
Log(HF):-0.78 (-1.59, 0.03), p < 0.1
Lag 3-4:
Log(SDNN):-0.22 (-0.55, 0.11)
LogJLF):-0.11 (-0.86,0.63)
Log(HF):-0.34(-1.17,0.49)
Lag 4-5:
Log(SDNN): -0.18 (-0.48, 0.12); Log(LF): -0.21
(-0.89, 0.48); Log(HF):-0.37 (-1.14,0.40)
Lag 5-6:
Log(SDNN): 0.17 (-0.14,0.49)
Log(LF): 0.54 (-0.18,1.25)
Log(HF): 0.00 (-0.80, 0.80)
No significant results for subjects without MetS.
January 2010
C-4
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Park et al.
(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
Range (Min, Max):
0.13,1.8
Statistical Analyses: Linear regression
Age Groups Analyzed:
21-81 yr
Sample Description:
497 men from the normative aging study in
Greater Boston area
Copollutant: NR
Increment: 0.24 ppm
% Change in HRV [Lower Cl, Upper Cl]
Lags examined: 4-h ma, 24-h ma, 48-h ma
Lag4-h ma:
SDNN(Log10):2.0(-2.9to7.3)
HF(Log10):8.8(-4.6to24.1)
LF(Log10):3.2(-7.0to14.6)
LF:HF(Log10):-5.1 (-13.5 to 4.1)
Lag 24-h ma:
SDNN(Log10):-2.2(-7.7to3.6)
HF(Log10):-13.2 (-25.4 to 1.0)
LF(Log10):-0.6(-11.9to12.1)
LF:HF(Log10): 14.5 (2.9-27.5)
Lag 48-h ma:
SDNN(Log10):-3.4(-10.2to3.9)
HF(Log10):-13.8 (-.28.9 to 4.4)
LF(Log10):-2.4(-16.2to13.6)
LF:HF(Log10): 13.2 (-1.1 to 29.6)
Author: Peters et al.
(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
Averaging Time: 24 h
Mean (SD) unit:
During air pollution episode:
4.54 mg/m
Outside air pollution episode:
A c-1 mn/m
f.vj \ Illy/Ill
Range (Min, Max):
Increment: 6.6 mg/m3
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)
Sample Description:
2681 men and women who participated in
the MONICA study
During air pollution episode:
2.39, 6.85
Outside air pollution episode:
0.91,11.51
Copollutant: NR
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)
Averaging Time: 24 h
Mean (SD) unit: 2.9 ppm
(personal monitor)
Range (Min, Max): 0.1, 18.0
Copollutant: NR
Increment: 1 ppm
Regression Coefficients [Lower Cl, Upper Cl]
Lags examined (per min): 5, 10
LagSmin:
HF: -0.006 (-0.023 to 0.010)
LF: -0.024 (-0.041 to -0.007)
VLF: -0.034 (-0.061 to -0.007)
Sample Description:
30 patients from the Outpatient Clinic of the
National Institute of Cardiology of Mexico
Notes: VLF = Very low frequency
Author: Schwartz et al.
(2005, 074317)
Period of Study: 1999
Location:
Boston, MA
Health Outcome: Measures of HRV via
Holler system
Study Design: Panel
Statistical Analyses: Linear regression
(hierarchical model)
Age Groups Analyzed:
61-89yr
Sample Description:
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 to10.5)
Lag 24 h:
SDNN:
.A 1 ^_n Rtr, _7 7V i-MQQn- _m 1 1.1 A'lr,
28 subjects living at or near an apartment
complex located on the same street as the
Harvard School of Public Health
-4.2 (-0.6 to -7.7); r
-17.4);
PNN50:
-14.8 (-3.0 to -25.2); LF:HF: 6.2 (-0.6 to 13.4)
January 2010
C-5
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
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:
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
Location:
3 Cities in Europe:
Amsterdam,
Netherlands ;Erfert,
Germany; Helsinki,
Finland
Health Outcome:
Stable CAD: Various measures of HRV via
ambulatory ECG (Holler system)
Study Design: Panel
Statistical Analyses: Linear regression
(mixed model)
Age Groups Analyzed: Mean age across 3
cities; 64-71 yr.
Sample Description:
131 subjects with stable CAD followed for 6
mo with biweekly clinical visits.
Averaging Time: 24 h
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfert: 0.4 mg/m
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
Increment: 1 mg/m3
Regression co-efficient [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2,3, 5-day avg
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 et al.
(2006, 088453)
Period of Study:
1999-2000
Location:
Atlanta, GA
Health Outcome: Various measures of HRV Averaging Time: 1 h
via Holler system
Mean (SD) unit:
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects models)
Age Groups Analyzed:
Mean 65 yr;IQR 55-73 yr
Sample Description:
18 subjects with COPD and 12 subjects with
recent Ml.
362.0 ppb
Range (Min, Max):
25th = 221.5; 75th = 398.1
Copollutant: correlation
PM25:r=0.43
Increment: NR
RR Estimate [Lower Cl, Upper Cl]; lag:
Lags examined (h ma): 1, 4,24
No CO results reported.
January 2010
C-6
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
ONSET OF CARDIAC ARRHYTHMIA
Author: Berger et al.
(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 76 yr)
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: Dockeryetal.
(2005, 078995)
Period of Study:
1995-2002
Location:
Boston, MA
Health Outcome:
Tachyarrhythmias:
Study Design: Panel
Statistical Analyses: Logistic regression
(GEE)
Age Groups Analyzed:
1 9-90 yr; mean 64 yr
Sample Description:
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
Lag2-dayma: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)
203 cardiac patients with ICDs within 40km
of air monitoring site at Harvard School of
Public Health, Boston
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
Copollutant: NR
Increment: 1 ppm
OR for Tachyarrhythmic event [Lower Cl, Upper
Cl]
Lags examined (days): 0
Results for all events
i *n n- n ooo in oin 1 m&\
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)
January 2010
C-7
-------
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 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 (8,201 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=0.68
03:r = -0.56
Averaging Time:
1 h and 24 h
Mean (SD) unit: NR
Range (percentiles):
1 h:
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 -2.1 2)
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 to 23-h ma:
0- to 23-h ma: 0.99 (0.80-1 .21)
January 2010
C-8
-------
Study
Author: Rich et al.
(2006, 088427)
Period of Study:
1995-1999
Location:
Boston, MA
Author: Sari et al.
(2008, 190315)
Period of Study: June
2007
Location:
Gaziantep, Turkey
Author: Sarnat et al.
(2006, 090489)
Period of Study:
24 wk during the
summer and fall of 2000
LQucHlun .
Steubenville, OH
Author: Vedal et al.
(2004, 055630)
Period of Study:
1997-2000
Location:
Vancouver, Canada
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
Health Outcome: P-wave dispersion
(predictors of atrial fibrillation, ventricular
arrhythmias and sudden death) via ECG
Study Design: Case control
Statistical Analyses: Pearson correlation
analysis
Age Groups Analyzed:
Barbecue workers mean age: 33.66 ± 9.43 yr
Control group mean age: 35.15 ± 6.78 yr
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
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 nonsmoking older adults
Health Outcome: Cardiac arrythmia via
patients with ICD
Study Design: Panel
Statistical Analyses:
Logistic regression (GEE)
Age Groups Analyzed: Range from
12-77 yr (mean age 53 yr)
Sample Description:
50 patients who experienced 1 or more
arrhythmia event during the 4yr
Concentrations
Averaging Time: 1 h
and 24 h
Mean (SD) unit: NR
Range (Min, Max):
1 h:
25th = 0.46; 75th = 1.04
24 h:
25th = 0.52; 75th = 1.03
Copollutant: NR
Averaging Time: NR
Mean (SD) unit: COHb%
Indoor barbecue workers:
6. 48% ±1.43
Control Group:
2.1 9% ±1.30
Range (Min, Max): NR
Copollutant: NR
Averaging Time: 24 h
Mean (SD) unit: 0.02 ppm
Range (Min, Max): -0.1, 1.5
Copollutant: correlation
PM2s' r = 0 45
S02:r = 0.62
N02:r=0.66
03:r = -0.37
Averaging Time: 24 h
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
CO Effect Estimates (95%
Increment:
Lag (hrs) 0:0.58 ppm
Lag (hrs) 0-23: 0.51 ppm
Cl)
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)
Increment: NR
Correlation coefficient for COHb [p-value]
Lags examined: NR
Pmin: -0.1 32 (0.245)
Pmax: 0.21 5 (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)
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:
Supraventricular ectopy
SVE: 0.99 (0.76-1 .29)
Ventricular ectopy
VE:1.05 (0.75-1.46)
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)
January 2010
C-9
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CARDIAC ARREST
Author: Levy et al.
(2001.017171)
Period of Study:
1988-1994
Location:
Seattle, WA
Author: Sullivan etal.
(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: 1.79 ppm
Range (Min, Max):
0.52, 5.92
Copollutant: correlation
PM10:r = 0.81
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)
Lag 1:0.97 (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
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
(Estimated 30-yr 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
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)
Nonfatal cases: 0.98 (0.82-1 .16)
Fatal cases: 1.22 (0.98-1 .52)
ln-hospitaldeath:1.16(0.89-1.51)
Out-of-hospital death: 1.36 (1.01 -1.84)
Increment: NR
OR Estimate [Lower Cl, Upper Cl]
5-yr avg exposure
All subjects (n = 301 ,273)
Sample Description: 43,275 Ml cases
during 1985-1996; 511,065 controls
Range (percentiles):
Cases: 5th = 7.3; 95th =267.4
Controls: 5th =6.1 ;95th=261.8
Copollutant: PM10, N02
All cases: 1.01 (0.97-1.05)
Nonfatal cases: 0.94 (0.89-1.00)
Fatal cases: 1.14 (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)
Nonfatal 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)
January 2010
C-10
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
CHANGES IN BLOOD PRESSURE
Author:
Ibalde-Mullietal. (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
Sample Description:
2,607 men and women 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 (-0.94to 2.31);
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. Health Outcome: BP
Period of Study:
1999-2001
Location:
Boston, MA
Study Design: Cohort/Panel
Statistical Analyses: Random effects
Age Groups Analyzed: 39-90 yr
Sample Description:
62 subjects with 631 total visits
Averaging Time:
1 hand 120 havg
Mean (SD) unit:
Same h:0.81 ppm
120-h avg: 0.66 ppm
Range (Min, Max):
Same h:
10th = 0.48; 90th = 1.22
120-h avg:
10th = 0.48; 90th = 0.86
Copollutant: NR
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
CO had no significant effect on BP
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(mean43yr)
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 = 4.31
Mar-May:
25th = 1.03; 75th = 2.14
Jun-Aug:
25th = 0.73; 75th = 1.58
Copollutant: NR
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 (-0.14 to
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 (-0.04to
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, or free protein S for all lag
periods.
January 2010
C-11
-------
Study
Author: Delfino etal.
(2008, 156390)
Period of Study:
2005-2006
Design
Health Outcome: Biomarkers of systemic
inflammation
Study Design: Panel
Statistical Analyses:
Concentrations
Averaging Time: 24 h
Mean (SD) unit:
0.78 ± 0.30 ppb
Range (Min, Max):
0.22,1.97
CO Effect Estimates (95% Cl)
Increment: NR
Estimated coefficient
Relationship to outdoor air pollutants:
CRP (ng/mL) : Lag 0 : 847 .52 ; 3-day avg : 728 .79 ; 9-
Location: Los Angeles,
CA
Age Groups Analyzed:
> 65 yr (mean 85.7 yr)
Sample Description: 29 nonsmoking
subjects with history of CAD living in
retirement communities
Copollutant (Outdoor):
EC: r= 0.84
OC:r=0.69
OCprimary:r = 0.73
N02:r=0.78
03:r = -0.35
PMo25:r = 0.84
PM2.5-io: r = 0.51
day avg: 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: 139.45; 9-day avg: 225.60
Relationship to indoor air pollutants:
CRP (ng/mL): Lag 0:695.39; 3-day avg: 527.37; 9-
day avg: 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: 107.95; 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
January 2010
C-12
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Delfino etal.
(2009, 200844)
Period of Study: Jul-
midOct and midOct-Feb
of 2005-2006 and
2006-2007
Location: Los Angeles,
CA
Health Outcome: Biomarkers of
inflammation
Study Design: Panel
Statistical Analyses: Linear mixed effects
models adjusted for confounders
Age Groups: 65+ (84.1 ± 5.60) yr
Sample Description: 60 subjects with
confirmed CAD history, nonsmoker,
unexposed to environmental tobacco smoke
Averaging Time: 24 h
Mean (SD) unit: 0.50 (0.25)
ppm
Range (min, max): 0.11,
1.30
Copollutant: N02, NOX, 03,
PM025, PM025-25, PM25-10, EC.
OC, BC, OCpri, SOC, PN/cm5
Increment: NR
Regression coefficients (95% Cl)
Subjects with positive responses:
Cu,Zn-SOD (U/g Hb): 1-day avg: 1441 (97, 2786),
3-day avg: 2634 (1416, 3854), 5-day avg: 4227
(2078, 6376), 7-day avg: 3474 (914,6034), 9-day
avg: 2954 (737,5172)
GPx-1 (U/g HB): 1-day avg: -0.97 (-4.45,2.50),
3-day avg: -2.21 (-6.48, 2.06), 5-day avg: 4.71
(-2.90,12.33), 7-day avg: 4.20 (-3.29,11.68), 9-day
avg: 4.76 (-1.58,11.10)
Subjects with negative responses:
Cu,Zn-SOD (U/g Hb): 1-day avg:-195 (-338, -52),
3-day avg: -242 (-399, -85), 5-day avg: -242 (-440,
-44), 7-day avg: -315 (-664,34), 9-day avg: -176
(-508,156)
GPx-1 (U/g HB): 1-day avg: -0.82 (-1.55, -0.08),
3-day avg: -0.85 (-1.66, -0.04), 5-day avg: -0.84
(-1.88, 0.21), 7-day avg: -1.04 (-2.85,0.78), 9-day
avg:-0.47 (-2.19,1.26)
All subjects:
IL-6 (pg/mL): 1-day avg.: 0.35 (0.17,0.54), 3-day
avg.: 0.40 (0.20,0.61), 5-day avg.: 0.54 (0.27,0.80),
7-day avg.: 0.34 (-0.06, 0.74), 9-day avg.: 0.31 (-
0.07, 0.70)
P-selectin (ng/mL): 1-day avg.: 3.33 (0.94,5.73),
3-day avg.: 3.65 (1.02, 6.29), 5-day avg.: 5.28 (1.86,
8.70), 7-day avg.: 11.2 (5.39,17.0), 9-day avg.: 10.4
(4.83,16.0)
TNF-RII (pg/mL): 1-day avg: 112 (13,211), 3-day
avg: 136 (29, 243), 5-day avg: 229 (88, 371), 7-day
avg: 132 (-86, 349), 9-day avg: 220 (19, 421)
TN F-a (pg/mL): 1 -day avg: 0.05 (-0.05, 0.16), 3-day
avg: 0.09 (-0.03, 0.20), 5-day avg: 0.14 (-0.01, 0.29),
7-day avg: 0.07 (-0.19, 0.33), 9-day avg: 0.14 (-0.11,
0.39)
CRP (ng/mL): 1-day avg: 780 (343,1217), 3-day
avg: 739 (255,1222), 5-day avg: 1117 (485,1749),
7-day avg: 126 (-800,1052), 9-day avg: 41 (-840,
923)
SOD (U/g Hb): 1-day avg: -62 (-231,108), 3-day
avg: -53 (-244,138), 5-day avg: -37 (-285, 211),
7-day avg: 98 (-314, 509), 9-day avg: 208 (-173,
590)
GPx-1 (U/g Hb): 1-day avg: -0.69 (-1.41, 0.03), 3-day
avg: -0.69 (-1.48,0.11), 5-day avg: -0.56 (-1.60,
0.48), 7-day avg: -0.56 (-2.34,1.21), 9-day avg: 0.05
(-1.63,1.72)
Effect modification by medication use:
TNF-RII (pg/mL): 1-day avg: All subjects: 125 (11,
239), Statins: 48 (-105, 201), No Statins: 199 (47,
352); 3-day avg: All subjects: 161 (39, 283), Statins:
1 (-170,171), No Statins: 306 (141, 472); 5-day avg:
All subjects: 257 (100, 413), Statins: 15 (-210, 240),
No Statins: 445 (240,649); 7-dayay avg: All subjects:
176 (-68, 419), Statins: 43 (-297, 382), No Statins:
283 (-23, 589); 9-dayay avg: 265 (41, 489), Statins:
160 (-158, 478), No Statins: 355 (65,646)
sP-selectin (ng/mL): 1-day avg: All subjects: 1.84
(-0.62, 4.30), Clopidogrel: 0.00 (-2.80,2.81), No
Clopidogrel: 1.72 (-0.42, 3.86); 3-day avg: All
subjects: 1.90 (-0.79, 4.60), Clopidogrel: -0.67 (-3.95,
2.60), No Clopidogrel: 1.60 (-0.76, 3.96); 5-day avg:
All subjects: 2.97 (-0.47, 6.41), Clopidogrel: -0.18
January 2010
C-13
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
(-4.38,4.01), No Clopidogrel: 3.04 (0.06, 6.01);
7-day avg: All subjects: 6.74 (0.75,12.73),
Clopidogrel: 2.24 (-4.22, 8.71), No Clopidogrel: 6.78
(1.60, 3.96); 9-day avg: All subjects: 6.96 (1.20,
12.72), Clopidogrel: 2.0 (-4.40, 8.48), No
Clopidogrel: 5.54 (0.46,10.6)
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,
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
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/m3
Increment: 0.6 ppm
Regression coefficients [SE]
Lags examined (days):1
Lag1:
Fibrinogen (mg/dL): -0.1 6 (0.67)
Factor VIII -C(%): 0.45 (0.42)
vWF%: -0.29 (0.50)
WBC(x103/mm3): 0.003 (0.01 7)
Albumin (g/dL): -0.018 (0.003)**
** p < 0.01
Increment: 0.34 mg/m3
Change of IL-6
% of overall mean per IQ range increase
Genotypes: 1 1,1 2,22
IL6 rs2069832
1 1 • o n in i T RV 1 o- n o LI j 1 i\-o o- o n
Germany; Barcelona,
Spain; Helsinki, Finland
Rome, Italy; Stockholm,
Sweden
(mean = 62.2 yr)
Sample Description: 955 subjects who had
experienced Ml between 4 mo and 6 yr
before start of the study
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
PM25:r=0.55
PNC: r = 0.67
(-4.7, 0.8); p-value: 0.03
IL6 rs2069840
11:2.0(0.3,3.8);12:0.4(-0.9,1.7);22:-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
11:1.0 (-0.7,2.7);! 2:0.7 (0.6, 2.0);2 2:0.4
(-1.9, 2.7); p-value: 0.64
FGBrs1800790
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
Author: Pekkanen et al. Health Outcome: Fibrinogen
(2000.013250) „..,„. „ u ,
Study Design: Cohort
Period of Study:
1991 -1993 Statistical Analyses:
Logistic regression
Location:
London, England
Age Groups Analyzed:
35-55 yr
Sample Description:
7,205 office workers
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
S02:r = 0.61
03:r = -0.45
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 3.19 g/l [p value]
Lags examined: 0,1,2,3
Lag 0:1.17 (0.05); Lag 1:1.09 (0.31);
Lag 2:1.14 (0.11); Lag 3:1.22 (<0.01)
January 2010
C-14
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Ruckerl et al.
(2006, 088754)
Period of Study:
2000-2001
Location:
Erfert, Germany
Health Outcome: Blood markers of
inflammation and coagulation
Study Design: Panel
Statistical Analyses:
Linear and logistic regression (fixed effects)
Age Groups Analyzed:
51-76yr(mean = 66yr)
Sample Description:
57 male patients with CHD
Averaging Time: 24 h
Mean (SD) unit:
0.52 mg/m
Range (Min, Max):
0.11,1.93
Copollutant correlation:
N02:r=0.82
Increment: 0.27 mg/m
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 VII)
0-23: -1.4 (-3.8 to 1.1); 24-47: -2.6 (-4.8 to 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/m3
Barcelona: 0.59 mg/m3
Helsinki: 0.31 mg/m
Rome: 1.40 mg/m3
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 (-6.84 to 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)
January 2010
C-15
-------
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 = 41 yr
Sample Description: 40 healthy individuals
Averaging Time: 24 h Increment: NR
Median (SD) unit: 333 ug/m3 Estimated Changes [Lower Cl, Upper Cl]
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
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); 172-96: -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); 124-48; -1.3 (-6.1,3.6);
148-72: -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);
0+10-96:0.1 (-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);
172-96:-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 = 46 yr
Sample Description: 3,659 subjects living
within 11 km of monitoring site
Averaging Time: 24 h
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
Increment: 0.3 ppm
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);
Lastwkavg:-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.
VARIOUS MEASURES OF CARDIOVASCULAR HEALTH
Author: Briet et al.
(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: 18-35yr
Sample Description: 40 healthy white male
nonsmokers
Averaging Time: 24 h
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
PM2.5, PM10, NO, N02, S02
Increment: NR
p-Coefficient [Lower Cl, Upper Cl]
Flow-mediated Brachial Artery Dilatation:
-0.68 (-1.22,-0.15)
Small Artery Reactive Hyperemia:
10.46(1.73,19.31)
January 2010
C-16
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
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:
PM15, PM10, NOX, SOX
Increment: NR
RR Estimate [Lower Cl, Upper Cl]
Lags examined: NR
No quantitative results presented
Author: Wellenius etal.
(2007, 092830)
Period of Study:
February 2002-March
2003
Location: Boston, MA
Health Outcome: Congestive heart failure
Study Design: Cohort (retrospective)
Statistical Analyses: Linear mixed models
Age Groups Analyzed: 33-88 yr.
Tai Chi Group mean age (n=14): 66 ± 13 yr.
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:
PM25: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
PM25:r = 0.44
PM10: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)
Author: Henrotin et al. (2007,
093270)
Period of Study: 1994-2004
Location:
Dijon, France
Health Outcome (ICD9 or ICD10):
Stroke (Ischemic & Hemorrhagic)
Study Design: Bidirectional case
crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: > 40 yr
Sample Description: NR
Averaging Time: 24 h
Mean (SD) unit: 683 ug/m3
Range (Min, Max): 0,4014
Copollutant: NR
Increment: 10 ug/m3
OR Estimate [Lower Cl, Upper Cl]
Lags (days) examined: 0,1,2, 3.
Ischemic:
Lag 0:0.999 (0.997-1.001)
Lag 1:0.998 (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)
Lag 1:1.001 (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.
January 2010
C-17
-------
Study
Author: Maheswaran et al.
(2005, 090769)
Period of Study: 1994-1 998
Location:
Sheffield, UK
Author: Tsai et al. (2003,
080133)
Period of Study:
1997-2000
Location:
Kaohsiung, Taiwan
Author: Villeneuve et al. (2006,
090191)
Period of Study: 1992-2002
Location:
Edmonton, Canada
Author: Wellenius et al. (2005,
088685)
Period of Study: NR
Location:
9 U.S. cities: Chicago, Detroit,
Pittsburgh, Cleveland,
Birmingham, New Haven,
Seattle, Minneapolis, Salt Lake
City
Design
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: 2 45 yr
Sample Description: 1 ,030 census
districts
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
ED Visits (within 5 hospitals)
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
ED Visits
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
Concentrations
Averaging Time: NR
Mean (SD) unit: Quintiles
Range (Min, Max): NR
Copollutant: NR
Averaging Time: 24 h
Mean (SD) unit: 0.79 ppm
Range (Min, Max): 0.24, 1.72
Copollutant: NR
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
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
CO Effect Estimates (95% Cl)
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 (1.1 7-1 .42)
5th: 1.37 (1.24-1 .52)
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 persisted when adjusting for
S02 and 03
IS results persisted when controlling for
PM10, S02 and 03
Increment: 0.5 ppm
OR Estimate [Lower Cl, Upper Cl]
Lags (days) examined: 0, 1 & 0-2
Ischemic (April-Sept)
Lag 0:1. 16 (1.00, 1.33)
Lag 1:1. 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.
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)
January 2010
C-18
-------
Study
Design
Concentrations
CO Effect Estimates (95%
Cl)
ISCHEMIC HEART DISEASE
Author: D'lppoliti et al. (2003,
074311)
Period of Study: 1995-1 997
Location:
Rome, Italy
Author: Hosseinpoor et al.
(2005, 087413)
Period of Study:
1996-2001
Location:
Tehran, Iran
Author: Lanki et al. (2006,
089788)
Period of Study: 1994-2000
Location:
5 European cities:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rnmo Itak/
r\U[Ilc, lldly
Stockholm, Sweden
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.
Health Outcome: Angina Pectoris
(ICD9:413;ICD10:I20)
Study Design: Time series
Statistical Analyses:
Poisson regression
Age Groups Analyzed: All
Sample Description: NR
Health Outcome: First AMI (ICD9:410;
ICD1 0:121, 122)
Study Design: Time series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed:
35+ yr
Sample Description:
26,854 Hospital Admissions
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
Averaging Time: 24 h
Mean (SD) unit: 10.8 mg/m3
Range (Min, Max): 1.6, 57.8
Copollutant: NR
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
95th - 0 ?• 75th - n 5
£UUI — "•'J, 'vjlll — U.vJ
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
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)
Lag 1:1. 020 (0.988-1 .054)
Lag 2: 1.033 (1.001 -1.066)
Lag 3: 1.01 0(0. 982-1 .040)
Lag 4: 1.025 (0.996-1 .055)
Lag 0-2: 1.044 (1.000- .089
Increment: 1 mg/m3
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0, 1 , 2, 3
Lag 1:1. 00957 (1.00600-1 .01 31 5)
Increment: 0.2 mg/m3
RR Estimate [Lower Cl, Upper Cl] ;
Lags examined: 0,1,2,3
All 5 cities:
Lag 0:1. 005 (1.000-1 .010)
Lag 1:1. 002 (0.996-1 .007)
Lag 2: 1.002 (0.997-1 .007)
Lag 3: 0.998 (0.992-1 .003)
3 cities with Hospital Discharge
Register(HDR):
Lag 0:1. 007 (1.001 -1.01 2)
Lag 1:1. 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:
Lag 0:1. 027 (1.006-1 .048)
Lag 1 : 1 .021 (1 .000-1 .042)
Lag 2: 1.018 (0.997-1 .039)
Lag 3: 1.015 (0.994-1 .037)
Non-Fatal:
Lag 0:1. 001 (0.995-1.008)
Lag 1:1. 000 (0.994-1 .007)
Lag 2: 1.004 (0.998-1 .011)
lag:
Lag 3:0.999 (0.992-1.006)
3 cities with HDR - > 75years
Fatal:
Lag 0:1.009 (0.992-1.006)
Lag 1:1.001 (0.985-1.018)
Lag 2:1.006 (0.990-1.023)
Lag 3:1.000 (0.983-1.017)
Non-Fatal:
Lag 0:1.015 (1.004-1.086)
Lag 1:1.006 (0.995-1.017)
Lag 2:0.995 (0.983-1.006)
Lag 3:0.998 (0.987-1.009)
January 2010
C-19
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Lee etal. (2003,
095552)
Period of Study: 1997-1999
Location:
Seoul, Korea
Study Design: Time-series
Health Outcome (ICD9 or ICD10):
Angina: ICD10:120
AMI: ICD10:121-123
Other Acute IHDs:ICD10:124
Statistical Analyses: Poisson
regression, GAM
Age Groups Analyzed: 64+ yr
Sample Description: 822 days
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
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: 1.19 (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)
Author: Maheswaran et al.
(2005, 090769)
Period of Study:
1994-1998
Location:
Sheffield, UK
Emergency Hospital Admission
Health Outcome (ICD9):
CH 0(410-414)
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)
Averaging Time: NR
Mean (SD) unit: Quintiles
Range (Min, Max): NR
Copollutant: NR
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: 1.15 (1.05-1.26)
4th: 1.19 (1.09-1.30)
5th: 1.20 (1.09-1.32)
Author: Mann etal. (2002,
036723)
Health Outcome (ICD9): IHD (IHD)
(410-414); Ml (410)
Period of Study: 1988-1995 Study Design: Time series
Location:
Southern California
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): 0.30,11.8
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)
Lag 1:1.51 (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)
Lag 1:3.34 (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)
Lag 1:1.45 (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 (-S.Oto.O)
Lag 1 not significant for all results
January 2010
C-20
-------
Study
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
Design
Health Outcome: Hospital cardiac (mi),
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
Concentrations
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
CO Effect Estimates (95% Cl)
Increment: 0.2 mg/m3 (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: 1.01 4 (1.001 -.026)
HEART FAILURE
Author: Lee etal. (2007,
090707)
Period of Study: 1996-2004
Location:
Kaohsiung City, Taiwan
Author: Symons et al. (2006,
091258)
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)
Hospital Admissions
Health Outcome: NR
Averaging Time: 24 h
Mean (SD) unit: 0.76 ppm
Range (Min, Max): 0.14, 1.72
Copollutant: NR
Averaging Time: 24 h
Mean (SD) unit: 0.4 ppm
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; 1.1 5 (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: 1.1 7 (1.07-1 .28)
<25°C: 1.36 (1.22-1 .51)
Increment: 0.2 ppm
OR Estimate [Lower Cl, Upper Cl]
Period of Study: 2002
(April-November)
Location:
Johns Hopkins Bayview
Medical Center, Baltimore, MD
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
398 Hospital Admissions for CHF
Range (Min, Max): 0.1,1.0
Copollutant: NR
Lags examined (days):
0,1,2,3, cum 1, cum 2, cum 3
Lag 0:0.86 (0.67-1.11)
Lag 1:0.90 (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)
Cum. Lag2:0.80 (0.54-1.17)
Cum. Lag3:0.27 (0.46-1.14)
Author: Wellenius et al. (2005,
087483)
Period of Study: 1987-1999
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)
January 2010
C-21
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Yang (2008,157160)
Period of Study: 1996-2004
Location: Taipei, Taiwan
Hospital Ad missions
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: 1.26ppm
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 (>20°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 (<20°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)
CARDIOVASCULAR DISEASES - NON-SPECIFIC
Author: Ballesteretal. (2001,
013257)
Period of Study: 1994-1996
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]; lag:
Lags examined (days): 0,1,2, 3,4, 5
All cardiovascular:
Lag 2:1.0077 (0.9912-1.0138)
Heart Disease:
Lag 1:1.0092 (0.9945-1.0242)
Cerebrovascular Disease:
Lag 1:0.9874 (0.9646-1.0107)
Author: Ballester et al. (2006,
088746)
Period of Study: 1995-1 999
Location: 14 Cities in Spain
Author: Barnett et al. (2006,
089770)
Period of Study: 1998-2001
Location:
Brisbane, Canberra,
Melbourne, Perth, Sydney
Australia
Health Outcome
(ICD9: All CVD (390-459) ;Heart
diseases (410-414, 427, 428)
Study Design: Time series
Statistical Analyses: GAM
Age Groups Analyzed: All
Sample Description: NR
Hospital Admissions with
CVDs
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
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
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
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)
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)
Zealand
Conditional logistic regression
Age Groups Analyzed:
15-64yr&>65yr
Sample Description: NR
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
Copollutant NR
: 1.8 (-0.7 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)
January 2010
C-22
-------
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
CVDs
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
Age Groups Analyzed: > 65 yr
Sample Description:
>9.3 million Medicare subjects
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
EC: r = 0.48
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
Health Outcome (ICD9): CVD Hospital
Admissions (410-429)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Averaging Time: 24 h
Mean(SD) unit: 1.37 ppm
Range (Min, Max): 0.37, 3.66
Copollutant: NR
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)
Age Groups Analyzed: All
Sample Description:
74,509 CVD hospital admissions
(47 Hospitals)
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 (1.194-1.272)
<20°C: 1.098 (1.034-1.165)
Adjusted for N02:
>20°C: 1.048 (1.003-1.095
<20°C: 0.983 (0.914-1.058)
Adjusted for 03:
>20°C: 1.196 (1.161-1.232)
<20°C: 1.092 (1.031-1.157)
Author: Filhol. (2008,190260)
Period of Study:
January 2001-July 2003
Location:
Sao Paulo, Brazil
ED Visits
Averaging Time: 8 h
Health Outcome (ICD10): Hypertension Mean (SD) unit: 2.7 ppm
and Cardiac Ischemic Disease (110-125)
v ' Range (Min, Max): 0.7,12.1
Study Design: Time series
Copollutant: correlation
Statistical Analyses: Linear Poisson PM10: r = 0.69
regression models
Age Groups Analyzed: >18yr
Sample Description: 45,000
Cardiovascular emergency room visits
from diabetic and non-diabetic patients
(tertiary referral teaching hospital)
N02:r = 0.58
S02:r=0.52
03:r = 0.07
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 (0.0418)
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 Ad missions of
CVDs
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)
January 2010
C-23
-------
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)
Lag 1:1.33 (0.47-2.20)
Lag 0-1:2.35 (1.39-3.32)
Cardiac Disease:
Lag 0:2.52 (1.50-3.54)
Lag 1:1.85 (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)
Lag 1:1.58 (0.01-3.19)
Lag 0-1:2.86 (1.07-4.68)
Stroke: No results were significant for
Stroke.
AIICVD:
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: 1993-1 997
Location:
Denver, CO
Author: Linn et al. (2000,
002839)
Period of Study: 1992-1 995
Location:
Los Angeles, CA
Hospital Admissions for CVD
Health Outcome (ICD9: M 1(410-
41 0.92) ; coronary atherosclerosis
(41 4-41 4.05) ; 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
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: 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
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: 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.
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 (0.01 4)
Spring: 0.1 07 (0.033)*
Summer: 0.030 (0.033)
Fall: 0.008 (0.012)
Ml
All: 0.040 (0.009)*
CHF
All: 0.025 (0.009)*
Cardiac Arrythmia
All: 0.023 (0.009)*
Stroke
All: 0.044 (0.009)*
Notes :*p< 0.05
January 2010
C-24
-------
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)
\Arithnntrur- 1 r\iA n 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 ma: 1.020 (1.01 0,1. 030)
Results for multi-pollutant models presented
graphically
Age Groups Analyzed: NR
Sample Description: 10,234,490 ED
Visits (238,360 CVD group)
January 2010
C-25
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Yang etal. (2004,
094376)
Health Outcome (ICD9:
CVDs (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
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Bell etal. (2007,
091059)
Period of Study: 1999-2002
Location:
Connecticut and
Massachusetts
Health Outcome:
Birth weight and LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
regression
Averaging Time: 24 h
Mean (SD) unit:
0.65 ppm (0.1 8)
Range (Min, Max): NR
Copollutant: NR
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)
Age Groups Analyzed: NA
Sample Description:
358,504 full-term live
singleton births (32-44 wk)
OR for LBW [Lower Cl, Upper Cl]
Entire pregnancy: 1.028 (0.983-1.074)
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
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)
Averaging Time: LUR 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
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
January 2010
C-26
-------
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 Syr 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
and 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 to 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 and valve, atrial septal,
pulmonary artery and valve, ventricular septal,
endocardia! cushion and 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
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):
Averaging Time: 24 h
Mean (SD) unit: NR
Range (Min, Max):
Percentiles:
25th: 0.99 ppm
75th: 1.41 ppm
Copollutant correlation:
TSP'r = 073
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
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)
January 2010
C-27
-------
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 postneonatal 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)
Birmingham: 0.64; Bristol: 1.01; Leeds:
0.73; Liverpool: 0.51; London: 0.77;
Manchester: 0.63; Middlesbrough: 0.37;
Newcastle: 0.67; Nottingham: 0.62;
Sheffield: 0.60
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
Increment: 1 mg/m
RR Estimate [Lower Cl, Upper Cl]
Lags examined (days): 0,1,2
All infant deaths: 1.02 (0.96,1.09)
Neonatal deaths: 0.99 (0.92,1.07)
Post-neonatal deaths: 1.09 (0.94,1.25)
City-specific results of all infant mortality displayed
graphically
Author: Huynh et al. (2006,
091240)
Period of Study: 1999-2000
Location:
California
Health Outcome:
PTB (24-36 wk gestation)
Study Design: Case-control
Statistical Analyses:
Conditional Logistic
regression
Age Groups Analyzed:
Cases = 24- to 36-wk
gestation; Controls = 39-to
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:
Perl 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:
Perl 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 Health Outcome: Oral clefts Averaging Time: 8 h
(2008,193794) (with or without palate)
v v H ' Mean (SD) unit: 0.69 (0.4)
Study Design: Case control
Period of Study: 2001-2003
Location: Taiwan
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: 100 ppb
RR for oral cleft [Lower Cl, Upper Cl]
Month 1:1.00 (0.96-1.04)
Month 2:1.00 (0.96-1.03)
Month 3:1.00 (0.96-1.03)
Author: Jalaludin et al. Health Outcome: PTB
(2007, 156601)
Study Design:
Period of Study: 1998-2000 Retrospective cohort
Location: Statistical Analyses:
Sydney, Australia Logistic regression
Averaging Time: 8 h
Mean (SD) unit: 0.9 ppm (0.68)
Range (Min, Max): NR
Copollutant correlation: PM10: r = 0.28
NO.,- r- nfin
Increment: 1 ppm
RR for PTB [Lower Cl, Upper Cl]
First mo:
All of Sydney: 0.89 (0.84-0.95)
Within 5km of site: 1 .03 (0.68-1 .54)
Age Groups Analyzed: NA
Sample Description:
123,840 full term live
singleton births (<42 wk)
S02: r = 0.24
03:r = -0.21
First trimester:
All of Sydney: 0.77 (0.71-0.83)
Within 5km of site: 1.24 (0.81-1.91)
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)
January 2010
C-28
-------
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
Design
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
388,1 05 full-term live
singleton births (37-44 wk)
Health Outcome:
PTB
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
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:
DM..- r = n 07
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
Age Groups Analyzed: NA
Sample Description:
52,113 live singleton births
N02:r=0.63
S02:r = 0.31
OR for PTB [Lower Cl, Upper Cl]
First Trimester:
Second quartile: 0.92 (0.81-1.05)
Third quartile: 1.14 (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.
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
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 3 km of monitoring
site.
Averaging Time: 24 h
Mean (SD) unit: 2.83 ppm
Range (Min, Max): 0.54, 10.25
Copollutant correlation:
PMi0: r = 0.71
N02:r=0.67
S02: r = 0.55
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
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.
January 2010
C-29
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Liu et al. (2003,
089548)
Health Outcome:
PTB, IUGR, LBW
Period of Study: 1985-1998 Study Design:
Retrospective cohort
Location:
Vancouver, BC, Canada Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
229,085 live singleton births
Averaging Time: 24 h
Mean (SD) unit: 1.0 ppm
Range (Min, Max): 25th: 0.7; 75th: 1.2
Copollutant: NR
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)
OR for PTB [Lower Cl, Upper Cl]
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 1:1.05 (1.00-1.10)
Trimester 2:0.97 (0.92-1.01)
Trimester 3:0.97 (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) (at least 20-wk
Averaging Time: 24 h
Mean (SD) unit: 1.1 ppm
Range (Min, Max): 25th: 0.6; 75th: 1 .3
Copollutant correlation:
PM25: 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 1 6% 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: 1.14 (0.83-1.58);
Third: 1.31 (1.06-1 .62)
Stratified results among African-Americans:
First: 1.43 (1.1 8-1 .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 trimeste: -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)
January 2010
C-30
-------
Study
Author: Medeiros et al.
(2005, 089824)
Period of Study: 1998-2000
Location:
Sao Paulo, Brazil
Design
Health Outcome: Birth
weight and LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic
Concentrations
Averaging Time: 24 h
Mean (SD) unit:
Daily mean shown in Figure (see paper)
Range (Min, Max): NR
Copollutant: NR
CO Effect Estimates (95% Cl)
Increment: 1 ppm
Regression coefficient for birth weight (g)
[Lower Cl, Upper Cl]
Trimesters:
First: -11 .9 (-15.5 to -8.2); Second: 4.9 (0.5-9.3);
Third: 12.1 (7.6-16.6)
regression
Age Groups Analyzed: NA
Sample Description:
311,735 full-term live
singleton births (37-41 wk)
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
Age Groups Analyzed: NA
Sample Description:
18,247 full-term live
singleton births (40 wk)
within 5 mi of a monitor
Increment: NR
Trimester specific results presented graphically
Single-pollutant Model for "sensitized to at least
one outdoor allergen"
OR adjusted for yr of birth and sex [Lower Cl,
Upper Cl]
Entire Pregnancy
24-havg: 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-havg: 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 foryr of birth and sex [SE]
Entire Pregnancy
24-havg: 1.33 (0.68)
Daily max:0.54 (0.27)
8-h max: 0.84 (0.42)
2nd Trimester
24-havg: 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
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)
January 2010
C-31
-------
Study
Author: Ritzetal. (2000,
012068)
Period of Study: 1989-1 993
Location:
Southern California
Design
Health Outcome: PTB
Study Design:
Retrospective Cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
Eligible study subjects were
singletons born at 26- to
44-wk gestation
Concentrations
Averaging Time: 6-9 a.m.
Mean (SD) unit: 2.70 ppm
Range (Min, Max):
0.36,9.12
Copollutant correlation:
PM10:r = 0.37
N02:r=0.60
03: r = -0.44
CO Effect Estimates (95% Cl)
Increment: 3 ppm
RR for PTB [Lower Cl, Upper Cl]
Adjusted for various risk factors and season of birth
and conception
6wk prior to birth: 1.04 (0.99-1.10)
1 st mo of pregnancy: 1 .04 (0.99-1 .09)
Adjusted for various risk factors
6 wk prior to birth: 1.06 (1.02-1 .10)
1 st mo of pregnancy: 1 .01 (0.97-1 .04)
Sample Description:
97,518 neonates born in
Southern California
Author: Ritzetal. (2002,
023227)
Period of Study: 1987-1993
Location:
Southern California
Health Outcome:
Birth defects (heart defects
and 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 and 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 and 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:
Postneonatal mortality
(28 days to 1 yr); all causes;
SIDS
Study Design: Case control
Statistical Analyses:
Conditional Logistic
regression
Sample Description:
Mothers residing within
16 km of monitoring site
Averaging Time: 24 h
Mean (SD) unit: 1.63 ppm
Range (Min, Max):
0.38, 3.44
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)
January 2010
C-32
-------
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
Author: Son etal. (2008,
190323)
Period of Study: NR
Location: Seoul, Korea
Author: Strickland et al.
(2009, 190324)
Period of Study: NR
Location : Atlanta, GA
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 58,31 6 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
Sample Description:
3,901 infants from the
California Children's Health
Study
Health Outcome:
Postneonatal mortality from
all causes
Study Design: Case
crossover and time series
Statistical Analyses:
Conditional logistic
regression
Age Groups Analyzed: NA
Sample Description: 1 ,286
first-born birth and infant
death records from
1999-2003 (only
postneonatal deaths)
Health Outcome:
Cardiovascular
malformations
Study Design:
Retrospective cohort
Statistical Analyses:
Poisson GLM
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
Averaging Time: 8 h
Mean (SD) unit: 1.01 ppm
Range (Min, Max): 0.29, 3.54
Copollutant:
PM10, N02, 03, S02
Averaging Time: 24-h
Mean (SD) unit:
By season of conception:
March-May: 0.9 ppm
June-August: 0.8 ppm
Sept.-Nov.:0.9ppm
Dec.-Feb.:0.7ppm
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 .1 7 (1 .08-1 .26) ; 1 .1 5 (1 .05-1 .26) ;
1.25(1.12-1.38)
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);
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)
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)
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)
Age Groups Analyzed: NA
Sample Description:
Pregnancies reaching at
least 20-wk gestation that
were conceived during
January 1,1986-March 12,
2003
By yr of conception:
1986-1991:0.7 ppm
1992-1997:0.8 ppm
1998-2003:0.7 ppm
Range (IQR): 0.3
Copollutant:
PM10(24h):r = 0.32
N02(24h):r = 0.41
03(8h):r = 0.07
S02(24h):r=0.23
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)
January 2010
C-33
-------
Study
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:
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
Concentrations
Averaging Time: 24 h
Mean (SD) unit: 8.27 ppmxIO
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
CO Effect Estimates
(95% Cl)
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:
<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:
Sample Description:
518,254 births within 4 mi of
a monitoring station. Varied
according to analyses.
Copollutant correlation:
First Trimester:
PM10:r = 0.12
PM25:r = 0.57
N02:r=0.81
S02:r = -0.31
<: 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, PM10)
OR for term LBW [Lower Cl, Upper Cl]
Third trimester:
<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.
Author: Woodruff etal.
(2008, 098386)
Period of Study: 1999-2002
Location:
U.S. counties with >250,000
residents
Health Outcome:
Postneonatal deaths
all causes; respiratory;
SIDS; ill-defined+ SIDS;
other causes.
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression (GEE)
Age Groups Analyzed: NA
Sample Description: NR
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: 1.14 (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)
Author: Yang etal. (2004,
094376)
Period of Study: 1994-2000
Location:
Taipei, Taiwan
Health Outcome:
Postneonatal mortality
(27 days-1 yr old)
Study Design: Case
crossover
Statistical Analyses:
Poisson regression
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 15.8 ppm x10
Range (Min, Max): 3.20, 48.4
Copollutant: NR
Increment: Interquartile range: 0.56 ppm
OR for Post-neonatal mortality [Lower Cl, Upper Cl]
Lag examined: 0-2
Lag 0-2: 1.038 (0.663-1 .624)
January 2010
C-34
-------
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
PM25: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
source)
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 mo
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
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
Health Outcome: Lung function (FVC,
FEV^FEWFVC.FEFs-^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
Age Groups Analyzed: 1st to 6th
grade children
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:
k\f\ • r — n QR n OQ
NU2. r - U.ob — U.yo
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)
Copollutant correlation:
PM10:r= 0.721
03:r= -0.204
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
FFV (m\ \
rcvi ^IIILJ
njl I, _.._
<;4-n 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
January 2010
C-35
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
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:
Nonsmoking 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/m
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;
N02
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
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
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
mivpH mnHpl
1 1 IIACU 1 1 IUUCI
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
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 J2.0) 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
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)
January 2010
C-36
-------
Study
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
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, Incheon, Busan, &
Jeju, Korea
Design
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: 10-1 7 yr
Sample Description:
A stratified cluster random sample of
students (n=69,367) from 1 ,139,452
students sampled nationwide
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
IHDpanel:7
Age Groups Analyzed:
COPD panel: 50-80 yr
Asthma panel :1 8-64 yr
IHD panel: 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
Concentrations
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 921 ug/m3
Range (Min, Max): (31 9, 1540)
Copollutant:
PM10;BS;N02;NO
Averaging Time: 8 h
Mean (SD) unit: NR
Range (min, max): NR
Copollutant:
NO, N02, NOX, 03, S02, PM10, PSI
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit:
Overall: 7.4 (6.2) mg/m3 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
PM10.25:r = -0.09
PM • r - 0 tt
r IVI10- 1 — 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
CO Effect Estimates (95% Cl)
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
concentrations
Estimate (p-value): 0.0750 (0.3336)
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), lag 0
Upper resp. symptoms:
1.006(1.003, 1.008), lag 0-2
Irritation symptoms:
1.004(1.001,1.006), lag 1-3
January 2010
C-37
-------
Study
Design
Concentrations
CO Effect Estimates (95% Cl)
Author: Mortimer et al.
(2008,187280)
Period of Study:
Nov 2000-Apr 2005
Location:
Fresno, California
Health Outcome: Allergic sensitization Averaging Time: 24-h avg, 24-h
~ . ~ • r, ,
Study Design: Panel
Statistical Analyses: Multistep
modeling
Age Groups Analyzed: 6-11 yr
Sample Description: 170 children
with physician diagnosed asthma
Mean (SD) unit: NR
IQ Range (24-h avg, 24-h max, 8-h
max): 0.28, 0.79,0.52
Copollutant: entire prenatal
correlation
N02:r = 0.74
03:r=-0.40
PM10:r=0.32
Increment: IQR
OR Estimate [Lower Cl, Upper Cl]; lag:
Lags examined: NR
Entire Pregnancy:
CO 24-h avg: 1.45 (1.02,2.07)
CO 24-h max: 1.53 (1.01, 2.33)
CO 24-h avg: 1.55 (1.01,2.37)
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
Acie Grouos Analvzed: 0-5 vr
Averaging Time: 8 h
Mean (SD) unit: NR
Range (min, max):
1 .3 |jg/m , 1 .83 ug/m
Copollutant: N02, S02, PM10
Increment: NR
Lags examined: NR
R Estimate:
Dry season: 0.13
Wet season: 0.25
Sample Description: 250 children
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
Health Outcome: respiratory
symptoms
Study Design: panel
Statistical Analyses: Mixed Effects
Models
Age Groups Analyzed: 5-12 yr
Sample Description:
861 children with persistent asthma
and atopy living in low-income census
tracts
Averaging Time: 8 h
Mean (SD) unit: NR
Range (10th-90th): 872.1 ppb
Copollutant:
PM10, S02, N02, 03
Increment: 872.1 ppb
Lags examined: NR
Change Estimate [Lower Cl, Upper Cl]:
FEV,:-0.56 (-1.31, 0.20)
PEFR:-0.49 (-1.24, 0.27)
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)
January 2010
C-38
-------
Study
Author: Penttinen et al.
(2001.030335)
Period of Study:
11/1996-4/1997
Location:
Helsinki, Finland
Author: Rabinovitch et al.
(2004, 096753)
Period of Study:
11/1999-3/2000;
11/2000-3/2001;
11/2001-3/2002
Location:
Denver, CO
Author: Ranzi et al. (2004,
089500)
Period of Study:
2/1999-5/1999
Location:
Emilia-Romagna, Italy
Author: Rodriguez et al.
(2007, 092842)
Period of Study:
1996-2003
Location:
Perth, Australia
Design
Health Outcome:
Lung function (PEF)
Study Design: Panel study
Statistical Analyses:
First order autoregressive linear model
Population:
57 nonsmoking adult asthmatics
Age Groups Analyzed: NR
Health Outcome:
Lung function (FEVi); asthma
exacerbation; bronchodilator use
Study Design: Panel study
Statistical Analyses:
Pulmonary function: Mixed effects
model; Asthma exacerbation and
medication use:GLM
Population:
Urban poor asthmatic children:
1999-2000:41
2000-2001 : 63
2001-2002:43
Age Groups Analyzed: 6-12 yr
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
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
Concentrations
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
PMio-2s' r = -0 30
PM25:r=0.32
piui • r- n IP
r ivii . i u.oc?
PNC: r = 0.44
NC0.01-0.1:r = 0.43
NC0.1-1:r = 0.47
NO' r = 0 60
MO • r - n A.A.
\\\J2- ' v ,*ri
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
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
Pollutant: CO
Averaging Time: 8-h avg
Mean (SD) unit: 1 .408 ppm
Range (Min, Max): (0.012, 8.031)
Copollutant: NR
CO Effect Estimates (95%
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 (1.1 9); 5-day avg
Afternoon
-0.4(0.43);0
-0.13(0.41);!
n 71 in 4iv o
U. / I ^U.t I / , £.
-3.03 (1.06); 5-day avg
Evsnino
-0.7 (0.45); 0;
-0.31 (0.44); 1
0.3 (0.44); 2
-3.62 (1.1 9); 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
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
Bronchodilator use:
1.065 (1.001 -1.1 33); 3-day ma
Cl)
The study did not present quantitative results
for CO.
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
January 2010
C-39
-------
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
January 2010
C-40
-------
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 yr
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
PNCO. 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
AFEV, (%) FEV, (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
January 2010
C-41
-------
Study
Author: von Klotetal.
(2002, 034706)
Period of Study:
9/1996-3/1997
Location:
Erfurt, Germany
Author: Yuetal. (2000,
013254)
Period of Study:
11/1993-8/1995
Location:
Seattle, Washington
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
Health Outcome:
Asthma symptoms (Wheezing,
coughing, chest tightness, shortness of
breath)
Study Design: Panel study
Statistical Analyses:
models (GEE)
Population: 133 mild-to-moderate
asthmatics
Age Groups Analyzed: 5-1 3 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' ~ '
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .6 ppm
Range (Min, Max): (0.65, 4.18)
Copollutant correlation:
PMi.0:r = 0.82
PM«: r = 0.86
S02:r=0.31
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
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);!
1.1 8 (1.04-1 .33); 2
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
Age Groups Analyzed:
Children : <1 4 yr, Adults : < 1 5 yr
Sample Description: Data
from daily number of
ambulance transports to ED for
asthma
Concentrations
Averaging Time: NR
Mean (SD) unit: 11.5ppm
Range (Min, Max): 3-44ppm
Copollutant: NR
Effect Estimates (95% Cl)
Increment: 0.1 ppm
ARIMA model for ambulance transports to ED for asthma
exacerbation among adults: p coefficient: 0.151, SE:
0.098, t statistic: 1 .537, P value: .125
ARIMA model for ambulance transports to ED for asthma
exacerbation among children: p coefficient: 0.01 9, SE:
0.034, t statistic: 0.549, P value: 0.583
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.
January 2010
C-42
-------
Study
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
Design
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)
Concentrations
Pollutant: CO
Averaging Time: Max 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;
PM2.5-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
Effect Estimates (95% Cl)
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: 2 65
CO: 1.024 (0.997-1 .053); 0-4
CO, PM10: 1.001 (0.961 -1.042); 0-4
flcthmg
Ane firniirv R-18
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: 5-18
yr;>65yr
CO: 1.104 (1.018-1.198); 0-5
CO, PM10:1.023 (0.911-1.149); 0-5
Author: Atkinson etal.
(1999.007882)
Period of Study:
1/1992-12/1994
Location:
London,
U.K.
ED Visits
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
0-1 4 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
0-1 4: 2. 92% (0.60,5.30);!
15-64: 2.15% (-0.27,4.63);!
> 65: 4.29% (1.15, 7.54); 0
Asthma vis its:
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
January 2010
C-43
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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., upper
and lower respiratory illness
including sinusitis, bronchitis,
and pneumonia)
Study Design: Time series
Statistical Analyses: Poisson
GAM, penalized splines
Age Groups Analyzed: <5 yr
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.
Author: Bell etal.
(2008, 091268)
Period of Study:
1/1995-12/2002
Location:
Taipei, Taiwan
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (486); asthma (493)
Study Design: Time series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SE) unit: 0.9 ppm
Range (Min, Max): (0.3, 3.6)
Copollutant: NR
Increment: 0.5 ppm
% Increase (Lower Cl, Upper Cl); lag
Asthma (avg correlation between monitor pairs = 0.75 (13
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
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
January 2010
C-44
-------
Study
Author: Bellini et al.
(2007, 097787)
Period of Study:
1996-2002
Location:
15 Italian cities
Design
Hospital Admissions
Health Outcome:
Respiratory conditions
Study Design:
Time-series; Meta-analysis
Statistical Analyses:
1 . GLM for city-specific
Concentrations
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: correlation NR
Effect Estimates (95% Cl)
Increment: 1 mg/m3
% 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
2. Bayesian random-effects for
meta analysis
Age Groups Analyzed:
All ages
Note: Estimates from Biggeri et al. (2004)
Author: Braga et al.
(2001.016275)
Period of Study:
1/1993-11/1997
Location:
Sao Paulo, Brazil
Hospital Admissions
Health Outcome (ICD9):
Respiratory (460-51 9)
Study Design: Time series
Statistical Analyses:
Poisson GAM, LOESS
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 4.8 (2.3) ppm
Range (Min, Max): (0.6, 19.1)
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
Age Groups Analyzed:
<2yr
3-5 yr
6-1 Syr
14-19 yr
0-19 yr
Copollutant: correlation
PM10:r = 0.60
03: r = -0.07
S02: r = 0.47
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
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
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.18 ppm
IQR (25th, 75th): (0.9, 1.4)
Copollutant: correlation
PM2.5:r = 0.49
PM10-25:r= 0.20
PM • r - n 4?
riving, i u.to
N02:r=0.55
S02: r = 0.37
03: r = -0.23
Pollutant: CO
Averaging Time: 1-havg
Mean (SD) unit: 1.9 ppm
IQR (25th, 75th): (1.3, 2.3)
Copollutant: correlation
03:r = 0.24
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:
Multipollutant model
CO S02, 03: 5.15%
CO,PM25,S02,03: 4.63%
CO, PM10.25, S02,03:5.25%
CO, PM10, S02, 03: 4.80%
CO,PM10.25,03: 4.00%
COPD:
Multipollutant model
CO, S02, 03: 3.02%
CO, PM25,S02, 03: 2.46%
CO, PM10.25, 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
January 2010
C-45
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Cakmak et al.
(2006, 093272)
Period of Study:
4/1993-3/2000
Location:
10 Canadian cities
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
Statistical Analyses:
1. Poisson
2. Restricted Maximum
Likelihood Method
Age Groups Analyzed:
All ages
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
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
Author: Cheng etal.
(2007, 093034)
Period of Study:
1996-2004
Location:
Kaohsiung, Taiwan
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-486)
Study Design: Bidirectional
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.76 ppm
Range (Min, Max): (0.1 4, 1.72)
Copollutant: correlation
nikfl
PM10
S02
N02
03
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
CO, 03, > 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
Health Outcome: pneumonia
HA
Study Design: case-crossover
Statistical Analyses:
Conditional Logistic regression
Age Groups Analyzed:
All ages
Sample Description:
152,594 HA for 47 hospitals in
Taipei city
Averaging Time: 24h
Mean (SD) unit: 1.26 ppm
Range (min, max): 0.12, 3.66
Copollutant: correlation
PM10:r = 0.34
S02: r = 0.57
N02:r=0.69
03:r = -0.31
Increment: 0.57 ppm (IQR)
OR Estimate [Lower Cl, Upper Cl]
Hag:
Lags examined: one wk before to one wk after
CO:
>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)
January 2010
C-46
-------
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: 0.01 3-0.482
Copollutant: N0;:;
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: Max 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, S02,03:5.10% (-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
January 2010
C-47
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Fung et al.
(2006, 089789)
Period of Study:
6/1995-3/1999
Location:
Vancouver,
Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory Illness
Study Design:
1. Dewanji and Moolgavkar
2. Time-series
3. Bidirectional case-crossover
Statistical Analyses:
1. Dewanji and Moolgavkar
2. Poisson
3. Conditional logistic
regression
Age Groups Analyzed: 2 65 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
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
Time-series
1.012 (1.000-1.023) ;0
1.017 (1.003-1.032); 0-2
1.017 (1.001-1.035); 0-4
1.016 (0.996-1.036); 0-6
Bidirectional 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
1.010J0.991 1.031; 0-6
Author: Fusco etal.
(2001.020631)
Period of Study:
1/199510/1997
Location:
Rome,
Italy
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
0-14 yr
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
S02: r = 0.56
03:2r = -0.57
Cold Season
S02'r= 0 37
N02:r = 0.41
03: r = -0.44
Warm Season
S02: r = 0.44
N02:r=0.59
03: r = -0.38
Increment: 1.5 mg/m
% 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% (0.90-10.40);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% (1.60-7.10);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% (2.90-14.60);0
COPD:
13.90% (6.80-21.50); 0
Age Group: 0-14
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
January 2010
C-48
-------
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
Ana f^rnimc Antilw^aH1
Pollutant: CO
Averaging Time: Max 8-h avg
Mean (SD) unit: 5.8 (2.4) ppm
Range (Min, Max): (1.3, 22.8)
Copollutant: correlation
PM • r- n RT,
rlvlio- 1 u.uo
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
January 2010
C-49
-------
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
(lMD) 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: Max 8-h avg
COPD (490.00-496.99
excluding asthma); pneumonia/ Mean (SD) umt:
influenza (480 00-489 99)' Al1 Year: 2-3 (1 -3) PPm;
Asthma (493) November-April: 2.2 (1.3) ppm;
v ' May-October: 2.4 (1.2) ppm
Study Design: Case crossover
Range (10th, 90th):
Statistical Analyses: All Year: (0.9, 4.2)
Conditional logistic regression November-April: (0.8, 4.2)
May-October: (1.1, 4.2)
Age Groups Analyzed:
All 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: Max 8-h avg
Lower respiratory tract
infections (466, 480-486) Mean (SD) unit: 1 .00 (0.30) ppm
Study Design: Time series Range (Min, Max): (0.51 , 1 .71)
Statistical Analyses: Copollutant: NR
1. General linear regression
2. Bayesian hierarchical
modeling
Age Groups Analyzed:
All Anac
MM rtyco
014\/r
- 1 f yl
1 *i fi4 \/r
i \j ut yi
> fi^ \/r
z. 00 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);!
0.40% (0.1 0-0.80); 2
January 2010
C-50
-------
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
Concentrations
Pollutant: CO
Averaging Time: Max 8-h avg
Mean (SD) unit:
All Season: 1.31 (0.43) ppm
Warm Months (April-September):
1.22 (0.32) ppm
Effect Estimates (95% Cl)
Increment: 1.3 ppm
Relative Risk (Lower Cl, Upper Cl); Lag
Warm months: 1 .15 (1 .07-1 .25); 0-1
Age Groups Analyzed:
All ages
Cold Months (October-March):
1.41 (0.5) ppm
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
Author: Karretal.
(2006, 088751)
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 ages
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 to 1 yr
Hospital Admissions
Health Outcome (ICD9):
Acute bronchiolitis (466.1)
Study Design: Case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Infants: 3 wkto! 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
0- r - n nfi?
3. 1 ~ U.UDO
*p < 0.05
Pollutant:CO
Averaging Time: 24-h avg
Mean (SD) unit:
Chronic:!, 770 ppb
Subchronic: 1,720 ppb
Range (Min, Max):
Chronic: (120, 8300)
Subchronic: (130, 5070)
Copollutant: NR
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
Range (Min, Max):
Lag 1:
Index*: (4, 9600) Referent*: (4, 9600)
Lag 4:
Index* (4, 8710) Referent* (4, 9600)
CoDollutant: NR
Increment: 10 ug/m3
RR Estimate [Lower Cl, Upper Cl] ; lag:
Lags examined: lag days 0-3
Single Pollutant: 0.9989 (0.985, 2.715), 2
Multi-pollutant: 0.998 (0.993, 1.004), 2
Winter, all ages: 1.027 (1.004, 1.051), 2
Winter, males 50-69:2.625 (1.048, 1.158)
Increment: 910 ppb, 960 ppb
Odds Ratio (Lower Cl, Upper Cl); lag:
Increment: 910 ppb
Subchronic broncholitis: 1 (0.97-1.03)
Increment: 960 ppb
Chronic broncholitis: 1 (0.97-1 .03)
Increment: 1361, 1400 ppb
Odds Ratio (Lower Cl, Upper Cl); Lag
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
Age Group:
Overall' 097 (094-1)' 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
* Index days: days lagged in
reference to date of hospitalization of
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.
January 2010
C-51
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Kimetal.
(2007, 092837)
Period of Study:
2002
Location:
Seoul,
Korea
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45 and J46)
Study Design:
Bidirectional case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All Ages
Pollutant: CO
Averaging Time: Max 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: JO.O, 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 5:0.99J0.94-1.04); 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: Kontos et al
(1999, 011326) Perio
of Study:
1/1987-12/1992
Location:
Piraeus, Greece
Author: Lee et al.
(2002, 034826)
Period of Study:
12/1997-12/1999
Location:
Seoul, Korea
Hospital Admissions
d
Health Outcome (ICD9):
Respiratory conditions
(laryngitis, bronchiolitis,
tonsillitis, acute
rhinopharyngitis, otitis,
bronchopneumonia,
pneumonia, asthma)
Study Design: Time series
Statistical Analyses:
Stochastic dynamical system
approach
Age Groups Analyzed: 0-1 4 yr
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45, J46)
Study Design: Time series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: <5 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean Range (SD) unit:
1987: 4.2 mg/m3
1992:3.6mg/m3
Range (Min, Max): NR
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
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
This study did not present quantitative results for CO.
Increment: 1.0 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
RR for asthma and exposure to various pollutants for
children under! 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
January 2010
C-52
-------
Study
Author: Lee et al.
(2006, 098248)
Period of Study:
1/2002-12/2002
Location:
Seoul,
Korea
Design
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45-46)
Study Design: Time series
Statistical Analyses:
GAM with stringent parameters
Concentrations
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
Effect Estimates (95% Cl)
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
Age Groups Analyzed: <15yr Range (Min, Max): NR
Copollutant: correlation
N02:r=0.55
S02: r = 0.72
PM10:r = 0.28
03: r = -0.36
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
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
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Bidirectional
case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All 3flPQ
MM ay co
ED Visits
Health Outcome (ICD9):
Respiratory illness (lower
respiratory illness, upper
respiratory illness, wheezing)
Study Design: Time series
Statistical Analyses: Poisson
Age Groups Analyzed: <3 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.77 ppm
Range (Min, Max): (0.23, 1.72)
Copollutant:
PM10
9O,
OW2
N02
n.
U3
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 5 ppm
Range (Min, Max): (1,1 2)
Copollutant: correlation
PM10:r = 0.50
N02: r= 0.35
S02: r = 0.56
03:r = 0.04
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
Lower Respiratory Illness
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
January 2010
C-53
-------
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
PM10.25: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:
Bidirectional case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: <5 yr
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
Oz'. 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
PM2.5:r = 0.10
PMio-2s: r = 0.06
PM10:r = 0.10
S02:r = 0.12
N02: r = 0.20
0- r - n 11
3.1 u. i i
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
ri\r\c
Olllb
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
January 2010
C-54
-------
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
January 2010
C-55
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Luginaah et
al. (2005. 057327)
Period of Study:
4/1995-12/2000
Location:
Windsor, ON, Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory illness (460-519)
Pollutant: CO
Averaging Time: 1-h max
Mean(SD)unit:1.3(1.0)ppm
Study Design1
Time series and case crossover ^"9® (Min, Max): (0,11.82)
Statistical Analyses:
1. Time-series: Poisson
2. Case-crossover: conditional
logistic regression
Age Groups Analyzed:
All ages
0-14 yr
15-64yr
>65yr
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
January 2010
C-56
-------
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
Author: McGowan et
al. (2002. 030325)
Period of Study:
6/1988-12/1998
Location:
Christchurch,
New Zealand
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 step-wise regression
Age Groups Analyzed: Adults
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
Concentrations
Pollutant: CO
Averaging Time: Max 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)
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.16 (1.51) mg/m3
Range (Min, Max): (0, 15.7)
Copollutant: NR
This study did not provide quantitative results for CO.
Statistical Analyses:
Generalized Additive Model
Age Groups Analyzed:
<15yr;>64yr
Author: Migliaretti et
al. (2007. 193772)
Period of Study:
1/1997-12/1999
Location:
Turin, Italy
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: 8-h median
Median (SD) unit: 3.36 (1 .57) mg/m3
Range (Min, Max): NR
Copollutant: correlation
TSP
Increment: 1 mg/m3
Odds Ratio (Lower Cl, Upper Cl); lag:
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)
January 2010
C-57
-------
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, PM10-25:
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, PM10.25:
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
January 2010
C-58
-------
Study
Author: Neidelletal.
(2004, 057330)
Period of Study:
1992-1998
Location:
California
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
Design
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
6-1 2 yr
12-1 Syr
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Time series
Statistical Analyses:
Semiparametric Poisson GAM
Age Groups Analyzed: <8 yr
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
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 .777 (1 .037) ppm
Range (Min, Max): NR
Copollutant: correlation
03
PM10
N02
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
Effect Estimates (95% Cl)
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: -0.01 0(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)
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
January 2010
C-59
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Sauerzapf et
al. (2009,180082)
Period of Study:
Jan 2006-Feb 2007
Location:
Norfolk county,
England
Hospital Admissions Averaging Time: 24 h
Health Outcome: COPD Mean (SD) unit:
Control days: 194.46 (80.93)
Study Design: Case crossover
Case days: 204.73 (11 9.97)
Statistical Analyses: Logistic
Regression Range (min, max):
Control days: 105.20, 408.10
Age Groups Analyzed:
18+ yr (90% of patients 60+ yr) Case days: 108.70, 432.20
Sample Description: Copollutant:
1 ,050 COPD admissions NO, N02, NOX, 03
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
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 Analyses:
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;
PM10.25: 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
PM,:r=0.63
PM25:r = 0.62
PM • r - 0 ?9
rlvlio- 1 — U.O£
PM10.2.5:r=0.32
Increment: 924 ppb
% Increase (Lower Cl, Upper Cl); Lag
CO: 6% (3, 9); 3
CO, PM15: 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);
1.06(1.00-1.111:3
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); 2
1.01 (0.93-1.10); 3
HospitalAdmissions:
All Respiratory Illnesses
Age Group:AII Ages:
0.99(0.95-1.02);1/1.00(0.96-1.04);2
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); 2
1.00 (0.91-1.11); 3
COPD
Age Group: Adults:
0.94 (0.86-1.03); 1 /1.04 (0.95-1.13); 2
0.97 (0.88-1.06); 3
January 2010
C-60
-------
Study
Author: Stiebetal.
(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:
<16yr; 16-55 yr
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;
PMzs: r = 0.27; H+:r = 0.23;
S04 . 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:
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.
January 2010
C-61
-------
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:1.414(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(1.125-1.393);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
January 2010
C-62
-------
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
Physician Visits
Health Outcome (ICD9):
Allergic rhinitis (177)
Study Design: Time series
Statistical Analyses: Poisson
GLM
Age Groups Analyzed: >65 yr
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Cross sectional
Statistical Analyses:
Spearman Rank Correlations
Age Groups Analyzed:
0-14 yr;<2yr; 2-5 yr;>5yr
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All ages
Hospital Admissions
Health Outcome (ICD9):
COPD: (490-492, 494, 496)
Study Design: Case crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
All ages
Concentrations
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.1 (0.4) ppm
Range (Min, Max): (0.0, 2.2)
Copollutant:
PM25
PM10
PM10-2.5
S02
N02
03
Pollutant: CO
Averaging Time: Monthly
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.33 ppm
Range (Min, Max): (0.32, 3.62)
Copollutant:
PM
rlVHo
S02
N02
3
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.33 ppm
Range (Min, Max): (0.32, 3.66) ppm
Copollutant:
PM
rlVHo
S02
N02
3
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(1.019-1.136);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 SO?
<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:1.169(1.102-1.240);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 SO?
<20°C: 0.895 (0.832-0.962); 0-2
>20°C: 1.274 (1.21 9-1 .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 (1.1 85-1 .285); 0-2
January 2010
C-63
-------
Study
Author: Yang etal.
(2005, 090184)
Period of Study:
1/1994-12/1998
Location:
Vancouver,
Canada
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
Design
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Time series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65yr
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
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
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-2s: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
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) ; 0-6
CO, 03, N02, S02: 1.10 (0.98-1 .23); 0-6
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); 1
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
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
January 2010
C-64
-------
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
(FEV,, cystic fibrosis pulmonary
exacerbation)
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population: 11,484 cystic fibrosis
patients
Age Groups Analyzed: >6 yr
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 0.692 (0.295) ppm
IQR (25th, 75th): (0.48,0.83)
Copollutant: NR
Increment: 1.0 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Two or more pulmonary exacerbations during 2000
1.02(0.85-1.22)
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 nonsmoking children
Age Groups Analyzed:
Middle-school children
(mean age = 13.Syr)
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)
January 2010
C-65
-------
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
(BHRj.FEV^SSropred.,
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: 6-1 Syr
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 J1.03-1.09;
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)
January 2010
C-66
-------
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: Meng etal.
(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 nonsmoking 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.
January 2010
C-67
-------
Study
Author: Mortimer et
al. (2008,122163)
Period of Study:
1989-2000
Location:
San Joaquin Valley,
CA
Design
Health Outcome: Lung function
(FVC,FEV1,PEF,FEF25-75,
FEWFVC, FEF25-75/FVC,
FEF25, FEF75)
Study Design: Cohort
Statistical Analyses:
1 . DSA algorithm
2. GEE
Concentrations
Pollutant: CO
Averaging Time:
8-h max monthly mean
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant; correlation:
Effect Estimates (95% Cl)
Increment: NR
Effect Size per IQR Increase in Pollutant (SE):
FEF25-75:
24-h avg CO exposure during 1st trimester
0.90% (0.0113)
FEWFVC
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)
FEWFVC
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/m3
Range (Min, Max): NR
Copollutant: NR
The study did not present quantitative results for CO.
January 2010
C-68
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Sole et al.
(2007, 090706)
Health Outcome: Symptoms of
asthma, rhinitis, and eczema
Period of Study: Study Design: Panel
Location:
Sao Paulo West, Sao
Paulo South, Santo
Andre, Curitba, &
Porto Alegre, Brazil
Statistical Analyses: Logistic
Regression
Age Groups Analyzed:
13-14 yr
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): NR
Copollutant: N02, S02,03
Increment: Risk in relation to center w/ lowest annual
mean (Porto Alegre = ret)
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 (0.91,1.16)
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)
Author: Wane
(1999.008105)
Period of Study:
10/1995-6/1996
Location:
Kaohsiung and
Pintong, Taiwan
et al. Health Outcome: Asthma
Study Design: Cross sectional
Statistical Analyses:
Multiple logistic regression
Population:
165,173 high school students
Age Groups Analyzed: 11-16yr
Pollutant: CO
Averaging Time: Annual median
Median (SD) unit: 0.80 ppm
Range (Min, Max): NR
Copollutant: NR
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)
Author: Wilhelmetal.
(2008, 191912)
Period of Study:
2000-2001
Location:
Los Angeles County
or San Diego County,
California
Health Outcome: Asthma
symptoms/ED visit/HA
Study Design: Panel
Statistical Analyses: Logistic
regression
Age Groups Analyzed: 0-1 7 yr
Sample Description:
612 children who reported a
physician diagnosis of asthma at
some point in their lives
Averaging Time: annual
Mean (SD) unit: 1 .0 ppm
Range (min, max): 0.34, 1 .8
Copollutant: correlation
03: r= -0.67
PM10:r=0.41
PM25:r= 0.60
N02: r= 0.57
traffic density: r= 0.02
Increment: NR
OR Estimate [Lower Cl, Upper Cl] ; lag:
Lags examined: NR
No associations observed between asthma symptom
outcome measures (no results shown)
January 2010
C-69
-------
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)
Design
Health Outcome (ICD9):
Mortality: All-cause
(nonaccidental) (<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
(nonaccidental) (<800);
cardiovascular (390-459);
respiratory (460-51 9)
Study Design: Meta-analysis
Statistical Analyses:
PniQQnn HI M
rUlooUII ULIVI
Age Groups Analyzed:
All ages
Health Outcome: Mortality
£tnHw Daciiin1 Pnhnrt
Concentrations
Pollutant: CO
Averaging Time: Max 8-h ma
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;
PM19.25: r = 0.10; BS:r = 0.77;
SO/: 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.
N02
03
PM10
Averaging Time: 24 h
Matin /£PN unit1 MoHian ralrnlator
Effect Estimates (95% Cl)
Increment: 1 .0 ppm
% Increase (Lower Cl, Upper Cl);
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);
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
\ % Phanno in Haik/ Mnntranma Hoa
lag:
lag:
the Fl nuior PI
Period of Study: 1992-2002
Location: Augsburg,
Germany; Barcelona, Spain;
Helsinki, Finland; Rome,
Italy; Stockholm, Sweden
Statistical Analyses: Poisson
regression analysis
Age Groups Analyzed: > 35 yr
Sample Description: First-
time Ml patients
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
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 15-day
avg.
January 2010
C-70
-------
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
(nonaccidental) (<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: Max 8-h ma
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/m3
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)
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)
Period of Study: 1991 -1993
Location:
Sao Paulo, Brazil
Author: Bremner et al.
(1999.007601)
Period of Study:
1/1992-12/1994
Location:
London, U.K.
Health Outcome (ICD9):
Mortality
Study Design:
Longitudinal study
Statistical Analyses:
State space model
Age Groups Analyzed: 2 65 yr
Health Outcome (ICD9):
Mortality: All-cause
(nonaccidental) (<800);
cardiovascular (390-459);
respiratory (460-519)
Study Design: Time series
Statistical Analyses: Poisson,
cubic splines
Age Groups Analyzed:
All ages
0-64 yr
265yr
65-74 yr
>75yr
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
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: TSP; N02; 03; S02
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 0.8 (0.4) ppm
Range (Min, Max): (0.2, 5.6)
Copollutant:
NO,-
INW2,
n •
US,
S02;
PU.»-
r IVI10,
BS
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)
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
January 2010
C-71
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Burnett et al. (2000, Health Outcome (ICD9):
010273) Mortality: All-cause
Period of Study: 1986-1996 (n°naCCidental)(<800)
Location:
8 Canadian cities
Study Design: Time series
Statistical Analyses:
1. Single-pollutant models:
Poisson GAM, LOESS
2. Multi-pollutant models:
Principal component regression
analysis
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.9 ppm
Range (Max): 7.2 ppm
Copollutant: correlation
03:r=-0.05
S02: r = 0.42
PM25:r = 0.44
PM10.25:r = 0.29
PM10:r=0.45
Increment: 0.9 ppm
% Increase (t-value); lag:
Temporally filtered daily nonaccidental mortality
(days in which PMio data available)
CO: 0.4 (0.4); 0; 2.0(2.3);1
CO, PM25:-0.7(-0.7);0;1.1(1.1);1
CO, PM10-25:0.1 (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, PM10.25, 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
Author: Chock etal. (2000,
010407)
Period of Study:
1989-1991
Location:
Pittsburgh, PA
Health Outcome (ICD9):
Mortality: All-cause
(nonaccidental) (<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
(nonaccidental) (<800); CVDs
J390-459); respiratory diseases
(460-519)
Study Design: Time series
Statistical Analyses: Poisson;
Random effects regression
model
Age Groups Analyzed:
All ages
<64yr
65-74 yr
75-84 yr
285yr
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: 24- h avg
Mean (SD) unit: 1 .02 ppm
Range (Min, Max): NR
Copollutant:
N02;
03;
S02;
PM2.5;
PMlO-2.5
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .29 ppm
Range (Min, Max): NR
Copollutant correlation:
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.
Pollutant: CO
Averaging Time: 1-havg
Mean (SD) unit: NR
Range (Min, Max): NR
copo Mutant:
PM •
r Ivlio,
PM2s;
03;
90 •
OW2,
NO,
INW2
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:
Nonaccidental:
5.88% (6.42); 1 ; 9.39% (6.89); 0-5
CO+PM10+03+S02:6.13% (4.34); 1
Age Group: £64
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
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
Fall: 0.00797(1. 14); 0
January 2010
C-72
-------
Study Design Concentrations Effect Estimates (95% Cl)
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
January 2010 C-73
-------
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
(nonaccidental) (<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
PM1(K,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, PM10-25:1.035 (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, PM10-25:1.049 (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, PM10-25:1.053 (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, PM10.25,S02, N02,03:
1.032 (4.6); 0-1
GAM Filtered Variables
CO: 1.030 (4.3); 0-1
CO, PM25, PM10.25,S02, N02,03:
1.022 (2.4); 0-1
GLM
CO: 1.023 (2.4); 0-1
CO, PM25, PM10.25,S02, N02,03:
1.013 (1.1); 0-1
Author: Conceicao et al.
(2001.016628)
Period of Study:
1994-1997
Location:
Sao Paulo, Brazil
Author: De Leon et al.
(2003, 055688)
Period of Study:
1/1985-12/1994
Location:
New York, NY
Health Outcome (ICD9):
Mortality: Respiratory diseases
(460-519)
Study Design: Time series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: <5 yr
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: Max 8-h ma
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
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
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
The study did not present quantitative results for CO.
January 2010
C-74
-------
Study
Author: Dominici et al.
(2003, 056116)
Period of Study:
1987-1994
Design
Health Outcome (ICD9):
Mortality: All-cause
(nonaccidental);
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
Health Outcome (ICD9):
Mortality: Respiratory;
cardiovascular
Study Design: Time series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time:
24-havg;Max8-havg
Median (SD) unit:
24-havg:1.4(1.0) ppm
Max 8-h avg: 2.1 (1.6) ppm
Range (Min, Max):
24-havg:(0.0,7.6)
Max 8-h avg: (0.2,2.5)
Copollutant: correlation
PM10:r= 0.609;
PM25:r= 0.435;
PM^25:r = 0.326;
COM: r= 0.736;
Increment: 2.2 ppm
Relative Risk (Lower Cl, Upper Cl);
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
Author: Fischer et al. (2003,
043739)
Period of Study:
1986-1994
Location: The Netherlands
Health Outcome (ICD9):
Mortality: Nonaccidental
(<800); pneumonia
(480-486) ;COPD (490-496);
cardiovascular (390-448)
Study Design: Time series
Statistical Analyses:
Poisson GAM, LOESS
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
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^f'ied 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
January 2010
C-75
-------
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 AMI (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; COM;
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
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
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
having
having
January 2010
C-76
-------
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);
CVDs (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):
MatharlanHc- 117 A OROm
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
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: 1.040 (1.020-1.060); 1
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 (nonaccidental) (<800)
Study Design: Time series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65yr
Pollutant: CO
Averaging Time: 24-h avg
Median (SD) unit: 1.6ppm
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)
January 2010
C-77
-------
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) ; nonaccidental
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 (nonaccidental)
(<800); respiratory;
cardiovascular
Study Design: Time series
Statistical Analyses: Poisson
GAM' LOESS
Age Groups Analyzed:
All ages
Health Outcome (ICD9):
Mortality: Nonaccidental
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: Max 8-h ma
Median (SD) unit:
Cold Season: 0.5 mg/m
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)
Author: Knoxetal. (2008,
193776)
Period of Study: 1996-2004
Location: 352 English local
authorities
Health Outcome (ICD9):
Mortality: Nonaccidental
(<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
Health Outcome: Mortality
Study Design: Cross sectional
Statistical Analyses: Linear
regression
Age Groups Analyzed: NR
Sample Description: Data
from Oxford Cancer
Intelligence Unit
Pollutant: CO
Averaging Time: 1-h max
Median (SD) unit:
1,310 (939.13) ppb
Range (Min, Max): (303.58, 7400)
Copollutant:
PM2s; PM10-2.5; 03; N02; S02; Acid;
EC; OC; S04; Oxygenated HCs;
NMHCs;N03
Averaging Time: NR
Meuan (SD) nit: NR
Range (Min, Max): NR
Copollutant: NR
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
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
January 2010
C-79
-------
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
Health Outcome (ICD10):
Mortality: Nonaccidental (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:
Step-wise regression
Age Groups Analyzed:
<65yr
>65yr
Pollutant: CO
Averaging Time: Max 8-h ma
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)
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
Copollutant:
NO;N02;03;S02;S042'
PM10;PM2.5
Joint model with CO
Philadelphia: 0.0059; 0-1
4 Pennsylvania Counties: 0.0089; 0-1
Pennsylvania + NJ: 0.0096; 0-1
Cardiac: 0.0135; 0-1;
Other causes:0.0084
<65:0.0154; 0-1;
> 65:0.0060; 0-1
January 2010
C-80
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lippmann et al.
(2000, 0119138)
Period of Study:
1985-1990
1992-1994
Location:
Detroit, Ml and Windsor, ON
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<800); circulatory (390-459);
respiratory (460-519)
Study Design: Time series
Statistical Analyses:
Poisson GLM
Age Groups Analyzed:
>65yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
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-PM1,: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;
PM^25:r = 0.24;
H+: r= 0.16; S042':r= 0.32;
03: r = 0.16; S02:r = 0.42;
N02: r = 0.68
Increment:
1985-1990:11.5 ppm; 1992-1994:8.4 ppm
Relative Risk (Lower Cl, Upper Cl); lag:
1985-1990
Total Mortality:
0.9842 (0.9667-1.002);0
1.0103 (0.9926-1.0284);1
1.0075 (0.9898-1.0254); 2
1.0145 (0.9967-1.0326); 3
0.9968 (0.9789-1.0151); 0-1
1.0105 (0.9925-1.0288); 1-2
1.0134 (0.9954-1.0317); 2-3
1.0003 (0.9823-1.0187); 0-2
1.0152 (0.9971-1.0336); 1-3
1.0053 (0.9873-1.0236); 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);!
1.0483 (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 J0.9950-1.0516); 0-3
1992-1994
Total Mortality
0.9933 (0.9636-1.024);0
1.0162 (0.9860-1.0473);1
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
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
January 2010
C-81
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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
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:
5:482ug/ms(>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
January 2010
C-82
-------
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:
5:482ug/ms(>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 (nonaccidental)
(<800); cardiovascular (390-
449)
Study Design: Time series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65yr
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;
PM^25: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
January 2010
C-83
-------
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.
January 2010
C-84
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Moolgavkar et al.
(2003, 051316)
Period of Study: 1987-1995
Location:
Cook County, Illinois & Los
Angeles County, California
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<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
Copollutant correlation:
Cook County:
N02:r=0.63;
03:r=-0.22;
S02:r = 0.35;
PM10:r=0.30
LA County:
N02:r=0.80;
03:r=-0.52;
S02:r = 0.78;
PM10:r=0.45;
PM25:r = 0.58
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;
1.0% (1.9); 3;/ 1.1% (2.1); 4;/ 1.4% (2.7); 5
Total Mortality Los Angeles County
CO:
1.3% (7.4); O;/ 1 .9% (10.5); 1 ; / 1 .6% (8.9); 2;
1 .4% (8.1); 3; / 1 .0% (5.9); 4; / 0.7% (4.1); 5
CO, PM10:
0% (0); 0; / 2.2% (4.8); 1 ; / 1 .4% (3.1); 2;
0.8% (1 .8); 3; / 0.7% (1 .6); 4; / 1 .3% (3.0); 5
CO, PM25:
-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% (1 .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):
1.2%(1.5);0;/2.1%(2.7);1;/0%(0);2;
0% (0);3;/-0.5% (-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 % (3.8); 3; / 1 .9% (1 .8); 4; / 1 .4% (1 .3); 5
Fall (CO):
1.8% (4.6)
0.6% (1 .5); 3; / 0.4% (1 .2); 4; / 0.2% (0.6); 5
1.8% (4.6); 0;/2.0% (5.1); 1 ;/ 1.0% (2.6); 2;
.6);
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% (7.1); 4;/ 1.4% (6.1); 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
January 2010
C-85
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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% (0); 3; /-0.8% (-0.7) ; / 4; 0% (0) ; 5
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
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<800); respiratory (460-519);
cardiovascular (393-440)
Study Design: Time series
Statistical Analyses: Poisson
GAM; LOESS
Age Groups Analyzed: >50 yr
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<800); respiratory (460-51 9);
cardiovascular (393-440)
Study Design: Time series
Statistical Analyses: Poisson
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: Max 8-h avg
Median unit: 1 .2 mg/m3
Range (Min, Max): (0,12.4)
Copollutant correlation:
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
GAM,LOESS
Age Groups Analyzed:
All ages
15-64yr
65-74 yr
>75
03: r=-0.46; N02:r = 0.59;
S02: r= 0.55; PM10:r= 0.45;
TSP:r=0.26;
TSP Blackness: r = 0.26
Cardiovascular Mortality
-2.48% (-4.30,-0.66);0
-0.84% (-2.61, 0.93); 1
-1.87% (-4.43, 0.69); 0-
Respiratory Mortality
-0.48% (-4.84, 3.87); 0
-0.14% (-4.43,4.15);!
-1.49% (-7.73, 4.74); 0-3
Author: Peters et al. (2000,
001756)
Period of Study:
1982-1994
Location:
Northern Bavaria (Rural
Germany) and the Coal
Basin of the Czech Republic
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Cardiovascular (390-
459); Respiratory (460-519);
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
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
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 J0.995,1.028); 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
CVD Mortality:
1.018 (0.994,1.044); 0; /1.012 (0.987,1.038); 1;
1.016 (0.991,1.041); 2; /1.004 (0.980,1.029); 3
January 2010
C-86
-------
Study
Author: Rainhametal.
(2003, 053202)
Period of Study:
1980-1996
Location:
Toronto, ON, Canada
Design
Health Outcome (ICD9):
Mortality: Cardiac (390-459);
Respiratory (480-519); Total
(non-accidental) (<800)
Study Design: Time-series
Statistical Analyses:Poisson
fiAM natural nihir snlinps
Concentrations
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
Effect Estimates (95% Cl)
The study did not present quantitative results for CO.
Location: Amsterdam
Age Groups Analyzed:
<65
>65
Author: Roemer et al.
(2001.019391)
Period of Study:
1/1987-11/1998
Health Outcome (ICD9): Pollutant: CO
Mortality: Total (non-accidental)
(<800) Averaging Time: 24-h avg
Study Design: Time-series Mean (SD) unit:
Increment:
Lag1 and 2: 100 ug/m3
Lag 0-6: 50 ug/m
Relative Risk (Lower Cl, Upper Cl); lag:
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
Air pollution background:
836 ug/m3
Air pollution traffic: 1805 ug/m
Range (10th, 90th):
Air pollution background:
(448,1315) ug/nf
Air pollution traffic:
(727, 3192) ug/m3
Copollutant:
BS;PM10;S02;N02;NO;03
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 (1.003-1.013); 2;
1.003 (0.999-1.007); 0-6
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
January 2010
C-87
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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.
098420)
Period of Study:
1990-1997
Location:
19 European Cities
(APHEA2)
(2007, Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Cardiovascular (390-
459)
Study Design: Time-series
Statistical Analyses:
Poisson and two-stage
hierarchical model
Age Groups Analyzed:
All ages
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:
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)
Increment: 1 mg/m
% 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
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
January 2010
C-88
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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
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)
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
January 2010
C-89
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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 (nonaccidental)
(<800)
Study Design: Time series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 1-havg
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.
Author: Sharovskyetal.
(2004,156976)
Period of Study:
1996-1998
Location:
Sao Paulo, Brazil
Health Outcome (ICD10):
Mortality: Ml (1.21)
Study Design: Time series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
35-109 yr
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
Increment: NR
px100(SE);lag:
CO: 1.42 (1.01)
CO, S02,PM10:0.97 (1.27)
Notes: The study did not present the lag used for
CO.
Author: Slaughter etal.
(2005, 073854)
Period of Study:
1/1995-6/2001
Location:
Spokane, WA
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<800); respiratory (460-519);
asthma (493) ;COPD (491,
492, 494, 496); pneumonia
(480-487); acute upper
respirator y tract infections
(464-466,490); cardiac
outcomes (390-459)
Study Design: Time series
Statistical Analyses:
Log-linear Poisson GLM,
natural splines for calendar
time
Age Groups Analyzed:
All ages
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
Range (Min, Max): NR
Copollutant correlation:
PM,: r = 0.63; PM25:r = 0.62;
PM10:r=0.32;
PM10.2.5:r = 0.32
The study did not present quantitative results for CO.
January 2010
C-90
-------
Study
Author: Stiebetal. (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 et al. (2002,
034835)
Period of Study:
1985-1995
Location:
Barcelona, Spain
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
Design Concentrations
Health Outcome (ICD9): Pollutant: CO
Mortality: Nonaccidental
Averaging Time: 24-h avg
Study Design: Meta-analysis
Mean (SD) unit: NR
Statistical Analyses: NR
J IQR (25th, 75th) : NR
Age Groups Analyzed:
All ages Copollutant: NR
Health Outcome (ICD9): Pollutant: CO
Mortality: Total (nonaccidental)
(<800); cardio-respiratory (390- Averaging Time: 24-h avg
459, 460-519, 785, 786) Megn (SD) „„„. „ 4? (Q 3g) mg/ms
Study Design: Time series |QR (25(h ?5th). (Q 23 „ 5?)
Statistical Analyses: Poisson Copo||utant corre|ation:
tjMM MC0.1-0.5:r=0.58;
Age Groups Analyzed: MCO.01-2.5: r = 0.57;
Allages P^.r^^NO^O.^;
Health Outcome (ICD9): Pollutant: CO
Mortality: COPD (491, 492,
494 496) Averaging Time: 8-h avg
Study Design: Mean (SD) unit: NR
Bidirectional case crossover D ,... .. , MD
Range (Mm, Max): NK
Statistical Analyses: Copollutant: PMin'N02-03
Conditional logistic regression "uh""lu"""- rmio, 11^2,^3
Age Groups Analyzed: >35yr
Health Outcome (ICD9): Pollutant: CO
Mortality: Respiratory mortality
Averaging Time: 24-h avg
Study Design: Case crossover ,
Median (SD) unit: 7.7 ug/m3
Statistical Analyses:
Conditional logistic regression Ran9e (Min> Max): (0.6, 66.0)
Age Groups Analyzed: >14yr Copollutant:
PM10;BS;N02;03;S02
Study population:
Asthmatic individuals: 5,610
Health Outcome (ICD9): Pollutant: CO
Mortality: Total (nonaccidental)
(<800) ; respiratory (460-51 9) ; Averaging Time: 24-h avg
circulatory (390-459) Megn (SD) unjt. „ 82? ppm
I!dtyc3cnase crossover Ran9e : <°-226< 1 770>
Statistical Analyses: no'50 cJ?M,~, r,
_ .... , , • i- • PMm SO? NO? 03
Conditional logistic regression ' ' '
Age Groups Analyzed:
All ages
Health Outcome (ICD9) : Pollutant: CO
Mortality: Total (nonaccidental)
(<8oo) Averaging Time: 24-h avg
Study Design: Case crossover Mean (SD) unit: 8.27 ppm
Statistical Analyses: Range (Min, Max): (2.26, 17.70)
Conditional logistic regression Copollutant1
Age Groups Analyzed: PM10; S02; 63; N02
27 days old to <1 yr of age
Effect Estimates (95% Cl)
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
Increment: 0.313 ppm
Odds Ratio (Lower Cl, Upper Cl); lag:
Total (nonaccidental): 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
January 2010
C-91
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Vedal et al. (2003,
039044)
Period of Study:
1/1994-12/1996
Location:
Vancouver, BC, Canada
Health Outcome (ICD9):
Mortality: Total (nonaccidental)
(<800); respiratory (460-519);
cardiovascular (390-459)
Study Design: Time series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
All ages
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
The study did not present quantitative results for CO.
Author: Villeneuve et al.
(2003, 055051)
Period of Study:
1986-1999
Location:
Vancouver, BC, Canada
Health Outcome (ICD9):
Mortality: Nonaccidental
(<800); cardiovascular
(401-440); respiratory
(460-51 9); cancer (140-239)
Study Design: Time series
Statistical Analyses:
Poisson, natural splines
Age Groups Analyzed: > 65 yr
Pollutant: CO
Averaging Time: 24- h avg
Mean (SD) unit: 1 .0 ppm
Range (Min, Max): (0.2, 4.9)
Copollutant:
PM25;PM10;PM10.25;TSP;
S04;CO;COH;03;N02;S02
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
January 2010
C-92
-------
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: Nonaccidental
(<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. (2004,
055603)
Period of Study:
1994-1998
Location:
Taipei, Taiwan
Health Outcome (ICD9):
Mortality: Nonaccidental
(<800); circulatory (390-459);
respiratory (460-519)
Study Design:
Bidirectional 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: 1.014 (0.925-1.110); 0-2
Circulatory: 0.996 (0.948-1.046); 0-2
January 2010
C-93
-------
Table C-8. Studies of long-term CO exposure and mortality.
Study
Author: Krewski et al.
(2009, 191193)
Period of Study:
1983-2000
Location:
United States
Author: Lipfertetal.
(2000, 004087)
Period of Study:
1975-1996
Location:
32 Veterans Hospitals,
USA
Design
Health Outcome: Mortality
Study Design: Cohort
Statistical Analyses: Random
effects Cox model
Age Groups Analyzed: 30+ yrs
Sample Description:
508,538 adults living in large US
cities
Mortality
Health Outcome (ICD9):
Nonaccidental
Study Design: Cohort
Study Population:
-90,000 hypertensive male U.S.
veterans
Statistical Analyses:
Staged regression
Age Groups Analyzed: NR
Concentrations
Averaging Time: 1980 annual avg
Mean (SD) unit:
1.68 (0.66) ppm
Range (min, max):
0.19,3.95
Co pollutant:
PM15, PM2.5, S02, S04, TSP, 03, N02
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1960-1974: 10.82 (5.15) ppm
1975-1 981: 7.64 (2.94) ppm
1982-1 988: 3. 42 (0.95) ppm
1989-1 996: 2. 36 (0.67) ppm
Range (Min, Max):
1960-1 974: (0.94, 35.30)
1975-1 981: (0.43, 22.38)
1982-1 988: (0.30, 15.20)
1989-1 996: (0.30, 7.10)
Copollutants; correlation:
1960-1974:
03:r = 0.004;
NO,: r = 0.690;
S04:r= 0.469
1975-1981:
03:r = 0.109;
NO,: r = 0.249;
S04:r= -0.155;
IP S042":r = 0.356;
PM25:r = 0.634;
PM10.25: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;
PM10.25:r=0.135
PM15:r = 0.284
1989-1996
03:r = 0.397;
NO,: r = 0.492;
S04:r= -0.551
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)
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: -0.01 70
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: -0.01 28
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.
January 2010
C-94
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lipfert and
Morris (2002,019217)
Period of Study:
1960-1997
Location:
U.S. counties
Mortality
Health Outcome (ICD9):
Nonaccidental
Study Design:
Ecological/ cross sectional
Statistical Analyses:
Staged regression
Age Groups Analyzed:
15-44yr
45-64 yr
65-74 yr
75-84 yr
>85yr
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit:
1960-1969:13.81 (8.47) ppm
1970-1974:9.64 (5.63) ppm
1979-1981:5.90 (3.54) ppm
1989-1991:2.69 (1.22) ppm
1995-1997:1.72 (0.76) ppm
Range (Min, Max): NR
Co pollutant:
TSP
S042'
S02
N02
03
Increment: NR
Attributable risk (SE):
Attributable Risks of mortality (1960-4)
Peak CO 1960-1964, All locations
Ages 15-44:0.1299 (0.0341)
Ages 45-64:0.0340 (0.0280)
Ages 65-74: -0.0058 (0.0220)
Ages 75-84:0.0121 (0.0188)
Ages > 85:0.0374 (0.0225)
Log Mean: 0.0365 (0.0149)
Attributable Risks of mortality (1970-4)
Peak CO 1970-1974, All locations
Ages 15-44:0.0553 (0.0240)
Ages 45-64:0.0181 (0.0148)
Ages 65-74:-0.0146 (0.0134)
Ages 75-84:-0.0128 (0.0098)
Ages > 85:-0.0151 (0.0093)
Log Mean: 0.0038 (0.0086)
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
January 2010
C-95
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
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)
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)
January 2010
C-96
-------
Study
Design
Concentrations
Effect Estimates (95% Cl)
Author: Lipfertetal.
(2006, 088218)
Period of Study:
1976-2001
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Nonaccidental
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
Peak03:r= 0.08
Peak S02:r =-0.05
PMzs:r = 0.08
S04 .r=-0.16
Note: VKTA= annual vehicle-km
traveled/km2
Increments ppm
Relative risk (Lower Cl, Upper Cl):
CO: 1.032 (0.954-1.117)
CO, InVKTA: 0.999 (0.923-1.081)
CO, InVKTA, N02:1.012 (0.923-1.110)
CO, InVKTA, N02+03:1.023 (0.939-1.115)
Author: Lipfertetal.
(2006, 088756)
Period of Study:
1997-2002
Location:
32 Veterans Hospitals,
USA
Author: Jerrettetal.
(2003, 087380)
Period of Study:
1982-1989
Location:
107 U.S. cities
Mortality
Health Outcome (ICD9):
Nonaccidental
Study Design: Cohort
Study Population:
-18,000 hypertensive male U.S.
veterans
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: NR
Mortality
Health Outcome (ICD9):
Cardiovascular; CHD;
Cerebrovascular disease
Study Design: Cohort
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1999-2001: 1.63 (0.84) ppm
1 999-2001 (STN sites only):
1.73(0.77)
Range (Min, Max):
1999-2001 '(040 67)
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
PM25 comp.:r = 0.133
N02:r=0.418
Peak03:r=0.172
Peak S02:r = 0.405
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 1.56 ppm
Range (Min, Max): (0.1 9, 3.95)
f^nnnlliitantc nnrratatinn1
Increment: NR
p coefficient (SE); t -statistic:
-0.00000536(0.0000324);-0.165
Increment: 1 ppm
Relative risk (Lower Cl, Upper Cl):
CO: 0.98 (0.92-1 .03)
CO, Sulfates: 0.97 (0.92-1 .03)
Study Population:
65,893 postmenopausal women
without previous CVD
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: > 30 yr
Sulfates: r= -0.07
N02
03
S02
January 2010
C-97
-------
Study
Author: Miller etal.
(2007, 090130)
Period of Study:
1994-1998
Location:
36 U.S. cities
Author: Pope et al.
(2002, 024689)
Period of Study:
1980-1998
Location:
All 50 States, Washington
DC, and Puerto Rico
Design
Mortality
Health Outcome (ICD9):
Cardiovascular; CHD;
Cerebrovascular disease
Study Design: Cohort
Study Population:
65,893 postmenopausal women
without previous CVD
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: 50-79 yr
Mortality
Health Outcome (ICD9):
Total (nonaccidental) (<800); lung
cancer (162); cardiopulmonary
(401-440,460-519)
Study Design: Prospective cohort
Concentrations
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutants:
PM2.5
PM 10-2.5
S02
N02
03
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
Effect Estimates (95% Cl)
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, PM2.5, PMio-2.5, S02, N02, 03: 0.93 (0.67, 1 .30)
The study presents results for CO graphically.
(ACS-CPS-II)
Statistical Analyses:
Cox proportional hazards model
Age Groups Analyzed: > 30 yr
Co pollutant:
PM2.5;PM10;TSP;S02;N02;03
January 2010
C-98
-------
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 Curr
Iss, 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, 120: 949-955. 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
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).
January 2010 C-99
-------
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
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
Botter 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 PFTN; 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
Chen L; 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
January 2010 C-100
-------
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 Curr Iss, 70: 2021-2026. 093034
Chiu HF; Cheng MH; Yang CY (2009). Air pollution and hospital admissions for pneumonia in a subtropical city: Taipei,
Taiwan. Inhal Toxicol, 21: 32-37. 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
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
Chuang KJ; Coull BA; Zanobetti A; Suh H; Schwartz J; Stone PH; Litonjua A; Speizer FE; Gold DR (2008). Particulate
Air Pollution as a Risk Factor for ST-Segment Depression in Patients With Coronary Artery Disease. Circulation,
118: 1314-1320. 155731
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
DTppoliti 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
Delfino RJ: Staimer N; Tjoa T; Gillen DL; Polidori A; Arhami M; Kleinman MT; Vaziri ND; Longhurst J; Sioutas C
(2009). Air pollution exposures and circulating biomarkers of effect in a susceptible population: clues to potential
causal component mixtures and mechanisms. Environ Health Perspect, 117: 1232-1238. 200844
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
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
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
Fischer P; Hoek G; Brunekreef B; Verhoeff A; van Wrjnen J (2003). Air pollution and mortality in the Netherlands: are the
elderly more at risk? Eur Respir J, 40: 34S-38S. 043739
January 2010 C-101
-------
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
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. Eur Respir 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
January 2010 C-102
-------
Ha E-H; Hong Y-C; Lee B-E; Woo B-H; Schwartz J; Christian! 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
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; Christiani 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; Kwon H-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
January 2010 C-103
-------
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
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, C A.
091262
Jalaludin B; Marines 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. J Expo Sci Environ Epidemiol, 16: 225-237. 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 Curr Iss, 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 SY; 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. Report Nr. 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
January 2010 C-104
-------
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; D'Ippoliti 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
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 andpreterm 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 hemo static/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
January 2010 C-105
-------
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
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. Health Effects Institute. Cambridge, MA. 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 T J; 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). Relacao entre baixo peso ao nascer e a poluicao do ar no Municipio de Sao Paulo
[Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil]. Rev Saude Publica, 39:
965-972. 089824
January 2010 C-106
-------
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
Min JY; Paek D; Cho SI; Min KB (2009). Exposure to environmental carbon monoxide may have a greater negative effect
on cardiac autonomic function in people with metabolic syndrome. Sci Total Environ, 407: 4807-4811. 199514
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. JAllergy ClinImmunol, 121: 1133-1139. 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
January 2010 C-107
-------
Parker JD; Woodruff TJ; Basu R; Schoendorf KG (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 JEpidemiol, 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
Pereira Filho MA; 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
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 CA III; 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
JEpidemiol, 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
January 2010 C-108
-------
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
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 MP
(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. NEngl JMed, 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
January 2010 C-109
-------
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
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
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
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
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
January 2010 C-110
-------
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
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. JExpo 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 Curr I ss, 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 Curr Iss, 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; 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
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
January 2010 C-111
-------
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
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. Health Effects Institute.
Boston, MA. 013912
Wilhelm M; Meng YY; Rull RP; English P; Balmes J; Ritz B (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: 110-115. 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 Curr Iss, 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 Curr Iss, 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 Curr Iss, 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 Curr Iss, 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
January 2010 C-112
-------
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. Circulation, 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
January 2010 C-113
-------
Annex D. Controlled Human
Exposure Studies
Table D-1. Controlled human exposure studies.
Study
Subjects
Exposure
Findings
Adiretal. (1999,
0010261
15 healthy nonsmokers Inhaled Concentration: Not provided
Gender: M Exposure Duration: 3 min 45 s
Age: 22-34 yr 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 postexposure
exercise duration (Bruce protocol) relative to clean air
exposure in 13 out of 15 subjects (p = 0.0012). Statistically
significant decreases in METs were also reported following CO
exposure (p = 0.0001). No CO-induced changes in HR, BP,
ECG parameters, or myocardial perfusion were observed.
Bathoorn et al. (2007,
1939631
19 former smokers with
COPD
Gender: 18 M/1 F
Age: 66-70 yr
Inhaled Concentration: 100 ppm (9 subjects) or
125ppm (10 subjects)
Exposure Duration: 2 h on each of 4
consecutive days
COHb Concentration: 2.7% (following 4th 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.
Following the 4th day of exposure, CO inhalation reduced
sputum eosinophils relative to room air and also increased the
provocative concentration of methacholine required to cause a
20% reduction in FEV,. 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 2 of
these patients experienced exacerbations of COPD during or
following CO exposure, with 1 patient requiring hospitalization
2 mo after exposure (initial symptoms first experienced 1 wk
postexposure).
Hanada et al. (2003, 20 healthy adults
Gender: M
Age: 26 + 1 yr
Inhaled Concentration: Not provided
Exposure Duration: 20 min
COHb Concentration: 20-24%
COHb Analysis: CO-oximeter (OSM-3)
15 subjects exposed for 20 min (10 min rest,
5 min handgrip exercise, 2 min postexercise
ischemia, 3 min recovery) under the following 4
conditions: (1) normoxia (inspiratory 02 fraction
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.
21.4%);
10.3%);
2) hypoxia (inspiratory 02 fraction
3) CO + normoxia; and (4) CO +
hyperoxia (inspiratory 02 fraction 95.9%). Trials
involving exposure to CO were conducted last in
this sequence. Each of the 4 conditions was
separated from the next by 20 min of rest.
5 subjects served as controls (4 consecutive 20
min periods of normoxia).
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).
January 2010
D-1
-------
Study
Subjects
Exposure
Findings
Kizakevich et al. (2000, 16 healthy nonsmokers Inhaled Concentration: Initial short term
0526911 _ . .. (4-6 min) exposure to 1,000 or 3,000 pap,
followed by exposures to 27, 55, 83, or 100 ppm
to maintain COHb concentration.
Gender: M
e:18-29yr
Exposure Duration: 4-6 min at 1,000 or
3,000 pap, followed by 20 min at 27, 55, 83, or
100 ppm.
Target COHb Concentrations: 5,10,15, and
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,
13 healthy nonsmokers Inhaled Concentration: 500 ppm
Gender: M Exposure Duration: 1 h
Age: 18-38 yr COHb Concentration: 7%
COHb Analysis: CO-oximeter (AVL 912)
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 (LPS).
Infusion of LPS 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.
Morse et al. (2008,
0979801
12 healthy nonsmokers
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
Exposures conducted on2 separate occasions to
room air (6 min) and CO. Subjects were exposed
to CO until COHb reached 6% (3- to 8-min
exposures).
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).
Ren etal. (2001,
1938501
12 healthy adults
(10 nonsmokers and 1
smoker)
Gender: 9 M/3 F
Age: 20-32 yr
Inhaled Concentration: 0.4% (4,000 ppm)
Exposure Duration: 10-30 min at 0.4% followed
by ~ 8-h with periodic exposure to maintain
COHb concentration
COHb Concentration: 10%
COHb Analysis: Not provided
Each subject underwent 4 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.
A statistically significant increase in ventilation was observed
following hypoxia, but no such increase was found following
any of the other 3 protocols, including exposure to CO. One
subject felt faint during the blood withdrawal protocol and did
not complete the study.
January 2010
D-2
-------
Study
Subjects
Exposure
Findings
Resch et al. (2005,
1938531
15 healthy nonsmokers
Gender: M
Age: 27 + 4 yr
Inhaled Concentration: 500 ppm
Exposure Duration: 1 h
COHb Concentration: -10%
COHb Analysis: CO-oximeter (AVL 912)
Exposures to CO and synthetic air control were
separated by a period of at least 1 wk.
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.
Vesely et al. (2004,
1940001
10 healthy nonsmokers
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)
Prior to and following exposure, subjects
performed hypoxic and hyperoxic rebreathing
tests. Four subjects were exposed to hypoxic
conditions first, while 6 subjects were exposed to
hyperoxic conditions first, both prior to and
following CO exposure.
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.
Zevinetal. (2001,
0211201
12 healthy smokers
Gender: M
Age: 27-47 yr
Inhaled Concentration: 1,200-1,500 ppm
Exposure Duration: 10 min each h, 16 h each
day, over 7 days
COHb Concentration: 5-6%
COHb Analysis: CO-oximeter (Ciba Corning
2500)
Exposures were conducted over 21 consecutive
days underS different protocols, with each
protocol lasting 7 days. In 1 protocol, subjects
smoked 20 cigarettes per day, 1 every 45 min. In
the other 2 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 3 protocols, with 6
subjects exposed sequentially to CO, smoking,
then air, and the other 6 exposed sequentially to
air, smoking, then CO.
COHb levels were similar during smoking and exposure to CO,
with average 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 3 exposures.
January 2010
D-3
-------
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: A pilot study. Eur Respir J, 3 0:
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
inhalation 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
Ren X; Dorrington KL; 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 P III; 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).
January 2010 D-4
-------
Annex E. Toxicological Studies
Table E-1. Human and animal studies.
Reference
Species / Model
Exposure
Duration
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 nonpregnant
myometrium. The HO activator hemin inhibited spontaneous and
oxytocin-induced contractility of the myometrium. Progesterone
induced HO-1 and HO-2 mRNA expression.
Achouha et al. (2008, Human arteries
1799181
Until equilibrium Approximately 30 pM
CO induced endothelium- and NO-independent relaxation of
precontracted human ITA and RA graft by partially stimulating
cGMP production. The mechanism and extent of relaxation
depended upon the tissue.
Ahmed et al. (2000, Human placenta
1938631
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.
Alexander et al. (2007, Rat
1938691 Sprague Dawley
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 et al. Rat
(2002,1923731 Sprague Dawley
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
Lawson ((2003,
1938711
Rat
Sprague Dawley
Adult female
Modulation of the HO/CO system in the anterior pituitary
female rat led to altered secretion of gonadotropins and |
I of the
prolactin.
Alexandreanu and
Lawson (2003,
1938761
Rat
Sprague Dawley
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 3 other electron transport proteins.
Andersen et al. (2006,
1804491
Rat
Long Evans
Male
Mouse
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.
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).
January 2010
E-1
-------
Reference Species / Model
Exposure
Duration
CO Concentration
Findings
Antonelli et al. (2006, Rat
1949601 Wistar
GD5-GD20
75ppm
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.
Appleton and Marks Human placenta
(2002, 1939351
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.
Ashfaq et al. (2003, Human placenta
1940021
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 90 or 180 ppm Skeletal abnormalities: Three pups (from n = 123) in the 180 ppm
exposure over CO group had deformities in their extremities at birth, whereas no
gestation 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.
(2002, 0431611
Human placenta
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 pM.
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- and 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.
Bainbridge and Smith Human placenta
(2005, 1939461
The role of HO in the placenta and during pregnancy is reviewed
in this article. The conflicting data on the activity, localization and
expression of HO in the placentas of pre-eclamptic women are
presented.
Bamberger et al. Human placenta
(2001,0162711
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.
Barber et al. (1999, Human myometrium
1939531
HO and NOS did not maintain human uterine quiescence during
pregnancy.
Barber et al. (2001, Human placenta
1938911
Women who had pregnancies with fetal growth restrictions (FGR)
produced term placenta with significant decreases in HO-2 vs
healthy pregnancies.
Baum et al. (2000,
0164351
Human
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, Rat
1938921 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
1939671 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.
January 2010
E-2
-------
Reference
Bingetal. (1995,
0794181
Burmester et al.
(2000, 099998)
Bye et al. (2008,
1937771
Cagianoetal. (1998,
0871701
Carmines and
Rajendran (2008,
1884401
Carratuetal. (1993,
0138121
Carratuetal. (1995,
079427)
Carratu et al. (2000,
0159351
Carratu et al. (2000,
0158391
Carraway et al. (2002,
0260181
Cella et al. (2006,
1932401
Species / Model
Rodent
Human
Mouse
Rat
Wistar
Female
Rat
Wistar
Female
Rat
Sprague Dawley
Rat
Wistar
Male pups
Rat
Wstar
Rat
Wstar
Rat
Wstar
Rat model of hypoxic
pulmonary vascular
remodeling
(Strain of rat not stated)
Rat
Sprague Dawley
Exposure CO Concentration Findings
Duration
Spatial learning in the Morris water maze was enhanced in
rodents exposed to the HO inhibitor tin protoporphyrin (Sn-PP).
Nb had a high oxygen affinity similar to Mb, and thus may
increase the availability of 02 to brain tissue.
100 h/wkfor 200 ppm CO-exposed (11-14.7% COHb) rats experienced a 24% decrease
18 mo in aerobic capacity evidenced by V02 max deficits. Left
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.
GDO-GD20 75 or 1 50 ppm 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 decrease in ejaculation frequency. Basal
extracellular dopamine 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 in CO-exposed rats.
GD6-GD1 9 of 600 ppm Significant decreases in birth weight were reported after CO
gestation for exposure. Maternal body weight was unchanged during
2 h/day gestation, but corrected terminal body weight (body weight minus
uterine weight) was significantly elevated in CO-exposed dams at
term.
GDO-GD20 75 or 150 ppm Prenatal CO exposure slowed the inactivation kinetics of
transient sodium current in the sciatic nerve fibers of 40-day-old
male rats. The maximum number of activatable Na channels at
normal resting potential was increased in CO exposed rats, and
the voltage-current relationship showed a negative shift of
sodium equilibrium potential.
150 ppm Sphingolipid homeostasis was disrupted in male offspring of
prenatally exposed rats, without a disruption in motor function.
GDO-GD20 150 ppm 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, and 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.
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.
3 wk Hypobaric hypoxia ± CO promoted remodeling and increased pulmonary vascular
50 ppm 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.
HO-1 production and HO concentration were shown to be
regulated by estrogen in the rat uterus.
Chen (2001,193985) Rat
Long Evans
Male
2 mo
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.
January 2010
E-3
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Chung et al. (2006, Rat
1939871 Sprague Dawley
Male
3-6%
CO inactivation of Mb did not induce any change in the
respiration rate, contractile function or high-energy phosphate
levels in perfused rat hearts.
Cronje et al. (2004, Rat
1804401 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
Cudmore et al. (2007, Human placenta
1939911
Human (HUVEC)
Mouse
(HO-1 deficient mouse on
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,
1939921
Human embryonic kidney 0-30 min
(HEK293) cells
20 pM
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,
1939941
Human
(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.
DeLucaetal. (1996, Rat GDO-GD20 75 or 150 ppm Prenatal CO (150 ppm) delayed development of the ion channels
0809111 Wistar responsible for passive and active membrane electrical
Female properties of skeletal muscle. CO-induced lower values of resting
Male pups 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
GDO-GD20 75 or 150 ppm Animals exposed to the higher dose of CO (150 ppm) in utero
had significantly impaired acquisition (at 3 and 18 mo) and
reacquisition (at 18 mo) of conditioned avoidance behavior.
Denschlag et al.
(2004, 1938941
Human
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. Rat GDO-GD20 75 and 150 ppm CO (150 ppm) reduced the minimum frequency of ultrasonic calls
(1993, 0138221 Wistar as well as decreased responsiveness to a challenge dose of
Female diazepam. There was no change in locomotion; however CO
impaired learning in a two-way active avoidance task.
Dubois et al. (2002,
1939111
Rat
Wistar
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 Ca '-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.
January 2010
E-4
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Dubois et al. (2005,
1804351
Rat
Wistar
Male
21 days
SOppm
CO attenuated PAHT by activating BKca channels in PA myocytes
and reduced hemodynamic changes of PAHT.
Dubois et al. (2003, Rat
1804391 Wistar
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
Continuous CO 150 ppm 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
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 et al.
0112941
Rat
Long Evans
Continuous CO 150 ppm
exposure
throughout
pregnancy
CO-exposed animals had cardiomegaly at birth (wet heart
weight) that dissipated by PND4.
Fechter and Annau
(1980, 0112951
Rat
Long Evans
Continuous CO 150 ppm
exposure
throughout
pregnancy
CO-exposed animals had decreased birth weight, impaired
righting reflexes, impaired negative geotaxis, and delayed
homing behavior.
Fechter etal. (1987,
0121941
Rat
Long-Evans
Male
1-4 mL/100 g BW(ip) 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.
Fechter etal. (1987,
0122591
Rat
Long Evans
Male
Continuous CO 75,150, or 300 ppm
exposure
throughout
pregnancy or from
GDOtoPNDIO
The neostriatum of each PND21 rat brains was 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).
Fechter et al. (1997, Guinea pigs
0813221
35 ml/kg gas (ip)
40% COHb
CO impairs high-frequency auditory sensitivity, shown by
increased compound action potential threshold at higher test
frequencies. Free radical inhibitors blocked this response.
Fechter et al.
0120301
Reviews the effects of carbon monoxide on brain development.
January 2010
E-5
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Garofolo et al. (2002,
1939301
Human infants
Rat
Rat: PND2-PND5
Human infants who die from SIDS showed 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.
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
50 ppm
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 an 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
exposure 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,
1939331
Rat 2 h/day, 7 days/wk Cigarette smoke:
Sprague Dawley by nose-only 150, 300, or 600 mg/m
inhalation
Total Particulate Matter
Males: 4 wk prior (TPM)
to and during
mating; and
Females: 2 wk
prior to mating;
during mating;
and through
weaning to
PND21
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.
Ghio et al. (2008,
0963211
Rat
Sprague Dawley
Adult male
24 h
50 ppm
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.
Human bronchial 2-24 h 10-100 ppm CO exposure for 24 h led to a dose-dependent decrease in
epithelial cells cellular non-heme iron, with the effect at 10 ppm statistically
(BEAS-2B) 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
Male and pregnant female
GDO-GD20 75 or 150 ppm This study showed that CO- exposed (75 and 150 ppm) 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
GD20 were reported (0 ppm: 1.6 + 0.1; CO 75 ppm: 7.36 + 0.2;
CO 150 ppm: 16.1+0.9).
January 2010
E-6
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Giustinoetal. (1993, Rat GDO-GD20 75or150ppm CO exposure in utero led to a reversible and dose-dependent
0138331 Wistar loss of function of splenic macrophages, with decreased killing
ability, decreased phagocytosis, and decreased ROS production
during the macrophage respiratory burst.
Giustinoetal. (1994, Rat GDO-GD20 75or150ppm CO (150 ppm) decreased the number of leukocyte common
0763431 Wistar antigen (LCA+) cells at PND21. This was reversed by PND540.
Male pups CO (75 ppm), and other measures of immunological changes
showed trends toward reduction (macrophages, T cells, B cells,
and MHC II cells).
Glabeetal. (1998,
0867041
Rat
Sprague Dawley
Male,
Myocardium
pCO = 0-107 Torr Increased pCO and increased COMb saturation did not alter
high-energy phosphate signals (ATP, phosphocreatine, Pj). MV02
began to decline at 87.6% COMb and is likely not due to
cytochrome c oxidase inhibition.
Graver et al. (2000,
0104651
Fetal lamb
(mixed breed)
10 min
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.
Hara et al. (2002, Rat 40 min 1,000-3,000 ppm CO exposure increased extracellular dopamine levels and
0374971 Sprague Dawley decreased its major metabolites in a Na -dependent pathway. CO
Male 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.
Harada et al. (2004,
1939201
Pig
Granulosa cells
In this porcine model, HO was able to augment granulosa cell
apoptosis allowing for proper follicular maturation.
Hendler and Baum Human
(2004, 1939251
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 Human
(1998, 0520191
Partitioning of 02 and CO in the human embryonic Hb is
discussed.
Iheagwaraetal.
(2007, 1938611
Mouse
C57BL/6
Male
3h
1,000 ppm
CO significantly reduced cytochrome c oxidase activity and Vmax
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.
Imaietal. (2001,
1938641
HO-1 transgenic mice
which specifically over-
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.
Ischiropoulos et al. Rat
(1996,0794911 Wistar
Male
200-290 g
60 min 1,000-3,000 ppm CO poisoning resulted in free NO in brains as measured by
electron paramagnetic resonance spectroscopy and in a 10-fold
40-60 min 1,000 ppm 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
following 40-min but not 60-min exposure to 1,000 ppm CO.
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.
Johnson and Johnson Rat
(2003, 0536111 Sprague Dawley
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. Pretreatment 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.
January 2010
E-7
-------
Reference
Species / Model
Exposure CO Concentration
Duration
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
Rats
WKY
Rats
Spontaneously
hypertensive (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 an 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
100 pM CO 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,
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
downregulation 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
24 h or
pretreatment for
1-2 h followed by
24 h post-
treatment
250 ppm
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.
January 2010
E-8
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Kim et al. (2008,
1939611
Primary rat hepatocytes 10-60 min
Primary mouse
hepatocytes
Respiration-deficient
human HepSB cells
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.
Kinobe et al. (2006, Sheep
1884471 Gravid and nongravid
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, Mouse
1919871
4h
Diesel emissions: Diesel exhaust enhanced vasoconstriction in veins but not
350 pg/m3 arteries. It were suggested that this is through the uncoupling of
eNOS.
Korres et al. (2007,
1909081
Human
Transient evoked otoacoustic emissions response and amplitude
at 4,000 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.
Kreiser et al. (2004, Human
1939481
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 |
0765241 Male
35 mL/kg (ip) CO increased the compound action potential threshold at high
frequencies. This could be blocked by inhibition of the glutamate
receptor.
Loennechen et al.
(1999, 0115491
Rat
Sprague Dawley
Female
220-240g
1 wk 100ppm
1 wk 100 ppm and
1 wk 200 pm
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
10'4M
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.
January 2010
E-9
-------
Reference
Lopez et al. (2008,
0973431
Species / Model Exposure
Duration
Rat Pregnant rats
Sprague Dawley exposed to CO
GD5-GD20
(Group A) or
GD5-GD20 plus
PND5-PND20
(Group B);
Group C (control
sir sxposursj.
CO Concentration Findings
25 ppm 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 showed vacuolization of
afferent terminals at the base of the cochlea. At PND3, group A
showed decreased synapsin-1 staining of the efferent nerve
terminals. At PND20, groups A and B showed decreased
neurofilament-IR (staining) in type I spiral ganglia neurons and
afferent nerve fibers. At PND12 and PND20, group B showed in-
creased HO-1 and SOD-1-IR in blood vessels of the stria
vasularis; group A was similar to controls. From PND3-PND20,
there was increased iNOS and increased nitrotyrosine-IR in
blood vessels of the cochlea.
10-18 h/day
Lopez et al. (2003,
1939011
Rat
Sprague Dawley
PND6 to weaning 12 or 25 ppm
(PND19-PND20)
In the cochlea, atrophy or vacuolization 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
Lund et al. (2009,
1802571
Mouse
ApoE"'"
Male
High-fat diet
Mouse
ApoE"'"
Male
High-fat diet
6 h/day, 7days/wk, 8, 40, or60|jg/m3PM
7 wk whole-gasoline exhaust;
or filtered exhaust with
gases matching the 60
pg/m concentration. CO
concentrations were 9,
50, and 80 ppm,
corresponding to the 8,
40, and 60 pg/m3 PM
whole-exhaust exposures
6 h/day, 1 or Gasoline engine exhaust
7 days containing 60 pg/m3 PM
and 80 ppm CO
Both whole-gasoline and filtered-gasoline 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/m3. 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.
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, Human
1939711
Wfomen with pre-eclampsia produced term placenta with
significant decreases in HO-2 vs women with healthy
pregnancies.
Lyall et al. (2000, Human
1939021 (placentas from 8-to19-
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 Rat
(1984, 0113551 Long Evans
McGregor et al. (1998, Guinea pig
0853421
Continuous 150 ppm
exposure
to CO over
gestation
GD23-GD25 until 200 ppm
term
(approximately
68 days)
10 h/day
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.
Aberrant respiratory responses (to asphyxia and C02) of
offspring with prenatal CO exposure. The authors hypothesized
that this may be related to changes in the brainstem. COHb was
measuerd in maternal (8.53 + 0.6% vs 0.25 + 0.1%) and fetal
blood (13.0 + 0.4% vs 1.6 + 0.1%) from CO-treated vs controls.
January 2010
E-10
-------
Reference Species / Model Exposure
Duration
CO Concentration
Findings
Mclaughlin et al.
(2001, 1938231
Human placenta
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.
Mclaughlin et al.
(2000, 0158151
Human 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
10wk)
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
Mereu et al (2000,
1938381
Middendorffetal.
(2000, 0158421
Montagnani et al.
(1996, 0809021
Rat 10wk
Dark Agouti
Male and female
Model of right ventricle
hypertrophy secondary to
chronic hypoxia (HH,
10wk)
Half of the animals were
exercise trained to induce
LV hypertrophy
Rat GDO-GD20
Wistar continuous CO
exposure
Human
Adult males aged 65-75 yr
Testicular tissue from
orchiectomy
Rat GDO-GD20
Wistar
Male pups
50 ppm alone In untrained animals, combined treatment with HH+CO led to
or concomitant with HH 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 (AWTd, PWTd), 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.
150 ppm 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.
75 or 150 ppm CO caused an increase in tetrodotoxin-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.
January 2010
E-11
-------
Reference
Naik and Walker
(2003, 1938521
Species / Model
Rat
Sprague Dawley
Male
Exposure
Duration
CO Concentration
21 OpL of CO/1 00 ml
of physiological
saline solution
Findings
Endogenous CO-mediated vasorelaxation involved cGMP-
independent activation^of vascular smooth muscle large-conduc-
tance Ca '-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 during gestation increased
the incidence of pup mortality and malformations including
gastroschisis. Zinc supplementation to a protein-deficient diet in
CO-exposed mice decreased fetal mortality and malformation.
Newby et al. (2005, Human
1939661 placental cells
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. (1998, Guinea pig
1939581
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, Rat
1938411 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 trigger apoptosis
of pre-meiotic germ cells and modulate spermatogenesis under
CdCI2 dependent oxidative stress.
Patel et al. (2003, Rat 30 min Buffer saturated with The ventricular glutathione content, both reduced and oxidized,
0431551 Sprague Dawley 0.01 and 0.05% CO decreased by 76% and 84% 90 min post-exposure to 0.01 % and
Male 0.05% CO, respectively. Treatment with antioxidants partially
262 ± 30 g blocked the decreases in glutathione. Increased creatine kinase
activity was observed in heart perfusate during and after
Isolated hearts treatment.
Penney etal. (1983, Rat
0113851 (strain not reported)
Penney etal. (1982, Rat
0113871 COBS
Piantadosi (2002,
0374631
GD17-GD22 157,166 or 200 ppm In utero CO exposure induced decreased fetal body weight,
decreased placental weight, increased wet heart weight at birth,
and altered cardiac enzymes at birth.
GDO-GD32 350 ppm PND1-PND3, Postnatal CO exposure decreased body weight, to a greater
then 425 ppm PND4- extent in male pups. The heart to body weight ratio and left
PND7, then 500 ppm ventricle plus interventricular septum and right ventricle weight
PND8-PND32 increased after birth in CO exposed pups. This persistent
cardiomegaly was not explained by increasing in DMA or
hydroxyproline.
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.
January 2010
E-12
-------
Reference
Species / Model
Exposure
Duration
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
concluded that CO caused mitochondrial pore stress
independently of its hypoxic effects
Prigge and Hochrainer Rat
(1977, 0123261 Wistar, SPF
(Raub and Benignus,
2002, 0416161
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).
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. Human
(2002,0375131 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 150 ppm
exposure
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.
Schwetzetal. (1979, Mouse
0118551 CF-1
Rabbit
New Zealand
7 or 24-h/day 250 ppm
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 et al. (1992, Mouse GD8-GD18 65,125, or 250 ppm CO exposure concomitant with a low-protein diet exacerbated the
0137591 CD-1 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 offspring survival and an additive effect on
malformations.
Singh (2006,1905121 Mouse
CD-1
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,
0138921
Mouse
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.
January 2010
E-13
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Singh (2003, 0536241 Mouse
Albino CD-1
GD8-GD18
500 ppm
CO decreased the mean implants per litter and 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-GD18 65, 125, 250, or 500 ppm
All concentration of CO decreased fetal weight in 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
at 65 ppm CO.
Singh (1986, 0128271 Mouse
Albino CD-1
GD7-GD18
65 or 125 ppm
Impaired aerial righting score at PND14 (65 and 125 ppm),
impaired negative geotaxis at PND10 and righting reflex or
PND1 (125 ppm)
Sitdikova et al. (2007, Frog neuro-muscular 20 min 96 pM
1804171 junctions
CO-induced acetylcholine release, without effects on the pre-
synaptic action potential or functional properties of post-synaptic
receptors in frog neuro-muscular preparations.
Song et al. (2002, Human 0-48 h
0375311 Primary human airway
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, Rat
1804141 Wistar
Female
169 +4.5 g
20 h/day,
x 5 days/wk,
x72wk
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.
Stevens and Wang Mouse
(1993, 1884581 C57/BI-6J
HO inhibition blocked long-term potentiation but not long-term
depression.
Rat
Sprague Dawley
Hippocampal brain slices
Stockard-Sullivan et Rat 22 h/day, 12,25,50, or 100 ppm
al. (2003,1909471 Sprague Dawley 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
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.
Storm and Fechter Rat
(1985, 0116521 Long Evans
Storm etal. (1986, Rat
0121361 Long Evans
GDO-GD20 75,150, and 300 ppm CO transiently decreased 5HT and NE in the pons/medulla and
increased NE in the cortex and hippocampus at PND42. CO
dose-dependently reduced cerebellum wet weight. Maternal
COHb: 2.5%, 11.5%, 18.5%, and 26.8% (0, 75,150, and 300
ppm, respectively).
GDO-PND10 CO decreased cerebellar weight (150-300 ppm at PND10,
75-300 ppm at PND21) and decreased total cerebellar GABA
75,150, and 300 ppm (150-300 ppm at PND10 and PND21). CO- exposed (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 1,100 ppm-58%
January 2010
E-14
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Suliman et al. (2007, Mouse
1937681 C57BL/6
Wild-type and eNOS
deficient
Male
Rat
Embryonic
cardiomyocytes H9c2
cells
1 h
50-1,250 ppm
OrHH
OMOOmM
dichlorom ethane
One-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 two- to threefold 24 h after exposure. CO
activated gene expression of these proteins involved in cardiac
mitochondrial biogenesis beginning at 2 h postexposure 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 postexposure 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
Tattoli etal. (1999,
0115571
Telfer etal. (2001,
1937691
Teran et al. (2005,
1937701
Mouse
Neuronal cultures
prepared from the
cerebral hemispheres of
16-day Charles River CD1
mouse embryos
Rat PND1-PND10 75 and 150 ppm
Wistar
Male and pregnant female
Human
Myometrium tissue
obtained from gravid (pre-
term [25- to 34 wk
gestation], term not in
labor or term in labor) and
non-gravid women
Rat lOOpM
Dahl/Rapp
salt-sensitive rats
Male
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.
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.
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.
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.
January 2010
E-15
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Thorn etal. (1994,
0764591
Rat
Wistar
Male
Isolated blood cells
1 h
Or
>1 h
SOmin
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.
Thorn and Ra
lschiropoulos(1997, Wistar
0856441 Male
200-290 g
1 h
30 minor2 h
1h
20-1,000 ppm
10-20 ppm
10-100 ppm
Platelet-rich plasma from
rats was used as the
source of platelets
Bovine pulmonary artery
endothelial cells
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 howNOwas 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
release of short-lived oxidants. This response was blocked by
NOS inhibition. Lysates from cells exposed to 50 and 100 ppm
CO had increased nitrotyrosine content. This response was
blocked by NOS inhibition. Cellular reduced sulfhydryls were not
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 51chromium.
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 by ethidium 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.
January 2010
E-16
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Thorn etal. (1997,
0843371
Bovine pulmonary artery
endothelial cells
30 min-4 h 10-100 ppm (11-110 nM) One-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 blocked 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
was blocked by NOS inhibition. Cytotoxicity of CO was evaluated
by the release of chromium. Cytotoxicity was evident 4 h
following a 2-h incubation with 110 nM 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 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 pretreated
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. (1999,
0167531
Rat 1 h
Wistar
Male
200-290 g
Some rats were fed a high
cholesterol diet
50-1,000 ppm 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
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 six-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.
January 2010
E-17
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Thorn etal. (1999,
0167571
Rat
Wistar
Male
200-290 g
1 h
50-1,000 ppm 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 timepoints following CO exposure. This
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
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 lungs 18 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 40 min-2 h
endothelial cells
11-110 nM Increased uptake of ethidium homodimer-1, a measure of
(10-100 ppm) 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
110nMCO.
The authors concluded that exposure to 11 nM CO led to an
adaptive response which protected cells from injury and
apoptosis resulting from NO-derived oxidants.
Thorn etal. (2001, Rat Until lost 1,000-3,000 ppm Neutrophils sequestration was observed in the brain vessels of
1937791 consciousness rats exposed to high-dose CO. CO also led to increased
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.
January 2010
E-18
-------
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, Guinea pig 1 0h/day over the 200 ppm
0159971 last 60% of
gestation
Tolcos et al. (2000, Guinea pig 10h/dayforthe 200 ppm
0104681 last 60% of
gestation
Fetal nd 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
immunoreactivityofthe medulla and no changes in muscarinic
acetylcholine receptor.
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.
Toyadaetal. (1996,
079945)
Tschugguel et al. Human
(2001, 193785)
v ' HUVEC
Vallone et al. (2004, Mouse protein
1939931
CO was generated by primary endothelial cells from human
umbilical veins and uterine arteries after exogenous 17-13 estra-
diol administration.
The authors presented the X-ray structure of CO-bound ferrous
murine Nb. When CO binds, the heme group slides deeper into
the protein crevice.
Villamoretal. (2000,
015838)
Vreman et al. (2000, Human
096915) Umbilical cord (artery and
HO activity was quantified in human umbilical cord and in the rat
vasculature (aorta and vena cavae). Human umbilical artery and
vein)
Rat
Aorta, vena cavae, liver
and heart
vein HO activity were equal. The rat aorta and vena cavae
produced equal amounts of HO activity (wet weight/g tissue) but
generated 3 times greater HO than the heart and 0.2 times of the
liver. HO activity in rat vasculature was 3 times 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.
January 2010
E-19
-------
Reference
Species / Model
Exposure CO Concentration
Duration
Findings
Vreman et al. (2005, Mouse
1937861 BALB/c
30 min 500 ppm
OR
Heme arginate
30 pmol/kg body weight
i.v.
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%
Injection of heme arginate resulted in a threefold increase in CO
excretion, reaching a maximum at 60 min. Animals were sacri-
ficed at 90 min. COHb levels were 0.9%. Tissue concentrations
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.
January 2010
E-20
-------
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
35ppm
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
Zamudioetal. (1995, Human
1939081
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.
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.
January 2010
E-21
-------
Reference
Species / Model
Exposure CO Concentration
Duration
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
postLPS treatment. Pretreatment of mice with the spin trap agent
PBN or the TNF a 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.
Zuckerbraun et al.
(2007, 1938841
Macrophages 10 min-24 h
RAW 264.7
THP-1 cells, wild-type and
respiration-deficient
50-500 ppm 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
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.
January 2010
E-22
-------
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. Mol Med,
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, 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, 228: 59-63. 193876
Alonso JR; Cardellach F; 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
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: 568-76.
194960
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; Janjua MZ; 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. Free Radic Biol Med, 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).
January 2010 E-23
-------
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. FASEB J, 15:
1158-68. 193891
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. Dev Brain Res, 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. Nature, 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. Toxicol Sci, 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
January 2010 E-24
-------
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA (2004). Carbon monoxide actuates O2-limited heme
degradation 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 Flt-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. J Cell Sci, 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
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. Mol Hum Reprod, 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
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
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 pulmonary
hypertension 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. J Cell Mol Med, 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; 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; 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
January 2010 E-25
-------
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: HI046-52. 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. 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
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. J Perinatol, 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
inDahlrats. 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
January 2010 E-26
-------
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. PNAS, 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 p3 Sbeta mitogen-activated protein kinase mediates the antiproliferative effect of carbon monoxide. PNAS, 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
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 TL; Lund AK; Lucas SN; Campen MJ (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling
of eNOS. ToxicolAppl Pharmacol, 230: 346-351. 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. J Perinatol, 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. ToxicolAppl 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. Pharmacol Toxicol, 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. J Neurosci Res, 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; Shohet R; 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
January 2010 E-27
-------
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-1 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. FASEB J, 14: 208-19. 193902
Lyall F; Myatt L (2002). The role of the placenta in pre-eclampsia - a workshop report. Placenta, 23: SI 42-145. 193971
Mactutus CF; FechterLD (1984). Prenatal exposure to carbon monoxide: learning andmemory deficits. Science, 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, 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 Curr Iss,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 HE; 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. J Clin Invest, 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
January 2010 E-28
-------
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
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 (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. Teratology, 19: 385-391. 011855
Shaoqing Y; Ruxin Z; Yinjian C; Jianqiu C; Chunsheng Z; Jiangfeng T; Genhong L (2008). Possible contribution of
endogenous carbon monoxide to the development of allergic rhinitis in guinea pigs. J Inflamm, 5: 23. 192384
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
Sitdikova GF; Islamov RR; Mukhamedyarov MA; Permyakova VV; Zefirov AL; Palotas A (2007). Modulation of
neurotransmitter release by carbon monoxide at the frog neuro-muscular junction. CurrDrug 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
Sorhaug 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
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
January 2010 E-29
-------
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
Styka PE; 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). Anew activating role for CO in cardiac mitochondrial
biogenesis. J Cell Sci, 120: 299-308. 193768
Sun Y; Jin K; Mao XO; Zhu Y; Greenberg DA (2001). Neuroglobin is up-regulated by and protects neurons from hypoxic-
ischemic injury. PNAS, 98: 15,306-15,311.026022
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. PNAS, 97: 1305-1310. 011574
Thorn SR; Ischiropoulos H (1997). Mechanism of oxidative stress from low levels of carbon monoxide. Health Effects
Institute. Boston, MA. 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
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
January 2010 E-30
-------
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. PNAS, 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
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
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. Am J Reprod Immunol, 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 STATS 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
Zhuo M; Small SA; Kandel ER; Hawkins RD (1993). Nitric oxide and carbon monoxide produce activity-dependent long-
term synaptic enhancement in hippocampus. Science, 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. FASEB J, 21: 1099-
1106. 193884
January 2010 E-31
------- |