United States
Wmm¦¦MEl Environmental Protection
KmI B * Agency
March 2009
EPA/600/R-09/019
Integrated Science Assessment for
Carbon Monoxide -
First External Review Draft
National Center for Environmental Assessment-RTP Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
March 2009
i
DRAFT-DO NOT QUOTE OR CITE
-------
March 2009 ii DRAFT-DO NOT QUOTE OR CITE
-------
Disclaimer
This document is the first external review draft for review purposes only and does not constitute
U.S. Environmental Protection Agency policy. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
March 2009
DRAFT-DO NOT QUOTE OR CITE
-------
March 2009 iv DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
Table of Contents v
List of Figures xi
Acronyms and Abbreviations xxii
Chapter 1. Introduction 1-1
1.1. Legislative Requirements 1-2
1.2. History of the NAAQS for CO 1-4
1.3. Document Development 1-5
1.4. Document Organization 1-5
1.5. Document Scope 1-6
1.6. EPA Framework for Causal Determination 1-7
1.6.1. Scientific Evidence Used in Establishing Causality 1-8
1.6.2. Association and Causation 1-8
1.6.3. Evaluating Evidence for Inferring Causation 1-9
1.6.4. Application of Framework for Causal Determination 1-13
1.6.5. Determination of Causality 1-14
1.6.6. Concepts in Evaluating Adversity of Health Effects 1-18
1.7. Summary 1-18
Chapter 2. Integrative Health Effects Overview 2-1
2.1. Ambient Concentrations, Sources, and Exposure 2-2
2.2. Dosimetry, Pharmacokinetics, and Mode of Action 2-4
2.2.1. Dosimetry and Pharmacokinetics 2-4
2.2.2. Mode of Action 2-4
2.3. Health Effects 2-6
2.3.1. Cardiovascular Morbidity 2-6
2.3.2. Central Nervous System Effects 2-7
2.3.3. Birth Outcomes and Developmental Effects 2-9
2.3.4. Respiratory Morbidity 2-10
2.3.5. Mortality 2-11
2.4. Public Health Impacts 2-13
2.4.1. Concentration-Response Relationship 2-13
2.4.2. Potentially Susceptible and Vulnerable Subpopulations 2-14
Chapter 3. Source to Exposure 3-1
3.1. Introduction 3-1
3.2. Sources and Emissions of CO 3-1
3.3. Physics and Chemistry of Atmospheric CO 3-8
3.4. Ambient Measurements 3-9
3.4.1. Ambient Measurement Instruments 3-10
3.4.2. Ambient Sampling Network Design 3-12
3.4.2.1. Monitor Siting Requirements 3-12
3.4.2.2. Spatial and Temporal Coverage 3-12
3.5. Environmental Concentrations 3-20
March 2009
v
DRAFT-DO NOT QUOTE OR CITE
-------
3.5.1. Spatial Variability 3-20
3.5.1.1. National Scale 3-20
3.5.1.2. Urban Scale 3-28
3.5.1.3. Neighborhood Scale 3-39
3.5.2. Temporal Variability 3-44
3.5.2.1. Multi-year Trends 3-44
3.5.2.2. Hourly Variation 3-46
3.5.3. Associations with Copollutants 3-50
3.5.4. Policy-Relevant Background 3-52
3.6. Issues in Exposure Assessment 3-54
3.6.1. Summary of Findings from 2000 CO AQCD 3-54
3.6.2. General Exposure Concepts 3-55
3.6.3. Monitoring Issues Associated with Exposure Assessment 3-58
3.6.3.1. Exposure Assessment using Community-Based Ambient Monitors 3-58
3.6.3.2. Personal Exposure Monitors 3-58
3.6.4. Indoor/Outdoor Relationships and Infiltration 3-59
3.6.5. Personal/Ambient Relationships 3-60
3.6.5.1. Panel and Population Exposure Studies 3-60
3.6.5.2. Commuting Time CO Exposure Studies 3-62
3.6.5.3. CO Exposure Assessment Variability and Error 3-65
3.6.6. Multi-pollutant Exposures 3-67
3.6.7. Exposure Modeling 3-68
3.7. Summary and Conclusions 3-71
3.7.1. Sources of CO 3-71
3.7.2. Physics and Chemistry of Atmospheric CO 3-71
3.7.3. Ambient CO Measurements 3-72
3.7.4. Environmental CO Concentrations 3-72
3.7.5. Exposure Assessment and Implications for Epidemiology 3-73
Chapter 4. Dosimetry and Pharmacokinetics of Carbon Monoxide 4-1
4.1. Introduction 4-1
4.2. Carboxyhemoglobin Formation 4-2
4.2.1. The Coburn-Forster-Kane and Other Models 4-2
4.2.2. Multicompartment Model 4-5
4.2.3. Mathematical Model Usage 4-6
4.3. Absorption, Distribution, and Elimination 4-7
4.3.1. Pulmonary Absorption 4-7
4.3.1.1. Mass Transfer of Carbon Monoxide 4-8
4.3.1.2. Lung Diffusion of Carbon Monoxide 4-9
4.3.2. Tissue Uptake 4-10
4.3.2.1. The Respiratory Tract 4-10
4.3.2.2. The Blood 4-10
4.3.2.3. Heart and Skeletal Muscle 4-12
4.3.2.4. Other Tissues 4-13
4.3.3. Pulmonary and Tissue Elimination 4-13
4.4. Conditions Affecting Uptake and Elimination 4-15
4.4.1. Environment and Activity 4-15
4.4.2. Altitude 4-16
4.4.3. Physical Characteristics 4-17
4.4.4. Health Status 4-18
4.5. Endogenous CO Production and Metabolism 4-18
4.6. Summary and Conclusions 4-20
Chapter 5. Integrated Health Effects 5-1
5.1. Mode of Action ofCOToxicity 5-1
5.1.1. Introduction 5-1
5.1.2. Hypoxic Mechanisms 5-1
March 2009
vi
DRAFT-DO NOT QUOTE OR CITE
-------
5.1.3. Non-Hypoxic Mechanisms 5-2
5.1.3.1. Non-Hypoxic Mechanisms Reviewed in the 2000 CO AQCD 5-2
5.1.3.2. Recent Studies of Non-Hypoxic Mechanisms 5-4
5.1.3.3. Implications of Non-Hypoxic Mechamisms 5-11
5.2. Cardiovascular Effects 5-13
5.2.1. Epidemiologic Studies 5-13
5.2.1.1. Epidemiologic Studies with Short-Term Exposure 5-14
5.2.1.2. Epidemiologic Studies with Long-Term Exposure 5-44
5.2.1.3. Summary of Epidemiologic Studies of Exposure to CO and Cardiovascular Effects 5-44
5.2.2. Controlled Human Exposure Studies 5-45
5.2.3. Toxicological Studies 5-48
5.2.3.1. Endothelial Dysfunction 5-48
5.2.3.2. Cardiac Remodeling Effects 5-50
5.2.3.3. Electrocardiographic Effects 5-52
5.2.3.4. Summary of Cardiovascular Toxicology 5-52
5.2.4. Summary of Cardiovascular Effects 5-53
5.3. Central Nervous System Effects 5-54
5.3.1. Controlled Human Exposure Studies 5-54
5.3.2. Toxicological Studies 5-55
5.3.3. Summary of Central Nervous System Effects 5-55
5.4. Birth Outcomes and Developmental Effects 5-56
5.4.1. Epidemiologic Studies 5-56
5.4.1.1. Preterm Birth 5-57
5.4.1.2. Birth Weight, Low Birth Weight, and Intrauterine Growth Restriction/Small for Gestational Age 5-61
5.4.1.3. Congenital Anomalies 5-71
5.4.1.4. Neonatal and Post-Neonatal Mortality 5-73
5.4.1.5. Summary of Epidemiologic Studies of Birth Outcomes and Developmental Effects 5-75
5.4.2. Toxicological Studies of Birth Outcomes and Developmental Effects 5-75
5.4.2.1. Birth Outcomes 5-75
5.4.2.2. Developmental Effects 5-83
5.4.3. Summary of Birth Outcomes and Developmental Effects 5-94
5.5. Respiratory Effects 5-95
5.5.1. Epidemiologic Studies with Short-Term Exposure 5-95
5.5.1.1. Pulmonary Function, Respiratory Symptoms, and Medication Use 5-95
5.5.1.2. Respiratory Hospital Admissions, ED Visits and Physician Visits 5-103
5.5.2. Epidemiologic Studies with Long-Term Exposure 5-113
5.5.2.1. Pulmonary Function 5-113
5.5.2.2. Asthma and Asthma Symptoms 5-114
5.5.2.3. Allergic Rhinitis 5-115
5.5.2.4. Summary of Associations between Long-Term Exposure to CO and Respiratory Morbidity 5-115
5.5.3. Controlled Human Exposure Studies 5-115
5.5.4. Toxicological Studies 5-116
5.5.5. Summary of Respiratory Health Effects 5-117
5.5.5.1. Short-Term Exposure to CO 5-117
5.5.5.2. Long-Term Exposure to CO 5-118
5.6. Mortality 5-119
5.6.1. Epidemiologic Studies with Short-Term Exposure to CO 5-119
5.6.1.1. Summary of Findings from 2000 CO AQCD 5-119
5.6.1.2. Multicity Studies 5-120
5.6.1.3. Meta-Analysis of All Criteria Pollutants 5-127
5.6.1.4. Single-City Studies 5-128
5.6.1.5. Summary of Mortality and Short-Term Exposure to CO 5-131
5.6.2. Epidemiologic Studies with Long-Term Exposure to CO 5-132
5.6.2.1. U.S. Cohort Studies 5-133
5.6.2.2. U.S. Cross-Sectional Analysis 5-137
5.6.2.3. Summary of Mortality and Long-Term Exposure to CO 5-138
5.7. Public Health Impacts 5-138
5.7.1. Concentration-Response Relationship 5-139
5.7.2. Potentially Susceptible and Vulnerable Populations 5-140
March 2009
vii
DRAFT-DO NOT QUOTE OR CITE
-------
5.7.2.1. Susceptibility Characteristics 5-141
5.7.2.2. Vulnerability Characteristics 5-144
Annex A. Atmospheric Science A-1
Annex B. Dosimetry Studies B-51
Annex C. Epidemiology Studies C-1
Annex D. Controlled Human Exposure Studies D-1
Annex E. Toxicological Studies E-1
References 1
March 2009
viii
DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Table 1-1. Aspects to aid in judging causality. 1-13
Table 1-2. Weight of evidence for causal determination. 1-15
Table 2-1. Causal determinations for health effects outcomes. 2-6
Table 3-1. Proximity to CO monitors for the total population by city. 3-19
Table 3-2. Proximity to CO monitors for adults aged 65 and older by city. 3-19
Table 3-3. Distribution of 1-h avg CO concentration (ppm) derived from AOS data. 3-23
Table 3-4. Distribution of 24-h avg CO concentration (ppm) derived from AOS data. 3-25
Table 3-5. Distribution of 1-h daily max CO concentration (ppm) derived from AOS data. 3-26
Table 3-6. Distribution of 8-h daily max CO concentration (ppm) derived from AOS data. 3-27
Table 3-7. Table of inter-sampler comparison statistics, as defined in the text, including Pearson r, P90 (ppm), COD and
d (km) for each pair of hourly CO monitors reporting to AOS for 2005-2007 in Phoenix, AZ. 3-31
Table 3-8. Table of inter-sampler comparison statistics, as defined in the text, including Pearson r, P90 (ppm), COD and
d (km) for each pair of hourly CO monitors reporting to AOS for 2005-2007 in Pittsburgh, PA. 3-34
Table 3-9. Percentage of time exposed to ambient CO (adjusted to reflect the absence of non-ambient CO from ETS and gas
cooking), average CO exposures, and percentage of exposure estimated for the population. 3-61
Table4-1a. CO concentration in pmol/100gwwtissue-human. 4-8
Table 4-1 b. CO concentration in pmol/mg fresh weight-adult mouse. 4-9
Table 5-1. Responses to low and moderate CO exposures. 5-9
Table 5-2. Tissue concentration of CO following inhalation exposure. 5-12
Table 5-3. Tissue concentration of CO following increased endogenous production. 5-12
Table 5-4. Summary of studies investigating the effect of CO exposure on HRV parameters. 5-17
Table 5.5. Summary of studies investigating the effect of CO exposure on cardiac arrhythmias. 5-20
Table 5-6. Summary of studies investigating the effect of CO exposure on blood markers of coagulation and inflammation. 5-24
Table 5-7. Summary of IHD hospital admission studies.1 5-29
Table 5-8. Summary of stroke hospital admission studies.1 5-34
Table 5-9. Summary of HF hospital admission studies. 5-37
Table 5-10. Association of ambient air pollution levels and cardiovascular morbidity in visits with and without specific
secondary conditions. 5-38
Table 5-11. Summary of non-specific CVD hospital admission studies. 5-41
Table 5-12. Brief summary of PTB studies. 5-61
Table 5-13. Brief summary of birth weight studies. 5-70
Table 5-14. Range of CO concentrations reported in key respiratory morbidity studies that examined effects associated with
short-term exposure to CO. 5-96
March 2009
ix
DRAFT-DO NOT QUOTE OR CITE
-------
Table 5-15. Range of CO concentrations reported in key respiratory hospital admission and ED visit studies that examine
effects associated with short-term exposure to CO. 5-104
Table 5-16. Range of CO concentrations reported in multi-city studies that examine mortality effects associated with short-term
exposure to CO. 5-120
Table 5-17. Range of CO concentrations reported in single-city studies that examine mortality effects associated with short-
term exposure to CO. 5-128
Table 5-18. Characteristics of susceptible subpopulations. 5-140
Table 5-19. Characteristics of vulnerable subpopulations. 5-141
Table A-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 AOS in Anchorage, AK. A-17
Table A-2. 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 AOS in Atlanta, GA. A-19
Table A-3. 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 AOS in Boston, MA. A-22
Table A-4. 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 AOS in Denver, CO. A-25
Table A-5. 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 AOS in Houston, TX. A-28
Table A-6. 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 AOS in Los Angeles, CA. A-31
Table A-7. 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 AOS in New York City, NY. A-37
Table A-8. 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 AOS in St. Louis, MO. A-41
Table B-1. Recent studies related to CO dosimetry and pharmacokinetics. B-51
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-8
Table C-3. Studies of CO exposure and neonatal and postneonatal outcomes. C-15
Table C-4. Studies of short-term CO exposure and respiratory morbidity C-22
Table C-5. Studies of short-term CO exposure and respiratory hospital admissions and ED visits. C-28
Table C-6. Studies of long-term CO exposure and respiratory morbidity. C-48
Table C-7. Studies of short-term CO exposure and mortality. C-51
Table C-8. Studies of long-term CO exposure and mortality. C-72
Table D-1. Controlled human exposure studies. D-1
Table E-1. Human and animal studies. E-1
March 2009
x
DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Figure 3-1. CO emissions (MT) in the U.S. by source sector in 2002. 3-2
Figure 3-2. Trends in anthropogenic CO emissions (MT) in the U.S. by source category for 1990 and 1996-2002. 3-3
Figure 3-3. CO concentrations measured by satellite at the 700 hectoPascal level (-10,000 feet above sea level) from
MOPITT for the period 15-23 July 2004 during intense wildfires in Alaska and Yukon. 3-6
Figure 3-4. Trends in sub-national CO emissions in the 10 U.S. EPA Regions for 1990 and 1996 to 2002. 3-6
Figure 3-5. CO emissions density map and distribution for the state of Colorado, and for selected counties in Colorado. 3-7
Figure 3-6. Map of-385 CO monitor locations in the U.S. in 2007. 3-14
Figure 3-7. Map of CO monitor locations with respect to population density in the Phoenix, AZCBSA, total population. 3-15
Figure 3-8. Map of CO monitor locations with respect to population density in the Phoenix, AZ CBSA, age 65 and older. 3-16
Figure 3-9. Map of CO monitor locations with respect to population density in the Pittsburgh, PACSA, total population. 3-17
Figure 3-10. Map of CO monitor locations with respect to population density in the Pittsburgh, PACSA, age 65 and older. 3-18
Figure 3-11. County-level map of second-highest 1-h avg CO concentrations in the U.S. in 2007. 3-21
Figure 3-12. County-level map of second-highest 8-h avg CO concentrations in the U.S. in 2007. 3-22
Figure 3-13. Map of CO monitor locations with AOS Site IDs for Phoenix, AZ. 3-29
Figure 3-14. Box plots illustrating the seasonal distribution of 2005-2007 hourly CO concentrations in Phoenix, AZ. Note: 1 =
winter, 2 = spring, 3 = summer, and 4 = fall on the x-axis. 3-30
Figure 3-15. Map of CO monitor locations with AOS Site IDs for Pittsburgh, PA. 3-32
Figure 3-16. Box plots illustrating the seasonal distribution of 2005-2007 hourly CO concentrations in Pittsburgh, PA. 3-33
Figure 3-17. Aerial view of the location of CO monitor C. 3-36
Figure 3-18. Aerial view of the location of CO monitor E. 3-37
Figure 3-19. Aerial view of the location of CO monitors A, B, and C. 3-38
Figure 3-20. Aerial view of the location of CO monitor F. 3-39
Figure 3-21. Relative concentrations of CO and copollutants at various distances from the 710 freeway in Los Angeles. 3-41
Figure 3-22. Dimensionless tracer gas concentration as a function of elevation at windward and leeward locations and street
canyon aspect ratios (H/W). 3-42
Figure 3-23. Inter-sampler correlations as a function of distance between CO monitors for samplers located within 4 km
(neighborhood scale) for Boston, Denver, Phoenix, and Pittsburgh. 3-43
Figure 3-24. (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). 3-44
Figure 3-25. 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 3-46
Figure 3-26. Diel plot generated from weekday hourly CO data (ppm) for the eleven CSAs and CBSAs 2005-2007. 3-48
March 2009 xi DRAFT-DO NOT QUOTE OR CITE
-------
Figure 3-27. Diel plot generated from weekend hourly CO data (ppm) for the eleven CSAs and CBSAs 2005-2007. 3-49
Figure 3-28. Seasonal plots of nationwide correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3,
(4) PM10, and (5) PM2.5 concentrations. 3-50
Figure 3-29. Map of the baseline monitor sites used in this assessment to compute policy-relevant background concentrations. 3-53
Figures 3-30. Monthly (circles) and annual (squares) average CO concentrations (ppb), 2005-2007 . 3-54
Figure 3-31. Distribution of time that a sample population spends in various environments, from the National Human Activity
Pattern Survey. 3-58
Figure 3-32. Comparison of in-vehicle (solid line) and engine (dotted line) results for (left) driving with windows closed and air
conditioner in recirculating air mode, and (right) driving with windows closed and air conditioner in fresh air mode. 3-63
Figure 3-33. Hourly personal vs. ambient CO concentrations obtained in Baltimore, MD. 3-66
Figure 3-34. Box plots of the ratio of personal to ambient concentrations obtained in Baltimore, MD. 3-66
Figure 4-1. Plot of fractional sensitivities of selected variables versus time of exposure. 4-4
Figure 4-2. Overall structure of the multicompartment model of storage and transport of CO. 4-6
Figure 4-3. Predicted COHb increments over endogenous levels resulting from 1,8,12, and 24 h CO exposures in a modeled
healthy human at rest. 4-7
Figure 4-4. Diagrammatic presentation of CO uptake and elimination pathways and CO body stores. 4-9
Figure 4-5. 02Hb dissociation curve of normal human blood, of blood containing 50% COHb, and of blood with only 50% Hb
because of anemia. 4-12
Figure 4-6. Changes in blood COHb after short-term and long-term exposure to CO, representing the biphasic nature of CO
elimination. 4-15
Figure 5-1. Summary of effect estimates (95% confidence intervals) associated with hospital admissions for various froms of
IHD. 5-28
Figure 5-2. Summary of effect estimates (95% confidence intervals) associated with ED visits and hospital admissions for
stroke. 5-33
Figure 5.3. Summary of effect estimates (95% confidence intervals) associated with hospital admissions for CHF. 5-36
Figure 5-4. Summary of effect estimates (95% confidence intervals) associated with hospital admissions for CVD. 5-41
Figure 5-5. Effect estimates from studies of ED visits and hospital admissions for CVD outcomes other than stroke from single
pollutant (CO only, closed circles) and copollutant (CO plus PM, NO2, SO2 and O3, open circles) models. 5-43
Figure 5-6. Summary of effect estimates (95% confidence intervals) for PTB associated with maternal exposure to ambient
CO. 5-60
Figure 5-7. Summary of change in birth weight (95% confidence intervals) associated with maternal exposure to ambient CO. 5-67
Figure 5-8. Summary of effect estimates (95% confidence intervals) for LBW associated with maternal exposure to ambient
CO. 5-68
Figure 5-9. Summary of effect estimates (95% confidence intervals) for SGA associated with maternal exposure to ambient
CO. 5-69
Figure 5-10. Estimates for FEV1 change expressed per SD change of the individual pollutants 5-98
Figure 5-11. Asthma symptoms, respiratory symptoms and medication use in asthmatic individuals associated with short-term
exposure to CO. 5-102
Figure 5-12. Summary of associations between short-term exposure to CO and respiratory hospital admissions.' 5-109
Figure 5-13. Summary of associations between short-term exposure to CO and respiratory ED visits. 5-112
March 2009 xii DRAFT-DO NOT QUOTE OR CITE
-------
Figure 5-14. Posterior means and 95%posterior intervals of national average estimates for CO effects on total (non-accidental)
mortality at lagsO, 1, and 2 within sets of the 90 U.S. cities with available pollutant data. 5-122
Figure 5-15. Summary of mortality risk estimates for short-term exposure to CO from multi-city studies. 5-127
Figure 5-16. Summary of mortality risk estimates for long-term exposure to CO. 5-136
Figure A-1. CO emissions density map and distribution for the state of Alaska and for Yukon-Koyukuk County in Alaska. A-1
Figure A-2. CO emissions density map and distribution for the state of Utah and for selected counties in Utah. A-2
Figure A-3. CO emissions density map and distribution for the state of Massachusetts and for selected counties in
Massachusetts. A-3
Figure A-4a. CO emissions density map and distribution for the state of Georgia and for selected counties in Georgia (1 of 2) A-4
Figure A-4b. CO emissions distribution for selected counties in Georgia (2 of 2) A-5
Figure A-5. CO emissions density map and distribution for the state of California and for selected counties in California. A-6
Figure A-6. CO emissions density map and distribution for the state of Alabama and for Jefferson County in Alabama. A-7
Figure A-7. Map of CO monitor locations with respect to population density in the Anchorage CBSA, total population. A-8
Figure A-8. Map of CO monitor locations with respect to population density in the Anchorage CBSA, ages 65 and older. A-8
Figure A-9. Map of CO monitor locations with respect to population density in the Atlanta CSA, total population. A-9
Figure A-10. Map of CO monitor locations with respect to population density in the Atlanta CSA, ages 65 and older. A-9
Figure A-11. Map of CO monitor locations with respect to population density in the Boston CSA, total population. A-10
Figure A-12. Map of CO monitor locations with respect to population density in the Boston CSA, ages 65 and older. A-10
Figure A-13. Map of CO monitor locations with respect to population density in the Denver CSA, total population. A-11
Figure A-14. Map of CO monitor locations with respect to population density in the Denver CSA, ages 65 and older. A-11
Figure A-15. Map of CO monitor locations with respect to population density in the Houston CSA, total population. A-12
Figure A-16. Map of CO monitor locations with respect to population density in the Houston CSA, ages 65 and older. A-12
Figure A-17. Map of CO monitor locations with respect to population density in the Los Angeles CSA, total population. A-13
Figure A-18. Map of CO monitor locations with respect to population density in the Los Angeles CSA, ages 65 and older. A-13
Figure A-19. Map of CO monitor locations with respect to population density in the New York City CSA, total population. A-14
Figure A-20. Map of CO monitor locations with respect to population density in the New York City CSA, ages 65 and older. A-14
Figure A-21. Map of CO monitor locations with respect to population density in the Seattle CSA, total population. A-15
Figure A-22. Map of CO monitor locations with respect to population density in the Seattle CSA, ages 65 and older. A-15
Figure A-23. Map of CO monitor locations with respect to population density in the St. Louis CSA, total population. A-16
Figure A-24. Map of CO monitor locations with respect to population density in the St. Louis CSA, ages 65 and older. A-16
Figure A-25. Map of CO monitor locations with AOS Site IDs for Anchorage, AK. A-17
Figure A-26. 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. A-18
Figure A-27. Map of CO monitor locations with AOS Site IDs for Atlanta, GA. A-19
March 2009
xiii
DRAFT-DO NOT QUOTE OR CITE
-------
Figure A-28. 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. A-20
Figure A-29. Map of CO monitor locations with AOS Site IDs for Boston, MA. A-21
Figure A-30. 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. A-23
Figure A-31. Map of CO monitor locations with AOS Site IDs for Denver, CO. A-24
Figure A-32. Box plots illustrating the seasonal distribution of hourly CO concentrations in Denver, CO. Note: 1 = winter, 2, =
spring, 3 = summer, and 4 = fall on the x-axis. A-26
Figure A-33. Map of CO monitor locations with AOS Site IDs for Houston, TX. A-27
Figure A-34. 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. A-29
Figure A-35. Map of CO monitor locations with AOS Site IDs for Los Angeles, CA. A-30
Figure A-36a. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los Angeles, CA. Note: 1 = winter, 2,
= spring, 3 = summer, and 4 = fall on the x-axis. A-33
Figure A-36b. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los Angeles, CA. Note: 1 = winter, 2,
= spring, 3 = summer, and 4 = fall on the x-axis. A-34
Figure A-36c. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los Angeles, CA. Note: 1 = winter, 2,
= spring, 3 = summer, and 4 = fall on the x-axis. A-35
Figure A-37. Map of CO monitor locations with AOS Site IDs for New York City, NY. A-36
Figure A-38. 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. A-38
Figure A-39. Map of CO monitor locations with AOS Site IDs for Seattle, WA. A-39
Figure A-40. Box plots illustrating the seasonal distribution of hourly CO concentrations in Seattle, WA. Note: 1 = winter, 2, =
spring, 3 = summer, and 4 = fall on the x-axis. A-40
Figure A-41. Map of CO monitor locaitons with AOS Site IDs for St. Louis, MO. A-41
Figure A-42. Box plots illustrating the seasonal distribution of hourly CO concentrations in St. Louis, MO. Note: 1 = winter, 2, =
spring, 3 = summer, and 4 = fall on the x-axis. A-42
Figure A-43. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Anchorage, AK. Also shown are correlations between 24-h average 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. A-43
Figure A-44. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Atlanta, GA. Also shown are correlations between 24-h average 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.) A-44
Figure A-45. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Boston, MA. Also shown are correlations between 24-h average 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.) Red bars denote
the median, and green stars denote the arithmetic mean. A-45
March 2009
xiv
DRAFT-DO NOT QUOTE OR CITE
-------
Figure A-46. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Denver, CO. Also shown are correlations between 24-h average 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.) Red bars
denote the median, and green stars denote the arithmetic mean. A-46
Figure A-47. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Los Angeles, CA. Also shown are correlations between 24-h average 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.) A-47
Figure A-48. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for New York City, NY. Also shown are correlations between 24-h average 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.) A-48
Figure A-49. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Phoenix, AZ. Also shown are correlations between 24-h average 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.) A-49
Figure A-50. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and
(5) PM2.5 concentrations for Seattle, WA. Also shown are correlations between 24-h average 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.) A-50
March 2009
xv
DRAFT-DO NOT QUOTE OR CITE
-------
Authors and Contributors
Authors
Dr. Thomas Long (CO Team Leader)—National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, 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, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens— National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Matthew Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
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
March 2009
xvi
DRAFT-DO NOT QUOTE OR CITE
-------
Contributors
Dr. Richard Baldauf—National Risk Management Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Vernon Benignus—National Health and Environmental Effects Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Lance McCluney—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Kris Novak—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Adam Reff—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Mark Schmidt—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Rhonda Thompson—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Reviewers
Dr. Richard Baldauf—National Risk Management Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Vernon Benignus—National Health and Environmental Effects Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Souad Benromdhane—Office of Air Quality Planning and Standards, Office of Air and Radiation,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Philip Bromberg—School of Medicine, University of North Carolina, Chapel Hill, NC
Dr. Matthew Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Daniel Costa—National Program Director for Air, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Andrew Ghio—National Health and Environmental Effects Laboratory, 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
March 2009
xvii
DRAFT-DO NOT QUOTE OR CITE
-------
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, 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 Rajan—Office of Air Quality Planning and Standards, Office of Air and Radiation, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Harvey Richmond—Office of Air Quality Planning and Standards, Office of Air and Radiation, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph Somers—Office of Transportation and Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Ann Arbor, MI
Dr. John Vandenberg—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Alan Vette—National Exposure Research Laboratory, 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, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lin Weaver—Department of Internal Medicine, LDS Hospital, Salt Lake City, UT
Dr. Lewis Weinstock—Office of Air Quality Planning and Standards, Office of Air and Radiation, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Ron Williams—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
CO Project Team
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
March 2009
xviii
DRAFT-DO NOT QUOTE OR CITE
-------
Scientific Staff
Dr. Thomas Long (CO Team Leader)—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, 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, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens— National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Technical Support Staff
Ms. Ellen Lorang— National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Connie Meacham—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Deborah Wales—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
March 2009
xix
DRAFT-DO NOT QUOTE OR CITE
-------
Clean Air Scientific Advisory Committee
CO NAAQS Review Panel
Chairperson
Dr. Joseph Brain*, Department of Environmnetal Health, Harvard School of Public Health, Harvard
University, Boston, MA
Members
Dr. James Crapo*, Department of Medicine, National Jewish Medical and Research Center, Denver, CO
Dr. Douglas Crawford-Brown, Department of Environmental Sciences and Engineering, University of
North Carolina, Chapel Hill, NC
Dr. Thomas Dahms, Department of Anesthesiology Research and Critical Care, St. Louis University
School of Medicine, St. Louis, MO
Dr. Russell R. Dickerson, Department of Meteorology, University of Maryland, College Park, MD
Dr. Laurence Fechter, Research Service, Department of Veterans Affairs, Loma Linda VA Medical
Center, Loma Linda, CA
Dr. H. Christopher Frey*, College of Engineering, Department of Civil, Construction, and Environmental
Engineering, North Carolina State University, Raleigh, NC
Dr. Milan Hazucha, Department of Medicine, Center for Environmental Medicine, Asthma and Lung
Biology, University of North Carolina, Chapel Hill, NC
Dr. Michael T. Kleinman, Department of Community & Environmental Medicine, University of
California-Irvine, Irvine, CA
Dr. Arthur Penn, Department of Comparative Biomedical Sciences, Louisiana State University School of
Veterinary Medicine, Baton Rouge, LA
Dr. Beate Ritz, School of Public Health, Epidemiology, University of California at Los Angeles, Los
Angeles, CA
Dr. Paul Roberts, Sonoma Technology, Inc., Petaluma, CA
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering, Georgia Institute of
Technology, Atlanta, GA
Dr. Stephen R. Thorn, Institute for Environmental Medicine, University of Pennsylvania, Philadelphia,
PA
* Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by the EPA
Administrator
March 2009
xx
DRAFT-DO NOT QUOTE OR CITE
-------
Science Advisory Board Staff
Dr. Ellen Rubin, Designated Federal Officer
1200 Pennsylvania Avenue, N.W.
Washington, DC, 20460
Phone: 202-343-9975
Fax: 202-233-0643
Email: rubin.ellen@epa.gov
Physical/Courier/FedEx Address:
Dr. Ellen Rubin, EPA Science Advisory Board Staff Office
Mail Code 1400F
Woodies Building, Room 3610E
1025 F Street, N.W.
Washington, DC 20004
March 2009
xxi
DRAFT-DO NOT QUOTE OR CITE
-------
Acronyms and Abbreviations
a alpha, ambient exposure factor (varies between 0 and 1)
a air exchange rate of the microenvironment
AA abdominal aorta(s)
ABR auditory brainstem response
ACS American Cancer Society
ACS-CPS-II ACS Cancer Prevention Study II
ADP adenosine diphosphate
AEFV area under the expiratory flow-volume curve
AGL above ground level
Akt Akt cell signaling pathway - protein family (of protein kinases B [PKB[)
AM, a.m. morning hours
AMI acute myocardial infarction
AMP adenosine monophosphate
ANOVA analysis of variance
APO E apolipoprotein E
APO E"'" mouse strain used as a model of atherosclerosis
ARI acute respiratory infection
AP action potential
APD action potential duration
APEX Air Pollution Exposure (population-based model)
APHEA Air Pollution and Health: A European Approach (multi-cities analyses)
APHEA2 extended analysis of APHEA
APTT activated partial thromboplastin time, (blood coagulability endpoint)
AQ air quality
AQCD Air Quality Criteria Document
AQS (EPA) Air Quality System (database)
AR gastronomy reared with artificial feeding system (milk substitutes using gastronomy-
feeding)
ARCO gastronomy reared + CO exposure
ARIC Atherosclerosis Risk in Communities (study)
ARID gastronomy reared with iron deficient diet
ARIDCO gastronomy reared with iron deficient diet + CO exposure
ATP adenosine triphosphate
ATS American Thoracic Society
avg average
AVP aortic valve prosthesis (surgically implantable biological prosthesis)
March 2009
xxii
DRAFT-DO NOT QUOTE OR CITE
-------
p
B lymphocytes
BALF
BC
BEAS-2B
BEIS
BELD
BHR
BKCa
BP
BQ-123
BS
BSP
C
Ca
CA
Ca2+
CAA
CAD
CALINE
CALPUFF
CAMP
cAMP
CAP(s)
CASAC
CASN
CAth
CBS A
CCGG
CD
Cd
CD-I
CDC
CdCl2
CFK
CFR
cGMP
ch2o
ch2o2
beta (beta coefficient, slope)
bursa-dependent lymphocytes
bronchoalveolar lavage fluid
black carbon
human bronchial epithelial cell line
Biogenic Emissions Inventory System
Biogenic Emissions Landcover Database
bronchial hyper-responsiveness
voltage and Ca2+-activated K+ channel(s)
blood pressure
a selective endothelin A (ETA) receptor antagonist
black smoke
black smoke particles
carbon
ambient concentration
cardiac arrhythmia
calcium ion
Clean Air Act
coronary artery disease
California Line Source Dispersion Model
Gaussian puff modeling system to simulate air quality dispersion and accesses long range
transport of pollutants. Distributed by TRC Solutions.
Childhood Asthma Management Program (study)
cyclic AMP
compound action potential(s)
Clean Air Scientific Advisory Committee
Cooperative Air Sampling Network
cardiac atherosclerosis
Core-Based Statistical Area (containing at least one urban area of 10,000 people [as
determined by the 2000 census]; replaces the older MSA definition from the 1990
census).
Carbon Cycle Greenhouse Gases Group (within ESRL)
cardiac dysrhthmias
cadmium
mouse strain
Centers for Disease Control and Prevention
cadmium chloride
Coburn-Forster-Kane (equation or modeling)
Code of Federal Regulations
cyclic GMP
formaldehyde
formic acid
March 2009
xxiii DRAFT-DO NOT QUOTE OR CITE
-------
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
C02 carbon dioxide
COD coefficient of divergence
CoH, COH coefficient of haze (a measurement of visibility interference in the atmosphere, as the
quantity of dust and smoke in a theoretical 1,000 linear feet of air).
COHb carboxyhemoglobin (% concentration measured in (mL CO/mL blood))
[COHb]0 carboxyhemoglobin concentration at time zero (0)
[COHb]t carboxyhemoglobin concentration at time t
COMb carboxymyoglobin
Complex IV mitochondrial cytochrome c oxidase complex assembly, respiratory chain complex IV in
the mitochondrial inner membrane.
Cong U.S. Congress
CONUS contiguous U.S. (United States of America)
COPD chronic obstructive pulmonary disease
CPS II Cancer Prevention Study II
C-R concentration-response (relationship)
CRC Coordinating Research Council
CrMP collapsin response mediator protein
CRP C-reactive protein
CSA Combined Statistical Area (an aggregate of adjacent CBSAs tied by specific commuting
behaviors), as defined by the 2000 census
Cu copper (element)
CVD cardiovascular disease
A delta, change, difference
d straight-line distance between monitor pairs
Dahl/Rapp salt-sensitive model of hypertension (substrain of Sprague Dawley rat)
D.C. District of Columbia
March 2009
xxiv
DRAFT-DO NOT QUOTE OR CITE
-------
df degrees of freedom
df/yr degrees of freedom per year
Dl lung diffusing capacity
DlCO lung diffusing capacity of CO (in mL/min/mmHg)
DmCO capacity for diffusion of CO into the muscle
DMT -1 divalent metal transporter-1 protein (transport and detoxification of metals)
DMV dorsal motor nucleus of the vagus nerve
DNA deoxyribonucleic acid
DOCA Deoxycorticosterone acetate
dP/dtLV left ventricular maximal and minimal first derived pressure (+dP/dtLV, -dP/dtLV)
dP/dtRV right ventricular maximal and minimal first derived pressure (+dP/dtRV, -dP/dtRV)
DSA deletion/substitution/addition
dt time spent in each location
S sigma, sum of a series
E exposure over some duration
Ea exposure to pollutant of ambient origin
E/A mitral ratio of peak early to late diastolic filling velocity (E/A).
EC elemental carbon
ED emergency department
EKG, ECG electrocardiogram; electrical activity of the heart over time, measured by an
electrocardiograph
Ena exposure to pollutant of non-ambient origin
eNOS endothelial nitric oxide synthase
EPA U.S. Environmental Protection Agency
EPO erythropoietin (stimulates production of new red blood cells)
EPR Electron Paramagnetic Resonance (spectroscopy), aka Electron Spin Resonance
(spectroscopy)
EPRI Electric Power Research Institute
ERK1/2 ERK-1 (MAPK p42) and ERK-2 (MAPK p44) (extracellular signal-regulated kinases [in
cell signaling pathway])
ESRL Earth System Research Laboratory (within GMD of NOAA)
ET-1 endothelin-1; vasoconstrictor that mediates regulation of vascular tone
ETa endothelin A (ETA) receptor subtype
ETS environmental tobacco smoke
EXPOLIS six-city European air pollution study
Factor VII FVII, a vitamin K-dependent serine protease glycoprotein (also known as stable factor or
proconvertin), initiates the process of coagulation.
Factor VIIIC blood-coagulation factor VIII. (antihemophilic factor, part of factor Vlll/von Willebrand
factor complex; acts in the intrinsic pathway of blood coagulation
FAS apoptosis stimulating fragment
FC interference filter
Fe iron
March 2009
xxv
DRAFT-DO NOT QUOTE OR CITE
-------
Fe2+ ferrous iron
Fe3+ ferric iron
FEF forced expiratory flow (Liters/second)
FEF25-75 determined from the time in seconds at which 25% and 75% of the vital capacity is
reached. The volume of air exhaled between these two times is the FEF 25-75. This value
reflects the status of medium and small sized airways.
FEM Federal equivalent method
FEV! forced expiratory volume in 1 second (volume of air exhaled into the spirometer
mouthpiece in one second, using maximum effort)
fi fraction of time spent indoors
FjCO fractional concentration of CO in ambient air (in ppm)
Finf infiltration factor
f0 fraction of time spent outdoors
FR Federal Register
FGR fetal growth restriction(s)
FRM Federal reference method
FSH follicle stimulating hormone
FVC forced vital capacity (volume of air that is expelled into the spirometer following a
maximum inhalation effort)
FVII Factor VII, a vitamin K-dependent serine protease glycoprotein (also known as stable
factor or proconvertin), initiates the process of coagulation
FW
fresh weight
G0/G1 phase
Phase of Cell Cycle in which cell commits to mitosis (division)
g= mg=
gram(s), milligram(s), microgram(s)
g/mole
unit of molar mass (grams/mole)
GAM
generalized additive model(s)
GD
gestational day
GEE
generalized estimating equations
GEM
gas extraction monitor
GFAP
glial fibrillary acidic protein
GFC
gas filter correlation
GLM
generalized linear models
GLMM
generalized linear mixed models
GMD
Global Monitoring Division (of NOAA)
GMP
guano sine monophosphate
GSH
glutathione
GSSG
oxidized glutathione
GTP
guanosine triphosphate
GWP(s)
global warming potential(s)
H
atomic hydrogen, hydrogen radical, height
h
hour
h2
molecular hydrogen
March 2009
xxvi
DRAFT-DO NOT QUOTE OR CITE
-------
H20 water
H202 hydrogen peroxide
H9c2 rat embryonic cardiomyocytes (cell line)
Hb hemoglobin
HC(s) hydrocarbon(s)
HCFC(s) hydrochlorofluorocarbon(s)
HCO formyl radical
HEAPSS Health Effects of Air Pollution among Susceptible Subpopulations (study)
hectoPascal atmospheric level measured by atmospheric pressure (e.g., 700 hectoPascals =10,000 feet
above sea level)
HEK293 human embryonic kidney cells (experimentally transformed cell line)
Hem hemorrhagic (stroke)
Hep3B Human hepatocarcinoma cell line
HF heart failure
HF high energy frequency (HRV parameter)
HFLFR high energy frequency versus low energy frequency ratio (HRV parameter)
HH hypobaric hypoxia (measured in torr)
HIF-la oxygen-responsive component of the hypoxia-inducible factor (HIF) 1 complex
his-Fe2+-his iron binding scheme in neuroglobin (Nb); hexacoordinated deoxy ferrous (Fe2+) form of
Nb, bound to histidine
HO heme oxygenase (microsomal enzyme that degrades heme protein to form endogenous
CO and iron (Fe2+))
H02 hydroperoxy radical, hydrogen dioxide
HO-1 inducible isoform of heme oxygenase (induced by stressors)
HO-2 constitutively expressed isoform of heme-oxygenase
HO/CO heme oxygenase/carbon monoxide system (signaling pathway)
HR heart rate (beats per minute), hazard ratio
H/R hypoxia followed by reoxygenation
HRV heart rate variability (beat-to-beat alterations in the heart determined by analyses of time
and frequency domains in ECG(s).
HUVEC(s) human umbilical vein endothelial cell(s)
hv photon
H7W height to width ratio
I electrical current
IARC International Agency for Research on Cancer
IC inferior colliculus (an auditory integrative section of the midbrain)
Ica L electrical current (I) through the L-type Ca2+ channel
ICAM-1 intercellular adhesion molecule
ICD implantable cardioverter defibrillators)
ICR substrain of CD1 mouse, Institute for Cancer Research
IDW inverse-distance-weighted
IHD ischemic heart disease
March 2009
xxvii
DRAFT-DO NOT QUOTE OR CITE
-------
IL-6
IL-8
INDAIR
I/O
IOM
i.p.
IQR
IR
IS
ISA
Isch
ITA
Ito
i.v.
IUGR
K+
k
kco
kHz
Km
km
Km/V|,i,ix
ko2
L, dL, mL, |xL
lag 0
lag 0-3
LBW
LCA+
LD
LDH
LDL
leeward
LF
LF/HF
LH
LOAEL
LOD
LOESS
LPS
LTP
interleukin-6
interleukin-8
Indoor Air Model
indoor to outdoor concentration ratio
Institute of Medicine
intraperitoneal injection
interquartile range
immunoreactivity
ischemic stroke
Integrated Science Assessment
ischemic (stroke)
internal thoracic artery of the heart (coronary artery bypass surgery graft)
transient outward current
intravenous (injection)
intrauterine growth restriction
potassium ion
dissociation rate, rate of CO loss from the microenvironment
dissociation rate of carbon monoxide from hemoglobin
kilohertz (1,000 cycles per second; frequency of an electrical signal)
Michaelis Constant; the substrate concentration which gives Vi Vmax (in the Michaelis-
Menten equation of enzyme kinetics)
kilometer(s)
slope of Michaelis-Menten equation; measures the efficency of an enzyme
Dissociation rate of oxygen from hemoglobin
Liter, deciLiter, milliLiter, microLiter
same day as the hospital, ED, clinic, physician visit
the three previous days before the hospital, ED, clinic, physician visit
low birth weight (<2,500 grams, (=51bs, 8 oz))
leucocyte common antigen cells
lactational day
lactate dehydrogenase (release of LDH is an indicator of cell membrane damage)
low-density lipoprotein
downwind
low energy frequency (HRV parameter)
low energy frequency to high energy frequency ratio (HRV parameter)
luetenizing hormone
lowest observed adverse effect level
limit of detection
locally weighted scatterplot smoothing (modeling method); aka LOWESS
lipopolysaccharide
long-term potentiation (hippocampal)
March 2009
xxviii
DRAFT-DO NOT QUOTE OR CITE
-------
LUR land use regression (model[s])
LV left ventricle (of heart)
LV+S left ventricular plus septum
LVDP left ventricular developed pressure
LVESP left ventricular end diastolic pressure
LVSF left ventricular shortening fraction
LVW left ventricular work
m, cm, |xm, nm meter(s), centimeter, micrometer(s) = micron(s), nanometer(s)
M, mM, |xM, nM, pM Molar, milliMolar, microMolar, nanoMolar, picoMolar
M Haldane coefficient representing the CO chemical affinity for Hb [or Mb]), Reaction
mediator.
ma moving average
MAPK mitogen-activated protein kinase(s), MAP kinase(s)
MAO-A monoamine oxidase A (mitochondrial enzyme; deaminates amine-neurotransmitters)
max maximum
Mb myoglobin
MC ultrafme particle mass concentration
METs metabolic equivalent unit(s)
mg/m3 unit of chemical concentration in air (milligrams per cubic meter)
MHC major histocompatibility complex
MI myocardial infarction, "heart attack"
Miller PEE Miller equation prediction equation estimates
min minute(s)
MIP-2 macrophage inflammatory protein-2
mitral E to A ratio mitral ratio of peak early to late diastolic filling velocity (E/A).
MMEF maximal midexpiratory flow
mmHg millimeters of mercury (unit of pressure)
MMP matrix metalloproteinase
MOA(s) mode(s) of Action
MOBILE6 (EPA) Mobile source emission factor model
MODIS Moderate Resolution Imaging Spectroradiometer
mol, nmol mole, nanomole
MONICA Monitoring of Trends and Determinants in Cardiovascular Disease (study)
MOPITT Measurement of Pollution in the Troposphere (instrument)
MPO myeloperoxidase (indicative of neutrophil degranulation)
MPT mitochondrial permeability transition
MR maternally reared; rodent pups reared with their respective dams
mRNA messenger RNA
ms millisecond(s)
MSA Metropolitan Statistical Area
MSNA muscle sympathetic nerve activity
March 2009
xxix
DRAFT-DO NOT QUOTE OR CITE
-------
MT million tons
MV02 myocardial oxygen consumption
n, N sample size
N number of monitor days
N2 nitrogen gas
NAAQS National Ambient Air Quality Standards
NADPH nicotinamide adenine dinucleotide phosphate
NADH-TR nicotinamide adenine dinucleotide - tetrazolium reductase; histochemical stain of muscle
tissue - the space between the myofibrils.
NAPAP National Acid Precipitation Assessment Program
NARSTO (formerly) North American Research Strategy for Tropospheric Ozone
NAS National Academy of Sciences
NASA National Aeronautics and Space Administration
Nb neuroglobin
NC ultrafme particle number concentration
NDIR nondispersive infrared (detection)
NE norepinephrine
NEI (EPA) National Emissions Inventory
NF -kB nuclear factor kappa B
Ni nickel (element)
NIHL noise-induced hearing loss
NMDA N-methyl-D-aspartate
NMHC(s) nonmethane hydrocarbon(s)
NMMAPS National Morbidity, Mortality, and Air Pollution Study
nmol nanomole
NN normal-to-normal (NN or RR) time interval between each QRS complex in the EKG
nNOS neuronal nitric oxide synthase (NOS)
NO nitric oxide
NO* nitric oxide free radicals, free radical species of nitric oxide
N02 nitrogen dioxide
NOAA National Oceanic and Atmospheric Administration
NOAEL no observed adverse effect level
NO*-Hb nitrosyl bound Hb
NO*-Mb nitrosyl bound Mb
NOx nitrogen oxides, oxides of nitrogen
NRC National Research Council
NTS nucleus of the solitary tract (in brainstem)
02 oxygen
03 ozone
02Hb oxyhemoglobin (% concentration in mL 02 / mL blood)
02Mb oxymyoglobin
March 2009
xxx
DRAFT-DO NOT QUOTE OR CITE
-------
OAE
otoacoustic emissions (testing)
OAQPS
Office of Air Quality Planning and Standards
OC
organic carbon
OH, OH OH*
hydroxyl group, hydroxyl radical(s)
OR
odds ratio(s)
OS
occlusive stroke
OSPM
Operational Street Pollution Model
P
penetration (of pollutant)
P,P
probability
P
number of paired hourly observations
PjCO
CO partial pressure in inhaled air (in mmHg)
P13K
phosphoinositide 3-kinase
p2jWafl/ciP1
inhibitor of CDK(s) (cyclin-dependent kinases) - regulates cell cycle arrest following
p38
P90
P450
Pa
PA
PACF
PACO
PAF
PAH
PAHT
PAN
Pa02
Pa02
PARP
Pb
PB
PBN
P-c
pCO
Pco
p-A
PDGF
PEE
PEF
PEFD(s)
DNA damage.
p38 MAP kinase
90th percentile of the absolute difference in concentrations
cytochrome P450
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 (CH3C03N02)
alveolar pressure for oxygen
arterial oxygen pressure
poly(ADP-ribose) polymerase; (involved in DNA repair; cleaved by caspases during
early apoptosis)
lead
barometric pressure (in mmHg)
the free radical inhibitors PBN (a spin trap) or N-tert-butyl-alpha-phenylnitrone (free
radical inhibitor) an organic spin trap agent designed specifically to form "stable" adducts
with free radicals in electron spin resonance studies.
average partial pressure in lung capillaries
partial pressure of CO in lung capillaries (in mmHg)
partial pressure of CO
average partial pressure of 02 in lung capillaries (mmHg)
platelet derived growth factor
prediction equation estimates
peak expiratory flow (L/min)
Personal Exposures Frequency Distributions
March 2009
xxxi DRAFT-DO NOT QUOTE OR CITE
-------
PEM(s) personal exposure monitor(s)
% percent
PHD pulmonary heart disease
Pj partial pressure of inhaled air
Pi inorganic phosphate
PIH primary intracerebral hemorrhage
PKB protein kinases B
PM particulate matter
p.m. afternoon/evening hours
PM2 5 fine particulate matter, particles with a nominal mean aerodynamic diameter less than or
equal to 2.5 |xm
PM10 particles with a mean aerodynamic diameter less than or equal to 10 |xm
PM10-2 5 coarse particulate matter, coarse fraction of PM10 (referred to as thoracic coarse particles
or coarse-fraction particles; generally including particles with a nominal mean
aerodynamic diameter greater than 2.5 |xm and less than or equal to 10 |xm).
PMN polymorphonuclear leukocytes
pmol picomole
PNC particle number concentration / count
PND post natal day
pNEM/CO probabilistic NAAQS Exposure Model for CO (preceded APEX)
PNN proportion of interval differences of successive normal-beat intervals in EKG
PNN50 proportion of interval differences of successive normal-beat intervals greater than 50 ms
in EKG
PNS peripheral nervous system
p02 partial pressure of oxygen in lung capillaries (in mmHg)
ppb parts per billion
ppm parts per million
PRB policy-relevant background
PT prothrombin time (blood coagulability endpoint)
PTB preterm birth (birth after the 20th week, but before the 38th week of human pregnancy)
PVCD peripheral vascular and cerebrovascular disease
Pv02 venous oxygen tension
PV02 peak oxygen consumption
Q cardiac output
QCP Quantitative Circulatory Physiology (model)
Qm blood flow to muscle
Qot blood flow to other tissues
r correlation coefficient
R2 coefficient of determination
R State R state of hemoglobin; structural shape of protein when binding oxygen (oxy state)
RA radial artery of the heart
RAW 264.7 mouse macrophage cell line
March 2009 xxxii DRAFT-DO NOT QUOTE OR CITE
-------
RBC red blood cell
redox status ratio of interconvertible reduced/oxidized forms of a molecule
rho(O) rho(O) cells (cells lacking mitochondrial DNA)
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 (study)
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
S second(s)
SA sphinganine
SAA serum amyloid A
SAB (EPA) Science Advisory Board
SBP systolic blood pressure
SDNN standard deviation normal-to-normal (NN or RR) time interval between each QRS
complex in the EKG
Se selenium
SEM standard error of mean
sEng soluble endoglin
SES socioeconomic status
Sess Session (of Congress)
SF6 sulfur hexafluoride (tracer gas)
sFlt soluble Fms-like tyrosine kinase-1
SGA small for gestational age
sGC soluble guanylate cyclase
SEEDS (EPA) Stochastic Eluman Exposure and Dose Simulation (model)
SE1R Spontaneously hypertensive rat strain
SIDS sudden infant death syndrome
SIPs State Implementation Plan(s)
siRNA small inhibitory RNA (silencing RNA)
SLAMS State and Local Air Monitoring Stations
SMC smooth muscle cell(s)
Sn tin
March 2009
xxxiii
DRAFT-DO NOT QUOTE OR CITE
-------
snMP tin-(IV)-mesoporphyrin (HO inhibitor)
SNP single-nucleotide polymorphism
SnPP-IX tin protoporphyrin IX, (HO inhibitor)
SO sphingosine
S02 sulfur dioxide
S042" sulfate
SOD superoxide dismutase
SOD1 cytoplasmic superoxide dismutase
SOD2 mitochondrial superoxide dismutase
SOPHIA Study of Particles and Health in Atlanta
ST-segment segment of EKG between QRS complex, and T wave. ST-segment elevation may indicate
myocardial infarction
STAT1 / STAT3 signal transducers and activators of transcription (transcription factors involved in cell
signaling)
STEMS Space-Time Exposure Modeling System
STN (EPA) Speciation Trends Network
STPD standard temperature and pressure, dry
SV stroke volume
SVEB supraventricular (atrium or atrioventricular node) ectopic beats
t photochemical lifetime
T trial
T lymphocytes thymus-dependent lymphocytes
T State T state of hemoglobin; structural shape of protein when not binding oxygen (deoxy state)
TBARS thiobarbituric acid reactive substances
TC total carbon
TFAM mitochondrial transcription factor A
Tg teragram(s)
TH tyrosine hydroxylase
THb total blood concentration of hemoglobin (in g Hb/mL blood)
THP-1 human monocyte-derived cell line, (can differentiate into macrophages)
TIA transient ischemic attack (mini-stroke)
TNF-a tissue necrosis factor alpha (WBC protein boosts immune system, too much can cause
inflammation)
torr unit of pressure, equal to 0.001316 (1/760) atmosphere
TPM total particulate matter
TSP total suspended particles
UFP ultrafme particle(s)
U.K. United Kingdom
ULTRA (Exposure and Risk Assessment for Fine and )Ultrafme Particles in Ambeint Air (Study)
URI upper respiratory infection
U. S. United States of America
USC U.S. Code
March 2009
xxx iv
DRAFT-DO NOT QUOTE OR CITE
-------
UV ultraviolet
V vanadium
V minute ventilation
•
V A alveolar ventilation (in mL/min at STPD)
VA alveolar volume, a mesuremnt of lung size
Vb blood volume (in mL)
VCo CO uptake rate (the product of DLCO and the mean PACO).
Vco endogenous CO production rate (in mL/min at STPD)
VD Dead space volume
VE ventilation rate
VEGF vascular endothelial growth factor
VLF very low energy frequency (HRV parameter)
Vmax maximum velocity (catalyzed by a fixed enzyme concentration, in the Michaelis-Menten
equation of enzyme kinetics)
V02 max maximum volume per time, of oxygen (maximal oxygen consumption, maximal oxygen
uptake or aerobic capacity)
VOC(s) volatile organic compound(s)
vol volume
VPB ventricular premature beat
V/Q ventilation-perfusion ratio
vWF von Willebrand factor (part of factor Vlll/von Willebrand factor complex; acts in the
intrinsic pathway of blood coagulation)
W width
WBC white blood cell
WHI Women's Health Initiative (Study)
windward upwind
WKY Wistar-Kyoto rat strain
ww wet weight
Xy, Xik observed hourly concentrations for time period i at sites j and k
yr year
Z/H elevation of the measurement (Z) scaled by height (H)
ZIP (code) Zone Improvement Plan (system of postal codes used in the U. S.)
Zn zinc
ZnPP IX Zn protoporphyrin IX, HO inhibitor
March 2009
xxxv
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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" (U.S. Code, 1970). 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 first external draft ISA thus serves to update and revise the
information available at the time of the previous AQCD for CO in 2000.
The Integrated Plan for Review of the NAAQS for CO (U.S. EPA, 2008b) identifies key policy-
relevant questions that provide a framework for this review of the scientific evidence. These questions
frame the entire review of the NAAQS and thus are informed by both science and policy considerations.
The ISA organizes and presents the scientific evidence such that it, when considered along with findings
from risk analyses and policy considerations, will help the U.S. Environmental Protection Agency (EPA)
address these questions during the NAAQS review:
¦ 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
March 2009
1-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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, multi-day distributed lags)?
¦ 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) govern the establishment and revision of the NAAQS.
Section 108 (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 (42 U.S.C. 7409) of the Clean Air Act 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)(1) 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
March 2009
1-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 (1980); 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 sensitive population(s) at risk, and the kind and degree of the
uncertainties that must be addressed. The selection of any particular approach to providing an 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)(1) 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
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."
March 2009
1-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 forCO, 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 (U.S. EPA, 1979a), which updated the scientific criteria upon which the initial CO standards
were based. A Staff Paper (U.S. EPA, 1979b) was prepared and, along with the AQCD, 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 (U.S. EPA, 1984b) and an updated Staff Paper (U.S. EPA,
1984a). 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 released a revised AQCD for CASAC
and public review. In a "closure letter" (McClellan, 1991) sent to the Administrator, the CASAC
concluded that the AQCD (U.S. EPA, 2000) ". . . 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) 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, 2000) and staff conclusions and
recommendations contained in the revised Staff Paper (U.S. EPA, 1992), 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 AQCD were initiated. A workshop was held in September 1998 to review
and discuss material contained in the revised 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) was
March 2009
1-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 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 (NRC, 2003) 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
review which started in 1997.
1.3. Document 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. Additional publications were identified by EPA scientists in a variety of disciplines by
combing through peer-reviewed scientific literature obtained through these ongoing literature searches,
reviewing previous EPA reports, and a review of reference lists from important publications. All relevant
epidemiologic, human clinical, and animal toxicological studies, including those related to exposure-
response relationships, mode(s) of action (MOA), or susceptible subpopulations published since the last
review were considered. Added to the body of research 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. A literature search conducted on the ecological effects of ambient CO did not identify
any relevant information. Further information was acquired from consultation with scientific experts and
the public.
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, 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
March 2009
1-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
the dosimetry and pharmacokinetics of CO, including formation and fate of carboxyhemoglobin (COHb).
Chapter 5 presents a discussion of the MOA of CO and evaluates and integrates epidemiologic, human
clinical, and animal toxicological information on the health effects of 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 the Annexes, 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.
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., carbon
monoxide 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 carbon
monoxide poisoning (Ernst and Zibrak, 1998; Penney, 2007; Raub et al., 2000). The possible influence of
other atmospheric pollutants on the interpretation of the role of CO in health effects studies is considered.
This includes other pollutants with the potential to co-occur in the environment (e.g., nitrogen dioxide
[N02], sulfur dioxide [S02], ozone [03], and particulate matter [PM]). The review also assesses relevant
scientific information associated with known or anticipated public welfare effects that may be identified.
March 2009
1-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
As discussed in Section 1.3, a literature review of the ecological effects of ambient CO identified no
relevant information, and thus these effects are not assessed. With regard to climate effects this review
includes updates and additional data available since the 2000 CO AQCD on the interaction of largely
urban CO emissions and hydroxyl radical concentrations, and on background concentrations in the U.S.
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 (IOM, 2008), 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 (IOM, 2008), International Agency for Research on Cancer
(IARC, 2006), EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005), Centers for Disease
Control and Prevention (CDC, 2004), and National Acid Precipitation Assessment Program (NAPAP,
1991). These formalized approaches offer guidance for assessing causality. The frameworks are similar in
March 2009
1-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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.
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 complement the clinical and observational data; these studies can help
characterize effects of concern, exposure-response relationships, susceptible subpopulations and MOAs.
In the absence of clinical 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). 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
March 2009
1-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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; IOM, 2008). 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 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. The 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 assessments 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 characterizing
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 meta-regression may be used.
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
March 2009
1-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 time. 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. 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 asthmatics or individuals with chronic
obstructive pulmonary disease (COPD) or 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 subpopulations.
Epidemiologic studies provide important information on the associations between health effects
and exposure of human populations to ambient air pollution. These studies also help to identify
susceptible or vulnerable subgroups and associated risk factors. There are important methodological
issues to be considered in evaluating results from air pollution epidemiologic studies, especially the
potential for confounding and/or effect modification; and exposure measurement error.
Scientific judgment is needed regarding sources and magnitude of potential confounding by
covariates, together with judgment about how well the existing constellation of study designs, results, and
analyses address this potential threat to inferential validity. One key consideration is evaluation of the
potential contribution of CO to health effects when it is a component of a complex air pollutant mixture.
There are multiple ways by which CO might cause or be associated with adverse health effects. First, the
reported CO effect estimates in epidemiologic studies may reflect independent CO effects on health.
Second, ambient CO may be serving as an indicator of complex ambient air pollution mixtures that share
the same source as CO (e.g., motor vehicle emissions). Finally, copollutants may mediate the effects of
CO, or CO may influence the toxicity of copollutants.
Epidemiologists use the term "interaction" or "effect modification" to denote the departure from
the observed joint risk from what might be expected based on the separate effects of the factors. In
addition, confounding can result in the production of an association between adverse health effects and
CO that is actually attributable to another factor that is associated with CO in a particular study.
Multivariate models are the most widely used strategy to address confounding in epidemiologic studies,
March 2009
1-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
but such models are not always easily interpreted when assessing effects of covarying pollutants such as
fine particulate matter (PM2 5) and N02.
Inferring causation requires consideration of potential confounders. In confounding, 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. When associations are found in epidemiologic
studies, one approach to remove spurious associations from possible confounders is to control for
characteristics that may differ between exposed and unexposed persons; this is frequently termed
"adjustment." 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. Further, the correlation between the air pollutant of interest and various copollutants may
show temporal and spatial discongruities that can influence exposures and health effects. Thus, results of
models that attempt to distinguish gaseous and particle effects must be interpreted with caution. Despite
these limitations, the use of multipollutant models is still the prevailing approach employed in most air
pollution epidemiologic studies and may provide 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. Stratified analysis can
also be used to examine potential effect modification. 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 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.
Measurement error is another problem encountered when adjusting for spurious associations.
Controlling for confounders, whether by adjustment or stratification, is only successful when the
confounder is well-measured. Considered together, the effects of a well-measured covariate may be
overestimated in contrast to a covariate measured with greater error. There are several components that
contribute to exposure measurement error in these studies, including the difference between true and
measured ambient concentrations, the difference between average personal exposure to ambient pollutants
March 2009
1-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
and ambient concentrations at central monitoring sites, and the use of average population exposure rather
than individual exposure estimates. Previous assessments have examined the role of measurement error in
time-series epidemiologic studies using simulated data and mathematical analyses and suggested that
"transfer of effects" would only occur under unusual circumstances (e.g., "true" predictors having high
positive or negative correlation; substantial measurement error; or extremely negatively correlated
measurement errors) (U.S. EPA, 2004).
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, multi-city 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. The number and degree of diversity of covariates, as
well as their relevance to the potential confounders, remain matters of scientific judgment. Intervention
studies, because of their 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
extended to explore the effects of putative toxicants on human health, especially through the study of
model systems in other species. 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.
Testable hypotheses about the causal nature of proposed mechanisms or modes of action are central to
utilizing experimental data in causal determinations.
Interpretations of experimental studies of air pollution effects in 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, there are
uncertainties associated with quantitative extrapolations at this time between laboratory animals and
humans of observed pollutant-induced pathophysiological alterations under the control of widely varying
biochemical, endocrine, and neuronal factors.
March 2009
1-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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"1 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 in 1965 and have been widely used (CDC, 2004; IARC, 2006;
IOM, 2008; U.S. EPA, 2005, 2008d). Several adaptations of the Hill aspects have been used in aiding
causality judgments in the ecological sciences (Adams, 2003; Buck et al., 2000; Collier, 2003; Fox, 1991;
Gerritsen et al., 1998). These aspects (Hill, 1965) have been modified (Table 1-1) for use in causal
determinations specific to health and welfare effects or pollutant exposures (U.S. EPA, 2008f).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.
Table 1-1. Aspects to aid in judging causality.
Consistency of the
observed association
Coherence
Biological plausibility
An inference of causality is strengthened when a pattern of elevated risks is observed across several independent studies. The
reproducibility of findings constitutes one of the strongest arguments for causality. If there are discordant results among
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 be drawn
from a variety of experimental approaches (e.g., greenhouse, laboratory, and field) and subdisciplines of ecology (e.g., community
ecology, biogeochemistry and paleological/historical reconstructions). The coherence of evidence from various fields greatly adds
to the strength of an inference of causality. The absence of other lines of evidence, however, is not a reason to reject causality.
An inference of causality tends to be strengthened by consistency with data from experimental studies or other sources
demonstrating plausible biological mechanisms. A proposed mechanistic linking between an effect and exposure to the agent is an
important source of support for causality, especially when data establishing the existence and functioning of those mechanistic
links are available. A lack of biologic understanding, however, is not a reason to reject causality.
1 The "aspects" described by Hill (1965) 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.
March 2009 1-13 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Strength of the observed
association
Biological gradient
(exposure-response
relationship)
Experimental evidence.
Temporal relationship of
the observed association
Analogy
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.
A clear exposure-response relationship (e.g., increasing effects associated with greater exposure) strongly suggests cause and
effect, especially when such relationships are also observed for duration of exposure (e.g., increasing effects observed following
longer exposure times). There are, however, many possible reasons that a study may fail to detect an exposure-response
relationship. Thus, although the presence of a biologic gradient may support causality, the absence ofan exposure-response
relationship does not exclude a causal relationship.
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.
Evidence of a temporal sequence between the introduction ofan agent, and appearance of the effect, constitutes another
argument in favor of causality.
Structure activity relationships and information on the agent's structural analogs can provide insight into whether an association is
causal. Similarly, information on mode of action for a chemical, as one of many structural analogs, can inform decisions regarding
likely causality.
While 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). 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.
Additionally, 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 (e.g., see discussion in
(CDC, 2004).
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
IOM's Improving the Presumptive Disability Decision-Making Process for Veterans (IOM, 2008), EPA's
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005), and the U.S. Surgeon General's smoking
reports (CDC, 2004). In the ISA, EPA uses a series of five descriptors to characterize the weight of
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.
March 2009
1-14
DRAFT-DO NOT QUOTE OR CITE
-------
evidence for causality (U.S. EPA, 2008d, f). 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
Ecological and Welfare Effects
Causal Evidence is sufficient to conclude that there is a causal relationship
relationship between relevant pollutant exposures and the health outcome. That is, a
positive association has been observed between the pollutant and the
outcome in studies in which chance, bias, and confounding could be
ruled out with reasonable confidence. Evidence includes, for example,
controlled human exposure studies; or 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 there is a causal relationship
between relevant pollutant exposure and the outcome. Causality is
supported when an association has been observed between the
pollutant and the outcome in studies in which chance, bias, and
confounding could be ruled out with reasonable confidence. Controlled
exposure (laboratory or small- to medium-scale field studies) provides
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.
Likely to be a Evidence is sufficient to conclude that a causal relationship is likely to
causal exist between relevant pollutant exposures and health outcome but
relationship important uncertainties remain. That is, a positive association has been
observed between the pollutant and the outcome in studies in which
chance and bias can be ruled out with reasonable confidence but
potential issues remain. For example: a) observational studies show
positive associations but copollutant exposures 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
evidence from multiple studies, sex, or species is positive but limited or
no human data are available. Evidence generally includes replicated and
high-quality studies by multiple investigators.
Evidence is sufficient to conclude that there is a likely causal
association between relevant pollutant exposures and the outcome.
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.
Suggestive of Evidence is suggestive of a causal relationship between relevant Evidence is suggestive of an association between relevant pollutant
a causal pollutant exposures and the health outcome, but is limited because exposures and the outcome, but chance, bias and confounding cannot
relationship chance, bias and confounding cannot be ruled out. For example, at least be ruled out. For example, at least one high-quality study shows an
one high-quality study shows a positive association but the results of association, but the results of other studies are inconsistent,
other studies are inconsistent.
Inadequate to Evidence is inadequate to determine that a causal relationship exists The available studies are of insufficient quality, consistency or
infer a causal between relevant pollutant exposures and health outcome. The available statistical power to permit a conclusion regarding the presence or
relationship studies are of insufficient quantity, quality, consistency or statistical absence of an association between relevant pollutant exposure and
power to permit a conclusion regarding the presence or absence of an the outcome,
association between relevant pollutant exposure and the outcome.
Suggestive of Evidence is suggestive of no causal relationship between relevant
no causal pollutant exposures and health outcome. Several adequate studies,
relationship covering the full range of levels of exposure that human beings are
known to encounter and considering susceptible or vulnerable
subpopulations, are mutually consistent in not showing a positive
association between exposure and the outcome at any level of
exposure.
Several adequate studies, examining relationships between relevant
exposures and outcomes, are consistent in failing to show an
association between exposure and the outcome at any level of
exposure.
Source: U.S. EPA (2008f)
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
March 2009
1-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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.
Effects on Human Populations
Important questions regarding quantitative relationships include:
¦ What is the concentration-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 subpopulations appear to be differentially affected i.e., more susceptible or vulnerable
to effects?
Addressing these questions requires evaluating the entirety of policy-relevant quantitative evidence
regarding the concentration-response relationships including levels of pollutant and exposure durations at
which effects were observed, and subpopulations that differ in susceptibility or vulnerability from the
general population. This integration of evidence resulted in identification of a study or set of studies that
best approximated the concentration response relationship for the U.S. population, given the current state
of knowledge and the uncertainties that surrounded these estimates.
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
March 2009
1-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
uncertainty (such as the low data density in the lower concentration range, possible influence of
measurement error, and individual differences in susceptibility to air pollution health effects) tend to
smooth and "linearize" the concentration-response function. Additionally, 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; Crump et al., 1976; Hoel,
1980). These attributes of population dose-response may explain why the available human data at ambient
concentrations for some environmental pollutants (e.g., PM, 03, 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).
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 typically characterize how the probability of ecological and welfare effects
change in response to exposure. A challenge to the quantification of exposure-response relationships for
ecological effects is the variability across ecosystems. Ecological responses are evaluated within the range
of observations, so a quantitative relationship may be determined for a given ecological system and scale.
There is great regional and local variability in ecosystems, thus an exposure-response relationship
generally cannot be determined at the larger 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.
March 2009
1-17
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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? (American Thoracic Society, 2000). This statement updated the guidance for
defining adverse respiratory health effects that had been published 15 years earlier
(American Thoracic Society, 1985), 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 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 myocardial
infarction and sudden death.
1.7. Summary
This first external review draft ISA is a concise evaluation and synthesis of the most
policy-relevant science for reviewing the NAAQS, and it communicates critical science judgments
relevant to the NAAQS review. It reviews the most policy-relevant evidence from health and
March 2009
1-18
DRAFT-DO NOT QUOTE OR CITE
-------
1 environmental effects studies, including mechanistic evidence from basic biological science. Annexes to
2 the ISA provide additional details of the literature published since the last review. A framework for
3 making critical judgments concerning causality was presented in this chapter. It relies on a widely
4 accepted set of principles and standardized language to express evaluation of the evidence. This approach
5 can bring rigor and clarity to the current and future assessments. This ISA should assist EPA and others,
6 now and in the future, to accurately represent what is presently known—and what remains unknown—
7 concerning the effects of CO on human health and public welfare.
March 2009
1-19 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Chapter 2. Integrative Health Effects
Overview
The subsequent chapters of this ISA present the most policy-relevant information related to the
review of the NAAQS for CO, including a synthesis of the evidence presented in the 2000 CO AQCD
with the results of more recent studies. This chapter integrates important findings from the disciplines
evaluated in the assessment of the CO scientific literature, which include atmospheric sciences, ambient
air data analyses, exposure assessment, dosimetry, and health effects research (animal toxicology,
controlled human exposure, 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 pollutant exposures and health effects.
The EPA framework applied here employs a five-level hierarchy for causal determination:
¦ Causal relationship
¦ Likely to be a causal relationship
¦ Suggestive of a causal relationship
¦ Inadequate to infer a causal relationship
¦ Suggestive of no causal relationship
This evaluation led to causal determinations for a range of health outcomes, and characterization of
the magnitude of these responses, including responses in susceptible or vulnerable subpopulations, 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 recently available scientific evidence along with key
findings and conclusions from the 2000 CO AQCD that best informs consideration of the policy-relevant
questions that frame this assessment as presented in Chapter 1. Section 2.1 discusses the trends in ambient
concentrations and sources of CO and provides a brief summary of factors influencing personal exposure
to ambient CO. Section 2.2 summarizes CO dosimetry and pharmacokinetics and describes what is
known regarding the hypoxic and non-hypoxic modes of action of CO. Section 2.3 integrates the evidence
for 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. Finally, Section 2.4
March 2009
2-1
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
presents the public health impacts associated with exposure to CO, and includes evidence of effects in
potentially susceptible and vulnerable populations from CO exposure.
2.1. Ambient Concentrations, Sources, and Exposure
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 -63 of -109 million tons (MT) of total CO emissions, based on the most recent
publicly available data from EPA's National Emissions Inventory (NEI). On-road mobile source emissions
decreased by 16% since 1997, which is the most recent year covered in the 2000 CO AQCD. In
metropolitan areas in the U.S., as much as 70-75% of all CO emissions result from on-road vehicle
exhaust. The majority of these on-road CO emissions are derived from gasoline-powered vehicles since
diesel vehicles emit relatively little CO. When emissions from incomplete combustion of fuels powering
non-road 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 sources of CO include wildfires, controlled vegetation burning, residential
biomass combustion, and industrial processes. While CO emissions from non-road mobile sources, fire,
and industry have remained fairly constant, on-road mobile source CO emissions have decreased by
roughly 5% per year since the early 1990s.
Significant reductions in ambient CO concentrations and in the number of NAAQS exceedances
have been observed over the past 25 years, a continuation of trends documented in the 2000 CO AQCD.
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. 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. 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 (yr). 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 years, a 1-h avg concentration in excess of 35 ppm was reported
only once, in 2007, in Ogden, UT (39 ppm), and there were only 7 reported 8-h avg values nationwide in
excess of 9 ppm in all 3 years. 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.
Policy-relevant background (PRB) concentrations of CO were computed 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
March 2009
2-2
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Oceanic and Atmospheric Administration's (NOAA) Global Monitoring Division (GMD) and are not part
of the EPA national regulatory network. The 3-year avg CO PRB averaged -0.13 ppm in Alaska,
-0.10 ppm in Hawaii, and -0.13 ppm over the contiguous U.S. (CONUS). (Note that 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 representative of the CONUS PRB.)
A person's total personal exposure to CO is a combination of ambient and non-ambient exposures.
Ambient exposure can be further broken down into direct exposure while outdoors and exposure to
ambient CO which has infiltrated into buildings and vehicles. Several studies have shown that in the
absence of indoor sources such as ETS or gas stoves, CO concentrations are highest in automobiles and
near highly traveled roadways. Specific concentrations vary as a function of vehicle ventilation rate and
CO emissions source strength and are related to, for example, on-road vehicle numbers and speed. In-
vehicle CO concentrations are typically reported to be between 2 and 5 times higher than ambient
concentrations measured at the roadside, but have been reported to be as much as 25 times higher. A
portion of these high concentrations comes from the vehicle's own engine emissions entrained with the
ambient concentrations from the roadway, although the exact fraction of this self-pollution in the in-
vehicle total varies with body integrity of the vehicle and is not known with certainty. Concentrations of
CO, like those of other on-road vehicle pollutants, often display an inverse relationship with distance
from the roadway. For CO, this fall-off with distance can be very sharp up to -70 meters (m), beyond
which the concentration increment from roadway emissions is largely indistinguishable from local
ambient concentrations. The siting and location instructions for federal regulatory CO monitors explicitly
recognize the need to measure at distances close to mobile sources with a requirement for monitored areas
that at least one monitor be sited to measure max concentration. This is often accomplished with a
monitor situated at the CFR-defined microscale of 2-10 m from the roadway; these microscale monitors
also have sample inlets mounted at 3 ± 0.5 m above ground level (AGL), unlike those at larger scale
distances whose inlet heights can vary between 2 and 15 m AGL. In 2007, there were at least 70 CO
monitors described as microscale reporting to AQS.
Although the correlations across CO monitors sited to sample at different scales can be greater than
0.8 in some areas, they also 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. At the same time,
personal exposure monitoring captures both ambient and non-ambient CO concentrations. Because
non-ambient CO exposure is not expected to be correlated with ambient CO concentrations at the
monitor, the non-ambient contribution to total personal CO exposure complicates interpretation of health
effects relationships observed in epidemiologic studies. For the general U.S. population, exposure error
analysis for epidemiologic studies, as summarized in Chapter 3, indicates that fixed-site measured
ambient CO concentration is generally a good indicator of ambient exposure to CO.
March 2009
2-3
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2.2. Dosimetry, Pharmacokinetics, and Mode of Action
2.2.1. Dosimetry and Pharmacokinetics
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
(02)-carrying capacity of blood and impairs the release of 02 from oxyhemoglobin (02Hb) to the tissues.
The 2000 CO AQCD has a detailed description of the well-established Coburn-Forster-Kane (CFK)
equation, which has been used for many years to model COHb formation. More recent work since then
has developed models 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 CO concentration 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, including health compromised individuals and developing fetuses,
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 C02.
2.2.2. Mode of Action
The diverse effects of CO are dependent upon dose, 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 mechanisms unrelated
to tissue hypoxia including 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 underlying
mechanism for both the hypoxic and non-hypoxic biological responses to CO.
March 2009
2-4
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
As discussed in the 2000 CO AQCD, 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
02-carrying capacity of blood, but it also impairs the release of 02 from 02Hb. Compensatory alterations
in hemodynamics, such as vasodilation and increased cardiac output, protect from tissue hypoxia. At
ambient CO concentrations, these compensatory changes are slight and likely tolerable in people with a
healthy cardiovascular system. However, people with cardiovascular detriments may be unable to endure
these small changes in hemodynamics which may lead to the presentation of health effects as described in
Sections 5.2.1 and 5.2.2. Binding of CO to Mb, as discussed in the 2000 CO AQCD and in Section
4.3.2.3, can also impair the delivery of 02 to tissues. Mb has a high affinity for CO; however,
physiological effects are seen only after high dose exposures to CO, resulting in COMb concentrations far
above baseline levels.
Non-hypoxic mechanisms underlying the biological effects of CO have been the subject of a
substantial amount of recent research since the 2000 CO AQCD. 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 cytochrome c oxidase, heme
loss from proteins, disruption of iron homeostasis, and alteration in cellular redox status. CO is a
ubiquitous cell signaling molecule and the physiological functions of HO-derived CO are numerous. The
endogenous generation and release of CO from HO-1 and HO-2 is tightly controlled, as is any
homeostatic process. Thus, exogenously-applied CO has the capacity to disrupt myriad 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. 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 can affect endogenous CO signaling pathways
and lead to adverse health effects is an open question for which there are no definitive answers at this
time.
March 2009
2-5
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2.3. Health Effects
Table 2-1. Causal determinations for health effects outcomes.
Outcome Category
Exposure Period
Causality Determination
Cardiovascular morbidity
Short-term
Likely to be 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
Respiratory morbidity
Short-term
Long-term
Suggestive of a causal relationship
Inadequate to infer a causal relationship
Mortality
Short-term
Suggestive of a causal relationship
Long-term
Suggestive of no causal relationship
2.3.1. 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 CAD (see Section 5.2). These studies, described in the 1991 and 2000 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 3-6%, with one multicenter study reporting similar
effects at COHb levels as low as 2.4%. No human clinical studies have evaluated the effect of controlled
exposures to CO resulting in COHb levels lower than 2.4%. Human clinical studies published since the
2000 CO AQCD have reported no association between CO and ST-segment changes or arrhythmia;
however, none of these studies included individuals with diagnoses of 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
were not representative of the most sensitive individuals in the population. In fact, the most sensitive
individuals may respond to levels of COHb lower than 2.4%. 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. First of all, endogenous CO production varies as described in Section 4.5, but
generally results in less than 1% COHb. Secondly, nonambient exposures to CO, such as exposure to
ETS, can increase COHb above baseline levels. Ambient exposures will result in an additive increase in
March 2009
2-6
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
COHb. Using mathematical modeling to predict changes in COHb in healthy inactive adults (Quantitative
Circulatory Physiology [QCP] model, Section 4.2.3), it can be estimated that exposure to 35 ppm CO for
1 h results in an increase of 0.6% COHb over baseline and exposure to 9 ppm CO for 8 h results in an
increase of 0.8% COHb over baseline. Furthermore, 24-h exposure to 3 ppm CO results in an increase of
0.4% COHb above baseline which can also be obtained following 1-h exposure to 30 ppm CO.
Consequently, exposure to CO at concentrations relevant to the NAAQS has the potential to increase
COHb to levels associated with adverse cardiovascular health effects in some individuals.
Findings of controlled human exposure studies are coherent with findings of recent epidemiologic
studies conducted since the 2000 CO AQCD, which observed associations between ambient CO
concentration and emergency department (ED) visits and hospital admissions for ischemic heart disease
(IHD), congestive heart failure (CHF) and all-cause cardiovascular disease (CVD). All but one of these
epidemiologic studies were conducted in locations where the entire distribution of CO concentrations
were at or below the level of the current NAAQS, with mean 24-h avg concentrations ranging from
0.5 ppm (Montreal, Canada) to 9.4 ppm (Tehran, Iran) (Table 5-7). A single study reported a negative
association between CO concentration and hospital admissions and ED visits for IHD among all ages; all
other associations were positive, with increases in hospital admissions and ED visits for IHD between
0.2% and 19.8% per standardized increase in CO concentration (Figure 5-1). These recent studies build
upon the conclusions of the 2000 CO AQCD that short-term variations in ambient CO concentrations are
associated with daily hospital admissions for heart disease.
These health outcomes are consistent with a role for CO in limiting 02 availability (i.e., hypoxic
mechanisms) in individuals with CAD. However, recent toxicological studies suggested that CO may also
act through non-hypoxic mechanisms by disrupting cellular signaling. Studies in healthy animals
demonstrated oxidative injury and inflammation in response to 50-100 ppm CO while studies in disease
models demonstrate effects on heart rhythm and exacerbation of cardiomyopathy and vascular remodeling
in response to 35-50 ppm CO. Furthermore, in utero exposure to 150 ppm CO alters postnatal
elecrophysiological maturation in rat cardiomyocytes. Further investigations will be useful in determining
the importance of non-hypoxic mechanisms following environmentally-relevant CO exposures. Taken
together, the evidence from epidemiologic, human clinical, and toxicological studies is Sufficient to
conclude that a causal relationship is likely to exist between relevant short-term CO exposures and
cardiovascular morbidity.
2.3.2. 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,
March 2009
2-7
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 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 employed rodent models to
show that low level CO exposure during the in utero 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 including active avoidance behavior (150 ppm CO),
non-spatial memory (75 and 150 ppm CO), spatial learning (endogenous CO inhibition), homing behavior
(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 in exposed male rodents. 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 with 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 which is mechanistically proposed to be mediated by
reactive oxygen species (ROS) (25 ppm CO). Functional tests of the auditory system of neonatally, low
level CO-exposed rodents, using otoacoustic emissions (OAE) testing (50 ppm CO) and amplitude
measurements of the 8th cranial nerve action potential (12, 25, 50, 100 ppm CO), revealed decrements in
auditory function at PND22 and permanent changes into adulthood using AP testing (50 ppm CO).
Together, these animal studies demonstrate that in utero exposure to CO can adversely affect adult
behavior, neuronal myelination, neurotransmission, and the auditory system in adult male rodents.
Considering the combined evidence from controlled human exposure and toxicological studies, the
March 2009
2-8
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
evidence is suggestive of a causal relationship between relevant short- and long-term CO exposures
and central nervous system effects.
2.3.3. 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, which included only two studies that examined the effect of ambient CO on low birth weight
(LBW). Since then, a number of studies have been conducted looking at varied outcomes, including PTB,
birth defects, fetal growth (including LBW), 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 results 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. 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.
This insult to the heart is coherent with results of human clinical studies demonstrating the heart as a
target for CO effects (Section 5.2). Animal toxicological studies provide additional evidence for such an
insult to the heart, and reported transient cardiomegaly at birth after continuous in utero CO exposure (60,
125, 250 and 500 ppm CO), delayed myocardial electrophysiolgical maturation (150 ppm CO), or
systemic splenic immunocompromise (75 or 150 ppm CO). Toxicological studies have also shown that
exogenous continuous in utero CO exposure (250 ppm) induced teratogenicity in rodent offspring in a
dose-dependent manner that was further exacerbated by dietary protein restriction (65 ppm CO) or zinc
depletion (500 ppm CO). Toxicological studies of exogenous 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
March 2009
2-9
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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.
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 do 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 in neonatal CO exposure. Auditory
functional testing using OAE (50 ppm CO) and ABR (12, 25, 50, 100 ppm CO) on rodents exposed
perinataly to CO showed that CO-exposed neonates had auditory decrements at PND22 (OAE and ABR)
and permanent changes into adulthood with ABR (50 ppm CO).
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 non-hypoxic mechanisms that could lead to such effects 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 a Causal relationship
between long-term exposures to relevant CO concentrations and developmental effects and birth
outcomes.
2.3.4. Respiratory Morbidity
New epidemiologic studies, supported by the body of literature summarized in the 2000 CO
AQCD, 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 (discussed in Section 5.5). However, the interpretation of the results from
epidemiologic studies is difficult due to the lack of extensive copollutant analyses along with the
moderate to high correlation between CO and other combustion/traffic generated pollutants. To date the
March 2009
2-10
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
majority of the literature has not extensively examined the association between CO and respiratory
morbidity due to studies focusing primarily on effects associated with exposure to other criteria
pollutants, namely PM and 03. This has contributed to the inability to disentangle the effects attributed to
CO from the larger complex air pollution mix. In addition, uncertainty as to a biological mechanism to
explain the respiratory-related effects observed in the epidemiologic literature further complicates the
interpretation of these results, especially considering the low ambient CO concentrations reported (24-h
avg: 0.35-2.1 ppm). However, animal toxicological studies do provide some evidence that short-term
exposure to CO (50-100 ppm) can cause oxidative injury and inflammation and alter pulmonary vascular
remodeling. Human clinical studies have not extensively examined the effect of short-term exposure to
CO on respiratory morbidity, specifically pulmonary function. The limited number of clinical studies that
have been conducted prior to and since the 2000 CO AQCD provide very little evidence of any adverse
effect of CO on the respiratory system at COHb levels <10%. Although human clinical studies have not
provided evidence to support CO-related respiratory health effects, the epidemiologic studies that
examined the effects of short-term exposure to CO and lung-related outcomes show positive associations
and animal toxicological studies demonstrate the potential for an underlying biological mechanism, which
together provide evidence that is suggestive of a causal relationship between short-term exposure to
relevant CO concentrations and respiratory morbidity.
Currently, only a few studies have been conducted that examine the association between long-term
exposure to CO and respiratory morbidity. 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 co-pollutant analyses to disentangle the respiratory effects associated with CO due to its high
correlation with N02 and other combustion-related pollutants. Overall, the evidence available is
inadequate to conclude that a causal relationship exists between long-term exposure to relevant
CO concentrations and respiratory morbidity.
2.3.5. Mortality
Among the gaseous pollutants examined in time-series mortality studies, CO is the least frequently
studied criteria air pollutant. Because CO was mostly treated as a potential confounder in these studies,
the information available regarding the nature of the association between short-term exposure to CO and
mortality is limited compared to the other pollutants. However, the recently available multi-city studies,
March 2009
2-11
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
which consist of larger sample sizes, and single-city studies generally confirmed the findings reported in
the 2000 CO AQCD (see Section 5.6).
The multi-city studies report comparable CO mortality risk estimates for total (non-accidental)
mortality with the APHEA2 European multi-city study showing slightly higher estimates for
cardiovascular mortality in single-pollutant models. However, when examining potential confounding by
copollutants these studies consistently showed that CO mortality risk estimates were reduced when N02
was included in the model, but this observation may not be "confounding" in the usual sense in that N02
may also be a surrogate marker of other pollutants or pollution sources (i.e., traffic).
Only one of the multi-city studies focused specifically on the CO-mortality association (the
APHEA study), examining: (1) model sensitivity; (2) the CO-mortality concentration-response (C-R)
relationship; and (3) potential effect modifiers of CO mortality risk estimates. The APHEA2 study
performed a sensitivity analysis, which indicated an approximate 50 - 80% difference in CO risk
estimates from a reasonable range of alternative models. In addition, the study examined the CO-mortality
C-R relationship through a grid search of varying threshold points, and found only weak evidence of a CO
threshold at 0.5 mg/m3 (0.43 ppm), but this result is complicated by the lowest 10% of the CO distribution
for seven of the 19 cities examined being at or above 2 mg/m3 (1.74 ppm). The examination of a variety
of city-specific variables to identify potential effect modifiers of the CO-mortality relationship found that
geographic region explained most of the heterogeneity in CO mortality risk estimates with the
CO-mortality associations being stronger in western and southern European cities than eastern cities. A
similar pattern has been reported for black smoke (BS) and S02 in previous APHEA studies, but the
geographic variability observed does not provide specific information which could be used to evaluate the
CO-mortality association.
The results from the single-city studies evaluated are generally consistent with the multi-city
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 other
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 new multi- and single-city studies evaluated provide evidence 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 and it is unclear if CO is acting alone or as a
surrogate for other combustion-related pollutants. In addition, the results also underscore the limitation of
current analytical methods to disentangle the health effects associated with one pollutant in the complex
air pollution mixture. Overall, the epidemiologic evidence is Suggestive of a Causal relationship
between short-term exposure to relevant CO concentrations and mortality.
March 2009
2-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The evaluation of new epidemiologic studies conducted since the 2000 CO AQCD 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. The re-analysis of the
American Cancer Society (ACS) data by Jerrett et al. (2003) found no association between long-term
exposure to CO and mortality. Similar results were obtained in an updated analysis conducted by the
original ACS investigators when using earlier (1980) CO data, but negative associations were found when
using more recent (1982-1998) data. The Women's Health Initiative (WHI) Study also found no
association between CO and CVD events (including mortality) using the 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. In addition, a cross-sectional
study of U.S. counties reported results that are generally consistent with the cohort studies: positive
associations between long-term exposure to PM2 5 and sulfate (S042 ) 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
subpopulations (i.e., post-menopausal women and hypertensive men) combined with (1) the lack of
evidence for respiratory and cardiovascular morbidity outcomes following long-term exposure to CO; and
(2) the absence of a proposed mechanism to explain the progression to mortality following long-term
exposure to CO provide supportive evidence that is suggestive of no causal relationship between long-
term exposure to relevant CO concentrations and mortality
2.4. Public Health Impacts
2.4.1. Concentration-Response Relationship
Currently, very limited information is available in the human clinical and epidemiologic literature
regarding the CO C-R relationship and the potential presence of a CO threshold. Two human clinical
studies described in the 2000 CO AQCD have evaluated the C-R relationship between CO and onset of
exercise-induced angina among individuals with CAD, but at the high end of CO concentrations (i.e., CO
levels above the current NAAQS). Anderson et al. (1973) exposed 10 adult men with stable angina for 4 h
to CO concentrations of 50 and 100 ppm, which resulted in average COHb levels 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 two concentrations of CO (average exposure concentrations of 117 and 253 ppm) resulting in average
March 2009
2-13
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
pre-exercise COHb levels of 2.4% and 4.7% (Allred et al., 1989). Relative to control (average COHb
0.7%), COHb levels of 2.4% and 4.7% were observed to decrease the time to onset of angina by 4.2% (p
= 0.054) and 7.1 % (p = 0.004), respectively. In addition, these investigators reported a statistically
significant decrease in the time to exercise-induced ST-segment depression with increasing COHb levels.
These findings provide some evidence of a significant C-R relationship at COHb concentrations between
2.4 and 4.7%. However, the human clinical literature has yet to evaluate the C-R relationship at lower CO
concentrations or COHb levels.
One study 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) 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)
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 seven 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.
2.4.2. Potentially Susceptible and Vulnerable Subpopulations
The examination of both susceptible and vulnerable subpopulations to CO exposure allows for the
NAAQS to provide an adequate margin of safety for both the general population and sensitive
subpopulations (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 or vulnerability of an individual to CO-related health effects. In this ISA, a susceptible
subpopulation is defined as those individuals with intrinsic biological characteristics that might exhibit an
adverse health effect to a pollutant at concentrations lower than those needed to elicit the same response
in the general population or those individuals that might elicit a more adverse health effect at the same
concentration. A vulnerable subpopulation is defined as those individuals with external, nonbiological
factors that increase the risk of adverse health effects, such as differential exposure or living at altitude.
March 2009
2-14
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
The most important susceptibility characteristic for increased risk due to CO exposure is CAD.
Individuals with heart disease may be at a greater risk from CO exposure since they may already have
compromised 02 delivery. CO is notable among air pollutants because it is especially harmful in
individuals with impaired cardiovascular systems. Persons with a normal cardiovascular system can
tolerate substantial concentrations of CO, if they vasodilate in response to the hypoxemia 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 IHD. Other medical conditions that confer increased susceptibility include obstructive
lung disease, which impairs airflow to the lungs, and conditions such as anemia that alter the blood 02
carrying capacity or content and result in a greater risk from COHb induced hypoxia and decreased tissue
02 delivery. Age is important for susceptibility in older adults, who have an increased elimination time for
COHb. Newborns can also be considered a susceptible subpopulation due to critical phases of
development and differences between fetal and maternal CO pharmacokinetics in utero.
Subpopulations considered vulnerable to CO-related health effects include people who spend a
considerable amount of time in or near traffic, live at high altitude, exercise and/or use medications which
may increase endogenous CO production. 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
experience elevated CO concentrations and therefore constitute a potentially vulnerable subpopulation
due to differential exposure. Vulnerable subpopulations also include individuals at high altitude, who
undergo physiologic changes that favor increased CO uptake and COHb formation. Exercising
individuals are also considered vulnerable because exercise facilitates CO uptake and transport by
increasing gas exchange efficiency and the COHb elimination rate decreases with physical activity.
Individuals who use certain medications which may increase endogenous CO production and the baseline
level of COHb are also considered vulnerable to CO-related health effects.
March 2009
2-15
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Chapter 3. Source to Exposure
3.1. Introduction
This chapter contains basic information about concepts and findings in atmospheric sciences and
exposure assessment pertaining to CO levels relevant for establishing a foundation for the detailed
discussions of health effects data in subsequent chapters. Section 3.2 provides an overview of the sources
of CO and examples of their spatial distribution. Atmospheric chemistry involved in the production and
removal of CO by oxidation processes is discussed in Section 3.3. Descriptions of CO measurement
methods and monitor siting requirements and locations are presented in Section 3.4. Data for ambient CO
concentrations are characterized in Section 3.5. Policy relevant background concentrations of CO,
i.e., those concentrations defined to result from uncontrollable emissions, are also presented in Section
3.5. Finally, factors related to human exposure to CO are discussed in Section 3.6.
3.2. Sources and Emissions of CO
CO is a colorless, odorless, tasteless gas consisting of one carbon (C) atom covalently bonded to
one 02 atom; its molar mass is 28.0101 g/mol. CO is formed primarily by incomplete combustion of
carbon-containing fuels and photochemical reactions in the atmosphere. In general, any increase in fuel
02 content, burn temperature, or mixing time in the combustion zone will tend to decrease production of
CO relative to C02; hence 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. Internal combustion engines used in mobile sources, by contrast, have widely varying
operating conditions and, thus, inherently higher and varying CO formation.
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 data. In the U.S., CO emissions data are tracked in the
National Emissions Inventory (U.S. EPA, 2006a), a composite of data from various sources including
industries and state, tribal, and local air agencies. 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 engineering calculations and estimates 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 (U.S. EPA, 2007). Although these
March 2009
3-1
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
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.
Nationally, on-road mobile sources in the NEI constituted more than half of total CO emissions in
2002, or ~63 MT of -109 MT total. For this reason, high concentrations of CO can often occur in areas of
heavy traffic. In metropolitan areas in the U.S., for example, as much as 70-75% of all CO emissions
came from on-road vehicle exhaust in the 2002 NEI (U.S. EPA, 2006a). The majority of these on-road CO
emissions derive from gasoline-powered vehicles since the 02 content, pressure, and temperature required
for diesel fuel ignition result in much less CO production. 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; see Figure 3-1.
Nationwide Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood
Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
0 100
Emissions (MT)
Source: U.S. EPA (2006a)
Figure 3-1. CO emissions (MT) in the U.S. by source sector in 2002.
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 includes chemical production,
~ 3
I
|1
|2
I
March 2009
3-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
petroleum refining, metals production, and industrial processes other than fuel combustion; (3) on-road
vehicles, which includes cars, trucks, buses, and motorcycles; and (4) non-road vehicles and engines, such
as farm and construction equipment, lawnmowers, chainsaws, boats, ships, snowmobiles, aircraft,
locomotive, and others. Using these NEI data, trends in the national CO emissions can be computed and
compared overtime. So, for example, the national-scale estimated anthropogenic CO emissions decreased
35% between 1990 and 2002; see Figure 3-2. The trend plot 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 to 2002 because only the 1990 data has been updated to be
comparable to the more recent inventories made since 1996.
160
140
Fuel combustion
GO
120
Other industrial processes
100
CO
On-road vehicles
CO
CO
Nonroad vehicles and engines
'90
'00
Year
Source: U.S. EPA (2006a, 2008c)
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).
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 included
Flarley et al. (2001; 2005), Parrish et al. (2002; 2006), Pollack et al. (2004), and Mobley et al. (2005). The
consistent conclusion from those investigations has been that annual average U.S. on-road vehicle CO
emissions decreased at a rate of -5% per year since the early 1990s. Additional analyses by Harley et al.
(2005), Parrish et al. (2002) and Parrish (2006) were also consistent with the suggestion in Pollack et al.
(2004) that the EPA MOBILE6 vehicle emissions model (http://www.epa.gov/otaa/m6.htm) now
overestimates vehicle CO emissions by a factor of ~2.
March 2009
3-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Biogenic emissions can vary widely from year to year (Bergamaschi et al., 2000). CO isotopes,
most notably 14CO, have been especially useful in helping refine emissions estimates and partition them
between combustion and other sources. Weinstock and Niki (1972) analyzed measurements of the 14C
content of CO in Tonawanda, NY (a suburb of Buffalo) obtained by MacKay et al. (1963) during the
winter of 1960 and found at that time that a large proprotion of the CO was produced by sources
containing modern carbon and not fossil carbon. The main source of this modern carbon CO was
conjectured by Weinstock and Niki (1972) to have been from photochemical oxidation of CH4; however,
CH4 has a long atmospheric life time (about 8 years) but does oxidize to compounds like methyl radical
and also is a contributor to 03 formation. Other work suggested that CO sources such as isoprene
oxidation and biomass burning were also important. For example, Conny (1998) reviewed studies using
measurements of 14CO to characterize winter sources of CO in Las Vegas, NV, and Albuquerque, NM.
These studies concluded that the contribution from residential wood burning could have been as high as
30% in the samples collected.
Estimates of non-anthropogenic CO emissions are made using the Biogenic Emissions Inventory
System (BEIS) model with data from the Biogenic Emissions Landcover Database (BELD) and annual
meteorological data; see http://www.epa.gov/ttnchiel/emch/biogenic. National biogenic emissions,
excluding fires, were estimated to contribute -5% of total CO emissions from all sources in 2002; fires in
2002 added another 13%, or -14.5 MT, to the national CO emissions total. 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.
On a global scale, however, the magnitude of their contribution is small relative to all 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). 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). Soil uptake of CO is driven by anaerobic bacteria (Inman et al., 1971). 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).
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 to 30% initiated by lightning (Andreae, 1991). However,
because fire management practices suppress natural wildfires, the buildup of fire fuels increases the
March 2009
3-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
susceptibility of forests to more severe but less frequent fires in the future. 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), 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; see Figure 3-3. The high CO concentration
measured by MOPITT coincided with the surface location of fires using aerosol plumes identified by the
Moderate Resolution Imaging Spectroradiometer (MODIS) also on Terra. Subsequent modeling by Pfister
et al. (2005) showed that the CO contribution from these fires in July 2004 was 30 (±5) teragrams (Tg)
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) 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 ten administrative Regions in the U.S. depicted in Figure 3-4 show a
more nuanced view of the national concentrations and trends described just 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.
On still finer scales, CO emissions from on-road mobile sources or from fires can dominate in
different places across the U.S. Figure 3-5 illustrates this variability with CO state-level emissions total
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, though emissions
varied strongly across counties with urban Denver County dominated by on-road vehicle emissions at
71% and rural Garfield County dominated by fire emissions at 67%.
March 2009
3-5
DRAFT - DO NOT CITE OR QUOTE
-------
-150 -120 -90 -60 -30 0
Longitude
100
110
120
130
140
150
160
170
180
190
200 ppbv
Source: Fishman et al. (2008)
Figure 3-3. CO concentrations measured by satellite at the 700 hectoPascal level (~10,000 feet
above sea level) from MOPITT for the period 15-23 July 2004 during intense wildfires
in Alaska and Yukon.
'90 '96 '97 '98 '99 '00 '01 02
Year
(D
O
0
Data are presented for 1990 EPA Regions
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.
Source: U.S. EPA (2006a, 2008c)
Figure 3-4. Trends in sub-national CO emissions in the 10 U.S. EPA Regions for 1990 and 1996
to 2002.
March 2009
3-6
DRAFT - DO NOT CITE OR QUOTE
-------
Denver
County
Garfield
County
iver
Cartucn Mcno»cte Emissions in 2002 (Tons per Square Mile)
"J 108 - 5.G9
I 5.45 - 1526
I "8.96 - TB7.70
Colorado State Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
ZJ 1,110,980
¦ 966,816
ZJ 385,460
Residential Wood Combustion
ID 78,308
Fossil Fuel Combustion
] 25,743
Electricity Generation
| 7,290
industrial Processes
| 6,838
Waste Disposal
527
Miscellaneous
334
Solvent Use
53
Denver County Emissions 2002
1,400,000
Emissions (Tons)
Garfieid County Emissions 2002
On-Road Vehicles
1129.554
Fires
0
Non-Road Equipment
Residential Wood Combustion
148,658
1,864
Fossil Fuel Combustion
905
Electricity Generation
391
Industrial Processes
305
Waste Disposal
5
Miscellaneous
36
Solvent Use
0
On-Road Vehicles
120,309
Fires
Non-Road Equipment
] 5,607
Residential Wood Combustion
13,121
Fossil Fuel Combustion
J 2,239
Electricity Generation
45
Industrial Processes
135
Waste Disposal
0
Miscellaneous
3
Solvent Use
0
Emissions (Tons)
Emissions (Tons)
Figure 3-5. CO emissions density map and distribution for the state of Colorado, and for
selected counties in Colorado.
March 2009
3-7
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.3. Physics and Chemistry of Atmospheric CO
In addition to being emitted directly by combustion sources, CO is produced by photooxidation of
methane (CH4) and other VOCs including nonmethane hydrocarbons (NMHCs) in the atmosphere, and of
organic molecules in surface waters and soils. CH4 oxidation is summarized in the following reaction
sequence:
CH4+OH ~^ch3 +h2o
CH3 + 02(+M) -> CH302(+M)
CH302 +NO -> CH30 + N02
CH30 + 02 ^ CH20 + H02
CH20 + hv^H2 +CO
or CH20 + hv^HC0+H
or CH20 + OH -> HCO + H20
HC0 + 02 ~^co+ho2
where M is a reaction mediator providing collisional energy but is neither created nor destroyed.
Photolysis of formaldehyde (CH20) 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 02 to form
hydroperoxy radical (H02) and CO. Reaction of methyl peroxy radical (CH302) with H02 radicals
(reaction not shown) to form methyl hydroperoxide (CH3OOH) is also operative, especially in low oxides
of nitrogen (NOx) conditions. Heterogeneous removal of the water-soluble intermediate products
CH3OOH, CH20, and radicals will decrease CO yields from CH4 oxidation.
While oxidation of CH20 nearly always produces CO and some small quantities of formic acid
(CH202) in the reaction of CH20 with H02 (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 02 in a termolecular recombination reaction to form peroxyacyl radicals,
which then can react with nitric oxide (NO) to form CH3 and C02; or the peroxyacyl radicals can react
with N02 to form peroxyacetyl nitrate (PAN), CH3C03N02. In this way, one carbon atom is oxidized
directly to C02 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 peroxyacyl
radicals also will react with other odd hydrogen radicals like OH.
Estimating the CO yield from oxidation of hydrocarbons (HCs) larger than CH4 requires computing
the yields of CH20, 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 with more than a few
March 2009
3-8
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
carbons is still unknown, and 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. It has long been known that as much as 30% of the carbon in HCs in many urban
areas is in the form of aromatics largely from mobile sources since gasoline contains significant quantities
of aromatics (Grosjean and Fung, 1984; Seila et al., 1989). Yet 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). Yields from other compounds range from less
than 0.1 for anthropogenic alkanes (Altshuller, 1991) to -0.7 for some other non-CFU HCs; see
Kanakidou and Crutzen (1999) for CO yields from other HCs.
The major pathway for removal of CO from the atmosphere is reaction with OH to produce C02
and H radicals that rapidly combine with 02 to form H02 radicals with a rate constant at 1 atm in air of
-2.4 x 10"13 cm3/molecule/s (Finlayson-Pitts and Pitts Jr). The mean tropospheric photochemical lifetime
(x) of CO in the northern hemisphere is -57 days (Khalil and Rasmussen, 1990; Thompson and Cicerone,
1986). Owing to variation in atmospheric water vapor, OH concentration, and insolation, shorter x 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. The CO x is shorter than the characteristic time
scale for mixing between the hemispheres of -1 year; hence a large gradient in concentrations can exist
between the hemispheres. In addition, the CO x at high latitudes is long enough to result in much smaller
gradients between 30° latitude and the pole of either hemisphere. The typical residence times of CO in
urban areas when assuming a diel-average OH concentration of 3 / 106/cm3 in urban areas is -16 days, so
CO will not 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 destroyed by indoor air reactions.
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), and
subsequently confirmed in the global assessments of climate change by the Intergovernmental Panel on
Climate Change (IPCC, 2001, 2007). CO is a weak contributor to greenhouse warming because its
fundamental absorption band near 4.63 (.un is far from the spectral maximum of earth's longwave
radiation at -10 (.im. However, because reaction with CO is also the major sink for OH, changes in CO
concentrations can lead to changes in the concentrations of other trace gases whose loss processes involve
OH attack. Some of these trace gases, CH4, for example, absorb infrared radiation from the earth's surface
and contribute to the greenhouse effect directly; others, including the hydrochlorofluorocarbons (HCFCs)
March 2009
3-9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
and methyl chloride and methyl bromide, can deplete stratospheric 03, increasing the surface-incident UV
flux.
This indirect effect of CO on stratospheric 03 concentrations is opposite in sign to the effect of CO
on 03 in the troposphere where CO reacts in a manner similar to other VOCs in the presence of NOx and
UV to create 03. (See the detailed description of 03 formation from VOCs and NOx in the 2008 NOx ISA
(U.S. EPA, 2008e). The long chemical lifetime and one-to-one stoichiometry of CO oxidation (whereby
one molecule of CO converts only one molecule of NO to N02) means that CO has significantly lower 03
forming potential than other VOCs in the troposphere. Carter (1998) computed a maximum incremental
reactivity for CO of 0.07 g 03 for 1 g CO, as compared to reactivities of total on-road vehicle exhaust
emissions in the range of 3 to 4. 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
greatly to 03 formation even though its photochemical processing is slow. Using data from instrumented
models including that of Jeffries (1995), NRC (1999) estimated, for example, that CO can contribute 15-
25% of the total 03 forming potential of gasoline exhaust emissions; this figure varies from region to
region. The contribution of CO to urban and regional 03 concentration is often less than 10% owing to its
very slow reactivity on these scales and to locally variable radical concentration ratios.
Because the greenhouse warming effects from CO are nearly completely indirect, and because CO
concentrations are spatially heterogeneous, neither the IPCC nor EPA computes global warming potentials
(GWPs) for CO, just as they do not for tropospheric 03, NO, N02, or VOCs (U.S. EPA, 2008a).
Additionally, urban and regional-scale oxidation of CO to C02 under current atmospheric conditions
proceeds very slowly and IPCC considers production of C02 through this pathway to be double counting
of CO effects (IPCC, 2007).
3.4. Ambient Measurements
3.4.1. Ambient Measurement Instruments
To promote uniform 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
December 2008, 19 automated FRMs and no FEMs had been approved for CO
(http://www.epa.gov/ttn/amtic/criteria.html').
March 2009
3-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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. The most sensitive trace-level versions of these instruments can detect minimum CO
concentrations of -0.04 ppm; the required lower detection limit for FRMs in the EPA network is 1.0 ppm
(40 CFR 53.20 Table B-l).
NDIR is based on the physics of CO's characteristic infrared absorption near 4.63 |im. 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. Earlier
concerns over zero-drift and nonlinear span responses have been addressed. 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), 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 C02, 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 (FC),
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 N2. Radiation transmitted through the
CO is completely attenuated at the wavelengths where CO absorbs strongly. The radiation transmitted
through the nitrogen gas (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.
March 2009
3-11
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.4.2. Ambient Sampling Network Design
3.4.2.1. Monitor Siting Requirements
CO monitoring is included at all active NCore sites and State and Local Air Monitoring Stations
(SLAMS) where continued measurements of CO using FRM are required until discontinuation is
approved by the EPA Regional Administrator. Where SLAMS CO monitoring is required, at least one of
the sites must be a max concentration site for that specific area. In 2007, there were -385 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 is often
done with a monitor situated at the CFR-defined microscale distance (2-10 m) 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 15 m. For the
CFR-defined middle (up to 500 m) and neighborhood (-500 m-4 km) scale monitoring, the minimum
monitor distance from a major roadway is inversely 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 middle scale CO monitor from a roadway with an average daily traffic count
of 50,000 vehicles per day is 135 m. More detail on siting requirements can be found in 40 CFR Part 58
Appendices D through E.
3.4.2.2. Spatial and Temporal Coverage
Figure 3-6 depicts the distribution of the -385 regulatory CO monitors operating in the U.S. in
2007. Data from 285 of the -385 CO monitors operating year-round in the years 2005-2007 met the 75%
data completeness criterion for inclusion in the multi-year ambient data analyses for this assessment. The
greatest density of monitors is in the CSAs for Los Angeles, CA and San Francisco, CA, and along the
Mid-Atlantic sea board. 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.
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
March 2009
3-12
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
region.1 As an example, Figures 3-7 through 3-10 display CO monitor density with respect to population
density (for total population and elderly adults aged 65 and over) for Phoenix and Pittsburgh. (Annex A
includes analogous plots for the other nine metropolitan regions.)
Although ambient monitors for CO and other criteria pollutants are explicitly not located in order
to monitor population exposures, it is instructive to test the utility of the current network for
characterizing this exposure. Tables 3-1 and 3-2 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. As concerns 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, whereas 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. Figures 3-7 through 3-
10 show that multiple CO monitors in Phoenix and Pittsburgh were in the city center, and that Pittsburgh
also had monitors in outlying areas of moderate to low population density. The CO monitors in Phoenix
appear to provide good coverage of areas of the highest total population density, while areas of high
elderly population density were located at greater distances from the monitors. CO monitors in Pittsburgh
appear to sample areas of high population density well for both the total and elderly populations.
Figures 3-13 and 3-15 in section 3.5 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 most
cases, the monitors were concentrated near the center of the CSA/CBSA. 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 CBS As tied by
specific commuting behaviors. The broader CSA definition was used when selecting monitors for the cities listed above with the exception of
Anchorage and Phoenix, which are not contained within a CSA. Therefore, the smaller CBSA definition was used for these metropolitan areas.
March 2009
3-13
DRAFT - DO NOT CITE OR QUOTE
-------
i CO Monitors - included in analysis, 2007
CO Monitors - all, 2007
I CSA/CBSA
Figure 3-6. Map of ~385 CO monitor locations in the U.S. in 2007. Locations are indicated with
triangles: filled triangles show locations of the 285 monitors used in data analysis
for this assessment; open triangles are at locations with monitors which did not
meet the data completeness requirements for analysis; blue lines mark the
boundaries of the 11 CSAs/CBSAs used in the data analysis for this assessment
March 2009
3-14
DRAFT - DO NOT CITE OR QUOTE
-------
Phoenix Core Based Statisical Area
i Kilometers
0 5 10 20 30 40
Q/
2005 Population Density
| Phoenix CO Monitors (5 km)
Population per 2.6 Sq Km
| 0-206
| 207 - 412
413 -2D60
2061 -4120
4121 - 10300
I 10301 -41200
i Kilometers
0 20 40 80 120 160
Figure 3-7. Map of CO monitor locations with respect to population density in the Phoenix, AZ
CBSA, total population.
March 2009
3-15
DRAFT - DO NOT CITE OR QUOTE
-------
Phoenix Core Based Statisical Area
1
ft.
Kilometers
0 5 10 20 30 40
Qt
2005 Population Density
| Phoenix CO Monitors (5 km)
Population > 65 per 2.6 Sq Km
| 0-56
| 57-111
112-557
556 - 1114
1115 - 2785
¦ 2786-11140
i Kilometers
0 20 40 80 120 160
Figure 3-8. Map of CO monitor locations with respect to population density in the Phoenix, AZ
CBSA, age 65 and older.
March 2009
3-16
DRAFT - DO NOT CITE OR QUOTE
-------
Pittsburgh Combined Statisical Area
¦ Kilometers
0 5 10 20 30 40
A
d:
2005 Population Density
| | Pittsburgh CO Monitors (5 km buffer)
Population per 2.6 Sq Km
| 16 - 530
531-105©
1060 - 5296
5297 - 10593
1D5S4 - 26481
I 26432 - 105925
i Kilometers
0 15 30 60 90 120
Figure 3-9. Map of CO monitor locations with respect to population density in the Pittsburgh,
PA CSA, total population.
March 2009 3-17 DRAFT-DO NOT CITE OR QUOTE
-------
Pittsburgh Combined Statisical Area
V
i Kilometers
Figure 3-10. Map of CO monitor locations with respect to population density in the Pittsburgh,
PA CSA, age 65 and older.
2005 Population Density
| | Pittsburgh CO Monitors (5 km buffer)
Population > 65 per 2.6 Sq Km
| 2-84
= 65 - 129
130 - 643
644-1287
1288 - 3217
I 3218-12867
i Kilometers
0 5 10 20 30 40
March 2009
3-18
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-1. Proximity to CO monitors for the total population by city.
Region
Total CSA/
CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Anchorage, AK
352,225
5,391
1.53
131,608
37.36
212,834
60.43
239,842
68.09
Atlanta, GA
5,316,742
5,480
0.10
149,772
2.82
672,701
12.65
1,444,986
27.18
Boston, MA
7,502,707
95,732
1.28
1,180,054
15.73
2,432,846
32.43
3,418,353
45.56
Denver, CO
2,952,039
26,096
0.88
497,598
16.86
1,091,444
36.97
1,720,360
58.28
Houston, TX
5,503,320
2,9068
0.53
599,796
10.90
1,669,117
30.33
2,506,830
45.55
Los Angeles, CA
17,655,319
202,340
1.15
4,064,309
23.02
11,928,427
67.56
15,074,972
85.38
New York, NY
22,050,940
201,350
0.91
3,711,369
16.83
8,385,801
38.03
12,454,837
56.48
Phoenix, AZ
3,818,147
47,478
1.24
503,433
13.19
1,033,102
27.06
1,581,887
41.43
Pittsburgh, PA
2,515,383
29,136
1.16
369,965
14.71
895,252
35.59
1,359,596
54.05
Seattle, WA
3,962,434
4,814
0.12
94,649
2.39
279,976
7.07
699,490
17.65
St. Louis, MO
2,869,955
16,638
0.58
255,499
8.90
886,412
30.89
1,303,636
45.42
Table 3-2. Proximity to CO monitors for adults aged 65 and older by city.
Region
Total CSA/
CBSA
<1 km
<5 km
<10 km
< 15 km
N
N
%
N
%
N
%
N
%
Anchorage, AK
17,742
361
2.03
8,986 50.65
12,038
67.85
12,990
73.22
Atlanta, GA
362,201
423
0.12
12,758
3.52
54,148
14.95
111,232
30.71
Boston, MA
945,790
8,272
0.87
131,198
13.87
297,392
31.44
430,502
45.52
Denver, CO
232,974
2,541
1.09
42,760
18.35
102,783
44.12
163,682
70.26
Houston, TX
377,586
1,703
0.45
42,312
11.21
130,567
34.58
182,049
48.21
Los Angeles, CA
1,626,663
17,974
1.10
380,079
23.37
1,069,188
65.73
1,355,461
83.33
New York, NY
2,710,675
29,534
1.09
427,601
15.77
940,121
34.68
1,429,215
52.73
Phoenix, AZ
388,150
2,877
0.74
35,839
9.23
77,244
19.90
125,300
32.28
Pittsburgh, PA
449,544
5,383
1.20
66,967
14.90
166,440
37.02
255,220
56.77
Seattle, WA
390,372
556
0.14
12,142
3.11
3,1036
7.95
69,858
17.90
St. Louis, MO
358,747
3,203
0.89
42,890
11.96
127,274
35.48
184,491
51.43
March 2009
3-19
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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 can only be exceeded once per year at
a given location. Figures 3-11 and 3-12 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 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. 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, the level of the current 1-h NAAQS; see Figure 3-11. Moreover, only two monitored locations,
one in Weber Co., UT and the other in Jefferson Co., AL (including Birmingham, AL), reported second-
highest 1-h CO concentrations above 18 ppm, or approximately one-half the level of the 1-h standard.
Figure 3-12 shows that no counties reported second-highest 8-h CO concentrations above 9 ppm, the level
of the 8-h NAAQS. Only nine counties reported second-highest 8-h CO concentrations above 4.5 ppm:
Jefferson Co., AL, Kay Co., OK, Imperial Co., CA, Weber Co., UT, Philadelphia Co., PA, Anchorage
Municipality, AK, Los Angeles Co., CA, Jefferson Co., OH, and San Diego Co., CA.
March 2009
3-20
DRAFT - DO NOT CITE OR QUOTE
-------
Population
(Millions)
2nd highest 1-hour average concentration (ppm) in 2007
. No Data 0—17 U >17 -35
~ >35-47 Hi >47-56 >58
300
Carbon Monoxide 2nd highest 1—hour average concentration
200-
Figure 3-11. 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 who live in counties with CO
concentrations in the range indicated. Note that approximately 150 million people
live in counties with no CO monitors.
March 2009
3-21
DRAFT - DO NOT CITE OR QUOTE
-------
0-1—
Population
(Millions)
2nd highest 8—hour average concentration (ppm) in 2007
~ No Data I 0-4.5 >4.5-9
H >9-12 >12-15 >15
300-
Carbon Monoxide 2nd highest 8 —hour average concentration
Figure 3-12. County-level map of second-highest 8-h avg CO concentrations in the U.S. in 2907.
The bar on the left shows the total U.S. population who live in counties with CO
concentrations in the range indicated. Note that approximately 150 million people
live in counties with no CO monitors.
March 2009
3-22
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-3. Distribution of 1 -h avg CO concentration (ppm) derived from AQS data.
Percentiles
n
Mean
Mln
1
5
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
2.1
39.0
2005
2,391,962
0.5
0.0
0.0
0.0
0.1
0.2
0.4
0.6
1.0
1.3
2.3
22.3
2006
2,402,153
0.5
0.0
0.0
0.0
0.1
0.2
0.4
0.6
0.9
1.2
2.1
35.3
2007
2,386,585
0.4
0.0
0.0
0.0
0.1
0.2
0.3
0.5
0.8
1.1
1.9
39.0
Winter (December - February)
1,752,340
0.6
0.0
0.0
0.0
0.1
0.3
0.4
0.7
1.2
1.6
2.7
20.0
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
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
285
0.5
0.0
0.0
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.0
1.5
2005
285
0.5
0.0
0.0
0.1
0.2
0.4
0.5
0.6
0.8
0.9
1.3
1.6
2006
285
0.5
0.0
0.0
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.2
1.4
2007
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.7
1.1
1.5
Winter (December - February)
285
0.6
0.0
0.0
0.2
0.2
0.4
0.5
0.7
0.9
1.1
1.5
1.6
Spring (March - May)
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.7
1.0
1.6
Summer (June - August)
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1.1
1.5
Fall (September - November)
285
0.5
0.0
0.0
0.1
0.2
0.4
0.4
0.6
0.8
0.9
1.1
1.5
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007j (N
= NUMBER OF OBSERVATIONS)
Anchorage3
25,672
1.1
0.0
0.1
0.2
0.3
0.5
0.7
1.3
2.3
3.1
5.0
13.1
Atlanta
76,683
0.5
0.0
0.0
0.2
0.2
0.3
0.4
0.6
0.8
1.1
1.6
10.8
Boston
171,975
0.4
0.0
0.0
0.0
0.1
0.2
0.4
0.5
0.7
0.9
1.4
10.0
Denver
129,038
0.5
0.0
0.0
0.1
0.2
0.3
0.4
0.6
1.0
1.3
2.2
9.3
Flouston
123,925
0.3
0.0
0.0
0.0
0.0
0.2
0.3
0.4
0.6
0.8
1.4
4.6
Los Angeles
592,960
0.5
0.0
0.0
0.0
0.1
0.2
0.3
0.6
1.0
1.4
2.3
8.4
New York
226,673
0.5
0.0
0.0
0.1
0.1
0.3
0.5
0.6
0.9
1.1
1.6
5.8
Phoenix
127,477
0.8
0.0
0.0
0.1
0.2
0.3
0.5
1.0
1.9
2.5
3.6
7.8
Pittsburgh
179,758
0.3
0.0
0.0
0.0
0.0
0.1
0.2
0.4
0.6
0.8
1.2
6.7
Seattle
25,818
0.8
0.0
0.1
0.2
0.3
0.4
0.6
0.9
1.3
1.6
2.5
5.9
St. Louis
77,142
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.9
1.4
5.7
Not in the 11 cities
5,449,251
0.5
0.0
0.0
0.0
0.1
0.2
0.4
0.6
0.9
1.2
2.1
39.0
!C0 monitoring is only available for quarters 1 and 4; since monitoring data is not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-3 contains the distribution of hourly CO measurements reported to AQS for 2005-2007. All
2 monitoring locations meeting the 75% data completeness criteria have been included in this table.
3 Anchorage did not meet the data completeness criteria since its monitoring sites were required to report
4 CO measurements during the first and fourth quarters of each year. Anchorage was included in the table,
5 however, for an approximate comparison with the other CSAs and CBSAs reporting year-round
6 measurements to AQS. (Anchorage was not, however, included in the national averages shown in the
March 2009
3-23
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
table.) 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-year period. The largest recorded second-highest 1-h concentration, 26.3 ppm, for this
period was reported in 2006 in Birmingham, AL (AQS site ID: 0107360004). The absolute highest 1-h
concentration, 39 ppm, between 2005 and 2007, was reported in Ogden, UT (AQS site ID: 490570006).
Concentrations 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 of Table 3-3 and illustrate the
distribution of the site average CO concentrations recorded at the 285 monitoring sites for 2005-2007 (see
Figure 3-6 for these sites). The site reporting the highest 3-year pooled 1-h avg CO concentration,
1.5 ppm, was located in San Juan, Puerto Rico (AQS site ID: 721270003). The eleven individual
CSAs/CBSAs discussed earlier are included in the table, none of which reported concentrations above the
value of the 1-h NAAQS. Four of the eleven 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.
March 2009
3-24
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-4. Distribution of 24-h avg CO concentration (ppm) derived from AQS data.
Percentiles
n
Mean
Min
1
5
10
25
50
75
90
95
99
Max
NATIONWIDE STATISTICS (N =
NUMBER OF OBSERVATIONS)
2005-2007
303,843
0.5
0.0
0.0
0.0
0.1
0.3
0.4
0.6
0.9
1.1
1.7
7.0
2005
101,184
0.5
0.0
0.0
0.0
0.1
0.3
0.4
0.6
0.9
1.1
1.8
5.8
2006
101,652
0.5
0.0
0.0
0.0
0.1
0.3
0.4
0.6
0.9
1.1
1.6
7.0
2007
101,007
0.4
0.0
0.0
0.0
0.1
0.2
0.4
0.5
0.8
1.0
1.6
6.9
Winter (December - February)
74,144
0.6
0.0
0.0
0.1
0.2
0.3
0.5
0.7
1.1
1.3
2.0
7.0
Spring (March - May)
77,317
0.4
0.0
0.0
0.0
0.1
0.2
0.4
0.5
0.7
0.9
1.4
6.4
Summer (June - August)
76,562
0.4
0.0
0.0
0.0
0.1
0.2
0.3
0.5
0.7
0.8
1.3
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
285
0.5
0.0
0.0
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.0
1.5
2005
285
0.5
0.0
0.0
0.1
0.2
0.4
0.5
0.6
0.8
0.9
1.3
1.6
2006
285
0.5
0.0
0.0
0.1
0.2
0.3
0.4
0.6
0.7
0.8
1.2
1.4
2007
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.7
1.1
1.5
Winter (December - February)
285
0.6
0.0
0.0
0.2
0.2
0.4
0.5
0.7
0.9
1.1
1.5
1.6
Spring (March - May)
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.7
1.0
1.6
Summer (June - August)
285
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1.1
1.5
Fall (September - November)
285
0.5
0.0
0.0
0.1
0.2
0.4
0.4
0.6
0.8
0.9
1.1
1.5
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007j (N
= NUMBER OF OBSERVATIONS)
Anchorage3
1,074
1.1
0.0
0.2
0.2
0.4
0.6
0.9
1.4
1.9
2.4
3.3
4.6
Atlanta
3,229
0.5
0.0
0.1
0.2
0.2
0.3
0.4
0.6
0.8
0.9
1.2
1.6
Boston
7,446
0.4
0.0
0.0
0.1
0.1
0.3
0.4
0.5
0.7
0.8
1.1
2.2
Denver
5,363
0.5
0.0
0.1
0.2
0.2
0.3
0.5
0.6
0.9
1.1
1.5
2.3
Flouston
5,188
0.3
0.0
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.9
1.9
Los Angeles
25,803
0.5
0.0
0.0
0.1
0.1
0.2
0.4
0.6
1.0
1.2
1.7
3.8
New York
9,513
5.1
0.0
0.0
0.1
0.2
0.4
0.5
0.6
0.8
1.0
1.3
2.5
Phoenix
5,348
0.8
0.0
0.1
0.2
0.3
0.4
0.6
1.1
1.6
1.9
2.5
3.4
Pittsburgh
7,497
0.3
0.0
0.0
0.0
0.0
0.1
0.2
0.4
0.6
0.7
1.0
1.9
Seattle
1,079
0.8
0.1
0.2
0.3
0.4
0.5
0.7
0.9
1.2
1.4
1.8
2.4
St. Louis
3,216
0.4
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.7
0.8
1.0
1.9
Not in the 11 cities
230,161
0.5
0.0
0.0
0.0
0.1
0.2
0.4
0.6
0.8
1.1
1.6
7.0
*C0 monitoring is only available for quarters 1 and 4; since monitoring data is not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-4 contains the distribution of 24-h avg CO concentrations derived from the 1-h
2 concentrations reported to AQS and summarized in Table 3-3. The nationwide mean, median, and
3 interquartile range for 24-h avg for 2005-2007 were 0.5, 0.4 and 0.3 ppm, respectively. These were
4 similar to those for the 1-h values. The maximum 24-h avg concentration in these years, 7 ppm, was
5 reported in Birmingham, AL (AQS site ID: 010736004).
March 2009
3-25
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-5. Distribution of 1 -h daily max CO concentration (ppm) derived from AQS data.
Percentiles
n
Mean
Mln
1
5
10
25
50
75
90
95
99
Max
NATIONWIDE STATISTICS (N
NUMBER OF OBSERVATIONS)
2005-2007
303,843
0.9
0.0
0.0
0.1
0.3
0.4
0.7
1.2
1.8
2.4
3.8
39.0
2005
101,184
1.0
0.0
0.0
0.2
0.3
0.5
0.8
1.3
2.0
2.6
4.1
22.3
2006
101,652
0.9
0.0
0.0
0.1
0.3
0.4
0.7
1.2
1.9
2.4
3.9
35.3
2007
101,007
0.8
0.0
0.0
0.1
0.2
0.4
0.7
1.1
1.7
2.1
3.4
39.0
Winter (December - February)
74,144
1.2
0.0
0.0
0.2
0.3
0.5
0.9
1.6
2.5
3.1
4.7
20.0
Spring (March - May)
77,317
0.8
0.0
0.0
0.1
0.3
0.4
0.7
1.0
1.6
2.0
3.0
35.3
Summer (June - August)
76,562
0.7
0.0
0.0
0.1
0.2
0.4
0.6
0.9
1.3
1.6
2.5
39.0
Fall (September - November)
75,820
1.0
0.0
0.0
0.2
0.3
0.5
0.8
1.3
2.0
2.5
3.8
24.1
NATIONWIDE STATISTICS, POOLED BY SITE (N
NUMBER OF SITES)
2005-2007
285
0.9
0.1
0.1
0.3
0.5
0.6
0.8
1.1
1.5
1.7
2.3
3.9
2005
285
1.0
0.1
0.1
0.4
0.5
0.7
0.9
1.2
1.6
2.0
2.5
3.7
2006
285
0.9
0.1
0.1
0.3
0.5
0.6
0.9
1.1
1.6
1.8
2.3
4.8
2007
285
0.8
0.1
0.1
0.3
0.4
0.6
0.8
1.0
1.4
1.6
2.0
3.1
Winter (December - February)
285
1.2
0.0
0.1
0.4
0.6
0.8
1.0
1.5
2.1
2.5
3.4
4.1
Spring (March - May)
285
0.8
0.1
0.1
0.3
0.4
0.6
0.8
1.0
1.3
1.5
2.1
4.0
Summer (June - August)
285
0.7
0.0
0.1
0.2
0.3
0.5
0.6
0.8
1.1
1.3
2.2
3.3
Fall (September - November)
285
1.0
0.1
0.1
0.3
0.5
0.7
0.9
1.2
1.7
2.0
2.4
4.1
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007j (N =
NUMBER OF OBSERVATIONS)
Anchorage3
1,074
2.6
0.0
0.3
0.6
0.8
1.3
2.2
3.5
5.0
6.1
7.6
13.1
Atlanta
3,229
0.8
0.0
0.2
0.3
0.3
0.4
0.7
1.1
1.4
1.7
2.2
10.8
Boston
7,446
0.7
0.0
0.1
0.2
0.3
0.4
0.6
0.9
1.2
1.6
2.6
10.0
Denver
5,363
1.2
0.1
0.2
0.4
0.5
0.7
1.0
1.5
2.2
2.7
3.9
9.3
Flouston
5,188
0.7
0.0
0.0
0.1
0.2
0.4
0.6
0.8
1.3
1.7
2.6
4.6
Los Angeles
25,803
1.0
0.0
0.1
0.2
0.3
0.5
0.8
1.3
2.0
2.6
4.0
8.4
New York
9,513
0.9
0.0
0.1
0.2
0.4
0.6
0.8
1.1
1.5
1.8
2.5
5.8
Phoenix
5,348
1.9
0.0
0.3
0.5
0.6
0.9
1.6
2.5
3.5
4.1
5.3
7.8
Pittsburgh
7,497
0.6
0.0
0.0
0.0
0.1
0.2
0.5
0.8
1.1
1.4
2.0
6.7
Seattle
1,079
1.5
0.2
0.4
0.5
0.7
0.9
1.3
1.8
2.4
2.9
4.3
5.9
St. Louis
3,216
0.8
0.0
0.1
0.3
0.4
0.5
0.6
0.9
1.3
1.7
2.7
5.7
Not in the 11 cities
230,161
0.9
0.0
0.0
0.1
0.2
0.4
0.7
1.2
1.8
2.4
3.8
39.0
*C0 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 3-5 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.
March 2009
3-26
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-6. 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
303,843
0.7
0.0
0.3
0.3
0.3
0.3
0.5
0.8
1.3
1.7
2.6
10.9
2005
101,184
0.7
0.0
0.3
0.3
0.3
0.3
0.6
0.9
1.4
1.8
2.8
9.7
2006
101,652
0.7
0.0
0.3
0.3
0.3
0.3
0.5
0.8
1.3
1.7
2.6
9.8
2007
101,007
0.6
0.0
0.3
0.3
0.3
0.3
0.5
0.8
1.2
1.5
2.3
10.9
Winter (December - February)
74,144
0.9
0.0
0.3
0.3
0.3
0.4
0.7
1.1
1.7
2.1
3.2
9.8
Spring (March - May)
77,317
0.6
0.0
0.3
0.3
0.3
0.3
0.5
0.7
1.1
1.3
2.0
9.6
Summer (June - August)
76,562
0.5
0.0
0.3
0.3
0.3
0.3
0.4
0.6
0.9
1.1
1.7
10.9
Fall (September - November)
75,820
0.7
0.0
0.3
0.3
0.3
0.3
0.6
0.9
1.4
1.8
2.7
9.0
NATIONWIDE STATISTICS, POOLED BY SITE (N = NUMBER OF SITES)
2005-2007
285
0.7
0.2
0.3
0.3
0.4
0.5
0.6
0.8
1.0
1.2
1.7
2.1
2005
285
0.7
0.3
0.3
0.3
0.4
0.5
0.6
0.9
1.1
1.4
1.9
2.2
2006
285
0.7
0.2
0.3
0.3
0.4
0.5
0.6
0.8
1.1
1.2
1.8
2.4
2007
285
0.6
0.2
0.3
0.3
0.4
0.5
0.6
0.7
1.0
1.1
1.6
2.0
Winter (December - February)
285
0.9
0.2
0.3
0.4
0.4
0.6
0.8
1.1
1.4
1.7
2.4
2.6
Spring (March - May)
285
0.6
0.2
0.3
0.3
0.4
0.4
0.5
0.7
0.9
1.1
1.6
2.2
Summer (June - August)
285
0.5
0.2
0.3
0.3
0.3
0.4
0.5
0.6
0.8
0.9
1.5
2.0
Fall (September - November)
285
0.7
0.2
0.3
0.3
0.4
0.5
0.6
0.9
1.2
1.3
1.8
2.2
STATISTICS FOR INDIVIDUAL CSAS/CBSAS (2005-2007j (N =
NUMBER OF OBSERVATIONS)
Anchorage3
1,074
1.7
0.3
0.3
0.4
0.6
0.9
1.5
2.3
3.3
3.9
5.0
6.5
Atlanta
3,229
0.6
0.0
0.2
0.2
0.3
0.4
0.5
0.8
1.1
1.3
1.7
2.5
Boston
7,446
0.6
0.3
0.3
0.3
0.3
0.3
0.5
0.7
0.9
1.1
1.8
5.8
Denver
5,363
0.8
0.3
0.3
0.3
0.3
0.5
0.7
1.0
1.4
1.8
2.4
3.4
Flouston
5,188
0.5
0.3
0.3
0.3
0.3
0.3
0.4
0.6
0.9
1.1
1.7
3.3
Los Angeles
25,803
0.7
0.3
0.3
0.3
0.3
0.3
0.6
0.9
1.5
1.8
2.7
6.2
New York
9,513
0.7
0.3
0.3
0.3
0.3
0.4
0.6
0.9
1.2
1.4
1.8
3.0
Phoenix
5,348
1.3
0.3
0.3
0.3
0.4
0.6
1.0
1.8
2.5
3.0
3.8
5.8
Pittsburgh
7,497
0.5
0.3
0.3
0.3
0.3
0.3
0.3
0.6
0.9
1.0
1.5
3.7
Seattle
1,079
1.1
0.3
0.3
0.4
0.5
0.7
1.0
1.4
1.8
2.2
3.2
4.0
St. Louis
3,216
0.6
0.3
0.3
0.3
0.3
0.3
0.5
0.7
0.9
1.2
1.9
4.2
Not in the 11 cities
230,161
0.7
0.0
0.3
0.3
0.3
0.3
0.5
0.8
1.3
1.6
2.5
10.9
!C0 monitoring is only available for quarters 1 and 4; since monitoring data is not available year-round, Anchorage is not included in the nationwide statistics shown in this table.
1 Table 3-6 contains the distribution of 8-h daily max concentrations derived from the 1-h CO
2 concentrations reported to AQS. This was done by first calculating the average concentration for each
3 successive 8-h period, thereby producing 24 8-h avg per day. The maximum of these values for a given
4 monitor within a given day (midnight-to-midnight) was used as the 8-h daily max statistic for that
5 monitor and day. The nationwide mean, median, and interquartile range for 8-h daily max concentrations
March 2009
3-27
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
reported for 2005-2007 were 0.7, 0.5, and 0.5 ppm, respectively. The highest 8-h daily max concentration,
10.9 ppm, was recorded at a monitor located five miles north of Newkirk, OK (AQS site ID: 400719010).
3.5.1.2. Urban Scale
This section covers urban variability in CO concentrations reported to AQS at the individual
CSA/CBSA level. Phoenix, AZ and Pittsburgh, PA 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 Phoenix CBSA and
Pittsburgh CSA shown in Figures 3-13 and 3-15, respectively, illustrate the location of all CO monitors
included in the analyses described above. Letters on the maps identify the individual monitor locations
and correspond with the letters provided in the accompanying box plots (Figures 3-14 and 3-16) and pair-
wise comparison tables (Tables 3-7 and 3-8). 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:
where Xij and Xik represent the observed hourly concentrations for time period i at sites j and k, andp 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. Similar maps, box plots and comparison tables for the nine remaining CSAs/CBSAs
described earlier in this chapter along with Phoenix and Pittsburgh are included in Annex A.
March 2009 3-28 DRAFT-DO NOT CITE OR QUOTE
Equation 3-1
-------
Phoenix Core Based Statistical Area
A
o)
• Phoenix CO Monitors
Phoenix Major Highways
Phoenix
0 15 30 60 90
120
I Kilometers
Figure 3-13. Map of CO monitor locations with AQS Site IDs for Phoenix, AZ,
March 2009
3-29
DRAFT - DO NOT CITE OR QUOTE
-------
A
B
C
D
E
Mean
0.93
0.84
0.58
0.76
0.79
Obs
25382
25589
25657
25414
25435
SD
0.95
0.88
0.64
0.72
0.64
£
Q_
Q.
C
o
c
-------
Table 3-7. Table of inter-sampler comparison statistics, as defined in the text, including
Pearson r, P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to
AQS for 2005-2007 in Phoenix, AZ.
Phoenix
A B C D E
1.00 0.89
0.80
0.86
0.84
0.0 0.7
1.1
0.8
0.9
0.0 0.37
0.43
0.39
0.37
0.0 1.6
8.9
3.9
3.5
1.00
0.81
0.88
0.89
0.0
0.9
0.6
0.7
0.0
0.38
0.34
0.24
0.0
9.4
3.4
4.9
1.00
0.81
0.85
0.0
0.7
0.6
0.0
0.41
0.36
0.0
6.6
6.8
1.00
0.83
0.0
0.6
0.0
0.33
Legend
0.0
5.2
r
1.00
P90
0.0
COD
0.0
d
0.0
March 2009
3-31
DRAFT - DO NOT CITE OR QUOTE
-------
Pittsburgh Combined Statistical Area
A
oi
A C
M «#
B
J
-a
X-^
Pittsburgh CO Monitors
Pittsburgh Major Highways
Pittsburgh
0 10 20
40
60
80
I Kilometers
Figure 3-15. Map of CO monitor locations with AQS Site IDs for Pittsburgh, PA.
March 2009
3-32
DRAFT - DO NOT CITE OR QUOTE
-------
Mean
Obs
SD
1.4 -
1.3-
1.2 -
1.1 -
1.0-
"E 0.9 -
Q.
S 0.8 -
I 0.7 -
CD
¦E 0.6-
-------
Table 3-8. Table of inter-sampler comparison statistics, as defined in the text, including
Pearson r, P90 (ppm), COD and d (km) for each pair of hourly CO monitors reporting to
AQS for 2005-2007 in Pittsburgh, PA.
Pittsburgh
A
B
c
D
E
F
G
>
O
O
0.52
0.43
0.05
0.11
0.42
0.18
0.0
0.5
0.6
0.7
0.6
0.5
0.7
0.0
0.69
0.68
0.74
0.72
0.73
0.88
0.0
2.2
1.8
41.9
66.8
34.4
45.8
B
1.00
0.73
0.30
0.42
0.29
0.35
0.00
0.40
0.50
0.40
0.50
0.50
0.00
0.39
0.54
0.51
0.62
0.86
0.00
0.70
43.40
68.00
33.60
43.70
C
1.00
0.20
0.39
0.25
0.30
0.0
0.7
0.6
0.7
0.8
0.0
0.56
0.51
0.65
0.88
0.0
43.4
68.2
33.3
44.1
D
1.00
0.16
0.09
0.02
0.0
0.6
0.6
0.7
0.0
0.57
0.69
0.87
0.0
27.5
75.0
84.1
E
1.00
0.0
0.0
0.0
0.26
0.5
0.68
101.0
0.21
0.6
0.87
104.1
F 1.00 0.09
0.0 0.5
0.0 0.90
Legend
0.0
37.8
r
1.00
P90
0.0
COD
0.0
d
0.0
1
2 The Phoenix CBSA in Figure 3-13 incorporates an area of 37,786 km2 with the CO monitors
3 densely packed near the urban center. The maximum straight-line distance between CO monitors in the
4 Phoenix CBSA is 9.4 km. By contrast, the Pittsburgh CSA in Figure 3-15 is less than half the size, having
5 an area of 14,627 km2, but the monitors are farther apart with a maximum straight-line distance between
6 them of 104.1 km. Of the eleven CSAs/CBSAs investigated, Phoenix has among the highest correlations
7 across monitors, while Pittsburgh has among the lowest. This discrepancy is partly due to the close
8 proximity of the monitors as noted here, but horizontal distance alone does not explain the range of
9 correlations observed between monitors. For example, the three monitors located in the downtown urban
10 core of Pittsburgh (sites A, B and C; AQS site IDs 420030010, 420030031 and 420030038; maximum
March 2009
3-34
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
distance = 2.2 km) exhibited considerably lower correlations (0.43 < r < 0.73) than any of the monitor
pairs in Phoenix (0.80 < r < 0.89), located up to 9.4 km apart.
The correlation structures for measurements at the monitors in each of the eleven 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. 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. Compare, for example, Figures 3-17 and 3-18,
the aerial views of monitors C (AQS site ID 040133002) and E (AQS site ID 040139997), respectively, in
Phoenix. Monitor C has a correlation value with monitor E of 0.85 even though C is less than 1 km from
the intersection of two major highways and E is surrounded by neighborhood residences. Because all
other correlations of monitors in Phoenix are also >0.8, the network of CO monitors there appears to have
captured the behavior of emitted and transported CO with high fidelity. In contrast, monitors in Pittsburgh
show a low degree of correlation, with only two values >0.5 and eight values < 0.2. The topography of
Pittsburgh is mountainous and therefore more complex than that of Phoenix, and this could account for
some of the disparity across space. But in addition, several CO monitors in Pittsburgh are sited at
distances far from the urban core, making them less likely to correlate with those monitors in the high-
density urban core. Figure 3-19 depicts the three monitors in the Pittsburgh urban core (monitor A: AQS
site ID 420030010, monitor B: AQS site ID 420030031, and monitor C: AQS site ID 420030038), which
demonstrated the highest correlation values in this CSA for 2005-2007. Figure 3-20 shows the location of
a more distant monitor, monitor F (AQS site ID 421250005), which does not correlate well with the three
monitors located downtown, very likely because it is located at a great distance from any strong CO
sources in the downtown urban core. This analysis demonstrates that agreement between monitors on an
urban scale is a complex function of monitor siting, location to sources, geography, and
micrometeorology.
In addition to high correlations, Phoenix exhibits similar average concentration levels across
monitors with 0.24 < COD < 0.43; see Table 3-7. The seasonal patterns shown in Figure 3-14 also
demonstrate homogeneity across monitors for Phoenix. By contrast, Pittsburgh had more variability
between monitored values at the sites shown in Figure 3-16 with the range in COD shifted upward
(0.39 < COD < 0.90); see Table 3-8. For the three sites located in the downtown urban core of Pittsburgh
(sites A, B and C), the COD is as high as 0.69, indicating greater spatial variability in CO concentrations
in Pittsburgh, even for these more proximal monitors.
The information in Table 3-8 should be used with some caution since many of the reported
concentrations for the years 2005-2007 are very near or even well below the monitors' stated lowest
detection limits. Because ambient concentrations are now in large part very near to the FRM detection
limit of 1.0 ppm and the coarsely reported measurement resolution is 0.1 ppm, the comparison statistics
shown in these tables might be biased to show a specious heterogeneity in the box plots. These cautions
March 2009
3-35
DRAFT - DO NOT CITE OR QUOTE
-------
1 are especially important for interpreting data from site G in Pittsburgh and selected other sites in other
2 metropolitan regions where one or more monitors consistently reported values below the stated detection
3 limit. Such monitors include Houston site D, AQS site ID 482011035, and New York site I, AQS site ID
4 361030009, included in Annex A.
Figure 3-17. Aerial view of the location of CO monitor C. AQS ID 040133002 (marked by the
yellow pin) in Phoenix, AZ, depicting its proximity to major roadways and
neighborhood residences.
March 2009
3-36
DRAFT - DO NOT CITE OR QUOTE
-------
ASSwfejrJ f,
Figure 3-18. Aerial view of the location of CO monitor E. AQS ID 040139997 (marked by the
yellow pin) in Phoenix, AZ, depicting its proximity to neighborhood residences and
secondary surface roads.
March 2009
3-37
DRAFT - DO NOT CITE OR QUOTE
-------
Figure 3-19. Aerial view of the location of CO monitors A, B, and C.AQS IDs 420030010,
420030031, and 420030038 (marked by yellow pins) in Pittsburgh, PA, depicting their
proximity to major roadways and areas of high commercial density.
March 2009
3-38
DRAFT - DO NOT CITE OR QUOTE
-------
Figure 3-20. Aerial view of the location of CO monitor F. AQS ID 421250005 (marked by the yellow
pin) in the Pittsburgh, PA CSA, depicting its proximity to secondary surface roads
and water ways.
3.5.1.3. Neighborhood Scale
1 Roadway density is an important determinant of the spatial distribution of CO concentrations in
2 urban areas. Mobile sources are the largest single source of CO, and their abundance and density affect
3 the magnitude of CO production. Urban topography around roadways also affects CO transport and
4 dispersion (e.g., Mfula et al., 2005; Rigby and Toumi, 2008). CO concentration gradients have been
5 observed in near-road environments (Baldauf et al., 2008; Pirjola et al,, 2006; Zhu et al., 2002) such that
6 CO concentrations are often higher on or near roadways compared with upwind sites, and can decrease
7 with increasing distance from the road. For example, Zhu et al. (2002) found that on-road CO
8 concentrations were approximately 10 tunes higher than at an upwind monitoring site, as shown m Figure
9 3-21. At 30 m downwind from the road, relative CO concentrations were decreased to 7 times above
10 upwind levels, and were approximately 2.5 times the upwind levels at 100 m. Concentrations continued to
March 2009
3-39
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
decrease and were still somewhat higher than upwind levels at the final monitoring site 300 m away.
Other traffic-related pollutants (BC, particle number) followed a similar pattern, although with different
downwind:upwind ratios. This indicates that near-road and on-road exposures are important in CO
exposure assessment.
Field measurements, computational modeling, and wind tunnel experiments have shown that
roadway design and roadside features can affect CO and other pollutant concentrations near roadways.
Field measurements reported by Baldauf et al. (2008) indicated that noise barriers could reduce near-road
pollutant concentrations by as much as 50 percent, although this effect was highly dependent on
meteorological conditons. This study also showed that the presence of mature vegetation, where the noise
barrier further reduced concentrations and flattened the concentration gradient away from the road. Urban
dispersion and wind-field modeling by Bowker et al. (2007) also demonstrated the influence of noise
barriers and vegetation on the concentrations and spatial variability of inert pollutants emitted from traffic
sources. Wind tunnel work reported by Baldauf et al. (In press) demonstrated the effects of noise barriers
as well as roadway design characteristics, such as the presence of cut and elevated roadway segments.
These results indicated that cut sections reduced concentrations and altered the gradient away from the
road for inert gases emitted from traffic sources such as CO. These results showed similar concentrations
as Zhu et al. (2002) for roadway segments at-grade with no obstructions to air flow as well as elevated
roadway segments with fill conditions.
The geometry of urban street canyons has a profound effect on the distribution of CO
concentrations on a micro-scale. 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; Chang and Meroney, 2003; Kastner-Klein and
Plate, 1999; So et al., 2005; Xiaomin et al., 2006). 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. Influential parameters include canyon height to width ratio (H/W), source
positioning, wind speed and direction, building shape, and upstream configuration of buildings.
Figure 3-22 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). 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. 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.
March 2009
3-40
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
0 8
e
c
5
u
0.6
1
9
U
> 04
Particle Number
Black Carbon
02
O—
CO
0.0 k . .
Upwind -200
-100
0 100
Distance to the 710 Freeway (m)
200
300 Downwind
Source: Zhu et al. (2002)
Figure 3-21. Relative concentrations of CO and copollutants at various distances from the 710
freeway in Los Angeles.
In a multi-site survey of curbside CO concentration in London, U.K., Croxford and Penn (1998)
observed differences in concentration related to the side of the street on which the monitor was positioned
relative to the wind direction. Bogo et al. (2001) measured CO concentrations in a street canyon with
building height-street width ratio of 1 in Buenos Aires, Argentina using a continuous CO monitor. Similar
to the Xiaomin et al. (2006) simulation results for a height-width ratio of 1, Bogo et al. (2001) 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.
For part of this study, they estimated aggregated emissions from several fixed sources, including thermal
power stations and home heating, and from traffic. Mobile source emissions were estimated to be more
than 800 times higher than fixed source emissions given assumptions regarding vehicle fleet. Moreover,
they estimated that up to 30% of the measured CO concentrations came from within the local street
canyon in which measurements were made.
March 2009
3-41
DRAFT - DO NOT CITE OR QUOTE
-------
^ windward,
measured
windward,
simulated
leeward,
measured
leeward,
simulated
0.8
0.8
0.6
0.6
A.
0.4
0.4
0.2
0.2
20
Dimensionless concentration
40
60
80
100
100 200 300 400 500 600
Dimensionless concentration
(b)
Source: Xiaomin et al. (2006)
Figure 3-22. Dimensionless tracer gas concentration as a function of elevation at windward and
leeward locations and street canyon aspect ratios (H/W). Z/H is the elevation of the
measurement (Z) scaled by building height (H). (a) Dimensionless concentration on
the windward and leeward sides of the canyon when H/W = 1 and wind speed = 3
m/s. (b) Dimensionless concentration on the windward and leeward sides of the
canyon when H/W = 2 and wind speed = 5 m/s. Computational fluid dynamics
modeling was performed, and measurements were obtained in wind tunnel
simulations.
1 Identification of neighborhood-scale variability is important for interpreting data from ambient CO
2 monitors. Figure 3-23 shows a scatter plot of CO correlation as a function of distance (Pearson r) for
3 monitors sited within 4 km of each other in four of the eleven CSAs and CBSAs. (The other seven CSAs
4 or CBSAs were excluded from this analysis because their monitors were spaced farther than 4 km from
5 each other.) The purpose of this plot is not to show a simple relationship between inter-sampler
6 correlation and distance but to illustrate how this relationship varies between cities at the neighborhood
7 scale. High correlation is seen for the monitors in the Phoenix, AZ CBSA, while those in the Pittsburgh.
8 PA, CSA have much lower inter-sampler correlations with steeper slope (Ar/Ad = -0.04 for Phoenix and
9 Ar/Ad = -0.27 for Pittsburgh). The inter-sampler correlation observed for the Boston, MA CSA is also
March 2009
3-42
DRAFT - DO NOT CITE OR QUOTE
-------
substantially lower than those observed for Phoenix. As discussed in section 3.5.1.2, these differences
could be due to a variety of factors, including but not limited to natural and urban topography, different
traffic conditions on roads located near monitors, and proximity of the monitors to the roadway.
0.0 0.5 1.0
O o
1.5 2.0 2.5
Distance (km)
A A
O Boston
~ Denver
A Phoenix
O Pittsburgh
3.0 3.5 4.0
Figure 3-23. Inter-sampler correlations as a function of distance between CO monitors for
samplers located within 4 km (neighborhood scale) for Boston, Denver, Phoenix,
and Pittsburgh.
March 2009
3-43
DRAFT - DO NOT CITE OR QUOTE
-------
3.5.2. Temporal Variability
3.5.2.1. Multi-year Trends
1 Figures 3-24 (top) shows ambient CO concentrations in ppm from 1980 to 2006 based on
2 continuous measurements averaged over 8-h time segments. The 8-h NAAQS is indicative of exposures
3 occurring over a sustained period of time, an outdoor worker's exposure over the course of a day. for
4 example. Figure 3-24 (bottom) depicts trends in the annual second-highest 8-h CO concentrations for 144
5 sites in 102 counties nationwide having data either in the State and Local Air Monitoring Stations
6 (SLAMS) network or from other special purpose monitors.
16r
90% of sites have concentrations below this line
kj
NAAQS = 9 ppm
v*jroGC
Median
10% of sites have
concentrations below this line
—i—i—i—i—r—
—i—i—i—i—i—i—i—
'80 '82 '84 *86
'90 '92 *94 '96
Year
•00 02 '04 '06
B0 "82 '84 '86
90 '92 '94
02 04 06
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 1980.
Source: U.S. EPA (2008c)
Figure 3-24. (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).
March 2009
3-44
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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
years; see Figure 3-24 (top). Since 1992, more than 90% of all these sites have reported CO
concentrations below the 8-h NAAQS of 9 ppm; see Figure 3-24. 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; see Figure 3-2. 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 CO concentrations
above the level of the CO NAAQS declined to zero over the same period; see Figure 3-24 (bottom).
Consistent with the nationwide trends in emissions and concentrations, CO concentrations in all ten
EPA Regions have steadily decreased since 1980, with reductions over this period ranging from 68% in
Region 7 to 85% in Region 1; see Figure 3-25. This is also consistent with declining emissions seen in
many regions of the U.S., shown in Figure 3-4. Reductions in anthropogenic CO emissions in the U.S.
and elsewhere in the extra-tropical northern hemisphere are a major cause for the observed decrease in
hemispheric and global-average CO concentrations observed since 1991 (Bakwin et al., 1994), although
-30% of the decline between 1991 and 2001 was attributed to decreases in CO following the eruption of
Mt. Pinatubo (Novelli et al., 2003).
March 2009
3-45 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
NAAQS = 9 ppm
-R1
— R2
-R3
-R4
— R5
- R6
R7
R8
-R9
— R10
—Nat'I
'80 '82 '84 '86 '
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.
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
EPA Regions
©
o
> a
® €»
O
©
©¦
Source: U.S. EPA (2008c)
Figure 3-25. 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
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 Figures 3-26 and 3-27, respectively among the
eleven CSAs and CBSAs examined in this assessment. The weekday data showed that the Anchorage
mean, median, 5th and 10th percentile CO concentration curves exhibit more pronounced morning and
evening rash 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 rash
hour and a slight trough between 1:00 and 5:00 AM. The Seattle plot shows a daytime plateau beginning
around 5:00 AM and lasting until roughly 10:00 PM, with higher concentrations dunng morning and
afternoon rush hour. Differences in hourly variation among the eleven CSAs and CBSAs reflect city-to-
March 2009
3-46
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
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 3-27 illustrates weekend diel trends for the eleven 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 shows that the mean and median CO concentrations remain
fairly constant throughout the day, with a slight reduction between 2:00 and 7:00 AM. The 90th and 95th
percentile curves illustrate more diel variation.
March 2009
3-47
DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage (N = 1270)
Atlanta (N = 2317)
Boston (N = 2488)
5.2 -i
3.9 -
E
Q_
2.6 -
o
o
1.3 -
16 12 18 24
Denver (N = 4592)
2.8
2.1
1.4
0.7
0
2.8 -]
2.8 -j
2.1 -
2.1 -
E
Q.
/\
5 1.4-
- <¦ 1.4-
1 * / X
o
1 » '¦ / N
, \
U 0.7 -
1 ,
1
' /
~
1
\
O
~«4
i
0 -
n
\
i
i
i
i
i
o
16 12 18 24
New York City (N =4651)
1
2.8
^ 2.1
E
Q_
-B 1-4
O
<¦> 0.7
0
2.8
^ 2.1
E
Q_
B 1.4
o
<¦> 0.7
0
6 12 18 24
Seattle (N = 1577)
2.8 !
2.1
1.4
0.7
0
6 12 18 24
Houston (N = 4212)
I !¦
I <
I 1
1
2.8 t
2.1 -
1.4 -
0.7-
2.8
2.1
1.4 -
0.7
0
-v-
16 12 18 24
Los Angeles (N = 12846)
6 12 18 24
Phoenix (N = 1021)
2.8
2.1
1.4
0.7
0
1
2.8 ~j
2.1 -
1.4 -
0.7
6 12 18 24
St. Louis (N = 2917)
6 12 18 24
Pittsburgh (N = 3842)
12 18 24
KEY
Median
Mean
90'th & 10'th
95'th & 5'th
12 18 24 1 6 12 18 24
Figure 3-26. Diel plot generated from weekday hourly CO data (ppm) for the eleven CSAs and
CBSAs 2005-2007, Included are the number of monitor days (N) and the median,
mean, 5th, 10th, 90th and 95th percentiles of CO concentration. Note that the y-axis
of the Anchorage CBSA plot is scaled to 5.2 ppm while the other plots are scaled to
2.8 ppm.
March 2009
3-48
DRAFT - DO NOT CITE OR QUOTE
-------
Anchorage (N = 505)
5.2 i
3.9 -
E
Q_
S 2.6 -
o
o 1.3
2.8
2.1
E
Q.
Q.
1.4 -
O
o
0.7 -
2.8 i
2.1 -
E
Q.
a 1.4-
O
o 0.7 -
6 12 18 24
Denver (N = 1897)
2.8
2.1
1.4
0.7
0
1
2.8 -
2.8 t
^ 2.1 -
2.1 -
E
Q_
3 14 "
1.4 -
o
> ~
^ 0.7 -
v /v /
°-7 ""
0 -
1 , 1 1 0 -¦
6 12 18 24
New York City (N = 1876)
16 12 18 24
Seattle (N = 639)
1
2.8
2.1
1.4
0.7
0
1
2.8
2.1
1.4
0.7
0
Atlanta (N = 932)
2.8 -j
2.1 -
1.4 -
07
i * i '-"i 11 i 0 --
6 12 18 24 1
Houston (N = 1706)
2.8
2.1 4
Boston (N = 1152)
6 12 18 24
Los Angeles (N = 5172)
1.4
0.7 -
6 12 18 24 1
Phoenix (N = 410)
2.8
2-1
\. * ' 1A ~
> . i
1 i
0.7 -
6 12 18 24
Pittsburgh (N = 1551)
6 12 18 24 1
St. Louis (N = 1178)
12 18 24
KEY
Median
Mean
90'th & 10th
95'th & 5th
12 18 24
12 18 24
Figure 3-27. Diel plot generated from weekend hourly CO data (ppm) for the eleven CSAs and
CBSAs 2005-2007, Included are the number of monitor days (N) and the median,
mean, 5th, 10th, 90th and 95th percentiles of CO concentration. Note that the y-axis
of the Anchorage CBSA plot is scaled to 5.2 ppm while the other plots are scaled to
2.8 ppm.
March 2009
3-49
DRAFT - DO NOT CITE OR QUOTE
-------
3.5.3. Associations with Copollutants
1 Associations between hourly CO and other copollutants, including S02, N02, 03, PMKI, and PM2 5
2 are provided in box plots for all U.S. monitors in Figure 3-28. The figure also shows the correlation of the
3 24-h avg CO concentration with the daily max 1-h and daily max 8-h CO concentrations as well as the
4 correlation between the daily max 1-h and daily max 8-h concentrations. AQS data were obtained from all
5 available co-located monitors across the U.S. after application of 75% completeness criteria. Pearson
6 correlation coefficients (r) were calculated using 2005-2007 data. Correlation plots analogous to Figure 3-
7 28 for select individual cities are provided in Annex A.
Winter
Spring
IE
I—
¦m
1 -
]—I
2 "
3 -
~—I
4 -
—I
5 "
I—BH
6 "
I-8H
7 "
bBH
8 "
1—\—1—1—1—1—1—1—1—
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0 4 0.6 o.e 1.0
Summer
—i
1 -
*l I 1
2 -
3 "
~ 1
4 -
~ 1
5 "
hSH
6 -
\-m
7 -
I—BDH
8 -
-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
Fall
~ OH
-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 3-28. Seasonal plots of nationwide correlations between hourly CO concentration with
hourly (1) SO2, (2) NO2, (3) O3, (4) PM10, and (5) PM2 5 concentrations. 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. (The numbers in this caption refer to those on the y-axis of each
seasonal plot) Red bars denote the median, and green stars denote the arithmetic
mean.
March 2009
3-50
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
The nationwide mean and median of the correlation with CO were positive for N02, PMi0, and
PM2.5; near zero for S02; and negative for 03. These findings might reflect common combustion sources
for CO, N02, and PM2 5. In all cases, a wide range of correlation values were reported. There were
significant positive correlations for all CO-CO (24-h avg, daily max 1-h, daily max 8-h) comparisons
rendered. Among those copollutants with positive associations, N02 had the highest mean and median
correlations, followed by PM2 5 and PMi0 (correlations vary by season). The correlations of CO with S02
and PMio were not significantly different from zero for any season; S02 would not be expected to
correlate well with CO because S02 emanates primarily from industrial sources. Correlations between CO
and N02 were significant and positive for winter, spring, and fall. Correlations between CO and PM2 5
were significant and positively correlated for winter and fall. Correlations between CO and 03 were
significant and negative for winter only, when CO emissions tend to be high and 03 formation is low. The
copollutant correlation plots for individual cities shown in Annex A illustrate higher and more statistically
significant correlations of CO with both PMi0 and PM2 5 in several but not all of the select cities
displayed. It is widely believed that sources of CO and PM2 5 are highly correlated because they are both
emitted directly during incomplete combustion and because secondary nitrate PM comes from NOx,
which is largely produced from mobile sources. The wide confidence intervals displayed in the
nationwide plots reflect the large pool of data in addition to the micrometeorological factors in each city.
Additionally, CO monitors tend to be located apart from other pollutant monitors based on the criteria
from the CFR of siting at least one monitor near the highest source of CO. Lack of co-location can affect
the correlation of reported CO with PMi0 and PM2 5 concentrations; this limitation also would inhibit the
significance of correlations between CO and PMio_2 5.
Several studies reported correlations between ambient CO and copollutants. Reported relationships
were generally consistent with the correlation data reported by the AQS. Sarnat et al. (2001) reported
significant positive Spearman's correlations of CO with N02 (r = 0.76) and PM2 5 (r = 0.69) and
significant negative correlations of CO with 03 (r = -0.67) in Baltimore. Correlation of CO with S02 was
insignificant (r = -0.12). The Sarnat et al. (2001) 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 N02 is expected given that both are closely related to
mobile source combustion emissions. Sarnat et al (2005) also reported significant year-round association
between CO and PM2 5 and significant associations between CO and S042" aerosols. Tolbert et al. (2007)
reported correlations between multiple pollutants in Atlanta and also showed the highest Spearman's
correlation for CO with N02 (r = 0.70). CO was also reported to have fairly high correlation with PM2 5
elemental carbon (EC) (r = 0.66), PM2 5 organic carbon (OC) (r = 0.59), and PM25 total carbon (TC)
(r = 0.63). Correlations were reported to be much lower for CO with 03 (r = 0.27) and PM2 5 S042
(r = 0.14). Kim et al. (2006) measured CO, N02, and PM2 5 at ambient fixed sites in Toronto, Canada and
March 2009
3-51
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
found associations, averaged over monitoring stations, of CO with PM2.5 (Spearman's r = 0.38, non-
significant) and of CO with N02 (r = 0.72, significant).
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 as policy-relevant background (PRB) concentrations. PRB
concentrations are those that are exclusive of anthropogenic emissions in the U.S., Canada and Mexico
(North America), and consist of world-wide biogenic emissions (including North America) and
anthropogenic emissions elsewhere in the world.
PRB concentrations of CO can best be 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) (http://www.esrl .noaa.gov/gmd/ccgg/iadv). CO data
through December 2007 are available with extensive quality assurance and control information from a
worldwide network of 72 nodes active in December 2008. ESRL GMD uses the highly sensitive gas
chromatography-mercury liberation photometric detection technique with precision to 1 parts per billion
(ppb) in 50 ppb (2 ppb in 200 ppb) and accuracy to 1.5 ppb in 500 ppb (2 ppb in 200 ppb) (Novelli et al.,
2008).
March 2009
3-52
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
*
Figure 3-29. Map of the baseline monitor sites used in this assessment to compute policy-
relevant background concentrations.
This assessment used data from 2005 through 2007 at 12 remote sites in the U.S. to determine PRB
(see the map in Figure 3-29): 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. Average concentrations for each month and for each
year, 2005-2007 are shown for each site in Figure 3-30. All sites demonstrate the well-known seasonality
in background CO with minima in the summer and fall and maxima in the winter and spring. Summer-
time minima may be related to photochemical reaction of CO with OH, as described in section 3.3.
Analysis for North American PRB is made here by segregating the three Alaska sites (based on their high
latitude) and the two Hawaii sites (based on their distance from the continent) and treating the remaining
seven sites as representative of the CONUS background. The 3-year avg CO PRB in Alaska ranged from
127 to 135 ppb with an average of 130 ppb; in Hawaii from 95.3 to 103.1 ppb with an average of 99.2
ppb; and over the CONUS from 118 to 146 ppb with an average of 132 ppb.
Wendoven Utah
Colorado
So uth em G reat P la in s,
Oklahoma
Barrovv AJaska
Biscayne, Florida
Shemya Island, Alaska
Cold Bay, Alaska
Mauna Loa, Hawaih *
* CO monitors
Monitor Location
Trinidad
March 2009
3-53
DRAFT - DO NOT CITE OR QUOTE
-------
¦aVv
Fill*, W1
BilCiyn*. FL
J F MAMJ JASON0J FMAM J J ASONDJ F MAMJ JASOND
2005 2006 2007
J FMAMJ J ASONDJ FMAMJ JASONDJ F MAMJ JASONO
2005 2006 2007
JFMAMJ JASONDJFMAMJ JASONOJ FMAMJ JASONO
2005 2006 2007
Figures 3-30. 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.
3.6. Issues in Exposure Assessment
3.6.1. Summary of Findings from 2000 CO AQCD
1 The 2000 CO AQCD (U.S. EPA, 2000) describes the results of studies completed prior to 1999 on
2 personal exposures and microetn ironmental concentrations of CO. Although these studies may no longer
3 be representative of current exposure levels due to declining ambient CO concentrations, the personal-
4 microenvironmental-ambient relationships are still instructive. Time spent commuting, particularly in
5 cars, was a major contributor to personal CO exposures. Many studies measured in-vehicle concentrations
6 of CO and found elevated concentrations compared to fixed-site monitors. Roadside CO monitors were
March 2009
3-54
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
elevated compared to ambient levels, and equal to or lower than in-vehicle levels (e.g., Ott, 1994; Rodes
et al., 1999). A small portion of the CO concentrations inside a vehicle cabin comes from the vehicle
itself, while a substantial fraction comes from roadway traffic 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. 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 non-ambient indoor sources, such as cooking, space heating, and smoking.
Separation of indoor CO into ambient and non-ambient components is important for determining the
effect of ambient CO concentrations, although this has not been done successfully in previous studies.
Two large studies done 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,
1985; Johnson, 1984). Non-ambient 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. Non-ambient exposures tend to obscure the relationship between ambient CO
concentrations and personal exposure.
3.6.2. General Exposure Concepts
A general framework for human exposure modeling was described in the 2008 Draft PM ISA, 2008
NOx ISA, and 2008 SOx ISA (U.S. EPA, 2008e, f). A brief conceptual model of human exposure is
provided here with respect to CO for the readers' convenience. An individual's daily exposure to CO can
be described based on a compartmentalization of the person's activities over a time period of interest:
E = |C jdt
Equation 3-2
where E = exposure over some duration, () = CO concentration at location /. and dt = time spent in
location j. This basic equation can be broken down into a microenvironmental model that accounts for
exposure to CO of ambient (Ea) and non-ambient (Ena) origin of the form:
March 2009
3-55
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
E = Ea+E,
Equation 3-3
This assessment focuses on the ambient component of exposure because the non-ambient portion of
exposure is subject to individual behavior. Daily Ea can be expressed in terms of the fraction of time spent
outdoors and indoors being exposed to ambient CO concentration, Ca (Wilson et al., 2000):
where/= fraction of time, subscript o = outdoor, subscript i = indoor, and Fmf = infiltration factor, /'inf
quantifies the equilibrium fraction of the CO concentration outside the microenvironment that penetrates
inside the microenvironment and remains in mixture. It is a function of the building air exchange
characteristics and the properties of the gas. Assuming steady state ventilation conditions, the infiltration
factor is a function of the penetration (/') of CO, the air exchange rate (a) of the microenvironment, and
the rate of CO loss (k) in the microenvironment; Fmf = Pa/(a+k). Given that k 0 for CO, ,¦ reduces to
P. Equation 3-4 is subject to the constraint fa + Zf = 1 to reflect the total exposure over a specified time
period. The indoor term has a summation because indoor exposure can occur in various
microenvironments throughout a time period of interest. "Outdoor" exposure can occur in parks, yards,
sidewalks, and on bicycles or motorcycles. "Indoor" 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). The complex human activity patterns that dictate
exposure across the population (all ages) are illustrated in Figure 3-31 (Klepeis et al., 2001). This figure
illustrates the diversity of daily activities among the population as well as the proportion of time spent in
each microenvironment. Although activities in Klepeis et al. (2001) are presented over a day, information
from this figure can be extracted to deduce variation over a 1-h or 8-h period.
As seen in equation 3-4, ambient exposure is a linear function of the ambient CO concentration.
The ambient exposure factor, a, is the ratio between the personal exposure to ambient CO and the ambient
concentration of CO (or the ambient exposure factor); i.e., a = EJCa. a is therefore the proportionality
factor in equation 3-4, i.e., a =fD +1(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 Fmf. If there are no significant local outdoor
sources and sinks of CO, then Ca can be approximated by the concentration measured by an ambient
monitor. If local sources and sinks exist and are significant, then the ambient component of outdoor air
must be estimated using dispersion models, land use regression models, receptor models, fine scale
chemistry-transport models or some combination of these techniques.
Equation 3-4
March 2009
3-56
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
A variety of approaches can be used to estimate exposure to ambient CO. In some cases, individual
personal exposures are measured with personal exposure monitors (PEMs), where personal samples are
taken to estimate population exposure. In other cases, ambient concentrations are used as an exposure
indicator. The ambient concentration may be based on measurements made at a single ambient monitor or
as the average of several ambient monitors.
Very few recent exposure assessment studies utilized measurements of CO concentration and none
of the recent exposure error studies used measurements of personal and ambient CO exposure data,
although Sheppard et al. (2005) presented a conceptual model for nonreactive pollutants that could
include CO. Many recent studies analyzing exposure error used PM data (e.g., Sheppard et al., 2005;
Wilson and Brauer, 2006; Zeger et al., 2000). The lack of recent CO data contributes to uncertainty
regarding personal-ambient relationships. However, review of exposure error studies for PM, under the
assumption that it is nonreactive, is instructive for discerning the relative influence of ambient and non-
ambient exposures if consideration is made for differences between CO and PM with respect to indoor
source variability and infiltration.
Wilson and Brauer (2006) showed significantly stronger associations between health effects and
ambient exposure than between health effects and total personal exposure. The use of personal exposure
in population exposure assessment studies could cause various errors in the health effect estimate because
the correlation between personal exposure and ambient concentration may be different for each subject in
a panel and may not be statistically significant (Wilson et al., 2007). The correlation between ambient
concentration and personal exposure may be high for some subjects, in which case the exposure error
caused by using ambient concentration instead of personal exposure may be small. In other subjects, the
correlation may be low or negative (and not statistically significant). In this case, the exposure error will
be high and may obscure relationships between ambient exposure and health effects. Differences in
correlation observed in these studies occurred in part because the ambient exposure factor, a, was
different for each subject. The relationships between ambient concentration and the ambient component
of personal exposure were statistically significant for all subjects. The analyses of Wilson and Brauer
(2006) and Wilson et al. (2007) focused on PM, which is subject to indoor-outdoor differences in size
distribution and chemical composition related to differential infiltration. However, this issue is still
relevant for CO because non-ambient CO exposure is uncorrelated with ambient CO exposure and
therefore could obscure health effects relationships observed in epidemiologic studies. Therefore, ambient
concentration is better than total personal exposure as an indicator of ambient CO exposure.
March 2009
3-57
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
5=1
O
Ph
%
a>
Pi
st
"5
100
90
SO
70
60
50
40
30
20
10
0
gaesaEeegeseeeeeeeEeeeeee
IS 5S I<3 Oi OS 03 IS cd IS OJ 03 03 ftpHftpHftQneHpHpH&eHftoS
oooooooooooooooooocooocoo
oooooooooooooooooocooocoo
rj r^i rn 'o ^6 r- oo 6^ o ^ ci 6o 6-! c oi
Source: Klepeis et al. (2001))
Figure 3-31. Distribution of time that a sample population spends in various environments, from
the National Human Activity Pattern Survey.
3.6.3. Monitoring Issues Associated with Exposure Assessment
3.6.3.1. Exposure Assessment using Community-Based Ambient Monitors
Instrumentation and associated monitoring errors are described in section 34; exposure error
related to instrumental measurement error in ambient monitors is described here. Because there will likely
be some random component to instrumental measurement error, the correlation of the measured CO
concentration with the true CO concentration will likely be less than 1. Sheppard et al. (2005) indicated
that instrument error in the individual or daily average concentrations have "the effect of attenuating the
estimate of a." However, Zeger et al. (2000) stated that the "instrument measurement error in the ambient
levels... is close to the Berkson type" and in order for this error to cause substantial bias in later estimation
of the health outcome, the error term (the difference between the true concentrations and the measured
concentrations) must be strongly correlated with the measured concentrations. Zeger et al. (2000)
suggested that, "Further investigations of this correlation in cities with many monitors are warranted/'
3.6.3.2. Personal Exposure Monitors
Portable monitors for measuring personal CO exposure include the Langan and Draeger monitors,
both of which use electrochemical oxidation-reduction techniques (Langan, 1992). These monitors
continuously log CO concentrations, making them suitable for use in personal monitoring studies. More
detail on personal CO monitoring is provided in the 2000 CO AQCD (U.S. EPA, 2000).
Other Outdoor
Residence-Tndoars
Residence-Outdoors
Office/Factory
Inside Vehicle
Near Vehicle
(Outdoors)
Bar,'Restaurant
Other
Indoor
School/
Public Bldg.
tmrrpTT IIIITTT11II11III TTIT II11111 ITTTfnrr
March 2009
3-58
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Measurement Error in Personal Exposure Monitors
Personal electrochemical CO monitors are subject to interference and drift, and have a relatively
high detection limit (approximately 1 ppm) relative to current ambient concentrations. Previous studies in
the 1980's and 1990's, when ambient levels were higher, were able to successfully deploy these monitors,
but more recent exposure studies have avoided personal CO measurements due to the high percentage of
non-detects. 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) evaluated the
Langan CO monitor as part of an air quality sampling manifold. At high (0.4-3.0 ppm) CO
concentrations, the instrument correlated well (R2 = 0.93) with a reference 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 = 0.4 in summer
and R2 = 0.59 in winter with the arithmetic average concentration underestimated by 47% in summer and
by 63% in winter. Chang et al. (2001) pointed out the need for frequent instrument zeroing to minimize
instrument drift. Abi Esber et al. (2007a) 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 in non-ambient exposures. This large
variation in personal measurements can result in high levels of classical measurement error (Sheppard et
al., 2005).
3.6.4. Indoor/Outdoor Relationships and Infiltration
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 measured I/O for CO.
Polidori etal. (2007) measured 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) reported I/O for CO measurements in the
absence of indoor sources in a school building in Athens, Greece and found that I/O varies with season.
During the summer, median I/O was reported to be 0.57 on weekdays, 0.91 on Saturdays, and 0.81 on
March 2009
3-59
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Sundays. In winter, median I/O was reported to be 0.82 during weekdays, 0.90 on Saturdays, and 0.74 on
Sundays. The authors attributed the lower weekday I/O values to traffic-based peaks in outdoor CO
concentrations that were not translated to indoors. In a related work, Chaloulakou et al. (2003) 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), 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
non-detects (Williams et al., 2003).
3.6.5. Personal/Ambient Relationships
3.6.5.1. Panel and Population Exposure Studies
Although several multi-pollutant exposure studies have been conducted recently in the U.S., (e.g.,
Sarnat, 2006), most have not included CO in the suite of pollutants, possibly due to limitations in personal
monitoring techniques. A few studies conducted in Europe and Canada measured personal-ambient
relationships for CO.
The EXPOLIS study (Georgoulis et al., 2002) 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. Regression coefficients greater than 1 may indicate location of a fixed
site monitor away from local sources (e.g., roadways), resulting in a lower ambient value than
experienced in the urban core. 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). 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) study, in the
absence of non-ambient CO from ETS and gas cooking, are shown in Table 3-9. The largest percentage of
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) found similar results. Bruinen
de Bruin et al. (2004) and Scotto di Marco et al. (2005) found that mobile source emissions were
important contributors to personal exposure, as described in the following subsection.
March 2009
3-60
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Table 3-9. Percentage of time exposed to ambient CO (adjusted to reflect the absence of
non-ambient CO from ETS and gas cooking), average CO exposures, and percentage
of exposure estimated for the population.
Percent of time exposed (%)
Exposure (ppm)
Percent of exposure (%)
INDOORS
89.6
81.1
Home
56.5
1.8
49.4
Work
29.1
1.9
26.8
Other
4.1
2.5
4.9
OUTDOORS
1.8
2.1
Home
0.2
2.3
0.2
Work
0.6
2.1
0.6
Other
1.0
2.6
1.2
IN-TRANSIT
8.5
16.8
Walking
3.0
3.0
4.4
Train/metro
0.7
3.0
1.0
Bus/tram
2.0
3.8
3.7
Motorbike
0.2
4.5
0.4
Car/taxi
2.6
5.7
7.2
Source: Bruinen de Bruin et al. (2004).
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; 2001)
reported higher personal exposures than ambient concentrations for children aged 3-6 years old in
Helsinki. Their mean daily max 1-h exposure was 5.2 ppm, compared to 1.4 ppm measured at a fixed-site
monitor. For daily max 8-h and 24-h avg 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 12% of the variability
in the 8-h max 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) 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 0.5-1.4 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
March 2009
3-61
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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) 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.
3.6.5.2. Commuting Time CO Exposure Studies
A number of studies focused on transit-time CO exposure, which can occur in a vehicle or while
walking or cycling. Kaur (2005) 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) found that average in-transit concentrations 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), trolley (winter: 9.6 ±1.9 ppm; summer: 8.2 ±
3 ppm), and rail (winter: 4 ± 0.6 ppm; summer: 3.4 ± 0.7 ppm). Duci et al. (2003) did not provide fixed
site CO concentrations but stated that in-transit exposures were higher in each case. 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; Scotto di Marco et al.,
2005). Scotto di Marco et al. (2005) found that, for a panel of 201 adult Helsinki residents (aged 25-55
years), subjects spent 8.1% of their time in transport, 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) found that, in the absence of non-ambient sources, the subjects
spent 8.5% 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 levels than urban residents with short commute times.
Gomez-Perales et al. (2004; 2007) measured CO exposures on buses, mini-buses, and metro cars in
Mexico City, Mexico to be 12 ppm, 15 ppm, and 7 ppm, respectively. These values are higher than
CONUS measurements and those presented by Kaur et al. (2005), but the relative difference between the
minibus and bus exposures in the Gomez-Perales et al. (2004; 2007) study are similar to those seen for the
taxi-to-bus or car-to-bus comparisons in Kaur et al. (2005). Aim et al. (1999) reported in-vehicle CO
March 2009
3-62
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
concentrations of 5.7 ppm in the morning and 3.1 ppm in the afternoon commute for Kuopio, Finland.
Abi Esber et al. (2007b) report results from CO concentration measurements taken within an automobile
in Beirut, Lebanon during the morning commute period of 7:30 to 9:30 a.m. Weekday trip CO levels
ranged from 10.8 ppm with the windows open and vents closed to 37.4 ppm when driving with windows
and vents closed. Mean and standard deviation for ambient CO concentrations, obtained using a roadside
monitor in Beirut during the periods September-December 2003, August-September 2004, and
May-August 2005 were 1.4 ± 0.7, 1.6 ± 0.4, and 1.1 ± 0.7 ppm, respectively.
Abi Esber and El-Fadel (2008) compared the amount of CO produced by an automobile, driving
the same route of Beirut described in Abi Esber et al. (Abi Esber et al., 2007b) above, by sampling CO
directly from the engine of the vehicle and separately from the cabin of the car under three different
ventilation conditions. For the case when one window was half-open and vents were closed, engine CO
concentrations averaged 12.6 ppm while in-vehicle concentrations averaged 17.7 ppm, which was a
40.5% increase. With windows closed and the air conditioner operating on "recirculating air" mode, CO
concentrations averaged 13 ppm from the engine and 30.2 ppm in the vehicle cabin, a 132% increase.
With windows closed and the air conditioner on "fresh air" mode, engine CO concentrations averaged
18.3 ppm while in-vehicle concentrations were 20.5 ppm, which was only a 12% increase. Figure 3-32
shows that the time series for the cabin and engine CO samples are very similar for the fresh air scenario,
but for the recirculating air ventilation the concentration increases as a logarithmic-type function as CO
builds up in the cabin of the vehicle.
a 50
eL
c.
O
u 25
g
50
20 30
Time(min)
0 ]0 20 30 40 50
Time (min)
Source: Abi Esber and El Fadel (2008)
Comparison of in-vehicle (solid line) and engine (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.
Figure 3-32.
March 2009
3-63
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Riedeker et al. (2003) measured CO concentrations inside patrol cars during shifts. Troopers
recorded in a time-activity diary the ventilation settings of their cars and exit/entry from the vehicle, and
the air conditioning was typically set to recirculation mode during the shifts. Riedeker et al. (2003) 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. Riedeker et al. (2003) noted that within-shift variability was higher than between-
shift variability, which underscores the variability in police officers' activities during a given shift. No
data was segregated by ventilation settings. Chang et al. (2000) measured CO concentrations during a
scripted activity study in Baltimore, MD in 1998 and 1999. Mean 1-h CO concentrations were near the
1 ppm detection limit of the Langan CO monitor. Microenvironmental CO concentrations were
significantly correlated with concentrations measured at a fixed-site ambient monitor for residential, other
indoor, in-vehicle, and outdoor near-road microenvironments during the winter. Significant correlations
were observed only for residential microenvironments during the summer. The location of the ambient
monitor near a roadway may have contributed to the lack of correlation with concentrations measured at
outdoor locations away from roadways. Microenvironmental concentrations inside vehicles were
significantly higher than those for other microenvironments.
Several factors may affect commuters' exposures. Van Wijnen et al. (1995) observed that while
bicycling, cyclists inhale 2.3 times as much air as pedestrians and drivers so that this group may, in fact,
inhale the most CO. Likewise, in their review of roadway exposures to CO and PM, Kaur et al. (2007)
listed a number of uncontrolled factors that may influence exposure. Vertical CO concentration gradients
have been documented with concentrations decreasing with height; lower breathing zone height among
children may make them more likely to be exposed to higher CO concentrations. With respect to
transportation, Kaur et al. (2007) suggested that vehicle ventilation, speed, position in traffic, and
start/stop activity influence in-vehicle exposures. Gomez-Perales et al. (2007) 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) made a similar observation in a study of
urban commuters' exposure within a vehicle.
Studies of vehicle self-pollution are instructive for considering potential for ambient CO infiltration
in vehicles. Behrentz et al. (2004) 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.
March 2009
3-64
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.6.5.3. CO Exposure Assessment Variability and Error
In the context of determining the effects of ambient pollutants on human health, the association
between the ambient component of personal exposures and ambient concentrations is more relevant than
the association between total personal exposures (ambient component + non-ambient component) and
ambient concentrations. If there are no non-ambient sources of a pollutant, the total personal exposure is
equal to the ambient personal exposure. However, non-ambient sources could significantly affect personal
exposures to CO. Unlike PM, CO has no chemical identifiers that can be used to apportion ambient and
non-ambient sources of CO (other than radioactive isotopes, whose detection is rare). For the general U.S.
population, exposure error analysis for epidemiologic studies indicates that fixed-site measured ambient
CO concentration is generally a good indicator of ambient exposure to CO, as discussed in more detail
below.
Figure 3-33 shows hourly vs. personal CO concentration data obtained by Chang et al. (2000) for a
1998-1999 multi-pollutant 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 non-ambient sources of CO.
Figure 3-34 is a box plot of the personal-to-ambient CO concentration ratio for the same five
microenvironments. Figure 3-34 shows that personal exposures in vehicles were on average 2.8 times
higher than ambient during the summer and 4.1 times higher than ambient in the winter. For the other four
microenvironments tested, the average ratio was around 1. Wide variability is seen in these plots,
particularly during the summer. Much of that variability could be due to the influence of non-ambient
sources, which would then result in poor correlation between total personal exposure and ambient
concentration.
Zeger et al. (2000) stated that there are three error terms in the estimate of a person's exposure: 1)
error from using pooled population data in lieu of individual data, 2) error between total exposure and
ambient concentration, and 3) error between the actual and measured ambient concentration. Zeger et al.
(2000) also described that errors related to individual variability and measurement (as stated above for the
latter) are of the Berkson type and therefore not expected to bias estimates of exposure or health effects.
Moreover, Sheppard et al. (2005) simulated ambient and non-ambient exposures to a non-reactive
pollutant and observed that non-ambient CO exposure has no effect on the association between ambient
CO exposure and health outcomes for the case where ambient and non-ambient sources were
independent.
March 2009
3-65
DRAFT - DO NOT CITE OR QUOTE
-------
E
I
3 -
g
a. 2
>>
i
o
X
+ + + ~
+~ ~
o
Indoor Residence
A
Indtx>r Other*
~
Outdoor near Roadway
V
Outdoor away from Road
~
In Vehicle
~
~
A
~
"°Se.»*0 ~~
^*8* ft °°
o o
Hourly Ambient CO (ppm)
Source: Chang et al. (2000)
Figure 3-33. Hourly personal vs. ambient CO concentrations obtained in Baltimore, MD. Summer
of 1998 in five settings: indoor residence, indoor other, outdoor near road, outdoor
away from road, in vehicle.
wswswswsw
Indoor Indoor Outdoor Outdoor In Vehicle
Residence Other Near away from
Road Road
Source: Adapted from Chang et al. (2000)
Figure 3-34. Box plots of the ratio of personal to ambient concentrations obtained in Baltimore,
MD. Summer of 1998 and winter of 1999 in five settings: indoor residence, indoor
other, outdoor near road, outdoor away from road, in vehicle. The dotted line shows
the mean, and the solid line shows the median. S = summer; W = winter.
March 2009
3-66
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
For community time-series epidemiology, the community averaged concentration, not the
concentration at each fixed monitoring site, is the concentration variable of concern (Zeger et al., 2000).
The correlation between the concentration at a central community ambient monitor and the community
averaged concentration depends on at least the following three factors. First, the distribution of monitors
across space: CO emissions from traffic might show spatial heterogeneity near roadway sources but have
a more homogeneous distribution farther downwind. It follows that use of a large number of samplers will
dampen inter-sampler variability. Second, the relationship between the measurement at the ambient
monitoring site and the community average concentration: if the site is selected to measure a "hot spot" or
pollution from a nearby source, estimates of community exposure could be skewed upwards. Third,
terrain features or source location: different terrain features or source locations across sub-communities
may differ in the temporal pattern of pollution. Intra-urban spatial heterogeneity was discussed in detail in
Section 3.5.1.2. Community exposure may not be well-represented when monitors cover large areas with
several sub-communities having different sources and topographies. For example, with the exception of
two closely located samplers, there is poor inter-sampler correlation among the Pittsburgh, PA CO
samplers. This variability reflects the wide topographical differences throughout the Pittsburgh CSA as
well as variation in traffic usage and distance to near-road sources at each point in contrast to the Phoenix
CBSA, where inter-sampler correlation was high.
Epidemiologic studies of long-term exposure rely on differences among communities in long-term
average ambient concentrations. If exposure errors are different in the different communities, the
differences in long-term ambient concentrations among communities may not represent the differences in
long-term average exposures. For example, there may be community to community differences in
measurement error, in the average ambient exposure factor (a) or the average non-ambient exposure. This
could happen if exposure to newly formed pollutants generated by vehicular traffic or pollutants from
other localized sources differed among the spatial areas. Thus, in a regression of health effects against
average concentration (as an indicator for average exposure) there could be a different amount of error
(either positive or negative) in the exposure indicator for each spatial area. This could add error and bias
the slope up or down. Geostatistical tools enable the use of concentration fields that include spatial
variations in concentration. However, it has not yet been possible to include individual or small-area
variations in the personal exposure to ambient concentrations or variations in personal exposures to
indoor-generated pollutants in long-term studies of the associations of pollutants with health effects.
3.6.6. Multi-pollutant Exposures
Since incomplete combustion is the primary source of ambient CO, exposure to ambient CO is
accompanied by exposure to other combustion-related pollutants, such as NOx and PM. Thus, ambient
March 2009
3-67
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 gas combustion. Correlations between ambient CO and ambient PM2 5,
PMio, N02, S02, and 03 from AQS data and the peer-reviewed literature were presented in section 3.5.3.
Nationwide ambient CO was most highly correlated with ambient N02 followed by PM2 5 and PMi0.
Correlations between CO and PM2 5 were not always statistically significant on a national basis;
correlations were insignificant for ambient CO with ambient S02 and ambient PMi0, and ambient CO was
negatively correlated with ambient 03. Correlations of ambient CO with PM2 5 and PMi0, were much
higher and more statistically significant at some of the select cities, as shown in Annex A.
Sarnat et al. (2001) analyzed the relationship between personal exposure to PM2 5 and ambient CO
concentrations and found significant associations during winter of ambient CO with personal exposure to
total PM2 5 (slope = 3.99) and to ambient PM2 5 (slope = 6.30). Ambient CO was not significantly
associated with personal PM2 5 exposure during summer. Sarnat et al. (2005) also reported significant
association between ambient CO concentrations and personal PM2 5 exposure for winter (three 12-day
sessions in 2000, dates not specified) only and significant year-round associations between ambient CO
concentrations and personal exposure to S042" aerosols.
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, PM2 5, and ultrafine PM in a street canyon, Kaur et al. (2005) found low
Pearson's correlation of personal CO exposure with personal PM2 5 exposure (r = 0.23). Personal CO
exposure had much better correlation with personal ultrafine PM exposure (r = 0.68). Chang et al. (2000)
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 vs. summer). In general, the correlations tended to
be stronger and more significant in the winter. Chang et al. (2000) pointed out that indoor air exchange
rates tend to be lower in winter, which could cause these correlations to be more sensitive to non-ambient
sources. Significant associations of CO with benzene and toluene were observed in-vehicle. Because CO
exposures most often occur together with exposure to other combustion-related pollutants, interpretation
of health studies using ambient CO data can be a challenge, as discussed further in Chapter 5.
3.6.7. Exposure Modeling
A number of modeling techniques to describe air pollutant exposures at the individual and
population level have been published since 2002. Ambient CO exposure models often assume that
March 2009
3-68
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
vehicular traffic is the sole source of emissions given that roughly 90% of urban CO is estimated to come
from traffic (Gulliver and Briggs, 2005). Individual-level models can predict personal exposure based on
a person's path of travel and on conditions of the surrounding environment using traffic and meteorology
data. One such example is the Space-Time Exposure Modeling System (STEMS) (Gulliver and Briggs,
2005). Population-based methods, such as the Air Pollution Exposure (APEX) model
(www.epa.gov/ttn/fera/human apcx.html). Stochastic Human Exposure and Dose Simulation (SHEDS)
(Burke et al., 2001), and EXPOLIS models (Kruize et al., 2003), involve stochastic treatment of the
model input factors. Another approach is to predict location-based exposures using a deterministic model
such as the CMAQ model, California Line Source Dispersion Model (CALINE), CALPUFF (long-range
plume transport model), or Operational Street Pollution Model (OSPM) for determination of street-level
pollution coupled with infiltration models to represent indoor exposure to ambient levels (Appel et al.,
2008; Gilliam et al.; Hering, 2007; Mensink and Cosemans, 2008; Wilson and Zawar-Reza, 2006).
Stochastic and deterministic methods are often combined, as described below. Land use regression (LUR)
models have also been developed to describe pollution levels as a function of geographic source behavior
(Gilliland et al., 2005; Ryan and LeMasters, 2007; Veen et al., 1997). Similarly, other spatial interpolation
techniques have been used to determine geographic distributions of exposure (Marshall et al., 2008).
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; 2006). A detailed description of various modeling techniques is provided in the 2008 NOx ISA and
2008 SOx ISA annexes (2008f, g), and an explanation of associated modeling errors is provided in the
2008 Draft PM ISA (U.S. EPA, 2008f). Applications of these models to assessment of CO exposure are
described in the following paragraphs.
The STEMS model maps exposures based on inputs for traffic levels, atmospheric dispersion,
background concentrations, and geography. Gulliver and Briggs (2005) 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) 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
PMio 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 micro-scales (Jerrett et al., 2004). Flachsbart (Flachsbart) tested numerous meteorological, traffic, and
background CO input variables in a regression approach to predicting CO exposure among individuals
March 2009
3-69
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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.
Bruinen de Bruin et al. (2004) utilized the EXPOLIS model to predict CO population exposures in
Milan, Italy. The simulation results showed that the U.S. 8-h NAAQS level was exceeded in one case out
of 1,000. The model also showed that exposures exceeded 20 ppm in one 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 probabilistic NAAQS Exposure Model for CO (pNEM/CO) was used
in the 1992 and 2000 CO AQCDs (U.S. EPA, 1992, 2000) to estimate personal CO exposures among
adult populations. Given reductions in ambient CO levels, details, and results from those efforts are
provided here. The current version of this model, now known as APEX, was used for exposure
assessment in the 03 NAAQS review (U.S. EPA, 2006b).
To examine indoor concentrations of ambient CO, Dimitroulopoulou et al. (2006) 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 non-ambient 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 non-source indoor concentrations was
attributed to the lack of CO loss mechanisms. With maximum CO concentrations modeled at 3.3 ppm in
the absence of sources, the probability density function of indoor exposure to ambient fell well below the
8-h NAAQS. In the Reducing Urban Pollution Exposure from Road Transport (RUPERT) study, Bell et
al. (2004) 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).
Another set of approaches to improve exposure estimates in urban areas involves construction of a
concentration surface over the geographic area. This does not estimate exposure directly because it does
not account for activity patterns or concentrations in different microenvironments. It provides an
improved estimate of the expected local outdoor concentration near residences, schools or workplaces,
and roadways across the area. 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-principles models, such as dispersion
models and chemical transport models, 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.
March 2009
3-70
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Marshall et al. (2008) 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 inverse-distance-
weighted (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 NO and N02, and indicated that LUR's higher spatial
precision reflects neighborhood-scale effects from nearby land use, but may not account for urban-scale
variation. These results highlight the variation in local concentration estimates with choice of estimation
technique.
3.7. Summary and Conclusions
3.7.1. Sources of CO
In the U.S., on-road mobile sources constituted more than half of total CO emissions in 2002, or
-63 MT of-109 MT total. In metropolitan areas in the U.S., for example, as much as 70-75% of all CO
emissions can come from on-road vehicle exhaust (U.S. EPA, 2006a). The majority of these on-road CO
emissions derive from gasoline-powered vehicles since the 02 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% of total CO emissions from all sources in 2002, and fires in 2002 added
another 13%, or -14.5 MT, to the national CO emissions total.
March 2009
3-71
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.7.2. Physics and Chemistry of Atmospheric CO
CO is produced by photooxidation of CH4 and other VOCs in the atmosphere. 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 C02 and H02. The mean photochemical lifetime (x)
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.
3.7.3. Ambient CO Measurements
As of March 2008, 19 automated FRMs and no FEMs had been approved for CO. All EPA FRMs
for CO operate on the principle of nondispersive infrared (NDIR) detection and can include gas filter
correlation (GFC). The required lower detection limit for FRMs in the EPA network is 1.0 ppm. In 2007,
there were 285 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. At
least 70 monitors across the U.S. have been positioned at microscale within 10 m of a road to capture
near-road concentrations. At larger scales, monitor distance from a road is inversely related to the road's
average daily traffic count to capture community averages.
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 2005-2007
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-year period. The nationwide mean, median, and interquartile range for 8-h daily max
concentrations reported between 2005-2007 were 0.7, 0.5, and 0.5 ppm, respectively. 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 years. 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 CSAs/CBSAs
examined for this assessment reveal a wide range of response between monitors in each city and among
March 2009
3-72
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
the cities judged against each other. 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. Likewise, micro- and neighborhood-
scale variation related to urban topography or microenvironmental source distribution may have a
significant effect on ambient CO concentrations with respect to street canyon concentrations. Anchorage,
AK had concentrations roughly twice those of the other metropolitan areas. Most of the CSAs/CBSAs
examined here had 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 required
FRM lower detection limit of 1.0 ppm coupled with the coarsely reported measurement resolution
(0.1 ppm) can artificially influence the comparison statistics shown in the tables and result in apparent
heterogeneity in the box plots (Figures 3-13 through Figures 3-16).
For nationwide copollutant correlation data, ambient CO had the highest mean and median
correlations with ambient N02, followed by PM25 (correlations vary by season). The S02 and PMi0
correlations did not appear to be statistically significantly different from zero for any season, and
statistical significance was only observed in winter and fall for PM2 5. Correlations between CO and N02
were significant for all seasons but summer. Correlations between CO and 03 were significant and
negatively correlated for winter only. The copollutant correlation plots shown in Annex A illustrate higher
and more statistically significant correlations between CO and PM in several but not all of the select cities
displayed.
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-year 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 CO concentration data. The studies of
personal exposure to CO presented here generally found that the largest fraction of an individual's
exposure to ambient CO occurs indoors but the highest CO exposure levels occur in transit. Among
commuters, exposures were highest for those traveling in automobiles in comparison with those traveling
March 2009
3-73
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
in buses and motorbikes and with those cycling or walking. A portion of that exposure is thought to come
from the vehicle in which the exposed person travels. Commuting time was an important predictor of
exposure, such that subjects living in low-CO concentration suburban areas and commuting to work
experienced higher CO levels than residents of higher-CO concentration urban areas with short commute
times. Additional analyses indicated that on-road mobile emissions were the most important source of CO
exposure for non-ETS-exposed subjects. Results have also indicated that time spent outdoors and
proximity to roadways may have been important in determining personal exposure in children.
Uncontrolled factors that may influence exposure include vertical CO concentration gradients; vehicle
ventilation, speed, position in traffic, and start/stop activity; and meteorology.
For the general U.S. population, exposure error analysis for epidemiologic studies indicates that
fixed-site measured ambient CO concentration is generally a good indicator of ambient exposure to CO.
Errors associated with exposure to non-ambient CO sources, for example smoking or gas cooking, are of
the Berkson type and therefore not expected to bias estimates of exposure or health effects. Simulations of
ambient and non-ambient exposures to a non-reactive generic pollutant demonstrated that non-ambient
exposure had no effect on health effects outcomes for the case where ambient and non-ambient sources
were independent. Likewise for community time-series epidemiology, selected monitoring sites are
thought to be representative of community averages. Measurement at a "hot spot" could skew community
exposure estimates upward. Topographical or source variability could produce differences in the temporal
pattern of pollution. Differences such as those produced by variation in local sources among communities
in long-term exposure studies may also produce error in estimation of the ambient exposure factor. Such
differences could add error, and therefore influence the slope of health effects estimate regressions up or
down.
Several studies have examined multi-pollutant exposure. Since incomplete combustion is the
primary source of ambient CO, exposure to ambient CO is accompanied by exposure to other
combustion-related pollutants, such as NOx and PM. High correlations of ambient CO with N02 and
PM2 5 have been observed in the peer-reviewed literature and AQS data. Thus, ambient CO is often
considered a surrogate for exposure to traffic-generated pollutants. Because CO exposures most often
occur together with exposure to other combustion-related pollutants, interpretation of health studies using
ambient CO data can be a challenge, as discussed further in Chapter 5.
March 2009
3-74
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 deoxyhemoglobin, resulting in COHb formation and consequent
reduction in 02-carrying capacity of blood and impaired 02 delivery to tissues. Research on the basics of
CO pharmacokinetics date 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). Relevant new data are included when available. New explanations of recent models of Hb binding
are included, as well as discussion on 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; Tyuma et al., 1981). 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.
March 2009
4-1
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
4.2. Carboxyhemoglobin Formation
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 COHb. These methods were reviewed in depth in the previous
AQCD (U.S. EPA, 2000). These 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) and Ott and
Mage (1978), as well as various others (Selvakumar et al., 1992; Singh et al., 1991). Using these simple
models, 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-
15 minute periods (Ott and Mage, 1978).
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). 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; Stewart et al., 1973b). 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 02Hb concentration is constant and is
as follows in Equation 4-1:
f \ f \
' b
d[COHb]t • [COHb]0P-O
VK — Vco f t;
CW2
dt [02Hb]M
1
1 PB - 47
¦ + -2
dlco v
V V /
P:CO
1 PB - 47
- + -2
DLC0 V
V V A J
Equation 4-1
where Vb is blood volume in milliliters (mL); [COHb]t is the COHb concentration at time t in mL CO/mL
blood, at standard temperature and pressure, dry (STPD); \co is the endogenous CO production rate in
mL/min, STPD; [COHb]0 is the COHb concentration at time zero in mL CO/mL blood, STPD; [02Hb] is
the 02Hb concentration in mL 02/mL blood, STPD; M is the Haldane coefficient representing the CO
March 2009
4-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
chemical affinity for Hb; P-02 is the average partial pressure of 02 in lung capillaries in mmHg; yA is
the alveolar ventilation in mL/min, STPD; DLCO is the lung diffusing capacity of CO in mL/min/mmHg,
STPD; PB is the barometric pressure in mmHg; and PiCO is the CO partial pressure in inhaled air in
mmHg.
This 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 incorporates the interdependence of COHb
and 02Hb levels since they are derived from the same pool of blood Hb. 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). The linear equation prediction of COHb concentration at or below 6% will only differ ± 0.5% 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 (McCartney,
1990). 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 FiCO 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.
March 2009
4-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1.00-1
1.00
F.CO
TH,
0.75 -
0.75
COHb,
0.50-
0.50
f 0.25-
(75
c
CD
s °-
03
C
o
0.25
D, CO
E -0.25-
LL
-0.25
-0.50-
-0.50
-0.75-
-0.75
-1.00
-1.00
0.6 s
6s
1 min
1.7 h
17 h
10 min
-2-1012 3
Log Time
Source: modified from McCartney (1990)
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 02 or CO saturation
versus p02 orpCO relation (McCartney, 1990). 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 to
7,500 ppm for 75 s to 5 min) were not accurately predicted by CFK modeling (Benignus et al., 1994;
Tikuisis et al., 1987a; Tikuisis et al., 1987b). Consistently, the predicted COHb value overpredicted
venous COHb (0.8 to 6%) and underpredicted arterial COHb (1.5 to 6.1%) and this disparity increased
after exercise. Individual differences between arterial and venous COHb varied from 2.3 to 12.1%
(Benignus et al., 1994). 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
March 2009
4-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
and venous blood and differences in cardiac output and lung wash-in. A modified CFK was created to
adjust for these issues and produce a more accurate COHb prediction (Smith et al., 1994).
4.2.2. Multicompartment Model
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). A five compartment 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; Bruce et al., 2008; 2006). This model structure is illustrated in Figure 4-2. This model includes the
dynamics of CO storage in the lung and its dependence on ventilation and CO pressure of mixed venous
blood, relaxes the assumption Hb is saturated by including the role of CO in altering the 02 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 assumes ventilation and Pa02 are constant. 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) compared the two
models and found similar results for long term exposure settings (1,000 min), however, the
multicompartment model predicted somewhat lower COHb levels compared to the CFK model over
transient CO uptake conditions when using data taken from Peterson and Stewart (1970).
March 2009
4-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
P,co
LUNGS
PACO PA02
Pp.2
'•'i Dlco
>
.
Q
I
MIXED
VENOUS
ARTERIAL
NONMUSCLE
TISSUE
....
v„
Qo,
MUSCLE TISSUE
Mb,02
I
Vtm,
Mb,CO
Mb;0;
t
Vtm-.
MbjCO
Source: Modified from Bruce et al. (2008)
Figure 4-2. Overall structure of the multicompartment model of storage and transport of CO.
Includes compartments for lung, arterial blood, venous blood, muscle tissue, and
nonmuscle tissue. The muscle compartment is divided into two subcompartments
for diffusion of gases within the tissue.
4.2.3. Mathematical Model Usage
Since measurements of COHb in the population are not readily available, mathematical models are
used to predict the resulting COHb levels from various CO exposure scenarios. Figure 4-3 illustrates the
predictions of COHb after 1, 8, 12, or 24 h of CO exposure at a range of concentrations in a healthy,
inactive, adult human. 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; Benignus et al., 1987; 2006). This dynamic model uses the
nonlinear CFK equation with modifications presented in Smith et al. (1994). The data in Figure 4-3 are
presented as the change in percent COHb from endogenous due to CO exposure to NAAQS relevant
levels from 0-35 ppm (A) and relevant ambient levels from 0-6 ppm (B). Endogenous CO production
varies as described in Section 4.5 but generally results in less than 1% COHb, with a QCP modeled value
of 0.39% at time zero. Figure 4-3 illustrates that 35 ppm CO for 1 h results in an increase of 0.56% COHb
over endogenous levels and 9 ppm CO for 8 h results in a 0.83% increase over endogenous levels. Also,
March 2009
4-6
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
these graphs show that long term, low concentration CO exposure results in equivalent COHb levels to
higher concentration, acute exposure. For example, COHb resulting from 35 ppm for 1 h (0.56%) is
approximately equivalent to 6 ppm for 8 h (0.55%) or 4 ppm for 24 h (0.57%).
A 1
5 " ° ¦
S & 0.7-
1 § 0.6-
5 8) °5"
S° 0.4-
^ £ 0.3-
8 0.2-
0.1 ¦
0 -
Figure 4-3. Predicted COHb increments over endogenous levels resulting from 1, 8,12, and 24 h
CO exposures in a modeled healthy human at rest. The QCP model used a dynamic
nonlinear CFK with affinity constant M = 230. The data are presented as the change
in percent COHb from endogenous levels due to CO exposure to 0-6 ppm (A) and
0-35 ppm (B). Note: Graph A is a subset of Graph B.
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 02 share various physico-chemical
properties, thus allowing for the extension of the knowledge about 02 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 (Vco) representing the product of DLCO and the mean alveolar pressure (PACO).
The importance of dead space volume, gas mixing and homogeneity, and ventilation/perfusion matching
were discussed in depth in the 2000 AQCD (U.S. EPA, 2000).
March 2009
4-7
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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 molecule 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,
1983). Then, CO diffuses in a "liquid" phase across the air-blood interface, binding red blood cell (RBC)
Hb. At environmental CO levels, CO uptake into RBC is limited by the reaction rate of binding of CO to
02Hb forming COHb. Pulmonary capillary RBC CO diffusion is rapidly achieved (Chakraborty et al.,
2004; Gibson and Roughton, 1955; Reeves and Park, 1992; Roughton and Forster, 1957). The rate and
level of COHb depends upon the pCO, p02 in the air, time of exposure, and the ventilation rate (Roughton
and Forster, 1957). 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, 1970a). Considerable
concentrations of CO have been measured in spleen, lung, kidney, liver, muscle, and heart (Vreman et al.,
2005; Vreman et al., 2006), whereas less CO is localized to fatty tissues, such as adipose and brain (Table
4-1). 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, 1970a). Transport pathways and body stores of CO are shown in Figure 4-4.
Table 4-1 a. CO concentration in pmol/100 g wwtissue- human.
Exposure
Adipose
Brain
Muscle
Heart
Kidney
Lung
Spleen
Blood
% COHb
Background
2 ± 1
3 ± 3
15 ± 9
31 ±23
23 ± 18
57 ±59
79 ±75
165±143
1.5 ± 1.2
Fire
5 ± 4
7 ± 5
24 ± 16
54 ±33
27 ±11
131 ±127
95 ±69
286±127
3.8 ±3.2
Fire + CO
18 ±29
17 ± 14
168±172
128 ±63
721 ±427
1097 ±697
2290±1409
3623 ±1975
40.7 ± 28.8
CO asphyxiation
25 ±27
72 ±38
265 ± 157
527 ±249
885 ± 271
2694 ±1730
3455±1347
5196 ± 2625
56.4 ±28.9
Source: Vreman et al. (2006)
March 2009
4-8
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-1 b. CO concentration in pmol/mg fresh weight - adult mouse.
Exposure
Brain
Muscle
Heart
Kidney
Lung
Spleen
Liver
Intestine
Testes
Blood
% COHb
Background
2 ± 0
10 ± 1
6 ± 1
7 ± 2
3 ± 1
6 ± 1
5 ± 1
4 ± 2
2 ± 1
45 ±5
0.5
500 ppm CO
18 ± 4
14 ± 1
100 ± 18
120 ± 12
250 ±2
229 ± 55
115 ± 31
9 ± 7
6 ± 3
2648 ± 400
28
30 pM heme
2 ± 0
7 ± 1
14 ± 3
7 ± 2
8 ± 3
11 ± 1
8 ± 3
3 ± 1
2 ± 0
88 ± 10
0.9
Source: Vreman et al. (2005)
Carbon Monoxide in the Ambient Air
RBC
Hemoglobin
Intravascular compartment
Alveolar Air
Plasma
Endogenous
production
of CO
Metabolism
of CO to CO-
Intracellular
enzymes
Myoglobin
Source: Adapted from Coburn (1967)
Figure 4-4. Diagrammatic presentation of CO uptake and elimination pathways and CO body
stores.
4.3.1.2. Lung Diffusion of Carbon Monoxide
1 Lung diffusion of CO is an entirely passive process of gas diffusion across the alveolo-capillary
2 membrane, through the plasma, across the RBC membrane and into the RBC stroma, where CO binding
3 to Hb rapidly occurs. Membrane and blood phase transfer are governed by physico-chemical laws,
4 including Fick's first law of diffusion. The diffusing capacity of the lung for CO, represented as DLCO, is
5 a measurement of the partial pressure difference between inspired and expired CO. Due to the rapid
6 binding of CO to Hb, a high pressure differential between air and blood exists when CO air levels are
7 increased. Inhalation of CO-free air reverses the pressure differential (higher CO pressure on the blood
8 side than the alveolar side), and then CO is released into the alveolar air. Since CO is also produced
9 endogenously, CO release will also be affected by this production pressure. However, the air-blood
March 2009
4-9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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, PAC02, body position, exercise, time of day, age, etc.
(Forster, 1966; Hsia, 2002). DLCO consistently decreases after intense bouts of exercise, likely due to the
redistribution of blood volume to the periphery (Hanel et al., 1997; Manier et al., 1991). 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). 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).
4.3.2. Tissue Uptake
4.3.2.1. The Respiratory Tract
The upper respiratory tract contributes little to the overall COHb uptake. The lung has nearly
constant exposure to CO; however, relatively little CO diffuses into the tissue except for 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) or in the monkey oronasal cavity after high CO exposure (Schoenfisch
et al., 1980).
4.3.2.2. The 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 02, meaning one part CO and 218
(210-250) parts 02 would form equal parts of 02Hb and COHb (Engel et al., 1969; Rodkey et al., 1969;
Roughton, 1970). This would happen when breathing air containing 21% 02 and 960 ppm CO. This
concept was presented by Haldane and Smith (1898) and later represented as the Haldane constant M
(210-250) in the Haldane equation by Douglas, Haldane, and Haldane (1912). M is relatively unaffected
by changes in physiological pH, C02, temperature, or 2,3-diphosphoglycerate:
March 2009
4-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
COHb -T- 02Hb = M x (pCO p02)
Equation 4-2
The Hb association rate for CO is 10% slower than 02 and occurs in a cooperative manner
(Chakraborty et al., 2004; Sharma et al., 1976). Hb is composed of four globin chains each containing a
heme group capable of binding CO or 02. 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). More simply, the greater the number of heme sites bound to CO, the
greater the affinity of free heme sites for 02, thus causing Hb to bind and retain 02 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 02 (kCo = 4X 10"4 £02), which accounts for the high affinity
values (Chakraborty et al., 2004). The half-time of dissociation reaction is about 11 seconds at 37 °C
(Holland, 1970). 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 et al., 1988).
CO binding to Hb also has effects on the 02 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 02 binding (Roughton, 1970). This is referred to as the "Haldane effect" and causes
tissues to have more trouble obtaining 02 from the blood, even compared to the same extent of reduced
Hb resulting from anemia. For example, Figure 4-5 (as explained in the 2000 AQCD) illustrates that an
acute anemia patient (50% of Hb) at a venous p02 of 26 mmHg (v'i), 5 vol % of 02 (50% saturation) was
extracted from the blood. In contrast, a CO poisoned person with 50% COHb, the venous p02 will have to
drop to 16 mmHg (v'2) to release the same 5 vol % 02. This more severe effect on 02 pressure may lead
to brain 02 depletion and loss of consciousness if any higher demand of 02 is needed (e.g., exercise).
March 2009
4-11
DRAFT - DO NOT CITE OR QUOTE
-------
Normal
5 mL/100 mL
50% COHb
5 mL/100 mL
50% Anemia
(02rib Capacity = lOmL/'IOOmL)
0 20 40 60 80 100
P02 (mmHg)
Source: U.S. EPA (1991)
Figure 4-5. 02Hb dissociation curve of normal human blood, of blood containing 50% COHb,
and of blood with only 50% Hb because of anemia.
4.3.2.3. Heart and Skeletal Muscle
1 Mb is a globular heme protein that facilitates 02 diffusion from the muscle sarcoplasm to
2 mitochondria, acting as an 02 supply buffer to maintain adequate p02 for mitochondria when the 02
3 supply changes, as in exercise. 02 has a greater affinity for Mb than Hb, which allows small changes in
4 tissue p02 to release large amounts of 02 from 02Mb (Wittenberg et al., 1975). Small reductions in 02
5 storage capacity of Mb, due to CO binding, may have a profound effect on the supply of 02 to the tissue.
6 Like Hb, Mb will undergo reversible CO binding, however the affinity constant is approximately
7 eight-times lower than Hb (M = 20-40 versus 218, respectively) (Haab, 1990). The association rate
March 2009
4-12
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
constant of CO and Mb is approximately 27 times lower than 02, however the dissociation rate constant is
approximately 630 times lower than 02 (Gibson et al., 1986) causing CO to be retained and possibly
stored in the muscle. CO levels have been measured in human muscle and heart tissues with less than 2%
COHb concentrations at background levels averaging 15 and 31 picomole (pmol) CO/100 grams ww,
respectively (Vreman et al., 2006). Under conditions of CO asphyxiation, tissue concentrations increased
to 265 and 527 pmol CO/100 grams ww muscle and heart tissue, respectively; however, heart tissue
concentrations varied widely between individuals (See Table 4-la). 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). COMb
concentrations in the heart and skeletal muscle increase with work load, causing an increase in
COMb/COHb that is not as greatly seen at rest (Sokal et al., 1984). 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).
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-02 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). This value is estimated to be 0.3-0.4% min"1 or 0.24 mL/min (Bruce and
Bruce, 2003; Prommer and Schmidt, 2007). The equilibration rate from blood to tissue is controversial.
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). Tables 4-la and 4-lb contain tissue CO concentrations from human and mouse under
different CO exposure conditions. The distribution of CO between the different 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). These results are in conflict with older papers suggesting that negligible
retention of CO occurs in the liver or brain (Sokal et al., 1984; Topping, 1975).
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 short-term CO exposure (Figure 4-6)
(Bruce and Bruce, 2006; Shimazu et al., 2000; Wagner et al., 1975). This event is modeled by a two-
March 2009
4-13
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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). The biphasic curve is
more obvious after short-term CO exposure (less than 1 h), whereas long-term CO exposure (5 h or more)
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). Differences in elimination kinetics could also be a result of the variation
in CO exposure duration (Weaver et al., 2000).
The elimination of COHb is affected by a number of factors, including duration of exposure, Pa02,
minute ventilation, the time post-exposure for analysis due to extravascular stores, as well as inter-
individual variability (Bruce and Bruce, 2006; Landaw, 1973; Shimazu, 2001). 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). In addition, in a series of poisoned patients, 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). COHb
elimination half-life falls as the fractional inspired 02 concentration increases. While breathing air at sea
level pressure, the expected half-life in adult males is approximately 285 minutes, but may be shorter in
adult females. With inhalation of normobaric 40% 02, the half-life falls to 75 minutes and futher to 21
minutes when breathing 100% 02 because of greater competition for Hb by 02 (Landaw, 1973). Another
study reports the half-life falls to 74 minutes (mean) after breathing 100% 02, although the range in this
particular study was 26 to 148 minutes (Weaver et al., 2000). In addition, COHb half-life will fall further
after normocapnic hyperoxic hyperpnea (i.e., hyperventilation while maintaining normal C02 pressure in
high 02) (Takeuchi et al., 2000).
March 2009
4-14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
100
— — Short-term
Long-term
£ 10 ¦
O
o
0
30
60
90
120
150
180
Time (min)
Adapted from Shimazu et al. (2000)
Figure 4-6. Changes in blood COHb after short-term and long-term exposure to CO,
representing the biphasic nature of CO elimination. Note: y-axis is log-scale.
4.4. Conditions Affecting Uptake and Elimination
4.4.1. Environment and Activity
Elevated CO exposure and COHb levels are dependent upon the changes in CO concentration in
the local environment. Pedestrians are exposed to high levels of CO for short time periods from vehicle
exhaust at busy intersections (see also Chapter 3.6). Higher exposure can also result from riding in an
automobile or stopping at busy intersections (Ott, 1994). Indoor exposure occurs from ETS and unvented
combustion appliances, such as natural gas cooking stoves, attached garages, and gas fireplaces, which
can accumulate to over 100 ppm (Dutton et al., 2001). Recreational exposure at levels exceeding 200 ppm
and peaks of 1,600 ppm could occur in indoor ice rinks using fossil fuel powered ice resurfacers and
coliseums housing malfunctioning equipment or poor ventilation (Levesque et al., 2000; Pelham et al.,
2002). Certain occupations provide instances and conditions for transient moderate-to-high CO levels,
including fire fighters and machinery operators. Such transient exposures have the ability to increase
COHb levels. For example, exposure for 5 minutes or less of a resting individual to 6,600 ppm CO will
result in up to 20% COHb (Benignus et al., 1994).
Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase in
gas exchange. 02 consumption can increase more than 10 fold during exercise. Similarly, ventilation,
March 2009
4-15
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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.
The COHb elimination rate decreases with physical activity (Joumard et al., 1981). Healthy
subjects exposed to CO and achieving COHb levels of approximately 4-5% observed a significant
detriment to exercise duration and maximal effort capability (measured by metabolic equivalent units)
(Adir et al., 1999). It is possible that CO lowers the anaerobic threshold, allowing earlier fatigue of the
skeletal muscles and decreased maximal effort capability.
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 AQCD and
other documents (U.S. EPA, 1978). In an effort to maintain proper 02 transport and supply, physiological
changes occur as compensatory mechanisms to combat the decreased barometric pressure and resulting
altitude induced hyperbaric hypoxia (HH). HH, unlike CO hypoxia, causes humans to hyperventilate,
which reduces arterial blood C02 (hypocapnia) and increases alveolar partial pressure of 02. Hypocapnia
will lead to difficulty of 02 dissociation and decreased blood flow, thus reducing tissue 02 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 02 bind
reduced Hb through a competitive-parallel reaction (Chakraborty et al., 2004). Breathing CO (9 ppm) at
rest at altitude produced higher COHb compared to sea level (McGrath et al., 1993), whereas high altitude
exposure with exercise caused a decrease in COHb levels versus similar exposure at sea level (Horvath et
al., 1988). 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).
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 early adaptive changes gradually revert
to sea level values. However, differences in people raised at high altitude persist even after
reacclimatization to sea level (Hsia, 2002).
March 2009
4-16
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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; Klasner et al.,
1998). After infancy, the COHb half-life increases with age, nearly doubling between 2 and 70 years
(Joumard et al., 1981). The rate of this increase in CO elimination is very rapid in the growing years (2 to
16 years of age), but slows beyond adolescence. Alveolar volume and DLCO increase with increasing
body length of infants and toddlers (Castillo et al., 2006), 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).
COHb concentrations are generally lower in female subjects than in male subjects (Horvath et al.,
1988) 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). 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). Women also
tended to be more resistant to altitude hypoxia (Horvath et al., 1988).
Fetal CO pharmacokinetics do not follow the same kinetics as maternal CO exposure, making it
difficult to estimate fetal COHb based on maternal levels. Human fetal Hb has a higher affinity for CO
than adult Hb (Di Cera et al., 1989). Maternal and fetal COHb concentrations have been modeled as a
function of time using a modified CFK equation (Hill et al., 1977). At steady-state conditions, the fetal
COHb is up to 10% 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 5.5%. The fetal CO uptake lags behind the maternal for
the first few hours but later may overtake the maternal values. Similarly, during washout, the fetal COHb
levels are maintained for longer, with a half-life of around 7.5 hours versus the maternal half-life of
around 4 hours.
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 diffusion capacity, DLCO (Pesola et al., 2004; Pesola et al., 2006). 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).
March 2009
4-17
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 02 carrying capacity or
content will result in a greater risk from COHb induced hypoxia and decreased tissue 02 delivery. The
severity of this effect depends upon the initial level of hypoxia.
Anemias are a group of diseases that result in insufficient blood 02 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., 1976a, b). 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 02 and 02 carrying capacity is reduced causing a shift
to the right in the 02 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 (V/Q)
inequality (Marthan et al., 1985), which could slow both CO uptake and elimination. Patients with
pulmonary sarcoidosis may have a decrease in lung volumes, a loss of DLCO, and gas exchange
abnormalities during exercise, including decreased arterial oxygen pressure (Pa02) and increased alveolar-
arterial oxygen pressure difference (Lamberto et al., 2004).
Individuals with heart disease may be at a greater risk from CO exposure since they may already
have compromised 02 delivery. Time to onset of angina was reduced after exposure to 100 ppm carbon
monoxide, compared to clean air (Kleinman et al., 1998). Hyperlipidemic patients may have decreased
CO diffusion capacity, a loss of V/Q gradient, and a decrease in Pa02 (Enzi et al., 1976).
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. This is due to endogenous CO production from heme protein catabolism
among other sources. In the natural degradation of RBC Hb, the porphyrin ring of heme is broken at the
methene bridge and catabolized in an 02 dependent manner by HO complexed with NADPH-cytochrome
P450 reductase and biliverdin reductase to biliverdin, Fe, and CO. Biliverdin is then further broken down
into bilirubin, a powerful endogenous antioxidant. Two main HO isoforms exist, HO-1 and HO-2.
March 2009
4-18
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Expression of HO-1 is inducible, whereas HO-2 is constitutively expressed. The major site of heme
catabolism, thus the major organ of CO production is the liver, followed by the spleen, brain, and
erythropoietic system (Berk et al., 1976). 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 16.4 |imol/h (0.42 mL/h)
and produces between 400-500 |imol CO per day (Coburn et al., 1964; Coburn et al., 1966; Coburn,
1970b). The endogenous rate of CO formation has been shown to vary little between different tissues,
ranging from 0.029 nmol/mg protein/h in chorionic villi of term human placenta to 0.28 nmol/mg
protein/h in rat olfactory receptor neurons in culture and in rat liver perfusate (Marks et al., 2002),
however these estimations are questionable 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).
HO mediated metabolism functions as the rate-limiting enzyme step in heme degradation and
endogenous CO production (Wu and Wang, 2005). Three isoforms of HO exist, but HO-1 is the only
inducible form (Maines and Kappas, 1974; Maines et al., 1986; McCoubrey et al., 1997). 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). High levels of CO (2,500 ppm) have been shown to
increase HO-1 activity in the brain of rats, as well as liberate intracellular heme to further stimulate
endogenous CO production (Cronje et al., 2004).
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). All of these sources result in a normal blood COHb concentration between 0.4 and
0.7% (Coburn et al., 1965). This baseline level of endogenous production can be altered by drugs or a
number of physiological conditions that alter RBC destruction, other hemoprotein breakdown, or bilirubin
production. Because of this sensitivity, ranges of endogenous COHb levels in the population are
uncertain. Nicotinic acid (Lundh et al., 1975), allyl-containing compounds (acetamids and barbiturates)
(Mercke et al., 1975b), diphenylhydantoin (Coburn, 1970b), progesterone (Delivoria-Papadopoulos et al.,
1974), and contraceptives (Mercke et al., 1975a) will increase CO production. Compounds such as carbon
disulfide and sulfur-containing chemicals (parathion and phenyltiourea) will increase CO by acting on
P450 system moieties (Landaw et al., 1970). The P450 system may also cause large increases in CO
produced from the metabolic degradation of dihalomethanes leading to very high (>10%) COHb levels
(Manno et al., 1992), which can be further enhanced by prior exposure to HCs or ethanol (Pankow et al.,
1991; Wirkner and Poelchen, 1996). HO can catalyze the release of CO from the auto-oxidation of
phenols, photo-oxidation of organic compounds, and lipid peroxidation of cell membrane lipids (Rodgers
et al., 1994).
March 2009
4-19
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Any disturbance in RBC hemostasis by acceleration of destruction of hemoproteins will lead to
increased production of CO. 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; Hampson, 2007; Meyer et al., 1998; Solanki et al., 1988). Patients with hemolytic
anemia exhibit COHb levels 2- to 3-times higher than healthy individuals and CO production rates 2- to
8-times higher (Coburn et al., 1966). Endogenous CO production rate varied from 0.70 to 3.44 mL/h in
anemic patients (Coburn et al., 1965). Women experience fluctuating COHb levels during pregnancy as
well as through the menstrual cycle when endogenous CO production doubles in the progesterone phase
(Delivoria-Papadopoulos et al., 1974; Mercke and Lundh, 1976).
Altitude has been shown to be positively associated with baseline COHb concentrations (McGrath,
1992; 1993). This increase in COHb with altitude induced hypoxia has also been associated with
increases in the mRNA, protein, and activity of HO-1 in rats and cells leading to enhanced endogenous
CO production (Carraway et al., 2002; Lee et al., 1997). 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.
Endogenous CO is removed from the body mainly by expiration and oxidation. CO will diffuse
across the alveolar-capillary membrane and then is exhaled. This event has been used as a noninvasive
measurement of endogenous CO and CO body load (Stevenson et al., 1979). CO can also be oxidized to
C02 by cytochrome c oxidase in the mitochondria (Fenn, 1970; Young and Caughey, 1986). 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).
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 02-carrying capacity of blood and
impairs the release of 02 from 02Hb to the tissues. COHb formation has been modeled mainly by the
CFK equation, but more recent models have included Mb and extravascular storage compartments, as
well as other dynamics of CO physiology. 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 CO
concentration 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
March 2009
4-20
DRAFT - DO NOT CITE OR QUOTE
-------
1 compromised individuals and developing fetuses, are at a greater risk from COHb induced health effects
2 due to altered CO kinetics, compromised cardiopulmonary processes, and increased baseline hypoxia
3 levels. Altitude may also significantly affect the kinetics of COHb formation. Compensatory mechanisms,
4 such as increased cardiac output, combat the decrease in barometric pressure. Altitude also increases the
5 endogenous production of CO through upregulation of HO-1. CO is considered a second messenger and
6 is endogenously produced from the catabolism of heme proteins by enzymes such as HO-1. Finally, CO is
7 removed from the body by expiration and oxidation to C02.
March 2009
4-21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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
mechanisms unrelated to tissue hypoxia including 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, 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 02-
carrying capacity of blood, but it also impairs the release of 02 from 02Hb. Compensatory alterations in
hemodynamics, such as vasodilation and increased cardiac output, protect from tissue hypoxia. At
ambient CO concentrations, these compensatory changes are slight and likely tolerable in people with a
healthy cardiovascular system. However, people with cardiovascular detriments may be unable to endure
these small changes in hemodynamics which may lead to the presentation of health effects as described in
Sections 5.2.1 and 5.2.2.
The 2000 CO AQCD reported changes in vasodilation due to CO levels between 500-2000 ppm
(Kanten et al., 1983; MacMillan, 1975). Vasodilation can be dependent on or independent of perturbations
in 02 supply (Koehler et al., 1982). For example, cerebral blood flow elevations that were independent of
02 availability were blocked by the inhibition of nitric oxide synthase (NOS) indicating a role for the free
radical species nitric oxide (NO*) in mediating vasodilation (Meilin et al., 1996).
Increased cardiac ouput was also discussed in the 2000 CO AQCD as a compensatory response to
CO-mediated tissue hypoxia. Findings of studies measuring hemodynamic alterations following CO
exposure were equivocal and sometimes contradictory (Penney, 1988). 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 resulting in COHb
levels around 9% (Ayres et al., 1973). There was no increase in cardiac output following acute exposure
March 2009
5-1
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
to 0.1% CO in this latter study. Increased heart rate and stroke volume (SV) were also observed in
response to CO exposure (Stewart et al., 1973b); however, some experiments found no change in SV in
humans with 18-20% COHb (Vogel and Gleser, 1972) or 12.5% COHb (Klausen et al., 1968). The 2000
CO AQCD reported that BP was generally unchanged in human CO exposure studies, while a number of
animal studies demonstrated CO-induced hypotension (Penney et al., 1988). No changes in forearm blood
flow, BP, or heart rate were reported in humans experiencing approximately 8% COHb (Hausberg and
Somers, 1997). However, high concentration animal exposures (3,000 to 10,000 ppm) showed diminished
organ blood flow (Brown and Piantadosi, 1992). In depth discussion of hemodynamic changes resulting
from CO exposure in recent human clinical studies can be found in Section 5.2.2.
Binding of CO to Mb, as discussed in the 2000 CO AQCD and in Section 4.3.2.3, can also impair
the delivery of 02 to tissues. Mb has a high affinity for CO; however, physiological 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).
Conversely, rat hearts perfused with 6% CO (60,000 ppm) exhibited no change in high energy phosphate
production, respiration rate, or contractile function (Chung et al., 2006; Glabe et al., 1998). These
conflicting studies employed CO levels that are not relevant for ambient exposure to CO.
5.1.3. Non-Hypoxic Mechanisms
Non-hypoxic mechanisms underlying the biological effects of CO were discussed in the 2000 CO
AQCD 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). 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 levels of CO.
5.1.3.1. Non-Hypoxic Mechanisms Reviewed in the 2000 CO AQCD
Inhibition of heme-containing proteins such as cytochrome c oxidase and cytochrome P450
oxidases 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). 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 are above 50% (Brown and Piantadosi, 1992). It is unlikely that this could arise
under physiological conditions or under conditions relevant to ambient exposures.
March 2009
5-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Studies indicated that CO exposure produces a pro-oxidant cellular environment by a second
mechanism. CO exposure at concentrations of 10-20 ppm and above caused platelets of laboratory rats, as
well as cultured bovine pulmonary endothelial cells, to release NO* (Thorn and Ischiropoulos, 1997). An
increase in available NO' was also seen in the lung and brain of CO-exposed rats (Ischiropoulos et al.,
1996; Thorn et al., 1999b). NO* combines with superoxide to form the highly active oxidant species,
peroxynitrite (Thorn et al., 1997), which can lead to the activation and sequestration of leukocytes in brain
vessels (Thorn et al., 2001b) and aorta, oxidation of plasma lipoproteins (Thorn et al., 1999a), and the
formation of protein nitrotyrosine (Ischiropoulos et al., 1996; Thorn et al., 1999a; Thorn et al., 1999b).
The mechanism leading to NO* release is likely the displacement of NO' from nitrosyl bound Mb (NO*-
Mb) or Hb (NO'-Hb) by CO. The rate of this event is slow; however it occurs at environmentally relevant
concentrations of CO (Thorn et al., 1997).
CO exposure also increases the production of other pro-oxidant species, including hydrogen
peroxide (H202) and hydroxyl radical (OH ). High level CO exposure (2,500 ppm) increases OH* in rat
brain and this response is distinct from tissue hypoxia (Piantadosi et al., 1997). The mechanism for
enhanced H202 production is unclear. The release of H202 in the lung of CO-exposed rats was dependent
upon the production of NO*, as it was inhibited by the pretreatment with a NOS inhibitor (Thorn et al.,
1999b). It is possible that peroxynitrite formed after CO exposure inhibited electron transport at
complexes I through III, or that cytochrome c oxidase interference led to mitochondrial dysfunction and
ROS production.
It has also been suggested that CO leads to vasorelaxation by three different mechanisms. First, CO
may inhibit the P450-dependent synthesis of vasoconstrictors (Wang, 1998). Vasodilation has been
demonstrated via this P450-mediated mechanism with high concentrations of CO (approximately
90,000 ppm) (Coceani et al., 1988). In the case of cytochrome P450 enzymes, tissue CO levels may have
to be abnormally increased to elicit a physiological response since the in vitro Warburg binding
coefficients, the ratio of CO to 02 at which half the reactive sites are occupied by CO, for cytochrome
P450s range from 0.1-12 (Piantadosi, 2002). P450 inhibition may also 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). CO may also interfere with the metabolism of
barbiturates and other drugs; however, this is probably due to the hypoxic actions of CO rather than to
P450 inhibition (Roth and Rubin, 1976a, b).
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 interacts with sGC and forms an unstable complex with the heme core of the protein.
The resulting porphyrin IX intermediate triggers cGMP production (Ndisang et al., 2004). CO causes
vascular relaxation, independent of the endothelium, in human arterial rings (Achouh et al., 2008), rat tail
March 2009
5-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
artery (Wang et al., 1997), and rat thoracic aorta (Lin and McGrath, 1988), but not cerebral vessels
(Andresen et al., 2006; Brian et al., 1994). Activation of sGC by CO has been linked to
neurotransmission, vasodilation, bronchodilation, inhibition of platelet aggregation, and inhibiton of
smooth muscle proliferation (Brune and Ullrich, 1987; Cardell et al., 1998; Morita et al., 1997; Verma et
al., 1993).
CO-mediated vasorelaxation can also be caused by activation of voltage-activated potassium (K+ )
channels or calcium (Ca2+)-activated K+ channels, which leads to membrane hyperpolarization, voltage-
dependent Ca2+ channel closing, reduction of resting Ca2+ concentration, and then vascular tissue
relaxation (Farrugia et al., 1993; Wang et al., 1997). This effect may be linked to sGC activity; however it
has also been reported to occur independently (Dubuis et al., 2003; Naik and Walker, 2003).
Developmental stage and tissue type will determine whether K+ channels or the sGC/cGMP pathway
plays more of a role in vasorelaxation (Ndisang et al., 2004).
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 pro-oxidant status following cytochrome c oxidase inhibition would likely
predispose towards cellular injury and death. The increase in free NO* following its release from
sequestered stores could also contribute to the pro-oxidant status. 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 one-thousand fold more potent with respect to vasodilation and sGC activation (Stone
and Marietta, 1994). CO could further contribute to endothelial dysfunction by this mechanism. Although
the 2000 CO AQCD made no definitive links between these non-hypoxic mechanisms of CO and
CO-mediated health effects, it did establish the potential for CO to interfere with basic cellular and
molecular processes which could lead to dysfunction and/or disease.
5.1.3.2. Recent Studies of Non-Hypoxic Mechanisms
Since the 2000 CO AQCD, new studies have provided additional evidence for non-hypoxic
mechanisms of CO which involve the binding of CO to reduced iron in heme proteins. These
mechanisms, which may be inter-related, are described below and include:
March 2009
5-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
¦ Alteration in NO* signaling
¦ Inhibition of cytochrome c oxidase
¦ Heme loss from protein
¦ Disruption of iron homeostasis
¦ Alteration in cellular redox status
This assessment evaluates these non-hypoxic 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 these
can potentially contribute to adverse health effects.
In addition, a large number of studies published since the 2000 CO AQCD has focused on the role
of COs as an endogenous signaling molecule and the potential therapeutic effects of exogenous CO
administered at high concentrations. This assessment addresses these topics only briefly, as they pertain to
the evaluation of health effects associated with environmentally-relevant CO exposures.
Alteration in NO* Signaling
Work by Thorup et al. (1999) demonstrated altered NO* signaling in isolated rat renal resistance
arteries. In one set of experiments, rapid release of NO* was observed in response to a supra-physiologic
dose of CO (100 nM). These findings are similar to those of Thorn and colleagues (Ischiropoulos et al.,
1996; Thorn et al., 1994; Thorn and Ischiropoulos, 1997; Thorn et al., 1997; Thorn et al., 1999a; Thorn et
al., 1999b; Thorn et al., 2000; Thorn et al., 2006) 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) demonstrated inhibition of NOS in
isolated rat renal resistance arteries. Here rapid NOS-dependent release of NO* following carbachol
stimulation was blocked by pretreatment with 100 nM NO*. Both sets of studies illustrate the potential of
CO to alter processes dependent on endogenous NO*. This could be critical in the case of cGMP-
mediated vasodilation since, as discussed above, NO* activates sGC to a greater extent than CO. Thus in
the presence of excess CO, NO'-dependent vasodilation may be significantly less.
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 cellular 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
March 2009
5-5
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
a study by Alonso et al. (2003), 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). In this
study, increased ROS were observed following exposure to 250 ppm, 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., 2006a) and exposure of mice
to 1,000 ppm CO for 3 h decreased cytochrome c oxidase activity in heart mitochondria (Iheagwara et al.,
2007). COHb levels were reported to be 61% in the latter study.
Heme Content Loss from Proteins
In addition to decreasing the activity of cyctochrome 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 (Iheagwara et al.,
2007). CO-mediated heme release was also seen in a study by Cronje et al. (2004), 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 (Ghio et al.,
2008). Furthermore in this same study, exposure of 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 observed in other
studies (Cronje et al., 2004; Iheagwara et al., 2007) might be expected to result in disruption of iron
homeostasis. Iron homeostasis is critical for the sequestration of free iron and the prevention of iron-
mediated redox cycling which can lead 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 oxidant stress in cultured cells (Kim et al., 2008; Zuckerbraun et al., 2007). A dose-dependent
increase in dichlorofluoroscein (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). 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). Recently, inhibition of cytochrome c oxidase was demonstrated
March 2009
5-6
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
in HEK-293 cells transfected with HO-1 and in macrophages with induced HO-1, and this effect was
attributed to endogenously produced CO (D'Amico et al., 2006). However inhibition of cytochrome c
oxidase at CO concentrations below 250 ppm has not been definitively demonstrated. 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). Significant
increases in apoptosis were also observed in this model. Thus, CO appeared to uncouple mitochondrial
respiration leading to 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; Patel et al., 2003). In addition, mitochondrial
redox stress was observed in livers of rats exposed to 50 ppm CO (Piantadosi et al., 2006). 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 (Thom et al., 2000) and mitochondrial biogenesis was
observed in hearts of mice exposed to 50 and 250 ppm CO for 1 h (Suliman et al., 2007).
Several mechanisms could contribute to the cellular redox stress elicited by CO exposure. As
discussed above, inhibition of cytochrome c oxidase results in the formation of superoxide from
mitochondrial substrates. However, interactions of CO with heme proteins can lead to the release of heme
and free iron which could also lead 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) and the cardiac mitochondrial biogenesis observed in mice exposed to CO
(Suliman et al., 2007) could be attributed specifically to non-hypoxic mechanisms of CO.
Recent studies also demonstrated non-hypoxic mechanisms of CO which do not directly involve
heme protein interactions. These mechanisms are described below and include:
¦ Alteration in ion channel activity
¦ Modulation of protein kinase pathways
Alteration in Ion Channel Activity
Work by Dubuis et al. (2002) demonstrated increased current through Ca2+-activated K+ channels in
smooth muscle cells from pulmonary arteries of rats exposed to 530 ppm CO for 3 weeks. These findings
provide further evidence for non cGMP-dependent vasodilatory actions of CO.
March 2009
5-7
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 (Ryter et al., 2006). These studies generally involved pretreatment with CO followed
by exposure to another agent 18-24 h later. There is extensive literature on this topic as reviewed by Ryter
et al. (2006) 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). A second example is provided by the study of Kim et al. (2005) where 250 ppm CO
inhibited PDGF- induced smooth muscle cell proliferation by upregulating p2iWafl/Cipl and caveolin-1, and
down-regulating cyclin A expression. These events were dependent upon cGMP and the p38 MAPK
pathway (Kim et al., 2005). Thirdly, rat endothelial cells exposed to 15 ppm CO escaped anoxia-
reoxygenation mediated 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). 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). While research focusing
on therapeutic applications of CO generally involves high level, short-term exposure to CO (250-
1,000 ppm for up to 24 h), some studies found effects below 20 ppm (Zhang et al., 2005). Few if any
studies on the therapeutic effects of CO have explored the dose-response relationship 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 non-hypoxic mechanisms is
much higher than what would be expected to result from exposures at ambient levels (Alonso et al., 2003;
Favory et al., 2006b; Iheagwara et al., 2007; Thorup et al., 1999). However in some studies the effects are
mimicked by upregulation of HO-1 which would result in increased local production of CO as well as of
iron and biliverdin (DAmico et al., 2006; Imai et al., 2001; Thorup et al., 1999). For example, HO-1
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; Thorup et al., 1999). In the study by
D'Amico et al. (2006), overexpression of HO-1 in cells inhibited cellular respiration by 12% and
decreased cytochrome c oxdase activity by 23%. It is not clear how comparable these conditions
March 2009
5-8
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
involving increased intracellular concentrations of endogenous CO are to increased intracellular
concentrations of CO resulting from ambient CO exposures.
There is a growing appreciation that non-hypoxic mechanisms may contribute to the effects
associated with CO poisoning (Ischiropoulos et al., 1996; Thorn et al., 1994; Weaver et al., 2007). On the
other hand, recent studies suggest that exogenous CO at low concentrations may have beneficial anti-
inflammatory, anti-proliferative and cytoprotective effects under certain circumstances (Ryter et al.,
2006). 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"
(50-500 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 non-hypoxic mechanisms are
demonstrated at CO concentrations of 50 ppm and less.
Table 5-1.
Responses to low and moderate CO exposures.
Reference
Model System
CO Exposure
Response
Notes
IN VITRO
Alonso et al.
(2003)
Human muscle
mitochondria
50,100, 500 ppm
5 min
Decreased cytochrome c oxidase activity
Thorn and
Ischiropoulos
(1997)
Rat
platelets
10 ppm
Increased free NO'
Thorn et al.
(1997)
Bovine pulmonary artery
endothelial cells
20 ppm
30-60 min
Increased free NO' and peroxynitrite
Reported to correspond to 7% COHb
Thorn et al.
(2000)
Bovine pulmonary artery
endothelial cells
10 ppm
40 min
Increased MnSOD and protection against toxic
effects of 100 ppm CO
Adaptive responses
Song et al.
(2002)
Human aortic smooth
muscle cells
50-500 ppm
24 h
Inhibition of cellular proliferation
Blocked activation of ERK1/2
pathway, independent of sGC and
other MAP kinases
Kim et al. (2005)
Rat pulmonary artery
smooth muscle cells
250 ppm
1 h
Inhibited PDGF- induced smooth muscle cell
proliferation
Upregulated p21Wa™'P1 and caveolin-
1, and down-regulated cyclin A
expression.
Kim et al. (2008)
Rat hepatocytes
250 ppm
1 h
2x per day
250 ppm
1 h
Blocked spontaneous apoptosis
Increased mitochondrial ROS generation, increased
mitochondrial glutathione oxidation, and decreased
cellular ascorbic acid
CO induced Akt phosphorylation via
ROS production
CO activated NFkB
Zhang et al.
(2005)
Rat pulmonary artery
endothelial cells
15 ppm
0.5-24 h
Blocked anoxia-reoxygenation mediated apoptosis
Modulation of PI3K/Akt/p38 MAP
kinase and STAT-1 and STAT-3
Zuckerbraun et
al. (2007)
Mouse macrophages
50 and 250 ppm
1 h
Increased ROS generation (dose dependent
response for 50-500 ppm CO)
Mitochondrial derived ROS and
cytochrome c oxidase inhibition
demonstrated for 250 ppm
March 2009
5-9
DRAFT - DO NOT CITE OR QUOTE
-------
Reference
Model System
CO Exposure
Response
Notes
Ghio et al. (2008) Human bronchial
epithelial cells
10-100 ppm
24 h
Dose-dependent decrease in cellular non-heme iron
(effect at 10 ppm was signficant, effect at 50 ppm
maximal)
Dose-dependent decrease in cellular ferritin at 50-
100ppm
50 ppm blocked iron uptake by cells
50 ppm increased iron release from cells
Compare with in vivo experiment in
same paper
IN VIVO
Ghio et al. (2008) Rats
50 ppm
24 h
Mild neutrophil accumulation in BALF
Increased lavage MIP-2, protein, LDFI
Lavage iron and ferritin were increased by CO
Lung non-heme iron was decreased by CO
Liver non-heme iron was increased by CO
Thorn et al.,
(1999b)
Rats
50 ppm
1 h
100 ppm
1 h
Increased nitrotyrosine in aorta
Leukocyte sequestration in aorta after 18 h
Albumin efflux from skeletal muscle
microvasculature 3 h after CO
LDL oxidation
Effects blocked by NOS inhibitor
Thorn et al.,
(1999a)
Rats
100 ppm
1 h
50 ppm
1 h
Elevated free NO' during CO exposure (EPR)
Elevated nitrotyrosine in lung homogenates
Lung capillary leakage 18 h after exposure
Inhibition of NOS abrogated CO
effects
Sorhaug et al.
(2006)
Rats
200 ppm
72 weeks
No changes in lung morphology
No pulmonary hypertension
No atherosclerotic lesions in systemic vessels
Ventricular hypertrophy
Loennechen et
al. (1999)
Rats
100 and 200 ppm
1-2 weeks
Increased ET-1 mRNA in the heart ventricles,
increased right and left ventricular weight
12 and 23% COHb
Favory et al.
(2006b)
Rats
250 ppm
90 min
Complex IV inhibition in myocardial fibers
Inhibition ofvasodilatory response to acetylcholine
and SNP, Increased coronary perfusion pressure
and contractility
11% COFIb
Piantadosi et al.
(2006)
Rats
50 ppm CO or FIFI
for 1, 3, or 7 days
Liver mitochondrial oxidative and nitrosative stress,
altered mitochondrial permeability pore transition
sensitivity
CO effects not mimicked by FIFI
Suliman et al.
(2007)
Mice
50 and 250 ppm
1 h
Cardiac mitochondrial biogenesis
Activation of GC involved. No role for
NOS. Increased mitochondrial FI2O2
and activation of Akt proposed
Wellenius et al.
(2004)
Rats
Model of Ml
35 ppm
1 h
Decreased delayed ventricular beat frequency
Altered arrhythmogenesis
Wellenius et al.
(2006a)
Rats
Model of Ml
35 ppm
1 h
Decreased supraventricular ectopic beats
Altered arrhythmogenesis
Carraway et al.
(2002)
Rats
Model of hypoxic
pulmonary vascular
remodeling
FHypobaric hypoxia
± 50 ppm CO
3 weeks
CO promoted remodeling and increased pulmonary
vascular resistance
Gautier et al.
(2007)
Rats
Model of right ventricle
hypertrophy secondary to
chronic hypoxia
3 weeks of FIFI with
50 ppm CO during
last week
Rats with pulmonary hypertension were more
sensitive to CO which altered the right ventricular
adaptive response to pulmonary hypertension
leading to ischemic lesions
Melin et al.
(2005)
Rats
Model of right ventricle
hypertrophy secondary to
chronic hypoxia
50 ppm
10 wk
CO increased cardiac dilation and decreased left
ventricular function
March 2009
5-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Reference
Model System CO Exposure
Response
Notes
Melinetal. Rats 50 ppr
(2002) Model of right ventricle 10 wk
hypertrophy secondary to
chronic hypoxia
50 ppm
CO increased right ventricular hypertrophy
decreased right ventricular diastolic function and
increased left ventricular weights
5.1.3.3. Implications of Non-Hypoxic Mechamisms
A key issue in understanding the biological effects of environmentally-relevant exposures to CO is
whether the resulting partial pressures of CO in cells and tissues can initiate cell signaling or perturb
signaling which normally is mediated by endogenously generated CO.
Several aspects need to be considered. One is that, for most cells and tissues, Hb acts as a buffer to
limit the availability of free CO. At the same time, COHb delivers CO to cells and tissues. CO
dissociation from Hb occurs as a function of CO diffusion down a concentration gradient. Hence, greater
release of CO from COHb will occur under conditions of low cell/tissue CO concentration.
A second consideration is the role played by 02 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 02 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).
A third consideration is whether certain cell types serve as primary targets for the effects of CO.
Besides the RBC, the first cells encountering CO which dissociates from Hb will be the endothelial cells
lining blood vessels. White blood cells in the circulation may also be first-line targets of Hb-dissociated
CO. 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 endothelial dysfunction, inflammatory cell activation and
respiratory epithelial injury may ensue as a result of preferential targeting of these cell types.
Fourth, it should be considered that adaptation to chronic exogenous CO exposure might occur and
that intermittent exogenous CO exposure might have unexpected effects.
Since there is potential for exogenous CO to affect endogenous pools of CO, it is important to
know the concentrations of CO in cells and tissues before and after exogenous exposures. Table 5-2
summarizes findings from 4 recent studies relevant to this issue. It should be noted that exposure to
50 ppm CO resulted in a 3-5 fold increased in tissue CO concentration.
March 2009 5-11 DRAFT-DO NOT CITE OR QUOTE
-------
Table 5-2. Tissue concentration of CO following inhalation exposure.
Reference
CO Exposure
Tissue CO Concentrations
COHb
Notes
Cronje, et al. (2004)
Rat
2,500 ppm
45 min
Blood: 27,500 (800) pmol/mg
Heart: 800 (300) pmol/mg
Muscle: 90 (80) pmol/mg
Brain: 60 (40) pmol/mg
Control levels in parentheses
66-72%
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)
Vreman, et al. (2005)
Mice
500 ppm
30 min
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 ± (4) pmol/mg
Testes: 6 ± 3 (2) pmol/mg
Control levels in parentheses
28%
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%
Piantadosi et al. (2006) Rats
50 ppm
1-7 days
Liver: 30-40 pmol/mg
Control liver 10 pmol/mg
4-5%
Control 1 %
CO concentration plateaued after 1 day
Suliman et al. (2007)
Mice
50, 250 and
1250 ppm
1 h
Heart (left ventricle)
50 ppm: 50 pmol/mg
250 ppm: 95 pmol/mg
1250 ppm: 160 pmol/mg
Control heart: 9 pmol/mg
No mention of COHb% but exposures were similar to
those in Cronje et al. (2004)
Furthermore, endogenous CO production is increased during inflammation, hypoxia, increased
heme availability and other conditions where HO-1 or HO-2 activity is increased. A few studies reported
increased COHb levels and/or cell and tissue concentrations of CO resulting from enhanced endogenous
CO production. Table 5-3 summarizes these findings.
Table 5-3. Tissue concentration of CO following increased endogenous production.
Reference Exposure
Tissue CO
COHb
Notes
Carrawayet al. (2000) Rats
Hypobaric Hypoxia for 21 days
1.5-2.8%
Control 0.5%
COHb highest after days 1 and 21
at 3-4 fold higher than controls
Piantadosi et al. (2006) Rats
Hypobaric Hypoxia
1-7 days
Liver: 5-12 pmol/mg
Control liver 10 pmol/mg
1-1.25%
Control 1%
CO concentration plateaued after 1 day
Vreman et al. (2005) Mice
30 pM heme
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
Control levels in parentheses
0.9%
CO concentration relative to 100% blood:
Heart: 16%
Spleen: 13%
Lung: 9%
Liver: 9%
Kidney: 8%
Muscle: 8%
Intestine: 3%
Brain: 2%
Testes: 2%
March 2009
5-12
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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, 2008), which could impact interpretation of results from studies in which HO-1
and HO-2 activities are increased. Hence the situations of increased endogenous CO production and that
of exogenous CO exposure are not equivalent. Nonetheless, additional measurements of CO levels in cells
and tissues following increased endogenous production and following inhalation of exogenous CO is
essential for a better understanding of the relationship between the CO tissue dose and the response.
In summary, CO is a ubiquitous cell signaling molecule and the physiological functions of HO-
derived CO are numerous. The endogenous generation and release of CO from HO-1 and HO-2 is tightly
controlled, as is any homeostatic process. Thus, exogenously-applied CO has the capacity to disrupt
myriad 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.
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 can affect
endogenous CO signaling pathways and lead to adverse health effects is an open question for which there
are no definitive answers at this time.
5.2. Cardiovascular Effects
This section characterizes the evidence from epidemiologic, controlled human exposure and animal
toxicological studies on the cardiovascular effects of CO. While epidemiologic studies evaluated the
effects of ambient exposures, experimental studies employed higher than ambient concentrations of CO
but not levels of exposure associated with poisoning.
5.2.1. Epidemiologic Studies
The 2000 CO AQCD concluded that epidemiologic studies provided evidence that short-term
variations in ambient CO concentrations were associated with daily hospital admissions for heart disease.
The following section reviews the literature since the 2000 CO AQCD, including new studies on
physiological cardiac endpoints and biomarkers and additional studies of daily hospital admissions for
heart disease that support the evidence in the 2000 CO AQCD.
March 2009
5-13
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
5.2.1.1. Epidemiologic Studies with Short-Term Exposure
Heart Rate and Heart Rate Variability
Heart rate variability (HRV) refers to the beat-to-beat alterations in the heart 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 (De
Jong and Randall, 2005; Deedwania et al., 2005; Huikuri et al., 1999; Rajendra Acharya et al., 2006).
Two 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 twenty-one 53-to
87-year old active residents and performed up to 12 ECG assessments on each subject over a period of 4
months (during summer 1997). Particles (PMi0, PM2 s) and several gaseous pollutants (03, N02, and S02)
were monitored at fixed sites (up to 4.8 miles from the study site) while CO was monitored 0.25 miles
from each participants' 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 03 were associated with HRV
parameters (Gold et al., 2000).
A similar study by the same group of researchers two years later involved 28 older subjects (aged
61-89) 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 weeks and HRV parameters (SDNN,
r-MSSD, PNN50, LF/HF ratio) were measured for 30 minutes each session. Data for PM25, black carbon
(BC), and CO were recorded at the Harvard School of Public Health (<1 km from the residence) while
data for N02, 03, and S02 were collected from government regulatory monitoring sites. There were
moderate correlations between CO and PM2 5 (r = 0.61) and N02 (r = 0.55), but not with S02 (r = 0.18) or
03 (r = 0.21). Similarly PM2 5 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).
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).
March 2009
5-14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
A later 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 to October 2003)
and the ECG data were linked to air pollution data for PM2.5, particle number concentration, BC, 03, N02,
S02, and CO. Lagged pollutant effects for a 4-h, 24-h, and 48-h moving avg were used. Since previous
studies established variable CO results, the main pollutant effects were with PM2 5 and 03 while CO was
not associated with HRV (Park et al., 2005b).
A study in Mexico City selected 30 subjects from the outpatient clinic at the National Institute of
Cardiology and followed them for -10 hours (starting at 0900 hrs) (Riojas-Rodriguez et al., 2006). Each
subject was connected to a Holter ECG monitor (e.g., a portable ECG monitor) and also given personal
PM2 5 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 (heart rate [HR], R-R, LF, HF) at various lags. In
copollutant models (PM2 5 and CO) 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 Mexico City, 34 residents from a nursing home underwent HRV analysis every other day for 3
months (Holguin et al., 2003). Exposure assessment for ambient PM2 5 was based on data recorded at a
monitor on the roof of the nursing home while exposures to ambient 03, N02, S02, 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 03 and PM2 5 were positively associated with HRV.
Wheeler et al. (2006) 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 PM2 5, 03, N02, S02 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 N02 and PM2 5, no quantitative results were reported for CO.
After reviewing 2,000 patient charts, Dales (2004) 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 weeks by a Holter monitoring system. Personal
air sampling for PM2 5 and CO was carried out for the same 24-h period whereby 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.
March 2009
5-15
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
HR was examined among a sub-sample of the Monitoring of Trends and Determinants in
Cardiovascular Disease (MONICA) study (n = 2,681) in Augsburg, Germany by Peters and colleagues
(1999a). Total suspended particles (TSP), S02, 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.
Liao et al (2004) investigated men and women aged 45-64 years 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
PMio, 03, CO, N02, and S02 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 Helsinki,
Finland, whereby a panel of subjects with CAD was followed for 6 months with biweekly clinical visits,
which included an ECG reading to assess HRV (Timonen et al., 2006). 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, PMi0, N02, CO) were
derived from data recorded at fixed monitoring site networks within each city. Correlation coefficients for
N02 and CO ranged from 0.32 to 0.86 in the three cities. CO was moderately correlated with PMi0 in
Helsinki (r = 0.40) and with PM2 5 in Amsterdam (r = 0.58) and more highly correlelated with PMi0 in
Erfurt (r = 0.77). Various lag periods were examined including lag 0 (24 h prior to the clinical visit)
through a 0-2-day avg lag and a 0-4-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).
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). The patients were asked to follow their usual daily activities,
but data were recorded only three times with 1-week 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.
March 2009
5-16
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
A study in Taiwan recruited 83 patients (aged 40-75) from the National Taiwan University
Hospital, Taipei and conducted ambulatory ECG readings using a Holter system (Chan et al., 2005).
Ambient air pollution exposures for PMi0, N02, S02, 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 N02
was associated with HRV parameters (SDNN and LF). CO was not associated with HRV.
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. Gold et al. (2005) recruited a
panel of 28 older adults living at or near an apartment complex located within 1 km of a monitoring site in
Boston, MA. Each subject underwent weekly ECGs for 12 weeks in summer 1999 with the main outcome
of interest being the ST-segment. Air pollution data in the form of PM25, black carbon (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 (1-24 h, 12 h and 24 h moving average [ma]) before the ECG. The final analyses included 24
subjects with 269 observations and results showed consistent negative associations of ST-segment level
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 depression, however only BC remained significant in
multipollutant models.
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 shows a summary of the reviewed studies.
Table 5-4. Summary of studies investigating the effect of CO exposure on HRV parameters.
Study
Location (Sample Size)
Cardiac Endpoint
Exposure
Assessment
Mean CO Level
(ppm)
Copollutants
Gold et al. (2000)
Boston, MA
(n = 21)
HR, SDNN,
r-MSSD
Ambient
Mean: 0.47(24 h)
Range: 0.12-0.82
PM10, PM25, O3, NO2,
S02
Schwartz et al.
(2005)
Boston, MA
(n = 28)
SDNN,
r-MSSD, PNN, LF/HF
Ambient
25th, 50th, 75th
percentiles:
0.38, 0.45, 0.54
PM2.5, BC, N02, 03
Park et al.
(2005b)
Boston, MA
(n = 4 97)
SDNN, LF, HF, LF/HF
Ambient
Mean: 0.50 (24 h)
Range: 0.13-1.8
PM2.5, BC, 03, N02, S02
Riojas-Rodriguez
et al. (2006)
Mexico City, Mexico
(n = 30)
HF, LF, VLF, HR, R-R
Personal
Mean: 2.9 (11 h)
Range: 0.1-18
PM2.5
Holguin et al.
(2003)
Mexico City, Mexico (n = 34)
HF,Double check
LF, LF/HF
Ambient
Mean: 3.3(24 h)
Range: 1.8-4.8
PM2.5,03, NO2, SO2
March 2009
5-17
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Study
Location (Sample Size)
Cardiac Endpoint
Exposure
Assessment
Mean CO Level
(ppm)
Copollutants
Wheeler et al.
(2006)
Atlanta, GA
(n = 30)
SDNN, r-MSSD, PNN, LF,
HF, LF/HF
Ambient
Mean: 362 ppb (4h)
25th, 50th, 75th
percentiles:
221.5,304.3,398.1
PM2.5,03, NO2, SO2
Dales et al.
(2004)
Toronto, Canada
(n = 36)
SDNN, HF, LF, LF/HF, N-N
Personal
Mean: 2.4*
Range: 0.4-16.5
PM2.5
Peters et al.
(1999a)
Augsburg, Germany (n = 2681)
HR
Ambient
Mean: 3.6
Range: 1.5-7.1
TSP, S02
Liao et al. (2004)
Maryland, North Carolina,
Minnesota,
(n = 4899-6784)
HR, SDNN, LF, HF
Ambient
Mean: 0.65 (24 h)
PM10, O3, NO2, SO2
Timonen et al.
(2006)
Amsterdam, the Netherlands;
Erfurt, Germany; Helsinki, Finland
(n = 131)
SDNN, HF, LF/HF
Ambient
Mean: 0.35-0.52
Range: 0.09-2.17
PM2.5, PM10, NO2
Ibald-Mulli et al.
(2004)
Amsterdam, the Netherlands;
Erfurt, Germany; FHelsinki, Finland
(n = 131)
BP, HR
Ambient
Mean: 0.35-0.52
Range: 0.09-2.17
UFP, PM10, PM25, N02,
S02
Tarkiainen et al.
(2003)
Kuopio, Finland
(n = 6)
PNN, SDNN, r-MSSD
Personal
Mean: 4.6
Range: 0.5-27.4
None
Chan et al. (2005) Taipei, Taiwan
(n = 83)
SDNN,
r-MSSD, LF
Ambient
Mean: 1.1
Range: 0.1-7.7
PM10, NO2, SO2
Gold et al. (2005)
Boston, MA
(n = 28)
ST-segment
Ambient
Median: 0.56 (12 h)
Maximum: 1.04
PM2.5, BC
*95th percentile of 24-h levels
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 of which can
be associated with 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/minute) while bradycardia refers to a slow heart beat
(e.g., <60 beats/minute). 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) 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, 03, N02, S02, S042,
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
March 2009
5-18
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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., 2000b). 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 PMi0 (r = 0.51) and PM2 5 (r = 0.56) and more highly
correlated with N02 (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., 2006b), Atlanta, GA (among 518 subjects) (Metzger et al., 2007), Boston, MA (among
203 subjects) (Rich et al., 2005), and Vancouver, Canada (Rich et al., 2004; Vedal et al., 2004) (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 also did not find an
association with ambient CO (Rich et al., 2005).
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 non-smoking older adults for 24 weeks during summer and fall (Sarnat et al.,
2006). Ambient exposures for up to 5 days prior to the health assessment (based on a 5-day moving
average) were calculated for PM2 5, S042, elemental carbon (EC), 03, N02, S02, 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 month period (Berger et al., 2006). Exposure estimates were calculated for ambient
ultrafine particles, PM2 5, CO, NO, N02, and S02 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 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, few studies have examined associations between CO and the occurrence of cardiac
arrhythmias, and these studies provided little evidence of a CO effect on cardiac arrhythmias. While more
March 2009
5-19
DRAFT - DO NOT CITE OR QUOTE
-------
studies analyzed data from ICDs, very few reported significant associations. This was similar for the
mixed results from the two 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
Cardiac Endpoint
Exposure
Assessment
Mean CO Level
(ppm)
Copollutants
ARRHYTHMIAS (AMONG PATIENTS WITH ICDS)
Dockery et al.
(2005)
Boston, MA
(n = 203)
Ventricular
Tachycardia
Ambient
25th, 50th, 75th, 95th, percentiles:
0.53,0.80,1.02,1.37 (2-day)
PM2.5, BC, 03, N02, S02, S042"
Peters et al.
(2000b)
Massachusetts,
(n = 100)
Ventricular fibrillation or
tachycardia
Ambient
Mean: 0.58 (24 h)
Max: 1.66
PM25, PM10, BC, O3, NO2, SO2
S042"
Rich et al. (2006b)
Boston, MA
(n = 56)
Ventricular arrhythmia
Ambient
25th, 50th, 75th percentiles:
0.4,0.5,0.6(24 h)
PM2.5, EC, 03, N02, S02
Metzgeret al.
(2007)
Atlanta, GA
(n = 518)
Ventricular
Tachycardia
Ambient
Mean: 1.7 (1 h)
Range: 0.1-7.7
PM10, PM2.5, O3, NO2, SO2
Rich et al. (2005)
Boston, MA
(n = 203)
Atrial fibrillation
Ambient
25th, 50th, 75th, 95th, percentiles:
0.53,0.80,1.02,1.37 (2-day)
PM2.5, BC, 03, N02, S02
Rich et al. (2004)
Vancouver,
Canada
(n = 34)
ICD discharge due to
arrhythmia
Ambient
Mean: 0.55 (24 h)
IQR: 0.16
PM25, PM10, EC, 03, N02, S02,
S04
Vedal et al. (2004)
Vancouver,
Canada
(n = 50)
ICD discharge due to
arrhythmia
Ambient
Mean: 0.6 (24 h)
Range: 0.3-1.6
PM10,03, NO2, SO2
ARRHYTHMIAS (VIA ECG)
Sarnat et al. (2006)
Steubenville, OH
(n = 32)
Atrial or ventricular
tachycardia
Ambient
Mean: 0.2 (24 h)
Range: 0.1,1.5
PM2.5, 03, N02, S02, S04, EC
Berger et al. (2006) Erfurt, Germany
(n = 57)
Atrial or ventricular
tachycardia
Ambient
Mean: 0.45 (24 h)
Min, Med, Max
0.10,0.38,1.68
PM10, PM2.5, N02, NO, S02, UF
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 few
studies have investigated the association between ambient CO exposure and the risk of cardiac arrest and
none reported a significant link between increased CO exposure and the occurrence of cardiac arrest.
March 2009
5-20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Two similar studies were conducted in Seattle, WA, and both did not report an association between
ambient CO and cardiac arrest. Both studies employed a case-crossover design and examined air pollution
exposures for black smoke particles (BSP), PMi0, S02, and CO. The correlation coefficient for PMi0 and
CO was 0.81. The first of these studies examined paramedic-attended out-of-hospital primary cardiac
arrests among 362 cases (between 1998-1994) in Seattle and King County, WA whereby lags of 0-5 days
were analyzed (Levy et al., 2001). The second of these studies examined out-of-hospital primary cardiac
arrest for a ten year period (1985-1994) among subjects within a health organization database (the Group
Health Cooperative of Puget Sound) whereby 0-day through 2-day lags were analyzed (Sullivan et al.,
2003).
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) 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, 03, N02, S02)
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 being 2 h
before the onset.
Blood Pressure
Only two studies have investigated whether short-term ambient CO influences BP. The earlier of
these two studies examined BP among 2607 men and women aged 25-64 who participated in the
Augsburg, Germany MONICA study (Ibald-Mulli et al., 2001). Exposures to ambient TSP, S02 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., 2004b). While results showed significant associations between increased BP and
ambient PM2 5, S02, 03, and BC, there was no significant effect for CO.
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,
March 2009
5-21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Factor VII, C-reactive protein (CRP), prothrombin, intercellular adhesion molecule (ICAM-1), and white
blood cell count (WBC).
Pekkanen et al. (2000) examined the association between daily concentrations of air pollution and
concentrations of plasma fibrinogen measured among 4982 male and 2223 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, 03, N02, S02, 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 and N02
(r = 0.81) and more moderate correlations with PMi0 (r = 0.57) and S02 (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 significantly associated with an increase in
plasma fibrinogen. Results were similar for N02 while all other pollutants were not associated with an
increase in plasma fibrinogen.
Liao et al. (2005) 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, N02,
S02, 03 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 CVD. CO was not associated with other hemostatic or inflammatory factors.
In Israel, Steinvil et al. (2008) examined WBC, fibrinogen, and CRP among 3,659 study subjects
enrolled in the Tel-Aviv Sourasky Medical Center inflammation survey, in which subjects lived
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
to 5 days (lags 1-4) was calculated for each patient. For CRP, the odds of observing CRP concentrations
above the 90th percentile 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 during lags 0-3 was associated with a decrease in FVII.
A similar study by Ruckerl and colleagues (2007) was conducted among 1,003 MI survivors across
6 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, PM2 5,
NO, N02, 03, S02, 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) 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 (PMi0, 03, N02, and CO) were obtained
from 53 fixed stations across the Lombardia region, which was divided into nine different study areas and
a network average for each pollutant was calculated across the available monitors within each of the nine
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, PMi0 and N02 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.
In summary, despite the small number of studies, there was some evidence of a significant link
between CO exposure and blood markers of coagulation and inflammation. Table 5.6 summarizes the
reviewed studies.
March 2009
5-23
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 5-6.
Summary of studies investigating the effect of CO exposure on blood markers of
coagulation and inflammation.
Study
Location,
Sample Size
Cardiac Endpoint
Exposure
Assessment
Mean CO Level (ppm)
Copollutants
Pekkanen et al (2000) London, U.K.
(n = 7205)
Plasma fibrinogen
Ambient
Mean: 1.22 (24 h)
10th, 50th, 90th, Max:
0.61,1.04,2.0,8.61
PM10, BS, 03, N02, S02
Liao et al (2005)
USA
(n = 10.208)
Fibrinogen, Vll-C, WBC,
albumin, vWF
Ambient
Mean: 1.4 (24 h)
PM10, O3, NO2, SO2
Steinvil et al (2008)
Tel-Aviv, Israel
(n = 3659)
CRP, fibrinogen,
WBC
Ambient
Mean: 0.8
25th, 50th, 75th percentiles:
0.7,0.8,1.0
PM10, O3, NO2, SO2
Ruckerl et al (2006)
Erfurt, Germany
(n = 57)
CRP, SAA, cell
adhesions and
coagulation
Ambient
Mean: 0.45 (24 h)
Range: 0.10,1.68
PM10, PM2.5, UFP, EC, N02
Rukerl et al (2007)
Six European cities
(n = 1003)
IL-6, CRP, fibrinogen
Ambient
Mean (24 h): 0.29-1.48
PM10, PM2.5, O3, NO2, SO2
Baccarelli et al (2007) Lombardia region, Italy
(n = 1218)
PT, APTT, fibrinogen,
anticoagulants
Ambient
Mean: 1.14-3.11
Max: 5.52-11.43
PM10, O3, NO2, SO2
Hospital Admissions and Emergency Department Visits
Since the 2000 CO AQCD there have been a number of studies investigating the effect of ambient
CO on hospital admissions and ED visits for cardiovascular diseases. Some of these studies have focused
solely on one specific CVD outcome, and these studies are discussed first. This is followed by a
discussion of studies that investigated admissions for all CVD outcomes (e.g., non-specific) or a variety
of specific CVD outcomes.
Ischemic Heart Disease, Myocardial Infarction, and Angina Pectoris
Ischemic heart disease (IHD), also known as coronary heart disease (CHD), is caused by
inadequate circulation of the blood to the heart muscle, which is a result of the heart arteries being
blocked by cholesterol deposits. IHD can lead to sudden episodes such as MI ("heart attack") or death, as
well as chronic conditions such as angina pectoris (chest pain).
Ischemic Heart Disease. Very few 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). Mann and colleagues (2002) 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 to 1995
in southern California. There were 54,863 visits analyzed and a 0.75 ppm increase in 8-h max CO
March 2009
5-24
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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) among persons without either
secondary diagnosis. Of all pollutants examined (PMi0, N02, 03, CO), only N02 showed similar positive
effects to CO and no multipollutant models were analyzed. The correlation coefficients between CO and
N02 ranged from 0.64 to 0.86 across the seven regions. This study indicated that people with IHD and
accompanying CHF and /or arrhythmia are a sensitive group in relation 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).
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) attributing to 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 /o [95 /o CI. 9.2—31.6]). \()¦ was the only other pollutant examined, and it too s 11ow cd significant
positive associations with ED visits for IHD for same-day exposure; however, no multipollutant models
were examined.
Lee and colleagues (2003b) examined daily counts of hospital admissions for IHD in Seoul, Korea
for the period from December 1997 to December 1999. Single-day lags 0 to 5 were analyzed and the lag
period with the strongest association for each pollutant was chosen. 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+ years of age (RR:
1.07 [95% CI: 1.01-1.13]). All other pollutants (PMi0, 03, N02) except S02 showed similar significant
effects and in a two-pollutant model with PMi0 the CO effect attenuated toward the null.
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 to 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., 2004a) used a time-series design and analyzed a
3-day moving average 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) examined the association of ambient air pollution
levels and cardiovascular morbidity in visits with and without specific secondary conditions (e.g., co-
morbidity). Within a time-stratified case-crossover design using the same lag structure already mentioned,
the main results showed that a 1 ppm increase in 1-h max CO concentration was associated with an
March 2009
5-25
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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)
analyzed data from 5 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 years and > 65 years 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, but only
among the elderly group (> 65 years).
The second of the Australian studies (Jalaludin et al., 2006) examined ED visits for CVD outcomes
in the elderly (65+years) in Sydney for the period 1997 to 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).
Myocardial Infarction. Linn et al. (2000) 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 years. 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 years) across 11 hospitals (Koken et al., 2003). Data between
July and August for the period 1993-1997 were analyzed. Single-day lags 0 to 4 were examined and CO
showed no association with hospital admissions for MI (quantitative results were not reported).
As part of the HEAPS S (Health Effects of Air Pollution among Susceptible Subpopulations) study,
Lanki et al. (2006) 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 N02 ranged from 0.43 to 0.75 across the five cities,
and for PMi0 the range was 0.21 to 0.56. A total of 26,854 hospitalizations were analyzed and pooled
estimates from all 5 cities showed that there was a weak positive association with AMI hospitalizations
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 cites (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
March 2009
5-26
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
<75-year age group (RR: 1.080 [95% CI: 1.017-1.144), and with non-fatal AMI in the > 75-year age
group (RR: 1.044 [95% CI: 1.011-1.076).
Further analyses within the HEAPSS cohort was conducted where the event of cardiac readmission
among the first MI survivors (n = 22, 006) was linked to ambient air pollution (von Klot et al., 2005). The
readmissions of interest were those with primary diagnosis of AMI, angina pectoris, dysrhythmia, and
heart failure that occurred at least 29 days after the index event. Single-day lags 0 to 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])
and this persisted in two-pollutant models that included either PMi0 or 03. Correlation coefficients with
CO ranged from 0.21 to 0.57 for PMi0 and 0.44 to 0.75 for N02.
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 etal., 2003). 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 for lag 2 (OR: 1.019 [95% CI: 1.001-1.037]). The other pollutants analyzed were N02 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 hospitalization (Barnett et al., 2006). 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 years).
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 to 2001 (Hosseinpoor
et al., 2005). Using a time-series approach, single-day lags of 0 to 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 N02,
PMio, and 03 with CO being the only significant pollutant (OR: 1.005 [95% CI: 1.001-1.008]).Figure 5-1
shows the effect estimates associated with daily admissions for various forms of IHD from selected
studies. Table 5.7 shows a summary of the IHD hospital admission studies that examined CO exposures.
In summary, the majority of studies reported significant increases in the daily number of
admissions for IHD 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).
March 2009
5-27
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location Lag
Mann et al. (2002)
California,
0-3
¦ CHF IHD
Mann et al. (2002)
California,
0-3
' ARR
Szyszkowicz (2007)
Montreal, Canada
0
| m All Anes
Szyszkowicz (2007)
Montreal, Canada
0
1
i - All Anes Males
|
Szyszkowicz (2007)
Montreal, Canada
0
1 t All Anes Females
Szyszkowicz (2007)
Montreal, Canada
0
m >64 vears
Szyszkowicz (2007)
Montreal, Canada
0
i , >64 vears Males
Szyszkowicz (2007)
Montreal, Canada
0
i m >64 vears Females
Lee et al. (2003b)
Seoul, Korea
5
, 'All Anes
Lee et al. (2003b)
Seoul, Korea
5
I , >64 vears
Metzgeret al. (2004a)
Atlanta, GA
0-2
1
|
Peel et al. (2007)
Atlanta, GA2002)
0-2
1
r*—
Barnett et al. (2006)
Australia, New Zealand
0-1
, m 15-64 years
Barnett et al. (2006)
Australia, New Zealand
0-1
i
>64 years
Jalaludin et al. (2006)
Sydney, Australia
0-1
i
i
Barnett et al. (2006)
Australia, New Zealand
0-1
i m 15-64years Ml
Barnett et al. (2006)
Australia, New Zealand
0-1
i
i m >64 years
|
Linn et al. (2000)
Los Angeles, CA
0
1
Lanki et al. (2006)
Multi-city, Europe
0
, m >34 years, All cities
Lanki et al. (2006)
Multi-city, Europe
0
i
fc <75 years, Non-fatal
i
Lanki et al. (2006)
Multi-city, Europe
0
Lanki et al. (2006)
Multi-city, Europe
0
1 , 75+years Non-fatal
Lanki et al. (2006)
Multi-city, Europe
0
, m Fatal
von Klot et al. (2005)
Multi-city, Europe
0
¦
1
D'lppoliti et al. (2003)
Rome, Italy
0-2
j m 18+years
D'lppoliti et al. (2003)
Rome, Italy
0-2
D'lppoliti et al. (2003)
Rome, Italy
0-2
i
. , fi
-------
Table 5-7. Summary of IHD hospital admission studies.1
Study
Location
Endpoints Examined
Copollutants
Lags Examined
CO Levels (ppm)
STUDIES THAT FOCUSED SOLELY ON IHD, Ml, OR ANGINA
Mann et al. (2002)
Southern California
(1988-1995)
IHD
PM10, NO2, O3
0,1,2, 2-4ma
Mean: 2.07 (8h)
Szyszkowicz (2007)
Montreal, Canada
(1997-2002)
IHD
NO2
0,1
Mean: 0.5 (24 h)
Lee et al. (2003b)
Seoul, Korea
(1997-1999)
IHD
PM10, NO2, SO2, O3
0,1,2,3,4,5
Mean: 1.8
Lanki et al. (2006)2
5 European cities
(1992-2000)
Ml (first acute)
PM10, N02, 03, PNC
0,1,2,3
Highest city was Rome.
25th = 1.5
75th = 2.9 mg/m3
von Klot et al. (2005)2
5 European cities
(1992-2001)
Ml, Angina, Cardiac*
PM10, N02, 03, PNC
0,1,2,3
Mean: highest city was Rome:
1.9(24 h)
D'lppoliti et al. (2003)2
Rome, Italy
(1995-1997)
Ml
TSP, N02, S02
0,1,2,3,4, 0-2
Mean: 3.8 (24 h)
Hosseinpoor et al.
(2005)2
Tehran, Iran
(1996-2001)
Angina
PM10, NO2, SO2, O3
0,1,2,3
Mean: 9.4 (24 h)
STUDIES THAT EXAMINED IHD, Ml, OR ANGINA AMONG OTHER CVDS
Metzgeret al. (2004a)
Atlanta, GA
(1993-2000)
IHD, All CVD, CD, CHF,
PVCD
PM10, NO2, SO2, O3
0-2ma
Mean: 1.5 (1 h)
Peel et al. (2007)
Atlanta, GA
(1993-2000)
IHD, All CVD, CD, CHF,
PVCD
PM10, NO2, SO2, O3
0-2ma
Mean: 1.5 (1 h)
Barnett et al. (2006)
Australia and New
Zealand
(1998-2001)
IHD, Ml, All CVD, CA,
Stroke
PM10, NO2, O3
Lag 0-1
Mean: (8 h)
0.5-2.1
Jalaludin et al. (2006)
Sydney, Australia
(1997-2001)
IHD, All CVD, Stroke,
Cardiac
PM10, NO2, SO2, O3
0,1,2,3,0-1
Mean: 0.82 (8 h)
Linn et al. (2000)
Los Angeles, CA
(1992-1995)
Ml, All CVD, CHF, CA, OS
PM10, NO2, O3
0
Mean: (24 h)
Winter 1.7, Spring 1.0,
Summer 1.2, Fall 2.1
Koken et al. (2003)
Denver, CO
(1993-1997)
Ml, CAth, PHD, CD, CHF
PM10, NO2, SO2, O3
0,1,2,3,4
Mean: 0.9 ppm (24 h)
1 Cardiac = AMI, angina, dysrhythmia, or HF; CA = Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhthmias; CHF = Congestive heart failure; PHD = Pulmonary
heart disease; OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average.
2These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which assumes standard
atmosphere and tempterature.
March 2009
5-29
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 mini-stroke, which has the same early symptoms as a
normal stroke but the symptoms disappear within 24 hours, leaving no apparent deficits.
A small number of air pollution studies have investigated hospital admissions for the three main
forms of stroke with the majority reporting positive associations with ambient CO and lag periods
between 0 and 3 days.
A U.S. study across 9 cities investigated hospital admissions for ischemic and hemorrhagic stroke
among Medicare beneficiaries aged 65+ years of age (155,503 ischemic and 19,314 hemorrhagic
admissions from the ED) (Wellenius et al., 2005a). Single-day lags 0 to 2 were examined and based on a
pooled estimate, same-day CO (lag 0) was associated with an increase in admissions of 1.98%
(95% CI: 0.86-3.12) per 0.5 ppm increase in 24-h CO concentration) for ischemic stroke admissions but
not hemorrhagic stroke admissions (-1.14%, 95% CI: -3.40 to 1.18). All other pollutants examined (PMi0,
N02, S02) were associated with an increase in ischemic stroke admissions, but not hemorrhagic stroke
admissions.
Villeneuve and colleagues (2006a) studied ED visits for hemorrhagic strokes, acute ischemic
strokes and transient ischemic attacks among individuals 65+ years of age at 5 hospitals within the
Edmonton area in Canada 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) and CO only had an effect on ischemic stroke during April-September. A 0.5 ppm increase the
CO concentration for a 3-day avg across lags 0-2 was associated with a 32% increase in risk (OR: 1.32
[95% CI 1.09-1.60]). 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, S02, and 03. When all seasons and all
strokes were combined there was no statistically significant association between all the pollutants
examined and increased admissions for stroke.
In Kaohsiung City, Taiwan, CO averaged over lags 0-2 was associated with increased admissions
for stroke across 63 hospitals (Tsai et al., 2003b). From 1997 to 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.
March 2009
5-30
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
These results persisted in two-pollutant models that included PMi0, S02, and 03, but did not persist when
controlling for N02.
Earlier research conducted in metropolitan Los Angeles examined hospital admissions for
cardiopulmonary illnesses from 1992-1995 (Linn et al., 2000). 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 years. When the analyses were
stratified by season there was a 1.8% increase during winter, a 4.55% increase during summer, and a 1.6%
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 to 2002 (Chan et al., 2006). Single-day
lags 0 to 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 03 and
PMio. There was no association with individual ischemic or hemorrhagic stroke. CO was moderately
correlated with PMi0 (r = 0.47) and PM2 5 (r = 0.44), and the correlation was higher with N02 (r = 0.77).
The only time-series study that focused specifically on stroke hospital admissions that did not
report a significant association with ambient CO was conducted in Dijon, France (Henrotin et al., 2007).
Hospital admissions for different types of first-ever stroke (e.g., ischemic, hemorrhagic) among subjects
over 40 years of age were analyzed for the period of 1994 to 2004. A bi-directional case-crossover study
design was employed where single-day lags of 0 to 3 were examined and CO had no significant
association across all lags. 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 (PMi0,
NOx, 03, S02, CO) only 03 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). A time-stratified case-crossover design was employed and
the age groups of 15-64 years and > 65 years 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+ years) in
Sydney for the period from 1997 to 2001 (Jalaludin et al., 2006). Using a time-series approach, single-day
lags of 0 to 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.
March 2009
5-31
DRAFT - DO NOT CITE OR QUOTE
-------
1 Figure 5-2 shows the effect estimates associated with daily admissions for stroke from selected
2 studies. Table 5-8 shows a summary of the stroke hospital admission studies that examined CO exposures.
3 In summary, there was some evidence that increased ambient CO concentrations were associated
4 with an increase in the number of hospital admissions for stroke. The largest positive effects came from
5 the Taiwan study in Kaohsiung (Tsai et al., 2003b) with slightly larger effects during the warmer period
6 (>20°C). Similarly, in the Canadian study by Villeneuve and colleagues (2006a) there was a stronger
7 effect during the warmer period (April-September).
March 2009
5-32 DRAFT-DO NOT CITE OR QUOTE
-------
Study Location Lac[
Wellenius et al. (2005a)
Multi-city, U.S.
0-2
f 65+years
Ischemic Stroke
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
"f" IS, 65+ years, all seasons
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
[ IS, 65+years, Apr-Sep
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
IS, 65+ years, Oct-Mar
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
""J1 CIS, 65+ years, all seasons
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
, CIS, 65+ years, Apr-Sep
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
* CIS, 65+ years, Oct-Mar
|
Tsai et al. (2003b)
Kaohsiung, Taiwan
0-2
20°CTemp
i
Linn et al. (2000)
Los Angeles, CA
0
b
i
¦
Chan et al. (2006)
Taipei, Taiwan
1
i
i
Henrotin et al. (2007)
Dijon, France
3
t
i
i
Wellenius et al. (2005a)
Multi-city, U.S.
0-2
^ 65+ years
Flemorrhagic Stroke
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
65+ years, all seasons
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
i
J * 65+years, Apr-Sep
Villeneuve et al. (2006a)
Edmonton, Canada
0-2
1
"*[ 65+years, Oct-Mar
Tsai et al. (2003b)
Kaohsiung, Taiwan
0-2
1
j " <20°CTemp
Tsai et al. (2003b)
Kaohsiung, Taiwan
0-2
| * >20°CTemp
Chan et al. (2006)
Taipei, Taiwan
1
i
Henrotin et al. (2007)
Dijon, France
1
*
i
Chan et al. (2006)
Taipei, Taiwan
2
i
i
i
^ Stroke (Non-Specific)
Jalaludin et al. (2006)
Sydney, Australia
0-1
k
¦
i
I
I 1 1 1 1 1 1 1 1
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
IS= Ischemic Stroke, CIS=Cerebral Ischemic Stroke Effect Estimate
Figure 5-2. 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.
March 2009
5-33
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 5-8. Summary of stroke hospital admission studies.1
Study
Location
Type Of Stroke
Examined
Copollutants
Lags Examined
CO Levels (ppm)
STUDIES THAT FOCUSED SOLELY ON STROKE
Wellenius et al. (2005a)
9 cities, USA
(1993-1999)
Isch, Flem
PM10, NO2, SO2
0,1,2
25th, 50th, 75th percentiles:
0.73,1.02,1.44
Villeneuve et al.
(2006a)
Edmonton, Canada (1992-
2002)
Isch, Flem, TIA
NO2, SO2, O3
0,1,0-2
Mean: 0.8 (24 h)
Tsai et al. (2003b)
Kaohsiung, Taiwan
(1997-2000)
Isch, Flem
PM10, NO2, SO2, O3
0-2
Mean: 0.79 (24 h)
Chan et al. (2006)
Taipei, Taiwan
(1997-2002)
All, Isch, Flem
PM10, NO2, SO2, O3
0,1,2,3
Mean: 1.7 (8h)
Henrotin et al. (2007)2
Dijon, France
(1994-2004)
Isch, Flem
PM10, NOx, SO2, O3
0,1,2,3
Mean: 0.59 (24 h)
STUDIES THAT EXAMINED STROKE AMONG OTHER CVDS
Linn et al. (2000)
Los Angeles, CA
(1992-1995)
Isch
PM10, NO2, O3
LagO
Mean: (24 h)
Winter 1.7, Spring 1.0,
Summer 1.2, Fall 2.1
Barnett et al. (2006)
Australia and New Zealand
(1998-2001)
All
PM10, NO2, O3
Lag 0-1
Mean: (8h)
0.5-2.1
Jalaludin et al. (2006)
Sydney, Australia
(1997-2001)
All
PM10, NO2, SO2, O3
0,1,2,3,0-1
Mean: 0.82 (8h)
11sch = Ischemic; Hem = Hemorrhagic; TIA = transient ischemic attack
2These studies presented CO concentrations in the units mg/m3. The concentrations were converted to ppm using the conversion factor 1 ppm = 1.15 mg/m3, which assumes standard
atmosphere and tempterature.
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.
Wellenius and colleagues (2005b) examined the rate of hospitalization for CHF among 55,019
Medicare recipients (aged > 65 years) residing in Allegheny County, PA, during 1987-1999. A time-
stratified case-crossover design was employed and single-day lags of 0 to 3 were analyzed and a 1 ppm
increase in 1-h max CO concentration on the same-day (lag 0) was associated with a 9.31%
(95% CI: 6.77-11.92) increase in the rate of hospitalization for CHF. This result persisted in two-pollutant
models that included PM10, N02, 03, and S02. CO was moderately correlated with S02 (r = 0.54) and
PM10 (r = 0.57) and more highly correlated with N02 (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). The patients were interviewed
after admission through the ED during their stays in overnight wards. The interview was designed to
March 2009
5-34
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
collect information about symptom onset, health conditions, and factors related to air pollution exposure.
Various lag periods (single day and cumulative days 0 to 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,
N02, 03) 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 (2000). 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 aged >30 years. 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, Colorado, investigated daily admissions for various CVDs among
older adults (>65 years) across 11 hospitals (Koken et al., 2003). Single-day lags 0 to 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 to 2000 were analyzed (Metzger et al., 2004a). A time-series design was
used and a 3-day moving average 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). When the analyses examined the same CVDs
among those with and without specific secondary conditions (e.g., co-morbidity) 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) (Peel etal., 2007).
In Kaohsiung city, Taiwan, a study analyzed 13,475 admissions for CHF across 63 hosptials for the
period 1996 through 2004 (Lee et al., 2007a). 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, S02, 03, and models with N02
only on warmer days, not with N02 on cooler days.
Figure 5-3 shows the effect estimates for associations between CO and daily admissions for HF
from selected studies. Table 5-9 summarizes the HF 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 significant associations at lags of 0 to 3 days.
March 2009
5-35
DRAFT - DO NOT CITE OR QUOTE
-------
Study Location Lag
Wellenius etal. (2005b) Pittsburgh, PA 0
Symons et al. (2006) Baltimore, MD 0-3
Linn et al. (2000)
Los Angeles, CA 0
Koken et al. (2003) Denver, CO
Metzger et al. (2004a) Atlanta, GA
0-2
Peel et al. (2007)
Atlanta, GA
0-2
Lee et al. (2007a) Kaohsiung, Taiwan 0-2
Lee et al. (2007a) Kaohsiung, Taiwan 0-2
<25 °C
>25 °C
0,75
1.25
1.75
2.75
Effect Estimate
Figure 5.3. 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.
March 2009
5-36
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 5-9. Summary of HF hospital admission studies.
Study
Location
Endpoints Examined
Copollutants
Lags Examined
CO Levels (ppm)
STUDIES THAT FOCUSED SOLELY ON HF
Wellenius et al. (2005b)
Pittsburgh, PA (1987-1999)
CHF
PM10, NO2, SO2, O3
0,1,2,3
Mean: 1.03 (24 h)
Symons et al. (2006)
Baltimore, MD (2002)
CHF
PM2.5, NO2, O3
0,1,2,3
Mean: 0.4 (24 h)
Lee et al. (2007a)
Kaohsiung, Taiwan
(1996-2004)
CHF
PM10, NO2, SO2, O3
0-2
Mean: 0.76 (24 h)
STUDIES THAT EXAMINED HF AMONG OTHER CVDS
Linn et al. (2000)
Los Angeles, CA (1992-1995)
CHF, Ml, All CVD, CA, OS
PM10, NO2, O3
0
Mean: (24 h)
Winter 1.7; Spring 1.0
Summer 1.2; Fall 2.1
Koken et al. (2003)
Denver, CO (1993-1997)
CHF, Ml, CAth, PHD, CD
PM10, NO2, SO2, O3
0,1,2,3
Mean: 0.9 (24 h)
Metzger et al. (2004a)
Atlanta, GA (1993-2000)
CHF, IHD, All CVD, CD, PVCD
PM10, NO2, SO2, O3
0-2ma
Mean 1.5 (1 h)
Peel et al. (2007)
Atlanta, GA (1993-2000)
CHF, IHD, All CVD, CD, PVCD
PM10, NO2, SO2, O3
0-2ma
Mean 1.5 (1 h)
*Cardiac = AMI, angina, dysrhythmia, or HF; CA = Cardiac arrhythmia; CAth = Cardiac atherosclerosis; CD = cardiac dysrhthmias; CHF = Congestive heart failure; PHD = Pulmonary
heart disease; OS = Occlusive stroke; PVCD = peripheral vascular and cerebrovascular disease, ma = moving average.
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., non-specific). Several of these studies also examined
specific CVDs and were briefly discussed in previous sections.
As discussed earlier, a study was conducted in Atlanta, GA where over 4 million ED visits from
31 hospitals for the period 1993 to 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 moving average over single-day lags 0-2 as the a
priori lag structure (Metzger et al., 2004a). 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 N02 and PM25.
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).
Within a time-stratified case-crossover design, a 3-day moving average 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 co-morbidity 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 the co-morbidities, 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.
March 2009
5-37
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Table 5-10.
Association of ambient air pollution levels and cardiovascular morbidity in visits with
and without specific secondary conditions.
Co-morbidity
IHD
Dysryhthmias
PVCD
CHF
HYPERTENSION
-With
1.007 (0.978-1.037)
1.065 (1.015-1.118)
1.038 (1.004-1.074)
1.037 (0.997-1.079)
- Without
1.022(1.000-1.043)
1.008 (0.988-1.029)
1.027 (1.002-1.054)
1.010 (0.985-1.037)
DIABETES
-With
0.985 (0.945-1.027)
1.058 (0.976-1.146)
1.065 (1.012-1.121)
1.020 (0.975-1.067)
- Without
1.023(1.004-1.042)
1.014(0.995-1.034)
1.025 (1.003-1.048)
1.018 (0.993-1.044)
COPD
-With
0.996 (0.938-1.057)
0.972 (0.878-1.077)
1.113 (1.027-1.205)
1.058 (1.003-1.115)
- Without
1.018(1.000-1.036)
1.018 (0.999-1.038)
1.026 (1.004-1.047)
1.011 (0.987-1.036)
CHF
-With
0.956 (0.907-1.007)
1.065 (0.968-1.173)
1.072 (0.981-1.172)
-
- Without
1.024(1.006-1.042)
1.015(0.996-1.034)
1.029 (1.008-1.051)
-
DYSRYHTHMIAS
-With
1.028 (0.985-1.072)
-
1.072 (1.011-1.138)
1.004 (0.960-1.051)
- Without
1.014(0.995-1.033)
-
1.026 (1.004-1.048)
1.023 (0.998-1.049)
PVCD - peripheral vascular and cerebrovascular disease, IHD = ischemic heart disease, CHF = congestive heart failure.
Source: Peel et al. (2007)
The third study utilizing the SOPHIA data extended the time period to include 1993 through 2004
(Tolbert et al., 2007) and focused on two large outcome groups: a respiratory diseases group and a
cardiovascular diseases group. The combined cardiovascular case group included the 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 N02, CO and TC, 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
aged >30 years (2000). When the analyses were stratified by season the significant CO effect was
strongest during winter (1.9% increase) followed by 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 to 3 days) and an increase in the number of daily cardiac hospital admissions
(quantitative results not reported) (Slaughter et al., 2005). 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). A total of 11,632 cardiac admissions were
March 2009
5-38
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
analyzed for the period of 1995 to 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 age group (lag 0) was -2.6 (95% CI: -6.2 to 3.3); and
for the 65+ age group, 0.4 (95% CI: -1.9 to 2.7). The authors reported moderate to low correlations with
N02 (r = 0.38), PMio (r = 0.21) and S02 (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 of
1997-2001 were analyzed (Chang et al., 2005). 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 PMi0, S02, or 03 in two-pollutant
models. An identical study in Kaohsiung analyzed 29,661 CVD admissions for the period 1997-2000
(Yang et al., 2004b). 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 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). 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+ years. Among those aged 15-64 years 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+ years) in Sydney for the period 1997 to 2001 (Jalaludin et al., 2006). 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) analyzed emergency hospital admissions
in Valencia, Spain for the period 1994 to 1996. The mean daily number of CVD admissions was 7 and
when using a time-series approach 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,
March 2009
5-39
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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).
Ballester et al. (2006) extended this research to include data from 14 Spanish cities for the period
of 1995 to 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 emergency 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 N02, 03 and S02.
Table 5.11 summarizes the non-specific CVD hospital admission studies that examined CO
exposures. Figure 5-4 shows the effect estimates associated with daily admissions for non-specific 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 significant 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.
March 2009
5-40
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location
Lag
Tolbert et al. (2007)*
Atlanta, GA
0-2
i
i*
i
i
|
Linn et al. (2000)
Los Angeles, CA
0
1
1
All year
i
Linn et al. (2000)
Los Angeles, CA
0
f- Spring
1
Linn et al. (2000)
Los Angeles, CA
0
1
r* Summer
i
Linn et al. (2000)
Los Angeles, CA
0
i
• Fall
i
Linn et al. (2000)
Los Angeles, CA
0
i
!~ Winter
i
i
Fung et al. (2005)
Windsor, Canada
0-2
i
i
i
• <65 years
i
Fung et al. (2005)
Windsor, Canada
0-2
i
'• 65+ years
i
i
Chang et al. (2005)
Taipei, Taiwan
0-2
i
i
i >20°CTemp
i
Chang et al. (2005)
Taipei, Taiwan
0-2
•i <20°CTemp
i
i
i
Yang et al. (2004a)
Kaohsiung, Taiwan
0-2
i
i
i
i
i
Yang et al. (2004a)
Kaohsiung, Taiwan
0-2
i >20°CTemp
i
i
Barnett et al. (2006)
Australia, New Zealand
0-1
i
i
•• 15-64 years
Barnett et al. (2006)
Australia, New Zealand
0-1
i
65+years
i
i
Jalaludin et al. (2006)
Sydney, Australia
0-1
i
i
i
i • 65+ years
i
i
i
Ballester et al. (2006)
Multi-city, Spain
0-1
i
i
i
i
0.9 1.1 1.3 1.5 1.7 1.9 2.1
Effect Estimate
"Represents the results reported by Metzger et al. (2004b) and Peel et al. (2007)
Figure 5-4. 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.
March 2009
5-41
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-11. Summary of non-specific CVD hospital admission studies.
Study
Location
CVD Codes
Copollutants
Lags Examined
CO Levels (ppm)
Metzgeret al.
(2004a)
Atlanta, GA
(1993-2000)
All CVD
PM10, NO2, SO2, O3
0-2ma
Mean: 1.5 (1 h)
Peel et al. (2007)
Atlanta, GA
(1993-2000)
All CVD
PM10, NO2, SO2, O3
0-2ma
Mean 1.5 (1 h)
Tolbert et al.
(2007)
Atlanta, GA
(1993-2004)
All CVD
PM10, NO2, SO2, O3
0-2ma
Mean 1.6 (1 h)
Linn et al. (2000)
Los Angeles, CA
(1992-1995)
All CVD
PM10, NO2, O3
0
Mean: (24 h)
Winter 1.7; Spring 1.0;
Summer 1.2; Fall 2.1
Slaughter et al.
(2005)
Spokane, WA
(1995-2001)
All CVD
(ICD9: 390-459)
PM10, PM2.5, CO
1,2,3
Mean: range across 5 monitors 0.42-
1.82 (24 h)
Fung et al. (2005)
Windsor, Canada
(1995-2000)
All CVD (HF, IHF, or
Dysrhythmia)
PM10, NO2, SO2, O3
0,0-1,0-2
Mean: 1.3 (24 h)
Chang et al. (2005) Taipei, Taiwan
(1997-2001)
All CVD
(ICD9: 410-429)
PM10, NO2, SO2, O3
0-2
Mean: 1.37 (24 h)
Yang et al. (2004b)
Kaohsiung, Taiwan
(1997-2000)
All CVD
(ICD9: 410-429)
PM10, NO2, SO2, O3
0-2
Mean: 0.79 (24 h)
Barnett et al.
(2006)
Australia and New
Zealand
(1998-2001)
All CVD
(ICD9: 390-459)
PM10, NO2, O3
0-1
Mean: (8h)
0.5-2.1
Jalaludin et al.
(2006)
Sydney, Australia
(1997-2001)
All CVD
(ICD9: 390-459)
PM10, NO2, SO2, O3
0,1,2,3,0-1
Mean: 0.82 (8h)
Ballester et al.
(2001)1
Valencia, Spain
(1994-1996)
All CVD
(ICD9: 390-459)
bs,no2, so2,o3
1,2,3,4,5
Mean: 0.54 (24 h)
Bal tester et al.
(2006)1
Multi-city, Spain
(1995-1999)
All CVD
(ICD9: 390-459)
BS, PM10, TSP, N02, S02,
03
0-1
Mean: range across 14 cities
0.12-0.24 (8h)
1 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 tempterature.
1
2
3
4
5
6
Figure 5-5 summarizes 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
PMio, PM2 5, N02, S02, and 03. In all but one instance (Lee et al., 2007a) (<25°C adjusted for N02) when
the single pollutant effect estimate was positive for CO, it remained positive after the addition of any of
the copollutants investigated.
March 2009 5-42 DRAFT-DO NOT CITE OR QUOTE
-------
Study
Outcome
Lag
Copollutant
Welleniuset al. (2005b)
CHF
0
CO alone
J •
PM10
CO+PM10
1
Chan et al. (2006)
CD
2
CO alone
CO+PM10
1 -©-
Lee et al. (2007a)
CHF
0
CO alone
1 •
>25°C
CO+PM10
1 e
>25°C
D'lppoliti et al. (2003)
HA
0-2
CO alone
1
CO+TSP
-e-
Chang et al. (2005)
CVD
0
CO alone
1 > 20°C
CO+PM10
1 —e— > 20°c
Chang et al. (2005)
CVD
0
CO alone
—+1— <20°C
CO+PM10
—e—- 20°c
Lee et a I. (2003b)
IHD
0
CO alone
j-#— 64+years
CO+PM10
-p©— 64+ years
Ballester et al. (2006)
CVD
0-1
CO alone
r®~~
CO+PM10
lO
von Klot et al. (2005)
Cardiac
0
CO alone
1 #
CO+PM10
1 0
Lee et al. (2007a)
CHF
0
CO alone
1 <25°C
CO+PM10
*25°C 0
1
Chan et al. (2006)
CD
2
CO alone
PM2.5
CO+PM2.5
\e-
Welleniuset al. (2005b)
CHF
0
CO alone
¦
1
¦
NO2
CO+NO2
1 —
Ballester eta I. (2006)
CVD
0-1
CO alone
CO+NO2
1—
Chang et al. (2005)
CVD
0
CO alone
1 > 20°C
CO+NO2
1—e— > 20°c
Chang et al. (2005)
CVD
0
CO alone
1— <20°C
CO+NO2
©J <20°C
Lee et al. (2007a)
CHF
0
CO alone
1
>25°C
CO+NO2
'
>25°C
Lee et al. (2007a)
CHF
0
CO alone
<25°C
CO+NO2
sr>R°r
1
Welleniuset al. (2005b)
CHF
0
CO alone
1 •
SO2
CO+SO2
1 e
Ballester eta I. (2006)
CVD
0-1
CO alone
i-e_
CO+SO2
i-e-
Chang et al. (2005)
CVD
0
CO alone
1 -m- > 20°c
CO+SO2
1 —e— >20°c
Chang et al. (2005)
CVD
0
CO alone
«' <20°C
CO+SO2
1 —e— <20°c
Lee et al. (2007a)
CHF
0
CO alone
1
>25°C
CO+SO2
1
1 e—
>25°C
Lee et al. (2007a)
CHF
0
CO alone
1 <25°C
CO+SO2
, <25°C
0
Welleniuset al. (2005b)
CHF
0
CO alone
1
O3
CO+O3
1 -
Chan et al. (2006)
CD
2
CO alone
1
CO+O3
1
Ballester eta I. (2006)
CVD
0-1
CO alone
1
CO+O3
1
Chang et al. (2005)
CVD
0
CO alone
1 > 20°C
CO+O3
' > 20°C
Chang et al. (2005)
CVD
0
CO alone
< 20°C
CO+O3
, <20°C
von Klot et al. (2005)
Cardiac
0
CO alone
1
CO+O3
-1
Lee et al. (2007a)
CHF
0
CO alone
1
>25°C
CO+O3
1
>25°C
Lee et al. (2007a)
CHF
0
CO alone
1 <25°C
CO+O3
1
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
5.2.1.2. Epidemiologic Studies with Long-Term Exposure
Only two studies examined CVD outcomes in association with long-term exposure to CO.
Rosenlund et al. (2006) investigated long-term exposure (30 years) to urban air pollution and the risk of
MI in Sweden. The study included 2,246 cases and 3,206 controls aged between 45 to 70 years and
residing in Stockholm County during 1992 to 1993. A detailed postal questionnaire was completed by
4067 subjects and all addresses inhabited during more than 2 years 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, N02, CO, PMi0,
and PM2 5, with the addition of S02 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 year 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-year avg CO
concentration). Similar results were reported for N02. The correlation between the 30-year N02 and CO
exposures was reasonably strong (r = 0.74) and multipollutant models with both these pollutants included
(N02, CO) were not examined. No other pollutants were significantly associated with all other MI
outcomes.
A small-area ecologic study analyzed mortality and hospital admissions for stroke across 1,030
census districts in Sheffield, U.K. (Maheswaran et al., 2005b). 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]).
5.2.1.3. 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 physiological 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 myocardial infarction, and increased BP.
March 2009
5-44
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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. There is little evidence that ambient CO is
associated with an increase in hospital admissions for ischemic stroke. Studies of hospital admissions and
ED visits for IHD and CHF provide the strongest evidence of ambient CO being associated with adverse
CVD outcomes. It is difficult to determine from this group of studies the extent to which CO is
independently associated with CVD outcomes or if 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, are coherent with
the effects demonstrated by controlled human exposure and animal toxicological studies, and supports a
direct effect of short-term CO exposure on CVD morbidity at ambient concentrations below the current
NAAQS level. Such direct effects are plausible considering that long-term, low concentration CO
exposure (See Section 4.2.3) could result in a COHB level approaching those used in controlled human
exposure studies (See Section 5.2.2).
5.2.2. Controlled Human Exposure Studies
Controlled human exposure studies provide valuable information related to the health effects of
exposure to air pollutants following short term exposures. These types of experiments are often referred to
as human clinical studies and are conducted in a laboratory setting under carefully regulated exposure
concentrations, environmental conditions, and subject activity levels. Human clinical studies are typically
conducted using a randomized crossover study design with subjects exposed to both the pollutant(s) of
interest and a clean air control. 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. Human clinical studies are limited by a number of factors
including a small sample size and relatively short exposure time. In addition, although health-
compromised individuals have been included in human clinical studies, all subjects participating must be
relatively healthy and do not represent the most sensitive individuals in the population.
Several human clinical studies cited in the 2000 CO AQCD observed changes in measures of
cardiovascular function among individuals with CAD following short term exposures to CO. In a multi-
laboratory study of men with stable angina, Allred et al. (1989, 1991) evaluated the effect of CO exposure
on exercise-induced angina and ST-segment changes indicative of myocardial ischemia. Relative to clean
March 2009
5-45
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
air exposure (COHb ~ 0.6%), exposures to CO resulting in COHb concentrations of 2.4% and 4.7% were
shown to decrease the time required to induce ST-segment changes by 5.1% (p = 0.02) and 12.1%
(p <0.001), respectively. These changes were well correlated with the onset of exercise-induced angina. 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% (Adams et al., 1988; Anderson et al., 1973) (Kleinman et al., 1989; Kleinman et al.,
1998). However, Sheps et al. (1987) 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). 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 relatively 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 presented very little evidence of CO-induced changes in cardiovascular
function in healthy adults. Davies and Smith (1980) 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, 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) 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 post-exposure
exercise. However, CO was demonstrated to decrease the post-exposure 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 02 consumption. These results
support the findings of several studies cited in the 2000 CO AQCD which observed decreases in exercise
duration and maximal aerobic capacity among healthy adults at COHb levels > 3% (Drinkwater et al.,
1974; Ekblom and Huot, 1972; Horvath et al., 1975; Raven et al., 1974). 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.
March 2009
5-46
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Kizakevich et al. (2000) 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 02 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) 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) 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 et al. (1997) 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). This protocol resulted in COHb concentrations of 5.6%
and 9.4% following exposures of 30 and 60 min, respectively, with 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).
In summary, controlled human exposures to CO among individuals with CAD have been shown to
consistently increase markers of myocardial ischemia at COHb levels between 3 and 6%, with one study
reporting similar effects following CO exposures resulting in COHb concentrations of 2.4%. 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
of as low as 5%, this effect has not been consistent across studies.
March 2009
5-47
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
5.2.3. Toxicological Studies
While novel toxicological research on environmental levels of CO was limited in the 2000 CO
AQCD, several sections reported potential relationships between CO exposure and cardiovascular effects.
Conflicting experimental data relating to the role of CO in promoting atherosclerotic vessel disease was
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; Thomsen, 1974;
Turner et al., 1979), other studies have failed to see this association (Penn et al., 1992; Stupfel and
Bouley, 1970). Ventricular hypertrophy has also been shown after chronic CO exposure (Penney et al.,
1984; Penney et al., 1988).
The following sections describe recent studies dealing with toxicity of relatively low levels 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 levels (35-250 ppm), has pathophysiological effects on the cardiovascular system
and on relatively ubiquitous cellular pathways.
CO exposure at environmentally-relevant levels is unlikely to overwhelm normal physiology in a
healthy cell; however, susceptibility may be rendered by disease or early development. A common theme
appears to be the vulnerability of vascular cells, especially the endothelium, which could be considered
the first organ of contact once taken up into the circulation. While relatively little research has been
conducted since the 2000 CO AQCD, several key studies conducted at near-environmental CO levels
provide important clues to the potential public health implications of ambient CO exposure.
5.2.3.1. Endothelial Dysfunction
While the preferential binding to heme and effective displacement of 02 by CO has been well
established for over a century, new information from various fields of study are beginning to elucidate
non-hypoxic mechanisms that may lead to cardiovascular abnormalities associated with CO exposure.
Research by Thorn, Ischiropoulos, and colleagues (Ischiropoulos et al., 1996; Thorn et al., 1994; Thorn
and Ischiropoulos, 1997; Thorn et al., 1997; Thorn et al., 1999a; Thorn et al., 1999b; Thorn et al., 2000;
Thorn et al., 2006) 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 NO*
from platelets with subsequent neutrophil activation and vascular injury (Ischiropoulos et al., 1996; Thorn
et al., 2006). The steps include (1) peroxynitrite generation from the reaction of NO* from platelets with
March 2009
5-48
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
neutrophil-derived superoxide followed by (2) stimulation of intravascular neutrophil degranulation that
can result in (3) myleoperoxidase deposition along the vascular lining. Products from myeloperoxidase-
mediated reactions can cause endothelial cell activation (Thorn et al., 2006) 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. The release of
free NO* was noted in isolated rat platelets exposed to 10-20 ppm CO (Thorn and Ischiropoulos, 1997).
Increased nitrotyrosine content of the aorta was observed in rats exposed to 50 ppm CO for 1 h (1999a;
Thorn et al., 1999b). 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., 1999b). Studies in cultured endothelial cells were also conducted using buffer
saturated with 10-100 ppm CO (Thorn et al., 1997). 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 release ofNO* from endothelial cells along with peroxynitrite
formation and delayed cell death was observed at CO concentrations of 22 nM and higher (Thorn et al.,
1997). A more recent study demonstrated adaptive responses in endothelial cells exposed to this same
range of CO concentrations (Thorn et al., 2000). 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 altered oxidative 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). In some
cases CO promotes vasoconstriction, which is thought to be mediated by inhibition of endothelial NOS
(Johnson et al., 2003; Thorup et al., 1999) 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 et al., 2003, 2004; Teran et al., 2005). Additional
studies will be useful in determining whether environmentally-relevant concentrations of CO have
detrimental effect on pre-existing conditions such as hypertension, metabolic syndrome or pregnancy.
March 2009
5-49
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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) and gasoline engine emissions (6 h/day x 1, 3, and 7 day at 80 ppm)
(Lund et al., 2009). Furthermore, evidence of vascular ROS following gasoline emissions has been shown
in certain animal models (6 h/day x 50 day at 8-80 ppm) (Lund et al., 2007). While none of these studies
examined the potential role of CO, alone or in combination, it is clearly a common factor in the various
combustion atmospheres, and future work will be needed to reveal its importance on vascular health.
5.2.3.2. 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 02 consumption.
Direct 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) demonstrated cardiac hypertrophy in rats exposed for 72
weeks to 200 ppm CO. COHb levels were reported to be 14.7%. Neither structural signs of hypertension
in the pulmonary arteries or atherosclerotic lesions in the systemic arteries were observed. Cardiac
hypertrophy was also demonstrated in rats exposed to 100-200 ppm CO for 1-2 weeks (Loennechen et al.,
1999). This response was accompanied by an increase in endothelin-1 expression. COHb levels were
reported to be 12-23% in this study.
Direct effects of CO on the healthy heart have also been demonstrated following short-term
exposures. In a study by Favory et al. (2006b), 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 hours of recovery, while COHb values decreased from 11% to
March 2009
5-50
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 were reduced by CO
exposure. The authors concluded that the discordant alterations in contractility (increased) and perfusion
(decreased) may place the heart at risk of 02 limitations following this exposure to CO.
Several studies examined the impact of lower levels (50 ppm) on pre-existing 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). 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 weeks 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
histopathological findings. The authors cited previous work demonstrating that exogenous CO can inhibit
NOS (Thorup et al., 1999), 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; 2005), Dark Agouti rats were exposed for 10 weeks 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 (2007) indicated that CO may interfere with normal homeostatic
responses to hypoxia. This 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) 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 |im) 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 Gautier et al. (2007) and Melin studies (2002; 2005) suggested that the combined effect of low levels
of CO with hypoxia is an enhanced right ventricle workload and an exacerbated cardiomyopathy related
March 2009
5-51
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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.
5.2.3.3. Electrocardiographic Effects
In two related studies, Wellenius et al. (2004; 2006b) examined the effect of CO on a rat model of
ischemia-related arrhythmia that was previously shown to produce significant results with exposures to
PM (Wellenius et al., 2002). ECG changes were observed during exposure to residual oil fly ash (ROFA)
particles in a rat model of MI. Thus, using an anesthetized model of post-infarction myocardial sensitivity,
Wellenius and colleagues tested the effects of 35 ppm CO (1-h exposure) on the induction of spontaneous
arrhythmias in Sprague Dawley rats (Wellenius et al., 2004). 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 post-exposure period
when administered alone. In a follow-up publication, results from the analysis of supraventricular ectopic
beats (SVEB) were provided (Wellenius et al., 2006b). 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.
5.2.3.4. Summary of Cardiovascular Toxicology
Recent 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 myocardial ischemia. Furthermore short-term exposure
to 50 ppm CO exacerbated cardiac pathology and impaired function in animal models of hypertrophic
cardiomyopathy and/or pulmonary hypertension. Ventricular hypertrophy was observed in healthy rats in
response to chronic exposures of 100-200 ppm CO. These studies provide a strong basis for the
development of adverse health effects resulting from exposures to CO at environmentally-relevant
concentrations.
March 2009
5-52
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
5.2.4. Summary of Cardiovascular Effects
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 (see Section 5.2). These studies, described in the 1991 and 2000 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 3-6%, with one multicenter study reporting similar
effects at COHb levels as low as 2.4%. No human clinical studies have evaluated the effect of controlled
exposures to CO resulting in COHb levels lower than 2.4%. Human clinical studies published since the
2000 CO AQCD have reported no association between CO and ST-segment changes or arrhythmia;
however, none of these studies included individuals with diagnoses of 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
were not representative of the most sensitive individuals in the population. In fact, the most sensitive
individuals may respond to levels of COHb lower than 2.4%. 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. First of all, endogenous CO production varies as described in Section 4.5, but
generally results in less than 1% COHb. Secondly, nonambient exposures to CO, such as exposure to
ETS, can increase COHb above baseline levels. Ambient exposures will result in an additive increase in
COHb. Using mathematical modeling to predict changes in COHb in healthy inactive adults (Quantitative
Circulatory Physiology [QCP] model, Section 4.2.3), it can be estimated that exposure to 35 ppm CO for
1 h results in an increase of 0.6% COHb over baseline and exposure to 9 ppm CO for 8 h results in an
increase of 0.8% COHb over baseline. Furthermore, 24 h exposure to 3 ppm CO results in an increase of
0.4% COHb above baseline which can also be obtained following 1-h exposure to 30 CO ppm.
Consequently, exposure to CO at concentrations relevant to the NAAQS has the potential to increase
COHb to levels associated with adverse cardiovascular health effects in some individuals.
Findings of controlled human exposure studies are coherent with findings of recent epidemiologic
studies conducted since the 2000 CO AQCD, which observed associations between ambient CO
concentration and ED visits and hospital admissions for IHD, CHF and all-cause cardiovascular disease.
All but one of these epidemiologic studies were conducted in locations where the entire distribution of
CO concentrations were at or below the level of the current NAAQS, with mean 24-h avg concentrations
ranging from 0.5 ppm (Montreal, Canada) to 9.4 ppm (Tehran, Iran) (Table 5-7). A single study reported a
negative association between CO concentration and hospital admissions and ED visits for IHD among all
March 2009
5-53
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
ages; all other associations were positive, with increases in hospital admissions and ED visits for IHD
between 0.2% and 19.8% per standardized increase in CO concentration (Figure 5-1). These recent
studies build upon the conclusions of the 2000 CO AQCD that short-term variations in ambient CO
concentrations are associated with daily hospital admissions for heart disease.
These health outcomes are consistent with a role for CO in limiting 02 availability (i.e., hypoxic
mechanisms) in individuals with CAD. However, recent toxicological studies suggested that CO may also
act through non-hypoxic mechanisms by disrupting cellular signaling. Studies in healthy animals
demonstrated oxidative injury and inflammation in response to 50-100 ppm CO while studies in disease
models demonstrate effects on heart rhythm and exacerbation of cardiomyopathy and vascular remodeling
in response to 35-50 ppm CO. Furthermore, in utero exposure to 150 ppm CO alters postnatal
elecrophysiological maturation in rat cardiomyocytes. Further investigations will be useful in determining
the importance of non-hypoxic mechanisms following environmentally-relevant CO exposures. Taken
together, the evidence from epidemiologic, human clinical and toxicological studies is sufficient to
conclude that a causal relationship is likely to exist between relevant short-term CO exposures and
cardiovascular morbidity.
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. 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; Fodor and Winneke, 1972; Horvath et al.,
1971; Putz et al., 1979). One study reported similar behavioral effects (time discrimination) among a
group of healthy volunteers with COHb levels <3% (Beard and Wertheim, 1967); however, subsequent
studies were unable to replicate these findings at such low exposure concentrations (Otto et al., 1979;
Stewart et al., 1973a). 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). It should be noted that all behavioral studies of
March 2009
5-54
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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. Toxicological Studies
The evidence for toxicological effects of CO exposure in laboratory animal models comes from in
utero or perinatal exposure to relatively low levels of CO (25 to 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 section of the ISA (Section 5.4.2).
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 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 employed rodent models to
show that low level 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 including active avoidance
behavior (150 ppm CO), non-spatial memory (75 and 150 ppm CO), spatial learning (endogenous CO
inhibition), homing behavior (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 in exposed male rodents.
March 2009
5-55
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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 with 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 which is mechanistically
proposed to be mediated by ROS (25 ppm CO). Functional tests of the auditory system of neonatally, low
level CO-exposed rodents, using OAE testing (50 ppm) and amplitude measurements of the 8th cranial
nerve action potential (12, 25, 50, 100 ppm), revealed decrements in auditory function at PND22 and
permanent changes into adulthood using action potential (AP) testing (50 ppm). 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 male rodents. Considering the
combined evidence from controlled human exposure and toxicological studies, the evidence is
suggestive of a causal relationship between relevant short- and long-term CO exposures 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. Various dichotomized measures of birth weight, such as LBW, SGA, and IUGR,
have been the most examined outcomes in air pollution research while preterm birth (PTB), congenital
malformations, and infant mortality are less studied.
In the 2000 CO AQCD only two studies were cited that examined the effect of ambient air
pollution on adverse birth outcomes and both of these studies investigated LBW as an endpoint
(Alderman et al., 1987; Ritz and Yu, 1999). At that time this area of research was in its infancy and since
then there has been increasing interest.
March 2009
5-56
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 weeks prior to birth among a cohort of 97,158 births in southern California
between 1989 and 1993 (Ritz et al., 2000). The exposure assessment within this study was based on data
from fixed site monitors that fell within a 2-mile 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 to 9:00 a.m.) during the
last 6 weeks 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 weeks prior to birth persisted (RR: 1.02 [95% CI: 1.01-1.03]).
Furthermore, when the analyses included variables for either season or other pollutants the CO effect
estimates generally were reduced.
Expanding on this research, Wilhelm and Ritz (2005) 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 weeks prior to birth. Among women living within a 1-mile 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 forN02 and 03 (RR: 1.05 [95% CI: 1.00-1.10]), but not with the inclusion of
PMio 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 weeks 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 (Ritz et al., 2007) were able to analyze
detailed maternal information from a survey of 2,543 of 6,374 women sampled 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, N02, 03, and PM2.5 during the first trimester and last 6 weeks 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, 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 vs. 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
March 2009
5-57
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
were wider due to the smaller sample. Any possible association between CO and PTB was less evident
during the last 6 weeks 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, though not statistically significant association, with 24-h CO
concentration during the entire pregnancy (OR: 1.03 [95% CI: 0.98-1.09] per 0.5 ppm increase), the first
month of gestation (OR: 1.05 [95% CI: 0.99-1.10] per 0.5 ppm increase), and the last 2 weeks of gestation
(OR: 1.00 [95% CI: 0.96-1.04] per 0.5 ppm increase) (Huynh et al., 2006). Although there was an
indication of an effect during early pregnancy, the small sample size (when compared to other studies)
may have influenced the lack of 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). This study
investigated maternal exposures to ambient gaseous pollutants (CO, N02, S02, 03) 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 2001-2002, CO concentrations during the
first trimester was 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). 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]). To assign the maternal exposures to CO, this study used a kriging technique, which
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.
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 1998-2000 and found no association between PTB and CO (Jalaludin et al., 2007).
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 three months of
March 2009
5-58
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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 last three months 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 5-6 shows the risk ratios 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; Ritz et al., 2007;
Wilhelm and Ritz, 2005). 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.
March 2009
5-59
DRAFT - DO NOT CITE OR QUOTE
-------
Study
Location
Exposure
Period
Exposure
Area
Ritz et al (2000)
California
Month 1
<2 miles of monitor
Ritz et al (2000)
California
Last 6 weeks
<2 miles of monitor
m-
Wilhelm & Ritz
(2005)
Los Angeles, CA
First trimester
ZIP code level
—
Wilhelm & Ritz
(2005)
Los Angeles, CA
Last 6 weeks
ZIP code level
Huynh et al (2006)
California
First month
County level
Huynh et al (2006)
California
Last two weeks
County level
Huynh et al (2006)
California
Entire pregnancy
County level
Liuet al. (2003)
Vancouver, Canada
First month
City wide
Liuet al. (2003)
Vancouver, Canada
Last month
City wide
Jalaludin et al.
(2007)
Sydney, Australia
First month
City wide
Jalaludin et al.
(2007)
Sydney, Australia
First trimester
City wide
Jalaludin et al.
(2007)
Sydney, Australia
Last month
City wide
1
0.75 1.00 1.25
Effect Estimate
Figure 5-6. 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.
March 2009 5-60 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 5-12. Brief summary of PTB studies.
Study
Location (Sample Size)
Mean CO (ppm)
Exposure Assessment
Exposure Window
Ritz et al. (2000)
California
(n = 97,158)
2.7 (6-9 a.m.)
<2 miles of monitor
Month 1
Last 6 weeks
Wilhelm and Ritz (2005)
Los Angeles, CA
(n = 106,483)
1.4(24 h)
Varying distances to monitor
Month 1
Last 6 weeks
Ritz et al. (2007)
Los Angeles, CA
(n = 58,316)
0.87 (24 h)
Nearest monitor to ZIP code
Entire pregnancy
Trimester 1
Last 6 weeks
Huynh et al. (2006)
California
(n = 42,692)
0.8 (24 h)
County level
Entire pregnancy
Month 1
Last 2 weeks
Liu etal. (2003)
Vancouver, Canada (n = 229,085)
1.0 (24 h)
City wide average
Month 1
Last month
Leem et al. (2006)
Incheon, Korea
(n = 52,113)
0.9 (24 h)
Residential address within Dong-based on
Kriging
First trimester
Last trimester
Jalaludin et al. (2007)
Sydney, Australia
(n = 123,840)
0.9 (8 h)
City wide average and
<5 km from monitor
First month
First trimester
Last trimester
Last month
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 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 low
birth weight (LBW) (<2,500g [5 lbs, 8 oz]) and SGA.
It should be noted that the terms small for gestational age (SGA), which is defined as a birth weight
<10th percentile for gestational age (and often sex), and intrauterine growth restriction (IUGR) are used
interchangeably. However, this definition of SGA does have limitations. For example, using this
definition of IUGR may overestimate the percentage of 'growth-restricted' neonates as it is unlikely that
10% of neonates have growth restriction (Wollmann, 1998). 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 Piatt, 2008).
Nevertheless, the terms SGA and IUGR are often used interchangeably and it therefore 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. Meaning that all IUGR neonates will be SGA, but
March 2009
5-61
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
not all SGA neonates will be IUGR (Wollmann, 1998). 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, in order to avoid overlap of the studies 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 and is briefly summarized here. Ritz and Yu (1999) examined the effect of ambient CO
during the last trimester on LBW 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 LBW.
Building upon this research, Wilhelm and Ritz (2005) reported similar results when extending this
study to include 136,134 births for the period of 1994-2000. Exposure to ambient CO during each
trimester was based on data recorded at monitoring stations of varying proximities to the mother's
residence. For women residing within 1 mile of a station, there was 36% (OR: 1.36 [95% CI: 1.04-1.76])
increased risk of having a term LBW baby for women in the highest CO exposure category
(1.84 ppm-75th percentile) for the third-trimester. There was also an increased risk of term LBW (OR:
1.28 [95% CI: 1.12-1.47]) among women in the highest exposure group when the analyses included
women within a 5 mile radius of a station. However, when the analyses included women within a 1-2
mile, or 2-4 mile 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 LBW (results not
reported).
Further validation in association with exposure times was observed in an analysis using a susbset of
participants in the Children's Health Study. Salam and colleagues (2005) found that CO only during the
first trimester was associated with reduced fetal growth. Their research examined birth weight, LBW, and
IUGR among a subset of participants in the Children's Health Study (Peters et al., 1999b) who were born
in California between 1975-1987 (n = 3901). The study examined term births with a gestational age
between 37-44 weeks. Exposures in this study were based on CO data from up to the three 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
March 2009
5-62
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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
significant association with LBW (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
weeks 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 miles of the mother's residence, exposures to CO and PM2 5 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 PM25. PM25 was
found to have a strong effect on birth weight within each trimester (Parker et al., 2005).
Two similar studies were conducted in the Northeast of the 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 1994-1996 found that exposure to ambient 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). 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 PMi0 and S02 were 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 32-44 weeks gestation between
1999-2002 in Connecticut and Massachusetts (Bell et al., 2007). 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, PM2 5, N02, and S02) 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]). 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 23.3 and 27.2 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. Whereas, in a previously cited California study by Huynh and
colleagues (Huynh et al., 2006) 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.
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 (37-42 weeks gestation) in Vancouver, BC
with monthly and trimester exposures to CO investigated in relation to LBW and IUGR (Liu et al., 2003).
March 2009
5-63
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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
LBW. The more recent of these 2 studies examined 386,202 singleton term births (37-42 weeks gestation)
in Calgary, Edmonton and Montreal between 1986 and 2000 (Liu et al., 2007). 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 N02 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 LBW 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). The second of these studies examined 311,735 singleton births (37-41 weeks
gestation) between 1998 and 2000 and reported a 6.0 g 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). It is important to note that neither of these studies found an association between CO exposure and
an increased risk of LBW. Therefore, despite CO during the first trimester being associated with reduced
birth weight, it was not associated with LBW.
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. For example, Ha and colleagues (2001) 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 LBW (RR: 1.10 [95% CI: 1.05-1.14] per 0.5 ppm increase
in 24-h CO concentration). When Lee and colleagues (2003a) extended this study to include singleton
term births (37-44 weeks gestation) for the period of 1996 to 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 LBW (OR: 1.04 [95% CI: 1.01-1.07] per 0.5 ppm increase). No associations were
March 2009
5-64
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
found in the third trimester for any of the pollutants. Monthly-specific exposures showed that the risk of
LBW tended to increase with CO exposure between months 2-5 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 1998-2000 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). However, this result did not persist after controlling for other pollutants (PMi0,
N02) and was only 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) 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 from each gestational age
group.
Of all studies reviewed, only two failed to find an association between maternal exposure to CO
and adverse birth outcomes. In northern Nevada, Chen and colleagues (2002) examined CO, PMi0, and 03
exposures among a cohort of 39,338 term births (37-44 weeks 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 LBW. 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 (37-44 weeks gestation) in
Taipei and Kaoshiung, Taiwan between 1995-1997, maternal exposures to CO, S02, 03, N02, and PMi0in
each trimester of pregnancy were examined and only S02 during the third trimester showed evidence of
attributing to LBW. Exposure assessment was based on data from the monitor closest to the centroid of
the mother's residential district and the final analyses only included 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 LBW 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., 2004a).
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 5-7 shows the change in birth weight (grams), Figure 5-8 shows the effect estimates for LBW, and
Figure 5-9 shows the effect estimates for SGA. In general the reported reductions in birth weight are
March 2009
5-65
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
small (ranging ~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 this
may be the result of correlated exposure periods.
Several studies examined various combinations of birth weight, LBW, 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 the decrease in birth weight did not
translate to an increased risk of LBW or SGA. However, it needs to be noted that a measureable change,
even if only a small one, on a population is different than an effect on a subset of susceptible births which
may increase the risk of IUGR/LBW/SGA.
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; Parker et al., 2005; Wilhelm and Ritz, 2005) while in some studies it did persist (Bell et al., 2007;
Gouveia et al., 2004; Liu et al., 2003, 2007), and other studies did not report results from multipollutant
models (Ha et al., 2001; Lee et al., 2003a; Maisonet et al., 2001; Medeiros and Gouveia, 2005). 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 LBW among African Americans but not
among whites and hispanics (Maisonet et al., 2001). 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). Parker and
colleagues (2005) also tested for interactions between race and found no significant association.
March 2009
5-66
DRAFT - DO NOT CITE OR QUOTE
-------
Exposure
Study Location Exposure Period ^rea
Bell et al (2007) Massachusetts Entire pregnancy County level
Salam et al. (2005) California First trimester ZIP code level
Salam et al. (2005) California Second trimester ZIP code level
Salam et al. (2005) California Third trimester ZIP code level
Salam et al. (2005) California Entire pregnancy ZIP code level
Chen et al. (2002) Northern NV First trimester County level
Chen et al. (2002) Northern NV Second trimester County level
Chen et al. (2002) Northern NV Third trimester County level
Chen et al. (2002) Northern NV Entire pregnancy County level
Mannes et al. (2005) Sydney, Australia First trimester City level
Mannes et al. (2005) Sydney, Australia Second trimester City level
Mannes et al. (2005) Sydney, Australia Third trimester City level
Mannes et al. (2005) Sydney, Australia Last month City level
Gouveia et al. (2004) Sao Paulo, Brazil First trimester City level
Gouveia et al. (2004) Sao Paulo, Brazil Second trimester City level
Gouveia et al. (2004) Sao Paulo, Brazil Third trimester City level
>
Medeiros et al. (2005) Sao Paulo, Brazil First trimester City level
Medeiros et al. (2005) Sao Paulo, Brazil Second trimester City level
Medeiros et al. (2005) Sao Paulo, Brazil Third trimester City level
-35 -25 -15 -5 0 5 10 15 20
Change in Birth weight (g)
Figure 5-7. 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.
March 2009
5-67
DRAFT - DO NOT CITE OR QUOTE
-------
Author
Location
Exposure Period
Exposure
Area
Bell et al (2007)
Connecticut &
Massachusetts
Entire pregnancy
County level
Maisonet et al. (2001)
Northeastern USA
First trimester
City level
Maisonet et al. (2001)
Northeastern USA
Second trimester
City level
Maisonet et al. (2001)
Northeastern USA
Third trimester
City level
Wilhelm and Ritz (2005)
Los Angeles County, CA
Third trimester
ZIP code level
Salam et al. (2005)
California
First trimester
ZIP code level
Salam et al. (2005)
California
Second trimester
ZIP code level
Salam et al. (2005)
California
Third trimester
ZIP code level
Salam et al. (2005)
California
Entire pregnancy
ZIP code level
Liuet al. (2003)
Vancouver, Canada
First month
City level
Liuet al. (2003)
Vancouver, Canada
Second month
City level
Ha et al (2001)
Seoul, Korea
First trimester
City level
Ha et al (2001)
Seoul, Korea
Third trimester
City level
Lee et al. (2003a)
Seoul, Korea
First trimester
City level
Lee et al. (2003a)
Seoul, Korea
Second trimester
City level
Lee et al. (2003a)
Seoul, Korea
Third trimester
City level
.
Lee et al. (2003a)
Seoul, Korea
Entire pregnancy
City level
I 1 1-
0.75 1.00 1.25
Effect Estimate
Figure 5-8. Summary of effect estimates (95% confidence intervals) for LBW 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.
March 2009 5-68 DRAFT-DO NOT CITE OR QUOTE
-------
Exposure Exposure
Period Area
Salam et al. (2005)
California
First trimester
ZIP code level
Salam et al. (2005)
California
Second trimester ZIP code level
Salam et al. (2005)
California
Third trimester
ZIP code level
Salam et al. (2005)
California
Entire pregnancy ZIP code level
Liu et al. (2007)
3 Cities, Canada
First trimester
City level
Liu et al. (2007)
3 Cities, Canada
Second trimester City level
Liu et al. (2007)
3 Cities, Canada
Third trimester
City level
Mannes et al. (2005) Sydney, Australia
Mannes et al. (2005) Sydney, Australia
Mannes et al. (2005) Sydney, Australia
Mannes et al. (2005) Sydney, Australia
Second trimester City level
Third trimester
First trimester
Last month
City level
City level
City level
Liuet al. (2003)
Liuet al. (2003)
Liuet al. (2003)
Liuet al. (2003)
Liuet al. (2003)
Vancouver, Canada First month
Vancouver, Canada First trimester
Vancouver, Canada Second trimester City level
Vancouver, Canada Third trimester
Vancouver, Canada Last month
City level
City level
City level
City level
0.75
1.00
Effect Estimate
1.25
Figure 5-9. 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.
March 2009
5-69
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-13. Brief summary of birth weight studies.
Study
Outcomes
Examined
Location (Sample Size)
Mean CO (ppm)
Exposure Assessment
Exposure Windows
UNITED STATES
Ritzand Yu (1999)
LBW
Los Angeles, CA
(n = 125, 573)
2.6 (6-9 a.m.)
<2 miles of monitor
Trimester 3
Wilhelm and Ritz
(2005)
LBW
Los Angeles County, CA
(n = 136,134)
1.4 (24 h)
Varying distances from monitor
Trimesters 1, 2, 3
Salam et al. (2005)
Birth weight
LBW
IUGR
California
(n = 3901)
1.8(24 h)
ZIP code level
Entire pregnancy
Trimesters 1, 2, 3
Parker et al. (2005)
Birth weight
SGA
California
(n = 18,247)
0.75 (8 h)
<5 miles from monitor
Entire pregnancy
Trimesters 1, 2, 3
Maisonet et al. (2001)
LBW
Boston, MA; Hartford, CT;
Philadelphia & Pittsburg, PA;
Washington DC
(n = 103,465)
1.1 (24 h)
City wide average
Trimesters 1, 2, 3
Belief al. (2007)
Birth weight
LBW
Connecticut and
Massachusetts,
(n = 358,504)
0.6(24 h)
County level average
Entire pregnancy
Trimesters 1, 3
Chen et al. (2002)
Birth weight
LBW
Northern Nevada,
(n = 36,305)
0.9 (8 h)
County level
Trimesters 1, 2, 3
CANADA
Liuet al. (2003)
LBW
IUGR
Vancouver, Canada
(n = 229,085)
1.0 (24 h)
City wide average
Trimester 1
Liu et al. (2007)
IUGR
Calgary, Edmonton, Montreal,
Canada
(n = 386,202)
1.1 (24 h)
City wide average
Trimesters 1, 2, 3
SOUTH AMERICA
Gouveia et al. (2004)
Birth weight
LBW
Sao Paulo, Brazil
(n = 179,460)
3.7(8 h)
City wide average
Trimesters 1, 2, 3
Medeiros et al. (2005) Birth weight
Sao Paulo, Brazil
3.0 (24 h)
City wide average
Trimesters 1, 2, 3
LBW
(n = 311,735)
(Presented in
graph)
AUSTRALASIA
Ha et al. (2001)
Birth weight
LBW
Seoul, Korea
(n = 276,763)
1.2 (24 h)
City wide average
Trimesters 1 and 3
Lee et al. (2003a)
LBW
Seoul, Korea
(n = ?)
1.2(24 h)
City wide average
Entire pregnancy
Trimesters 1, 2, 3
Mannes et al. (2005)
Birth weight
Sydney, Australia
0.8(8 h)
City wide average and
Trimesters 1, 2, 3
SGA
(n = 138,056)
<5 km from monitor
Last 30 days
Lin et al. (2004b)
LBW
Taipei, Kaoshiung, Taiwan
(n = 92,288)
Taipei 1.1,
Kaoshiung 8.1
<3 km of monitor
Entire pregnancy
Trimesters 1, 2, 3
March 2009
5-70
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
5.4.1.3. Congenital Anomalies
Despite the growing evidence of an association between ambient air pollution and various adverse
birth outcomes, only three studies investigated the effect of temporal variations in ambient air pollution
on congenital anomalies. Given the higher prevalence and associated mortality, heart defects have been
the main focus of two of these three recent air pollution studies. The other study's focus was cleft
lip/palate.
The first of these studies was conducted in southern California (Ritz et al., 2002). Exposure to
ambient CO, N02, 03 and PMi0 during each of the first three months 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 and 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 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 03 during the second month of pregnancy, which was
associated with aortic artery and valve defects.
The second study was conducted in Texas (Gilboa et al., 2005), where exposure to ambient CO,
N02, S02, 03 and PMi0 during the 3rd to 8th week of gestation was examined among births between
1997-2000. 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 8.6-14.2 km with maximum distances
ranging from 35.5-54.5 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 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). Whereas, 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 S02
and ventricular septal defects.
March 2009
5-71
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The last of these three studies was a case-control study that examined maternal exposure to various
air pollutants during the first three months 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,
S02, NOx, 03, and CO were averaged over each of the first three months of pregnancy. Interestingly, of
all the pollutants examined, only 03 during the first two months 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).
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 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, and on the contrary, reported results of a protective effect.
Interestingly, similar inconsistencies were also found for PMi0 between these studies. For example,
PM10 in Texas was associated with an increased risk of atrial septal defects, 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) provide little discussion toward the
inconsistent results with the southern California study. One suggestion is 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) also suggested the inconsistency was due to different exposure periods. In Texas the
exposures were averaged over the 3rd to 8th week 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 little 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.
March 2009
5-72
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
5.4.1.4. Neonatal and Post-Neonatal Mortality
A handful of studies examined the effect of ambient air pollution on neonatal and post-neonatal
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 of 1998-2000 in association with concurrent day exposure to S02, CO, 03,
PMio. Moving averages from 2 to 7 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 S02 (r = 0.55), only PMi0 and S02 were associated with an increase in
the daily rate of neonatal deaths (Lin et al., 2004a).
Post-Neonatal
Two studies in the U.S. examined the potential association between ambient CO and post-neonatal
(from 28 days to 1 year 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 from a monitoring site within
the South Coast Air Basin of California. Exposures for 2-week, 1-month, 2-month, and 6-month 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). Ambient 1-h max CO concentrations averaged over
the 2 months 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, N02, 03) the positive CO mortality effect
decreased by around 50% and failed to reach statistical significance. Based on exposure from 2 weeks
before death, CO was associated with an increased risk of respiratory related post-neonatal deaths
occurring 28 days to 1 year afterbirth (OR: 1.14 [95% CI: 1.03-1.25] per 1 ppm increase) and 28 days to
3 months afterbirth (OR: 1.20 [95% CI: 1.02-1.40]), but no effect was observed for respiratory related
deaths occurring 4-12 months 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 LBW infants.' When these two strata were analyzed, CO was
March 2009
5-73
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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).
The second of these 2 studies examined 3,583,495 births, including 6,639 post-neonatal deaths
occurring in 96 counties throughout the U.S. (in counties with more than 250,000 residents) between
1989 and 2000 (Woodruff et al., 2008). Only exposure during the first two months of life was examined
and these were 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 two months 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 post-neonatal
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.
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 post-neonatal deaths and ambient
air pollutants, including CO. The data analyzed were from the cities of Taipei (Yang et al., 2006) and
Kaohsiung (Tsai et al., 2006b) with ambient CO concentrations being 1.6 ppm and 0.8 ppm respectively.
Both studies examined deaths for the period of 1994-2000 and employed a case-crossover design that
compared air pollution levels 1 week before and after each infant's death.
Similarly, another study in South Korea examined post-neonatal mortality for the period of 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 post-neonatal
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).
In general, the inconsistent exposure periods examined among these studies allows for limited
direct comparison and interpretation. Nevertheless, there is limited evidence that CO is associated with an
increased risk of infant mortality during the post-neonatal period. The exposure periods examined varied
from the same-day CO to lag periods up to a 6 month period prior to birth with one study alternatively
exploring exposures averaged over the first two months 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 exposures,
neonatal exposures, and post-neonatal exposures, and the possible interaction that birth weight and
gestational age may have on the results yielded from these examined exposure periods, needs further
attention within this field of research.
March 2009
5-74
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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. Only two studies examined the effects of CO on birth defects.
Both studies found maternal exposure to CO to be associated with an increased risk of cardiac birth
defects. This insult to the heart is coherent with the CO effects on the heart characterized in Section 5.2.
In general, there is limited evidence that CO is associated with an increased risk of infant mortality during
the post-neonatal period.
There is evidence of ambient CO during pregnancy having a negative effect on fetal growth. In
general, the reviewed studies (Figures 5.7 through 5.9) reported small reductions in birth weight (ranging
-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, LBW, 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 LBW 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.
5.4.2. Toxicological Studies of Birth Outcomes and Developmental
Effects
5.4.2.1. Birth Outcomes
Decreased Birth Weight
Multiple reports have been published associating low level CO exposure in laboratory animals and
decrements in birth weight; some of these studies noted reduced growth rates persisting in the neonatal
period. Prigge et al. (1977) saw significant decreases in near-term fetal body weight (GD21) after 21 days
of continuous exposure of pregnant Wistar rats to CO (125, 250, or 500 ppm). Fechter and Annau (1977)
exposed pregnant rats to 150 ppm CO continuously during gestation via inhalation and found 5%
significantly decreased birth weights at PND1 in gestationally exposed pups versus control animals with
weight decrements persisting to weaning; lactational cross fostering did not ameliorate the reduced
growth rates. Dams exposed to CO during gestation had COHb over gestation of 15% with control dams
March 2009
5-75
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 dams versus air-exposed control dams.
Penney et al. (1983) exposed pregnant rats to CO (157, 166, and 200 ppm) over GD6-GD19 and
found significant decreases in near term fetal rat weight at GD20; gestation in rats is ~ 22 days. Carmines
et al. (2008) exposed Sprague-Dawley rats to -600 ppm CO 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. Dam COHb was 30% immediately after
exposure. The half life in rats at this exposure level is ~ 42 minutes. Maternal body weight was
unchanged during gestation, but corrected terminal body weight (body weight minus uterine weight) was
significantly elevated in CO-exposed dams at term. There were no external malformations (teratogenicity)
seen. Singh et al. (1992; 2006) found significant decreases in birth weight in gestationally CO-exposed
mouse pups (65, 125, 250 or 500 ppm) in two studies. In the 2006 study, mice were exposed to CO for 6
h/day for the first 2 weeks of pregnancy, and in the 1992 sutdy, dams were exposed to CO (GD0-GD18)
longer with fetuses collected at GD18. Astrup et al. (1972) found significant decreases (11 and 20%,
respectivelty) in birth weight of rabbits exposed to either 90 or 180 ppm CO continuously over the
duration of gestation. Fechter and Annau (1977) reported decreased birth weight, albeit not significant, of
Long-Evans rats exposed in utero to 150 ppm CO continuously throughout gestation (dam COHb 15%),
but by PND4 and through pre-weaning to PND21, exposed pups weighed significantly less than controls.
Tolcos et al. (2000a) found significant decreases in body weight and crown to rump length in guinea pig
fetuses after being exposed 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). However, in other studies, Tolcos
found no significant differences in birth weight of guinea pig pups after a similar exposure (GD23-GD25
to term). During pregnancy, fetal and maternal COHb levels were 13% and 8.5%, respectively.
Pregnancy Loss and Perinatal Death
Schwetz et al. (1979) exposed CF-1 mice or New Zealand rabbits to 250 ppm CO for either 7 h/day
or 24 h/day over GD6-GD15 (mice) or GD6-GD18 (rabbits), yielding 4 exposure paradigms. The fetuses
were then collected by C-section at the termination of exposure, which was 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 7 h/day CO had a significant increase in the
number of resorbed pups (not seen in 24 h/day CO exposure). Fetal mouse weight was significantly
greater than control in the 7 h exposures and significantly less in the 24 h exposure groups with
corresponding significant differences in crown to rump length in the two groups. Rabbits seemed to be
less affected by CO exposure manifesting with no significant perinatal death or pregnancy loss.
Astrup et al. (1972) studied the effect of CO on fetal development after continuous CO exposure
(90 or 180 ppm CO) over the duration of gestation in rabbits. COHb was 16-18% and 8-9% in the 180
March 2009
5-76
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
and 90 ppm exposure groups, respectively. 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
hours 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 180 ppm CO exposure level. Gestation length was
unchanged with CO exposure.
Fechter and Annau (1977) 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 or pup number at PND1.
Maternal Diet
Maternal Protein Intake and Neonatal Mouse Mortality
Pregnant CD-I mice were exposed intermittently (6 h/day for first two weeks 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 CO exposure coupled with low-protein
diet on neonatal mortality at 1 week of age (Singh, 2006). 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 week of
age, pup mortality was significantly increased by CO-exposure. Dietary protein restriction also induced
pup mortality at 1 week of age; all pups in the 4% protein diet died by 1 week of age. CO exposure
(65 ppm only) combined with a normal protein diet significantly increased pup mortality at 1 week.
Animals receiving supplemental protein diets with CO exposure (65 and 125 ppm) had pups that had
significant increases in mortality at 1 week versus control air pups. Control protein diet pups had
significantly increased pup mortality at 1 week with CO exposure (65 ppm only) versus control air
animals (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 week of age versus control (0 ppm CO). In
summary, these data show that in utero CO exposure induced increased neonatal mouse deaths at 1 week
in supplemental protein and normal protein diet exposure groups.
Maternal Low Protein Diet and CO-Dependent Teratogenicity
The role of diet as a contributor to teratogenicity of CO (0, 65, 125, or 250 ppm CO) in CD-I mice
given a various protein diets (27%, 16%, 8% or 4% protein) during pregnancy was explored by Singh et
al. (1992). Timed pregnant CD-I mice were exposed continuously to CO from GD8-GD18 of pregnancy
at which point animals were sacrificed and fetuses collected. Subsequent 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 et al., 1992; 2003). As expected, all levels of CO and the lowest protein diet (8 or 4% protein)
March 2009
5-77
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
given to the dams during gestation resulted in significantly decreased fetal weight of normal fetuses at
GDI8. 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 malformations including jaw changes. All concentrations of
CO exposure significantly increased the percentage of litters with malformations. There was a CO dose-
dependent increase in percent of litters with malformations within each maternal dietary protein level.
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 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.
Zinc-Supplemented Protein-Deficient Maternal Diet and Mortality and Teratogenicity
Further studies by Neggers and Singh (2006) explored how teratogenicity and fetal mortality were
affected by zinc (Zn) modulation in CO-exposed pregnant dams (CD-I mouse) given protein insufficient
diets. Developmental toxicitiy of CO was attenuated by protein supplementation, i.e., protein
supplemented animals (27%) showed a significantly lower incidence of fetal mortality versus 8% and
16% protein groups. Further, dietary restriction of both protein and Zn with co-exposure to CO during
gestation increased the incidence of pup mortality and malformations including gastroschisis.
Earlier studies by Schwetz et al. (1979) found fetal skeletal alterations in lumbar ribs or spurs were
significantly increased in the 7 h/day and 24 h/day gestationally CO-exposed fetal CF-1 mice collected
near term after CO exposure (250 ppm), over GD6-GD15 (dam gestational COHb 10-15% for 7 h/day
exposure, 24 h/day dam COHb not measured); these changes were not seen in similarly exposed fetal
rabbits (1979).
Astrup et al. (1972) 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 (n = 123) in the 180 ppm CO group had deformities in their extremities at birth,
whereas no control and no 90 ppm CO-exposed animals manifested with this malformation.
Skeletal Abnormalities
Earlier studies by Schwetz et al. (1979) found fetal skeletal alterations in lumbar ribs or spurs were
significantly increased in the 7 h/day and 24 h/day gestationally CO-exposed fetal CF-1 mice collected
near term after CO exposure (250 ppm), over GD6-GD15 (dam gestational COHb 10-15% for 7h/day
exposure, 24 h/day dam COHb not measured); these changes were not seen in similarly exposed fetal
rabbits.
March 2009
5-78
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Astrup et al. (1972) 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 (n = 123) in the 180 ppm CO group had deformities in their extremities at birth,
whereas no control and no 90 ppm CO-exposed animals manifested with this malformation.
Endogenous CO
Spontaneous Abortions
Idiopathic spontaneous abortions are more frequent in women with HO-1 polymorphisms in their
genome (Denschlag et al., 2004). To evaluate the role of HO-1 in spontaneous abortion, a mouse model
that spontaneously undergoes abortion (CBA/J x DBA/2 J mice) (Zenclussen et al., 2006) was used with
and without HO adenovirus treatment to see if pregnancy outcome could be modulated by changing HO
concentration. HO-1 is known to protect organs from rejection (Kotsch et al., 2006) and thus, HO activity
may protect the developing fetus from rejection by the non-self maternal immune system. Pregnancy
outcome was significantly better (abortion rate significantly decreased) in mice overexpressing HO due to
adenovirus transfer. Thus, in this model, it appears that upregulation of the HO/CO system is able to
protect the developing fetus from spontaneous abortion.
Depressed Vascular Reactivity during Pregnancy and CO
CO through the production of soluble guanylate cyclase is able to stimulate the relaxation of
vascular smooth muscle (Villamor et al., 2000). Further, the role of HO expression is important in the
maintenance and outcome of pregnancy and lactation. During pregnancy, there is increased blood volume
without a concurrent increase in systemic BP; this is accomplished by a decrease in total peripheral
vascular resistance, to which CO contributes (Zhao et al., 2008). In humans, genetic polymorphisms in
HO-1 (microsatellite polymorphisms associated with altered HO-1 transcription) are linked to idiopathic
recurrent miscarriages (Denschlag et al., 2004) and administering HO-inhibitors to pregnant rodents
induced total litter loss (Alexandreanu and Lawson, 2002). Various pathologies of pregnancy including
IUGR and pre-eclampsia are associated with significant decreases in placental HO activity (Denschlag et
al., 2004; McLaughlin et al., 2003). Thus, the HO/CO system appears to be crucial in maintaining
pregnancy.
CO and Vasuiar Relaxation during Pregnancy
Isolated rat aortic rings and tail artery rings from pregnant dams (Sprague-Dawley rats) can be
relaxed by submersion in exogenous CO solutions (Longo et al., 1999). Further, the administration of the
HO inhibitor SnMP induced increased maternal BP (systolic, diastolic, and mean arterial pressure) during
pregnancy in FVB mice (Zhao et al., 2008). Zhao also showed pregnancy induced increased total body
CO excretion as measured in inhalation chambers, and that this increased CO production could be
March 2009
5-79
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
significantly decreased by SnMP administration. Pregnant dam abdominal aortas (AA) are significantly
dilated with pregnancy and SnMP treatment leads to AA vasoconstriction to levels similar to non-
pregnant mice. Zhao looked at blood flow changes using Doppler technology and observed a significant
increase in uterine artery blood flow velocity following SnMP administration. 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.
Placenta
Women living at high altitude (chronic hypoxia exposure), women with pre-eclampsia, or women
who had pregnancies with fetal growth restrictions (FGR) produced term placenta with significant
decreases in HO-2 versus women living at lower altitude with uncomplicated pregnancies (Barber et al.,
2001; Lyall et al., 2000). FGR and pre-eclampsia HO-2 changes were detected in endothelial cells; HO-1
immunostaining was very low in the placenta. Women living at high altitude have an increased risk of
adverse pregnancy outcomes versus women living at lower altitudes (Zamudio et al., 1995). Oxygenation
is important in early pregnancy and triggers trophoblast invasion of the spiral arteries (Kingdom and
Kaufmann, 1997). Isolated human placenta exposed to solutions containing CO demonstrated a
concentration-dependent decrease in perfusion pressure (Bainbridge et al., 2002) further demonstrating
the role of CO in maintaining basal vasculature tone.
The endogenous generation of CO in the placenta has been demonstrated in chorionic villi of term
placenta (McLaughlin et al., 2001) with HO activity highest in the placenta near term (McLean et al.,
2000). CO can be generated from multiple endogenous sources; 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
(Ahmed et al., 2005). Term human placental cell types including syncytiotrophoblasts and
cytotrophoblasts were grown in cell culture under basal and hypoxic conditions to explore changes in HO
expression (Newby et al., 2005). HO-1 was expressed at significantly lower amounts in
syncytiotrophoblasts versus cytrotrophoblasts at normoxic conditions. 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. Nonetheless,
these data should be interpreted remembering that isolated cell culture models lack the interaction with
the intact placenta and may lead to different physiological outcomes.
Uterus at Parturition
The addition of exogenous CO to isolated human and rat uterine tissue during pregnancy failed to
induce relaxation of uterine tissue (Longo et al., 1999). Thus, exogenous CO failed to quiet the
spontaneous contractility of rat or human myometrium (uterine smooth muscle). CO is not able to relax
March 2009
5-80
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
all types of vascular smooth muscle (Brian et al., 1994), and pregnancy appears to modulate the response
of tissues to CO (Katoue et al., 2005).
HO-1 induction and HO concentration have been shown to be regulated by estrogen in the rat
(Sprague Dawley) uterus (Cella et al., 2006) during pregnancy and in non-gravid rats. This agrees with
work by Tschugguel et al. (2001) in which CO was generated by primary endothelial cells from human
umbilical veins and uterine arteries after exogenous 17-(3 estradiol administration.
Ovaries
Ovarian Follicular Atresia. As a part of normal follicular maturation in the ovaries, the majority of
follicles undergo artresia via apoptosis prior to ovulation. Harada et al. (2004) 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.
Ovarian Steroidogenesis. HO-1 and HO-2 are localized in the ovaries in rats (Sprague Dawley)
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 (Alexandreanu and Lawson, 2002). CrMP significantly
decreased ovarian production of gonadotrophin-induced androstenedione and progesterone without
affecting estradiol levels. Hemin treatment caused androstenedione and estradiol production from rat
ovaries in vitro. Thus, the HO/CO pathway may play an important role in rat ovarian steroidogenesis.
Anterior Pituitary and Heme Oxygenase
HO-1 and HO-2 are expressed in rat (Sprague Dawley) anterior pituitary and the secretion of
gonadotropins and prolactin is affected by HO inhibitors and HO substrates (Alexandreanu and Lawson,
2003a). The estrogen-induced afternoon surge of LH was advanced forward in time by chronic
administration of the HO inhibitor CrMP and this advance could be reversed by concomitant
administration of hemin, a HO substrate. The serum FSH surge was unaffected by CrMP 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 CrMP and CrMP+hemin
and CrMP significantly decreased prolactin release. In vitro studies using pituitary explants showed that
LH release was significantly increased by CrMP administration and unaffected by hemin. Modulation of
the HO/CO system in the anterior pituitary of the female rat led to altered secretion of gonadotropins and
prolactin.
March 2009
5-81
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Lactation and Pregnancy and HO Modulation
The role of the HO/CO system in estrous cyclicity, pregnancy, and lactation were monitored using
a Sprague Dawley rat model (Alexandreanu and Lawson, 2002). Pregnancy outcomes used dams that
were injected daily (GD5-GD16) with hemin, a HO substrate, or CrMP, a HO inhibitor that were followed
until 2-days post-term at which point they were euthanized and uteri examined for implantation sites and
fetuses. Another group of female rats were monitored for estrous cycle changes after receiving daily
hemin (10 or 11 days of inj ections) or CrMP (11 or 14 days of inj ections). A third group of animals was
monitored during lactation after receiving either CrMP or hemin exclusively during lactation (lactation
days LD5-LD15); the lactational effect was monitored by weighing pups out to weaning (LD21). Hemin
had no effect on any of the outcomes in this study, except a slight effect on litter weight gain during
lactational exposure. CrMP decreased time in estrous in a dose-dependent manner. CrMP exposure
significantly increased uterine fetal resorptions and led to no live births, possibly due to vasoconstriction
and associated ischemia of the placental vascular bed. CrMP induced decreased litter weight gain during
lactation, which the authors attributed to decreased maternal milk production or ejection problems as
cross-fostered pups regained weight lost during nursing on CrMP dams. The lactational effects seen in
this model may be explained by changes in prolactin as previously reported (Alexandreanu and Lawson,
2002). These data indicate that the HO/CO pathway is important in estrous cyclicity, parturition, and
lactation. Nonetheless, CrMP may also inhibit NO* production, a mechanism that is distinct from
CO-dependent effects.
Summary of Toxicological Studies on Birth Outcomes
Endogenous CO production by various organs and the vascular system during pregnancy is
important for maintaining pregnancy and for the proper development of the fetus and/or offspring
postnatally. Exogenous CO addition has been shown to alter perinatal effects and birth outcomes in
various toxicological studies. Endogenous CO production is required for maintenance of a pregnancy as
has been shown by use of HO inhibitors; further, genetic polymorphisms in HO-1, which are associated
with increased HO expression upon stress, are associated with idiopathic recurrent miscarriages in a study
of a Caucasian population. A rodent study of spontaneous abortion demonstrated better pregnancy
outcome in animals in which HO-1 levels were increased by adenovirus transfer.
Rodent studies have explored the role of exogenous CO exposure on birth outcomes. Exogenous
CO exposure to rodent fetuses in utero significantly increased postnatal mortality. CO exposure in utero
induced teratogenicity in offspring in a dose-dependent manner that was further exacerbated by dietary
protein restriction or zinc depletion. Two studies noted skeletal abnormalities following in utero CO
exposure (180-250 ppm).
CO induces cGMP production by vascular smooth muscle beds, which can regulate vascular tone.
Thus, endogenous CO production is associated with maintenance of decreased peripheral vascular
March 2009
5-82
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
resistance during pregnancy that prevents the increased blood volume of pregnancy from affecting BP.
This has been demonstrated in the human placenta and cells from the placenta with exogenous CO
addition further decreasing perfusion pressure. However, exogenous CO addition to human and rat uterine
tissue during pregnancy was unable to relax spontaneous myometrial contraction.
Estrogens have been shown to upregulate HO-1 levels in the uterus in non-gravid and pregnant rat
uteri; other studies have shown isolated cell culture systems administered 17-beta estradiol can generate
CO. In the female reproductive tract, the endogenous HO/CO system is involved in proper follicular
maturation, ovarian steroidogenesis, the secretion of gonadotropin and prolactin by the anterior pituitary,
in maintenance of lactation, and estrous cyclicity in rodent studies.
5.4.2.2. Developmental Effects
CNS Developmental Effects
Behavioral
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 GD0-GD20 (De Salvia et al., 1995). 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), 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) found there were no significant behavioral impairments in the low dose animals (75 ppm).
Animals exposed to the higher dose of CO (150 ppm) in utero had significantly impaired acquisition (at
3 and 18 months of age) and reacquisition (at 18 months of age) of conditioned avoidance behavior. 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 low level CO exposure in utero can lead to permanent behavioral changes in male
offspring.
Mactutus and Fechter (1984) also found that acquisition 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%) 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. Prenatal CO exposure induced learning and memory
deficits in male and female offspring.
March 2009
5-83
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Habituation and Non-Spatial Working Memory. Investigators have used animal models to study
how low dose CO exposure (75 or 150 ppm) during gestation can affect behavioral outcomes
(habituation, exploration of novel objects, and motor activity) in offspring as adults. Giustino et al. (1999)
exposed primparious pregnant Wistar rats to CO (0, 75 or 150 ppm) by inhalation from GD0-GD20.
Blood COHb concentrations (mean %± SEM) on GD20 have been reported (0 ppm: 1.6 ± 0.1; CO
75 ppm: 7.36 ± 0.2; CO 150 ppm: 16.1 ± 0.9). At age 40 weeks, male offspring were given two 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. Exploration time was defined as time exploring objects
during each trial. Global habituation was quantified as a comparison of the time spent exploring the two
objects in Tl to 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 time spent exploring the
new object. This object recognition tests for the preference that rats have for investigating novel objects in
lieu of familiar objects and is a measurement of non-spatial working memory. The results of this study
showed gestationally CO (75 and 150 ppm) exposed animals at 40 days of age have a significantly
decreased Tl during exploration of novel objects. Global habituation of control and 75 ppm CO rats
showed exploratory times with T2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
activity at PND1 and PND4 (both after subcutaneous L-DOPA administration to induce movement) and at
PND14, but not at PND21. CO-exposed pups showed significantly less activity than air-exposed controls
through the pre-weaning window.
Fechter and Annau (1980) exposed rats to CO (150 ppm) continuously throughout gestation and
found that exposed offspring manifested with poorer than normal performance on the negative geotaxis, a
reflexive response that results in a directional movement with or against gravity. In these studies, negative
geotaxis was defined as performing a 180° turn to face the top of an inclined plane. Continuous prenatal
CO exposure (125 ppm, GD7-GD18) in CD-I mice impaired negative geotaxis at PND 10 in a study by
Singh (1986). 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).
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-1 mice on PND 1. 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) (Singh, 1986) at PND14.
These behavioral tests indicated neuromuscular, vestibular or postural effects in the CO-exposed neonate.
Earlier studies by Fechter and Annau (1977) identified an early window of sensitivity for CO-dependent
motor activity defecits of PND1-PND4, with recovery by PND21. Carratu's (2000) PND40 and PND90
monitoring of motor activity of in utero CO-exposed animals may be too late to detect CO-dependent
changes.
Neuronal
Sciatic Nerve Myelination. In utero exposure (GD0-GD20) to low levels of CO (0, 75 or 100 ppm)
and its effect on sciatic nerve myelination in male offspring was studied in Wistar rats (Carratu et al.,
2000). 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), 150 ppm (GD10: 14.42 ±
0.52; GD20: 16.08 ± 0.88). There were significant increases in %COHb in all CO-exposed animals. 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 mice at 40 and 90 days. In conclusion, in utero
exposure of male rats to CO induces decreased myelination of the sciatic nerve without concomitant
decreases in motor activity.
Brain and Peripheral Nervous System Myelination. The effect of in utero exposure (GD0-GD20)
of Wistar rats to CO (0 or 150 ppm) on pup neuronal myelination was studied in the peripheral nervous
system and the brain of male offspring at 90 days of age (Carratu et al., 2000). CO exposure during
development induces hypoxia, and hypoxia can induce sphingomyelin changes which could lead to
impaired myelination and motor activity decrements. This study reported maternal COHb (mean % ±
March 2009
5-85
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
SEM) as 1.9 ± 0.04 and 16.02 ± 0.98 in control and 150 ppm CO-exposed animals, respectively. Prenatal
CO exposure had no effect on brain sphinganine (SA) or sphingosine (SO) levels in male offspring at
90 days of age. However, the sciatic nerve had significant increases in SO after CO exposure, no changes
in SA and a significant decrease in the SA/SO ratio 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. SO is an intermediate in sphingolipid turnover and SA is an intermediate of de novo sphingolipid
biosynthesis. These results demonstrate that sphingolipid homeostasis in the PNS but not CNS is
interrupted in offspring exposed in utero to CO without a manifestation of changes in motor activity.
Neurotransmitter Changes
Medullar Cholinergic and Catecholaminergic Changes after In Utero CO Exposure and SIDS.
One theory related to Sudden Infant Death Syndrome (SIDS) relates to aberrant development of brain
stem nuclei controlling respiratory, cardiovascular, and arousal activity. To address these changes in an
animal model, Tolcos et al. (2000a) exposed pregnant guinea pigs to CO (0 or 200 ppm) for 10 hours per
day over the last 60% of gestation leading to fetal and maternal COHb levels of 13% (versus control of
0.25%) and 8.5 % (versus control of 1.6%), respectively. Guinea pigs are a good model because they
display a similar time course of CNS development to humans with the majority of development occurring
in utero. Pups and sows were collected near term and CO-exposed pups were found to have significant
decrements in body, brain, and liver weights as well as decreased crown to rump length when compared
to control pups. Medullar volume was also significantly decreased in CO-exposed pups. Neurotransmitter
systems were affected after CO exposure. Specifically, the catecholaminergic system of the brainstem
displayed significant decreases in immunoreactivity for tyrosine hydroxylase (TH), which is likely due to
decreased cell number in specific medullar regions. This was consistent with earlier work showing
aberrant respiratory responses (to asphyxia and C02) of offspring with prenatal CO exposure (McGregor
et al., 1998). The cholinergic system is also affected by prenatal CO exposure with significant increases in
choline acetyl-transferase (ChAT) immunoreactivity of the medulla and no changes in muscarinic
acetylcholine receptor. This is in contrast to human infants with SIDS that show decreased brainstem
muscarinic receptor binding (Kinney et al., 1995). ChAT changes in this study (Tolcos et al., 2000a) were
from areas of the medulla associated with tongue innervation, which is crucial to swallowing and which
may be impaired in SIDS infants.
Mesolimbic Dopaminergic Effects and Impaired Sexual Behavior. Exposure (GD0-GD20) of
pregnant Wistar rats to CO (0, 75 or 150 ppm) and its effects on adult (5 and 10 months of age) male
offspring sexual behavior and mesolimbic dopaminergic function was accessed by Cagiano et al. (1998).
Maternal COHb at GD10 was 1, 7, and 15% (0, 75 and 150 pm CO, respectively) and at GD20 was 1.5, 7,
and 16% (0, 75, 150 ppm CO, respectively). Using the aforementioned parameters, only animals at the
150 ppm CO exposure were significantly affected by CO exposure. At 5 months of age, CO-exposed male
March 2009
5-86
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 change CO-induced
increases in mount to intromission latency or decreases in mount to intromission 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. This demasculination of the CO-exposed offspring
paralleled earlier studies of mice exposed gestationally to hypoxia (Hermans et al., 1993). When rats were
followed at ten months of age, there were no significant changes in copulatory activity or neurochemical
parameters of the CO-exposed male offspring showing a recovery from earlier decrements. In summary,
in utero exposure to CO induced impairment of copulatory sexual behavior in male offspring with
accompanying changes in the mesolimbic dopaminergic system at five months of age. By ten months of
age, these changes were no longer detectable in CO-exposed males.
Dopamine in the Neostriatum Is Affected by In Utero Plus Perinatal CO Exposure. 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).
Cerebellum Weight and Neurotransmitter Changes after In Utero and Perinatal CO Exposure.
Long Evans dams and pups exposed to CO (75, 150, or 300 ppm with COHb 9%, 15% and 24%) over the
duration of gestation and out to PND10, during the period of cerebellular development in the rat which is
equivalent to the in utero cerebellar development in the human, yielded a dose-dependent reduction in
cerebellum wet weight (significant at 150 and 300 ppm) at PND21 with increases in norepinephrine (NE)
concentration in the extrinsic noradrenergic system (a system that terminates in the cerebellum) in
CO-exposed animals monitored from PND14 to PND42, which is in contrast to a lack of NE change seen
in the cerebral cortex with in utero only CO exposure (Storm and Fechter, 1985a, b). With the same
exposures as Storm and Fechter, 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 and Fechter, 1985b). Thus, in these studies, it appeared
that the cerebellum and neurotransmitters of the cerebellum were significantly affected by CO exposure
during cerebellar development.
Neonatal Hyperthermia Effects on Neurotransmitters and Neuroglia. To explore the interaction
of hyperthemia and hypoxia (CO-induced), two risk factors for SIDS, pregnant guinea pigs were exposed
to CO (0 or 200 ppm) for 10 h/day for the last 60% of gestation. At PND4 male pups were exposed to
hyperthermia or ambient temperature as a control. Brains were then collected at 1 and 8 weeks of age. CO
exposure had no effect on litter size or litter birth weight. In utero CO exposure sensitized some areas of
the brain to future hyperthermic insults. Specifically, CO plus hyperthermia induced significant increases
March 2009
5-87
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
in serotonin in multiple brain regions (NTS, DMV, and hypoglossal nucleus) at 1 week of age; this change
was no longer evident at 8 weeks of age. Hyperthermia exposure alone induced decreased met-enkephalin
neurotransmitter immunoreactivity at 1 week of age that was absent at 8 weeks 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 weeks 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
leading to generation of necrotic lesions in the brain and changes in neurotransmitter levels.
Glutamatergic Transmission Impairment by In Utero CO Exposure. Pregnant Wistar rats were
exposed continuously to CO (75 ppm) during gestation (GD5-GD20) (Antonelli et al., 2006). Primary cell
cultures obtained from the cerebral cortex of exposed offspring (PND1) had decreased extracellular
glutamate (basal and K+-evoked) levels versus air exposed controls, which may functionally impair
cortical glutamatergic transmission in CO-exposed offspring, possibly affecting learning and memory.
Electrophysiological Changes
Gestational exposure of pregnant Wistar rats to continuous CO (0, 75 or 150 ppm with dam COHb
for 150 ppm CO of 15%) for the entirety of gestation yielded electrophysiological changes in the
peripheral nervous system with reversible changes (present at PND40 and absent at PND270) in sodium
channel inactivation kinetics and irreversible changes in the sodium equilibrium potential (Carratu et al.,
1993) in CO-exposed pups (75 and 150 ppm) versus control pups. These voltage clamp studies showed
that in utero CO exposure had both reversible and irreversible effects on sodium channels, which are
essential for proper electrophysiological function of the PNS.
The Developing Auditory System
The developing auditory system of rodents has recently been recognized to be a target of CO
exposure. Low concentrations of CO exposure via inhalation (0, 12.5, 25, or 50 ppm) over PND8-PND22
during the period of extensive auditory development and synaptogenesis was studied by Webber, who
looked specifically at the inferior colliculus (IC), an auditory integrative section of the midbrain. Rodent
pups were allowed to be maternally reared (MR) or were removed from their respective dams and
nutritionally supported with gastronomy-reared nutrition (AR), an artificial feeding system or received
AR plus CO exposure (ARCO). AR reared pups were fed a milk substitute comparable to natural rat milk
via intragastric cannulation. AR allowed nursing pups to be exposed to CO without possible
CO-dependent developmental changes being confounded by materal CO co-exposure or lactational
March 2009
5-88
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
exposures. Half of the pups were collected at PND27 and half were collected at PND75-PND77. Brains
were sectioned and stained for c-Fos, a marker of neuronal activation in the nervous system. AR and MR
exposed animals were found to have no significant differences in c-FOS staining. Further, c-Fos
immunoreactivity in the central IC was significantly decreased in the ARCO-exposed animals at both
PND27 and PND75-PND77 over all dose groups (12.5, 25, or 50 ppm CO); immnunostaining of other
subregions of the IC were not affected by CO. These studies showed exposure to low concentrations of
CO during development can lead to permanent changes in the auditory system of mice that persist into
adulthood.
Others have shown that mild CO exposure early in development impacts cochlear development
(Lopez et al., 2003). Sprague Dawley rats were reared from PND6 to weaning (PND19-PND20) on milk
substitutes using gastronomy-feeding (AR) or maternally reared (MR). These pups were exposed to low
levels of CO (12 or 25 ppm, ARCO) via inhalation from PND6-PND27. At PND27, the animals were
sacrificed and examinations of the cochlea were performed to determine the effect of mild CO exposure.
In ARCO animals, there was no evidence of damage to the inner or outer hair cells. After CO exposure
(25 ppm), there was atrophy or vacuolization of the nerve cells that innervate the inner (not outer) hair
cells. Also, fibers of the 8th cranial nerve at the level of the internal auditory canal of the ARCO animals
exposed to 25 ppm CO had distorted myelination and vacuolization of the axoplasm. Energy production
markers in the organ of corti and spiral ganglion neurons including cytochrome oxidase and NADH-TR
were significantly decreased in 25 but not 12 ppm CO exposure groups versus control (AR and MR).
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 when compared to controls. Together, these
findings show that specific areas of the cochlea were affected after low level developmental CO exposure.
Two non-invasive measurements of auditory function, otoacoustic emissions testing (OAE) and
auditory brainstem response (ABR) are routinely performed on the majority of human newborns in
hospitals throughout the U.S. to detect early hearing loss. OAE requires the insertion of a microphone and
earphone into a sleeping newborn's ear; when a sound is generated, an echo is recorded on the
microphone in a functional ear and absent or reduced in an affected ear. The ABR measures brainwaves
generated after exposure to a sound. Otherwise healthy, term human infants born to smoking mothers
have impaired cochlear development, albeit mild, with decreased amplitudes of transient evoked
otoacoustic emissions versus newborns born to non-smokers (Korres et al., 2007); CO is one of many
potential affective components of cigarette smoke.
Stockard-Sullivan et al. (2003) used these two functional tests in Sprague-Dawley rat pups
receiving ARCO (12, 25, or 50 ppm CO) to determine how continuous (22 h/day) perinatal CO exposure
(PND6-PND22) functionally affected hearing in the developing rat. OAEs were measured in two ARCO
groups (25 and 50 ppm) at PND22. At 50 ppm, significant reductions in OAE amplitude were detected at
specific frequencies (7.13 and 8.01 kHz). ABR conduction time was not affected in CO-exposed animals
March 2009
5-89
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
(12, 25, 50, 100 ppm). Using another otoacoustic test revealed significant attenuation of the action
potential of the 8th cranial nerve with ARCO exposure (12, 25, and 50 ppm CO) versus AR controls at
PND22, but this is complicated by the finding that AR control animals had significant attenuation of the
8th cranial nerve AP versus MR control animals, implying that artificial diet contributes to AP changes
before CO was introduced. Nonetheless, the ARCO-dependent attenuation of the 8th cranial nerve AP
(versus AR control) was permanent, persisting until adulthood in the 50 ppm CO exposure group (the
only CO group monitored); OAE was not measured in adult animals. These functional tests reported that
neonatal exposure to low concentrations of CO can induce auditory functional changes in rodents.
Multiple high dose CO studies have shown a role for ROS in CO-dependent ototoxicity. Studies
using high dose CO (i.p. injection of pure CO [35 mL/kg] inducing 40% COHb) have shown
CO-dependent otoxoxicity in adult guinea pigs as measured by loss of threshold of cochlear compound
action potentials (CAP) could be attenuated using the free radical inhibitors PBN (a spin trap) or
allopurinol (a superoxide scavenger), implicating ROS in high dose CO dependent cochlear damage
(Fechter et al., 1997). Fechter et al. (2002) found that noise-induced hearing loss (NIHL) was potentiated
by CO (500 ppm) coexposure. In the same manuscript, Fechter et al. (2002) showed potentiation of NIHL
post-CO exposure (1,200 ppm for 30 minutes in adult rodents) with significant elevations seen in free
radical production above control animals only in the CO+NIHL group; an interesting finding of this study
was that the acute high dose CO exposure did not induce significant increases in free radical generation
with this high dose acute CO exposure alone. A possible mechanism for this high-dose CO cochlear
damage is via glutamate release. Use of an NMDA inhibitor attenuated acute CO-dependent (1,200 ppm
or i.p. injection of pure CO [35 mL/kg]) CAP threshold disruption at 15 minutes post CO-exposure.
Glutamate is a known ROS generator (Liu and Fechter, 1995) . The glutamate receptor blockade did not
afford long-term protection (1 h or 4 weeks post-CO exposure) from high doses of CO (1,200 ppm or i.p.
injection of 35ml/kg pure CO) (Chen et al., 2001; Liu and Fechter, 1995)
Cognizant of the ROS-dependent contribution to the auditory changes seen with high dose
CO-exposure in animal toxicology models, Webber et al. investigated whether limiting iron availability
through dietary iron restriction could confer protection against low dose perinatal CO-dependent auditory
decrements. Briefly, rat pups were reared from PND6 through the duration of lacation with gastronomy-
reared nutrition with adequate iron (AR) or AR with iron-deficient diets (ARID) or pups were reared with
their respective dams (MR). Animals exposed to CO were denoted as ARCO and ARIDCO and received
either 25 or 100 ppm CO from PND9 to PND24 with all animals collected on PND27. ARID, ARIDCO,
and ARCO groups were compared to controls (MR and AR) to determine differences in the auditory
system after low iron exposure and/or CO exposure. Coincidentally, ARIDCO mice became anemic but
ARID mice were not. Cochlea were stained for transferrin, an iron-transport protein, and low iron
exposure (ARID and ARIDCO) induced increased transferrin. Neurofilament loss from the spiral
gangilian neurons and somas after ARCO treatment was rescued (no detectable neurofilament loss) with
March 2009
5-90
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
low iron (ARIDCO); ARID 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 IC, but c-fos levels were unchanged
after low iron diet concomitant with CO exposure (ARIDCO). These authors postulated that ROS
generated via the interaction of peroxide and iron (via the Fenton reaction or Haber Weiss chemistry) can
lead to an oxidative stress which leads to decreased c-fos expression in the IC and decreased numbers of
neurofilaments; decreasing the available iron to the developing animal 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. The changes in c-fos are not completely explained as ARID
mice also show significant decreases in c-fos IC staining.
Summary of Toxicological Studies on Developmental Central Nervous System Effects
Toxicological studies employing rodent models have shown that low level CO exposure during the
in utero 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 in male offspring after in utero CO exposure. Other studies have shown
that after in utero CO exposure, adult male offspring manifest with altered behavior including a lack of
habituation and an absence of familiarity with previously viewed objects (non-spatial memory). In two
separate studies, in utero CO exposure (75 and 150 ppm) was associated with significant PNS
myelination decrements without associated changes in motor activity in adult animals. With in utero CO
exposure peripheral myelination was decreased and the sphingolipid homeostasis (150 ppm CO) was
disrupted; CNS myelination was not affected. Multiple studies demonstrated that in utero CO exposure
affected glutamatergic (75 ppm), cholinergic (200 ppm), catecholaminergic (200 ppm), and dopaminergic
(75 ppm) neurotransmitter levels or transmission in exposed male 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, in utero CO exposure has been
shown to affect the developing auditory system of rodents, inducing permanent changes into adulthood.
The IC showed decreased c-Fos staining indicating decreased neuronal activation post-perinatal CO
exposure (12 and 25 ppm CO). The cochlea also showed adverse effects after in utero CO exposure with
atrophy of nerve cells innervating the inner hair cells (25 ppm CO), impaired myelination of the 8th
cranial nerve (25 ppm CO), and decreased energy production in the organ of corti (25 ppm CO). Auditory
functional testing using OAE (50 ppm CO) and attenuation of the 8th cranial nerve AP (12, 25, 50,
100 ppm CO) on rodents exposed perinataly to CO showed that CO-exposed nenonates had auditory
decrements at PND22 (OAE and AP) and into adulthood with AP decrements. Further studies limiting
March 2009
5-91
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
iron consumption in the diet protected CO-exposed mice from cochlear damage at the spiral ganglion,
possibly through a diminuation of ROS. Together, these animal studies demonstrate that in utero or
perinatal exposure to CO can adversely affect adult behavior, neuronal myelination, neurotransmission,
and the auditory system in rodents.
Cardiovascular and Systemic Developmental Effects
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). Specifically, at 4 weeks 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 It0 (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 weeks of age. These
CO-dependent changes resolved by 8 weeks of age, reflecting a delayed maturation. The authors noted
that human SIDS is increased in infants whose mothers are smokers and cigarette smoke is composed of
approximately 4% CO. 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. However, no SIDS-like cardiac abberations
were followed in intact Holter-monitored rats in this study.
Heart Morphological changes after In Utero or Perinatal CO Exposure
Multiple authors have reported cardiomegaly following in utero low level CO exposure. Prigge and
Hochrainer (1977) 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) found that prenatal
exposure to CO affected cardiac development in exposed offspring. Long Evans rats that were exposed to
CO continuously (150 ppm) during gestation manifested with significant elevations in wet heart weight at
PND1 as well as heart weight in relation to body weight; body weight was significantly decreased in
CO-exposed pups. At no other age measured (PND4, PND 14, or PND21) did the CO-exposed pups show
increased heart weight (absolute or relative wet or dry weight) versus control. At PND 14, control pups
that had significantly increased body weight, also had significantly increased absolute heart weights (dry
and wet) versus CO-exposed offspring. 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) studied
CO-dependent (500 ppm) cardiomegaly in neonates (continuous CO exposure for 32 days starting at
March 2009
5-92
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
PND1). Other studies of adult male Charles River derived rats exposed to CO for 6 weeks (at 400 or
500 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).
Systemic Immune Toxicology after In Utero CO Exposure
In utero exposure (GD0-GD20) of male Wistar rats to relatively low CO (0, 75, or 150 ppm)
concentrations induced reversible changes in macrophage function (Giustino et al., 1993). The killing of
Candida albicans (yeast) by splenic macrophages was significantly decreased at PND15 in in utero
CO-exposed male offspring (75 and 150 ppm) but recovered and maintained function by PND21 and
PND60. Macrophage phagocytosis of Candida 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 dose 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 Wistar rats to CO
exposure induced changes in the frequency of splenic immunocompotent cells (Giustino et al., 1994).
Specifically, there was a significant decrease in the number of leucocyte common antigen cells (LCA+)
cells in PND21 male rats exposed during gestation to 150 ppm CO; other cell subpopulations including
macrophages, Major Histocompatibility (MHC) II cells, T and B lymphocytes did not show significant
decreases in cell numbers with with CO exposure.
Summary of Toxicological Studies of Cardiovascular and Systemic Development
In utero CO exposure is associated with various adverse, albeit non-persistent, cardiac aberrations.
Exposure to 150 ppm induced a delayed maturation of the cardiac action potential in CO-exposed
offspring. Specifically, APD failed to shorten at 4 weeks of age but shortened to control duration by 8
weeks of age. Also, ion channels in CO-exposed animals, which are related to APD, were significantly
different from controls at 4 weeks but matured to control levels by 8 weeks of age. In other studies,
continuous in utero CO exposure (60, 125, 250, and 500 ppm) induced cardiomegaly at PND1 which was
transient and regressed by PND4. Systemic immunocompromise was displayed in two studies focusing on
splenic cells. In the first study, a reversible and dose dependent loss of function of splenic macrophages
with decreased killing ability, decreased phagocytosis, and decreased ROS production during the
macrophage respiratory burst was observed (150 ppm but not 75 ppm CO). In the second study, the
distribution of splenic immunocompotent cells was 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
low doses of CO (60, 125, 150, 250, or 500 ppm) is able to induce transient changes in cardiac
March 2009
5-93
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
morphology, cardiac action potentials, and systemic immunity that may make the immediate neonatal
period a time for a CO-exposed animal more susceptible to other outside stressors.
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, which
included only two studies that examined the effect of ambient CO on LBW. Since then, a number of
studies have been conducted looking at varied outcomes, including PTB, birth defects, fetal growth
(including LBW), 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 results 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. 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.
This insult to the heart is coherent with results of human clinical studies demonstrating the heart as a
target for CO effects (Section 5.2). Animal toxicological studies provide additional evidence for such an
insult to the heart, and reported transient cardiomegaly at birth after continuous in utero CO exposure (60,
125, 250 and 500 ppm CO), delayed myocardial electrophysiolgical maturation (150 ppm CO), or
systemic splenic immunocompromise (75 or 150 ppm CO). Toxicological studies have also shown that
exogenous continuous in utero CO exposure (250 ppm) induced teratogenicity in rodent offspring in a
dose-dependent manner that was further exacerbated by dietary protein restriction (65 ppm CO) or zinc
depletion (500 ppm CO). Toxicological studies of exogenous 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 SGA/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
March 2009
5-94
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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.
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 do 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 abberations 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 OAE (50 ppm CO) and 8th cranial nerve AP amplitude measurements (12, 25, 50, 100 ppm
CO) on rodents exposed perinataly to CO showed that CO-exposed nenonates had auditory decrements at
PND22 (OAE and AP) and permanent changes into adulthood with AP (50 ppm CO).
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 non-hypoxic mechanisms that could lead to such effects 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 a causal relationship
between long-term exposures to relevant CO concentrations and developmental effects and birth
outcomes
5.5. Respiratory Effects
5.5.1. Epidemiologic Studies with Short-Term Exposure
5.5.1.1. Pulmonary Function, Respiratory Symptoms, and Medication Use
The 2000 CO AQCD 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,
March 2009
5-95
DRAFT - DO NOT CITE OR QUOTE
-------
1 or respiratory disease. As a result, the 2000 CO AQCD did not conclude that a causal association exists
2 between short-term exposure to CO and respiratory health effects. The following section evaluates the
3 current literature that examines the potential association between short-term exposure to CO and
4 respiratory health effects. Table 5-14 lists the studies evaluated in this section along with the respiratory
5 health outcomes examined and CO concentrations reported.
Table 5-14. Range of CO concentrations reported in key respiratory morbidity studies that
examined effects associated with short-term exposure to CO.
Author
Location
Years
Health Outcome
Metric
Mean
Concentration
(ppm)
Middle/Upper Percentile
Concentrations (ppm)
Rabinovitch et al.
(2004)
Denver, CO
Year 1: n = 41
Year 2: n = 63
Year 3: n = 43
11/1999-3/2000;
11/2000-3/2001;
11/2001-3/2002
Pulmonary function; 24-h avg
Medication use
1.0
50th: 0.9
75th: 1.2
Maximum: 3.5
Silkoff et al.
(2005)
Denver, CO
Year 1: n = 16
Year 2: n = 18
1999-2000 (winter);
2000-2001 (winter)
Pulmonary function; 24-h avg
Medication use
1999-2000:1.2
2000-2001:1.1
1999-2000
50th: 1.10
75th: 1.43
Maximum: 3.79
2000-2001
50th: 0.975
75th: 1.34
Maximum: 2.81
Fischer etal.
(2002)1
The Netherlands
n = 68
March - April2
Pulmonary function
24-h avg
0.80
Max: 1.34
Ranzi et al.
(2004)1
Emilia-Romagna
Region, Italy
n = 120
2/1999-5/1999
Pulmonary function; 24-h avg
Respiratory
symptoms;
Medication use
Urban: 1.34
Rural: 1.06
NR
Lagorio et al.
(2006)1
Rome, Italy
(n = 29)
5/1999-6/1999;
11/1999-12/1999
Pulmonary
Function
24-h avg
Spring: 1.83
Winter: 10.7
Overall: 6.4
Overall
Max: 25.1
Penttinen et al.
(2001)1
FHelsinki, Finland
n = 57
11/1996-4/1997
Pulmonary function
24-h avg
NR
50th: 0.35
75th: 0.43
Maximum: 0.96
Timonen et al.
(2002)1
Kuopio, Finland
n = 33
2/1994-4/1994
Pulmonary function
24-h avg
0.52
Maximum: 2.43
Chen et al.
(1999)
Taiwan
n = 941
5/1995-1/1996
Pulmonary function
1-h max;
24-h avg
NR
1-h max
Maximum: 3.6
Delfino et al.
(2003)
Los Angeles, CA
n = 22
11/1999-1/2000
Asthma symptoms
1-h max;
8-h max
1-h max: 7.7
8-h max: 5.0
1-h max
90th: 12.0
Maximum: 17
8-h max
90th: 7.9
Maximum: 10
Slaughter et al.
(2003)
Seattle, WA
n = 133
12/1993-8/1995
Asthma symptoms;
Medication use
24-h avg
NR
50th: 1.47
75th: 1.87
Yu et al. (2000)
Seattle, WA
n = 133
11/1993-8/1995
Asthma symptoms
24-h avg
1.6
50th: 1.47
Maximum: 4.18
Schildcrout et al.
(2006)
8 North American
cities
n = 990
11/1993-9/1995
Asthma symptoms;
Medication use
24-h avg
NR
50th: 0.63-1.49
75th: 0.77-1.90
90th: 0.95-2.40
March 2009
5-96
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Author
Location
Years
Health Outcome
Metric
Mean
Concentration
(ppm)
Middle/Upper Percentile
Concentrations (ppm)
von Klot et al.
(2002)1
Erfurt, Germany
n = 53
10/1996-3/1997
Asthma symptoms;
Medication use
24-h avg
0.78
50th: 0.70
75th: 1.04
Maximum: 2.60
Park et al.
(2005a)
Incheon, Korea
n = 64
3/2002-6/2002
Asthma symptoms;
Medication use
24-h avg
Control days: 0.64
Dust days: 0.65
NR
Rodriguez et al.
(2007)
Perth, Australia
n = 263
6/1996-7/1998
Symptoms
associated with
respiratory illness
8-h max
1.41
Maximum: 8.03
de Hartog et al.
(2003)1
Amsterdam, the
Netherlands
n = 37
Erfurt, Germany
n = 47
Helsinki, Finland
n = 47
1998-1999 (winter)
Respiratory
symptoms
24-h avg
Amsterdam: 0.52
Erfurt: 0.35
Maximum:
Amsterdam: 1.39
Helsinki: 0.35
Erfurt: 2.17
Helsinki: 0.87
1 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 tempterature.
2 These studies did not provide the year(s) in which air quality data was collected.
Pulmonary Function
Rabinovitch et al. (2004) examined the association between exposure to ambient air pollutants and
asthma exacerbation in a panel of urban minority children, 6-12 years old, with moderate to severe asthma
over three winters in Denver, CO. The investigators examined pulmonary function by measuring forced
expiratory volume in 1 second (FEV,) and peak expiratory flow (PEF) in the morning on school days, and
also at night on weekends or other nonschool days. Using a 3-day moving average (lag 0-2) for all
pollutants, Rabinovitch et al. (2004) did not find an association between CO and either lung function
parameter during the morning or at night.
Silkoff et al. (2005) also examined lung function in Denver during the winter months, but in a
panel of former smokers that were at least 40 years old and had been diagnosed with COPD. In this study
CO concentrations were similar to those reported in Rabinovitch et al. (2004). The authors examined the
association between exposure to air pollutants and lung function (i.e., FEV, and PEF) in both the morning
and the evening. Silkoff et al. (2005) found contradictory results when examining the effects of CO for
each of the winter periods separately, 1999-2000 and 2000-2001. During the analysis of the first winter
(i.e., 1999-2000) CO was not found to be associated with lung function decrements at any lag, but an
increase in FEV, during the morning was observed at lag 1. For the second winter (i.e., 2000-2001) the
authors found a significant negative association between CO exposure and FEV, in the evening at lag 2
(Figure 5-10). CO was not found to be associated with PEF at any lag during either winter period. Silkoff
et al. (2005) postulated that the difference in 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,
PM10, PM2 5, and N02 concentrations all increased. The decrements in FEV, observed in the second
March 2009
5-97
DRAFT - DO NOT CITE OR QUOTE
-------
1 winter, therefore, may have been due to the slightly worse, although not significantly different, baseline
2 lung function of the panel of subjects used during the second winter (Silkoff et al., 2005).
AM
0.06
= 0.04
O
Q.
§ 0.02
o
Q
<1>
Q.
V
O)
c
(0
>
LU
-0.02
PM
-0.04
PM2.5
PM10
CO
N02
PM2.5 PM10
CO
N02
0.06
= 0.04
o
Q.
o
c»
c
CO
0.02
-0.02 ¦¦
-0.04
AM
PM
O^-CM O-t-CM O *- CN Ot-CM
DIOOl 0)05 0) 0)00) 0)0)0)
cq co co cotoco toroto w ^ _co
PM2.5 PM10
CO
N02
PM2.5
PM10
CO
N02
Source: Silkoff et al. (2005)
Figure 5-10. Estimates for FEVi change expressed per SD change of the individual pollutants
PM2.5, PM10, CO, and NO2 for the 1999-2000 (top) and 2000-2001 (bottom) winters at
lags 0,1, and 2. The asterisk indicates a significant association (p <0.05).
March 2009
5-98
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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. The results from these studies
contradict those reported in the U.S.-based studies previously discussed. 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) examined the association between CO and lung function in a panel
consisting of 57 non-smoking 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 afternoon ((3 = -1.52, SE = 0.29)
and evening ((3 = -1.81, SE = 0.27) for a 5-day average. 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 a two-
pollutant model at lag 0. In this study, high correlations between UFP and other traffic generated
pollutants (e.g., CO and NOx) make it difficult to attribute the observed respiratory effects to a specific
pollutant.
Lagorio et al. (2006) also conducted a study that examined the association between CO and lung
function in adults. In this study, 3 panels of subjects that resided in Rome, Italy, were selected with
underlying asthma, COPD, or IHD. The ages of the subjects varied depending on the panel, but overall
the subjects ranged from 18-80 years old. In single-pollutant models CO was found to be negatively
associated with both FVC (forced vital capacity) and FEVi 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) did observe high correlations between CO and PM2 5, but not N02.
Unfortunately, copollutant models were not conducted in this analysis to identify whether the CO
associations observed are confounded by other pollutants.
Timonen et al. (2002) examined the effect of CO on bronchial responsiveness and pulmonary
function (i.e., FVC, FEVi, MMEF, and AEFV) in a panel of children 7-12 years 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 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. However, CO was not associated with exercise induced changes in lung function. Overall,
Timonen et al. (2002) found that high concentrations of combustion byproducts (i.e., BS, PMi0, particle
numbers, N02, and CO) were associated with impairment in baseline lung function. These associations,
along with the high correlation between pollutants, 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.
March 2009
5-99
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Chen et al. (1999) examined the effect of CO on lung function in a panel of 8-13 year 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 0800 to
1800) in a single-pollutant model. In addition, the authors found a high correlation between CO and N02
concentrations (r = 0.86-0.98), but did not conduct multipollutant analyses to examine the effect of each
pollutant.
One additional study, Fischer et al. (2002), examined the association between CO and respiratory
health, specifically lung function in a non-selected cohort study of 68 children ages 10-11 that live in an
urban environment (i.e., 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) found negative associations between CO and
FEVi, PEF, and MMEF at both lags 1 and 2. Additonally, the authors found an association between CO
and an increase in eNO at lag 1. However, the study did not present the correlations between pollutants or
examine copollutant models.
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 results were observed across
studies. Studies consisting of children enrolled in the Childhood Asthma Management Program (CAMP)
study found that CO was positively associated with asthma symptoms. Yu et al. (2000) found a 14%
increase in asthma symptoms ([95% CI: 5-23] per 0.5 ppm increase in 24-h avg CO concentrations at lag
1) in a population of 5-13 year old children (n = 133) with asthma in Seattle, WA. These effects persisted
when controlling for previous day's asthma symptoms (12% [95% CI: 5-19] at lag 1). Using the same
population of children, Slaughter et al. (2003) found a significant 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]), respectively.
Schildcrout et al. (2006) examined the association between air pollutants and asthma symptoms in 990
children ages 5-12 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 N02, PMi0, and S02 where included in joint pollutant models.
Two additional U.S. studies also found positive associations between CO and asthma symptoms,
Rabinovitch et al. (2004) and Delfino et al. (2003). Rabinovitch et al. (2004) found a positive association
between 24-h avg CO concentrations for a 3-day moving average (lag 0-2) and asthma exacerbtions (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) also observed positive associations
March 2009
5-100
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
between CO and asthma symptoms in a population of Hispanic children with asthma in a Los Angeles,
CA, community. However, positive associations were found only when using the previous day's
maximum 8-h avg CO concentration as the exposure metric. These results are in contrast to those studies
reported above which found positive associations when using 24-h avg CO concentrations as the exposure
metric. It should be noted that in comparison to Rabinovitch et al. (2004) and the other studies discussed
above, the mean ambient concentrations for 1-h maximum and maximum 8-h avg reported by Delfino et
al. (2003) 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 a panel study consisting of 53 adults with asthma or asthma symptoms in Germany, von Klot
et al. (2002) 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 UFP, making it unclear if the effect observed
was due to CO alone. Additionally, Park et al. (2005a) in a panel study of individuals 16-75 years old in
Incheon, Korea with bronchial asthma did not find an association between CO and nighttime asthma
symptoms or cough. Figure 5-11 summarizes the results from studies that provided usable quantitative
results and examined the association between short-term exposure to CO and asthma or respiratory
symptoms in asthmatic individuals.
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 pollution levels. The majority of
U.S.-based studies (i.e., Rabinovitch (2004), Slaughter et al. (2003), and Schildcrout et al. (2006))
focused on rescue inhaler use in children with ages ranging from 5-13 years old. Rabinovitch et al. (2004)
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 6-12 year old urban minority children with moderate to severe
asthma in the winter in Denver, CO, which is evident by the large confidence intervals surrounding the
estimate. Slaughter et al. (2003) in a population of 5-12 year old children with asthma in Seattle, WA
found a positive and significant 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 multi-city study conducted by Schildcrout et al. (2006), which analyzed
rescue inhaler use in 990 children ages 5-13 with asthma in eight North American cities. Schildcrout et al.
(2006) 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 an increase in other pollutants
(i.e., N02, PMio, and S02) when included in joint models with CO. However, both Slaughter et al. (2003)
and Schildcrout et al. (2006) attributed the associations observed to other combustion byproducts.
Additional studies (Park et al., 2005a; Silkoff et al., 2005; von Klot et al., 2002) conducted in Denver,
March 2009
5-101
DRAFT - DO NOT CITE OR QUOTE
-------
CO; Erfurt, Germany; and Incheon, Korea, respectively, found results that are consistent with those
previously reported, but in populations with combined ages ranging from 16-77. Figure 5-11 presents the
risk estimates from studies that examined the association between short-term exposure to CO and
medication use in asthmatic individuals.
Reference Location Population Age Lag
E
o
"S.
E
>.
CO
o
Q.
CO
2 (1-h max)
Symptom Score >2 (max 8-h avg)
_ Wheeze
Inhaler Use
«- Inhaler Useb
r Inhaler Usec
Inhaler Use
— Inhaler Use (|32-agonist)
—•—Inhaler Use (Corticosteroid)
Adults
Children
Adults
1.0
—I—
1.5
—I 1 1 1 1—I
2.0 2.5 3.0 4.0
Odds Ratios
Figure 5-11. Asthma symptoms, respiratory symptoms and medication use in asthmatic
individuals associated with short-term exposure to CO1. 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 Non-Asthmatic Individuals
In addition to examining the association between short-term exposure to CO and respiratory
symptoms (e.g., cough, wheeze, shortness of breath, etc.) in asthmatic populations some studies examined
these respiratory effects in individuals classified as non-asthmatics. Rodriguez et al. (2007) and de Hartog
1 Effect estimates from Park et al. (2005a) were not included in this figure because the study did not provide the increment at which the effect
estimates were calculated.
March 2009
5-102
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
et al. (2003) examined the effect of CO on respiratory symptoms in a panel of 263 children 0-5 years old
at high risk for developing asthma in Perth, Australia, and a 3-city panel of individuals >50 years of age
with CHD during the winter in Amsterdam, the Netherlands, Erfurt, Germany, and Helsinki, Finland,
respectively. Rodriguez et al., (2007) 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 Perth, Australia. It is
unclear which pollutant is driving the effect observed by Rodriguez et al. (2007) because multipollutant
models were not examined and additional analyses were not conducted to further characterize the
associations observed.
In a panel of elderly individuals with CHF in three European locations, de Hartog et al. (2003)
observed some marginal associations, specifically between CO concentration and the prevalence of the
respiratory symptom shortness of breath, at lag 3 and 0-5-days. Although a marginal association was
observed, the authors found that the associations between air pollution exposure and respiratory
symptoms were stronger for PM2 5 than for gaseous air pollutants.
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 outcomes. Among asthmatics, the studies reviewed generally find positive
associations between short-term exposure to CO and respiratory-related health effects (i.e., decrements in
lung function/lung function growth, respiratory symptoms, and medication use). It is difficult to
determine from this group of studies if CO is independently associated with respiratory outcomes or if
CO is an indicator for other traffic-related 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-related health effects in near-road locations. A lack 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 which 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) evaluated a limited amount of literature that examined the
association between short-term exposure to CO and respiratory hospital admissions, ED visits, and
physician visits in the U.S. (i.e., Seattle, WA, Reno, NY, and Anchorage, AK) and Europe (i.e., Barcelona,
Spain). From these studies, the 2000 CO AQCD concluded that positive associations were observed for
short-term exposure to CO with several respiratory outcomes, including asthma and COPD. However, the
March 2009
5-103
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
lack of a biologically plausible mechanism for CO-induced respiratory morbidity at that time brought into
question the validity of the results observed.
Since the 2000 CO AQCD, the number of studies that examined the association between short-term
exposure to CO and respiratory morbidity has increased; however, the total number of studies published is
still considerably less than the number that examine the health effects associated with exposure to other
criteria air pollutants. This section focuses primarily on those studies conducted in the U.S. and Canada
which examined the potential respiratory health effects associated with CO at concentrations at or similar
to those found in the U.S. Unlike previous sections, which also evaluated studies conducted outside of the
U.S., the expansive U.S.-based respiratory hospital admission and ED visits literature provides adequate
evidence to examine the association between short-term exposure to CO and respiratory HA and ED
visits. Collectively, the studies conducted outside of the U.S. observed associations that are consistent
with those observed in the U.S.-based studies evaluated below (see Annex C).
Overall, this section focuses on respiratory-related hospital admissions because the majority of the
literature examines hospital admissions as opposed to ED visits or physician visits (Table 5-15 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 hospital admissions, 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 hospital admission, ED visit, and physician visit studies, each outcome is
evaluated in individual sections. In addition, each section presents results separately for respiratory health
outcomes which includes 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 5-15.
Range of CO concentrations reported in key respiratory hospital admission and ED
visit studies that examine effects associated with short-term exposure to CO.
Author
Location
Type of Visit (ICD9)
Metric
Mean Concentration (ppm)
Middle/Upper Percentile
Concentrations (ppm)
Cakmak et al.
(2006a)
10 Canadian
cities
Hospital Admissions: Respiratory
disease (i.e., Acute bronchitis and
bronchiolitis; Pneumonia; Chronic and
unspecific bronchitis; Emphysema;
Asthma; Bronchiectasis; Chronic
airway obstruction)
24-h avg
0.8
Maximum: 6.5
Linn et al. (2000)
Los Angeles,
CA
Hospital Admissions: Pulmonary;
Asthma; COPD
24-h avg
Winter: 1.7; Spring: 1.0;
Summer: 1.2; Fall: 2.1
Maximum:
Winter: 5.3; Spring: 2.2;
Summer: 2.7; Fall: 4.3;
March 2009
5-104
DRAFT - DO NOT CITE OR QUOTE
-------
Author
Location
Type of Visit (ICD9)
Metric
Mean Concentration (ppm)
Middle/Upper Percentile
Concentrations (ppm)
Slaughter et al.
(2005)
Spokane, WA
ED Visits and Hospital Admissions:
Respiratory; Asthma; COPD;
Pneumonia; Acute Respiratory
Infection
24-h avg
Hamilton St.: 1.73
Backdoor Tavern: 1.29
Spokane Club: 1.41
Third and Washington: 1.82
Rockwood: 0.42
95th: 3.05
Burnett et al.
(2001)
Toronto, ON,
Canada
Hospital Admissions: Respiratory
disease (i.e., Asthma; Acute
bronchitis/bronchiolitis; Croup;
Pneumonia)
1-h max
1.9
50th: 1.8; 75th: 2.3;
95th: 3.3; 99th: 4.0
Maximum: 6.0
Yang et al. (2003)
Vancouver, BC,
Canada
Hospital Admissions: Respiratory
diseases
24-h avg
0.98
50th: 0.82; 75th: 1.16
Maximum: 4.90
Lin et al. (2003)
Toronto, ON,
Canada
Hospital Admissions: Asthma
24-h avg
1.18
50th: 1.10; 75th: 1.40
Maximum: 6.10
Lin et al. (2004c)
Vancouver, BC,
Canada
Hospital Admissions: Asthma
24-h avg
0.96
50th: 0.80; 75th: 1.12
Maximum: 4.90
Moolgavkar
(2003a) (re-
analysis of
Moolgavkar
2000a)
Cook County,
IL; Los Angeles
County, CA
Hospital Admissions: COPD
24-h avg
NR
Cook:
50th: .99; 75th: 1.25
Maximum: 3.91
Los Angeles:
50th: 1.35; 75th: 2.16
Maximum: 5.96
Yang et al. (2005)
Vancouver, BC,
Canada
Hospital Admissions: COPD
24-h avg
0.71
50th: 0.64
Maximum: 2.48
Karretal. (2006)
South Coast Air
Basin, CA
Hospital Admissions: Acute
bronchiolitis
24-h avg
Lag 1:
Index: 1.730
Referrent: 1.750
Lag 4:
Index: 1.760
Referrent: 1.790
Lag 1:
Index:
50th: 1.52; 75th: 2.26;
90th: 3.16
Maximum: 9.60
Referrent:
50th: 1.51; 75th: 2.29;
90th: 3.23
Maximum: 9.60
Lag 4:
Index:
50th: 1.54; 75th: 2.31;
90th: 3.23
Maximum: 8.71
Referrent:
50th: 1.55; 75th: 2.35;
90th: 3.30
Maximum: 9.60
Karretal. (2007)
South Coast Air
Basin, CA
Hospital Admissions: Acute
bronchiolitis
24-h avg;
Monthly avg
24-h avg: 1.720
Monthly: 1.770
24-h avg:
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
Zanobetti and
Schwartz (2006)
Boston, MA
Hospital Admissions: Pneumonia
24-h avg
NR
50th: 0.48; 75th: 0.60;
95th: 0.88
Lin et al. (2005)
Toronto, ON,
Canada
Hospital Admissions: Respiratory
infections
24-h avg
1.16
50th: 1.05; 75th: 1.37
Maximum: 2.45
March 2009
5-105
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Author
Location
Type of Visit (ICD9)
Metric
Mean Concentration (ppm)
Middle/Upper Percentile
Concentrations (ppm)
Peel et al. (2005)
Atlanta, GA
ED Visits: All respiratory; Asthma;
COPD; URI; Pneumonia
1-h max
1.8
90th: 3.4
Tolbert et al.
(2007)
Atlanta, GA
ED Visits: Respiratory diseases
(i.e., Asthma; COPD; URI; Pneumonia;
Bronchiolitis)
1-h max
1.6
50th: 1.3; 75th: 2.0; 90th: 3.0
Maximum: 7.7
Ito et al. (2007)
New York, NY
ED Visits: Asthma
8-h max
1.31
50th: 1.23; 75th: 1.52;
95th: 2.11
Villeneuve et al.
(2006b)
Toronto, ON,
Canada
Physicians Visits:
Allergic rhinitis
24-h avg
1.1
Maximum: 2.2
Hospital Admissions
Respiratory Disease
All but two of the published hospital admission studies that examined the association between
short-term exposure to CO and all respiratory diseases from North America were conducted in Canada,
and only one study presented results from a combined analysis of multiple cities (Cakmak et al., 2006b).
In a study of 10 of the largest Canadian cities, Cakmak et al. (2006b) examined respiratory hospital
admissions (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). U.S.-based studies (Los Angeles and Spokane) reported similarly weak or null associations for
respiratory disease hospital admissions (Linn et al., 2000; Slaughter et al., 2005). In a study conducted in
Toronto, Canada for the time period 1980-1994, Burnett et al. (2001) reported a relatively strong
association between 1-h max CO and respiratory disease hospital admissions in children less than two
years 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 hospital admissions for a
2-day avg (lag 0-1) per 1 ppm increase in 1-h max CO. In the two-pollutant model analysis, the estimates
for both CO and 03 remained elevated. Yang et al. (2003) 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 years of age)
respiratory disease (ICD-9: 460-519) admissions in Vancouver for the time period 1986-1998. Yang et al.
(2003) also reported elevated associations with 24-h avg CO and respiratory hospital admissions (ICD-9:
codes 460-519) for ages 65 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 risk estimate remained elevated when 03
was included in the model.
1 To determine the lag for the combined estimate across all 10 cities, Cakmak et al. averaged the strongest associations from lags 0-5 days from
each city.
March 2009
5-106
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Asthma
Some studies that examined the effect of short-term exposure to CO on asthma hospital admissions
conducted all age and age-stratified analyses, specifically to examine effects in children. In two of these
hospital admission studies conducted in Canada, evidence was observed for increased pediatric (ages
6-12) asthma hospital admissions (ICD-9: 493) in boys, but not girls (Lin et al., 2003; 2004c); however, a
biological explanation was not provided which could explain this difference. Lin et al. (2003) used a bi-
directional case-crossover analysis in Toronto, Canada for the years 1981-1993 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 hospitalization. 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 a copollutant analysis, the estimates for boys were
essentially unchanged when adjusting for all PM indices (Lin et al., 2003). It should be noted that this
study used a bi-directional case-crossover analysis, which may be biased (Levy et al., 2001). 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). In an
additional analysis conducted by Lin et al. (2004c), the authors found less consistent evidence for a
greater effect in boys versus girls in Vancouver during the years 1987-1998 using a time-series study
design that stratified results by socioeconomic status (SES). In one additional study that examined asthma
hospital admissions for all ages and genders combined, Slaughter et al. (2005) observed some evidence of
an increase in asthma hospital admissions (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, 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 hospital admissions. Moolgavkar (2003a) (a reanalysis of (Moolgavkar, 2000a)
examined hospital admissions 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 percent
increases of 0.53-1.20% for all ages in Los Angeles County, and 0.17-1.41% forages 65 and older in
Cook County, for a 0.5 ppm increase in 24-h avg CO and lags ranging from 0 to 5 days. Yang et al. (2005)
reported similar results for COPD hospital admissions (ICD-9 490-492, 494, 496) in Vancouver for ages
65 and older for the years 1994-1998 for a moving average of 0-6 day lags (RR =1.14 [95% CI: 1.03-
1.23] per 0.5 ppm increase in 24-h avg CO). However, Slaughter et al. (2005) found no association
between short-term exposure to CO and COPD hospital admissions (ICD-9 491, 492, 494, 496) in
March 2009
5-107
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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.
Acute Bronchiolitis in Infants
Two studies (Karr et al., 2006; 2007) from the South Coast Air Basin in California examined both
short term (lag 0 or 1) and longer term levels of CO in relation to acute bronchiolitis (ICD-9: 466)
hospital admissions during the first year of life from 1995-2000. Karr et al. (2006) found no evidence of a
short-term association between ambient CO concentrations and hospital admissions 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), which examined longer term exposures (average in the month prior to hospitalization and lifetime
average) in a matched case-control study, also did not provide any evidence of an association with CO.
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) examined pneumonia hospital admissions (ICD-9 480-487) in Boston, MA, for the
years 1995-1999 for ages 65 and older using a time-stratified case-crossover analysis. The authors
reported an increase in pneumonia hospital admissions 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) 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 N02 (r = 0.67), and moderately correlated with PM2 5 (r = 0.52).
Instead of examining the effect of CO on pneumonia hospital admissions separately as was done by
Zanobetti and Schwartz (2006), Lin et al. (2005) presented results for the overall effect of CO on
respiratory infection hospital admissions (ICD-9: 464, 466, 480-487). In this analysis, Lin et al. (2005)
examined the potential increase in respiratory hospital admissions in children less than 15 years of age in
Toronto, Canada for 1998-2001 using a bi-directional case-crossover approach. The authors reported
elevated estimates for boys (OR= 1.15 [95% CI: 1.02-1.29] 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:
0.92-1.21]). Lin et al. (2005) did not provide an explanation as to why the estimates are stronger for boys
than for girls. It should be noted that this study used a bi-directional case-crossover analysis, which may
be biased (Levy et al., 2001). Studies that have 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).
March 2009
5-108
DRAFT - DO NOT CITE OR QUOTE
-------
Endpoint
Reference
Location
Age
Lag
Burnett et al. (2001)
Toronto, CAN
<2
0-1
1
¦
Yang et al. (2003)
Vancouver, CAN
<3
1
I
1
All
Linn et al. (2000)
Los Angeles, CA
30+
0
¦
h
i
Respiratory
Yang et al. (2003)
Vancouver, CAN
65+
1
•
i
Cakmak et al. (2006b)
10 Canadian Cities
All
2.8
t
>
Slaughter et al. (2005)
Spokane, WA
All
2
i
i
¦
i
i
Lin et al. (2003)
Toronto, CAN
6-12
1
! T Rnyq
Asthma
Lin et al. (2003)
Slaughter et al. (2005)
Toronto, CAN
Spokane, WA
6-12
1
{ fiirk
All
2
¦
1
1
¦
COPD
Slaughter et al. (2005)
Yang et al. (2003)
Spokane, WA
Vancouver, CAN
15+
65+
2
0-6
1
1
¦
1
1
1 *
1
¦
m
1
1
1
I
Pneumonia
Zanobetti and Schwartz
(2006)
Boston, MA
65+
0
1
1 _
1
Acute
Bronchiolitis
Karret al. (2006)
SQAB, CA
<1
1
1
1
1
1
¦
-¦r
i
i
¦
i
Respiratory
Infection
Lin et al. (2005)
Lin et al. (2005)
Toronto, CAN
Toronto, CAN
<15
0-5
i
J . Rnyq
<15
0-5
¦
¦ r
i
1 1 1 1 1
0.9 1.0 1.1 12 1.3 1.4
Effect Estimates
Figure 5-12. Summary of associations between short-term exposure to CO and respiratory
hospital admissions.12 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.
Emergency Department Visits
Respiratory Disease
1 Peel et al. (2005) conducted a large single-city ED study in Atlanta, GA, which included data from
2 31 hospitals for the time period 1993-2000. In this study, results were reported for various respiratory-
1 Risk estimates from Moolgavkar (2003a) 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.
2 Risk estimates from Lin et al. (2004c) 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).
March 2009
5-109
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
related visits (ICD-9 460-466, 477, 480-486, 491-493, 496, 786.09). In an all ages analysis, the authors
found a RR of 1.011 (95% CI: 1.004-1.019) for a 3-day avg (lag 0-2) per 1 ppm increase in 1-h max CO
concentration for all respiratory diseases. Tolbert et al. (2007) expanded the time period used in the Peel
et al. (2005) study to include ED visits through 2004, and reported similar results for respiratory 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 03, N02, or PM were added to the model (). In
addition, Tolbert et al. (2007) reported high correlations between CO and N02 (r = 0.70) and EC
(r = 0.66); and a moderate correlation with PM2 5 (r = 0.51). One additional ED visits study that also
examined respiratory disease (Slaughter et al., 2005) presented essentially null results at lag 1 and 2, but
found similar results to Peel et al. (2005) and Tolbert et al. (2007) at lag 3 (RR = 1.02 [95% CI: 1.00-1.03]
per 0.5 ppm increase in 24-h avg CO).
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). In this study the authors reported results from
distributed lag models including lags 0-13 in addition to a moving average 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 moving average CO concentrations (RR =
1.010 [95% CI: 0.999-1.022] for lags 0-2 compared to RR = 1.076 [95% CI: 1.047-1.105] 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 years) (RR= 1.019 [95% CI: 1.004-1.035] 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). Slaughter et al. (2005), 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) also provide evidence of
increased ED visits for asthma (ICD-9 493) for all ages in New York City for 1999-2002, but quantitative
results were not provided. In addition, Ito et al. (2007) stated that increased CO effect estimates were
attenuated when N02 was included in the model, but effect estimates remained elevated in two-pollutant
models with either PM2 5 or 03.
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) 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 a moving average of lag 0-2) with similar results for the
distributed lag model (RR = 1.03 [95% CI: 0.98-1.09). However, results from Slaughter et al. (2005) from
March 2009
5-110
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Spokane 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 hospital admission analysis conducted by Zanobetti and Schwartz (2006) discussed
above, Peel et al. (2005) 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 three-day
moving average (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 (days 0-13), Peel et al. (2005) observed evidence
of an association (RR = 1.045 [95% CI: 1.01-1.08]). An examination of URI ED visits, the largest of the
respiratory ED groups, found slightly increased risk estimates for both the three-day moving average
(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 PMi0, 03, or N02 were included in the model. Upon conducting an age-stratifed
analysis, Peel et al. (2005) also found that infant (less than one year of age) and pediatric (ages 2-18) URI
ED visit CO risk estimates were substantially stronger than all age risk estimates.
March 2009
5-111
DRAFT - DO NOT CITE OR QUOTE
-------
Endpoint
Reference Location Age Lag
All
Respiratory
Slaughter et al. (2005) Spokane, WA All 3
Peel et al. (2005) Atlanta, GA All 0-2
Tolbert et al. (2007) Atlanta, GA All 0-2
Asthma
Peel etal. (2005) Atlanta, GA 2-18 0-2
Peel etal. (2005) Atlanta, GA All 0-2
Peel etal. (2005) Atlanta, GA All 0-13*
Slaughter et al. (2005) Spokane, WA All 3
COPD
Slaughter et al. (2005) Spokane, WA 15+ 1
Peel etal. (2005) Atlanta, GA All 0-2
Peel etal. (2005) Atlanta, GA All 0-13*
Pneumonia
Peel etal. (2005) Atlanta, GA All 0-2
Peel etal. (2005) Atlanta, GA All 0-13*
Respiratory
Infection
Peel etal. (2005) Atlanta, GA All 0-2
Peel etal. (2005) Atlanta, GA All 0-13*
* Unconstrained distributed lag a® 0JJ6 100 105 110 '¦
Effect Estimates
Figure 5-13. Summary of associations between short-term exposure to CO and respiratory ED
visits. Effect estimates were standardized depending on the averaging time used in
the study: 0.5 ppm for24-h avg, 0.75 ppm for max 8-h avg, and 1.0 ppm for 1-h max.
Physician Visits
1 Although hospital admissions and ED visits are the two most well studied measures of morbidity, a
2 few studies also examined the effect of CO on unscheduled physician visits. In a time-series study,
3 Villeneuve et al. (2006b) examined the effect of CO on physician visits for allergic rhinitis in individuals
4 65 and older in Toronto, Canada. Although quantitative results were only presented in figures, upon
5 observation it was evident that estimates were consistent with a null association for lags 0-6 (Villeneuve
6 et al., 2006b). In an additional study, Sinclair et al. (2004) reported results for urgent care visits for asthma
7 and respiratory infections in a health maintenance organization in Atlanta, GA; however, the study only
8 reported statistically significant results, of which none were for CO.
March 2009
5-112
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Summary of Associations between Short-Term Exposure to CO and Respiratory
Hospital Admissions, ED Visits, and Physicians Visits
Relatively few studies evaluated the association between short-term exposure to ambient CO and
hospital admissions and ED visits for various respiratory outcomes compared to other criteria air
pollutants (e.g., 03 and PM). Although evidence for consistent positive associations (See Figures 5-12 and
5-13) has been found across these studies, various issues surrounding the association between short-term
exposure to CO and respiratory-related health effects have not been addressed due to: the lack of studies
that examined potential confounders of the CO-respiratory hospital admission and ED visits relationship;
and uncertainty as to a biologically plausible mechanism which could explain the association between CO
exposure and respiratory-related health effects. Some of the studies evaluated suggest that CO acts as an
indicator of combustion (e.g., traffic), which is supported by the moderate to high correlation between CO
and other traffic-related pollutants such as N02, PM2.5, EC, or BC and in addition complicates the results
presented. Only two studies examined potential confounding of CO risk estimates by other pollutants
through copollutant models, and found that CO risk estimates were robust or attenuated but remained
positive in two-pollutant models with 03, N02, or PM indices.
5.5.2. Epidemiologic Studies with Long-Term Exposure
The 2000 CO AQCD 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.
5.5.2.1. Pulmonary Function
Mortimer et al. (2008) examined the effect of prenatal and lifetime exposures to air pollutants on
pulmonary function in 232 asthmatic children that 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 individual pollutant metrics
on pulmonary function (Mortimer et al., 2008). In an attempt to remedy this problem 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). This methodology uses an
exploratory process to identify the best predictive model for each outcome of interest. Using this
approach, Mortimer et al. (2008) found that exposure to CO early in life (ages 0-3) was negatively
associated with FEVi/FVC and FEF25.75/FVC resulting in an effect size of -2.5% and -4.8%, respectively,
March 2009
5-113
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
per IQR increase in CO.1 Additional negative associations were observed between exposure to CO during
the first 6 years of life and FEF25 (-6.7%), and prenatal, not trimester-specific exposure, and FEF25.75 (%
reduction not reported). Overall, Mortimer et al. (2008) found that the effects were limited to subgroups,
including African Americans, individuals diagnosed with asthma before the age of 2 years, and
individuals exposed to maternal smoking during pregnancy. 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). 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 N02, which together are markers for pollutants generated by urban
combustion sources (e.g., mobile sources) (Mortimer et al., 2008).
5.5.2.2. Asthma and Asthma Symptoms
All of the studies that examined the association between long-term exposure to CO and asthma
and/or asthma symptoms presented consistent, positive results, which is similar to what was observed
when evaluating the effects attributed to short-term exposure to CO. Two U.S.-based studies, Goss et al.
(2004) and Meng et al. (2007), were evaluated which examined the effect of long-term exposure to CO
on: (1) pulmonary exacerbations in a cohort of individuals with cystic fibrosis >6 years of age, and (2)
poorly controlled asthma in a population of asthmatics >18 years old that resided close to cross-street
intersections in Los Angeles County and San Diego County, respectively. Of these two studies only Goss
et al. (2004) observed a positive association between a respiratory outcome (i.e., pulmonary
exacerbations) and long-term exposure to CO. However, it is unclear if the effects observed are due to CO
alone because the authors did not conduct co-pollutant analyses.
When evaluating studies conducted in other countries, Hirsch et al. (1999), in a study conducted in
Germany, and Guo et al. (1999), Wang et al. (1999), and Hwang et al. (2005), in studies conducted in
Taiwan, all found positive associations between long-term exposure to CO and asthma or asthma
symptoms in populations ranging from 6-16 years old. In these studies, the authors addressed the
observed associations differently. Guo et al. (1999) and Hwang et al. (2005) both concluded that it is
unlikely CO directly affects the respiratory system; Hirsch et al. (1999) attributed the increase in the
prevalence of cough and bronchitis to exposure to traffic-related air pollutants (i.e., N02, CO, and
benzene); and Wang et al. (1999) did not interpret the association between long-term exposure to CO and
adolescent asthma. Only Hwang et al. (2005) conducted a co-pollutant analysis and found that the asthma
effects observed were robust to the inclusion of PMi0, S02 and 03 in the model. This is of note because
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.
March 2009
5-114
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
inclusion of the traffic-related pollutants NOx or N02 in some of the aforementioned studies complicated
the overall results because NOx and N02 are highly correlated with CO, which makes it difficult to
separate the effects attributed to each pollutant.
5.5.2.3. Allergic Rhinitis
Hwang et al. (2006) and Lee et al. (2003c) 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) and Lee et al. (2003c) 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.
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, brings into question
the validity of the associations observed.
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) 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). The 2000 CO
AQCD reported no evidence of CO-induced changes in exercise ventilation at COHb levels <15% during
submaximal exercise (Koike et al., 1991). 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; Vesely et al., 2004)'. The
results of these studies demonstrate that the hypoxia- and C02-induced increases in pulmonary ventilation
March 2009
5-115
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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). Subjects were
exposed to both CO at concentrations of 100-125 ppm as well as room air for 2 h on each of four
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).
5.5.4. Toxicological Studies
As discussed in Section 5.2.3., the work of Thom, Ischiropoulos and colleagues (Ischiropoulos et
al., 1996; Thom and Ischiropoulos, 1997; Thom et al., 1997; Thom et al., 1999a; Thom et al., 1999b)
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, 1-h exposure of rats to 50 ppm CO resulted in
increased lung capillary leakage 18 h later (Thom et al., 1999a). 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 H202 and nitrotyrosine. All of these effects were blocked by inhibition of NOS. These results,
which were partially discussed in the 2000 CO AQCD, demonstrate the potential for exogenous CO to
interact with NO'-mediated pathways and to lead to pathophysiological effects in the lung.
Recent work by Ghio et al. (2008) showed a disruption of cellular iron homeostasis following
exposure to a low level of CO (50 ppm x 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,
March 2009
5-116
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 non-specific
application of exogenous CO will have little capacity to discriminate between excessive and/or
inappropriate iron which catalyzes oxidative stress and iron which may be required for normal
homeostasis.
A chronic inhalation study by Sorhaug et al. (2006) demonstrated no alterations in lung
morphology in Wistar rats exposed to 200 ppm CO for 72 weeks. COHb levels were reported to be 14.7%
and morphological changes were noted in the heart as described in Section 5.2.3.
A recent study by Carraway et al. (2002) 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 (j,m) 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.
In summary, one older study (Thorn et al., 1999a) and two new studies (Carraway et al., 2002; Ghio
et al., 2008) 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).
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, 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. However, the interpretation of the results from epidemiologic studies is
difficult due to the lack of extensive copollutant analyses along with the moderate to high correlation
between CO and other combustion/traffic generated pollutants. To date the majority of the literature has
not extensively examined the association between CO and respiratory morbidity due to studies focusing
March 2009
5-117
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
primarily on effects associated with exposure to other criteria pollutants, namely PM and 03. This has
contributed to the inability to disentangle the effects attributed to CO from the larger complex air
pollution mix (particularly motor vehicle emissions). In addition, uncertainty as to a biological
mechanism to explain the respiratory-related effects observed in the epidemiologic literature further
complicates the interpretation of these results, especially considering the low ambient CO concentrations
reported (24-h avg: 0.35-2.1 ppm). However, animal toxicological studies do provide some evidence that
short-term exposure to CO (50-100 ppm) can cause oxidative injury and inflammation and alter
pulmonary vascular remodeling. Human clinical studies have not extensively examined the effect of
short-term exposure on respiratory morbidity, specifically pulmonary function. The limited number of
human clinical studies that have been conducted prior to and since the 2000 CO AQCD provide very little
evidence of any adverse effect of CO on the respiratory system at COHb levels <10%. Although human
clinical studies have not provided evidence to support CO-related respiratory health effects, the
epidemiologic studies that examined the effects of short-term exposure to CO and lung-related outcomes
show positive associations and animal toxicological studies demonstrate the potential for an underlying
biological mechanism, which together provide evidence that is suggestive of a causal relationship
between short-term exposure to relevant CO concentrations and respiratory morbidity.
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. 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 co-pollutant analyses to disentangle the respiratory effects associated with CO due to its high
correlation with N02 and other combustion-related pollutants. Overall, the evidence available is
inadequate to conclude that a causal relationship exists between long-term exposure to relevant
CO concentrations and respiratory morbidity.
March 2009
5-118
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
5.6. Mortality
5.6.1. Epidemiologic Studies with Short-Term Exposure to CO
The relationship between short-term exposure to CO and mortality has not been extensively
examined over the years due to the majority of epidemiologic studies focusing on mortality effects
associated with exposure to PM and 03. As a result, a clear understanding of the association between
short-term exposure to CO and mortality has yet to be developed. This section summarizes the main
findings of the 2000 CO AQCD, 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 examined the association between short-term exposure to CO and mortality
through the analysis of twelve single-city time-series studies, and one multi-city 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.
The reasons for this conclusion, which are shared with those studies that examined the effect of short-
term exposure to CO on morbidity, were due to: internal inconsistencies and lack of coherence of the
reported results within and across studies; the representativeness of the average ambient CO levels of
spatially heterogeneous ambient CO values derived from fixed monitoring sites 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 levels (<5.0 ppm, 1-h daily
max) reported in the epidemiologic studies (U.S. EPA, 2000). Additionally, some epidemiologic studies
have also suggested that CO is acting as an indicator for other combustion-related pollutants, which has
led to investigators questioning the CO-mortality relationship even when associations have been
observed.
To date the aforementioned issues have not been addressed primarily due to the majority of the
recent time-series mortality studies focusing on the effects of only PM and 03. As such, CO has usually
been considered one of the potential confounding copollutants in air pollution epidemiologic studies. As a
result, the available CO information from these PM and 03 studies most frequently consists of risk
estimates from single- and multipollutant models. Given the limitation that most of these studies were not
March 2009
5-119
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
conducted to examine CO, the goal of this review is to evaluate the CO-mortality association, and
specifically the: magnitude of associations; evidence of confounding; and evidence of effect modification.
5.6.1.2. Multi-City 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 multi-city 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, and thus avoiding potential
publication bias.1 Table 5-16 lists the multi-city studies evaluated along with the mean CO concentrations
reported in each study.
Table 5-16. Range of CO concentrations reported in multi-city studies that examine mortality
effects associated with short-term exposure to CO.
Author
Location
Years
Averaging Time
Mean Concentration
(ppm)
Range of Mean
Concentrations Across
Cities (ppm)
Dominici et al. (2003b;
2005a) (reanalysis of
Samet et al., 2000b)
82 U.S. cities1
(NMMAPS)
1987-1994
24-h avg
1.02
Baton Rouge = 0.43
Spokane = 2.19
Burnett et al. (2004)
12 Canadian cities
1981-1999
24-h avg
1.02
Winnipeg = 0.58
Toronto = 1.31
Samoli et al. (2007)2
19 European cities
(APHEA2)
1990-19973
8-h max
2.12
Basel = 0.52
Athens = 5.3
1 The study actually consisted of 90 U.S. cities, but only 82 had CO data.
2This 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.
3The 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 largest 90 U.S. cities (82 cities for CO) in the National Morbidity,
Mortality, and Air Pollution Study (NMMAPS) Dominici et al. (2003b; 2005a) (reanalysis of Samet et al.,
2000b) is by far the largest multi-city study conducted to date to investigate the mortality effects of air
pollution, but the study primarily focused on PM10. The range in 24-h avg CO concentrations in 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
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.
March 2009
5-120
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
observed in the estimates generated using GAM models with default convergence criteria (Dominici et
al., 2002), Dominici et al. (2003b; 2005a) (reanalysis of Samet et al., 2000b) used the data using GAM
with stringent convergence criteria as well as GLM.
Focusing on the results obtained using GLM, PMi0 and 03 (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 of CO, S02, or N02, with total (non-accidental) mortality. However, as with
PMio, the gaseous pollutants CO, S02, and N02 each showed the strongest association at a 1-day lag (for
03, a 0-day lag). Figure 5-14 presents the total mortality risk estimates for CO from Dominici et al.
(2003b). 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 PMi0 and 03 in the
model did not reduce CO risk estimates. However, the confidence bands were wider in the multipollutant
models, but this could be attributed to: (1) PMi0 data in many of the cities being collected every 6th day,
as opposed to daily data for gaseous pollutants; and (2) 03 being collected in some cities only during
warm months. The addition of N02 (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 cities1) were available for the CO multipollutant analysis with PMi0 and N02 compared to the
single-pollutant CO analysis (82 cities). However, most of the cities that did not have N02 data (26 out of
32), and subsequently were not included in the multipollutant analysis, 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 CO risk estimate in the CO multipollutant analysis with
PM10 and N02. In comparison to the PMi0 risk estimates, which were not reduced in multipollutant
models, the CO risk estimates from multipollutant models indicate less consistent associations with
mortality.
1 One city was excluded from the multipollutant analysis because it contained N02 data, but did not contain CO data.
March 2009 5-121 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
ABCDEABCDEABCDE
Models
Source: Dominici et al. (2003b)
Figure 5-14. Posterior means and 95% posterior intervals of national average estimates for CO
effects on total (non-accidental) mortality at lags 0,1, and 2 within sets of the 90 U.S.
cities with available pollutant data. Models A = CO alone; B = CO + PM10; C = CO +
PM10 + O3; D = CO + PM10 + NO2; E = CO + PM10 + SO2.
Canadian Multi-City Studies
Since the 2000 CO AQCD two Canadian multi-city studies have been published that examined the
association between mortality and short-term exposure to air pollutants: (1) an analysis of PM10, PM25,
PM10_2 5, and gaseous pollutants in 8 cities from 1986 to 1996 (Burnett et al., 2000); and (2) an analysis of
PM10, PM2 5, PM10-2 5, and gaseous pollutants in 12 cities from 1981 to 1999 (Burnett et al., 2004). The
2000 study utilized GAM with default convergence criteria, and upon reanalysis only examined PM
indices (Burnett and Goldberg, 2003).
Burnett et al. (2004) is the most extensive Canadian multi-city study conducted to date, both in
terms of the length of the study and the number of cities covered. Although the study focused on N02
because it was the best predictor of short-term mortality fluctuations among the pollutants examined
(N02, 03, S02, CO, PM2 5, and PM10_2 5), it did present single- and copollutant results for all pollutants
included in the analysis. The mean CO concentrations reported by Burnett et al. (2004) are similar to
those reported in NMMAPS (see Table 5-14).
Burnett et al. (2004) examined the effect of short-term exposure to CO on total (non-accidental)
mortality. The authors found the strongest mortality association at lag 1-day for CO, S02, PM2 5, PM10_25,
PM10 (arithmetic addition of PM2 5 and PM10.2.5), and CoH, whereas for N02, it was the 3-day moving
average (i.e., average of 0-, 1-, and 2-day lags), and for 03, it was the 2-day moving average. In this study,
Burnett et al. (2004) used 24-h avg pollutant concentrations because these values showed stronger
associations than the daily 1-h max values for all of the gaseous pollutants and CoH, but not for 03. In a
single-pollutant model the CO risk estimate for total mortality was 0.33% (95% CI: 0.12-0.54) per
March 2009
5-122
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
0.5 ppm increase in 24-h avg CO with a 1-day lag. After adjusting for N02, the CO risk estimate was
reduced to 0.04% (95% CI: -0.19 to 0.26), while the N02 risk estimate was only slightly affected
(increased from 2.25% to 2.35%) when including CO in the model. In this analysis, a copollutant model
including both CO and PM was not presented. The similarity between the results presented in this
Canadian multi-city and NMMAPS is that, in both analyses, CO risk estimates appeared to be sensitive to
the addition of N02 in the regression model. However, interpretation of these results requires some
caution because: (1) N02 tends to have a more spatially uniform distribution within a city compared to
CO; (2) CO and N02 share common sources (e.g., traffic); and (3) CO and N02 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), S02, N02, and 03, 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), and N02 One study, Gryparis et al. (Gryparis et al., 2004)
did consider CO as a potential confounder in an analysis of 03, and found that the addition of CO
increased 03 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 multi-city studies), the
Samoli et al. (2007) analysis focused specifically on CO.
Samoli et al. (2007) investigated the effect of short-term exposure to CO on total (non-accidental)
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 for the APHEA2 study of 2.12 ppm is higher than the estimated max 8-h avg CO
concentrations reported in U.S. (Dominici et al., 2003b, 2005b) and Canadian (Burnett et al., 2004) cities
of 1.53 ppm.1 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 the correlation between CO and 1-h max N02
(r = 0.03-0.68).
To examine the CO-mortality relationship, Samoli et al. (2007) 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
1 The max 8-h avg concentration for the Dominici et al. (2003b) and Burnett et al. (2004) 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.
March 2009
5-123
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
temporal confounding. These methods consisted of: (1) confining the extent of smoothing to 8 degrees of
freedom per year (df/yr); and (2) selecting the appropriate extent of smoothing through minimization of
the absolute value of the sum of partial auto-correlation functions (PACF) of the residuals, which resulted
in the analysis using on average 5 df/yr for total mortality and 4 df/yr for cardiovascular mortality. The
authors also conducted copollutant analyses using PMi0, BS, S02, N02, or 03 (1 h). In the second stage
model Samoli et al. (2007) examined heterogeneity in CO risk estimates between cities by regressing risk
estimates from individual cities on potential effect modifiers including: a) the air pollution level and mix
in each city (i.e., mean levels of pollutants, ratio PMi0/NO2); b) the exposure (number of CO monitors,
correlation between monitors' measurements); c) variables describing the health status of the population
(e.g., crude mortality rate); d) the geographic area (northern, western, and central-eastern European
cities); and e) the climatic conditions (mean temperature and relative humidity levels).
Samoli et al. (2007) found that CO was associated with total 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 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
mortality and 0.70% (95% CI: 0.31-1.09) for cardiovascular mortality. In the sensitivity analysis, Samoli
et al. (2007) used 8 or 12 df/yr to adjust for temporal confounding. Both approaches led to similar risk
estimates, but using PACF to choose the extent of smoothing generally resulted in larger CO risk
estimates (by -50 to 80%).
During the examination of results obtained from the copollutant models, the authors noted that
there was indication of confounding of CO risk estimates by BS and N02, but not PMi0. These results are
consistent with CO, BS, and N02 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 N02 were: 0.26% (-0.09 to 0.61) for total mortality and 0.37% (-0.05 to
0.80) for cardiovascular mortality. Thus, the inclusion of N02 in the model nearly halved the CO risk
estimates (whereas the N02 risk estimate was not sensitive to the inclusion of CO in the model). A similar
magnitude of reduction in the CO risk estimates was also observed when including BS in the model.
Overall, the sensitivity of CO risk estimates to the inclusion of N02 in the model is consistent with the
results presented in NMMAPS and the Canadian multi-city studies.
In the second stage model, Samoli et al. (2007) 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
March 2009
5-124
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
cities where the standardized mortality rate was lower. Earlier APHEA studies also reported a regional
pattern of air pollution associations for BS and S02, 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) 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) 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 seven of the 19 European cities examined, the lowest 10% of the
CO distribution was at or above 2 mg/m3 (1.74 ppm). Thus, the interpretation of the suggestive indication
of a threshold is limited.
In summary, the APHEA2 analysis of CO in 19 cities found an association between CO and total
and cardiovascular mortality in single-pollutant models, but the associations were substantially reduced
when N02 or BS was included in copollutant models. The evidence for potential confounding of CO risk
estimates by N02 is consistent with the findings from the NMMAPS and Canadian 12 cities studies. In
addition, Samoli et al. (2007) 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 weak evidence of a CO threshold.
Other European Multi-City Studies
An additional European multi-city study was conducted by Biggeri et al. (2005) 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, because all of the pollutants showed positive
associations with the mortality endpoints examined and the correlations among the pollutants were not
presented, it is unclear if the observed associations are shared or confounded.
Summary of Multi-City 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 APHEA2
study being slightly larger (0.57%) (Figure 5-15). In both the NMMAPS and Canadian studies, a 1-day
lag showed the strongest association; but the APHEA2 study used an a priori exposure window, the
average of 0- and 1-day lags, which has been found to be the exposure window most strongly associated
with mortality in PM analyses.
March 2009
5-125
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
The APHEA2 risk estimates presented in Figure 5-15 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 to 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 N02 in a copollutant model (0.11, 0.03, and 0.26%, for the NMMAPS, Canadian 12-cities
study, and APHEA2, respectively). Thus, these results suggest confounding by N02. However, this
interpretation is further complicated because as with CO, N02 itself may be an indicator of combustion
sources, such as traffic. Because CO measurements tend to reflect more local impacts, due to the location
of monitors, than N02 (which is a secondary pollutant and therefore more spatially uniform) it is also
possible that CO, the less precisely measured pollutant in terms of spatial distribution, may "lose" in the
multipollutant model. Thus, it may not be accurate to interpret these results as evidence of "confounding
by N02."
March 2009
5-126
DRAFT - DO NOT CITE OR QUOTE
-------
Dominici et al. (2003a; 2003b)
(re-analysis by 2000a)
NMMAPS, lag 1
CO only (82 cities)
+ PM10 (82 cities)
+ PM10 and NO2 (57 cities)
Burnett et al. (2004)
12 Canadian cities, lag 1
CO only
-N02=
Samoli et al. (2007)
APHEA2 (19 cities), lag 0-1
CO only
- PM10
-BS
-NO2
e-
-©-
aN02 is the average of 0,1, and 2 day lags.
~i 1 1 1 1 1 1 r~
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
% Increase
Figure 5-15. Summary of mortality risk estimates for short-term exposure to CO from multi-city
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
1 Stieb et al. (2002) reviewed the time-series mortality studies published between 1985 and 2000,
2 and conducted a meta-analysis to estimate combined effects for PMi0, CO, N02, 03, and S02. Because
3 many of the studies reviewed in the 2000 analysis used GAM with default convergence criteria,
4 Stieb et al. (2003) updated the estimates from the meta-analysis by separating the GAM versus non-GAM
5 studies. In this meta-analysis the authors also presented separate combined estimates for single- and
6 multipollutant models. Overall, there were more GAM estimates than non-GAM estimates for all of the
March 2009
5-127
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
pollutants except S02. 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 rather small.
5.6.1.4. Single-City Studies
In addition to the multi-city 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
multi-city studies, often focused on the PM-mortality association, but also provided additional
information that is not available in the multi-city studies. Because the sample size used in each single-city
study is small, and subsequently results in wide confidence bands, a quantitative comparison of the results
from single- and multi-city studies is difficult. In addition, some studies do not present CO results
quantitatively adding to the inability to adequately compare studies. Table 5-17 lists the single-city studies
evaluated along with the mean CO concentrations reported in each study.
Table 5-17. Range of CO concentrations reported in single-city studies that examine mortality
effects associated with short-term exposure to CO.
Study
Location
Years
Averaging
Time
Mean Concentration
(ppm)
Upper Percentile
Concentrations (ppm)
De Leon et al. (2003)
New York, NY
1985-1994
24-h avg
2.45
95th: 4.04
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
(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 PMi0. In single-pollutant models,
PMio, CO, S02, and N02 all showed the same pattern of association with circulatory mortality for
individuals > 75, 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) analyzed 15 air pollutants for their associations with mortality in Atlanta, GA,
for a two-year period starting in August 1998. These pollutants included PM2.5, PM10-2.5, UFP surface area
and counts, aerosol acidity, EC, OC, S042, 03, CO, S02, and N02. 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 all-cause mortality, the
strongest association, which was also statistically significant, was found for PM2 5. CO, S042, and PMi0.2 5
also showed positive associations with all-cause mortality, but they were not significant (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), and these
associations remained when PM2 5 was included in the model (results were presented graphically).
Single-City Studies Conducted in Canada
Vedal et al. (2003) examined the association between short-term exposure to "low levels" of air
pollution (i.e., PMi0, 03, N02, S02, and CO) and daily morality 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 03 in the summer and N02 in the winter showed the strongest associations with
mortality. Vedal et al. (2003) found that CO was positively, but not significantly associated with mortality.
Additionally, the association between short-term exposure to N02 and mortality was found to be
consistent with the results from the Canadian multi-city study conducted by Burnett et al. (2004).
Villeneuve et al. (2003) 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, 03, CO, S02, andN02 were examined for their associations with
all-cause, cardiovascular, and respiratory mortality. When examining the association between gaseous
pollutants and all-cause mortality in this data set, N02 and S02 showed the strongest associations, while
the association between CO and all-cause mortality were generally weaker than those for N02 and S02.
For cardiovascular mortality, S02 risk estimates were smaller than those for N02 or CO, while for
respiratory mortality, S02 showed the strongest associations. However, the wider confidence bands of
March 2009
5-129
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
these categories and the smaller daily counts make it difficult to assess CO associations with cause-
specific mortality outcomes.
Goldberg et al. (2003) contrasted associations between air pollution and mortality in individuals
with underlying CHF vs. mortality in individuals who were identified as having CHF one year 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) examined associations between PM2 5, CoH,
S042", 03, CO, S02, and N02, and mortality. The authors found no association between any of the air
pollutants and mortality with underlying CHF. However, Goldberg et al. (2003) found positive
associations between air pollution and mortality in individuals diagnosed with CHF one year prior to
death. Of the air pollutants examined, CoH, N02, and S02 were most consistently associated with
mortality for ages 65 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) conducted a
multi-city 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) re-analyzed by (Hoek, 2003) examined associations between air pollution and
all-cause, 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 to 1994. The air
pollutants analyzed included BS, PMi0, 03, N02, S02, CO, S042" and N03~. In the single-pollutant
models, all of the pollutants were significantly associated with all-cause 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, S042" and N03" were included in the model while the risk estimates for
these copollutants remained significantly positive. BS, CO, and N02 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). The authors found that 03 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) study is
somewhat in contrast to the result from the Samoli et al. (2007) multi-city study in that, in the Hoek et al.
March 2009
5-130
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(2000) 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) reanalysis by (Hoek, 2003) 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 03, BS, PMi0, CO, S02, and N02 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. While the results
suggested larger impacts of air pollution on more specific cardiovascular causes, the authors did not
provide evidence of an association that was specific to any particular pollutant which would aid in
addressing the question of confounding.
Summary of Single-City Studies
Overall, it is difficult to identify a clear pattern of CO-mortality associations from the single-city
studies evaluated because of the relatively small sample sizes, expected variation in the air pollution mix
across cities, and differences in the analytical approaches used across studies. However, in all of these
studies, CO showed, at least to some extent, positive associations with all-cause or circulatory/
cardiovascular deaths. In addition the one study that examined specific cardiovascular causes of death
(Hoek et al., 200preanalysis by (Hoek, 2003) did not find an association that is specific to CO. Although
the extent of sensitivity of CO-mortality associations varied across studies, it could be attributed to the
likely variation in the correlation between CO and copollutants.
5.6.1.5. Summary of Mortality and Short-Term Exposure to CO
Among the gaseous pollutants examined in time-series mortality studies, CO is the least frequently
studied criteria air pollutant. Because CO was mostly treated as a potential confounder in these studies,
the information available regarding the nature of the association between short-term exposure to CO and
mortality is limited compared to the other pollutants. However, the recently available multi-city studies,
which consist of larger sample sizes, and single-city studies generally confirmed the findings reported in
the 2000 COAQCD.
The multi-city studies which were evaluated, reported comparable CO mortality risk estimates for
total (non-accidental) mortality with the APHEA2 European multi-city study (Samoli et al., 2007)
showing slightly higher estimates for cardiovascular mortality in single-pollutant models. However, when
examining potential confounding by copollutants these studies consistently showed that CO mortality risk
estimates were reduced when N02 was included in the model, but this observation may not be
"confounding" in the usual sense in that N02 may also be an indicator of other pollutants or pollution
sources (i.e., traffic).
March 2009
5-131
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Only one of the multi-city studies focused specifically on the CO-mortality association (Samoli et
al., 2007), the APHEA2 study), 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 APHEA study
performed a sensitivity analysis, which indicated an approximate 50 - 80% difference in CO risk
estimates from a reasonable range of alternative models. In addition, the study examined the CO-mortality
concentration-response relationship through a grid search of varying threshold points, and found only
weak evidence of a CO threshold at 0.5 mg/m3 (0.43 ppm), but this result was complicated by the lowest
10% of the CO distribution for seven of the 19 cities examined being at or above 2 mg/m3 (1.74 ppm).
The examination of a variety of city-specific variables to indentify potential effect modifiers of the
CO-mortality relationship found that geographic region explained most of the heterogeneity in CO
mortality risk estimates with the CO-mortality associations being stronger in western and southern
European cities than eastern cities (Samoli et al., 2007). A similar pattern has been reported for black
smoke (BS) and S02 in previous APHEA studies, but the geographic variability observed does not
provide specific information, which could be used to evaluate the CO-mortality association.
The results from the single-city studies are generally consistent with the multi-city 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 other 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; and it is unclear if CO is acting alone or as an
indicator for other combustion-related pollutants. In addition, the results underscore the limitation of
current analytical methods to disentangle the health effects associated with one pollutant in the complex
air pollution mixture. Overall, the epidemiologic evidence is Suggestive of a Causal relationship
between short-term exposure to relevant CO concentrations and mortality.
5.6.2. Epidemiologic Studies with Long-Term Exposure to CO
The 2000 CO AQCD 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,
but it should be noted that these studies primarily focused on PM, and CO was only considered in these
March 2009
5-132
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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.
5.6.2.1. U.S. Cohort Studies
American Cancer Society Cohort Studies
Pope et al. (1995) 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 5 and S042" were
associated with total (non-accidental), cardiopulmonary, and lung cancer mortality, but not with mortality
for all other causes (i.e., non-accidental minus cardiopulmonary and lung cancer). Gaseous pollutants
were not analyzed in Pope et al. (1995). Jerrett et al. (2003) (using data from (Krewski et al., 2000)
conducted an extensive sensitivity analysis of the Pope et al. (1995). ACS data, augmented with
additional gaseous pollutants data. Due to the smaller number of CO monitors available compared to
S042, the number of metropolitan statistical areas (MSAs) included in the CO analysis were reduced
(from 151 with S042) to 107. The mean annual CO concentrations in these MSAs ranged from 0.19 to
3.95 ppm. CO was weakly negatively correlated with S042" (r = -0.07). Among the gaseous pollutants
examined (CO, N02, 03, S02), only S02 showed positive associations with mortality, and in addition was
the only copollutant that reduced S042" risk estimates. For CO, the relative risk estimates for total (non-
accidental) mortality in single and copollutant models with S042" 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) 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. (1995) study. In
addition to PM2 5, 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 S042" were
both associated with total, 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) (See Figure 5-16) 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
1 These results were presented graphically in Pope et al. (2002) and were estimated for Figure 5-16.
March 2009
5-133
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
evidence from other mortality studies that examined the association between long-term exposure to CO
and mortality does not suggest that CO elicits a protective effect.
Women's Health Initiative Cohort Study
Miller et al. (2007) studied 65,893 postmenopausal women between the ages of 50 and 79 years
without previous CVD in 36 U.S. metropolitan areas from 1994 to 1998. They 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 annual mean levels 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 to 100 meters)." 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 events, including 261 cardiovascular-related deaths. Hazard ratios for the initial
cardiovascular event were estimated. The following results are for models that only included subjects
with non-missing 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-f.ig/nr1 increase in annual average), followed by S02
(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, PM10-2.5, S02, N02, and 03), 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. This study did not examine the correlations among pollutants and, therefore, the extent of
confounding could not be examined, but PM2 5 was clearly the best predictor of cardiovascular events.
The Washington University-EPRI Veterans' Cohort Mortality Studies
Lipfert et al. (2000a) 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
years (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, S042, CO, 03, N02, S02, and Pb. Lipfert et al. (2000a) estimated
exposures by indentifying the county of residence at the time of entry to the study. Four exposure periods
March 2009
5-134
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
(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. (2000a) reported that indications of concurrent mortality
risks (i.e., associations between mortality and air quality for the same period) were found forN02 and
peak 03. The estimated CO mortality risks were all negative, but not significant.
Lipfert et al. (2006b) examined associations between traffic density and mortality in the same
Veterans' Cohort, but 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, N02, 03. PM2 5, S042, non-S042" PM25, and PMi0.2.5). In the multipollutant models, mortality
risk estimates were not statistically significant for any of the other pollutants, except 03. The natural log
of the traffic density variable (VKTA = vehicle-km traveled per year) 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.
(2006b) note 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. (2006b) 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 non-negative effects observed.
Lipfert et al. (2006a) extended the analysis of the Veterans Cohort data to include the EPA's
Speciation Trends Network (STN) data, which collected chemical components of PM2 5. The authors
analyzed the STN data for the year 2002, and again used county-level averages. In addition, they analyzed
PM2 5 and gaseous pollutants data for 1999 through 2001. As in the other Lipfert et al. (2006b) study,
traffic density was the most important predictor of mortality, but associations were also observed for EC,
vanadium (V), nickel (Ni), and N03~. Ozone, N02, and PMi0 also showed positive, but weaker
associations. The authors found no association between the mean annual 95th percentile of hourly CO
March 2009
5-135
DRAFT - DO NOT CITE OR QUOTE
-------
values and mortality (RR = 0.995 [95% CI: 0.988-1.001] per 1 ppm increase in CO concentration) in a
single-pollutant model. The study did not present multipollutant model results for CO.
Endpoint
Reference
Years of
Mortality Year(s)ofAQ
Cohort Data Data
All-cause
Cardio-pulmonary
Cardiovascular
Pope etal. (2002)= ACS 1982-1998 1980
Pope etal. (2002)= ACS 1982-1998 1982-1998
Lipfert et al. (2006b) Veterans 1976-2001 1976-2001
Lipfert et al. (2006b) Veterans 1976-2001 1976-2001
Lipfert et al. (2006a) Veterans 1997-2001 1999-2001
Jerrett et al. (2003) ACS 1982-1998 1982-=
Jerrett et al. (2003) ACS
Pope etal. (2002)= ACS
Pope etal. (2002)= ACS
Miller et al. (2007)b WHI
Miller et al. (2007)b WHI
1982-1998 1982-=
1982-1998 1980
1982-1998 1982-1998
1994-1998 2000
1994-1998 2000
T
- ln(VKTA)d
H S042"
r + PM2 5, PM10-2 5, SO2, NO2, O3
J
T
a The study does not present CO results quantitatively. This effect estimate and
95% confidence interval were estimated from Figure 5 in Pope et al. (2002).
b Effect estimate is only for subjects with non-missing exposure data for all pollutants
c 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.
d Natural log ofvehicle-km traveled per year (VKTA).
0.80 0-B5 0.90 0.95 140 1.05 1.10 1.15 120
Effect Estimates
Figure 5-16. 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.
March 2009
5-136
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 (Kuenzli and Tager, 1997). In this sense, cross-sectional studies may be useful in evaluating the
correlation among exposure variables.
Lipfert and Morris (2002) 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+) all-cause (excluding AIDS and trauma)
mortality on air pollution, demography, climate, SES, lifestyle, and diet. The authors analyzed TSP, PMi0,
PM2 5, S042, S02, CO, N02, and 03. 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 PM25 between
1979-1984 and 1999. In response to the varying number of counties with valid air pollution data by
pollutant and time, Lipfert and Morris (2002) 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. Overall, there were too many results to summarize
because of the large number of age groups, lagged exposure time windows, and mortality study periods
examined in the study, but similar to the ACS cohort studies, PM2 5 and S042" were found to be
consistently and positively associated with mortality. Lipfert and Morris (2002) 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). Also, consistent with the
Lipfert et al. (2000a) and the Pope et al. (2002) 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) often showed a positive
association with CO, but considering the small number of deaths attributed to this age group (less than
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.
March 2009
5-137
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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 (2000) 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. The re-analysis
of the ACS data (Pope et al., 1995) by Jerrett et al. (2003) found no association between long-term
exposure to CO and mortality. Similar results were obtained in an updated analysis conducted by Pope
et al. (2002) when using earlier (1980) CO data, but negative associations were found when using more
recent (1982-1998) data. The Women's Health Initiative (WHI) Study also found no association between
CO and CVD events (including mortality) using the data from recent years (1994-1998) (Miller et al.,
2007), 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., 2006a; Lipfert et al.,
2006b). One additional study was identified that used a cross-sectional study design, Lipfert and Morris
(2002), 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 S042" 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
subpopulations (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 is suggestive of no causal relationship between long-term
exposure to CO and mortality.
5.7. Public Health Impacts
This section addresses several issues relating to the broader public health impact from exposure to
ambient CO through a discussion on: (1) the shape of the concentration-response (C-R) relationship for
CO, which is based primarily on controlled human exposure studies; and (2) the identification of
subpopulations which may experience increased risks from CO exposures, through either enhanced
susceptibility (e.g., as a result of pre-existing disease, genetic factors, age) and/or vulnerability associated
with external factors or differential exposure (e.g., altitude, close proximity to sources, activity patterns).
March 2009
5-138
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
5.7.1. Concentration-Response Relationship
Currently, very limited information is available in the human clinical and epidemiologic literature
regarding the CO C-R relationship and the potential presence of a CO threshold. Two human clinical
studies described in the 2000 CO AQCD had evaluated the C-R relationship between CO and onset of
exercise-induced angina among individuals with CAD, but at the high end of CO concentrations (i.e., CO
levels above the current NAAQS). Anderson et al. (1973) exposed 10 adult men with stable angina for 4 h
to CO concentrations of 50 and 100 ppm, which resulted in average COHb levels 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 two concentrations of CO (average exposure concentrations of 117 and 253 ppm) resulting in average
pre-exercise COHb levels of 2.4% and 4.7% (Allred et al., 1989). Relative to control (average COHb
0.7%), COHb levels of 2.4% and 4.7% were observed to decrease the time to onset of angina by 4.2%
(p = 0.054) and 7.1 % (p = 0.004), respectively. In addition, these investigators reported a statistically
significant decrease in the time to exercise-induced ST-segment depression with increasing COHb levels.
These findings provide some evidence of a significant C-R relationship at COHb concentrations between
2.4 and 4.7%. However, the human clinical literature has yet to evaluate the C-R relationship at lower CO
concentrations or COHb levels.
One study 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) 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)
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 seven 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.
March 2009
5-139
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
5.7.2. Potentially Susceptible and Vulnerable Populations
Interindividual variation in human responses to air pollutants indicates that not all individuals
exposed to pollutants respond similarly. That is, some subpopulations are at increased risk to the
detrimental effects of pollutant exposure. The NAAQS are intended to provide an adequate margin of
safety for both general populations and susceptible and vulnerable subpopulations, or those subgroups
potentially at increased risk for ambient air pollution health effects. For the purposes of this document, a
susceptible subpopulation is defined as one with an intrinsic biological characteristic (e.g., disease) that
might result in an adverse health effect at pollutant concentrations below those needed to elicit the same
response in the general population, or result in a more adverse effect at the same concentration. A
vulnerable subpopulation is one with a non-biological characteristic (e.g., differential exposure, living at
altitude) that results in increased incidence of health effects from an ambient pollutant relative to the
general population. The previous review of the CO NAAQS identified certain groups within the
population that may be more susceptible to the effects of CO exposure, including individuals (particularly
older adults) with CAD and other vascular diseases, anemia patients, patients with obstructive lung
disease, and infants. Other subgroups considered to be somewhat vulnerable in the previous review
include individuals living at altitude and those using certain medications that can increase endogenous CO
production. Tables 5-18 and 5-19 provide an overview of the characteristics that contribute to susceptible
and vulnerable subpopulations, respectively, which have been observed in the examination of the NAAQS
for all criteria pollutants. Those characteristics of susceptible/vulnerable subpopulations exposed
specifically to CO, as mentioned in the literature that encompasses this ISA, are discussed below.
Table 5-18. Characteristics of susceptible subpopulations.
Susceptibility Characteristics1
Pre-existing disease: Cardiovascular diseases, Anemia, Obesity, Diabetes, Respiratory diseases (e.g., asthma, obstructive lung disease)
Age: Children, Older Adults (65+)
Infants: Premature, Low Birth Weight
Gender
Pregnancy
Birth Defects
Race/Ethnicity
Genetic Factors
Nutritional status
1 Susceptible (i.e., intrinsic) refers to biological characteristics of an individual, which can include life stage, genetics, and pre-existing disease.
March 2009
5-140
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 5-19. Characteristics of vulnerable subpopulations.
Vulnerability Characteristics1
Altitude
Level of Exercise
Proximity to Roadways
Medication Use
Geographic Location
Work Environment (e.g., outdoor workers)
Air Conditioning Use
SES
Education Level
Nutritional status
1 Vulnerable (i.e., extrinsic) refers to non-biological variables associated with an individual that can result in a health effect.
5.7.2.1. Susceptibility Characteristics
Cardiovascular Disease
Individuals with heart disease may be at a greater risk from CO exposure since they may already
have compromised 02 delivery. CO is notable among air pollutants because it is especially harmful in
individuals with impaired cardiovascular systems. Persons with a normal cardiovascular system can
tolerate substantial concentrations of CO, if they vasodilate in response to the hypoxemia 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 IHD.
An estimated 81 million American adults (1 in 3) have one or more type of cardiovascular disease
(CVD), with an estimated 47% of these being 60 or more years of age. CVD is the leading cause of death
in the U.S. with nearly 2,400 deaths each day-an average of one death every 37 seconds (Rosamond et al.,
2008). For the major diseases within the category of total CVD, about 73 million Americans have high
BP, 16 million have CHD, 5 million have heart failure, 5 million have stroke, and the estimated
prevalence of congenital cardiovascular defects is estimated to be between 650,000 to 1.3 million
(Rosamond et al., 2008). In the U.S., IHD is the largest major killer, causing 1 in 5 deaths. Because the
numbers of affected people are so high, even relatively small percent increases in cardiovascular mortality
or morbidity in the population could have a large impact on public health.
Each year in the U.S. approximately 780,000 people experience a new or recurrent stroke, with the
majority of these being a first stroke (77%). On average this equates to someone in the U.S. having a
stroke every 40 seconds with a death occurring every 3 to 4 minutes. When considered separately from
March 2009
5-141
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
other CVDs, stroke ranks third among all causes of death, behind diseases of the heart and cancer
(Rosamond et al., 2008). While epidemiologic evidence is weak regarding associations between ambient
CO and stroke, diseased individuals could exhibit increased sensitivity to CO exposure.
Obstructive Lung Disease
Patients with obstructive lung disease, such as COPD, were identified as a susceptible
subpopulation in the 2000 CO AQCD. COPD is a progressive disease resulting in decreased air flow to
the lungs, and is especially prevalent among smokers. A majority of COPD patients have exercise
limitations as demonstrated by a decrease in 02 saturation during mild to moderate exercise. This makes
individuals with hypoxia resulting from COPD particularly sensitive to CO during submaximal exercise
typical of normal daily activity. COPD patients who are smokers may have baseline COHb levels of
4-8%, leaving little reserve for increases in COHb due to ambient exposure. COPD is often accompanied
by a number of changes in gas exchange, including increased VD and V/Q inequality (Marthan et al.,
1985), which could slow both CO uptake and elimination. Patients with pulmonary sarcoidosis may have
a decrease in lung volumes, a loss of DLCO, and gas exchange abnormalities during exercise, including
decreased Pa02 and increased alveolar-arterial oxygen pressure difference (Lamberto et al., 2004). A
controlled human exposure study on 19 individuals with COPD (Bathoorn et al., 2007) 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. The few epidemiologic studies that
evaluated the relationship between ambient CO and increased hospital admissions or ED visits for COPD
show weak positive associations. Epidemiologic results were similar for asthmatics, who can also
experience exercise-induced airflow limitation.
Anemia
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 02 carrying capacity or
content will result in a greater risk from COHb induced hypoxia and decreased tissue 02 delivery. The
severity of this effect depends upon the initial level of hypoxia. Anemias are a group of diseases that
result in insufficient blood 02 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., 1976a, b). 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 02 and 02 carrying capacity is reduced causing a shift to the right in the 02 dissociation curve. It is
March 2009
5-142
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
well documented that African-American populations have a higher incidence of sickle cell anemia, which
may be a risk factor for CO hypoxia.
Any disturbance in RBC hemostasis by acceleration of destruction of hemoproteins will lead to
increased production of CO. 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; Hampson, 2007; Meyer et al., 1998; Solanki et al., 1988). Patients with hemolytic
anemia exhibit COHb levels 2- to 3-times higher than healthy individuals and CO production rates 2- to
8-times higher (Coburn et al., 1966).
Age and Gender
Age and gender alter the variables that influence the uptake, distribution, and elimination of CO.
COHb levels decline more rapidly in young children than adults after CO exposure (Joumard et al., 1981;
Klasner et al., 1998). After infancy, the COHb half-life increases with age, practically doubling between 2
and 70 years (Joumard et al., 1981). The rate of this decrease in CO elimination is very rapid in the
growing years (2 to 16 years of age), but slows beyond adolescence. Alveolar volume and DLCO
increased with increasing body length of infants and toddlers (Castillo et al., 2006), 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).
COHb concentrations are generally lower in female subjects than in male subjects (Horvath et al.,
1988) and 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). 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). Women also
tended to be more resistant to altitude hypoxia (Horvath et al., 1988).
Newborns and Young Infants
Fetal CO pharmacokinetics do not follow the same kinetics as maternal CO exposure, making it
difficult to estimate fetal COHb based on maternal levels. Human fetal Hb has a higher affinity for CO
than adult Hb (Di Cera et al., 1989). Maternal and fetal COHb concentrations have been modeled as a
function of time using a modified CFK equation (Hill et al., 1977). At steady-state conditions, the fetal
COHb is up to 10% 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 5.5%. The fetal CO uptake lags behind the maternal for
the first few hours but later may overtake the maternal values. Similarly, during washout, the fetal COHb
levels are maintained for longer, with a half-life of around 7.5 hours versus the maternal half-life of
around 4 hours. In addition, women experience fluctuating COHb levels during pregnancy as well as
March 2009
5-143
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
through the menstrual cycle when endogenous CO production doubles in the progesterone phase
(Delivoria-Papadopoulos et al., 1974; Mercke and Lundh, 1976).
Epidemiologic studies of birth outcomes have examined well-established clinical metrics of infant
health. Preterm birth (PTB, <37 weeks gestation) and low birth weight (LBW, birth weight <2,500 g)
have been established as strong predictors of infant mortality and morbidity (Barker et al., 2002;
Berkowitz and Papiernik, 1993; Li et al., 2003; Mclntire et al., 1999). In 2004, 36.5 percent of all infant
deaths in the U.S. were preterm-related (MacDorman et al., 2007). 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 LBW was
8.2%, which has risen 17% since 1990 (Martin et al.).
Congenital anomalies remain the leading cause of infant death in the U.S. (Martin et al.). In 2004,
for every 1,000 live births in the U.S. the infant (up to 1 year of age) mortality rate was 6.8 deaths, while
the neonatal (under 28 days of age) mortality rate was 4.5 deaths, and the postneonatal (between 28 days
and 1 year of age) mortality rate was 2.4 deaths (National Center for Health Statistics, 2007). From 1968
to 1995, the proportion of infant mortality attributable to birth defects increased from 14.5% to 22.2 %
(Centers for Disease Control and Prevention, 1998). Limited epidemiologic research has been conducted
on congenital anomalies because of their rare occurrence, which makes it difficult to detect small effects
within defined exposure periods. For example, in 2005 the rates for spina bifida and anencephaly in the
U.S. were 18.0 and 11.3 per 100,000 births respectively (Martin et al., 2007) and as a result of folic acid
fortification the prevalence of these defects declined considerably from 1995 to 2002 (Williams et al.,
2005). Survival rate among infants with spina bifida had also improved due to folic acid fortification (Bol
et al., 2006). The rate of cleft lip/palate is higher at approximately 79.1 per 100,000 births (Martin et al.,
2007).
The rate of cardiovascular defects is much higher. Data from the Metropolitan Atlanta Congenital
Defects Program (MACDP), which is one of the most comprehensive birth defect registries in the U.S.,
showed that the prevalence of congenital heart defects had increased between 1968 and 1997. During
1995-1997 the rate was 90.2 per 10,000 births (0.9%) and this had increased from 58.7 per 10,000 births
since 1986-1972 (Botto et al., 2001). Cardiovascular defects are the single largest contributor to infant
mortality attributable to birth defects (Centers for Disease Control and Prevention, 1998). Between 1979
and 1997, 1 in 10 infant deaths (9.8%) was associated with a congenital heart defect, and 1 in 13 infant
deaths (7.4%) was due to a congenital heart defect (Boneva et al., 2001).
5.7.2.2. Vulnerability Characteristics
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
March 2009
5-144
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
and other documents (U.S. EPA, 1978). In an effort to maintain proper 02 transport and supply,
physiological changes occur as compensatory mechanisms to combat the decreased barometric pressure
and resulting altitude induced HH (HH). HH, unlike CO hypoxia, causes humans to hyperventilate, which
reduces arterial blood C02 (hypocapnia) and increases alveolar partial pressure of 02. Hypocapnia will
lead to difficulty of 02 dissociation and decreased blood flow, thus reducing tissue 02 supply. HH
increases BP and cardiac output and leads to redistribution of blood from skin to organs and from blood 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 02 bind reduced Hb through a
competitive-parallel reaction (Chakraborty et al., 2004). Breathing CO (9 ppm) at rest at altitude produced
higher COHb compared to sea level (McGrath et al., 1993), whereas high altitude exposure with exercise
caused a decrease in COHb levels versus similar exposure at sea level (Horvath et al., 1988). 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. Early 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).
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, differences in people raised at high altitude persist even after
reacclimatization to sea level (Hsia, 2002).
Altitude has been shown to be positively associated with baseline COHb concentrations (McGrath,
1992; McGrath et al., 1993). This increase in COHb with altitude induced hypoxia has also been
associated with increases in the mRNA, protein, and activity of HO-1 in rats and cells leading to
enhanced endogenous CO production (Carraway et al., 2002; Chin et al., 2007). 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.
Activity Patterns
Exercise is an important determinant of CO kinetics and toxicity due to the extensive increase in
gas exchange. 02 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.
The COHb elimination rate decreases with physical activity (Joumard et al., 1981). Healthy
subjects exposed to CO and achieving COHb levels of approximately 4-5% observed a significant
detriment to exercise duration and maximal effort capability (measured by metabolic equivalent units)
March 2009
5-145
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
(Adir et al., 1999). It is possible that CO lowers the anaerobic threshold, allowing earlier fatigue of the
skeletal muscles and decreased maximal effort capability.
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 experience elevated CO
concentrations, as discussed in Chapter 3, and therefore constitute a potentially vulnerable subpopulation
due to differential exposure. Studies of commuters have shown that commuting time is an important
predictor of CO exposure for those traveling by car, cycling, and walking, and that on-road CO
concentrations are typically two to five times higher than concentrations measured at the roadside.
Medication Use
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). All of these sources result in a blood COHb concentration between 0.4 and 0.7%
(Coburn et al., 1965). This baseline level of endogenous production can be altered by drugs or a number
of physiological conditions that alter RBC destruction, other hemoprotein breakdown, or bilirubin
production. Nicotinic acid (Lundh et al., 1975), allyl-containing compounds (acetamids and barbiturates)
(Mercke et al., 1975b), diphenylhydantoin (Coburn, 1970b), progesterone (Delivoria-Papadopoulos et al.,
1974), and contraceptives (Mercke et al., 1975a) will increase CO production. Compounds such as carbon
disulfide and sulfur-containing chemicals (parathion and phenyltiourea) will increase CO by acting on
P450 system moieties (Landaw et al., 1970). The P450 system may also cause large increases in CO
produced from the metabolic degradation of dihalomethanes leading to very high (>10%) COHb levels
(Manno et al., 1992), which can be further enhanced by prior exposure to HCs or ethanol (Pankow et al.,
1991; Wirkner and Poelchen, 1996). HO can catalyze the release of CO from the auto-oxidation of
phenols, photo-oxidation of organic compounds, and lipid peroxidation of cell membrane lipids (Rodgers
et al., 1994).
March 2009
5-146
DRAFT - DO NOT CITE OR QUOTE
-------
Annex A. Atmospheric Science
etfrfi
Yukon
Koyukuk
Carbon Mono^de Emissions in 2002 (Tons per Square Mile) i ' 0.06 — 0.33
I 10.35 - 101
I—1 131 - 55.53
Alaska State Emissions 2002
On-Road Vehicles
] 145,340
Fires
18.403.218
Non-Road Equipment
84,990
Residential Wood Combustion
12,746
Fossil Fuel Combustion
27,954
Electricity Generation
6,060
Industrial Processes
4,256
Waste Disposal
1,698
Miscellaneous
40
Solvent Use
0
0 12,00
Emissions (Tons)
Yukon-Koyukuk County Emissions 2002
On-Road Vehicles
2,494
Fires
I 8.095.363
Non-Road Equipment
3,192
Residential Wood Combustion
398
Fossil Fuel Combustion
8
Electricity Generation
0
Industrial Processes
0
Waste Disposal
52
Miscellaneous
0
Solvent Use
0
Emissions (Tons)
Figure A-1. CO emissions density map and distribution for the state of Alaska and for Yukon-
Koyukuk County in Alaska.
March 2009
A-1
DRAFT-DO NOT QUOTE OR CITE
-------
Weber
Utah
Grand
Carbon Monoxide Emissions in 2002 (Tons per Squere Mile)
Utah State Emissions 2002
On-Road Vehicles
1768.164
Fires
|257,199
Non-Road Equipment
Residential Wood Combustion
| 203,475
|12,031
Fossil Fuel Combustion
] 32,365
Electricity Generation
4,234
Industrial Processes
~ 42,466
Waste Disposal
152
Miscellaneous
277
Solvent Use
17
1 159 - 5.90
=1 6,97 - 16.34
=l S.56 - 395.50
Weber County Emissions 2002
On-Road Vehicles
146.841
Fires
916
Non-Road Equipment
112.491
Residential Wood Combustion
1,046
Fossil Fuel Combustion
497
Electricity Generation
56
Industrial Processes
12,079
Waste Disposal
0
Miscellaneous
24
Solvent Use
3
Emissions (Tons)
Utah County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
189.535
| 2,439
120.195
Residential Wood Combustion
12,048
Fossil Fuel Combustion
11,919
Electricity Generation
1
Industrial Processes
1,346
Waste Disposal
1
Miscellaneous
44
Solvent Use
0
Emissions (Tons^
Grand County Emissions 2002
On-Road Vehicles
Fires
113,242
1123.963
Non-Road Equipment
11,771
Residential Wood Combustion
44
Fossil Fuel Combustion
298
Electricity Generation
0
Industrial Processes
32
Waste Disposal
1
Miscellaneous
2
Solvent Use
0
Emissions (Tons^
Emissions (Tons^
Figure A-2. CO emissions density map and distribution for the state of Utah and for selected
counties in Utah.
March 2009
A-2
DRAFT-DO NOT QUOTE OR CITE
-------
Middlesex
Suffolk
~
Woreeste
Norfolk
Carbon Monoxide Emisacos in 2002 (Tens per Square Mile)
49,37 - 117.27
13875 - 198.59
252.78 - 225154
Massachusetts State Emissions 2002
Middlesex County Emissions 2002
I 926.079
I 214.556
Fires
0
Fires
0
Non-Road Equipment
1490.378
Non-Road Equipment
|101.247
Residential Wood Combustion
1104,462
Residential Wood Combustion
492
Fossil Fuel Combustion
] 28,111
Fossil Fuel Combustion
] 7,063
Electricity Generation
11,569
Electricity Generation
492
Industrial Processes
1,948
Industrial Processes
287
Waste Disposal
3,877
Waste Disposal
323
Miscellaneous
482
Miscellaneous
77
Solvent Use
77
Solvent Use
7
1,20C
000
260
Emissions (Tons)
Emissions (Tons)
Norfolk County Emissions 2002
Suffolk County Emissions 2002
1154.098
164.900
Fires
0
Fires
0
Non-Road Equipment
144.873
Non-Road Equipment
I 63.447
Residential Wood Combustion
J 8,803
Residential Wood Combustion
2,326
Fossil Fuel Combustion
2,728
Fossil Fuel Combustion
1,689
Electricity Generation
189
Electricity Generation
225
Industrial Processes
208
Industrial Processes
143
Waste Disposal
285
Waste Disposal
1
Miscellaneous
24
Miscellaneous
110
Solvent Use
8
Solvent Use
0
260
000 0 260
Emissions (Tons)
Emissions (Tons)
Figure A-3. CO emissions density map and distribution for the state of Massachusetts and for
selected counties in Massachusetts.
March 2009
A-3 DRAFT-DO NOT QUOTE OR CITE
-------
Chatham
Fulton
On-Road Vehicles
Fires
Non-Road Equipment
I 223.633
0
180,197
Residential Wood Combustion
12,262
Fossil Fuel Combustion
12,850
Electricity Generation
0
Industrial Processes
275
Waste Disposal
279
Miscellaneous
92
Solvent Use
0
Emissions (Tons)
On-Road Vehicles
1142,648
Fires
0
Non-Road Equipment
Residential Wood Combustion
168,502
12,010
Fossil Fuel Combustion
11,907
Electricity Generation
0
Industrial Processes
104
Waste Disposal
254
Miscellaneous
5
Solvent Use
7
Emissions (Tons)
On-Road Vehicles
I 2.604.609
Fires
] 53,412
Non-Road Equipment
Residential Wood Combustion
1720,812
147,461
Fossil Fuel Combustion
| 54,857
Electricity Generation
9,824
Industrial Processes
149,802
Waste Disposal
~ 175,000
Miscellaneous
500
Solvent Use
16
0 3,500,000
Emissions (Tons)
Fulton County Emissions 2002 Dekalb County Emissions 2002
Dekalb
Carbon Monoxide Emissions in 2002 (Tons per Square Mile) i ' 4.62 — 22.32
I 1 22.45 - 49.92
1=^ 5111 - 803,67
Georgia State Emissions 2002
Liberty
Figure A-4a. CO emissions density map and distribution for the state of Georgia and for selected
counties in Georgia (1 of 2).
March 2009
A-4
DRAFT-DO NOT QUOTE OR CITE
-------
Chatham County Emissions 2002
Liberty County Emissions 2002
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
67,821
125,904
1,849
On-Road Vehicles
Fires
Non-Road Equipment
Residential Wood Combustion
Fossil Fuel Combustion
Electricity Generation
Industrial Processes
Waste Disposal
Miscellaneous
Solvent Use
19,711
4,832
55.427
Emissions (Tons)
Glynn County Emissions 2002
On-Road Vehicles
127,024
Fires
0
Non-Road Equipment
~ 15,467
lesidential Wood Combustion
134
Fossil Fuel Combustion
3,056
Electricity Generation
377
Industrial Processes
] 6,619
Waste Disposal
1,129
Miscellaneous
4
Solvent Use
0
Emissions (Tons)
Emissions (Ton$
Figure A-4b. CO emissions distribution for selected counties in Georgia (2 of 2).
March 2009
A-5
DRAFT-DO NOT QUOTE OR CITE
-------
Orange
San
Diego
San Dfego~
Cafbon M¦encode Emissions in 2002 (Tons per Square Mile) ' ' 0.35 —
1720
I 1 17.29 -
34.04
[=1 34 72 - "848.52
California State Emissions 2002
Los Angeles County Emissions 2002
On-Road Vehicles
On-Road Vehicles
1 3.436.059
I 824.676
Fires
Fires
Zl 62,576
/-A-A WA
Non-Road Equipment
Residential Wood Combustion
Non-Road Equipment
Residential Wood Combustion
11.191.454
[226.349
Z3 287,944
7,220
Fossil Fuel Combustion
] 98,990
Fossil Fuel Combustion
10,795
Electricity Generation
28,471
Electricity Generation
9,814
Industrial Processes
34,908
Industrial Processes
6,960
Waste Disposal
22,308
Waste Disposal
143
Miscellaneous
3,548
Miscellaneous
674
Solvent Use
379
Solvent Use
27
Emissions (Tons)
San Diego County Emissions 2002
Emissions (Tons)
Orange County Emissions 2002
On-Road Vehicles
I 283.597
On-Road Vehicles
Fires
Zl 62,192
Fires
978
Non-Road Equipment
1101.515
Non-Road Equipment
1105,638
Residential Wood Combustion
] 20,821
Residential Wood Combustion
1,660
Fossil Fuel Combustion
114,855
Fossil Fuel Combustion
2,871
Electricity Generation
895
Electricity Generation
655
Industrial Processes
231
Industrial Processes
134
Waste Disposal
162
Waste Disposal
107
Miscellaneous
226
Miscellaneous
184
Solvent Use
0
Solvent Use
8
Emissions (Tons)
Emissions (Tons)
Figure A-5. CO emissions density map and distribution for the state of California and for
selected counties in California.
March 2009
A-6
DRAFT-DO NOT QUOTE OR CITE
-------
n
Jefferson
et
Carbon M<>n«
-------
Anchorage Core Based Statisical Area
200£ Papulation Density
[ 1 An^CT»9» CO Uorttcr> (5 ton feiNO
Popol*lK>n pet 2.8 Sq Km
Figure A-7. Map of CO monitor locations with respect to population density in the Anchorage
CBSA, total population.
Anchorage Core Based Statisical Area
j** *
' i
Milk
o so m tag no an
2004 Population OcnsMy
I I *i*fw»)»COM»ilw|Stoiev
Population £ 85 per 2.6 Sq Kin
| 10 19
»-»
Figure A-8. Map of CO monitor locations with respect to population density in the Anchorage
CBSA, ages 65 and older.
March 2009
A-8
DRAFT-DO NOT QUOTE OR CITE
-------
Atlanta Combined Statisical Area
Figure A-9. Map of CO monitor locations with respect to population density in the Atlanta CSA,
total population.
»0S Population Densrty
| Maria CO Usn«H» 15 km I
Population per 2.6 Sq Km
^¦a-230
¦¦ 231
B 460-KS5
2298-4590
4591 -11475
¦¦ 11478 - 4S8CO
Atlanta Combined Statisical Area
EEIH=BB Kiemttert
0 15 30 63 90 120
Figure A-10. Map of CO monitor locations with respect to population density in the Atlanta CSA,
ages 65 and older.
March 2009
A-9
DRAFT-DO NOT QUOTE OR CITE
-------
Boston Combined Stalisical Area
2Q0£ Population Density
[ | frman CO i.5 ten U^tyj
Population per 2.8 Sq Km
0 1530 60 90 120
Figure A-11. Map of CO monitor locations with respect to population density in the Boston CSA,
total population.
Boston Combined Statisical Area
Kilometers
0 1S30 eo 90 120
Figure A-12. Map of CO monitor locations with respect to population density in the Boston CSA,
ages 65 and older.
March 2009
A-10
DRAFT-DO NOT QUOTE OR CITE
-------
Denver Combined Statisical Area
p>
r
''J
9r
1MI
0 5 10 20 30 40
2005 Population Density
I 1 CO Uon
-------
Houston Combined Statisical Area
K*>m«1ers
0 a 50 1K> >50 200
Figure A-15. Map of CO monitor locations with respect to population density in the Houston CSA,
total population.
Houston Combined Statisical Area
0 5 10 3D 30 40
0 25 50 100 ISO 200
2004 Population Ocnsrty
I I xwtion CO "*<««» (S km tur*>>
Population 2 85 per 2.6 5q Km
o-w
19-37
p— a - ifrt
1S5-3S7
¦I 368-91B
j^B »<» • wi
Figure A-16. Map of CO monitor locations with respect to population density in the Houston CSA,
ages 65 and older.
March 2009
A-12
DRAFT-DO NOT QUOTE OR CITE
-------
Los Angeles Combined Statisical Area
0 5 10 20 30 ¦»
2004 Population Density
Kitomalws
0 30 60 120 180 240
Figure A-17. Map of CO monitor locations with respect to population density in the Los Angeles
CSA, total population.
Los Angeles Combined Statisical Area
¦ ¦ Kibmcico
0 5 10 20 30 40
2004 Population Density
I I CO Uanter* iS*m u/Vi
Population 2 85 per 2.6 Sq Km
0 -
¦"»<»
¦ '¦ '"3 p 199-M0
& b 1H1 49,0
CEHH=aHl«onHm
0 30 60 120 180 240
Figure A-18. Map of CO monitor locations with respect to population density in the Los Angeles
CSA, ages 65 and older.
March 2009
A-13
DRAFT-DO NOT QUOTE OR CITE
-------
New York Combined Statisical Area
EX^EZZBHlCOIMm
0 5 10 20 33 40
2404 Population Density
- *30tOQ
0 JO 40 SO 120 160
Figure A-19. Map of CO monitor locations with respect to population density in the New York City
CSA, total population.
New York Combined Statisical Area
¦ ¦ Ki'orr ctf i
0 20 40 80 t20 160
Figure A-20. Map of CO monitor locations with respect to population density in the New York City
CSA, ages 65 and older.
March 2009
A-14
DRAFT-DO NOT QUOTE OR CITE
-------
Seattle Combined Statisical Area
K4om»]era
0 1530 CO 90 120
Figure A-21. Map of CO monitor locations with respect to population density in the Seattle CSA,
total population.
Seattle Combined Statisical Area
0 1530 60 90 120
Figure A-22. Map of CO monitor locations with respect to population density in the Seattle CSA,
ages 65 and older.
March 2009
A-15
DRAFT-DO NOT QUOTE OR CITE
-------
St. Louis Combined Statisical Area
iKjIonrters
2004 Population Density
I Si low* CO MtrttSf* |S «ro Mttt)
Population per 1.6 Sq Km
-141
Kitometer*
0 15 30 60 90 >20
Figure A-23. Map of CO monitor locations with respect to population density in the St. Louis
CSA, total population.
St. Louis Combined Statisical Area
Kilometers
0 15 30 60 SO 120
Figure A-24. Map of CO monitor locations with respect to population density in the St. Louis
CSA, ages 65 and older.
March 2009
A-16
DRAFT-DO NOT QUOTE OR CITE
-------
Anchorage Core Based Statistical Area
• Anchorage CO Monitors
— Anchorage Major Highways
| Anchorage
0 15 30 60 90 120
¦ ¦ Kilometers
Figure A-25. Map of CO monitor locations with AQS Site IDs for Anchorage, AK.
Table A-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 repotting to
AQS in Anchorage, AK.
Anchorage
A B
1.00
0.73
0.0
1.1
0.00
0.32
0.0
9.0
Legend
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-17
DRAFT-DO NOT QUOTE OR CITE
-------
A
B
Mean
1.04
1.10
Obs
12969
12703
SD
0.94
1.04
4 -
3-
Q.
Cl
c
o
o-
I I I I I I I I I
1 2 3 4 1 2 3 4
season
Figure A-26. 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.
March 2009
A-18
DRAFT-DO NOT QUOTE OR CITE
-------
Atlanta Combined Statistical Area
M
Atlanta Major Highways
| Atlanta
Kilometers
Figure A-27. Map of CO monitor locations with AQS Site IDs for Atlanta, GA.
Table A-2. 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 repotting to
AQS in Atlanta, GA.
Atlanta
A
B C
A
1.00
0.60 0.12
0.0
O
cn
o
--j
0.00
0.27 0.37
0.0
22.5 74.7
B
1.00 0.10
0.0 0.7
0.00 0.38
Legend
0.0 61.7
C
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009 A-19 DRAFT-DO NOT QUOTE OR CITE
-------
Mean
Obs
SD
1.7-
1.6-
1.5-
1.4-
1.3-
.1.2-
1.1 -
1,0-
0.9-
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
0,0-
A
0.53
25531
0.35
0.58
25440
0.30
C
0.30
25712
0.13
I [ I I I I I I II I I I I
1234 1234 1234
season
Figure A-28. 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.
March 2009
A-20
DRAFT-DO NOT QUOTE OR CITE
-------
Boston Combined Statistical Area
Boston CO Monitors
Boston Major Highways
Kilometers
Figure A-29. Map of CO monitor locations with AQS Site IDs for Boston, MA.
March 2009
A-21
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-3. 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 repotting to
AQS in Boston, MA.
Boston
A
B
c
D
E
F
G
>
O
O
0.35
0.50
0.49
0.41
0.50
0.40
0.0
0.4
0.4
0.5
0.4
0.6
0.4
0.00
0.58
0.48
0.42
0.41
0.44
0.40
0.0
37.2
39.7
57.9
41.3
18.3
89.1
B
1.00
0.52
0.34
0.27
0.35
0.34
0.0
0.4
0.6
0.5
0.7
0.4
0.00
0.56
0.59
0.58
0.60
0.55
0.0
2.5
58.0
78.2
55.1
60.2
C
1.00
0.37
0.26
0.38
0.36
0.0
0.5
0.5
0.6
0.4
0.00
0.45
0.45
0.46
0.47
0.0
58.9
80.7
57.5
58.7
D
1.00
0.40
0.46
0.34
0.0
0.4
0.5
0.5
0.00
0.28
0.25
0.39
0.0
85.8
61.5
58.9
E
1.00
0.49
0.29
0.0
0.5
0.4
0.00
0.30
0.37
0.0
26.1
128.6
F 1.00 0.43
0.0 0.6
0.00 0.39
Legend
0.0 102.6
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-22
DRAFT-DO NOT QUOTE OR CITE
-------
E
Q.
Q.
C
o
Mean
Obs
SD
1.9-
1.8-
1.7 -
1.6-
1.5-
1.4 -
1.3-
1.2-
1.1 -
1.0-
0.9-
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
o.o-
A
B
C
D
E
F
G
0.33
0.26
0.36
0.53
0.45
0.60
0.34
24362
24134
24260
24446
25197
25869
23707
0.22
0.24
0.26
0.23
0.27
0.37
0.22
I I I I I I I I I
1234 1234
I I I I I I I I
1234 1234
season
I I I I I I I I I
1234 1234
Figure A-30. 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.
March 2009
A-23
DRAFT-DO NOT QUOTE OR CITE
-------
Denver Combined Statistical Area
Denver CO Monitors
Denver Major Highways
Kilometers
Figure A-31. Map of CO monitor locations with AQS Site IDs for Denver, CO.
March 2009
A-24
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-4. 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 repotting to
AQS in Denver, CO.
Denver
A
B
C
D
E
A
1.00
0.53
0.59
0.64
0.52
0.0
0.6
0.6
0.5
0.6
0.00
0.43
0.29
0.37
0.34
0.0
38.5
10.1
10.9
68.2
B
1.00
0.46
0.49
0.54
0.0
0.7
0.7
0.6
0.00
0.44
0.47
0.43
0.0
46.9
47.0
44.6
C
1.00
0.76
0.45
0.0
0.5
0.7
0.00
0.34
0.36
0.0
1.3
78.3
D
1.00
0.46
0.0
0.7
0.00
0.42
Legend
0.0
79.0
E
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-25
DRAFT-DO NOT QUOTE OR CITE
-------
Mean
Obs
SD
3-
E
a.
CL
c
o
'•*—>
CO
c
s
c
o
o
2 -
A
0.52
25920
0.36
0.42
25559
0.38
C
0.65
25959
0.42
D
0.52
25552
0.46
E
0.55
26048
0.46
III
I I I II I I I I II I I I II I I I
1234 1234 1234 1234
season
rm
1234
Figure A-32. Box plots illustrating the seasonal distribution of hourly CO concentrations in
Denver, CO. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
March 2009
A-26
DRAFT-DO NOT QUOTE OR CITE
-------
Houston Combined Statistical Area
M
Houston Major Highways
| Houstoi
OJ
Kilometers
Figure A-33. Map of CO monitor locations with AQS Site IDs for Houston, TX.
March 2009 A-27 DRAFT-DO NOT QUOTE OR CITE
-------
Table A-5. 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 repotting to
AQS in Houston, TX.
Houston
A B C D E
A
1.00
0.72
0.45
0.56
0.68
0.0
0.3
0.4
0.5
0.3
0.00
0.29
0.47
0.73
0.24
0.0
17.5
16.7
19.8
32.2
B
1.00
0.56
0.65
0.63
0.0
0.4
0.5
0.4
0.00
0.47
0.73
0.29
0.0
16.3
25.2
39.7
C
1.00
0.53
0.43
0.0
0.5
0.4
0.00
0.74
0.47
0.0
9.3
23.5
D
1.00
0.57
0.0
0.4
0.00
0.72
Legend
0.0
14.5
E
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-28
DRAFT-DO NOT QUOTE OR CITE
-------
A
B
C
D
E
Mean
0.42
0.39
0.35
0.14
0.33
Obs
23997
25241
24922
25285
24480
SD
0.27
0.33
0.26
0.22
0.16
1.5
1.4
1.3
1.2
1.1
1.0
0.9-
ro 0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1 -
0.0
rm
12 34
I I I I II I I I
12 34 12 34
I I I I II I I I
12 34 12 34
season
Figure A-34. 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.
March 2009
A-29
DRAFT-DO NOT QUOTE OR CITE
-------
Los Angeles Combined Statistical Area
• Los Angeles CO Monitors
— Los Angeles Major Highways
| Los Angeles
200
h Kilometers
Figure A-35. Map of CO monitor locations with AQS Site IDs for Los Angeles, CA.
March 2009
A-30
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-6. 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 repotting to
AQS in Los Angeles, CA.
Los Angeles
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
s
T
u
V
w
X
A
1.00
0.45
0.43
0.46
0.44
0.21
0.57
0.41
0.17
0.19
0.43
0.27
0.29
0.14
0.42
0.34
0.54
0.33
0.47
0.56
0.17
0.24
0.61
0.48
0.0
0.5
1.0
0.8
0.7
1.5
0.8
0.8
1.1
0.7
0.5
0.6
0.6
0.7
0.5
0.8
0.7
0.5
0.8
0.4
0.6
0.6
0.5
0.6
0.00
0.38
0.35
0.38
0.33
0.44
0.33
0.43
0.47
0.72
0.47
0.62
0.43
0.64
0.52
0.33
0.33
0.36
0.43
0.38
0.62
0.50
0.30
0.35
0
50.0
36.5
29.0
56.5
35.0
17.7
18.7
42.4
51.0
61.9
62.2
33.9
51.2
60.6
23.6
52.6
131.4
49.2
74.8
117.7
68.4
27.4
59.9
B
1.00
0.69
0.71
0.64
0.50
0.63
0.49
0.43
0.59
0.36
0.42
0.59
0.47
0.57
0.59
0.57
0.35
0.55
0.55
0.25
0.41
0.50
0.53
0.0
0.8
0.6
0.7
1.3
0.7
0.8
0.9
0.6
0.6
0.7
0.5
0.7
0.6
0.6
0.7
0.6
0.7
0.5
0.7
0.6
0.6
0.6
0.00
0.38
0.38
0.34
0.50
0.41
0.49
0.48
0.69
0.50
0.60
0.39
0.59
0.50
0.37
0.42
0.38
0.46
0.39
0.62
0.50
0.40
0.41
0
18.9
21.2
17.8
26.4
65.0
31.6
35.2
11.4
37.4
74.9
53.7
64.4
85.9
48.5
98.1
178.1
96.1
112.0
161.1
115.3
76.4
109.1
C
1.00
0.84
0.74
0.70
0.78
0.67
0.62
0.64
0.39
0.59
0.72
0.70
0.69
0.75
0.70
0.40
0.52
0.58
0.35
0.58
0.59
0.65
0.0
0.5
0.6
0.9
0.5
0.7
0.8
1.0
1.1
1.1
0.7
0.9
1.0
0.6
0.6
1.2
0.8
1.0
1.2
1.0
0.7
0.7
0.00
0.29
0.28
0.35
0.26
0.39
0.41
0.72
0.52
0.65
0.41
0.63
0.56
0.26
0.29
0.46
0.42
0.46
0.67
0.53
0.29
0.35
0
14.7
20.0
29.1
53.5
18.1
40.8
27.0
30.1
57.4
51.8
66.3
84.6
43.6
88.2
167.7
85.3
106.5
142.7
97.8
63.8
96.3
D
1.00
0.64
0.74
0.77
0.67
0.60
0.59
0.34
0.53
0.70
0.58
0.63
0.74
0.62
0.32
0.49
0.53
0.34
0.51
0.57
0.58
0.0
0.7
1.0
0.5
0.7
0.8
0.8
0.9
0.9
0.6
0.8
0.8
0.5
0.6
1.0
0.8
0.8
1.0
0.8
0.6
0.6
0.00
0.34
0.41
0.33
0.44
0.45
0.73
0.52
0.63
0.40
0.63
0.55
0.31
0.37
0.46
0.45
0.46
0.66
0.53
0.36
0.39
0
31.8
15.3
43.8
11.7
27.1
22.7
44.7
67.5
37.3
51.6
70.2
29.8
77.4
157.4
75.1
93.4
143.4
95.9
55.2
87.9
E
1.00
0.49
0.70
0.54
0.44
0.52
0.53
0.53
0.61
0.57
0.55
0.61
0.73
0.42
0.62
0.66
0.31
0.55
0.58
0.71
0.0
1.2
0.6
0.8
0.9
0.9
0.8
0.9
0.7
0.8
0.8
0.7
0.6
0.9
0.7
0.8
1.0
0.8
0.6
0.6
0.00
0.41
0.30
0.42
0.44
0.72
0.49
0.62
0.41
0.62
0.54
0.29
0.29
0.40
0.40
0.40
0.64
0.50
0.30
0.32
0
42.1
73.4
38.0
52.3
29.1
20.4
64.0
68.3
80.7
101.2
61.5
108.0
187.6
105.2
125.1
158.2
115.6
83.8
116.3
F
1.00
0.65
0.53
0.70
0.63
0.23
0.55
0.71
0.67
0.57
0.78
0.54
0.28
0.35
0.39
0.33
0.51
0.39
0.44
0.0
1.0
1.0
1.0
1.5
1.6
1.6
1.1
1.3
1.5
0.9
1.1
1.7
1.2
1.6
1.7
1.5
1.3
1.3
0.00
0.29
0.39
0.39
0.76
0.60
0.72
0.51
0.70
0.65
0.30
0.30
0.56
0.45
0.56
0.73
0.61
0.35
0.42
0
45.0
23.8
11.8
20.4
58.2
82.5
27.4
38.6
59.5
23.8
74.8
154.5
73.7
86.0
152.6
103.4
57.0
88.6
G
1.00
0.65
0.52
0.51
0.47
0.61
0.69
0.60
0.70
0.77
0.78
0.47
0.60
0.69
0.40
0.58
0.78
0.68
0.0
0.6
0.8
1.0
0.9
0.9
0.7
0.9
0.9
0.5
0.4
1.0
0.7
0.8
1.1
0.9
0.5
0.6
0.00
0.34
0.38
0.73
0.53
0.66
0.44
0.66
0.58
0.20
0.19
0.47
0.38
0.46
0.68
0.54
0.18
0.31
0
35.4
48.6
63.9
79.6
75.4
31.4
46.3
48.9
24.3
35.0
114.2
31.8
58.1
113.3
62.4
12.0
44.2
H
1.00
0.41
0.40
0.32
0.45
0.54
0.50
0.55
0.58
0.53
0.39
0.39
0.47
0.31
0.46
0.53
0.52
0.0
0.9
1.0
1.0
1.0
0.8
0.9
0.9
0.7
0.7
1.0
0.8
0.9
1.1
0.9
0.7
0.7
0.00
0.46
0.75
0.58
0.70
0.52
0.68
0.62
0.37
0.35
0.54
0.46
0.53
0.71
0.59
0.35
0.41
0
34.7
34.4
46.2
59.6
37.7
54.0
69.5
28.1
70.1
149.6
67.2
89.2
131.7
84.3
46.0
78.6
I
1.00
0.59
0.15
0.43
0.63
0.64
0.48
0.62
0.46
0.24
0.31
0.36
0.24
0.43
0.28
0.39
0.0
1.0
1.2
1.1
0.8
0.9
1.1
0.7
0.8
1.2
1.0
1.1
1.3
1.1
0.9
0.9
0.00
0.75
0.60
0.69
0.48
0.64
0.58
0.37
0.39
0.53
0.48
0.54
0.70
0.59
0.41
0.46
0
26.4
69.4
94.0
23.2
29.5
52.1
24.6
74.1
152.4
74.0
81.1
159.7
109.6
60.3
90.1
J
1.00
0.24
0.39
0.58
0.60
0.42
0.59
0.45
0.26
0.43
0.43
0.18
0.41
0.27
0.43
0.0
0.6
0.6
0.6
0.5
0.5
0.8
0.9
0.5
0.9
0.5
0.6
0.6
0.9
0.8
0.00
0.73
0.74
0.71
0.69
0.73
0.72
0.73
0.70
0.74
0.70
0.75
0.72
0.73
0.72
0
48.7
84.4
47.4
55.8
78.3
44.2
95.0
174.8
93.7
106.1
166.0
118.6
75.8
108.0
K
1.00
0.40
0.29
0.25
0.28
0.34
0.54
0.40
0.46
0.56
0.25
0.39
0.53
0.50
0.0
0.5
0.7
0.7
0.5
0.9
0.8
0.4
0.9
0.3
0.5
0.5
0.6
0.6
0.00
0.62
0.53
0.67
0.59
0.52
0.53
0.45
0.56
0.47
0.62
0.54
0.51
0.51
0
48.4
81.8
96.2
114.6
73.4
114.5
192.2
110.8
135.3
148.8
110.7
88.3
119.3
L
1.00
0.56
0.55
0.51
0.63
0.56
0.41
0.41
0.50
0.45
0.67
0.49
0.50
0.0
0.7
0.6
0.4
0.9
0.9
0.4
0.9
0.4
0.4
0.4
0.7
0.7
0.00
0.60
0.66
0.61
0.64
0.67
0.56
0.67
0.57
0.62
0.57
0.65
0.64
0
94.7
112.0
122.7
84.2
104.6
171.9
99.2
132.5
104.0
75.4
77.9
100.4
M
1.00
0.75
0.67
0.81
0.62
0.37
0.49
0.53
0.31
0.54
0.47
0.55
0.0
0.5
0.5
0.5
0.7
0.7
0.8
0.6
0.8
0.6
0.7
0.6
0.00
0.55
0.46
0.38
0.46
0.41
0.49
0.42
0.62
0.50
0.45
0.46
0
17.3
32.9
10.5
51.3
129.2
51.7
58.7
144.7
93.7
41.6
68.5
March 2009
A-31
DRAFT-DO NOT QUOTE OR CITE
-------
1.00
0.62
0.72
0.56
0.34
0.40
0.46
0.28
0.55
0.36
0.50
0.0
0.5
0.7
0.8
0.6
0.9
0.5
0.7
0.6
0.8
0.7
0.00
0.56
0.62
0.66
0.59
0.67
0.59
0.69
0.62
0.65
0.64
0
23.7
27.9
57.1
129.6
59.3
55.1
158.5
107.5
54.9
76.9
1.00
0.67
0.61
0.44
0.43
0.51
0.33
0.45
0.57
0.51
0.0
0.8
0.8
0.3
0.9
0.3
0.4
0.4
0.7
0.6
0.00
0.54
0.58
0.47
0.59
0.50
0.63
0.56
0.56
0.56
0
41.5
43.3
107.9
47.5
32.4
152.3
102.5
52.6
64.6
1.00
0.66
0.40
0.50
0.55
0.34
0.58
0.54
0.55
0.0
0.6
1.0
0.8
0.8
1.0
0.8
0.6
0.7
0.00
0.24
0.42
0.38
0.43
0.66
0.51
0.25
0.33
0
51.0
130.6
50.2
63.7
137.1
86.4
35.8
65.7
1.00
0.48
0.67
0.72
0.35
0.55
0.70
0.73
0.0
0.9
0.6
0.7
1.0
0.8
0.5
0.5
0.00
0.47
0.37
0.46
0.68
0.54
0.20
0.30
0
80.1
6.1
30.5
110.6
62.8
27.4
21.3
R
1.00
0.40
0.49
0.29
0.38
0.47
0.46
0.0
0.9
0.3
0.3
0.4
0.7
0.7
0.00
0.50
0.32
0.55
0.44
0.43
0.43
0
82.4
75.7
123.6
102.8
104.2
73.3
S
1.00
0.66
0.19
0.38
0.54
0.64
0.0
0.8
1.1
0.9
0.6
0.6
0.00
0.49
0.69
0.58
0.37
0.41
0
36.64
105.4
57.05
22.78
17.67
T
1.00
0.30
0.49
0.68
0.69
0.0
0.4
0.4
0.6
0.6
0.00
0.58
0.47
0.43
0.42
0
137.8
92.3
54.8
48.0
u
1.00
0.43
0.36
0.31
0.0
0.4
0.8
0.8
0.00
0.59
0.66
0.65
0
51.0
103.7
90.2
V 1.00 0.47 0.51
0.0 0.7 0.6
0.00 0.52 0.52
0 52.7 44.9
w
1.00 0.64
0.0 0.5
0.00 0.29
Legend
0 32.7
X
r
1.00
P90
0.0
COD
0.00
d
0
March 2009
A-32
DRAFT-DO NOT QUOTE OR CITE
-------
2.5001
rrn
1234
i i i i l i i i i
1234 1234
I I I I I I I I I
1234 1234
season
I I I I I I I I I
1234 1234
Figure A-36a. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los
Angeles, CA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
March 2009
A-33
DRAFT-DO NOT QUOTE OR CITE
-------
Mean 0.69
Obs 24259
SD 0.56
4.5-
4.0-
3.5-
3.0 -
2.5-
E
CL
Q-
C
o
ro
~ 2.0 i
(1)
o
c
° 1 5 H
o 1
1.0-
0.5-
o.o-
J
0.24
24965
0.37
K
0.30
24860
0.25
L
0.23
24135
0.29
M
0.42
24264
0.46
N
0.31
24760
0.47
0
0.26
24831
0.25
!li III! Il.l
I I I I
1234
I I I I I I I I I
1234 1234
season
1 234
P
0.62
24705
0.55
il
I I I I I I I I I
1234 1234
Figure A-36b. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los
Angeles, CA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
March 2009
A-34
DRAFT-DO NOT QUOTE OR CITE
-------
Q
R
S
T
U
V
W
X
Mean
0.67
0.25
0.60
0.29
0.17
0.30
0.59
0.53
Obs
24885
24938
24778
24792
24105
24796
24767
24844
SD
0.42
0.14
0.46
0.20
0.17
0.28
0.32
0.38
4.5-
4.0-
3.5-
£
Q.
Q.
C
o
3.0-
2.5-
i 2.0-
c
03
O
c
o
CJ
1.5-
1.0"
0.5"
o.o-
I I I I II I I I
1234 1234
I I I I
1234
I I I I II I I I
1234 1234
season
Illl
llll
1 234
I I I I I II I I
1234 1234
Figure A-36c. Box plots illustrating the seasonal distribution of hourly CO concentrations in Los
Angeles, CA. Note: 1 = winter, 2, = spring, 3 = summer, and 4 = fall on the x-axis.
March 2009
A-35
DRAFT-DO NOT QUOTE OR CITE
-------
New York Combined Statistical Area
• New York CO Monitors
— New York Major Highways
| New York
120
m Kilometers
Figure A-37. Map of CO monitor locations with AQS Site IDs for New York City, NY.
March 2009
A-36
DRAFT-DO NOT QUOTE OR CITE
-------
Table A-7. 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 repotting to
AQS in New York City, NY.
New York
A
B
c
D
E
F G
H
I
A
1.00
0.58
0.65
0.55
0.40
0.56 0.56
0.41
0.30
0.0
0.4
0.7
0.4
0.4
O
4^
O
4^
0.5
0.8
0.00
0.22
0.28
0.25
0.25
0.24 0.23
0.28
0.75
0.0
7.0
15.9
45.8
70.6
43.7 10.5
17.8
76.5
B
1.00
0.64
0.55
0.35
0.54 0.54
0.59
0.49
0.0
0.8
0.4
0.5
O
o
0.4
0.7
0.00
0.29
0.23
0.26
0.23 0.23
0.23
0.74
0.0
16.8
45.4
71.5
38.1 15.0
24.5
82.9
C
1.00
0.54
0.32
0.48 0.52
0.43
0.31
0.0
0.9
0.9
0.9 0.7
0.9
1.3
0.00
0.35
0.34
0.34 0.24
0.35
0.81
0.0
29.9
55.0
35.7 8.9
20.5
85.5
D
1.00
0.50
0.57 0.41
0.46
0.33
0.0
0.4
O
o
0.4
0.7
0.00
0.24
0.23 0.28
0.27
0.72
0.0
27.5
36.7 37.5
45.1
107.8
E
1.00
0.47 0.33
0.33
0.32
0.0
O
4^
O
4^
0.4
0.6
0.00
0.23 0.25
0.27
0.73
0.0
61.9 61.0
65.0
120.3
F
1.00 0.41
0.34
0.31
O
o
o
0.4
0.7
0.00 0.26
0.27
0.72
0.0 43.6
55.8
119.7
G
1.00
0.46
0.29
0.0
0.4
0.7
0.00
0.26
0.77
0.0
12.3
76.8
H
1.00
0.43
0.0
0.6
0.00
0.73
Legend
0.0
65.1
1
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-37
DRAFT-DO NOT QUOTE OR CITE
-------
A
B
C
D
E
F
G
H
!
Mean
0.55
0.52
0.85
0.48
0.50
0.49
0.62
0.47
0.12
Obs
23113
25150
25646
25028
25727
25691
25547
25022
25749
SD
0.27
0.30
0.43
0.27
0.24
0.25
0.21
0.23
0.17
E
Q_
Q.
C
0
03
i—
c
01
o
c
o
u
2.0
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
1 2 34
in
1111 m i m
1234 1234
I I I I
1234
I I I I I I l I l
1234 1234
season
I I I I I i I I i
1234 1234
Figure A-38. 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.
March 2009
A-38
DRAFT-DO NOT QUOTE OR CITE
-------
Seattle Combined Statistical Area
Seattle CO Monitors
Seattle Major Highways
| Seattle
120
Kilometers
Figure A-39. Map of CO monitor locations with AQS Site IDs for Seattle, WA.
March 2009
A-39
DRAFT-DO NOT QUOTE OR CITE
-------
Figure A-40. 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.
March 2009
A-40
DRAFT-DO NOT QUOTE OR CITE
-------
St Louis Combined Statistical Area
St Louis Major Highways
| StLoi
Kilometers
Figure A-41. Map of CO monitor locaitons with AQS Site IDs for St. Louis, MO.
Table A-8. 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 repotting to
AQS in St. Louis, MO.
St. Louis
ABC
1.00
0.19
0.60
0.0
0.5
0.3
0.00
0.40
0.24
0.0
21.2
9.5
1.00
0.19
0.0
0.5
0.00
0.42
Legend
0.0
19.8
r
1.00
P90
0.0
COD
0.00
d
0.0
March 2009
A-41
DRAFT-DO NOT QUOTE OR CITE
-------
A
B
C
Mean
0.44
0.43
0.42
Obs
25325
25879
25938
SD
0.25
0.25
0.29
1.3
1.2
1.1
1.0
0.9
0.8
0.7
ro 0.6
o 0.5
0.4
0.3
0.2
0.1
0.0
-0.1
I I I I
12 3 4
III]
12 3 4
season
I I I I
12 3 4
Figure A-42. 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.
March 2009
A-42
DRAFT-DO NOT QUOTE OR CITE
-------
Winter
Spring
1 -
1 ¦
2 •
2 ¦
3 ¦
3 ¦
4 ¦
o
o
4 ¦
o
5 •
o
5 ¦
o
6 ¦
(SD
6 •
m
7 ¦
GID
7 ¦
©
8 ¦
8 ¦
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
T
1.0
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4
0.6
1.0
Fall
o o
o
00©
©
GD
t—¦—i—'—i—¦—i—¦—i—¦——¦—i—•—i—'—i—¦—i—1—r
-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-43. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2,
(2) NO2, (3) O3, (4) PM10, and (5) PM2.5 concentrations for Anchorage, AK. Also shown
are correlations between 24-h average 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.
March 2009
A-43
DRAFT-DO NOT QUOTE OR CITE
-------
Winter
Spring
1 ¦
1 ¦
2 •
oo
2 •
o o
3 ¦
3 •
o o
4 '
4 ¦
5 ¦
o o
5 ¦
o o
6 •
o
6 •
GOO
7 ¦
-------
Winter
Spring
1 ¦
OGDO
1 ¦
OCED
2 •
OGD
2 •
©
3 ¦
o
3 •
00
4 '
O O O
4 ¦
OO
5 ¦
oo
5 ¦
oo o
6 •
OK®
6 •
(OxOGD
7 ¦
(Hi
7 •
(OS)
8 •
0©
8 ¦
OT3D
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
ODD
OGSS>
O ®D
-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)
cm
OGDO
o
o o
OQO
uyy/
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-45. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2,
(2) NO2, (3) O3, (4) PM10, and (5) PM2.5 concentrations for Boston, MA. Also shown are
correlations between 24-h average 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.) Red bars denote the median, and green stars denote the arithmetic
mean.
March 2009
A-45
DRAFT-DO NOT QUOTE OR CITE
-------
Winter
Spring
1 ¦
GD
1 ¦
<~>
2 •
O
2 •
O
3 ¦
i i i i i
8 ¦
i i i i i i i i i i
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
1 ¦
o
1 •
OO
2 •
oo
2 •
O
3 '
<0)00
3 "
-------
Winter
Spring
OOGDOdD
ODD
QBE) OO
(3®
W
SC'K: i')
I< I O il).®
K I > )
©k* i >
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
_1—,—|—,—|—i—|—,—|—i——i—|—i—|—i—|—i—|—i—r
-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 ¦
00
GOOD
1
2 ¦
oocM»)
2
3 ¦
OU'IOXO
3
4 ¦
DOOMED
4
5 •
C
5
6 ¦
6
7 '
(Ml
7
8 ¦
8
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 06 0.8 1.0
r (correlation coefficient)
(jOOGDO
tu <•>: ij)»
mmm
K'MJI
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4
r (correlation coefficient)
0.8 1.0
Figure A-47. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2,
(2) NO2, (3) O3, (4) PM10, and (5) PM2.5 concentrations for Los Angeles, CA. Also
shown are correlations between 24-h average 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.)
March 2009
A-47
DRAFT-DO NOT QUOTE OR CITE
-------
Winter
Spring
m> o
imoo
WD
©o o
((<:
o«¦>
(a®
((COB
CD CD)
QUE)
a <: ¦ •)»
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
~i—¦—i—¦—i—¦—i—¦—i—• ¦—i—«—i—1—i—«—i—i—r
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
:• («: ~ )>
-------
Winter
1 ¦
O OO
1
2 ¦
O 0©
2
3 ¦
3
4
o cs®
4
5 ¦
CD
5
6 ¦
ox®
6
7 ¦
(fJJil
7
8 ¦
(T«B»
8
!
GSE>
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0 2 04 0.6 0.8 1.0
~i—1—i—»—i—1—i—1—i—1——1—i—1—i—1—i—1—i—1—r~
-1.0 -0.8 -06 -04 -0.2 00 0.2 0.4 06 0.8 1.0
Summer
Fall
1 -
©
1 ¦
O O O
2 •
GO
2 •
O CD
3 ¦
OO O
OO
3 ¦
GD
4 ¦
c
4 •
5 ¦
OO
5 ¦
C2D>
6 ¦
CHE)
6 ¦
(SSD
7 ¦
o ©
7 ¦
CD
8 ¦
O (EBD
8 ¦
OO®
-1.0 -0,8 -0.6 -0.4 -0,2 0.0 0.2 04 06 0.8 1,0
r (correlation coefficient)
-1.0 -0.8 -0.6 -04 -0.2 0.0 0.2 0.4 0,6 08 1.0
r (correlation coefficient)
Figure A-49. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2,
(2) NO2, (3) O3, (4) PM10, and (5) PM2.5 concentrations for Phoenix, AZ. Also shown
are correlations between 24-h average 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.)
March 2009
A-49
DRAFT-DO NOT QUOTE OR CITE
-------
Winter
Spring
1 ¦
o
1 ¦
o
2 ¦
o
2 ¦
o
3 ¦
o
3 ¦
o
4 '
4 '
5 ¦
5 ¦
6 ¦
(O
6 ¦
OSS)
7 •
(M)
7 ¦
8 ¦
-10 -08 -06 -04 -02 00 0.2 04 06 08 1.0
-1.0 -0.8 -06 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
r (correlation coefficient)
r (correlation coefficient)
Figure A-50. Seasonal plots of correlations between hourly CO concentration with hourly (1) SO2,
(2) NO2, (3) O3, (4) PM10, and (5) PM2.5 concentrations for Seattle, WA. Also shown are
correlations between 24-h average 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.)
March 2009
A-50
DRAFT-DO NOT QUOTE OR CITE
-------
Annex B. Dosimetry Studies
Table B-1. Recent studies related to CO dosimetry and pharmacokinetics.
Reference
Purpose
Findings
Alcantara et al. (2007)
To use a quantum mechanics/molecular mechanics
approach to understand the cooperativity of Hb ligand
binding and 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)
To determine if low concentrations of CO would affect
exercise performance and myocardial perfusion in
young healthy men.
Men with COHb levels between 4-6% had decreased exercise
performance measured by decreased mean duration of exercise (1.52
min) and maximal effort described by metabolic equivalent units
(2.04). No changes were seen in lactate/pyruvate ratio, arrhythmias,
or myocardial perfusion.
Andersson et al. (2000) To investigate if CO could be endogenously produced Both nose and paranasal sinuses contained HO-like immunoreactivity,
in the nose and paranasal sinuses. mostly in the respiratory epithelium, indicating local CO production in
the upper respiratory airways.
Bruce and Bruce (2003)
Bruce and Bruce (2006)
To create a mathematical model to predict uptake and This model contains 5 compartments: lung, arterial blood, venous
distribution of CO in both vascular and tissue blood, muscle tissue, and nonmuscle tissue. It was constructed to
compartments during constant or variable inhalation include tissue compartment flux and difference between venous and
levels of CO. arterial COHb for short exposures which is not possible with the CFK
model.
To use their mathematical multicompartment model
along with experimental data to predict the factors that
influence the washout rates of CO, along with
predicting the rates of CO uptake, distribution in
vascular and extravascular (muscle and non-muscle
tissue) compartments, and washout over a range of
exposure and conditions.
Rates of CO washout follow a biphasic elimination where washout
was faster immediately post exposure. The difference in rates is likely
due to slow equilibration between vascular and extravascular
compartments. Important factors contributing to washout kinetics
include: peak COHb level, exposure duration and concentration, time
after exposure samples were obtained, and individual variability.
Bruce et al. (2008)
To develop a mathematical model able to integrate a
large body of indirect experimental findings on the
uptake and distribution of CO by accounting for
arteriole to venule shunting via intra-tissue pathways
and diffusion of blood gases into tissues from pre-
capillary vessels like arterioles.
The former model of Bruce and Bruce (2006) was altered by adding a
mass balance equation for O2 so p02 is directly calculated in the
compartments and the muscle compartment is divided into two sub-
compartments of muscle and non-muscle tissue. CO uptake from
blood by muscle is much slower than O2, thus COHb% will fall rapidly
while COMb% could remain high.
Carraway et al. (2000) To test the hypothesis that HO-1 gene expression and Rats were exposed to HH (17,000 ft) for 1 -21 days. COHb increased
protein are upregulated in the lungs of rats during after 1 day and progressively after 14 days. HO-1 protein and activity
chronic hypoxia. were upregulated during early chronic hypoxia. This HO-1 was
localized to inflammatory cells and then to newly muscularized
arterioles.
Castillo et al. (2006) To describe a new method for measurement of CO
diffusing capacity (DlCO) and alveolar volume (Va) in
sleeping infants (6-22 months old), using a single
4-sec 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)
To present an analytical expression for diffusing
capacity of CO, NO, CO2, and O2 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 O2
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.
Cronje et al. (2004)
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 CO/O2 ratio.
Results indicate that tissue and blood [CO] dissociate during CO
inhalation, but [CO] does not follow blood [CO] or 1/p02 as in the
Warburg theory during intake or elimination. Tissue [CO] increases
later during the resolution period and varies significantly among
animals and tissues. The deviation from the predicted values in the
brain is likely due to the release of heme and increase in NADPH
stimulating endogenous CO production by HO.
March 2009
B-51
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Purpose
Findings
Dutton et al. (2001)
To monitor CO, NO2, 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, NO2, and PAH
indoors when the fireplaces were used. CO concentrations could
exceed 100 ppm. NO2 concentrations average 0.36 ppm over 4 hours.
PAH 4-h time average reached 35 ng/m3.
Hampson (2007)
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%.
Hsia (2002)
To review the current concepts and practical
relevance of the diffusing capacity/cardiac output
interaction, in hopes of aiding in the interpretation of
diffusing capacity, membrane diffusing capacity, and
capillary blood volume.
This review helped to understand the determinants of changes in
diffusing capacity, including hematocrit, erythrocyte distribution, blood
volume, lung volume, cardiac output, etc.
Lamberto et al. (2004)
To evaluate which component, alveolar membrane
diffusing capacity (Dm) and pulmonary capillary blood
volume (Vc), is responsible for decreased resting
DlCO in sarcoidosis patients and which component is
the best predictor of gas exchange abnormalities.
Patients with pulmonary sarcoidosis had decreased lung volumes, a
loss in DlCO, and gas exchange abnormalities during exercise
including decreased Pa02 and increased alveolar-arterial oxygen
pressure difference. Dm accounted for the majority of the decrease in
DlCO and was predictive for gas exchange abnormalities.
Levesque et al. (2000)
To describe the results of air quality monitoring in an
indoor ice skating rink during Monster Truck and car
demolition exhibitions.
Maximum time-weighted average levels of CO were 100 ppm with
several peaks exceeding 200 ppm (maximum: 1,600 ppm).
Marks et al. (2002)
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.
Merx et al. (2001)
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 PO2. Partially
02-saturated Mb (87% 02Mb) exposed to 20% CO had significantly
decreased LVDP (12%) and PVO2 (30%) in wild-type but not myo-'-
hearts.
Pelham et al. (2002)
To review the literature on exposure and effects of
mainly CO and NO2 in enclosed ice rinks.
CO levels as high as 300 ppm were recorded after episodes of
malfunctioning ice resurfacing equipment or inadequate ventilation.
Pesola et al. (2004)
To determine if healthy African Americans may be
misdiagnosed has having respiratory deficient due to
comparison using Caucasian-derived prediction
equation estimates of DlCO.
The lung volume of African American individuals is 10-15% lower than
Caucasians. The measured DlCO was consistently significantly lower
in African Americans than what would be predicted, thus the authors
suggest a race correction reduction of the Miller PEE for diffusion of
12%.
Pesola et al. (2006)
To determine if healthy Asians may be misdiagnosed
has having respiratory deficient due to comparison
using Caucasian-derived prediction equation
estimates of DlCO.
The lung volume of Asian individuals is 10-15% lower than
Caucasians, thus a Chinese derived prediction for DlCO should be
used.
Prommer and Schmidt (2007)
To determine the error in total Hb mass
measurements using the optimized CO-rebreathing
method due to loss of CO to Mb
Optimal blood mixing (when venous and arterial blood COHb% are
equivalent) was determined to be after 6 min. A small volume of
administered CO leaves the vascular space (0.32% per min). 2.3%
increase in total Hb mass would be found if CO diffusion was not
included.
Richardson et al. (2002)
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% O2 + 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.
Shimazu et al. (2000)
To study the effects of short-term (minutes) or long-
term (several hours) 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)
To discuss the findings of Weaver et al. (2000) on
COHb ti/2.
The authors discuss that CO elimination is biphasic and is heavily
affected by duration of exposure which was not taken into account in
the Weaver, et al. (2000) paper.
Takeuchi et al. (2000)
To examine the relationship between minute Patients were exposed to 400 to 1,000 ppm CO, resulting in 10-12%
ventilation and rate of COHb reduction during COHb. The half-time of COHb reduction was 78 ± 24 min during
breathing 100% O2 and during normocapnic hyperoxic 100% O2 treatment and 31 ± 6 min during normocapnic hyperpnea
hyperpnea. with O2 treatment.
Vreman et al. (2005)
To develop a sensitive and reproducible method of
CO quantification in rodent (mouse and rat) tissue
pre- and post-exposure in hopes of understanding
endogenous CO production.
Tissues were sonicated mixed with sulfosalicylic acid for 30 min at
0°C and then liberated CO was analyzed by gas chromatograph.
Blood contained the highest CO concentration. Lowest concentrations
were found in brain, testes, intestine, and lung (endogenously).
March 2009
B-52
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Purpose
Findings
Vreman et al. (2006) To test a method of CO quantification in frozen
postmortem human tissues from 3 determined
categories of fatalities: trauma with no suspected CO
exposure (controls), fire-related, and CO asphyxiation.
CO levels were analyzed in adipose, brain, muscle, heart, kidney,
lung, spleen, and blood (ordered from approximate low to high tissue
concentration). It was suggested that blood, muscle, brain, lung, and
kidney are suitable for diagnosing death due to lethal CO exposure
due to regression analysis against COHb values.
Weaver et al. (2000) To determine in COHb half-life is influenced by CO
poisoning vs. experimental CO exposure, loss of
consciousness, concurrent tobacco smoking, or P£>i.
COHb ti/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 O2 treatment.
Wu and Wang (2005) To review the endogenous production of CO through
HO, as well as discuss physiological roles for CO both
toxic and therapeutic.
CO is produced endogenously by HO-1 and -2 and acts as a
gasotransmitter, inducing cell signaling cascades. The review
discusses possible roles for CO in the various organ systems. Also, it
discusses the potential pharmacological and therapeutic applications
for CO.
March 2009
B-53
DRAFT-DO NOT QUOTE OR CITE
-------
Annex C. Epidemiology Studies
Table C-1. Studies of CO exposure and cardiovascular morbidity.
Study
Design
Concentrations
CO Effect Estimates (95% CI)
CHANGES IN HEART RATE AND HEART RATE VARIABILITY
Author: Chan et al. (2005)
Period of Study:
December 2001 -
February 2002
Location:
Taipei, Taiwan
Health Outcome: Various measures of
HRV via ambulatory ECG (Holter system)
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects)
Age Groups Analyzed:
40 - 75 years
Sample Description:
83 patients from the National Taiwan
University Hospital
Averaging Time:
1-h ma
Mean (SD) unit:
1.1 ppm
Range (Min, Max):
0.1,7.7
Copollutant: NR
Increment: NR
RR Estimate [Lower CI, Upper CI]
Lags examined (-h ma): 1, 2, 3, 4, 5, 6, 7, 8
CO had no statistically significant effect on SDNN, rMSSD,
LF, HF.
Author: Dales et al. (2004)
Period of Study: NR
Location:
Toronto, Canada
Health Outcome: Various measures of
HRV via Holter system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects)
Age Groups Analyzed:
51 - 88 years (mean 65 years)
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
PMzs: r= 0.17
Increment: NR
Regression co-efficient [Lower CI, Upper CI]
Lags examined: NR
CO had no statistically significant effect on LF, HF, HFLFR,
SDNN among those taking Beta-blockers. Whereas CO
had a positive effect on SDNN among those not taking
Beta-blockers. Slope = 0.0111 (0.002-0.020, p=0.02)
Author: Gold et al. (2000)
Period of Study:
June - September 1997
Location:
Boston, MA
Health Outcome (ICD9 or ICD10): Heart
Rate and various measures of HRV via
Holter system
Study Design: Panel/Cohort
Statistical Analyses: Linear regression
(fixed effects/random effects)
Age Groups Analyzed:
53 - 87 years
Sample Description: 21 active Boston
residents observed up to 12 times.
Averaging Time: 24-h
Mean (SD) unit:
0.47 ppm
Range (Min, Max): 012,
0.82
Copollutant: NR
Increment: 0 6 ppm
% Change [Lower CI, Upper CI]
Lags examined : 24-h
No significant effect with CO (no results recorded)
Author: Gold et al. (2005)
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:
61-88 years
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 CI, Upper CI]
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: Holguin et al.
(2003)
Period of Study:
February-April 2000
Location:
Mexico City, Mexico
Health Outcome: Various measures of
HRV via ECG
Study Design: Panel
Statistical Analyses: GEE
Age Groups Analyzed:
60-96 years (mean age 79 years)
Sample Description:
34 patients who were permanent residents
of a nursing home in the Northeast
metropolitan area.
Averaging Time: 24-h
Mean (SD) unit: 3 3 ppm
Range (Min, Max): 1 8
4.8
Copollutant: NR
Increment: 10 ppm
Regression Coefficients [Lower CI, Upper CI]
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)
March 2009
C-1
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Liao et al. (2004)
Period of Study:
1996- 1998
Location:
Forsyth County, NC;
Selected suburbs of
Minneapolis, MN;
Jackson, Ml
Health Outcome: Heart Rate & various
rates of HRV.
Study Design: Cohort
Statistical Analyses: Linear regression
Age Groups Analyzed:
45 - 64 years (mean 62 years)
Sample Description:
6784 study subjects from the
atherosclerosis risk in communities study
Averaging Time: 24-h
Mean (SD) unit:
0.65 ppm (0.44)
Range (Min, Max): NR
Copollutant: NR
Increment: 0.44 ppm
Regression coefficients Lags examined : 1
Lag 1 :
HF (log transformed): -0.033
LF (log transformed): 0.006
SDNN : -0.274
Heart Rate (bpm): 0.404*
Confidence Intervals not recorded
*p < 0.05
Author: Park et al. (2005b) Health Outcome: Various measures of
HRV via ECG
Study Design: Panel/Cohort
Statistical Analyses: Linear regression
Age Groups Analyzed:
21 - 81 years
Sample Description:
497 men from the Normative aging study
in Greater Boston
Period of Study:
2000 - 2003
Location:
Boston, MA
Averaging Time: 24-h
Mean (SD) unit:
0.50 ppm
Range (Min, Max):
0.13,1.8
Copollutant: NR
Increment: 0.24 ppm
% Change in HRV [Lower CI, Upper CI]
Lags examined: 4-h ma, 24-h ma, 48-h ma
Lag 4-h ma:
SDNN (Log 10): 2.0 (-2.9 to 7.3)
HF (Log 10): 8.8 (-4.6 to 24.1)
LF(Log10): 3.2 (-7.0 to 14.6)
LF :HF(Log10): -5.1 (-13.5 to 4.1)
Lag 24-h ma:
SDNN (Log 10): -2.2 (-7.7 to 3.6)
HF (Log10): -13.2 (-25.4 to 1.0)
LF(Log 10): -0.6 (-11.9 to 12.1)
LF :HF(Log10): 14.5 (2.9-27.5)
Lag 48-h ma:
SDNN(LoglO):-3.4 (-10.2 to 3.9)
HF (Log10): -13.8 (-.28.9 to 4.4)
LF (Log10): -2.4 (-16.2 to 13.6)
LF :HF (Log 10): 13.2 (-1.1 to 29.6)
Author: Peters et al.
(1999a)
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 years
Sample Description:
2681 men & women who participated in
the MONICA study
Averaging Time: 24-h
Mean (SD) unit:
During air pollution
episode: 4.54 mg/m3
Outside air pollution
episode: 4.51 mg/m3
Range (Min, Max):
During air pollution
episode: 2.39, 6.85
Outside air pollution
episode: 0.91,11.51
Respectively
Copollutant: NR
Increment: 6.6 mg/m3
Mean Change in Heart Rate (beats/min) [Lower CI,
Upper CI]
Lags examined: 0, 5-day avg
All
Lag 0:0.97 (0.02-1.91)
Lag 5-day avg : 0.70 (-0.09 to 1.48)
Men
Lag 0 : 0.95 (-0.37 to 2.27)
Lag 5-day avg : 0.91 (-0.25 to 2.07)
Women
Lag 0 : 0.98 (-0.37 to 2.34)
Lag 5-day avg : 0.52 (-0.55 to 1.59)
Author: Riojas-Rodriguez
et al. (2006)
Period of Study:
December 2001 - April
2002
Location:
Mexico City, Mexico
Health Outcome: Various measures of
HRV via Holter system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects models)
Age Groups Analyzed:
25 - 76 years (mean 55 years)
Sample Description:
30 patients from the Outpatient clinic of
the National Institute of Cardiology of
Mexico
Averaging Time: 24-h
Mean (SD) unit: 2 9 ppm
(personal monitor)
Range (Min, Max): 01,
18.0
Copollutant: NR
Increment: 1 ppm
Regression Coefficients [Lower CI, Upper CI]
Lags examined (per minutes): 5,10
Lag 5 minutes :
HF:-0.006 (-0.023 to 0.010)
LF : -0.024 (-0.041 to -0.007)
VLF : -0.034 (-0.061 to -0.007)
Notes: VLF = Very low frequency
Author: Schwartz et al.
(2005)
Period of Study: 1999
Location:
Boston, MA
Health Outcome: Measures of HRV via
Holter system
Study Design: Panel
Statistical Analyses: Linear regression
(hierarchical model)
Age Groups Analyzed:
61 - 89 years
Sample Description:
28 subjects living at or near an apartment
complex located on the same street at the
Harvard School of Public Health
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max): ppm
25th = 0.38; 75th = 0.54
Copollutant: correlation
PM2.5: r= 0.61
NO2: r = 0.55
S02: r = -0.18
03: r= 0.21
Increment: 016 ppm
% Change in HRV [Lower CI, Upper CI]
Lags examined : 24-h, 1-h
Lag 1-h:
SDNN : -2.6 (-5.6 to 0.5); rMSSD : -3.9 (-10.6 to 3.3);
PNN50 : -3.5 (-13.7 to 8.0); LF :HF : 4.5 (-1.2 to 10.5)
Lag 24-h:
SDNN : -4.2 (-0.6 to -7.7); rMSSD : -10.2 (-2.4 to -17.4);
PNN50 : -14.8 (-3.0 to -25.2); LF :HF : 6.2 (-0.6 to 13.4)
March 2009
C-2
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Tarkiainen et al.
(2003)
Period of Study:
October 1997 - May 1998
Location:
Kuopio, Finland
Health Outcome: Various measures of
HRV via Ambulatory ECG (Holter system)
Study Design: Panel
Statistical Analyses: ANOVA for
repeated errors (GLM)
Age Groups Analyzed:
Age 55 - 68 years
Sample Description: 6 male patients with
angiographically verified CAD
Averaging Time: 24-h
Mean (SD) unit:
4.6 ppm (maximum of CO
episode) (personal
monitoring)
Range (Min, Max): 0 5,
27.4 (maximum of CO
episode)
Copollutant: NR
Increment: NR
RR Estimate [Lower CI, Upper CI]
Lags examined : 5 minute prior to CO episode, 5 minute
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)
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 (Holter system)
Study Design: Panel
Statistical Analyses: Linear regression
(mixed model)
Age Groups Analyzed: Mean age across
3 cities; 64-71 years.
Sample Description:
131 subjects with Stable CAD followed for
6 months with bi-weekly clinical visits.
Averaging Time: 24-h
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfert: 0.4 mg/m3
Helsinki: 0.4 mg/m3
Range (Min, Max):
Amsterdam: 0.4,1.6
Erfert: 0.1, 2.5
Helsinki: 0.1,1.0
Copollutant: correlation
Amsterdam:
PM25: r= 0.58
N02: r = 0.76
Erfert:
PM10: r = 0.77
N02: r = 0.86
Helsinki:
PM10: r = 0.40
N02: r = 0.32
Increment: 1 mg/m3
Regression co-efficient [Lower CI, Upper CI]
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)
Period of Study:
1999-2000
Location:
Atlanta, GA
Health Outcome: Various measures of
HRV via Holter system
Study Design: Panel
Statistical Analyses: Linear regression
(mixed effects models)
Age Groups Analyzed:
Mean 65 years - IQR 55- 73 years.
Sample Description:
18 subjects with COPD and 12 subjects
with recent Ml.
Averaging Time: 1-h
Mean (SD) unit:
362.0 ppb
Range (Min, Max):
25th = 221.5; 75th = 398.1
Copollutant: correlation
PM2.5: r= 0.43
Increment: NR
RR Estimate [Lower CI, Upper CI]; lag :
Lags examined (h ma): 1, 4, 24
No CO results reported.
ONSET OF CARDIAC ARRHYTHMIA
Author: Berger et al.
(2006)
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 years (mean 76years)
Sample Description:
57 men with CHD
Averaging Time: 24-h
Mean (SD) unit: 0.52
mg/m3
Range (Min, Max): 011,
1.93
Copollutant: correlation
NR
Increment:
All: 0.27 mg/m3
5-day avg : 0.22 mg/m3
RR Estimate [Lower CI, Upper CI]
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 CI, Upper CI]
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)
March 2009
C-3
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Dockery et al.
(2005)
Period of Study:
1995-2002
Location:
Boston, MA
Health Outcome:
Tachyarrhythmias:
Study Design: Panel
Statistical Analyses: Logistic regression
(GEE)
Age Groups Analyzed:
19 - 90 years; mean age 64 years
Sample Description:
203 cardiac patients with ICDs within
40km of air monitoring site at Harvard
School of Public Health, Boston
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
25th = 0.53; 75th = 1.02
Copollutant: NR
Increment: 0 48 ppm
OR for Ventricular Arrhythmia [Lower CI, Upper CI]
Lags examined (days): 0,1, 2, 3
Lag 2day ma: 1.14 (0.95-1.29)
Among those who had an Arrhythmia -
within 3 days : 1.65 (1.17-2.33)
later than 3 days : 1.04 (0.83-1.29)
Author: Metzgeret al.
(2007)
Period of Study:
1993-2002
Location:
Atlanta, GA
Health Outcome:
Cardiac Arrhythmia, ICD,
Ventricular tachyarrhythmia
Study Design: Panel
Statistical Analyses:
Logistic regression (GEE)
Age Groups Analyzed:
15 - 88 years
Sample Description:
518 patients with ICDs with at least one
ventricular tachyarrhythmic event
Averaging Time: 1-h
Mean (SD) unit: 17 ppm
Range (Min, Max): 01,
7.7
Copollutant: NR
Increment: 1 ppm
OR for Tachyarrhythmic event [Lower CI, Upper CI]
Lags examined (days): 0
Results for All events
Lag 0 : 0.999 (0.970-1.028)
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)
Author: Peters et al.
(2000b)
Period of Study:
1995- 1997
Location:
Eastern Massachusetts
Health Outcome:
Defibrillated discharges for ventricular
tachycardia or fibrillation
Study Design: Panel
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Mean age of 62 years
Sample Description:
100 patients with ICDs
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
PM25: r= 0.56
N02: r = 0.71
S02: r = 0.41
O3: r = -0.40
Increment:
0.65 ppm (Lags 0,1, 2, 3); 0.42 ppm (Lag 5-day mean)
OR for Defibrillated Discharge [Lower CI, Upper CI]
Lags examined (days): 0,1, 2 ,3, 5-day mean
At least one discharge:
Lag 0:1.07 (0.62-1.86); Lag 1 :1.06 (0.61-1.85);
Lag 2:1.05 (0.62-1.77); Lag 3 : 0.09 (0.65-1.83);
Lag 5-day mean : 1.23 (0.71 - 2.12)
At least 10 discharges:
Lag 0:1.12 (0.54-2.32); Lag 1 :1.13 (0.54-2.33);
Lag 2:1.62 (0.85-3.09); Lag 3 :1.98 (1.05-3.72);
Lag 5-day mean : 1.94 (1.01 -.75)
Author: Rich et al. (Rich et
al.)
Period of Study:
February - December
2000
Location:
Vancouver, Canada
Health Outcome: Cardiac arrhythmia via
patients ICD
Study Design:
Case-crossover
Statistical Analyses:
Conditional Logistic regression
Age Groups Analyzed:
15 - 85 years
Sample Description:
34 patients who experienced at least 1
ICD discharge (8201 person days)
Averaging Time: 24-h
Mean (SD) unit:
553.8 ppb
Range (Min, Max):
IQR: 162 7
Copollutant: correlation
PM10: r = 0.40
SO2: r =0.75
N02: r =0.68
O3: r = -0.56
Increment: NR
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1, 2, 3
No significant effect (results not reported in table).
Author: Rich et al. (2005)
Period of Study:
1995- 1999
Location:
Boston, MA
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
Averaging Time:
1-h & 24-h
Mean (SD) unit: NR
Range (percentiles):
1-h:
0.46
1.04
25th
75th
24-h
25th
75th
Copollutant: NR
= 0.52
= 1.03
Increment: 0.56 ppm; 0.54; 0.51; 0.49 respectively for
results shown below
OR Estimate [Lower CI, Upper CI]
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)
March 2009
C-4
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Rich et al. (2006b)
Period of Study:
2001 & 2002
Location:
St. Louis, MO
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.
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
25th = 0.4; 75th = 0.6
Copollutant: NR
Increment: 0 2 ppm
OR for Ventricular Arrhythmia [Lower CI, Upper CI]
Lags examined : 0-23 h-ma
0-23h-ma : 0.99 (0.80-1.21)
Author: Rich et al. (2006a)
Period of Study:
1995- 1999
Location:
Boston, MA
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
Averaging Time: 1-h &
24-h
Mean (SD) unit: NR
Range (Min, Max):
1-h:
25th = 0.46: 75th
24-h:
25th = 0.52; 75th =
Copollutant: NR
1.04
1.03
Increment:
Lag (hrs) 0 : 0.58 ppm
Lag (hrs) 0-23 : 0.51 ppm
OR for Episode of Atrial Fibrillation [Lower CI, Upper
CI]
Lags (h): 0, 0-23
Lag 0:0.87 (0.56-1.37)
Lag 0-23:0.71 (0.39-1.28)
Author: Sarnat et al.
(2006)
Period of Study:
24 weeks during the
Summer and Fall of 2000
Location:
Steubenville, OH
Health Outcome: Arrhythmia via ECG
measurements
Study Design: Panel
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
53-90 years (mean age 71)
Sample Description:
32 non-smoking older adults
Averaging Time: 24-h
Mean (SD) unit:
0.02 ppm
Range (Min, Max): -01,
1.5
Copollutant: correlation
PM25: r= 0.45
S02: r = 0.62
N02: r = 0.66
03: r = -0.37
Increment: 0 2 ppm
RR Estimate [Lower CI, Upper CI]; 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)
Author: Vedal et al. (2004)
Period of Study:
1997-2000
Location:
Vancouver, Canada
Health Outcome: Cardiac arrythmia via
patients with ICD
Study Design: Panel
Statistical Analyses:
Logistic regression (GEE)
Age Groups Analyzed: Range from
12-77 (mean age 53)
Sample Description:
50 patients who experienced 1 or more
arrhythmia event days during the four
years
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
O3: r = -0.52
Increment: 0 2 ppm
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1, 2, 3
No significant effect for CO (results shown in plots)
CARDIAC ARREST
Author: Levy et al. (2001)
Period of Study:
1988-1994
Location:
Seattle, WA
Health Outcome: Out of hospital primary
cardiac arrest
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
25-75 years
Sample Description:
362 cases
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
Increment: NR
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1
Lag 1 : 0.99 (0.83-1.18)
Author: Sullivan et al.
(2003)
Period of Study:
1985-1994
Location:
Washington State
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 Increment: 1 02 ppm
Mean (SD) unit:
1.92 ppm
Range (Min, Max):
0.52,7.21
Copollutant: NR
OR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1, 2
LagO
Lag 1
Lag 2
0.95 (0.85-1.05)
0.97 (0.87-1.08)
0.99 (0.89-1.11)
March 2009
C-5
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
MYOCARDIAL INFARCTION
Author: Peters et al.
(2001)
Period of Study:
1995- 1996
Location:
Boston, MA
Health Outcome: Onset of Ml:
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
772 participants
Averaging Time: 24-h
Mean (SD) unit: 1 09
Range
(percentiles): ppm
5th = 0.49
95th = 1.78
Copollutant: NR
Increment: 2 H - 1 ppm; 24 h - 0.6 ppm
OR Estimate [Lower CI, Upper CI]
Onset of Ml:
2 h prior: 1.22 (0.89-1.67)
24 h prior: 0.98 (0.70-1.36)
Author: Rosenlund et al.
(2006)
Period of Study:
1992-1994
Location:
Stockholm, Sweden
Health Outcome: Ml
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
45 - 70 years
Sample Description:
1,397 cases, 1,870 controls
Averaging Time:
Mean (SD) unit:
66.8 (jg/m3
(est. 30yr residential
exposure)
Range (percentiles):
5th = 13.9; 95th = 295.7
Copollutant: NR
Increment: 300 |jg/m3
OR Estimate [Lower CI, Upper CI]; lag :
Estimated 30 yr avg. exposure
All cases: 1.04 (0.89-1.21)
Non-fatal cases : 0.98 (0.82-1.16)
Fatal cases : 1.22 (0.98-1.52)
In-hospital death : 1.16 (0.89-1.51)
Out-of-hospital death : 1.36 (1.01-1.84)
CHANGES IN BLOOD PRESSURE
Author:
Health Outcome: BP - SPB
Averaging Time: 24-h
Increment: Lag 0 : 5.6 mg/m3
Ibalde-Mulli et al. (2001)
Study Design: Cohort
Mean (SD) unit:
5-day prior avg.
Period of Study:
1984-1985
Statistical Analyses:
4.1 mg/m3
Mean Change [Lower CI, Upper CI]
Gaussian regression for repeated
Range (Min, Max):
SPB mmHg
Location:
measures
1.7,8.2
Augsburg, Germany
Age Groups Analyzed:
25-64 years
Copollutant: NR
Lag 0 (days):
All: 0.53 (-0.66 to 1.72); Men : 0.68 (-0.94 to 2.31);
Women : 0.51 (-1.31 to 2.19)
Sample Description:
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)
2,607 men & women aged 25-64 years
Author: Zanobetti et al.
(2004a)
Period of Study:
1999-2001
Location:
Boston, MA
Health Outcome: BP
Study Design: Cohort/Panel
Statistical Analyses: Random effects
Age Groups Analyzed: 39 - 90 years
Sample Description:
62 subjects with 631 total visits
Averaging Time:
1-h & 120-h avg
Mean (SD) unit:
Same Hr: 0.81 ppm
120 Hr av: 0.66 ppm
Range (Min, Max):
Same h:
10th = 0.48; 90th = 1.22
120-h av:
10th = 0.48; 90th = 0.86
Copollutant: NR
Increment: NR
RR Estimate [Lower CI, Upper CI]
CO had no significant effect on BP
CHANGES IN BLOOD MARKERS OF COAGULATION AND INFLAMMATION
Author: Baccarelli et al.
(2007)
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:
11-84 years (mean 43years)
Sample Description:
1,218 healthy individuals who were
partners or friends of patients with
thrombosis who attended the thrombosis
center of the University of Milan.
Averaging Time: 1-h
Mean (SD) unit: NR
Range (percentiles):
Sept-Nov:
25th = 1.36; 75th
Dec-Feb:
25th = 2.00; 75th
Mar-May:
25th = 1.03; 75th
Jun-Aug:
25th = 0.73; 75th
Copollutant: NR
3.52
4.31
2.14
1.58
Increment: NR
Regression co-efficient [Lower CI, Upper CI]
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 (-0.04 to 0.11); Lag
30:0.06 (-0.01 to 0.14)
Notes: CO had no effect on fibrinogen, functional
antithrombin, functional protein C, protein C antigen,
functional protein S, free protein S for all lag periods.
Author: Liao et al. (2005)
Period of Study:
1996 to 1998
Location:
Forsyth County, NC;
Selected suburbs of
Minneapolis, MN;
Jackson, Ml
Health Outcome: Various measures of
hemostasis/ inflammation
Study Design: Cohort
Statistical Analyses:
Linear regression
Age Groups Analyzed:
45 - 64 years
Sample Description:
10208 subjects from the Atherosclerosis
Risk in Communities Study
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: 0 6 ppm
Regression coefficients [SE]
Lags examined (days): 1
Lag 1:
Fibrinogen (mg/dL): -0.16 (0.67)
Factor VIII-C (%): 0.45 (0.42)
vWF % : -0.29 (0.50)
WBC(x103/mm3): 0.003 (0.017)
Albumin (g/dL):-0.018 (0.003)**
**p < 0.01
March 2009
C-6
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Pekkanen et al.
(2000)
Period of Study:
1991 - 1993
Location:
London, England
Health Outcome: Fibrinogen
Study Design: Cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed:
35 - 55 years
Sample Description:
7,205 office workers
Averaging Time: 8-h
Mean (SD) unit:
1.4mg/m3
Range (Min, Max):
Min = NR, Max = 9.9
Copollutant correlation:
PM10: r = 0.57
N02: r = 0.81
S02: r = 0.61
O3: 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)
Author: Ruckerl et al.
Health Outcome: Blood markers of
Averaging Time: 24-h
Increment: 0.27 mg/m3
(2006)
inflammation and coagulation
Mean (SD) unit:
OR Estimate for blood marker >90th percentile [Lower
Period of Study:
Study Design: Panel
0.52 mg/m3
CI, Upper CI]
2000-2001
Statistical Analyses:
Range (Min, Max):
Lags examined (h): 0-23, 24-47, 48-71,
Location:
Linear and logistic regression (fixed
0.11,1.93
5-day avg.
Erfert, Germany
effects)
Copollutant correlation:
CRP (C-reactive protein)
Age Groups Analyzed:
N02: r = 0.82
0-23: 0.9 (0.7-1.2); 24-47:1.0 (0.7-1.5);
51-76 years (mean age 66 years)
48-71 :1.5 (1.1-2.1); 5-day avg 1.1 (0.8-1.6)
Sample Description:
57 male patients with CHD
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)
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 years
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/m3
Rome: 1.40 mg/m3
Stockholm: 0.29 mg/m3
Range (Min, Max): NR
Copollutant: NR
Increment: 0.34 mg/m3
% Change in mean [Lower CI, Upper CI]
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)
Author: Steinvil et al.
(2008)
Period of Study:
2003 - 2006
Location:
Tel Aviv, Israel
Health Outcome: Various measures of
inflammation sensitive biomarkers
Study Design: Cohort
Statistical Analyses:
Linear regression
Age Groups Analyzed:
mean age 46 years
Sample Description:
3659 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
NO2: r = 0.857
S02: r = 0.671
O3: r= -0.656
Increment: 0 3 ppm
Regression co-efficient [Lower CI, Upper CI]
Lags examined (days): 0,1, 2, 3, 4, 5, 6, 7, last week avg
Fibrinogen - Men
Lag 0 : -3.3 (-6.1 to -0.6); Lag 1 : -2.6 (-5.5 to 0.4);
Lag 2 : -3.4 (-6.6 to -0.3); Lag 3 : -3.4 (-6.5 to -0.2);
Lag 4 : -5.9 (-8.9 to -2.9); Lag 5 : -4.7 (-7.8 to -1.6);
Lag 6 : -2.0 (-5.1 to 1.0); Lag 7 : -2.7 (-5.7 to 0.2);
Last week avg: -7.7 (-12.1 to -3.3)
Notes: No effect on fibrinogen among women. CO had no
effect on CRP among men and no effect on CRP and WBC
among women for all Lag times examined.
March 2009
C-7
DRAFT-DO NOT QUOTE OR CITE
-------
Table C-2. Studies of CO exposure and cardiovascular hospital admissions and ED visits.
Study
Design
Concentrations
CO Effect Estimates (95% CI)
STROKE
Author: Chan et al. (2006)
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
NO2: r = 0.77
PM25: r = 0.44
PM10: r= 0.47
Increment: 0 8 ppm
OR Estimate [Lower CI, Upper CI]
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)
Period of Study: 1994 - 2004
Location:
Dijon, France
Health Outcome (ICD9 or ICD10):
Stroke (Ischemic & Hemorrhagic)
Study Design: Bi-directional Case-
crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: a 40
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 683 (jg/m3
Range (Min, Max): 0, 4014
Copollutant: NR
Increment: 10 |jg/m3
OR Estimate [Lower CI, Upper CI]
Lags (days) examined: 0,1,2, 3.
Ischemic:
Lag 0
Lag 1
Lag 2
Lag 3
0.999 (0.997-1.001)
0.998 (0.997-1.001)
0.999 (0.998-1.001)
1.000 (0.998-1.001)
Hemorrhagic:
Lag 0
Lag 1
Lag 2
Lag 3
1.000 (0.996-1.004)
1.001 (0.997-1.005)
0.999 (0.995-1.004)
0.998 (0.994-1.002)
Also not significant when stratified by sex.
Author: Maheswaran et al.
Health Outcome (ICD9 or ICD10):
Averaging Time: NR
Increment:
(2005b)
Stroke deaths (ICD9: 430-438); Stroke
Mean (SD) unit: Quintiles
NR - Quintiles of exposure
Period of Study: 1994- 1998
Hospital admissions (ICD10:160-I69)
Range (Min, Max): NR
RR Estimate [Lower CI, Upper CI]
Location:
Study Design: Ecological
Copollutant: NR
Adjusted for sex, age, deprevation, smoking.
Sheffield, UK
Statistical Analyses:
Poisson regression
Age Groups Analyzed: a 45 years
Sample Description:
1,030 census districts
Quintiles:
2nd: 1.04 (0.94-1.16)
3rd: 1.01 (0.91-1.13)
4th : 1.10 (0.99-1.23)
5th : 1.11 (0.99-1.25)
Adjusted for sex, age:
2nd: 1.11 (1.01-1.22)
3rd: 1.15 (1.04-1.27)
4th : 1.29 (1.17-1.42)
5th : 1.37 (1.24-1.52)
Author: Tsai et al. (2003b)
Period of Study:
1997-2000
Location:
Kaohsiung, Taiwan
Study Design: Case-crossover
Health Outcome (ICD9 or ICD10):
Cerebrovascular diseases: ICD9: 430 to
438 (Subarachnoid hemorrhagic stroke
430, Primary intracerebral hemorrhage
(PIH): 431-432, Ischemic stroke (IS): 433-
435).
Statistical Analyses: NR
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 0 79 ppm
Range (Min, Max): 0 24,1 72
Copollutant: NR
Increment: 0.8 ppm (IQR)
RR Estimate [Lower CI, Upper CI]
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 1.18 (0.80-0.72)
IS : OR 1.77 (1.31-2.39)
Notes:
2-pollutant models:
PIH results persisted when adjusting for SO2
and O3
IS results persisted when controlling for PM10,
SO2 and O3
March 2009
C-8
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Villeneuve et al.
(2006a)
Period of Study: 1992-2002
Location:
Edmonton, Canada
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+
Sample Description:
12,422 visits
Averaging Time: 24-h
Mean (SD) unit: 0 8 ppm
Range (percentiles):
25th = 0.5; 75th = 1.0
Copollutant correlation :
O3: r= -0.54
PM2.5: r = 0.43
PM10: r = 0.30
Increment: 0 5 ppm
OR Estimate [Lower CI, Upper CI]
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.
Author: Wellenius et al. (2005a)
Period of Study: NR
Location:
9 U.S. cities: Chicago, Detroit,
Pittsburgh, Cleveland,
Birmingham, New Haven,
Seattle, Minneapolis, Salt Lake
City
ED Visits
Health Outcome:
Stroke among Medicare beneficiaries:
(Ischemic, hemorrhagic)
Study Design: Time-series
Statistical Analyses:
Logistic regression
Age Groups Analyzed: a 65 years
Sample Description: 155,503 visits
Averaging Time: NR
Mean (SD) unit: NR
Range (percentiles):
25th = 0.73; 50th = 1.02;
75th = 1.44 (ppm)
Copollutant: correlation
PM10: r= 0.43
Increment: 0 71 ppm
% Change [Lower CI, Upper CI]
Lag : 0
Ischemic: 2.83 (1.23-4.46)
Hemorrhagic : -1.61 (-4.79 to 1.68)
ISCHEMIC HEART DISEASE
Author: D'lppoliti et al. (2003)
Period of Study: 1995 - 1997
Location:
Rome, Italy
Hospital Admissions
Health Outcome (ICD9): Ml (410)
Study Design: Case-crossover
Statistical Analyses: Conditional logistic
regression
Age Groups Analyzed: 18+
Sample Description:
6,531 patients.
Averaging Time: 24-h
Mean (SD) unit: 4 4 mg/m3
Range (percentiles):
25th = 2.8; 75th = 4.3
Copollutant: correlation
TSP: r = 0.35
S02: r= 0.56
N02: r = 0.31
Increment: 1 mg/m3
OR Estimate [Lower CI, Upper CI]; 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 3:1.010(0.982-1.040)
Lag 4:1.025 (0.996-1.055)
Lag 0-2:1.044(1.000- .089
Author: Hosseinpoor et al.
(2005)
Period of Study:
1996-2001
Location:
Tehran, Iran
Health Outcome:
Angina Pectoris (ICD9: 413; ICD10:120)
Study Design: Time-series
Statistical Analyses:
Poisson regression
Age Groups Analyzed: All
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 10.8 mg/m3
Range (Min, Max): 16, 57 8
Copollutant: NR
Increment: 1 mg/m3
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1,2,3
Lag 1 :1.00957 (1.00600-1.01315)
March 2009
C-9
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Lanki et al. (2006)
Period of Study: 1994 - 2000
Location:
5 European cities:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Health Outcome:
First AMI (ICD9: 410; ICD10:121,122)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed:
35+ years
Sample Description:
26,854 Hospital Admissions
Averaging Time: 24-h
Mean (SD) unit: NR
Unit: mg/m3
Range (percentiles):
Augsburg, Germany
25th = 0.7; 75th = 1.1
Barcelona, Spain
25th = 0.6; 75th = 1.4
Helsinki, Finland
25th = 0.3; 75th = 0.5
Rome, Italy
25th = 1.7; 75th = 2.9
Stockholm, Sweden
25th = 0.3; 75th = 0.5
Copollutant: correlation
PMio:r= 0.21 -0.56
N02:r = 0.43-0.75
03:r = -.023-020
Increment: 0.2 mg/m3
RR Estimate [Lower CI, Upper CI]; lag :
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(1.001-1.012)
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 - s 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 3: 0.999 (0.992-1.006)
3 cities with HDR - a 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
Lag 1
Lag 2
Lag 3
1.015(1.004-1.086)
1.006 (0.995-1.017)
0.995 (0.983-1.006)
0.998 (0.987-1.009)
Author: Lee et al. (2003b)
Study Design: Time-series
Averaging Time:
Period of Study: 1997 - 1999
Health Outcome (ICD9 or ICD10):
Daily max
Location:
Angina: ICD10:120
Mean (SD) unit:
Seoul, Korea
AMI: ICD10:121-123
1.8 ppm
Other Acute IHDs: ICD10:124
Range (percentiles):
Statistical Analyses:
25th = 1.2
Poisson regression, GAM
75th = 2.2
Age Groups Analyzed: 64 +
Copollutant: correlation
Sample Description: 822 days
PM20: 0.60
S02: 0.81
N02: 0.79
03: -0.39
Author: Maheswaran et al.
Emergency Hospital Admission
Averaging Time: NR
(2005b)
Health Outcome (ICD9):
Mean (SD) unit: Quintiles
Period of Study:
CHD (410-414)
Range (Min, Max): NR
1994-1998
Study Design: Ecological
Copollutant: NR
Location:
Statistical Analyses: Poisson regression
Sheffield, UK
Age Groups Analyzed: 45+ years
Sample Description:
11,407 Emergency Hospital Admissions
for CHD in patients 45+ years (within
1,030 census districts)
Increment: 1 ppm (IQR)
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1, 2, 3, 4,
All year:
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)
5,6
Increment: NA
RR Estimate [Lower CI, Upper CI]
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)
March 2009
C-10
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Mann et al. (2002)
Period of Study: 1988 - 1995
Location:
Southern California
Health Outcome (ICD9): IHD (IHD) (410-
414); Myocardial Infarction (Ml) (410)
Study Design: Time-series
Statistical Analyses:
Poisson regression, GAM
Age Groups Analyzed: All
Sample Description:
54,863 IHD admissions among Southern
California Kaiser- Permanente members
(within 20km of monitor)
Averaging Time: 8-h
Mean (SD) unit: 2 07 ppm
Range (Min, Max): 0 30,11.1
Copollutant: correlation
Ranging across 7 regions:
N02: r = 0.64, 0.86
03:r = -0.37, 0.28
PMio:r= 0.15, 0.40
Increment: 1 ppm
% Change [Lower CI, Upper CI]
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)
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: 4979 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 CI, Upper CI]; 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 a 64
All Patients: Lag 0 : 4.9 (1.3-8.7)
Males: Lag 0 : 7.5 (2.6-12.6)
Females: Lag 0 : 2.4 (-3.0 to.0)
Lag 1 not significant for all results
Author: von Klot et al. (2005)
Period of Study: 1992 - 2001
Location:
5 European cities:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Health Outcome: Hospital Cardiac
(Myocardial Infacrtion (Ml), Angina,
Dysrythmia, Heart Failure) Re-admissions
Study Design:
Prospective Cohort
Statistical Analyses: Poisson regression
Age Groups Analyzed: All
Sample Description:22006 survivors of
first Ml
Averaging Time: 24-h
Unit: mg/m3
Mean (SD) unit:
Augsburg, Germany: 0.93
Barcelona, Spain: 1.00
Helsinki, Finland: 0.42
Rome, Italy: 2.21
Stockholm, Sweden: 0.43
Range (Min, Max): NR
Copollutant: correlation
PM10: r= 0.21 -0.57
NO2: r = 0.44 — 0.75
03: r = -.027-0.47
Increment: 0.2 mg/m3 (0.172 ppm)
RR Estimate [Lower CI, Upper CI]
Lags examined (days): 0,1,2,3
Lag 0:
Ml: 1.022 (0.998-.047)
Angina: 1.009 (0.992-.02)
Cardiac : 1.014 (1.001-.026)
HEART FAILURE
Author: Lee et al. (2007a)
Period of Study: 1996 - 2004
Location:
Kaohsiung City, Taiwan
Hospital Admissions
Health Outcome (ICD9): CHF (428)
Study Design: Case-crossover
Statistical Analyses: Conditional logistic
regression
Age Groups Analyzed: All
Sample Description:
13475 Hospital Admissions (63 Hospitals)
Averaging Time: 24-h
Mean (SD) unit: 0.76 ppm
Range (Min, Max): 014,1 72
Copollutant: NR
Increment: 0 31 ppm
OR Estimate [Lower CI, Upper CI]
Lag examined (days): 0-2
>25°C: 1.19(1.09-1.31)
<25°C: 1.39 (1.24-1.54)
Adjusted for PM10:
>25°C; 1.15 (1.04-1.27)
<25°C; 1.21 (1.206-1.38)
Adjusted for SO2:
>25°C: 1.23 (1.11-1.36)
<25°C: 1.39 (1.24-1.55)
Adjusted for NO2:
>25°C: 1.22 (1.08-1.39)
<25°C: 0.94 (0.81-1.10)
Adjusted for O3:
>25°C: 1.17 (1.07-1.28)
<25°C: 1.36 (1.22-1.51)
March 2009
C-11
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Symons et al. (2006)
Period of Study: 2002 (April -
November)
Location:
Johns Hopkins Bayview Medical
Center, Baltimore, MD
Hospital Admissions
Health Outcome: NR
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
Sample Description:
398 Hospital Admissions for CHF
Averaging Time: 24-h
Mean (SD) unit: 0 4 ppm
Range (Min, Max): 01,10
Copollutant: NR
Increment: 0 2 ppm
OR Estimate [Lower CI, Upper CI]
Lags examined (days):
0,1, 2, 3, cum 1, cum 2, cum 3
Lag 0
Lag 1
Lag 2
Lag 3
Cum. Lag1
Cum. Lag2
Cum. Lag3
0.86 (0.67-1.11)
0.90 (0.70-1.17)
0.96 (0.73-1.26)
0.88 (0.67-1.16)
' 0.82 (0.60-1.13)
0.80 (0.54-1.17)
0.27 (0.46-1.14)
Author: Wellenius et al. (2005b)
Period of Study: 1987 - 1999
Location:
Pennsylvania, PA
Hospital Admissions
Health Outcome (ICD9): CHF
(428, 428.1)
Study Design: Case-crossover
Statistical Analyses: Conditional logistic
regression
Age Groups Analyzed: 65+
Sample Description:
54019 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
O3: r= -0.25
S02: r= 0.54
Increment: 0 55 ppm
% Change [Lower CI, Upper CI]
Lags examined (days): 0,1,2,3
Lag 0:
Single pollutant model: 4.55 (3.33-5.79)
Adjusted for PM10: 5.18 (3.49-6.89)
Adjusted for NO2: 4.84 (3.06-6.66)
Adjusted for O3: 4.35 (3.08-5.64)
Adjusted for SO2: 4.51 (3.15-5.90)
CARDIOVASCULAR DISEASES - NON-SPECIFIC
Author: Ballester et al. (2001)
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 CI, Upper CI]; 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)
Period of Study: 1995 - 1999
Location: 14 Cities in Spain
Health Outcome (ICD9:
All CVD (390-459);Heart Diseases (410-
414,427, 428)
Study Design: Time-series
Statistical Analyses: GAM
Age Groups Analyzed: All
Sample Description: NR
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
Increment: 1 mg/m3
% Change [Lower CI, Upper CI]
Lags examined (days): 0-1
All CVD: Lag 0-1 : 2.06 (0.65-3.48)
Heart Disease: Lag 0-1 : 4.15 (1.31-7X
Author: Barnett et al. (2006)
Period of Study: 1998 - 2001
Location:
Brisbane, Canberra, Melbourne,
Perth, Sydney Australia
Auckland & Christchurch, New
Zealand
Hospital Admissions with
Cardiovascular Diseases
Health Outcome (ICD9: Arrythmia (247);
Cardiac Disease (390 - 429); Cardiac
Failure (428); IHD (410-413); Ml (410);
Total CVD (390-459)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
15-64 years & a 65 years
Sample Description: NR
Averaging Time: 8-h
Mean (SD) unit: ppm
Brisbane: 1.7
Canberra: 0.9
Melbourne: 1.0
Perth: 1.0
Sydney: 0.8
Auckland: 2.1
Christchurch: 0.5
Range (Min, Max): ppm
Brisbane: 0.0, 7.0
Canberra: 0.0, 5.8
Melbourne: 0.1, 8.0
Perth: 0.1, 4.0
Sydney: 0.0, 4.5
Auckland: 0.2, 7.9
Christchurch: 0.0, 5.4
Copollutant: NR
Increment: 0 9 ppm
% Change [Lower CI, Upper CI]
Lags examined (days): 0-1
15-64 years
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)
Ml: 1.8 (-0.7 to 4.3)
Total CVD: 1.2(0.3-2.1)
> 65 years
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)
March 2009
C-12
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Chang et al. (2005)
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
Age Groups Analyzed: All
Sample Description:
74,509 CVD hospital admissions
(47 Hospitals)
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 CI, Upper CI]
Lag examined (days): 0-2
>20°C: 1.090 (1.064-1.118)
<20°C: 0.984 (0.927-1.044)
Adjusted for PM10:
>20°C: 1.171 (1.132-1.211)
<20°C: 0.946 (0.892-1.003)
Adjusted for SO2:
>20°C: 1.232 (1.194-1.272)
<20°C: 1.098 (1.034-1.165)
Adjusted for NO2:
>20°C: 1.048 (1.003-1.095
<20°C: 0.983 (0.914-1.058)
Adjusted for O3
>20°C: 1.196 (1.161-1.232)
<20°C: 1.092 (1.031-1.157)
Author: Fung et al. (2005)
Period of Study: 1995 - 2000
Location:
Windsor, Ontario,
Canada
Hospital Admissions of
Cardiovascular Diseases
Health Outcome (ICD9): CHF (428);
IHD (410 - 414); Dysrythmias (427)
Study Design: Time-series
Statistical Analyses: GLM
Age Groups Analyzed: All
Sample Description:
11,632 Cardiac hospital admissions
Averaging Time: 24-h
Mean (SD) unit: 1 3 ppm
Range (Min, Max): 0 0,118
Copollutant: correlation
PM10: r= 0.21
N02: r = 0.38
S02: r= 0.16
03: r = 0.10
Increment: 1 2 ppm
% Change [Lower CI, Upper CI]
Lags examined (days): 0,0-1, 0-2
<65 years
Lag 0 : -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)
> 65 years
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)
Author: Jalaludin et al. (2006) ED Visits
Period of Study:
1997-2001
Location:
Sydney, Australia
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+ years
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 CI, Upper CI]
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.
All Cardiovascular Disease:
Cool period: Lag 0 : 3.26 (2.00-4.53)
Cardiac Disease:
Cool period: Lag 0 : 3.43 (1.95-4.93)
IHD:
Cool period: Lag 0 : 3.64 (1.28-6.06)
Warm period: Lag 0 : 2.29 (0.01-4.62)
Stroke:
Cool period: Lag 0 : 3.54 (0.78-6.37)
Notes:
Cool: May to October
Warm : November to April
Author: Koken et al. (2003)
Period of Study: 1993 - 1997
Location:
Denver, CO
Hospital Admissions for
Cardiovascular Disease
Health Outcome (ICD9: Ml (410-410 92);
Coronary Atherosclerosis (414-414.05);
Pulmonary Heart Disease (416 - 416.9);
Cardiac Dysrythmia (427 - 427.9);
CHF (428)
Study Design: Time-series
Statistical Analyses: GLM
Age Groups Analyzed: >65 years
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit: 0 9 ppm
Range (Min, Max): 0 3,16
Copollutant: correlation
PM10: r= 0.25
N02: r = 0.73
SO2: r = 0.21
O3: r= -0.40
Increment: 0 3 ppm
% Change [Lower CI, Upper CI]
Lags examined (days): 1, 2, 3, 4
CHF: Lag 3:10.5 (0.1-22.0)
CO not significantly associated with other Lag
periods.
March 2009
C-13
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Linn et al. (2000)
Period of Study: 1992 - 1995
Location:
Los Angeles, CA
Health Outcome:
Hospital Admissions for Cardiovascular,
Cerebrovascular, Pulmonary.
Study Design: Time-series
Statistical Analyses:
Ordinary least squares regression;
Poisson regression
Age Groups Analyzed:
>30 years
Sample Description: NR
Averaging Time: 24-h
Mean (SD) unit:
Winter: 1.7; Spring: 1.0;
Summer: 1.2; Fall: 2.1
Range (Min, Max):
Winter: 0.5, 5.3; Spring: 0.4, 2.2;
Summer: 0.3, 2.7; Fall: 0.2, 4.3
Copollutant: correlation
Winter:
PM10: r = 0.78; NO2: r = 0.89;
03: r= -0.43;
Spring:
PM10: r = 0.54; NO2:r = 0.92;
03: 0.29
Summer:
PM10: r = 0.72; NO2: r = 0.94;
03: 0.03
Fall:
PM10: r = 0.58; NO2:r = 0.84;
03: r= -0.36
Increment: 1 ppm
Co-efficient [SE]
Lags examined (lags): 0,1
LagO:
Cardiovascular
All: 0.032 (0.003)* (e.g. 3.2% increase)
Winter: 0.038 (0.006)*
Spring: 0.010 (0.015)
Summer: 0.035 (0.014)*
Fall: 0.027 (0.006)*
Cerebrovascular
All: 0.009 (0.007)
Winter: -0.008 (0.014)
Spring: 0.107 (0.033)*
Summer: 0.030 (0.033)
Fall: 0.008 (0.012)
Myocardial Infarction
All: 0.040 (0.009) *
CHF
All: 0.025 (0.009)*
Cardiac Arrythmia
All: 0.023 (0.009)*
Stroke
All: 0.044 (0.009)*
Notes:* p < 0 05
Author: Metzger et al. (2004a)
Period of Study: 1993 - 2000
Location:
Atlanta, GA
ED Visits (from 31 hospitals)
Health Outcome (ICD9: Cardiovascular:
IHD (410-414); Acute Ml (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
SO2: r =0.26
03: r =0.20
Increment: 1 ppm
RR Estimate [Lower CI, Upper CI]
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)
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)
Study Design: Case-crossover
Statistical Analyses: Conditional logistic
regression
Age Groups Analyzed: All
Sample Description: 4,407,535 visits
Averaging Time: 1-h
Mean (SD) unit: 1 8 ppm
Range (SD): SD: 12
Copollutant: NR
Increment: 1 2 ppm
OR Estimate [Lower CI, Upper CI]
Lags examined (days): 0-2ma
IHD:
Without Diabetes : 1.023 (1.004-1.420)
Without CHF: 1.024(1.006-1.042)
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)
Period of Study: 1995 - 2001
Location:
Spokane, WA
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
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
Increment: NR
RR Estimate [Lower CI, Upper CI]; lag :
Lags examined (days): 1, 2, 3
No significant association. Results not
reported.
March 2009
C-14
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Yang et al. (2004a)
Period of Study: 1997 - 2000
Location:
Kaohsiung City, Taiwan
Health Outcome (ICD9:
Cardiovascular diseases (410-429)
Study Design: Case-crossover
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 CI, Upper CI]
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 SO2:
>25°C: 1.406 (1.327-1.489)
<25°C: 1.3450 (1.352-1.555)
Adjusted for NO2:
>25°C: 1.246 (1.166-1.332)
<25°C: 0.905 (0.819-0.999)
Adjusted for O3:
>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% CI)
Author: Bell et al Health Outcome:
(2007)
Period of Study:
1999-2002
Location:
Connecticut and
Massachusetts
Birth weight and LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic regression
Age Groups Analyzed: NA
Sample Description:
358,504 full term live singleton
births (32-44 weeks)
Averaging Time: 24-h
Mean (SD) unit:
0.65 ppm (0.18)
Range (Min, Max): NR
Copollutant: NR
Increment: Interquartile range - 0.30 ppm
Regression co-efficient for birth weight (g) [Lower CI, Upper CI]
Entire pregnancy: -16.2 (-19.7 to -12.6)
Stratified by race.
Black mother: -10.9 (-20.2 to -1.6)
White mother: -17.5 (-21.3 to -13.7)
OR for LBW [Lower CI, Upper CI]
Entire pregnancy : 1.028 (0.983-1.074)
Author: Braueret
al. (2008)
Period of Study:
1999-2004
Location:
Vancouver, Canada
Health Outcome: LBW, PTB
and SGA
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
70249 live singleton births
Averaging Time:
Land use regression model
Mean (SD) unit: 633 |jg/m3
Range (Min, Max):
124,1409
Copollutant: correlation:
PM10: r = 0.73
NO2: r = 0.75
S02: r = 0.82
03: r = -0.39
Increment: 100 |jg/m3
OR for SGA [Lower CI, Upper CI];
Entire pregnancy : 1.06 (1.03-1.08)
OR for term LBW [Lower CI, Upper CI];
Entire pregnancy : 1.02 (0.96-1.09)
OR PTB [Lower CI, Upper CI];
Entire pregnancy : 1.16 (1.01-1.33)
Author: Chen et al.
(2002)
Period of Study:
1991 - 1999
Location:
Northern Nevada
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 weeks)
Averaging Time: 8-h
Mean (SD) unit: 0 98 ppm
Range (Min, Max):
0.25, 4.87
Copollutant: NR
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
Author: Conceicao
et al. (2001)
Period of Study:
1994-1997
Location:
Sao Paulo, Brazil
Health Outcome: Child
mortality, under 5 years of age
Study Design: Time-series
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 -
[SE];
Lags examined : 0,1, 2, 3
Lag 2: 0.0306 (0.0076) (p < 0.01)
Lag chosen for best fitting model
under 5 years of age
March 2009
C-15
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Gilboa et
al. (2005)
Period of Study:
1997-2000
Location:
Texas
Health Outcome:
Birth defects (heart defects &
orofacial clefts)
Study Design: Case-control
Statistical Analyses:
Conditional Logistic regression
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: Exposure categories (ppm):<0.4; 0.4-0.5; 0.5 - 0.7; >0.7
OR for Birth Defects [Lower CI, Upper CI];
Exposure period : weeks 3 - 8 of pregnancy
Conotruncal defects:
1.00; 1.38 (0.97-1.97); 1.17 (0.81-1.70); 1.46 (1.03-2.08)
Tetralogy of Fallot:
1.00; 0.92 (0.52-1.62); 1.27 (0.75-2.14); 2.04 (1.26-3.29)
Notes: CO was not associated with the following defects: Aortic artery S
valve, atrial septal, pulmonary artery & valve, ventricular septal,
endocardial cushion & mitral valve , cleft lip, cleft palate, aortic valve
stenosis, coarctation of the aorta, ostium secundum.
Author: Gouveia et
al. (2004)
Period of Study:
1997
Location:
Sao Paulo, Brazil
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 weeks)
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 CI, Upper CI]
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 CI, Upper CI]
4th quartile exposure (compared to lowest quartile):
First: 1.02 (0.82-1.27); Second : 1.07 (0.88-1.30);
Third: 0.93 (0.76-1.12)
Author: Ha et al
(2001)
Period of Study:
1996- 1997
Location:
Seoul, South Korea
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 weeks)
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
Percentiles:
25th: 0.99 ppm
75th: 1.41 ppm
Copollutant correlation: TSP:
r = 0.73
NO2: r = 0.75
S02: r = 0.82
03: r = -0.39
Increment: 0 42 ppm
RR for LBW [Lower CI, Upper CI]
Trimesters:
First: 1.08 (1.04,1.12)
Third : 0.91 (0.87, 0.96)
Author: Ha et al
(2003)
Period of Study:
1995- 1999
Location:
Seoul, South Korea
Health Outcome:
Post-neonatal mortality
(1 month - 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: 24-h
Mean (SD) unit: 12 ppm
Range (Min, Max):
0.39, 3.38
Copollutant correlation:
PM10: r = 0.63
N02: r = 0.72
SO2: r = 0.75
O3: r = -0.46
Increment: 0 57 ppm
RRfor Post-neonatal mortality (1 month -1 yr)
[Lower CI, Upper CI]
Lags examined : 0
Total mortality:
Lag 0:1.020 (0.976-1.067)
Respiratory mortality:
Lag 0:1.388 (1.009-1.911)
Author: Huynh et
al. (2006)
Period of Study:
1999-2000
Location:
California
Health Outcome:
PTB (24-36 weeks gestation)
Study Design: Case-control
Statistical Analyses:
Conditional Logistic regression
Age Groups Analyzed:
Cases = 24-36 weeks gestation;
Controls = 39-44 weeks
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 month and last two
weeks 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 weeks of pregnancy were
similar.
OR for PTB [Lower CI, Upper CI]
First month of gestation:
Per 1 ppm increase : 1.10 (0.99-1.20)
Second quartile : 0.94 (0.88-1.01)
Third quartile : 1.04 (0.97-1.11)
Fourth quartile : 1.05 (0.96-1.14)
Last two weeks of gestation:
Per 1 ppm increase : 1.00 (0.93-1.09)
Second quartile : 1.03 (0.97-1.10)
Third quartile : 1.04 (0.97-1.12)
Fourth quartile : 0.99 (0.91-1.08)
Entire pregnancy:
Per 1 ppm increase : 1.06 (0.95-1.18)
Second quartile : 0.97 (0.91-1.04)
Third quartile : 0.99 (0.92-1.05)
Fourth quartile : 1.02 (0.94-1.09)
Lowest quartile used as reference group
March 2009
C-16
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Hwang
and Jaakkola
(2008)
Period of Study:
2001-2003
Location: Taiwan
Health Outcome: Oral clefts
(with or without palate)
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
6,530 cases from 721,289
newborns
Averaging Time: 8-h
Mean (SD) unit: 0.69 (0.4)
Range (Min, Max): 0.25, 2 7
Copollutant correlation:
PM10: r = -0.19
NOx: r = 0.82
S02: r = 0.24
03: r = -0.19
Increment: 100 ppb
RR for oral cleft [Lower CI, Upper CI]
Month 1
Month 2
Month 3
1.00 (0.96-1.04)
1.00 (0.96-1.03)
1.00 (0.96-1.03)
Author: Jalaludin
et al. (2007)
Period of Study:
1998-2000
Location:
Sydney, Australia
Health Outcome: PTB
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
123,840 full term live singleton
births (<42 weeks)
Averaging Time: 8-h
Mean (SD) unit:
0.9 ppm (0.68)
Range (Min, Max): NR
Copollutant correlation:
PM10: r = 0.28
N02: r = 0.60
S02: r = 0.24
03: r = -0.21
Increment: 1 ppm
RR for PTB [Lower CI, Upper CI]
First month:
All of Sydney : 0.89 (0.84-0.95)
Within 5km of site : 1.03 (0.68-1.54)
First trimester:
All of Sydney : 0.77 (0.71-0.83)
Within 5km of site : 1.24 (0.81-1.91)
1 month prior to birth:
All of Sydney: 0.96 (0.88-1.04)
Within 5km of site : 1.00 (0.86-1.15)
3 months prior to birth:
All of Sydney: 0.99 (0.90-1.09)
Within 5km ofsite : 1.11 (0.94-1.31)
Author: Lee et al
(2003a)
Period of Study:
1996- 1998
Location:
Seoul, South Korea
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
388,105 full term live singleton
births (37-44 weeks)
Averaging Time: 24-h
Mean (SD) unit: 12 ppm
Range (Min, Max): 0.4, 3 4
Copollutant correlation:
PM10: r = 0.47
N02: r = 0.77
S02: r = 0.79
Increment: 0 5 ppm
OR for LBW [Lower CI, Upper CI]
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)
Author: Leem et
Health Outcome:
Averaging Time:
Increment:
al. (2006)
PTB
Kriging was used to estimate
Exposure level - Quartiles of exposure for first trimester (mg/m3)
Period of Study:
Study Design:
exposure
First: 0.47-0.63; Second : 0.6 -0.77;
2001 - 2002
Retrospective cohort
Mean (SD) unit: NR
Third : 0.78-0.90; Fourth : 0.91-1.27
Location:
Statistical Analyses:
Range (Min, Max): NR
- exposure groups for third trimester was similar
Incheon, Korea
Logistic regression
Copollutant correlation:
OR for PTB [Lower CI, Upper CI]
Age Groups Analyzed: NA
PM10: r = 0.27
First Trimester:
Sample Description:
52,113 live singleton births
N02: r = 0.63
S02: r = 0.31
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.
Health Outcome:
Averaging Time: 24-h
Increment: NR
(2004a)
Period of Study:
1998-2000
Location:
Sao Paulo, Brazil
Neonatal death (within first
28 days of life)
Study Design: Time-series
Statistical Analyses:
Poisson regression (GAM)
Age Groups Analyzed: NA
Sample Description: NR
Mean (SD) unit: 2.83 ppm
Range (Min, Max):
0.54,10.25
Copollutant correlation:
PM10: r = 0.71
N02: r = 0.67
SO2: r = 0.55
03: r= 0.03
Regression coefficent for neonatal death [SE]
Lags examined : 0
Lag 0:0.0061 (0.0110)
March 2009
C-17
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Lin et al.
(2004b)
Period of Study:
1995- 1997
Location:
Taipei &
Kaoshiung, Taiwan
Health Outcome:
LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
92288 full term live singleton
births (>37 weeks) within 3km of
monitoring site.
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: Exposure groups
M = Median exposure 1.1-14.2 ppm
H = High exposure >14.2 ppm
OR for LBW [Lower CI, Upper CI]
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.
Author: Liu et al.
(2003)
Period of Study:
1985- 1998
Location:
Vancouver,
BC,Canada
Health Outcome:
PTB, IUGR, LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
229,085 live singleton births
Averaging Time: 24-h
Mean (SD) unit: 10 ppm
Range (Min, Max):
25th: 0.7; 75th: 1.2
Copollutant: NR
Increment: 10 ppm
OR for LBW [Lower CI, Upper CI]
Month of pregnancy:
First month: 1.01 (0.93-1.09)
Last month: 0.96 (0.88-1.04)
OR for PTB [Lower CI, Upper CI]
First month : 0.95 (0.89-1.01)
Last month : 1.08 (1.01-1.15)
OR for IUGR [Lower CI- Upper CI]
First month : 1.06 (1.01-1.10)
Last month : 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)
Period of Study:
1995-2000
Location:
Calgary, Edmonton,
and Montreal
Canada
Health Outcome: IUGR
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
386,202 live singleton births
Averaging Time: 24-h
Mean (SD) unit: 11 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
Increment: 1 ppm
RR for LBW [Lower CI, Upper CI]
Notes: CO was associated with an increased risk of IUGR of
approximately 16% and 23% in the first and nine month of pregnancy.
(All results presented in Figures)
Author: Maisonet
et al. (2001)
Period of Study:
1994-1996
Location:
Northeastern USA
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 weeks)
Averaging Time: 24-h
Mean (SD) unit: NR
Range (Min, Max):
Pprppnt||pc;•
25th: 0.93 ppm; 75th: 1.23 ppm
Copollutant: NR
Increment: 1 ppm
OR for LBW [Lower CI, Upper CI]
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 (1.18-1.74); Second : 1.27 (0.87-1.86);
Third : 1.75 (1.50-2.04)
Notes: CO had no effect on whites or Hispanics
March 2009
C-18
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Mannes
et al. (2005)
Period of Study:
1998-2000
Location:
Sydney, Australia
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 weeks gestation)
Averaging Time: 8-h
Mean (SD) unit: 0 8 ppm
Range (Min, Max): 0 0, 46
Copollutant: correlation
PM10: r = 0.26
NO2: r = 0.57
Os: r = -0.20
Increment: 1 ppm
Regression coefficients for birth weight (g) [Lower CI, Upper CI]
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 month prior to birth : -15.28 (-25.59 to -4.97)
Births within 5 km of monitor:
First trimester: -8.56 (-28.60 to 10.68)
Second trimester: -28.87 (-50.98 to -6.76)
Third trimester: -22.88 (-44.58 to -1.18)
One month prior to birth : -10.41 (-30.03 to 9.21)
OR for SGA [Lower CI, Upper CI]
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 month 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 month prior to birth : 1.10 (0.96-1.27)
Author: Medeiros
et al. (2005)
Period of Study:
1998-2000
Location:
Sao Paulo, Brazil
Health Outcome: Birth weight
and LBW
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic regression
Age Groups Analyzed: NA
Sample Description:
311,735 full term live singleton
births (37-41 weeks)
Averaging Time: 24-h
Mean (SD) unit:
Daily mean shown in Figure
(see paper)
Range (Min, Max): NR
Copollutant: NR
Increment: 1 ppm
Regression co-efficient for birth weight (g) [Lower CI, Upper CI]
Trimesters:
First: -11.9 (-15.5 to -8.2); Second : 4.9 (0.5-9.3);
Third : 12.1 (7.6-16.6)
OR for LBW [Lower CI, Upper CI]
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: Parker et
al. (2005)
Period of Study:
2000
Location:
California
Health Outcome:
Birth weight & SGA
Study Design:
Retrospective cohort
Statistical Analyses:
Linear and logistic regression
Age Groups Analyzed: NA
Sample Description:
18,247 full term live singleton
births (40 weeks) within 5 miles
of a monitor
Averaging Time: 24-h
Mean (SD) unit: 0 78 ppm
Range (Min, Max): NR
Copollutant: NR
Increment: Quartiles of exposure for first trimester
First: <0.57; Second : 0.57-0.76 ;
Third : 0.76- 0.93; Fourth : >0.93
- exposure groups for other trimesters were similar
Regression co-efficient for birth weight (g) [Lower CI, Upper CI]
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 CI, Upper CI]
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)
Author: Ritz et al.
Health Outcome: PTB
Averaging Time:
Increment: 3 ppm
(2000)
Study Design:
6-9 a.m.
RR for PTB [Lower CI, Upper CI]
Period of Study:
1989- 1993
Retrospective Cohort
Mean (SD) unit: 2 70 ppm
Adjusted for various risk factors and season of birth and
Statistical Analyses:
Range (Min, Max):
conception
Location:
Logistic regression
0.36,9.12
6 weeks prior to birth : 1.04 (0.99-1.10)
Southern California
Age Groups Analyzed:
Copollutant correlation:
1 st month of pregnancy : 1.04 (0.99-1.09)
Eligible study subjects were
PM10: r = 0.37
Adjusted for various risk factors
singletons born at 26-44 weeks
N02: r = 0.60
6 weeks prior to birth : 1.06 (1.02-1.10)
gestation
03: r = -0.44
1st month of pregnancy: 1.01 (0.97-1.04)
Sample Description:
97,518 neonates born in
Southern California
March 2009
C-19
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Ritz et al.
(2002)
Period of Study:
1987- 1993
Location:
Southern California
Health Outcome:
Birth defects (heart defects &
orofacial clefts)
Study Design: Case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description: NR
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: Exposure categories: ppm
<1.14; 1.14-1.57; 1.57- 2.39; >2.39
OR for Birth defects [Lower CI, Upper CI]:
Period of exposure - Second month of pregnancy.
Aortic artery & valve defects:
1.00 (ref group); 1.10 (0.73-1.66); 1.25 (0.74-2.13);
0.93 (0.47-1.85)
Pulmonary artery & valve anomalies:
1.00 (ref group); 1.09 (0.69-1.73); 0.92 (0.50-1.70);
1.00 (0.46-2.17)
Ventricular septal defects:
1.00 (ref group); 1.62 (1.05-2.48); 2.09 (1.19-3.67);
2.95(1.44-6.05)
Conotruncal defects:
1.00 (ref group); 0.79 (0.47-1.32); 0.73 (0.36-1.47);
0.95 (0.38-2.38)
Notes: Results also presented for more specific defects, however CO
showed no association (see paper Table 3.). CO not associated with
orofacial clefts)
Author: Ritz et al.
(2006)
Period of Study:
1989-2000
Location:
Southern California
Health Outcome: Post-neonatal
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
O3: r = -0.57
Increment: 1 ppm
OR for Post-neonatal death [Lower CI, Upper CI]
Exposure period : 2 weeks prior to death, 1 month prior to death,
2 months prior to death, 6 months prior to death
All causes:
2 weeks prior to death : 1.14 (1.03-1.25)
2 months prior to death : 1.11 (1.06-1.16)
SIDS:
2 months prior to death : 1.19 (1.10-1.28)
Term/normal weight births
2 months 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 months 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, NO2,
PM10, O3)
Author: Ritz et al.
(2007)
Period of Study:
January-
December 2003
Location:
Los Angeles, CA
Health Outcome: PTB
Study Design:
Nested case-control
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
Asurvey of 2,543 of 6,37 4
women sampled from a cohort of
58,316 eligible births in Los
Angeles county.
Averaging Time: 24-h
Mean (SD) unit: NR
Copollutant correlation:
TSP: r = 0.73
NO2: r = 0.75
S02: r = 0.82
03: r = -0.39
Increment: Exposure categories (ppm):
Less than 0.58: 0.59-0.91; 0.92-1.25; >1.25
RR for LBW [Lower CI, Upper CI]
First trimester:
1.00 (Ref group); 1.17 (1.08-1.26); 1.15 (1.05-1.26); 1.25 (1.12-1.38)
6 weeks 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)
Author: Salam et
al. (2005)
Period of Study:
1975- 1987
Location:
California
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
Averaging Time: 24-h
Mean (SD) unit:
1.8 ppm (0.9)
(Entire pregnancy)
Range: NR
Copollutant: correlation
PM10: r = 0.41
N02: r = 0.69
03: r = -0.27
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 CI, Upper CI]
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 CI, Upper CI]
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 CI, Upper CI]
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)
March 2009
C-20
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Tsai et al.
(2006b)
Period of Study:
1994-2000
Location:
Kaoshiung, Taiwan
Health Outcome:
Postneonatal death
(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:
8.27 ppm x10
Range (Min, Max):
2.26,17.7
Copollutant: NR
Increment: Interquartile range : 0.31 ppm
OR for Post-neonatal mortality [Lower CI, Upper CI]
Lag examined : 0-2
Lag 0-2:1.051 (0.304-3.630)
Author: Wilhelm et
al. (2005)
Period of Study:
1994-2000
Location:
Los Angeles, CA
Health Outcome:
Term LBW and PTB
Study Design:
Retrospective cohort
Statistical Analyses:
Logistic regression
Age Groups Analyzed: NA
Sample Description:
518,254 births within 4 miles of
a monitoring station. Varied
according to analyses.
Averaging Time: 24-h
Mean (SD) unit:
Trimester 1:1.42 ppm
Results for third trimester and 6
weeks prior to birth were
similar to first trimester
Range (Min, Max):
0.26, 2.82
Copollutant correlation:
First Trimester:
PM10: r = 0.12
PM25: r= 0.57
NO2: r = 0.81
S02: r = -0.31
Increment: 1 ppm
RR for PTB [Lower CI, Upper CI]
First trimester:
Less than 1 mile: 1.06 (1.00-1.12)
1-2 miles: 1.06 (1.03-1.10)
2-4 miles: 1.08 (1.06-1.09)
ZIP code level: 1.04(1.01-1.07)
6 weeks prior to birth:
Less than 1 mile: 1.04(0.98-1.09)
1-2 miles: .04(1.01-1.08)
2-4 miles: 1.01 (0.99-1.02)
ZIP code level: 1.03 (1.00-1.06)
Notes: All results above did not persist in multipollutant model (CO,
NO2, O3, PM10)
OR for term LBW [Lower CI, Upper CI]
Third trimester:
Less than 1 mile: 1.10 (0.98-1.23)
1-2 miles: 1.05 (0.99-1.13)
2-4 miles: 1.06 (1.02-1.10)
ZIPcode level: 1.12(1.05-1.19)
Notes: All results above did not persist in multipollutant model (CO,
NO2, O3, PM10)
See paper for results based on exposure category groupings.
Author: Woodruff
et al. (2008)
Period of Study:
1999-2002
Location:
U.S. counties with
>250,000 residents
Health Outcome:
Post-neonatal deaths
All causes; respiratory; SIDS; i
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
O3: r = -0.46
Increment: 0.39 ppm
OR for Post-neonatal mortality [Lower CI, Upper CI]
Avg exposure over the first 2 months 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 et al.
(2004b)
Period of Study:
1994-2000
Location:
Taipei, Taiwan
Health Outcome:
Post-neonatal mortality
(27 days-1 yr old)
Study Design: Case-crossover
Statistical Analyses:
Poisson regression
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 CI, Upper CI]
Lag examined : 0-2
Lag 0-2:1.038 (0.663-1.624)
March 2009
C-21
DRAFT-DO NOT QUOTE OR CITE
-------
Table C-4. Studies of short-term CO exposure and respiratory morbidity
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Chen et al. (1999)
Health Outcome: Lung function (FVC,
Pollutant: CO
Increment: NR
Period of Study: 5/1995 -
FEV1, FEV1/FVC, FEF25-75%, PEF)
Averaging Time:
p Coefficient (SE); lag:
1/1996
Study Design: Cross-sectional survey
1-h maximum; 24-h avg
FVC (mL)
Location:
Statistical Analyses:
Mean (SD) unit: NR
24-h avg
3 Taiwan communities
Multivariate linear model
Population:
Range (Min, Max):
1-h maximum: (0.4, 3.6)
-66.6 (40.73); 1
-147.71 (64.48); 2
2.2 (48.13); 7
1-h maximum
941 children (Boys: 453; Girls: 488)
Copollutant correlation:
Age Groups Analyzed: 8-13
N02:r = 0.86-0.98
Note: To represent the
schoolchildren's exposure the daytime
-33.25 (20.74); 1
-16.48 (19.67); 2
-5.18 (16.48); 7
avg and peak concentrations were
FEV1 (mL)
24-h avg
20.55 (38.24); 1
-82.42 (60.95); 2
48.23 (45.58); 7
1-h maximum
1.2(19.48); 1
-1.44 (18.57); 2
20.96 (15.67); 7
measured from 0800 to 1800.
Author: Chen et al. (2000)
Period of Study:
8/1996-6/1998
Location:
Washoe County, NV
Health Outcome: School absenteeism
Study Design: Time-series
Statistical Analyses: Maximum
likelihood
Population: 1st to 6th qrade children:
27,793
Age Groups Analyzed: 1st to 6th grade
children
Pollutant: CO
Averaging Time: 1-h maximum
Mean (SD) unit:
2.73 (1.154) ppm
Range (Min, Max): (0.65, 2.73)
Copollutant correlation:
PM10: r = 0.721
03: r = -0.204
Increment: 1 0 ppm
% Increase (Lower CI, Upper CI); lag:
3.79% (1.04-6.55); 0
Author: de Hartog et al.
(2003)
Period of Study:
1998- 1999
Location:
Amsterdam, Netherlands;
Erfurt, Germany;
Helsinki, Finland
Health Outcome: Respiratory
symptoms (shortness of breath, being
awakened by breathing problems,
phlegm, wheezing, tripping heart)
Study Design: Time-series
Statistical Analyses:
Logistic regression
Population:
Non-smoking individuals with CHD:
Amsterdam: 37
Erfurt: 47
Helsinki: 47
Age Groups Analyzed:
>50
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Amsterdam: 0.6 mg/m3
Erfurt: 0.4 mg/m3
Helsinki: 0.4 mg/m3
Range (Min, Max):
Amsterdam: (0.4,1.6)
Erfurt: (0.1,2.5)
Helsinki: (0.1,1.0)
Copollutant:
PM25;
N02
Increment: 0.25 mg/m3
Odds Ratio (Lower CI, Upper CI); lag:
Incidence of symptoms
Shortness of breath
1 (0.92-1.1); 0
0.96 (0.88-1.05); 1
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
1.03 (0.93-1.15); 2
1.11 (1-1.22); 3
1.16 (0.98-1.37); 0-4
Phlegm
,_..0
; 1
; 2
; 3
; 0-4
1.05 (0.93-1.19);
1.02 (0.91-1.14);
1.08 (0.96-1.22);
1.09 (0.97-1.22);
1.13 (0.94-1.35);
Prevalence of symptoms
Shortness of breath
1 (0.94-1.06); 0
0.99 (0.94-1.05)
0.99 (0.93-1.05)
1.01 (0.95-1.07)
0.98 (0.9-1.07); 0-4
Being awakened by breathing problems
1.01 (0.93-1.1); 1
0.99 (0.91-1.08); 2
1.1 (1.02-1.19); 3
1.13 (1-1.29); 0-4
March 2009
C-22
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Delfino et al.
(2003)
Period of Study:
11/1999- 1/2000
Location:
Los Angeles, CA
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 mixed
model
Population:
22 asthmatic Hispanic children
Age Groups Analyzed: 10-15
Pollutant: CO
Averaging Time:
1-h maximum; 8-h maximum
Mean (SD) unit:
1-h maximum: 7.7 (3.1) ppb
8-h maximum: 5.0 (2.0) ppb
Range (Min, Max):
1-h maximum: (2,17)
8-h maximum: (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
Increment: 5.0 ppb & 3.0 ppb
Odds Ratio (Lower CI, Upper CI); lag:
1-maximum
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); 1
8-h maximum
Increment: 3 0 ppb
Symptom scores >1
0.95 (0.55-1.62); 0
1.2 (0.77-1.86); 1
Symptom scores >2
0.53 (0.10-2.92); 0
1.43 (0.41-5.00); 1
Author: Estrella et al.
(2005)
Period of Study: 1/2000 -
4/2000
Location:
Quito, Ecuador
Health Outcome:
Acute respiratory infection
Study Design: Prospective study
Statistical Analyses:
Logistic regression; Poisson
Population: 960 children
Age Groups Analyzed: 6-11
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Odds Ratio (Lower CI, Upper CI); 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)
Author: Fischer et al.
(2002)
Period of Study: NR
Location:
Utrecht, Netherlands
Health Outcome:
Lung function (FVC, FEVi, PEF, MMEF)
Study Design: Panel study
Statistical Analyses:
Restricted maximum likelihood linear
model
Population: 68 children
Age Groups Analyzed: 10-11
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 921 |jg/m3
Range (Min, Max): (319,1540)
Copollutant:
PM10; BS; N02; NO
Increment: 100 |jg/m3
mL (SE); lag:
FVC: 0.5 (0.4); 1; 0.1 (0.2); 2
FEVi: -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
March 2009
C-23
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Lagorio et al.
(2006)
Period of Study:
5/1999-6/1999;
11/1999- 12/1999
Location:
Rome, Italy
Health Outcome:
Lung function (FVC, FEVi)
Study Design:
Time-series panel study
Statistical Analyses:
Generalized estimating equations (GEE)
Population:
COPD panel: 11
Asthma panel: 11
IHD panel: 7
Age Groups Analyzed:
COPD panel: 50-80
Asthma panel: 18-64
IHD panel: 40-64
Notes: Asthma panel was restricted to
never smokers, while COPD and IHD
panels include former smokers if
smoking cessation occurred at least 1
year prior to enrollment.
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:
PM2.5: r = 0.67
PM10-2.5: r = -0.09
PM10: r = 0.55
N02: r = 0.05
O3: r = -0.87
S02: r = 0.65
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/
0.132 (0.165); 0-2
FEV1 (% of predicted)
0.204 (0.120); 0
0.114 (0.142); 0-1
0.159 (0.194); 0-2
Author: Park et al. (2002) Health Outcome: School absenteeism
Period of Study:
3/1996- 12/1999
Location:
Seoul, Korea
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;
NO2: r = 0.70;
S02: r = 0.67;
O3: r = -0.46
Increment: 0 52 ppm
Relative Risk (Lower CI, Upper CI); 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 et al. (2005a)
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
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 CI, Upper CI); 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)
March 2009
C-24
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Penttinen et al.
(2001)
Period of Study: 11/1996 -
4/1997
Location:
Helsinki, Finland
Health Outcome:
Lung function (PEF)
Study Design: Panel study
Statistical Analyses:
First order autoregressive linear model
Population:
57 non-smoking adult asthmatics
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time: 24-h avg
Median unit: 0.4 mg/m3
Range (Min, Max):
(0.1,1.1) mg/m3
Copollutant correlation:
PM10: r = -0.03
PM10-25: r = -0.30
PM25: r = 0.32
PM1: r= 0.39
PNC: r = 0.44
NC0.01-0.1: r = 0.43
NC0.1-1: r = 0.47
NO: r= 0.60
NO2: r = 0.44
Increment: 0.2 mg/m3
p Coefficient (SE); lag:
PEF Deviations (L/min)
Morning
0.27 (0.38); 0
-1.08 (0.36); 1
0.23 (0.38); 2
-1.11 (1.19); 5-day avg
Afternoon
-0.4 (0.43); 0
-0.13 (0.41); 1
-0.71 (0.41);2
-3.03 (1.06); 5-day avg
Evening
-0.7 (0.45); 0;
-0.31 (0.44); 1
0.3 (0.44); 2
-3.62 (1.19); 5-day avg
Co-pollutant models with PNC
Morning: -0.67 (0.64); 1
Afternoon: -0.46 (0.69); 0
Evening: -0.46 (0.73); 0
Author: Rabinovitch et al.
(2004)
Period of Study: 11/1999 -
3/2000; 11/2000-3/2001;
11/2001 -3/2002
Location:
Denver, CO
Health Outcome:
Lung function (FEV1); asthma
exacerbation; bronchodilator use
Study Design: Panel study
Statistical Analyses:
Pulmonary function: Mixed effects
model; Asthma exacerbation and
medication use: GLM
Population:
Urban poor asthmatic children:
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; NO2; SO2; O3
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 CI, Upper CI); lag:
Asthma exacerbation:
1.012 (0.913-1.123); 3-day ma
Bronchodilator use:
1.065 (1.001-1.133); 3-day ma
1999-2000
2000-2001
2001-2002
Age Groups Analyzed: 6-12
Author: Ranzi et al. (2004)
Period of Study:
2/1999-5/1999
Location:
Emilia-Romagna, Italy
Health Outcome:
Lung function; respiratory symptoms,
medication use
Study Design: Panel study
Statistical Analyses: GLM
Population:
120 "asthma-like" school children
Age Groups Analyzed: 6-11
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: NO2; TSP; PM2.5
The study did not present quantitative
results for CO.
Author: Rodriguez et al.
(2007)
Period of Study:
1996-2003
Location:
Perth, Australia
Health Outcome:
Respiratory symptoms (body
temperature, cough, wheeze/rattle
chest, runny/blocked nose)
Study Design: Panel study
Statistical Analyses:
Logistic regression, GEE
Population:
263 children at high risk of developing
asthma
Age Groups Analyzed: 0-5
Pollutant: CO
Averaging Time: 8-h avg
Mean (SD) unit: 1 408 ppm
Range (Min, Max): (0.012, 8.031)
Copollutant: NR
Increment: NR
Odds Ratio (Lower CI, Upper CI); 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
1.136 (1.016-1.26); 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
March 2009
C-25
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Schildcrout et al.
(2006)
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Co pollutant:
NO2; O3; PM10; SO2
Increment: 1 0 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Asthma Symptoms
1.08 (1.01-1.14); 0
1.07 (0.99-1.16); 1
1.08 (1.02-1.15); 2
1.05 (1.01-1.09); 0-2
Asthma Symptoms
+ 20 ppb increase in NO2
1.07 (1-1.14); 0
1.04(0.96-1.11); 1
1.09 (1.02-1.16); 2
1.04 (1-1.08); 0-2
+ 25 pg/m3 increase in PM10
1.08 (1.01-1.15); 0
1.06 (0.99-1.14); 1
1.08 (1.02-1.14); 2
1.05 (1.01-1.08); 0-2
+ 10 ppb increase in SO2
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
1.06 (1.01-1.1); 2
1.04 (1.01-1.07); 0-2
Rescue Inhaler Use
+ 20 ppb increase in NO2
1.05 (0.99-1.12); 0
1.04 (0.98-1.11); 1
1.07 (1.02-1.12); 2
1.04(1-1.07), 0-2
+ 25 pg/m3 increase in PM10
1.06 (0.99-1.13); 0
1.05 (0.99-1.11); 1
1.05 (1.01-1.09); 2
1.03 (1-1.07); 0-2
+ 10 ppb increase in SO2
1.04 (0.96-1.12); 0
1.04 (0.97-1.1); 1
1.08 (1.03-1.13); 2
1.04 (1-1.08); 0-2
Author: Silkoffet al. (2005)
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:
1st winter: 16 with a history of more than
10 pack years 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 years 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: a 40
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.
March 2009
C-26
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: Slaughter et al.
(2003)
Period of Study:
12/1994-8/1995
Location:
Seattle, WA
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
Pollutant: CO
Averaging Time: 24-h avg
Median unit: 1 47 ppm
IQR (25th, 75th): (0 23,1 87)
Copollutant: NR
Increment:
Increased asthma attack severity: 0.67 ppm
Increased rescue inhaler use: 1.0 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Increased asthma attack severity:
Without transition: 1.21; 1
With transition: 1.17; 1
Increased rescue inhaler use:
Without transition: 1.09(1.03-1.16); 1
With transition: 1.06 (1.01-1.1); 1
Author: Steerenberg et al.
(2001)
Period of Study: NR
Location:
Bilthoven and Utrecht,
the Netherlands
Health Outcome:
Lung function (PEF); exhaled nitric
oxide; inflammatory nasal markers
Study Design: Panel study
Statistical Analyses:
Restricted maximum likelihood linear
model
Population: 126 children
Age Groups Analyzed: 8-13
Notes: The study was only conducted
for a two month period: February and
March.
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
The study did not present quantitative
results for CO.
Author: Timonen et al.
(2002)
Period of Study: 2/1994 -
4/1994
Location:
Kuopio, Finland
Health Outcome:
Exercise induced bronchial
responsiveness; Lung function (FVC,
FEVi, MMEF, AEFV)
Study Design: Panel study
Statistical Analyses: Linear regression
Population:
33 children with chronic respiratory
symptoms
Age Groups Analyzed: 7-12
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
PNCo. 1-0.3: r = 0.71
PNC0.3-1.0: r = 0.60
PNC1.0-3.2: r = 0.84
PNC3 2-10: r = 0.79
N02: r = 0.85
Increment: 0.32 mg/m3
p Coefficient (SE); lag:
Exercise induced responsiveness
AFEV1 (%)
-0.081 (0.647); 0
0.03 (0.262); 1
0.087 (0.26); 2
-0.091 (0.275); 3
0.19 (0.599); 0-3
AMMEF (%)
0.442 (1.79); 0
0.52 (0.723); 1
0.313 (0.719); 2
-0.616 (0.75); 3
0.096 (1.64); 0-3
AAEFV (%)
0.287 (1.19); 0
0.281 (0.482); 1
0.904 (0.474); 2
0.15 (0.483); 3
1.6 (1.05); 0-3
FVC (mL)
0.064 (10.9); 0
-4.79 (4.51); 1
-9.78 (4.24); 2
-13.9 (4.7); 3
-29.4 (10.1); 0-3
FEV1 (mL)
19.2 (13.2); 0
-9.04 (5.45); 1
-9.15 (5.21); 2
-11.7 (5.77); 3
-17.5 (12.5); 0-3
MMEF (mL/s)
22.2 (36.9); 0
-23(15.2); 1
-4.63 (14.7); 2
-30.9 (16); 3
-24.9 (34.8); 0-3
AEFV (L2/s)
-0.093 (0.088); 0
-0.068 (0.036); 1
-0.06 (0.035); 2
-0.05 (0.039); 3
-0.076 (0.083); 0-3
March 2009
C-27
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
CO Effect Estimates (95% CI)
Author: von Klot et al.
(2002)
Period of Study:
9/1996-3/1997
Location:
Erfurt, Germany
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0 9 mg/m3
Range (Min, Max): (0.3, 3.0)
Copollutant correlation:
NCo.01-0.1: r = 0.66
NCo.i-o.5. r = 0.79
NCo.5-2.5. r = 0.46
MCo.i-o.5. r = 0.66
MC001-25: r = 0.65
PM2.5-IO: r = 0.42
PM10: r = 0.69
N02: r = 0.82
S02: r = 0.32
Increment:
0 and 5-day avg lag: 0.6 mg/m3
14-day avg lag: 0.54 mg/m3
Odds Ratio (Lower CI, Upper CI); lag:
Prevalence: Inhaled p2-agonist use
0.98 (0.93-1.03)
1.04 (0.97-1.12)
0.93 (0.86-1.01)
0
0-4
0-13
Prevalence: Inhaled corticosteroid use
1.05 (1-1.11); 0
1.25 (1.17-1.34); 0-4
1.06 (0.97-1.15); 0-13
Prevalence: Wheezing
1.03 (0.97-1.08); 0
1.13 (1.05-1.22); 0-4
1.14 (1.05-1.25); 0-13
Co-pollutant models
Inhaled p2-agonist use
CO+MCO.01-2.5:
1 (0.91-1.11); 0-4
CO+NC00.01-0.1:
1.01 (0.91-1.11); 0-4
Inhaled corticosteroid use
CO+MCO.01-2.5:
0.89 (0.81-0.98); 0-13
CO+NC: 0.01-0.
1:0.81 (0.72-0.91); 0-13
Wheezing
CO+MCO.01-2.5:
1.15 (1.04-1.27); 0-4
CO+NC0.01-0.1:
1.09 (0.98-1.22); 0-4
Author: Yu et al. (2000)
Period of Study:
11/1993-8/1995
Location:
Seattle, Washington
Health Outcome:
Asthma symptoms (Wheezing,
coughing, chest tightness, shortness of
breath)
Study Design: Panel study
Statistical Analyses:
Repeated measures logistic regression
models (GEE)
Population:
133 mild-to-moderate asthmatics
Age Groups Analyzed: 5-13
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 6 ppm
Range (Min, Max): (0.65, 418)
Copollutant correlation:
PM10: r = 0.82
PM10: r = 0.86
S02: r = 0.31
Increment: 1 0 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Marginal GEE
1.22 (1.03-1.45); 0
1.3(1.11-1.52); 1
1.26 (1.09-1.46); 2
Transition GEE
1.18 (1.02-1.37); 0
1.25 (1.1-1.42); 1
1.18 (1.04-1.33); 2
Table C-5. Studies of short-term CO exposure and respiratory hospital admissions and ED visits.
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Anderson et al Hospital Admission
(2001)
Period of Study:
10/1994-12/1996
Location:
West Midlands; U.K.
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:
AO ages
0-14
15-64
>65
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 0.8 (0.7) ppm
Range (Min, Max): (0.2,10)
Copollutant: correlation
PM10: r = 0.55;
PM2.5: r= 0.54;
PM25-10: r = 0.10;
BS: r = 0.77;
S042-: r = 0.17;
N02: r = 0.73;
03: r = -0.29;
S02: r = 0.49
Increment: 10 ppm
% Increase (Lower CI, Upper CI); lag:
Respiratory Diseases
Age Group
0.3% (-1.10 to 1.70); 0-1
1.50% (-0.60 to 3.60); 0-1
-0.70% (-3.60 to 2.30); 0-1
0.00% (-2.10 to 2.10); 0-1
All ages
0-14:
15-64:
>65:
Asthma
Age Group
0-14:
15-64:
COPD
Age Group
>65:
3.90% (-0.50 to 8.50); 0-1
-4.90% (-10.60 to 1.10); 0-1
1.00% (-2.50 to 4.60); 0-1
March 2009
C-28
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Andersen et al.
(2007)
Period of Study:
1/1999-12/2004
Location:
Copenhagen,
Denmark
Hospital Admission
Health Outcome (ICD10):
Respiratory diseases: Chronic
bronchitis (J41-42), Emphysema
(J43), COPD (J44), Asthma
(J45), Status asthmaticus (J46),
Pediatric asthma (J45),
Pediatric asthmaticus (J46)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: 5-18;
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.3 (0.1) ppm
IQR (25th, 75th): (0.22, 0.34)
Copollutant; correlation:
PM10: r = 0.45
Increment: 0 12 ppm
Relative Risk (Lower CI, Upper CI); lag:
Respiratory Disease
Age Group: > 65
CO: 1.024(0.997-1.053); 0-4
CO, PM10:1.001 (0.961-1.042); 0-4
Asthma
Age Group: 5-18
CO: 1.104(1.018-1.198); 0-5
CO, PM10:1.023 (0.911-1.149); 0-5
Author: Atkinson et al.
(1999)
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:
AOaqes
0-14
15-64
>65
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 CI, Upper CI); lag:
Respiratory complaints
Age Group
All ages: 0.76% (-0.83,2.38); 1
0-14: 2.92% (0.60, 5.30); 1
15-64:2.15% (-0.27, 4.63); 1
>65: 4.29% (1.15, 7.54); 0
Asthma visits:
Single-pollutant model
Age Group:
All ages: 3.32% (0.56,6.16); 1
0-14: 4.13% (-0.11, 8.54); 0
15-64: 4.41% (0.46,8.52); 1
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, SO2: 2.34% (-1.94, 6.81); 0
CO, PM10: 2.93% (-1.53, 7.58); 0
CO, BS: 4.19% (-0.04, 8.60); 0
Author: Bedeschi et al ED Visits
(2007)
Period of Study:
1/2001-3/2002
Location:
Reggio Emilia,
Italy
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: <15
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
NO2: r = 0.77
The study did not provide quantitative results for CO.
March 2009
C-29
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Bell et al.
(2008)
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:
AO 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 CI, Upper CI); 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 > 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% (-3.54 to 3.42); 1
0.37% (-2.95 to 3.8); 2
-1.08% (-4.34 to 2.3); 3
0.3% (-4.71 to 5.57); 0-3
Author: Bellini et al.
(2007)
Period of Study:
1996-2002
Location:
15 Italian cities
Hospital Admissions
Health Outcome:
Respiratory Conditions
Study Design:
Time-series; Meta-analysis
Statistical Analyses:
1. GLM for city-specific
estimates
2. Bayesian random-effects for
meta analysis
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: correlation NR
Increment: 1 mg/m3
% Increase (Lower CI, Upper CI); Lag
Respiratory conditions
All ages:
Season:
Winter: 0.58%; 0-1
Summer: 3.47%; 0-1
All Season: 125%; 0-3
Note: Estimates from Biggeri et al. (2004)
Author: Braga et al.
(2001)
Period of Study:
1/1993-11/1997
Location:
Sao Paulo, Brazil
Hospital Admissions
Health Outcome (ICD9):
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
<2
3-5
6-13
14-19
0-19
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 4.8 (2.3) ppm
Range (Min, Max): (0.6,191)
Copollutant:
correlationCopollutant: correlation
PM10: r = 0.60
O3: r = -0.07
S02: r = 0.47
Increment: 3 ppm
% Increase (Lower CI, Upper CI); lag:
Respiratory
Age Group:
<2:5.00% (3.30-6.80); 0-6
3-5: 4.90% (1.40-8.50); 0-6
6-13:1.00% (-2.50 to4.60); 0-6
14-19:11.30% (5.90-16.80); 0-6
0-19: 4.90% (3.50-6.40); 0-6
March 2009
C-30
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Burnett et al.
(1999)
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
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
PM10: r = 0.43
NO2: r = 0.55
S02: r = 0.37
03: r = -0.23
Increment: 1 18 ppm
% Increase (t-value); lag:
Asthma: 5.35% (3.92); 0
COPD: 2.93% (1.48); 0
Respiratory Infection: 5.00% (4.25); 0
Asthma:
Multi-pollutant model
CO, S02,03: 5.15%
CO, PM25.SO2.O3: 4.63%
CO. PM10-25. S02. 03: 5.25%
CO. PM10. S02. 03: 4.80%
CO. PM10-2 5. 03: 4.00%
COPD:
Multi-pollutant model
CO. SO2. 03: 3.02%
CO. PM25.SO2.O3: 2.46%
CO. PM10-25. S02. 03: 3.00%
CO. PM10. S02. 03: 2.75%
CO. PM10-25.O3: 3.00%
Author: Burnett et al.
(2001)
Period of Study:
1/1980-12/1994
Location:
Toronto, ON,
Canada
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
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 19 ppm
IQR (25th, 75th): (1.3, 2.3)
Copollutant: correlation
03: r= 0.24
Increment: 1 9 ppm
% Increase (Lower CI, Upper CI); lag
Respiratory problems
CO: 19.20%; 0-1
CO, 03:14.30%; 0-1
Author: Cakmak et al.
(2006b)
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 CI, Upper CI); lag:
Respiratory disease
CO: 0.60% (0.20,1); 2.8
CO, S02, N02,03: -0.20% (-0.70- 0.30); 2.£
Author: Cheng et al.
(2007)
Period of Study:
1996-2004
Location:
Kaohsiung, Taiwan
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-486)
Study Design: Bi-directional
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.76 ppm
Range (Min, Max): (0.14,1.72)
Copollutant: correlation
PM10
S02
N02
03
Increment: 0.31 ppm
Odds Ratio (Lower CI, Upper CI); 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: 1.18 (1.14-1.23); 0-2
CO, <25 °C: 1.47 (1.41-1.53); 0-2
CO, PM10, a 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
CO,03, <25 °C: 1.44 (1.38-1.5); 0-2
March 2009
C-31
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Cho et al.
(2000)
Period of Study:
1/1996-12/1996
Location:
3 South Korea cities:
Hospital Admissions
Health Outcome (ICD9):
Bronchial asthma; COPD;
Bronchitis
Study Design: Time-series
Statistical Analyses: Poisson
GAM, LOESS
Age Groups Analyzed: All
Ages
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;
O3 Max: r = -0.069
Ulsan
S02: r = 0.108; N02:r= 0.446;
TSP: r = 0.286; 03: r = -0.195;
O3 Max: r = -0.107
Suwon
S02: r = 0.556; N02:r= 0.291;
TSP: r = 0.496; 03: r = -0.371;
O3 Max: r = -0.365
Increment: 1,000 ppm
Relative Risk (Lower CI, Upper CI); lag:
Estimates obtained using dummy variables to apply
environmental indicators to the model
Daejeon
CO: 1.26 (1.08-1.47)
TSP, SO2, N02, 03:1.21 (1.02-1.44)
Ulsan
CO: 3.55 (1.65-7.63)
TSP, SO2, N02, 03: 2.51 (1.06-5.93)
Suwon
CO: 1.24(0.97-1.59)
TSP, SO2, N02, 03:1.19 (0.88-1.61)
Estimates obtained using actual measured integrated
environmental pollution indicator values
Daejeon
CO: 1.34(1.14-1.58)
Ulsan
CO: 1.27 (0.94-1.71)
Suwon
CO: 3.55 (1.27-9.93)
Author: Farhat et al.
(2005)
Period of Study:
8/1996-8/1997
Location:
Sao Paulo, Brazil
Hospital Visits & ED Visits
Health Outcome (ICD9):
Pneumonia/bronchopneumonia
(480-486); Asthma (493);
Bronchiolitis (466)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 3.8 (1.6) ppm
Range (Min, Max): (1.1,11.4)
Copollutant: correlation
PM10: r = 0.72;
S02: r = 0.49;
N02: r = 0.59;
O3: r = -0.8
Increment: 1 8 ppm
% Increase (Lower CI, Upper CI); 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
CO, 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
Author: Fung et al.
(2006)
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. Bi-directional case-crossover
Statistical Analyses:
1. Dewanji and Moolgavkar
2. Poisson
3. Conditional logistic regression
Age Groups Analyzed: a 65
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; O3: r = -0.53;
N02: r= 0.74; S02: r = 0.61;
PM10: r = 0.46; PM2.5: r = 0.23;
PM10-2.5: r = 0.51
Increment: 0 24 ppm
Relative Risk (Lower CI, Upper CI); 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
Bi-directional case-crossover
1.010 (0.006-1.023); 0
1.012 (0.996-1.027); 0-2
1.012 (0.995-1.03); 0-4
1.010(0.991 1.031; 0-6
March 2009
C-32
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Fusco et al.
(2001)
Period of Study:
1/1995 10/1997
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:
AOaqes
0-14
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
N02: r = 0.31
O3: r = -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: 15 mg/m3
% Increase (Lower CI, Upper CI); lag:
Age Group: All Ages
Respiratory conditions
2.80% (1.30-4.30); 0
1.80% (0.20-3.30); 1
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, NO2: 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, NO2:6.90 (0.80-13.40); 1
Asthma
6.30 (-0.50 to 13.50); 0
8.20 (1.10-15.70); 1;
-0.70 (-7.30 to 6.30); 2
3.50 (-3.20 to 10.60); 3;
4.80 (-1.90 to 12.00); 4
CO, NO2:3.30 (-4.20 to 11.30); 1
Author: Gouveia and
Fletcher (2000a)
Period of Study:
11/1992-9/1994
Location:
Sao Paulo,
Brazil
Hospital Admissions
Health Outcome (ICD9): All
respiratory diseases Pneumonia
(480-486); Asthma (493);
Bronchitis (466, 490, 491)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
<1
<5
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 5.8 (2.4) ppm
Range (Min, Max): (1 3, 22.8)
Copollutant: correlation
PM10: r = 0.63
S02: r = 0.65
N02: r = 0.35
Increment: 6 9 ppm
Relative Risk (Lower CI, Upper CI); lag:
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
Asthma
Age Group:
<5:1.081 (0.98-1.192); 0
March 2009
C-33
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Hajat et al.
(1999)
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, SOS-
SOS, 510-515, 518, 519, 786)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
AOaqes
0-14
15-64
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
All year: 0.8 (0.4) ppm
Warm Season
(April-September): 0.7 (0.3) ppm
Cool Season
(October-March): 10 (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 CI, Upper CI); 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 to 5.20); 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 Group & 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
a 65 & Warm Season: 15.60% (3.10-29.60); 2
a 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
a 65 & Warm Season: 1.00% (-1.60 to 3.80); 2
a 65 & Cold Season: -2.20% (-6.50 to 2.40); 3
Author:
Hajat et al. (2002)
Period of Study:
1/1992-12/1994
Location:
London, U.K.
General Practitioner Visits
Health Outcome (ICD9):
Upper Respiratory Diseases
(URD)
Study Design: Time-series
Statistical Analyses:
Poisson, GAM, LOESS
Age Groups Analyzed:
0-14
15-64
a 65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
All year: 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: (0.4,10)
Cool Season: (0.5, 16)
Copollutant: NR
Increment: 0.6 ppm, 0.8 ppm, & 1.1 ppm
% Increase (Lower CI, Upper CI); 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); 1
>65: 4.90% (-1.80 to 12.10); 3
Cold Season, Increment: 1 1 ppm
Age Group
0-14: -2.50% (-4.90 to 0.10); 1
14-64:0.60% (-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); 1
>65:5.80% (2.40 to 9.30); 3
March 2009
C-34
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Hapcioglu et
al. (2006)
Period of Study:
1/1997-12/2001
Location:
Istanbul,
Turkey
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494-496)
Study Design: Cross-sectional
Statistical Analyses:
Pearson Correlation Coefficient
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: Monthly
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Correlation Coefficient:
Between CO exposure and COPD: 0.57
Between CO exposure and COPD when controlling for
temperature: 0.25
Author: Hinwood et al.
(2006)
Period of Study:
1/1992-12/1998
Location:
Perth,
Australia
Hospital Admissions
Health Outcome (ICD9):
COPD (490.00-496.99
excluding asthma)
Pneumonia/influenza (480.00-
489.99); Asthma (493)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
All Year: 2.3 (1.3) ppm;
November-April: 2.2 (1.3) ppm;
May-October: 2.4 (1.2) ppm
Range (10th, 90th):
All Year: (0.9, 4.2)
November-April: (0.8, 4.2)
May-October: (1.1, 4.2)
Copollutant: correlation
All Year:
NO2: r = 0.57
O3: r= 0.00
November-April:
NO2: r = 0.55
O3: r= 0.00
May-October:
NO2: r = 0.57
03: r= 0.16
Increment: 2 3 ppm
Odds Ratio (Lower CI, Upper CI); Lag
Pneumonia
0.99999 (0.9737-
1.00650 (0.9806-
1.00351 (0.9779-
1.00424 (0.9790-
1.00581 (0.9752-
1.01005 (0.9755-
1.00805 (0.9701-
COPD
0.99915
1.00205
0.98630
0.98970
0.99960
0.99260
0.99160
(0.9693-
(0.9727-
(0.9577-
(0.9619-
(0.9647-
(0.9538-
(0.9493-
1.0268)
1.0331)
1.0298)
1.0301)
1.0374)
1.0458)
1.0474)
1.0297)
1.0323)
1.0158)
1.0182)
1.0357)
1.0329)
1.0357)
0
1
2
3
0-1
0-2
0-3
0
1
2
3
0-1
0-2
0-3
Author: Hwang and
Chan (2002)
Period of Study:
1998
Location:
50 communities in
Taiwan
Clinic Visits
Health Outcome (ICD9):
Lower respiratory tract
infections (466, 480-486)
Study Design: Time-series
Statistical Analyses:
1. General linear regression
2. Bayesian hierarchical
modeling
Age Groups Analyzed:
AO Ages
0-14
15-64
>65
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
1.00 (0.30) ppm
Range (Min, Max): (0.51,1.71)
Copollutant: NR
Increment: 01 ppm
% Increase (Lower CI, Upper CI); 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: 0-14
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); 1
0.20% (-0.10 to 0.40); 2
Age Group: > 65
1.10% (0.80-1.50)
0.60% (0.30-1.00)
0.40% (0.10-0.80)
Author: Ito et al. (2007)
Period of Study:
1999-2002
Location:
New York City, NY
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
GLM
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
All Season: 1.31 (0.43) ppm
Warm Months (April-September):
1.22 (0.32) ppm
Cold Months (October-March): 1 41
(0.5) ppm
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
Increment: 13 ppm
Relative Risk (Lower CI, Upper CI); Lag
Warm months: 1.15 (1.07-1.25); 0-1
March 2009
C-35
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Karr et al.
(2007)
Period of Study:
1995-2000
Location:
South Coast Air Basin,
CA
Hospital Admissions
Health Outcome (ICD9):
Acute bronchiolitis (466.1)
Study Design: Matched case-
control
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Infants: 3 weeks -1 yr
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Chronic: 1,770 ppb
Subchronic: 1,720 ppb
Range (Min, Max):
Chronic: (120, 8300)
Subchronic: (130, 5070)
Copollutant: NR
Increment: 910 ppb, 960 ppb
Odds Ratio (Lower CI, Upper CI); lag:
Increment: 910 ppb
Subchronic broncholitis: 1 (0.97-1.03)
Increment: 960 ppb
Chronic broncholitis: 1 (0.97-1.03)
Author: Karr et al
(2006)
Period of Study:
1995-2000
Location:
South Coast Air Basin,
CA
Hospital Admissions
Health Outcome (ICD9):
Acute bronchiolitis (466.1)
Study Design: Case-Crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
Infants: 3 weeks - 1 year
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)
Copollutant: NR
* Index days: days lagged in reference
to date of hospitalization of a case.
Referent days: are for each case and
includes all days that are the same
day of week and in the same month as
the index day for that case for CO.
Increment: 1361,1400 ppb
Odds Ratio (Lower CI, Upper CI); Lag
Increment: 1361 ppb
Age Group:
Overall: 0.99 (0.96-1.02); 1
25-29 weeks: 0.86 (0.68-1.1); 1
29 1/7-34 weeks: 1 (0.86-1.15); 1
341/7 - 37 weeks: 0.95 (0.87-1.04);
37 1/7-44 weeks: 1 (0.97-1.03); 1
Increment: 1400 ppb
Age Group:
Overall: 0.97 (0.94-1); 4
25-29 weeks: 0.93 (0.72-1.2); 4
29 1/7-34 weeks: 0.89 (0.77-1.03);
341/7 - 37 weeks: 0.98 (0.90-1.08);
37 1/7-44 weeks: 0.97 (0.94-1); 4
Author: Kim et al.
(2007)
Period of Study:
2002
Location:
Seoul,
Korea
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45 and J46)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
AO Ages
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit:
Daily Concentration: 8.6 (4.6) ppm
Relevant Concentration:
2.8(2.8) ppm
Range (Min, Max):
Daily Concentration: (0.8, 44.0)
Relevant Concentration: (0.0, 30.4)
Copollutant: NR
Quintile 3:1.04(1.02-1.07)
Quintile 4:1.10(1.06-1.15)
Quintile 5:1.06 (1.03-1.09)
Relative Risk (Lower CI, Upper CI); lag:
Individual Level SEP
Quintile 1:1.06 (1.02-1.09); 1-3 ma
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
1-3 ma
1-3 ma
1-3 ma
Overall: 1.06 (1.04-1.07); 1-3 ma
Relative Effect Modification for SES
Individual Level SEP
Quintile 1:1
Quintile 2:1 (0.95.1.04); 1-3 ma
Quintile 3: 0.99 (0.94-1.03); 1-3 ma
Quintile 4:1.02 (0.97-1.06); 1-3 ma
Quintile 5: 0.99 (0.94-1.04); 1-3 ma
Regional Level SEP
Quintile 1:1
Quintile 2:1.05 (0.97-1.14); 1-3 ma
Quintile 3:1.03 (0.96-1.11); 1-3 ma
Quintile 4:1.08 (1-1.16); 1-3 ma
Quintile 5:1.05 (0.97-1.13); 1-3 ma
March 2009
C-36
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Kontos et al.
(1999)
Period of Study:
1/1987-12/1992
Location:
Piraeus, Greece
Hospital Admissions
Health Outcome (ICD9):
Respiratory conditions
(laryngitis, bronchiolitis,
tonsillitis, acute rhinopharyngitis,
otitis, bronchopneumonia,
pneumonia, asthma)
Study Design: Time-series
Statistical Analyses:
Stochastic dynamical system
approach
Age Groups Analyzed: 0-14
Pollutant: CO
Averaging Time: 24-h avg
Mean Range (SD) unit:
1987: 4.2 mg/m3
1992: 3.6 mg/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
This study did not present quantitative results for CO.
Author: Lee et al.
(2002)
Period of Study:
12/1997-12/1999
Location:
Seoul, Korea
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45, J46)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: <15
Pollutant: CO
Averaging Time: 1-h maximum
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
NO2: r = 0.785
03:r = -0.388
Increment: 1.0 ppm
Relative Risk (Lower CI, Upper CI); lag:
RR for asthma and exposure to various pollutants for
children under 15 years old
Pollutant:
CO: 1.16 (1.10-1.22); 2-3 avg
CO, PM10: 1.13 (1.07-1.20); 2-3 avg
CO, S02: 1.17 (1.08-1.27); 2-3 avg
CO, N02: 1.04 (0.95-1.14); 2-3 avg
CO, 03: 1.16 (1.11-1.22); 2-3avg
CO, 03, PM10:1.148 (1.084-1.217); 2-3 avg
CO, 03, PM10, S02:1.168 (1.075-1.269); 2-3 avg
CO, 03, PM10, S02, N02:1.098 (0.994-1.214); 2-3 avg
Author: Lee et al
(2006)
Period of Study:
1/2002-12/2002
Location:
Seoul,
Korea
Hospital Admissions
Health Outcome (ICD10):
Asthma (J45-46)
Study Design: Time-series
Statistical Analyses:
GAM with stringent parameters
Age Groups Analyzed: <15
Pollutant: CO
Averaging Time:
Maximum 2-h avg
Mean (SD) unit:
High SES: 6.08 (2.10) ppb
Moderate SES: 6.35 (2.44) ppb
Low SES: 6.67 (2.59) ppb
Range (Min, Max): NR
Copollutant: correlation
NO2: r = 0.55
S02: r = 0.72
PM10: r = 0.28
03: r = -0.36
Increment: 3.01 ppb, 0.26 ppb, 4.52 ppb, 3.68 ppb
Relative Risk (Lower CI, Upper CI); lag:
Increment: 3.01 ppb
Overall: 1.07 (0.96-1.20); 0
Increment: 0.26 ppb
High SES: 1.06 (0.96-1.17); 0
Increment: 4.52 ppb
Moderate SES: 0.96 (0.84-1.10); 0
Increment: 3.68 ppb
Low SES: 1.02 (0.85-1.24); 0
Author: Lee et al
(2007b)
Period of Study:
1996-2003
Location:
Kaohsiung, Taiwan
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Bi-directional
case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.77 ppm
Range (Min, Max): (0.23,1 72)
Copollutant:
PM10
S02
N02
Os
Increment: 0.29 ppm
Odds Ratio (Lower CI, Upper CI); lag:
CO
<25°C : 1
> 25°C : 1
CO. PM10
<25°C : 1
> 25°C : 1
CO, SO2
<25°C : 1
> 25°C : 1
CO, NO2
<25°C : 0
> 25°C : 1
CO, 03
<25°C : 1
> 25°C : 1
398 (1.306-1.496); 0-2
189 (1.123-1.259); 0-2
257 (1.152-1.371); 0-2
149 (1.079-1.224); 0-2
396 (1.295-1.504); 0-2
241 (1.161-1.326); 0-2
973 (0.877-1.080); 0-2
.196 (1.104-1.297); 0-2
378 (1.286-1.477); 0-2
.170 (1.105-1.239); 0-2
March 2009
C-37
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Lin et al.
(1999)
Period of Study:
5/1991-4/1993
Location:
Sao Paulo, Brazil
ED Visits
Health Outcome (ICD9):
Respiratory illness (lower
respiratory illness, upper
respiratory illness, wheezing)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed: <13
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 5 ppm
Range (Min, Max): (1,12)
Copollutant: correlation
PM10: r = 0.50
N02: r = 0.35
S02: r = 0.56
O3: r= 0.04
Increment: NR
Relative Risk (Lower CI, Upper CI); 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
Author: Lin et al.
(2003)
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
Age Groups Analyzed: 6-12
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.18 (0.50) ppm
Range (Min, Max): (0, 610)
Copollutant: correlation
S02: r = 0.37
NO2: r = 0.55
03: r = -0.16
PM2.5: r= 0.45
PM10-2.5: r = 0.17
PM10: r = 0.38
Increment: 0 5 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Boys:
Adjusting for Daily Weather Variables
1.05(1-1.11); 1/1.07 (1.01-1.14); 2
1.08 (1.01-1.16); 3/1.08 (1-1.17); 4
1.07 (0.99-1.16); 5/1.07 (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); 1/1.08 (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); 1/1.01 (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)
0.99 (0.90-1.09); 3/0.97 (0.87-1.08)
0.99 (0.89-1.11); 5/1.02 (0.90-1.15)
1.05 (0.93-1.20); 7
Author: Lin et al.
(2004c)
Period of Study:
1/1987-12/1998
Location:
Vancouver, BC
Canada
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
GAM, LOESS
Age Groups Analyzed: 6-12
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.96 (0.52) ppm
Range (Min, Max): (0.23, 4.90)
Copollutant: correlation
SO2: r = 0.67
N02: r = 0.73
O3: r = -0.35
Increment: 0 5 ppm
Relative Risk (Lower CI, Upper CI); 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
Low SES:
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
Low SES:
1.01 (0.92-1.11); 1 / 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
March 2009
C-38
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Lin et al.
(2005)
Period of Study:
1998-2001
Location:
Toronto,
Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory Infections (464,
466, and 480-487)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: <15
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.16 (0.38) ppm
Range (Min, Max): (0.38, 2.45)
Copollutant: correlation
PM2.5: r= 0.10
PM10-2.5: r = 0.06
PM10: r = 0.10
S02: r = 0.12
N02: r = 0.20
03: r = -0.11
Increment: 0.44 ppm
Odds Ratio (Lower CI, Upper CI); Lag
Boys
No adjustment:
1.11 (1.01-1.22); 0-3/1.10 (1.00-1.22); 0-5
Adjustment for weather variables:
1.13 (1.03-1.24); 0-3/1.13 (1.02-1.25); 0-5
Adjustment for weather variables and PM:
1.08 (0.98-1.20); 0-3 /1.08 (0.97-1.20); 0-5
Girls
No adjustment:
0.99 (0.89-1.10); 0-3/1.00 (0.89-1.13); 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 (0.90-1.13); 0-3/1.02 (0.90-1.15); 0-5
Total
No adjustment:
1.06 (0.98-1.14); 0-3 /1.06 (0.98-1.15); 0-5
Adjustment for weather variables:
1.09 (1.01-1.17); 0-3/1.10 (1.01-1.19); 0-5
Adjustment for weather variables and PM:
1.05 (0.97-1.14); 0-3 /1.06 (0.97-1.15); 0-5
Author: Linn et al.
(2000)
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,
>30
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;
Os: 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;
O3: r = -0.36
Increment: 1.0 ppm
P(SE); lag:
Pulmonary
Age Group: > 30
AO 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
March 2009
C-39
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Luginaah et al.
(2005)
Period of Study:
4/1995-12/2000
Location:
Windsor, ON, Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory illness (460-519)
Study Design:
Time-series and Case-crossover
Statistical Analyses:
1. Time-series: Poisson
2. Case-crossover: conditional
logistic regression
Age Groups Analyzed:
AO ages
0-14
15-64
>65
Pollutant: CO
Averaging Time: 1-h maximum
Mean (SD) unit: 1.3 (1.0) ppm
Range (Min, Max): (0,11 82)
Copollutant: correlation
N02: r = 0.38
S02: r = 0.16
03: r= 0.10
CoH: r= 0.31
PM10: r = 0.21
Increment: 1 17 ppm
Relative Risk (Lower CI, Upper CI); Lag
Females and Case-crossover study design
Age Group: All ages:
1.037 (0.968-1.111); 1
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
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
1.007 (0.859-1.181); 2
1.032 (0.858-1.240); 3
Age Group: > 65:
1.014(0.922-1.116); 1
1.024(0.907-1.156); 2
1.035 (0.893-1.200); 3
Males and Case-crossover study design
Age Group: All Ages:
0.950 (0.884-1.020)
0.945 (0.862-1.036)
0.965 (0.866-1.075)
Age Group: 0-14:
1.003 (0.904-1.113); 1
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)
0.991 (0.760-1.293)
Age Group: > 65:
0.867 (0.775-0.970)
0.865 (0.752-0.994)
0.946 (0.807-1.109)
Female and Time-series study design
Age Group: All Ages:
1.049 (0.993-1.108); 1
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
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
1.025 (0.944-1.112); 2
1.081 (0.963-1.213); 3
Age Group: > 65:
1.029 (0.957-1.118); 1
1.030 (0.928-1.144); 2
1.013 (0.899-1.142); 3
Male and Time-series study design
Age Group: All Ages:
0.989 (0.932-1.049)
0.986 (0.946-1.029)
0.987 (0.929-1.048)
Age Group: 0-14:
1.034(0.949-1.126)
0.996 (0.933-1.062)
0.968 (0.881-1.064)
Age Group: 15-64:
0.994(0.854-1.157)
0.988 (0.884-1.104)
0.951 (0.806-1.121)
Age Group: > 65:
0.901 (0.817-0.994)
0.904(0.803-1.019)
0.963 (0.845-1.098)
March 2009
C-40
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Martins et al.
(2002)
Period of Study:
5/1996-9/1998
Location:
Sao Paulo, Brazil
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
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Mean (SD) unit: 3.7 (1.7) ppm
Range (Min, Max): (10,12 6)
Copollutant: correlation
N02: r = 0.62;
S02: r = 0.51;
PM10: r = 0.73;
O3: r= 0.07
Increment: 1 63 ppm
P(SE); lag:
Chronic Lower Respiratory Diseases
Age Group
>64:0.0489 (0.0274); 2
Author: Masjedi et al.
(2003)
Period of Study:
9/1997-2/1998
Location:
Tehran, Iran
ED Visits
Health Outcome (ICD9):
Total acute respiratory
conditions; Asthma (493);
COPD (490-492, 494, 496)
Study Design: Time-series
Statistical Analyses:
Multiple stepwise regression
Age Groups Analyzed: Adults
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 8.85 ppm
Range (Min, Max): (215, 23.8)
Copollutant: NR
Increment: NR
p (p-value); lag;
Asthma: -0.779 (0.12)
COPD: 0.012 (0.71)
Acute Respiratory conditions: -0.086 (0.400)
Correlation coefficients:
Mean 3-day CO levels and asthma: -0.300 (0.149)
Mean weekly CO level and asthma: -0.14 (0.2)
Mean 10-day CO levels and asthma: -0.05 (0.43)
Author: McGowan et
al. (2002)
Period of Study:
6/1988- 12/1998
Location:
Christchurch,
New Zealand
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-487); Acute
respiratory infections (460-466);
Chronic lung Diseases (491 -
492, 494-496); Asthma (493)
Study Design: Time-series
Statistical Analyses:
Generalized Additive Model
Age Groups Analyzed:
<15; >64
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.
Author: Migliaretti et al.
(2007)
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:
>15
15-64
>64
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 CI, Upper CI); 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(1.011-1.099)
March 2009
C-41
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Moolgavkar
(2000a)
Period of Study:
1987- 1995
Location:
3 U.S. counties:
Los Angeles County,CA
Cook County, IL
Maricopa County, AZ
Hospital Admissions
Health Outcome (ICD9):
COPD plus asthma (490-496)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
AO Ages
0-19
20-64
>65
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)
Maricopa: (269, 4777)
Copollutant: correlation
Cook County:
N02: r= 0.63; S02:r = 0.35;
03: r = -0.28
LA County:
N02: r= 0.80; S02:r = 0.78;
O3: r = -0.52
Maricopa County:
N02: r= 0.66; S02:r = 0.53;
03: r = -0.61
Increment: 1.0 ppm
% Increase (t-statistic); lag:
Age Group: > 65
Cook County
CO:
2.60 (1.9); 0;/3.00 (2.2); 1; /1.30 (1.0); 2;
1.40 (1.1); 3;/1.10 (0.8); 4;/2.30 (1.8); 5
Los Angeles County
CO:
5.40 (11.3); 0; / 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;
CO, PM10:
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, PMzs:
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(H); 4;/4.90 (3.8); 5
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); 4; / 4.80 (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, PMio-2s:
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(H); 5
Age Group: 20-64
Los Angeles County
CO:
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); 0;/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); 0;/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, PMio-2s:
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
Author: Moolgavkar
(2003b)
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:
AO Ages;
>65
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;
O3: r = -0.52
Increment: 1 ppm
% Increase (t-statistic); lag:
COPD - Los Angeles County
CO-
GAM-30 (108):
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 (108):
2.37 (8.67); 0;/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); 0; / 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 s):
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
March 2009
C-42
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Neidell et al.
(2004)
Period of Study:
1992-1998
Location:
California
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
Linear Regression
Age Groups Analyzed:
0-1
1-3
3-6
6-12
12-18
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.777 (1.037) ppm
Range (Min, Max): NR
Copollutant: correlation
03
PM10
N02
Increment: NR
P(SE); lag;
Single-pollutant model
Age Group
0-1:-0.007 (0.009);
1-3: 0.027 (0.009);
3-6:0.053 (0.010);
6-12: 0.047 (0.009);
12-18:0.025 (0.008)
Fixed effect controlling for03, PM10, and NO2
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 O3, PM10, NO2 and Avoidance
Behavior
Age Group
0-1:-0.010 (0.010);
1-3:0.027 (0.011);
3-6:0.051 (0.011);
6-12:0.025 (0.011);
12-18:0.021 (0.009)
Author: Norris et al.
ED Visits
Pollutant: CO
Increment: 0 6 ppm
(1999)
Health Outcome (ICD9):
Averaging Time: 24-h avg
Relative Risk (Lower CI, Upper CI); Lag
Period of Study:
9/1995- 12/1996
Location:
Seattle, WA
Asthma (493)
Mean (SD) unit: 1.6 (0.5) ppm
High Utilization: 1.04(0.93-1.16); 1
Study Design: Time-series
Range (Min, Max): (0 6, 41)
Low Utilization: 1.15 (1.05-1.28); 1
Statistical Analyses:
Semiparametric Poisson GAM
Age Groups Analyzed: <18
Copollutant: correlation
PM10: r = 0.74
NO2 (1-h max): r = 0.47
NO2 (24-h avg.): r= 0.66
SO2 (1-h max): r= 0.15
S02 (24-h avg.): r = 0.32
All: 1.10(1.02-1.19); 1
Author: Peel et al.
(2005)
Period of Study:
1/1993- 8/2000
Location:
Atlanta, GA
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
Analysis: 2-18
Pollutant: CO
Averaging Time: 1-h maximum
Mean (SD) unit: 1.8 (1.2) ppm
Range (10th, 90th): (0 5, 3 4)
Copollutant: NR
Increment: 10 ppm
Relative Risk (Lower CI, Upper CI); Lag
Health Condition
All respiratory illnesses: 1.011 (1.004-1.019); 0-2
URI:
1.012 (1.003-1.021); 0-2 /1.066 (1.045-1.087); 0-13
Asthma:
1.010 (0.999-1.022); 0-2
1.076 (1.047-1.105); 0-13
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
AOS (1/1/93-8/31/00): 1.011 (1.004-1.019); 0-2
AOS (8/1/98- 8/31/00): 1.010 (1.000-1.021); 0-2
ARIES (8/1/98- 8/31/00): 1.018 (1.003-1.033); 0-2
Author: Sheppard et
al. (1999)
Period of Study:
1987-1994
Location:
Seattle, WA
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed: <65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1831 ppb
IQR (25th, 75th): (1277, 2201)
Copollutant: correlation
PM10: r = 0.83; PM2.5: r = 0.78;
PM10-25: r = 0.56; O3: r = -0.18;
S02: r = 0.24
Increment: 924 ppb
% Increase (Lower CI, Upper CI); Lag
CO: 6% (3, 9); 3
CO, PM2.5: 5% (1,8); 3
March 2009
C-43
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Slaughter et al.
(2005)
Period of Study:
1/1995-6/2001
Location:
Spokane, WA
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:
Poisson GLM, Natural Splines
Age Groups Analyzed:
AO ages,
Adults
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (5th, 95th): (1 25, 3 05)
Copollutant: correlation
PMi: r = 0.63
PM25: r= 0.62
PM10: r = 0.32
PM10-2.5: r = 0.32
Increment: 10 ppm
Relative Risk (Lower CI, Upper CI); lag:
ED Visits
All Respiratory Illnesses
Age Group: All Ages:
0.99 (0.96-1.02); 1 /1.01 (0.98-1.04); 2
1.03 (1.00-1.06); 3
Asthma
Age Group: All Ages:
1.00 (0.95-1.06); 1 /1.01 (0.96-1.07); 2
1.06 (1.00-1.11); 3
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
Hospital Admissions:
All Respiratory Illnesses
Age Group: All 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: All 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
Author: Steib et al.
(2007)
Period of Study:
7/1992- 3/1996
Location:
Saint John,
Canada
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
Pollutant: CO
Averaging Time:
24-h avg
1-h maximum
Mean (SD) unit:
All year: 0.5 (0.3) ppm
May-September: 0.6 (0.3) ppm
All year: 1.6 (1.1) ppm,
May-September: 1.7 (0.9) ppm
Range (Min, Max): NR
Copollutant: correlation
H2S: r= -0.10; NO2:r = 0.68;
03: r = -0.05; S02: r = 0.31;
TRS: r= 0.07; PM10: r= 0.28;
PM25: r= 0.27; Hr r = 0.23;
S042-: r = 0.27; CoH:r= 0.55
Increment: 0.5 & 1.7 ppm
% Increase (Lower CI, Upper CI); lag:
All Respiratory Illnesses
Increment: 0 5 ppm
All Year: -3.40; 7
Increment: 1 7 ppm
May- September: -5.70
Author: Sun et al
(2006)
Period of Study:
1/2004- 12/2004
Location:
Taiwan
ED Visits
Health Outcome (ICD9):
Asthma (493)
Study Design: Cross-sectional
Statistical Analyses:
Pearson correlation analysis
Age Groups Analyzed:
<16; 16-55
Pollutant: CO
Averaging Time: Monthly
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Correlation Coefficient:
Asthma
Age Group:
<16:0.653
16-55:0.425
March 2009
C-44
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Tenias et al.
(2002)
Period of Study:
1/1994- 12/1995
Location:
Valencia, Spain
ED Visits
Health Outcome (ICD9):
COPD (491,492,494, 496)
Study Design: Time-series
Statistical Analyses:
1. Poisson autoregressive
2. Sensitivity: GAM, LOESS
Age Groups Analyzed: >14
Pollutant: CO
Averaging Time:
24-h avg
1-h maximum
Mean (SD) unit:
24-h avg
All year: 3.1 mg/m3
Warm Months: 2.5 mg/m3
Cold Months: 3.7 mg/m3
1-h avg
All year: 6.7 mg/m3
Warm Months: 5.4 mg/m3
Cold Months: 8.0 mg/m3
Range (Min, Max):
24-h avg: (0.9, 7.1)
1-h maximum: (1.6,17.2)
Copollutant: correlation
S02:r = 0.734; N02:r = 0.180;
03:r = -0.517
Increment: 1 mg/m3
Relative Risk (Lower CI, Upper CI); Lag
24-h avg
All Year: 1.074 (0.998- 1156); 1
Cold Months: 1.070 (0.991-1.156); 1
Warm Months: 1.129 (0.960-1.329); 1
1-h maximum
All Year: 1.039 (1.014-1.066); 1
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); 1
All Year: GAM, LOESS:
1.042(1.019-1.066); 1
Author: Thompson et
al. (2001)
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;
O3: r = -0.52; NOx (log): r = 0.74;
NO (log): r= 0.71; N02: r = 0.69
Increment: NR
Relative Risk (Lower CI, Upper CI); 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.
Author: Tolbert et al.
(2007)
Period of Study:
1/1993- 12/2004
Location:
Atlanta, GA
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
Pollutant: CO
Averaging Time: 1-h maximum
Mean (SD) unit: 16 ppm
Range (Min, Max): (01,7 7)
Copollutant: correlation
PM10: r = 0.51; 03: r = 0.27;
N02: r= 0.70; S02:r = 0.28;
Coarse PM: r= 0.38; PM2s: r= 0.47;
S04: r = 0.14;EC: r = 0.66;
OC: r = 0.59; TC:r= 0.63;
OHC: r= 0.29
Increment: 1 22 ppm
Relative Risk (Lower CI, Upper CI); lag:
Respiratory Diseases: 1.016 (1.009-1.022); 3
Note: The study only provides results of the multi-pollutant
models in figures, not quantitatively.
Author: Trapasso and
Keith (1999)
Period of Study:
1/1994- 12/1994
Location:
Bowling Green, KY
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Time-series
Statistical Analyses:
Spearman Rank Correlation
Coefficient
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: NR
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Correlation Coefficient (lag)
CO Mean: r = 0.19
CO Mean: r = 0.27
CO Mean: r = 0.21
CO Max:
CO Max
CO Max
r = 0.26; 0
r = 0.36; 1
r = 0.24; 2
March 2009
C-45
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Tsai et al.
(2006a)
Period of Study:
1996-2003
Location:
Kaohsiung, Taiwan
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0 77 ppm
Range (Min, Max): (0 23,1 72)
Co pollutant:
PM10
S02
NO2
03
Increment: 0 29 ppm
Odds Ratio (Lower CI, Upper CI); Lag
OR for getting asthma and exposure to various pollutants for
all ages at either <25°C or a 25°C
CO
<25°C : 1.414(1.300-1.537); 0-2
> 25°C : 1.222 (1.138-1.312); 0-2
CO, PM10
<25°C : 1.251 (1.125-1.393); 0-2
>25°C: 1.178 (1.088-1.274); 0-2
CO, S02
<25°C : 1.207 (1.076-1.354); 0-2
>25°C: 1.290 (1.188-1.400); 0-2
CO, N02
<25°C : 0.916 (0.807-1.039); 0-2
>25°C: 1.249 (1.127-1.384); 0-2
CO, 03
<25°C : 1.396 (1.282-1.520); 0-2
>25°C: 1.195 (1.113-1.284); 0-2
Author: Vigotti et al.
(2007)
Period of Study:
1/2000- 12/2000
Location:
Pisa, Italy
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: <10;
>65
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
Increment: 1mg/m3
% Increase (Lower CI, Upper CI); Lag
Age Group
<10:18.60% (-6.90 to 51.10); 1
>65: 26.50% (3.40-54.80); 4
Author: Villeneuve et
al. (2006b)
Period of Study:
1995-2000
Location:
Toronto, ON,
Canada
Physician Visits
Health Outcome (ICD9):
Allergic rhinitis (177)
Study Design: Time-series
Statistical Analyses: Poisson
GLM
Age Groups Analyzed: >65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.1 (0.4) ppm
Range (Min, Max): (0.0,2.2)
Copollutant:
PM2.5
PM10
PM10-2.5
S02
N02
03
Increment: 0 4 ppm
Odds Ratio (Lower CI, Upper CI); Lag
The study did not present quantitative results for CO.
Author: Xirasagar et al.
(2006)
Period of Study:
1998-2001
Location:
Taiwan
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Cross-sectional
Statistical Analyses:
Spearman Rank Correlations
Age Groups Analyzed:
0-14; <2; 2-5; >5
Pollutant: CO
Averaging Time: Monthly
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Correlation Coefficient (Lag)
Age Group:
<2: r= -0.208
2-5: r = -0.281
>5: r= -0.134
Author: Yang et al.
(2007)
Period of Study:
1996-2003
Location:
Taipei,
Taiwan
Hospital Admissions
Health Outcome (ICD9):
Asthma (493)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
AO ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 33 ppm
Range (Min, Max): (0.32, 3.62)
Copollutant:
PM10
S02
N02
03
Increment: 0.53 ppm
Odds Ratio (Lower CI, Upper CI); Lag
CO
<25°C : 1
> 25°C : 1
CO, PM10
<25°C : 1
> 25°C : 1
CO, S02
<25°C : 1
> 25°C : 1
CO, N02
<25°C : 0
> 25°C : 1
C0,03
<25°C : 1
> 25°C : 1
076 (1.019-1.136); 0-2
277 (1.179-1.383); 0-2
050 (0.983-1.122); 0-2
332 (1.216-1.459); 0-2
131 (1.059-1.207); 0-2
278 (1.174-1.392); 0-2
915 (0.839-0.997); 0-2
.177 (1.049-1.320); 0-2
169 (1.102-1.240); 0-2
.275 (1.177-1.382); 0-2
March 2009
C-46
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Yang et al.
(2007)
Period of Study:
1996-2003
Location:
Taipei, Taiwan
Hospital Admissions
Health Outcome (ICD9):
COPD: (490-492, 494, 496)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 33 ppm
Range (Min, Max):
(0.32, 3.66) ppm
Co pollutant:
PM10
S02
NO2
03
Increment: 0.53 ppm
Odds Ratio (Lower CI, Upper CI); Lag
CO
<20°C: 0.975 (0.921,1.033); 0-2
>20°C: 1.227 (1.178-1.277); 0-2
CO, PM10
<20°C: 0.925 (0.863-0.992); 0-2
>20°C: 1.177 (1.123-1.235); 0-2
CO, SO2
<20°C: 0.895 (0.832-0.962); 0-2
>20°C: 1.274 (1.219-1.331); 0-2
CO, NO2
<20°C: 1.000 (0.910-1.099); 0-2
>20°C: 1.061 (0.998-1.129); 0-2
CO, 03
<20°C: 0.935 (0.875-0.999); 0-2
>20°C: 1.234 (1.185-1.285); 0-2
Author: Yang et al.
(2005)
Period of Study:
1/1994- 12/1998
Location:
Vancouver,
Canada
Hospital Admissions
Health Outcome (ICD9):
COPD (490-492, 494, 496)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: .71 (0.28) ppm
Range (Min, Max): (0.30, 2.48)
Copollutant: correlation
O3: r = -0.56
N02: r = 0.73
S02: r = 0.67
PM10: r = 0.50
Increment: 0 3 ppm
Relative Risk (Lower CI, Upper CI); 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 (1.02-1.12); 0-5
1.08 (1.02-1.13); 0-6
Multi pollutant:
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
Author: Yang et al.
(2003)
Period of Study:
1/1986- 12/1998
Location:
Vancouver, BC,
Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory diseases (460-519)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
<3;
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.98 (0.54) ppm
IQR (25th, 75th): (0 62,1.16)
Copollutant: correlation
O3: r = -0.52
CoH
N02
S02
Increment: 0 54 ppm
Odds Ratio (Lower CI, Upper CI); 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); 1
CO, 03: 1.04(1.01-1.07); 1
CO, 03, CoH, N02, S02: 1.02 (0.96-1.08); 1
Age Group: > 65
CO alone: 1.02 (1.00-1.04); 1
CO, 03: 1.02 (1.00-1.04); 1
CO, 03, CoH, N02, S02: 0.96 (0.93-1.00); 1
Author: Yang et al.
(2004b)
Period of Study:
6/1/1995-3/31/1999
Location:
Vancouver,
Canada
Hospital Admissions
Health Outcome (ICD9):
Respiratory diseases (460-519);
Pneumonia (480-486); Asthma
(493)
Study Design: Case-control
Statistical Analyses:
Pearson's correlation coefficient
Age Groups Analyzed: <3
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; PM2.5: r = 0.24;
PM10-25: r = 0.33; O3: r = -0.53;
N02: r= 0.74; S02: r = 0.61
This study did not present quantitative results for CO.
Author: Zanobetti and
Schwartz (2006)
Period of Study:
1995-1999
Location:
Boston, MA
Hospital Admissions
Health Outcome (ICD9):
Pneumonia (480-487)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All
ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
IQR (25th, 75th): (0.39, 0.60)
Copollutant: correlation
PM25: r= 0.52; BC:r=0.82;
N02: r = 0.67; 03:r = -0.30
Increment: 0.475 ppm
% Increase (Lower CI, Upper CI); lag:
5.45 (1.10, 9.51); 0
5.12 (0.83, 9.16); 0-1
March 2009
C-47
DRAFT-DO NOT QUOTE OR CITE
-------
Table C-6. Studies of long-term CO exposure and respiratory morbidity.
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Goss et al
(2004)
Period of Study:
1999-2000
Location:
U.S.
Health Outcome: Lung function
(FEVi, Cystic fibrosis pulmonary
exacerbation)
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population:
11,484 cystic fibrosis patients
Age Groups Analyzed: >6
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 CI, Upper CI); lag:
Two or More Pulmonary Exacerbations During 2000
1.02 (0.85-1.22)
Author: Guo et al
(1999)
Period of Study:
10/1995-5/1996
Location:
Taiwan
Health Outcome: Asthma
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population:
331,686 non-smoking children
Age Groups Analyzed:
Middle-school children
(mean age: 13.8)
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 853 (277) ppb
Range (Min, Max): (381,1610)
Copollutant: NR
Increment: 326 ppb
% Increase (Lower CI, Upper CI); lag:
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)
Author: Hirsch et al.
(1999)
Period of Study:
Population:
9/1995-6/1996
Air:
4/1994-4/1995
Location:
Dresden, Germany
Health Outcome:
Asthma symptoms in the past 12
months (wheeze, morning cough);
Doctor's diagnosis (asthma,
bronchitis); Lung function (bronchial
hyperresponsiveness (BHR), FEVi
<85% pred., FEF25-75% <70% pred.)
Study Design: Cross-sectional
Statistical Analyses:
Multiple logistic regression
Population:
5-7: 2,796; 9-11: 2,625
Age Groups Analyzed:
5-7 and 9-11
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 |jg/m3
Prevalence Odds Ratio (Lower CI, Upper CI); lag:
Symptoms in the past 12 months: 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; 9-11:112 (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
Age Groups: 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: FEVi <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 months: Wheeze
Age Groups: 5-7; 9-11
Atopic children: 1 (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: 1 (0.86-1.16)
Nonatopic children: 1.21 (1.1-1.33)
Notes: Atopic Children were defined as those children
_with_s£ecificj£E_Jo_aeroa]ler2ens_>0;^J
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Children were defined as those children with specific IgE
to aeroallergens s 0.7 kll-L-1.
Author: Hwang et al.
(2006)
Period of Study:
2001
Location:
Taiwan
Health Outcome:
Allergic rhinitis
Study Design: Cross-sectional
Statistical Analyses:
Two-stage hierarchical model
(logistic and linear regression)
Population:
32,143 Taiwanese school children
Age Groups Analyzed: 6-15
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit: 664 (153) ppb
Range (Min, Max): (416, 964)
Copollutant: correlation
NOx: r = 0.88
03: r= -0.37
PM10: r= 0.27
S02: r= 0.40
Increment: 100 ppb
Adjusted Odds Ratio (Lower CI, Upper CI); 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, Os: 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: a 12:1.06 (1.03-1.09)
ETS: Yes: 1.06 (1.03-1.08); ETS: No: 1.05 (1.02-1 .C
Visible Mold: Yes: 1.07(1.03-1.11)
Visible Mold: No: 1.05 (1.03-1.07)
Author: Hwang et al.
(2005)
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
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 664 (153) ppb
Range (Min, Max): (416, 964)
Copollutant: correlation
NOx: r = 0.88
03: r= -0.37
PM10: r= 0.27
S02: r= 0.40
Increment: 100 ppb
Adjusted Odds Ratio (Lower CI, Upper CI); 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, PMio,03: 1.119 (1.084-1.155)
Male: 1.49 (1.37-1.63); Female: 1
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: a 12: 2.43 (1.9-3.09)
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
Health Outcome:
Pollutant: CO
The study did not present quantitative results for CO.
(2003c)
Allergic rhinitis
Averaging Time:
Period of Study:
Study Design: Cohort
Annual avg
10/1995-5/1996
Statistical Analyses:
Mean (SD) unit: 853 (277) ppb
Location:
Taiwan
Multiple logistic regression
Range (Min, Max): (381,1610)
Population:
331,686 non-smoking children
Age Groups Analyzed: 12-14
Copollutant: NR
Author: Meng et al.
(2007)
Period of Study:
11/2000-9/2001
Location:
Los Angeles County
and San Diego County,
California
Health Outcome: Asthma
Study Design: Cohort
Statistical Analyses:
Logistic regression
Population:
1,609 physician-diagnosed
asthmatics
Age Groups Analyzed: a 18
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; PM2.5: r = 0.52;
N02: r = 0.55
The study did not present quantitative results for CO.
March 2009
C-49
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Mortimer et al.
(2008)
Period of Study:
1989-2000
Location:
San Joaquin Valley,
CA
Health Outcome: Lung function
(FVC, FEVi, PEF, FEF25-75,
FEVi/FVC, FEF25-75/FVC, FEF25,
FEF75)
Study Design: Cohort
Statistical Analyses:
1. DSA algorithm
2. GEE
Population: 232 asthmatic children
Age Groups Analyzed: 6-11
Pollutant: CO
Averaging Time:
8-h maximum monthly mean
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant; correlation:
Lifetime
NO2 (24-h avg): r= 0.68
O3 (8-h maximum): r = -0.40
PM10 (24-h avg): r = 0.05
Prenatal
CO (8-h maximum): r = 0.52
NO2 (24-h avg): r= 0.37
O3 (8-h maximum): r = -0.16
PM10 (24-h avg): r = -0.05
Increment: NR
Effect Size per IQR Increase in Pollutant (SE):
FEF2575:
24-h avg CO exposure during 1st trimester
0.90% (0.0113)
FEV1/FVC
Daily maximum CO exposure during ages 0 to 3
-2.50% (0.0016)
FEF2575/FVC
24-h avg CO exposure during ages 0 to 6 and diagnosed
with asthma <2 years old
-4.80% (0.0446)
FEF25
24-h avg CO exposure during ages 0 to 6 and diagnosed
with asthma <2 years 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)
FEF2575
Lifetime 24-h avg CO exposure
-0.94454 (0.3975)
FEF25-75/FVC
-0.1090 (0.0303)
FEV1/FVC
Prenatal 8-h maximum CO exposure: 0.1711 (0.0653)
Lifetime 1-h maximum CO exposure: -0.3242 (0.0919)
24-h avg CO exposure during ages 0-3 and diagnosed
with asthma <2 years old: -0.1814 (0.0599)
FEF25
24-h avg CO exposure during ages 0-6 and diagnosed
with asthma <2 years old: -1.0460 (0.1953)
FEF75
Lifetime 8-h maximum CO exposure: -0.4214 (0.1423)
Author: Singh et al.
(2003)
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
Pollutant: CO
Averaging Time:
Annual avg
Mean (SD) unit:
Roadside: 3,175 |jg/m3
Campus: 2,150 |jg/m3
Range (Min, Max): NR
Copollutant: NR
The study did not present quantitative results for CO.
Author: Wang et al.
(1999)
Period of Study:
10/1995-6/1996
Location:
Kaohsiung and
Pintong, Taiwan
Health Outcome: Asthma
Study Design: Cross-sectional
Statistical Analyses:
Multiple logistic regression
Population:
165,173 high school students
Age Groups Analyzed: 11-16
Pollutant: CO
Averaging Time: Annual median
Median (SD) unit: 0 80 ppm
Range (Min, Max): NR
Copollutant: NR
Increment: NR
Adjusted Odds Ratio (Lower CI, Upper CI); lag:
CO Concentrations: <0.80 ppm: 1.0
CO Concentrations a 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)
March 2009
C-50
DRAFT-DO NOT QUOTE OR CITE
-------
Table C-7. Studies of short-term CO exposure and mortality.
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Anderson et al.
(2001)
Period of Study: 10/1994 -
12/1996
Location:
West Midlands,
United Kingdom
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit: 0.8 (0.7) ppm
Range (Min, Max): (0.2,10.0)
Copollutant correlation:
PM10: r= 0.55; PM25:r = 0.54;
PM10-25: r = 0.10; BS: r= 0.77;
S042-: r = 0.17; N02: r= 0.73;
Os: r= -0.29; S02:r = 0.49
Increment: 1 0 ppm
% Increase (Lower CI, Upper CI); lag:
All-cause
0.8% (-0.6 to 2.2); 0-1
Cardiovascular
2.5% (0.4-4.6); 0-1
Respiratory
1.2% (-2.1 to 4.6); 0-1
Author: Bellini et al (2007) Health Outcome (ICD9):
Period of Study: 1996 - 2002 M°^A'^se (non
Location:
15 Italian cities
accidental) (
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Meta-analysis
Statistical Analyses:
Poisson GLM
Age Groups Analyzed: All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
S02
N02
03
PM10
Increment: 1 mg/m3
% Increase (Lower CI, Upper CI); lag:
All-cause
1.19% (0.61-1.72); 0-1
Respiratory
0.66% (-1.46 to 2.88); 0-1
Cardiovascular
0.93% (-0.10 to 1.77); 0-1
Author: Biggeri et al. (2005)
Period of Study: 1990-1999
Location:
8 Italian Cities (Turin, Milan,
Verona, Bologna, Ravenna,
Florence, Rome, and
Palermo)
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular (390-459);
Respiratory (460-519); Cardio-
respiratory
Study Design: Meta-analysis
Statistical Analyses: Poisson
GLM, cubic splines
Age Groups Analyzed: All ages
Pollutant: CO
Averaging Time:
Maximum 8-h moving average
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)
Milan, 1995-1997: (NR, 12.3)
Verona, 1995-1999: (NR, 10.2)
Ravenna, 1991-1995: (NR, 7.0)
Bologna, 1996-1998: (NR, 11.1)
Florence, 1996-1998: (NR, 8.7)
Rome, 1992-1994: (NR, 22.3)
Rome, 1995-1997: (NR, 18.5)
Palermo, 1997- 1999: (NR, 8.0)
Copollutant: NR
Increment: 1.0 mg/m3
% Increase (Lower CI, Upper CI); 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: Botter et al. (2002)
Period of Study: 1991-1993
Location:
Sao Paulo, Brazil
Health Outcome (ICD9):
Mortality
Study Design:
Longitudinal study
Statistical Analyses:
State space model
Age Groups Analyzed: a 65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: TSP; NO2; O3; SO2
Increment: NR
P (SE):
Model 1
Model 2
Model 3
Model 4
0.0053 (0.0036)
0.0046 (0.0028)
0.0040 (0.0028)
0.0032 (0.0028)
March 2009
C-51
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Bremner et al. (1999) Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (
Period of Study:
1/1992-12/1994
Location:
London, U.K.
Cardiovascular (390-459);
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Poisson,
cubic splines
Age Groups Analyzed:
All aqes
0-64
>65
65-74
>75
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.4) ppm
Range (Min, Max): (0.2, 5.6)
Copollutant:
N02;
Os;
S02;
PM10;
BS
Increment: 0 8 ppm
% Increase (Lower CI, Upper CI); 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
a 75: 2.3% (-0.5 to 5.3); 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); 1
CO, 03: 3.98% (0.85-7.21); 1
CO, PM10: 0.62% (-0.59 to 1.85); 1
CO, BS: 1.29% (-1.53 to 4.19); 1
Author: Burnett et al. (2000)
Period of Study: 1986-1996
Location:
8 Canadian cities
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800)
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
O3: r= -0.05
S02: r= 0.42
PM2.5: r = 0.44
PM10-2 5: r = 0.29
PM10: r= 0.45
Increment: 0 9 ppm
% Increase (t-value); lag:
Temporally filtered daily non-accidental mortality (days
in which PM10 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, PM10-2.5:0.1 (0.2); 0; 1.8 (2.1); 1
CO, PM10: -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, NO2, S02: 0.7 (1.9)
Model 2: CO, S04, Ni, Fe, Zn, 03, N02: 0.7 (1.7)
Author: Burnett et al (2004) Health Outcome (ICD9):
Period of Study: 1981 - 1999 Mor'allty: ^-c®uns® (non"
' accidental) (<800)
Location:
12 Canadian cities
Study Design: Time-series
Statistical Analyses:
1. Poisson, natural splines
2. Random effects regression
model
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 02 ppm
Range (Min, Max): NR
Copollutant:
N02;
Os;
S02;
PM2.5;
PM10-2.5
Increment: 1 02 ppm
% Increase (t-value); lag:
0.68% (3.12); 1
CO, N02: 0.07% (0.30); 1
March 2009
C-52
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Cakmak et al
Period of Study:
1/1997- 12/2003
Location:
Chile - 7 cities
(2007) Health Outcome (ICD9):
Mortality: All-cause (non-
accidental) (<800);
Cardiovascular diseases (390-
459); Respiratory diseases (460-
519)
Study Design: Time-series
Statistical Analyses: Poisson;
Random effects regression
model
Age Groups Analyzed:
All aqes
<64
65-74
75-84
>85
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 29 ppm
Range (Min, Max): NR
Copollutant correlation:
O3: 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.
Increment: 1 29 ppm
% Increase (t-value); lag:
Non-accidental:
5.88% (6.42); 1; 9.39% (6.89); 0-5
CO+PM10+O3+SO2: 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
March 2009
C-53
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Chock et al. (2000)
Period of Study:
1989- 1991
Location:
Pittsburgh, PA
Health Outcome (ICD9):
Mortality: Respiratory (480-486,
490-496, 507); Cardiovascular
(390-448); Influenza (487)
Study Design: Time-series
Statistical Analyses:
Poisson GAM; Cubic B-spline
basis functions
Age Groups Analyzed:
All ages
<75
>75
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Co pollutant:
PM10;
PM25;
Os;
S02;
N02
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;
-0.00219 (-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: 0.01655 (1.90); 0
Summer: 0.00155 (0.14); 0
Fall: 0.00797 (1.14); 0
CO, PM10
Winter: -0.00563 (-0.50); 0
Spring: 0.01233 (0.99); 0
Summer: -0.00712 (-0.48); 0
Fall: 0.00661 (0.73); 0
CO, PM10, 03, S02, N02
Winter: -0.01326 (-0.95); 0
Spring: 0.02501 (1.54); 0
Summer: 0.01874 (0.92); 0
Fall: 0.01011 (0.88); 0
Age Group:>75
CO Alone:-0.0035 (-0.67); 0
CO, PM10: -0.0104 (-1.67); 0
CO, PM10, N02:-0.0128 (-1.80); 0
CO, PM10, 03, S02, N02: -0.0144 (-1.99); 0
CO
-0.00025 (-0.05); -3; / -0.00242 (-0.42); -2;
-0.00238 (-0.41); -1; / -0.00302 (-0.54); 0;
-0.00116 (-0.20); 1; /-0.00508 (-0.88); 2;
-0.00251 (-0.48); 3
CO, PM10, 03, S02, N02
-0.00123 (-0.17); -3; / -0.00876 (-1.13); -2;
-0.00682 (-0.88); -1; / -0.01248 (-1.66); 0;
-0.00672 (-0.86); 1;/-0.00181 (-0.23); 2;
-0.00515 (-0.69); 3
Season
CO
Winter: -0.00304 (-0.43); 0
Spring: 0.00482 (0.54); 0
Summer: 0.01178 (1.07); 0
Fall:-0.01011 (-1.43); 0
CO, PM10
Winter: -0.02303 (-2.03); 0
Spring: -0.00517 (-0.40); 0
Summer: 0.00735 (0.50); 0
Fall:-0.01042 (-1.14); 0
CO, PM10, 03, S02, N02
Winter: -0.03370 (-2.41); 0
Spring: -0.00652 (-0.39); 0
Summer: 0.01258 (0.61); 0
Fall:-0.01250 (-1.07); 0
March 2009
C-54
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Cifuentes et al.
(2000)
Period of Study:
1988- 1996
Location:
Santiago, Chile
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 2.5 ppb
Range (5th, 95th): (0.6, 6.2)
Health Outcome (ICD9):
Mortality: All causes (non-
accidental) (<800)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, GAM with filtered variables Copollutant correlation:
& GLM r = °-80
Age Groups Analyzed: SO^ r = 0 62 ^
Alla9es N02: r = 0 65
03: r= -0.01
Increment:
All year: 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, SO2:1.038 (6.0); 0-1
CO, N02:1.026 (3.9); 0-1
CO, 03:1.036 (4.8); 0-1
Winter
CO: 1.052 (5.9); 0-1
CO, PM25:1.025 (2.1); 0-1
CO, PM10-2.5:1.049 (4.3); 0-1
CO, SO2:1.049 (5.0); 0-1
CO, N02:1.027 (2.6); 0-1
CO, Os: 1.051 (4.4); 0-1
Summer
CO: 1.053 (6.0); 0-1
CO, PM2.5:1.053 (5.3); 0-1
CO, PM10-25:1.053 (5.3); 0-1
CO, SO2:1.050 (5.2); 0-1
CO, NO2:1.047 (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, SO2, NO2, O3:
1.032 (4.6); 0-1
GAM Filtered Variables
CO: 1.030 (4.3); 0-1
CO, PM25, PM10-25, SO2, NO2, O3:
1.022 (2.4); 0-1
GLM
CO: 1.023 (2.4); 0-1
CO, PM25, PM10-25, SO2, NO2, O3:
1.013 (1.1); 0-1
Author: Conceicao et al.
Health Outcome (ICD9):
Pollutant: CO
Increment: NR
(2001)
Mortality: Respiratory diseases
(460-519)
Averaging Time:
P (SE); lag:
Period of Study:
Maximum 8-h moving avg
CO: 0.0306 (0.0076); 2
1994- 1997
Study Design: Time-series
Mean (SD) unit:
CO,S02, PM10, 03: 0.0259 (0.0116); 2
Location:
Statistical Analyses: Poisson
Total: 4.4 (2.2) ppm
Model 1: Pollutant concentration:
Sao Paulo, Brazil
GAM
1994: 5.1 (2.4) ppm
0.0827 (0.0077); 2
Age Groups Analyzed:
1995: 5.1 (2.4) ppm
1996: 3.9 (2.0) ppm
1997: 3.7 (1.6) ppm
Range (Min, Max): NR
Copollutant:
PM10; SO2; O3
Model 2:1+loess(time):
<5
0.0285 (0.0074); 2
Model 3: 2+loess(temperature)+humidity:
0.0309 (0.0076); 2
Model 4: 3+nonrespiratory counts:
0.0306 (0.0076); 2
Model 5: 4+autoregressive parameters:
0.0292 (0.0118); 2
Author: De Leon et al.
Period of Study:
1/1985- 12/1994
Location:
New York, NY
(2003) Health Outcome (ICD9):
Mortality: Circulatory (390-459);
Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
<75
>75
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 2.45 ppm
IQR (25th, 75th): (1 80,2 97)
Copollutant:
PM10;
Os;
S02;
N02
The study did not present quantitative results for
CO.
March 2009
C-55
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Dominici et al.
(2003b)
Period of Study:
1987- 1994
Location:
90 U.S. cities (NMMAPS)
Health Outcome (ICD9):
Mortality: All-cause (non-
accidental); Cardiovascular;
Respiratory
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Co pollutant:
03; N02; S02; CO
Increment: 1 ppm
% Increase (Lower CI, Upper CI); Lag
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: Fairley et al. (1999)
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-h avg; Maximum 8-h avg
Median (SD) unit:
24-h avg: 1.4 (1.0) ppm
Maximum 8-h avg: 2.1 (1.6) ppm
Range (Min, Max):
24-h avg: (0.0, 7.6)
Maximum 8-h avg: (0.2, 2.5)
Copollutant: correlation
PMio:r= 0.609;
PM2.5: r = 0.435;
PM10-25: r = 0.326;
COH:r = 0.736;
N03:r = 0.270;
S04: r = 0.146; 03: r = -0.215
Increment: 2 2 ppm
Relative Risk (Lower CI, Upper CI); lag:
1980-1986
CO: 1.04; 0;
CO: 1.05; 1;
CO.COH: 1.00; 1;
CO, N03:1.03;
CO, N03,03,C0H: 1.00
1989-1996
CO: 1.02; 0;
CO: 1.04; 1;
CO, PM25: 0.98;
CO, N03:1.01;
CO, N02, 03, N03:1.06
Respiratory mortality: CO: 1 08; 1
Cardiovascular mortality: CO: 1 04; 1
Author: Fischer et al. (2003)
Period of Study:
1986- 1994
Location: The Netherlands
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Pneumonia
(480-486); COPD (490-496);
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
<45
45-64
65-74
>75
Pollutant: CO
Averaging Time: 24-h avg
Median (SD) unit: 406 |jg/m3
Range (Min, Max): (174, 2620)
Copollutant:
PM10; BS; 03; N02; SO2
Increment: 1,200 |jg/m3
Relative Risk (Lower CI, Upper CI); 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
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)
Period of Study:
1998-2000
Location:
Rome, Italy
Health Outcome (ICD9):
Mortality: IHD (410-414)
Study Design:
Time-stratified case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: >35
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 2.4 (1.0) mg/m3
IQR (25th, 75th): (1 7, 2 9)
Copollutant correlation:
PNC: r = 0.89; PM10: r = 0.34;
N02: r = 0.54; S02:r = 0.52;
03: r = 0.01
Increment: 12 mg/m3
% Increase (Lower CI, Upper CI); lag:
6.5% (1.0-12.3); 0
4.7% (-0.9 to 10.7); 1
2.6% (-3.0 to 8.5); 2
-0.1% (-5.5 to 5.5); 3
7.0% (0.8-13.7); 0-1
March 2009
C-56
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Forastiere et al.
(2007)
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
IQR (25th, 75th): NR
Co pollutant:
PM10; PM2.5; NOx; Benzene
This study did not present quantitative results for
CO.
Author: Goldberg et al.
(2001)
Period of Study:
1984- 1993
Location: Montreal, Quebec,
Canada
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:
<65; > 65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (01, 51)
Co pollutant:
TSP; PM10; PM25; Sulfates; COH; S02;
N02; NO; 03
The study did not present quantitative results for
CO
Author: Goldberg et al Health Outcome (ICD9):
(2003) Mortality: CHF (428)
Period of Study: 1984- 1993 Study Design: Time-series
Location: Statistical Analyses: Poisson
Montreal, Quebec, Canada GLM, natural splines
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (01, 51)
Co pollutant:
PM2.5; Sulfate; SO2; NO2; O3
Increment: 0 50 ppm
% Increase (Lower CI, Upper CI); lag:
Daily mortality from CHF
-0.99% (-6.31 to 4.63); 0
0.12% (-5.29 to 5.84); 1
-1.38% (-8.81 to 6.66); 0-2
Daily mortality among persons classified as having
CHF before death
2.10% (-0.24 to 4.49)
2.28% (-0.09 to 4.72)
2.86% (-0.46 to 6.29)
0
1
0-2
Author: Goldberg et al.
(2006)
Period of Study:
1984- 1993
Location:
Montreal, Quebec, Canada
Health Outcome (ICD9):
Mortality: Diabetes (250)
Study Design: Time-series
Statistical Analyses: Poisson,
natural splines
Age Groups Analyzed: a 65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 0.8 (0.5) ppm
Range (Min, Max): (01, 51)
Co pollutant:
PM2.5;
Sulfate;
S02;
NO2;
03
Increment: 0 50 ppm
% Increase (Lower CI, Upper CI); lag:
Daily mortality from diabetes
2.64% (-2.56 to 8.12); 0
6.54% (1.31-12.03); 1
8.08% (1.03-15.62); 0-2
Daily mortality among persons classified as having
diabetes before death
1.15% (-1.69 to 4.07); 0
1.30% (-1.58 to 4.27); 1
2.63% (-1.42 to 6.85); 0-2
Author: Gouveia et al.
(2000b)
Period of Study:
1991 - 1993
Location:
Sao Paulo, Brazil
Health Outcome (ICD9):
Mortality: Respiratory;
Cardiovascular; All other causes
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
All ages
>65
<5
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit: 5.8 (2.1) ppm
Range (Min, Max): (13,16 2)
Co pollutant:
PM10; SO2; NO2; O3
Increment: 51 ppm
Relative Risk (Lower CI, Upper CI); 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
March 2009
C-57
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Gwynn et al. (2000)
Period of Study:
5/1988- 10/1990
Location:
Buffalo, NY
Health Outcome (ICD9):
Mortality: Respiratory (466, 480-
486); Circulatory (401-405, 410-
414, 415-417); All non-accidental
causes (<800)
Study Design: Time-series
Statistical Analyses:
Poisson GLM
Age Groups Analyzed: All ages
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;
NO2: r = 0.65
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
Author: Hoek et al. (2001)
Period of Study:
1986-1994
Location: The Netherlands
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
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant:
03; BS; PM10; S02; N02
Increment: 120 |jg/m3
Relative Risk (Lower CI, Upper CI); 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: 1.109 (1.012-1.216); 0-6
Cerebrovascular mortality:
1.066 (1.029-1.104); 0-6
Embolism, thrombosis: 1.065 (0.926-1.224); 0-6
Author: Hoek et al. (2000)
Period of Study:
1986- 1994
Location: The Netherlands
Health Outcome (ICD9):
Mortality: Pneumonia
(480-486); COPD (490-496);
Cardiovascular diseases (CVD)
(390-448)
Study Design: Time-series
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Statistical Analyses:
Poisson GAM, LOESS
Netherlands: 457 |jg/m3
Four Major Cities: 589 |jg/m3
Range (Min, Max):
Netherlands: (174, 2620)
Four Major Cities: (202, 4621)
Age Groups Analyzed: All ages copollutant correlation:
PM10: r= 0.64; BS: r = 0.89;
O3: r = -0.48; NO2: r = 0.89;
SO2: r = 0.65; SO42-: r = 0.55;
NO3-: r = 0.58
Increment:
Single-day lag (1): 1,500 |jg/m3
Weekly avg (0-6): 1200 |jg/m3
Relative Risk (Lower CI, Upper CI); Lag
CO
Four Major Cities: 1.022 (0.995-1.050); 1
Four Major Cities: 1.044 (1.008-1.082); 0-6
Netherlands w/o Major Cities: 1.040 (1.020-1.060); 1
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, SO42-
Total mortality: 0.990 (0.951-1.030); 0-6
CVD: 0.999 (0.939-1.063); 0-6
Author: Honda et al. (2003)
Period of Study:
1976-1990
Location:
Tokyo, Japan
Health Outcome (ICD9):
Mortality:
Total (non-accidental) (<800)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24-h avg
Median (SD) unit: 1 6 ppm
Range (Min, Max): (0, 6.8)
Copollutant correlation:
NO: r = 0.403; N02: r = 0.415;
Oxidant: r = 0.396; SO2: r = 0.675
Increment: NR
Rate Ratio (Lower CI, Upper CI); 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)
Author: Hong et al. (2002b)
Period of Study:
1/1991-12/1997
Location:
Seoul, Korea
Health Outcome (ICD9):
Mortality: Hemorrhagic and
ischemic stroke (431-434)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
All ages
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
Increment: 0.76 ppm
Relative Risk (Lower CI, Upper CI); lag:
1.06 (1.02,1.09); 1
Multipollutant:
CO, TSP: 1.07 (1.03,1.11); 1
CO, N02:1.06(1.00,1.11); 1
CO, SO2:1.05(1.01,1.10); 1
CO, 03:1.09 (1.05,1.13); 1
March 2009
C-58
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Hong et al. (1999a)
Period of Study:
1/1995- 12/1995
Location:
Inchon, Korea
Health Outcome (ICD9):
Mortality: Cardiovascular (400-
440); Respiratory
(460-519); Non-accidental
causes (<800)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.7 (0.8) ppm
Range (Min, Max): (0 3, 51)
Co pollutant:
SO2; NO2; O3
Increment: 1 ppm
Relative Risk (Lower CI, Upper CI); lag:
Total mortality:
0.993 (0.950,1.037); 0-4
Cardiovascular mortality:
0.965 (0.892,1.044); 0-4
Author: Hong et al. (2002a)
Period of Study:
1/1995- 12/1998
Location:
Seoul, Korea
Health Outcome (ICD9):
Mortality: Stroke (160-169)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
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; NO2:r = 0.64;
S02: r= 0.90; O3:r=-0.35
Increment: 0 3 ppm
% Increase (Lower CI, Upper CI); lag:
CO: 2.2% (0.4, 4.1); 2
CO (stratified by PM10 concentration):
65
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Median (SD) unit:
Cold Season: 0.5 mg/m3
Warm Season: 0.4 mg/m3
Range (Min, Max):
Cold Season: (0.1, 2.4)
Warm Season: (0.1, 1.1)
Copollutant: correlation
Cold Season:
PM2.5: r = 0.32; UFP:r = 0.47
Warm Season:
PM2.5: r = 0.24; UFP:r = 0.39
Increment: 0.2 mg/m3
% Increase (Lower CI, Upper CI); 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
March 2009
C-59
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Klemm et al. (2004)
Period of Study:
8/1998-7/2000
Location:
Fulton County and DeKalb
County, GA (ARIES)
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(390-459); Respiratory
(460-519); Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson GLM, natural cubic
splines
Age Groups Analyzed:
<65; > 65
Pollutant: CO
Averaging Time: 1-h maximum
Median (SD) unit:
1,310 (939.13) ppb
Range (Min, Max):
(303.58, 7400)
Co pollutant:
PM25; PM10-25; O3; NO2; SO2; Acid; EC;
OC; S04;
Oxygenated HCs;
NMHCs; NOs
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
Author: Kwon et al. (2001)
Period of Study:
1994-1998
Location:
Seoul, Korea
Health Outcome (ICD9):
Mortality: CHF (428);
Cardiovascular (390-459)
Study Design:
1. Time-series
2. Bi-directional case-crossover
Statistical Analyses:
1. Poisson GLM, LOESS
2. Conditional logistic regression
Age Groups Analyzed:
<55
55-64
65-74
75-84
>85
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit: 12 4 ppb
Range (Min, Max): (4.1, 38.0)
Copollutant correlation:
PM10: r= 0.713; N02:r = 0.744;
SO2: r = 0.843; O3: r = -0.367
Increment: 0.59 ppm
Odds Ratio (Lower CI, Upper CI); 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
>75:1.116 (0.989-1.258); 0
Time from admission to death
4 or less weeks: 1.088 (0.907-1.306); 0
>4 weeks: 1.017 (0.920-1.124); 0
Total mortality: 1.033 (0.946-1.127); 0
Cardiovascular mortality: 1.033 (0.920-1.160); 0
Cardiac death: 1.052 (0.919-1.204); 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 et al. (2007c)
Period of Study:
1/2000-12/2004
Location:
Seoul, Korea
Health Outcome (ICD10):
Mortality: Non-accidental (A00-
R99)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed: All ages
Pollutant: CO
Averaging Time:
Maximum 8-h moving avg
Mean (SD) unit:
w/Asian dust days: 0.92 (0.42) ppm
w/o Asian dust days:0.92 (0.41) ppm
Asian dust days only: 1.00 (0.47) ppm
Range (Min, Max): NR
Copollutant: PM10; NO2; SO2; O3
Increment: 0.54 ppm
% Increase (Lower CI, Upper CI); 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
March 2009
C-60
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Lipfert et al. (2000b) Health Outcome (ICD9):
Period of Study: !Sa!'^R^at°!4n „
5/1992 - 9/1995 (460-519); Cardiac (390-448);
Cancer; Other causes (<800)
Location: , _ .
Philadelphia, PA, three nearby Study Design. Time-series
suburban Pennsylvania Statistical Analyses:
counties, and three nearby Stepwise regression
New Jersey counties
Age Groups Analyzed:
<65
>65
Pollutant: CO
Averaging Time:
24-h avg; 1-h maximum
Mean (SD) unit:
Camden:
24-h avg: 0.75 (0.40) ppm
Philadelphia:
24-h avg: 0.63 (0.40) ppm
1-h maximum: 1.44 (1.04)
Range (Min, Max):
Camden: (0.10, 3.8)
Philadelphia:
24-h avg: (0.10, 3.3)
1-h maximum: (0.0, 7.8)
Co pollutant:
NO; N02; 03; S02; S042-;
PM10; PM2.5
Increment: NR
Attributable Risk; lag:
Peak CO
All-cause
Philadelphia: 0.0054; 0-1
4 Pennsylvania Counties: 0.0081; 0-1
Pennsylvania + NJ: 0.0085; 0-1
CO
All seven counties in Pennsylvania and New Jersey
All ages
Respiratory: -0.0067; Cardiac: 0.0131;
Other: 0.0078
All-cause:
<65: 0.0148; 0-1; >65: 0.0054; 0-1
Joint model with CO
Philadelphia: 0.0059; 0-1
4 Pennsylvania Counties: 0.0089; 0-1
Pennsylvania + NJ: 0.0096; 0-1
Cardiac: 0.0135; 0-1;
Other causes: 0.0084
<65:0.0154; 0-1;
> 65: 0.0060; 0-1
Author: Lippmann et al.
(2000)
Period of Study:
1985-1990
1992-1994
Location:
Detroit, Ml and Windsor, ON Poisson GLM
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Circulatory (390-459);
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses:
Age Groups Analyzed:
>65
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-PM10: r = 0.02;
TSP-S042-: r = 0.18;
03: r= -0.22; S02:r = 0.36;
N02: r = 0.58
1992-1994
PM10: r= 0.38; PM2.5: r = 0.38;
PM10-2 5: r = 0.24;
H+: r = 0.16; S042-:r = 0.32;
03: r = 0.16; S02: r = 0.42;
N02: r = 0.68
Increment:
1985-1990:11.5 ppm; 1992-1994: 8.4 ppm
Relative Risk (Lower CI, Upper CI); lag:
1985-1990
Total Mortality:
0.9842 (0.9667-1.002); 0
1.0103 (0.9926-1.0284) '
1.0075 (0.9898-1.0254)
1.0145 (0.9967-1.0326)
0.9968 (0.9789-1.0151)
1.0105 (0.9925-1.0288)
1.0134 (0.9954-1.0317)
1.0003 (0.9823-1.0187)
1.0152 (0.9971-1.0336)
1.0053 (0.9873-1.0236)
1
2
3
0-1
1-2
2-3
0-2
1-3
0-3
0
1
2
3
0-1
1-2
2-3
0-2
1-3
0-3
Circulatory Mortality:
0.9818 (0.9574-1.0068)
0.9991 (0.9745-1.0243)
0.9980 (0.9735-1.0232)
1.0088 (0.9841-1.0341)
0.9888 (0.9640-1.0144)
0.9981 (0.9732-1.0237)
1.0042 (0.9792-1.0298)
0.9900 (0.9650-1.0157)
1.0029 (0.9777-1.0287)
0.9944 (0.9692-1.0202)
Respiratory Mortality;
0.9644 (0.9042-1.0287); 0
1.0142 (0.9518-1.0808); 1
1.0483 (0.9845-1.1164); 2
1.0468 (0.9828-1.1149); 3
0.9868 (0.9248-1.053); 0-1
1.0372 (0.9730-1.1056); 1-2
1.0554 (0.9904-1.1246); 2-3
1.0088 (0.9457-1.0762); 0-2
1.0466 (0.9817-1.1158); 1-3
1.0205 (0.9569-1.0884); 0-3
Total minus respiratory and circulatory mortality:
0.9939 (0.9668-1.0217); 0
1.0278 (1.0001-1.0562); 1
1.0178 (0.9902-1.0461); 2
1.0227 (0.9948-1.0514); 3
1.0127 (0.9860-1.0412); 0-1
1.0269 (0.9989-1.0556); 1-2
1.0249 (0.9968-1.0538); 2-3
1.0172 (0.9893-1.0458); 0-2
1.0322 (1.0041-1.0612); 1-3
1.0229 (0.9950-1.0516); 0-3
1992-1994
Total Mortality
0.9933 (0.9636-1.024); 0
March 2009
C-61
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
1.0162 (0.9860-1
1.0116 (0.9816-1
0.9947 (0.9648-1
1.0056 (0.9756-1
1.0165 (0.9864-1
1.0038 (0.9739-1
1.0098 (0.9796-1
1.0104 (0.9862-1
1.0064 (0.9755-1
0473); 1
0426); 2
0254); 3
0366); 0-1
0476); 1-2
0476); 2-3
0409); 0-2
0414); 1-3
0382); 0-3
Circulatory Mortality
1.0076 (0.9640-1.0531)
1.0307 (0.9865-1.0768)
1.0142 (0.9705-1.0598)
0.9523 (0.9102-0.9964)
1.0229 (0.9788-1.0688)
1.0267 (0.9827-1.0727)
0.9802 (0.9375-1.0248)
1.0243 (0.9801-1.0726)
0.9987 (0.9553-1.0441)
1.0019 (0.9573-1.0487)
Respiratory Mortality
0.9894 (0.8912-1.0984)
0.9474 (0.8521-1.0533)
0.9652 (0.8682-1.0732)
0.9931 (0.8934-1.1040)
0.9626 (0.8668-1.0691)
0.9485 (0.8535-1.0541)
0.9752 (0.8775-1.0838)
0.9555 (0.8802-1.0615)
0.9567 (0.8607-1.0635)
0.9584 (0.9604-1.0675)
0
1
2
3
0-1
1-2
2-3
0-2
1-3
0-3
0
1
2
3
0-1
1-2
2-3
0-2
1-3
0-3
Total minus respiratory and circulatory mortality:
0.9769 (0.9332-1.0227)
1.0135 (0.9682-1.0609)
1.0195 (0.9747-1.0664)
1.0429 (0.9974-1.0905)
0.9940 (0.9494-1.0406)
1.0197 (0.9746-1.0670)
1.0371 (0.9918-1.0845)
1.0045 (0.9596-1.0515)
1.0353 (0.9896-1.0831)
1.0215 (0.9749-1.0702)
0
1
2
3
0-1
1-2
2-3
0-2
1-3
0-3
Author: Maheswaran et al.
(2005a)
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:
>45
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Co pollutant:
NOx;
PM10
Notes: Quintiles represent the
following mean CO concentrations and
category limits:
5: 482 (jg/m3 (a 455)
4: 443 (jg/m3 (a 433 to <455)
3: 426 (jg/m3 (a 419 to <433)
2: 405 (jg/m3 (a 387 to <419)
1: 360 (jg/m3 (<387)
Increment: NR
Rate Ratios (Lower CI, Upper CI):
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
March 2009
C-62
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Maheswaran et al.
(2005b)
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:
>45
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Quintile:
5: 482 (jg/m-
4: 443 (jg/m-
3: 426 (jg/m-
2: 405 (jg/m-
1: 360 (jg/m-
Range (Min, Max): NR
Copollutant correlation :
PM10: r= 0.88; NOx:r = 0.87
Notes: Quintiles represent the
following mean CO concentrations and
category limits:
5: 482 (jg/m3 (a 455)
4: 443 (jg/m3 (a 433 to <455)
3: 426 (jg/m3 (a 419 to <433)
2: 405 (jg/m3 (a 387 to <419)
1: 360 (jg/m3 (<387)
Increment: NR
Rate Ratios (Lower CI, Upper CI); 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);
1.22 (1.07, 1.39)
0.95 (0.83, 1.09)
0.97 (0.85, 1.11)
1 (lowest): 1.00
Author: Mar et al. (2000)
Period of Study:
1995-1997
Location:
Phoenix, AZ
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Cardiovascular (390-449)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups Analyzed:
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.5 (0.8) ppm
Range (Min, Max):
1995: (0.5, 4.0) ppm
1996: (0.3, 4.0) ppm
1997: (0.3, 3.7) ppm
Copollutant correlation:
PM2.5: r = 0.85;
PM10: r= 0.53;
PM10-25: r = 0.34;
N02: r = 0.87;
O3: r =-0.40;
S02: r =0.53
Increment: 1 19 ppm
Relative Risk (Lower CI, Upper CI); lag:
Total Mortality (CO exposure):
1.06 (1.02,1.09); 0;
1.05 (1.01,1.09); 1
Cardiovascular Mortality (CO exposure):
1.05 (1.00,1.11); 0;
1.10(1.04,1.15); 1;
1.07 (1.02,1.12); 2;
1.07 (1.02,1.12); 3;
1.08 (1.03,1.13); 4
Author: Moolgavkar et al.
(2000b)
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
NO2:
Cook: r= 0.63;
LA: r = 0.80;
Maricopa: r = 0.66
SO2:
Cook: r= 0.35;
LA: r = 0.78;
Maricopa: r = 0.53
03:
Increment: 1 ppm
% Change (Lower CI, Upper CI); 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); 0;/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
CO and PM25
0.43 (-1.35, 2.20); 0; / 2.88 (1.16, 4.60); 1;
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); 1;
March 2009
C-63
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Cook: r= -0.28;
LA: r = -0.52;
Maricopa: r = -0.61
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); 0;/2.80 (-1.60,7.19); 1;
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); 1;
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); 0;/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); 1;
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); 0; / 3.50 (0.60, 6.41); 1;
3.52 (0.66, 6.38); 2;/4.61 (1.85, 7.37); 3;
4.78 (2.10, 7.46); 4;/5.15 (2.45, 7.84); 5
Notes: Total Mortality effect estimates were not
presented quantitatively.
Author: Moolgavkar et al.
(2003b)
Period of Study: 1987-1995
Location:
Cook County, Illinois & Los
Angeles County, California
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Circulatory (390-448)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All Ages
Pollutant: CO
Averaging Time: 24-h avg
Median unit:
Cook County: 993 ppb
LA County: 1347 ppb
Range (Min, Max):
Cook County: (224, 3912) ppb
LA County: (237, 5955) ppb
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;
O3: r= -0.52;
SO2: r = 0.78;
PM10: r= 0.45;
PM2.5: 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); 0;/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; 12.2% (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); 0;/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); 0; / 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); 0; /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
March 2009
C-64
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Ostro et al. (1999)
Period of Study:
1989-1992
Location:
Coachella Valley, California
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-519);
Cardiovascular (393-440)
Study Design: Time-series
Statistical Analyses: Poisson
GAM; LOESS
Age Groups Analyzed: >50
Pollutant: CO
Averaging Time: 1-h maximum
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
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; /2.0% (5.1); 1; /1.0% (2.6); 2;
0.6% (1.5); 3; / 0.4% (1.2); 4; / 0.2% (0.6); 5
Winter (CO):
0% (0); 0; 10.8% (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); 0;/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;
0.1% (1.1); 3;/1.4% (1.9); 4;/1.6% (2.1); 5
CVD Mortality Los Angeles County
CO:
1.6% (6.3); 0; /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, PM2.5:
-2.2% (-2.7); 0; /1.5% (1.8); 1; /1.9% (2.0); 2;
1.9% (2.2); 3; / 2.1% (2.6); 4; /3.7%(4.5); 5
CVD Mortality (Season Specifid) Cook County
Spring (CO):
0.7% (0.5); 0; /1.4% (1.1); 1; /0.3% (0.3); 2;
1.1% (0.9); 3; / 0.4% (3.1); 4; /0.1% (0.6); 5
Summer (CO):
-2.6% (-1.4); 0; / 2.5% (1.4); 1; / 6.5% (3.7); 2;
0.9% (0.5); 3; /-1.9% (-1.1); 4; / -1.0% (-0.6); 5
Fall (CO):
0% (0); 0; 12.9% (2.5); /1; 0% (0); 2;
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
Increment: NR
P(SE); lag:
CO: 0.0371 (0.0157); 2
CO, PM10: 0.0300 (0.0194); 2
March 2009
C-65
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Penttinen et al.
(2004)
Period of Study:
1988-1996
Location:
Helsinki, Finland
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-519);
Cardiovascular (393-440)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, LOESS
Age Groups Analyzed:
All ages
15-64 years
65-74 years
>75
Pollutant: CO
Averaging Time:
Maximum 8-h avg
Median unit: 12 mg/m3
Range (Min, Max): (0,12 4)
Copollutant correlation:
03: r= -0.46; NO2:r = 0.59;
S02: r= 0.55; PM10: r= 0.45; TSP: r =
0.26;
TSP Blackness: r= 0.26
Increment: 1 mg/m3
% Increase (Lower CI, Upper CI); lag:
Total Mortality
-1.50% (-2.78,-0.22); 0
0.15% (-1.09, 1.39); 1
-1.00% (-2.80, 0.81); 0-3
Cardiovascular Mortality
-2.48% (-4.30, -0.66); 0
-0.84% (-2.61,0.93); 1
-1.87% (-4.43, 0.69); 0-
Respiratory Morality
-0.48% (-4.84, 3.87)
-0.14% (-4.43, 4.15)
-1.49% (-7.73, 4.74)
0
1
0-3
Author: Peters et al. (2000a)
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;
PMzs: r = 0.42
Increment: 1 mg/m3
Relative Risk (Lower CI, Upper CI); lag:
Coal Basin of the Czech Republic
Total Mortality:
1.016 (0.998,1.035); 0;/ 1.016 (0.998,1.034); 1; 1.013
(0.996,1.030); 2;/1.012 (0.995,1.028); 3
Northeast Bavaria
Total Mortality:
1.014(0.994,1.034); 0;/ 1.023 (1.005,1.041); 1; 1.013
(0.995,1.031); 2;/1.003 (0.985,1.021); 3
Cardiovascular Disease Mortality:
1.018 (0.994,1.044); 0;/ 1.012 (0.987,1.038); 1; 1.016
(0.991,1.041); 2;/1.004 (0.980,1.029); 3
Author: Rainham et al (2003) Health Outcome (ICD9):
Period of Study:
1980-1996
Location:
Toronto, ON, Canada
Mortality: Cardiac (390-459);
Respiratory (480-519); Total
(non-accidental) (<800)
Study Design: Time-series
Statistical Analyses:Poisson
GAM, natural cubic splines
Age Groups Analyzed:
<65
>65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1.0 (0.4) ppm
Range (Min, Max): (0 0, 4 0)
Copollutant: O3; N02; S02
The study did not present quantitative results for
CO
Author: Roemer et al. (2001)
Period of Study:
1/1987- 11/1998
Location: Amsterdam
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Air pollution background:
836 (jg/m3
Air pollution traffic: 1805 |jg/m3
Range (10th, 90th):
Air pollution background:
(448,1315) (jg/m3
Air pollution traffic:
(727, 3192) (jg/m3
Copollutant:
BS; PM10; S02; N02; NO; 03
Increment:
Lag 1 and 2:100 |jg/m3
Lag 0-6: 50 |jg/m3
Relative Risk (Lower CI, Upper CI); lag:
Total Population using Background sites
1.002 (1.000-1.004); 1;
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;
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)
1.000 (1.000-1.001)
1;
2;
0-6
March 2009
C-66
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Samet et al. (2000b)
Period of Study:
1987-1994
Location:
20 U.S. Cities: Los Angeles,
CA; New York, NY; Chicago,
IL; Dallas, TX; Houston, TX;
San Diego, CA; Anaheim, CA;
Phoenix, AZ; Detroit, Ml;
Miami, FL; Philadelphia, PA;
Minneapolis, MN; Seattle, WA;
San Jose, CA; Cleveland, OH;
San Bernardino, CA;
Pittsburgh, PA; Oakland, CA;
Atlanta, GA; San Antonio, TX
Health Outcome (ICD9):
Mortality: Cardiovascular (390-
459); Respiratory
(460-519); Other (non-accidental)
(<800)
Study Design: Time-series
Statistical Analyses:
Two-stage log linear regression
model
Age Groups Analyzed:
<65
65-74
>75
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
Los Angeles: 15.1 ppm
New York: 20.4 ppm
Chicago: 7.9 ppm
Dallas: 7.4 ppm
Houston: 8.9 ppm
San Diego: 11.0 ppm
Anaheim: 12.3 ppm
Phoenix: 12.6 ppm
Detroit: 6.6 ppm
Miami: 10.6 ppm
Philadelphia: 11.8 ppm
Minneapolis: 11.8 ppm
Seattle: 17.8 ppm
San Jose: 9.4 ppm
Cleveland: 8.5 ppm
San Bernardino: 10.3 ppm
Pittsburgh: 12.2 ppm
Oakland: 9.1 ppm
Atlanta: 8.0 ppm
San Antonio: 10.1 ppm
Range (10th, 90th):
Los Angeles: (5.9, 28.3)
New York: (14.8, 27.6)
Chicago: (4.5,11.9)
Dallas: (3.6,12.0)
Houston: (4.0,14.2)
San Diego: (4.5, 20.5)
Anaheim: (3.7, 25.2)
Phoenix: (5.4, 22.6)
Detroit: (3.2,11.1)
Miami: (6.5,15.9)
Philadelphia: (7.0,17.2)
Minneapolis: (7.0,17.0)
Seattle: (10.5, 26.4)
San Jose: (1.7, 21.3)
Cleveland: (3.7,13.8)
San Bernardino: (4.0,17.5)
Pittsburgh: (6.1,19.8)
Oakland: (2.9,17.0)
Atlanta: (3.2,14.3)
San Antonio: (4.1, 17.3)
Copollutant correlation:
PM10: r= 0.45; 03: r= -0.19;
N02: r = 0.64; S02:r = 0.41
This study did not provide quantitative results for
CO.
Author: Samoli et al. (2007) Health Outcome (ICD9):
Pollutant: CO
Period of Study:
1990-1997
Location:
19 European Cities
(APHEA2)
K1*1 (nonf Rental) Averaging Time: 24-h avg
<800 ; Cardiovascular 390-459 M M M
Mean Range (unit - mg/m3):
Study Design: Time-series
Statistical Analyses:
Poisson and two-stage
hierarchical model
Age Groups Analyzed:
All ages
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)
Rome: (2.5, 5.9)
_^tockholnTjO_5J_2]^^^_^^^_
Increment: 1 mg/m3
% Increase (Lower CI, Upper CI); lag:
Non-accidental mortality
8 Degrees of Freedom per year
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, SO2: 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, SO2: 0.68% (0.47-0.90); 0-1
CO, 03:1.12% (0.93-1.31); 0-1
CO, N02: 0.72% (0.50-0.95); 0-1
March 2009
C-67
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
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
SO2: r= 0.35 to 0.82
N02: r = 0.03 to 0.68
O3: 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
CO, 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
CO, 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
Mean O3:
25th Percentile: 1.04% (0.67-1.41); 0-1
75th Percentile: 0.82% (0.55-1.10); 0-1
Standardized Mortality Rate:
25th Percentile: 1.06% (0.71-1.42); 0-1
75th Percentile: 0.61% (0.30-0.93); 0-1
Population >75 years of age (%):
25th Percentile: 0.58% (0.25-0.92); 0-1
_75th_Percentilei_0ii94%_^0i64i1_i242J_0^1^^^_^^^
March 2009
C-68
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
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)
Mean 63:
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 >75 years of age (%):
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)
Period of Study:
1989-1995
Location:
Spokane, WA
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 1-h avg
Mean (SD) unit:
Dust Storm Days:
09/08/1990: 6.37 ppm
09/12/1990:3.40 ppm
10/04/1990: 3.15 ppm
11/09/1990:2.45 ppm
11/23/1990:2.50 ppm
09/13/1991:4.60 ppm
10/16/1991: 2.10 ppm
10/21/1991:2.20 ppm
09/04/1992:3.43 ppm
09/12/1992:1.80 ppm
09/13/1992:1.65 ppm
09/25/1992:2.95 ppm
09/26/1992:4.30 ppm
10/08/1992: 3.85 ppm
09/11/1993:1.88 ppm
11/3/1993: 5.33 ppm
07/24/1994: 2.10 ppm
08/30/1996: 2.85 ppm
Range (Min, Max): NR
Copollutant: PM10
The study did not present quantitative results for
CO
Author: Sharovsky et al.
(2004)
Period of Study:
1996-1998
Location:
Sao Paulo, Brazil
Health Outcome (ICD10):
Mortality: Myocardial Infarction
(1.21)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed: 35-109
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 3.7 (1.6) ppm
Range (Min, Max): (10,118)
Copollutant: correlation
S02: r= 0.73; PM10: r= 0.51
Increment: NR
px 100 (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 et al.
(2005)
Period of Study:
1/1995-6/2001
Location:
Spokane, WA
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-519);
Asthma (493); COPD (491, 492,
494, 496); Pneumonia (480-487);
Acute Upper Respiratory Tract
Infections (464-466, 490);
Cardiac Outcomes (390-459)
Study Design: Time-series
Statistical Analyses:
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 :
PM1: r = 0.63; PM25:r=0.62;
PM10: r= 0.32;
PM10-25: r = 0.32
The study did not present quantitative results for
CO.
March 2009
C-69
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Stieb et al. (2003)
Period of Study:
1985-2000
Location: All locations
Health Outcome (ICD9):
Mortality: Non-accidental
Study Design: Meta-analysis
Statistical Analyses: NR
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: NR
IQR (25th, 75th): NR
Age Groups Analyzed: All ages Copollutant: NR
Increment: 1.1 ppm
% Excess Mortality (Lower CI, Upper CI); 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)
Author: Stolzel et al. (2006)
Period of Study:
9/1995-8/2001
Location:
Erfurt, Germany
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800);
Cardio-respiratory (390-459, 460-
519,785,786)
Study Design: Time-series
Statistical Analyses: Poisson
GAM
Age Groups Analyzed:
All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
0.47 (0.39) mg/m3
IQR (25th, 75th): (0 23, 0 57)
Copollutant correlation:
MCo. 1-0.5: r = 0.58;
MCo01-25: r = 0.57;
PM10: r= 0.50; NO: r= 0.70;
N02: r = 0.71
Increment: 0.34 mg/m3
Relative Risk (Lower CI, Upper CI); 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 (0.974-1.017); 5
Author: Sunyeret al. (2001)
Period of Study:
1990-1995
Location:
Barcelona, Spain
Health Outcome (ICD9):
Mortality: COPD
(491, 492, 494, 496)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:>35
Pollutant: CO
Averaging Time: 8-h avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutant: PM10; NO2; O3
Increment: 4 5 |jg/m3
Odds Ratio (Lower CI, Upper CI); lag:
CO: 1.052 (0.990-1.117); 0-2
CO, PM10:1.017 (0.947-1.091); 0-2
Author: Sunyer et al. (2002)
Period of Study:
1985-1995
Location:
Barcelona, Spain
Health Outcome (ICD9):
Mortality: Respiratory mortality
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:>14
Study population:
Asthmatic individuals: 5,610
Pollutant: CO
Averaging Time: 24-h avg
Median (SD) unit: 7 7 |jg/m3
Range (Min, Max): (0 6, 66 0)
Copollutant:
PM10; BS; N02; 03; S02
Increment: 7 2 |jg/m3
Odds Ratio (Lower CI, Upper CI); lag:
Asthmatic individuals with 1 ED visit
1.127 (0.895-1.418); 0-2
Asthmatic individuals with >1 ED visit
1.125 (0.773-1.638); 0-2
Asthma/COPD individuals with >1 ED visit
0.815 (0.614-1.082); 0-2
Author: Tsai et al. (2003a)
Period of Study:
1994-2000
Location:
Kaohsiung, Taiwan
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800); Respiratory (460-519);
Circulatory (390-459)
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: 0.827 ppm
Range (Min, Max): (0.226,1.770)
Copollutant:
PM10; SO2; NO2; O3
Increment: 0.313 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Total (non-accidental): 1.003 (0.968-1.039); 0-2
Respiratory: 1.011 (0.883-1.159); 0-2
Circulatory: 0.986 (0.914-1.063); 0-2
Author: Tsai et al. (2006b)
Period of Study:
1994-2000
Location:
Kaohsiung, Taiwan
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<800)
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed:
27 days old to <1 yr of age
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 8 27 ppm
Range (Min, Max): (2.26,17.70)
Copollutant:
PM10; SO2; O3; NO2
Increment: 0.31 ppm
Odds Ratio (Lower CI, Upper CI); lag:
Postneonatal Mortality
1.051 (0.304-3.630); 0-2
Author: Vedal et al. (2003)
Period of Study:
1/1994-12/1996
Location:
Vancouver, BC, Canada
Health Outcome (ICD9):
Mortality: Total (non-accidental)
(<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:
PMio:r = 0.71; 03: r = 0.12;
N02: r = 0.81 ;S02:r = 0.67
Winter:
PM10: r= 0.76; O3: r= -0.65;
N02: r = 0.78; S02:r = 0.83
The study did not present quantitative results for
CO.
March 2009
C-70
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Villeneuve et al.
(2003)
Period of Study:
1986-1999
Location:
Vancouver, BC, Canada
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(401-440); Respiratory
(460-519); Cancer (140-239)
Study Design: Time-series
Statistical Analyses:
Poisson, natural splines
Age Groups Analyzed: a 65
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 10 ppm
Range (Min, Max): (0.2, 4.9)
Co pollutant:
PM25; PM10; PM10-25; TSP;
S04; CO; COH; 03; N02; SO2
Increment: 1 1 ppb
% Increase (Lower CI, Upper CI); 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
Qancer
-2.8% (-7.6 to 2.4); 0-2; I -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: Wichmann et al.
(2000)
Period of Study:
9/1995- 12/1998
Location:
Erfurt, Germany
Health Outcome (ICD9):
Mortality: Non-accidental
(<800); Cardiovascular
(401-440); Respiratory
(460-519)
Study Design: Time-series
Statistical Analyses:
Poisson GAM, LOESS
Age Groups Analyzed:
<70
70-79
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 CI, Upper CI); 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); 1
Log-transformed: 1.027 (0.973-1.083); 1
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 et al. (2004a)
Period of Study:
1994-1998
Location:
Taipei, Taiwan
Health Outcome (ICD9):
Mortality:
Non-accidental (<800);
Circulatory (390-459);
Respiratory (460-519)
Study Design:
Bi-directional case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups Analyzed: All ages
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit: 1 16 ppm
Range (Min, Max): (0.24, 4.42)
Copollutant:
PM10; SO2; NO2; O3
Increment: 0.52 ppm
Odds Ratio (Lower CI, Upper CI); 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
March 2009
C-71
DRAFT-DO NOT QUOTE OR CITE
-------
Table C-8. Studies of long-term CO exposure and mortality.
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Lipfert et al.
(2000b)
Period of Study:
1975-1996
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-90,000 hypertensive male U.S.
veterans
Statistical Analyses:
Staged regression
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1960-1974: "
1975-1981
1982-1988
1989-1996
10.82 (5.15) ppm
7.64 (2.94) ppm
3.42 (0.95) ppm
2.36 (0.67) ppm
Range (Min, Max):
1960-1974
1975-1981
1982-1988
1989-1996
(0.94, 35.30)
(0.43, 22.38)
(0.30,15.20)
(0.30,7.10)
Copollutants; correlation:
1960-1974:
03:r= 0.004;
N02:r = 0.690;
S042": r = 0.469
1975-1981:
03:r = 0.109;
N02:r = 0.249;
S042-: r = -0.155;
IP S042-: r = 0.356;
PMzs:r= 0.634;
PM10-25: r = 0.498;
PM15: r = 0.626
1982-1988
03: r= 0.158; N02:r =
r = -0.518;
IP S042-: r = 0.075;
PMzs:r= 0.296;
PM10-25: r = 0.135
PM15: r = 0.284
1989-1996
03:r= 0.397;
N02:r = 0.492;
S042-: r = -0.551
0.413; S042-:
Increment: NR
Coefficient:
Baseline Model
Exposure Period: up to 1975
Single Period: -0.000
Deaths, 1976-81:0.0043
Deaths, 1982-88:-0.0002
Deaths after 1988: -0.0041
Exposure Period: 1975-81
Single Period: -0.013
Deaths, 1976-81: -0.0170
Deaths, 1982-88:-0.0217
Deaths after 1988: -0.0240
Exposure Period: 1982-88
Single Period: -0.028
Deaths, 1976-81:-0.0294
Deaths, 1982-88:-0.0484
Deaths after 1988: -0.0424
Exposure Period: 1989-96
Single Period: -0.046
Deaths, 1976-81: -0.0590
Deaths, 1982-88:-0.0581
Deaths after 1988: -0.0536
Final Model w/ Ecological Variables
Exposure Period: up to 1975
Single Period: -0.001
Deaths, 1976-81:0.0013
Deaths, 1982-88:-0.0022
Deaths after 1988: -0.0061
Exposure Period: 1975-81
Single Period: -0.008
Deaths, 1976-81:-0.0128
Deaths, 1982-88:-0.0186
Deaths after 1988: -0.0203
Exposure Period: 1982-88
Single Period: -0.009
Deaths, 1976-81: -0.0007
Deaths, 1982-88:-0.0246
Deaths after 1988: -0.0216
Exposure Period: 1989-96
Single Period: -0.009
Deaths, 1976-81:-0.0106
Deaths, 1982-88:-0.0136
Deaths after 1988: -0.0078
Notes: Mortality risks based on mean concentrations
of pollutants less estimated background weighted by
the number of subjects in each county, but The study
did not present this value for each pollutant.
Author: Lipfert and
Morris (2002)
Period of Study:
1960-1997
Location:
U.S. counties
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design:
Ecological/ cross-sectional
Statistical Analyses:
Staged regression
Age Groups Analyzed:
15-44
45-64
65-74
75-84
>85
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
Copollutant:
TSP
SO42-
S02
N02
Os
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 a 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)
March 2009
C-72
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
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 a 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 a 85: 0.0187 (0.0135)
Log Mean: 0.0084 (0.0149)
Peak C01979-1981, 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 a 85: -0.0201 (0.0089)
Log Mean: -0.0165 (0.0089)
Peak C01979-1981, 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 a 85: -0.0073 (0.0170)
Log Mean: 0.0109 (0.0218)
Peak C01979-1981, 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 a 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 a 85: -0.0209 (0.0073)
Log Mean: -0.0178 (0.0098)
Peak CO 1979-1981, All locations
Ages 15-44: 0.0522 (0.0227)
Ages 45-64: -0.0047 (0.0121)
Ages 65-74: -0.0165 (0.0078)
Ages 75-84: -0.0268 (0.0068)
Ages a 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 a 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 a 85:-0.0138 (0.0113)
Log Mean: 0.0041 (0.0176)
Attributable Risks of mortality (1995-97)
Peak CO 1995-1997, All locations
March 2009
C-73
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
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 a 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 a 85: -0.0158 (0.0162)
Log Mean: 0.0162 (0.0149)
Author: Lipfert et al.
(2006b)
Period of Study:
1976-2001
Location:
32 Veterans Hospitals
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-70,000 hypertensive male U.S.
veterans
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1976-1981
1982-1988
1989-1996
1997-2001
7.6 (2.9) ppm
3.4 (9.5) ppm
2.4(0.67) ppm
1.6 (5.6) ppm
Range (Min, Max): NR
Copollutants correlation:
In(VKTA): r = -0.06
Avg NO2: r= 0.43
Peak O3: r= 0.08
Peak SO2: r = -0.05
PM25: r = 0.08
S042": r = -0.16
Note: VKTA=annual vehicle-km
traveled/km2
Increment: 2 ppm
Relative risk (Lower CI, Upper CI):
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, NO2+O3:1.023 (0.939-1.115)
March 2009
C-74
DRAFT-DO NOT QUOTE OR CITE
-------
Study
Design
Concentrations
Effect Estimates (95% CI)
Author: Lipfert et al.
(2006a)
Period of Study:
1997-2002
Location:
32 Veterans Hospitals,
USA
Mortality
Health Outcome (ICD9):
Non-accidental
Study Design: Cohort
Study Population:
-18,000 hypertensive male U.S.
veterans
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: NR
Pollutant: CO
Averaging Time:
95th Percentile Annual avg
Mean (SD) unit:
1999-2001:1.63 (0.84) ppm
1999-2001 (STN sites only): 1.73
(0.77)
Range (Min, Max):
1999-2001: (0.40, 6.7)
1999-2001 (STN sites only): (0.47, 4.2)
Copollutants correlation:
ln(traffic 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
NOs-:r=-0.088
^PlVhscomp.: r= 0.133
N02: r = 0.418
Peak 03: r= 0.172
Peak SO2: r = 0.405
Increment: NR
p coefficient (SE); t-statistic:
-0.00000536 (0.0000324);-0.165
Author: Jerrett et al.
(2003)
Period of Study:
1982-1989
Location:
107 U.S. cities
Mortality
Health Outcome (ICD9):
Cardiovascular; CHD;
Cerebrovascular disease
Study Design: Cohort
Study Population:
65, 893 postmenopausal women
without previous cardiovascular
disease
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: a 30
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: 1 56 ppm
Range (Min, Max): (019, 3.95)
Copollutants correlation:
Sulfates: r= -0.07
N02
03
SO2
Increment: 1 ppm
Relative risk (Lower CI, Upper CI):
CO: 0.98 (0.92-1.03)
CO, Sulfates: 0.97 (0.92-1.03)
Author: Miller et al.
(2007)
Period of Study:
1994-1998
Location:
36 U.S. cities
Mortality
Health Outcome (ICD9):
Cardiovascular; CHD;
Cerebrovascular disease
Study Design: Cohort
Study Population:
65, 893 postmenopausal women
without previous cardiovascular
disease
Statistical Analyses:
Cox proportional-hazards model
Age Groups Analyzed: 50-79
Pollutant: CO
Averaging Time: Annual avg
Mean (SD) unit: NR
Range (Min, Max): NR
Copollutants:
PM2.5
PM10-2.5
S02
N02
03
Increment: 1 ppm
Hazard ratio (Lower CI, Upper CI):
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, PM10-2.5, S02, N02, 03: 0.93 (0.67,1.30)
Author: Pope et al.
(2002)
Period of Study:
1980-1998
Location:
All 50 States, Washington
DC, and Puerto Rico
(ACS-CPS-II)
Mortality
Health Outcome (ICD9):
Total (non-accidental) (<800); Lung
Cancer (162); Cardiopulmonary (401-
440, 460-519)
Study Design:
Prospective cohort
Statistical Analyses:
Cox proportional hazards model
Age Groups Analyzed: a 30
Pollutant: CO
Averaging Time: 24-h avg
Mean (SD) unit:
1980:1.7(0.7) ppm
1982-1998:1.1 (0.4) ppm
Range (Min, Max): NR
Copollutant:
PM2.5; PM10; TSP; S02; N02; 03
The study presents results for CO graphically.
March 2009
C-75
DRAFT-DO NOT QUOTE OR CITE
-------
Annex D. Controlled Human
Exposure Studies
Table D-1. Controlled human exposure studies.
Study
Subjects
Exposure
Findings
Adiretal.
(1999)
15 healthy non-
smoking males;
22-34 years old
Subjects exposed to both CO and room air for 3 min 45 sec. Exposure to CO resulted in a decrease in post-exposure
Actual inhaled CO concentration not provided. Subjects
exposed to CO concentration required to produce a venous
COHb of 4-6%. Exposures were separated by 1 month with
the order of exposure randomly assigned.
exercise duration (Bruce protocol) relative to clean air
exposure in 13 out of 15 subjects (p=0.0012). Statistically
significant decreases in metabolic equivalent units (METs)
were also reported following CO exposure (p = 0.0001). No
CO-induced changes in heart rate (HR), BP, ECG
parameters, or myocardial perfusion were observed.
Bathoorn et al. 19 former smokers Exposures to both CO and room air conducted for 2 h on
(2007) with COPD; each of four consecutive days in a randomized crossover
18 males/1 female; design. 9 subjects were exposed to 100 ppm and
66-70 years old 10 subjects were exposed to 125 ppm. CO and room air
exposures were separated by at least 1 week.
Following the fourth 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 FEVi. Neither of these effects
were shown to reach statistical significance. No changes in
sputum neutrophils, white blood cell counts or serum C-
reactive protein (CRP) were observed. Although this study
appears to demonstrate some evidence of an anti-
inflammatory effect of CO among subjects with COPD, it
must be noted that two of these patients experienced
exacerbations of COPD during or following CO exposure,
with one patient requiring hospitalization two months after
exposure (initial symptoms first experienced 1-week post-
exposure).
Hanada et al. 20 healthy males; 15 subjects exposed for 20 min (10 min rest, 5 min handgrip
(2003) 26 ± 1 years old exercise, 2 min post-exercise ischemia, 3 min recovery)
under the following four conditions: (1) normoxia (inspiratory
02 fraction 21.4%), (2) hypoxia (inspiratory 02 fraction
10.3%), (3) CO + normoxia, (4) CO + hyperoxia (inspiratory
02 fraction 95.9%). Inhaled CO administered to achieve
-20% COHb in venous blood. Each of the four conditions
were separated by 20 min of rest. 5 subjects served as
controls (four consecutive 20 min periods of normoxia).
Blood oxygenation, BP, HR and respiratory rate were
measured during exposure. Muscle sympathetic nerve
activity (MSNA) and leg hemodynamics were evaluated in
two subsets of the study group (n = 8 and 7, respectively).
Arterial oxygen saturation (pulse oximetry) was significantly
lower, and resting HR and ventilation significantly higher
during the period of hypoxia compared to the other periods;
none of these measures were affected by exposure to CO.
MSNA was shown to increase during hypoxia and CO
exposure relative to normoxia. Neither hypoxia nor CO was
found to affect leg blood flow or vasoconstriction.
Kizakevich 16 healthy non- Subjects exposed on 4 separate days to increasing CO
et al. (2000) smoking males; concentrations during either upper-body exercise (hand-
18-29 years old crank) or lower-body exercise (treadmill). Targeted COHb
levels were initially attained using short term (4-6 min)
exposures to CO at concentrations between 1,000 and
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).
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 a
5%, and increases in both cardiac output and cardiac
contractility reached statistical significance at COHb
concentrations a 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 s 20%.
Mayretal. 13 healthy non- Subjects exposed to both 500 ppm CO and clean air for 1 h,
(2005) smoking males; with exposures separated by a 6-week period. Immediately
18-38 years old following exposure, subjects were administered an
intravenous bolus dose (2 ng/kg) of lipopolysaccharide
(LPS).
The average COHb concentration was 7% following the 1 h
CO exposure. 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. 12 healthy non- Exposures conducted on two separate occasions to both
(2008) smoking males; room air (6 min) and CO. Subjects were exposed to
25 ± 2.9 years old 3,000 ppm CO until COHb reached 6% (3-8 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)
March 2009
D-1
DRAFT-DO NOT QUOTE OR CITE
-------
Study Subjects
Exposure
Findings
Ren et al. 12 healthy subjects; Each subject underwent four different 8 h experimental
(2001) 11 nonsmokers; protocols: (1) isocapnic hypoxia (end-tidal P02 held at
9 males / 3 females; 55 mmHg), (2) withdrawal of 500 mL of venous blood at the
20-32 years old 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.
Resch et al.
(2005)
15 healthy non-
smoking males;
27 ± 4 years old
Subjects exposed to 500 ppm CO and synthetic air for 1 h at COHb levels averaged 5.6% after 30 min and 9.4% after 60
rest in a randomized crossover study design. Exposures
were separated by a period of at least 1 week.
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. 10 healthy non- Each subject was exposed to CO for 30-45 min to achieve a
(2004). smoking males; COHb level of 10%. Prior to and following exposure,
22-52 years old subjects performed hypoxic and hyperoxic rebreathing tests.
Four subjects were exposed to hypoxic conditions first,
while six subjects were exposed to hyperoxic conditions
first, both prior to and following CO exposure.
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 CO2-
induced stimulation ofventilation.
Zevin et al. 12 healthy male Exposures were conducted over 21 consecutive days under
(2001) smokers; three different protocols, with each protocol lasting 7 days.
27-47 years old In one protocol, subjects smoked 20 cigarettes per day, one
every 45 min. In the other two protocols, every 45 min (20
times per day) subjects breathed either air or CO
(-1200 ppm) from a 1 liter bag once per min for 10 min at a
time. Subjects completed all three protocols, with six
subjects exposed sequentially to CO, smoking, then air, and
the other six exposed sequentially to air, smoking, then CO.
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 three exposures.
March 2009
D-2
DRAFT-DO NOT QUOTE OR CITE
-------
Annex E. Toxicological Studies
Table E-1. Human and animal studies.
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Acevedo and Ahmed
(1998)
Human
HO-1 and HO-2 (mRNA and protein) were upregulated in
pregnant myometrium when compared to non-pregnant
myometrium. The HO activator hemin inhibited spontaneous
and oxytocin-induced contractility of the myometrium.
Progesterone induced HO-1 and HO-2 mRNA expression.
Achouh et al. (2008)
Human arteries
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
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 versus 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)
Human
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
(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.
(2002)
Rat
(Sprague Dawley)
The role of the HO/CO system in estrous cyclicity, pregnancy
and lactation was evaluated using HO inhibitors and stubstrates.
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 (2003a)
Rat
(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 and
Lawson (2003b)
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)
Human muscle tissue
mitochondria
5 min
50-500 ppm
CO significantly reduced muscle mitochondrial cytochrome c
oxidase activity by 20%, 42%, and 55% after treatment with 50,
100, and 500 ppm CO respectively but did not change the
activity of three other electron transport proteins.
Andresen et al. (2006)
Rat
(Long Evans)
Male
Mouse
(C57BL/6J)
Male
Cerebral vessels
1-100 pM
CO did not dilate rat or mouse cerebral arteries until 100 |jM,
which is not a physiological concentration. Also, the HO
inhibitors constricted vessels in a nonspecific manner.
Antonelli et al. (2006)
Rat
(Wistar)
GD5-GD20
75 ppm
Pups exposed to CO in utero had significant impairment of
cortical neuronal glutamanergic transmission at PND1 in both
neurons at rest and in neurons stimulated with depolarization.
Appleton and Marks
(2002)
Human placenta
Endogenous CO production by HO in the human placenta was
regulated by O2 availability. Placental HO activity was directly
dependent on O2 availability; this does not vary between pre-
eclamptic and normotensive placentas.
March 2009
E-1
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Ashfaq et al. (2003)
Human plancenta
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.
Astrup (1972)
Rabbit
(strain not identified)
Continuous CO
exposure over
gestation
90 or 180 ppm
Skeletal abnormalities: Three pups (from n = 123) in the 180
ppm CO group had deformities in their extremities at birth,
whereas no control and no 90 ppm CO-exposed animals
manifested with this malformation.
Bainbridge et al.
(2002)
Human placenta
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. Human placenta
(2006)
6 h
Starting concentrations
of CO: 3.9 (j M CO in cell
culture media (control)
and CO-exposed groups:
116 ijM, 145 ijM,
181 pM.
After 3 hours, the CO in
the culture media was
3.7 (jM (control), and
CO-exposed cells
10.2,12, and 15.9 |jM.
C-section placentas were collected from healthy term
pregnancies.Villous explants of placentas were cultured under
hypoxia followed by reoxygenation (H/R). H/R + CO-exposed
placental tissue had decreased apoptosis and decreased PARP
(a protein marker of apoptosis) versus control H/R exposed
cells. Secondary necrosis of the placental tissue post H/R was
inhibited by CO treatment.
Bainbridge and Smith
(2005)
Human placenta
The role of HO in the placenta and during pregnancy are
reviewed in this article. The conflicting data on the activity,
localization, and expression of HO in the placentas of pre-
eclamptic women are presented.
Bamberger et al.
(2001)
Human placenta
Expression and tissue localization of soluble guanylyl cyclase in
human placenta using antibody localization were characterized.
These tools can be used in future studies to elucidate the
NO'/CO/cGMP pathway.
Barber et al. (1999)
Human
HO and NOS did not maintain human uterine quiescence during
pregnancy.
Barber et al. (2001)
Human placenta
Women who had pregnancies with fetal growth restrictions
(FGR) produced term placenta with significant decreases in HO-
2 versus healthy pregnancies.
Baum et al. (2000)
Human
End-tidal CO measurements in women with pregnancy-induced
hypertension and pre-eclampsia were significantly lower than in
normotensive pregnant women.
Bergeron et al. (1998)
Rat
Brain
To address the developmental changes of HO staining in the
brain, immunohistochemical staining for HO-1 was performed
on the developing rat brain at PND7, PND14, and PND21. HO-1
staining was most intense at PND7 and by PND21 reached its
adult pattern of staining localizing to the hippocampus, thalamic
and hypothalamic nuclei, with virtually no staining of
endothelium, white matter and cortex. HO-2 is the dominant HO
isoform in the brain.
Bing et al. (1995)
Rodent
Spatial learning in the Morris water maze was enhanced in
rodents exposed to the HO inhibitor tin protoporphyrin (Sn-PP).
Burmester et al. (2000) Human
and
Mouse
Nb had a high oxygen affinity similar to Mb, thus may increase
the availabilty of 62to brain tissue.
Bye et al. (2008)
Rat
(Wistar)
Female
100 h/wk for 18 200 ppm
mo
CO-exposed (11-14.7% COHb) rats experienced a 24%
decrease in aerobic capacity evidenced by VO2 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.
Cagiano et al. (1998)
Rat
GD0-GD20 75 or 150 ppm
At 5 months 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 with CO-exposed rats.
March 2009
E-2
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Carmines and
Rajendran (2008)
Rat
(Sprague Dawley)
GD6-GD19 of
gestation for
2 h/day
600 ppm
Significant decreases in birth weight were reported after CO
exposure. Maternal body weight was unchanged during
gestation, but corrected terminal body weight (body weight
minus uterine weight) was significantly elevated in CO-exposed
dams at term.
Carratu et al. (1995) Rat
(Wistar)
150 ppm
Sphingoipid homeostasis was disrupted in male offspring of
prenatally exposed rats, without a disruption in motor function.
Carratu et al. (2000) Rat
(Wistar)
GD0-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, no changes in SA at 90 days of age. Motor
activity, which could be affected by changes in myelination,
showed no differences between CO and control animals at
90 days of age.
Carratu et al. (2000)
Rat GD0-GD20
(Wistar)
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.
Carraway et al. (2002)
Rat model of hypoxic 3 weeks
pulmonary vascular
remodeling
Strain of rat not stated
Hypobaric hypoxia ±
50 ppm
CO promoted remodeling and increased pulmonary vascular
resistance in response to HH. The number of small muscular
vessels was increased compared with HH alone. Changes in
cell proliferation, apoptosis, actin and HO-1 gene and protein
expression correlated with structural changes. COHb levels
were <0.5% in controls, 1.5-2.8% in the HH treatment group and
3.5-3.9% in the HH + CO treatment group.
Cella et al. (2006)
Rat
(Sprague Dawley)
HO-1 production and HO concentration were shown to be
regulated by estrogen in the rat uterus.
Chung et al. (2006)
Rat
(Sprague Dawley)
Male
3-6%
CO inactivation of Mb does not induce any change in the
respiration rate, contractile function, or high-energy phosphate
levels in perfused rat hearts.
Cronje et al. (2004) Rat
(Sprague Dawley)
Male
240-325 g
45 min
2,500 ppm
Results indicate that tissue and blood [CO] (66-72% COHb)
dissociate during CO inhalation, but tissue [CO] does not follow
blood [CO] or 1/p02 as in the Warburg theory during intake or
elimination. Tissue [CO] increases later during the resolution
period and varies significantly among animals and tissues. The
deviation from the predicted values in the brain is likely due to
the release of heme and increase in NADPH stimulating
endogenous CO production by HO. Immediately following expo-
sure, 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)
Cudmore et al. (2007) Human placenta
Human (HUVEC)
Mouse
(HO-1 deficient mouse
on 129/SV x C57BL/6
background)
Pig
(Porcine aortic
endothelial cells)
HUVEC cells, porcine aortic endothelial cells, HO-1 null mice
and placental villous explants (normotensive and pre-eclamptic
pregnancies) were used in this study. The HO-1/CO system
inhibited sFlt-1 and sEng release, two factors upregulated in
pre-eclampsia.
D'Amico et al. (2006) Human embryonic 0-30 min 20 |jM
kidney (HEK293) cells
Exogenous CO inhibited respiration in HEK293 cells under
ambient O2 concentration (21%). Inhibition was enhanced under
hypoxic conditions. Increased endogenous CO resulting from
HO-1 overexpression inhibited respiration by 12% and
cytochrome c oxidase activity by 23%. This effect was enhanced
under hypoxic conditions.
Dani et al. (2007)
Human
(neonatal blood)
CO was lower at birth and 48-72 hours postpartum in infants
born by elective C-section and higher in vaginally born infants.
March 2009
E-3
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
De Salvia et al. (1995)
Rat
(Wistar)
GD0-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 months) and
reacquisition (at 18 months) of conditioned avoidance behavior.
Denschlag et al.
(2004)
Human
Genetic polymorphisms in human HO-1 are linked to idiopathic
recurrent miscarriages.
Dewilde et al. (2001)
Nb exists as a reversibly hexacoordinated Hb type with a His-
Fe2*-His binding scheme. Dissociation of the internal ligand by
O2 or CO is the rate limiting step.
Dubuis et al. (2002)
Rat
(Wistar)
Adult female
250 g
3 weeks
530 ppm
Intrapulmonary resistance artery smooth muscle cells were
isolated from control and exposed rats. Electrophysiological re-
cordings provided evidence of increased Ca2*-activated K*
current consequent to chronic CO exposure. The authors
speculated that this could in part explain the vasodilatory effect
of CO in the pulmonary circulation.
Dubuis et al. (2005)
Rat
(Wistar)
Male
21 days
50 ppm
CO attenuated PAHT by activating BKca channels in PA
myocytes and reduced hemodynamic changes of PAHT.
Dubuis et al. (2003)
Rat
(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.
Favoryetal. (2006b) Rat 90 min 250 ppm CO inhibited myocardial permeabilized fiber respiration
250-300 a (complex IV), increased coronary perfusion pressure and left
. ventricular developed pressure (LVDP) first derivative and
Strain not stated 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 O2 supply-to-utilization which could potentially lead to
myocardial hypoxia because of the increased O2 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.
FechterandAnnau
(1977)
Rat
(Long Evans)
Continuous CO
exposure
throughout
pregnancy
150 ppm CO
The authors found a 5% significantly decreased birth weights at
PND1 in gestationally CO-exposed pups versus 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 dams versus air-exposed control dams.
Fechteretal. (1980)
Rat
(Long Evans)
Continous CO
exposure
throughout
pregnancy
150 ppm
CO-exposed animals had cardiomegaly at birth (wet heart
weight) that dissipated by PND4.
FechterandAnnau
(1980)
Rat
Continous CO
exposure
throughout
pregnancy
150 ppm
CO-exposed animals had decreased birth weight, impaired
righting reflexes, impaired negative geotaxis, and delayed
homing behavior.
Fechteretal. (1987)
Rat
(Long Evans)
Continuous CO
exposure
throughout
pregnancy or from
GD0 to PND10
75,150, or 300 ppm
The neostriatum of PND21 rat brains were collected and
showed disrupted development following CO exposure (GD0-
PND10 group, 300 ppm CO). Dopamine levels were also
significantly elevated in CO-exposed animals (GD0-PND10, 150
and 300 ppm CO).
March 2009
E-4
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure qq Concentration
Duration
Findings
Garofolo et al. (2002) Human infants
Rat
Rat: PND2-PND5
Human infants who die from SIDS show decreased brainstem
muscarinic receptor binding versus infants dying from other
causes. S-adrenergic modulation of muscarinic receptors in
developing heart was observed.
Rodent S-adrenergic agonists at PND2-PND5 induced
muscarinic receptor decrement in adenylyl cyclase.
Gautier et al. (2007)
Rat
(Wistar)
Adult male
Model of right
ventricular hypertrophy
secondary to chronic
hypoxia
3 weeks of HH ±
CO in final week
Or 1 week of CO
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 CO+ HH led to an in-
creased left ventricular shortening fraction (LVSF). CO alone led
to a decrease in LVSF and the mitral E-to-A ratio, indicative of a
LV filling impairment. Hypobaric hypoxia decreased the relative
RV perfusion and increased the relative LV perfusion. These
effects were prevented with concomitant exposure to CO
although exposure to CO alone had no effects on myocardial
perfusion. Morphologic and histologic analysis demonstrated RV
hypertrophy in both the HH group and the CO+HH group and
fibrotic lesions in the CO+HH group. The authors concluded that
the 1-week exposure to 50 ppm CO had a deleterious effect on
RV myocardial perfusion adaptation to chronic hypoxia and
pressure overload. Although the reduced RV pressure overload
was beneficial it was counterbalanced by impaired RV perfusion
and redistribution of perfusion toward the LV.
Gaworski et al. (2004)
Rat
(Sprague Dawley)
2 h/day,
7 days/week by
nose-only
inhalation
Males: 4 weeks
prior to and during
mating; and
Females: 2 weeks
prior to mating,
during mating,
and through
weaning to
PND21
Cigarette smoke:
150, 300, or 600 mg/m3
Total Particulate Matter
(TPM)
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.
Ghioet al. (2008)
Rat
(Sprague Dawley)
Adult male
Human bronchial
epithelial cells
(BEAS-2B)
24 h
2-24 h
50 ppm
10-100 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.
CO exposure for 24 h led to a dose-dependent decrease in
cellular non-heme iron, with the effect at 10 ppm statistically
significant and the effect at 50 ppm maximal. This effect was
reversible since removing the cells after 2 h of CO and incu-
bating 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) Rat From GD0-GD20 75 or 150 ppm This study showed that CO (75 and 150 ppm) exposed male
(Wistar) of pregnancy animals at 40 days of age had a significantly decreased time of
Male and exploration of novel objects. The 150 ppm CO group showed a
Pregnant female 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).
March 2009
E-5
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Giustino et al. (1993)
Rat
(Wistar)
GD0-20
75, or 150 ppm
CO exposure in utero led to a reversible and dose dependent
loss of function of splenic macrophages with decreased killing
ability, decreased phagocytosis, and decreased ROS production
during the macrophage respiratory burst.
Glabe et al. (1998)
Rat
(Sprague Dawley)
Male,
Myocardium
Pco = 0 -107 Torr
Increased Pco and increased COMb saturation did not alter high
energy phosphate signals (ATP, phosphocreatine, Pi). MV02
began to decline at 87.6% COMb and is likely not due to
cytochrome c oxidase inhibition.
Graver et al. (2000)
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
(Sprague Dawley)
Male
40 min
1,000-3,000 ppm
CO exposure increased extracellular dopamine levels and
decreased its major metabolites in a Na*-dependent pathway.
CO 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)
Pig
(Granulosa cells)
In this porcine model, HO was able to augment granulosa cell
apoptosis allowing for proper follicular maturation.
Hendlerand Baum
(2004)
Human
End-tidal breath CO measurements in pregnant women with
contractions (term and pre-term) were lower than those
measurements in non-contracting women.
Hofmann and Brittain
(1998)
Human
Partitioning of O2 and CO in the human embryonic Hb is dis-
cussed.
Iheagwara et al.
(2007)
Mouse
(C57BI6)
Male
3 h
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.
Imai et al. (2001)
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.
(1996)
Rat
(Wistar)
Male
200-290 g
60 min
40-60 min
1,000-3,000 ppm
1,000 ppm
CO poisoning resulted in free NO' in brains as measured by
electron paramagnetic resonance spectroscopy and in a 10-fold
increase in nitrotyrosine as measured by immunohistochemical
staining. These responses were blocked by pretreatment with a
NOS inhibitor but not by neutrophil depletion.
Brain nitrotyrosine formation was blocked by platelet depletion
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 for xanthine oxidase
activation.
The authors concluded that perivascular reactions mediated by
peroxynitrite are key to CO poisoning effects in brain.
Johnson and Johnson
(2003)
Rat
(Sprague Dawley)
Male
250-300 g
0-100 pM
CO produced a concentration dependent, endothelium-
dependent vasoconstriction in isolated gracilis muscle arterioles,
evident at 1 |jM CO. Pre-treatment with a NOS substrate
prevented this response while pretreatment with a NOS inhibitor
converted this response to a vasodilation. The authors
concluded that exogenous CO was acting through NOS
inhibition.
March 2009
E-6
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Johnson et al. (2003)
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)
Rat
(Sprague Dawley)
Male
Deoxycorticosterone
acetate (DOCA)-salt
hypertension model
WKY rats
Spontaneously
hypertensive rats
(SHR)
Salt-sensitive DOCA rats, but not SHR, had elevated aortic HO-
1 expression and blood COHb levels. Both had elevated mean
arterial BP compared with controls. Acetylcholine-mediated
vasodilation of isolated gracilis muscle arterioles was attenuated
in DOCA rats but not SHR. Pretreatment with a HO inhibitor
restored the response in DOCA rats. The authors concluded
that HO-1-derived CO contributes to endothelial dysfunction in
DOCA but not SHR.
Johnson et al. (2006)
Rat
(Zucker)
Lean and obese
Male
100 (jM 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)
Rat
(Wistar)
Pregnancy-induced modulation of calcium mobilization and
down-regulation of Rho-kinase expression contributed to
attenuated vasopressin-induced contraction of the rat aorta.
Khan et al. (2006)
Nb overexpressing
BDF x CD1 mice
Cerebral and myocardial infarcts were decreased in neuroglobin
overexpressing mice, decreasing ischemic injury.
Kim et al. (2005)
Primary rat pulmonary
artery smooth muscle
cells
Inbred LEW rat
(Sprague Dawley)
200-250 g
24 h or
pretreatment for
1-2 h followed by
24 h
posttreatment
250 ppm
Exposure of cells in culture to 250 ppm CO for 24-h inhibited
serum-stimulated cell proliferation, increased expression of
p21 wafi/ciPi anc| 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
p21 wafi/ciPi, CyC|jn 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.
March 2009
E-7
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Kim et al. (2008)
Primary rat
hepatocytes
Primary mouse
hepatocytes
Respiration-deficient
human Hep3B cells
10-60 min
250 ppm
Exposure of cells in culture to 250 CO for 1 h twice a day
prevented spontaneous hepatocyte death over 6 days in culture.
CO also decreased caspase-3 activity. Cell death was deter-
mined to be partly due to apoptosis. CO also increased ROS as
measured by dichlorofluorescein fluorescence in rat
hepatocytes, mouse hepatocytes and Hep3B cells but not in
respiration-deficient Hep3B 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
(gravid and non-gravid
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 et al. (2008)
Mouse
4 h
Diesel emissions: 350
pg/m3
Diesel exhaust enhanced vasoconstriction in veins but not
arteries. It was suggested that this is through the uncoupling of
eNOS.
Kreiseretal. (2004)
Human
End tidal CO concentrations were lower in pregnant women with
gestational hypertension and pre-eclampsia than normotensive
women.
Lash et al. (2003)
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)
Mouse
(ICR [CD-1 ])
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
upregualtion post maternal LPS treatment.
Loennechen et al.
(1999)
Rat
(Sprague Dawley)
Female
220-240g
1 week
1 week 100 ppm
and
1 week 200 pm
100ppm
100-200 ppm
Endothelin-1 expression increased by 53% and 54% in the left
and right ventricle respectively during the 2 week exposure and
by 43% and 12% in the left and right ventricle respectively
during the 1 week exposure. Right ventricular to body weight
ratio was increased by 18% and 16% in the 2 week and 1 week
exposure groups respectively. COHb levels were 23% and 12%
in the 2 week and 1 week exposure groups respectively.
Longo et al. (1999)
Rat
(Sprague Dawley)
Human
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.
Lopez et al. (2008)
Rat
(Sprague-Dawley)
Pregnant rats
exposed to CO
GD5-GD20
(Group A) or
GD5-GD20 plus
PND5-PND20
(Group B);
Group C (control
air exposure).
10 -18 h/day
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 show vacuolization of
afferent terminals at the base of the cochlea. At PND3, group A
shows decreased synapsin-1 staining of the efferent nerve
terminals. At PND20, groups A and B show decreased neurofila-
ment-IR (staining) in type I spiral ganglia neurons and afferent
nerve fibers. At PND12 and PND20, group B shows increased
HO-1 and SOD-1-IR in blood vessels of the stria vasularis;
group A is similar to controls. From PND3-PND20, there is
increased iNOS and increased nitrotyrosine-IR in blood vessels
of the cochlea.
Lopez et al. (2003)
Rat
(Sprague-Dawley)
PND6 to weaning
(PND19-PND20)
12 or 25ppm
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 versus control.
Expression of the calcium-mediated myosin ATPase in the
organ of corti and spiral ganglion neurons was significantly de-
creased in the 25 ppm CO exposure group versus controls.
March 2009
E-8
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Lund et al. (2007)
Mouse
(ApoE+)
Male
High fat diet
6 h/day,
7days/week,
7 weeks
8, 40, or 60 pg/m3 PM
whole gasoline exhaust;
or filtered exhaust with
gases matching the 60
pg/m3 concentration. CO
concentrations were 9,
50, and 80 ppm
corresponding to the 8,
40, and 60 pg/m3 PM
whole exhaust
exposures
Both whole and filtered exhaust increased aortic mRNA
expression of matrix metalloproteinase-3 (MMP-3), MMP-7, and
MMP-9, tissue inhibitor of metalloproteinases-2, endothelin-1
and HO-1 at 60 pg/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.
Lund et al. (2009)
Mouse
(ApoE-'-)
Male
High fat diet
6 h/day, 1 or
7 days
Gasoline engine exhaust
containing 60 pg/m3 PM
and 80 ppm CO
Gasoline exhaust exposure increased aortic MMP-2/9 activity at
1 and 7 days. Protein levels of aortic MMP-9, MMP-2, TMP-2
and plasma MMP-9 were also increased after 7 days. Lipid
peroxidation in aorta resulting from gasoline exhaust exposure
was inhibited by treatment with the antioxidant Tempol, while
increases in mRNA for ET-1 and MMP-9 in aortas were inhibited
by treatment with BQ-123, an antagonist of ETa receptor.
Treatment with BQ-123 also reduced aortic MMP-2/9 activity in
aortas following gasoline exhaust exposure. The authors
concluded that ETa receptor pathway is a key mediator of
gasoline engine exhaust effects in the vasculature.
Lyall and Myatt (2002)
Human
Women with pre-eclampsia, produced term placenta with
significant decreases in HO-2 versus women with healthy
pregnancies.
Lyallet al. (2000)
Human
(placentas from 8-19
weeks pregnancy and
term placentas)
The use of a HO inhibitor, ZnPP, increased placental perfusion
pressure. HO-1 and HO-2 were expressed in the placenta and
placental bed and vary in expression over the course of
pregnancy. HO may thus be involved in trophoblast invasion,
placental function and perfusion pressure.
Mactutus and Fechter
(1984)
Rat
(Long Evans)
Continous
exposure
to CO over
gestation
150 ppm
Acquisition as measured in a two-way conditioned avoidance
(flashing light warnings followed by mild footshock) test failed to
improve with age of in utero CO-exposed (150 ppm, dam COHb
15%) rats (male and female offspring) in contrast to air-exposed
controls who improved with age/maturation, indicating a failure
in the associative process of learning. The authors also found
impairments in reacquisition performance, an index of retention,
in PND31 rats that had received continuous in utero CO
exposure. Prenatal CO exposure induced learning and memory
deficits in male and female offspring.
McGregor and
Westcott (1998)
Guinea pig
GD23-GD25 until
term
(approximately
68 days)
10 h/day
200 ppm
Aberrant respiratory responses (to asphyxia and CO2) of
offspring with prenatal CO exposure. The authors hypothesized
this may be related to changes in the brainstem. COHb in
maternal (8.53 ± 0.6% versus 0.25 ± 0.1%) and fetal blood
(13.0 ± 0.4% versus 1.6 ± 0.1%) from CO-treated versus
controls.
McLaughlin et al.
(2001)
Human
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 chroionic villi of term placenta.
McLaughlin et al.
(2000)
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)
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 versus normotensive pregnancies.
McLean et al. (2000) Human placenta
HO activity was highest in the placenta near term.
March 2009
E-9
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Melin et al. (2002)
Rat 10 weeks
(Dark Agouti)
Male
Model of right ventricle
hypertrophy secondary
to chronic hypoxia
(HH 10 weeks)
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. Echo-
cardiography 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/cHrv,, -dP/dtm) and right ventricular work (RVW).
CO significantly enhanced the effects of HH on RVEDP and
significantly diminished the effects of HH on dP/dtm 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)
Rat
(Dark Agouti)
Male and female
Model of right ventricle
hypertrophy secondary
to chronic hypoxia (HH,
10 weeks)
Half of the animals
were exercise- trained
to induce LV
hypertrophy
10 weeks
50 ppm alone
or concomitant with HH
In untrained animals, combined treatment with HH+CO led to
increased LV+S and RV weights compared with HH treatment
alone. HH+CO led to several changes in measured
echocardiographic parameters including increased anterior and
posterior wall thickness in diastole (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.
Mereu et al (2000)
Rat
(Wistar)
GD0-GD20
continuous CO
exposure
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.
Middendorff et al.
(2000)
Human
Adult males aged 65-
75. Testicular tissue
from orchiectomy
Zn protoporphryin (ZnPP) and Hb both significantly reduced
seminiferous tubular cGMP generation, suggesting a role for CO
in human testicular tissue.
Naik and Walker
(2003)
Rat
(Sprague-Dawley)
Male
210 (jL of CO/100 mL
of physiological
saline solution
Endogenous CO mediated vasorelaxation involved cGMP-
independent activation of vascular smooth muscle large-
conductance Ca2*-activated K* channels. However exogenous
CO vasodilation was cGMP dependent.
Ndisang et al. (2004)
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.
Neggers and Singh
(2006)
Mouse
(CD-1)
GD7-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 versus
8% and 16% protein groups. Further, dietary restriction of both
protein and zinc with CO-exposure to CO during gestation
increased the incidence of pup mortality and malformations
including gastroschisis. Zinc supplementation to protein
deficient diet in CO-exposed mice decreased fetal mortality and
malformation.
March 2009
E-10
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Newby et al. (2005)
Human
(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
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
(Wistar)
Adult male
The role of HO-1 in spermatogenesis was explored. CdCh
induced testicular HO-1 and reduced HO-2 protein in rats.
Pretreatment with ZnPPIX attenuated CdCh-dependent
apoptosis. Leydig cells use HO-1 derived CO to tirgger
apoptosis of pre-meiotic germ cells and modulate spermato-
genesis under CdCh dependent oxidative stress.
Patel et al. (2003)
Rat
(Sprague Dawley)
Male
262 ± 30 g
Isolated hearts
30 min
Buffer saturated with
0.01 and 0.05% CO
The ventricular glutathione content, both reduced and oxidized,
decreased by 76% and 84% 90 min post-exposure to 0.01%
and 0.05% CO, respectively. Treatment with antioxidants
partially blocked the decreases in glutathione. Increased
creatine kinase activity was observed in heart perfusate during
and after treatment.
Penney et al. (1983)
Rat
(strain not reported)
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.
Piantadosi (2002)
Reviews the biochemical activities of CO, including various
heme protein binding. The review stresses the importance of the
CO/O2 ratio in determining the physiological effects of CO.
Piantadosi (2008)
Reviews the physiologic responses to exogenous and
endogenous CO and biochemical effects including the binding
to heme proteins, the generation of reactive O2 species and
activation related signaling pathways.
Piantadosi et al.
(2006)
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 mitochon-
dria from MPT. This effect was restored in the presence of a
strong reductant. The authors conclude that CO causes
mitochondrial pore stress independently of its hypoxic effects
Prigge and Hochrainer
(1977)
Rat
(Wistar, SPF)
GD0-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 (NOAEL 125 ppm CO).
Raub and Benignus
(2002)
Reviews the physiology of CO and the effects on the nervous
system. It is estimated that COHb would have to rise to 15-20%
before a 10% reduction in any behavioral or visual
measurement could be observed.
Richardson et al.
(2002)
Human
Male
20% COHb
20% COHb did not influence 02Mb binding indicated by
unaltered deoxy-myoblobin signal. Resting skeletal muscle
metabolic rate was unaffected by 20% COHb. VO2 max was
decreased. No decrement in intracellular PO2 was found. 20%
COHb altered exercising bioenergetics, pH, PCr, and ATP
levels.
Ryter et al. (2006)
Reviews the basic science of exogenous and endogenous CO
including HO-1 regulation. It also reviews some therapeutic
applications for CO.
March 2009
E-11
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species / Model
Exposure
Duration
CO Concentration
Findings
Sartiani et al. (2004)
Rat
(Wistar)
In utero inhalation
exposure
150 ppm
At 4 weeks 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 Ik, (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 weeks of age. All of these CO-dependent
changes were no longer different from controls at 8 weeks of
age, showing a delayed maturation.
Schwetz et al. (1979)
Mouse
(CF-1)
Rabbit
(New Zealand)
7 or 24-h/day
GD6-GD15 (Mice)
GD6-GD18
(Rabbits)
250 ppm
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 versus control. Increased birth
weight in mice exposed to 7 h/day CO versus controls. No
similar effects were seen in rabbits.
Singh et al. (1992)
Mouse
(CD-1)
GD8-GD18
65,125, or 250 ppm
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 have a synergistic effect on offspring survival and an
additive effect on malformations.
Singh (2006)
Mouse
(CD-1)
6 h/day during
the first
2 weeks
of pregnancy
65 or 125 ppm
Modulating dam protein intake during in utero CO exposure
altered pup mortality.
Sitdikova et al. (2007)
Frog neuro-muscular
junctions
20 min
96 pM
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)
Fluman
Primary human aortic
smooth muscle cells
0-48 h
10-250 ppm
CO inhibited SMC proliferation at concentrations from 50-
500 ppm. The cell cycle arrest occurred at the G0/G1 phase of
the cell cycle. CO increased expression of the cell cycle inhibitor
p21c'P1 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
(Wistar)
Female
169 ± 4.5 g
20 h/day,
x 5 days/week,
x 72 weeks
200 ppm
COHb was 14.7% in CO-exposed animals and 0.3% in controls.
Total Fib 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.
Stocka rd-S u 11 ivan et al.
(2003)
Rat
(Sprague-Dawley)
22 h/day, PND6-
PND22
12, 25, 50, or 100 ppm
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)
versus controls at PND22.
March 2009
E-12
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Suliman et al. (2007)
Mouse 1 h
(C57BL/6 wild-type and
eNOS deficient)
Male
Rat
(embryonic
cardiomyocytes H9c2
cells)
50-1,250 ppm
OrHH
Or 100 mM
dichloromethane
1-h exposure of mice to 1,250 ppm CO increased cardiac
mitochondrial content of all 5 respiratory complexes 24 h later.
The volume density of interfibrillar mitochondria was increased
by 30% after 24 h demonstrating that CO caused cardiac
mitochondrial biogenesis. The CO concentration in heart
increased from 9 pmol/mg to 50-150 pmol/mg in mice exposed
to 50-1,250 ppm CO for 1 h. These levels declined to baseline
by 6 h. Peroxisome proliferator-activated receptor gamma
coactivator 1 alpha (PGC-1 a) expression was increased 6 h
following exposure to 50-1,250 ppm CO. Expression of DNA
polymerase and mitochondrial transcription factor A (TFAM) was
increased 6 and 24 h after exposure, while mitochondrial DNA
was increased 2-3 fold 24 h after exposure. CO activated gene
expression of these proteins involved in cardiac mitochondrial
biogenesis beginning at 2 h post exposure for PGC-1a, nuclear
respiratory factors 1 and 2 (NRF-1 and -2) and at 6 h
postexposure for TFAM. These effects were independent of
NOS and not seen with HH. CO exposure resulted in
phosphorylation of p38 MAPK and Akt at 2 and 6 h post-
exposure to 1,250 ppm CO for 1 h. Inhibition of p38 activation
failed to inhibit the CO-mediated increase in cardiac
mitochondrial biogenesis.
In cell culture experiments, CO derived from dichloromethane
metabolism resulted in increased cGMP, protein levels of SOD2,
TFAM, NRF-1, NRF-2, PGC-1,mitochondrial ROS, Akt
phosphorylation, and mitochondrial DNA. Inhibition of GC or
PI3K/Akt, but not p38, blocked the responses to CO. A role for
mitochondrial H2O2 in Akt regulation was demonstrated.
Mitochondrial H2O2 and the PI3K/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 et al. (2001)
Mouse
Neuronal cultures
prepared from the
cerebral hemispheres
of 16-day Charles River
CD1 mouse embryos
Nb expression was increased by neuronal hypoxia in vitro and
focal cerebral ischemia in vivo. Inhibiting Nb reduced neuronal
survival after hypoxia whereas Nb overexpression enhanced
neuronal survival.
Tattoli et al. (1999)
Rat
(Wistar)
Male and
Pregnant female
PND1-PND10 75 and 150 ppm
Cognitive function was assessed in rats after postnatal CO
exposure at 3 and 18 months of age. Postnatal CO exposure
did not affect the acquisition and reacquisition of an active
avoidance task. This is different from previous findings by the
same laboratory indicating that in utero exposure to CO (75 and
150 ppm) induced long-lasting learning and memory deficits.
Telferetal. (2001) Human
Myometrium tissue
obtained from gravid
[pre-term (25-34 weeks
gestation), term not in
labor or term in labor]
and non-gravid women
cGMP was monitored in various myometrial tissues. cGMP was
significantly higher than that from nonpregnant tissue and
decreased at term, especially in tissue from laboring women.
Teran et al. (2005)
Rat
(Dahl/Rapp
salt-sensitive rats)
Male
100 pM
A high salt diet for 1-4 weeks 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 weeks. 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 weeks 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.
March 2009
E-13
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
1 h
1,000 ppm
OR
OR
>1 h
1,000-3,000 and
30 min
higher ppm
0.5 mL of pure CO
Thorn et al. (1994)
Rat
(Wistar)
Male
Isolated blood cells
CO poisoning inhibited B2 integrin-dependent PMN adherence
in heparinized blood obtained from rats immediately after expo-
sure. 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
Ischiropoulos (1997)
Rat
(Wistar)
Male
200-290 g
Platelet-rich plasma
from rats was used as
the source of platelets
Bovine pulmonary
artery endothelial cells
1 h
30 min or 2 h
1 h
20-1,000 ppm
10-20 ppm
10-100 ppm
Platelets isolated from rats exposed to 20-1,000 ppm CO for 1-h
released NO' in a dose-dependent manner. COHb levels were
0.7% in controls, and 3.2%, 7.8% and 51.0% in 20,100 and
1,000 ppm exposure groups respectively.
Isolated platelets released NO' when incubated for 30 min with
20-100 ppm CO. NOS activity was not enhanced by 100 ppm
CO. Platelets released NO' in response to 10-100 ppm CO after
30 min pretreatment with a NOS inhibitor, suggesting that CO
displaces NO' from heme-binding sites. Longer incubations (2
h) with the NOS inhibitor led to a diminished response to
100 ppm CO. There appears to be a discrepancy in the results
depending on how NO' was measured (electrode versus 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.
March 2009
E-14
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Thorn et al. (1997) Bovine pulmonary 30min-4h 10-100 ppm (11-110 nM) 1-h exposure to 111-110 nM CO led to a dose-dependent
artery endothelial cells increase in NO' release, as measured by nitrite+nitrate. Signifi-
cance was achieved at 22 nM (corresponding to an interstitial
partial pressure of 20 ppm and a blood COHb level of 7%). NOS
inhibition blocked the response to 110 nM CO. A dose-
dependent increase in cellular nitrotyrosine was also observed
following a 2-h exposure to CO, with significance achieved at 55
nM CO. NOS inhibition blacked the response to 110 nM. CO
exposure failed to decrease the concentration of reduced
sulfhydryls, but did result in the extracellular release of a short-
lived oxidant species which was blocked by NOS inhibition.
Dihydrorhodamine oxidation, a measure of peroxynitrite
formation, occurred in response to 110 nM CO, an effect which
blocked by NOS inhibition. Cytotoxicity of CO was evaluated by
the release of 51chromium. 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, al-
though NOS inhibition had protective effects under conditions of
continuous CO exposure of 4 h. Exposure to 110 nM CO for 2 h
led to the loss of membrane integrity, measured by ethidium
homodimer-1 staining, 18 h later. This response was blocked by
NOS inhibition. Exposure to 110 nM CO had no effect on O2
consumption, production of intracellular H202or cellular redox
activity. Exposure to 110 nM did not alter arginine transport or
NOS activity. NO' release from cells which had been pre-treated
with a NOS inhibitor and then exposed briefly to 5% CO was
measured using a NO-selective electrode suggesting that CO
competed with intracellular binding sites of NOV
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, et al. (1999b)
Rat
(Wistar)
Male
200-290 g
Some rats fed a high
cholesterol diet
1 h
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.
Platelet 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 6-fold increase
in LDL oxidation immediately after 1-h exposure to 100 ppm CO.
This effect was not blocked by NOS inhibition.
The authors concluded that CO can alter vascular status by
several mechanisms linked to NO'-derived oxidants.
March 2009
E-15
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Thorn et al. (1999a)
Rat
(Wistar)
Male
200-290 g
1 h
50-1,000 ppm Leakage of albumin into lung parenchyma occurred 18 h, but
not at earlier time points, after rats were exposed to 100 ppm
CO for 1 h. 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 H2O2 was elevated by exposure to 100 ppm CO
for 1 h and this effect was blocked when rats were pretreated
with a NOS inhibitor. Elevated nitrotyrosine content was
observed in lung homogenates 2-4 h following 1-h exposure of
rats to 100 ppm CO. This effect was also blocked by
pretreatment with a NOS inhibitor. No leukocyte sequestration
was observed in Iungs18 h following exposure to 100 ppm CO.
CO-induced lung leak was not affected by neutrophil depletion.
The authors concluded that CO causes lung vascular injury
which is dependent on NO'.
Thorn et al. (2000)
Bovine pulmonary
artery endothelial cells
40 min-2 h
11-110 nM
(10-100 ppm)
Increased uptake ofethidium homodimer-1, a measure of
decreased membrane integrity and cell death, was observed in
endothelial cells 18 h after exposure to 110 nM for 60-120 min.
Exposures of 20-40 nM were ineffective in this regard. Ethidium
uptake was also increased by 2-h exposure to 88 nM CO.
Preincubation for 2 h with an inhibitor of eNOS, an antioxidant,
and an inhibitor of peroxynitrite reactions blocked the
CO-mediated cell death. Morphological changes in cells were
observed 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-
exposure of cells to 11 nM CO for 40 min followed by a 3-h
incubation period resulted in an increased level of MnSOD and
protection against cell death 18 h following a 2-h exposure to
110 nM CO. The authors concluded that exposure to 11 nM CO
led to an adaptive response which protected cells from injury
and apoptosis resulting from NO'-derived oxidants.
Thorn et al. (2001a) Rat
Until lost 1,000-3,000 ppm Neutrophils sequestration was observed in the brain vessels of
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)
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.
March 2009
E-16
DRAFT-DO NOT QUOTE OR CITE
-------
Reference Species I Model n *¦ CO Concentration
~ niiratmn
Findings
Thorup et al. (1999) Rat 0.01 -10 pM Perfusion of Isolated rat renal resistance arteries with
(Sprague Dawley) CO-containing buffer (0.001-10 pM) led to the biphasic release
Males of NO', peaking at 100 nM and declining to undetectable
200-250 g responses at 10 pM. 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. (2000b)
Guinea pig
10 h/day over the 200 ppm
last 60% of
gestation
Fetal and maternal COHb were 13% and 8.5% respectively.
Neurotransmitter systems were affected after CO exposure. The
catecholaminergic system of the brainstem displayed significant
decreases in immunoreactivity for tyrosine hydroxylase (TH),
which was likely due to decreased cell number in specific
medullar regions. The cholinergic system was also affected by
prenatal CO exposure with significant increases in ChAT
immunoreactivity of the medulla and no changes in muscarinic
acetylcholine receptor.
Tolcos et al. (2000a)
Guinea pig
10 h/day for the 200 ppm
last 60% of
gestation
Brains were collected at 1 and 8 weeks of age. These data
showed that CO exposure in utero sensitized the brain to hyper-
thermia at PND4 leading to generation of necrotic lesions in the
brain and changes in neurotransmitter levels.
Tschuqquel et al.
(2001)
Human
HUVEC
CO was generated by primary endothelial cells from human
umbilical veins and uterine arteries after exogenous 17-S estra-
diol administration.
Vallone et al. (2004)
Mouse protein
The authors present the X-ray structure of CO-bound ferrous
murine Nb. When CO binds, the heme group slides deeper into
the protein crevice.
Vreman et al. (2000)
Human umbilical cord
(artery and vein)
Rat
Aorta, vena cavae, liver
and heart
HO activity was quantified in human umbilical cord and in the rat
vasculature (aorta and vena cavae). Human umbilical artery and
vein HO activity were equal. The rat aorta and vena cavae
produced equal amounts of HO activity (wet weight/g tissue) but
generated 3x greater HO than the heart and 0.2x of the liver.
HO activity in rat vasculature was 3x that of the human cord
tissues. Use of the HO inhibitor CrMP effectively blocked HO
activity in the rat liver and heart but was less effective at
blocking HO activity in the human umbilical cord or the rat
vasulature (only 50% effective). The activity of HO in the
umbilical vessels may provide a role for CO in control of
vasculature tone during pregnancy.
March 2009
E-17
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Vreman et al. (2005) Mouse
(BALB/c)
30 min 500 ppm
OR
Heme arginate
30 pmol/kg body weight
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 3-fold increase in CO
excretion reaching a maximaum 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 etal.(2007) Human
Acute CO poisoning Mean COHb in humans with acute CO poisoning was 35%.
Hyperbaric O2 reduces cognitive sequelae in a randomized
clinical trial of CO-poisoned patients. Risk factors for cognitive
sequelae without hyperbaric O2 included older age and longer
CO exposures. Patients with loss of consciousness or high
initial COHb levels should also be treated with hyperbaric O2.
Webber et al. (2003) Rat
Strain not stated
PND8-PND22 12.5, 25, or 50 ppm
Immunostating ofc-Fos, a marker of neuronal activation in the
nervous system was followed. C-Fos immunoreactivity in the
central IC 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 IC
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.
Webber et al. (2005) Rat
Strain not stated
PND9-PND24 25 or 100 ppm
Neurofilament loss from the spiral gangilian 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 IC, but c-fos levels were un-
changed after low iron diet concomitant with CO exposure
(ARIDCO).
March 2009
E-18
DRAFT-DO NOT QUOTE OR CITE
-------
Reference
Species I Model
Exposure
Duration
CO Concentration
Findings
Wellenius et al. (2004)
Rat
(Sprague Dawley)
250 g
Diazepam-sedated
Model of acute Ml
induced by
thermocoagulation
1 h, 12-18 h after
surgery
35 ppm
CO exposure decreased ventricular premature beat frequency
by 60.4% during the exposure period compared to controls. 1-h
exposure to CAPs (318 |jg/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.
(2006b)
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 et al. (2001) Human
HO localization in human endometrium and its changes in
expression over the menstral cycle were explored in this study.
HO-1 was constitutively expressed throughout the menstral
cycle and HO-2 was greater in the secretory than the
proliferative phase of the menstral 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.
Zamudio et al. (1995) Human
Women living at high altitude had an increased risk of adverse
pregnancy outcomes versus women living at lower altitudes.
Zenclussen et al.
(2006)
Mouse
(CBA/J x DBA/2 J)
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) Rat 8-28 h 15 ppm Exposure to 15 ppm CO during anoxia resulted in decreased
(pulmonary artery £hTfTWao?? °lSTfT 1 and lncreased phosphorylation of
endothelial cells) STAT3 at 8-24 h. Similar responses were observed when 24 h
anoxia was followed by a period of reoxygenation (0.5-4 h).
DNA binding of STAT1 was decreased while that of STAT3 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 PI3K and 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 STAT3 pathways.
The authors concluded that CO blocks anoxia-reoxygenation
mediated apoptosis through modulation of PI3K/Akt/p38 MAPK
and STAT1 and STAT3.
Zhang et al. (2007)
Mouse
A single dose of LPS administered to pregnant mice induced up-
regulation of HO-1 but not HO-2 in the mouse placenta 12-48 h
post-LPS treatment. Pre-treatment of mice with the spin trap
agent PBN or the TNFa inhibitor pentoxifylline prevented the
LPS-dependent HO-1 upregulation. Thus ROS may mediate the
LPS-dependent upregulation of HO-1.
Zhao et al. (2008)
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.
Zhuo et al. (1993)
Rodent
Hippocampal LTP of brain sections is significantly affected by
CO exposure with ZnPP IX, a HO inhibitor, blocking
hippocampal LTP.
March 2009
E-19
DRAFT-DO NOT QUOTE OR CITE
-------
Reference Species I Model n *¦ CO Concentration
~ niiratmn
Findings
Zuckerbraun et al. Macrophages 10 min - 24 h 50-500ppm
(2007) RAW 264.7
THP-1 cells, wild-type
and respiration-
deficient
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 antimycin C, 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.
March 2009
E-20
DRAFT-DO NOT QUOTE OR CITE
-------
References
Abi Esber AL; El-Fadel M; Shihadeh A. (2007a). 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.
Abi Esber LA; El-Fadel M; Nuwayhid I; Saliba N. (2007b). The effect of different ventilation modes on in-vehicle
carbon monoxide exposure. Atmos Environ 41: 3644-3657.
Abi Esber LA; El-Fadel M. (2008). In-vehicle CO ingression: Validation through field measurements and mass
balance simulations. The Science of the total environment 394: 75-89.
Abram SR; Hodnett BL; Summers RL; Coleman TG; Hester RL. (2007). Quantitative Circulatory Physiology: an
integrative mathematical model of human physiology for medical education. Advances in Physiology
Education 31: 202.
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.
Achouh PE; Simonet S; Fabiani JN; Verbeuren TJ. (2008). Carbon monoxide induces relaxation of human internal
thoracic and radial arterial grafts. Interactive Cardiovascular and Thoracic Surgery 7: 959-962.
Adams KF; Koch G; Chatteijee B; Goldstein GM; O'Neil JJ; Bromberg PA; Sheps DS. (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.
Adams SM. (2003). Establishing causality between environmental stressors and effects on aquatic ecosystems. Hum
Ecol Risk Assess 9: 17-35.
Adir Y; Merdler A; BenHaim S; Front A; Harduf R; BittermanH. (1999). Effects of exposure to low concentrations
of carbon monoxide on exercise performance and myocardial perfusion in young healthy men.
Occupational and Environmental Medicine 56: 535-538.
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.
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 (Noisy-le-grand) 51: 447-451.
Akland GG, Hartwell, T.D., Johnson, T.R., Whitmore, R.W. (1985). Measuring human exposure to carbon
monoxide in Washington, D.C. and Denver, Colorado during the winter of 1982-1983. Environmental
Science and Technology 19: 911-918.
Alcantara RE; Xu C; Spiro TG; Guallar V. (2007). A quantum-chemical picture of hemoglobin affinity. Proceedings
of the National Academy of Sciences 104: 18451-18455.
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.
Alexander PG; Chau L; Tuan RS. (2007). Role of nitric oxide in chick embryonic organogenesis and
dysmorphogenesis. Birth Defects Res A Clin Mol Teratol 79: 581-594.
Alexandreanu IC; 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.
Alexandreanu IC; Lawson DM. (2003a). Heme oxygenase in the rat anterior pituitary: immunohistochemical
localization and possible role in gonadotropin and prolactin secretion. Exp Biol Med (Maywood) 228: 64-
69.
Alexandreanu IC; Lawson DM. (2003b). Heme oxygenase in the rat ovary: immunohistochemical localization and
possible role in steroidogenesis. Exp Biol Med (Maywood) 228: 59-63.
AllredEN; BleeckerER; ChaitmanBR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; SelvesterRH; 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.
AllredEN; BleeckerER; ChaitmanBR; Dahms TE; Gottlieb SO; Hackney JD; Pagano M; SelvesterRH; Walden
SM; Warren J. (1991). Effects of carbon monoxide on myocardial ischemia. Environ Health Perspect 91:
89-132.
Aim S; Jantunen MJ; Vartiainen M. (1999). Urban commuter exposure to particle matter and carbon monoxide
inside an automobile. J Expo Anal Environ Epidemiol 9: 237-244.
March 2009
R-1
DRAFT-DO NOT QUOTE OR CITE
-------
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.
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.
Alonso JR; CardellachF; Lopez S; Casademont J; Miro O. (2003). Carbon monoxide specifically inhibits
cytochrome c oxidase of human mitochondrial respiratory chain. Pharmacol Toxicol 93: 142-146.
Altshuller AP. (1991). Estimating product yields of carbon-containing products from the atmospheric
photooxidation of ambient air alkenes. Journal of Atmospheric Chemistry 13: 131-154.
American Thoracic Society. (1985). Guidelines as to what constitutes an adverse respiratory health effect, with
special reference to epidemiologic studies of air pollution. Am Rev RespirDis 131: 666-668.
American Thoracic Society. (2000). What constitutes an adverse health effect of air pollution? . American Journal of
Respiratory and Critical Care Medicine 161: 665-673.
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.
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.
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. Occupational and Environmental Medicine 58: 504-510.
Andersson JA; Uddman R; Cardell LO. (2000). Carbon monoxide is endogenously produced in the human nose and
paranasal sinuses. J Allergy Clin Immunol 105: 269-273.
Andreae MO. (1991). Biomass burning: Its history, use, and distribution and its impact on environmental quality and
global climate. Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications (JS Levine,
Ed.). In Ecological studies, Voll. 84. Analysis and synthesis. The MIT Press, Cambridge.
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: H223-230.
Antonelli T; Tomasini MC; Tattoli M; Cassano T; Finetti S; Mazzoni E; Trabace L; Carratu MR; Cuomo V;
Tanganelli S; Ferraro L. (2006). Prenatal exposure to the cannabinoid receptor agonist WIN 55,212-2 and
carbon monoxide reduces extracellular glutamate levels in primary rat cerebral cortex cell cultures.
Neurochem Int 49: 568-576.
Appel KW; Bhave PV; Gilliland AB; Sarwar G; Roselle SJ. (2008). Evaluation of the community multiscale air
quality (CMAQ) model version 4.5: Sensitivities impacting model performance; Part II—particulate matter.
Atmos Environ 42: 6057-6066.
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: H2055-2059.
Ashfaq M; Janjua MZ; Nawaz M. (2003). Effects of maternal smoking on placental morphology. J Ayub Med Coll
Abbottabad 15: 12-15.
Astrup P; Olsen HM; Trolle D; Kjeldsen K. (1972). Effect of moderate carbon-monoxide exposure on fetal
development. Lancet 2: 1220-1222.
Atkinson RW; Anderson HR; Strachan DP; Bland JM; Bremner SA; Ponce de Leon A. (1999). Short-term
associations between outdoor air pollution and visits to accident and emergency departments in London for
respiratory complaints. EurRespir J 13: 257-265.
Ayres SM; Giannelli S, Jr.; Mueller H; Criscitiello A. (1973). Myocardial and systemic vascular responses to low
concentration of carboxyhemoglobin. Ann Clin Lab Sci 3: 440-447.
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. Journal of
Thrombosis and Haemostasis 5: 252-260.
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.
Bainbridge SA; Smith GN. (2005). HO in pregnancy. Free Radic Biol Med 38: 979-988.
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.
March 2009
R-2
DRAFT-DO NOT QUOTE OR CITE
-------
Bakwin PS; Tans PP; Novelli PC. (1994). Carbon monoxide budget in the Northern Hemisphere. Geophys Res Lett
21.
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.
Baldauf RW, N. Watkins, D. Heist, C. Bailey, P. Rowley, R. Shores. 2009. Near-Road Air Quality Monitoring:
Factors Affecting Network Design and Interpretation of Data, J. of Air Quality, Atmosphere, & Health (in
press). (In press). Near-road air quality monitoring: Factors affecting network Design and Interpretation of
data. J of Air Quality, Atmosphere, & Health.
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.
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 admissions association in Spain: results within the EMECAS project. J
Epidemiol Community Health 60: 328-336.
Bamberger AM; Koglin M; 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 alpha(2)
subunit. J Clin Endocrinol Metab 86: 909-912.
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.
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: 1158-1168.
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.
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.
Bathoorn E; Slebos DJ; Postma DS; Koeter GH; van Oosterhout AJ; van der Toorn M; Boezen HM; Kerstjens HA.
(2007). Anti-inflammatory effects of inhaled carbon monoxide in patients with COPD: a pilot study. Eur
Respir J 30: 1131-1137.
BaumM; Schiff E; KreiserD; Dennery PA; Stevenson DK; Rosenthal T; SeidmanDS. (2000). End-tidal carbon
monoxide measurements in women with pregnancy-induced hypertension and preeclampsia. Am J Obstet
Gynecol 183: 900-903.
Beard RR; Wertheim GA. (1967). Behavioral impairment associated with small doses of carbon monoxide. In
American Journal of Public Health (Vol. 57, pp. 2012).
Bedeschi E; Campari C; Candela S; Collini G; Caranci N; Frasca G; Galassi C; Francesca G; Vigotti MA. (2007).
Urban air pollution and respiratory emergency visits at pediatric unit, Reggio Emilia, Italy. J Toxicol
Environ Health A 70: 261-265.
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.
Bell MC; Namdeo AK; Chen H; Ashmore MR; Terry AC; Dimitroulopoulou S. (2004). Reducing Urban Pollution
Exposure from Road Transport (RUPERT). In Urban Transport (pp. 829-838). Southampton, UK: WIT
Press.
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and
Massachusetts. Environ Health Perspect 115: 1118-1124.
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions
in Taipei, Taiwan. Occupational and Environmental Medicine 65: 104-111.
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.
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.
Benignus VA. (1993). Importance of experimenter-blind procedure in neurotoxicology. Neurotoxicol Teratol 15: 45-
49.
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.
March 2009
R-3
DRAFT-DO NOT QUOTE OR CITE
-------
Benignus VA; Coleman T; Eklund CR; KenyonEM. (2006). A general physiological and toxicokinetic (GPAT)
model for simulating complex toluene exposure scenarios in humans. Toxicol Mech Methods 16: 27-36.
Bergamaschi P; Hein R; Heimann M; Crutzen PJ. (2000). Inverse modeling of the global CO cycle 1. Inversion of
CO mixing ratios. Journal of Geophysical Research 105: 1909-1927.
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; Wichmann HE; Peters A. (2006). Runs of
ventricular and supraventricular tachycardia triggered by air pollution in patients with coronary heart
disease. J Occup Environ Med 48: 1149-1158.
Bergeron M; Ferriero DM; Sharp FR. (1998). Developmental expression of heme oxygenase-1 (HSP32) in rat brain:
an immunocytochemical study. Brain Res Dev Brain Res 105: 181-194.
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.
Berk PD; Blaschke TF; Scharschmidt BF; Waggoner JG; Berlin NI. (1976). A new approach to quantitation of the
various sources of bilrubin in man. J Lab Clin Med 87: 767-780.
Berkowitz GS; PapiernikE. (1993). Epidemiology of Preterm Birth. Epidemiol Rev 15: 414-443.
Biggeri A; Baccini M; Bellini P; Terracini B. (2005). Meta-analysis of the Italian studies of short-term effects of air
pollution (MISA), 1990-1999. Int J Occup Environ Health 11: 107-122.
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.
Bogo H; Gomez DR; Reich SL; Negri RM; San Roman E. (2001). Traffic pollution in a downtown site of Buenos
Aires City. Atmos Environ 35: 1717-1727.
Bol KA; Collins JS; Kirby RS. (2006). Survival of infants with neural tube defects in the presence of folic acid
fortification. Pediatrics 117: 803-813.
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.
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.
Botter DA; Jorgensen B; Peres AA. (2002). A longitudinal study of mortality and air pollution for Sao Paulo, Brazil.
J Expo Anal Environ Epidemiol 12: 335-343.
Botto LD; Correa A; Erickson JD. (2001). Racial and temporal variations in the prevalence of heart defects.
Pediatrics 107: E32.
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 37.
Braga AL; Saldiva PH; Pereira LA; Menezes JJ; Conceicao GM; 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.
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.
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. Occupational and
Environmental Medicine 56: 237-244.
Brian JE, Jr.; Heistad DD; Faraci FM. (1994). Effect of carbon monoxide on rabbit cerebral arteries. Stroke 25: 639-
643; discussion 643-634.
Brown SD; Piantadosi CA. (1992). Recovery of energy metabolism in rat brain after carbon monoxide hypoxia. J
Clin Invest 89: 666-672.
Bruce EN; Bruce MC. (2003). A multicompartment model of carboxyhemoglobin and carboxymyoglobin responses
to inhalation of carbon monoxide. J Appl Physiol 95: 1235-1247.
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.
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.
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.
Brune B; Ullrich V. (1987). Inhibition of platelet aggregation by carbon monoxide is mediated by activation of
guanylate cyclase. Mol Pharmacol 32: 497-504.
March 2009
R-4
DRAFT-DO NOT QUOTE OR CITE
-------
Buck S; Dodds W; Fisher J; Hart D; Parker A; Stevenson J; Watson V; Welch E. (2000). Nutrient criteria technical
guidance manual, rivers and streams: US Environmental Protection Agency, Office of Water. EPA-822-B-
00-002,
Burke JM; Zufall MJ; Ozkaynak H. (2001). A population exposure model for particulate matter: case study results
for PM (2.5) in Philadelphia, PA. J Expo Anal Environ Epidemiol 11: 470-489.
Burmester T; Weich B; Reinhardt S; Hankeln T. (2000). A vertebrate globin expressed in the brain. Nature 407:
520-523.
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air pollution
on cardiorespiratory hospitalizations. Archives of Environmental Health 54: 130-139.
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.
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. American Journal of Epidemiology 153: 444-452.
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities.
Revised Analyses of Time-Series Studies of Air Pollution and Health Special Report Boston: Health
Effects Institute: 85-90.
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. Archives of Environmental Health
59: 228-236.
Bye A; Sorhaug S; Ceci M; Hoydal MA; Stolen T; Heinrich G; Tjonna AE; Najjar SM; Nilsen OG; Catalucci D;
Grimaldi S; Contu R; Steinshamn S; Condorelli G; Smith GL; Ellingsen O; WaldumH; Wisloff U. (2008).
Carbon monoxide levels experienced by heavy smokers impair aerobic capacity and cardiac contractility
and induce pathological hypertrophy. Inhalation Toxicology 20: 635-646.
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.
Cakmak S; Dales RE; Judek S. (2006a). Respiratory Health Effects of Air Pollution Gases: Modification by
Education and Income. Archives of Environmental Health 61: 5-10.
Cakmak S; Dales RE; Judek S. (2006b). Do gender, education, and income modify the effect of air pollution gases
on cardiac disease? J Occup Environ Med 48: 89-94.
Cakmak S; Dales RE; Vidal CB. (2007). Air pollution and mortality in Chile: susceptibility among the elderly.
Environ Health Perspect 115: 524-527.
Cardell LO; 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.
Carmines EL; Rajendran N. (2008). Evidence for carbon monoxide as the major factor contributing to lower fetal
weights in rats exposed to cigarette smoke. Toxicol Sci 102: 383-391.
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.
Carratu MR; Cagiano R; De Salvia MA; Trabace L; Cuomo V. (1995). Developmental neurotoxicity of carbon
monoxide. Arch Toxicol Suppl 17: 295-301.
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.
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.
Carraway MS; Ghio AJ; Carter JD; Piantadosi CA. (2000). Expression of heme oxygenase-1 in the lung in chronic
hypoxia. American journal of physiology Lung cellular and molecular physiology 278: L806.
Carraway MS; Ghio AJ; Suliman HB; Carter JD; Whorton AR; Piantadosi CA. (2002). Carbon monoxide promotes
hypoxic pulmonary vascular remodeling. American Journal of Physiology- Lung Cellular and Molecular
Physiology 282: 693-702.
Carter WPL. (1998). Updated maximum incremental reactivity scale for regulatory applications. Preliminary Report
to California Air Resources BoardContract 95308.
March 2009
R-5
DRAFT-DO NOT QUOTE OR CITE
-------
Castillo A; Llapur CJ; Martinez T; Kisling J; Williams-Nkomo T; Coates C; TepperRS. (2006). Measurement of
single breath-hold carbon monoxide diffusing capacity in healthy infants and toddlers. Pediatr Pulmonol
41: 544-550.
Cavallin-Stahl E; Mercke C; LundhB. (1976a). Erythropoiesis and Carbon Monoxide Production inHodgkin's
Disease. Br J Haematol 32: 167-176.
Cavallin-Stahl E; Mercke C; Lundh B. (1976b). Carbon monoxide production in patients with breast carcinoma. Br J
Haematol 32: 177-182.
CDC. (2004). The health consequences of smoking: a report of the Surgeon General. U.S. Department of Health and
Human Services, National Center for Chronic Disease Prevention and Health Promotion, Office on
Smoking and Health. Available from:
http://www.cdc.gov/tobacco/data statistics/sar/sar 2004/chapters.htm. Atlanta, GA. Available from:
http://www.cdc.gov/tobacco/sgr/sgr 2004/index.htm.
Cella M; Farina MG; Keller Sarmiento MI; 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.
Centers for Disease Control and Prevention. (1998). Trends in infant mortality attributable to birth defects-United
States, 1980-1995. MMWR Morb Mortal Wkly Rep 47: 773-778.
Chakraborty S; Balakotaiah V; Bidani A. (2004). Diffusing capacity reexamined: relative roles of diffusion and
chemical reaction in red cell uptake of 02, CO, C02, and NO. J Appl Physiol 97: 2284-2302.
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.
Chaloulakou A; Mavroidis I; Duci A. (2003). Indoor and outdoor carbon monoxide concentration relationships at
different microenvironments in the Athens area. Chemosphere 52: 1007.
Chan CC; Chuang KJ; Su TC; Lin LY. (2005). Association between nitrogen dioxide and heart rate variability in a
susceptible population. Eur J Cardiovasc Prev Rehabil 12: 580-586.
Chan CC; Chuang KJ; Chien LC; Chen WJ; Chang WT. (2006). Urban air pollution and emergency admissions for
cerebrovascular diseases in Taipei, Taiwan. Eur Heart J 27: 1238-1244.
Chang CC; Tsai SS; Ho SC; Yang CY. (2005). Air pollution and hospital admissions for cardiovascular disease in
Taipei, Taiwan. Environ Res 98: 114-119.
Chang CH; Meroney RN. (2003). Concentration and flow distributions in urban street canyons: wind tunnel and
computational data. Journal of Wind Engineering & Industrial Aerodynamics 91: 1141-1154.
Chang LT; Koutrakis P; Catalano PJ; Suh HH. (2000). Hourly personal exposures to fine particles and gaseous
pollutants - Results from Baltimore, Maryland. J Air Waste Manage Assoc 50: 1223-1235.
Chang LT; 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 0-3, PM2.5, and CO. J Air Waste Manage
Assoc 51: 1414-1422.
Chen GD; Kong J; Reinhard K; Fechter LD. (2001). NMDA receptor blockage protects against permanent noise-
induced hearing loss but not its potentiation by carbon monoxide. Hearing Res 154: 108-115.
Chen H; Namdeo A; Bell M. (2008). Classification of road traffic and roadside pollution concentrations for
assessment of personal exposure. Environmental Modelling & Software 23: 282-287.
Chen L; Jennison BL; Yang W; Omaye ST. (2000). Elementary school absenteeism and air pollution. Inhalation
Toxicology 12: 997-1016.
Chen L; Yang W; Jennison BL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada,
1991-1999. Inhalation Toxicology 14: 141-157.
Chen PC; Lai YM; Chan CC; Hwang JS; Yang CY; Wang JD. (1999). Short-term effect of ozone on the pulmonary
function of children in primary school. Environ Health Perspect 107: 921-925.
Cheng MF; Tsai SS; Wu TN; Chen PS; Yang CY. (2007). Air pollution and hospital admissions for pneumonia in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A 70: 2021-2026.
Chevalier RB; Krumholz RA; Ross JC. (1966). Reaction of nonsmokers to carbon monoxide inhalation.
Cardiopulmonary responses at rest and during exercise. Journal of the American Medical Association 198:
1061-1064.
Chin BY; Jiang G; Wegiel B; Wang HJ; MacDonald T; Zhang XC; Gallo D; Cszimadia E; Bach FH; Lee PJ. (2007).
Hypoxia-inducible factor lalpha stabilization by carbon monoxide results in cytoprotective
preconditioning. Proceedings of the National Academy of Sciences 104: 5109.
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.
March 2009
R-6
DRAFT-DO NOT QUOTE OR CITE
-------
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 Manage Assoc 50: 1481-1500.
Chung Y; Huang SJ; Glabe A; Jue T. (2006). Implication of CO inactivation on myoglobin function. American
Journal of Physiology- Lung Cellular and Molecular Physiology 290: C1616-1624.
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 Manage Assoc 50: 1287-1298.
Clewell HJ; Crump KS. (2005). Quantitative Estimates of Risk for Noncancer Endpoints. An International Journal
25: 285-289.
Coburn RF; Williams WJ; Forster RE. (1964). Effect of Erythrocyte Destruction on Carbon Monoxide Production in
Man. J Clin Invest 43: 1098-1103.
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.
Coburn RF; Williams WJ; Kahn SB. (1966). Endogenous carbon monoxide production in patients with hemolytic
anemia. J Clin Invest 45: 460-468.
Coburn RF. (1967). Endogenous carbon monoxide production and body CO stores. Acta Med Scand Suppl 472:
269-282.
Coburn RF. (1970a). The carbon monoxide body stores. Ann N Y Acad Sci 174: 11-22.
Coburn RF. (1970b). Enhancement by phenobarbital and diphenylhydantoin of carbon monoxide production in
normal man. N Engl J Med 283: 512-515.
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. Circul Res 62: 471-477.
Collier TK. (2003). Forensic Ecotoxicology: Establishing Causality between Contaminants and Biological Effects in
Field Studies. HumEcol Risk Assess 9: 259 - 266.
Conceicao GM; Miraglia SG; Kishi HS; Saldiva PH; Singer JM. (2001). Air pollution and child mortality: a time-
series study in Sao Paulo, Brazil. Environ Health Perspect 109 Suppl 3: 347-350.
Conny JM. (1998). The isotopic characterization of carbon monoxide in the troposphere. Atmos Environ 32: 2669-
2683.
Conrad R; Seiler W. (1985). Influence of temperature, moisture, and organic carbon on the flux of H 2 and CO
between soil and atmosphere: Field studies in subtropical regions. Journal of Geophysical Research 90:
5699-5709.
Cronje FJ; Carraway MS; Freiberger JJ; Suliman HB; Piantadosi CA. (2004). Carbon monoxide actuates 0(2)-
limited heme degradation in the rat brain. Free Radic Biol Med 37: 1802-1812.
Croxford B; Penn A. (1998). Siting considerations for urban pollution monitors. Atmos Environ 32: 1049-1057.
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.
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 115: 1789-1797.
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.
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.
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery
disease. J Occup Environ Med 46: 1217-1221.
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.
Davies DM; Smith DJ. (1980). Electrocardiographic changes in healthy men during continuous low-level carbon
monoxide exposure. Environ Res 21: 197-206.
Davies RF; Topping DL; Turner DM. (1976). The effect of intermittent carbon monoxide exposure on experimental
atherosclerosis in the rabbit. Atherosclerosis 24: 527-536.
de Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; van Wijnen 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. American Journal of
Epidemiology 157: 613-623.
March 2009
R-7
DRAFT-DO NOT QUOTE OR CITE
-------
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 196-187.
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality
associations with air pollution. American Journal of Respiratory and Critical Care Medicine 167: 1117-
1123.
De Salvia MA; Cagiano R; Carratu MR; Di Giovanni V; Trabace L; Cuomo V. (1995). Irreversible impairment of
active avoidance behavior in rats prenatally exposed to mild concentrations of carbon monoxide.
Psychopharmacology (Berl) 122: 66-71.
Deedwania PC; Stein PK; Koren A; Mukheijee 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 Pharmacotherapy (pp. 3029): McGraw-Hill.
Delfino RJ; Gong H, Jr.; 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.
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.
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 Human Reprod 10: 211-214.
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.
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 F0. Biochemistry 28: 2631-2638.
Dimitroulopoulou C; Ashmore MR; Byrne MA; Kinnersley RP. (2001). Modelling of indoor exposure to nitrogen
dioxide in the UK. Atmos Environ 35: 269-279.
Dimitroulopoulou C; Ashmore MR; Hill MTR; Byrne MA; Kinnersley R. (2006). INDAIR: A probabilistic model
of indoor air pollution in UK homes. Atmos Environ 40: 6362-6379.
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.
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. American Journal of Epidemiology 156: 193-203.
Dominici F; Sheppard L; Clyde M. (2003a). Health Effects of Air Pollution: A Statistical Review. International
Statistical Review 71: 243-276.
Dominici F; McDermott A; Daniels M; Zeger SL; Samet JM. (2003b). Mortality among residents of 90 cities.
Revised Analyses of Time-Series Studies of Air Pollution and Health Special Report Boston, MA: Health
Effects Institute: 9-24.
Dominici F; Zanobetti A; Zeger SL; Schwartz J; Samet JM. (2005a). The National Morbidity, Mortality, and Air
Pollution Study Part IV: Hierarchical Bivariate Time-Series Models-A Combined Analysis of PM~ 1~ 0
Effects on Hospitalization and Mortality. Health Effects Institute,
Dominici F; McDermott A; Daniels M; Zeger SL; Samet JM. (2005b). Revised analyses of the National Morbidity,
Mortality, and Air Pollution Study: mortality among residents of 90 cities. J Toxicol Environ Health A 68:
1071-1092.
Douglas CG; Haldane JS; Haldane JBS. (1912). The laws of combination of haemoglobin with carbon monoxide
and oxygen. The Journal of Physiology 44: 275.
Drinkwater BL; Raven PB; Horvath SM; Gliner JA; Ruhling RO; Bolduan NW; Taguchi S. (1974). Air pollution,
exercise, and heat stress. Archives of Environmental Health 28: 177-181.
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. American Journal
of Physiology-Lung Cellular and Molecular Physiology 283: L120-129.
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.
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.
March 2009
R-8
DRAFT-DO NOT QUOTE OR CITE
-------
Duci A; Chaloulakou A; Spyrellis N. (2003). Exposure to carbon monoxide in the Athens urban area during
commuting. Sci Total Environ 309: 47-58.
Dutton SJ; Hannigan MP; Miller SL. (2001). Indoor pollutant levels from the use of unvented natural gas fireplaces
in Boulder, Colorado. J Air Waste Manage Assoc 51: 1654-1661.
Ekblom B; Huot R. (1972). Response to submaximal and maximal exercise at different levels of
carboxyhemoglobin. Acta Physiol Scand 86: 474-482.
Engel LA. (1983). Gas mixing within the acinus of the lung. J Appl Physiol 54: 609-618.
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.
Enzi G; BevilacquaM; Crepaldi G. (1976). Disturbances in pulmonary gaseous exchange in primary
hyperlipoproteinemias. Bull Eur Physiopathol Respir 12: 433-442.
Ernst A; Zibrak JD. (1998). Carbon Monoxide Poisoning. New Engl J Med 339: 1603-1608.
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:
607-611.
Fairley D. (1999). Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ Health
Perspect 107: 637-641.
Farhat SC; Paulo RL; Shimoda TM; Conceicao GM; Lin CA; Braga AL; Warth MP; Saldiva PH. (2005). Effect of
air pollution on pediatric respiratory emergency room visits and hospital admissions. Braz J Med Biol Res
38: 227-235.
Farrugia G; Irons WA; Rae JL; SarrMG; Szurszewski JH. (1993). Activation of whole cell currents in isolated
human jejunal circular smooth muscle cells by carbon monoxide. Am J Physiol 264: G1184-1189.
Favory R; Lancel S; Marechal X; Tissier S; Neviere R. (2006a). Cardiovascular protective role for activated protein
C during endotoxemia in rats. Intensive Care Med 32: 899-905.
Favory R; Lancel S; Tissier S; Mathieu D; Decoster B; Neviere R. (2006b). Myocardial dysfunction and potential
cardiac hypoxia in rats induced by carbon monoxide inhalation. American Journal of Respiratory and
Critical Care Medicine 174: 320-325.
Fechter LD; Annau Z. (1977). Toxicity of mild prenatal carbon monoxide exposure. Science 197: 680-682.
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.
Fechter LD; Annau Z. (1980). Prenatal carbon monoxide exposure alters behavioral development. Neurobehav
Toxicol 2: 7-11.
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.
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.
Fechter LD; Chen G; Rao D. (2002). Chemical asphyxiants and noise. Noise Health 4: 49.
Fenn WO. (1970). The burning of CO in tissues. Ann N Y Acad Sci 174: 64-71.
Finlayson-Pitts BJ; Pitts Jr JN. (2000). Chemistry of the Upper and Lower Atmosphere Theory, Experiments, and
Applications. San Diego: Academic Press.
Fischer P; Hoek G; Brunekreef B; Verhoeff A; van Wijnen J. (2003). Air pollution and mortality in the Netherlands:
are the elderly more at risk? European Respiratory Journal 21: 34S-38S.
Fischer PH; Steerenberg PA; Snelder JD; van Loveren H; van Amsterdam JG. (2002). Association between exhaled
nitric oxide, ambient air pollution and respiratory health in school children. Int Arch Occup Environ Health
75: 348-353.
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.
Fishman J; Bowman KW; Burrows JP; Richter A; Chance KV; Edwards DP; Martin RV; Morris GA; Pierce RB;
Ziemke JR. (2008). Remote Sensing of Tropospheric Pollution from Space. Bulletin of the American
Meteorological Society 89: 805-821.
Flachsbart PG. Models of exposure to carbon monoxide inside a vehicle on a Honolulu highway. Journal of
Exposure Analysis and Environmental Epidemiology 9: 245.
Fodor GG; Winneke G. (1972). Effect of low CO concentrations on resistance to monotony and on psychomotor
capacity. Staub Reinhaltung Der Luft 32: 46-54.
March 2009
R-9
DRAFT-DO NOT QUOTE OR CITE
-------
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. American Journal of Respiratory and Critical Care Medicine 172: 1549-
1555.
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.
ForsterRE. (1966). Diffusion of gases in the lungs. Physiologist 9: 110-122.
Fox GA. (1991). Practical causal inference for ecoepidemiologists. J Toxicol Environ Health 33: 359-373.
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.
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhalation Toxicology 18: 1005-1011.
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.
Garofolo MC; Seidler FJ; Auman JT; Slotkin TA. (2002). B-Adrenergic modulation of muscarinic cholinergic
receptor expression and function in developing heart. American Journal of Physiology- Regulatory,
Integrative and Comparative Physiology 282: 1356-1363.
Gautier M; Antier D; Bonnet P; Net JLL; 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.
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.
Gee GC; Payne-Sturges DC. (2004). Environmental health disparities: a framework integrating psychosocial and
environmental concepts. Environ Health Perspect 112: 1645-1653.
Georgoulis LB; Hanninen O; Samoli E; Katsouyanni K; Kunzli N; Polanska L; Bruinen de Bruin Y; Aim S;
Jantunen M. (2002). Personal carbon monoxide exposure in five European cities and its determinants.
Atmos Environ 36: 963-974.
Gerritsen J; Carlson RE; Dycus DL; Faulkner C; Gibson GR. (1998). Lake and reservoir bioassessment and
biocriteria: Technical guidance document. EPA 841-B-98-007. US Environmental Protection Agency,
Office of Water, Washington, DC. http://www.epa.gov/owow/monitoring/tech/lakes.html.
Ghio 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.
Gibson QH; Roughton FJ. (1955). The kinetics of dissociation of the first oxygen molecule from fully saturated
oxyhaemoglobin in sheep blood solutions. Proc R Soc Lond B Biol Sci 143: 310-334.
Gibson QH; Olson JS; McKinnie RE; Rohlfs RJ. (1986). A kinetic description of ligand binding to sperm whale
myoglobin. J Biol Chem 261: 10228-10239.
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. American
Journal of Epidemiology 162: 238-252.
Gilliam RC; Huber AH; Raman S; Neogi D. (2005). Metropolitan-Scale Transport and Dispersion of Materiel
Released from the New York World Trade Center Following September 11, 2001—Part II: The CALPUFF
Plume Model. Pure and Applied Geophysics 162: 2005-2028.
Gilliland F; Avol PK; Kinney PL; Jerrett M; Dvonch T; Lurmann F; Buckley T; Breysse P; Keeler G; de Villiers T;
McConnellR. (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: 1447-1454.
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. Pharmacol Toxicol 73: 274-278.
Giustino A; Carratu MR; Siro BG; 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.
March 2009
R-10
DRAFT-DO NOT QUOTE OR CITE
-------
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.
Glabe A; Chung Y; Xu D; Jue T. (1998). Carbon monoxide inhibition of regulatory pathways in myocardium. Am J
Physiol 274: H2143-2151.
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. Circulation 101: 1267-1273.
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.
Goldberg MS; Burnett RT; Bailar JC, 3rd; Tamblyn R; Ernst P; Flegel K; Brook J; Bonvalot Y; Singh R; Valois
MF; Vincent R. (2001). Identification of persons with cardiorespiratory conditions who are at risk of dying
from the acute effects of ambient air particles. Environ Health Perspect 109 Suppl 4: 487-494.
Goldberg MS; Burnett RT; Valois MF; Flegel K; Bailar JC, 3rd; 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.
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. Environmental Research 100: 255.
Goldstein AH; Galbally IE. (2007). Known and unexplored organic constituents in the earth's atmosphere.
Environmental Science and Technology 41: 1514-1521.
Gomez-Perales JE; Colvile RN; Nieuwenhuijsen MJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-
Figueroa VH; Bianco-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.
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez-Avedoy V; Paramo-Figueroa VH; Bianco-
Jimenez S; Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo-Navarro M. (2007). Bus,
minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City. Atmos Environ
41: 890-901.
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. American Journal of Respiratory and Critical
Care Medicine 169: 816-821.
Gouveia N; Fletcher T. (2000a). Respiratory diseases in children and outdoor air pollution in Sao Paulo, Brazil: a
time series analysis. Occupational and Environmental Medicine 57: 477-483.
Gouveia N; Fletcher T. (2000b). Time series analysis of air pollution and mortality: effects by cause, age and
socioeconomic status. J Epidemiol Community Health 54: 750-755.
Gouveia N; Bremner SA; Novaes HM. (2004). Association between ambient air pollution and birth weight in Sao
Paulo, Brazil. J Epidemiol Community Health 58: 11-17.
Grosjean D; Fung K. (1984). Hydrocarbons and carbonyl in Los Angeles air. J Air Pollut Control Assoc 34.
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. American Journal of Physiology- Lung Cellular
and Molecular Physiology 278: L779-784.
Gryparis A; Forsberg B; Katsouyanni K; Analitis A; Touloumi G; Schwartz J; Samoli E; Medina S; Anderson HR;
Niciu EM. (2004). Acute Effects of Ozone on Mortality from the" Air Pollution and Health: A European
Approach" Project. American Journal of Respiratory and Critical Care Medicine 170: 1080.
Gulliver J; Briggs DJ. (2005). Time-space modeling of journey-time exposure to traffic-related air pollution using
GIS. Environ Res 97: 10-25.
Guo YL; Lin YC; Sung FC; Huang SL; Ko YC; Lai JS; Su HJ; Shaw CK; Lin RS; Dockery DW. (1999). Climate,
traffic-related air pollutants, and asthma prevalence in middle-school children in taiwan. Environ Health
Perspect 107: 1001-1006.
Guyatt AR; Holmes MA; Cumming G. (1981). Can carbon monoxide be absorbed from the upper respiratory tract in
man? Eur J Respir Dis 62: 383-390.
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.
Ha EH; Hong YC; Lee BE; Woo BH; Schwartz J; Christiani DC. (2001). Is air pollution a risk factor for low birth
weight in Seoul? Epidemiology 12: 643-648.
March 2009
R-11
DRAFT-DO NOT QUOTE OR CITE
-------
Ha EH; Lee JT; Kim H; Hong YC; Lee BE; Park HS; Christiani DC. (2003). Infant susceptibility of mortality to air
pollution in Seoul, South Korea. Pediatrics 111: 284-290.
Haab P. (1990). The effect of carbon monoxide on respiration. Experientia 46: 1202-1206.
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.
Hajat S; Anderson HR; Atkinson RW; Haines A. (2002). Effects of air pollution on general practitioner
consultations for upper respiratory diseases in London. Occupational and Environmental Medicine 59: 294-
299.
Haldane J. (1898). Some improved methods of gas analysis. The Journal of Physiology 22: 465.
Hampson NB. (2007). Carboxyhemoglobin elevation due to hemolytic anemia. J Emerg Med 33: 17-19.
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.
Hanel B; Teunissen I; Rabol A; Warberg J; Secher NH. (1997). Restricted postexercise pulmonary diffusion
capacity and central blood volume depletion (Vol. 83, pp. 11-17): Am Physiological Soc.
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 and
Built Environment 15: 147-153.
Hara 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.
Harada T; Koi H; Kubota T; Aso T. (2004). Haem oxygenase augments porcine granulosa cell apoptosis in vitro. J
Endocrinol 181: 191-205.
Harley RA; McKeen SA; Pearson J; Rodgers MO; Lonneman WA. (2001). Analysis of motor vehicle emissions
during the Nashville/Middle Tennessee Ozone Study. Journal of Geophysical Research 106: 3559-3567.
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. Environmental Science and Technology 39: 5356-
5362.
Hausberg M; Somers VK. (1997). Neural circulatory responses to carbon monoxide in healthy humans.
Hypertension 29: 1114-1118.
HegewaldMJ; Crapo RO. (2007). Socioeconomic status and lung function. Chest 132: 1608-1614.
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 Perinatal 24: 275-278.
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.
Occupational and Environmental Medicine 64: 439-445.
Hering SV. (2007). Using Regional Data and Building Leakage to Assess Indoor Concentrations of Particles of
Outdoor Origin. Aerosol Science and Technology 41: 639-654.
Hermans RH; McGivern RF; Chen W; Longo LD. (1993). Altered adult sexual behavior in the male rat following
chronic prenatal hypoxia. Neurotoxicol Teratol 15: 353-363.
Hill AB. (1965). The Environment and Disease: Association or Causation? Proc R Soc Med 58: 295-300.
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: H311-323.
Hinwood AL; De Klerk N; Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett
JCEH. (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.
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.
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.
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.
Hoek G. (2003). Daily mortality and air pollution in The Netherlands. In Revised analyses of time-series studies of
air pollution and health. Health Effects Institute, Boston, MA.
Hoel DG. (1980). Incorporation of background in dose-response models. Fed Proc 39: 73-75.
March 2009
R-12
DRAFT-DO NOT QUOTE OR CITE
-------
Hofmann OM; Brittain T. (1998). Partitioning of oxygen and carbon monoxide in the three human embryonic
hemoglobins. Hemoglobin 22: 313-319.
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.
Holland RA. (1970). Reaction rates of carbon monoxide and hemoglobin. Ann N Y Acad Sci 174: 154-171.
Hong YC; Leem JH; Ha EH. (1999a). Air pollution and daily mortality in Inchon, Korea. J Korean Med Sci 14: 239-
244.
Hong YC; Leem JH; Ha EH; Christiani DC. (1999b). PM(10) exposure, gaseous pollutants, and daily mortality in
Inchon, South Korea. Environ Health Perspect 107: 873-878.
Hong YC; Lee JT; Kim H; Ha EH; Schwartz J; Christiani DC. (2002a). Effects of air pollutants on acute stroke
mortality. Environ Health Perspect 110: 187-191.
Hong YC; Lee JT; Kim H; Kwon HJ. (2002b). Air pollution: a new risk factor in ischemic stroke mortality. Stroke
33: 2165-2169.
Horvath SM; Dahms TE; O'Hanlon JF. (1971). Carbon monoxide and human vigilance. A deleterious effect of
present urban concentrations. Archives of Environmental Health 23: 343-347.
Horvath SM; Raven PB; Dahms TE; Gray DJ. (1975). Maximal aerobic capacity at different levels of
carboxyhemoglobin. J Appl Physiol 38: 300-303.
Horvath SM; Bedi JF; Wagner JA; Agnew J. (1988). Maximal aerobic capacity at several ambient concentrations of
CO at several altitudes. J Appl Physiol 65: 2696-2708.
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.
Hsia CC. (2002). Recruitment of lung diffusing capacity: update of concept and application. Chest 122: 1774-1783.
Huikuri HV; Jokinen V; Syvanne M; Nieminen MS; Airaksinen KE; Ikaheimo MJ; Koistinen JM; Kauma H;
Kesaniemi AY; Majahalme S; Niemela KO; Frick MH. (1999). Heart rate variability and progression of
coronary atherosclerosis. Atertio Thromb Vase Biol 19: 1979-1985.
Hutcheon JA; Piatt RW. (2008). The missing data problem in birth weight percentiles and thresholds for "Small-for-
Gestational-Age". American Journal of Epidemiology 167: 786-792.
Huynh M; Woodruff TJ; Parker JD; Schoendorf KC. (2006). Relationships between air pollution and preterm birth
in California. Paediatr Perinat Epidemiol 20: 454-461.
Hwang B; Jaakkola JK. (2008). Ozone and Other Air Pollutants and the Risk of Oral Clefts Environ Health Perspect
116:1411-1415
Hwang BF; Lee YL; Lin YC; Jaakkola JJ; Guo YL. (2005). Traffic related air pollution as a determinant of asthma
among Taiwanese school children. Thorax 60: 467-473.
Hwang BF; Jaakkola JJ; Lee YL; Lin YC; Guo YL. (2006). Relation between air pollution and allergic rhinitis in
Taiwanese schoolchildren. RespirRes 7: 23.
Hwang JS; Chan CC. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness.
American Journal of Epidemiology 155: 1-10.
IARC. (2006). IARC monographs on the evaluation of carcinogenic risks to humans: Preamble. . International
Agency for Research on Cancer (IARC). Lyon, France.
http ://mono graphs, iarc .fr/ENG/Preamble/CurrentPreamble .pdf
Ibald-Mulli A; Stieber J; WichmannHE; Koenig W; Peters A. (2001). Effects of air pollution on blood pressure: a
population-based approach. Am J Public Health 91: 571-577.
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.
Iheagwara KN; Thom SR; Deutschman CS; Levy RJ. (2007). Myocardial cytochrome oxidase activity is decreased
following carbon monoxide exposure. Biochim Biophys Acta 1772: 1112-1116.
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. Circul Res 89: 55-62.
Inman RE; Ingersoll RB; Levy EA. (1971). Soil: A Natural Sink for Carbon Monoxide. Science 172: 1229-1231.
IOM (2008). Improving the Presumptive Disability Decision-Making Process for Veterans. In Samet James M,
Bodurow CC (Eds.),
IPCC. (2001). Climate Change 2001: The Scientific Basis. Cambridge Univ. Press, Cambridge, UK.
IPCC. (2007). Climate Change 2007: The Physical Science Basis. Cambridge Univ. Press, Cambridge, UK.
March 2009
R-13
DRAFT-DO NOT QUOTE OR CITE
-------
Ischiropoulos H; Beers MF; Ohnishi ST; Fisher D; Garner SE; Thom SR. (1996). Nitric oxide production and
perivascular nitration in brain after carbon monoxide poisoning in the rat. J Clin Invest 97: 2260-2267.
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:
S45-60.
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.
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.
Jeffries HE. (1995). Photochemical air pollution. In Singh HB (Ed.), Composition, chemistry, and climate of the
atmosphere (pp. 308-348). New York, NY: Van Nostrand Reinhold.
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of the
air pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health A 66:
1735-1777.
Jerrett M; Arain A; Kanaroglou P; Beckerman B; Potoglou D; Sahsuvaroglu T; Morrison J; Giovis C. (2004). A
review and evaluation of intraurban air pollution exposure models. Journal of Exposure Science and
Environmental Epidemiology 15: 185-204.
Johnson FK; Johnson RA. (2003). Carbon monoxide promotes endothelium-dependent constriction of isolated
gracilis muscle arterioles. American Journal of Physiology- Regulatory, Integrative and Comparative
Physiology 285: 536-541.
Johnson FK; Durante W; Peyton KJ; Johnson RA. (2003). Heme oxygenase inhibitor restores arteriolar nitric oxide
function in Dahl rats. Hypertension 41: 149-155.
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.
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. American Journal of Physiology- Regulatory, Integrative and Comparative Physiology
290: 601-608.
Johnson T. (1984). A Study of Personal Exposure to Carbon Monoxide in Denver, Colorado. Research Triangle
Park, NC: US Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report
no. EPA-600/4-84-014,
Joumard R; Chiron M; Vidon R; Maurin M; Rouzioux JM. (1981). Mathematical models of the uptake of carbon
monoxide on hemoglobin at low carbon monoxide levels. Environ Health Perspect 41: 277-289.
Kanakidou M; Crutzen PJ. (1999). The photochemical source of carbon monoxide: Importance, uncertainties and
feedbacks. Chemosphere-Global Change Science 1: 91-109.
Kanten WE; Penney DG; Francisco K; Thill JE. (1983). Hemodynamic responses to acute carboxyhemoglobinemia
in the rat. Am J Physiol 244: H320-327.
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.
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. American Journal of Epidemiology 165: 553-
560.
Kastner-Klein P; Plate EJ. (1999). Wind-tunnel study of concentration fields in street canyons. Atmos Environ 33:
3973-3979.
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: 220-227.
Katoue MG; Khan I; Oriowo MA. (2006). Pregnancy-induced modulation of calcium mobilization and down-
regulation of Rho-kinase expression contribute to attenuated vasopressin-induced contraction of the rat
aorta. Vascul Pharmacol 44: 170-176.
Katsouyanni K; Touloumi G; Samoli E; Gryparis A; Le Tertre A; Monopolis Y; Rossi G; Zmirou D; Ballester F;
Boumghar A. (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.
March 2009
R-14
DRAFT-DO NOT QUOTE OR CITE
-------
Kaur S; NieuwenhuijsenM; Colvile R. (2005). Personal exposure of street canyon intersection users to PM2.5,
uhrafine particle counts and carbon monoxide in Central. London, UK. Atmos Environ 39: 3629-3641.
Kaur S; NieuwenhuijsenMJ; Colvile RN. (2007). Fine particulate matter and carbon monoxide exposure
concentrations in urban street transport microenvironments. Atmos Environ 41: 4781-4810.
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.
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.
Khalil MAK; Rasmussen RA. (1990). The global cycle of carbon monoxide: trends and mass balance. Chemosphere
20: 227-242.
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.
Proceedings of the National Academy of Sciences 103: 17944-17948.
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. Journal of Exposure Science and Environmental Epidemiology 16: 172-183.
Kim HP; Wang X; Nakao A; Kim SI; Murase N; Choi ME; Ryter SW; Choi AM. (2005). Caveolin-1 expression by
means of p38beta mitogen-activated protein kinase mediates the antiproliferative effect of carbon
monoxide. Proceedings of the National Academy of Sciences 102: 11319-11324.
Kim HS; Loughran PA; Rao J; Billiar TR; Zuckerbraun BS. (2008). Carbon monoxide activates NF-kappaB via
ROS generation and Akt pathways to protect against cell death of hepatocytes. American Journal of
Physiology Gastrointestinal and Liver Physiology 295: G146-G152.
Kim SY; O'Neill MS; Lee JT; Cho Y; Kim J; KimH. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health 80: 701-710.
King GM. (1999). Characteristics and significance of atmospheric carbon monoxide consumption by soils.
Chemosphere-Global Change Science 1: 53-63.
Kingdom JCP; Kaufmann P. (1997). Oxygen and placental villous development: Origins of fetal hypoxia. Placenta
18: 613-621.
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.
Kinobe RT; Vlahakis JZ; Soong JM; Szarek WA; Brien JF; Longo LD; 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.
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.
Klasner AE; Smith SR; Thompson MW; Scalzo AJ. (1998). Carbon monoxide mass exposure in a pediatric
population. Acad Emerg Med 5: 992-996.
KlausenK; Rasmussen B; GjellerodH; MadsenH; Petersen E. (1968). Circulation, metabolism and ventilation
during prolonged exposure to carbon monoxide and to high altitude. Scand J Clin Lab Invest Suppl 103:
26-38.
KleinmanMT; Davidson DM; Vandagriff RB; Caiozzo VJ; Whittenberger JL. (1989). Effects of short-term
exposure to carbon monoxide in subjects with coronary artery disease. Archives of Environmental Health
44: 361-369.
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. Archives of Environmental
Health 53: 388-397.
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data
from ARIES. Inhalation Toxicology 16 Suppl 1: 131-141.
Klepeis NE; Nelson WC; Ott WR; Robinson JP; Tsang 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 Anal Environ Epidemiol 11:231-252.
KlimischHJ; Chevalier HJ; Harke HP; Dontenwill W. (1975). Uptake of carbon monoxide in blood of miniture pigs
and other mammals. Toxicology 3: 301-310.
March 2009
R-15
DRAFT-DO NOT QUOTE OR CITE
-------
Knuckles TL; Lund AK; Lucas SN; Campen MJ. (2008). Diesel exhaust exposure enhances venoconstriction via
uncoupling of eNOS. Toxicol Appl Pharmacol 230: 346-351.
Koehler RC; Jones MD; Traystman RJ. (1982). Cerebral circulatory response to carbon monoxide and hypoxic
hypoxia in the lamb. American Journal of Physiology- Heart and Circulatory Physiology 243: 27-32.
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 85: 169-183.
Koken PJ; 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.
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.
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.
Kotsch K; Francuski M; Pascher A; Klemz R; Seifert M; Mittler J; Schumacher G; Buelow R; Volk HD; Tullius SG.
(2006). Improved Long-Term Graft Survival after HO-1 Induction in Brain-Dead Donors. Am J
Transplantation 6: 477.
Kreider JC; Blumberg MS. (2005). Geotaxis and beyond: commentary on Motz and Alberts (2005). Neurotoxicol
Teratol 27: 529-533.
KreiserD; BaumM; SeidmanDS; Fanaroff A; ShahD; Hendlerl; Stevenson DK; Schiff E; DruzinML. (2004). End
tidal carbon monoxide levels are lower in women with gestational hypertension and pre-eclampsia. J
Perinatal 24:213-217.
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerret M. (2000). Reanalysis of the Harvard Six
Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality. . Health
Effects, Institute, Boston, MA.
Kruize H; Hanninen O; Breugelmans O; Lebret E; Jantunen M. (2003). Description and demonstration of the
EXPOLIS simulation model: two examples of modeling population exposure to particulate matter. J
ExpoAnalEnvironEpidemiol 13: 87-99.
Kuenzli 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.
Kwon HJ; Cho SH; 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.
Lagorio S; Forastiere F; Pistelli R; Iavarone 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.
Lai HK; Kendall M; Ferrier H; Lindup I; Aim S; Hanninen O; Jantunen M; Mathys P; Colvile R; Ashmore MR;
CullinanP; NieuwenhuijsenMJ. (2004). Personal exposures and microenvironment concentrations of
PM2.5, VOC, N02 and CO in Oxford, UK. Atmos Environ 38: 6399-6410.
Lamberto C; Nunes H; Le ToumelinP; DuperronF; 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.
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.
Landaw SA. (1973). The effects of cigarette smoking on total body burden and excretion rates of carbon monoxide.
JOccupMed 15: 231-235.
LanganL. (1992). Portability in measuing exposure to carbon monoxide. Journal of Exposure Analysis and
Environmental Epidemiology Suppl 1: 223-239.
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. (2006). Associations of traffic related air
pollutants with hospitalisation for first acute myocardial infarction: the HEAPSS study. Occupational and
Environmental Medicine 63: 844-851.
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: H160-167.
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003a). Exposure to air pollution during different
gestational phases contributes to risks of low birth weight. Hum Reprod 18: 638-643.
March 2009
R-16
DRAFT-DO NOT QUOTE OR CITE
-------
Lee IM; Tsai SS; Ho CK; Chiu HF; Yang C. (2007a). Air pollution and hospital admissions for congestive heart
failure in a tropical city: Kaohsiung, Taiwan. Inhalation Toxicology 19: 899-904.
Lee IM; Tsai SS; Chang CC; Ho CK; Yang CY. (2007b). Air pollution and hospital admissions for chronic
obstructive pulmonary disease in a tropical city: Kaohsiung, Taiwan. Inhalation Toxicology 19: 393-398.
Lee JT; Kim H; Song H; Hong YC; Cho YS; Shin SY; Hyun YJ; Kim YS. (2002). Air pollution and asthma among
children in Seoul, Korea. Epidemiology 13: 481-484.
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003b). Air pollution and hospital admissions for ischemic
heart diseases among individuals 64+ years of age residing in Seoul, Korea. Archives of Environmental
Health 58: 617-623.
Lee JT; Son JY; Kim H; Kim SY. (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.
Lee JT; Son JY; Cho YS. (2007c). 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. Environmental Research 105:
409-413.
Lee PJ; Jiang BH; Chin BY; Iyer NV; Alam J; Semenza GL; Choi AM. (1997). Hypoxia-inducible factor-1 mediates
transcriptional activation of the heme oxygenase-1 gene in response to hypoxia. J Biol Chem 272: 5375-
5381.
Lee YL; Shaw CK; Su HJ; Lai JS; Ko YC; Huang SL; Sung FC; Guo YL. (2003c). Climate, traffic-related air
pollutants and allergic rhinitis prevalence in middle-school children in Taiwan. Eur Respir J 21: 964-970.
Leem JH; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect 114: 905-910.
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.
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 Anal Environ Epidemiol 10: 58-65.
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.
Li CI; Daling JR; Emanuel I. (2003). Birthweight and risk of overall and cause-specific childhood mortality.
Paediatr Perinat Epidemiol 17: 164-170.
Li XY; Zhang C; Wang SF; Ji YL; Wang H; Zhao L; Xu DX. (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.
Liao D; Duan Y; Whitsel EA; Zheng ZJ; Heiss G; Chinchilli VM; Lin HM. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. American
Journal of Epidemiology 159: 768-777.
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria
pollutants with plasma hemostatic/inflammatory markers: a population-based study. J Expo Anal Environ
Epidemiol 15: 319-328.
Lin CA; Martins MA; Farhat SC; Pope CA, 3rd; Conceicao GM; Anastacio VM; Hatanaka M; Andrade WC;
Hamaue WR; Bohm GM; Saldiva PH. (1999). Air pollution and respiratory illness of children in Sao Paulo,
Brazil. Paediatr Perinat Epidemiol 13: 475-488.
Lin CA; Pereira LA; Nishioka DC; Conceicao GM; Braga AL; Saldiva PH. (2004a). Air pollution and neonatal
deaths in Sao Paulo, Brazil. Braz J Med Biol Res 37: 765-770.
Lin CM; Li CY; Yang GY; Mao IF. (2004b). Association between maternal exposure to elevated ambient sulfur
dioxide during pregnancy and term low birth weight. Environmental Research 96: 41-50.
LinH; McGrath JJ. (1988). Vasodilating effects of carbon monoxide. Drug Chem Toxicol 11: 371-385.
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.
Lin M; Chen Y; Villeneuve PJ; Burnett RT; Lemyre L; Hertzman C; McGrail KM; Krewski D. (2004c). Gaseous air
pollutants and asthma hospitalization of children with low household income in Vancouver, British
Columbia, Canada. American Journal of Epidemiology 159: 294-303.
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: e235-240.
March 2009
R-17
DRAFT-DO NOT QUOTE OR CITE
-------
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.
Lipfert F; Baty J; Miller J; Wyzga R. (2006a). PM 2.5 Constituents and Related Air Quality Variables As Predictors
of Survival in a Cohort of US Military Veterans. Inhalation Toxicology 18: 645-657.
Lipfert FW; Perry HM; Miller JP; Baty JD; Wyzga RE; Carmody SE. (2000a). The Washington University-EPRI
Veterans' Cohort Mortality Study. Preliminary Results. Inhalation Toxicology 12: 41-73.
Lipfert FW; Morris SC; Wyzga RE. (2000b). Daily mortality in the Philadelphia metropolitan area and size-
classified particulate matter. J Air Waste Manage Assoc 50: 1501-1513.
Lipfert FW; Morris SC. (2002). Temporal and spatial relations between age specific mortality and ambient air
quality in the United States: regression results for counties, 1960-97. Occupational and Environmental
Medicine 59: 156-174.
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006b). 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.
Lippmann M; Ito K; Nadas A; Burnett RT. (2000). Association of particulate matter components with daily
mortality and morbidity in urban populations. Res Rep Health Eff Inst: 5-72, discussion 73-82.
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.
Liu S; Krewski D; Shi Y; Chen Y; Burnett RT. (2007). Association between maternal exposure to ambient air
pollutants during pregnancy and fetal growth restriction. Journal of Exposure Science and Environmental
Epidemiology 17: 426-432.
Liu Y; Fechter LD. (1995). MK-801 protects against carbon monoxide-induced hearing loss. Toxicol Appl
Pharmacol 132: 196-202.
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 J (Ed.),
Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications (pp. 289-304). Cambridge,
Massachusetts: The MIT Press.
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.
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.
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: 666-675.
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.
Luginaah IN; Fung KY; 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.
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.
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. Atertio Thromb Vase Biol.
Lundh B; Cavallin-Stahl E; Mercke C. (1975). Nicotinic acid and the endogenous production of carbon monoxide.
Acta Med Scand 197: 173-176.
Luomanmaki K; Coburn RF. (1969). Effects of metabolism and distribution of carbon monoxide on blood and body
stores. Am J Physiol 217: 354-363.
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-
219.
Lyall F; Myatt L. (2002). The role of the placenta in pre-eclampsia~a workshop report. Placenta 23 Suppl A: S142-
145.
March 2009
R-18
DRAFT-DO NOT QUOTE OR CITE
-------
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.
MacKay C; Pandow M; Wolfgang R. (1963). On the chemistry of natural radiocarbon. J geophys Res 68: 3929-
3931.
MacMillan V. (1975). Regional cerebral blood flow of the rat in acute carbon monoxide intoxication. Can J Physiol
Pharmacol 53: 644-650.
Mactutus CF; Fechter LD. (1984). Prenatal exposure to carbon monoxide: learning and memory deficits. Science
223:409-411.
Mactutus CF; Fechter LD. (1985). Moderate prenatal carbon monoxide exposure produces persistent, and apparently
permanent, memory deficits in rats. Teratology 31:1.
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005a). 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.
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005b). Outdoor air
pollution and stroke in Sheffield, United Kingdom: a small-area level geographical study. Stroke 36: 239-
243.
Maines MD; Kappas A. (1974). Cobalt induction of hepatic heme oxygenase; with evidence that cytochrome P-450
is not essential fortius enzyme activity. Proceedings of the National Academy of Sciences 71: 4293-4297.
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 Chem261: 411-419.
Maisonet M; Bush TJ; Correa A; Jaakkola JJ. (2001). Relation between ambient air pollution and low birth weight
in the Northeastern United States. Environ Health Perspect 109 Suppl 3: 351-356.
Manier G; Moinard J; Techoueyres P; Varene N; Guenard H. (1991). Pulmonary diffusion limitation after prolonged
strenuous exercise. Respir Physiol 83: 143-153.
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.
Mannes T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on
birth weight in Sydney, Australia. Occupational and Environmental Medicine 62: 524-530.
Manno M; Rugge M; Cocheo V. (1992). Double fatal inhalation of diehloromethane. Hum Exp Toxicol 11: 540-
545.
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.
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.
Marshall JD; Nethery E; Brauer M. (2008). Within-urban variability in ambient air pollution: comparison of
estimation methods. Atmos Environ 42: 1359-1369.
Marthan R; Castaing Y; Manier G; Guenard H. (1985). Gas exchange alterations in patients with chronic obstructive
lung disease. Chest 87: 470-475.
Martin JA; Hamilton BE; Sutton PD; Ventura SJ; Menacker F; Kirmeyer S; Munson ML. (2007). Births: final data
for 2005. In National vital statistics reports. National Center for Health Statistics, Hyattsville, MD. 48-11.
Martins LC; Latorre Mdo R; Saldiva PH; Braga AL. (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.
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.
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. American Journal of Respiratory and
Critical Care Medicine 171: 354-360.
McCartney ML. (1990). Sensitivity analysis applied to Coburn-Forster-Kane models of carboxyhemoglobin
formation. Am Ind Hyg Assoc J 51: 169-177.
McClean PA; Duguid NJ; Griffin PM; Newth CJ; Zamel N. (1981). Changes in exhaled pulmonary diffusing
capacity at rest and exercise in individuals with impaired positional diffusion. Bull Eur Physiopathol Respir
17: 179-186.
McClellan RO. (1991). CASAC Closure on the 1991 AQCD for Carbon Monoxide: US EPA-SAB-CASAC-LTR-
92-015, Washington, DC.
March 2009
R-19
DRAFT-DO NOT QUOTE OR CITE
-------
McClellan RO. (1992). CASAC Closure on the OAQPS Staff Paper for Carbon Monoxide: US EPA-SAB-CASAC-
LTR-92-016, Washington, DC.
McCoubrey WK, Jr.; Huang TJ; Maines MD. (1997). Isolation and characterization of a cDNA from the rat brain
that encodes hemoprotein heme oxygenase-3. Eur J Biochem 247: 725-732.
McGowan JA; Hider RN; Chacko E; Town GI. (2002). Particulate air pollution and hospital admissions in
Christchurch, New Zealand. Aust N Z J Public Health 26: 23-29.
McGrath JJ. (1992). Effects of altitude on endogenous carboxyhemoglobin levels. J Toxicol Environ Health 35:
127-133.
McGrath JJ; Schreck RM; Lee PS. (1993). Carboxyhemoglobin levels in humans: effects of altitude Inhalation
Toxicology 5: 241-249.
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.
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.
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.
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.
McLaughlin BE; Lash GE; Smith GN; Marks GS; Nakatsu K; Graham CH; Brien JF. (2003). Heme oxygenase
expression in selected regions of term human placenta. Exp Biol Med (Maywood) 228: 564-567.
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.
Medeiros A; Gouveia N. (2005). Relationship between low birthweight and air pollution in the city of Sao Paulo,
Brazil. Rev Saude Publica 39: 965-972.
Meilin S; Rogatsky GG; Thom 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.
Melin A; Obert P; Rebocho M; Bonnet P. (2002). Cardiac morphology and function following long-term exposure
to carbon monoxide at high altitude in rats. J Toxicol Environ Health A 65: 1981-1998.
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.
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.
Mensink C; Cosemans G. (2008). From traffic flow simulations to pollutant concentrations in street canyons and
backyards. Environmental Modelling & Software 23: 288-295.
Mercke C; Cavallin-Stahl E; Lundh B. (1975a). Carbon monoxide production and reticulocyte count in normal
women. Effect of contraceptive drugs and smoking. Acta Med Scand 198: 155-160.
Mercke C; Cavallin-Stahl E; Lundh B. (1975b). Heme catabolism during short-term treatment with phenobarbital,
diazepam and oxazepam. Acta Med Scand 198: 149-154.
Mercke C; Lundh B. (1976). Erythrocyte filterability and heme catabolism during the menstrual cycle. Ann Intern
Med 85: 322-324.
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.
Merx MW; Flogel U; Stumpe T; Godecke A; Decking UK; Schrader J. (2001). Myoglobin facilitates oxygen
diffusion. FASEB J 15: 1077-1079.
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2004a).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology 15: 46-56.
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Mulholland JA. (2004b). Ambient air pollution and
arrhythmic events inpatients with implanted cardioverter defibrillators, 1993-2002. Epidemiology 15: S22-
S23.
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.
March 2009
R-20
DRAFT-DO NOT QUOTE OR CITE
-------
Meyer J; Prien T; Van Aken H; Bone HG; 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.
Mfula AM; Kukadia V; Griffiths RF; Hall DJ. (2005). Wind tunnel modelling of urban building exposure to outdoor
pollution. Atmos Environ 39: 2737-2745.
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.
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.
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. New Engl J Med 356: 447.
Mobley D; Deslauriers M; Rojas-Brachos L. (2005). Improving Emission Inventories for Effective Air-Quality
Management across North America—a NARSTO Assessment. NARSTO, Pasco WA. NARSTO-05-001.
http://www.narsto.org/section.src?SID=8#5.
Moolgavkar SH. (2000a). Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
metropolitan areas in the United States. Inhalation Toxicology 12 Suppl 4: 75-90.
Moolgavkar SH. (2000b). Air pollution and daily mortality in three U.S. counties. Environ Health Perspect 108:
777-784.
Moolgavkar SH. (2003a). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook Counties
in Health Effects Institute Special Report: Revised Analyses of Time-Series Studies of Air Pollution and
Health.
Moolgavkar SH. (2003b). Air pollution and daily mortality in two U.S. counties: season-specific analyses and
exposure-response relationships. Inhalation Toxicology 15: 877-907.
Morita T; Mitsialis SA; Koike H; Liu Y; Kourembanas S. (1997). Carbon monoxide controls the proliferation of
hypoxic vascular smooth muscle cells. J Biol Chem 272: 32804-32809.
Morse CI; Pritchard LJ; Wust RC; Jones DA; Degens H. (2008). Carbon monoxide inhalation reduces skeletal
muscle fatigue resistance. Acta Physiol (Oxf) 192: 397-401.
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.
Naik JS; Walker BR. (2003). Heme oxygenase-mediated vasodilation involves vascular smooth muscle cell
hyperpolarization. Am J Physiol Heart Circ Physiol 285: H220-228.
NAPAP. (1991). The experience and legacy of NAPAP. Report of the Oversight Review Board. : National Acid
Precipitation Program., Washington, DC.
National Center for Health Statistics. (2007). Health, United States, 2007: With chartbook on trends in health of
Americans. Center for Disease Control, Hyattsville, MD,
Ndisang JF; Tabien HE; Wang R. (2004). Carbon monoxide and hypertension. J Hypertens 22: 1057-1074.
Neas LM; Schwartz J. (1996). The determinants of pulmonary diffusing capacity in a national sample of U.S. adults.
American Journal of Respiratory and Critical Care Medicine 153: 656-664.
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.
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.
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.
Norris G; YoungPong 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.
Novelli PC; Masarie KA; Lang PM; Hall BD; Myers RC; Elkins JW. (2003). Reanalysis of tropospheric CO trends:
Effects of the 1997-1998 wildfires. Journal of Geophysical Research-Atmospheres 108: 4464.
Novelli PC; Masarie KA; Atmospheric Carbon Monoxide Dry Air Mole Fractions from the NOAA ESRL Carbon
Cycle Cooperative Global Air Sampling Network -V—; ftp://ftp.cmdl.noaa.gov/ccg/co/flask/event/. (2008).
Atmospheric Carbon Monoxide Dry Air Mole Fractions from the NOAA ESRL Carbon Cycle Cooperative
Global Air Sampling Network, 1988-2007..
NRC. (1999). Ozone-Forming Potential of Reformulated Gasoline. National Research Council, National Academies
Press, Washington, D.C.
March 2009
R-21
DRAFT-DO NOT QUOTE OR CITE
-------
NRC. (2003). Managing carbon monoxide pollution in meteorological and topographical problem areas.
Washington, D.C.: National Academies Press.
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.
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. Environmental Research 81: 231-238.
Ott W, Switzer, P., Willits, N. (1994). Carbon monoxide exposures inside an automobile traveling on an urban
arterial highway. J Air Waste Manage Assoc 44: 1010-1018.
Ott WR; Mage DT. (1978). Interpreting urban carbon monoxide concentrations by means of a computerized blood
COHb model. J AirPollut Control Assoc 28: 911-916.
Otto DA; Benigus VA; Prah JD. (1979). Carbon monoxide and human time discrimination: failure to replicate
Beard-Wertheim experiments. Aviation Space and Environmental Medicine 50: 40-43.
Ozawa N; Goda N; Makino N; Yamaguchi T; Yoshimura Y; Suematsu M. (2002). Leydig cell-derived heme
oxygenase-1 regulates apoptosis of premeiotic germ cells in response to stress. J Clin Invest 109: 457-467.
Pankow D; MatschinerF; WeigmannHJ. (1991). Influence of aromatic hydrocarbons on the metabolism of
dichloromethane to carbon monoxide in rats. Toxicology 68: 89-100.
Park H; Lee B; Ha EH; Lee JT; Kim H; Hong YC. (2002). Association of air pollution with school absenteeism due
to illness. Arch Pediatr Adolesc Med 156: 1235-1239.
Park JW; Lim YH; Kyung SY; An CH; Lee SP; Jeong SH; Ju YS. (2005a). 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.
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005b). Effects of air pollution on heart rate
variability: the VA normative aging study. Environ Health Perspect 113: 304-309.
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in
California. Pediatrics 115: 121-128.
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. Journal of
Geophysical Research 107: 4140.
Parrish DD. (2006). Critical evaluation of US on-road vehicle emission inventories. Atmos Environ 40: 2288-2300.
Patel AP; John Moody A; Handy RD; Robert Sneyd J. (2003). Carbon monoxide exposure in rat heart: glutathione
depletion is prevented by antioxidants. Biochem Biophys Res Commun 302: 392-396.
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2005).
Ambient air pollution and respiratory emergency department visits. Epidemiology 16: 164-174.
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. American Journal of
Epidemiology 165: 625-633.
Pekkanen J; Brunner EJ; Anderson HR; Tiittanen P; Atkinson RW. (2000). Daily concentrations of air pollution and
plasma fibrinogen in London. Occupational and Environmental Medicine 57: 818-822.
Pelham TW; Holt LE; Moss MA. (2002). Exposure to carbon monoxide and nitrogen dioxide in enclosed ice arenas.
Occupational and Environmental Medicine 59: 224-233.
Penn A; Currie J; Snyder C. (1992). Inhalation of carbon monoxide does not accelerate arteriosclerosis in cockerels.
Eur J Pharmacol 228: 155-164.
Penney DG; BaylerianMS; Thill JE; Fanning CM; Yedavally S. (1982). Postnatal carbon monoxide exposure:
immediate and lasting effects in the rat. American Journal of Physiology- Heart and Circulatory Physiology
243: 328-339.
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: H289-297.
Penney DG; Barthel BG; Skoney JA. (1984). Cardiac compliance and dimensions in carbon monoxide-induced
cardiomegaly. Cardiovasc Res 18: 270-276.
Penney DG. (1988). Hemodynamic response to carbon monoxide. Environ Health Perspect 77: 121-130.
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.
Penney DG. (2007). Carbon Monoxide Poisoning. Boca Raton, FL CRC Press.
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.
March 2009
R-22
DRAFT-DO NOT QUOTE OR CITE
-------
Penttinen P; Tiittanen P; Pekkanen J. (2004). Mortality and air pollution in metropolitan Helsinki, 1988—1996.
Scand J Work Environ Health 30 Suppl 2: 19-27.
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.
Pesola GR; Huggins G; Sherpa TY. (2006). Abnormal predicted diffusion capacities in healthy Asians: an inequality
with a solution. Respiration 73: 799-807.
Peters A; Perz S; Doring A; Stieber J; Koenig W; Wichmann HE. (1999a). Increases in heart rate during an air
pollution episode. American Journal of Epidemiology 150: 1094-1098.
Peters A; Skorkovsky J; Kotesovec F; Brynda J; Spix C; Wichmann HE; Heinrich J. (2000a). Associations between
mortality and air pollution in central Europe. Environ Health Perspect 108: 283-287.
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery
DW. (2000b). Air pollution and incidence of cardiac arrhythmia. Epidemiology 11: 11-17.
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of
myocardial infarction. Circulation 103: 2810-2815.
Peters JM; Avol E; Navidi W; London SJ; Gauderman WJ; Lurmann F; Linn WS; Margolis H; Rappaport E; Gong
H; Thomas DC. (1999b). A study of twelve Southern California communities with differing levels and
types of air pollution. I. Prevalence of respiratory morbidity. American Journal of Respiratory and Critical
Care Medicine 159: 760-767.
Petersen LC. (1977). The effect of inhibitors on the oxygen kinetics of cytochrome c oxidase. Biochim Biophys
Acta 460: 299-307.
Peterson JE; Stewart RD. (1970). Absorption and elimination of carbon monoxide by inactive young men. Archives
of Environmental Health 21: 165-171.
Peterson JE; Stewart RD. (1975). Predicting the carboxyhemoglobin levels resulting from carbon monoxide
exposures. J Appl Physiol 39: 633-638.
Pfister G; Hess PG; Emmons LK; Lamarque JF; Wiedinmyer C; Edwards DP; Petron G; Gille JC; Sachse GW.
(2005). Quantifying CO emissions from the 2004 Alaskan wildfires using MOPITT CO data. Geophys Res
Lett 32: 1-5.
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.
Piantadosi CA. (2002). Biological chemistry of carbon monoxide. Antioxid Redox Signal 4: 259-270.
Piantadosi CA; Carraway MS; Suliman HB. (2006). Carbon monoxide, oxidative stress, and mitochondrial
permeability pore transition. Free Radic Biol Med 40: 1332-1339.
Piantadosi CA. (2008). Carbon monoxide, reactive oxygen signaling, and oxidative stress. Free Radic Biol Med 45:
562-569.
Piijola L; Paasonen P; Pfeiffer D; Hussein T; Hameri K; Koskentalo T; Virtanen A; Ronkko T; Keskinen J;
Pakkanen TA. (2006). Dispersion of particles and trace gases nearby a city highway: mobile laboratory
measurements in Finland. Atmos Environ 40: 867-879.
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
Manage Assoc 57: 366-379.
Pollack A; Lindhjem C; Stoeckenius T; Tran C; Mansell G; Jimenez M; Wilson G; Coulter-Burke S. (2004).
Evaluation of the US EPA MOBILE6 Highway Vehicle Emissions Factor Model. ENVIRON International,
Novato, CA. CRC Project E-64.
Pope CA, 3rd; ThunMJ; Namboodiri MM; Dockery DW; Evans JS; SpeizerFE; Heath CW, Jr. (1995). Particulate
air pollution as a predictor of mortality in a prospective study of U.S. adults. American Journal of
Respiratory and Critical Care Medicine 151: 669-674.
Pope CA, 3rd; 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. Journal of the
American Medical Association 287: 1132-1141.
Prigge E; Hochrainer D. (1977). Effects of carbon monoxide inhalation on erythropoiesis and cardiac hypertrophy in
fetal rats. Toxicol Appl Pharmacol 42: 225-228.
Prommer N; Schmidt W. (2007). Loss of CO from the intravascular bed and its impact on the optimised CO-
rebreathing method. Eur J Appl Physiol 100: 383-391.
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 2: 97-112.
March 2009
R-23
DRAFT-DO NOT QUOTE OR CITE
-------
Rabinovitch N; 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.
Rainham DG; Smoyer-Tomic KE. (2003). The role of air pollution in the relationship between a heat stress index
and human mortality in Toronto. Environmental Research 93: 9-19.
Rajendra Acharya U; Paul Joseph K; Kannathal N; Lim CM; Suri JS. (2006). Heart rate variability: a review. Med
BiolEng Comput44: 1031-1051.
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.
Raub JA; Mathieu-Nolf M; Hampson NB; Thom SR. (2000). Carbon monoxide poisoning—a public health
perspective. Toxicology 145: 1-14.
Raub JA; Benignus VA. (2002). Carbon monoxide and the nervous system. Neurosci Biobehav Rev 26: 925-940.
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.
Reeves RB; Park HK. (1992). CO uptake kinetics of red cells and CO diffusing capacity. Respir Physiol 88: 1-21.
Ren X; Dorrington KL; Robbins PA. (2001). Respiratory control in humans after 8 h of lowered arterial P02,
hemodilution, or carboxyhemoglobinemia. J Appl Physiol 90: 1189-1195.
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.
RichDQ; Schwartz J; MittlemanMA; LinkM; Luttmann-Gibson H; Catalano PJ; SpeizerFE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. American
Journal of Epidemiology 161: 1123-1132.
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR; Dockery
DW. (2006a). Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in
ambient air pollution. Environ Health Perspect 114: 120-123.
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006b). Association of
ventricular arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the
St Louis, Missouri metropolitan area. Occupational and Environmental Medicine 63: 591-596.
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. Inhalation Toxicology 16: 363-372.
Richardson RS; Noyszewski EA; Saltin B; Gonzalez-Alonso J. (2002). Effect of mild carboxy-hemoglobin on
exercising skeletal muscle: intravascular and intracellular evidence. American Journal of Physiology-
Regulatory, Integrative and Comparative Physiology 283: R1131-1139.
RiedikerM; Williams R; Devlin R; Griggs T; Bromberg P. (2003). Exposure to Particulate Matter, Volatile Organic
Compounds, and Other Air Pollutants Inside Patrol Cars. Environmental Science and Technology 37: 2084-
2093.
Rigby M; Toumi R. (2008). London air pollution climatology: Indirect evidence for urban boundary layer height and
wind speed enhancement. Atmos Environ 42: 4932-4947.
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. Journal of Exposure Science and Environmental
Epidemiology 16: 131-137.
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.
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.
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in
Southern California. American Journal of Epidemiology 155: 17-25.
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics
118:493-502.
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. American Journal of Epidemiology
166: 1045-1052.
March 2009
R-24
DRAFT-DO NOT QUOTE OR CITE
-------
Rodes C; Sheldon L; Whitaker D; Clayton A; Fitzgerald K. (1999). Measuring concentrations of selected air
pollutants inside California vehicles. Final report. Sierra Research Institute, Sacramento, CA. PB--99-
161028/XAB.
Rodgers PA; VremanHJ; Dennery PA; Stevenson DK. (1994). Sources of carbon monoxide (CO) in biological
systems and applications of CO detection technologies. Semin Perinatal 18: 2-10.
Rodkey FL; O'Neal JD; CollisonHA. (1969). Oxygen and carbon monoxide equilibria of human adult hemoglobin
at atmospheric and elevated pressure. Blood 33: 57-65.
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.
Roemer WH; van Wijnen JH. (2001). Daily mortality and air pollution along busy streets in Amsterdam, 1987-1998.
Epidemiology 12: 649-653.
Rosamond W; Flegal K; Furie K; Go A; Greenlund K; Haase N; Hailpern SM; Ho M; Howard V; Kissela B; Kittner
S; Lloyd-Jones D; McDermott M; Meigs J; Moy C; Nichol G; O'Donnell C; Roger V; Sorlie P; Steinberger
J; Thom T; Wilson M; Hong Y. (2008). Heart disease and stroke statistics - 2008 update - A report from the
American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 117:
E25-E146.
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.
Roth RA; Rubin RJ. (1976a). 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.
Roth RA; Rubin RJ. (1976b). 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.
RothmanKJ; Greenland S. (1998). Modern epidemiology (2 ed.). Philadelphia, PA: Lippincott-Raven Publishers.
Roughton FJ; 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.
Roughton FJ; Forster RE. (1957). Relative importance of diffusion and chemical reaction rates in determining rate
of exchange 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.
Roughton FJ. (1970). The equilibrium of carbon monoxide with human hemoglobin in whole blood. Ann N Y Acad
Sci 174: 177-188.
Ruckerl R; Ibald-Mulli A; Koenig W; Schneider 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. American Journal of Respiratory and Critical Care Medicine 173: 432-441.
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
Jacquemin B; von Klot 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.
Ryan PH; LeMasters GK. (2007). A Review of Land-use Regression Models for Characterizing Intraurban Air
Pollution Exposure. Inhalation Toxicology 19: 127-133.
Ryter SW; Alam J; Choi AM. (2006). Heme oxygenase-1/carbon monoxide: from basic science to therapeutic
applications. Physiol Rev 86: 583-650.
Salam MT; Millstein J; Li YF; 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.
Samet JM; Dominici F; Curriero FC; Coursac I; Zeger SL. (2000a). Fine particulate air pollution and mortality in 20
U.S. cities, 1987-1994. N Engl J Med 343: 1742-1749.
Samet JM; Dominici F; Zeger SL; Schwartz J; Dockery DW. (2000b). The National Morbidity, Mortality, and Air
Pollution Study. Part I: Methods and methodologic issues. Res Rep Health Eff Inst: 5-14; discussion 75-84.
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.
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology:
confounders or surrogates? Environ Health Perspect 109: 1053-1061.
March 2009
R-25
DRAFT-DO NOT QUOTE OR CITE
-------
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.
Sarnat SE. (2006). Factors affecting the association between ambient concentrations and personal exposures to
particles and gases. Environ Health Perspect 114: 649-654.
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. Occupational and Environmental
Medicine 63: 700-706.
Sartiani L; Cerbai E; Lonardo G; DePaoli P; Tattoli M; Cagiano R; Carratu MR; 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.
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. American Journal of Epidemiology 164: 505-517.
Schoenfisch WH; Hoop KA; Struelens BS. (1980). Carbon monoxide absorption through the oral and nasal mucosae
of cynomolgus monkeys. Archives of Environmental Health 35: 152-154.
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.
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.
Schwetz BA; Smith FA; Leong BK; Staples RE. (1979). Teratogenic potential of inhaled carbon monoxide in mice
and rabbits. Teratology 19: 385-392.
Scotto di Marco G; Kephalopoulos S; Ruuskanen J; Jantunen M. (2005). Personal carbon monoxide exposure in
Helsinki, Finland. Atmos Environ 39: 2697-2707.
Seila RL; Lonneman WA; Meeks SA. (1989). Determination of C2 to C12 ambient air hydrocarbons in 39 US cities
from 1984 through 1986. PB-89-214142/XAB, Computer Sciences Corp., Research Triangle Park, NC
(USA),
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.
Sharma VS; Schmidt MR; Ranney HM. (1976). Dissociation of CO from carboxyhemoglobin. J Biol Chem 251:
4267-4272.
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.
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.
Sheppard L; Slaughter JC; Schildcrout J; Liu LJ; Lumley T. (2005). Exposure and measurement contributions to
estimates of acute air pollution effects. Journal of Exposure Analysis and Environmental Epidemiology 15:
366.
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.
Archives of Environmental Health 42: 108-116.
Sheps DS; Herbst MC; Hinderliter AL; Adams KF; Ekelund LG; O'Neil JJ; Goldstein GM; Bromberg PA; Dalton
JL; Ballenger MN; et al. (1990). Production of arrhythmias by elevated carboxyhemoglobin in patients with
coronary artery disease. Ann Intern Med 113: 343-351.
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.
Shimazu T. (2001). Half-life of blood carboxyhemoglobin. Chest 119: 661-663.
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.
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 Manage Assoc 54: 1212-1218.
Singh J. (1986). Early behavioral alterations in mice following prenatal carbon monoxide exposure.
Neurotoxicology 7: 475-481.
March 2009
R-26
DRAFT-DO NOT QUOTE OR CITE
-------
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.
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.
Singh MP; Sharan M; Selvakumar S. (1991). A mathematical model for the computation of carboxyhaemoglobin in
human blood as a function of exposure time. Philos Trans R Soc Lond B Biol Sci 334: 135-147.
Singh V; Khandelwal R; Gupta AB. (2003). Effect of air pollution on peak expiratory flow rate variability. J Asthma
40: 81-86.
Sitdikova GF; Islamov RR; Mukhamedyarov MA; Permyakova W; Zefirov AL; Palotas A. (2007). Modulation of
neurotransmitter release by carbon monoxide at the frog neuro-muscular junction. Curr Drug Metab 8: 177-
184.
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: 346-3 53.
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 Anal
Environ Epidemiol 15: 153-159.
Smith MV. (1990). Comparing solutions to the linear and nonlinear CFK equations for predicting COHb formation.
MathBiosci 99: 251-263.
Smith MV; Hazucha MJ; Benignus VA; Bromberg PA. (1994). Effect of regional circulation patterns on observed
HbCO levels. J Appl Physiol 77: 1659-1665.
So ESP; Chan ATY; Wong AYT. (2005). Large-eddy simulations of wind flow and pollutant dispersion in a street
canyon. Atmos Environ 39: 3573-3582.
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. Arch Toxicol 56: 106-108.
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.
Song R; Mahidhara RS; Liu F; Ning W; Otterbein LE; Choi AM. (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.
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.
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. Archives of Environmental Health 56: 167-174.
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air-pollution
and inflammation-sensitive biomarkers. Environmental Research 106: 51-61.
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.
Stewart RD; Newton PE; Hosko MJ; Peterson JE. (1973a). Effect of carbon monoxide on time perception. Archives
of Environmental Health 27: 155-160.
Stewart RD; Peterson JE; Fisher TN; Hosko MJ; Baretta ED; Dodd HC; Herrmann AA. (1973b). Experimental
human exposure to high concentrations of carbon monoxide. Archives of Environmental Health 26: 1-7.
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 Manage Assoc 52:
470-484.
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 Manage Assoc 53: 258-261.
Stieb DM; Evans GJ; Sabaliauskas K; Chen LI; Campbell ME; Wheeler AJ; Brook JR; Guay M. (2007). A scripted
activity study of the impact of protective advice on personal exposure to ultra-fine and fine particulate
matter and volatile organic compounds. Journal of Exposure Science and Environmental Epidemiology 18:
495-502.
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.
March 2009
R-27
DRAFT-DO NOT QUOTE OR CITE
-------
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Heinrich J; Wichmann HE; Peters A. (2006). Daily
mortality and particulate matter in different size classes in Erfurt, Germany. Journal of Exposure Science
and Environmental Epidemiology 17: 458-467.
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.
Storm JE; Fechter LD. (1985a). Alteration in the postnatal ontogeny of cerebellar norepinephrine content following
chronic prenatal carbon monoxide. J Neurochem 45: 965-969.
Storm JE; Fechter LD. (1985b). Prenatal carbon monoxide exposure differentially affects postnatal weight and
monoamine concentration of rat brain regions. Toxicol Appl Pharmacol 81: 139-146.
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.
Styka PE; Penney DG. (1978). Regression of carbon monoxide-induced cardiomegaly. American Journal of
Physiology- Heart and Circulatory Physiology 235: 516-522.
Suliman HB; Carraway MS; Tatro LG; Piantadosi CA. (2007). A new activating role for CO in cardiac
mitochondrial biogenesis. J Cell Sci 120: 299.
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. American Journal of Epidemiology 157: 501-509.
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.
Sun Y; Jin K; Mao XO; Zhu Y; Greenberg DA. (2001). Neuroglobin is up-regulated by and protects neurons from
hypoxic-ischemic injury. Proceedings of the National Academy of Sciences 98: 15306-15311.
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.
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.
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. American Journal of Epidemiology 164: 421-433.
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.
Takeuchi A; Vesely A; Rucker J; Sommer LZ; Tesler J; Lavine E; Slutsky AS; Maleck WH; Volgyesi G; Fedorko
L; Iscoe S; Fisher JA. (2000). A simple "new" method to accelerate clearance of carbon monoxide.
American Journal of Respiratory and Critical Care Medicine 161: 1816-1819.
Tarkiainen TH; Timonen KL; Vanninen EJ; Aim S; Hartikainen JE; Pekkanen J. (2003). Effect of acute carbon
monoxide exposure on heart rate variability in patients with coronary artery disease. Clin Physiol Funct
Tmaging 23: 98-102.
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.
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.
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. Archives of Environmental Health 57: 41-47.
TeranFJ; 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. American Journal of Physiology- Regulatory,
Integrative and Comparative Physiology 288: 615-622.
Thom 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.
Thom SR; Ischiropoulos H. (1997). Mechanism of oxidative stress from low levels of carbon monoxide. Res Rep
Health Elf Inst: 1-19; discussion 21-17.
Thom SR; Xu YA; Ischiropoulos H. (1997). Vascular endothelial cells generate peroxynitrite in response to carbon
monoxide exposure. ChemRes Toxicol 10: 1023-1031.
March 2009
R-28
DRAFT-DO NOT QUOTE OR CITE
-------
Thom SR; Ohnishi ST; Fisher D; Xu YA; Ischiropoulos H. (1999a). Pulmonary vascular stress from carbon
monoxide. Toxicol Appl Pharmacol 154: 12-19.
Thom SR; Fisher D; Xu YA; Garner S; Ischiropoulos H. (1999b). Role of nitric oxide-derived oxidants in vascular
injury from carbon monoxide in the rat. Am J Physiol 276: H984-992.
Thom SR; Fisher D; Xu YA; Notarfrancesco K; Ischiropoulos H. (2000). Adaptive responses and apoptosis in
endothelial cells exposed to carbon monoxide. Proceedings of the National Academy of Sciences 97: 1305-
1310.
Thom SR; Fisher D; Manevich Y. (2001a). Roles for platelet-activating factor and? NO-derived oxidants causing
neutrophil adherence after CO poisoning. American Journal of Physiology- Heart and Circulatory
Physiology 281: 923-930.
Thom SR; Fisher D; Manevich Y. (2001b). Roles for platelet-activating factor and *NO-derived oxidants causing
neutrophil adherence after CO poisoning. Am J Physiol Heart Circ Physiol 281: H923-930.
Thom SR; Bhopale VM; Han ST; Clark JM; Hardy KR. (2006). Intravascular neutrophil activation due to carbon
monoxide poisoning. American Journal of Respiratory and Critical Care Medicine 174: 1239-1248.
Thompson AJ; Shields MD; Patterson CC. (2001). Acute asthma exacerbations and air pollutants in children living
in Belfast, Northern Ireland. Archives of Environmental Health 56: 234-241.
Thompson AM; Cicerone RJ. (1986). Possible perturbations to atmospheric CO, CH4, and OH. Journal of
Geophysical Research 91: 10853-10864.
ThomsenHK. (1974). Carbon monoxide-induced atherosclerosis in primates. An electron-microscopic study on the
coronary arteries of Macacatrus monkeys. Atherosclerosis 20: 233-240.
Thorup C; Jones CL; Gross SS; Moore LC; Goligorsky MS. (1999). Carbon monoxide induces vasodilation and
nitric oxide release but suppresses endothelial NOS. American Journal of Physiology- Renal Physiology
277: 882-889.
Tikuisis P; Buick F; Kane DM. (1987a). Percent carboxyhemoglobin in resting humans exposed repeatedly to 1,500
and 7,500 ppm CO. J Appl Physiol 63: 820-827.
Tikuisis P; Madill HD; Gill BJ; Lewis WF; Cox KM; Kane DM. (1987b). A critical analysis of the use of the CFK
equation in predicting COHb formation. Am Ind Hyg Assoc J 48: 208-213.
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. Occupational and Environmental
Medicine 59: 129-134.
Timonen KL; Vanninen E; de Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Heinrich J; Hoek G; Lanki T;
Peters A. (2006). Effects of ultrafine and fine particulate and gaseous air pollution on cardiac autonomic
control in subjects with coronary artery disease: The ULTRA study. Journal of Exposure Science and
Environmental Epidemiology 16: 332-341.
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. Journal of Exposure Science and Environmental
Epidemiology 17 Suppl2: S29-35.
Tolcos M; Mallard C; McGregor H; Walker D; Rees S. (2000a). 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.
Tolcos M; McGregor H; Walker D; Rees S. (2000b). 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.
Topping DL. (1975). Acute effects of carbon monoxide on the metabolism of perfused rat liver. Biochem J 152:
425-427.
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.
Tsai SS; Huang CH; Goggins WB; Wu TN; Yang CY. (2003a). Relationship between air pollution and daily
mortality in a tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A 66: 1341-1349.
Tsai SS; Goggins WB; Chiu HF; Yang CY. (2003b). Evidence for an association between air pollution and daily
stroke admissions in Kaohsiung, Taiwan. Stroke 34: 2612-2616.
Tsai SS; Cheng MH; Chiu HF; Wu TN; Yang CY. (2006a). Air pollution and hospital admissions for asthma in a
tropical city: Kaohsiung, Taiwan. Inhalation Toxicology 18: 549-554.
Tsai SS; Chen CC; Hsieh HJ; Chang CC; Yang CY. (2006b). Air pollution and postneonatal mortality in a tropical
city: Kaohsiung, Taiwan. Inhalation Toxicology 18: 185-189.
March 2009
R-29
DRAFT-DO NOT QUOTE OR CITE
-------
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 (Vol. 86, pp. 3833-3839): Endocrine Soc.
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.
Tyuma I; Ueda Y; Imaizumi K; Kosaka H. (1981). Prediction of the carbonmonoxyhemoglobin levels during and
after carbon monoxide exposures in various animal species. Jpn J Physiol 31: 131-143.
U.S. Code. (1970). Clean Air Act, §108, air quality criteria and control techniques. Retrieved, from.
U.S. EPA. (1978). Altitude as a factor in air pollution. Office of Research and Development, Environmental Criteria
and Assessment Office, Research Triangle Park, NC. EPA 600/9-78-015.
U.S. EPA. (1979a). Air quality criteria for carbon monoxide. Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, U.S. Environmental Protection Agency, Research Triangle
Park. NC. EPA-600/8-79-022.
U.S. EPA. (1979b). Assessment of Adverse Health Effects from Carbon Monoxide and Implications for Possible
Modifications of the Standard. Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC.
U.S. EPA. (1984a). 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.
U.S. EPA. (1984b). Revised evaluation of health effects associated with carbon monoxide exposure: an addendum to
the 1979 EPA air quality criteria document for carbon monoxide. Research Triangle Park, NC: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office; report no. EPA-
600/8-83-033F.
U.S. EPA. (1991). Air quality criteria for carbon monoxide. EPA/600/8-90/045F, Office of Health and
Environmental Assessment., Environmental Criteria and Assessment Office. Research Triangle Park, NC,
Available from: NTIS, Springfield, VA, 1991. PB93-167492,
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. Research Triangle Park, NC: Office of Air
Quality Planning and Standards; report no. EPA-452/R-92-004.
U.S. EPA. (2000). Air quality criteria for carbon monoxide. National Center for Environmental Assessment,
Research Triangle Park, N.C. EPA/600/P-99/001F. http://www.epa.gov/ncea/pdfs/coaacd.pdf.
U.S. EPA. (2004). Air Quality Criteria for Particulate Matter. U.S. Environmental Protection Agency, Washington,
D.C. EPA/600/P-99/002aF-bF.
U.S. EPA. (2005). Guidelines for Carcinogen Risk Assessment. U.S. Environmental Protection Agency,
Washington, D.C. EPA/630/P-03/001F.
U.S. EPA. (2006a). 2002 National Emissions Inventory Data and Documentation. U.S. Environmental Protection
Agency, Office of Air and Radiation, Office of Air Quality Planning and Standards, Research Triangle
Park, NC. http://www.epa.gov/ttn/chief/net/2002inventorv.html.
U.S. EPA. (2006b). Air Quality Criteria for Ozone and Related Photochemical Oxidants. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Research Triangle Park, NC. EPA 600/R-05/004aF.
U.S. EPA. (2007). Documentation For The Final 2002 Mobile National Emissions Inventory - Version 3
U.S. EPA. (2008a). Inventory of U.S. Greenhouse Gas Emissions and Sinks, 1990-2006. U.S. Environmental
Protection Agency, Office of Atmospheric Programs, Washington, D.C. EPA 430-R-08-005.
U.S. EPA. (2008b). Plan for Review of the National Ambient Air Quality Standards for Carbon Monoxide. U.S.
Environmental Protection Agency, National Center for Environmental Assessment, Office of Research and
Development, Research Triangle Park, NC.
U.S. EPA. (2008c). U.S. EPA's 2008 Report on the Environment (Final Report). U.S. Environmental Protection
Agency, Office of Research and Development, National Center for Environmental Assessment,
Washington, DC. EPA/600/R-07/045F.
U.S. EPA. (2008d). Integrated Review Plan for the National Ambient Air Quality Standards for Particulate Matter.
U.S. Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Research Triangle Park, NC.
U.S. EPA. (2008e). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Research Triangle Park, NC. EPA/600/R-08/071.
March 2009
R-30
DRAFT-DO NOT QUOTE OR CITE
-------
U.S. EPA. (2008f). Integrated Science Assessment for Particulate Matter, First External Review Draft. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Research Triangle Park, NC. EPA/600/R-08/139.
U.S. EPA. (2008g). Integrated Science Assessment for Sulfur Oxides-Health Criteria. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Research Triangle Park, NC. EPA/600/R-08/047F-FA.
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. Proceedings of the National Academy of
Sciences 101: 17351-17356.
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. Atmospheric Chemistry and Physics 6:
3423-3441.
van Wijnen JH; Verhoeff AP; Jans HW; van Bruggen M. (1995). The exposure of cyclists, car drivers and
pedestrians to traffic-related air pollutants. Int Arch Occup Environ Health 67: 187-193.
Vedal S; Brauer M; White R; Petkau J. (2003). Air pollution and daily mortality in a city with low levels of
pollution. Environ Health Perspect 111: 45-52.
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with
implantable cardioverter defibrillators. Inhalation Toxicology 16: 353-362.
Veen AVD; Briggs DJ; Collins S; Elliott P; Fischer P; Kingham S; Lebret E; Pryl K; Van Reeuwijk H; Smallbone
K. (1997). Mapping urban air pollution using GIS: A regression-based approach. International Journal of
Geographical Information Science 11: 699-718.
Verma A; Hirsch DJ; Glatt CE; Ronnett GV; Snyder SH. (1993). Carbon monoxide: a putative neural messenger.
Science 259: 381-384.
Vesely AE; Somogyi RB; Sasano H; Sasano N; Fisher JA; Duffin J. (2004). The effects of carbon monoxide on
respiratory chemoreflexes in humans. Environmental Research 94: 227-233.
Vigotti MA; Chiaverini F; Biagiola P; Rossi G. (2007). Urban air pollution and emergency visits for respiratory
complaints in Pisa, Italy. J Toxicol Environ Health A 70: 266-269.
VillamorE; PERez-VizcalNo 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.
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 Anal Environ
Epidemiol 13: 427-435.
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006a). Associations between outdoor air pollution and emergency
department visits for stroke in Edmonton, Canada. Eur J Epidemiol 21: 689-700.
Villeneuve PJ; Doiron MS; Stieb D; Dales R; Burnett RT; Dugandzic R. (2006b). Is outdoor air pollution associated
with physician visits for allergic rhinitis among the elderly in Toronto, Canada? Allergy 61: 750-758.
Vogel JA; Gleser MA. (1972). Effect of carbon monoxide on oxygen transport during exercise. J Appl Physiol 32:
234-239.
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. European
Respiratory Journal 20: 691-702.
von Klot S; Peters A; Aalto P; Bellander T; Berglind N; D'lppoliti D; Elosua R; Hormann A; Kulmala M; Lanki T;
Lowel H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F. (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.
VremanHJ; WongRJ; Kim EC; NabsethDC; Marks GS; Stevenson DK. (2000). Haem oxygenase activity in
human umbilical cord and rat vascular tissues. Placenta21: 337-344.
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.
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.
Wagner JA; Horvath SM; Dahms TE. (1975). Carbon monoxide elimination. Respir Physiol 23: 41-47.
Wang R; Wang Z; Wu L. (1997). Carbon monoxide-induced vasorelaxation and the underlying mechanisms. Br J
Pharmacol 121: 927-934.
March 2009
R-31
DRAFT-DO NOT QUOTE OR CITE
-------
Wang R. (1998). Resurgence of carbon monoxide: an endogenous gaseous vasorelaxing factor. Can J Physiol
Pharmacol 76: 1-15.
WangTN;Ko YC; Chao YY; Huang CC; LinRS. (1999). Association between indoor and outdoor air pollution and
adolescent asthma from 1995 to 1996 in Taiwan. Environmental Research 81: 239-247.
Weaver LK; Howe S; Hopkins R; Chan KJ. (2000). Carboxyhemoglobin half-life in carbon monoxide-poisoned
patients treated with 100% oxygen at atmospheric pressure. Chest 117: 801-808.
Weaver LK; Valentine KJ; Hopkins RO. (2007). Carbon Monoxide Poisoning: Risk Factors for Cognitive Sequelae
and the Role of Hyperbaric Oxygen. American Journal of Respiratory and Critical Care Medicine: 200701.
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.
Webber DS; Lopez I; Korsak RA; 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.
Weinstock B; Niki H. (1972). Carbon monoxide balance in nature (Vol. 176, pp. 290-292): The American
Association for the Advancement of Science.
Wellenius GA; Saldiva PHN; Batalha JRF; Krishna Murthy GG; 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 (Vol. 66, pp. 327-335): Soc Toxicology.
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.
Wellenius GA; Schwartz J; Mittleman MA. (2005a). Air pollution and hospital admissions for ischemic and
hemorrhagic stroke among medicare beneficiaries. Stroke 36: 2549-2553.
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005b). Particulate air pollution and the rate of
hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania.
American Journal of Epidemiology 161: 1030-1036.
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006a). Effects of ambient particles and
carbon monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhalation
Toxicology 18: 1077-1082.
Wellenius GA; Schwartz J; Mittleman MA. (2006b). Particulate air pollution and hospital admissions for congestive
heart failure in seven United States cities. Am J Cardiol 97: 404-408.
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.
Wichmann HE; 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. 1041-5505 (Print).
http://www.ncbi.nlm.nih.gov/entrez/auerv.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list uids=119
18089.
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.
Williams LJ; Rasmussen SA; Flores A; Kirby RS; Edmonds LD. (2005). Decline in the prevalence of spina bifida
and anencephaly by race/ethnicity: 1995-2002. Pediatrics 116: 580-586.
Williams R; Vette A; Croghan C; Sheldon L; Rodes C; Thornburg J; Ejire A; Herbst M; Sanders W; Suggs J; Rea
A; Leovic K. (2003). The Research Triangle Park particulate matter panel study: PM mass concentration
relationships. Atmos Environ 37: 5349-5363.
Wilson JG; Zawar-Reza P. (2006). Intraurban-scale dispersion modelling of particulate matter concentrations:
Applications for exposure estimates in cohort studies. Atmos Environ 40: 1053-1063.
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. Journal of the Air & Waste
Management Association (1995) 50: 1167.
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter exposure
from a personal monitoring panel study. Journal of Exposure Science and Environmental Epidemiology 16:
264-274.
March 2009
R-32
DRAFT-DO NOT QUOTE OR CITE
-------
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. Journal of
exposure science & environmental epidemiology 17: Sll.
Wirkner K; Poelchen W. (1996). Influence of long-term ethanol treatment on rat liver aniline and p-nitrophenol
hydroxylation. Alcohol 13: 69-74.
Wittenberg BA; Wittenberg JB; Caldwell PR. (1975). Role of myoglobin in the oxygen supply to red skeletal
muscle. J Biol Chem 250: 9038-9043.
Wittenberg BA; Wittenberg JB. (1993). Effects of carbon monoxide on isolated heart muscle cells. Health Effects
Institute, Boston, Massachusetts. Report No. 62.
http://www.ncbi.nlm.nih.gov/entrez/querv.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list uids=815
5298.
Wollmann HA. (1998). Intrauterine growth restriction: Definition and etiology. Hormone Res 49: 1-6.
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.
Wu L; Wang R. (2005). Carbon monoxide: endogenous production, physiological functions, and pharmacological
applications. Pharmacol Rev 57: 585-630.
Xiaomin X; Zhen H; Jiasong W. (2006). The impact of urban street layout on local atmospheric environment. Build
Environ 41: 1352-1363.
Xirasagar S; LinHC; Liu TC. (2006). Seasonality in pediatric asthma admissions: the role of climate and
environmental factors. Eur J Pediatr 165: 747-752.
Yang CY; Chang CC; Chuang HY; Tsai SS; Wu TN; Ho CK. (2004a). Relationship between air pollution and daily
mortality in a subtropical city: Taipei, Taiwan. Environ Int 30: 519-523.
Yang CY; Chen YS; Yang CH; Ho SC. (2004b). Relationship between ambient air pollution and hospital admissions
for cardiovascular diseases in kaohsiung, taiwan. J Toxicol Environ Health A 67: 483-493.
Yang CY; Hsieh HJ; Tsai SS; Wu TN; Chiu HF. (2006). Correlation between air pollution and postneonatal
mortality in a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A 69: 2033-2040.
Yang CY; Chen CC; Chen CY; Kuo HW. (2007). Air pollution and hospital admissions for asthma in a subtropical
city: Taipei, Taiwan. J Toxicol Environ Health A 70: 111-117.
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A 70: 1214-1219.
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. Inhalation Toxicology 15: 1297-1308.
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. Environmental Research
99: 99-105.
Yoshiki N; Kubota T; Aso T. (2001). Identification of heme oxygenase in human endometrium. J Clin Endocrinol
Metab 86: 5033-5038.
Young LJ; Caughey WS. (1986). Oxygenation of carbon monoxide by bovine heart cytochrome c oxidase.
Biochemistry 25: 152-161.
Yu O; Sheppard L; Lumley T; Koenig JQ; Shapiro GG. (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.
Zamudio S; Palmer SK; Regensteiner JG; Grindlay Moore L. (1995). High altitude and hypertension during
pregnancy. Am J Hum Biol 7: 183-193.
Zanobetti A; Schwartz J; Ridker PM. (2004a). Air pollution and markers of cardiovascular risk. Epidemiology 15:
S22.
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold
DR. (2004b). Ambient pollution and blood pressure in cardiac rehabilitation patients. Circulation 110:
2184-2189.
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community
Health 60: 890-895.
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.
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. J Reprod Immunol 69: 35-52.
March 2009
R-33
DRAFT-DO NOT QUOTE OR CITE
-------
Zevin S; Saunders S; Gourlay SG; Jacob P; Benowitz NL. (2001). Cardiovascular effects of carbon monoxide and
cigarette smoking. J Am Coll Cardiol 38: 1633-1638.
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.
Zhang X; Shan P; Alam J; Fu XY; Lee PJ. (2005). Carbon monoxide differentially modulates STAT1 and STAT3
and inhibits apoptosis via a phosphatidylinositol 3-kinase/Akt and p38 kinase-dependent STAT3 pathway
during anoxia-reoxygenation injury. J Biol Chem 280: 8714-8721.
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.
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.
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.
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.
March 2009
R-34
DRAFT-DO NOT QUOTE OR CITE
------- |