United States
Environmental Protection
Agency
Office of Research
and Development
Washington, DC 20460
EPA600/P-99/001B
October 1999
Second External Review Draft
EPA
Air Quality Criteria for
Carbon Monoxide
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on
its technical accuracy and policy implications.
-------
EPA 600/P-99/001B
October 1999
Second External Review Draft
Air Quality Criteria for
Carbon Monoxide
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Disclaimer
This document is an external draft for review purposes only and does not constitute Agency
policy. Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
October 1999 ii DRAFT-DO NOT QUOTE OR CITE
-------
Preface
The U.S. Environmental Protection Agency (EPA) promulgates the National Ambient Air
Quality Standards (NAAQS) on the basis of an up-to-date compilation of scientific knowledge
about the relationship between various concentrations of ambient air pollutants and their adverse
effects on man and the environment. These air quality criteria are published in criteria
documents. In 1970, the first air quality criteria document for carbon monoxide (CO) was issued
by the National Air Pollution Control Administration, a predecessor of EPA. On the basis of
scientific information in that document, NAAQS were promulgated for CO at levels of 9 ppm for
an 8-h average and 35 ppm for a 1-h average. Periodic scientific assessments of the published
literature were completed by EPA in 1979 and, again, in 1984. The last full-scale CO criteria
document revision was published in 1991. Although the air quality criteria have changed over
the past two decades, the NAAQS for CO have remained the same. This revised criteria
document consolidates and updates the current scientific basis for another reevaluation of the CO
NAAQS in accordance with the provisions identified in Sections 108 and 109 of the Clean Air
Act.
This document was prepared and reviewed by experts from state and federal government
offices, academia, and industry for use by EPA in support of decision making on potential public
health risks of CO; it describes the nature, sources, distribution, measurement, and concentrations
of CO in both the outdoor (ambient) and indoor environments and evaluates the latest data on the
health effects in exposed human populations. Although not intended to be an exhaustive
literature review, this document is intended to cover all pertinent literature through early 1999.
The National Center for Environmental Assessment—Research Triangle Park, NC,
acknowledges the contributions provided by the authors, contributors, and reviewers and the
diligence of its staff and contractors in the preparation of this document.
October 1999 iii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
Page
List of Tables xi
List of Figures xv
Authors, Contributors, and Reviewers xxi
U.S. Environmental Protection Agency, Science Advisory Board, Clean Air Scientific
Advisory Committee, Carbon Monoxide Review Panel xxxi
U.S. Environmental Protection Agency Project Team for Development of Air Quality
Criteria for Carbon Monoxide xxxiii
EXECUTIVE SUMMARY E-l
1. INTRODUCTION 1-1
1.1 LEGISLATIVE REQUIREMENTS 1-1
1.2 REGULATORY BACKGROUND 1-2
1.3 RATIONALE FOR THE EXISTING CARBON MONOXIDE STANDARDS .... 1-5
1.3.1 Carboxyhemoglobin Levels of Concern 1-5
1.3.2 Relationship Between Carbon Monoxide Exposure and
Carboxyhemoglobin Levels 1-7
1.3.3 Estimating Population Exposure 1-7
1.3.4 Decision on the Primary Standards 1-8
1.4 ISSUES OF CONCERN FOR THE CURRENT CRITERIA DEVELOPMENT ... 1-8
1.4.1 Sources and Emissions 1-8
1.4.2 Atmospheric Chemistry 1-9
1.4.3 Global Cycle 1-10
1.4.4 Measurement Technology 1-10
1.4.5 Ambient Air Quality 1-10
1.4.6 Indoor Emissions and Concentrations 1-10
1.4.7 Exposure Assessment 1-11
1.4.8 Mechanisms of Action 1-11
1.4.9 Health Effects 1-12
1.4.10 Carbon Monoxide Interaction with Drugs 1-13
1.4.11 Subpopulations at Risk 1-14
1.5 METHODS AND PROCEDURES FOR DOCUMENT PREPARATION 1-14
1.6 ORGANIZATION AND CONTENT OF THE DOCUMENT 1-16
REFERENCES 1-18
2. ANALYTICAL METHODS FOR MONITORING CARBON MONOXIDE 2-1
2.1 INTRODUCTION 2-1
2.2 OVERVIEW OF TECHNIQUES FOR MEASUREMENT OF AMBIENT
CARBON MONOXIDE 2-2
2.3 GAS STANDARDS FOR CALIBRATION 2-3
2.3.1 Pre-madeMixtures 2-5
2.3.1.1 Standard Reference Materials 2-5
October 1999 v DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
Page
2.3.1.2 National Institute of Standards and Technology Traceable
Reference Materials 2-6
2.3.1.3 U.S. Environmental Protection Agency Protocol Gases 2-6
2.3.1.4 Dutch Bureau of Standards 2-6
2.3.1.5 Commercial Blends 2-7
2.3.2 Laboratory Blended Mixtures 2-7
2.3.3 Other Methods 2-8
2.3.4 Intercomparisons of Standards 2-8
2.3.5 Infrared Absorption 2-10
2.4 MEASUREMENT IN AMBIENT AIR 2-10
2.4.1 Sampling System Components 2-10
2.4.2 Quality Assurance Procedures for Sampling 2-11
2.4.3 Sampling Schedules 2-13
2.4.4 Continuous Analysis 2-13
2.4.4.1 Nondispersive Infrared Photometry 2-13
2.4.4.2 Gas Chromatography-Flame lonization 2-16
2.4.4.3 Mercury Liberation 2-16
2.4.4.4 Tunable Diode Laser Spectroscopy 2-17
2.4.4.5 Resonance Fluorescence 2-17
2.4.5 Intercomparisons of Methods 2-18
2.4.6 Other Methods of Analysis 2-18
2.5 MEASUREMENT USING PERSONAL MONITORS 2-19
2.6 BIOLOGICAL MONITORING 2-20
2.6.1 Carboxyhemoglobin Measurements 2-20
2.6.2 Breath Carbon Monoxide Measurements 2-22
2.6.3 Relationships of Breath Carbon Monoxide to Blood
Carboxyhemoglobin 2-23
2.6.4 Summary of the Relationship Between Biological Measurements
of Carbon Monoxide 2-24
2.7 SUMMARY 2-26
REFERENCES 2-28
3. SOURCES, EMISSIONS, AND CONCENTRATIONS OF CARBON
MONOXIDE IN AMBIENT AND INDOOR AIR 3-1
3.1 INTRODUCTION 3-1
3.2 THE GLOBAL CYCLE OF CARBON MONOXIDE 3-2
3.2.1 Global Background Concentrations of Carbon Monoxide 3-3
3.2.2 Sources and Global Emissions Estimates of Carbon Monoxide 3-7
3.2.3 The Atmospheric Chemistry of Carbon Monoxide 3-10
3.3 NATIONWIDE CARBON MONOXIDE EMISSIONS ESTIMATES 3-16
3.4 CARBON MONOXIDE CONCENTRATIONS IN AMBIENT AIR 3-23
3.4.1 Nationwide Trends in Ambient Carbon Monoxide Concentrations 3-24
October 1999 vi DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
age
3.4.2 Orcadian Patterns in Carbon Monoxide Concentrations 3-30
3.4.3 Characterization of the Spatial and Temporal Variability in Carbon
Monoxide Concentrations in Selected U.S. Cities 3-34
3.5 SOURCES, EMISSIONS, AND CONCENTRATIONS OF CARBON
MONOXIDE IN INDOOR ENVIRONMENTS 3-51
3.5.1 Sources and Emissions of Carbon Monoxide in Indoor Environments
Prior to 1991 3-51
3.5.2 Unvented Combustion Sources and Estimated Emissions Rates 3-52
3.5.2.1 Gas Cooking Ranges and Ovens and Furnaces 3-52
3.5.2.2 Emissions from Unvented Space Heaters 3-55
3.5.2.3 Woodstoves and Fireplaces 3-57
3.5.2.4 Environmental Tobacco Smoke 3-58
3.5.3 Sources-Related Concentrations of Carbon Monoxide in Indoor
Environments Prior to 1991 3-59
3.5.4 Indoor Concentrations of Carbon Monoxide 3-60
3.5.4.1 Factors Affecting Carbon Monoxide Concentrations 3-60
3.5.4.2 Models for Carbon Monoxide Concentrations 3-65
3.5.4.3 Microenvironmental Monitoring Studies 3-66
3.6 SUMMARY 3-74
REFERENCES 3-79
4. POPULATION EXPOSURE TO CARBON MONOXIDE 4-1
4.1 INTRODUCTION 4-1
4.2 BRIEF SUMMARY OF POPULATION EXPOSURE STUDIES PRIOR
TO 1991 4-3
4.2.1 Sensitive Populations 4-4
4.2.2 Estimates of Population Exposure Based on Fixed-Site Monitors 4-4
4.2.3 Surveys of Population Exposure Using Personal Monitors 4-6
4.2.4 Population Exposure Models 4-8
4.3 SURVEY OF RECENT EXPOSURE STUDIES OF NONSMOKERS 4-15
4.3.1 Nonoccupational Exposures 4-16
4.3.1.1 Exposure to Carbon Monoxide from Motor Vehicles 4-16
4.3.1.2 Exposure to Carbon Monoxide in Recreational Vehicles 4-23
4.3.1.3 Residential Exposure to Carbon Monoxide 4-24
4.3.1.4 Exposure to Carbon Monoxide at Commercial Facilities 4-28
4.3.1.5 Studies of Breath Carbon Monoxide in Populations:
The Effects of Exposure to Carbon Monoxide 4-31
4.3.1.6 Nonoccupational Exposure to Methylene Chloride 4-34
4.3.2 Occupational Exposures 4-36
4.3.2.1 Exposures to Carbon Monoxide in the Workplace 4-36
4.3.2.2 Exposures to Methylene Chloride in the Workplace 4-37
October 1999 vii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
age
4.3.3 Activity Pattern Studies 4-39
4.3.3.1 Activity Patterns of California Residents 4-40
4.3.3.2 Activity Patterns of Children in Six States 4-43
4.3.3.3 A Comparative Study Between California and the Nation 4-43
4.3.3.4 An English Study 4-44
4.3.3.5 A Boston Study of Household Activities, Life Cycle, and
Role Allocation 4-44
4.3.3.6 The National Human Activity Pattern Survey 4-45
4.4 MAJOR FACTORS AFFECTING POPULATION EXPOSURE 4-45
4.4.1 Federal Policies Affecting Transportation and Air Quality
in Urban Areas 4-46
4.4.2 Federal and State Policies Affecting Temporal Trends in Exposure 4-48
4.4.2.1 Effects of Motor Vehicle Emission Standards on
Unintentional Death Rates 4-48
4.4.2.2 Effects of Motor Vehicle Emission Standards on Passenger
Cabin Exposure 4-50
4.4.3 Social Changes Affecting Human Activity Patterns 4-54
4.5 CONCLUSIONS 4-56
REFERENCES 4-60
5. PHARMACOKINETICS AND MECHANISMS OF ACTION OF
CARBON MONOXIDE 5-1
5.1 INTRODUCTION 5-1
5.2 ABSORPTION, DISTRIBUTION, AND PULMONARY ELIMINATION 5-2
5.2.1 Pulmonary Uptake 5-2
5.2.1.1 Mass Transfer of Carbon Monoxide 5-2
5.2.1.2 Effects of Dead Space and Ventilation/Perfusion Ratio 5-3
5.2.1.3 Lung Diffusion of Carbon Monoxide 5-4
5.2.2 Tissue Uptake 5-5
5.2.2.1 The Lung 5-5
5.2.2.2 The Blood 5-6
5.2.2.3 Heart and Skeletal Muscle 5-8
5.2.2.4 The Brain and Other Tissues 5-9
5.2.3 Pulmonary and Tissue Elimination 5-10
5.3 TISSUE PRODUCTION AND METABOLISM OF CARBON MONOXIDE .... 5-11
5.4 CONDITIONS AFFECTING CARBON MONOXIDE UPTAKE AND
ELIMINATION 5-14
5.4.1 Environment and Activity 5-14
5.4.2 Altitude 5-15
5.4.3 Physical Characteristics 5-17
5.4.4 Health Status 5-18
October 1999 viii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
age
5.5 MODELING CARBOXYHEMOGLOBIN FORMATION 5-20
5.5.1 The Coburn-Forster-Kane and Other Regression Models 5-20
5.5.1.1 Empirical Regression Models 5-20
5.5.1.2 Linear and Nonlinear Coburn-Forster-Kane Differential
Equations 5-21
5.5.1.3 Confirmation Studies of Coburn-Forster-Kane Models 5-23
5.5.1.4 Application of Coburn-Forster-Kane Models 5-25
5.6 INTRACELLULAR EFFECTS OF CARBON MONOXIDE 5-27
5.6.1 Introduction 5-27
5.6.2 Inhibition of Hemoprotein Function 5-28
5.6.3 Free Radical Production 5-29
5.6.4 Stimulation of Guanylate Cyclase 5-31
5.7 MECHANISMS OF CARBON MONOXIDE TOXICITY 5-32
5.7.1 Alterations in Blood Flow 5-32
5.7.2 Mitochondri al Dysfuncti on and Altered Producti on of High-Energy
Intermediates 5-33
5.7.3 Vascular Insults Associated with Exposure to Carbon Monoxide 5-34
5.8 OTHEREFFECTS OF CARBON MONOXIDE 5-36
5.9 SUMMARY 5-37
REFERENCES 5-39
6. HEALTH EFFECTS OF EXPOSURE TO AMBIENT CARBON MONOXIDE 6-1
6.1 INTRODUCTION 6-1
6.2 CARDIOVASCULAR EFFECTS 6-3
6.2.1 Epidemiologic Studies 6-3
6.2.1.1 Introduction 6-3
6.2.1.2 Ambient Carbon Monoxide and Exacerbation of
Heart Disease 6-10
6.2.1.3 Ambient Carbon Monoxide and Daily Mortality 6-32
6.2.1.4 Ambient Carbon Monoxide and Low Birth Weight 6-36
6.2.1.5 Ambient Carbon Monoxide and Frequency of
Respiratory Illness 6-38
6.2.2 Controlled Laboratory Studies 6-42
6.3 CENTRAL NERVOUS SYSTEM AND BEHAVIORAL EFFECTS 6-47
6.3.1 Brain Oxygen Metabolism 6-47
6.3.1.1 Whole Brain 6-47
6.3.1.2 Subregions of the Brain 6-48
6.3.2 Behavioral Effects of Carbon Monoxide 6-50
6.4 DEVELOPMENTAL TOXICITY 6-55
6.5 ACUTE PULMONARY EFFECTS 6-57
6.6 OTHER SYSTEMIC EFFECTS OF CARBON MONOXIDE 6-58
October 1999 ix DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
age
6.7 PHYSIOLOGIC RESPONSES TO CARBON MONOXIDE EXPOSURE 6-59
6.8 COMBINED EXPOSURE OF CARBON MONOXIDE WITH OTHER
POLLUTANTS, DRUGS, AND ENVIRONMENTAL FACTORS 6-60
6.8.1 High-Altitude Effects 6-60
6.8.2 Interaction with Drugs 6-61
6.8.3 Interaction with Other Air Pollutants and Environmental Factors 6-62
6.8.4 Tobacco Smoke 6-64
6.9 SUMMARY 6-65
REFERENCES 6-69
7. INTEGRATIVE SUMMARY AND CONCLUSIONS 7-1
7.1 INTRODUCTION 7-1
7.2 ENVIRONMENTAL SOURCES 7-2
7.3 ENVIRONMENTAL CONCENTRATIONS 7-3
7.4 CARBOXYHEMOGLOBIN LEVELS IN THE POPULATION 7-5
7.5 MECHANISMS OF CARBON MONOXIDE ACTIVITY 7-7
7.6 HEALTH EFFECTS OF CARBON MONOXIDE 7-8
7.7 SUBPOPULATIONS POTENTIALLY AT RISK FROM EXPOSURE
TO AMBIENT CARBON MONOXIDE 7-9
7.7.1 Age, Gender, and Pregnancy as Risk Factors 7-10
7.7.2 Preexisting Disease as a Risk Factor 7-11
7.7.2.1 Subjects with Heart Disease 7-12
1.122 Subjects with Other Vascular Diseases 7-12
7.7.2.3 Subjects with Anemia and Other Hematologic Disorders 7-14
7.7.2.4 Subjects with Obstructive Lung Disease 7-15
7.7.3 Subpopulations at Risk from Combined Exposure to Carbon Monoxide
and Other Chemical Substances 7-15
7.7.3.1 Interactions with Drugs 7-15
7.7.3.2 Interactions with Other Chemical Substances in the
Environment 7-16
7.7.4 Subpopulations Exposed to Carbon Monoxide at High Altitudes 7-16
7.8 CONCLUSIONS 7-17
REFERENCES 7-18
APPENDIX A: Abbreviations and Acronyms A-l
October 1999 x DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Number Page
1-1 National Ambient Air Quality Standards for Carbon Monoxide 1-3
2-1 Performance Specifications for Automated Analytical Methods for
Carbon Monoxide 2-4
2-2 Suggested Performance Specifications for Monitoring Carbon Monoxide
in Nonurban Environments 2-5
3-1 Summary of Major Sources and Sinks of Carbon Monoxide 3-2
3-2 Annual Global Carbon Monoxide Emissions Estimates 3-8
3-3 U.S. Nationwide Carbon Monoxide Emissions 3-17
3-4 Sites Not Meeting the 8-Hour Carbon Monoxide National Ambient Air Quality
Standard, 1993 to 1997 3-26
3-5 Running-Average Exceedances of the 9-ppm 8-Hour Carbon Monoxide
Standard, 1988, 1996, and 1997 3-32
3-6 Annual Orcadian Pattern of 8-Hour Average Carbon Monoxide Concentrations
Culminating in Values Greater Than 9.5 ppm in Lynwood and Hawthorne, CA,
During 1996 3-33
3-7 Ranges in Average Carbon Monoxide Emission Rates for Residential Sources ... 3-53
3-8 Sources of Carbon Monoxide in the Indoor Environment 3-55
3-9 Combustible Fuels in Homes in the United States in 1995 3-60
3-10 Prevalence of Gas Cooking Ranges 3-62
3-11 Residential Air Exchange Rates 3-64
3-12 Carbon Monoxide Descriptive Statistics for All Homes 3-68
3-13 Carbon Monoxide Concentrations in Smoking and Nonsmoking Areas
in Real-Life Situations 3-72
October 1999 xi DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number Page
4-1 Carbon Monoxide Concentrations in Selected Microenvironments of
Denver, CO, 1982 and 1983 4-9
4-2 Summary of Studies of In-Vehicle Exposure and Ambient Carbon Monoxide
Concentrations, 1965 and 1992 4-18
4-3 Mean Breath Carbon Monoxide Levels and Sample Sizes Across Smoking
Categories and Job Types 4-33
4-4 Studies of Occupational Exposures and Dosages 4-38
4-5 Time Spent in Different Microenvironments by Californians, 1987 to 1990 4-41
4-6 Percentage of Californians Who Use or Who Are in Proximity to Potential
Sources of Either Carbon Monoxide or Methylene Chloride on a Given Day,
1987 to 1990 4-42
4-7 Motor Vehicle Carbon Monoxide Emission Standards, Typical In-Vehicle
Carbon Monoxide Exposures, and Unintentional Carbon Monoxide-Related
Death Rates in the United States 4-49
4-8 Typical Net Mean Carbon Monoxide Concentration Ranges by Travel Mode
for Cities in Three Countries 4-53
6-1 Modeled Relative Risks of Interquartile-Range Increases in Ambient Pollutant
Concentrations, for Daily Heart Disease Admissions in Persons >65 Years Old,
Detroit, MI, 1986 to 1989 6-14
6-2 Modeled Effects of 5th- to 95th-Percentile Increments in Ambient Air Pollutant
Concentrations on Daily Numbers of Cardiac and Respiratory Hospital
Admissions in Single-Pollutant Models, by Season, Athens, Greece, 1998 6-15
6-3 Acute Myocardial Infarction: One and Two Pollutant Models with Cool and
Warm Season, London, UK, April 1, 1987 to March 31, 1994 6-21
6-4 Modeled Percentage Increases in Hospital Admissions for Heart Disease,
Associated with Interquartile Range Increases in Ambient PM10 and CO Levels,
in Eight Locations, Across These Locations, and in Tucson, AZ, 1988 to 1990 . . . 6-27
October 1999 xii DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number Page
6-5 Modeled Percentage Increases in Hospitalizations at Mean Pollutant
Concentrations in Single-Pollutant and Multi-Pollutant Models,
Toronto, Canada, 1980 to 1994 6-31
6-6 Estimated Lowest-Observed-Effect Levels for Exposure of Laboratory
Animals to Carbon Monoxide 6-46
6-7 Key Health Effects of Exposure to Ambient Carbon Monoxide 6-66
7-1 Predicted Carbon Monoxide Exposures in the Population 7-7
October 1999 xiii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Number Page
2-1 Schematic diagram of gas filter correlation monitor for carbon monoxide 2-15
2-2 The correlation between an end-tidal breath carbon monoxide concentration
after a 10-second breathhold and blood carboxyhemoglobin levels 2-25
3-1 Latitudinal and seasonal variability in carbon monoxide concentrations obtained
in the National Oceanic and Atmospheric Administration Climate Monitoring
Diagnostics Laboratory monitoring network 3-4
3-2 Global background average carbon monoxide levels and growth rates for global
background average 3-6
3-3 Locations of sites in the nationwide ambient carbon monoxide monitoring
network, 1997 3-25
3-4 Nationwide composite average of the annual second highest 8-hour carbon
monoxide concentrations, 1978 to 1997 3-28
3-5 Variability in the annual second highest 8-hour carbon monoxide concentrations
across all sites in the United States, 1988 to 1997 3-29
3-6 Composite average of the annual second highest 8-hour carbon monoxide
concentrations for rural, suburban, and urban sites, 1988 to 1997 3-30
3-7 Diurnal variation of nationwide composite hourly average carbon monoxide
concentrations for winter, 1987 to 1996 3-31
3-8 Map of Denver showing locations of carbon monoxide monitoring sites 3-36
3-9 Average diurnal variation in carbon monoxide at the Denver-Broadway site
for weekdays during the winter season 3-37
3-10 Monthly average diurnal variation in carbon monoxide at the
Denver-Broadway site for weekdays from May 1986 through May 1987 3-37
3-11 Monthly average diurnal variation in carbon monoxide at the
Denver-Broadway site for weekdays from May 1995 through May 1996 3-38
3-12 Central tendency statistics for the daily 8-hour maximum carbon monoxide
concentration at the Denver-Broadway site during the winter season from
1986 to 1995 3-38
October 1999 xv DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number Page
3-13 Map of Los Angeles showing locations of carbon monoxide monitoring sites .... 3-39
3-14 Average diurnal variation in carbon monoxide at the Los Angeles-Lynwood
site for weekdays during the winter season 3-40
3-15 Central tendency statistics for the daily 8-hour maximum carbon monoxide
concentration at the Los Angeles-Hawthorne site during the winter season
from 1986 to 1995 3-41
3-16 Central tendency statistics for the daily 8-hour maximum carbon monoxide
concentration at the Los Angeles-Barstow site during the winter season from
1986 to 1995 3-42
3-17 Average diurnal variation in carbon monoxide at the Los Angeles-Hawthorne
site for weekdays during the winter season 3-43
3-18 Average diurnal variation in carbon monoxide at the Los Angeles-El Toro
site for weekdays during the winter season 3-43
3-19 Monthly average diurnal variation in carbon monoxide at the
Los Angeles-Anaheim site for weekdays from May through May 1986 to
1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 3-44
3-20 Map of New York showing locations of carbon monoxide monitoring sites 3-45
3-21 Monthly average diurnal variation in carbon monoxide at the
New York-Flatbush site for weekdays from May through May 1986 to
1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 3-46
3-22 Average diurnal variation in carbon monoxide at the New York-Manhattan
site for weekdays during the winter season 3-46
3-23 Monthly average diurnal variation in carbon monoxide at the
New York-Manhattan site for weekdays from May through May 1986 to
1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 3-47
3-24 Monthly average diurnal variation in carbon monoxide at the
New York-Morristown, NJ, site for weekdays from May through May
1986 to 1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 3-47
3-25 Map of Phoenix showing locations of carbon monoxide monitoring sites 3-48
October 1999 xvi DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number Page
3-26 Monthly average diurnal variation in carbon monoxide at the
Phoenix-Central site for weekdays from May through May 1986 to
1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 3-49
3-27 Central tendency statistics for the daily 8-hour maximum carbon monoxide
concentration at the Phoenix-East Butler site during the winter season from
1986 to 1995 3-49
3-28 Monthly average diurnal variation in carbon monoxide at the Phoenix-West
site for weekdays from May through May 1986 to 1987, 1989 to 1990,
1992 to 1993, and 1995 to 1996 3-50
3-29 Monthly average diurnal variation in carbon monoxide at the Phoenix-South
site for weekdays from May through May 1986 to 1987, 1989 to 1990,
1992 to 1993, and 1995 to 1996 3-50
3-30 Percentage of U.S. households using unvented combustion heaters, by type of
fuel, stratified by region 3-63
3-31 Modeled indoor carbon monoxide concentration distributions in houses with
only one indoor combustion pollutant source 3-66
3-32 Arithmetic mean carbon monoxide concentrations by presence or absence of
combustion source 3-69
4-1 Conceptual overview of the probabilistic National Ambient Air Quality
Standards Exposure Model 4-12
4-2 Observed versus simulated 8-hour daily maximum exposure for persons
residing in homes with gas stoves in Denver, CO 4-13
4-3 Observed versus simulated 8-hour daily maximum exposure for persons
residing in homes without gas stoves in Denver, CO 4-14
4-4 Log-probability plot of the maximum 1-hour indoor minus outdoor carbon
monoxide concentrations based on a random sample of 277 homes that
used gas appliances in California, 1991 and 1992 4-26
October 1999 xvii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number Page
4-5 Log-probability plot of the maximum 8-hour indoor minus outdoor carbon
monoxide concentrations based on a random sample of 277 homes that
used gas appliances in California, 1991 and 1992 4-27
4-6 Excess carbon monoxide concentrations in the exhaled air of nonsmoking
control subjects, untreated asthmatics, and treated asthmatics 4-35
4-7 Trends in ambient carbon monoxide concentrations and in-vehicle carbon
monoxide exposures, 1965 to 1992 4-51
5-1 Diagrammatic presentation of carbon monoxide uptake and elimination
pathways and carbon monoxide body stores 5-4
5-2 Oxyhemoglobin dissociation curve of normal human blood, of blood containing
50% carboxyhemoglobin, and of blood with only 50% hemoglobin because of
anemia 5-7
5-3 Plot of fractional sensitivities of selected variables versus time of exposure 5-23
6-1 Lowess smoothing of the association between ambient levels of CO and hospital
admissions for congestive heart failure (CHF) among elderly people after
adjustment for temperature, month, day of week, and year, 1986 through
1989 6-12
6-2 Nonparametric smoothed curve of adjusted hospitalization rates for congestive
heart failure in the elderly plotted against daily 1-h maximum CO levels in
10 Canadian cities, 1981 to 1991 6-18
6-3 The relative risk associated with the exposure percentiles of carbon monoxide
at specific temperature strata, based on results of the multi-pollutant and
single-pollutant generalized linear models of hospital admissions for heart
failure among the elderly in Chicago, IL, 1986 to 1989 6-24
6-4 The effect of carbon monoxide exposure on time to onset of angina 6-44
6-5 The relationship between carboxyhemoglobin and the cerebral metabolic rate
for oxygen 6-49
6-6 The relationship between carboxyhemoglobin and behavior 6-52
October 1999 xviii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number Page
6-7 The relationship between carboxyhemoglobin and behavior: effects in
humans compared to rats 6-53
7-1 Predicted carboxyhemoglobin levels resulting from 1- and 8-h exposures
to carbon monoxide at rest and with light exercise are based on the
Coburn-Forster-Kane equation 7-6
7-2 Percentage breakdown of deaths from cardiovascular diseases in the
United States 7-13
7-3 Estimated prevalence of cardiovascular disease by age and sex for the
United States, 1988 to 1991 7-14
October 1999 xix DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
CHAPTER 1. INTRODUCTION
Principal Author
Mr. James A. Raub—Project Manager and Coordinator for Heath Effects
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors
Dr. Robert S. Chapman—Coordinator for Epidemiology Studies
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly M. Comfort—Coordinator for Indoor Air Emissions and Concentrations
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William G. Ewald—Coordinator for Measurement Methods
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David T. Mage—Coordinator for Population Exposure
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—Coordinator for Atmospheric Chemistry, Sources, and Emissions
National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Mr. C. Barnett—New York Healthy Schools Network, 96 South Swan Street
Albany, NY 12210
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
October 1999 xxi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Dr. David J. McKee—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Mr. Harvey M. Richmond—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 2. ANALYTICAL METHODS FOR MONITORING CARBON MONOXIDE
Principal Authors
Dr. Russell R. Dickerson—Department of Meteorology, The University of Maryland
College Park, MD 20742
Dr. David T. Mage—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors
Mr. William G. Ewald—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lance Wallace—National Exposure Research Laboratory
U.S. Environmental Protection Agency, Reston, VA 22092
Reviewers
Dr. Michael G. Apte—Indoor Environment Department
Lawrence Berkeley National Laboratory, Berkeley, CA 94720
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935
October 1999 xxii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Dr. Thomas E. Dahms—Department of Anesthesiology, School of Medicine
St. Louis University Medical Center, St. Louis, MO 63110
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Milan J. Hazucha—Department of Medicine
Center for Environmental Medicine and Lung Biology, The University of North Carolina
Chapel Hill, NC 27599
Dr. Michael T. Kleinman—Department of Community and Environmental Medicine
California College of Medicine, University of California, Irvine, CA 92697
Mr. Leon Langan—Langan Products, Inc., 2660 California Street, San Francisco, CA 94115
Dr. William A. McClenny—National Exposure Research Laboratory (MD-44)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Dr. Leonard Newman—Environmental Chemistry Division
Brookhaven National Laboratory, Upton, NY 11973
Dr. Paul Roberts—Sonoma Technology, Inc., Petaluma, CA 94954
CHAPTER 3. SOURCES, EMISSIONS, AND CONCENTRATIONS OF
CARBON MONOXIDE IN AMBIENT AND INDOOR AIR
Principal Authors
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly M. Comfort—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Michael P. Zelenka—National Exposure Research Laboratory (MD-56)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
October 1999 xxiii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors
Mr. Warren P. Freas—Office of Air Quality Planning and Standards (MD-14)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Alan H. Huber—National Exposure Research Laboratory (MD-56)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Dr. Michael G. Apte—Indoor Environment Department
Lawrence Berkeley National Laboratory, Berkeley, CA 94720
Dr. Irwin H. Billick—WEC Consulting, Ltd., Potomac, MD 20854
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Mr. Tom Chappie—State of Alaska, Department of Environmental Conservation
555 Cordova Street, Anchorage, AK 99501
Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Peter G. Flachsbart—Department of Urban and Regional Planning
University of Hawaii at Manoa, Honolulu, HI 96822
Dr. Lawrence J. Folinsbee—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Mary Good—Municipality of Anchorage, Department of Health and Human Services
825 L Street, P.O. Box 196650, Anchorage, AK 99519
Mr. Hank Hove—Fairbanks North Star Borough, 809 Pioneer Road, P.O. Box 71267
Fairbanks, AK 99707
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
Dr. Kai-Shen Liu—Environmental Health Laboratory
California Department of Health Services, Berkeley, CA 94704
October 1999 xxiv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Mr. Thomas R. McCurdy—National Exposure Research Laboratory (MD-56)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Robert Morris—Department of Family Medicine
Tufts University School of Medicine, Boston, MA 02111
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Dr. Leonard Newman—Environmental Chemistry Division
Brookhaven National Laboratory, Upton, NY 11973
Dr. Paul Roberts—Sonoma Technology, Inc., Petaluma, CA 94954
Dr. Jed Waldman—California Department of Health Services, Berkeley, CA 94704
CHAPTER 4. POPULATION EXPOSURE TO CARBON MONOXIDE
Principal Author
Dr. Peter G. Flachsbart—Department of Urban and Regional Planning
University of Hawaii at Manoa, Honolulu, HI 96822
Contributor
Dr. David T. Mage—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Dr. Michael G. Apte—Indoor Environment Department
Lawrence Berkeley National Laboratory, Berkeley, CA 94720
Mr. C. Barnett—New York Healthy Schools Network, 96 South Swan Street
Albany, NY 12210
Dr. Irwin H. Billick—WEC Consulting, Ltd., Potomac, MD 20854
October 1999 xxv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Milan J. Hazucha—Department of Medicine
Center for Environmental Medicine and Lung Biology, The University of North Carolina
Chapel Hill, NC 27599
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
Dr. Michael T. Kleinman—Department of Community and Environmental Medicine
California College of Medicine, University of California, Irvine, CA 92697
Dr. Kai-Shen Liu—Environmental Health Laboratory
California Department of Health Services, Berkeley, CA 94704
Mr. Thomas R. McCurdy—National Exposure Research Laboratory (MD-56)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Robert Morris—Department of Family Medicine
Tufts University School of Medicine, Boston, MA 02111
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Mr. Harvey M. Richmond—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jed Waldman—California Department of Health Services, Berkeley, CA 94704
October 1999 xxvi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 5. PHARMACOKINETICS AND MECHANISMS OF
ACTION OF CARBON MONOXIDE
Principal Authors
Dr. Milan J. Hazucha—Department of Medicine
Center for Environmental Medicine and Lung Biology, The University of North Carolina
Chapel Hill, NC 27599
Dr. Stephen R. Thorn—Institute for Environmental Medicine and
Department of Emergency Medicine, University of Pennsylvania, Philadelphia, PA 19104
Reviewers
Mr. C. Barnett—New York Healthy Schools Network, 96 South Swan Street
Albany, NY 12210
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935
Dr. Thomas E. Dahms—Department of Anesthesiology
School of Medicine, St. Louis University Medical Center, St. Louis, MO 63110
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
Dr. Michael T. Kleinman—Department of Community and Environmental Medicine
California College of Medicine, University of California, Irvine, CA 92697
Dr. James J. McGrath—Department of Physiology, School of Medicine
Texas Tech University Health Sciences Center, Lubbock, TX 79430
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Dr. Peter Tikuisis—Defence and Civil Institute of Environmental Medicine
1133 Sheppard Avenue, West, Toronto, Ontario, M3M 3B9, Canada
October 1999 xxvii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 6. HEALTH EFFECTS OF EXPOSURE TO AMBIENT CARBON MONOXIDE
Principal Authors
Mr. James A. Raub—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Robert S. Chapman—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Mr. C. Barnett—New York Healthy Schools Network, 96 South Swan Street
Albany, NY 12210
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935
Dr. Thomas E. Dahms—Department of Anesthesiology, School of Medicine
St. Louis University Medical Center, St. Louis, MO 63110
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Lawrence J. Folinsbee—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Milan J. Hazucha—Department of Medicine
Center for Environmental Medicine and Lung Biology, The University of North Carolina
Chapel Hill, NC 27599
Dr. Jon M. Heuss—Air Improvement Resources, Inc., 7355 Rickett Drive
Washington, MI 48094
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
Dr. Michael T. Kleinman—Department of Community and Environmental Medicine
California College of Medicine, University of California, Irvine, CA 92697
October 1999 xxviii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Dr. Victor G. Laties—Environmental Medicine, University of Rochester Medical Center
School of Medicine and Dentistry, Rochester, NY 14642
Dr. James J. McGrath—Department of Physiology, School of Medicine
Texas Tech University Health Sciences Center, Lubbock, TX 79430
Dr. Robert Morris—Department of Family Medicine
Tufts University School of Medicine, Boston, MA 02111
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
CHAPTER 7. INTEGRATIVE SUMMARY AND CONCLUSIONS
Principal Author
Mr. James A. Raub—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors
Dr. Robert S. Chapman—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly M. Comfort—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William G. Ewald—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David T. Mage—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
October 1999 xxix DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Reviewers
Dr. Michael G. Apte—Indoor Environment Department
Lawrence Berkeley National Laboratory, Berkeley, CA 94720
Ms. Kelly M. Brown—Ford Motor Company, 17225 Federal Drive, Suite 145
Allen Park, MI 48101
Mr. Albert Donnay—MCS Referral & Resources, Inc., 508 Westgate Rd., Baltimore, MD 21229
Dr. Lawrence J. Folinsbee—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Sandra E. Inkster—U.S. Consumer Product Safety Commission, 4330 East West Highway
Room 600, Bethesda, MD 02814
Dr. James J. McGrath—Department of Physiology, School of Medicine
Texas Tech University Health Sciences Center, Lubbock, TX 79430
Ms. Nancy W. Newkirk—American Petroleum Institute, 1220 L Street, NW
Washington, DC 20005
Dr. Stephen R. Thorn—Institute for Environmental Medicine and
Department of Emergency Medicine, University of Pennsylvania, Philadelphia, PA 19104
Dr. Vanessa Vu—National Center for Environmental Assessment (860ID)
U.S. Environmental Protection Agency, 401 M St. SW, Washington, DC 20460
Dr. Jed Waldman—California Department of Health Services, Berkeley, CA 94704
October 1999 xxx DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Science Advisory Board
Clean Air Scientific Advisory Committee
Carbon Monoxide Review Panel
Chair
Dr. Joe Mauderly—Director of External Affairs, Senior Scientist, and Director of
National Environmental Respiratory Center, Lovelace Respiratory Research Institute
Albuquerque, NM 87108
Members
Mr. John Elston—Administrator, Office of Air Quality Management, State of New Jersey
Department of Environmental Protection and Energy, Trenton, NJ 08625
Dr. Philip K. Hopke—R.A. Plane Professor of Chemistry, Clarkson University
Potsdam, NY 13699
Dr. Eva J. Pell—Steimer Professor of Agriculture Sciences, Buckhout Laboratory
The Pennsylvania State University, University Park, PA 16802
Dr. Arthur C. Upton—Director, Independent Peer Review, CRESP
Environmental and Occupational Health Sciences Institute, Piscataway, NJ 08854
Dr. Sverre Vedal—Professor of Medicine, Vancouver General Hospital
Vancouver, BC Canada V57 3J5
Dr. Warren White—Senior Research Associate, Chemistry Department, Washington University
St. Louis, MO 63130
Consultants
Dr. Stephen M. Ayres—Director, International Health Programs
Virginia Commonwealth University/Medical College of Virginia, Richmond, VA 23284
Dr. Thomas E. Dahms—Professor and Director, Anesthesiology Research
Department of Anesthesiology, St. Louis University School of Medicine, St. Louis, MO 63110
Dr. Victor G. Laties—Professor Emeritus, Department of Environmental Medicine
University of Rochester Medical Center, Rochester, NY 14642
October 1999 xxxi DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Science Advisory Board
Clean Air Scientific Advisory Committee
Carbon Monoxide Review Panel
(cont'd)
Dr. Brian Leaderer—Professor, Division of Environmental Health Sciences
Yale University School of Medicine, New Haven, CT 06519
Dr. Lawrence D. Longo—Professor, School of Medicine, Department of Physiology
Center for Perinatal Biology, Departments of Physiology and Gynecology and Obstetrics
School of Medicine, Loma Linda University, Loma Linda, CA 92354
Designated Federal Official
Mr. Robert Flaak—Designated Federal Officer and Team Leader, Committee Operations Staff
U.S. Environmental Protection Agency, Science Advisory Board (1400), Washington, DC 20460
Staff Assistant
Ms. Diana Pozun—Management Assistant, U.S. Environmental Protection Agency
Science Advisory Board (1400), Washington, DC 20460
October 1999 xxxii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Project Team for Development of
Air Quality Criteria for Carbon Monoxide
Scientific Staff
Mr. James A. Raub—Project Manager and Coordinator for Health Effects
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—Coordinator for Atmospheric Chemistry, Sources, and Emissions
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Dr. Robert S. Chapman—Coordinator for Epidemiology Studies
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Mr. William G. Ewald—Coordinator for Measurement Methods
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Ms. Beverly M. Comfort—Coordinator for Indoor Air Emissions and Concentrations
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Dr. David T. Mage—Coordinator for Population Exposure
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Ms. Ellie Speh—Office Manager, Environmental Media Assessment Group
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Technical Support Staff
Mr. Douglas B. Fennell—Technical Information Specialist
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Ms. Diane H. Ray—Technical Information Manager
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
Mr. Richard N. Wilson—Clerk
National Center for Environmental Assessment (MD-52), Research Triangle Park, NC 27711
October 1999 xxxiii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. Environmental Protection Agency
Project Team for Development of
Air Quality Criteria for Carbon Monoxide
(cont'd)
Document Production Staff
Mr. John R. Barton—Document Processing Coordinator
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Ms. Yvonne Harrison—Word Processor
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Ms. Bettye Kirkland—Word Processor
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Mr. David E. Leonhard—Graphic Artist
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Ms. Carolyn T. Perry—Word Processor
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Ms. Veda E. Williams—Graphic Artist
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Technical Reference Staff
Mr. R. David Belton—Reference Specialist
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Mr. John Bennett—Technical Information Specialist
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Mr. William Hardman—Reference Retrieval and Database Entry Clerk
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Ms. Sandra L. Hughey—Technical Information Specialist
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
Mr. Jian Ping Yu—Reference Retrieval and Database Entry Clerk
OAO Corporation, Chapel Hill-Nelson Highway, Beta Building, Suite 210, Durham, NC 27713
October 1999 xxxiv DRAFT-DO NOT QUOTE OR CITE
-------
i EXECUTIVE SUMMARY
2
3 Air Quality Criteria for Carbon Monoxide
4 External Review Draft
5
6 The purpose of this document is to present air quality criteria for carbon monoxide (CO), in
7 accordance with Sections 108 and 109 of the Clean Air Act (CAA), that reflect the latest
8 scientific information useful in indicating the kind and extent of all identifiable effects on public
9 health and welfare that may be expected from the presence of CO in ambient air. This document
10 is an update of Air Quality Criteria for Carbon Monoxide, published by the U.S. Environmental
11 Protection Agency (EPA) in 1991, and will be used as the scientific basis for reevaluating the
12 current national ambient air quality standards (NAAQS) for CO. This executive summary
13 concisely summarizes key findings from the present document.
14
15
16 Summary Findings
17
18 Monitoring
19
20 Reliable methods are identified in Chapter 2 for monitoring CO concentrations in ambient
21 air to determine compliance with the NAAQS and the potential effects on overall air quality and
22 for monitoring the impact of ambient CO exposure on human populations.
23
24 • Several adequate techniques exist for highly reliable monitoring of CO to ensure compliance
25 with the NAAQS. The most reliable method for continuous measurement of CO in ambient air
26 is the nondispersive infrared (NDIR) optical transmission technique, the technique on which
27 the EPA-designated analytical reference methods are based. One category of NDIR monitors,
28 the gas filter correlation monitor, is still the single most widely used analyzer for fixed-site
29 monitoring stations.
30
31 • Determining CO levels at many nonurban locations requires substantially better performance
32 than that required to demonstrate compliance with the NAAQS. Commercial CO-monitoring
33 instruments, sometimes with minor modifications, can meet the measurement needs for
34 supplying useful data on the distribution and trends of ambient CO and for modeling
35 photochemical smog in places where ambient levels are significantly below the NAAQS.
36
37 • There are at this time commonly used and accepted procedures for generating CO measurement
38 standards that are accurate to better than ±2% in the parts-per-million range and about ±10% in
39 the range of concentrations found in the clean troposphere. Several CO measurement
40 techniques have been intercompared and found reliable.
41
42 • Several electrochemical and passive sampling methods are available. These techniques are
43 currently not equivalent to compliance monitoring methods but are useful for personal
44 exposure studies and for measuring CO concentrations in indoor, outdoor, and in transit
45 microenvironments.
October 1999 E-l DRAFT-DO NOT QUOTE OR CITE
-------
1 • Blood carboxyhemoglobin (COHb) level and CO concentration in exhaled breath are
2 biological indicators of CO exposure. Although the use of optical methods (e.g., CO-Oximetry
3 [CO-Ox]) is common for population sampling and clinical analyses of COHb, gas
4 chromatography is the method of choice for measuring low COHb saturations (<5%) that are
5 expected to occur with ambient CO exposures. The measurement of CO in exhaled breath has
6 practical advantages for population exposure sampling but has a greater potential for error in
7 the estimation of COHb than a direct measurement of COHb.
8
9 Global Tropospheric Chemistry
10
11 Current information about the abundance and distribution, the nature of sources and sinks,
12 and the chemistry of CO in environments ranging from the global background to indoor air is
13 summarized in Chapter 3. The importance of CO for atmospheric chemistry also is discussed in
14 this chapter.
15
16 • In nonurban areas, tropospheric CO has a significant role in affecting the oxidizing capacity of
17 the earth's atmosphere. Reaction with CO is a principal process by which hydroxyl radicals
18 are removed from the atmosphere. Reaction with hydroxyl radicals is also the primary process
19 for removing many other man-made and natural compounds, including CO, from the
20 atmosphere.
21
22 • Carbon monoxide is linked closely to the cycle of tropospheric ozone and may be responsible
23 for 20 to 40% of the ozone formed in nonurban areas. Ozone is an oxidant, a greenhouse gas,
24 and a precursor of hydroxyl radicals. On balance, if CO increases, the net effect is to decrease
25 hydroxyl radicals.
26
27 • Carbon monoxide is, therefore, an intermediary in determining the future concentrations of
28 many environmentally important trace gases. The future of methane, a greenhouse gas, cannot
29 be evaluated adequately or predicted without an accurate understanding of the global CO
30 budget, which is not presently available. Similarly, predicting future concentrations of other
31 environmentally important gases, such as the hydrochlorofluorocarbons that can deplete
32 stratospheric ozone, depends on how well we understand the CO budget.
33
34 • Global background CO concentrations average about 120 and 40 ppb in remote marine areas of
35 the Northern and Southern Hemispheres, respectively, that are not affected by local sources.
36 Results from flask and in situ monitoring stations show no discernible trend in CO levels since
37 1993.
38
39 • The average lifetime of CO in the atmosphere is about 2 mo, longer at high latitudes and
40 shorter at low latitudes.
41
42 • In addition to direct emissions from fossil fuel and biomass burning, CO is produced in the
43 atmosphere by the photochemical oxidation of anthropogenic and biogenic hydrocarbons.
44 Because of uncertainties in reaction kinetics, the identification of reaction products, and the
45 effects of heterogeneous processes, the accuracy of estimates of photochemical sources of CO
46 are limited.
October 1999 E-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The global emissions of CO are about 2.3 x 109 metric tons per year, amounting to an annual
2 source of about 1.0 x io9 metric tons of carbon in the atmosphere, compared with a global
3 anthropogenic input of 7.1 x IO9 metric tons per year of carbon in carbon dioxide. Estimates of
4 individual CO sources are uncertain by a factor of two or more; however, the total production
5 of CO is known to within 25%, based on its estimated rate of destruction because of reactions
6 with hydroxyl radicals.
7
8 • Emissions from various sources in developing countries are likely to be very significant but are
9 not known at present.
10
11 Regional and Urban Air Quality
12
13 Emissions, concentrations, and effects of CO on air quality within the United States are
14 discussed in Chapter 3 and its appendix.
15
16 • Carbon monoxide plays an important role in atmospheric photochemistry in regional and urban
17 environments. In urban areas, CO either can produce or destroy ozone, depending on the
18 concentrations of nitrogen oxides and hydrocarbons. In numerical simulations of several urban
19 air sheds, CO was found to be responsible for production of 10 to 20% of the ozone.
20
21 • The nationwide average annual second highest 8-h ambient CO concentration decreased from
22 lOppmin 1978 to 4 ppm in 1997.
23
24 • On- and nonroad mobile sources account for approximately 80% of the 1997 nationwide
25 emissions inventory for CO. Declines in ambient CO levels in the United States follow
26 approximately the decline in motor vehicle emissions of CO. However, the relative
27 importance of nonroad sources has increased over the past decade from 12.7% of total
28 emissions in 1988 to 19.2% in 1997.
29
30 • There were 41 exceedances of the 8-h NAAQS for CO at 12 U.S. monitoring sites in 1997.
31 These sites, in descending order, were located in Calexico and Los Angeles-Long Beach, CA;
32 Fairbanks, AK; Steubenville, OH; El Paso, TX; and Phoenix, AZ.
33
34 • Carbon monoxide levels in the geographically diverse metropolitan statistical areas (MSAs) of
35 Denver, CO; Los Angeles, CA; New York, NY; and Phoenix, AZ, have decreased from 1986
36 through 1995. However, the nature of the diurnal and seasonal variations has remained
37 essentially the same. These variations result largely from the interaction among motor vehicle
38 emissions, traffic patterns, and meteorological parameters, such as wind speed and mixing
39 height. [An analysis of air quality in Fairbanks, AK, is in progress.]
40
41 • In general, the spatial distribution of CO within these four air sheds was highly heterogeneous.
42 For instance, correlations between time series of daily 8-h average maximum CO
43 concentrations at different monitoring sites in the Denver, New York, and Los Angeles MSAs
44 ranged from about 0 to 0.7, whereas corresponding correlations ranged from about 0.3 to 0.7 in
45 the Phoenix MSA.
46
October 1999 E-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 Indoor Air Quality
2
3 Indoor CO exposure may represent a significant portion of the total human exposure to CO.
4 The sources, emissions, and concentrations of CO found in indoor microenvironments also are
5 discussed in Chapter 3.
6
7 • Carbon monoxide occurs indoors directly through emissions from various indoor combustion
8 sources or indirectly as a result of infiltration or ventilation from outdoor sources. In the
9 absence of indoor sources, average CO concentrations generally will equal those in the
10 surrounding ambient environment.
11
12 • Emissions of CO from the use of adequately vented combustion appliances (e.g., gas and oil
13 furnaces, gas water heaters, gas dryers) will not contaminate indoor air unless the units or
14 venting systems are malfunctioning.
15
16 • The major sources of CO in residential microenvironments are tobacco smoke, vehicle startup
17 and idling in attached garages, and unvented or partially vented combustion appliances.
18 Factors affecting emissions of CO from the use of combustion appliances in the home include
19 the type of source (e.g., gas cooking stoves, unvented space heaters, wood stoves, fireplaces),
20 appliance design, type of fuel used, fuel consumption rate, and source operating condition.
21 Carbon monoxide concentrations in the indoor environment will vary based on the source
22 emission rate, use pattern, ambient CO concentration, air exchange rate, building volume, and
23 air mixing within the indoor compartments.
24
25 • Carbon monoxide emissions from gas stoves depend on their use pattern, operating condition,
26 and fuel consumption rate. Ranges with standing pilot lights emit more CO than ranges with
27 electronic pilot lights. The contribution of gas cooking stoves to CO concentrations in the
28 indoor environment is expected to be negligible because of the intermittent nature of the
29 stoves' use, unless gas stoves are used as a heat source.
30
31 • Carbon monoxide emissions from unvented space heaters vary as a function of unit design and
32 operating condition, type of fuel used and consumption rate, air currents near the space heater,
33 and use pattern. Carbon monoxide concentrations in environments using space heaters depend
34 on the type of space heater, emission rate, air exchange/infiltration rate, and frequency and
35 duration of use. Reported indoor CO concentrations are higher in homes using unvented space
36 heaters as the primary source of heat.
37
38 • Wood stoves and fireplaces emit CO during fire start-ups and maintenance, through leaks in
39 the stove or venting system, and from back drafting. Carbon monoxide emissions are higher
40 during the first stage of a fire because of increased fuel usage and lower combustion
41 temperatures.
42
43 • Carbon monoxide emissions from tobacco smoke depend on the type of tobacco product
44 (e-g-, cigarette, cigar) and the degree to which tobacco is actively smoked. Concentrations of
45 CO from the use of tobacco products will exceed background concentrations, but will vary
46 based on differences in ventilation, the number of cigarettes or cigars smoked, and the smoking
October 1999 E-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 rate. For example, it is possible for cigar smokers to raise indoor CO concentrations to more
2 than 9 ppm above ambient levels measured outside.
3
4 Population Exposure
5
6 The reduction in automotive emissions brought about by the CAA have reduced in-traffic
7 CO exposures and traffic-related ambient CO concentrations well below those measured in the
8 past decade. Chapter 4 describes the reduction in human exposures to CO over the past decade,
9 brought about by the reduction in automotive emissions of CO. However, people are still
10 exposed to CO at concentration levels above the NAAQS in areas of high traffic density and in
11 indoor locations where tobacco is smoked and where combustion devices (e.g., stoves, heaters)
12 are not adequately vented.
13
14 • Fixed site monitors often are used in urban areas to measure the ambient concentrations to
15 which individuals in the surrounding areas may be exposed. These measurements tend to
16 overestimate 8-h exposure values for people living in areas of lower traffic and underestimate
17 the exposure of people living in areas of higher traffic.
18
19 • Neighborhood scale, fixed-site ambient CO monitoring may provide a reasonable estimate of
20 the average CO exposures for some people who are not exposed to tobacco smoke or other
21 sources of CO in their homes and occupations.
22
23 • Nonsmokers exposed to tobacco smoke, heavy traffic fumes, and indoor sources of CO will
24 have higher body burdens of CO (COFtb) than would be predicted from ambient data alone.
25
26 • Emission reductions in CO mandated by the CAA amendments have led to significant
27 reductions in ambient CO concentrations and lower traffic-related exposures to CO from motor
28 vehicle exhaust, suggesting that estimates of current population exposure based on pre-1990
29 exposure studies may no longer apply. There currently is not a good estimate of CO exposure
30 distribution for the population.
31
32 • Personal CO exposures that exceed the level of the NAAQS will still occur in some
33 nonsmokers exposed to sources of CO not controlled by the CAA (e.g., recreational vehicles,
34 garages, poorly vented or malfunctioning indoor combustion sources) or exposed in their
35 occupations or hobbies to CO or to organic solvents that are metabolized to CO (e.g.,
36 methylene chloride).
37
38 • Modern CO exposure models adequately predict the average general population exposure but
39 still underpredict high CO exposures, indicating that further work is required to understand the
40 activities and emissions associated with these higher exposures.
41
42 Pharmacokinetics and Mechanisms of Action
43
44 The action of CO in the body and the factors influencing its uptake, distribution to vital
45 tissues, and elimination, provide the foundation for measuring or predicting effects on organ
October 1999 E-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 function. In Chapter 5, the basic principles of CO pharmacokinetics are reviewed, and the
2 possible mechanisms for pathophysiologic effects at the cellular level are discussed.
3
4 • The most prominent pathophysiological effect of CO is hypoxemia caused by the binding of
5 CO to hemoglobin. The formation of COHb reduces the oxygen-carrying capacity of blood
6 and impairs release of oxygen from red blood cells to tissues. The brain and heart are
7 especially sensitive to CO-induced hypoxia and cytotoxicity because these tissues have the
8 highest resting oxygen requirements.
9
10 • Carbon monoxide is produced endogenously through heme degradation, metabolic processes of
11 drugs, and degradation of unsaturated fatty acids, inhaled solvents and other xenobiotics. High
12 altitude and many disorders, especially anemias and inflammatory lung diseases, also may
13 increase endogenous levels of CO.
14
15 • The amount of COHb formed from exogenous exposure is dependent on the CO concentration
16 and duration of exposure, minute ventilation, lung diffusion capacity, and ambient pressure, as
17 well as the health status and metabolism of the exposed individual. The formation of COHb is
18 reversible, but, because of a small blood-to-air CO pressure gradient and tight binding of CO to
19 Hb, the elimination half-time is quite long, varying from 2 to 6.5 h.
20
21 • The physical and physiological variables affecting the rate of COHb formation and elimination
22 have been integrated into empirical and mathematical models for estimating COHb levels from
23 different conditions of exposure. The nonlinear Coburn-Forster-Kane equation is the most
24 widely used predictive model of COHb formation and is still considered the best all-around
25 model for COHb prediction.
26
27 • Intracellular binding of CO to hemoproteins, particularly myoglobin (Mb) found in heart and
28 skeletal muscle, would be favored under conditions of low intracellular oxygen tension as
29 COHb levels rise. The impact of ambient CO on intracellular CO uptake by Mb is not well
30 understood.
31
32 • New investigations have expanded on the physiological effects of CO in two areas. First, there
33 is a growing recognition that CO may play a role in normal neurotransmission and vasomotor
34 control. Second, there also is increased interest in the ability of CO to cause free-radical-
35 mediated changes in tissues. The impact of ambient CO on these processes and the role they
36 may have in pathophysiology is not well understood.
37
38 Health Effects
39
40 Concerns about the potential health effects of exposure to CO are addressed in Chapter 6 by
41 examining the published results of extensive controlled-exposure studies and more limited
42 population-exposure studies. Emphasis is placed on the current understanding of quantifiable
43 health effects that are likely to occur in humans at the low COHb levels (<5 %) that are predicted
44 to result from typical ambient CO exposures.
45
October 1999 E-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Blood COHb levels are the best known indicators of potential health risk; however, the lowest-
2 observed-effect levels depend on the method used for analysis. Gas chromatography (GC)
3 generally is regarded as more accurate than CO-Ox for measuring low (<5%) COHb
4 saturations.
5
6 • Maximal exercise duration and performance in healthy individuals has been shown to be
7 reduced at COHb levels of >2.3 and >4.3% (GC), respectively. Performance decrements are
8 small, however, and likely to affect only competing athletes. No effects were observed during
9 submaximal exercise in healthy individuals at COHb levels as high as 15 to 20%.
10
11 • Decreased exercise tolerance has been observed consistently in patients with coronary artery
12 disease (CAD) and reproducible exercise-induced angina (chest pain) at COHb levels of 3 to
13 6% COHb (CO-Ox). The indicators of myocardial ischemia during exercise, such as
14 electrocardiographic changes and associated chest pain, were statistically significant in one
15 large multicenter clinical study at >2.4% COHb (GC) and showed a dose-response relationship
16 with increasing COHb.
17
18 • An increase in the number and complexity of exercise-related arrhythmias (irregular heart
19 beats) has been observed at >6% COHb (CO-Ox) in some people with CAD and a high level
20 of baseline ectopy (a chronic arrhythmia) that may present an increased risk of sudden death.
21
22 • In epidemiologic studies, daily fluctuations in ambient CO concentration have been associated
23 consistently with fluctuations in heart disease exacerbation. These associations have been
24 observed at ambient CO levels at or below the current CO standards. The influence on these
25 associations of ambient CO exposure, relative to CO exposure from nonambient sources, has
26 not been determined.
27
28 • Epidemiologic studies also suggest associations of short-term ambient CO exposure with
29 nonaccidental daily mortality, the great majority of which occurs in people at least 65 years of
30 age. As above, the relative influences on these associations of ambient and nonambient CO
31 have not been determined and the possibility that CO is acting as a marker for other
32 combustion-related pollutants cannot be ruled out.
33
34 • Laboratory animal studies indicate that acute CO poisoning can affect the growth and function
35 of the developing fetus. Epidemiologic studies show a limited association of subchronic
36 ambient CO exposure with low birth weight; however, these studies are not conclusive.
37
38 • Recent analyses indicate that significant behavioral impairments in healthy individuals should
39 not be expected until COHb levels exceed 20%; however, mild central nervous system effects
40 have been reported in the historical CO literature at COHb levels between 5 and 20%.
41
42 • Ambient levels of CO are not known to have any direct effects on lung tissue. Observed
43 epidemiologic associations of short-term ambient CO levels with daily respiratory illness
44 frequency cannot yet be interpreted with confidence.
45
October 1999 E-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Carbon monoxide has the potential to interact with other stressors. These include (1) visitation
2 to high altitudes, especially for patients with CAD; (2) use of psychoactive drugs or alcohol;
3 (3) use of specific medications, especially nitric oxide and calcium channel blockers;
4 (4) prolonged exposure to heat; and (5) exposure to other pollutants.
5
6 Subpopulations Potentially At Risk
1
8 On the basis of monitored ambient CO concentrations and quantifiable CO concentration-
9 response relationships for health effects demonstrated in humans, the following conclusions are
10 made in Chapter 7 regarding subpopulations potentially at risk from exposure to ambient CO.
11
12 • Young, healthy nonsmokers are not at immediate risk from ambient CO exposure because only
13 limitations at maximal exercise performance have been demonstrated at low COHb levels
14 (<5%) that are predicted to result from ambient exposures. Effects have not been demonstrated
15 on healthy individuals performing submaximal exercise that is more typical of daily human
16 activity.
17
18 • Patients with reproducible exercise-induced angina (chest pain) are a sensitive group within the
19 general population that is at increased risk of experiencing decreased exercise tolerance
20 because of exacerbation of cardiovascular symptoms at ambient or near-ambient CO-exposure
21 concentrations that result in COHb levels of 2.4% (GC) or higher.
22
October 1999 E-8 DRAFT-DO NOT QUOTE OR CITE
-------
i 1. INTRODUCTION
2
3
4 This document is an update of Air Quality Criteria for Carbon Monoxide, published by the
5 U.S. Environmental Protection Agency (EPA) in 1991, and will serve as the basis for
6 reevaluating the current National Ambient Air Quality Standards (NAAQS) for carbon monoxide
7 (CO) set in 1994. Carbon monoxide is one of six ubiquitous ambient air pollutants covered by
8 the Federal Clean Air Act (CAA) requiring an assessment of the latest scientific knowledge as a
9 requisite step in the development of standards to protect public health and welfare. The present
10 document is not intended as a complete and detailed literature review, but it does summarize
11 relevant key information from the previous 1991 document and evaluates new information
12 relevant to the CO NAAQS criteria development, based on pertinent published literature
13 available through early 1999.
14 Carbon monoxide, a trace constituent of the troposphere, is produced by both natural
15 processes and human activities. Because plants can both metabolize and produce CO, trace
16 levels are considered a normal constituent of the natural environment. Although ambient
17 concentrations of CO in the vicinity of urban and industrial areas can exceed global background
18 levels, there are no reports of these currently measured levels of CO producing any adverse
19 effects on plants or microorganisms. Ambient concentrations of CO, however, may be
20 detrimental to human health and welfare, depending on the levels that occur in areas where
21 humans live and work and on the susceptibility of exposed individuals to potentially adverse
22 effects.
23 This chapter presents a brief summary of the legislative and regulatory history of the
24 CO NAAQS and the rationale for the existing standards and gives an overview of the issues,
25 methods, and procedures utilized in the preparation of the present document.
26
27
28 1.1 LEGISLATIVE REQUIREMENTS
29 Two sections of the CAA govern the establishment, review, and revision of NAAQS.
30 Section 108 (U.S. Code, 1991) directs the Administrator of EPA to identify and issue air quality
October 1999 1-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 criteria for pollutants that may reasonably be anticipated to endanger public health or welfare.
2 These air quality criteria are to reflect the latest scientific information useful in indicating the
3 kind and extent of all identifiable effects on public health or welfare that may be expected from
4 the presence of the pollutant in ambient air.
5 Section 109(a) of the CAA (U.S. Code, 1991) directs the Administrator of EPA to propose
6 and promulgate primary and secondary NAAQS for pollutants identified under Section 108.
7 Section 109(b)(l) defines a primary standard as one that the attainment and maintenance of
8 which, in the judgment of the Administrator, based on the criteria and allowing for an adequate
9 margin of safety, is requisite to protect the public health. The secondary standard, as defined in
10 Section 109(b)(2), must specify a level of air quality that the attainment and maintenance of
11 which, in the judgment of the Administrator, based on the criteria, is requisite to protect the
12 public welfare from any known or anticipated adverse effects associated with the presence of the
13 pollutant in ambient air.
14 Section 109(d) of the CAA (U.S. Code, 1991) requires periodic review and, if appropriate,
15 revision of existing criteria and standards. If, in the Administrator's judgment, EPA's review and
16 revision of criteria make appropriate the proposal of new or revised standards, such standards are
17 to be revised and promulgated in accordance with Section 109(b). Alternatively, the
18 Administrator may find that revision of the standards is inappropriate and conclude the review by
19 leaving the existing standards unchanged.
20
21
22 1.2 REGULATORY BACKGROUND
23 On April 30, 1971, EPA promulgated identical primary and secondary NAAQS for CO at
24 levels of 10 mg/m3 (9 ppm) for an 8-h average and 40 mg/m3 (35 ppm) for a 1-h average, not to
25 be exceeded more than once per year. The scientific basis for the primary standard, as described
26 in the first criteria document (National Air Pollution Control Administration, 1970), was a study
27 suggesting that low levels of CO exposure resulting in carboxyhemoglobin (COHb)
28 concentrations of 2 to 3% were associated with neurobehavioral effects in exposed subjects
29 (Beard and Wertheim, 1967).
30 In accordance with Sections 108 and 109 of the CAA, EPA periodically has reviewed and
31 revised the criteria on which the existing NAAQS for CO (Table 1) are based. On August 18,
October 1999 1 -2 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 1. NATIONAL AMBIENT AIR QUALITY STANDARDS
FOR CARBON MONOXIDE
Date of Promulgation
August 1, 1994
Primary NAAQS
9ppma(10mg/m3)
35ppma(40mg/m3)
Averaging Time
8-hb
l-hb
al ppm = 1.145 mg/m3, 1 mg/m3 = 0.873 ppmat25 °C, 760 mmHg.
bNot to be exceeded more than once per year.
Source: Federal Register (1994).
1 1980, EPA proposed certain changes in the standards based on scientific evidence reported in the
2 revised criteria document for CO (U.S. Environmental Protection Agency, 1979). Such evidence
3 indicated that the Beard and Wertheim (1967) study no longer should be considered as a sound
4 scientific basis for the standard. Additional medical evidence accumulated since 1970, however,
5 indicated that aggravation of angina pectoris and other cardiovascular diseases would occur at
6 COHb levels as low as 2.7 to 2.9%. On August 18, 1980, EPA proposed changes to the standard
7 (Federal Register, 1980) based on the findings of the revised criteria. The proposed changes
8 included (1) retaining the 8-h primary standard level of 9 ppm, (2) revising the 1-h primary
9 standard level from 35 ppm to 25 ppm, (3) revoking the existing secondary CO standards
10 (because no adverse welfare effects have been reported at or near ambient CO levels),
11 (4) changing the form of the primary standards from deterministic to statistical, and (5) adopting
12 a daily interpretation for exceedances of the primary standards, so that exceedances would be
13 determined on the basis of the number of days on which the 8- or 1-h average concentrations are
14 above the standard levels.
15 The 1980 proposal was based in part on health studies conducted by Dr. Wilbert Aronow.
16 In March 1983, EPA learned that the Food and Drug Administration (FDA) had raised serious
17 questions regarding the technical adequacy of several studies conducted by Dr. Aronow on
18 experimental drugs, leading FDA to reject use of the Aronow drug study data. Therefore, EPA
19 convened an expert committee to examine the Aronow CO studies before any final decisions
20 were made on the NAAQS for CO. In its report (Horvath et al., 1983), the committee concluded
October 1999 1 -3 DRAFT-DO NOT QUOTE OR CITE
-------
1 that EPA should not rely on Dr. Aronow' s data because of concerns regarding the research that
2 substantially limited the validity and usefulness of the results.
3 An addendum to the 1979 criteria document for CO (U.S. Environmental Protection
4 Agency, 1984) reevaluated the scientific data concerning health effects associated with exposure
5 to CO at or near ambient exposure levels in light of the committee recommendations and taking
6 into account findings reported subsequent to those previously reviewed. On September 13, 1985,
7 EPA issued a final notice (Federal Register, 1985) announcing retention of the existing primary
8 NAAQS for CO and rescinding the secondary NAAQS for CO.
9 The criteria review process was initiated again on July 22, 1987, and notice of availability
10 of the revised draft criteria document was published in the Federal Register (Federal Register,
11 1990) on April 19, 1990. This draft document included discussion of several new studies of
12 effects of CO on angina patients that had been initiated in light of the controversy discussed
13 above. The Clean Air Scientific Advisory Committee (CASAC) reviewed the draft criteria
14 document at a public meeting held on April 30, 1991. The EPA carefully considered comments
15 received from the public and from CAS AC in preparing the final criteria document (U.S.
16 Environmental Protection Agency, 1991). On July 17, 1991, CASAC sent to the EPA
17 Administrator a "closure letter" outlining key issues and recommendations and indicating that the
18 document provided a scientifically balanced and defensible summary of the available knowledge
19 of effects of CO. A revised "staff paper" based on the scientific evidence was released for public
20 review in February 1992, followed by two CASAC review meetings held on March 5 and on
21 April 28, 1992. The CASAC came to closure on the final staff paper (U.S. Environmental
22 Protection Agency, 1992) in a letter to the Administrator dated August 11, 1992, indicating that it
23 provided a scientifically adequate basis for EPA to make a regulatory decision on the appropriate
24 primary NAAQS for CO. On August 1, 1994, EPA issued a final decision (Federal Register,
25 1994) that revisions of the NAAQS for CO were not appropriate at that time.
26 In keeping with the requirements of the CAA, EPA's National Center for Environmental
27 Assessment has begun to review and once again revise the criteria for CO.
28
29
30
October 1999 1 -4 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3 RATIONALE FOR THE EXISTING CARBON MONOXIDE
2 STANDARDS
3 The following discussion describing the bases for the existing CO NAAQS set in 1994 has
4 been excerpted and adapted from "National Ambient Air Quality Standards for Carbon
5 Monoxide—Final Decision" (Federal Register, 1994). The discussion includes the rationale for
6 selection of the level and averaging time for the NAAQS that would be protective of adverse
7 effects in the most sensitive subpopulation and EPA's assessment that led to a decision not to
8 revise the existing standards for CO.
9
10 1.3.1 Carboxyhemoglobin Levels of Concern
11 In selecting the appropriate level and averaging time for the primary NAAQS for CO, the
12 EPA Administrator must first determine the COHb levels of concern, taking into account a large
13 and diverse health effects database. Based on the assessments provided in the criteria document
14 (U.S. Environmental Protection Agency, 1991) and in the staff paper (U.S. Environmental
15 Protection Agency, 1992), judgments were made to identify the most useful studies for
16 establishing a range of COHb levels to be considered for standard setting. In addition, the more
17 uncertain or less quantifiable evidence was reviewed to determine the lower end of the range that
18 would provide an adequate margin of safety from effects of clear concern. The following
19 discussion summarizes the most critical considerations for the Administrator's 1994 decision on
20 the CO NAAQS.
21 The Administrator of EPA concluded that cardiovascular effects, as measured by decreased
22 time to onset of angina pain and by decreased time to onset of significant electrocardiogram
23 (ECG) ST-segment depression, were the health effects of greatest concern to be clearly
24 associated with CO exposures at levels observed in the ambient air. These effects were
25 demonstrated in angina patients at postexposure COHb levels that were elevated to 2.9 to 5.9%
26 (CO-Oximetry [CO-Ox] measurement), representing incremental increases of 1.5 to 4.4% from
27 baseline levels. Time to onset of significant ECG ST-segment change, which is indicative of
28 myocardial ischemia in patients with documented coronary artery disease and a more objective
29 indicator of ischemia than angina pain, provided supportive evidence of health effects occurring
30 at exposures as low as 2.9 to 3.0% COHb (CO-Ox). The clinical importance of cardiovascular
31 effects associated with exposures to CO resulting in COHb levels less than 2.9% remains less
October 1999 1 -5 DRAFT-DO NOT QUOTE OR CITE
-------
1 certain and was considered only in evaluating whether the current CO standards provide an
2 adequate margin of safety.
3 The Administrator of EPA also considered the following factors in evaluating the adequacy
4 of the current CO NAAQS.
5 • Short-term reductions in maximal work capacity were measured in trained athletes
6 exposed to CO sufficient to produce COHb levels as low as 2.3%.
7 • The wide range of human susceptibility to CO exposures and ethical considerations in
8 selecting subjects for experimental purposes, taken together, suggest that the most
9 sensitive individuals have not been studied.
10 • Animal studies of developmental toxicity and human studies of the effects of maternal
11 smoking provide evidence that exposures to high concentrations of CO can be
12 detrimental to fetal development, although little is known about the effects of ambient
13 CO exposures on the developing human fetus.
14 • Although little is known about the effects of CO on other potentially sensitive
15 populations besides those with coronary artery disease, there is reason for concern about
16 visitors to high altitudes, individuals with anemia or respiratory disease, and the elderly.
17 • Impairment of visual perception, sensorimotor performance, vigilance, and other central
18 nervous system effects have not been demonstrated to be caused by CO concentrations
19 commonly found in ambient air; however, short-term peak CO exposures may be
20 responsible for impairments that could be a matter of concern for complex activities such
21 as automobile driving.
22 • Limited evidence suggests concern for individuals exposed to CO concurrently with drug
23 use (e.g., alcohol), heat stress, or coexposure to other pollutants.
24 • Large uncertainties remain regarding modeling COHb formation and estimating human
25 exposure to CO that could lead to over- or underestimation of COHb levels associated
26 with attainment of a given CO NAAQS in the population.
27 • Measurement of COHb made using the CO-Ox technique may not reflect the COHb
28 levels in angina patients studied, thereby creating uncertainty in establishing a lowest
29 effects level for CO.
30 The Administrator concluded that the lowest COHb level at which adverse effects have been
31 demonstrated in persons with angina is around 2.9 to 3.0%, representing an increase of 1.5%
October 1999 1 -6 DRAFT-DO NOT QUOTE OR CITE
-------
1 COHb above baseline when using the CO-Ox to measure COHb. These data serve to establish
2 the upper end of the range of COHb levels of concern. Taking into account the above data
3 uncertainties, the less significant health endpoints, and less quantifiable data on other potentially
4 sensitive groups, the lower end of the range was established at 2.0% COHb.
5
6 1.3.2 Relationship Between Carbon Monoxide Exposure and
7 Carboxyhemoglobin Levels
8 In order to set ambient CO standards based on an assessment of health effects at various
9 COHb levels, it is necessary to estimate the ambient CO concentrations that are likely to result in
10 COHb levels of concern. The best all-around model for predicting COHb levels is the Coburn,
11 Foster, Kane (CFK) differential equation (U.S. Environmental Protection Agency, 1991).
12 Baseline estimates of COHb levels expected to be reached by nonsmokers exposed to various
13 constant concentrations of CO can be determined by the CFK equation (U.S. Environmental
14 Protection Agency, 1992). There are, however, two major uncertainties involved in estimating
15 COHb levels resulting from exposure to CO concentrations. First, the large distribution of
16 physiological parameters used in the CFK equation across the population of interest is sufficient
17 to produce noticeable deviations in the COHb levels. Second, predictions based on exposure to
18 constant CO concentrations can under- or overestimate responses of individuals exposed to
19 widely fluctuating CO levels that typically occur in the ambient environment.
20
21 1.3.3 Estimating Population Exposure
22 The EPA review included an analysis of CO exposures expected to be experienced by
23 residents of Denver, CO, under air quality scenarios where the 8-h NAAQS is just attained.
24 Although the exposure analysis included passive smoking and gas stove CO emissions as indoor
25 sources of CO, it did not include other sources that may be of concern to high-risk groups (e.g.,
26 lawn equipment, wood stoves, fireplaces, faulty furnaces). The analysis indicated that, at the 8-h
27 standard, fewer than 0.1% of the nonsmoking cardiovascular-disease population would
28 experience a COHb level >2.1% (U.S. Environmental Protection Agency, 1992). A smaller
29 population was estimated to exceed higher COHb percentages.
30
31
October 1999 1 -7 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3.4 Decision on the Primary Standards
2 Based on the exposure analysis results described above, the Administrator of EPA
3 concluded that relatively few people of the cardiovascular sensitive population group analyzed
4 would experience COHb levels >2.1% when exposed to CO levels in the absence of indoor
5 sources when the current ambient standards were attained. Although indoor sources of CO may
6 be of concern to high-risk groups, their contribution cannot be effectively mitigated by ambient
7 air quality standards.
8 The Administrator of EPA also determined that both the 1-h and 8-h averaging times for
9 CO were valid because the 1-h standard provided reasonable protection from health effects that
10 might be encountered from very short duration peak (acute) exposures in the urban environment,
11 and the 8-h standard provided a good indicator for tracking continuous exposures that occur
12 during any 24-h period. The Administrator concurred with staff recommendations (U.S.
13 Environmental Protection Agency, 1992) that both averaging times be retained for the primary
14 CO standards.
15 For these reasons, the EPA Administrator determined under CAA Section 109(d)(l) that
16 revisions to the current 1-h (35 ppm) and 8-h (9 ppm) primary standards for CO were not
17 appropriate at that time (Federal Register, 1994).
18
19
20 1.4 ISSUES OF CONCERN FOR THE CURRENT CRITERIA
21 DEVELOPMENT
22 The following is a brief summary of scientific issues that are addressed in the revised air
23 quality criteria document for CO. These issues are based on findings presented at symposia and
24 workshops that were convened to assess the current state of understanding of the sources,
25 atmospheric cycle, and health effects of CO and revised, as appropriate, by peer review
26 comments received on earlier draft chapters of the criteria document.
27
28 1.4.1 Sources and Emissions
29 Detailed descriptions of the processes forming CO during combustion were presented in the
30 previous CO document. These descriptions have been reviewed for accuracy in the revised
31 document; however, a good deal of uncertainty exists regarding the correct values for CO
October 1999 1 -8 DRAFT-DO NOT QUOTE OR CITE
-------
1 emissions from transportation sources. Emissions from transportation have been revised upward
2 in the current emissions and trends report (U.S. Environmental Protection Agency, 1996) from
3 those used in the previous document. Emissions estimates for CO from various sources are
4 highly uncertain, especially those for transportation sources. The potential of relatively new
5 techniques (e.g., inverse modeling) for testing and improving emissions estimates needs to be
6 evaluated.
7
8 1.4.2 Atmospheric Chemistry
9 Much of the material discussed in the previous criteria document is already available in
10 standard textbooks and does not need to be reviewed. New information, however, is needed in
11 this current review regarding the chemistry of CO formation from the oxidation of methane and
12 nonmethane hydrocarbons (NMHCs). For example, the fractional yields of CO resulting from
13 the oxidation of NMHCs, especially isoprene and monoterpenes, need to be established. The
14 importance of CO for ozone formation in the urban and nonurban atmosphere also needs to be
15 highlighted.
16 There are a number of ways to express the amount of a substance in the atmosphere.
17 Perhaps the most commonly used measure is concentration, which is the amount, or mass, of a
18 substance in a given volume divided by that volume (e.g., mol/m3 in SI units). Often in the
19 literature, however, quantities of gaseous substances are expressed as volume mixing ratios, such
20 as parts per million (ppm) or parts per billion (ppb). These terms are technically not
21 "concentrations", but rather refer properly to the molar mixing ratio of a substance (equivalent to
22 volumetric mixing ratio for an ideal gas), which is the ratio of the concentration of a substance in
23 mol/m3 to the concentration in mol/m3 of all gaseous components in a given air volume (Seinfeld
24 and Pandis, 1998). Thus, mixing ratio is a mole fraction that in SI units should be expressed as
25 //mol/mol for ppm, and nmol/mol for ppb. Throughout this document, however, mixing ratio
26 will be referred to as concentration of CO in ppm or ppb because these terms have been
27 extensively referred to in the human exposure, lexicological, and epidemiological literature and
28 as the basis for CO compliance monitoring for the NAAQS.
29
30
31
October 1999 1 -9 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.4.3 Global Cycle
2 Global trends in tropospheric CO concentrations declined from about 1988 to 1993 after
3 several years of annual increases, as determined by different networks of surface observations.
4 Carbon monoxide levels apparently have stabilized since 1993. The reasons for the changes in
5 CO trends still need to be determined.
6
7 1.4.4 Measurement Technology
8 The discussion on measurement methods for CO in the previous document has been
9 reviewed, older methods have been removed, and newer methods for monitoring CO from
10 various environmental sources are presented.
11
12 1.4.5 Ambient Air Quality
13 Because of the everchanging nature of atmospheric concentrations, levels in various
14 environments (rural, urban, and suburban) have been reanalyzed for different regions of the
15 United States. The temporal variability of CO levels from daily to seasonal time scales also has
16 been characterized. Relations between urban concentrations of CO and regional and global
17 background levels also are examined, as well as background levels of CO for use in different
18 applications.
19
20 1.4.6 Indoor Emissions and Concentrations
21 Indoor concentrations of CO are a function of outdoor concentrations, indoor sources,
22 infiltration, ventilation, and air mixing. In the absence of indoor sources, concentrations of CO
23 in the indoor environment are similar to those in ambient air; however, personal CO exposure
24 studies have shown that CO concentrations in excess of 9 ppm can occur in certain indoor and
25 in-transit microenvironments associated with transportation sources that are not considered part
26 of the ambient air. Unvented, improperly installed, or poorly maintained combustion appliances,
27 downdrafts during unstable weather conditions, and depressurization from the operation of
28 exhaust systems and fireplaces also may contribute to potentially high CO concentrations
29 indoors. Further research is still needed, however, to determine the contribution of nonambient
30 sources to total human exposure to CO.
October 1999 1-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.4.7 Exposure Assessment
2 Compliance with the NAAQS is determined by measurements taken at fixed-site, ambient
3 monitors, yet exposure monitoring in the field and modeling studies indicate that individual
4 personal exposure does not correlate directly with CO concentrations determined by the
5 fixed-site monitors alone. This is because of the mobility of people and the spatial and temporal
6 variability of CO concentrations across a given area. The nature of differences between fixed-
7 site and personal monitoring results should be given greater attention, especially in regard to
8 interpreting the results of epidemiology studies.
9 Data from population field studies can be used to construct and test models of human
10 exposure that account for time and activity patterns known to affect exposure to CO. New
11 information from field monitoring studies needs to be incorporated into exposure models to
12 better capture the observed personal exposure distributions, including the higher exposures found
13 in the tail of exposure distribution.
14 A unique feature of CO exposure is that the dose an individual receives can be estimated by
15 measuring COHb. The reader should note, however, that such exposure estimates are affected by
16 the time interval between peak CO exposure and blood sampling and by the use of any
17 supplemental oxygen therapy. It also has been shown that the method chosen for measurement
18 of COHb can be a source of considerable error, particularly at the low end of the CO dissociation
19 curve, where COHb levels are <5%. The sensitivity of COHb measurement techniques will,
20 therefore, have an influence on the lowest-observed-effect level for CO. Gas chromatography
21 (GC) is regarded as more accurate than CO-Ox for measuring the lower COHb levels.
22
23 1.4.8 Mechanisms of Action
24 The principle cause of CO toxicity is tissue hypoxia caused by CO binding to hemoglobin.
25 Secondary mechanisms related to intracellular uptake of CO have been the focus of recent
26 research. Current knowledge summarized in this document suggests that the most likely protein
27 other than hemoglobin to be inhibited functionally at relevant levels of COHb is myoglobin,
28 found in heart and skeletal muscle. The extent of effects caused by CO molecules in solution
29 needs to be evaluated in relation to typical ambient CO exposures in the population. Other
30 mechanisms of interest, which have not yet been demonstrated to occur at ambient CO levels, are
October 1999 1-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 cytotoxic effects (e.g., vasomotor control, free oxygen radicals) independent of impaired oxygen
2 delivery.
3
4 1.4.9 Health Effects
5 There are many published studies on acute experimental and accidental exposures to CO;
6 however, there is not enough reliable information on chronic exposures to low concentrations
7 from either ambient population-exposure studies or from occupational studies. Further work is
8 needed, therefore, to determine potential long-term exposures in the population and to develop
9 reliable dose-response relationships for at-risk groups. This information currently is missing
10 from the published literature. Some of the issues associated with acute CO exposures are
11 discussed below.
12
13 Cardiovascular Effects
14 Maximal exercise duration is reduced in young, healthy, nonsmoking individuals at COHb
15 levels as low as 2.3% (GC), but this effect is small and would be of concern mainly for
16 competing athletes. Clinical studies on subjects with reproducible exercise-induced angina have
17 confirmed that adverse effects occur with postexposure COHb levels as low as 2.4% (GC).
18 Thus, aggravation of coronary artery disease continues to provide the best scientific basis in
19 support for the current (9-ppm, 8-h and 35-ppm, 1-h) NAAQS for CO. More recent
20 epidemiology studies in the United States, Canada, and Europe have suggested that day-to-day
21 variations in ambient CO concentrations are related to cardiovascular hospital admissions and
22 daily mortality, especially for individuals over 65 years of age. It is not clear, however, if the
23 observed association results from CO or from combustion-related particles or, perhaps, from
24 some other, unmeasured pollutant exposure that covaries in time with CO.
25
26 Cerebrovascular Effects
27 Carbon monoxide hypoxia increases cerebral blood flow in healthy subjects, even at very
28 low exposure levels. Behaviors that require sustained attention or performance are most
29 sensitive to levels of COHb >5%. Disease or injury that impairs compensatory increases in
30 blood flow may increase the probability of effects, but little is known about the susceptibility of
31 compromised individuals to ambient levels of CO. Accidental exposures to high-level CO have
October 1999 1-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 been shown to cause neurological problems weeks after recovery from the acute episode. It is
2 not known, however, if these late neurological sequelae, described as intellectual deterioration;
3 memory impairment; and cerebral, cerebellar, and mid-brain damage, result from long-term
4 exposure to low ambient levels of CO.
5
6 Developmental Toxicity
1 Relatively high CO exposures of 150 to 200 ppm during gestation, leading to
8 approximately 15 to 25% COHb, produce reductions in birth weight, cardiomegaly, delays in
9 behavioral development, and disruption in cognitive function in newborn laboratory animals of
10 several species. Little data exist on humans exposed to CO for predicting a lowest-observed-
11 effect level for developmental effects. Studies relating human CO exposures from ambient
12 sources or cigarette smoking to reduced birth weight are of concern because of the risk for
13 developmental disorders; however, many of these studies have not considered all sources of CO
14 and may be confounded by other variables (e.g., smoke components, maternal behavior,
15 nutrition, genetics). Nevertheless, some health professionals have considered this evidence
16 sufficient to identify pregnant women, and the developing fetus, as at risk to ambient levels of
17 CO.
18
19 High-Altitude Effects
20 There are relatively few reports on the effects of inhaling CO at high altitudes. Current
21 knowledge supports the possibility that the effects of hypoxic hypoxia and CO-hypoxia are at
22 least additive. The potential additive effects of CO exposure in sensitive individuals visiting at
23 high altitudes need to be considered.
24
25 1.4.10 Carbon Monoxide Interaction with Drugs
26 There remains little direct information on the possible enhancement of CO toxicity by
27 concomitant illegal and prescription drug use or abuse; however, there are some data on
28 psychoactive drugs that suggest cause for concern.
29
30
31
October 1999 1-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.4.11 Subpopulations at Risk
2 On the basis of known effects described, heart disease patients with reproducible exercise-
3 induced angina appear to be best established as a sensitive group within the general population
4 that is at increased risk for experiencing health effects of concern at ambient or near-ambient CO
5 exposure concentrations resulting in COHb levels <5%. Certain other groups are at potential risk
6 from exposure to CO, but further research is required to specify health effects associated with
7 ambient or near-ambient CO exposures in these probable risk groups.
8
9
10 1.5 METHODS AND PROCEDURES FOR DOCUMENT PREPARATION
11 The procedures that were followed for developing the revised criteria document for CO are
12 different from those that have been used for recent criteria documents. For example, the previous
13 CO criteria document (U.S. Environmental Protection Agency, 1991) was a more comprehensive
14 scientific review of available information on the nature, sources, distribution, measurement, and
15 concentrations of CO in the environment and on the known and anticipated health effects that
16 CO would have on at-risk population groups. In lieu of a comprehensive review of the literature,
17 emphasis in the present criteria document has been placed on the development of a concise
18 summary of key information and a more interpretative discussion of the new scientific and
19 technological data available since the previous criteria were evaluated. The resulting document
20 is more of an update, in accordance with the recommendations made by CASAC.
21 The main focus of this revised criteria document is on the evaluation and interpretation of
22 more recent air quality, human exposure, and health effects issues. Therefore, the techniques
23 used to present this information vary according to the state-of-science for the respective topics.
24 For example, the analysis of ambient air quality is based on newly obtained air monitoring data
25 and utilizes the previous analysis only for showing trends over time. As a result of the relatively
26 dramatic decrease in ambient CO concentrations, population exposure to ambient CO also has
27 declined. Human exposure studies conducted in the early 1980s and earlier distributions of
28 COHb levels in the U.S. population that were relied on heavily in the previous assessment are no
29 longer relevant to the current picture of ambient CO exposure in the 1990s. Thus, key
30 information on population exposure must focus on the newer studies and on modeling results.
October 1999 1-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 On the other hand, the health effects literature on CO has remained relatively static since the
2 previous 1991 assessment, except for provocative publications on cellular mechanisms of CO
3 action and on epidemiologic associations of ambient CO with mortality and morbidity in the
4 elderly population. Newly published studies on most of the other health outcomes reconfirm the
5 conclusions made in the last document and are incorporated into the previous summaries by
6 reference only.
7 One of the first steps used in the development of this revised document was to convene
8 symposia or workshops to review the key scientific issues and to focus on the selection of
9 material that could be included in the document as the basis for the development of
10 standard-setting criteria. Both EPA and non-EPA scientific experts were utilized for this effort.
11 An interdisciplinary scientific symposium was held in Portland, OR, in December 1997, to
12 assess current scientific understanding of the atmospheric cycle of CO, including its sources,
13 sinks, and distribution. The three main subject areas covered in the symposium relate to the
14 distribution and spatial and temporal variability of CO, the atmospheric budget of CO, and direct
15 or indirect effects of CO on human health. Results from papers presented at the symposium are
16 included, by reference, in this revised criteria document.
17 A mini workshop, jointly organized by EPA, the Gas Research Institute, and the Health
18 Effects Institute, was convened in Chicago, IL, on April 24 and 25, 1998, to provide expert
19 scientific discussion on the public health significance of exposures to low levels (<50 ppm) of
20 CO. The three main topics covered in the meeting were human exposure patterns and trends in
21 CO exposure, pharmacokinetics and mechanisms of action of CO, and health effects.
22 A summary of the discussion by participants and conclusions drawn from the meeting were used
23 by authors in preparation of the draft criteria document chapters.
24 Finally, a public peer-review workshop was convened on September 17 and 18, 1998, to
25 define key issues, to review early drafts of the criteria document chapters, and to ascertain and
26 discuss any pertinent new literature. The respective authors of the draft chapters or sections of
27 the document revised them on the basis of the workshop recommendations. The revised chapters
28 of the document were incorporated into the first external review draft released for public
29 comment and reviewed by CAS AC on June 9, 1999. Necessary revisions were made in response
30 to public comments and CASAC recommendations before this Second External Review Draft of
31 the criteria document was released.
October 1999 1-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.6 ORGANIZATION AND CONTENT OF THE DOCUMENT
2 The updated air quality criteria document for CO critically evaluates and assesses scientific
3 information on air quality, human exposure, and health effects associated with exposure to the
4 concentrations found in the environment. Emphasis has been placed on the development of a
5 concise review of key information and a more interpretative discussion of the new scientific and
6 technological data available since completion of the previous criteria document (U.S.
7 Environmental Protection Agency, 1991). The references cited in the document should be
8 reflective of the state of knowledge through early 1999 on those issues most relevant to review of
9 the NAAQS for CO.
10 To aid in the concise development of this document, summaries of the relevant published
11 literature and selective discussion of the literature has been undertaken. Studies that were
12 presented in the previous criteria document and whose data were judged to be significant because
13 of their usefulness in deriving the current NAAQS are discussed briefly in the text. The reader,
14 however, is mainly referred back to the more extensive discussion of these "key" studies in the
15 previous document. Other, older studies are discussed in the text if they are open to
16 reinterpretation because of newer data, or potentially useful in deriving revised standards for CO.
17 Generally, only published information that has undergone scientific peer review has been
18 included in the revised criteria document. However, some newer studies not yet published in the
19 open literature but meeting high standards of scientific reporting have been included for a few
20 areas.
21 The structure of the present document follows the general outline of the previous criteria
22 document (U.S. Environmental Protection Agency, 1991), especially for topics that have changed
23 little since the last criteria review. The resulting sequence of discussion should help the reader to
24 find and contrast similar sections. There are, however, a few exceptions where some topics have
25 been consolidated into a single chapter in order to present a more concise document. The
26 executive summary at the beginning of the document provides a concise presentation of key
27 information and conclusions from all subsequent chapters.
28 The document begins with this introduction (Chapter 1), which provides the regulatory
29 history of CO and an understanding of the scientific basis for the current CO NAAQS.
30 Information on analytical methods for monitoring CO (Chapter 2) covers the measurement of CO
31 in ambient (outdoor) and indoor air, as well as methods for measuring breath CO and blood CO
October 1999 1-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 levels in exposed individuals. Chapter 3 provides information on the atmospheric chemistry of
2 CO and typical sources, emissions, and concentrations found in the ambient and indoor
3 environments, topics addressed in separate chapters of the previous document. The remaining
4 chapters are similar to the previous document, covering topics on population exposure to CO
5 (Chapter 4), pharmacokinetics and mechanisms of action (Chapter 5), and health effects
6 (Chapter 6). The last chapter (Chapter 7) provides an overall integrative summary of key
7 findings and an evaluation of subpopulations potentially at risk from exposure to CO.
October 1999 1-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Beard, R. R.; Wertheim, G. A. (1967) Behavioral impairment associated with small doses of carbon monoxide.
3 Am J. Public Health 57: 2012-2022.
4 Federal Register. (1980) Carbon monoxide; proposed revisions to the national ambient air quality standards:
5 proposed rule. F. R. (August 18) 45: 55,066-55,084.
6 Federal Register. (1985) Review of the national ambient air quality standards for carbon monoxide; final rule.
7 F. R. (September 13) 50: 37,484-37,501.
8 Federal Register. (1990) Draft criteria document for carbon monoxide; notice of availability of external review
9 draft. F. R. (April 19) 55: 14,858.
10 Federal Register. (1994) National ambient air quality standards for carbon monoxide—final decision. F. R.
11 (August 1) 59: 38,906-38,917.
12 Horvath, S. M; Ayres, S. M; Sheps, D. S.; Ware, J. (1983) [Letter to Dr. Lester Grant, including the peer-review
13 committee report onDr. Aronow's studies]. Washington, DC: U.S. Environmental Protection Agency,
14 Central Docket Section; Docket no. OAQPS-79-7 IV. H.58.
15 National Air Pollution Control Administration. (1970) Air quality criteria for carbon monoxide. Washington, DC:
16 U.S. Department of Health, Education, and Welfare, Public Health Service; report no. NAPCA-PUB-AP-62.
17 Available from: NTIS, Springfield, VA; PB-190261.
18 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
19 New York, NY: John Wiley & Sons, Inc.
20 U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
21 quality standards. U. S. C. 42: §§7408-7409.
22 U.S. Environmental Protection Agency. (1979) Air quality criteria for carbon monoxide. Research Triangle Park,
23 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
24 no. EPA-600/8-79-022. Available from: NTIS, Springfield, VA; PB81-244840.
25 U.S. Environmental Protection Agency. (1984) Revised evaluation of health effects associated with carbon
26 monoxide exposure: an addendum to the 1979 EPA air quality criteria document for carbon monoxide.
27 Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
28 Assessment Office; report no. EPA-600/8-83-033F. Available from: NTIS, Springfield, VA;
29 PB85-103471/HSU.
30 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
31 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
32 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
33 U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
34 monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
3 5 Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available from:
36 NTIS, Springfield, VA; PB93-157717.
37 U.S. Environmental Protection Agency. (1996) National air quality and emissions trends report, 1995. Research
3 8 Triangle Park, NC: Office of Air Quality Planning and Standards, Emissions Monitoring and Analysis
39 Division; report no. EPA/454/R-96-005. Available from: NTIS, Springfield, VA; PB97-127500/XAB.
October 1999 1-18 DRAFT-DO NOT QUOTE OR CITE
-------
i 2. ANALYTICAL METHODS FOR
2 MONITORING CARBON MONOXIDE
3
4
5 2.1 INTRODUCTION
6 Investigations into relationships between ambient carbon monoxide (CO) levels and human
7 health outcomes and public health warnings of potentially harmful CO levels require accurate,
8 precise, and representative measurements of CO. Reliable measurement methods also are needed
9 to evaluate the effects of ambient CO on overall air quality. This chapter will review methods
10 for monitoring CO in ambient air for conditions ranging from clean continental environments to
11 polluted urban ones. Biological methods for monitoring the impact of ambient CO exposure on
12 human populations also will be reviewed.
13 To promote uniform enforcement of the air quality standards set forth under the Clean Air
14 Act as amended (U.S. Code, 1991), the U.S. Environmental Protection Agency (EPA) has
15 established provisions under which analytical methods can be designated as "reference" or
16 "equivalent" methods (Code of Federal Regulations, 199 la). Either a reference method or
17 equivalent method for air quality measurements is required for acceptance of measurement data
18 for National Ambient Air Quality Standards (NAAQS) compliance. An equivalent method for
19 monitoring CO can be so designated when the method is shown to produce results equivalent to
20 the approved reference monitoring method based on absorption of infrared radiation from a
21 nondispersed beam.
22 The EPA-designated reference methods are automated methods utilizing the nondispersive
23 infrared (NDIR) technique, generally accepted as being the most reliable, continuous method for
24 the measurement of CO in ambient air. The official EPA reference methods (Code of Federal
25 Regulations, 199la) include eleven reference methods designated for use in determining
26 compliance for CO. Before a particular NDIR instrument can be used in a reference method, it
27 must be designated by the EPA as approved in terms of manufacturer, model number,
28 components, operating range, etc. Several NDIR instruments have been so designated (Code of
29 Federal Regulations, 199la), including the gas filter correlation (GFC) technique, which was
October 1999 2-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 developed through EPA-sponsored research (Burch et al., 1976). No equivalent method using a
2 principle other than NDIR has been designated for measuring CO in ambient air.
3
4
5 2.2 OVERVIEW OF TECHNIQUES FOR MEASUREMENT OF AMBIENT
6 CARBON MONOXIDE
7 The NDIR technique is an automated and continuous method that is based on the specific
8 absorption of infrared radiation by the CO molecule (Feldstein, 1967). Most commercially
9 available analyzers incorporate a gas filter to minimize interferences from other gases; they
10 operate near atmospheric pressure, and the most sensitive analyzers are able to detect minimum
11 CO concentrations of about 0.02 ppm. Interferences because of carbon dioxide (CO2) and water
12 vapor can be dealt with so as not to affect the data quality; a particle filter (Teflon® or nylon
13 composition is recommended), and desiccant in the inlet line improves reliability. Nondispersive
14 infrared analyzers are relatively insensitive to flow rate, require no wet chemicals, are sensitive
15 over wide concentration ranges, and have short response times. Nondispersive infrared analyzers
16 of the newer GFC type have overcome zero and span problems, as well as minor problems
17 caused by vibrations.
18 A more sensitive method for measuring low background levels is gas chromatography
19 (Bergman et al., 1975; Bruner et al., 1973; Dagnall et al., 1973; Porter and Volman, 1962;
20 Feldstein, 1967; Smith et al., 1975; Swinnerton et al., 1968; Tesarik and Krejci, 1974). This
21 technique is an automated, semicontinuous method where CO is separated from water, CO2, and
22 hydrocarbons other than methane (CH4) by a stripper column. Carbon monoxide and CH4 then
23 are separated on an analytical column, and the CO is passed through a catalytic reduction tube,
24 where it is converted to CH4. The CO (converted to CH4) passes through a flame ionization
25 detector (FID), and the resulting signal is proportional to the concentration of CO in the air.
26 Mercury liberation detectors offer greater sensitivity and ease of operation than FID's.
27 (Section 2.4.4.3). These methods have no known interferences and can be used to measure levels
28 from 0.02 to 45 ppm.
29 Whichever method or instrument is used, it is essential that the results be evaluated by
30 frequent calibration with samples of known composition (Commins et al., 1977; Goldstein, 1977;
31 National Bureau of Standards, 1975). Chemical analyses can be relied on only after the analyst
October 1999 2-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 has achieved acceptable accuracy in the analysis of such standard samples through an audit
2 program.
3 The performance specifications for automated CO analyzers currently in use are shown in
4 Table 2-1. The normal full-scale operating range for reference methods is 0 to 50 ppm (0 to
5 57 mg/m3). Some instruments offer higher ranges, typically 0 to 100 ppm (0 to 115 mg/m3),
6 or lower ranges such as 0 to 20 ppm (0 to 23 mg/m3). Higher ranges up to 1,000 ppm
7 (1,145 mg/m3) are used to measure CO concentrations in vehicular tunnels and parking garages.
8 Although CO is one of the criteria pollutants, it is also a precursor to ozone and a useful
9 tracer of combustion-derived pollutants (Carter, 1991; Ryan et al., 1998). These additional roles
10 for CO make its detection at levels well below the NAAQS highly desirable. At many existing
11 monitoring sites, the mixing ratio is frequently below the lower detectable limit specified in
12 Table 2-1. Chemical Transport Models (CTMs) developed to understand air pollution and often
13 required to test abatement strategies for photochemical smog, rely on accurate data for
14 concentrations of source gases including nitrogen oxides, non-methane hydrocarbons, and CO.
15 Boundary layer CO concnetrations ratios in urban areas are typically 100s of ppb (Seinfeld and
16 Pandis, 1998; Moy et al., 1994; Morales-Morales, 1998). A CO monitor with precision of
17 500 ppb would be adequate to prove compliance with the CO standard, but would not provide
18 adequate input data for CTMs. This chapter, therefore, will review methods for measuring CO in
19 ambient air that provide sensitivity adequate to quantify the content of clean continental
20 boundary layer air, that is with uncertainty on the order of 10 ppb and has a detection limit
21 around 50 ppb in addition to methods in current use. Suggested performance specifications for
22 monitoring CO in nonurban environments are shown in Table 2-2.
23
24
25 2.3 GAS STANDARDS FOR CALIBRATION
26 There are basically two different types of calibration gas mixture, pre-made blends and
27 mixtures prepared in the laboratory. Certain types of pre-made blends can be purchased with
28 recognized and accepted certification and traceability information. Other pre-made blends can be
29 purchased without certification or with certification of limited acceptance. There is no
30 mechanism to provide accepted certification for mixtures made in the laboratory. The EPA
31 accepts only the first four types of gas mixtures described below.
October 1999 2-3 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-1. PERFORMANCE SPECIFICATIONS FOR AUTOMATED
ANALYTICAL METHODS FOR CARBON MONOXIDE
Range 0 to 50 ppm (0 to 57 mg/m3)
Noise 0.5 ppm (0.6 mg/m3)
Lower detectable limit 1.0 ppm (1.2 mg/m3)
Interference equivalent
Each interfering substance ±1.0 ppm (±1.2 mg/m3)
Total interfering substances 1.5 ppm (1.7 mg/m3)
Zero drift
12 h ±1.0 ppm (±1.2 mg/m3)
24 h ±1.0 ppm (±1.2 mg/m3)
Span drift, 24 h
20% of upper range limit ± 10.0%
80% of upper range limit ± 2.5%
Lag time lOmin
Rise time 5 min
Fall time 5 min
Precision
20% of upper range limit 0.5 ppm (0.6 mg/m3)
80% of upper range limit 0.5 ppm (0.6 mg/m3)
Definitions:
Range: Nominal minimum and maximum concentrations that a method is capable of measuring.
Noise: The standard deviation about the mean of short duration deviations in output that are not caused by input
concentration changes.
Lower detectable limit: The minimum pollutant concentration that produces a signal of twice the noise level.
Interference equivalent: Positive or negative response caused by a substance other than the one measured.
Zero drift: The change in response to zero pollutant concentration during continuous unadjusted operation.
Span drift: The percent change in response to an upscale pollutant concentration during continuous unadjusted
operation.
Lag time: The time interval between a step change in input concentration and the first observable corresponding
change in response.
Rise time: The time interval between initial response and 95% of final response.
Fall time: The time interval between initial response to a step decrease in concentration and 95% of final response.
Precision: Variation about the mean of repeated measurements of the same pollutant concentration expressed as
one standard deviation about the mean.
Source: Code of Federal Regulations (1991a).
October 1999 2-4 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-2. SUGGESTED PERFORMANCE SPECIFICATIONS FOR MONITORING
CARBON MONOXIDE IN NONURBAN ENVIRONMENTS
Range
Noise
Lower detectable limit
Interference equivalent
Each interfering substance
Total interfering substances
Zero drift
12 h
24 h
Zero interval/ maximum
Span drift, 24 h
20% of upper range limit
80% of upper range limit
Lag time
Rise time
Fall time
Precision
20% of upper range limit
80% of upper range limit
0 to 50 ppm (0 to 57 mg/m3)
0.05 ppm (0.06 mg/m3)
0.05 ppm (0.06 mg/m3)
±0.05 ppm (±0.06 mg/m3)
0.10 ppm (0.12 mg/m3)
±0.1 ppm (±0.12 mg/m3)
±0.1 ppm (±0.12 mg/m3)
1 h
±5.0%
±2%
1 min
5 min
5 min
0.2 ppm (0.24 mg/m3)
0.2 ppm (0.24 mg/m3)
aZero interval is the interval between measuring chemical zeros.
Source: Adapted from Code of Federal Regulations (1991a).
1
2
3
4
5
6
7
2.3.1 Pre-made Mixtures
2.3.1.1 Standard Reference Materials
Calibration gas standards of CO in air (certified at levels of approximately 12, 23, and
46 mg/m3 or (10, 20, and 40 ppm, respectively) or in nitrogen (N2; 10 ppm to 13%) are
obtainable from the Standard Reference Material Program of the National Institute of Standards
and Technology (NIST), formerly the National Bureau of Standards, Gaithersburg, MD 20899.
These Standard Reference Materials (SRMs) are supplied as compressed gas mixtures at about
135 bar (1,900 psi) in high-pressure aluminum cylinders containing 800 L (28 ft3) of gas at
standard temperature and pressure, dry (STPD) (National Bureau of Standards, 1975; Guenther
October 1999
2-5
DRAFT-DO NOT QUOTE OR CITE
-------
1 et al., 1996). Each cylinder is supplied with a certificate stating concentration and uncertainty.
2 The concentrations are certified to be accurate to ±1% relative to the stated values. Because of
3 the resources required for their certification, SRMs are not intended for use as daily working
4 standards, but rather as primary standards against which transfer standards can be calibrated.
5
6 2.3.1.2 National Institute of Standards and Technology Traceable Reference Materials
7 Calibration gas standards of CO in air or N2, in the concentrations indicated above, are
8 obtainable from specialty gas companies. Information as to whether a specialty gas company
9 supplies such mixtures is obtainable from the specific company, or the information may be
10 obtained from the Standard Reference Material Program of NIST. These NIST Traceable
11 Reference Materials (NTRMs) are purchased directly from industry and are supplied as
12 compressed gas mixtures at about 135 bar (1,900 psi) in high pressure aluminum cylinders
13 containing 4,000 L (140 ft3) of gas at STPD. Each cylinder is supplied with a certificate stating
14 concentration and uncertainty. The concentrations are certified to be accurate to ±1% relative to
15 the stated values (Guenther et al., 1996).
16
17 2.3.1.3 U.S. Environmental Protection Agency Protocol Gases
18 Calibration gas standards of CO in air or CO in N2 at approximately the same
19 concentrations as SRMs and NTRMs can be purchased from specialty gas companies as EPA
20 Protocol Gases. These gases are blended and analyzed according to an EPA protocol document
21 and are supplied as gas mixtures in high pressure aluminum cylinders. These mixtures are
22 supplied with certificates stating concentration and uncertainty (U.S. Environmental Protection
23 Agency, 1997).
24
25 2.3.1.4 Dutch Bureau of Standards
26 Calibration gas standards of CO in air over a wide concentration range also can be
27 purchased from the Dutch Bureau of Standards, which is the Nederland Meetinstituut (NMi)
28 Holland (fax 31-15-261-2971). These are Primary Reference Materials (PRMs) or Certified
29 Reference Materials (CRMs). These Reference Materials (PRMs or CRMs) are supplied as
30 compressed gas mixtures at about 135 bar (1,900 psi) in high pressure aluminum cylinders
31 containing 800 L of gas at STPD. Each cylinder is supplied with a certificate stating
October 1999 2-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentration and uncertainty. The NIST and EPA recognize the equivalency of specific NMi
2 standards with NIST standards on the strength of the NIST/NMi Declaration of Equivalency
3 Document.
4
5 2.3.1.5 Commercial Blends
6 Calibration gas mixtures of CO in air or N2 over a wide concentration range also can be
7 purchased commercially from many specialty gas companies. Some mixtures may have
8 "certification" documentation and some may not. These mixtures can be ordered in cylinders of
9 almost any size. Mild steel cylinders are to be avoided (U.S. Environmental Protection Agency,
10 1991).
11 The nominal values for CO concentration supplied by the vendor should be verified by
12 intercomparison with an SRM or other validated standard sample. A three-way intercomparison
13 has been made among the NIST SRM's, commercial gas blends, and an extensive set of standard
14 gas mixtures prepared by gravimetric blending at EPA (Paulsell, 1976). Results of the
15 comparison showed that commercial gas blends are within ±2% of the true value represented by
16 a primary standard. Another study on commercial blends (Elwood, 1976) found poorer accuracy.
17 To achieve compatible results in sample analyses, different laboratories should interchange and
18 compare their respective working standards frequently.
19
20 2.3.2 Laboratory Blended Mixtures
21 Mixtures of CO in almost any matrix gas can be blended in the laboratory. One can start
22 with gaseous CO or mixtures of CO and dilute these to any concentration desired. The three
23 common procedures for blending mixtures into containers are the gravimetric (weighing)
24 procedure, the manometric (pressure) technique, and the volumetric method. One also can use
25 dynamic dilution to prepare standards that are not stored in containers but are used at the time of
26 preparation. There are advantages and disadvantages to each procedure, and one must evaluate
27 the application, standards requirements, and laboratory equipment before choosing the method of
28 standards preparation.
29 Standard samples of CO in air also can be prepared by flowing gas dilution techniques.
30 In a versatile system designed for this purpose (Hughes et al., 1973), air at a pressure of about
31 0.7 to 7.0 bars (about 10 to 100 psi) above ambient is first purified and dried by passage through
October 1999 2-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 cartridges of charcoal and silica gel, then is passed through a sintered metal filter into a flow
2 control and flowmeter system. The CO (or a mixture of CO in air that is to be diluted further),
3 also under pressure, is passed through a similar flow control and flowmeter system.
4 Dynamic dilution employed to make CO standards often relies on mass flow controllers.
5 When performing a calibration with this technique, care should be taken to control the
6 temperature and pressure of the flow controllers. Investigations into the performance of several
7 brands of mass flow controllers on aircraft has revealed that, for large pressure changes, some
8 instruments experience errors in the output well beyond the specifications (Weinheimer and
9 Ridley, 1990).
10
11 2.3.3 Other Methods
12 Permeation tubes have been used for preparing standard mixtures of such pollutant gases as
13 sulfur dioxide (SO2) and nitrogen dioxide (O'Keeffe and Ortman, 1966; Scaringelli et al., 1970).
14 Permeation tubes are not used routinely in the United States for making CO standard samples and
15 are not recommended. In the permeation tube techniques, a sample of the pure gas under
16 pressure is allowed to diffuse through a calibrated partition at a defined rate into a diluent gas
17 stream to give a standard sample of known composition.
18 Another possible way to liberate known amounts of CO into a diluent gas is by thermal
19 decomposition of nickel tetracarbonyl [Ni(CO)4]. However, an attempt to use this as a
20 gravimetric calibration source showed that the relation between CO output and weight loss of the
21 Ni(CO)4 is nonstoichiometric (Stedman et al., 1976).
22
23 2.3.4 Intercomparisons of Standards
24 Initial efforts to establish the absolute uncertainty of CO standards and to put various
25 research groups around the world on the same scale revealed systematic errors in some of the
26 standards. Careful preparation of gas standards and repeated intercomparison of calibration
27 gases and measurements on ambient air have since led to general agreement within the
28 international community on both a reference scale and on analytical methods. Calibration
29 standards now generally agree to within 5%, and atmospheric measurements made with a variety
30 of analytical techniques agree to 10 ppb or better.
October 1999 2-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 The National Aeronautics and Space Administration (NASA), as part of the Chemical
2 Instrumentation Test and Evaluation Project, intercompared a tunable diode laser spectroscopy
3 (TDLS) technique and several "grab"-sample gas chromatography-flame ionization detection
4 (GC-FID) techniques (Hoell et al., 1984, 1985). Initial results indicated a high degree of
5 correlation among the various instruments, but agreement on the absolute concentration was only
6 about 15%; differences were as large as 38%. When the intercomparison was repeated (Hoell
7 et al., 1987), calibration standards agreed within 95% confidence levels. Measurements of
8 ambient air samples under actual field conditions demonstrated agreement within experimental
9 uncertainty (on the order of 10 ppb) for CO concentrations ratios from 60 to 170 ppb. When data
10 from the various instruments were regressed, however, slopes again differed from unity by as
11 much as 14%.
12 Careful intercomparisons of calibration gases indicate that accurate and consistent
13 standards can be made. Hughes et al. (1991) compared primary gas standards of CO in
14 N2 produced by NIST and the National Physical Laboratory in the United Kingdom. These
15 standards, prepared gravimetrically, contained concentrations ratios ranging from 10 ppm to 8%.
16 In a blind intercomparison, the mean difference was 0.2%, well within the experimental
17 uncertainty of the techniques. Novell! et al. (1991) gravimetrically produced CO in zero air in
18 the range of 25 to 1,000 ppb from both pure CO and a NIST SRM; they found agreement to
19 within 1%. Agreement with commercially available NIST-traceable standards was within 3%.
20 Reasonable consistency (6% or better) was found with standards used by Australian, German,
21 Brazilian, and several American institutions. One Australian standard was found to be 22%
22 lower, although trouble with this standard had been reported previously (Weeks et al., 1989).
23 A reevaluation of the reference scale in the range of nonurban ambient concentrations (Novell!
24 et al., 1994) confirmed agreement to within 5% or better for the National Oceanic and
25 Atmospheric Administration (NOAA), NASA, and German groups.
26 Intercomparisons of TDLS and NDIR GFC techniques (Poulida et al., 1991; Fried et al.,
27 1991) indicated agreement within experimental uncertainty (better than 10% for typical
28 tropospheric concentrations of 100 to 1,000 ppb), when NIST-based standards were used to
29 calibrate both instruments. These experiments demonstrated good agreement in ambient and
30 compressed air. These results, as well as results form spiking tests, indicated no significant
October 1999 2-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 interferences in either monitor. The intercomparisons also established linearity for both
2 techniques in the range from 100 ppm to 10 ppb.
3 Recent standards normalization and intercomparisons of TDLS, mercury liberation,
4 GC-FID, and NDIR techniques are described by Novell! et al. (1998). For concentrations ratios
5 down to the lowest expected in the boundary layer, about 50 ppb, agreement among groups was
6 typically better than 10 ppb; for higher mixing ratios the typical agreement was about 5%.
7
8 2.3.5 Infrared Absorption
9 The TDLS can provide an independent measurement of the concentration of a CO standard.
10 Fried et al. (1991) used the high-resolution transmission (HITRAN) molecular absorption
11 database for the line parameters to calculate the concentration based on direct absorption. Their
12 results agreed with a NIST-certified gas standard to within 1.6%, well within the uncertainty of
13 the absorption measurement.
14
15
16 2.4 MEASUREMENT IN AMBIENT AIR
17 This section discusses several important aspects of the continuous and intermittent
18 measurement of CO in the atmosphere, including sampling techniques, sampling schedules, and
19 recommended analytical methods for CO measurement.
20
21 2.4.1 Sampling System Components
22 Carbon monoxide monitoring requires a sample introduction system, an analyzer system,
23 and a data recording system. A sample introduction system consists of a sampling probe,
24 an intake manifold, tubing, and air movers. This system is needed to collect the air sample from
25 the atmosphere and to transport it to the analyzer without altering the original concentration.
26 It also may be used to introduce known gas concentrations in order to periodically check the
27 reliability of the analyzer output. Construction materials for the sampling probe, intake
28 manifold, and tubing should be tested to demonstrate that the test atmosphere composition or
29 concentration is not altered significantly. It is recommended that sample introduction systems be
30 fabricated from borosilicate glass or fluorinated ethylene propylene Teflon® (Code of Federal
October 1999 2-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 Regulations, 1991b) if several pollutants are to be monitored. However, in monitoring for CO
2 only, it has been reported (Wohlers et al., 1967) that no measurable pollutant losses were
3 observed at the high (>1 L/min) sampling flow rates when sampling systems were constructed of
4 tygon, polypropylene, polyvinylchloride, aluminum, or stainless steel piping. The sample
5 introduction system should be constructed so that it presents no pressure drop to the analyzer.
6 At low flow and low concentrations, such operation may require validation.
7 The analyzer system consists of the analyzer itself and any sample-preconditioning
8 components that may be necessary. Sample preconditioning might require a moisture control
9 system such as a Nafion® drying tube to help minimize the false positive response of the analyzer
10 (e.g., the NDIR analyzer) to water vapor and a particulate filter to help protect the analyzer from
11 clogging and possible chemical interference caused by paniculate buildup in the sample lines or
12 analyzer inlet. The sample preconditioning system also may include a flow metering and control
13 device to control the sampling rate to the analyzer.
14
15 2.4.2 Quality Assurance Procedures for Sampling
16 The accuracy and validity of data collected from a CO monitoring system must be ensured
17 through a quality assurance program. Such a program consists of procedures for calibration,
18 operational and preventive maintenance, data handling, and auditing; the procedures should be
19 documented fully in a quality assurance program manual maintained by the monitoring
20 organization.
21 Calibration procedures consist of periodic multipoint primary calibration and secondary
22 calibration, both of which are prescribed to minimize systematic error. Primary calibration
23 involves the introduction of test atmospheres of known concentration to an instrument in its
24 normal mode of operation for the purpose of producing a calibration curve.
25 A calibration curve is derived from the analyzer response obtained by introducing several
26 successive test atmospheres of different known concentrations. One recommended method for
27 generating CO test atmospheres is to use air containing no CO along with several known
28 concentrations of CO in air or N2 contained in high-pressure gas cylinders and verified by
29 NIST-certified SRMs wherever possible (Code of Federal Regulations, 199 la). The CO can be
30 removed from an air stream by oxidation to CO2 on a catalyst (Dickerson and Delany, 1988;
31 Parrish et al., 1994). The number of standard gas mixtures (cylinders) necessary to establish a
October 1999 2-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 calibration curve depends on the nature of the analyzer output. A multipoint calibration at five or
2 six different CO concentrations covering the operating range of the analyzer is recommended by
3 EPA (Code of Federal Regulations, 1991b; Federal Register, 1978). Alternatively, the multipoint
4 calibration is accomplished by diluting a known high-concentration CO standard gas with zero
5 gas in a calibrated flow dilution system.
6 Secondary calibration consists of a zero and upscale span of the analyzer. This is
7 recommended to be performed daily (Federal Register, 1978). If the analyzer response differs by
8 more than 2% from the certified concentrations, then the analyzer is adjusted accordingly.
9 Complete records of secondary calibrations should be kept to aid in data reduction and for use in
10 auditing. For high-sensitivity measurement, hourly zeros and weekly calibrations are
11 recommended.
12 Specific criteria for data selection and several instrument checks are available (Smith and
13 Nelson, 1973). Data recording involves recording in a standard format for data storage,
14 interchange of data with other agencies, or data analysis. Data analysis and interpretation usually
15 include a mathematical or statistical analysis of air quality data and a subsequent effort to
16 interpret results in terms of exposure patterns, meteorological conditions, characteristics of
17 emission sources, and geographic and topographic conditions.
18 Auditing procedures consist of several quality control checks and subsequent error analyses
19 to estimate the accuracy and precision of air quality measurements. The quality control checks
20 for CO include a data processing check, a control sample check, and a water vapor interference
21 check, which should be performed by a qualified individual independent of the regular operator.
22 The error analysis is a statistical evaluation of the accuracy and precision of air quality data.
23 Guidelines have been published by EPA (Smith and Nelson, 1973) for calculating an overall bias
24 and standard deviation of errors associated with data processing, measurement of control
25 samples, and water vapor interference, from which the accuracy and precision of CO
26 measurements can be determined. Since January 1, 1983, all state and local agencies submitting
27 data to EPA must provide estimates of accuracy and precision of the CO measurements based on
28 primary and secondary calibration records (Federal Register, 1978). The precision and accuracy
29 audit results through 1985 indicate that the 95% national probability limits for precision are
30 ±9%, and the 95% national probability limits for accuracy are within ±1.5% for all audit levels
October 1999 2-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 up to 85 ppm. The results (accuracy) for CO exceed comparable results for other criteria
2 pollutants with national ambient air quality standards (Rhodes and Evans, 1987).
3
4 2.4.3 Sampling Schedules
5 Carbon monoxide concentrations in the atmosphere exhibit large temporal variations
6 because of changes in the time and rate that CO is emitted by different sources and because of
7 changes in meteorological conditions that govern the amounts of transport and dilution that take
8 place. During a 1-year period, an urban CO station may monitor hourly concentrations of CO
9 ranging from below the minimum detection limit to as high as 45 ppm (52 mg/m3). The NAAQS
10 for CO are based on the second highest 1- and 8-h average concentrations; violations represent
11 extreme events. In order to measure the highest two values from the distribution of 8,760 hourly
12 values in a year, the "best" sampling schedule to employ is continuous monitoring 24 h per day,
13 365 days per year. Even so, continuous monitors rarely operate for long periods without data
14 losses because of malfunctions, upsets, and routine maintenance. Data losses of 5 to 10% are
15 common. Consequently, the data must be interpreted in terms of the likelihood that the NAAQS
16 were attained or exceeded. Statistical methods can be employed to interpret the results (Garbarz
17 etal., 1977; Larsen, 1971).
18 Compliance with 1- and 8-h NAAQS requires continuous monitoring. Statistically valid
19 sampling could be performed on random or systematic schedules, however, if annual averages or
20 relative concentration levels were of importance. Most investigations of various sampling
21 schedules have been conducted for paniculate air pollution data (Hunt, 1972; Ott and Mage,
22 1975; Phinney and Newman, 1972), but the same schedules also could be used for CO
23 monitoring. However, most instruments do not perform reliably in intermittent sampling.
24
25 2.4.4 Continuous Analysis
26 2.4.4.1 Nondispersive Infrared Photometry
27 Carbon monoxide has a characteristic infrared absorption near 4.6 jim. The absorption of
28 infrared radiation by the CO molecule therefore can be used to measure CO concentration in the
29 presence of other gases. The NDIR method is based on this principle.
October 1999 2-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nondispersive infrared systems have several advantages. They are not sensitive to flow
2 rate, they require no wet chemicals, they are reasonably independent of ambient air temperature
3 changes, they are sensitive over wide concentration ranges, and they have short response times.
4 Further, NDIR systems may be operated by nontechnical personnel. Gas filter correlation
5 spectroscopy analyzers are used most frequently now in documenting compliance with ambient
6 air standards.
7
8 Gas-Filter Correlation Spectroscopy
9 A GFC monitor (Burch et al., 1976) has the advantages of an NDIR instrument and the
10 additional advantages of smaller size, no interference from CO2, and very small interference from
11 water vapor. A top schematic view of the GFC monitor is shown in Figure 2-1, showing the
12 components of the optical path for CO detection. During operation, air flows continuously
13 through the sample cell. Radiation from the source is directed by optical transfer elements
14 through the two main optical subsystems: (1) the rotating gas filter and (2) the optical multipass
15 (sample) cell. The beam exits the sample cell through interference filter FC, which limits the
16 spectral passband to a few of the strongest CO absorption lines in the 4.6-|im region. Detection
17 of the transmitted radiation occurs at the infrared detector, C.
18 The gas correlation cell is constructed with two compartments: one compartment is filled
19 with 0.5 atm CO, and the second compartment is filled with pure N2. Radiation transmitted
20 through the CO is completely attenuated at wavelengths where CO absorbs strongly. The
21 radiation transmitted through the N2 is reduced by coating the exit window of the cell with a
22 neutral attenuator so that the amounts of radiation transmitted by the two cells are made
23 approximately equal in the passband that reaches the detector.
24 In operation, radiation passes alternately through the two cells as they are rotated to
25 establish a signal modulation frequency. If CO is present in the sample, the radiation transmitted
26 through the CO is not appreciably changed, whereas that through the N2 cell is changed. This
27 imbalance is linearly related to CO concentrations in ambient air.
28
29 Enhanced Performance
30 Although commercial CO monitors were designed to meet the performance specifications
31 shown in Table 2-1, several instruments have the potential for much greater sensitivity.
October 1999 2-14 DRAFT-DO NOT QUOTE OR CITE
-------
A.
B.
Plated Pattern
Cell Containing N2 Neutral Attentuator
Chopper
Cell Containing
Carbon Monoxide
Figure 2-1. Schematic diagram of GFC monitor for CO. A = optical layout (M denotes
mirror reflector, and L denotes lens); B = detail of correlation cell.
Source: Chancy and McClenny (1977).
1 Modifications of commercially available NDIR monitors (Dickerson and Delany, 1988; Parrish
2 et al., 1994) have been made to enhance their performance, but the manufacturers have continued
3 to improve instruments and offer "high-sensitivity" options that could meet the requirements of
4 monitoring clean continental air (i.e., a detection limit of about 50 ppb and resolution of 10 ppb).
October 1999
2-15
DRAFT-DO NOT QUOTE OR CITE
-------
1 The principal constraints on the lower detectable limits of commercially available NDIR
2 CO monitors are detector noise, water vapor interference, and drift in the background. Several
3 methods have been developed by researchers to improve detector noise, such as cooling the
4 preamplifier and improving the optics. More recent improvements made by the manufacturers,
5 such as gold-coated mirrors and selected infrared (IR) radiation detectors have been effective in
6 reducing detector noise.
7 Water vapor produces a negative artifact such that a volume mixing ratio of 1% would
8 reduce apparent CO mixing ratio measurement by 50 ppb. This interference can be reduced to
9 within tolerances by drying the sample air with a cold trap, desiccant, or drying tube (Dickerson
10 and Delany, 1988). Alternatively, the zero can be checked frequently enough so that changes in
11 ambient humidity are unlikely to produce a significant error (Parrish et al., 1994).
12 The greatest source of potential error in monitoring CO in the 0.1-ppm range is background
13 drift. The stability of the instruments with respect to changes in calibration (span) is adequate,
14 but the background (zero) drifts on time scales of minutes to hours in response to, among other
15 factors, instrument temperature. This drift can be accounted for most easily by frequent chemical
16 zeroing with a oxidizer that converts CO to CO2.
17
18 2.4.4.2 Gas Chromatography-Flame lonization
19 Carbon monoxide can be measured in either ambient air samples collected every few
20 minutes or in air from grab samples stored under pressure in inert canisters. Carbon monoxide in
21 air samples is dried, preconcentrated, reduced to methane, and detected by flame ionization
22 (GC/FID) (Heidt, 1978; Greenberg et al., 1984; Hoell et al., 1987). Uncertainty on the order of
23 10 ppb or 10% of the observation can be obtained routinely.
24
25 2.4.4.3 Mercury Liberation
26 This technique, involving reaction with hot mercuric oxide to give elemental mercury
27 vapor, was developed early this century (Moser and Schmid, 1914; Beckman et al., 1948;
28 McCullough et al., 1947; Mueller, 1954; Palanos, 1972; Robbins et al., 1968) and is now
29 available commercially (e.g., Trace Analytical Inc., Menlo Park, CA). The method is
30 temperature and pressure sensitive, and operation in the continuous mode requires elimination of
31 interferences from sulfur dioxide, hydrogen, and hydrocarbons (Seller et al., 1980). Successful
October 1999 2-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 continuous operation has been reported with response time on the order of 20 s and detection
2 limits near 20 ppb (Fishman et al., 1980; Brunke et al., 1990).
3 As a GC detector, mercury liberation (GC/ML) offers high sensitivity, without the
4 interferences inherent in continuous measurements (e.g., Novell! et al., 1991, 1998). Air samples
5 are collected in glass bottles and injected into a gas chromatograph with two columns. The CO is
6 then detected with a commercial mercuric oxide reduction detector (e.g., Trace Analytical Inc.,
7 Menlo Park, CA). The system is linear from 10 to over 1,000 ppb, has a detection limit below
8 10 ppb, and the reported uncertainty is about 2%.
9
10 2.4.4.4 Tunable Diode Laser Spectroscopy
11 Tunable diode lasers (TDLs) produce IR radiation with a line width that is narrow
12 compared with typical absorption lines of atmospheric trace gases. Absorption of IR radiation by
13 a single rotational line in the 4.6-|im band can be exploited to measure CO with high precision
14 and rapid response and without interferences; the sharp focus on a narrow spectral region
15 provides great selectivity. Air samples are measured over open paths through the ambient air
16 (Chaney et al., 1979) or by pulling air samples through an orifice into a long-path cell maintained
17 at a pressure well below ambient (Sachse et al., 1987; Fried et al., 1991; Roths et al., 1996).
18 Radiation from a TDL is modulated over a very narrow wavelength region such that absorption
19 by CO produces an AC signal. The background is measured by catalytic oxidation of CO to CO2.
20 Instruments based on TDLS are currently the fastest and most sensitive extant, with a
21 typical detection limit of a few ppb and a response time of a few seconds. For long-term
22 monitoring, the high cost and need for a skilled operator on site are disadvantages.
23
24 2.4.4.5 Resonance Fluorescence
25 Resonance fluorescence of CO in the vacuum UV has been used for a highly sensitive and
26 rapidly responding instrument (Volz and Kley, 1985; Gerbig et al., 1996). Excitation is
27 represented by the following reaction (Equation 2-1):
CO (X1^) + hv^CO (X'n). (2-1)
28
October 1999 2-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 Atmospheric CO absorbs radiation in the 150-nm range from a radio frequency discharge lamp,
2 and fluorescence from the excited CO is detected by a photo-multiplier tube. The lamp generates
3 a plasma in a continuous flow of CO2 in argon. Limits to the sensitivity of this instrument are set
4 by interference from water vapor, continuum Raman scattering by oxygen (O2), and by drift in
5 the lamp intensity. The pressure in the fluorescence chamber must be maintained between 7 and
6 9 mbar air to balance interference from O2 and signal from CO.
7 Recent improvements (Gerbig et al., 1999) have reduced the detection limit to 3 ppb for a
8 response time of as short as a few seconds. The high sensitivity and small size of the instrument
9 are desirable for measurements from aircraft. Before the instrument is practical for air pollution
10 monitoring, its stability must be improved. As the lamp window degrades, sensitivity is lost,
11 such that after about 200 h of operation, a factor of two loss in the span can be expected.
12
13 2.4.5 Intercomparisons of Methods
14 Several techniques have been evaluated in rigorous intercomparisons under field
15 conditions. For unpolluted tropospheric air, a number of instruments employing different
16 analytical principles have consistently measured concentrations ratios that agree within
17 experimental uncertainty. These techniques include TDLS, NDIR/GFC, GC-FID, and GC-ML
18 (Hoell et al., 1987; Fried et al., 1991; Poulida et al., 1991; Novell! et al., 1998). Additional
19 details can be found in Section 2.2.4.
20
21 2.4.6 Other Methods of Analysis
22 Color changes induced by reaction of a solid or liquid date back to Haldane (1897-1898)
23 and were reviewed extensively in the previous criteria document. Examples include the colored
24 silver sol method, the NIST colorimetric indicating gel, the length-of-stain indicator tube, and
25 frontal analysis (U.S. Environmental Protection Agency, 1991).
26 More recently developed electrochemical techniques show highly improved resolution and
27 specificity (e.g., Langan, 1992; Lee et al., 1992a, Ott et al., 1995). Electrochemical sensors
28 operate by measuring the current of a small fuel cell, and because of their reduced size and power
29 requirements have been used extensively in exposure and indoor research studies (see
30 Section 2.5). Precision of from 0.2 to 2 ppm has been reported. Further independent evaluation
October 1999 2-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 and intercomparison, followed by publication in the reviewed scientific literature is called for to
2 determine the sensitivity, stability, and selectivity of electrochemical methods, in order to
3 establish equivalency to the NDIR instrument by EPA for use in compliance monitoring.
4
5
6 2.5 MEASUREMENT USING PERSONAL MONITORS
7 Monitors at fixed locations provide useful information on ambient CO concentrations and
8 their variability and trends, but cannot measure personal exposure. Information on personal
9 exposure, including home, in-transit, and work-related concentrations is needed for
10 epidemiologic studies. The previous criteria document (U.S. Environmental Protection Agency,
11 1991) reviewed the state of the science of personal monitors as of about 1986. Since that time,
12 the devices have been further developed and refined.
13 One technique involves an ion-exchange Y-type zeolite, with zinc ion as the adsorbent.
14 The adsorbent is desorbed thermally, converted to methane, and analyzed using GC-FID (e.g.,
15 Lee et al., 1992b,c; Lee and Yanagisawa, 1992, 1995). Apte (1997) reviewed several of these
16 devices and described passive samplers based on transition metal compound color changes
17 measured spectrochemically. The method suffers an interference from ethylene, but provides
18 adequate performance (sensitivity of 10 ppm-h and precision of 20% or better) for health studies.
19 Substantial work remains for most passive samplers on stability and response to temperature,
20 humidity, and interferences. These techniques lack the response speed and sensitivity for
21 ambient air monitoring.
22 Numerous field studies on personal exposure have been conducted with electrochemical
23 sensors (Akland et al., 1985; Ott et al., 1986; Wallace et al., 1988; Ott et al., 1994; Klepeis et al.,
24 1999; McBride et al., 1999); some are described in Chapters 3 and 4. These studies show that
25 spatial and temporal variability, and the effects of microenvironments can have a major impact
26 on the level of CO exposure.
27 Carbon monoxide emissions from vehicles can be measured rapidly with a remote sensing
28 IR technique (Bishop et al., 1989; Stedman and Bishop, 1996) in which CO is measured near
29 4.3 //m and CO2 at 3.6 //m; a third wavelength is used as a reference for intensity. The
30 instrument has been evaluated in a double blind intercomparison with on-board NDIR, and the
31 two methods agreed well within experimental uncertainty (Lawson et al., 1990). Surveys
October 1999 2-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 conducted with this technique reveal that a majority of CO is emitted by a minority of vehicles.
2 The method has been used to evaluate the efficacy of inspection and maintenance programs and
3 oxygenated fuels (Beaton et al., 1995; Stedman et al., 1997, 1998). The sum of measurements
4 indicates a general decrease in fleet-averaged CO emissions over the past decade (Bradley et al.,
5 1999).
8 2.6 BIOLOGICAL MONITORING
9 A unique feature of CO exposure is that there is a biological marker of the dose that the
10 individual has received—the blood level of CO. This level may be calculated by measuring
11 blood carboxyhemoglobin (COHb) or by measuring CO in end-tidal exhaled breath after a
12 standardized breathhold maneuver, with a required correction for the background CO inhaled
13 prior to a breathhold (Smith, 1977; Wallace, 1983). The measurement methods for COHb and
14 breath CO were reviewed extensively in the previous criteria document (U.S. Environmental
15 Protection Agency, 1991). This section provides an update on advances in analytical methods for
16 measuring blood COHb and breath CO that have been published in the literature since the
17 previous review. New studies reporting breath CO or blood COHb in population studies are
18 discussed in Chapter 4, along with other new CO exposure assessments.
19
20 2.6.1 Carboxyhemoglobin Measurements
21 Direct reading of COHb usually is performed in the clinical or hospital setting through the
22 use of a direct-reading spectrophotometer, such as a CO-Oximeter (CO-Ox). For clinical
23 purposes, precision on the order of ±1% COHb is not of primary importance, because of the need
24 to differentiate between conditions of low levels of COHb and the much higher levels of COHb
25 that indicate treatment for CO poisoning. The concern in this setting, for example, is to rapidly
26 distinguish between 1 and 10% COHb, not between 1 and 2% COHb. Marshall et al. (1995)
27 showed a wide range of threshold COHb values (measured in the blood by CO-Ox, not estimated
28 from breath CO) used to determine treatment in a sample of 23 Boston, MA, area laboratories.
29 For example, eight laboratories accepted values of 5 to 6% COHb as normal in nonsmokers,
30 a value that cannot be supported by the modern scientific literature. The authors recommended
October 1999 2-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 the use of threshold limits of 3% COHb for nonsmokers and 10% COHb for smokers when
2 classifying subjects for treatment.
3 The performance of the various early versions of the CO-Ox instruments for measuring
4 blood COHb was reviewed in Section 8.5 of U.S. Environmental Protection Agency (1991).
5 These and later instruments, of different design from different manufacturers, used several
6 wavelengths of light for simultaneous measurement of Hb, O2Hb, COHb and methemoglobin
7 (Freeman and Steinke, 1993; Gong, 1995; Bailey et al., 1997). Vreman et al. (1993) and
8 Mahoney et al. (1993) confirmed that considerable difficulties were encountered for COHb
9 concentrations below 5% (a region with which most environmental studies of nonsmokers are
10 concerned), and the authors concluded that the CO-Ox is unreliable for environmental studies.
11 Some versions of the CO-Ox also were found to be influenced by bilirubin and by fetal
12 hemoglobin, presenting difficulty in diagnosing newborn infants with jaundice (Vreman and
13 Stevenson, 1994; Stevenson and Vreman, 1997). Shepard and McMahan (1996) present a highly
14 detailed analysis of the causes and effects of oximeter errors in blood gas analyses.
15 Recent CO-Ox developments have been a new six-wavelength instrument (Instrumentation
16 Laboratory, 1999) used by Kimmel et al. (1999) and a 128-wavelength instrument (Krarup, 1998)
17 both of which identify and correct for possible interferences. The latter instrument is still under
18 formal independent evaluation, and, although peer-reviewed published results of comparison
19 testing are expected to be forthcoming shortly, the only article currently in press is in German,
20 in a non-peer reviewed journal (Krarup, 1999). It is possible that comparison of the results on
21 the same sample using the new multi-wavelength instrument and older fewer-wavelength
22 instruments may show that the new instruments measure lower COHb if they better correct for
23 the positive interferences of various non-COHb species in the blood, such as varying fractions of
24 fetal hemoglobin and sulfhemoglobin. This would be consistent with the report that some
25 laboratories, as cited above, accepted 5 to 6% COHb from oximeter readings as normal for
26 nonsmokers.
27 For a research study to relate health effects or breath CO to COHb, the method of choice is
28 GC analysis of the CO gas released from the blood when COHb is dissociated (U.S.
29 Environmental Protection Agency, 1991; Van Dam and Daenenes, 1994; Lloyd and Rowe,
30 1999). The reader, therefore, is alerted to the difference between end-tidal breath CO to blood
31 COHb relationships when the COHb is determined by CO-Ox or by GC. A calibration curve
October 1999 2-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 relating exhaled end-tidal breath CO to COHb should be based on a standard breath-hold
2 maneuver for the CO collection and the GC method of COHb analysis. It is beyond the scope of
3 this chapter to reanalyze the early COHb literature and estimate the effect of the possible positive
4 interferences that were not accounted for by the early CO-Ox instruments because each
5 instrument and its on-site calibration procedure would create a different bias that cannot be
6 known with certainty. However, in general, the levels of COHb associated with low levels of
7 ambient CO exposure in field studies may have been overestimated in the past.
8
9 2.6.2 Breath Carbon Monoxide Measurements
10 Carbon monoxide in the breath can be measured by all techniques used to measure ambient
11 CO concentrations, as described in U.S. Environmental Protection Agency (1991). A common
12 type of instrument in use for rapidly screening large numbers of people for CO exposures or
13 measuring breath CO distributions is the electrochemical analyzer. The subject performs an
14 inhalation-breathhold maneuver and exhales through a mouthpiece into the instrument inlet. The
15 end tidal breath is retained for analysis, and the reading in ppm CO can be converted to COHb
16 through a calibration curve or nomogram provided with the instrument.
17 Vreman et al. (1993) presented evidence to show that a serious positive interferent in the
18 electrochemical method (hydrogen gas) is present in the exhaled breath of some persons as a
19 result of metabolism of certain foods. Because this could have affected many previous studies,
20 including the very large EPA studies in Washington, DC, and Denver, CO (Akland et al., 1985),
21 it would be desirable to determine the fraction of the population so affected. Because of the
22 general decline of ambient CO, this potential interference takes on more importance in any future
23 studies, which must account for this problem if employing electrochemical devices to measure
24 breath CO.
25 Lee et al. (1991) developed a TDLS system that was well suited for measuring low levels of
26 CO in breath. The system also can detect the abundance of isotopic CO (13C16O), with a
27 preliminary finding of a slight enrichment over atmospheric abundance in breath. Lee et al.
28 (1994) employed the instrument in a study correlating breath CO and blood COHb in people
29 living near Boulder, CO (described in Chapter 4).
30 The passive CO sampler developed by Lee and Yanagisawa (1992, 1995) (see Section 2.5)
31 has a reusable sampling system that allows the collection of only the last 5 mL of a breath
October 1999 2-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 expelled after breath holding for 20 s, thus obtaining alveolar air undiluted by dead space air.
2 The sampler was unaffected by humidity; however, the rather low efficiency of collection (50%)
3 and the resulting fairly high detection limit of 3.2 ppm may limit the utility of the sampler for
4 environmental studies.
5
6 2.6.3 Relationships of Breath Carbon Monoxide to Blood
7 Carboxyhemoglobin
8 The end-tidal breath CO versus COHb relationships reviewed in the previous criteria
9 document (see Table 8-14 in U.S. Environmental Protection Agency, 1991) and in studies
10 published in the literature since then are often at variance because they use either a 10-s, 15-s,
11 or 20-s breathhold step in the breath collection; use either GC or CO-Ox for the blood COHb
12 measurements; or may not correct for the CO content of the inhaled air. The use of a 20-s
13 breathhold, as recommended by Jones et al. (1958), with a correction for the CO content of the
14 inhaled air (Smith, 1977; Wallace, 1983), would improve the reproducibility of the CO breath
15 measurements, and the use of GC would improve the accuracy of the corresponding COHb
16 measurement. The 20-s breathhold is preferable, because it maximizes the approach to
17 equilibrium and minimizes the magnitude of the required correction for CO in inhaled air.
18 Therefore, specific details regarding the length of the breathhold, corrections for inhaled CO, and
19 the method of COHb analysis should be provided in the published discussions of studies of the
20 CO-COHb relationship so that differences among study results can be evaluated.
21 One comprehensive review article on CO-COHb relationships (Vreman et al., 1995)
22 discusses physical and chemical properties, endogenous and exogenous sources of CO, body
23 burden and elimination, toxicity and treatment, clinical chemistry, measurement methods, and the
24 relationship of CO and COHb to bilirubin and jaundice in neonates. A second, less
25 comprehensive review from the same investigators focuses on the production of CO and bilirubin
26 in equal amounts by heme degradation and on the physiological significance of CO as a neuronal
27 messenger (Rodgers et al., 1994).
28 Lee et al. (1994) performed a study of CO-COHb relationships at altitude in Boulder.
29 A total of 13 nonsmoking adults were exposed to 9 ppm CO for both 1 and 8 h. Blood was
30 sampled and end-tidal breath samples were taken after a 10-s breathhold. Mean COHb values
31 prior to exposure were 0.65% and, after exposure for the 1- and 8-h periods were 1.2 and 2.2%,
October 1999 2-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 respectively. The corresponding mean CO levels in the breath samples were about 2.4, 4.4, and
2 8.2 ppm (uncorrected for the ~0 ppm ambient CO in inhaled air), respectively, as shown in
3 Figure 2-2. The slope of 3.65 ppm per 1% COHb saturation after a 10-s breathhold at altitude is
4 somewhat smaller than previous estimates of about 5 ppm CO per 1% COHb, but the previous
5 estimates were based on a 20-s breathhold near sea level that maximizes the end-tidal CO if the
6 inhaled air had a CO concentration below the 20-s end-tidal breath CO (Jones et al., 1958).
7
8 2.6.4 Summary of the Relationship Between Biological Measurements of
9 Carbon Monoxide
10 The use of CO-Ox to measure COHb provides useful information regarding values of
11 COHb in populations being studied for clinical diagnosis. However, the range of COHb values
12 obtained with this optical method for blood collected from nonsmokers is greater than that
13 obtained from a split sample analyzed for COHb by research laboratory GC. Therefore, the
14 greater potential exists with the CO-Ox for having an incorrect absolute value for COHb, as well
15 as an incorrectly broadened range of values, when used in population studies. In addition, it is
16 not clear exactly how sensitive the CO-Ox techniques are to small changes in COHb at the low
17 CO end of the COHb dissociation curve. Interferences (e.g., from variable levels of oxygen
18 saturation of hemoglobin [O2Hb]) and nonlinear phenomena appear to have a very significant
19 influence on the COHb reading at low COHb concentrations in a sample, suggesting nonlinearity
20 or a disproportionality in the absorption spectra of different species of Hb (e.g., HbA [adult],HbF
21 [fetal], HbS [sickle], HbZH [Zurich]). Gas chromatography continues to be the method of choice
22 for measuring COHb in a research setting, although, with care, a CO-Ox can be specially
23 calibrated by GC analysis of calibration-standard blood samples prepared with low COHb
24 concentrations (Allred et al., 1991).
25 The measurement of exhaled breath CO has the advantages of ease, speed, precision
26 (provided the required correction for CO in the inhaled air is made), and greater subject
27 acceptance than the invasive measurement of blood COHb. Breath CO measurement on
28 randomly chosen people can be related to the blood COHb by use of an empirical relationship
29 developed by simultaneous measurements of COHb (preferably by GC) and breath CO using the
30 identical procedure for the breath collection that is used in the population study. The empirical
31 relationships developed with different breath holding techniques will differ from the theoretical
October 1999 2-24 DRAFT-DO NOT QUOTE OR CITE
-------
10
Q.
Q.
O
O
CO
CD
8-
6-
4-
2-
0
• Oh
• 1 h, 9 ppm CO
A 8 h, 9 ppm CO
r= 0.945
0
1 2
Blood Carboxyhemoglobin /0
Figure 2-2. The correlation between an end-tidal breath CO concentration after a 10-s
breathhold and blood COHb levels expressed as individual data points as well
as mean ± standard deviation The breath concentration was not corrected for
the concentration of CO in the inhaled air (Smith, 1977; Wallace, 1983).
Source: Lee et al. (1991, 1994).
1 Haldane equilibrium relationship for the reaction CO + O2Hb <->• O2 + COHb, which depends on
2 the ratio of adult- to fetal-hemoglobin (HbA:HbF). This is because the Haldane relationship is
3 for in vitro static equilibrium, and the empirical end-tidal breath CO-blood COHb relationship is
4 for an in vivo dynamic equilibrium that depends on how long the breath is held and on the
5 correction for the CO in inhaled air.
October 1999
2-25
DRAFT-DO NOT QUOTE OR CITE
-------
1 2.7 SUMMARY
2 The review of the state of the science for this criteria document yields several major points
3 concerning analytical techniques for CO measurement.
4 Several adequate techniques exist for highly reliable monitoring of CO to ensure
5 compliance with the NAAQS. Determination of the actual mean ambient air concentration
6 requires substantially better performance than does the minimum required to demonstrate
7 compliance with the NAAQS. Commercial instruments, sometimes with minor modifications,
8 can meet the measurement needs for supplying useful data on the emission, distribution, and
9 trends of ambient CO and for modeling photochemical smog.
10 Use of enhanced instruments for monitoring of actual CO concentrations with reasonable
11 precision is needed if CO levels in clean continental air outside of urban environments are to be
12 quantified adequately. Commonly used calibration standards and measurement techniques have
13 in the past failed to meet the criteria of precise measurement, but there is now general agreement
14 on procedures for generating standards with absolute accuracy better than about 2% in the parts
15 per million range and about 10% in the range of mixing ratios found in the clean troposphere.
16 Compressed air mixtures, traceable to NIST or NMi provide reliable means of precise
17 calibration.
18 The NDIR, GC/ML, GC/FID, and TDL techniques have undergone careful evaluation with
19 synthetic air mixtures and ambient air, and are deemed reliable. The methods were
20 intercompared in both open and blind studies with designated "disinterested third party" referees.
21 Early problems were identified and corrected, and the most recent intercomparisons indicate
22 general agreement on calibration standards and ambient air measurements over a broad range of
23 concentrations. New techniques should undergo the same rigorous evaluation.
24 Several new electrochemical and passive sampling methods have become available. These
25 techniques are not yet equivalent to the NDIR method for compliance monitoring or precise
26 enough for CO measurements in background ambient air (<0.5 ppm CO) for inputs to Chemical
27 Transport Models (CTM), but they are very useful for personal exposure and indoor research
28 studies. Further work on the stability and specificity of the electrochemical methods to obtain
29 EPA equivalency is warranted.
30 The level of COHb in the blood may be determined directly by blood analysis or indirectly
31 by measuring CO in exhaled breath. The use of CO-Ox to measure COFIb can provide useful
October 1999 2-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 information regarding mean values in populations being studied or as an aid in clinical diagnosis.
2 It has been shown, however, that the range of values obtained with this optical method will be
3 greater than that obtained with other more accurate methods, especially at COHb levels <5%.
4 Gas chromatography continues to be the method of choice for measuring COHb.
5 The measurement of exhaled breath has the advantages of ease, speed, precision, and
6 greater subject acceptance than measurement of blood COHb. However, the accuracy of the
7 breath measurement procedure and the validity of the in vitro Haldane relationship between
8 breath and blood still remains in question, especially at low environmental CO concentrations.
9
October 1999 2-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
3 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
4 Technol. 19: 911-918.
5 Allied, E. N.; Bleecker, E. R.; Chaitman, B. R.; Damns, T. E.; Gottlieb, S. O.; Hackney, J. D.; Pagano, M;
6 Selvester, R. H.; Walden, S. M.; Warren, J. (1991) Effects of carbon monoxide on myocardial ischemia.
7 Environ. Health Perspect. 91: 89-132.
8 Apte, M. G. (1997) A population-based exposure assessment methodology for carbon monoxide: development of a
9 carbon monoxide passive sampler and occupational dosimeter [dissertation]. Berkeley, CA: U.S. Department
10 of Energy, Lawrence Berkeley National Laboratory, Environmental Energy Technologies Division;
11 LBNL-40838. Available from: NTIS, Springfield, VA; DE98052714.
12 Bailey, S. R.; Russell, E. L.; Martinez, A. (1997) Evaluation of the AVOXimeter: precision, long-term stability,
13 linearity, and use without heparin. J. Clin. Monit. 13: 191-198.
14 Beaton, S. P.; Bishop, G. A.; Zhang, Y.; Ashbaugh, L. L.; Lawson, D. R.; Stedman, D. H. (1995) On-road vehicle
15 emissions: regulations, costs, and benefits. Science (Washington, DC) 268: 991-993.
16 Beckman, A. O.; McCullough, J. D.; Crane, R. A. (1948) Microdetermination of carbon monoxide in air: a portable
17 instrument. Anal. Chem. 20: 674-677.
18 Bergman, I.; Coleman, J. E.; Evans, D. (1975) A simple gas chromatograph with an electrochemical detector for the
19 measurement of hydrogen and carbon monoxide in the parts per million range, applied to exhaled air.
20 Chromatographia 8: 581-583.
21 Bishop, G. A.; Starkey, J. R.; Ihlenfeldt, A.; Williams, W. J.; Stedman, D. H. (1989) IR long-path photometry:
22 a remote sensing tool for automobile emissions. Anal. Chem. 61: 671A-677A.
23 Bradley, K. S.; Stedman, D. H.; Bishop, G. A. (1999) A global inventory of carbon monoxide emissions from motor
24 vehicles. Chemosphere: in press.
25 Bruner, F.; Ciccioli, P.; Rastelli, R. (1973) The determination of carbon monoxide in air in the parts per billion
26 range by means of a helium detector. J. Chromatogr. 77: 125-129.
27 Brunke, E.-G.; Scheel, H. E.; Seiler, W. (1990) Trends of tropospheric CO, N2O, and CH4 as observed at Cape
28 Point, South Africa. Atmos. Environ. 24A: 585-595.
29 Burch, D. E.; Gates, F. J.; Pembrook, J. D. (1976) Ambient carbon monoxide monitor. Research Triangle Park, NC:
30 U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; EPA report no.
31 EPA-600/2-76-210. Available from: NTIS, Springfield, VA; PB-259577.
32 Carter, W. P. L. (1991) Development of ozone reactivity scales forvolatile organic compounds. Research Triangle
33 Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
34 Laboratory; report no. EPA/600/3-91/050. Available from: NTIS, Springfield, VA; PB91-243386.
3 5 Chaney, L. W.; McClenny, W. A. (1977) Unique ambient carbon monoxide monitor based on gas filter correlation:
36 performance and application. Environ. Sci. Technol. 11: 1186-1190.
37 Chaney, L. W.; Rickel, D. G.; Russwurm, G. M.; McClenny, W. A. (1979) Long-path laser monitor of carbon
38 monoxide: system improvements. Appl. Opt. 18: 3004-3009.
39 Code of Federal Regulations. (1991a) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
40 Code of Federal Regulations. (1991b) National primary and secondary ambient air quality standards.
41 C. F. R. 40: §50.
42 Commins, B. T.; Berlin, A.; Langevin, M.; Peal, J. A. (1977) Intercomparison of measurement of
43 carboxyhaemoglobin in different European laboratories and establishment of the methodology for the
44 assessment of COHb levels in exposed populations. Luxembourg, Sweden: Commission of the European
45 Communities, Health and Safety Directorate; doc. no. V/F/1315/77e.
46 Dagnall, R. M.; Johnson, D. J.; West, T. S. (1973) A method for the determination of carbon monoxide, carbon
47 dioxide, nitrous oxide and sulphur dioxide in air by gas chromatography using an emissive helium plasma
48 detector. Spectrosc. Lett. 6: 87-95.
49 Dickerson, R. R.; Delany, A. C. (1988) Modification of a commercial gas filter correlation CO detector for
50 enhanced sensitivity. J. Atmos. Oceanic Technol. 5: 424-431.
51 Elwood, J. H. (1976) A calibration procedural system and facility for gas turbine engine exhaust emission analysis.
52 In: Calibration in air monitoring: a symposium; August 1975; Boulder, CO. Philadelphia, PA: American
53 Society for Testing and Materials; pp. 255-272; ASTM special technical publication no. 598.
October 1999 2-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Federal Register. (1978) Air quality surveillance and data reporting: proposed regulatory revisions. F. R. (August 7)
2 43: 34892-34934.
3 Feldstein, M. (1967) Methods for the determination of carbon monoxide. Prog. Chem. Toxicol. 3: 99-119.
4 Fishman, J.; Seiler, W.; Haagenson, P. (1980) Simultaneous presence of O3 and CO bands in the troposphere. Tellus
5 32:456-463.
6 Freeman, G. L.; Steinke, J. M. (1993) Evaluation of two oximeters for use in cardiac catheterization laboratories.
7 Catheterization Cardiovasc. Diagn. 30:51-57.
8 Fried, A.; Henry, B.; Parrish, D. D.; Carpenter, J. R.; Buhr, M. P. (1991) Intercomparison of tunable diode laser and
9 gas filter correlation measurements of ambient carbon monoxide. Atmos. Environ. 25A: 2277-2284.
10 Garbarz, J.-J.; Sperling, R. B.; Peache, M. A. (1977) Time-integrated urban carbon monoxide measurements.
11 San Francisco, CA: Environmental Measurements, Inc.
12 Gerbig, C.; Kley, D.; Volz-Thomas, A.; Kent, J.; Dewey, K.; McKenna, D. S. (1996) Fast response resonance
13 fluorescence CO measurements aboard the C-130: instrument characterization and measurements made
14 during North Atlantic Regional Experiment 1993. J. Geophys. Res. [Atmos.] 101: 29,229-29,238.
15 Gerbig, C.; Schmitgen, S.; Kley, D.; Volz-Thomas, A.; Dewey, K.; Haaks, D. (1999) An improved fast-response
16 vacuum-UV resonance fluorescence CO instrument. J. Geophys. Res. [Atmos.] 104: 1699-1704.
17 Goldstein, G. M. (1977) COHb measurements, results of discussion between Drs. A. Berlin, CEC and G. Goldstein,
18 EPA on ECE doc V/F/1315/77e [letter to Mr. Jack Thompson]. Research Triangle Park, NC: U.S.
19 Environmental Protection Agency; September 23, 1977.
20 Gong, A. K. (1995) Near-patient measurement of methemoglobin, oxygen saturation, and total hemoglobin:
21 evaluation of a new instrument for adult and neonatal intensive care. Crit. Care Med. 23: 193 -201.
22 Greenberg, J. P.; Zimmerman, P. R.; Heidt, L.; Pollock, W. (1984) Hydrocarbon and carbon monoxide emissions
23 from biomass burning in Brazil. J. Geophys. Res. [Atmos.] 89: 1350-1354.
24 Guenther, F. R.; Dorko, W. D.; Miller, W. R.; Rhoderick, G. C. (1996) The NIST traceable reference material
25 program for gas standards. Washington, DC: U.S. Department of Commerce, National Institute of Standards
26 and Technology; NIST special publication 260-126.
27 Haldane, J. (1897-1898) Some improved methods of gas analysis. J. Physiol. (London) 22: 465-480.
28 Heidt, L. E. (1978) Whole air collection and analysis. Atmos. Technol. (9): 3-8.
29 Hoell, J. M.; Gregory, G. L.; Carroll, M. A.; McFarland, M.; Ridley, B. A.; Davis, D. D.; Bradshaw, J.; Rodgers,
30 M. O.; Torres, A. L.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.; Campbell, M. C.; Farmer,
31 J. C.; Sheppard, J. C.; Wang, C. C.; Davis, L. I. (1984) An intercomparison of carbon monoxide, nitric oxide,
32 and hydroxyl measurement techniques: overview of results. J. Geophys. Res. [Atmos.] 89: 11,819-11,825.
33 Hoell, J. M., Jr.; Gregory, G. L.; McDougal, D. S.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.
34 (1985) An intercomparison of carbon monoxide measurement techniques. J. Geophys. Res. [Atmos.]
35 90: 12,881-12,889.
36 Hoell, J. M., Jr.; Gregory, G. L.; McDougal, D. S.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.
37 (1987) Airborne intercomparison of carbon monoxide measurement techniques. J. Geophys. Res. [Atmos.]
38 92:2009-2019.
39 Hughes, E. E.; Dorko, W. D.; Scheide, E. P.; Hall, L. C.; Beilby, A. L.; Taylor, J. K. (1973) Gas generation systems
40 for the evaluation of gas detecting devices. Washington, DC: U.S. Department of Commerce, National
41 Bureau of Standards; report no. NBSIR 73-292.
42 Hughes, E. E.; Davenport, A. J.; Woods, P. T.; Zielinski, W. L., Jr. (1991) Intercomparison of a range of primary
43 gas standards of carbon monoxide in nitrogen and carbon dioxide in nitrogen prepared by the National
44 Institute of Standards and Technology and the National Physical Laboratory. Environ. Sci. Technol.
45 25:671-676.
46 Hunt, W. F., Jr. (1972) The precision associated with the sampling frequency of log-normally distributed air
47 pollutant measurements. J. Air Pollut. Control Assoc. 22: 687-691.
48 Instrumentation Laboratory. (1999) IL 682 CO-Oximeter: ready when you are. Lexington, MA: Instrumentation
49 Laboratory, Critical Care Technical Support Group.
50 Jones, R. H.; Ellicott, M. F.; Cadigan, J. B.; Gaensler, E. A. (1958) The relationship between alveolar and blood
51 carbon monoxide concentrations during breathholding. J. Lab. Clin. Med. 51: 553-564.
52 Kimmel, E. C.; Carpenter, R. L.; Reboulet, J. E.; Still, K. R. (1999) A physiological model for predicting
53 carboxyhemoglobin formation from exposure to carbon monoxide in rats. J. Appl. Physiol. 86: 1977-1983.
54 Klepeis, N. E.; Ott, W. R.; Repace, J. L. (1999) The effect of cigars on indoor air quality. J. Exposure Anal.
55 Environ. Epidemiol.: in press.
October 1999 2-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 Krarup, T. (1998) The ABL™700 Oximeter: new standards in interference detection. Copenhagen, Denmark:
2 Radiometer Medical; bulletin no. 3.
3 Krarup, T. (1999) ABL 700 series [email to D. Mage]. Denmark, Copenhagen: Radiometer Medical; September 2.
4 Langan, L. (1992) Portability in measuring exposure to carbon monoxide. In: Measuring, understanding, and
5 predicting exposures in the 21st century: proceedings of the conference; November 1991; Atlanta, GA.
6 J. Exposure Anal. Environ. Epidemiol. l(suppl. 1): 223-239.
7 Larsen, R. I. (1971) A mathematical model for relating air quality measurements to air quality standards. Research
8 Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; report no. AP-89.
9 Available from: NTIS, Springfield, VA; PB-205277.
10 Lawson, D. R.; Groblicki, P. J.; Stedman, D. H.; Bishop, G. A.; Guenther, P. L. (1990) Emissions from in-use
11 motor vehicles in Los Angeles: a pilot study of remote sensing and the inspection and maintainence program.
12 J. Air Waste Manage. Assoc. 40: 1096-1105.
13 Lee, K.; Yanagisawa, Y. (1992) Development of a sampler for carbon monoxide in expired air. Presented at:
14 85th annual meeting and exhibition of the Air & Waste Management Association; June; Kansas City, MO.
15 Pittsburgh, PA: Air & Waste Management Association; paper no. 92-145.03.
16 Lee, K.; Yanagisawa, Y. (1995) Sampler for measurement of alveolar carbon monoxide. Environ. Sci. Technol.
17 29: 104-107.
18 Lee, P. S.; Majkowski, R. F.; Perry, T. A. (1991) Tunable diode laser spectroscopy for isotope analysis - detection
19 of isotopic carbon monoxide in exhaled breath. IEEE Trans. Biomed. Eng. 38: 966-973.
20 Lee, K.; Yanagisawa, Y.; Spengler, J. D. (1992a) Comparison of carbon monoxide measurement methods.
21 In: Ozone: papers from the 85th Annual Meeting & Exhibition, Air & Waste Management Association; June;
22 Kansas City, MO. [Proceedings of Air & Waste Management Association Annual Meeting; 1992, v. 14,
23 92-80.04].
24 Lee, K.; Yanagisawa, Y.; Hishinuma, M.; Spengler, J. D.; Billick, I. H. (1992b) A passive sampler for measurement
25 of carbon monoxide using a solid adsorbent. Environ. Sci. Technol. 26: 697-702.
26 Lee, K.; Yanagisawa, Y.; Spengler, J. D.; Billick, I. H. (1992c) Measurement of personal carbon monoxide
27 exposures by mailed passive sampler. J. Air Waste Manage. Assoc. 42: 1212-1213.
28 Lee, P. S.; Schreck, R. M.; Hare, B. A.; McGrath, J. J. (1994) Biomedical applications of tunable diode laser
29 spectrometry: correlation between breath carbon monoxide and low level blood carboxyhemoglobin
30 saturation. Ann. Biomed. Eng. 22: 120-125.
31 Lloyd, W. G.; Rowe, D. R. (1999) Estimation of carboxyhemoglobin. Environ. Sci. Technol. 33: 782-785.
32 Mahoney, J. J.; Vreman, H. J.; Stevenson, D. K.; Van Kessel, A. L. (1993) Measurement of carboxyhemoglobin and
3 3 total hemoglobin by five specialized spectrophotometers (CO-oximeters) in comparison with reference
34 methods. Clin. Chem. (Washington, DC) 39: 1693-1700.
35 Marshall, M. D.; Kales, S. N.; Christiani, D. C.; Goldman, R. H. (1995) Are reference intervals for
36 carboxyhemoglobin appropriate? A survey of Boston area laboratories. Clin. Chem. (Washington, DC)
37 41: 1434-1438.
38 McBride, S. J.; Ferro, A.; Ott, W.; Switzer, P.; Hildemann, L. M. (1999) Investigations of the proximity effect for
39 pollutants in the indoor environment. J. Exposure Anal. Environ. Epidemiol.: accepted.
40 McCullough, J. D.; Crane, R. A.; Beckman, A. O. (1947) Determination of carbon monoxide in air by use of red
41 mercuric oxide. Anal. Chem. 19: 999-1002.
42 Morales-Morales, R. (1998) Carbon monoxide, ozone, and hydrocarbons in the Baltimore metropolitan area
43 [dissertation]. College Park, MD: University of Maryland. Available from: University Microfilms
44 International, Ann Arbor, MI; publication no. 9908993.
45 Moser, L.; Schmid, O. (1914) Die Bestimmung des Kohlenoxyds durch gelbes quecksilberoxyd [The determination
46 of carbon monoxide through yellow mercury oxide]. Z. Anal. Chem. 53: 217-233.
47 Moy, L. A.; Russell, R.; Ryan, W. F. (1994) Relationship between back trajectories and tropospheric trace gas
48 concentrations in rural Virginia. Atmos. Environ. 28: 2789-2800.
49 Mueller, R. H. (1954) A supersensitive gas detector permits accurate detection of toxic or combustible gases in
50 extremely low concentrations. Anal. Chem. 26: 39A-42A.
51 National Bureau of Standards. (1975) Catalog of NBS standard reference materials, 1975-76 edition. Washington,
52 DC: U.S. Department of Commerce, National Bureau of Standards; NBS special publication no. 260.
53 Novelli, P. C.; Elkins, J. W.; Steele, L. P. (1991) The development and evaluation of a gravimetric reference scale
54 for measurements of atmospheric carbon monoxide. J. Geophys. Res. [Atmos.] 96: 13,109-13,121.
October 1999 2-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 Novell!, P. C.; Collins, J. E., Jr.; Myers, R. C.; Sachse, G. W.; Scheel, H. E. (1994) Reevaluation of the
2 NOAA/CMDL carbon monoxide reference scale and comparisons with CO reference gases at
3 NASA-Langley and the Fraunhofer Institut. J. Geophys. Res. [Atmos.] 99: 12,833-12,839.
4 Novelli, P. C.; Connors, V. S.; Reichle, H. G., Jr.; Anderson, B. E.; Brenninkmeijer, C. A. M.; Brunke, E. G.;
5 Doddridge, B. G.; Kirchhoff, V. W. J. H.; Lam, K. S.; Masarie, K. A.; Matsuo, T.; Parrish, D. D.; Scheel,
6 H. E.; Steele, L. P. (1998) An internally consistent set of globally distributed atmospheric carbon monoxide
7 mixing ratios developed using results from an intercomparison of measurements. J. Geophys. Res. [Atmos.]
8 103: 19,285-19,293.
9 O'Keeffe, A. E.; Ortman, G. C. (1966) Primary standards for trace gas analysis. Anal. Chem. 38: 760-763.
10 Ott, W. R.; Mage, D. T. (1975) Random sampling as an inexpensive means for measuring average annual air
11 pollutant concentrations in urban areas. Presented at: 68th annual meeting of the Air Pollution Control
12 Association; June; Boston, MA. Pittsburgh, PA: Air Pollution Control Association; paper no. 75-14.3.
13 Ott, W. R.; Rodes, C. E.; Drago, R. J.; Williams, C.; Burmann, F. J. (1986) Automated data-logging personal
14 exposure monitors for carbon monoxide. J. Air Pollut. Control Assoc. 36: 883-887.
15 Ott, W.; Switzer, P.; Willits, N. (1994) Carbon monoxide exposures inside an automobile traveling on an urban
16 arterial highway. J. Air Waste Manage. Assoc. 44: 1010-1018.
17 Ott, W. R.; Vreman, H. J.; Switzer, P.; Stevenson, D. K. (1995) Evaluation of electrochemical monitors for
18 measuring carbon monoxide concentrations in indoor, in-transit, and outdoor microenvironments.
19 In: Measurement of toxic and related air pollutants: proceedings of an international symposium; May;
20 Research Triangle Park, NC. Pittsburgh, PA: Air & Waste Management Association; pp. 172-177.
21 (A&WMA publication VIP-50).
22 Palanos, P. N. (1972) A practical design for an ambient carbon monoxide mercury replacement analyzer.
23 Anal. Instrum. 10: 117-125.
24 Parrish, D. D.; Holloway, J. S.; Fehsenfeld, F. C. (1994) Routine, continuous measurement of carbon monoxide
25 with parts per billion precision. Environ. Sci. Technol. 28: 1615-1618.
26 Paulsell, C. D. (1976) Use of the National Bureau of Standards standard reference gases in mobile source emissions
27 testing. In: Calibration in air monitoring: a symposium; August 1975; Boulder, CO. Philadelphia, PA:
28 American Society for Testing and Materials; pp. 232-245; ASTM special technical publication 598.
29 Phinney, D. E.; Newman, J. E. (1972) The precision associated with the sampling frequencies of total paniculate at
30 Indianapolis, Indiana. J. Air Pollut. Control Assoc. 22: 692-695.
31 Porter, K.; Volman, D. H. (1962) Flame ionization detection of carbon monoxide for gas chromatographic analysis.
32 Anal. Chem. 34: 748-749.
33 Poulida, O.; Dickerson, R. R.; Doddridge, B. G.; Holland, J. Z.; Wardell, R. G.; Watkins, J. G. (1991) Trace gas
34 concentrations and meteorology in rural Virginia. 1. Ozone and carbon monoxide. J. Geophys. Res. [Atmos.]
35 96:22,461-22,475.
36 Rhodes, R. C.; Evans, E. G. (1987) Precision and accuracy assessments for state and local air monitoring networks
37 1985. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring
3 8 Systems Laboratory; EPA report no. EPA-600/4-87-003. Available from: NTIS, Springfield, VA;
39 PB87-145447.
40 Robbins, R. C.; Borg, K. M.; Robinson, E. (1968) Carbon monoxide in the atmosphere. J. Air Pollut. Control
41 Assoc. 18: 106-110.
42 Rodgers, P. A.; Vreman, H. J.; Dennery, P. A.; Stevenson, D. K. (1994) Sources of carbon monoxide (CO) in
43 biological systems and applications of CO detection technologies. Semin. Perinatal. 18: 2-10.
44 Roths, J.; Zenker, T.; Parchatka, U.; Wienhold, F. G.; Harris, G. W. (1996) Four-laser airborne infrared
45 spectrometer for atmospheric trace gas measurements. Appl. Opt. 35: 7075-7084.
46 Ryan, W. F.; Doddridge, B. G.; Dickerson, R. R.; Morales, R. M.; Hallock, K. A.; Roberts, P. T.; Blumenthal,
47 D. L.; Anderson, J. A.; Civerolo, K. (1998) Pollutant transport during a regional O3 episode in the
48 mid-Atlantic states. J. Air Waste Manage. Assoc. 48: 786-797.
49 Sachse, G. W.; Hill, G. F.; Wade, L. O.; Perry, M. G. (1987) Fast-response, high-precision carbon monoxide sensor
50 using a tunable diode laser absorption technique. J. Geophys. Res. [Atmos.] 92: 2071-2081.
51 Scaringelli, F. P.; Rosenberg, E.; Rehme, K. A. (1970) Comparison of permeation devices and nitrite ion as
52 standards for the colorimetric determination of nitrogen dioxide. Environ. Sci. Technol. 4: 924-929.
53 Seiler, W.; Giehl, H.; Roggendorf, P. (1980) Detection of carbon monoxide and hydrogen by conversion of mercury
54 oxide to mercury vapor. Atmos. Technol. (12): 40-45.
October 1999 2-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
2 New York, NY: John Wiley & Sons, Inc.
3 Shepherd, A. P.; McMahan, C. A. (1996) Role of oximeter error in the diagnosis of shunts. Catheterization
4 Cardiovasc. Diagn. 37: 435-446.
5 Smith, N. J. (1977) End-expired air technic for determining occupational carbon monoxide exposure. J. Occup.
6 Med. 19: 766-769.
7 Smith, F.; Nelson, A. C., Jr. (1973) Guidelines for development of a quality assurance program: reference method
8 for the continuous measurement of carbon monoxide in the atmosphere. Research Triangle Park, NC: U.S.
9 Environmental Protection Agency, Quality Assurance and Environmental Monitoring Laboratory; EPA
10 report no. EPA-R4-73-028a. Available from: NTIS, Springfield, VA; PB-222512.
11 Smith, R. G.; Bryan, R. J.; Feldstein, M; Locke, D. C.; Warner, P. O. (1975) Tentative method for constant
12 pressure volumetric gas analysis for O2, CO2, CO, N2, hydrocarbons (ORSAT). Health Lab. Sci. 12: 177-181.
13 Stedman, D. H.; Bishop, G. A. (1996) Measuring the emissions of passing cars. Ace. Chem. Res. 29: 489-495.
14 Stedman, D. H.; Kok, G.; Delumyea, R.; Alvord, H. H. (1976) Redundant calibration of nitric oxide, carbon
15 monoxide, nitrogen dioxide, and ozone air pollution monitors by chemical and gravimetric techniques.
16 In: Calibration in air monitoring. Philadelphia, PA: American Society for Testing and Materials; pp. 337-344.
17 (ASTM special technical publication 598).
18 Stedman, D. H.; Bishop, G. G.; Aldrete, P.; Slott, R. S. (1997) On-road evaluation of an automobile emission test
19 program. Environ. Sci. Technol. 31: 927-931.
20 Stedman, D. H.; Bishop, G. A.; Slott, R. S. (1998) Repair avoidance and evaluating inspection and maintenance
21 programs. Environ. Sci. Technol. 32: 1544-1545.
22 Stevenson, D. K.; Vreman, H. J. (1997) Carbon monoxide and bilirubin production in neonates. Pediatrics
23 100:252-254.
24 Swinnerton, J. W.; Linnenbom, V. J.; Cheek, C. H. (1968) A sensitive gas chromatographic method for determining
25 carbon monoxide in seawater. Limnol. Oceanogr. 13: 193-195.
26 Tesarik, K.; Krejci, M. (1974) Chromatographic determination of carbon monoxide below the 1 ppm level.
27 J. Chromatogr. 91: 539-544.
28 U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
29 quality standards. U. S. C. 42: §§7408-7409.
30 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
31 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
32 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
33 U.S. Environmental Protection Agency. (1997) EPA traceability protocol for assay and certification of gaseous
34 calibration standards. Washington, DC: National Center for Environmental Research and Quality Assurance;
35 report no. EPA/600/R-97/121. Available from: NTIS, Springfield, VA; PB98-115314.
36 Van Dam, J.; Daenens, P. (1994) Microanalysis of carbon monoxide in blood by head-space capillary gas
37 chromatography. J. Forensic Sci. 39: 473-478.
3 8 Volz, A.; Kley, D. (1985) A resonance-fluorescence instrument for the in-situ measurement of atmospheric carbon
39 monoxide. J. Atmos. Chem. 2: 345-357.
40 Vreman, H. J.; Stevenson, D. K. (1994) Carboxyhemoglobin determined in neonatal blood with a CO-oximeter
41 unaffected by fetal oxyhemoglobin. Clin. Chem. (Washington, DC) 40: 1522-1527.
42 Vreman, H. J.; Mahoney, J. J.; Stevenson, D. K. (1993) Electrochemical measurement of carbon monoxide in
43 breath: interference by hydrogen. Atmos. Environ. Part A 27: 2193-2198.
44 Vreman, H. J.; Mahoney, J. J.; Stevenson, D. K. (1995) Carbon monoxide and Carboxyhemoglobin. Adv. Pediatr.
45 42:303-334.
46 Wallace, L. A. (1983) Carbon monoxide in air and breath of employees in an underground office. J. Air Pollut.
47 Control Assoc. 33: 678-682.
48 Wallace, L.; Thomas, J.; Mage, D.; Ott, W. (1988) Comparison of breath CO, CO exposure, and Coburn model
49 predictions in the U.S. EPA Washington-Denver (CO) study. Atmos. Environ. 22: 2183-2193.
50 Weeks, I. A.; Galbally, I. E.; Fraser, P. J.; Matthews, G. (1989) Comparison of the carbon monoxide standards used
51 at Cape Grim and Aspendale. In: Forgan, B. W.; Ayres, G. P., eds. Baseline Atmosheric Program (Australia)
52 1987. Canberra, New South Wales: Department of Administrative Services, Bureau of Meteorology;
53 pp. 21-25.
54 Weinheimer, A. J.; Ridley, B. A. (1990) A cautionary note on the use of some mass flow controllers. J. Geophys.
55 Res. [Atmos.] 95: 9817-9821.
October 1999 2-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wohlers, H. C.; Newstein, H.; Daunis, D. (1967) Carbon monoxide and sulfur dioxide adsorption on- and
2 desorption from glass, plastic, and metal tubings. J. Air Pollut. Control Assoc. 17: 753-756.
October 1999 2-33 DRAFT-DO NOT QUOTE OR CITE
-------
i 3. SOURCES, EMISSIONS, AND
2 CONCENTRATIONS OF CARBON MONOXIDE
3 IN AMBIENT AND INDOOR AIR
4
5
6 3.1 INTRODUCTION
7 This chapter summarizes current information about the abundance and distribution, the
8 nature of sources and sinks and the chemistry of carbon monoxide (CO) in various environments,
9 ranging from the global background to indoor air. Carbon monoxide is studied in these widely
10 varied environments for different reasons. In indoor environments with sources such as properly
11 operating unvented combustion sources, malfunctioning or misused combustion appliances,
12 automobile exhaust from attached garages, and cigarettes and, in outdoor environments such as
13 urban areas where emissions from motor vehicles or wood burning can cause high concentrations
14 to exist, carbon monoxide is of direct concern because of health effects resulting from human
15 exposure to these high concentrations. Human exposures to CO are discussed in Chapter 4.
16 Carbon monoxide in less polluted air is of interest because of its importance to atmospheric
17 chemistry. Carbon monoxide can affect the formation of ozone (O3) and other photochemical
18 oxidants in the atmosphere. Carbon monoxide strongly influences the abundance of hydroxyl
19 radicals (OH), thus affecting the global cycles of many biogenic and anthropogenic trace gases
20 that affect the abundance of stratospheric O3 and the energy budget of the atmosphere. Changes
21 in CO levels, therefore, may contribute to widespread changes in atmospheric chemistry and
22 indirectly affect global climate. In this chapter, the global scale aspects of CO are discussed first,
23 and then the discussion proceeds to successively smaller scales. An overview of the major
24 sources and sinks of CO and the resulting CO distribution on a global basis and the importance
25 of CO to tropospheric chemistry is presented in Section 3.2, followed by a discussion of
26 nationwide emissions of CO in Section 3.3. Nationwide trends in ambient CO levels and related
27 discussions on CO air quality are presented in Section 3.4, and concentrations and sources of CO
28 in indoor environments are discussed in Section 3.5.
29
30
October 1999 3-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.2 THE GLOBAL CYCLE OF CARBON MONOXIDE
2 The major sources and sinks of CO are summarized in Table 3-1. Examples of major
3 activities leading to the emissions of CO from each source category are shown in the second
4 column of Table 3-1. Many of these sources have natural components. As can be seen from
5 Table 3-1, CO is produced as a primary pollutant during the combustion of fossil and biomass
6 fuels. Vegetation also can emit CO directly into the atmosphere as a metabolic byproduct.
TABLE 3-1. SUMMARY OF MAJOR SOURCES AND SINKS OF
CARBON MONOXIDE
Sources and Sinks Notes
Sources
Fossil fuel combustion Transportation and coal, oil, and natural gas burning.
Biomass burning Agricultural clearing, wood and refuse burning, and forest fires.a
CH4 oxidation Wetlands,3 agriculture (rice cultivation, animal husbandry, and biomass
burning), landfills, coal mining, and natural gas and petroleum industry.
NMHC oxidation Transportation (alkanes, alkenes, and aromatic compounds) and
vegetation3 (isoprene and terpenes).
Organic matter oxidation3 Humic and other organic substances in surface waters and soils.
Vegetation3 Metabolic by-product.
Sinks
Reaction with OH radicals Hydroxyl radicals are ubiquitous scavengers of many atmospheric
pollutants.
Soil microorganisms3 Responsible microorganisms still need to be catalogued.
3Sources and sinks that have large natural components.
1 Carbon monoxide is formed as an intermediate product during the photochemical oxidation
2 of methane and non-methane hydrocarbons (NMHCs) to CO2. Major sources of methane are
3 summarized in the second column of Table 3-1. Likewise, major sources of NMHCs, whose
4 oxidation produces CO, are given. In addition, the photooxidation of organic matter in surface
5 waters (oceans, lakes, and rivers) and on the soil surface occurs. Carbon monoxide is lost
6 primarily by reaction with atmospheric OH radicals and by uptake by soil microorganisms.
October 1999 3-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 More detailed descriptions of the nature of individual sources of primary CO shown in
2 Table 3-1 and estimates of the strengths of these sources, along with similar material for
3 nonchemical sinks of CO, are given in Section 3.2.2. Carbon monoxide concentrations and
4 trends in the background atmosphere are discussed in Section 3.2.1, and the atmospheric
5 chemistry of CO, including the formation of secondary CO, is discussed in Section 3.2.3.
6
7 3.2.1 Global Background Concentrations of Carbon Monoxide
8 In common usage, the term "background levels" refers to concentrations observed in
9 remote areas relatively unaffected by local pollution sources. However, several definitions of
10 background levels are possible (see Chapter 6, U.S. Environmental Protection Agency, 1996).
11 The two definitions chosen in that document as being most relevant for regulatory purposes and
12 for providing corrections to assessments of the health risks posed by exposure to CO are based
13 on estimates of contributions from uncontrollable sources that can affect CO levels in the United
14 States. The first definition includes anthropogenic and natural sources outside North America,
15 and natural sources within North America. The second definition includes only natural sources
16 within and outside North America. These background levels refer to concentrations that would
17 be present because of the presence of these sources alone. Because of long-range transport from
18 anthropogenic source regions in North America, it may be impossible to obtain background
19 levels defined above solely on the basis of direct measurement. However, some inferences about
20 what these levels may be can be made with the help of numerical models and historical data.
21 Surface measurements of CO concentrations are made routinely as part of the National
22 Oceanic and Atmospheric Administration's Climate Monitoring Diagnostics Laboratory
23 (NOAA/CMDL) Global Cooperative Air Sampling Network (e.g., Hofmann et al., 1996).
24 Carbon monoxide flask samples are collected weekly in flasks or continuously with in situ gas
25 chromatographs at about 40 remote sites around the world. These sites are located primarily in
26 the marine boundary layer, with a few located in continental areas. The latitudinal and seasonal
27 variations in CO levels are summarized in the three-dimensional diagram shown in Figure 3-1
28 (National Oceanic and Atmospheric Administration, 1999). Annual average CO mixing ratios
29 are about 120 ppb in the Northern Hemisphere and about 40 ppb in the Southern Hemisphere.
30 Seasonal maxima in CO mixing ratios occur during late winter in both hemispheres, and minima
October 1999 3-3 DRAFT-DO NOT QUOTE OR CITE
-------
2.50 -,
Figure 3-1. Latitudinal and seasonal variability in CO concentrations obtained in the
NOAA/CMDL monitoring network.
Source: National Oceanic and Atmosphere Administration (1999a).
1 occur during late summer, with about a factor of two variation between maximum and minimum
2 values. Carbon monoxide is well mixed in high latitudes of both the Northern Hemisphere and
3 the Southern Hemisphere. A steep gradient in CO mixing ratios exists between about 30° north
4 (N) latitude and about 10° south (S) latitude. Carbon monoxide concentrations range from a
5 minimum of about 30 ppb during summer in the Southern Hemisphere to about 200 ppb at high
6 latitudes in the Northern Hemisphere during winter. Thus, CO concentrations in remote areas of
7 the Northern Hemisphere are only a small fraction (~ 1 to 2%) of those of concern to human
8 health (as given by the National Ambient Air Quality Standards [NAAQS] for CO of 9 ppm for
9 the second highest, nonoverlapping 8-h average concentration).
October 1999
5-4
DRAFT-DO NOT QUOTE OR CITE
-------
1 There were sufficient data on tropospheric air quality trends to suggest that CO
2 concentrations measured at global background sites were increasing at 1.2 ± 0.6% per year from
3 1981 to 1986, based on data collected by the Oregon Graduate Institute (OGI), this increase
4 presumably resulted from anthropogenic activities (Khalil and Rasmussen, 1988a). From 1987
5 until 1992, global background levels of CO declined at a rate of about -2.6 ± 0.8% per year
6 (Khalil and Rasmussen, 1994), whereas Novell! et al. (1994) determined a rate of decrease in CO
7 of -6.5 ± 0.8% per year from 1990 to 1993 and later results reported by NOAA/CMDL indicate
8 that background levels of CO continued to decline although at a lower rate of -2.6 ± 0.2% per
9 year from 1990 to 1995 (Novell! et al., 1998). Hypotheses to explain these observations include
10 reductions in fossil fuel combustion (Bakwin et al., 1994; Novell! et al., 1994; Khalil and
11 Rasmussen, 1994) and tropical biomass burning (Yung et al., 1999). Possible increases in
12 tropospheric OH levels resulting from enhanced transmission of solar ultraviolet radiation caused
13 by stratospheric O3 depletion may have been an additional factor (Fuglestvedt et al., 1994; Bekki
14 et al., 1994). More recent data for changes in global background CO levels are shown in
15 Figure 3-2. The data from 1993 to 1997 do not show clearly any stable upward or downward
16 trend.
17 Because direct measurements of sufficient precision for defining trends have come into use
18 only within the last 15 to 20 years, estimates of longer term trends in CO levels must come from
19 indirect means. Rinsland and Levine (1985) derived an increase in the mean tropospheric CO
20 abundance of about 2% per year from 1950 to 1984, based on an examination of solar spectra
21 captured on photographic plates in Europe. Column CO abundances obtained over Zvenigorod,
22 Russia from 1974 through 1997 have increased by about 1% per year (Yurganov et al., 1999),
23 while measurements obtained with a similar technique over the Alps have indicated decreases in
24 CO of-0.18 ±0.16% per year from 1984 to 1995 compared to a change of -0.95 ± 0.32% per
25 year from 1984 to 1993. The difference arises mainly from a pronounced minimum during the
26 second half of 1992 and 1993 (depending on the fitting function (Mahieu et al., 1997). Thus,
27 there is still considerable uncertainty in defining global trends for CO based on differences in
28 trends found in specific regions.
29 Carbon monoxide concentrations measured in air bubbles trapped in the ice sheets of
30 Greenland and Antarctica have been used as proxies for CO concentrations in ambient air at the
October 1999 3-5 DRAFT-DO NOT QUOTE OR CITE
-------
.a
o.
O
O
125
100
75
.Q
a_
O
O
0
-4
-8
Global Average
Global Growth Rate
1990 1991 1992 1993 1994 1995 1996 1997
Year
Figure 3-2. Global background average CO levels (upper) and growth rates for global
background average CO (lower).
Source: National Oceanic and Atmospheric Administration (1999b).
October 1999
DRAFT-DO NOT QUOTE OR CITE
-------
1 time the air bubbles were sealed from the atmosphere (Haan et al., 1996). Carbon monoxide
2 concentrations derived this way for the preindustrial era (roughly corresponding to the year 1850,
3 when anthropogenic activities should not have influenced significantly the atmospheric
4 composition) are about 90 ppb for the high-latitude Northern Hemisphere and are about 50 ppb
5 for the high-latitude Southern Hemisphere. Some enhancement of Northern Hemispheric values
6 over Southern Hemispheric values during the preindustrial era is likely because of the greater
7 mass of vegetation that can emit NMHCs in the Northern Hemisphere. However, it should be
8 noted that the CO in the trapped air bubbles also may result from the decomposition of organic
9 compounds also trapped in the same air bubbles, and that it is difficult to extract CO from the air
10 bubbles without contamination. Both factors tend to cause positive artifacts in the CO
11 concentrations reported. In addition, the Northern Hemispheric value derived from the ice cores
12 is higher than that predicted by atmospheric model studies of the preindustrial era that indicate
13 CO mixing ratios of about 50 ppb (Thompson and Cicerone, 1986; Pinto and Khalil, 1991;
14 Thompson etal., 1993).
15
16 3.2.2 Sources and Global Emissions Estimates of Carbon Monoxide
17 Global CO emission estimates are summarized in Table 3-2. Motor vehicles contribute
18 most of the emissions from fossil fuel combustion on the global scale according to the entries by
19 Logan et al. (1981) and Dignon et al. (1998). Bradley et al. (1999) estimated global emissions
20 from motor vehicles of 213 Tg/year in 1991 based on roadside remote sensing measurements
21 around the world. They also calculated a decrease of 17% in global motor vehicle emissions
22 from 1991 to 1995. Their estimated uncertainty in both figures is about 20%. Variables
23 controlling the formation of CO during combustion of any fuel are oxygen concentration, flame
24 temperature, gas residence time at high temperature, and mixing in the combustion zone.
25 In general, increases in all four factors result in lower amounts of CO produced relative to carbon
26 dioxide (CO2). Carbon monoxide is produced primarily during conditions of incomplete
27 combustion. The estimates for fossil fuel emissions by stationary sources, shown in the footnotes
28 to Table 3-2, do not include significant contributions from power plants, because fuels are burned
29 with high efficiency in modern power plants. Rather, they are based on estimates of CO emitted
30 in small, hand-fired furnaces used for domestic purposes (e.g., cooking, heating, water
31 sterilization) and in inefficient boilers and furnaces used in small-scale industrial operations.
October 1999 3-7 DRAFT-DO NOT QUOTE OR CITE
-------
0
£
VO
VO
VO
oo
DRAFT-DO NOT
0
0
H
W
O
O
H
W
TABLE 3-2. ANNUAL GLOBAL CARBON MONOXIDE EMISSIONS ESTIMATES
Allen et al.
(1996)
Sources
Fossil fuel combustion 329
Biomass burning 370
Natural NMHC oxidation 618
Anthropogenic NMHC oxidation —
Methane oxidation 722
Oceans —
Soils —
Vegetation —
Total 2,039
Sinks
Soils
OH reaction
Total
(in teragrams [Tg] per year)
Logan et al. Seiler and Conrad Pacyna and Graedel
(1981) (1987) (1995)
450a 640 ± 200 440 ± 150
655 1,000 ± 600 700 ± 200
560 900 ± 500 800 ± 400
90 — —
810 600 ±300 600 ±200
40 100 ±90 50 ± 40
— — —
130 75 ±25 75 ±25
2,735 3,315 ± 1,700 2,700 ± 1,000
Dignon et al.
(1998)
600b(2.1)c
600 (2.7)
300 (2.3)
200 (2.3)
600 (2)
10
30
200 (4)
2,500 (1.5)
300 (3)
2,300 (1.4)
2,600
"Estimate includes 150 Tg/year from stationary sources.
bEstimate includes 100 Tg/year from stationary sources.
°Values in parentheses represent ratio of maximum to minimum estimate of source term.
-------
1 This latter source is of significance only in eastern Europe and in developing countries of Africa
2 and Asia (especially China). However, it also should be noted that the importance of this source
3 has been declining as heating needs are met increasingly by centralized power plants.
4 Biomass burning consists of wildfires and the burning of vegetation to clear new land for
5 agriculture and population resettlement; to control the growth of unwanted plants on pasture
6 land; to dispose of agricultural and domestic waste; and as fuel for cooking, heating, and water
7 sterilization. Most wildfires are the result of human activities with only a small fraction (10 to
8 30%) initiated by lightning (Andreae, 1991). Biomass burning exhibits strong seasonality, with
9 most biomass burned during the local dry season. The smoldering phase of combustion yields
10 higher emissions factors than the flaming phase. Lobert et al. (1991) found, in controlled
11 combustion chamber experiments with a wide variety of vegetation types, that, on average,
12 84% of CO was produced during the smoldering phase and 16% during the flaming phase of
13 combustion. Smoldering conditions are more prevalent during the burning of large pieces of
14 vegetation, such as trees, compared with grasses. Nonetheless, most CO is produced in the
15 tropics by savanna burning (mainly in Africa), followed by burning forests, fuel wood, and
16 agricultural waste. Less than 20% of the CO produced by biomass burning originates in middle
17 and high latitudes (Andreae, 1991).
18 The other sources of CO shown in Table 3-2 all have large natural components. Carbon
19 monoxide may be evolved from the photodecomposition of organic matter in surface waters
20 (such as oceans, rivers, and lakes) and the soil surface. Soils can act as a source or a sink for
21 carbon monoxide, depending on soil moisture, the intensity of sunlight reaching the soil surface,
22 and soil temperature (e.g., Inman et al., 1971; Conrad and Seller, 1985). Soil uptake of CO
23 occurs because of anaerobic bacteria (Inman et al., 1971). Emissions of CO from soils appear to
24 occur by abiotic processes, such as thermodecomposition or photodecomposition of organic
25 matter. In general, warm and moist conditions found in most soils favor CO uptake, whereas hot
26 and dry conditions as found in deserts and some savannas favor the release of CO (King, 1999).
27 The value reported for soil emissions in Table 3-2 is based on very limited data, and hence it is
28 difficult even to assign uncertainty bounds (Conrad, 1996). Moxley and Cape (1997)
29 hypothesized that from 20 to 80% of CO in the stable nocturnal boundary layer (calculated
30 heights between 40 and 220 m) could have been depleted by soil microorganisms during
31 transport inland 100 km from the Scottish coast.
October 1999 3-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Estimates of the magnitude of the soil sink range from 250 to 640 Tg/year (Logan et al.,
2 1981; Cicerone, 1988), with a current "best" estimate of 300 Tg/year, with an uncertainty range
3 of a factor of three (Dignon et al., 1998). More extensive field measurements, perhaps based on
4 the eddy correlation technique (Ritter et al., 1994), are needed to characterize the variability and
5 the direction of the CO flux to the soil surface. Most CO in the atmosphere is lost by its
6 oxidation to CO2 by OH radicals. Reaction with OH radicals accounts for a loss of
7 2,300 Tg/year, with an uncertainty factor of 1 .4 (Dignon et al., 1998). Because of large
8 uncertainties in individual sources and sinks, the imbalance between sources (2,500 Tg/year) and
9 sinks (2,600 Tg/year) is not significantly different from zero. By using a mean value of 80% for
10 biomass burning produced by human activity and a value of two-thirds for the fraction of CH4
1 1 produced by human activity (Houghton et al., 1992), it can be seen that approximately two-thirds
12 of CO is produced globally as the result of human activities.
13
14 3.2.3 The Atmospheric Chemistry of Carbon Monoxide
1 5 Carbon monoxide is produced by the photooxidation of CH4 and other organic compounds
16 (including NMHCs) in the atmosphere and of organic molecules in surface waters and soils
17 (Table 3-1 and 3-2). Estimates of CH4 emissions from the various source categories shown in
18 Table 3-1 can be found in the Intergovernmental Panel on Climate Change report (Houghton
19 et al., 1992). Methane oxidation can be summarized by the following sequence of reactions:
20
21 CH4 + OH - CH3 + H2O
22 CH3 + O2 (+M) - CH3O2
23 CH3O2 + NO - CH3O + NO2
24 CH3O + O2 - CH2O + HO2
25 CH2O + hu - H2 + CO,
26 or CH2O + hu - HCO + H,
27 or CH2O + OH - HCO + H2O
28 HCO + O2 - CO + HO2,
29
30 where M is a mediator (e.g., nitrogen, O2, argon, CO2). The photolysis of formaldehyde (CH2O)
3 1 proceeds by two pathways, the first yields molecular hydrogen (H2) plus CO (55%), and the
October 1999 3-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 second yields atomic hydrogen (H) plus the formyl radical (HCO) (45%), where the percentages
2 are given for overhead sun conditions (Rogers, 1990). Formyl radicals then react with molecular
3 oxygen (O2) to form the hydroperoxy radical (HO2) plus CO. In addition, the reaction of the
4 methyl peroxy radical (CH3O2) with HO2 radicals, forming methyl hydroperoxide (CH3OOH),
5 needs to be considered, especially in low nitrogen oxide (NOX) environments. The heterogeneous
6 removal of soluble intermediate products, such as CH3OOH, CH2O, and radicals, decreases the
7 yield of CO from the oxidation of CH4.
8 Although the oxidation of CH2O nearly always results in CO formation (except for the
9 formation of small quantities of formic acid in the reaction of CH2O with HO2), the oxidation of
10 acetaldehyde (CH3CHO) does not always yield two CO molecules. The photolysis of CH3CHO
11 also involves pathways that produce molecules and radicals, namely CH4 + CO and CH3 + HCO.
12 Estimates of the yield of CO from the photooxidation of CH4 and CH3 are subject to the same
13 considerations outlined above. The reaction of CH3CHO with OH radicals can yield acetyl
14 radicals (CH3CO). The acetyl radicals then will participate with O2 in a termolecular
15 recombination reaction to form acetyl peroxy radicals, which then can react with nitric oxide
16 (NO) to form CH3 + CO2 (or the acetyl peroxy radicals can react with nitrogen dioxide (NO2) to
17 form peroxy acetyl nitrate [PAN]). Thus, one of the carbon atoms can be oxidized directly to
18 CO2 without passing through CO. The yield of CO depends on the OH concentration and the
19 photolysis rate of CH3CHO, as well as on the abundance of NO, as acetyl peroxy radicals also
20 can react with HO2 and other hydrogen-bearing radicals.
21 Estimates of the yield of CO from the oxidation of more complex hydrocarbons requires
22 the calculation of the yields of CH2O, CH3CHO, CH3CO, and analogous radicals from the
23 oxidation of the parent molecule. Likewise, the extent of heterogeneous removal of soluble
24 intermediate products needs to be considered in the oxidation of more complex hydrocarbons.
25 However, in contrast to simple hydrocarbons containing one or two carbon atoms, detailed
26 kinetic information is lacking about the gas phase oxidation pathways of many anthropogenic
27 hydrocarbons (e.g., aromatic compounds, such as benzene and toluene), biogenic hydrocarbons
28 (e.g., isoprene, the monoterpenes), and their intermediate oxidation products (e.g., epoxides,
29 nitrates, carbonyl compounds). As much as 30% of the carbon in hydrocarbons in many urban
30 areas is in the form of aromatic compounds (Grosjean and Fung, 1984; Seila et al., 1989). Yet
31 mass balance analyses performed on irradiated smog chamber mixtures of aromatic hydrocarbons
October 1999 3-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 indicate that only about one-half of the carbon is in the form of compounds that can be identified.
2 Reactions that have condensible products, such as those occurring during the oxidation of
3 terpenes, also need to be considered because these reactions produce secondary organic
4 particulate matter, thereby reducing the potential yield of CO.
5 The yield of CO from the oxidation of CH4 is about 0.9, and it is about 0.4 from the
6 oxidation of ethane and propane, on a per carbon basis from estimates based on atmospheric
7 model results (Kanakidou et al., 1991). Jacob and Wofsy (1990) estimated that 1 mole of CO is
8 produced by the oxidation of 1 mole of isoprene (corresponding to a conversion factor of 0.2 on a
9 per carbon basis) for low NOX levels. For higher NOX levels, they estimated that 3 moles of CO
10 are produced per mole of isoprene oxidized (corresponding to a conversion factor of 0.6 on a per
11 carbon basis). Isoprene accounts for most of the CO produced by the photochemical oxidation of
12 NMHCs shown in Table 3-2.
13 The major pathway removing CO from the atmosphere is by its reaction with OH radicals.
14 There have been numerous determinations of the rate coefficient for this reaction. The most
15 recent evaluation of kinetics data for use in atmospheric modeling (National Aeronautics and
16 Space Administration, Panel for Data Evaluation, 1997) recommends a value of 1.5 x 10"13
17 (1 + 0.6 Patm) cm3 molecules"1 s"1, with a value of 0 ± 300 K for E/R for the reaction
18
19 OH + CO - Products.
20
21 This reaction proceeds through two channels. The bimolecular channel yields H + CO2, whereas
22 the addition channel leads to the formation of a carboxyl radical (HOCO). In the presence of O2,
23 the HOCO intermediate is converted to HO2 + CO2. Therefore, for atmospheric purposes, the
24 products of the reaction OH + CO can be taken to be HO2 and CO2.
25 Estimates of OH radical concentrations can be used along with the rate coefficient given
26 above to calculate the lifetime of CO in the atmosphere. Measurements of OH radical
27 concentrations in situ (Hard et al., 1992; Mount and Williams, 1997; Poppe et al., 1994) in the
28 lower troposphere show that their levels are highly site specific and are highly variable in space
29 and time. Typical mid-latitude noontime values during summer (when OH concentrations are at
30 their highest levels) range from about 5 to 10 x 106 OH/cm3 and are much lower during other
31 times of the day and during other seasons. As a result, it is difficult to derive average values that
October 1999 3-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 would be meaningful for use in calculating the atmospheric lifetime of long-lived species that
2 react with OH radicals, based on direct measurements. Modeling the atmospheric distribution of
3 methyl chloroform (CH3CC13) has been used to derive diurnal and global average OH values for
4 calculating the atmospheric lifetimes of long-lived species by comparing predictions to
5 observations (Prinn et al., 1992). Average OH values derived in this manner are about 8 x 10s
6 OH/cm3. By further adjusting the OH fields derived in a simulation of the CH3CC13 distribution,
7 to optimize the fit between the measurements and simulations of CH3CC13 concentrations, Krol
8 et al. (1998) derived concentrations of 1.00 x 106 OH/cm3 in 1978 and 1.07 x 106 OH/cm3 in
9 1993. The resulting trend in OH values is estimated to be 0.46 ± 0.6% year'1. Krol et al. (1998)
10 also used a three dimensional model of atmospheric chemistry to examine the sensitivity of the
11 OH trend to stratospheric ozone depletion, decreases in CO emissions, increases in tropical water
12 vapor, and NOX and CH4 emissions, and derived an overall change of 6% in OH levels from 1978
13 to 1993. However, it should be noted that many of the changes used as input to the model
14 calculations are highly uncertain and thus, the results should be viewed only as a sensitivity study
15 (Law, 1999). It should also be noted that Prinn et al. (1995) had found little or no trend in OH
16 based on methychloroform data.
17 The resulting globally averaged atmospheric lifetime of CO is then approximately 2 mo.
18 Shorter lifetimes are found in the tropics, whereas longer lifetimes are found at higher latitudes.
19 During winter at high latitudes, CO is essentially inert. The CO lifetime is shorter than the
20 characteristic time scale for mixing between the hemispheres (about 1 year), and hence a large
21 gradient in concentrations can exist between the hemispheres (see Figure 3-1). In addition, the
22 chemical lifetime of CO at high latitudes is long enough to result in much smaller gradients
23 between 30° latitude and the pole of either hemisphere. However, the lifetime of CO is much
24 longer than typical residence times of CO in urban areas (assuming a diurnally averaged CO level
25 of 3 x io6 in urban areas) and in indoor environments, where OH levels are expected to be orders
26 of magnitude lower.
27 Reaction with CO, is in turn, the major reaction of OH radicals. The reaction of CO with
28 OH radicals constitutes at least 50% of the tropospheric sink of OH radicals (e.g., Collins et al.,
29 1997). Thus, changes in the abundance of CO could lead to changes in the abundance of a
30 number of trace gases whose major loss process involves reaction with OH radicals. These trace
31 gases can absorb infrared radiation from the earth's surface and contribute to the greenhouse
October 1999 3-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 effect (e.g., CH4) or can deplete stratospheric O3 (e.g., methyl chloride [CH3C1], methyl bromide
2 [CH3Br], and hydrochlorofluorocarbons, such as difluorochloromethane). Because of the
3 importance of CO in determining OH levels, interest has focused on the possible effects of
4 increases in anthropogenic CO emissions on the concentrations of gases such as those listed
5 above (Sze, 1977; Chameides et al., 1977; Thompson and Cicerone, 1986). For instance,
6 Thompson and Cicerone (1986) found in numerical simulations, in which the CO mixing ratio at
7 the surface was allowed to increase by 1% per year from 1980 to 2000, while holding CH4
8 emissions constant, that the mixing ratio of CH4 at the surface increased by about 12%
9 (corresponding to a mean increase of 0.56% per year), and that the mixing ratio of CH4 at the
10 surface increased by about 30% (corresponding to a mean increase of 1.3% per year) for an
11 increase in surface CO mixing ratio of 2% per year. However, based on the trend results reported
12 earlier, Briihl and Crutzen (1998) have examined the consequences of decreases in CO levels for
13 atmospheric chemistry. They found, by decreasing CO emissions linearly by about 20% between
14 1990 and 2000, that the CO reductions could lead to a significant decrease (-25%) in the growth
15 rate of CH4 and even to a decrease in CH4 levels in the case of constant CH4 emissions.
16 An accurate knowledge of the sources and sinks of carbon monoxide in the atmosphere is
17 therefore necessary for assessing the effects of future increases in anthropogenic CO emissions
18 on the concentrations of the above-mentioned radiatively and photochemically important trace
19 species. However, because of nonlinearities introduced into the calculation of OH radical
20 concentrations by short-lived NOX (e.g., Hameed et al., 1979), an accurate assessment of these
21 effects awaits the development of three-dimensional chemistry and transport models
22 incorporating the spatial variability of NOX (e.g., Kanakidou and Crutzen, 1993).
23 In the free troposphere, in the absence of significant quantities of NMHCs, the effects of
24 CO on tropospheric O3 can be summarized as shown below.
25
Atmospheric Reactions Leading to O3 Production
CO + OH - CO2 + H
H + O2 (+M) - HO2 (+M)
HO2 + NO - NO2 + OH
NO2 + hv - O + NO
O + O2 (+M) - O3 (+M)
Net CO + 2O2 - CO2 + O3
October 1999 3-14 DRAFT-DO NOT QUOTE OR CITE
-------
Atmospheric Reactions Leading to O3 Destruction
CO + OH - CO2 + H
H + O2 (+M) - HO2 (+M)
HO2 + O3 - OH + 2O2
Net CO + O3 - CO2 + O2
1 The oxidation of CO by OH could lead to the production or destruction of O3, depending on
2 the ratio of NO to HO2 concentrations. Based on current values of rate coefficients for the
3 reactions of HO2 with NO and O3, in regions where NO levels are greater than about 10 ppt, the
4 oxidation of CO leads to O3 formation, whereas, in areas where NO levels are less than about
5 10 ppt, the oxidation of CO leads to O3 destruction. Nitric oxide levels less than 10 ppt typically
6 are found over the tropical oceans (Carroll et al., 1990). A rough estimate of the fraction of
7 O3 production resulting from CO in the remote troposphere can be made by taking the overall
8 rate of the reaction of CO with OH radicals and then correcting for the fraction of HO2 radicals
9 that do not react with NO, based on free radical balances presented by Collins et al. (1997). This
10 quantity (i.e., the rate of conversion of NO to NO2 by HO2 radicals produced by the reaction of
11 CO with OH radicals) represents 20 to 40% of the production of O3 on a global basis.
12 The effects of CO on O3 photochemistry in environments with abundant hydrocarbons
13 (e.g., cities, tropical rain forests) require a much more complex treatment that includes the
14 competition for OH radicals by CO and NMHCs and the effects of this competition on the
15 overall budget of hydrogen-containing radicals (i.e., OH, HO2). In urban environments, reaction
16 with OH radicals represents the major loss process for NMHCs and initiates the sequence of
17 further reactions leading to the formation of O3 and CO itself. Detailed analyses of the radical
18 balances (i.e., production and loss rates in each reaction) in urban air chemistry models, as
19 performed by Jeffries (1995), can give the amount of O3 formed because of the reaction of CO.
20 However, only a few such analyses have been performed. Jeffries (1995), presented the results of
21 a numerical simulation of an O3 episode in Atlanta, GA on June 6, 1988, and found that reaction
22 with CO constituted 33% of the loss of OH radicals. It was found, by tracking sources of various
23 radicals produced by the oxidation of VOCs and that oxidize NO to NO2, that CO accounted for
24 about 17.5% of the O3 formed in this example (compared to about 82.5% for VOCs). Obviously,
25 more analyses of this sort are needed to characterize regional differences in the importance of CO
26 in different cities in the United States, which may have very different combinations of CO,
October 1999 3-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 NMHC, and NOX concentrations than those used in these examples. Because of nonlinearities in
2 the production rate of O3 involving each of the above species, caution should be exercised in
3 attempting to estimate the effects of variations in CO levels on O3 production rates in the case
4 studies cited above.
5
6
7 3.3 NATIONWIDE CARBON MONOXIDE EMISSIONS ESTIMATES
8 Total estimated primary CO emissions in the United States for the period of 1988 through
9 1997 summarized from the National Emissions Inventory Trends (NET) database are shown in
10 Table 3-3 (U.S. Environmental Protection Agency, 1998). These emissions are shown in the
11 original units used in their calculation (i.e., thousands of short tons per year) and with the same
12 number of significant figures. A short ton is equal to 2,000 Ib or 9.08 x 10s g. Table 3-3 shows
13 that total CO emissions decreased by 24.7% from 1988 to 1997; however, the fractional
14 contribution of transportation (the major source of CO both then and now) remained relatively
15 constant at 77%. The term "transportation" includes both onroad and nonroad sources. Onroad
16 sources consist of automobiles, trucks, and buses. Nonroad sources consist of categories such as
17 lawn equipment, construction equipment, trains, aircraft, boats, ships, recreational equipment,
18 logging equipment, and agricultural equipment. From 1988 to 1997, the contribution of onroad
19 sources decreased from 61 to 57%, while the contribution of nonroad sources increased from
20 13 to 19%. In addition, there are several categories, such as fuel consumption by electric utilities
21 and industry, in which emissions have increased over the same period. Total CO emissions for
22 the United States were reported to be 66,189 thousand short tons for 1990, the last year reported
23 in the previous air quality criteria document (AQCD) for CO. It can be seen from inspection of
24 Table 3-3 that the values for 1990 have been revised upward in the interim to 95,794 thousand
25 short tons. The upward revision in values for 1990 is primarily the result of changes in the
26 methods for calculating motor vehicle emissions. The MOBILES emissions factor model (U.S.
27 Environmental Protection Agency, 1993) replaced the earlier MOBILE4.1 version (U.S.
28 Environmental Protection Agency, 199la). The most significant change was in using EVI240 data
29 to replace FTP testing of recruited in-use vehicles. Exhaust emission rates were found to
30 increase significantly, especially for older higher mileage vehicles. In addition, changes were
31 made in methods for calculating inputs to the model (e.g., temperatures, operating mode) and
October 1999 3-16 DRAFT-DO NOT QUOTE OR CITE
-------
October ]
VO
VO
VO
OJ
1
^
o
^
H
1
O
O
2|
0
H
O
Cj
O
w
0
o
1 — I
h-1
TABLE 3-3. U.S. NATIONWIDE CARBON MONOXIDE
(thousands of short tons)
Source Category
Fuel Combustion Electrical Utility3
Fuel Combustion Industrial1"
Fuel Combustion Residential, Commercial,
Institutional0
Commercial/Institutional Coal
Commercial/Institutional Oil
Commercial/Institutional Gas
Miscellaneous Fuel Combustion (except residential)
Residential Wood (fireplaces, woodstoves)
Residential Other
Chemical and Allied Product Manufacturing11
Ferrous and Nonferrous Metal Processing6
Petroleum and Related Industries'
Other Industrial Processes5
Solvent Utilization
Storage and Transport
Waste Disposal and Recycling
Incineration
Conical wood burner
Municipal incinerator
Industrial
Commercial/institutional
1988
314
669
6,390
15
18
47
55
6,086
168
1,917
2,101
441
711
2
56
1,806
903
19
35
10
38
1989
321
672
6,450
15
17
49
55
6,161
153
1,925
2,132
436
716
2
55
1,747
876
19
35
9
39
1990
363
879
4,269
14
18
44
149
3,781
262
1,183
2,640
333
537
5
76
1,079
372
6
16
9
19
1991
349
920
4,587
14
17
44
141
4,090
281
1,127
2,571
345
548
5
28
1,116
392
7
17
10
20
1992
350
955
4,849
15
18
51
141
4,332
292
1,112
2,496
371
544
5
17
1,138
404
6
15
10
21
EMISSIONS
1993
363
1,043
4,181
15
18
53
143
3,679
274
1,093
2,536
371
594
5
51
1,248
497
6
14
87
21
1994
370
1,041
4,108
15
18
54
147
3,607
268
1,171
2,475
338
600
5
24
1,225
467
6
14
48
21
1995
372
1,056
4,506
15
19
54
145
3,999
273
1,223
2,380
348
624
6
25
1,185
432
6
15
10
21
1996
394
1,072
4,513
15
19
54
163
3,993
269
1,223
2,378
348
635
6
25
1,203
443
6
15
10
22
1997
406
1,110
3,301
16
19
56
168
2,278
264
1,287
2,465
364
663
6
26
1,242
467
6
16
11
23
-------
October 1999
V°
oo
o
£
H
6
o
0
H
O
o
H
W
O
O
HH
H
W
TABLE 3-3 (cont'd). U.S. NATIONWIDE CARBON
(thousands of short tons)
Source Category
Waste Disposal and Recycling (cont'd)
Residential
Other
Open Burning
Industrial
Commercial/institutional
Residential
Other
On-Road Vehicles
Light-Duty Gas Vehicles and Motorcycles
Light-duty gas vehicles
Motorcycles
Light-Duty Gas Trucks
Heavy-Duty Gas Vehicles
Diesels
Heavy-duty diesel vehicles
Light-duty diesel vehicles
Non-Road Engines and Vehicles
Non-Road Gasoline
Recreational
Construction
1988
800
2
903
21
4
877
NA
71,081
45,553
45,367
186
17,133
7,072
1,322
1,290
32
14,698
12,464
318
401
1989
773
2
870
21
5
845
NA
66,050
42,234
42,047
187
15,940
6,506
1,369
1,336
34
14,820
12,537
321
398
1990
294
27
706
14
46
509
137
57,848
37,407
37,198
209
13,816
5,360
1,265
1,229
36
15,376
13,088
359
355
1991
312
26
722
14
18
516
144
62,074
40,267
40,089
177
15,014
5,459
1,334
1,298
36
15,368
13,065
365
329
MONOXIDE EMISSIONS
1992
324
28
731
15
50
523
144
59,859
39,370
39,190
180
14,567
4,569
1,352
1,315
37
15,652
13,305
370
334
1993
340
29
749
15
52
529
153
60,202
39,163
38,973
190
15,196
4,476
1,367
1,328
38
15,828
13,454
374
348
1994
347
30
755
15
54
533
153
61,833
37,507
37,312
195
17,350
5,525
1,451
1,411
38
16,050
13,638
378
382
1995
351
29
750
15
52
536
147
54,106
33,701
33,500
200
14,829
4,123
1,453
1,412
39
16,271
13,805
382
393
1996
360
30
757
16
53
539
149
53,262
28,732
28,543
189
19,271
3,766
1,493
1,453
35
16,409
13,935
386
400
1997
380
31
772
16
55
545
156
50,257
27,036
26,847
189
18,364
3,349
1,508
1,468
35
16,755
14,242
389
423
-------
October 1999
OJ
1
VO
O
>
H
6
o
0
H
O
o
s
0
o
a
TABLE 3-3 (cont'd). U.S. NATIONWIDE CARBON MONOXIDE EMISSIONS
(thousands of short tons)
Source Category
Non-Road Engines and Vehicles (cont'd)
Industrial
Lawn and garden
Farm
Light commercial
Logging
Airport service
Recreational marine vessels
Non-Road Diesel
Recreational
Construction
Industrial
Lawn and garden
Farm
Light commercial
Logging
Airport service
Railway maintenance
Recreational marine vessels
Aircraft
Marine Vessels
Railroads
1,207
5,866
92
3,219
31
144
1,185
1,129
3
634
150
23
176
44
58
35
2
4
931
56
118
1,227
5,929
63
3,223
33
147
1,195
1,149
3
655
148
25
177
45
58
31
2
4
955
59
121
1988
1,387
6,501
213
2,428
32
116
1,698
1,180
3
677
146
27
178
46
58
38
2
4
904
83
121
1989
1,350
6,599
170
2,385
33
114
1,720
1,207
3
699
146
30
179
48
58
38
2
4
888
87
120
1990
1,374
6,684
199
2,453
34
118
1,739
1,236
3
721
147
32
180
49
57
38
2
5
901
85
125
1991
1,371
6,770
209
2,472
34
119
1,757
1,268
3
744
149
35
181
51
57
40
3
5
905
81
120
1992
1,404
6,823
175
2,551
36
121
1,769
1,300
3
766
152
38
183
53
56
41
3
5
915
82
114
1993
1,436
6,895
145
2,621
40
129
1,763
1,329
4
788
155
41
184
54
56
39
3
5
942
82
114
1994
1,446
6,949
150
2,658
41
131
1,775
1,330
3
789
156
44
182
56
52
40
3
5
949
82
112
1995
1,510
7,009
152
2,787
44
141
1,788
1,301
3
768
154
47
176
56
45
43
3
5
1,012
85
115
-------
a
o
cr
VO
VO
VO
OJ
K>
O
IAIXLJ^ j-j ^coni uj. u.a. r>Ai«jr> \\IVJL \^AKJ>\JW WL\JW\JA.LV& j^ivnaaujrMa
(thousands of short tons)
Source Category
Miscellaneous
Other Combustion
Structural fires
Agricultural fires
Slash/prescribed burning
Forest wildfires
Other
Total All Sources
Major Subcategories:
"Coal Burning
bl\Tntnrfl1 rrn« Riimina
1988
15,895
15,895
242
612
4,332
10,709
NA
116,081
1989
8,153
8,153
242
571
4,332
3,009
NA
103,480
1990
11,208
11,207
164
415
4,668
5,928
32
95,794
1991
8,751
8,751
166
413
4,713
3,430
28
97,790
1992
7,052
7,052
168
421
4,760
1,674
30
94,400
1993
7,013
7,013
169
415
4,810
1,586
34
94,526
1994
9,614
9,613
170
441
4,860
4,114
28
98,854
1995
7,050
7,049
171
465
4,916
1,469
28
89,151
1996
9,463
9,462
142
475
4,955
3,863
27
90,929
1997
9,568
9,568
143
501
5,033
3,863
28
87,451
H
6
o
2
o
H
O
'Residential wood burning
"tarbon black manufacturing
Terrous metal production
'Refineries
gWood, paper, and pulp
Notes:
NA = not available. For several source categories, emissions either prior to or beginning with 1985 are not available at the more detailed level but are contained in the more
aggregate estimate. "Other" categories may contain emissions that could not be allocated accurately to specific source categories. To convert emissions to gigagrams
(thousands of metric tons), multiply the above values by 0.9072.
Source: Adapted from U.S. Environmental Protection Agency (1998).
o
&
o
-------
1 in the method for calculating vehicle miles traveled. Additional differences relate to the use of
2 county-level statistics for vehicle registration, as well as the use of temperature data from
3 individual counties. The value of 6.2 x 107 short tons (56 Tg) shown in Table 3-3 for emissions
4 from onroad vehicles for 1991 may be compared to a value of 4.0 x 107 short tons (36 Tg)
5 derived from remote sensing of vehicle exhausts for 1991 (Bradley et al., 1999).
6 In addition, it should be noted that Table 3-3 does not include formation of secondary CO,
7 such as from the oxidation of isoprene. Annual emissions of isoprene in the contiguous United
8 States are about 17.2 Tg/year (Pierce and Dudek, 1996). A source of CO of 7.1 Tg/year can be
9 calculated using the conversion factor of 0.20 for C in isoprene to C in CO estimated by Jacob
10 and Wofsy (1990). This value would add about 9% to the estimated U.S. emissions for CO in
11 1995, shown in Table 3-3. The oxidation of anthropogenic and other natural NMHCs may
12 supply an additional 2 to 3 Tg CO per year.
13 A number of techniques, such as roadside tunnel sampling and the remote sensing of
14 individual motor vehicle emissions have been applied in the past several years at a number of
15 locations throughout the United States to test CO emissions estimates and to derive emissions
16 factors (i.e., emissions per unit distance traveled). Two major points have been realized on the
17 basis of these studies: first, that a small percentage of motor vehicles are responsible for most of
18 the emissions, and, second, that CO and hydrocarbon emissions had been systematically
19 underestimated by as much as a factor of two in emissions factor models. As a result of these
20 studies, a number of revisions have been made to emissions inventories.
21 Roadside remote sensing data indicate that about 50% of CO and NMHC emissions are
22 produced by only about 10% of the vehicles (Lawson et al., 1990; Stephens and Cadle, 1991).
23 These "superemitters" are typically older, poorly maintained vehicles. Bishop and Stedman
24 (1996) also found that the most important variables governing CO emissions are fleet age and
25 owner maintenance. There are also a surprising number of newer vehicles that are classified as
26 superemitters. Possible reasons are related to tampering with emissions control systems to
27 improve milage, the use of contaminated fuels that may interfere with the proper operation of
28 emissions control systems, and the lack of maintenance of emissions control equipment and the
29 failure of emissions control equipment. In addition to the above activities, so-called off-cycle
30 operations also can result in enhanced emissions relative to those conditions for which emissions
31 testing is usually done. For example, rapid accelerations have been shown to increase emissions
October 1999 3-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 relative to less stressful driving modes. Roadside remote sensing of motor vehicle emissions
2 have also been used to evaluate the effectiveness of inspection and maintenance programs
3 (Zhang et al., 1996; Stedman et al., 1997; Stedman et al., 1998).
4 A comparison of emissions factors computed on the basis of tunnel measurements in
5 Van Nuys, CA, during the South Coast Air Quality Study in 1987 with those calculated by
6 emissions inventory models indicated that CO emissions were underpredicted by emissions
7 models (i.e., the Emissions Factor 7C [EMFAC7C] model, which is similar to MOBILES) by a
8 factor of 2.7, and hydrocarbon emissions were underestimated by a factor of 3.8 (Ingalls et al.,
9 1989; Pierson et al., 1990). A reinterpretation of the Van Nuys tunnel data by Pollack et al.
10 (1998) indicated that emissions factors calculated using MOBILESa were only a few percent
11 greater than the ambient tunnel data indicated (21.3 versus 20.9 g/mi), compared with a factor of
12 two difference using EMFAC7F (9.6 versus 20.9 g/mi). Likewise, a comparison of emission
13 factors computed on the basis of measurements in the Fort McHenry, MD, and Tuscarora, PA,
14 tunnels with those calculated by emissions models (MOBILE4.1 and MOBILES) indicated that
15 both versions of the MOBILE model gave predictions within ±50% of observations most of the
16 time (Pierson et al., 1996). However, it should be noted that emissions in tunnels arise from
17 vehicles in warmed-up or hot-stabilized operation. Cold and hot start emissions, which are
18 important components of the emissions inventory, generally are not present in tunnels and thus
19 are not evaluated in these studies.
20 Comparisons of ambient air quality data with predictions of emissions factor models have
21 been made for conditions when ambient concentrations result primarily from local emissions
22 with minimal photochemical processing and minimal transport from locations with different
23 source characteristics. The optimal time to obtain such conditions is during the early morning,
24 when ambient concentrations of CO, non-methane organic compounds (NMOCs), and NOX
25 typically peak and are dominated by local mobile source emissions (Fujita et al., 1992). These
26 comparisons have been performed in California for the Los Angeles Basin (Fujita et al., 1992),
27 the San Joaquin Valley, and San Francisco Bay Area (Magliano et al., 1993), and for the
28 Lake Michigan air quality region (Korc et al., 1993). A fairly consistent picture of
29 underpredictions of ambient CO to NOX and NMOC to NOX ratios by emissions factor models,
30 after allowing for the effects of atmospheric processing and transport, emerged from these
31 studies. In the Los Angeles Basin study, ambient CO to NOX ratios were factors of 1.3 to
October 1999 3-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.9 higher than corresponding emissions inventory ratios during summer, and factors of 1.2 to
2 2.4 higher than predicted by emissions models during fall. In the San Joaquin Valley study,
3 ambient CO to NOX ratios ranged from factors of 1.1 to 7.2 higher than predicted by emission
4 models. In the Lake Michigan area study, ambient CO to NOX ratios ranged from factors of
5 1.7 to 4.7 higher than predicted by emissions models. However, more recent comparisons
6 between ambient and emission inventory CO/NOX for Los Angeles and the San Joaquin Valley
7 show better agreement than in the previous studies mentioned above (Croes et al., 1996; Ipps and
8 Popejoy, 1998; Haste et al., 1998). These improvements have largely arisen through the process
9 of model development, evaluation and further refinement.
10 Stationary sources account for approximately 23% of nationwide CO emissions. Indoor
11 sources represented in Table 3-3 by residential combustion of wood and other fuels account for
12 only about 3% of annually averaged, nationwide CO emissions. However, on a local basis where
13 wood burning is widespread, these sources can account for significant fractions of the CO present
14 in ambient air. Khalil and Rasmussen (1988, 1989) have shown that during the winter in
15 Medford, OR, and in Olympia, WA, the contribution of wood burning to CO levels was of
16 comparable importance to automobiles. Khalil and Rasmussen (1999) have found that biomass
17 burning which takes the form of agricultural burning during the Fall and residential wood
18 burning during the Winter account for 20 to 40% of excess over nonurban background CO levels.
19
20
21 3.4 CARBON MONOXIDE CONCENTRATIONS IN AMBIENT AIR
22 The U.S. Environmental Protection Agency's (EPA's) Aerometric Information Retrieval
23 System (AIRS) receives data from the National Air Monitoring Stations (NAMS) and the State
24 and Local Air Monitoring Stations (SLAMS). Current NAAQS define 1- and 8-h average
25 concentrations that should not be exceeded more than once per year. The standards are met if the
26 second highest 1-h value is less than or equal to 35 ppm (40 mg/m3), and the second highest
27 nonoverlapping 8-h value is less than or equal to 9 ppm (10 mg/m3). Nationwide trends in
28 ambient CO concentrations are presented in Section 3.4.1, diurnal variations in ambient CO
29 levels are presented in Section 3.4.2, and a more detailed characterization of the spatial and
30 temporal variability in ambient CO concentrations in selected urban areas are presented in
31 Section 3.4.3. The analyses in Section 3.4.3 were performed for the Denver, CO (Shadwick
October 1999 3-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 et al., 1997); Los Angeles; New York City, NY; and Phoenix, AZ, Metropolitan Statistical Areas
2 (Shadwick et al., 1997, 1998a,b,c). The results of similar analyses for the Fairbanks, AK,
3 Metropolitan Statistical Area will be included in this section when they become available.
4
5 3.4.1 Nationwide Trends in Ambient Carbon Monoxide Concentrations
6 In 1997, 538 monitoring sites reported ambient CO air quality data to EPA's AIRS. Most
7 CO monitoring stations in the United States are located in larger urban areas. Figure 3-3 displays
8 the geographic locations of the monitoring sites reporting CO data to AIRS for 1997. On the
9 map, the sites are identified as NAMS, SLAMS, or "other". The NAMS were established by
10 EPA to ensure a long-term national network for urban-area-oriented ambient monitoring and to
11 provide a systematic, consistent database for air quality comparisons and trends analysis. The
12 SLAMS allow state or local governments to develop networks tailored for their immediate
13 monitoring needs. These NAMS and SLAMS sites conform to uniform criteria for monitor
14 siting, instrumentation, and quality assurance. "Other" monitors may be Special Purpose
15 Monitors, monitors at industrial sites, monitors on tribal lands, etc. Although state and local air
16 programs may require extensive monitoring to document and measure the local impacts of CO
17 emissions, only two NAMS sites are required in urbanized areas with populations greater than
18 500,000. Two categories of NAMS sites are required: (1) peak concentration areas (microscale),
19 such as major traffic corridors, street canyons, and major arterial streets, and (2) areas with high
20 population and traffic densities (middle scale or neighborhood scale).
21 Twenty-seven sites in 15 areas failed to meet the 8-h CO NAAQS in at least 1 year of the
22 5-year period, 1993 to 1997. Only six of the sites shown in Figure 3-3 failed to meet the 8-h
23 standard of 9 ppm, and none of the 538 monitoring sites exceeded the 1-h standard of 35 ppm in
24 1996. The locations of these 27 sites and the second highest 8-h CO concentrations and the
25 number of exceedances by year, are given in Table 3-4.
26 Figure 3-4 shows the consistent, downward trend in the nationwide composite average of
27 the annual second highest 8-h CO concentration during the past 20 years (1978 through 1997).
28 This statistic relates directly to the averaging time and form of the current CO NAAQS and
29 complies with the recommendations of the Intra-Agency Task Force on Air Quality Indicators
30 (U.S. Environmental Protection Agency, 1981). The dashed curve in Figure 3-4 tracks the trend
October 1999 3-24 DRAFT-DO NOT QUOTE OR CITE
-------
NAMS
SLAMS
Other
Figure 3-3. Locations of sites in the nationwide ambient CO monitoring network, 1997.
Source: U.S. Environmental Protection Agency (1998).
1 in the composite mean of the annual second highest 8-h average concentration for
2 184 monitoring sites that reported ambient air quality data in at least 17 of the past 20 years,
3 1978 to 1997. All monitoring sites are weighted equally when computing the nationwide
4 composite mean concentration. This selection criterion maximizes the number of sites available
5 for trend analyses. This subset of sites yields good geographical coverage with sites from more
6 than 90 cities in 39 states. Each year, site leases are lost, or sites are discontinued, and new sites
7 come online; therefore, the 184 long-term-trend sites compose only one-third of the currently
8 active CO monitors. The solid line in Figure 3-4 shows the trend in the composite mean for a
9 larger database of 368 sites that have reported ambient CO monitoring data in at least 8 of the
10 past 10 years. Missing annual second-highest CO concentration data for the second through
October 1999
3-25
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-4. SITES NOT MEETING THE 8-HOUR CARBON MONOXIDE
a
o
cr
VO
VO
VO
OJ
I
to
ON
O
>
H
6
o
o
H
O
d
o
w
o
o
HH
i
W
NATIONAL AMBIENT AIR
1993
Location Airs Site ID
Anchorage, AK 020200017
020200018
020200037
Denver, CO 080310002
Detroit, MI 261630014
El Paso, TX 481410027
481410044
Fairbanks, AK 020900002
020900013
020900020
Flathead CO, MT 300290045
Calexico, CA 060250005
060250006
Jersey City, NJ 340175001
Las Vegas, NV/AZ 320030557
Los Angeles-Long Beach, CA 060371002
060371201
060371301
2nd
Maxa
7.7
9.7
9.9
10.4
5.6
7.4
10.6
10.1
9.6
9
n/d
n/d
n/d
7.6
9.9
8.1
8
13.8
No.
Exc.bb
0
2
2
2
0
0
2
5
2
1
n/d
n/d
n/d
0
3
0
0
20
QUALITY STANDARD, 1993 TO
1994
2nd
Max
8.3
8.6
11
8.2
10.3
7.1
7.6
10.2
8.5
9.8
n/d
12.9
n/d
10.7
10.6
10.2
9.9
15.3
No.
Exc.
0
0
2
1
2
1
0
3
1
3
n/d
10
n/d
3
5
5
3
24
1995
2nd
Max
7.6
7.4
8.4
9.5
5.6
7.9
7.5
11.8
10.6
11.6
6.5
19.7
n/d
8.1
9.2
11
9.4
11.6
No.
Exc.
0
0
0
2
0
0
0
9
3
7
0
15
n/d
1
1
5
1
14
1997
1996
2nd
Max
9.6
8.7
10.5
7.3
4.5
10.3
9.1
8.6
8.4
8.6
11.1
14.1
7.8
6.7
10.1
8.5
6.7
14.5
No.
Exc.
3
0
3
0
0
2
1
1
0
0
2
9
0
0
3
0
0
22
1997
2nd
Max
6.8
7.1
7
5.5
n/dc
7.9
7.2
12.1
10.8
10.6
4.9
16.7
9.6
6.7
6.3
7.2
7.7
15
No.
Exc.
0
0
0
0
n/d
1
1
3
2
4
0
12
2
0
0
0
1
12
-------
o
a
o
a-
VO
VO
VO
OJ
i
to
o
TABLE 3-4 (cont'd). SITES NOT MEETING THE 8-HOUR CARBON
NATIONAL AMBIENT AIR QUALITY STANDARD, 1993 TO
1993
Location
Newark, NJ
Phoenix-Mesa, AZ
Provo-Orem, UT
Spokane, WA
Steubenville-Weirton,
OH-WW
a Annual second highest
Airs Site ID
060375001
340390003
040130019
040130022
490490004
530630040
530630044
540290009
540290011
2nd
Maxa
9.6
6
8
n/d
9.6
9.8
11.8
8.3
9.4
No.
Exc.b
2
0
0
n/d
2
2
4
1
1
1994
2nd
Max
11.3
11.3
9.6
n/d
9.3
8.1
8.8
9.6
17.1
No.
Exc.
6
2
2
n/d
1
0
0
2
5
1995
2nd
Max
8.7
7.7
8.4
9.9
7.1
8.4
11.2
6
6.7
No.
Exc.
0
0
0
3
0
0
4
0
1
MONOXIDE
1997
1996
2nd
Max
10.5
6
8.2
10
9.1
9
8.4
6.2
3.6
No.
Exc.
5
0
0
2
1
1
1
0
0
1997
2nd No.
Max Exc.
7.9 1
5.1 0
7 0
7.8 1
n/d n/d
6.3 0
n/d n/d
8.8 1
2.5 0
nonoverlapping 8-h average CO concentration.
[> b Number of exceedances of the 8-h CO NAAQS.
H
1
O
O
0
H
O
0 n/d = no data.
Source: U.S. Environmental Protection Agency's
Aeormetric
Information Retrieval
System (AIRS).
O
HH
H
W
-------
14-
12-
E
0.10-
Q.
I
O
O 6-
o
O
4-
1978-87 (184 sites)
1988-97 (368 sites)
I I I I I I I I I I I I I I I I I I
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
Year
Figure 3-4. Nationwide composite average of the annual second highest 8-h CO
concentrations, 1978 to 1997.
Source: U.S. Environmental Protection Agency (1998).
1 ninth years are estimated by linear interpolation from the surrounding years. Missing endpoints
2 are replaced with the nearest valid year of data. This latter procedure explains the discrepancy
3 between the two curves in 1988. Specific computational details are described elsewhere (U.S.
4 Environmental Protection Agency, 1998). This larger data set permits the examination of the
5 inter-site variability in peak CO concentrations. Figure 3-5 presents the 10th, 50th, and
6 90th percentile concentrations and the composite mean concentrations across these 368 sites.
7 The 10th, 50th, and 90th percentile concentrations CO concentrations for each year are indicated,
8 respectively, by the bottom, middle, and top lines of each box. For example, 10% of the
October 1999
3-28
DRAFT-DO NOT QUOTE OR CITE
-------
o
"co
CD
o
o
O
O
O
1O-
5-
o -I
Key:
— 9Oth Percentile
— Mean
— Median
— 10th Percentile
368 Sites
NAAQS
88 89 9O 91 92 93 94 95 96 97
Vear
Figure 3-5. Variability in the annual second highest 8-h CO concentrations across all sites
in the United States, 1988 to 1997.
Source: U.S. Environmental Protection Agency (1998).
1
2
3
4
5
9
10
11
368 trends sites reported 1988 second highest 8-h CO concentrations lower than the bottom of
the first bar in Figure 3-5. The yearly composite mean across all 368 sites is indicated by
the"x" in each bar. Figure 3-6 shows trends in CO concentrations in each of the different
sampling environments (urban, suburban, and rural sites). As can be seen from Figure 3-6, the
downward trend in ambient CO concentrations occurred at monitoring sites in urban, suburban,
and rural environments. An interesting feature of the data shown in Figures 3-4 to 3-6 is the
increase in CO levels from 1993 to 1994, which is the only year over year increase except for
1985 to 1986. The increase corresponds to an increase in mobile source and wildfire emissions
presented in Table 3-3. The decrease in ambient CO levels measured in populated areas over the
past decade is also reflected at least at one continental background site at Shenandoah National
Park, VA (Hallock-Waters et al., 1999).
October 1999
3-29
DRAFT-DO NOT QUOTE OR CITE
-------
Rural (12 sites)
Suburban (145 sites)
Urban (2O8 sites)
88
89
Year
Figure 3-6. Composite average of the annual second highest 8-h CO concentrations for
rural, suburban, and urban sites, 1988 to 1997.
Source: U.S. Environmental Protection Agency (1998).
1 3.4.2 Orcadian Patterns in Carbon Monoxide Concentrations
2 The circadian variation in winter time, composite, hourly CO concentrations from 1987
3 through 1996 is shown in Figure 3-7 (Cohen and Iwamiya, 1998). It can be seen that hourly
4 mean CO concentrations peak during the morning rush hours (7 to 9 a.m.). This peak results
5 mainly from CO emitted into the relatively shallow morning boundary layer by motor vehicles
6 (e.g., Fujita et al., 1992). The CO mixing ratios decline towards mid-afternoon, as the height of
7 the atmospheric mixing layer increases and then increase again with the onset of the evening rush
8 hour. Carbon monoxide levels fall off less rapidly after the afternoon peak because the mixing
9 layer height decreases during evening and nighttime. There is a general decrease in CO levels
10 during the night because of a lack of fresh emissions combined with processes such as mixing
11 with CO-poor areas and deposition to the surface. The downward trend in CO concentrations
12 from 1987 to 1996 is apparent for all times of the day. Especially notable is the decease in 7 to
October 1999
3-30
DRAFT-DO NOT QUOTE OR CITE
-------
Q.
Q.
g
2
•i—»
I
o
O
2.5-
2.0-
1.5-
1.0-
0.5-
i i i i i n
1234567
n i i i i i i i i i i i i i i
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Figure 3-7. Diurnal variation of nationwide composite hourly average CO concentrations
for winter (December to February), 1987 to 1996.
Source: Cohen and Iwamiya (1998).
1
2
3
4
5
9
10
11
12
13
14
15
9 a.m. CO concentrations, which is consistent with the decrease in motor vehicle emissions that
was noted earlier for the same period. During the period from 1987 to 1996, the 24-h nationwide
composite average CO concentration decreased from 2.0 to 1.2 ppm.
The circadian pattern of 8-h average CO concentrations that exceed 9.5 ppm is somewhat
different from the hourly average concentration pattern shown in Figure 3-7. One way to
estimate the time of day when 8-h average CO concentrations are likely to exceed 9.5 ppm is to
sum the numbers of these events by hour of the day over the course of a year. Hourly average
CO data were taken from AIRS to construct running 8-h averages for 1996.
In the previous AQCD for CO (U.S. Environmental Protection Agency, 1991b), which
summarized ambient monitoring data through 1988, six stations in six major cities, with
prominent patterns of 8-h exceedances, were selected to demonstrate the variability in the
circadian patterns of exceedances using the above technique; some peaked in daylight hours and
others in nighttime hours. Five of those six monitors are still in operation; Table 3-5 summarizes
and compares their 1988 record with 1996 data.
October 1999
3-31
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-5. RUNNING-AVERAGE EXCEEDANCES OF THE 9 ppm 8-HOUR
CARBON MONOXIDE STANDARD, 1988,1996, and 1997
Location
Lynwood, CA
Hawthorne, CA
Las Vegas, NV
New York City, NY
Steubenville, OH
Spokane, WA
AIRS Site ID No.
06037 1301
060375001
32 003 0557
360610081
39081 1012
53 063 0040
1988
392
163
102
123
152
169
1996
110
19
3
NAa
0
1
1997
62
2
0
0
0
0
1997 data completeness
Jan-Dec/97
Jan-Dec/97
Jan-Mar/97
Jan-June, Oct-Dec/97
Jan-Sep/97
Jan-Dec/97
a NA = not available
Note: As of June 14, 1999, the 1997 data for the Las Vegas, New York City, and Steubenville stations were incomplete. Based on 1996
data, Steubenville probably is not a problem; the New York City station is missing summer months, and the "zero" is probably warranted;
and the Las Vegas station is missing the fall 1997 months; hence, the conclusion is indeterminant.
1 Note that in Table 3-5 and in the analysis of 1996 data shown in Table 3-6 a "running-
2 average" exceedance is defined as any hour that culminates in an 8-h average higher than
3 9.5 ppm. This definition differs from that used in the construction of Table 3-4 in that the
4 number of nonoverlapping exceedances was used in Table 3-4. A formal violation of the 8-h
5 standard occurs when, in a given year, a second 8-h average exceeds 9.5 ppm but does not
6 overlap the first 8-h exceedance. Exceedances culminating in any hour are treated here because
7 an individual's cumulative exposure to a level greater than 9.5 ppm could occur in any hour.
8 At the Lynwood station, the running-average exceedances have declined from 392 in 1988
9 to 110 in 1996 (22 nonoverlapping exceedances); the majority of exceedances in 1996 occurred
10 in the hours between midnight and sunrise, as they had in 1988. They occurred in the months of
11 January, February, November, and December.
12 At the Hawthorne station, running-average exceedances have declined from 163 in 1988 to
13 19 in 1996 (five nonoverlapping exceedances). These are clustered around sunrise when
14 dispersion is most likely to be at a minimum. The exceedances at this station also occur in the
15 winter quarter.
16 Only a small number of stations have several running-average exceedances; however, these
17 recurrent high levels are attributed to unusual local situations. A prime example is the
October 1999 3-32 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-6. ANNUAL ORCADIAN PATTERN OF 8-HOUR AVERAGE CARBON
MONOXIDE CONCENTRATIONS CULMINATING IN VALUES GREATER THAN
9.5 ppm IN LYNWOOD AND HAWTHORNE, CA, DURING 1996
Ending of 8-h Period
Midnight
1 a.m.
2 a.m.
3 a.m.
4 a.m.a
5 a.m.
6 a.m.
7 a.m.
8 a.m.
9 a.m.
10a.m.
11 a.m.
Noon
1 p.m.
2p.m.
3 p.m.
4 p.m.
5 p.m.
6 p.m.
7 p.m.
8 p.m.
9 p.m.
10p.m.
11 p.m.
Total
Lynwood
9
9
10
11
8
7
8
7
8
8
6
O
2
1
1
0
0
0
0
0
1
1
4
6
110
Hawthorne
0
0
2
O
4
4
2
2
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19
"Calibrations normally are done at 4 a.m., thus values are interpolated.
1 monitoring station in Calexico, which is several blocks away from a major U.S.-Mexico border
2 crossing and the route leading to it; nine nonoverlapping exceedances were recorded in 1996.
3 Reportedly, there are often long lines of idling vehicles waiting to cross the border, including
4 vehicles of Mexican registration that are not equipped with the emission control equipment
5 required on vehicles sold in the United States. Such situations will need to be addressed on a
6 local, case-by-case basis.
7
October 1999 3-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.4.3 Characterization of the Spatial and Temporal Variability in Carbon
2 Monoxide Concentrations in Selected U.S. Cities.
3 The spatial and temporal variability of ambient carbon monoxide was characterized in four
4 cities in the continental United States (New York, NY; Denver, CO; Phoenix, AZ; and
5 Los Angeles, CA) and results will be reported for Fairbanks, AK, as they become available.
6 These five cities were chosen to characterize the spatial and temporal variability in CO in widely
7 different geographic regions. New York City is characterized by urban canyons. Denver is a
8 rapidly growing, high-altitude city. Phoenix is a rapidly growing city in an arid environment.
9 Los Angeles is characterized by emissions which are confined to a mountain basin. Fairbanks is
10 located in a mountain valley with a much higher potential for air stagnation than the continental
11 U.S. cities. Each of these cities has been in nonattainment of the 8-h NAAQS for CO at some
12 time within the past 5 years. In addition, the four cities in the continental Unites States has been
13 the locations of studies either characterizing personal exposure to CO or relating health outcomes
14 to air pollution levels.
15 Hourly average CO data obtained from EPA's AIRS were used to calculate running 8-h
16 averages for 1986 to!996. Only valid hourly average values were used to compose the 8-h
17 average. In the case that less than six valid hourly average values were used to compose the 8-h
18 average, the 8-h average was set to a missing value. The six valid hourly average values in an
19 8-h window corresponds to 75% data capture in the 8-h window.
20 The 24 running 8-h averages assigned to a day were used to compute the daily maximum
21 8-h average. A daily maximum 8-h average was considered to be valid if at least 18 of the 8-h
22 running averages for the day were valid as described in the preceding paragraph. The 18 valid
23 8-h running averages in a day corresponds to a 75% data capture. In the case that a valid daily
24 maximum 8-h running average could not be computed, a missing value was assigned to the daily
25 maximum 8-h average. The summary statistics were computed without regard to data capture.
26 Summary statistics (aside from the total number of observations) should be regarded as
27 representative if at least 75% of the possible data values were valid. Statistics on central
28 tendency and correlation were tabulated for all of the sites in each MSA for both the hourly and
29 8-h running averages. The statistics were analyzed by year, season, day of week, and hour of the
30 day.
October 1999 3-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 A map of Denver showing the locations of CO monitoring sites is given in Figure 3-8. The
2 average diurnal variation in CO concentrations at the Denver-Broadway site during the winter
3 months of November through February for the period 1986 through 1995 is given in Figure 3-9.
4 The monthly average diurnal variation in CO at this site for weekdays from May 1986 through
5 May 1987 is shown in Figure 3-10, whereas the same quantity for the period 1995 to 1996 is
6 shown in Figure 3-11. Central tendency statistics for the daily 8-h max CO concentration for the
7 winter season from 1986 to 1995 at the Broadway site are given in Figure 3-12.
8 A map of Los Angeles showing the locations of CO monitoring sites is shown in
9 Figure 3-13. The average diurnal variation in CO concentrations at the Los Angeles-Lynwood
10 site during the winter months of November through February for the period 1986 through 1995 is
11 given in Figure 3-14. Central tendency statistics for the daily 8-h max CO concentration for the
12 winter season from 1986 to 1995 at the Broadway site are given in Figure 3-15 and for the
13 Los Angeles-Barstow site are given in Figure 3-16. The average diurnal variation in CO
14 concentrations at the Los Angeles-Hawthorne site during the winter months of November
15 through February for the period 1986 through 1995 is given in Figure 3-17 and for the
16 Los Angeles-El Toro site in Figure 3-18. The monthly average diurnal variation at the
17 Los Angeles-Anaheim site from May through May 1986 to 1987, 1989 to 1990, 1992 to 1993,
18 and 1995 to 1996 is shown in Figure 3-19.
19 A map of the New York Metropolitan area showing the locations of CO monitoring sites is
20 given in Figure 3-20. The monthly average diurnal variation at the New York-Flatbush site from
21 May through May 1986 to 1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 is shown in
22 Figure 3-21. The average diurnal variation in CO concentrations at the New York-Manhattan
23 site during the winter months of November through February for the period 1986 through 1995 is
24 given in Figure 3-22. The monthly average diurnal variation at the New York-Manhattan site
25 from May through May 1986 to 1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 is shown in
26 Figure 3-23 and for the New York-Morristown, NJ, site in Figure 3-24.
27 A map of Phoenix showing the locations of CO monitoring sites is given in Figure 3-25.
28 The monthly average diurnal variation at the Los Angeles-Anaheim site from May through May
29 1986 to 1987, 1989 to 1990, 1992 to 1993, and 1995 to 1996 is shown in Figure 3-26. Central
30 tendency statistics for the daily 8-h max CO concentration for the winter season from 1986 to
31 1995 at the east Butler site are given in Figure 3-27. The monthly average diurnal variation at the
October 1999 3-35 DRAFT-DO NOT QUOTE OR CITE
-------
.'I.
Greeley
Boulder^Marine
Arvada
Welby r "-;
J :-
4 ; •' ;
«••. « •"•»••-..,.„•.»
De H ye r» ^ roa d way
: Denver-Julian;
^••-Littleton
Population Density
0 to 300 pcrsonsi'krn2
301 to 1000 pcrsons/km2
1001 to 2000 persons'km2
2001 to 5000 persons/km2
^B 5001 to 10,000 persons/kmS
| over 10,000 persons/km 2
® CO Monitoring Sites
'\ / Counties
Interstates/Divided Highways
Major Roads
Water
10
10
20
Figure 3-8. Map of Denver showing locations of CO monitoring sites.
October 1999 3-36 DRAFT-DO NOT QUOTE OR CITE
-------
ret
Caior Scale in
Figure 3-9. Average diurnal variation in CO at the Denver-Broadway site for weekdays
during the winter season (November through February). The abscissa shows the time of
day from midnight to midnight, the ordinate shows years from the winter of 1986/87
through 1995/96, and the z-axis shows CO concentration in ppm.
} 4 5 6
Color Scale ir. PPW
Figure 3-10. Monthly average diurnal variation in CO at the Denver-Broadway site for
weekdays from May 1986 through May 1987. The abscissa shows the time of day, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
October 1999
3-37
DRAFT-DO NOT QUOTE OR CITE
-------
KnUi
Figure 3-11. Monthly average diurnal variation in CO at the Denver-Broadway site for
weekdays from May 1995 through May 1996. The abscissa shows the time of day, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
YEAR
n Median
o Outliers
o Extremes
--H--P50
1985 1986 1987 1988 1989
1990 1991
YEAR
1992 1993 1994 1995 1996
Figure 3-12. Central tendency statistics for the daily 8-h max CO concentration at the
Denver-Broadway site during the winter season from 1986 to 1995. The top graph shows
box plots (with 10, 25, 50, 75, and 90 percentile values) for the entire time series. Each
circle (outlier) or diamond (extreme) refers to an individual observation that is either three
or four standard deviations (SDs) from the mean, and the horizontal line shows the current
8-h NAAQS for CO. The lower graph again shows the 25, 50, and 75 percentile values (P25,
P50, and P75) from the upper graph.
October 1999
3-38
DRAFT-DO NOT QUOTE OR CITE
-------
V
+
i.
£&&*>
(-jReseda
V
Barstow:?
<£MB urban k
Angeles
Azusa
Population Density
Q to 300 persons/km2
301 to 1000 persons/km2
1001 to 2000 per5ons/km2
| 2001 to 5000 persons/kmZ
^H 5001 to 10,000 persons/km2
over 10,000 pcrsons/km2
® CO Monitoring Sites
A. / Counties
States
Interstates/Divided Highways
Major Roads
Water
10 0 10 20 3040km
;: c iwt
4 Hi
i K.
Figure 3-13. Map of Los Angeles showing locations of CO monitoring sites.
October 1999
3-39
DRAFT-DO NOT QUOTE OR CITE
-------
Hour
Figure 3-14. Average diurnal variation in CO at the Los Angeles-Lynwood site for
weekdays during the winter season (November through February). The abscissa shows the
time of day from midnight to midnight, the ordinate shows years from the winter of 1986/87
through 1995/96, and the z-axis shows CO concentration in ppm.
1
2
3
4
5
6
7
8
9
10
11
1995
Phoenix-West site from May through May 1986 to 1987, 1989 to 1990, 1992 to 1993, and
to 1996 is shown in Figure 3-28 and at the Phoenix-South site in Figure 3-29.
Analysis of ambient CO data obtained in the four geographically diverse metropolitan
statistical areas of Denver, Los Angeles, New York, and Phoenix has shown that urban CO levels
have decreased over the past 10 years. However, there have been instances where the downward
trend has reversed itself on a year-to-year basis. Although the number of violation days has
declined for these cities, and the seasonally averaged peak concentrations generally do not exceed
8 ppm, at least one exceedance of 9 ppm for the maximum daily 8-h average for CO has occurred
in 1995/1996 (the final year in this analysis) in all four of these cities.
Data obtained from different monitoring sites within a given MSA show a large degree of
variability. During 1996, for example, annual mean CO concentrations ranged from 0.4 to
October 1999
3-40
DRAFT-DO NOT QUOTE OR CITE
-------
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
YEAR
n Median
_ 10
S 9
fe 8
z 7
| 6
z 4
g 3
§ 2
O 1
1986 1987 1988 1989 1990 1
YEAR
1992 1993 1994 1995
—o— P75
--D-- P50
-•C--P25
Figure 3-15. Central tendency statistics for the daily 8-h max CO concentration at the Los
Angeles-Hawthorne site during the winter season from 1986 to 1995. The top graph shows
box plots (with 10, 25, 50, 75, and 90 percentile values) for the entire time series. The
horizontal line shows the current 8-h NAAQS for CO. The lower graph again shows the 25,
50, and 75 percentile values (P25, P50, and P75) from the upper graph.
1 1.5 ppm in the Denver MSA, 0.4 to 3.2 ppm in the Los Angeles CMSA, 0.6 to 3.7 in the New
2 York CMSA, and 0.7 to 3.4 in the Phoenix MSA. Carbon monoxide concentrations during the
3 cold season (November through February) range from 5 to 20% higher than the annual average in
4 each MSA. However, it should be noted that, despite decreasing CO concentrations, the nature
5 of the diurnal and seasonal variation observed at each monitoring site has remained remarkably
6 constant over the 10-year period covered in this analysis. At all the sites investigated here, it is
7 clear that the diurnal and seasonal variations in CO observed in these MSAs result largely from
8 the interaction between motor vehicle emissions and meteorological parameters that, at times,
9 can be conducive to the buildup of CO near the surface. The diurnal concentration profiles in
10 most cases show a very distinctive two-peaked structure for weekdays. The peaks correspond to
11 both the morning and evening rush hour commutes. Frequently, the morning peak is higher than
12 the evening peak at any given site because the height of the mixed layer is much lower during the
13 morning, thus inhibiting vertical mixing that would have diluted CO. In the late afternoon and
October 1999
3-41
DRAFT-DO NOT QUOTE OR CITE
-------
z
o
o;
z
LJJ
O
Z
o
o
14
12
10
8
6
4
2
g* 10
a. 9
^ 8
I 7
<: 6
C£ 5
h-
4
3
2
1
0
Z
LJJ
O
z
o
o
1986
o
1987
1988
1989
1990 1991
YEAR
1992
1993
1994
1995
o Median
o Outliers
1986 1987
1988 1989
1990 1991
YEAR
1992 1993
1994
1995
^3— P75
--n-- PSD
- -0- •• P25
Figure 3-16. Central tendency statistics for the daily 8-h max CO concentration at the
Los Angeles-Barstow site during the winter season from 1986 to 1995. The top graph
shows box plots (with 10, 25, 50, 75, and 90 percentile values) for the entire time series.
Each circle (outlier) refers to an individual observation that is three SDs from the mean,
and the horizontal line shows the current 8-h NAAQS for CO. The lower graph again
shows the 25, 50, and 75 percentile values (P25, P50, and P75) from the upper graph.
1 into early evening, increased atmospheric turbulence resulting from solar heating raises the
2 height of the mixed layer, resulting in generally lower CO concentrations compared with those of
3 the morning.
4 Regional differences in atmospheric processes also may play a role in producing the
5 nighttime behavior of CO observed at numerous sites in the Los Angeles and Phoenix MSAs
6 compared with either the nationwide composite average diurnal cycle of CO shown in
7 Figure 3-13 or other locations, such as the Denver or New York MSAs. In the Los Angeles and
8 Phoenix metropolitan areas CO concentrations often remain until midnight at levels reached
9 during the evening rush hour. Then, although these concentrations gradually diminish
October 1999
3-42
DRAFT-DO NOT QUOTE OR CITE
-------
Hour
Figure 3-17. Average diurnal variation in CO at the Los Angeles-Hawthorne site for
weekdays during the winter season (November through February). The abscissa shows the
time of day from midnight to midnight, the ordinate shows years from the winter of 1986/87
through 1995/96, and the z-axis shows CO concentration in ppm.
2 3 4 S
Scale In PHI
Figure 3-18. Average diurnal variation in CO at the Los Angeles-El Toro site for weekdays
during the winter season (November through February). The abscissa shows the time of
day from midnight to midnight, the ordinate shows years from the winter of 1986/87
through 1995/96, and the z-axis shows CO concentration in ppm.
October 1999
3-43
DRAFT-DO NOT QUOTE OR CITE
-------
19»(r 19*7
1989- 1WO
1MJ
6s»"
13«-199fi
Figure 3-19. Monthly average diurnal variation in CO at the Los Angeles-Anaheim site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
1 throughout the night, they do not drop to the low afternoon levels (typically no more than 1 to
2 2 ppm, often less than this amount) before they begin to increase again because of the morning
3 rush hour. This pattern is shown quite well in Figure 3-26, which depicts the seasonal, diurnal
4 concentration profile for the Central Phoenix site.
5 In general, the highest values of ambient CO were found during the wintertime (defined as
6 the months of November through February, inclusive) in all of the MSAs included here. There
7 were a few sites in the New York Metropolitan Area where a wintertime peak in CO was not
8 discernable; the site on Flatbush Avenue in Brooklyn (Figure 3-21) is an excellent example of
9 this. It is not clear without further analysis, what combination of seasonal variations in emissions
10 and meteorological parameters gave rise to this result.
11
October 1999
3-44
DRAFT-DO NOT QUOTE OR CITE
-------
v. };
/*?>.A".*l
.• Jersey Ci
'
.
'.
A >?^
PerH|Ainbi
, ^iR^ -
x-v - / ^
*^<.~£
.4 *
Id^
Population Densily © CO Monitoring Sites
0 to 300 personsi'km2 /'v1' Bounties
301 to 1000 per&ons/km2 I I States
1001 to 2000 persons/km 2
| 2001 to 5000 persons/krn2
| 5001 to 10,000 person5''km2 -
over 10,000 pereon^krn2
Interstates/Divided Highways
Major Roads
Water
10 20 30km
» t' 3 t
:: to
OH
AM
Nny.-nl»r 1»h,
Figure 3-20. Map of New York showing locations of CO monitoring sites.
October 1999
3-45
DRAFT-DO NOT QUOTE OR CITE
-------
1986- 19B7
IMS 1996
1 2 345 6
Oalor Scal« in FfW
8 9 10
Figure 3-21. Monthly average diurnal variation in CO at the New York-Flatbush site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
Figure 3-22. Average diurnal variation in CO at the New York-Manhattan site for
weekdays during the winter season (November through February). The abscissa shows
the time of day from midnight to midnight, the ordinate shows years from the winter of
1986/87 through 1995/96, and the z-axis shows CO concentration in ppm.
October 1999
3-46
DRAFT-DO NOT QUOTE OR CITE
-------
1JW
Figure 3-23. Monthly average diurnal variation in CO at the New York-Manhattan site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
I'JIIf, I?*.'
19K'
1993 1393
1W5- 19H
Figure 3-24. Monthly average diurnal variation in CO at the New York-Morristown,
NJ, site for weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and
1995 to 1996. On each graph, the abscissa shows the time of day from midnight to
midnight, the ordinate shows the month of the year, and the z-axis on each graph shows CO
concentration in ppm.
October 1999
3-47
DRAFT-DO NOT QUOTE OR CITE
-------
"I* 4Mt-j
fK Jt
E. Butler Dr.
©
West Phoenix
S9«r ^ N. Miller Rd.
Central PhoenixSS
South Phoenix
.
Density ® ^ Monitoring Sites
0 to 300 person 5,'km 2 / v*' Counties
301 to 1000 personsykm2 I I States
1001 to 2000 persons/km2 Interstates/Divided Highways
2001 to 5000 per sons,'km 2 Major Roads
5001 to 10,000 pcrsons.-kin2 ^H Water
over 10,000 p*rsons/km2
^
10 0 10km
>1 ll'iiiw
D4Q
h. 1*M
Figure 3-25. Map of Phoenix showing locations of CO monitoring sites.
October 1999 3-48 DRAFT-DO NOT QUOTE OR CITE
-------
11WT- IMD
L9SJ- 1993
l«fi
0 1 2 3 I 5 fi
Color Sc»l» in
Figure 3-26. Monthly average diurnal variation in CO at the Phoenix-Central site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
o
o
1986
1987
1988
1989
1990 1991
YEAR
1992
1993
1994
1995
n Median
o Outliers
—o- P75
--D-- P50
1987 1988
1991
1995
YEAR
Figure 3-27. Central tendency statistics for the daily 8-h max CO concentration at the
Phoenix-East Butler site during the winter season from 1986 to 1995. The top graph shows
box plots (with 10, 25, 50, 75, and 90 percentile values) for the entire time series. Each
circle (outlier) refers to an individual observation that is three SDs from the mean, and the
horizontal line shows the current 8-h NAAQS for CO. The lower graph again shows the
25, 50, and 75 percentile values (P25, P50, and P75) from the upper graph.
October 1999
3-49
DRAFT-DO NOT QUOTE OR CITE
-------
,-:,.
012345S7
Osier Scale ID FEW
Figure 3-28. Monthly average diurnal variation in CO at the Phoenix-West site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
in
IMS 1393
19JS-19M
0 1 2 3 4 5 6
Golar Scale ID PHI
Figure 3-29. Monthly average diurnal variation in CO at the Phoenix-South site for
weekdays from May through May 1986 to 1987,1989 to 1990,1992 to 1993, and 1995 to
1996. On each graph, the abscissa shows the time of day from midnight to midnight, the
ordinate shows the month of the year, and the z-axis shows CO concentration in ppm.
October 1999
3-50
DRAFT-DO NOT QUOTE OR CITE
-------
1 3.5 SOURCES, EMISSIONS, AND CONCENTRATIONS OF CARBON
2 MONOXIDE IN INDOOR ENVIRONMENTS
3 The general United States population spends up to 95% of its time indoors. In recent years,
4 more emphasis has been placed on the evaluation of pollutant sources, emissions, and
5 concentrations in indoor environments to aid in the evaluation of total human exposure. It is
6 particularly important to evaluate carbon monoxide concentrations in indoor environments
7 because indoor exposure may represent a significant portion of the total CO exposure.
8 The following sections focus on the sources and emission rates of CO in indoor
9 environments. Concentrations of CO in various indoor environments also will be discussed.
10 Emphasis is placed on the evaluation of manufacturer recommended uses of combustion
11 appliances and consumer products and the resulting CO emissions and concentrations.
12 Accidental sources and concentrations will be mentioned only briefly. This section will only
13 summarize those studies discussed in the previous criteria document. The reader is referred to
14 the 1991 Air Quality Criteria for Carbon Monoxide (U.S. Environmental Protection Agency,
15 1991b).
16
17 3.5.1 Sources and Emissions of Carbon Monoxide in Indoor Environments
18 Prior to 1991
19 Gas cooking stoves, unvented space heaters, cigarette combustion, woodburning stoves,
20 and automotive vehicles represent sources of CO in indoor environments. Approximately, 45%
21 of the homes in the United States used gas for cooking, drying clothes, and heating water in
22 1981. Emissions from gas cooking stoves were highly variable among burners on a singles gas
23 cooking range and between gas stoves. Improperly adjusted gas stoves (yellow-tipping flame)
24 could result in a greater than fivefold increase in emissions over that of properly adjusted stoves
25 (blue flame). However, on the average, emissions were comparable for top burners, ovens, and
26 pilot lights. Vented gas dryers and furnaces contribute a negligible amount of CO to the indoor
27 environment.
28 Carbon monoxide emissions from unvented gas and kerosene space heaters were also
29 variable from heater to heater. Carbon monoxide emissions from unvented gas space heaters
30 were higher for maltuned units and varied with the method of emission testing. Among the types
31 of unvented space heaters, emissions were higher for the infrared gas space heaters versus
October 1999 3-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 convective and catalytic unvented gas units. For unvented kerosene space heaters, emissions
2 were higher for the radiant heaters.
3 Limited information was available on CO emissions from woodburning stoves. However,
4 nonairtight woodburning stoves can contribute substantial amounts of CO into the indoor
5 environment. Carbon monoxide from cigarette combustion showed little variability among the
6 different brands, but could also be substantial. In 1987, 29% of the United States population
7 smoked. Carbon monoxide emissions from two cigarettes smoked over an hour approached that
8 of one range top burner operating under blue-flame conditions. Ranges in average emissions for
9 CO sources in the indoor environment are listed in Table 3-7.
10
11 3.5.2 Unvented Combustion Sources and Estimated Emissions Rates
12 Carbon monoxide occurs in indoor environments directly through emissions from various
13 indoor combustion sources or indirectly as a result of infiltration or ventilation from outdoor
14 sources. Unvented, partially vented, and improperly vented combustion appliances and
15 consumer products represent the primary sources of CO emissions in the indoor environment.
16 Table 3-8 lists the various sources of CO in the indoor environment. Emissions of CO from use
17 of combustion appliances will depend on several factors. These factors include the source (e.g.,
18 gas cooking stoves, unvented space heaters, woodstoves, fireplaces), appliance design, type of
19 fuel used, fuel consumption rate, use pattern, and operating condition.
20 Two different approaches are used to evaluate CO emissions for combustion appliances:
21 the direct or sampling-hood approach and the mass-balance/chamber approach. For details on
22 these two approaches, see U.S. Environmental Protection Agency (1991b).
23
24 3.5.2.1 Gas Cooking Ranges and Ovens and Furnaces
25 Emissions of CO from gas top ranges will depend on the use pattern, operating condition,
26 fuel consumption rate, and air infiltration into the microenvironment. Average annual household
27 fuel consumption has been estimated at 5,000 ft3 for ranges with standing pilots (Johnson et al.,
28 1992). Menkedick et al. (1993) reported annual household fuel consumption of 2,180 ft3
29 (±890 ft3) for burners, based on actual fuel consumption measurements taken on 103 gas ranges
30 individually metered over a 2-year period in Illinois. Fuel consumption for burner and standing
31 pilots was 5,710 ft3 (±1,830 ft3). The average fuel consumption was also affected by the age of
October 1999 3-52 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-7. RANGES IN AVERAGE CARBON MONOXIDE
EMISSION RATES FOR RESIDENTIAL SOURCES
Fuel
Unit Type Typea
Gas Ranges0
Top burners NG
Ovens NG
Burner pilots NG
Oven pilots NG
Gas Space Heaters
Convective NG
P
Infrared NG
P
Catalytic NG
Kerosene Heaters
Convective
Radiant
Two-stage
Fuel Consumption
Rate (kJ/minb) Flame
— Blue
— Yellow-
tipping
— Blue
— Yellow-
tipping
— Blue
— Blue
Yellow-
tipping
131-784 Blue
353-660 Blue
260-368 Infrared
258 Infrared
207 Blue
37-202 —
85-168 —
132-182 —
Ranges in Average
Emission Rates (Mg/kJ)
15-215
92-197
12-257
53-62
28-56
209-322
40
3-33
16
45-69
45
9-14
4-60
27-173
9-54
Source
Himmel and DeWerth (1974)
Traynoretal. (1982)
Borrazzo et al. (1987)
Coteetal. (1974)
Moschandreas et al. (1985)
Fortmanetal. (1984)
Himmel and DeWerth (1974)
Coteetal. (1974)
Moschandreas et al. (1985)
Himmel and DeWerth (1974)
Traynoretal. (1982)
Borrazzo et al. (1987)
Fortmanetal. (1984)
Himmel and DeWerth (1974)
Himmel and DeWerth (1974)
Himmel and DeWerth (1974)
Moschandreas et al. (1985)
Traynoretal. (1984, 1985)
Moschandreas et al. (1985)
Thrasher and DeWerth (1979)
Zawackietal. (1984)
Traynoretal. (1984, 1985)
Traynoretal. (1984, 1985)
Moschandreas et al. (1985)
Traynoretal. (1984, 1985)
Moschandreas et al. (1985)
Leaderer(1982)
Traynoretal. (1983)
Moschandreas et al. (1985)
Leaderer(1982)
Traynoretal. (1983)
Moschandreas et al. (1985)
Traynoretal. (1983)
October 1999
3-53
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-7 (cont'd). RANGES IN AVERAGE CARBON MONOXIDE
EMISSION RATES FOR RESIDENTIAL SOURCES
Unit Type
Gas Dryer
Water heater
Woodstoves and
Fireplaces
Tobacco Smoke
Cigarette
Fuel Fuel Consumption Ranges in Average
Type3 Rate (kJ/minb) Flame Emission Rates (pig/kJ)
NG — — 40-69
NG — — 25-77 ppm
— — — 0.08-2. 18 g/h
— — 40-67 mg/cigarette
(ms + ss)d
Source
Moschandreas et al. (1985)
Belles etal. (1979)
Traynoretal. (1987)
National Research Council
(1986)
Rickertetal. (1984)
aNG = natural gas, P = propane.
bOne kJ (kiloJoule) is the equivalent of 3.485 ft3 of natural gas.
Tuel consumption rates not provided for most studies.
dms + ss = mainstream and sidestream smoke.
1 the occupants (older adults used the range more frequently for preparing meals than did young
2 adults) and the presence or absence of a standing pilot, and showed a seasonal trend. A recent
3 study by Spicer and Billick (1996) evaluated CO emissions from a gas top burner both with and
4 without a load. An indoor monitoring study of 293 homes conducted by the Gas Research
5 Institute, Pacific Gas and Electric, and Southern California Gas showed increased CO emissions
6 from the use of older gas ranges with standing pilot lights (Billick et al., 1994, 1996). Study
7 details are included in the section on indoor concentrations.
8 Carbon monoxide emissions from vented gas furnaces are generally negligible (Borrazzo
9 et al., 1987); however, emissions may vary based on the working condition and efficiency of the
10 unit. Ryan and McCrillis (1994) evaluated CO emissions from two gas furnaces; one furnace
11 was an older model with an energy efficiency of 60 to 70% and the other was a newer furnace
12 with an energy efficiency of 94%. The furnaces were operated for 10 min then allowed to cool
13 for 5 to 10 min. The cycle was repeated 12 to 18 times during the course of each test. The
14 CO emission rate was >1,000 //g/kJ for the older unit, compared with 6 //g/kJ for the newer,
15 more efficient model.
16
October 1999
3-54
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-8. SOURCES OF CARBON MONOXIDE IN THE
INDOOR ENVIRONMENT
Source Comments
Outdoor (ambient air) Carbon monoxide is produced as a primary pollutant during the combustion of
fossil and biomass fuel and as a secondary gas in the photochemical oxidation
of methane and other organic compounds in the atmosphere. Carbon monoxide
enters indoor compartments through mechanical ventilation systems and
infiltration through the building envelope.
Gas cooking ranges Emissions of CO from gas ranges depends on the use pattern, unit operating
condition, and fuel consumption rate. Gas ranges with standing pilots emit
more CO than do units with electronic pilots. Poorly tuned burners emit more
CO than well-tuned burners.
Gas space heaters Emissions of CO from gas space heaters are affected by the fuel type and
consumption rate, type of burner (convective, radiant, or catalytic), operating
condition, and duration of use.
Kerosene space heaters Emissions vary based on unit type (convective, radiant, or two-stage), operating
condition, and duration of use.
Environmental tobacco smoke The majority of CO entering indoor compartments from the combustion of
tobacco products is through sidestream smoke.
Fireplaces and woodstoves Carbon monoxide is emitted during fire start-ups, leaks in stoves and pipes, and
during backdrafting resulting from depressurization.
Gas furnaces, clothes dryers, Gas furnaces and dryers generally are vented and do not emit CO in the indoor
and water heaters environment unless the unit is malfunctioning.
Motor vehicles Operating motor vehicles in enclosed spaces can be significant sources of CO
in indoor environments.
1 3.5.2.2 Emissions from Unvented Space Heaters
2 Carbon monoxide emissions from unvented space heaters will vary as a function of
3 appliance design and condition, manner of operation, and fuel type and consumption rate.
4 Higher emissions have been reported for infrared gas space heaters versus the convective or
5 catalytic units. Other factors that may affect emissions from unvented space heaters include air
6 circulation near the heater, primary aeration, air infiltration and exchange, and use pattern.
7 Hedrick and Krug (1995) determined the emissions of CO from four different gas space
8 heaters in operation and the emissions from the pilot lights. The study was conducted in a test
9 house in Chicago, IL. The house was a 1-story, 3-bedroom with a full basement, single-family
10 dwelling. Total square footage was 2,300 sq ft. Eight burner experiments and three pilot light
October 1999 3-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 experiments were conducted. Heaters were (1) 10,000 BTU-h blue-flame convective,
2 (2) 15,000 BTU-h (max.) radiant-tile (infrared), (3) 14,000 BTU-h (max.) fan-forced blue-flame
3 convective, and (4) 16,000 BTU-h perforated-tube convective heaters. Heaters were operated for
4 8-h followed by a 15-h decay period. The pilot studies were conducted over a 48-h period. The
5 emission rates for CO varied from 7.9 //g/BTU for fan-forced models to 182.1 //g/BTU for the
6 perforated tube (CO emissions enhanced by leaky gas pressure regulating valve) and
7 57.9 //g/BTU for the infrared unit. Carbon monoxide concentrations in the test house are
8 discussed in the section addressing CO concentrations associated with indoor sources. Spicer
9 and Billick (1996) reported CO emissions indexes of 19.1 and 28.7 //g/kJ for a convective
10 blue-flame space heater. An emission index of 44.1 //g/kJ was noted for a radiant unit.
11 Fan et al. (1997) reported the average pollutant emission rates for a new portable gas stove,
12 a used kerosene radiant space heater, kerosene lamp, an oil lamp, and several candles. The fuel
13 used for the portable gas stove was butane and 1-K grade kerosene was used in the kerosene
14 space heater. Both 1-K grade kerosene and citronella patio torch fuel were use for the oil lamp
15 test. The lamp wick was tested at the normal height (1 in.) and at high flame (2 in.). Four
16 7.6-cm diameter candles were burned together in each candle test. The test for the portable stove
17 and the kerosene heater was conducted in a room with a volume of 19 m3. Both the lamp tests
18 and the candle tests were conducted in a 0.15 m3 chamber. The tests were run for 30 min to 2 h
19 with a 30-min decay period. Estimates of the emission rates were done using a
20 single-compartment mass balance model. The emission rates for the portable gas stove and
21 kerosene heater were 33.6 ± 15.0 and 226.7 ± 100 mg/h, respectively. Emission rates of 8.2 ±
22 1.1, 7.1 ±0.8, and 4.7 ± 3.0 mg/h were established for the kerosene lamp, oil lamp, and candles.
23 The height of the wick did not affect the emission of CO from either the kerosene lamp or the oil
24 lamp. Several recent studies on CO emissions evaluated emissions from prototype units; a brief
25 discussion of the results of these studies follows.
26 The Gas Research Institute evaluated the emissions of CO, NO, NO2, and unburned
27 hydrocarbons (UHC) from a vented and an unvented prototype space heater. The heater design
28 was a Pyrocore ceramic fiber radiant burner. Emission goals were 5.5 //g/kJ (20 ppm) CO,
29 4.0 //g/kJ (13.5 ppm) NO, 0.7 //g/kJ (1.5 ppm) NO2, and 0.8 //g/kJ (5.0 ppm) UHC. Both units
30 had cross-flow room air circulation fans. The samples collected from the unvented space heater
31 were corrected for dilution effects. Emission rates were estimated using a single-equation
October 1999 3-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 mass-balance model. Carbon monoxide emissions ranged from 0.6 to >27.5 //g/kJ (2.0 to
2 >100 ppm), based on the amount of excess air and the firing rate values (Duret and Tidball,
3 1990).
4 Apte and Traynor (1993) determined the emission rates of combustion pollutants for a
5 radiant-fiber-matrix gas burner prototype. Fuel consumption ranged from 333 to 527 kJ/min.
6 Carbon monoxide emission rates were generally low, with an average of 3.0 //g/kJ when the unit
7 was operated with 10% excess air, and 7.4 //g/kJ with 40% excess air.
8 The Institute of Gas Technology, along with Maxon Corporation, designed and tested an air
9 heater based on the cyclonic combustion concept. This technology included premixed high-
10 excess air, and cyclonic combustion with flame stabilization, in conjunction with optimized
11 nozzle velocity control. Carbon monoxide emissions were 1.6 to 5.9 //g/kJ (1.0 to 3.6 ppm at
12 15% O2) (Xiong et al., 1991).
13
14 3.5.2.3 Woodstoves and Fireplaces
15 A few studies have evaluated the emission of CO from woodstove or fireplace use. Carbon
16 monoxide may enter the indoor environment during fire start-up and tending and through leaks in
17 the stove or venting system. Carbon monoxide emissions are higher during the first stage of a
18 fire because of increasing amounts of fuel being burned and inadequate temperature conditions.
19 Such intermittent emissions makes it difficult to accurately determine CO emission rates.
20 Mueller Associates (1985) reported CO emission ranges of 0.07 to 0.375 g/h for wood heaters.
21 Jaasma et al. (1995) conducted a study designed to evaluate the effectiveness of custom-built
22 glass doors for fireplaces in reducing CO emissions under conditions of negative pressure. The
23 glass doors decreased spillage of CO; however, decreasing the leakiness of the glass doors did
24 not always reduce CO spillage. Tests with the glass doors closed had CO emission rates of 2 to
25 36 g/h (highest levels represented leaking glass doors). Carbon monoxide emissions on the order
26 of 70 g/h were noted for glass-door-opened tests under negative pressure.
27 Carbon monoxide also may enter the indoor environment through backdrafting when the
28 natural draft is overcome by depressurization. Depressurization generally occurs during fire
29 start-up, but also may occur during operation of other equipment such as kitchen and bathroom
30 exhaust fans and clothes dryers. Nagda et al. (1996) summarized the results of several studies on
31 emissions of pollutants into living compartments as a result of house depressurization. Carbon
October 1999 3-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 monoxide emissions were found to be insignificant. Tiegs and Bighouse (1994) evaluated CO
2 spillage from a woodstove under chamber and in-house conditions. They reported CO leakage
3 into the indoor environment by nonairtight woodstoves during conditions of negative pressure.
4
5 3.5.2.4 Environmental Tobacco Smoke
6 Carbon monoxide emissions from the combustion of tobacco occurs in the indoor
7 environment when smokers exhale the inhaled or mainstream smoke (ms) and from the emission
8 of sidestream smoke (ss) from smoldering tobacco products. The majority of the CO emissions
9 are from sidestream smoke. Ott et al. (1992) reported a sidestream-to-mainstream ratio of ~3.
10 The amount of CO emitted will vary based on the type of tobacco product (cigarette/cigar), the
11 degree to which tobacco is actively smoked, and the amount of smoke being absorbed by the
12 lungs (Klepeis et al., 1996; Akbar-Khanzadeh and Greco, 1996).
13 The Federal Trade Commission compiled data on 933 varieties of cigarette manufactured
14 and sold in the United States in 1992. These data were provided by the various cigarette
15 manufacturers. Carbon monoxide emission rates for the brands of cigarettes reported ranged
16 from <0.5 to 23.0 mg per cigarette (cigarettes emitting 23.0 mg were unfiltered brands)
17 (Federal Trade Commission, 1994). Klepeis et al. (1995, 1996) measured CO concentrations in
18 airport smoking lounges under real life conditions. They estimated CO emissions to be 78 mg
19 per cigarette (ms and ss) on the basis of an average CO emission rate of 11.1 mg/min and a
20 smoking duration of 7 min. An estimated total CO emission rate of 81.2 mg for three cigarettes
21 (ms and ss) was reported by Ott et al. (1992). Lofroth et al. (1989) estimated a CO emission rate
22 of 67 mg per cigarette (ss) based on a cigarette weight of ~1.2 g and a smoking duration of
23 12 min. Large cigars emit substantially more CO than do cigarettes. Emission rates of 82 to
24 200 mg CO/g (mass smoked; ms and ss) were reported by Klepeis et al. (1999) (smoked by
25 machine and by a person). Cigar mass ranged from 5.9 to 16.7 g, and the smoking time was 7 to
26 40 min for the machine-smoked test and 78 and 90 min for the test measuring emissions for a
27 cigar being smoked by a person.
28
29
30
October 1999 3-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.5.3 Source-Related Concentrations of Carbon Monoxide in Indoor
2 Environments Prior to 1991
3 Microenvironments associated with motor vehicles usually result in the highest
4 concentrations of CO. Carbon monoxide concentrations of up to 28 ppm were reported in indoor
5 parking garages or indoor environments associated with attached garages (Akland et al., 1985;
6 Johnson et al., 1984; Wallace, 1983; Flachsbart and Ott, 1984). Carbon monoxide
7 concentrations inside moving vehicles can exceed the 8-h, 9 ppm and 1-h, 35 ppm NAAQS for
8 CO (Flachsbart et al., 1987; Chaney, 1978; Ziskind et al., 1981) and are generally higher in
9 personal vehicles than in public transportation (Flachsbart et al., 1987; Cortese and Spengler,
10 1976).
11 Based on the intermittent use of gas cooking stoves, average long-term concentrations of
12 CO are not expected to be significant (Research Triangle Institute, 1990; Koontz and Nagda,
13 1987). However, short-term peak concentrations of CO of 1.8 to 120 ppm have been reported
14 from the use of gas cooking stoves (Research Triangle Institute, 1990; Koontz and Nagda, 1987;
15 Leaderer et al., 1984; Moschandreas and Zabransky, 1982; Sterling and Sterling, 1979).
16 The use of unvented gas space heaters as primary heat sources is expect to exhibit higher
17 long-term concentrations of CO ranging from 0.26 to 9.49 ppm (mean) (McCarthy et al., 1987;
18 Koontz and Nagda, 1988). Peak CO concentrations from the use of unvented gas heaters were
19 also generally higher than unvented kerosene heaters and gas cooking stoves (Koontz and Nagda,
20 1987; Research Triangle Institute, 1990; Leaderer et al., 1984; Davidson et al., 1987).
21 Indoor concentrations of CO from the use of nonairtight woodburning stoves can contribute
22 as much as 9 ppm to the average indoor CO level (Traynor et al., 1984). Airtight stoves have
23 been shown to contribute from 0.1 to 2.0 ppm CO to the average CO background level
24 (Humphreys et al., 1986; Traynor et al., 1984)
25 Concentrations of CO in environments with smoking is highly variable, depending on the
26 type of environment, number of cigarettes smoked, and the type and amount of ventilation. Peak
27 CO concentrations of 32 (mechanical ventilation) and 41 ppm (natural ventilation) have been
28 measured in automobiles (Harke and Peters, 1974). However, while cigarettes are expected to
29 contribute to the indoor CO concentrations, the additions are not expected to be substantial
30 except when heavy smoking occurs in small spaces.
31
October 1999 3-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.5.4 Indoor Concentrations of Carbon Monoxide
2 3.5.4.1 Factors Affecting Carbon Monoxide Concentrations
3 A number of factors can affect indoor CO concentrations: the presence of a source and its
4 use pattern, pollutant emission rate, ambient air concentrations, infiltration through the building
5 envelope, air exchange rate (AER), building volume, and air mixing within the indoor
6 compartments. Information on the number of households using combustion appliances in the
7 United States in 1995 appear in Table 3-9.
TABLE 3-9. COMBUSTIBLE FUELS IN HOMES IN THE UNITED STATES IN 1995
Combustible Fuel Type
Heatinga
Cooking"
Central Air
Conditioning0
Clothes Dryerd
Water Heater6
Piped Gas
49,203,000
(848,000)
35,001,000
2,971,000
15,998,000
50,558,000
Bottled Gas Fuel Oil
4,251,000 10,974,000
(558,000) (451,000)
4,217,000 —
— —
3,239,000 5,808,000
Kerosene/Other
Liquid Fuel
1,055,000
(1,159,000)
301,000
—
331,000
Wood
3,533,000
(7,949,000)
33,000
—
44,000
aBased on 96,650,000 occupied housing units with heating fuel. Values in parenthesis represent use of fuel type
as secondary heating fuel.
bBased on 97,406,000 occupied housing units with cooking fuel.
"Based on 46,577,000 occupied housing units with central air conditioning.
dBased on 70,756,000 occupied housing units with central air conditioning.
eBased on 97,522,000 occupied housing with hot piped water.
Source: U.S. Census Bureau (1999).
1 The major sources of CO in residential environments are unvented gas or kerosene
2 appliances. Because gas cooking ranges are used intermittently for cooking purposes, it is not
3 likely that the use of gas ranges would result in substantial increases in CO over long periods of
4 time, except in households where gas cooking stoves have continuously burning pilots or are
October 1999
3-60
DRAFT-DO NOT QUOTE OR CITE
-------
1 used improperly as a primary or secondary source of heat. Koontz et al. (1992) reported the
2 results of a survey conducted in 1985 and 1991 designed to determine the prevalence of kitchen
3 fans and the factors affecting their use and the impact of other cooking appliances (i.e.,
4 microwave ovens, toaster ovens, hot plates) on the use of the range for cooking. The authors
5 reported a 27% increase in the use of gas ranges without standing pilot lights between 1985 and
6 1991 and a 20% reduction in the use of both electric and gas stoves for cooking. Ninety-five
7 percent of the households surveyed reported having another form of cooking appliance in
8 addition to the gas range, and, of this number, 55 to 65% reported using the stove less often.
9 There were, however, more people using a gas range for purposes of supplemental heating than
10 there were using electric ranges for that purpose (11% versus 3.6%). Estimates of gas cooking
11 stove usage range from 30 to 60 min/day. Table 3-10 contains information on the prevalence of
12 gas cooking stoves in the United States.
13 The use of unvented space heaters represent a significant source of CO in indoor
14 environments. Data from the National Health and Nutrition Examination Survey estimated that
15 13.7 million adults used unvented combustion space heaters between 1988 and 1994. Based on
16 the information obtained in the survey, an estimated 13.2% of the adult population in the
17 southern United States used unvented combustion space heaters. An estimated 5.9% of the adult
18 population in the Midwest, 4.2% in the Northeast, and 2.5% in the West used unvented space
19 heaters (Figure 3-30) (Slack and Heumann, 1997). The use of more unvented combustion space
20 heaters in the South also was reported in earlier studies by U.S. Department of Housing and
21 Urban Development (1987), U.S. Environmental Protection Agency (1992), and Williams et al.
22 (1992). This may be because in areas with relatively mild winters, combustion space heaters are
23 used frequently as the primary source of heat. The U.S. Census Bureau estimated that
24 1,055,000 and 1,159,000 households used kerosene or another liquid fuel as primary and
25 secondary heating fuels, respectively, in 1995 (U.S. Census Bureau, 1999). The U.S.
26 Environmental Protection Agency (1990) estimated that kerosene heaters are used 16.7 h/day in
27 southern states as primary sources of heat and, in regions where the heaters are used as secondary
28 heat sources, estimated use ranges from 2.6 to 10.7 h/day.
29 The AER, the balance of the flow of air in and out of a microenvironment, is based on the
30 fraction of air that enters the microenvironment through infiltration through unintentional
31 openings in the building envelope, natural ventilation through any designed opening in the
October 1999 3-61 DRAFT-DO NOT QUOTE OR CITE
-------
0
o
1
VO
VO
VO
LtJ
ON
to
o
£
H
1
O
o
2;
o
H
0
0
H
W
0
O
H
w
TABLE 3-10. PREVALENCE OF GAS COOKING RANGES
Fuel Region
Gas U.S. households
(1985)
Natural gas U.S. households
(1993)
Natural gas California
(1991, 1992)
Natural gas Los Angeles, CA
(1995)
Denver, CO
(1995)
Natural gas Washington, DC
(1982-1983)
Denver, CO
(1982-1983)
Natural gas U.S. households
(1995)
Number of
Households Percentage of
(total) Households Comments
929
(2,323)
33,813,000
(94,363,000)
142
(293)
2,614,000
(3,165,200)
151,300
(770,600)
609,029
(953,714)
85,542
(345,163)
35,001,000
(97,406,000)
4 1 Twenty percent of the 924 households had
electronic pilots in 1985. Twenty-seven percent of
the households had electronic pilots in 1991, based
on 886 respondents from the 1985 survey.
36
52 Fifty-two percent of the households examined used
gas for cooking. Fifty-one percent of the
households using gas cooking stoves had electronic
pilots.
82
20
64 Results are estimates of the pollution using gas
cooking stoves compared to the total population.
25
40
Reference
Koontz etal. (1992)
U.S. Census Bureau
(1998)
Wilson etal. (1993)
U.S. Census Bureau
(1998)
Johnson (1984)
Hartwell etal. (1984)
U.S. Census (1999)
-------
6
5-
8) 4H
&
0)
O
o-
Kerosene
Propane
Natural Gas
Northeast Midwest
South
West
Figure 3-30. Percentage of U.S. households using unvented combustion heaters, by type of
fuel, stratified by region (Third National Health and Nutrition Examination
Survey, 1988 to 1994).
Source: Slack and Heumann (1997).
1 building envelope (doors, windows), and forced ventilation systems. Infiltration is the dominant
2 mechanism for residential air exchange. Forced ventilation is typically the dominant mechanism
3 for air exchange in nonresidential buildings. Natural ventilation, airflow through doors and
4 opened windows, is seasonal (Koontz and Rector, 1995). Air exchange rates for some residential
5 buildings are listed in Table 3-11. Air exchange rates varied depending on the outside
October 1999
3-63
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-11. RESIDENTIAL AIR EXCHANGE RATES
Geographic Location
Riverside, CA
Los Angeles, CA
Northern California
San Diego, CA
United States
California
New York
(Suffolk and
Onondaga Counties)
Sample Size
175
571
(1 week)
426
(1 week)
372
(2 days)
75
(2 days)
128
(2 days)
85
(2 days)
2,844
293
245
AER
(h-1)
0.87 (day)
0.75 (night)
0.62
0.78
(March)
1.05
1.51
(July)
0.47
0.58
(January)
0.63
0.79
(winter)
0.41
(winter)
0.46
(winter)
0.76
(all seasons)
0.55
(winter)
0.65
(spring)
1.50
(summer)
0.41
(fall)
0.58
0.59
Type
Arithmetic
Geometric
Arithmetic
Geometric
Arithmetic
Geometric
Arithmetic
Geometric
Arithmetic
Arithmetic
Arithmetic
Arithmetic
Arithmetic
Arithmetic
Standard
Deviation
—
1.95
0.63
2.39
1.47
1.97
0.47
1.97
0.57
0.34
0.34
0.88
0.46
0.57
1.53
0.58
0.43
0.03
Reference
Sheldon etal. (1992)
Wilson etal. (1996)
Murray and
Burmaster (1995)
Colome etal. (1994)
Research Triangle
Institute (1990)
1 temperature, geographical location, type of cooking fuel used, type of heating system used, and
2 building type (Colome et al., 1994). Homes with gas cooking stoves with standing pilots and gas
3 wall furnaces had the highest AER. Homes with gas stoves without pilots had higher AER than
4 electric stoves. Also, homes with forced air furnaces had higher AER than electric homes.
October 1999
3-64
DRAFT-DO NOT QUOTE OR CITE
-------
1 Air exchange rates for nonresidential microenvironments also have been measured. Lagus
2 Applied Technology, Inc. (1995) reported AERs for 49 nonresidential buildings (14 schools,
3 22 offices, and 13 retail establishments) in California. Average mean (median) AERs were
4 2.45 (2.24), 1.35 (1.09), and 2.22 (1.79) h'1 for schools, offices, and retail establishments,
5 respectively. Air infiltration rates for 40 of the 49 buildings were 0.32, 0.31, and 1.12 IT1 for
6 schools, offices, and retail establishments, respectively. Air exchange rates for 40 nonresidential
7 buildings in Oregon and Washington (Turk et al., 1989) averaged 1.5 h'1 (mean [median =
8 1.3 tr1]). The geometric mean of the AERs for six garages was 1.6 h'1 (Marr et al., 1998). Park
9 et al. (1998) reported AERs for three stationary cars (cars varied by age) under different
10 ventilation conditions. Air exchange rates ranged from 1.0 to 3.0 h'1 for windows closed and fan
11 off, 13.3 to 23.5 tr1 for window opened and fan off, 1.8 to 3.7 h'1 for window closed and fan on
12 recirculation (two cars tested), and 36.2 to 47.5 h'1 for windows closed and fan on fresh air (one
13 car tested). An average AER of 13.1 h'1 was reported by Ott et al. (1992) for a station wagon
14 moving at 20 mph with the windows closed.
15
16 3.5.4.2 Models for Carbon Monoxide Concentrations
17 Indoor concentrations of CO can be estimated using the mass-balance model. The
18 mass-balance model estimates the concentration of a pollutant over time. The simplest form of
19 the model is represented by the following differential equation for a perfectly mixed
20 microenvironment.
21
^f =vCOUT+-^-vCIN. (3-1)
22
23 A more in-depth discussion of the mass-balance model may be found in U.S. Environmental
24 Protection Agency (1991b) and Nagda et al. (1987).
25 Traynor et al. (1989) used a model to predict CO concentrations in residential environments
26 for one pollutant source. Model inputs included ambient air concentrations, source emission
27 rates and usage characteristics, compartment volume, AERs, and outside temperatures. The
28 model combined the steady-state version of the mass-balance model used in indoor air quality
29 studies, a source-usage model for space heating appliances, and an air exchange model.
October 1999 3-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 A combination of the Monte Carlo and deterministic techniques was used to predict indoor
2 concentration distributions. Based on the modeled results, the use of kerosene heaters, unvented
3 gas space heaters, and gas ovens and ranges for heating produced the highest concentrations of
4 CO in the indoor environment (see Figure 3-31). The findings illustrated in Figure 3-31 are for
5 only a limited number of model runs, sources, and building conditions.
i i i i i
i — i —
i — i
i i i i i i i 11 r
P. - - Gas Oven/Range
~~° ' ' Without Hood
• 1 Gas Oven/Range
"^ ' With Hood
1% 10% GM
I—I © 1 1 Smoking
I—I—6—I—I Gas Domestic Hot Water
I—I—6—I—I Gas Dryer
-e-
-e-
Radiant
Kerosene Heater
• Gas Oven/Range
for Heating
Convective
Kerosene Heater
I 1
H e-
Infrared Uriverited
Gas Space Heater
Convective Unvented
Gas Space Heater
I h
-e-
H Nonairtight Woodstove
I h
-e-
H Airtight Woodstove
i—i—e
i—i—e-i-
H Gas Wall/Floor Furnace
1 Gas Boiler
i—i—e
H Gas Forced-Air Furnace
H 0
H Oil Forced-Air Furnace
I—I—9—I—I Electric
\ I I I I I I
0.3 1
\ \ I I I I I
10
CO Concentration (ppm)
\ Ii i i i i
100
\ \ r
600
Figure 3-31. Modeled indoor CO concentration distributions in houses with only one
indoor combustion pollutant source.
Source: Traynoretal. (1989).
1 3.5.4.3 Microenvironmental Monitoring Studies
2 Residential Carbon Monoxide Concentrations Related to Indoor Sources
3 Peak CO concentrations of 5.0 ppm (over ambient concentrations) from the use of gas
4 ranges and stoves were reported by Davidson et al. (1987), based on a survey of the literature.
October 1999
3-66
DRAFT-DO NOT QUOTE OR CITE
-------
1 The Gas Research Institute, Pacific Gas and Electric Company, San Diego Gas and Electric
2 Company, and Southern California Gas Company initiated indoor monitoring of 293 randomly
3 selected homes. Monitoring was for a single 48-h period. Carbon monoxide levels indoors were
4 reported to be closely associated with levels outdoors for most of the residences monitored.
5 However, 13 homes had CO concentrations above 9 ppm, and concentrations in one home
6 exceeded 35 ppm. Homes with gas ranges with standing pilot lights had higher CO
7 concentrations than did homes with gas ranges with electronic pilot lights or electric ranges.
8 Homes with standing pilots had a 0.56-ppm increase in net CO. Indoor minus outdoor CO
9 concentrations for six averaging times were used to ranked homes. Using that criterion, 21 of the
10 293 homes studied were selected for case studies. The higher CO seen in these homes possibly
11 was associated with occupant smoking, the use of gas stoves for heating purposes, infiltration
12 from attached garages, the type of heating system used (homes with gas wall furnaces had higher
13 CO), the building type and size (smaller multi-family homes had higher CO than larger single-
14 family homes), use of gas appliances, and more than one CO source. The average AER varied by
15 type of heating system (wall furnaces > forced-air > electric) and building type (multi-family
16 units > single-family units) (Billick et al., 1994, 1996; Colome et al., 1994). The CO descriptive
17 statistics for homes in this study are given in Table 3-12.
18 The Research Triangle Institute (1990) monitored CO concentrations in 400 homes for
19 3 days in Suffolk and Onondaga Counties, NY. Homes in Suffolk County were monitored in
20 January and February. The selected homes in Onondaga County were monitored in February,
21 March, and April. The average room volume was assumed to be 50 m3. The average AER was
22 0.59 h"1. Carbon monoxide monitors were placed in the primary living space and close to the
23 source. Approximately half of the homes used gas cooking stoves. Kerosene heaters had to be
24 operated at least 3 h/day to qualify as a CO source, and the woodstove or fireplace had to be
25 operated an average of 2 h/day. Any reported usage of gas stoves qualified them as sources. The
26 average CO concentration in the primary living area was 2.23 ±0.17 ppm (results for
27 209 homes). Use of both gas stoves and kerosene space heaters was associated with increased
28 CO. Homes using woodstoves or fireplaces had lower CO than did homes without woodstoves
29 or fireplaces (see Figure 3-32). An explanation for the finding of higher CO in homes without
30 wood stoves and fireplaces was not provided. Lower CO levels may have been associated with
31 an increased air exchange rate. Also, CO emissions for the use of wood stoves and fireplaces are
October 1999 3-67 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-12. CARBON MONOXIDE DESCRIPTIVE STATISTICS FOR ALL HOMES (number = 277; in ppm)
r+
O
cr
VO
VO
VO
OJ
ON
OO
O
H
6
o
0
H
O
O
0
o
All Homes
Arithmetic Mean
Standard Error
Mode
Percentiles
Maximum
95th
75th
50th
25th
5th
Minimum
Source: Modified
48-h
1.6
0.1
1.0
12.9
4.3
1.8
1.2
0.7
0.1
0.0
from Wilson et
Max
10-min
5.2
0.3
2.0
37.9
15.1
6.6
3.5
2.0
1.0
0.1
al. (1993).
Indoor Average
Max
30-min Maxl-h Max 8-h
4.8 4.5 2.9
0.3 0.3 0.2
2.0 2.0 1.0
36.7 35.8 23.5
14.2 13.2 8.3
6.0 5.8 3.4
3.1 3.0 2.0
2.0 2.0 1.2
1.0 1.0 0.5
0.0 0.0 0.0
Outdoor Average
Max Max
48-h 10-min 30-min Max 1-h
1.0 5.5 4.3 3.8
0.1 0.4 0.3 0.2
0.1 2.0 2.0 1.0
10.8 68.7 31.5 27.3
2.7 16.1 12.3 10.6
1.3 6.1 5.2 4.8
0.8 3.3 2.9 2.6
0.3 2.0 1.9 1.5
0.1 1.1 1.0 1.0
0.0 0.2 0.1 0.0
Max 8-h
2.0
0.1
1.0
17.3
6.3
2.2
1.4
0.9
0.3
0.0
-------
Q.
Q.
O
o
c
I
O
O
4.0
3.0
2.0 -
1.0 -
D Presence of Source
• Absence of Source
3.86
2.25
1.85
Onondaga County Suffolk County
Gas Stove
Suffolk County Onondaga County
Woodstove/
Fireplaces
Kerosene
Heaters
Figure 3-32. Arithmetic mean CO concentrations by presence or absence of combustion
source. Homes in Suffolk County were monitored in January and February
and in Onondaga County in February, March, and April.
Source: Research Triangle Institute (1990).
1 intermittent and will generally only occur during fire start-up and tending or through leaks in the
2 stove or venting system.
3 Hedrick and Krug (1995) reported CO concentrations from the use of unvented gas space
4 heaters and pilot lights in a test house in Chicago, IL. The gas space heaters included blue-flame
5 convective, radiant-tile, fan-forced blue-flame and perforated-tube convective units. Emission
October 1999
3-69
DRAFT-DO NOT QUOTE OR CITE
-------
1 rates for the units are discussed in Section 3.5.2.2. The house was a single-family, one-story,
2 three-bedroom dwelling with a full basement, comprising 1,150 ft2 per level. Eight burner
3 experiments and three pilot light experiments were conducted. Heaters were operated for
4 8 h followed by a 15-h decay period. The pilot studies were conducted over a 48-h period. Fans
5 were used to distribute emissions throughout the house, excluding the basement. The highest CO
6 concentrations were seen with the radiant-tile heater (13.4 ppm), and the lowest CO
7 concentrations were reported with the fan-forced unit (0.9 ppm). Two of the tests using the
8 blue-flame convective units were affected by gas leakage, resulting in CO concentrations of
9 4.7 and 4.8 ppm. The test not affected by the leakage had a CO concentration of 2.7 ppm. The
10 maximum CO concentration in the test house from use of the perforated-tube convective heater
11 was 31.9 ppm. Carbon monoxide concentrations during the pilot light experiments did not
12 exceed 2.0 ppm.
13 Moderately high concentrations of CO have been reported in homes where unvented
14 kerosene heaters are in use. Burton et al. (1990) monitored both the inside and outside of two
15 mobile homes for pollutants emitted from the operation of a radiant and a convective unvented
16 kerosene space heater. No other sources of combustion by-products were in the homes. Heaters
17 were operated from 4:00 to 9:00 p.m. daily. Six random sampling periods were conducted, three
18 with heaters on and three with heaters off. Measurements were made until 11:00 p.m. Average
19 CO concentrations while the heaters were in operation were 12 ppm for the convective heater and
20 4 ppm for the radiant heater. When the heaters were not in use, average CO concentrations were
21 5 and 1 ppm, respectively. Ambient CO concentrations in the mobile home park were reported
22 to be negligible.
23 Carbon monoxide concentrations in eight single-wide mobile homes (150 to 255 m3) were
24 reported by Mumford et al. (1990, 1991). Convective kerosene heaters were used in four of the
25 homes. Three of the homes used radiant heaters, and one home used a convective/radiant heater.
26 Monitoring was conducted 2.6 to 9.5 h/day (average 6.5 h/day) for 2 weeks with heaters on and
27 2 weeks with heaters off. Fuel consumption rates ranged from 252 to 295 kJ/min for convective
28 units, 105 to 168 kJ/min for radiant units, and 120 kJ/min for the convective/radiant unit.
29 Average AERs were 0.47 h"1 with heater on and 0.48 h"1 with heater off. Monitoring began
30 when heater use began and continued for 2 h after the heaters were turned off. Carbon
31 monoxide concentrations were above 9 ppm in four of the eight homes. In one home with a
October 1999 3-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 convective heater, CO concentrations peaked at 51 ppm. The average CO concentration with
2 heaters off was 1.4 ppm.
3 Williams et al. (1992) reported CO concentrations in eight all-electric mobile homes (each
4 <100 m2) from the use of kerosene heaters over a 6-day measurement period. The space heaters
5 were used for an average of 4.5 h/day between 4:00 to 11:00 p.m. Measured CO in homes
6 0.5 h prior to kerosene heater use ranged from 0 to 8 ppm. Average CO during nonuse days was
7 1.4 ppm ±0.3. Peak CO values ranged from 0.3 to 50.2 ppm. The 8-h average CO concentration
8 in the homes was 7.4 ±1.4 ppm. Peaks usually were observed at the end of the combustion
9 period. The AER averaged 0.47 h"1 when the unit was in operation. Homes with the radiant and
10 multistage units had higher CO than homes with convective heaters. Average 1-h levels of CO
11 for convective units ranged from 1.3 to 5.3 ppm. One-hour average CO concentrations for
12 radiant and multistage units ranged from 1.1 to 28.3 ppm and 16.0 to 50.0 ppm, respectively.
13
14 Carbon Monoxide Concentrations Related to Environmental Tobacco Smoke
15 Carbon monoxide concentrations in environments where smoking occurs exceed
16 background CO concentrations. The indoor concentrations will depend on the size of the indoor
17 space, number of cigarettes smoked, smoking rate, CO emission rate, differences in ventilation,
18 and the ambient CO concentrations (Turner et al., 1992). The U.S. Centers for Disease Control
19 and Prevention reported an estimated 61 million smokers in the United States in 1995,
20 representing 29% of the population. The percentage of smokers between 1994 and 1995 was
21 unchanged. In 1994, an estimated 1.5 million American became daily smokers. The estimated
22 number of new smokers per year has not changed since 1980 (Tobacco Information and
23 Prevention Source, 1996). The United States Department of Agricultural estimated that
24 487 billion cigarettes were sold in 1995 (Federal Trade Commission, 1997). An estimated
25 4.4 billion cigars were sold in the United States in 1997, compared to 3.8 billion in 1996 (Federal
26 Trade Commission, 1999). Ott et al. (1992, 1995) conducted a series of monitoring experiments
27 in a one-story house during development of a multi-compartment indoor mass-balance model to
28 predict the pollutant concentrations from environmental tobacco smoke. Smoking time ranged
29 from 6.5 to 9.5 min. Carbon monoxide concentrations were measured in three locations in the
30 bedroom after three cigarettes were smoked over a 9-h period. Concentrations ranged from
31 0.4 to 0.6 ppm. The only ventilation was a partially opened window covered with a shade
October 1999 3-71 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
12
13
14
(AER = 1.2 h4). Klepeis et al. (1995) reported a range of 0.41 to 1.2 ppm CO (average 0.75 ppm
CO) in airport smoking lounges based on 10 sampling periods ranging from 60 to 146 min. The
average number of people smoking during the period ranged from 2.8 to 13.5. The room
volumes ranged from 238 to 803 m3, with AERs of 12.8 and 15.8 h4. Holcomb (1993) reviewed
the literature on tobacco smoke in various indoor environments and evaluated those data the
authors defined as generated under real-life conditions. Carbon monoxide concentrations ranged
from 0.1 to 10.2 ppm depending on the indoor environment. The results are outlined in
Table 3-13. Lofroth et al. (1989) reported CO concentrations in a chamber test for cigarettes
(1 cigarette) smoked every 15 to 30 min. The chamber volume was 13.6 m3, with a set AER of
3.55 h"1. The cigarette mass was 1.2 g, and mass smoked was 0.9 to 1.0 g. Smoking duration was
-12 min per cigarette. Carbon monoxide concentrations averaged 1.56 and 2.17 ppm for
the 30-min tests and 4.16 ppm for the 15-min test.
TABLE 3-13. CARBON MONOXIDE CONCENTRATIONS (ppm) IN SMOKING (S)
AND NONSMOKING (NS) AREAS IN REAL-LIFE SITUATIONS
Category
Offices and
Public
Buildings
Restaurants
Taverns/Bars
Trains
Buses
Autos
No. of
Studies
13
5
2
2
1
1
Smoking Nonsmoking
Sample Sample
Size Mean Range Size Mean Range
697 2.95 0.1-8.7 275 2.99 0.7-4.0
107 3.6 0.4-9.0 — —
5 6.4 — — —
18 2.2 1.0-5.2 10 1.30 0.5-2.9
35 6.0 3.7-10.2 — —
— — — 213 11.6 8.8-22.3
Diff. in
Means
S-NS
-0.04
—
—
0.90
—
—
Source: Holcomb (1993).
1 Carbon monoxide concentrations approaching 20 ppm were reported by Ott et al. (1992) in
2 the compartment of a moving vehicle after three cigarettes were smoked over a 60 min period.
October 1999
3-72
DRAFT-DO NOT QUOTE OR CITE
-------
1 The vehicle traveled at 20 mi/h with the windows closed and the air conditioned set at
2 recirculation. Carbon monoxide concentrations reached almost 17 ppm after the first cigarette
3 had been smoked. An averaged CO concentration of 9 ppm, over a 200-min period, was reported
4 by Klepeis et al. (1999) during a cigar banquet. More than 100 cigars were smoked by
5 approximately 30 people. Concentrations were reported to range from 3 to 19 ppm under various
6 smoking conditions.
7
8 Carbon Monoxide Concentrations Associated with Motor Vehicles
9 In the United States, motor vehicles dominate total anthropogenic emissions of CO. Older
10 vehicles are likely to remit more CO than newer models. However, when the newer models with
11 catalytic converters are started, the fuel-air mixture is rich to facilitate ignition and to improve
12 cold engine operation. Also, the catalytic converter is not warm enough to function efficiently
13 (Marr et al., 1998). Emission rates for combustion vehicles are discussed in Section 3.3 of this
14 chapter.
15 Several studies have monitored the CO concentrations inside a moving vehicle under
16 various operating conditions. Chan et al. (1989, 1991) evaluated CO, NOX, O3, and VOCs inside
17 two moving vehicles (a 1983 and a 1987 model). Tailpipe emissions were higher for the older
18 model. Driving routes were selected to represent three distinct traffic patterns; urban traffic, high
19 traffic, and rural traffic. Inside ventilation was (1) windows and vents closed and air
20 conditioning on; (2) windows and vents closed and fan on; or (3) front windows half opened,
21 vent and fan on. Average CO concentrations for 70 samples, including both cars and all driving
22 routes, were 11.3 ppm. The in-vehicle concentrations were almost four times higher than
23 ambient CO (3.0 ppm). Carbon monoxide concentrations in the rural traffic pattern were
24 significantly less than either the urban and high-traffic patterns. Carbon monoxide in urban and
25 high-traffic patterns was not significantly different. The lowest CO concentrations were
26 measured when the windows were open; however, the concentration difference between the
27 different ventilation modes was only 1.0 ppm. Ott et al. (1994) reported an average CO
28 concentration of 10.2 ppm in the inside compartment of a moving vehicle for 93 trips under
29 urban highway conditions at varying times of the day. A more detailed discussion of this study
30 appears in the chapter on human exposure. Higher CO concentrations were reported both inside
31 and outside of moving vehicles in Sacramento and Los Angeles than those measured at roadside
October 1999 3-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 and ambient monitoring stations. Carbon monoxide concentrations inside vehicles ranged from
2 nondetectable to 6 ppm (Rodes et al., 1998). In-vehicle CO from deployment of four different air
3 bags was reported by Wheatley et al. (1997). The time-weighted-average CO concentration
4 ranged form 174 to 370 ppm. Peak CO concentrations occurred 2 min after deployment.
5 Kern et al. (1990) measured CO concentrations in a poorly sealed, detached garage from
6 operation of an emissions-controlled (catalytic reactor and oxygen sensor) and an emissions-
7 uncontrolled vehicle (carbureted, without a catalytic reactor). Two tests were conducted: Test 1,
8 poorly sealed garage door with a 3-in. crack, and Test 2, garage door sealed with rags. Carbon
9 monoxide concentrations in the poorly sealed garage reached 4,700 ppm for the uncontrolled car
10 versus 2,000 ppm for the controlled car. When the garage was better sealed, CO concentrations,
11 after 110 min of operation, reached 8,400 ppm for the uncontrolled vehicle versus 3,600 ppm for
12 the controlled vehicle.
13 Amendola and Hanes (1984) evaluated the concentration of CO in automotive repair shops
14 based on seasonal conditions and as a function of work environment size. Monitoring was
15 conducted in a small service station (1 to 2 bays), a large service station (>2 bays), and an
16 automobile dealership. The 8-h time-weighted average during warm weather ranged from 3.3 to
17 16.2 ppm, 3.4 to 21.6 ppm, and 12.1 to 20.8 ppm for the small and large service stations and the
18 dealership; however, the authors noted that CO concentrations were affected by the type of
19 ventilation used in the facility, volume and type of repairs, and employee work habits, such as
20 minimizing engine run time.
21
22
23 3.6 SUMMARY
24 Carbon monoxide is produced by the incomplete combustion of burning fossil and biomass
25 fuels. Approximately 70% of the CO produced globally is the result of human activities. Carbon
26 monoxide in the atmosphere is of both primary and secondary origin. The photochemical
27 oxidation of CH4 and NMHCs accounts for almost one-half of the total source strength of CO.
28 The uncertainty in estimates of the magnitudes of individual CO sources ranges from a factor of
29 two to three.
30 Atmospheric CO concentrations in remote areas of the world have been increasing at the
31 rate of about 1% per year throughout most of the industrial era. This increase reflects the growth
October 1999 3-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 of anthropogenic emissions from the combustion of fossil and biomass fuels and increased
2 agriculture to feed the expanding world population. However, CO levels decreased for several
3 years from the late 1980s to the early 1990s. The reasons for this decline are not clear, although
4 several factors may be involved. Since then, there has been no clear trend in CO levels.
5 Carbon monoxide plays an important role in atmospheric chemistry because it is the major
6 reactant for OH radicals. Reaction with OH radicals is the loss mechanism for many trace gases
7 that are responsible for contributing to the greenhouse effect (e.g., CH4) and for depleting
8 stratospheric O3 (e.g., CH3C1, and CH3Br). Thus, increases in CO concentrations can suppress
9 OH radical levels and allow the concentrations of these trace gases contributing to global-scale
10 environmental problems to increase, even if trace gas emissions are constant. Conversely,
11 decreases in global average CO concentrations can stabilize or even reverse the growth rates of
12 the trace gases mentioned above.
13 Carbon monoxide may be responsible for 20 to 40% of the O3 formed in the background or
14 "clean" troposphere. In addition, CO may have been responsible for 10 to 20% of the O3 formed
15 during smog episodes in the few urban areas that have been examined. Obviously, the
16 photochemistry must be examined in more cities before any more general statements about the
17 importance of CO for urban air chemistry can be made.
18 Emissions from transportation dominate other sources of CO within the United States.
19 Even though CO in urban areas results largely from motor vehicle emissions, a sizable fraction of
20 the CO observed in rural air may be produced by the photochemical oxidation of isoprene and
21 other NMHCs, at least according to one model study. Residential wood burning may be an
22 important source of CO in a number of urban areas. The uncertainties in the magnitudes of
23 individual sources in the nationwide and worldwide emission inventories are comparable (i.e.,
24 roughly a factor of between two and three).
25 There has been a consistent decrease in the nationwide annual second highest maximum
26 8-h composite average ambient CO concentration over the past 20 years, from about 11 ppm in
27 1977 to about 4 ppm in 1996. This improvement in CO quality occurred despite a 121% increase
28 in vehicle miles traveled, a 29% increase in population, and a 104% increase in gross domestic
29 product in the United States over the same period. During the past 10 years, the composite mean
30 annual CO second maximum 8-h concentration decreased 37% at 190 urban sites, 37% at
31 142 suburban locations, and 48% at 10 rural monitoring sites. Hourly average CO concentrations
October 1999 3-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 decreased from 2.0 ppm to 1.2 ppm over the past 10 years. Despite uncertainties in the
2 calculations of CO emissions, the decline in ambient CO concentrations in the United States
3 reflects the controls placed on automotive emissions. These declines are seen clearly in the
4 trends in CO levels during times of day when CO concentrations result mainly from mobile
5 source emissions.
6 The patterns and trends of observed CO reflect reductions in the CO emissions of the past
7 11 years. However, it is important to note that the reported concentrations from the monitoring
8 sites are representative only of the air quality in their neighborhoods. Also, although personal
9 exposure to CO from mobile sources also should be decreasing, the CO values from the
10 monitoring sites are not equivalent to personal exposures. The same ratios of personal to
11 monitored CO from past studies in urban areas with CO emissions dominated by mobile sources
12 may remain applicable today, but continued validation is needed.
13 Carbon monoxide occurs in indoor environments directly through emissions from various
14 indoor combustion sources or indirectly as a result of infiltration or ventilation from outdoor
15 sources. Carbon monoxide concentrations in the indoor compartment is influenced by the CO
16 emission rate of the unvented combustion source, the ambient CO concentration, infiltration
17 through the building envelope, building volume, AER, and air mixing within the indoor
18 compartments. In the absence of an indoor source, CO concentrations generally will equal those
19 in the surrounding ambient environment.
20 In 1995, an estimated 40% of the United States population used gas for cooking. Carbon
21 monoxide emissions from gas ranges vary from range to range for both the top burners and the
22 oven burners and are dependent on the type of pilot light, the fuel consumption rate, the
23 frequency of use, and the operating condition. Older gas ranges with standing pilot lights emit
24 more CO than do newer units with electronic pilot lights. Carbon monoxide emissions from the
25 use of gas cooking stoves are 15 to 215 and 53 to 344 //g/kJ for range top burners with
26 blue-flame and yellow-tipping operating conditions, respectively. Emissions from oven use
27 ranged from 12 to 226 //g/kJ for blue-flame and 54 to 62 //g/kJ for yellow-tipping. Carbon
28 monoxide emissions for both burner and oven standing pilots were only reported for blue flame
29 operating conditions and ranged from 28 to 56 and 209 to 322 //g/kJ for burners and ovens,
30 respectively. Estimates of fuel consumption range from 5,000 to 5,710 ft3/year for burners and
31 standing pilots. Annual pilot light fuel consumption for standing pilots has been estimated to be
October 1999 3-76 DRAFT-DO NOT QUOTE OR CITE
-------
1 3,530 ft3/y. A steady decrease in the emission of CO from pilot lights will likely occur with the
2 replacement of older gas ranges with new models without standing pilots. Given the decrease in
3 the number of individuals using gas stoves, the amount of time spent cooking on gas stoves, and
4 the replacement of gas stoves with standing pilots with electronic pilots, it is likely that CO
5 emission from gas stoves will continue to decrease. Also, with the advent of other cooking
6 appliances (e.g., microwaves, toaster ovens, heating plates), the use of ranges in meal preparation
7 is decreasing. Estimates of gas stove usage range from 30 to 60 min/day. Carbon monoxide
8 emissions from gas water heaters and furnaces are generally negligible, but less efficient models
9 will emit more CO than the more efficient ones.
10 The use of well-maintained, energy-efficient gas stoves will result in only intermittent,
11 small increases in CO concentrations. In one report, CO concentrations from gas stove usage
12 ranged from 0.65 to >9.0 ppm. The high levels of CO likely were associated with the presence of
13 multiple CO sources, the use of the gas stove as a supplemental heat source, infiltration from
14 ambient sources, and the building type (multi-family units versus single-family units).
15 Short-term peak concentrations have been noted as high as 120 ppm.
16 Emissions from unvented space heaters are a function of the appliance design, combustion
17 efficiency, length and frequency of use, and the fuel type and consumption rate. Unvented gas
18 and kerosene space heaters generate the highest CO emissions and concentrations in the indoor
19 environment. Data covering the years 1988 to 1994 indicate that 13.7 million adults used
20 unvented space heaters in the United States. The largest percentage of unvented space heaters
21 were in southern states. Data from a recent survey by the U.S. Census Bureau indicates that
22 1.06 million households used unvented space heaters (heating fuel described and kerosene or
23 other liquid fuel) during 1995 as the primary heat source and 1.2 million households used
24 unvented space heaters as a secondary heat source. Catalytic gas and convective kerosene space
25 heaters emit the smallest amount of CO. Radiant and infrared unvented space heaters emit
26 higher amounts of CO. The CO concentration from the use of a radiant-tile gas unit was
27 13.4 ppm, compared with 0.9 ppm for a convective gas unit. However, CO concentrations also
28 depend on the frequency and duration of use of the space heater. Higher CO is found in homes
29 where unvented space heaters are the primary heat source (up to 37 ppm) versus homes where
30 unvented space heaters are used to supplement another heat source (<9.0 ppm). The decreasing
31 usuage of unvented space heaters is likely to result in decreased CO emissions from this source.
October 1999 3-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 Available studies on woodstoves and fireplaces indicate that CO is emitted from fire
2 start-up and maintenance, leaks in the stove and venting system, and through backdrafting when
3 the natural draft is overcome by depressurization. The average CO source strength for airtight
4 stoves ranged from 0.08 to 0.27 g/h (10 to 140 cnrVh) versus 0.32 to 2.18 g/h (220 to
5 1,800 cm3/h) for nonairtight stoves. Carbon monoxide emissions of 0.07 to 0.375 g/h were
6 reported for wood heaters.
7 Carbon monoxide from the combustion of tobacco products occurs in the indoor
8 environment primarily through sidestream smoke. The Center for Disease Control estimated that
9 29% of the U.S. population smoked in 1995. An estimated 487 billion cigarettes were sold
10 during that year. An estimated 4.4 billion cigars were sold in the U.S. in 1997, compared to
11 3.8 billion in 1996. Emissions will vary based on the type and brand of tobacco product.
12 Concentrations of CO in environments where smoking occurs will exceed background
13 concentrations but will be dependent on the CO emission rate of the tobacco product, number of
14 cigarettes smoked, smoking rate, size of the indoor compartment, ventilation rate, and ambient
15 CO concentrations.
16 Carbon monoxide emissions from the use of combustion engines may produce significant
17 increases in CO in the microenvironments where the engines are being operated. Emissions from
18 the operation of motor vehicles in enclosed, inadequately ventilated spaces such as garages and
19 repair shops have resulted in 8-h time-weighted average CO concentrations of up to 22 ppm in
20 the summer and 110 ppm during the winter.
21
October 1999 3-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Akbar-Khanzadeh, F.; Greco, T. M. (1996) Health and social concerns of restaurant/bar workers exposed to
3 environmental tobacco smoke. Med. Lav. 87: 122-132.
4 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
5 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
6 Technol. 19:911-918.
7 Allen, D. I; Kasibhatla, P.; Thompson, A. M.; Rood, R. B.; Doddridge, B. G.; Pickering, K. E.; Hudson, R. D.; Lin,
8 S.-J. (1996) Transport-induced interannual variability of carbon monoxide determined using a chemistry and
9 transport model. J. Geophys. Res. [Atmos.] 101: 28,655-28,669.
10 Amendola, A. A.; Hanes, N. B. (1984) Characterization of indoor carbon monoxide levels produced by the
11 automobile. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of the 3rd international
12 conference on indoor air quality and climate, v. 4, chemical characterization and personal exposure; August;
13 Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building Research; pp. 97-102. Available
14 from: NTIS, Springfield, VA; PB85-104214.
15 Andreae, M. O. (1991) Biomass burning: its history, use, and distribution and its impact on environmental quality
16 and global climate. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
17 implications. Cambridge, MA: MIT Press; pp. 1-21.
18 Apte, M. G.; Traynor, G. W. (1993) Pollutant emission rates from a radiant fiber-matrix gas burner. Air Waste
19 43:223-225.
20 Bakwin, P. S.; Tans, P. P.; Novelli, P. C. (1994) CO budget in the northern hemisphere. Geophys. Res. Lett.
21 21:433-436.
22 Bekki, S.; Law, K. S.; Pyle, J. A. (1994) Effect of ozone depletion on atmospheric and CO. Nature 371: 595-597.
23 Belles, F. E.; DeWerth, D. W.; Himmel, R. L. (1979) NOx emissions of furnaces and water heaters predominant in
24 the California market. Presented at: 72nd annual meeting of the Air Pollution Control Association; June;
25 Cincinnati, OH. Pittsburgh, PA: Air Pollution Control Association; paper no. 79-60.1.
26 Billick, L; Johnson, D.; Moschandreas, D.; Relwani, S. (1984) An investigation of operational factors that influence
27 emission rates from gas appliances. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of
28 the 3rd international conference on indoor air quality and climate, v. 4, chemical characterization and
29 personal exposure; August; Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building
30 Research; pp. 181-188. Available from: NTIS, Springfield, VA; PB85-104214.
31 Billick, I. H.; Behrens, D. W.; Becker, E. W.; Garrison, C. A.; Wilson, A. L.; Colome, S. D.; Tian, Y. (1996)
32 California residential indoor air quality study: an overview. In: Proceedings of the 1995 international gas
33 research conference; November 1995; Cannes, France. Rockville, MD: Government Institutes, Inc.;
34 pp. 1707-1716.
35 Bishop, G. A.; Stedman, D. H. (1996) Measuring the emissions of passing cars. Ace. Chem. Res. 29: 489-495.
36 Borrazzo, J. E.; Davidson, C. I.; Hendrickson, C. T. (1987) A statistical analysis of published gas stove emission
37 factors for CO, NO, and NO2. In: Seifert, B.; Esdorn, H.; Fischer, M.; Rtiden, H.; Wegner, J., eds. Indoor air
38 '87: proceedings of the 4th international conference on indoor air quality and climate. Volume 1: volatile
39 organic compounds, combustion gases, particles and fibres, microbiological agents; August; Berlin, Federal
40 Republic of Germany. Berlin, Federal Republic of Germany: Institute for Water, Soil and Air Hygiene;
41 pp. 316-320.
42 Bradley, K. S.; Stedman, D. H.; Bishop, G. A. (1999) A global inventory of carbon monoxide emissions from motor
43 vehicles. Chemosphere: in press.
44 Briihl, C.; Crutzen, P. J. (1999) Reductions in the anthropogenic emissions of CO and their effect on CH4.
45 Chemosphere: in press.
46 Burton, R. M.; Seila, R. A.; Wilson, W. E.; Pahl, D. A.; Mumford, J. L.; Koutrakis, P. (1990) Characterization of
47 kerosene heater emissions inside two mobile homes. In: Indoor air '90: precedings of the 5th international
48 conference on indoor air quality and climate, volume 2, characteristics of indoor air; July-August; Toronto,
49 ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air Quality and Climate, Inc.;
50 pp. 337-342.
51
October 1999 3-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 Carroll, M. A.; Hastie, D. R.; Ridley, B. A.; Rodgers, M. O.; Torres, A. L.; Davis, D. D.; Bradshaw, J. D.;
2 Sandholm, S. T.; Schiff, H. L; Karecki, D. R.; Harris, G. W.; Mackay, G. L; Gregory, G. L.; Condon, E. P.;
3 Trainer, M.; Hubler, G.; Montzka, D. D.; Madronich, S.; Albritton, D. L.; Singh, H. B.; Beck, S. M.;
4 Shipham, M. C.; Bachmeier, A. S. (1990) Aircraft measurement of NOX over the eastern Pacific and
5 continental United States and implications for ozone production. J. Geophys. Res. [Atmos.] 95:
6 10,205-10,233.
7 Chameides, W. L.; Liu, S. C.; Cicerone, R. J. (1977) Possible variations in atmospheric methane. J. Geophys. Res.
8 82: 1795-1798.
9 Chan, C.-C.; Ozkaynak, H.; Spengler, J. D.; Sheldon, L.; Nelson, W.; Wallace, L. (1989) Commuter's exposure to
10 volatile organic compounds, ozone, carbon monoxide, and nitrogen dioxide. Presented at: 82nd annual
11 meeting of the Air and Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air and Waste
12 Management Association; paper no. 89-34A.4.
13 Chan, C.-C.; Ozkaynak, H.; Spengler, J. D.; Sheldon, L. (1991) Driver exposure to volatile organic compounds, CO,
14 ozone and NO2 under different driving conditions. Environ. Sci. Technol. 25: 964-972.
15 Chaney, L. W. (1978) Carbon monoxide automobile emissions measured from the interior of a traveling automobile.
16 Science (Washington, DC) 199: 1203-1204.
17 Cicerone, R. J. (1988) How has the atmospheric concentration of CO changed? In: Rowland, F. S.; Isaksen, I. S. A.,
18 eds. The changing atmosphere: report of the Dahlem workshop on the changing atmosphere; November 1987;
19 Berlin, Federal Republic of Germany. Chichester, United Kingdom: John Wiley & Sons, Ltd.; pp. 49-61.
20 (Physical, chemical, and earth sciences research report no. 7).
21 Cohen, J.; Iwamiya, R. (1998) Analyses of diurnal patterns in hourly and eight-hourly average ambient carbon
22 monoxide concentrations. Research Triangle Park, NC: U.S. Environmental Protection Agency, Air Quality
23 Trends Analysis Group; contract no. EPA 68-D7-0066.
24 Collins, W. J.; Stevenson, D. S.; Johnson, C. E.; Derwent, R. G. (1997) Tropospheric ozone in a global-scale
25 three-dimensional Lagrangian model and its response to NO^ emission controls. J. Atmos. Chem.
26 26:223-274.
27 Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
28 monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research Institute;
29 report no. GRI-93/0224.3. Available from: NTIS, Springfield, VA; PB96-159421.
30 Conrad, R. (1996) Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CH4, OCS, N2O and
31 NO). Microbiol. Rev. 60: 609-640.
3 2 Conrad, R.; Seiler, W. (1985) Influence of temperature, moisture, and organic carbon on the flux of H2 and CO
33 between soil and atmosphere: field studies in subtropical regions. J. Geophys. Res. [Atmos.] 90: 5699-5709.
34 Cortese, A. D.; Spengler, J. D. (1976) Ability of fixed monitoring stations to represent personal carbon monoxide
35 exposure. J. AirPollut. Control Assoc. 26: 1144-1150.
36 Cote, W. A.; Wade, W. A., Ill; Yocom, J. E. (1974) A study of indoor air quality. Washington, DC: U.S.
37 Environmental Protection Agency, Office of Research and Development; EPA report no. EPA-650/4-74-042.
38 Available from: NTIS, Springfield, VA; PB-238556.
39 Croes, B. E.; Effa, R. C.; Pasek, R. J. (1996) A comparison of the California emission inventory with 1995 ambient
40 concentration data in Los Angeles. Presented at: 6th CRC on-road vehicle emissions workshop; March; San
41 Diego, CA. Sacramento, CA: California Air Resources Board.
42 Davidson, C. I.; Borrazzo, J. E.; Hendrickson, C. T. (1987) Pollutant emission factors for gas stoves: a literature
43 survey. Research Triangle Park, NC: U.S. Environmental Protection Agency, Air and Energy Engineering
44 Research Laboratory; EPA report no. EPA-600/9-87-005. Available from: NTIS, Springfield, VA;
45 PB87-171328.
46 Dignon, J.; Ando, T.; Brenninkmeijer, C.; Conny, J.; Granier, C.; Khalil, A.; King, G.; King, S.; Law, K.; Levine,
47 J. S.; Rockmann; Stedman, D.; Yung, Y.; Zafiriou, O. (1998) Carbon monoxide sources and sinks: working
48 group summary report. In: Khalil, M. A. K.; Pinto, J. P.; Shearer, M. J., eds. Proceedings of the international
49 conference on atmospheric carbon monoxide and its environmental effects; December, 1997; Portland,
50 Oregon. Washington, DC: U.S. Environmental Protection Agency; report no. EPA/600/R-98/047;
51 pp. 359-374.
52 Duret, M. J.; Tidball, R. K. (1990) Concept development for low emissions space heater: final report (May 1, 1988 -
53 August 31, 1989). Chicago, IL: Gas Research Institute; report no. GRI-90/0164.
54
October 1999 3-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fan, C.; Zhang, J.; Cheung, J. P. (1997) Air pollutants emitted from household combustion sources.
2 In: Measurement of toxic and related air pollutants. Volume 2: proceedings of an A&WMA/EPA specialty
3 conference; April-May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste Management Association;
4 pp. 837-846. (A&WMA publication VIP-74).
5 Federal Trade Commission. (1994) Tar, nicotine, and carbon monoxide of the smoke of 933 varieties of domestic
6 cigarettes. Washington, DC: Federal Trade Commission.
7 Federal Trade Commission. (1997) Federal Trade Commission report to Congress for 1995, pursuant to the Federal
8 Cigarette Labeling and Advertising Act. Washington, DC: Federal trade Commission; August 7. Available:
9 www.ftc.gov/os/1997/9708/cigrep97.htm.
10 Federal Trade Commission. (1999) Cigar sales and advertising and promotional expenditures for calendar years
11 1996 and 1997. Washington, DC: Federal Trade Commission. Available:
12 www.ftc.gov/os/1999/9907/cigarreportl999.htm.
13 Flachsbart, P. G.; Ott, W. R. (1984) Field surveys of carbon monoxide in commercial settings using personal
14 exposure monitors. Washington, DC: U.S. Environmental Protection Agency, Office of Monitoring Systems
15 and Quality Assurance; EPA report no. EPA-600/4-84-019. Available from: NTIS, Springfield, VA;
16 PB84-211291.
17 Flachsbart, P. G.; Mack, G. A.; Howes, J. E.; Rodes, C. E. (1987) Carbon monoxide exposures of Washington
18 commuters. JAPCA 37: 135-142.
19 Fortmann, R. C.; Borrazzo, J. E.; Davidson, C. I. (1984) Characterization of parameters influencing indoor pollutant
20 concentrations. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of the 3rd
21 international conference on indoor air quality and climate, v. 4, chemical characterization and personal
22 exposure; August; Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building Research;
23 pp. 259-264. Available from: NTIS, Springfield, VA; PB85-104214.
24 Fuglestvedt, J. S.; Jonson, J. E.; Isaksen, I. S. A. (1994) Effect of reductions in stratospheric ozone on tropospheric
25 chemistry through changes in photolysis rates. Tellus B 46: 172-192.
26 Fujita, E. M.; Croes, B. E.; Bennett, C. L.; Lawson, D. R.; Lurmann, F. W.; Main, H. H. (1992) Comparison of
27 emission inventory and ambient concentration ratios of CO, NMOG, and NOX in California's south coast air
28 basin. J. Air Waste Manage. Assoc. 42: 264-276.
29 Grosjean, D.; Fung, K. (1984) Hydrocarbons and carbonyls in Los Angeles air. J. Air Pollut. Control Assoc.
30 34: 537-543.
31 Haan, D.; Martinerie, P.; Raynaud, D. (1996) Ice core data of atmospheric carbon monoxide over Antarctica and
32 Greenland during the last 200 years. Geophys. Res. Lett. 23: 2235-2238.
33 Hallock-Waters, K. A.; Doddridge, B. G.; Dickerson, R. R.; Spitzer, S.; Ray, J. D. (1999) Carbon monoxide in the
34 U.S. mid-Atlantic troposphere: evidence for a decreasing trend. Geophys. Res. Lett. 26: 2861-2864.
35 Hameed, S.; Pinto, J. P.; Stewart, R. W. (1979) Sensitivity of the predicted CO-OH-CH4 perturbation to
36 tropospheric NO^ concentrations. J. Geophys. Res. C: Oceans Atmos. 84: 763-768.
37 Hard, T. M.; Mehrabzadeh, A. A.; Chan, C. Y.; O'Brien, R. J. (1992) PAGE measurements of tropospheric HO with
38 measurements and model of interferences. J. Geophys. Res. [Atmos.] 97: 9795-9817.
39 Hard, T.; Rasmussen, R. A.; Bergamaschi, P.; Gros, V.; Hahn, M.; Herman, R.; Hoor, P.; Klouda, G.; Novelli, P.;
40 Raynaud, D.; Reichle, H.; Shearer, M.; Waibel, A.; Yurganov, L. (1998) Distributions and trends: working
41 group summary report. In: Khalil, M. A. K.; Pinto, J. P.; Shearer, M. J., eds. Proceedings of the international
42 conference on atmospheric carbon monoxide and its environmental effects; December, 1997; Portland,
43 Oregon. Washington, DC: U.S. Environmental Protection Agency; report no. EPA/600/R-98/047;
44 pp. 173-183.
45 Harke, H.-P.; Peters, H. (1974) Zum Problem des Passivrauchens. III. Uber den EinfluB des Rauchens auf die
46 CO-Konzentration im Kraftfahrzeug bei Fahrten im Stadtgebiet [The problem of passive smoking. III. The
47 influence of smoking on the CO concentration in driving automobiles]. Int. Arch. Arbeitsmed. 33: 221-229.
48 Hartwell, T. D.; Clayton, C. A.; Ritchie, R. M.; Whitmore, R. W.; Zelon, H. S.; Jones, S. M.; Whitehurst, D. A.
49 (1984) Study of carbon monoxide exposure of residents of Washington, DC and Denver, Colorado. Research
50 Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory;
51 EPA report no. EPA-600/4-84-031. Available from: NTIS, Springfield, VA; PB84-183516.
52
October 1999 3-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 Haste, T. L.; Chinkin, L. R.; Kumar, N.; Lurmann, F. W.; Hurwitt, S. B. (1998) Use of ambient data to evaluate a
2 regional emission inventory for the San Joaquin Valley in California. Chow, J.; Koutrakis, P., eds. PM2.5:
3 a fine particle standard: proceedings of an international specialty conference; January; Long Beach, CA., v. 1.
4 Pittsburgh, PA: Air & Waste Management Association; pp. 451-462. (AWMA specialty conference;
5 VIP-81).
6 Hedrick, R. L.; Krug, E. K. (1995) Conventional Research House measurements of emissions from unvented gas
7 space heaters. Chicago, IL: Gas Research Institute; report no. GRI-95/0205. Available from: NTIS,
8 Springfield, VA; PB95-266466
9 Himmel, R. L.; DeWerth, D. W. (1974) Evaluation of the pollutant emissions from gas-fired ranges. Project
10 EP-1-23: analysis of flue products from gas-fired appliances. Cleveland, OH: American Gas Association
11 Laboratories; report no. 1492.
12 Hofmann, D. J.; Peterson, J. T.; Rosson, R. M, eds. (1996) Climate monitoring and diagnostics laboratory no. 23.
13 Summary report 1994-95. Boulder, CO: U.S. Department of Commerce, National Oceanic and Atmospheric
14 Administration, Environmental Research Laboratories.
15 Holcomb, L. C. (1993) Indoor air quality and environmental tobacco smoke: concentration and exposure. Environ.
16 Int. 19: 9-40.
17 Houghton, J. T.; Callander, B. A.; Varney, S. K. (1992) Climate change 1992: the supplementary report to the IPCC
18 scientific assessment. New York, NY: Cambridge University Press.
19 Humphreys, M. P.; Knight, C. V.; Pinnix, J. C. (1986) Residential wood combustion impacts on indoor carbon
20 monoxide and suspended particulates In: Proceedings of the 1986 EPA/APCA symposium on measurement
21 of toxic air pollutants; April; Raleigh, NC. Research Triangle Park, NC: U.S. Environmental Protection
22 Agency, Environmental Monitoring Systems Laboratory; pp. 736-747; report no. EPA-600/9-86-013.
23 Available from: NTIS, Springfield, VA; PB87-182713.
24 Ingalls, M. N.; Smith, L. R.; Kirksey, R. E. (1989) Measurement of on-road vehicle emission factors in the
25 California South Coast Air Basin: volume I, regulated emissions. Atlanta, GA: Coordinating Research
26 Council, Inc.; report no. CRC-APRAC-AP-4-SCAQS-l. Available from: NTIS, Springfield, VA;
27 PB89-220925.
28 Inman, R. E.; Ingersoll, R. B.; Levy, E. A. (1971) Soil: a natural sink for carbon monoxide. Science (Washington,
29 DC) 172: 1229-1231.
30 Ipps, D. T.; Popejoy, D. (1998) An updated verification of the on-road vehicle emissions inventory for the Los
31 Angeles basin using ambient air quality data. Presented at: The 8th CRC on-road vehicle emissions
32 workshop; April; San Diego, CA. Sacramento, CA: California Air Resources Board.
33 Ito, K.; Thurston, G. D.; Nadas, A.; Lippmann, M. (1998) Multiple pollutants' spatio-temporal exposure
34 characterization errors. In: Measurement of toxic and related air pollutants: proceedings of a specialty
35 conference, volume I; September; Gary, NC; pp. 53-63. (A&WMA publication VIP-85).
36 Jaasma, D. R.; Stern, C. H.; Honeycutt, D. J.; Fugler, D. (1995) Effect of glass doors on carbon monoxide spillage
3 7 from masonry fireplaces. Presented at: 88th annual meeting and exhibition of the Air & Waste Management
38 Association; June; San Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no.
39 95-RP137.01.
40 Jacob, D. J.; Wofsy, S. C. (1990) Budgets of reactive nitrogen, hydrocarbons, and ozone over the Amazon forest
41 during the wet season. J. Geophys. Res. [Atmos.] 95: 16,737-16,754.
42 Jeffries, H. E. (1995) Photochemical air pollution. In: Singh, H. B., ed. Composition, chemistry, and climate of the
43 atmosphere. New York, NY: Van Nostrand-Reinhold; p. 337.
44 Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle Park,
45 NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; EPA report no.
46 EPA-600/4-84-014. Available from: NTIS, Springfield, VA; PB84-146125.
47 Johnson, D.; Billick, I.; Moschandreas, D.; Relwani, S. (1984) Emission rates from unvented gas appliances.
48 In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of the 3rd international conference on
49 indoor air quality and climate, v. 4, chemical characterization and personal exposure; August; Stockholm,
50 Sweden. Stockholm, Sweden: Swedish Council for Building Research; pp. 367-373. Available from: NTIS,
51 Springfield, VA; PB85-104214.
52 Kanakidou, M.; Crutzen, P. J. (1993) Scale problems in global tropospheric chemistry modeling: comparison of
53 results obtained with a three-dimensional model, adopting longitudinally uniform and varying emissions of
54 NOX and NMHC. Chemosphere 26: 787-801.
October 1999 3-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kanakidou, M; Singh, H. B.; Valentin, K. M.; Crutzen, P. J. (1991) A two-dimensional study of ethane and
2 propane oxidation in the troposphere. J. Geophys. Res. [Atmos.] 96: 15,395-15,413.
3 Kern, R. A.; Knoedler, J. A.; Trester, R. J. (1990) Automotive carbon monoxide emissions in a closed garage.
4 Presented at: International congress and exposition; February-March; Detroit, MI. Warrendale, PA: Society
5 of Automotive Engineers; pp. 33-38. (SAE technical paper series no. 900499).
6 Khalil, M. A. K.; Rasmussen, R. A. (1988a) Carbon monoxide in the earth's atmosphere: indications of a global
7 increase. Nature (London) 332: 242-245.
8 Khalil, M. A. K.; Rasmussen, R. A. (1988b) Carbon monoxide in an urban environment: application of a receptor
9 model for source apportionment. JAPCA 38: 901-906.
10 Khalil, M. A. K.; Rasmussen, R. A. (1989) Urban carbon monoxide: contributions of automobiles and wood
11 burning. Chemosphere 19: 1383-1386.
12 Khalil, M. A. K.; Rasmussen, R. A. (1994) Global decrease in atmospheric carbon monoxide concentration. Nature
13 (London) 370: 639-641.
14 Khalil, M. A. K.; Rasmussen, R. A. (1999) Sources of carbon monoxide in urban areas: measurements in the
15 GLOBE program. Chemosphere: in press.
16 King, G. M. (1999) Characteristics and significance of atmospheric carbon monoxide consumption by soils.
17 Chemosphere: in press.
18 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
19 monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
20 meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh,
21 PA: Air & Waste Management Association; paper no. 95-MPH.03. Available from: NTIS, Springfield, VA;
22 PB95-225496.
23 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
24 lounges. Environ. Sci. Technol. 30: 2813-2820.
25 Klepeis, N. E.; Ott, W. R.; Repace, J. L. (1999) The effect of cigar smoking on indoor levels of carbon monoxide
26 and particles. J. Exposure Anal. Environ. Epidemiol.: accepted.
27 Koontz, M. D.; Nagda, N. L. (1987) Survey of factors affecting NO2 concentrations. In: Seifert, B.; Esdorn, H.;
28 Fischer, M.; Rtiden, H.; Wegner, J., eds. Indoor air '87: proceedings of the 4th international conference on
29 indoor air quality and climate, v. 1, volatile organic compounds, combustion gases, particles and fibres,
30 microbiological agents; August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of Germany:
31 Institute for Water, Soil and Air Hygiene; pp. 430-434.
32 Koontz, M. D.; Nagda, N. L. (1988) A topical report on a field monitoring study of homes with unvented gas space
33 heaters. Volume III—methodology and results. Chicago, IL: Gas Research Institute; report no.
34 GRI-87/0044.2. Available from: NTIS, Springfield, VA; PB88-216791.
3 5 Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
36 Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics;
37 contract nos. 68-D9-0166 and 68-D3-0013.
3 8 Koontz, M. D.; Mehegan, L. L.; Nagda, N. L. (1992) Distribution and use of cooking appliances that can affect
39 indoor air quality: topical report (November 1985~December 1991). Chicago, IL: Gas Research Institute;
40 report no. GRI-93/0013. Available from: NTIS, Springfield, VA; PB93-185452/XAB.
41 Korc, M. E.; Roberts, P. T.; Chinkin, L. R.; Main, H. H. (1993) Comparison of emission inventory and ambient
42 concentration ratios of NMOC, NOX, and CO in the Lake Michigan air quality region. Santa Rosa, CA:
43 Sonoma Technology, Inc.; report no. STI-90218-1357-FR.
44 Krol, M.; Van Leeuwen, P. J.; Lelieveld, J. (1998) Global OH trend inferred from methylchloroform measurements.
45 J. Geophys. Res. [Atmos.] 103: 10,697-10,711.
46 Lagus Applied Technology, Inc. (1995) Air change rates in non-residential buildings in California. Sacramento, CA:
47 California Energy Commission; contract no. 400-91-034; July.
48 Law, K. S. (1999) Theoretical studies of carbon monoxide distributions, budgets and trends. Chemosspere: in press.
49 Lawson, D. R.; Groblicki, P. J.; Stedman, D. H.; Bishop, G. A.; Guenther, P. L. (1990) Emissions from in-use
50 motor vehicles in Los Angeles: a pilot study of remote sensing and the inspection and maintainence program.
51 J. Air Waste Manage. Assoc. 40: 1096-1105.
52 Leaderer, B. P. (1982) Air pollutant emissions from kerosene space heaters. Science (Washington, DC)
53 218: 1113-1115.
54
October 1999 3-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 Leaderer, B. P.; Stolwijk, J. A. J.; Zagraniski, R. T.; Qing-Shan, M. (1984) A field study of indoor air contaminant
2 levels associated with unvented combustion sources. Presented at: 77th annual meeting of the Air Pollution
3 Control Association; June; San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association;
4 paper no. 84-33.3.
5 Levine, J. S.; Pinto, J. P. (1998) The production of CO by biomass burning. In: Khalil, M. A. K.; Pinto, J. P.;
6 Shearer, M. J., eds. Proceedings of the international conference on atmospheric carbon monoxide and its
7 environmental effects; December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection
8 Agency; report no. EPA/600/R-98/047; pp. 251-256.
9 Lobert, J. M.; Scharffe, D. H.; Hao, W.-M; Kuhlbusch, T. A.; Seuwen, R.; Warneck, P.; Crutzen, P. J. (1991)
10 Experimental evaluation of biomass burning emissions: nitrogen and carbon containing compounds.
11 In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge,
12 MA: MIT Press; pp. 289-304.
13 Lofroth, G.; Burton, R. M.; Forehand, L.; Hammond, S. K.; Seila, R. L.; Zweidinger, R. B.; Lewtas, J. (1989)
14 Characterization of environmental tobacco smoke. Environ. Sci. Technol. 23: 610-614.
15 Logan, J. A.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B. (1981) Tropospheric chemistry: a global perspective.
16 J. Geophys. Res. C: Oceans Atmos. 86: 7210-7254.
17 Magliano, K. L.; Lurmann, F. W.; Watson, J. G. (1993) Reconciliation of emissions inventory and ambient data in
18 the San Joaquin Valley and the San Francisco Bay area. Presented at: 86th annual meeting & exhibition of
19 the Air & Waste Management Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management
20 Association; paper no. 93-WP-100.8.
21 Mahieu, E,; Zander, R.; Delbouille, L.; Demoulin, P.; Roland, G.; Servais, C. (1997) Observed trends in total
22 vertical column abundances of atomspheric gases from IR solar spectra recorded at the jungfraujoch.
23 J. Atmos. Chem. 28: 227-243.
24 Marr, L. C.; Morrison, G. C.; Nazaroff, W. W.; Harley, R. A. (1998) Reducing the risk of accidental death due to
25 vehicle-related carbon monoxide poisoning. J. Air Waste Manage. Assoc. 48: 899-906.
26 McCarthy, S. M.; Yarmac, R. F.; Yocom, J. E. (1987) Indoor nitrogen dioxide exposure: the contribution from
27 unvented gas space heaters. In: Seifert, B.; Esdorn, H.; Fischer, M.; Rtiden, H.; Wegner, J., eds. Indoor air
28 '87: proceedings of the 4th international conference on indoor air quality and climate, v. 1, volatile organic
29 compounds, combustion gases, particles and fibres, microbiological agents; August; Berlin, Federal Republic
30 of Germany. Berlin, Federal Republic of Germany: Institute for Water, Soil and Air Hygiene; pp. 478-482.
31 Moschandreas, D. J.; Zabransky, J., Jr. (1982) Spatial variation of carbon monoxide and oxides of nitrogen
32 concentrations inside residences. Environ. Int. 8: 177-183.
33 Moschandreas, D. J.; Relwani, S. M.; O'Neill, H. J.; Cole, J. T.; Elkins, R. H.; Macriss, R. A. (1985)
34 Characterization of emission rates from indoor combustion sources. Chicago, IL: Gas Research Institute;
35 report no. GRI 85/0075. Available from: NTIS, Springfield, VA; PB86-103900.
36 Mount, G. H.; Williams, E. J. (1997) An overview of the tropospheric OH photochemistry experiment, Fritz
37 Peak/Idaho Hill, Colorado, fall 1993. J. Geophys. Res. [Atmos.] 102: 6171-6186.
3 8 Moxley, J. M.; Cape, J. N. (1997) Depletion of carbon monoxide from the nocturnal boundary layer. Atmos.
39 Environ. 31: 1147-1155.
40 Mueller Associates. (1985) Indoor air quality environmental information handbook: combustion sources.
41 Washington, DC: U.S. Department of Energy, Office of Environmental Analysis; report no.
42 DOE/EV/10450-1. Available from: NTIS; Springfield, VA; DE85006589.
43 Mumford, J. L.; Lewtas, J.; Burton, R. M.; Svendsgaard, D. B.; Houk, V. S.; Williams, R. W.; Walsh, D. B.;
44 Chuang, J. C. (1990) Unvented kerosene heater emissions in mobile homes: studies on indoor air particles,
45 semivolatile organics, carbon monoxide, and mutagenicity. In: Indoor air '90: precedings of the 5th
46 international conference on indoor air quality and climate, volume 2, characteristics of indoor air;
47 July-August; Toronto, ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air Quality
48 and Climate, Inc.; pp. 257-262.
49 Mumford, J. L.; Williams, R. W.; Walsh, D. B.; Burton, R. M.; Svendsgaard, D. J.; Chuang, J. C.; Houk, V S.;
50 Lewtas, J. (1991) Indoor air pollutants from unvented kerosene heater emissions in mobile homes: studies on
51 particles, semivolatile organics, carbon monoxide, and mutagenicity. Environ. Sci. Technol. 25: 1732-1738.
52 Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
53 parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
54
October 1999 3-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nagda, N. L.; Koontz, M. D.; Fortmann, R. C.; Billick, I. H. (1987) Detailed evaluation of range-exhaust fans.
2 In: Seifert, B.; Esdorn, H.; Fischer, M.; Rueden, H.; Wegner, J., eds. Indoor air '87: proceedings of the
3 4th international conference on indoor air quality and climate, v. 3, developing countries, guaranteeing
4 adequate indoor air quality, control measures, ventilation effectiveness, thermal climate and comfort, policy
5 and strategies; August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of Germany: Institute
6 for Water, Soil and Air Hygiene; pp. 324-328.
7 Nagda, N. L.; Koontz, M. D.; Billick, I. H.; Leslie, N. P.; Behrens, D. W. (1996) Causes and consequences of
8 backdrafting of vented gas appliances. J. Air Waste Manage. Assoc. 46: 838-846.
9 National Aeronautics and Space Administration, Panel for Data Evaluation. (1997) Chemical kinetics and
10 photochemical data for use in stratospheric modeling: evaluation number 12. Pasadena, CA: National
11 Aeronautics and Space Administration, Jet Propulsion Laboratory; JPL publication 97-4.
12 National Oceanic and Atmospheric Administration. (1999a) Global distribution of carbon monoxide [CO rug].
13 Boulder, CO: Climate Monitoring & Diagnostics Laboratory, Carbon Cycle Group. Available:
14 www.cmdl.noaa.gov/ccgg/figures/figures.html.
15 National Oceanic and Atmospheric Administration. (1999b) CO trends and growth rates. Boulder, CO: Climate
16 Monitoring & Diagnostic Laboratory, Carbon Cycle Group. Available:
17 www.cmdl.noaa.gov/ccgg/figures/figures.html.
18 National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health
19 effects. Washington, DC: National Academy Press.
20 Novelli, P. C.; Masarie, K. A.; Tans, P. P.; Lang, P. M. (1994) Recent changes in atmospheric carbon monoxide.
21 Science (Washington, DC) 263: 1587-1590.
22 Novelli, P. C.; Connors, V. S.; Reichle, H. G., Jr.; Anderson, B. E.; Brenninkmeijer, C. A. M.; Brunke, E. G.;
23 Doddridge, B. G.; Kirchhoff, V. W. J. H.; Lam, K. S.; Masarie, K. A.; Matsuo, T.; Parrish, D. D.; Scheel,
24 H. E.; Steele, L. P. (1998) An internally consistent set of globally distributed atmospheric carbon monoxide
25 mixing ratios developed using results from an intercomparison of measurements. J. Geophys. Res. [Atmos.]
26 103: 19,285-19,293.
27 Ott, W.; Langan, L.; Switzer, P. (1992) A time series model for cigarette smoking activity patterns: model validation
28 for carbon monoxide and respirable particles in a chamber and an automobile. In: Measuring, understanding,
29 and predicting exposures in the 21st century: proceedings of the conference; November 1991; Atlanta, GA.
30 J. Exposure Anal. Environ. Epidemiol. 2(suppl. 2): 175-200.
31 Ott, W.; Switzer, P.; Willits, N. (1994) Carbon monoxide exposures inside an automobile traveling on an urban
32 arterial highway. J. Air Waste Manage. Assoc. 44: 1010-1018.
33 Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
34 functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
35 June; San Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no. 95-WP84B.03.
36 Pacyna, J. M.; Graedel, T. E. (1995) Atmospheric emissions inventories: status and prospects. Annu. Rev. Energy
37 Environ. 20: 265-300.
38 Park, J.-H.; Spengler, J. D.; Yoon, D.-W.; Dumyahn, T.; Lee, K.; Ozkaynak, H. (1998) Measurement of air
39 exchange rate of stationary vehicles and estimation of in-vehicle exposure. J. Exposure Anal. Environ.
40 Epidemiol. 8: 65-78.
41 Peters, A.; Kotesovec, F.; Skorkovsky, J.; Brynda, J.; Heinrich, J. (1999) Acute effects of suspended particle
42 concentrations in the atmosphere on mortality - a study comparing northeast Bavaria and north Bohemia.
43 Final report. Bavaria, Czechoslovakia: Institute for Epidemiology; report no. GSF-EP S 1/99.
44 Pierce, T. E.; Dudek, M. P. (1996) Biogenic emission estimates for 1995. In: The emission inventory: key to
45 planning, permits, compliance, and reporting; proceedings of a specialty conference; September; New
46 Orleans, LA. Pittsburgh, PA: Air & Waste Management Association; pp. 309-326. (AWMA publication
47 VIP-65).
48 Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with other on-road
49 vehicle emission data. J. Air Waste Manage. Assoc. 40: 1495-1504.
50 Pierson, W. R.; Gertler, A. W.; Robinson, N. F.; Sagebiel, J. C.; Zielinska, B.; Bishop, G. A.; Stedman, D. H.;
51 Zweidinger, R. B.; Ray, W. D. (1996) Real-world automotive emissions—summary of studies in the Fort
52 McHenry and Tuscarora mountain tunnels. In: Parrish, D.; Trainer, M.; Rao. S. T.; Solomon, P. A., eds.
53 A&WMA international specialty conference on regional photochemical measurements and modeling, part 2;
54 November 1993; San Diego, CA. Atmos. Environ. 30: 2233-2256.
October 1999 3-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pinto, J. P.; Khalil, M. A. K. (1991) The stability of tropospheric OH during ice ages, inter-glacial epochs and
2 modern times. Tellus 43B: 347-352.
3 Pinto, J. P.; Rasmussen, R. A. (1998) Sources of CH4, CO, and H2 in Prague, Czech Republic. In: Khalil, M. A. K.;
4 Pinto, J. P.; Shearer, M. J., eds. Proceedings of the international conference on atmospheric carbon monoxide
5 and its environmental effects; December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental
6 Protection Agency; report no. EPA/600/R-98/047; pp. 313-319.
7 Pollack, A. K.; Dunker, A. M.; Fieber, J. K.; Heiken, J. G.; Cohen, J. P.; Shepard, S. B.; Schleyer, C. H.; Yarwood,
8 G. (1998) Revision of light-duty vehicle emission inventories using real-world measurements-Auto/Oil
9 Program, phase II. J. Air Waste Manage. Assoc. 48: 291-305.
10 Poppe, D.; Zimmermann, J.; Bauer, R.; Brauers, T.; Briining, D.; Callies, J.; Dorn, H.-P.; Hofzumahaus, A.; Johnen,
11 F.-J.; Khedim, A.; Koch, H.; Koppmann, R.; London, H.; Mtiller, K.-P.; Neuroth, R.; Plass-Dtilmer, C.; Platt,
12 U.; Rohrer, F.; Roth, E.-P.; Rudolph, J.; Schmidt, U.; Wallasch, M.; Ehhalt, D. H. (1994) Comparison of
13 measured OH concentrations with model calculations. J. Geophys. Res. [Atmos.] 99: 16,633-16,642.
14 Prinn, R.; Cunnold, D.; Simmonds, P.; Alyea, F.; Boldi, R.; Crawford, A.; Fraser, P.; Gutzler, D.; Hartley, D.;
15 Rosen, R.; Rasmussen, R. (1992) Global average concentration and trend for hydroxyl radicals deduced from
16 ALE/GAGE trichloroethane (methyl chloroform) data for 1978-1990. J. Geophys. Res. [Atmos.] 97:
17 2445-2461.
18 Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.; Alyea, F. N.; Cunnold, D. M. (1995) Atmospheric trends and
19 lifetime of CH3CC13 and global OH concentrations. Science 269: 187-192.
20 Research Triangle Institute. (1990) An investigation of infiltration and indoor air quality. Albany, NY: New York
21 State Energy Research and Development Authority; report no. NYERDA 90-11. Available from: NTIS,
22 Springfield, VA;PB91-119156.
23 Rickert, W. S.; Robinson, J. C.; Collishaw, N. (1984) Yields of tar, nicotine, and carbon monoxide in the sidestream
24 smoke from 15 brands of Canadian cigarettes. Am. J. Public Health 74: 228-231.
25 Rinsland, C. P.; Levine, J. S. (1985) Free tropospheric carbon monoxide concentrations in 1950 and 1951 deduced
26 from infrared total column amount measurements. Nature (London) 318: 250-254.
27 Ritter, J. A.; Barrick, J. D. W.; Watson, C. E.; Sachse, G. W.; Gregory, G. L.; Anderson, B. E.; Woerner, M. A.;
28 Collins, J. E., Jr. (1994) Airborne boundary layer flux measurements of trace species over Canadian boreal
29 forest and northern wetland regions. J. Geophys. Res. [Atmos.] 99: 1671-1685.
30 Rodes, C.; Sheldon, L.; Whitaker, D.; Clayton, A.; Fitzgerald, K.; Flanagan, J.; DiGenova, F.; Hering, S.; Frazier,
31 C. (1998) Measuring concentrations of selected air pollutants inside California vehicles [final report].
32 Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no. 95-339.
33 Rogers, J. D. (1990) Ultraviolet absorption cross sections and atmospheric photodissociation rate constants of
34 formaldehyde. J. Phys. Chem. 94: 4011-4015.
35 Seila, R. L.; Lonneman, W. A.; Meeks, S. A. (1989) Determination of C2 to C12 ambient air hydrocarbons in 39 U.S.
36 cities from 1984 through 1986. Research Triangle Park, NC: U.S. Environmental Protection Agency,
37 Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/3-89/058. Available from:
38 NTIS, Springfield, VA; PB89-214142/HSU.
39 Seiler, W.; Conrad, R. (1987) Contribution of tropical ecosystems to the global budget of trace gases especially
40 CH4, H2, CO and N2O. In: Dickinson, R. E., ed. The geophysiology of Amazonia: vegetation and climate
41 interactions. New York, NY: John Wiley; pp. 133-162. (Glantz, H. H., ed. Wiley series in climate and
42 biosphere).
43 Shadwick, D.; Glen, G.; Lakkadi, Y.; Lansari, A.; del Valle-Torres, M. (1997) Analysis of carbon monoxide for the
44 Denver, Colorado MSA. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
45 Research and Development, National Exposure Research Laboratory; contract no. 68-D5-0049; December.
46 Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998a) Analysis of carbon monoxide for the Los Angeles, California
47 CMSA. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and
48 Development, National Exposure Research Laboratory; contract no. 68-D5-0049; June.
49 Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998b) Analysis of carbon monoxide for the New York, New York
50 CMSA. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and
51 Development, National Exposure Research Laboratory; contract no. 68-D5-0049; June.
52 Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998c) Analysis of carbon monoxide for the Phoenix, Arizona MSA.
53 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and Development,
54 National Exposure Research Laboratory; contract no. 68-D5-0049; June.
October 1999 3-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sheldon, L.; Clayton, A.; Jones, B.; Keever, I; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992) Indoor
2 pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board, Research
3 Division; report no. ARB/R-92/486. Available from: NTIS, Springfield, VA; PB92-235837.
4 Slack, H. H.; Heumann, M. A. (1997) Use of unvented residential heating appliances-United States, 1988-1994.
5 Morb. Mortal. Wkly. Rep. 46: 1221-1224.
6 Spicer, C. W.; Billick, I. H. (1996) Persistence and fate of natural gas appliance emissions. In: Proceedings of the
7 1995 international gas research conference; November 1995; Cannes, France. Rockville, MD: Government
8 Institutes, Inc.; pp. 1752-1761.
9 Stedman, D. H.; Bishop, G. G.; Aldrete, P.; Slott, R. S. (1997) On-road evaluation of an automobile emission test
10 program. Environ. Sci. Technol. 31: 927-931.
11 Stedman, D. H.; Bishop, G. A.; Slott, R. S. (1998) Repair avoidance and evaluating inspection and maintenance
12 programs. Environ. Sci. Technol. 32: 1544-1545.
13 Stephens, R. D.; Cadle, S. H. (1991) Remote sensing measurements of carbon monoxide emissions from on-road
14 vehicles. J. Air Waste Manage. Assoc. 41: 39-46.
15 Sterling, T. D.; Sterling, E. (1979) Carbon monoxide levels in kitchens and homes with gas cookers. J. Air Pollut.
16 Control Assoc. 29: 238-241.
17 Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
18 concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
19 Sze, N. D. (1977) Anthropogenic CO emissions: implications for the atmospheric CO-OH-CH4 cycle. Science
20 (Washington, DC) 195: 673-675.
21 Thompson, A. M.; Cicerone, R. J. (1986) Possible perturbations to atmospheric CO, CH4, and OH. J. Geophys. Res.
22 [Atmos.] 91: 10,853-10,864.
23 Thompson, A. M.; Chappellaz, J. A.; Fung, I. Y.; Kucsera, T. L. (1993) The atmospheric CH4 increase since the last
24 glacial maximum. (2). Interactions with oxidants. Tellus 45B: 242-257.
25 Thrasher, W. H.; DeWerth, D. W. (1979) Evaluation of the pollutant emissions from gas-fired room heaters.
26 Cleveland, OH: American Gas Association, Cleveland Laboratories; research report no. 1515.
27 Tiegs, P.; Bighouse, R. (1994) Spillage of combustion byproducts from woodstoves operated in negative pressure
28 environments. Presented at: 87th annual meeting and exhibition of the Air & Waste Management
29 Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association;
30 paper no. 94-FA150.04.
31 Tobacco Information and Prevention Source. (1996) 1995 national household survey on drug abuse, tobacco related
32 statistics, SAMHSA, August 1996. Atlanta, GA: U.S. Centers for Disease Control and Prevention. Available:
33 www.cdc.gov/tobacco/samhsa.htm.
34 Traynor, G. W.; Anthon, D. W.; Hollowell, C. D. (1982) Technique for determining pollutant emissions from a
35 gas-fired range. Atmos. Environ. 16: 2979-2987.
36 Traynor, G. W.; Allen, J. R.; Apte, M. G.; Girman, J. R.; Hollowell, C. D. (1983) Pollutant emissions from portable
37 kerosene-fired space heaters. Environ. Sci. Technol. 17: 369-371.
38 Traynor, G. W.; Apte, M. G.; Carruthers, A. R.; Dillworth, J. F.; Grimsrud, D. T. (1984) Pollutant emission rates
39 from unvented infrared and convective gas-fired space heaters. Berkeley, CA: U.S. Department of Energy,
40 Lawrence Berkeley Laboratory; report no. LBL-18258. Available from: NTIS, Springfield, VA;
41 DE85 010647/XAB.
42 Traynor, G. W.; Girman, J. R.; Apte, M. G.; Dillworth, J. F.; White, P. D. (1985) Indoor air pollution due to
43 emissions from unvented gas-fired space heaters. J. Air Pollut. Control Assoc. 35: 231-237.
44 Traynor, G. W.; Apte, M. G.; Carruthers, A. R.; Dillworth, J. F.; Grimsrud, D. T.; Gundel, L. A. (1987) Indoor air
45 pollution due to emissions from wood-burning stoves. Environ. Sci. Technol. 21: 691-697.
46 Traynor, G. W.; Aceti, J. C.; Apte, M. G.; Smith, B. V.; Green, L. L.; Smith-Reiser, A.; Novak, K. M.; Moses,
47 D. O. (1989) Macromodel for assessing residential concentrations of combustion-generated pollutants: model
48 development and preliminary predictions for CO, NO2, and respirable suspended particles. Berkeley, CA:
49 U.S. Department of Energy, Lawrence Berkeley Laboratory; report no. LBL-25211. Available from: NTIS,
5 0 Springfield, VA; DE89013 3 96/XAB.
51 Turk, B. H.; Grimsrud, D. T.; Brown, J. T.; Geisling-Sobotka, K. L.; Harrison, J.; Prill, R. J. (1989) Commercial
52 building ventilation rates and particle concentrations. ASHRAE Trans. 95(part 1): 422-433.
53 Turner, S.; Cyr, L.; Gross, A. J. (1992) The measurement of environmental tobacco smoke in 585 office
54 environments. Environ. Int. 18: 19-28.
October 1999 3-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 U.S. Census Bureau. (1998) American housing survey: detailed tables from the 1993 AHS-N data chart. How do
2 you fuel your home? - Table 2-5. Washington, DC: U.S. Department of Commerce. Available online at:
3 www.census.gov/hhes/housing/ahs/tab2-5.html.
4 U.S. Census Bureau. (1999) USA Statistics in brief-law, education, communications, transportation, housing.
5 Washington, DC: U.S. Department of Commerce. Available: www.census.gov/statab/www/part2.html.
6 U.S. Department of Housing and Urban Development. (1987) Indoor air pollution: the role of heating and cooking.
7 Washington, DC: Office of the Assistant Secretary for Policy Development and Research.
8 U.S. Environmental Protection Agency. (1981) U.S. Environmental Protection Agency Intra-Agency Task Force on
9 Air Quality Indicators. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
10 report no. EPA-450/4-81-015. Available from: NTIS, Springfield, VA; PB81-177982.
11 U.S. Environmental Protection Agency. (1990) National air quality and emissions trends report, 1988. Research
12 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/450/4-90/002. Available
13 from: NTIS, Springfield, VA; PB90-200114/XAB.
14 U.S. Environmental Protection Agency. (1991a) User's guide to MOBILE4.1 (mobile source emission factor
15 model). Ann Arbor, MI: Office of Mobile Sources; report no. EPA-AA-TEB-91-01. Available from: NTIS,
16 Springfield, VA; PB91-228759.
17 U.S. Environmental Protection Agency. (1991b) Air quality criteria for carbon monoxide. Research Triangle Park,
18 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
19 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
20 U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
21 monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
22 Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available from:
23 NTIS, Springfield, VA; PB93-157717.
24 U.S. Environmental Protection Agency. (1993) MOBILE5a: highway vehicle emissions factor model (on diskette).
25 Ann Arbor, MI: Emission Planning and Strategies Division; report no. EPA/SW/MT-94/120. Available from:
26 NTIS, Springfield, VA; PB95-500179INC (DOS format), PB95-500187INC (Macintosh format), or
27 PB95-500138INC (mag tape, EBCDIC character set).
28 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
29 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00 laF-cF. 3v.
30 Available from: NTIS, Springfield, VA; PB96-168224.
31 U.S. Environmental Protection Agency. (1998) National air quality and emissions trends report, 1997. Research
32 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-98-016. Available:
33 www.epa.gov/oar/aqtrnd97/.
34 Wallace, L. A. (1983) Carbon monoxide in air and breath of employees in an underground office. J. Air Pollut.
35 Control Assoc. 33: 678-682.
36 Wheatley, A. D.; Sadhra, S.; Beach, J. R. (1997) Exposure to toxic gas and particle phase pollutants evolved during
37 deployment of airbags in vehicles. Indoor Environ. 6: 134-139.
38 Williams, R.; Walsh, D.; White, J.; Jackson, M.; Mumford, J. (1992) Effect on carbon monoxide levels in mobile
3 9 homes using unvented kerosene heaters for residential heating. Indoor Environ. 1: 272-278.
40 Wilson, A. L.; Colome, S. D.; Tian, Y. (1993) California residential indoor air quality study. Volume 1:
41 methodology and descriptive statistics. Appendices. Chicago, IL: Gas Research Institute; report no.
42 GRI-93/0224.2. Available from: NTIS, Springfield, VA; PB94-166055.
43 Wilson, A. L.; Colome, S. D.; Tian, Y.; Becker, E. W.; Baker, P. E.; Behrens, D. W.; Billick, I. H.; Garrison, C. A.
44 (1996) California residential air exchange rates and residence volumes. J. Exposure Anal. Environ.
45 Epidemiol. 6:311-326.
46 Xiong, T.; Khinkis, M. J.; Coppin, W. P. (1991) The development of an ultra-low-emission gas-fired combustor for
47 space heaters. Presented at: 14th World Energy Engineering and 1991 Environmental Engineering Congress;
48 October; Atlanta, GA. Chicago, IL: Institute of Gas Technology.
49 Yarwood, G.; Morris, R. (1998) Regional scale CO modeling. In: Khalil, M. A. K.; Pinto, J. P.; Shearer, M. J., eds.
50 Proceedings of the international conference on atmospheric carbon monoxide and its environmental effects;
51 December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection Agency; report no.
52 EPA/600/R-98/047; pp. 413-431.
53 Yung, Y. L.; Shia, C.; Herman, R. L. (1999) Is the biomass burning source of CO decreasing? Chemosphere:
54 in press.
October 1999 3-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 Yurganov, L. N.; Grechleo, E. I.; Dzhola, A. V. (1999) Zvenigorod carbon monoxide time series: 27 years of
2 measurements. Chemosphere: in press.
3 Zawacki, T. S.; Cole, J. T.; Huang, V. M. S.; Banasiuk, H.; Macriss, R. A. (1984) Efficiency and emissions
4 improvement of gas-fired space heaters. Task 2. Unvented space heater emission reduction [final report].
5 Chicago, IL: Gas Research Institute; report no. GRI-84/0021. Available from: NTIS, Springfield, VA;
6 PB84-237734.
7 Zawacki, T. S.; Cole, J. T.; Jasionowski, W. J.; Macriss, R. A. (1986) Measurement of emission rates from gas-fired
8 space heaters. Chicago, IL: Gas Research Institute; report no. GRI-86/0245. Available from: NTIS,
9 Springfield, VA; PB87-153995.
10 Zhang, Y.; Stedman, D. H.; Bishop, G. A.; Beaton, S. P.; Guenther, P. L. (1996) On-road evaluation of
11 inspection/maintenance effectiveness. Environ. Sci. Technol. 30: 1445-1450.
12 Ziskind, R. A.; Rogozen, M. B.; Carlin, T.; Drago, R. (1981) Carbon monoxide intrusion into sustained-use
13 vehicles. Environ. Int. 5: 109-123.
October 1999 3-89 DRAFT-DO NOT QUOTE OR CITE
-------
i 4. POPULATION EXPOSURE TO CARBON MONOXIDE
2
3
4 4.1 INTRODUCTION
5 National Ambient Air Quality Standards (NAAQS) have been set to protect public health
6 and welfare. The NAAQS for carbon monoxide (CO), which are not to be exceeded more than
7 once per year, are 9 ppm for an 8-h average and 35 ppm for a 1-h average. These standards
8 include a margin of safety to protect the population from adverse effects of CO exposure.
9 Accordingly, this chapter reviews studies of population exposure to CO concentrations from
10 different sources and explains why CO exposure studies are necessary and how they are done.
11 It also discusses how population exposures are estimated, describes typical levels and durations
12 of CO exposure in various microenvironments, and examines how CO exposures have changed
13 over time in the United States.
14 Because Americans spend substantial amounts of time indoors, it is important to determine
15 the total population exposure to CO from both indoor and outdoor CO sources. In this chapter,
16 "outdoor" concentrations are those measured in air that immediately surrounds an indoor
17 microenvironment. Because of air exchange, "outdoor" CO concentrations have a direct effect
18 on CO concentrations measured indoors. "Ambient" concentrations are those measured at
19 fixed-site, air quality monitoring stations that are used to determine compliance with the
20 NAAQS.
21 As discussed in Chapter 3, one cannot assume that "outdoor" and "ambient" concentrations
22 are similar because of the ubiquitous presence of local motor vehicle traffic emissions that do not
23 impact directly the monitoring station (e.g., downwind of it). However, one generally can
24 assume that indoor and "outdoor" CO concentrations are approximately the same, except for
25 situations when CO is emitted by indoor sources (e.g., gas appliances inside a home), or when
26 CO emissions from an immediate outdoor source directly contaminate indoor microenvironments
27 (e.g., when a vehicle's undiluted exhaust infiltrates the passenger cabin of that vehicle or a
28 following vehicle).
29 After inhalation, CO binds with hemoglobin (Hb) in the blood to form carboxyhemoglobin
30 (COHb). Besides endogenous CO production developed within the body from Hb catabolism,
October 1999 4-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 everyone is exposed to a global background level of CO in the ambient air on the order of
2 0.1 ppm (see Section 3.2). These combined sources constitute a reference or baseline exposure
3 as reflected in an endogenous COHb level on the order of 0.5% that varies individually, based on
4 physiological differences. These differences reflect variation in basal metabolisms and other
5 metabolic sources, as discussed in detail in Sections 5.3 and 5.4. This chapter discusses the
6 exposure of nonsmokers to CO. Smokers are excluded, because they represent a source of CO,
7 and because of their higher baseline levels of COHb and adaptive response to elevated COHb.
8 The study of population exposure is multidisciplinary, and the definition of personal
9 exposure has evolved over time (Ott, 1982; Duan, 1982; Lioy, 1990; U.S. Environmental
10 Protection Agency, 1992a; Last et al., 1995; Zartarian et al., 1997). A recent definition offered
11 by Zartarian et al. (1997) states that exposure is the contact between an agent and a target at a
12 specified contact boundary, defined as a surface in space containing at least one exposure point
13 (a point at which contact occurs). Using this definition, this chapter assumes that an inhaled CO
14 molecule (the agent) reaches a human (the target) at the lining of the lung (the contact boundary)
15 where CO exchange takes place between air and blood. In actual field studies, the air in the
16 immediate vicinity of the target is often assumed to be well mixed, such that a measured CO
17 concentration in the air can be assumed to represent a non-smoking person's actual exposure
18 from CO inhalation.
19 This chapter is concerned only with CO exposures that occur at concentrations capable of
20 increasing COHb levels above a reference baseline level. Besides exposure to CO concentrations
21 above the background level, human COHb levels can be elevated because of metabolic
22 degradation of many drugs, solvents (e.g., methylene chloride), and other compounds to CO.
23 For details see Section 5.3. Because the endogenous production of CO from drugs and solvents
24 may continue for several hours, it can prolong any cardiovascular stress from COHb. Moreover,
25 the maximum COHb level from an endogenous CO production can last up to twice as long a
26 comparable COHB levels caused by comparable exposures to exogenous CO (Wilcosky and
27 Simonsen, 1991; Agency for Toxic Substances and Disease Registry, 1993). Hence, the
28 literature on exposure to methylene chloride also is discussed in this chapter.
29 Descriptive studies of exposure typically report: average and peak concentrations to which
30 people are exposed, the temporal aspects of exposure (i.e., averaging times), where exposures
31 occurred (i.e., outdoor and indoor microenvironments), and the sources of microenvironmental
October 1999 4-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposures that add to ambient air intrusion (e.g., motor vehicles, gas appliances, etc.).
2 Explanatory studies try to identify factors that affect or contribute to exposure, because that
3 information may enable mitigation of high-level exposures. Unlike epidemiologic or clinical
4 studies of health effects, exposure studies rarely identify the health outcomes (e.g., headache,
5 dizziness, nausea, etc.) associated with measured exposures.
6 For completeness, this chapter also briefly discusses high-level nonambient CO exposures
7 that can lead to CO poisoning and death. Although such effects are not health outcomes
8 presently used in setting the NAAQS for CO, they may be affected by reduced emissions from
9 sources regulated by the U.S. Environmental Protection Agency (EPA) to meet the CO NAAQS
10 (e.g., accidental exposure to undiluted motor vehicle exhaust). Chapter 6 discusses relevant
11 health effects due to exposure to ambient CO concentrations.
12 This chapter is organized as follows: the first section summarizes the state of knowledge
13 on population exposure to CO as of 1991, when EPA published the last prior CO air quality
14 criteria document (AQCD). Next follows a discussion of more recently published studies of
15 population exposure to all sources of CO (except the active inhalation of tobacco smoke), which
16 delineates typical and peak levels of exposure as people engage in daily activities, including
17 those related to an occupation. Factors affecting trends in population exposure are then
18 described, including factors such as public policies affecting motor vehicle emissions, travel
19 behavior, and societal changes in human activity patterns, particularly those related to motor
20 vehicles (which are a major source of CO emissions as shown in Chapter 3). The conclusion
21 section summarizes findings of this assessment and discusses their implications for CO exposure
22 models, such as the probabilistic NAAQS Exposure Model (NEM) used for evaluating CO
23 exposures under different levels of the NAAQS. The conclusion also examines the extent to
24 which CO exposures have changed since the 1991 AQCD.
25
26
27 4.2 BRIEF SUMMARY OF POPULATION EXPOSURE STUDIES
28 PRIOR TO 1991
29 This section briefly reviews key population exposure studies that were completed by 1990.
30 It identifies populations sensitive to CO exposure and discusses studies of population exposure
31 based on fixed-site and personal monitors, as well as relevant population exposure models. This
October 1999 4-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 section does not discuss many pre-1990 exposure studies for two reasons: (1) the pre-1990
2 studies are reviewed in the 1991 CO AQCD, with the primary purpose of this edition of the
3 AQCD being to focus on more recent studies; and (2) the results of older studies may no longer
4 be indicative of current population exposures, given the major reductions in vehicle emissions
5 during the last 30 years, social changes affecting commuting patterns, and other factors. Factors
6 affecting changes in population exposure are discussed in Section 4.4.
7
8 4.2.1 Sensitive Populations
9 The NAAQS are intended to protect the general public, including probable high-risk groups
10 of the general population. These groups differ from one air pollutant to another. In the case of
11 CO, these groups include the elderly; pregnant women; fetuses; young infants; and those
12 suffering from anemia or certain other blood, cardiovascular, or respiratory diseases. People
13 currently thought to be at greatest risk from exposure to ambient CO levels are those with
14 ischemic heart disease who have stable exercise-induced angina pectoris (cardiac chest pain).
15 Individuals with this disease represented about 3% of the U.S. population in 1994. Studies show
16 that earlier time to onset of cardiac chest pain occurred in these people while they exercised
17 during exposures to CO concentrations that produced levels of COHb in the bloodstream in the
18 range of 2 to 3% (U.S. Environmental Protection Agency, 1991). The National Health and
19 Nutrition Examination Survey (NHANES) II study reported that 6.4% of the U.S. population
20 who never smoked had COHb levels above 2.1%, based on a national random sample of people
21 (n = 3,141) ranging in age from 12 to 74 years (Radford and Drizd, 1982). The NHANES II
22 study was done in the late 1970s when ambient CO concentrations were much higher (see
23 Figure 3-3).
24
25 4.2.2 Estimates of Population Exposure Based on Fixed-Site Monitors
26 In the United States, NAAQS attainment is based on ambient air quality measurements
27 recorded at a nationwide network of fixed-site monitors. Based on this network, EPA's Office of
28 Air Quality Planning and Standards estimated that 12.7 million people lived in seven counties
29 where ambient CO levels exceeded the NAAQS in 1996 and, thus, were at increased risk of
30 exposure to CO levels above the NAAQS (U.S. Environmental Protection Agency, 1998a). The
October 1999 4-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 estimate was made by combining census data on county populations with data on violations of
2 the CO NAAQS recorded by stationary monitors. Previous studies have shown why such
3 estimates should not be interpreted as assessments of population exposure to CO, for the two
4 reasons discussed below:
5 (1) Ambient CO concentrations are not spatially homogeneous within the area monitored.
6 For example, Ott and Eliassen (1973) reported average CO levels ranging from 5.2 to
7 14.2 ppm for sidewalks along congested streets of downtown San Jose, CA. Corresponding
8 CO averages at fixed-site monitors were only 2.4 to 6.2 ppm. A decade later, Ott and
9 Flachsbart (1982) found a narrower gap between simultaneous CO measurements from
10 fixed-site and personal exposure monitors deployed at indoor and outdoor commercial
11 settings in five California cities.
12 (2) In the absence of indoor CO sources and immediate outdoor sources (i.e., idling motor
13 vehicles), indoor CO concentrations tend to equal outdoor concentrations over the long term.
14 In buildings with mechanical ventilation systems, the timing and scheduling of outdoor
15 "make-up" air into the building affects ratios of indoor-outdoor concentrations both in the
16 short and long term (Yocom, 1982). For example, when make-up air was introduced into an
17 air-conditioned building during morning rush hours (when outdoor CO levels were high),
18 indoor CO concentrations exceeded outdoor levels for the remainder of the day (Yocom
19 et al., 1971). In the presence of indoor sources such as gas appliances, indoor CO
20 concentrations often exceed the outdoor levels (U.S. Environmental Protection Agency,
21 1991). Many American homes use gas (natural gas and liquid propane) for space heating,
22 cooking, heating water, and drying clothes. In a 1985 Texas study of a low-socioeconomic
23 population, CO concentrations were greater than or equal to 9 ppm in 12% of surveyed
24 homes. Residential CO concentrations were high where unvented gas space heaters were
25 used as the primary heat source (Koontz and Nagda, 1988).
26 These facts take on added significance given that many Americans spend most of their time
27 indoors (Szalai, 1972; Chapin, 1974; Meyer, 1983; Johnson, 1989; Schwab et al., 1990). Hence,
28 studies of actual personal exposure to CO are preferred over crude estimates of population
29 exposure to ambient CO, in determining what risk CO poses to public health from a total
30 exposure perspective (Sexton and Ryan, 1988).
31
October 1999 4-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.2.3 Surveys of Population Exposure Using Personal Monitors
2 With the development of personal exposure monitors (PEMs) in the 1970s, researchers
3 began to measure either the total human exposure of a population or the exposures of
4 subpopulations in microenvironments that posed higher risks of CO exposure, such as inside a
5 motor vehicle moving slowly in congested traffic. In theory, a microenvironment exists if the
6 CO concentration at a particular location and time is sufficiently homogeneous yet significantly
7 different from the concentrations at other locations (Duan, 1982).
8 Human exposure studies of target populations typically use either a direct or an indirect
9 approach. In the direct approach, PEMs are distributed either to a representative or
10 "convenience" (nonrandom) sample of a human population. Population exposure parameters
11 cannot be estimated from a convenience sample, because it does not represent the population
12 from which it was drawn. Using PEMs, people record exposures to selected air pollutants as
13 they engage in their regular daily activities. In the indirect approach, trained technicians use
14 PEMs to measure pollutant concentrations in specific microenvironments or populations. This
15 information then must be combined with additional data on human activity patterns to estimate
16 the time spent in those microenvironments (Duan, 1982; Sexton and Ryan, 1988). For further
17 discussion of these topics, see Section 8.2 of U.S. Environmental Protection Agency (1991) and
18 Mage (1991).
19 Sexton and Ryan (1988) discuss types of personal monitors and research methods used by
20 the direct and indirect approaches. Although small passive monitors may be placed near a
21 person's oral/nasal cavity where exposure actually occurs, larger monitors must be carried by a
22 person or placed nearby. Using data from PEMs, one can construct exposure-time profiles for a
23 particular activity such as commuting or the integrated exposure between two points in time.
24 From this information, one can determine the average concentration to which a person has been
25 exposed for a given time period. Based on the superposition principle, one also can determine a
26 net microenvironmental concentration by subtracting the outdoor concentration, as measured by
27 an appropriate fixed-site monitor, from a microenvironmental concentration measured by a
28 personal monitor (Duan, 1982). Because ambient CO concentrations are not spatially
29 homogeneous at any given moment, the net microenvironmental concentration can be either
30 positive or negative in value. A negative net value can occur, for example, in homes with no CO
31 sources during morning periods when ambient CO concentrations from rising traffic emissions
October 1999 4-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 on highways have not yet diffused into residential areas. A negative net value simply indicates
2 that the microenvironment has a lower positive CO concentration than the outdoor environment
3 at a given moment.
4 In an early pilot study in Los Angeles, using the direct approach, subjects recorded their
5 exposures and corresponding activities in diaries (Ziskind et al., 1982). Because this was
6 cumbersome and potentially distorted the activity, later studies used data loggers to store
7 concentrations electronically, as done by major studies of the urban populations of Denver, CO,
8 and Washington, DC (Akland et al., 1985). In these studies, subjects still used diaries to record
9 pertinent information about their activities in specified microenvironments while monitoring
10 personal exposures. Data were then transferred electronically from data loggers and manually
11 from diaries to computer files for analysis.
12 The direct approach, which uses the total exposure assessment methodology, provides a
13 frequency distribution of air pollutant concentrations for a sample of people, selected randomly
14 from either a general or a specific population (defined by demographic, occupational and health-
15 risk factors) for a particular time period of interest (e.g., a day). Studies using the direct approach
16 enable researchers to assess what percentage of a large population is exposed to pollutant
17 concentrations in excess of ambient air quality standards (Akland et al., 1985). Studies using the
18 indirect approach may focus on situations that bring large numbers of people in contact with high
19 concentrations in specific microenvironments. For example, Flachsbart and Brown (1989)
20 determined what percentage of employees were exposed to CO concentrations in excess of
21 national and state ambient air quality standards at a large shopping center attached to a parking
22 garage in Honolulu, HI.
23 Direct studies of general populations are rare because of their expense and the logistical
24 problems of monitor distribution. Two examples for CO were those done in Denver and
25 Washington during the winter of 1982 and 1983 (Akland et al., 1985). In both studies, the target
26 population included noninstitutionalized, nonsmoking residents, 18 to 70 years of age, who lived
27 in the city's metropolitan area, an estimated 1.2 million adults in Washington and 500,000 in
28 Denver. In both cities, the composite network of fixed-site monitors overestimated the 8-h
29 exposures of people with low-level personal exposures and underestimated the 8-h exposures of
30 people with high-level personal exposures. With respect to the underestimates, over 10% of the
31 daily maximum 8-h personal exposures in Denver exceeded the NAAQS of 9 ppm, and about 4%
October 1999 4-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 of maximum 8-h personal exposures in Washington exceeded 9 ppm. The end-expired breath
2 CO levels were in excess of 10 ppm, which is roughly equivalent to about 2% COHb in about
3 12.5% of the Denver participants and about 10% (after corrections were made for instrumental
4 measurement drift) of the Washington participants. Simultaneous CO measurements at fixed-site
5 monitors exceeded 9 ppm only 3% of the time in Denver and never exceeded 9 ppm in
6 Washington (Akland et al., 1985).
7 The Denver and Washington studies identified certain activities associated with higher CO
8 exposures. The two highest average CO concentrations occurred when subjects were inside a
9 parking garage and when traveling by car. Those who commuted 6 h or more per week had
10 higher average exposures than those who commuted less than 6 h per week. Table 4-1 shows
11 that higher mean CO concentrations occurred for travel by motor vehicle (motorcycle, bus, car,
12 and truck) than by walking and bicycle, and that high indoor concentrations above the 8-h
13 NAAQS of 9 ppm occurred in public garages, service stations or motor vehicle repair facilities.
14 Denver had higher average CO concentrations than Washington for all microenvironments
15 because of Denver's higher altitude and colder winter climate (Ott et al., 1992).
16
17 4.2.4 Population Exposure Models
18 Many studies developed computer models to predict exposure in both general and special
19 populations (U.S. Environmental Protection Agency, 1991). These models are important because
20 it is impossible and impractical to know the hourly and daily exposure of every person in a
21 population on a real-time basis. Models of human exposure are empirically derived
22 mathematical relationships, theoretical algorithms, or hybrids of these two. To support policy
23 decisions related to the setting of ambient and emission standards, EPA supported development
24 of two general population exposure models: (1) the NAAQS Exposure Model (NEM) and (2) the
25 Simulation of Human Activity and Pollutant Exposure (SHAPE) model. These models assume
26 that an individual's total CO exposure over a specified time interval can be estimated as the sum
27 of the average concentration within a microenvironment, multiplied by the amount of time spent
28 in that microenvironment (Duan, 1982).
29 The SHAPE model used a stochastic approach to simulate the exposure of an individual
30 over a 24-h period (Ott, 1984). The model replicates a person's daily activity pattern by
31 sampling from probability distributions representing the chance of entry, time of entry, and time
October 1999 4-8 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-1. CARBON MONOXIDE CONCENTRATIONS IN SELECTED
MICROENVIRONMENTS OF DENVER, CO, 1982 AND 1983
(Listed in descending order of mean CO concentration)
Microenvironment
In-Transit
Motorcycle
Bus
Car
Truck
Walking
Bicycling
Outdoor
Public garages
Residential garages or carports
Service stations or vehicle repair facilities
Parking lots
Other locations
School grounds
Residential grounds
Sports arenas, amphitheaters
Parks, golf courses
Indoor
Public garages
Service stations or vehicle repair facilities
Other locations
Other repair shops
Shopping malls
Residential garages
Restaurants
Offices
Auditoriums, sports arenas, concert halls
Number
22
76
3,632
405
619
9
29
22
12
61
126
16
74
29
21
116
125
427
55
58
66
524
2,287
100
Mean1
(ppm)
9.79
8.52
8.10
7.03
3.88
1.34
8.20
7.53
3.68
3.45
3.17
1.99
1.36
0.97
0.69
13.46
9.17
7.40
5.64
4.90
4.35
3.71
3.59
3.37
Standard Error
(ppm)
1.74
0.81
0.16
0.49
0.27
1.20
0.99
1.90
1.10
0.54
0.49
0.85
0.26
0.52
0.24
1.68
0.83
0.87
1.03
0.85
0.87
0.19
0.002
0.48
October 1999
4-9
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-1 (cont'd). CARBON MONOXIDE CONCENTRATIONS IN SELECTED
MICROENVIRONMENTS OF DENVER, CO, 1982 AND 1983
(Listed in descending order of mean CO concentration)
Microenvironment
Indoor (cont'd)
Stores
Health care facilities
Other public buildings
Manufacturing facilities
Homes
Schools
Churches
Number
734
351
115
42
21,543
426
179
Meana
(ppm)
3.23
2.22
2.15
2.04
2.04
1.64
1.56
Standard Error
(ppm)
0.21
0.23
0.30
0.39
0.02
0.13
0.25
a An observation was recorded whenever a person changed a microenvironment and on every hour; thus, each
observation had an averaging time of 60 min or less.
Sources: Johnson (1984) and Akland et al. (1985), as reported in U.S. Environmental Protection Agency (1991).
1 spent in 22 different microenvironments. Transition probabilities determine a person's
2 movement from one microenvironment to another. The model assumes that microenvironmental
3 concentrations reflect the contribution of an ambient concentration and a component representing
4 CO sources within each microenvironment. Because SHAPE relies on field surveys of
5 representative populations, the data requirements of the model are fairly extensive.
6 The SHAPE model can estimate the frequency distribution of maximum standardized
7 exposures to CO for an urban population and the cumulative frequency distribution of maximum
8 exposures for both 1-h and 8-h periods, thereby allowing estimates of the proportion of the
9 population that is exposed to CO concentrations above the NAAQS. An evaluation of SHAPE
10 by Ott et al. (1988), using survey data from the aforementioned Denver study, showed that the
11 observed and predicted arithmetic means of the 1-h and 8-h maximum average CO exposures
12 were in close agreement. However, SHAPE overpredicted low-level exposures and
13 underpredicted high-level exposures.
14
October 1999 4-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 Unlike SHAPE, which uses diary data from each person in a population, the NEM model
2 aggregates people into cohorts. The NEM model has evolved over time from deterministic to
3 probabilistic versions. As described elsewhere (Johnson and Paul, 1983; Paul and Johnson,
4 1985), the deterministic version of NEM simulates movements of selected groups (cohorts) of an
5 urban population through a set of exposure districts or neighborhoods and through different
6 microenvironments. Cohorts are identified by district of residence and, if applicable, district of
7 employment, as well as by age-occupation group and activity pattern subgroup. The NEM uses
8 empirical adjustment factors for indoor and in-transit microenvironments, and accumulates
9 exposure over 1 year. Although deterministic NEM was able to estimate central tendencies in
10 total exposure accurately, it did less well estimating the associated uncertainty caused by
11 variation in time spent in various microenvironments (Quackenboss et al., 1986) or to variation
12 in microenvironmental concentrations (Akland et al., 1985). Paul et al. (1988) discussed
13 advancements in the deterministic version of NEM.
14 In recent years, EPA developed the probabilistic NEM for CO (pNEM/CO); see Johnson
15 et al. (1992) for a description of the assumptions and algorithms of pNEM/CO, as those details
16 are beyond the scope of this chapter. Figure 4-1 shows the conceptual overview of the logic and
17 data flow of the pNEM/CO model. It shows how any alternative CO NAAQS can be evaluated
18 by establishing the distributions of personal exposures to CO when that alternative CO standard
19 is met. The inputs to the model (e.g., activity patterns, ambient monitoring data, air exchange
20 rates) are in the round-cornered boxes, and the model calculations are shown in the other boxes.
21 McCurdy (1995) examined the history of both the NEM and pNEM models and the role
22 they have played in reviews of criteria air pollutants such as CO. The EPA used pNEM/CO,
23 rather than the SHAPE model, in its previous review of the CO NAAQS (U.S. Environmental
24 Protection Agency, 1992b). At the request of the Clean Air Scientific Advisory Committee, as
25 part of that review, EPA performed a limited evaluation of the predictions of pNEM/CO against
26 observed data for subjects of the Denver, CO, study (Johnson et al., 1992). That evaluation
27 concluded that there was generally good agreement between the distributions of observed and
28 predicted 1-h daily maximum exposures, but that the model tended to underpredict the highest
29 8-h daily maximum exposures (i.e., >12 ppm) and overpredict the lowest 8-h daily maximum
30 exposures (i.e., <5 ppm). Extending the earlier evaluation of the pNEM/CO model, Law et al.
31 (1997) performed 20 simulated runs, the average values of which were used for evaluation
October 1999 4-11 DRAFT-DO NOT QUOTE OR CITE
-------
Air Exchange Rates and
Building Volumes
Emission Rates and
Use Patterns for
Indoor Sources
(e.g., gas appliances,
passive smoking)
Ambient Fixed-Site
Concentrations
Air Quality Specification
Seasonal Considerations
(Temperature)
Mass Balance Model
for Indoor
Microenvironments
Outdoor
Microenvironment
Concentrations
Human Activity and
Exertion Patterns
Population and
Commuting Data
Exposure Algorithms
Distribution of People and
Occurrences of Exposures
Linked with Breathing Rate
Figure 4-1. Conceptual overview of pNEM.
Source: Johnson etal. (1999).
1 purposes. Based on this evaluation, Law et al. (1997) reported the predicted and observed
2 population exposure cumulative frequency distributions (CFD), with and without gas stove use.
3 Figures 4-2 and 4-3 show that, regardless of gas stove use, pNEM/CO overpredicted the CFD at
4 low exposures and underpredicted the CFD at high exposures for 8-h daily maximum exposures.
5 The proportion of the Denver population exposed at or above 9 ppm (8-h NAAQS) was 12.7%
6 (observed) versus 9.9% (highest predicted) for those with gas stoves and 13.3% (observed)
7 versus 8.8% (highest predicted) for those without gas stoves. Similar results were reported for
8 the 1-h daily maximum exposures, with and without gas stove use. The proportion of the Denver
9 population exposed at or above 35 ppm (1-h NAAQS) was 3.2% (observed) versus 1.2% (highest
October 1999
4-12
DRAFT-DO NOT QUOTE OR CITE
-------
50'
40 •
30-
.—s 20'
E
Q.
a.
c
o
vs 10-
ro
o
c
o
O
CD
T3
c
o
.a
ro
O
5 -
1 -
0.5
Observed Exposure (n = 591)
Simulated Exposure
5 10
30 50 70
I I
90 95
99
99.9 99.99 99.999
Cumulative Probability (Percent)
Figure 4-2. Observed versus simulated 8-h daily maximum exposure for persons residing
in homes with gas stoves in Denver, CO.
Source: Law etal. (1997).
1
2
3
4
5
predicted) for those with gas stoves and 2.1% (observed) versus 1.1% (highest predicted) for
those without gas stoves. Relatively close agreement between simulated and observed PEM data
occurred for CO concentrations near the average exposure, within the range of 6 to 13 ppm for
the 1-h case and within 5.5 to 7.0 ppm for the 8-h case.
October 1999
4-13
DRAFT-DO NOT QUOTE OR CITE
-------
50 '
40 •
Q.
Q.
C
g
"75 10-
§
c
o
O 5
CD
.— 4 —
g
c ,
O 2
^
CO
O
1 —
0.5
Observed Exposure (n = 188)
Simulated Exposure
i i
99.9 99.99 99.999
iiiiiriiiii[
5 10 30 50 70 90 95 99
Cumulative Probability (Percent)
Figure 4-3. Observed versus simulated 8-h daily maximum exposure for persons residing
in homes without gas stoves in Denver.
Source: Law etal. (1997).
1 Law et al. (1997) proposed four factors that may explain why pNEM/CO underpredicted
2 the observed CFD of people with high-level exposures (>90th percentile).
3 (1) pNEM/CO modeled only two indoor CO sources (passive smoking and gas stoves), and
4 omitted other home combustion sources (e.g., attached garages, fireplaces, kerosene heaters,
5 wood stoves, etc.) that may have impacted observed levels using personal monitors.
October 1999
4-14
DRAFT-DO NOT QUOTE OR CITE
-------
1 (2) While people with high-level exposures were observed for two consecutive days, the model
2 randomly sampled each day separately thereby diluting the chances of sampling that person
3 on both days.
4 (3) The model used activity pattern data from three cities (Cincinnati, OH; Denver, CO; and
5 Washington, DC) to predict exposures of Denver residents. In reality, activity patterns of
6 Cincinnati or Washington residents may not reflect those residing in Denver.
7 (4) The model used a constant vehicular air-exchange rate and a constant secondary-smoke CO
8 value that are known to vary in reality.
9 Presently, EPA's Office of Air Quality Planning and Standards is in the process of revising
10 the pNEM/CO model.
11 Given this evaluation and EPA's mandate under the Clean Air Act to protect public health,
12 research is needed to improve the abilities of both SHAPE and pNEM to predict high-end
13 exposures. One possible improvement is to consider addressing autocorrelation of inputs for
14 time and concentration, particularly inputs of possibly nonindependent microenvironments. For
15 example, the commuting time and CO exposure of someone going from home to work may not
16 be independent of their commuting time and exposure from work to home. Also, both SHAPE
17 and pNEM may need to consider CO emission sources, microenvironments and activity patterns
18 excluded from previous versions of the models, in order to improve characterization of high-end
19 exposures. Model refinements should recognize that high-end exposures may come from either
20 regulated or unregulated CO emissions under the Clean Air Act.
21
22
23 4.3 SURVEY OF RECENT EXPOSURE STUDIES OF NONSMOKERS
24 This section focuses primarily on population exposure studies published or accepted for
25 publication in the scientific, peer-reviewed literature since the last CO AQCD was published in
26 1991, and includes some earlier studies that were omitted. It compares results from both old and
27 new studies wherever possible. However, the scope is limited to the exposure of nonsmokers.
28 Studies involving smokers were excluded because it is impossible to measure the concentration
29 and amount of CO inhaled by a smoker in a field study. As such, smokers have higher baseline
30 levels of COHb than nonsmokers and an adaptive response to elevated COHb.
October 1999 4-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 The section is subdivided into three parts. The first part discusses nonoccupational
2 exposure studies, because the NAAQS are intended to protect the general public. The second
3 part reviews occupational studies, because urban population exposure models (e.g., pNEM/CO)
4 must account for that portion of total exposure that occurs in occupational settings. The third
5 part describes recent activity pattern studies because they focus on the important temporal
6 component of exposure.
7
8 4.3.1 Nonoccupational Exposures
9 This subsection focuses on nonoccupational CO exposures that occur because of a variety
10 of human activities that require contact with sources of CO emissions such as motor vehicles and
11 fuel-burning tools and appliances. Section 3.5 discusses studies where CO has been measured by
12 area monitors in indoor microenvironments because such measurements constitute an indirect
13 exposure estimate. In addition, this section discusses studies of breath CO in populations and
14 studies of exposure to methylene chloride because it increases COHb levels.
15
16 4.3.1.1 Exposure to Carbon Monoxide from Motor Vehicles
17 This subsection presents nonoccupational studies of exposure to CO concentrations from
18 motor vehicles. Because these studies date to the mid-1960s, many of them were reviewed in the
19 1991 CO AQCD, and, therefore, they are not again reviewed here. Subsection 4.4.2.2 examines
20 the effects of progressively tighter CO emission standards on passenger cabin exposure. This
21 subsection focuses on other factors that affect the CO exposure of motorists and bicyclists.
22 Several studies done prior to 1991 reported passenger exposure to engine or tailpipe
23 emissions of CO (Amiro, 1969; Clements, 1978; Ziskind et al., 1981). More recently, Hampson
24 and Norkool (1992) reported that 20 of 68 children were treated for accidental CO poisoning
25 after riding in the back of eight pickup trucks in Seattle, WA, between 1986 and 1991.
26 Seventeen of the 20 children rode under a rigid closed canopy attached to the bed of the truck,
27 and the other three rode beneath a tarpulin. Six pickups had known exhaust system leaks; three
28 had rear-end tailpipes, and three had side-mounted tailpipes. For the 20 children, average COHb
29 levels measured in an emergency room were 18.2% (mean) (2.4%) (standard error), and levels
30 ranged from 1.6 to 37.0% COHb.
October 1999 4-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Studies done both before and after 1991 continue to show that fixed-site monitors
2 underestimate in-vehicle CO exposures. Flachsbart (1995) reported that 14 of 16 in-vehicle
3 exposure studies performed in the United States between 1965 and 1992 simultaneously
4 measured both ambient and passenger cabin concentrations. Regardless of the study, Table 4-2
5 shows that the mean CO concentrations inside vehicles always exceeded the mean ambient CO
6 concentrations measured at fixed-site monitors. The ratio between a study's mean in-vehicle CO
7 concentration and its mean ambient CO concentration fell between 2 and 5 for most studies,
8 regardless of when the study was done, but exceeded 5 for two studies done during the early
9 1980s. Of the more recent studies, Chan et al. (1991) found that median CO concentrations were
10 11 ppm inside test vehicles driven on hypothetical routes in Raleigh, NC during August and
11 September 1988, but median ambient concentrations were only 2.8 ppm at fixed-site monitors.
12 Fixed-site samples were collected about 30 to 100 m from the midpoint of each route.
13 Flachsbart (1995) proposed that the results of the 16 studies could be explained partly by
14 different study approaches (direct versus indirect) and by other aspects of study design, including
15 choice of city, season of the year, the surveyed road's functional type and location, travel mode,
16 and vehicular ventilation. For example, by pairing three direct studies with three indirect studies
17 (as shown in Table 4-2) that were done at the same time, Flachsbart showed that the mean
18 in-vehicle exposure measured by the direct approach was always lower than that measured by the
19 indirect approach. Although direct studies sampled real populations engaged in a variety of trips
20 in all types of traffic, most indirect studies focused on hypothetical commuters with higher
21 exposures in rush hour traffic. In another example, a comparison of Studies 8 through 10 in
22 Table 4-2 shows the effect of roadway type. Study 9 had a sizeable component of residential
23 driving, which may explain why the mean in-vehicle CO exposure of 7.7 ppm for Study 9 was
24 lower than the mean exposures for Studies 8 and 10. When Study 9 data were disaggregated by
25 roadway location, mean CO concentrations were 10 ppm for major commuting routes and
26 5.5 ppm for drives in residential areas. The mean concentration of 10 ppm for major commuting
27 routes in Study 9 is similar to the mean CO concentrations reported for arterial highways by other
28 studies (i.e., 9.8 ppm for Study 8 and 10.6 ppm for Study 10), which also were done during the
29 early 1980s.
30 Like earlier studies, recent ones also have looked at effects of different routes and travel
31 modes on CO exposure. Chan et al. (1991) reported significantly different in-vehicle exposures
October 1999 4-17 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-2. SUMMARY OF STUDIES OF IN-VEHICLE EXPOSURE AND
AMBIENT CARBON MONOXIDE CONCENTRATIONS, 1965 AND 1992
Study
Sites
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
Study
Period
= 1965
= 1966
April 1966 - June 1967
~ Summer 1973
October 1974 - February 1975
July - August 1978
October 1979
January 1980 -February 1981
January - March 1981
November 1981 -May 1982
November 1982 - February 1983
January - March 1983
May 1987 -March 1988
August - September 1988
January 1991 - March 1992
November 1991 - December 1992
Mean In- Vehicle
CO Concentration
(ppm)
37.0
31.8b
25.4 c
17.5
13.4
3.8
13.1
9.8
7.7 d
10.6
6.5 f
11.7
8.6
11.3
4.6
3.0
Variation of In- Vehicle
CO Concentration
(ppm)a
23-58
7-77
18-36
15-20
±5.5
±7.2
10.9-15.3
±5.8
±3.6 -±7.7
5-30
0.14-0.32B
±2.2 -±9.0
±4.95
±5.1
±2.1
±2.9
Typical Ambient
CO (ppm)
20-30
14.3
No data
6.0
6.0
3.5
3.4
<1.5
2.5-8.4 e
1.1
3.2-6.6 e
2.3
3.7
2.9
<1.0
No data
Study Sites:
1. Los Angeles, CA
2. Chicago, IL; Cincinnati, OH; Denver, CO; St.
Louis, MO; Washington, DC
3. 14 cities
4. Los Angeles, CA
5. Boston, MA
6. Washington, DC
7. Los Angeles, CA
8. Menlo Park, Palo Alto, and Los Altos, CA
9. Denver, CO; Los Angeles, CA; Phoenix, AZ;
Stamford, CT
10. Honolulu, HI
11. Denver, CO; Washington, DC
12. Washington, DC
13. Los Angeles, CA
14. Raleigh, NC
15. Menlo Park, Palo Alto, and Los Altos, CA
16. New Jersey Turnpike and Route 18, NJ
Notes (numbers shown in parentheses are CO concentrations in ppm):
"Range or one standard deviation except as noted.
"•Chicago (37); Cincinnati (21); Denver (40); St. Louis (36); Washington (25).
'Atlanta (29); Baltimore (21); Chicago (24); Cincinnati (23); Denver (29); Detroit (25); Houston (23); Los Angeles (36); Louisville (20);
Minneapolis-St. Paul (28); New York (34); Phoenix (29); St. Louis (18); Washington (19).
•"Denver (10.3); Los Angeles (8.5); Phoenix (6.7); Stamford (5.2); commuting and residential driving microenvironments weighted by sample
size.
'Range across all cities studied.
"Denver (8.0); Washington (5.0).
8Range in standard error for all cities studied.
Source: Flachsbart (1995).
October 1999
4-18
DRAFT-DO NOT QUOTE OR CITE
-------
1 to CO for standardized drives on three routes that varied in traffic volume and speed. The
2 median in-vehicle CO concentration was 13 ppm for 30 samples in the downtown area of
3 Raleigh, which had heavy traffic volumes, slow speeds, and frequent stops. The next highest
4 concentrations (median =11 ppm, n = 34) occurred on an interstate beltway that had moderate
5 traffic volumes and high speeds, and the lowest concentrations (median = 4 ppm, n = 6) occurred
6 on rural highways with low traffic volumes and moderate speeds. Similarly, Dor et al. (1995)
7 reported CO exposures of 12 ppm for 19 trips lasting an average of 82 min on a route through
8 central Paris, France, which was 2 to 3 ppm higher than the mean exposure for 30 trips split
9 between two suburban routes. In terms of travel modes, both Joumard (1991) and Dor et al.
10 (1995) found differences in CO exposures for public and private modes of travel in French cities
11 and towns, confirming findings made earlier in the United States by Flachsbart et al. (1987).
12 Ott et al. (1994) developed statistical models of passenger cabin exposure to CO
13 concentrations from highway emissions, based on 88 trips taken during a 13.5-mo period in 1980
14 and 1981. All trips occurred in one vehicle with windows set in a "standard position" as it
15 traveled an arterial highway (El Camino Real) in the San Francisco Bay Area of California. The
16 models are noteworthy because they examined the explanatory power of nine variables. The best
17 model predicted the average CO exposure per trip as a function of just two variables: (1) traffic
18 conditions as measured by the proportion of travel time stopped and (2) a seasonal trend term
19 expressed as a cosine function of the day of the year on which the trip was taken (adjusted
20 R2 = 0.67). A model that included ambient CO concentrations from a fixed-site monitor slightly
21 improved the explanatory power of the model (adjusted R2 = 0.71).
22 Flachsbart (1999a) developed a set of statistical models of passenger exposure to CO
23 concentrations inside a motor vehicle as it traveled a coastal artery (the Kalanianaole Highway)
24 in Honolulu, HI. All trips occurred during morning periods over a 6-mo period in 1981 and
25 1982. The 6.2-km study site was divided into three links. The models predicted the average CO
26 concentration inside the vehicle's passenger cabin on the third link as a function of several
27 variables. Based on data for 80 trips, the three most powerful models (adjusted R2 = 0.69) were
28 nonlinear combinations of several variables: the average CO concentration inside the cabin for
29 the second link; wind speed and direction; and either travel time, vehicle speed, or the estimated
30 motor vehicle CO emission factor for the third link. Several nonlinear models were based on
31 data for 62 trips for which nonzero, ambient CO concentrations were available. For this smaller
October 1999 4-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 database, the most practical models (adjusted R2 = 0.67) combined three variables: (1) the
2 ambient CO concentration; (2) the second-link travel time; and (3) either the travel time, vehicle
3 speed, or CO emission factor for the third link. The models showed that cabin exposure was
4 strongly affected by travel time and average vehicle speed, both of which were affected by the
5 time that the test vehicle entered each link of the highway. This implied that stochastic
6 simulations of exposure (e.g., the SHAPE model by Ott et al. [1988]) should not assume that trip
7 times and commuter exposures are independent of trip-starting times.
8 The most recent U.S. study of in-vehicle CO exposure reported results for hypothetical
9 commutes on standardized routes in Los Angeles and Sacramento, CA, during early fall, 1997
10 (Rodes et al., 1998). Continuous CO concentrations were measured over 2-h periods at different
11 times of day for 29 trips by two test vehicles (one vehicle following the other). The CO levels
12 were measured both inside and outside (at the base of the windshield) of the vehicles, and
13 outdoors along surveyed routes and at nearby fixed monitoring stations. Since the minimum
14 quantification limit (MQL) of the portable monitor was 2 ppm, all data below the MQL were
15 treated as zero concentrations. The research design employed a balanced factorial design to
16 determine the range of CO concentrations encountered in each city under different scenarios.
17 The scenarios represented combinations of different test vehicle types and ventilation settings,
18 roadway types, level of roadway congestion, and time of day. The study is unique in that the lead
19 test vehicle, mounted with a video camera, followed vehicles with dectectable emissions (by eye
20 or nose) whenever possible. Although the lead vehicle frequently trailed city buses and heavy
21 duty diesel trucks, it also targeted visibly gross-emitting, gasoline-powered vehicles.
22 Because the California study design purposely emphasized scenarios likely to result in high
23 in-vehicle emissions, the study results can not easily be compared to results of the 16 studies in
24 Table 4-2. Mean CO concentrations inside both test vehicles were reported for each scenario
25 based on only two to four commutes in each city. These concentrations ranged from less than
26 MQL to 2.6 ppm, based on a total of 13 trips in Sacramento and from 3.0 to 6.0 ppm, based on a
27 total of 16 trips in Los Angeles. As in similar studies, the means of ambient CO concentrations
28 measured at fixed-site monitors fell below the means of in-vehicle CO levels in both cities, and
29 typically were less than the MQL of the portable monitor. Because ambient CO levels vary from
30 city to city, the study computed microenvironmental concentrations of test vehicles by
31 subtracting ambient from interior CO levels to estimate CO exposure resulting solely from
October 1999 4-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 roadway emissions. They found average microenvironmental concentrations to be higher in
2 Los Angeles (4.6 to 4.9 ppm) than in Sacramento (2.1 to 3.1 ppm) during rush hour trips on
3 freeways; but, there was little difference in average microenvironmental concentrations between
4 freeway and arterial trips during rush hour trips in both cities. Based on preliminary analysis of
5 five trips, the study reported that CO concentrations could reach short-term peaks, ranging from
6 ~15 to 70 ppm, when the vehicle trailed gasoline-powered delivery trucks and older sedans.
7 The CO exposure of cycling as a travel mode has been studied and compared to the
8 exposure of motorists. In England, Bevan et al. (1991) reported that the mean CO exposure of
9 cyclists in Southhampton was 10.5 ppm, based on 16 runs over two 6-mi routes that took an
10 average of 35 min to complete. Note that the CO exposures of European cyclists may not be
11 comparable to cyclists' exposures in the United States because installation of catalytic converters
12 on new cars in Europe occurred in 1988, about 13 years after their introduction in the United
13 States (Faiz et al., 1996). In The Netherlands, Van Wijnen et al. (1995) compared exposures of
14 volunteers serving as both car drivers and cyclists on several routes in Amsterdam during winter
15 and spring. For a given route, the mean personal 1-h CO concentrations were always higher for
16 car drivers than for cyclists regardless of when sampling occurred during the year. However, a
17 volunteer breathed 2.3 times more air per minute on average as a cyclist than as a car driver.
18 When adjusted for variation in breathing rate, the range in median 1-h averaged uptakes of CO of
19 cyclists (2.4 to 3.2 mg) approached that of car drivers (2.7 to 3.4 mg).
20 Studies have quantified the effect of traffic volume and speed on in-vehicle CO exposure.
21 Flachsbart et al. (1987) reported that in-vehicle CO exposures fell by 35% when test vehicle
22 speeds increased from 10 to 60 mph on eight commuter routes in Washington, DC. In a similar
23 study of typical commuter routes in central Riyadh, Saudi Arabia, Koushki et al. (1992) found
24 that in-vehicle CO exposures fell by 36% when vehicle speeds increased from 14 to 55 km/h
25 (8.7 to 34.2 mph). They also found that mean in-vehicle CO concentrations increased by 71.5%
26 when traffic volumes increased from 1,000 to 5,000 vehicles per hour. Mean CO levels ranged
27 from 30 to 40 ppm averaged over trips of 25 to 43 min during peak hours, and ranged from 10 to
28 25 ppm for trips of 15 to 20 min during off-peak hours.
29 The effects of diurnal and seasonal variation on in-vehicle CO exposure were not discussed
30 completely in the 1991 CO AQCD. Studies of diurnal effects on in-vehicle exposure during peak
31 travel periods have been inconclusive because they did not control for covariation in traffic
October 1999 4-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 volumes and speeds, ambient CO concentrations, or meteorological conditions (e.g.,
2 temperatures, wind speeds) during different periods of the day. In Los Angeles, CA,
3 Haagen-Smit (1966) found evidence that CO exposures during afternoon commutes were greater
4 than those during morning commutes. Similar results were later found by Cortese and Spengler
5 (1976) in Boston, MA, by Wallace (1979) in Washington, DC, and by Dor et al. (1995) in Paris,
6 France. However, contrary evidence was reported by Holland (1983) for four U.S. cities, and by
7 Joumard (1991) in France. Recently, Rodes et al. (1998) compared commuter exposures for
8 morning and evening rush hour periods for 12 freeway trips in two California cities. Ambient
9 CO levels were subtracted from in-vehicle CO levels to estimate the vehicle's
10 microenvironmental concentration. In Los Angeles, microenvironmental concentrations during
11 evening commutes were about 25% lower than morning values because of higher wind speeds.
12 In Sacramento, such concentrations during evening trips were slightly higher than morning
13 values because of higher traffic congestion levels. In another recent study, Aim et al. (1999)
14 reported that the geometric mean CO concentration of 11 morning trips (3.1 ppm) exceeded that
15 of 12 afternoon trips (2.0 ppm) on a standard route in Kuopio, Finland, but attributed this
16 difference to weather and traffic variables.
17 Seasonal variation in ambient temperatures, wind conditions, and traffic volumes affect
18 passenger cabin exposure to CO, as shown in studies by Shikiya et al. (1989) in southern
19 California, Ott et al. (1994) in northern California, Dor et al. (1995) in France, and Flachsbart
20 (1999a) in Hawaii. Ott et al. (1994) and Dor et al. (1995), who both measured exposures for an
21 entire year, reported that they were generally higher in the fall and winter than in the spring and
22 summer. Such results usually are attributed to colder temperatures in temperate climates, which
23 increase CO emissions per vehicle mile during winter months. In Hawaii, where temperatures
24 are never cold enough to have a substantial effect on motor vehicle emissions, Flachsbart (1999a)
25 found that traffic flows and wind speeds had reinforcing effects on passenger cabin exposures to
26 CO concentrations on a coastal highway in Honolulu. During late fall, exposures were low
27 because traffic flows were light and wind speeds low. During winter and spring, exposures were
28 relatively higher because traffic flows were greater and winds calmer than during the fall.
29 In 1984, the U.S. Department of Transportation ordered a gradual phase-in of airbag
30 technology on passenger cars for safety reasons, and made dual airbags mandatory for all cars
31 beginning September 1, 1997. Based on deployment of four different airbags, Wheatley et al.
October 1999 4-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 (1997) reported that the time-weighted-average (TWA) CO concentration ranged from 174 to
2 370 ppm. Peak CO concentrations occurred 2 min after deployment.
3
4 4.3.1.2 Exposure to Carbon Monoxide in Recreational Vehicles
5 Two studies examined personal exposure to CO in the exhaust of recreational vehicles.
6 In the first study, Simeone (1991) collected CO concentrations in the passenger areas of large
7 power boats with side-mounted exhausts during routine cruises offshore of Boston and
8 Annapolis, MD. In Boston Harbor, CO concentrations averaged 56 ppm during a 60-min cruise
9 and 28 ppm after a 30-min cruise. For the Chesapeake Bay cruises near Annapolis, average
10 stabilized CO concentrations at the helm ranged from 93 to 170 ppm over 20- to 30-min periods
11 and 272 ppm over 30 min on the rear deck near the transom of the boat. In both studies, exhaust
12 gas was affected significantly by airflow about the boat under certain head winds. At head wind
13 speeds of 10 to 30 knots, turbulent mixing occurred in closer proximity to the rear of the boats,
14 enabling exhaust gases to migrate freely into each boat.
15 In the second study, Snook (1996) studied the CO exposure of a snowmobiler while
16 traveling in the wake of a lead snowmobiler on a 2- to 3-mi straight trail over level terrain in
17 Grand Teton National Park, WY. The CO exposure of the following snowmobiler was measured
18 under stable atmospheric conditions in Tedlar™ bags. The distance between the two
19 snowmobiles ranged from 25 to 125 ft and speeds ranged from 10 to 40 mph. The follower's
20 maximum average centerline exposure was 23.1 ppm, which occurred at 10 mph and 25 ft behind
21 the lead snowmobile. Although Snook (1996) reported no averaging times for exposures, one
22 can estimate that these times ranged from 3 to 18 min from the data given on the snowmobiler's
23 travel distance and vehicle speed. At distances greater than 25 ft, centerline exposures tended to
24 increase with greater speeds. At 15 ft off centerline, average concentrations fell sharply to levels
25 of 0 to 7.5 ppm. When the snowmobiler drove alone, the average concentration minus the
26 background concentration was 1.3 to 3.0 ppm. Background concentrations ranged from 0.2 to
27 0.5 ppm.
28 Snowmobile tourism has become a booming business across the nation and in several
29 national parks. Over 87,000 tourists traveled by snowmobile in Yellowstone National Park
30 (Wilkinson, 1995) during the winter of 1993 and 1994. Under steady-state conditions, a
31 snowmobile may emit from 10 to 20 g/mi of CO, while a modern U.S. automobile equipped with
October 1999 4-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 a catalyst emits far less (0.01 to 0.04 g/mi) at speeds of 10 to 40 mph. There are no federal laws
2 regulating the exhaust from snowmobile engines, and states are preempted from implementing
3 snowmobile emission standards. The typical snowmobile utilizes a two-stroke engine, because it
4 is less expensive than a four-stroke engine and provides a high powerweight ratio. However, a
5 two-stroke engine produces relatively high emissions of CO (Snook and Davis, 1997).
6
7 4.3.1.3 Residential Exposure to Carbon Monoxide
8 Residential sources of CO concentrations include motor vehicle operation inside a home
9 garage and the use of unventilated or poorly ventilated kerosene space heaters, gas appliances,
10 and charcoal grills and hibachis in the living area of the home. Studies of exposures to nonfatal
11 concentrations are discussed first, followed by studies of unintentional deaths caused by high
12 indoor concentrations.
13 According to the Barbecue Industry Association, 44 million American households owned a
14 charcoal grill in 1989, and an estimated 600 million charcoal-barbecuing events take place
15 annually (Hampson et al., 1994). An early study showed that the air stream from charcoal grills
16 contains 20 to 2,000 ppm of CO, with 75% of grills emitting 200 ppm and above (Yates, 1967).
17 Gasman et al. (1990) reported COHb levels ranging from 6.9 to 17.4% in a family of four people
18 in northern California who had been exposed to smoke from cooking indoors on a barbecue grill,
19 which was found by fire fighters in the middle of the living room.
20 Mumford et al. (1991) and Williams et al. (1992) assessed CO exposure to emissions from
21 unvented portable kerosene heaters in eight small mobile homes with no gas appliances and low
22 air exchange rates. Each home was monitored for an average of 6.5 h per day for 3 days per
23 week for 4 weeks. For 2 weeks, the heater was on, and, for 2 weeks, it was off. When the heater
24 was turned on, it was in use for an average of 4.5 h. When the heater was in use, study
25 participants (all nonsmokers) spent most of their time in the family room or kitchen. Sampling
26 took place in the living area about 1.5 to 3 m from the heater. The mean 8-h CO concentrations
27 were 7.4 ppm (1-h peak =11.5 ppm) when the heater was on and 1.4 ppm (1-h peak =1.5 ppm)
28 when it was off. Peaks usually were observed at the end of the combustion period. The ambient
29 CO level measured 0.5 h prior to heater use ranged from 0 to 8 ppm. When the heater was on,
30 three of the eight homes had 8-h average CO levels that exceeded the NAAQS, and one home
31 routinely had levels of 30 to 50 ppm.
October 1999 4-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 In joint studies, Wilson et al. (1993a,b) and Colome et al. (1994) reported CO exposures for
2 a random sample of California homes that used gas appliances during a 48-h period from
3 December 1991 to April 1992. For periods of 48 h, the median CO concentration was 1.2 ppm
4 (indoors) and 0.8 ppm (outdoors), and the median of the maximum 8-h average CO concentration
5 was 2.0 ppm (indoors) and 1.4 ppm (outdoors). Of surveyed homes, 13 of 286 homes (4.5%) had
6 indoor CO concentrations above the NAAQS of 9 ppm for 8 h, and 8 of 282 homes (2.8%) had
7 outdoor CO concentrations above this standard. Although most of the exceedances occurred in
8 the Los Angeles basin, these percentages could be low because the basin was under-represented
9 in the statewide sample. The study did not translate these percentages into statewide estimates.
10 Figures 4-4 and 4-5 show log-probability plots of the 1-h and 8-h maximum indoor minus
11 outdoor CO concentrations, respectively, for a common sample of 277 homes. These Figures
12 show that 17 homes (6.1%) had 1-h maximum concentrations indoors that were at least 5 ppm
13 higher than outdoor levels, and that 10 homes (3.6%) had 8-h maximum CO concentrations
14 indoors that were at least 5 ppm higher than outdoor levels. They suggest that a small percentage
15 of California homes would still have indoor CO problems even if outdoor CO levels at these
16 homes complied with federal ambient standards. Using univariate regression analysis, outdoor
17 CO concentrations explained approximately 55% of the variation found in indoor CO
18 concentrations. Higher net indoor CO levels (indoor minus outdoor CO concentrations) were
19 traced to several factors: space heating with gas ranges and gas-fired wall furnaces, use of gas
20 ranges with continuous gas pilot lights, small home volumes, and smoking cigarettes. The study
21 reported that several factors may have contributed to higher indoor net CO levels: malfunctioning
22 gas furnaces, automobile exhausts leaking into the home from attached garages and carports,
23 improper use of gas appliances (e.g., gas fireplaces), and improper installation of gas appliances
24 (e.g., forced air unit ducts).
25 In an area monitoring study, Kern et al. (1990) measured CO concentrations in a poorly
26 sealed, detached garage from operation of an emissions controlled (catalytic reactor and oxygen
27 sensor) and an emissions uncontrolled vehicle (carbureted and no catalytic reactor). Two tests
28 were conducted: Test 1 involved a poorly sealed garage door with a 3-in. opening, and Test 2
29 sealed the garage door with rags. The CO concentrations in the poorly sealed garage reached
30 4,700 ppm for the car without emission controls versus 2,000 ppm for the car with controls.
October 1999 4-25 DRAFT-DO NOT QUOTE OR CITE
-------
100.0'
Q.
a.
O
o 10.0
x
ro
O
1.0
0.1
.01 .1 .412 5 10 20 304050 6070 80 90 95 98 9999.6 99.9 99.99
Percentage
Figure 4-4. Log-probability plot of the maximum 1-h indoor minus outdoor CO
concentrations based on a random sample of 277 homes that used gas
appliances in California, 1991 and 1992. (Only those indoor minus outdoor
values greater than or equal to 0.1 ppm are shown.)
Source: Wilson et al. (1993a).
1 When the garage was better sealed, CO concentrations reached 8,400 ppm for the uncontrolled
2 vehicle versus 3,600 ppm for the controlled vehicle.
3 Unintentional deaths caused by CO poisonings have been studied in California,
4 New Mexico, and Washington. Two California studies collected data for the 1979 to 1988
5 period. In the first, Liu et al. (1993) reported that 13.3% of 444 deaths were caused by improper
6 use of charcoal grills and hibachies, of which 54% occurred inside motor vehicles (e.g., a van or
7 recreational camper) and 46% in residential structures (e.g., homes, apartments, shacks, tents).
8 Relative to their share of the state population, higher death rates occurred among Asians, blacks,
9 males, and people aged 20 to 39. In the second study, Girman et al. (1998) identified specific
October 1999
4-26
DRAFT-DO NOT QUOTE OR CITE
-------
100.0
Q.
o.
o
O 10.0-
00
E
X
03
O
1.0'
0.1-1
.01
_L
.412 5 10
20 30 40 50 60 70 80
Percentage
90 95 98 99 99.6 99.9 99.99
Figure 4-5. Log-probability plot of the maximum 8-h indoor minus outdoor CO
concentrations based on a random sample of 277 homes that used gas
appliances in California, 1991 and 1992. (Only those indoor minus outdoor
values greater than or equal to 0.1 ppm are shown.)
Source: Wilson et al. (1993a).
1 factors that contributed to unintentional deaths caused by CO from several combustion sources
2 (e.g., charcoal grills and hibachis, other heating and cooking appliances, motor vehicles, small
3 engines, and camping equipment). There was a strong association between alcohol use and CO
4 poisoning from motor vehicles. Typically, motorists under the influence of alcohol, would pull
5 into their garages, leave the engine running to listen to cassette tapes, and then fall asleep. Faulty
6 heating equipment used during winter months was implicated in about 50% of all unintentional
7 deaths in studies by both Girman et al. (1998) in California and Yoon et al. (1998) in New
8 Mexico.
9
October 1999
4-27
DRAFT-DO NOT QUOTE OR CITE
-------
1 Based on data for 10 counties in Washington, Hampson et al. (1994) reported features of
2 unintentional CO poisoning cases that occurred between 1982 and 1993. Most cases occurred
3 when electrical power was interrupted during fall and winter months, because of either regional
4 storms or unpaid utility bills. Of 509 patients treated with hyperbaric oxygen, 79 (16%) were
5 exposed when charcoal briquets were burned for heating or cooking in 32 separate incidents.
6 Non-English speaking Hispanic whites and Asians were disproportionately represented among
7 the cases. The COHb levels averaged 21.6% and ranged from 3.0 to 45.8%.
8 The National Center for Health Statistics (NCHS) and the U. S. Consumer Product Safety
9 Commission (CPSC) estimated that 212 deaths in 1992 were caused by fuel-burning appliances
10 used in the home. Of these deaths, 13 involved use of gasoline-powered appliances (National
11 Center for Health Statistics and U.S. Consumer Product Safety Commission, 1992). The CPSC
12 also estimated that 3,900 CO injury accidents occurred in 1994, of which about 400 were
13 associated with the use of gasoline-powered engines or tools (National Institute for Occupational
14 Safety and Health, 1996). In response to the problem, several federal government agencies
15 issued a joint alert concerning exposure to CO emitted by these sources (National Institute for
16 Occupational Safety and Health, 1996). These sources involved use of pressure washers, air
17 compressors, concrete-cutting saws, electric generators, floor buffers, power trowels, water
18 pumps, and welding equipment. Unintentional CO poisonings frequently happened indoors even
19 when people took precautions to ventilate the building.
20
21 4.3.1.4 Exposure to Carbon Monoxide at Commercial Facilities
22 Although motorists may turn off their engines during refueling, people still may be exposed
23 to CO concentrations from idling cars and other cars entering and leaving the fueling area of the
24 station. Wilson et al. (1991) randomly sampled 100 self-service filling stations and, for
25 comparison, took convenience samples at 10 parking garages and 10 nearby office buildings in
26 Los Angeles, Orange, Riverside, and San Bernardino counties of Southern California. They took
27 5-min samples of 13 motor vehicle air pollutants including CO in each microenvironment and in
28 the ambient environment. Microenvironmental and ambient concentrations were measured on
29 the same day but not simultaneously. The highest median CO concentration occurred in parking
30 garages (11.0 ppm), followed by service stations (4.3 ppm), and office buildings (4.0 ppm). The
31 median ambient CO concentration was 2.0 ppm.
October 1999 4-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ice skating, motocross, and tractor pulls are sporting events in which significant quantities
2 of CO may be emitted in short periods of time by machines in poorly ventilated indoor arenas.
3 The CO is emitted by several sources, including ice resurfacing machines and ice edgers during
4 skating events; motor vehicles at tractor-pull, monster-truck, and motocross competitions; and
5 gas-powered radiant heaters used to heat viewing stands. At these events, a "tractor" is a truck or
6 other vehicle modified to look like a farm tractor that is powered with aircraft turbines or
7 supercharged automobile engines. These competitions usually involve many motor vehicles with
8 no emission controls.
9 Several studies of CO exposure in commercial facilities were not cited in the 1991 CO
10 AQCD. First, Kwok (1981) reported episodes of CO poisoning among skaters inside four arenas
11 in Ontario, Canada. Mean CO levels ranged from 4 to 81 ppm for periods of about 80 min. The
12 CO levels in the spectator areas ranged from 90 to 100% of levels on the rink. The ice
13 resurfacing machines lacked catalytic emission controls. Second, both Sorensen (1986) and
14 Miller et al. (1989) reported CO concentrations greater than 100 ppm in ice rinks from the use of
15 gas resurfacing machines. High concentrations were attributed to poorly maintained machines
16 and to an ice rink without sufficient ventilation. Third, based on data collected in the Quebec
17 city area, Levesque et al. (1990) developed a linear relationship between CO exposure and the
18 CO concentration in exhaled breath (see Section 4.3.1.5 for discussion of CO exposure and
19 breath CO relationships) but could not eliminate other factors affecting the relationship. In a
20 later study, Levesque et al. (1991) measured the alveolar CO of 14 male adult nonsmokers who
21 played ice hockey, but who were not exposed in occupational settings. Rink CO concentrations
22 ranged from 0 to 76.2 ppm. The study again found a linear relationship between exposure and
23 absorbed CO such that, for each 10 ppm of CO in the indoor air, the players absorbed enough CO
24 to raise alveolar CO by 4.1 ppm or about 0.76% COHb.
25 In the United States, surveys of CO exposure were done at ice arenas in Vermont,
26 Massachusetts, Wisconsin, and Washington. For a rink in Massachusetts, Lee et al. (1993)
27 showed that excessive CO concentrations can occur even with well-maintained equipment and
28 fewer resurfacing operations if ventilation is inadequate. Average CO levels were less than
29 20 ppm over 14 h, with no significant source of outdoor CO. Ventilation systems could not
30 disperse pollutants emitted and trapped by temperature inversions and low air circulation at ice
31 level. In another study, Lee et al. (1994) reported that CO concentrations measured inside six
October 1999 4-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 enclosed rinks in the Boston area during a 2-h hockey game ranged from 4 to 117 ppm, whereas
2 outdoor levels were about 2 to 3 ppm, and the alveolar CO of hockey players increased by an
3 average of 0.53 ppm per 1 ppm CO exposure over 2 h. Fifteen years earlier, Spengler et al.
4 (1978) found CO levels ranging from 23 to 100 ppm in eight enclosed rinks in the Boston area,
5 which suggests that CO exposure levels in ice arenas have not improved.
6 In a letter, Paulozzi et al. (1991) reported that 25 people exposed to CO during a Vermont
7 high-school ice hockey game had mean COHb levels of 8.9%, but did not report if any of them
8 were smokers. Although Paulozzi et al. (1991) was unable to measure CO concentrations at the
9 game, Smith et al. (1992) reported CO levels of 150 ppm (no averaging time was given) at an
10 indoor ice-hockey rink in Wisconsin. To document the extent of the problem in Vermont,
11 Paulozzi et al. (1993) measured CO during eight high-school games in the state, and reported that
12 average CO levels for the entire game ranged from <5 to 101 ppm, with a mean of 35 ppm.
13 Hampson (1996) reported a maximum CO level of 354 ppm inside an ice arena in Seattle in
14 March 1996. Based on data for 17 persons whose tobacco use was not reported, the average
15 COHb level was 8.6% (range 3.3 to 13.9%). The source of CO was a malfunction in a
16 20-year-old ice resurfacing machine. Hampson also reported that CO may have diffused into an
17 adjacent bingo hall through an open door. In view of these studies, the State of Minnesota
18 declared in Regulation No. 4635 that CO measurements taken 20 min after ice resurfacing must
19 be less than 30 ppm.
20 Studies also have been done in sports arenas that allow motor vehicles. Boudreau et al.
21 (1994) reported CO levels for three indoor sporting events (i.e., monster-truck competitions and
22 tractor pulls) in Cincinnati, OH. The CO measurements were taken before and during each event
23 at different elevations in the public seating area of each arena with most readings obtained at the
24 midpoint elevation where most people were seated. Average CO concentrations over 1 to 2-h
25 ranged from 13 to 23 ppm before the event to 79 to 140 ppm during the event. Measured CO
26 levels were lower at higher seating levels. The ventilation system was operated maximally, and
27 ground-level entrances were completely open.
28 High CO levels also have been found at motor vehicle competitions in Canada. In a study
29 not cited in the 1991 CO AQCD, Luckurst and Solkoski (1990) recorded CO concentrations at
30 two tractor-pull events in Winnipeg, Manitoba. The mean of instantaneous concentrations at
31 25 locations in the arena ranged from 68 ppm at the start of the first event to 262 ppm by the end.
October 1999 4-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 At the second event, the range was 78 to 436 ppm. Levesque et al. (1997) reported CO levels at
2 an indoor motocross competition held in a skating rink in the Quebec city region. The May 1994
3 event lasted from roughly 8 p.m. to midnight. Average CO concentrations were determined at
4 five stations located at different points in the arena. The TWA concentrations ranged from
5 19.1 to 38.0 ppm, with higher levels during the second half of the show. High CO concentrations
6 forced a health official to interrupt the event seven times to help clear the air. Covariance
7 analysis showed that CO levels were related to the initial CO concentration, the event duration,
8 motor size, and especially the number of motorcycles on the track.
9
10 4.3.1.5 Studies of Breath Carbon Monoxide in Populations: The Effects of Exposure to
11 Carbon Monoxide
12 The concentration of CO in the end-tidal breath of a nonsmoker, after a standardized
13 breathholding maneuver, can be related to an exponentially time-weighted average of the
14 previous CO exposures (U.S. Environmental Protection Agency, 1991). As described in
15 Chapter 2, the breath concentration of CO can also be related to percentage of blood COHb by a
16 calibration curve constructed by simultaneous blood sampling and end-tidal breath sampling.
17 Because the CO air quality standards are intended to protect the nonsmoking public from
18 attaining a level of 2.1% COHb (U.S. Environmental Protection Agency, 1992b), these end-tidal
19 breath CO measurements can demonstrate whether nonsmoking subjects have been recently
20 protected from reaching this level. Such measurements are more informative than an 8-h
21 personal CO exposure measurement because a nominal 8-h average of 9 ppm can attain different
22 COHb concentrations. For example, a subject starting with 0.5% COHb will reach a higher
23 COHb level after a 4-h exposure to 3 ppm CO followed by a 4-h exposure to 15 ppm CO than if
24 exposed to 4-h at 15 ppm CO followed by 4-h at 3 ppm CO.
25 The EPA reviewed the pre-1990 literature reports of breath CO measurements in various
26 populations (U.S. Environmental Protection Agency, 1991; Section 8.5.2.2). These data and the
27 more recent data on breath CO in the following part of this section often are collected with
28 different breathhold-time, often are uncorrected for the CO content of the inhaled air (Smith,
29 1977; Wallace, 1983), and also may be subject to a positive hydrogen-interference if the breath
30 CO is analyzed electrochemically (Vreman et al., 1993) (see Section 2.6.2 of this document).
31 Consequently, this should be taken as a caveat by the reader that a portion of the variance among
October 1999 4-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 the results of different studies may be related to different breath collection methods and different
2 breath CO measurement techniques.
3 Lando et al. (1991) collected breath samples of 4,647 workers using MiniCO breath kits
4 (Model 1000, Catalyst Research Corporation, Owings Mills, MD). The latter part of a breath
5 was collected in a balloon following a 15-s breathhold but the method of analysis was not
6 described. Although the authors cite Jarvis et al. (1980) for this method, Jarvis et al. (1980) used
7 the Jones et al. (1958) method that requires a 20-s breathhold. Furthermore, these data are
8 uncorrected for the amount of CO in the maximal inhalation prior to the breathhold step (Smith,
9 1977; Wallace, 1983). Consequently, these data are not compatible with other studies using
10 20-s for the breathhold time and corrected data. Mean CO levels (Table 4-3) ranged from
11 4.2 (±1.66 standard deviation [SD]) ppm for never-smokers to 33.3 (±11.22 SD) ppm for heavy
12 smokers (25 cigarettes/day or more). Based on cutoffs of 3 and 6 ppm above ambient, a larger
13 number of ex-smokers (1.7 to 3.3%) than never-smokers (0.4 to 1.9%) appeared to be falsely
14 reporting their smoking status.
15 Chung et al. (1994) employed the Lee and Yanagisawa (1992, 1995) sampler to measure
16 personal exposure to CO of 15 Korean housewives using charcoal briquettes for cooking. The
17 COHb levels also were measured using a CO-Oximeter (CO-Ox). Although the personal
18 sampler had somewhat high imprecision based on four duplicate samples (average of 2.1 ppm
19 difference), the investigators were able to document a higher level of both exposure to CO and
20 blood COHb when the charcoal briquettes were used. Levels of COHb were generally high, even
21 without use of the briquettes, leading the experimenters to hypothesize that the high prevalence
22 of smoking (all 15 subjects had smokers in their homes) had elevated the level above the levels
23 found in the U.S. among nonsmokers.
24 Seufert and Kiser (1996) measured CO levels in the end-tidal breath after a 10-s breathhold
25 of 126 crew members of a nuclear submarine just before and just after a 62-h submerged period.
26 The CO level in the submarine (called "ambient" by the authors) increased from 2.6 ppm to
27 9.2 ppm in the fan room and in two other spaces. The authors state that the increase was caused
28 primarily by cigarette smoke from the 40 smokers aboard because auxiliary diesel engines were
29 not used during the submersion period. The nonsmokers' breath CO increased from 9 to 21 ppm.
30 Although the authors did not comment on the considerable difference between the nonsmokers'
31 breath CO of 21 ppm and the measured "ambient" concentration of only 9.2 ppm, it may have
October 1999 4-32 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-3. MEAN (M) BREATH CARBON MONOXIDE LEVELS AND SAMPLE
SIZES ACROSS SMOKING CATEGORIES AND JOB TYPES3
Smoking Category
Never-smokers
Quitters
Occasional smokers
Light smokers
(1 to 15 cigarettes/day)
Moderate smokers
(16 to 24 cigarettes/day)
Heavy smokers
(25 cigarettes/day or more)
M
SD
n
M
SD
n
M
SD
n
M
SD
n
M
SD
n
M
SD
n
Total
4.2
1.66
2,328
4.6
3.10
1,148
7.6
6.12
178
14.3
8.40
238
24.7
10.47
351
33.3
11.22
273
Blue Collar
4.5
1.89
294
5.1
5.03
217
7.6
3.98
22
15.6
8.94
48
24.6
11.72
97
32.6
9.61
95
Job Type
Clerical
4.1
1.69
958
4.4
2.19
427
8.0
7.05
90
14.0
8.70
131
25.4
10.06
180
34.1
12.73
117
White Collar
4.1
1.55
1,076
4.5
2.63
504
7.1
5.34
66
13.79
7.21
59
23.4
9.67
74
33.0
10.50
61
aSample size refers to those with CO measurements; CO measurements were taken on 97.2% of those interviewed.
The CO levels are in ppm. Data are for end-tidal breath collected after a 15-s breathhold without the required
correction for the CO in the inhaled air (Smith, 1977; Wallace, 1983)
Source: Lando etal. (1991).
1 been because of higher smoking rates in nonmonitored duty sections than in monitored sections,
2 the absence of a correction for higher CO concentrations in air inhaled for the 10-s breathhold
3 than in the end-tidal breath CO, and the end-tidal breath CO and "ambient" CO measurements
4 being made with two different instrumental systems. Operation of an atmospheric revitalization
5 system that removed CO also may have contributed to a lower monitored "ambient" CO than the
6 CO nonsmokers were exposed to in the nonmonitored duty sections.
7 Zayasu et al. (1997) present the first study showing that asthmatics untreated by
8 corticosteroids have higher 20-s breathhold end-tidal breath CO than either healthy controls or
9 treated asthmatics, as determined by subtracting the background level from the observed reading
October 1999
4-33
DRAFT-DO NOT QUOTE OR CITE
-------
1 (Figure 4-6). This is not the required correction for CO in the inhaled air reported by Smith
2 (1977) and Wallace (1983), so these data are not consistent with those studies where this
3 correction was made. They attribute the higher levels to lung inflammation, leading to a possible
4 increase in heme oxygenase, which creates endogenous CO. (For more information on
5 endogenous CO production, see Section 5.3.)
6 Shenoi et al. (1998) tested 470 youths (aged 5 to 20 years) in hospital admissions for CO in
7 breath, using the electrochemical Vitalograph Breath CO monitor (Vitalograph, Inc., Lenexa, KS).
8 The results, showing that 1.9% (9 of 470) had end-tidal breath CO levels greater than 9 ppm after
9 a 20-s breathhold, were confirmed by CO-Ox testing of blood. Five of the nine patients with the
10 higher breath CO were believed to be cigarette smokers, one may have been exposed to fumes
11 from a faulty furnace, and three were believed to be exposed to environmental tobacco smoke or
12 traffic exhaust. No corrections were made for the ppm of CO in the air inhaled for the
13 breathhold.
14
15 4.3.1.6 Nonoccupational Exposure to Methylene Chloride
16 Nonoccupational exposure to methylene chloride, which can be metabolized to CO in the
17 body, potentially occurs when the chemical is found in contaminated ambient air and
18 groundwater used as drinking water and in consumer products that contain the chemical as a
19 solvent, flame-retardant additive, or propellant. Exposure to methyl ene chloride in the home
20 primarily occurs through use of paint and varnish removers. Exposure also may occur through
21 use of aerosol propellants such as those found in hair sprays, antiperspirants, air fresheners, and
22 spray paints. The Agency for Toxic Substances and Disease Registry (1993) reported that some
23 aerosol products may contain up to 50% methylene chloride. However, the current extent of
24 methylene chloride in aerosol products apparently has not been studied recently; nor are typical
25 population exposures to methylene chloride from consumer products known.
26 Ambient exposure may occur near production and use facilities or near hazardous waste
27 sites that store methylene chloride. Ambient concentrations of methylene chloride near organic
28 solvent cleaning and paint and varnish removal operations range from 7.1 to 14.3 ppb, averaged
29 over 1 year (Systems Applications, Inc., 1983), and ambient levels at other locations were
30 reported by the U.S. Environmental Protection Agency (1985). Although methylene chloride
31 readily disperses when released into the air, it may remain in groundwater for years and be
October 1999 4-34 DRAFT-DO NOT QUOTE OR CITE
-------
15n
Q.
Q.
O
O
CD
.C
X
LU
10-
5-
0
O
O
O
O
O
00
O
O
O
Control
Untreated
Asthmatic
Treated
Asthmatic
Figure 4-6. Excess CO concentrations in the exhaled air of nonsmoking control subjects
(n = 30), untreated asthmatics (n = 30), and treated asthmatics (n = 30). The
values shown were determined by subtracting the background level from the
observed reading. "Untreated" means no inhaled corticosteroids, "Treated"
refers to regularly inhaled corticosteroids, and the horizontal bar indicates the
mean value.
Source: Zayasuetal. (1997).
October 1999
4-35
DRAFT-DO NOT QUOTE OR CITE
-------
1 ingested in drinking water or inhaled when it volatilizes during showering and laundering
2 (Agency for Toxic Substances and Disease Registry, 1993).
3 Exposure to about 500 ppm of methylene chloride for several hours can elevate COHb
4 levels to 15%. Increases in COHb levels can be detected in the blood of nonsmokers about
5 30 min after exposure to methylene chloride. Stewart et al. (1972) demonstrated that elevated
6 COHb levels were proportional to a series of controlled exposures to methylene chloride. In a
7 controlled experiment, Stewart and Hake (1976) observed postexposure levels of COHb ranging
8 from 5 to 10% after 3 h of use of a liquid-gel paint remover containing 80% methylene chloride
9 and 20% methanol by weight. Concurrent exposure to methylene chloride and methanol
10 prolongs the period of elevated COHb in the body (Stewart and Hake, 1976; Buie et al., 1986;
11 Wilcosky and Simonsen, 1991). Peterson (1978) reported COHb levels of up to about 10%
12 saturation after inhalation of methylene chloride concentrations ranging from 50 to 500 ppm over
13 5 days for 7.5 h per day.
14
15 4.3.2 Occupational Exposures
16 This subsection discusses occupational exposures to CO and methylene chloride.
17
18 4.3.2.1 Exposures to Carbon Monoxide in the Workplace
19 A survey by the National Institute for Occupational Safely and Health (NIOSH) found that
20 3.5 million workers in the private sector potentially are exposed to CO primarily from motor
21 exhaust. The number of persons potentially exposed to CO in the work environment is greater
22 than that for any other physical or chemical agent (Pedersen and Sieber, 1988). In 1992, there
23 were 900 work-related CO poisonings resulting in death or illness in private industry as reported
24 by the U.S. Bureau of Labor Statistics (as cited in National Institute for Occupational Safety and
25 Health, 1996). Three risk factors affect industrial occupational exposure: (1) the work
26 environment is located in a densely populated area that has high background (i.e., ambient) CO
27 concentrations; (2) the work environment produces CO as a product or by-product of an
28 industrial process, or the work environment tends to accumulate CO concentrations that may
29 result in occupational exposures; and (3) the work environment involves exposure to methylene
30 chloride, which is metabolized to CO in the body. Proximity to fuel combustion of all types
31 elevates CO exposure for certain occupations: airport employees; auto mechanics; small
October 1999 4-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 gasoline-powered tool operators (e.g., users of chainsaws); charcoal meat grillers; construction
2 workers; crane deck operators; firefighters; forklift operators; parking garage or gas station
3 attendants; policemen; taxi, bus, and truck drivers; toll booth and roadside workers; and
4 warehouse workers (U.S. Environmental Protection Agency, 1991).
5 Studies of firefighters are discussed briefly below because these studies were not discussed
6 in the 1991 CO AQCD. Other occupational studies of CO exposure are summarized in
7 Table 4-4, which shows CO concentrations for each study (typical values or ranges), averaging
8 periods, and the measured or estimated percent COHb levels for nonsmokers, if reported.
9 Lees (1995) reviewed studies of firefighter exposures to combustion products including
10 CO. During severe fires, firefighters were exposed to CO concentrations in excess of 500 ppm in
11 approximately 29% of 1,329 min sampled by Burgess et al. (1977) and in 48% of measurements
12 taken by Barnard and Weber (1979). Gold et al. (1978) reported a geometric mean concentration
13 of 110 ppm for a log-normal distribution of 65 samples with average duration of less than
14 10 min. The short-term exposure limit, designed to prevent acute effects of CO exposure, is
15 400 ppm averaged over 15 min. In three studies, the STEL was exceeded in 15 to 33% of
16 measurements (Treitman et al., 1980; Brandt-Rauf et al., 1988; Jankovic et al., 1991). Inside a
17 self-contained breathing apparatus, CO measurements ranged from 1 to 105 ppm in six samples
18 (Jankovic et al., 1991).
19 Firefighters are exposed to lower CO levels when they suppress bushfires, wildland fires,
20 and forest fires. For bushfires, Brotherhood et al. (1990) estimated that Australian firefighters
21 were exposed to CO levels averaging 17 ppm based on COHb measurements taken afterwards.
22 In a study of wildland fires in California, Materna et al. (1992) reported an average CO level of
23 14.4 ppm over a 3.5-h period (range 1.4 to 38 ppm) during fireline mop-up and a prescribed burn.
24 Concentrations were higher during evening hours, when inversions occurred, and could range up
25 to 300 ppm near gasoline-powered pumping engines. Materna et al. (1993) found comparable
26 results using different methods. For forest fires, the National Institute for Occupational Safety
27 and Health (1994) reported an average CO concentration of 11.5 ppm for a 9-h period.
28
29 4.3.2.2 Exposures to Methylene Chloride in the Workplace
30 Certain occupations expose workers to organic solvents such as methylene chloride. The
31 solvent is widely used as a degreaser, paint remover, aerosol propellant, and blowing agent for
October 1999 4-37 DRAFT-DO NOT QUOTE OR CITE
-------
o
o
t—t-
8-
VO
VO
VO
TABLE 4-4. STUDIES OF OCCUPATIONAL EXPOSURES AND DOSAGES3
J^.
I
oo
H
6
O
o
H
O
o
H
W
O
O
HH
H
W
Occupational Category
Airport workers
Bus drivers
Chainsaw/gas tool
operators
Charcoal meat grillers
Firefighters
Forklift operators and
workers in facilities
with forklifts
Garage mechanics
Manufacturing jobs
Traffic/roadway workers
CO Concentration (ppm)
and Averaging Period
5.0-13.6 (0.25 h)
5-300 (0.1-1.7 h) (INT)
5.8-12.5 TWA (0.5-1.0 h)
NA
>200 (<2 min)
16.2-24.3 TWA (8 h)
NA (3.0-5.6 h)
NA
14-20 (0.5-3. 0 + h)
0-105 ppm (IM) (0-1. 7 h)
14.4 TWA (3. 5 h)
1.2-24.2 (9 h)
250-300 (5 h)
NA (4.4 h)
370-386 (NA)
25-47 TWA (8-12 h)
3-34 (8 h)
42.6%>35(lh)
0-83 TWA (4 h)
2-7 (8 h)
1-4.3 (8 h)
5-42 (2 s)
Measured or Estimated
Percent COHb
NA
NA
NA
9.2-75.6 in 5 farmers
NA
>4 in 10 NS
5.7-7.0 in 56 NS
2.45 in 207 NS
3-7 in 9 NS
NA
NA
NA
5-22 for 4 NS
4.2-28.7 for 7 NS
21.1 ±0.7
6.3-13. 3 for 4 NS
NA
>5in45%ofNS
>3.5in71.4%ofNS
NA
NA
<5
State/Country
Massachusetts, U.S.
U.S.
France
U.S.
U.S.
Germany
Bahrain
Maryland, U.S.
Australia
U.S.
California, U.S.
California, U.S.
North Carolina, U.S.
North Carolina, U.S.
North Carolina, U.S.
Colorado, U.S.
California, U.S.
Ontario, Canada
Seven European countries
Four states, U.S.
Denmark
Massachusetts, U.S.
References
Bellin and Spengler (1980)
McCammon et al. (1981)
Limasset et al. (1993)
Kahleretal. (1993)
National Institute for Occupational
Safety and Health (1996)
Hunger etal. (1997)
Madanietal. (1992)
Radford and Levine (1976)
Brotherhood etal. (1990)
Jankovic et al. (1991)
Materna etal. (1992, 1993)
National Institute for Occupational
Safety and Health (1994)
Baucom etal. (1987)
Fawcettetal. (1992)
Ely etal. (1995)
McCammon et al. (1996)
Apte (1997)
Gourdeau et al. (1995)
Gardiner etal. (1992)
Boeniger(1995)
Raaschou-Nielsen et al. (1995)
Kamei and Yanagisawa (1997)
aNA = not available, INT = interior of vehicle, NS = nonsmokers, and IM = inside mask.
Source: Adapted from Apte (1997) and updated.
-------
1 polyurethane foams. It is used as an extractant for foods and spices, a grain fumigant, and a
2 low-pressure refrigerant. It also is used in the manufacturing of synthetic fibers, photographic
3 film, polycarbonate plastics, Pharmaceuticals, printed circuit boards, and inks. More than one
4 million workers have significant potential for exposure to methylene chloride (Agency for Toxic
5 Substances and Disease Registry, 1993). Moreover, the highest levels of exposure to methylene
6 chloride often occur in the workplace. To protect worker health, the 8-h TWA threshold limit
7 value for methylene chloride was set at 50 ppm by the American Conference of Governmental
8 Industrial Hygienists. Exposure at this concentration leads to COHb levels of about 1.9% in
9 experimental subjects. Exposure to 500 ppm for several hours may elevate COHb levels as high
10 as 15%. An 8-h exposure to about 500 mg/m3 (3.5 mg/m3 = 1 ppm) of methylene chloride vapor
11 is equivalent to an 8-h exposure to 35 ppm of CO (U.S. Environmental Protection Agency,
12 1985).
13 Methylene chloride stored in tissue may continue to metabolize to CO after several hours of
14 acute exposure. In such cases, COHb levels will continue to rise and peak as high as 25% about
15 5 to 6 h after exposure (Agency for Toxic Substances and Disease Registry, 1993). Shusterman
16 et al. (1990) reported an apparent linear elevation of COHb as a function of hours worked by a
17 furniture refinisher who used paint stripper containing methylene chloride. Ghittori et al. (1993)
18 reported a significant linear correlation (correlation coefficient [r] = 0.87) between methylene
19 chloride concentration in air and CO in alveolar air of nonsmoking and sedentary factory workers
20 in Italy. Exposure to 600 mg/m3 of methylene chloride for 7.5 h was associated with a COHb
21 level of 6.8% in eight volunteers. Exposure to methylene chloride also can be fatal. Leikin et al.
22 (1990) reported fatalities of two people who were exposed to unknown concentrations of
23 methylene chloride while they removed paint in an enclosed space. Death was caused not by CO
24 poisoning, but by solvent-induced narcosis. Before they died, their COHb levels continued to
25 rise following cessation of exposure despite treatment by high levels of oxygen.
26
27 4.3.3 Activity Pattern Studies
28 In assessing population exposure, studies of human activity patterns over a fixed time
29 period (e.g., 24 h) are necessary to determine how many people potentially are exposed to
30 sources of an air pollutant, and how long people spend in activities that involve use of these
31 sources. Accordingly, this section reviews studies of human activity patterns that pertain to
October 1999 4-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 population exposure to CO and methylene chloride. Previous studies reviewed in the 1991 CO
2 AQCD reported that many Americans spent most of their time indoors at home, school, or work,
3 etc. (Szalai, 1972; Chapin, 1974; Meyer, 1983; Johnson, 1989; Schwab et al., 1990). Although
4 more recent activity pattern studies largely confirm this finding, their sampling and questionnaire
5 designs provide new insights. This section reviews, in chronological order, newer studies that
6 include two surveys of activity patterns in California, a similar survey of preadolescents in six
7 states, a comparative study between California and the nation, a study in southwest England,
8 a Boston study, and a recent survey at the national level.
9
10 4.3.3.1 Activity Patterns of California Residents
11 The California Air Resources Board (CARB) conducted two surveys to determine the
12 activity patterns of California residents. In each case, projectable probability samples were
13 drawn from English-speaking households who had telephones. The first surveyed 1,762 adults
14 and adolescents over 11 years of age from fall 1987 through summer 1988 (Wiley et al., 1991a;
15 Jenkins et al., 1992), and the second surveyed 1,200 children under age 11 from April 1989
16 through February 1990 (Wiley et al., 1991b; Phillips et al., 1991). Using telephone interviews,
17 both surveys asked participants to complete a 24-h diary for the preceding day. People ages
18 9 and over responded directly to the interview, and the primary adult careprovider responded for
19 young children. The diaries enabled estimates of time spent in various activities and locations,
20 and determinations of whether respondents used or were near sources of pollutants, including
21 consumer products, combustion appliances, and motor vehicles.
22 Similar to previous studies, the results showed that all age groups spent most of their time
23 indoors. Adults and adolescents in California spent, on average, 87% of their time indoors (62%
24 at home and 25% elsewhere), and only 6% of their time outdoors. They also spent 7% of their
25 time in transit mostly in a car, van, or pickup truck. Compared to adults and adolescents,
26 children spent a similar amount of time indoors (86%), but more time at home (76%) and
27 outdoors (10%), and less time indoors elsewhere (10%) and in transit by car, van or pickup truck
28 (4%). About 46% of nonsmoking adults/adolescents reported being near others' tobacco smoke
29 at some time during the day.
30 Table 4-5 summarizes results of the two California surveys for various microenvironments
31 pertinent to CO exposure. For each microenvironment, the table shows the mean and range in
October 1999 4-40 DRAFT-DO NOT QUOTE OR CITE
-------
o
a
o
VO
VO
VO
TABLE 4-5. TIME SPENT IN DIFFERENT MICROENVIRONMENTS BY CALIFORNIANS, 1987 TO 1990
(minutes per day; weighted)51
Adults/Adolescents
Microenvironment
Motor Vehicle
Inside a garage
Inside an auto repair shop,
parking garage, or gas station
Inside a vehicle:
Car
Van or pickup truck
Bus
Potential Gas Appliance
Kitchen
Utility/laundry room
Basement
Industrial plant/factory
Restaurant
Bar/nightclub
Outdoor Transit
Walking
Bicycle/skates
Motorcycle, scooter
Bus/train/ride stop
Stroller/carried
Other
Population
Mean
9
11
73
18
4
74
3
35
28
8
10
1
1
NA
1
Percent
Doers
9.4
11.6
74.2
17.5
3.4
75.3
5.0
0.5
8.9
34.5
4.6
26.4
3.1
1.9
NA
1.3
Doer
Mean
97
91
99
102
114
98
53
79
393
81
174
38
41
62
NA
38
Doer
Range
1-845
1-685
1-585
2-785
5-1,320
1-930
3-380
5-180
4-750
1-885
5-825
1-360
5-160
5-430
NA
5-270
Population
Mean
2
43
13
3
47
<}
6
NA
6
1
<}
<1
Children
Percent
Doers
4.1
1
67.1
10.4
7.4
70.0
0.6
<0.2
12.7
NA
24.7
4.4
2.6
2.0
0.4
Doer
Mean
40
11
65
129
39
66
34
75
34
49
NA
24
22
15
40
72
Doer
Range
2-300
3-47
1-630
1-985
1-134
1-320
5-180
75-75
15-45
3-255
NA
1-195
2-155
3-40
1-195
10-110
NA = not available.
a Note: To generalize the results of the survey sample to the entire state population, the data were weighted to correct for nonuniform probabilities of including
certain individuals in the sample.
Source: Adapted from Jenkins et al. (1992); Phillips et al. (1991).
-------
1
2
3
4
5
9
10
time spent per day by both the entire sample and by those who actually did an activity in the
microenvironment (i.e., "doers"). The results show the disparity in mean time spent by the
population and by doers of an activity, which has implications for calculating population
exposure in risk assessment. Table 4-6 gives the percentage of each sampled population who
reported use of or proximity to potential sources of either CO or methylene chloride on a given
day. The study did not measure CO or methylene chloride concentrations from these sources in
microenvironments. Also, the surveys did not indicate whether respondents lived in a home
where combustion appliances were vented.
TABLE 4-6. PERCENTAGE OF CALIFORNIANS WHO USE OR WHO ARE
IN PROXIMITY TO POTENTIAL SOURCES OF EITHER CARBON MONOXIDE
OR METHYLENE CHLORIDE ON A GIVEN DAY, 1987 TO 1990 (WEIGHTED)3
Potential Pollutant Source
Consumer Products'"
Personal care aerosols
Scented room fresheners
Solvents
Oil-based paints
Activities/Places
Went to a gas station, parking garage, or
auto repair shop
Pumped gasoline
Have attached garage0
Had vehicle in attached garage0
Took a hot showerb
Near Combustion Appliances
Had gas heat on0
Had gas oven/range on
Environmental Tobacco Smoke (ETS)
Nonsmokers near ETS at any time during the day:
Adults (18 years or older)
Youths (12 to 17 years)
Adults and youths (12 years and older)
Youths (0 to 1 1 years)
Adults/Adolescents
1987-88
40
31°
12
5
26
15
62
37
77
26
35
43
64
46
Children
1989-90
36
37
3
2
11
1
63
36
26
24
29
38
aNote: To generalize the results of the survey sample to the entire state population, the data were weighted to
correct for nonuniform probabilities of including certain individuals in the sample.
bPotential methylene chloride exposure.
°Data presented for adult respondents (age 18 years or older) only.
Source: Adapted from Jenkins et al. (1992); Phillips et al. (1991).
October 1999
4-42
DRAFT-DO NOT QUOTE OR CITE
-------
1 4.3.3.2 Activity Patterns of Children in Six States
2 In 1990 and 91, Silvers et al. (1994) surveyed the activities of preadolescent children
3 (ages 5 to 12 years) from a projectable probability sample of 1,000 households in six states.
4 These states included three on the East Coast (New Jersey, New York, and Pennsylvania) and
5 three on the West Coast (California, Oregon, and Washington). Comparisons between this study
6 known as the Children's Activity Survey (CAS) and the CARB children's study are possible
7 because both were done over an entire year at about the same time, and both used a retrospective
8 time diary for a 24-h day. Both studies reported very similar results in terms of the mean hours
9 per day spent by preadolescent children for locations designated "indoors" (21.5 h for CARB
10 versus 21.7 h for CAS) and "at home" (18.0 h for CARB versus 17.8 h for CAS). For each
11 study, these results varied by ±1 h for different seasons of the year. There was variation in
12 specific activities (e.g., the CAS study reported that preadolescents spent less time per day
13 "riding in a vehicle" in California [0.52 h] than they did in the five other states [0.82 h], when the
14 five were combined as a group). The CAS study did not report time spent near other CO sources.
15
16 4.3.3.3 A Comparative Study Between California and the Nation
17 Robinson and Thomas (1991) compared results of activity pattern studies, one conducted
18 by CARB in California in 1987 and 1988 (Wiley et al., 1991a; Jenkins et al., 1992), and the other
19 done at the national level in 1985 (Cutler, 1990; Cornish et al., 1991). Although the two surveys
20 used different methods of gathering and coding data, the data were receded to enable
21 comparisons. The comparison showed that Californians averaged more time at work and in
22 commuting to work than was the case nationally, but averaged less time doing housework and
23 caring for children. California men also spent more time traveling. The national study appeared
24 to show greater time spent at home and in the yard, but, these results might be explained by
25 differences in location codes between the two studies, rather than by actual differences in
26 participant activity patterns. For example, the national study did not ask participants to identify
27 whether they worked indoors or outdoors. Because the national study was not designed for
28 exposure assessment, the authors proposed that the CARB study become a model for a future
29 national study oriented to exposure assessment. Such a study is discussed in Section 4.3.3.6.
30
October 1999 4-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.3.3.4 An English Study
2 Farrow et al. (1997) studied time spent inside the home from a sample of 170 households
3 in Avon, England from November 1990 through June 1993. A pregnant woman lived in each
4 household at the start of the study. Households completed a weekly diary for 1 year that covered
5 roughly the last 6 mo of the woman's pregnancy and the first 6 mo of the new infant's life. The
6 results indicated that the average amount of time spent inside the home per day varied by family
7 member as follows: mothers, 18.4 h (76.7%); fathers, 14.7 h (61.3%); and infants, 19.3 h
8 (80.4%). Infants spent more time at home during winter than summer. Although fathers spent
9 more time at home on weekends, mothers and infants spent less time. The applicability of the
10 study results for U.S. households was not determined, and it is hard to judge without comparative
11 information about the two countries. However, the study in England indicates that exposure may
12 be a function of a parent's gender or household role, supporting a similar conclusion based on a
13 nationwide study of U.S. activity patterns (see Section 4.3.3.6).
14
15 4.3.3.5 A Boston Study of Household Activities, Life Cycle, and Role Allocation
16 Using activity diary data from 150 households that participated in a 1991 Boston survey,
17 Vadarevu and Stopher (1996) tested several hypotheses about household travel. One study
18 hypothesis was that there are significant differences in mean time allocations of activities among
19 different "life-cycle groups" based on age, working status, and household size. They tested the
20 theory that life-cycle stage affects which activities fall into mandatory, flexible, and optional
21 categories; how much time can be allocated to different activities; and which household member
22 does each activity. They found that time allocated by households to specific activities varied
23 according to whether the household consisted of a single working adult, multiple adults, a young
24 family, an older family, or a nonworking adult. However, they found no significant differences
25 among the life-cycle groups or between any life-cycle group and the population mean in terms of
26 the total time spent in mandatory activities (work, work-related, school, and certain at-home
27 activities), which required on average 21 h per day. The amount of time spent in all flexible,
28 optional, and travel activities was about 3 h per day.
29
October 1999 4-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.3.3.6 The National Human Activity Pattern Survey
2 The U.S. Environmental Protection Agency's National Human Activity Pattern Survey
3 (NHAPS) collected 24-h diary data of activities and locations provided by 9,386 respondents
4 interviewed nationwide in the United States between October 1992 and September 1994 (Klepeis
5 et al., 1996). To enable projections to a larger population, the sample was weighted by the 1990
6 U.S. Census data to account for disproportionate sampling of certain population groups defined
7 by age and gender. Results were analyzed across a dozen subgroups: gender, age, race,
8 Hispanic, education, employment, census region, day-of-week, season, asthma, angina, and
9 bronchitis/emphysema. The weighted results showed that, on average, 86.9% of a person's day
10 was spent indoors (68.7% at residential locations), 7.2% of the day was spent in or near vehicles,
11 and 5.9% of the day was spent in outdoor locations.
12 The study also reported unweighted descriptive statistics and percentiles for both the full
13 population and various subpopulations (i.e., people who actually did certain activities or who
14 spent time in certain microenvironments) (Tsang and Klepeis, 1996). Of all 9,386 respondents,
15 38.3% reported having a gas range or oven at home, and another 23.7% said that the range/oven
16 had a burning pilot light. In terms of motor vehicle use, 10% of 6,560 people (7.0% of total
17 sample) spent more than 175 min per day inside a car, and 10% of 1,172 people (1.2% of total
18 sample) spent more than 180 min inside a truck, pickup, or van. Of those who were inside a
19 car and knew they had angina (n = 154 respondents), 10% of them spent more than 162 min
20 per day inside a car. The survey also asked about sources of household pollutants.
21 Of 4,723 respondents, 10.5% were exposed to solvents, 10.4% to open flames, and 8.4% to
22 "gas-diesel" powered equipment; 6.3% of these respondents were in a garage or indoor parking
23 lot; and 5.7% reported that someone smoked cigarettes at home. Only 1.8% of 4,663 respondents
24 reported having a kerosene space heater at home.
25
26
27 4.4 MAJOR FACTORS AFFECTING POPULATION EXPOSURE
28 This section discusses major factors that have and may continue to affect population
29 exposure to CO. These factors include public policies affecting urban transportation planning
30 and air quality, motor vehicle emissions, and social and technological changes affecting human
31 activity patterns.
October 1999 4-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.4.1 Federal Policies Affecting Transportation and Air Quality in
2 Urban Areas
3 In the United States, the national effort to improve air quality can be traced to the Clean Air
4 Act (CAA) amendments of 1970, 1977, and 1990. As discussed in Chapter 3, the effect of these
5 CAA amendments on ambient CO concentrations has been substantial. Moreover, emissions
6 from on-road vehicles have declined since 1970, even as other socioeconomic indicators of
7 growth have increased. Between 1970 and 1995, nationwide emissions of CO from on-road
8 vehicles fell 33.4% (U.S. Environmental Protection Agency, 1996), despite compound annual
9 growth rates of 1.0% in the nation's population and 3.2% in vehicle miles of travel (VMT)
10 during the same period (U.S. Department of Transportation, 1996). The faster growth rate of
11 VMT can be attributed to many factors that have decentralized housing and jobs within urban
12 regions since World War II.
13 Since the mid-1960s, major construction projects intended to expand highway capacities
14 have been opposed in some metropolitan areas. Opponents claimed that these projects promoted
15 urban sprawl and induced motor vehicle travel that raised regional air pollutant emissions.
16 To address these concerns, the 1990 CAA amendments state that transportation actions (plans,
17 programs, and projects) cannot create new NAAQS violations, increase the frequency or severity
18 of existing NAAQS violations, nor delay attainment of the NAAQS (U.S. Code, 1990). Pursuant
19 thereto, the U.S. Environmental Protection Agency promulgated its Transportation Conformity
20 Rule. Complementary provisions of the 1991 Intermodal Surface Transportation Efficiency Act
21 offered financial incentives under the Congestion Management and Air Quality (CMAQ)
22 improvement program. Under CMAQ, metropolitan planning organizations were offered federal
23 funds to improve air quality by implementing transportation control measures (TCMs).
24 Examples of TCMs include programs to promote car and van pooling, flextime, special lanes for
25 high occupancy vehicles, and parking restrictions.
26 Austin et al. (1994) examined how TCMs have changed travel activity, including number
27 of trips, vehicle miles of travel, vehicle speed, travel time, and the extent to which commuters
28 have shifted travel from peak to off-peak periods. Using an emission factors model (i.e.,
29 MOBILES [for a description of MOBILES, see U.S. Environmental Protection Agency, 1998b]),
30 the study inferred how much TCMs would change average speeds of motor vehicles and CO
31 emissions therefrom. The direct effect of TCMs on commuter exposure to CO has received only
October 1999 4-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 limited study. Flachsbart (1989) found that priority (with-flow and contra-flow) lanes were
2 effective in reducing exposure to motor vehicle exhaust on a coastal artery in Honolulu.
3 Compared to commuter CO exposure in adjacent but congested lanes, exposure in priority lanes
4 was about 18% less for those in carpools, 28% less for those in high-occupancy vehicles (e.g.,
5 vanpools), and 61% less for those in express buses. These differences occurred possibly because
6 commuters in priority lanes traveled faster than those in the congested lanes. Faster vehicles
7 created more air turbulence, which may have helped to disperse pollutants surrounding vehicles
8 in priority lanes. Furthermore, these differences existed even though the priority lanes were often
9 downwind of the congested lanes. Although higher speeds were related to lower exposures in
10 priority lanes, differences in exposure also could have been caused by differences in vehicle type
11 and ventilation, both of which were not controlled.
12 More recently, Rodes et al. (1998) compared the CO concentrations of two test vehicles
13 driven on standardized routes that included a freeway carpool lane in Los Angeles. One vehicle
14 used the carpool lane and the other the unrestricted lanes. Each vehicle repeatedly drove its route
15 for both the morning and evening rush-hour periods of one day, and CO measurements were
16 taken continuously both inside and outside of each vehicle. Because the vehicles had different
17 air exchange rates, comparisons of external CO levels measured at the base of the windshield are
18 appropriate. Based on these measurements, the average CO concentration of the vehicle in the
19 noncarpool lane (5.6 ppm) was twice as high as that of the vehicle in the carpool lane (2.8 ppm).
20 For a hypothetical 30-mi commute, exposure in the noncarpool lane (measured in parts per
21 million-minutes) was estimated to be 187% greater than exposure in the carpool lane. The study
22 suggests that carpoolers may have lower total CO exposure for their entire commute, because
23 they are exposed to lower CO concentrations and spend less time commuting in heavy traffic.
24 However, the study did not account for the extra time to collect nonfamily members of a carpool.
25 Models to estimate the direct effects of TCMs on commuter CO exposure are not apparent
26 in the literature. However, Flachsbart (1999a) developed a series of statistical models to predict
27 passenger cabin exposure to CO based on trip variables for a 3.85-mi, Honolulu artery divided
28 into three links. Based on data for 80 trips, the most practical models of third-link exposure
29 (adjusted correlation coefficient [R2] = 0.69) combined three variables: (1) the ambient CO
30 concentration; (2) the second-link travel time; and (3) either the travel time, vehicle speed, or CO
31 emission factor for the third link. The models showed that the vehicle's travel time and average
October 1999 4-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 speed and the CO emission factor for a given link of the roadway had equal ability to predict
2 passenger cabin exposure to CO on the third link because of mathematical relationships among
3 these three predictor variables.
4
5 4.4.2 Federal and State Policies Affecting Temporal Trends in Exposure
6 Studies show significant temporal trends in population exposure to CO concentrations from
7 motor vehicle emissions based on different indicators. One indicator is unintentional death rates
8 from CO poisoning, and another is based on direct measurements of passenger cabin exposure to
9 CO concentrations from traffic emissions. Table 4-7 summarizes data on these indicators from
10 several U.S. studies and shows the federal and California tailpipe CO emission standards by
11 model year for comparison. In Table 4-7, the net mean CO concentration value represents the
12 microenvironmental component of total exposure. This value equals the mean in-vehicle CO
13 concentration, minus the mean ambient CO concentration as recorded simultaneously at a
14 fixed-site monitor.
15
16 4.4.2.1 Effects of Motor Vehicle Emission Standards on Unintentional Death Rates
17 Based on death certificate reports compiled by the National Center for Health Statistics,
18 Cobb and Etzel (1991) reported statistics on the annual rate of unintentional deaths from CO
19 poisoning in the United States. As shown in Table 4-7, the annual death rate per 100,000
20 population declined from 0.67 in 1979 to 0.39 in 1988. Motor vehicle exhaust gas accounted for
21 6,552 deaths or 56.7% of the total 11,547 unintentional deaths occurring during the 10-year
22 period. The highest death rates per 100,000 persons occurred among males, blacks, the elderly,
23 and residents of northern states. Monthly variation in death rates indicated a seasonal pattern,
24 with January fatalities routinely about two to five times higher than in July.
25 Although mortality is not a health effect used in setting the NAAQS for CO, the Cobb and
26 Etzel (1991) study still has value in its revelations about cofactors of personal exposure to high
27 CO levels. Moreover, the study speculated that declining death rates could be attributed in part
28 to automaker compliance with the motor vehicle CO emission standards of the CAA. They
29 argued that tighter CO emission standards may enable cars to emit exhaust into an enclosed space
30 for a longer period of time before CO builds up to toxic levels.
31
October 1999 4-48 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-7. MOTOR VEHICLE CARBON MONOXIDE EMISSION STANDARDS,
TYPICAL IN-VEHICLE CARBON MONOXIDE EXPOSURES, AND
UNINTENTIONAL CARBON MONOXIDE-RELATED DEATH RATES
IN THE UNITED STATES
New Passenger Car
CO Emission Standard
Year
Pre-control
= 1965
1966
1968
1970
1972
1973
1974
1974-75
1975
1976
1977
1978
1979
1980
1981
1981
1981
1981
1981-82
1982
1982-83
1982-83
1983
1984
1985
1986
1987
1987-88
1988
1989
1990
1991-92
1992
Federal
(g/mi)
84.0
84.0
84.0
51.0
34.0
28.0
28.0
28.0
15.0
15.0
15.0
15.0
15.0
15.0
7.0
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
3.4
California
(g/mi)
84.0
84.0
51.0
51.0
34.0
34.0
34.0
34.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
7.0
Net Mean
In- Vehicle CO
Concentration3
(ppm)
12.0
17.5
11.5
7.4
10.3
9.7
8.3
5.2
4.3
2.9
2.9
9.5
1.4
1.8
9.4
4.9
8.4
-3.6
<3.0
CO Exposure Study
Location
Los Angeles
Five U.S. cities
Los Angeles
Boston
Washington
Los Angeles
Santa Clara Co., CA
Denver
Los Angeles
Phoenix
Stamford
Honolulu
Denver
Washington
Washington
Los Angeles
Raleigh
Santa Clara Co., CA
New Jersey suburbs of
New York City, NY
U.S. Unintentional
CO-Related Annual
Death Rate per
100,000 Population
0.67
0.55
0.58
0.58
0.58
0.58
0.56
0.53
0.49
0.49
0.44
0.39
0.39
aMean in-vehicle CO concentration minus mean ambient CO concentration.
Source: Johnson (1988); Cobb and Etzel (1991); Flachsbart (1995); and Faiz et al. (1996).
October 1999
4-49
DRAFT-DO NOT QUOTE OR CITE
-------
1 4.4.2.2 Effects of Motor Vehicle Emission Standards on Passenger Cabin Exposure
2 Based on a review of 16 U.S. studies that occurred between 1965 and 1992, Flachsbart
3 (1995) reported a long-term, downward trend in commuter exposure levels (Table 4-2).
4 Evidence of this downward trend appears in Figure 4-7, which shows the ambient (lower line)
5 and mean CO concentrations inside vehicles (top line) for these studies. These lines do not
6 imply that CO concentrations can be inferred from points on the lines themselves, or that
7 relationships exist between results for different cities. Studies reported typical (mean or median)
8 CO concentrations for trips, most of which lasted an hour or less. Mean CO concentrations fell
9 from 37 ppm in 1965, as reported by Haagen-Smit (1966) for a study in Los Angeles to 3 ppm in
10 1992 for a study by Lawryk et al. (1995) in the New Jersey suburbs of New York City. If one
11 assumes that these results are representative of commuter CO exposures in cities during these
12 time periods, then exposures fell 92% over this 27-year period. This reduction implies that CO
13 exposure levels reported in the past for a particular place and time in the United States may not
14 be indicative of current exposures.
15 In the United States, the effect of progressively tighter CO emission standards on in-vehicle
16 CO exposures over time is readily apparent in Table 4-7. Prior to 1968, each new passenger car
17 emitted 84 g/mi of CO, but by the 1981 model year and thereafter, each new car sold outside of
18 California emitted only 3.4 g/mi of CO, a reduction of 96% (Johnson, 1988). This reduction in
19 certified CO emissions for new passenger cars is roughly equal to the 92% reduction in
20 commuter exposure reported above for the same period. Further analysis reveals that net mean
21 exposure data and the applicable emission standard data in Table 4-7 are highly correlated
22 (r = 0.74, p < 0.0005 for a one-tailed test of the hypothesis). In this analysis, the applicable
23 emission standard (federal or California) was determined by the location of the exposure study.
24 Because the exposure studies did not adhere to a standard protocol, Flachsbart (1995)
25 recommended that future in-vehicle CO exposure studies should use standard protocols to
26 facilitate comparisons and to document the effect on exposure of future measures taken under
27 motor vehicle emission control programs.
28 Two of the 16 studies did follow a standard protocol. Ott et al. (1994) measured in-vehicle
29 CO concentrations on 88 standardized trips over a 1-year period in 1980-1981 on a suburban
30 highway near San Jose. They reported a mean CO concentration of 9.8 ppm for trips of 35 to
31 45 min. In 1991-1992, Ott et al. (1993) resurveyed this highway using a methodology similar to
October 1999 4-50 DRAFT-DO NOT QUOTE OR CITE
-------
Q.
a.
c
o
c
CD
o
o
O
O
O
c
CO
CD
'65 '67 '69 71 73 75 77 79 '81 '83 '85 '87 '89 '91 '93
Year
o Exposure-direct n Ambient-direct * Exposure-indirect • Ambient-indirect
1 Los Angeles, CA
2 Chicago, IL; Cincinnati, OH; Denver, CO;
St. Louis, MO; Washington, DC
3 14 cities
4 Los Angeles, CA
5 Boston, MA
6 Washington, DC
7 Los Angeles, CA
8 Menlo Park, Palo Alto, and Los Altos, CA
9 Denver, CO; Los Angeles, CA; Phoenix, AZ;
Stamford, CT
10 Honolulu, HI
11 a Denver, CO
11 b Washington, DC
12 Washington, DC
13a Los Angeles, CA (summer)
13b Los Angeles, CA (winter)
14 Raleigh, NC
15 Menlo Park, Palo Alto, and Los Altos, CA
16 New Jersey Turnpike and Route 18, NJ
Figure 4-7. Trends in ambient CO concentrations and in-vehicle CO exposures, 1965 to
1992. (The upper and lower lines are provided to make a clear distinction
between exposure and ambient CO data reported for each city; these lines do
not imply that results for cities are related.)
Source: Flachsbart (1995).
1 their previous study to determine in-vehicle exposure trends. They reported that the mean
2 in-vehicle CO concentration had dropped to 4.6 ppm or 47% of the mean value estimated
3 11 years earlier. They attributed the exposure reduction to replacement of older vehicles with
4 newer ones that have lower CO emission factors. This reduction is particularly significant, as
October 1999
4-51
DRAFT-DO NOT QUOTE OR CITE
-------
1 daily traffic volumes on this highway grew by 19.1% during the intervening period, according to
2 estimates by Yu et al. (1996).
3 For this highway, Yu et al. (1996) developed a mathematical model known as the STREET
4 model to predict trends in CO emissions and exposures. Based on fleet turnover and no changes
5 in the 1990 California motor vehicle CO emission standards, the model predicted that the median
6 CO concentrations would drop from 3.9 ppm in 1991 and 1992 to 1.6 to 1.8 ppm in 2002 and
7 2003. At the 99% percentile, the model predicted that the CO concentrations would drop from
8 10 ppm in 1991 and 1992 to 4.0 to 4.6 ppm in 2002 and 2003. This prediction was based on an
9 additional expected reduction of up to 60% in tailpipe emissions of CO primarily because of
10 continued replacement of older cars with newer, low-emission vehicles. However, these
11 predictions could be too low because the study did not anticipate the phenomenal growth in sport
12 utility vehicle (SUV) use in California during the 1990s. The certified CO emissions of SUVs
13 exceed that of standard passenger cars.
14 Similar studies of commuter CO exposure were done by Flachsbart et al. (1987) in the
15 United States, Koushki et al. (1992) in Saudi Arabia, Fernandez-Bremauntz and Ashmore
16 (1995a,b) in Mexico, and Dor et al. (1995) in France. These studies used similar methods of data
17 collection and analysis, with one exception. Smoking was allowed for some trips in the Saudi
18 study, but was not allowed in the other studies. Table 4-8 shows typical values of the net mean
19 CO concentration by travel mode for three of the studies. The net mean CO concentration for the
20 Saudi study could not be determined. The net CO concentrations for each travel mode in Mexico
21 City were much higher than for comparable modes in both Washington and Paris, where net CO
22 concentrations were similar. The similarity between the U.S. and French studies occurred even
23 though catalytic converters existed on 62% of American cars in 1982 (U.S. Department of
24 Commerce, 1983) but were not yet common on French cars in 1992 (Dor et al., 1995).
25 The reasons for the similarity in results between the U.S. and French studies are not readily
26 apparent. However, passenger cabin exposure levels in North and Central America can be
27 explained partly by comparing the history of automotive emission standards in the United States
28 and Mexico. The United States initiated nationwide emission standards on new passenger cars in
29 1968 and adopted progressively tighter controls throughout the 1970s (Johnson, 1988). By the
30 1975 model year, catalytic converters became standard equipment on new passenger cars.
31
October 1999 4-52 DRAFT-DO NOT QUOTE OR CITE
-------
0 TABLE 4-8. TYPICAL NET MEAN CARBON MONOXIDE CONCENTRATION RANGES BY TRAVEL MODE
o FOR CITIES IN THREE COUNTRIES3 b
rt>
vo
VO
VO
Travel Mode
Automobile
Diesel bus
Rail transit
Washington, DC, USA
(1983)
Net Mean CO
Concentrations
(ppm)
7-12
2-6
0-3
Averaging Times
(min)
34-69
82-115
27-48
Mexico City, Mexico
(1991)
Net Mean CO
Concentrations
(ppm)
37-47
14-27
9-13
Averaging Times
(min)
35-63
40-99
39-59
Paris, France
(1991-92)
Net Mean CO
Concentrations Averaging Times
(ppm) (min)
7-10
2-3
1
82-106
NA
NA
H
6
o
2
o
H
O
a"Typical" means do not include outlier values that can be attributed to unusual circumstances.
bNet mean CO concentration = mean in-vehicle CO concentration, minus mean ambient CO concentration.
NA = not available.
Source: Adapted from Flachsbart (1999b).
o
&
o
-------
1 Mexico adopted a tailpipe CO emission standard of 47.0 g/mi for the 1975 model year, and,
2 by the 1993 model year, Mexico finally reached parity with the 1981 U.S. standard of 3.4 g/mi.
3
4 4.4.3 Social Changes Affecting Human Activity Patterns
5 Between 1965 and 1985, the Americans' Use of Time Project at the University of
6 Maryland reported that average time spent in travel for leisure trips increased from 2.7 to
7 3.1 h per week (Cornish et al., 1991). Despite this trend, there is evidence that average
8 commuting times between home and work have remained stable. The decennial census
9 collected travel time data for the first time in 1980. By 1990, the census showed that the
10 nation's average commuting time of 21.7 min in 1980 increased only 40 s to 22.4 min in 1990.
11 Although the number of workers who commuted 45 min or more increased from 10.9 million in
12 1980 to 13.9 million in 1990, the mean travel time of this commuter cohort actually decreased
13 slightly from 59.6 min in 1980 to 58.5 min in 1990. One reason for this is that more people
14 were taking their morning commute from home to work during the "shoulder hours" from 6 to
15 7 a.m. or from 8 to 9 a.m. than during the "peak hour" from 7 to 8 a.m. In 1990, the shoulder
16 hours accounted for about 37% of worker trip starts, whereas the "peak hour" accounted for
17 only 32% of all trip starts (Pisarski, 1992).
18 Typically, average commuting times in large metropolitan regions are greater than those
19 nationwide. In the Washington, DC metropolitan region, for example, the average daily
20 commuting time for all modes of travel between home and work was 62 min per day in 1957
21 (Bello, 1958). This value increased to 69 min per day by 1968 and fell slightly to 68.3 min in
22 1987 and 1988 when the region was resurveyed. The stability in daily commuting times
23 between the 1968 and the 1987 and 1988 surveys was achieved by an increase in travel speed.
24 This increase in travel speed offset greater travel distances between home and work that
25 occurred during that 20-year period. For those who commuted alone by automobile in
26 Washington, the average Euclidean, round-trip travel distance from home to work increased
27 from 13.8 miles in 1968 to 16.1 miles in 1987 and 1988. However, the average trip speeds of
28 solo commuters in Washington also increased (i.e., by 10.7% for home-to-work trips and by
29 20.3% for work-to-home trips) to offset increased commuting distance (Levinson and Kumar,
30 1994). This increase in trip speeds is significant because passenger cabin exposure to CO
October 1999 4-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations has been shown in a separate study to be inversely related to travel speed in the
2 Washington area (Flachsbart et al., 1987).
3 In another study of the Washington area based on the same data and time period, Levinson
4 and Kumar (1995) observed an 85% overall increase in the number of jobs and a decline in
5 average household size from 3.34 to 2.67 people. During the 20-year period from 1968 to 1988,
6 vehicle registrations increased 118%, but road capacity increased only 13%. The average
7 number of autos per household increased from 1.6 to 2.0. However, the most important change
8 was a higher percentage of women in the work force, which forced readjustments and
9 reallocations of time spent in household activities. Specifically, workers had more per capita
10 income but spent less time at home and engaged in more travel for nonwork trips during peak
11 travel periods. Compared to 1968, working men spent 20 min less time at home in 1988, and
12 working women spent about 40 min less. Commuters made multiple stops (i.e., trip chaining)
13 on their way home from work (e.g., visiting health clubs, picking up children at day-care,
14 shopping, eating at restaurants). In 1968, such errands and activities usually were done after the
15 primary worker returned home with the household car. By 1988, these trips often were made in
16 separate vehicles by each household member on their way home from work. By 1988, average
17 time spent daily in travel per person in the Washington area had increased by 14 min for
18 workers and by 11 min for nonworkers over 1968. Levinson and Kumar (1995) said that these
19 results do not support the hypothesis "that individuals spend a fixed amount of time per day
20 (just over 1 h) in transportation, and make all budget allocation adjustment on non-travel times."
21 Instead, they suggested that some urban households have been spending more time in travel and
22 less time at home, and have been buying more household services outside the home.
23 On the other hand, Levinson and Kumar (1995) anticipate that some people will spend
24 more time at home in the future. They noted: "Several factors suggest that work at home,
25 telecommuting, and teleshopping may be on the verge of wide-spread adoption. The technology
26 is coming into place with the long-awaited advent of videophones, and of the 'information
27 superhighway', that is, broad-band two-way communications facilitated by the recent
28 consolidations in the telecommunications and entertainment industries." The percentage of
29 people working at home increased from 2.3% in 1980 to 3% in 1990 (Pisarski, 1992).
30 Currently, an estimated 52 million Americans are self-employed to some extent, working either
31 in home offices for themselves or for companies as telecommuters. In 1975, only 2.5 million
October 1999 4-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 Americans worked at home. The U.S. General Services Administration expects to see
2 60,000 telecommuters working for the federal government by the end of 1998. In 1994, there
3 were fewer than 4,000 such workers (Webster, 1998). This employment shift could have
4 tremendous implications for population exposure.
5
6
7 4.5 CONCLUSIONS
8 This chapter has reviewed studies of population exposure to carbon monoxide, including
9 some key studies from the previous CO AQCD, and studies that have been published in the
10 peer-reviewed scientific literature since 1991. This section draws several conclusions from this
11 review, and identifies both the extent to which CO exposures have changed since the last AQCD
12 and some current gaps in knowledge about population exposure to CO. The previous CO AQCD
13 concluded that, on an individual basis, personal exposure is poorly correlated with ambient CO
14 concentrations as measured at fixed-site monitors, because of personal mobility relative to the
15 monitor's fixed location, and the spatial and temporal variability of CO concentrations (U.S.
16 Environmental Protection Agency, 1991). Like earlier studies, more recent ones indicate that
17 the extent and magnitude of observed personal CO exposures may be greater than those
18 predicted from fixed-site monitors used to determine compliance with the NAAQS (Wilson
19 et al., 1993a,b; Colome et al., 1994). Yet, when ambient CO levels are either high or low on a
20 given day, fixed-site monitors still reflect the corresponding high or low aggregates of personal
21 exposures on those days. Otherwise, the stations do not adequately represent the CO exposures
22 of community residents while they are exposed to motor vehicle exhaust during commuting,
23 occupational and residential sources of unvented fuel combustion, or tobacco smoke. The mean
24 COHb level of people exposed to CO from these sources will be greater than their mean COHb
25 level predicted solely from exposure to CO of ambient origin.
26 Implementation of motor vehicle emission standards, catalytic converters, motor vehicle
27 inspection/maintenance programs, and cleaner burning fuels during the past three decades has
28 reduced the CO exposures of urban commuters (Flachsbart, 1995, 1999b). This conclusion has
29 important implications, because it suggests that modeled estimates of current commuter
30 exposure, based on data inputs from pre-1990 exposure studies, may be too high. Moreover, the
31 Yu et al. (1996) study indicates that the average CO concentrations in passenger cabins of motor
October 1999 4-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 vehicles are expected to drop further in the near future. However, those projections could be too
2 low because the study did not anticipate or account for the phenomenal growth in sport utility
3 vehicle (SUV) use during the 1990s. The certified CO emissions of SUVs presently exceed
4 those of standard passenger cars, and the U.S. Environmental Protection Agency did not propose
5 tighter emission standards for SUVs until 1999.
6 Likewise, there are uncertainties over the extent to which population exposure to CO has
7 changed in other ways since the 1991 AQCD. First, there are no apparent trend studies of CO
8 exposure in other important microenvironments (e.g., indoor parking garages, pedestrian
9 sidewalks on commercial streets, or home environments affected by greater use of microwaves
10 for cooking in lieu of gas ranges). Second, the net effect of various travel behavior trends on
11 commuter CO exposure is uncertain. Trends noted in this chapter include disproportionately
12 high growth rates in vehicle miles of travel, growth in travel during "shoulder" hours of
13 peak-traffic periods, and growing use of personal computers for telecommuting and
14 teleshopping from home in lieu of trips by motor vehicles. These trends and their implications
15 for exposure suggest that the results of earlier personal CO monitoring studies, such as those of
16 Akland et al. (1985) summarized in Table 4-1, are probably no longer indicative of present CO
17 levels and population exposures. These types of personal exposure studies would need to be
18 redone to determine current CO exposure levels of similar urban populations.
19 The 1991 CO AQCD reported that people are exposed to elevated CO levels in certain
20 indoor microenvironments (e.g., unventilated parking garages, motor vehicles with leaky
21 exhaust systems, and small homes with unvented gas stoves and space heaters). More recent
22 studies in California homes indicate that elevated CO concentrations still exist and can be
23 caused by several factors such as attached garages and carports, ranges with continuous gas pilot
24 lights, improper use and installation of gas appliances, especially in small homes (Wilson
25 et al.,1993a,b; Colome et al.,1994). Also, recent studies have found elevated CO concentrations
26 when people ride certain types of recreational vehicles (i.e., snowmobiles and power boats),
27 gather indoors to barbecue food (sometimes to cope with electrical power outages), and watch
28 sporting events held at indoor arenas. High-level exposures may occur inside arenas when they
29 are used for ice skating, motocross and monster-truck competitions, and tractor-pulls. Vehicles
30 used in these competitions often lack any type of emission controls. In some cases, ventilation
31 alone has not lowered CO sufficiently to safe levels at these events. Moreover, recent studies
October 1999 4-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 report that high-level CO exposures can occur when people use unregulated gasoline-powered
2 appliances, engines, and tools, (e.g., chainsaws), even under ventilated conditions.
3 The previous CO AQCD reported that many Americans spend most of their time indoors.
4 This finding still appears to be true, according to more recent studies of activity patterns in
5 California and nationwide (e.g., the NHAPS study). They indicate that Americans now spend,
6 on average, between 87 and 89% of their day indoors and about 7% of their time in or near
7 vehicles. However, activity patterns have shifted since the pioneering studies of the early 1970s.
8 Recent study of travel behavior in Washington, DC, indicates that some people are spending
9 relatively more time in travel and less time at home compared to the past, because of growth in
10 the service sector of the nation's economy. This growth has enabled a growing number of
11 people to buy services outside the home that once were provided by household members.
12 Buyers often visit one or more retailers (i.e., trip chaining) as part of the daily commute to and
13 from work (Levinson and Kumar, 1995) which extends their trip times.
14 In analyzing activity patterns, the California and national studies both analyzed activities
15 by conventional social categories (e.g., gender, age, race, etc.). In a travel behavior study,
16 Vadarevu and Stopher (1996) used a different social construct that revealed significant
17 differences in activity patterns among different life-cycle groups defined by age, working status,
18 and household size. Hence, although their study did not focus on CO exposure assessment
19 per se, it still has useful lessons for the design and analysis of activity pattern studies. Their
20 study also revealed that modern urban people make continuous tradeoffs in their activity
21 patterns and household-role allocations (in terms of who gets to do what and when) in response
22 to ongoing social and technological changes. However, the effects of these activity and role
23 adjustments by householders on personal CO exposure still needs to be documented by
24 empirical study.
25 In light of the above, population exposure models (e.g., pNEM/CO and SHAPE) may need
26 to sample from distributions that more accurately represent current microenvironmental CO
27 concentrations and time budgets and add certain high exposure level microenvironments (e.g.,
28 tobacco smoke exposure while in a vehicle, or sporting events involving motor vehicles at
29 indoor arenas) to their current list. In the future, simulation models of exposure should consider
30 that trip times and commuter exposures are not independent of trip-starting times, and that the
31 distribution of CO exposure is homogeneous for all types of commuters. As evidence of the
October 1999 4-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 former, Flachsbart (1999a) showed that a commuter's travel time and CO exposure inside a
2 passenger car for a trip from home to work was related to trip departure time. Not surprisingly,
3 travel during off-peak hours (i.e., the "shoulder" periods) to avoid congested traffic resulted in
4 both less travel time and CO exposure. As evidence of the latter, carpoolers can reduce their
5 CO exposure if they use high-occupancy vehicle lanes on highways, as shown by both
6 Flachsbart (1989) in Hawaii and by Rodes et al. (1998) in Southern California.
7
October 1999 4-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Agency for Toxic Substances and Disease Registry. (1993) Methylene chloride toxicity. Am. Fam. Phy.
3 47: 1159-1166.
4 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
5 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
6 Technol. 19: 911-918.
7 Aim, S.; Jantunen, M. J.; Vartiainen, M. (1999) Urban commuter exposure to particle matter and carbon monoxide
8 inside an automobile. J. Exposure Anal. Environ. Epidemiol. 9: 237-244.
9 Amiro, A. G. (1969) Carbon monoxide presents public health problem. Presented at: 33rd annual educational
10 conference, National Association of Sanitarians; June; Houston, TX. J. Environ. Health 32: 83-88.
11 Apte, M. G. (1997) A population-based exposure assessment methodology for carbon monoxide: development of
12 a carbon monoxide passive sampler and occupational dosimeter [dissertation]. Berkeley, CA: U.S.
13 Department of Energy, Lawrence Berkeley National Laboratory, Environmental Energy Technologies
14 Division; LBNL-40838. Available from: NTIS, Springfield, VA; DE98052714.
15 Apte, M. G.; Traynor, G. W. (1993) Pollutant emission rates from a radiant fiber-matrix gas burner. Air Waste
16 43:223-225.
17 Austin, B. S.; Heiken, J. G.; Shepard, S. B.; Duvall, L. L. (1994) Methodologies for estimating emission and travel
18 activity effects of TCMs. Washington, DC: U.S. Environmental Protection Agency, Office of Mobile
19 Sources; report no. EPA-420-R-94-002.
20 Barnard, R. J.; Weber, J. S. (1979) Carbon monoxide: a hazard to firefighters. Arch. Environ. Health 34: 255-257.
21 Baucom, C. D.; Freeman, J. I.; MacCormack, J. M. (1987) Carbon monoxide poisoning in a
22 garment-manufacturing plant - North Carolina. Morb. Mortal. Wkly. Rep. 36: 543-545.
23 Bellin, P.; Spengler, J. D. (1980) Indoor and outdoor carbon monoxide measurements at an airport. J. Air Pollut.
24 Control Assoc. 30: 392-394.
25 Bello, F. (1958) The city and the car. In: The exploding metropolis. Garden City, NJ: Doubleday Anchor Books;
26 pp. 32-61.
27 Bevan, M. A. J.; Proctor, C. J.; Baker-Rogers, J.; Warren, N. D. (1991) Exposure to carbon monoxide, respirable
28 suspended particulates, and volatile organic compounds while commuting by bicycle. Environ. Sci. Technol.
29 25:788-791.
30 Boeniger, M. F. (1995) Description of facilities and employee exposure to carbon monoxide at four different
31 bridge and tunnel toll authorities in the United States. Cincinnati, OH: National Institute for Occupational
32 Safety and Health, Division of Surveillance, Field Studies, Hazard Evaluations; report no. IWS-45-02.
33 Available from: NTIS, Springfield, VA; PB96-106539.
34 Boudreau, D. R.; Spadafora, M. P.; Wolf, L. R.; Siegel, E. (1994) Carbon monoxide levels during indoor sporting
35 events—Cincinnati, 1992-1993 [reprint fromMMWR 43: 21-23]. JAMA J. Am. Med. Assoc. 271: 419.
36 Brandt-Rauf, P. W.; Fallen, L. F., Jr.; Tarantini, T.; Idema, C.; Andrews, L. (1988) Health hazards of fire fighters:
37 exposure assessment. Br. J. Ind. Med. 45: 605-612.
38 Brotherhood, J. R.; Budd, G. M.; Jeffery, S. E.; Hendrie, A. L.; Beasley, F. A.; Costin, B. P.; Wu, Z. (1990) Fire
39 fighters' exposure to carbon monoxide during Australian bushfires. Am. Ind. Hyg. Assoc. J. 51: 234-240.
40 Buie, S. E.; Pratt, D. S.; May, J. J. (1986) Diffuse pulmonary injury following paint remover exposure. Am. J.
41 Med. 81:702-704.
42 Btinger, J.; Bombosch, F.; Mesecke, U.; Hallier, E. (1997) Monitoring and analysis of occupational exposure to
43 chain saw exhausts. Am. Ind. Hyg. Assoc. J. 58: 747-751.
44 Burgess, W. A.; Sidor, R.; Lynch, J. J.; Buchanan, P.; Clougherty, E. (1977) Minimum protection factors for
45 respiratory protective devices for firefighters. Am. Ind. Hyg. Assoc. J. 38: 18-23.
46 Chan, C.-C.; Ozkaynak, H.; Spengler, J. D.; Sheldon, L. (1991) Driver exposure to volatile organic compounds,
47 CO, ozone and NO2 under different driving conditions. Environ. Sci. Technol. 25: 964-972.
48 Chapin, F. S., Jr. (1974) Human activity patterns in the city: things people do in time and in space. New York, NY:
49 John Wiley & Sons.
50 Chung, Y.; Park, S. E.; Lee, K.; Yanagisawa, Y.; Spengler, J. D. (1994) Determinations of personal carbon
51 monoxide exposure and blood carboxyhemoglobin levels in Korea. Yonsei Med. J. 35: 420-428.
52 Clements, J. S. (1978) School bus carbon monoxide intrusion. Washington, DC: U.S. Department of
53 Transportation, National Highway Traffic Safety Administration; report no. DOT-HS-803 705. Available
54 from: NTIS, Springfield, VA; PB-291156/8.
October 1999 4-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 Cobb, N.; Etzel, R. A. (1991) Unintentional carbon monoxide-related deaths in the United States, 1979 through
2 1988. JAMA J. Am. Med. Assoc. 266: 659-663.
3 Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
4 monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research
5 Institute; report no. GRI-93/0224.3. Available from: NTIS, Springfield, VA; PB96-159421.
6 Communication Sciences Group. (1998) Smoke-free bar patron survey: summary of findings. Los Angeles, CA:
7 Los Angeles County Department of Health.
8 Cornish, E.; Willard, T.; Fields, D. (1991) How Americans use time: an interview with sociologist John P.
9 Robinson. Futurist Sept./Oct: 23-27.
10 Cortese, A. D.; Spengler, J. D. (1976) Ability of fixed monitoring stations to represent personal carbon monoxide
11 exposure. J. AirPollut. Control Assoc. 26: 1144-1150.
12 Cutler, B. (1990) Where does the free time go? Am. Demograph. (November): 36-39.
13 Dor, F.; Le Moullec, Y.; Festy, B. (1995) Exposure of city residents to carbon monoxide and monocyclic aromatic
14 hydrocarbons during commuting trips in the Paris metropolitan area. J. Air Waste Manage. Assoc.
15 45: 103-110.
16 Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
17 Ely, E. W.; Moorehead, B.; Haponik, E. F. (1995) Warehouse workers' headache: emergency evaluation and
18 management of 30 patients with carbon monoxide poisoning. Am. J. Med. 98: 145-155.
19 Faiz, A.; Weaver, C. S.; Walsh, M. P.; Gautam, S. P.; Chan, L.-M. (1996) Emission standards and regulations.
20 In: Air pollution from motor vehicles: standards and technologies for controlling emissions. Washington,
21 DC: World Bank; pp. 1-24.
22 Farrow, A.; Taylor, H.; Golding, J. (1997) Time spent in the home by different family members. Environ. Technol.
23 18:605-613.
24 Fawcett, T. A.; Moon, R. E.; Fracica, P. J.; Mebane, G. Y.; Theil, D. R.; Piantadosi, C. A. (1992) Warehouse
25 workers' headache: carbon monoxide poisoning from propane-fueled forklifts. J. Occup. Med. 34: 12-15.
26 Fernandez-Bremauntz, A. A.; Ashmore, M. R. (1995a) Exposure of commuters to carbon monoxide in Mexico
27 City—I. Measurement of in-vehicle concentrations. Atmos. Environ. 29: 525-532.
28 Fernandez-Bremauntz, A. A.; Ashmore, M. R. (1995b) Exposure of commuters to carbon monoxide in Mexico
29 City II. Comparison of in-vehicle and fixed-site concentrations. J. Exposure Anal. Environ. Epidemiol.
30 5:497-510.
31 Field Research Corporation. (1998) A survey of California bar patrons about smoking policies and smoke-free
32 bars. Sacramento, CA: California Department of Health Services; March.
33 Flachsbart, P. G. (1989) Effectiveness of priority lanes in reducing travel time and carbon monoxide exposure.
34 Inst. Transport. Eng. J. 59: 41-45.
3 5 Flachsbart, P. G. (1995) Long-term trends in United States highway emissions, ambient concentrations, and
36 in-vehicle exposure to carbon monoxide in traffic. J. Exposure Anal. Environ. Epidemiol. 5: 473-495.
37 Flachsbart, P. G. (1999a) Models of exposure to carbon monoxide inside a vehicle on a Honolulu highway.
38 J. Exposure Anal. Environ. Epidemiol. 9: 245-260.
39 Flachsbart, P. G. (1999b) Human exposure to carbon monoxide from mobile sources. Chemosphere: accepted.
40 Flachsbart, P. G.; Brown, D. E. (1989) Employee exposure to motor vehicle exhaust at a Honolulu shopping
41 center. J. Architect. Plan. Res. 6: 19-33.
42 Flachsbart, P. G.; Mack, G. A.; Howes, J. E.; Rodes, C. E. (1987) Carbon monoxide exposures of Washington
43 commuters. JAPCA 37: 135-142.
44 Gardiner, K.; Trethowan, W. N.; Harrington, J. M.; Calvert, I. A.; Glass, D. C. (1992) Occupational exposure to
45 carbon monoxide and sulphur dioxide during the manufacture of carbon black. Ann. Occup. Hyg.
46 36: 363-372.
47 Gasman, J. D.; Varon, J.; Gardner, J. P. (1990) Revenge of the barbecue grill: carbon monoxide poisoning. West.
48 J. Med. 153:656-657.
49 Ghittori, S.; Marraccini, P.; Franco, G.; Imbriani, M. (1993) Methylene chloride exposure in industrial workers.
50 Am. Ind. Hyg. Assoc. J. 54:27-31.
51 Girman, J. R.; Chang, Y.-L.; Hayward, S. B.; Liu, K.-S. (1998) Causes of unintentional deaths from carbon
52 monoxide poisonings in California. West. J. Med. 168: 158-165.
53 Gold, A.; Burgess, W. A.; Clougherty, E. V. (1978) Exposure of firefighters to toxic air contaminants. Am. Ind.
54 Hyg. Assoc. J. 39: 534-539.
October 1999 4-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gordon, B. (1998) Easing the transition to smoke-free bars. Contra Costa-Solano Counties, CA: American Lung
2 Association of Contra Costa-Solano.
3 Gourdeau, P.; Parent, M; Soulard, A. (1995) Exposition a 1'oxyde de carbone dans les garages d'automobiles:
4 evaluation chez les mecaniciens [Exposure of carbon monoxide in car repair shops: evaluation of
5 mechanics]. Can. J. Public Health 86: 414-417.
6 Haagen-Smit, A. J. (1966) Carbon monoxide levels in city driving. Arch. Environ. Health 12: 548-551.
7 Hampson, N. B. (1996) Carbon monoxide poisoning at an indoor ice arena and bingo hall—Seattle, 1996 [reprint
8 of MMWR 1996; 45: 265-267]. JAMA J. Am. Med. Assoc. 275: 1468-1469.
9 Hampson, N. B.; Norkool, D. M. (1992) Carbon monoxide poisoning in children riding in the back of pickup
10 trucks. JAMA J. Am. Med. Assoc. 267: 538-540.
11 Hampson, N. B.; Kramer, C. C.; Dunford, R. G.; Norkool, D. M. (1994) Carbon monoxide poisoning from indoor
12 burning of charcoal briquets. JAMA J. Am. Med. Assoc. 271: 52-53.
13 Holland, D. M. (1983) Carbon monoxide levels in microenvironment types of four U.S. cities. Environ. Int.
14 9:369-377.
15 Jankovic, J.; Jones, W.; Burkhart, J.; Noonan, G. (1991) Environmental study of firefighters. Ann. Occup. Hyg.
16 35:581-602.
17 Jarvis, M. J.; Russell, M. A. H.; Saloojee, Y. (1980) Expired air carbon monoxide: a simple breath test of tobacco
18 smoke intake. Br. Med. J. 281: 484-485.
19 Jenkins, P. L.; Phillips, T. J.; Mulberg, E. J.; Hui, S. P. (1992) Activity patterns of Californians: use of and
20 proximity to indoor pollutant sources. Atmos. Environ. Part A 26: 2141-2148.
21 Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle
22 Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; EPA
23 report no. EPA-600/4-84-014. Available from: NTIS, Springfield, VA; PB84-146125.
24 Johnson, J. H. (1988) Automotive emissions. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the
25 automobile, and public health. Washington, DC: National Academy Press; pp. 39-75.
26 Johnson, T. (1989) Human activity patterns in Cincinnati, Ohio [final report]. Palo Alto, CA: Electric Power
27 Research Institute; project no. RP940-06; report no. EN-6204.
28 Johnson, T.; Paul, R. A. (1983) The NAAQS Exposure Model (NEM) applied to carbon monoxide. Research
29 Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
30 EPA report no. EPA-450/5-83-003. Available from: NTIS, Springfield, VA; PB84-242551.
31 Johnson, T.; Capel, J.; Paul, R.; Wijnberg, L. (1992) Estimation of carbon monoxide exposures and associated
32 carboxyhemoglobin levels in Denver residents using a probabilistic version of NEM. Research Triangle
33 Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; contract
34 no. 68-DO-0062.
3 5 Johnson, T.; Mihlan, G.; LaPointe, J.; Fletcher, K.; Capel, J. (1999) Estimation of carbon monoxide exposures and
3 6 associated carboxyhemoglobin levels in Denver residents using pNEM/CO (version 2.0) [draft]. Research
37 Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
38 March 15.
39 Jones, R. H.; Ellicott, M. F.; Cadigan, J. B.; Gaensler, E. A. (1958) The relationship between alveolar and blood
40 carbon monoxide concentrations during breathholding. J. Lab. Clin. Med. 51: 553-564.
41 Joumard, R. (1991) Pollution de 1'air dans les transports [Air pollution from transportation]. Pollut. Atmos.
42 33:20-24.
43 Kahler, M.; Kuhse, W.; Wintermeyer, L. A. (1993) Unintentional carbon monoxide poisoning from indoor use of
44 pressure washers—Iowa, January 1992-January 1993. Morb. Mortal. Wkly. Rep. 42: 777-779, 785.
45 Kamei, M.; Yanagisawa, Y. (1997) Estimation of CO exposure of road construction workers in tunnel. Ind. Health
46 35: 119-125.
47 Kern, R. A.; Knoedler, J. A.; Trester, R. J. (1990) Automotive carbon monoxide emissions in a closed garage.
48 Presented at: International congress and exposition; February-March; Detroit, MI. Warrendale, PA: Society
49 of Automotive Engineers; pp. 33-38. (SAE technical paper series no. 900499).
50 Klepeis, N. E.; Tsang, A. M.; Behar, J. V. (1996) Analysis of the national human activity pattern survey (NHAPS)
51 respondents from a standpoint of exposure assessment. Washington, DC: U.S. Environmental Protection
52 Agency, Office of Research and Development; report no. EPA/600/R-96/074.
53 Koontz, M. D.; Nagda, N. L. (1988) A topical report on a field monitoring study of homes with unvented gas space
54 heaters. Volume III—methodology and results. Chicago, IL: Gas Research Institute; report no.
55 GRI-87/0044.2. Available from: NTIS, Springfield, VA; PB88-216791.
October 1999 4-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 Koushki, P. A.; Al-Dhowalia, K. H.; Niaizi, S. A. (1992) Vehicle occupant exposure to carbon monoxide. J. Air
2 Waste Manage. Assoc. 42: 1603-1608.
3 Kwok, P. W. (1981) Reduction of carbon monoxide in indoor skating arenas. Environ. Health Rev. 25: 60-63.
4 Lando, H. A.; McGovern, P. G.; Kelder, S. H.; Jeffery, R. W.; Forster, J. L. (1991) Use of carbon monoxide breath
5 validation in assessing exposure to cigarette smoke in a worksite population. Health Psychol. 10: 296-301.
6 Last, J. M; Abramson, J. H.; Friedman, G. D.; Porta, M; Spasoff, R. A.; Thuriaux, M. (1995) A dictionary of
7 epidemiology. 3rd ed. New York, NY: Oxford University Press; pp. 5, 60.
8 Law, P. L.; Lioy, P. J.; Zelenka, M. P.; Huber, A. H.; McCurdy, T. R. (1997) Evaluation of a probabilistic
9 exposure model applied to carbon monoxide (pNEM/CO) using Denver personal exposure monitoring data.
10 J. Air Waste Manage. Assoc. 47: 491-500.
11 Lawryk, N. J.; Lioy, P. J.; Weisel, C. P. (1995) Exposure to volatile organic compounds in the passenger
12 compartment of automobiles during periods of normal and malfunctioning operation. J. Exposure Anal.
13 Environ. Epidemiol. 5: 511-531.
14 Lee, K.; Yanagisawa, Y. (1992) Development of a sampler for carbon monoxide in expired air. Presented at: 85th
15 annual meeting and exhibition of the Air & Waste Management Association; June; Kansas City, MO.
16 Pittsburgh, PA: Air & Waste Management Association; paper no. 92-145.03.
17 Lee, K.; Yanagisawa, Y. (1995) Sampler for measurement of alveolar carbon monoxide. Environ. Sci. Technol.
18 29: 104-107.
19 Lee, K.; Yanagisawa, Y.; Spengler, J. D. (1993) Carbon monoxide and nitrogen dioxide levels in an indoor ice
20 skating rink with mitigation methods. Air Waste 43: 769-771.
21 Lee, K.; Yanagisawa, Y.; Spengler, J. D.; Nakai, S. (1994) Carbon monoxide and nitrogen dioxide exposures in
22 indoor ice skating rinks. J. Sports Sci. 12: 279-283.
23 Lees, P. S. J. (1995) Combustion products and other firefighter exposures. Occup. Med.: State of the Art Rev.
24 10:691-706.
25 Leikin, J. B.; Kaufman, D.; Lipscomb, J. W.; Burda, A. M.; Hryhorczuk, D. O. (1990) Methylene chloride: report
26 of five exposures and two deaths. Am. J. Emerg. Med. 8: 534-537.
27 Levesque, B.; Dewailly, E.; Lavoie, R.; Prud'Homme, D.; Allaire, S. (1990) Carbon monoxide in indoor ice skating
28 rinks: evaluation of absorption by adult hockey players. Am. J. Public Health 80: 594-598.
29 Levesque, B.; Lavoie, R.; Dewailly, E.; Prud'Homme, D.; Allaire, S. (1991) An experiment to evaluate carbon
30 monoxide absorption by hockey players in ice skating rinks. Vet. Hum. Toxicol. 33: 5-8.
31 Levesque, B.; Allaire, S.; Prud'homme, H.; Rhainds, M.; Lebel, G.; Bellemarre, D.; Dupuis, K. (1997) Indoor
32 motocross competitions: air quality evaluation. Am. Ind. Hyg. Assoc. J. 58: 286-290.
33 Levinson, D. M.; Kumar, A. (1994) The rational locator: why travel times have remained stable. J. Am. Plan.
34 Assoc. 60:319-332.
35 Levinson, D.; Kumar, A. (1995) Activity, travel, and the allocation of time. J. Am. Plan. Assoc. 61: 458-470.
36 Licavoli, K. (1998) Bar survey results for select cities in San Mateo county. San Francisco, CA: American Lung
3 7 Association of San Francisco and San Mateo Counties.
38 Limasset, J.-C.; Diebold, F.; Hubert, G. (1993) Exposition des conducteurs de bus urbains aux polluants de la
39 circulation automobile [Assessment of bus drivers' exposure to the pollutants of urban traffic]. Sci. Total
40 Environ. 134: 39-49.
41 Lioy, P. J. (1990) Assessing total human exposure to contaminants: a multidisciplinary approach. Environ. Sci.
42 Technol. 24: 938-945.
43 Liu, K.-S.; Girman, J. R.; Hayward, S. B.; Shusterman, D.; Chang, Y.-L. (1993) Unintentional carbon monoxide
44 deaths in California from charcoal grills and hibachis. J. Exposure Anal. Environ. Epidemiol.
45 3(suppl. 1): 143-151.
46 Luckhurst, D. G.; Solkoski, G. (1990) Carbon monoxide levels in indoor "tractor-pull" events—Manitoba. Can.
47 Dis. Wkly. Rep. 16-17: 79-81.
48 Madani, I. M.; Khalfan, S.; Khalfan, H.; Jidah, J.; Aladin, M. N. (1992) Occupational exposure to carbon
49 monoxide during charcoal meat grilling. Sci. Total Environ. 114: 141-147.
50 Mage, D. T. (1991) A comparison of the direct and indirect methods of human exposure. In: Gledhill, B. L.;
51 Mauro, F., eds. New horizons in biological dosimetry: proceedings of the international symposium on trends
52 in biological dosimetry; October 1990; Lerici, Italy. New York, NY: Wiley-Liss, Inc.; pp. 443-454.
53 (Progress in clinical and biological research: v. 372).
54 Materna, B. L.; Jones, J. R.; Sutton, P. M.; Rothman, N.; Harrison, R. J. (1992) Occupational exposures in
55 California wildland fire fighting. Am. Ind. Hyg. Assoc. J. 53: 69-76.
October 1999 4-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 Materna, B. L.; Koshland, C. P.; Harrison, R. J. (1993) Carbon monoxide exposure in wildland firefighting:
2 a comparison of monitoring methods. Appl. Occup. Environ. Hyg. 8: 479-487.
3 McCammon, C. S.; Halperin, W. F.; Lemen, R. A. (1981) Carbon monoxide exposure from aircraft fueling
4 vehicles. Arch. Environ. Health 36: 136-138.
5 McCammon, J. B.; McKenzie, L. E.; Heinzman, M. (1996) Carbon monoxide poisoning related to the indoor use
6 of propane-fueled forklifts in Colorado workplaces. Appl. Occup. Environ. Hyg. 11: 192-198.
7 McCurdy, T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
8 models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
9 Meyer, B. (1983) Indoor air quality. Reading, MA: Addison-Wesley Publishing Company, Inc.
10 Miller, R. K.; Ryan, M. C.; Bilowus, P. (1989) Carbon monoxide poisoning in indoor ice skating arenas. Va. Med.
11 116:74-76.
12 Mumford, J. L.; Williams, R. W.; Walsh, D. B.; Burton, R. M.; Svendsgaard, D. J.; Chuang, J. C.; Houk, V. S.;
13 Lewtas, J. (1991) Indoor air pollutants from unvented kerosene heater emissions in mobile homes: studies
14 on particles, semivolatile organics, carbon monoxide, and mutagenicity. Environ. Sci. Technol.
15 25: 1732-1738.
16 National Institute for Occupational Safety and Health. (1994) Health hazard evaluation report HETA
17 90-0365-2415, U.S. Department of the Interior, National Park Service, Yosemite National Park, California.
18 Cincinnati, OH: U.S. Department of Health and Human Services, Public Health Service, Centers for
19 Disease Control and Prevention. Available from: NTIS, Springfield, VA; PB95-242541.
20 National Institute for Occupational Safety and Health. (1996) Preventing carbon monoxide poisoning from small
21 gasoline-powered engines and tools [NIOSH alert]. Cincinnati, OH: U.S. Department of Health and Human
22 Services, Public Health Service; report no. DHHS (NIOSH) publication no. 96-118. Available from: NTIS,
23 Springfield, VA; PB97-147078.
24 Ott, W. R. (1982) Concepts of human exposure to air pollution. Environ. Int. 7: 179-196.
25 Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
26 21:97-128.
27 Ott, W.; Eliassen, R. (1973) A survey technique for determining the representativeness of urban air monitoring
28 stations with respect to carbon monoxide. J. Air Pollut. Control Assoc. 23: 685-690.
29 Ott, W.; Flachsbart, P. (1982) Measurement of carbon monoxide concentrations in indoor and outdoor locations
30 using personal exposure monitors. Environ. Int. 8: 295-304.
31 Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
32 exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
33 Environ. 22:2101-2113.
34 Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities
35 for human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
36 Ott, W.; Switzer, P.; Willits, N.; Hildemann, L.; Yu, L. (1993) Trends of in-vehicle CO exposures on a California
37 arterial highway over one decade. Research Triangle Park, NC: U.S. Environmental Protection Agency,
3 8 Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/A-93/220. Available
39 from: NTIS, Springfield, VA; PB93-234508.
40 Ott, W.; Switzer, P.; Willits, N. (1994) Carbon monoxide exposures inside an automobile traveling on an urban
41 arterial highway. J. Air Waste Manage. Assoc. 44: 1010-1018.
42 Paul, R. A.; Johnson, T. (1985) The NAAQS exposure model (NEM) applied to carbon monoxide: addendum.
43 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
44 Standards; report no. EPA 450/5-85-004. Available from: NTIS, Springfield, VA; PB85-235182.
45 Paul, R. A.; Johnson, T.; McCurdy, T. (1988) Advancements in estimating urban population exposure. Presented
46 at: 81st annual meeting of the Air Pollution Control Association; June; Dallas, TX. Pittsburgh, PA: Air
47 Pollution Control Association; paper no. 88-127.1.
48 Paulozzi, L. J.; Satink, F.; Spengler, R. F. (1991) A carbon monoxide mass poisoning in an ice arena in Vermont
49 [letter]. Am. J. Public Health 81: 222.
50 Paulozzi, L. J.; Spengler, R. F.; Vogt, R. L.; Carney, J. K. (1993) A survey of carbon monoxide and nitrogen
51 dioxide in indoor ice arenas in Vermont. J. Environ. Health 56(5): 23-25.
52 Pedersen, D. H.; Sieber, W. K. (1988) National occupational exposure survey. Volume III: analysis of management
53 interview responses. Cincinnati, OH: U.S. Department of Health and Human Services, National Institute for
54 Occupational Safety and Health; DHHS (NIOSH) publication no. 89-103. Available from: NTIS,
55 Springfield, VA; PB90-193640.
October 1999 4-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 Peterson, J. E. (1978) Modeling the uptake, metabolism and excretion of dichloromethane by man. Am. Ind. Hyg.
2 Assoc. J. 39: 41-47.
3 Phillips, T. J.; Jenkins, P. L.; Mulberg, E. J. (1991) Children in California: activity patterns and presence of
4 pollutant sources. In: Health risk and risk communication: papers from the 84th annual meeting &
5 exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada. Pittsburgh, PA:
6 Air & Waste Management Association; volume 17, paper no. 91-172.5.
7 Pisarski, A. E. (1992) New perspectives in commuting: based on early data from the 1990 decennial census and the
8 1990 Nationwide Personal Transportation Study. Washington, DC: U.S. Department of Transportation,
9 Federal Highway Administration; report no. FHWA-PL/92-026. Available from: NTIS, Springfield, VA;
10 PB97-165567.
11 Quackenboss, J. J.; Spengler, J. D.; Kanarek, M. S.; Letz, R.; Duffy, C. P. (1986) Personal exposure to nitrogen
12 dioxide: relationship to indoor/outdoor air quality and activity patterns. Environ. Sci. Technol. 20: 775-783.
13 Raaschou-Nielsen, O.; Nielsen, M. L.; Gehl, J. (1995) Traffic-related air pollution: exposure and health effects in
14 Copenhagen street cleaners and cemetery workers. Arch. Environ. Health 50: 207-213.
15 Radford, E. P.; Drizd, T. A. (1982) Blood carbon monoxide levels in persons 3-74 years of age: United States,
16 1976-80. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service, National
17 Center for Health Statistics; publication no. (PHS) 82-1250. (Advance data from vital and health statistics:
18 no. 76).
19 Radford, E. P.; Levine, M. S. (1976) Occupational exposures to carbon monoxide in Baltimore firefighters.
20 J. Occup. Med. 18: 628-632.
21 Robinson, J. P.; Thomas, J. (1991) Time spent in activities, locations, and microenvironments: a
22 California-national comparison. Las Vegas, NV: U.S. Environmental Protection Agency, Environmental
23 Monitoring Systems Laboratory; report no. EPA/600/4-91/006. Available from: NTIS, Springfield, VA;
24 PB92-140789.
25 Schwab, M.; Colome, S. D.; Spengler, J. D.; Ryan, P. B.; Billick, I. H. (1990) Activity patterns applied to pollutant
26 exposure assessment: data from a personal monitoring study in Los Angeles. Toxicol. Ind. Health
27 6:517-532.
28 Seufert, K. T.; Kiser, W. R. (1996) End-expiratory carbon monoxide levels as an estimate of passive smoking
29 exposure aboard a nuclear-powered submarine. South. Med. J. 89: 1181-1183.
30 Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
31 models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public health.
32 Washington, DC: National Academy Press; pp. 207-238.
33 Shenoi, R.; Stewart, G.; Rosenberg, N. (1998) Screening for carbon monoxide in children. Pediatr. Emer. Care
34 14:399-402.
35 Shusterman, D.; Quinlan, P.; Lowengart, R.; Cone, J. (1990) Methylene chloride intoxication in a furniture
3 6 refinisher: a comparison of exposure estimates utilizing workplace air sampling and blood
37 carboxyhemoglobin measurements. J. Occup. Med. 32: 451-454.
38 Silvers, A.; Florence, B. T.; Rourke, D. L.; Lorimor, R. J. (1994) How children spend their time: a sample survey
39 for use in exposure and risk assessments. Risk Anal. 14: 931-944.
40 Simeone, L. F. (1991) The intrusion of engine exhaust into the passenger areas of recreational power boats.
41 Washington, DC: U.S. Coast Guard, Office of Navigation Safety and Waterway Services; report no.
42 DOT-VNTSC-CG-91-1. Available from: NTIS, Springfield, VA; AD-A254146.
43 Smith, N. J. (1977) End-expired air technic for determining occupational carbon monoxide exposure. J. Occup.
44 Med. 19: 766-769.
45 Smith, W.; Anderson, T.; Anderson, H. A.; Remington, P. L. (1992) Nitrogen dioxide and carbon monoxide
46 intoxication in an indoor ice arena - Wisconsin, 1992. Morb. Mortal. Wkly. Rep. 41: 383-385.
47 Snook, L. M. (1996) An investigation of driver exposure to carbon monoxide while traveling in the wake of a
48 snowmobile [dissertation]. Knoxville, TN: University of Tennessee. Available from: Xerox University
49 Microfilms, Ann Arbor MI; order no. AAD97-09060.
50 Snook, L. M.; Davis, W. T. (1997) An investigation of driver exposure to carbon monoxide while traveling in the
51 wake of a snowmobile. Presented at: 90th annual meeting & exhibition of the Air & Waste Management
52 Association; Toronto, ON, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no.
53 97-RP143.02.
54 Sorensen, A. J. (1986) The importance of monitoring carbon monoxide levels in indoor ice skating rinks. J. Am.
55 Coll. Health 34: 185-186.
October 1999 4-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 Spengler, J. D.; Stone, K. R.; Lilley, F. W. (1978) High carbon monoxide levels measured in enclosed skating
2 rinks. J. Air Pollut. Control Assoc. 28: 776-779.
3 Stewart, R. D.; Hake, C. L. (1976) Paint-remover hazard. JAMA J. Am. Med. Assoc. 235: 398-401.
4 Stewart, R. D.; Fisher, T. N; Hosko, M. J.; Peterson, J. E.; Baretta, E. D.; Dodd, H. C. (1972) Experimental human
5 exposure to methylene chloride. Arch. Environ. Health 25: 342-348.
6 Systems Applications, Inc. (1983) Human exposure to atmospheric concentrations of selected chemicals.
7 Volume II. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
8 Planning and Standards; EPA contract no. 68-02-3066. Available from: NTIS, Springfield, VA;
9 PB83-265249.
10 Szalai, A., ed. (1972) The use of time: daily activities of urban and suburban populations in twelve countries.
11 The Hague, The Netherlands: Mouton & Co.
12 Treitman, R. D.; Burgess, W. A.; Gold, A. (1980) Air contaminants encountered by firefighters. Am. Ind. Hyg.
13 Assoc. 141:796-802.
14 Tsang, A. M.; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
15 activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
16 Research and Development; report no. EPA/600/R-96/148.
17 U.S. Code. (1990) Clean Air Act, as amended by PL 101-549, November 15, 1990. U. S. C. 42: §§7401-7671q.
18 U.S. Department of Commerce. (1983) Statistical abstract of the United States: 1982-83. National data book and
19 guide to sources. 103rd ed. Washington, DC: Bureau of the Census; p. 618.
20 U.S. Department of Transportation. (1996) Transportation and air quality: selected facts and figures. Washington,
21 DC: Federal Highway Administration; FHWA/PD/96-006; p. 6. Available from: NTIS, Springfield, VA;
22 PB97-110613.
23 U.S. Environmental Protection Agency. (1985) Health assessment document for dichloromethane (methylene
24 chloride). Final report. Research Triangle Park, NC: Office of Health and Environmental Assessment,
25 Environmental Criteria and Assessment Office; report no. EPA-600/8-82/004F. Available from: NTIS,
26 Springfield, VA; PB85-191559.
27 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
28 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office;
29 report no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
30 U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
31 monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research
32 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available
33 from: NTIS, Springfield, VA; PB93-157717.
34 U.S. Environmental Protection Agency. (1996) National air pollutant emission trends, 1900-1995. Research
3 5 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-96-007. Available
36 from: NTIS, Springfield, VA; PB97-155329.
37 U.S. Environmental Protection Agency. (1998a) National air quality and emissions trends report, 1996. Research
38 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-97-013. Available
39 from: NTIS, Springfield, VA; PB98-135155. Available: www.epa.gov/oar/aqtrnd96.
40 U.S. Environmental Protection Agency. (1998b) Compilation of air pollutant emission factors. Volume II:
41 Mobile Sources (AP-42), pending 5th edition. Ann Arbor, MI: Office of Mobile Sources. Available:
42 www.epa.gov/omswww/models.htm.
43 Vadarevu, R. V.; Stopher, P. R. (1996) Household activities, life cycle, and role allocation. Trans. Res. Rec.
44 1556:77-85.
45 Van Wijnen, J. H.; Verhoeff, A. P.; Jans, H. W. A.; van Bruggen, M. (1995) The exposure of cyclists, car drivers
46 and pedestrians to traffic-related air pollutants. Int. Arch. Occup. Environ. Health 67: 187-193.
47 Vreman, H. J.; Mahoney, J. J.; Stevenson, D. K. (1993) Electrochemical measurement of carbon monoxide in
48 breath: interference by hydrogen. Atmos. Environ. Part A 27: 2193-2198.
49 Wallace, L. (1979) Use of personal monitor to measure commuter exposure to carbon monoxide in vehicle
50 passenger compartments. Presented at: 72nd annual meeting of the Air Pollution Control Association; June;
51 Cincinnati, OH. Pittsburgh, PA: Air Pollution Control Association; paper no. 79-59.2.
52 Wallace, L. A. (1983) Carbon monoxide in air and breath of employees in an underground office. J. Air Pollut.
53 Control Assoc. 33: 678-682.
54 Webster, H. (1998) Would you like to work at home? Read. Dig. 152: 128-134.
October 1999 4-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wheatley, A. D.; Sadhra, S.; Beach, J. R. (1997) Exposure to toxic gas and particle phase pollutants evolved
2 during deployment of airbags in vehicles. Indoor Environ. 6: 134-139.
3 Wilcosky, T. C.; Simonsen, N. R. (1991) Solvent exposure and cardiovascular disease. Am. J. Ind. Med.
4 19:569-586.
5 Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991a) Study of children's activity
6 patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
7 Available from: NTIS, Springfield, VA; PB94-106903.
8 Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin G. (1991b) Activity
9 patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
10 ARB/R93/487. Available from: NTIS, Springfield, VA; PB94-108719.
11 Wilkinson, T. (1995) Snowed under. Natl. Parks 69: 32-37.
12 Williams, R.; Walsh, D.; White, J.; Jackson, M.; Mumford, J. (1992) Effect on carbon monoxide levels in mobile
13 homes using unvented kerosene heaters for residential heating. Indoor Environ. 1: 272-278.
14 Wilson, A. L.; Colome, S. D.; Tian, Y. (1991) Air toxics microenvironment exposure and monitoring study. Final
15 report. El Monte, CA: South Coast Air Quality Management District; February.
16 Wilson, A. L.; Colome, S. D.; Tian, Y. (1993a) California residential indoor air quality study. Volume 1.
17 Methodology and descriptive statistics. Chicago, IL: Gas Research Institute; report no. GRI-93/0224.1.
18 Available from: NTIS, Springfield, VA; PB94-166048.
19 Wilson, A. L.; Colome, S. D.; Tian, Y. (1993b) California residential indoor air quality study. Volume 1:
20 methodology and descriptive statistics. Appendices. Chicago, IL: Gas Research Institute; report no.
21 GRI-93/0224.2. Available from: NTIS, Springfield, VA; PB94-166055.
22 Yates, M. W. (1967) A preliminary study of carbon monoxide gas in the home. J. Environ. Health 29: 413-420.
23 Yocom, J. E. (1982) Indoor-outdoor air quality relationships: a critical review. J. Air Pollut. Control Assoc.
24 32:500-520.
25 Yocom, J. E.; Clink, W. L.; Cote, W. A. (1971) Indoor/outdoor air quality relationships. J. Air Pollut. Control
26 Assoc. 21:251-259.
27 Yoon, S. S.; Macdonald, S. C.; Parrish, R. G. (1998) Deaths from unintentional carbon monoxide poisoning and
28 potential for prevention with carbon monoxide detectors. JAMA J. Am. Med. Assoc. 279: 685-687.
29 Yu, L. E.; Hildemann, L. M.; Ott, W. R. (1996) A mathematical model for predicting trends in carbon monoxide
30 emissions and exposures on urban arterial highways. J. Air Waste Manage. Assoc. 46: 430-440.
31 Zartarian, V. G.; Ott, W. R.; Duan, N. (1997) A quantitative definition of exposure and related concepts.
32 J. Exposure Anal. Environ. Epidemiol. 7: 411-437.
33 Zayasu, K.; Sekizawa, K.; Okinaga, S.; Yamaya, M.; Ohrui, T.; Sasaki, H. (1997) Increased carbon monoxide in
34 exhaled air of asthmatic patients. Am. J. Respir. Crit. Care Med. 156: 1140-1143.
35 Ziskind, R. A.; Rogozen, M. B.; Carlin, T.; Drago, R. (1981) Carbon monoxide intrusion into sustained-use
36 vehicles. Environ. Int. 5: 109-123.
37 Ziskind, R. A.; Fite, K.; Mage, D. T. (1982) Pilot field study: carbon monoxide exposure monitoring in the general
38 population. Environ. Int. 8: 283-293.
October 1999 4-67 DRAFT-DO NOT QUOTE OR CITE
-------
i 5. PHARMACOKINETICS AND MECHANISMS
2 OF ACTION OF CARBON MONOXIDE
3
4
5 5.1 INTRODUCTION
6 Basic research on the physiology, pharmacokinetics, and toxicology of carbon monoxide
7 (CO) that ended in the late seventies was followed by studies focused primarily on the
8 cardiopulmonary effects of CO as an ambient air pollutant. Although research in this area
9 continues, more recent studies have refocused on the mechanisms of action and
10 pathophysiological effects of CO at a cellular level and on its role as a cytotoxic agent and neural
11 messenger. In this chapter, the sections discussing basic pharmacokinetics draw heavily from
12 Chapter 9 of the previous CO criteria document (U.S. Environmental Protection Agency, 1991).
13 However, all sections were revised and consolidated, many were expanded, and several new
14 sections were added. In particular, sections on tissue production and metabolism of CO and
15 intracellular effects of CO have been revised extensively and expanded. The new section on
16 conditions affecting uptake and elimination of CO discusses the influence of physical activity,
17 altitude, physical characteristics, and health status on carboxyhemoglobin (COHb) formation.
18 Also, new sections on the mechanisms of CO and a review of the developing concepts have been
19 added.
20 Although the focus of this document is on the effects of ambient and near ambient levels of
21 CO leading to low COHb levels (<5%), this chapter discusses, where appropriate, findings of a
22 selected number of human studies carried out at moderate COHb levels (<20%). Also discussed
23 are observations from a limited number of relevant animal studies at even higher COHb levels.
24 The purpose for the inclusion of such observations from human studies at higher CO
25 concentrations, and animal studies in general, is to facilitate the understanding of CO kinetics,
26 related pathophysiologic processes, and mechanisms of cytotoxicity. Furthermore, the animal
27 data at least partially fill the knowledge gaps for which no human data are currently available in
28 these areas of research. Despite some well recognized interspecies differences in the
29 toxicokinetics of CO, the basic mechanisms of CO toxicity between animals and humans are
30 similar and in many respects close to identical (Tyuma et al., 1981; Benignus and Annau, 1994;
October 1999 5-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kimmel et al., 1999). The discussion of interspecies differences as they relate to humans may
2 aid in interpretation of data and elucidation of mechanisms; however, for the material presented
3 in this chapter, it is not essential and is well beyond the scope of this document. Despite
4 interspecies differences, especially in the uptake and elimination kinetics of CO, extrapolation of
5 observations from animals to man as applied in this chapter, even with its many assumptions,
6 may be useful in identifying potential pathophysiologic and histotoxic processes associated with
7 CO exposure.
8
9
10 5.2 ABSORPTION, DISTRIBUTION, AND PULMONARY ELIMINATION
11 5.2.1 Pulmonary Uptake
12 Although CO is not one of the respiratory gases, the similarity of physico-chemical
13 properties of CO and oxygen (O2) permits an extension of the findings of studies on the kinetics
14 of transport of O2 to those of CO. The rate of formation and elimination of COHb, its
15 concentration in blood, and its catabolism is controlled by numerous physical factors and
16 physiological mechanisms. The relative contribution of these mechanisms to the overall COHb
17 kinetics will depend on the environmental conditions, the physical activity of an individual, and
18 many other physiological processes, some of which are complex and still poorly understood
19 (see Section 5.4 for details). All of the pulmonary uptake occurs at the respiratory bronchioles,
20 alveolar ducts, and sacs. The rate of CO uptake depends on the rate of COHb formation. At the
21 low concentration of CO in inhaled air, the rate of uptake and the rate of COHb formation could,
22 for all practical purposes, be considered to be qualitatively the same.
23
24 5.2.1.1 Mass Transfer of Carbon Monoxide
25 The mass transport of CO between the airway opening (mouth and nose) and the red blood
26 cell (RBC) hemoglobin (Hb) is predominantly controlled by physical processes. The CO transfer
27 to the Hb-binding sites is accomplished in two sequential steps: (1) transfer of CO in a gas
28 phase, between the airway opening and the alveoli; and (2) transfer in a "liquid" phase, across the
29 air-blood interface, including the RBC. In the gas phase, the key mechanisms of transport are
30 convective flow, by the mechanical action of the respiratory system, and diffusion in the acinar
October 1999 5-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 zone of the lung (Engel et al., 1973). Subsequent molecular diffusion of CO across the
2 alveolo-capillary membrane along the CO pressure gradient, plasma, and RBC is the virtual
3 mechanism of the liquid phase. The principal transport pathways and body stores of CO are
4 shown in Figure 5-1 (Coburn, 1967).
5
6 5.2.1.2 Effects of Dead Space and Ventilation/Perfusion Ratio
7 The effectiveness of alveolar gas exchange depends on effective gas mixing and matching
8 of ventilation and perfusion. During normal tidal breathing, the inhaled gas is not distributed
9 uniformly across the tracheobronchial tree. With increased inspiratory flow, as during exercise,
10 intrapulmonary gas distribution becomes more uniform, but gas concentration inhomogeneity
11 still will persist. Considering that almost 90% of gas is contained within the acinar zone of the
12 lung, any increase in gas inhomogeneity in this terminal region will have about the same negative
13 effect as an additional increase in the alveolar dead space or a decrease in the alveolo-capillary
14 diffusion capacity (Engel et al., 1973).
15 The inefficiency of gas mixing and a consequent decrease in the effectiveness of alveolar
16 gas exchange is aggravated by ventilation/perfusion (VA/Q) mismatch. Because of the gravity
17 dependence of ventilation and even more of perfusion in an upright posture, regional VA/Q ratios
18 will range from 0.6 (at the base of the lung) to 3.0 (at the apex), the overall value being 0.85.
19 As a result, the VA/Q ratio is the principal variable controlling gas exchange, and any inequalities
20 not only will impair transfer of gases to the blood but also will interfere with unloading of gases
21 from the blood into the alveolar air. In humans, a change in posture to recumbent or horizontal,
22 or exercise will increase the uniformity of VA/Q ratios and promote more efficient gas exchange,
23 whereas increased resting lung volume, increased airway resistance, decreased lung compliance,
24 and, generally, any lung abnormality will aggravate VA/Q ratio inequality.
25 The simplest indicator of the VA/Q ratio inequalities is the volume of physiological dead
26 space (VD), which comprises both the anatomical and alveolar dead space. The alveolar dead
27 space results from reduced perfusion of alveoli, relative to their ventilation (Singleton et al.,
28 1972). Both right-to-left and physiological shunts under normal conditions contribute little to
29 VA/Q inequality (West, 1990a). An increase in tidal volume or respiratory frequency, or both,
October 1999 5-3 DRAFT-DO NOT QUOTE OR CITE
-------
htenoxide -n the Ambient Air
Figure 5-1. Diagrammatic presentation of CO uptake and elimination pathways and
CO body stores.
Source: Adapted from Coburn (1967).
1 will increase moderately to substantially the VD in healthy subjects and in individuals with lung
2 function impairment, respectively (Lifshay et al., 1971).
3
4 5.2.1.3 Lung Diffusion of Carbon Monoxide
5 The next step in the transfer of gases across the alveolar air-Hb barrier is accomplished by
6 gas diffusion which is an entirely passive process. In order to reach the Hb-binding sites, CO and
7 other gas molecules have to pass across the alveolo-capillary membrane, diffuse through the
8 plasma, pass across the RBC membrane, and finally enter the RBC stroma before reaction
9 between CO and Hb can take place. The molecular transfer across the membrane and the blood
10 phase is governed by general physico-chemical laws, particularly by Pick's first law of diffusion
11 (West, 1990b). The exchange and equilibration of gases between the two compartments (air and
12 blood) is very rapid. The dominant driving force is a partial pressure differential of CO across
13 this membrane. For example, inhalation of a bolus of air containing levels of CO above blood
October 1999
5-4
DRAFT-DO NOT QUOTE OR CITE
-------
1 baseline rapidly increases blood COHb by immediate and tight binding of CO to Hb. The
2 rapidity of CO binding to Hb keeps a low partial pressure of CO within the RBC, thus
3 maintaining a high pressure differential between air and blood and consequent diffusion of CO
4 into blood. Subsequent inhalation of CO-free air reverses the gradient (higher CO pressure on
5 the blood side than alveolar air), and CO is released into alveolar air. The air-blood pressure
6 gradient for CO is usually much higher than the blood-air gradient; therefore, the CO uptake will
7 be a proportionately faster process than CO elimination. The rate of CO release also will be
8 affected by metabolic and endogenous production of CO.
9 Diurnal variations in CO diffusion capacity of the lung (DLCO) related to variations in Hb
10 concentration have been reported in normal, healthy subjects (Frey et al., 1987). Others found
11 the changes to be related also to physiological factors such as oxyhemoglobin (O2Hb), COHb,
12 partial pressure of alveolar CO2, ventilatory pattern, O2 consumption, blood flow, functional
13 residual capacity, etc. (Forster, 1987). Diffusion seems to be relatively independent of lung
14 volume within the midrange of vital capacity. However, at extreme volumes, the differences in
15 diffusion rates could be significant; at total lung capacity, diffusion is higher, whereas, at residual
16 volume, it is lower than the average (McClean et al., 1981). In a supine position at rest, DLCO
17 has been shown to be significantly higher than that at rest in a sitting position (McClean et al.,
18 1981). Carbon monoxide diffusion increases with exercise, and, at maximum work rates,
19 diffusion will be maximal regardless of body position. This increase is attained not only by
20 increases in both the diffusing capacity of the alveolar-capillary membrane and the pulmonary
21 capillary blood flow (Stokes et al., 1981) but also by increased V^/Q uniformity (Harf et al.,
22 1978). Under pathologic conditions, where several components of the air-blood interface might
23 be affected severely, and the \/Q ratio inequality also may increase as in emphysema, and
24 fibrosis, or edema, both the uptake and elimination of CO will be affected (Barie et al., 1994).
25
26 5.2.2 Tissue Uptake
27 5.2.2.1 The Lung
28 Although the lung in its function as a transport system for gases is exposed continuously to
29 CO, very little CO actually diffuses into the lung tissue itself (as dissolved CO), except for the
30 alveolar region where it diffuses into blood. The epithelium of the conductive zone
31 (nasopharynx and large airways) presents a significant barrier to diffusion of CO. Therefore,
October 1999 5-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 diffusion and gas uptake by the tissue, even at high CO concentration, will be slow; most of this
2 small amount of CO will be dissolved in the mucosa of the airways. Diffusion into the
3 submucosal layers and interstitium will depend on the concentration of CO and duration of
4 exposure. Experimental exposures of the oronasal cavity in monkeys to very high concentrations
5 of CO for a very short period of time (5 s) increased the blood COHb level to <3.5%.
6 Comparative exposures of the whole lung, however, elevated COHb to almost 60% (Schoenfisch
7 et al., 1980). Thus diffusion of CO across the airway mucosa will contribute little if at all to
8 overall COHb concentration.
9
10 5.2.2.2 The Blood
11 The rate of CO binding with Hb is about 20% slower, and the rate of dissociation from Hb
12 is an order of magnitude slower than are these rates for O2. However, the CO chemical affinity
13 (represented by the Haldane coefficient, M) for Hb is about 218 (210 to 250) times greater than
14 that of O2 (Roughton, 1970; Rodkey et al., 1969). Under steady-state conditions (gas exchange
15 between blood and atmosphere remain constant), one part of CO and 218 parts of O2 would form
16 equal parts of O2Hb and COHb (50% of each), which would be achieved by breathing air
17 containing 21% oxygen and 650 ppm CO. Moreover, the ratio of COHb to O2Hb is proportional
18 to the ratio of their respective partial pressures, PCO and PO2. The relationship between the
19 affinity constant M and PO2 and PCO, first expressed by Haldane (1897-1898), has the following
20 form:
COHb/02Hb = M x (PCO/P02). (5-1)
21
22 At equilibrium, when Hb is maximally saturated by O2 and CO at their respective gas tensions,
23 the M value for all practical purposes is independent of pH, CO2, temperature, and
24 2,3-diphosphoglycerate (Wyman et al., 1982; Gronlund and Garby, 1984).
25 Under dynamic conditions, competitive binding of O2 and CO to Hb is complex; simply
26 said, the greater the number of heme molecules bound to CO, the greater is the affinity of free
27 hemes for O2. However, CO not only occupies O2-binding sites, molecule for molecule, thus
28 reducing the amount of available O2, but also alters the characteristic relationship between O2Hb
29 and PO2, which in normal blood is S-shaped. Figure 5-2 illustrates the basic mechanisms of CO
30 toxicity operating at any CO concentration. The a and a' points represent the arterial values of
October 1999 5-6 DRAFT-DO NOT QUOTE OR CITE
-------
20
100
T3
O
o
o
5
E,
c
CD
x
O
15 H
10 -
S
o5
Q_
(U
E
-^ 5
50% Anemia
(O2 Hb Capacity = 10mL/100mL)
20
40 60
PO2 (mm Hg)
80
100
Figure 5-2. Oxyhemoglobin dissociation curve of normal human blood, of blood
containing 50% COHb, and of blood with only 50% Hb because of anemia.
See the text for additional details.
Source: U.S. Environmental Protection Agency (1991).
1 PO2. The v represents the venous PO2 of healthy individuals after extraction of 5 vol % of O2.
2 With increasing concentration of COHb in blood, the dissociation curve is shifted gradually to
3 the left, and its shape is transformed into a near rectangular hyperbola. Because the shift occurs
October 1999
5-7
DRAFT-DO NOT QUOTE OR CITE
-------
1 over a critical saturation range for release of O2 to tissues, a reduction in O2Hb by CO binding
2 will have more severe effects on the release of O2 than the equivalent reduction in Hb caused by
3 anemia. Thus, in an acute anemia patient (50% of Hb) at a venous PO2 of 26 torr (vj), 5 vol % of
4 O2 (50% desaturation) was extracted from blood, an amount sufficient to sustain tissue
5 metabolism. In contrast, in a person poisoned with CO (50% COHb), the venous PO2 will have
6 to drop to 16 torr (v2; severe hypoxia) to release the same, 5 vol % O2. Any higher demand on
7 O2 under these conditions (e.g., by exercise) might result in brain oxygen depletion and loss of
8 consciousness of the CO-poisoned individual.
9 Because so many cardiopulmonary factors determine COHb formation, the association
10 between COHb concentration in blood and duration of exposure is not linear but S-shaped. With
11 progression of exposure, the initial slower COHb formation gradually accelerates, but, as COHb
12 approaches equilibrium, the build-up slows down again. The S-shape form becomes more
13 pronounced with higher CO levels or with exercise (Benignus et al., 1994; Tikuisis et al., 1992).
14 As Fig. 5-1 shows, CO not only is exchanged between alveolar air and blood but also is
15 distributed by blood to other tissues. Studies on dogs found that over the range of 2 to
16 35% COHb, 77% of total body CO remains in the vascular compartment on the average. The
17 rest of CO diffused to extravascular tissues, primarily skeletal muscle where it is bound to Mb.
18 Compared to dogs, the extravascular CO stores in men are smaller and account for <16% of total
19 body CO. Similarly to animals, no shift between blood and extravascular compartments was
20 found at low <4% COHb (Coburn, 1967; Luomanmaki and Coburn, 1969).
21
22 5.2.2.3 Heart and Skeletal Muscle
23 Myoglobin (Mb), as a respiratory hemoprotein of muscular tissue, will undergo a reversible
24 reaction with CO in a manner similar to O2. Greater affinity of O2 for Mb than Hb (hyperbolic
25 versus S-shaped dissociation curve) is, in this instance, physiologically beneficial because a small
26 drop in tissue PO2 will release a large amount of O2 from O2Mb. The main function of Mb is
27 thought to be a temporary store of O2 and to act as a diffusion facilitator between hemoglobin and
28 the tissues (Peters et al., 1994).
29 Myoglobin has a CO affinity constant approximately eight-times lower than hemoglobin
30 (M = 20 to 40 versus 218, respectively) (Haab and Durand-Arczynska, 1991; Coburn and
October 1999 5-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 Mayers, 1971). As with Hb, the combination velocity constant between CO and Mb is only
2 slightly lower than that for O2, but the dissociation velocity constant is much lower than that for
3 O2. The combination of greater affinity (Mb is 90% saturated at PO2 of 20 mmHg) and lower
4 dissociation velocity constant for CO favors retention of CO in the muscular tissue. Thus, a
5 considerable amount of CO potentially can be stored in the skeletal muscle (Luomanmaki and
6 Coburn, 1969). The binding of CO to Mb (COMb) in heart and skeletal muscle in vivo has been
7 demonstrated at levels of COHb below 2% in heart and 1% in skeletal muscle (Coburn, 1973;
8 Coburn and Mayers, 1971). At rest, the COMb/COHb ratio (0.4 to 1.2) does not increase with an
9 increase in COHb up to 50% saturation and appears to be independent of the duration of
10 exposure (Sokal et al., 1984). During exercise, the relative rate of CO binding increases more for
11 Mb than for Hb, and CO will diffuse from blood to skeletal muscle (Werner and Lindahl, 1980);
12 consequently, the COMb/COHb will increase for both skeletal and cardiac muscles (Sokal et al.,
13 1986). A similar shift in CO has been observed under hypoxic conditions because a fall in
14 myocyte intracellular PO2 below a critical level will increase the relative affinity of Mb to CO
15 (Coburn and Mayers, 1971). Consequent reduction in O2 storage capacity of Mb may have a
16 profound effect on the supply of O2 to the tissue. Apart from Hb and Mb, other hemoproteins
17 will react with CO. However, the exact role of such compounds on O2-CO kinetics still needs to
18 be ascertained. For more discussion on this topic, see Section 5.6.1.
19
20 5.2.2.4 The Brain and Other Tissues
21 The concentration of CO in brain tissue has been found to be about 30- to 50-times lower
22 than that in blood. During the elimination of CO from brain, the above ratio of concentrations
23 was still maintained (Sokal et al., 1984). However, the energy requirement of brain tissue is very
24 high and varies greatly with local metabolism. Because oxygen demand also is coupled to local
25 functional activity, which at times may be very high, and because the brain's oxygen storage is
26 minimal, any degree of hypoxia if uncompensated will have a detrimental effect on brain
27 function. The primary effects of low, ambient concentrations of CO on other organs (e.g., liver,
28 kidney) is via hypoxic mechanisms (see Section 6.6).
29
October 1999 5-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.2.3 Pulmonary and Tissue Elimination
2 An extensive amount of data available on the rate of CO uptake and the formation of COHb
3 contrast sharply with the limited information available on the dynamics of CO washout from
4 body stores and blood. Although almost all of the studies investigating CO elimination pattern
5 and processes involved in modulating the rate of COHb decline in blood were done at moderate
6 COHb levels, the physiologic mechanisms involved in CO elimination kinetics are also effected
7 at lower blood COHb including levels resulting from ambient exposures. The elimination rate of
8 CO from an equilibrium state will follow a monotonically decreasing, second-order (logarithmic
9 or exponential) function (Pace et al., 1950). The rate, however, may not be constant when the
10 steady-state conditions have not yet been reached. Particularly after very short and high CO
11 exposures, it is possible that COHb decline could be biphasic, and it can be approximated best by
12 a double-exponential function; the initial rate of decline or "distribution" might be considerably
13 faster than the later "elimination" phase (Wagner et al., 1975). The reported divergence of the
14 COHb decline rate in blood and in exhaled air suggests that the CO elimination rates from
15 extravascular pools are slower than those reported for blood (Landaw, 1973). Although the
16 absolute elimination rates are associated positively with the initial concentration of COHb, the
17 relative elimination rates appear to be independent of the initial concentration of COHb (Wagner
18 etal., 1975).
19 The same factors that govern CO uptake will affect CO elimination. This suggests that the
20 Coburn-Forster-Kane (CFK) model (see Section 5.5.1) may be suitable to predict CO elimination
21 as well. Surprisingly, few studies tested this application. When breathing air, the CFK model
22 predicted very well the COHb decline. However, at a higher partial pressure of O2 in humidified
23 inspired air (PjC^) or under hyperbaric O2 conditions, the key CFK equation parameters,
24 particularly the DLCO value, must be adjusted for hyperoxic conditions so that CFK will predict
25 more accurately the elimination of CO (Tikuisis, 1996; Tikuisis et al., 1992; Tyuma et al., 1981).
26 The half-time of CO disappearance from blood under normal recovery (air) showed a
27 considerable between-individual variance. For COHb concentrations of 2 to 10%, the half-time
28 ranged from 3 to 5 h (Landaw, 1973); others reported the range to be 2 to 6.5 h for slightly higher
29 initial concentrations of COHb (Peterson and Stewart, 1970). The CO elimination half-time in
30 nonsmokers is considerably longer in men (4.5 h) than in women (3.2 h). During sleep, the
31 elimination rate slowed in both sexes, but, in men, it became almost twice as slow (8.0 h) as
October 1999 5-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 during waking hours. Although no ventilation variables were obtained during the study, the
2 day-to-night differences have been attributed to lower ventilation rates at sleep. The authors
3 speculate that the sex differences in elimination half-time are related to the skeletal muscle mass
4 and intrinsically to the amount of Mb (Deller et al., 1992). The half-time elimination rate
5 appears to be independent of the CO exposure source (e.g., fire, CO intoxication). Normobaric
6 O2 administered to fire victims and CO-poisoned individuals resulted in about the same CO
7 elimination half-time, 91 and 87 min, respectively (Levasseur et al., 1996).
8 Increased inhaled concentrations of O2 accelerated elimination of CO; by breathing
9 100% O2, the half-time was shortened by almost 75% (Peterson and Stewart, 1970). The average
10 half-life of COHb in individuals with very low COHb level (1.16%) breathing hyperbaric O2 was
11 26 min, compared with 71 min when breathing normobaric O2 (Jay and McKindly, 1997). The
12 elevation of PO2 to 3 atm reduced the half-time to about 20 min, which is approximately a
13 14-fold decrease over that seen when breathing room air (Britten and Myers, 1985; Landaw,
14 1973). Although the washout of CO can be somewhat accelerated by an admixture of 5% carbon
15 dioxide (CO2) in O2, hyperbaric O2 treatment is more effective in facilitating displacement of CO.
16 Therefore, hyperbaric oxygen is used as a treatment of choice in CO poisoning. A mathematical
17 model of COHb elimination that takes into account PjO2 has been developed but not yet validated
18 (Singh et al., 1991; Selvakumar et al., 1993).
19
20
21 5.3 TISSUE PRODUCTION AND METABOLISM OF CARBON
22 MONOXIDE
23 In the process of natural degradation of RBC hemoglobin to bile pigments, a carbon atom is
24 separated from the porphyrin nucleus and, subsequently, is catabolized by heme oxygenase (HO)
25 into CO. The major site of heme breakdown and, therefore, the major production organ of
26 endogenous CO is the liver (Berk et al., 1976). The spleen and the erythropoietic system are
27 other important catabolic generators of CO. Because the amount of porphyrin breakdown is
28 stoichiometrically related to the amount of endogenously formed CO, the blood level of COHb or
29 the concentration of CO in the alveolar air have been used with mixed success as quantitative
30 indices of the rate of heme catabolism (Landaw et al., 1970; Solanki et al., 1988). Diurnal
31 variations in endogenous CO production are significant, reaching a maximum around noon and a
October 1999 5-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 minimum around midnight (Levitt et al., 1994; Mercke et al., 1975a). Week-to-week variations
2 of CO production are greater than day-to-day or within-day variations for both males and females
3 (Lynch and Moede, 1972; Mercke et al., 1975b).
4 Any disturbance leading to accelerated destruction of RBCs and accelerated breakdown of
5 other hemoproteins would lead to increased production of CO. Hematomas, intravascular
6 hemolysis of RBCs, blood transfusion, and ineffective erythropoiesis all will elevate COHb
7 concentration in blood. In females, COHb levels fluctuate with the menstrual cycle; the mean
8 rate of CO production in the premenstrual, progesterone phase is almost doubled
9 (Delivoria-Papadopoulos et al., 1974; Mercke and Lundh, 1976). Neonates and pregnant women
10 also showed a significant increase in endogenous CO production related to increased breakdown
11 of RBCs. Degradation of RBCs under pathologic conditions such as anemia (hemolytic,
12 sideroblastic, and sickle cell), thalassemia, Gilbert's syndrome with hemolysis, and other
13 hematological diseases also will accelerate CO production (Berk et al., 1974; Solanki et al.,
14 1988). In patients with hemolytic anemia, the CO production rate was 2- to 8-times higher, and
15 blood COHb concentration was 2- to 3-times higher than in healthy individuals (Coburn et al.,
16 1966). Anemias also may develop under many pathophysiologic conditions characterized by
17 chronic inflammation, such as malignant tumors or chronic infections (Cavallin-Stahl et al.,
18 1976) (see also Section 5.4.3).
19 Not all endogenous CO comes from RBC degradation. Other hemoproteins, such as Mb,
20 cytochromes, peroxidases, and catalase, contribute approximately 20 to 25% to the total amount
21 of CO (Berk et al., 1976). Approximately 0.4 mL/h of CO is formed by Hb catabolism, and
22 about 0.1 mL/h originates from nonhemoglobin sources (Coburn et al., 1963; 1964). This will
23 result in a blood COHb concentration between 0.4 and 0.7% (Coburn et al., 1965).
24 A large variety of drugs will affect endogenous CO production. Generally, any drug that
25 will increase bilirubin production, primarily from the catabolism of Hb, will promote endogenous
26 production of CO. Nicotinic acid (Lundh et al., 1975), allyl-containing compounds (acetamids
27 and barbiturates) (Mercke et al., 1975c), diphenylhydantoin (Coburn, 1970a), progesterone
28 (Delivoria-Papadopoulos et al., 1974), contraceptives (Mercke et al., 1975b) will all elevate
29 tissue bilirubin and, subsequently, CO production.
30 Another mechanism that will increase CO production is a stimulation of HO and
31 subsequent degradation of cytochrome P-450-dependent, mixed-function oxidases. Several types
October 1999 5-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 of compounds such as a carbon disulfide and sulfur-containing chemicals (parathion and
2 phenylthiourea) will act on different moieties of the P-450 system leading to an increase in
3 endogenous CO (Landaw et al., 1970). Other sources of CO involving HO activity include
4 auto-oxidation of phenols, photooxidation of organic compounds and lipid peroxidation of cell
5 membrane lipids (Rodgers et al., 1994). The P-450 system also is involved in oxidative
6 dehalogenation of dihalomethanes, widely used solvents in homes and industry (Kim and Kim,
7 1996). Metabolic degradation of these solvents and other xenobiotics results in the formation of
8 CO that can lead to very high (>10%) COHb levels (Manno et al., 1992; Pankow, 1996).
9 Ascent to high elevations will increase the endogenous level of COHb in both humans and
10 animals (McGrath, 1992; McGrath et al., 1993). The baseline COHb level has been shown to be
11 positively dependent on altitude (McGrath, 1992). Assuming the same endogenous production
12 of CO at altitude as at sea level, the increase in COHb most likely is consequent to a decrease in
13 PO2 (McGrath et al., 1993). Whether other variables, such as an accelerated metabolism or a
14 greater pool of hemoglobin, transient shifts in body stores, or a change in the elimination rate of
15 CO are contributing factors, remains to be explored. Animal studies suggest that the elevated
16 basal COHb production is not a transient phenomenon but persists through a long-term
17 adaptation period (McGrath, 1992).
18 In recent years, new discoveries in molecular biology identified the CO molecule as being
19 involved in many physiological responses, such as smooth muscle relaxation, inhibition of
20 platelet aggregation, and as a neural messenger in the brain (for details, see Sections 5.6 and 5.7).
21 Most recently, several studies reported yet another function of CO, that of a possible marker of
22 inflammation in individuals with upper respiratory tract infection (Yamaya et al., 1998) and
23 bronchiectasis (Horvath et al., 1998a), and in asthmatics (Zayasu et al., 1997; Horvath et al.,
24 1998b). In the Zayasu et al. (1997) study, the investigators found that exhaled concentrations of
25 CO in asthmatics taking corticosteroids were about the same as in healthy individuals (1.7 and
26 1.5 ppm, respectively), whereas, in asthmatics who did not use corticosteroids, the average CO
27 concentration was 5.7 ppm. The authors speculate that one of the anti-inflammatory effects of
28 corticosteroids is the down-regulation of HO. Whether asthmatics have an increased COHb level
29 was not measured in this study or reported in other studies. Patients with chronic inflammatory
30 lung disease, such as bronchiectasis may produce a substantial amount of CO (e.g., 11.8 ppm).
31 As with asthma, induction of heme oxygenase appears to be the primary mechanism involved in
October 1999 5-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 the production of CO (Horvath et al., 1988). Critical illness also seems to be associated with
2 elevated production of CO (Meyer et al., 1998). When compared with controls, ill patients (not
3 characterized) had higher COHb in both arterial and central venous blood not attributable to an
4 elevated inspired concentration of O2 used to treat patients. Moreover, the higher COHb in
5 arterial blood than in central venous blood measured in both ill and control individuals has lead
6 the authors to speculate that a positive arterio-venous COHb difference results from the
7 up-regulation of the inducible isoform of heme oxygenase (HO-1) in the lung and subsequent
8 production of CO (see Section 5.6.4).
9
10
11 5.4 CONDITIONS AFFECTING CARBON MONOXIDE UPTAKE
12 AND ELIMINATION
13 5.4.1 Environment and Activity
14 During exercise, increased demand for O2 requires adjustment of the cardiopulmonary
15 system, so that an increased demand for O2 is met with an adequate supply of O2. Depending on
16 the intensity of exercise, the physiologic changes may range from minimal, involving primarily
17 the respiratory system, to substantial, involving extensively the respiratory, cardiovascular, and
18 other organ systems, inducing local as well as systemic changes. Exercise will improve the VA/Q
19 ratio in the lung, increase the respiratory exchange ratio (RER), increase cardiac output, increase
20 DL^O, mobilize RBC reserves from the spleen and induce other compensatory changes. Heavy
21 exercise will cause a decrease in plasma volume leading to hemoconcentration and a subsequent
22 decrease in blood volume. Of the many mechanisms operating during exercise, the two most
23 important physiologic variables are (1) the alveolar ventilation (VA) and (2) cardiac output.
24 Although some physiologic changes during exercise may impair CO loading into blood (e.g.,
25 relative decrease in DLCO during severe exercise), the majority of the changes will facilitate CO
26 transport. Thus, by increasing gas exchange efficiency, exercise also will promote CO uptake.
27 Consequently, the rate of CO uptake and of COHb formation will be proportional to the intensity
28 of exercise. During a transition period from rest to exercise while exposed to CO
29 (500 ppm/10 min), the diffusing capacity and CO uptake were reported to rise faster than O2
30 consumption for each exercise intensity (Kinker et al., 1992).
October 1999 5-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 Apart from physiological factors, the concentration of CO, as well as the rate of change of
2 CO concentration in an individual's immediate environment, can have a significant impact on
3 COHb. For example, at intersections with idling and accelerating cars, pedestrians will be
4 exposed for a short period of time to higher CO concentrations than those present at other places
5 on the same street. Around home, an individual working with a chain saw, lawnmower, or other
6 gasoline-powered tools will be exposed transiently to higher, and occasionally to much higher
7 (e.g., breathing near the exhaust of a chain saw), concentrations of CO (up to 400 ppm) (Biinger
8 et al., 1997). In indoor environments, exposure to elevated CO from unventilated gas appliances
9 or from environmental tobacco smoke may increase transiently the COHb level of a previously
10 unexposed individual. Occupationally, there are many instances and conditions under which
11 workers may be exposed briefly to moderate-to-high levels of CO from operating equipment or
12 other sources. Despite the shortness of each exposure episode, such transients may result in a
13 relatively high COHb concentration. As an example, exposure for 5 min or less of a resting
14 individual to 7,600 ppm CO in inhaled air will result in almost 20% COHb (Benignus et al.,
15 1994). On repeated brief exposures to high CO, the COHb will increase further until the
16 concentrations in inhaled CO and in blood reach equilibrium. Once the distribution of CO within
17 body stores is complete, the COHb will remain constant, unless the ambient CO concentration
18 changes (either up or down) again. As is the CO uptake, so is the elimination of CO from blood
19 governed by the gas concentration gradient between blood and alveolar air. However, the
20 elimination of CO from blood is a much slower process (see Section 5.2.3) and, therefore, will
21 take many hours of breathing clean air before the baseline COHb value is reached.
22 Recently, a unique source of CO exposure was identified. It has been found repeatedly that
23 the use of volatile anesthetics (fluranes) in closed-circuit anesthetic machines, when CO2
24 absorbent (soda lime) is dry, can result in a significant production of CO caused by a degradation
25 of the anesthetic and subsequent exposure of a patient to CO (up to 7.0% COHb) (Woehlck et al.,
26 1997a,b).
27
28 5.4.2 Altitude
29 Altitude may have a significant influence on the COHb kinetics (U.S. Environmental
30 Protection Agency, 1978). These changes are consequent to compensatory and adaptive
31 physiologic mechanisms. At sea level, at a body temperature of 37 °C, barometric pressure (PB)
October 1999 5-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 of 760 torr, and air (gas) saturated with water vapor (BTPS conditions) the PjO2 is 149 torr.
2 At an altitude of 3,000 m (9,840 ft; PB = 526 torr), the PA is only 100 torr, resulting in an acute
3 hypoxic hypoxia. Direct measurements of blood gases on over 1,000 nonacclimatized
4 individuals at this altitude found the partial pressure of O2 in alveolar air to be only 61 torr
5 (Boothby et al., 1954). The hypoxic drive will trigger a complement of physiological
6 compensatory mechanisms (to maintain O2 transport and supply), the extent of which will depend
7 on elevation, exercise intensity, and the length of a stay at the altitude. During the first several
8 days, the pulmonary ventilation at a given O2 uptake (work level) will increase progressively
9 until a new quasi-steady-state level is achieved (Bender et al., 1987; Burki, 1984). The DLCO
10 will not change substantially at elevations below 2,200 m but was reported to increase above that
11 altitude, and the spirometric lung function (FVC, FEF25.75)will be reduced as well (Ge et al.,
12 1997). The maximal aerobic capacity and total work performance will decrease, and the RER
13 will increase (Horvath et al., 1988). Redistribution of blood from skin to organs and within
14 organs from blood into extravascular compartments, as well as an increase in cardiac output, will
15 promote CO loading (Luomanmaki and Coburn, 1969). Because of a decrease in plasma volume
16 (hemoconcentration), the Fib concentration will be higher than at sea level (Messmer, 1982). The
17 blood electrolytes and acid-base equilibrium will be readjusted, facilitating transport of O2.
18 Thus, for the same CO concentration as at sea level, these compensatory changes will favor CO
19 uptake and COHb formation (Burki, 1984). By the same token, the adaptive changes will affect
20 not only CO uptake but CO elimination as well. Carboxyhemoglobin levels at altitude has been
21 shown to be increased in both laboratory animals and humans (McGrath, 1992; McGrath et al.,
22 1993). Breathing CO (9 ppm) at rest at altitude has produced higher COHb levels than those at
23 sea level (McGrath et al., 1993). Surprisingly, exercise in a CO atmosphere (50 to 150 ppm) at
24 altitude appeared either to suppress COHb formation or to shift the CO storage, or both. The
25 measured COHb levels were lower than those found under similar conditions of exercise and
26 exposure at sea level (Horvath et al., 1988).
27 The short-term acclimatization (within a week or two) will stabilize the compensatory
28 changes. During a prolonged stay at high altitude (over a few months), most of the early adaptive
29 changes gradually will revert to the sea level values, and long-term adaptive changes, such as an
30 increase in tissue capillarity and myoglobin content in the skeletal muscle, begin to take place.
31 Smokers appear to tolerate short-term hypoxic hypoxia caused by high altitude (7,620 m
October 1999 5-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 [25,000 ft]) much better than nonsmokers, who experience more severe subjective symptoms and
2 a greater decline in task performance (Yoneda and Watanabe, 1997). Perhaps smokers, because
3 of chronic hypoxemia (because of chronically elevated COHb), develop partial tolerance to
4 hypoxic hypoxia. Although the mechanisms of COHb formation in hypoxic hypoxia and CO
5 hypoxia are different, the resultant decrease in O2 saturation and activation of compensatory
6 mechanisms (e.g., an increased cerebral blood flow) appear to be at least additive (McGrath,
7 1988). Psychophysiological studies, in particular, seem to support the possibility of
8 physiological equivalency of hypoxic effects, whether induced by altitude at equlibrium or
9 ambient CO concentration. However, it must be remembered that, although some of the
10 mechanisms of action of hypoxic hypoxia and CO hypoxia are the same, CO elicits other toxic
11 effects not necessarily related to O2 transport mechanisms (Ludbrook et al., 1992; Zhu and
12 Weiss, 1994). Recently, Kleinman et al. (1998) demonstrated that the effects of CO and
13 simulated altitude were not synergistic but additive.
14
15 5.4.3 Physical Characteristics
16 Physical characteristics (e.g., sex, age, race, pregnancy) are not known to directly modify
17 the basic mechanisms of CO uptake and COHb formation and elimination. However, the
18 baseline values of many cardiopulmonary variables that may influence COHb kinetics are known
19 to change with physical characteristics.
20 The CO uptake and elimination rates either at rest or exercise decrease with age. During
21 the growing years (2 to 16 years of age), the COHb elimination half-time increases rapidly with
22 age in both sexes and is relatively shorter for boys than for girls. Beyond teenage years, the
23 half-time for CO elimination continues to grow longer but at a lower rate. In contrast to the
24 adolescent period, the COHb half-life during the adult years was found to be persistently shorter
25 (-6%) in females than that in males (Joumard et al., 1981). Furthermore, it has been well
26 established that, with age, the DLCO, one of the key determinants, decreases (Guenard and
27 Marthan, 1996). The rate of DLCO decline is lower in middle-aged women than it is in men,
28 however, at older ages, the rates evened and are about the same for both sexes (Neas and
29 Schwartz, 1996). The decrease in DLCO, combined with an increase in VA/Q mismatch, which
30 increases with age, means that it will take longer to both load and eliminate CO from blood.
October 1999 5-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 In pregnancy, increased requirement for iron may lead to iron deficiency and anemia
2 (for further details see Section 5.4.3). Pregnant women who smoked showed a more pronounced
3 shift of the O2 dissociation curve to the left (-5% COHb) than one would expect from the same
4 COHb concentration in nonpregnant women. Thus, increased O2 affinity, combined with
5 decreased O2-carrying capacity of blood of CO-exposed women, may promote fetal hypoxia
6 (Grote et al., 1994). Animal studies found that protein deficiency in pregnant mice had no
7 modulating effect on maternal COHb but resulted in a greater concentration of placental COHb
8 (Singh et al., 1993; Singh and Moore-Cheatum, 1993; Singh et al., 1992).
9 Young women were found to be more resistant to altitude hypoxia than were men, but the
10 physiological factors for this difference remain unexplored (Horvath et al., 1988).
11 Carboxyhemoglobin levels, although elevated at altitude, were found to be about the same for
12 both males and females (McGrath et al., 1993).
13 Whether the dynamics of COHb formation and elimination or the absolute COHb levels for
14 the same exposure conditions are different in any way between races have not been studied.
15 Blacks have lower diffusion capacity than whites (Neas and Schwartz, 1996), which transiently
16 will slow CO loading and unloading. It also is well documented that the black population has a
17 higher incidence of sickle cell anemia, which may be a risk factor for CO hypoxia (see
18 Section 5.4.4 below).
19
20 5.4.4 Health Status
21 An individual with any pathophysiologic condition that reduces the blood O2 content will
22 be at a greater risk from CO exposure because additional reduction in the O2-carrying capacity of
23 blood resulting from COHb formation will increase hypoxemia. Depending on the severity of
24 initial hypoxia, exposure to CO may lower the O2 content to the point where O2 delivery to the
25 tissues becomes insufficient.
26 One group of disorders that encompasses a range of etiologically varied diseases
27 characterized by a reduction in total blood Hb and subsequent insufficiency to meet O2 demands
28 are the anemias. Anemia is a result of either impaired formation of RBCs or increased loss or
29 destruction of RBCs. The former category includes disorders of altered O2 affinity,
30 methemoglobinemias, and diseases with functionally abnormal and unstable hemoglobins.
31 By far, the most prevalent disorder in this group is a single-point mutation of Hb (Hb S), causing
October 1999 5-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 sickle cell diseases, the most typical of which is a sickle cell anemia. The O2-carrying capacity of
2 individuals afflicted with sickle cell anemia is reduced not only because of a smaller amount of
3 Hb, but also the O2 dissociation curve is shifted to the right, reducing the O2 affinity as well.
4 Initial compensation involves primarily the cardiovascular system. The cardiac output will
5 increase as both heart rate and stroke volume increase.
6 The opposite condition of anemia is polycythemia, an increased number of RBCs in blood.
7 Although in polycythemia the total amount of hemoglobin generally is elevated, under certain
8 conditions the arterial O2 saturation may be decreased, leading to a higher risk of additional
9 hypoxia when exposed to CO (Foster et al., 1978; Stork et al., 1988).
10 One of the characteristic symptoms of chronic obstructive pulmonary disease (COPD) is
11 increased VD and VA/Q inequality (Marthan et al., 1985). Subsequently, impaired gas mixing
12 because of poorly ventilated lung zones will result in decreased arterial O2 saturation and
13 hypoxemia. These pathophysiologic conditions will slow both CO uptake and elimination.
14 Any COHb formation will further lower the O2 content of blood and increase hypoxemia.
15 Because COPD patients very often operate at the limit of their O2 transport capability, exposure
16 to CO may compromise severely tissue oxygenation.
17 Because O2 extraction by the myocardium is high, a greater O2 demand by the myocardium
18 of healthy individuals is met by an increased coronary blood flow. Patients with coronary artery
19 disease (CAD) have a limited ability to increase coronary blood flow in response to increased
20 O2 demand during physical activity. If this compensatory mechanism is further compromised by
21 decreased O2 saturation from CO inhalation, the physical activity of patients with CAD may be
22 restricted severely consequent to more rapid development of myocardial ischemia.
23 Individuals with congestive heart failure, right-to-left shunt in congenital heart disease, or
24 cerebrovascular disease also may be at a greater risk from CO exposure because of already
25 compromised O2 delivery.
26
27
October 1999 5-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.5 MODELING CARBOXYHEMOGLOBIN FORMATION
2 5.5.1 The Coburn-Forster-Kane and Other Regression Models
3 5.5.1.1 Empirical Regression Models
4 The most direct approach to establishing a prediction equation for COHb is to regress
5 observed COHb values against the concentration and duration of exogenous CO exposure.
6 Inclusion of other predictors such as initial COHb level and VA generally will improve the
7 precision of the predictions. Most of the CO regression models are purely empirical and have no
8 physiological basis. Their applicability therefore is limited to more or less exact conditions that
9 were used to collect the data on which they are based.
10 Peterson and Stewart (1970) developed a regression equation for estimating percent COHb
11 following a 15-min to 8-h exposure of resting nonsmokers to moderate levels of CO (25 to
12 523 ppm):
13
14 Log % COHb = 0.858 Log CO + 0.630 Log t- 0.00094 t'- 2.295, (5-2)
15
16 where CO refers to the concentration of CO in inhaled ambient air in parts per million, t is the
17 exposure duration in minutes, and t' is the postexposure time in minutes (set to 0 until the end of
18 exposure). Data from a subsequent study were used to derive a new empirical formula for much
19 higher concentrations of CO (1,000 to 35,600 ppm) but much shorter exposure times (45 s to
20 10 min) (Stewart et al., 1973). These equations still are used occasionally in field conditions to
21 quickly estimate COHb concentration.
22 To predict changes in COHb as a function of ambient CO concentration in an urban setting,
23 Ott and Mage (1978) developed a linear differential equation where only ambient CO
24 concentration varied with time. All other parameters were empirically derived constants. With
25 this simple model, they were able to show that the presence of CO spikes in data averaged over
26 hourly intervals may lead to underestimating the COHb concentration by as much as 21% of the
27 true value. Consequently, they recommended that monitored CO be averaged over 10 to 15 min
28 periods. Based on a similar approach, other empirical models have been developed but not
29 validated (Chung, 1988; Forbes et al., 1945). Comparison of predicted COHb values by these
30 two models revealed a progressive divergence of the estimated COHb curves between models as
October 1999 5-20 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
exposure (100 ppm) progressed, with absolute differences approaching almost 7% COHb. Such
wide variations in predicted COHb best demonstrate the inaccuracy of these types of models
when applied outside of a narrowly defined range and question their utility in practical
applications (Tikuisis, 1996).
Several mathematical models have been developed to predict COHb as a function of
exposure time (Singh et al., 1991; Sharan et al., 1990) or altitude (Selvakumar et al., 1992). The
physiological variables used by Peterson and Stewart (1970) were employed to test these models
and compare the results to the CFK predictions. The agreement between predicted COHb values
by these models and the CFK model was very good; however, the models have not been
validated by experimental studies.
5.5.1.2 Linear and Nonlinear Coburn-Forster-Kane Differential Equations
In 1965, Coburn, Forster, and Kane developed a differential equation to describe the major
physical and physiological variables that determine the concentration of COHb in blood for the
examination of the endogenous production of CO. The equation, referred to as the CFK model,
either in its original form or adapted to special conditions is still much in use today for the
prediction of COHb consequent to inhalation of CO. Equation 5-3 represents the linear CFK
model that assumes O9Hb is constant:
d[COHb]t
dt
= Vco-
[COHb]0P502
[02Hb]M
1
1 1
DL v
V L v A )
+
P,CO
1 1
DL V
V L v A )
(5-3)
where Vb is blood volume in milliliters; [COHb], is the COHb concentration at time t in
milliliters CO per milliliter blood, standard temperature and pressure, dry (STPD); Vco is the
endogenous CO production rate in milliliters per minute, STPD; [COHb]0 is the COHb
concentration at time zero in milliliters CO per milliliter blood, STPD; [O2Hb] is the
oxyhemoglobin concentration in milliliters O2 per milliliter blood, STPD; PcO2 is the average
partial pressure of O2 in lung capillaries in millimeters of mercury; VA is the alveolar ventilation
in ml/min, STPD; DL is the lung diffusing capacity of CO in ml/min/mmHg, STPD; and PjCO is
October 1999
5-21
DRAFT-DO NOT QUOTE OR CITE
-------
1 the CO partial pressure in inhaled air in millimeters of mercury. The model also assumes an
2 instant equilibration of gases in the lung, COHb concentration between venous and arterial
3 blood, and COHb concentrations between blood and extravascular tissues. Because O2 and CO
4 combine with Hb from the same pool, higher COHb values do affect the amount of Hb available
5 for bonding with O2. Such interdependence can be modeled by substituting (1.38 Hb- [COHb])
6 for [O2Hb], where Hb refers to the number of grams of Hb per milliliter of blood (Tikuisis et al.,
7 1987a). The CFK differential equation (Equation 5-3) then becomes nonlinear:
d[COHb]t Vco 1 [COHb]0Pc-02
dt Vb Vb/3l [02Hb]M '
9
10 where P is (1/DL) + (PB - 47)/VA, and PB is the barometric pressure in millimeters of mercury.
1 1 The nonlinear CFK model is more accurate physiologically but has no explicit solution.
12 Therefore, interactive or numerical integration methods must be used to solve the equation
13 (Muller and Barton, 1987; Johnson et al., 1992). One of the requirements of the method is that
14 the volumes of all gases be adjusted to the same conditions (e.g., STPD) (Muller and Barton,
15 1987; Tikuisis et al., 1987a,b).
16 In general, the linear CFK equation is a better approximation to the nonlinear equation
17 during the uptake of CO than during the elimination of CO. As long as the linear CFK equation
18 is used to predict COHb levels at or below 6% COHb, the solution to the nonlinear CFK model
19 will deviate no more than ±0.5% COHb (Smith, 1990). Over the years, it has been empirically
20 determined that minute ventilation and the DLCO have the greatest influence on the CO uptake
21 and elimination. The relative importance of other physiologic variables will vary with exposure
22 conditions and health status. A comprehensive evaluation of fractional sensitivities of
23 physiologic variables for both the linear and nonlinear CFK equation shows that each variable
24 will exert its maximal influence at different times of exposure (McCartney, 1990). The analysis
25 found that only the fractional concentration of CO in inhaled air, in parts per million (FjCO) and
26 Vco will not affect the rate at which equilibrium is reached. Figure 5-3 illustrates the temporal
27 changes in fractional sensitivities of the principal determinants of CO uptake for the linear form
28 of the CFK equation; THb is the total blood concentration of hemoglobin. The fractional
October 1999 5-22 DRAFT-DO NOT QUOTE OR CITE
-------
1.00
1.00
(D
CD
CD -0-50 H
-0.75-
-1.00
-0.25-
0 1
Log Time (min)
1.7h
2
17h
3
-1.00
Figure 5-3. Plot of fractional sensitivities of selected variables versus time of exposure (see
text for details).
Source: Modified from McCartney (1990).
1
2
3
4
5
6
1
sensitivity of unity means that, for example, a 5% error in the selected variable induces a 5%
error in predicted COHb value by the nonlinear model.
5.5.1.3 Confirmation Studies of Coburn-Forster-Kane Models
Since the publication of the original paper (Coburn et al., 1965), several investigators have
tested the fit of both the linear and nonlinear CFK model to experimental data using different CO
exposure profiles, a variety of experimental conditions and different approaches to evaluating the
October 1999
5-23
DRAFT-DO NOT QUOTE OR CITE
-------
1 parameters of the model. In all of these studies, almost all of the physiologic coefficients either
2 were assumed or estimated based on each individual's physical characteristics; the COHb values
3 were both measured directly and calculated for each individual.
4 Stewart et al. (1970) tested the CFK linear differential equation on 18 resting, healthy
5 subjects exposed to 25 different CO exposure profiles for periods of 0.5 to 24 h and to CO
6 concentrations ranging from 1 to 1,000 ppm. In a later study, they tested the nonlinear CFK
7 equation on 22 subjects at various levels of exercise while being exposed to up to 200 ppm CO
8 for up to 5.25 h (Peterson and Stewart, 1975). From the obtained values, they concluded that
9 either at rest or exercise the agreement between the predicted and measured COHb values was
10 good (correlation coefficient [r] > 0.74).
11 The first study to test both the linear and nonlinear CFK models for CO uptake and
12 elimination in pedestrians and car passengers exposed to ambient CO levels in a city was
13 conducted by Joumard et al. (1981). The two cohorts exposed for 2 h to street and traffic
14 concentrations of CO, respectively, comprised 73 nonsmokers (18 to 60 years of age). Blood
15 COHb samples were taken only at the beginning and the end of each j ourney, where the COHb
16 value reached 2.7%, on average. Both equations performed well in estimating accurately COHb
17 levels, although the value for males was underestimated slightly.
18 The predictive strength of the CFK model under variable CO concentrations was tested by
19 Hauck and Neuberger (1984). A series of experiments was performed on four subjects exposed
20 to a total of 10 different CO exposure profiles at several exercise levels, so that each exposure
21 was a unique combination of CO concentration and exercise pattern. The ventilation and COHb
22 values (measured and calculated) were obtained at 1-min intervals. The agreement between
23 measured and predicted COHb under these varied conditions was very good; the mean difference
24 was only 7.4% of the nominal (maximal predicted) value.
25 A series of studies has tested the accuracy of the CFK equation under transient exposure
26 conditions that would violate several assumptions of the CFK model, specifically the assumption
27 of a single, well-mixed vascular compartment. These studies were designed to simulate everyday
28 conditions (e.g., environmental, occupational, military) that may involve frequent but brief
29 (75 s to 5 min) exposures to high (667 to 7,500 ppm) CO concentrations at rest and exercise.
30 Moreover, the experiments were designed to test the accuracy of the CFK equation under
31 transient exposure conditions during the CO uptake and early elimination phase from arterial and
October 1999 5-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 venous blood. Attempts were made to measure directly some of the key physiologic parameters
2 of the CFK equation for each subject (Tikuisis et al., 1987a,b; Benignus et al., 1994). The
3 studies have shown that during and immediately following exposure, the arterial COHb was
4 considerably higher (1.5 to 6.1%), and the venous COHb was considerably lower (0.8 to 6%)
5 than the predicted COHb. The observed COHb differences between arterial and venous blood
6 ranged from 2.3 to 12.1% COHb among individuals (Benignus et al., 1994). The overprediction
7 of venous COHb increased during exercise (-10% of the true value). Provided that the total CO
8 dose (concentration x time) is the same and within the time constant for the CO uptake and
9 elimination, the COHb value was found to be the same, regardless of the pattern of exposure.
10 Because VA affects both the equilibrium and the rate at which it is achieved, inconsistencies in
11 the estimates or conversion of gas volumes (ATPS and BTPS to STPD) will affect the predicted
12 values. The interindividual and intraindividual disparities between measured and predicted
13 COHb values were attributed primarily to delays in mixing of arterial and venous blood and
14 differences in cardiac output; but, other factors such as lung wash-in also contribute to this
15 phenomenon. Modification of the CFK equation by adjusting for regional differences in blood
16 flow produced a model that predicted with much greater accuracy both the arterial (<0.7% COHb
17 difference) and venous (<1% COHb difference) COHb during transient uptake and elimination of
18 CO from blood (Smith et al., 1994).
19 Although the CO concentrations used in these studies are several orders of magnitude
20 higher than the usual CO concentrations found in ambient air, under certain conditions (see
21 Section 5.4.1) people can be exposed briefly (<10 min) to such (or even higher) levels of CO in
22 their immediate environment. Because the physiologic mechanisms (but not the kinetics) of
23 COHb formation are independent of CO concentration, high COHb transients, particularly in
24 at-risk individuals, could be of clinical importance. Even briefly, higher arterial COHb may lead
25 to functional impairment of the hypoxia-sensitive brain and heart (see Sections 5.2.2.3 and
26 5.2.2.4). In these situations, the predicted instantaneous arterial COHb level will be substantially
27 underestimated.
28
29 5.5.1.4 Application of Coburn-Forster-Kane Models
30 To obviate measurements of CFK equation parameters, many of which are complex
31 techniques, attempts were made to simplify the CFK equation, because it may be difficult or even
October 1999 5-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 impossible to measure directly some of these parameters, particularly during physical activity.
2 In one study, by relating physiological parameters to the O2 uptake by the body, which was in
3 turn related to an activity level, a simplified linear form of the CFK model was developed
4 (Bernard and Duker, 1981). The model was used subsequently to draw simple nomograms of
5 predictive relationship between pairs of variables, but the accuracy of the nomograms was not
6 experimentally tested.
7 The need for more accurate COHb prediction under more complex physiologic or exposure
8 conditions requires either modification or expansion of the CFK model. Benignus (1995)
9 combined a physiological model of respiratory gas exchange, MACPUF (Ingram et al., 1987),
10 with the CFK model. The new model allows for continuous output and input of
11 60 cardiopulmonary variables, including FjCO. The usefulness of the model is particularly in its
12 ability to continuously update COHb concentration in response to dynamically changing
13 physiologic parameters. The model also allows COHb prediction under conditions that otherwise
14 would be very difficult to duplicate in the laboratory.
15 A fundamental modification of the CFK model was made by Hill et al. (1977) to study the
16 effects of CO inspired by the mother on the level of fetal COHb. The Hill equation combines the
17 CFK equation (for maternal COHb), with a term denoting COHb transfer from a placenta into the
18 fetus. Comparative evaluation of predicted and measured fetal COHb concentrations under
19 time-varying and steady-state conditions in both women and animals showed acceptable
20 agreement only under steady-state conditions (Hill et al., 1977; Longo and Hill, 1977).
21 As mentioned in Section 5.5.1.3, Smith et al. (1994) expanded the CFK model to allow for
22 prediction of arterial and venous COHb during a transient CO uptake and early elimination
23 phase. The model incorporated regional differences in blood flow, particularly in the forearm,
24 because the forearm is used most frequently for blood sampling. This more elaborate model
25 performed extremely well in predicting blood COHb. Although the model was validated on a
26 small number of subjects using the same experimental setting, the validation was not performed
27 under more demanding conditions of physical activity and varying CO concentrations.
28 To accurately predict COHb in individuals exposed to dihalomethanes, which are a source
29 of endogenous CO (see Section 5.3), the CFK model was extended to account for the CO
30 production caused by oxidation of a parent chemical (Andersen et al., 1991). The model
31 developed and validated on rats employed a variety of exposure scenarios to dichloromethane.
October 1999 5-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 It subsequently was tested on six volunteers exposed to dichloromethane, where, after adjustment
2 of a few parameters, the COHb level was predicted remarkably well. After further validation,
3 this model has potential use in predicting accurately COHb caused by exogenous and endogenous
4 CO originating from different sources (e.g., Hb degradation, metabolism of dihalomethanes,
5 inhaled CO).
6 Reexpression of the solution of the CFK model from units of percent COHb to parts per
7 million of CO allows the examination of a variety of CO concentration profiles, while keeping a
8 simple preselected target COHb as a constant. Application of the transformed model to urban
9 hourly averaged CO concentrations that just attained alternative 1-h and 8-h CO NAAQS showed
10 that, depending on the air quality pattern used, between 0.01 to 10% of the population may
11 exceed a target 2.1% COHb level in blood without ambient CO concentrations ever exceeding
12 the standard. By including transients, the models predicted COHb more accurately, particularly
13 when built into the 8-h running averages (Venkatram and Louch, 1979; Biller and Richmond,
14 1982, 1992). Actually, the ambient CO concentrations could be averaged over any time period
15 less than or equal to the half-life of COHb (Saltzman and Fox, 1986).
16
17
18 5.6 INTRACELLULAR EFFECTS OF CARBON MONOXIDE
19 5.6.1 Introduction
20 Traditional concepts for CO pathophysiology have been based on the high affinity of CO
21 for deoxyhemoglobin and consequent reduction of O2 delivery. This mechanism is relevant for
22 high CO concentrations, but it is less likely to be relevant to the concentrations of CO currently
23 found in the ambient environment. This section will summarize recently published information
24 on biochemical mechanisms that is not related to an impairment of oxygen delivery from
25 elevations in COHb. Some of the studies outlined in this section were done with cells in culture
26 and others with laboratory rats. To be relevant to human exposures from environmental
27 contamination, it is important to note what concentrations of CO are likely to occur in vivo.
28 Lung parenchyma represents a special situation where cells may be exposed to ambient CO
29 without the reduction in concentration associated with hemoglobin-bound CO. Elsewhere in the
30 body, only a fraction of COHb will dissociate to elevate extravascular CO concentrations. This
October 1999 5-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 elevation is in the range of approximately 2 to 10 nM when the COHb concentration is from
2 0.8 to 3.8% (Coburn, 1970b; Gothert, 1970). The COHb values near steady state conditions in
3 laboratory rats are close to values for humans (Kimmell, 1999). This strengthens the potential
4 for human relevance in recent animal studies showing that newly identified biochemical
5 mechanisms do have adverse physiological effects. However, caution still is warranted because
6 direct evidence for the occurrence of these mechanisms in humans has not been shown.
7
8 5.6.2 Inhibition of Hemoprotein Function
9 Carbon monoxide can inhibit a number of hemoproteins found in cells, such as myoglobin,
10 cytochrome c oxidase, cytochrome P-450, dopamine P hydroxylase, and tryptophan oxygenase
11 (Coburn and Forman, 1987). Inhibition of these enzymes could have adverse effects on cell
12 function.
13 Carbon monoxide acts as a competitive inhibitor, hence biological effects depend on the
14 partial pressures of both CO and O2. The cellular hemoprotein with the highest relative affinity
15 for CO over that for O2 is myoglobin. Carbon monoxide will inhibit myoglobin-facilitated
16 oxygen diffusion, but physiological compromise is seen only with high concentrations of COMb.
17 Wittenberg and Wittenberg (1993) found that high-energy phosphate production was inhibited in
18 isolated cardiac myocytes, maintained at a physiologically relevant oxygen concentration, when
19 COMb exceeded 40%. The authors estimated that formation of sufficient COMb to impair
20 oxidative phosphorylation in vivo would require a COHb level of 20 to 40%.
21 Coefficients for binding CO versus O2 among cytochrome P-450-like proteins vary between
22 0.1 and approximately 12, and there have been recent discussions suggesting that CO-mediated
23 inhibition of these proteins could cause smooth muscle relaxation in vivo (Coburn and Forman,
24 1987; Wang et al., 1997a; Wang, 1998). The issue relates to inhibition of cytochrome
25 P-450-dependent synthesis of several potent vasoconstricting agents (Wang, 1998). Vasodilation
26 has been shown via this mechanism with high concentrations of CO (ca. 90,000 ppm) (Coceani
27 et al., 1988). It is unclear, however, whether this could arise under physiological conditions and
28 CO concentrations produced endogenously. The competition between CO and O2 for
29 cytochrome c oxidase was well outlined in the previous review (U.S. Environmental Protection
30 Agency, 1991), but some additional information has been published. Based on its Warburg
31 partition coefficient of between 5 and 15, CO binding is favored only in situations where oxygen
October 1999 5-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 tension is extremely low (Coburn and Forman, 1987). Carbon monoxide binding to cytochrome
2 c oxidase in vivo will occur when COHb is high (ca. 50%), a level that causes both systemic
3 hypotension as well as impaired oxygen delivery (Brown and Piantadosi, 1992). Mitochondrial
4 dysfunction, possibly linked to cytochrome inhibition, has been shown to inhibit energy
5 production, and it also may be related to enhanced free radical production (Piantadosi et al.,
6 1995; 1997a). There has been no new information published since the last air quality criteria
7 document that pertains to the effects of CO on dopamine P hydroxylase or tryptophan oxygenase.
8
9 5.6.3 Free Radical Production
10 Laboratory animal studies indicate that nitrogen- and oxygen-based free radicals are
11 generated in vivo during CO exposures. Exposure to CO at concentrations of 20 ppm or more
12 for 1 h will cause platelets to become a source of the nitric oxide free radical ('NO) in the
13 systemic circulation of rats (Thorn et al., 1994, Thorn and Ischiropoulos, 1997). Studies with
14 cultured bovine pulmonary endothelial cells have demonstrated that exposures to CO at
15 concentrations as low as 20 ppm cause cells to release 'NO, and the exposure will cause death by
16 a 'NO process that is manifested 18 to 24 h after the CO exposure (Thorn et al., 1997; Thorn and
17 Ischiropoulos, 1997). The mechanism is based on elevations in steady-state 'NO concentration
18 and production of peroxynitrite (Thorn et al., 1994; Thorn et al., 1997). Peroxynitrite is a
19 relatively long-lived, strong oxidant that is produced by the near diffusion-limited reaction
20 between 'NO and superoxide radical (Huie and Padjama, 1993).
21 The mechanism by which CO concentrations of 11 nM or more cause an elevation of
22 steady-state 'NO concentrations appears to be based on altered intracellular "routing" of'NO in
23 endothelial cells and platelets. It is well established that the association and dissociation rate
24 constants of'NO with hemoproteins exceed the rate constants for O2 or CO (Gibson et al., 1986).
25 However, Moore and Gibson (1976) found that when CO was incubated with nitrosyl
26 (•NO)-myoglobin or 'NO-hemoglobin, CO slowly displaced the 'NO. Carbon monoxide
27 replacement occurred even when there was excess 'NO-heme protein, and replacement rates were
28 enhanced by increasing the CO concentration or by carrying out the reaction in the presence of
29 agents such as thiols, which will react with the liberated 'NO. These conditions, including the
30 presence of thiols, exist in cells exposed to environmentally relevant concentrations of CO.
31 Exposures to up to 1,070 nM CO do not alter the rate of production of'NO by platelets and
October 1999 5-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 endothelial cells, yet liberation of'NO was enhanced by CO (Thorn and Ischiropoulos, 1997;
2 Thorn et al., 1994; Thorn et al., 1997).
3 Carbon monoxide will increase the concentration of'NO available to react with in vivo
4 targets in both lung and brain, based on electron paramagnetic resonance studies with rats
5 exposed to 50 ppm CO or more (Ischiropoulos et al., 1996; Thorn et al., 1999a). The
6 concentrations of nitric oxide synthase isoforms in lung were not altered because of CO, and the
7 mechanism for elevation in 'NO was thought to be the same as that found in isolated cells (Thorn
8 et al., 1994; 1997). Exposure to 50 to 100 ppm CO also will increase hydrogen peroxide (H2O2)
9 production in lungs of rats (Thorn et al., 1999a). The phenomenon depended on 'NO production,
10 as it was inhibited in rats pretreated with Nwnitro-L-arginine methyl ester (L-NAME), a nitric
11 oxide synthase inhibitor. Production of'NO-derived oxidants also is increased in CO-exposed
12 rats, based on measurements of nitrotyrosine, a major product of the reaction of peroxynitrite
13 with proteins (Ischiropoulos et al., 1996; Thorn et al., 1998, 1999a,b).
14 The mechanism for enhanced H2O2 production in lungs of CO-exposed rats is not clear.
15 It is possible that 'NO or peroxynitrite may perturb mitochondrial function. Peroxynitrite inhibits
16 electron transport at complexes I through HI, and 'NO targets cytochrome oxidase (Cassina and
17 Radi, 1996; Lizasoain et al., 1996; Poderoso et al., 1996). It is important to note, however, that
18 alterations in mitochondrial function and an increase of cellular H2O2 were not found in studies
19 where cultured bovine endothelial cells were exposed to similar CO concentrations (Thorn et al.,
20 1997). An alternative possible mechanism to mitochondrial dysfunction is that exposure to CO
21 may inhibit antioxidant defenses. Mechanisms linked to elevations in 'NO could be responsible
22 for inhibiting one or more enzymes. Nitric oxide-derived oxidants can inhibit manganese
23 superoxide dismutase and glyceraldehyde-3-phosphate dehydrogenase and deplete cellular stores
24 of reduced glutathione (Ischiropoulos et al., 1992; Luperchio et al., 1996).
25 Exposure to high CO concentrations (2,500 to 10,000 ppm) cause mitochondria in brain
26 cells to generate hydroxyl-like radicals (Piantadosi et al., 1995; 1997a). An additional source of
27 partially reduced O2 species found in animals exposed to CO is xanthine oxidase. Conversion of
28 xanthine dehydrogenase, the enzyme normally involved with uric acid metabolism, to xanthine
29 oxidase, the radical-producing form of the enzyme, occurred in the brains of rats exposed to
30 approximately 3,000 ppm CO (Thorn, 1992). Lower CO concentrations did not trigger this
31 change. Therefore, xanthine oxidase is unlikely to be a free radical source following exposures
October 1999 5-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 to CO at concentrations found in ambient air. Moreover, enzyme conversion was not a primary
2 effect of CO. Rather, it only occurred following sequestration and activation of circulating
3 leukocytes (Thorn, 1993).
4
5 5.6.4 Stimulation of Guanylate Cyclase
6 In recent years, CO has been demonstrated to play a physiological role in vasomotor control
7 and neuronal signal transduction (Morita et al., 1995; Ingi et al., 1996). Carbon monoxide is
8 produced endogenously by oxidation of organic molecules, but the predominant source is from
9 the degradation of heme (Rodgers et al., 1994). The rate-limiting enzyme for heme metabolism
10 is heme oxygenase (HO), which converts heme to biliverdin, free iron, and CO. Three isoforms
11 of HO have been characterized. The HO-1 is an inducible enzyme found in vascular endothelial
12 cells, smooth muscle cells, bronchoalveolar epithelium, and pulmonary macrophages. The HO-1
13 is induced by its substrate, heme, as well as 'NO, H2O2, several cytokines, and lipopolysaccharide
14 (Arias-Diaz et al., 1995; Durante et al., 1997; Morita et al., 1995; Motterlini et al., 1996). The
15 HO-2 is a constitutive enzyme found in certain neurons within the central nervous system,
16 testicular cells, and vascular smooth muscle cells (Marks, 1994). Little is known about HO-3,
17 which was recently identified in homogenates from a number of organs (McCoubrey et al.,
18 1997).
19 A main physiological role for CO is thought to be regulation of soluble guanylate cyclase
20 activity. Both CO and 'NO can activate guanylate cyclase, although activation by CO is
21 approximately 30-fold lower (Stone and Marietta, 1994). In neuronal cells possessing both heme
22 oxygenase and nitric oxide synthase, regulation of cyclic guanosine monophosphate (cGMP)
23 synthesis is mediated in a reciprocal fashion by producing either CO or 'NO (Ingi et al., 1996;
24 Maines et al., 1993). A compensatory interrelationship between nitric oxide synthase and heme
25 oxygenase also has been found in endothelial cells and activated macrophages, although its
26 functional significance is unknown (Kurata et al., 1996; Seki et al., 1997). In macrophages,
27 cGMP synthesis promotes chemotaxis, and cGMP-mediated synthesis and secretion of tumor
28 necrosis factor a has been linked to both CO and 'NO (Arias-Diaz et al., 1995; Belenky et al.,
29 1993). Carbon monoxide causes smooth muscle relaxation by stimulating soluble guanylate
30 cyclase (Utz and Ullrich, 1991; Wang et al., 1997b). Smooth muscle relaxation also may occur
31 because of activation of calcium dependent potassium channels, although this effect may be
October 1999 5-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 linked to guanylate cyclase activity (Trischmann et al., 1991; Wang et al., 1997a). Carbon
2 monoxide-mediated smooth muscle relaxation is involved with control of microvascular hepatic
3 portal blood flow (Goda et al., 1998; Pannen and Bauer, 1998) and suppressing contractions in
4 the gravid uterus (Acevedo and Ahmed, 1998). It also may play a role in gastrointestinal motility
5 (Farrugia et al., 1998).
6
7
8 5.7 MECHANISMS OF CARBON MONOXIDE TOXICITY
9 5.7.1 Alterations in Blood Flow
10 Carbon monoxide from environmental pollution may exert similar effects in vivo to those
11 of endogenously produced CO, because the nanomolar tissue concentrations resulting from
12 inhalation of CO are comparable or greater than concentrations produced by cells possessing
13 heme oxygenase. Liver parenchyma has been estimated to generate approximately 0.45 nmol
14 CO/gram liver/min (Goda et al., 1998). Carbon monoxide synthesis by smooth muscle cells is
15 approximately 8 pmol/mg protein/min for human aorta and 23 to 37 pmol/mg protein/min for rat
16 aorta (Cook et al., 1995; Grundemar et al., 1995). Carbon monoxide production by unstimulated
17 pulmonary macrophages is 3.6 pmol/mg protein/min, and, after stimulation with
18 lipopolysaccharide, it increases to about 5.1 pmol/mg protein/min (Arias-Diaz et al., 1995). The
19 rate of synthesis of CO varies widely for nerve cells. Cerebellar granule cells generate
20 approximately 3 fmol/mg protein/min, olfactory nerve cells produce 4.7 pmol/mg protein/min,
21 and rat cerebellar homogenates can generate as much as 56.6 pmol/mg protein/min (Ingi and
22 Ronnett, 1995; Ingi et al., 1996; Maines, 1988; Nathanson et al., 1995).
23 Vasodilation is a well-established effect caused by exposure to environmental CO. At high
24 CO concentrations, on the order of 500 to 2,000 ppm, the mechanism is related to impairment of
25 O2 delivery (Kanten et al., 1983; MacMillan, 1975). However, a portion of the observed
26 increases in cerebral blood flow are independent of pertubations in O2 supply (Koehler et al.,
27 1982). In a setting where cellular oxidative metabolism was not impaired by CO, elevations in
28 cerebral blood flow appeared to be mediated by 'NO (Meilin et al., 1996). Whether the
29 mechanism was the same as that outlined in the section above, which causes oxidative stress,
30 remains to be determined.
October 1999 5-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 It is unclear whether disturbances in vascular tone by environmental CO is a generalized,
2 systemic response, and the impact of variables such as the duration of exposure have not been
3 adequately investigated. Although cerebral vasodilation mediated by 'NO was reported with
4 exposures to 1,000 ppm CO, 1,000 ppm CO did not alter pulmonary vasoconstriction in an
5 isolated-perfused rat lung model (Cantrell and Tucker, 1996). Exposure to 150,000 ppm CO
6 caused no changes in pulmonary artery pressure in isolated blood-perfused lungs, although CO
7 did inhibit hypoxic pulmonary vasoconstriction (Tamayo et al., 1997). Humans exposed to CO
8 for sufficient time to achieve COHb levels of approximately 8% were not found to have
9 alterations in forearm blood flow, blood pressure, or heart rate (Hausberg and Somers, 1997).
10 Animals exposed to high CO concentrations (e.g., 3,000 to 10,000 ppm) have diminished
11 organ blood flow, which contributes to CO-mediated tissue injury (Brown and Piantadosi, 1992;
12 Ginsberg and Meyers, 1974; Okeda et al., 1981; Song et al., 1983; Thorn, 1990). The mechanism
13 is based on CO-mediated hypoxic stress and cardiac dysfunction; therefore, these effects do not
14 arise at CO concentrations relevant to ambient air quality.
15
16 5.7.2 Mitochondrial Dysfunction and Altered Production of High-Energy
17 Intermediates
18 When exposed to 10,000 ppm CO, rats exhibit impaired high-energy phosphate synthesis
19 and production of hydroxyl free radicals because of mitochondrial dysfunction (Brown and
20 Piantadosi, 1992; Piantadosi et al., 1995). Exposure to 2,500 ppm CO also will cause hydroxyl
21 radicals to be produced, apparently by mitochondria, because of a process that could not be
22 related to hypoxic stress (Piantadosi et al., 1997a). Evidence for mitochondrial dysfunction has
23 not been observed in vivo at lower CO concentrations. However, under conditions of high
24 metabolic demand, exposure to even 1,000 ppm CO in the absence of an overt hypoxic stress will
25 result in impaired energy production in brain (Meilin et al., 1996).
26 Carbon monoxide binding to mitochondrial cytochromes of respiring cells in vitro has been
27 documented only when either the CO concentration was extraordinarily high, or O2 tension was
28 extremely low, such that the CO/O2 ratio exceeded 12:1 (Coburn and Forman, 1987). Following
29 CO exposure and removal to fresh air, CO diffuses out from cells, and mitochondrial function is
30 restored. This process is enhanced by inspiration of hyperbaric oxygen (Brown and Piantadosi,
31 1992). Studies in mice indicate that high CO concentrations inhibit synthesis of high-energy
October 1999 5-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 phosphates during exposure to 5,000 ppm CO for 15 min and these changes do not persist
2 following removal to fresh air (Matsuoka et al., 1993). In summary, mitochondrial dysfunction
3 and impaired high-energy phosphate synthesis have been shown by several independent
4 laboratories to occur during exposures to high CO concentrations. Current information suggests
5 that this alteration does not occur at CO concentrations relevant to ambient air quality, and that
6 changes in energy production are not persistent for long periods of time following CO exposure.
7
8 5.7.3 Vascular Insults Associated with Exposure to Carbon Monoxide
9 There are two primary variables that impact on CO toxicity. One is the concentration of
10 CO, the other is the duration of exposure. Traditionally, these two variables have been viewed as
11 merely alternative ways of elevating COHb concentration in the body. The concentration of CO
12 breathed dictates the duration of exposure required to achieve a particular blood level of COHb
13 or tissue level of CO. This view is predicated on the notion that CO pathophysiology is
14 determined by its binding to one or another hemoprotein and to inhibition of oxygen delivery or
15 oxidative metabolism.
16 There is a substantial body of literature to suggest that, at least with regard to vascular
17 effects, the duration of exposure has a greater impact on the magnitude of CO pathophysiology
18 than what is predicted based on the concentration of CO that is inspired. For example, the lungs
19 are the first site for potential action of environmental CO. Results from investigations have been
20 conflicting regarding the risk for pulmonary injury from CO. Because of the lack of consensus
21 and also the absence of a recognized biochemical mechanism, low concentrations of CO have
22 been viewed as posing little risk to lung physiology (U.S. Environmental Protection Agency,
23 1991, 1992). When animals have been exposed to high CO concentrations sufficient to raise
24 COHb levels to 56 to 90%, exposures have lasted for only minutes because of the hypoxic stress.
25 In these studies, evidence of increased capillary permeability was inconsistent (Fein et al., 1980;
26 Niden and Schulz, 1965; Penney et al., 1988), and no other alterations in lung physiology were
27 detected (Fisher et al., 1969; Robinson et al., 1985; Shimazu et al., 1990; Sugi et al., 1990).
28 In contrast, when human beings or experimental animals have been exposed to relatively low CO
29 concentrations for many hours at a time, capillary leakage of macromolecules from the lungs and
30 systemic vasculature have been documented, but the presence of an hypoxic stress was
October 1999 5-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 questioned (Kjeldsen et al., 1972; Maurer, 1941; Parving et al., 1972; Siggaard-Andersen et al.,
2 1968).
3 In light of the physiological role for CO in vasomotor control, protracted exposures may be
4 prone to disturb vascular homeostasis, giving rise to pathophysiological responses. Monkeys
5 exposed to 250 ppm CO for 2 weeks exhibited coronary artery damage consisting of
6 subendothelial edema, fatty streaking, and lipid-loaded cells (Thomsen, 1974). This study and
7 others (Armitage et al., 1976; Davies et al., 1976; Turner et al., 1979; Webster et al., 1968) have
8 suggested a link between atherosclerosis and chronic CO exposure. However, other studies have
9 failed to find evidence for an association (Hugod et al., 1978; Penn et al., 1992).
10 Carbon monoxide may cause vascular insults. Leakage of albumin and leukocyte
11 sequestration has been shown following exposures of rats to 50 ppm or more for 1 h, and the
12 process is mediated by 'NO-derived oxidants (Ischiropoulos et al., 1996; Thorn, 1993; Thorn
13 et al., 1998, 1999a,b). Brain oxidative stress associated with this mechanism has been shown
14 with rats exposed to 1,000 to 3,000 ppm CO for 1 h (Ischiropoulos et al., 1996; Thorn, 1993).
15 However, it is unclear whether the flux of'NO, resulting from exposures to lower CO
16 concentrations contribute to oxidative or nitrosative stress in vivo. Important differences in the
17 patterns of leakage from pulmonary and systemic vascular beds suggest that they may be caused
18 by different mechanisms. For example, systemic vascular leakage was present for several hours
19 after CO exposure, and the leakage resolved within 18 h, whereas pulmonary vascular leakage
20 was not measurable until 18 h after CO exposure, and it resolved by 48 h. Both pulmonary and
21 systemic vascular leakage occurred after hour-long exposures to CO, but not when exposures
22 lasted for only 30 min, and vascular changes were not different whether rats were exposed to just
23 50 ppm or as much as 1,000 ppm CO. These are recent observations and further investigations
24 are required before their relevance to environmental CO contamination can be assessed
25 adequately. Moreover, it should be emphasized that the vascular leakage observed in lungs and
26 systemic microvasculature following exposures to CO at concentrations as low as 50 ppm may
27 have no pathophysiological impact if regional lymphatics can sustain a higher flow so that edema
28 does not occur (Thorn et al., 1998, 1999a,b).
29
30
October 1999 5-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.8 OTHER EFFECTS OF CARBON MONOXIDE
2 Among the most concerning pathophysiological effects of CO is its propensity for causing
3 brain damage. There has been considerable effort focussed on potential mechanisms for this
4 process. With regard to ambient air standards, however, it is important to note that recent studies
5 were done with high CO concentrations. Carbon monoxide poisoning is not a "pure"
6 pathological process, as injuries may be precipitated by a combination of cardiovascular effects
7 linked to hypoperfusion or frank ischemia, COHb-mediated hypoxic stress, and intracellular
8 effects, including free radical production and oxidative stress. For example, CO poisoning
9 causes elevations of glutamate and dopamine in experimental models and human fatalities
10 (Arranz et al., 1997; Ishimaru et al., 1991, 1992; Nabeshima et al., 1990, 1991; Newby et al.,
11 1978; Piantadosi et al., 1997b). These elevations occur because of the CO-associated
12 cardiovascular compromise and, possibly, other direct CO-mediated effects. Based on the effects
13 of agents that block the N-methyl-D-aspartate (NMD A) receptor, elevations of glutamate in
14 experimental CO poisoning have been linked to a delayed type (but not an acute type) of
15 amnesia, to loss of CA1 neurons in the hippocampus of mice, and to loss of glutamate-dependent
16 cells in the inner ear of rats (Ishimaru et al., 1991, 1992; Liu and Fechter, 1995; Nabeshima et al.,
17 1990, 1991). Antioxidants can protect against CO-mediated cytotoxicity of glutamate-dependent
18 nerve cells (Fechter et al., 1997). Mechanisms of glutamate neurotoxicity include excessive
19 calcium influx, free radical-mediated injury that may include calcium-calmodulin-dependent
20 activation of cytosolic NO synthase, and lipid peroxidation. Moderate stimulation by excitatory
21 amino acids may cause mitochondrial dysfunction with impaired synthesis of adenosine
22 triphosphate and production of reactive O2 species (Beal, 1992). Cell death can be through
23 necrosis or programmed cell death, depending on the intensity of the stimulus (Gwag et al.,
24 1995). There also may be a synergistic injury with other forms of oxidative stress because
25 reactive O2 species can intensify excitotoxicity (Bridges et al., 1991; Pellegrini-Giampietro et al.,
26 1990). Glutamate also can injure cells in the central nervous system that do not have NMDA
27 receptors by competing for cysteine uptake, which inhibits synthesis of glutathione (Lipton et al.,
28 1997; Murphy et al., 1989; Oka et al., 1993).
29
30
October 1999 5-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.9 SUMMARY
2 The most prominent pathophysiological effect of CO is hypoxemia caused by avid binding
3 of CO to Hb. Formation of COHb reduces O2-carrying capacity of blood and impairs release of
4 O2 from O2Hb to tissues. In addition to tissue hypoxia, ultimate diffusion of CO to cells may
5 affect adversely their function. The brain and heart tissues are particularly sensitive to
6 CO-induced hypoxia and cytotoxicity. The rate of COHb formation and elimination depends on
7 many physical and physiological factors. The same factors that govern CO uptake determine CO
8 elimination as well. The flow of CO between blood and either alveolar air or the tissues, and
9 vice versa, is governed by the CO concentration gradient between these compartments. Because
10 of a small blood-to-air CO gradient, and tight binding of CO to Hb, the elimination half-time is
11 quite long, varying from 2.0 to 6.5 h. Apart from the CO concentration in ambient air, the
12 principal determinants of CO uptake are minute ventilation and lung diffusion capacity. Thus,
13 any physiological conditions that affect these variables (e.g., exercise, age) also will affect the
14 kinetics of COHb. Both the physical and physiological variables have been integrated into many
15 empirical and mathematical models of COHb formation and elimination under static and
16 dynamic conditions of ambient CO concentration and physiologic function. The nonlinear CFK
17 equation is the most widely used predictive model of COHb formation, and it still is considered
18 the best all-around model for COHb prediction. Altitude may have a significant influence on
19 COHb kinetics. The effects of hypoxic hypoxia (altitude) and CO-induced hypoxia appear to be
20 additive. Adaptation to altitude will moderate COHb formation. In addition to exogenous
21 sources of CO, the gas also is produced endogenously through catabolism of Hb, metabolic
22 processes of drugs, and degradation of inhaled solvents and other xenobiotics. This last source
23 may lead to very high (up to 50%) COHb concentrations. Many disorders, particularly anemias
24 of any etiology, will predispose affected individuals to CO hypoxia. Furthermore, patients with
25 a variety of cardiopulmonary (e.g., COPD, CAD) and chronic inflammatory diseases may be at
26 increased risk because of elevated COHb.
27 Apart from impaired O2 delivery to the tissues because of COHb formation, recent studies
28 of CO pathophysiology suggest cytotoxic effects independent of O2. New investigations have
29 expanded the understanding of CO in two areas. First, there is a growing recognition of the role
30 that CO may play in normal neurophysiology and in microvascular vasomotor control. The
31 impact of CO from ambient air on these processes has not been investigated adequately. Hence,
October 1999 5-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 there is insufficient information available to influence decisions on ambient air quality standards.
2 The second area of investigation of CO is related to its propensity for causing
3 free-radical-mediated changes in tissues. Mechanisms for these changes have been linked to both
4 mitochondria and to a CO-mediated disturbance of intracellular "buffering" of endogenously
5 generated free radicals (e.g., 'NO). The role these mechanisms play in pathophysiology currently
6 is being investigated. Where dose-response studies are available, the concentrations of CO that
7 cause adverse effects in animals exceed current NAAQS.
October 1999 5-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Acevedo, C. H.; Ahmed, A. (1998) Hemeoxygenase-1 inhibits human myometrial contractility via carbon monoxide
3 and is upregulated by progesterone during pregnancy. J. Clin. Invest. 101: 949-955.
4 Andersen, M. E.; Clewell, H. M, III; Gargas, M. L.; MacNaughton, M. G.; Reitz, R. H.; Nolan, R. J.; McKenna,
5 M. J. (1991) Physiologically based pharmacokinetic modeling with dichloromethane, its metabolite, carbon
6 monoxide, and blood carboxyhemoglobin in rats and humans. Toxicol. Appl. Pharmacol. 108: 14-27.
7 Arias-Diaz, J.; Vara, E.; Garcia, C.; Villa, N.; Balibrea, J. L. (1995) Evidence for a cyclic guanosine
8 monophosphate-dependent, carbon monoxide-mediated, signaling system in the regulation of TNF-cc
9 production by human pulmonary macrophages. Arch. Surg. (Chicago) 130: 1287-1293.
10 Armitage, A. K.; Davies, R. F.; Turner, D. M. (1976) The effects of carbon monoxide on the development of
11 atherosclerosis in the White Carneau pigeon. Atherosclerosis 23: 333-344.
12 Arranz, B.; Blennow, K.; Eriksson, A.; Mansson, J.-E.; Marcusson, J. (1997) Serotonergic, noradrenergic, and
13 dopaminergic measures in suicide brains. Biol. Psychiatry 41: 1000-1009.
14 Barie, P. S.; Wu, W.; Hariri, R. J.; Halebian, P. H.; Shires, G. T. (1994) Alterations of pulmonary gas exchange
15 after superimposed carbon monoxide poisoning in acute lung injury. Surgery (St. Louis) 115: 678-686.
16 Beal, M. F. (1992) Does impairment of energy metabolism result in excitotoxic neuronal death in neurodegenerative
17 illnesses? Ann. Neurol. 31: 119-130.
18 Belenky, S. N.; Robbins, R. A.; Rubinstein, I. (1993) Nitric oxide synthase inhibitors attenuate human monocyte
19 chemotaxis in vitro. J. Leukocyte Biol. 53: 498-503.
20 Bender, P. R.; Weil, J. V.; Reeves, J. T.; Moore, L. G. (1987) Breathing pattern in hypoxic exposures of varying
21 duration. J. Appl. Physiol. 62: 640-645.
22 Benignus, V. A. (1995) A model to predict carboxyhemoglobin and pulmonary parameters after exposure to O2,
23 CO2, and CO. Aviat. Space Environ. Med. 66: 369-374.
24 Benignus, V. A.; Annau, Z. (1994) Carboxyhemoglobin formation due to carbon monoxide exposure in rats.
25 Toxicol. Appl. Pharmacol. 128: 151-157.
26 Benignus, V. A.; Hazucha, M. J.; Smith, M. V.; Bromberg, P. A. (1994) Prediction of carboxyhemoglobin
27 formation due to transient exposure to carbon monoxide. J. Appl. Physiol. 76: 1739-1745.
28 Berk, P. D.; Rodkey, F. L.; Blaschke, T. F.; Collison, H. A.; Waggoner, J. G. (1974) Comparison of plasma
29 bilirubin turnover and carbon monoxide production in man. J. Lab. Clin. Med. 83: 29-37.
30 Berk, P. D.; Blaschke, T. F.; Scharschmidt, B. F.; Waggoner, J. G.; Berlin, N. I. (1976) A new approach to
31 quantitation of the various sources of bilirubin in man. J. Lab. Clin. Med. 87: 767-780.
32 Bernard, T. E.; Duker, J. (1981) Modeling carbon monoxide uptake during work. Am. Ind. Hyg. Assoc. J.
33 42:361-364.
34 Biller, W. F.; Richmond, H. M. (1982) Sensitivity analysis on Coburn model predictions of COHb levels associated
35 with alternative CO standards. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
36 of Air Quality Planning and Standards; EPA contract no. 68-02-3600.
37 Biller, W. F.; Richmond, H. M. (1992) COHb module for a probabilistic CO NEM. In: Johnson, T.; Capel, J.;
3 8 Paul, R.; Wijnberg, L. Estimation of carbon monoxide exposures and associated carboxyhemoglobin levels in
39 Denver residents using a probabilistic version of NEM. Research Triangle Park, NC: U.S. Environmental
40 Protection Agency, Office of Air Quality Planning and Standards; contract no. 68-DO-0062; appendix C.
41 Boothby, W. M.; Lovelace, W. R., II; Benson, O. O., Jr.; Strehler, A. F. (1954) Volume and partial pressures of
42 respiratory gases at altitude. In: Boothby, W. M., ed. Respiratory physiology in aviation. Randolph Field, TX:
43 U.S. Air Force School of Aviation Medicine; pp. 39-49.
44 Bridges, R. J.; Koh, J.-Y.; Hatalski, C. G.; Cotman, C. W. (1991) Increased excitotoxic vulnerability of cortical
45 cultures with reduced levels of glutathione. Eur. J. Pharmacol. 192: 199-200.
46 Britten, J. S.; Myers, R. A. M. (1985) Effects of hyperbaric treatment on carbon monoxide elimination in humans.
47 Undersea Biomed. Res. 12: 431-438.
48 Brown, S. D.; Piantadosi, C. A. (1992) Recovery of energy metabolism in rat brain after carbon monoxide hypoxia.
49 J. Clin. Invest. 89: 666-672.
50 Btinger, J.; Bombosch, F.; Mesecke, U.; Hallier, E. (1997) Monitoring and analysis of occupational exposure to
51 chain saw exhausts. Am. Ind. Hyg. Assoc. J. 58: 747-751.
52 Burki, N. K. (1984) Effects of acute exposure to high altitude on ventilatory drive and respiratory pattern. J. Appl.
53 Physiol. Respir. Environ. Exercise Physiol. 56: 1027-1031.
October 1999 5-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 Cantrell, J. M; Tucker, A. (1996) Low-dose carbon monoxide does not reduce vasoconstriction in isolated rat
2 lungs. Exp. Lung Res. 22: 21-32.
3 Cassina, A.; Radi, R. (1996) Differential inhibitory action of nitric oxide and peroxynitrate on mitochondrial
4 electron transport. Arch. Biochem Biophys. 328: 309-316.
5 Cavallin-Stahl, E.; Mercke, C.; Lundh, B. (1976) Carbon monoxide production in patients with breast carcinoma.
6 Br. J. Haematol. 32: 177-182.
7 Chung, S. J. (1988) Formulas predicting carboxyhemoglobin resulting from carbon monoxide exposure. Vet. Hum.
8 Toxicol. 30: 528-532.
9 Coburn, R. F. (1967) Endogenous carbon monoxide production and body CO stores. Acta Med. Scand. Suppl.
10 472:269-282.
11 Coburn, R. F. (1970a) Enhancement by phenobarbital and diphenylhydantoin of carbon monoxide production in
12 normal man. N. Engl. J. Med. 283: 512-515.
13 Coburn, R. F. (1970b) The carbon monoxide body stores. In: Coburn, R. F., ed. Biological effects of carbon
14 monoxide. Ann. N. Y. Acad. Sci. 174: 11-22.
15 Coburn, R. F. (1973) Mean myoglobin oxygen tension in skeletal and cardiac muscle. In: Bicher, H. I.; Bruley,
16 D. F., eds. Oxygen transport to tissue: instrumentation, methods, and physiology. New York, NY: Plenum
17 Press; pp. 571-577. (Advances in experimental medicine and biology: v. 37A-B).
18 Coburn, R. F.; Forman, H. J. (1987) Carbon monoxide toxicity. In: Fishman, A. P.; Farhi, L. E.; Tenney, S. M.;
19 Geiger, S. R., eds. Handbook of physiology: a critical, comprehensive presentation of physiological
20 knowledge and concepts. Section 3: The respiratory system. Volume IV. Gas exchange. Bethesda, MD:
21 American Physiological Society; pp. 439-456.
22 Coburn, R. F.; Mayers, L. B. (1971) Myoglobin O2 tension determined from measurements of carboxymyoglobin in
23 skeletal muscle. Am. J. Physiol. 220: 66-74.
24 Coburn, R. F.; Blakemore, W. S.; Forster, R. E. (1963) Endogenous carbon monoxide production in man. J. Clin.
25 Invest. 42: 1172-1178.
26 Coburn, R. F.; Williams, W. J.; Forster, R. E. (1964) Effect of erythrocyte destruction on carbon monoxide
27 production in man. J. Clin. Invest. 43: 1098-1103.
28 Coburn, R. F.; Forster, R. E.; Kane, P. B. (1965) Considerations of the physiological variables that determine the
29 blood carboxyhemoglobin concentration in man. J. Clin. Invest. 44: 1899-1910.
30 Coburn, R. F.; Williams, W. J.; Kahn, S. B. (1966) Endogenous carbon monoxide production in patients with
31 hemolytic anemia. J. Clin. Invest. 45: 460-468.
32 Coceani, F.; Breen, C. A.; Lees, J. G.; Falck, J. R.; Olley, P. M. (1988) Further evidence implicating a cytochrome
33 P-450-mediated reaction in the contractile tension of the lamb ductus arteriosus. Circ. Res. 62: 474-477.
34 Cook, M. N.; Nakatsu, K.; Marks, G. S.; McLaughlin, B. E.; Vreman, H. J.; Stevenson, D. K.; Brien, J. F. (1995)
3 5 Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can. J.
36 Physiol. Pharmacol. 73:515-518.
37 Davies, R. F.; Topping, D. L.; Turner, D. M. (1976) The effect of intermittent carbon monoxide exposure on
38 experimental atherosclerosis in the rabbit. Atherosclerosis 24: 527-536.
39 Delivoria-Papadopoulos, M.; Coburn, R. F.; Forster, R. E. (1974) Cyclic variation of rate of carbon monoxide
40 production in normal women. J. Appl. Physiol. 36: 49-51.
41 Deller, A.; Stenz, R.; Forstner, K.; Konrad, F. (1992) Die elimination von kohlenmonoxydhamoglobin -
42 geschlechtsspezifische und zirkadiane einflusse [The elimination of carboxyhemoglobin COHB sex-specific
43 and circadian influences in healthy volunteers]. Infusionstherapie 19: 121-126.
44 Durante, W.; Kroll, M. H.; Christodoulides, N.; Peyton, K. J.; Schafer, A. I. (1997) Nitric oxide induces heme
45 oxygenase-1 gene expression and carbon monoxide production in vascular smooth muscle cells. Circ. Res.
46 80:557-564.
47 Engel, L. A.; Wood, L. D. H.; Utz, G.; Macklem, P. T. (1973) Gas mixing during inspiration. J. Appl. Physiol.
48 35: 18-24.
49 Farrugia, G.; Miller, S. M.; Rich, A.; Liu, X.; Maines, M. D.; Rae, J. L.; Szurszewski, J. H. (1998) Distribution of
50 heme oxygenase and effects of exogenous carbon monoxide in canine jejunum. Am. J. Physiol.
51 274: G350-G358.
52 Fechter, L. D.; Liu, Y.; Pearce, T. A. (1997) Cochlear protection from carbon monoxide exposure by free radical
53 blockers in the guinea pig. Toxicol. Appl. Pharmacol. 142: 47-55.
54 Fein, A.; Grossman, R. F.; Jones, J. G.; Hoeffel, J.; McKay, D. (1980) Carbon monoxide effect on alveolar
5 5 epithelial permeability. Chest 78: 726-731.
October 1999 5-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fisher, A. B.; Hyde, R. W.; Baue, A. E.; Reif, J. S.; Kelly, D. F. (1969) Effect of carbon monoxide on function and
2 structure of the lung. J. Appl. Physiol. 26: 4-12.
3 Forbes, W. H.; Sargent, F.; Roughton, F. J. W. (1945) The rate of carbon monoxide uptake by normal men. Am. J.
4 Physiol. 143: 594-608.
5 Forster, R. E. (1987) Diffusion of gases across the alveolar membrane. In: Fishman, A. P.; Farhi, L. E.; Tenney,
6 S. M; Geiger, S. R., eds. Handbook of physiology: a critical, comprehensive presentation of physiological
7 knowledge and concepts. Section 3: the respiratory system. Volume IV. Gas exchange. Bethesda, MD:
8 American Physiological Society; pp. 71-88.
9 Foster, L. J.; Corrigan, K.; Goldman, A. L. (1978) Effectiveness of oxygen therapy in hypoxic polycythemic
10 smokers. Chest 73: 572-576.
11 Frey, T. M.; Crapo, R. O.; Jensen, R. L.; Elliott, C. G. (1987) Diurnal variation of the diffusing capacity of the lung:
12 is it real? Am. Rev. Respir. Dis. 136: 1381-1384.
13 Ge, R.-L.; Matsuzawa, Y.; Takeoka, M.; Kubo, K.; Sekiguchi, M.; Kobayashi, T. (1997) Low pulmonary diffusing
14 capacity in subjects with acute mountain sickness. Chest 111: 58-64.
15 Gibson, Q. H.; Olson, J. S.; McKinnie, R. E.; Rohlfs, R. J. (1986) A kinetic description of ligand binding to sperm
16 whale myoglobin. J. Biol. Chem. 261: 10,228-10,236.
17 Ginsberg, M. D.; Myers, R. E. (1974) Experimental carbon monoxide encephalopathy in the primate. I. Physiologic
18 and metabolic aspects. Arch. Neurol. 30: 202-208.
19 Goda, N.; Suzuki, K.; Naito, M.; Takeoka, S.; Tsuchida, E.; Ishimura, Y.; Tamatani, T.; Suematsu, M. (1998)
20 Distribution of heme oxygenase isoforms in rat liver: topographic basis for carbon monoxide-mediated
21 microvascular relaxation. J. Clin. Invest. 101: 604-612.
22 Gothert, M.; Lutz, F.; Malorny, G. (1970) Carbon monoxide partial pressure in tissue of different animals. Environ.
23 Res. 3: 303-309.
24 Grenlund, J.; Garby, L. (1984) Numerical values of the classical Haldane coefficient. J. Appl. Physiol. 57: 850-859.
25 Grote, J.; Ball, P.; Oltmanns, K.; Stolp, W. (1994) The effect of increased blood carbon monoxide levels on the
26 hemoglobin oxygen affinity during pregnancy. In: Vaupel, P.; Zander R.; Bruley, D. F., eds. Oxygen
27 transport to tissue XV: proceedings of the twentieth annual meeting of the International Society on Oxygen
28 Transport to Tissue; August 1992; Mainz, Germany. New York, NY: Plenum Press; pp. 145-150. (Advances
29 in experimental medicine and biology v. 345).
30 Grundemar, L.; Johansson, M.-B.; Ekelund, M.; Hogestatt, E. D. (1995) Haem oxygenase activity in blood vessel
31 homogenates as measured by carbon monoxide production. Acta Physiol. Scand. 153: 203-204.
32 Guenard, H.; Marthan, R. (1996) Pulmonary gas exchange in elderly subjects. Eur. Respir. J. 9: 2573-2577.
33 Gwag, B. J.; Lobner, D.; Koh, J. Y.; Wie, M. B.; Choi, D. W. (1995) Blockade of glutamate receptors unmasks
34 neuronal apoptosis after oxygen-glucose deprivation in vitro. Neuroscience 68: 615-619.
35 Haab, P. E.; Durand-Arczynska, W. Y. (1991) Carbon monoxide effects on oxygen transport. In: Crystal, R. G.;
36 West, J. B.; Barnes, P. J.; Cherniack, N. S.; Weibel, E. R., eds. The lung: scientific applications, v. 2. New
37 York, NY: Raven Press; pp. 1267-1275.
3 8 Haldane, J. (1897-1898) Some improved methods of gas analysis. J. Physiol. (London) 22: 465-480.
39 Harf, A.; Pratt, T.; Hughes, J. M. B. (1978) Regional distribution ofVAX^ in man at rest and with exercise measured
40 with krypton-Sim. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 44: 115-123.
41 Hauck, H.; Neuberger, M. (1984) Carbon monoxide uptake and the resulting carboxyhemoglobin in man. Eur. J.
42 Appl. Physiol. 53: 186-190.
43 Hausberg, M.; Somers, V. K. (1997) Neural circulatory responses to carbon monoxide in healthy humans.
44 Hypertension 29: 1114-1118.
45 Hill, E. P.; Hill, J. R.; Power, G. G.; Longo, L. D. (1977) Carbon monoxide exchanges between the human fetus and
46 mother: a mathematical model. Am. J. Physiol. 232: H311-H323.
47 Horvath, S. M.; Agnew, J. W.; Wagner, J. A.; Bedi, J. F. (1988) Maximal aerobic capacity at several ambient
48 concentrations of carbon monoxide at several altitudes. Cambridge, MA: Health Effects Institute; research
49 report no. 21.
50 Horvath, I.; Loukides, S.; Wodehouse, T.; Kharitonov, S. A.; Cole, P. J.; Barnes, P. J. (1998a) Increased levels of
51 exhaled carbon monoxide in bronchiectasis: a new marker of oxidative stress. Thorax 53: 867-870.
52 Horvath, I.; Donnelly, L. E.; Kiss, A.; Paredi, P.; Kharitonov, S. A.; Barnes, P. J. (1998b) Raised levels of exhaled
53 carbon monoxide are associated with an increased expression of heme oxygenase-1 in airway macrophages in
54 asthma: a new marker of oxidative stress. Thorax 53: 668-672.
October 1999 5-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hugod, C.; Hawkins, L. H.; Kjeldsen, K.; Thomsen, H. K.; Astrap, P. (1978) Effect of carbon monoxide exposure
2 on aortic and coronary intimal morphology in the rabbit. Atherosclerosis 30: 333-342.
3 Huie, R. E.; Padmaja, S. (1993) The reaction of NO with superoxide. Free Radical Res. Commun. 18: 195-199.
4 Ingi, T.; Ronnett, G. V. (1995) Direct demonstration of a physiological role for carbon monoxide in olfactory
5 receptor neurons. J. Neurosci. 15: 8214-8222.
6 Ingi, T.; Cheng, J.; Ronnett, G. V. (1996) Carbon monoxide: an endogenous modulator of the nitric oxide-cyclic
7 GMP signaling system. Neuron 16: 835-842.
8 Ingram, D.; Dickinson, C. J.; Ahmed, K. (1987) MACPUF graphics version: a simulation of human respiration, gas
9 exchange and control. London, United Kingdom: CHIME, UCL Medical School. Available:
10 http://crnserver.chime.ucl.ac.uk/Models/macpuf.htm.
11 Ischiropoulos, H.; Zhu, L.; Chen, J.; Tsai, M.; Martin, J. C.; Smith, C. D.; Beckman, J. S. (1992)
12 Peroxynitrate-mediated tyrosine nitration catalyzed by superoxide dismutase. Arch. Biochem. Biophys.
13 298:431-437.
14 Ischiropoulos, H.; Beers, M. F.; Ohnishi, S. T.; Fisher, D.; Garner, S. E.; Thorn, S. R. (1996) Nitric oxide
15 production and perivascular tyrosine nitration in brain after carbon monoxide poisoning in the rat. J. Clin.
16 Invest. 97: 2260-2267.
17 Ishimaru, H.; Nabeshima, T.; Katoh, A.; Suzuki, H.; Fukuta, T.; Kameyama, T. (1991) Effects of successive carbon
18 monoxide exposures on delayed neuronal death in mice under the maintenance of normal body temperature.
19 Biochem. Biophys. Res. Commun. 179: 836-840.
20 Ishimaru, H.; Katoh, A.; Suzuki, H.; Fukuta, T.; Kameyama, T.; Nabeshima, T. (1992) Effects of
21 N-methyl-D-aspartate receptor antagonists on carbon monoxide-induced brain damage in mice. J. Pharmacol.
22 Exp. Ther. 261:349-352.
23 Jay, G. D.; McKindley, D. S. (1997) Alterations in pharmacokinetics of carboxyhemoglobin produced by oxygen
24 underpressure. UnderseaHyperbaric Med. 24: 165-173.
25 Johnson, T.; Capel, J.; Paul, R.; Wijnberg, L. (1992) Estimation of carbon monoxide exposures and associated
26 carboxyhemoglobin levels in Denver residents using a probabilistic version of NEM. Research Triangle Park,
27 NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no.
28 68-DO-0062.
29 Joumard, R.; Chiron, M.; Vidon, R.; Maurin, M.; Rouzioux, J.-M. (1981) Mathematical models of the uptake of
30 carbon monoxide on hemoglobin at low carbon monoxide levels. Environ. Health Perspect. 41: 277-289.
31 Kanten, W. E.; Penney, D. G.; Francisco, K.; Thill, J. E. (1983) Hemodynamic responses to acute
32 carboxyhemoglobinemia in the rat. Am. J. Physiol. 244: H320-H327.
33 Kim, S. K.; Kim, Y. C. (1996) Effect of a single administration of benzene, toluene or /w-xylene on
34 carboxyhaemoglobin elevation and metabolism of dichloromethane in rats. J. Appl. Toxicol. 16: 437-444.
35 Kimmel, E. C.; Carpenter, R. L.; Reboulet, J. E.; Still, K. R. (1999) A physiological model for predicting
36 carboxyhemoglobin formation from exposure to carbon monoxide in rats. J. Appl. Physiol. 86: 1977-1983.
37 Kinker, J. R.; Haffor, A.-S.; Stephan, M.; Clanton, T. L. (1992) Kinetics of CO uptake and diffusing capacity in
38 transition from rest to steady-state exercise. J. Appl. Physiol. 72: 1764-1772.
39 Kjeldsen, K.; Astrup, P.; Wanstrup, J. (1972) Ultrastructural intimal changes in the rabbit aorta after a moderate
40 carbon monoxide exposure. Atherosclerosis 16: 67-82.
41 Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
42 effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
43 Health 53: 388-397.
44 Koehler, R. C.; Jones, M. D., Jr.; Traystman, R. J. (1982) Cerebral circulatory response to carbon monoxide and
45 hypoxic hypoxia in the lamb. Am. J. Physiol. 243: H27-H32.
46 Kurata, S.; Matsumoto, M.; Yamashita, U. (1996) Concomitant transcriptional activation of nitric oxide synthase
47 and heme oxygenase genes during nitric oxide-mediated macrophage cytostasis. J. Biochem. 120: 49-52.
48 Landaw, S. A. (1973) The effects of cigarette smoking on total body burden and excretion rates of carbon
49 monoxide. J. Occup. Med. 15: 231-235.
50 Landaw, S. A.; Callahan, E. W., Jr.; Schmid, R. (1970) Catabolism of heme in vivo: comparison of the simultaneous
51 production of bilirubin and carbon monoxide. J. Clin. Invest. 49: 914-925.
52 Levasseur, L.; Galliot-Guilley, M.; Richter, F.; Scherrmann, J. M.; Baud, F. J. (1996) Effects of mode of inhalation
53 of carbon monoxide and of normobaric oxygen administration on carbon monoxide elimination from the
54 blood. Hum. Exp. Toxicol. 15: 898-903.
October 1999 5-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 Levitt, M. D.; Ellis, C.; Levitt, D. G. (1994) Diurnal rhythm of heme turnover assessed by breath carbon monoxide
2 concentration measurements. J. Lab. Clin. Med. 124: 427-431.
3 Lifshay, A.; Fast, C. W.; Glazier, J. B. (1971) Effects of changes in respiratory pattern on physiological dead space.
4 J. Appl. Physiol. 31: 478-483.
5 Lipton, S. A.; Kim, W.-K.; Choi, Y.-B.; Kumar, S.; D'Emilia, D. M.; Rayudu, P. V.; Arnelle, D. R.; Stamler, J. S.
6 (1997) Neurotoxicity associated with dual actions of homocysteine at the jV-methyl-D-aspartate receptor.
7 Proc. Natl. Acad. Sci. USA 94: 5923-5928.
8 Liu, Y.; Fechter, L. D. (1995) MK-801 protects against carbon monoxide-induced hearing loss. Toxicol. Appl.
9 Pharmacol. 132: 196-202.
10 Lizasoain, I.; Moro, M. A.; Knowles, R. G.; Darley-Usmar, V.; Moncada, S. (1996) Nitric oxide and peroxynitrate
11 exert distinct effects on mitochondria! respiration which are differentially blocked by glutathione or glucose.
12 Biochem. J. 314: 877-880.
13 Longo, L. D.; Hill, E. P. (1977) Carbon monoxide uptake and elimination in fetal and maternal sheep. Am. J.
14 Physiol. 232: H324-H330.
15 Ludbrook, G. L.; Helps, S. C.; Gorman, D. F.; Reilly, P. L.; North, J. B.; Grant, C. (1992) The relative effects of
16 hypoxic hypoxia and carbon monoxide on brain function in rabbits. Toxicology 75: 71-80.
17 Lundh, B.; Cavallin-Stahl, E.; Mercke, C. (1975) Nicotinic acid and the endogenous production of carbon
18 monoxide. ActaMed. Scand. 197: 173-176.
19 Luomanmaki, K.; Coburn, R. F. (1969) Effects of metabolism and distribution of carbon monoxide on blood and
20 body stores. Am. J. Physiol. 217: 354-363.
21 Luperchio, S.; Tamir, S.; Tannenbaum, S. R. (1996) NO-induced oxidative stress and glutathione metabolism in
22 rodent and human cells. Free Radical Biol. Med. 21: 513-519.
23 Lynch, S. R.; Moede, A. L. (1972) Variation in the rate of endogenous carbon monoxide production in normal
24 human beings. J. Lab. Clin. Med. 79: 85-95.
25 MacMillan, V. (1975) Regional cerebral blood flow of the rat in acute carbon monoxide intoxication. Can. J.
26 Physiol. Pharmacol. 53:644-650.
27 Maines, M. D. (1988) Heme oxygenase: function, multiplicity, regulatory mechanisms, and clinical applications.
28 FASEB J. 2: 2557-2568.
29 Maines, M. D.; Mark, J. A.; Ewing, J. F. (1993) Heme oxygenase, a likely regulator of cGMP production in the
30 brain: induction in vivo of HO-1 compensates for depression in NO synthase activity. Mol. Cell. Neurosci.
31 4:398-405.
32 Manno, M.; Rugge, M.; Cocheo, V. (1992) Double fatal inhalation of dichloromethane. Hum. Exp. Toxicol.
33 11:540-545.
34 Marks, G. S. (1994) Heme oxygenase: the physiological role of one of its metabolites, carbon monoxide and
3 5 interactions with zinc protoporphyrin, cobalt protoporphyrin and other metalloporphyrins. Cell. Mol. Biol.
36 (Paris) 40: 863-870.
37 Marthan, R.; Castaing, Y.; Manier, G.; Guenard, H. (1985) Gas exchange alterations in patients with chronic
38 obstructive lung disease. Chest 87: 470-475.
39 Matsuoka, M.; Igisu, H.; Tanaka, I.; Hori, H.; Koga, M. (1993) Brain energy metabolites in mice after an acute
40 exposure to carbon monoxide. Res. Commun. Chem. Pathol. Pharmacol. 81: 15-20.
41 Maurer, F. W. (1941) The effects of carbon monoxide anoxemia on the flow and composition of cervical lymph.
42 Am. J. Physiol. 133: 170-179.
43 McCartney, M. L. (1990) Sensitivity analysis applied to Coburn-Forster-Kane models of carboxyhemoglobin
44 formation. Am. Ind. Hyg. Assoc. J. 51: 169-177.
45 McClean, P. A.; Duguid, N. J.; Griffin, P. M.; Newth, C. J. L.; Zamel, N. (1981) Changes in exhaled pulmonary
46 diffusing capacity at rest and exercise in individuals with impaired positional diffusion. Clin. Respir. Physiol.
47 17: 179-186.
48 McCoubrey, W. K., Jr.; Huang, T. J.; Maines, M. D. (1997) Isolation and characterization of a cDNA from the rat
49 brain that encodes hemoprotein heme oxygenase-3. Eur. J. Biochem. 247: 725-732.
50 McGrath, J. J. (1988) Carbon monoxide studies at high altitude. Neurosci. Biobehav. Rev. 12:311-314.
51 McGrath, J. J. (1992) Effects of altitude on endogenous carboxyhemoglobin levels. J. Toxicol. Environ. Health
52 35: 127-133.
53 McGrath, J. J.; Schreck, R. M.; Lee, P. S. (1993) Carboxyhemoglobin levels in humans: effects of altitude.
54 Inhalation Toxicol. 5:241-249.
October 1999 5-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 Meilin, S.; Rogatsky, G. G.; Thorn, S. R.; Zarchin, N.; Guggenheimer-Furman, E.; Mayevsky, A. (1996) Effects of
2 carbon monoxide on the brain may be mediated by nitric acid. J. Appl. Physiol. 81: 1078-1083.
3 Mercke, C.; Lundh, B. (1976) Erythrocyte filterability and heme catabolism during the menstrual cycle. Ann. Intern.
4 Med. 85: 322-324.
5 Mercke, C.; Cavailin-Stahl, E.; Lundh, B. (1975a) Diurnal variation in endogenous production of carbon monoxide.
6 ActaMed. Scand. 198: 161-164.
7 Mercke, C.; Cavallin-Stahl, E.; Lundh, B. (1975b) Carbon monoxide production and reticulocyte count in normal
8 women: effect of contraceptive drugs and smoking. Acta Med. Scand. 198: 155-160.
9 Mercke, C.; Cavallin-Stahl, E.; Lundh, B. (1975c) Heme catabolism during short-term treatment with phenobarbital,
10 diazepamandoxazepam. ActaMed. Scand. 198: 149-154.
11 Messmer, K. (1982) Oxygen transport capacity. In: Brendel, W.; Zink, R. A., eds. High altitude physiology and
12 medicine. New York, NY: Springer-Verlag; pp. 117-122.
13 Meyer, J.; Prien, T.; Van Aken, H.; Bone, H.-G.; Waurick, R.; Theilmeier, G.; Booke, M. (1998) Arterio-venous
14 carboxyhemoglobin difference suggests carbonmonoxide production by human lungs. Biochem. Biophys.
15 Res. Commun. 244: 230-232.
16 Moore, E. G.; Gibson, Q. H. (1976) Cooperativity in the dissociation of nitric oxide from hemoglobin. J. Biol.
17 Chem. 251:2788-2794.
18 Morita, T.; Perrella, M. A.; Lee, M.-E.; Kourembanas, S. (1995) Smooth muscle cell-derived carbon monoxide is a
19 regulator of vascular cGMP. Proc. Natl. Acad. Sci. U. S. A. 92: 1475-1479.
20 Motterlini, R.; Foresti, R.; Intaglietta, M.; Winslow, R. M. (1996) NO-mediated activation of heme oxygenase:
21 endogenous cytoprotection against oxidative stress to endothelium. Am. J. Physiol. 270: H107-H114.
22 Muller, K. E.; Barton, C. N. (1987) A nonlinear version of the Coburn, Forster and Kane model of blood
23 carboxyhemoglobin. Atmos. Environ. 21: 1963-1967.
24 Murphy, T. H.; Miyamoto, M.; Sastre, A.; Schnaar, R. L.; Coyle, J. T. (1989) Glutamate toxicity in a neuronal cell
25 line involves inhibition of cystine transport leading to oxidative stress. Neuron 2: 1547-1558.
26 Nabeshima, T.; Yoshida, S.; Morinaka, H.; Kameyama, T.; Thurkauf, A.; Rice, K. C.; Jacobson, A. E.; Monn, J. A.;
27 Cho, A. K. (1990) MK-801 ameliorates delayed amnesia, but potentiates acute amnesia induced by CO.
28 Neurosci. Lett. 108: 321-327.
29 Nabeshima, T.; Katoh, A.; Ishimaru, H.; Yoneda, Y.; Ogita, K.; Murase, K.; Ohtsuka, H.; Inari, K.; Fukuta, T.;
30 Kameyama, T. (1991) Carbon monoxide-induced delayed amnesia, delayed neuronal death and change in
31 acetylcholine concentration in mice. J. Pharmacol. Exp. Ther. 256: 378-384.
32 Nathanson, J. A.; Scavone, C.; Scanlon, C.; McKee, M. (1995) The cellular Na+ pump as a site of action for carbon
33 monoxide and glutamate: a mechanism for long-term modulation of cellular activity. Neuron 14: 781-794.
34 Neas, L. M.; Schwartz, J. (1996) The determinants of pulmonary diffusing capacity in a national sample of U.S.
35 adults. Am. J. Respir. Crit. Care Med. 153: 656-664.
36 Newby, M. B.; Roberts, R. J.; Bhatnagar, R. K. (1978) Carbon monoxide- and hypoxia-induced effects on
37 catecholamines in the mature and developing rat brain. J. Pharmacol. Exp. Ther. 206: 61-68.
3 8 Niden, A. H.; Schulz, H. (1965) The ultrastructural effects of carbon monoxide inhalation on the rat lung. Virchows
39 Arch. Pathol. Anat. Physiol. 339: 283-292.
40 Oka, A.; Belliveau, M. J.; Rosenberg, P. A.; Volpe, J. J. (1993) Vulnerability of oligodendroglia to glutamate:
41 pharmacology, mechanisms, and prevention. J. Neurosci. 13: 1441-1453.
42 Okeda, R.; Funata, N.; Takano, T.; Miyazaki, Y.; Higashino, F.; Yokoyama, K.; Manabe, M. (1981) The
43 pathogenesis of carbon monoxide encephalopathy in the acute phase—physiological and morphological
44 correlation. Acta Neuropathol. 54: 1-10.
45 Ott, W. R.; Mage, D. T. (1978) Interpreting urban carbon monoxide concentrations by means of a computerized
46 blood COHb model. J. Air Pollut. Control Assoc. 28: 911-916.
47 Pace, N.; Strajman, E.; Walker, E. L. (1950) Acceleration of carbon monoxide elimination in man by high pressure
48 oxygen. Science (Washington, DC) 111: 652-654.
49 Pankow, D. (1996) Carbon monoxide formation due to metabolism of xenobiotics. In: Penney, D. G., ed. Carbon
50 monoxide. Boca Raton, FL: CRC Press; pp. 25-43.
51 Pannen, B. H.; Bauer, M. (1998) Differential regulation of hepatic arterial and portal venous vascular resistance by
52 nitric oxide and carbon monoxide in rats. Life Sci. 62: 2025-2033.
53 Parving, H.-H.; Ohlsson, K.; Buchardt Hansen, H. J.; Rorth, M. (1972) Effect of carbon monoxide exposure on
54 capillary permeability to albumin and cc2-macroglobulin. Scand. J. Clin. Lab. Invest. 29: 381-388.
October 1999 5-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pellegrini-Giampietro, D. E.; Cherici, G.; Alesiani, M.; Carla, V.; Moroni, F. (1990) Excitatory amino acid release
2 and free radical formation may cooperate in the genesis of ischemia-induced neuronal damage. J. Neurosci.
3 10: 1035-1941.
4 Perm, A.; Currie, J.; Snyder, C. (1992) Inhalation of carbon monoxide does not accelerate arteriosclerosis in
5 cockerels. Eur. J. Pharmacol. 228: 155-164.
6 Penney, D. G.; Davidson, S. B.; Gargulinski, R. B.; Caldwell-Ayre, T. M. (1988) Heart and lung hypertrophy,
7 changes in blood volume, hematocrit and plasma renin activity in rats chronically exposed to increasing
8 carbon monoxide concentrations. J. Appl. Toxicol. 8: 171-178.
9 Peters, T.; Jtirgens, K. D.; Gunther-Jtirgens, G.; Gros, G. (1994) Determination of myoglobin-diffusivity in intact
10 skeletal muscle fibers: an improved microscope-photometrical approach. In: Vaupel, P.; Zander, R.; Bruley,
11 D. F., eds. Oxygen transport to tissue XV. New York, NY: Plenum Press; pp. 677-683. (Advances in
12 experimental medicine and biology v. 345).
13 Peterson, J. E.; Stewart, R. D. (1970) Absorption and elimination of carbon monoxide by inactive young men.
14 Arch. Environ. Health 21: 165-171.
15 Peterson, J. E.; Stewart, R. D. (1972) Human absorption of carbon monoxide from high concentrations in air.
16 Am. Ind. Hyg. Assoc. J. 33: 293-297.
17 Peterson, J. E.; Stewart, R. D. (1975) Predicting the carboxyhemoglobin levels resulting from carbon monoxide
18 exposures. J. Appl. Physiol. 39: 633-638.
19 Piantadosi, C. A.; Tatro, L.; Zhang, J. (1995) Hydroxyl radical production in the brain after CO hypoxia in rats.
20 Free Radical Biol. Med. 18: 603-609.
21 Piantadosi, C. A.; Zhang, J.; Demchenko, I. T. (1997a) Production of hydroxyl radical in the hippocampus after CO
22 hypoxia or hypoxic hypoxia in the rat. Free Radical Biol. Med. 22: 725-732.
23 Piantadosi, C. A.; Zhang, J.; Levin, E. D.; Folz, R. J.; Schmechel, D. E. (1997b) Apoptosis and delayed neuronal
24 damage after carbon monoxide poisoning in the rat. Exp. Neurol. 147: 103-114.
25 Poderoso, J. J.; Cerreras, M. C.; Lisdero, C.; Riobo, N.; Schopfer, F.; Boveris, A. (1996) Nitric oxide inhibits
26 electron transfer and increases superoxide radical production in rat heart mitochondria and submitochondrial
27 particles. Arch. Biochem. Biophys. 328: 85-92.
28 Robinson, N. B.; Barie, P. S.; Halebian, P. H.; Shires, G. T. (1985) Distribution of ventilation and perfusion
29 following acute carbon monoxide poisoning. In: 41st Annual forum on fundamental surgical problems held at
30 the 71st annual clinical congress of the American College of Surgeons; October; Chicago, IL. Surg. Forum
31 36:115-118.
32 Rodgers, P. A.; Vreman, H. J.; Dennery, P. A.; Stevenson, D. K. (1994) Sources of carbon monoxide (CO) in
33 biological systems and applications of CO detection technologies. Semin. Perinatal. 18: 2-10.
34 Rodkey, F. L.; O'Neal, J. D.; Collison, H. A. (1969) Oxygen and carbon monoxide equilibria of human adult
3 5 hemoglobin at atmospheric and elevated pressure. Blood 33: 57-65.
36 Roughton, F. J. W. (1970) The equilibrium of carbon monoxide with human hemoglobin in whole blood.
37 In: Coburn, R. F., ed. Biological effects of carbon monoxide. Ann. N. Y. Acad. Sci. 174: 177-188.
38 Saltzman, B. E.; Fox, S. H. (1986) Biological significance of fluctuating concentrations of carbon monoxide.
39 Environ. Sci. Technol. 20: 916-923.
40 Schoenfisch, W. H.; Hoop, K. A.; Struelens, B. S. (1980) Carbon monoxide absorption through the oral and nasal
41 mucosae of cynomolgus monkeys. Arch. Environ. Health 35: 152-154.
42 Seki, T.; Naruse, M.; Naruse, K.; Yoshimoto, T.; Tanabe, A.; Imaki, T.; Hagiwara, H.; Hirose, S.; Demura, H.
43 (1997) Interrelation between nitric oxide synthase and heme oxygenase in rat endothelial cells. Eur. J.
44 Pharmacol. 331:87-91.
45 Selvakumar, S.; Sharan, M.; Singh, M. P. (1992) Mathematical model for the exchange of gases in the lungs with
46 special reference to carbon monoxide. Med. Biol. Eng. Comput. 30: 525-532.
47 Selvakumar, S.; Sharan, M.; Singh, M. P. (1993) A mathematical model for the elimination of carbon monoxide in
48 humans. J. Theor. Biol. 162: 321-336.
49 Sharan, M.; Selvakumar, S.; Singh, M. P. (1990) Mathematical model for the computation of alveolar partial
50 pressure of carbon monoxide. Int. J. Biomed. Comput. 26: 135-147.
51 Shimazu, T.; Ikeuchi, H.; Hubbard, G. B.; Langlinais, P. C.; Mason, A. D., Jr.; Pruitt, B. A., Jr. (1990) Smoke
52 inhalation injury and the effect of carbon monoxide in the sheep model. J. Trauma 30: 170-175.
53 Siggaard-Andersen, J.; Bonde Petersen, F.; Hansen, T. L; Mellemgaard, K. (1968) Plasma volume and vascular
54 permeability during hypoxia and carbon monoxide exposure. Scand. J. Clin. Lab. Invest. Suppl. 103: 39-48.
October 1999 5-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 Singh, J.; Moore-Cheatum, L. (1993) Gestational protein deficiency enhances fetotoxicity of carbon monoxide.
2 In: Keen, C. L.; Bendich, A.; Willhite, C. C., eds. Maternal nutrition and pregnancy outcome. Ann. N. Y.
3 Acad. Sci. 678: 366-368.
4 Singh, M. P.; Sharan, M; Selvakumar, S. (1991) A mathematical model for the computation of
5 carboxyhaemoglobin in human blood as a function of exposure time. Philos. Trans. R. Soc. Lond.
6 B 334: 135-147.
7 Singh, J.; Smith, C. B.; Moore-Cheatum, L. (1992) Additivity of protein deficiency and carbon monoxide on
8 placental carboxyhemoglobin in mice. Am. J. Obstet. Gynecol. 167: 843-846.
9 Singh, J.; Aggison, L., Jr.; Moore-Cheatum, L. (1993) Teratogenicity and developmental toxicity of carbon
10 monoxide in protein-deficient mice. Teratology 48: 149-159.
11 Singleton, G. J.; Olsen, C. R.; Smith, R. L. (1972) Correction for mechanical dead space in the calculation of
12 physiological dead space. J. Clin. Invest. 51: 2768-2772.
13 Smith, M. V. (1990) Comparing solutions to the linear and nonlinear CFK equations for predicting COHb
14 formation. Math. Biosci. 99: 251-263.
15 Smith, M. V.; Hazucha, M. J.; Benignus, V. A.; Bromberg, P. A. (1994) Effect of regional circulation patterns on
16 observed HbCO levels. J. Appl. Physiol. 77: 1659-1665.
17 Sokal, J. A.; Majka, J.; Palus, J. (1984) The content of carbon monoxide in the tissues of rats intoxicated with
18 carbon monoxide in various conditions of acute exposure. Arch. Toxicol. 56: 106-108.
19 Sokal, J.; Majka, J.; Palus, J. (1986) Effect of work load on the content of carboxymyoglobin in the heart and
20 skeletal muscles of rats exposed to carbon monoxide. J. Hyg. Epidemiol. Microbiol. Immunol. 30: 57-62.
21 Solanki, D. L.; McCurdy, P. R.; Cuttitta, F. F.; Schechter, G. P. (1988) Hemolysis in sickle cell disease as measured
22 by endogenous carbon monoxide production: a preliminary report. Am. J. Clin. Pathol. 89: 221-225.
23 Song, S.-Y.; Okeda, R.; Funata, N.; Higashino, F. (1983) An experimental study of the pathogenesis of the selective
24 lesion of the globus pallidus in acute carbon monoxide poisoning in cats: with special reference to the
25 chronologic change in the cerebral local blood flow. Acta Neuropathol. 61: 232-238.
26 Stewart, R. D.; Peterson, J. E.; Baretta, E. D.; Bachand, R. T.; Hosko, M. J.; Herrmann, A. A. (1970) Experimental
27 human exposure to carbon monoxide. Arch. Environ. Health 21: 154-164.
28 Stewart, R. D.; Peterson, J. E.; Fisher, T. N.; Hosko, M. J.; Baretta, E. D.; Dodd, H. C.; Herrmann, A. A. (1973)
29 Experimental human exposure to high concentrations of carbon monoxide. Arch. Environ. Health 26: 1-7.
30 Stokes, D. L.; Maclntyre, N. R.; Nadel, J. A. (1981) Nonlinear increases in diffusing capacity during exercise by
31 seated and supine subjects. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 51: 858-863.
32 Stone, J. R.; Marietta, M. A. (1994) Soluble guanylate cyclase from bovine lung: activation with nitric oxide and
33 carbon monoxide and spectral characterization of the ferrous and ferric states. Biochemistry 33: 5636-5640.
34 Stork, R. L.; Bredle, D. L.; Chapler, C. K.; Cain, S. M. (1988) Regional hemodynamic responses to hypoxia in
35 polycythemic dogs. J. Appl. Physiol. 65: 2069-2074.
36 Sugi, K.; Theissen, J. L.; Traber, L. D.; Herndon, D. N.; Traber, D. L. (1990) Impact of carbon monoxide on
37 cardiopulmonary dysfunction after smoke inhalation injury. Circ. Res. 66: 69-75.
38 Tamayo, L.; Lopez-Lopez, J. R.; Castaneda, J.; Gonzalez, C. (1997) Carbon monoxide inhibits hypoxic pulmonary
39 vasoconstriction in rats by a cGMP-independent mechanism. Pfluegers Arch. 434: 698-704.
40 Thorn, S. R. (1990) Carbon monoxide-mediated brain lipid peroxidation in the rat. J. Appl. Physiol. 68: 997-1003.
41 Thorn, S. R. (1992) Dehydrogenase conversion to oxidase and lipid peroxidation in brain after carbon monoxide
42 poisoning. J. Appl. Physiol. 73: 1584-1589.
43 Thorn, S. R. (1993) Leukocytes in carbon monoxide-mediated brain oxidative injury. Toxicol. Appl. Pharmacol.
44 123:234-247.
45 Thorn, S. R.; Ischiropoulos, H. (1997) Mechanism of oxidative stress from low levels of carbon monoxide.
46 Cambridge, MA: Health Effects Institute; research report no. 80.
47 Thorn, S. R.; Ohnishi, S. T.; Ischiropoulos, H. (1994) Nitric oxide released by platelets inhibits neutrophil B2
48 integrin function following acute carbon monoxide poisoning. Toxicol. Appl. Pharmacol. 128: 105-110.
49 Thorn, S. R.; Xu, Y. A.; Ischiropoulos, H. (1997) Vascular endothelial cells generate peroxynitrite in response to
50 carbon monoxide exposure. Chem. Res. Toxicol. 10: 1023-1031.
51 Thorn, S. R.; Garner, S.; Fisher, D.; Ischiropoulos, H. (1998) Vascular nitrosative stress from CO exposure.
52 In: Program and abstracts [of the] Undersea and Hyperbaric Medical Society annual scientific meeting:
53 pre- and post-courses; May; Seattle, WA. Undersea Hyperbaric Med. 25(suppl.): 47.
54 Thorn, S. R.; Ohnishi, S. T.; Fisher, D.; Xu, Y. A.; Ischiropoulos, H. (1999a) Pulmonary vascular stress from
55 carbon monoxide. Toxicol. Appl. Pharmacol. 154: 12-19.
October 1999 5-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 Thorn, S. R.; Fisher, D.; Xu, Y. A.; Garner, S.; Ischiropoulos, H. (1999b) Role of nitric oxide-derived oxidants in
2 vascular injury from carbon monoxide in the rat. Am. J. Physiol. 276: H984-H992.
3 Thomsen, H. K. (1974) Carbon monoxide-induced atherosclerosis in primates: an electron-microscopic study on the
4 coronary arteries ofMacaca irus monkeys. Atherosclerosis 20: 233-240.
5 Tikuisis, P. (1996) Modeling the uptake and elimination of carbon monoxide. In: Penney, D. G., ed. Carbon
6 monoxide. Boca Raton, FL: CRC Press; pp. 45-67.
7 Tikuisis, P.; Buick, F.; Kane, D. M. (1987a) Percent carboxyhemoglobin in resting humans exposed repeatedly to
8 1,500 and 7,500 ppm CO. J. Appl. Physiol. 63: 820-827.
9 Tikuisis, P.; Madill, H. D.; Gill, B. J.; Lewis, W. F.; Cox, K. M.; Kane, D. M. (1987b) A critical analysis of the use
10 of the CFK equation in predicting COHb formation. Am. Ind. Hyg. Assoc. J. 48: 208-213.
11 Tikuisis, P.; Kane, D. M.; McLellan, T. M.; Buick, F.; Fairburn, S. M. (1992) Rate of formation of
12 carboxyhemoglobin in exercising humans exposed to carbon monoxide. J. Appl. Physiol. 72: 1311-1319.
13 Trischmann, U.; Klockner, U.; Isenberg, G.; Utz, J.; Ullrich, V. (1991) Carbon monoxide inhibits
14 depolarization-induced Ca rise and increases cyclic GMP in visceral smooth muscle cells. Biochem.
15 Pharmacol. 41:237-241.
16 Turner, D. M.; Lee, P. N.; Roe, F. J. C.; Gough, K. J. (1979) Atherogenesis in the White Carneau pigeon: further
17 studies of the role of carbon monoxide and dietary cholesterol. Atherosclerosis 34: 407-417.
18 Tyuma, I.; Ueda, Y.; Imaizumi, K.; Kosaka, H. (1981) Prediction of the carbonmonoxyhemoglobin levels during
19 and after carbon monoxide exposures in various animal species. Jpn. J. Physiol. 31: 131-143.
20 U.S. Environmental Protection Agency. (1978) Altitude as a factor in air pollution. Research Triangle Park, NC:
21 Office of Research and Development, Environmental Criteria and Assessment Office; report no. EPA
22 600/9-78-015. Available from: NTIS, Springfield, VA; PB-285645.
23 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
24 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
25 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
26 U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
27 monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
28 Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available from:
29 NTIS, Springfield, VA; PB93-157717.
30 Utz, J.; Ullrich, V. (1991) Carbon monoxide relaxes ileal smooth muscle through activation of guanylate cyclase.
31 Biochem. Pharmacol. 41: 1195-1201.
32 Venkatram, A.; Louch, R. (1979) Evaluation of CO air quality criteria using a COHb model. Atmos. Environ.
33 13:869-872.
34 Wagner, J. A.; Horvath, S. M.; Dahms, T. E. (1975) Carbon monoxide elimination. Respir. Physiol. 23: 41-47.
3 5 Wang, R. (1998) Resurgence of carbon monoxide: an endogenous gaseous vasorelaxing factor. Can. J. Physiol.
36 Pharmacol. 76: 1-15.
37 Wang, R.; Wu, L.; Wang, Z. (1997a) The direct effect of carbon monoxide onKCa channels in vascular smooth
38 muscle cells. Pfluegers Arch. 434: 285-291.
39 Wang, R.; Wang, Z.; Wu, L. (1997b) Carbon monoxide-induced vasorelaxation and the underlying mechanisms.
40 Br. J. Pharmacol. 121: 927-934.
41 Webster, W. S.; Clarkson, T. B.; Lofland, H. B. (1968) Carbon monoxide-aggravated atherosclerosis in the squirrel
42 monkey. Exp. Mol. Pathol. 13: 36-50.
43 Werner, B.; Lindahl, J. (1980) Endogenous carbon monoxide production after bicycle exercise in healthy subjects
44 and in patients with hereditary spherocytosis. Scand. J. Clin. Lab. Invest. 40: 319-324.
45 West, J. B. (1990a) Ventilation/blood flow and gas exchange. 5th ed. Oxford, United Kingdom: Blackwell
46 Scientific.
47 West, J. B. (1990b) Respiratory physiology - the essentials. 4th ed. Baltimore, MD: Williams & Wilkins.
48 Wittenberg, B. A.; Wittenberg, J. B. (1993) Effects of carbon monoxide on isolated heart muscle cells. Cambridge,
49 MA: Health Effects Institute; research report no. 62.
50 Woehlck, H. J.; Dunning, M. B., Ill; Kulier, A. H.; Sasse, F. J.; Nithipataikom, K.; Henry, D. W. (1997a) The
51 response of anesthetic agent monitors to trifluoromethane warns of the presence of carbon monoxide from
52 anesthetic breakdown. J. Clin. Monit. 13: 149-155.
53 Woehlck, H. J.; Dunning, M., Ill; Connolly, L. A. (1997b) Reduction in the incidence of carbon monoxide
54 exposures in humans undergoing general anesthesia. Anesthesiology 87: 228-234.
October 1999 5-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wyman, J.; Bishop, G.; Richey, B.; Spokane, R.; Gill, S. (1982) Examination of Haldane's first law for the partition
2 of CO and O2 to hemoglobin A0. Biopolymers 21: 1735-1747.
3 Yamaya, M; Sekizawa, K.; Ishizuka, S.; Monma, M.; Mizuta, K.; Sasaki, H. (1998) Increased carbon monoxide in
4 exhaled air of subjects with upper respiratory tract infections. Am. J. Respir. Crit. Care Med. 158:311-314.
5 Yoneda, I.; Watanabe, Y. (1997) Comparisons of altitude tolerance and hypoxia symptoms between nonsmokers
6 and habitual smokers. Aviat. Space Environ. Med. 68: 807-811.
7 Zayasu, K.; Sekizawa, K.; Okinaga, S.; Yamaya, M.; Ohrui, T.; Sasaki, H. (1997) Increased carbon monoxide in
8 exhaled air of asthmatic patients. Am. J. Respir. Crit. Care Med. 156: 1140-1143.
9 Zhu, N.; Weiss, H. R. (1994) Effect of hypoxic and carbon monoxide-induced hypoxia on regional myocardial
10 segment work and O2 consumption. Res. Exp. Med. 194: 97-107.
11
October 1999 5-48 DRAFT-DO NOT QUOTE OR CITE
-------
i 6. HEALTH EFFECTS OF EXPOSURE TO
2 AMBIENT CARBON MONOXIDE
3
4
5 6.1 INTRODUCTION
6 Concerns about the potential health effects of exposure to carbon monoxide (CO) have
7 been addressed in extensive controlled-exposure studies and more limited population-exposure
8 studies. Under varied experimental protocols, considerable information has been obtained on the
9 toxicity of CO, its direct effects on blood and tissues, and the manifestations of these effects in
10 the form of changes in organ function. This chapter assesses the current understanding of health
11 effects that may occur in individuals breathing CO in ambient air. Additional discussion of
12 studies reporting CO-associated health effects can be found in the previous document,
13 Air Quality Criteria for Carbon Monoxide (U.S. Environmental Protection Agency, 1991) and in
14 a number of excellent reviews (Kleinman, 1992; Bascom et al., 1996; Penney, 1996a).
15 Although evidence from laboratory animal studies indicates that CO can adversely affect
16 the cardiovascular and nervous systems of both mature animals and developing offspring, the
17 concentrations of CO used during exposure and consequent levels of carboxyhemoglobin
18 (COHb) saturation are much higher than typically would be experienced by ambient exposures of
19 humans. The laboratory animal studies, therefore, must be interpreted with caution. However,
20 they can be useful for exploring the properties and possible mechanisms of an effect much more
21 thoroughly and extensively than is possible in humans. An effort is made in this chapter to
22 compare the health effect levels for CO found in laboratory animal and human controlled-
23 exposure studies.
24 A review of the health effects literature on CO since the last assessment was published in
25 1991 finds many published articles on CO poisoning, possibly reflecting increased media
26 attention to this topic. Many of these articles, however, reported effects at CO levels far higher
27 than in ambient air. Severe effects from acute exposure to high levels of CO are not directly
28 germane to problems associated with exposure to current ambient levels of CO and, thus, are not
29 discussed in detail in this chapter. They are, however, mentioned briefly in the summary of this
30 and the following chapter to a snapshot of the full range of CO toxicity and to provide public
October 1999 6-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 health information about potential effects of accidental exposure to CO, particularly those
2 exposures occurring indoors.
3 The next section of this chapter (Section 6.2) addresses cardiovascular effects of CO. The
4 section begins with a discussion of epidemiologic studies (Section 6.2.1), because of their
5 potential importance in assessing community health effects of ambient CO exposure.
6 Section 6.2.1 emphasizes studies of ambient CO and heart disease exacerbation, because
7 short-term ambient CO concentrations have been associated more frequently with such
8 exacerbation than with other health-related endpoints. For purposes of continuity in the
9 epidemiologic discussion, Section 6.2.1 also addresses studies that have evaluated ambient CO in
10 relation to health indices other than heart disease exacerbation. These are daily mortality,
11 incidence of low birth weight, and daily frequency of respiratory illness.
12 The remainder of Section 6.2 (Section 6.2.2) summarizes controlled-exposure studies of
13 CO effects on maximal exercise performance and in subjects with reproducible exercise-induced
14 angina. In 1991, these studies formed a major scientific basis for U.S. Environmental Protection
15 Agency (EPA) review of the levels and adequacy of existing national ambient air quality
16 standards (NAAQS) for CO. Although these areas have changed little since 1991, controlled-
17 exposure studies continue to provide the most quantitative evidence on low-level CO effects in
18 humans.
19 Next in importance to cardiovascular effects, but of questionable impact for the young and
20 healthy population, are studies of neurobehavioral effects, which had earlier provided the
21 scientific basis for the first CO NAAQS. Subsequent assessments of the neurobehavioral
22 literature, however, have raised questions about both the methods and results of the early studies.
23 Articles published since the last assessment in 1991 have been mostly reviews that attempt to
24 explain the equivocal results found at low COHb levels, and to provide a physiological basis for
25 behavioral effects. The section on neurobehavioral effects (Section 6.3) illustrates the difficulty
26 in studying an organ system with exquisite compensatory responses to a reduced oxygen supply
27 (hypoxia).
28 The rest of the chapter summarizes current knowledge about developmental toxicity
29 (Section 6.4), acute pulmonary effects (Section 6.5), other systemic effects of CO (Section 6.6),
30 physiologic responses to CO exposure (Section 6.7), and combined exposure of CO with other
31 pollutants, drugs, and environmental factors (Section 6.8). Little new information has been
October 1999 6-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 published on these areas of CO toxicity, and their summaries remain essentially the same as
2 published in the previous criteria document (U.S. Environmental Protection Agency, 1991).
3 Significant new studies have been added to the summaries. No new published studies are known
4 to draw into question the conclusions drawn from the previous literature review of these topics.
5 Finally, a summary section (Section 6.9) provides a concise review of the key human health
6 effects most clearly associated with exposure to ambient CO.
7
8
9 6.2 CARDIOVASCULAR EFFECTS
10 Cardiovascular disease (CVD) is the leading cause of death in the United States (American
11 Heart Association, 1997; U.S. Centers for Disease Control and Prevention, 1997). An estimated
12 58 million people in the United States (-20% of the population) have one or more types of CVD
13 (American Heart Association, 1997). For the major diseases within this category of total CVD,
14 about 50 million Americans have high blood pressure, 14 million have ischemic (coronary) heart
15 disease, 5 million have heart failure, 4 million have cerebrovascular disease (stroke), and
16 1.8 million have rheumatic fever or heart disease. Because the numbers of affected people are so
17 high, even relatively small percentage increases in cardiovascular mortality or morbidity in the
18 population could have a large impact on both public health and health care costs. In the
19 following discussion, the effects of CO on potentially susceptible population groups are explored
20 through epidemiologic and controlled laboratory studies (for further discussion of subpopulations
21 at risk, see Chapter 7).
22
23 6.2.1 Epidemiologic Studies
24 6.2.1.1 Introduction
25 In recent years, many epidemiologic studies have shown associations of short-term ambient
26 air pollution exposure with mortality, exacerbation of preexisting illness, and physiologic
27 changes. These studies have increased concern as to whether ambient air pollution exposure can
28 promote, and perhaps even produce, harmful health outcomes, even when ambient pollutant
29 concentrations do not exceed current U.S. air quality standards.
October 1999 6-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 The epidemiologic database regarding short-term ambient air pollution exposure is growing
2 rapidly, and its interpretation is changing over time. As recently as the mid-1990s, many
3 epidemiologic studies had reported associations of mortality and exacerbation of pre-existing
4 disease with ambient levels of paniculate matter (PM), but relatively few had investigated or
5 reported such associations with gaseous pollutants, including CO. Since then, however,
6 investigators have given more thorough consideration to PM and gaseous pollutants, and have
7 frequently observed positive, statistically significant associations of harmful effects with both.
8 Thus, although associations of PM with harmful effects continue to be observed, the role of
9 gaseous pollutants appears to be stronger than it did in previous evaluations.
10 Another important recent change relates to the biological plausibility of observed
11 epidemiologic associations of ambient air pollution with harmful health outcomes. A few years
12 ago, it was frequently argued that associations of low-level ambient air pollution with harmful
13 health outcomes were biologically implausible. Now, with a considerable amount of new
14 experimental evidence in hand, and after much new and sharpened thought on this issue, it is
15 possible to hypothesize ways by which ambient exposure to multiple air pollutants could
16 plausibly be involved in the complex chain of biological events leading to harmful health effects
17 in the human population.
18 In epidemiologic studies of ambient air pollution, small positive estimates of air pollutant
19 health effects have been observed quite consistently. These estimates have frequently been
20 statistically significant at a = 0.05. If ambient air pollution actually promotes or produces
21 harmful health effects, relatively small effect estimates would generally be expected on
22 biological and epidemiologic grounds. Also, the magnitudes and significance levels of observed
23 air pollution-related effects estimates have varied somewhat from place to place. This would
24 also be expected if the observed epidemiologic associations denote actual effects, because
25 populations differ in characteristics that could affect susceptibility to air pollution health effects.
26 These characteristics include demographic and socioeconomic factors, underlying health status,
27 indoor-outdoor activities, diet, medical care systems and access to them, and exposure to risk
28 factors other than ambient air pollution, such as extreme weather conditions.
29 Thus, though it has been argued that epidemiologic studies are trustworthy only if they
30 show relatively large effects estimates (e.g., large relative risks), and that the observed effects
31 estimates for ambient air pollution are not sufficiently constant across epidemiologic studies,
October 1999 6-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 these arguments have only limited weight in relation to ambient air pollution studies. It should
2 also be borne in mind that in any large population exposed to ambient air pollution, even a small
3 relative risk for a prevalent health disorder could calculate to a substantial public health burden
4 attributable to air pollution exposure.
5 At the same time, important biological, epidemiologic, and statistical uncertainties remain
6 in the current epidemiologic database for ambient CO and other air pollutants. Biologically and
7 epidemiologically, it has not been confirmed that the magnitudes of observed statistical health
8 effects estimates for ambient CO are quantitatively commensurate with actual underlying
9 population susceptibility to CO exposure. Also, it has not been confirmed that the observed
10 spatial variation in air pollution effects estimates reflects actual cross-population differences in
11 susceptibility.
12 The ambient atmosphere contains numerous air pollutants, and there is increasing
13 realization that health effects associated statistically with any single pollutant may actually be
14 mediated by multiple components of the complex ambient mix. Specific attribution of effects to
15 any single pollutant may therefore convey an overly simplistic notion of biological reality.
16 Carbon monoxide is one of many air pollutants generated by combustion sources, including
17 mobile sources. These pollutants include CO, PM, and nitrogen oxides, which have been
18 considered in epidemiologic studies to date. These pollutants also include numerous volatile or
19 semivolatile organic compounds, which have not yet been systematically considered in relation to
20 the non-cancer health outcomes usually associated with exposure to criteria air pollutants.
21 In available epidemiologic studies, harmful health outcomes are frequently associated with
22 multiple combustion-related or mobile source-related air pollutants. Many investigators have
23 raised the possibility that CO may be a surrogate or marker for a larger subset of the overall
24 ambient air pollution mix. Some investigators have argued that CO is a biologically passive
25 surrogate, and that statistical associations of CO with health effects actually reflect effects of
26 pollutants other than CO. However, most investigators have reserved judgment on this issue, and
27 several investigators have emphasized the need for further research on CO.
28 The health effects of long-term exposure to CO and other air pollutants have received little
29 attention in epidemiologic studies, and are not well understood as yet. Health effects of long-
30 term exposure at present-day ambient pollutant levels, which are generally lower than past levels,
31 are especially uncertain. Also, it is not known whether long-term exposure to ambient CO plays
October 1999 6-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 a role in the induction (incidence) of new cases of illness. Further research on long-term health
2 effects will be important both to understand better how the public may be protected against such
3 effects, and to provide an epidemiologic context in which to assess the plausibility, accuracy, and
4 validity of the short-term effects estimates.
5 Important statistical uncertainties also remain in the available epidemiologic database for
6 CO and other air pollutants. Many of these are especially pertinent to daily time series studies,
7 which are the majority of the epidemiologic studies available for ambient CO, and which are a
8 large part of the epidemiologic database for other criteria air pollutants. These statistical
9 uncertainties, coupled with existing biological and epidemiologic uncertainties, raise some
10 questions as to the quantitative accuracy of the pollutant effects estimates themselves, and
11 perhaps in some cases as to their qualitative validity. This, in turn, contributes to scientific
12 uncertainty regarding estimates of potential public health impacts of current U.S. CO levels.
13 Unresolved statistical issues include those described below.
14 (1) Large proportional differences in air pollutant effects estimates with different model
15 specifications. As discussed above, small health effects estimates have generally been
16 observed for ambient air pollutants, and small effects would indeed be expected on
17 biological and epidemiologic grounds. At the same time, because the effects estimates are
18 small, they are subject to large proportional differences with different choices of statistical
19 models. They are also subject to changes in statistical significance level. As the following
20 sections show, for CO and other air pollutants, effects estimates are sensitive to many model
21 selection choices: (1) for different metrics for the same pollutant; (2) for different choices of
22 covariates in single-pollutant models; (3) for a given pollutant in single-pollutant and multi-
23 pollutant models; (4) for a given pollutant in different multi-pollutant models; (5) for
24 different lag or moving-average structures for CO, co-pollutants, and covariates; (6) for the
25 functional form of the modeled concentration-response relationship; and (7) with different
26 adjustments for autocorrelation and overdispersion in the statistical estimation procedure.
27 It seems quite likely that pollutant effects estimates in multi-pollutant models are more
28 biologically and epidemiologically sound than those in single-pollutant models. However,
29 even this has not been demonstrated, and scientific consensus as to optimal modeling
30 strategies for time series air pollution studies has not yet been achieved. Therefore, the
October 1999 6-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 choice of effects estimates to employ in risk assessments for short-term ambient air pollution
2 effects remains open to question.
3 (2) Potential confounding of air pollution and weather effects. Meteorologic events and
4 ambient air pollutant concentrations may be highly correlated on short time scales, even
5 after longer time trends have been filtered. It is essential to model joint effects of air
6 pollution and weather with great care. Such joint modeling has been conducted only rarely
7 in time series studies of ambient air pollution. To date, simple additive or proportional
8 assumptions have generally been made in modeling health effects of air pollutants and
9 weather. These assumptions are not fully adequate, largely because health effects estimates
10 for air pollutants are small and subject to large proportional differences with different model
11 specifications.
12 (3) Insufficient reporting of statistical uncertainty. In available studies, statistical uncertainty
13 has generally been assessed rather superficially, without formal consideration of the model
14 tuning performed by the investigators. For example, lag times and averaging times for air
15 pollutant metrics have usually been selected to maximize statistical effects estimates for
16 pollutants. This may have led not only to unrealistically large reported effects estimates, but
17 also to unduly narrow confidence intervals for these estimates. In future studies, uncertainty
18 arising from model tuning should be more carefully assessed. In this effort, resampling or
19 simulation procedures, which would recreate the entire model estimation process, should be
20 considered.
21 (4) Health effects averaged over extended time periods. None of the available time series air
22 pollution studies is capable of assessing the incremental effect of pollutants over extended
23 time periods. For example, current models cannot determine whether reduction in pollution
24 will decrease monthly rates of hospital admissions or mortality, even if they predict a
25 reduction of admissions on days with low pollution. This public health-related issue cannot
26 be addressed by daily time series analysis, using only admission or mortality counts.
27 In future studies, investigators could consider time-averaged health effects over, say, one
28 month or three months, in relation to pollution exposure metrics for the corresponding
29 periods. Consideration of extended time-averaged health effects would also tend to
30 attenuate any short-term "harvesting" that might be observed in daily analyses. (In time
31 series studies of air pollution, "harvesting" is a short-term elevation in the frequency of a
October 1999 6-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 health outcome during or just after a short period of elevated ambient air pollution, followed
2 quickly by a short-term reduction in frequency of the same outcome below baseline
3 frequency, then by a return to baseline. It has been argued that presence of harvesting would
4 suggest that elevated air pollution exposure hastens occurrence of the health outcome by a
5 short time, but brings about little or no net increase in occurrence of the outcome. It also has
6 been argued that absence of harvesting would suggest that without the elevated air pollution
7 exposure, the outcome might have been delayed for a long time, or might not have occurred
8 at all.)
9 To date, short-term fluctuations in ambient CO have been examined in time series studies
10 of daily exacerbations of preexisting cardiovascular disease, mortality, and changes in respiratory
11 illness frequency. Subchronic ambient CO levels, averaged over about 3 mo, also have been
12 examined in relation to frequency of low birth weight. Associations have been observed most
13 frequently for short-term ambient CO fluctuations and exacerbation of heart disease
14 (a subcategory of cardiovascular disease), as usually indexed by daily hospital admissions. These
15 associations are biologically plausible to some degree. In heart disease patients, the coronary
16 arteries usually are narrowed and unable to dilate normally. Increased hypoxic stress resulting
17 from small increases in COHb saturation could conceivably lead to clinically apparent worsening
18 of their illness.
19 If observed associations of ambient CO with heart disease exacerbation prove to be real and
20 specific to CO, they would be of genuine public health concern. In the U.S. in 1996, there were
21 about 4,239,000 hospital discharges with heart disease as the first-listed diagnosis (Graves and
22 Owings, 1998). Among these, about 2,262,000 (53.4%) were for ischemic heart disease,
23 825,000 (19.5%) for myocardial infarction or heart attack (a subcategory of ischemic heart
24 disease), 870,000 (20.5%) for congestive heart failure, and 618,000 (14.6%) for cardiac
25 dysrhythmias. Also, there were 733,361 deaths caused by heart disease (Peters et al., 1998).
26 Even a small percentage reduction in admissions or deaths caused by heart disease would result
27 in a large number of avoided cases.
28 Even so, fluctuations in ambient CO levels would be expected to produce only very small
29 changes in COHb saturation, and only small changes in tissue oxygenation. Thus, the observed
30 associations of ambient CO with heart disease exacerbation remain difficult to rationalize
31 pathophysiologically. Also, such exacerbation has been associated not only with ambient CO,
October 1999 6-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 but also with other combustion-related ambient pollutants such as NO2 and PM. Thus, the extent
2 to which such exacerbation is truly attributable to ambient CO exposure is not yet clear.
3 Studies of short-term ambient CO levels and daily mortality have yielded mixed results.
4 Observed associations of ambient CO with mortality are of potential public health concern
5 because associations of ambient CO with heart disease exacerbation have been observed
6 frequently, and because heart disease is the leading cause of death in the United States. For
7 example, in the United States in 1996, 733,361 (31.7%) of all 2,314,690 deaths were caused by
8 heart disease (Peters et al., 1998). Again, however, there is not a firm pathophysiologic basis for
9 the observed CO-mortality associations, and the degree of correspondence between statistical
10 effects estimates in populations and real effects in individuals has not yet been determined.
11 Two available U.S. studies suggest that subchronic ambient CO exposure may be
12 associated with increased incidence of low birth weight. These studies raise concern, because
13 fetal hemoglobin binds CO somewhat more strongly than does adult hemoglobin, and because
14 increased COHb saturation would be expected to impair tissue O2 delivery more in the fetus than
15 in the child or the adult (Longo, 1976). At the same time, these studies are not conclusive. They
16 are subject to potential confounding by unmeasured factors, such as maternal smoking, that are
17 known to influence birth weight. Also, outdoor CO levels may be correlated with indoor levels
18 of CO and other pollutants, which could be higher than outdoor levels, and which were not
19 measured in these studies. Common socioeconomic factors could be associated with both
20 ambient CO levels and such potential confounding variables.
21 Some investigators have observed associations of short-term fluctuations in ambient CO
22 with daily frequency of respiratory illness. In most cases, exacerbation of preexisting respiratory
23 illness has been assessed, though some cases of acute respiratory infection constitute occurrence
24 of new illness, especially in young people. The biological plausibility of these associations is
25 tenuous, because there are as yet no demonstrated mechanisms by which ambient-level CO
26 exposure could produce or promote harmful respiratory effects. Also, in epidemiologic studies to
27 date, associations of ambient CO with respiratory disease frequency have been observed less
28 consistently than with heart disease exacerbation. Nevertheless, the observed associations of
29 ambient CO with respiratory disease frequency are worthy of discussion, because they suggest
30 that ambient CO exposure may not be specifically linked with heart disease.
October 1999 6-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Overall, then, recent epidemiologic studies have raised the level of concern regarding
2 potential harmful effects of present-day ambient CO exposure, especially with respect to heart
3 disease exacerbation, possibly with respect to mortality and low birth weight, and even,
4 conceivably, with respect to increased frequency of respiratory illness. However, further research
5 on the health effects of long-term and short-term ambient CO exposure is strongly warranted.
6 This research should assess effects not only of changes in COHb levels and tissue oxygenation,
7 but also of CO dissolved in the blood, and of CO in tissues other than blood. It should assess
8 effects of CO alone and as a component of the complex ambient air pollution mix.
9 Individual epidemiologic studies that have considered ambient CO are discussed below in
10 the following order of health outcomes: (1) daily exacerbation of heart disease, (2) daily
11 mortality, (3) incidence of low birth weight, and (4) daily frequency of respiratory illness.
12
13 6.2.1.2 Ambient Carbon Monoxide and Exacerbation of Heart Disease
14 Recent epidemiologic studies in the United States, Canada, and Europe suggest that short-
15 term variation in ambient CO levels is associated with daily hospital admissions for heart
16 disease. In several studies of such admissions, effects of lagged ambient air pollutant levels have
17 been examined, in addition to effects of air pollutant levels on the same day as the admissions
18 (0-day lag). When averaging times for ambient pollution metrics have been 24 h or shorter,
19 modeled effects of CO generally have been strongest with a 0-day lag. When averaging times
20 have been longer than 24 h, CO effects generally have been strongest when the last day of the
21 averaging period is lagged 0 days.
22 As mentioned above, observed associations of ambient CO with heart disease exacerbation
23 have some biological plausibility and are of potential public health concern. However, these
24 associations should be interpreted cautiously. The average daily maximum CO concentrations
25 measured by stationary monitors in the epidemiologic studies have generally been low (<5 ppm).
26 Any increase over endogenous COHb levels produced by a 1-h exposure to <10 ppm exogenous
27 CO, for example, would be difficult even to measure accurately. Even 8 h of exposure to 10 ppm
28 CO with light to moderate exercise (ventilation about 20 L/min) would be expected to produce
29 only a 1.0-2.0% increase in COHb saturation over the baseline level of about 0.5% COHb.
30 Pathophysiologically, it remains difficult to reconcile such small expected ambient CO-induced
31 changes in COHb saturation with overt exacerbation of heart disease in the community setting.
October 1999 6-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 Also, epidemiologic studies of ambient CO have relied heavily on pollutant measurements
2 at stationary outdoor monitors. CO levels at these monitors may not be well correlated with
3 personal CO exposures and doses, especially in compromised persons such as cardiac patients,
4 who spend much of their time indoors. Additional information on personal CO exposure and
5 other individual characteristics, such as active and passive smoking, would be highly desirable in
6 future epidemiologic studies of ambient CO. Furthermore, as discussed in Chapter 3, in most
7 U.S. metropolitan areas there is considerable spatial variation in simultaneous CO measurements
8 made at different monitoring sites. In most epidemiologic studies to date, exposure metrics have
9 consisted of CO measurements averaged across sites. The effects of such multi-site averaging of
10 CO levels on statistical health effects estimates are not yet well understood.
11 The diagnostic category "heart disease," is smaller and more specific than the category
12 "cardiovascular disease," which comprises heart disease and other disorders such as
13 cerebrovascular disease (including stroke), hypertension (high blood pressure), and diseases of
14 the blood vessels. To date, short-term variation in ambient CO levels has been more strongly
15 associated with heart disease exacerbation than exacerbation of other cardiovascular diseases.
16 At the same time, heart disease itself comprises several diagnostic subcategories, such as
17 ischemic heart disease (including myocardial infarction, coronary artery atherosclerosis, and
18 angina), heart failure, and disturbances of cardiac rhythm (dysrhythmias). The available
19 epidemiologic database is not entirely consistent regarding the specific heart disease
20 subcategories with which ambient CO levels are most strongly associated.
21 Morris et al. (1995) conducted a time-series analysis of ambient levels of gaseous air
22 pollutants (CO, nitrogen dioxide [NO2], sulfur dioxide [SO2], and ozone [O3]), in relation to
23 Medicare hospital admissions for congestive heart failure (CHF) in seven U.S. cities (Chicago,
24 IL; Detroit, MI; Houston, TX; Los Angeles, CA; Milwaukee, WI; New York City, NY; and
25 Philadelphia, PA) during the 4-year period, 1986 to 1989. The average daily maximum 1-h CO
26 levels (mean ± standard deviation [SD]) ranged from 1.8 (±1.0) ppm in Milwaukee to
27 5.6 (±1.7) ppm in New York City. The relative risk of admissions associated with a 10-ppm
28 increase in ambient CO ranged from 1.10 in New York City to 1.37 in Los Angeles. All seven
29 cities showed similar patterns of increasing admissions with increasing ambient CO
30 concentrations. In almost all analyses, CO effects were stronger on the day of admission
31 (0-day lag) than on previous days. In single-pollutant models, the effect of CO was statistically
October 1999 6-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 significant in all cities but Houston. In multi-pollutant models, the CO effect was significant in
2 all cities but New York and Milwaukee. In the transition from single-pollutant to multi-pollutant
3 models, CO effects were more stable, and more frequently retained statistical significance, than
4 effects of the other pollutants. Figure 2 from Morris et al. (1995), which shows exposure-
5 response curves for ambient CO levels and CHF admission rates in each city, is reproduced in
6 Figure 6-1. The authors estimated that each year, approximately 3,250 hospital admissions for
7 congestive heart failure could be attributed to the observed association with short-term elevations
8 in ambient CO levels.
10-
O
CD
A
c
o
'
8-
g
0
0.
O -,
o 7
o
o"
6-
r-
E 5
O 4-
Philadelphia
Milwaukee
Detroit
New York
4 6 8 10
Carbon Monoxide (ppm)
12
14
Figure 6-1. Lowess smoothing of the association between ambient levels of CO and
hospital admissions for congestive heart failure (CHF) among elderly people
after adjustment for temperature, month, day of week, and year, 1986 through
1989.
Source: Morris et al. (1995).
October 1999
6-12
DRAFT-DO NOT QUOTE OR CITE
-------
1 Schwartz and Morris (1995) examined air pollution and hospital admissions for heart
2 disease (congestive heart failure, ischemic heart disease, and cardiac dysrhythmias) in people
3 aged 65 years or older in the Detroit metropolitan area during the years 1986 through 1989.
4 Air quality data were available for PM10 on 82% of possible days and for O3 on 85% of possible
5 days. Data were available for SO2 and CO on all days during the study period. The mean PM10
6 was 48.0 mg/m3, the mean O3 was 41.0 ppb, the mean SO2 was 25.4 ppb, and the mean CO was
7 2.4 ppm. Data were analyzed with Poisson auto-regressive models, with independent variables
8 for temperature, dew point, month, and linear and quadratic time trends. For each pollutant, the
9 difference between the 25th and 75th percentiles of the distribution of ambient concentrations
10 during the study period (the "interquartile range") was calculated. Relative risks for the health
11 outcomes assessed were reported per increment in ambient concentration equal to each
12 pollutant's interquartile range.
13 Daily admissions for ischemic heart disease were associated with interquartile range
14 increases of 32 //g/m3 for PM10 (relative risk [RR] = 1.018; 95% confidence interval [CI] 1.005,
15 1.032), 18ppbforSO2(RR= 1.014; 95% CI 1.003, 1.026), and 1.28 ppm for CO (RR= 1.010;
16 95% CI 1.001, 1.018). However, both the CO and SO2 effects lost statistical significance after
17 controlling for PM10, whereas PM10 remained significant after controlling for the other pollutants.
18 Daily admissions for heart failure were associated independently and significantly with
19 interquartile range increases in both PM10 (RR = 1.024; 95% CI 1.004, 1.044) and CO
20 (RR = 1.022; 95% CI 1.010, 1.034). These results are summarized in Table 6-1. Effects of CO
21 were stronger on the day of admission (0-day lag) than on previous days. No pollutant was a
22 significant risk factor for dysrhythmia admissions, and O3 was not a significant risk factor for
23 admissions in any heart disease category.
24 Ambient CO could be a surrogate for general combustion-related or mobile-source air
25 pollution. In some locations, CO is highly correlated with PM during the winter months. The
26 ambient PM10 level was associated with heart disease exacerbation in Schwartz and Morris
27 (1995), but was not assessed in Morris et al. (1995). Also, ambient PM concentrations were
28 associated with hospital admissions for both heart failure and ischemic heart disease in Ontario,
29 Canada (Burnett et al., 1995).
30 Pantazopoulou et al. (1995) investigated cardiac and respiratory disease exacerbation in the
31 area of Athens, Greece, in relation to short-term ambient air pollution levels during 1988. The
October 1999 6-13 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6-1. MODELED RELATIVE RISKS OF INTERQUARTILE-RANGE
INCREASES IN AMBIENT POLLUTANT CONCENTRATIONS, FOR
DAILY HEART DISEASE ADMISSIONS IN PERSONS*65 YEARS OLD,
DETROIT, MI, 1986 TO 1989
Two-Pollutant Models
Single-Pollutant Models (including PM10 and CO or SO2)a
Disease Category CO SO2 O3 PM10 CO PM10 SO2
Congestive Heart Failure 1.022b 1.002 1.022 1.032b 1.022b 1.024b
(Table 6)
Ischemic Heart Disease 1.010b 1.014b 1.010 1.018b 1.006 1.016b
(Table 4) 1.015b 1.009
Dysrhythmias No Significant Pollutant Effects
Tor congestive heart failure, no model with PM10 and SO2 was reported.
bStatistically significant at a = 0.05.
Source: Modified from Schwartz and Morris (1995).
1 health outcomes analyzed were daily outpatient emergency department visits, and daily hospital
2 admissions, for cardiac and respiratory causes. Pollutant metrics were daily maximum 8-h
3 moving average CO, daily maximum hourly NO2, and 24-h average black smoke (BS), each
4 averaged over multiple monitoring stations in the Athens area. Mean levels of CO for all
5 available monitoring stations were 4.5 mg/m3 in winter and 3.4 mg/m3 in summer. Data were
6 analyzed with multiple linear regression, with adjustment for meteorological and chronological
7 variables. Only single-pollutant models were reported. The report did not mention lagged
8 analysis, so effects were presumably reported for pollutant levels on the same day as the visit or
9 admission. Separate analyses were conducted for "winter" (January 1 to March 21 and
10 September 22 to December 31) and "summer" (March 22 to September 21).
11 Pollutant effects were reported as the modeled increment in the number of visits or
12 admissions from the 5th to 95th percentile of the pollutant's concentration distribution during
13 1988. Winter and summer findings from Pantazopoulou et al. (1995) are summarized in
14 Table 6-2. No pollutant had a statistically significant effect on any health outcome in summer.
October 1999 6-14 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6-2. MODELED EFFECTS OF 5TH-TO 95TH-PERCENTILE INCREMENTS IN AMBIENT AIR
a
o
cr
VO
VO
VO
POLLUTANT CONCENTRATIONS ON DAILY NUMBERS OF CARDIAC AND RESPIRATORY
HOSPITAL ADMISSIONS IN SINGLE-POLLUTANT MODELS, BY SEASON, ATHENS, GREECE, 1998
Winter (January 1 -March 21 and September 22-December 3 1) Summer (March 22-September 21)
5th to 95th Increase in
Disease Percentile Increment Number of
Category Pollutant C"g/m3)a Admissions'3 95% CI
ON
i
O
Tj
H
6
o
0
H
O
Cardiac CO 8
Black
Smoke
NO2
Respiratory CO 8
Black
Smoke
NO2
a5th- to 95th-percentile increments in //g/m3
bThe average daily numbers of total cardiac
were 41.8 in winter and 41.4 in summer.
,200 11.0 3.4, 18.5
235 11.8 4.7, 18.9
76 11.2 3.3, 19.2
,200 9.9 2.5, 17.4
235 9.2 2.0, 16.3
76 10.4 2.7, 18.2
as measured at two sites in central Athens.
admissions were 73.8 in winter and 63.9 in summer.
5th to 95th Increase in
Percentile Number of
Increment (//g/m3)a Admissions'5 95% CI
5,600 1.4 -5.5, 8.4
134 3.0 -2.9,8.9
108 -0.06 -6.6,6.5
5,600 2.2 -5.3, 11.5
134 3.2 -4.0, 10.3
108 3.9 -5.9, 10.2
The average daily numbers of total respiratory admissions
Source: Modified from Pantazopoulou et al. (1995).
O
HH
H
W
-------
1 In contrast, all three pollutants had positive, statistically significant effects on both cardiac and
2 respiratory "emergency admissions" (unscheduled hospital admissions) in winter. Pollution
3 effects were stronger for admissions than outpatient visits, and accounted for a higher proportion
4 of respiratory admissions than cardiac admissions.
5 Separate analyses were conducted for CO effects when ambient CO levels were averaged
6 across two, three, or five monitoring stations (mean CO levels were 8.2, 6.5, and 5.2 mg/m3,
7 respectively). Interestingly, estimated 5th- to 95th-percentile increments in winter cardiac
8 admissions were nearly identical in these different analyses (11.0, 11.4, and 11.2 admissions,
9 respectively). Corresponding CO-related increments in winter respiratory admissions were also
10 similar (9.9, 12.1, and 11.3 admissions, respectively). Effects of BS varied somewhat more than
11 those of CO when averaged across different numbers of monitoring stations. These observations
12 suggest that spatial variation in ambient CO may be smaller in Athens than in many U.S.
13 locations, or that spatial heterogeneity in ambient CO levels may not always greatly distort
14 CO-related health effects estimates, even when CO levels are averaged across multiple
15 monitoring sites. The observations also underscore that further research is required on the effects
16 of different choices of ambient pollutant metrics in time series studies.
17 Schwartz (1997) examined relationships of short-term ambient air pollution levels with
18 cardiovascular hospital admissions in people at least 65 years old in Tucson, AZ, from 1988
19 through 1990. The analyzed range of diagnoses (International Classification of Diseases,
20 Version 9 [ICD-9] codes 390 to 429) included heart disease, hypertension, rheumatic fever, and
21 pulmonary circulatory disorders. It did not include cerebrovascular disease or peripheral blood
22 vessel diseases. Thus, heart disease constituted the large majority, but not all, of the analyzed
23 disorders. The author assessed effects of CO, SO2, O3, NO2, and PM10, as measured at a single
24 monitoring station. Poisson regression models included air pollution metrics, temperature,
25 humidity, and day of week. Nonparametric smoothing was used to adjust for long-term temporal
26 patterns. Exposure-to-admission lags of 0, 1, and 2 days were apparently assessed in different
27 models, and effects estimates of pollutant levels on the same day as admission (0-day lag) were
28 apparently reported. During the study, the median of maximum hourly CO concentrations and
29 the median of 24-h average PM10 concentrations were 3.03 ppm and 39 //g/m3, respectively. The
30 correlation between PM10 and SO2 was lower than in eastern U.S. cities.
October 1999 6-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 In Schwartz (1997), relative risk estimates of CO and PM10 for admissions were of similar
2 magnitude, independent, additive, and statistically significant at a = 0.05. In a model assessing
3 both pollutants simultaneously, the estimated percentage increases in admissions (PIAs) across
4 the interquartile ranges of CO and PM10 levels were 2.33 and 2.37%, respectively. Effects
5 estimates for both pollutants appeared to be quite stable across seasons, and did not appear to be
6 confounded with the meteorologic parameters assessed. There were no appreciable associations
7 of admissions with ambient levels of SO2, O3, or NO2.
8 Burnett et al. (1997a) examined temporal relationships between short-term ambient air
9 pollution levels and hospitalizations for congestive heart failure (CHF) in the elderly (individuals
10 >64 years of age) in 10 Canadian cities, during the 11-year period 1981 through 1991. The
11 average daily number of CHF admissions was 39 in the 134 catchment hospitals. A time series
12 analysis adjusted for long-term time trends, seasonal and subseasonal temporal variation, and
13 day-of-week effects was used to explore the relationship between cardiopulmonary illness and
14 the ambient air pollutants CO, NO2, SO2, O3, and coefficient of haze (COH, an optical index of
15 ambient PM concentration). After stratifying by month of the year and adjusting for temperature,
16 dew point, and other pollutants, the log of the daily 1-h maximum CO concentration on the day
17 of admission (0-day lag) had the strongest statistical association with hospitalization for CHF.
18 The relative risk was 1.065 (95% CI = 1.028, 1.104) for an increase from 1 to 3 ppm CO
19 (the interquartile range of the ambient CO concentration distribution). Associations of other
20 pollutants with admissions also were observed in single-pollutant models. However, risk
21 estimates for these other pollutants were more sensitive to simultaneous adjustment for multiple
22 pollutants and weather variables than were the estimates for CO. Burnett et al. (1997a) present a
23 nonparametrically smoothed exposure-response function for daily maximum hourly CO level and
24 adjusted daily CHF admissions (Figure 6-2). The authors noted that CO could be acting as a
25 marker for pollution from transportation sources in general, and that independent effects of
26 non-CO pollutants could not be ruled out.
27 Burnett et al. (1997b) investigated summertime ambient air pollution in relation to
28 unscheduled hospital admissions for cardiac and respiratory diseases in Toronto, Ontario,
29 Canada, in the summers of 1992, 1993, and 1994 (total 388 days). Hourly measurements of CO,
30 O3, SO2, NO2, and COH were available from multiple monitoring stations. Daily measurements
31 of fine and coarse PM (8:00 a.m.-8:00 a.m.) were available from a dichotomous sampler at a
October 1999 6-17 DRAFT-DO NOT QUOTE OR CITE
-------
1.10-
o
'
E
1.05-
Q.
O
_C
_>,
'ro
1.00-
0.95-
246
Daily 1 hour maximum CO (ppm)
Figure 6-2. Nonparametric smoothed curve of adjusted hospitalization rates for
congestive heart failure in the elderly plotted against daily 1-h maximum CO
levels in 10 Canadian cities, 1981 to 1991. Hospitalization rates are adjusted
for long-term time trends, seasonal and subseasonal variations, and day of
week effects and are normalized to unity for each city separately.
Source: Burnett et al. (1997a).
1 downtown site chosen to be representative of the Toronto area. Measurements of COH were also
2 available. The PM10 levels were calculated as the sum of daily fine and coarse PM mass.
3 Ambient CO levels were low; the mean and 95th percentile of the daily 1-h maximum CO
4 concentration were 1.8 and 3.2 ppm, respectively.
5 In data analysis, pollutant concentrations were lagged 0 to 4 days before admission.
6 Additional pollutant metrics were computed as multiday average ambient concentrations, with
October 1999
6-18
DRAFT-DO NOT QUOTE OR CITE
-------
1 the last day of the averaging period lagged 0 to 2 days before admission. The number of days in
2 the averaging period, and its last day, varied from pollutant to pollutant. In single-pollutant
3 models, there were positive, usually significant, effects of all pollutants on both cardiac and
4 respiratory admissions. Many two-pollutant models, each with one particulate metric and one
5 gaseous pollutant metric, were constructed. In these, there was little evidence of a CO effect on
6 cardiac or respiratory admissions. Effects of CO remained slightly positive, but were not
7 statistically significant. Effects of PM metrics adjusted for CO were similar to PM effects in
8 single-pollutant models. The gaseous pollutant least sensitive to adjustment for PM metrics
9 was O3. The gaseous pollutant that most attenuated PM effects was NO2. The authors
10 hypothesized that the absence of a CO may result from the fact that summertime ambient CO
11 levels are low and emphasized the potential importance of the overall ambient air pollution mix,
12 recommending that "all available air pollution measures be considered in assessing the effects of
13 any single pollutant on health."
14 Poloniecki et al. (1997) investigated cardiovascular hospital admissions in relation to
15 ambient air pollution concentrations one day before admission in London, UK, from April 1,
16 1987, to March 31, 1994. Pollutant metrics were 24-h mean CO, BS, SO2, and NO2, and hourly
17 mean O3 between 9:00 a.m. and 5:00 p.m. Measurements of all gaseous pollutants were taken
18 from a single site in central London. Measurements of BS were taken from one central site and
19 four suburban sites. Median and 90th-percentile CO concentrations were 0.9 and 1.8 ppm,
20 respectively. Corresponding BS concentrations were 12 and 22//g/m3. Health outcomes
21 considered were admissions for all cardiovascular diseases and for the following seven
22 diagnostic subgroups: myocardial infarction (MI), other ischemic heart disease, heart failure,
23 angina, cardiac dysrhythmia, cerebrovascular disease, and other circulatory diseases.
24 Analytical models were adjusted for day of week, holidays, an influenza epidemic in 1989,
25 and several temporal cycles ranging from 20 days to the whole study period. Pollutant
26 concentration, temperature, and humidity one day before admission were also entered into the
27 models (1-day lag). Single-pollutant Poisson models were constructed for each health outcome.
28 In these models, CO was positively and statistically significantly associated with admissions for
29 all cardiovascular diseases, MI, and other circulatory diseases. Black smoke was significantly
30 associated with admissions for all cardiovascular diseases, MI, and angina. Admissions for all
31 cardiovascular diseases and MI were significantly associated with all pollutants except O3, which
October 1999 6-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 was generally negatively, but not significantly, associated with these admissions. The only
2 diagnostic subgroup signficantly associated with more than one pollutant was MI. Admissions
3 for heart failure, other ischemic heart disease, and cerebrovascular disease were not significantly
4 associated with any pollutants.
5 Additional single-pollutant and two-pollutant Poisson models were constructed for MI
6 during the warm season (April through September) and the cool season (October through March).
7 In the warm season, there were no significant associations of air pollution with MI admissions
8 (p > 0.25). In the cool season, CO was positively and significantly associated with MI in the
9 single-pollutant model and the two-pollutant model with O3. The same was true for all other
10 pollutants except O3, and SO2 was significant in all two-pollutant models. P-values and
11 regression coefficients from Table 4 of Poloniecki et al. (1997), which show pollutant effects in
12 single-pollutant and two-pollutant models, and which show the marked differences in modeled
13 air pollution effects between the cool and warm seasons, are presented in Table 6-3.
14 This study suggests that, in London, short-term elevations in ambient air pollution levels
15 are related to cardiac hospital admissions, especially MI admissions, 1 day later. The authors
16 systematically evaluated pollution effects on different diagnostic subgroups of CV disease, and,
17 unlike North American investigators, reported no association of ambient CO with congestive
18 heart failure. Even so, there is some consistency between the European and North American
19 studies. Specifically, these studies show positive, statistically significant relationships of
20 ambient CO concentration, and other pollutant concentrations, with hospital admissions for heart
21 disease, and these relationships tend to be sronger for heart disease than other types of
22 cardiovascular disease such as stroke. Also, effects of CO on heart disease exacerbation have
23 consistently been strongest with a 0-day lag for the ambient CO metric. Poloniecki et al. (1997)
24 evaluated effects of air pollution only with a 1-day lag. Thus, though it is not certain, modeled
25 CO effects may have been biased downward in this study.
26 Poloniecki et al. (1997) present a balanced discussion of consistencies and inconsistencies
27 in the available epidemiologic database regarding ambient air pollution. They emphasize the
28 potential for confounding of air pollution and meteorologic effects, despite conscientious efforts
29 to model these effects appropriately. They state, "Significant associations [of air pollution with
October 1999 6-20 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6-3. ACUTE MYOCARDIAL INFARCTION: ONE AND TWO POLLUTANT
MODELS WITH COOL AND WARM SEASON3 b, LONDON, UK,
APRIL 1,1987 TO MARCH 31,1994
Test Pollutant
03 (ppb)
Cool season:
P Value
Coefficient
Warm season:
P Value
Coefficient
N02 (ppb)
Cool season:
P Value
Coefficient
Warm season:
P Value
Coefficient
S02(ppb)
Cool Season:
P Value
Coefficient
Warm season:
P Value
Coefficient
CO (ppb)
Cool season:
P Value
Coefficient
Warm season:
P Value
Coefficient
Black smoke (//g/m3)
Cool Season:
P Value
Coefficient
Warm season:
P Value
Coefficient
03
0.22
-0.0013
0.48
-0.0005
0.0004C
0.0022
0.49
0.0004
0.0005°
0.0025
0.70
0.0004
o.oor
0.0324
0.63
0.0112
0.0006°
0.0033
0.84
-0.0003
NO2
0.72
0.0004
0.39
-0.0006
0.0020'
0.0013
0.53
0.0003
0.03C
0.0015
0.49
0.0009
0.15
0.0205
0.63
0.0154
0.23
0.0014
0.96
0.0001
Second pollutant
SO2
0.91
0.0001
0.26
-0.0009
0.15
0.0008
0.65
-0.0003
0.0004"
0.0021
0.53
0.0006
0.39
0.0083
0.55
-0.0160
0.10
0.0015
0.86
0.0003
CO
0.93
-0.0001
0.34
-0.0009
0.84
0.0002
0.90
0.0001
0.02C
0.0020
0.32
0.0013
0.02e
0.0227
0.40
0.0187
0.13
0.0019
0.79
-0.0006
BS
0.95
0.0001
0.34
-0.0007
0.23
0.0008
0.63
0.0003
0.03C
0.0015
0.58
0.0006
0.38
0.0100
0.39
0.0243
0.002'
0.0024
0.82
0.0003
"Diagonal elements (bold italics) are single-pollutant models; off-diagonal elements are test pollutant modeled with a second pollutant
bCool season is October to March, and warm season is April to September. Coefficient = Poisson regression coefficient—for example,
percent admissions per unit of pollutant = [exp (coeff) -I] x 100.
°P<0.05.
Source: Modified from Poloniecki et al. (1997).
October 1999
6-21
DRAFT-DO NOT QUOTE OR CITE
-------
1 health outcomes] might...arise solely because factors that affect the clearance and creation of air
2 pollution—such as the weather and driving motor vehicles—also have a direct effect on human
3 behavior and on acute health outcomes...Although we have modelled temperature and humidity,
4 there are other factors—for example, wind and rain—which affect air pollution and also affect
5 the amount of physical exercise taken and the duration of time spent outdoors, where levels of air
6 pollution are generally different from indoors. With so much potential for confounding, and with
7 a good many conflicting results to be found in the scientific literature, we chose to place the
8 statistical emphasis on the direction and consistency of the relations (P values) rather than on
9 estimating the size of effects from the models which produce the largest effects, as has usually
10 been done." Thus, these authors recognize the high degree of uncertainty inherent in ostensibly
11 quantitative estimates of effects of individual pollutants in time series studies.
12 Poloniecki et al. (1997) also present the following assessment of observed associations of
13 CO with heart disease exacerbation: "The present results...make a case for further study of CO.
14 Average concentrations over one hour at a kerbside monitor can reach 25 ppm [in the UK].
15 Breathing CO at 50 ppm for one hour increased carboxyhaemoglogin by a factor of two, and
16 reduced symptom limited exercise capacity in patients with stable angina...Thus the risk of
17 myocardial ischaemia when pushing a lawn mower or or shovelling snow is increased by the
18 proximity of an internal combustion engine. We have shown that fluctuations in...CO are an
19 additional risk factor for myocardial infarction in the population, and this effect could be due to
20 an increased risk of infarction during exercise when environmental concentrations of CO have
21 risen within the preceding 24 hours."
22 Morris and Naumova (1998) investigated joint effects of short-term ambient CO
23 concentration and ambient temperature on daily hospital admissions for CHF in people >64 years
24 of age, in Cook County (Chicago), IL, from 1986 through 1989. Data were analyzed with
25 general linear models (GLM) and general additive models (GAM). The pollutant metrics
26 assessed were daily maximum hourly levels of CO, NO2, SO2, and O3, as well as 24-h average
27 PM10. For each day of the study, gaseous pollutant measurements were averaged across Cook
28 County's eight monitoring stations, six of which were in downtown Chicago. The PM10 levels
29 were measured at only one station on 80% of study days. In addition to these pollutant variables,
30 models included variables for daily maximum hourly temperature, day of week, month of year,
31 and year of study.
October 1999 6-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 In single-pollutant GLMs, the level of each pollutant except O3 was associated positively
2 and statistically significantly with CHF admissions on the same day. In a GLM that included all
3 pollutants, only CO was associated significantly with admissions. In this model, the RR for CHF
4 admission at the 75th percentile of maximum hourly CO concentration was 1.08 (95% CI 1.03,
5 1.12), as compared to RR = 1 at CO concentration of zero. Associations of CO and other
6 pollutants with admissions were strongest on the day of admission (0-day lag), and weakened
7 rapidly with successively longer lag times.
8 The authors conducted a detailed and thoughtful analysis to assess effects of ambient CO at
9 different temperatures. These effects were analyzed in three ways: (1) inclusion of a
10 CO-temperature interaction term in GLM; (2) simultaneous inclusion of a CO term and a
11 temperature term in GAM, generating an additive CO-temperature effects surface; and
12 (3) analysis with stratification on daily maximum ambient temperature (<40°, 40 to 75°, and
13 >75° F). Effects of CO on CHF admissions were consistently associated inversely with
14 temperature (stronger effects at lower temperatures). For example, in a multipollutant GLM,
15 RRs for CHF admissions at the 75th percentile of maximum hourly CO concentration were
16 1.09(1.01, 1.18), 1.07(1.01, 1.13), and 1.01 (0.92, 1.11) when maximum temperature was <40,
17 40 to 75°, and >75° F, respectively. Figure 3 in Morris and Naumova (1998) shows
18 temperature-specific exposure-response curves for percentiles of ambient CO distributions and
19 relative risk of CHF admission in single-pollutant and multi-pollutant models. This figure is
20 reproduced in Figure 6-3. The authors hypothesized that CHF patients may be unusually
21 susceptible to CO effects, and that cardiovascular and other stresses imposed by cold weather
22 may heighten this susceptibility.
23 Morris and Naumova (1998) is one of the few time series studies that have jointly analyzed
24 effects of ambient air pollution and weather, and thereby not forced statistical models to generate
25 single effects estimates for air pollutant metrics. It is an unusually strong time series study, and
26 its findings regarding ambient CO effects on CHF admissions are therefore of unusual interest
27 and concern. However, even these findings should be interpreted cautiously. For example, as
28 the authors point out, fixed-site CO measurements do not give exact estimates of individual
29 subjects' total CO exposures. Discrepancies between estimated and actual total CO exposures
30 may be greatest in cold weather, when indoor CO levels may be higher than in warm weather,
31 and when CHF patients may spend more time indoors. Also, a given day-to-day difference in
October 1999 6-23 DRAFT-DO NOT QUOTE OR CITE
-------
1.20-
1.15-
to
*k_
CD
£ 1.10-1
m
CD
1.05-
1.00-
A
20 40 60 80
CO exposure percentile
100
1.20-
1.15-
CD
£ 1.10
CO
1.05-
1.00-
B
20 40 60 80
CO exposure percentile
100
Figure 6-3. The relative risk associated with the exposure percentiles of CO at specific
temperature strata, based on results of the (A) multi-pollutant and
(B) single-pollutant generalized linear models of hospital admissions for heart
failure among the elderly in Chicago, IL, 1986 to 1989.
Source: Morris and Naumova (1998).
1 ambient CO level could reflect a greater day-to-day difference in total CO exposure in cold
2 weather than in warm weather. Thus, the observed stronger associations of ambient CO with
3 CHF in colder weather might reflect an underlying effect of increased total CO dose in colder
4 weather, and not an effect of ambient CO exposure per se. Finally, the observed associations
5 could indeed reflect a true effect of ambient CO fluctuations, superimposed on higher baseline
6 CO doses from nonambient sources during colder weather.
October 1999
6-24
DRAFT-DO NOT QUOTE OR CITE
-------
1 As mentioned above, Burnett et al. (1997b) did not observe an association of ambient CO
2 with cardiac admissions. However, their results do not contradict those of Morris and Naumova
3 (1998), who observed a strong association of CO with congestive heart failure admissions at cold
4 temperatures but little or no association at warm ones. Pantazopoulou et al. (1995) and
5 Poloniecki et al. (1997) also observed associations of CO and other air pollutants with cardiac
6 admissions during cooler weather but not during warmer weather. Thus, these four studies
7 generate the hypothesis that ambient CO effects on heart disease exacerbation are stronger during
8 cooler weather than during warmer weather. This hypothesis should be tested further in future
9 studies.
10 Yang et al. (1998) studied cardiovascular hospital admissions in relation to short-term CO
11 levels in Reno-Sparks, NV, over 6 years (1989 to 1994). The study area is about 4,400 ft above
12 sea level. At this elevation, the effect of COHb on tissue O2 delivery might differ from that at sea
13 level. The studied range of ICD-9 codes, 390 to 459.99, included heart disease and other
14 cardiovascular disorders such as stroke and hypertension. The mean measured ambient CO level
15 during the study period was 3.09 ppm. Data were analyzed with weighted least-squares
16 regression and autoregressive integrated moving average (ARTMA) regression for time series
17 data. The CO metric was the average of the highest value of the maximum hourly CO levels at
18 all five local monitoring stations from midnight to noon. Models were adjusted with dummy
19 variables for day of week and month of year. Wind speed and the previous day's minimum
20 temperature were also considered. CO effects estimates were reported for the ambient CO level
21 on the day of admission (0-day lag). Pollutants other than CO were not considered in the
22 analysis.
23 The authors stated, "All hospital admissions for CV [cardiovascular disease] and IHD
24 [ischemic heart disease] were significantly associated with CO concentrations..." Tabular
25 displays of results were limited to all CV admissions and IHD admissions. No results were
26 reported for other diagnostic subcategories. Modeled CO effects on admissions were larger in
27 males than females, and did not differ significantly by age. The significance levels of CO effects
28 are somewhat difficult for the reader to discern. The tabular summary of regression results
29 (Table 5) suggests very high significance levels (p < 0.006); units of regression coefficients were
30 not given. However, the text states, "According to the ARUVIA models, CV and IHD hospital
31 admissions increased 1.19% (95% CI: 0.99, 1.39%) and 2.83% (95% CI: 2.07, 3.60%),
October 1999 6-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 respectively, for each 1 ppm increase in the 1-h maximum CO level...." This suggests only
2 marginal significance of the effect of CO on total cardiovascular admissions, and clear
3 significance of the CO effect on IHD admissions. On balance, this report suggests an association
4 of short-term ambient CO levels with cardiovascular admissions, especially IHD admissions, at
5 an elevation of 4,400 ft. The report does not enable inference as to effects of pollutants other
6 than CO.
7 Schwartz (1999) evaluated effects of short-term ambient CO and PM10 exposure on daily
8 hospital admissions for cardiovascular disease (ICD-9 codes 390 to 429) in persons aged at least
9 65 years, in eight U.S. counties over 3 years, 1988 to 1990. The analyzed health outcomes were
10 the same as in Schwartz (1997) (i.e., ICD-9 codes 390 to 429). The eight study locations were
11 Chicago, IL; Colorado Springs, CO; Minneapolis and St. Paul MN; New Haven, CT; and Seattle,
12 Spokane, and Tacoma, WA. Previous findings from Tucson, AZ, also were summarized
13 (Schwartz, 1997). During the study period, 50th-percentile CO and PM10 levels ranged from
14 2.0 to 4.7 ppm and 23-37 //g/m3, respectively, across the study locations. Pollutants other than
15 CO and PM10 were not analyzed, though the author stated: "The emphasis is on locations in the
16 western United States, because western locations have been less studied, because the correlation
17 between PM10 and O3 usually is negative versus positive in the east, and because SO2
18 concentrations are much lower in the west." (Even so, the dataset included the four nonwestern
19 locations of New Haven, Chicago, Minneapolis, and St. Paul.) Daily maximum 1-h CO levels
20 and 24-h average PM10 levels were considered, apparently only on the same day as admissions
21 (0-day lag). Aerometric data were taken from population-oriented CO monitors, as presumably
22 opposed to compliance monitors, which may be located away from population centers. Poisson
23 regression was employed, with LOESS smoothing to adjust for effects of temperature, dew point,
24 and seasonal patterns. Effects estimates were reported as percentage increases in admissions
25 (PIAs) over approximate interquartile ranges of ambient pollutant concentrations.
26 In all study locations, both CO and PM10 levels were positively associated with daily
27 cardiac admissions. CO effects were statistically significant in seven of nine locations, and PM10
28 effects were significant in six of nine locations. Relative effects of CO and PM10 differed widely
29 among locations. Overall, the interquartile PIA for CO was 2.79% (95% C.I. 1.89 to 3.68), and
30 that for PM10 was 2.48% (1.81 to 3.14). Effects estimates for CO and PM10 were not related to
31 the location-specific correlation between these two pollutants' concentrations, suggesting that
October 1999 6-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 their effects on cardiac admissions were at least partially independent. Table 3 of Schwartz
2 (1999), which shows modeled ambient CO and PM10 effects in all eight study locations and in
3 Tucson, is reproduced with slight modification in Table 6-4. In two-pollutant models, effects
4 estimates for CO and PM10 were somewhat smaller than in single-pollutant models. Pollution
5 effects estimates were not reported for diagnostic subcategories (e.g., ischemic heart disease and
6 heart failure). This study's findings leave a general impression of similar, partially independent
7 statistical effects of short-term ambient CO and PM10 on total cardiac admissions in the elderly.
TABLE 6-4. MODELED PERCENTAGE INCREASES IN HOSPITAL ADMISSIONS
FOR HEART DISEASE, ASSOCIATED WITH INTERQUARTILE RANGE
INCREASES IN AMBIENT PM10 AND CO LEVELS, IN EIGHT LOCATIONS,
ACROSS THESE LOCATIONS, AND IN TUCSON, AZ, 1988 TO 1990
City
Chicago, IL
Colorado Springs, CO
Minneapolis, MN
New Heaven, CT
St. Paul, MN
Seattle, WA
Spokane, WA
Tacoma, WA
Across Locations
Tucson, AZ
PM10
2.31
2.76
2.03
2.87
4.19
1.77
3.28
2.63
2.48
2.99
95% CL
1.31,3.33
-3.2,9.09
-1.87,6.09
1.04,4.73
1.44,7.00
-0.07,3.64
0.43,6.21
0.47,2.63
1.81,3.14
0.55,5.50
CO
2.84
0.51
4.09
3.04
0.74
4.22
2.71
1.84
2.79
2.94
95% CL
1.59,4.10
-2.41,3.51
1.59,6.65
1.18,4.93
-1.84,3.39
2.44,6.02
0.69,4.78
0.24,3.46
1.89,3.68
0.54,5.71
Window3
76
180
143
172
191
161
145
120
NA
83
"Size of window (in days) for LOESS smoothing of time to control season.
Source: Modified from Schwartz (1999).
1 At the same time, it is not clear why CO exerted a substantially larger statistical effect than
2 PM10 on admissions in Minneapolis, while the reverse was true in the adjacent location of
3 St. Paul. It is also not clear why the LOESS seasonal smoothing window was 34% longer in
October 1999 6-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 St. Paul than in Minneapolis. Similarly, CO had a larger effect than PM10 on admissions in
2 Seattle, while the reverse was true in nearby Tacoma. Furthermore, the seasonal smoothing
3 window in Seattle was 34% longer than in Tacoma.
4 Schwartz (1999) points out that while CO is a known "cardiovascular toxin," cardiac
5 effects in experimental studies of humans have been observed only at higher than ambient CO
6 levels. He states that one would expect to observe CO effects at lower levels in epidemiologic
7 studies, because epidemiology examines the entire population, which comprises a broader range
8 of disease states and exposure to potential effects modifiers than do experimental subject groups.
9 He also raises the possibility that ambient CO may be a marker for automotive pollution, and
10 states, "One potential constituent [of automotive pollution]...that may have acute cardiovascular
11 toxicity is the volatile and semivolatile organic aerosols...CO may be serving as a proxy for these
12 compounds. Nevertheless, associations between cardiovascular hospital admissions and CO
13 have been seen in more than 20 North American cities,...suggesting that a better understanding of
14 this pollutant should have a high priority."
15 Burnett et al. (1999) assessed daily unscheduled hospital admissions for eight types of
16 cardiovascular and respiratory disease exacerbation in Toronto, Canada, in relation to short-term
17 ambient air pollution exposure over 15 years, 1980 to 1994. Daily average concentrations of CO,
18 NO2, SO2, and O3 were obtained from four monitoring stations. Daily average concentrations of
19 fine PM (PM2 5), coarse PM (PM10_2 5), and PM10 were estimated from TSP and sulfate levels
20 measured at a single central station. Daily average COH levels were measured at the same
21 station. The cardiovascular disease types were ischemic heart disease, heart failure, cardiac
22 dysrhythmias, cerebrovascular disease (including stroke), and peripheral circulatory disorders.
23 The respiratory disease types were asthma, obstructive lung disease, and respiratory infection.
24 Data were analyzed with general additive models, with accommodation for over- or under-
25 dispersion from Poisson variation. The data were prefiltered with a 31-day LOESS smoother.
26 The LOESS smoothers also were employed to adjust for effects of temperature and relative
27 humidity. Models were adjusted for day-of-week effects. Exploratory models were run to
28 provide parsimonious choices of climate variables in final models. The report did not present
29 seasonal air pollution effects estimates. Pollution metrics were 1-, 2-, or 3-day averages, with the
30 final day of the averaging period lagged 0 to 2 days before admission. All possible single-
31 pollutant models were run for each disease type. Pollutant metrics were constructed to give the
October 1999 6-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 log of relative risk per unit change in pollutant concentration. Pollutant effects were reported as
2 the percentage increase in total daily hospital admissions, from zero concentration to the mean
3 pollutant concentration during the study period.
4 In the single-pollutant models (Table 3 in Burnett et al., 1999), pollutant effects estimates
5 were weak or negative for cerebrovascular disease and peripheral circulatory disease, so these
6 two disease types were not considered in subsequent multi-pollutant analyses. This left only
7 three heart disease categories and three respiratory disease categories for further consideration.
8 In single-pollutant models, the CO effect in each of these six categories was strongest for a
9 1- or 2-day averaging time, ending on the same day as the admission (0-day lag).
10 The authors constructed multi-pollutant models in an effort to ascertain effects of the
11 overall ambient air pollution mix, and to estimate health benefits that would follow air pollution
12 reduction. They selected pollutant metrics from the single-pollutant models for inclusion in the
13 multi-pollutant models. For each pollutant, they selected the averaging period and lag that
14 yielded the largest ratio of log-relative risk to standard error (largest T-ratio or t-statistic). They
15 constructed two sets of multi-pollutant models. In the first set (Table 4 in Burnett et al., 1999),
16 four models were run for each of the six cardiac and respiratory disease categories. In one of
17 these four models, all PM metrics were excluded. In each of the other three models, the
18 investigators forced inclusion of one and only one PM metric (PM25, PM10_25, or PM10). After
19 these constraints were imposed, gaseous pollutants were entered into each of the 24 models using
20 stepwise regression. Specific gaseous pollution metrics in final models were selected using the
21 Akaike Information Criterion (AIC).
22 In this first set of multi-pollutant models, inclusion of gaseous pollutants led to
23 substantially smaller PM effects estimates than observed in single-pollutant models. Also,
24 modeled percentage increases in admissions (PIAs) with a PM metric in the model were only
25 moderately larger than corresponding increases with no PM metric in the model. The maximum
26 diagnosis-specific percentage increase in the PIA with a PM metric in the model, over the PIA
27 with no PM metric, ranged from 7.4% for ischemic heart disease to 28.1% for respiratory
28 infection. These increases were larger for respiratory diseases (average = 21.0%) than for heart
29 diseases (average = 8.5%).
30 In the second set of multi-pollutant models (Table 5 in Burnett et al., 1999), all gaseous and
31 PM metrics competed equally for inclusion, and no metric was constrained to be included or
October 1999 6-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 excluded. In these models, percentage increases in diagnosis-specific hospital admissions, at the
2 mean concentrations of included pollutants, ranged from 9.29% (IHD) to 14.45% (respiratory
3 infection). Effects of CO and NO2, which are both generated by mobile sources, were largely
4 confounded. CO but not NO2 was included in models for asthma, obstructive lung disease, and
5 dysrhythmias, and NO2 but not CO was included for IHD. Both CO and NO2 were included as
6 independent predictors for heart failure. Taken together, effects estimates for CO and NO2 were
7 larger, and more frequently statistically significant, for heart diseases than for respiratory
8 diseases.
9 Table 6-5 summarizes modeled effects of CO in single-pollutant models, and modeled
10 effects of CO and other pollutants in multi-pollutant models, as taken from Tables 3 and 5 of
11 Burnett et al. (1999). No PM metric was included in multi-pollutant models for the heart
12 diseases IHD and heart failure (which accounted for 73.9% of all U.S. hospital admissions for
13 heart disease in 1996). When a PM metric was included, the contribution of the modeled PM
14 effect to the overall pollution-associated increase in admissions ranged from 19.3%
15 (dysthythmias) to 42.1% (respiratory infection). The average contribution of PM metrics to the
16 air pollution-associated percentage increase in heart disease admissions was thus only 6.4%
17 ([0% + 0% + 19.3%] / 3). If weighted by the number of admissions for each of the three
18 analyzed heart disease categories in the United States in 1996 (see Section 6.2.1.1), this
19 contribution would calculate to only 3.2%. The average contribution of PM metrics to
20 percentage increases in respiratory disease admissions was 34.2%. Fine PM was included in
21 models for dysrhythmias and respiratory infection, and coarse PM was included for asthma and
22 obstructive lung disease. PM10 was not included for any disease, and no more than one PM
23 metric was included for any disease. Ozone was associated much more strongly with respiratory
24 disease admissions than with heart disease admissions. Single-pollutant model effects of SO2
25 could be largely explained by inclusion of other pollutants.
26 In Burnett et al. (1999), modeled PM metrics were not measured directly. Thus, the
27 findings are subject to quantitative uncertainty regarding specific pollutant effects, and regarding
28 gaseous pollution effects relative to PM effects. Modeled effects of PM may have been biased
29 downward, but this is not certain. Also, the extent to which findings in Toronto can be
30 generalized to other locations is not clear. Nevertheless, this report presents the most
31 comprehensive effort to date to examine multi-pollutant effects across a variety of health
October 1999 6-30 DRAFT-DO NOT QUOTE OR CITE
-------
October 1999
Oi
OJ
O
H
6
o
0
H
O
O
0
o
TABLE 6-5. MODELED PERCENTAGE INCREASES IN HOSPITALIZATIONS AT MEAN POLLUTANT
CONCENTRATIONS3 IN SINGLE-POLLUTANT AND MULTI-POLLUTANT MODELS,
TORONTO, CANADA, 1980 TO 1994
Single-Pollutant Models (Table 3) Multi-Pollutant Models (Table 5)
Disease
Category CO NO2 SO2 O3 PM10 PM25 PM10.25 CO NO2 SO2 O3 PM
Heart Disease
Dysrhythmias 8.99(3.60) 5.33(1.73) 0.80(1.43) 3.51(1.71) 5.00(3.03) 4.33(2.91) 2.47(1.88) 7.00(2.50) — b — 3.34(1.63) 2.47C(1.49)
Heart Failure 8.33(5.71) 9.48(6.33) 1.93(3.85) 1.42(1.42) 5.75(3.51) 4.70(3.20) 3.77(2.79) 4.09(2.08) 6.89(3.44) — — —
Ischemic 7.31(6.46) 9.73(8.40) 2.32(6.13) 0.61(0.99) 4.97(5.55) 5.73(6.08) 1.81(3.02) — 8.34(6.10) 0.95(2.07) — —
Heart Disease
Respiratory Disease
Asthma 5.35(3.92) 3.33(2.37) 1.01(1.76) 6.32(4.63) 5.27(3.39) 4.60(3.22) 5.25(4.20) 4.00(3.86) — — 4.99(3.48) 4.00d(3.04)
Obstructive 2.93(1.48) 2.21(1.07) 0.03(0.05) 7.29(4.23) 4.11(2.44) 3.42(1.89) 6.07(3.26) 3.00(1.52) — — 6.08(2.74) 3.86d(1.90)
Lung Disease
Respiratory 5.00(4.25) 6.89(5.53) 2.40(5.04) 4.42(4.29) 8.35(5.96) 7.64(6.09) 4.44(4.00) — 4.44(3.31) — 3.93(3.80) 6.08C(4.46)
Infection
T-statistics for pollutant effects estimates are in parentheses, and are approximately equal to Z-statistics. Two-tailed Z-statistics of 1.96 or greater, and 2.58
denote statistical significance at a = 0.05 and a = 0.01, respectively.
bA dash denotes that the indicated pollutant was not included in the multi-pollutant model.
The PM metric included was PM2 s (fine PM).
The PM metric included was PM2 5_10 (coarse PM).
Source: Modified from Burnett et al. (1999).
All
Pollutants
12.81
10.98
9.29
12.99
12.94
14.45
or greater,
-------
1 outcomes. This study, like many other recent studies, emphasizes the importance of considering
2 both gaseous and particulate pollutants in data analyses and underscores the importance of
3 gaseous pollutants as contributors to health effects attributable to the overall ambient air
4 pollution mix, especially in relation to heart disease exacerbation. Also, unless both gaseous
5 pollutants and PM are taken into account, modeled effects estimates for single pollutants are
6 likely to be inaccurate. Also, using effects estimates derived from multipollutant models may
7 accommodate, at least partly, for confounding of effects among individual pollutants in
8 estimating potential public health risks.
9
10 6.2.1.3 Ambient Carbon Monoxide and Daily Mortality
11 Epidemiologic studies of the relationship between CO exposure and daily mortality are not
12 conclusive. Early studies in Southern California (Goldsmith and Landaw, 1968; Cohen et al.,
13 1969; Hexter and Goldsmith, 1971) suggested an association between atmospheric levels of CO
14 and increased mortality from cardiovascular disease, but potential confounders were not
15 controlled thoroughly. In contrast, Kuller et al. (1975) observed no association between ambient
16 CO levels and cardiovascular disease or sudden death in Baltimore, MD.
17 Kinney and Ozkaynak (1991) investigated effects of short-term variation in ambient air
18 pollution levels on daily nonaccidental, nonviolent deaths in Los Angeles County over 10 years,
19 1970 through 1979. The mean daily death count was 152; of these, 87 (57%) and 8 (5%) resulted
20 from cardiovascular and respiratory causes, respectively. Pollution metrics were daily maximum
21 CO, daily maximum hourly total oxidants (O3 in 1979), 24-h average SO2, NO2, and KM
22 (a particulate metric similar to British Smoke, related to elemental carbon), and visual extinction
23 coefficient (Bext, related to fine particulate). In multiple regression models, adjusted for
24 meteorology and temporal patterns, there were statistically significant associations of total and
25 cardiovascular mortality with temperature, oxidants (lagged 1 day), and the automotive pollutants
26 CO, NO2, and KM (each lagged 0 days). Levels of the latter three pollutants were too highly
27 intercorrelated to enable confident assessment of their separate effects. Respiratory mortality
28 was associated with temperature but not with any pollution metric, though the power to test for
29 pollutant effects was limited by small numbers of daily respiratory deaths.
30 More recent time series studies in North and South America and in Europe have also
31 yielded mixed results in relating day-to-day variations in CO levels with daily mortality.
October 1999 6-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 For example, no relationship was found between CO and daily mortality in Los Angeles,
2 Chicago, or Philadelphia (Ito et al., 1995; Kinney et al., 1995; Ito and Thurston, 1996; Kelsall
3 et al., 1997) after adjusting for particles (i.e., PM10), time trends, and weather. Verhoeff et al.
4 (1996) found no relationship between 24-h average CO concentrations and daily mortality in
5 Amsterdam, with or without adjustment for PM10 and other pollutants.
6 Three other studies (Touloumi et al., 1994; Salinas and Vega, 1995; Wietlisbach et al.,
7 1996) showed small, statistically significant relationships between CO and daily mortality.
8 However, effects of other pollutants (e.g., total suspended particles [TSP], SO2, NO2, black
9 smoke) and of meteorologic variables (e.g., temperature, relative humidity) were also significant.
10 Further research will be needed to determine whether low-level CO exposure actually is
11 increasing mortality (particularly in the elderly population), whether CO is a surrogate marker for
12 some other mobile-source or combustion-related pollutant, or whether CO is a surrogate for the
13 overall combustion-related or automotive pollution mix.
14 Touloumi et al. (1994) investigated air pollution and daily all-cause mortality in Athens
15 from 1984 through 1988. Daily mean pollution indicators for SO2, black smoke, and CO were
16 averaged over all available monitoring stations. Autoregressive models were used, with log-
17 transformed daily mortality as the dependent variable, and with adjustment for temperature,
18 relative humidity, year, season, day of week, and serial correlations in mortality. Separate
19 models for log(SO2), log(smoke), and log(CO) yielded statistically significant effects estimates
20 (p < 0.001). Air pollution measurements lagged by 1 day were most strongly associated with
21 daily mortality. In a multipollutant model, SO2 and smoke were independent predictors of
22 mortality, though to a lesser extent than temperature and relative humidity. Addition of an
23 independent variable for CO concentration did not improve this model's ability to predict daily
24 mortality, suggesting that CO may be a surrogate marker for other mobile-source or combustion-
25 related pollutants.
26 In one of the Air Pollution and Health—A European Approach (APHEA) studies in
27 Athens, Touloumi et al. (1996) observed a distinct positive association of ambient CO levels
28 with daily mortality. Ambient CO concentrations were compiled from three fixed outdoor
29 monitoring stations. Median, mean, and maximum 8-h CO levels were 6.1 mg/m3 (5.3 ppm),
30 6.6 mg/m3 (5.8 ppm), and 24.9 mg/m3 (21.7 ppm), respectively. The relative risk for daily
31 mortality of a 10 mg/m3 (9 ppm) increase in the daily ambient air CO concentration was
October 1999 6-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.10 (95% CI = 1.05, 1.15). This finding may be attributable to yet-unknown health effects of
2 low levels of CO, to the presence of highly compromised susceptible groups in the population,
3 or, again, to CO acting as a surrogate for other combustion-generated air pollutants.
4 Salinas and Vega (1995) examined the effect of urban air pollution on daily mortality in
5 Metropolitan Santiago, Chile, from 1988 through 1991. Measurements of maximum 8-h average
6 CO; maximum hourly O3; daily mean SO2, PM10, and PM2 5; and meteorologic variables were
7 obtained from five monitoring stations. Total and respiratory disease-specific deaths were
8 assessed, calculating the risk of death by municipality and month of the year using age-adjusted
9 standardized mortality ratios, and controlling for socioeconomic status. Daily death counts were
10 regressed on pollutant levels, using Poisson regression and controlling for temperature and
11 relative humidity. There were geographic differences, independent of socioeconomic and living
12 conditions, in both total mortality and disease-specific mortality (e.g., pneumonia, chronic
13 obstructive pulmonary disease, asthma). The daily number of deaths was significantly associated
14 with ambient CO, ambient suspended particles, and humidity, and also was associated with
15 temperature when the model included all days with complete data during the 4-year period. The
16 associations remained significant for those days with PM2 5 levels below 150 //g/m3.
17 Wietlisbach et al. (1996) assessed the association between daily mortality and air pollution
18 in metropolitan Zurich, Basel, and Geneva, Switzerland, from 1984 through 1989. Daily counts
19 were obtained for total mortality, mortality in people 65 years of age or older, and respiratory and
20 cardiovascular disease mortality. Daily measurements of weather variables and TSP, SO2, NO2,
21 CO, and O3 were obtained in each city. Poisson models were used to regress daily death counts
22 on pollutant levels, controlling for time trends, seasonal factors, and weather variables. Positive,
23 statistically significants associations were found between daily mortality and TSP, SO2, and NO2.
24 The strongest associations were observed with 3-day moving averages. Somewhat weaker
25 associations were observed in each city between mortality in people 65 years of age or older and
26 measured average daily CO concentrations (mean = 1 to 2 mg/m3 [1 to 2 ppm], max = 5 to
27 8 mg/m3 [4 to 7 ppm]). Associations with O3 were weak and inconsistent. When all pollutants
28 were modeled simultaneously, the regression coefficients were unstable and not statistically
29 significant.
30 In two recent studies, Burnett and colleagues investigated associations of CO and other
31 pollutants with daily nonaccidental mortality in Canada. In one study (Burnett et al., 1998a), the
October 1999 6-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 investigators assessed the roles of average daily concentrations of ambient CO, other gaseous
2 pollutants, sulfates, TSP, COH, estimated PM2 5 and PM10, and meteorology in Toronto from
3 1980 through 1994. The time series was adjusted for long-term trends and temporal cycles.
4 Effects of several different exposure-to-mortality lags were explored, and the final choice of lags
5 was based on the Akaike Information Criterion. A 2-day moving average was selected as the
6 optimum metric for CO, but not for all pollutants. Final models included same-day dew point
7 temperature. In single-pollutant models, ambient levels of all pollutants except O3 were
8 associated positively and statistically significantly with daily mortality, and this association was
9 strongest for CO. Two-pollutant models also were constructed, each including CO and one of
10 the other pollutants. In these models, the magnitudes of relative risks for CO differed little from
11 that in the single-pollutant model for CO. In contrast, the relative risks for other pollutants
12 generally decreased appreciably. Also, the relative risks for CO remained statistically significant
13 in all two-pollutant models. Although the relative risk of CO was highest for deaths from cardiac
14 causes, there was also a clear positive association of CO with deaths from other causes.
15 Burnett et al. (1998b) also examined associations of ambient levels of gaseous pollutants
16 (CO, NO2, O3, and SO2) with daily nonaccidental mortality in 11 Canadian cities from
17 1980 through 1991. In single-pollutant models, relative risks of CO for mortality were more
18 consistent across cities than were relative risks of the other pollutants. However, in
19 multipollutant models, CO-associated relative risks decreased substantially, and NO2 and SO2
20 appeared to explain much of the CO effect on mortality. The estimated percentage increase in
21 mortality risk attributable to combined exposure to all four pollutants differed widely among
22 cities, ranging from 3.6% in Edmonton and Windsor to 11.0% in Quebec. The authors reasoned
23 that reductions in gaseous pollutant levels might be more effective than reductions in PM levels
24 in reducing mortality. It is not possible to interpret this study quantitatively, because direct
25 measurements of PM and PM constituents were not included in the analyses. At the same time,
26 these results underscore the need for measurement and statistical treatment of a broad range of
27 pollutants, and for further systematic assessment and comparison of the public health importance
28 of exposure to ambient CO, other ambient gaseous pollutants, and PM.
29 There have been few studies of ambient CO and mortality in children. Saldiva et al. (1994,
30 1995) observed no association between CO and daily mortality among children or the elderly in
31 Sao Paulo, Brazil, after adjusting for nitrogen oxides and PM10, though Pereira et al. (1998) did
October 1999 6-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 observe a limited relationship of ambient CO concentration with intrauterine mortality.
2 Interestingly, in the latter study, COHb levels in cord blood were correlated with short-term
3 ambient CO levels, even though intrauterine mortality was associated somewhat less strongly
4 with CO than with other pollutants. At the same time, Pereira et al. (1998) is difficult to interpret
5 because the investigators assessed fetal loss occurring only after 28 weeks of gestation, whereas
6 the large majority of spontaneous abortions occur before that time.
7
8 6.2.1.4 Ambient Carbon Monoxide and Low Birth Weight
9 Low birth weight (typically defined as birth weight <2,500 g) is associated with infant
10 mortality and childhood morbidity, and may predict increased risk of morbidity into adulthood
11 (Joseph and Kramer, 1996; Institute of Medicine, 1985). Though low birth weight is probably
12 not a direct cause of these harmful outcomes, it is a useful marker for developmental
13 disturbances that are more directly responsible (Weinberg and Wilcox, 1998).
14 Alderman et al. (1987) conducted a case-control study of birth weight in relation to ambient
15 CO concentration in Denver, CO, from 1975 through 1983. The CO metric was the time-
16 weighted geometric mean ambient CO level as measured in the mother's neighborhood during
17 the last 3 mo of gestation. The large majority of mothers lived within 2 mi of their neighborhood
18 monitoring sites. Median and 95th percentile CO concentrations ranged among monitoring sites
19 from 0.5 to 3.6 ppm and 0.8 to 4.8 ppm, respectively. Air pollutants other than CO were not
20 considered in analysis. Individual-level data on maternal age, race, education, marital status,
21 parity, and prior pregnancy history were available from birth certificates, but data on mother's
22 personal CO exposure and smoking were not available. Ambient CO exposures were divided
23 into quintiles and analyzed with Mantel-Haenszel methods, adjusting for mother's race and
24 education. No association of ambient CO level with frequency of birth weight <2,500 g was
25 observed in these analyses.
26 After consultation with exposure assessment experts, the investigators divided subjects into
27 two groups for whom the monitoring data were considered to reflect true CO exposure more and
28 less accurately. Separate analyses of these two groups were conducted. Interestingly, there was
29 more suggestion of a positive, monotonic CO effect in the former group (p value for chi-square
30 test of trend equaled 0.07 [marginally significant]) than in the latter group (p = 0.56 [not
31 significant]).
October 1999 6-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ritz and Yu (1999) assessed low birth weight in southern California, in relation to ambient
2 CO levels, between 1989 and 1993. The health outcome analyzed was incidence of birth weight
3 <2,500 g in singleton babies born at term (37 to 44 weeks of gestation), treated as a dichotomous
4 variable. Birth weights <1,000 g and >5,500 g were excluded. The main exposure variable was
5 the average ambient CO level from 6:00 to 9:00 a.m. during the third trimester of pregnancy, as
6 measured at the South Coast Air Quality Management District (SCAQMD) monitoring station
7 nearest to the mother's residence. This metric was chosen after consultation with SCAQMD
8 aerometric experts. Births were included only if the mother's residential zip code was entirely or
9 largely within 2 mi of one of 18 SCAQMD stations. The large majority of study mothers lived
10 within 2 mi of the nearest station. Thus, the study design, like that of Alderman et al. (1987),
11 addressed the fact that in the United States at least, ambient CO levels are often heterogeneous
12 over large distances. The overall average third-trimester ambient CO level was 2.6 ppm; average
13 CO levels ranged from 1.44 to 3.72 ppm across the study stations. Data were analyzed with
14 logistic regression, adjusting for gestational age, parity, time since previous birth, infant's gender,
15 and mother's age, educational level, and ethnic group. Ecological variables for commuting
16 habits, constructed from census data for the respective zip codes, also were included.
17 A total of 125,573 term births (92.1% of 136,376 eligible births) was included in analysis.
18 Birth weight was <2,500 g in 2,813 (2.2%) of these. The analysis predicted a 22% increase in
19 incidence of low birth weight among babies born to mothers with average ambient CO exposure
20 above the 95th percentile of 5.5 ppm (odds ratio 1.22, 95% CI1.03, 1.44). This rose to 33%
21 when first births were excluded and to 54% for mothers <20 years old. Average ambient CO
22 levels during the first and second trimesters, and during all trimesters combined, were not
23 associated with low birth weight incidence.
24 Measurements of O3, NO2, and PM10 were made at 6 of the 18 monitoring stations.
25 Multi-pollutant models were constructed for subjects living near these 6 stations. In these
26 models, ambient third-trimester ambient CO levels were categorized into 0 to 50th percentile
27 (reference category), 50 to 95th percentile (2.2 to 5.5 ppm), and >95th percentile (>5.5 ppm).
28 The authors stated, "The effects of CO appeared more pronounced after adjustment for
29 concurrent exposures to NO2, PM10, and ozone," although specific effects estimates for the
30 non-CO pollutants were not reported. In the multi-pollutant models, incidence of low birth
31 weight consistently increased with increasing ambient CO level. Effects estimates for the highest
October 1999 6-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 CO exposure category achieved statistical significance for births after the first birth, and for
2 births to mothers <20 years old, but not for all births.
3 Ritz and Yu (1999) provide justification for their choices of study population, exposure
4 period (third trimester), pollution averaging time (6:00 to 9:00 a.m.), and allowable distance of
5 subjects' residences from monitoring stations. They identify important, relevant unmeasured
6 factors such as maternal smoking, nutrition, prepregnancy weight, adverse pregnancy experience,
7 and occupational history. They acknowledge that confounding because of these factors is
8 possible. They also argue that their study design and analytical approach render serious
9 confounding unlikely.
10 Taken together, Alderman et al. (1987) and Ritz and Yu (1999) raise concern as to whether
11 contemporary ambient CO exposure is a risk factor for low birth weight. The findings have some
12 biological plausibility because the CO binding affinity of fetal hemoglobin is somewhat greater
13 than that of adult hemoglobin, and, at a given level of CO exposure, tissue O2 delivery is reduced
14 more in the fetus than in the child or the adult, in whom fetal hemoglobin has been replaced by
15 adult hemoglobin (Longo, 1976). The observation by Pereira et al. (1998) of an association of
16 ambient CO concentration with cord blood COHb level reinforces this concern. (Of course, the
17 fetus might also be unusually susceptible to hypoxia from exposure to agents other than CO.)
18 Both Alderman et al. (1987) and Ritz and Yu (1999) recommend further research with
19 individual-level measurements of CO exposure and relevant covariates.
20
21 6.2.1.5 Ambient Carbon Monoxide and Frequency of Respiratory Illness
22 Short-term variation in ambient CO levels has been associated in several studies with daily
23 variation in indices of respiratory illness frequency. In most cases, these indices reflect
24 exacerbation of preexisting respiratory illness. Associations of ambient CO with respiratory
25 illness frequency have been observed less frequently than with heart disease exacerbation. Also,
26 there is as yet no demonstrated biological mechanism by which CO at ambient exposure levels
27 could plausibly promote respiratory illness exacerbation or new respiratory illness. Therefore,
28 observed associations of ambient CO with such exacerbation should be interpreted very
29 cautiously, and should by no means be considered confirmed. At the same time, it is appropriate
30 to discuss these associations because they indicate that ambient CO exposure is not specifically
31 linked epidemiologically with heart disease. The correct interpretation of this lack of specificity
October 1999 6-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 is not yet known. On one hand, it could suggest that short-term ambient CO exposure effects are
2 not confined to the cardiovascular system. On the other hand, it could be taken as a caveat on the
3 observed associations of ambient CO with heart disease exacerbation, because even though the
4 pathophysiologic connection of CO with respiratory disease is more tenuous than with heart
5 disease, statistical associations of CO with respiratory disease frequency are nevertheless
6 observed.
7 Sunyer et al. (1991) investigated daily emergency department (ED) visits for chronic
8 obstructive pulmonary disease (COPD) in Barcelona, Spain, in relation to short-term ambient air
9 pollution levels, during 1985 and 1986. ED visits for asthma were excluded. Pollutants
10 considered were daily average and daily maximum hourly SO2, daily average black smoke (BS),
11 and daily maximum hourly CO, NO2, and O3. Ambient levels of CO were quite low during the
12 study (yearly mean of daily maximum hourly levels = 5.4 mg/m3, and 98th percentile
13 = 14.9 mg/m3). In single-pollutant regression models adjusted for meteorology, season, and day
14 of week, positive, statistically significant effects estimates were most consistently observed for
15 SO2 lagged 0 or 1 day, but not 2 days. Effects were significant for both daily average and
16 maximum hourly SO2 levels. Effects of BS and CO were also positive and statistically
17 significant (p < 0.01). Effects of NO2 and O3 were not significant. Air pollution effects on visits
18 were weaker in fall than in other seasons. Effects of SO2 remained positive and significant even
19 when days with daily average SO2 above 72 mg/m3 (181 days, 24.8% of all 730 study days) were
20 excluded from analysis.
21 Gordian et al. (1996) examined relationships between short-term ambient air pollution
22 levels and daily outpatient visits for asthma, bronchitis, and upper respiratory illness (e.g., sore
23 throat, sinusitis, earache, rhinitis) in Anchorage, AK, from May 1992 to March 1994. Numbers
24 of visits were derived from medical insurance claims by state and municipal employees, and their
25 dependents. Measurements of CO were made only during winter months. Briefly, an increase of
26 10 mg/m3 in PM10 was associated with a 3 to 6% increase in visits for asthma, and a 1 to 3%
27 increase in visits for upper respiratory illness. Winter CO concentrations were associated with
28 increased numbers of visits for bronchitis and upper respiratory illness, but not for asthma.
29 At the same time, these CO concentrations were tightly correlated with overall automobile
30 exhaust emissions, including NO2, fine particles, and VOCs such as benzene. Thus, visits for
31 respiratory illness could not be linked specifically to ambient CO exposure.
October 1999 6-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 Yang et al. (1997) investigated asthma emergency room visits in Reno, NV, in relation to
2 short-term ambient air pollution levels over 3 years, 1992 to 1994. Analytical methods were
3 similar to those used in Yang et al. (1998). Briefly, there was a positive, statistically significant
4 association of visits with the ambient O3 level lagged 0 days, but not 1 or 2 days. Visits were not
5 associated with ambient CO or PM10 levels.
6 Two reports investigating asthma exacerbation in relation to short-term ambient levels of
7 CO and other air pollutants have been published in 1999 by University of Washington
8 investigators (Sheppard et al., 1999; Norris et al., 1999). Taken together, these reports highlight
9 the difficulties inherent in efforts to specify single-pollutant effects in populations exposed to the
10 complex ambient air pollution mix. Sheppard et al. (1999) considered admissions for asthma to
11 23 greater Seattle hospitals from 1987 through 1994 in persons aged <65 years. Fifty-four
12 percent of these admissions occurred in persons less than 20 years old. Pollutants considered
13 were daily average CO, PM10, PM2 5 (fine PM), PM10_25 (coarse PM), and SO2, and daily
14 maximum 8-h average O3. The CO monitors were located in street canyons, not residential areas.
15 PM10 was measured with both the EPA reference method and with light-scattering nephelometry.
16 Fine PM levels were largely estimated from PM10 measurements. Coarse PM levels were
17 calculated as the difference between PM10 and fine PM levels. The PM measurements from
18 residential sites were given higher weight (80%) than those from an industrial site (20%).
19 During the study period, 50th-percentile ambient levels of average CO, PM10, fine PM, and
20 coarse PM were quite moderate (1.7 ppm and 27, 13, and 14 //g/m3, respectively).
21 Data were analyzed to treat all measured pollutants even-handedly. Semiparametric
22 Poisson regression models were used and included dummy variables for day of week. Multiple
23 lag times were considered; for each pollutant, the lag time showing the strongest statistical
24 association with admissions was selected. Single- and two-pollutant models were constructed,
25 and pollutant effects estimates were reported over interquartile ranges of ambient pollutant
26 concentrations.
27 In single-pollutant models, CO lagged 3 days and O3 lagged 2 days were associated most
28 strongly with asthma admissions. Associations of admissions with PM10, fine PM, and coarse
29 PM, each lagged 1 day, were also positive and statistically significant. The association of
30 admissions with SO2 (lagged 0 days) was positive but not significant. In season-specific,
31 two-pollutant models, generally similar effects were observed for CO lagged 3 days and fine PM
October 1999 6-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 lagged 1 day. Over all seasons, effects of both pollutants were positive and significant. Both
2 pollutants were positively, and at least marginally significantly, associated with admissions in
3 fall, winter, and spring. Both pollutants were negatively associated with admissions in summer.
4 Sheppard et al. (1999) stated, "In striving for a balanced approach to all measured pollutants...,
5 we observed unexpected associations for CO that dominated the PM effects. Nevertheless,...,
6 there is no evidence for an effect [of CO] on the underlying physiology of asthma. CO may be an
7 important environmental indicator of incomplete combustion, particularly from mobile sources."
8 Norris et al. (1999) investigated short-term ambient air pollution and emergency
9 department (ED) visits for asthma in Seattle children aged <18 years, from September 1995
10 through December 1996. Analyses were conducted for the entire study population, for urban
11 children (in whom the reported overall annual hospitalization rate was more than 600/100,000)
12 and for suburban children (in whom the reported overall rate was less than 100/100,000).
13 Pollutants considered were daily average CO, PM10, and PMX (as measured by nephelometer),
14 hourly average SO2, and daily maximum hourly average and daily average NO2. During the
15 study period, mean ambient levels of CO, PM10, and fine PM were 1.6 ppm, 21.7 //g/m3, and
16 12 //g/m3 (estimated), respectively. There were too few O3 measurements to include in analytical
17 models. Relative risks were reported over interquartile ranges of pollutant concentrations. Lag
18 times of 0 through 4 days were considered. Models were adjusted with dummy variables for day
19 of week, smoothing splines for time trends (approximately a 2-mo moving average), ambient
20 temperature, and dew point temperature.
21 Pollutant effects on asthma visits were usually reported with a 1-day lag. In single-
22 pollutant models, effects of PMX and PM10 were consistently positive and at least marginally
23 statistically significant. Effects of CO were also consistently positive, but not significant in the
24 urban children. Effects of daily average NO2, lagged 2 days, were positive and marginally
25 significant. Mixed results were observed for other pollution metrics. In multipollutant models,
26 effects of PMX and PM10 remained positive and statistically significant, but SO2 and NO2 effects
27 did not retain significance. Multipollutant models did not include CO because CO was assumed
28 a priori to be a surrogate for stagnant conditions. Norris et al. (1999) stated, "In summary, this
29 study found a small but significant association between air pollution and increased ED visits for
30 asthma...PM and CO concentrations...were associated with increased childhood ED visits for
31 asthma and represent the daily variation in incomplete combustion products...."
October 1999 6-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 The exclusion of CO from multi-pollutant models by Norris et at. (1999) appears dubious,
2 because both they and Sheppard et al. (1999) observed statistically significant associations of
3 both PM and CO with asthma exacerbation in single-pollutant models. The assumption that CO
4 is a surrogate for stagnant conditions may well be valid, but no rationale was presented for why
5 the same assumption could not be made for the other combustion-related air pollutants that were
6 included in the multipollutant models. Also, if ambient CO is merely a biologically inert fellow-
7 traveler with ambient PM, it would be difficult to understand why, in Sheppard et al. (1999), the
8 CO-asthma association was strongest when lagged 3 days, whereas the PM-asthma association
9 was strongest when lagged only 1 day. Furthermore, although it is true that there is no known
10 biological mechanism by which CO at ambient levels could exacerbate asthma, the mechanistic
11 linkage of combustion-related, nonbiological PM with asthma exacerbation remains to be more
12 clearly elucidated.
13 Norris et al. (1999) conclude, "Results from this study...show significant increases of ED
14 asthma visits...with daily PM25 concentrations substantially below the...standard of 15 //g/m3
15 annually." Although this statement is reasonable enough, it does not mention that associations of
16 asthma exacerbation with ambient CO levels were also observed, that ambient CO levels were
17 also below the existing CO standards, and that PM2 5 levels were largely estimated, whereas CO
18 levels were measured directly. Finally, the authors' assumption that children under 18 years old
19 are more "susceptible" than the general population is somewhat questionable, because short-term
20 elevations in ambient air pollution levels have been most strongly associated with unequivocally
21 harmful health effects in the elderly, and because asthma death rates and case-fatality rates are
22 higher in adults than in children. Taken together, Sheppard et al. (1999) and Norris et al. (1999)
23 show generally similar degrees of association of short-term fluctuations in ambient CO and PM
24 levels with frequency of asthma exacerbation in Seattle.
25
26 6.2.2 Controlled Laboratory Studies
27 The most extensive human experimental studies on the cardiovascular effects of CO have
28 been those conducted in predominantly young, healthy, nonsmoking subjects during exercise.
29 Previous assessments of these effects (U.S. Environmental Protection Agency, 1979, 1984, 1991;
30 Horvath, 1981; Shephard, 1983, 1984) have identified what appears to be a linear relationship
31 between the level of COFtb in the blood and decrements in human exercise performance,
October 1999 6-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 measured as maximal oxygen uptake. Short-term maximal exercise performance significantly
2 decreases at COHb levels ranging from 5 to 20% (Pirnay et al., 1971; Vogel and Gleser, 1972;
3 Ekblom and Huot, 1972; Weiser et al., 1978; Stewart et al., 1978; Klein et al., 1980; Koike and
4 Wasserman, 1992). One study (Horvath et al., 1975) observed a marginal decrease in maximal
5 exercise performance at a COHb level as low as 4.3% COHb. Short-term maximal exercise
6 duration also has been shown to be significantly reduced at COHb levels ranging from 2.3 to
7 20% (Ekblom and Huot, 1972; Drinkwater et al., 1974; Raven et al., 1974a,b; Horvath et al.,
8 1975; Weiser et al., 1978; Koike and Wasserman, 1992). The observed decreases in maximal
9 exercise performance and duration, however, are so small that they are only of concern primarily
10 for competing athletes, rather than for healthy people conducting everyday activities at less than
11 maximal exercise levels. In fact, no significant effects on oxygen uptake or on exercise
12 ventilation and heart rate were reported during submaximal exercise at COHb saturations as high
13 as 15 to 20% (see Section 10.3.2 in U.S. Environmental Protection Agency, 1991), especially at
14 work rates below the metabolic acidosis threshold (Koike et al., 1991).
15 Of greater concern at more typical ambient CO levels are certain cardiovascular effects
16 during exercise that are likely to occur in a smaller, but sizeable, segment of the general
17 population having a deficiency of blood supply (ischemia) to the heart muscle. This group of
18 patients with coronary artery disease (CAD) and reproducible exercise-induced angina (chest
19 pain) is regarded as the most sensitive risk group for CO-exposure effects. Several important
20 studies (Anderson et al., 1973; Sheps et al., 1987; Adams et al., 1988; Kleinman et al., 1989;
21 Allred et al., 1989a,b, 1991) have provided the cardiovascular database for CO in CAD patients.
22 In these studies, discussed in detail in the previous document (see Section 10.3.2 in U.S.
23 Environmental Protection Agency, 1991), significant ischemia was measured subjectively by the
24 time of exercise required for the development of angina (time of onset of angina) and objectively
25 by the time required to demonstrate a 1-mm change in the ST segment of the electrocardiogram.
26 Adverse effects were found with postexposure COHb levels as low as 3 to 6% when compared
27 on the basis of optical measurements (Figure 6-4). This represents incremental increases of
28 1.5 to 4.4% COHb from preexposure baseline levels. Effects on silent ischemia episodes
29 (no chest pain), which represent the majority of episodes in these patients, have not been studied.
30 Only one new study has become available since publication of the 1991 document. As part
31 of an investigation of CO exposure at high altitude, 17 men with documented CAD and stable
October 1999 6-43 DRAFT-DO NOT QUOTE OR CITE
-------
(0
_c
c
^
£
-------
1 5.40 min) during exercise after exposure to CO. The results are in good agreement with those
2 observed in the previously reported studies (see Figure 6-4). There was no statistically
3 significant effect on ST segment change, on the duration of angina, or on hemodynamic factors
4 such as blood pressure and heart rate.
5 Despite clearly demonstrable effects of low-level CO exposure in patients with ischemic
6 heart disease, the adverse health consequences of these types of effects are very difficult to
7 predict in the at-risk population of individuals with heart disease. There is a wide distribution of
8 professional judgments on the clinical significance of small performance decrements occurring
9 with the levels of exertion and CO exposure defined in the studies noted above. The decrements
10 in performance that have been described at the lowest levels (<3% COHb) are in the range of
11 reproducibility of the test and may not be alarming to some physicians. On the other hand, the
12 consistency of the responses in time to onset of angina across the studies and the dose-response
13 relationship described by Allred et al. (1989a,b, 1991) between COHb and time to ST segment
14 changes strengthen the argument in the minds of other physicians that, although small, the effects
15 could limit the activity of these individuals and affect their quality of life. In addition, it has been
16 argued by Bassan (1990) that 58% of cardiologists believe recurrent episodes of exertional
17 angina are associated with a substantial risk of precipitating a myocardial infarction (heart
18 attack), a fatal arrhythmia (abnormal heart rhythm), or slight but cumulative myocardial damage.
19 Exposures to low levels of CO resulting in 5 to 20% COHb do not produce significant
20 changes in cardiac rhythm or conduction during rest or exercise in healthy humans (Davies and
21 Smith, 1980; Kizakevich et al., 1994). Effects of CO on resting and exercise-induced ventricular
22 arrhythmia in patients with CAD are dependent on their clinical status. Hinderliter et al. (1989)
23 reported no effects of 4 and 6% COHb in patients with ischemic heart disease who did not have
24 chronic arrhythmia (ectopy) during baseline monitoring. In more severely compromised
25 individuals with higher levels of baseline ectopy, exposures to CO that produce 6% COHb have
26 been shown to significantly increase the number and complexity of arrhythmias (Sheps et al.,
27 1990), but not at lower COHb levels (Sheps et al., 1990, 1991; Chaitman et al., 1992; Dahms
28 et al., 1993). This finding, combined with epidemiologic evidence of CO-related morbidity and
29 mortality noted above, and the morbidity and mortality studies of workers who are routinely
30 exposed to combustion products (e.g., Stern et al., 1981, 1988; Edling and Axelson, 1984;
31 Sardinas et al., 1986; Michaels and Zoloth, 1991; Koskela, 1994; Melius, 1995; Strom et al.,
October 1999 6-45 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
12
13
14
15
1995) suggest that CO exposure may provide an increased risk of hospitalization or death in
patients with more severe heart disease.
There also is evidence from experimental studies with laboratory animals that CO can
adversely affect the cardiovascular system. The lowest-observed-effect level (LOEL) varies,
depending on the exposure regime used and species tested (see Table 6-6). Results from animal
studies (reviewed in U.S. Environmental Protection Agency, 1979, 1991; Turino, 1981;
McGrath, 1982; Penney, 1988, 1996a) suggest that inhaled CO can cause disturbances in cardiac
rhythm and conduction in healthy and cardiac-impaired animals that are consistent with the
human data. Results from animal studies (U.S. Environmental Protection Agency, 1991) also
indicate that inhaled CO can increase hemoglobin concentration and hematocrit ratio, probably
representing compensation for the reduction in oxygen transport caused by CO. At high CO
concentrations, excessive increases in hemoglobin and hematocrit may impose an additional
workload on the heart and compromise blood flow to the tissues.
TABLE 6-6. ESTIMATED LOWEST-OBSERVED-EFFECT LEVELS FOR
EXPOSURE OF LABORATORY ANIMALS TO CARBON MONOXIDE
LOEL
Health Effect Category
Cardiovascular effects
Cardiac rhythm
Cardiomegaly
Hemodynamics
Hematology
Atherosclerosis and thrombosis
Schedule-controlled behavior
Developmental effects
Lung morphology and function
CO (ppm)
50
200
150
100
250
330
60
5,000
COHb (%)
2.6
15.8
7.5
9.3
20.0
25.0
6.0
60.0
Duration
6 weeks"
30 days"
30 min
46 days"
10 weeksb
2h
21 days"
15 min
Species
Dog
Rat
Rat
Rat
Rabbit
Rat
Rat
Rat
Reference
Preziosi etal. (1970)
Penney etal. (1974)
Kanten etal. (1983)
Penney etal. (1974)
Daviesetal. (1976)
Merigan and
Mclntire (1976)
Prigge and
Hochrainer(1977)
Niden and Schulz
(1965)
"Continuous daily exposure.
Intermittent daily exposure, 4 h/day.
October 1999
6-46
DRAFT-DO NOT QUOTE OR CITE
-------
1 There is conflicting evidence that CO exposure enhances development of atherosclerosis in
2 laboratory animals, but most studies show no measurable effect when the animals are fed normal
3 diets without added cholesterol, even at high (-20%) COHb saturations (U.S. Environmental
4 Protection Agency, 1979, 1991; Penn et al., 1992; Penn, 1993; Mennear, 1993; Smith and
5 Steichen, 1993; Strom et al., 1995). Similarly, the possibility that CO promotes significant
6 changes in lipid metabolism that may accelerate atherosclerosis is suggested in only a few
7 laboratory animal studies (see Table 10-7 in U.S. Environmental Protection Agency, 1991) but
8 not in humans (Leaf and Kleinman, 1996b); however, any such effect must be subtle at most.
9 More recent in vitro studies utilizing cell culture techniques have explored the hypothesis that
10 CO causes cellular oxidative stress and leads to injuries of the vascular endothelium that may
11 precipitate atherosclerosis (Thorn and Ischiropoulos, 1997; Thorn et al., 1997). Unfortunately,
12 the ability of environmentally relevant CO concentrations to mediate this activity in the intact
13 organism has not been evaluated. Finally, CO probably inhibits rather than promotes platelet
14 aggregation (U.S. Environmental Protection Agency, 1991; Min et al., 1992), lending support to
15 forensic observations that thrombosis is not a prominent feature of CO-mediated injury.
16 In general, there are few data to indicate that an atherogenic effect of exposure is likely to occur
17 in human populations at frequently encountered levels of ambient CO.
18
19
20 6.3 CENTRAL NERVOUS SYSTEM AND BEHAVIORAL EFFECTS
21 6.3.1 Brain Oxygen Metabolism
22 6.3.1.1 Whole Brain
23 It has been documented amply in the literature that, as COHb is formed, vasodilation in the
24 brain (and increased blood supply) occurs in such proportions so as to keep the supply of oxygen
25 (O2) to the brain constant (Helfaer and Traystman, 1996; U.S. Environmental Protection Agency,
26 1991). The increased blood flow is sufficient to compensate not only for the oxygen supply
27 decrease caused by reduced arterial O2 content (CaO2) but is also sufficient to compensate for the
28 increased difficulty of extraction of O2 because of the shifted oxyhemoglobin dissociation curve.
29 This compensatory vasodilation appears to be effective from low levels to very high levels of
30 COHb (at least up to 60%) and is similar in the fetus, neonate, and healthy adult.
October 1999 6-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 Despite the compensatory regulation of O2 supply to the brain, it appears that
2 O2 consumption, measured as the cerebral metabolic rate for O2 (CMRO2), is reduced as COHb
3 rises. The reason for this is unclear, but the fact is well documented (Doblar et al., 1977; Jones
4 and Traystman, 1984; U.S. Environmental Protection Agency, 1991; Langston et al., 1996). The
5 amount of reduction in CMRO2 as a function of COHb can be seen by combining the information
6 of Doblar et al. (1977) from goats and Langston et al. (1996) from sheep into one graph (see
7 Figure 6-5). Although information from Jones and Traystman (1984) and associated studies was
8 expressed as a function of CaO2, not COFIb, and was difficult to incorporate into Figure 6-5 and
9 the associated analysis, their data corroborate those of the other workers.
10 From Figure 6-5, it may be seen that the CMRO2 does not decrease to 90% of baseline until
11 -27% COFIb (95% confidence limits were -21 to 32% COHb). The data from sheep and goats
12 agreed with the results of Paulson et al. (1973), who reported that the mean human CMRO2 did
13 not decrease significantly, even for COFIb up to 20%. Because Paulson et al. (1973) did not
14 report the value of their means, it was not possible to include their results as data points in
15 Figure 6-5.
16
17 6.3.1.2 Subregions of the Brain
18 There are a number of reports of the blood-flow response to COFIb of subregions of the
19 brain (U.S. Environmental Protection Agency, 1991). The results generally demonstrate that
20 some areas of the brain have less baseline blood flow than others, and that the COHb-
21 compensatory increase in blood flow is not the same for all areas. Generally, however, the
22 percent increases over baseline are nearly the same for all areas except the neurohypophysis
23 (Hanley et al., 1986). It is important to note that the latter area serves homoeostatic and not
24 ongoing behavioral functions. Thus, it would appear that the subregions of the brain have
25 compensatory increased blood flow in the presence of COHb similar to the whole brain. To be
26 sure, all possible regions of the brain have not been tested, but no evidence to indicate otherwise
27 has been found.
28 Work by Sinha et al. (1991), measuring regional capillary perfusion and blood flow in the
29 presence of COHb elevation, indicates that the problem of compensation for COHb-reduced
30 CaO2 is not as simple as indicated above. Blood flow was measured using radiolabeled dye and
31 capillary morphology was measured by fluorescence microscopy. With these methods, there
October 1999 6-48 DRAFT-DO NOT QUOTE OR CITE
-------
1.1-
0)
a*
03
.Q
C
o
"^
Q.
E
3
(0
c
o
o
c
0)
O)
>l
X
o
0.1-
0.0-
COHb(%)
Figure 6-5. The relationship between COHb and CMRO2. Means from Doblar et al.
(1977) were taken from their Tables 1 and 3, and CMRO2 values were
transformed to percent of baseline. Figures 1 and 3 of Langston et al. (1996)
were converted digitally (Summasketch III graphical to digital conversion) and
also were transformed to percent of baseline. Data from both sources were
merged into the same database and a logit function was fitted to the data using
SAS, PROC NLIN (SAS Institute Inc., 1990). The solid line is the best fit,
dashed lines are the 95% confidence limits, and the points plotted are means
from the published studies.
October 1999
6-49
DRAFT-DO NOT QUOTE OR CITE
-------
1 appeared to be an increase in the number of perfused capillaries and in the amount of blood flow
2 as COHb increased. Thus, not only may more blood be delivered, but increased capillary
3 perfusion would decrease the diffusion distance to the tissue.
4 Presumably the compensatory mechanisms in subareas of the brain would work in a
5 manner similar to that of the whole brain and thus would show similar decreases in CMRO2 as
6 COHb increases. No corroborative studies, however, have been reported in the literature.
7 Better and more detailed documentation of regional CMRO2in humans as well as other
8 species seems appropriate, but does not have high priority because not much evidence exists to
9 suggest that the results would differ from whole-brain results. It appears that what is needed is
10 not more descriptive work, but an effort should be made to understand the mechanism by which
11 COHb elevation reduces CMRO2. Furthermore, information is needed about brain conditions
12 under which brain compensatory mechanisms might be impaired (e.g., injury, inflammation,
13 ailments associated with aging and co-exposure to other pollutants). If such information were
14 available, specific theoretical (biologically based) predictions could be made, and behavioral
15 experiments designed to test them.
16
17 6.3.2 Behavioral Effects of Carbon Monoxide
18 The effects of CO on behavior, especially the ability to perform certain time discrimination
19 tasks, provided the scientific basis for the first CO NAAQS in 1971 (see Section 1.2). As further
20 research data became available, however, the results on human behavior at low levels of CO
21 exposure (<5% COHb) were called into question and subsequently dismissed as the basis for the
22 standard (U.S. Environmental Protection Agency, 1979). After reviewing available studies, the
23 previous criteria document (U.S. Environmental Protection Agency, 1991) concluded that effects
24 on behavior were demonstrated unambiguously in both humans and laboratory animals at COHb
25 elevations above 20%. Below this level, the results were less consistent. The document also
26 concluded, however, that it seems unwise to ignore the historical evidence in favor of effects on
27 human performance at COHb levels between 5 and 20% (e.g., Horvath et al., 1971; Fodor and
28 Winneke, 1972; Putz et al., 1976, 1979; Benignus et al., 1987). Even if behavioral effects are
29 small or occasional, they may be important to the performance of critical tasks.
30 Behavioral experiments with the effects of elevated COHb frequently have been marred by
31 methodological problems. In particular, experiments employing single-blind designs were shown
October 1999 6-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 to be 2.5 times as likely to find significant results as similar studies employing double-blind
2 methods (Benignus, 1993, 1996). This problem was noted previously, and reports of findings of
3 behavioral effects of CO were summarized with respect to whether a double-blind procedure had
4 been followed (see Table 10-25 of U.S. Environmental Protection Agency, 1991). From this
5 summary, it was concluded that, at most, there was credible evidence for effects on only three
6 (somewhat artificially defined) categories of behavior: (1) tracking, (2) vigilance, and
7 (3) continuous performance. Even within these categories, considering only double-blind
8 studies, it was noted that less than 50% of all studies found significant effects. Furthermore,
9 most of the double-blind experiments reporting significant results were unreplicable.
10 Benignus (1994) performed extensive meta-analyses of the CO-behavioral literature.
11 Because single-blind studies were likely to include many type I errors (results erroneously
12 declared significant), only double-blind human CO studies were included. In this report, two
13 dose-effect curves were estimated from the literature by converting all behavioral endpoints to
14 percent of baseline. A dose-effect curve for COHb and behavior was estimated from rat
15 experiments and corrected for the effects of hypothermia. The COHb for rats was estimated from
16 exposure conditions by use of a rat-specific version of the Coburn-Forster-Kane Equation
17 (Benignus and Annau, 1994). Another dose-effect curve was estimated from the human
18 literature on hypoxic hypoxia, which was converted to equivalent COHb via equal arterial
19 oxygen contents and corrected for effects of hypocapnia. These two curves virtually overlapped
20 each other. Human data points from CO-behavior experiments then were plotted onto the curve
21 fitted to the rat CO data (no curve was fitted to human data because of the small effect sizes and
22 small COHb levels). The conclusion from this meta-analysis was that human behavioral
23 impairments of 10% (ED-10) should not be expected until COHb exceeds 20%.
24 Data for the rat studies from Benignus (1994) were refitted for present purposes using the
25 same dose-effect model (a logit) as for the CMRO2 data above (originally a different function
26 was used). The results are plotted in Figure 6-6, in which the different studies are coded by
27 letters. With the logit function, it is estimated that a 10% decrement should be produced in rats
28 by ca. 25% COHb (95% confidence limits of -20 to 30%). Data from all available double-blind
29 human studies also were converted to percent of baseline and plotted (Figure 6-7) along with the
30 logit curve fitted to the rat data (from Figure 6-6). The human data plotted in Figure 6-7 were
31 mostly not statistically significant (thin solid lines) and seem, as a group, not to have a
October 1999 6-51 DRAFT-DO NOT QUOTE OR CITE
-------
1.H
0)
0.3-
COHb (%)
Figure 6-6. The relationship between COHb and behavior: effects in rats. Points plotted
are the hypothermia-corrected means from four studies of effects of COHb in
rats in which COHb was estimated by a rat-specific version of the Coburn-
Forster-Kane Equation (see Benignus, 1994). The solid line is a best-fit logit
curve to the rat data. The dashed lines are 95% confidence limits. The data
points are means from various studies coded by letter as follows: A = Ator
(1982), B = Ator et al. (1976), C = Smith et al. (1976), and D = Merigan and
Mclntire (1976).
October 1999
6-52
DRAFT-DO NOT QUOTE OR CITE
-------
1.H
\
\
0.7H
0
I '
10
I I I I
20
COHb (%)
I r
30
40
Figure 6-7. The relationship between COHb and behavior: effects in humans compared to
rats. Points plotted are means from various human studies. The heavy dashed
line is the line fitted to rat data as shown in Figure 6-6. The points for the
human data are connected by thin lines to depict the dose-effect curve found in
each study. Solid thin lines are from studies in which no significant effect was
found; dashed lines from studies in which a significant effect was reported.
The data points are coded such that alphabetic symbols are from
nonsignificant studies and other symbols from significant studies.
Investigators are coded as follows: X = Benignus et al. (1990), C = Benignus
et al. (1977), KC = Groll-Knapp et al. (1982), U = Harbin et al. (1988), O = Otto
et al. (1979), W = Ramsey (1973), P = Roche et al. (1981), E = Stewart et al.
(1973), V = Stewart et al. (1970), P = Weir et al. (1973), Z = Wright et al.
(1973), + = Benignus et al. (1987), $ = Putz et al. (1976), * = Putz et al. (1979),
and = Ramsey (1973).
October 1999
6-53
DRAFT-DO NOT QUOTE OR CITE
-------
1 dose-related trend of decrements, and it could be argued reasonably that effects in humans cannot
2 be shown to differ from those in rats. Some of the human data, however, at low levels of COHb
3 (4 to 10%) do appear below baseline and were declared statistically significant (thin dashed lines)
4 by the authors of the original reports.
5 The low-COHb significant results plotted in Figure 6-3 were invariably reported in studies
6 in which only a few levels of COHb were evaluated. Studies in which more and higher COHb
7 levels were tested invariably did not find statistically significant effects, even at much higher
8 levels. Furthermore, for every study reporting low-COHb level impairments, other studies failed
9 to replicate the findings, or highly similar studies failed to find effects.
10 In summary, no reliable evidence demonstrating decrements in neural or behavioral
11 function in healthy young adult humans has been reported for COHb levels below 20%, and even
12 these studies are untested by replications. The low-COHb behavioral effects that have
13 sometimes been reported cannot be taken at face value because they are not reliably repeatable,
14 and they do not fit into a wider range, dose-effect pattern reported in other studies. It is more
15 reasonable to conclude that no statistically detectable behavioral impairments occur until COHb
16 exceeds 20 to 30%. The conclusion, based on behavioral evidence alone, is bolstered by the
17 findings that whole-brain CMRO2 is not reduced by a similar amount until COHb rises to 21 to
18 32%. Because a dose-effect curve has been fitted, any level of effect may be considered
19 (e.g., ED-5 or ED-20). The interpolation of a curve to an ED-5 point would imply that the COHb
20 levels for such an effect size would be 15 to 26%. Such an interpolation is more speculative than
21 an ED-10, however, because the experimental verification would be difficult, requiring large
22 numbers of subjects and careful control of error variance. Additionally, as interpolation
23 approaches small effect sizes, the error possibility because of statistical model selection
24 (threshold versus continuous) increases dramatically.
25 Unless the effort is to find a behavioral paradigm that would yield replicable low-level
26 COHb decrements in behavior that were part of a wider range, dose-effect curve, it would seem
27 unfruitful to continue behavioral work. The lack of such dose-effect information within a study
28 has contributed to the problems of interpreting published literature. Behavioral work should be
29 encouraged, in an effort to determine whether reliable decrements in behavior are associated with
30 low levels of COHb. However, any new behavioral experiments should involve several CO
31 exposure levels, including a level high enough to produce changes. Failing that, inclusion of
October 1999 6-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 some other procedure or a reference dose of some other active substance would be useful to
2 verify the sensitivity of the behaviors under study. In addition, other experiments should be
3 designed to contribute to our understanding of how CMRO2 relates to COHb elevation and
4 behavioral changes.
5 Behaviors implicated by the research findings involve detection of infrequent events
6 (vigilance), hand-eye coordination (compensatory tracking), and other forms of continuous
7 performance (U.S. Environmental Protection Agency, 1991). Because of the unreliability of the
8 findings, however, it is questionable whether these behaviors should be cited as effects. Until
9 reliable behavioral effects are demonstrated in a dose-related manner, it is premature to speculate
10 about the kind of behavioral effects and, thus, to lend credence to unreliable findings.
11 Because COHb elevates brain blood flow, it has the possibility of altering the delivery of
12 other toxicants to the brain or altering the biotransformation or elimination of toxicants (e.g., Doi
13 and Tanaka, 1984; Kim and Carlson, 1983; Roth and Rubin, 1976a,b). In combination with
14 exercise or hypoxic hypoxia, the interactions would become even more complex. Disease and
15 ailments associated with aging concomitant with all of the above also could be important.
16 Interactions such as these are understood from physiological theory and could be given
17 quantitative estimates by the use of physiological simulation using whole-body physiological
18 models that are currently under development.
19
20
21 6.4 DEVELOPMENTAL TOXICITY
22 An issue directly relevant to the protection of public health is the potential effect of CO on
23 growth and function of the developing fetus, infant, and child. Results obtained from new
24 research on this specific outcome of CO exposure (e.g., Carratu et al., 1993, 1996; Di Giovanni
25 et al., 1993; De Salvia et al., 1995; De Luca et al., 1996) have not changed the conclusions
26 presented in Section 10.5 of the previous criteria document (U.S. Environmental Protection
27 Agency, 1991). From all of the laboratory animal studies, it is clear that severe, acute CO
28 poisoning can be fetotoxic, although specification of maternal and fetal COHb levels is difficult
29 because such exposures rarely involve the achievement of steady-state COHb levels or permit
30 careful and rapid determination of COHb levels. Available data (reviewed in U.S.
31 Environmental Protection Agency, 1991; Annau and Fechter, 1994; Carratu et al., 1995; Penney,
October 1999 6-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 1996b) provide strong evidence that maternal CO exposures of 150 to 200 ppm, leading to
2 approximately 15 to 25% COHb, produce reductions in birth weight, cardiomegaly, delays in
3 behavioral development, and disruption of cognitive function in laboratory animals of several
4 species. Isolated experiments (Prigge and Hochrainer, 1977; Abbatiello and Mohrmann, 1979;
5 Singh, 1986) suggest that some of these effects may be present at concentrations as low as 60 to
6 65 ppm (approximately 6 to 11% COHb) maintained throughout gestation (see Table 6-6).
7 Studies relating human CO exposure from ambient sources or cigarette smoking to reduced birth
8 weight (e.g., Martin and Bracken, 1986; Rubin et al., 1986; Alderman et al., 1987; Wouters
9 et al., 1987; Brooke et al., 1989; Spitzer et al., 1990; Wen et al., 1990; Peacock et al., 1991a;
10 Zaren et al., 1996; Jedrychowski and Flak, 1996; Seeker-Walker et al., 1997) are of concern
11 because of the risk for developmental disorders (Olds et al., 1994a,b; Olds, 1997); however,
12 many of these studies have not considered all sources of CO exposure, other pollutants (Wang
13 et al., 1997), or other risk factors during gestation (Peacock et al., 1991b; Luke, 1994; Robkin,
14 1997).
15 Results from laboratory animal studies suggest that exposure to lower levels of CO, leading
16 to < 10% COHb, should not have much of an effect on the developing fetus until possibly later in
17 gestation when the embryo is much larger and more dependent on transport of oxygen by red
18 blood cells (Robkin, 1997). In addition, results from a multicenter, prospective study (Koren
19 et al., 1991) of fetal outcome following mild to moderate accidental CO poisoning in pregnancy
20 suggest that hypoxemia associated with measured COHb saturations of up to 18% (or even
21 higher estimated levels) does not impair the growth potential of the fetus when pregnancy
22 continues normally. Therefore, it is very unlikely that ambient levels of CO typically
23 encountered by pregnant women would cause increased fetal risk. It is necessary, however, to
24 consider the combined effects of CO with the other common risk factors that may cause adverse
25 fetal outcome (e.g., tobacco and alcohol consumption, genetic background, maternal general
26 health, obstetric history).
27 One of the more important determinants of the course and outcome of pregnancy that was
28 not previously discussed is maternal-fetal nutrition (Luke, 1994). Laboratory animal studies
29 conducted to determine the combined effect of gestational CO exposure and nutritional
30 deficiency suggest that CO has a greater effect on the fetus in protein-deficient mice (Singh and
31 Moore-Cheatum, 1993; Singh et al., 1993). Reductions in the rate of pregnancy, lower fetal
October 1999 6-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 weights, and increased fetal malformations were reported at CO concentrations as low as 65 ppm
2 maintained between 6 and 23 h per day during the first trimester of pregnancy (gestational
3 days 8 through 18). Previous evidence of the fetotoxic and teratogenic effects of CO in
4 laboratory animals (U.S. Environmental Protection Agency, 1991) came largely from high levels
5 of exposure (i.e., in the range of 500 ppm for rodents).
6 There are studies (e.g., Schoendorf and Kiely, 1992; Scragg et al., 1993; Mitchell et al.,
7 1993; Klonoff-Cohen et al., 1995; Blair et al., 1996; Hutter and Blair, 1996; MacDorman et al.,
8 1997) linking maternal cigarette smoking with sudden infant death syndrome (SIDS), but the role
9 of CO is uncertain, especially in relation to other known risk factors for SIDS, such as
10 developmental abnormalities (Schwartz et al., 1998), prone sleeping (Kahn et al., 1993; Franco
11 et al., 1996), overheating (Douglas et al., 1996), and soft bedding (Ponsonby et al., 1993; Kemp
12 et al., 1998). Data from human populations (Hoppenbrouwers et al., 1981) suggesting a link
13 between ambient CO exposures and SIDS are weak, but further study should be encouraged.
14 Children may experience neurological symptoms such as dizziness or fainting after an acute
15 episode of CO poisoning (>15% COHb) or, in some cases, neurological impairment may develop
16 days to weeks after very high exposures (Crocker and Walker, 1985). Human data from these
17 cases of accidental high CO exposures are difficult to use in identifying a LOEL for CO because
18 of the small number of cases reviewed and problems in documenting exposure levels. However,
19 such data, if systematically gathered and reported, could be useful in identifying possible ages of
20 special sensitivity to CO and co-factors or other risk factors that might identify sensitive
21 subpopulations.
22
23
24 6.5 ACUTE PULMONARY EFFECTS
25 It is unlikely that CO has any direct effects on lung tissue, except for extremely high
26 concentrations that can cause cell damage and edema (Niden and Schulz, 1965; Fein et al., 1980;
27 Burns et al., 1986). No new information has been published in the literature to change this
28 conclusion drawn from Section 10.2 of the previous criteria document (U.S. Environmental
29 Protection Agency, 1991). Experimental studies on the effects of CO exposures producing
30 COHb saturations up to 56% failed to find any consistent effects on pulmonary cells and tissue or
31 on the vasculature of the lung (Fisher et al., 1969; Weissbecker et al., 1969; Hugod, 1980; Chen
October 1999 6-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 et al., 1982; Shimazu et al., 1990). Human studies on the pulmonary function effects of CO are
2 complicated by the lack of adequate exposure information, the small number of subjects studied,
3 and the short exposures explored. Decrements in lung function have been observed with
4 increasing severity of CO poisoning (Kolarzyk, 1994a,b, 1995). For example, occupational or
5 accidental exposure to the products of combustion and pyrolysis, particularly indoors, may lead
6 to acute decrements in lung function if COHb levels are greater than 17% (Sheppard et al., 1986)
7 but not at saturations less than 2% (Cooper and Alberti, 1984; Hagberg et al., 1985; Evans et al.,
8 1988). It is difficult, however, to separate the potential effects of CO from the effects of other
9 respiratory irritants in smoke and exhaust. Community population studies on CO in ambient air
10 have not found strong relationships with pulmonary function, symptomatology, and disease
11 (Lutz, 1983; Robertson and Lebowitz, 1984; Lebowitz et al., 1987).
12
13
14 6.6 OTHER SYSTEMIC EFFECTS OF CARBON MONOXIDE
15 Laboratory animal studies (reviewed in Section 10.6 of U.S. Environmental Protection
16 Agency, 1991) suggest that enzyme metabolism and the P-450-mediated metabolism of
17 xenobiotic compounds may be affected by CO exposure (e.g., Montgomery and Rubin, 1971;
18 Pankow et al., 1974; Roth and Rubin, 1976a,b,c). Most of the authors of these studies have
19 concluded, however, that effects on metabolism at low COHb levels (< 15%) are attributable
20 entirely to tissue hypoxia produced by increased levels of COHb because the effects are no
21 greater than those produced by comparable levels of hypoxia produced by insufficient oxygen
22 delivery. No new studies have been published at CO levels relevant to ambient exposures.
23 At higher levels of exposure, where COHb concentrations exceed 15 to 20%, there may be direct
24 inhibitory effects of CO on the activity of mixed-function oxidases, but more basic research is
25 needed. The decreases in xenobiotic metabolism shown with CO exposure may be important to
26 individuals receiving drug treatment.
27 Inhalation of high levels of CO, leading to COHb concentrations greater than 10 to 15%,
28 have been reported to cause a number of other systemic effects in laboratory animals and effects
29 in humans suffering from acute CO poisoning. Tissues of highly active oxygen metabolism, such
30 as heart, brain, liver, kidney, and muscle, may be particularly sensitive to CO poisoning. The
31 impairment of function in the heart and brain caused by CO exposure is well known and has been
October 1999 6-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 described above. Other systemic effects of CO poisoning are not as well known and are therefore
2 less certain. There are reports of effects on liver (Katsumata et al., 1980), kidney (Kuska et al.,
3 1980), bone (Zebro et al., 1983), and immune capacity in the lung and spleen (Snella and
4 Rylander, 1979). It generally is agreed that these effects are caused by the severe tissue damage
5 occurring during acute CO poisoning resulting from one or more of the following: ischemia
6 resulting from the formation of COHb, inhibition of oxygen release from oxyhemoglobin,
7 inhibition of cellular cytochrome function (e.g., cytochrome oxidases), and metabolic acidosis.
8
9
10 6.7 PHYSIOLOGIC RESPONSES TO CARBON MONOXIDE EXPOSURE
11 The only evidence for short- or long-term compensation to increased COHb levels in the
12 blood is indirect. Experimental animal data (reviewed in Section 10.7 of U.S. Environmental
13 Protection Agency, 1991) indicate that incremental increases in COHb produce physiological
14 responses that tend to offset the deleterious effects of CO exposure on oxygen delivery to the
15 tissues. Experimental human data (presented in a report by Kizakevich et al., 1994) indicate that
16 compensatory cardiovascular responses to submaximal upper- and lower-body exercise
17 (e.g., increased heart rate, cardiac contractility, cardiac output) occur after CO exposures. These
18 changes were highly significant for exposures attaining 20% COHb. Other compensatory
19 responses are increased coronary blood flow, cerebral blood flow, hemoglobin (through increased
20 hemopoiesis), and oxygen consumption in muscle.
21 Short-term compensatory responses in blood flow or oxygen consumption may not be
22 complete or may even be absent in certain persons. For example, from the laboratory animal
23 studies, it is known that coronary blood flow is increased with COHb, and, from human clinical
24 studies, it is known that subjects with ischemic heart disease respond to the lowest levels of
25 COHb (6% or less). The implication is that, in some cases of cardiac impairment, the short-term
26 compensatory mechanism is impaired.
27 From neurobehavorial studies (see Section 6.3.2 of the present document), it is apparent
28 that decrements resulting from CO exposure have not been consistent in all subjects, even in the
29 same studies, and have not demonstrated a dose-response relationship with increasing COHb
30 levels. The implication from these data suggests there may be some threshold or time lag in a
October 1999 6-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 compensatory mechanism such as increased blood flow. Without direct physiological evidence
2 in either laboratory animals or humans, this concept can be only hypothesized.
3 The mechanism by which long-term adaptation may occur, if it can be demonstrated in
4 humans, is assumed to be increased hemoglobin concentration via an increase in hemopoiesis.
5 This alteration in hemoglobin production has been demonstrated repeatedly in laboratory animal
6 studies, but no recent studies have been conducted that indicate the occurrence of some
7 adaptational benefit. Even if the hemoglobin increase is a signature of adaptation, it has not been
8 demonstrated at low ambient concentrations of CO.
9
10
11 6.8 COMBINED EXPOSURE OF CARBON MONOXIDE WITH OTHER
12 POLLUTANTS, DRUGS, AND ENVIRONMENTAL FACTORS
13 6.8.1 High-Altitude Effects
14 Although there are many studies comparing and contrasting the effects of inhaling CO with
15 those produced by short-term, high-altitude exposure, there are relatively few reports on the
16 combined effects of inhaling CO at high altitudes. There are data (reviewed in Section 11.1 of
17 U.S. Environmental Protection Agency, 1991) to support the possibility that the effects of these
18 two hypoxic factor episodes are at least additive. Most of these early data were obtained at CO
19 concentrations too high to have much meaning for regulating the amount of CO in ambient air.
20 More recent studies by Kleinman et al. (1998) evaluated the combined effects of lower levels of
21 CO at high altitude. In general, the results confirm the additivity of hypoxic effects at a
22 simulated altitude of 2.1 km and CO exposures resulting in 4% COHb.
23 There are even fewer studies of the long-term effects of CO at high altitude. These studies,
24 identified in Table 11-2 of the previous criteria document (U.S. Environmental Protection
25 Agency, 1991), indicate few changes at CO concentrations below 100 ppm and altitudes below
26 4,572 m (15,000 ft). The fetus may be particularly sensitive to the effects of CO at altitude
27 (Longo, 1976), as is especially true with the high levels of CO associated with maternal smoking
28 (Moore et al., 1982).
29 The potential effects on human health of inhaling CO at high altitudes are complex (see
30 Section 5.4.1) Whenever CO binds to hemoglobin (Hb), it reduces the amount of Hb available to
31 carry oxygen. People visiting high altitudes (where the partial pressure of oxygen in the
October 1999 6-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 atmosphere is lower) will experience reduced levels of oxygen in the blood (hypoxemia) because
2 of a relative hypoventilation that occurs, particularly during sleep. Carbon monoxide, by binding
3 to Hb, intensifies the hypoxemia existing at high altitudes by further reducing transport of oxygen
4 to the tissues. In addition, COHb saturations are higher at altitude than at sea level because,
5 in part, of changes in elimination of endogenous CO and of more rapid uptake of exogenous CO
6 (McGrath, 1992; McGrath et al., 1991, 1993). However, within hours of arrival at high altitude,
7 certain physiological adjustments begin to take place (Grover et al., 1986), and, over several
8 days, these mechanisms will operate to lessen the initial impact of atmospheric hypoxia.
9 Hemoconcentration occurs, and the increased Hb concentration offsets the decreased blood
10 oxygen saturation and restores oxygen concentrations to former levels. Consequently, the simple
11 additive model of COHb and altitude hypoxemia may be valid only during early altitude
12 exposure. The new visitor to higher altitudes, especially the elderly and those with CAD
13 (Kleinman et al., 1998; Leaf and Kleinman, 1996a), may be at greater risk from the added effects
14 of ambient CO than the adapted resident. The period of increased risk probably is prolonged in
15 the elderly because adaptation to high altitude proceeds more slowly with increasing age (Dill
16 etal., 1985).
17
18 6.8.2 Interaction with Drugs
19 There remains little direct information on the possible enhancement of CO toxicity by
20 concomitant drug use or abuse; however, there are some data suggesting cause for concern.
21 There is some evidence that interactions of drug effects with CO exposure can occur in both
22 directions, that is, CO toxicity may be enhanced by drug use, and the toxic or other effects of
23 drugs may be altered by CO exposure. Nearly all published data available on CO combinations
24 with drugs concern psychoactive drugs (Montgomery and Rubin, 1971, 1973; McMillan and
25 Miller, 1974; Medical College of Wisconsin, 1974; Pankow et al., 1974; Rockwell and
26 Weir, 1975; Roth and Rubin, 1976a,b,c; Mitchell et al., 1978; Topping et al., 1981; Kim and
27 Carlson, 1983; Engen, 1986; Knisely et al., 1987, 1989). Descriptions of these studies were
28 provided in Section 11.2 of the previous criteria document (U.S. Environmental Protection
29 Agency, 1991). The following summary, excerpted from the last review, still applies because
30 nothing significant has appeared in the recently published literature.
October 1999 6-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 The use and abuse of psychoactive drugs and alcohol are widespread. Because of the effect
2 of CO on brain function, interactions between CO and psychoactive drugs could be anticipated.
3 However, very little systematic research has addressed this question. In addition, very little of
4 the research that has been done has utilized models for expected effects from treatment
5 combinations. Thus, often it is not possible to assess whether the combined effects of drugs and
6 CO exposure are additive or differ from additivity. It is important to recognize that even additive
7 effects of combinations can be of clinical significance, especially when the individual is unaware
8 of the combined hazard. The greatest evidence for a potentially important interaction of CO
9 comes from studies with alcohol in both laboratory animals and humans, where at least additive
10 effects have been obtained (Mitchell et al., 1978; Knisely et al., 1987, 1989). The significance of
11 these effects is augmented by the probable high incidence of combined alcohol use and CO
12 exposure in the population.
13 Besides interaction with psychoactive drugs, there is growing concern that prescribed
14 medications, especially nitric oxide blockers and calcium channel blockers, could interact with
15 CO. There are no known published data available, however, on CO combinations with these
16 drugs.
17
18 6.8.3 Interaction with Other Air Pollutants and Environmental Factors
19 Much of the data concerning the combined effects of CO and other pollutants found in
20 ambient air are based on laboratory animal experiments that were discussed in Section 11.3 of
21 the previous criteria document (U.S. Environmental Protection Agency, 1991). More recent
22 studies published since then have confirmed the conclusions made at that time and are included
23 here for completeness. Only a few controlled-exposure studies of humans are available, and the
24 results were discussed in more detail in the previous document. These early studies in healthy
25 human subjects (Drinkwater et al., 1974; Raven et al., 1974a,b; Gliner et al., 1975; Hackney
26 et al., 1975a,b; DeLucia et al., 1983) on relevant concentrations of common air pollutants such as
27 CO, NO2, O3, and peroxyacetylnitrate failed to show any interaction from combined exposure.
28 The more recent epidemiology studies (e.g., Morris et al., 1995; Schwartz and Morris, 1995;
29 Schwartz, 1997, 1999; Burnett et al., 1997a,b; Morris and Naumova, 1998; Burnett et al., 1999)
30 suggest an association between hospital admissions for cardiovascular disease and ambient
31 exposure to multiple pollutants, including CO and PM. In animal studies, no interaction was
October 1999 6-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 observed following combined exposure of CO and common air pollutants such as NO2 and SO2
2 (Busey, 1972; Murray et al., 1978; Hugod, 1979). However, an additive effect on learning
3 behavior was observed following combined exposure of high levels (>100 ppm) of CO and NO
4 (Groll-Knapp et al., 1988), and a synergistic dose effect (increased COHb) was observed after
5 combined exposure to CO and O3 (Murphy, 1964).
6 Toxicological interactions of combustion products, primarily CO, carbon dioxide, NO2, and
7 hydrogen cyanide (HCN), at levels typically produced by indoor and outdoor fires, have shown a
8 synergistic effect on mortality following CO plus CO2 exposure (Rodkey and Collison, 1979;
9 Levin et al., 1987a) and CO plus NO2 exposure (Levin, 1996), and an additive effect with HCN
10 (Levin et al., 1987b). Additive effects on mortality also were observed when CO, HCN, and low
11 oxygen were combined; adding CO2 to this combination was synergistic (Levin et al., 1988).
12 Finally, laboratory animal studies (Young et al., 1987; Yang et al., 1988; Fechter et al.,
13 1988, 1997; Fechter, 1995; Gary et al., 1997) suggest that combinations of environmental factors
14 such as heat stress and noise may be important determinants of health effects occurring in
15 combination with CO exposure. Of the effects described, one potentially most relevant to typical
16 human exposures is a greater decrement in the exercise performance seen when heat stress is
17 combined with 50 ppm CO (Drinkwater et al., 1974; Raven et al., 1974a,b; Gliner et al., 1975).
18 All of studies discussed above involve interactions with exogenous exposure to CO. There
19 are endogenous sources of CO (e.g., heme degradation, peroxidative degradation of unsaturated
20 fatty acids, xenobiotic metabolism) that also can lead to increased COHb saturation. They are
21 discussed in more detail in Section 5.3 of this document. Possibly one of the greatest concerns
22 regarding potential risk in the population comes from inhalation exposure to the halogenated
23 hydrocarbons widely used as solvents, especially the dihalomethanes (e.g., methylene chloride
24 [dichloromethane], dibromomethane, diiodomethane, and bromochloromethane). There is some
25 indication from the older literature (Fodor and Roscovanu, 1976) that oral exposure to
26 trihalogenated methane derivatives also will lead to increased COHb. Other volatile solvents
27 (e.g., carbon tetrachloride, chloroform, methanol) were tested in laboratory animals, but none
28 produced increased levels of COHb (Pankow, 1996).
29 Methylene chloride provides the greatest potential exposure to the population because it has
30 been widely used as a paint remover, degreaser, and aerosol propellant (Wilcosky and Simonsen,
31 1991). When inhaled, it will undergo metabolic breakdown by cytochrome P-450 in liver to
October 1999 6-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 form CO, chloride, and CO2. Increased levels of CO from metabolic breakdown of exogenous
2 chemicals will increase COHb measured in the blood and add to the increased COHb levels
3 resulting from exogenous CO exposure (DiVincenzo and Kaplan, 1981a,b; Kurppa et al., 1981).
4 The metabolism to CO can be saturated, leading to a slower elimination of COHb than after CO
5 exposure (Pankow, 1996). Also, any co-exposures to other chemicals or drugs that affect
6 cytochrome P-450 also will affect COHb saturation (Kim and Kim, 1996; Wirkner et al., 1997).
7
8 6.8.4 Tobacco Smoke
9 Although tobacco smoke is another source of CO for smokers as well as nonsmokers, it is
10 also a source of other chemicals (e.g., nicotine, NO2, HCN, polyaromatic hydrocarbons [PAHs],
11 aldehydes, ketones) that could interact with environmental CO. Available data suggest that some
12 of these components can affect the cardiovascular system. For example, nicotine clearly
13 aggravates the decrease in oxygen capacity induced by CO through an increase in the oxygen
14 demand of the heart (Khosla et al., 1994; Benowitz, 1997), and PAHs have been implicated in
15 atherosclerosis (Glantz and Parmley, 1991). Little is known, however, about the relative
16 importance of CO compared with the other components of tobacco smoke.
17 The association between active smoking and CVD is fully established (Surgeon General of
18 the United States, 1983). Passive smoking exposes an individual to all components in the
19 cigarette smoke, but the CO component dominates heavily because only 1% or less of the
20 nicotine is absorbed from environmental tobacco smoke (ETS), compared with 100% in an active
21 smoker (Wall et al., 1988; Jarvis, 1987). Therefore, passive smoking will be closer to pure CO
22 exposure than active smoking, even if the resultant levels of COHb are low (about 1 to 2%)
23 (Jarvis, 1987). The relationship between passive smoking and increased risk of CVD is
24 controversial. Early studies on this relationship were reviewed in the 1986 report of the Surgeon
25 General of the United States (1986) and by the National Research Council (1986). Since that
26 time, the epidemiological evidence linking passive smoking exposure to heart disease has
27 expanded rapidly. The available literature on the relationship between passive exposure to ETS
28 in the home and the risk of cardiovascular-associated morbidity or mortality in the nonsmoking
29 spouse of a smoker consists of numerous published reports (e.g., Glantz and Parmley, 1991;
30 Steenland, 1992; Wells, 1994; Kritz et al., 1995; LeVois and Layard, 1995; Steenland et al.,
31 1996; Kawachi et al., 1997; Howard et al., 1998; He et al., 1999). The data suggest that
October 1999 6-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 nonsmokers exposed to ETS had a relative risk of CVD of approximately 1.3 (95% CI of 1.2 to
2 1.4). The association of CVD with prolonged exposure to ETS could be caused by any number
3 of biochemical mechanisms, including greater platelet aggregation, endothelial cell damage,
4 reduced oxygen supply, greater oxygen demand, and the direct effects of CO (Kalmaz et al.,
5 1993; Zhu and Parmley, 1995; Weiss, 1996; Werner and Pearson, 1998). Unfortunately, given
6 the size of this association (25 to 30%) compared to active smoking (-75%), and the inherent
7 problems with the studies, it is sill not known, with accuracy, how much or even whether
8 exposure to ETS increases the risk of CVD (Bailar, 1999).
9
10
11 6.9 SUMMARY
12 The effects of exposure to low CO concentrations, such as the levels found in ambient air,
13 are far more subtle and considerably less threatening than those occurring in frank poisoning
14 from high CO concentrations. Because the COHb level of the blood is the best indicator of
15 potential health risk, symptoms of exposures to excessive ambient air levels of CO are described
16 here in terms of associated COHb levels. The LOEL, however, depends on the method used for
17 analysis of COHb. Gas chromatography (GC) is the method of choice for measuring COHb,
18 particularly at saturation levels <5%, because of the large variability and potential high bias of
19 the optical methods such as CO-Ox. The key human health effects most clearly demonstrated to
20 be associated with exposure to ambient CO are summarized in Table 6-7.
21 Maximal exercise duration and performance in healthy individuals have been shown to be
22 reduced at COHb levels of >2.3% and >4.3% (GC), respectively. The decrements in
23 performance at these levels are small and likely to affect only competing athletes rather than
24 people engaged in everyday activities. In fact, no effects were observed during submaximal
25 exercise in healthy individuals at COHb levels as high as 15 to 20%.
26 Adverse effects have been observed in individuals with CAD at 3 to 6% COHb by optical
27 methods of measurement. At these levels, individuals with reproducible exercise-induced angina
28 (chest pain) are likely to experience a reduced capacity to exercise because of decreased time to
29 onset of angina. The indicators of myocardial ischemia during exercise, which is detectable by
October 1999 6-65 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 6-7. KEY HEALTH EFFECTS OF EXPOSURE TO AMBIENT
CARBON MONOXIDE
Target
Organ
Health Effects3*
Tested Population0
References
Lungs
Heart
Heart
Heart
Heart
Brain
Reduced maximal exercise duration with 1-h
peak CO exposures resulting in >2.3% COHb
(GC)
Reduced time to ST segment change of the
ECG (earlier onset of myocardial ischemia)
with peak CO exposures resulting in >2.4%
COHb (GC)
Reduced exercise duration because of increased
chest pain (angina) with peak CO exposures
resulting in >3% COHb (CO-Ox)
Healthy individuals
Individuals with
coronary artery
disease
Individuals with
coronary artery
disease
Increased number and complexity of
arrhythmias (abnormal heart rhythm) with
peak CO exposures resulting in >6% COHb
(CO-Ox)
Increased hospital admissions associated
with ambient pollutant exposures
Central nervous system effects, such as
decrements in hand-eye coordination (driving
or tracking) and in attention or vigilance
(detection of infrequent events), with 1-h peak
CO exposures (~5 to 20% COHb)
Individuals with
coronary artery
disease and high
baseline ectopy
(chronic arrhythmia)
Individuals >65 years
old with
cardiovascular
disease
Healthy individuals
Drinkwateretal. (1974)
Raven etal. (1974b)
Horvathetal. (1975)
Allredetal. (1989a,b;
1991)
Anderson et al. (1973)
Shepsetal. (1987)
Adams etal. (1988)
Kleinmanetal. (1989,
1998)
Allredetal. (1989a,b;
1991)
Sheps etal. (1990)
Schwartz and Morris
(1995)
Morris etal. (1995)
Schwartz (1997)
Burnett etal. (1997a)
Horvathetal. (1971)
Fodor and Winneke (1972)
Putzetal. (1976, 1979)
Benignus et al. (1987)
aThe EPA has set significant harm levels of 50 ppm (8-h average), 75 ppm (4-h average), and 125 ppm (1-h
average). Exposure under these conditions could result in COHb levels of 5 to 10% and cause significant health
effects in sensitive individuals.
bMeasured blood COHb level after CO exposure.
Tetuses, infants, pregnant women, elderly people, and people with anemia or with a history of cardiac or
respiratory disease may be particularly sensitive to CO.
1 electrocardiographic (ECG) changes (ST depression) and associated angina, were statistically
2 significant in one study at >2.4% COHb (GC) and showed a dose-response relationship with
3 increasing COHb. An increase in the number and complexity of exercise-related arrhythmias
October 1999
6-66
DRAFT-DO NOT QUOTE OR CITE
-------
1 also has been observed at >6% COHb (CO-Ox) in some people with CAD and high levels of
2 baseline ectopy (a chronic arrhythmia) that may present an increased risk of sudden death.
3 In community epidemiologic studies, short-term fluctuations in ambient CO concentration
4 have been associated quite frequently with heart disease exacerbation, especially in the elderly.
5 Short-term ambient CO concentrations have been associated somewhat less frequently with daily
6 mortality. Results from experimental laboratory studies, and reports of increased morbidity and
7 mortality in workers routinely exposed to combustion products, provide some qualitative support
8 for these community-level studies. However, the available epidemiologic database is
9 inconclusive as to whether observed statistical associations reflect actual and specific effects of
10 ambient CO exposure. Further research is needed on short-term and long-term exposure to
11 ambient CO and other combustion-related air pollutants, and on the relative influence of
12 exposure to pollutants from non-ambient sources.
13 Central nervous system effects, including reductions in hand-eye coordination (driving or
14 tracking) and in attention or vigilance, have been reported at peak COHb levels of 5% and
15 higher, but later work indicates that significant behavioral impairments in healthy individuals
16 should not be expected until COHb levels exceed 20%. It must be emphasized, however, that
17 even a 5% COHb level is associated with 1-h CO concentrations of 100 ppm or higher. Thus, at
18 typical ambient air levels of CO, no observable central nervous system effects would be expected
19 to occur in the healthy population.
20 The current ambient air quality standards for CO (9 ppm for 8 h and 35 ppm for 1 h) are
21 intended to keep COHb levels below 2.1% to protect the most sensitive members of the general
22 population (i.e., individuals with CAD). Individuals in motor vehicles are at the greatest risk
23 from ambient CO exposure, followed by pedestrians, bicyclists, and joggers in the proximity of
24 roadways and the rest of the general urban population exposed to vehicle exhaust. Several hours
25 of exposure to peak ambient CO concentrations found occasionally at downtown urban sites
26 during periods of heavy traffic would be required to produce COHb levels of concern in the most
27 sensitive nonsmokers. Carbon monoxide levels occurring outside the downtown urban locations
28 are expected to be lower and are probably more representative of levels found in residential areas
29 where most people live. Significant health effects from ambient CO exposure are not likely
30 under these latter exposure conditions. Active cigarette smoking increases the risk for
31 developing cardiovascular and pulmonary disease, and passive smoking also can elevate COHb
October 1999 6-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 levels in nonsmokers under conditions of poor ventilation, putting nonsmoking co-workers and
2 family members at increased risk. Carbon monoxide poisoning from indoor exposures to higher
3 than ambient CO levels occurs frequently, has more severe consequences and often is
4 overlooked.
5
October 1999 6-68 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abbatiello, E. R.; Mohrmann, K. (1979) Effects on the offspring of chronic low exposure carbon monoxide during
3 mice pregnancy. Clin. Toxicol. 14: 401-406.
4 Adams, K. F.; Koch, G.; Chatterjee, B.; Goldstein, G. M; O'Neil, J. I; Bromberg, P. A.; Sheps, D. S.;
5 McAllister, S.; Price, C. J.; Bissette, J. (1988) Acute elevation of blood carboxyhemoglobin to 6% impairs
6 exercise performance and aggravates symptoms in patients with ischemic heart disease. J. Am. Coll. Cardiol.
7 12:900-909.
8 Alderman, B. W.; Baron, A. E.; Savitz, D. A. (1987) Maternal exposure to neighborhood carbon monoxide and risk
9 of low infant birth weight. Public Health Rep. 102: 410-414.
10 Allred, E. N.; Bleecker, E. R.; Chaitman, B. R.; Dahms, T. E.; Gottlieb, S. O.; Hackney, J. D.; Pagano, M.;
11 Selvester, R. H.; Walden, S. M.; Warren, J. (1989a) Short-term effects of carbon monoxide exposure on the
12 exercise performance of subjects with coronary artery disease. N. Engl. J. Med. 321: 1426-1432.
13 Allred, E. N.; Bleecker, E. R.; Chaitman, B. R.; Dahms, T. E.; Gottlieb, S. O.; Hackney, J. D.; Hayes, D.;
14 Pagano, M.; Selvester, R. H.; Walden, S. M.; Warren, J. (1989b) Acute effects of carbon monoxide exposure
15 on individuals with coronary artery disease. Cambridge, MA: Health Effects Institute; research report no. 25.
16 Allred, E. N.; Bleecker, E. R.; Chaitman, B. R.; Dahms, T. E.; Gottlieb, S. O.; Hackney, J. D.; Pagano, M.;
17 Selvester, R. H.; Walden, S. M.; Warren, J. (1991) Effects of carbon monoxide on myocardial ischemia.
18 Environ. Health Perspect. 91: 89-132.
19 American Heart Association. (1997) 1998 heart and stroke statistical update. Dallas, TX: American Heart
20 Association. Available online at: www.amhrt.org/Scientific/HSstats98/index.html.
21 Anderson, E. W.; Andelman, R. J.; Strauch, J. M.; Fortuin, N. J.; Knelson, J. H. (1973) Effect of low-level carbon
22 monoxide exposure on onset and duration of angina pectoris: a study in ten patients with ischemic heart
23 disease. Ann. Intern. Med. 79: 46-50.
24 Annau, Z.; Fechter, L. D. (1994) The effects of prenatal exposure to carbon monoxide. In: Needleman, H. L.;
25 Bellinger, D., eds. Prenatal exposure to toxicants: developmental consequences. Baltimore, MD: Johns
26 Hopkins University Press; pp. 249-267. (Annau, Z., ed. The Johns Hopkins series in environmental
27 toxicology).
28 Ator, N. A. (1982) Modulation of the behavioral effects of carbon monoxide by reinforcement contingencies.
29 Neurobehav. Toxicol. Teratol. 4: 51-61.
30 Ator, N. A.; Merigan, W. H., Jr.; Mclntire, R. W. (1976) The effects of brief exposures to carbon monoxide on
31 temporally differentiated responding. Environ. Res. 12: 81-91.
32 Bailar, J. C., III. (1999) Passive smoking, coronary heart disease, and meta-analysis [letter]. N. Engl. J. Med.
33 340: 958-959.
34 Bascom, R.; Bromberg, P. A.; Costa, D. L.; Devlin, R.; Dockery, D. W.; Frampton, M. W.; Lambert, W.; Samet,
35 J. M.; Speizer, F. E.; Utell, M. (1996) Health effects of outdoor air pollution [part 2]. Am. J. Respir. Crit.
36 Care Med. 153: 477-498.
37 Bassan, M. M. (1990) Sudden cardiac death [letter to the editor]. N. Engl. J. Med. 322: 272.
38 Benignus, V. A. (1993) Importance of experimenter-blind procedure in neurotoxicology. Neurotoxicol. Teratol.
39 15:45-49.
40 Benignus, V. A. (1994) Behavioral effects of carbon monoxide: meta analyses and extrapolations. J. Appl. Physiol.
41 76: 1310-1316.
42 Benignus, V. A. (1996) Behavioral effects of carbon monoxide exposure: results and mechanisms. In: Penney,
43 D. G., ed. Carbon monoxide. Boca Raton, FL: CRC Press; pp. 211-238.
44 Benignus, V. A.; Annau, Z. (1994) Carboxyhemoglobin formation due to carbon monoxide exposure in rats.
45 Toxicol. Appl. Pharmacol. 128: 151-157.
46 Benignus, V. A.; Otto, D. A.; Prah, J. D.; Benignus, G. (1977) Lack of effects of carbon monoxide on human
47 vigilance. Percept. Mot. Skills 45: 1007-1014.
48 Benignus, V. A.; Muller, K. E.; Barton, C. N.; Prah, J. D. (1987) Effect of low level carbon monoxide on
49 compensatory tracking and event monitoring. Neurotoxicol. Teratol. 9: 227-234.
50 Benignus, V. A.; Muller, K. E.; Malott, C. M. (1990) Dose-effects functions for carboxyhemoglobin and behavior.
51 Neurotoxicol. Teratol. 12: 111-118.
52 Benowitz, N. L. (1997) The role of nicotine in smoking-related cardiovascular disease. Preventive Med.
53 26:412-417.
October 1999 6-69 DRAFT-DO NOT QUOTE OR CITE
-------
1 Blair, P. S.; Fleming, P. I; Bensley, D.; Smith, I.; Bacon, C.; Taylor, E.; Berry, I; Golding, I; Tripp, J. (1996)
2 Smoking and the sudden infant death syndrome: results from 1993-5 case-control study for confidential
3 inquiry into stillbirths and deaths in infancy. Br. Med. J. 313: 195-198.
4 Brooke, O. G.; Anderson, H. R.; Bland, J. M; Peacock, J. L.; Stewart, C. M. (1989) Effects on birth weight of
5 smoking, alcohol, caffeine, socioeconimic factors, and psychosocial stress. Br. Med. J. 298: 795-801.
6 Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
7 paniculate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
8 142: 15-22.
9 Burnett, R. T.; Dales, R. E.; Brook, J. R.; Raizenne, M. E.; Krewski, D. (1997a) Association between ambient
10 carbon monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 Canadian cities.
11 Epidemiology 8: 162-167.
12 Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997b) The role of paniculate size and chemistry in the
13 association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
14 Environ. Health Perspect. 105: 614-620.
15 Burnett, R. T.; Cakmak, S.; Raizenne, M. E.; Stieb, D.; Vincent, R.; Krewski, D.; Brook, J. R.; Philips, O.;
16 Ozkaynak, H. (1998a) The association between ambient carbon monoxide levels and daily mortality in
17 Toronto, Canada. J. Air Waste Manage. Assoc. 48: 689-700.
18 Burnett, R. T.; Cakmak, S.; Brook, J. R. (1998b) The effect of the urban ambient air pollution mix on daily
19 mortality rates in 11 Canadian cities. Can. J. Public Health 89: 152-156.
20 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate and gaseous air
21 pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
22 Burns, T. R.; Greenberg, S. D.; Cartwright, J.; Jachimczyk, J. A. (1986) Smoke inhalation: anultrastructural study
23 of reaction to injury in the human alveolar wall. Environ. Res. 41: 447-457.
24 Busey, W. M. (1972) Chronic exposure of albino rats to certain airborne pollutants. New York, NY: Coordinating
25 Research Council, Inc.; report no. CRC-APRAC-CAPM-6-68-6. Available from: NTIS, Springfield, VA;
26 PB-212794.
27 Carratu, M. R.; Renna, G.; Giustino, A.; De Salvia, M. A.; Cuomo, V. (1993) Changes in peripheral nervous system
28 activity produced in rats by prenatal exposure to carbon monoxide. Arch. Toxicol. 67: 297-301.
29 Carratu, M. R.; Cagiano, R.; De Salvia, M. A.; Trabace, L.; Cuomo, V. (1995) Developmental neurotoxicity of
30 carbon monoxide. In: Degen, G. H.; Seiler, J. P.; Bentley, P., eds. Toxicology in transition: proceedings of
31 the 1994 EUROTOX congress; August 1994; Basel, Switzerland. Arch. Toxicol. Suppl. 17: 295-301.
32 Carratu, M. R.; Cagiano, R.; Cortese, L; Renna, G.; Siro Brigiani, G.; Trabace, L.; Cuomo, V. (1996) Myelination
33 disorders produced by prenatal exposure to moderate carbon monoxide concentrations in rat offspring.
34 Fundam. Clin. Pharmacol. 10: 172.
35 Gary, R.; Clarke, S.; Delic, J. (1997) Effects of combined exposure to noise and toxic substances—critical review of
36 the literature. Ann. Occup. Hyg. 41: 455-465.
37 Chaitman, B. R.; Dahms, T. E.; Byers, S.; Carroll, L. W.; Younis, L. T.; Wiens, R. D. (1992) Carbon monoxide
3 8 exposure of subjects with documented cardiac arrhythmias. Cambridge, MA: Health Effects Institute;
39 research report no. 52.
40 Chen, S.; Weller, M. A.; Penney, D. G. (1982) A study of free lung cells from young rats chronically exposed to
41 carbon monoxide from birth. Scanning Electron Microsc. 2: 859-867.
42 Cohen, S. L; Deane, M.; Goldsmith, J. R. (1969) Carbon monoxide and survival from myocardial infarction. Arch.
43 Environ. Health 19: 510-517.
44 Cooper, K. R.; Alberti, R. R. (1984) Effect of kerosene heater emissions on indoor air quality and pulmonary
45 function. Am. Rev. Respir. Dis. 129: 629-631.
46 Crocker, P. J.; Walker, J. S. (1985) Pediatric carbon monoxide toxicity. J. Emerg. Med. 3: 443-448.
47 Dahms, T. E.; Younis, L. T.; Wiens, R. D.; Zarnegar, S.; Byers, S. L.; Chaitman, B. R. (1993) Effects of carbon
48 monoxide exposure in patients with documented cardiac arrhythmias. J. Am. Coll. Cardiol. 21: 442-450.
49 Davies, D. M.; Smith, D. J. (1980) Electrocardiographic changes in healthy men during continuous low-level carbon
50 monoxide exposure. Environ. Res. 21: 197-206.
51 Davies, R. F.; Topping, D. L.; Turner, D. M. (1976) The effect of intermittent carbon monoxide exposure on
52 experimental atherosclerosis in the rabbit. Atherosclerosis 24: 527-536.
53 De Luca, A.; Pierno, S.; Tricarico, D.; Carratu, M. R.; Cagiano, R.; Cuomo, V; Camerino, D. C. (1996)
54 Developmental changes of membrane electrical properties of rat skeletal muscle fibers produced by prenatal
55 exposure to carbon monoxide. Environ. Toxicol. Pharmacol. 2: 213-221.
October 1999 6-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 DeLucia, A. J.; Whitaker, J. H.; Bryant, L. R. (1983) Effects of combined exposure to ozone and carbon monoxide
2 (CO) in humans. In: Lee, S. D.; Mustafa, M. G.; Mehlman, M. A., eds. International symposium on the
3 biomedical effects of ozone and related photochemical oxidants; March; Pinehurst, NC. Princeton, NJ:
4 Princeton Scientific Publishers, Inc.; pp. 145-159. (Advances in modern environmental toxicology: v. 5).
5 De Salvia, M. A.; Cagiano, R.; Carratu, M. R.; Di Giovanni, V.; Trabace, L.; Cuomo, V. (1995) Irreversible
6 impairment of active avoidance behavior in rats prenatally exposed to mild concentrations of carbon
7 monoxide. Psychopharmacology (Berlin) 122: 66-71.
8 Di Giovanni, V.; Cagiano, R.; De Salvia, M. A.; Giustino, A.; Lacomba, C.; Renna, G.; Cuomo, V. (1993)
9 Neurobehavioral changes produced in rats by prenatal exposure to carbon monoxide. Brain Res.
10 616: 126-131.
11 Dill, D. B.; Alexander, W. C.; Myhre, L. G.; Whinnery, J. E.; Tucker, D. M. (1985) Aerobic capacity of D.B. Dill,
12 1928-1984. Fed. Proc. 44: 1013.
13 DiVincenzo, G. D.; Kaplan, C. J. (1981a) Effect of exercise or smoking on the uptake, metabolism, and excretion of
14 methylene chloride vapor. Toxicol. Appl. Pharmacol. 59: 141-148.
15 DiVincenzo, G. D.; Kaplan, C. J. (198Ib) Uptake, metabolism, and elimination of methylene chloride vapor by
16 humans. Toxicol. Appl. Pharmacol. 59: 130-140.
17 Doblar, D. D.; Santiago, T. V.; Edelman, N. H. (1977) Correlation between ventilatory and cerebrovascular
18 responses to inhalation of CO. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 43: 455-462.
19 Doi, R.; Tanaka, H. (1984) Effects of carbon monoxide exposure on the distribution of methylmercury in mice.
20 J. Toxicol. Sci. 9: 11-22.
21 Douglas, A. S.; Allan, T. M.; Helms, P. J. (1996) Seasonality and the sudden infant death syndrome during 1987-9
22 and 1991-3 in Australia and Britain. Br. Med. J. 312: 1381-1383.
23 Drinkwater, B. L.; Raven, P. B.; Horvath, S. M.; Gliner, J. A.; Ruhling, R. O.; Bolduan, N. W.; Taguchi, S. (1974)
24 Air pollution, exercise, and heat stress. Arch. Environ. Health 28: 177-181.
25 Edling, C.; Axelson, O. (1984) Risk factors of coronary heart disease among personnel in a bus company. Int. Arch.
26 Occup. Environ. Health 54:181-183.
27 Ekblom, B.; Huot, R. (1972) Response to submaximal and maximal exercise at different levels of
28 carboxyhemoglobin. Acta Physiol. Scand. 86: 474-482.
29 Engen, T. (1986) The combined effect of carbon monoxide and alcohol on odor sensitivity. Environ. Int.
30 12:207-210.
31 Evans, R. G.; Webb, K.; Homan, S.; Ayres, S. M. (1988) Cross-sectional and longitudinal changes in pulmonary
32 function associated with automobile pollution among bridge and tunnel officers. Am. J. Ind. Med. 14: 25-36.
33 Fechter, L. D. (1995) Combined effects of noise and chemicals. Occup. Med. State of the Art Rev. 10: 609-621.
34 Fechter, L. D.; Young, J. S.; Carlisle, L. (1988) Potentiation of noise induced threshold shifts and hair cell loss by
35 carbon monoxide. Hear. Res. 34: 39-47.
36 Fechter, L. D.; Liu, Y.; Pearce, T. A. (1997) Cochlear protection from carbon monoxide exposure by free radical
37 blockers in the guinea pig. Toxicol. Appl. Pharmacol. 142: 47-55.
38 Fein, A.; Grossman, R. F.; Jones, J. G.; Hoeffel, J.; McKay, D. (1980) Carbon monoxide effect on alveolar
3 9 epithelial permeability. Chest 78: 726-731.
40 Fisher, A. B.; Hyde, R. W.; Baue, A. E.; Reif, J. S.; Kelly, D. F. (1969) Effect of carbon monoxide on function and
41 structure of the lung. J. Appl. Physiol. 26: 4-12.
42 Fodor, G. G.; Roscovanu, A. (1976) Erhoehter Blut-CO-Gehalt bei Mensch und Tier durch inkorperierte
43 halogenierte Kohlenwasserstoffe [Increased blood CO content in humans and animals caused by incorporated
44 halogenated hydrocarbons]. Zentralbl. Bakteriol. Parasitenkd. Infektionskrankh. Hyg. Abt. 1: Orig. Reihe
45 B 162: 34-40.
46 Fodor, G. G.; Winneke, G. (1972) Effect of low CO concentrations on resistance to monotony and on psychomotor
47 capacity. Staub Reinhalt. Luft 32: 46-54.
48 Franco, P.; Groswasser, J.; Sottiaux, M.; Broadfield, E.; Kahn, A. (1996) Decreased cardiac responses to auditory
49 stimulation during prone sleep. Pediatrics 97: 174-178.
50 Glantz, S. A.; Parmley, W. W. (1991) Passive smoking and heart disease: epidemiology, physiology, and
51 biochemistry. Circulation 83: 1-12.
52 Gliner, J. A.; Raven, P. B.; Horvath, S. M.; Drinkwater, B. L.; Sutton, J. C. (1975) Man's physiologic response to
53 long-term work during thermal and pollutant stress. J. Appl. Physiol. 39: 628-632.
54 Goldsmith, J. R.; Landaw, S. A. (1968) Carbon monoxide and human health. Science (Washington, DC)
55 162: 1352-1359.
October 1999 6-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gordian, M. E.; Ozkaynak, H.; Xue, J.; Morris, S. S.; Spengler, J. D. (1996) Paniculate air pollution and respiratory
2 disease in Anchorage, Alaska. Environ. Health Perspect. 104: 290-297.
3 Graves, E. J.; Owings, M. F. (1998) 1996 summary: National Hospital Discharge Survey. Advance data from vital
4 and health statistics; no. 301. Hyattsville, MD: National Center for Health Statistics; DHHS publication no.
5 (PHS) 98-1250. Available: www.cdc.gov/nchswww/releases/98facts/98sheets/ineinj.htm.
6 Groll-Knapp, E.; Haider, M.; Jenkner, H.; Liebich, H.; Neuberger, M.; Trimmel, M. (1982) Moderate carbon
7 monoxide exposure during sleep: neuro- and psychophysiological effects in young and elderly people.
8 Neurobehav. Toxicol. Teratol. 4: 709-716.
9 Groll-Knapp, E.; Haider, M.; Kienzl, K.; Handler, A.; Trimmel, M. (1988) Changes in discrimination learning and
10 brain activity (ERP's) due to combined exposure to NO and CO in rats. Toxicology 49: 441-447.
11 Grover, R. F.; Weil, J. V.; Reeves, J. T. (1986) Cardiovascular adaptation to exercise at high altitude. Exercise
12 Sport Sci. Rev. 14: 269-302.
13 Hackney, J. D.; Linn, W. S.; Mohler, J. G.; Pedersen, E. E.; Breisacher, P.; Russo, A. (1975a) Experimental studies
14 on human health effects of air pollutants: II. four-hour exposure to ozone alone and in combination with
15 other pollutant gases. Arch. Environ. Health 30: 379-384.
16 Hackney, J. D.; Linn, W. S.; Law, D. C.; Karuza, S. K.; Greenberg, H.; Buckley, R. D.; Pedersen, E. E. (1975b)
17 Experimental studies on human health effects of air pollutants: III. two-hour exposure to ozone alone and in
18 combination with other pollutant gases. Arch. Environ. Health 30: 385-390.
19 Hagberg, M.; Kolmodin-Hedman, B.; Lindahl, R.; Nilsson, C.-A.; Norstrom, A. (1985) Irritative complaints,
20 carboxyhemoglobin increase and minor ventilatory function changes due to exposure to chain-saw exhaust.
21 Eur. J. Respir. Dis. 66: 240-247.
22 Hanley, D. F.; Wilson, D. A.; Traystman, R. J. (1986) Effect of hypoxia and hypercapnia on neurohypophyseal
23 blood flow. Am. J. Physiol. 250: H7-H15.
24 Harbin, T. J.; Benignus, V. A.; Muller, K. E.; Barton, C. N. (1988) The effects of low-level carbon monoxide
25 exposure upon evoked cortical potentials in young and elderly men. Neurotoxicol. Teratol. 10: 93-100.
26 He, J.; Vupputuri, S.; Allen, K.; Prerost, M.; Hughes, J.; Whelton, P. K. (1999) Passive smoking and the risk of
27 coronary heart disease~a meta-analysis of epidemiologic studies. N. Engl. J. Med. 340: 920-926.
28 Helfaer, M. A.; Traystman, R. J. (1996) Cerebrovascular effects of carbon monoxide. In: Penney, D. G., ed. Carbon
29 monoxide. Boca Raton, FL: CRC Press; pp. 69-86.
30 Hexter, A. C.; Goldsmith, J. R. (1971) Carbon monoxide: association of community air pollution with mortality.
31 Science (Washington, DC) 172: 265-267.
32 Hinderliter, A. L.; Adams, K. F., Jr.; Price, C. J.; Herbst, M. C.; Koch, G.; Sheps, D. S. (1989) Effects of low-level
3 3 carbon monoxide exposure on resting and exercise-induced ventricular arrhythmias in patients with coronary
34 artery disease and no baseline ectopy. Arch. Environ. Health 44: 89-93.
35 Hoppenbrouwers, T.; Calub, M.; Arakawa, K.; Hodgman, J. E. (1981) Seasonal relationship of sudden infant death
36 syndrome and environmental pollutants. Am. J. Epidemiol. 113: 623-635.
37 Horvath, S. M. (1981) Impact of air quality in exercise performance. Exercise Sport Sci. Rev. 9: 265-296.
38 Horvath, S. M.; Dahms, T. E.; O'Hanlon, J. F. (1971) Carbon monoxide and human vigilance: a deleterious effect of
39 present urban concentrations. Arch. Environ. Health 23: 343-347.
40 Horvath, S. M.; Raven, P. B.; Dahms, T. E.; Gray, D. J. (1975) Maximal aerobic capacity at different levels of
41 carboxyhemoglobin. J. Appl. Physiol. 38: 300-303.
42 Howard, G.; Wagenknecht, L. E.; Burke, G. L.; Diez-Roux, A.; Evans, G. W.; McGovern, P.; Nieto, J.; Tell, G. S.
43 (1998) Cigarette smoking and progression of atherosclerosis: the atherosclerosis risk in communities (ARIC)
44 study. JAMA J. Am. Med. Assoc. 279: 119-124.
45 Hugod, C. (1979) Effect of exposure to 0.5 ppm hydrogen cyanide singly or combined with 200 ppm carbon
46 monoxide and/or 5 ppm nitric oxide on coronary arteries, aorta, pulmonary artery, and lungs in the rabbit. Int.
47 Arch. Occup. Environ. Health 44: 13-23.
48 Hugod, C. (1980) The effect of carbon monoxide exposure on morphology of lungs and pulmonary arteries in
49 rabbits: a light- and electron-microscopic study. Arch. Toxicol. 43: 273-281.
50 Hutter, C. D. D.; Blair, M. E. (1996) Carbon monoxide-does fetal exposure cause sudden infant death syndrome?
51 Med. Hypotheses 46: 1-4.
52 Institute of Medicine. (1985) Preventing low birth weight. Washington, DC: National Academy Press.
53 Ito, K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk subpopulations.
54 J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
October 1999 6-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ito, K.; Kinney, P.; Thurston, G. D. (1995) Variations in PM-10 concentrations within two metropolitan areas and
2 their implications for health effects analyses. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the
3 colloquium on paniculate air pollution and human mortality and morbidity, part II; January 1994; Irvine, CA.
4 Inhalation Toxicol. 7: 735-745.
5 Jarvis, M. J. (1987) Uptake of environmental tobacco smoke. In: O'Neill, I. K.; Brunnemann, K. D.; Dodet, B.;
6 Hoffmann, D., eds. Environmental carcinogens: methods of analysis and exposure measurement, v. 9 -
7 passive smoking. Lyon, France: International Agency for Research on Cancer; pp. 43-58. (IARC scientific
8 publications no. 81).
9 Jedrychowski, W.; Flak, E. (1996) Confronting the prenatal effects of active and passive tobacco smoking on the
10 birth weight of children. Cent. Eur. J. Public Health 4: 201 -205.
11 Jones, M. D., Jr.; Traystman, R. J. (1984) Cerebral oxygenation of the fetus, newborn, and adult. Semin. Perinatal.
12 8:205-216.
13 Joseph, K. S.; Kramer, M. S. (1996) Review of the evidence on fetal and early childhood antecedents of adult
14 chronic disease. Epidemiol. Rev. 18: 158-174.
15 Kahn, A.; Groswasser, J.; Sottiaux, M.; Rebuffat, E.; Franco, P.; Dramaix, M. (1993) Prone or supine body position
16 and sleep characteristics in infants. Pediatrics 91: 1112-1115.
17 Kalmaz, E. E.; Saunders, P. J.; Kalmaz, G. D. (1993) Effects of carbon monoxide on arterial endothelial
18 permeability and its possible relation to arterial disease. J. Occup. Med. Toxicol. 2:31-39.
19 Kanten, W. E.; Penney, D. G.; Francisco, K.; Thill, J. E. (1983) Hemodynamic responses to acute
20 carboxyhemoglobinemia in the rat. Am. J. Physiol. 244: H320-H327.
21 Katsumata, Y.; Aoki, M.; Oya, M.; Yada, S.; Suzuki, O. (1980) Liver damage in rats during acute carbon monoxide
22 poisoning. Forensic Sci. Int. 16: 119-123.
23 Kawachi, I.; Colditz, G. A.; Speizer, F. E.; Manson, J. E.; Stampfer, M. J.; Willett, W. C; Hennekens, C. H. (1997)
24 A prospective study of passive smoking and coronary heart disease. Circulation 95: 2374-2379.
25 Kelsall, J. E.; Samet, J. M.; Zeger, S. L.; Xu, J. (1997) Air pollution and mortality in Philadelphia, 1974-1988.
26 Am. J. Epidemiol. 146: 750-762.
27 Kemp, J. S.; Livne, M.; White, D. K.; Arfken, C. L. (1998) Softness and potential to cause rebreathing: differences
28 in bedding used by infants at high and low risk for sudden infant death syndrome. J. Pediatr. 132: 234-239.
29 Khosla, S.; Laddu, A.; Ehrenpreis, S.; Somberg, J. C. (1994) Cardiovascular effects of nicotine: relation to
30 deleterious effects of cigarette smoking. Am. Heart J. 127: 1669-1672.
31 Kim, Y. C.; Carlson, G. P. (1983) Effect of carbon monoxide inhalation exposure in mice on drug metabolism in
32 vivo. Toxicol. Lett. 19: 7-13.
33 Kim, S. K.; Kim, Y. C. (1996) Effect of a single administration of benzene, toluene or /w-xylene on
34 carboxyhaemoglobin elevation and metabolism of dichloromethane in rats. J. Appl. Toxicol. 16: 437-444.
35 Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
36 Environ. Res. 54: 99-120.
37 Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
3 8 In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
39 mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
40 Kizakevich, P. N.; Hazucha, M.; VanHoose, L.; McCartney, M. L.; Bolick, K.; Jochem, W. J.; Clayton, C. A.;
41 Alexander, B. V; Boehleche, B.; Hackney, A. J.; Sheps, D. S.; Clapp, L. (1994) Noninvasive ambulatory
42 assessment of cardiac function and myocardial ischemia in healthy subjects exposed to carbon monoxide.
43 Frederick, MD: U.S. Army Medical Research and Materiel Command; contract no. DAMD17-91-C-1007.
44 Available from: NTIS, Springfield, VA; AD-A286 014.
45 Klein, J. P.; Forster, H. V.; Stewart, R. D.; Wu, A. (1980) Hemoglobin affinity for oxygen during short-term
46 exhaustive exercise. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 48: 236-242.
47 Kleinman, M. T. (1992) Health effects of carbon monoxide. In: Lippmann, M., ed. Environmental toxicants: human
48 exposures and their health effects. New York, NY: Van Nostrand Reinhold; pp. 98-118.
49 Kleinman, M. T.; Leaf, D. (1991) Effects of exposure to low-level carbon monoxide at high altitude in sensitive
50 subjects. Sacramento, CA: California Air Resources Board. Available from: NTIS, Springfield, VA;
51 PB95-103842.
52 Kleinman, M. T.; Davidson, D. M.; Vandagriff, R. B.; Caiozzo, V. J.; Whittenberger, J. L. (1989) Effects of
53 short-term exposure to carbon monoxide in subjects with coronary artery disease. Arch. Environ. Health
54 44:361-369.
October 1999 6-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
2 effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
3 Health 53: 388-397.
4 Klonoff-Cohen, H. S.; Edelstein, S. L.; Lefkowitz, E. S.; Srinivasan, I. P.; Kaegi, D.; Chang, J. C.; Wiley, K. J.
5 (1995) The effect of passive smoking and tobacco exposure through breast milk on sudden infant death
6 syndrome. JAMA J. Am. Med. Assoc. 273: 795-798.
7 Knisely, J. S.; Rees, D. C.; Salay, J. M.; Balster, R. L.; Breen, T. J. (1987) Effects of intraperitoneal carbon
8 monoxide on fixed-ratio and screen-test performance in the mouse. Neurotoxicol. Teratol. 9: 221-225.
9 Knisely, J. S.; Rees, D. C.; Balster, R. L. (1989) Effects of carbon monoxide in combination with behaviorally
10 active drugs on fixed-ratio performance in the mouse. Neurotoxicol. Teratol. 11: 447-452.
11 Koike, A.; Wasserman, K. (1992) Effect of acute reduction in oxygen transport on parameters of aerobic function
12 during exercise. Ann. Acad. Med. Singapore 21: 14-22.
13 Koike, A.; Wasserman, K.; Armon, Y.; Weiler-Ravell, D. (1991) The work-rate-dependent effect of carbon
14 monoxide on ventilatory control during exercise. Respir. Physiol. 85: 169-183.
15 Kolarzyk, E. (1994a) The effect of acute carbon monoxide poisoning on the respiratory system efficiency. I. Values
16 of spirometric parameters in different degrees of poisoning. Int. J. Occup. Med. Environ. Health 7: 225-235.
17 Kolarzyk, E. (1994b) The effect of acute carbon monoxide poisoning on the respiratory system efficiency. II. Types
18 of ventilatory disorder and dynamics of changes according to the severity of carbon monoxide poisoning. Int.
19 J. Occup. Med. Environ. Health 7: 237-243.
20 Kolarzyk, E. (1995) Regulation of breathing in cases of acute carbon monoxide poisoning. Int. J. Occup. Med.
21 Environ. Health 8: 89-101.
22 Koren, G.; Sharav, T.; Pastuszak, A.; Garrettson, L. K.; Hill, K.; Samson, L; Rorem, M.; King, A.; Dolgin, J. E.
23 (1991) A multicenter, prospective study of fetal outcome following accidental carbon monoxide poisoning in
24 pregnancy. Reprod. Toxicol. 5: 397-403.
25 Koskela, R.-S. (1994) Cardiovascular diseases among foundry workers exposed to carbon monoxide. Scand.
26 J. Work Environ. Health 20: 286-293.
27 Kritz, H.; Schmid, P.; Sinzinger, H. (1995) Passive smoking and cardiovascular risk. Arch. Intern. Med.
28 155: 1942-1948.
29 Kuller, L. H.; Radford, E. P.; Swift, D.; Perper, J. A.; Fisher, R. (1975) Carbon monoxide and heart attacks. Arch.
30 Environ. Health 30: 477-482.
31 Kurppa, K.; Kivisto, H.; Vainio, H. (1981) Dichloromethane and carbon monoxide inhalation: carboxyhemoglobin
32 addition, and drug metabolizing enzymes in rat. Int. Arch. Occup. Environ. Health 49: 83-87.
33 Kuska, J.; Kokot, F.; Wnuk, R. (1980) Acute renal failure after exposure to carbon monoxide. Mater. Med. Pol.
34 (Engl. Ed.) 12: 236-238.
35 Langston, P.; Gorman, D.; Runciman, W.; Upton, R. (1996) The effect of carbon monoxide on oxygen metabolism
36 in the brains of awake sheep. Toxicology 114: 223-232.
37 Leaf, D. A.; Kleinman, M. T. (1996a) Urban ectopy in the mountains: carbon monoxide exposure at high altitude.
38 Arch. Environ. Health 51: 283-290.
39 Leaf, D. A.; Kleinman, M. T. (1996b) Acute exposure to carbon monoxide does not affect plasma lipids,
40 lipoproteins, and apolipoproteins. Angiology 47: 337-341.
41 Lebowitz, M. D.; Collins, L.; Holberg, C. J. (1987) Time series analyses of respiratory responses to indoor and
42 outdoor environmental phenomena. Environ. Res. 43: 332-341.
43 Levin, B. C. (1996) New research avenues in toxicology: 7-gas N-gas model, toxicant suppressants, and genetic
44 toxicology. In: Chaturvedi, A. K.; Sanders, D. C.; Higgins, E. A., eds. Proceedings of the international
45 colloquium on advances in combustion toxicology; April 1995; Oklahoma City, OK. Toxicology
46 115:89-106.
47 Levin, B. C.; Paabo, M.; Gurman, J. L.; Harris, S. E.; Braun, E. (1987a) Toxicological interactions between carbon
48 monoxide and carbon dioxide. Presented at: 16th conference of toxicology, Air Force Medical Research
49 Laboratory; October 1986; Dayton, OH. Toxicology 47: 135-164.
50 Levin, B. C.; Paabo, M.; Gurman, J. L.; Harris, S. E. (1987b) Effects of exposure to single or multiple combinations
51 of the predominant toxic gases and low oxygen atmospheres produced in fires. Fundam. Appl. Toxicol.
52 9:236-250.
October 1999 6-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 Levin, B. C.; Paabo, M; Gurman, J. L.; Clark, H. M; Yoklavich, M. F. (1988) Further studies of the lexicological
2 effects of different time exposures to the individual and combined fire gases—carbon monoxide, hydrogen
3 cyanide, carbon dioxide and reduced oxygen. In: Polyurethanes 88: proceedings of the SPI 31st annual
4 technical/marketing conference; October; Philadelphia, PA. Society of the Plastics Industry, Inc.;
5 pp. 249-252.
6 LeVois, M. E.; Layard, M. W. (1995) Publication bias in the environmental tobacco smoke/coronary heart disease
7 epidemiologic literature. Regul. Toxicol. Pharmacol. 21: 184-191.
8 Longo, L. D. (1976) Carbon monoxide: effects on oxygenation of the fetus in utero. Science (Washington, DC)
9 194: 523-525.
10 Longo, L. D. (1977) The biological effects of carbon monoxide on the pregnant woman, fetus, and newborn infant.
11 Am. J. Obstet. Gynecol. 129: 69-103.
12 Luke, B. (1994) Maternal-fetal nutrition. Clin. Obstet. Gynecol. 37: 93-109.
13 Lutz, L. J. (1983) Health effects of air pollution measured by outpatient visits. J. Fam. Pract. 16: 307-313.
14 MacDorman, M. F.; Cnattingius, S.; Hoffman, H. J.; Kramer, M. S.; Haglund, B. (1997) Sudden infant death
15 syndrome and smoking in the United States and Sweden. Am. J. Epidemiol. 146: 249-257.
16 Martin, T. R.; Bracken, M. B. (1986) Association of low birth weight with passive smoke exposure in pregnancy.
17 Am. J. Epidemiol. 124: 633-642.
18 May, D. S.; Kelly, J. J.; Mendlein, J. M.; Garbe, P. L. (1991) Surveillance of major causes of hospitalization among
19 the elderly, 1988. Morb. Mortal. Wkly. Rep. 40: 7-21.
20 McGrath, J. J. (1982) Physiological effects of carbon monoxide. In: McGrath, J. J.; Barnes, C. D., eds. Air pollution
21 - physiological effects. New York, NY: Academic Press; pp. 147-181.
22 McGrath, J. J. (1992) Effects of altitude on endogenous carboxyhemoglobin levels. J. Toxicol. Environ. Health
23 35: 127-133.
24 McGrath, J. J.; Lee, P.; Schreck, R. M. (1991) Effects of altitude on expired breath carbon monoxide levels.
25 Presented at: 84th annual meeting and exhibition of the Air & Waste Management Association; June;
26 Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-138.2.
27 McGrath, J. J.; Schreck, R. M.; Lee, P. S. (1993) Carboxyhemoglobin levels in humans: effects of altitude.
28 Inhalation Toxicol. 5: 241-249.
29 McMillan, D. E.; Miller, A. T., Jr. (1974) Interactions between carbon monoxide and <5?-amphetamine or
30 pentobarbital on schedule-controlled behavior. Environ. Res. 8: 53-63.
31 Medical College of Wisconsin. (1974) Exposure of humans to carbon monoxide combined with ingestion of ethyl
32 alcohol and the comparison of human performance when exposed for varying periods of time to carbon
33 monoxide. Milwaukee, WI: Medical College of Wisconsin; report no. MCOW-ENVM-CO-74-2. Available
34 from: NTIS, Springfield, VA; PB-242099.
35 Melius, J. M. (1995) Cardiovascular disease among firefighters. Occup. Med.: State of the Art Rev. 10: 821-827.
36 Mennear, J. H. (1993) Carbon monoxide and cardiovascular disease: an analysis of the weight of evidence. Regul.
37 Toxicol. Pharmacol. 17: 77-84.
3 8 Merigan, W. H.; Mclntire, R. W. (1976) Effects of carbon monoxide on responding under a progressive ratio
39 schedule in rats. Physiol. Behav. 16: 407-412.
40 Michaels, D.; Zoloth, S. R. (1991) Mortality among urban bus drivers. Int. J. Epidemiol. 20: 399-404.
41 Min, H. Y.; Wei, Z.; Quan, Q. M.; Rong, Z. (1992) Correlation between hemorheological parameters and
42 viscoelasticity of thrombus in rats after exposure to carbon monoxide. Clin. Hemorheol. 12: 593-598.
43 Mitchell, D. S.; Packham, S. C.; Fitzgerald, W. E. (1978) Effects of ethanol and carbon monoxide on two measures
44 of behavioral incapacitation of rats. Proc. West. Pharmacol. Soc. 21: 427-431.
45 Mitchell, E. A.; Ford, R. P. K.; Stewart, A. W.; Taylor, B. J.; Becroft, D. M. O.; Thompson, J. M. D.; Scragg, R.;
46 Hassall, I. B.; Barry, D. M. J.; Allen, E. M.; Roberts, A. P. (1993) Smoking and the sudden death syndrome.
47 Pediatrics 91: 893-896.
48 Montgomery, M. R.; Rubin, R. J. (1971) The effect of carbon monoxide inhalation on in vivo drug metabolism in
49 the rat. J. Pharmacol. Exp. Ther. 179: 465-473.
50 Montgomery, M. R.; Rubin, R. J. (1973) Oxygenation during inhibition of drug metabolism by carbon monoxide or
51 hypoxic hypoxia. J. Appl. Physiol. 35: 505-509.
52 Moore, L. G.; Rounds, S. S.; Jahnigen, D.; Grover, R. F.; Reeves, J. T. (1982) Infant birth weight is related to
53 maternal arterial oxygenation at high altitude. J. Appl. Physiol.: Respir. Environ. Exercise Physiol.
54 52:695-699.
October 1999 6-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 Morris, R. D.; Naumova, E. N. (1998) Carbon monoxide and hospital admissions for congestive heart failure:
2 evidence of an increased effect at low temperatures. Environ. Health Perspect. 106: 649-653.
3 Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and hospitalization for congestive
4 heart failure among elderly people in seven large US cities. Am. J. Public Health 85: 1361-1365.
5 Murphy, S. D. (1964) A review of effects on animals of exposure to auto exhaust and some of its components.
6 J. Air Pollut. Control Assoc. 14: 303-308.
7 Murray, F. J.; Schwetz, B. A.; Crawford, A. A.; Henck, J. W.; Staples, R. E. (1978) Teratogenic potential of sulfur
8 dioxide and carbon monoxide in mice and rabbits. In: Mahlum, D. D.; Sikov, M. R.; Hackett, P. L.; Andrew,
9 F. D., eds. Developmental toxicology of energy-related pollutants: proceedings of the seventeenth annual
10 Hanford biology symposium; October 1977; Richland, WA. Oak Ridge, TN: U.S. Department of Energy,
11 Technical Information Center; pp. 469-478. Available from: NTIS, Springfield, VA; CONF-771017.
12 National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health
13 effects. Washington, DC: National Academy Press.
14 Niden, A. H.; Schulz, H. (1965) The ultrastructural effects of carbon monoxide inhalation on the rat lung. Virchows
15 Arch. Pathol. Anat. Physiol. 339: 283-292.
16 Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V.; Shappard, L.; Stout, J. W. (1999) An association
17 between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
18 Perspect. 107: 489-493.
19 Olds, D. (1997) Tobacco exposure and impaired development: a review of the evidence. Ment. Retard. Dev. Disabil.
20 Res. Rev. 3:257-269.
21 Olds, D. L.; Henderson, C. R.; Tatelbaum, R. (1994a) Intellectual impairment in children of women who smoke
22 cigarettes during pregnancy. Pediatrics 93: 221-227.
23 Olds, D. L.; Henderson, C. R.; Tatelbaum, R. (1994b) Prevention of intellectual impairment in children of women
24 who smoke cigarettes during pregnancy. Pediatrics 93: 228-233.
25 Otto, D. A.; Benigus, V. A.; Prah, J. D. (1979) Carbon monoxide and human time discrimination: failure to
26 replicate Beard-Wertheim experiments. Aviat. Space Environ. Med. 50: 40-43.
27 Pankow, D. (1996) Carbon monoxide formation due to metabolism of xenobiotics. In: Penney, D. G., ed. Carbon
28 monoxide. Boca Raton, FL: CRC Press; pp. 25-43.
29 Pankow, D.; Ponsold, W.; Fritz, H. (1974) Combined effects of carbon monoxide and ethanol on the activities of
30 leucine aminopeptidase and glutamic-pyruvic transaminase in the plasma of rats. Arch. Toxicol. 32: 331-340.
31 Pantazopoulou, A.; Katsouyanni, K.; Kourea-Kremastinou, J.; Trichopoulos, D. (1995) Short-term effects of air
32 pollution on hospital emergency outpatient visits and admissions in the greater Athens, Greece area. Environ.
33 Res. 69:31-36.
34 Paulson, O. B.; Parving, H.-H.; Olesen, J.; Skinhoj, E. (1973) Influence of carbon monoxide and of hemodilution on
35 cerebral blood flow and blood gases in man. J. Appl. Physiol. 35: 111-116.
36 Peacock, J. L.; Bland, J. M.; Anderson, H. R.; Brooke, O. G. (1991a) Cigarette smoking and birthweight: type of
37 cigarette smoked and a possible threshold effect. Int. J. Epidemiol. 20: 405-412.
38 Peacock, J. L.; Bland, J. M.; Anderson, H. R. (1991b) Effects on birthweight of alcohol and caffeine consumption
39 in smoking women. J. Epidemiol. Commun. Health 45: 159-163.
40 Perm, A. (1993) Determination of the atherogenic potential of inhaled carbon monoxide. Cambridge, MA: Health
41 Effects Institute; research report number 57.
42 Perm, A.; Currie, J.; Snyder, C. (1992) Inhalation of carbon monoxide does not accelerate arteriosclerosis in
43 cockerels. Eur. J. Pharmacol. 228: 155-164.
44 Penney, D. G. (1988) A review: hemodynamic response to carbon monoxide. Environ. Health Perspect 77: 121-130.
45 Penney, D. G., ed. (1996a) Carbon monoxide. Boca Raton, FL: CRC Press.
46 Penney, D. G. (1996b) Effects of carbon monoxide exposure on developing animals and humans. In: Penney, D. G.,
47 ed. Carbon monoxide. Boca Raton, FL: CRC Press; pp. 109-144.
48 Penney, D.; Benjamin, M.; Dunham, E. (1974) Effect of carbon monoxide on cardiac weight as compared with
49 altitude effects. J. Appl. Physiol. 37: 80-84.
50 Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.; Singer, J. M.;
51 Bohm, G. M.; Saldiva, P. H. N. (1998) Association between air pollution and intrauterine mortality in Sao
52 Paulo, Brazil. Environ. Health Perspect. 106: 325-329.
53 Peters, K. D.; Kochanek, K. D.; Murphy, S. L.; (1998) Deaths: final data for 1996. Atlanta, GA: U.S. Centers for
54 Disease Control and Prevention, National Center for Health Statistics. (National vital statistics reports: v. 47,
55 no. 9). Available: www.cdc.gov/nchswww/releases/98facts/98sheets/finmort.htm.
October 1999 6-76 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pirnay, F.; Dujardin J.; Deroanne, R.; Petit, J. M. (1971) Muscular exercise during intoxication by carbon
2 monoxide. J. Appl. Physiol. 31: 573-575.
3 Poloniecki, J. D.; Atkinson, R. W.; Ponce de Leon, A.; Anderson, H. R. (1997) Daily time series for cardiovascular
4 hospital admissions and previous day's air pollution in London, UK. Occup. Environ. Med. 54: 535-540.
5 Ponsonby, A.-L.; Dwyer, T.; Gibbons, L. E.; Cochrane, J. A.; Wang, Y.-G. (1993) Factors potentiating the risk of
6 sudden infant death syndrome associated with the prone position. N. Engl. J. Med. 329: 377-382.
7 Preziosi, T. J.; Lindenberg, R.; Levy, D.; Christenson, M. (1970) An experimental investigation in animals of the
8 functional and morphologic effects of single and repeated exposures to high and low concentrations of
9 carbon monoxide. Ann. N. Y. Acad. Sci. 174: 369-384.
10 Prigge, E.; Hochrainer, D. (1977) Effects of carbon monoxide inhalation on erythropoiesis and cardiac hypertrophy
11 in fetal rats. Toxicol. Appl. Pharmacol. 42: 225-228.
12 Putz, V. R.; Johnson, B. L.; Setzer, J. V. (1976) Effects of CO on vigilance performance: effects of low level
13 carbon monoxide on divided attention, pitch discrimination, and the auditory evoked potential. Cincinnati,
14 OH: U.S. Department of Health, Education, and Welfare, National Institute for Occupational Safety and
15 Health; report no. NIOSH-77-124. Available from: NTIS, Springfield, VA; PB-274219.
16 Putz, V. R.; Johnson, B. L.; Setzer, J. V. (1979) A comparative study of the effects of carbon monoxide and
17 methylene chloride on human performance. J. Environ. Pathol. Toxicol. 2: 97-112.
18 Raven, P. B.; Drinkwater, B. L.; Horvath, S. M.; Ruhling, R. O.; Gliner, J. A.; Sutton, J. C.; Bolduan, N. W.
19 (1974a) Age, smoking habits, heat stress, and their interactive effects with carbon monoxide and
20 peroxyacetylnitrate on man's aerobic power. Int. J. Biometeorol. 18: 222-232.
21 Raven, P. B.; Drinkwater, B. L.; Ruhling, R. O.; Bolduan, N.; Taguchi, S.; Gliner, J.; Horvath, S. M. (1974b) Effect
22 of carbon monoxide and peroxyacetyl nitrate on man's maximal aerobic capacity. J. Appl. Physiol.
23 36:288-293.
24 Ritz, B.; Yu, F. (1999) The effect of ambient carbon monoxide on low birth weight among children born in southern
25 California between 1989 and 1993. Environ. Health Perspect. 107: 17-25.
26 Ramsey, J. M. (1973) Effects of single exposures of carbon monoxide on sensory and psychomotor response.
27 Am. Ind. Hyg. Assoc. J. 34: 212-216.
28 Robertson, G.; Lebowitz, M. D. (1984) Analysis of relationships between symptoms and environmental factors over
29 time. Environ. Res. 33: 130-143.
30 Robkin, M. A. (1997) Carbon monoxide and the embryo. Int. J. Dev. Biol. 41: 283-289.
31 Roche, S.; Horvath, S.; Gliner, J.; Wagner, J.; Borgia, J. (1981) Sustained visual attention and carbon monoxide:
32 elimination of adaptation effects. Hum. Factors 23: 175-184.
33 Rockwell, T. J.; Weir, F. W. (1975) The interactive effects of carbon monoxide and alcohol on driving skills.
34 Columbus, OH: The Ohio State University Research Foundation; CRC-APRAC project CAPM-9-69.
3 5 Available from: NTIS, Springfield, VA; PB-242266.
36 Rodkey, F. L.; Collison, H. A. (1979) Effects of oxygen and carbon dioxide on carbon monoxide toxicity.
37 J. Combust. Toxicol. 6: 208-212.
38 Roth, R. A., Jr.; Rubin, R. J. (1976a) Role of blood flow in carbon monoxide- and hypoxic hypoxia-induced
39 alterations in hexobarbital metabolism in rats. Drug Metab. Dispos. 4: 460-467.
40 Roth, R. A., Jr.; Rubin, R. J. (1976b) Comparison of the effect of carbon monoxide and of hypoxic hypoxia.
41 II. Hexobarbital metabolism in the isolated, perfused rat liver. J. Pharmacol. Exp. Ther. 199: 61-66.
42 Roth, R. A., Jr.; Rubin, R. J. (1976c) Comparison of the effect of carbon monoxide and of hypoxic hypoxia.
43 I. In vivo metabolism, distribution and action of hexobarbital. J. Pharmacol. Exp. Ther. 199: 53-60.
44 Rubin, D. H.; Leventhal, J. M.; Krasilnikoff, P. A.; Weile, B.; Berget, A. (1986) Effect of passive smoking on
45 birth-weight. Lancet (8504): 415-417.
46 Saldiva, P. H. N.; Lichtenfels, A. J. F. C.; Paiva, P. S. O.; Barone, I. A.; Martins, M. A.; Massad, E.; Pereira,
47 J. C. R.; Xavier, V. P.; Singer, J. M.; Bohm, G. M. (1994) Association between air pollution and mortality
48 due to respiratory diseases in children in Sao Paulo, Brazil: a preliminary report. Environ. Res. 65: 218-225.
49 Saldiva, P. H. N.; Pope, C. A., Ill; Schwartz, J.; Dockery, D. W.; Lichtenfels, A. J.; Salge, J. M.; Barone, I.; Bohm,
50 G. M. (1995) Air pollution and mortality in elderly people: a time-series study in Sao Paulo, Brazil. Arch.
51 Environ. Health 50: 159-163.
52 Salinas, M.; Vega, J. (1995) The effect of outdoor air pollution on mortality risk: an ecological study from Santiago,
53 Chile. World Health Stat. Q. 48: 118-125.
54 Sardinas, A.; Wang Miller, J.; Hansen, H. (1986) Ischemic heart disease mortality of fireman and policemen. Am. J.
55 Public Health 76: 1140-1141.
October 1999 6-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 SAS Institute Inc. (1990) The NLIN procedure. In: SAS/STAT® user's guide, version 6, fourth edition. Volume 2.
2 Gary, NC: SAS Institute Inc.; pp. 1135-1193.
3 Schoendorf, K. C.; Kiely, J. L. (1992) Relationship of sudden infant death syndrome to maternal smoking during
4 and after pregnancy. Pediatrics 90: 905-908.
5 Schwartz, J. (1997) Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology
6 8:371-377.
7 Schwartz, J. (1999) Air pollution and hospital admissions for heart disease in eight U.S. counties. Epidemiology
8 10: 17-22.
9 Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
10 Michigan. Am. J. Epidemiol. 142: 23-35.
11 Schwartz, P. J.; Stramba-Badiale, M.; Segantini, A.; Austoni, P.; Bosi, G.; Giorgetti, R.; Grancini, F.; Marni, E. D.;
12 Perticone, F.; Rosti, D.; Salice, P. (1998) Prolongation of the QT interval and the sudden infant death
13 syndrome. N. Engl. J. Med. 338: 1709-1761.
14 Scragg, R.; Mitchell, E. A.; Taylor, B. J.; Stewart, A. W.; Ford, R. P. K.; Thompson, J. M. D.; Allen, E. M.;
15 Becroft, D. M. O. (1993) Bed sharing, smoking, and alcohol in the sudden infant death syndrome. Br. Med J.
16 307: 1312-1318.
17 Seeker-Walker, R. H.; Vacek, P. M.; Flynn, B. S.; Mead, P. B. (1997) Smoking in pregnancy, exhaled carbon
18 monoxide, and birth weight. Obstet. Gynecol. 89: 648-653.
19 Shephard, R. J. (1983) Carbon monoxide: the silent killer. Springfield, IL: Charles C. Thomas. (American lecture
20 series no. 1059).
21 Shephard, R. J. (1984) Athletic performance and urban air pollution. Can. Med. Assoc. J. 131: 105-109.
22 Sheppard, D.; Distefano, S.; Morse, L.; Becker, C. (1986) Acute effects of routine firefighting on lung function.
23 Am. J. Ind. Med. 9: 333-340.
24 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
25 nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
26 Sheps, D. S.; Adams, K. F., Jr.; Bromberg, P. A.; Goldstein, G. M.; O'Neil, J. J.; Horstman, D.; Koch, G. (1987)
27 Lack of effect of low levels of carboxyhemoglobin on cardiovascular function in patients with ischemic heart
28 disease. Arch. Environ. Health 42: 108-116.
29 Sheps, D. S.; Herbst, M. C.; Hinderliter, A. L.; Adams, K. F.; Ekelund, L. G.; O'Neil, J. J.; Goldstein, G. M.;
30 Bromberg, P. A.; Dalton, J. L.; Ballenger, M. N.; Davis, S. M.; Koch, G. G. (1990) Production of
31 arrhythmias by elevated carboxyhemoglobin in patients with coronary artery disease. Ann. Intern. Med.
32 113:343-351.
33 Sheps, D. S.; Herbst, M. C.; Hinderliter, A. L.; Adams, K. F.; Ekelund, L. G.; O'Neil, J. J.; Goldstein, G. M.;
34 Bromberg, P. A.; Ballenger, M.; Davis, S. M.; Koch, G. (1991) Effects of 4 percent and 6 percent
3 5 carboxyhemoglobin on arrhythmia production in patients with coronary artery disease. Cambridge, MA:
3 6 Health Effects Institute; research report no. 41.
37 Shimazu, T.; Ikeuchi, H.; Hubbard, G. B.; Langlinais, P. C.; Mason, A. D., Jr.; Pruitt, B. A., Jr. (1990) Smoke
3 8 inhalation injury and the effect of carbon monoxide in the sheep model. J. Trauma 30: 170-175.
39 Singh, J. (1986) Early behavioral alterations in mice following prenatal carbon monoxide exposure.
40 Neurotoxicology 7: 475-481.
41 Singh, J.; Moore-Cheatum, L. (1993) Gestational protein deficiency enhances fetotoxicity of carbon monoxide.
42 In: Keen, C. L.; Bendich, A.; Willhite, C. C., eds. Maternal nutrition and pregnancy outcome. Ann. N. Y.
43 Acad. Sci. 678: 366-368.
44 Singh, J.; Aggison, L., Jr.; Moore-Cheatum, L. (1993) Teratogenicity and developmental toxicity of carbon
45 monoxide in protein-deficient mice. Teratology 48: 149-159.
46 Sinha, A. K.; Klein, J.; Schultze, P.; Weiss, J.; Weiss, H. R. (1991) Cerebral regional capillary perfusion and blood
47 flow after carbon monoxide exposure. J. Appl. Physiol. 71: 1196-1200.
48 Smith, C. J.; Steichen, T. J. (1993) The atherogenic potential of carbon monoxide. Atherosclerosis (Shannon, Irel.)
49 99: 137-149.
50 Smith, M. D.; Merigan, W. H.; Mclntire, R. W. (1976) Effects of carbon monoxide on fixed-consecutive-number
51 performance in rats. Pharmacol. Biochem. Behav. 5: 257-262.
52 Snella, M.-C.; Rylander, R. (1979) Alteration in local and systemic immune capacity after exposure to bursts of CO.
53 Environ. Res. 20: 74-79.
October 1999 6-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 Spitzer, W. O.; Lawrence, V.; Dales, R.; Hill, G.; Archer, M. C.; Clark, P.; Abenhaim, L.; Hardy, J.; Sampalis, I;
2 Pinfold, S. P.; Morgan, P. P. (1990) Links between passive smoking and disease: a best-evidence synthesis.
3 Clin. Invest. Med. 13: 17-42.
4 Steenland, K. (1992) Passive smoking and the risk of heart disease. JAMA J. Am. Med. Assoc. 267: 94-99.
5 Steenland, K.; Thun, M.; Lally, C.; Heath, C., Jr. (1996) Environmental tobacco smoke and coronary heart disease
6 in the American Cancer Society CPS-II cohort. Circulation 94: 622-628.
7 Stern, F. B.; Lemen, R. A.; Curtis, R. A. (1981) Exposure of motor vehicle examiners to carbon monoxide:
8 a historical prospective mortality study. Arch. Environ. Health 36: 59-66.
9 Stern, F. B.; Halperin, W. E.; Hornung, R. W.; Ringenburg, V. L.; McCammon, C. S. (1988) Heart disease mortality
10 among bridge and tunnel officers exposed to carbon monoxide. Am. J. Epidemiol. 128: 1276-1288.
11 Stewart, R. D.; Peterson, J. E.; Baretta, E. D.; Bachand, R. T.; Hosko, M. J.; Herrmann, A. A. (1970) Experimental
12 human exposure to carbon monoxide. Arch. Environ. Health 21: 154-164.
13 Stewart, R. D.; Newton, P. E.; Hosko, M. J.; Peterson, J. E. (1973) Effect of carbon monoxide on time perception.
14 Arch. Environ. Health 27: 155-160.
15 Stewart, R. D.; Newton, P. E.; Kaufman, J.; Forster, H. V.; Klein, J. P.; Keelen, M. H., Jr.; Stewart, D. J.; Wu, A.;
16 Hake, C. L. (1978) The effect of a rapid 4% carboxyhemoglobin saturation increase on maximal treadmill
17 exercise. New York, NY: Coordinating Research Council, Inc.; report no. CRC-APRAC-CAPM-22-75.
18 Available from: NTIS, Springfield, VA; PB-296627.
19 Strom, J.; Alfredsson, L.; Malmfors, T. (1995) Carbon monoxide: causation and aggravation of cardiovascular
20 diseases - a review of the epidemiologic and toxicologic literature. Indoor Environ. 4: 322-333.
21 Sunyer, J.; Anto, J. M.; Murillo, C.; Saez, M. (1991) Effects of urban air pollution on emergency room admissions
22 for chronic obstructive pulmonary disease. Am. J. Epidemiol. 134: 277-286.
23 Surgeon General of the United States. (1983) The health consequences of smoking: cardiovascular disease - a report
24 of the Surgeon General. Rockville, MD: U. S. Department of Health and Human Services, Public Health
25 Service, Office on Smoking and Health; publication no. DHHS(PHS) 84-50204.
26 Surgeon General of the United States. (1986) The health consequences of involuntary smoking: a report of the
27 Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Office on Smoking and
28 Health; publication no. DHHS (CDC) 87-8398.
29 Thorn, S. R.; Ischiropoulos, H. (1997) Mechanism of oxidative stress from low levels of carbon monoxide.
30 Cambridge, MA: Health Effects Institute; research report no. 80.
31 Thorn, S. R.; Xu, Y. A.; Ischiropoulos, H. (1997) Vascular endothelial cells generate peroxynitrite in response to
32 carbon monoxide exposure. Chem. Res. Toxicol. 10: 1023-1031.
33 Topping, D. L.; Fishlock, R. C.; Trimble, R. P.; Storer, G. B.; Snoswell, A. M. (1981) Carboxyhaemoglobin inhibits
34 the metabolism of ethanol by perfused rat liver. Biochem. Int. 3: 157-163.
35 Touloumi, G.; Pocock, S. J.; Katsouyanni, K.; Trichopoulos, D. (1994) Short-term effects of air pollution on daily
36 mortality in Athens: a time-series analysis. Int. J. Epidemiol. 23: 957-967'.
37 Touloumi, G.; Samoli, E.; Katsouyanni, K. (1996) Daily mortality and "winter type" air pollution in Athens,
38 Greece—a time series analysis within the APHEA project. In: St Leger, S., ed. The APHEA project. Short
3 9 term effects of air pollution on health: a European approach using epidemiological time series data.
40 J. Epidemiol. Commun. Health 50(suppl. 1): S47-S51.
41 Turino, G. M. (1981) Effect of carbon monoxide on the cardiorespiratory system. Carbon monoxide toxicity:
42 physiology and biochemistry. Circulation 63: 253A-259A.
43 U.S. Centers for Disease Control and Prevention. (1997) Mortality patterns—preliminary data, United States, 1996.
44 Morb. Mortal. Wkly. Rep. 46: 941-944.
45 U.S. Environmental Protection Agency. (1979) Air quality criteria for carbon monoxide. Research Triangle Park,
46 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
47 no. EPA-600/8-79-022. Available from: NTIS, Springfield, VA; PB81-244840.
48 U.S. Environmental Protection Agency. (1984) Revised evaluation of health effects associated with carbon
49 monoxide exposure: an addendum to the 1979 EPA air quality criteria document for carbon monoxide.
50 Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
51 Assessment Office; report no. EPA-600/8-83-033F. Available from: NTIS, Springfield, VA;
52 PB85-103471/HSU.
53 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
54 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
55 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
October 1999 6-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 Verhoeff, A. P.; Hoek, G.; Schwartz, J.; Van Wijnen, J. H. (1996) Air pollution and daily mortality in Amsterdam.
2 Epidemiology 7: 225-230.
3 Vogel, J. A.; Gleser, M. A. (1972) Effect of carbon monoxide on oxygen transport during exercise. J. Appl. Physiol.
4 32:234-239.
5 Wall, M. A.; Johnson, J.; Jacob, P.; Benowitz, N. L. (1988) Cotinine in the serum, saliva, and urine of nonsmokers,
6 passive smokers, and active smokers. Am. J. Public Health 78: 699-701.
7 Wang, X.; Ding, H.; Ryan, L.; Xu, X. (1997) Association between air pollution and low birth weight:
8 a community-based study. Environ. Health Perspect. 105: 514-520.
9 Weinberg, C. R.; Wilcox, A. J. (1998) Reproductive epidemiology. In: Rothman, K. J.; Greenland, S., eds. Modern
10 epidemiology. 2nd ed. Philadelphia, PA: Lippincott-Raven Publishers; pp. 585-608.
11 Weir, F. W.; Rockwell, T. H.; Mehta, M. M.; Attwood, D. A.; Johnson, D. F.; Herrin, G. D.; Anglen, D. M.;
12 Safford, R. R. (1973) An investigation of the effects of carbon monoxide on humans in the driving task: final
13 report. Columbus, OH: The Ohio State University Research Foundation; contract no. 68-02-0329 and
14 CRC-APRAC project CAPM-9-69. Available from: NTIS, Springfield, VA; PB-224646.
15 Weiser, P. C.; Morrill, C. G.; Dickey, D. W.; Kurt, T. L.; Cropp, G. J. A. (1978) Effects of low-level carbon
16 monoxide exposure on the adaptation of healthy young men to aerobic work at an altitude of 1,610 meters.
17 In: Folinsbee, L. J.; Wagner, J. A.; Borgia, J. F.; Drinkwater, B. L.; Gliner, J. A.; Bedi, J. F., eds.
18 Environmental stress: individual human adaptations. New York, NY: Academic Press, Inc.; pp. 101-110.
19 Weiss, S. T. (1996) Cardiovascular effects of environmental tobacco smoke. Circulation 94: 599.
20 Weissbecker, L.; Carpenter, R. D.; Luchsinger, P. C.; Osdene, T. S. (1969) In vitro alveolar macrophage viability:
21 effect of gases. Arch. Environ. Health 18: 756-759.
22 Wells, A. J. (1994) Passive smoking as a cause of heart disease. J. Am. Coll. Cardiol. 24: 546-554.
23 Wen, S. W.; Goldenberg, R. L.; Cutter, G. R.; Hoffman, H. J.; Cliver, S. P.; Davis, R. O.; DuBard, M. B. (1990)
24 Smoking, maternal age, fetal growth, and gestational age at delivery. Am. J. Obstet. Gynecol. 162: 53-58.
25 Werner, R. M.; Pearson, T. A. (1998) What's so passive about smoking? Secondhand smoke as a cause of
26 atherosclerotic disease [editorial]. JAMA J. Am. Med. Assoc. 279: 157-158.
27 Wietlisbach, V.; Pope, C. A., Ill; Ackermann-Liebrich, U. (1996) Air pollution and daily mortality in three Swiss
28 urban areas. Soz. Praeventivmed. 41: 107-115.
29 Wilcosky, T. C.; Simonsen, N. R. (1991) Solvent exposure and cardiovascular disease. Am. J. Ind. Med.
30 19:569-586.
31 Wirkner, K.; Damme, B.; Poelchen, W.; Pankow, D. (1997) Effect of long-term ethanol pretreatment on the
32 metabolism of dichloromethane to carbon monoxide in rats. Toxicol. Appl. Pharmcol. 143: 83-88.
33 Wouters, E. J. M.; de Jong, P. A.; Cornelissen, P. J. H.; Kurver, P. H. J.; van Oel, W. C.; van Woensel, C. L. M.
34 (1987) Smoking and low birth weight: absence of influence by carbon monoxide? Eur. J. Obstet. Gynecol.
35 Reprod. Biol. 25: 35-41.
36 Wright, G.; Randell, P.; Shephard, R. J. (1973) Carbon monoxide and driving skills. Arch. Environ. Health
37 27: 349-354.
38 Yang, L.; Zhang, W.; He, H.; Zhang, G. (1988) Experimental studies on combined effects of high temperature and
39 carbon monoxide. J. Tongji Med. Univ. 8: 60-65.
40 Yang, W.; Jennison, B. L.; Omaye, S. T. (1997) Air pollution and asthma emergency room visits in Reno, Nevada.
41 Inhalation Toxicol. 9: 15-29.
42 Yang, W.; Jennison, B. L.; Omaye, S. T. (1998) Cardiovascular disease hospitalization and ambient levels of carbon
43 monoxide. J. Toxicol. Environ. Health 55A: 185-196.
44 Young, J. S.; Upchurch, M. B.; Kaufman, M. J.; Fechter, L. D. (1987) Carbon monoxide exposure potentiates
45 high-frequency auditory threshold shifts induced by noise. Hear. Res. 26: 37-43.
46 Zaren, B.; Lindmark, G.; Gebre-Medhin, M. (1996) Maternal smoking and body composition of the newborn. Acta
47 Paediatr. 85: 213-219.
48 Zebro, T.; Wright, E. A.; Littleton, R. J.; Prentice, A. I. D. (1983) Bone changes in mice after prolonged continuous
49 exposure to a high concentration of carbon monoxide. Exp. Pathol. 24: 51-67.
50 Zhu, B.-Q.; Parmley, W. W. (1995) Hemodynamic and vascular effects of active and passive smoking. Am. Heart J.
51 130:1270-1275.
October 1999 6-80 DRAFT-DO NOT QUOTE OR CITE
-------
i 7. INTEGRATIVE SUMMARY AND CONCLUSIONS
2
3
4 7.1 INTRODUCTION
5 Carbon monoxide (CO) is a colorless, tasteless, odorless, and nonirritating gas that is a
6 product of incomplete combustion of carbon-containing fuels. It also is produced within living
7 organisms by the natural degradation of hemoproteins (e.g., hemoglobin, myoglobin,
8 cytochromes) or as a by-product of xenobiotic metabolism, especially the breakdown of inhaled
9 organic solvents containing halomethanes (e.g., methylene bromide, iodide, or chloride). With
10 external exposure to additional CO, subtle health effects can begin to occur, and exposure to very
11 high levels can result in death.
12 The health significance of CO in the air largely results from CO being absorbed readily
13 from the lungs into the bloodstream, there forming a slowly reversible complex with hemoglobin
14 (Hb), known as carboxyhemoglobin (COHb). The presence of significant levels of COHb in the
15 blood causes hypoxia (i.e., reduced availability of oxygen to body tissues). The blood COHb
16 level, therefore, represents a useful physiological marker to predict the potential health effects of
17 CO exposure. The amount of COHb formed is dependent on the CO concentration and duration
18 of exposure, exercise (which increases both the amount of air inhaled and exhaled per unit of
19 time), the pulmonary diffusing capacity for CO, ambient pressure, health status, and the specific
20 metabolism of the exposed individual. The formation of COHb is a reversible process, but,
21 because of the high affinity of CO for Hb, the elimination half-time is quite long, varying from
22 2 to 6.5 h depending on the initial COHb levels. This may lead to accumulation of COHb,
23 especially if exposure is to varying concentrations of CO over extended periods of time.
24 Fortunately, mechanisms exist in normal, healthy individuals to compensate for the reduction in
25 tissue oxygen caused by increasing levels of COHb. Cardiac output increases and blood vessels
26 dilate to carry more blood so that the tissue can extract adequate amounts of oxygen from the
27 blood. There are several medical disorders, however, that can make an individual more
28 susceptible to the potential adverse effects of low levels of CO, especially during exercise.
29 Occlusive vascular disease (e.g., coronary heart disease and cerebrovascular disease) limit blood
30 flow to the tissues, obstructive lung disease (e.g., bronchitis, emphysema, asthma) causes
October 1999 7-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 gas-exchange abnormalities that limit the amount of oxygen that diffuses into the blood, and
2 anemia reduces the oxygen-carrying capacity of the blood. Under any of these conditions,
3 exposure to CO could reduce further the amount of oxygen available to the affected body tissues.
4 A reduction in oxygen delivery due to elevated COHb levels, combined with impaired air or
5 blood flow to the diseased tissues, will further reduce organ system function and limit exercise
6 capacity.
7 The existing national ambient air quality standards (NAAQS) for CO of 9 ppm for 8 h and
8 35 ppm for 1 h (Federal Register, 1994) have been established to reduce the risk of adverse
9 health effects in the population groups most sensitive to the presence of CO in the ambient air.
10 The term "ambient air" is interpreted to mean outdoor air at ground level where people live and
11 breathe. A great majority of people, however, spend most of their time indoors. A realistic
12 assessment of the health effects from exposure to ambient CO, therefore, must be set in the
13 context of total exposure, a major component of which is indoor exposure.
14 This chapter provides a summary of the key factors discussed in Chapters 2 through 6 of
15 the present document that determine what risk ambient CO poses to public health. An effort also
16 is made to qualitatively delineate key factors that contribute to anticipated health risks from
17 ambient CO in special subpopulations that form a significant proportion of the population at
18 large. Risk factors such as age, gender, and pregnancy are discussed, as well as preexisting heart,
19 lung, vascular, and hematologic diseases. Subpopulations at risk because of exposure to ambient
20 CO alone, or combined with other environmental factors, are identified. This information will be
21 used by the U.S. Environmental Protection Agency's Office of Air Quality Planning and
22 Standards for development of their staff paper and associated assessments that will help to
23 determine the adequacy of the existing CO NAAQS.
24
25
26 7.2 ENVIRONMENTAL SOURCES
27 Carbon monoxide comes from both natural and anthropogenic processes. About half of the
28 atmospheric CO is released at the earths's surface from fossil fuel and biomass burning, and the
29 rest is produced as the result of photochemical reactions in the atmosphere. About two-thirds of
30 the CO in the atmosphere arises from human activities; natural processes account for the
31 remaining one-third. The background concentration of CO in the troposphere influences the
October 1999 7-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 abundance of hydroxyl radicals (OH), thus affecting the global cycles of many natural and
2 anthropogenic trace gases, such as methane, that are removed from the atmosphere by reacting
3 with OH. During the 1980s, CO concentrations in remote marine areas increased at
4 approximately 1% per year. More recent reports, however, show that CO concentrations in these
5 locations declined rapidly between 1988 and 1993. Since 1993, the downward trend in CO has
6 slowed or leveled off, depending on the measurement laboratory, and it is not clear whether CO
7 will continue to decline or will increase.
8
9
10 7.3 ENVIRONMENTAL CONCENTRATIONS
11 The annual average CO concentration is about 0.13 ppm at monitoring sites located in the
12 marine boundary layer of the Pacific Ocean in the mid-latitudes of the Northern Hemisphere.
13 These sites are remote from local pollutant sources, and the values obtained at these sites are
14 thought to represent global background values for CO. Because of seasonal variations in the
15 emissions and chemical loss of CO through reaction with OH radicals, mean global background
16 CO levels vary between about 0.09 ppm in summer and about 0.16 ppm in winter. Annual 24-h
17 average CO concentrations obtained at U.S. monitoring sites in rural areas away from
18 metropolitan areas are typically about 0.20 ppm, compared with an annual 24-h average of
19 1.2 ppm across all monitoring sites in the Aerometric Information Retrieval System network in
20 1996.
21 In the United States, ambient air 8-h average CO concentrations monitored at fixed-site
22 stations in metropolitan areas are generally below 9 ppm and have decreased significantly since
23 1990 when the last CO criteria document was completed (U.S. Environmental Protection
24 Agency, 1991). In the latest year of record, 1997, annual mean CO concentrations were all less
25 than 9 ppm. However, in spite of the vehicle emission reductions responsible for the decrease in
26 ambient CO, high short-term peak CO concentrations still can occur in certain outdoor locations
27 and situations associated with motor vehicles and other combustion engine sources, for example,
28 riding behind high emitters (from both on- or nonroad vehicles) or in a vehicle with a defective
29 exhaust system and using lawnmowers, weeders, tillers, or other garden equipment. Also, air
30 quality data from fixed-site monitoring stations underestimate the short-term peak CO levels in
31 heavy traffic environments.
October 1999 7-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 Indoor and in-transit concentrations of CO can be significantly different from the typically
2 low ambient CO concentrations. The CO levels in homes without combustion sources are
3 usually lower than 5 ppm. The highest residential concentrations of CO that have been reported
4 are associated with vehicle startup and idling in attached garages and the use of unvented gas or
5 kerosene space heaters where peak concentrations of CO as high or higher than 50 ppm have
6 been reported. Carbon monoxide concentrations also have exceeded 9 ppm for 8 h in several
7 homes with gas stoves and, in one case, 35 ppm for 1 h; however, these higher CO
8 concentrations were in homes with older gas ranges that had pilot lights that burn continuously.
9 Newer or remodeled homes have gas ranges with electronic pilot lights. Also, the availability of
10 other cooking appliances (e.g., microwaves, heating plates) has decreased the use of gas ranges in
11 meal preparation.
12 Average CO concentrations as high as 10 to 12 ppm have been reported in human exposure
13 studies for in-vehicle compartments of moving automobiles. Carbon monoxide concentrations
14 will depend, however, on the season and traffic pattern, and the findings of more recent studies
15 suggest that pre-1990 study results are no longer applicable. For example, commuter exposure to
16 motor vehicle exhaust fell from a high of 37 ppm CO for a Los Angeles, CA, study in 1965 to a
17 low of 3 ppm CO for a New Jersey Turnpike study in 1992. For San Francisco, CA, using the
18 same data collection protocol, typical commuter exposures fell about 50% in the 11-year period
19 from 1980 to 1991, despite a 19% increase in average daily traffic. Carbon monoxide levels in
20 other indoor environments affected by engine exhaust (e.g., parking garages, tunnels) follow
21 similar trends but tend to be higher than in other indoor environments.
22 Because indoor and outdoor air quality differ substantially, and because people spend much
23 of their time indoors, ambient air quality measurements alone do not provide accurate estimates
24 of personal or population exposure to CO from ambient and nonambient sources. Whereas the
25 ambient monitoring data reflect exposure to ambient sources of CO only, the measurement of
26 CO from personal monitors reflects more accurately the actual total human population exposure
27 to CO.
28
29
30
31
October 1999 7-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.4 CARBOXYHEMOGLOBIN LEVELS IN THE POPULATION
2 Carbon monoxide diffuses rapidly across the alveolar and capillary membranes and more
3 slowly across the placental membrane. At equilibrium, approximately 95% of the absorbed CO
4 binds with hemoglobin to form COHb that, when elevated above the endogenous level, is a
5 specific biomarker of CO exposure. The remaining 5% is distributed extravascularly. During
6 continuous exposure to a fixed ambient concentration of CO, the COHb concentration increases
7 rapidly at the onset of exposure, starts to level off after 3 h, and approaches a steady state after
8 6 to 8 h of exposure. Therefore, an 8-h COHb value should be closely representative of any
9 longer continuous exposures. In real-life situations, prediction of individual COHb levels is
10 difficult because of large spatial and temporal variations in both indoor and outdoor levels of CO
11 and temporal variations of alveolar ventilation rates. Because COHb measurements are not
12 readily available in the exposed population, mathematical models have been developed to predict
13 COHb levels from known CO exposures under a variety of circumstances (see Figure 7-1).
14 Evaluation of human CO exposure situations indicates that occupational exposures in some
15 workplaces, or exposures in homes with faulty or unvented combustion sources, can exceed
16 100 ppm CO, leading to COHb levels of 4 to 5% with 1-h exposure and 10% or more with
17 continued exposure for 8 h or longer (see Table 7-1). Such high exposure levels are encountered
18 rarely by the general public under ambient conditions. More frequently, short-term exposures to
19 less than 25 to 50 ppm CO occur in the general population, and, at the low exercise levels usually
20 engaged in under such circumstances, resulting COHb levels typically remain below 2 to 3%
21 among nonsmokers. Those levels can be compared to the physiological baseline for nonsmokers,
22 which is estimated to be in the range of 0.3 to 0.7% COHb. Unfortunately, no new data have
23 become available on the distribution of COHb levels in the U.S. population since large-scale
24 nationwide surveys (e.g., National Health and Nutrition Examination Survey II [Radford and
25 Drizd, 1982]) and human exposure field studies (e.g., Denver, CO, and Washington, DC [Akland
26 et al., 1985]) were conducted in the late 1970s and early 1980s.
27 The major source of total exposure to CO for smokers comes from active tobacco smoking.
28 Baseline COHb concentrations in smokers average 4%, with a usual range of 3 to 8% for
29 one- to two-pack-per-day smokers, reflecting absorption of CO from inhaled smoke.
30 Carboxyhemoglobin levels as high as 15% have been reported for chain smokers. Exposure to
31 tobacco smoke not only increases COHb concentrations in smokers, but, under some
October 1999 7-5 DRAFT-DO NOT QUOTE OR CITE
-------
0)
Q.
A 8 h, 20 Umin
O 8h, 10 Umin
• 1 h, 20 L/min
• 1 h, 10 Umin
CO (ppm)
100
Figure 7-1. Predicted COHb levels resulting from 1- and 8-h exposures to CO at rest
(minute ventilation rate of 10 L/min) and with light exercise (20 L/min) are
based on the Coburn-Forster-Kane equation, using the following assumed
parameters for nonsmoking adults: altitude = 0 ft, initial COHb level = 0.5%,
Haldane coefficient = 218, blood volume = 5.5 L, Hb level = 15 g/100 mL, lung
diffusivity = 30 mL/torr/min, and endogenous rate of CO production =
0.007 mL/min.
1 circumstances, it also can affect nonsmokers. In some of the studies cited in this document,
2 neither the smoking habits of the subjects, nor their passive exposure to tobacco smoke, have
3 been taken into account. In addition, as the result of their higher baseline COHb levels, smokers
4 actually may be exhaling more CO into the air than they are inhaling from the ambient
5 environment when they are not smoking. Smokers may even show an adaptive response to the
6 elevated COHb levels, as evidenced by increased red blood cell volumes or reduced plasma
7 volumes. As a consequence, it is not clear if incremental increases in COHb caused by typical
8 ambient exposures actually would raise the chronically elevated COHb levels resulting from
9 smoking.
October 1999
7-6
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 7-1. PREDICTED CARBON MONOXIDE EXPOSURES
IN THE POPULATION
Predicted COHb Response'
,a,b
Exposure Conditions0 1 h, Light Exercise 8 h, Light Exercise
Nonsmoking adults 2 to 3% 4 to 7%
exposed to 25 to 50 ppm
Workplace or home 4 to 5% 12 to 13%
with faulty combustion
appliances at ==100 ppm
a See Figure 7-1 for assumed parameters of the Coburn-Forster-Kane equation (Coburn et al., 1965).
b Light exercise at 20 L/min.
0 Exposures are steady state.
1 7.5 MECHANISMS OF CARBON MONOXIDE ACTIVITY
2 A clear mechanism of action underlying the effects of low-level CO exposure is the
3 decreased oxygen-carrying capacity of blood and subsequent interference with oxygen release at
4 the tissue level that is caused by the binding of CO with Hb, producing COHb. The resulting
5 impaired delivery of oxygen can interfere with cellular respiration and cause tissue hypoxia. The
6 critical tissues (e.g., brain, heart) of healthy subjects have intrinsic physiologic mechanisms (e.g.,
7 increased blood flow and oxygen extraction) to compensate for CO-induced hypoxia.
8 In compromised subjects, or as CO levels increase, these compensatory mechanisms may be
9 overwhelmed, and tissue hypoxia, combined with impaired tissue perfusion and systemic
10 hypotension induced by hypoxia, may cause identifiable health effects.
11 Carbon monoxide will bind to intracellular hemoproteins such as myoglobin (Mb),
12 cytochrome oxidase, mixed-function oxidases (e.g., cytochrome P-450), tryptophan oxygenase,
13 and dopamine hydroxylase. Hemoprotein binding to CO would be favored under conditions of
14 low intracellular partial pressure of oxygen (PO2), particularly in brain and myocardial tissue
15 when intracellular PO2 decreases with increasing COHb levels. The hemoprotein most likely to
16 be inhibited functionally at relevant levels of COHb is Mb, found predominantly in heart and
17 skeletal muscle. The physiological significance of CO uptake by Mb is uncertain, but sufficient
18 concentrations of carboxymyoglobin potentially could limit maximal oxygen uptake of
October 1999 7-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 exercising muscle. There is also some evidence that binding of CO to intracellular hemoproteins
2 may secondarily precipitate oxidative stress. The health risks associated with this mechanism
3 have not been clearly established.
4
5
6 7.6 HEALTH EFFECTS OF CARBON MONOXIDE
7 This document deals primarily with the relatively low concentrations of CO that may
8 induce effects in humans at or near the lower margin of detection by current technology. Yet, the
9 health effects associated with exposure to this pollutant range from the more subtle
10 cardiovascular and neurobehavioral effects at low-ambient concentrations, as identified in the
11 preceding chapter, to unconsciousness and death following acute exposure to high
12 concentrations. The morbidity and mortality resulting from the latter exposures are described in
13 several recent reports (Jain, 1990; Penney, 1996; Ernst and Zibrak, 1998).
14 The health effects from exposure to low CO concentrations, such as the levels found in
15 ambient air, are considerably less threatening than those occurring in frank poisoning from high
16 CO concentrations. Effects of exposure to excessive ambient air levels of CO are summarized
17 here in terms of COHb levels; however, the lowest-observed-effect level depends on the method
18 used for analysis of COHb. Gas chromatography (GC) is the method of choice for measuring
19 COHb at saturation levels <5%, because of the large variability and potentially high bias of
20 optical methods such as CO-Oximetry (CO-Ox). Health effects are possible in sensitive
21 nonsmoking individuals exposed to ambient CO if peak concentrations are high enough, or of
22 sufficient duration, to raise the COHb saturation to critical levels above their physiological
23 baseline of 0.3 to 0.7% (GC). At 2.3% COHb (GC) or higher, some (predominantly young and
24 healthy) individuals may experience decreases in maximal exercise duration. At 2.4% COHb
25 (GC) or higher, patients with coronary artery disease (CAD) experience reduced exercise time
26 before the onset of acute myocardial ischemia, which is detectable either by symptoms (angina)
27 or by electrocardiographic changes (ST segment depression). At 5% COHb (CO-Ox) or higher,
28 some healthy individuals may experience impaired psychomotor performance; however, there is
29 large variability in response across studies that tested the same concentrations of CO, and work
30 conducted since the last criteria document review (U.S. Environmental Protection Agency, 1991)
31 indicates that significant behavioral impairments in healthy individuals should not be expected
October 1999 7-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 until COHb levels exceed 20% (CO-Ox). At 6% COHb (CO-Ox) or higher, some people with
2 CAD and high levels of baseline ectopy (chronic arrhythmia) may experience an increase in the
3 number and complexity of exercise-related arrhythmias.
4 Epidemiologic studies have associated elevated ambient CO levels with increased
5 exacerbation of heart disease in the population, but overall findings are not conclusive, possibly
6 because personal exposures may not be represented adequately by the CO concentrations
7 measured by fixed-site monitors. For example, exposure to cigarette smoke or to combustion
8 exhaust gases from small engines and recreational vehicles typically raises COHb levels much
9 higher than levels resulting from mean ambient CO exposures, and, for most people, exposures
10 to indoor sources of CO will exceed controllable outdoor exposures.
11 Health effects are more likely to occur, therefore, in individuals who are physiologically
12 stressed, either by exercise or by medical conditions that can make them more susceptible to low
13 levels of CO. The specific subpopulations potentially at risk from exposure to ambient CO are
14 discussed next.
15
16
17 7.7 SUBPOPULATIONS POTENTIALLY AT RISK FROM EXPOSURE
18 TO AMBIENT CARBON MONOXIDE
19 Most of the known quantifiable concentration-response relationships regarding the human
20 health effects of CO come from two carefully defined population groups: (1) healthy,
21 predominantly male, young adults and (2) patients with diagnosed CAD. On the basis of the
22 effects described, patients with reproducible exercise-induced angina appear to be best
23 established as a sensitive group within the general population that is at increased risk of
24 experiencing the health effects (i.e., decreased exercise duration because of exacerbation of
25 cardiovascular symptoms) of concern at ambient or near-ambient CO-exposure concentrations
26 that result in COHb levels as low as 2.4% (GC). A smaller sensitive group of healthy individuals
27 experience decreased exercise duration at similar levels of CO exposure, but only during
28 short-term maximal exercise. Decrements in exercise duration in the healthy population,
29 therefore, primarily would be a concern for athletes, rather than for people performing everyday
30 activities.
October 1999 7-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 It can be hypothesized, however, from both clinical and theoretical work and from
2 experimental research in laboratory animals, that certain other groups in the population are at
3 potential risk to exposure from CO. Probable risk groups that have not been studied adequately,
4 but that could be expected to be susceptible to CO because of gender differences, aging, or
5 preexisting disease or because of the use of medications or alterations in their environment
6 include fetuses and young infants; pregnant women; the elderly, especially those with
7 compromised cardiovascular function; individuals with partially obstructed coronary arteries, but
8 not yet manifesting overt symptomatology of CAD; those with heart failure; people with
9 peripheral vascular or cerebrovascular disease; individuals with hematologic diseases (e.g.,
10 anemia) that affect oxygen-carrying capacity or transport in the blood; individuals with
11 genetically unusual forms of hemoglobin associated with reduced oxygen-carrying capacity;
12 those with chronic obstructive pulmonary disease; people using medicinal or recreational drugs
13 with central nervous system depressant properties; individuals exposed to other chemical
14 substances (e.g., methylene chloride) that increase endogenous formation of CO; and individuals
15 who have not adapted to high altitude and are exposed to a combination of high altitude and CO.
16 Little empirical evidence is available by which to specify health effects associated with ambient
17 or near-ambient CO exposures in these probable risk groups.
18
19 7.7.1 Age, Gender, and Pregnancy as Risk Factors
20 The fetus and newborn infant are theoretically susceptible to CO exposure for several
21 reasons. Fetal circulation is likely to have a higher COHb level than the maternal circulation
22 because of differences in uptake and elimination of CO from fetal Hb. Because the fetus
23 normally has a lower oxygen tension in the blood than does the mother, a drop in fetal oxygen
24 tension resulting from the presence of COHb could have potentially serious effects. The
25 newborn infant, with a comparatively high rate of oxygen consumption and lower
26 oxygen-transport capacity for Hb than those of most adults, also would be potentially susceptible
27 to the hypoxic effects of increased COHb. Data from laboratory animal studies on the
28 developmental toxicity of CO suggest that prolonged exposure to high levels (>60 ppm) of
29 CO during gestation may produce a reduction in birth weight, cardiomegaly, and delayed
30 behavioral development. Limited epidemiologic findings suggest some association of subchronic
31 ambient CO exposure with low birth weight, but are not conclusive. Additional studies are
October 1999 7-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 needed to determine if chronic exposure to CO, particularly at low, near-ambient levels, can
2 compromise the already marginal conditions existing in the fetus and newborn infant. The
3 effects of CO on maternal-fetal relationships are not well understood.
4 In addition to fetuses and newborn infants, pregnant women also represent a susceptible
5 group because pregnancy is associated with increased alveolar ventilation and an increased rate
6 of oxygen consumption that serves to increase the rate of CO uptake from inspired air. Perhaps a
7 more important factor is that pregnant women experience an expanded blood volume associated
8 with hemodilution and thus are anemic because of the disproportionate increase in plasma
9 volume compared with erythrocyte volume. This group may be at increased risk and, therefore,
10 should be studied to evaluate the effects of ambient CO exposure and elevated COHb levels.
11 Changes in metabolism with age may make the aging population particularly susceptible to
12 the effects of CO. Maximal oxygen uptake declines with age. The rate of decline varies widely
13 among individuals because of the many confounding factors such as heredity, changes in body
14 mass and composition, and level of fitness.
15
16 7.7.2 Preexisting Disease as a Risk Factor
17 7.7.2.1 Subjects with Heart Disease
18 As introduced in Chapter 6, heart disease is a smaller and more specific subcategory of
19 cardiovascular disease that comprises, besides heart disease, other disorders such as
20 cerebrovascular disease (including stroke), hypertension (high blood pressure), and diseases of
21 the blood vessels. At the present time, short-term variations in ambient CO have been more
22 strongly associated with heart disease exacerbation than exacerbation of other cardiovascular
23 diseases. Heart disease, in turn, has other diagnostic subcategories, such as ischemic heart
24 disease (including CAD, myocardial infarction, and angina), heart failure, and disturbances of
25 heart rhythm (arrhythmia). Patients with heart disease have a markedly reduced circulatory
26 capacity and, therefore, may be very sensitive to limitations in oxygen-carrying capacity. Thus,
27 exposure to CO will reduce their exercise capacity and could have more serious consequences.
28 Coronary heart disease (CHD) remains the major cause of death and disability in
29 industrialized societies. In the United States, CHD is the single largest killer of males and
30 females, causing a total of 481,000 deaths in 1995 (American Heart Association, 1997),
31 two-thirds of all deaths from heart disease (U.S. Centers for Disease Control and Prevention,
October 1999 7-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 1997) and about half of all deaths from cardiovascular disease (see Figure 7-2). Almost
2 14 million Americans have a history of this disease, with much greater prevalence in both males
3 and females at increasing ages (see Figure 7-3). Individuals with CHD have myocardial
4 ischemia, which occurs when the heart muscle receives insufficient oxygen delivered by the
5 blood. For some, exercise-induced angina pectoris (chest pain) can occur. In all patients with
6 diagnosed CAD, however, the predominant type of ischemia, as identified by ST segment
7 depression, is asymptomatic (i.e., silent). Thus, patients who experience angina typically have
8 additional ischemic episodes that are asymptomatic. Unfortunately, some individuals in the
9 population have CAD but are totally asymptomatic and, therefore, do not know they are
10 potentially at risk. It has been estimated that 5% of middle-aged men show signs of ischemia
11 during an exercise stress test; a significant number of these men will have angiographic evidence
12 of CAD. Persons with both asymptomatic and symptomatic CAD have a limited coronary flow
13 reserve and, therefore, will be sensitive to a decrease in oxygen-carrying capacity induced by CO
14 exposure.
15 Heart failure is a major and growing public health problem. Almost 5 million Americans
16 have heart failure, and about 400,000 new cases occur each year (American Heart Association,
17 1997). Because the prevalence of heart failure is known to increase with age, improvements in
18 the average life expectancy of the general population would be expected to increase the
19 magnitude of the problem over the next few decades. The etiology of heart failure is diverse, but
20 the most common secondary conditions observed in hospitalized patients are CHD, hypertension,
21 chronic obstructive pulmonary disease (COPD), diabetes, and cardiomyopathy (Croft et al.,
22 1997). The exacerbation of some of these secondary conditions by CO are not well known;
23 however, any heart failure patients with CAD, for example, might be even more sensitive to CO
24 exposure.
25
26 7.7.2.2 Subjects with Other Vascular Diseases
27 Vascular disease, including cerebrovascular disease, is present in both males and females
28 and is more prevalent above 65 years of age because of the increasing likelihood of adverse
29 effects from atherosclerosis or thickening of the artery walls. Atherosclerosis is a leading cause
30 of many deaths from heart attack and stroke (American Heart Association, 1997). In fact, when
31 considered separately from other cardiovascular diseases, stroke ranks as the third leading cause
October 1999 7-12 DRAFT-DO NOT QUOTE OR CITE
-------
50% Coronary Heart Disease
22% Other
1% Rheumatic Fever/
Rheumatic Heart Disease
1% Congenital Heart Defects
2% Atherosclerosis
4% Congestive Heart Failure
4% High Blood Pressure
16% Stroke
Figure 7-2. Percentage breakdown of deaths from cardiovascular diseases in the United
States (1996 mortality statistics).
Source: American Heart Association (1997); National Center for Health Statistics (1995).
1 of death behind heart disease and cancer (U.S. Centers for Disease Control and Prevention,
2 1997). Vascular diseases are associated with a limited blood flow capacity and, therefore,
3 patients with these diseases should be sensitive to CO exposure. It is not clear, however, how
4 low levels of exposure to CO will affect these individuals. For example, only one study,
5 reviewed in the previous criteria document (U.S. Environmental Protection Agency, 1991), has
6 been reported on patients with peripheral vascular disease. Ten men with diagnosed intermittent
7 claudication (lameness) experienced a significant decrease in time to onset of leg pain when
8 exercising on a bicycle ergometer after breathing 50 ppm CO for 2 h (2.8% COHb).Further
9 research is needed, therefore, to better determine the sensitivity of individuals with vascular
10 disease to CO.
October 1999
7-13
DRAFT-DO NOT QUOTE OR CITE
-------
c
Q.
O
80
60
79
40
I
(D
Q_
20
0
59
6 3 6 4
13
9
29
17
45
45
56
"
70
72
-
70
18-19 20-29 30-39 40-49 50-59 60-69 70-79 >80
Ages
Males
Females
Figure 7-3. Estimated prevalence of cardiovascular disease by age and sex for the United
States, 1988 to 1991.
Source: American Heart Association (1997); Collins (1997); Adams and Marano (1995).
1 7.7.2.3 Subjects with Anemia and Other Hematologic Disorders
2 Clinically diagnosed low values of Hb, characterized as anemia, are a relatively prevalent
3 condition throughout the world. If the anemia is mild to moderate, an inactive person is often
4 asymptomatic. However, because of the limitation in the oxygen-carrying capacity resulting
5 from the low Hb values, an anemic person should be more sensitive to low-level CO exposure
6 than would be a person with normal Hb levels. Anemia is more prevalent in pregnant women
7 and in the elderly, two already potentially high-risk groups. An anemic person also will be more
8 sensitive to the combination of CO exposure and high altitude.
9 Individuals with hemolytic anemia often have higher baseline levels of COHb because the
10 rate of endogenous CO production from heme catabolism is increased. One of the many causes
11 of anemia is the presence of abnormal Hb in the blood. For example, in sickle-cell disease, the
October 1999
7-14
DRAFT-DO NOT QUOTE OR CITE
-------
1 average lifespan of red blood cells with abnormal hemoglobin S is 12 days compared to an
2 average of 120 days in healthy individuals with normal Hb. As a result, baseline COHb levels
3 can be as high as 4%. In subjects with Hb Zurich, where affinity for CO is 65 times that of
4 normal Hb, COHb levels range from 4 to 7%. Presumably, exogenous exposure to CO, in
5 conjunction with higher endogenous CO levels, could result in critical levels of COHb.
6 However, it is not known how ambient or near-ambient levels of CO would affect individuals
7 with these disorders.
8
9 1.1.2A Subjects with Obstructive Lung Disease
10 Chronic obstructive pulmonary disease (COPD) is a prevalent disease especially among
11 smokers, and a large number (>50%) of these individuals have limitations in their exercise
12 performance demonstrated by a decrease in oxygen saturation during mild to moderate exercise.
13 As a consequence, individuals with hypoxia resulting from COPD such as bronchitis and
14 emphysema may be susceptible to CO during submaximal exercise typical of normal daily
15 activity. In spite of their symptoms, many of them (-30%) continue to smoke and already may
16 have COHb levels of 4 to 8%. The COPD patients with hypoxia are also more likely to have a
17 progression of the disease resulting in severe pulmonary insufficiency, pulmonary hypertension,
18 and right heart failure.
19 Hospital admissions for asthma have increased considerably in the past few years,
20 particularly among individuals less than 18 years of age. Because asthmatics also can experience
21 exercise-induced airflow limitation, it is likely that they also would experience hypoxia during
22 attacks and be susceptible to CO. It is not known, however, how exposure to CO actually would
23 affect these individuals. Observed epidemiologic associations of short-term ambient CO levels
24 with respiratory disease frequency cannot yet be interpreted with confidence.
25
26 7.7.3 Subpopulations at Risk from Combined Exposure to Carbon Monoxide
27 and Other Chemical Substances
28 7.7.3.1 Interactions with Drugs
29 There is an almost complete absence of data on the possible toxic consequences of
30 combined CO exposure and drug use. Because of the diverse classes of both cardiovascular and
31 psychoactive drugs, and the many other classes of drugs that have not been examined at all, it
October 1999 7-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 must be concluded that this is an area of concern that is difficult to address meaningfully at the
2 present time.
3
4 7.7.3.2 Interactions with Other Chemical Substances in the Environment
5 Besides direct exposure to ambient CO, there are other chemical substances in the
6 environment that can lead to increased COHb saturation when inhaled. Halogenated
7 hydrocarbons used as organic solvents undergo metabolic breakdown by cytochrome P-450 to
8 form CO and inorganic halide. Possibly the greatest concern regarding potential risk in the
9 population comes from exposure to one of these halogenated hydrocarbons, methylene chloride,
10 and some of its derivatives that could result in potentially harmful levels of COHb in individuals
11 at risk.
12
13 7.7.4 Subpopulations Exposed to Carbon Monoxide at High Altitudes
14 For patients with CAD, restricted coronary blood flow limits oxygen delivery to the
15 myocardium. Carbon monoxide also has the potential for compromising oxygen transport to the
16 heart. For this reason, such patients have been identified as the subpopulation most sensitive to
17 the effects of CO. A reduction in PO2 in the atmosphere, as at high altitude, also has the
18 potential for compromising oxygen transport. Therefore, patients with coronary artery disease
19 who visit higher elevations may be unusually sensitive to the added effects of atmospheric CO.
20 It is important to distinguish between the long-term resident of high altitude and the newly
21 arrived visitor from low altitude. Specifically, the visitor will be more hypoxemic than the fully
22 adapted resident. The combination of high altitude with CO will pose the greatest risk to persons
23 newly arrived at high altitude who have underlying cardiopulmonary disease, particularly because
24 they are usually older individuals.
25 It is known that low birth weights occur both in infants born at altitudes above 6,000 ft and
26 in infants born near sea level, whose mothers had elevated COHb levels because of cigarette
27 smoking. It also has been shown that COHb levels in smokers at high altitude are higher than
28 those in smokers at sea level. Although it is probable that the combination of hypoxic hypoxia
29 and hypoxia resulting from ambient exposure to CO could reduce birth weight further at high
30 altitude and possibly modify future development, no data are available to evaluate this
31 hypothesis.
October 1999 7-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 7.8 CONCLUSIONS
2 Ambient CO concentrations measured at central, fixed-site monitors in metropolitan areas
3 of the United States have decreased significantly since the late 1980s, when air quality was
4 reviewed in the previous criteria document (U.S. Environmental Protection Agency, 1991). The
5 decline in ambient CO follows approximately the decline in motor vehicle emissions. Exposure
6 to tobacco smoke, to CO indoors from unvented or partially vented combustion sources, and to
7 CO from uncontrolled outdoor sources (e.g, small combustion engines) may represent a
8 significant portion of an individual's total CO exposure. Unfortunately, there is not a good
9 estimate of CO exposure distribution for the current population.
10 Health assessment information provided in the present document does not warrant
11 changing the conclusions of the previous document. The principal cause of CO-induced effects
12 at low levels of exposure still is thought to be increased COHb formation and the consequent
13 reduction of oxygen delivery to the body's organs and tissues. The air quality criteria used to
14 support the existing CO NAAQS were primarily those data obtained from experimental studies
15 of nonsmoking coronary artery disease patients during exercise. These studies identified adverse
16 effects with CO exposures that lead to COHb levels of 2.4% (GC) or higher. Young, healthy
17 individuals appear to be at little or no health risk because of ambient CO exposure. In these
18 individuals, the only observed effect of CO exposures resulting in <5% COHb has been reduction
19 of maximal exercise. No effects of CO exposures in this range have been observed in healthy
20 individuals performing submaximal exercise at levels typical of normal human activities.
21 Greater concern, therefore, has focused on subpopulations in which biological and
22 pathophysiologic considerations would suggest increased susceptibility to low-level CO
23 exposure. Indeed, recent epidemiologic studies that have become available since publication of
24 the previous document raised the level of scientific concern regarding ambient CO as a potential
25 risk factor for heart disease exacerbation, mortality, and low birth weight. Results of these
26 studies make a strong scientific case for further research on the health effects of ambient CO
27 exposure. This research should address CO alone and as a component of the overall ambient air
28 pollution mix. Nevertheless, the epidemiologic studies remain subject to considerable biological
29 and statistical uncertainty, and the available epidemiologic database does not provide convincing
30 evidence that further selective reduction of ambient CO levels would substantially benefit public
31 health.
October 1999 7-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2
3 Adams, P. F.; Marano, M. A. (1995) Current estimates from the National Health Interview Survey, 1994.
4 Hyattsville, MD: U. S. Department of Health and Human Services, Public Health Service, National Center for
5 Health Statistics; publication no. 96-1521. (Vital and health statistics: v. 10, no. 193). Available:
6 www.cdc.gov/nchswww/products/pubs/pubd/series/ser.htm.
7 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
8 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
9 Technol. 19: 911-918.
10 American Heart Association. (1997) 1998 heart and stroke statistical update. Dallas, TX: American Heart
11 Association. Available: www.amhrt.org/Scientific/HSstats98/index.html.
12 Coburn, R. F.; Forster, R. E.; Kane, P. B. (1965) Considerations of the physiological variables that determine the
13 blood carboxyhemoglobin concentration in man. J. Clin. Invest. 44: 1899-1910.
14 Collins, J. G. (1997) Prevalence of selected chronic conditions: United States, 1990-1992. Hyattsville, MD: U.S.
15 Department of Health and Human Services, Public Health Service, National Center for Health Statistics;
16 publication no. 97-1522. (Vital and health statistics: v. 10, no. 194). Available:
17 www.cdc.gov/nchswww/products/pubs/pubd/series/ser.htm.
18 Croft, J. B.; Giles, W. H.; Pollard, R. A.; Casper, M. L.; Anda, R. F.; Livengood, J. R. (1997) National trends in the
19 initial hospitalization for heart failure. J. Am. Geriatr. Soc. 45: 270-275.
20 Ernst, A.; Zibrak, J. D. (1998) Carbon monoxide poisoning. N. Engl. J. Med. 339: 1603-1608.
21 Federal Register. (1994) National ambient air quality standards for carbon monoxide—final decision. F. R.
22 (August 1) 59: 38,906-38,917.
23 Jain, K. K., ed. (1990) Carbon monoxide poisoning. St. Louis, MO: Warren H. Green, Inc.
24 National Center for Health Statistics. (1995) Health, United States, 1994. Hyattsville, MD: U.S. Department of
25 Health and Human Services, Public Health Service; publication no. 95-1232. Available:
26 www.cdc.gov/nchswww/products/pubs/pubd/hus/2010/2010.htm.
27 Penney, D. G., ed. (1996) Carbon monoxide. Boca Raton, FL: CRC Press.
28 Radford, E. P.; Drizd, T. A. (1982) Blood carbon monoxide levels in persons 3-74 years of age: United States,
29 1976-80. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service, National
30 Center for Health Statistics; DHHS publication no. (PHS) 82-1250. (Advance data from vital and health
31 statistics: no. 76).
32 U.S. Centers for Disease Control and Prevention. (1997) Mortality patterns—preliminary data, United States, 1996.
33 Morb. Mortal. Wkly. Rep. 46: 941-944.
34 U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
3 5 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
36 no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
October 1999 7-18 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
AC
AER
AIRS
AQCD
BTPS
C
CAA
CAD
CaO2
CARB
CAS
CASAC
CFD
CFK
cGMP
CH3
CH4
CH3Br
CH3CC1
CH3CH(
CH3C1
CH3CO
CH2O
CH3O2
CH3OOI
APPENDIX A
Abbreviations and Acronyms
Alternating current
Air exchange rate
Aerometric Information Retrieval System
Air quality criteria document
Body temperature and pressure, saturated with water vapor at 37 °C
Carbon
Clean Air Act
Coronary artery disease
Arterial oxygen content
California Air Resources Board
Children's Activity Survey
Clean Air Scientific Advisory Committee
Cumulative frequency distribution
Coburn-Forster-Kane
Cyclic guanosine monophosphate
Methyl radical
Methane
Methyl bromide
Methyl chloroform
Acetaldehyde
Methyl chloride
Acetyl radical
Formaldehyde
Methyl peroxy radical
Methyl hydroperoxide
October 1999
A-l
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
CHD
CHF
CI
CMAQ
CMRO2
CMSA
CO
CO2
COH
COHb
COMb
CO-Ox
COPD
CPSC
CRM
CTM
CVD
DLCO
ECG
ED
EMFAC7
EPA
ETS
FAF
F^O
FDA
FID
GAM
Coronary heart disease
Congestive heart failure
Confidence interval
Congestion Management and Air Quality
Cerebral metabolic rate for oxygen
Consolidated metropolitan statistical area
Carbon monoxide
Carbon dioxide
Coefficient of haze
Carboxyhemoglobin
Carboxymyoglobin
CO-Oximetry or CO-Oximeter
Chronic obstructive pulmonary disease
Consumer Product Safety Commission
Certified Reference Material
Chemical Transport Model
Cardiovascular disease
Diffusing capacity for carbon monoxide
Electrocardiogram
Effective dose for a specific decrement in function
Emissions Factor 7
U.S. Environmental Protection Agency
Environmental tobacco smoke
Forced-air furnace
Fractional concentration of carbon monoxide in inhaled air
Food and Drug Administration
Flame ionization detection or detector
General additive model
October 1999
A-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
GC
GFC
GLM
hu
H
H2
Hb
HCN
HCO
HO
HO2
H202
HOCO
IR
LOEL
M
M
Mb
MDL
MI
MSA
n
N
N2
NAAQS
NAMS
NASA
Ni(CO)4
Gas chromatography or gas chromatograph
Gas filter correlation
General linear model
Photon
Atomic hydrogen
Molecular hydrogen
Hemoglobin
Hydrogen cyanide
Formyl radical
Heme oxygenase
Hydroperoxy radical
Hydrogen peroxide
Carboxyl radical
Infrared
Lowest-observed-effect level
Haldane coefficient
Mediator
Myoglobin
Minimum detection limit
Myocardial infarction
Metropolitan Statistical Area
Number
North
Molecular nitrogen
National Ambient Air Quality Standards
National Air Monitoring Station
National Aeronautics and Space Administration
Nickel tetracarbonyl
October 1999
A-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
NDIR
NEM
NHANES
NIST
NMDA
NMHC
NMi
NMOC
NO
•NO
NO2
NOX
NOAA/CMDL
NTRM
O
02
03
OGI
OH
O2Hb
P
"atm
PB
PAH
PAN
PCO
PEM
Nondispersive infrared
National Ambient Air Quality Standards Exposure Model
National Health and Nutrition Examination Survey
National Institute of Standards and Technology
N-methyl-D-aspartate
Non-methane hydrocarbon
Nederland Meetinstitut (i.e., Dutch Bureau of Standards)
Non-methane organic compounds
Nitric oxide
Nitric oxide free radical
Nitrogen dioxide
Nitrogen oxides
National Oceanic and Atmospheric Administration Climate Monitoring
Diagnostics Laboratory
National Institute of Standards and Technology Traceable
Reference Material
Atomic oxygen
Molecular oxygen
Ozone
Oregon Graduate Institute
Hydroxyl radical
Oxyhemoglobin
Probability
Pressure in atmospheres
Barometric pressure
Polyaromatic hydrocarbon
Peroxyacetly nitrate
Partial pressure of carbon monoxide
Personal exposure monitor
October 1999
A-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
PM
PM2,
PM10
pNEM
PA
P02
PRM
Q
r
R2
RBC
RER
RR
S
SD
SHAPE
SIDS
SLAMS
SO2
SRM
ST
STPD
TCM
TDL
TDLS
Tg
TSP
TWA
Particulate matter
Particulate matter with an aerodynamic diameter <2.5 jim
Particulate matter with an aerodynamic diameter < 10 jim
Probabilistic National Ambient Air Quality Standards Exposure Model
Partial pressure of oxygen in humidified inspired air
Partial pressure of oxygen
Primary Reference Material
Perfusion
Linear regression correlation coefficient
Multiple correlation coefficient
Red blood cell
Respiratory exchange ratio
Relative risk
South
Standard deviation
Simulation of Human Activity and Pollutant Exposure
Sudden infant death syndrome
State and Local Air Monitoring Station
Sulfur dioxide
Standard Reference Materials
Segment of the electrocardiogram
Standard temperature and pressure, dry
Transportation Control Measure
Tunable diode laser
Tunable diode laser spectroscopy
Teragram
Total suspended particulate
Time-weighted average
October 1999
A-5
DRAFT-DO NOT QUOTE OR CITE
-------
1 UHC Unburned hydrocarbon
2 UV Ultraviolet
3 UVGSH Unvented gas space heater
4 VA Alveolar ventilation
5 VD Volume of physiological dead space
6 VMT Vehicle miles of travel
7 W/F Wall or floor furnace
October 1999 A-6 DRAFT-DO NOT QUOTE OR CITE
------- |