United States
                  Environmental Protection
                  Agency
                   Office of Research
                   and Development
                   Washington, DC 20460
EPA 600/P-99/001
February 1999
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-99/001
                                            February 1999
                                       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.
February 15, 1999                          ii         DRAFT-DO NOT QUOTE OR CITE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 1998.
     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.
February 15, 1999                           iii         DRAFT-DO NOT QUOTE OR CITE

-------
                               Table of Contents

                                                                           Page

List of Tables	xi
List of Figures 	xiii
Authors, Contributors, and Reviewers	 xvii
U.S. Environmental Protection Agency Project Team for Development of Air Quality
    Criteria for Carbon Monoxide	 xxv

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	1-9
        1.4.4   Measurement Technology	1-9
        1.4.5   Ambient Air Quality	1-9
        1.4.6   Indoor Emissions and Concentrations	1-10
        1.4.7   Exposure Assessment	1-10
        1.4.8   Mechanisms of Action 	1-11
        1.4.9   Health Effects	1-11
        1.4.10  Carbon Monoxide Interaction with Drugs  	1-13
        1.4.11  Subpopulations atRisk	1-13
   1.5  METHODS AND PROCEDURES FOR DOCUMENT PREPARATION	1-13
   1.6  ORGANIZATION AND CONTENT OF THE DOCUMENT	1-15
REFERENCES  	1-17

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
              2.3.1.2  National Institute of Standards and Technology Traceable
                     Reference Materials	2-6

February 15, 1999                        v         DRAFT-DO NOT QUOTE OR CITE

-------
                                Table of Contents
                                     (cont'd)
                                                                               age
              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
              2.4A.I  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 Standards  	2-18
        2.4.6  Other Methods of Analysis	2-18
   2.5  MEASUREMENT USING PERSONAL MONITORS  	2-19
   2.6  BIOLOGICAL MONITORING  	2-19
        2.6.1  Carboxyhemoglobin Measurements	2-20
        2.6.2  Breath Carbon Monoxide Measurements	2-21
        2.6.3  Relationships of Breath Carbon Monoxide to Blood
              Carboxyhemoglobin  	2-22
        2.6.4  Summary of the Relationship Between Biological Measurements
              of Carbon Monoxide	2-23
   2.7  SUMMARY 	2-24
REFERENCES  	2-26

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	3-6
        3.2.3  The Atmospheric Chemistry of Carbon Monoxide	3-9
   3.3  NATIONWIDE CARBON MONOXIDE EMISSIONS  ESTIMATES 	3-15
   3.4  CARBON MONOXIDE CONCENTRATIONS IN AMBIENT AIR 	3-19
        3.4.1  Nationwide Trends in Ambient Carbon Monoxide Concentrations	3-20
        3.4.2  Orcadian Patterns in Carbon Monoxide Concentrations	3-22

February 15, 1999                         vi        DRAFT-DO NOT QUOTE OR CITE

-------
                              Table of Contents
                                    (cont'd)
                                                                            age
   3.5   SOURCES, EMISSIONS, AND CONCENTRATIONS OF CARBON
        MONOXIDE IN INDOOR ENVIRONMENTS	3-28
        3.5.1   Unvented Combustion Sources and Estimated Emissions Rates	3-28
              3.5.1.1  Gas Cooking Ranges and Ovens and Furnaces 	3-29
              3.5.1.2  Emissions from Unvented Space Heaters  	3-33
              3.5.1.3  Woodstoves and Fireplaces	3-34
              3.5.1.4  Environmental Tobacco Smoke	3-34
        3.5.2   Indoor Concentrations of Carbon Monoxide 	3-35
              3.5.2.1  Factors Affecting Carbon Monoxide Concentrations 	3-35
              3.5.2.2  Models for Carbon Monoxide Concentrations 	3-40
              3.5.2.3  Microenvironmental Monitoring Studies	3-43
   3.6   SUMMARY 	3-52
REFERENCES  	3-56

APPENDIX 3 A: THE SPATIAL AND TEMPORAL VARIABILITY OF CARBON
              MONOXIDE IN SELECTED URBAN AREAS OF THE UNITED
              STATES 	 3A-1
   3A.1  INTRODUCTION	 3A-1
   3 A.2  GENERAL METHODOLOGY FOR DATA COLLECTION AND
         ANALYSES 	 3A-2
   3A.3  RESULTS AND DISCUSSION	 3A-3
         3A.3.1 Denver	 3A-3
         3A.3.2 Los Angeles	 3A-7
         3A.3.3 New York City 	 3A-11
         3A.3.4 Phoenix	 3A-15
   3A.4  SUMMARY 	 3A-16
REFERENCES  	 3A-20

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 	4-3
        4.2.2   Estimates of Population Exposure Based on Fixed-Site Monitors	4-4
        4.2.3   Surveys of Population Exposure Using Personal Monitors 	4-5
        4.2.4   Population Exposure Models 	4-7
   4.3   SURVEY OF RECENT EXPOSURE STUDIES OF NONSMOKERS 	4-10
        4.3.1   Nonoccupational Exposures	4-10
              4.3.1.1  Exposure to Carbon Monoxide from Motor Vehicles	4-12
              4.3.1.2  Exposure to Carbon Monoxide in Recreational Vehicles 	4-17
              4.3.1.3  Residential Exposure to Carbon Monoxide	4-18

February 15, 1999                        vii        DRAFT-DO NOT QUOTE OR CITE

-------
                                 Table of Contents
                                       (cont'd)
                                                                                  age
               4.3.1.4  Exposure to Carbon Monoxide at Commercial Facilities 	4-21
               4.3.1.5  Studies of Breath Carbon Monoxide in Populations:
                       The Effects of Exposure to Carbon Monoxide  	4-23
               4.3.1.6  Nonoccupational Exposure to Methylene Chloride	4-26
        4.3.2   Occupational Exposures	4-28
               4.3.2.1  Exposures to Carbon Monoxide in the Workplace  	4-28
               4.3.2.2  Exposures to Methylene Chloride in the Workplace	4-31
        4.3.3   Activity Pattern Studies 	4-32
               4.3.3.1  Activity Patterns of California Residents	4-32
               4.3.3.2  Activity Patterns of Children in Six States  	4-33
               4.3.3.3  A Comparative Study Between California and the Nation 	4-35
               4.3.3.4  An English Study	4-36
               4.3.3.5  A Boston Study of Household Activities, Life Cycle, and
                       Role Allocation 	4-36
               4.3.3.6  The National Human Activity Pattern Survey   	4-37
   4.4  MAJOR FACTORS AFFECTING POPULATION EXPOSURE	4-38
        4.4.1   Federal Policies Affecting Transportation and Air Quality
               in Urban Areas  	4-38
        4.4.2   Federal and State Policies Affecting Temporal Trends in Exposure 	4-40
               4.4.2.1  Effects of Motor Vehicle Emission Standards on
                       Unintentional Death Rates	4-40
               4.4.2.2  Effects of Motor Vehicle Emission Standards on Passenger
                       Cabin Exposure 	4-42
        4.4.3   California's No-Smoking Policy	4-45
        4.4.4   Social Changes Affecting Human Activity Patterns	4-46
   4.5  CONCLUSIONS	4-48
REFERENCES  	4-51

5.  PHARMACOKINETICS AND MECHANISMS OF ACTION OF
   CARBON MONOXIDE 	5-1
   5.1  INTRODUCTION	5-1
   5.2  ABSORPTION, DISTRIBUTION, AND PULMONARY ELIMINATION 	5-1
        5.2.1   Pulmonary Uptake	5-1
               5.2.1.1  Mass Transfer of Carbon Monoxide 	5-2
               5.2.1.2  Effects of Dead Space and Ventilation/Perfusion Ratio 	5-2
               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	5-5
               5.2.2.3  Heart and Skeletal Muscle	5-7
               5.2.2.4  The Brain and Other Tissues	5-8

February 15, 1999                         viii        DRAFT-DO NOT QUOTE OR CITE

-------
                               Table of Contents
                                     (cont'd)
                                                                               age
        5.2.3  Pulmonary and Tissue Elimination	5-9
   5.3   TISSUE PRODUCTION AND METABOLISM OF CARBON MONOXIDE .... 5-10
   5.4   CONDITIONS AFFECTING CARBON MONOXIDE UPTAKE AND
        ELIMINATION  	5-13
        5.4.1  Environment and Activity	5-13
        5.4.2  Altitude	5-14
        5.4.3  Physical Characteristics  	5-16
        5.4.4  Health Status	5-17
   5.5   MODELING CARBOXYHEMOGLOBIN FORMATION	5-18
        5.5.1  The Coburn-Forster-Kane and Other Regression Models  	5-18
              5.5.1.1  Empirical Regression Models 	5-18
              5.5.1.2  Linear and Nonlinear Coburn-Forster-Kane Differential
                      Equations	5-20
              5.5.1.3  Confirmation Studies of Coburn-Forster-Kane Models	5-21
              5.5.1.4  Application  of Coburn-Forster-Kane Models 	5-24
   5.6   INTRACELLULAR EFFECTS OF CARBON MONOXIDE	5-26
        5.6.1  Inhibition of Hemoprotein Function	5-26
        5.6.2  Free Radical Production	5-27
        5.6.3  Stimulation of Guanylate Cyclase	5-29
   5.7   MECHANISMS OF CARBON MONOXIDE TOXICITY 	5-30
        5.7.1  Alterations in Blood Flow	5-30
        5.7.2  Mitochondrial Dysfunction and Altered Production of High-Energy
              Intermediates	5-31
        5.7.3  Vascular Insults Associated with Exposure to Carbon Monoxide	5-32
   5.8   OTHER EFFECTS OF CARBON MONOXIDE	5-34
   5.9   SUMMARY 	5-35
REFERENCES  	5-37

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  Carbon Monoxide and Hospital Admissions	6-5
              6.2.1.3  Carbon Monoxide and Daily Mortality	6-10
        6.2.2  Controlled Laboratory Studies  	6-14
   6.3   CENTRAL NERVOUS SYSTEM AND BEHAVIORAL EFFECTS	6-19
        6.3.1  Brain Oxygen Metabolism  	6-19
              6.3.1.1  Whole Brain	6-19
              6.3.1.2  Subregions of the Brain	6-19
        6.3.2  Behavioral Effects of Carbon Monoxide 	6-21

February 15, 1999                         ix        DRAFT-DO NOT QUOTE OR CITE

-------
                              Table of Contents
                                   (cont'd)
                                                                          age
   6.4  DEVELOPMENTAL TOXICITY  	6-25
   6.5  ACUTE PULMONARY EFFECTS	6-28
   6.6  OTHER SYSTEMIC EFFECTS OF CARBON MONOXIDE 	6-28
   6.7  PHYSIOLOGIC RESPONSES TO CARBON MONOXIDE EXPOSURE	6-29
   6.8  COMBINED EXPOSURE OF CARBON MONOXIDE WITH OTHER
       POLLUTANTS, DRUGS, AND ENVIRONMENTAL FACTORS 	6-30
       6.8.1  High-Altitude Effects  	6-30
       6.8.2  Interaction with Drugs  	6-31
       6.8.3  Interaction with Other Air Pollutants and Environmental Factors	6-32
       6.8.4  Environmental Tobacco Smoke  	6-33
   6.9  SUMMARY 	6-34
REFERENCES  	6-38

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	7-4
   7.5  MECHANISMS OF CARBON MONOXIDE TOXICITY  	7-7
   7.6  HEALTH EFFECTS OF CARBON MONOXIDE	7-8
   7.7  SUBPOPULATIONS POTENTIALLY AT RISK FROM EXPOSURE
       TO AMBIENT CARBON MONOXIDE  	7-10
       7.7.1  Age, Gender, and Pregnancy as Risk Factors	7-11
       7.7.2  Preexisting Disease as Risk Factors  	7-12
             7.7.2.1   Subjects with Coronary Heart Disease	7-12
             7.7.2.2   Subjects with Congestive Heart Failure	7-12
             7.7.2.3   Subjects with Other Vascular Diseases	7-14
             7.7.2.4   Subjects with Anemia and Other Hematologic Disorders	7-15
             7.7.2.5   Subjects with Obstructive Lung Disease	7-15
       7.7.3  Subpopulations at Risk from Combined Exposure to Carbon Monoxide
             and Other Chemical Substances	7-16
             7.7.3.1   Interactions with Drugs  	7-16
             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  	7-17
   7.8  CONCLUSIONS	7-17
REFERENCES  	7-19

APPENDIX A: Abbreviations and Acronyms  	  A-l
February 15, 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	3-7

3-3     Nationwide Carbon Monoxide Emissions Estimates, 1987 to 1996  	3-16

3-4     Sites Not Meeting the Carbon Monoxide National Ambient Air Quality
        Standards in 1996	3-22

3-5     Running-Average Exceedances of the 9-ppm 8-Hour Carbon Monoxide
        Standard, 1988 Versus 1996	3-26

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  	3-27

3-7     Sources of Carbon Monoxide in the Indoor Environment  	3-29

3-8     Ranges in Average Carbon Monoxide Emission Rates for Residential Sources  ... 3-31

3-9     Prevalence of Gas Cooking Ranges 	3-37

3-10    Residential Air Exchange Rates	3-39

3-11    Input Parameters for Carbon Monoxide	3-44

3-12    Carbon Monoxide Descriptive Statistics for All Homes  	3-47

3-13    Carbon Monoxide Concentrations in Smoking and Nonsmoking Areas
        in Real-Life Situations 	3-51

3A-1    Kendall Tau Spatial Correlations for the Daily Maximum 8-Hour Average
        Carbon Monoxide Data in the Denver Metropolitan Statistical Area	  3 A-6


February 15, 1999                          xi          DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables
                                       (cont'd)

Number                                                                          Page
3 A-2    Kendall Tau Spatial Correlations for the Daily Maximum 8-Hour Average
        Carbon Monoxide Data in the Los Angeles Consolidated Metropolitan
        Statistical Area 	  3A-9

3 A-3    Kendall Tau Spatial Correlations for the Daily Maximum 8-Hour Average
        Carbon Monoxide Data in the New York City Consolidated Metropolitan
        Statistical Area 	 3A-13

3 A-4    Kendall Tau Spatial Correlations for the Daily Maximum 8-Hour Average
        Carbon Monoxide Data in the Phoenix Metropolitan Statistical Area	 3A-17

4-1     Mean Breath Carbon Monoxide Levels and Sample Sizes Across Smoking
        Categories and Job Types	4-25

4-2     Studies of Occupational Exposures and Dosages	4-30

4-3     Time Spent in Different Microenvironments by Californians, 1987 to 1990	4-34

4-4     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	4-35

4-5     Motor Vehicle Carbon Monoxide Emission Standards, Typical In-Vehicle
        Carbon Monoxide Exposures, and Unintentional Carbon Monoxide-Related
        Death Rates in the United States	4-41

4-6     Typical Net Mean Carbon Monoxide Concentration Ranges by Travel Mode
        for Cities in  Three Countries	4-44

6-1     Estimated Lowest-Observed-Effect Levels for Exposure of Laboratory
        Animals to Carbon Monoxide	6-18

6-2     Key Health Effects of Exposure to Ambient Carbon Monoxide 	6-36

7-1     Predicted Carbon Monoxide Exposures in the Population	7-7
February 15, 1999                         xii        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	2-23

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      Locations of sites in the nationwide ambient carbon monoxide monitoring
         network, 1996	3-21

3-3      Nationwide composite average of the annual second highest 8-hour carbon
         monoxide concentrations, 1977 to 1996	3-23

3-4      Variability in the annual second highest 8-hour carbon monoxide concentrations
         across all sites in the United States, 1987 to 1996 	3-24

3-5      Composite average of the annual second highest 8-hour carbon monoxide
         concentrations for rural, suburban, and urban sites, 1987 to 1996	3-25

3-6      Diurnal variation of nationwide composite  hourly average carbon monoxide
         concentrations for winter, from 1987 to 1996	3-26

3-7      Annual trend in gas  stove burner fuel use	3-30

3-8      Percentage of U.S. households using unvented combustion heaters, by type of
         fuel, stratified by region 	3-38

3-9      Modeled indoor carbon monoxide concentration distributions in houses with
         only one indoor combustion pollutant source	3-43

3-10     Arithmetic mean carbon monoxide concentrations by presence or absence of
         combustion source	3-48

3A-1     Map of Denver showing locations of carbon monoxide monitoring sites 	  3A-21

3 A-2     Average diurnal variation in carbon monoxide at the Denver-Broadway site
         for weekdays during the winter season	  3 A-22
February 15, 1999                          xiii         DRAFT-DO NOT QUOTE OR CITE

-------
                                  List of Figures
                                       (cont'd)

Number                                                                         Page
3 A-3    Monthly average diurnal variation in carbon monoxide at the
        Denver-Broadway site for weekdays from May 1986 through May 1987  	  3A-22

3 A-4    Monthly average diurnal variation in carbon monoxide at the
        Denver-Broadway site for weekdays from May 1995 through May 1996  	  3A-23

3 A-5    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	  3A-23

3A-6    Map of Los Angeles showing locations of carbon monoxide monitoring sites  ..  3A-24

3 A-7    Average diurnal variation in carbon monoxide at the Los Angeles-Lynwood
        site for weekdays during the winter season	  3 A-25

3 A-8    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	  3A-25

3 A-9    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	  3A-26

3 A-10   Average diurnal variation in carbon monoxide at the Los Angeles-Hawthorne
        site for weekdays during the winter season	  3 A-26

3A-11   Average diurnal variation in carbon monoxide at the Los Angeles-El Toro
        site for weekdays during the winter season	  3 A-27

3A-12   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 	  3A-27

3A-13   Map of New York showing locations of carbon monoxide monitoring sites ....  3A-28

3A-14   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 	  3A-29

3 A-15   Average diurnal variation in carbon monoxide at the New York-Manhattan
        site for weekdays during the winter season	  3 A-29

February 15, 1999                         xiv        DRAFT-DO NOT QUOTE OR CITE

-------
                                  List of Figures
                                       (cont'd)

Number                                                                         Page
3 A-16   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  	  3A-30

3 A-17   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	  3A-30

3 A-18   Map of Phoenix showing locations of carbon monoxide monitoring sites	  3A-31

3 A-19   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  	  3A-32

3 A-20   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	  3A-32

3A-21   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	  3A-33

3 A-22   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	  3A-33

4-1     Conceptual overview of the probabilistic National Ambient Air Quality
        Standards Exposure Model	4-9

4-2     Observed versus simulated 8-hour daily maximum exposure for persons
        residing in homes with gas stoves in Denver 	4-11

4-3     Observed versus simulated 8-hour daily maximum exposure for persons
        residing in homes without gas stoves in Denver	4-12

4-4     Trends in ambient carbon monoxde concentrations and in-vehicle carbon
        monoxide exposures, 1965 to 1992  	4-14

4-5     Excess carbon monoxide concentrations in the exhaled air of nonsmoking
        control subjects, untreated asthmatics, and treated asthmatics	4-27
February 15, 1999                         xv         DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                       (cont'd)

Number                                                                          Page
5-1     Diagrammatic presentation of carbon monoxide uptake and elimination
        pathways and carbon monoxide body stores	5-3

5-2     Oxyhemoglobin dissociation curve of normal human blood, of blood containing
        50% carboxyhemoglobin, and of blood with only 50% amount of hemoglobin
        because of anemia	5-7

5-3     Plot of fractional sensitivities of selected variables versus time of exposure	5-22

6-1     The effect of carbon monxide exposure on time to onset of angina	6-16

6-2     The relationship between carboxyhemoglobin and the cerebral metabolic rate
        for oxygen	6-20

6-3     The relationship between carboxyhemoglobin and behavior	6-24

7-1     Relationship between carbon monoxide exposure and carboxyhemoglobin
        levels in the blood	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
February 15, 1999                         xvi        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

Dr. David J. McKee—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Harvey M. Richmond—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
February 15, 1999                        xvii        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
  CHAPTER 2. ANAL YTICAL 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

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

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

Dr. William A. McClenny—National Exposure Research Laboratory (MD-44)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
February 15, 1999                        xviii        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
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

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

Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935

Dr. Peter G. Flachsbart—Department of Urban and Regional Planning
University of Hawaii at Manoa, Honolulu, HI 96822
February 15, 1999                         xix         DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Dr. Lawrence J. Folinsbee—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

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

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

Dr. Irwin H. Billick—WEC Consulting, Ltd., Potomac, MD 20854

Dr. Steven D. Colome—Integrated Environmental Services, Irvine, CA 92612-2935

February 15, 1999                         xx        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
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

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

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
            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

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

February 15, 1999                         xxi        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
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
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

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

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. Michael T. Kleinman—Department of Community and Environmental Medicine
California College of Medicine, University of California, Irvine, CA 92697

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
February 15, 1999                         xxii        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Dr. Robert Morris—Department of Family Medicine
Tufts University School of Medicine, Boston, MA  02111
            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

Reviewers

Dr. Michael G. Apte—Indoor Environment Department
Lawrence Berkeley National Laboratory, Berkeley, CA  94720

Dr. Lawrence J. Folinsbee—National Center for Environmental Assessment (MD-52)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. James J. McGrath—Department of Physiology
School of Medicine, Texas Tech University Health Sciences Center, Lubbock, TX  79430
February 15, 1999                        xxiii        DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
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
February 15, 1999                        xxiv       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
February 15, 1999                         xxv        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
February 15, 1999                        xxvi        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 inadequate for compliance monitoring or precise measurements in ambient air but are
44       useful for personal exposure studies. Further work to improve stability and specificity is
45       necessary.

       February 15, 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 than COHb
 7       for measurement error.
 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 130 and 50 ppb in remote areas of the
35       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 over
37       the past 10 years.
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.

        February 15, 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       11 ppm in 1977 to 4 ppm in 1996. Between 1987 and 1996, this statistic decreased by 37% at
23       190 urban sites, 37% at 142 suburban sites, and 48% at 10 rural sites. During this same period,
24       the nationwide average annual 24-h  CO levels decreased by about 40%, from 2.0 ppm to about
25       1.2 ppm.
26
27      • On- and off-road mobile sources account for approximately 80% of the 1996 nationwide
28       emissions inventory for CO. Declines in ambient CO levels in the United States follow
29       approximately the decline in motor vehicle emissions of CO.
30
31      • There were 51 exceedances of the 8-h NAAQS for CO at nine U.S. monitoring sites in 1996.
32       These sites, in descending order, were located in Lynwood, CA; Calexico and Hawthorne, CA;
33       Anchorage, AK (two sites); Las Vegas, NV; Phoenix, AZ; Kalispell, MT; and El Paso, TX.
34
35      • Carbon monoxide levels in the four  geographically diverse metropolitan statistical areas
36       (MSAs) of Denver, CO; Los Angeles, CA; New York, NY; and Phoenix, AZ, have decreased
37       from  1986 through 1995. However, the nature of the diurnal and seasonal variations has
38       remained essentially the same. These variations result largely from the interaction among
39       motor vehicle emissions, traffic patterns, and meteorological  parameters, such as wind speed
40       and mixing height.
41
42      • In general, the spatial distribution of CO within these four air sheds was highly heterogeneous.
43       For instance, correlations between time series of daily 8-h  average maximum CO
44       concentrations at different monitoring sites in the Denver, New York, and Los Angeles MSAs
45       ranged from about zero to 0.7, whereas corresponding correlations ranged from about 0.3 to
46       0.7 in the Phoenix MSA.

        February 15, 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 unvented and partially vented
17        combustion appliances.  Factors affecting emissions of CO from the use of combustion
18        appliances in the home include the type of source (e.g., gas cooking stoves, unvented space
19        heaters, wood stoves, fireplaces), appliance design, manufacturer, type of fuel used, fuel
20        consumption rate, and source operating condition.  Carbon monoxide concentrations in the
21        indoor environment will vary based on the source emission rate, use pattern, ambient CO
22        concentration, air exchange rate, building volume, and air mixing within the indoor
23        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 space heaters as
36        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
        February 15, 1999                          E-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1       based on differences in ventilation, the number of cigarettes or cigars smoked, and the smoking
 2       rate.
 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 impact on human populations of lower ambient CO
 9     exposures and the remaining risks to CO exposure above the level of the NAAQS in areas of
10     high traffic density and in other, indoor locations where combustion devices (e.g., stoves,
11     heaters) are not properly vented.
12
13     • Fixed site monitors often are used in urban areas to measure the ambient concentrations to
14       which individuals in the surrounding areas may be exposed. These measurements tend to
15       overestimate 8-h exposure values for people living in areas of lower traffic and underestimate
16       the exposure of people living in areas of higher traffic.
17
18     • Fixed-site ambient CO monitoring may provide a reasonable estimate of the average CO
19       exposures for some people who are not exposed to tobacco smoke or other sources of CO in
20       their homes and occupations.
21
22     • Nonsmokers exposed to tobacco smoke, heavy traffic fumes, and indoor sources of CO will
23       have higher body burdens of CO (COHb) than would be predicted from ambient data alone.
24
25     • Emission reductions in CO mandated by the CAA amendments have led to significant
26       reductions in ambient CO concentrations and lower traffic-related exposures to CO from motor
27       vehicle exhaust, suggesting that estimates of current population exposure based on pre-1990
28       exposure studies may no longer apply. There currently is not a good estimate of CO  exposure
29       distribution for the population.
30
31     • Personal CO exposures that exceed the level of the NAAQS will still occur in some
32       nonsmokers exposed to sources of CO not controlled by the CAA (e.g., recreational vehicles,
33       poorly vented or malfunctioning indoor combustion sources) or exposed in their occupations or
34       hobbies to CO or to organic solvents that are metabolized to CO (e.g., methylene chloride).
35
36     • Modern CO exposure models adequately predict the average general population exposure but
37       still underpredict high CO exposures, indicating that further work is required to understand the
38       activities and emissions associated with these higher exposures.
39
40     Pharmacokinetics and Mechanisms of Action
41
42           The action of CO in the body and the factors influencing its uptake, distribution to vital
43     tissues, and elimination, provide the foundation for measuring or predicting effects on organ
44     function.  In Chapter 5, the basic principles of CO pharmacokinetics are reviewed, and the
45     possible mechanisms for pathophysiologic effects at the cellular level are discussed.
46

       February  15,  1999                         E-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      • The most prominent pathophysiological effect of CO is hypoxemia caused by the binding of
 2       CO to hemoglobin. The formation of COHb reduces the oxygen-carrying capacity of blood
 3       and impairs release of oxygen from red blood cells to tissues. The brain and heart are
 4       especially sensitive to CO-induced hypoxia and cytotoxicity because these tissues have the
 5       highest resting oxygen requirements.
 6
 7      • The amount of COHb formed is dependent on the CO concentration and duration of exposure,
 8       minute ventilation, lung diffusion capacity, and ambient pressure, as well as the health status
 9       and metabolism of the exposed individual. The formation of COHb is reversible, but, because
10       of a small blood-to-air CO pressure gradient and tight binding of CO to Hb, the elimination
11       half-time is quite long, varying from 2 to 6.5 h.
12
13      • The physical and physiological variables affecting the rate of COHb formation and elimination
14       have been integrated into empirical and mathematical models for estimating COHb levels from
15       different conditions of exposure.  The nonlinear Coburn-Forster-Kane equation is the most
16       widely used predictive model of COHb formation and is still considered the best all-around
17       model for COHb prediction.
18
19      • Intracellular binding of CO to hemoproteins, particularly myoglobin (Mb) found in heart and
20       skeletal muscle, would be favored under conditions of low intracellular oxygen tension as
21       COHb levels rise. The impact of ambient CO on intracellular CO uptake by Mb is not well
22       understood.
23
24      • New investigations have expanded on the physiological effects of CO in two areas. First, there
25       is a growing recognition that CO may play a role in normal neurotransmission and in
26       microvascular vasomotor control.  Second, there also is increased interest in the ability of CO
27       to cause free-radical-mediated changes in tissues. The impact of ambient CO on these
28       processes and the role they may have in pathophysiology is not well understood.
29
30      Health Effects
31
32           Concerns about the potential health effects of exposure to CO are addressed in Chapter 6 by
33      examining the published results of extensive controlled-exposure studies and more limited
34      population-exposure studies. Emphasis is placed on the current understanding of quantifiable
35      health effects that are likely to occur in humans at the low COHb levels (<5 %) that are predicted
36      to result from typical ambient CO exposures.
37
38      • Blood COHb levels are the best indicator of potential health risk; however, the lowest-
39       observed-effect levels depend on the method used for analysis. Gas chromatography (GC)
40       generally is regarded as more accurate than CO-Ox  for measuring low COHb saturations
41       (<5%).
42
43      • Maximal exercise duration and performance in healthy individuals has been shown to be
44       reduced at COHb levels of >2.3 and >4.3% (GC), respectively. Performance decrements are
45       small, however, and likely to affect only competing athletes. No  effects were observed during
46       submaximal exercise in healthy individuals at COHb levels as high as  15 to 20%.

        February 15, 1999                          E-6         DRAFT-DO NOT QUOTE OR CITE

-------
 1      • Decreased exercise tolerance has been observed consistently in patients with coronary artery
 2       disease (CAD) and reproducible exercise-induced angina (chest pain) at COHb levels of 3 to
 3       6% COHb (CO-Ox). The indicators of myocardial ischemia during exercise, such as
 4       electrocardiographic changes and associated chest pain, were statistically significant in one
 5       large multicenter study at >2.4% COHb (GC) and showed a dose-response relationship with
 6       increasing COHb.
 7
 8      • An increase in the  number and complexity of exercise-related arrhythmias (irregular heart
 9       beats) has been observed at >6% COHb (CO-Ox) in some people with CAD and a high level
10       of baseline ectopy  (chronic arrhythmia) that may present an increased risk of sudden death.
11
12      • In epidemiologic studies, daily fluctuations in ambient CO  concentration have been associated
13       consistently with fluctuations in hospital admissions for congestive heart failure.  This
14       association has been observed at ambient CO levels at or below the current CO standards. The
15       influence on this association of ambient CO exposure, relative to CO exposure from
16       nonambient sources, has not been determined.
17
18      • Epidemiologic studies also suggest associations of short-term ambient CO exposure with total
19       daily cardiovascular hospital admissions and with total nonaccidental daily mortality, the great
20       majority of which occurs in people at least 65 years of age. As above, the relative influences
21       on these associations of ambient and nonambient CO have  not been determined.
22
23      • Recent analyses indicate that significant behavioral impairments in healthy individuals should
24       not be expected until COHb levels exceed 20%; however, mild central nervous system effects
25       have been reported in the historical CO literature at COHb  levels between 5 and 20%.
26
27      • Acute CO poisoning can affect the growth and function of the developing fetus, infant, and
28       child; however, it is very unlikely that ambient levels of CO typically encountered by pregnant
29       women would cause increased fetal risk.
30
31      • Ambient levels of CO are not known to have any direct effects on lung tissue.
32
33      • Carbon monoxide  has the potential to interact with other stressors. These include (1) visitation
34       to high altitudes, especially for patients with CAD; (2) use  of psychoactive drugs  or alcohol;
35       (3) use  of specific medications, especially nitric oxide and calcium channel blockers;
36       (4) prolonged exposure to heat; and (5) exposure to other pollutants.
37
3 8      Subpopulations Potentially At Risk
39
40           On the basis of monitored  ambient CO concentrations and quantifiable CO concentration-
41      response relationships for health effects demonstrated in humans, the following conclusions are
42      made in Chapter 7 regarding subpopulations potentially at risk from exposure to ambient CO.
43
44      • Young, healthy nonsmokers are  not at immediate risk from ambient CO exposure because only
45       limitations at maximal exercise performance have been demonstrated at low COHb levels
46       (<5%) that are predicted to result from ambient exposures.  Effects have not been demonstrated

        February  15, 1999                         E-7        DRAFT-DO NOT QUOTE OR CITE

-------
1       on healthy individuals performing submaximal exercise that is more typical of daily human
2       activity.
3
4     • Patients with reproducible exercise-induced angina (chest pain) are a sensitive group within the
5       general population that is at increased risk of experiencing decreased exercise tolerance
6       because of exacerbation of cardiovascular symptoms at ambient or near-ambient CO-exposure
7       concentrations that result in COHb levels of 2.4% (GC) or higher.
      February 15, 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  1998.
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, can 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

       February 15,  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,
        February 15, 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

        February 15, 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, CAS AC 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  CAS AC  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, EP A's National Center for Environmental
27     Assessment has begun to review and once  again revise the criteria for CO.
28
29
30
       February 15, 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
        February 15, 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
        February 15, 1999                          1-6        DRAFT-DO NOT QUOTE OR CITE

-------
 1     The Administrator concluded that the lowest COHb level at which adverse effects have been
 2     demonstrated in persons with angina is around 2.9 to 3.0%, representing an increase of 1.5%
 3     COHb above baseline when using the CO-Ox to measure COHb. These data serve to establish
 4     the upper end of the range of COHb levels of concern.  Taking into account the above data
 5     uncertainties, the less significant health endpoints, and less quantifiable data on other potentially
 6     sensitive groups, the lower end of the range was established at 2.0% COHb.
 7
 8     1.3.2  Relationship Between Carbon Monoxide Exposure and
 9            Carboxyhemoglobin Levels
10          In order to set ambient CO standards based on an assessment of health effects at various
11     COHb levels, it is necessary to estimate the ambient CO concentrations that are likely to result in
12     COHb levels of concern. The best all-around model for predicting COHb levels is the Coburn,
13     Foster, Kane (CFK) differential equation (U.S. Environmental Protection Agency, 1991).
14     Baseline estimates of COHb levels expected to be reached by nonsmokers exposed to various
15     constant concentrations of CO can be determined by the CFK equation (U.S. Environmental
16     Protection Agency, 1992). There are, however, two major uncertainties involved in estimating
17     COHb levels resulting from exposure to CO concentrations.  First, the large distribution of
18     physiological parameters used in the CFK equation across the population of interest is sufficient
19     to produce noticeable deviations in the COHb levels.  Second, predictions based on exposure to
20     constant CO concentrations can under- or overestimate responses of individuals exposed to
21     widely fluctuating CO levels that typically occur in the ambient environment.
22
23     1.3.3  Estimating Population Exposure
24          The EPA review included an analysis of CO exposures expected to be experienced by
25     residents of Denver, CO, under air quality scenarios where the 8-h NAAQS is just attained.
26     Although the exposure analysis included passive smoking and gas stove CO emissions as indoor
27     sources of CO, it did not include other sources that may be of concern to high-risk groups (e.g.,
28     wood stoves, fireplaces, faulty furnaces). The analysis indicated that, at the 8-h standard, fewer
29     than 0.1% of the nonsmoking cardiovascular-disease population would experience a COHb  level
30     >2.1% (U.S. Environmental Protection Agency, 1992). A smaller population was estimated to
31     exceed higher COHb percentages.
       February 15, 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 (bolus) 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 102(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
        February 15, 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
17     1.4.3  Global Cycle
18          Global trends in tropospheric CO concentrations declined from about 1988 to 1993 after
19     several years of annual increases, as  determined by different networks of surface observations.
20     Carbon monoxide levels apparently have stabilized since 1993.  The reasons for the changes in
21     CO trends still need to be determined.
22
23     1.4.4  Measurement Technology
24          The discussion on measurement methods for CO in the previous document has been
25     reviewed, older methods have been removed, and newer methods for monitoring CO from
26     various environmental sources are presented.
27
28     1.4.5  Ambient Air Quality
29          Because of the everchanging nature of atmospheric concentrations, levels in various
30     environments (rural, urban, and suburban) have been reanalyzed for different regions of the

       February 15, 1999                          1-9         DRAFT-DO NOT QUOTE OR CITE

-------
 1      United States. The temporal variability of CO levels from daily to seasonal time scales also has
 2      been characterized. Relations between urban concentrations of CO and regional and global
 3      background levels also are examined, as well as background levels of CO for use in different
 4      applications.
 5
 6      1.4.6 Indoor Emissions and Concentrations
 7           Indoor concentrations of CO are a function of outdoor concentrations, indoor sources,
 8      infiltration, ventilation, and air mixing. In the absence of indoor sources, concentrations of CO
 9      in the indoor environment are similar to those in ambient air; however, personal CO exposure
10      studies have shown that CO concentrations in excess of 9 ppm can occur in certain indoor and
11      in-transit microenvironments associated with transportation sources that are not considered part
12      of the ambient air. Unvented, improperly installed, or poorly maintained combustion appliances,
13      downdrafts during unstable weather conditions, and depressurization from the operation of
14      exhaust systems and fireplaces also may contribute to potentially high CO concentrations
15      indoors.  Further research is still needed,  however, to determine the contribution of nonambient
16      sources to total human exposure to CO.
17
18      1.4.7 Exposure Assessment
19           Compliance with the NAAQS is determined by measurements taken at fixed-site, ambient
20      monitors, yet exposure monitoring in the field and modeling studies indicate that individual
21      personal exposure does not correlate directly with CO concentrations determined by the
22      fixed-site monitors alone.  This is because of the mobility of people and the spatial and temporal
23      variability of CO concentrations across a given area. The nature of differences between fixed-
24      site and personal monitoring results should be given greater attention, especially in regard to
25      interpreting the results of epidemiology studies.
26           Data from population field studies can be used to construct and test models of human
27      exposure that account for time and activity patterns known to affect exposure to CO. New
28      information from field monitoring studies needs to be incorporated into exposure models  to
29      better capture the observed personal exposure distributions, including the higher exposures found
30      in the tail of exposure distribution.

        February 15, 1999                         1-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1           A unique feature of CO exposure is that the dose an individual receives can be estimated by
 2      measuring COHb.  It has been shown, however, that the method chosen for measurement of
 3      COHb can be a source of considerable error, particularly at the low end of the CO dissociation
 4      curve, where COHb levels are <5%.  The sensitivity of COHb measurement techniques will,
 5      therefore, have an influence on the lowest-observed-effect level for CO.
 6
 7      1.4.8 Mechanisms of Action
 8           The principle cause of CO toxicity is tissue hypoxia caused by CO binding to hemoglobin.
 9      Secondary mechanisms related to intracellular uptake of CO have been the focus of recent
10      research.  Current knowledge summarized in this document suggests that the most likely protein
11      other than hemoglobin to be inhibited functionally at relevant levels of COHb is myoglobin,
12      found in heart and  skeletal muscle. The extent of effects caused by free oxygen radicals and lipid
13      peroxidation needs to be evaluated in relation to typical ambient CO exposures in the population.
14
15      1.4.9 Health Effects
16           There  are many published studies on acute experimental and accidental exposures to CO;
17      however,  there is not enough reliable information on chronic exposures to low concentrations
18      from either ambient population-exposure studies or from occupational studies. Further work is
19      needed, therefore, to determine potential long-term exposures in the population  and to develop
20      reliable dose-response relationships for at-risk groups. This information currently is missing
21      from the published literature.  Some  of the issues associated with acute CO exposures are
22      discussed below.
23
24      Cardiovascular Effects
25           Maximal exercise performance is reduced in young, healthy, nonsmoking  individuals at
26      COHb levels as low as  2.3%, but this effect is small and would be of concern mainly for
27      competing athletes. Clinical studies  on subjects with reproducible exercise-induced angina have
28      confirmed that adverse effects occur with postexposure COHb levels as low as 3%.  Thus,
29      aggravation of coronary artery disease continues to provide the best scientific basis in support for
30      the current (9-ppm, 8-h and 35-ppm, 1-h) NAAQS for CO. More recent epidemiology studies in

        February  15, 1999                          1-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the United States, Canada, and Europe have suggested that day-to-day variations in ambient CO
 2      concentrations are related to cardiovascular hospital admissions and daily mortality, especially
 3      for individuals over 65 years of age. It is not clear, however, if the observed association results
 4      from CO or from combustion-related particles or, perhaps, from some other, unmeasured
 5      pollutant exposure that covaries in time with CO.
 6
 7      Cerebrovascular Effects
 8           Carbon monoxide hypoxia increases cerebral blood flow in healthy subjects, even at very
 9      low exposure levels. Thus, effects of CO on behavior are not consistent. Behaviors that require
10      sustained attention or performance are most sensitive to levels of COHb >5%.  Disease or injury
11      that impairs compensatory increases in blood flow may increase the probability of effects, but
12      little is known about the susceptibility of compromised individuals to ambient levels of CO.
13      Accidental exposures to CO have been shown to cause neurological problems weeks after
14      recovery from the acute episode.  It is not known, however, if these late neurological sequelae,
15      described as intellectual deterioration; memory impairment; and cerebral, cerebellar, and
16      mid-brain damage, result from long-term exposure to low ambient levels of CO.
17
18      Developmental Toxicity
19           Relatively high CO exposures of 150 to 200 ppm during gestation, leading to
20      approximately 15 to 25% COHb, produce reductions in birth weight, cardiomegaly, delays in
21      behavioral development,  and disruption in cognitive function in newborn laboratory animals of
22      several species. Little data exist on humans exposed to CO for predicting a lowest-observed-
23      effect level for developmental effects.  Studies relating human CO exposures from ambient
24      sources or cigarette smoking to reduced birth weight are  of concern because of the risk for
25      developmental disorders; however, many of these studies have not considered all sources of CO.
26      Nevertheless, some health professionals have considered this evidence sufficient to identify
27      pregnant women, and the developing fetus, as at risk to ambient levels of CO.
28
29      High-Altitude Effects
30           There are relatively few reports on the effects of inhaling CO at high altitudes. Current
31      knowledge supports the possibility that the effects of hypoxic hypoxia and CO-hypoxia are at

        February 15, 1999                          1-12        DRAFT-DO NOT  QUOTE OR CITE

-------
 1     least additive.  The potential additive effects of CO exposure in sensitive individuals visiting at
 2     high altitudes needs to be highlighted.
 3
 4     1.4.10  Carbon Monoxide Interaction with Drugs
 5          There remains little direct information on the possible enhancement of CO toxicity by
 6     concomitant drug use or abuse; however, there are some data on psychoactive drugs that suggest
 7     cause for concern.
 8
 9     1.4.11  Subpopulations at Risk
10          On the basis of known effects described, patients with reproducible exercise-induced
11     angina appear to be best established as a sensitive group within the general population that is at
12     increased risk for experiencing health effects of concern at ambient or near-ambient CO exposure
13     concentrations resulting in COHb levels <5%. Certain other groups are at potential risk from
14     exposure to CO, but further research is required to specify health effects associated with ambient
15     or near-ambient CO exposures in these probable risk groups.
16
17
18     1.5  METHODS AND PROCEDURES FOR DOCUMENT PREPARATION
19          The procedures that were followed for developing the revised criteria document for CO are
20     different from those that have been used for recent criteria documents. For example, the previous
21     CO criteria document (U.S. Environmental Protection Agency, 1991) was a more comprehensive
22     scientific review of available information on the  nature, sources, distribution, measurement, and
23     concentrations of CO in the environment and on the known and anticipated health effects that
24     CO would have on at-risk population groups. In lieu of a comprehensive review of the literature,
25     emphasis in the present criteria document has been placed on the development of a concise
26     summary of key information and a more interpretative discussion of the new scientific and
27     technological data available since the previous criteria were evaluated.  The resulting document
28     is more of an update, in accordance with the recommendations made by CASAC.
29          The main focus of this revised criteria document is on the evaluation and interpretation of
30     more recent air quality, human exposure, and health effects issues.  Therefore, the techniques

       February 15, 1999                          1-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      used to present this information vary according to the state-of-science for the respective topics.
 2      For example, the analysis of ambient air quality is based on newly obtained air monitoring data
 3      and utilizes the previous analysis only for showing trends over time.  As a result of the relatively
 4      dramatic decrease in ambient CO concentrations, population exposure to ambient CO also has
 5      declined.  Human exposure studies conducted in the early 1980s and earlier distributions of
 6      COHb levels in the U.S. population that were relied on heavily in the previous assessment are no
 7      longer relevant to the current picture of ambient CO exposure in the  1990s. Thus, key
 8      information on population exposure must focus on the newer studies and on modeling results.
 9      On the other hand, the health effects literature on CO has remained relatively static since the
10      previous 1991 assessment, except for provocative publications on cellular mechanisms of CO
11      action and on epidemiologic associations of ambient CO with mortality and morbidity in the
12      elderly population. Newly published studies on most of the other health outcomes reconfirm the
13      conclusions made in the last document and are incorporated into the previous  summaries by
14      reference only.
15           One of the first steps used in the development of this revised document was to convene
16      symposia or workshops to review the key scientific issues and to focus on the selection of
17      material that could be included in the document as the basis for the development of
18      standard-setting criteria. Both EPA and non-EPA scientific experts were utilized for this effort.
19           An interdisciplinary scientific symposium was held in Portland, OR, in December 1997, to
20      assess current scientific understanding of the atmospheric cycle of CO, including its sources,
21      sinks, and distribution.  The three main subject areas covered in the symposium relate to the
22      distribution and spatial and temporal variability of CO, the atmospheric budget of CO, and  direct
23      or indirect effects of CO on human health.  Results from papers presented at the symposium are
24      included, by reference,  in this revised criteria document.
25           A mini workshop, jointly organized by EPA, the Gas Research Institute, and the Health
26      Effects Institute, was convened in Chicago, IL, on April 24 and 25, 1998, to provide expert
27      scientific discussion  on the public health significance of exposures to low levels (<50 ppm) of
28      CO.  The three main topics covered in the meeting were human exposure patterns and trends in
29      CO exposure, pharmacokinetics and mechanisms of action of CO, and health effects.
30      A summary of the discussion by participants and conclusions drawn from the meeting were used
31      by authors in preparation of the draft criteria document chapters.

        February  15, 1999                          1-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Finally, a public peer-review workshop was convened on September 17 and 18, 1998, to
 2      define key issues, to review early drafts of the present criteria document chapters, and to
 3      ascertain and discuss any pertinent new literature.  The respective authors of the draft chapters or
 4      sections of the document revised them on the basis of the workshop recommendations.  The
 5      revised chapters of the document have been incorporated into this first external review draft
 6      being released for public comment and review by CASAC. Necessary revisions will be made in
 7      response to public comments and CASAC recommendations before the final version of the
 8      criteria document is released.
 9
10
11      1.6   ORGANIZATION AND CONTENT OF THE DOCUMENT
12           The updated air quality criteria document for CO critically evaluates and assesses scientific
13      information on air quality, human exposure, and health effects associated with exposure to the
14      concentrations found in the environment. Emphasis has been placed on the development of a
15      concise review of key information and a more interpretative discussion of the new scientific and
16      technological data available since completion of the previous criteria document (U.S.
17      Environmental Protection Agency, 1991). The references cited in the document should be
18      reflective of the state of knowledge through 1998 on those issues most relevant to review of the
19      NAAQSforCO.
20           To aid in the concise development of this document, summaries of the relevant published
21      literature and selective discussion of the literature has been undertaken.  Studies that were
22      presented in the previous criteria document and whose data were judged  to be significant because
23      of their usefulness in deriving the current NAAQS are discussed briefly in the text.  The reader,
24      however, has been referred to the more extensive discussion of these "key" studies in the
25      previous document. Other, older studies are discussed in the text if they  are open to
26      reinterpretation because of newer data, or potentially useful in deriving revised standards for  CO.
27      Generally, only published information that has undergone scientific peer review has been
28      included in the revised criteria document. However, some newer studies not yet published in the
29      open literature but meeting high standards of scientific reporting have been included for a few
30      areas.

        February 15, 1999                         1-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1          The structure of the present document follows the general outline of the previous criteria
 2     document (U.S. Environmental Protection Agency, 1991), especially for topics that have changed
 3     little since the last criteria review. The resulting sequence of discussion should help the reader to
 4     find and contrast similar sections.  There are, however, a few exceptions where some topics have
 5     been consolidated into a single chapter in order to present a more concise document.  The
 6     executive summary at the beginning of the document provides a concise presentation of key
 7     information and conclusions from all subsequent chapters.
 8          The document begins with this introduction (Chapter 1),  which provides the regulatory
 9     history of CO and an understanding of the scientific basis for the current CO NAAQS.
10     Information on analytical methods for monitoring CO (Chapter 2) covers the measurement of CO
11     in ambient (outdoor) and indoor air, as well as methods for measuring breath CO and blood CO
12     levels in exposed individuals.  Chapter 3 provides information on the atmospheric chemistry of
13     CO and typical sources, emissions, and concentrations found in the ambient and indoor
14     environments, topics addressed in separate chapters of the previous document. The remaining
15     chapters are similar to the previous document, covering topics on population exposure to CO
16     (Chapter 4), pharmacokinetics and mechanisms of action (Chapter 5), and health effects
17     (Chapter 6).  The last chapter (Chapter 7) provides an overall integrative summary of key
18     findings and an evaluation of subpopulations potentially at risk from exposure to CO.
       February 15, 1999                         1-16        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. F. R.
  7             (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       U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
19             quality standards. U. S. C. 42: §§7408-7409.
20       U.S. Environmental Protection Agency. (1979) Air quality criteria for carbon monoxide. Research Triangle Park,
21             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
22             no. EPA-600/8-79-022.  Available from: NTIS, Springfield, VA; PB81-244840.
23       U.S. Environmental Protection Agency. (1984) Revised evaluation of health effects associated with carbon
24             monoxide exposure: an addendum to the  1979 EPA air quality criteria document for carbon monoxide.
25             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
26             Assessment Office; report no. EPA-600/8-83-033F. Available from: NTIS, Springfield, VA;
27             PB85-103471/HSU.
28       U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
29             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
30             no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
31       U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
32             monoxide:  1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
33             Park, NC: Office of Air Quality Planning and Standards; report no.  EPA-452/R-92-004. Available from:
34             NTIS, Springfield, VA; PB93-157717.
35       U.S. Environmental Protection Agency. (1996) National air quality and emissions trends report,  1995. Research
3 6             Triangle Park, NC: Office of Air Quality Planning and Standards, Emissions Monitoring and Analysis
37             Division; report no. EPA/454/R-96-005. Available from: NTIS,  Springfield, VA; PB97-127500/XAB.
         February 15, 1999                             1-17         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
       February 15, 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
 6           AMBIENT 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 (see
27      Section 2.3.4.3). 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

        February 15, 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 etal., 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 mixing ratios in urban areas are typically 100s of ppb (Seinfeld and Pandis,
16      1998; Dickerson et al., 1992; Morales, 1998). A CO monitor with precision of 500 ppb would be
17      adequate to prove compliance with the CO standard, but would not provide adequate input data
18      for CTMs. This chapter, therefore, will review methods for measuring CO in ambient air that
19      provide sensitivity adequate to quantify the content of clean continental boundary layer air, that is
20      with uncertainty on the order of 10 ppb and has a detection limit around 50 ppb in addition to
21      methods in current use.  Suggested performance specifications for monitoring CO in nonurban
22      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.
        February 15, 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).
February 15, 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/cm3 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
      February 15, 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

        February 15, 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
        February 15, 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.
15      In the permeation tube techniques, a sample of the pure gas under pressure is allowed to diffuse
16      through a calibrated partition at a defined rate into a diluent gas stream to give a  standard sample
17      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.

        February 15, 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 mixing ratios from 60 to 170 ppb.  When data from
10      the various instruments were regressed, however, slopes again differed from unity by as much as
11      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 mixing ratios ranging from 10 ppm to 8%.  In a
16      blind intercomparison, the mean difference was 0.2%, well within the experimental uncertainty
17      of the techniques. Novell! et al. (1991) gravimetrically produced CO in  zero air in the range of
18      25 to 1,000 ppb from both pure CO and a NIST SRM; they found agreement to within 1%.
19      Agreement with commercially available NIST-traceable standards was within 3%. Reasonable
20      consistency (6% or better) was found with standards used by Australian, German, Brazilian,  and
21      several American institutions.  One Australian standard was found to be 22% lower, although
22      trouble with this standard had  been reported previously (Weeks et al., 1989). A reevaluation of
23      the reference scale in the range of nonurban ambient concentrations (Novell! et al., 1994)
24      confirmed agreement to within 5% or better for the National Oceanic and Atmospheric
25      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


        February 15, 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 mixing ratios down to
 5     the lowest expected in the boundary layer, about 50 ppb, agreement among groups was typically
 6     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

       February 15, 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, 199la).  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
        February 15, 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


        February  15, 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.
       February 15, 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.

        February 15, 1999                         2-14        DRAFT-DO NOT QUOTE OR CITE

-------
       A.
              Detector C[
                    LC
                   Source
                                                        Sample Cell
                                    ixJ^Correlation
                                           Cell
                                   n
                                                           Electronics
                                                                                  innnMM«_
       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).
       February 15, 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 (TR) 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      (Heidt, 1978; Greenberg et al., 1984; Hoell et al., 1987). Uncertainty on the order of 10 ppb or
23      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

        February 15, 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 offers high sensitivity, without the interferences
 4      inherent in continuous measurements (e.g., Novell! et al., 1991, 1998). Air samples are collected
 5      in glass bottles and injected into a gas chromatograph with two columns. The CO is then
 6      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
1 1           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. Ambient air samples are measured over open paths through the
16      ambient air (Chaney et al., 1979) or by pulling air samples through an orifice into a long-path cell
17      maintained at a pressure well below ambient (Sachse et al., 1987; Fried et al., 1991; Roths et al.,
18      1996). Radiation from a  TDL is modulated over a very narrow wavelength region such that
19      absorption by CO produces an AC signal.  The background is measured by catalytic oxidation of
20      CO to CO2.
21           Instruments based on TDLS are currently the fastest and most sensitive extant, with a
22      typical detection limit of a few ppb and a response time of a few seconds. For long-term
23      monitoring, the high cost and need for a skilled operator on site are disadvantages.
24
25      2.4.4.5  Resonance Fluorescence
26           Resonance fluorescence of CO in the vacuum UV has been used for a highly sensitive and
27      rapidly responding instrument (Volz and Kley, 1985; Gerbig et al., 1996). Excitation is
28      represented by the following reaction (Equation 2-1):
                                                                                         (2-1)
29

        February 15, 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., 1998) have reduced the detection limit to 3 ppb for a
 8      response time of 1 s. The high sensitivity and small size of the instrument are desirable for
 9      measurements from aircraft. Before the instrument is practical for air pollution monitoring, its
10      stability must be improved. As the lamp window degrades, sensitivity is lost, such that after
11      about 200 h of operation, a factor of two loss in the span can be expected.
12
13      2.4.5 Intercomparisons of Standards
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 mixing ratios that agree within experimental
17      uncertainty. These techniques include TDLS, NDIR/GFC, GC-FID, and GC-ML (Hoell et al.,
18      1987; Fried et al., 1991; Poulida et al.,  1991; Novell! et al., 1998). Additional details can be
19      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). More recently developed
26      electrochemical techniques show improved resolution and specificity (e.g., Ott et al., 1994;
27      Wallace  et al., 1988). These techniques are inadequate for ambient air monitoring to demonstrate
28      compliance but have use in personal exposure studies. Further research on the selectivity,
29      stability, and sensitivity of these methods is needed.
30

        February 15, 1999                         2-18       DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.5  MEASUREMENT USING PERSONAL MONITORS
 2          Monitors at fixed locations provide useful information on ambient CO concentrations and
 3     their variability and trends, but cannot measure personal exposure.  Information on personal
 4     exposure, including home, in-transit, and work-related concentrations is needed for
 5     epidemiological studies.  The previous criteria document (U.S. Environmental Protection
 6     Agency, 1991) reviewed the state of the science of personal monitors as of about 1986.  Since
 7     that time, the devices have been further developed and refined.
 8          One technique involves an ion-exchange Y-type zeolite, with zinc ion as the adsorbent.
 9     The adsorbent is desorbed thermally, converted to methane, and analyzed using GC-FID (e.g.,
10     Lee et al., 1992a,b; Lee and Yanagisawa, 1992;  1995).
11          Apte (1997) reviewed several of these devices and described passive samplers based on
12     transition metal compound color changes measured spectrochemically. The method suffers an
13     interference from ethylene, but provides adequate performance (sensitivity of 10 ppm-h and
14     precision of 20% or better) for health studies. Substantial work remains for most passive
15     samplers on stability and response to temperature, humidity, and interferences.  These techniques
16     lack the response speed and sensitivity for ambient air monitoring.
17
18
19     2.6  BIOLOGICAL MONITORING
20          A unique feature of CO exposure is that there is a biological marker of the dose that the
21     individual has received—the blood level of CO.  This level may be calculated by measuring
22     blood carboxyhemoglobin (COHb) or by measuring CO in end-tidal exhaled breath after a
23     standardized breathhold maneuver, with a required correction for the background CO inhaled
24     prior to a breathhold (Smith, 1977; Wallace, 1983). The measurement methods for COHb and
25     breath CO were reviewed extensively in the previous criteria document (U.S. Environmental
26     Protection Agency, 1991).  This section provides an update on advances in analytical methods for
27     measuring blood COHb and breath CO that have been published in the literature since the
28     previous review. New studies reporting breath CO or blood COHb in population studies are
29     discussed in Chapter 4, along with other new CO exposure assessments.
30

       February 15, 1999                         2-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.6.1  Carboxyhemoglobin Measurements
 2          Direct reading of COHb is usually performed in the clinical or hospital setting through the
 3     use of a direct-reading spectrophotometer, such as a CO-Oximeter (CO-Ox). For this use,
 4     precision on the order of ±1% COHb is not of primary importance, because of the need to
 5     differentiate between conditions of low levels of COHb and much higher levels of COHb that
 6     indicate treatment for CO poisoning. The concern in this setting, for example, is to rapidly
 7     distinguish between 1 and 10% COHb, not between 1 and 2% COHb.
 8          The performance of various CO-Ox instruments for measuring blood COHb was
 9     investigated by Vreman et al. (1993) and Mahoney et al. (1993) after the previous review
10     (U.S. Environmental Protection Agency, 1991).  They confirmed that considerable difficulties
11     still were encountered for COHb concentrations below 5% (a region with which most
12     environmental studies of nonsmokers are concerned), and the authors concluded that CO-Ox is
13     unreliable for environmental studies. The CO-Ox also was found to be influenced by bilirubin
14     and by fetal hemoglobin, presenting difficulty in diagnosing newborn infants with jaundice
15     (Stevenson and Vreman, 1997); however, a new commercial device was found to be unaffected
16     by these two problems (Vreman and Stevenson, 1994).
17          Marshall et al. (1995) showed a very wide range of "threshold" COHb values (measured in
18     the blood by CO-Ox, not estimated from breath CO) was used to determine treatment in a sample
19     of 23 Boston, MA, area laboratories. For example, eight laboratories accepted values of 5 to 6 %
20     COHb as normal in nonsmokers, a value that cannot be supported by the scientific literature.
21     The authors recommended the use of threshold limits of 3% COHb for nonsmokers and 10%
22     COHb for smokers when classifying subjects for treatment.
23          For a research study, the method of choice is GC for analysis of the CO gas released from
24     the blood when COHb is dissociated (U.S. Environmental Protection Agency, 1991).  The reader
25     is alerted to the difference between breath CO-blood COHb relationships when the COHb is
26     determined by CO-Ox or by GC, and a calibration curve relating exhaled breath CO to COHb
27     should be based on  a standard breathhold maneuver for the CO collection and the GC method of
28     COHb analysis.
29
30

       February 15, 1999                        2-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.6.2 Breath Carbon Monoxide Measurements
 2           Carbon monoxide in the breath can be measured by all techniques used to measure ambient
 3      CO concentrations, as described in U.S. Environmental Protection Agency (1991). A common
 4      type of instrument in use for rapidly screening large numbers of people for CO exposures or
 5      measuring breath CO distributions is the electrochemical analyzer. The subject performs an
 6      inhalation-breathhold maneuver and exhales through a mouthpiece into the instrument inlet. The
 7      end tidal breath is retained for analysis, and the reading in ppm CO can be converted to COHb
 8      through a calibration curve or nomogram provided with the instrument.
 9           Vreman et al. (1993) presented evidence to show that a serious positive interferent in the
10      electrochemical method (hydrogen gas) is present in the exhaled breath of some persons as a
11      result of metabolism of certain foods. Because this could have affected many previous studies,
12      including the very large EPA studies in Washington, DC, and Denver, CO (Akland et al., 1985),
13      it would be desirable to determine the fraction of the population so affected.  Because of the
14      general decline of ambient CO, this potential interference takes on more importance in any future
15      studies, which must account for this problem if employing electrochemical devices to measure
16      breath CO.
17           Lee et al. (1991) developed a TDLS system that was well suited for measuring low levels of
18      CO in breath.  The system also can detect the abundance of isotopic CO (13C16O), with a
19      preliminary finding of a slight enrichment over atmospheric abundance in breath. Lee et al.
20      (1994) employed the instrument in a study correlating breath CO and blood COHb in people
21      living near Boulder, CO (described in Chapter 4).
22           The passive CO sampler developed by Lee and Yanagisawa (1992, 1995) (see Section 2.5)
23      has a reusable sampling system that allows the collection of only the last 5 mL of a breath
24      expelled after breath holding for 20 s, thus obtaining alveolar air undiluted by dead space air.
25      The sampler was unaffected by humidity; however, the rather low efficiency of collection (50%)
26      and the resulting fairly high detection limit of 3.2 ppm may limit the utility of the sampler for
27      environmental studies.
28
29
30

        February 15, 1999                        2-21         DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.6.3  Relationships of Breath Carbon Monoxide to Blood
 2            Carboxyhemoglobin
 3          The end-tidal breath CO versus COHb relationships reviewed in the previous criteria
 4     document (see Table 8-14 in U.S. Environmental Protection Agency, 1991) and in studies
 5     published in the literature  since then are often at variance because they use either a 10-s, 15-s,
 6     or 20-s breathhold step in  the breath collection; use either GC or CO-Ox for the blood COHb
 7     measurements; or may not correct for the CO content of the inhaled air. The use of a 20-s
 8     breathhold, as recommended by Jones et al. (1958), with a correction for the CO content of the
 9     inhaled air (Smith,  1977; Wallace, 1983), would improve the reproducibility of the CO breath
10     measurements, and the use of GC would improve the accuracy of the corresponding COHb
11     measurement. The 20-s breathhold is preferable, because it maximizes the approach to
12     equilibrium and minimizes the magnitude of the required correction for CO in inhaled air.
13     Therefore, specific details regarding the length of the breathhold, corrections for inhaled CO, and
14     the method of COHb analysis should be provided in the published discussions of studies of the
15     CO-COHb relationship so that differences among study results can be evaluated.
16          One comprehensive  review article on CO-COHb relationships (Vreman et al., 1995)
17     discusses physical and chemical properties, endogenous and exogenous sources of CO, body
18     burden and elimination, toxicity and treatment, clinical chemistry, measurement methods, and the
19     relationship of CO  and COHb to bilirubin and jaundice in neonates. A second, less
20     comprehensive review from the same investigators focuses on the production of CO and bilirubin
21     in equal amounts by heme degradation and on the physiological significance of CO as a neuronal
22     messenger (Rodgers et al., 1994).
23          Lee et al. (1994) performed a study of CO-COHb relationships at altitude in Boulder.
24     A total of 13 nonsmoking  adults were exposed to 9 ppm CO for both 1 and 8 h. Blood was
25     sampled  and end-tidal breath  samples were taken after a 10-s breathhold. Mean COHb values
26     prior to exposure were 0.65% and, after exposure for the 1- and 8-h periods were 1.2 and 2.2%,
27     respectively.  The corresponding mean CO levels in the breath samples were about 2.4, 4.4, and
28     8.2 ppm  (uncorrected for the ~0 ppm ambient CO in inhaled air), respectively, as shown in
29     Figure 2-2.  The slope of 3.65 ppm per 1% COHb saturation after a 10-s breathhold at altitude is
30     somewhat smaller than previous estimates of about 5 ppm CO per 1% COHb, but the previous
       February 15, 1999                         2-22        DRAFT-DO NOT QUOTE OR CITE

-------
         10
      Q.
      Q_


      O
      O
      CO
      CD
          8-
     6-
     4-
          2-
          0
             0
               •  Oh
               •  1  h, 9 ppm CO
               A  8 h, 9 ppm CO
               r= 0.945
                               1
                               Blood Carboxyhemoglobin (%
      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

2

3


4

5
estimates were based on a 20-s breathhold near sea level that maximizes the end-tidal CO if the

inhaled air had a CO concentration below the 20-s end-tidal breath CO (Jones et al., 1958).



2.6.4 Summary of the Relationship Between Biological Measurements of

      Carbon Monoxide

     The use of CO-Ox to measure COHb provides useful information regarding values of

COHb in populations being studied for clinical diagnosis. However, the range of COHb values
      February 15, 1999
                                     2-23
DRAFT-DO NOT QUOTE OR CITE

-------
 1      obtained with this optical method for blood collected from nonsmokers is greater than that
 2      obtained from a split sample analyzed for COHb by research laboratory GC.  Therefore, the
 3      greater potential exists with the CO-Ox for having an incorrect absolute value for COHb, as well
 4      as an incorrectly broadened range of values, when used in population studies.  In addition, it is
 5      not clear exactly how sensitive the CO-Ox techniques are to small changes in COHb at the low
 6      CO end of the COHb dissociation curve.  Interferences (e.g., from variable levels of oxygen
 7      saturation of hemoglobin [O2Hb]) and nonlinear phenomena appear to have a very significant
 8      influence on the COHb reading at low COHb concentrations in a sample, suggesting nonlinearity
 9      or a disproportionality in the absorption spectra of different species of Hb (e.g., HbA [adult],
10      HbF [fetal], HbS [sickle], HbZH [Zurich]). Gas chromatography continues to be the method of
11      choice for measuring COHb in a research setting, although, with care, a CO-Ox can be specially
12      calibrated by GC analysis of calibration-standard blood samples prepared with low COHb
13      concentrations (Dahms and Horvath, 1974).
14           The measurement of exhaled breath CO has the advantages of ease, speed, precision
15      (provided the required correction for CO  in the inhaled air is made), and greater subject
16      acceptance than the invasive measurement of blood COHb.  Breath CO measurement on
17      randomly chosen people can be related to the blood COHb by use of an empirical relationship
18      developed by simultaneous measurements of COHb (preferably by GC) and breath CO using the
19      identical procedure for the breath collection that is used in the  population study. The empirical
20      relationships developed with different breath holding techniques will differ from the theoretical
21      Haldane equilibrium relationship for the reaction CO + O2Hb <->• O2 + COHb, which depends on
22      the ratio of adult- to fetal-hemoglobin (HbA:HbF).  This is because the Haldane relationship is
23      for static equilibrium, and the empirical end-tidal breath  CO-blood COHb relationship is for a
24      dynamic equilibrium that depends on how long the breath is held and on the correction for the
25      CO in inhaled air.
26
27
28      2.7  SUMMARY
29           The review of the state of the science for this criteria document yields several major points
30      concerning analytical techniques for CO measurement.

        February 15, 1999                          2-24        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Several adequate techniques exist for highly reliable monitoring of CO to ensure
 2      compliance with the NAAQS. Determination of the actual mean ambient air concentration
 3      requires substantially better performance than does the minimum required to demonstrate
 4      compliance with the NAAQS. Commercial instruments, sometimes with minor modifications,
 5      can meet the measurement needs for supplying useful data on the emission, distribution, and
 6      trends of ambient CO  and for modeling photochemical smog.
 7           Use of enhanced instruments for monitoring of actual CO concentrations with reasonable
 8      precision is needed if CO levels  in clean continental air outside of urban environments are to be
 9      quantified adequately. Commonly used calibration standards and measurement techniques have
10      in the past failed to meet the criteria of precise measurement, but there is now general agreement
11      on procedures for generating standards with absolute accuracy better than about 2% in the parts
12      per million range and about 10% in the range of mixing ratios found in the clean troposphere.
13      Several monitoring techniques have been intercompared and found reliable.
14           Several electrochemical and passive sampling methods have become available.  These
15      techniques are inadequate for compliance monitoring or precise measurements in ambient air,  but
16      may prove useful for personal exposure studies. Further work on the stability and specificity of
17      these methods is warranted. Accepted methods of analytical chemistry should be applied, and
18      the results should be published in the reviewed scientific literature.
19           The level of COHb in the blood may be determined directly by blood analysis or indirectly
20      by measuring CO in exhaled breath. The use of CO-Ox to measure COHb can provide useful
21      information regarding mean values  in populations being studied or as an aid in clinical diagnosis.
22      It has been shown, however, that the range of values obtained with this optical method will be
23      greater than that obtained with other more accurate methods, especially at COHb levels <5%.
24      Gas chromatography continues to be the method of choice for measuring COHb.
25           The measurement of exhaled breath has the advantages of ease, speed, precision, and
26      greater subject acceptance than measurement of blood COHb.  However, the accuracy of the
27      breath measurement procedure and  the validity of the Haldane relationship between breath and
28      blood still remains in question, especially at low environmental CO concentrations.
29
        February 15, 1999                         2-25        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       Apte, M. G. (1997) A population-based exposure assessment methodology for carbon monoxide: development of a
  6             carbon monoxide passive sampler and occupational dosimeter [dissertation]. Berkeley, CA: U.S. Department
  7             of Energy, Lawrence Berkeley National Laboratory, Environmental Energy Technologies Division;
  8             LBNL-40838. Available from: NTIS, Springfield, VA; DE98052714.
  9       Beckman, A. O.; McCullough, J. D.; Crane, R. A. (1948) Microdetermination of carbon monoxide in air: a portable
10             instrument. Anal. Chem. 20: 674-677.
11       Bergman, I.; Coleman, J. E.; Evans, D. (1975) A simple gas chromatograph with an electrochemical detector for the
12             measurement of hydrogen and carbon monoxide in the parts per million range, applied to exhaled air.
13             Chromatographia 8: 581-583.
14       Bruner, F.; Ciccioli, P.; Rastelli, R. (1973) The determination of carbon monoxide in air in the parts per billion
15             range by means of a helium detector. J. Chromatogr. 77: 125-129.
16       Brunke, E.-G.; Scheel, H. E.; Seiler, W.  (1990) Trends of tropospheric CO, N2O, and CH4 as observed at Cape
17             Point, South Africa. Atmos. Environ. 24A: 585-595.
18       Burch, D. E.; Gates, F. J.; Pembrook, J. D. (1976) Ambient carbon monoxide monitor. Research Triangle Park, NC:
19             U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; EPA report no.
20             EPA-600/2-76-210. Available from: NTIS, Springfield, VA; PB-259577.
21       Carter, W. P. L. (1991) Development of ozone reactivity scales for volatile organic compounds. Research Triangle
22             Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
23             Laboratory; report no. EPA/600/3-91/050. Available from: NTIS, Springfield, VA; PB91-243386.
24       Chaney, L. W.; McClenny, W. A. (1977) Unique ambient carbon monoxide monitor based on gas filter correlation:
25             performance and application. Environ. Sci. Technol. 11: 1186-1190.
26       Chaney, L. W.; Rickel, D. G.; Russwurm, G. M.; McClenny, W. A. (1979) Long-path laser monitor of carbon
27             monoxide: system improvements. Appl. Opt. 18: 3004-3009.
28       Code of Federal Regulations. (1991a) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
29       Code of Federal Regulations. (1991b) National primary and secondary ambient air quality standards. C. F. R.
30             40: §50.
31       Commins, B. T.; Berlin, A.; Langevin, M.; Peal, J. A. (1977) Intercomparison of measurement of
32             carboxyhaemoglobin in different European laboratories and establishment of the methodology for the
33             assessment of COHb levels in exposed populations. Luxembourg, Sweden: Commission of the European
34             Communities, Health and Safety Directorate; doc. no. V/F/1315/77e.
3 5       Dagnall, R. M.; Johnson, D. J.; West, T. S. (1973) A method for the determination of carbon monoxide, carbon
3 6             dioxide, nitrous oxide and sulphur dioxide in air by gas chromatography using an emissive helium plasma
37             detector. Spectrosc. Lett. 6: 87-95.
38       Dahms, T. E.; Horvath, S. M. (1974) Rapid, accurate technique for determination of carbon monoxide in blood.
39             Clin. Chem. (Washington, DC) 20: 533-537.
40       Dickerson, R. R.; Delany, A. C.  (1988) Modification of a commercial gas filter correlation CO detector for
41             enhanced sensitivity. J. Atmos. Oceanic Technol. 5: 424-431.
42       Elwood, J. H. (1976) A calibration procedural system and facility for gas turbine engine exhaust emission analysis.
43             In: Calibration in air monitoring: a symposium; August 1975; Boulder, CO. Philadelphia, PA: American
44             Society for Testing and Materials; pp. 255-272; ASTM special technical publication no. 598.
45       Federal Register. (1978) Air quality surveillance and data reporting: proposed regulatory revisions. F. R. (August 7)
46             43: 34892-34934.
47       Feldstein, M. (1967) Methods for the determination of carbon monoxide. Prog. Chem. Toxicol. 3: 99-119.
48       Fishman, J.; Seiler, W.; Haagenson, P. (1980) Simultaneous presence of O3 and CO bands in the troposphere. Tellus
49             32:456-463.
50       Fried, A.; Henry, B.; Parrish, D. D.; Carpenter, J. R.; Buhr, M. P. (1991) Intercomparison of tunable diode laser and
51             gas filter correlation measurements of ambient carbon monoxide. Atmos. Environ. 25A: 2277-2284.
52       Garbarz, J.-J.; Sperling, R. B.; Peache, M.  A. (1977) Time-integrated urban carbon monoxide measurements.
53             San Francisco, CA: Environmental Measurements, Inc.
         February 15,  1999                             2-26         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Gerbig, C.; Kley, D.; Volz-Thomas, A.; Kent, J.; Dewey, K.; McKenna, D. S. (1996) Fast response resonance
  2             fluorescence CO measurements aboard the C-130: instrument characterization and measurements made
  3             during North Atlantic Regional Experiment 1993. J. Geophys. Res. [Atmos.]  101: 29,229-29,238.
  4       Gerbig, C.; Schmitgen, S.; Kley, D.; Volz-Thomas, A.; Dewey, K.; Haaks, D. (1998) An improved fast-response
  5             vacuum-UV resonance fluorescence CO instrument. In press.
  6       Goldstein, G. M. (1977) COHb measurements, results of discussion between Drs. A. Berlin, CEC and G. Goldstein,
  7             EPA on ECE doc V/F/1315/77e [letter to Mr. Jack Thompson]. Research Triangle Park, NC: U.S.
  8             Environmental Protection Agency; September 23, 1977.
  9       Greenberg, J. P.; Zimmerman, P. R.; Heidt, L.; Pollock, W. (1984) Hydrocarbon and carbon monoxide emissions
10             from biomass burning in Brazil. J. Geophys. Res. [Atmos.] 89: 1350-1354.
11       Guenther, F. R.; Dorko, W. D.; Miller, W. R.; Rhoderick, G. C. (1996) The NIST traceable reference material
12             program for gas standards. Washington, DC: U.S. Department of Commerce, National Institute of Standards
13             and Technology; NIST special publication 260-126.
14       Haldane, J. (1897-1898) Some improved methods of gas analysis. J. Physiol. (London) 22: 465-480.
15       Heidt, L. E. (1978) Whole air collection and analysis. Atmos. Technol. (9): 3-8.
16       Hoell, J. M.; Gregory, G. L.; Carroll, M. A.; McFarland, M.; Ridley, B. A.; Davis, D. D.; Bradshaw, J.; Rodgers,
17             M. O.; Torres, A. L.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.; Campbell, M. C.; Farmer,
18             J. C.; Sheppard, J. C.; Wang, C. C.; Davis, L. I. (1984) An intercomparison of carbon monoxide, nitric oxide,
19             and hydroxyl measurement techniques: overview of results. J. Geophys. Res.  [Atmos.] 89: 11,819-11,825.
20       Hoell, J. M., Jr.; Gregory, G. L.; McDougal, D. S.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.
21             (1985) An intercomparison of carbon monoxide measurement techniques. J.  Geophys. Res. [Atmos.]
22             90: 12,881-12,889.
23       Hoell, J. M., Jr.; Gregory, G. L.; McDougal, D. S.; Sachse, G. W.; Hill, G. F.; Condon, E. P.; Rasmussen, R. A.
24             (1987) Airborne intercomparison of carbon monoxide measurement techniques. J. Geophys. Res. [Atmos.]
25             92:2009-2019.
26       Hughes, E. E.; Dorko, W. D.; Scheide, E. P.; Hall, L. C.; Beilby, A. L.; Taylor, J. K. (1973) Gas generation systems
27             for the evaluation of gas detecting devices. Washington, DC: U.S. Department of Commerce, National
28             Bureau of Standards; report no. NBSIR 73-292.
29       Hughes, E. E.; Davenport, A. J.; Woods, P. T.; Zielinski, W. L., Jr. (1991) Intercomparison of a range of primary
3 0             gas standards of carbon monoxide in nitrogen and carbon dioxide in nitrogen prepared by the National
31             Institute of Standards and Technology and the National Physical Laboratory. Environ. Sci. Technol.
32             25:671-676.
33       Hunt, W. F., Jr. (1972) The precision associated with the sampling frequency of log-normally distributed air
34             pollutant measurements. J. Air Pollut.  Control Assoc. 22: 687-691.
35       Jones, R. H.; Ellicott, M. F.; Cadigan, J. B.; Gaensler, E. A. (1958) The relationship between alveolar and blood
36             carbon monoxide concentrations during breathholding. J. Lab. Clin. Med. 51: 553-564.
37       Larsen, R. I. (1971) A mathematical model for relating air quality measurements to air quality standards. Research
3 8             Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; report no. AP-89.
39             Available from: NTIS, Springfield, VA; PB-205277.
40       Lee, K.; Yanagisawa, Y. (1992) Development of a sampler for carbon monoxide in expired air. Presented at:
41             85th annual meeting and exhibition of the Air & Waste Management Association; June; Kansas City, MO.
42             Pittsburgh, PA: Air & Waste Management Association; paper no.  92-145.03.
43       Lee, K.; Yanagisawa, Y. (1995) Sampler for measurement of alveolar carbon monoxide. Environ. Sci. Technol.
44             29: 104-107.
45       Lee, P. S.; Majkowski, R. F.; Perry, T. A. (1991) Tunable diode laser spectroscopy for isotope analysis - detection
46             of isotopic carbon monoxide in exhaled breath. IEEE Trans. Biomed. Eng. 38: 966-973.
47       Lee, K.; Yanagisawa, Y.; Hishinuma, M.; Spengler, J. D.; Billick, I. H. (1992a) A passive sampler for measurement
48             of carbon monoxide using a solid adsorbent. Environ. Sci. Technol. 26: 697-702.
49       Lee, K.; Yanagisawa, Y.; Spengler, J. D.; Billick, I. H. (1992) Measurement of personal carbon monoxide
50             exposures by mailed passive sampler. J. Air Waste Manage. Assoc. 42: 1212-1213.
51       Lee, P. S.; Schreck, R. M.; Hare, B. A.; McGrath, J. J. (1994) Biomedical applications of tunable diode laser
52             spectrometry: correlation between breath carbon monoxide and low level blood carboxyhemoglobin
53             saturation. Ann. Biomed. Eng. 22: 120-125.
54
         February 15, 1999                             2-27         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Mahoney, J. J.; Vreman, H. J.; Stevenson, D. K.; VanKessel, A. L. (1993) Measurement of carboxyhemoglobin and
  2             total hemoglobin by five specialized spectrophotometers (CO-oximeters) in comparison with reference
  3             methods. Clin. Chem. (Washington, DC) 39: 1693-1700.
  4       Marshall, M. D.; Kales, S. N.; Christian!, D. C.; Goldman, R. H. (1995) Are reference intervals for
  5             carboxyhemoglobin appropriate? A survey of Boston area laboratories. Clin. Chem. (Washington, DC)
  6             41: 1434-1438.
  7       McCullough, J. D.; Crane, R. A.; Beckman, A. O. (1947) Determination of carbon monoxide in air by use of red
  8             mercuric oxide. Anal. Chem. 19: 999-1002.
  9       Morales, R. M. (1998) Carbon monoxide, ozone, and hydrocarbons in the Baltimore/Washington Area
10             [dissertation]. College Park, MD: The University of Maryland, Department of Chemistry.
11       Moser, L.; Schmid, O. (1914) Die Bestimmung des Kohlenoxyds durch gelbes quecksilberoxyd [The determination
12             of carbon monoxide through yellow mercury oxide]. Z. Anal. Chem. 53: 217-233.
13       Mueller, R. H. (1954) A supersensitive gas detector permits accurate detection of toxic or combustible gases in
14             extremely low concentrations. Anal. Chem. 26: 39A-42A.
15       National Bureau of Standards. (1975) Catalog of NBS standard reference materials, 1975-76 edition. Washington,
16             DC: U.S. Department of Commerce, National Bureau of Standards; NBS special publication no. 260.
17       Novelli, P. C.; Elkins, J.  W.; Steele, L.P.(1991) The development and evaluation of a gravimetric reference scale
18             for measurements of atmospheric carbon monoxide. J.  Geophys. Res. [Atmos.] 96: 13,109-13,121.
19       Novelli, P. C.; Collins, J. E., Jr.; Myers, R. C.; Sachse, G. W.; Scheel, H. E. (1994) Reevaluation of the
20             NOAA/CMDL carbon monoxide reference scale and comparisons with CO reference gases at
21             NASA-Langley and the Fraunhofer Institut. J. Geophys. Res. [Atmos.] 99: 12,833-12,839.
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       O'Keeffe, A.  E.; Ortman, G. C. (1966) Primary standards for trace gas analysis. Anal. Chem. 38: 760-763.
28       Ott, W. R.; Mage, D. T. (1975) Random sampling as an inexpensive means for measuring average annual air
29             pollutant concentrations in urban areas. Presented at: 68th annual meeting of the Air Pollution Control
30             Association; June; Boston, MA. Pittsburgh, PA: Air Pollution Control Association; paper no. 75-14.3.
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       Palanos, P. N. (1972) A practical design for an ambient carbon monoxide mercury replacement analyzer. Anal.
34             Instrum. 10: 117-125.
35       Parrish, D. D.; Holloway, J. S.; Fehsenfeld, F. C. (1994) Routine, continuous measurement of carbon monoxide
36             with parts per billion precision.  Environ. Sci. Technol. 28: 1615-1618.
37       Paulsell, C. D. (1976) Use of the National Bureau of Standards standard reference gases in mobile source emissions
38             testing. In: Calibration in air monitoring: a symposium; August 1975; Boulder, CO. Philadelphia, PA:
39             American Society for Testing and Materials; pp. 232-245; ASTM special technical publication 598.
40       Phinney, D. E.; Newman, J. E. (1972) The precision associated with the sampling frequencies of total paniculate at
41             Indianapolis, Indiana. J. Air Pollut. Control Assoc. 22: 692-695.
42       Porter, K.; Volman, D. H. (1962) Flame ionization detection of carbon monoxide for gas chromatographic analysis.
43             Anal. Chem. 34: 748-749.
44       Poulida,  O.; Dickerson, R. R.; Doddridge, B. G.; Holland, J. Z.; Wardell, R. G.; Watkins, J. G. (1991) Trace gas
45             concentrations and meteorology in rural Virginia. 1. Ozone and carbon monoxide. J. Geophys. Res. [Atmos.]
46             96:22,461-22,475.
47       Rhodes, R. C.; Evans, E. G. (1987) Precision and accuracy assessments for state and local air monitoring networks
48             1985. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring
49             Systems Laboratory; EPA report no. EPA-600/4-87-003. Available from: NTIS, Springfield, VA;
50             PB87-145447.
51       Robbins, R. C.; Borg, K. M.; Robinson, E. (1968) Carbon monoxide in the atmosphere. J. Air Pollut. Control
52             Assoc. 18: 106-110.
53       Rodgers, P. A.; Vreman, H. J.; Dennery, P. A.; Stevenson, D. K. (1994) Sources of carbon monoxide (CO) in
54             biological systems and applications of CO detection technologies. Semin. Perinatol. 18: 2-10.
         February 15, 1999                             2-28         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Roths, J.; Zenker, T.; Parchatka, U.; Wienhold, F. G.; Harris, G. W. (1996) Four-laser airborne infrared
  2             spectrometer for atmospheric trace gas measurements. Appl. Opt. 35: 7075-7084.
  3       Ryan, W. F.; Doddridge, B. G.; Dickerson, R. R.; Morales, R. M; Hallock, K. A.; Roberts, P. T.; Blumenthal,
  4             D. L.; Anderson, J. A.; Civerolo, K. (1998) Pollutant transport during a regional O3 episode in the
  5             mid-Atlantic states. J. Air Waste Manage. Assoc. 48: 786-797.
  6       Sachse, G. W.; Hill, G. F.; Wade, L. O.; Perry, M. G. (1987) Fast-response, high-precision carbon monoxide sensor
  7             using a tunable diode laser absorption technique. J. Geophys. Res. [Atmos.] 92: 2071-2081.
  8       Scaringelli, F. P.; Rosenberg, E.; Rehme, K. A. (1970) Comparison of permeation devices and nitrite ion as
  9             standards for the colorimetric determination of nitrogen dioxide. Environ. Sci. Technol. 4: 924-929.
10       Seiler, W.; Giehl, H.; Roggendorf, P. (1980) Detection of carbon monoxide and hydrogen by conversion of mercury
11             oxide to mercury vapor. Atmos. Technol. (12): 40-45.
12       Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
13             New York, NY: John Wiley & Sons, Inc.
14       Smith, N. J. (1977) End-expired air technic for determining occupational carbon monoxide exposure. J. Occup.
15             Med. 19: 766-769.
16       Smith, F.; Nelson, A. C., Jr. (1973) Guidelines for development of a quality assurance program: reference method
17             for the continuous measurement of carbon monoxide in the atmosphere. Research Triangle Park, NC: U.S.
18             Environmental Protection Agency, Quality Assurance and Environmental Monitoring Laboratory; EPA
19             report no. EPA-R4-73-028a. Available from: NTIS, Springfield, VA; PB-222512.
20       Smith, R. G.; Bryan, R. J.; Feldstein, M.; Locke, D. C.; Warner, P. O. (1975) Tentative method for constant
21             pressure volumetric gas analysis for O2, CO2, CO, N2, hydrocarbons (ORSAT). Health Lab. Sci. 12: 177-181.
22       Stedman, D. H.; Kok, G.; Delumyea, R.; Alvord, H. H. (1976) Redundant calibration of nitric oxide, carbon
23             monoxide, nitrogen dioxide, and ozone air pollution monitors by chemical and gravimetric techniques.
24             In: Calibration in air monitoring. Philadelphia, PA: American Society for Testing and Materials; pp. 337-344.
25             (ASTM special technical publication 598).
26       Stevenson, D. K.; Vreman, H. J. (1997) Carbon  monoxide and bilirubin production in neonates. Pediatrics
27             100:252-254.
28       Swinnerton, J. W.; Linnenbom, V. J.; Cheek, C. H. (1968) A sensitive gas chromatographic method for determining
29             carbon monoxide in seawater. Limnol. Oceanogr. 13: 193-195.
30       Tesarik, K.; Krejci, M. (1974) Chromatographic determination of carbon monoxide below the  1 ppm level.
31             J. Chromatogr. 91: 539-544.
32       U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques,  §109, national ambient air
33             quality standards. U. S. C. 42: §§7408-7409.
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.
37       U.S. Environmental Protection Agency. (1997) EPA traceability protocol for assay and certification of gaseous
3 8             calibration standards. Washington, DC: National Center for Environmental Research and Quality Assurance;
39             report no. EPA/600/R-97/121. Available from: NTIS, Springfield, VA; PB98-115314.
40       Volz, A.; Kley, D. (1985) A resonance-fluorescence instrument for the in-situ measurement of atmospheric carbon
41             monoxide. J. Atmos. Chem.  2: 345-357.
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.; Stevenson, D. K. (1994) Carboxyhemoglobin determined in neonatal blood with a CO-oximeter
45             unaffected by fetal oxyhemoglobin. Clin. Chem. (Washington, DC) 40: 1522-1527.
46       Vreman, H. J.; Mahoney, J. J.; Stevenson, D. K. (1995) Carbon monoxide and Carboxyhemoglobin.  Adv. Pediatr.
47             42:303-334.
48       Wallace, L. A. (1983) Carbon monoxide in air and breath of employees in an underground office. J. Air Pollut.
49             Control Assoc. 33: 678-682.
50       Wallace, L.; Thomas, J.; Mage, D.; Ott, W. (1988) Comparison of breath CO, CO exposure, and Coburn model
51             predictions in the U.S. EPA  Washington-Denver (CO) study. Atmos. Environ. 22: 2183-2193.
52       Weeks, I. A.; Galbally, I. E.; Fraser, P. J.; Matthews, G. (1989) Comparison of the carbon monoxide standards used
53             at Cape Grim and Aspendale. In: Forgan, B. W.; Ayres, G. P., eds. Baseline Atmosheric Program (Australia)
54             1987. Canberra, New South  Wales: Department of Administrative Services, Bureau of Meteorology;
55             pp. 21-25.


         February 15, 1999                             2-29         DRAFT-DO NOT QUOTE OR CITE

-------
1      Weinheimer, A. J.; Ridley, B. A. (1990) A cautionary note on the use of some mass flow controllers. J. Geophys.
2            Res. [Atmos.] 95: 9817-9821.
3      Wohlers, H. C.; Newstein, H.; Daunis, D. (1967) Carbon monoxide and sulfur dioxide adsorption on- and
4            desorption from glass, plastic, and metal tubings. J. Air Pollut. Control Assoc. 17: 753-756.
       February 15, 1999                            2-30         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          Current information about the abundance and distribution, the nature of sources and sinks
 8     and the chemistry of carbon monoxide (CO) in various environments, ranging from the global
 9     background to indoor air, is summarized in this chapter. Carbon monoxide is studied in each of
10     these widely varied environments for very different reasons.  In indoor environments with
11     sources such as malfunctioning or misused combustion appliances, automobile exhaust from
12     attached garages, and cigarettes and, in outdoor environments such as urban areas where
13     emissions from motor vehicles or wood burning can cause high concentrations to exist, carbon
14     monoxide is of direct concern because of health effects resulting from human exposure to these
15     high concentrations. Human exposures to CO are discussed in Chapter 4.
16          Carbon monoxide in the atmosphere outside of these situations is of interest because of its
17     importance to atmospheric chemistry. Carbon monoxide can affect the formation of ozone (O3)
18     and other photochemical oxidants in the atmosphere.  Carbon monoxide strongly influences the
19     abundance of hydroxyl  radicals (OH), thus affecting the global cycles of many biogenic and
20     anthropogenic trace gases, such as methane (CH4), that affect the abundance of stratospheric
21     O3 and the energy budget of the atmosphere. Changes in CO levels, therefore, may contribute to
22     widespread changes in atmospheric chemistry and indirectly affect global climate. In this
23     chapter, the global scale aspects of CO are discussed first, and then the  discussion proceeds to
24     successively smaller scales. An overview of the major sources and sinks of CO and the resulting
25     CO distribution on a global basis and the importance of CO to tropospheric chemistry is
26     presented in Section 3.2, followed by a discussion of nationwide emissions of CO in Section 3.3.
27     Nationwide trends in ambient CO levels and related discussions on CO air quality are presented
28     in Section 3.4, and concentrations and sources of CO in indoor environments are discussed in
29     Section 3.5.
30

       February 15, 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.
7
                   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, aromatic and 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
       February 15, 1999                           3-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1     waters (oceans, lakes, and rivers) and on the soil surface occurs. Carbon monoxide is lost
 2     primarily by reaction with atmospheric OH radicals and by uptake by soil microorganisms.
 3           More detailed descriptions of the nature of individual sources of primary CO shown in
 4     Table 3-1 and estimates of the strengths of these sources, along with similar material for
 5     nonchemical sinks of CO, are given in Section 3.2.2. Carbon monoxide concentrations and
 6     trends in the background atmosphere are discussed in Section 3.2.1, and the atmospheric
 7     chemistry of CO, including the formation of secondary CO, is discussed in  Section 3.2.3.
 8
 9     3.2.1 Global Background Concentrations of Carbon Monoxide
10           In common usage, the term "background levels" refers to concentrations observed in
11     remote areas relatively unaffected by local pollution sources. However, several definitions of
12     background levels  are possible (see Chapter 6, U.S. Environmental Protection Agency, 1996).
13     The two definitions chosen in that document as being most relevant for regulatory purposes and
14     for providing corrections to assessments of the health risks posed by exposure to CO are based
15     on estimates of contributions from uncontrollable sources that can affect CO levels in the United
16     States. The first definition includes anthropogenic and natural sources outside North America,
17     and natural sources within North America. The second definition includes  only natural sources
18     within and outside North America. These background  levels refer to concentrations that would
19     be present because of the presence of these sources  alone. Because of long-range transport from
20     anthropogenic source regions in North America, it may be impossible to obtain background
21     levels defined above solely on the basis of direct measurement. However, some inferences about
22     what these levels may be can be made with the help of numerical models and historical data.
23           Surface measurements of CO concentrations are made routinely as part of the National
24     Oceanic and Atmospheric Administration's Climate Monitoring Diagnostics Laboratory
25     (NOAA/CMDL) Global Cooperative Air Sampling Network (e.g., Hofmann et al., 1996).
26     Carbon monoxide flask samples are collected weekly in flasks or continuously with in situ gas
27     chromatographs at about 40 remote sites around the world. These sites are  located primarily in
28     the marine boundary layer, with a few located in continental  areas.  The latitudinal and seasonal
29     variations in CO levels are summarized in the three-dimensional diagram shown in Figure 3-1
30     (Novell! et al., 1998). Annual average CO mixing ratios are about 130 ppb in the Northern

       February 15,  1999                          3-3        DRAFT-DO NOT QUOTE OR CITE

-------
       Q.
       Q.
          250 -1
      o  100 J

\
                         o
         LATITUDE    30° S
                         60°S
                            90"S
                                        91
     92
 93     94
      YEAR
                               95
                                        96
      Figure 3-1.  Latitudinal and seasonal variability in CO concentrations obtained in the
                  NOAA/CMDL monitoring network.
      Source: Novell! etal. (1998).
1     Hemisphere and about 50 ppb in the Southern Hemisphere. Seasonal maxima in CO mixing
2     ratios occur during late winter in both hemispheres, and minima occur during late summer, with
3     about a factor of two variation between maximum and minimum values. Carbon monoxide is
4     well mixed in high latitudes of both the Northern Hemisphere and the Southern Hemisphere.
5     A steep gradient in CO mixing ratios exists between about 30° north (N) latitude and about
6     10° south (S) latitude. Carbon monoxide concentrations range from a minimum of about 45 ppb
7     during summer in the Southern Hemisphere to about 220 ppb at high latitudes in the Northern
8     Hemisphere during winter.  Thus, CO concentrations in remote areas of the Northern Hemisphere
9     are only a small fraction («1 to 2%) of those of concern to human health (as given by the
      February 15, 1999
  3-4
DRAFT-DO NOT QUOTE OR CITE

-------
 1      National Ambient Air Quality Standards [NAAQS] for CO of 9 ppm for the second highest,
 2      nonoverlapping 8-h average concentration).
 3           During the 1980s, there were sufficient data on tropospheric air quality trends to suggest
 4      that CO concentrations measured in remote areas were increasing globally at approximately
 5      1% per year, based on data collected by the Oregon Graduate Institute (OGI) (Khalil and
 6      Rasmussen, 1988).  This increase presumably resulted from anthropogenic activities. From
 7      about 1988 until 1992, global average CO concentrations declined rapidly at a rate of about
 8      -2.6 ± 0.8% per year (Khalil and Rasmussen, 1994), perhaps caused by reductions in fossil fuel
 9      combustion and tropical biomass burning. Possible increases in tropospheric OH levels resulting
10      from enhanced transmission of solar ultraviolet radiation caused by stratospheric O3  depletion
11      may have been an additional factor.  Since 1993, these investigators have found that CO
12      concentrations have leveled off. However, these results differ from those reported by
13      NOAA/CMDL. Taken collectively over the last 10 years (as of December 1997), data collected
14      at in situ and flask monitoring stations by OGI and NOAA/CMDL show no significant trend
15      (Hard etal., 1998).
16           Because direct measurements of sufficient precision for defining trends have come into use
17      only within the last  15 to 20 years, estimates of longer term trends in CO levels must come from
18      indirect means. Rinsland and Levine (1985) derived an increase in  the mean tropospheric CO
19      abundance of about 2% per year from 1950 to 1984, based on an examination of solar spectra
20      captured on photographic plates in Europe. Carbon monoxide concentrations measured in air
21      bubbles trapped in the ice sheets of Greenland and Antarctica have been used as proxies for CO
22      concentrations in ambient air at the time the air bubbles were sealed from the atmosphere (Haan
23      et al., 1996). Carbon monoxide concentrations  derived this way for the preindustrial era (roughly
24      corresponding to the year 1850, when anthropogenic activities should not have influenced
25      significantly the atmospheric composition) are about 90 ppb for the high-latitude Northern
26      Hemisphere and are about 50 ppb for the high-latitude Southern Hemisphere. However, it should
27      be noted that the CO in the trapped air bubbles also may result from the decomposition of
28      organic compounds also trapped in the same air bubbles, and that it is difficult to extract CO
29      from the air bubbles without contamination. Both factors tend to cause positive artifacts in the
30      CO concentrations reported. In addition, the Northern Hemispheric value derived from the ice
31      cores is higher than that predicted by atmospheric model studies of the preindustrial  era that

        February 15, 1999                          3-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1     indicate CO mixing ratios of about 50 ppb (Thompson and Cicerone, 1986; Pinto and Khalil,
 2     1991; Thompson et al., 1993). Some enhancement of Northern Hemispheric values over
 3     Southern Hemispheric values during the preindustrial era is likely because of the greater mass of
 4     vegetation that can emit NMHCs in the Northern Hemisphere.
 5
 6     3.2.2  Sources and Global Emissions Estimates of Carbon Monoxide
 7          Global CO emission estimates are summarized in Table 3-2.  By far, the largest
 8     contribution from fossil fuel combustion on the global scale arises from motor vehicles.
 9     Variables controlling the formation of CO during combustion of any fuel  are oxygen
10     concentration, flame temperature, gas residence time at high temperature, and mixing in the
11     combustion zone. In general, increases in all four factors result in lower amounts of CO
12     produced relative to carbon dioxide (CO2). Carbon monoxide is produced primarily during
13     conditions of incomplete combustion. The estimates for fossil fuel emissions by stationary
14     sources, shown in Table 3-2, do not include significant contributions from power plants, because
15     fuels are burned with high efficiency in modern power plants. Rather, they are based on
16     estimates of CO emitted in small, hand-fired furnaces used for domestic purposes (e.g., cooking,
17     heating, water sterilization) and in inefficient boilers and furnaces used in small-scale industrial
18     operations.  This latter source is of significance only in eastern Europe and in developing
19     countries of Africa and Asia (especially China).  For example, Pinto and Rasmussen (1998) have
20     estimated that as much as 50% of CO observed during a severe air pollution episode in Prague,
21     Czech Republic, in February 1993 may have come from inefficient coal burning.  However, it
22     also should be noted that the importance of this source has been declining as heating needs are
23     met increasingly by centralized power plants.
24          Biomass burning consists of burning vegetation to clear new land for agriculture and
25     population resettlement; to control the growth of unwanted plants on pasture land; to dispose of
26     agricultural and domestic waste; and as fuel for cooking, heating, and water sterilization.
27     Biomass burning exhibits strong seasonality, with most biomass burned during the local dry
28     season.  The smoldering phase of combustion yields higher emissions factors than the flaming
29     phase. Lobert et al. (1991) found, in controlled combustion chamber experiments with a wide
30     variety of vegetation types, that, on average, 84% of CO was produced during the smoldering
31     phase and 16% during the flaming phase of combustion. Smoldering conditions are more
       February 15, 1999                          3-6        DRAFT-DO NOT QUOTE OR CITE

-------
TABLE 3-2. ANNUAL GLOBAL CARBON MONOXIDE EMISSIONS ESTIMATES
cr
i
«!
y,
VO
VO
VO








s



O
^
H
O'
0
2;
0
H
O
O
H
W
O
O
H
W
(in teragrams [Tg] per year)
Allen et al. Logan et al. Seiler and Conrad Pacyna and Graedel
(1996) (1981) (1987) (1995)
Sources


Fossil fuel combustion 329 450a 640 ±200 440 ±150
Biomass burning 370 655 1,000 ±600 700 ±200
Natural NMHC oxidation 618 560 900 ±500 800 ±400
Anthropogenic NMHC oxidation — 90 — —
Methane oxidation 722 810 600 ± 300 600 ±200
Oceans — 40 100 ± 90 50 ±40
Soils — — — —
Vegetation — 130 75 ±25 75 ± 25
Total 2,039 2,735 3,3 15 ±1,700 2,700 ±1,000
Sinks
Soils
OH reaction


Total

"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.











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















-------
 1      prevalent during the burning of large pieces of vegetation, such as trees, compared with grasses.
 2      Nonetheless, most CO is produced in the tropics by savanna burning (mainly in Africa), followed
 3      by burning forests, fuel wood, and agricultural waste.  Less than 20% of the CO produced by
 4      biomass burning originates in middle and high latitudes (Andreae, 1991; Levine and Pinto,
 5      1998).
 6           The other sources of CO shown in Table 3-2 all have large natural components. Carbon
 7      monoxide may be evolved from the photodecomposition of organic matter in surface waters
 8      (such as oceans, rivers, and lakes) and the soil surface.  Soils can act as a source or a sink for
 9      carbon monoxide, depending on soil moisture, the intensity of sunlight reaching the soil surface,
10      and soil temperature (e.g., Inman et al., 1971; Conrad and Seller, 1985). Soil uptake of CO
11      occurs because of anaerobic bacteria and potentially can be  an important sink, as soils exposed to
12      test atmospheres containing 100 ppm CO could reduce the CO concentrations to near zero within
13      a few hours (Inman et al., 1971). Emissions of CO from soils appear to occur by abiotic
14      processes, such as thermodecomposition or photodecomposition of organic matter. In general,
15      warm and moist conditions favor CO uptake, whereas hot and dry conditions favor the release of
16      CO. Liebl and Seller (1976) have estimated a mean deposition velocity of 0.04 cm/s for CO.
17      However, despite the small mean value for the deposition velocity of CO, uptake by soil
18      microbes still may be capable of depleting CO in the stable nocturnal boundary layer (Moxley
19      and Cape, 1997).  The value reported for soil emissions in Table 3-2 is based on very limited
20      data, and hence it is difficult even to assign uncertainty bounds (Conrad, 1996).
21          Estimates of the magnitude of the soil sink range from 250 to 640 Tg/year (Logan et al.,
22      1981; Cicerone, 1988), with a current "besf'estimate of 300 Tg/year, with an uncertainty range of
23      a factor of three (Dignon et al., 1998).  More extensive field measurements, perhaps based on the
24      eddy correlation technique (Ritter et al., 1994), are needed to characterize the variability and the
25      direction of the CO flux to the soil surface. Most CO in the atmosphere is lost by its oxidation to
26      CO2 by OH radicals. Reaction with OH radicals accounts for a loss of 2,300 Tg/year, with an
27      uncertainty factor of 1.4 (Dignon et al., 1998). Because of large uncertainties in individual
28      sources and sinks, the imbalance between sources (2,500  Tg/year) and sinks (2,600 Tg/year) is
29      not significantly different from zero.
30
31

        February 15, 1999                          3-8         DRAFT-DO NOT QUOTE OR CITE

-------
 1     3.2.3  The Atmospheric Chemistry of Carbon Monoxide
 2          Carbon monoxide is produced by the photooxidation of CH4 and other organic compounds
 3     (including NMHCs) in the atmosphere and of organic molecules in surface waters and soils
 4     (Table 3-1 and 3-2). Estimates of CH4 emissions from the various source categories shown in
 5     Table 3-1 can be found in the Intergovernmental Panel on Climate Change report (1996).
 6     Methane oxidation can be summarized by the following sequence of reactions:
 7
 8                                   CH4 + OH - CH3 + H2O
 9                                   CH3 + O2 (+M) - CH3O2 (+M)
10                                   CH3O2 + NO - CH3O + NO2
11                                   CH3O + O2 - CH2O + HO2
12                                   CH2O + hu - H2 + CO,
13                              or    CH2O + hu - HCO + H,
14                              or    CH2O + OH - HCO  + H2O
15                                   HCO + O2 - CO + HO2,
16
17     where M is a mediator (e.g., nitrogen, O2, argon, CO2).  The photolysis of formaldehyde (CH2O)
18     proceeds by two pathways, the first yields molecular hydrogen (H2) plus CO (55%), and the
19     second yields atomic hydrogen  (H) plus the formyl radical (HCO) (45%), where the percentages
20     are given for overhead sun conditions (Rogers, 1990).  Formyl radicals then react with molecular
21     oxygen (O2) to form the hydroperoxy radical (HO2) plus CO. In addition, the reaction of the
22     methyl peroxy radical (CH3O2)  with HO2 radicals, forming methyl hydroperoxide (CH3OOH),
23     needs to be considered, especially in low nitrogen oxide (NOX) environments.  The heterogeneous
24     removal of soluble intermediate products, such as CH3OOH, CH2O, and radicals, decreases the
25     yield of CO from the oxidation of CH4.
26          Althought the oxidation of CH2O nearly always results in CO formation (except for the
27     formation of small quantities of formic acid in the reaction of CH2O with HO2), the oxidation of
28     acetaldehyde (CH3CHO) does not always yield two CO molecules.  The photolysis of CH3CHO
29     also involves pathways that produce molecules and radicals, namely CH4 + CO and CH3 + HCO.
30     Estimates of the yield of CO from the photooxidation of CH4 and CH3 are  subject to the same
31     considerations outlined above.  The reaction of CH3CHO with OH radicals can yield acetyl
       February 15, 1999                         3-9         DRAFT-DO NOT QUOTE OR CITE

-------
 1      radicals (CH3CO). The acetyl radicals then will participate with O2 in a termolecular
 2      recombination reaction to form acetyl peroxy radicals, which then can react with nitric oxide
 3      (NO) to form CH3 + CO2 (or the acetyl peroxy radicals can react with nitrogen dioxide (NO2) to
 4      form peroxy acetyl nitrate [PAN]).  Thus, one of the carbon atoms can be oxidized directly to
 5      CO2 without passing through CO.  The yield of CO depends on the OH concentration and the
 6      photolysis rate of CH3CHO, as well as on the abundance of NO, as acetyl peroxy radicals also
 7      can react with HO2 and other hydrogen-bearing radicals.
 8          Estimates of the yield of CO from the oxidation of more complex hydrocarbons requires
 9      the calculation of the yields of CH2O, CH3CHO, CH3CO, and analogous radicals from the
10      oxidation of the parent molecule. Likewise, the extent of heterogeneous removal of soluble
11      intermediate products needs to be considered in the oxidation of more complex hydrocarbons.
12      However, in contrast to simple hydrocarbons containing one or two carbon atoms, detailed
13      kinetic information is lacking about the gas phase oxidation pathways of many anthropogenic
14      hydrocarbons (e.g., aromatic compounds, such as benzene and toluene), biogenic hydrocarbons
15      (e.g., isoprene, the monoterpenes), and their intermediate oxidation products (e.g., epoxides,
16      nitrates, carbonyl compounds). As much as 30% of the carbon in hydrocarbons in many urban
17      areas is in the form of aromatic compounds (Grosjean and Fung, 1984; Seila et al., 1989). Yet
18      mass balance analyses performed on irradiated smog chamber mixtures of aromatic hydrocarbons
19      indicate that only about one-half of the carbon is in the form of compounds that can be identified.
20      Reactions that have condensible products, such as those occurring during the oxidation of
21      terpenes, also need to be considered because these reactions produce secondary organic
22      particulate matter, thereby reducing the potential yield of CO.
23          The yield of CO from the oxidation of CH4 is about 0.9, and it is about 0.4 from the
24      oxidation of ethane and propane, on a per carbon basis from estimates based on atmospheric
25      model results (Kanakidou et al., 1991). Jacob and Wofsy (1990) estimated that 1 mole of CO is
26      produced by the oxidation of 1 mole of isoprene (corresponding to a conversion factor of 0.2 on a
27      per carbon basis) for low NOX levels.  For higher NOX levels, they  estimated that 3 moles of CO
28      are produced per mole of isoprene oxidized (corresponding to a conversion factor of 0.6  on a per
29      carbon basis).  Isoprene accounts for most of the CO produced by the photochemical oxidation of
30      NMHCs shown in Table 3-2. An indication of the possible importance of the oxidation of
31      isoprene as a source of CO is provided by Yarwood and Morris (1998).  They found that the

        February  15, 1999                         3-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      oxidation of isoprene accounted for over 50% of the CO in many rural areas in a simulation of an
 2      O3 episode (July 7 to 18, 1995) in the eastern half of the United States.
 3           Although the source of secondary CO may be important on regional and global scales,
 4      it may not be as important when considering CO in urban areas. In analyzing the results of a
 5      numerical simulation of an O3 episode in California's San Joaquin Valley from August 3 to 4,
 6      1990, Jeffries (1998) found that the peak photochemical oxidation of NMHCs produced only
 7      2.5 ppb CO per hour, compared with an average CO mixing ratio of 260 ppb. Thus, it would
 8      take about 100  h of uninterrupted oxidation of NMHCs under optimal conditions to produce the
 9      observed CO. The lack of importance of photochemical sources in this air shed reflects the
10      combination of concentrated, local, automotive emissions and the lack of complete oxidation of
11      the NMHCs emitted in the model domain.
12           The major pathway removing CO from the atmosphere is by its reaction with OH radicals.
13      There have been numerous determinations of the rate coefficient for this reaction. The most
14      recent evaluation of kinetics data for use in atmospheric modeling (National Aeronautics and
15      Space Administration, Panel for Data Evaluation, 1997) recommends a value of 1.5 x  10"13
16      (1 + 0.6 Patm) cm3 molecules"1 s"1, with a value of 0 ± 300 K for E/R for the reaction
17
18                                       OH + CO - Products.
19
20      This reaction proceeds through two channels. The bimolecular channel yields H + CO2, whereas
21      the addition channel leads to the formation of a carboxyl radical (HOCO). In the presence of O2,
22      the HOCO intermediate is converted to HO2 + CO2. Therefore, for atmospheric purposes, the
23      products of the reaction OH + CO can be taken to be HO2 and CO2.
24           Estimates of OH radical concentrations can be used along with the rate coefficient given
25      above to calculate the lifetime of CO in the atmosphere. Measurements of OH radical
26      concentrations in situ (Hard et al., 1992; Mount and Williams,  1997; Poppe et al., 1994) in the
27      lower troposphere show that their levels are highly  site specific and are highly variable in space
28      and time. Typical mid-latitude noontime values during summer (when OH concentrations are at
29      their highest levels) range from about 5 to 10 x 106 OH/cm3 and are much lower during other
30      times of the day and during other seasons.  As a result, it is difficult to derive average values that
31      would be meaningful for use in calculating the atmospheric lifetime of long-lived species that

        February 15, 1999                         3-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      react with OH radicals, based on direct measurements.  Modeling the atmospheric distribution of
 2      methyl chloroform (CH3CC13) has been used to derive diurnal and global average OH values for
 3      calculating the atmospheric lifetimes of long-lived species by comparing predictions to
 4      observations (Prinn et al., 1992). Average OH values derived in this manner are about
 5      8 x 10s OH/cm3.  By further adjusting the OH fields derived in a simulation of the CH3CC13
 6      distribution, to optimize the fit between the measurements and simulations of CH3CC13
 7      concentrations, Krol et al. (1998) derived concentrations of 1.00 x 106 OH/cm3 in  1978 and
 8      1.07  x 106 OH/cm3 in 1993. The resulting trend in OH values is estimated to be
 9      0.46 ± 0.5% year"1.  The resulting atmospheric lifetime of CO is approximately 2 mo in middle
10      latitudes (range, 1 to 4 mo,  depending on latitude and season, with highest values during winter
11      and at higher latitudes). These lifetimes  are shorter than the characteristic time  scale for mixing
12      between the hemispheres (about 1 year),  and hence a large gradient in concentrations can exist
13      between the hemispheres (see Figure 3-1). In addition, the chemical lifetime of CO at high
14      latitudes is long enough to result in much smaller gradients between 30°  latitude and the pole of
15      either hemi sphere.
16           Reaction with CO, is in turn, the major reaction of OH  radicals. The reaction of CO with
17      OH radicals constitutes at least 50% of the tropospheric sink of OH radicals (e.g.,  Collins et al.,
18      1997). Thus, changes in the abundance of CO could lead to changes in the abundance of a
19      number of trace gases whose major loss process involves reaction with OH radicals. These trace
20      gases can absorb infrared radiation from  the earth's surface and contribute to the greenhouse
21      effect (e.g., CH4) or can deplete stratospheric O3 (e.g., methyl chloride [CH3C1], methyl bromide
22      [CH3Br], and hydrochlorofluorocarbons,  such as difluorochloromethane). Because of the
23      importance of CO in determining OH levels, interest has focused on the possible effects of
24      increases in anthropogenic CO emissions on the concentrations of gases such as CH4 (Sze,  1977;
25      Chameides et al., 1977; Thompson and Cicerone,  1986). For instance, Thompson and Cicerone
26      (1986) found in numerical simulations, in which the CO mixing ratio at the surface was allowed
27      to increase by 1% per year from 1980 to  2000, while holding CH4 emissions constant, that the
28      mixing ratio of CH4 at the surface increased by about 12% (corresponding to a mean increase of
29      0.56% per year), and that the mixing ratio of CH4 at the surface increased by about 30%
30      (corresponding to a mean increase of 1.3% per year) for an increase in surface CO mixing ratio
31      of 2% per year. However, based on the trend results reported earlier, Briihl and Crutzen (1998)

        February 15, 1999                         3-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1     have examined the consequences of decreases in CO levels for atmospheric chemistry. They
 2     found, by decreasing CO emissions linearly by about 20% between 1990 and 2000, that the CO
 3     reductions could lead to a significant decrease (-25%) in the growth rate of CH4 and even to a
 4     decrease in CH4 levels in the case of constant CH4 emissions. An accurate knowledge of the
 5     sources and sinks of carbon monoxide in the atmosphere is therefore necessary for assessing the
 6     effects of future increases in anthropogenic CO emissions on the concentrations of the
 7     above-mentioned radiatively and photochemically important trace species. However, because of
 8     nonlinearities introduced into the calculation of OH radical concentrations by short-lived NOX
 9     (e.g., Hameed et al., 1979), an accurate assessment of these effects awaits the development of
10     three-dimensional chemistry and transport models incorporating the spatial variability of NOX
11     (e.g., Kanakidou and Crutzen, 1993).
12          In the free troposphere, in the absence of significant quantities of NMHCs, the effects of
13     CO on tropospheric O3 can be summarized as shown below.
14
                     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
                    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
       February 15, 1999                         3-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      O3 production resulting from CO in the remote troposphere can be made by taking the overall
 2      rate of the reaction of CO with OH radicals and then correcting for the fraction of HO2 radicals
 3      that do not react with NO, based on free radical balances presented by Collins et al. (1997). This
 4      quantity (i.e., the rate of conversion of NO to NO2 by HO2 radicals produced by the reaction of
 5      CO with OH radicals) represents 20 to 40% of the production of O3 on a global basis.
 6           The effects of CO on O3 photochemistry in environments with abundant hydrocarbons
 7      (e.g., cities, tropical rain forests) require a much more complex treatment that includes the
 8      competition for OH radicals by CO and NMHCs and the effects of this competition on the
 9      overall budget of hydrogen-containing radicals (i.e., OH, HO2). In urban environments, reaction
10      with OH radicals represents the major loss process for NMHCs and initiates the sequence of
11      further reactions leading to the formation of O3 and CO itself.  Detailed analyses of the radical
12      balances (i.e.,  production and loss rates in each reaction) in urban air chemistry models, as
13      performed by Jeffries (1998), can give the amount of O3 formed because of the reaction of CO.
14      However, only a few such analyses have been performed. Jeffries (1998), in a numerical
15      simulation of an O3 episode in the San Joaquin Valley from August 3 to 4,  1990, found that
16      reaction with CO constituted 27% of the loss of OH radicals. It was found, by tracking sources
17      of various radicals produced by the oxidation of VOCs and that oxidize NO to NO2, that CO
18      accounted for  about 14% of the O3 formed in this example (compared to about 86% for VOCs).
19      He found that  CO was responsible for 18% of the O3 production in another simulation of an
20      O3 episode in Charlotte, NC. Methane accounted for 6%, paraffins (with more than three carbon
21      [C] atoms) for 25%, ethane for 6%, olefins for 4%, toluene for 2%, xylenes for 3%, isoprene for
22      10%, aldehydes for 21%, methylglyoxyl for 1%, PAN for 3%, and other species for about 1% of
23      the O3 produced in this episode. In examining grid cells downwind of New York City, NY,
24      during an O3 episode, Jeffries (1998) also found that CO accounted for about 11% of the
25      O3 formed. Obviously, more analyses of this sort are needed to characterize regional differences
26      in the importance of CO in different cities in the United States, which may have very different
27      combinations of CO, NMHC, and NOX concentrations than those used in these examples.
28      Because of nonlinearities in the production rate of O3 involving each of the above species,
29      caution should be exercised in attempting to estimate the effects of variations in CO levels on
30      O3 production rates in the case  studies cited above.
31

        February  15, 1999                         3-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1     3.3  NATIONWIDE CARBON MONOXIDE EMISSIONS ESTIMATES
 2           Total primary CO emissions were estimated for the period of 1987 through 1996 using the
 3     National Emissions Inventory Trends database (U.S. Environmental Protection Agency, 1998).
 4     These emissions (see Table 3-3) are shown in the original units used in their calculation
 5     (i.e., thousands of short tons per year).  A short ton is equal to 2,000 Ib or 9.08 x 10s g.  Table 3-3
 6     shows that total CO emissions decreased by 18.4% from 1987 to 1996; however, the fractional
 7     contribution of transportation (the major source of CO both then and now) remained constant at
 8     79%. In addition, there are several categories, such as fuel consumption by electric utilities and
 9     industry and nonroad mobile sources, in which emissions have increased over the same period.
10     Total CO emissions for the United States were reported to be 66,189 thousand short tons for
11     1990, the last year reported in the previous air quality criteria document (AQCD) for CO. It can
12     be seen from inspection of Table 3-3 that the values for 1990 have been revised upward in the
13     interim  to 96,535 thousand short tons. The upward revision in values  for 1990 is primarily the
14     result of changes in the methods for calculating motor vehicle emissions. The MOBILES
15     emissions factor model (U.S. Environmental Protection Agency, 1993) replaced the earlier
16     MOBILE4.1 version (U.S. Environmental Protection Agency, 1991a).  Changes were made in
17     methods for calculating inputs to the model (e.g., temperatures and operating mode) and in the
18     method for calculating vehicle miles traveled. Additional differences  relate to the use of
19     county-level statistics for vehicle registration, as well as the use of temperature data from
20     individual counties.
21           In addition, emissions of secondary CO, such as from the oxidation of isoprene also should
22     also be  mentioned. Annual emissions of isoprene in the contiguous United States are about
23     17.2 Tg/year (Pierce and Dudek, 1996). A source of CO of 8.9 Tg/year is calculated using the
24     conversion factor of 0.25 for C in isoprene to C in CO estimated by Yarwood and Morris (1998).
25     This value represents about 11% of the estimated U.S. emissions for CO in 1995, as shown in
26     Table 3-3.  The oxidation of anthropogenic and other natural NMHCs may supply an additional
27     2 to 3 Tg CO per year.
28           A number of techniques, such as roadside tunnel sampling and the remote sensing of
29     individual motor vehicle emissions have been applied in the past several years at a number of
30     locations throughout the United States to test CO emissions estimates  and to derive emissions
31     factors (i.e., emissions per unit distance traveled). Two major points have been realized on the
       February 15, 1999                         3-15        DRAFT-DO NOT QUOTE OR CITE

-------
TABLE 3-3. NATIONWIDE CARBON MONOXIDE EMISSIONS ESTIMATES, 1987 to 1996
cr
i
VO
VO
VO






1
Oi

O
>
H
6
o
o
H
0
0
H
W
0
O
H
w
(thousands of short tons per year)
Source Category
Fuel Combustion
Electric Utilities
Coal
Oil
Gas
Internal combustion
Industrial
Coal
Oil
Gas
Other
Internal combustion
Other
Residential wood
Other
Industrial Processes
Chemical and Allied Processing
Metals Processing
Petroleum and Related
Industries

1987
6,967
307
223
20
53
10
649
85
46
252
171
96
6,011
5,719
292
6,851
1,798
1,984
455


1988
7,379
320
236
25
48
11
669
87
46
265
173
98
6,389
6,086
303
7,034
1,917
2,101
441


1989
7,449
327
239
26
51
11
672
87
46
271
173
96
6,449
6,161
288
7,013
1,925
2,132
436


1990
5,510
363
234
20
51
57
879
105
74
226
279
195
4,269
3,781
488
5,852
1,183
2,640
333


1991
5,856
349
234
19
51
45
920
101
60
284
267
208
4,587
4,090
497
5,740
1,127
2,571
345


1992
6,155
350
236
15
51
47
955
102
64
300
264
227
4,849
4,332
517
5,683
1,112
2,496
371


1993
5,586
363
246
16
49
51
1,043
101
66
322
286
268
4,181
3,679
502
5,898
1,093
2,536
371


1994
5,519
370
247
15
53
55
1,041
100
66
337
287
251
4,109
3,607
502
5,839
1,171
2,475
338


1995
5,934
372
250
10
55
58
1,056
98
71
345
297
245
4,505
3,999
506
5,790
1,223
2,380
348


1996
5,962
377
263
11
44
59
1,072
99
72
348
305
247
4,513
3,993
520
5,817
1,223
2,378
348



-------
February 15, '.
VO
VO
VO






OJ
1
^

o
H
6
O
o
H
0
0
H
W
0
O
H
w
TABLE 3-3 (cont'd). NATIONWIDE CARBON MONOXIDE EMISSIONS ESTIMATES, 1987
(thousands of short tons per year)
Source Category
Other Industrial Processes
Solvent Utilization
Storage and Transport
Waste Disposal and Recycling
Transportation
On-Road Vehicles
Nonroad Sources
Miscellaneous
Structural Fires
Agricultural Fires
Prescribed Burning
Forest Wildfires
Other
Total All Sources
1987
713
2
50
1,850
86,209
71,250
14,959
8,852
242
483
4,332
3,795
NA
108,879
1988
711
2
56
1,806
86,861
71,081
15,780
15,895
242
612
4,332
10,709
NA
117,169
1989
716
2
55
1,747
81,832
66,050
15,781
8,153
242
571
4,332
3,009
NA
104,447
1990
537
5
76
1,079
73,965
57,848
16,117
11,208
164
415
4,668
5,928
32
96,535
1991
548
5
28
1,116
78,114
62,704
16,040
8,751
166
413
4,713
3,430
28
98,461
1992
544
5
17
1,138
76,233
59,859
16,374
7,052
168
421
4,760
1,674
30
95,123
1993
594
5
51
1,248
76,794
60,202
16,592
7,013
169
415
4,810
1,586
34
95,291
1994
600
5
24
1,225
78,706
61,833
16,873
9,614
170
441
4,860
4,114
28
99,677
to 1996
1995
624
6
25
1,185
70,947
54,106
16,841
7,050
171
465
4,916
1,469
28
89,721

1996
635
6
25
1,203
69,946
52,944
17,002
7,099
172
475
4,955
1,469
27
88,822
Note: Some columns may not sum to totals because of rounding.
Source: U.S. Environmental Protection Agency






(1998).




































-------
 1      basis of these studies: first, that a small percentage of motor vehicles are responsible for most of
 2      the emissions, and, second, that CO and hydrocarbon emissions have been systematically
 3      underestimated by as much as a factor of two in emissions factor models.
 4           Roadside remote sensing data indicate that about 50% of CO and NMHC emissions are
 5      produced by only about 10% of the vehicles (Lawson et al., 1990; Stephens and Cadle, 1991).
 6      These "superemitters" are typically older, poorly maintained vehicles.  There are also a surprising
 7      number of newer vehicles that are classified as superemitters. Possible reasons are related to
 8      tampering with emissions control systems to improve milage, the use of contaminated fuels that
 9      may interfere with the proper operation of emissions control systems, and the lack of
10      maintenance of emissions control equipment. In addition to the above activities, so-called
11      off-cycle operations also can result in enhanced emissions relative to those conditions for which
12      emissions testing is usually done. For example, rapid accelerations have been  shown to increase
13      emissions relative to less  stressful driving modes.
14           A comparison of emissions factors computed on the basis of tunnel measurements in
15      Van Nuys, CA, during the South Coast Air  Quality Study in 1987 with those calculated by
16      emissions inventory models indicated that CO emissions were underpredicted by emissions
17      models (i.e., the Emissions Factor 7C [EMFAC7C] model, which is similar to  MOBILES) by a
18      factor of 2.7, and hydrocarbon emissions were underestimated by a factor of 3.8 (Ingalls et al.,
19      1989; Pierson et al., 1990). A reinterpretation of the Van Nuys tunnel data by Pollack et al.
20      (1998) indicated that emissions factors calculated using MOBILESa were only a few percent
21      more than the ambient tunnel data indicated (21.3 versus 20.9 g/mi), compared with a factor of
22      two difference using EMFAC7F (9.6 versus 20.9 g/mi). Likewise, a comparison of emission
23      factors  computed on the basis of measurements in the Fort McHenry, MD, and Tuscarora, PA,
24      tunnels with those calculated by emissions models (MOBILE4.1 and MOBILES) illustrated the
25      extent of model dependence on the intercomparisons (Pierson et al., 1996). In general, both
26      versions of the  model gave predictions within ±50% of observations most of the time. However,
27      both versions showed different tendencies to either over- or underpredict, suggesting that the
28      model predictions were within the range of uncertainty of the model.
29           Comparisons of ambient air quality data with predictions of emissions factor models have
30      been made for conditions when ambient concentrations result primarily from local emissions
31      with minimal photochemical processing and minimal transport from locations with different

        February 15, 1999                         3-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      source characteristics. The optimal time to obtain such conditions is during the early morning,
 2      when ambient concentrations of CO, non-methane organic compounds (NMOCs), and NOX
 3      typically peak and are dominated by local mobile source emissions (Fujita et al., 1992). These
 4      comparisons have been performed in California for the Los Angeles Basin (Fujita et al., 1992),
 5      the San Joaquin Valley, and San Francisco Bay Area (Magliano et al.,  1993), and for the
 6      Lake Michigan air quality region (Korc et al., 1993). A fairly consistent picture of
 7      underpredictions of ambient CO to NOX and NMOC to NOX ratios by emissions factor models,
 8      after allowing for correction for the artifacts mentioned above, emerges from these studies.
 9      In the Los Angeles Basin study, ambient CO to NOX ratios were factors of 1.3 to 2.9 higher than
10      corresponding emissions inventory ratios during summer, and factors of 1.2 to 2.4 higher than
11      predicted by emissions models during fall. In the San Joaquin Valley study, ambient CO to NOX
12      ratios ranged from factors of 1.1 to 7.2 higher than predicted by emission models. In the Lake
13      Michigan area study, ambient CO to NO ratios ranged from factors of 1.7 to 4.7 higher than
14      predicted by emissions models. However, more recent comparisons between ambient and
15      emission inventory CO/NOX for Los Angeles and the San Joaquin Valley show better agreement
16      than in previous studies (Croes  et al., 1996; Ipps and Popejoy, 1998; Haste et al.,  1998).
17
18
19      3.4  CARBON MONOXIDE CONCENTRATIONS IN AMBIENT AIR
20          The U.S. Environmental Protection Agency's (EPA's) Aerometric Information Retrieval
21      System (AIRS) receives data from the National Air Monitoring Stations (NAMS) and the State
22      and Local Air Monitoring Stations (SLAMS).  Current NAAQS define 1- and 8-h average
23      concentrations that should not be exceeded more than once per year. The standards are met if the
24      second highest 1-h value is less than 35 ppm (40 mg/m3), and the second highest nonoverlapping
25      8-h value is less than 9 ppm (10 mg/m3). Nationwide trends in ambient CO concentrations are
26      presented in Section 3.4.1, diurnal variations in ambient CO levels are presented in Section 3.4.2,
27      and a more detailed characterization of the spatial and temporal variability in ambient CO
28      concentrations in selected urban areas are presented in Appendix 3-A.  The analyses in
29      Appendix 3-A were performed  for the Denver, CO;  Los  Angeles; New York City, NY; and
30      Phoenix, AZ, Metropolitan Statistical Areas.
31
        February 15, 1999                         3-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      3.4.1 Nationwide Trends in Ambient Carbon Monoxide Concentrations
 2           In 1996, 554 monitoring sites reported ambient CO air quality data to EPA's AIRS. Most
 3      CO monitoring stations in the United States are located in larger urban areas. Figure 3-2 displays
 4      the geographic locations of the monitoring sites reporting CO data to AIRS for 1996.  On the
 5      map, the sites are identified as NAMS, SLAMS, or "other". The NAMS were established by
 6      EPA to ensure a long-term national network for urban-area-oriented ambient monitoring and to
 7      provide a systematic, consistent database for air quality comparisons and trends analysis. The
 8      SLAMS allow state or local governments to develop networks tailored for their immediate
 9      monitoring needs. These NAMS and SLAMS sites conform to uniform criteria for monitor
10      siting, instrumentation, and quality assurance.  "Other" monitors may be Special Purpose
11      Monitors, monitors at industrial sites, monitors on tribal lands, etc. Although state and local air
12      programs may require extensive monitoring to document and measure the local impacts of CO
13      emissions, only two NAMS sites  are required in urbanized areas with populations greater than
14      500,000.  Two categories of NAMS sites are required: (1) peak concentration areas (microscale),
15      such as major traffic corridors,  street canyons, and major arterial streets, and (2) areas with high
16      population and traffic densities (middle scale or neighborhood scale).
17           Only nine of the sites shown in Figure 3-2 failed to meet the 8-h standard of 9 ppm, and
18      none of the 554 monitoring sites exceeded the 1-h standard of 35 ppm in 1996. The locations of
19      these nine sites, along with the  second highest 8-h CO concentration and the number of
20      exceedances for 1996, are given in Table 3-4.
21           Figure 3-3 shows the consistent, downward trend in the nationwide composite average of
22      the annual second highest 8-h CO concentration during the past 20 years (1977 through 1996).
23      This statistic relates directly to  the averaging time and form of the current CO NAAQS and
24      complies with the recommendations of the Intra-Agency Task Force on Air Quality Indicators
25      (U.S. Environmental Protection Agency,  1981). The dashed curve in Figure 3-3 tracks the trend
26      in the composite mean of the annual second highest 8-h average concentration for  102
27      monitoring sites that reported ambient air quality data in at least 17 of the past 20 years, 1977 to
28      1996. All monitoring sites are  weighted equally when computing the nationwide composite
29      mean concentration. This selection criterion maximizes the number of sites available for trend
30      analyses.  This subset of sites yields good geographical coverage with sites from more than
31      50 cities in 28 states. Each year,  site leases are lost, or sites are discontinued, and new sites come
        February  15, 1999                         3-20        DRAFT-DO NOT QUOTE OR CITE

-------
       Figure 3-2. Locations of sites in the nationwide ambient CO monitoring network, 1996.
 1     online; therefore, the 102 long-term-trend sites compose less than 20% of the currently active CO
 2     monitors. The solid line in Figure 3-3 shows the trend in the composite mean for a larger
 3     database of 345 sites that have reported ambient CO monitoring data in at least 8 of the past
 4     10 years.  Missing annual second-highest CO concentration data for the second through ninth
 5     years are estimated by linear interpolation from the surrounding years.  Missing endpoints are
 6     replaced with the nearest valid year of data.  This latter procedure explains the discrepancy
 7     between the two curves in 1987 and 1988.  Specific computational details are described
 8     elsewhere (U.S. Environmental Protection Agency, 1998).  This larger data set permits the
 9     examination of the inter-site variability in peak CO concentrations.  Figure 3-4 presents the 10th,
10     50th, and 90th percentile concentrations and the composite mean concentrations across these
11     345  sites. The 10th, 50th, and 90th percentile concentrations CO concentrations for each year are
       February 15, 1999
3-21
DRAFT-DO NOT QUOTE OR CITE

-------
                  TABLE 3-4.  SITES NOT MEETING THE CARBON MONOXIDE
                    NATIONAL AMBIENT AIR QUALITY STANDARDS IN 1996
1996
Location
Lynwood, CA°
Calexico, CA
Kalispell, MT
Hawthorne, CAC
Anchorage, AK
El Paso, TX
Las Vegas, NV
Phoenix, AZ
Anchorage, AK
AIRS Site ID No.
06037 1301
06 025 0005
30 029 0045
06 037 5001
02 020 0037
48 141 0027
32 003 0557
04 013 0022
020200017
2nd Maxa
14.5
14.1
11.1
10.5
10.5
10.3
10.1
10.0
9.6
No. Exc.b
22
9
2
5
3
2
3
2
3
       aSecond highest 8-h average CO concentration.
       bNumber of exceedances of CO NAAQS.
       "Both sites are located in Los Angeles County.
 1     indicated, respectively, by the bottom, middle, and top lines of each box. For example, 10% of
 2     the 345 trends sites reported 1987 second highest 8-h CO concentrations lower than the bottom
 3     of the first bar in Figure 3-4. The yearly composite mean across all 345 sites is indicated by the
 4     "x" in each bar. Figure 3-5 shows trends in CO concentrations in each of the different sampling
 5     environments (urban, suburban, and rural sites). As can be seen from Figure 3-5, the downward
 6     trend in ambient CO concentrations occurred at monitoring sites in urban, suburban, and rural
 7     environments.
 9     3.4.2  Orcadian Patterns in Carbon Monoxide Concentrations
10          The diurnal variation in winter time, composite, hourly CO concentrations is shown in
11     Figure 3-6 from 1987 through 1996 (Cohen and Iwamiya,  1998). It can be seen that hourly mean
12     CO concentrations peak during the morning rush hours (7  to 9 a.m.). This peak results mainly
13     from CO emitted into the relatively shallow morning boundary layer by motor vehicles (e.g.,
14     Fujita et al., 1992).  The CO mixing ratios decline towards mid-afternoon, as the height of the
       February 15, 1999
3-22
DRAFT-DO NOT QUOTE OR CITE

-------
            Q.
            g
            v^
            03
            -i— •
            §
            c
            o
            O
                14
                12
                10
                                               1977-1996  1987-1996
                                               (102 sites)  (345 sites)
                     I   I    I   I
                                                                            I    I   I    I
                    77  78  79  80  81  82  83  84  85  86  87 88  89 90  91  92  93  94  95 96
                                                    Year
       Figure 3-3.  Nationwide composite average of the annual second highest 8-h CO
                   concentrations, 1977 to 1996.
       Source: U.S. Environmental Protection Agency (1998).
1      atmospheric mixing layer increases and then increase again with the onset of the evening rush
2      hour. Carbon monoxide levels fall off less rapidly after the afternoon peak because the mixing
3      layer height decreases during evening and nighttime. There is a general decrease in CO levels
4      during the night because of a lack of fresh emissions combined with processes such as mixing
5      with CO-poor areas and deposition to the surface.  The downward trend in CO concentrations
6      from 1987 to 1996 is apparent for all times of the day.  Especially notable is the decease in 7 to
7      9 a.m. CO concentrations, which is consistent with the decrease in motor vehicle emissions that
8      was noted earlier for the same period.  During the period from 1987 to 1996, the 24-h nationwide
9      composite average CO concentration decreased from 2.0 to 1.2 ppm.
      February 15, 1999
3-23
DRAFT-DO NOT QUOTE OR CITE

-------
               15H
           Q.  10 -
           Q.
           o
           "CD
           
-------
             Q_
             Q_
             o
            O
            O
            O
                  8
                  6-
                  4H
2-
                                                	Rural
                                                	Suburban
                                                      Urban
                      87    88    89    90
                               91     92
                                Year
   93
94    95    96
       Figure 3-5. Composite average of the annual second highest 8-h CO concentrations for
                  rural, suburban, and urban sites, 1987 to 1996.
       Source: U.S. Environmental Protection Agency (1998).
 1     others in nighttime hours.  Five of those six monitors are still in operation; Table 3-5 summarizes
 2     and compares their 1988 record with 1996 data.
 3          Note that in Table 3-5 and in the analysis of 1996 data shown in Table 3-6 a
 4     "running-average" exceedance is defined as any hour that culminates in an 8-h average higher
 5     than 9.5 ppm. This definition differs from that used in the construction of Table 3-4 in that the
 6     number of nonoverlapping exceedances was used in Table 3-4. A formal violation of the 8-h
 7     standard occurs when, in a given year, a second 8-h average exceeds 9.5 ppm but does not
 8     overlap the first 8-h exceedance. Exceedances culminating in any hour are treated here because
 9     an individual's cumulative exposure to a level greater than 9.5 ppm could occur in any hour.
10          At the Lynwood station, the running-average exceedances have declined from 392 in 1988
11     to 110 in 1996 (22 nonoverlapping exceedances); the majority of exceedances in 1996 occurred
       February 15, 1999
                              3-25
DRAFT-DO NOT QUOTE OR CITE

-------
 Q.
 Q.
g

2
-i—»



I
o
O
    2.5-
    2.0-
    1.5-
    1.0-
    0.5-
           \   \   \   \   \    \
       1234567
                                 \   \   \   \    \   \   \   \   \    \   \   \   \    \   \

                                9  10 11   12  13  14 15  16  17  18 19  20  21  22  23  24
                                           Hour
 Figure 3-6.  Diurnal variation of nationwide composite hourly average CO concentrations

            for winter (December to February), from 1987 to 1996.



 Source: Cohen and Iwamiya (1998).
     TABLE 3-5.  RUNNING-AVERAGE EXCEEDANCES OF THE 9 ppm 8-HOUR

               CARBON MONOXIDE STANDARD, 1988 VERSUS 1996
Location
Lynwood, CA
Hawthorne, CA
Las Vegas, NV
New York City, NY
Steubenville, OH
Spokane, WA
AIRS Site ID No.
06 037 1301
06 037 5001
32 003 0557
36 061 0081
39 081 1012
53 056 0040
1988
392
163
102
123
152
169
1996
110
19
O
NAa
0
1
aNA = not available.
February 15, 1999
                                      3-26
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     in the hours between midnight and sunrise, as they had in 1988. They occurred in the months of
2     January, February, November, and December.
3          At the Hawthorne station, running-average exceedances have declined from 163 in 1988 to
4     19 in 1996 (five nonoverlapping exceedances).  These are clustered around sunrise when
5     dispersion is most likely to be at a minimum. The exceedances at this station also occur in the
6     winter quarter.

      February 15, 1999                        3-27        DRAFT-DO NOT QUOTE OR CITE

-------
 1          Only a small number of stations have several running-average exceedances; however, these
 2     recurrent high levels are attributed to unusual local situations. A prime example is the
 3     monitoring station in Calexico, which is several blocks away from a major U.S.-Mexico border
 4     crossing and the route leading to it; nine nonoverlapping exceedances were recorded in 1996.
 5     Reportedly, there are often long lines of idling vehicles waiting to cross the border, including
 6     vehicles of Mexican registration that are not equipped with the emission control equipment
 7     required on vehicles sold in the United States.  Such situations will need to be addressed on a
 8     local, case-by-case basis.
 9
10
11     3.5   SOURCES, EMISSIONS, AND CONCENTRATIONS OF CARBON
12           MONOXIDE IN INDOOR ENVIRONMENTS
13          The general U.S. population spends up to 85% of its time indoors.  In recent years, more
14     emphasis has been placed on the evaluation of pollutant sources, emissions, and concentrations
15     in indoor environments to aid in the evaluation of total human exposure. It is particularly
16     important to evaluate carbon monoxide concentrations in indoor environments because indoor
17     exposure may represent a significant portion of the total CO exposure.
18          The proceeding sections focus on the sources and emission rates of CO in indoor
19     environments. Concentrations of CO in various indoor environments also will be discussed.
20     Emphasis is placed on the evaluation of those currently accepted uses of combustion appliances
21     and consumer products and the resulting CO emissions and concentrations. Accidental or
22     unintentional sources and concentrations will be mentioned only briefly.
23
24     3.5.1  Unvented Combustion Sources and Estimated Emissions Rates
25          Carbon monoxide occurs in indoor environments directly through emissions from various
26     indoor combustion sources or indirectly as a result of infiltration or ventilation from outdoor
27     sources. Unvented, partially vented, and improperly vented combustion appliances and
28     consumer products represent the primary sources of CO emissions in the indoor environment.
29     Table 3-7 lists the various sources of CO in the indoor environment. Emissions of CO from use
30     of combustion appliances will depend on several factors.  These factors include the source (e.g.,

       February 15,  1999                         3-28        DRAFT-DO NOT QUOTE OR CITE

-------
                            TABLE 3-7. SOURCES OF CARBON MONOXIDE
                                    IN THE INDOOR ENVIRONMENT
         Source
                                                         Comments
         Outdoor (ambient air)
         Gas cooking ranges
         Gas space heaters
         Kerosene space heaters
         Environmental tobacco smoke
         Fireplaces and woodstoves
         Gas furnaces, clothes dryers,
         and water heaters

         Motor vehicles
                             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.

                             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.

                             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.

                             Emissions vary based on unit type (convective, radiant, or two-stage), operating
                             condition, and duration of use.

                             The majority of CO entering indoor compartments from the combustion of
                             tobacco products is through sidestream smoke.

                             Carbon monoxide is emitted during fire start-ups, leaks in stoves and pipes, and
                             during backdrafting resulting from depressurization.

                             Gas furnaces and dryers generally are vented and do not emit CO in the indoor
                             environment unless the unit is malfunctioning.

                             Operating motor vehicles in enclosed spaces can be significant sources of CO
                             in indoor environments.
 1

 2

 3

 4

 5

 6

 7

 8

 9

10
gas cooking stoves, unvented space heaters, woodstoves, fireplaces), appliance design,

manufacturer, type of fuel used, fuel consumption rate, use pattern, and operating condition.

      Two different approaches are used to evaluate CO emissions for combustion appliances:

the direct or sampling-hood approach and the mass-balance/chamber approach. For details on

these two approaches, see U.S. Environmental Protection Agency (1991b).


3.5.1.1 Gas Cooking Ranges  and Ovens and Furnaces

      Emissions of CO from gas top ranges will depend of the use pattern, operating condition,

fuel consumption rate, and air infiltration into the microenvironment.  Average annual household

fuel consumption has been estimated at 5,000 ft3 for ranges with standing pilots (Johnson et al.,
        February 15, 1999
                                             3-29
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 9
10
11
1992). Menkedick et al. (1993) reported annual household fuel consumption of 2,180 ft3
(±890 ft3) for burners, based on actual fuel consumption measurements taken on 103 gas ranges
individually metered over a 2-year period in Illinois. Fuel consumption for burner and standing
pilots was 5,710 ft3 (±1,830 ft3).  The average fuel consumption also was affected by the age of
the occupants (older adults used the range more frequently for preparing meals than did young
adults) and the presence or absence of a standing pilot, and it showed a seasonal trend.
Figure 3-7 illustrates the seasonal trend for daily gas stove fuel consumption. Carbon monoxide
emissions from gas ranges (burners and ovens) are summarized in Table 3-8 as a function of
operating conditions.
            8.0
            7.5-
         CD
        £
         o
        !Q
         CD
            6.5-
         
-------
                TABLE 3-8. RANGES IN AVERAGE CARBON MONOXIDE
                     EMISSION RATES FOR RESIDENTIAL SOURCES
                   Fuel    Fuel Consumption
 Unit Type        Type"    Rate (kJ/minb)
                          Flame
                          Ranges in Average
                        Emission Rates (,ug/kJ)
                                         Source
 Gas Ranges0

   Top burners
NG
               Blue
   Ovens
NG
   Burner pilots     NG


   Oven pilots       NG


 Gas Space Heaters

   Convective       NG
           131-784
                                            Yellow-
                                            tipping
               Blue
                         Yellow-
                         tipping

                           Blue
                           Blue
               Blue
   Infrared
NG
   Catalytic
NG
                              351-703
260-368
                              258-264
  207
                           Blue
Infrared
                         Infrared
 Blue
                 16-200
                                          92-197
                 16-622
                                                               62
                               28-56
                              209-322
                 3-192
                                 16
5-69
                                4-45
9-14
              Himmel and DeWerth
                 (1974)
              Traynoretal. (1982)
              Borrazzo et al. (1987)
              Cote etal. (1974)
              Moschandreas et al. (1985)
              Spicer and Billick (1996)
              Billicketal. (1984)

              Himmel and DeWerth
                 (1974)
              Cote etal. (1974)
              Moschandreas et al. (1985)

              Himmel and DeWerth
                 (1974)
              Traynoretal. (1982)
              Borrazzo et al. (1987)

              Himmel and DeWerth
                 (1974)

              Himmel and DeWerth
                 (1974)

              Himmel and DeWerth
                 (1974)
Traynoretal. (1984, 1985)
Moschandreas et al. (1985)
Thrasher and DeWerth
  (1979)
Zawackietal. (1984)
Hednck and Krug (1995)
Spicer and Billick (1996)
Zawackietal. (1986)

Traynoretal. (1984, 1985)
Zawackietal. (1986)

Traynoretal. (1984, 1985)
Moschandreas et al. (1985)
Spicer and Billick (1996)
Zawackietal. (1986)

Traynoretal. (1984, 1985)
Zawackietal. (1986)

Moschandreas et al. (1985)
February 15, 1999
                             3-31
                                DRAFT-DO NOT QUOTE OR CITE

-------
                 TABLE 3-8 (cont'd). RANGES IN AVERAGE CARBON MONOXIDE
                           EMISSION RATES FOR RESIDENTIAL SOURCES

                      Fuel   Fuel Consumption            Ranges in Average
        Unit Type      Type3     Rate (kJ/minb)     Flame   Emission Rates Cug/kJ)             Source
        Kerosene Heaters

          Convective
          Radiant
          Two-stage

        Furnaces
               NG
        Woodstoves     —
          and
          Fireplaces

        Tobacco Smoke

          Cigarette
          Cigar
                           37-193
                           85-168
                          132-182
    4-272
   43-264
    54-9

  6-> 1,000

2.0-1,748.6 g/h
                                                O.5-78 mg/cigarette
                                                   130-200 mg/gd
Leaderer(1982)
Traynoretal. (1983)
Moschandreas et al. (1985)
Traynoretal. (1990)
Mumfordetal. (1991)

Leaderer(1982)
Traynoretal. (1983)
Moschandreas et al. (1985)
Traynoretal. (1990)
Mumfordetal. (1991)

Traynoretal. (1983)

Ryan and McCrillis (1994)

Traynoretal. (1987)
Mueller Associates (1985)
Jaasmaetal. (1995)
                  National Research Council
                    (1986)
                  Klepeisetal. (1995, 1996)
                  Ottetal. (1992)
                  Lofrothetal. (1989)

                  Klepeisetal. (1999)
       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.
       dEmission rate based on cigar mass of 4.9 to 8.8 g.
1

2

3

4

5

6
was an older model with an energy efficiency of 60 to 70% and the other was a newer furnace
with an energy efficiency of 94%. The furnaces were operated for 10 min then allowed to cool

for 5 to 10 min.  The cycle was repeated 12 to 18 times during the course of each test. The
CO emission rate was >1,000 //g/kj for the older unit, compared with 6 //g/kJ for the newer,

more efficient model.
       February 15, 1999
                                              3-32
        DRAFT-DO NOT QUOTE OR CITE

-------
 1      3.5.1.2 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           Tables 7-5 and 7-6 in the previous AQCD for CO (U.S. Environmental Protection Agency,
 8      1991b) summarize the available emissions data from earlier studies on gas and kerosene space
 9      heaters.  The available data show variability in CO emissions based on unit design, maintenance,
10      and the type of fuel used. Several recent studies on CO emissions evaluated emissions from
11      prototype units; a brief discussion of the results of these studies follows.
12           The Gas Research Institute evaluated the emissions of CO, NO, NO2, and unburned
13      hydrocarbons (UHC) from a vented and an unvented prototype space heater.  The heater design
14      was a Pyrocore ceramic fiber radiant burner. Emission goals were  5.5 //g/kJ (20 ppm) CO,
15      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
16      had cross-flow room air circulation fans. The samples collected from the unvented  space heater
17      were corrected for dilution effects. Emission rates were estimated using a single-equation
18      mass-balance model. Carbon monoxide emissions ranged from 0.6 to >27.5 //g/kJ (2.0 to
19      >100 ppm), based on the amount of excess air and the firing rate values (Duret and Tidball,
20      1990).
21           Apte and Traynor (1993) determined the  emission rates of combustion pollutants for a
22      radiant-fiber-matrix gas burner prototype.  Fuel consumption ranged from 333 to 527 kJ/min.
23      Carbon monoxide emission rates were generally low, with an average of 3.0 //g/kJ when the unit
24      was operated with 10% excess air, and 7.4 //g/kJ with 40% excess air.
25           The Institute of Gas Technology, along with Maxon Corporation, designed and tested an air
26      heater based on the cyclonic combustion concept. This technology included premixed high-
27      excess air, and cyclonic combustion  with flame stabilization, in conjunction with optimized
28      nozzle velocity control.  Carbon monoxide emissions were 1.6 to 5.9 //g/kJ (1.0 to 3.6 ppm at
29      15% O2) (Xiong et al.,  1991).  Carbon monoxide emissions from unvented space heaters are
30      summarized in Table 3-8 as a function of unit type, fuel type and consumption rate,  and
31      appliance operating condition.

        February 15, 1999                         3-33        DRAFT-DO NOT QUOTE OR CITE

-------
 1      3.5.1.3 Woodstoves and Fireplaces
 2           A few studies have evaluated the emission of CO from woodstove or fireplace use. Carbon
 3      monoxide may enter the indoor environment during fire start-up and tending and through leaks in
 4      the stove or venting system. Carbon monoxide emissions are higher during the first stage of a
 5      fire because of increasing amounts of fuel being burned and inadequate temperature conditions.
 6      Such intermittent emissions makes it difficult to accurately determine CO emission rates. In an
 7      early study by Traynor et al. (1987), the average CO source strength for three airtight stoves was
 8      found to be 0.13 to 0.175 g/h (104 to 140 cm3/h), whereas the average source strengths for a
 9      nonairtight stove ranged from 0.275 to 2.25 g/h (220 to  1,800 cm3/h). Mueller Associates (1985)
10      reported CO emission ranges of 0.07 to 0.375 g/h for wood heaters. Jaasma et al. (1995)
11      conducted a study designed to evaluate the effectiveness of custom-built glass doors for
12      fireplaces in reducing CO emissions under conditions of negative pressure. The glass doors
13      decreased spillage of CO; however, decreasing the leakiness of the glass doors did not always
14      reduce CO spillage. Tests with the glass doors closed had CO emission rates of 2 to 36 g/h
15      (highest levels represented leaking glass doors).  Carbon monoxide emissions on the order of
16      70 g/h were noted for glass- door-opened tests under negative pressure.
17           Carbon monoxide also may enter the indoor environment through backdrafting when the
18      natural draft is overcome by depressurization. Depressurization generally occurs during fire
19      start-up, but also may occur during operation of other equipment such as kitchen and bathroom
20      exhaust fans and clothes dryers.  Nagda et al. (1996) summarized the results of several studies on
21      emissions of pollutants into living compartments as a result of house depressurization. Carbon
22      monoxide emissions were found to be insignificant. Tiegs and Bighouse (1994) evaluated CO
23      spillage from a woodstove under chamber and in-house conditions. They reported CO leakage
24      into the indoor environment by nonairtight woodstoves during conditions of negative pressure.
25
26      3.5.1.4 Environmental Tobacco Smoke
27           Carbon monoxide emissions from the combustion of tobacco occurs in the indoor
28      environment when smokers exhale the inhaled or mainstream smoke and from the emission of
29      sidestream smoke from smoldering tobacco products.  The majority of the CO  comes from the
30      sidestream smoke.  The amount of CO emitted will vary based  on the type of tobacco product


        February 15, 1999                         3-34       DRAFT-DO NOT QUOTE OR CITE

-------
 1      (cigarette/cigar), the degree to which tobacco is actively smoked, and the amount of smoke being
 2      absorbed by the lungs (Klepeis et al., 1996; Akbar-Khanzadeh and Greco, 1996).
 3           The National Academy of Science report on environmental tobacco smoke estimated a CO
 4      emission range of 40 to 67 mg per cigarette (National Research Council, 1986).  The Federal
 5      Trade Commission compiled data on 933 varieties of cigarette manufactured and sold in the
 6      United States in  1992. These data were provided by the various cigarette manufacturers.  Carbon
 7      monoxide emission rates for the brands of cigarettes reported ranged from <0.5 to 23.0 mg per
 8      cigarette (cigarettes emitting 23.0 mg were unfiltered brands) (Federal Trade Commission, 1994).
 9      Klepeis et al. (1995, 1996) measured CO concentrations in airport smoking lounges under real
10      life conditions. They estimated CO emissions to be 78 mg per cigarette on the basis of an
11      average CO emission rate of  11.1 mg/min and a smoking duration of 7  min. An estimated total
12      CO emission rate of 81.2 mg for three cigarettes was reported by Ott et al. (1992).  Lofroth et al.
13      (1989) estimated an emission rate of 67 mg per cigarette based on a cigarette weight of ~ 1.2 g
14      and a smoking duration of 12 min.  Large cigars emit substantially more CO than do cigarettes.
15      Emission rates of 130 to 200  mg CO/g (mass smoked) were reported by Klepeis et al. (1999)
16      (smoked by machine and by a person). Cigar mass ranged from 4.9 to 8.8 g, and the smoking
17      time was 7.8 to 24 min for the machine-smoked test and 76 min for the test measuring emissions
18      for a cigar being smoked by a person.
19
20      3.5.2 Indoor  Concentrations of Carbon Monoxide
21      3.5.2.1  Factors  Affecting Carbon Monoxide Concentrations
22           A number of factors can affect indoor CO concentrations:  the presence of a source and its
23      use pattern, pollutant emission rate, ambient air concentrations, infiltration through the building
24      envelope, air exchange rate (AER), building volume, air mixing within the indoor compartments,
25      and the presence and effectiveness of a pollutant removal  system.
26           The major  sources of CO in residential environments are unvented gas or kerosene
27      appliances.  Because gas cooking ranges are used intermittently, it is not likely that the use of gas
28      ranges would result in substantial increases in CO over long periods of time, except in
29      households where gas cooking stoves are used improperly as a primary or secondary source of
30      heat.  Also, the replacement of gas ranges with standing pilots lights with ranges with electronic
31      pilots likely will  have an influence on the CO concentrations from gas range usage. Koontz et al.
        February 15, 1999                         3-35        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (1992) reported the results of a survey conducted in 1985 and 1991. The purposes of the survey
 2      were to determine the prevalence of kitchen fans and the factors affecting their use and to
 3      determine the impact of other cooking appliances (i.e., microwave ovens, toaster ovens, hot
 4      plates) on the use of the range for cooking.  The authors reported a 35% increase in the use of gas
 5      ranges without standing pilot lights between 1985 and 1991 and a 20% reduction in the use of
 6      stoves for cooking. Ninety-five percent of the households surveyed reported having another form
 7      of cooking appliance in addition to the gas range, and, of this number, 55 to 65% reported using
 8      the stove less often. There were, however, more people using a gas range for purposes of
 9      supplemental heating than there were using electric ranges for that purpose (11% versus 3.6%).
10      Estimates of gas cooking stove usage range from 30 to 60 min/day. Table 3-9 contains
11      information on the prevalence of gas cooking stoves in the United States.
12          Data from the National Health and Nutrition Examination Survey estimated that
13      13.7 million adults used unvented combustion space heaters between 1988 and 1994.  Based on
14      the information obtained in the survey, an estimated 13.2% of the adult population in the
15      southern United States used unvented combustion space heaters. An estimated 5.9% of the adult
16      population in the Midwest, 4.2% in the Northeast, and 2.5% in the West used unvented space
17      heaters (Figure 3-8) (Slack and Heumann, 1997).  The use of more combustion space heaters in
18      the South also was reported in earlier studies by U.S. Department of Housing and Urban
19      Development (1987), U.S. Environmental Protection Agency (1992), and Williams et al. (1992).
20      This may be because, in areas with relatively mild winters, combustion space heaters are used
21      frequently as the primary source of heat. The U.S. Environmental Protection Agency (1990)
22      estimated that kerosene heaters are used 16.7 h/day in southern  states and, in regions where the
23      heaters are used as secondary heat sources, estimated a use range of 2.6 to 10.7 h.
24          The AER, the balance of the flow  of air in and out of a microenvironment, is based on the
25      fraction of air that enters the microenvironment through infiltration through unintentional
26      openings in the building envelope, natural ventilation through any designed opening in the
27      building envelope (doors, windows), and forced ventilation systems.  Infiltration is the dominant
28      mechanism for residential air exchange. Forced ventilation is typically the dominant mechanism
29      for air exchange in nonresidential buildings. Natural ventilation, airflow through doors and
30      opened windows, is seasonal (Koontz and Rector, 1995). Air exchange rates for residential


        February 15, 1999                         3-36        DRAFT-DO NOT QUOTE OR CITE

-------
TABLE 3-9. PREVALENCE OF GAS COOKING RANGES
w
i
^2
^
VO
VO
VO








OJ
3


o

^
H
1
O
0
o
H
0
0
H
W
O
O
H
W


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
(1982-1993)

Denver, CO
(1982-1983)














Number of
Households
(total)
959
(2,323)


33,813,000
(94,363,000)
150
(293)
2,614,000
(3,165,200)
151,300
(770,600)
609,029
(953,714)

85,542
(345,163)















Percentage of
Households Comments Reference
4 1 Twenty percent of households had electronic pilots Koontz et al. ( 1 992)
in 1985. Twenty-seven percent of the households
had electronic pilots in 1991, based on
886 respondents from the 1985 survey.
36 U.S. Census Bureau
(1998)
52 Twenty-five percent of households had electronic Wilson et al. (1993)
pilots; 3% had combination pilots.
82 U.S. Census Bureau
(1998)
20

64 Results are based on a survey of 1987 participants in Johnson (1984)
Washington and 1,139 participants in Denver. Hartwell et al. (1984)

25
















-------
           6
           5-
      §>   4H
      03
            o-
                      Kerosene
                      Propane
                      Natural Gas
                   Northeast     Midwest
                                                South
                 West
      Figure 3-8. 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

2

3

4
buildings are in Table 3-10. Air exchange rates vary depending on the outside temperature,

geographical location, type of cooking fuel used, type of heating system used, and building type

(Colome et al., 1994).
      February 15, 1999
                                     3-38
DRAFT-DO NOT QUOTE OR CITE

-------
                     TABLE 3-10. 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
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 et al.
(1992)
Wilson et al.
(1996)









Murray and
Burmaster (1995)




Colome et al.
(1994)
Research Triangle
Institute (1990)
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
4     2.45 (2.24), 1.35 (1.09), and 2.22 (1.79)/h for schools, offices, and retail establishments,
      February 15, 1999                         3-39        DRAFT-DO NOT QUOTE OR CITE

-------
 1      respectively. Air infiltration rates for 40 of the 49 buildings were 0.32, 0.31, 1 . 12/h for schools,
 2      offices, and retail establishments, respectively. Air exchange rates for 40 nonresidential
 3      buildings in Oregon and Washington (Turk et al., 1989) averaged 1.5/h (mean [median = 1.3/h]).
 4      Park et al. (1998) reported AERs for three stationary cars (cars varied by age) under different
 5      ventilation conditions.  Air exchange rates ranged from 1.0 to 3.0/h for windows closed and fan
 6      off, 13.3 to 23.5/h for window opened and fan off, 1.8 to 3.7/h for window closed and fan on
 7      recirculation (two cars tested), and 36.2 to 47.5/h for windows closed and fan on fresh air (one
 8      car tested).  An average AER of 13.1/h was reported by Ott et al. (1992) for a station wagon
 9      moving at 20 mph with the windows closed.
10
1 1      3.5.2.2 Models for Carbon Monoxide Concentrations
12           The concentration of CO in indoor environments will depend on the source emission rate,
13      source use characteristics, building characteristics, outdoor concentrations and the rate of
14      penetration into indoor compartments, distribution between compartments, and the pollutant
15      removal rate.
16           Indoor concentrations of CO can be estimated using the mass-balance model. The
17      mass-balance model estimates the concentration of a pollutant over time. The simplest form of
18      the model is represented by the following differential equation for a perfectly mixed
19      microenvironment.
                                                                       -    IN
                                                                            ~
                                   B
20
21      where Cm is the indoor pollutant concentration in mass per unit volume, FB is the deposition
22      fraction of outdoor pollutant infiltration through the building envelope, v is the AER in I/time,
23      COUT is the outdoor pollutant concentration in mass/volume, S is the indoor pollutant source
24      emission rate in mass/time, c is the effective indoor volume fraction and is assumed to equal 1,
25      V is the volume of microenvironment (national mean residential volume is 380 m3; Murray and
26      Burmaster,  1995), Fd is the coefficient for decay in I/time (zero for CO), q is the flow rate into
27      the air cleaning device in volume/time, and F is the efficiency of the air cleaning device.
        February 15, 1999                         3-40        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Equation 3-1 can be simplified because the deposition and decay rates for CO are negligible, and
 2      assuming the absence of an air cleaning device, the equation becomes
 3
                                   dCIN             S
                                   -^- = vCOUT + — - vCIN.                           (3-2)
 4
 5      A more in-depth discussion of the mass-balance model may be found in U.S. Environmental
 6      Protection Agency (1991b) and Nagda et al. (1987).
 7           Traynor et al. (1989) used a macromodel to predict CO concentrations in residential
 8      environments for one pollutant source. Model inputs included ambient air concentrations, source
 9      emission rates and usage characteristics, compartment volume, AERs, and outside temperatures.
10      The macromodel combined the steady-state version of the mass-balance model used in indoor air
11      quality studies, a source-usage model for space heating appliances, and an air exchange model.
12      A combination of the Monte  Carlo and deterministic techniques was used to predict indoor
13      concentration distributions.
14           The mass-balance model for CO is expressed as
15
16                                    CIN=COUT+S/vV.                               (3-3)
17
18      The source-strength model was expressed as
19
20                                          S = QEFV,                                    (3-4)
21
22      where Q is the source usage rate,  same as house heating requirements for space heaters, in
23      kiloJoules per hour or cigarettes per hour; E is the source emission rate in micrograms per
24      kiloJoule or micrograms per  cigarette; and Fv is the appliance venting factor (unitless; 1 = 100%
25      unvented).  The space heater source usage rate (Q) is based on the heating requirements of the
26      heater and is equal to the heat loss through conduction and infiltration minus the "free" heat,
27      modified by the unit efficiency and the occupant life-style factor (under- or overheating and
28      heating only certain rooms).  The source usage rate was expressed as

        February 15, 1999                         3-41        DRAFT-DO NOT QUOTE OR CITE

-------
 1                                 Q = -(UAAT + vVqAT-Qf),                            (3-5)

 2
 3      where b is the life-style factor (unitless), e is the appliance efficiency (unitless; 1 = 100%
 4      efficient), U is the building envelope thermal conductance in kiloJoules per hour per square
 5      meter per degree centrigrade, UAAT accounts  for the conduction losses through the building
 6      envelope, A is the total house surface area including ceiling and floor in square meters, AT is the
 7      indoor/outdoor temperature difference in degrees centigrade, V is the house volume in cubic
 8      meters, q is the heat content of dry air (1.2 kJ/m3 °C), and Qf is the house free heat in kiloJoules
 9      per hour.  House free heat includes solar heat gain, internal-appliance heat, and human generated
10      heat; it was expressed as
11
12                                  Qf = Qs + Qappl +Qpeople,                              (3-6)
13
14      where Qs is the solar heat gain in kiloJoules per hour, Qappl is the interior heat generated by
15      appliances (the American  Society of Heating, Refrigerating, and Air Conditioning Engineers
16      standard is 2,000 kJ/h), and Qpeople is the interior heat generated by people (estimated to be
17      1,000 kJ/h for a house with two adults and two children). The air exchange model included an
18      additional term to account for increased AERs because of the use of woodstoves and was
19      expressed as
20
91                          v = (ELA2fs2AT + ELA2fwV + Z2appl) °'5
                                                  V

22      and
23                                          ELA = SLAxAf,
24
25      where ELA is the effective leakage area in square meters, fs is the reference stack parameter for
26      the infiltration model (430 m/h °C05), fw is the reference wind parameter for the infiltration
27      model (120,000 m2/km2), u is the wind speed in kilometers per hour, Zappl is the woodstove-flue

        February 15, 1999                          3-42       DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
flow rate (76 m3/h), SLA is the specific leakage area (unitless), and Af is the floor surface area in
square meters. Based on the modeled results, the use of kerosene heaters, unvented gas space
heaters, and gas ovens and ranges for heating produced the highest concentrations of CO in the
indoor environment (see Figure 3-9).  The model input parameters are given in Table 3-11.
            0.3
                            -e-
                                    , Gas Oven/Range
                                   ~" With Hood
                                       •  Gas Oven/Range
                                       ~"  Without Hood
                                                                                   1% 10% GM 90% £
                    i—i—e-
                          H Smoking
                 I—|—0—|—| Gas Domestic Hot Water
                 I—I—0—I—I Gas Dryer
                                           -e-
                                                              .  Radiant
                                                                Kerosene Heater
                                           -e-
                                  i—i—
                              H	e-
                                     -e-
                Infrared Unvented
                Gas Space Heater
                                                                       • Gas Oven/Range
                                                                        for Heating
                                                                     . Convective
                                                                      Kerosene Heater
                                                      H
                      I	h
                       -e-
                      i—i-
                      -e-
—I Nonairtight Woodstove
H Airtight Woodstove
                   Convective Unvented
                   Gas Space Heater
                 i—i—e-+
                                    H Gas Wall/Floor Furnace
                                    	1 Gas Boiler
                i—i—e
                 H—e—i-
                                 H Gas Forced-Air Furnace
                                 H Oil Forced-Air Furnace
                 I—I—©—I—I Electric
                                                  10
                                            CO Concentration (ppm)
                                                                           100
                                                                                              600
       Figure 3-9.  Modeled indoor CO concentration distributions in houses with only one
                    indoor combustion pollutant source.
       Source: Traynoretal. (1989).
1      3.5.2.3 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.
       February 15, 1999
                                             3-43
                    DRAFT-DO NOT QUOTE OR CITE

-------
                                    TABLE 3-11.  INPUT PARAMETERS FOR CARBON MONOXIDE
        Input Parameter
                                                    Data
                                                    Points
Geometric
  Mean
    Geometric
Standard Deviation
    Empirical
Distribution Range
VO
VO
VO
HEATING APPLIANCE EFFICIENCY (%)
  Gas forced-air furnace (FAF)
  Gas boiler
  Gas wall/floor (W/F) furnace
  Oil FAF
  Airtight Woodstove
  Nonairtight Woodstove

SMOKING FREQUENCY (cigarettes/h)
                                                                                                   1.5
                                             55-95
                                             60-95
                                             75-85
                                             70-90
                                             45-65
                                             35-45




OJ
-U
"^


o

£
H
6
o
!2
O
H
O

0
H
W
O
O
HH
H
W
POLLUTANT EMISSION RATES CwgM or
A(g/cigarette)
Gas FAF
Gas boiler
Gas W/F
Unvented gas space heater (UVGSH) infrared
UVGSH convective
Gas stove top
Gas stove oven
Gas dryer
Gas water heater
Oil FAF
Wood airtight
Wood nonairtight
Kerosene radiant
Kerosene convective
Cigarettes

NONZERO VENT FACTORS FOR VENTED
GAS/OIL SPACE HEATERS


RANGE HOOD VENT FACTOR (unitless)
POLLUTANT PENETRATION FACTORS
(unitless)



11
32
—
4
12
53
56
2
55
30
12
4
7
4
6

5



1
—




11.4
4.4
8.4
4.6
27.9
81.3
38.6
52.0
5.8
6.3
3.9
3.2
67.9
42.1
714.0






1.0




6.4
1.8
2.3
1.1
2.4
3.1
4.1
1.5
1.9
11.7
2.1
2.0
1.4
3.4
1.3

6.8%



0.3




-------
TABLE 3-11 (cont'd). INPUT PARAMETERS FOR CARBON MONOXIDE
w
i
(^
VO
VO
VO





OJ
^yi

O
s
H
6
0
o
H
0
0
H
W
O
O
H
W
Data
Input Parameter Points
MARKET PENETRATION OF APPLIANCE (%)
PRIMARY HEATING APPLIANCE DISTRIBUTION
(%)
NONHEATING APPLIANCE DISTRIBUTION (%)
NON-SPACE-HEATING APPLIANCE FUEL —
USAGE (kJ/h)
HOUSE VOLUME (m3) 48
OUTDOOR POLLUTANT CONCENTRATION (ppm) 2
BUILDING U-VALUES (kJ/h m2 °C)
Walls and ceilings —
Floors —
Windows and storm windows —
Doors and storm doors —
OUTDOOR TEMPERATURE ( ° C) —
WIND SPEED (m/s) —
SOLAR GAESP (kJ/h m2) —
NUMBER OF STORIES —
SPECIFIC LEAKAGE AREA (10'4 m2/m2) 50
aRASS sample size = 350 responses out of 2,074 possible (17%).
bData details found in Traynor et al. (1989).
°Range of solar gain in houses with and without storm windows.
Source: Traynor etal. (1989).

Geometric Geometric Empirical
Mean Standard Deviation Distribution Range
b
b
b
b

325-775
0.7 1.2
b
b
b
b


-9-25
4.4-16.2
39-455
1-2
2.84 1.44




-------
 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 determine the highest ranked homes. Using
10      that technique, 21 of the 293 homes studied were selected for case studies. The higher CO seen
11      in these homes possibly was associated with occupant smoking, the use of gas stoves for heating
12      purposes, infiltration from attached garages, the type of heating system used (homes with gas
13      wall furnaces had higher CO), the building type and size (smaller multi-family homes had higher
14      CO than larger single-family homes), use of gas appliances, and more than one CO source.  The
15      average AER varied by type of heating system (wall furnaces > forced-air > electric) and building
16      type (multi-family units > single-family units) (Billick et al., 1994, 1996; Colome et al.,  1994).
17      The CO descriptive 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. The average room volume was assumed to be
20      50 m3. The average AER was 0.59 h"1. Carbon monoxide monitors were placed in the primary
21      living space and close to the source. Approximately half of the homes used gas cooking stoves.
22      Kerosene heaters had to be operated at least 3 h/day to qualify as a CO source, and the woodstove
23      or fireplace had to be operated an average of 2 h/day. Any reported usage  of gas stoves qualified
24      them as sources. The average CO concentration in the  primary living area was 2.23 ± 0.17 ppm
25      (results for 209 homes). Use of both gas stoves and kerosene space heaters was associated with
26      increased CO. Homes using woodstoves  or fireplaces had  lower CO than did homes without
27      woodstoves or fireplaces (see Figure 3-10).
28           Hedrick and Krug (1995) reported CO concentrations from the use of unvented gas space
29      heaters and pilot lights in a test house in Chicago, IL. The  gas space heaters included blue-flame
30      convective, radiant-tile, fan-forced blue-flame and perforated-tube convective units.  Emission


        February 15, 1999                          3-46        DRAFT-DO NOT QUOTE OR CITE

-------
TABLE 3-12. CARBON MONOXIDE DESCRIPTIVE STATISTICS FOR ALL HOMES (number = 277; in ppm)
2
VO
VO
VO




OJ
rv
^J

O
H
6
o
o
H
0
0
H
W
0
O
H
w
All Homes

Arithmetic Mean
Standard Error
Mode
Percentiles
Maximum
95th
75th
50th
25th
5th
Minimum


48-h
1.6
0.1
1.0

12.9
4.3
1.8
1.2
0.7
0.1
0.0
Source: Modified from Wilson et al











Max
10-min
5.2
0.3
2.0

37.9
15.1
6.6
3.5
2.0
1.0
0.1
. (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 Max 8-h
1.0 5.5 4.3 3.8 2.0
0.1 0.4 0.3 0.2 0.1
0.1 2.0 2.0 1.0 1.0

10.8 68.7 31.5 27.3 17.3
2.7 16.1 12.3 10.6 6.3
1.3 6.1 5.2 4.8 2.2
0.8 3.3 2.9 2.6 1.4
0.3 2.0 1.9 1.5 0.9
0.1 1.1 1.0 1.0 0.3
0.0 0.2 0.1 0.0 0.0







-------
       Q.
       a.
       O
       O
       O
       '•4-<
       2
       -§-<
       I
       o
       O
            4.0
      3.0 -
      2.0 -
             1.0 -
                                                                         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 3-10. Arithmetic mean CO concentrations by presence or absence of combustion
                   source.
       Source: Research Triangle Institute (1990).
1
2
3
4
5
6
rates for the units are discussed in Section 3.5.1.2. The house was a single-family, one-story,
three-bedroom dwelling with a full basement, comprising 1,150 ft2 per level.  Eight burner
experiments and three pilot light experiments were conducted. Heaters were operated for
8 h followed by a 15-h decay period. The pilot studies were conducted over a 48-h period. Fans
were used to distribute emissions throughout the house, excluding the basement. The highest CO
concentrations were seen with the radiant-tile heater (13.4 ppm), and the lowest CO
      February 15, 1999
                                          3-48
 DRAFT-DO NOT QUOTE OR CITE

-------
 1      concentrations were reported with the fan-forced unit (0.9 ppm). Two of the tests using the
 2      blue-flame convective units were affected by gas leakage, resulting in CO concentrations of
 3      4.7 and 4.8 ppm. The test not affected by the leakage had a CO concentration of 2.7 ppm.  The
 4      maximum CO concentration in the test house from use of the perforated-tube convective heater
 5      was 31.9 ppm.  Carbon monoxide concentrations during the pilot light experiments did not
 6      exceed 2.0 ppm.
 7          Koontz and Nagda (1988) reported CO concentrations in 82 homes using unvented gas
 8      space heaters as the primary heat source and in 29 homes using unvented gas space heaters as a
 9      secondary heat source. Forty-one of the homes monitored did not use a gas space heater.  Carbon
10      monoxide was monitored continuously (consisting of sequential  15-min averages) close to the
11      geometric center of the house over an average period of 5 days.  The highest CO concentrations
12      were seen in homes using the space heaters as the primary heat source (highest concentration was
13      36.6 ppm).  Carbon monoxide concentrations did not exceed 9.0 ppm in homes using space
14      heaters as a secondary heat source or in homes without gas space heaters.  Table 7-13 of the
15      previous AQCD for CO presents a comparison of the CO concentrations in the homes studied
16      (U.S. Environmental Protection Agency, 1991b).
17          Moderately high concentrations of CO have been reported in homes where unvented
18      kerosene heaters are in use.  Burton et al. (1990) monitored both the inside and outside of two
19      mobile homes for pollutants emitted from the operation of a radiant and a convective unvented
20      kerosene space heater. No other sources of combustion by-products were in the homes. Heaters
21      were operated from 4:00 to 9:00 p.m. daily.  Six random sampling periods were conducted, three
22      with heaters on and three with heaters off. Measurements were made until 11:00 p.m.  Average
23      CO concentrations while the heaters were in operation were 12 ppm for the convective heater and
24      4 ppm for the radiant heater.  When the heaters were not in use, average CO concentrations were
25      5 and 1 ppm, respectively.  Ambient CO concentrations in the mobile home park were reported
26      to be negligible.
27          Carbon monoxide concentrations in eight single-wide mobile homes (150 to 255 m3) were
28      reported by Mumford et al. (1990, 1991).  Convective  kerosene heaters were used in four of the
29      homes. Three of the homes used radiant heaters, and one home used a convective/radiant heater.
30      Monitoring was conducted 2.6 to 9.5 h/day (average 6.5 h/day) for 2 weeks with heaters on and
31      2 weeks with heaters off. Fuel consumption rates ranged from 252 to 295 kJ/min for convective

        February 15, 1999                          3-49       DRAFT-DO NOT QUOTE OR CITE

-------
 1     units, 105 to 168 kJ/min for radiant units, and 120 kJ/min for the convective/radiant unit.
 2     Average AERs were 0.47 h"1 with heater on and 0.48  h"1 with heater off.  Monitoring began
 3     when heater use began and continued for 2 h after the heaters were turned off. Carbon
 4     monoxide concentrations were above 9 ppm in four of the eight homes. In one home with a
 5     convective heater, CO concentrations peaked at 51 ppm. The average CO concentration with
 6     heaters off was 1.4 ppm.
 7          Williams et al. (1992) reported CO concentrations in eight all-electric mobile homes (each
 8     <100 m2) from the use of kerosene heaters over a 6-day  measurement period.  The space heaters
 9     were used for an average of 4.5 h/day between 4:00 to 11:00 p.m. Background CO ranged from
10     0 to 8 ppm. Average CO during nonuse days was 1.4 ppm ± 0.3. Peak CO values ranged from
11     0.3 to 50.2 ppm.  The 8-h average CO concentration in the homes was  7.4 ±1.4 ppm. Peaks
12     usually were observed at the end of the combustion period. The AER averaged 0.47 h"1 when the
13     unit was in operation.  Homes with the radiant and multistage units had higher CO than homes
14     with convective heaters. Average 1-h levels of CO for convective units ranged from 1.3 to
15     5.3 ppm. One-hour average CO concentrations  for radiant and multistage units ranged from
16     1.1 to 28.3 ppm and 16.0 to 50.0 ppm, respectively.
17
18     Carbon Monoxide Concentrations Related to Environmental Tobacco Smoke
19          Carbon monoxide concentrations in environments where smoking occurs exceed
20     background CO concentrations. All the concentrations will depend on the size of the indoor
21     space, number of cigarettes smoked, smoking rate, CO emission rate, differences in ventilation,
22     and the ambient CO concentrations (Turner et al., 1992). Ott et al. (1992, 1995) conducted a
23     series of monitoring experiments in a one-story  house during development of a
24     multi-compartment indoor mass-balance model  to predict the pollutant concentrations from
25     environmental tobacco smoke. Smoking time ranged from 6.5 to 9.5 min. Carbon monoxide
26     concentrations were measured in three locations in the bedroom after three cigarettes were
27     smoked over a 9-h period. Concentrations ranged from  0.4 to 0.6 ppm. The only ventilation was
28     a partially opened window covered with a shade (AER =1.2 h"1). Klepeis et al. (1995) reported a
29     range of 0.41 to 1.2 ppm CO (average 0.75 ppm CO)  in  airport smoking lounges based on
30     10 sampling periods ranging from 60 to 146 min.  The average number of people smoking during
31     the period ranged from 2.8 to 13.5.  The room volumes ranged from 238 to 803 m3, with AERs of

       February 15, 1999                         3-50       DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
12.8 and 15.8 h"1. 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
9 test.
10
11
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
Smoking
No. of Sample Sample
Studies Size Mean Range Size
13 697 2.95 0.1-8.7 275


5 107 3.6 0.4-9.0 —
2 5 6.4 — —
2 18 2.2 1.0-5.2 10
1 35 6.0 3.7-10.2 —
1 — — — 213
Diff. in
Nonsmoking Means
Mean Range S - NS
2.99 0.7-4.0 -0.04


— —
— —
1.30 0.5-2.9 0.90
— —
11.6 8.8-22.3 —
        Source: Holcomb (1993).
1      Carbon Monoxide Concentrations from the Operation of Motor Vehicles
2          Kern et al. (1990) measured CO concentrations in a poorly sealed, detached garage from
3      operation of an emissions-controlled (catalytic reactor and oxygen sensor) and an emissions-
4      uncontrolled vehicle (carbureted, without a catalytic reactor).  Two tests were conducted: Test 1,
      February 15, 1999
                                         3-51
DRAFT-DO NOT QUOTE OR CITE

-------
 1      poorly sealed garage door with a 3-in. crack, and Test 2, garage door sealed with rags. Carbon
 2      monoxide concentrations in the poorly sealed garage reached 4,700 ppm for the uncontrolled car
 3      versus 2,000 ppm for the controlled car. When the garage was better sealed, CO concentrations,
 4      after 110 min of operation, reached 8,400 ppm for the uncontrolled vehicle versus 3,600 ppm for
 5      the controlled vehicle.
 6          Amendola and Hanes (1984) evaluated the concentration of CO in automotive repair shops
 7      based on seasonal conditions and as a function of work environment size.  Monitoring was
 8      conducted in a small service station (1 to 2 bays), a large service  station (>2 bays), and an
 9      automobile dealership.  The 8-h time-weighted average during warm weather ranged from 3.3 to
10      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
11      dealership; however, the authors noted that CO concentrations were affected by the type of
12      ventilation used in the facility, volume and type of repairs, and employee work habits, such as
13      minimizing engine run time.
14
15
16      3.6   SUMMARY
17          Carbon monoxide is produced by the incomplete combustion of burning fossil and biomass
18      fuels.  Approximately 80% of the CO produced  globally is the result of anthropogenic activities.
19      Carbon monoxide in the atmosphere is of both primary and secondary origin. The photochemical
20      oxidation of CH4 and NMHCs accounts for almost one-half of the total  source strength of CO.
21      The uncertainty in estimates of the magnitudes of individual CO  sources ranges from a factor of
22      two to three.
23          Atmospheric CO concentrations in remote areas of the world have been increasing at the
24      rate of about 1% per year throughout most of the industrial era. This increase reflects the growth
25      of anthropogenic emissions from the combustion of fossil and biomass fuels and increased
26      agriculture to feed the expanding world population. However, CO levels decreased for several
27      years from the late 1980s to the early 1990s.  The reasons for this decline are not clear, although
28      several factors may  be involved. Levels of CO may have started  increasing again over the past
29      few years, although the significance of this trend is debatable.
30          Carbon monoxide plays an important role in atmospheric chemistry because it is the major
31      reactant for  OH radicals. Reaction with OH radicals is the loss mechanism for many trace gases
        February 15, 1999                         3-52        DRAFT-DO NOT QUOTE OR CITE

-------
 1      that are responsible for contributing to the greenhouse effect (e.g., CH4) and for depleting
 2      stratospheric O3 (e.g., CH3C1, and CH3Br).  Thus, increases in CO concentrations can suppress
 3      OH radical levels and allow the concentrations of these trace gases contributing to global-scale
 4      environmental problems to increase, even if trace gas emissions are constant. Conversely,
 5      decreases in global average CO concentrations can stabilize or even reverse the growth rates of
 6      the trace gases mentioned above.
 7           Carbon monoxide may be responsible for 20 to 40% of the O3 formed in the background or
 8      "clean" troposphere.  In addition, CO may have been responsible for 10 to 20% of the O3 formed
 9      during smog episodes in the few urban areas that have been examined. Obviously, the
10      photochemistry must be examined in more cities before any more general statements about the
11      importance of CO for urban air chemistry can be made.
12           Emissions from transportation dominate other sources of CO within the United States.
13      Even though CO in urban areas results largely from motor vehicle emissions, a sizable fraction of
14      the CO observed in rural air may be produced by the photochemical oxidation of isoprene and
15      other NMHCs, at least according to one model study. The uncertainties in the nationwide and
16      worldwide emission inventories are comparable (i.e., roughly a factor of between two  and three).
17           There has been a consistent decrease in the nationwide annual second highest maximum
18      8-h composite average ambient CO concentration over the past 20 years, from about 11 ppm in
19      1977 to about 4 ppm in 1996. This improvement in CO quality occurred despite a 121% increase
20      in vehicle miles traveled, a 29% increase in population,  and a 104% increase in gross domestic
21      product in the United States over the same period. During the past 10 years, the composite mean
22      annual CO second maximum 8-h concentration decreased 37% at 190 urban sites, 37% at
23      142 suburban locations, and 48% at 10 rural monitoring sites. Hourly average CO concentrations
24      decreased from 2.0 ppm to 1.2  ppm over the past 10 years.  Despite uncertainties in the
25      calculations of CO emissions, the decline in ambient CO concentrations in the United States
26      reflects the controls placed on automotive emissions.  These declines are seen clearly in the
27      trends in CO levels during times of day when CO concentrations result mainly from mobile
28      source emissions.
29           The patterns and trends of observed CO reflect reductions in the CO emissions of the past
30      11 years.  However, it is important to note that the reported concentrations from the monitoring
31      sites are representative  only of the air quality in their neighborhoods.  Also, although personal

        February 15, 1999                         3-53        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposure to CO from mobile sources also should be decreasing, the CO values from the
 2      monitoring sites are not equivalent to personal exposures. The same ratios of personal to
 3      monitored CO from past studies in urban areas with CO emissions dominated by mobile sources
 4      may remain applicable today, but continued validation is needed.
 5           Carbon monoxide occurs in indoor environments directly through emissions from various
 6      indoor combustion sources or indirectly as a result of infiltration or ventilation from outdoor
 7      sources. Carbon monoxide concentrations in the indoor compartment is influenced by the CO
 8      emission rate of the unvented combustion appliance, the ambient CO concentration, infiltration
 9      through the building envelope, building volume, AER, and air mixing within the indoor
10      compartments.  In the absence of an indoor source, CO concentrations generally will equal those
11      in the surrounding ambient environment.
12           Carbon monoxide emissions from gas ranges vary from range to range for both the top
13      burners and the oven burners and will depend on the type of pilot light, the fuel consumption
14      rate, the frequency of use,  and the operating condition.  Older gas ranges with standing pilot
15      lights emit more CO than do newer units with electronic pilot lights. Estimates of fuel
16      consumption range from 5,000 to 5,710 ft3/year for burners and standing pilots.  Annual pilot
17      light  fuel consumption for standing pilots has been estimated to be 3,530 ft3. A steady decrease
18      in the emission of CO from pilot lights likely will occur with the replacement of older gas ranges
19      with  new models without standing pilots. Also, with the advent of other cooking appliances
20      (e.g., microwaves, toaster ovens, heating plates), the use of ranges  in meal preparation is
21      decreasing. Estimates of gas stove usage range from 30 to 60 min/day. Carbon monoxide
22      emissions from furnaces are generally negligible, but less efficient models will emit more CO
23      than the more efficient ones.
24           The use of well-maintained,  energy-efficient gas stoves will result in only intermittent,
25      small increases in CO concentrations. Carbon monoxide concentrations from gas stove usage
26      have  been reported to range from 0.65 to >9.0 ppm.  The high levels of CO  likely were
27      associated with the presence of multiple CO sources, the use of the gas stove as a supplemental
28      heat source, infiltration from ambient sources, and the building type (multi-family units versus
29      single-family units).
30           Emissions from unvented space heaters are a function of the appliance design, combustion
31      efficiency, length and frequency of use, and the fuel type and consumption rate.  The combustion

        February 15, 1999                          3-54        DRAFT-DO NOT QUOTE OR CITE

-------
 1      appliances generating the highest CO emissions and concentrations are unvented gas and
 2      kerosene space heaters. Convective space heaters emit the smallest amount of CO.  Radiant and
 3      infrared unvented space heaters emit higher amounts of CO. The CO concentration from the use
 4      of a radiant-tile gas unit was 13.4 ppm, compared with 0.9 ppm for a convective gas unit.
 5      However, CO concentrations also depend on the frequency and duration of use of the space
 6      heater. Higher CO is found in homes where space heaters are the primary heat source (up to
 7      37 ppm) versus homes where space heaters are used to supplement another heat source
 8      (<9.0 ppm).
 9           Available studies on woodstoves and fireplaces indicate that CO is emitted from fire
10      start-up and maintenance, leaks in the stove and venting system, and through backdrafting when
11      the natural draft is overcome by depressurization.  The average CO source strength for airtight
12      stoves ranged from 10 to 140 cnrVh (9.7 to 136 g/h) versus 220 to 1,800 cnrVh (213.7 to
13      1,748.6 g/h) for nonairtight stoves.
14           Carbon  monoxide from the combustion of tobacco products occurs in the indoor
15      environment primarily through sidestream smoke. Emissions will vary based on the type and
16      brand of tobacco product. Concentrations of CO in environments where smoking occurs will
17      exceed background concentrations but will be dependent on the CO emission rate of the tobacco
18      product, number of cigarettes smoked, smoking rate, size  of the indoor compartment, ventilation
19      rate, and ambient CO concentrations.
20           Carbon  monoxide emissions from the use of combustion engines may produce significant
21      increases in CO in the  microenvironments where the engines are being operated. Emissions from
22      the operation  of motor vehicles in enclosed, inadequately  ventilated spaces such as garages and
23      repair shops have resulted in 8-h time-weighted average CO concentrations of up to 22 ppm in
24      the summer and 110 ppm during the winter.
25
       February 15, 1999                         3-55        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       Allen, D. I; Kasibhatla, P.; Thompson, A. M.; Rood, R. B.; Doddridge, B. G.; Pickering, K. E.; Hudson, R. D.; Lin,
  5             S.-J. (1996) Transport-induced interannual variability of carbon monoxide determined using a chemistry and
  6             transport model. J. Geophys. Res. [Atmos.] 101: 28,655-28,669.
  7       Amendola, A. A.; Hanes, N. B. (1984) Characterization of indoor carbon monoxide levels produced by the
  8             automobile. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of the 3rd international
  9             conference on indoor air quality and climate, v. 4, chemical characterization and personal exposure; August;
10             Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building Research; pp. 97-102. Available
11             from: NTIS, Springfield, VA; PB85-104214.
12       Andreae, M. O. (1991) Biomass burning: its history, use, and distribution and its impact on environmental quality
13             and global climate. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
14             implications. Cambridge, MA: MIT Press; pp. 1-21.
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       Billick, I.; Johnson, D.; Moschandreas, D.; Relwani, S. (1984) An investigation of operational factors that influence
18             emission rates from gas appliances. In: Berglund, B.; Lindvall, T.; Sundell, J., eds. Indoor air: proceedings of
19             the 3rd international conference on indoor air quality and climate, v. 4, chemical characterization and
20             personal exposure; August; Stockholm, Sweden. Stockholm, Sweden: Swedish Council for Building
21             Research; pp. 181-188. Available from: NTIS, Springfield, VA; PB85-104214.
22       Billick, I. H.; Wilson, A. L.; Colome, S. D.; Tian, Y. (1994) California residential indoor air quality study:
23             methodology. Presented at: 87th annual meeting and exhibition of the Air & Waste Management Association;
24             June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-WP90.05.
25       Billick, I. H.; Behrens, D. W.; Becker, E. W.; Garrison, C. A.; Wilson, A. L.;  Colome, S. D.; Tian, Y. (1996)
26             California residential indoor air quality study: an overview. In: Proceedings of the 1995 international gas
27             research conference; November 1995; Cannes, France. Rockville, MD: Government Institutes, Inc.;
28             pp. 1707-1716.
29       Borrazzo, J. E.; Davidson, C. I.; Hendrickson, C. T. (1987) A statistical analysis of published gas stove emission
30             factors  for CO, NO, and NO2. In: Seifert, B.; Esdorn, H.; Fischer, M.; Rtiden, H.; Wegner, J., eds. Indoor air
31             '87: proceedings of the 4th international conference on indoor air quality and climate. Volume 1: volatile
32             organic compounds, combustion gases, particles and fibres, microbiological agents; August; Berlin, Federal
33             Republic of Germany. Berlin, Federal Republic of Germany: Institute for Water, Soil and Air Hygiene;
34             pp. 316-320.
35       Briihl, C.; Crutzen, P. J. (1998) Atmospheric CO—current issues. In: Khalil, M. A. K.; Pinto, J. P.; Shearer, M. J.,
36             eds. Proceedings of the international conference on atmospheric carbon monoxide and its environmental
37             effects; December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection Agency; report
38             no. EPA/600/R-98/047; pp. 13-23.
39       Burton, R. M.; Seila, R. A.; Wilson, W. E.; Pahl, D. A.; Mumford, J. L.; Koutrakis, P. (1990) Characterization of
40             kerosene heater emissions inside two mobile homes. In: Indoor air '90: precedings of the 5th international
41             conference on indoor air quality and climate, volume 2, characteristics of indoor air; July-August; Toronto,
42             ON, Canada. Ottawa, ON, Canada:  International Conference on Indoor Air Quality and Climate, Inc.;
43             pp. 337-342.
44       Carroll, M. A.; Hastie, D. R.; Ridley, B. A.; Rodgers, M. O.; Torres, A. L.; Davis, D. D.; Bradshaw, J. D.;
45             Sandholm, S. T.;  Schiff, H. I.; Karecki, D. R.; Harris,  G. W.; Mackay, G. I.; Gregory, G. L.; Condon, E. P.;
46             Trainer, M.; Hubler, G.; Montzka, D. D.; Madronich,  S.; Albritton, D. L.; Singh, H. B.; Beck, S. M.;
47             Shipham, M. C.; Bachmeier, A. S. (1990) Aircraft measurement of NOX over the eastern Pacific and
48             continental United States and implications for ozone production. J. Geophys. Res. [Atmos.]
49             95: 10,205-10,233.
50       Chameides, W. L.; Liu,  S. C.; Cicerone, R. J. (1977) Possible variations in atmospheric methane. J. Geophys. Res.
51             82:1795-1798.
         February 15, 1999                              3-56          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Cicerone, R. J. (1988) How has the atmospheric concentration of CO changed? In: Rowland, F. S.; Isaksen, I. S. A.,
  2             eds. The changing atmosphere: report of the Dahlem workshop on the changing atmosphere; November 1987;
  3             Berlin, Federal Republic of Germany. Chichester, United Kingdom: John Wiley & Sons, Ltd.; pp. 49-61.
  4             (Physical, chemical, and earth sciences research report no. 7).
  5       Cohen, J.; Iwamiya, R. (1998) Analyses of diurnal patterns in hourly and eight-hourly average ambient carbon
  6             monoxide concentrations. Research Triangle Park, NC: U.S. Environmental Protection Agency, Air Quality
  7             Trends Analysis Group; contract no. EPA 68-D7-0066.
  8       Collins, W. J.; Stevenson, D. S.; Johnson, C. E.; Derwent, R. G. (1997) Tropospheric ozone in a global-scale
  9             three-dimensional Lagrangian model and its response to NOX emission controls. J. Atmos. Chem.
10             26:223-274.
11       Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
12             monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research Institute;
13             report no. GRI-93/0224.3. Available from: NTIS, Springfield,  VA; PB96-159421.
14       Conrad, R. (1996) Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CH4, OCS, N2O and
15             NO). Microbiol. Rev. 60: 609-640.
16       Conrad, R.; Seiler, W. (1985) Influence of temperature, moisture, and organic carbon on the flux of H2 and CO
17             between soil and atmosphere: field studies in subtropical regions. J. Geophys. Res. [Atmos.] 90: 5699-5709.
18       Cote, W. A.; Wade, W. A., Ill; Yocom, J. E. (1974) A study of indoor air quality. Washington, DC: U.S.
19             Environmental Protection Agency, Office of Research and Development; EPA report no. EPA-650/4-74-042.
20             Available from: NTIS, Springfield, VA; PB-238556.
21       Croes, B.; Effa, R.; Pasek, R. (1996) A comparison of the California emissions inventory with 1995 ambient
22             concentration data in Los Angeles. Presented at: The 6th CRC on-road vehicle emissions workshop; March;
23             San Diego, CA. Sacramento, CA: California Air Resources Board.
24       Davidson, C. I.; Borrazzo, J. E.; Hendrickson, C. T. (1987) Pollutant emission factors for gas stoves: a literature
25             survey.  Research Triangle Park, NC: U.S. Environmental Protection Agency, Air and Energy Engineering
26             Research Laboratory; EPA report no. EPA-600/9-87-005. Available from: NTIS, Springfield, VA;
27             PB87-171328.
28       Dignon, J.; Ando, T.; Brenninkmeijer, C.; Conny, J.; Granier, C.; Khalil, A.; King, G.; King, S.; Law, K.; Levine,
29             J. S.; Rockmann;  Stedman, D.; Yung, Y.; Zafiriou, O. (1998) Carbon monoxide sources and sinks: working
30             group summary report. In: Khalil, M. A. K.; Pinto, J. P.;  Shearer, M. J., eds. Proceedings of the international
31             conference on atmospheric carbon monoxide and its environmental effects; December, 1997; Portland,
32             Oregon. Washington, DC: U.S. Environmental Protection Agency; report no. EPA/600/R-98/047;
33             pp. 359-374.
34       Duret, M. J.; Tidball, R.  K. (1990) Concept development for low emissions space heater: final report (May 1, 1988 -
35             AugustSl, 1989). Chicago, IL: Gas Research Institute; report no. GRI-90/0164.
3 6       Federal Trade  Commission. (1994) Tar, nicotine, and carbon monoxide of the smoke  of 93 3 varieties of domestic
37             cigarettes. Washington, DC: Federal Trade Commission.
38       Fujita, E. M.; Croes, B. E.; Bennett, C. L.; Lawson, D. R.; Lurmann, F.  W.; Main, H.  H.  (1992) Comparison of
3 9             emission inventory and ambient concentration ratios of CO, NMOG, and NOX in California's south coast air
40             basin. J. Air Waste Manage. Assoc. 42: 264-276.
41       Grosjean, D.; Fung, K. (1984) Hydrocarbons and carbonyls in Los Angeles air. J. Air Pollut. Control Assoc.
42             34: 537-543.
43       Haan, D.; Martinerie, P.; Raynaud, D. (1996) Ice core data of atmospheric carbon monoxide over Antarctica and
44             Greenland during the last 200 years. Geophys. Res. Lett. 23: 2235-2238.
45       Hameed, S.; Pinto, J. P.; Stewart, R. W. (1979) Sensitivity of the predicted CO-OH-CH4 perturbation to
46             tropospheric NO^ concentrations. J. Geophys. Res. C: Oceans Atmos. 84: 763-768.
47       Hard, T. M.; Mehrabzadeh, A. A.; Chan, C. Y.; O'Brien, R. J. (1992) PAGE measurements of tropospheric HO with
48             measurements and model of interferences. J. Geophys. Res. [Atmos.] 97: 9795-9817.
49       Hard, T.; Rasmussen, R. A.; Bergamaschi, P.;  Gros, V.; Hahn, M.; Herman, R.; Hoor, P.; Klouda, G.; Novelli, P.;
50             Raynaud, D.; Reichle, H.; Shearer, M.; Waibel, A.; Yurganov, L. (1998) Distributions and trends: working
51             group summary report. In: Khalil, M. A. K.; Pinto, J. P.;  Shearer, M. J., eds. Proceedings of the international
52             conference on atmospheric carbon monoxide and its environmental effects; December, 1997; Portland,
53             Oregon. Washington, DC: U.S. Environmental Protection Agency; report no. EPA/600/R-98/047;
54             pp. 173-183.
         February 15, 1999                             3-57         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hartwell, T. D.; Clayton, C. A.; Ritchie, R. M; Whitmore, R. W.; Zelon, H. S.; Jones, S. M.; Whitehurst, D. A.
  2             (1984) Study of carbon monoxide exposure of residents of Washington, DC and Denver, Colorado. Research
  3             Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory;
  4             EPA report no. EPA-600/4-84-031. Available from: NTIS, Springfield, VA; PB84-183516.
  5       Haste, T. L.; Chinkin, L. R.; Kumar, N.; Lurmann, F.; Hurwitt, S. (1998) Use of ambient data to evaluate a regional
  6             emission inventory for the San Joaquin Valley in California. Presented at: PM2.5: a fine particle standard
  7             conference; January; Long Beach, CA.
  8       Hedrick, R. L.; Krug, E. K. (1995) Conventional Research House measurements of emissions from unvented gas
  9             space heaters. Chicago, IL: Gas Research Institute; report no. GRI-95/0205. Available from: NTIS,
10             Springfield, VA; PB95-266466
11       Himmel, R. L.; DeWerth,  D. W. (1974) Evaluation of the pollutant emissions from gas-fired ranges. Project
12             EP-1-23: analysis of flue products from gas-fired appliances. Cleveland, OH: American Gas Association
13             Laboratories; report no. 1492.
14       Hofmann, D. J.; Peterson, J. T.; Rosson, R. M., eds. (1996) Climate monitoring and diagnostics laboratory no. 23.
15             Summary report 1994-95. Boulder, CO: U.S. Department of Commerce, National Oceanic and Atmospheric
16             Administration, Environmental Research Laboratories.
17       Holcomb, L. C. (1993) Indoor air quality and environmental tobacco smoke: concentration and exposure. Environ.
18             Int. 19:9-40.
19       Ingalls, M. N.; Smith, L. R.; Kirksey, R. E. (1989) Measurement of on-road vehicle emission factors in the
20             California South Coast Air Basin: volume I, regulated emissions. Atlanta, GA: Coordinating Research
21             Council, Inc.; report no. CRC-APRAC-AP-4-SCAQS-l. Available from: NTIS, Springfield, VA;
22             PB89-220925.
23       Inman, R. E.; Ingersoll, R. B.; Levy, E. A. (1971) Soil: a natural sink for carbon monoxide. Science (Washington,
24             DC) 172: 1229-1231.
25       Intergovernmental Panel on Climate Change.  (1996) Climate change 1995:  the science of climate change. Summary
26             for policymakers and technical summary of the working group I report. Presented at: the fifth session of the
27             IPCC Working Group I; November; Madrid, Spain. Geneva, Switzerland: World Meteorological Association.
28       Ipps, D. T.; Popejoy, D. (1998) An updated verification of the on-road vehicle emissions inventory for the
29             Los Angeles basin using ambient air quality data. Presented at: The 8th CRC on-road vehicle emissions
30             workshop; April; San Diego, CA. Sacramento, CA: California Air Resources Board.
31       Jaasma, D. R.; Stern, C. H.; Honeycutt, D. J.; Fugler, D.  (1995) Effect of glass doors on carbon monoxide spillage
32             from masonry fireplaces. Presented at: 88th annual meeting and exhibition of the Air & Waste Management
33             Association; June;  San Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no.
34             95-RP137.01.
3 5       Jacob, D. J.; Wofsy, S. C. (1990) Budgets of reactive nitrogen, hydrocarbons, and ozone over the Amazon forest
36             during the wet season. J. Geophys. Res. [Atmos.] 95:  16,737-16,754.
37       Jeffries, H. (1998) Sources and sinks of CO: urban hydrocarbon oxidation.  In: Khalil, M. A. K.; Pinto, J. P.;
3 8             Shearer, M. J., eds. Proceedings of the international conference on atmospheric carbon monoxide and its
39             environmental effects; December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection
40             Agency; report no.  EPA/600/R-98/047; pp. 377-387.
41       Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle Park,
42             NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; EPA report no.
43             EPA-600/4-84-014. Available from: NTIS, Springfield, VA; PB84-146125.
44       Johnson, T.; Capel,  J.; Paul, R.; Wijnberg, L. (1992) Estimation of carbon monoxide exposures and associated
45             carboxyhemoglobin levels in Denver residents using a probabilistic version of NEM. Research Triangle Park,
46             NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no.
47             68-DO-0062.
48       Kanakidou, M.; Crutzen, P. J. (1993) Scale problems in global tropospheric chemistry modeling: comparison of
49             results obtained with a three-dimensional model, adopting longitudinally uniform and varying emissions of
5 0             NOX and NMHC. Chemosphere 26: 787-801.
51       Kanakidou, M.; Singh, H. B.; Valentin, K. M.; Crutzen, P. J. (1991) A two-dimensional study of ethane and
52             propane oxidation in the troposphere. J. Geophys. Res. [Atmos.] 96:  15,395-15,413.
53       Kern, R. A.; Knoedler, J. A.; Trester, R. J. (1990) Automotive carbon monoxide emissions in a closed garage.
54             Presented at:  International congress and exposition; February-March; Detroit, MI. Warrendale, PA:  Society
55             of Automotive Engineers; pp. 33-38. (SAE technical paper series no. 900499).


         February 15, 1999                             3-58         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Khalil, M. A. K.; Rasmussen, R. A. (1988) Carbon monoxide in the earth's atmosphere: indications of a global
  2             increase. Nature (London) 332: 242-245.
  3       Khalil, M. A. K.; Rasmussen, R. A. (1994) Global decrease in atmospheric carbon monoxide concentration. Nature
  4             (London) 370: 639-641.
  5       Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
  6             monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
  7             meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh,
  8             PA: Air & Waste Management Association; paper no. 95-MPH.03. Available from: NTIS, Springfield, VA;
  9             PB95-225496.
10       Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
11             lounges. Environ. Sci. Technol. 30: 2813-2820.
12       Klepeis, N. E.; Ott, W. R.; Repace, J. L. (1999) The effect of cigars on indoor air quality. J. Exposure Anal.
13             Environ. Epidemiol. [in press].
14       Koontz,  M. D.; Nagda, N. L. (1988) A topical report on a field monitoring  study of homes with unvented gas space
15             heaters. Volume Ill-methodology and results. Chicago, IL: Gas Research Institute; report no.
16             GRI-87/0044.2. Available from: NTIS, Springfield,  VA; PB88-216791.
17       Koontz,  M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
18             Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics;
19             contract nos. 68-D9-0166 and 68-D3-0013.
20       Koontz,  M. D.; Mehegan, L. L.; Nagda, N. L. (1992) Distribution and use of cooking appliances that can affect
21             indoor air quality: topical report (November 1985—December 1991). Chicago, IL: Gas Research Institute;
22             report no. GRI-93/0013. Available from: NTIS, Springfield, VA; PB93-185452/XAB.
23       Korc, M. E.; Roberts, P. T.; Main,  H. H.; Chinkin, L. R.  (1993) Comparison of emission inventory and ambient
24             concentration ratios of CO, NMOC, and NOx in the Lake Michigan air quality region. Santa Rosa,  CA:
25             Sonoma Technology, Inc.; report no. STI-90218-1358.
26       Krol, M.; Van Leeuwen, P. J.; Lelieveld, J. (1998) Global OH trend inferred from methylchloroform measurements.
27             J. Geophys. Res. [Atmos.] 103: 10,697-10,711.
28       Lagus Applied Technology, Inc. (1995) Air change rates in non-residential buildings in California. Sacramento, CA:
29             California Energy Commission; contract no. 400-91-034; July.
30       Lawson, D. R.; Groblicki, P. J.; Stedman, D. H.; Bishop, G. A.; Guenther, P.  L. (1990) Emissions from in-use
31             motor vehicles in Los Angeles: a pilot study of remote sensing and the inspection and maintainence program.
32             J. Air Waste Manage. Assoc. 40: 1096-1105.
33       Leaderer, B. P. (1982) Air pollutant emissions from kerosene space heaters. Science (Washington, DC)
34             218: 1113-1115.
35       Levine, J. S.; Pinto, J. P. (1998) The production of CO by biomass burning. In: Khalil, M. A. K.; Pinto, J. P.;
36             Shearer, M. J., eds. Proceedings of the international  conference on atmospheric carbon monoxide and its
37             environmental effects; December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection
3 8             Agency; report no. EPA/600/R-98/047; pp. 251-256.
39       Liebl, K. H.; Seiler, W. (1976)  CO and H2 destruction at the soil surface. In: Schlegel, H. G.; Gottschalk, G.;
40             Pfennig, N., eds. Microbial production and utilization of gases (H2, CH4, CO): proceedings of the
41             symposium; September 1975; Goettingen, Federal Republic of Germany. Goettingen, Federal Republic of
42             Germany: E. Goltze KG; pp. 215-229.
43       Lobert, J. M.;  Scharffe, D. H.; Hao, W.-M.; Kuhlbusch, T.  A.; Seuwen, R.; Warneck, P.; Crutzen, P. J. (1991)
44             Experimental evaluation of biomass burning emissions: nitrogen and carbon containing compounds.
45             In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge,
46             MA: MIT Press; pp. 289-304.
47       Lofroth, G.; Burton, R. M.; Forehand, L.; Hammond, S. K.; Seila, R. L.; Zweidinger, R. B.; Lewtas, J. (1989)
48             Characterization of environmental tobacco smoke. Environ.  Sci. Technol. 23: 610-614.
49       Logan, J. A.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B.  (1981) Tropospheric chemistry: a global perspective.
50             J. Geophys. Res. C: Oceans Atmos.  86: 7210-7254.
51       Magliano, K. L.; Lurmann, F. W.; Watson, J. G.  (1993) Reconciliation of emissions inventory and ambient data in
52             the San Joaquin Valley and the San Francisco Bay area. Presented at: 86th annual meeting & exhibition of
53             the Air & Waste Management Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management
54             Association; paper no. 93-WP-100.8.
         February 15, 1999                             3-59         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Menkedick, J.; Niemuth, N.; Hartford, P.; Landstrom, D. K. (1993) Metered ranges, cooktops and ovens in the
  2             northern Illinois gas residential load study data base: topical report. Chicago, IL: Gas Research Institute;
  3             report no. GRI-93/0204; July.
  4       Moschandreas, D. J.; Relwani, S. M; O'Neill, H. J.; Cole, J. T.; Elkins, R. H.; Macriss, R. A. (1985)
  5             Characterization of emission rates from indoor combustion sources. Chicago, IL: Gas Research Institute;
  6             report no. GRI 85/0075. Available from: NTIS, Springfield, VA; PB86-103900.
  7       Mount, G. H.; Williams, E. J. (1997) An overview of the tropospheric OH photochemistry experiment, Fritz
  8             Peak/Idaho Hill, Colorado, fall 1993. J. Geophys. Res.  [Atmos.] 102: 6171-6186.
  9       Moxley, J. M.; Cape, J. N. (1997) Depletion of carbon monoxide from the nocturnal boundary layer. Atmos.
10             Environ. 31: 1147-1155.
11       Mueller Associates. (1985) Indoor air quality environmental information handbook: combustion sources.
12             Washington, DC: U.S. Department of Energy, Office of Environmental Analysis; report no.
13             DOE/EV/10450-1. Available from: NTIS; Springfield, VA; DE85006589.
14       Mumford, J. L.; Lewtas, J.; Burton, R. M.; Svendsgaard, D. B.; Houk, V. S.; Williams, R. W.; Walsh, D. B.;
15             Chuang, J. C. (1990) Unvented kerosene heater emissions in mobile homes:  studies on indoor air particles,
16             semivolatile organics, carbon monoxide, and mutagenicity.  In: Indoor air '90: precedings of the 5th
17             international conference on indoor air quality and climate, volume 2, characteristics of indoor air;
18             July-August; Toronto, ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air Quality
19             and Climate, Inc.; pp. 257-262.
20       Mumford, J. L.; Williams, R. W.; Walsh, D. B.; Burton, R. M.; Svendsgaard, D. J.; Chuang, J. C.; Houk, V. S.;
21             Lewtas, J. (1991) Indoor air pollutants from unvented kerosene heater emissions in mobile homes: studies on
22             particles, semivolatile organics, carbon monoxide, and mutagenicity. Environ. Sci. Technol. 25: 1732-1738.
23       Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
24             parametric distributions by season and climatic region. Risk Anal.  15: 459-465.
25       Nagda, N. L.; Koontz, M. D.; Fortmann, R. C.; Billick, I. H. (1987) Detailed evaluation of range-exhaust fans.
26             In: Seifert, B.; Esdorn, H.; Fischer, M.; Rueden, H.; Wegner, J., eds. Indoor  air '87: proceedings of the
27             4th international conference on indoor air quality and climate, v. 3, developing countries, guaranteeing
28             adequate indoor air quality, control measures, ventilation effectiveness, thermal climate and comfort, policy
29             and strategies; August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of Germany: Institute
3 0             for Water, Soil and Air Hygiene;  pp. 324-328.
31       Nagda, N. L.; Koontz, M. D.; Billick, I.  H.; Leslie, N. P.; Behrens, D. W. (1996) Causes and consequences of
32             backdrafting of vented gas appliances. J. Air Waste Manage. Assoc. 46: 838-846.
33       National Aeronautics and Space Administration, Panel for Data Evaluation. (1997)  Chemical kinetics and
34             photochemical data for use in stratospheric modeling: evaluation number 12. Pasadena,  CA: National
3 5             Aeronautics and Space Administration, Jet Propulsion Laboratory; JPL publication 97-4.
36       National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health
37             effects. Washington, DC: National Academy Press.
38       Novelli, P. C.; Connors, V.  S.; Reichle, H. G., Jr.;  Anderson, B. E.; Brenninkmeijer, C. A. M.; Brunke, E. G.;
39             Doddridge, B. G.; Kirchhoff, V. W. J. H.; Lam, K. S.; Masarie, K. A.; Matsuo, T.; Parrish,  D. D.; Scheel,
40             H. E.; Steele, L.  P. (1998) An internally consistent set of globally distributed atmospheric carbon monoxide
41             mixing ratios developed using results from an intercomparison of measurements. J. Geophys. Res. [Atmos.]
42             103: 19,285-19,293.
43       Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities for
44             human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
45       Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
46             functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
47             June; San Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no. 95-WP84B.03.
48       Pacyna, J. M.; Graedel, T. E. (1995) Atmospheric emissions inventories:  status and prospects.  Annu. Rev. Energy
49             Environ. 20: 265-300.
50       Park, J.-H.; Spengler, J. D.; Yoon, D.-W.; Dumyahn, T.; Lee, K.; Ozkaynak, H. (1998) Measurement of air
51             exchange rate of stationary vehicles and estimation  of in-vehicle exposure. J. Exposure Anal. Environ.
52             Epidemiol. 8: 65-78.
53
         February 15, 1999                              3-60          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Pierce, T. E.; Dudek, M. P. (1996) Biogenic emission estimates for 1995. In: The emission inventory: key to
  2             planning, permits, compliance, and reporting; proceedings of a specialty conference; September;
  3             New Orleans, LA. Pittsburgh, PA: Air & Waste Management Association; pp. 309-326. (AWMA publication
  4             VIP-65).
  5       Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with other on-road
  6             vehicle emission data. J. Air Waste Manage. Assoc. 40: 1495-1504.
  7       Pierson, W. R.; Gertler, A. W.; Robinson, N. F.; Sagebiel, J. C.; Zielinska, B.; Bishop, G. A.; Stedman, D. H.;
  8             Zweidinger, R. B.; Ray, W. D. (1996) Real-world automotive emissions—summary of studies in the Fort
  9             McHenry and Tuscarora mountain tunnels. In: Parrish, D.; Trainer, M.; Rao. S. T.; Solomon, P. A.,  eds.
10             A&WMA international specialty conference on regional photochemical measurements and modeling, part 2;
11             November 1993; San Diego, CA.  Atmos. Environ. 30: 2233-2256.
12       Pinto, J. P.; Khalil, M. A. K. (1991) The stability of tropospheric OH during ice ages, inter-glacial epochs and
13             modern times. Tellus 43B: 347-352.
14       Pinto, J. P.; Rasmussen, R. A. (1998) Sources of CH4, CO, and H2 in Prague, Czech Republic. In: Khalil, M. A. K.;
15             Pinto, J. P.; Shearer, M. J., eds. Proceedings of the international conference on atmospheric carbon monoxide
16             and its environmental effects;  December,  1997; Portland, Oregon. Washington, DC: U.S. Environmental
17             Protection Agency; report no.  EPA/600/R-98/047; pp. 313-319.
18       Pollack, A. K.; Dunker, A. M.; Fieber, J. K.; Heiken, J. G.; Cohen, J. P.; Shepard, S.  B.; Schleyer, C. H.;
19             Yarwood, G. (1998) Revision of light-duty vehicle emission inventories using real-world
20             measurements-Auto/Oil Program, phase II. J. Air Waste Manage. Assoc. 48: 291-305.
21       Poppe, D.; Zimmermann, J.; Bauer, R.; Brauers, T.; Briining, D.; Callies, J.; Dorn, H.-P.; Hofzumahaus, A.; Johnen,
22             F.-J.; Khedim, A.; Koch, H.; Koppmann, R.; London, H.; Mtiller, K.-P.; Neuroth, R.; Plass-Dtilmer, C.;
23             Platt, U.; Rohrer, F.; Roth, E.-P.; Rudolph, J.; Schmidt, U.; Wallasch, M.; Ehhalt, D. H. (1994) Comparison
24             of measured OH concentrations with model calculations. J. Geophys. Res.  [Atmos.] 99: 16,633-16,642.
25       Prinn, R.; Cunnold, D.; Simmonds, P.; Alyea, F.; Boldi, R.; Crawford, A.; Fraser, P.; Gutzler, D.; Hartley,  D.;
26             Rosen, R.; Rasmussen, R. (1992)  Global average concentration and trend for hydroxyl radicals deduced from
27             ALE/GAGE trichloroethane (methyl chloroform) data for 1978-1990. J. Geophys. Res. [Atmos.]
28             97:2445-2461.
29       Research Triangle Institute. (1990) An investigation of infiltration and indoor air quality. Albany, NY: New York
30             State Energy Research and Development Authority; report no. NYERDA 90-11. Available from: NTIS,
31             Springfield, VA; PB91-119156.
32       Rinsland, C. P.; Levine, J. S. (1985) Free tropospheric carbon monoxide concentrations in 1950 and 1951 deduced
33             from infrared total column amount measurements. Nature (London) 318: 250-254.
34       Ritter, J. A.; Barrick, J. D. W.; Watson, C. E.; Sachse, G. W.; Gregory, G. L.; Anderson, B. E.; Woerner, M. A.;
3 5             Collins, J. E., Jr. (1994) Airborne boundary layer flux measurements of trace species over Canadian boreal
36             forest and northern wetland regions. J. Geophys. Res. [Atmos.] 99: 1671-1685.
37       Rogers, J. D. (1990) Ultraviolet absorption cross sections and atmospheric photodissociation rate constants of
38             formaldehyde. J. Phys. Chem. 94: 4011-4015.
39       Ryan, J. V.; McCrillis, R. C. (1994) Analysis of emissions from residential natural gas furnaces. Presented at: 87th
40             annual meeting and exhibition of the Air & Waste Management Association; June; Cincinnati, OH.
41             Pittsburgh, PA: Air & Waste Management Association; paper no. 94-WA75A.04.
42       Seila, R. L.; Lonneman, W. A.; Meeks, S. A. (1989) Determination of C2 to C12 ambient air hydrocarbons in 39 U.S.
43             cities from 1984 through 1986. Research Triangle Park, NC: U.S. Environmental Protection Agency,
44             Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/3-89/058. Available from:
45             NTIS, Springfield, VA; PB89-214142/HSU.
46       Seiler, W.; Conrad,  R. (1987) Contribution of tropical ecosystems to the global budget of trace gases especially
47             CH4, H2, CO and N2O. In: Dickinson, R. E., ed. The geophysiology of Amazonia: vegetation and climate
48             interactions. New York, NY: John Wiley; pp. 133-162. (Glantz, H. H., ed. Wiley series in climate and
49             biosphere).
50       Sheldon, L.; Clayton, A.; Jones, B.; Keever, J.; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992) Indoor
51             pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board, Research
52             Division; report no. ARB/R-92/486. Available from: NTIS, Springfield, VA; PB92-235837.
53       Slack, H. H.; Heumann, M. A. (1997) Use of unvented residential heating appliances—United States,  1988-1994.
54             Morb.  Mortal. Wkly. Rep. 46: 1221-1224.
55


         February 15, 1999                              3-61         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Spicer, C. W.; Billick, I. H. (1996) Persistence and fate of natural gas appliance emissions. In: Proceedings of the
  2             1995 international gas research conference; November 1995; Cannes, France. Rockville, MD: Government
  3             Institutes, Inc.; pp. 1752-1761.
  4       Stephens, R. D.; Cadle, S. H. (1991) Remote sensing measurements of carbon monoxide emissions from on-road
  5             vehicles. J. Air Waste Manage. Assoc. 41: 39-46.
  6       Sze, N. D. (1977) Anthropogenic CO emissions: implications for the atmospheric CO-OH-CH4 cycle. Science
  7             (Washington, DC) 195: 673-675.
  8       Thompson, A. M; Cicerone, R. J. (1986) Possible perturbations to atmospheric CO, CH4, and OH. J. Geophys. Res.
  9             [Atmos.] 91: 10,853-10,864.
10       Thompson, A. M.; Chappellaz, J. A.; Fung, I. Y.; Kucsera, T. L. (1993) The atmospheric CH4 increase since the last
11             glacial maximum. (2). Interactions with oxidants. Tellus 45B: 242-257.
12       Thrasher, W. H.; DeWerth, D. W. (1979) Evaluation of the pollutant emissions from gas-fired room heaters.
13             Cleveland, OH: American Gas Association Laboratories; research report no. 1515.
14       Tiegs, P.; Bighouse, R. (1994) Spillage of combustion byproducts from woodstoves operated in negative pressure
15             environments. Presented at:  87th annual meeting and exhibition of the Air & Waste Management
16             Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no.
17             94-FA150.04.
18       Traynor, G. W.; Anthon, D. W.;  Hollowell, C. D. (1982) Technique for determining pollutant emissions from a
19             gas-fired range. Atmos. Environ. 16: 2979-2987.
20       Traynor, G. W.; Allen, J. R.; Apte, M. G.; Girman, J. R.; Hollowell, C. D. (1983) Pollutant emissions from portable
21             kerosene-fired space heaters. Environ. Sci. Technol. 17: 369-371.
22       Traynor, G. W.; Apte, M. G.; Carruthers, A. R.; Dillworth, J. F.; Grimsrud, D. T. (1984) Pollutant emission rates
23             from unvented infrared and convective gas-fired space heaters. Berkeley, CA: U.S. Department of Energy,
24             Lawrence Berkeley Laboratory; report no. LBL-18258. Available from: NTIS, Springfield, VA;
25             DE85 010647/XAB.
26       Traynor, G. W.; Girman, J. R.; Apte, M.  G.; Dillworth, J. F.; White, P. D. (1985) Indoor air pollution due to
27             emissions from unvented gas-fired space heaters. J. Air Pollut. Control Assoc. 35: 231-237.
28       Traynor, G. W.; Apte, M. G.; Carruthers, A. R.; Dillworth, J. F.; Grimsrud, D. T.; Gundel, L. A. (1987) Indoor air
29             pollution due to emissions from wood-burning stoves. Environ. Sci. Technol. 21: 691-697.
30       Traynor, G. W.; Aceti, J. C.; Apte, M. G.; Smith, B. V.; Green, L. L.; Smith-Reiser, A.; Novak, K. M.; Moses,
31             D. O. (1989) Macromodel for assessing residential concentrations of combustion-generated pollutants: model
32             development and preliminary predictions for CO, NO2, and respirable suspended particles. Berkeley, CA:
33             U.S. Department of Energy, Lawrence Berkeley Laboratory; report no. LBL-25211. Available from: NTIS,
3 4             Springfield, VA; DE89013 3 96/XAB.
35       Traynor, G. W.; Apte, M. G.; Sokol, H. A.; Chuang, J. C.; Tucker, W. G.; Mumford, J. L. (1990) Selected organic
36             pollutant emissions from unvented kerosene space heaters. Environ. Sci. Technol. 24: 1265-1270.
37       Turk, B. H.; Grimsrud, D. T.; Brown, J. T.; Geisling-Sobotka, K. L.; Harrison, J.; Prill, R. J. (1989) Commercial
38             building ventilation rates and particle concentrations. ASHRAE Trans. 95(part 1): 422-433.
39       Turner, S.; Cyr, L.; Gross, A. J. (1992) The measurement of environmental tobacco smoke in 585 office
40             environments. Environ. Int.  18: 19-28.
41       U.S. Census Bureau. (1998) American housing survey: detailed tables from the 1993 AHS-N data chart. How do
42             you fuel your home? - Table 2-5. Washington, DC: U.S. Department of Commerce. Available online at:
43             www.census.gov/hhes/housing/ahs/tab2-5.html.
44       U.S. Department of Housing and Urban Development. (1987) Indoor air pollution: the role of heating and cooking.
45             Washington, DC: Office of the Assistant Secretary for Policy Development and Research.
46       U.S. Environmental Protection Agency. (1981) U.S. Environmental Protection Agency Intra-Agency Task Force on
47             Air Quality Indicators. Research Triangle Park, NC: Office of Air Quality Planning and Standards;  report no.
48             EPA-450/4-81-015.  Available from: NTIS, Springfield, VA; PB81-177982.
49       U.S. Environmental Protection Agency. (1990) National air quality and emissions trends report, 1988. Research
50             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/450/4-90/002. Available
51             from: NTIS, Springfield, VA; PB90-200114/XAB.
52       U.S. Environmental Protection Agency. (1991a) User's guide to MOBILE4.1 (mobile source emission factor
53             model). Ann Arbor, MI: Office of Mobile Sources; report no. EPA-AA-TEB-91-01. Available from: NTIS,
54             Springfield, VA; PB91-228759.
         February 15, 1999                             3-62         DRAFT-DO NOT QUOTE OR CITE

-------
  1       U.S. Environmental Protection Agency. (1991b) Air quality criteria for carbon monoxide. Research Triangle Park,
  2             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
  3             no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
  4       U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
  5             monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
  6             Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available from:
  7             NTIS, Springfield, VA; PB93-157717.
  8       U.S. Environmental Protection Agency. (1993) MOBILESa: highway vehicle emissions factor model (on diskette).
  9             Ann Arbor, MI: Emission Planning and Strategies Division; report no. EPA/SW/MT-94/120. Available from:
10             NTIS, Springfield, VA; PB95-500179INC (DOS format), PB95-500187INC (Macintosh format), or
11             PB95-500138INC (mag tape, EBCDIC character set).
12       U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
13             NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
14             Available from: NTIS, Springfield, VA; PB96-168224.
15       U.S. Environmental Protection Agency. (1998) National air quality and emissions trends report, 1996. Research
16             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-97-013. Available
17             from: NTIS, Springfield, VA; PB98-135155. Available  online at: www.epa.gov/oar/aqtrnd96/.
18       Williams, R.; Walsh, D.; White, J.; Jackson, M.; Mumford, J. (1992) Effect on carbon monoxide levels in mobile
19             homes using unvented kerosene heaters for residential heating. Indoor Environ.  1: 272-278.
20       Wilson, A. L.; Colome, S. D.;  Tian, Y. (1993) California residential indoor air quality study. Volume 1.
21             Methodology and descriptive statistics, appendix. Chicago, IL: Gas Research Institute; report no.
22             GRI-93/0224.2. Available from: NTIS, Springfield, VA; PB94-166055.
23       Wilson, A. L.; Colome, S. D.;  Tian, Y.; Becker, E. W.; Baker,  P. E.; Behrens, D. W.; Billick, I. H.; Garrison, C. A.
24             (1996) California residential air exchange rates and residence volumes. J. Exposure Anal. Environ.
25             Epidemiol. 6:311-326.
26       Xiong, T.; Khinkis, M. J.; Coppin, W. P. (1991) The development of an ultra-low-emission gas-fired combustor for
27             space heaters. Presented at: 14th World Energy Engineering and 1991 Environmental Engineering Congress;
28             October; Atlanta, GA. Chicago, IL: Institute of Gas Technology.
29       Yarwood, G.; Morris, R. (1998) Regional scale CO modeling.  In: Khalil, M. A. K.; Pinto, J. P.; Shearer, M. J., eds.
3 0             Proceedings of the international conference on atmospheric carbon monoxide and its environmental effects;
31             December, 1997; Portland, Oregon. Washington, DC: U.S. Environmental Protection Agency; report no.
32             EPA/600/R-98/047; pp. 413-431.
33       Zawacki, T. S.; Cole, J. T.; Huang, V M. S.; Banasiuk, H.; Macriss, R. A. (1984) Efficiency and emissions
34             improvement of gas-fired space heaters. Task 2. Unvented space heater emission reduction [final report].
35             Chicago, IL: Gas Research Institute; report no. GRI-84/0021. Available from: NTIS, Springfield, VA;
36             PB84-237734.
37       Zawacki, T. S.; Cole, J. T.; Jasionowski, W. J.; Macriss, R. A. (1986) Measurement of emission rates from gas-fired
38             space heaters. Chicago, IL: Gas Research Institute; report no. GRI-86/0245. Available from: NTIS,
39             Springfield, VA; PB87-153995.
         February 15, 1999                             3-63         DRAFT-DO NOT QUOTE OR CITE

-------
 i                                  APPENDIX 3A:
 2
 3                The Spatial and Temporal Variability of
 4         Carbon Monoxide in  Selected Urban Areas of the
 5                                    United States
 8     3A.1 INTRODUCTION
 9          The objective of this appendix is to present a general overview of the mean concentrations
10     and the temporal and spatial variability of carbon monoxide (CO) in selected major metropolitan
11     areas located in different geographic regions of the United States. Statistical analyses (Shadwick
12     et al., 1997,  1998a,b,c) were performed on ambient CO data obtained in the Denver, CO; Los
13     Angeles, CA; New York City, NY; and Phoenix, AZ, Metropolitan Statistical Areas (MSAs).
14     These MSAs were chosen on the basis of their geographical diversity and their potential for
15     showing exceedances of the CO National Ambient Air Quality Standards (NAAQS) during the
16     years 1994 through 1996. The current NAAQS for CO was not met in the Los Angeles and
17     Phoenix MSAs in  1996.  In addition, the Denver and the New York City MSAs exceeded the CO
18     NAAQS in 1995 and 1994, respectively. The analyses presented here focus on characterizing the
19     variability in CO at different sites within the above MSAs for the winter season (defined as
20     November 1 through the last day of February) for the years 1986 through 1996. Ambient CO
21     levels are generally highest in the winter season because vehicle emissions are highest during
22     cold weather, and there is a higher incidence of enhanced stability in the atmospheric boundary
23     layer, leading to reduced vertical mixing (dilution) of emissions from the surface.
24          Carbon monoxide data were obtained from the U.S. Environmental Protection Agency's
25     Aerometric Information Retrieval System (AIRS). Data from the network described in
26     Section 3.4 are archived routinely in AIRS. Carbon monoxide data collected at the monitoring
27     sites are meant to be used to characterize levels in environments ranging from "hot spots" (i.e.,
28     sites located at or very close to major  sources) to ambient levels found in residential areas, and,
29     hence, care should be exercised in matching the spatial scale over which the data obtained at a


       February 15, 1999                       3A-1       DRAFT-DO NOT QUOTE OR CITE

-------
 1      monitoring site are meant to be representative with the spatial scale required by a particular use
 2      of the data.
 3           Motor vehicle emissions are the major source of CO in each of these MS As, and several of
 4      the sites in each MSA are located at or near roadways with high emissions. However, there are
 5      special siting criteria that guide the location of monitoring stations near source-specific high
 6      concentration clusters.  Therefore, although CO concentrations at many monitoring stations in
 7      urban areas may be dominated by mobile sources, the concentrations are not at levels that a
 8      person would experience within or near an operating automobile. Time series in CO
 9      concentrations obtained at monitoring stations in urban areas could be used to monitor the
10      progress in controlling emissions and to aid in evaluating relative changes in personal exposure
11      to ambient CO, but they cannot be used alone to provide concentrations for estimating personal
12      exposures.
13
14
15      3A.2 GENERAL METHODOLOGY FOR DATA COLLECTION AND
16           ANALYSES
17           Hourly average CO data were used to calculate running 8-h averages for 1986 to!996.
18      Only valid hourly average values were used to compose the 8-h average.  In the case that less
19      than six valid hourly average values were used to compose the 8-h average, the 8-h average was
20      set to a missing value. The six valid hourly average values in an 8-h window corresponds to 75%
21      data capture in the 8-h window.
22           The 24 running 8-h averages assigned to a day were used to compute the daily maximum
23      8-h average.  A daily maximum 8-h average was considered to be valid if at least 18 of the 8-h
24      running averages for the day were valid as described in the preceding paragraph.  The 18 valid
25      8-h running averages in a day corresponds to a 75% data capture. In the case that a valid daily
26      maximum 8-h running average could not be computed, a missing value was assigned to the daily
27      maximum 8-h average. The summary statistics were computed without regard to data capture.
28      Summary statistics (aside from the total number of observations) should be regarded as
29      representative if at least 75% of the possible data values were valid. Statistics on central
30      tendency and correlation were tabulated for all of the sites in each MSA for both the hourly and
       February 15, 1999                         3A-2        DRAFT-DO NOT QUOTE OR CITE

-------
 1      8-hour running averages.  The statistics were analyzed by year, season, day of week, and hour of
 2      the day.
 3          The Kendall tau correlation statistic was used to evaluate the relations in the daily
 4      maximum 8-h averages among the different sites in the four MSAs chosen above. Kendall tau is
 5      a nonparametric measure of correlation between two ranked time series.  The Kendall tau
 6      measures the similarity in rankings within two time series.  It is plus one (+1) if the
 7      corresponding values in the two time series are ranked exactly the same way, and minus one (-1)
 8      if the corresponding values in the two time series are ranked exactly in reverse order.
 9
10
11      3A.3 RESULTS AND DISCUSSION
12      3A.3.1  Denver
13          The locations of the eight monitors that had valid hourly average CO data for 1986 to 1996
14      are shown in Figure 3 A-l.  The CO monitors in the Denver MSA are located in predominantly
15      urban or suburban locations and represent six of the seven counties in the Denver MSA. The
16      average hourly CO concentration among all the sites ranged from 0.32 to 1.52 ppm in 1996. The
17      CO time series at each of these monitors show nuances unique to each site, but there are also
18      similarities common to all of the sites. These similarities will be the focus for discussion;
19      consequently, the downtown Denver monitoring site for CO will be used as the illustrative
20      example of the trends in ambient CO concentrations in Denver over the last decade. This  site is
21      designated as Broadway (AIRS identification number 080310002).
22          Figure 3A-2 is a three-dimensional plot of the average diurnal CO concentration time series
23      for weekdays for the winter seasons  (November through February, as defined above) from
24      1986/87 through 1995/96.  It is important to note that this graph shows the average
25      concentrations during a 24-h period for each years' winter season and, as such, is not indicative
26      of the CO time series of concentrations for any given  day.
27          The most notable features in the CO concentration trends for Denver, as seen in
28      Figure 3A-2, are the double peaks: (1) one maximum starting at approximately 6 a.m.  and
29      peaking at about 9 a.m. and (2) the second maximum  occurring at approximately 6 p.m., after
30      beginning at about 3 p.m. These peaks correspond to rush hour traffic and result almost entirely

        February 15, 1999                         3A-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      from emissions from motor vehicles. For the 10-year period in this figure, 1986 through 1996,
 2      the average CO concentrations were higher during the afternoon peak.  Also, the CO
 3      concentrations fall off much less rapidly following the afternoon peak than they do following the
 4      morning peak. Indeed, the average concentrations seen after midnight are still higher than those
 5      measured at noon, just 3 h after the morning peak. The nighttime behavior of CO may reflect the
 6      presence of additional sources, such as wood burning, and the influence of atmospheric
 7      dynamics, such as reduced vertical mixing in the atmospheric surface layer at night.  However,
 8      wood burning has been restricted in the Denver area in recent years and so the reduction in
 9      atmospheric boundary layer height during the night may more than compensate for the decline in
10      motor vehicle emissions. There is also a very noticeable downward trend in CO concentrations
11      from 1986 to 1996 that was observed at all sites.  At the Broadway site, for example, weekday
12      hourly average CO concentrations during the winter season fell 60% from 1986 to 1996, whereas
13      weekend values dropped by 47% over the same period.
14          Figures 3 A-3 and 3 A-4 show average monthly CO concentration data for the periods from
15      May 1986 through May  1987 and May 1995 through May 1996, respectively. The dominant
16      feature in Figure 3A-3 is the large peak occurring between 5 p.m. and 6 p.m. in December of
17      1986. A similar peak, although reduced in magnitude, for December 1995 is  seen in
18      Figure 3 A-4. Overall, the concentrations observed in Figure 3 A-4 are reduced substantially in
19      relation to their counterparts in Figure 3A-3.  However, the general shape of the CO time series
20      shows little change from 1986/87 (Figure 3A-3) to 1995/96 (Figure 3A-4). The morning and
21      early evening peaks are evident in both figures.  These peaks, primarily because of motor vehicle
22      emissions, occur throughout the entire year. However, the concentrations are elevated from
23      about November though March compared with the rest of the year. In addition to being higher
24      than those in 1995/96, the concentrations in 1986/87 also show greater variability. This is
25      especially true for the concentrations observed during the off-peak hours (again, the peak hours
26      are centered on 9  a.m. and 6 p.m.).  The diurnal CO concentrations for May 1995 through May
27      1996 (Figure 3 A-4) show less variability, particularly during the off-peak hours, than do those in
28      the May 1986 through May 1987 period (Figure 3A-3).
29          Figure 3A-5 shows two graphs of the central tendency of the wintertime daily maximum
30      8-h average for weekdays at the Broadway monitoring site. The top graph shows a box plot for
31      the entire range of data.  The bottom graph focuses on the interquartile range of the same data

        February 15,  1999                        3 A-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      shown in the top graph.  The horizontal line across the center of the top graph shows the current
 2      NAAQS 8-h standard for CO.  The P75, P50, and P25 in the legend for the lower graph refer to
 3      the 75-, 50-, and 25-percentiles, respectively. Figure 3 A-5 shows the reduction in CO
 4      concentrations observed for the winter months of 1986/87 through 1995/96. Also apparent from
 5      this figure is the reduction in the "spread" or variability of the 8-h average CO concentrations for
 6      wintertime from 1986/87 through  1995/96. Each circle or diamond in the top graph represents an
 7      individual 8-h "observation".  The designation of outliers (circles) or extremes (diamonds)
 8      relates to whether the observation is three or four standard deviations from the mean,
 9      respectively. It is clear from the top graph that the number of wintertime "exceedances" (the
10      occurrence of an 8-h average concentration that is greater than 9 ppm) at the Broadway site has
11      been reduced drastically in the period from 1986/87 to 1995/96.
12           Table 3A-1 shows the Kendall tau correlation coefficients for the daily maximum 8-h CO
13      concentration at the sites in the Denver MSA.  The correlation is strongest among the downtown
14      sites, which also are located the closest together. A number of factors contribute to the generally
15      low values that are seen.  These factors include the influence of local sources, differences in
16      meteorological conditions across the MSA, nonoverlapping averaging periods, and measurement
17      error. The correlation coefficients for the site pairs generally increased from the  1986/87 to
18      1995/96 winter seasons (Table 3A-1).  It is believed that, as the populated area of the  Denver
19      MSA spreads out and there is more motor vehicle traffic from increasingly distant suburbs, the
20      ambient CO profiles for these distant communities become more homogeneous.  Littleton, CO,
21      provides a good example of the increase in correlation with several of the other CO sites in the
22      Denver area. Littleton (1990 population of approximately 33,700 people) is a city located
23      approximately 20 km south of downtown Denver.  Its correlation coefficient has  shown large
24      increases, along with the Broadway and Julian Denver CO sites, as well as with the Arvada and
25      Greeley, CO, sites. These increased correlations were for the weekday data and were all of at
26      least one order of magnitude (Table 3A-1). One order of magnitude increases also were seen for
27      the weekend data between Littleton and the Arvada and Greeley sites. Arvada and Greeley are
28      approximately 25 km and 95 km, respectively, from Littleton. Therefore, it appears that the
29      increase in correlation coefficients is related to the population spreading out from downtown
30      Denver, resulting in the surrounding suburbs becoming more similar relative to traffic patterns
31      and, hence, to the distribution of the ambient CO.

        February 15, 1999                         3A-5       DRAFT-DO NOT QUOTE OR CITE

-------
         TABLE 3A-1.  KENDALL TAU SPATIAL CORRELATIONS FOR THE
    DAILY MAXIMUM 8-HOUR AVERAGE CARBON MONOXIDE DATA IN THE
                   DENVER METROPOLITAN STATISTICAL AREA
                                    1986/87 Winter Season
                                    Boulder-    Denver-   Denver-  Denver-
                   Welby  Littleton    Marine    Broadway   Albion    Julian   Arvada  Greeley
                                        WEEKDAY
 Welby

 Littleton

 Boulder-Marine

 Denver-Broadway

 Denver-Albion

 Denver-Julian

 Arvada

 Greeley
                                        WEEKEND
                                    1995/96 Winter Season
                                    Boulder-    Denver-   Denver-  Denver-
                  Welby   Littleton    Marine   Broadway   Albion    Julian    Arvada  Greeley

                                        WEEKDAY
 Welby

 Littleton

 Boulder-Marine

 Denver-Broadway

 Denver-Albion

 Denver-Julian

 Arvada

 Greeley
                                        WEEKEND
February 15,  1999
3A-6
DRAFT-DO NOT QUOTE OR CITE

-------
 1      3A.3.2  Los Angeles
 2           The Los Angeles CMS A (Consolidated Metropolitan Statistical Area) consists of five
 3      counties surrounding and including metropolitan Los Angeles (Los Angeles, Orange, Riverside,
 4      San Bernardino, and Ventura counties).  Forty-two monitors had at least one year's worth of
 5      hourly average CO data in the years from 1986 through 1996, with 12 sites (Figure 3A-6) having
 6      CO time-series data that completely covered the entire 10-year period.
 7           The monitoring data reflect pronounced variation in CO between sites throughout the
 8      Los Angeles CMSA. There are substantial differences in CO levels between sites.  Hourly
 9      average CO concentrations ranged from 0.38 to 3.16 ppm in 1996. At the Lynwood site
10      (Figure 3A-7), there are violations of the daily maximum 8-h average in every year of the
11      monitoring period analyzed here. In contrast, the Barstow site (not shown), located about
12      190 km from downtown Los Angeles, shows no violations over the same 10 years of monitoring.
13      Lynwood is located approximately 15 km south-southeast of downtown Los Angeles, about
14      halfway between downtown Los Angeles and Long Beach.  It is a densely populated area
15      intersected by several major interstate highways.
16           Trends determined from the time-series graphs range from weak downward trends (usually
17      found at those sites with the highest levels of CO, as shown in Figure 3 A-8 for the Hawthorne
18      site) to no discernable change in CO levels over the 10-year period (usually the case when CO
19      levels are already low; as in the case of Barstow, shown in Figure 3A-9).  However, the trend in
20      CO levels is not uniformly downward at any given site. The CO time-series for the wintertime,
21      weekday data for the 1989/90 winter season showed generally elevated CO levels at all sites,
22      whereas  the 1994/95 winter season had the lowest CO levels at most sites. This leads to a
23      perceived upswing in CO levels for the 1995/96 winter season. However, the year-to-year
24      fluctuations in CO do not mean necessarily that the CO concentrations were on the rise following
25      the 1995/96 winter season.  In general, there has been a decrease in the number of violations over
26      time at most of the sites.
27           Wintertime diurnal CO patterns are also quite diverse.  Many sites in the Los Angeles
28      CMSA, such as the Hawthorne site (Figure 3 A-10), show the afternoon peak beginning to rise
29      rather late in the day (after 6 p.m.), with the CO levels rising gradually at some locations  and
30      more abruptly at others, and finally peaking around midnight with elevated levels of CO
31      persisting until the morning rush hour.  Other sites, like that at El Toro, approximately 65 km
        February 15,  1999                         3A-7       DRAFT-DO NOT  QUOTE OR CITE

-------
 1      southeast of downtown Los Angeles, show a more "classic" CO pattern (see Figure 3A-11), one
 2      where the afternoon CO begins to rise between 4 p.m. and 5 p.m., peaks at about 6 p.m., and
 3      gradually but steadily declines until about 1 a.m. the following morning. The CO concentration
 4      then maintains an average daily minimum value, usually around 1 or 2 ppm, until the next day's
 5      morning rush hour.
 6           Each of the twelve sites in this analysis exhibited a wintertime maximum in CO
 7      concentrations. This wintertime peak in ambient CO is illustrated by Figure 3A-12, the monthly
 8      diurnal plots for Anaheim.  The plots for 1986/87 and 1989/90 illustrate two points quite well:
 9      (1) the pronounced rush hour peaks and (2) the distinct peak in CO during the wintertime
10      (peaking in the month of December).  In these two respects, Los Angeles shows  remarkable
11      similarity to Denver.
12           Table 3 A-2 shows the correlations for the  daily maximum 8-h average CO concentration
13      between sites for the Los Angeles CMS A. The  upper table shows the correlations for the
14      1986/87 winter season, whereas the table at the  bottom shows the correlations for the 1995/96
15      winter season. With some exceptions, the between-site correlations generally increased for each
16      site-pair in 1995/96 compared with those in 1986/87. This was analogous to the Denver site
17      correlations.  The city-pair that had the highest correlation in the 1986/87 analysis remained
18      unchanged for the  1995/96 analysis in 8 out of the 11 city-pairs in the weekday data, and in 7  out
19      of the 11 city-pairs in the weekend data.  This is indicative of the Los Angeles CMS A having
20      several distinct airsheds. The monitoring sites that "reside" within a particular air shed likely
21      will exhibit similar characteristics in ambient CO. For example, the Burbank, Hawthorne,
22      Lynwood, and Reseda monitors show a history of violations of the daily maximum 8-h average
23      over the last 5 years of monitoring data. These sites are located in heavily urbanized areas of
24      Los Angeles County, with numerous major freeways nearby. The spatial correlations among
25      these four sites are weak to moderate (range: 0.3 to 0.7, for the 1995/96 weekday data).
26      Peripheral to the area containing the urbanized monitors, are the monitoring sites in La Habra,
27      Long Beach, Anaheim, and Riverside.  These sites had violations  before 1991, but have not had
28      any violations of the daily maximum 8-h average since 1991. The last group of  sites, consisting
29      of Asuza, Barstow, El Toro, and West Los Angeles, did not have any violations  over the 10-year
30      period.
31

        February 15, 1999                         3A-8        DRAFT-DO NOT QUOTE OR CITE

-------
VO

VO

VO
H

6
o

*
o
H

O
c
o
H
W

O
&
O

H
W
          TABLE 3A-2. KENDALL TAU SPATIAL CORRELATIONS FOR THE DAILY MAXIMUM 8-HOUR AVERAGE

         CARBON MONOXIDE DATA IN THE LOS ANGELES CONSOLIDATED METROPOLITAN STATISTICAL AREA


                                                    1986/87 Winter Season
                  Azusa
West                              Long                        El

 LA    Burbank  Reseda   Lynwood    Beach   Hawthorne  Anaheim   Toro
 La

Habra   Riverside  Barstow

-------
VO
VO
VO
H

6
O

*
O
H

O
c
O
H
W

O
?d

O

H
W
        TABLE 3A-2 (cont'd). KENDALL TAU SPATIAL CORRELATIONS FOR THE DAILY MAXIMUM 8-HOUR AVERAGE

          CARBON MONOXIDE DATA IN THE LOS ANGELES CONSOLIDATED METROPOLITAN STATISTICAL AREA


                                                      1995/96 Winter Season
                   Azusa
West

 LA
Burbank   Reseda
          Long

Lynwood    Beach    Hawthorne
Anaheim
 El

Toro
 La

Habra
Riverside   Barstow

-------
 1           A notable exception to the general increase in the between-site correlations was seen in the
 2      correlations between Barstow and each of the other sites for the weekday data, where the
 3      correlations were uniformly lower in 1995/96 compared to the correlations in 1986/87.
 4      However, this site is outside the Los Angeles Basin (lying across the San Bernardino Mountains
 5      from Los Angeles),  and hence there is probably very little relationship between ambient CO in
 6      Barstow and the other Los Angeles CO sites.
 7
 8      3A.3.3 New York City
 9           The CO monitors in the New York City CMS A are located in 10 of the 30 counties
10      spanning the three states (New York, New Jersey, and Connecticut) that comprise that CMS A
11      (Figure 3 A-13).  Thirteen of the 29 monitors had time series that cover the time period from 1986
12      to 1996 and therefore were used in both the annual and seasonal data descriptions. Hourly
13      average CO concentrations in the New York CMSA ranged from 0.63 to 3.65 ppm in 1996.  The
14      CO monitors were all  in areas categorized as either urban or suburban. Despite this, the
15      wintertime, daily maximum 8-h average CO did not reach the levels seen in Denver or
16      Los Angeles (see above descriptions of these cities). One of the sites with the highest CO values
17      was that located on Flatbush Avenue in downtown Brooklyn (AIRS identification number
18      360470071, Figure 3A-14).  Yet, this site exceeded the NAAQS only twice since 1990, once in
19      November 1991 (daily maximum 8-h average = 10.9 ppm) and once in February  1995 (daily
20      maximum 8-h average = 9.8 ppm).
21           An interesting feature  of the Flatbush Avenue site is the persistence of elevated CO levels
22      after the morning rush hour, thereby making it difficult to discern between the morning and
23      evening rush hours (Figure 3 A-14).  This feature also occurred at a site in downtown Manhattan
24      (Figures 3 A-15 and  16).  The elevated CO levels between the rush hour peaks remained a feature
25      at these sites despite a steady decline in the CO concentrations with passing years at both sites,
26      and the elevated levels of CO between rush hours occurs throughout the year and are not limited
27      to a particular season or time of year. Two possible explanations for this phenomenon are
28      (1) that the sources of CO do not diminish after the "typical" rush hour has come to an end, and
29      (2) that meteorological processes that normally dilute CO concentrations in surface air during
30      midday (increased surface winds and mixing layer height) are not effective at these sites, perhaps
31      because of microenvironmental conditions such as the "urban canyon" effect. Both of these sites
        February 15, 1999                         3A-11       DRAFT-DO NOT QUOTE OR CITE

-------
 1      are located close to the approaches to major bridges spanning the East River, with heavy traffic
 2      throughout the day. In stark contrast to this finding is the observation that the CO concentrations
 3      fall off rapidly after midnight and reach minimal levels (1 to 2 ppm or less) at all of the sites
 4      investigated, even those in heavily urbanized areas.
 5           Another interesting feature of the New York CMS A data was the lack of the seasonality in
 6      CO concentrations that was seen in the Denver and Los Angeles data. Many of the sites,
 7      particularly the urban sites, did not exhibit the pronounced wintertime peak in mean CO. This
 8      can be seen in the seasonal graphs for the two urban New York sites discussed above
 9      (Figures 3A-14 and 16) and for the suburban site at Morristown, NJ (Figure 3A-17).  The reasons
10      for these differences are not clear without data to characterize seasonal variations in mixing layer
11      heights and other meteorological variables and emissions.
12           Table 3 A-3 shows the Kendall tau correlation coefficients calculated for the daily
13      maximum 8-h average CO concentration  for the New York CMS A monitoring sites for the
14      earliest year and the most recent year in these analyses. The correlation coefficients exhibit a
15      wide range (from a correlation coefficient of zero [i.e., no correlation at all] to a correlation
16      coefficient of 0.8) of values indicating heterogeneity in the 8-h maximum CO concentrations
17      among the monitoring sites.  Although the correlation coefficients generally rose from the
18      1986/87 winter season to the 1995/96 winter season, the increase was not as pronounced as those
19      in either Denver or Los Angeles, and there were numerous exceptions where  the correlations
20      decreased during the same time period.
21           Care should be taken in attempting to draw any conclusions based on correlation
22      coefficients involving any two sites. For  example, the site that has the highest correlation
23      coefficient (0.622) for the weekday data with the Flatbush Avenue site (the site in Brooklyn that
24      had some of the highest measured CO values in the New York CMS A analysis) for the 1986/87
25      winter season was located in Morristown, a suburban, mostly residential community with
26      seemingly little in common with Brooklyn. Interestingly, Morristown also had the highest
27      correlation (0.615) with Brooklyn for the  weekend data.  Then, for weekdays in the 1995/96
28      winter season, Brooklyn had the highest correlation (0.60) with Elizabeth, NJ, whereas, for
29      weekends, Brooklyn again had the highest correlation (0.54) with the Elizabeth site.  The
30      population density in  downtown Brooklyn is more like that of Elizabeth than Morristown;
31      however, they are still quite far apart. Apparently, characteristics shared by the New Jersey sites

        February  15, 1999                         3A-12       DRAFT-DO NOT QUOTE OR CITE

-------
VO
VO
VO
H

6
O

*
O
H

O
c
O
H
W

O
?d

O

H
W
        TABLE 3A-3. KENDALL TAU SPATIAL CORRELATIONS FOR THE DAILY MAXIMUM 8-HOUR AVERAGE CARBON

       	MONOXIDE DATA IN THE NEW YORK CITY CONSOLIDATED METROPOLITAN STATISTICAL AREA	


                                                           1986/87 Winter Season
Bridgeport  Stamford   Fort Lee
                                                               Jersey   Perth

                                             Hackensack  Newark    City   Amboy   Freehold  Morristown
                  7th

Elizabeth   Flatbush   Avenue  Manhattan
                                                               WEEKDAY
                                            WEEKEND

-------
                              TABLE 3A-3 (cont'd). KENDALL TAU SPATIAL CORRELATIONS FOR THE

                              DAILY MAXIMUM 8-HOUR AVERAGE CARBON MONOXIDE DATA IN THE

                              NEW YORK CITY CONSOLIDATED METROPOLITAN STATISTICAL AREA
VO

VO

VO
>

£
H

6
o

z
o
H

O
c
o
H
W

O
&
O

H
W
Bridgeport



Stamford



Fort Lee



Hackensack



Newark



Jersey City



Perth Amboy



Freehold



Morristown



Elizabeth



Flatbush



7th Avenue



Manhattan
                                                               1995/96 Winter Season
             Bridgeport   Stamford   Fort Lee  Hackensack
        Jersey    Perth

Newark   City   Amboy  Freehold  Morristown
                    7th

Elizabeth  Flatbush  Avenue  Manhattan
                                                                   WEEKDAY



                                                                    0.51    0.55
                                                           WEEKEND

-------
 1      with the Flatbush Avenue site, rather than proximity, are influencing the correlations between
 2      them.
 3
 4      3A.3.4 Phoenix
 5           The Phoenix MSA consists of Maricopa and Final counties (Figure 3A-18). These two
 6      counties include all of metropolitan Phoenix.  A total of 12 CO monitors (all were located in
 7      Maricopa county) were in the MSA during the study period; however, only five had complete
 8      time-series for the entire 10-year period.  The analysis here will focus on these five sites.  Hourly
 9      average CO concentrations ranged from 0.72 to 3.39 ppm at these sites in 1996.
10           There was a small decrease in average ambient CO levels at the five Phoenix sites from
11      1986 to 1996. At most of the sites, the morning rush hour peak showed a greater reduction than
12      did the evening peak.  In fact, only small improvement was seen in the evening rush hour peak
13      levels at any of the sites. This is illustrated by the seasonal, diurnal pattern for Central Phoenix
14      (Figure 3 A-19).  The evening rush hour peak showed little improvement, if any, from the
15      1987/88 to the 1995/96 time period. Although the sites did not exhibit a strong downward trend
16      in concentrations, they experienced a trend toward decreased variance in the distribution of their
17      CO values, as illustrated by the box plots for the East Butler Drive site (Figure 3A-20).
18      Therefore, although the median  CO values at these sites have shown either no improvement or
19      only modest decreases, the highest values have decreased substantially in the 10-year period.
20           Much like the Denver and Los Angeles data, the CO patterns in Phoenix exhibited seasonal
21      variations, with the wintertime concentrations being higher than those at other times of the year.
22      However, note that although this statement is true for the morning and evening rush hour peaks,
23      it is less true for the period of very low CO levels observed between noon and about 5 p.m.
24      Take, for example, the site in West Phoenix (Figure 3A-21). There are pronounced morning and
25      evening rush hour peaks at this site for all four time periods in this figure. Yet, during the
26      interim period, approximately between noon and 5 p.m., the CO levels are almost negligible,
27      even for the earliest period in the figure (i.e., May 1987 to May 1988). The obvious rationale is
28      that the dynamic process of turbulent mixing in the lower troposphere is diluting rapidly and
29      carrying away the ambient CO produced  during the morning rush hour.  Coinciding with the
30      effects of atmospheric dynamics on the midday CO levels is the likely scenario that the source of
31      the morning CO peak (almost certainly motor vehicle emissions) is effectively "turned off after
        February 15, 1999                        3 A-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      about 8 a.m. at this site, resulting in a drastic reduction in the measured CO levels.  This
 2      minimum of CO during midday was seen at four of the five sites, the one exception being the
 3      South Phoenix site (Figure 3 A-22). And, in fact, as can be seen in Figure 3 A-22, it was only the
 4      earliest time period (May 1987 to May 1988) where the CO levels remained elevated between the
 5      rush hour peaks.  Subsequent to that 1-year period, the midday CO levels were quite low, as
 6      described above for the West Phoenix site.
 7           The West Phoenix site (Figure 3A-21) illustrates another interesting feature of the seasonal,
 8      diurnal pattern for CO.  Similar to the pattern seen in many of the Los Angeles sites, the Phoenix
 9      sites display a wintertime, evening peak whose elevated CO levels persist throughout the night,
10      often not returning to their preevening peak levels before they begin to rise again for the morning
11      peak. Obviously, the higher the evening peak, the higher the CO levels will be as they persist
12      throughout the night, but, even for evening peak CO levels of around 5 ppm, the CO generally
13      persisted through the night, remaining above preevening peak levels. The reason that the evening
14      peak is much broader and persists much longer than does the morning peak is largely because the
15      factors  affecting the dispersion of the CO in the atmosphere (primarily the wind speed and the
16      height of the mixed layer) are reduced greatly after sunset.
17           Table 3A-4 contains the Kendall tau correlation coefficients for the Phoenix MSA
18      monitoring  sites for the earliest year and the most recent year in these analyses.  For the 10 pairs
19      of sites for the weekday data, six pairs had correlations that rose between the 1986/87 winter
20      season and the 1995/96 winter season, and four pairs remained about the same.  For the 10 pairs
21      of sites for the weekend data, only three pairs of correlations increased, five pairs decreased, and
22      two pairs remained about the same. No discernable pattern is evident in the matrix of
23      correlations in Table 3 A-4. Indications are that the Phoenix airshed is more well mixed than are
24      the others investigated  here.
25
26
27      3A.4 SUMMARY
28           An analysis of ambient CO data obtained in the four geographically diverse metropolitan
29      statistical areas of Denver, Los Angeles,  New York, and Phoenix has shown that urban CO levels
30      have decreased over the past 10 years. However, there have been instances where the downward
31      trend has reversed itself on a year-to-year basis. Although the number  of violation  days has
        February 15, 1999                        3 A-16       DRAFT-DO NOT QUOTE OR CITE

-------
     TABLE 3A-4. KENDALL TAU SPATIAL CORRELATIONS FOR THE DAILY
        MAXIMUM 8-HOUR AVERAGE CARBON MONOXIDE DATA IN THE
                  PHOENIX METROPOLITAN STATISTICAL AREA
                                  1986/87 Winter Season
                  South Phoenix   West Phoenix   East Butler Drive
                       Central
                       Phoenix
                                      WEEKDAY
 South Phoenix
 West Phoenix
 East Butler Drive
 Central Phoenix
 North Miller Road
                                       WEEKEND
                      North Miller
                         Road
                                  1995/96 Winter Season
                  South Phoenix   West Phoenix   East Butler Drive
                       Central
                       Phoenix
                                      WEEKDAY
 South Phoenix
 West Phoenix
 East Butler Drive
 Central Phoenix
 North Miller Road
                                       WEEKEND
                      North Miller
                         Road
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
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
February 15, 1999
3 A-17
DRAFT-DO NOT QUOTE OR CITE

-------
York CMS A, and 0.7 to 3.4 in the Phoenix MSA.  Carbon monoxide concentrations during the
cold season (November through February) range from 5 to 20% higher than the annual average in
each MSA. However, it should be noted that, despite decreasing CO concentrations, the nature
of the diurnal and seasonal variation observed at each monitoring site has remained remarkably
constant over the 10-year period covered in this analysis.  At all the sites investigated here, it is
clear that the diurnal and seasonal variations in CO observed in these MSAs result largely from
the interaction between motor vehicle emissions and meteorological parameters that, at times,
can be conducive to the buildup of CO near the surface. The diurnal concentration profiles in
most cases show a very distinctive two-peaked structure for weekdays. The peaks correspond to
both the morning and evening rush hour commutes.  Frequently, the morning peak is higher than
the evening peak at any given site because the height of the mixed layer is much lower during the
morning, thus inhibiting vertical mixing that would have diluted CO.  In the late afternoon and
into early evening, increased atmospheric turbulence resulting from solar heating raises the
height of the mixed layer, resulting in generally lower CO concentrations compared with those of
the morning.
     Regional differences in atmospheric processes  also may play a role in producing the
nighttime behavior of CO observed at numerous sites in the Los Angeles and Phoenix MSAs
compared with either the nationwide  composite average diurnal cycle of CO shown in
Figure 3 A-6 or other locations, such as the Denver or New York MSAs. In the Los Angeles and
Phoenix metropolitan areas CO concentrations often remain until midnight at levels  reached
during the evening rush hour.  Then,  although these concentrations gradually diminish
throughout the night, they do not drop to the low afternoon levels (typically no more than  1 to
2 ppm, often less than this amount) before they begin to increase again because of the morning
rush hour. This pattern is shown quite well in Figure 3 A-19, which depicts the seasonal, diurnal
concentration profile for the Central Phoenix site.
     In general, the highest values of ambient CO were found during the wintertime (defined as
the months of November through February, inclusive) in all of the MSAs included here. There
were a few sites in the New York Metropolitan Area where a wintertime peak in CO was not
discernable; the site on Flatbush Avenue in Brooklyn (Figure 3A-14) is an excellent example of
this. It is not clear without further analysis, what combination of seasonal variations in emissions
and meteorological parameters gave rise to this result.

February 15, 1999                         3 A-18       DRAFT-DO NOT QUOTE OR CITE

-------
     There were large variations in the Kendall tau correlation coefficients calculated for the
daily maximum 8-h average CO concentrations between individual sites in a given MSA during
the winters of 1986/87 and 1995/96, suggesting a high degree of heterogeneity in the daily
maximum 8-h average ambient CO levels in the MSAs that were characterized, likely indicating
that daily maximum 8-h average CO levels observed at particular monitoring sites may not be
related to CO levels occurring some distance away from the monitoring site.  Further analyses
will determine whether the correlation coefficients will increase significantly with an increase in
averaging time to 24 h.  Based on the above results, caution should be exercised in using the
daily maximum 8-h average data to characterize the exposure of the general population within a
given MSA (e.g., for studies relating health outcomes to ambient CO levels or in health studies in
which CO may be viewed as a confounding variable).
February 15, 1999                         3 A-19       DRAFT-DO NOT QUOTE OR CITE

-------
REFERENCES

Shadwick, D.; Glen, G.; Lakkadi, Y.; Lansari, A.; del Valle-Torres, M. (1997) Analysis of carbon monoxide for the
      Denver, Colorado MSA. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
      Research and Development, National Exposure Research Laboratory; contract no. 68-D5-0049; December.
Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998a) Analysis of carbon monoxide for the Los Angeles, California
      CMS A. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and
      Development, National Exposure Research Laboratory; contract no. 68-D5-0049; June.
Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998b) Analysis of carbon monoxide for the New York, New York
      CMS A. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and
      Development, National Exposure Research Laboratory; contract no. 68-D5-0049; June.
Shadwick, D.; Glen, G.; King, J.; Chen, X. (1998c) Analysis of carbon monoxide for the Phoenix, Arizona MSA.
      Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and Development,
      National Exposure Research Laboratory; contract no. 68-D5-0049; June.
February 15, 1999                            3A-20        DRAFT-DO NOT QUOTE OR CITE

-------
                          .'4.
                                             Greeley
                       *
                Boulder>Marine
           Arvada
                        i© ©
                 Welby  r   "^
                          •      r~
                 i      ;••'"•'      '•
                 (P r.,    J ......
             ": Dehyer-Bro^dway
       Denver-Julian*
                             ^Dfenver-Albion
                               fc*-» /r"
                                t T.!


                               -Littleton
    Population Density
        0 to 300 porsonsl'km2
        301 to 1000 person s.!'km2
        1001 to2000 persons'k!Ti2
      | 2001 to 5000 persons/km2
    Hi 5001 *° 10,000 persons/kma
      I over 10,000 persons,'km2
                  CO Monitoring Sites
                 . Counties
                  States
                  Interstates/Divided Highways
                  Major Roads
                  Water
   10
10
20 30km
                                     -M cam
                                     u
                                      DM     ProdLKWl Ncfmirtxf 1 Mh. 1«S
Figure 3A-1. Map of Denver showing locations of CO monitoring sites.

February 15, 1999                 3A-21      DRAFT-DO NOT QUOTE OR CITE

-------
                                               I  'V
                                Color Seal* In MM
Figure 3A-2. 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 mixing ratios in ppm.
                               >  «  B I  i  §
                              Color Scale in PFW
Figure 3A-3. 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 mixing ratios in ppm.
February 15, 1999
3A-22
DRAFT-DO NOT QUOTE OR CITE

-------
Figure 3A-4. 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 mixing ratios in ppm.
             1986
                  1987
                        1988
                             1989
                                   1990  1991

                                    YEAR
                                              1992
                                                    1993
                                                         1994
                                                               1995
                                                                      n  Median
                                                                      o  Outliers
                                                                      o  Extremes
         10
        H  6

        I?
        Z  o
        LU  °
        O  2
        o  o
---n	n	a-
-••-.c	o	o-
                                                                     --H--P50
          1985   1986   1987  1988  1989
                                   1990  1991
                                     YEAR
                                             1992  1993  1994   1995   1996
Figure 3A-5.  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.
February 15, 1999
      3A-23
DRAFT-DO NOT QUOTE OR CITE

-------
         I
                                                          V

                                                          Barstowt?
           ^AjPHfc
        ©Reseda
           -,~-
              *- aeurbank   -

                                      1—.r
                                      d*      -^&
                            «
                  Los AngeN*
     Population Density
          0 to 300 persons/km2
          w i\j jvu |jci -a"uiri i-a^ F\Iiij.        •__i—
          301 to 1000 porsons/km2     [_] statas
     CO Monitoring Sites
/'; / Counties
          1001 to 2000 personsJ'km2
          2001 to 5000 persons/km2
          5001 to 10,000 per5ons/hm2
          over 10,000 pcrsons.'kfn2
     tnterstates/Divided Highways
     Major Roads
     Water
     10  0   10  20  3040km

ISrVlSl
 jLUJJllill lllll-li
                                              E. i. J r.rr
                                              4DD
Figure 3A-6.  Map of Los Angeles showing locations of CO Monitoring sites.
February 15, 1999
3A-24
DRAFT-DO NOT QUOTE OR CITE

-------
                                                               I ,T
                            o>  i  a  a  4  s
                             Color  Scale in. PHS
  Figure 3A-7. 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 concentrations in ppm.
                1986
                     1987   1988
                               1989
                                    1990  1991
                                      YEAR
                                               1992
                                                    1993
                                                         1994   1995
                                                                    a  Median
                      1987  1988   1989  1990  1991   1992  1993  1994   1995
                                      YEAR
                                                                    —o— P75
                                                                    --D-- P50
                                                                    -o—P25
Figure 3A-8.  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.
 February 15, 1999
3A-25
DRAFT-DO NOT QUOTE OR CITE

-------
12
10
8
6
4
2
0

^i*S$$5$iaJ
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
G Median
o Outliers
YEAR
              10
                                    1990   1991
                                      YEAR
                                             1992  1993
Figure 3A-9. 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.
                            0121456
                             COlcc Scale in PM

Figure 3A-10. 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 concentrations in ppm.
February  15, 1999                       3A-26       DRAFT-DO NOT QUOTE OR CITE

-------
                                                         IVM
Figure 3A-11. 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 concentrations in ppm.
                                                1969- 1*»
                            14-IM
                                                   1996
Figure 3A-12. 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.
February 15, 1999
3A-27
DRAFT-DO NOT QUOTE OR CITE

-------
               7f\       '?
     \      f    *7
       *      p       •_    ^' -.
       ',  X'V
     * v ^f
          ^% -*
             '*"•.  -F
     ^          '>"-.r..
     Population Density
         0 to 300 person&/krn2
         301 to 1000 persons/kmS
         1001 to 2000 persons/km2
        | 2001 to 5000 persons/km2
        | 5001 to 10,000 person5'km2
         over 10,000 person5,'km2
    CO Monitoring Sites
    f Counties
    States
    Interstates/Divided Highways
    Major Roads
    Water
                      20  30km
                                            QH
                                            uH
                       Nnwntw Iftlh. 1Wfl
Figure 3A-13.  Map of New York showing locations of CO monitoring sites.
February 15, 1999
3A-28
DRAFT-DO NOT QUOTE OR CITE

-------
                           1966-
                           1992- I&93
                                               IMS, 1996
                              123456
                            Color Seal*  in KM
         6 9 10
Figure 3A-14. 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 3A-15. 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 concentrations in ppm.
February 15, 1999
5A-29
DRAFT-DO NOT QUOTE OR CITE

-------
                                                1W 1HO
                                1  2  3  4  E
                                   scolc in Fm
 Figure 3A-16. 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.
                                                   19W
                            1W2. 1993
                                                19SS- 19K
                                 a a  4567
                                  sc*i* in pm
              IP
Figure 3A-17. 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.
February 15, 1999
3A-30
DRAFT-DO NOT QUOTE OR CITE

-------
                                     _
                     E. Butler Dr.
      West Pho

      Central Phoenix
                                      N.  Miller Rd.
                South Phoenix
                              if
                   .—k
   Population Density             ® co Monitoring Sites
       0 to 300 persons/kfn2      / V Counties
       301 to 1000 person s/km2    I   I States
       1001 to 2000 persQns/km2       Interstates/Divided Highways
   HI 2001 to 5000 persons,'km2       Major Roads
       5001 to 10,000 persons.''km2   • Water
       over 10.000 personsj'km2               	
    10
                                     HW4CW
                                     M ItMl
                                                      1«fl
Figure 3A-18. Map of Phoenix showing locations of CO monitoring sites.
February 15, 1999
3A-3 1
                                     DRAFT-DO NOT QUOTE OR CITE

-------
                              ;i't.
 Figure 3A-19. 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.
           I 10
           fc  8
           |  6

           I  4
                1986
                     1987
                          1988
                               1989
                                    1990  1991
                                      YEAR
                                              1992
                                                   1993
                                                        1994
                                                             1995
                                                                    Q  Median
                                                                    o  Outliers
- — - I U
9
fe 8
Z 7
0
i= 6
< F
I 5
ii
0 1
o n

.



-
^r^ — "--^^-^^
- --^^^^^=:=r^^^^^^^^

: . i j i . , .
1BRR 1SR7 1flRR IflRB 1flnn 19B1 19fl9 mfl?i 1Sfl4 ICIflS







^^ P75
--D-- P50
—*>-• P25
                                      YEAR
Figure 3A-20. 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.
February 15, 1999
3A-32
DRAFT-DO NOT QUOTE OR CITE

-------
                           •*SB5~i
                            **&:. IB
"ti£
                               i«a- mn
                                                 199HSM
                               0123155
                               Cbior Scale in PPM
 Figure 3A-21. 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.
                              L9B7.
                             1WS 1593
                                                USSS- L9K
                                                    .:.
                                   Seal* ID  PM
Figure 3A-22. 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.
February 15, 1999
           3A-33
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,

        February 15, 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 potential source
 7      of CO, 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; Federal Register, 1992; Last
10      et al.,  1995; Zartarian et al., 1997). A recent definition offered by Zartarian et al. (1997) states
11      that exposure is the contact between an  agent and a target at a specified contact boundary,
12      defined as a surface in  space containing at least one exposure point (a point at which contact
13      occurs).  Under this definition, exposure can occur instantaneously at a single point in space, or it
14      can occur over time or space. For purposes of reviewing the literature in this chapter, the target
15      is the human, the agent is CO, and the contact boundary is assumed to be the lining of the lung
16      where CO exchange takes place between the air and the blood.
17           This chapter is concerned only with CO exposures that occur at concentrations capable of
18      increasing COHb levels above a reference baseline level.  Besides  exposure to CO concentrations
19      above the background level,  human COHb levels can be elevated because of metabolic
20      degradation of many drugs, solvents (e.g., methylene chloride), and other compounds to CO.
21      For details see Section 5.3. The maximum COHb level resulting from an exposure to methylene
22      chloride has twice the half-life of the identical maximum COHb  level produced by direct
23      exposure to a CO molecule (Wilcosky and Simonsen,  1991).  Therefore, people exposed to
24      methylene chloride will be more susceptible to reaching any given COHb concentration resulting
25      from CO exposure.  Hence, the literature on exposure to methylene chloride also is discussed in
26      this chapter.
27           The chapter is organized as follows. The first section summarizes the state of knowledge
28      on population exposure to CO as of 1991, when the U.S. Environmental Protection Agency
29      (EPA) published the last CO air quality  criteria document (AQCD).  This is followed by a
30      discussion of more recently published studies of population exposure to all sources of CO, except
31      those related to active inhalation of tobacco smoke.  This discussion describes typical levels of

        February 15,  1999                          4-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposure as people engage in daily activities, including those related to an occupation. Next is a
 2      description of factors affecting trends in population exposure to carbon monoxide.  These factors
 3      include public policies affecting motor vehicle emissions, travel behavior, and smoking and
 4      societal changes in human activity patterns.  The conclusion summarizes findings of this review
 5      and discusses their implications for CO exposure models such as the probabilistic NAAQS
 6      Exposure Model.
 7
 8
 9      4.2  BRIEF SUMMARY OF POPULATION EXPOSURE STUDIES
10           PRIOR TO 1991
11           This section briefly reviews key population exposure studies that were completed by 1991.
12      It identifies populations who are sensitive to CO exposure, discusses studies of population
13      exposure based on fixed-site and personal monitors, and relevant population exposure models.
14
15      4.2.1 Sensitive Populations
16           The NAAQS are intended to protect certain sensitive and probable risk groups of the
17      general population.  These groups differ from one air pollutant to another. In the case of CO,
18      these groups include anemics; the elderly; pregnant women; fetuses; young infants; and those
19      suffering from certain blood, cardiovascular,  or respiratory diseases. People currently thought to
20      be at greatest risk from exposure to ambient CO levels are those with ischemic heart disease who
21      have stable exercise-induced angina pectoris  (cardiac chest pain).  Individuals with this disease
22      represented about 3% of the U.S. population in 1994. Studies show that earlier time to onset of
23      cardiac chest pain occurred in these people while they exercised during exposures to CO
24      concentrations that produced levels of COHb in the bloodstream in the range of 2 to 3% (U.S.
25      Environmental Protection Agency, 1991). The National Health and Nutrition Examination
26      Survey (NHANES) n study reported that 6.4% of the U.S. population who never smoked had
27      COHb levels above 2.1%, based on a national random sample of people (n = 3,141) ranging in
28      age from 12 to 74 years (Radford  and Drizd,  1982).  The NHANES H study was done in the late
29      1970s when ambient CO concentrations were much higher (see Figure 3-3).
30
31
        February 15,  1999                         4-3         DRAFT-DO NOT QUOTE OR CITE

-------
 1     4.2.2  Estimates of Population Exposure Based on Fixed-Site Monitors
 2          In the United States, attainment of the NAAQS is based on ambient air quality
 3     measurements recorded at a nationwide network of fixed-site monitors.  Based on this network,
 4     the Office of Air Quality Planning and Standards of EPA estimated that 12.7 million people lived
 5     in seven counties where CO levels exceeded the NAAQS in 1996 (U.S. Environmental
 6     Protection Agency, 1998a). The estimate was made by combining census data on county
 7     populations with data on violations of the CO NAAQS recorded by stationary monitors.
 8     Previous studies have shown that such estimates are insufficient for estimating population
 9     exposure to CO for the two reasons discussed below:
10     (1) Ambient CO concentrations are not spatially homogeneous within the area monitored. For
11         example, Ott and Eliassen (1973) reported average CO levels ranging from 5.2 to 14.2 ppm
12         for sidewalks along congested streets of downtown San Jose, CA.  Corresponding CO
13         averages at fixed-site monitors were only 2.4 to 6.2 ppm. A decade later, Ott and Flachsbart
14         (1982) found a narrower gap between simultaneous CO measurements from fixed-site and
15         personal exposure monitors deployed at indoor and outdoor commercial settings in five
16         California cities.
17     (2) In the absence of indoor CO sources and immediate outdoor sources (i.e., idling motor
18         vehicles), indoor CO concentrations tend to equal outdoor concentrations over the long term.
19         In buildings with mechanical ventilation systems, the timing and scheduling of outdoor
20         "make-up"  air into the building affects ratios of indoor-outdoor concentrations both in the
21         short and long term (Yocom, 1982).  For example, when make-up air was introduced into an
22         air-conditioned building during morning rush hours (when outdoor CO levels were high),
23         indoor CO concentrations exceeded outdoor levels for the remainder of the day (Yocom
24         et al., 1971). In the presence of indoor sources such as gas appliances, indoor CO
25         concentrations often exceed the outdoor levels (U.S. Environmental Protection Agency,
26         1991). Many American homes use gas (natural gas and liquid propane) for space heating,
27         cooking, heating water, and drying clothes. In a 1985 Texas study of a low-socioeconomic
28         population, CO concentrations were greater than or equal to 9 ppm in  12% of surveyed
29         homes.  Residential CO concentrations were high where unvented gas space heaters were
30         used as the  primary heat source (Koontz and Nagda, 1988).

       February 15, 1999                          4-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      These facts take on added significance given that many Americans spend most of their time
 2      indoors (Szalai, 1972; Chapin, 1974; Meyer, 1983; Johnson, 1987; Schwab et al., 1990). Hence,
 3      studies of actual personal exposure to CO, rather than crude estimates of population exposure to
 4      ambient CO, are necessary to determine what risk CO poses to public health (Sexton and Ryan,
 5      1988).
 6
 7      4.2.3 Surveys  of Population Exposure Using Personal Monitors
 8           With the development of personal exposure monitors (PEMs) in the 1970s, researchers
 9      began to measure either the total human exposure of a population or the exposures of
10      subpopulations in microenvironments that posed higher risks of CO exposure, such as inside a
11      motor vehicle moving slowly in congested traffic. In theory, a microenvironment exists if the
12      CO concentration at a particular location and time is sufficiently homogeneous yet significantly
13      different from the concentrations at other locations (Duan, 1982).
14           Human exposure studies of target populations typically use either a direct or an indirect
15      approach. In the direct approach, PEMs are distributed either to a representative or
16      "convenience" (nonrandom) sample of a human population. Population exposure parameters
17      cannot be estimated from a convenience sample, because it does not represent the population
18      from which it was drawn. Using PEMs, people record exposures to selected air pollutants as
19      they engage in their regular daily activities. In the indirect approach, trained technicians use
20      PEMs to measure pollutant concentrations in specific microenvironments or populations.  This
21      information then must be combined with additional data on human activity patterns to estimate
22      the time spent in those microenvironments.
23           Sexton and Ryan (1988) discuss types of personal monitors and research methods used by
24      the direct and indirect approaches. Although small passive monitors may be placed near a
25      person's oral/nasal cavity where exposure actually occurs, larger monitors must be carried by a
26      person or placed nearby. Using data  from PEMs, one  can construct exposure-time profiles for a
27      particular activity such as commuting or the integrated exposure between two points in time.
28      From this information, one can determine the average concentration to which a person has been
29      exposed for a given time period. Based on the superposition principle, one also can determine a
30      net microenvironmental concentration by subtracting the  outdoor concentration, as measured by
31      an appropriate fixed-site monitor, from a microenvironmental concentration measured by a
        February 15, 1999                         4-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1      personal monitor.  Because ambient CO concentrations are not spatially homogeneous at any
 2      given moment, the net microenvironmental concentration can be either positive or negative in
 3      value. A negative net value can occur, for example, in homes with no CO sources during
 4      morning periods when ambient CO concentrations from rising traffic emissions on highways
 5      have not yet diffused into residential areas. A negative net value simply indicates that the
 6      microenvironment has a lower positive CO concentration than the outdoor environment at a
 7      given moment.
 8           In an early pilot study in Los Angeles, using the direct approach, subjects recorded their
 9      exposures and corresponding activities in diaries (Ziskind et al., 1982). Because this was
10      cumbersome and potentially distorted  the activity, later studies used data loggers to store
11      concentrations electronically, as done  by major studies of the urban populations of Denver, CO,
12      and Washington, DC (Akland et al., 1985).  In these studies, subjects still used diaries to record
13      pertinent information about their activities in specified microenvironments while monitoring
14      personal exposures.  Data were then transferred electronically from data loggers and manually
15      from diaries to computer files for analysis.
16           The direct approach, which uses the total exposure assessment methodology, provides a
17      frequency distribution of air pollutant  concentrations for a sample of people, selected  randomly
18      from either a general or a specific population (defined by demographic, occupational and health-
19      risk factors) for a particular time period of interest (e.g., a day).  Studies using the direct approach
20      enable researchers to assess what percentage of a large population is exposed to pollutant
21      concentrations in excess of ambient air quality standards (Akland et al., 1985). Studies using the
22      indirect approach may focus on situations that bring large numbers of people in contact with high
23      concentrations in specific microenvironments.  For example, Flachsbart and Brown (1989)
24      determined what percentage of employees were exposed to CO concentrations in excess of
25      national  and state ambient air quality standards at a large shopping center attached to a parking
26      garage in Honolulu,  HI.
27           Direct studies of general populations are rare because of their expense and the logistical
28      problems of monitor distribution. Two examples for CO were those done in Denver and
29      Washington during the winter of 1982-1983 (Akland et al., 1985).  In both studies, the target
30      population included  noninstitutionalized, nonsmoking residents,  18 to 70  years of age, who lived
31      in the city's metropolitan area. This was estimated to be  about 1.2 million adults in Washington

        February 15, 1999                          4-6         DRAFT-DO NOT QUOTE  OR CITE

-------
 1     and 500,000 in Denver. In both cities, the composite network of fixed-site monitors
 2     overestimated the 8-h exposures of people with low-level personal exposures and underestimated
 3     the 8-h exposures of people with high-level personal exposures. With respect to the
 4     underestimates, over 10% of the daily maximum 8-h personal exposures in Denver exceeded the
 5     NAAQS of 9 ppm, and about 4% of maximum 8-h personal exposures in Washington exceeded
 6     9 ppm.  The end-expired breath CO levels were in excess of 10 ppm, which is roughly equivalent
 7     to about 2% COHb in about 12.5% of the Denver participants and about 10% (after corrections
 8     were made for instrumental measurement drift) of the Washington participants.  Simultaneous
 9     CO measurements at fixed-site monitors exceeded 9 ppm only 3% of the time in Denver and
10     never exceeded 9 ppm in Washington (Akland et al., 1985).
11          The Denver and Washington studies identified certain activities associated with higher CO
12     exposures.  The two highest average CO concentrations occurred when subjects were inside a
13     parking garage and when traveling by car. Those who commuted 6 h or more per week had
14     higher average exposures than those who commuted less than 6 h per week.  Higher mean CO
15     concentrations occurred for travel by  motor vehicle (car, bus, truck, etc.) than by walking. High
16     exposures also were traced to indoor and occupational sources involving use of or proximity to
17     gas appliances and gasoline engines.  High indoor concentrations above the 8-h NAAQS of
18     9 ppm occurred in public garages, service stations, and motor vehicle repair facilities. Denver
19     had higher average CO concentrations than Washington for all microenvironments because of
20     Denver's higher altitude and colder winter climate (Ott et al., 1992).
21
22     4.2.4  Population Exposure Models
23          Many studies developed models to predict exposure in both general and special populations
24     (U.S. Environmental Protection Agency, 1991). These models are important because it is
25     impossible and impractical to know the hourly and daily exposure of every person in a
26     population on a real-time basis. Models of human exposure are empirically derived
27     mathematical relationships, theoretical algorithms, or hybrids of these two. To support policy
28     decisions related to the setting of ambient and emission standards, EPA supported development
29     of two general population exposure models: (1) the NAAQS Exposure Model (NEM) and (2) the
30     Simulation of Human Activity and Pollutant Exposure (SHAPE) model.  These models assume
31     that an individual's total CO exposure over a specified time interval can be estimated as the sum
       February 15,  1999                         4-7         DRAFT-DO NOT QUOTE OR CITE

-------
 1      of the average concentration within a microenvironment, multiplied by the amount of time spent
 2      in that microenvironment (Duan, 1982).
 3           The SHAPE model used a stochastic approach to simulate the exposure of an individual
 4      over a 24-h period (Ott, 1984).  The model replicates a person's daily activity pattern by sampling
 5      from  probability distributions representing the chance of entry, time of entry, and time spent in
 6      22 different microenvironments. Transition probabilities determine a person's movement from
 7      one microenvironment to another.  The model assumes that microenvironmental concentrations
 8      reflect the contribution of an ambient concentration and a component representing CO sources
 9      within each microenvironment. Because SHAPE relies on field surveys of representative
10      populations, the data requirements of the model are fairly extensive.
11           The SHAPE model can estimate the frequency distribution of maximum standardized
12      exposures to CO for an urban population and the cumulative frequency distribution of maximum
13      exposures for both 1-h and 8-h periods, thereby allowing estimates of the proportion of the
14      population that is exposed to CO concentrations above the NAAQS.  An evaluation of SHAPE
15      by Ott et al. (1988), using survey data from the aforementioned Denver study, showed that the
16      observed and predicted arithmetic means of the 1-h and 8-h maximum  average CO exposures
17      were  in close agreement.  However, SHAPE overpredicted low-level exposures and
18      underpredicted high-level exposures.
19           Compared to SHAPE, which focuses on an individual's exposure, NEM has a higher level
20      of aggregation based on cohorts of people. The NEM model has evolved over time from
21      deterministic to probabilistic versions.  As described elsewhere (Johnson and Paul, 1983; Paul
22      and Johnson, 1985), the deterministic version of NEM simulates movements of selected groups
23      (cohorts) of an urban population through a set of exposure districts or neighborhoods and through
24      different microenvironments. Cohorts are identified by district of residence and, if applicable,
25      district of employment, as well  as by age-occupation group and activity pattern subgroup. The
26      NEM uses empirical adjustment factors for indoor and in-transit microenvironments, and
27      accumulates exposure over 1 year. Although deterministic NEM was able to estimate central
28      tendencies in total exposure accurately, it did less well estimating the associated uncertainty
29      caused by variation in time spent in various microenvironments (Quackenboss et al., 1986) or to
30      variation in microenvironmental concentrations (Akland et al., 1985). Paul et al.  (1988)
31      discussed advancements in the deterministic version of NEM.

        February 15, 1999                          4-8         DRAFT-DO NOT QUOTE OR CITE

-------
1            In recent years, EPA developed the probabilistic NEM for CO (pNEM/CO); see Johnson

2      et al. (1992) for a description of the assumptions and algorithms of pNEM/CO, as those details

3      are beyond the scope of this chapter.  Figure 4-1 shows the conceptual overview of the logic and

4      data flow of the pNEM/CO model. It shows how any alternative CO NAAQS can be evaluated

5      by establishing the distributions of personal exposures to CO when that alternative CO standard

6      is met.  The inputs to the model (e.g., activity patterns, ambient monitoring data, air exchange

7      rates) are in the round-cornered boxes, and the model calculations are shown in the other boxes.
         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).
       February 15, 1999
                    4-9
     DRAFT-DO NOT QUOTE OR CITE

-------
 1          McCurdy (1995) examined the history of both the NEM and pNEM models and the role
 2     they have played in reviews of criteria air pollutants such as CO. The EPA used pNEM/CO in its
 3     previous review of the CO NAAQS (U.S. Environmental Protection Agency, 1992). At the
 4     request of the Clean Air Science Advisory Committee, EPA evaluated the predictions of
 5     pNEM/CO against observed data for subjects of the Denver CO study. Based on this evaluation,
 6     Law et al. (1997) reported the predicted and observed population exposure cumulative frequency
 7     distributions (CFD), with and without gas stove use.  Figures 4-2 and 4-3 show that, regardless of
 8     gas stove use, pNEM/CO overpredicted the CFD at low exposures and underpredicted the CFD
 9     at high exposures for 8-h daily maximum exposures. Similar results were reported for the 1-h
10     daily maximum exposures, with and without gas stove use.  Relatively close agreement between
11     simulated and observed PEM data occurred for CO concentrations near the average exposure,
12     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.
13          In summary, there is a need to improve the abilities of both SHAPE and pNEM to predict
14     high-end exposures. This perhaps could be achieved by introducing autocorrelation of inputs for
15     time and concentration for nonindependent microenvironments (e.g., the commuting time and
16     CO exposure from home to work is not independent of the commuting time and CO exposure
17     from work to home).  Also, both SHAPE and pNEM could consider CO emission sources,
18     microenvironments and activity patterns that may have been excluded from the models.
19
20
21     4.3   SURVEY OF RECENT EXPOSURE STUDIES OF NONSMOKERS
22          This section discusses some population exposure studies that were excluded from the
23     1991 CO AQCD and studies that were published or accepted for publication in the scientific,
24     peer-reviewed literature since then.  The section is divided into nonoccupational and
25     occupational subsections.
26
27     4.3.1  Nonoccupational Exposures
28          This subsection focuses on nonoccupational CO exposures that occur because of a variety
29     of human activities that require contact with sources of CO emissions such as motor vehicles and
30     fuel-burning tools and appliances.  Section 3.5 discusses studies where CO has been measured by

       February 15, 1999                        4-10       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
area monitors in indoor microenvironments because such measurements constitute an indirect


exposure estimate. In addition, this section discusses studies of breath CO in populations and


studies of exposure to methylene chloride because it increases COHb levels.
      February 15, 1999
                                        4-11
DRAFT-DO NOT QUOTE OR CITE

-------
           Q.
                50'



                40 -




                30-







                20'
—    4 —
           C
           g

           "03   10-



           §
           c
           o
           O
           0)
           73

           "I

           o
            c
            o
           .a

            ro
           O
                1 -
               0.5
                                         Observed Exposure (n = 188)

                                         Simulated Exposure
                                                                        i      i
                                                                       99.9   99.99  99.999
                    iiiiriiiii[
                   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     4.3.1.1 Exposure to Carbon Monoxide from Motor Vehicles


2          This subsection presents nonoccupational studies of exposure to CO concentrations from


3     motor vehicles. Because these studies date to the mid-1960s, many of them were reviewed in the


4     1991 CO AQCD, and, therefore, they are not reviewed here. The studies in this section
      February 15, 1999
                                  4-12
DRAFT-DO NOT QUOTE OR CITE

-------
 1      examined the CO exposure of motorists and bicyclists. The first study examined passenger
 2      exposure to CO from air bags.
 3           Air bags that use explosives for rapid deployment are a required safety feature on cars sold
 4      in the United States.  Based on deployment of four different air bags, Wheatley et al. (1997)
 5      reported that the time-weighted-average (TWA) CO concentration ranged from 174 to 370 ppm.
 6      Peak CO concentrations occurred 2 min after deployment.  Several studies done prior to 1991
 7      reported passenger exposure to engine or tailpipe emissions of CO (Amiro, 1969; Clements,
 8      1978; Ziskind et al., 1981). More recently, Hampson and Norkool (1992) reported that 20 of
 9      68 children were treated for accidental CO poisoning after they rode in the back of pickup trucks
10      in Seattle, WA.  All trucks had exhaust systems with previously known leaks or tailpipes that
11      exited at the rear rather than the side of the truck.  In 17 trucks, children rode under a rigid closed
12      canopy attached to the bed, and in the other three trucks they rode beneath a tarpaulin.  Average
13      COHb levels measured in an emergency room were  18.2% ± 2.4% (mean, plus or minus the
14      standard error of the mean) and ranged from 1.6 to 37.0%.
15           Studies done both before and after 1991 continue to show that fixed-site monitors
16      underestimate in-vehicle CO exposures. Flachsbart (1995) reported that 14 of 16 in-vehicle
17      exposure studies performed in the United States between 1965 and 1992 simultaneously
18      measured both ambient and passenger cabin concentrations. Regardless of the study,
19      Figure 4-4 shows that the mean CO concentrations inside vehicles always exceeded the mean
20      ambient CO concentrations measured at fixed-site monitors. The lines connecting points on the
21      upper and lower lines of Figure 4-4 do not imply relationships between results for different cities.
22      The lines are shown to make a clear distinction between exposure data from indirect studies,
23      represented by the upper line, and the ambient data from  indirect studies, represented by the
24      lower line. The  ratio between a study's mean in-vehicle CO concentration and its mean ambient
25      CO concentration generally fell between 2 and 5 for most studies, regardless of when the study
26      was done, but exceeded 5 for two studies done during the early 1980s. Of the more recent
27      studies, Chan et  al. (1991)  found that median CO concentrations were 11 ppm inside test vehicles
28      driven on hypothetical routes in Raleigh, NC, during August and September 1988, but median
29      ambient concentrations were only 2.8 ppm at fixed-site monitors. Fixed-site  samples were
30      collected about 100 to 300 ft away from the midpoint of each route.
31

        February 15, 1999                         4-13        DRAFT-DO NOT QUOTE OR CITE

-------
            Q.
            a.

            c
            o
           •jz;
            03
           -i—'
            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 4-4.  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           Flachsbart (1995) proposed that the results of the 16 studies could be explained partly by

2      different study approaches (direct versus indirect) and by other aspects of study design, including

3      choice of city, season of the year, the surveyed road's functional type and location, travel mode,

4      and vehicular ventilation.  For example, by pairing three direct studies with three indirect studies

5      (as shown in Figure 4-4) that were done at the same time, Flachsbart showed that the mean
       February 15, 1999
4-14
DRAFT-DO NOT QUOTE OR CITE

-------
 1      in-vehicle exposure measured by the direct approach was always lower than that measured by the
 2      indirect approach.  Although direct studies sampled real populations engaged in a variety of trips
 3      in all types of traffic, most indirect studies focused on hypothetical commuters with higher
 4      exposures in rush hour traffic. In another example, a comparison of Studies 8 through 10 in
 5      Figure 4-4 shows the effect of roadway type. Study 9 had a sizeable component of residential
 6      driving, which may explain why the mean in-vehicle CO exposure of 7.7 ppm for Study 9 was
 7      lower than the mean exposures for Studies 8 and 10.  When the data of Study 9 were
 8      disaggregated by roadway location, the mean CO concentrations were 10 ppm for major
 9      commuting routes and 5.5 ppm for drives in residential areas.  The mean concentration of
10      10 ppm for major commuting routes in Study 9 is similar to the mean CO concentrations reported
11      for arterial highways by other studies (i.e., 9.8 ppm for Study 8 and 10.6 ppm for Study 10),
12      which also were done during the early 1980s.
13           Like earlier studies, recent ones also have looked at the effect of different routes and travel
14      modes on CO exposure.  Chan et al. (1991) reported significantly different in-vehicle exposures
15      to CO for standardized drives on three routes that varied in traffic volume and speed. The
16      median in-vehicle CO concentration was 13 ppm for 30 samples in the downtown area of
17      Raleigh, which had heavy traffic volumes, slow speeds, and frequent stops. The next highest
18      concentrations (median =11  ppm, n = 34) occurred on an interstate beltway that had moderate
19      traffic volumes and high speeds, and the lowest concentrations (median = 4 ppm, n = 6) occurred
20      on rural highways that had low traffic volumes and moderate speeds. Similarly, Dor et al. (1995)
21      reported CO exposures of 12 ppm for 19 trips lasting an average of 82 min on a route through
22      central Paris, France, which was 2 to 3 ppm higher than the mean exposure for 30 trips split
23      between two suburban routes. In terms of travel modes, both Joumard  (1991) and Dor et al.
24      (1995) found differences in CO exposures for public and private modes of travel in French cities
25      and towns. Their findings confirmed those made earlier in the United States by Flachsbart et al.
26      (1987).
27           The CO exposure of cycling as a travel mode has been studied and compared to the
28      exposure of motorists. In England, Bevan et al. (1991) reported that the mean CO exposure of
29      cyclists in Southhampton was 10.5 ppm, based on 16 runs over two 6-mi routes that took an
30      average of 35 min to complete. Keep in mind that the CO exposures of European cyclists may
31      not be comparable  to the exposures of cyclists in the United States because the installation of

        February 15, 1999                         4-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      catalytic converters on new cars in Europe occurred in 1988, about 13 years after their
 2      introduction in the United States (Faiz et al., 1996). In The Netherlands, Van Wijnen et al.
 3      (1995) compared exposures of volunteers serving as both car drivers and cyclists on several
 4      routes in Amsterdam during winter and spring.  For a given route, the mean personal 1-h
 5      concentrations were always higher for car drivers than for cyclists regardless of when sampling
 6      occurred during the year. However, a volunteer breathed 2.3 times more air per minute on
 7      average as a cyclist than as a car driver. When adjusted for variation in breathing rate, the range
 8      in 1-h median CO concentrations of car drivers (1.4 to 1.9 ppm) was only slightly higher than
 9      that for cyclists (1.2 to 1.7 ppm).
10           Studies have quantified the effect of traffic volume and speed on in-vehicle CO exposure.
11      Flachsbart et al. (1987) reported that in-vehicle CO exposures fell by 35% when test vehicle
12      speeds increased from 10 to 60 mph on eight commuter routes in Washington. In a  similar study
13      of typical commuter routes in central Riyadh, Saudi Arabia, Koushki et al. (1992) found that
14      in-vehicle CO exposures fell by 36% when vehicle speeds increased from  14 to 55 km/h (8.7 to
15      34.2 mph).  They also found that mean in-vehicle CO concentrations increased by 71.5% when
16      traffic volumes increased from 1,000 to 5,000 vehicles per hour.  Mean CO levels ranged from
17      30 to 40 ppm averaged over trips of 25 to 43 min during peak hours,  and ranged from 10 to
18      25 ppm for trips of 15 to 20 min during off-peak hours.
19           The effects  of diurnal and seasonal variation on in-vehicle CO exposure were not
20      completely discussed in the 1991 CO AQCD.  Studies of diurnal effects on in-vehicle exposure
21      during peak travel periods have been inconclusive because studies were unable to control for
22      covariation in traffic volumes and speeds, ambient CO concentrations, or meteorological
23      conditions (e.g., temperatures, wind speeds) during different periods of the day. In Los Angeles,
24      CA, Haagen-Smit (1966) found evidence that CO exposures  during afternoon commutes were
25      greater than those during morning commutes.  Similar results were later found by Cortese and
26      Spengler (1976) in Boston, MA, by Wallace (1979) in Washington, and by Dor et al. (1995) in
27      Paris.  However, Holland (1983) found contrary evidence in four U.S. cities. In all of these
28      studies, differences in exposures by time of day were not statistically significant. Recently,
29      Aim et al. (1998) reported that the geometric mean CO concentration of 11 morning trips
30      (3.1 ppm) exceeded that of 12 afternoon trips (2.0 ppm) on a standard route in Kuopio, Finland.
31      This result was attributed to weather and traffic variables.

        February 15, 1999                         4-16       DRAFT-DO NOT QUOTE OR CITE

-------
 1           Studies in different parts of the world have shown that seasonal variation in ambient
 2      temperatures, wind conditions, and traffic volumes affect passenger cabin exposure to CO.
 3      Studies by Ott et al. (1994) in northern California and by Dor et al. (1995) in France both
 4      measured exposures for an entire year. Both studies reported that passenger cabin exposures
 5      were generally higher in fall and winter than they were in spring and summer.  Such results
 6      usually are attributed to colder temperatures in temperate climates, which increase CO emissions
 7      per vehicle-mile during winter months. In a year-long study, Aim et al. (1998) reported that the
 8      highest CO concentrations for motorists in Kuopio occurred during morning rush hour periods in
 9      winter. In Hawaii, where temperatures are never cold enough to have a substantial effect on
10      motor vehicle emissions, Flachsbart (1998a) reported that light traffic flows on an arterial
11      highway in Honolulu and strong winds worked together to lower passenger cabin CO exposures
12      during late fall, whereas heavy traffic flows and calmer winds elevated cabin exposures during
13      winter and spring months.
14
15      4.3.1.2 Exposure to Carbon Monoxide in Recreational Vehicles
16           Two studies examined personal exposure to CO in the exhaust of recreational vehicles.
17      In the first study, Simeone (1991) collected CO concentrations in the passenger areas of large
18      power boats with side-mounted exhausts  during routine cruises offshore of Boston and
19      Annapolis, MD.  In Boston Harbor, CO concentrations averaged 56 ppm during a 60-min cruise
20      and 28 ppm after a 30-min cruise.  For the Chesapeake Bay cruises near Annapolis, average
21      stabilized CO concentrations at the helm  ranged from 93 to 170 ppm over 20- to 30-min periods
22      and 272 ppm over 30 min on the rear deck near the transom of the boat. In both  studies, exhaust
23      gas was affected significantly by airflow about the boat under certain head winds. At head-wind
24      speeds of 10 to 30 knots, turbulent mixing occurred in closer proximity to the rear of the boats,
25      enabling exhaust gases to migrate freely into each boat.
26           In the second study, researchers studied the CO exposure of a snowmobiler while traveling
27      in the wake of a lead snowmobiler on a 2- to 3-mi straight trail over level terrain in Grand Teton
28      National Park, WY (Snook, 1996). The CO exposure of the follower was collected under stable
29      atmospheric conditions in Tedlar bags. The distance between the two snowmobiles ranged from
30      25 to 125 ft and speeds ranged from  10 to 40 mph. The follower's maximum average centerline
31      exposure was 23.1 ppm, which occurred at 10 mph and 25 ft behind the lead snowmobile.

        February 15, 1999                         4-17        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Although Snook (1996) reported no averaging times for exposures, one can estimate that these
 2      times ranged from 3 to 18 min from the data given on the snowmobiler's travel distance and
 3      vehicle speed. At distances greater than 25 ft, centerline exposures tended to increase with
 4      greater speeds.  At 15 ft off centerline, average concentrations fell sharply to levels of 0 to
 5      7.5 ppm.  When the snowmobiler drove alone, the average concentration minus the background
 6      concentration was 1.3 to 3.0 ppm.  Background concentrations ranged from 0.2 to 0.5 ppm.
 7           Snowmobile tourism has become a booming business across the nation and in several
 8      national parks.  Over 87,000 tourists traveled by snowmobile in Yellowstone National Park
 9      (Wilkinson, 1995) during the winter of 1993-1994.  A snowmobile may emit from 10 to
10      20 g CO/mi, whereas a modern U.S. automobile emits far less (0.01 to 0.04 g/mi).  There are no
11      federal laws regulating the exhaust from snowmobile engines, and states are preempted from
12      implementing snowmobile emission standards. The typical snowmobile utilizes a two-stroke
13      engine, because it is less expensive than a four-stroke engine and provides a high power/weight
14      ratio. However, a two-stroke engine produces relatively high emissions of CO (Snook and
15      Davis, 1997).
16
17      4.3.1.3 Residential Exposure to Carbon Monoxide
18           Residential  sources of CO concentrations include motor vehicle operation inside a home
19      garage and the use of unventilated or poorly ventilated kerosene space heaters, gas appliances,
20      and charcoal grills and hibachis in the living area of the home.  Studies of exposures to nonfatal
21      concentrations are discussed first, followed by studies of unintentional deaths caused by high
22      indoor concentrations.
23           According to the Barbecue Industry Association, 44 million American households owned a
24      charcoal grill in 1989, and  an estimated 600 million charcoal-barbecuing events take place
25      annually (Hampson et al., 1994).  An early study showed that the air stream from charcoal grills
26      contains 20 to 2,000 ppm of CO, with 75% of grills emitting 200 ppm and above (Yates, 1967).
27      Gasman et al. (1990) reported COHb levels ranging from 6.9 to 17.4% in a family of four people
28      in northern California who had been exposed to smoke from cooking indoors on a barbecue grill,
29      which was found by fire fighters in the middle of the living room.
30           Mumford et al. (1991) and Williams  et al. (1992) assessed CO exposure to emissions from
31      unvented portable kerosene heaters in eight small mobile homes with no gas appliances and low

        February 15, 1999                        4-18         DRAFT-DO NOT QUOTE OR CITE

-------
 1      air exchange rates. Each home was monitored for an average of 6.5 h per day for 3 days per
 2      week for 4 weeks. For 2 weeks, the heater was on, and, for 2 weeks, it was off. When the heater
 3      was turned on, it was in use for an average of 4.5 h.  When the heater was in use, study
 4      participants (all nonsmokers) spent most of their time in the family room or kitchen.  Sampling
 5      took place in the living area about 1.5 to 3 m from the heater. The mean 8-h CO concentrations
 6      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)
 7      when it was off.  Peaks usually were observed at the end of the combustion period. The ambient
 8      CO level measured 0.5 h prior to heater use ranged from 0 to 8 ppm. When the heater was on,
 9      three of the eight homes had 8-h average CO levels that exceeded the NAAQS, and one home
10      routinely had levels of 30 to 50 ppm.
11          In joint studies, Wilson et al. (1993a,b) and Colome et al. (1994) reported CO exposures for
12      a random sample of California homes that used gas appliances during a 48-h period from
13      December 1991 to April 1992.  For periods of 48 h, the median CO concentration was 1.2 ppm
14      (indoors) and 0.8 ppm (outdoors),  and the median of the maximum 8-h average CO
15      concentrations was 2.0 ppm (indoors) and 1.4 ppm (outdoors). Of 277 surveyed homes,
16      13 (4.3%) had indoor CO concentrations and 8 (2.9%) had outdoor CO concentrations above the
17      NAAQS of 9 ppm for 8 h. The study did not translate these percentages to statewide estimates.
18      Using univariate regression analysis, outdoor CO concentrations explained approximately 55%
19      of the variation found in indoor CO concentrations.  Higher net indoor CO levels (indoor minus
20      outdoor CO concentrations) were traced to several factors:  space heating with a gas range and
21      gas-fired wall furnaces, use of gas ranges with continuous gas pilot lights, small home volumes,
22      and  smoking cigarettes.  The study reported that several factors may have contributed to higher
23      indoor net CO levels:  malfunctioning gas furnaces, automobile exhausts leaking into the home
24      from attached garages and carports, improper use of gas appliances (e.g., gas fireplaces), and
25      improper installation of gas appliances (e.g., forced air unit ducts).
26          In an area monitoring study, Kern et al. (1990) measured CO concentrations in a poorly
27      sealed,  detached garage from operation of an emissions controlled (catalytic reactor, oxygen
28      sensor) and an emissions uncontrolled vehicle (carbureted, no catalytic reactor).  Two tests were
29      conducted:  Test 1 involved a poorly sealed garage door with a 3-in opening, and Test 2 sealed
30      the garage door with rags.  The CO concentrations in the poorly sealed garage reached 4,700 ppm
31      for the car without emission controls versus 2,000 ppm for the car with controls. When the

        February 15, 1999                         4-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      garage was better sealed, CO concentrations reached 8,400 ppm for the uncontrolled vehicle
 2      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 period from 1979
 5      to 1988. In the first study, Liu et al.  (1993) reported that 13.3% of 444 deaths were caused by
 6      improper use of charcoal grills and hibachies, of which 54% occurred inside motor vehicles
 7      (e.g., a van or recreational camper), and 46% occurred in residential structures (e.g., homes,
 8      apartments, shacks, or tents).  Relative to their share of the state population, higher death rates
 9      occurred among Asians, blacks, males, and people aged 20 to 39. In the second study, Girman
10      et al. (1998) identified specific factors that caused or contributed to unintentional deaths caused
11      by CO from several combustion sources, including charcoal grills and hibachis,  other heating and
12      cooking appliances, motor vehicles,  small engines, and camping equipment.  There was a strong
13      association between alcohol use and CO poisoning from motor vehicles. Faulty heating
14      equipment used during winter months was implicated in about 50% of all unintentional deaths in
15      studies by both Girman et al. (1998) in California and Yoon et al. (1998) in New Mexico.
16           Based on data for 10 counties in Washington, Hampson et al. (1994) reported
17      characteristics of unintentional CO poisoning cases that occurred between 1982  and 1993. Most
18      cases occurred when electrical power was interrupted during fall and winter months.  Service was
19      either interrupted because of regional storms or discontinued because of unpaid utility bills.
20      Of 509 patients treated with hyperbaric oxygen, 79 (16%) were exposed when charcoal briquets
21      were burned for either heating or cooking in 32 separate incidents. Non-English speaking
22      Hispanic whites and Asians were disproportionately represented among the cases. The COHb
23      levels averaged 21.6% and ranged from 3.0 to 45.8%.
24           The National Center for Health Statistics and the U.S. Consumer Product Safety
25      Commission (CPSC) estimated that 212 deaths in 1992 were caused by fuel-burning appliances
26      used in the home. Of these deaths, 13 involved use of gasoline-powered appliances (National
27      Center for Health Statistics and U.S. Consumer Product Safety Commission, 1992).  The CPSC
28      also estimated that 3,900 CO injury accidents occurred in 1994, of which about 400 were
29      associated with the use of gasoline-powered engines or tools (National Institute for Occupational
30      Safety and Health, 1996).  In response to the problem,  several federal government agencies
31      issued a joint alert concerning exposure to CO emitted by these sources (National Institute for

        February 15, 1999                         4-20        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Occupational Safety and Health, 1996). These sources involved use of pressure washers, air
 2      compressors, concrete-cutting saws, electric generators, floor buffers, power trowels, water
 3      pumps, and welding equipment. Unintentional CO poisonings frequently happened indoors even
 4      when people took precautions to ventilate the building.
 5
 6      4.3.1.4 Exposure to Carbon Monoxide at Commercial Facilities
 7           Increasingly, people pump their own fuel at service stations.  This practice potentially puts
 8      people at greater risk of exposure to CO emissions from motor vehicles if nearby engines are
 9      idling. Wilson et al. (1991) randomly sampled 100 self-service filling stations and, for
10      comparison, took convenience samples at 10 parking garages and 10 nearby office buildings in
11      Los Angeles, Orange, Riverside, and San Bernardino counties of Southern California. They took
12      5-min samples of 13 motor vehicle air pollutants including CO in each microenvironment and in
13      the ambient environment.  Microenvironmental and ambient concentrations were measured on
14      the same day but not simultaneously.  The highest median CO concentration occurred in parking
15      garages (11.0 ppm), followed by service stations (4.3 ppm), and  office buildings (4.0 ppm). The
16      median ambient CO concentration was 2.0 ppm.
17           Ice skating, motocross, and tractor pulls are sporting events in which significant quantities
18      of CO may be emitted in short periods of time by machines in poorly ventilated indoor arenas.
19      The CO is emitted by several sources, including ice resurfacing machines and ice edgers during
20      skating events; motor vehicles at tractor-pull, monster-truck, and motocross competitions; and
21      gas-powered radiant heaters used to heat viewing stands. At these events, a "tractor" is a truck or
22      other vehicle modified to look like a farm tractor that is powered with aircraft turbines or
23      supercharged automobile engines. These competitions usually involve many motor vehicles with
24      no emission controls.
25           Several studies were not cited in the 1991 CO AQCD. First, Kwok (1981) reported
26      episodes of CO poisoning among skaters inside four arenas in Ontario, Canada.  Mean CO levels
27      ranged from 4 to 81 ppm for periods of about 80 min.  The CO levels in the spectator areas
28      ranged from 90 to 100% of levels on the rink. The ice resurfacing machines lacked catalytic
29      emission controls.  Second, both Sorensen (1986) and Miller et al. (1989) reported CO
30      concentrations greater than 100 ppm in ice rinks from the use of gas resurfacing machines. High
31      concentrations were attributed to poorly maintained machines and to an ice rink without

        February 15, 1999                         4-21        DRAFT-DO NOT QUOTE OR CITE

-------
 1      sufficient ventilation. Third, based on data collected in the Quebec city area, Levesque et al.
 2      (1990) developed a linear relationship between CO exposure and the CO concentration in
 3      exhaled breath (see Section 4.3.1.5 for discussion of CO exposure and breath CO relationships)
 4      but could not eliminate other factors affecting the relationship.  In a later study, Levesque et al.
 5      (1991) measured the alveolar CO of 14 male adult nonsmokers who played ice hockey, but who
 6      were not exposed in occupational settings. Rink CO concentrations ranged from 0 to 76.2 ppm.
 7      The study again found a linear relationship between exposure and absorbed CO such that, for
 8      each 10 ppm of CO in the indoor air, the players absorbed enough CO to raise alveolar CO by
 9      4.1 ppm or about 0.76% COHb.
10          In the United States, surveys of CO exposure were done at ice arenas in the northern states
11      of Vermont, Massachusetts, Wisconsin, and Washington. For a rink in Massachusetts, Lee et al.
12      (1993) showed that excessive CO concentrations can occur even with well-maintained equipment
13      and fewer resurfacing operations if ventilation is inadequate. Average CO levels were less than
14      20 ppm over 14 h. There was no significant source of outdoor CO.  Ventilation systems could
15      not disperse pollutants emitted and trapped by temperature inversions and low air circulation at
16      ice level. In another study, Lee et al. (1994) reported that CO concentrations measured inside six
17      enclosed rinks in the Boston area during a 2-h hockey game ranged from 4 to 117 ppm, whereas
18      outdoor concentrations were about 2 to 3 ppm, and the alveolar CO of hockey players  increased
19      by an average of 0.53 ppm per 1 ppm CO exposure over 2 h. Fifteen years earlier, Spengler et al.
20      (1978) found CO levels ranging from 23 to 100 ppm in eight enclosed rinks in the Boston area.
21      This suggests that CO exposure levels in ice arenas have not improved.
22          In a letter, Paulozzi et al. (1991) reported that 25 people, exposed to CO during a
23      high-school ice hockey game in Vermont, had mean COHb levels of 8.9%, but did not report
24      whether any of them were smokers.  Although Paulozzi et al. (1991) was unable to measure CO
25      concentrations at the game, Smith et al. (1992) reported CO levels of 150 ppm (no averaging
26      time was given) at an indoor ice-hockey rink in Wisconsin.  To document the extent of the
27      problem in Vermont, Paulozzi et al. (1993) took CO measurements during eight high-school
28      games played in the state, and reported that average CO levels for the entire game ranged from
29      <5 to 101 ppm, with a mean of 35 ppm. Hampson (1996) reported a maximum CO level of
30      354 ppm inside an ice arena in Seattle in March 1996.  Based on data for 17 persons whose
31      tobacco use was not reported, the average COHb level was 8.6%, with a range of 3.3 to 13.9%.

        February 15, 1999                        4-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      The source of the CO was a malfunction in a 20-year-old ice resurfacing machine.  Hampson also
 2      reported that CO may have diffused to an adjacent bingo hall through an open door. In view of
 3      all these studies, the State of Minnesota declared in Regulation No. 4635 that CO measurements
 4      taken 20 min after ice resurfacing must be less than 30 ppm.
 5           Studies also have been done in sports arenas that allow motor vehicles.  Boudreau et al.
 6      (1994) reported CO levels for three indoor sporting events (i.e., monster-truck competitions and
 7      tractor pulls) in Cincinnati, OH. The CO measurements were taken before and during each event
 8      at different elevations in the public seating area of each arena with most readings obtained at the
 9      midpoint elevation where most people were seated. Average CO concentrations over 1 to
10      2 h ranged from 13 to 23 ppm before the event to 79 to 140 ppm during the event.  Measured CO
11      levels were lower at higher seating levels.  Attendance was approximately 40% of the arena's
12      capacity for each event. The ventilation system was operated maximally, and ground-level
13      entrances were completely open.
14           High CO concentrations also have been reported at motor vehicle competitions in Canada.
15      In an important study not cited in the 1991 CO AQCD, Luckurst and Solkoski (1990) recorded
16      CO concentrations at two tractor-pull events in Winnipeg, Manitoba.  The mean of instantaneous
17      concentrations at 25 locations in the  arena ranged from 68 ppm  at the start of the first event to
18      262 ppm by the end.  At the second event, the range was 78 to 436 ppm. Levesque et al. (1997)
19      reported CO levels at an indoor motocross competition that was held in a skating rink in  the
20      Quebec city region. The event lasted from roughly 8 p.m. to midnight in May 1994. The average
21      CO concentrations were determined  at five stations located at different points in the arena. The
22      TWA concentrations ranged from 19.1 to 38.0 ppm, with concentrations higher during the
23      second half of the show. High CO concentrations forced a health official to interrupt the event
24      seven times to help clear the air. Covariance analysis showed that CO levels were  related to the
25      initial CO concentration, the event duration, motor size, and especially the number of
26      motorcycles on the track.
27
28      4.3.1.5  Studies of Breath Carbon Monoxide in Populations: The Effects of Exposure to
29             Carbon Monoxide
30           The EPA reviewed the pre-1990 literature reports of breath CO measurements in various
31      populations (U.S. Environmental Protection Agency, 1991; Section 8.5.2.2).  These data and the

        February 15, 1999                         4-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      more recent data on breath CO in the following part of this section often are collected with
 2      different breathhold-time, often are uncorrected for the CO content of the inhaled air (Smith,
 3      1977; Wallace, 1983), and also may be subject to a positive hydrogen-interference if the breath
 4      CO is analyzed electrochemically (Vreman et al., 1993) (see Section 2.6.2 of this document).
 5      Consequently, this should be taken as a caveat by the reader that a portion of the variance
 6      between the results of different studies may be related to different breath collection methods and
 7      different breath CO measurement techniques.
 8           Lando et al. (1991) collected breath samples of 4,647 workers using MiniCO breath kits
 9      (Model 1000, Catalyst Research Corporation, Owings Mills, MD). The latter part of a breath
10      was collected in a balloon following a 15-s breathhold but the method of analysis was not
11      described.  Although the authors cite Jarvis et al. (1980) for this method, Jarvis et al. (1980) used
12      the Jones et al. (1958) method that requires a 20-s breathhold. Furthermore, these data are
13      uncorrected for the amount of CO in the maximal inhalation prior to the breathhold step (Smith,
14      1977; Wallace, 1983). Consequently, these data are not compatible with other studies using
15      20-s for the breathhold time and corrected data. Mean CO levels (Table 4-1) ranged from
16      4.2 (±1.66 standard deviation [SD]) ppm for never-smokers to 33.3 (±11.22 SD) ppm for heavy
17      smokers (25 cigarettes/day or more). Based on cutoffs of 3 and 6 ppm above ambient, a larger
18      number of ex-smokers (1.7 to 3.3%) than never-smokers (0.4 to 1.9%) appeared to be falsely
19      reporting their smoking status.
20           Chung et al. (1994) employed the Lee and Yanagisawa (1992, 1995) sampler to measure
21      personal exposure to CO of 15 Korean housewives using charcoal briquettes for cooking.  The
22      COHb levels also were measured using a CO-Oximeter (CO-Ox). Although the personal
23      sampler had somewhat high imprecision based on four duplicate samples (average of 2.1 ppm
24      difference), the investigators were able to document a higher level of both exposure to CO and
25      blood COHb when the charcoal briquettes were used. Levels of COHb were generally high, even
26      without use of the briquettes, leading the experimenters to hypothesize that the high prevalence
27      of smoking (all 15 subjects had smokers in their homes) had elevated the level above the levels
28      found in the U.S. among nonsmokers.
29           Seufert and Kiser (1996) measured CO levels in the end-tidal breath after a 10-s breathhold
30      of 126 crew members of a nuclear submarine just before and just after a 62-h submerged period.
31      The CO level in the submarine (called "ambient" by the authors) increased from 2.6 ppm to

        February 15, 1999                        4-24       DRAFT-DO NOT QUOTE OR CITE

-------
         TABLE 4-1.  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     9.2 ppm in the fan room and in two other spaces. The authors state that the increase was caused
2     primarily by cigarette smoke from the 40 smokers aboard because diesel engines were not used
3     during the submersion period. The nonsmokers' breath CO increased from 9 to 21 ppm.
4     Although the authors did not comment on the considerable difference between the nonsmokers'
5     breath CO of 21 ppm and the measured "ambient" concentration of only 9.2 ppm, it may have
6     been because of higher smoking rates in nonmonitored duty sections than in the monitored
7     sections, the absence of a correction for higher CO concentrations in the air inhaled for the 10-s
8     breathhold than in the end-tidal breath CO, and the end-tidal breath CO and "ambient" CO
9     measurements were made with two different instrumental  systems.  The operation of an
      February 15, 1999
4-25
DRAFT-DO NOT QUOTE OR CITE

-------
 1      atmospheric revitalization system that removed CO also may have contributed to a lower
 2      monitored "ambient" CO than the CO nonsmokers were exposed to in the nonmonitored duty
 3      sections.
 4           Zayasu et al. (1997) present the first study showing that asthmatics untreated by
 5      corticosteroids have higher 20-s breathhold end-tidal breath CO than either healthy controls or
 6      treated asthmatics, as determined by subtracting the background level from the observed reading
 7      (Figure 4-5). This is not the required correction for CO in the inhaled  air reported by Smith
 8      (1977) and Wallace (1983), so these data are not consistent with those studies where this
 9      correction was made.  They attribute the higher levels to lung inflammation, leading to a possible
10      increase in heme oxygenase, which creates endogenous CO. (For more information on
11      endogenous CO production, see Section 5.3.)
12           Shenoi et al. (1998) tested 470 youths in hospital admissions (aged 5 to 20 years) for CO in
13      breath using the electrochemical Vitalograph Breath CO monitor (Vitalograph, Inc., Lenexa,
14      KS).  The results, showing that 1.9% (9 out of 470) had end-tidal breath CO levels greater than
15      9 ppm after a 20-s breathhold, were confirmed by CO-Ox testing of blood.  Five of the 9 patients
16      with the higher breath CO were believed to be cigarette smokers, one may have been exposed to
17      fumes from a faulty furnace, and three were believed to be exposed to  environmental tobacco
18      smoke or traffic exhaust.  No corrections were made for the parts per million of CO in the air
19      inhaled for the breathhold.
20
21      4.3.1.6 Nonoccupational Exposure to Methylene Chloride
22           Nonoccupational exposure to methylene chloride, which can be metabolized to CO in the
23      body, potentially occurs when the chemical is found in contaminated ambient air and
24      groundwater used as drinking water and in consumer products that contain the chemical as a
25      solvent, flame-retardant additive, or propellant. Exposure to methylene chloride in the home
26      primarily occurs through use of paint and varnish removers. Exposure also may occur through
27      use of aerosol propellants such as those found in hair sprays, antiperspirants,  air fresheners, and
28      spray paints.  The Agency for Toxic Substances and Disease Registry of the U.S. Public Health
29      Service reported that some aerosol products may contain up to 50% methylene chloride (Agency
30      for Toxic Substances and Disease Registry, 1993). However, the current extent of methylene


        February 15, 1999                         4-26        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 4-5. 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).
February 15, 1999
                               4-27
DRAFT-DO NOT QUOTE OR CITE

-------
 1      chloride in aerosol products apparently has not been studied recently. Also unknown are typical
 2      population exposures to methylene chloride from consumer products.
 3           Ambient exposure may occur near production and use facilities or near hazardous waste
 4      sites that store methylene chloride. Ambient concentrations of methylene chloride near organic
 5      solvent cleaning and paint and varnish removal operations range from 7.1 to 14.3 ppb, averaged
 6      over 1 year (Systems Applications Inc., 1983), and ambient levels at other locations were
 7      reported by the U.S. Environmental Protection Agency (1985).  Although methylene chloride
 8      readily disperses when released into the air, it may remain in groundwater for years and be
 9      ingested in drinking water or inhaled when it volatilizes during showering and laundering
10      (Agency for Toxic Substances and Disease Registry, 1993).
11           Exposure to about 500 ppm of methylene chloride for several hours can elevate COHb
12      levels to 15%. Increases in COHb levels can be detected in the blood of nonsmokers about
13      30 min after exposure to methylene chloride.  Stewart et al. (1972) demonstrated that elevated
14      COHb levels were proportional to a series of controlled exposures to methylene chloride.  In a
15      controlled experiment, Stewart and Hake (1976) observed postexposure levels of COHb ranging
16      from  5 to 10% after 3  h of use of a liquid-gel paint remover containing 80% methylene chloride
17      and 20% methanol by weight. Concurrent exposure to methylene chloride and methanol
18      prolongs the period of elevated COHb in the body (Stewart and Hake, 1976; Buie et al., 1986;
19      Wilcosky and Simonsen,  1991).  Peterson (1978) reported COHb levels of up to about 10%
20      saturation after inhalation of methylene chloride concentrations ranging from 50 to 500 ppm over
21      5 days for 7.5 h per day.
22
23      4.3.2 Occupational Exposures
24           This subsection discusses occupational exposures to CO and methylene chloride.
25
26      4.3.2.1  Exposures to Carbon Monoxide in the Workplace
27           A survey by the National Institute for Occupational  Safety and Health found that
28      3.5 million workers in the private sector potentially are exposed to CO primarily from motor
29      exhaust. The number of persons potentially exposed to CO in the work environment is greater
30      than that for any other physical or chemical  agent (Pedersen and Sieber, 1988). In 1992, there
31      were  900 work-related CO poisonings resulting in death or illness in private industry as reported
        February 15,  1999                         4-28         DRAFT-DO NOT QUOTE OR CITE

-------
 1      by the U.S. Bureau of Labor Statistics (as cited in National Institute for Occupational Safety and
 2      Health, 1996). Three risk factors affect industrial occupational exposure:  (1) the work
 3      environment is located in a densely populated area that has high background CO concentrations;
 4      (2) the work environment produces CO as a product or by-product of an industrial process, or the
 5      work environment tends to accumulate CO concentrations that may result in occupational
 6      exposures; and (3) the work environment involves exposure to methylene chloride, which is
 7      metabolized to CO in the body.  Proximity to fuel combustion of all types elevates CO exposure
 8      for certain occupations:  airport employees; auto mechanics; small gasoline-powered tool
 9      operators (e.g., users of chainsaws); charcoal meat grillers; construction workers; crane deck
10      operators; firefighters; forklift operators; parking garage or gas station attendants; policemen;
11      taxi, bus, and truck drivers; toll booth and roadside workers; and warehouse workers (U.S.
12      Environmental Protection Agency, 1991).
13           Studies of firefighters are discussed briefly below because these studies were not discussed
14      in the 1991 CO AQCD.  Other occupational studies of CO exposure are not discussed here, but
15      are summarized in Table 4-2, which shows CO concentrations for each study (typical values or
16      ranges), averaging periods, and the measured or estimated percent COHb levels for nonsmokers,
17      if reported.
18           Lees (1995) reviewed studies of firefighter exposures to combustion products including
19      CO. During severe fires, firefighters were exposed to CO concentrations in excess of 500 ppm in
20      approximately 29% of 1,329 minutes sampled by Burgess et al. (1977) and in 48% of
21      measurements taken by Barnard and Weber (1979). Gold et al. (1978) reported a geometric
22      mean concentration of 110 ppm for a log-normal distribution of 65 samples with average
23      duration of less than 10 min. The short-term exposure limit, designed to prevent acute effects of
24      CO exposure, is 400 ppm averaged over 15 min. In three studies, the STEL was exceeded in
25      15 to 33% of measurements (Treitman et al., 1980; Brandt-Rauf et al., 1988; Jankovic et al.,
26      1991).  Inside a self-contained breathing apparatus, CO measurements ranged from 1 to 105 ppm
27      in six samples (Jankovic et al., 1991).
28           Firefighters are exposed to lower CO levels when they suppress bushfires, wildland fires,
29      and forest fires. For bushfires, Brotherhood et al. (1990) estimated that Australian firefighters
30      were exposed to CO levels averaging  17 ppm based on COHb measurements taken afterwards.
31      In a study of wildland fires in California, Materna et al. (1992) reported an average CO

        February 15, 1999                         4-29        DRAFT-DO NOT QUOTE OR CITE

-------
February 15.
VO
VO
VO









J^.
1
OJ
r~^
TABLE 4-2. STUDIES OF
Occupational Category
Airport workers
Bus drivers
Chainsaw/gas tool
operators


Charcoal meat grillers
Firefighters




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)
OCCUPATIONAL EXPOSURES AND DOSAGES3
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
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.
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)
o
o
2
o
H
O
c
o
H
W
O
V
O
HH
H
W
         Forklift operators and
           workers in facilities
           with forklifts
Garage mechanics

Manufacturing jobs

Traffic/roadway workers
   250-300 (5 h)
    NA (4.4 h)
   370-386 (NA)
25-47 TWA (8-12 h)
     3-34 (8 h)

  42.6% > 35  (1 h)

  0-83 TWA (4 h)

     2-7 (8 h)
     1-4.3 (8 h)
     5-42 (2 s)
   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
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.
Baucometal. (1987)
Fawcettetal. (1992)
Ely etal. (1995)
McCammon et al. (1996)
Apte (1997)

Gourdeauetal. (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      concentration of 14.4 ppm over a 3.5-h period (range 1.4 to 38 ppm) during fireline mop-up and
 2      a prescribed burn. Concentrations were higher during evening hours, when inversions occurred,
 3      and could range up to 300 ppm near gasoline-powered pumping engines. Materna et al. (1993)
 4      found comparable results using different methods. For forest fires, the National Institute for
 5      Occupational Safety and Health (1994) reported an average CO concentration of 11.5 ppm for a
 6      9-h period.
 7
 8      4.3.2.2 Exposures to Methylene Chloride in the Workplace
 9           Certain occupations expose workers to organic solvents such as methylene chloride.  The
10      solvent is widely used as a degreaser, paint remover, aerosol propellant, and blowing agent for
11      polyurethane foams. It is used as an extractant for foods and spices,  a grain fumigant, and a
12      low-pressure refrigerant. It also is used in the manufacturing of synthetic fibers, photographic
13      film, polycarbonate plastics, Pharmaceuticals, printed  circuit boards, and inks. More than one
14      million workers have significant potential for exposure to methylene chloride (Agency for Toxic
15      Substances  and Disease Registry, 1993). Moreover, the highest levels of exposure to methylene
16      chloride often occur in the workplace. To protect worker health, the 8-h TWA threshold limit
17      value for methylene chloride was set at 50 ppm by the American Conference of Governmental
18      Industrial Hygienists. Exposure at this concentration leads to COHb levels of about 1.9% in
19      experimental subjects.  Exposure to 500 ppm for several hours may elevate COHb levels as high
20      as  15%. An 8-h exposure to about 500 mg/m3 (3.5 mg/m3 = 1 ppm) of methylene chloride vapor
21      is equivalent to an 8-h exposure to 35 ppm of CO (U.S. Environmental Protection Agency,
22      1985).
23           Methylene chloride stored in tissue may continue to metabolize to CO after several hours of
24      acute exposure.  In such cases, COHb levels will continue to rise and peak as high as 25% about
25      5 to 6 h after exposure (Agency for Toxic Substances  and Disease Registry, 1993). Shusterman
26      et al. (1990) reported an apparent linear elevation of COHb as a function of hours worked by a
27      furniture refinisher who used paint stripper containing methylene chloride.  Ghittori et al. (1993)
28      reported a significant linear correlation (correlation coefficient [r]  = 0.87) between methylene
29      chloride concentration in air and CO in alveolar air of nonsmoking and sedentary factory workers
30      in Italy. Exposure to 600 mg/m3 of methylene chloride for 7.5 h was associated with a COHb
31      level of 6.8% in eight volunteers.  Exposure to methylene chloride also can be fatal.  Leikin et al.

        February 15, 1999                          4-31         DRAFT-DO NOT QUOTE OR CITE

-------
 1      (1990) reported fatalities of two people who were exposed to unknown concentrations of
 2      methylene chloride while they removed paint in an enclosed space.  Death was caused not by CO
 3      poisoning, but by solvent-induced narcosis.  Before they died, their COHb levels continued to
 4      rise following cessation of exposure despite treatment by high levels of oxygen.
 5
 6      4.3.3 Activity Pattern Studies
 7           In assessing population exposure, studies of human activity patterns over a fixed time
 8      period (e.g., 24 h) are necessary to determine how many people potentially are exposed to
 9      sources of an air pollutant, and how long people spend in activities that involve use of these
10      sources.  Accordingly, this section reviews studies of human activity patterns that pertain to
11      population exposure to CO and methylene chloride. Previous studies reported that many
12      Americans spent most of their time indoors at home, school, or work, etc. (Szalai, 1972;  Chapin,
13      1974; Meyer, 1983; Johnson, 1987; Schwab et al.,  1990). Although more recent activity pattern
14      studies largely confirm this finding, their sampling and questionnaire designs provide new
15      insights. This section reviews, in chronological order, these newer studies that include two
16      surveys of activity patterns in California, a similar survey of preadolescents in six states,
17      a comparative study between California and the nation, a study in southwest England, a Boston
18      study, and a recent survey at the national level.  As a body, these studies were not discussed in
19      the 1991 CO AQCD.
20
21      4.3.3.1 Activity Patterns of California Residents
22           The California Air Resources Board (CARB) conducted two surveys to determine the
23      activity patterns of California residents. In each case, projectable probability samples were
24      drawn from English-speaking households who had telephones. The first surveyed 1,762  adults
25      and adolescents over 11 years of age from fall 1987 through summer 1988 (Wiley et al.,  1991a;
26      Jenkins et al.,  1992), and the second surveyed 1,200 children under age 11 from April  1989
27      through February 1990 (Wiley et al., 1991b; Phillips et al., 1991). Using telephone interviews,
28      both surveys asked participants to complete a 24-h diary for the preceding day. People ages
29      9 and over responded directly to the interview, and the primary adult careprovider responded for
30      young children.  The diaries enabled estimates of time spent in various activities and locations,
        February 15, 1999                         4-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1      and determinations of whether respondents used or were near sources of pollutants, including
 2      consumer products, combustion appliances, and motor vehicles.
 3           Similar to previous studies, the results showed that all age groups spent most of their time
 4      indoors.  Adults and adolescents spent, on average, 87% of their time indoors (62% at home and
 5      25% elsewhere), and only 6% of their time outdoors.  They also spent 7% of their time in transit
 6      mostly in a car, van, or pickup truck. Compared to adults and adolescents, children spent a
 7      similar amount of time indoors (86%), but more time at home (76%) and outdoors (10%),  and
 8      less time indoors elsewhere (10%) and in transit by car, van or pickup truck (4%). About 46% of
 9      nonsmoking adults/adolescents reported being near others' tobacco smoke at some time during
10      the day.
11           Table 4-3 summarizes results of the two California surveys for various microenvironments
12      pertinent to CO exposure.  For each microenvironment, the table shows the mean and range in
13      time spent per day by both the entire sample and by those who actually did an activity in the
14      microenvironment (i.e., "doers"). The results show the disparity in mean time spent by the
15      population and by doers of an activity, which has implications for calculating population
16      exposure in risk assessment. Table 4-4 gives the percentage of each sampled population who
17      reported use of or proximity to potential sources of either CO or methylene chloride on a given
18      day. The study did not measure CO or methyl ene chloride concentrations from these sources in
19      microenvironments.  Also, the surveys did not indicate whether respondents lived in a home
20      where combustion appliances were vented.
21
22      4.3.3.2 Activity Patterns  of Children in Six States
23           In 1990-91, Silvers et al. (1994) surveyed the activities of preadolescent children (ages 5 to
24      12 years) from a projectable probability sample of 1,000 households in six states.  These states
25      included three on the East Coast (New Jersey, New York, and Pennsylvania) and three on the
26      West Coast (California, Oregon, and Washington).  Comparisons between this study known as
27      the Children's Activity Survey (CAS) and the CARB children's  study are possible because both
28      were done over an entire year at about the same time, and both used a retrospective time diary for
29      a 24-h day. Both studies reported very similar results in terms of the mean hours per day spent
30      by preadolescent children for locations designated "indoors" (21.5 h for  CARB versus  21.7 h for
31      CAS) and "at home" (18.0 h for CARB versus 17.8 h for CAS). For each study, these  results

        February 15,  1999                         4-33       DRAFT-DO NOT QUOTE OR CITE

-------
VO
VO
VO
00
           TABLE 4-3.  TIME SPENT IN DIFFERENT MICROENVIRONMENTS BY CALIFORNIANS, 1987 TO 1990
                                                  (minutes per day; weighted)
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
o
6

-------
            TABLE 4-4. 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)
Potential Pollutant Source
Consumer Products3
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 garageb
Had vehicle in attached garageb
Took a hot shower8
Near Combustion Appliances
Had gas heat onb
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
31b
12
5

26

15
62
37
77

26
35


43
64
46

Children
1989-90

36
37
o
J
2

11

1
63
36
26

24
29





38
      Potential methylene chloride exposure.
      bData presented for adult respondents (age 18 years or older) only.
      Source: Adapted from Jenkins et al. (1992); Phillips et al. (1991).
1
2
3
4
5
varied by ±1 h for different seasons of the year.  There was variation in specific activities (e.g.,
the CAS study reported that preadolescents spent less time per day "riding in a vehicle" in
California [0.52 h] than they did in the five other states [0.82 h], when the five were combined as
a group).  The CAS study did not report time spent near other CO sources.

4.3.3.3 A Comparative Study Between California and the Nation
     Robinson and Thomas  (1991) compared results of activity pattern studies, one conducted
by CARS in California in 1987-1988 (Wiley et al., 1991a; Jenkins et al.,  1992), and the other
      February 15, 1999
                                         4-35
DRAFT-DO NOT QUOTE OR CITE

-------
 1      done at the national level in 1985 (Cutler, 1990; Cornish et al., 1991). Although the two surveys
 2      used different methods of gathering and coding data, the data were receded to enable
 3      comparisons. The comparison showed that Californians averaged more time at work and in
 4      commuting to work than was the case nationally, but averaged less time doing housework and
 5      caring for children. California men also spent more time traveling.  The national study appeared
 6      to show greater time spent at home and in the yard. However, these results could be explained by
 7      differences in location codes between the two studies, rather than by actual differences in
 8      participant activity patterns between them. For example, the national  study did not ask
 9      participants to identify whether they worked indoors or outdoors. Because the national study was
10      not designed for exposure assessment, the authors  proposed that the CARB study become a
11      model for a future national study oriented to exposure assessment. Such a study is discussed in
12      Section 4.3.3.6.
13
14      4.3.3.4 An English Study
15          Farrow et al. (1997) determined how much time was spent inside the home from a sample
16      of 170 households living in Avon, England, from November 1990 through June 1993.
17      A pregnant woman lived in each household at the start of the study. Households completed a
18      weekly diary for 1 year that covered roughly the last 6 mo of the woman's pregnancy and the first
19      6 mo of the new infant's life. The results indicated that the average amount of time spent inside
20      the home per day varied by family member as follows: mothers, 18.4 h (76.7%); fathers, 14.7 h
21      (61.3%); and infants, 19.3 h (80.4%).  Infants spent more time at home during winter than
22      summer. Although fathers spent more time at home on weekends, mothers and infants spent less
23      time. The applicability of the study results for U.S. households was not determined, and it is
24      hard to judge without comparative information about the two countries.
25
26      4.3.3.5 A Boston Study of Household Activities, Life Cycle, and Role Allocation
27          Using activity diary data from 150 households that participated in a 1991 Boston survey,
28      Vadarevu and Stopher (1996) tested several hypotheses  about household travel. One study
29      hypothesis was  that there are significant differences in mean time allocations of activities among
30      different life-cycle groups based on age, working status, and household size. They tested the
31      theory that life-cycle stage affects which activities  fall into mandatory, flexible, and optional

        February 15,  1999                        4-36       DRAFT-DO NOT QUOTE OR CITE

-------
 1      categories; how much time can be allocated to different activities; and which household member
 2      does each activity. They found that time allocated by households to specific activities varied
 3      according to whether the household consisted of a single working adult, multiple adults, a young
 4      family, an older family, or a nonworking adult.  However, they found no significant differences
 5      among the life-cycle groups or between any life-cycle group and the population mean in terms of
 6      the total time spent in mandatory activities (work, work-related, school, and certain at-home
 7      activities), which required on average 21 h per day. The amount of time spent in all flexible,
 8      optional, and travel activities was about 3 h per  day.
 9
10      4.3.3.6 The National Human Activity Pattern Survey
11           The National Human Activity Pattern Survey collected 24-h diary data of activities and
12      locations provided by 9,386 respondents interviewed nationwide in the United States between
13      October 1992 and September 1994 (Klepeis et al., 1996). To enable projections to a larger
14      population, the sample was weighted by the 1990 U.S. Census data to account for
15      disproportionate sampling of certain population groups defined by age and gender. Results were
16      analyzed across a dozen subgroups: gender, age, race, Hispanic, education, employment, census
17      region, day-of-week, season, asthma, angina, and bronchitis/emphysema.  The weighted results
18      showed that, on average, 86.9% of a person's day was spent indoors (68.7% at residential
19      locations), 7.2% of the day was spent in or near vehicles, and 5.9% of the day was spent in
20      outdoor locations.
21           The study also reported unweighted descriptive statistics and percentiles for both the full
22      population and various subpopulations (i.e., people who actually did certain activities or who
23      spent time in certain microenvironments) (Tsang and Klepeis, 1996). Of all 9,386 respondents,
24      38.3% reported having a gas range or oven at home, and another 23.7% said that the range/oven
25      had a burning pilot light. In terms of motor vehicle use, 10% of 6,560 people (7.0% of total
26      sample) spent more than 175 min per day inside a car,  and 10% of 1,172 people (1.2% of total
27      sample) spent more than 180 min inside a truck, pickup, or van.  Of those who were inside a
28      car and knew they had angina (n = 154 respondents), 10% of them spent more than  162 min
29      per day inside a car. The survey also asked about sources of household pollutants.
30      Of 4,723 respondents, 10.5% were exposed to solvents, 10.4% to open flames, and 8.4% to
31      gas-diesel powered equipment; 6.3% of these respondents were in a garage or indoor parking lot;

        February 15, 1999                         4-37        DRAFT-DO NOT QUOTE OR CITE

-------
 1     and 5.7% reported that someone smoked cigarettes at home. Only 1.8% of 4,663 respondents
 2     reported having a kerosene space heater at home.
 3
 4
 5     4.4   MAJOR FACTORS AFFECTING POPULATION EXPOSURE
 6          This section discusses major factors that have and may continue to affect population
 7     exposure to CO.  These factors include public policies affecting urban transportation planning
 8     and air quality, motor vehicle emissions, and smoking in public places and social changes
 9     affecting human activity patterns.
10
11     4.4.1  Federal Policies Affecting Transportation and Air Quality in
12            Urban Areas
13          In the United States, the national effort to improve air quality can be traced to the Clean Air
14     Act (CAA) amendments of 1970, 1977, and 1990.  As discussed in Chapter 3, the effect of these
15     CAA amendments on ambient CO concentrations has been substantial. Moreover, emissions
16     from on-road vehicles have declined since 1970, even as other socioeconomic indicators of
17     growth have increased. Between 1970  and 1995, nationwide emissions of CO from on-road
18     vehicles fell 33.4% (U.S. Environmental Protection Agency, 1996), despite compound annual
19     growth rates of 1.0% in the nation's population and 3.2% in vehicle miles of travel (VMT)
20     during the same period (U.S. Department of Transportation, 1996). The faster growth rate of
21     VMT can be attributed to many factors that have decentralized housing and jobs within urban
22     regions since World War II.
23          Since the mid-1960s, major construction projects intended to expand highway capacities
24     have been opposed in some metropolitan areas. Opponents claimed that these projects promoted
25     urban sprawl and induced motor vehicle travel that raised regional air pollutant emissions.
26     To address these concerns, the  1990 CAA amendments state that transportation actions (plans,
27     programs, and projects) cannot create new NAAQS violations, increase the frequency or severity
28     of existing NAAQS violations, nor delay attainment of the NAAQS (U.S. Code, 1990). Pursuant
29     thereto, the U.S. Environmental Protection Agency promulgated its Transportation Conformity
30     Rule. Complementary provisions of the 1991 Intermodal Surface Transportation Efficiency Act
31     offered financial incentives under the Congestion Management and Air Quality (CMAQ)
       February 15, 1999                         4-38        DRAFT-DO NOT QUOTE OR CITE

-------
 1      improvement program. Under CMAQ, metropolitan planning organizations were offered federal
 2      funds to improve air quality by implementing transportation control measures (TCMs).
 3      Examples of TCMs include programs to promote car and van pooling, flextime, special lanes for
 4      high occupancy vehicles, and parking restrictions.
 5           The U.S. Environmental Protection Agency (1994) examined how TCMs have changed
 6      travel activity, including number of trips, vehicle miles of travel, vehicle speed, travel time, and
 7      the extent to which commuters have shifted travel from peak to off-peak periods. Using an
 8      emission factors model (i.e., MOBILES [for a description of MOBILES, see U.S. Environmental
 9      Protection Agency, 1998b]), the study inferred how much TCMs would change average speeds of
10      motor vehicles  and CO emissions therefrom.  The direct effect of TCMs on commuter exposure
11      to CO has received only limited study. Flachsbart (1989) found that priority (with-flow and
12      contra-flow) lanes were effective in reducing exposure to motor vehicle exhaust on a coastal
13      artery in Honolulu. Compared to commuter CO exposure in adjacent but congested lanes,
14      exposure in priority lanes was about 18% less for those in carpools, 28% less for those in
15      high-occupancy vehicles (e.g., vanpools), and 61% less for those in express buses.  These
16      differences occurred possibly because commuters in priority lanes traveled faster than those in
17      the congested lanes. Faster vehicles created more air turbulence, which may have helped to
18      disperse pollutants surrounding vehicles in priority lanes. Furthermore, these differences existed
19      even though the priority lanes were often downwind of the congested lanes. Although higher
20      speeds were related to lower exposures in priority lanes,  differences in exposure also could have
21      been caused by differences in vehicle type and ventilation, both of which were not controlled.
22           Models to estimate the direct effects of TCMs on commuter CO exposure are not apparent
23      in the literature. However, Flachsbart (1998a) developed a series of statistical models to predict
24      passenger cabin exposure to  CO based on trip variables for a 3.85-mi, Honolulu artery divided
25      into three links.  Based on data for 80 trips, the most practical  models of third-link exposure
26      (adjusted correlation coefficient [R2] = 0.69) combined three variables:  (1) the ambient CO
27      concentration; (2) the second-link travel time; and (3)  either the travel time, vehicle speed, or CO
28      emission factor for the third link. The models showed that the vehicle's travel time and average
29      speed and the CO emission factor for a given link of the roadway had equal ability to predict
30      passenger cabin exposure to  CO on the third link because of mathematical relationships among
31      these three predictor variables.

        February 15, 1999                         4-39        DRAFT-DO NOT QUOTE OR CITE

-------
 1      4.4.2 Federal and State Policies Affecting Temporal Trends in Exposure
 2           Studies show significant temporal trends in population exposure to CO concentrations from
 3      motor vehicle emissions based on different indicators. One indicator is unintentional death rates
 4      from CO poisoning, and another is based on direct measurements of passenger cabin exposure to
 5      CO concentrations from traffic emissions. Table 4-5 summarizes data on these indicators from
 6      several U.S. studies and shows the federal and California tailpipe CO emission standards by
 7      model year for comparison. In Table 4-5, the net mean CO concentration value represents the
 8      microenvironmental component of total exposure. This value equals the mean in-vehicle CO
 9      concentration, minus the mean ambient CO concentration as recorded simultaneously at a
10      fixed-site monitor.
11
12      4.4.2.1  Effects of Motor Vehicle Emission Standards  on Unintentional Death Rates
13           Based on death certificate reports compiled by the National Center for Health Statistics,
14      Cobb and Etzel (1991) reported statistics on the annual rate of unintentional deaths from CO
15      poisoning in the United States. As shown in Table 4-5, the annual death rate per  100,000
16      population declined from 0.67 in 1979 to 0.39 in 1988. Motor vehicle exhaust gas accounted for
17      6,552 deaths or 56.7% of the total 11,547 unintentional deaths occurring during the 10-year
18      period. The highest death rates per 100,000 persons occurred among males, blacks, the elderly,
19      and those residing in northern states. Monthly variation in death rates indicated a seasonal
20      pattern, with January fatalities routinely about two to five times higher than those occurring in
21      July.
22           Cobb and Etzel (1991) speculated that declining death rates could be attributed in part to
23      reductions in population CO exposure brought about by automaker compliance with the motor
24      vehicle CO emission standards of the CAA. They argued that tighter CO emission standards may
25      enable cars to emit exhaust into an enclosed space for a longer period of time before CO builds
26      up to toxic levels. To test this hypothesis, Pearson's correlation coefficient (r) was computed
27      using the federal emission standard data and annual death rate data shown in Table 4-5.  The
28      value of the correlation coefficient r = 0.54, probability [p] <  0.05 for a one-tailed test of the
29      hypothesis)  supports their argument.
30

        February 15, 1999                         4-40        DRAFT-DO NOT QUOTE OR CITE

-------
  TABLE 4-5. 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
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).
February 15, 1999                       4-41       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. Studies reported
 4      typical (mean or median) CO concentrations for trips, most of which lasted an hour or less.
 5      Mean CO concentrations fell from 37 ppm in 1965, as reported by Haagen-Smit (1966) for a
 6      study in Los Angeles to 3 ppm in 1992 for a study by Lawryk et al. (1995) in the New Jersey
 7      suburbs of New York City. If one assumes that these results are representative of commuter CO
 8      exposures in cities during these time periods, then exposures fell 92% over this 27-year period.
 9      This reduction implies that CO exposure levels reported in the past for a particular place and time
10      in the United States may not be indicative of current exposures.
11           In the  United States,  the effect of progressively tighter CO emission standards on in-vehicle
12      CO exposures over time is readily apparent in Table 4-5. Prior to 1968, each new passenger car
13      emitted 84 g/mi of CO, but by the 1981  model year and thereafter, each new car sold outside of
14      California emitted only 3.4 g/mi of CO, a reduction of 96% (Johnson, 1988). This reduction in
15      certified CO emissions for new passenger cars is roughly equal to the 92% reduction in
16      commuter exposure reported above for the same  period. Further analysis reveals that net mean
17      exposure data and the applicable emission standard data in Table 4-5 are highly correlated
18      r = 0.74, p < 0.0005 for a one-tailed test of the hypothesis). In this analysis, the applicable
19      emission standard (federal or California) was determined by the location of the exposure study.
20      Because the exposure studies did not adhere to a standard protocol, Flachsbart (1995)
21      recommended that future in-vehicle CO exposure studies should use standard protocols to
22      facilitate comparisons and to document  the effect on exposure of future measures taken under
23      motor vehicle emission control programs.
24           Two of the 16 studies did follow a standard protocol. Ott et al. (1994) measured in-vehicle
25      CO concentrations on 88 standardized trips over  a 1-year period in 1980-1981 on a suburban
26      highway near San Jose.  They reported a mean CO concentration of 9.8 ppm for trips of 35 to
27      45 min.  In 1991-1992, Ott et al. (1993) resurveyed this highway using a methodology similar to
28      their previous study to determine in-vehicle exposure trends.  They reported that the mean
29      in-vehicle CO concentration had dropped to 4.6 ppm or 46.9% of the mean value estimated
30      11 years earlier.  They attributed the exposure reduction to  replacement of older vehicles with
31      newer ones that have lower CO emission factors.  This reduction is particularly significant, as

        February 15, 1999                         4-42        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 1991-1992 to 1.6 to 1.8 ppm in 2002-2003.
 7      At the 99% percentile, the model predicted that the CO concentrations would drop from 10 ppm
 8      in 1991-1992 to 4.0 to 4.6 ppm in 2002-2003. This prediction was based on an additional
 9      expected reduction of up to 60% in tailpipe emissions of CO primarily because of continued
10      replacement of older cars with newer, low-emission vehicles. This prediction suggests that
11      CO exposure levels inside passenger cars will virtually disappear with vehicle fleet turnover in
12      the next few years.
13          Similar studies of commuter CO exposure were done by Flachsbart et al. (1987) in the
14      United States, Koushki et al. (1992) in Saudi Arabia, Fernandez-Bremauntz and Ashmore
15      (1995a,b) in Mexico, and Dor et al.  (1995) in France.  These studies used similar methods of data
16      collection and analysis, with one exception. Smoking was allowed for some trips in the Saudi
17      study, but was not allowed in the other studies. Table 4-6 shows typical values of the net mean
18      CO concentration by travel mode for three of the studies. The net mean CO concentration for the
19      Saudi study could not be determined. The net CO concentrations for each travel mode in Mexico
20      City were much higher than for comparable modes in both Washington and Paris, where net CO
21      concentrations were similar.  The similarity between the U.S. and French studies occurred even
22      though catalytic converters existed on 62% of American cars in 1982 (U.S. Department of
23      Commerce, 1983) but were not yet common on French cars in 1992 (Dor et al., 1995).
24          The reasons for the similarity in results between the U.S. and French studies are not readily
25      apparent. However, passenger cabin exposure levels in North and Central America can be
26      explained partly by comparing the history of automotive emission standards in the United States
27      and Mexico.  The United  States initiated nationwide emission standards on new passenger cars in
28      1968 and adopted progressively tighter controls throughout the 1970s (Johnson, 1988).  By the
29      1975 model year,  catalytic converters became standard equipment on new passenger cars.
30      Mexico adopted a tailpipe CO emission standard of 47.0 g/mi for the 1975 model year, and,
31      by the 1993 model year, Mexico finally reached parity with the 1981 U.S. standard of 3.4 g/mi.

        February 15,  1999                         4-43        DRAFT-DO NOT QUOTE OR CITE

-------
          TABLE 4-6.  TYPICAL NET MEAN CARBON MONOXIDE CONCENTRATION RANGES BY TRAVEL MODE
cr
3
FOR CITIES IN THREE COUNTRIES3 b
^2 Washington, DC, USA
£ (1983)
VO
VO
VO
Travel Mode
Automobile
Diesel bus
Rail transit
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
(ppm)
7-10
2-3
1
Averaging Times
(min)
82-106
NA
NA
       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 (1998b).
H

6
o

*
o
H

O
c
o
H
W

O
&
O

H
W

-------
 1      4.4.3 California's No-Smoking Policy
 2           A California law prohibiting smoking in bars, lounges, nightclubs, and restaurants with
 3      bars took effect on January 1, 1998. The law is the first statewide ban of its kind in the nation.
 4      An assessment of the reduction in CO exposure from smoking because of the ban could not be
 5      found in the literature. However, several surveys were made in California to assess public
 6      support and proprietor compliance with the ban amidst media reports of noncompliance. The
 7      surveys did not assess whether total CO exposure from tobacco smoking in California has
 8      declined because of the ban, nor did they determine the extent to which smoking has shifted to
 9      or increased in locations not regulated by the ban.  The results of the surveys have not yet been
10      published in the scientific literature.
11           Shortly after the law took effect, several nonprofit organizations opposed to smoking did
12      field surveys in the San Francisco Bay Area to verify proprietor compliance with the law. The
13      first survey, which occurred in mid-January 1998, in San Mateo County, showed that 36 of
14      42 bars (85.7%) were in compliance (Licavoli, 1998). The second survey focused on bars
15      located in areas with tourist attractions. It reported that 96% of 36 bars in San Francisco's
16      Fisherman's Wharf district were in compliance, as were all  11 of those surveyed in Oakland's
17      Jack London Square (Gordon, 1998).
18           A statewide survey of 1,001 adults (age 21 years and older) was conducted in February
19      and March 1998 to determine public opinion and behavior regarding smoking in bars and
20      similar establishments.  The poll showed that a majority (64%) of respondents visited some kind
21      of bar in the past year. Only 11% of them claimed to visit bars at least once or more per week.
22      Patrons spent an average of 1.5 h per visit, although regular users spent more time than that.
23      The survey found that 75% of bar patrons did not smoke in bars, and that 65% said that  a law
24      intended to prohibit smoking in bars would not affect their patronage (California Department of
25      Health Services, 1998).
26           In February 1998, Communication Sciences Group (1998) reported results of a telephone
27      survey in Los Angeles using a random-digit dialing method. The survey included
28      English-speaking adults  (ages 21 and over) who had visited a nightclub, bar, lounge, or
29      restaurant/bar combination in California since the law took effect.  Of 455 respondents,
30      102 (22%) said they were smokers.  The survey found that 98% of the respondents were aware
31      of the new law; 85% reported that they would either be more likely to visit a smoke-free
        February 15, 1999                        4-45         DRAFT-DO NOT QUOTE  OR CITE

-------
 1     establishment, or that it made no difference; 78% of frequent patrons (one or more visits to a bar
 2     per week) reported either an increased intention or no effect on intention to visit smoke-free
 3     bars; 61% either strongly or somewhat approved the smoking ban; and 70% reported that it was
 4     important to have a smoke-free environment inside these places.
 5
 6     4.4.4  Social Changes Affecting Human Activity Patterns
 7          Between 1965 and 1985, the Americans' Use of Time Project at the University of
 8     Maryland reported that average time spent in travel for all kinds of trips increased from 2.7 to
 9     3.1 h per week (Cornish et al., 1991). Despite this trend, there is evidence that average
10     commuting times between home and work have remained stable. The decennial census
11     collected travel time data for the first time in 1980. By 1990, the census showed that the
12     nation's average commuting time of 21.7 min in 1980 increased only 40 s to 22.4 min in 1990.
13     Although the number of workers who traveled 45 min or more increased from 10.9 million in
14     1980 to 13.9 million in 1990, their mean travel time actually decreased from 59.6  min in 1980 to
15     58.5 min in 1990. One reason for this is that more people were taking their morning commute
16     from home to work during the "shoulder hours" from 6 to 7 a.m.  or from 8 to 9 a.m. than during
17     the "peak hour" from 7 to 8  a.m. In 1990, the shoulder hours accounted for about 37% of
18     worker trip starts, whereas the "peak hour" accounted for only 32% of all trip starts (Pisarski,
19     1992).
20          Typically, average commuting times in large cities are greater than those nationwide.
21     Even so, some cities still show a stable trend in commuting times. For example, the average
22     commuting time for the Washington metropolitan region was 33.5 min in 1957. By 1968, the
23     average commuting time stood at 32.5 min, and it remained at 32.5 min for 1987-1988 when the
24     city was resurveyed. This stability in travel times occurred despite an increase in average
25     commuting distance between 1968 and 1988. In Washington, the average distance of those who
26     commuted alone by automobile from home to work increased from 6.9 miles in 1968 to
27     7.8 miles in 1987-1988, but  their average trip speeds also increased by more than  10.7% to
28     offset the increase in distance (Levinson and Kumar, 1994). Such higher average  trip speeds
29     should contribute to lower commuter exposures, because passenger cabin exposure in the
30     Washington area was shown to be inversely related to travel speed by passenger car (Flachsbart
31     etal., 1987).
       February 15,  1999                         4-46        DRAFT-DO NOT QUOTE OR CITE

-------
 1          In another study of the Washington area based on the same data and time period, Levinson
 2      and Kumar (1995) observed an 85% overall increase in the number of jobs and a decline in
 3      average household size from 3.34 to 2.67 people.  During the 20-year period from 1968 to 1988,
 4      vehicle registrations increased 118%, but road capacity increased only 13%, which led to more
 5      traffic  congestion. The average number of autos per household increased from 1.6 to 2.0.
 6      However, the most important change was a higher percentage of women in the work force,
 7      which  forced readjustments and reallocations of time spent in household activities. Specifically,
 8      workers had more per capita income but spent less time at home  and engaged in more travel for
 9      nonwork trips during peak travel periods.  Compared to 1968, working men spent 20 min less
10      time at home in 1988, and working women spent about 40 min less. Commuters made multiple
11      stops (i.e., trip chaining) on their way home from work (e.g., visiting health clubs, picking up
12      children at day-care, shopping, eating at restaurants). In  1968, such errands and activities
13      usually were done after the primary worker returned home with the household car. By 1988,
14      these trips often were made in separate vehicles by each household member on their way home
15      from work. By 1988, average time spent daily in travel per person in the Washington area had
16      increased by 14 min for workers and by 11  min for nonworkers over 1968.  Levinson and
17      Kumar (1995) said that these results do not support the hypothesis "that individuals spend a
18      fixed amount of time per day (just over 1 h) in transportation, and make all budget allocation
19      adjustment on non-travel times." Instead, they suggested that some urban households have been
20      spending more time in travel and less time  at home,  and have been buying more household
21      services outside the home.
22          On the other hand, Levinson and Kumar (1995) anticipate that some people will spend
23      more time at home in the future.  They noted: "Several factors suggest that work at home,
24      telecommuting, and teleshopping may be on the verge of wide-spread adoption. The technology
25      is coming into place with the long-awaited  advent of videophones, and of the 'information
26      superhighway', that is, broad-band two-way communications facilitated by the recent
27      consolidations in the telecommunications and entertainment industries." The percentage of
28      people working at home increased from 2.3% in 1980 to 3% in 1990 (Pisarski, 1992).
29      Currently, an estimated 52 million Americans are  self-employed  to some extent, working either
30      in home offices for themselves or for companies as telecommuters. In 1975, only 2.5 million


        February 15,  1999                         4-47        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Americans worked at home.  The U.S. General Services Administration expects to see 60,000
 2      telecommuters working for the federal government by the end of 1998.  In 1994, there were
 3      fewer than 4,000 such workers (Webster, 1998).  This employment shift could have tremendous
 4      implications for population exposure.
 5
 6
 7      4.5  CONCLUSIONS
 8           Several conclusions emerge from this review of new studies of population exposure to
 9      carbon monoxide.  First, fixed-site ambient CO monitors do not necessarily represent the
10      outdoor CO concentrations in a given urban location because of spatial and temporal variations
11      in traffic patterns.  Study has shown that fixed-site CO monitors often underestimate outdoor
12      CO concentrations at the locations nearer traffic sources and overestimate the CO
13      concentrations at the locations further from traffic sources than the fixed-site location. If the
14      average ambient CO concentration measured at a fixed monitoring station adequately represents
15      the average outdoor CO concentration in the surrounding community (e.g, within ±1 or 2 ppm),
16      the station may provide a reasonable estimate of the community's mean exposure to CO  of
17      ambient origin. The station will not adequately represent the exposures to CO of the people
18      residing in the community while they are exposed to tobacco smoke; motor vehicle exhaust
19      during commutes to and from work, school, etc.;  and occupational and residential sources of
20      fuel combustion. The mean COHb level of people exposed to these sources will be greater than
21      their mean COHb level predicted solely from  exposure to CO of ambient origin.
22           Second, implementation of motor vehicle emission standards, catalytic converters,
23      inspection/maintenance programs, and cleaner burning fuels over the past three decades has
24      reduced the CO exposures of motorists, pedestrians,  and bicyclists on streets and highways.
25      This suggests that estimates of current population exposure based on pre-1990 exposure  studies
26      may no longer apply. Presently, there is no good estimate of the CO exposure distribution for
27      the current population.  Moreover, average CO concentrations in passenger cabins  of motor
28      vehicles are expected to drop further in the near future, based on mathematical projections for a
29      standardized  trip on an arterial highway (Yu et al., 1996). Further changes in exposure may
30      occur because of certain trends in travel behavior caused by growth rates in vehicle miles of
31      travel, growth in travel during the shoulder hours of peak traffic periods, and greater use of
        February 15,  1999                          4-48        DRAFT-DO NOT QUOTE OR CITE

-------
 1      computers to telecommute and teleshop from home.  The effects of these trends on population
 2      exposure have not been studied fully. Also, the effect of transportation control measures on
 3      commuter exposure has received only limited study.
 4           Third, personal CO exposures that exceed the NAAQS still occur among nonsmokers in
 5      certain locations and atypical situations.  These include occasions when people (usually
 6      children) ride in the back of vehicles with leaky rear doors (e.g., school bus, covered pickup
 7      truck, ambulance) or ride certain recreational vehicles (i.e., snowmobiles, power boats). It also
 8      includes a growing number of events where people gather to barbecue food or watch sporting
 9      events held at indoor arenas used for ice skating, motocross and monster-truck competitions,
10      and tractor-pulls, because the vehicles in these events lack emission controls.  In some cases,
11      ventilation has not lowered CO concentrations sufficiently. High CO exposures also still occur
12      in small homes with malfunctioning and poorly ventilated gas and kerosene appliances.
13      Moreover, unintentional poisonings and fatalities are still reported when people use
14      gasoline-powered appliances, engines, and tools (e.g., chainsaws), even under ventilated
15      conditions.
16           Fourth, CO exposures are still high in certain occupations such as charcoal meat grillers,
17      garage mechanics, firefighters, forklift operators, and those who work in facilities adjacent to
18      forklift operations.  Also, more than one million people have high exposure to methylene
19      chloride in the workplace, where it is used as a degreaser, paint remover, aerosol propellant, and
20      blowing agent. Another concern is concurrent exposure to  methylene  chloride and methanol,
21      which prolongs the period of elevated COHb in the body.
22           Fifth, the California and national activity pattern studies indicate that Americans spend, on
23      average, between 87 and 89% of their day indoors and about 7% of time in or near vehicles.
24      However, U.S. Census and metropolitan travel behavior surveys indicate that, although average
25      commuting times from home to work have fallen, some Americans in large cities are spending a
26      growing amount of time in surface travel compared with the past and less time at home. As a
27      result, they have been buying more household services outside the home.  This suggests that
28      individuals do not have fixed daily travel-time budgets in the long term.
29           Finally, there are implications of this literature review for activity pattern studies and
30      exposure models.  The California and national activity pattern studies described time spent by
31      conventional social categories (e.g., gender, age, race, etc.) These studies were not guided by

        February 15, 1999                         4-49         DRAFT-DO NOT QUOTE OR CITE

-------
 1     any theoretical framework as was the Boston study, which was able to explain activity patterns
 2     more convincingly. Hence, although the Boston study did not focus on exposure assessment, it
 3     still has profound implications for any study that analyzes activity pattern data.  The lesson for
 4     simulation models is that they need to sample from distributions that more accurately represent
 5     current microenvironmental pollutant concentrations and time budgets, which have been
 6     affected by public policies affecting motor vehicle emissions, travel behavior in cities, and
 7     smoking in public places.  Moreover, exposure models should not assume that time distributions
 8     in different microenvironments are independent of each other.  Modern households appear to be
 9     making continuous tradeoffs in their activity patterns and role allocations, in terms of who gets
10     to do what and when, to better adapt to ongoing technological and social changes affecting
11     society.
12
       February 15, 1999                         4-50        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. (1998) Urban commuter exposure to particle matter and carbon monoxide
  8             inside an automobile. J. Exposure Anal. Environ. Epidemiol. [accepted].
  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 a
12             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       Barnard, R. J.; Weber, J. S. (1979) Carbon monoxide: a hazard to firefighters. Arch. Environ. Health 34: 255-257.
16       Baucom, C. D.; Freeman, J. I.; MacCormack, J. M. (1987) Carbon monoxide poisoning in a
17             garment-manufacturing plant - North Carolina. Morb. Mortal. Wkly. Rep. 36: 543-545.
18       Bellin, P.; Spengler, J. D. (1980) Indoor and outdoor carbon monoxide measurements at an airport. J. Air Pollut.
19             Control Assoc. 30: 392-394.
20       Bevan, M.  A. J.; Proctor, C. J.; Baker-Rogers, J.; Warren, N. D. (1991) Exposure to carbon monoxide, respirable
21             suspended particulates, and volatile organic compounds while commuting by bicycle. Environ. Sci. Technol.
22             25:788-791.
23       Boeniger, M. F. (1995) Description of facilities and employee exposure to  carbon monoxide at four different
24             bridge and tunnel toll authorities in the United States. Cincinnati, OH: National Institute for Occupational
25             Safety and Health, Division of Surveillance, Field Studies, Hazard Evaluations; report no. IWS-45-02.
26             Available from: NTIS, Springfield, VA; PB96-106539.
27       Boudreau, D. R.; Spadafora, M. P.; Wolf, L. R.; Siegel, E. (1994) Carbon monoxide levels during indoor sporting
28             events—Cincinnati, 1992-1993 [reprint fromMMWR 43: 21-23]. JAMA J. Am. Med. Assoc. 271:  419.
29       Brandt-Rauf, P. W.; Fallen, L. F., Jr.; Tarantini, T.; Idema, C.; Andrews, L. (1988) Health hazards of fire fighters:
30             exposure assessment. Br. J. Ind. Med. 45: 605-612.
31       Brotherhood, J. R.; Budd, G. M.; Jeffery, S. E.; Hendrie, A. L.; Beasley, F. A.; Costin, B. P.; Wu, Z. (1990) Fire
32             fighters' exposure to carbon monoxide during Australian bushfires. Am. Ind. Hyg. Assoc.  J. 51:  234-240.
33       Buie, S. E.; Pratt, D. S.; May, J. J. (1986) Diffuse pulmonary injury following paint remover exposure. Am. J.
34             Med. 81:702-704.
35       Btinger, J.; Bombosch, F.; Mesecke, U.; Hallier, E. (1997) Monitoring and analysis of occupational exposure to
36             chain saw exhausts. Am. Ind. Hyg. Assoc. J. 58: 747-751.
37       Burgess, W. A.; Sidor, R.; Lynch, J. J.; Buchanan, P.; Clougherty, E. (1977) Minimum protection factors for
38             respiratory protective devices for firefighters. Am. Ind. Hyg. Assoc. J. 38: 18-23.
39       California Department of Health Services. (1998) A survey of California bar patrons about smoking policies and
40             smoke-free bars. Sacramento, CA: Tobacco Control Section.
41       Chan, C.-C.;  Ozkaynak, H.; Spengler, J. D.; Sheldon, L. (1991) Driver exposure to volatile organic compounds,
42             CO, ozone and NO2 under different driving conditions. Environ. Sci. Technol. 25: 964-972.
43       Chapin, F.  S., Jr. (1974) Human activity patterns in the city: things people do in time and in space. New York, NY:
44             John Wiley  & Sons.
45       Chung, Y.; Park, S. E.; Lee, K.; Yanagisawa, Y.; Spengler, J. D. (1994) Determinations of personal carbon
46             monoxide exposure and blood carboxyhemoglobin levels in Korea. Yonsei Med. J. 35: 420-428.
47       Clements, J. S. (1978) School bus carbon monoxide intrusion. Washington, DC: U.S. Department of
48             Transportation, National Highway Traffic Safety Administration; report no. DOT-HS-803 705. Available
49             from: NTIS, Springfield, VA; PB-291156/8.
50       Cobb, N.; Etzel, R. A. (1991) Unintentional carbon monoxide-related deaths in the United States, 1979 through
51             1988. JAMA J. Am. Med. Assoc. 266: 659-663.
52
         February 15, 1999                             4-51          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Colome, S. D.; Wilson, A. L.; Tian, Y. (1994) California residential indoor air quality study. Volume 2. Carbon
  2             monoxide and air exchange rate: an univariate and multivariate analysis. Chicago, IL: Gas Research
  3             Institute; report no. GRI-93/0224.3. Available from: NTIS, Springfield, VA; PB96-159421.
  4       Communication Sciences Group. (1998) Smoke-free bar patron survey: summary of findings. Los Angeles, CA:
  5             Los Angeles County Department of Health.
  6       Cornish, E.; Willard, T.; Fields, D. (1991) How Americans use time: an interview with sociologist John P.
  7             Robinson. Futurist Sept./Oct: 23-27.
  8       Cortese, A. D.; Spengler, J. D. (1976) Ability of fixed monitoring stations to represent personal carbon monoxide
  9             exposure. J. Air Pollut. Control Assoc. 26: 1144-1150.
10       Cutler, B. (1990) Where does the free time go? Am. Demograph. (November): 36-39.
11       Dor, F.; Le Moullec, Y.; Festy, B. (1995) Exposure of city residents to carbon monoxide and monocyclic aromatic
12             hydrocarbons during commuting trips in the Paris metropolitan area. J. Air Waste Manage. Assoc.
13             45: 103-110.
14       Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
15       Ely, E. W.; Moorehead, B.; Haponik, E. F. (1995) Warehouse workers' headache: emergency evaluation and
16             management of 30 patients with carbon monoxide poisoning. Am. J. Med. 98:  145-155.
17       Faiz, A.; Weaver, C.  S.; Walsh, M. P.; Gautam, S. P.; Chan, L.-M. (1996) Emission standards and regulations.
18             In: Air pollution from motor vehicles:  standards and technologies for controlling emissions. Washington,
19             DC: World Bank; pp.  1-24.
20       Farrow, A.; Taylor, H.; Golding, J. (1997) Time spent in the home by different family members. Environ. Technol.
21             18:605-613.
22       Fawcett, T. A.; Moon, R. E.; Fracica, P. J.; Mebane, G. Y.; Theil, D. R.; Piantadosi, C. A. (1992) Warehouse
23             workers' headache: carbon monoxide poisoning from propane-fueled forklifts.  J.  Occup. Med. 34: 12-15.
24       Federal Register. (1992) Guidelines for exposure assessment. F. R. (May 29) 57: 22887-22938.
25       Fernandez-Bremauntz, A. A.; Ashmore, M. R. (1995a) Exposure of commuters to carbon monoxide in Mexico
26             City—I. Measurement of in-vehicle concentrations. Atmos. Environ.  29: 525-532.
27       Fernandez-Bremauntz, A. A.; Ashmore, M. R. (1995b) Exposure of commuters to carbon monoxide in Mexico
28             City II. Comparison of in-vehicle and fixed-site concentrations. J. Exposure Anal. Environ. Epidemiol.
29             5:497-510.
30       Flachsbart, P. G. (1989) Effectiveness of priority lanes in reducing travel time and carbon monoxide exposure.
31             Inst. Transport. Eng. J. 59: 41-45.
32       Flachsbart, P. G. (1995) Long-term trends in United States highway emissions, ambient concentrations, and
33             in-vehicle exposure to carbon monoxide in traffic. J. Exposure Anal.  Environ. Epidemiol. 5: 473-495.
34       Flachsbart, P. G. (1998a) Models of exposure to carbon monoxide inside a vehicle on a Honolulu highway.
35             J. Exposure Anal. Environ. Epidemiol. [accepted].
36       Flachsbart, P. G. (1998b) Human exposure to carbon monoxide from mobile sources. Chemosphere [accepted].
37       Flachsbart, P. G.; Brown, D.  E. (1989) Employee exposure to motor vehicle exhaust at a Honolulu shopping
38             center. J. Architect. Plan. Res. 6: 19-33.
39       Flachsbart, P. G.; Mack, G. A.; Howes, J. E.; Rodes, C. E. (1987) Carbon monoxide exposures of Washington
40             commuters.  JAPCA 37: 135-142.
41       Gardiner, K.; Trethowan, W. N.; Harrington,  J. M.; Calvert, I. A.; Glass, D.  C. (1992) Occupational exposure to
42             carbon monoxide and  sulphur dioxide  during the manufacture of carbon black. Ann. Occup. Hyg.
43             36: 363-372.
44       Gasman, J. D.; Varon, J.; Gardner, J. P. (1990) Revenge of the barbecue grill: carbon monoxide poisoning.
45             West.  J. Med.  153: 656-657.
46       Ghittori, S.; Marraccini, P.; Franco, G.; Imbriani, M. (1993) Methylene chloride exposure in industrial workers.
47             Am. Ind. Hyg. Assoc.  154:27-31.
48       Girman, J. R.; Chang, Y.-L.;  Hayward, S. B.; Liu, K.-S. (1998) Causes of unintentional deaths from carbon
49             monoxide poisonings in California. West. J. Med. 168:  158-165.
50       Gold, A.; Burgess, W. A.; Clougherty, E. V. (1978) Exposure of firefighters to toxic air contaminants. Am. Ind.
51             Hyg. Assoc. J. 39: 534-539.
52       Gordon, B. (1998) Easing the transition to smoke-free bars. Contra Costa-Solano Counties, CA: American Lung
53             Association of Contra Costa-Solano.
54
         February 15, 1999                             4-52          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Gourdeau, P.; Parent, M; Soulard, A. (1995) Exposition a 1'oxyde de carbone dans les garages d'automobiles:
  2             evaluation chez les mecaniciens [Exposure of carbon monoxide in car repair shops: evaluation of
  3             mechanics]. Can. J. Public Health 86: 414-417.
  4       Haagen-Smit, A. J. (1966) Carbon monoxide levels in city driving. Arch. Environ. Health 12: 548-551.
  5       Hampson, N. B. (1996) Carbon monoxide poisoning at an indoor ice arena and bingo hall-Seattle, 1996 [reprint of
  6             MMWR 1996; 45: 265-267]. JAMA J. Am. Med. Assoc. 275: 1468-1469.
  7       Hampson, N. B.; Norkool, D. M. (1992) Carbon monoxide poisoning in children riding in the back of pickup
  8             trucks. JAMA J. Am. Med. Assoc. 267: 538-540.
  9       Hampson, N. B.; Kramer, C. C.; Dunford, R. G.; Norkool, D. M. (1994) Carbon monoxide poisoning from indoor
10             burning of charcoal briquets. JAMA J. Am. Med. Assoc. 271: 52-53.
11       Holland, D. M. (1983) Carbon monoxide levels in microenvironment types of four U.S. cities. Environ. Int.
12             9:369-377.
13       Jankovic, J.; Jones, W.; Burkhart, J.; Noonan, G. (1991) Environmental study of firefighters. Ann. Occup.
14             Hyg. 35: 581-602.
15       Jarvis, M. J.; Russell, M. A. H.; Saloojee, Y. (1980) Expired air carbon monoxide: a simple breath test of tobacco
16             smoke intake. Br. Med. J. 281: 484-485.
17       Jenkins, P. L.; Phillips, T. J.; Mulberg, E. J.; Hui, S. P. (1992) Activity patterns of Californians: use of and
18             proximity to indoor pollutant sources. Atmos. Environ.  Part A 26: 2141-2148.
19       Johnson, T. (1987) A study of human activity patterns in Cincinnati, Ohio. Palo Alto, CA: Electric Power
20             Research Institute; contract no. RP940-06.
21       Johnson, J. H. (1988) Automotive emissions. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the
22             automobile, and public health. Washington, DC: National Academy Press; pp. 39-75.
23       Johnson, T.; Paul, R. A. (1983) The NAAQS Exposure Model (NEM) applied to carbon monoxide. Research
24             Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
25             EPA report no. EPA-450/5-83-003.  Available from: NTIS, Springfield, VA; PB84-242551.
26       Johnson, T.; Capel, J.; Paul, R.; Wijnberg,  L. (1992) Estimation of carbon monoxide exposures and associated
27             carboxyhemoglobin levels in Denver residents using a probabilistic version of NEM. Research Triangle
28             Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; contract
29             no. 68-DO-0062.
30       Johnson, T.; Mihlan, G.; LaPointe, J.; Fletcher, K.; Capel, J. (1999) Estimation of carbon monoxide exposures and
31             associated carboxyhemoglobin levels in Denver residents using pNEM/CO (version 2.0) [draft]. Research
32             Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
33       Jones, R.  H.; Ellicott, M. F.; Cadigan, J. B.; Gaensler, E. A. (1958) The relationship between alveolar and blood
34             carbon monoxide concentrations during breathholding.  J. Lab. Clin. Med. 51: 553-564.
35       Joumard,  R. (1991) Pollution de 1'air dans les transports [Air pollution from transportation]. Pollut. Atmos.
36             33:20-24.
37       Kahler, M.; Kuhse, W.; Wintermeyer, L. A. (1993) Unintentional carbon monoxide poisoning from indoor use of
38             pressure washers—Iowa, January 1992-January 1993. Morb. Mortal. Wkly. Rep. 42: 777-779, 785.
39       Kamei, M.; Yanagisawa, Y. (1997) Estimation of CO exposure of road construction workers in tunnel. Ind. Health
40             35: 119-125.
41       Kern, R. A.; Knoedler, J. A.; Trester, R. J.  (1990) Automotive carbon monoxide emissions in a closed garage.
42             Presented at:  International congress  and exposition; February-March; Detroit, MI. Warrendale, PA:  Society
43             of Automotive Engineers; pp. 33-38. (SAE technical paper series no. 900499).
44       Klepeis, N. E.; Tsang, A. M.; Behar, J. V. (1996) Analysis of the national human activity pattern survey (NHAPS)
45             respondents from a standpoint of exposure assessment.  Washington, DC: U.S. Environmental Protection
46             Agency, Office of Research and Development;  report no. EPA/600/R-96/074.
47       Koontz, M. D.; Nagda, N. L. (1988) A topical report on a field monitoring study of homes with unvented gas space
48             heaters. Volume III—methodology and results.  Chicago, IL: Gas Research Institute; report no.
49             GRI-87/0044.2. Available from: NTIS, Springfield, VA; PB88-216791.
50       Koushki, P. A.; Al-Dhowalia, K. H.; Niaizi, S. A. (1992) Vehicle occupant exposure to carbon monoxide. J. Air
51             Waste Manage. Assoc. 42: 1603-1608.
52       Kwok, P.  W. (1981) Reduction of carbon monoxide in indoor  skating arenas. Environ. Health Rev. 25: 60-63.
53       Lando, H. A.; McGovern, P. G.; Kelder, S. H.; Jeffery, R. W.; Forster,  J. L. (1991) Use of carbon monoxide breath
54             validation in assessing exposure to cigarette smoke in a worksite population. Health Psychol. 10: 296-301.
         February 15, 1999                             4-53          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Last, J. M; Abramson, J. H.; Friedman, G. D.; Porta, M.; Spasoff, R. A.; Thuriaux, M. (1995) A dictionary of
  2             epidemiology. 3rd ed. New York, NY: Oxford University Press; pp. 5, 60.
  3       Law, P. L.; Lioy, P. J.; Zelenka, M. P.; Huber, A. H.; McCurdy, T. R. (1997) Evaluation of a probabilistic
  4             exposure model applied to carbon monoxide (pNEM/CO) using Denver personal exposure monitoring data.
  5             J. Air Waste Manage. Assoc. 47: 491-500.
  6       Lawryk, N. J.; Lioy, P. J.; Weisel, C. P. (1995) Exposure to volatile organic  compounds in the passenger
  7             compartment of automobiles during periods of normal and malfunctioning operation. J. Exposure Anal.
  8             Environ. Epidemiol. 5: 511-531.
  9       Lee, K.; Yanagisawa, Y. (1992) Development of a sampler for carbon monoxide in expired air. Presented at:
10             85th annual meeting and exhibition of the Air & Waste Management Association; June; Kansas City, MO.
11             Pittsburgh, PA: Air & Waste Management Association; paper no. 92-145.03.
12       Lee, K.; Yanagisawa, Y. (1995) Sampler for measurement of alveolar carbon monoxide. Environ. Sci. Technol.
13             29:  104-107.
14       Lee, K.; Yanagisawa, Y.; Spengler, J. D. (1993) Carbon monoxide and nitrogen dioxide levels in an indoor ice
15             skating rink with mitigation methods. Air Waste 43:  769-771.
16       Lee, K.; Yanagisawa, Y.; Spengler, J. D.; Nakai, S. (1994)  Carbon monoxide and nitrogen dioxide exposures in
17             indoor ice skating rinks. J. Sports Sci. 12: 279-283.
18       Lees, P. S. J. (1995) Combustion products and other firefighter exposures. Occup. Med.: State of the Art Rev.
19             10:691-706.
20       Leikin, J. B.; Kaufman, D.; Lipscomb, J. W.; Burda, A. M.; Hryhorczuk, D.  O. (1990) Methylene chloride: report
21             of five exposures and two deaths. Am. J. Emerg. Med. 8: 534-537.
22       Levesque,  B.; Dewailly, E.;  Lavoie, R.; Prud'Homme, D.; Allaire, S. (1990) Carbon monoxide in indoor ice skating
23             rinks: evaluation of absorption by adult hockey players.  Am. J. Public Health 80: 594-598.
24       Levesque,  B.; Lavoie, R.; Dewailly, E.; Prud'Homme, D.; Allaire, S. (1991) An experiment to evaluate carbon
25             monoxide absorption by hockey players in ice skating rinks. Vet. Hum. Toxicol. 33: 5-8.
26       Levesque,  B.; Allaire, S.; Prud'homme, H.; Rhainds, M.; Lebel, G.; Bellemarre, D.; Dupuis, K. (1997) Indoor
27             motocross competitions: air quality evaluation. Am.  Ind. Hyg. Assoc. J. 58: 286-290.
28       Levinson, D. M.; Kumar, A. (1994) The rational locator: why travel times have remained stable. J. Am. Plan.
29             Assoc. 60:319-332.
30       Levinson, D.; Kumar, A. (1995) Activity, travel, and the  allocation of time. J. Am. Plan. Assoc. 61: 458-470.
31       Licavoli, K. (1998) Bar survey results for select cities in San Mateo county.  San Francisco, CA: American Lung
32             Association of San Francisco and San Mateo Counties.
33       Limasset, J.-C.; Diebold, F.; Hubert, G. (1993) Exposition des conducteurs de bus urbains aux polluants de la
34             circulation automobile [Assessment of bus drivers' exposure to the pollutants of urban traffic]. Sci. Total
35             Environ. 134:  39-49.
36       Lioy, P. J.  (1990) Assessing total human exposure to contaminants: a multidisciplinary approach. Environ. Sci.
37             Technol. 24: 938-945.
38       Liu, K.-S.; Girman, J. R.; Hayward, S. B.; Shusterman, D.;  Chang, Y.-L. (1993) Unintentional carbon monoxide
39             deaths in California from charcoal grills and hibachis. J. Exposure Anal. Environ. Epidemiol.
40             3(suppl. 1): 143-151.
41       Luckhurst, D. G.; Solkoski,  G. (1990) Carbon monoxide levels in indoor "tractor-pull" events—Manitoba. Can.
42             Dis. Wkly. Rep. 16-17: 79-81.
43       Madani, I.  M.; Khalfan,  S.; Khalfan, H.; Jidah, J.; Aladin, M. N. (1992) Occupational exposure to carbon
44             monoxide during  charcoal meat grilling. Sci. Total Environ. 114: 141-147.
45       Materna, B. L.; Jones, J. R.; Sutton, P. M.; Rothman, N.; Harrison, R. J. (1992) Occupational exposures in
46             California wildland fire fighting. Am. Ind. Hyg. Assoc. J. 53: 69-76.
47       Materna, B. L.; Koshland, C. P.;  Harrison, R. J. (1993) Carbon monoxide exposure in wildland firefighting:
48             a comparison of monitoring methods. Appl. Occup. Environ. Hyg. 8: 479-487.
49       McCammon, C. S.; Halperin, W. F.; Lemen, R. A. (1981) Carbon monoxide exposure from aircraft fueling
50             vehicles. Arch. Environ. Health 36: 136-138.
51       McCammon, J. B.; McKenzie, L. E.; Heinzman, M. (1996) Carbon monoxide poisoning related to the indoor use
52             of propane-fueled forklifts in Colorado workplaces.  Appl. Occup. Environ. Hyg. 11: 192-198.
53       McCurdy,  T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
54             models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
55       Meyer, B. (1983)  Indoor air quality. Reading, MA: Addison-Wesley Publishing Company, Inc.


         February 15, 1999                             4-54          DRAFT-DO NOT QUOTE  OR CITE

-------
  1       Miller, R. K.; Ryan, M. C.; Bilowus, P. (1989) Carbon monoxide poisoning in indoor ice skating arenas. Va. Med.
  2             116:74-76.
  3       Mumford, J. L.; Williams, R. W.; Walsh, D. B.; Burton, R. M.; Svendsgaard, D. I; Chuang, J. C.; Houk, V. S.;
  4             Lewtas, J. (1991) Indoor air pollutants from unvented kerosene heater emissions in mobile homes: studies
  5             on particles, semivolatile organics, carbon monoxide, and mutagenicity. Environ. Sci. Technol.
  6             25: 1732-1738.
  7       National Center for Health Statistics and U.S. Consumer Product Safety Commission. (1992) Death certificate file.
  8             Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease
  9             Control and Prevention.
10       National Institute for Occupational Safety and Health. (1994) Health hazard evaluation report HETA
11             90-0365-2415, U.S. Department of the Interior, National Park Service, Yosemite National Park, California.
12             Cincinnati, OH: U.S. Department of Health and Human Services, Public Health Service, Centers for
13             Disease Control and Prevention. Available from: NTIS, Springfield, VA; PB95-242541.
14       National Institute for Occupational Safety and Health. (1996) Preventing carbon monoxide poisoning from small
15             gasoline-powered engines and tools [NIOSH alert]. Cincinnati, OH: U. S. Department of Health and Human
16             Services, Public Health Service; report no. DHHS (NIOSH) publication no. 96-118. Available from: NTIS,
17             Springfield, VA; PB97-147078.
18       Ott, W. R. (1982) Concepts of human exposure to air pollution. Environ. Int. 7: 179-196.
19       Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
20             21:97-128.
21       Ott, W.; Eliassen, R.  (1973) A survey technique for determining the  representativeness of urban air monitoring
22             stations with respect to carbon monoxide. J. Air Pollut. Control Assoc. 23: 685-690.
23       Ott, W.; Flachsbart, P. (1982) Measurement of carbon monoxide concentrations in indoor and outdoor locations
24             using personal exposure monitors. Environ. Int. 8: 295-304.
25       Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
26             exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
27             Environ. 22:2101-2113.
28       Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities
29             for human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
30       Ott, W.; Switzer, P.; Willits, N.; Hildemann, L.; Yu, L. (1993) Trends of in-vehicle CO exposures on a California
31             arterial highway over one decade. Research Triangle Park, NC: U.S. Environmental Protection Agency,
32             Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/A-93/220. Available
33             from: NTIS, Springfield, VA; PB93-234508.
34       Ott, W.; Switzer, P.; Willits, N. (1994) Carbon monoxide exposures inside an automobile traveling on an urban
35             arterial highway. J. Air Waste Manage. Assoc. 44: 1010-1018.
36       Paul, R. A.; Johnson, T. (1985) The NAAQS exposure model (NEM) applied to carbon monoxide: addendum.
37             Research Triangle Park, NC: U.S. Environmental Protection Agency,  Office of Air Quality Planning and
38             Standards; report no. EPA 450/5-85-004. Available from: NTIS, Springfield, VA; PB85-235182.
39       Paul, R. A.; Johnson, T.; McCurdy, T. (1988) Advancements in estimating urban population exposure. Presented
40             at:  81st annual meeting of the Air Pollution Control Association; June; Dallas, TX. Pittsburgh, PA: Air
41             Pollution Control Association; paper no. 88-127.1.
42       Paulozzi,  L. J.; Satink, F.; Spengler, R. F. (1991) A carbon monoxide mass poisoning in an ice arena in Vermont
43             [letter]. Am. J. Public Health 81: 222.
44       Paulozzi,  L. J.; Spengler, R. F.; Vogt, R. L.; Carney, J. K. (1993) A survey of carbon monoxide and nitrogen
45             dioxide in indoor ice arenas in Vermont. J. Environ. Health 56(5): 23-25.
46       Pedersen, D. H.;  Sieber, W. K. (1988) National occupational exposure survey. Volume III: analysis of management
47             interview responses. Cincinnati, OH: U.S. Department of Health and Human Services, National Institute for
48             Occupational Safety and Health; DHHS (NIOSH) publication no.  89-103. Available from: NTIS,
49             Springfield, VA; PB90-193640.
50       Peterson,  J. E. (1978) Modeling the uptake, metabolism and excretion of dichloromethane by man. Am. Ind. Hyg.
51             Assoc. J. 39: 41-47.
52       Phillips, T. J.; Jenkins, P. L.;  Mulberg, E. J. (1991) Children in California: activity patterns and presence of
53             pollutant sources. In: Health risk and risk communication: papers from the 84th annual meeting &
54             exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada. Pittsburgh, PA:
55             Air & Waste Management Association; volume 17, paper no. 91-172.5.


         February 15, 1999                            4-55         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Pisarski, A. E. (1992) New perspectives in commuting: based on early data from the 1990 decennial census and the
  2             1990 Nationwide Personal Transportation Study. Washington, DC: U.S. Department of Transportation,
  3             Federal Highway Administration; report no. FHWA-PL/92-026. Available from: NTIS, Springfield, VA;
  4             PB97-165567.
  5       Quackenboss, J. J.; Spengler, J. D.; Kanarek, M. S.; Letz, R.; Duffy, C. P. (1986) Personal exposure to nitrogen
  6             dioxide: relationship to indoor/outdoor air quality and activity patterns. Environ. Sci. Technol. 20: 775-783.
  7       Raaschou-Nielsen, O.; Nielsen, M. L.; Gehl, J.  (1995) Traffic-related air pollution: exposure and health effects in
  8             Copenhagen street cleaners and cemetery workers. Arch. Environ. Health 50: 207-213.
  9       Radford, E. P.; Drizd, T. A. (1982) Blood carbon monoxide levels in persons 3-74 years of age: United States,
10             1976-80. Hyattsville, MD: U.S. Department of Health and Human Services, National Center for Health
11             Statistics; DHHS publication no. (PHS)  82-1250. (Advance data from vital and health statistics: no. 76).
12       Radford, E. P.; Levine, M. S. (1976) Occupational exposures to carbon monoxide in Baltimore firefighters.
13             J. Occup. Med. 18: 628-632.
14       Robinson, J. P.; Thomas, J. (1991) Time spent in activities, locations, and microenvironments:
15             a California-national comparison. Las Vegas, NV: U.S. Environmental Protection Agency, Environmental
16             Monitoring Systems Laboratory; report no. EPA/600/4-91/006. Available from: NTIS, Springfield, VA;
17             PB92-140789.
18       Schwab, M.; Colome, S. D.; Spengler, J. D.; Ryan, P. B.; Billick, I. H. (1990) Activity patterns applied to pollutant
19             exposure assessment: data from a personal monitoring study in Los Angeles. Toxicol. Ind. Health
20             6:517-532.
21       Seufert, K. T.; Kiser, W. R. (1996) End-expiratory carbon monoxide levels as an estimate of passive smoking
22             exposure aboard a nuclear-powered submarine. South. Med. J. 89: 1181-1183.
23       Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
24             models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile,  and public health.
25             Washington, DC: National Academy Press; pp. 207-238.
26       Shenoi, R.; Stewart, G.; Rosenberg, N. (1998) Screening for carbon monoxide in children. Pediatr. Emer. Care
27             [in press].
28       Shusterman, D.; Quinlan, P.; Lowengart, R.; Cone, J. (1990) Methylene chloride intoxication in a furniture
29             refinisher: a comparison of exposure estimates utilizing workplace air sampling and blood
30             carboxyhemoglobin measurements. J. Occup. Med. 32: 451-454.
31       Silvers, A.; Florence, B. T.; Rourke, D. L.; Lorimor, R. J. (1994) How children spend their time: a sample survey
32             for use in exposure and risk assessments. Risk Anal. 14: 931-944.
33       Simeone, L. F.  (1991) The intrusion of engine exhaust into the passenger areas of recreational power boats.
34             Washington, DC: U.S. Coast Guard, Office of Navigation Safety and Waterway Services; report no.
35             DOT-VNTSC-CG-91-1. Available from: NTIS,  Springfield, VA; AD-A254146.
36       Smith, N. J. (1977) End-expired air technic for determining occupational carbon monoxide exposure. J. Occup.
37             Med. 19: 766-769.
38       Smith, W.; Anderson, T.; Anderson, H. A.; Remington, P. L. (1992) Nitrogen dioxide and carbon monoxide
39             intoxication in an indoor ice arena - Wisconsin,  1992. Morb. Mortal. Wkly. Rep. 41: 383-385.
40       Snook, L. M. (1996) An investigation of driver exposure to carbon monoxide while traveling in the wake of a
41             snowmobile [dissertation]. Knoxville, TN: University of Tennessee. Available from: Xerox University
42             Microfilms, Ann Arbor MI; order no. AAD97-09060.
43       Snook, L. M.; Davis, W. T. (1997) An investigation of driver exposure to carbon monoxide while traveling in the
44             wake of a snowmobile. Presented at: 90th annual meeting & exhibition of the Air & Waste Management
45             Association; Toronto, ON, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no.
46             97-RP143.02.
47       Sorensen, A. J. (1986) The importance of monitoring carbon monoxide levels in indoor ice skating rinks. J. Am.
48             Coll. Health 34:  185-186.
49       Spengler, J. D.; Stone, K. R.; Lilley, F. W. (1978) High carbon monoxide levels measured in enclosed skating
50             rinks. J. Air Pollut. Control Assoc. 28: 776-779.
51       Stewart, R. D.; Hake, C. L. (1976) Paint-remover hazard. JAMA J. Am. Med. Assoc. 235: 398-401.
52       Stewart, R. D.; Fisher, T. N; Hosko, M. J.; Peterson,  J.  E.; Baretta, E. D.; Dodd, H. C. (1972) Experimental human
53             exposure to methylene chloride. Arch. Environ. Health 25: 342-348.
54
         February 15, 1999                             4-56         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Systems Applications, Inc. (1983) Human exposure to atmospheric concentrations of selected chemicals.
  2             Volume II. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
  3             Planning and Standards; EPA contract no. 68-02-3066. Available from: NTIS, Springfield, VA;
  4             PB83-265249.
  5       Szalai, A., ed. (1972) The use of time: daily activities of urban and suburban populations in twelve countries.
  6             The Hague, The Netherlands: Mouton & Co.
  7       Treitman, R. D.; Burgess, W. A.; Gold, A. (1980) Air contaminants encountered by firefighters. Am. Ind. Hyg.
  8             Assoc. 141:796-802.
  9       Tsang, A. M; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
10             activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
11             Research and Development; report no. EPA/600/R-96/148.
12       U.S. Code. (1990) Clean Air Act,  as amended by PL 101-549, November 15, 1990. U. S. C. 42: §§7401-7671q.
13       U.S. Department of Commerce. (1983) Statistical abstract of the United States: 1982-83. National data book and
14             guide to sources. 103rd  ed.  Washington, DC: Bureau of the Census; p. 618.
15       U.S. Department of Transportation. (1996) Transportation and air quality: selected facts and figures. Washington,
16             DC: Federal Highway Administration; FHWA/PD/96-006; p. 6. Available from: NTIS, Springfield, VA;
17             PB97-110613.
18       U.S. Environmental Protection Agency. (1985) Health assessment document for dichloromethane (methylene
19             chloride). Final report. Research Triangle Park, NC: Office of Health and Environmental Assessment,
20             Environmental Criteria and Assessment Office; report no.  EPA-600/8-82/004F. Available from:
21             NTIS, Springfield, VA; PB85-191559.
22       U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
23             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office;
24             report no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
25       U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
26             monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research
27             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available
28             from: NTIS, Springfield, VA; PB93-157717.
29       U.S. Environmental Protection Agency. (1994) Methodologies for estimating emission and travel activity effects of
3 0             TCMs.  Washington, DC: Office of Mobile Sources; report no. EPA-420-R-94-002.
31       U.S. Environmental Protection Agency. (1996) National air pollutant emission trends, 1900-1995. Research
32             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-96-007. Available
33             from: NTIS, Springfield, VA; PB97-155329.
34       U.S. Environmental Protection Agency. (1998a) National air quality and emissions trends report, 1996. Research
35             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-97-013. Available
36             from: NTIS, Springfield, VA; PB98-135155. Available online at: www.epa.gov/oar/aqtrnd96/.
37       U.S. Environmental Protection Agency. (1998b) Compilation of air pollutant emission factors. Volume II:
38             Mobile Sources (AP-42), pending 5th edition. Ann Arbor, MI: Office of Mobile Sources. Available online
39             at: www.epa.gov/omswww/models.htm.
40       Vadarevu, R. V.; Stopher, P. R. (1996) Household activities, life  cycle, and role allocation. Trans. Res. Rec.
41             1556:77-85.
42       Van Wijnen, J. H.; Verhoeff, A. P.; Jans, H. W. A.; van Bruggen, M. (1995) The exposure of cyclists, car drivers
43             and pedestrians to traffic-related air pollutants. Int. Arch. Occup. Environ. Health 67: 187-193.
44       Vreman, H.  J.; Mahoney, J. J.; Stevenson, D. K. (1993) Electrochemical measurement of carbon monoxide in
45             breath:  interference by hydrogen. Atmos. Environ. Part A 27: 2193-2198.
46       Wallace, L.  (1979) Use of personal monitor to measure commuter exposure to carbon monoxide in vehicle
47             passenger compartments. Presented at: 72nd annual meeting of the Air Pollution Control Association; June;
48             Cincinnati, OH. Pittsburgh, PA: Air Pollution Control Association; paper no. 79-59.2.
49       Wallace, L.  A. (1983) Carbon monoxide in air and breath of employees in an underground office. J. Air Pollut.
50             Control Assoc. 33: 678-682.
51       Webster, H. (1998) Would you like to work at home? Read. Dig.  152: 128-134.
52       Wheatley, A. D.; Sadhra, S.; Beach, J. R. (1997) Exposure to toxic gas and particle phase pollutants evolved
53             during deployment of airbags in vehicles. Indoor Environ.  6: 134-139.
54       Wilcosky, T. C.; Simonsen, N. R.  (1991) Solvent exposure and cardiovascular disease. Am. J. Ind. Med.
55             19:569-586.


         February 15, 1999                            4-57          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991a) Study of children's activity
  2             patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
  3             Available from: NTIS, Springfield, VA; PB94-106903.
  4       Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin G. (1991b) Activity
  5             patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
  6             ARB/R93/487. Available from: NTIS,  Springfield, VA; PB94-108719.
  7       Wilkinson, T. (1995) Snowed under. Natl. Parks 69: 32-37.
  8       Williams, R.; Walsh, D.; White, J.; Jackson, M.; Mumford, J. (1992) Effect on carbon monoxide levels in mobile
  9             homes using unvented kerosene heaters for residential heating. Indoor Environ. 1: 272-278.
10       Wilson,  A. L.; Colome, S. D.; Tian, Y. (1991) Air toxics microenvironment exposure and monitoring study.
11             El Monte, CA: South Coast Air Quality Management District.
12       Wilson,  A. L.; Colome, S. D.; Tian, Y. (1993a) California residential indoor air quality study. Volume 1.
13             Methodology and descriptive statistics. Chicago, IL:  Gas Research Institute; report no. GRI-93/0224.1.
14             Available from: NTIS, Springfield, VA; PB94-166048.
15       Wilson,  A. L.; Colome, S. D.; Tian, Y. (1993b) California residential indoor air quality study. Volume 1.
16             Methodology and descriptive statistics, appendix. Chicago, IL: Gas Research Institute; report no.
17             GRI-93/0224.2. Available from: NTIS, Springfield, VA; PB94-166055.
18       Yates, M. W. (1967) A preliminary study of carbon monoxide gas in the home. J. Environ. Health 29: 413-420.
19       Yocom,  J. E. (1982) Indoor-outdoor air quality relationships: a critical review. J.  Air Pollut. Control Assoc.
20             32:500-520.
21       Yocom,  J. E.; Clink, W. L.; Cote, W. A. (1971) Indoor/outdoor air quality relationships. J. Air Pollut. Control
22             Assoc. 21:251-259.
23       Yoon, S. S.; Macdonald, S. C.; Parrish, R. G.  (1998) Deaths from unintentional carbon monoxide poisoning and
24             potential for prevention with carbon monoxide detectors. JAMA J. Am. Med. Assoc. 279: 685-687.
25       Yu, L. E.; Hildemann, L. M.; Ott, W. R. (1996) A mathematical model for predicting trends in carbon monoxide
26             emissions and exposures on urban arterial highways. J. Air Waste  Manage. Assoc. 46: 430-440.
27       Zartarian, V. G.; Ott, W. R.; Duan, N. (1997)  A quantitative definition of exposure and related concepts.
28             J. Exposure Anal. Environ. Epidemiol. 7: 411-437.
29       Zayasu,  K.; Sekizawa, K.; Okinaga, S.; Yamaya, M.; Ohrui, T.; Sasaki, H. (1997) Increased carbon monoxide in
30             exhaled air of asthmatic  patients. Am. J. Respir. Crit. Care Med. 156: 1140-1143.
31       Ziskind, R. A.; Rogozen, M. B.; Carlin, T.; Drago, R. (1981) Carbon monoxide intrusion into sustained-use
32             vehicles. Environ. Int. 5: 109-123.
33       Ziskind, R. A.; Fite, K.; Mage,  D. T. (1982) Pilot field study: carbon monoxide exposure monitoring in the general
34             population. Environ. Int. 8: 283-293.
35
         February 15, 1999                             4-58          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, pharmacodynamics, and toxicology of carbon monoxide
 7     (CO) that ended in 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 pharmacodynamics 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
21
22     5.2   ABSORPTION, DISTRIBUTION, AND PULMONARY
23           ELIMINATION
24     5.2.1 Pulmonary Uptake
25          Although CO is not one of the respiratory gases, the similarity of physico-chemical
26     properties of CO and oxygen (O2) permits an extension of the findings of studies on the kinetics
27     of transport of O2 to those of CO. The rate of formation and elimination of COHb, its
28     concentration in blood, and its catabolism is controlled by numerous physical factors and
29     physiological mechanisms. The relative contribution of these mechanisms to the overall COHb
30     kinetics will depend on the environmental conditions, the physical activity of an individual, and

       February  15, 1999                         5-1         DRAFT-DO NOT QUOTE OR CITE

-------
 1      many other physiological processes, some of which are complex and still poorly understood
 2      (see Section 5.4 for details). All of the pulmonary uptake occurs at the respiratory bronchioles,
 3      alveolar ducts, and sacs.  The rate of CO uptake depends on the rate of COHb formation.  At the
 4      low concentration of CO in inhaled air, the rate of uptake and the rate of COHb formation could,
 5      for all practical purposes, be considered to be qualitatively the same.
 6
 7      5.2.1.1 Mass Transfer of Carbon Monoxide
 8           The mass transport of CO between the airway opening (mouth and nose) and the red blood
 9      cell (RBC) hemoglobin (Hb) is predominantly controlled by physical processes. The CO transfer
10      to the Hb-binding sites is accomplished in two sequential steps: (1) transfer of CO in a gas
11      phase, between the airway opening and the alveoli; and (2) transfer in a "liquid" phase, across the
12      air-blood interface, including the RBC.  In the gas phase, the key mechanisms of transport are
13      convective flow, by the mechanical action of the respiratory system, and diffusion in the acinar
14      zone of the lung (Engel et al., 1973).  Subsequent molecular diffusion of CO across the
15      alveolo-capillary membrane, plasma, and RBC  is the virtual mechanism of the liquid phase. The
16      principal transport pathways and body stores of CO are shown in Figure 5-1  (Coburn, 1967).
17
18      5.2.1.2 Effects of Dead Space and Ventilation/Perfusion Ratio
19           The effectiveness of alveolar gas exchange depends on effective gas mixing and matching
20      of ventilation and perfusion. During normal tidal breathing, the inhaled gas is not distributed
21      uniformly across the tracheobronchial tree. With increased inspiratory flow, as during exercise,
22      intrapulmonary gas  distribution becomes more uniform, but gas concentration inhomogeneity
23      still will persist. Considering that almost 90% of gas is contained within the acinar zone of the
24      lung, any increase in gas inhomogeneity in this  terminal region will have about the same negative
25      effect as an additional increase in the alveolar dead space or a decrease in the alveolo-capillary
26      diffusion capacity (Engel et al., 1973).
27           The inefficiency of gas mixing and a consequent decrease in the effectiveness of alveolar
28      gas exchange is aggravated by ventilation/perfusion (VA/Q) mismatch.  Because of the gravity-
29      dependence of ventilation and even more of perfusion, regional VA/Q ratios will range from
30      0.6 (at the base of the lung) to 3.0 (at the apex), the overall value being 0.85. Because the VA/Q

        February 15, 1999                          5-2        DRAFT-DO NOT QUOTE OR CITE

-------
                                     Cswbewi (fenastiele ;-n (fw Anibiaril Air
                                              Alveolar Air
                            pr'nMlwi't*!
                            "
         Figure 5-1.  Diagrammatic presentation of CO uptake and elimination pathways and
                     CO body stores.
         Source:  Adapted from Coburn (1967).
 1      ratio is the principal variable controlling gas exchange, any inequalities not only will impair
 2      transfer of gases to the blood but also will interfere with unloading of gases from the blood into
 3      the alveolar air. In humans, a change in posture to recumbent or horizontal, or exercise will
 4      increase the uniformity of VA/Q ratios and promote more efficient gas exchange, whereas
 5      increased resting lung volume, increased airway resistance, decreased lung compliance, and,
 6      generally, any lung abnormality will aggravate VA/Q ratio inequality.
 7           The simplest indicator of the VA/Q ratio inequalities is the volume of physiological dead
 8      space (VD), which comprises both the anatomical and alveolar dead space. The alveolar dead
 9      space results from reduced perfusion of alveoli, relative to their ventilation (Singleton et al.,
10      1972). An increase in tidal volume or respiratory frequency, or both, will increase moderately to
11      substantially the VD in healthy subjects and in individuals with  lung function impairment,
12      respectively (Lifshay et al., 1971).
       February 15, 1999
5-3
DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.2.1.3 Lung Diffusion of Carbon Monoxide
 2           Although the above mechanisms controlling the rate of formation of blood COHb are
 3      predominantly active processes, the second key mechanism, diffusion of gases across the alveolar
 4      air-Hb barrier, is an entirely passive process. In order to reach the Hb-binding sites, CO and
 5      other gas molecules have to pass across the alveolo-capillary membrane, diffuse through the
 6      plasma, pass across the RBC membrane, and finally enter the RBC stroma before reaction
 7      between CO and Hb can take place. The molecular transfer across the membrane and the blood
 8      phase is governed by general physico-chemical laws, particularly by Pick's first law of diffusion.
 9      The exchange and equilibration of gases between the two compartments (air and blood) is very
10      rapid. The dominant driving force is a partial pressure differential of CO across this membrane.
11      For example, inhalation of a bolus of air containing a high level of CO rapidly increases blood
12      COHb by immediate and tight binding of CO to Hb. The rapidity of CO binding to Hb keeps a
13      low partial pressure of CO within the RBC, thus maintaining a high pressure differential between
14      air and blood and consequent diffusion of CO into blood.  Subsequent inhalation of CO-free air
15      reverses the gradient (higher CO pressure on the blood side than alveolar air), and CO is released
16      into alveolar air.  The air-blood pressure gradient for CO is usually much higher than the
17      blood-air gradient; therefore, the CO uptake will be a proportionately faster process than CO
18      elimination.  The rate of CO release will be further affected by tissue metabolism and
19      endogenous production of CO.
20           Diurnal variations in CO diffusion related to variations in Hb have been reported in normal,
21      healthy subjects (Frey et al., 1987).  Others found the changes to be related also to physiological
22      factors such as oxyhemoglobin (O2Hb), COHb, partial pressure of alveolar CO2, ventilatory
23      pattern, O2 consumption, blood flow, functional residual capacity, etc. (Forster, 1987). In a
24      supine position at rest, CO diffusion has been shown to be significantly higher than that at rest in
25      a sitting position (McClean et al., 1981). Carbon monoxide diffusion increases with exercise,
26      and, at maximum work rates, diffusion will be maximal regardless of body position.  This
27      increase is attained by increases in both the diffusing capacity of the alveolar-capillary membrane
28      and the pulmonary capillary blood flow (Stokes et al., 1981). Diffusion seems to be relatively
29      independent of lung volume within the midrange of vital capacity. However, at extreme
30      volumes, the differences in diffusion rates  could be significant; at total lung capacity, diffusion is
31      higher, whereas, at residual volume, it is lower than the average (McClean et al., 1981).  Under

        February 15, 1999                          5-4         DRAFT-DO NOT QUOTE OR CITE

-------
 1      pathologic conditions, where one of several components of the air-blood interface might be
 2      affected severely, as in emphysema, fibrosis, or edema, both the uptake and elimination of CO
 3      will be affected (Barie et al., 1994).
 4
 5      5.2.2 Tissue Uptake
 6      5.2.2.1 The Lung
 7           Although the lung in its function as a transport system  for gases is exposed continuously to
 8      CO, very little CO actually diffuses into the lung tissue itself (as dissolved CO), except for the
 9      alveolar region.  The epithelium of the conductive zone (nasopharynx and large airways) presents
10      a significant barrier to diffusion of CO.  Therefore, diffusion and gas uptake by the tissue, even at
11      high CO concentration, will be slow; most of this small amount of CO will be dissolved in the
12      mucosa of the airways. Diffusion into the submucosal layers and interstitium will depend on the
13      concentration of CO and duration of exposure. Experimental exposures of the oronasal cavity in
14      monkeys to very high concentrations of CO for a very short period of time (5 s) increased the
15      blood COHb level to <3.5%. Comparative exposures of the whole lung, however, elevated
16      COHb to almost 60% (Schoenfisch et al.,  1980).  Thus diffusion of CO across the airway mucosa
17      will contribute little if at all to overall COHb concentration.  In the transitional zone (>20th
18      generation), where both conductive and diffusive transport take place, diffusion of CO into lung
19      interstitium will be much easier and, at times, more complete. In the respiratory zone (alveoli),
20      which is the most effective interface for CO transfer, diffusion into the lung interstitium will be
21      complete.
22
23      5.2.2.2 The Blood
24           The rate of CO binding with Hb is about 20% slower, and the rate of dissociation from Hb
25      is an order of magnitude slower than are these rates for O2. However, the CO chemical affinity
26      (represented by the Haldane coefficient, M) for Hb is about 218 (210 to 250)-times greater than
27      that of O2 (Roughton, 1970; Rodkey et al., 1969).  Under steady-state conditions (gas exchange
28      between blood and atmosphere remain constant), one part of CO and 218 parts of O2 would form
29      equal parts of O2Hb and COHb (50% of each), which would be achieved by breathing air
30      containing 21% oxygen and 650 ppm CO. Moreover, the ratio of COHb to O2Hb is proportional
31      to the ratio of their respective partial pressures, PCO and PO2.  The relationship between the
        February 15, 1999                          5-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1      affinity constant M and PO2 and PCO, first expressed by Haldane (1897-1898), has the following
 2      form:
                              COHb/O2Hb = M x (PCO/PO2).                       (5-1)
 3
 4      At equilibrium, when Hb is maximally saturated by O2 and CO at their respective gas tensions,
 5      the M value for all practical purposes is independent of pH, CO2, temperature, and
 6      2,3-diphosphoglycerate that bind to the deoxygenated form of Hb and so lower its O2 affinity
 7      over a wide range of PCO/PO2 ratios (Wyman et al., 1982; Gronlund and Garby, 1984).
 8           Under dynamic conditions, competitive binding of O2 and CO to Hb is complex; simply
 9      said, the greater the number of heme molecules bound to CO, the greater is the affinity of free
10      hemes for O2. Any decrease in the amount of available Hb for O2 transport (CO poisoning,
11      bleeding, anemia, blood diseases, etc.) will reduce the quantity of O2 carried by blood to the
12      tissue. However, CO not only occupies O2-binding sites, molecule for molecule, thus reducing
13      the amount of available O2, but also alters the characteristic relationship between O2Hb and PO2,
14      which in normal blood is S-shaped (v) (Figure 5-2).  With increasing concentration of COHb in
15      blood, the dissociation curve is shifted gradually to the left, and its  shape is transformed into a
16      near rectangular hyperbola. Because the shift occurs over a critical saturation range for release of
17      O2 to tissues, a reduction in O2Hb by CO poisoning will have more severe effects on the release
18      of O2 than the equivalent reduction in Hb caused by anemia.  Thus, in an acute anemia patient
19      (50% of Hb) at a tissue PO2 of 26 torr (v'), 5 vol % of O2 (50% desaturation) might be extracted
20      from blood, an amount sufficient to sustain tissue metabolism. In contrast, in a person poisoned
21      with CO (50% COHb), the tissue PO2 will have to drop to 16 torr (v2; severe hypoxia) to release
22      the same, 5 vol % O2. Any higher demand on O2 under these conditions (e.g.,  by exercise) might
23      result in brain oxygen depletion  and loss of consciousness of the CO-poisoned individual.
24           Because so many cardiopulmonary factors determine COHb formation, the association
25      between COHb concentration in blood and duration of exposure is  not linear but S-shaped. With
26      progression of exposure, the initial slower COHb formation gradually accelerates, but, as COHb
27      approaches equilibrium, the build-up slows down again. The S-shape form becomes more
28      pronounced with higher CO levels or with exercise (Benignus et al., 1994; Tikuisis et al., 1992).
        February 15, 1999                          5-6         DRAFT-DO NOT QUOTE OR CITE

-------
                 20
                                                                    100
                 15 -
o
o
5
E,
§
i
o
1
0)
Q_
CD
E
                 10
                                          50% Anemia
                                          (O2 Hb Capacity = 10ml_/100 mL)
                                20
40
 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.
      Source:  U.S. Environmental Protection Agency (1991).
1     5.2.2.3  Heart and Skeletal Muscle
2          Myoglobin (Mb), as a respiratory hemoprotein of muscular tissue, will undergo a reversible
3     reaction with CO in a manner similar to O2. Greater affinity of O2 for Mb than Hb (hyperbolic
4     versus S-shaped dissociation curve) is, in this instance, physiologically beneficial because a small
5     drop in tissue PO2 will release a large amount of O2 from O2Mb.  The main function of Mb is
6     thought to be a temporary store of O2 and to act as a diffusion facilitator between hemoglobin and
7     the tissues (Peters et al., 1994).
      February 15, 1999
                                    5-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Myoglobin has a CO affinity constant approximately eight-times lower than hemoglobin
 2      (M = 20 to 40 versus 218, respectively) (Haab and Durand-Arczynska, 1991; Coburn and
 3      Mayers, 1971). As with Hb, the combination velocity constant between CO and Mb is only
 4      slightly lower than that for O2, but the dissociation velocity constant is much lower than that for
 5      O2.  The combination of greater affinity (Mb is 90% saturated at PO2 of 20 mmHg) and lower
 6      dissociation velocity constant for CO favors retention of CO in the muscular tissue.  Thus, a
 7      considerable amount of CO potentially can be stored in the skeletal muscle. The binding of CO
 8      to Mb (COMb) in heart and skeletal muscle in vivo has been demonstrated at levels of COHb
 9      below 2% in heart and  1% in skeletal muscle (Coburn, 1973; Coburn and Mayers, 1971). At rest,
10      the COMb/COHb ratio (0.4 to 1.2) does not increase with an increase in COHb up to
11      50% saturation and appears to be independent of the duration of exposure (Sokal et al., 1984).
12      During  exercise, the relative rate of CO binding increases more for Mb than for Hb, and CO will
13      diffuse from blood to skeletal muscle (Werner and Lindahl, 1980); consequently, the
14      COMb/COHb will increase for both skeletal and cardiac muscles (Sokal et al., 1986). A similar
15      shift in  CO has been observed under hypoxic conditions because a fall in intracellular PO2 below
16      a critical level will increase the relative affinity of Mb to CO (Coburn and Mayers,  1971).
17      Consequent reduction in O2-carrying capacity of Mb may have a profound effect on the supply of
18      O2 to the tissue. Apart  from Hb and Mb, other hemoproteins will react with CO. However, the
19      exact role of such compounds on O2-CO kinetics still needs to be ascertained.  For more
20      discussion on this topic, see Section 5.6.1.
21
22      5.2.2.4  The Brain and Other Tissues
23           The concentration of CO in brain tissue has been found to be about 30- to 50-times lower
24      than that in blood. During the elimination of CO from brain, the above ratio of concentrations
25      was still maintained (Sokal et al., 1984). However, the energy requirement of brain tissue is very
26      high and varies greatly  with local metabolism.  Because oxygen demand also is coupled to local
27      functional activity, which at times may be very high, and because the brain's oxygen storage is
28      minimal, any degree of hypoxia will have a detrimental effect on brain function. The primary
29      effects of CO on other organs (e.g., liver, kidney) is via hypoxic mechanisms.
30
31

        February 15, 1999                         5-8         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. The elimination rate of CO from an equilibrium state will follow a
 5      monotonically decreasing, second-order (logarithmic or exponential) function (Pace et al., 1950).
 6      The rate, however, may not be constant when the steady-state conditions have not yet been
 7      reached. Particularly after very short and high CO exposures, it is possible that COHb decline
 8      could be biphasic, and it can be approximated best by a double-exponential function; the initial
 9      rate of decline or "distribution" might be considerably faster than the later "elimination" phase
10      (Wagner et al., 1975). The reported divergence of the COHb decline rate in blood and in exhaled
11      air suggests that the CO elimination rates from  extravascular pools are slower than those reported
12      for blood (Landaw, 1973). Although the absolute elimination rates are associated positively with
13      the initial concentration of COHb, the relative elimination rates appear to be independent of the
14      initial concentration of COHb (Wagner et al., 1975).
15           The same factors that govern CO uptake will affect CO elimination. This suggests that the
16      Coburn-Forster-Kane (CFK) model (see  Section 5.5.1) may be  suitable to predict CO elimination
17      as well.  Surprisingly, few studies tested this application.  When breathing air, the CFK model
18      predicted very well the COHb decline. However, at a higher partial pressure of O2 in humidified
19      inspired air (Pf)2) or under hyperbaric O2 conditions, the key CFK equation parameters,
20      particularly the diffusing capacity for CO (DLCO) value, must be adjusted for hyperoxic
21      conditions so that CFK will predict more accurately the elimination of CO (Tikuisis, 1996;
22      Tikuisis et al., 1992; Tyuma et al., 1981). The half-time of CO disappearance from blood under
23      normal recovery (air) showed a considerable between-individual variance. For COHb
24      concentrations of 2 to 10%, the half-time ranged from 3 to 5 h (Landaw, 1973); others reported
25      the range to be 2 to 6.5 h for slightly higher initial concentrations of COHb (Peterson and
26      Stewart, 1970). The CO elimination half-time in nonsmokers is considerably longer in men
27      (4.5 h) than in women (3.2 h).  During sleep, the elimination rate slowed in both  sexes, but,
28      in men, it became almost twice as slow (8.0 h) as during waking hours.  Although no ventilation
29      variables were obtained during the study, the day-to-night differences have been  attributed to
30      lower ventilation rates at sleep.  The authors speculate that the sex differences in elimination
31      half-time are related to the skeletal muscle mass and intrinsically to the amount of stored Mb
        February 15, 1999                          5-9         DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Deller et al., 1992). The half-time elimination rate appears to be independent of the CO
 2      exposure source (e.g., fire, CO intoxication).  Normobaric O2 administered to fire victims and
 3      CO-poisoned individuals resulted in about the same CO elimination half-time, 91 and 87 min,
 4      respectively (Levasseur et al., 1996).
 5          Increased inhaled concentrations of O2 accelerated elimination of CO; by breathing
 6      100% O2, the half-time was shortened by almost 75% (Peterson and Stewart, 1970). The average
 7      half-life of COHb in individuals with very low COHb level (1.16%) breathing hyperbaric O2 was
 8      26 min, compared with 71 min when breathing normobaric O2 (Jay and McKindly,  1997). The
 9      elevation of PO2 to 3 atm reduced the half-time to about 20 min, which is approximately a
10      14-fold decrease over that seen when breathing room air (Britten and Myers, 1985;  Landaw,
11      1973). Although the washout of CO can be somewhat accelerated by an admixture of 5% carbon
12      dioxide (CO2) in O2, hyperbaric O2 treatment is more effective in facilitating displacement of CO.
13      Therefore, hyperbaric oxygen is used as a treatment of choice in CO poisoning. A mathematical
14      model of COHb elimination that takes into account PjO2 has been developed but not yet validated
15      (Singh etal., 1991).
16
17
18      5.3   TISSUE PRODUCTION AND  METABOLISM OF CARBON
19           MONOXIDE
20          In the process of natural degradation of RBC hemoglobin to bile pigments, a carbon atom
21      (a-bridge C) is separated from the porphyrin nucleus and, subsequently, is catabolized by
22      microsomal  heme oxygenase (HOX) into CO.  The major site of heme breakdown and, therefore,
23      the major production organ of endogenous CO is the liver (Berk et al., 1976).  The spleen and the
24      erythropoietic system are other important catabolic generators of CO.  Because the amount of
25      porphyrin breakdown is stoichiometrically related to the amount of endogenously formed CO, the
26      blood level of COHb or the concentration of CO in the alveolar air have been used with mixed
27      success as quantitative indices of the rate of heme catabolism (Landaw et al.,  1970; Solanki  et al.,
28      1988). Diurnal variations in endogenous CO production are significant, reaching a maximum
29      around noon and a minimum around midnight (Levitt et al., 1994; Mercke et al., 1975a).
30      Week-to-week variations of CO production are greater than day-to-day or within-day variations
31      for both males and females.

        February 15, 1999                          5-10        DRAFT-DO NOT QUOTE OR  CITE

-------
 1           Any disturbance leading to accelerated destruction of RBCs and accelerated breakdown of
 2      other hemoproteins would lead to increased production of CO.  Hematomas, intravascular
 3      hemolysis of RBCs, blood transfusion, and ineffective erythropoiesis all will elevate COHb
 4      concentration in blood. In females, COHb levels fluctuate with the menstrual  cycle; the mean
 5      rate of CO production in the premenstrual, progesterone phase is almost doubled
 6      (Delivoria-Papadopoulos et al., 1974; Mercke and Lundh, 1976).  Neonates and pregnant women
 7      also showed a significant increase in endogenous CO production related to increased breakdown
 8      of RBCs. Degradation of RBCs under pathologic conditions such as anemia (hemolytic,
 9      sideroblastic, and sickle cell), thalassemia, Gilbert's syndrome with hemolysis, and other
10      hematological diseases also will accelerate CO production (Berk et al.,  1974; Solanki et al.,
11      1988). In patients with hemolytic anemia, the CO production rate was 2- to 8-times higher, and
12      blood COHb concentration was 2- to 3-times higher than in healthy individuals (Coburn et al.,
13      1966). Anemias also may develop under many pathophysiologic conditions characterized by
14      chronic inflammation, such as malignant tumors or chronic infections (Cavallin-Stahl et al.,
15      1976) (see also Section 5.4.3).
16           Not all endogenous CO comes from RBC  degradation. Other hemoproteins, such as Mb,
17      cytochromes, peroxidases, and catalase, contribute approximately 20 to 25% to the total amount
18      of CO (Berk et al., 1976).  Approximately 0.4 mL/h of CO is formed by Hb catabolism, and
19      about 0.1 mL/h originates from nonhemoglobin sources (Coburn et al.,  1964).
20           A large variety of drugs will affect endogenous CO production. Generally, any drug that
21      will increase bilirubin production, primarily from the catabolism of Hb, will promote endogenous
22      production of CO. Nicotinic acid (Lundh et al., 1975), allyl-containing compounds (acetamids
23      and barbiturates) (Mercke et al., 1975b), diphenylhydantoin (Coburn, 1970a), progesterone
24      (Delivoria-Papadopoulos et al., 1974), contraceptives (Mercke et al., 1975c) will all elevate
25      tissue bilirubin and, subsequently, CO production.
26           Another mechanism that will increase CO production is a stimulation of HOX and
27      subsequent degradation of cytochrome P-450-dependent, mixed-function oxidases. Several types
28      of compounds such as a carbon disulfide and sulfur-containing chemicals (parathion and
29      phenylthiourea) will act on different moieties of the P-450 system leading to an increase in
30      endogenous CO (Landaw et al., 1970).  Other sources of CO involving HOX activity include
31      auto-oxidation of phenols, photooxidation of organic compounds  and lipid peroxidation of cell

        February 15, 1999                         5-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      membrane lipids (Rodgers et al., 1994).  The P-450 system also is involved in oxidative
 2      dehalogenation of dihalomethanes, widely used solvents in homes and industry (Kim and Kim,
 3      1996).  Metabolic degradation of these solvents and other xenobiotics results in the formation of
 4      CO that can lead to very high (>10%) COHb levels (Manno et al., 1992; Pankow, 1996).
 5           Ascent to high elevations will increase the endogenous level of COHb in both humans and
 6      animals (McGrath, 1992; McGrath et al., 1993). The basal CO level has been shown to be
 7      positively dependent on altitude (McGrath, 1992).  Assuming the same endogenous production
 8      of CO at altitude as at sea level, the increase in COHb most likely is consequent to a decrease in
 9      PO2 (McGrath et al., 1993).  Whether other variables, such as an accelerated metabolism or a
10      greater pool of hemoglobin,  transient shifts in body stores, or a change in the elimination rate of
11      CO are contributing factors,  remains to be explored. Animal studies suggest that the elevated
12      basal COHb production is not a transient phenomenon but persists through a long-term
13      adaptation period (McGrath, 1992).
14           In recent years, new discoveries in molecular biology identified the CO molecule as being
15      involved in many physiological responses, such as smooth muscle relaxation, inhibition of
16      platelet aggregation, and as a neural messenger in the brain (for details, see Sections 5.6 and 5.7).
17      Most recently, several studies reported yet another function of CO, that of a possible marker of
18      inflammation in individuals  with upper respiratory tract infection (Yamaya et al., 1998) and in
19      asthmatics (Zayasu et al., 1997). In the latter study, the investigators found that exhaled
20      concentrations of CO in asthmatics taking corticosteroids were about the same as in healthy
21      individuals (1.7  and 1.5 ppm, respectively), whereas, in asthmatics who did not use
22      corticosteroids, the average CO concentration was 5.7 ppm.  The authors speculate that one of
23      the anti-inflammatory effects of corticosteroids is the down-regulation of HOX. Whether
24      asthmatics have an increased COHb level was not measured in this  study or reported in other
25      studies. Critical illness also  seems to be  associated with elevated production of CO (Meyer et al.,
26      1998).  When compared with controls, ill patients (not characterized) had higher COHb in both
27      arterial and central venous blood not attributable to an elevated inspired concentration of O2 used
28      to treat patients. Moreover, the higher COHb in arterial blood than  in central venous blood
29      measured in both ill and control individuals has lead the authors to speculate that a positive
30      arterio-venous COHb difference results from the up-regulation of the inducible isoform of heme
31      oxygenase (HO-1) in the lung and subsequent production of CO (see Section 5.6.3).

        February 15, 1999                          5-12        DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.4  CONDITIONS AFFECTING CARBON MONOXIDE UPTAKE
 2           AND ELIMINATION
 3      5.4.1  Environment and Activity
 4          During exercise, increased demand for O2 requires adjustment of the cardiopulmonary
 5      system, so that an increased demand for O2 is met with an adequate supply of O2.  Depending on
 6      the intensity of exercise, the physiologic changes may range from minimal, involving primarily
 7      the respiratory system, to substantial, involving extensively the respiratory, cardiovascular, and
 8      other organ systems, inducing local as well as systemic changes. Exercise will improve the VA/Q
 9      ratio in the lung, increase the  respiratory exchange ratio (RER), increase cardiac output, increase
10      DLCO, mobilize RBC reserves from the spleen and induce other compensatory changes.  Heavy
11      exercise will cause a decrease in plasma volume leading to hemoconcentration and a subsequent
12      decrease in blood volume.  Of the many mechanisms operating during exercise, the two most
13      important physiologic variables are (1) the alveolar ventilation (VA) and (2) cardiac output.
14      Although some physiologic changes during exercise may impair CO loading into blood (e.g.,
15      relative decrease in DLCO during severe exercise), the majority of the changes will facilitate CO
16      transport.  Thus, by increasing gas exchange efficiency, exercise also will promote CO uptake.
17      Consequently, the rate of CO uptake and of COHb formation will be proportional to the intensity
18      of exercise.  During a transition period from rest to exercise while exposed to CO (500 ppm), the
19      diffusing capacity and CO uptake were  reported to rise faster than O2 consumption for each
20      exercise intensity (Kinker et al., 1992).  Carbon monoxide hypoxemia (15% COHb), at a
21      constant exercise intensity above the lactic acid threshold, was found to accelerate muscle
22      deoxygenation compared with normoxic conditions (Maehara et al., 1997).
23          Apart from physiological factors, the concentration of CO, as well as the rate of change of
24      CO concentration in an individual's immediate  environment, can have a significant impact on
25      COHb. For example, at intersections with idling and accelerating cars, pedestrians will be
26      exposed for a short period of time to higher CO concentrations than those present at other places
27      on the  same street. Around home, an individual working with a chain saw, lawnmower, or other
28      gasoline-powered tools will be exposed transiently to higher, and occasionally to much higher
29      (e.g., breathing near the exhaust of a chain saw), concentrations of CO (up to 400 ppm) (Biinger
30      et al., 1997). In indoor environments, exposure to elevated CO from unventilated gas appliances

        February 15, 1999                        5-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1      or from environmental tobacco smoke may increase transiently the COHb level of a previously
 2      unexposed individual. Occupationally, there are many instances and conditions under which
 3      workers may be exposed briefly to moderate-to-high levels of CO from operating equipment or
 4      other sources. Despite the shortness of each exposure episode, such transients may result in a
 5      relatively high COHb concentration. As an example, exposure for 5 min or less of a resting
 6      individual to 7,600 ppm CO in inhaled air will result in almost 20% COHb (Benignus et al.,
 7      1994).  On repeated brief exposures to high CO, the COHb will increase further until the
 8      concentrations in inhaled CO and in blood reach equilibrium.  Once the distribution of CO within
 9      body stores is complete, the COHb  will  remain constant, unless the ambient CO concentration
10      changes (either up or down) again.  As is the CO uptake, so is the elimination of CO from blood
11      governed by the gas concentration gradient between blood and alveolar air.  However, the
12      elimination CO from blood is a much slower process (see Section 5.2.3) and, therefore, will take
13      many hours of breathing clean air before the baseline COHb value is reached.
14           Both hypoxic hypoxia and CO-induced hypoxia lowered the core temperature and
15      metabolic rate of rats. The effects were  more pronounced as ambient temperature decreased.
16      Both types of hypoxia elicited compensatory hyperventilation, but, only during hypoxic hypoxia,
17      was the minute ventilation progressively increased as ambient PO2 decreased (Gautier and
18      Bonora, 1994).
19           Recently, a unique source of CO exposure was identified. It has been found repeatedly that
20      the use of volatile anesthetics (fluranes) in closed-circuit anesthetic machines, when CO2
21      absorbent (soda lime) is dry, can result in a significant production of CO caused by a degradation
22      of the anesthetic and subsequent exposure of a patient to CO (up to 7.0% COHb) (Woehlck et al.,
23      1997a,b).
24
25      5.4.2 Altitude
26           Altitude may have a significant influence on the COHb kinetics (U.S. Environmental
27      Protection Agency, 1978). These changes are consequent to compensatory and adaptive
28      physiologic mechanisms. At sea level, at a barometric pressure (PB) of 760 torr, the PjO2 (body
29      temperature and pressure, saturated with water vapor at 37  °C [BTPS] conditions) is 149 torr.
30      At an altitude of 3,000 m (9,840 ft;  PB = 526 torr), the PA is only 100 torr, resulting in an acute
31      hypoxic hypoxia. Direct measurements  of blood gases on over 1,000 nonacclimatized
        February  15,  1999                         5-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1      individuals at this altitude found the partial pressure of O2 in alveolar air to be only 61 torr
 2      (Boothby et al., 1954).  The hypoxic drive will trigger a complement of physiological
 3      compensatory mechanisms (to maintain O2 transport and supply), the extent of which will depend
 4      on elevation, exercise intensity, and the length of a stay at the altitude. During the first several
 5      days, the pulmonary ventilation at a given O2 uptake (work level) will increase progressively
 6      until a new quasi-steady-state level  is achieved (Bender et al., 1987; Burki, 1984). The DLCO
 7      will not change substantially at elevations below 2,200 m but was reported to increase above that
 8      altitude, and the spirometric lung function will be reduced as well (Ge et al., 1997).  The
 9      maximal aerobic capacity and total  work performance will decrease, and the RER will increase
10      (Horvath et al., 1988).  Redistribution of blood from skin to organs and within organs from blood
11      into extravascular compartments, as well as an increase in cardiac output, will promote CO
12      loading (Luomanmaki and Coburn,  1969). Because of a decrease in plasma volume
13      (hemoconcentration), the Hb concentration will be higher than at sea level (Messmer, 1982). The
14      blood electrolytes and acid-base equilibrium will be readjusted, facilitating transport of O2.
15      Thus, for the same CO  concentration as at sea level, these compensatory changes will favor CO
16      uptake and COHb formation (Burki, 1984). By the same token, the adaptive changes will affect
17      not only CO uptake but CO elimination as well. Carboxyhemoglobin levels at altitude has been
18      shown to be increased in both laboratory animals and humans (McGrath, 1992; McGrath et al.,
19      1993). Breathing CO (9 ppm) at rest at altitude has produced higher COHb levels than those at
20      sea level (McGrath et al., 1993).  Surprisingly, exercise in a CO atmosphere (50 to 150  ppm) at
21      altitude appeared either to suppress  COHb formation or to shift the CO storage, or both. The
22      measured COHb levels were lower  than those found under similar conditions of exercise and
23      exposure at sea level (Horvath et al., 1988).
24           The short-term acclimatization (within a week or two) will stabilize the compensatory
25      changes. During a prolonged stay at high altitude (over a few months), most of the early adaptive
26      changes gradually will revert to the  sea level values,  and long-term adaptive changes, such as an
27      increase in tissue capillarity and myoglobin content in the skeletal muscle, begin to take place.
28      Smokers appear to tolerate short-term hypoxic hypoxia caused by high altitude (7,620 m
29      [25,000 ft]) much better than nonsmokers, who experience more severe subjective symptoms and
30      a greater decline in task performance (Yoneda and Watanabe, 1997). Perhaps smokers, because
31      of chronic hypoxemia (because of chronically elevated COHb), develop partial tolerance to

        February 15, 1999                         5-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1      hypoxic hypoxia. Although the mechanisms of COHb formation in hypoxic hypoxia and CO
 2      hypoxia are different, the resultant decrease in O2 saturation and activation of compensatory
 3      mechanisms (e.g., an increased cerebral blood flow) appear to be at least additive  (McGrath,
 4      1988).  Psychophysiological studies, in particular, seem to support the possibility of
 5      physiological equivalency of hypoxic effects, whether induced by altitude at equlibrium or
 6      ambient CO concentration. However, it must be remembered that, although some of the
 7      mechanisms of action of hypoxic hypoxia and CO hypoxia are the same, CO elicits other toxic
 8      effects not necessarily related to O2 transport mechanisms (Ludbrook et al., 1992; Zhu and
 9      Weiss,  1994).  Recently, Kleinman et al. (1998) demonstrated that the effects of CO and
10      simulated altitude were not synergistic but additive.
11
12      5.4.3 Physical Characteristics
13           Physical characteristics (e.g., sex, age, race, pregnancy) are not known to directly modify
14      the basic mechanisms of CO uptake and COHb formation and elimination.  However, the
15      baseline values of many cardiopulmonary variables that may influence COHb kinetics are known
16      to change with physical characteristics.
17           The CO uptake and elimination rates either at rest or exercise decrease with age. During
18      the growing years (2 to 16 years  of age), the COHb  elimination half-time increases rapidly with
19      age in both sexes and is relatively shorter for boys than for girls.  Beyond teenage years, the
20      half-time for CO elimination continues to grow longer but at a lower rate.  In contrast to the
21      adolescent period, the COHb half-life  during the adult years was found to be persistently shorter
22      (-6%) in females than that in males (Joumard et al., 1981). Furthermore, it has been well
23      established that, with age, the DLCO, one of the key determinants, decreases (Guenard and
24      Marthan, 1996).  The rate of DLCO decline is lower in middle-aged women than it is in men,
25      however, at older ages, the rates  evened and are about the same for both sexes (Neas and
26      Schwartz, 1996).  The decrease in DLCO, combined with an increase in VA/Q mismatch, which
27      increases with age, means that it will take longer to both load and eliminate CO from blood.
28           In pregnancy, increased requirement for iron may lead to iron deficiency and anemia
29      (for further details see Section 5.4.3).  Pregnant women who smoked showed a more pronounced
30      shift of the O2 dissociation curve to the left (-5% COHb) than one would expect from the same

        February 15, 1999                         5-16        DRAFT-DO NOT QUOTE OR CITE

-------
 1      COHb concentration in nonpregnant women.  Thus, increased O2 affinity, combined with
 2      decreased O2-carrying capacity of blood of CO-exposed women, may promote fetal hypoxia
 3      (Grote et al., 1994). Animal studies found that protein deficiency in pregnant mice had no
 4      modulating effect on maternal COHb but resulted in a greater concentration of placental COHb
 5      (Singh et al., 1993; Singh and Moore-Cheatum, 1993; Singh et al.,  1992).
 6           Young women were found to be more resistant to altitude hypoxia than were men, but the
 7      physiological factors for this difference remain unexplored (Horvath et al., 1988).
 8      Carboxyhemoglobin levels, although elevated at altitude, were found to be about the same for
 9      both males and females (McGrath et al., 1993).
10           Whether the dynamics of COHb formation and elimination or the absolute COHb levels for
11      the same exposure conditions are different in any way between races have not been studied.
12      Blacks have lower diffusion capacity than whites (Neas and Schwartz, 1996), which transiently
13      will slow CO loading and unloading. It also is well documented that the black population has a
14      higher incidence of sickle cell anemia,  which may be a risk factor for CO hypoxia (see
15      Section 5.4.3 below).
16
17      5.4.4 Health Status
18           An individual with any pathophysiologic condition that reduces the blood O2 content will
19      be at  a greater risk from CO exposure because additional reduction in the O2-carrying capacity of
20      blood resulting from COHb formation will increase hypoxemia.  Depending on the severity of
21      initial hypoxia, exposure to CO may lower the O2 content to the point where O2 delivery to the
22      tissues becomes insufficient.
23           One group of disorders that encompasses a range of etiologically varied diseases
24      characterized by a decreased O2-carrying capacity of blood are the anemias. In most general
25      terms, anemia is any condition that will lead to a reduction in the concentration of RBCs in
26      blood. Anemia is a result of either impaired formation of RBCs or increased loss or destruction
27      of RBCs. The former category includes disorders of altered O2 affinity, methemoglobinemias,
28      and diseases with functionally abnormal and unstable hemoglobins. By far, the most prevalent
29      disorder in this group is a single-point mutation of Hb (Hb  S), causing sickle cell diseases, the
30      most  typical of which is a sickle cell anemia.  The O2-carrying capacity of individuals  afflicted
31      with sickle cell anemia is reduced not only because of a smaller amount of Hb, but also the
        February 15, 1999                          5-17        DRAFT-DO NOT QUOTE OR CITE

-------
 1      O2 dissociation curve is shifted to the right, reducing the O2 affinity as well. Initial compensation
 2      involves primarily the cardiovascular system.  The cardiac output will increase as both heart rate
 3      and stroke volume increase.
 4           The opposite condition of anemia is polycythemia, an increased number of RBCs in blood.
 5      Although in polycythemia the total amount of hemoglobin generally is elevated, under certain
 6      conditions the arterial O2 saturation may be decreased, leading to a higher risk of additional
 7      hypoxia when exposed to CO.
 8           One of the characteristic symptoms of chronic obstructive pulmonary disease (COPD) is
 9      increased VD and VA/Q inequality (Marthan et al., 1985).  Subsequently, impaired gas mixing
10      because of poorly ventilated lung zones will result in decreased arterial O2 saturation and
11      hypoxemia.  These pathophysiologic conditions will slow both CO uptake and elimination.
12      Any COHb formation will further lower the O2 content of blood and increase hypoxemia.
13      Because COPD patients very often operate at the limit of their O2 transport capability, exposure
14      to CO may compromise severely tissue oxygenation.
15           Because O2 extraction by the myocardium is high, a greater O2 demand by the myocardium
16      of healthy individuals is met by an increased coronary blood flow. Patients with coronary artery
17      disease (CAD) have a limited ability to increase coronary blood flow in response to increased
18      O2 demand during physical activity. If this compensatory mechanism is further compromised by
19      decreased O2 saturation from CO inhalation, the physical activity of patients with CAD may be
20      restricted severely consequent to more rapid development of myocardial ischemia.
21           Individuals with congestive heart failure, right-to-left shunt in congenital heart disease, or
22      cerebrovascular disease also may be at a greater risk from CO exposure because of already
23      compromised O2 delivery.
24
25
26      5.5   MODELING CARBOXYHEMOGLOBIN FORMATION
27      5.5.1 The Coburn-Forster-Kane and Other Regression Models
28      5.5.1.1  Empirical Regression Models
29           The most direct approach to establishing a prediction equation for COHb is to regress
30      observed COHb values against the concentration and duration of exogenous CO exposure.

        February 15, 1999                         5-18        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Inclusion of other predictors such as initial COHb level and VA generally will improve the
 2      precision of the predictions. All regression models are purely empirical and have no
 3      physiological basis. Their applicability therefore is limited to more or less exact conditions that
 4      were used to collect the data on which they are based.
 5          Peterson and Stewart (1970) developed a regression equation for estimating percent COHb
 6      following a 15-min to 8-h exposure of resting nonsmokers to moderate levels of CO (25 to
 7      523 ppm):

                Log % COHb  = 0.858 Log CO + 0.630 Log t - 2.295 - 0.00094 t',         (5-2)
 8     where CO refers to the concentration of CO in inhaled ambient air in parts per million, t is the
 9     exposure duration in minutes, and t' is the postexposure time in minutes. Data from a
10     subsequent study were used to derive a new empirical formula for much higher concentrations of
11     CO (1,000 to 35,600 ppm) but much shorter exposure times (45 s to 10 min) (Stewart et al.,
12     1973). These equations still are used occasionally in field conditions to quickly estimate COHb
13     concentration.
14           Several mathematical  models have been developed to predict COHb as a function of
15     exposure time (Singh et al.,  1991) or altitude (Selvakumar et al., 1992).  The physiological
16     variables used by Peterson and  Stewart (Peterson and Stewart, 1970) were employed to test the
17     models and compare the results to the CFK predictions (Coburn et al., 1965).  The agreement
18     between predicted COHb values by these models and the  CFK model was very good; however,
19     the models have not been validated by experimental studies.
20           To predict changes in  COHb as a function of ambient CO concentration in an urban setting,
21     Ott and Mage (1978) developed a linear differential equation where only ambient CO
22     concentration varied with time. All other parameters were empirically derived constants. With
23     this simple model, they were able to show that the presence of CO spikes in data averaged over
24     hourly intervals may lead to underestimating the COHb concentration by as much as 21% of the
25     true value. Consequently, they recommended that monitored CO be averaged over 10 to 15 min
26     periods. Based on a similar approach, other empirical models have been developed but not
27     validated (Chung, 1988; Forbes et al., 1945). Comparison of predicted COHb values by these
       February 15,  1999                         5-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
models revealed a progressive divergence of the COHb curves as exposure progressed, with
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 et al., 1992).

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 O2Hb is constant:
                    d[COHb]t
                       dt
                       =  Vco-
[COHb]0Pc-02
[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; and PjCO is the CO partial
pressure in inhaled air in millimeters of mercury.  The model also assumes an instant
equilibration of gases in the lung,  COHb concentration between venous and arterial blood, and
COHb concentrations between blood and extravascular tissues. Because O2 and CO combine
with Hb from the same pool, higher COHb values do affect the amount of Hb available for
bonding with O2. Such interdependence can be modeled by substituting (1.38 Hb- [COHb]) for
[O2Hb], where Hb refers to the number of grams of Hb per milliliter of blood (Tikuisis et al.,
1987a).  The CFK differential equation (Equation 5-3) then becomes nonlinear:
        February 15, 1999
                                         5-20
DRAFT-DO NOT QUOTE OR CITE

-------
                        d[COHb]t   Vco     1  [        [COHb]0P502 ]
                            dt     -  Vb    Vhp(^U-    [02Hb]M   )
 1
 2      where P is (1/DL) + (PB - 47)/VA, and PB is the barometric pressure in millimeters of mercury.
 3      The nonlinear CFK model is more accurate physiologically but has no explicit solution.
 4      Therefore, interactive or numerical integration methods must be used to solve the equation
 5      (Muller and Barton, 1987; Johnson et al., 1992). One of the requirements of the method is that
 6      the volumes of all gases be adjusted to the same conditions (e.g., STPD) (Muller and Barton,
 7      1987; Tikuisis et al., 1987a,b).
 8          In general, the linear CFK equation is a better approximation to the nonlinear equation
 9      during the uptake of CO than during the elimination of CO. As long as the linear CFK equation
10      is used to predict COHb levels at or below 6% COHb, the solution to the nonlinear CFK model
11      will deviate no more than ±0.5% COHb (Smith, 1990). Over the years, it has been empirically
12      determined that minute ventilation and the lung diffusion  capacity for CO have the greatest
13      influence on the CO uptake and elimination.  The relative importance of other physiologic
14      variables will vary with exposure conditions and health status.  A comprehensive evaluation of
15      fractional sensitivities of physiologic variables for both the linear and nonlinear CFK equation
16      shows that each variable will exert its maximal influence at different times of exposure
17      (McCartney, 1990). The analysis found that only the fractional concentration of CO in inhaled
18      air, in parts per million (FjCO) and Vco will not affect the rate at which equilibrium is reached.
19      Figure 5-3 illustrates the temporal changes in fractional sensitivities of the principal determinants
20      of CO uptake for the linear form of the CFK equation. The fractional sensitivity  of unity means
21      that, for example, a 5% error in the selected variable induces a 5% error in predicted COHb
22      equilibrium value by the nonlinear model.
23
24      5.5.1.3 Confirmation Studies of Coburn-Forster-Kane Models
25          Since the publication of the original paper (Coburn et al., 1965), several investigators have
26      tested the fit of both the linear and nonlinear CFK model to experimental data using different CO
27      exposure profiles, a variety of experimental conditions and different approaches to evaluating the
28      parameters of the model.  In all of these studies, almost all of the physiologic coefficients either
        February 15, 1999                         5-21        DRAFT-DO NOT QUOTE OR CITE

-------
       (D
       CD
       CD     -0-50 H
              -0.75-
              -1.00
              -0.25-
                                              0            1
                                           Log Time (min)
                        1.7h
                         2
                                         -1.00
                         17h
                          3
       Figure 5-3. Plot of fractional sensitivities of selected variables versus time of exposure (see
                  text for details).
       Source: Modified from McCartney (1990).
1     were assumed or estimated based on each individual's physical characteristics; the COHb values
2     were both measured directly and calculated for each individual.
3          Stewart et al. (1970) and Peterson and Stewart (1972) tested the CFK linear differential
4     equation on 18 resting subjects exposed to 25 different CO exposure profiles for periods of 0.5 to
5     24 h and to CO concentrations ranging from 1 to 1,000 ppm. In a later study, they tested the
6     nonlinear CFK equation on 22 subjects at various levels of exercise while being exposed to up to
7     200 ppm CO for up to 5.5 h (Peterson and Stewart, 1975). From the obtained values, they
      February 15, 1999
5-22
DRAFT-DO NOT QUOTE OR CITE

-------
 1      concluded that either at rest or exercise the agreement between the predicted and measured
 2      COHb values was good (correlation coefficient [r] > 0.74).
 3           The first study to test both the linear and nonlinear CFK models for CO uptake and
 4      elimination in pedestrians and car passengers exposed to ambient CO levels in a city was
 5      conducted by Joumard et al. (1981).  The two cohorts  exposed for 2 h to street and traffic
 6      concentrations of CO, respectively, comprised 73 nonsmokers (18 to 60 years of age).  Blood
 7      COHb samples were taken only at the beginning and the end of each journey, where the COHb
 8      value reached 2.7%, on average. Both equations performed well in estimating accurately COHb
 9      levels, although the value for males was underestimated slightly.
10           The predictive strength of the CFK model under variable CO concentrations was tested by
11      Hauck and Neuberger (1984). A series of experiments was performed on four subjects exposed
12      to a total of 10 different CO exposure profiles at several exercise levels, so that each exposure
13      was a unique combination of CO concentration and exercise pattern.  The ventilation and COHb
14      values (measured and calculated) were obtained at 1-min intervals.  The agreement between
15      measured and predicted COHb under these varied conditions was very good; the mean difference
16      was only 7.4% of the nominal (maximal predicted) value.
17           A series of studies has tested the accuracy of the CFK equation under transient exposure
18      conditions that would violate several assumptions of the CFK model,  specifically the assumption
19      of a single, well-mixed vascular compartment.  These studies were designed to simulate everyday
20      conditions (e.g., environmental, occupational, military) that may involve frequent but brief
21      (75 s  to 5  min) exposures to high (667 to 7,500 ppm) CO concentrations at rest and exercise.
22      Moreover, the experiments were designed to test the accuracy of the CFK equation under
23      transient exposure conditions during the CO uptake and early elimination phase from arterial and
24      venous blood. Attempts were made to measure directly some of the key physiologic parameters
25      of the CFK equation for each subject (Tikuisis et al., 1987a,b; Benignus et al., 1994). The
26      studies have shown that during and immediately following exposure, the arterial COHb was
27      considerably higher (1.5 to 6.1%), and the venous COHb was considerably lower (0.8 to 6%)
28      than the predicted COHb. The observed individual COHb differences between arterial and
29      venous blood ranged from 2.3 to 12.1% COHb among individuals (Benignus et  al., 1994).  The
30      overprediction of venous COHb increased during exercise («10% of the true value). Provided
31      that the total CO dose (concentration x  time) is the same and within the time constant for the

        February  15, 1999                         5-23        DRAFT-DO NOT QUOTE OR CITE

-------
 1      CO uptake and elimination, the COHb value was found to be the same, regardless of the pattern
 2      of exposure.  Because the CFK model appeared to be most sensitive to VA, inconsistencies in the
 3      estimates or conversion of gas volumes (ATPS and BTPS to STPD) will affect the predicted
 4      values. The interindividual and intraindividual disparities between measured and predicted
 5      COHb values were attributed primarily to delays in mixing of arterial and venous blood and
 6      differences in cardiac output; but, other factors such as lung wash-in also contribute to this
 7      phenomenon. Modification of the CFK equation by adjusting for regional differences in blood
 8      flow produced a model that predicted with much greater accuracy both the arterial (<0.7% COHb
 9      difference) and venous (<1% COHb difference) COHb during transient uptake and elimination of
10      CO from blood (Smith et al., 1994).
11           Although the CO concentrations used in these studies are several orders of magnitude
12      higher than the usual CO concentrations found in ambient air, under certain conditions (see
13      Section 5.4.1) people can be exposed briefly (<10 min) to such (or even higher) levels of CO in
14      their immediate environment.  Because the physiologic mechanisms (but not the kinetics) of
15      COHb formation are independent of CO concentration, high COHb transients, particularly in
16      at-risk individuals, could be of clinical importance. Even briefly, higher arterial  COHb may lead
17      to functional impairment of the hypoxia-sensitive brain and heart (see Sections 5.2.2.3 and
18      5.2.2.4). In these situations, the predicted arterial COHb level will be seriously underestimated.
19
20      5.5.1.4 Application of Coburn-Forster-Kane Models
21           To obviate measurements of CFK equation parameters, many of which are complex
22      techniques, attempts were made to simplify the CFK equation, because it may be difficult or even
23      impossible to measure directly some of these parameters, particularly during physical activity.
24      In one study, by relating physiological parameters to the O2 uptake by the body, which was in
25      turn related to an activity level, a simplified linear form of the CFK model was developed
26      (Bernard and Duker, 1981). The model was used subsequently to draw simple nomograms of
27      predictive relationship between pairs of variables, but the accuracy of the nomograms was not
28      experimentally tested.
29           The need for more accurate COHb prediction under more complex physiologic or exposure
30      conditions requires either modification or expansion of the CFK model.  Benignus (1995)
31      combined a physiological model of respiratory gas exchange, MACPUF  (Ingram et al., 1987),
        February 15, 1999                         5-24       DRAFT-DO NOT QUOTE OR CITE

-------
 1     with the CFK model. The new model allows for continuous output and input of
 2     60 cardiopulmonary variables, including FjCO.  The usefulness of the model is particularly in its
 3     ability to continuously update COHb concentration in response to dynamically changing
 4     physiologic parameters.  The model also allows COHb prediction under hypothetical conditions
 5     that otherwise would be very difficult to duplicate in the laboratory.
 6           A fundamental modification of the CFK model was made by Hill et al. (1977) to study the
 7     effects of CO inspired by the mother on the level of fetal COHb. The Hill equation combines the
 8     CFK  equation (for maternal COHb), with a term denoting COHb transfer from a placenta into the
 9     fetus.  Comparative evaluation of predicted and measured fetal COHb concentrations under
10     time-varying and steady-state conditions in both women and animals showed acceptable
11     agreement only under steady-state conditions (Hill et al., 1977; Longo and Hill, 1977).
12           As mentioned in Section 5.5.1.4, Smith et al. (1994) expanded the CFK model to allow for
13     prediction of arterial and venous COHb during a transient CO uptake and early elimination
14     phase. The model incorporated regional differences in blood flow, particularly in the forearm,
15     because the forearm is used most frequently for blood sampling. This more elaborate model
16     performed extremely well in predicting blood COHb.  Although the model was validated on a
17     small number of subjects using the same experimental setting, the validation was not performed
18     under more demanding conditions of physical activity and varying CO concentrations.
19           To accurately predict COHb in individuals exposed to dihalomethanes, which are a source
20     of endogenous CO (see Section 5.3), the CFK model was extended to account for the CO
21     production caused by oxidation of a parent chemical (Andersen et al., 1991). The model
22     developed and validated on rats employed a variety of exposure scenarios to dichloromethane.
23     It subsequently was tested on six volunteers exposed to dichloromethane, where, after adjustment
24     of a few parameters, the COHb level was predicted remarkably well. After further validation,
25     this model has potential use in predicting accurately COHb caused by exogenous and endogenous
26     CO originating from different sources.
27           Reexpression of the solution of the CFK model from units of percent COHb to parts per
28     million of CO allows the examination of a variety of CO concentration profiles, while keeping a
29     simple preselected target COHb as a constant.  Application of the transformed model to urban
30     hourly averaged CO concentrations that just attained alternative 1-h and 8-h CO NAAQS showed
31     that, depending on the air quality pattern used, between 0.01 to 10% of the population may

       February 15, 1999                         5-25        DRAFT-DO NOT  QUOTE OR CITE

-------
 1     exceed a target 2.1% COHb level in blood without CO ever exceeding the standard.
 2     By including transients, the models predicted COHb more accurately, particularly when built into
 3     the 8-h running averages (Venkatram and Louch, 1979; Biller and Richmond, 1982, 1992).
 4     Actually, the ambient CO concentrations could be averaged over any time period less than or
 5     equal to the half-life of COHb (Saltzman and Fox, 1986).
 6
 7
 8     5.6   INTRACELLULAR EFFECTS OF CARBON MONOXIDE
 9     5.6.1  Inhibition of Hemoprotein Function
10          Traditional concepts for CO pathophysiology have been based on the high affinity of CO
11     for deoxyhemoglobin and consequent reduction of O2 delivery.  Carbon monoxide also can
12     inhibit a number of hemoproteins found in cells, such as myoglobin, cytochrome c oxidase,
13     cytochrome P-450, dopamine P hydroxylase, and tryptophan oxygenase (Coburn and Forman,
14     1987). Inhibition of these enzymes could have adverse effects on cell function.
15          Carbon monoxide acts as a competitive inhibitor, hence biological effects depend on the
16     partial pressures of both CO and O2. The cellular hemoprotein with the highest Warburg
17     partition coefficient for CO is myoglobin. Carbon monoxide will inhibit myoglobin-facilitated
18     oxygen diffusion, but physiological compromise is seen only with high concentrations of COMb.
19     Wittenberg and Wittenberg (1993) found that high-energy phosphate production was inhibited in
20     isolated cardiac myocytes, maintained at a physiologically relevant oxygen concentration, when
21     COMb exceeded 40%. The authors estimated that formation of sufficient COMb to impair
22     oxidative phosphorylation in vivo would require a COHb level of 20 to 40%.
23          Warburg partition coefficients for cytochrome P-450-like proteins vary between 0.1 and
24     approximately 12, and there have been recent discussions suggesting that CO-mediated inhibition
25     of these proteins could cause  smooth muscle relaxation in vivo (Coburn and Forman, 1987;
26     Wangetal., 1997a;Wang,  1998). The issue relates to inhibition of cytochrome P-450-dependent
27     synthesis of several potent vasoconstricting agents (Wang, 1998). Vasodilation has been shown
28     via this mechanism with high concentrations of CO (ca. 90,000 ppm) (Coceani et al., 1988).  It is
29     unclear, however, whether this could arise under physiological conditions  and CO concentrations
30     produced endogenously.  The competition between CO and O2 for cytochrome c oxidase was

       February 15, 1999                         5-26        DRAFT-DO NOT QUOTE OR CITE

-------
 1      well outlined in the previous review (U.S. Environmental Protection Agency, 1991), but some
 2      additional information has been published. Based on its Warburg partition coefficient of
 3      between 5 and  15, CO binding is favored only in situations where oxygen tension is extremely
 4      low (Coburn and Forman, 1987). Carbon monoxide binding to cytochrome c oxidase in vivo
 5      will occur when COHb is high (ca.  50%), a level that causes both systemic hypotension as well
 6      as impaired oxygen delivery (Brown and Piantadosi, 1992). Mitochondrial dysfunction, possibly
 7      linked to cytochrome inhibition, has been shown to inhibit energy production, and it also may be
 8      related to enhanced free radical production (Piantadosi et al., 1995;  1997a).  There has been no
 9      new information published since the last air quality criteria document that pertains to the effects
10      of CO on dopamine P hydroxylase or tryptophan oxygenase.
11
12      5.6.2 Free Radical Production
13           Laboratory animal studies indicate that nitrogen- and oxygen-based free radicals are
14      generated in vivo during CO exposures. Exposure to CO at concentrations of 20 ppm or more
15      for 1  h will cause platelets to become a source of the nitric oxide free radical ('NO) in the
16      systemic circulation of rats (Thorn et al., 1994, Thorn and Ischiropoulos, 1997). Studies with
17      cultured bovine pulmonary endothelial cells have demonstrated that exposures to CO at
18      concentrations  as low as 20 ppm cause cells to release 'NO, and the exposure will cause death by
19      a 'NO process that is manifested 18 to 24 h after the CO exposure (Thorn et al., 1997; Thorn and
20      Ischiropoulos, 1997).  The mechanism is based on elevations in steady-state'NO concentration
21      and production of peroxynitrite (Thorn et al.,  1994; Thorn et al., 1997). Peroxynitrite is a
22      relatively long-lived, strong oxidant that is produced by the near diffusion-limited reaction
23      between 'NO and superoxide radical (Huie and Padjama, 1993).
24           The mechanism by which CO concentrations of 20 ppm or more cause an elevation of
25      steady-state 'NO concentrations appears to be based on altered intracellular "routing" of'NO in
26      endothelial cells and platelets.  It is well established that the association and dissociation rate
27      constants of'NO with hemoproteins exceed the rate constants for O2 or CO (Gibson et al., 1986).
28      However, Moore and Gibson (1976) found that when CO was incubated with nitrosyl ('NO)-
29      myoglobin or 'NO-hemoglobin, CO slowly displaced the 'NO. Carbon monoxide replacement
30      occurred even when there was excess 'NO-heme protein, and replacement rates were enhanced by
31      increasing the CO concentration or by carrying out the reaction in the presence of agents such as
        February 15, 1999                         5-27         DRAFT-DO NOT QUOTE OR CITE

-------
 1      thiols, which will react with the liberated 'NO.  These conditions, including the presence of
 2      thiols, exist in cells exposed to environmentally relevant concentrations of CO. Exposures to up
 3      to 1,000 ppm CO do not alter the rate of production of'NO by platelets and endothelial cells,
 4      yet liberation of'NO was enhanced by CO (Thorn and Ischiropoulos, 1997; Thorn et al., 1994;
 5      Thorn etal., 1997).
 6           Carbon monoxide will increase the concentration of'NO available to react with in vivo
 7      targets in both lung and brain, based on electron paramagnetic resonance studies of experimental
 8      animals  (Ischiropoulos et al., 1996; Thorn et al., 1999a).  The concentrations of nitric oxide
 9      synthase isoforms in lung were not altered because of CO, and the mechanism for elevation in
10      'NO was thought to be the same as that found in isolated cells (Thorn et al., 1994; 1997).
11      Exposure to 50 to 100 ppm CO also will increase hydrogen peroxide (H2O2) production in lungs
12      of rats (Thorn et al., 1999a). The phenomenon depended on 'NO production, as it was inhibited
13      in rats pretreated with Nwnitro-L-arginine methyl ester (L-NAME), a nitric oxide  synthase
14      inhibitor. Production of'NO-derived oxidants  also is increased in CO-exposed rats,  based on
15      measurements of nitrotyrosine, a maj or product of the reaction of peroxynitrite with proteins
16      (Ischiropoulos et al., 1996; Thorn et al., 1998, 1999a,b).
17           The mechanism for enhanced H2O2 production in lungs of CO-exposed rats is not clear.
18      It is possible that 'NO or peroxynitrite may perturb mitochondrial function. Peroxynitrite inhibits
19      electron transport at complexes I through HI, and 'NO targets cytochrome oxidase (Cassina and
20      Radi, 1996; Lizasoain et al., 1996;  Poderoso et al., 1996).  It is important to note, however, that
21      alterations in mitochondrial function and an increase of cellular H2O2 were not found in studies
22      where cultured endothelial cells were exposed to similar CO concentrations (Thorn et al., 1997).
23      An alternative possible mechanism to mitochondrial dysfunction is that exposure to CO may
24      inhibit antioxidant defenses. Mechanisms linked to elevations in 'NO could be responsible for
25      inhibiting one or more enzymes.  Nitric oxide-derived oxidants can inhibit manganese
26      superoxide dismutase and glyceraldehyde-3-phosphate dehydrogenase and deplete cellular stores
27      of reduced glutathione (Ischiropoulos et al., 1992; Luperchio et al., 1996).
28           Exposure to high CO concentrations (2,500 to 10,000 ppm) cause mitochondria in brain
29      cells to generate hydroxyl-like radicals (Piantadosi et al., 1995;  1997a). An additional source of
30      partially reduced O2 species found  in animals exposed to CO is  xanthine oxidase. Conversion of
31      xanthine dehydrogenase, the enzyme normally  involved with uric acid metabolism, to xanthine

        February 15, 1999                          5-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1      oxidase, the radical-producing form of the enzyme, occurred in the brains of rats exposed to
 2      approximately 3,000 ppm CO (Thorn, 1992).  Lower CO concentrations did not trigger this
 3      change. Therefore, xanthine oxidase is unlikely to be a free radical source following exposures
 4      to CO at concentrations found in ambient air.  Moreover, enzyme conversion was not a primary
 5      effect of CO. Rather, it only occurred following sequestration and activation of circulating
 6      leukocytes (Thorn, 1993).
 7
 8      5.6.3 Stimulation of Guanylate Cyclase
 9           In recent years, CO has been demonstrated to play a physiological role in vasomotor control
10      and neuronal signal transduction (Morita et al., 1995; Ingi et al.,  1996). Carbon monoxide is
11      produced  endogenously by oxidation of organic molecules, but the predominant source is from
12      the degradation of heme (Rogers et al., 1994).  The rate-limiting enzyme for heme metabolism is
13      heme oxygenase (HO), which converts heme to biliverdin, free iron, and CO.  Three isoforms of
14      HO have been characterized.  The HO-1 is an inducible enzyme found in vascular endothelial
15      cells, smooth muscle cells, bronchoalveolar epithelium, and pulmonary macrophages. The HO-1
16      is induced by its substrate, heme, as well as 'NO, H2O2, several cytokines, and lipopolysaccharide
17      (Arias-Diaz et al., 1995; Durante et al., 1997;  Morita et al., 1995; Motterlini et al., 1996).  The
18      HO-2 is a constitutive enzyme found in certain neurons within the central nervous system,
19      testicular  cells, and vascular smooth muscle cells (Marks, 1994). Little is known about HO-3,
20      which was recently identified in homogenates from a number of organs (McCoubrey et al.,
21      1997).
22           A main physiological role for CO is thought to be regulation of soluble guanylate cyclase
23      activity. Both CO and 'NO can activate guanylate cyclase, although activation by CO is
24      approximately 30-fold lower (Stone and Marietta, 1994). In neuronal cells possessing both heme
25      oxygenase and nitric oxide synthase, regulation of cyclic guanosine monophosphate (cGMP)
26      synthesis is mediated in a reciprocal fashion by producing either CO or 'NO (Ingi et al., 1996;
27      Maines et al., 1993).  A compensatory interrelationship between nitric oxide synthase and heme
28      oxygenase also has been found in endothelial  cells and activated macrophages, although its
29      functional significance is unknown (Kurata et al.,  1996; Seki et al., 1997).  In macrophages,
30      cGMP synthesis promotes chemotaxis, and cGMP-mediated synthesis and secretion of tumor
31      necrosis factor a has been linked to both CO and 'NO (Arias-Diaz et al., 1995; Belenky et al.,
        February 15, 1999                         5-29        DRAFT-DO NOT QUOTE OR CITE

-------
 1      1993). Carbon monoxide causes smooth muscle relaxation by stimulating soluble guanylate
 2      cyclase (Utz and Ullrich, 1991; Wang et al., 1997b).  Smooth muscle relaxation also may occur
 3      because of activation of calcium dependent potassium channels, although this effect may be
 4      linked to guanylate cyclase activity (Trischmann et al., 1991; Wang et al., 1997a). Carbon
 5      monoxide-mediated smooth muscle relaxation is involved with control of microvascular hepatic
 6      portal blood flow (Goda et al., 1998; Pannen and Bauer, 1998) and suppressing contractions in
 7      the gravid uterus (Acevedo and Ahmed, 1998). It also may play a role in gastrointestinal motility
 8      (Farrugia et al., 1998).
 9
10
11      5.7  MECHANISMS OF CARBON MONOXIDE TOXICITY
12      5.7.1  Alterations in Blood Flow
13          Carbon monoxide from environmental pollution may exert similar effects in vivo to those
14      of endogenously produced CO, because the tissue concentrations resulting from inhalation of CO
15      are comparable or greater than concentrations produced by cells possessing heme oxygenase.
16      Environmental exposures to CO result in interstitial CO concentrations in the nanomolar range.
17      Coburn demonstrated this relation to be true for COHb levels up to 50%. For example, at a
18      COHb of 3.8%, the extravascular fluid CO concentration is approximately 11 nmol (Coburn,
19      1970b; Gothert et al., 1970).  Liver parenchyma has been estimated to generate approximately
20      0.45 nmol CO/gram liver/min (Goda et al., 1998).  Carbon monoxide synthesis by smooth
21      muscle cells is approximately 8 pmol/mg protein/min for human aorta and 23 to 37 pmol/mg
22      protein/min for rat aorta (Cook et al.,  1995; Grundemar et al., 1995).  Carbon monoxide
23      production by unstimulated pulmonary macrophages is 3.6 pmol/mg protein/min, and, after
24      stimulation with lipopolysaccharide, it increases to about 5.1 pmol/mg protein/min (Arias-Diaz
25      et al., 1995). The rate of synthesis of CO varies widely for nerve cells.  Cerebellar granule  cells
26      generate  approximately 3  fmol/mg protein/min, olfactory nerve cells produce 4.7 pmol/mg
27      protein/min, and rat cerebellar homogenates can generate as much as 56.6 pmol/mg protein/min
28      (Ingi and Ronnett, 1995; Ingi et al., 1996; Maines, 1988; Nathanson et al., 1995).
29          Vasodilation is a well-established effect caused by exposure to environmental CO. At high
30      CO concentrations, the mechanism is related to impairment of O2 delivery (Kanten et al., 1983;

        February 15, 1999                          5-30        DRAFT-DO NOT QUOTE OR CITE

-------
 1     MacMillan, 1975). At low CO concentrations, the increases in cerebral blood flow exceed those
 2     predicted to be required to maintain adequate O2 supply (Koehler et al., 1982).  In a setting where
 3     cellular oxidative metabolism was not impaired by CO, elevations in cerebral blood flow
 4     appeared to be mediated by 'NO (Meilin et al., 1996).  Whether the mechanism was the same as
 5     that outlined in the section above, which causes oxidative stress, remains to be determined.
 6           It is unclear whether disturbances in vascular tone by environmental CO is a generalized,
 7     systemic response, and the impact of variables such as the duration of exposure have not been
 8     adequately investigated.  Although cerebral vasodilation mediated by 'NO was reported with
 9     exposures to 1,000 ppm CO,  1,000 ppm CO did not alter pulmonary vasoconstriction in an
10     isolated-perfused rat lung model (Cantrell and Tucker, 1996).  Exposure to 150,000 ppm CO
11     caused no changes in pulmonary artery pressure in isolated blood-perfused lungs, although CO
12     did inhibit hypoxic pulmonary vasoconstriction (Tamayo et al., 1997). Humans exposed to CO
13     for sufficient time to achieve COHb levels of approximately 8% were not found to have
14     alterations in forearm blood flow, blood pressure, or heart rate (Hausberg and  Somers, 1997).
15           Animals exposed to high CO concentrations (e.g., 3,000 to 10,000 ppm) have diminished
16     organ blood flow, which contributes to CO-mediated tissue injury (Brown and Piantadosi, 1992;
17     Ginsberg and Meyers, 1974; Okeda et al., 1981; Song et al., 1983; Thorn, 1990). The mechanism
18     is based on CO-mediated hypoxic stress and cardiac dysfunction; therefore, these effects do not
19     arise  at CO concentrations relevant to  ambient air quality.
20
21     5.7.2 Mitochondrial Dysfunction and Altered Production of High-Energy
22            Intermediates
23           When exposed to 10,000 ppm CO, rats exhibit impaired high-energy  phosphate synthesis
24     and production of hydroxyl free radicals because of mitochondrial dysfunction (Brown and
25     Piantadosi, 1992; Piantadosi et al., 1995). Exposure to 2,500 ppm CO also will cause hydroxyl
26     radicals to be produced, apparently by mitochondria, because of a process that could not be
27     related to hypoxic stress (Piantadosi et al.,  1997a). Evidence for mitochondrial dysfunction has
28     not been observed in vivo at lower CO concentrations. However, under conditions of high
29     metabolic demand, exposure to even 1,000 ppm CO in the absence of an overt hypoxic stress will
30     result in impaired energy production in brain (Meilin et al., 1996).
       February 15, 1999                         5-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Carbon monoxide binding to mitochondrial cytochromes of respiring cells in vitro has been
 2      documented only when either the CO concentration was extraordinarily high, or O2 tension was
 3      extremely low, such that the CO/O2 ratio exceeded 12:1 (Coburn and Forman, 1987). Following
 4      CO exposure and removal to fresh air, CO diffuses out from cells, and mitochondrial function is
 5      restored. This process is enhanced by inspiration of hyperbaric oxygen (Brown and Piantadosi,
 6      1992).  Studies in mice indicate that high CO concentrations inhibit synthesis of high-energy
 7      phosphates during exposure to  5,000 ppm CO for  15 min and these changes do not persist
 8      following removal to fresh air (Matsuoka et al., 1993). In summary, mitochondrial dysfunction
 9      and impaired high-energy phosphate synthesis have been shown by several independent
10      laboratories to occur during exposures to high CO concentrations. Current information suggests
11      that this alteration does not  occur at CO concentrations relevant to ambient air quality, and that
12      changes in energy production are not persistent for long periods of time following CO exposure.
13
14      5.7.3 Vascular Insults Associated with Exposure to Carbon Monoxide
15           There are two primary variables that impact  on CO toxicity. One is the concentration of
16      CO, the other is the duration of exposure. Traditionally, these two variables have been  viewed as
17      merely alternative ways of elevating COHb concentration in the body.  The concentration of CO
18      breathed dictates the duration of exposure required to achieve a particular blood level of COHb
19      or tissue level of CO. This view is predicated on the notion that CO pathophysiology is
20      determined by its binding to one or another hemoprotein and to inhibition of oxygen delivery or
21      oxidative metabolism.
22           There is a substantial body of literature to suggest that, at least with regard to vascular
23      effects, the duration of exposure has a greater impact on the magnitude of CO pathophysiology
24      than what is predicted based on the concentration  of CO that is inspired.  For example, the lungs
25      are the first site for potential action of environmental CO. Results from investigations have been
26      conflicting regarding the risk for pulmonary injury from CO.  Because of the lack of consensus
27      and also the absence of a recognized biochemical mechanism, low concentrations of CO have
28      been viewed as posing little risk to lung physiology (U.S. Environmental Protection Agency,
29      1992). When animals have been exposed to high CO concentrations sufficient to raise COHb
30      levels to 56 to 90%, exposures  have lasted for only minutes because of the hypoxic stress.
31      In these studies, evidence of increased capillary permeability was inconsistent (Fein et al., 1980;
        February 15,  1999                         5-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Niden and Schulz, 1965; Penney et al., 1988), and no other alterations in lung physiology were
 2      detected (Fisher et al., 1969; Robinson et al., 1985; Shimazu et al., 1990;  Sugi et al., 1990).
 3      In contrast, when human beings or experimental animals have been exposed to relatively low CO
 4      concentrations for many hours at a time, capillary leakage of macromolecules from the lungs and
 5      systemic vasculature have been documented, but the presence of an hypoxic stress was
 6      questioned (Kjeldsen et al., 1972; Maurer, 1941; Parving et al., 1972; Siggaard-Andersen et al.,
 7      1968).
 8           In light of the physiological role for CO in vasomotor control, protracted exposures may be
 9      prone to disturb vascular homeostasis, giving rise to pathophysiological responses.  Intermittent
10      exposures to CO for  1 day increased glomerular filtration rate by 50% (Pauli et al., 1968).
11      Monkeys exposed to 250 ppm CO for 2 weeks exhibited coronary artery damage consisting of
12      subendothelial edema, fatty streaking, and lipid-loaded cells (Thomsen, 1974). This study and
13      others (Armitage et al., 1976; Davies et al., 1976; Turner et al., 1979; Webster et al., 1968) have
14      suggested a link between atherosclerosis and chronic CO exposure. However, other studies have
15      failed to find evidence for an association (Hugod et al., 1978; Penn et al.,  1992).
16           Carbon monoxide may cause vascular insults, including leakage of albumin and leukocyte
17      sequestration, because of a process mediated by 'NO-derived oxidants (Ischiropoulos et al., 1996;
18      Thorn, 1993; Thorn et al., 1998, 1999a,b). Brain oxidative stress associated with this mechanism
19      has been shown with exposures of 1,000 to 3,000 ppm CO (Ischiropoulos et al., 1996; Thorn,
20      1993). However, it is unclear whether the flux of'NO, resulting from  exposures to lower CO
21      concentrations contribute to oxidative or nitrosative stress in vivo.  Important differences in the
22      patterns of leakage from pulmonary and systemic vascular beds suggest that they may be caused
23      by different mechanisms. For example, systemic vascular leakage  was present for several hours
24      after CO exposure, and the leakage resolved within 18 h, whereas pulmonary vascular leakage
25      was not measurable until 18 h after CO exposure, and it resolved by 48 h. Both pulmonary and
26      systemic vascular leakage occurred after hour-long exposures to CO, but not when exposures
27      lasted for only 30 min, and vascular changes were not different whether rats were exposed to just
28      50 ppm or as much as 1,000 ppm CO.  These are recent observations and  further investigations
29      are required before their relevance to environmental CO contamination can  be assessed
30      adequately. Moreover, it should be emphasized that the vascular leakage  observed in lungs and
31      systemic microvasculature following exposures to CO at concentrations as low as 50 ppm may

        February 15, 1999                         5-33        DRAFT-DO NOT QUOTE OR CITE

-------
 1      have no pathophysiological impact if regional lymphatics can sustain a higher flow so that edema
 2      does not occur (Thorn et al., 1998, 1999a,b).
 3
 4
 5      5.8   OTHER EFFECTS OF CARBON MONOXIDE
 6           Among the most concerning pathophysiological effects of CO is its propensity for causing
 7      brain damage. There has been considerable effort focussed on potential mechanisms for this
 8      process. With regard to ambient air standards, however, it is important to note that recent studies
 9      were done with high CO concentrations. Carbon monoxide poisoning is not a "pure"
10      pathological process, as injuries may be precipitated by a combination of cardiovascular effects
11      linked to hypoperfusion or frank ischemia, COHb-mediated hypoxic stress, and intracellular
12      effects, including free radical production and oxidative stress. For example, CO poisoning
13      causes elevations of glutamate and dopamine in experimental models and human fatalities
14      (Arranz et al., 1997; Ishimaru et al., 1991, 1992; Nabeshima et al., 1990,  1991; Newby et al.,
15      1978; Piantadosi et al., 1997b).  These elevations occur because of the CO-associated
16      cardiovascular compromise and, possibly, other direct CO-mediated effects. Based on the effects
17      of agents that block the N-methyl-D-aspartate (NMD A) receptor, elevations of glutamate in
18      experimental CO poisoning have been linked to a delayed type (but not an acute type) of
19      amnesia, to loss of CA1 neurons in the hippocampus of mice, and to loss of glutamate-dependent
20      cells in the inner ear of rats (Ishimaru et al., 1991, 1992; Liu and Fechter, 1995; Nabeshima et al.,
21      1990, 1991). Antioxidants can protect against CO-mediated cytotoxicity of glutamate-dependent
22      nerve cells (Fechter et al., 1997). Mechanisms of glutamate neurotoxicity include excessive
23      calcium influx, free radical-mediated injury that may include calcium-calmodulin-dependent
24      activation of cytosolic NO synthase,  and lipid peroxidation. Moderate stimulation by excitatory
25      amino acids may cause mitochondrial dysfunction with impaired synthesis of adenosine
26      triphosphate and production of reactive O2 species (Beal, 1992). Cell death can be through
27      necrosis or programmed cell death, depending on the intensity of the stimulus (Gwag et al.,
28      1995).  There also may be a synergistic injury with other forms of oxidative stress because
29      reactive O2 species can intensify excitotoxicity (Bridges et al., 1991; Pellegrini-Giampietro et al.,
30      1990).  Glutamate also can injure cells in the central nervous system that do not have NMDA

        February 15, 1999                         5-34        DRAFT-DO NOT QUOTE OR CITE

-------
 1      receptors by competing for cysteine uptake, which inhibits synthesis of glutathione (Lipton et al.,
 2      1997; Murphy et al., 1989; Oka et al., 1993).
 3
 4
 5      5.9   SUMMARY
 6           The most prominent pathophysiological effect of CO is hypoxemia caused by avid binding
 7      of CO to Hb.  Formation of COHb reduces O2-carrying capacity of blood and impairs release of
 8      O2 from O2Hb to tissues. In addition to tissue hypoxia, ultimate diffusion of CO to cells may
 9      affect adversely their function.  The brain and heart tissues are particularly sensitive to
10      CO-induced hypoxia and cytotoxicity. The rate of COHb formation and elimination depends on
11      many physical and physiological factors.  The same factors that govern CO uptake determine CO
12      elimination as well.  The flow of CO between blood and either alveolar air or the tissues, and
13      vice versa, is governed by the CO concentration gradient between these compartments.  Because
14      of a small blood-to-air CO gradient, and tight binding of CO to Hb, the elimination half-time is
15      quite long, varying from 2.0 to 6.5  h. Apart from the CO concentration in ambient air, the
16      principal determinants of CO uptake are minute ventilation and lung diffusion capacity.  Thus,
17      any physiological conditions that affect these variables (e.g., exercise, age) also will affect the
18      kinetics of COHb. Both the physical and  physiological variables have been  integrated into many
19      empirical and mathematical models of COHb formation and elimination under static and
20      dynamic conditions of ambient CO concentration and physiologic function.  The nonlinear CFK
21      equation is the most widely used predictive model of COHb  formation, and  it still is considered
22      the best all-around model for COHb prediction. Altitude may have a significant influence on
23      COHb kinetics.  The effects of hypoxic hypoxia (altitude) and CO-induced hypoxia appear to be
24      additive. Adaptation to altitude will moderate COHb  formation.  In addition to exogenous
25      sources of CO, the gas also is produced endogenously through catabolism of Hb, metabolic
26      processes of drugs,  and degradation of inhaled solvents and other xenobiotics. This last source
27      may lead to very high (up to 50%)  COHb  concentrations. Many disorders, particularly anemias
28      of any etiology, will predispose affected individuals to CO hypoxia. Furthermore, patients with
29      a variety of cardiopulmonary (e.g., COPD, CAD) and chronic inflammatory diseases may be at
30      increased risk because of elevated COHb.

        February 15, 1999                         5-35       DRAFT-DO NOT QUOTE OR CITE

-------
 1          Apart from impaired O2 delivery to the tissues because of COHb formation, recent studies
 2     of CO pathophysiology suggest cytotoxic effects independent of O2. New investigations have
 3     expanded the understanding of CO in two areas. First, there is a growing recognition of the role
 4     that CO may play in normal neurophysiology and in microvascular vasomotor control.  The
 5     impact of CO from ambient air on these processes has not been investigated adequately. Hence,
 6     there is insufficient information available to influence decisions on ambient air quality standards.
 7     The second area of investigation of CO is related to its propensity for causing free-radical -
 8     mediated changes in tissues.  Mechanisms for these changes have been linked to both
 9     mitochondria and to a CO-mediated disturbance of intracellular "buffering" of endogenously
10     generated free radicals (e.g., 'NO). The role these mechanisms play in pathophysiology currently
11     is being investigated.  Where dose-response studies are available,  the concentrations of CO that
12     cause adverse effects in animals exceed current NAAQS.
13
       February 15, 1999                         5-36        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-a
  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             illness. 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.; Hazucha, M. J.; Smith, M. V.; Bromberg, P. A. (1994) Prediction of carboxyhemoglobin
25             formation due to transient exposure to carbon monoxide. J. Appl. Physiol. 76: 1739-1745.
26       Berk, P. D.; Rodkey, F. L.; Blaschke, T. F.;  Collison, H. A.; Waggoner, J. G. (1974) Comparison of plasma
27             bilirubin turnover and carbon monoxide  production in man. J. Lab. Clin. Med. 83: 29-37.
28       Berk, P. D.; Blaschke, T. F.; Scharschmidt, B. F.; Waggoner, J. G.;  Berlin, N. I. (1976) A new approach to
29             quantitation of the various sources of bilirubin in man. J. Lab. Clin. Med. 87: 767-780.
30       Bernard, T. E.; Duker, J. (1981) Modeling carbon monoxide uptake during work. Am. Ind. Hyg. Assoc. J.
31             42:361-364.
32       Biller, W. F.; Richmond, H. M. (1982) Sensitivity analysis on Coburn model predictions of COHb levels associated
33             with alternative CO standards. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
34             of Air Quality Planning and Standards; EPA contract no. 68-02-3600.
35       Biller, W. F.; Richmond, H. M. (1992) COHb module for a probabilistic CO NEM. In:  Johnson, T.; Capel, J.;
36             Paul, R.; Wijnberg, L. Estimation of carbon monoxide exposures and associated carboxyhemoglobin levels
37             in Denver residents using a probabilistic version of NEM. Research Triangle Park, NC: U.S.  Environmental
3 8             Protection Agency, Office of Air Quality Planning and Standards; contract no. 68-DO-0062; appendix C.
39       Boothby, W. M.; Lovelace, W. R., II; Benson, O. O., Jr.; Strehler, A. F. (1954) Volume and partial pressures of
40             respiratory gases at altitude. In: Boothby, W. M., ed.  Respiratory physiology in aviation. Randolph Field,  TX:
41             U.S. Air Force School of Aviation Medicine; pp. 39-49.
42       Bridges, R. J.; Koh, J. Y.; Hatalski, C. G.; Cotman, C. W. (1991) Increased excitotoxic vulnerability of cortical
43             cultures with reduced levels of glutathione. Eur. J. Pharmacol. 192: 199-200.
44       Britten, J. S.; Myers, R. A. M. (1985) Effects of hyperbaric treatment on carbon monoxide elimination in humans.
45             Undersea Biomed. Res. 12: 431-438.
46       Brown, S. D.; Piantadosi, C. A. (1992) Recovery of energy metabolism in rat brain after carbon monoxide hypoxia.
47             J. Clin. Invest. 89: 666-672.
48       Btinger, J.; Bombosch, F.; Mesecke, U.; Hallier, E. (1997) Monitoring and analysis of occupational exposure to
49             chain saw exhausts. Am. Ind. Hyg. Assoc. J. 58: 747-751.
50       Burki, N. K. (1984) Effects of acute exposure to high altitude on ventilatory drive and respiratory pattern. J. Appl.
51             Physiol. Respir. Environ. Exercise Physiol. 56: 1027-1031.
52       Cantrell, J. M.; Tucker, A. (1996) Low-dose carbon monoxide does not reduce vasoconstriction in isolated rat
53             lungs. Exp. Lung Res. 22: 21-32.
         February 15, 1999                              5-37         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Cassina, A.; Radi, R. (1996) Differential inhibitory action of nitric oxide and peroxynitrate on mitochondrial
  2             electron transport. Arch. Biochem Biophys. 328: 309-316.
  3       Cavallin-Stahl, E.; Mercke, C.; Lundh, B. (1976) Carbon monoxide production in patients with breast carcinoma.
  4             Br. J. Haematol. 32:  177-182.
  5       Chung, S. J. (1988) Formulas predicting carboxyhemoglobin resulting from carbon monoxide exposure. Vet. Hum.
  6             Toxicol. 30: 528-532.
  7       Coburn, R. F. (1967) Endogenous carbon monoxide production and body CO stores. Acta Med. Scand. Suppl.
  8             472:269-282.
  9       Coburn, R. F. (1970a) Enhancement by phenobarbital and diphenylhydantoin of carbon monoxide production in
10             normal man. N. Engl. J. Med. 283: 512-515.
11       Coburn, R. F. (1970b) The carbon monoxide body stores. In: Coburn, R. F., ed. Biological effects of carbon
12             monoxide. Ann. N. Y. Acad. Sci. 174: 11-22.
13       Coburn, R. F. (1973) Mean myoglobin oxygen tension in skeletal and cardiac muscle. In: Bicher, H. I.; Bruley,
14             D. F., eds. Oxygen transport to tissue: instrumentation, methods, and physiology. New York, NY:
15             Plenum Press; pp. 571-577. (Advances in experimental medicine  and biology: v. 37A-B).
16       Coburn, R. F.; Forman, H. J. (1987) Carbon monoxide toxicity. In: Fishman, A. P.; Farhi, L. E.; Tenney, S. M.;
17             Geiger, S. R., eds. Handbook of physiology: a critical, comprehensive presentation of physiological
18             knowledge and concepts. Section 3: The respiratory system. Volume IV. Gas exchange. Bethesda, MD:
19             American Physiological Society; pp. 439-456.
20       Coburn, R. F.; Mayers, L. B. (1971) Myoglobin O2 tension determined from measurements of carboxymyoglobin in
21             skeletal muscle. Am. J. Physiol. 220: 66-74.
22       Coburn, R. F.; Williams, W. J.; Forster, R. E. (1964) Effect of erythrocyte destruction on carbon monoxide
23             production in man. J. Clin. Invest. 43: 1098-1103.
24       Coburn, R. F.; Forster, R. E.; Kane, P. B. (1965) Considerations of the physiological variables that determine the
25             blood carboxyhemoglobin concentration in man. J. Clin. Invest. 44: 1899-1910.
26       Coburn, R. F.; Williams, W. J.; Kahn, S. B. (1966) Endogenous carbon monoxide production in patients with
27             hemolytic anemia. J. Clin. Invest. 45: 460-468.
28       Coceani, F.; Breen,  C. A.; Lees, J. G.; Falck, J. R.; Olley, P. M. (1988) Further evidence implicating a cytochrome
29             P-450-mediated reaction in the contractile tension of the lamb ductus arteriosus. Circ. Res. 62: 471-477.
30       Cook, M. N.; Nakatsu, K.; Marks, G. S.; McLaughlin, B.  E.; Vreman, H. J.; Stevenson, D. K.; Brien, J. F. (1995)
31             Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can.
32             J. Physiol. Pharmacol. 73:515-518.
33       Davies, R. F.; Topping, D. L.; Turner, D. M. (1976) The effect of intermittent carbon monoxide exposure on
34             experimental atherosclerosis in the rabbit. Atherosclerosis 24: 527-536.
35       Delivoria-Papadopoulos, M.; Coburn, R. F.; Forster,  R. E. (1974) Cyclic variation of rate of carbon monoxide
36             production in normal women. J. Appl. Physiol. 36: 49-51.
37       Deller, A.; Stenz, R.; Forstner, K.; Konrad, F. (1992) Die elimination von kohlenmonoxydhamoglobin -
3 8             geschlechtsspezifische und zirkadiane einflusse [The elimination of carboxyhemoglobin COHB sex-specific
39             and circadian influences in healthy volunteers]. Infusionstherapie 19: 121-126.
40       Durante, W.; Kroll, M. H.; Christodoulides, N.; Peyton, K. J.; Schafer, A.  I. (1997) Nitric oxide induces heme
41             oxygenase-1  gene expression and carbon monoxide production in vascular smooth muscle cells. Circ. Res.
42             80:557-564.
43       Engel, L. A.; Wood, L. D. H.; Utz, G.; Macklem, P. T. (1973) Gas mixing during inspiration. J. Appl. Physiol.
44             35: 18-24.
45       Farrugia, G.; Miller, S. M.; Rich, A.; Liu, X.; Maines, M. D.; Rae, J. L.; Szurszewski, J. H.  (1998) Distribution of
46             heme oxygenase and effects of exogenous carbon monoxide in canine jejunum. Am. J. Physiol.
47             274: G350-G358.
48       Fechter, L. D.; Liu, Y.; Pearce, T. A. (1997) Cochlear protection from carbon monoxide exposure by free radical
49             blockers in the guinea pig. Toxicol. Appl. Pharmacol. 142: 47-55.
50       Fein, A.; Grossman, R. F.; Jones, J. G.; Hoeffel, J.; McKay, D. (1980) Carbon monoxide effect on alveolar
51             epithelial permeability. Chest 78: 726-731.
52       Fisher, A. B.; Hyde, R. W.; Baue, A. E.; Reif, J. S.; Kelly, D. F. (1969) Effect of carbon monoxide on function and
53             structure of the lung. J. Appl. Physiol. 26: 4-12.
54       Forbes, W. H.; Sargent, F.; Roughton, F. J. W. (1945) The rate of carbon monoxide uptake by normal men. Am. J.
55             Physiol. 143: 594-608.


         February 15, 1999                              5-38         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Forster, R. E. (1987) Diffusion of gases across the alveolar membrane. In: Fishman, A. P.; Farhi, L. E.; Tenney,
  2             S. M; Geiger, S. R., eds. Handbook of physiology: a critical, comprehensive presentation of physiological
  3             knowledge and concepts. Section 3: the respiratory system. Volume IV. Gas exchange. Bethesda, MD:
  4             American Physiological Society; pp. 71-88.
  5       Frey, T. M.; Crapo, R. O.; Jensen, R. L.; Elliott, C. G. (1987) Diurnal variation of the diffusing capacity of the lung:
  6             is it real? Am. Rev. Respir. Dis. 136: 1381-1384.
  7       Gautier, H.; Bonora, M. (1994) Ventilatory and metabolic responses to cold and CO-induced hypoxia in awake rats.
  8             Respir. Physiol. 97: 79-91.
  9       Ge, R.-L.; Matsuzawa, Y.; Takeoka, M.; Kubo, K.; Sekiguchi, M.; Kobayashi, T. (1997) Low pulmonary diffusing
10             capacity in subjects with acute mountain sickness. Chest 111: 58-64.
11       Gibson, Q. H.; Olson, J. S.; McKinnie, R. E.; Rohlfs, R. J. (1986) A kinetic description of ligand binding to sperm
12             whale myoglobin. J. Biol. Chem. 261: 10,228-10,236.
13       Ginsberg, M. D.; Myers, R. E. (1974) Experimental carbon monoxide encephalopathy in the primate.
14             I. Physiologic and metabolic aspects. Arch. Neurol.  30: 202-208.
15       Goda, N.; Suzuki, K.; Naito, M.; Takeoka, S.; Tsuchida, E.; Ishimura, Y.; Tamatani, T.; Suematsu, M. (1998)
16             Distribution of heme oxygenase isoforms in rat liver: topographic basis for carbon monoxide-mediated
17             microvascular relaxation. J. Clin. Invest. 101: 604-612.
18       Gothert, M.; Lutz, F.; Malorny, G. (1970) Carbon monoxide partial pressure in tissue of different animals.  Environ.
19             Res. 3:  303-309.
20       Gronlund,  J.; Garby, L. (1984) Numerical values of the classical Haldane coefficient. J. Appl. Physiol.: Respir.
21             Environ. Exercise Physiol. 57: 850-859.
22       Grote, J.; Ball, P.; Oltmanns, K.;  Stolp, W. (1994) The effect of increased blood carbon monoxide levels on the
23             hemoglobin oxygen affinity during pregnancy. In: Vaupel, P.; Zander R.; Bruley, D. F., eds. Oxygen
24             transport to tissue XV: proceedings of the twentieth annual meeting of the International Society on Oxygen
25             Transport to Tissue; August 1992; Mainz, Germany. New York, NY: Plenum Press; pp. 145-150. (Advances
26             in experimental medicine and biology v. 345).
27       Grundemar, L.; Johansson, M.-B.; Ekelund, M.; Hogestatt, E. D. (1995)  Haem oxygenase activity in blood vessel
28             homogenates as measured by carbon monoxide production. Acta Physiol. Scand. 153:  203-204.
29       Guenard, H.; Marthan, R. (1996)  Pulmonary gas exchange in elderly subjects. Eur. Respir. J. 9: 2573-2577.
30       Gwag, B. J.; Lobner, D.; Koh, J. T.; Wie, M. B.; Choi, D. W. (1995) Blockade of glutamate receptors unmasks
31             neuronal apoptosis after oxygen-glucose deprivation in vitro. Neuroscience 68: 615-619.
32       Haab, P. E.; Durand-Arczynska, W. Y. (1991) Carbon monoxide effects on oxygen transport. In: Crystal, R. G.;
33             West, J. B.; Barnes, P. J.; Cherniack, N. S.; Weibel, E. R., eds. The lung: scientific applications, v. 2.
34             New York, NY: Raven Press; pp. 1267-1275.
35       Haldane, J. (1897-1898) Some improved methods of gas analysis. J. Physiol. (London) 22: 465-480.
36       Hauck, H.; Neuberger, M. (1984) Carbon monoxide uptake and the  resulting carboxyhemoglobin in man. Eur. J.
37             Appl. Physiol. 53:  186-190.
38       Hausberg, M.; Somers, V. K. (1997) Neural circulatory responses to carbon monoxide in healthy humans.
39             Hypertension 29: 1114-1118.
40       Hill, E. P.; Hill, J. R.; Power, G. G.; Longo, L. D. (1977) Carbon monoxide exchanges between the human fetus and
41             mother: a mathematical model.  Am. J. Physiol.  232: H311-H323.
42       Horvath, S. M.; Agnew, J. W.; Wagner, J. A.; Bedi, J. F. (1988) Maximal aerobic capacity at several ambient
43             concentrations of carbon monoxide at several altitudes. Cambridge, MA: Health Effects Institute; research
44             report no. 21.
45       Hugod, C.; Hawkins, L. H.; Kjeldsen, K.; Thomsen, H. K.; Astrup, P. (1978) Effect of carbon monoxide exposure
46             on aortic and coronary intimal morphology in the rabbit. Atherosclerosis 30: 333-342.
47       Huie, R. E.; Padmaja, S. (1993) The reaction of NO with superoxide. Free Radical Res. Commun.  18: 195-199.
48       Ingi, T.; Ronnett, G. V. (1995) Direct demonstration of a physiological role for carbon monoxide in olfactory
49             receptor neurons. J. Neurosci. 15: 8214-8222.
50       Ingi, T.; Cheng, J.; Ronnett, G. V. (1996) Carbon monoxide: an endogenous modulator of the nitric oxide-cyclic
51             GMP signaling system. Neuron  16: 835-842.
52       Ingram, D.; Dickinson, C. J.; Ahmed, K. (1987) MACPUF: a simulation of human respiration, gas exchange and
53             control. Graphic version. Oxford, United Kingdom: IRL Press Limited.
         February 15, 1999                              5-39         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Ischiropoulos, H.; Zhu, L.; Chen, I; Tsai, M.; Martin, J. C.; Smith, C. D.; Beckman, J. S. (1992)
  2             Peroxynitrate-mediated tyrosine nitration catalyzed by superoxide dismutase. Arch. Biochem. Biophys.
  3             298: 431-437.
  4       Ischiropoulos, H.; Beers, M. F.; Ohnishi, S. T.; Fisher, D.; Garner, S. E.; Thorn, S. R. (1996) Nitric oxide
  5             production and perivascular tyrosine nitration in brain after carbon monoxide poisoning in the rat. J. Clin.
  6             Invest. 97: 2260-2267.
  7       Ishimaru, H.; Nabeshima, T.; Katoh, A.; Suzuki, H.; Fukuta, T.; Kameyama, T. (1991) Effects of successive carbon
  8             monoxide exposures on delayed neuronal death in mice under the maintenance of normal body temperature.
  9             Biochem. Biophys. Res. Commun. 179: 836-840.
10       Ishimaru, H.; Katoh, A.; Suzuki, H.; Fukuta, T.; Kameyama, T.; Nabeshima, T. (1992) Effects of
11             N-methyl-D-aspartate receptor antagonists on carbon monoxide-induced brain damage in mice.
12             J. Pharmacol. Exp. Ther. 261: 349-352.
13       Jay, G. D.; McKindley, D. S. (1997) Alterations in pharmacokinetics of carboxyhemoglobin produced by oxygen
14             underpressure. UnderseaHyperbaric Med. 24: 165-173.
15       Johnson, T.; Capel, J.; Paul, R.; Wijnberg, L. (1992) Estimation of carbon monoxide exposures and associated
16             carboxyhemoglobin levels in Denver residents using a probabilistic version of NEM. Research Triangle Park,
17             NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; contract no.
18             68-DO-0062.
19       Joumard, R.; Chiron, M.; Vidon, R.; Maurin, M.; Rouzioux, J.-M. (1981) Mathematical models  of the uptake of
20             carbon monoxide on hemoglobin at low carbon monoxide levels. Environ. Health Perspect. 41: 277-289.
21       Kanten, W. E.; Penney, D. G.; Francisco, K.; Thill,  J. E. (1983) Hemodynamic responses to acute
22             carboxyhemoglobinemia in the rat. Am. J. Physiol. 244: H320-H327.
23       Kim, S. K.; Kim, Y. C. (1996) Effect of a single administration of benzene, toluene or /w-xylene on
24             carboxyhaemoglobin elevation and metabolism of dichloromethane in rats. J. Appl. Toxicol. 16: 437-444.
25       Kinker, J. R.; Haffor, A.-S.; Stephan, M.; Clanton, T. L. (1992) Kinetics of CO uptake and diffusing capacity in
26             transition from rest to steady-state exercise. J. Appl. Physiol. 72: 1764-1772.
27       Kjeldsen, K.; Astrup, P.; Wanstrup, J. (1972) Ultrastructural intimal changes in the rabbit aorta after a moderate
28             carbon monoxide exposure. Atherosclerosis  16: 67-82.
29       Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
30             effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
31             Health [in press].
32       Koehler, R. C.; Jones, M. D., Jr.; Traystman, R. J. (1982) Cerebral circulatory response to carbon monoxide and
33             hypoxic hypoxia in the lamb. Am. J. Physiol. 243: H27-H32.
34       Kurata, S.; Matsumoto, M.; Yamashita, U. (1996) Concomitant transcriptional activation of nitric oxide synthase
35             and heme oxygenase genes during nitric oxide-mediated macrophage cytostasis. J. Biochem. 120: 49-52.
36       Landaw, S. A. (1973) The effects of cigarette smoking on total body burden and excretion rates  of carbon
37             monoxide. J. Occup. Med. 15: 231-235.
38       Landaw, S. A.; Callahan, E. W., Jr.; Schmid, R. (1970) Catabolism of heme in vivo: comparison of the simultaneous
39             production of bilirubin and carbon monoxide. J. Clin. Invest. 49: 914-925.
40       Levasseur, L.; Galliot-Guilley, M.; Richter, F.; Scherrmann, J. M.; Baud, F. J. (1996) Effects of mode of inhalation
41             of carbon monoxide and of normobaric oxygen administration on carbon monoxide elimination from the
42             blood. Hum. Exp. Toxicol. 15:  898-903.
43       Levitt, M. D.; Ellis, C.; Levitt, D. G. (1994) Diurnal rhythm of heme turnover assessed by breath carbon monoxide
44             concentration measurements. J. Lab. Clin. Med. 124: 427-431.
45       Lifshay, A.; Fast, C. W.; Glazier, J. B. (1971) Effects of changes in respiratory pattern on physiological dead space.
46             J. Appl. Physiol. 31:  478-483.
47       Lipton, S. A.; Kim, W.-K.; Choi, Y.-B.; Kumar, S.;  D'Emila, D. M.; Rayudu, P. V.; Arnelles, D. R.; Stamler, J. S.
48             (1997) Neurotoxicity associated with dual actions of homocysteine at the N-methyl-D-aspartate receptor.
49             Proc. Natl. Acad. Sci. USA 94: 5923-5928.
50       Liu, Y.; Fechter, L. D. (1995) MK-801 protects against carbon monoxide-induced hearing loss. Toxicol. Appl.
51             Pharmacol. 132: 196-202.
52       Lizasoain, I.; Moro, M. A.;  Knowles, R. G.; Darley-Usmar, V.; Moncada, S. (1996) Nitric oxide and peroxynitrate
53             exert distinct effects  on mitochondria! respiration which are differentially blocked by glutathione or glucose.
54             Biochem. J. 314:877-880.
         February 15, 1999                              5-40          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Longo, L. D.; Hill, E. P. (1977) Carbon monoxide uptake and elimination in fetal and maternal sheep. Am. J.
  2             Physiol. 232: H324-H330.
  3       Ludbrook, G. L.; Helps, S. C.; Gorman, D. F.; Reilly, P. L.; North, J. B.; Grant, C. (1992) The relative effects of
  4             hypoxic hypoxia and carbon monoxide on brain function in rabbits. Toxicology 75: 71-80.
  5       Lundh, B.; Cavallin-Stahl, E.; Mercke, C. (1975) Nicotinic acid and the endogenous production of carbon
  6             monoxide. ActaMed. Scand. 197:  173-176.
  7       Luomanmaki, K.; Coburn, R. F. (1969) Effects of metabolism and distribution of carbon monoxide on blood and
  8             body stores. Am. J. Physiol. 217: 354-363.
  9       Luperchio, S.; Tamir, S.; Tannenbaum, S. R. (1996) NO-induced oxidative stress and glutathione metabolism in
10             rodent and human cells. Free Radical Biol. Med. 21: 513-519.
11       MacMillan, V. (1975) Regional cerebral blood flow of the rat in acute carbon monoxide intoxication. Can. J.
12             Physiol. Pharmacol. 53:644-650.
13       Maehara, K.; Riley, M; Galassetti, P.; Barstow,  T. J.; Wasserman, K. (1997) Effect of hypoxia and carbon
14             monoxide on muscle oxygenation during exercise. Am. J. Respir. Crit. Care Med. 155: 229-235.
15       Maines, M. D. (1988) Heme oxygenase: function, multiplicity, regulatory mechanisms, and clinical applications.
16             FASEB J. 2: 2557-2568.
17       Maines, M. D.; Mark, J. A.; Ewing, J. F. (1993)  Heme oxygenase, a likely regulator of cGMP production in the
18             brain: induction in vivo of HO-1 compensates for depression in NO synthase activity. Mol.  Cell. Neurosci.
19             4:398-405.
20       Manno, M.; Rugge, M.; Cocheo, V. (1992) Double fatal inhalation of dichloromethane. Hum. Exp. Toxicol.
21             11:540-545.
22       Marks, G. S. (1994) Heme oxygenase: the physiological role  of one of its metabolites, carbon monoxide and
23             interactions with zinc protoporphyrin, cobalt protoporphyrin and other metalloporphyrins. Cell.  Mol. Biol.
24             (Paris) 40: 863-870.
25       Marthan, R.; Castaing, Y.; Manier, G.; Guenard, H. (1985) Gas exchange alterations in patients with chronic
26             obstructive lung disease. Chest 87: 470-475.
27       Matsuoka, M.; Igisu, H.; Tanaka, L; Hori, H.; Koga, M. (1993) Brain energy metabolites in mice after an acute
28             exposure to carbon monoxide. Res. Commun. Chem. Pathol. Pharmacol. 81: 15-20.
29       Maurer, F. W. (1941) The effects of carbon monoxide anoxemia on the flow and composition of cervical lymph.
30             Am. J. Physiol. 133:  170-179.
31       McCartney, M. L. (1990) Sensitivity analysis applied to Coburn-Forster-Kane models of carboxyhemoglobin
32             formation. Am. Ind. Hyg. Assoc. J. 51: 169-177.
33       McClean, P. A.; Duguid, N. J.; Griffin, P. M.; Newth, C. J. L.; Zamel, N. (1981) Changes in exhaled pulmonary
34             diffusing capacity at rest and exercise in individuals with impaired positional diffusion. Clin. Respir. Physiol.
35             17: 179-186.
36       McCoubrey, W. K., Jr.; Huang, T. J.; Maines, M. D. (1997) Isolation and characterization of a cDNA from the rat
37             brain that encodes hemoprotein heme oxygenase-3. Eur. J. Biochem. 247: 725-732.
3 8       McGrath, J. J. (1988) Carbon monoxide studies at high altitude. Neurosci. Biobehav. Rev. 12:311-314.
39       McGrath, J. J. (1992) Effects of altitude on endogenous carboxyhemoglobin levels. J. Toxicol. Environ. Health
40             35: 127-133.
41       McGrath, J. J.; Schreck, R. M.; Lee, P. S. (1993) Carboxyhemoglobin levels in humans: effects of altitude.
42             Inhalation Toxicol. 5:241-249.
43       Meilin, S.; Rogatsky, G. G.; Thorn, S. R.; Zarchin, N.; Guggenheimer-Furman, E.; Mayevsky, A. (1996) Effects of
44             carbon monoxide on the brain may be mediated by nitric acid. J. Appl. Physiol. 81: 1078-1083.
45       Mercke, C.; Lundh, B. (1976) Erythrocyte filterability and heme catabolism during the menstrual cycle. Ann. Int.
46             Med. 85: 322-324.
47       Mercke, C.; Cavallin-Stahl, E.; Lundh, B. (1975a) Diurnal variation in endogenous production of carbon monoxide.
48             ActaMed. Scand. 198: 161-164.
49       Mercke, C.; Cavallin-Stahl, E.; Lundh, B. (1975b) Heme catabolism during short-term treatment with phenobarbital,
50             diazepamandoxazepam. ActaMed. Scand. 198: 149-154.
51       Mercke, C.; Cavallin-Stahl, E.; Lundh, B. (1975c) Carbon monoxide production and reticulocyte count in normal
52             women: effect of contraceptive drugs and smoking. Acta Med. Scand. 198: 155-160.
53       Messmer, K. (1982) Oxygen transport capacity. In: Brendel, W.; Zink, R. A., eds. High altitude physiology and
54             medicine. New York, NY: Springer-Verlag; pp. 117-122.
         February 15, 1999                             5-41         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Meyer, J.; Prien, T.; Van Aken, H.; Bone, H. G.; Waurick, R.; Theilmeier, G.; Booke, M. (1998) Arterio-venous
  2             carboxyhemoglobin difference suggests carbon monoxide production by human lungs. Biochem. Biophys.
  3             Res. Commun. 244: 230-232.
  4       Moore, E. G.; Gibson, Q. H. (1976) Cooperativity in the dissociation of nitric oxide from hemoglobin. J. Biol.
  5             Chem. 251:2788-2794.
  6       Morita, T.; Perrella, M. A.; Lee, M.-E.; Kourembanas, S. (1995) Smooth muscle cell-derived carbon monoxide is a
  7             regulator of vascular cGMP. Proc. Natl. Acad. Sci. U. S. A. 92: 1475-1479.
  8       Motterlini, R.; Foresti, R.; Intaglietta, M.; Winslow, R. M. (1996) NO-mediated activation of heme oxygenase:
  9             endogenous cytoprotection against oxidative stress to endothelium. Am. J. Physiol. 270: H107-H114.
10       Muller, K. E.; Barton, C. N. (1987) A nonlinear version of the Coburn, Forster and Kane model of blood
11             carboxyhemoglobin. Atmos. Environ. 21: 1963-1967.
12       Murphy, T. H.; Miyamoto, M.; Sastre, A.; Schnaar, R. L.; Coyle, J. T. (1989) Glutamate toxicity in a neuronal cell
13             line involves inhibition of cystine transport leading to oxidative stress. Neuron 2: 1547-1558.
14       Nabeshima, T.; Yoshida, S.; Morinaka, H.; Kameyama, T.; Thurkauf, A.; Rice, K. C.; Jacobson, A. E.; Monn, J. A.;
15             Cho, A. K. (1990) MK-801 ameliorates delayed amnesia, but potentiates acute amnesia induced by CO.
16             Neurosci. Lett. 108: 321-327.
17       Nabeshima, T.; Katoh, A.; Ishimaru, H.; Yoneda, Y.; Ogita, K.; Murase, K.; Ohtsuka, H.; Inari, K.; Fukuta, T.;
18             Kameyama, T. (1991) Carbon monoxide-induced delayed amnesia, delayed neuronal death and change in
19             acetylcholine concentration in mice. J. Pharmacol. Exp. Ther. 256: 378-384.
20       Nathanson, J. A.; Scavone, C.; Scanlon, C.; McKee,  M. (1995) The cellular Na+ pump as a site of action for carbon
21             monoxide and glutamate: a mechanism for long-term modulation of cellular activity. Neuron 14: 781-794.
22       Neas, L. M.;  Schwartz, J. (1996) The determinants of pulmonary diffusing capacity in a national sample of U.S.
23             adults. Am. J. Respir. Crit. Care Med. 153: 656-664.
24       Newby, M. B.; Roberts, R. J.; Bhatnagar, R. K. (1978) Carbon monoxide- and hypoxia-induced effects on
25             catecholamines in the mature and developing rat brain. J. Pharmacol. Exp.  Ther. 206: 61-68.
26       Niden, A. H.; Schulz, H. (1965) The ultrastructural effects of carbon monoxide inhalation on the rat lung. Virchows
27             Arch. Pathol. Anat. Physiol. 339: 283-292.
28       Oka, A.; Belliveau, M. J.; Rosenberg, P. A.; Volpe, J. J. (1993) Vulnerability of oligodendroglia to glutamate:
29             pharmacology, mechanisms, and prevention. J. Neurosci. 13: 1441-1453.
30       Okeda, R.; Funata, N.; Takano, T.; Miyazaki,  Y.; Higashino, F.; Yokoyama, K.; Manabe, M. (1981) The
31             pathogenesis of carbon monoxide encephalopathy  in the acute phase—physiological and morphological
32             correlation. Acta Neuropathol. 54: 1-10.
33       Ott, W. R.; Mage, D. T.  (1978) Interpreting urban carbon monoxide concentrations by means of a computerized
34             blood COHb model. J. Air Pollut. Control Assoc. 28: 911-916.
35       Pace, N.; Strajman, E.; Walker, E. L. (1950) Acceleration of carbon monoxide elimination in man by high pressure
36             oxygen. Science (Washington, DC) 111: 652-654.
37       Pankow, D. (1996) Carbon monoxide formation due  to metabolism of xenobiotics. In: Penney, D. G., ed. Carbon
3 8             monoxide. Boca Raton, FL: CRC Press; pp. 25-43.
39       Pannen, B. H.; Bauer, M. (1998) Differential regulation of hepatic arterial and portal venous vascular resistance by
40             nitric oxide and carbon monoxide in rats. Life Sci. 62: 2025-2033.
41       Parving, H.-H.; Ohlsson, K.; Buchardt Hansen, H. J.; Rorth, M. (1972) Effect of carbon monoxide exposure on
42             capillary permeability to albumin and a2-macroglobulin. Scand. J. Clin. Lab. Invest. 29:  381-388.
43       Pauli, H.  G.; Truniger, B.; Larsen, J. K.; Mulhausen, R. O. (1968) Renal function during prolonged exposure to
44             hypoxia and carbon monoxide. I. Glomerular filtration and plasma flow. Scand. J. Clin. Lab. Invest. Suppl.
45             103:55-60.
46       Pellegrini-Giampietro, D. E.; Cherici, G.; Alesiani, M.; Carla, V.; Moroni, F. (1990) Excititory amino acid release
47             and free radical formation may cooperate in the genesis of ischemia-induced neuronal damage. J. Neurosci.
48             10: 1035-1941.
49       Perm, A.; Currie,  J.; Snyder, C. (1992) Inhalation of carbon monoxide does not accelerate arteriosclerosis in
50             cockerels. Eur. J. Pharmacol. 228: 155-164.
51       Penney, D. G.; Davidson, S. B.; Gargulinski, R. B.; Caldwell-Ayre, T. M. (1988) Heart and lung hypertrophy,
52             changes in blood volume, hematocrit and plasma renin activity in rats chronically exposed to increasing
53             carbon monoxide concentrations. J. Appl. Toxicol. 8: 171-178.
54
         February 15, 1999                              5-42          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Peters, T.; Jiirgens, K. D.; Gunther-Jtirgens, G.; Gros, G. (1994) Determination of myoglobin-diffusivity in intact
  2             skeletal muscle fibers: an improved microscope-photometrical approach. In: Vaupel, P.; Zander, R.; Bruley,
  3             D. F., eds. Oxygen transport to tissue XV. New York, NY: Plenum Press; pp. 677-683. (Advances in
  4             experimental medicine and biology v. 345).
  5       Peterson, J. E.; Stewart, R. D. (1970) Absorption and elimination of carbon monoxide by inactive young men. Arch.
  6             Environ. Health 21: 165-171.
  7       Peterson, J. E.; Stewart, R. D. (1972) Human absorption of carbon monoxide from high concentrations in air.
  8             Am. Ind. Hyg. Assoc. 133:293-297.
  9       Peterson, J. E.; Stewart, R. D. (1975) Predicting the carboxyhemoglobin levels resulting from carbon monoxide
10             exposures. J. Appl. Physiol. 39: 633-638.
11       Piantadosi, C. A.; Tatro, L.; Zhang, J. (1995) Hydroxyl radical production in the brain after CO hypoxia in rats.
12             Free Radical Biol. Med.  18: 603-609.
13       Piantadosi, C. A.; Zhang, J.; Demchenko, I. T. (1997a) Production of hydroxyl radical in the hippocampus after CO
14             hypoxia or hypoxic hypoxia in the rat. Free Radical Biol. Med. 22: 725-732.
15       Piantadosi, C. A.; Zhang, J.; Levin, E. D.; Folz, R. J.; Schmechel, D. E. (1997b) Apoptosis and delayed neuronal
16             damage after carbon monoxide poisoning in the rat. Exp.  Neurol. 147: 103-114.
17       Poderoso, J. J.; Cerreras, M. C.; Lisdero, C.; Riobo, N.;  Schopfer, F.; Boveris, A. (1996) Nitric oxide inhibits
18             electron transfer and increases superoxide radical production in rat heart mitochondria and submitochondrial
19             particles. Arch. Biochem. Biophys. 328: 85-92.
20       Robinson, N. B.; Barie, P. S.; Halebian, P. H.; Shires, G. T. (1985) Distribution of ventilation and perfusion
21             following acute  carbon monoxide poisoning. In: 41st Annual forum on fundamental surgical problems held at
22             the 71st annual clinical congress of the American College of Surgeons; October; Chicago, IL. Surg. Forum
23             36: 115-118.
24       Rodgers, P. A.; Vreman, H. J.; Dennery, P. A.; Stevenson, D. K. (1994) Sources of carbon monoxide (CO) in
25             biological systems and applications of CO detection technologies. Semin. Perinatal. 18: 2-10.
26       Rodkey, F. L.; O'Neal, J. D.; Collison, H. A. (1969) Oxygen and carbon monoxide equilibria of human adult
27             hemoglobin at atmospheric and elevated pressure. Blood 33: 57-65.
28       Roughton, F. J. W. (1970) The equilibrium of carbon monoxide with human hemoglobin in whole blood.
29             In: Coburn, R. F., ed. Biological effects of carbon monoxide. Ann. N. Y. Acad. Sci. 174:  177-188.
30       Saltzman, B. E.; Fox, S. H. (1986) Biological significance of fluctuating concentrations of carbon monoxide.
31             Environ. Sci. Technol. 20: 916-923.
32       Schoenfisch, W. H.; Hoop, K. A.; Struelens, B. S. (1980) Carbon monoxide absorption through the oral and nasal
33             mucosae of cynomolgus monkeys. Arch. Environ. Health 35: 152-154.
34       Seki, T.; Naruse, M.; Naruse, K.; Yoshimoto, T.; Tanabe, A.; Imaki, T.; Hagiwara, H.; Hirose, S.; Demura, H.
3 5             (1997) Interrelation between nitric oxide synthase and heme oxygenase in rat endothelial cells. Eur. J.
36             Pharmacol.  331:87-91.
37       Selvakumar, S.; Sharan, M.; Singh, M. P. (1992) Mathematical model for the exchange of gases in the lungs with
38             special reference to carbon monoxide. Med. Biol. Eng. Comput. 30: 525-532.
39       Shimazu, T.; Ikeuchi, H.; Hubbard, G. B.; Langlinais, P. C.; Mason, A. D., Jr.; Pruitt, B. A.,  Jr. (1990) Smoke
40             inhalation injury and the effect of carbon monoxide in the sheep model. J. Trauma 30: 170-175.
41       Siggaard-Andersen, J.; Bonde Petersen, F.; Hansen, T. L; Mellemgaard, K. (1968) Plasma volume and vascular
42             permeability during hypoxia and carbon monoxide exposure. Scand. J. Clin. Lab. Invest. Suppl.  103: 39-48.
43       Singh, J.; Moore-Cheatum, L. (1993) Gestational protein deficiency enhances fetotoxicity of carbon monoxide.
44             In: Keen, C. L.;  Bendich, A.; Willhite, C. C., eds. Maternal nutrition and pregnancy outcome. Ann. N. Y.
45             Acad. Sci. 678: 366-368.
46       Singh, M. P.; Sharan, M.; Selvakumar, S. (1991) A mathematical model for the computation of
47             carboxyhaemoglobin in human blood as a function of exposure time. Philos. Trans. R. Soc. Lond. B
48             334: 135-147.
49       Singh, J.; Smith, C. B.; Moore-Cheatum, L. (1992) Additivity of protein deficiency and carbon monoxide on
50             placental carboxyhemoglobin in mice. Am. J. Obstet. Gynecol. 167: 843-846.
51       Singh, J.; Aggison, L.,  Jr.; Moore-Cheatum, L. (1993) Teratogenicity and developmental toxicity of carbon
52             monoxide in protein-deficient mice. Teratology 48:  149-159.
53       Singleton, G. J.; Olsen, C. R.; Smith, R. L. (1972) Correction for mechanical dead space in the calculation of
54             physiological dead space. J. Clin. Invest. 51: 2768-2772.
         February 15, 1999                              5-43          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Smith, M. V. (1990) Comparing solutions to the linear and nonlinear CFK equations for predicting COHb
  2             formation. Math. Biosci. 99:251-263.
  3       Smith, M. V.; Hazucha, M. J.; Benignus, V. A.; Bromberg, P. A. (1994) Effect of regional circulation patterns on
  4             observed HbCO levels. J. Appl. Physiol. 77: 1659-1665.
  5       Sokal, J. A.; Majka, J.; Palus, J. (1984) The content of carbon monoxide in the tissues of rats intoxicated with
  6             carbon monoxide in various conditions of acute exposure. Arch. Toxicol. 56:  106-108.
  7       Sokal, J.; Majka, J.; Palus, J. (1986) Effect of work load on the content of carboxymyoglobin in the heart and
  8             skeletal muscles of rats exposed to carbon monoxide. J. Hyg. Epidemiol. Microbiol. Immunol. 30: 57-62.
  9       Solanki, D. L.; McCurdy, P. R.; Cuttitta, F. F.; Schechter, G. P. (1988) Hemolysis in sickle cell disease as measured
10             by endogenous carbon monoxide production: a preliminary report. Am. J. Clin. Pathol.
11             89:221-225.
12       Song, S.-Y.; Okeda, R.; Funata, N.; Higashino, F. (1983) An experimental study of the pathogenesis of the selective
13             lesion of the globus pallidus in acute carbon monoxide poisoning in cats: with special reference to the
14             chronologic change in the cerebral local blood flow. Acta Neuropathol. 61: 232-238.
15       Stewart, R. D.;  Peterson, J. E.; Baretta, E. D.; Bachand, R. T.; Hosko, M. J.; Herrmann, A. A.  (1970) Experimental
16             human exposure to carbon monoxide. Arch. Environ. Health 21: 154-164.
17       Stewart, R. D.;  Peterson, J. E.; Fisher, T. N.; Hosko, M. J.; Baretta, E. D.; Dodd, H. C.; Herrmann, A. A. (1973)
18             Experimental human exposure to high concentrations of carbon monoxide. Arch. Environ. Health 26:  1-7.
19       Stokes, D. L.; Maclntyre, N. R.; Nadel, J. A. (1981) Nonlinear increases in diffusing capacity  during exercise by
20             seated and supine subjects. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 51: 858-863.
21       Stone, J. R.; Marietta, M. A. (1994) Soluble guanylate cyclase from bovine lung: activation with nitric oxide and
22             carbon monoxide and spectral characterization of the ferrous and ferric states. Biochemistry 33: 5636-5640.
23       Sugi, K.; Theissen, J. L.; Traber, L. D.; Herndon, D. N.; Traber, D. L. (1990) Impact of carbon monoxide on
24             cardiopulmonary dysfunction after smoke inhalation injury. Circ. Res. 66: 69-75.
25       Tamayo, L.; Lopez-Lopez, J. R.; Castaneda, J.; Gonzalez, C. (1997) Carbon monoxide inhibits hypoxic pulmonary
26             vasoconstriction in rats by a cGMP-independent mechanism. Pfluegers Arch. 434: 698-704.
27       Thorn, S. R. (1990) Carbon monoxide-mediated brain lipid peroxidation in the rat. J. Appl. Physiol. 68: 997-1003.
28       Thorn, S. R. (1992) Dehydrogenase conversion to oxidase and lipid peroxidation in brain after carbon monoxide
29             poisoning. J. Appl. Physiol. 73: 1584-1589.
30       Thorn, S. R. (1993) Leukocytes in carbon monoxide-mediated brain oxidative injury. Toxicol. Appl. Pharmacol.
31             123:234-247.
32       Thorn, S. R.; Ischiropoulos, H. (1997) Mechanism of oxidative stress from low levels of carbon monoxide.
33             Cambridge, MA: Health Effects Institute; research report no. 80.
34       Thorn, S. R.; Ohnishi, S. T.; Ischiropoulos, H. (1994)  Nitric oxide released by platelets inhibits neutrophil B2
3 5             integrin function following acute carbon monoxide poisoning. Toxicol. Appl.  Pharmacol. 128:  105-110.
36       Thorn, S. R.; Xu, Y. A.; Ischiropoulos, H. (1997) Vascular endothelial cells generate peroxynitrite in response to
37             carbon monoxide exposure. Chem. Res. Toxicol. 10: 1023-1031.
38       Thorn, S. R.; Garner, S.; Fisher, D.; Ischiropoulos, H. (1998) Vascular nitrosative stress from  CO exposure.
39             In: Program and abstracts [of the] Undersea and Hyperbaric Medical Society annual scientific meeting:
40             pre- and post-courses; May; Seattle, WA. Undersea Hyperbaric Med. 25(suppl.): 47.
41       Thorn, S. R.; Ohnishi, S. T.; Fisher, D.; Xu, Y. A.; Ischiropoulos, H. (1999a) Pulmonary vascular stress from
42             carbon monoxide. Toxicol. Appl. Pharmacol. 154: 12-19.
43       Thorn, S. R.; Fisher, D.; Xu, Y. A.; Garner, S.; Ischiropoulos, H. (1999b) Role of nitric oxide-derived oxidants in
44             vascular injury from carbon monoxide in the rat. Am. J. Physiol.: in press.
45       Thomsen, H. K. (1974) Carbon monoxide-induced atherosclerosis in primates: an electron-microscopic study on the
46             coronary arteries of Macaco irus monkeys. Atherosclerosis 20: 233-240.
47       Tikuisis, P. (1996) Modeling the uptake and elimination of carbon monoxide. In: Penney, D. G., ed. Carbon
48             monoxide. Boca Raton, FL: CRC Press; pp. 45-67.
49       Tikuisis, P.; Buick, F.; Kane, D. M. (1987a) Percent carboxyhemoglobin in resting humans exposed repeatedly to
50             1,500 and 7,500 ppm CO. J. Appl. Physiol. 63: 820-827.
51       Tikuisis, P.; Madill, H. D.; Gill, B. J.; Lewis, W. F.; Cox, K. M.; Kane, D. M. (1987b)  A critical analysis of the use
52             of the CFK equation in predicting COHb formation. Am. Ind. Hyg. Assoc. J. 48: 208-213.
53       Tikuisis, P.; Kane, D. M.; McLellan, T. M.; Buick, F.; Fairburn, S. M. (1992) Rate of formation of
54             carboxyhemoglobin in exercising humans exposed to carbon monoxide. J. Appl. Physiol. 72: 1311-1319.
         February 15, 1999                              5-44          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Trischmann, U.; Klockner, U.; Isenberg, G.; Utz, J.; Ullrich, V. (1991) Carbon monoxide inhibits
  2             depolarization-induced Ca rise and increases cyclic GMP in visceral smooth muscle cells. Biochem.
  3             Pharmacol. 41:237-241.
  4       Turner, D. M; Lee, P. N.; Roe, F. J. C.; Gough, K. J. (1979) Atherogenesis in the White Carneau pigeon: further
  5             studies of the role of carbon monoxide and dietary cholesterol. Atherosclerosis 34: 407-417.
  6       Tyuma, I.; Ueda, Y.; Imaizumi, K.; Kosaka, H. (1981) Prediction of the carbonmonoxyhemoglobin levels during
  7             and after carbon monoxide exposures in various animal species. Jpn. J. Physiol. 31: 131-143.
  8       U.S. Environmental Protection Agency. (1978) Altitude as a factor in air pollution. Research Triangle Park, NC:
  9             Office of Research and Development, Environmental Criteria and Assessment Office; report no. EPA
10             600/9-78-015. Available from: NTIS, Springfield, VA; PB-285645.
11       U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
12             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
13             no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
14       U.S. Environmental Protection Agency. (1992) Review of the national ambient air quality standards for carbon
15             monoxide: 1992 reassessment of scientific and technical information. OAQPS staff paper. Research Triangle
16             Park, NC: Office of Air Quality Planning and Standards; report no. EPA-452/R-92-004. Available from:
17             NTIS, Springfield, VA; PB93-157717.
18       Utz, J.; Ullrich, V. (1991) Carbon monoxide relaxes ileal smooth muscle through activation of guanylate cyclase.
19             Biochem. Pharmacol. 41: 1195-1201.
20       Venkatram, A.; Louch, R. (1979) Evaluation of CO air quality criteria using a COHb model. Atmos. Environ.
21             13:869-872.
22       Wagner, J. A.; Horvath, S. M.; Dahms, T. E. (1975) Carbon monoxide elimination. Respir. Physiol. 23: 41-47.
23       Wang, R. (1998) Resurgence of carbon monoxide: an endogenous gaseous vasorelaxing factor. Can. J. Physiol.
24             Pharmacol. 76: 1-15.
25       Wang, R.; Wu, L.; Wang, Z. (1997a) The direct effect of carbon monoxide on KCa channels in vascular smooth
26             muscle cells. Pfluegers Arch. 434: 285-291.
27       Wang, R.; Wang, Z.; Wu, L. (1997b) Carbon monoxide-induced vasorelaxation and the underlying mechanisms.
28             Br. J. Pharmacol. 121: 927-934.
29       Webster, W. S.;  Clarkson, T. B.; Lofland, H. B. (1968) Carbon monoxide-aggravated atherosclerosis in the squirrel
30             monkey. Exp. Mol. Pathol. 13: 36-50.
31       Werner, B.; Lindahl, J. (1980) Endogenous carbon monoxide production after bicycle exercise in healthy subjects
32             and in patients with hereditary spherocytosis. Scand. J. Clin. Lab. Invest. 40:  319-324.
33       Wittenberg, B. A.; Wittenberg,  J. B. (1993) Effects of carbon monoxide on isolated heart muscle cells. Cambridge,
34             MA: Health Effects Institute; research report no. 62.
35       Woehlck, H. J.; Dunning, M. B., Ill; Kulier, A. H.; Sasse, F. J.; Nithipataikom, K.; Henry, D. W. (1997a) The
3 6             response of anesthetic agent monitors to trifluoromethane warns of the presence of carbon monoxide from
37             anesthetic breakdown. J. Clin. Monit. 13:  149-155.
38       Woehlck, H. J.; Dunning, M., Ill; Connolly, L. A. (1997b) Reduction in the incidence of carbon monoxide
39             exposures in humans undergoing general anesthesia. Anesthesiology 87: 228-234.
40       Wyman, J.; Bishop, G.; Richey, B.; Spokane, R.; Gill, S. (1982) Examination of Haldane's first law for the partition
41             of CO and O2 to hemoglobin A0. Biopolymers 21: 1735-1747.
42       Yamaya, M.; Sekizawa, K.; Ishizuka, S.; Monma, M.; Mizuta, K.; Sasaki, H. (1998) Increased carbon monoxide in
43             exhaled air of subjects with upper respiratory tract infections. Am. J. Respir.  Crit. Care Med. 158:311-314.
44       Yoneda, I.; Watanabe, Y. (1997) Comparisons of altitude tolerance and hypoxia symptoms between nonsmokers
45             and habitual smokers. Aviat. Space Environ. Med. 68: 807-811.
46       Zayasu, K.; Sekizawa, K.; Okinaga, S.; Yamaya, M.; Ohrui, T.; Sasaki, H. (1997) Increased carbon monoxide in
47             exhaled air of asthmatic patients. Am. J. Respir. Crit. Care Med. 156: 1140-1143.
48       Zhu, N.; Weiss, H. R. (1994) Effect of hypoxic and carbon monoxide-induced hypoxia on regional myocardial
49             segment work and O2 consumption. Res. Exp. Med. 194: 97-107.
         February 15, 1999                             5-45         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 summarizes the current understanding of
11     health effects that may occur in individuals breathing CO in ambient air. A more detailed
12     discussion of studies reporting CO-associated health effects can be found in the previous
13     document, Air Quality Criteria for Carbon Monoxide (U.S. Environmental Protection Agency,
14     1991) and in a number of excellent reviews (Kleinman, 1992; Bascom et al., 1996; Penney,
15     1996a).
16          Although evidence from laboratory animal studies indicates that CO can adversely affect
17     the cardiovascular and nervous systems of both mature animals and developing offspring, the
18     concentrations of CO used during exposure and consequent levels of carboxyhemoglobin
19     (COHb) saturation are much higher than typically would be experienced by ambient exposures of
20     humans. The laboratory animal studies, therefore, must be interpreted with caution.  However,
21     they can be useful for exploring the properties and possible mechanisms of an effect much more
22     thoroughly and extensively than is possible in humans.  An effort will be made in this chapter to
23     compare the health  effect levels for CO found in laboratory animal and human controlled-
24     exposure studies.
25          A review of the health effects literature on CO since the last assessment was published in
26     1991 finds many published articles on CO poisoning, possibly reflecting more media attention to
27     this topic. Many of these articles, however, reported effects at CO levels far higher than in
28     ambient air.  Severe effects from acute exposure to high levels of CO are not directly germane to
29     the problems associated with exposure to current ambient levels of CO and, thus, are not
30     discussed in detail in this chapter.  They are, however, mentioned briefly in the summary of this

       February 15, 1999                         6-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      chapter and in the following chapter to present a complete understanding of what is known about
 2      CO toxicity and to provide public health information about potential effects of accidental
 3      exposure to CO, particularly those exposures occurring indoors.
 4           The next section of this chapter (Section 6.2) addresses cardiovascular effects of CO. The
 5      section begins (Section 6.2.1) with a discussion of epidemiologic studies because of their
 6      potential importance in assessing health effects of low-level, ambient CO exposure. In available
 7      epidemiologic studies, associations of short-term ambient CO concentrations with frequency of
 8      cardiovascular hospital admissions, especially for congestive heart failure (CHF), have been
 9      observed quite consistently.  An association of short-term CO levels with mortality also has been
10      observed in several studies.  This association does not appear to be as specific as that of CO with
11      CHF admissions.
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 the studies on neurobehavioral effects that provided the scientific basis
21      for the first CO NAAQS.  Subsequent assessments of the neurobehavioral literature, however,
22      have questioned both the methods and results of the early studies. Articles published since the
23      last assessment in 1991 have been mostly reviews that attempted to understand the equivocal
24      results found at low COHb levels and to provide a physiological basis for behavioral effects.
25      This  section (Section 6.3) discusses the difficulty in studying an organ system with exquisite
26      compensatory responses to a reduced oxygen supply (hypoxia).
27           The rest of the chapter summarizes current knowledge about developmental toxicity
28      (Section 6.4), acute pulmonary effects (Section 6.5), other systemic effects of CO (Section 6.6),
29      physiologic responses to CO exposure (Section 6.7), and combined exposure of CO with other
30      pollutants, drugs, and environmental factors (Section 6.8).  Little new information has been
31      published on these areas of CO toxicity, and their summaries remain essentially the same as

        February 15, 1999                          6-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1      published in the previous criteria document (U.S. Environmental Protection Agency, 1991).
 2      Significant new studies have been added to the summaries, where applicable, if these studies
 3      provide confirmation of the conclusions.  No new published studies are known to refute or draw
 4      into question the conclusions drawn from the previous literature review of these topics. Finally,
 5      a summary section (Section 6.9) provides a concise review of the key human health effects most
 6      clearly demonstrated to be associated with ambient exposure to 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.  In fact, more than one in five males and
17      females have some form of CVD. Because the numbers of affected people are so high, any true
18      relationship between CO exposure and increased cardiovascular mortality or morbidity in the
19      population could have a large impact on both public health and health care costs. In the
20      following discussion, the effects of CO on potentially susceptible population groups are explored
21      through epidemiologic and controlled laboratory studies (for a  more detailed discussion of
22      subpopulations at risk, see Chapter 7).
23
24      6.2.1 Epidemiologic Studies
25      6.2.1.1  Introduction
26           In recent years, many epidemiologic studies have shown associations of ambient air
27      pollution exposure with mortality, morbidity, and physiologic changes. These studies have
28      increased concern that ambient air pollution can promote, and perhaps even produce, adverse
29      health outcomes, even when concentrations of individual pollutants are at or below current U.S.
30      air quality standards. At the same time, the relevant epidemiologic database is growing rapidly,

        February 15,  1999                         6-3         DRAFT-DO  NOT QUOTE OR CITE

-------
 1      and its interpretation is changing over time. For example, as recently as the mid-1990s,
 2      numerous epidemiologic studies had reported associations of mortality and morbidity with
 3      exposure to ambient particulate matter (PM), but relatively few had investigated or reported such
 4      associations with gaseous pollutants such as CO. Since then, however, investigators have given
 5      more thorough and balanced consideration to PM and gaseous pollutants, and although
 6      statistically significant associations of harmful health outcomes with PM continue to be
 7      observed, the role of gaseous pollution exposure appears stronger than it did even a few years
 8      ago.
 9           Recent epidemiologic studies of CO exemplify this point. To date, several well-conducted
10      studies in the United States and Canada have reported an association of ambient CO exposure
11      with hospital admissions for CHF. Current evidence suggests that this association is quite
12      specific in two important ways. First, consistent and strong associations have been found
13      between CFIF admissions and CO, whereas associations with other criteria pollutants are less
14      clear.  Second, CO does not appear to be especially strongly associated with admissions for
15      cardiovascular causes other than CHF or with respiratory disorders. Also, the association of CO
16      with CHF has considerable biological plausibility, particularly because it appears to be strongest
17      during cold weather, which may heighten underlying susceptibility to the worsening of CHF.
18      On balance, the observed association of ambient CO exposure with CHF admissions is as
19      biologically and epidemiologically credible as any other association of ambient air pollution with
20      harmful health outcome yet observed in time series studies of ambient air pollution.
21      An association of ambient CO exposure with mortality also has been reported in epidemiologic
22      studies, although not as consistently or specifically as with CHF admissions.
23           At the same time, there remain important uncertainties in the current epidemiologic
24      database for ambient CO and other air pollutants.  For example, the health effects of long-term
25      ambient CO exposures are not known. Also, there is increasing realization that health effects
26      observed in association with any single pollutant may actually be mediated by multiple
27      components of the complex ambient mix.  Specific attribution of effects to CO or any other
28      single pollutant thus may yield an oversimplistic picture of biological reality. Also, observed
29      effects of CO and other pollutants appear to be subtle in relation to effects of other  risk factors
30      such as smoking, socioeconomic status, and demographic factors.  This lends uncertainty to the


        February 15, 1999                          6-4         DRAFT-DO NOT QUOTE OR CITE

-------
 1      potential public health consequences of further reductions of ambient CO or any other single
 2      ambient air pollutant at present-day U.S.  levels.
 3           Individual epidemiologic studies of ambient CO exposure are discussed below, first for
 4      hospital admissions (largely CHF admissions) and then for daily mortality.
 5
 6      6.2.1.2 Carbon Monoxide and Hospital Admissions
 7           Kurt et al. (1978) observed a relationship between ambient CO levels and emergency room
 8      visits for cardiovascular complaints in Denver, CO.  However, correlations were relatively weak,
 9      and environmental factors other than CO were not evaluated. A number of recent studies in the
10      United States (Morris et al., 1995; Schwartz and Morris, 1995; Schwartz, 1997; Morris and
11      Naumova, 1998), Canada (Burnett et al.,  1997), and Europe (Pantazopoulou et al., 1995;
12      Poloniecki et al.,  1997) have suggested that daily variation in ambient CO levels is related to
13      daily hospital admissions for cardiovascular illness,  especially for both CHF and people over
14      64 years of age.
15           As discussed above, the observed association of ambient CO exposure with CHF
16      admissions has biological plausibility. However, full understanding of this observation will
17      require resolution of several issues. Congestive heart failure is a major cause  of hospital
18      admissions.  In 1996, there were an estimated 4,239,000 hospital admissions for heart disease in
19      the United States, and 870,000 (20.5%) of these had a discharge diagnosis of CHF (Graves and
20      Owings, 1998). In addition, CHF is one of the most common indications for hospitalization in
21      adults over 64 years old (May et al., 1991; Croft et al., 1997; Haldeman et al., 1998).  At the
22      same time, readmission rates are substantial in elderly CHF patients,  ranging from 29 to 47%.
23      Behavioral factors (e.g., smoking, noncompliance with medications and dietary guidelines) and
24      social factors (e.g., isolation) contribute to CHF readmissions, suggesting that a portion of such
25      admissions could be prevented. Some admissions with a diagnosis of CHF actually may result
26      from other causes such as noncardiogenic pulmonary edema and pneumonia.  Although these
27      factors would probably not seriously confound observed relationships of short-term ambient CO
28      exposure and CHF admissions, they lend uncertainty as to the actual number of CHF patients
29      potentially at risk. Additional characterization of individual subjects (e.g., smoking habits,
30      personal CO exposure, specific clinical findings) would be highly desirable in future
31      epidemiologic studies of ambient CO exposure. Also, as mentioned in Chapter 3, in most U.S.

        February 15, 1999                          6-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1      metropolitan areas there is considerable spatial variation among simultaneous CO measurements
 2      made at different monitoring sites. In most epidemiologic studies to date, exposure metrics have
 3      consisted of CO measurements averaged across sites. The effects of such averaging on statistical
 4      health effects estimates are not yet well understood.
 5           The average daily maximum CO concentrations measured by stationary monitors in the
 6      epidemiologic studies are generally very low (<5 ppm). Any increase over endogenous COHb
 7      levels produced by a 1-h exposure to <10 ppm exogenous CO, for example, would be difficult to
 8      measure. Even 8 h of exposure to 10 ppm CO with light to moderate exercise ventilation
 9      (20 L/min) would produce only 1.5% COHb, assuming a baseline of 0.5% COHb.  In addition,
10      CO levels measured at stationary monitors usually are not highly correlated with personal CO
11      exposures,  especially for compromised persons (such as cardiac patients) who spend much of
12      their time indoors.
13           Morris  et al. (1995) performed  a time-series analysis of ambient levels of gaseous air
14      pollutants (CO, nitrogen dioxide [NO2], sulfur dioxide [SO2], and ozone [O3]) and Medicare
15      hospital admissions for CHF in seven U.S. cities (Chicago, IL; Detroit, MI; Houston, TX;
16      Los Angeles, CA; Milwaukee, WI; New York City, NY; and Philadelphia, PA) during the 4-year
17      period from 1986 through 1989. The average daily maximum  1-h CO levels (mean ± standard
18      deviation [SD]) ranged from 1.8 (±1.0) ppm in Milwaukee to 5.6 (±1.7) ppm in New York City.
19      The relative risk of admissions associated with a 10 ppm increase in CO ranged from 1.10 in
20      New York  City to 1.37 in Los Angeles. All seven cities showed similar patterns of increasing
21      admissions with increasing ambient CO concentrations. The authors estimated that
22      approximately 3,250 hospital admissions for congestive heart failure (5.7% of all such
23      admissions) each year could be attributed to the observed association with CO levels.
24           Schwartz and Morris (1995) examined air pollution and cardiovascular hospital admissions
25      for people aged 65 years or older in the Detroit metropolitan area during the years 1986 through
26      1989.  Air quality data were available for PM < 10 jim (PM10) on 82% of possible days and for
27      O3 on 85% of possible days.  Data were available for SO2 and CO on all days during the study
28      period. The mean PM10 was 48.0 //g/m3, the mean O3 was 41.0 ppb, the mean SO2 was 25.4 ppb,
29      and the mean CO was 2.4 ppm. A Poisson auto-regressive model was used to analyze the data,
30      with independent variables for temperature, dew point, month, and linear and quadratic time
31      trends. For each pollutant, the difference between the 25th and 75th percentiles of the

        February 15,  1999                          6-6         DRAFT-DO NOT QUOTE OR CITE

-------
 1      distribution of ambient concentrations during the study period (the "interquartile range") was
 2      calculated.  Relative risks for the health outcomes assessed were reported per increment in
 3      ambient concentration equal to each pollutant's interquartile range. Daily admissions for
 4      ischemic heart disease were associated with interquartile range increases of 32 |ig/m3 for PM10
 5      (relative risk [RR] = 1.018; 95% confidence interval [CI] 1.005 to 1.032), 18 ppb for SO2
 6      (RR= 1.014; 95% CI 1.003 to 1.026), and 1.28 ppm for CO  (RR= 1.010; 95% CI 1.001 to
 7      1.018); however, both SO2 and CO became insignificant after controlling for PM10, whereas PM10
 8      remained significant after controlling for the other pollutants. Daily admissions for heart failure
 9      were associated independently with the interquartile range increases for both PM10 (RR = 1.024;
10      95% CI 1.004 to  1.044) and CO (RR = 1.022; 95% CI 1.010 to 1.034).  Ozone was not a
11      significant risk factor for cardiovascular admissions, and no  pollutant was a significant risk factor
12      for dysrhythmia admissions.
13           It is possible that CO could be a surrogate for general mobile-source pollution.  In some
14      locations, CO is highly correlated with particles during the winter months.  Particles (PM10) were
15      found to be associated with CVD in the Schwartz and Morris (1995) study, but were not assessed
16      in the Morris et al. (1995) analysis.  Also, particles previously have been shown to be associated
17      with hospital admissions for both heart failure and ischemic  heart disease in Ontario (Burnett
18      etal., 1995).
19           Schwartz (1997) examined relationships of short-term  ambient air pollution levels with
20      cardiovascular hospital admissions in people at least 65 years old in Tucson, AZ, from 1988
21      through 1990. The author assessed effects of CO, SO2, O3, NO2, and PM10, as measured at a
22      single monitoring station, as well as of meteorologic variables.  Poisson regression was
23      conducted on temperature, humidity, day of week, and  air  pollution levels.  Long-term temporal
24      patterns were removed by nonparametric smoothing. It appears that exposure-to-admission lags
25      of 0, 1, and 2 days were assessed in different models, and that effects estimates for levels on the
26      same day as admission were reported. During the study, the median maximum hourly CO
27      concentration and the median 24-h average PM10 concentration were 3.03 ppm and 39 |ig/m3,
28      respectively. The correlation between PM10 and SO2 was lower than in eastern U.S. cities.
29      Relative risk estimates of CO and PM10 for cardiovascular admissions were of similar magnitude,
30      independent, additive, and statistically significant at a 5%  level of significance (p). In a model
31      assessing both pollutants simultaneously, the estimated percentage increases in admissions across

        February 15, 1999                          6-7        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the interquartile ranges of CO and PM10 levels were 2.33% and 2.37%, respectively.  Relative
 2      risk estimates for both pollutants appeared to be quite stable across seasons, and to be
 3      unconfounded with the meteorologic parameters assessed. There was no appreciable association
 4      of SO2, O3, or NO2 levels with cardiovascular admissions in the elderly.
 5           Burnett et al. (1997) examined temporal relationships between ambient air pollutants and
 6      hospitalizations among the elderly (individuals >64 years of age) in 10 Canadian cities during the
 7      11-year period from 1981 through 1991. Out of a study population of 12.6 million, an average of
 8      39 elderly patients were admitted daily to 134 hospitals because of CHF.  A time series analysis
 9      adjusted for long-term time trends, seasonal and subseasonal temporal variations, and day-of-
10      week effects was used to explore the association between cardiopulmonary illness and the
11      ambient air pollutants CO, NO2, SO2, and O3 and coefficient of haze (COH). After stratifying by
12      month of the year and adjusting for temperature, dew point, and other pollutants, the log of the
13      daily 1-h maximum CO concentration recorded on the day of admission had the strongest and
14      most consistent statistical association with hospitalization for CHF. The relative risk was 1.065
15      (95% CI = 1.028 to 1.104) for an increase from 1 to 3 ppm CO (the interquartile range of the
16      exposure distribution). Associations of levels of other pollutants with admissions also were
17      observed in single-pollutant models. However, risk estimates for these other pollutants were
18      more sensitive to simultaneous adjustment for multiple pollutants or weather variables than were
19      the estimates for CO.  The authors noted that CO could possibly be acting as a marker for
20      pollution from transportation sources in general, and that independent effects of non-CO
21      pollutants could not be ruled out.
22           Pantazopoulou et al. (1995) studied the daily number of emergency outpatient visits and
23      admissions for cardiac and respiratory causes to all major hospitals in the greater Athens, Greece,
24      area during 1988.  Air pollutant concentrations were obtained from the Greek Ministry of the
25      Environment for smoke, CO, and NO2. Mean levels of CO for all available monitoring stations
26      were 4.5 ± 1.6 (SD) mg/m3 (4.0 ± 1.4 ppm) in winter and 3.4 ± 1 (SD) mg/m3 (3.0 ± 0.9 ppm) in
27      summer.  Multiple linear regression modeling was used, controlling, separately for winter and
28      summer, for the potential effects of meteorological and chronological variables. A positive
29      association was found between the daily number of emergency admissions for cardiac and
30      respiratory causes and all measured pollutants during the winter, but not during the summer.


        February 15, 1999                           6-8        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Poloniecki et al. (1997) studied emergency admissions to London, England, hospitals for
 2      circulatory diseases from 1987 through 1994. Time series relationships were assessed with daily
 3      measurements of O3, NO2, SO2, CO, and black smoke on the day prior to admission.  Long-term
 4      and day-of-week trends, temperature, humidity, an influenza epidemic in 1989, and cyclical
 5      covariations with periodicity >20 days in daily measures of pollution and admissions were
 6      accounted for in the analysis.  No associations were found between O3 and circulatory diseases;
 7      the remaining pollutants were associated with combined acute myocardial infarction (MI) and
 8      circulatory diseases.  The associations with acute MI were significant only in the cool season;
 9      attributable cases (95% CIs) and p values were 2.5% (0.8 to 4.3%), p = 0.003 for black smoke;
10      2.7% (0.8 to 4.6%), p = 0.002 for NO2; 2.1% (0.7 to 3.5%), p = 0.001 for CO; and 1.7% (0.7 to
11      2.6%), p = 0.0006 for SO2.  There were also associations between black smoke and angina
12      (p = 0.02), NO2 and arrhythmia (p = 0.04), and CO  and other circulatory diseases  (p = 0.004).
13      However, unlike several U.S.  studies (Morris et al., 1995; Schwartz and Morris, 1995; Morris
14      and Naumova, 1998), the ambient CO level was not associated with admissions for heart failure
15      or ischemic heart disease.
16           Morris and Naumova (1998) investigated joint effects of short-term  ambient CO exposure
17      and ambient temperature on daily hospital admissions for CHF in people >64 years of age, in
18      Cook County (Chicago), IL, from 1986 through 1989. Ambient air pollution data were taken
19      from EPA's Aerometric Information Retrieval System database.  Data were analyzed with
20      general linear models (GLM) and general additive models (GAM). Pollutant metrics assessed in
21      analytical models were daily maximum hourly levels of CO, NO2, SO2, and O3, as well as 24-h
22      average PM10. For each day of the study, gaseous pollutant measurements were averaged across
23      Cook County's eight monitoring stations, six of which were in downtown Chicago. Only one
24      station collected daily PM10 samples (on 80% of study days).  In addition to these pollutant
25      variables, models included variables for daily maximum hourly temperature, day of week, month
26      of year, and year of study.  In single-pollutant GLMs, the level of each measured pollutant except
27      O3 was associated positively and statistically significantly (p = 0.05) with  admissions on the same
28      day. In a GLM that included all pollutants, only CO was associated significantly with
29      admissions. In this model, the RR for CHF admission at the 75th percentile of maximum hourly
30      CO concentration was 1.08 (95% CI1.03 to 1.12), compared with RR = 1 at CO concentration


        February 15,  1999                           6-9         DRAFT-DO NOT QUOTE OR CITE

-------
 1      of zero. Associations of CO and other pollutants with admissions were strongest with a lag time
 2      of 0 days and weakened rapidly with successively longer lag times.
 3           Effects of ambient CO on CHF admissions at different temperatures were analyzed in three
 4      ways: (1) inclusion of a CO-temperature interaction term in GLM; (2) simultaneous inclusion of
 5      a CO term and a temperature term in GAM, generating an additive CO-temperature effects
 6      surface; and (3) analysis with stratification on daily maximum ambient temperature (<40,
 7      40 to 75, and >75 °F).  Effects of CO on CHF admissions consistently were associated inversely
 8      with temperature (stronger effects at lower temperatures).  For example, in a multipollutant
 9      GLM, RRs for CFIF admissions at the 75th percentile of maximum hourly CO concentration
10      were 1.09(1.01 to 1.18), 1.07(1.01 to 1.13), and 1.01 (0.92 to 1.11) when maximum temperature
11      was <40, 40 to 75, and  >75 °F, respectively.
12           The authors pointed out that fixed-site CO measurements give inexact estimates of
13      individual subjects' total CO exposures and could well underestimate them during cold weather.
14      Thus, a given ambient CO level could reflect higher total CO exposure in cold weather than in
15      warm weather.  Also, a given day-to-day difference in ambient CO level could reflect a greater
16      day-to-day difference in total CO exposure in cold weather than in warm weather. The authors
17      also hypothesized, plausibly, that persons with CFIF may be unusually susceptible to CO, and
18      that cardiovascular and other stresses imposed by cold weather may heighten this susceptibility.
19      In any event, this study adds to the epidemiologic database linking ambient CO exposure with
20      aggravation of preexisting CHF. It also lends considerable weight to the hypothesis that
21      CO-mediated aggravation of CHF is most likely to occur during cold weather.
22
23      6.2.1.3 Carbon Monoxide and Daily Mortality
24           Epidemiologic studies of the relationship between CO exposure and daily mortality are not
25      yet conclusive.  Early studies in Southern California (Goldsmith and Landaw, 1968; Cohen et al.,
26      1969; Hexter and Goldsmith, 1971) suggested an association between atmospheric levels of CO
27      and increased mortality from CVD, but potential confounders were not controlled thoroughly.
28      In contrast, a study in Baltimore, MD  (Kuller et al., 1975), showed no association between
29      ambient CO levels and CVD or sudden death. More recent time series studies in North and
30      South America and in Europe also have yielded mixed results in relating day-to-day variations in
31      CO levels with daily mortality. For example, no relationship was  found between CO and daily

        February 15, 1999                         6-10       DRAFT-DO NOT QUOTE OR CITE

-------
 1      mortality in Los Angeles, Chicago, or Philadelphia (Ito et al., 1995; Kinney et al., 1995; Ito and
 2      Thurston, 1996; Kelsall et al., 1997) after adjusting for particles (i.e., PM10), time trends, and
 3      weather.  Verhoeff et al. (1996) found no relationship between 24-h average CO concentrations
 4      and daily mortality in Amsterdam, with or without adjustment for PM10 and other pollutants.
 5      Sal diva et al. (1994, 1995) found no association between CO and daily mortality among children
 6      or the elderly in Sao Paulo, Brazil, after adjusting for nitrogen oxides and PM10, though Pereira
 7      et al. (1998) did observe a limited relationship of ambient CO concentration with intrauterine
 8      mortality. Interestingly, in the latter study, COHb levels in cord blood were correlated with
 9      short-term ambient CO levels, even though intrauterine mortality was associated somewhat more
10      strongly with other pollutants than with CO. At the same time, Pereira et al. (1998) is difficult to
11      interpret because the investigators assessed fetal loss occurring only after 28 weeks of gestation,
12      whereas the large majority of spontaneous abortions occur before that time.
13           Three other studies (Touloumi et al., 1994; Salinas and Vega, 1995; Wietlisbach et al.,
14      1996) showed small, statistically significant relationships between CO and daily mortality.
15      However, effects of other pollutants (e.g., total suspended particles [TSP],  SO2, NO2, black
16      smoke) and of meteorologic variables (e.g., temperature, relative humidity) were also significant.
17      Further research will be needed to determine whether low-level CO exposure actually is
18      increasing mortality, particularly in the elderly population, whether CO  is a surrogate marker for
19      some other mobile-source or combustion-related pollutant, or whether CO  is a surrogate for the
20      overall combustion-related or automotive pollution mix.
21           Touloumi et al. (1994) investigated air pollution and daily all-cause mortality in Athens
22      from 1984 through 1988. Daily mean pollution indicators for SO2, black smoke, and CO were
23      averaged over all available monitoring stations. Autoregressive models with log-transformed
24      daily mortality as the dependent variable were used to adjust for temperature, relative humidity,
25      year, season, day of week, and for serial correlations in mortality.  Separate models for log(SO2),
26      log(smoke), and log(CO) yielded statistically significant effects estimates (p < 0.001).  Air
27      pollution measurements lagged by 1 day were most strongly associated with daily mortality.
28      A multipollutant model showed that SO2 and smoke were independent predictors of mortality,
29      though to a lesser extent than temperature and relative humidity.  Addition of an independent
30      variable for CO concentration did not improve this model's ability to predict daily mortality,


        February 15, 1999                           6-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      suggesting that CO may be a surrogate marker for other mobile-source or combustion-related
 2      pollutants.
 3           In one of the European Air Pollution and Health—A European Approach studies in Athens,
 4      Touloumi et al. (1996) observed a distinct positive association of ambient CO levels with daily
 5      mortality.  Ambient CO concentrations were compiled from three fixed outdoor monitoring
 6      stations. Median, mean, and maximum 8-h CO levels were 6.1 mg/m3 (5.3 ppm), 6.6 mg/m3
 7      (5.8 ppm), and 24.9 mg/m3 (21.7 ppm), respectively. The relative risk for daily mortality of a
 8      10 mg/m3 (9 ppm) increase in the daily ambient air CO concentration was 1.10 (95% CI = 1.05 to
 9      1.15). This finding is of interest, but it has not yet been confirmed in other epidemiologic
10      studies. It may be explained by yet unknown health effects of low levels of CO, by the presence
11      of highly compromised susceptible groups in the population, or, again, by CO acting as a
12      surrogate for other combustion-generated air pollutants.
13           Salinas and Vega (1995) examined the effect of urban air pollution on daily mortality in
14      Metropolitan Santiago, Chile, from 1988 through 1991. Measurements of maximum 8-h average
15      CO; maximum hourly O3; daily mean SO2, PM10, and PM25; and meteorologic variables were
16      obtained from five monitoring stations. Total and respiratory disease-specific deaths were
17      assessed, calculating the risk of death by municipality and month of the year using age-adjusted
18      standardized mortality ratios and controlling for socioeconomic status. Daily death counts were
19      regressed on pollutant levels, using Poisson regression and controlling for temperature and
20      relative humidity.  A clear pattern in the geographic distribution of risk of death was found,  both
21      for total mortality and disease-specific mortality (e.g., pneumonia, chronic obstructive pulmonary
22      disease, asthma), regardless of socioeconomic and living  conditions.  The number of deaths was
23      significantly associated with humidity, CO,  and suspended particles and also was associated with
24      temperature when the model included all days with available data during the 4-year period.  The
25      associations remained significant for those days with PM2 5 levels below 150 |ig/m3.
26           Wietlisbach et al. (1996) assessed the  association between daily mortality and air pollution
27      in metropolitan Zurich, Basel, and Geneva,  Switzerland, from 1984 through 1989. Daily counts
28      were obtained for total mortality, mortality in people 65 years of age or older, and respiratory and
29      cardiovascular disease mortality. Daily measurements of weather variables and TSP, SO2, NO2,
30      CO, and O3 were obtained in each city. Poisson models were used to regress daily death counts
31      on pollutant levels, controlling for time trends, seasonal factors, and weather variables.

        February 15, 1999                         6-12         DRAFT-DO NOT QUOTE OR CITE

-------
 1      A positive, statistically significant association was found between daily mortality and TSP, SO2,
 2      andNO2.  The strongest associations were observed with 3-day moving averages. Somewhat
 3      weaker associations were observed in each city between mortality in people 65 years of age or
 4      older and measured average daily CO concentrations (mean = 1 to 2 mg/m3 [1 to 2 ppm],
 5      max = 5 to 8 mg/m3 [4 to 7 ppm]). Associations with O3 were weak and inconsistent.  When all
 6      pollutants were modeled simultaneously, the regression coefficients were unstable and not
 7      statistically significant.
 8           In two recent studies, Burnett and colleagues investigated associations of CO and other
 9      pollutants with daily nonaccidental mortality in Canada. In one study (Burnett et al., 1998a), the
10      investigators assessed the roles of average daily concentrations of ambient CO, other gaseous
11      pollutants, sulfates, TSP, COH, estimated PM2 5 and PM10, and meteorology in Toronto from
12      1980 through 1994. The time series was adjusted for long-term trends and temporal cycles.
13      Effects of several different exposure-to-mortality lags were explored, and the final choice of lags
14      was based on the Akaike Information Criterion.  A two-day moving average was selected as the
15      optimum lag for CO,  but not for all pollutants. Final models included same-day  dew point
16      temperature.  In single-pollutant models, ambient levels of all pollutants except O3 were
17      associated positively  and statistically significantly with daily mortality, and this association was
18      strongest for ambient CO.  Two-pollutant models also were constructed, each including CO and
19      one of the other pollutants. In these models, the magnitudes of relative risks for  CO differed
20      little from that in the  single-pollutant model for CO. In contrast, the relative risks for other
21      pollutants generally decreased appreciably. Also, the relative risks for CO remained statistically
22      significant (at p = 0.05) in all two-pollutant models. Although the relative risk of CO was
23      highest for deaths from cardiac causes, there was also a clear positive association of CO with
24      deaths from other causes.
25           Burnett et al.  (1998b) also examined associations of ambient levels of gaseous pollutants
26      (CO, NO2, O3, and  SO2) with daily nonaccidental mortality in 11 Canadian cities from
27      1980 through 1991. In single-pollutant models, relative risks of CO for mortality were more
28      consistent across cities than were relative risks of the other pollutants.  However, in
29      multipollutant models, CO-associated relative risks decreased substantially, and NO2 and SO2
30      appeared to explain much of the CO effect on mortality. The estimated percentage increase in
31      mortality risk attributable to combined exposure  to all four pollutants differed widely from city to

        February 15, 1999                          6-13         DRAFT-DO NOT QUOTE OR CITE

-------
 1      city, ranging from 3.6% in Edmonton and Windsor to 11.0% in Quebec. The authors reasoned
 2      that reductions in gaseous pollutant levels might be more effective than reductions in PM levels
 3      in lessening mortality rates.  It is difficult to interpret this study quantitatively, because direct
 4      measurements of PM and PM constituents were not included in the analyses.  At the same time,
 5      the results underscore the need for measurement and statistical treatment of a broad range of
 6      pollutants and for further systematic assessment and comparison of the public health importance
 7      of exposure to CO, other ambient gaseous pollutants, and PM.
 8
 9      6.2.2 Controlled Laboratory Studies
10           The most extensive human experimental studies on the cardiovascular effects of CO have
11      been those conducted in predominantly young, healthy, nonsmoking subjects  during exercise.
12      Previous assessments of these effects (U.S. Environmental Protection Agency, 1979,  1984, 1991;
13      Horvath, 1981; Shephard, 1983, 1984) have identified what appears to be a linear relationship
14      between the level of COHb in the blood and decrements in human exercise performance,
15      measured as maximal oxygen uptake. Short-term maximal exercise performance significantly
16      decreases at COHb levels ranging from 5 to 20% (Pirnay et al., 1971; Vogel and Gleser, 1972;
17      Ekblom and Huot, 1972; Weiser et al., 1978; Stewart et al., 1978; Klein et al., 1980; Koike and
18      Wasserman, 1992).  One study (Horvath et al., 1975) observed a marginal decrease in maximal
19      exercise performance at a COHb level as low as 4.3% COHb. Short-term maximal exercise
20      duration also has been shown to be significantly reduced at COHb levels ranging from 2.3 to
21      20% (Ekblom and Huot, 1972; Drinkwater et al., 1974; Raven et al., 1974a,b; Horvath et al.,
22      1975; Weiser et al.,  1978; Koike and Wasserman,  1992).  The observed decreases in maximal
23      exercise performance and duration, however, are so small that they are only of concern primarily
24      for competing athletes, rather than for healthy people conducting everyday activities at less than
25      maximal exercise levels. In fact, no significant effects on oxygen uptake or on exercise
26      ventilation and heart rate were reported during submaximal exercise at COHb saturations as high
27      as 15 to 20% (see Section 10.3.2 in U.S. Environmental Protection Agency, 1991), especially at
28      work rates  below the metabolic acidosis threshold  (Koike et al., 1991). Cigarette smoking has a
29      similar effect on cardiopulmonary response to exercise in nonathletic human subjects, indicating
30      a reduced ability for sustained work.

        February 15, 1999                          6-14        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Of greater concern at more typical ambient CO levels are certain cardiovascular effects
 2      during exercise that are likely to occur in a smaller, but sizeable, segment of the general
 3      population having a deficiency of blood supply (ischemia) to the heart muscle. This group of
 4      patients with coronary artery disease (CAD) and reproducible exercise-induced angina (chest
 5      pain) is regarded as the most sensitive risk group for CO-exposure effects. Several important
 6      studies (Anderson et al., 1973; Sheps et al., 1987; Adams et al., 1988; Kleinman et al., 1989;
 7      Allred et al., 1989a,b, 1991) have provided the cardiovascular database for CO in CAD patients.
 8      In these studies, discussed in detail in the previous document (see Section 10.3.2 in U.S.
 9      Environmental Protection Agency, 1991), significant ischemia was measured subjectively by the
10      time of exercise required for the development of angina (time of onset of angina) and objectively
11      by the time required to demonstrate a 1-mm  change  in the ST segment of the electrocardiogram.
12      Adverse effects were found with postexposure COHb levels as low as 3 to 6% when compared
13      on the basis of optical measurements (Figure 6-1). This represents incremental increases of
14      1.5 to 4.4% COHb from preexposure baseline levels. Effects on silent ischemia episodes (no
15      chest pain), which represent the maj ority of episodes in these patients, have not been studied.
16           Only one new study has become available since publication of the 1991 document.  As part
17      of an investigation of CO exposure at high altitude,  17 men with documented CAD and stable
18      angina performed exercise stress tests after random 2-h exposures to either clean air or 100 ppm
19      CO  at sea level (Kleinman et al., 1998; Leaf and Kleinman,  1996a; Kleinman and Leaf, 1991).
20      The methods used were similar to those previously reported by Kleinman et al. (1989). Group
21      mean COHb levels measured by CO-Oximetry (CO-Ox)were 0.6 ± 0.3 (SD)% and
22      3.9 ± 0.5 (SD)% for clean air and CO exposures, respectively. Repeated measures analysis of
23      variance for a subgroup (number =13) with  angina on all test days demonstrated a statistically
24      significant (p < 0.05) decrease of 9.1 ± 0.6% in the time to onset of angina (from 5.94 to
25      5.40 min) during exercise after exposure to CO. The results are in good agreement with those
26      observed in the previously reported studies (see Figure 6-1).  There was no statistically
27      significant effect on ST segment change, on  the duration of angina, or on  hemodynamic factors
28      such as blood pressure and heart rate.
29           Despite clearly demonstrable effects of low-level CO exposure in patients with ischemic
30      heart disease, the adverse health consequences of these types of effects are very  difficult to
31      predict in the at-risk population of individuals with heart disease. There is a wide distribution of

        February 15, 1999                         6-15        DRAFT-DO NOT QUOTE OR CITE

-------
CO
c
'en
c

CD
o
Ll
CD

30^

25-

20-


15-
10-
5-
0-


-5-
I U
i i i i


T T
Anderson et al.
• (1973) ^
T

L L
Kleinman T ^SS^?- I
etal • T Allred et al. (1989a,b, •
,iqom T> (1989a,b, 1991) J_ Adamsetal.
^ ; J_ -L 1991) ¥ (1988)
1 i Sheps et al.
(1987)

_L
I I I I
                      246
                                  Percent COHb by Optical Methods


       Figure 6-1.  The effect of CO exposure on time to onset of angina. For comparison across
                  studies, data are presented as mean percent differences between air- and
                  CO-exposure days for individual subjects calculated from each study. Bars
                  indicate calculated standard errors of the mean. The COHb levels were
                  measured at the end of exposure; however, because of protocol differences
                  among studies and lack of precision in optical measurements of COHb,
                  comparisons must be interpreted with caution.

       Source: Modified from U.S. Environmental Protection Agency (1991); Allred et al. (1989b,1991).
1     professional judgments on the clinical significance of small performance decrements occurring

2     with the levels of exertion and CO exposure defined in the studies noted above. The decrements

3     in performance that have been described at the lowest levels (<3% COHb) are in the range of

4     reproducibility of the test and may not be alarming to some physicians.  On the other hand, the

5     consistency of the responses in time to onset of angina across the studies and the dose-response

6     relationship described by Allred et al. (1989a,b, 1991) between COHb and time to ST segment

7     changes strengthen the argument in the minds of other physicians that, although small, the effects
      February 15, 1999
6-16
DRAFT-DO NOT QUOTE OR CITE

-------
 1      could limit the activity of these individuals and affect their quality of life. In addition, it has been
 2      argued by Bassan (1990) that 58% of cardiologists believe recurrent episodes of exertional
 3      angina are associated with a substantial risk of precipitating a myocardial infarction (heart
 4      attack), a fatal arrhythmia (abnormal heart rhythm), or slight but cumulative myocardial damage.
 5          Exposures to low levels of CO resulting in 5 to 20% COHb do not produce significant
 6      changes in cardiac rhythm or conduction during rest or exercise in healthy humans (Davies and
 7      Smith, 1980; Kizakevich et al., 1994).  Effects of CO on resting and exercise-induced ventricular
 8      arrhythmia in patients with CAD are dependent on their clinical status. Hinderliter et al. (1989)
 9      reported no effects of 4 and 6% COHb in patients with ischemic heart disease who did not have
10      chronic arrhythmia (ectopy) during baseline monitoring.  In more severely compromised
11      individuals with higher levels of baseline ectopy, exposures to CO that produce 6% COHb have
12      been shown to significantly increase the number and complexity of arrhythmias (Sheps et al.,
13      1990), but not at lower COHb levels (Sheps et al., 1990,  1991; Chaitman et al., 1992; Dahms
14      et al., 1993). This finding, combined with epidemiologic evidence of CO-related morbidity and
15      mortality noted above, and the morbidity and mortality studies of workers who are routinely
16      exposed to combustion products (e.g., Stern et al., 1981,  1988; Edling and Axelson, 1984;
17      Sardinas et al., 1986; Michaels and Zoloth, 1991; Koskela, 1994; Melius, 1995; Strom et al.,
18      1995)  suggest that CO exposure may provide an increased risk of hospitalization or death in
19      patients with more severe heart disease.
20          There also is evidence from experimental studies with laboratory animals that CO can
21      adversely affect  the cardiovascular system.  The lowest-observed-effect level (LOEL) varies,
22      depending on the exposure regime used and species tested (see Table 6-1). Results from animal
23      studies (reviewed in U.S. Environmental Protection Agency, 1979, 1991; Turino, 1981;
24      McGrath, 1982;  Penney, 1988, 1996a) suggest that inhaled CO can cause disturbances in cardiac
25      rhythm and conduction in healthy and cardiac-impaired animals that are consistent with the
26      human data.  Results from animal studies (U.S. Environmental Protection Agency, 1991) also
27      indicate that inhaled CO can increase hemoglobin concentration and hematocrit ratio, probably
28      representing compensation for the reduction in oxygen transport caused by CO.  At high CO
29      concentrations, excessive increases in hemoglobin and hematocrit may impose an additional
30      workload on the heart and compromise blood flow to the tissues.
31

        February 15, 1999                         6-17        DRAFT-DO NOT QUOTE OR CITE

-------
              TABLE 6-1.  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
12.0
7.5
9.3
20.0
25.0

6.0

60.0

Duration

6 weeks
30 days
30 min
45 days
10 weeks
2h

21 days

15 min

Species

Dog
Rat
Rat
Rat
Rabbit
Rat

Rat

Rat

Reference

Preziosi et al. (1970)
Penney etal. (1974)
Kantenetal. (1983)
Penney etal. (1974)
Davies etal. (1976)
Merigan and
Mclntire (1976)
Prigge and
Hochrainer(1977)
Niden and Schulz
(1965)
 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.
       February 15, 1999
6-18
DRAFT-DO NOT QUOTE OR CITE

-------
 1     6.3   CENTRAL NERVOUS SYSTEM AND BEHAVIORAL EFFECTS
 2     6.3.1  Brain Oxygen Metabolism
 3     6.3.1.1 Whole Brain
 4          It has been documented amply in the literature that, as COHb is formed, there is
 5     vasodilation in the brain (and increased blood supply) in such proportion as to keep the supply of
 6     oxygen (O2) to the brain constant (Helfaer and Traystman, 1996; U.S. Environmental Protection
 7     Agency, 1991). The increased blood flow is sufficient to compensate not only for the oxygen
 8     supply decrease caused by reduced arterial O2 content (CaO2) but is also sufficient to compensate
 9     for the increased difficulty of extraction of O2 because of the shifted oxyhemoglobin dissociation
10     curve.  This compensatory vasodilation appears to be effective from low levels to very high
11     levels of COHb (at least up to 60%) and is similar in the fetus, neonate, and healthy adult.
12          Despite the compensatory regulation of O2 supply to the brain, it appears that
13     O2 consumption, measured as the cerebral metabolic rate for O2 (CMRO2), is reduced as COHb
14     rises. The reason for this is unclear, but the fact is well documented (Doblar et al., 1977; Jones
15     and Traystman, 1984; U.S.  Environmental Protection Agency, 1991; Langston et al.,  1996). The
16     amount of reduction in CMRO2 as a function of COHb can be shown by combining the
17     information of Doblar et al. (1977) from goats and Langston et al. (1996)  from sheep into one
18     graph (see Figure 6-2). Although information from Jones and Traystman  (1984) and associated
19     studies was expressed as a function of CaO2, not COHb, and was difficult to incorporate into
20     Figure 6-2 and the associated analysis, their data corroborate those of the  other workers.
21          From Figure 6-2, it may be seen that the CMRO2 does not decrease to 90% of baseline until
22     -27% COHb (95% confidence limits were ~2\ to 32% COHb).  The data from sheep and goats
23     agreed with the results of Paulson et al. (1973), who reported that the mean human CMRO2 did
24     not decrease significantly, even for COHb up to 20%.  Because Paulson et al. (1973)  did not
25     report the value of their means, it was not possible to include their results as data points on
26     Figure 6-2.
27
28     6.3.1.2 Subregions of the  Brain
29          There are a number of reports of the blood-flow response to COHb of subregions of the
30     brain (U.S. Environmental Protection Agency, 1991).  The results generally demonstrate that

       February 15, 1999                         6-19        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 6-2. 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.
February 15, 1999
6-20
DRAFT-DO NOT QUOTE OR CITE

-------
 1      some areas of the brain have less baseline blood flow than others, and that the COHb-
 2      compensatory increase in blood flow is not the same for all areas. Generally, however, the
 3      percent increases over baseline are nearly the same for all areas except the neurohypophysis
 4      (Hanley et al., 1986). It is important to note that the latter area serves homoeostatic and not
 5      ongoing behavioral functions. Thus, it would appear that the subregions of the brain have
 6      compensatory increased blood flow in the presence of COHb similar to the whole brain.  To be
 7      sure, all possible regions of the brain have not been tested, but no evidence to indicate otherwise
 8      has been found.
 9          Work by Sinha et al. (1991), measuring regional capillary perfusion and blood flow in the
10      presence of COHb elevation, indicates that the problem of compensation for COHb-reduced
11      CaO2 is not as simple as indicated above. Blood flow was measured using  radiolabeled dye and
12      capillary morphology was measured by fluorescence microscopy. With these methods, there
13      appeared to be an increase in the number of perfused capillaries and in the amount of blood flow
14      as COHb increased.  Thus, not only may more blood be delivered, but increased capillary
15      perfusion would decrease the diffusion distance to the tissue.
16          Presumably the compensatory mechanisms in subareas of the brain would work in a
17      manner similar to that of the whole brain and thus would show similar decreases in CMRO2 as
18      COHb increases.  No corroborative studies, however, have been reported in the literature.
19          Better and more detailed documentation of regional CMRO2in humans as well as other
20      species seems appropriate, but does not have high priority because not much evidence exists to
21      suggest that the results would differ from whole-brain results. It appears that what is  needed is
22      not more descriptive work, but an effort should be made to understand the mechanism by which
23      COHb elevation reduces CMRO2.  Furthermore, information is needed about brain conditions
24      under which brain compensatory mechanisms might be impaired (e.g., injury, inflammation,
25      ailments associated with aging and co-exposure to other pollutants). If such information were
26      available, specific theoretical (biologically based) predictions could be made, and behavioral
27      experiments designed to test them.
28
29      6.3.2  Behavioral Effects of Carbon Monoxide
30          The effects of CO on behavior, especially the ability to perform certain time discrimination
31      tasks, provided the scientific basis for the first air quality standard in 1971 (see Section 1.2).
        February 15,  1999                         6-21        DRAFT-DO NOT QUOTE OR CITE

-------
 1      As further research data became available, however, the results on human behavior at low levels
 2      of CO exposure (<5% COHb) were called into question and subsequently dismissed as the basis
 3      for the standard (U.S. Environmental Protection Agency, 1979).  After reviewing available
 4      studies, the previous criteria document (U.S. Environmental Protection Agency, 1991) concluded
 5      that effects on behavior were demonstrated unambiguously in both humans and laboratory
 6      animals at COHb elevations above 20%. Below this level, the results were less consistent.  The
 7      document also concluded, however, that it seems unwise to ignore the historical evidence in
 8      favor of effects on human performance at COHb levels between 5 and 20% (e.g., Horvath et al.,
 9      1971;  Fodor and Winneke,  1972; Putz et al., 1976, 1979; Benignus et al., 1987).  Even if
10      behavioral effects are small or occasional, they may be important to the performance of critical
11      tasks.
12          Behavioral experiments with the effects of elevated COHb  frequently have been marred by
13      methodological problems. In particular, experiments employing  single-blind designs were shown
14      to be 2.5 times as likely to find significant results as similar studies employing double-blind
15      methods (Benignus, 1993, 1996). This problem was noted previously, and reports of findings of
16      behavioral effects of CO were summarized with respect to whether a double-blind procedure had
17      been followed (see Table 10-25 of U.S. Environmental Protection Agency, 1991). From this
18      summary, it was concluded  that, at most, there was credible evidence for effects on only three
19      (somewhat artificially defined) categories of behavior: (1) tracking, (2) vigilance, and
20      (3) continuous performance. Even within these categories, considering only double-blind
21      studies, it was noted that less than 50% of all studies found significant effects. Furthermore,
22      most of the double-blind experiments reporting significant results were unreplicable.
23          Benignus (1994) performed extensive meta-analyses of the CO-behavioral literature.
24      Because single-blind studies were likely to include many type I errors (results erroneously
25      declared significant), only double-blind human CO studies were included. In this report, two
26      dose-effect curves were estimated from the literature by converting all behavioral endpoints to
27      percent of baseline. A dose-effect curve for COHb and behavior was estimated from
28      experiments with rats and was corrected for the effects of hypothermia. Another dose-effect
29      curve was estimated from the human literature on hypoxic hypoxia, which was converted to
30      equivalent COHb via equal  arterial oxygen contents and corrected for effects of hypocapnia.
31      These two curves virtually overlapped each other. Human data points from CO-behavior

        February  15,  1999                         6-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      experiments then were plotted onto the curve fitted to the rat CO data (no curve was fitted to
 2      human data because of the small effect sizes and small COHb levels). The conclusion from this
 3      meta-analysis was that human behavioral impairments of 10% (ED-10) should not be expected
 4      until COHb exceeds 20%.
 5          Data for the rat studies from Benignus (1994) were refitted for the present purposes using
 6      the same dose-effect model (a logit) as for the CMRO2 data above (originally a different function
 7      was used). Data from all available double-blind human studies also were converted to percent of
 8      baseline and plotted on the same graph as the rat data and the logit curve fitted to the rat data
 9      (Figure 6-3).  With the logit function, it is estimated that a 10% decrement should be produced in
10      rats by ca. 25% COHb (95% confidence limits of -20 to 30%). The human data plotted in
11      Figure 6-3 (asterisks) seem, as a group, not to have a dose-related trend of decrements, and it
12      could be argued reasonably that effects in humans cannot be shown to differ from those in rats.
13      Some  of the human data, however, at low levels of COHb (4 to 10%) do appear below baseline
14      and were declared statistically significant by the authors of the original reports.
15          The low-COHb significant results plotted in Figure 6-3 were invariably reported in studies
16      in which only a few levels of COHb were evaluated.  Studies in which more and higher COHb
17      levels  were tested invariably did not find statistically significant effects, even at much higher
18      levels. Furthermore, for every study reporting low-COHb level impairments, other studies failed
19      to replicate the findings, or highly similar studies failed to find effects.
20          In summary, no reliable evidence demonstrating decrements in neural  or behavioral
21      function in healthy young adult humans has been reported for COHb levels below 20%, and even
22      these studies  are untested by replications. The low-COHb behavioral effects that have
23      sometimes been reported cannot be taken at face value because they are not reliably repeatable,
24      and they do not fit into a wider range, dose-effect pattern reported in other studies. It is more
25      reasonable to conclude that no statistically detectable behavioral impairments occur until COHb
26      exceeds 20 to 30%. The conclusion, based on behavioral evidence alone, is bolstered by the
27      findings that whole-brain CMRO2 is not reduced by a similar amount until COHb rises to  21 to
28      32%.  Because a dose-effect curve has been fitted, any level of effect may be considered
29      (e.g., ED-5 or ED-20). The interpolation of a curve to an ED-5 point would imply that the COHb
30      levels  for such an effect size would be 15 to 26%. Such an interpolation is more speculative than
31      an ED-10, however, because the experimental verification would be difficult, requiring large

        February 15,  1999                         6-23        DRAFT-DO NOT QUOTE OR CITE

-------
 0)
      0.0-
                                      COHb(%)

Figure 6-3. The relationship between COHb and behavior. Open circles are the
           hypothermia-corrected means from four studies of effects of COHb in rats (see
           Benignus, 1994). The solid line is a best-fit logit curve to the rat data. The
           dashed lines are 95% confidence limits. Asterisks are data from double-blind
           human studies (see Benignus, 1994) that were plotted in the figure but were
           not used to fit the logit function.
February 15, 1999
6-24
DRAFT-DO NOT QUOTE OR CITE

-------
 1      numbers of subjects and careful control of error variance.  Additionally, as interpolation
 2      approaches small effect sizes, the error possibility because of statistical model selection
 3      (threshold versus continuous) increases dramatically.
 4           Unless the effort is to find a behavioral paradigm that would yield replicable low-level
 5      COHb decrements in behavior that were part of a wider range, dose-effect curve, it would seem
 6      unfruitful to continue behavioral work. It is important, however, that any new behavioral
 7      experiments demonstrate a dose-effect curve that includes higher COHb levels. The lack of such
 8      dose-effect information within a study has contributed to the problems of interpreting published
 9      literature. Because of the difficulties in replication, it would seem desirable that behavioral
10      experiments be designed to verify prediction from biologically based information (e.g.,
11      mechanistic theory about CMRO2 and how it is affected by COHb elevation).
12           The behaviors implicated by the research findings involve such abilities as detection of
13      infrequent events (vigilance), hand-eye coordination (compensatory tracking), and other forms of
14      continuous performance (U.S. Environmental Protection Agency, 1991). Because of the
15      unreliability of the findings, however, it is questionable whether these behaviors should be cited
16      as effects.  Until reliable behavioral effects are demonstrated in  a dose-related manner,  it seems
17      premature to speculate about the kind of behavioral effects and thus lend credence to unreliable
18      findings.
19           Because COHb elevates brain blood flow, it has the possibility of altering the delivery of
20      other toxicants to the brain or altering the biotransformation or elimination of toxicants (e.g., Doi
21      and Tanaka, 1984; Kim and Carlson, 1983; Roth and Rubin, 1976a,b). In combination with
22      exercise or hypoxic hypoxia, the interactions would become even more complex.  Disease and
23      ailments associated with aging concomitant with all of the above could be important.
24      Interactions such as these are understood from physiological theory and can  be given quantitative
25      estimates by the use of physiological simulation using whole-body physiological simulators.
26
27
28      6.4  DEVELOPMENTAL TOXICITY
29           An issue directly relevant to the protection of public health is the potential effect of CO on
30      growth and function of the developing fetus,  infant, and child. Results obtained from new
31      research on this  specific outcome of CO exposure (e.g., Carratu et al., 1993, 1996; Di Giovanni
        February 15, 1999                          6-25         DRAFT-DO NOT QUOTE OR CITE

-------
 1      et al., 1993; De Salvia et al., 1995; De Luca et al., 1996) have not changed the conclusions
 2      presented in Section 10.5 of the previous criteria document (U.S. Environmental Protection
 3      Agency, 1991). From all of the laboratory animal studies, it is clear that severe, acute CO
 4      poisoning can be fetotoxic, although specification of maternal and fetal COHb levels is difficult
 5      because such exposures rarely involve the achievement of steady-state COHb levels or permit
 6      careful  and rapid determination of COHb levels. Available data (reviewed in U.S.
 7      Environmental Protection Agency, 1991; Annau and Fechter, 1994; Carratu et al., 1995; Penney,
 8      1996b)  provide strong evidence that maternal CO exposures of 150 to 200 ppm, leading to
 9      approximately 15 to 25% COHb, produce reductions in birth weight, cardiomegaly, delays in
10      behavioral development, and disruption of cognitive function in laboratory animals of several
11      species. Isolated experiments (Prigge and Hochrainer, 1977; Abbatiello and Mohrmann, 1979;
12      Singh, 1986) suggest that some of these effects may be present at concentrations as low as 60 to
13      65 ppm (approximately 6 to 11% COHb) maintained throughout gestation (see Table 6-1).
14      Studies relating human CO exposure from ambient sources or cigarette smoking to reduced birth
15      weight  (e.g., Martin and Bracken, 1986; Rubin et al., 1986; Alderman et al., 1987; Wouters
16      et al., 1987; Brooke et al., 1989; Spitzer et al., 1990; Wen et al., 1990; Peacock et al., 1991a;
17      Zaren et al., 1996; Jedrychowski and Flak, 1996; Seeker-Walker et al., 1997) are of concern
18      because of the risk for developmental disorders (Olds et al.,  1994a,b; Olds, 1997); however,
19      many of these studies have not considered all sources of CO exposure, other pollutants (Wang
20      et al., 1997), or other risk factors during gestation (Peacock et al., 1991b; Luke, 1994; Robkin,
21      1997).
22           Results from laboratory animal studies indicate that CO, by itself, should not have much of
23      an effect on the developing fetus until later in gestation when the embryo is much larger and
24      more dependent on transport of oxygen by red blood cells.  In addition, results from a
25      multicenter, prospective study (Koren et al., 1991) of fetal outcome following mild to moderate
26      accidental  CO poisoning in pregnancy suggests that hypoxemia associated with measured COHb
27      saturations of up to 18% (or even higher estimated levels) does not impair the growth potential of
28      the fetus when pregnancy continues normally. Therefore, it is very unlikely that ambient levels
29      of CO typically encountered by pregnant women would cause increased fetal risk. It is
30      necessary,  however, to consider the combined effects of CO with the other common risk factors


        February 15, 1999                          6-26        DRAFT-DO NOT QUOTE OR CITE

-------
 1      that may cause adverse fetal outcome (e.g., tobacco and alcohol consumption, genetic
 2      background, maternal general health and obstetric history).
 3           One of the more important determinants of the course and outcome of pregnancy that was
 4      not previously discussed is maternal-fetal nutrition (Luke, 1994). Laboratory animal studies
 5      conducted to determine the combined effect of gestational CO exposure and nutritional
 6      deficiency suggest that  CO has a greater effect on the fetus in protein-deficient mice (Singh and
 7      Moore-Cheatum, 1993; Singh et al.,  1993). Reductions in the rate of pregnancy, lower fetal
 8      weights, and increased  fetal malformations were reported at CO concentrations as low as 65 ppm
 9      maintained for 6 h per day during the first trimester of pregnancy. Previous evidence of the
10      fetotoxic and teratogenic effects of CO in laboratory animals (U.S. Environmental Protection
11      Agency, 1991) came largely from high levels of exposure (i.e., in the range of 500 ppm for
12      rodents).
13           There are studies  (e.g., Schoendorf and Kiely, 1992; Scragg et al., 1993; Mitchell et al.,
14      1993; Klonoff-Cohen et al., 1995; Blair et al., 1996; Hutter and Blair, 1996; MacDorman et al.,
15      1997) linking maternal  cigarette smoking with sudden infant death syndrome (SIDS), but the role
16      of CO is uncertain, especially in relation to other known risk factors for SIDS, such as
17      developmental abnormalities (Schwartz et al., 1998), prone sleeping (Kahn et al., 1993; Franco
18      et al., 1996), overheating (Douglas et al., 1996), and soft bedding (Ponsonby et al., 1993; Kemp
19      et al., 1998). Data from human populations (Hoppenbrouwers et al., 1981) suggesting a link
20      between ambient CO exposures and  SIDS are weak, but further study should be encouraged.
21      Children may experience neurological symptoms such as dizziness or fainting after an acute
22      episode of CO poisoning (>15% COHb) or, in some cases, neurological impairment may develop
23      days to weeks  after very high exposures (Crocker and Walker, 1985). Human data from these
24      cases of accidental high CO exposures  are difficult to use in identifying a LOEL for CO because
25      of the small number of cases reviewed  and problems in  documenting exposure levels. However,
26      such  data, if systematically gathered  and reported, could be useful in identifying possible ages of
27      special sensitivity to CO and co-factors or other risk factors that might identify sensitive
28      subpopulations.
29
        February 15, 1999                         6-27        DRAFT-DO NOT QUOTE OR CITE

-------
 1     6.5  ACUTE PULMONARY EFFECTS
 2          It is unlikely that CO has any direct effects on lung tissue, except for extremely high
 3     concentrations that can cause cell damage and edema (Niden and Schulz, 1965; Fein et al., 1980;
 4     Burns et al., 1986). No new information has been published in the literature to change this
 5     conclusion drawn from Section 10.2 of the previous criteria document (U.S. Environmental
 6     Protection Agency, 1991). Experimental studies on the effects of CO exposures producing
 7     COHb saturations up to 39% failed to find any consistent effects on pulmonary cells and tissue or
 8     on the vasculature of the lung (Fisher et al., 1969;  Weissbecker et al., 1969; Hugod, 1980; Chen
 9     et al., 1982). Human studies on the pulmonary function effects of CO are complicated by the
10     lack of adequate exposure information, the small number of subjects studied, and the short
11     exposures explored. Decrements in lung function have been observed with increasing severity of
12     CO poisoning (Kolarzyk, 1994a,b, 1995).  For example, occupational or accidental exposure to
13     the products of combustion and pyrolysis, particularly indoors, may lead to acute decrements in
14     lung function if COHb levels are greater than 17% (Sheppard et al., 1986) but not  at saturations
15     less than 2% (Cooper and Alberti, 1984; Hagberg  et al., 1985; Evans et al.,  1988).  It is difficult,
16     however, to separate the  potential effects of CO from the effects of other respiratory irritants in
17     smoke and exhaust. Community population studies on CO in ambient air have not found strong
18     relationships with pulmonary function, symptomatology, and disease (Lutz, 1983;  Robertson and
19     Lebowitz, 1984; Lebowitz et al., 1987).
20
21
22     6.6  OTHER SYSTEMIC EFFECTS OF CARBON MONOXIDE
23          Laboratory animal  studies (reviewed in Section 10.6 of U.S. Environmental Protection
24     Agency, 1991) suggest that enzyme metabolism and the P-450-mediated metabolism of
25     xenobiotic compounds may be affected by CO exposure (e.g., Montgomery and Rubin, 1971;
26     Pankow et al.,  1974; Roth and Rubin, 1976a,b,c).  Most of the authors of these studies have
27     concluded, however, that effects on metabolism at low COHb levels (< 15%) are attributable
28     entirely to tissue hypoxia produced by increased levels of COHb because the effects are no
29     greater than those produced by comparable levels  of hypoxia produced by insufficient oxygen
30     delivery. At higher levels of exposure, where COHb concentrations exceed  15 to 20%, there may

       February 15, 1999                         6-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1     be direct inhibitory effects of CO on the activity of mixed-function oxidases, but more basic
 2     research is needed. The decreases in xenobiotic metabolism shown with CO exposure may be
 3     important to individuals receiving drug treatment.
 4          Inhalation of high levels of CO, leading to COHb concentrations greater than 10 to 15%,
 5     have been reported to cause a number of other systemic effects in laboratory animals and effects
 6     in humans suffering from acute CO poisoning.  Tissues of highly active oxygen metabolism, such
 7     as heart, brain, liver, kidney, and muscle, may be particularly sensitive to CO poisoning. The
 8     impairment of function in the heart and brain caused by CO exposure is well known and has been
 9     described above. Other systemic effects of CO poisoning are not as well known and are therefore
10     less certain.  There are reports of effects on liver (Katsumata et al., 1980), kidney (Kuska et al.,
11     1980), bone (Zebro et al., 1983), and immune capacity in the lung and spleen (Snella and
12     Rylander, 1979). It generally is agreed that these effects are caused by the severe tissue damage
13     occurring during acute CO poisoning resulting from one or more of the following:  ischemia
14     resulting from the formation of COHb, inhibition of oxygen release from oxyhemoglobin,
15     inhibition of cellular cytochrome function (e.g., cytochrome oxidases), and metabolic acidosis.
16
17
18     6.7  PHYSIOLOGIC RESPONSES TO CARBON MONOXIDE
19           EXPOSURE
20          The only evidence for short- or long-term compensation to increased COHb levels in the
21     blood is indirect. Experimental animal data (reviewed in Section 10.7 of U.S. Environmental
22     Protection Agency, 1991) indicate that incremental increases in COHb produce physiological
23     responses that tend to offset the deleterious effects of CO exposure on oxygen delivery to the
24     tissues. Experimental human data (presented in a report by Kizakevich et al., 1994) indicate that
25     compensatory cardiovascular responses to submaximal upper- and lower-body exercise
26     (e.g., increased heart rate, cardiac contractility, cardiac output) occur after CO exposures.  These
27     changes were highly significant for exposures attaining 20% COHb. Other compensatory
28     responses are increased coronary blood flow, cerebral blood flow, hemoglobin (through increased
29     hemopoiesis), and oxygen consumption in muscle.
30          Short-term compensatory responses in blood flow or oxygen consumption may not be
31     complete or may even be absent in certain persons. For example, from the laboratory animal

       February 15, 1999                         6-29        DRAFT-DO NOT QUOTE OR CITE

-------
 1     studies, it is known that coronary blood flow is increased with COHb, and, from human clinical
 2     studies, it is known that subjects with ischemic heart disease respond to the lowest levels of
 3     COHb (6% or less). The implication is that, in some cases of cardiac impairment, the short-term
 4     compensatory mechanism is impaired.
 5          From neurobehavorial studies (see Section 6.3.2 of the present document), it is apparent
 6     that decrements resulting from CO exposure have not been consistent in all subjects, even in the
 7     same studies, and have not demonstrated a dose-response relationship with increasing COHb
 8     levels. The implication from this data suggests there may be some threshold or time lag in a
 9     compensatory mechanism such as increased blood flow. Without direct physiological evidence
10     in either laboratory animals or humans, this concept can be only hypothesized.
11          The mechanism by which long-term adaptation may occur, if it can be demonstrated in
12     humans, is assumed to be increased hemoglobin concentration via an increase in hemopoiesis.
13     This alteration in hemoglobin production has been demonstrated repeatedly in laboratory animal
14     studies, but no recent studies have been conducted that indicate the occurrence of some
15     adaptational benefit. Even if the hemoglobin increase is a signature of adaptation, it has not been
16     demonstrated at low ambient concentrations of CO.
17
18
19     6.8  COMBINED EXPOSURE OF CARBON MONOXIDE WITH OTHER
20           POLLUTANTS,  DRUGS, AND ENVIRONMENTAL FACTORS
21     6.8.1  High-Altitude Effects
22          Although there are many studies comparing and contrasting the effects of inhaling CO with
23     those produced by short-term, high-altitude exposure, there are relatively few reports on the
24     combined effects of inhaling CO at high altitudes.  There are data (reviewed in Section 11.1 of
25     U.S. Environmental Protection Agency, 1991) to support the possibility that the effects of these
26     two hypoxic factor episodes are at least additive. Most of these early data were obtained at CO
27     concentrations too high to have much meaning for regulating the amount of CO in ambient air.
28     More recent studies by Kleinman et al. (1998) evaluated the combined effects of lower levels of
29     CO at high altitude. In general, the results confirm the additivity of hypoxic effects at a
30     simulated altitude of 2.1 km and CO exposures resulting in 4% COHb.

       February 15, 1999                         6-30       DRAFT-DO NOT QUOTE OR CITE

-------
 1           There are even fewer studies of the long-term effects of CO at high altitude.  These studies,
 2      identified in Table 11-2 of the previous criteria document (U.S. Environmental Protection
 3      Agency, 1991), indicate few changes at CO concentrations below 100 ppm and altitudes below
 4      4,572 m (15,000 ft).  The fetus may be particularly sensitive to the effects of CO at altitude, as is
 5      especially true with the high levels of CO associated with maternal smoking (Moore et al., 1982).
 6           The potential effects on human health of inhaling CO at high altitudes are complex (see
 7      Section 5.4.1) Whenever CO binds to hemoglobin (Hb), it reduces the amount of Hb available to
 8      carry oxygen.  People visiting high altitudes (where the partial pressure of oxygen in the
 9      atmosphere is lower) will experience reduced levels of oxygen in the blood (hypoxemia) because
10      of a relative hypoventilation that occurs, particularly during sleep.  Carbon monoxide, by binding
11      to Hb, intensifies the hypoxemia existing at high altitudes by further reducing transport  of oxygen
12      to the tissues.  In addition, COHb saturations are higher at altitude than at sea level because,
13      in part, of changes in elimination of endogenous CO and of more rapid uptake of exogenous CO
14      (McGrath, 1992; McGrath et al., 1991, 1993).  However, within hours of arrival at high altitude,
15      certain physiological adjustments begin to take place (Grover et al., 1986), and, over several
16      days, these mechanisms will operate to lessen the initial impact of atmospheric hypoxia.
17      Hemoconcentration occurs, and the increased Hb concentration offsets the decreased blood
18      oxygen saturation and restores  oxygen concentrations to former levels. Consequently, the simple
19      additive model of COHb and altitude hypoxemia may be valid only during early altitude
20      exposure. The new visitor to higher altitudes,  especially the elderly and those with CAD
21      (Kleinman et al., 1998; Leaf and Kleinman, 1996a), may be at greater risk from the added effects
22      of ambient CO than the adapted resident. The period of increased  risk probably is prolonged in
23      the elderly because adaptation to high altitude proceeds more slowly with increasing age (Dill
24      etal., 1985).
25
26      6.8.2 Interaction with Drugs
27           There remains little direct information on the possible enhancement of CO toxicity by
28      concomitant drug use or abuse; however, there are some data suggesting cause for concern.
29      There is some evidence that interactions of drug effects with CO exposure can occur in  both
30      directions, that is, CO toxicity may be enhanced by drug use, and the toxic or other effects of
31      drugs may be altered by CO exposure. Nearly all published data available on CO combinations
        February 15, 1999                         6-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1     with drugs concern psychoactive drugs (Montgomery and Rubin, 1971, 1973; McMillan and
 2     Miller, 1974; Medical College of Wisconsin, 1974; Pankow et al., 1974; Rockwell and
 3     Weir, 1975; Roth and Rubin, 1976a,b,c; Mitchell et al., 1978; Topping et al.,  1981; Kim and
 4     Carlson, 1983; Engen, 1986; Knisely et al., 1987, 1989). Descriptions of these studies were
 5     provided in Section 11.2 of the previous criteria document (U.S. Environmental Protection
 6     Agency, 1991). The following summary has been excerpted from the last review.
 7          The use and abuse of psychoactive drugs and alcohol are widespread. Because of the effect
 8     of CO on brain function, interactions between CO and psychoactive drugs could be anticipated.
 9     However, very little systematic research has addressed this question. In addition, very little of
10     the research that has been done has utilized models for expected effects from treatment
11     combinations. Thus, often it is not possible to assess whether the combined effects of drugs and
12     CO exposure are additive or differ from additivity. It is important to recognize that even additive
13     effects of combinations can be of clinical significance, especially when the individual is unaware
14     of the combined hazard. The greatest evidence for a potentially important interaction of CO
15     comes from studies with alcohol in both laboratory animals and humans, where at least  additive
16     effects have been obtained (Mitchell et al., 1978; Knisely et al., 1987, 1989).  The significance of
17     these effects is augmented by the probable high incidence of combined alcohol use and  CO
18     exposure in the population.
19          Besides interaction with psychoactive drugs, there is growing concern that prescribed
20     medications, especially nitric oxide blockers and calcium channel blockers, could interact with
21     CO. There are no known published data available, however, on CO combinations with  these
22     drugs.
23
24     6.8.3  Interaction with Other Air Pollutants and Environmental Factors
25          Much of the data concerning the combined effects of CO and other pollutants found in
26     ambient air are based on laboratory animal experiments that were discussed in Section 11.3 of
27     the previous criteria document (U.S. Environmental Protection Agency, 1991).  More recent
28     studies published since then have confirmed the conclusions made at that time and are included
29     here for completeness.  Only a few controlled-exposure studies of humans are available, and the
30     results were discussed in more detail in the previous  document.  These early studies in healthy
31     human subjects (Drinkwater et al., 1974; Raven et al., 1974a,b; Gliner et al., 1975; Hackney
       February 15, 1999                         6-32        DRAFT-DO NOT QUOTE OR CITE

-------
 1      et al., 1975a,b; DeLucia et al., 1983) on common air pollutants such as CO, NO2, O3, and
 2      peroxyacetylnitrate failed to show any interaction from combined exposure. The more recent
 3      epidemiology studies (e.g., Morris et al., 1995; Schwartz and Morris, 1995; Schwartz, 1997;
 4      Burnett et al., 1997) suggest an association between hospital admissions for cardiovascular
 5      disease and ambient exposure to multiple pollutants, including CO and PM. In animal studies,
 6      no interaction was observed following combined exposure of CO and common air pollutants
 7      such as NO2 and SO2 (Busey, 1972; Murray et al., 1978; Hugod,  1979).  However, an additive
 8      effect on learning behavior was observed following combined exposure of high levels of CO and
 9      NO (Groll-Knapp et al., 1988), and a synergistic dose effect (increased COHb) was observed
10      after combined exposure to CO and O3 (Murphy, 1964).
11          Toxicological interactions of combustion products, primarily CO, carbon dioxide, NO2, and
12      hydrogen cyanide (HCN), at levels typically produced by indoor  and outdoor fires, have  shown a
13      synergistic effect on mortality following CO plus CO2 exposure (Rodkey and Collison, 1979;
14      Levin et al.,  1987a) and CO plus NO2 exposure (Levin, 1996), and an additive effect with HCN
15      (Levin et al., 1987b). Additive effects on mortality also were observed when CO, HCN, and low
16      oxygen were combined; adding CO2 to this combination was  synergistic (Levin et al., 1988).
17          Finally, laboratory animal studies (Young et al., 1987; Yang et al., 1988; Fechter et al.,
18      1988, 1997;  Fechter,  1995; Gary et al., 1997) suggest that combinations of environmental factors
19      such as heat stress and noise may be important determinants of health effects occurring in
20      combination with CO exposure.  Of the effects described, one potentially most relevant to typical
21      human exposures is a greater decrement in the exercise performance seen when heat stress is
22      combined with 50 ppm CO (Drinkwater et al., 1974; Raven et al., 1974a,b; Gliner et al.,  1975).
23
24      6.8.4  Environmental Tobacco Smoke
25          Although tobacco smoke is another source of CO for smokers as well as nonsmokers, it is
26      also a source of other chemicals (e.g., nicotine, NO2, HCN, polyaromatic hydrocarbons [PAHs],
27      aldehydes, ketones) that could interact with environmental CO. Available  data suggest that some
28      of these components can affect the cardiovascular system. For example, nicotine clearly
29      aggravates the decrease in oxygen capacity induced by CO through an increase in the oxygen
30      demand of the heart (Khosla et al., 1994; Benowitz, 1997), and PAHs have been implicated in

        February 15, 1999                         6-33        DRAFT-DO NOT QUOTE OR CITE

-------
 1      atherosclerosis (Glantz and Parmley, 1991). Little is known, however, about the relative
 2      importance of CO compared with the other components of tobacco smoke.
 3           The association between active smoking and CVD is fully established (Surgeon General of
 4      the United States, 1983). Passive smoking exposes an individual to all components in the
 5      cigarette smoke, but the CO component dominates heavily because only 1% or less of the
 6      nicotine is absorbed from environmental tobacco smoke (ETS), compared with 100% in an active
 7      smoker (Wall et al., 1988; Jarvis, 1987). Therefore, exposure to ETS will be the closest to pure
 8      CO exposure, even if the resultant levels of COHb are low (about 1 to 2%) (Jarvis, 1987).
 9      The relationship between passive smoking and increased risk of CVD is controversial.  Early
10      studies on this relationship were reviewed in the 1986 report of the Surgeon General of the
11      United States (1986) and by the National Research Council (1986).  Since that time, the
12      epidemiological evidence linking passive smoking exposure to heart disease has expanded
13      rapidly. The available literature on the relationship between passive  exposure to ETS in the
14      home and the risk of cardiovascular-associated morbidity or mortality in the nonsmoking spouse
15      of a smoker consists of numerous published reports (e.g., Glantz and Parmley, 1991; Steenland,
16      1992; Wells, 1994; Kritz et al., 1995; LeVois and Layard, 1995; Steenland et al., 1996; Kawachi
17      et al., 1997; Howard et al., 1998). The data suggest an increase of approximately 1.3 in the
18      relative risk of CVD (95% CI of 1.2 to 1.4) is associated with prolonged exposure to ETS that
19      may be caused by any number of biochemical mechanisms, including greater platelet
20      aggregation, endothelial cell damage, reduced oxygen supply, greater oxygen demand, and the
21      direct effects of CO (Kalmaz et al., 1993; Zhu and Parmley, 1995; Weiss, 1996; Werner and
22      Pearson, 1998).
23
24
25      6.9   SUMMARY
26           As noted at the beginning of this chapter, the lethality of CO that results from exposure to
27      very high concentrations is well known to health professionals and the public.  Symptoms of
28      acute CO  poisoning cover a wide range, depending on severity of exposure: headache, dizziness,
29      weakness, nausea, vomiting, disorientation,  confusion, collapse, and coma.  Perhaps the most
30      insidious effects of CO poisoning are the delayed development of neurological impairment that
31      occur within 1 to 3 weeks after exposure and the neurobehavioral consequences, especially in
        February 15, 1999                          6-34       DRAFT-DO NOT QUOTE OR CITE

-------
 1      children.  Carbon monoxide poisoning during pregnancy results in high risk for the mother by
 2      increasing the short-term complication rate and, for the fetus, by causing reduced birth weight,
 3      heart and brain disorders, and possibly death.
 4           The effects of exposure to low concentrations, such as the levels found in ambient air, are
 5      far more subtle and considerably less threatening than those occurring in frank poisoning from
 6      high CO concentrations. Because the COHb level of the blood is the best indicator of potential
 7      health risk, symptoms of exposures to excessive ambient air levels of CO are described here in
 8      terms of associated COHb levels. The LOEL, however, depends on the method used for analysis
 9      of COHb. Gas chromatography (GC) is the method of choice for measuring COHb, particularly
10      at saturation levels <5%, because of the large variability and potential high bias of the optical
11      methods such as CO-Ox. The key human health effects most clearly demonstrated to be
12      associated with exposure to ambient CO are summarized in Table 6-2.
13           Maximal exercise duration and performance in healthy individuals have been shown to be
14      reduced at COHb levels of >2.3% and >4.3% (GC), respectively.  The decrements in
15      performance at these levels are small and likely to affect only competing athletes rather than
16      people engaged in everyday activities.  In fact, no effects were observed during submaximal
17      exercise in healthy individuals at COHb levels as high as 15 to 20%.
18           Adverse effects have been observed in individuals with CAD at 3 to 6% COHb by optical
19      methods of measurement.  At these levels, individuals with reproducible exercise-induced angina
20      (chest pain) are likely to experience a reduced capacity to exercise because of decreased time to
21      onset of angina. The indicators of myocardial ischemia during exercise, which is detectable by
22      electrocardiographic (ECG) changes (ST depression) and associated angina, were statistically
23      significant in one study at >2.4% COHb (GC) and showed a dose-response relationship with
24      increasing COHb.  An increase in the number and complexity of exercise-related arrhythmias
25      also has been observed at >6% COHb (CO-Ox) in some people with CAD and high levels of
26      baseline ectopy (chronic arrhythmia) that may present an increased risk of sudden death.
27           Results from the controlled laboratory studies and reports of increased morbidity and
28      mortality in workers routinely exposed to combustion products provide support for the recent
29      epidemiology studies suggesting that day-to-day variations in ambient CO concentrations are
30      related to cardiovascular hospital admissions and daily mortality, especially for individuals over


        February 15, 1999                         6-35        DRAFT-DO NOT QUOTE OR CITE

-------
                 TABLE 6-2. KEY HEALTH EFFECTS OF EXPOSURE TO AMBIENT
                                           CARBON MONOXIDE
         Target
         Organ                 Health Effectsa'b
                                            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 et al. (1988)
Kleinmanetal. (1989,
  1998)
Allredetal. (1989a,b;
  1991)

Sheps etal. (1990)
Schwartz and Morris
  (1995)
Morris etal. (1995)
Schwartz (1997)
Burnett etal. (1997)

Horvathetal. (1971)
Fodor and Winneke  (1972)
Putzetal. (1976, 1979)
Benignus etal. (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      65 years of age. The relative influences on these associations of ambient and nonambient CO

2      have not been determined.

3            Central nervous system effects, including reductions in hand-eye coordination (driving or

4      tracking) and in attention or vigilance, have been reported at peak COHb levels of 5% and
       February 15, 1999
                                     6-36
       DRAFT-DO NOT QUOTE OR CITE

-------
 1      higher, but later work indicates that significant behavioral impairments in healthy individuals
 2      should not be expected until COHb levels exceed 20%.  It must be emphasized, however, that
 3      even a 5% COHb level is associated with 1-h CO concentrations of 100 ppm or higher. Thus, at
 4      typical ambient air levels of CO, no observable central nervous system effects would be expected
 5      to occur in the healthy population.
 6          The current ambient air quality standards for CO (9 ppm for 8 h and 35 ppm for 1 h) are
 7      intended to keep COHb levels below 2.1% to protect the most sensitive  members of the general
 8      population (i.e., individuals with CAD). Individuals in motor vehicles are at the greatest risk
 9      from ambient CO exposure, followed by pedestrians, bicyclists, and joggers in the proximity of
10      roadways and the rest of the general urban population exposed to vehicle exhaust. Several hours
11      of exposure to peak ambient CO concentrations found occasionally at downtown urban sites
12      during periods of heavy traffic would be required to produce COHb levels of concern in the most
13      sensitive nonsmokers. Carbon monoxide levels occurring outside the downtown urban locations
14      are expected to be lower and are probably representative of levels found in residential areas
15      where most people live.  Significant health effects from ambient CO exposure  are not likely
16      under these latter exposure conditions.  Active cigarette smoking increases the risk for
17      developing cardiovascular and pulmonary disease, and passive smoking also can elevate COHb
18      levels  in nonsmokers under conditions of poor ventilation, putting nonsmoking co-workers and
19      family members at increased risk.  Carbon monoxide poisoning from indoor exposures to higher
20      than ambient CO levels occurs frequently, has more severe consequences, and  often is
21      overlooked. Efforts in prevention through public and medical education should be encouraged.
22
       February 15, 1999                         6-37        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.; Pagano,
14             M.; Selvester, R. H.; Walden, S. M.; Warren, J. (1989b) Acute effects of carbon monoxide exposure on
15             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       Bascom, R.; Bromberg, P. A.; Costa, D. L.; Devlin, R.; Dockery, D. W.; Frampton, M. W.; Lambert, W.; Samet,
29             J. M.; Speizer, F. E.; Utell, M. (1996) Health effects of outdoor air pollution [part 2]. Am. J. Respir. Crit.
30             Care Med. 153:477-498.
31       Bassan, M. M. (1990) Sudden cardiac death [letter to the  editor]. N. Engl. J. Med. 322: 272.
32       Benignus, V. A. (1993) Importance of experimenter-blind procedure in neurotoxicology. Neurotoxicol. Teratol.
33             15:45-49.
34       Benignus, V. A. (1994) Behavioral effects of carbon monoxide: meta analyses and extrapolations. J.  Appl. Physiol.
35             76:  1310-1316.
36       Benignus, V. A. (1996) Behavioral effects of carbon monoxide exposure: results and mechanisms. In: Penney,
37             D. G., ed. Carbon monoxide. Boca Raton, FL: CRC Press; pp. 211-238.
38       Benignus, V. A.; Muller, K. E.; Barton, C. N.; Prah, J. D. (1987) Effect of low level carbon monoxide on
39             compensatory tracking and event monitoring. Neurotoxicol. Teratol. 9: 227-234.
40       Benowitz, N. L. (1997) The role of nicotine in smoking-related cardiovascular disease. Preventive Med.
41             26:412-417.
42       Blair, P. S.; Fleming, P. J.; Bensley, D.; Smith, L; Bacon, C.; Taylor, E.; Berry, J.; Golding, J.; Tripp, J. (1996)
43             Smoking and the sudden infant death syndrome: results from 1993-5 case-control study for confidential
44             inquiry into stillbirths and deaths in infancy. Br. Med. J. 313: 195-198.
45       Brooke, O. G.; Anderson, H. R.; Bland, J. M.; Peacock, J. L.; Stewart, C. M. (1989) Effects on birth weight of
46             smoking, alcohol, caffeine, socioeconimic factors,  and psychosocial stress. Br.  Med. J. 298: 795-801.
47       Burnett, R. T.; Dales, R.; Krewski, D.; Vincent, R.; Dann, T.; Brook, J. R. (1995) Associations between ambient
48             paniculate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am. J. Epidemiol.
49             142: 15-22.
50       Burnett, R. T.; Dales, R. E.; Brook, J. R.; Raizenne, M. E.; Krewski, D. (1997) Association between ambient carbon
51             monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 Canadian cities.
52             Epidemiology 8: 162-167.
         February 15, 1999                              6-38          DRAFT-DO NOT QUOTE OR CITE

-------
  1       Burnett, R. T.; Cakmak, S.; Raizenne, M. E.; Stieb, D.; Vincent, R.; Krewski, D.; Brook, J. R.; Philips, O.;
  2             Ozkaynak, H. (1998a) The association between ambient carbon monoxide levels and daily mortality in
  3             Toronto, Canada. J. Air Waste Manage. Assoc. 48: 689-700.
  4       Burnett, R. T.; Cakmak, S.; Brook, J. R. (1998b) The effect of the urban ambient air pollution mix on daily
  5             mortality rates in 11 Canadian cities. Can. J. Public Health 89: 152-156.
  6       Burns, T. R.; Greenberg, S. D.; Cartwright, J.; Jachimczyk, J. A.  (1986) Smoke inhalation: anultrastructural study
  7             of reaction to injury in the human alveolar wall. Environ. Res. 41: 447-457.
  8       Busey, W. M. (1972) Chronic exposure of albino rats to certain airborne pollutants. New York, NY: Coordinating
  9             Research Council, Inc.; report no. CRC-APRAC-CAPM-6-68-6. Available from: NTIS, Springfield, VA;
10             PB-212794.
11       Carratu, M. R.; Renna, G.; Giustino, A.; De Salvia, M. A.; Cuomo, V. (1993) Changes in peripheral nervous system
12             activity produced in rats by prenatal exposure to carbon monoxide. Arch. Toxicol. 67: 297-301.
13       Carratu, M. R.; Cagiano, R.; De Salvia, M. A.; Trabace, L.; Cuomo, V. (1995) Developmental neurotoxicity of
14             carbon monoxide. In: Degen, G. H.; Seiler,  J. P.; Bentley,  P., eds. Toxicology in transition: proceedings of
15             the 1994 EUROTOX congress; August 1994; Basel, Switzerland. Arch. Toxicol. Suppl. 17: 295-301.
16       Carratu, M. R.; Cagiano, R.; Cortese, L; Renna, G.; Siro Brigiani, G.; Trabace, L.; Cuomo, V. (1996) Myelination
17             disorders produced by prenatal exposure to  moderate carbon monoxide concentrations in rat offspring.
18             Fundam. Clin. Pharmacol. 10: 172.
19       Gary, R.; Clarke, S.; Delic, J. (1997) Effects of combined exposure to noise and toxic substances—critical review of
20             the literature. Ann. Occup. Hyg. 41: 455-465.
21       Chaitman, B. R.; Dahms, T. E.; Byers, S.; Carroll, L. W.; Younis, L. T.; Wiens, R. D. (1992) Carbon monoxide
22             exposure of subjects with documented cardiac arrhythmias. Cambridge, MA: Health Effects Institute;
23             research report no. 52.
24       Chen, S.; Weller, M. A.; Penney, D. G. (1982) A study of free lung cells from young rats chronically exposed to
25             carbon monoxide from birth. Scanning Electron Microsc. 2: 859-867.
26       Cohen, S. L; Deane, M.; Goldsmith, J. R. (1969) Carbon monoxide and survival from myocardial infarction. Arch.
27             Environ. Health 19: 510-517.
28       Cooper, K. R.; Alberti, R. R. (1984) Effect of kerosene heater emissions on indoor air quality and pulmonary
29             function. Am. Rev. Respir. Dis. 129: 629-631.
30       Crocker, P. J.; Walker, J. S. (1985) Pediatric carbon monoxide toxicity. J. Emerg. Med. 3: 443-448.
31       Croft, J. B.; Giles, W. H.; Pollard, R. A.; Casper, M. L.; Anda, R. F.; Livengood, J. R. (1997) National trends in the
32             initial hospitalization for heart failure. J. Am. Geriatr. Soc. 45: 270-275.
33       Dahms,  T. E.; Younis, L. T.; Wiens, R. D.; Zarnegar, S.; Byers, S. L.; Chaitman, B. R. (1993) Effects of carbon
34             monoxide exposure in patients with documented cardiac arrhythmias. J. Am. Coll. Cardiol. 21: 442-450.
35       Davies, D. M.; Smith, D. J. (1980) Electrocardiographic changes in healthy men during continuous low-level carbon
36             monoxide exposure. Environ. Res. 21: 197-206.
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       De Luca, A.; Pierno, S.; Tricarico, D.; Carratu, M. R.; Cagiano, R.; Cuomo, V.; Camerino, D. C.  (1996)
40             Developmental changes of membrane electrical properties of rat skeletal muscle fibers produced by prenatal
41             exposure to carbon monoxide. Environ. Toxicol. Pharmacol. 2: 213-221.
42       DeLucia, A. J.; Whitaker, J. H.; Bryant, L. R. (1983) Effects of combined exposure to ozone and carbon monoxide
43             (CO) in humans. In: Lee,  S. D.; Mustafa, M. G.; Mehlman, M. A., eds. International symposium on the
44             biomedical effects of ozone and related photochemical oxidants; March; Pinehurst, NC. Princeton, NJ:
45             Princeton Scientific Publishers, Inc.; pp. 145-159. (Advances in modern environmental toxicology: v. 5).
46       De Salvia, M. A.; Cagiano, R.; Carratu, M. R.; Di  Giovanni, V.; Trabace, L.; Cuomo, V. (1995) Irreversible
47             impairment of active avoidance behavior in rats prenatally exposed to mild concentrations of carbon
48             monoxide. Psychopharmacology (Berlin) 122: 66-71.
49       Di Giovanni, V.; Cagiano, R.; De Salvia, M. A.; Giustino, A.; Lacomba, C.; Renna, G.; Cuomo, V. (1993)
50             Neurobehavioral changes produced in rats by prenatal exposure to carbon monoxide. Brain Res.
51             616:126-131.
52       Dill, D. B.; Alexander, W. C.; Myhre, L. G.; Whinnery, J. E.; Tucker, D. M. (1985) Aerobic capacity of D.B. Dill,
53             1928-1984. Fed. Proc. 44: 1013.
54       Doblar, D. D.; Santiago, T. V.; Edelman, N. H. (1977) Correlation between ventilatory and cerebrovascular
55             responses to  inhalation of CO. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 43: 455-462.


         February 15, 1999                              6-39         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Doi, R.; Tanaka, H. (1984) Effects of carbon monoxide exposure on the distribution of methylmercury in mice.
  2             J. Toxicol. Sci. 9: 11-22.
  3       Douglas, A. S.; Allan, T. M; Helms, P. J. (1996) Seasonality and the sudden infant death syndrome during 1987-9
  4             and 1991-3 in Australia and Britain. Br. Med. J. 312: 1381-1383.
  5       Drinkwater, B. L.; Raven,  P. B.; Horvath, S. M.; Gliner, J. A.; Ruhling, R. O.; Bolduan, N. W.; Taguchi, S. (1974)
  6             Air pollution, exercise, and heat stress. Arch. Environ. Health 28:  177-181.
  7       Edling, C.; Axelson, O. (1984) Risk factors of coronary heart disease among personnel in a bus company. Int. Arch.
  8             Occup. Environ. Health 54:181-183.
  9       Ekblom, B.; Huot, R. (1972) Response to submaximal and maximal exercise at different levels of
10             carboxyhemoglobin. ActaPhysiol. Scand. 86: 474-482.
11       Engen, T. (1986) The combined effect of carbon monoxide and alcohol on odor sensitivity. Environ. Int.
12             12:207-210.
13       Evans, R. G.; Webb, K.; Homan, S.; Ayres, S. M. (1988) Cross-sectional and longitudinal changes in pulmonary
14             function associated with automobile pollution among bridge and tunnel officers. Am. J. Ind. Med. 14: 25-36.
15       Fechter, L. D. (1995) Combined effects of noise and chemicals. Occup. Med. State of the Art Rev. 10: 609-621.
16       Fechter, L. D.; Young, J. S.;  Carlisle,  L. (1988) Potentiation of noise induced threshold shifts and hair cell loss by
17             carbon monoxide. Hear. Res. 34: 39-47.
18       Fechter, L. D.; Liu,  Y.; Pearce, T. A. (1997) Cochlear protection from carbon monoxide exposure by free radical
19             blockers in the guinea pig. Toxicol. Appl. Pharmacol. 142: 47-55.
20       Fein, A.; Grossman, R. F.; Jones, J. G.; Hoeffel, J.; McKay, D. (1980) Carbon monoxide effect on alveolar
21             epithelial permeability. Chest 78: 726-731.
22       Fisher, A. B.; Hyde, R. W.; Baue, A. E.; Reif, J. S.; Kelly, D. F. (1969) Effect of carbon monoxide on function and
23             structure of the lung. J. Appl. Physiol. 26: 4-12.
24       Fodor, G. G.; Winneke, G. (1972) Effect of low CO concentrations on resistance to monotony and on psychomotor
25             capacity. Staub Reinhalt. Luft 32: 46-54.
26       Franco, P.; Groswasser, J.; Sottiaux, M.; Broadfield, E.; Kahn, A. (1996) Decreased cardiac responses to auditory
27             stimulation during prone sleep. Pediatrics 97: 174-178.
28       Glantz, S. A.;  Parmley, W. W. (1991) Passive smoking and heart disease: epidemiology, physiology, and
29             biochemistry. Circulation 83: 1-12.
30       Gliner, J. A.; Raven, P. B.; Horvath, S. M.; Drinkwater, B. L.; Sutton, J. C. (1975) Man's physiologic response to
31             long-term work during thermal and pollutant stress. J. Appl. Physiol. 39: 628-632.
32       Goldsmith, J. R.; Landaw, S. A. (1968) Carbon monoxide and human health. Science (Washington, DC)
33             162: 1352-1359.
34       Graves, E. J.; Owings, M.  F. (1998) 1996 summary: National Hospital Discharge Survey. Advance data from vital
3 5             and health statistics; no. 301. Hyattsville, MD: National Center for Health Statistics; DHHS publication no.
36             (PHS) 98-1250. Available online at: http://www.cdc.gov/nchswww/releases/98facts/98sheets/ineinj.htm.
37       Groll-Knapp, E.; Haider, M.; Kienzl, K.; Handler, A.; Trimmel, M. (1988) Changes in discrimination learning and
38             brain activity (ERP's)  due to  combined exposure to NO and CO in rats. Toxicology 49: 441-447.
39       Grover, R. F.; Weil, J. V.;  Reeves, J. T. (1986) Cardiovascular adaptation to exercise at high altitude. Exercise
40             Sport Sci. Rev. 14:  269-302.
41       Hackney, J. D.; Linn, W. S.;  Mohler, J. G.; Pedersen, E. E.; Breisacher, P.; Russo, A. (1975a) Experimental studies
42             on human health effects of air pollutants: II. four-hour exposure to ozone alone and in combination with
43             other pollutant gases.  Arch. Environ. Health 30: 379-384.
44       Hackney, J. D.; Linn, W. S.;  Law, D. C.; Karuza, S. K.; Greenberg, H.; Buckley, R. D.; Pedersen, E. E. (1975b)
45             Experimental studies on human health effects of air pollutants: III. two-hour exposure to ozone alone and in
46             combination with other pollutant gases. Arch. Environ. Health 30: 385-390.
47       Hagberg, M.; Kolmodin-Hedman, B.; Lindahl, R.; Nilsson, C.-A.; Norstrom, A. (1985) Irritative complaints,
48             carboxyhemoglobin increase and minor ventilatory function changes due to exposure to chain-saw exhaust.
49             Eur. J. Respir. Dis.  66: 240-247.
50       Haldeman, G. A.; Rashidee,  A.; Horswell, R. (1998) Changes in mortality from heart failure—United States,
51             1980-1995. Morb. Mortal. Wkly. Rep. 47: 633-637.
52       Hanley, D. F.; Wilson, D.  A.; Traystman, R. J. (1986) Effect of hypoxia and hypercapnia on neurohypophyseal
53             blood flow. Am. J. Physiol. 250: H7-H15.
54       Helfaer, M. A.; Traystman, R. J. (1996) Cerebrovascular effects of carbon monoxide. In: Penney, D. G., ed. Carbon
55             monoxide. Boca Raton, FL: CRC Press; pp. 69-86.


         February 15, 1999                              6-40         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Hexter, A. C.; Goldsmith, J. R. (1971) Carbon monoxide: association of community air pollution with mortality.
  2             Science (Washington, DC) 172: 265-267.
  3       Hinderliter, A. L.; Adams, K. F., Jr.; Price, C. J.; Herbst, M. C.; Koch, G.; Sheps, D. S. (1989) Effects of low-level
  4             carbon monoxide exposure on resting and exercise-induced ventricular arrhythmias in patients with coronary
  5             artery disease and no baseline ectopy. Arch. Environ. Health 44: 89-93.
  6       Hoppenbrouwers, T.; Calub, M.; Arakawa, K.; Hodgman, J. E. (1981) Seasonal relationship of sudden infant death
  7             syndrome and environmental pollutants. Am. J. Epidemiol. 113: 623-635.
  8       Horvath, S. M. (1981) Impact of air quality in exercise performance. Exercise Sport Sci. Rev.  9: 265-296.
  9       Horvath, S. M.;  Damns, T. E.; O'Hanlon, J. F. (1971) Carbon monoxide and human vigilance: a deleterious effect of
10             present urban concentrations. Arch. Environ. Health 23: 343-347.
11       Horvath, S. M.;  Raven, P. B.; Dahms, T. E.; Gray, D. J. (1975) Maximal aerobic capacity at different levels of
12             carboxyhemoglobin. J. Appl. Physiol. 38: 300-303.
13       Howard, G.; Wagenknecht, L. E.; Burke, G. L.; Diez-Roux, A.; Evans, G. W.; McGovern, P.;  Nieto, J.; Tell, G. S.
14             (1998) Cigarette smoking and progression of atherosclerosis: the atherosclerosis risk in communities (ARIC)
15             study. JAMA J. Am. Med. Assoc. 279: 119-124.
16       Hugod, C. (1979) Effect of exposure to 0.5 ppm hydrogen cyanide singly or combined with 200 ppm carbon
17             monoxide and/or 5 ppm nitric oxide on coronary arteries, aorta, pulmonary artery, and lungs in the rabbit. Int.
18             Arch. Occup. Environ. Health 44: 13-23.
19       Hugod, C. (1980) The effect of carbon monoxide exposure on morphology of lungs and pulmonary arteries in
20             rabbits: a light- and electron-microscopic study. Arch. Toxicol. 43: 273-281.
21       Hutter, C. D. D.; Blair, M. E. (1996) Carbon monoxide—does fetal exposure cause sudden infant death syndrome?
22             Med. Hypotheses 46: 1-4.
23       Ito, K.; Thurston, G. D. (1996) Daily PM10/mortality associations: an investigation of at-risk subpopulations.
24             J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
25       Ito, K.; Kinney,  P.; Thurston, G.  D. (1995) Variations in PM-10 concentrations within two metropolitan areas and
26             their implications for health effects analyses. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the
27             colloquium on paniculate air pollution and human mortality and morbidity, part II; January 1994; Irvine, CA.
28             Inhalation Toxicol. 7: 735-745.
29       Jarvis, M. J. (1987) Uptake of environmental tobacco smoke. In: O'Neill, I. K.; Brunnemann, K. D.; Dodet, B.;
30             Hoffmann, D., eds. Environmental carcinogens: methods of analysis and exposure measurement, v. 9 -
31             passive smoking. Lyon, France: International Agency for Research on Cancer; pp. 43-58.  (IARC scientific
32             publications no. 81).
33       Jedrychowski, W.; Flak, E. (1996) Confronting the prenatal effects of active and passive tobacco smoking on the
3 4             birth weight of children. Cent. Eur. J. Public Health 4: 201 -205.
35       Jones, M. D., Jr.; Traystman, R. J. (1984) Cerebral oxygenation of the fetus, newborn, and adult. Semin. Perinatal.
36             8:205-216.
37       Kahn, A.; Groswasser, J.; Sottiaux, M.; Rebuffat, E.; Franco, P.; Dramaix, M. (1993) Prone or supine body position
38             and sleep characteristics in infants. Pediatrics 91: 1112-1115.
39       Kalmaz, E. E.; Saunders, P. J.; Kalmaz, G. D. (1993) Effects of carbon monoxide on arterial endothelial
40             permeability and its possible relation to arterial disease. J. Occup. Med. Toxicol. 2:31-39.
41       Kanten, W. E.; Penney, D. G.; Francisco, K.; Thill, J. E. (1983) Hemodynamic responses to acute
42             carboxyhemoglobinemia in the rat. Am. J. Physiol. 244: H320-H327.
43       Katsumata, Y.; Aoki, M.; Oya, M.; Yada, S.; Suzuki, O. (1980) Liver damage in rats during acute carbon monoxide
44             poisoning. Forensic Sci. Int. 16: 119-123.
45       Kawachi, L; Colditz, G. A.; Speizer, F. E.; Manson, J. E.; Stampfer, M. J.; Willett, W. C.; Hennekens, C. H. (1997)
46             A prospective study of passive smoking and coronary heart disease. Circulation 95: 2374-2379.
47       Kelsall, J. E.; Samet, J. M.; Zeger, S. L.; Xu, J. (1997) Air pollution and mortality in Philadelphia, 1974-1988.
48             Am. J. Epidemiol. 146: 750-762.
49       Kemp, J. S.; Livne, M.; White, D. K.; Arfken, C. L. (1998) Softness and potential to cause rebreathing: differences
50             in bedding used by infants at high and low risk for sudden infant death syndrome. J. Pediatr. 132: 234-239.
51       Khosla, S.; Laddu, A.; Ehrenpreis, S.; Somberg, J. C. (1994) Cardiovascular effects of nicotine: relation to
52             deleterious effects of cigarette smoking. Am. Heart J. 127: 1669-1672.
53       Kim, Y. C.; Carlson, G. P. (1983) Effect of carbon monoxide inhalation exposure in mice on drug metabolism in
54             vivo. Toxicol. Lett. 19: 7-13.
         February 15, 1999                              6-41         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
  2             In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
  3             mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
  4       Kizakevich, P. N.; Hazucha, M; Van Hoose, L.; McCartney, M. L.; Bolick, K.; Jochem, W. I; Clayton, C. A.;
  5             Alexander, B. V.; Boehleche, B.; Hackney, A. J.; Sheps, D. S.; Clapp, L. (1994) Noninvasive ambulatory
  6             assessment of cardiac function and myocardial ischemia in healthy subjects exposed to carbon monoxide.
  7             Frederick, MD: U.S. Army Medical Research and Materiel Command; contract no. DAMD17-91-C-1007.
  8             Available from: NTIS, Springfield, VA; AD-A286 014.
  9       Klein, J. P.; Forster, H. V.; Stewart, R. D.; Wu, A. (1980) Hemoglobin affinity for oxygen during short-term
10             exhaustive exercise. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 48: 236-242.
11       Kleinman, M. T. (1992) Health effects of carbon monoxide. In: Lippmann, M., ed. Environmental toxicants: human
12             exposures and their health effects. New York, NY: Van Nostrand Reinhold; pp. 98-118.
13       Kleinman, M. T.; Leaf, D. (1991) Effects of exposure to low-level carbon monoxide at high altitude insensitive
14             subjects.  Sacramento, CA: California Air Resources Board. Available from: NTIS, Springfield, VA;
15             PB95-103842.
16       Kleinman, M. T.; Davidson, D. M.; Vandagriff, R. B.; Caiozzo, V J.; Whittenberger, J. L. (1989) Effects of
17             short-term exposure to carbon monoxide in subjects with coronary artery disease.  Arch. Environ. Health
18             44:361-369.
19       Kleinman, M. T.; Leaf, D. A.; Kelly, E.; Caiozzo, V.; Osann, K.; O'Niell, T. (1998) Urban angina in the mountains:
20             effects of carbon monoxide and mild hypoxemia on subjects with chronic stable angina. Arch. Environ.
21             Health [in press].
22       Klonoff-Cohen, H.  S.; Edelstein, S. L.; Lefkowitz, E. S.;  Srinivasan, I. P.; Kaegi, D.; Chang, J. C.; Wiley, K. J.
23             (1995) The effect of passive smoking and tobacco exposure through breast milk on sudden infant death
24             syndrome. JAMA J. Am. Med. Assoc. 273: 795-798.
25       Knisely, J. S.; Rees, D. C.; Salay, J. M.; Balster, R. L.; Breen, T. J. (1987) Effects of intraperitoneal carbon
26             monoxide on fixed-ratio and screen-test performance in the mouse. Neurotoxicol. Teratol. 9: 221-225.
27       Knisely, J. S.; Rees, D. C.; Balster, R. L. (1989) Effects of carbon monoxide in combination with behaviorally
28             active drugs on fixed-ratio performance in the mouse. Neurotoxicol. Teratol.  11: 447-452.
29       Koike, A.; Wasserman, K. (1992) Effect of acute reduction in oxygen transport on parameters of aerobic function
30             during exercise. Ann. Acad. Med. Singapore 21: 14-22.
31       Koike, A.; Wasserman, K.; Armon, Y.; Weiler-Ravell, D. (1991) The work-rate-dependent effect of carbon
32             monoxide on ventilatory control during exercise. Respir. Physiol. 85: 169-183.
33       Kolarzyk, E. (1994a) The effect of acute carbon monoxide poisoning on the respiratory system efficiency. I. Values
34             of spirometric parameters in different degrees of poisoning. Int. J. Occup. Med. Environ. Health 7: 225-235.
3 5       Kolarzyk, E. (1994b) The effect of acute carbon monoxide poisoning on the respiratory system efficiency. II. Types
36             of ventilatory disorder and dynamics of changes according to the severity of carbon monoxide poisoning. Int.
37             J. Occup. Med. Environ. Health 7: 237-243.
3 8       Kolarzyk, E. (1995) Regulation of breathing in cases of acute carbon monoxide poisoning. Int. J. Occup. Med.
39             Environ. Health 8: 89-101.
40       Koren, G.; Sharav, T.; Pastuszak, A.; Garrettson, L. K.; Hill, K.; Samson, I.; Rorem, M.; King, A.; Dolgin, J. E.
41             (1991) A multicenter, prospective study of fetal outcome following accidental carbon monoxide poisoning in
42             pregnancy. Reprod. Toxicol. 5: 397-403.
43       Koskela, R.-S. (1994) Cardiovascular diseases among foundry workers exposed to carbon monoxide. Scand. J.
44             Work Environ. Health 20: 286-293.
45       Kritz, H.; Schmid, P.; Sinzinger, H. (1995) Passive smoking and cardiovascular risk. Arch. Intern. Med.
46             155: 1942-1948.
47       Kuller, L. H.; Radford, E. P.; Swift, D.; Perper, J. A.; Fisher, R. (1975) Carbon monoxide and heart attacks. Arch.
48             Environ. Health 30: 477-482.
49       Kurt, T. L.; Mogielnicki, R. P.; Chandler, J. E. (1978) Association of the frequency of acute cardiorespiratory
50             complaints with ambient levels of carbon monoxide. Chest 74: 10-14.
51       Kuska, J.; Kokot, F.; Wnuk, R. (1980) Acute renal failure after exposure to carbon monoxide. Mater. Med. Pol.
52             (Engl. Ed.) 12: 236-238.
53       Langston, P.; Gorman, D.; Runciman, W.; Upton, R. (1996) The effect of carbon monoxide on oxygen metabolism
54             in the brains  of awake sheep. Toxicology 114: 223-232.
         February 15, 1999                              6-42         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Leaf, D. A.; Kleinman, M. T. (1996a) Urban ectopy in the mountains: carbon monoxide exposure at high altitude.
  2             Arch. Environ. Health 51: 283-290.
  3       Leaf, D. A.; Kleinman, M. T. (1996b) Acute exposure to carbon monoxide does not affect plasma lipids,
  4             lipoproteins, and apolipoproteins. Angiology 47: 337-341.
  5       Lebowitz, M. D.; Collins, L.; Holberg, C. J. (1987) Time series analyses of respiratory responses to indoor and
  6             outdoor environmental phenomena. Environ. Res. 43: 332-341.
  7       Levin, B. C. (1996) New research avenues in toxicology: 7-gas N-gas model, toxicant suppressants, and genetic
  8             toxicology. In: Chaturvedi, A. K.; Sanders, D. C.; Higgins, E. A., eds. Proceedings of the international
  9             colloquium on advances in combustion toxicology; April 1995; Oklahoma City, OK. Toxicology
10             115:89-106.
11       Levin, B.C.; Paabo, M.; Gurman, J. L.; Harris, S. E.; Braun, E. (1987a) Toxicological interactions between carbon
12             monoxide and carbon dioxide. Presented at: 16th conference of toxicology, Air Force Medical Research
13             Laboratory; October 1986; Dayton, OH. Toxicology 47: 135-164.
14       Levin, B.C.; Paabo, M.; Gurman, J. L.; Harris, S. E. (1987b) Effects of exposure to single or multiple combinations
15             of the predominant toxic gases and low oxygen atmospheres produced in fires. Fundam. Appl. Toxicol.
16             9:236-250.
17       Levin, B. C.; Paabo, M.; Gurman, J. L.; Clark, H. M.; Yoklavich, M.  F. (1988) Further studies of the lexicological
18             effects of different time exposures to the individual and combined fire gases—carbon monoxide, hydrogen
19             cyanide, carbon dioxide and reduced oxygen. In: Polyurethanes 88: proceedings of the SPI 31st annual
20             technical/marketing conference; October; Philadelphia, PA. Society of the Plastics Industry, Inc.; pp.
21             249-252.
22       LeVois, M. E.; Layard, M. W. (1995) Publication bias in the environmental tobacco smoke/coronary heart disease
23             epidemiologic literature. Regul. Toxicol. Pharmacol. 21: 184-191.
24       Luke, B. (1994) Maternal-fetal nutrition. Clin. Obstet. Gynecol. 37: 93-109.
25       Lutz, L. J. (1983) Health effects of air pollution measured by outpatient visits. J. Fam. Pract. 16: 307-313.
26       MacDorman, M. F.; Cnattingius, S.; Hoffman, H. J.; Kramer, M. S.; Haglund, B. (1997) Sudden infant death
27             syndrome and smoking in the United States and Sweden. Am.  J.  Epidemiol. 146: 249-257.
28       Martin, T. R.; Bracken, M. B. (1986) Association of low birth weight with passive smoke exposure in pregnancy.
29             Am. J. Epidemiol. 124: 633-642.
30       May, D. S.; Kelly, J.  J.; Mendlein, J. M.; Garbe, P. L. (1991) Surveillance of major causes of hospitalization among
31             the elderly, 1988. Morb. Mortal. Wkly. Rep. 40: 7-21.
32       McGrath, J. J. (1982) Physiological effects of carbon monoxide. In: McGrath, J. J.; Barnes, C. D., eds.  Air
33             pollution - physiological effects. New York, NY: Academic Press; pp. 147-181.
34       McGrath, J. J. (1992) Effects of altitude on endogenous carboxyhemoglobin levels. J. Toxicol. Environ. Health
35             35: 127-133.
36       McGrath, J. J.; Lee, P.; Schreck, R. M. (1991) Effects of altitude on expired breath carbon monoxide levels.
3 7             Presented at: 84th annual meeting and exhibition of the Air &  Waste Management Association; June;
38             Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-138.2.
39       McGrath, J. J.; Schreck, R. M.; Lee, P. S. (1993) Carboxyhemoglobin levels in humans: effects of altitude.
40             Inhalation Toxicol. 5:241-249.
41       McMillan, D. E.; Miller, A. T., Jr. (1974) Interactions between carbon monoxide and J-amphetamine or
42             pentobarbital on schedule-controlled behavior. Environ. Res. 8: 53-63.
43       Medical College of Wisconsin. (1974) Exposure of humans to carbon monoxide combined with ingestion of ethyl
44             alcohol and the comparison of human performance when exposed for varying periods of time to carbon
45             monoxide. Milwaukee, WI: Medical College of Wisconsin; report no. MCOW-ENVM-CO-74-2. Available
46             from: NTIS, Springfield, VA; PB-242099.
47       Melius, J. M. (1995)  Cardiovascular disease among firefighters. Occup. Med.: State of the Art Rev. 10: 821-827.
48       Mennear, J. H. (1993) Carbon monoxide and cardiovascular disease:  an analysis of the weight of evidence. Regul.
49             Toxicol. Pharmacol.  17: 77-84.
50       Merigan, W. H.; Mclntire, R. W. (1976) Effects of carbon monoxide on responding under a progressive ratio
51             schedule in rats. Physiol. Behav. 16: 407-412.
52       Michaels, D.; Zoloth, S. R. (1991) Mortality among urban bus drivers. Int.  J. Epidemiol. 20: 399-404.
53       Min, H. Y.; Wei, Z.;  Quan, Q. M.; Rong,  Z. (1992) Correlation between hemorheological parameters and
54             viscoelasticity of thrombus in rats after exposure to carbon monoxide. Clin. Hemorheol.  12: 593-598.
         February 15, 1999                             6-43         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Mitchell, D. S.; Packham, S. C.; Fitzgerald, W. E. (1978) Effects of ethanol and carbon monoxide on two measures
  2             of behavioral incapacitation of rats. Proc. West. Pharmacol. Soc. 21: 427-431.
  3       Mitchell, E. A.; Ford, R. P. K.; Stewart, A. W.; Taylor, B. I; Becroft, D. M. O.; Thompson, J. M. D.; Scragg, R.;
  4             Hassall, I. B.; Barry, D. M. J.; Allen, E. M.; Roberts, A. P. (1993) Smoking and the sudden death syndrome.
  5             Pediatrics 91: 893-896.
  6       Montgomery, M. R.; Rubin, R. J. (1971) The effect of carbon monoxide inhalation on in vivo drug metabolism in
  7             the rat.  J. Pharmacol. Exp. Ther. 179: 465-473.
  8       Montgomery, M. R.; Rubin, R. J. (1973) Oxygenation during inhibition of drug metabolism by carbon monoxide or
  9             hypoxic hypoxia. J. Appl. Physiol. 35: 505-509.
10       Moore, L. G.;  Rounds,  S. S.; Jahnigen, D.; Grover, R. F.; Reeves, J. T. (1982) Infant birth weight is related to
11             maternal arterial oxygenation at high altitude. J. Appl. Physiol.: Respir. Environ. Exercise Physiol.
12             52:695-699.
13       Morris, R. D.; Naumova, E. N. (1998) Carbon monoxide and hospital admissions for congestive heart failure:
14             evidence of an increased effect at low temperatures. Environ. Health Perspect. 106: 649-653.
15       Morris, R. D.; Naumova, E. N.; Munasinghe, R. L. (1995) Ambient air pollution and hospitalization for congestive
16             heart failure among elderly people in seven large US cities. Am. J. Public Health 85: 1361-1365.
17       Murphy, S. D. (1964) A review of effects on animals of  exposure to auto exhaust and some of its components.
18             J. Air Pollut. Control Assoc. 14: 303-308.
19       Murray, F. J.;  Schwetz, B. A.; Crawford, A. A.; Henck, J. W.; Staples, R. E. (1978) Teratogenic potential of sulfur
20             dioxide and carbon monoxide  in mice and rabbits. In: Mahlum, D. D.; Sikov, M. R.; Hackett, P. L.; Andrew,
21             F. D., eds. Developmental toxicology of energy-related pollutants: proceedings of the seventeenth annual
22             Hanford biology symposium; October 1977; Richland, WA. Oak Ridge, TN: U.S. Department of Energy,
23             Technical Information Center; pp. 469-478. Available from: NTIS, Springfield, VA; CONF-771017.
24       National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health
25             effects. Washington, DC: National Academy Press.
26       Niden, A. H.;  Schulz, H. (1965) The  ultrastructural effects of carbon monoxide inhalation on the rat lung. Virchows
27             Arch. Pathol. Anat. Physiol. 339: 283-292.
28       Olds, D. (1997) Tobacco exposure and impaired development: a review of the evidence. Ment. Retard. Dev. Disabil.
29             Res. Rev. 3:257-269.
3 0       Olds, D. L.; Henderson, C. R.; Tatelbaum, R. (1994a) Intellectual impairment in children of women who smoke
31             cigarettes during pregnancy. Pediatrics 93: 221-227.
3 2       Olds, D. L.; Henderson, C. R.; Tatelbaum, R. (1994b) Prevention of intellectual impairment in children of women
33             who smoke  cigarettes during pregnancy. Pediatrics 93: 228-233.
34       Pankow, D.; Ponsold, W.; Fritz, H. (1974) Combined effects of carbon monoxide and ethanol on the activities of
3 5             leucine aminopeptidase and glutamic-pyruvic transaminase in the plasma of rats. Arch. Toxicol.
36             32:331-340.
37       Pantazopoulou, A.; Katsouyanni, K.; Kourea-Kremastinou, J.; Trichopoulos, D. (1995) Short-term effects of air
3 8             pollution on hospital emergency outpatient visits and admissions in the greater Athens, Greece area. Environ.
39             Res. 69:31-36.
40       Paulson, O. B.; Parving, H.-H.; Olesen, J.; Skinhoj, E. (1973) Influence of carbon monoxide and of hemodilution on
41             cerebral blood flow and blood gases in man. J. Appl. Physiol. 35: 111-116.
42       Peacock, J. L.; Bland, J. M.; Anderson, H. R.; Brooke, O. G. (1991a) Cigarette smoking and birthweight: type of
43             cigarette smoked and a possible threshold effect. Int. J. Epidemiol.  20: 405-412.
44       Peacock, J. L.; Bland, J. M.; Anderson, H. R. (1991b) Effects on birthweight of alcohol and caffeine consumption
45             in smoking women. J. Epidemiol. Commun. Health 45: 159-163.
46       Perm,  A. (1993) Determination of the atherogenic potential of inhaled carbon monoxide. Cambridge, MA: Health
47             Effects Institute; research report number 57.
48       Perm,  A.; Currie, J.; Snyder, C. (1992) Inhalation of carbon monoxide does not accelerate arteriosclerosis in
49             cockerels. Eur. J. Pharmacol. 228: 155-164.
50       Penney, D. G. (1988) A review: hemodynamic response  to carbon monoxide. Environ. Health Perspect
51             77:121-130.
52       Penney, D. G., ed.  (1996a) Carbon monoxide. Boca Raton, FL: CRC Press.
53       Penney, D. G. (1996b) Effects of carbon monoxide exposure on developing animals and humans. In: Penney,
54             D. G., ed. Carbon monoxide. Boca Raton, FL: CRC Press; pp. 109-144.
         February 15, 1999                             6-44         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Penney, D.; Benjamin, M.; Dunham, E. (1974) Effect of carbon monoxide on cardiac weight as compared with
  2             altitude effects. J. Appl. Physiol. 37: 80-84.
  3       Pereira, L. A. A.; Loomis, D.; Conceicao, G. M. S.; Braga, A. L. F.; Areas, R. M.; Kishi, H. S.; Singer, J. M.;
  4             Bohm, G. M.; Saldiva, P. H. N. (1998) Association between air pollution and intrauterine mortality in Sao
  5             Paulo, Brazil. Environ. Health Perspect. 106: 325-329.
  6       Pirnay, F.; Dujardin J.; Deroanne, R.; Petit, J. M. (1971) Muscular exercise during intoxication by carbon
  7             monoxide. J. Appl. Physiol. 31: 573-575.
  8       Poloniecki, J. D.; Atkinson, R. W.; Ponce de Leon, A.; Anderson, H. R. (1997) Daily time series for cardiovascular
  9             hospital admissions and previous day's air pollution in London, UK. Occup. Environ. Med. 54: 535-540.
10       Ponsonby, A.-L.; Dwyer, T.; Gibbons, L. E.; Cochrane, J. A.; Wang, Y.-G. (1993) Factors potentiating the risk of
11             sudden infant death syndrome associated with the prone position. N. Engl. J. Med. 329: 377-382.
12       Preziosi, T. J.; Lindenberg, R.; Levy, D.; Christenson, M. (1970) An experimental investigation in animals of the
13             functional and morphologic effects of single and repeated exposures to high and low concentrations of
14             carbon monoxide.  Ann.  N. Y. Acad. Sci. 174: 369-384.
15       Prigge, E.; Hochrainer, D. (1977) Effects of carbon monoxide inhalation on erythropoiesis and cardiac hypertrophy
16             in fetal rats. Toxicol. Appl. Pharmacol. 42: 225-228.
17       Putz, V. R.; Johnson, B. L.; Setzer, J. V. (1976) Effects of CO on vigilance performance: effects of low level
18             carbon monoxide on divided attention, pitch discrimination, and the auditory evoked potential. Cincinnati,
19             OH: U.S. Department of Health, Education, and Welfare, National Institute for Occupational Safety and
20             Health; report no. NIOSH-77-124. Available from: NTIS, Springfield, VA; PB-274219.
21       Putz, V. R.; Johnson, B. L.; Setzer, J. V. (1979) A comparative study of the effects of carbon monoxide and
22             methylene chloride on human performance. J. Environ. Pathol. Toxicol. 2: 97-112.
23       Raven, P. B.; Drinkwater, B. L.; Horvath, S. M.; Ruhling, R. O.; Gliner, J. A.; Sutton, J.  C.; Bolduan, N. W.
24             (1974a) Age, smoking habits, heat stress, and their interactive effects with carbon monoxide and
25             peroxyacetylnitrate on man's aerobic power. Int. J. Biometeorol. 18: 222-232.
26       Raven, P. B.; Drinkwater, B. L.; Ruhling, R. O.; Bolduan, N.; Taguchi, S.; Gliner, J.; Horvath, S. M. (1974b) Effect
27             of carbon monoxide and peroxyacetyl nitrate on man's maximal aerobic capacity. J. Appl. Physiol.
28             36:288-293.
29       Robertson, G.; Lebowitz, M. D. (1984) Analysis of relationships between symptoms and environmental factors over
30             time. Environ. Res. 33: 130-143.
31       Robkin, M. A. (1997) Carbon monoxide and the embryo. Int. J.  Dev. Biol. 41: 283-289.
32       Rockwell, T. J.; Weir, F.  W.  (1975) The interactive effects of carbon monoxide and alcohol on driving skills.
3 3             Columbus, OH: The Ohio State University Research Foundation; CRC-APRAC project CAPM-9-69.
34             Available from: NTIS, Springfield, VA; PB-242266.
35       Rodkey,F.L.; Collison, H. A.  (1979) Effects of oxygen and carbon dioxide on carbon monoxide toxicity.
36             J. Combust. Toxicol. 6:  208-212.
37       Roth, R. A., Jr.; Rubin, R. J.  (1976a) Role of blood flow in carbon monoxide- and hypoxic hypoxia-induced
3 8             alterations in hexobarbital metabolism in rats. Drug Metab. Dispos. 4: 460-467.
39       Roth, R. A., Jr.; Rubin, R. J.  (1976b) Comparison of the effect of carbon monoxide and of hypoxic hypoxia.
40             II. Hexobarbital metabolism in the isolated, perfused rat liver. J. Pharmacol. Exp. Ther. 199: 61-66.
41       Roth, R. A., Jr.; Rubin, R. J.  (1976c) Comparison of the effect of carbon monoxide and of hypoxic hypoxia.
42             I. In vivo metabolism, distribution and action of hexobarbital. J. Pharmacol. Exp. Ther. 199: 53-60.
43       Rubin, D. H.; Leventhal,  J. M.; Krasilnikoff, P. A.; Weile, B.; Berget, A. (1986) Effect of passive smoking on
44             birth-weight. Lancet (8504): 415-417.
45       SAS Institute Inc. (1990) The NLIN procedure. In: SAS/STAT® user's guide, version 6, fourth edition.  Volume 2.
46             Cary, NC: SAS Institute Inc.; pp. 1135-1193.
47       Saldiva, P. H. N.; Lichtenfels, A. J. F.  C.; Paiva, P. S. O.; Barone, I. A.; Martins, M. A.; Massad, E.; Pereira, J. C.
48             R.; Xavier, V. P.; Singer, J. M.; Bohm, G. M. (1994) Association between air pollution and mortality due to
49             respiratory diseases in children in Sao Paulo, Brazil: a preliminary report. Environ. Res. 65: 218-225.
50       Saldiva, P. H. N.; Pope, C. A., Ill; Schwartz, J.; Dockery, D. W.; Lichtenfels, A. J.; Salge, J. M.; Barone, I.; Bohm,
51             G. M. (1995) Air pollution and mortality in elderly people: a time-series study in Sao Paulo, Brazil. Arch.
52             Environ. Health 50: 159-163.
53       Salinas, M.; Vega, J. (1995) The effect of outdoor air pollution on mortality risk: an ecological study from Santiago,
54             Chile. World Health Stat. Q. 48: 118-125.
55


         February 15, 1999                              6-45         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Sardinas, A.; Wang Miller, J.; Hansen, H. (1986) Ischemic heart disease mortality of fireman and policemen.
  2             Am. J. Public Health 76: 1140-1141.
  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.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular disease in Detroit,
  8             Michigan. Am. J. Epidemiol. 142: 23-35.
  9       Schwartz, P. J.; Stramba-Badiale, M.; Segantini, A.; Austoni, P.; Bosi, G.; Giorgetti, R.; Grancini, F.; Marni,
10             E. D.; Perticone, F.;  Rosti, D.; Salice, P. (1998) Prolongation of the QT interval and the sudden infant death
11             syndrome. N. Engl. J. Med. 338: 1709-1761.
12       Scragg, R.; Mitchell, E. A.; Taylor, B. J.; Stewart, A. W.; Ford, R. P. K.; Thompson, J. M. D.; Allen, E. M.;
13             Becroft, D. M. O. (1993) Bed sharing, smoking, and alcohol in the sudden infant death syndrome.
14             Br. Med J. 307: 1312-1318.
15       Seeker-Walker, R. H.; Vacek, P. M.; Flynn, B. S.; Mead, P. B. (1997) Smoking in pregnancy,  exhaled carbon
16             monoxide, and birth weight. Obstet. Gynecol. 89: 648-653.
17       Shephard, R. J. (1983) Carbon monoxide: the silent killer. Springfield, IL: Charles C. Thomas. (American lecture
18             series no. 1059).
19       Shephard, R. J. (1984) Athletic performance and urban air pollution. Can. Med. Assoc. J. 131: 105-109.
20       Sheppard, D.; Distefano, S.; Morse, L.; Becker, C. (1986) Acute effects of routine firefighting on lung function.
21             Am. J. Ind. Med. 9: 333-340.
22       Sheps, D. S.; Adams, K. F., Jr.; Bromberg, P. A.; Goldstein, G. M.; O'Neil, J. J.; Horstman, D.; Koch, G. (1987)
23             Lack of effect of low levels of carboxyhemoglobin on cardiovascular function in patients with ischemic heart
24             disease. Arch. Environ. Health 42: 108-116.
25       Sheps, D. S.; Herbst, M. C.; Hinderliter, A. L.; Adams, K. F.; Ekelund, L. G.; O'Neil, J. J.; Goldstein, G.  M.;
26             Bromberg, P. A.; Dalton, J. L.; Ballenger, M. N.; Davis, S. M.; Koch, G. G. (1990) Production of
27             arrhythmias by elevated carboxyhemoglobin in patients with coronary artery disease. Ann. Intern. Med.
28             113:343-351.
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.; Ballenger, M.; Davis, S. M.; Koch, G. (1991) Effects of 4 percent and 6 percent
31             carboxyhemoglobin  on arrhythmia production in patients with coronary artery disease. Cambridge, MA:
32             Health Effects Institute; research report no. 41.
33       Singh, J. (1986) Early behavioral alterations in mice following prenatal carbon monoxide exposure.
34             Neurotoxicology 7: 475-481.
35       Singh, J.; Moore-Cheatum, L. (1993) Gestational protein deficiency enhances fetotoxicity of carbon monoxide.
36             In: Keen, C. L.; Bendich, A.; Willhite, C. C., eds. Maternal nutrition and pregnancy outcome. Ann. N. Y.
37             Acad. Sci. 678: 366-368.
38       Singh, J.; Aggison, L., Jr.; Moore-Cheatum, L. (1993) Teratogenicity and developmental toxicity of carbon
39             monoxide in protein-deficient mice. Teratology 48: 149-159.
40       Sinha, A. K.; Klein,  J.; Schultze, P.; Weiss, J.; Weiss, H. R. (1991) Cerebral regional capillary perfusion  and blood
41             flow after carbon monoxide exposure. J.  Appl. Physiol. 71: 1196-1200.
42       Smith, C. J.; Steichen, T. J. (1993) The atherogenic potential of carbon monoxide. Atherosclerosis (Shannon, Irel.)
43             99: 137-149.
44       Snella, M.-C.; Rylander, R. (1979) Alteration in local and systemic immune capacity after exposure to bursts of CO.
45             Environ. Res. 20: 74-79.
46       Spitzer,  W. O.; Lawrence, V.; Dales, R.; Hill, G.; Archer, M. C.; Clark, P.; Abenhaim, L.; Hardy, J.; Sampalis, J.;
47             Pinfold, S. P.; Morgan, P. P. (1990) Links between passive smoking and disease: a best-evidence synthesis.
48             Clin. Invest. Med. 13: 17-42.
49       Steenland, K.  (1992) Passive smoking and the risk of heart disease. JAMA J. Am. Med. Assoc. 267: 94-99.
50       Steenland, K.; Thun, M.; Lally, C.; Heath, C., Jr. (1996) Environmental tobacco smoke and coronary heart disease
51             in the American Cancer Society CPS-II cohort. Circulation 94: 622-628.
52       Stern, F. B.; Lemen, R. A.;  Curtis, R. A. (1981) Exposure of motor vehicle examiners to carbon monoxide:
53             a historical prospective mortality study. Arch. Environ. Health 36: 59-66.
54
         February 15, 1999                              6-46         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Stern, F. B.; Halperin, W. E.; Hornung, R. W.; Ringenburg, V. L.; McCammon, C. S. (1988) Heart disease mortality
  2             among bridge and tunnel officers exposed to carbon monoxide. Am. J. Epidemiol.
  3             128: 1276-1288.
  4       Stewart, R. D.; Newton, P. E.; Kaufman, J.; Forster, H. V.; Klein, J. P.; Keelen, M. H., Jr.; Stewart, D. J.; Wu, A.;
  5             Hake, C. L. (1978) The effect of a rapid 4% carboxyhemoglobin saturation increase on maximal treadmill
  6             exercise. New York, NY: Coordinating Research Council, Inc.; report no. CRC-APRAC-CAPM-22-75.
  7             Available from: NTIS, Springfield, VA; PB-296627.
  8       Strom, J.; Alfredsson, L.; Malmfors, T. (1995) Carbon monoxide: causation and aggravation of cardiovascular
  9             diseases - a review of the epidemiologic and toxicologic literature. Indoor Environ. 4: 322-333.
10       Surgeon General of the United States. (1983) The health consequences of smoking: cardiovascular disease -
11             a report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Public
12             Health Service, Office on Smoking and Health; publication no. DHHS(PHS) 84-50204.
13       Surgeon General of the United States. (1986) The health consequences of involuntary smoking: a report of the
14             Surgeon General. Rockville, MD: U. S. Department of Health and Human Services, Office on Smoking and
15             Health; publication no. DHHS (CDC) 87-8398.
16       Thorn, S. R.; Ischiropoulos, H. (1997) Mechanism of oxidative stress from low levels of carbon monoxide.
17             Cambridge, MA: Health Effects Institute; research report no. 80.
18       Thorn, S. R.; Xu, Y. A.; Ischiropoulos, H. (1997) Vascular endothelial cells generate peroxynitrite in response to
19             carbon monoxide exposure. Chem. Res. Toxicol. 10: 1023-1031.
20       Topping, D. L.; Fishlock, R. C.; Trimble, R. P.; Storer, G. B.; Snoswell, A. M. (1981) Carboxyhaemoglobin inhibits
21             the metabolism of ethanol by perfused rat liver. Biochem. Int. 3: 157-163.
22       Touloumi, G.; Pocock,  S. J.; Katsouyanni, K.; Trichopoulos, D. (1994) Short-term effects of air pollution on daily
23             mortality in Athens: a time-series analysis. Int. J. Epidemiol. 23: 957-967.
24       Touloumi, G.; Samoli, E.; Katsouyanni, K. (1996) Daily mortality and "winter type" air pollution in Athens,
25             Greece—a time series analysis within the APHEA project. In: St Leger, S., ed. The APHEA project. Short
26             term effects of air pollution on health: a European approach using epidemiological time series data.
27             J. Epidemiol. Commun. Health 50(suppl. 1): S47-S51.
28       Turino, G. M. (1981) Effect of carbon monoxide on the cardiorespiratory system.  Carbon monoxide toxicity:
29             physiology and biochemistry. Circulation 63: 253A-259A.
30       U.S. Centers for Disease Control and Prevention. (1997) Mortality patterns—preliminary data, United States, 1996.
31             Morb. Mortal. Wkly. Rep. 46: 941-944.
32       U.S. Environmental Protection Agency. (1979) Air quality criteria for carbon monoxide. Research Triangle Park,
33             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
34             no. EPA-600/8-79-022. Available from: NTIS, Springfield,  VA; PB81-244840.
35       U.S. Environmental Protection Agency. (1984) Revised evaluation of health effects associated with carbon
36             monoxide exposure: an addendum to the 1979 EPA air quality criteria document for carbon monoxide.
37             Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
3 8             Assessment Office; EPA report no. EPA-600/8-83-033F.  Available from: NTIS, Springfield, VA;
39             PB85-103471/HSU.
40       U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
41             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
42             no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
43       Verhoeff, A. P.;  Hoek,  G.; Schwartz, J.; Van Wijnen, J. H. (1996) Air pollution and daily mortality in Amsterdam.
44             Epidemiology 7: 225-230.
45       Vogel, J. A.; Gleser, M. A. (1972) Effect of carbon monoxide on oxygen transport during exercise. J. Appl. Physiol.
46             32:234-239.
47       Wall, M. A.; Johnson, J.; Jacob, P.; Benowitz, N. L. (1988) Cotinine in the serum, saliva, and urine of nonsmokers,
48             passive smokers, and active smokers. Am. J. Public Health 78: 699-701.
49       Wang, X.; Ding, H.; Ryan, L.; Xu, X. (1997) Association between air pollution and low birth weight:
50             a community-based study. Environ. Health Perspect.  105: 514-520.
51       Weiser, P. C.; Morrill, C. G.; Dickey, D. W.; Kurt, T. L.; Cropp, G.  J. A. (1978) Effects of low-level  carbon
52             monoxide exposure on the adaptation of healthy young men to aerobic work at an altitude of 1,610 meters.
53             In: Folinsbee, L. J.; Wagner, J. A.; Borgia, J. F.; Drinkwater, B. L.; Gliner, J. A.; Bedi, J. F., eds.
54             Environmental stress: individual human adaptations. New York, NY: Academic Press, Inc.; pp. 101-110.
55       Weiss, S. T. (1996) Cardiovascular effects of environmental tobacco smoke. Circulation 94: 599.


         February 15, 1999                              6-47         DRAFT-DO NOT QUOTE OR CITE

-------
  1       Weissbecker, L.; Carpenter, R. D.; Luchsinger, P. C.; Osdene, T. S. (1969) In vitro alveolar macrophage viability:
  2             effect of gases. Arch. Environ. Health 18: 756-759.
  3       Wells, A. J. (1994) Passive smoking as a cause of heart disease. J. Am. Coll. Cardiol. 24: 546-554.
  4       Wen, S. W.; Goldenberg, R. L.; Cutter, G. R.; Hoffman, H. J.; Cliver, S. P.; Davis, R. O.; DuBard, M. B. (1990)
  5             Smoking, maternal age, fetal growth, and gestational age at delivery. Am. J. Obstet. Gynecol. 162: 53-58.
  6       Werner, R. M.; Pearson, T. A. (1998) What's so passive about smoking? Secondhand smoke as a cause of
  7             atherosclerotic disease [editorial]. JAMA J. Am. Med. Assoc. 279: 157-158.
  8       Wietlisbach, V.; Pope, C. A., Ill; Ackermann-Liebrich, U. (1996) Air pollution and daily mortality in three Swiss
  9             urban areas. Soz. Praeventivmed. 41: 107-115.
10       Wouters, E. J. M.; de Jong, P. A.; Cornelissen, P. J. H.; Kurver, P. H. J.; van Del, W. C.; van Woensel, C. L. M.
11             (1987) Smoking and low birth weight: absence of influence by carbon monoxide? Eur. J. Obstet. Gynecol.
12             Reprod. Biol.  25: 35-41.
13       Yang, L.; Zhang, W.; He, H.; Zhang, G. (1988) Experimental studies on combined effects of high temperature and
14             carbon monoxide. J. Tongji Med. Univ. 8: 60-65.
15       Young, J. S.; Upchurch, M. B.; Kaufman, M. J.; Fechter, L. D. (1987) Carbon monoxide exposure potentiates
16             high-frequency auditory threshold shifts induced by noise. Hear. Res. 26: 37-43.
17       Zaren, B.; Lindmark, G.; Gebre-Medhin, M. (1996) Maternal smoking and body composition of the newborn. Acta
18             Paediatr. 85: 213-219.
19       Zebro, T.; Wright, E. A.; Littleton, R. J.; Prentice, A. I. D. (1983) Bone changes in mice after prolonged continuous
20             exposure to a high concentration of carbon monoxide. Exp. Pathol. 24: 51-67.
21       Zhu, B.-Q.; Parmley, W. W. (1995) Hemodynamic and vascular effects of active and passive smoking. Am. Heart J.
22             130:  1270-1275.
23
         February 15, 1999                             6-48         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 public health significance of CO in the air largely results from CO being absorbed
13      readily from the lungs into the bloodstream, there forming a slowly reversible complex with
14      hemoglobin (Hb), known as carboxyhemoglobin (COHb). The presence of significant levels of
15      COHb in the blood causes hypoxia (i.e., reduced availability of oxygen to body tissues). The
16      blood COHb level, therefore, represents a useful physiological marker to predict the potential
17      health effects of CO exposure. The amount of COHb formed is dependent on the CO
18      concentration and duration of exposure, exercise (which increases both the amount of air inhaled
19      and exhaled per unit of time and the pulmonary diffusing capacity for CO), ambient pressure,
20      health status, and the specific metabolism of the exposed individual.  The formation of COHb is
21      a reversible process, but, because of tight binding of CO to Hb, the elimination half-time is quite
22      long, varying from 2 to 6.5 h depending on the initial COHb levels.  This may lead to
23      accumulation of COHb, and even relatively low inhaled concentrations of CO can produce
24      substantial blood levels of COHb. Fortunately, mechanisms exist in normal, healthy individuals
25      to compensate for the reduction in tissue oxygen caused by increasing levels of COHb. Cardiac
26      output increases and blood vessels dilate to carry more blood so that the tissue can extract
27      adequate amounts of oxygen from the blood.  There are several medical conditions, however, that
28      can make an individual more susceptible to the potential adverse effects of low levels of CO,
29      especially during exercise. Occlusive vascular disease may restrict the  increase in blood flow to
30      the tissues, chronic obstructive lung disease causes gas-exchange abnormalities that limit the

        February 15, 1999                         7-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      amount of oxygen that diffuses into the blood, and anemia reduces the oxygen-carrying capacity
 2      of the blood. Under any of these conditions, exposure to CO could reduce further the amount of
 3      oxygen available to the affected tissues.
 4           The prevailing national ambient air quality standards (NAAQS) for CO of 9 ppm for 8 h
 5      and 35 ppm for 1  h (Federal Register, 1994) have been established to prevent adverse health
 6      effects in the most sensitive population groups associated with the presence of CO in the ambient
 7      air. The term "ambient air" is interpreted to mean outdoor air at ground level where people live
 8      and breathe.  A great majority of people, however, spend most of their time indoors. A realistic
 9      assessment of the health effects from exposure to ambient CO, therefore, must be set in the
10      context of total exposure, a major component of which is indoor exposure.
11           This chapter provides an overall summary of the key factors presented in Chapters 2
12      through 6 of the present document that determine what risk ambient CO poses to public health.
13      An effort also is made to qualitatively delineate key factors that contribute to anticipated health
14      risks  from ambient CO in special subpopulations that form  a significant proportion of the
15      population at large. Risk factors such as age, gender, and pregnancy are discussed, as well as
16      preexisting heart,  lung, vascular, and hematologic diseases. Subpopulations at risk because of
17      exposure to  ambient CO alone, or combined with other environmental factors are identified.
18      This information will be used by the U.S. Environmental Protection Agency's Office of Air
19      Quality Planning and Standards for development of their staff paper and associated assessments
20      that will help to determine the adequacy of the existing CO NAAQS.
21
22
23      7.2  ENVIRONMENTAL SOURCES
24           Carbon monoxide comes from both natural and anthropogenic processes.  About half of the
25      atmospheric CO is released at the earths's surface from fossil fuel and biomass burning, and the
26      rest is produced naturally in the atmosphere. About 80% of the CO emitted at the surface is from
27      human activities, whereas natural processes account for the remaining 20%.  The background
28      concentration of CO in the troposphere influences the abundance of hydroxyl radicals (OH), thus
29      affecting the global cycles of many natural and anthropogenic trace gases such as methane that
30      are removed from the atmosphere by reacting with OH.  During the 1980s, global CO
31      concentrations increased at approximately 1% per year.  More recent reports, however, show that
        February 15, 1999                         7-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1      global CO concentrations declined rapidly between 1988 and 1993.  Since 1993, the downward
 2      trend in global CO has leveled off, and it is not clear whether CO will continue to decline or will
 3      increase.
 4
 5
 6      7.3   ENVIRONMENTAL CONCENTRATIONS
 7           The annual average CO concentration is about 0.13 ppm at monitoring sites located in the
 8      marine boundary layer of the Pacific Ocean in the mid-latitudes of the Northern Hemisphere.
 9      These sites are remote from local pollutant sources, and the values obtained at these sites are
10      meant to represent global background values for CO.  Because of seasonal variations in the
11      emissions and chemical loss of CO through reaction with OH radicals, mean global background
12      CO levels vary between about 0.09 ppm in summer and about 0.16 ppm in winter. Annual 24-h
13      average CO concentrations obtained at U.S. monitoring sites in rural areas away from
14      metropolitan areas are typically about 0.20 ppm, compared with an annual 24-h average of
15      1.2 ppm across all monitoring sites in the Aerometric Information Retrieval System network in
16      1996.
17           In the United States, ambient air 8-h average CO concentrations monitored from fixed-site
18      stations in metropolitan areas are generally below 9 ppm and have decreased significantly since
19      1990 when the last CO criteria document was completed (U.S. Environmental Protection
20      Agency, 1991). In the latest year of record, 1997, the annual mean CO concentrations were all
21      less than 9 ppm.  However, in spite of the vehicle emission reductions responsible for the
22      decrease in ambient CO, high  short-term peak CO concentrations still can occur in certain
23      outdoor locations and situations associated with motor vehicles, for example, riding behind high
24      emitters or in a vehicle with a defective exhaust system. Also, air quality data from fixed-site
25      monitoring stations underestimate the short-term  peak CO levels in heavy traffic environments.
26           Indoor and in-transit concentrations of CO can be significantly different from the typically
27      low ambient CO concentrations.  The CO levels in homes without combustion sources are
28      usually lower than 5 ppm. The highest residential concentrations of CO that have been reported
29      in the scientific literature are associated with the use of unvented gas or kerosene space heaters
30      where peak concentrations of CO as high as 50 ppm have been reported.  Carbon monoxide
31      concentrations also have exceeded 9 ppm for 8 h  in several homes with gas stoves and, in one
        February 15, 1999                         7-3         DRAFT-DO NOT QUOTE OR CITE

-------
 1      case, 35 ppm for 1 h; however, these higher CO concentrations were in homes with older gas
 2      ranges that had pilot lights that burn continuously.  Newer or remodeled homes have gas ranges
 3      with electronic pilot lights. Also, the availability of other cooking appliances (e.g., microwaves,
 4      heating plates) has decreased the use of gas ranges in meal preparation.
 5          Average CO concentrations as high as 10 to 12  ppm have been reported in human exposure
 6      studies for in-vehicle compartments of moving automobiles. Carbon monoxide concentrations
 7      will depend, however, on the season and traffic pattern, and the findings of more recent studies
 8      imply that pre-1990 study results are no longer applicable. For example, commuter exposure to
 9      motor vehicle exhaust fell from a high of 37 ppm CO for a Los Angeles, CA, study in 1965 to a
10      low of 3 ppm CO for a New Jersey Turnpike study in 1992, and for  San Francisco, CA, using the
11      same data collection protocol, typical commuter exposures fell about 50% in the 11-year period
12      from 1980 to 1991, despite a 19% increase in average daily traffic. Carbon monoxide levels in
13      other indoor environments affected by engine exhaust (e.g., parking garages, tunnels) follow
14      similar trends but tend to be higher than in other indoor environments.
15          Because indoor and outdoor air quality differ substantially, and because people  spend much
16      of their time indoors, ambient air quality measurements alone do not provide accurate estimates
17      of personal or population exposure to CO from ambient and nonambient sources.  Whereas the
18      ambient monitoring data reflect exposure to ambient  sources of CO  only, the measurement of
19      CO from personal monitors reflects more accurately the actual total  human population exposure
20      to CO.
21          Occupational 8-h exposure levels are generally  below 35 ppm, the recommended
22      occupational  exposure limit for CO (Federal Register, 1989; American Conference of
23      Governmental Industrial Hygienists, 1994; Code of Federal Regulations, 1998). Workers
24      exposed to exhaust gases, especially from older vehicles, may  have peak exposures over 200 ppm
25      in enclosed spaces.
26
27
28      7.4  CARBOXYHEMOGLOBIN LEVELS IN THE POPULATION
29          Carbon monoxide diffuses rapidly across the alveolar and capillary membranes  and more
30      slowly across the placental membrane.  At equilibrium, approximately 95% of the absorbed CO
31      binds with hemoglobin to form  COHb that, when elevated above the endogenous level, is a
        February 15,  1999                          7-4        DRAFT-DO NOT QUOTE OR CITE

-------
 1      specific biomarker of CO exposure in blood. The remaining 5% is distributed extravascularly.
 2      During continuous exposure to a fixed ambient concentration of CO, the COHb concentration
 3      increases rapidly at the onset of exposure, starts to level off after 3 h, and approaches a steady
 4      state after 6 to 8 h of exposure. Therefore, an 8-h COHb value would be closely representative of
 5      any longer continuous exposures. In real-life situations, the prediction of individual COHb levels
 6      is difficult because of large spatial and temporal variations in both indoor and outdoor levels of
 7      CO and temporal variations of alveolar ventilation rates. Because COHb measurements are not
 8      readily available in the exposed population, mathematical models have been developed to predict
 9      COHb levels from known CO exposures under a variety of circumstances (see Figure 7-1).
10          Evaluation of human CO exposure situations indicates that occupational exposures in some
11      workplaces or exposures in homes with faulty or unvented combustion appliances can exceed
12      100 ppm CO, often leading to COHb levels of 4 to 5% with 1-h exposure and 10% or more with
13      continued exposure for 8 h or longer (see Table 7-1).  Such high exposure levels are encountered
14      much less frequently by the general public under ambient conditions.  More frequently,
15      short-term exposures to less than 25 to 50 ppm CO occur among the general population, and, at
16      the low exercise levels usually engaged in under such circumstances, resulting COHb levels
17      typically remain below 2 to 3% among nonsmokers.  Those levels can be compared to the
18      physiological baseline for nonsmokers, which is estimated to be in the range of 0.3 to 0.7%
19      COHb. Unfortunately, no new data have become available on the distribution of COHb levels in
20      the U.S. population since large-scale nationwide surveys (e.g., National Health and Nutrition
21      Examination Survey n [Radford and Drizd, 1982]) and human  exposure field studies
22      (e.g., Denver, CO, and Washington, DC  [Akland et al., 1985]) were conducted in the late 1970s
23      and early 1980s.
24          The major source of total exposure to CO for smokers comes from active tobacco smoking.
25      Baseline COHb concentrations in smokers average 4%, with a usual range of 3 to 8% for
26      one- to two-packs-per-day smokers, reflecting absorption of CO from inhaled smoke.
27      Carboxyhemoglobin levels as high as 15% have been reported for chain smokers. Exposure to
28      tobacco smoke not only increases COHb concentrations in smokers, but, under some
29      circumstances, it also can affect nonsmokers. In some of the studies cited in this  document,
30      neither the smoking habits of the subjects nor their exposure to passive smoking have been  taken
31      into account. In addition, as the result of their higher baseline COHb levels, smokers actually

        February 15, 1999                         7-5        DRAFT-DO NOT QUOTE OR CITE

-------
       0)
       Q.
       O
       o
                     A  8 h, 20 L/min
                     O  8h, 10l_/min
                     •  1 h, 20 L/min
                     •  1 h, 10 L/min
                                                                                      100
                                             CO (ppm)
      Figure 7-1. Relationship between CO exposure and COHb levels in the blood. Predicted
                 COHb levels resulting from 1- and 8-h exposures to CO at rest (alveolar
                 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     may be exhaling more CO into the air than they are inhaling from the ambient environment when
2     they are not smoking.  Smokers may even show an adaptive response to the elevated COHb
3     levels, as evidenced by increased red blood cell volumes or reduced plasma volumes.  As a
4     consequence, it is not clear if incremental increases in COHb caused by typical ambient
5     exposures actually would be additive to the chronically elevated COHb levels resulting from
6     tobacco smoke.
      February 15, 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 Conditions               1 h, Light Exercise            8 h, Light Exercise
        Nonsmoking adults                     2 to 3%                      4 to 7%
        exposed to 25- to 50-ppm
        peak CO levels
        Workplace or home                     4 to 5%                     12 to 13%
        with faulty combustion
        appliances  at ~ 100 ppm	
       a See Figure 7-1 for details on the Coburn-Forster-Kane equation (Coburn et al., 1965).
       b Light exercise at 20 L/min.
 1     7.5   MECHANISMS OF CARBON MONOXIDE TOXICITY
 2          A clear mechanism of action underlying the potentially toxic effects of low-level CO
 3     exposure is the decreased oxygen-carrying capacity of blood and subsequent interference with
 4     oxygen release at the tissue level that is caused by the binding of CO with Hb, producing COHb.
 5     The resulting impaired delivery of oxygen can interfere with cellular respiration and cause tissue
 6     hypoxia. The critical tissues (e.g., brain, heart) of healthy subjects have intrinsic physiologic
 7     mechanisms (e.g., increased blood flow and oxygen extraction) to compensate for CD-induced
 8     hypoxia. In compromised subjects, or as CO levels increase, these compensatory mechanisms
 9     may be overwhelmed, and tissue hypoxia, combined with impaired tissue perfusion and
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     where intracellular PO2 decreases with increasing COHb levels.  The most likely hemoprotein 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

       February 15, 1999                          7-7        DRAFT-DO NOT QUOTE OR CITE

-------
 1      exercising muscle.  There is also some evidence that binding of CO to intracellular hemoprotein
 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           The majority of this document deals with the relatively low concentrations of CO that
 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 concentrations
12      of CO. The morbidity and mortality resulting from the latter exposures are described briefly here
13      to complete the picture of CO exposure in present-day society.
14           Carbon monoxide is reported to be the cause of more than half of the fatal poisonings that
15      are reported in many countries.  Fatal cases also are grossly under-reported or misdiagnosed by
16      medical professionals.  Therefore, the precise number of individuals who have suffered from
17      CO intoxication is not known. The symptoms, signs, and prognosis of acute CO poisoning
18      correlate poorly with the  level of COHb measured at the time of hospital admission; however,
19      because CO poisoning is  a diagnosis frequently overlooked, the importance of the early
20      symptoms (headache, dizziness, weakness, nausea, confusion, disorientation, and visual
21      disturbances) has to be emphasized, especially if they recur with a certain periodicity or in certain
22      circumstances. Complications occur frequently in CO poisoning.  Immediate death is most likely
23      cardiac in origin because myocardial tissues are highly  sensitive to the hypoxic effects of CO.
24      Severe poisoning results  in marked hypotension and lethal arrhythmias and other
25      electrocardiographic changes. Pulmonary edema is a fairly common feature. Neurological
26      manifestation of acute CO poisoning includes disorientation, confusion, and coma. Perhaps the
27      most insidious effect of CO poisoning is the delayed development of neuropsychiatric
28      impairment within 1 to 3  weeks and the neurobehavioural consequences, especially in children.
29      Carbon monoxide poisoning during pregnancy results in high risk for the mother by increasing
30      the short-term complication rate and, for the fetus, by causing fetal death, developmental
31      disorders, and chronic cerebral lesions.
        February 15, 1999                          7-8        DRAFT-DO NOT QUOTE OR CITE

-------
 1           The health effects from exposure to low CO concentrations, such as the levels found in
 2      ambient air, are far more subtle and considerably less threatening than those occurring in frank
 3      poisoning from high CO concentrations. Effects of exposure to excessive ambient air levels of
 4      CO are described here in terms of COHb levels; however, the lowest-observed-effect level
 5      depends on the method used for analysis of COHb.  Gas chromatography (GC) is the method of
 6      choice for measuring COHb at saturation levels <5% because of the large variability and
 7      potential high bias of optical methods such as CO-Oximetry (CO-Ox). Health effects are
 8      possible in sensitive nonsmoking individuals exposed to ambient CO if peak concentrations are
 9      high enough or of sufficient duration to raise the COHb saturation to critical levels above their
10      physiological baseline of 0.3 to 0.7% (GC). At 2.3% COHb (GC) or higher, some
11      (predominantly young and healthy) individuals may experience decreases in maximal exercise
12      duration.  At 2.4% COHb (GC) or higher, patients with coronary artery disease (CAD)
13      experience reduced  exercise time before the onset of acute myocardial ischemia, which is
14      detectable either by  symptoms (angina) or by electrocardiographic changes (ST  depression).
15      At 5% COHb (CO-Ox) or higher, some healthy individuals may experience impaired
16      psychomotor performance; however, there is large variability in response across studies that
17      tested the same concentrations of CO, and later work indicates that significant behavioral
18      impairments in healthy individuals should not be expected until COHb levels exceed 20%
19      (CO-Ox). At 6% COHb (CO-Ox) or higher, some people with CAD and high levels of baseline
20      ectopy (chronic arrhythmia) may experience an increase in the number and complexity of
21      exercise-related arrhythmias.
22           Epidemiologic studies have associated elevated ambient CO levels with increased
23      incidence of cardiovascular illness in the population, but overall findings are inconclusive,
24      possibly because personal exposures may not be represented adequately by the CO
25      concentrations measured by fixed-site monitors.  For example, exposure to cigarette smoke or to
26      combustion exhaust gases from small engines and recreational vehicles typically raises COHb
27      levels much higher than levels resulting from mean ambient CO exposures, and, for most people,
28      exposures to indoor sources of CO will exceed controllable outdoor exposures.
29           Health effects are more likely to occur, therefore, in individuals who are physiologically
30      stressed, either by exercise or by medical conditions that can make them more susceptible to low


        February 15, 1999                          7-9         DRAFT-DO NOT QUOTE OR CITE

-------
 1      levels of CO.  The specific subpopulations potentially at risk from exposure to ambient CO are
 2      discussed next.
 3
 4
 5      7.7  SUBPOPULATIONS POTENTIALLY AT RISK FROM EXPOSURE
 6           TO AMBIENT CARBON MONOXIDE
 7           Most of the known quantifiable concentration-response relationships regarding the human
 8      health effects of CO come from two carefully defined population groups:  (1) healthy,
 9      predominantly male, young adults and (2) patients with diagnosed CAD.  On the basis of the
10      effects described, patients with reproducible exercise-induced angina appear to be best
11      established as a sensitive group within the general population that is at increased risk of
12      experiencing the health effects (i.e., decreased exercise duration because of exacerbation of
13      cardiovascular symptoms) of concern at ambient or near-ambient CO-exposure concentrations
14      that result in COHb levels as low as 3%. A smaller sensitive group of healthy individuals
15      experience decreased exercise duration at similar levels of CO exposure, but only during
16      short-term maximal exercise. Decrements in exercise duration in the healthy population,
17      therefore, primarily would be a concern for athletes, rather than for people performing everyday
18      activities.
19           It can be hypothesized, however, from both clinical and theoretical work and from
20      experimental research in laboratory animals, that certain other groups in the population are at
21      potential risk to exposure from CO. Probable risk groups that have not been studied adequately,
22      but that could be expected to be susceptible to CO because of gender differences, aging, or
23      preexisting disease or because of the use of medications or alterations in their environment
24      include fetuses and young infants; pregnant women; the elderly, especially those with
25      compromised cardiovascular function; individuals with partially obstructed coronary arteries, but
26      not yet manifesting overt symptomatology of CAD; those with congestive heart failure (CHF);
27      people with peripheral vascular or cerebrovascular disease; individuals  with hematological
28      diseases (e.g., anemia) that affect oxygen-carrying capacity or transport in the blood; individuals
29      with genetically unusual forms of hemoglobin associated with reduced oxygen-carrying capacity;
30      those with chronic obstructive pulmonary disease; people using medicinal or recreational drugs
31      having central nervous system depressant properties;  individuals exposed to other chemical

        February  15, 1999                         7-10        DRAFT-DO NOT QUOTE OR CITE

-------
 1      substances (e.g., methylene chloride) that increase endogenous formation of CO; and individuals
 2      who have not adapted to high altitude and are exposed to a combination of high altitude and CO.
 3      Little empirical evidence is available by which to specify health effects associated with ambient
 4      or near-ambient CO exposures in these probable risk groups.
 5
 6      7.7.1  Age, Gender, and Pregnancy as Risk Factors
 7          The fetus and newborn infant are theoretically susceptible to CO exposure for several
 8      reasons.  Fetal circulation is likely to have a higher COHb level than the maternal circulation
 9      because of differences in uptake and elimination of CO from fetal Hb. Because the fetus
10      normally has a lower oxygen tension in the blood than does the mother, a drop in fetal  oxygen
11      tension resulting from the presence of COHb could have potentially serious effects. The
12      newborn infant, with a comparatively high rate of oxygen consumption and lower
13      oxygen-transport capacity for Hb than those of most adults, also would be potentially susceptible
14      to the hypoxic effects of increased COHb. Data from laboratory animal studies on the
15      developmental toxicity of CO suggest that prolonged exposure to high levels (>60 ppm) of
16      CO during gestation may produce a reduction in birth weight, cardiomegaly, and delayed
17      behavioral development. Human data are scant and more difficult to evaluate, and further
18      research is warranted. Additional studies are needed to determine if chronic exposure to CO,
19      particularly at  low, near-ambient levels, can compromise the already marginal conditions existing
20      in the fetus and newborn infant.  The effects of CO on maternal-fetal relationships are not well
21      understood.
22          In addition to fetuses and newborn infants, pregnant women also represent a susceptible
23      group because pregnancy is associated with increased alveolar ventilation and an increased rate
24      of oxygen consumption that serves to increase the rate of CO uptake from inspired air.  Perhaps a
25      more important factor is that pregnant women experience an expanded blood volume associated
26      with hemodilution and thus are anemic because of the disproportionate increase in plasma
27      volume compared with erythrocyte volume. This group may be at increased risk and, therefore,
28      should be studied to evaluate the effects of ambient CO exposure and elevated COHb levels.
29          Changes in metabolism with age may make the aging population particularly susceptible to
30      the effects of CO.  Maximal oxygen uptake declines with age.  The rate of decline varies widely

        February 15, 1999                          7-11        DRAFT-DO NOT QUOTE OR CITE

-------
 1      among individuals because of the many confounding factors such as heredity, changes in body
 2      mass and composition, and level of fitness.
 3
 4      7.7.2 Preexisting Disease as Risk Factors
 5      7.7.2.1 Subjects with Coronary Heart Disease
 6           Coronary heart disease (CHD) remains the major cause of death and disability in
 7      industrialized societies.  In the United States, CHD is the single largest killer of males and
 8      females, causing a total of 481,000 deaths in 1995 (American Heart Association, 1997),
 9      two-thirds of all deaths from heart disease (U.S. Centers for Disease Control and Prevention,
10      1997) and about half of all deaths from cardiovascular disease (see Figure 7-2). Almost
11      14 million Americans have a history of this disease, with much greater prevalence in both males
12      and females at increasing ages (see Figure 7-3). Individuals with CHD have myocardial
13      ischemia, which occurs when the heart muscle receives  insufficient oxygen delivered by the
14      blood. For some, exercise-induced angina pectoris (chest pain) can occur.  In all patients with
15      diagnosed CAD, however, the predominant type of ischemia, as identified by ST segment
16      depression, is asymptomatic (i.e., silent).  In other words, patients who experience angina
17      typically have additional ischemic episodes that are asymptomatic.  Unfortunately, some
18      individuals in the population have coronary artery disease but are totally asymptomatic and,
19      therefore, do not know they are potentially at risk. It has been estimated that 5% of middle-aged
20      men show signs of ischemia during an exercise stress test; a significant number of these men will
21      have angiographic evidence of CAD. Persons with both asymptomatic and symptomatic CAD
22      have a limited coronary flow reserve and, therefore, will be sensitive to a decrease in
23      oxygen-carrying capacity induced by CO exposure. In addition, CO might exert a direct effect on
24      vascular smooth muscle, particularly in those individuals with an already damaged vascular
25      endothelium.
26
27      7.7.2.2 Subjects with Congestive Heart Failure
28           Congestive heart failure is a major and growing public health problem.  Almost 5 million
29      Americans have CHF, and about 400,000 new cases occur each year (American Heart
30      Association,  1997). Because the prevalence of heart failure is known to increase with age,

        February 15, 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     improvements in the average life expectancy of the general population would be expected to
2     increase the magnitude of the problem over the next few decades.
3           Patients with CHF have a markedly reduced circulatory capacity and, therefore, may be
4     very sensitive to limitations in oxygen-carrying capacity.  Thus, exposure to CO will reduce their
5     exercise capacity and, especially because of its arrhythmogenic activity, may have serious
6     consequences.  The etiology of heart failure is diverse, but the most common secondary
7     conditions observed in hospitalized patients are coronary heart disease, hypertension, chronic
8     obstructive pulmonary disease, diabetes, and cardiomyopathy (Croft et al., 1997).  Many heart
9     failure patients with CAD, therefore, might be even more sensitive to CO exposure.
      February 15, 1999
7-13
DRAFT-DO NOT QUOTE OR CITE

-------
          c
          Q.
          O
                 80
                 60
                                                                                        79
                 40
          I
          (D
          CL
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 Other Vascular Diseases
 2          Vascular disease, including cerebrovascular disease, is present in both the male and female
 3     population and is more prevalent above 65 years of age because of the increasing likelihood of
 4     adverse effects from atherosclerosis or thickening of the artery walls.  Atherosclerosis is a
 5     leading cause of many deaths from heart attack and stroke (American Heart Association, 1997).
 6     In fact, when considered separately from other cardiovascular diseases, stroke ranks as the third
 7     leading cause of death behind heart disease and cancer (U.S. Centers for Disease Control and
 8     Prevention, 1997).  Vascular diseases are associated with a limited blood flow capacity and,
 9     therefore, patients with these diseases  should be sensitive to CO exposure.  It is not clear,
10     however, how low levels of exposure to CO will affect these individuals. For example, only one
       February 15, 1999
                                7-14
DRAFT-DO NOT QUOTE OR CITE

-------
 1      study reviewed in the previous criteria document (U.S. Environmental Protection Agency, 1991),
 2      has been reported on patients with peripheral vascular disease.  Ten men with diagnosed
 3      intermittent claudication experienced a significant decrease in time to onset of leg pain when
 4      exercising on a bicycle ergometer after breathing 50 ppm CO for 2 h (2.8% COHb).  Further
 5      research is needed, therefore, to better determine the sensitivity of individuals with vascular
 6      disease to CO.
 7
 8      7.7.2.4 Subjects with Anemia and Other Hematologic Disorders
 9           Clinically diagnosed low values of Hb, characterized as anemia, are a relatively prevalent
10      condition throughout the world.  If the anemia is mild to moderate, an inactive person is often
11      asymptomatic. However, because of the limitation in the oxygen-carrying capacity resulting
12      from the low Hb values, an anemic person should be more sensitive to low-level CO exposure
13      than would be a person with normal Hb levels.  If anemia is combined with other prevalent
14      diseases, such as CAD, the individual also will be at an increased risk to CO exposure. Anemia
15      is more prevalent in women and in the elderly, two already potentially high-risk groups.
16      An anemic person also will be more sensitive to the combination of CO exposure and high
17      altitude.
18           Individuals with hemolytic anemia often have higher baseline levels of COHb because the
19      rate of endogenous CO production from heme catabolism is increased.  One of the many causes
20      of anemia is the presence of abnormal Hb in the blood. For example, in sickle-cell disease, the
21      average lifespan of red blood cells with abnormal hemoglobin S is 12 days compared to an
22      average of 88 days in healthy individuals with normal Hb. As a result, baseline COHb levels can
23      be as high as 4%. In subjects with Hb Zurich, where affinity for CO is 65 times that of normal
24      Hb, COHb levels range from 4 to 7%.  Presumably, exogenous exposure to CO, in conjunction
25      with higher endogenous CO levels, could result in critical levels of COHb. However, it is not
26      known how ambient or near-ambient levels of CO would affect individuals with these disorders.
27
28      7.7.2.5 Subjects with Obstructive Lung Disease
29           Chronic obstructive pulmonary disease (COPD) is a prevalent disease especially among
30      smokers, and a large number (>50%) of these individuals have limitations in their exercise
31      performance demonstrated by a decrease in oxygen saturation during mild to moderate exercise.

        February 15, 1999                         7-15         DRAFT-DO NOT QUOTE OR CITE

-------
 1     As a consequence, individuals with hypoxia resulting from COPD such as bronchitis and
 2     emphysema may be susceptible to CO during submaximal exercise typical of normal daily
 3     activity. In spite of their symptoms, many of them («30%) continue to smoke and already may
 4     have COHb levels of 4 to 8%. The COPD patients with hypoxia are also more likely to have a
 5     progression of the disease resulting in severe pulmonary insufficiency, pulmonary hypertension,
 6     and right heart failure.
 7          Hospital admissions for asthma have increased considerably in the past few years,
 8     particularly among individuals less than 18 years of age. Because asthmatics also can experience
 9     exercise-induced airflow limitation, it is likely that they also would experience hypoxia during
10     attacks and be susceptible to CO. It is not known, however, how  exposure to CO actually would
11     affect these individuals.
12
13     7.7.3  Subpopulations at Risk from Combined Exposure to Carbon Monoxide
14            and Other Chemical Substances
15     7.7.3.1 Interactions with Drugs
16          There is an almost complete absence of data on the possible toxic consequences of
17     combined CO exposure and drug use.  Because of the diverse classes of both cardiovascular and
18     psychoactive  drugs, and the many other classes of drugs that have not been examined at all, it
19     must be concluded that this is an area of concern that is difficult to meaningfully address at the
20     present time.
21
22     7.7.3.2 Interactions with Other Chemical Substances in the Environment
23          Besides direct exposure to ambient CO, there are other chemical substances in the
24     environment that can lead to increased COHb saturation when inhaled. Halogenated
25     hydrocarbons used as organic solvents undergo metabolic breakdown by cytochrome P-450 to
26     form CO and inorganic halide. Possibly the greatest concern regarding potential risk in the
27     population comes from exposure to one of these halogenated hydrocarbons,  methylene chloride,
28     and some of its derivatives that could result in potentially harmful levels of COHb in individuals
29     at risk.
30
31

       February 15,  1999                        7-16         DRAFT-DO NOT QUOTE OR CITE

-------
 1      7.7.4 Subpopulations Exposed to Carbon Monoxide at High Altitudes
 2           For patients with CAD, restricted coronary blood flow limits oxygen delivery to the
 3      myocardium. Carbon monoxide also has the potential for compromising oxygen transport to the
 4      heart.  For this reason, such patients have been identified as the subpopulation most sensitive to
 5      the effects of CO. A reduction in PO2 in the atmosphere, as at high altitude, also has the
 6      potential for compromising oxygen transport.  Therefore, patients with coronary artery disease
 7      who visit higher elevations may be unusually sensitive to the added effects of atmospheric CO.
 8           It is important to distinguish between the long-term resident of high altitude and the newly
 9      arrived visitor from low altitude.  Specifically, the visitor will be more hypoxemic than the fully
10      adapted resident. The combination of high altitude with CO will pose the greatest risk to persons
11      newly arrived at high altitude who have underlying cardiopulmonary disease, particularly because
12      they are usually older individuals.
13           It is known that low birth weights occur both in infants born at altitudes above 6,000 ft and
14      in infants born near sea level, whose mothers had elevated COHb levels because of cigarette
15      smoking. It also has been shown that COHb levels in smokers at high altitude are higher than
16      those in smokers at sea level.  Although it is probable that the combination of hypoxic hypoxia
17      and hypoxia resulting from ambient exposure to CO could reduce birth weight further at high
18      altitude and possibly modify future development, no data are available to evaluate this
19      hypothesis.
20
21
22      7.8   CONCLUSIONS
23           Ambient CO concentrations measured at central, fixed-site monitors in metropolitan areas
24      of the United States have decreased significantly since the late 1980s, when air quality was
25      reviewed in the previous criteria document (U.S. Environmental Protection Agency, 1991).  The
26      decline in ambient CO follows approximately the decline in motor vehicle emissions.  Exposure
27      to tobacco smoke, to CO indoors from unvented or partially vented combustion appliances, and
28      to CO from uncontrolled outdoor sources (e.g, small combustion engines) may represent a
29      significant portion of an individual's total CO exposure. Unfortunately, there is not a good
30      estimate of CO exposure distribution for the current population.

        February 15,  1999                         7-17        DRAFT-DO NOT QUOTE OR  CITE

-------
 1          Health assessment information provided in the present document supports and substantiates
 2     the conclusions drawn in the previous document.  Young, healthy individuals are not at
 3     immediate risk from ambient CO exposure because only limitations at maximal exercise
 4     performance have been demonstrated at <5% COHb saturation levels. Effects have not been
 5     demonstrated on heathy individuals performing submaximal exercise that is more germane to
 6     typical daily human activity.  The greater concern regarding the effect of ambient exposure has,
 7     therefore, focused on subpopulations who are particularly susceptible to CO.  The air quality
 8     criteria used to support the existing CO NAAQS for the general population were primarily those
 9     data obtained from the exposure to CO of nonsmoking subjects with CAD while exercising. The
10     principal cause of CO-induced effects at low levels of ambient exposure is still thought to be
11     increased COHb formation and the subsequent impaired delivery of oxygen to critical tissues.
12     Available health effects information on the subpopulation with CAD has identified adverse
13     effects with CO exposures, leading to COHb levels of 2.4% (GC) or higher.
14
       February 15, 1999                          7-18        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,
  4             1994.Hyattsville, MD: U.S. Department of Health and Human Services; publication no. 96-1521. Vital
  5             Health Stat. 10(193). Available online at: http://www.cdc.gov/nchswww/products/pubs/pubd/series/ser.htm.
  6       Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
  7             monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
  8             Technol.  19:911-918.
  9       American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
10             substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
11             Governmental Industrial Hygienists.
12       American Heart Association. (1997) 1998 heart and stroke statistical update. Dallas, TX: American Heart
13             Association. Available online at: www.amhrt.org/Scientific/HSstats98/index.html.
14       Coburn, R. F.; Forster, R. E.; Kane, P. B. (1965) Considerations of the physiological variables that determine the
15             blood carboxyhemoglobin concentration in man. J. Clin. Invest. 44: 1899-1910.
16       Code of Federal Regulations. (1998) Subpart Z—toxic and hazardous substances; air contaminants. Table Z-l,
17             limits for air contaminants, carbon monoxide and ozone. C. F. R. 29: §1910.1000, table Z-l. Available online
18             at: http://www.osha-slc.gov/OshStd_data/1910_1000_TABLE_Z-l.html.
19       Collins, J. G. (1997) Prevalence of selected chronic conditions: United States, 1990-1992. Hyattsville, MD: U.S.
20             Department of Health and Human Services; publication no. 97-1522. Vital Health Stat.  10(194). Available
21             online at: http://www.cdc.gov/nchswww/products/pubs/pubd/series/ser.htm.
22       Croft, J. B.; Giles, W. H.; Pollard, R. A.; Casper, M. L.; Anda, R. F.; Livengood, J. R. (1997) National trends in the
23             initial hospitalization for heart failure. J. Am. Geriatr. Soc. 45: 270-275.
24       Federal Register. (1989) Air contaminants; final rule, carbon monoxide. F. R. (January 19) 54: 2651-2652.
25       Federal Register. (1994) National ambient air quality standards for carbon monoxide-final decision. F. R.
26             (August 1) 59: 38906-38917.
27       National Center for Health Statistics. (1995) Health, United States, 1994. Hyattsville, MD: U.S. Department of
28             Health and Human Services, Public Health Service; publication no. 95-1232. Available online at:
29             http://www.cdc.gov/nchswww/products/pubs/pubd/hus/2010/2010.htm.
30       Radford, E. P.; Drizd, T. A. (1982) Blood carbon monoxide levels in persons 3-74 years of age: United States,
31             1976-80.  Hyattsville, MD: U.S. Department of Health and Human Services, National Center for Health
32             Statistics; DHHS publication no. (PHS)  82-1250. (Advance data from vital and health statistics: no. 76).
33       U.S. Centers for Disease Control and Prevention. (1997) Mortality patterns—preliminary data, United States, 1996.
34             Morb. Mortal. Wkly. Rep. 46: 941-944.
35       U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide.  Research Triangle Park,
36             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
37             no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
         February 15, 1999                              7-19          DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29




AC
AER
AIRS
AQCD
BTPS
C
CAA
CAD
CaO2
CARB
CAS
CASAC
CFD
CFK
cGMP
CH3
CH4
CH3Br
CH3CC13
CH3CHO
CH3C1
CH3CO
CH2O
CH3O2
CH3OOH
                              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-Foster-Kane
                   Cyclic guanosine monophosphate
                   Methyl radical
                   Methane
                   Methyl bromide
                   Methyl chloroform
                   Acetaldehyde
                   Methyl chloride
                   Acetyl radical
                   Formaldehyde
                   Methyl peroxy radical
                   Methyl hydroperoxide
February 15, 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
February 15, 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
HOX
H202
HOCO
IR
LOEL
M
M
Mb
MDL
MI
MSA
n
N
N2
NAAQS
NAMS
NASA
                     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



                     Microsomal heme oxygenase



                     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
February 15, 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
Ni(CO)4
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
                     Nickel tetracarbonyl

                     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
February 15, 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
PEM
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
                     Personal exposure monitor



                     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
February 15, 1999
A-5
DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
6
7
TWA
UHC
UV
UVGSH
VA
VD
VMT
W/F
Time-weighted average
Unburned hydrocarbon
Ultraviolet
Unvented gas space heater
Alveolar ventilation
Volume of physiological dead space
Vehicle miles of travel
Wall or floor furnace
      February 15, 1999
                                       A-6
                               DRAFT-DO NOT QUOTE OR CITE

-------