, ' C. I
oEPA
United States
Environmental Protection
Agency
Office of Health and
Environmental Assessment
Washington DC 20460
EPA-600/8-S3-033A
August 1983
External Review Draft
Research and Development
Revised Evaluation of Review
Health Effects
Associated with
Carbon Monoxide
Exposure:
An Addendum to the
1979 Air Quality
Criteria Document for
Carbon Monoxide
Draft
(Do Not
Cite or Quote)
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-600/8-83-033A
August 1983
External Review Draft
Draft
Do Not Quote or Cite
Revised Evaluation of Health
Effects Associated with Carbon
Monoxide Exposure:
An Addendum to the 1979 EPA
Air Quality Criteria Document 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.
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office
Research Triangle Park, NC 27711
-------
DRAFT
8/22/83
CONTRIBUTORS AND REVIEWERS
The following authors contributed to the writing of this addendum.
Dr. Lester D. Grant, Environmental Criteria and Assessment Office,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James A. Raub, Environmental Criteria and Assessment Office,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Vernon A. Benignus, Neurotoxicology Division, Health Effects Research
Laboratory, U.S. Environmental Protection Agency, Building 224H, University
of North Carolina, Chapel Hill, NC 27514.
Dr. David 0. McKee, Strategies and Air Standards Division, Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency,
Durham, NC 27701
A working draft of this addendum was circulated for preliminary peer-review
to the individuals listed below, and comments received were taken into account
in preparing the present draft addendum. The views expressed in the present
addendum should not be taken, however, as representing those of any single
individual or group listed here.
Dr. Richard Ayres
Natural Resources Defense Council
1725 I Street, NW
Suite 600
Washington, DC 20006
Dr. Laurence Fechter
Department of Environmental Health Sciences
School of Hygiene and Public Health
John Hopkins University
615 North Wolfe Street
Baltimore, MD 21205
Dr. George Goldstein,
Clinical Studies Branch, Health
Effects Research Laboratory
U.S. Environmental Protection Agency
Building 224H, University of
North Carolina
Chapel Hill, NC 27514
Dr. Jack Hackney
Environmental Health Service
Rancho Los Amigos Hospital
7601 Imperial Highway
Downey, CA 90242
i i
-------
DRAFT
8/22/83
CONTRIBUTORS AND REVIEWERS
Dr. Stephen M. Horvath
Institute of Environmental Stress
University of California
Santa Barbara, CA 93106
Dr. Victor G. Laties
Environmental Health Sciences Center
University of Rochester, School of Medicine
Rochester, NY 14642
Dr. Pat Mihevic
Institute of Environmental Stress
University of California
Santa Barbara, CA 93106
Dr. John O'Neil
Clinical Studies Branch, Health
Effects Research Laboratory
U.S. Environmental Protection Agency
Building 224H, University of
North Carolina
Chapel Hill, NC 27515
Dr. David S. Sheps
Department of Cardiology
University of North Carolina
School of Medicine
Chapel Hill, NC 27514
Dr. Jaroslav J. Vostal
Biomedical Science Department
General Motors Research Laboratories
Warren, MI, 48090-9055
Dr. Jeames Wagner
Institute of Environmental Stress
University of California
Santa Barbara, CA 93106
i i i
-------
DRAFT
8/22/83
TABLE OF CONTENTS
Page
INTRODUCTION 1
MECHANISMS OF ACTION 2
HEALTH EFFECTS OF LOW LEVEL CO EXPOSURES 6
CARDIOVASCULAR EFFECTS 7
NEUROBEHAVIORAL EFFECTS 11
EFFECTS OF CO EXPOSURE ON FIBRINOLYSIS 14
PERINATAL CO EFFECTS 17
POPULATION AT RISK 19
SUMMARY AND CONCLUSIONS 23
REFERENCES 27
APPENDIX A A-l
APPENDIX B B-l
-------
DRAFT
8/22/83
INTRODUCTION
On April 30, 1971, the Environmental Protection Agency promulgated (36 FR
8186) national ambient air quality standards (NAAQS) for carbon monoxide (CO)
under section 109 of the Clean Air Act. Identical primary and secondary
standards were set at levels of 9 ppm, 8-hour average, and 35 ppm, 1-hour
average, neither to be exceeded more than once per year. The scientific bases
for these standards are contained in the document Air Quality Criteria for
Carbon Monoxide (U.S. Department of Health, Education, and Welfare, March,
1970, AP-62). The 1971 standards were primarily based on work by Beard and
Wertheim (1967) suggesting that low-level CO exposures resulting in carboxy-
hemoglobin (COHb) levels of 2 to 3 percent are associated with impairment of
ability to discriminate time intervals, a central nervous system (CNS) effect.
The revised Air Quality Criteria Document for Carbon Monoxide (U.S. EPA, 1979)
indicated that this study is no longer considered to provide credible evidence
for such CNS effects occurring at 2-3% COHB and, therefore, does not represent
a sound scientific basis for the standard as discussed in an August 18, 1980
EPA proposal notice (45 FR 55066). However, medical evidence published since
1970 indicated at the time of proposal that aggravation of angina and other
cardiovascular diseases may occur at COHb levels as low as 2.7 to 2.9 percent.
Assessment of this and other medical evidence led EPA to propose, on August
18, 1980, retention of the 8-hour primary standard level of 9 ppm and revision
of the 1-hour standard level from 35 ppm to 25 ppm (45 FR 55066).
The 1980 proposal was based in part on several health studies conducted
by Dr. Wilbert Aronow (Aronow et al., 1972; Aronow and Isbell, 1973; Aronow et
al., 1974; Aronow et al., 1974; Aronow and Cassidy, 1975; Aronow et al., 1977;
Aronow, 1978). Based on evaluation of these studies in 1979 by EPA staff,
their expert consultants, and the Agency's Science Advisory Board, it was
concluded that these studies demonstrate human health effects of carbon
monoxide that should be considered by the Agency in reconfirming existing or
proposing new NAAQS for CO. The Aronow studies were an important element in
the judgment of what blood carboxyhemoglobin (COHb) levels represent a health
concern for sensitive individuals. This "critical" range was defined as
2.7-3.0 percent. An additional study by Aronow (1981) later reported findings
-------
DRAFT
8/22/83
suggesting that aggravation of angina symptoms, i.e., small but statistically
significant decreases (~10%) in time to onset of exercise-induced angina, may
occur in angina patients at COHb levels as low as 2.0 percent.
Since the CO standard was proposed by EPA in 1980, news media reports
appearing in early 1983 indicated that the Food and Drug Administration (FDA)
raised questions regarding the technical adequacy of several studies conducted
by Dr. Aronow on experimental drugs, leading to FDA rejection of use of the
drug study data. While there was no specific direct evidence that similar
problems might exist for the CO studies conducted by Dr. Aronow, EPA judged
that an independent assessment of these studies was advisable prior to a final
NAAQS decision on CO. An expert committee was empaneled by EPA and met with
Dr. Aronow to discuss his studies and to examine limited available data and
records from his CO studies. In their report, the committee (chaired by Dr.
Stephen M. Horvath, Director of the Institute of Environmental Stress, Uni-
versity of California-Santa Barbara) concluded that EPA should not rely on Dr.
Aronow1s data due to concerns regarding problems associated with the studies
which substantially limit the validity and usefulness of those study results
(Horvath et al., 1983).
The main purpose of the present addendum is to re-evaluate the scientific
data base concerning health effects associated with exposure to CO at ambient
or near ambient exposure levels, in light of the diminished value of the
Aronow studies and taking into account any new findings that have become
available beyond those reviewed in the revised Air Quality Criteria Document
for CO (U.S. EPA, 1979). This addendum is, accordingly, organized to provide:
(1) a concise summary of key health effects information discussed in the 1979
document as pertinent to characterization of health effects associated with
relatively low level CO exposures; and (2) an overview of the limited new
evidence bearing on the subject which has become available in the past several
years.
MECHANISMS OF ACTION
The 1979 Criteria Document discussed extensive evidence indicating that
the binding of CO to hemoglobin, producing COHb and decreased oxygen carrying
capacity, results in decreased oxygen transport and uptake in most body tissues.
The resulting hypoxic state (impairing normal biochemical-physiological cellular
-------
DRAFT
8/22/83
processes as a function of increasing external CO exposure and consequently
increasing blood COHb concentrations), it was concluded, probably represents
the main mechanism of action underlying the induction of toxic effects by low
level CO exposures.
Several important relationships between COHb levels and other physio-
logical parameters discussed in the 1979 Criteria Document have continued to
be the subject of evaluation since then. Of much importance is the relation-
ship between external CO exposure levels and consequent increases in blood
COHb levels. Many factors, discussed in the 1979 Criteria Document, can
affect the rate at which COHb increases above pre-existing endogenous levels
of COHb in response to inhalation of exogenous CO. These include, for example,
the pattern of external CO exposure, as in the case of acute short-term exposures
to high CO concentrations versus longer term exposure to relatively low levels
of CO. CO exposure-COHb concentration relationships have been modeled by
Coburn (Coburn et al. , 1965), taking into account several pertinent factors
(see Chapter 9 of the 1979 EPA CO Criteria Document for discussion of the
Coburn model equations). COHb levels predicted by the Coburn equations, as
depicted in Figure 1, are widely accepted as the currently best available
modeled estimates of COHb levels likely to result from varying CO concentra-
tions, exposure durations and exercise levels. It should be noted that some
questions have been raised regarding the specific mathematical approach employed
by Coburn in solving his equations to predict COHb concentration as a function
of time, considering appropriate physiological parameters (Venkatram and
touch, 1979; Ott and Mage, 1980; Marcus, 1980; Joumard et al., 1981). However,
the proposed alternative approaches yield very similar estimated COHb levels
to those projected by Coburn1s approach. In addition, actual blood COHb
concentrations observed in response to particular external CO exposure situa-
tions have been consistent with those predicted by Coburn (Peterson and Stewart,
1975), although further experimental verification would be useful to demonstrate
that the Coburn equation accurately predicts uptake and excretion of CO under
widely varying conditions.
The exact mechanisms responsible for the hypoxia induced by CO are not
known. The most widely accepted mechanism of CO toxicity has been attributed
to the preferential binding of CO to hemoglobin which produces hypoxia by re-
ducing 09 transport by red blood cells to the tissues and impedes the dis-
-------
DRAFT
8/22/83
20 LP« U m CD
10 im to PPW co
20 tP« 20 PPH CO
10 1PM 20 PPH CO
20 LPH XO PPM CO
10 LPH 10 PPM CO
Figure 1. Blood COHb concentrations predicted by Coburn equations to occur as
a function of exposure duration and ambient carbon monoxide con-
centrations under resting (10 LPM), light exercise (20 LPM), or
heavy exercise (50 LPM) conditions. LPM = liters perjninute venti-
lation rate. Source: U.S. EPA (1979).
-------
DRAFT
8/22/83
sociation of 0^ from hemoglobin in the capillaries. However, other mechanisms
have been postulated for reducing oxygen transport. It is possible for CO to
bind to intracellular hemoproteins such as myoglobin and cytochrome oxidase,
which depends on the relationship of oxygen tension (POp) and CO tension (PCO)
to CO binding constants (Coburn, 1979). The affinity of cytochrome oxidase
for CO is similar to that for oxygen compared to myoglobin (30-50x) and hemo-
globin (220x) which would make it less likely to be responsible for impairment
of facilitated diffusion of oxygen to the mitochondria. However, if steep
oxygen tension gradients exist between the extracellular and intracellular
environment, then the P02 surrounding the mitochondria! terminal oxidase would
be low enough to have increased binding with CO. This hypothesis was tested
by Coburn (1979) in studies on isolated vascular smooth muscle. He concluded
that significant CO binding to cytochrome oxidase was unlikely to be an i_n
vivo mechanism of CO toxicity in that particular tissue. Myoglobin was also
unlikely because it is absent or present in only small quantities. It is
possible that CO binds to hemoproteins other than hemoglobin, myoglobin, or
cytochrome oxidase. Cytochrome P-450, tryptophan deoxygenase, and tryptophan
catalase all have high enough binding affinities for CO in specific tissues to
be considered as possible candidates (Coburn, 1979).
The binding of CO to myoglobin in heart and skeletal muscle may be high
enough to reduce intracellular oxygen transport in those tissues (Coburn,
1979; Agostoni et a!., 1980). Using a computer simulation of a three-
compartment model (arterial blood, venous capillary blood, and tissue myo-
globin), Agostoni et al. (1980) predicted that conditions would be favorable
for formation of carboxymyoglobin at COHb levels of 5-10%, particularly in
areas where the P02 was physiologically low (e.g., in subendocardium) and when
conditions of hypoxia, ischemia, or increased metabolic demands were present.
This could provide theoretical support for experimental evidence of myocardial
ischemia, such as electrocardiographic irregularities and decrements in work
capacity discussed later. However, it is not known whether binding of CO to
myoglobin could cause health effects (e.g., decreases in maximal oxygen con-
sumption during exercise) occurring at COHb levels as low as 4-5%. Additional
research is needed before this possibility can be more definitively evaluated.
Whatever the specific biochemical-molecular mechanisms involved in the
induction of CO toxicity in specific tissues or organ systems, it is thought
-------
DRAFT
8/22/83
that COHb concentrations in the blood represent a meaningful and useful
physiological marker by which to gauge the internal CO dose present at a given
time due to the combined cumulative contributions of: (1) baseline endogenous
production of CO by internal body processes; and (2) added CO body burden(s)
resulting from inhalation exposure to exogenous sources of CO.
HEALTH EFFECTS OF LOW-LEVEL CO EXPOSURES
In evaluating CO-induced health effects in humans, a crucial question
that must be addressed concerns blood COHb levels demonstrated to be associated
with effects on specific organ systems. The COHb levels of concern vary with
the patterns and concentrations of external CO exposure pertinent to the
objectives of any particular health risk evaluation. For example, as discussed
in the 1979 CO Criteria Document (U.S. EPA, 1979), some occupations inherently
involve exposure to high levels of CO, at times including frequent exposures
to CO levels well in excess of 100-200 ppm under certain workplace conditions.
On the other hand, it is relatively unusual for members of the general public
to encounter such high CO exposure levels.
As examples of possible extreme CO exposure situations encountered by the
general public, the 1979 Criteria Document noted that the following scenarios
may result in exposure to unusually high ambient levels of CO: (1) On a large
city freeway where traffic has come to a halt, the ambient CO level may exceed
44 ppm; (2) Inside a closed automobile where cigarettes are being smoked, CO
concentrations may exceed 87 ppm; (3) In enclosed, unventilated garages, CO
levels in excess of 100 ppm have been found; and (4) In a heavily traveled
vehicular tunnel, a 1-hour maximum of 218 ppm CO was recorded. Under relatively
mild exercise conditions likely to occur in such exposure situations for any
sustained period of time, the Coburn equations predict COHb blood concentrations
of <10 percent, assuming exposure durations of less than 8 hrs. More often,
the general (non-smoking) population is exposed to substantially lower CO
levels (<20-50 ppm) sustained over 1 to 8 hr. periods, potentially resulting
in maximum COHb levels of 6-7 percent but much more often in COHb levels below
2-3 percent. The present evaluation is, therefore, focused mainly on health
effects observed at blood COHb levels below 10 percent. The latter COHb
-------
DRAFT
8/22/83
levels are most pertinent for present objectives, i.e., the development of
criteria for ambient air quality standards.
Since the writing of the 1979 Criteria Document, several new studies have
been published which contribute to the CO health effects data base. These new
studies, including both human and animal toxicology data, are concisely reviewed
below within a context of integrating the new information with evidence previou-
sly reviewed in the 1979 Criteria Document.
1. Cardiovascular Effects
The most extensive studies on the cardiovascular effects of CO have been
those involving maximum aerobic capacity (V0? ). Previous data reviewed in
the revised CO Criteria Document (U. S. EPA, 1979) demonstrated statistically
significant decreases in VOp when COHb levels ranged from 7-20% under con-
ditions of short-term maximal exercise (Ekblom and Huot, 1972; Pirnay et al.,
1971; Vogel and Gleser, 1972). In another study (Horvath et al. , 1975), the
critical level at which COHb marginally influenced (P < 0.10) VOp was
approximately 4.3%. In this study, work time to exhaustion was also reduced by
4.9 and 7% when COHb levels had attained 3.3 and 4.3%, -respectively. Reductions
in VXL following exhaustive treadmill exercise have since been confirmed at
5% COHb. In a double blind experiment (Stewart et al., 1978; Klein et al., 1980),
6 physically fit male fire fighters were randomly exposed to either CO or fil-
tered air twice a week for 3 weeks and exercised to exhaustion. On exposure
days, the subjects breathed a bolus of 20,000 ppm CO for 47 seconds followed by
30 ppm CO for 4 hours. This resulted in a sustained elevation of COHb at 5.0-
5.5% saturation for 4 hours. Similar decrements in maximal exhaustion times were
noted following the acute exposure and at the end of 4 hours. No adaptation to
this hypoxic stress was observed after 4 hours of exposure or over the 3 weeks
of testing. Significant decreases in total exercise time (3.8%) and VO, (3%)
were also previously reported by Weiser et al. (1978) in a project designed to
determine the effect of 5% COHb exposure on healthy young men residing in Denver
at an altitude of 1610m. Blood COHb concentrations were -quickly increased to
5.1% from a resting level of 1.0% by adding a bolus of 100% CO to a closed-
circuit system from which the subjects rebreathed. The decrement in VO- x was
consistent with those reported above and the authors concluded that changes in
-------
DRAFT
8/22/83
exercise performance following CO exposure at this altitude were similar to,
but not greater than, changes occurring at sea level.
The effects of lower CO exposure levels have also been investigated under
conditions of short-term maximum exercise duration (Drinkwater et al., 1974;
Raven et al. , 1974a,b). In this series of studies, a walking test with pro-
gressively increasing grade was used on subjects continuously breathing 50 ppm
CO at either of two ambient temperatures, 25°C or 35°C, with a relative humidity
of 20%. The two populations consisted of young (23+ years) and middle-aged
(48+ years) subjects, both smokers and nonsmokers. During the duration of the
test, COHB levels in nonsmokers increased from 0.7 to approximately 2.8%,
while levels in smokers rose from 2.6-3.2% to 4.1-4.5%. Control studies
conducted on these subjects while they breathed filtered air indicated that
COHb decreased in both smokers and nonsmokers in the absence of experimental
CO exposures. These studies did not find any reduction in maximum aerobic
capacity. In fact, the only statistically significant effect related to CO
was a statistically significant decrease (<5%) in absolute exercise time con-
sistently observed in the nonsmoking subjects but not in the smokers. These
observations extend those found earlier by Ekblom and Huot (1972), who reported
a large decrease (38%) in work time at 7% COHb.
The revised CO Criteria Document (U. S. EPA, 1979) also noted that oxygen
uptake during short exposures and submaximal work was apparently not affected
even at COHb concentrations of 15-20%. Recently, DeLucia et al. (1983) reported
that COHb levels of 7.3% in nonsmokers and 9.3% in smokers did not induce any
effects involving subjective symptoms, pulmonary function, exercise metabolism,
or blood parameters in healthy subjects. In a controlled experimental study
designed to test for potential synergism between 03 and CO, 24 male and female
volunteers were evaluated while performing moderate aerobic exercise at 50% of
V02 . After 100 ppm CO exposure for approximately 1 hr., COHb levels in non-
smokers rose from 1.0-2.1% to 6.0-9.6% and COHb levels in smokers rose from
1.9-5.1% to 6.6-11.8%. DeLucia et al. (1983) attributed the lack of CO effects
as possibly being due to large cardiovascular reserves found in healthy subjects.
It should be noted that healthy young subjects were used in most of the
above studies evaluating the effects of CO on work capacity. A recent study
by Calverley et al. (1981) demonstrated a decrease in walking distance in 15
-------
DRAFT
8/22/83
patients with severe chronic bronchitis and emphysema at a mean COHb concen-
tration of 12.3%. They evaluated 11 men and 4 women with severe reversible
airway obstruction [FEV1 Q = 0.56 ± 0.2 (SD) liters; FVC = 1.54 ± 0.4 (SD)
liters] who were hypoxic [Pa02 = 5.2 ± 4.9 (SD) mm Hg]. All patients were
medically stable at the time of the study and smokers were asked to stop
smoking for 12 hours before each session. Each subject walked while breathing
air and oxygen before and after exposure to 200 ppm CO in air, which raised
their COHb concentrations from 1.1-5.4% to 9.6-14.9%. There was a significant
reduction in walking distance when the patients breathed either air or oxygen
after exposure to CO. A significant increase in walking distance when the
patients breathed oxygen after exercise was abolished by CO exposure. There
was no relationship to normal smoking habits of the patients. It is therefore
quite possible that individuals with hypoxia due to bronchitis or emphysema
are more susceptible to CO during submaximal work loads typical of everyday
exercise.
Other cardiovascular effects of CO are thought to be of greater concern,
i.e., those affecting individuals suffering from chronic angina. However, the
precise COHb levels at which such cardiovascular effects occur in angina patients
are much less well defined than COHb concentrations associated with various
health effects discussed above. Angina pectoris is a cardiovascular disease in
which mild exercise or excitement produces symptoms of pressure and pain in the
chest because of insufficient oxygen supply to the heart muscle. Angina patients
exposed to low levels of CO while resting have been reported to exhibit statis-
tically significantly reduced time to onset of exercise-induced angina at mean
COHb levels of 2.9 (range, 1.3-3.8 percent) and 4.5 percent (range, 2.8-5.4
percent) and to experience significantly increased duration of angina attacks
during exercise at a mean COHb level of 4.5 percent (Anderson et a!., 1973).
Certain questions have been raised regarding the design and conduct of the
study, the small number (N=10) of subjects studied, and the absence of credible
independent confirmation of its findings. A thorough reevaluation of the
Anderson et al. (1973) study, addressing major points of concern, has been con-
ducted recently (see Appendix A), and found that the study, in fact, provides
reasonably good evidence for the hastening of angina occurring in angina patients
at COHb levels of 2.9 to 4.5 percent. The Aronow et al. (1973) and Aronow (1981)
studies similarly reported decreased time to onset of angina in exercising patients
at COHb levels of 2.0-3.0 percent. However, these Aronow findings are now most
-------
DRAFT
8/22/83
appropriately interpreted as providing suggestive evidence for such effects occur-
ring at COHb levels below 3 percent. More conclusive statements regarding this
issue will not be possible until the results of independent studies attempting to
replicate such findings become available.
Another cardiovascular effect of possible concern is that of increased
blood flow that occurs as a compensatory response to CO exposures (Ayres et
al., 1969; Ayres et al., 1970; 1979). This response might result in coronary
damage or other vascular effects due to added stress on the cardiovascular
system. However, inconclusive results have been obtained in community epidemi-
ology studies examining the relationship between CO exposure, mortality from
myocardial infarction (heart attack), sudden death due to arteriosclerotic
heart disease, and cardiorespiratory complaints (Goldsmith and Landau, 1968;
Kurt, et al., 1978; Kurt, et al., 1979). Hence, the possibility of such an
association remains in question, and further research is also needed in order
to clarify this issue.
The results of one controlled human exposure study reported by Davies and
Smith (1980) are suggestive of possible effects on cardiac function at low to
moderate CO exposure in healthy individuals. In a series of replicated experi-
ments, six matched groups of young human subjects lived in a closed-environment
exposure chamber for 18 days. They were exposed continuously to 0, 15, or 50 ppm
of CO in air during the middle 8 days. Standard 12-lead electrocardiograms were
recorded from each subject during the control, exposure, and recovery periods.
Unequivocal P-wave changes were seen during the CO exposure period in 3 of 15
subjects at 15 ppm CO (2.4% COHb) and 6 of 15 at 50 ppm (7.1% COHb) compared to
none of 14 at 0 ppm (0.5% COHb). The changes were evenly distributed among non-
smoking subjects and subjects who had stopped smoking 3 days before the start
of CO exposure. In addition, one subject, later identified as having evidence
of myocardial ischemia, showed marked S-T changes at 15 ppm. In a separate pilot
study by these investigators with both smokers and nonsmokers at 75 ppm CO (10.9%
COHb in nonsmokers; 14.9% COHb in smokers) significant EKG changes were demon-
strated in 7 of 10 subjects. In most cases, the CO-induced changes remained
after exposure ceased. The authors concluded that P-wave abnormalities demon-
strated in this study were due to interference of normal atria! pacemaking or
conducting tissue activity by CO. In addition, they speculate that CO has a
specific toxic effect on the myocardium rather than (or in addition to) a
generalized decrease in 02 transport to the tissue.
10
-------
DRAFT
8/22/83
Most animal studies on the cardiovascular effects of CO have been con-
ducted at exposure concentrations resulting in rather high (>15-20 percent)
COHb levels. Only two studies are relevant for present discussion. Becker
and Haak (1979) exposed 11 adult mongrel dogs to increasing concentrations of
CO 1 hr. after coronary artery ligation. These sequential exposures produced
step-wise increases in the COHb level from 4.9% to 17.0%. Myocardial ischemia,
as indicated by the amount of S-T segment elevation in epicardial electrocar-
diograms, increased significantly at the lowest COHb level and increased
further with increasing CO exposure. These changes occurred in the absence of
altered heart rate, blood pressure, left atria! pressure, cardiac output, or
blood flow to ischemic myocardium. Flow to non-ischemic myocardium increased
with CO exposure at a rate approximately double the increase in COHb. They
concluded that low level exposure to CO can significantly augment ischemia in
acute myocardial infarction, apparently through a reduction in oxygen supplied
to the ischemic tissue. They suggested, however, that the hypoxia induced by
CO was more severe than could be accounted for by a reduction in tissue 0«
delivery alone.
Foster (1981) investigated the arrhythmogenic effects of CO during the
initial minutes of acute myocardial ischemia in 8 mongrel dogs. Since each
dog served as its own control, brief occlusions of the coronary artery were
performed sequentially both before and after the administration of 100 ppm CO
which raised COHb levels to 10.4%. Bipolar epicardial electrograms were
recorded in each experiment from the ischemic and non-ischemic myocardial
zones. An additional 6 dogs were used to confirm the reproducibility of
ischemic conduction slowing during successive occlusions in the absence of CO.
There was no significant increase in ischemic myocardial conduction slowing
after CO. This lack of arrhythmogenic effect of CO was supported by the
absence of increased incidence of spontaneous ventricular tachycardia during
ischemia after CO administration. The author concludes that clinically en-
countered COHb levels may not have sufficient arrhythmogenic effect to be of
significant health importance during the initial minutes of myocardial ischemia.
2. Neurobehavioral Effects
The 1979 CO Criteria Document noted that statistically significant effects
on central nervous system (CNS) functions have been most clearly shown to occur
11
-------
DRAFT
8/22/83
at COHb levels of 5-17 percent. This is indicated by studies which demonstra-
ted decrements in vigilance, visual perception, manual dexterity, learning
ability, and performance of complex sensorimotor tasks such as driving
(Bender et a!., 1971; Schutte, 1973; O'Donnell et a!., 1971; McFarland et al.,
1944; McFarland, 1973; Putz et al., 1976; Salvatore, 1974; Wright et al.,
1973; Rockwell and Weir, 1975; Rummo and Sarlanis, 1974). An evaluative
review by Laties and Merigan (1979) substantially agreed with these conclu-
sions. Also, as reviewed in the 1979 Criteria Document, some studies (Horvath
et al., 1971; Fodor and Winneke, 1972; Groll-Knapp et al., 1972; Putz et al.,
1976) have reported significant decrements in vigilance performance (defined as
the ability to detect small changes in one's external environment occurring at
unpredictable times) to be associated with COHb levels in the 3.0-7.6% range;
and one study by Beard and Grandstaff (1975) reported that vigilance effects may
occur at levels as low as 1.8 percent COHb. The lowest COHb levels at which
vigilance decrements occur, however, are a matter of considerable dispute in
view of numerous other studies not finding such effects at COHb levels below
5.0 percent (Haider et al., 1975; Winneke, 1974; Winneke et al., 1976;
Christensen et al., 1977; Benignus and Otto, 1977).
Since the writing of the 1979 Criteria Document, several new studies have
been published which contribute to the data base of CO-related neurobehavioral
effects. In the following, each of the several neurobehavioral endpoints are
re-evaluated by integrating new findings.
Vigilance—Si nee the 1979 document was written, several new pieces of research
on CO and vigilance have appeared. In an evaluative review of the CO-vigilance
literature, Benignus et al. (1983) concluded that all of the studies (except
one) cited in the 1979 document had serious credibility flaws due to (a) non-
replication of the work,.(b) gross statistical abuse, (c) non-concentration-
related effects or (d) combinations of the above.
The single study from the 1979 document which had no serious flaws was
that of Putz et al. (1976), which demonstrated vigilance decrements at 5%
COHb. Since that time the same group of researchers have replicated these
results on independent subjects (Putz et al. , 1979) and have reported their
earlier results in peer-reviewed literature (Putz, 1979). The fact that the
experimental design and details of the tasks were not the same in the two
12
-------
DRAFT
8/22/83
experiments implies that the effects were robust and thus lends further cre-
dibility to these results. The credibility of these findings would be appre-
ciably increased if an independent group of researchers also were to success-
fully replicate the study.
Three other articles on CO and vigilance have appeared since 1979 (Benignus
et a!., 1983, Davies et al.} 1981, Roche et al., 1981). All of them achieved
COHb levels of 5-7%. None of the studies found significant effects on vigi-
lance. It is noteworthy that all of them used vigilance paradigms different
from those of Putz et al (1976). Quite possibly the parameters and conditions
under which vigilance was studied are sensitive so that unless the proper con-
ditions exist, the effects of such COHb levels will not be detected.
It appears safe to conclude that, at least under some conditions, reliable
but small decrements in vigilance occur at about 5% COHb. The fact that 5% and
higher levels of COHb were not observed to produce vigilance decrements in many
studies is probably a reflection of (a) low experimental test sensitivity in the
those cases or (b) the small effects of CO at these levels coupled with what
is probably a rather low-slope concentration-effects curve in the region of
COHb less than 20%. It must be emphasized that these explanations for the many
no-effect studies are conjectural.
Sensory and Time Discrimination—Benignus et al. (1983) concluded that no
highly reliable evidence for time discrimination decrements exists. The ele-
gant concentration-related decrements in dark adaption beginning at 5% COHb
demonstrated by McFarland et al. (1944) remain to be replicated. The importance
of replication should not be overlooked since the decrements were (a) dose-related
and (b) showed a decrement at the lowest non-zero COHb level.
Davies et al. (1981) tested visual sensitivity in dark adapted subjects
who had been exposed continuously to 50 ppm CO for a total of 5 days (COHb of
7% by the end of each day). They reported no effects of CO. Luria and McKay
(1979) reported that untrained observers showed no decrement in a night vision
test, eye movement or visual evoked potential at COHb levels of 9 percent. The
fact that McFarland et al. (1944) used bolus exposure methods, whereas the
above investigators used continuous low level exposure, is perhaps significant.
McFarland et al. (1944) also used highly trained observers, which had the effect
13
-------
DRAFT
8/22/83
of reducing variance and thus increasing the sensitivity of their study. Cer-
tainly the McFarland et al. (1944) study cannot be said to have been invalidated
by newer data.
Some support for visual sensitivity decrements due to CO has been provided
by a study measuring the electroretinogram in anesthetized cats (Ingenito and
Durlacher, 1979). When cats were exposed to 1000 ppm CO the electroretinogram
was decreased in amplitude by 30 minutes after the start of exposure, at which
time the mean COHb was 7.5%. Further decreases were concentration-related.
Complex Sensorimotor Performance and Driving--The conclusion of the 1979
document, that driving-like tasks are impaired at COHb levels of 5% or greater,
has since been substantially strengthened. Three articles have been published
(Putz et al., 1976, 1979; Putz, 1979), using variations of the same task and
experimental design but in two independent groups of subjects. In all c^ses, it
was reported that 5% COHb produced decrements in compensatory tracking, a hand-
eye-coordination task. In -all cases, however, the decrements occurred only during
high task difficulty. Other tasks such as reciprocal tapping and digit manipula-
tion were not affected by COHb levels of up to 5% (Mihevic et al., 1982).
Sleep and Activity—Although there has been one new publication in this area
(Groll-Knapp et al., 1982), the conclusions remain the same. Marginal increases
in deep sleep and concomitant decreases in rapid-eye-movement sleep were
reported at 8% COHb due to exposure to 100 ppm CO for 8 hrs. during sleep. As
before, the changes did not reach statistical significance when appropriate
corrections are made (Benignus and Muller, 1982).
Central Nervous System Electrical Activity—No significant changes have occurred
in this area of research since 1979. Marginal, nonsignificant effects have
been reported by Benignus et al. (1983) in the electroencephalogram alpha
frequency band and by Groll-Knapp et al. (1982) on evoked potentials. Both
groups of investigators produced COHb levels of 5-6 percent.
3. Effects of CO Exposure on Fibrinolysis
An area of growing interest, which was not extensively discussed in the
1979 Criteria Document, concerns possible effects of CO on fibrinolysis. The
14
-------
DRAFT
8/22/83
fibrinolytic system is an integral part of homeostatic mechanisms and it has
been suggested that derangement of that system may contribute to the patho-
genesis of thrombosis. The plasminogen activator may be released from the
blood vessel wall by a variety of stimuli, including anoxia, electric shock,
and either local or systemic administration of vasoactive substances such as
epinephrine, acetylcholine, serotonin, or histamine. The finding that anoxia
can release plasminogen activator has suggested to investigators that the
presence of COHb with its associated anoxia could also be associated with
increased plasminogen activator.
Resting fibrinolytic activity is often low and its measurement may yield
conflicting results that do not readily permit intra-individual comparisons.
The relationship of tissue activator to vascular activator suggests that they
are components of two separate fibrinolytic mechanisms: (1) the vascular acti-
vator being part of a humoral system whose main role is maintenance of vascular
patency and (2) the basic activator concerned with tissue repair and wound
healing.
Animal Studies—Fibrinolytic activity was studied in 12 rabbits (6 control
rabbits) exposed for 8 weeks to an ambient concentration of 50 ppm CO (Kalmaz
et al., 1977). The control animals had COHb levels of approximately 6%, while
the CO-exposed animals' COHb, for some unexplained reasons, continued to
increase, finally stabilizing at 30.9%. Significant increases in whole blood
clotting time, serum fibrin/fibrinogen degradation products, and acceleration
of whole blood clot lysis occurred. Euglobulin lysis time was significantly
accelerated by the end of the first week of exposure. Although these fibrino-
lytic activity changes are of interest, their relationship to CO exposure is
far from clear.
In another study, Kalmaz et al. (1978) exposed rabbits to 50 ppm for 8
weeks or to 300 ppm for 4 weeks. A second group of rabbits exposed to 300 ppm
CO were given epsilonaminocaproic acid. Acceleration of the whole blood clot
lysis and euglobulin lysis times was observed. Microscopic examination of
large vessels showed endothelial damage - a possible source for a plasminogen
activator release. A more recent study by Kalmaz et al. (1980) involved
exposing rabbits to ambient air, 50 ppm CO for 24 hours/day continuously (8
weeks), and 300 ppm CO for 24 hours/day continuously (4 weeks). They found a
15
-------
DRAFT
8/22/83
consistent change in circulating platelet quantity for CO-exposed rabbits.
However, their conclusion that prolonged exposure to low levels of CO may
influence changes in circulating platelet counts and/or congenital platelet
function disorders in man is yet to be confirmed.
In a recent in-vitro study by Hartiala et al. (1982) 5-minute exposure to
100% CO was found to have neither an effect on the decrease in prostacyclin
(PGI2) production nor a direct effect on ADP-induced aggregability of human
platelet rich plasma. This led the authors to conclude that CO is not respon-
sible for the temporary increase in platelet aggregability after cigarette
smoking.
Human Studies—Workers chronically exposed to high levels of carbon monoxide
(approximately 100 ppm with occasional levels of 200-400 ppm) were studied by
El-Attar and Sairo (1968). They found an accelerated clot lysis time suggestive
of an enhancement of blood fibrinolytic activity. Levels of COHb present were
only crudely determined and were not related to any specific individual. CO
poisoning of 21 workers was suggested by the presence of subjective symptoms,
headache, blurred vision, etc. Twenty-eight workers, presumably some of the
first group of 21, were restudied after one day exposure to ambient CO (levels
not given). This group also exhibited accelerated lysis times but to a lesser
degree. In 15 control subjects no fibrinolytic activity could be detected.
This study, although suggestive, was too poorly controlled to be of real
value. A similar suggestive study was made by Alexieva et al. (1975) on 100
workers in a coke-chemical plant. Some increase in fibrogen was observed in
these chronically exposed individuals. Panchenko et al. (1977) have also
suggested that a relationship between blood coagulation and CO existed. The
data presented are far from conclusive.
The possibility that exposure to other substances in addition to CO
resulted in alterations of fibrinolytic activity was evaluated by Janzon and
Nilsson (1975). Smokers and nonsmokers were studied by techniques superior to
those used by the above investigators. Smokers and nonsmokers had the same
fibrinolytic activity when smokers were studied after 12 hours abstention from
smoking. Smoking 6 cigarettes during 3 hours was associated with an increased
fibrinolytic activity in blood. They believed that this increase was probably
due to the combined effects of nicotine and carbon monoxide. Mansouri and
16
-------
DRAFT
8/22/83
Perry (1982) studied the alteration of platelet aggregation by cigarette smoke
and carbon monoxide. Inhalation by healthy adults of CO sufficient to raise
the COHb level to between 4.5% and 11% was found to be responsible for inhi-
bition of platelet aggregation, which returned to normal after five hours.
However, there was no consistent correlation reported between COHb levels and
alterations in platelet aggregation.
Brinkhouse (1977) exposed 23 men (15 nonsmokers) for 4 hours to either 0,
50, or 100 ppm CO. COHb levels on the days of carbon monoxide exposure reached
2.17% (50 ppm) and 4.15% (100 ppm). Studies on blood coagulation were made by
the best and most sophisticated techniques. Platelet count, prothrombin time,
partial thromboplastin time, thrombin time, fibrin split products, factor VIII,
and platelet aggregation were determined before and after the exposure.
Coagulation parameters were not significantly affected by the CO exposures. A
review by Haft (1979) discusses the role of platelets in the etiology and
natural history of coronary artery disease. It is clear that smoking increases
the activity of platelets and cigarette smokers have shortened platelet survival
time. Endothelial injury may be facilitated by serum CO, by levels of circulating
catecholamines, and other factors.
In conclusion, the effects on fibrinolytic activity of exposure to carbon
monoxide are far from clear. The studies on acute exposure to CO do not spe-
cifically implicate this pollutant in the observed alterations in fibrinolytic
activity, and the studies on chronic exposure are too poorly controlled to
confirm any definite effects on the blood coagulation system.
4. Suggestive Evidence for Perinatal CO Effects
The Criteria Document (U.S. EPA, 1979) provides discussion of results
from certain animal toxicology studies which point toward the possibility of
CO exerting perinatal effects on the fetus or newborn. With long-term exposures
of pregnant animals to CO, fetal COHb levels have been shown to be higher than
maternal COHb levels and fetal elimination of CO was slower than maternal CO
elimination. The ability of CO to decrease the oxygen transport capacity of
maternal and fetal hemoglobin may result in interference in fetal tissue
oxygenation during important developmental stages. Whereas normal adults have
reserve capacity and compensatory responses which enable them to handle moder-
ately high COHb levels without irreversible consequences, the fetus may under
17
-------
DRAFT
8/22/83
normal situations be operating close to critical levels in terms of tissue
oxygen supply. Thus, even moderate CO exposures may have a deleterious effect
on fetal development (Longo, 1977) but this, too, remains to be demonstrated
along with pertinent dose-response relationships. In several animal studies
in which pregnant females were exposed to CO, deleterious effects were gen-
erally reported in the offspring (e.g., reduced birthweight, increased newborn
mortality, and lower behavioral activity levels) even when no effects on the
mothers were detected. In human studies, similar effects have been reported
in children of mothers who smoked cigarettes during pregnancy, suggesting that
expectant mothers and their unborn children may also represent population
groups at special risk for CO effects, but this remains to be more clearly
defined along with any pertinent dose-response relationships.
Sudden infant death syndrome (SIDS) is characterized by the sudden,
normally unexplained death of an infant. Research has suggested numerous
possible etiological factors (e.g., disease, temperature, maternal smoking,
and pollutant levels), but numerous questions remain regarding which are the
most significant factors involved. Seasonal incidence of SIDS has been studied
in several epidemiology studies (Peterson, 1966; Bergman et a!., 1972; Bonser
et a!., 1978). A pattern which appears to be consistent across countries in
the northern hemisphere is increasing incidence in October, peaking in December
and January, remaining high until May, and then declining sharply in June and
remaining low until October.
CO has been hypothesized to be associated with seasonal variations in
SIDS incidence rates, based on certain epidemiologic data. Hoppenbrouwers et
al. (1981) have reported that increased seasonal incidence of SIDS in Los
Angeles County during winter may be at least partially explained by higher
levels of CO, sulphur dioxide (S02), nitrogen dioxide (N02) and hydrocarbons
(HC). They suggest that higher ambient levels of these pollutants may be
implicated in chronic hypoxia which often precedes death from SIDS. Related
to this hypothesis, they found that SIDS cases in LA were correlated to daily
mean levels of the above pollutants and peaks in these pollutant levels pre-
ceded seasonal increases in SIDS by seven weeks. The authors further report
that infants dying from SIDS lived longer (1) if born in low rather than high
pollution areas and (2) if born in months of low versus high pollution.
18
-------
DRAFT
8/22/83
Finally, a direct proportionality was reported between exposure to pollution
for infants from conception to two months of age and bimonthly rate of SIDS.
In an editorial letter assessing the results of the Hoppenbrouwers et al.
study, Goldstein (1982) commented on the presence of indoor sources of CO, NO-
and lead. She further pointed out that during colder months infants spend
most of their time indoors where concentrations of these pollutants can far
exceed ambient levels. Thus, she suggests that the relationship between SIDS
and ambient pollution levels may be only coincidental and the evidence for it
is at best suggestive and in need of further confirmation before any causal
relationships might be inferred.
Maternal smoking has been related to SIDS in several studies (Bergman and
Wiesner, 1976; Lewak et al., 1979; Peterson, 1981). Because CO is only one of
numerous pollutants found in cigarette smoke, however, it is difficult to
infer a causal relationship between CO and SIDS. Other factors associated
with SIDS include passive smoking, younger maternal age, short intervals
between pregnancies, gestational age of less than 40 weeks, birth weight of
less than 3000 g., lower socioeconomic status, and male sex. Thus, the number
.of potentially confounding factors makes finding an association between CO and
SIDS extremely difficult.
POPULATIONS AT RISK
The 1979 Criteria Document directed attention toward identification of
sensitive population groups at special risk for CO-induced health effects.
One key concept in defining special risk groups for CO effects is the idea
that any preexisting or concommitant physiological or pathological condition
which interferes or interacts with oxygen absorption into blood or its transport
to and perfusion of body tissues can logically be expected to exacerbate
CO-induced health effects associated with the hypoxic effects of CO. Thus,
certain large segments of the general population can be reasonably hypothe-
sized as likely to be at greater risk for experiencing CO-induced health effects
than healthy, non-smoking adults.
These probable risk groups include: (1) fetuses and young infants; (2) the
elderly, especially those with reduced cardiopulmonary functions attributable to
a variety of factors associated with typical aging processes; (3) other, younger
19
-------
DRAFT
8/22/83
individuals with overt, severe cardiac damage or acutely severe respiratory
diseases, e.g., pneumonia; (4) individuals with chronic bronchitis or emphysema;
(5) individuals with symptoms (e.g., angina) indicative of chronic cardiovascular
disease; (6) individuals with hematological diseases, e.g., anemia that affect
oxygen carrying capacity or transport in the blood; and (7) persons with geneti-
cally unusual hemoglobin forms associated with decreased oxygen capacity. In
addition to the above, one might reasonably expect that individuals under the
influence of certain drugs, used for recreational or medicinal purposes, may be
at greater risk for CO-induced effects due to interactive effects between CO and
certain pharmacological agents. Lastly, under high altitude conditions (where
reduced levels of atmospheric oxygen exist), increased vulnerability to CO health
effects of both the above sensitive population groups and otherwise, non-sensitive
healthy individuals might be expected.
As noted in the 1979 Criteria Document, relatively little concrete
experimental or observational evidence currently exists by which most groups
(or conditions) listed abo.ve have been clearly demonstrated to be associated
with increased risk for CO-induced health effects. Nor have clear-cut
quantitative lowest-observed-effect levels or dose-response relationships been
delineated for the occurrence of CO effects among the above "at risk" groups
or in conjunction with special interacting circumstances (i.e., drug usage or
high altitude living) that might exacerbate CO effects. Only very limited
discussion, therefore, can be provided here regarding CO risk factors and
sensitive groups likely at special risk for CO effects.
In regard to the first group, fetuses and young infants, certain evidence
was alluded to earlier from animal toxicology studies indicating that higher
levels of COHb and increased CO excretion time occur among fetuses in compari-
son to their mothers exposed to CO; and some deficits in postnatal growth and
development were noted in the offspring of dams exposed to CO during pregnancy.
Also, analogous perinatal effects were noted in human infants born to mothers
who smoked during pregnancy, and associations between CO exposure and SIDS have
been hypothesized. However, insufficient evidence exists at this time by which
to estimate CO exposure levels at which any CO effects on human fetuses or new-
born infants may occur or whether the latter types of effects seen in smoking
mothers are due specifically to CO versus other components of tobacco smoke
either singly or in combination.
20
-------
DRAFT
8/22/83
Little specific evidence directly demonstrates the increased vulnera-
bility of the elderly for CO health effects in precise quantitative terms.
Given the increased vulnerability of the aged to many different kinds of
stress, including thermal stress (hot or cold) which taxes decreased reserve
capacities to maintain adequate cardiovascular delivery of oxygen to body
tissues, it must be expected that CO exposure would render the elderly more
vulnerable to the effects of other types of cardiovascular stresses. Also,
conversely, it is reasonable to hypothesize that interactive effects involving
other stress factors might lead to exacerbation of CO-induced health effects
or their occurrence at lower external CO exposure levels than in younger,
healthy adults. Similarly, younger individuals with overt, severe cardiac
damage or insufficiency or severe acute respiratory diseases, e.g., pneumonia,
can be expected to be more vulnerable to CO (i.e., either CO exacerbation of
other disease effects or, conversely, increased susceptability to CO health
effects) due to reduced reserve capacities to cope with stress generally or
increased sensitivity of already compromised organs or tissuos, e.g., heart
muscle, to the hypoxia induced by CO.
Turning to individuals with chronic bronchitis and emphysema, again,
reduced reserve capacities for dealing with cardiovascular stresses and al-
ready reduced oxygenation of blood should exacerbate or hasten the onset of
health effects associated with CO-induced hypoxia. Analogously, angina
patients or others with obstructed coronary arteries but not manifesting overt
symptoms such as angina, should be at greater risk for CO health effects. Both
the Anderson et al. (1973) and several Aronow publications on angina patients
reported findings previously accepted as demonstrating the increased vulnera-
bility of angina patients to CO in terms of hastening of the onset of exercise-
induced angina. These papers, furthermore, appeared to confidently establish
2-3% blood COHb as the range of COHb values associated with the onset of statis-
tically significant effects indicative of CO exacerbation of angina and, pro-
bably, associated hypoxic effects on cardiac muscle. It is now clear that the
Anderson findings (as of yet not independently confirmed) can be appropriately
interpreted as providing reasonably good evidence for exacerbation of angina
symptoms occurring at approximately 2.9 to 4.5 percent COHb; and the possibility
of such effects occurring at lower COHb levels (as hinted at by the Aronow studies)
cannot be ruled out at this time.
21
-------
DRAFT
8/22/83
In regard to individuals with anemia being at special risk for CO health
effects, the 1979 Criteria Document noted that CO poisoning is similar to
anemia, wherein the oxygen capacity of the blood is decreased because the
affinity of hemoglobin (Hb) for binding oxygen is reduced. The 02 dissociation
curve for anemics is similar to that for normals except that it is shifted to
the right. However, when curves from individuals with 50% reductions in Hb con-
tent are compared with dissociation curves for blood with 50% COHb content, there
are striking differences. Consequently, care must be taken to avoid overly
simplistic extrapolations regarding the likely impact of particular CO
exposures on anemia patients due to additional (CO-induced) reductions of b*
carrying capacity beyond the reductions already evident in their blood. The
soecific CO exposure levels and associated blood COHb concentrations at which
anemia patients may be at increased risk for specific CO health effects remain
to be clearly delineated, but little doubt exists regarding the likelihood
that the effective CO exposure (and COHb) levels are distinctly lower than for
healthy, non-anemic individuals.
Another logically-hypothesized "at risk" group for CO effects are individuals
with unusual hemoglobin types that result in chronic elevations of COHb blood
levels even in the absence of external CO exposure. Normal adult hemoglobin
has a relative affinity or equilibrium constant (M) for CO of about 200 for
most animal species, but has been reported to be as high as 240 to 250 in
humans (Roughton, 1970). There are approximately 350 human hemoglobin variants,
including those found in the fetus, sickle cell anemics, and other individuals
with hemoglobinopathy (hemoglobin disorders). Approximately one fourth of the
hemoglobin variants now known are considered unstable, i.e., they denature and
precipitate when red blood cells or hemolyzates are exposed to heat, red
oxdyes, or isopropanol (Bunn et al., 1977). One of these variants, hemoglobin
Zurich (HbZ) has been found to have an affinity for CO which is approximately
65 times that of normal hemoglobin (Zinkham et al., 1980; Giacometti et al.,
1980; Zinkham et al., 1983). This results in chronic elevation of endogenous
COHb levels ranging from 3.9 to 6.7% in nonsmoking HbZ individuals.
As for possible drug-induced enhancement of CO effects, whereas there
exists little specific data directly supporting the idea, it is logical to
suspect that individuals who use certain drugs would be at increased risk for
experiencing health effects associated with CO exposure. For example, drugs
22
-------
DRAFT
8/22/83
with primary or secondary CMS depressant effects should be expected to exacer-
bate neurobehavioral effects of CO; and drugs which have primary or secondary
cardiac stimulant effects might worsen CQ-related cardiac effects. Any vaso-
constrictive drugs would be expected to reduce 0,, delivery to various organs,
thus also exacerbating CO effects on exercise, cardiac or neurobehavioral
function. Further speculations about drug-CO interactions are possible, but
it should be emphasized that these are predictions based on theoretical grounds
that rely heavily on our current understanding of hypoxia being the likely
main mechanism underlying the induction of CO-induced health effects. Unfor-
tunately, few data currently exist by which to judge the likely validity of
such speculations.
Hypothesized enhancement or exacerbation of CO health effects under high
altitude conditions, similarly, rests heavily on hypoxia being the key mechan-
ism of CO toxicity. The interactive effects of high altitude hypoxia and CO
exposure, however, appear to be more complex than might be simplistically
expected (Collier and Goldsmith, 1983). For example, the 1979 Criteria Document
notes that adaptation to high altitudes occurs and alludes to the fact that,
analogously, adaptation to CO may alter the position of the 02 dissociation
curve as a function of extensive prior CO exposure. Thus, individuals living
in high altitude situations with frequently elevated ambient air CO levels may
adapt to both the high altitude and CO-induced hypoxia, whereas other indivi-
duals newly entering high altitude and elevated ambient CO conditions might
experience CO effects at concentrations below those effective at lower alti-
tudes.
SUMMARY AND CONCLUSIONS
As was stated at the outset, the main purpose of the present addendum is
to re-evaluate the scientific data base concerning health effects associated
with exposure to CO at ambient or near-ambient exposure levels. The re-
evaluation includes both (1) summarization of information contained in the
revised EPA Air Quality Criteria Document for CO (U.S. EPA, 1979) and (2) new
information and studies that have become available beyond what was reviewed in
the 1979 document. The most important points of information reviewed and key
conclusions derived from this evaluation of the CO health effects data base are
summarized below.
23
-------
DRAFT
8/22/83
Mechanisms of Action—The binding of CO to hemoglobin, producing COHb and de-
creasing the oxygen carrying capacity of blood, appears to be the main mechanism
of action underlying the induction of toxic effects of low-level CO exposures.
The precise mechanisms by which toxic effects are induced via COHb formation
are not yet fully understood, but likely include the induction of a hypoxic state
in many tissues of diverse organ systems. Alternative mechanisms of CO-induced
toxicity (besides COHb) have been hypothesized, but none have yet been demon-
strated to operate at relatively low (near-ambient) CO exposure levels. Blood
COHb levels, then, are currently accepted as representing a useful physiological
marker by which to estimate internal CO burdens due to the combined contribu-
tion of (1) endogenously derived CO and (2) exogenously derived CO resulting
from exposure to external sources of CO. COHb levels likely to result from
particular patterns (concentrations, durations, etc.) of external CO exposure
can be reasonably well estimated from equations developed by Coburn.
CO Exposure Levels—Evaluation of human CO exposure situations indicates that
occupational exposures in some workplace situations can regularly exceed 100
ppm CO, often leading to COHb levels of 10 percent or more. In contrast, such
high exposure levels are much less commonly encountered by the non-occupationally
exposed general public. More frequently, exposures to less than 25-50 ppm CO
for any extended period of time occur among the general population and, at the
low exercise levels usually engaged in under such circumstances, resulting COHb
levels most typically remain below 2-3 percent among non-smokers. Those levels
can be compared to the physiologic norm for non-smokers, which is estimated to
be in the range of 0.3 to 0.7 percent COHb. Baseline COHb concentrations in
smokers, however, often greatly exceed 1 percent, reflecting absorption of CO
from inhaled smoke.
Health Effects of Low Level CO Exposures—Four types of health effects reported
or hypothesized to be associated with CO exposures (especially those producing
COHb levels below 10 percent) were evaluated: (1) cardiovascular effects; (2)
neurobehavioral effects; (3) fibrinolysis effects; and (4) perinatal effects.
In regard to cardiovascular effects, decreased oxygen uptake capacity and
resultant decreased work capacity under maximal exercise conditions have been
clearly shown to occur in healthy young adults starting at 5.0 percent COHb;
24
-------
DRAFT
8/22/83
and one study observed small decreases in work capacity at COHb levels as low
as 3.3 to 4.3%. These cardiovascular effects may have health implications
for the general population in terms of potential curtailment of certain physi-
cally demanding occupational or recreational activities under circumstances of
sufficiently high CO exposure. However, of greater concern at more typical
ambient CO exposure levels are certain cardiovascular effects (i.e., aggravation
of angina symptoms during exercise) likely to occur in a smaller, but sizeable,
segment of the general population. This group, chronic angina patients, is pre-
sently viewed as the most sensitive risk group for CO exposure effects, based on
evidence for aggravation of angina occurring in patients at COHb levels of 2.9
to 4.5 percent COHb. Such aggravation of angina is thought to represent an
adverse health effect for several reasons articulated in the 1980 proposal pre-
amble (45 FR 55066), and the Clean Air Scientific Advisory Committee (CASAC)
concurred with EPA's judgment on this matter (see Appendix B). Dose-response
relationships for cardiovascular effects in cororary artery disease patients
remain to be more conclusively defined, and the possibility cannot be ruled out
at this time that such effects may occur at levels below 2.9 percent COHb (as
hinted at by the results of the now-questioned Aronow studies).
No reliable evidence demonstrating decrements in neurobehavioral function
in healthy young adults has been reported at COHb levels below 5%. Much of
the research at 5% COHb did not show any effect even when behaviors similar
to those affected in other studies were involved. However, if any CO effects
on neurobehavioral functions in fact occur below 5% COHb, then none of the sig-
nificant-effects studies would have found such decrements, because none of them
used COHb levels below 5%. Other workers who failed to find CO decrements at 5%
or higher COHb levels may have employed tests not sufficiently sensitive to
reliably detect small effects of CO. From the empirical evidence, then, it can
be said that the COHb levels in the 5% region do produce decrements in neuro-
behavioral function. However, it cannot be said confidently that COHb levels
lower than 5% would be without effect. One important point made in the 1979
document should be reiterated here. Only young, healthy adults have been stu-
died using demonstrably sensitive tests and COHb levels of 5% or greater. The
question of groups at special risk for CNS effects, therefore, has not been
explored. Of special note are those individuals who are'taking drugs which have
primary or secondary depressant effects which would be expected to exacerbate
CO-related neurobehavioral decrements. Other groups at possibly increased risk
25
-------
DRAFT
8/22/83
for CO-induced neurobehavioral effects are the aged and ill but these groups,
also, have not been evaluated for such risk.
In contrast to the data available which demonstrates associations between
cardiovascular and neurobehavioral effects and relatively low-level CO expo-
sures, much less clear evidence exists for other types of health effects being
associated with low-level CO exposures. For example, only relatively weak
evidence points towards possible CO effects on fibrinolytic activity and, then,
generally only at rather high CO exposure levels. Similarly, whereas certain
data are also suggestive of perinatal effects (e.g. reduced birth weight, slowed
postnatal development, Sudden Infant Death Syndrome) being associated with CO
exposure, insufficient evidence presently exists by which to either confirm
such associations qualitatively or to establish any pertinent exposure-effect
relationships.
Population Groups at Risk for Ambient CO Exposure Effects—Angina patients or
others with obstructed coronary arteries, but not yet manifesting overt symp-
tomatology of coronary artery disease, appear to be best established as a
sensitive group within the general population that is at increased risk for
experiencing health effects (i.e. exacerbation of cardiovascular symptoms) of
concern at ambient or near-ambient CO exposure levels. Several other probable
risk groups were identified, i.e.: (1) fetuses and young infants; (2) the
elderly (especially those with compromised cardiopulmonary functions); (3)
younger individuals with severe cardiac or acutely severe respiratory diseases;
(4) individuals with chronic bronchitis or emphysema; (5) individuals with
hematological diseases (e.g. anemia) that affect oxygen carrying capacity or
transport in the blood; (6) individuals with genetically unusual forms of hemo-
globin associated with reduced oxygen carrying capacity; and (7) individuals
using medicinal or recreational drugs having CNS depressant properties. However,
little emperical evidence currently is available by which to specify particular
COHb levels at which such individuals are likely to experience specific health
effects associated with ambient or near-ambient CO exposures. Nor does unambi-
guous evidence yet exist which clearly establishes that healthy non-sensitive
individuals or those in the above probable risk categories are affected at lower
CO exposure levels under high altitude conditions than CO exposure concentrations
effective at lower altitudes.
26
-------
DRAFT
8/22/83
REFERENCES
Agostoni, A., R. Stabilini, G. Viggiano, M. Luzzana, and M. Samaja.
Influence of capillary and tissue P02 on carbon monoxide binding to
myoglobin: a theoretical evaluation. Microvasc. Res. 20:81-7, 1980.
Alexieva, T., J. Boyadjieva, S. Ivanova, and E. Marinova. Some of the
early criteria in the diagnosis of chronic carbon monoxide poisoning.
Prob. Khygiennta !_: 73-78, 1975.
Anderson, E. W., R. J. Andelman, J. M. Strauch, N. J. Fortuin, and J. H.
Knelson. Effect of low-level carbon monoxide exposure on onset and duration
of angina pectoris: A study on 10 patients with ischemic heart disease. Ann.
Intern. Med. 79:46-50, 1973.
Aronow, W. S., D. N. Harris, M. W. Isbell, S. N. Rokaw, and B. Imparato.
Effect of freeway travel on angina pectoris. Ann. Intern. Med. 77:669-676,
1972.
Aronow, W. S., and M. W. Isbell. Carbon monoxide effect on exercise-induced
angina pectoris. Ann. Intern. Med. 79:392-395, 1973.
Aronow, W. S., E. A. Stemmer, and M. W. Isbell. Effect of carbon monoxide
exposure on intermittent claudication. Circulation 49:415-417, 1974.
Aronow, W. S., J. Cassidy, J. S. Vangrow, H. March, J. C. Kern, J. R. Goldsmith,
M. Khemka, J. Pagano, and M. Vawter. Effect of cigarette smoking and breathing
carbon monoxide on cardiovascular hemodynamics on anginal patients. Circulation
50:340-347, 1974.
Aronow, W. S., and J. Cassidy. Effect of carbon monoxide on maximal treadmill
exercise: A study in normal persons. Ann. Intern. Med. 83:496-499, 1975.
Aronow, W. S., J. Ferlinz, and F. Glauser. Effect of carbon monoxide on
exercise performance in chronic obstructive pulmonary disease. Am. J. Med.
63:904-908, 1977.
Aronow, W. S. Effect of passive smoking on angina pectoris. N. Eng. J. Med.
299:21-24, 1978.
Aronow, W. S. Aggravation of angina pectoris by two percent carboxyhemoglobin.
Am. Heart J. 101:154-157, 1981.
Ayres, S. M., H. S. Mueller, J. J. Gregory, S. Giannelli, Jr., and J. L. Penny.
Systemic and myocardial hemodynamic responses to relatively small concentrations
of carboxyhemoglobin (COHb). Arch. Environ. Health J8:699-709, 1969.
Ayres, S. M., S. Giannelli, Jr., and H. Mueller. Myocardial and systemic
responses to carboxyhemoglobin. In: Biological Effects of Carbon Monoxide,
Proceedings of a Conference, New York, Academy of Sciences, New York,
January 12-14, 1970. Ann. NY Acad. Sci. U4:268-293, 1970.
27
-------
DRAFT
8/22/83
Ayres, S. M., R. G. Evans, and M. E. Buehler. The prevalence of carboxyhemoglo-
binemia in New Yorkers and its effects on the coronary and systemic
circulation, Prev. Med 8:323-332, 1979.
Beard, R. R., and G. A. Wertheim. Behavioral impairment associated with
small doses of carbon monoxide. Am. J. Public Health 57:2012-2022, 1967.
Beard, R. R., and N. W. Grandstaff. Carbon monoxide and human functions.
In: Behavioral Toxicology. B. Weiss and V. G. Laties, eds., Plenum Press,
New York, 1975. pp. 1-26.
Becker, L. C., and E. D. Haak, Jr. Augmentation of myocardial ischemia
by low level carbon monoxide exposure in dogs. Arch. Environ. Health.
34:274-9, 1979.
Bender, W. W. Gothert, G. Malorny, and P. Sebbesse. Effects of low carbon
monoxide concentrations in man. Arch. Toxicol. 27:142-158, 1971.
Benignus, V. A., D. A. Otto, J. D. Prah, and G. Benignus. Lack of effects
of carbon monoxide on human vigilance. Percept. Mot. Skills 45:1007-1014,
1977.
Benignus, V. A., and K. E. Muller. Bad statistics are worse than none.
Neurotox. 3:153-154, 1982.
Benignus, V. A., K. E. Muller, C. N. Barton, J. D. Prah, and L. L. Benignus.
Neurobehavioral effects of CO exposure in humans. Protocol 1, Final report.
U.S. Army Medical Research and Development Command, Fort Detrick, Frederick,
MD. Project Order 1812 January, 1983.
Bergman, A. G., C. G. Ray, and M. A. Pomeroy. Studies of the sudden infant
death syndrome in King County Washington: III. Epidemiology. Pediatrics
49:860-70, 1972.
Bergman, A. B. and L. A. Wiesner. Relationship of passive cigarette
smoking to sudden infant death syndrome. Pediatrics 58:665-668, 1976.
Bonser, R. S. A., B. H. Knight, and R. R. West. Sudden infant death
syndrome in Cardiff, association with epidemic influenza and with tempera-
ture - 1955-1974. Int. J. Epidemial 7:335-40, 1978.
Brinkhouse, K. M. Effects of low level carbon monoxide exposure. Blood
lipids and coagulation parameters. U.S. Dept. Commerce, PB-269340, Spring-
field, Virginia, 1977.
Bunn, H. F., H. Forget, H. M. Ranney. Hemoglobinopathies (Saunders,
Philadelphia), p. 184-213, 1977.
Calverley, P. M., R. J. Leggett, and D. C. Flenley. Carbon monoxide and
exercise tolerance in chronic bronchitis and emphysema. Br. Med. J. [Clin
Res] 283:878-80, 1981.
Christensen, C. L., J. A. Gliner, S. M. Horvath, and J. A. Wagner. Effects
of three kinds of hypoxias on vigilance performance. Av. Sp. Env. Med.
48:491-496, 1977.
28
-------
DRAFT
8/22/83
Coburn, R. F., W. J. Williams, and R. E. Forster. Effect of erythrocyte
destruction on carbon monoxide production in man. J. Clin. Invest.
43:1098-1103, 1964.
Coburn, R. F., R. E. Forster, and P. B. Kane. Considerations of the physio-
logical variables that determine the blood carboxyhemoglobin concentration
in man. J. Clin. Invest. 44:1899-1910, 1965.
Coburn, R. F. Mechanisms of carbon monoxide toxicity. Prev. Med.
8:310-322, 1979.
Collier, C. R. and J. R. Goldsmith. Interactions of carbon monoxide and hemo-
globin at high altitude. Atmos. Environ. 17:723-728, 1983.
Davies, D. M. and D. J. Smith. Electrocardiographic changes in healthy
men during continuous low-level carbon monoxide exposure. Environ. Res.
21:9-15, 1980.
Davies, D. M., E. J. Jolly, R. J. Pethybridge, and W. P. Colquhoun. The
effects of continuous exposure to carbon monoxide on auditory vigilance in
man. Int. Arch. Occup. Environ. Health 48:25-34, 1981.
DeLucia, A. J., J. H. Whitaker, and L. R. Bryant. Effects of combined
exposure to ozone and carbon monoxide (CO) in humans. In: Advances in
Modern Environmental Toxicology, Vol. V, International Symposium on the
Biomedical Effects of Ozone and Related Photochemical Oxidants. Lee, S. D.,
M. G. Mustafa, M. A. Men!man (eds). Princeton Scientific Publishers:
Princeton, N. J., pp. 145-159, 1983.
Drinkwater, B. L., P. B. Raven, S. M. Horvath, J. A. Gliner, R. 0. Ruhling,
N. W. Bolduan, and S. Taguchi. Air pollution, exercise, and heat stress.
Arch. Environ. Health 28:177-181, 1974.
El-Attar, 0. A., and D. M. Sairo. Effect of carbon monoxide on the whole
fibrinolytic activity. Ind. Med. Surg. 37:774-777, 1968.
Ekblom, B., and R. Huot. Response to submaximal and maximal exercise at
different levels of carboxyhemoglobin. Acta Physiol. Scand. 86:474-482,
1972.
Fodor, G. G., and G. Winneke. Effect of low CO concentrations on resistance
to monotony and on psychomotor capacity. Staub Reinhalt. Luft 32:45-54, 1972.
Foster, J. R. Arrhythmogenic effects of carbon monoxide in experimental
acute myocardial ischemia: lack of slowed conduction and ventricular
tachycardia. Am. Heart J. 1_02:876-82, 1981.
Giacometti, G. M., M. Brunori, E. Antonini, E. E. Dilorio, and K. H.
Winterhalter. The reaction of hemoglobin Zurich with oxygen and carbon
monoxide. J. Biol. Chem. 255:6160-6165, 1980.
Goldsmith, J. R., and S. A. Landaw. Carbon monoxide and human health.
Science J62:1352-1359, 1968.
29
-------
DRAFT
8/22/83
Goldstein, I. Letter to the editor: Seasonal relationship of sudden
infant death syndrome and environmental pollutants. Am. J. Epid. 116,
189-191, 1982.
Groll-Knapp, E. , H. Wagner, H. Hauck, and M. Haider. Effects of low carbon
monoxide concentrations on vigilance and computer-analyzed brain potentials.
Staub Reinhalt. Luft 32:64-68, 1972.
Groll-Knapp, E., M. Haider, H. Jenkner, H. Liebich, M. Neurberger, and
M. Trimmel. Moderate carbon monoxide exposure during sleep: Neuro-and
psychophysical effects in young and elderly people. Neurobehav. Toxicol.
Teratol. 4:709-716, 1982.
Haft, J. I. Role of blood platelets in coronary artery disease. Am.
Cardiol. 43:1197-1206, 1979.
J.
Haider, M., E. Groll-Knapp, H. Hoeller, M. Neuberger, and H. Stidl. Effects
of moderate CO dose on the central nervous system—electrophysiological and
behaviour data and clinical relevance. In: Clinical Implications of Air
Pollution Research, Air Pollution Medical Research Conference, American
Medical Association, San Francisco, California, December 5-6, 1974. A. J.
Finkel and W. C. Duel, eds., Publishing Sciences Group, Inc., Acton, MA,
1976. pp. 217-232.
Hartiala, J., N. Simberg, and P. Uotila. Exposure to carbon monoxide or to
nicotine does not inhibit PGI2 formation by rat arterial rings incubated
with human platelet-rich plasma. Artery. JO:412-419, 1982.
Hoppenbrouwers, T., M. Calub, K. Arakawa, and J. E. Hodgman. Seasonal
relationship of sudden infant death syndrome and environmental pollutants.
Am. J. Epid. 113:623-635, 1981.
Horvath, S. M., T. E. Dahms, and J. F. O'Hanlon. Carbon monoxide and human
vigilance. A deleterious effect of present urban concentrations. Arch.
Environ. Health 23:343-347, 1971.
Horvath, S. M., P. B. Raven, T. E. Dahms, and D. J. Gray. Maximal aerobic
capacity at different levels of carboxyhemoglobin. J. Appl. Physio!. 38:300-303,
1975.
Horvath, S. M., S. M. Ayres, D. S. Sheps, and J. Ware. [Letter to Dr. Lester
D. Grant, including the peer-review committee report on Dr. Aronow's studies].
May 25, 1983. Available from: U.S. Environmental Protection Agency, Central
Docket Section, Washington, D. C.; docket no. OAQPS-79-7 IV. H.58, 1983.
Ingenito, A. J. and L. Durlacher. Effects of carbon monoxide on the B-wave
of the cat electroretinogram: Comparisons with nitrogen hypoxia, epinephrine
vasodilator drugs, and changes in respiratory tidal volume. J. Pharmacol.
Exp. Ther. 2Th 638-646, 1979.
Janzon, L., and I. M. Nilsson.
51:1120-1123, 1975.
Smoking and fibrinolysis. Circulation
30
-------
DRAFT
8/22/83
Kalmaz, E. V., L. W. Carter, C. A. Nau, and J. W. Hampton. Effect of
carbon monoxide exposure on blood. Fibrinolytic activity in rabbits.
Environ. Res. 14:194-202, 1977.
Kalmaz, E. V., L. W. Carter, R. L. Coleman, and J. W. Hampton. Enhance-
ment of fibrinolysis in rabbits exposed to low and moderate levels of carbon
monoxide inhibited by epsilon amino caproic acid. Thromb. Haemostasis 39:109-121,
1978.
Kalmaz, E. V., L. W. Canter, and J. W. Hamptom. Effect of long-term low and
moderate levels of carbon monoxide exposure on platelet counts of rabbits.
J. Environ. Pathol. Toxicol. 4:351-8, 1980.
Klein, J. P., H. V. Forster, R. D. Stewart, and A. Wu. Hemoglobin affinity
for oxygen during short-term exhaustive exercise. J. Appl. Physio!.
48:236-42, 1980.
Kurt, T. I., R. P. Mogielnicki, and J. E. Chandler. Association of the
frequency of acute cardiorespiratory complaints with ambient levels of carbon
monoxide. Chest 74:10-14, 1978.
Kurt, T. L., R. P.-Mogielnicki, J. E. Chandler, and K. Hirst. Ambient
carbon monoxide levels and acute cardiorespiratory complaints: an explora-
tory study. Am. J. Public Health 69:360-363, 1979.
Laties, V. G. and W. H. Merigan. Behavioral effects of carbon monoxide on
animals and man. Ann. Rev. Pharmacol. Toxicol. 19:357-392, 1979.
Lewak, N., B. J. Van den Berg, and J. B. Beckwith. Sudden infant death
syndrome risk factors. Clinical Pediatrics J_8:404-410, 1979.
Longo, L. D. The biological effects of carbon monoxide on the pregnant
woman, fetus, and newborn infant. Am. J. Obstet. Gynecol. 1_29:69-103, 1977.
Luria, S. M. and C. L. McKay. Effects of low levels of carbon monoxide on
visions of smokers and nonsmokers. Arch. Environ. Health 34:38-44, 1979.
Marcus, A. H. Mathematical models for carboxyhemoglobin. Atroos. Environ.
14:841-844, 1980.
Mansouri, A. and C. A. Perry. Alteration of platelet aggregation by
cigarette smoke and carbon monoxide. Thromb. Haemost. 48:286-8, 1982.
McFarland, R. A., F. J. W. Roughton, M. H. Halperin, and J. I. Niven.
The effects of carbon monoxide and altitude on visual thresholds. J. Aviat.
Med. l_5:381-394, 1944.
McFarland, R. A. Low level exposure to carbon monoxide and driving per-
formance. Arch. Environ. Health 27:355-359, 1973.
Mihevic, P. M., J. A. Gliner, and S. M. Horvath. Carbon monoxide exposure and
information processing during perceptual motor performance. Int. Arch.
Occup. Environ. Health 51:355-363, 1983.
31
-------
DRAFT
8/22/83
O'Donnell, R. D., P. Mikulka, P. Heinig, and J. Theodore. Low level carbon
monoxide exposure and human psychomotor performance. Toxicol. Appl. Pharma-
col. lj*: 593-602, 1971.
Ott, W. and D. Mage. Further discussions: Evaluation of CO air quality
criteria using a COHb model. Atmos. Environ. 1_4:979-980, 1980.
Panchenko, E. N., I. F. Nalcha, N. I. Dzyuba, and N. P. Luk1 Yanova. The
effect of industrial factors in the development of cerebral atherosclerosis
in workers in a methanol plant. Vrach. Delo. 4:109-113, 1977.
Peterson, D. R. Sudden, unexpected death in infants. An epidemiologic
study. Am. J. Epid. 84:478-482, 1966.
Peterson, D. The sudden infant death syndrome - reassessment of growth
retardation in relation to maternal smoking and the hypoxia hypothesis.
Am. J. Epid. VI3:583-589, 1981.
Peterson, J. E., and R. D. Stewart. Predicting the carboxyhemoglobin levels
resulting from carbon monoxide exposures. J. Appl. Physiol. 39:633-638, 1975.
Pirnay, F., J. Dujardin, R. Deroanne, and J. M. Petit. Muscular exercise during
intoxication by carbon monoxide J. Appl. Physiol. 31:573-575, 1971.
Putz, V. R., B. L. Johnson, and J. V. Setzer. Effects of CO on Vigilance
Performance. Effects of Low Level Carbon Monoxide on Divided Attention,
Pitch Discrimination, and the Auditory Evoked Potential. DHEW (NIOSH)
Publication No. 77-124, U.S. Department of Health, Education, and Welfare,
National Institute of Occupational Safety and Health, Cincinnati, OH,
November, 1976.
Putz, V. R. The effects of carbon monoxide on dual-task performance. Human
Factors 21:13-24, 1979.
Putz, V. R., B. L. Johnson, and J. V. Setzer. A comparative study of the
effects of carbon monoxide and methylene chloride on human performance.
In: Proc. 1st. Ann. NIOSH Sci. Symp. Chicago: Pathotox Publishing Co., 1979.
Raven, P. B., B. L. Drinkwater, S. M. Horvath, R. 0. Ruhling, J. A. Gliner,
J. C. Sutton, and N. W. Bolduan. Age, smoking habits, heat stress, and their
interactive effects with carbon monoxide and peroxyacetylnitrate on man's
aerobic power. Int. J. Biometeorol. J8:222-232, 1974.
Raven, P. B., B. L. Drinkwater, R. 0. Ruhling, N. W. Bolduan, S. Taguchi,
J. A. Gliner, and S. M. Horvath. Effect of carbon monoxide and peroxyacetyl
nitrate on man's maximal aerobic capacity. J. Appl. Physiol. 36:288-293, 1974.
Roche, S., S. Horvath, J. Gliner, J. Wagner, and J. Borgia. Sustained
visual attention and carbon monoxide: elimination of adaptation effects.
Human Factors 23:175-184, 1981.
Rockwell, T. J., and F. W. Weir. The Interactive Effects of Carbon Monoxide
and Alcohol on Driving Skills. Ohio State University Research Foundation,
Columbus, Oh, January 1975.
32
-------
DRAFT
8/22/83
Roughton, F. J. W. The equilibrium of carbon monoxide with human hemo-
globin in whole blood. Ann. N. Y. Acad. Sci. J74:177-188, 1970.
Rummo, N. , and K. Sarlanis. The effect of carbon monoxide on several measures
of vigilance in a simulated driving task. J. Saf. Res. 6:126-130, 1974.
Salvatore, S. Performance decrement caused by mild carbon monoxide levels
on two visual functions. J. Saf. Res. 6:131-134, 1974.
Schulte, J. H. Effects of mild carbon monoxide intoxication. Arch.
Environ. Health 7:524-530, 1973.
Stewart, R. D., P. E. Newton, J. Kaufman, H. V. Forster, and J. P. Klein.
Effect of a rapid 4 percent carboxyhemoglobin saturation increase on maximal
treadmill exercise. CRC-APRAC-CAPM-22-75, 1978.
U. S. Environmental Protection Agency. Air Quality Criteria for Carbon
Monoxide, U. S. EPA Publication No. 600/8-79-022. October, 1979.
Venkatram, A. and R. Louch. Evaluation of CO air quality criteria using a
COHb model. Atmos. Environ. 1_3:869~872, 1979.
Vogel, J. A. and M. A. Gleser. Effects of carbon monoxide on oxygen transport
during exercise. J. Appl..Physio!. 32:234-239, 1972.
Weiser, P. C., C. G. Merrill, D. W. Dickey, T.L. Kurt, and G.J.A. Cropp. Effects
of low-level carbon monoxide exposure on the adaptation of healthy young
men to aerobic work at an altitude of 1,610 meters. In: Environmental
Stress. Individual Human Adaptations. L. J. Folinsbee, J. A. Wagner
J. F. Borgia, B. L. Drinkwater, J. A. Gliner, and J. F. Bedi, eds.,
Academic Press, New York, 1978. pp. 101-110.
Winneke, G. Behavioral effects of methylene chloride and carbon monoxide as
assessed by sensory and psychomotor performance. In: Behavioral Toxicology.
Early Detection of Occupational Hazards, Proceedings of a Workshop, National
Institute for Occupational Safety and Health and University of Cincinnati,
Cincinnati, Ohio, June 24-29, 1973. C. Xintaras, B. L. Johnson, and I. de Groot,
eds., HEW Publication No. (NIOSH) 74-126, U. S. Department of Health, Edu-
cation and Welfare.
Winneke, G., G. Fodor, and H. Schlipkoter. Carbon monoxide, trichloroethylene,
and alcohol: reliability and validity of neurobehavioral effects. In:
Multidisciplinary Perspectives in Event-Related Brain Potential Research,
Proceedings of the Fourth International Congress on Event-Related Slow
Potentials of the Brain (EPIC IV), Univ of NC and U.S. EPA, Hendersonville,
NC, April 4-10, 1976. D. A. Otto, ed., EPA-600/9-77-043, U.S. EPA, RTP,
NC, December 1978. pp. 461-469.
Wright, G., P. Randell, and R. J. Shephard. Carbon monoxide and driving
skills. Arch. Environ. Health 27:349-354, 1973.
Zinkham, W. H., R. A. Houtchens, and W. S. Caughey. Carboxyhemoglobin
levels in an unstable hemoglobin disorder (Hb Zurich): effect on phenotypic
expression. Science 209:406-408, 1980.
33
-------
DRAFT
8/22/83
APPENDIX A
EPA HEALTH EFFECTS RESEARCH LABORATORY
REEVALUATION OF ANDERSON ET AL. (1973) STUDY
A-l
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
HEALTH EFFECTS RESEARCH LABORATORY
' «o RESEARCH TRIANGLE PARK
NORTH CAROLINA 27711
DATE: August 19, 1983
SUBJECT: Evaluation of Anderson et al. (1973) Study of Carbon Monoxide (CO)
Effects on Angina Patients.
FROM: John J. O'Neil, Ph.D., Chief
Clinical Research Branch, ITD/HERL
TO: Lester D. Grant, Ph.D.
Director, ECAO (MD-52)
THRU: F. Gordon Hueter, Ph.D., Director
Health Effects Research Laboratory, RTP (MD-58)
Thank you for the opportunity to comment on the paper by Einar Anderson
and his co-workers which is entitled "Effect of Low-Level Carbon Monoxide
Exposure on Onset and Duration of Angina Pectoris." I have re-read it,
have discussed it with several colleagues, especially Dr. Vernon Benignus,
and would offer the following observations by way of review.
This is a good paper. The work appears to be carefully done, the
data seem reasonable, and the paper is well written. Critical review
reveals some flaws in the design and conduct of the research, but this is
typical of most scientific work and, therefore, significance of these
comments may be interpreted differently by different readers.
DESIGN
1) ST segment depression reported in this paper is not useful
and does not constitute any physiological support to the observations
regarding the onset and duration of angina. The authors recognize this
shortcoming. The important data in the paper are the measurement of time
to onset of pain and the measurement of the duration of pain.
2) The paper would have been considerably strengthened by more
careful selection of the subject population. For example, 5 subjects
were smokers and 5 were non-smokers and one subject was taking digitalis.
In my opinion, the study would have been strengthened considerably had
the subject population been more homogeneous.
3) It is unclear to me if the study is really double blind. That
is, did the investigator who administered the gases also interact with
the subjects by, for example, placing the mask on the subject. Such
interactions would make the study not a true double blind design.
4) The number of subjects (n) is very small. Only ten subjects
were studied and of these there are three for whom there are missing data
points. I have been told anecdotally that this study was intended as
a pilot study. It is unfortunate that it was not possible to design
and complete the proper follow-up study. Given the situation, this is
valuable data to have published and should be used to develop our under-
standing of the response of angina patients to CO.
A-2
-------
METHODS AND RESULTS
1) It is disconcerting that the measured COHb levels appear to be
lower than we would predict on the basis of our knowledge of CO-COHb
kinetics. Why is this measured value lower than we would expect?
a) Measurement of COHb. The measurements were done in
triplicate using the Buchwald analysis. As best as I can determine,
these estimates of COHb are accurate and there is no reason to doubt
their validity on the basis of technique or equipment.
b) It is also possible that the exposure mask used in this study
was loose fitting and leaked. This would reduce the exposure level
and the final COHb levels.
c) Though not reported in the paper, the subjects were apparently
allowed to rest ad libitum during the study. This amounted to approxi-
mately 10-15 minutes out of the hour. This "rest" period would have
reduced the exposure time an
-------
Table 2. Time to Onset of Pain
Mean (Range)
Mon (Air) 294 (215-480)
Air (Control) 325 (220-435)
50 ppm 264.5 (65-390)
100 ppm 263.5 (85-425)
Friday (Air) 300 (185-485)
The exposures on Monday and Friday were included to demonstrate the com-
parison of two air exposures separated by the study itself. The air
control and the exposures to 50 or 100 ppm CO were randomized on Tuesday
Wednesday and Thursday. The data for Monday and Friday are both considerably
closer to the data for 50 and 100 ppm CO exposure than to that for the air
control. This is, I believe, associated with the small number of subjects
used in the study. Dr. Benignus and his colleagues analyzed the data for
the different air days and did not show any significant differences
between these days.
STATISTICAL ANALYSIS
Dr. Benignus has developed an analysis of the paper by Anderson et
al. and I have excerpted part of his critique here.
Critique of the Study by Anderson et a!U (1973)
Vernon A. Benignus
The statistical tests used in the study were anticonserv-
ative (Benignus and Muller, 1982) and thus would have tended
toward showing a significant effect even if none were present
in the population.
The data of Anderson et al. were reanalyzed using one
multivariate analyses of variance for each of the two measures
(time to onset of angina and duration of angina). Multivariate
tests were used because the data for 0, 50 and 100 ppm exposure
were collected from the same subjects. Since two separate
significance tests were to be run (one for time to onset and
one for duration) Bonferroni corrections were used to keep
experimentwo.se a = .05. Each test would then be evaluated
at a/2 = .025.
A-4
-------
Table A. Overall Test Results
Variable
Time to Onset
Duration
F
7.72
3.10
df
2,8
2,7
P<
.014
.11
Table A shows the results of the overall tests. This table
agrees with Anderson et al. that time to onset of angina
would have been significantly affected by CO exposure but
does not agree with Anderson e_t al. that duration of Angina
is affected by CO exposure. Stepdown tests of the time to
onset data revealed that the 50 ppm exposure showed a CO
effect, p<.017 and the 100 ppm exposure showed a CO
effect p<.002.
Table B. Means of Time to Onset Data
CO Level Mean time to onset of Angina
0 ppm 310
50 ppm 265
100 ppm 264
Table B shows the mean times to onset for the 0, 50, and
100 ppm exposure days. While both the 50 and 100 ppm
exposures produced shorter times to onset, they did not
differ among themselves. (The fact that the p value for
the 100 ppm day was lower than for the 50 ppm day was due
to the fact that the data on the 100 ppm day were more
closely correlated with the 0 ppm values. Thus Table B
shows a puzzling non-dose-related finding.)
Inspection of Table 1 in Anderson e_t al_. (1973),
reveals that all COHb values were related to the exposure
level so that the non-dose-related findings in time to
onset remain unexplained. To be sure, the variability;
was high on all days and with the small number of subjects,
such non dose related findings can occur even if the effects
A-5
-------
are dose related in the population. This is especially
true since the COHb values were both close to the lower
limits for CO effects and the general dose-effects curve
can be plausibly argued to have a rather low slope.
The tests of signficance employed in this critique are
quite conservative. Thus, the fact that significant results
would have been reported even if appropriate and conservative
tests had been done, lends credibility to the study. On the
other hand, the non dose related findings are mildly disturbing.
Considering (1) the small number of subjects, (2) the high
variability of the data, (3) the fact that both doses were
near the no-effects level and (4) the low p values on the
significance tests, the data are strongly suggestive of an
effect but are sorely in need of replication and extension.
More subjects should be run and a wider dose range
should be studied to be able adequately to quantify a
dose-effects curve.
Conclusions
Several aspects of the study were of above average quality. The study
was conducted in a double-blind fashion; very commendable, especially by
comparison to the extant literature. Only a small number of variables
were studied, thus minimizing the chances of finding some spurious variable
which accidentally covaried with CO exposure; another positive quality.
When compared to an ideal study the study by Anderson et al. has
several flaws in its design and execution and the results have inconsistencies.
However, when compared to the extant literature, the design and execution
of this study is commendable. None of the inconsistencies are of a major
nature and several plausible explanations exist for them. The results of
this study suggest that angina is exacerbated by small increases in COHb.
This study is sorely in need of replication and extension. More subjects
should be run and a wider dose range should be studied to be able to adequately
quantify the dose response relationships. Even if no inconsistencies were
present in this study, it would be rash to rely entirely upon one study
with 10 subjects. It would be equally irresponsible to disregard these
findings. The greatest imprudence would be to fail to do the studies
sugested by these results.
A-6
-------
DRAFT
8/22/83
APPENDIX B
CASAC LETTER (AUGUST 31, 1982) TO
EPA ADMINISTRATOR CONCERNING ISSUES
INVOLVED IN SETTING OF NAAQS FOR CO
B-l
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON. D.C. 20A60
August 31, 1982
OFFICE OF
THE ADMINISTRATOR
Mrs. Anne M. Gorsuch
Admini strator
U.S. Environmental Protection Agency
Washington, D.C. 20460
Dear Mrs. Gorsuch:
The Clean Air Scientific Advisory Committee (CASAC) met
on July 6 to provide its advice on several issues related to
the ambient air quality standard for carbon monoxide. The
Committee had previously advised the Administrator of the
scientific adequacy of the criteria document and staff paper
in a closure memorandum dated October 9, 1979.
At its most recent meeting the Committee provided advice
to the Agency on four issues. These included: 1) setting a
revised eight-hour carbon monoxide standard that includes five
allowable exceedances; 2) the role and significance of the 1981
study published by Dr. Wilbert Aronow; 3) sensitivity analysis
and exposure analysis predictions of carboxyhemoglobin (COHb)
levels and ambient CO concentrations under alternative air
quality standards;. 4) range of scientifically acceptable
alternative standards for CO.
I would like to briefly summarize for you the Committee's
views on each of these issues.
1. Development of a Multiple Exceedance 8-Hour Standard.
The CASAC reached a consensus that a multiple exceedance
standard has both scientific as well as administrative merit.
From a scientific point of view this approach recognizes the
stochastic or random-like character of meteorological events;
administratively, it reduces the exement of chance in determining
compliance with the standard. In recommending that you adopt
a multiple exceedance standard, the Committee notes that an
increase in the number of allowable exceedances will, in effect
relax the existing standard if the standard level 'remains
unchanged. In order to provide protection to the public health
with an adequate margin of safety you should consider the impact
of a multiple exceedance standard upon ambient CO concentrations
and levels of blood COHb.
B-2
-------
2. Role of the 1981 Aronow Study.
CASAC reached no overall consensus on the significance
which the Agency ought to attribute to the Aronow study. The
study reported a 10 percent reduction in the time to onset of
angina during treadmill exercise at blood carboxyhemoglobin
levels of 2 percent. CASAC discussed the fact that the response
observed at 2.0% COHb was more subtle than that observed at
higher levels (2.7 - 2.9% COHb) and speculated that even more
subtle responses might be found at COHb levels below 2.0%. - The
Committee concluded that there may be no physiological response
threshold for carbon monoxide. One CASAC consultant, while
noting that the study data are solid and irrefutable, concluded
that activity and exposure patterns of angina patients are far
different from the general population. He also observed that
there is no reason to believe that changes in the time of
onset of angina during treadmill exercise are a valid biologic
endpoint for the determination of an adverse health effect.
Another Committee consultant, however, concluded that shortening
of exercise time prior to the onset of an angina attack clearly
is an adverse health effect.
While reaching no consensus on the role of this study, the
Committee's earlier position as stated in the October 9, 1979
closure memorandum — that.the critical effects level for COHb
occurs between 2.7% - 3.0% and that the onset of angina represents
an -adverse effect — remains as the CASAC consensus on this
issue.
3. Scientific and Technical Adequacy of Sensitivity and
Exposure Analyses.
The sensitivity and exposure analyses were prepared by the
Agency to compare the relationship between ambient CO concentrations
and various levels of blood COHb. In addition, the analyses
estimated the number and distribution of individuals who were
projected to experience various COHb levels under alternative
CO standards.
CASAC has reviewed the exposure and sensitivity analyses
and has concluded that both are scientifically acceptable given
the current etate-of-the-art of the scientific community's
ability to model physiological and other parameters related to
this pollutant. Specifically, the Committee would draw to your
attention two of its conclusions on these analyses: 1) the
Agency's use of the Haldane constant with a value set at 218 is
a reasonable selection among a variety of physiological parameters
B-3
-------
discussed in the sensitivity analysis? and 2) the draft preamble
states that an Agency objective is to keep 99% of the population
below a COHb level of 2.5%. Since there may be no threshold
concentration level for carbon monoxide below which no adverse
effects will be experienced by anyone, and since one hundred
percent protection is not feasible, a social policy choice must
be made to limit societal risk from this pollutant. From a
scientific standpoint the 99% objective is within the realm of
reason, but there may be other than scientific factors you wish
to consider in reaching a decision on this particular issue.
4. Scientifically Acceptable Range for the 8-Hour CO
Standard.
In commenting upon the staff's proposals for a revised
8-hour CO standard set at 9 parts per million (p. p.m.) with
five exceedances, or 12 p. p.m. with one exceedance, the Committee
made the following consensus observations:
o a standard set at 12 p. p.m. with 5 exceedances
is not scientifically acceptable
o a standard established at 12 p. p.m. with 1 exceedance
would provide a very small margin of safety
o the scientific evidence alone cannot identify an
exact level at which to set a standard for carbon monoxide.
Given the need to protect sensitive members of the population
from this pollutant, the Committee advises you to choose a
standard level and a corresponding number of exceedances that
will limit COHb below the critical effects level of 2.7 - 3.0%,
with an adequate margin of safety.
. The Committee appreciates the opportunity to advise you on
the carbon monoxide standard and hopes that its comments will
be useful as you finalize the standard. We urge you to proceed
expeditiously in this matter because the criteria document and
staff paper, reviewed by CASAC more than three years ago, will
be increasingly subject to challenge because of any newly published
literature on this pollutant. In addition, both the private
sector and individual citizens need to know the standard level
for the next five years for planning purposes and for reassurance
that public health is being adequately protected.
Sincerely yours,
Sheldon K. Friedlander
Chairman, Clean Air Scientific
Advisory Committee *
cc: Dr. John W. Hernandez
Kathleen Bennett
Dr. Terry F. Yosie
B-4
------- |