728 x 90
728 x 90
728 x 90
728 x 90
728 x 90

Clearing the Air on Carbon Monoxide: Fatal Scientific Flaws in the EPA Crackdown on Denver and Other

May 1, 1988

Issue Paper


By C. Edwin Baker, Gordon Swanby

Executive Summary:

  • Environmental analysts Baker and Swanby contend that the EPA is strong-arming Denver and scores of other cities with unduly strict control measures against carbon monoxide levels that are really a nonproblem in health terms.
  • Maximum allowable CO levels are set far lower than technical studies can justify, providing an excessive and burdensome margin of safety. Colorado uses a CO monitoring system that exaggerates actual hazards and violates EPA guidelines.
  • The authors find no scientific warrant for frightening the public into expensive and disruptive lifestyle changes. The regulations seem designed to satisfy abstractions and centralize power, not to protect real people from real risks.
  • EPA should raise the violation standard to a scientifically valid level and trigger it after the fifth bad day rather than the first. Also, since cleaner cars seem to have accounted for most of the recent CO improvement, other control measures should be suspended for a year to confirn whether this is so.
  • Rebuttals by the EPA Regional Administrator and the Metro Air Quality Council are in the Appendices I and II.

Introduction: Myth and Reality

Denver is one of 59 U.S. cities that are under scrutiny by the U.S. Environmental Protection Agency, facing economic sanctions in the form of withheld federal highway funds and pollution control grants. This action is being contemplated by the government because these cities have not achieved compliance with EPAs ambient air quality standard for carbon monoxide (CO), as mandated by the Clean Air Act.

This standard was to have been met by the end of 1987, but Congress decided to allow an additional eight months so that appropriate priate measures for dealing with non compliant cities could be strategized. In the interim, EPA has attempted to bring a measure of sanity to the matter by suggesting only those cities that have not made a sincere effort to comply with the standards should be penalized.

Further, they recommended that if a non-compliant city shows minimal, but steady, improvement in air quality, (e.g., three percent per year), then that city should be deemed to be making satisfactory progress toward complying with the standard.

But on the other hand, it is the very stringent EPA CO standard that has created the problem in the first place. Misleadingly described as health-based, the standard turns out to have been set by such purist criteria, with associated non-representative monitoring requirements, that it will be wrenchingly difficult, if not impossible for most cities ever to come into compliance.

The threat of economic sanctions by the EPA has brought about a spate of CO control strategies in Denver during the past few years. The associated uproar has created the erroneous notion with the public that our air quality is worse now than ever. The truth is that Denvers CO levels have been declining steadily over the past several years, and are continuing to improve. In that sense, Denvers air is cleaner and safer now than at any time since the agencies began keeping air quality records.

The reason Denver is said to have a CO problem is that the overly stringent ambient standard has not been met, and projections are grim for meeting the standard any time soon. In analyzing this problem, the important questions are (1) what is the standard, (2) what is the basis for the standard, (3) what are the health consequences of exceeding the standard, (4) how is ambient CO measured, and (5) how is compliance determined?

This issue paper will examine each question in turn, demonstrating why the authors believe the standard has a faulty scientific basis, the monitoring programs are flawed, and the compliance policy is unjustified.

Carbon Monoxide: Definite Dangers. Debatable Response

Carbon monoxide is an odorless, colorless gas present to some degree in the atmosphere (usually less than 1 ppm) and in buildings at all times. Most CO is the result of incomplete combustion of organic (carbon-containing) materials, which if completely oxidized would yield carbon dioxide and water.

Carbon monoxide is produced when combustion processes are not highly efficient or when there is not enough oxygen present to combine with the available carbon atoms. Carbon monoxide in the atmosphere cannot be seen, and therefore does not contribute to visible pollution.

Inhalation of CO is considered a health issue because it competes with oxygen in the blood for attachment to the hemoglobin molecule. Obviously, if the inhaled concentration of CO is extremely high, considering CO has an affinity for hemoglobin at least 200 times that of oxygen, death can result in a very short time. However, concentrations of around 4000 ppm are required for the manifestation of lethal effects.

It is this characteristic of CO that causes the gas to be referred to by the media as “poisonous” or “deadly” when they report on pollution episodes or CO control measures. However, it must be emphasized that atmospheric concentrations of CO experienced today in Denver, or in any other U. S. city (less than 20 ppm), in no way approach deadly or poisonous proportions.

Rather, the public and political concern about CO pollution results from heavy media coverage of the failure of numerous cities to meet a very stringent air quality standard.

The EPA ambient standards for carbon monoxide are 35 ppm one-hour average and 9 ppm eight-hour average; the eight-hour standard is computed on a “moving average” basis, and is not to be exceeded more than once per year. Denver no longer has a problem meeting the onehour standard, and so this paper will address only the eight-hour standard.

Carbon monoxide is monitored at several locations in a metropolitan area, representing urban, middle-scale and neighborhood areas (Denvers monitoring scheme is presented in Figure 1). The monitors are located so as to measure worstcase exposures because EPA intends for the maximally exposed and most sensitive individuals to be protected by the ambient standards. The EPA does not allow averaging of measured values from the several monitors in an area, nor do they allow averaging of data from any one monitor beyond the eight-hour limit.

On an annual or seasonal basis, a city may be below the standard, and theoretically in compliance, but wintertime temperature inversions and the proximity of passing automobiles to CO sensors sometimes make it impossible for every monitor to remain below the standard. However, it is interesting to note, particularly in the case of Denver, that during even the worst temperature inversions, some of the area monitors do remain below the standard. Tn fact, some monitors never register violations.

It only requires one monitor registering an exceedance of the standard, and then that same or a different monitor registering another exceedance on a different day during the same calendar year, to result in a state of noncompliance with the standard.

The amount by which the standard is exceeded is of no consequence to the EPA for compliance purposes. A city is in as much trouble with the EPA for 10 ppm as it is for 20 ppm. Its pass or fail–no middle ground–even though most cities today are not exceeding the standard very often nor by large increments.

From a health standpoint, it is not the concentration of CO in the air that is important, but rather how much CO is inhaled and then absorbed by the blood stream. When CO combines with hemogloblin molecules in the blood, the resultant chemical substance is called carboxyhemoglobin (COHb). The risk to health generally is expressed as the percentage of COHb in the blood.

There is a lack of general agreement among the scientific community regarding the relationship between concentration of CO in ambient air and the resulting percentage of COHb in the blood. Neither is there consensus regarding the lowest percentage of COHb that represents a threat to human health. However, there is agreement that healthy persons can tolerate higher percentages of COHb in the bloodstream than weaker individuals (e.g., those with cardiac or pulmonary diseases).


Air Pollution Monitoring Stations in the Denver, Colorado Metropolitan Area

1. WELBY 78th & Steele
2. HIGHLANDS S. University & County Line Road
3. ARVADA 57th & Garrison
4. CARRIAGE 23rd & Julian
5. NATIONAL JEWISH HOSPITAL Colfax & Colorado Blvd.
6. CAMP 21st & Broadway
7. AURORA 50 S. Peoria
8. BOULDER 2320 Marine Street

NOTE: Since this map was prepared, the Aurora station has been relocated to Englewood, near the intersection of S. Huron Street and Eastman Avenue, north of Cinderella City Mall.

SOURCE: “PSI”, a pamphlet published by the Colorado Department of Health, Air Pollution Control Division, January 1987.

Since the EPA is charged under the Clean Air Act to protect all individuals in society, air quality standards are set conservatively to assure a corresponding level of protection. While this goal is a noble one, it will become apparent later in this paper that ambient levels of carbon monoxide found in the air of Denver and other U. S. cities today, having improved drastically over the past decade, do not represent a threat to the health of normal individuals, and only theoretically to the health of weaker persons.

Why the Standard is Scientifically Suspect

The Occupational Safety and Health Administration has set a workplace CO standard of 50 ppm ,(l) averaged over eight hours; this standard was set with a margin of safety. The American Conference of Governmental Industrial hygienists recommends a short-term exposure limit (SThL) for CO of 400 ppm(2) Table 1 summarizes the gaping disparities between the CO standards of other entities and EPAs 9 ppm standard.

The EPA 9 ppm standard dates back to the early 1970s, prior to the congressional mandate to establish health-based standards. Nevertheless, at that time, EPA took a health-based approach to defining and setting the standard. As they are still practicing today, the agency collected and reviewed all of the scientific literature on CO effects, and then selected the data base that reflected the most cautious view (which incidentally cannot be corroborated by the vast majority of published research on the subject).

Interestingly, however, the studies forming the basis for the original standard-setting subsequently were discredited, (3) and the EPA was forced to search for further justification, or else risk losing the standard to a higher value.

The EPA did, in fact, come up with another group of studies that helped them “back into” the extant standard. These studies were performed and published by Dr. Wilbert Aronow, who coincidentally also provided drug testing information used by the Food and Drug Administration (FDA).

In 1983, the FDA found serious technical difficulties with Dr. Aronows work, leading to the rejection of his drug study data. Upon further investigation, the EPA also found problems with Aronows CO studies, and once again found themselves in the position of having to discard the technical basis for the 9 ppm.

But not to worry–enter the Anderson study. E. W. Anderson et al. published a paper in 1973 that demonstrated a CO-related “health effect” experienced by persons suffering from angina pectoris.(4) By studying the length of time to onset of chest pain in these patients, Anderson concluded that inhalation of CO for four hours prior to exercise on a treadmill decreased the time to onset of chest pain by approximately 20%.


Comparison of EPA Carbon Monoxide Standards with Standards and Criteria Published by Other Authorities

(all values in ppm)









OSHA(c )








DOP (g)







(a) Short-term exposure limit (usually 15 minutes)

(b) Since the EPA-determined threshold of health effects is 18 ppm, the 9 ppm standard represents a 100% margin of safety.

(c) Occupational Safety and Health Administration

(d) National Institute of Occupational Safety and Health

(e) American Conference of Governmental Industrial Hygienists

(f) 25 ppm recommended for persons with cardiac or pulmonary disease

(g) Department of Transportation

(h) N. Irving Sax: “Dangerous Properties of Industrial Materials”, sixth edition

Notwithstanding it is stretching definitions to classify premature onset of chest pain during exercise as a “health effect”, several serious problems have surfaced regarding the Anderson study, which now provides the sole technical basis for the 9 ppm standard.

Anderson used only 10 subjects in his study, and these persons did not represent a homogeneous study group in any sense of the word. Five were smokers and one subject was taking digitalis. For his conclusions and application of statistics, Anderson used average values, even though the ranges of results among the test groups overlapped in some cases.

Additionally, in order to produce the observed effects, patients were administered CO concentrations of 50 and 100 ppm, which would be expected to result in COHb levels of about 8% and 15%, respectively. However, the COHb levels he measured were in the range of only 3% and 4.5%.

It generally is conceded that carefully controlled inhalation experiments of 50 and 100 ppm CO would produce higher COHb levels in the blood. Therefore, the EPA concluded either the COHb levels in reality were higher, reflecting an error in blood chemistry analysis, or the patients did not actually breathe the stated concentrations of CO. Because the latter would justify the standard for a lower ambient concentration of CO, that explanation was accepted.

Then, by back-calculating from the most conservative model available (the Coburn model), the EPA surmised the ambient level of CO that imparts 3% COHb to the blood, under moderate exercise, was approximately 18 ppm. By applying a margin of safety to this concentration, the 9 ppm standard once again was “justified”. The Clean Air Act stipulates the margin of safety is to be “adequate”, and there certainly is no question the difference between 9 ppm and 18 ppm qualifies as adequate!

The problems surrounding the Anderson study led to two important statements by reviewers. The first, made by Dr. Vernon Benignus, is very revealing: “The statistical tests used in the study… .would have tended toward showing a significant effect even if none were present…” (3)

The second is equally revealing, albeit more provocative, and is taken from a 1982 letter to EPA administrator Ann Gorsuch from the chairman of EPA s hand-picked Clean Air Scientific Advisory Committee