Tobacco Documents Online

COMMENTS ON THE RISK ASSESSMENT PORTION OF THE MAY 1990 EPA DRAFT REPORT ENTITLED "HEALTH EFFECTS OF PASSIVE SMOKING: ASSESSMENT OF LUNG CANCER IN ADULTS AND RESPIRATORY DISORDERS IN CHILDREN" PREPARED BY ANTHONY V. COLUCCI, Sc.D. COLUCCI AND ASSOCIATES, INC. 823 W. FIFTH STREET WINSTON-SALEM, NC 27101 SEPTEMBER 29, 1990

I. Introduction Over the last half decade, considerable attention has been paid to studies of environmental tobacco smoke (ETS) and a variety of diseases, including lung cancer. Over that time period, several attempts have been made to generate a risk function for ETS that would allow assessment of the number of potential lung cancer deaths which could be estimated in the population of non-smokers exposed to ETS. The first of the risk assessment attempts was performed by Repace and Lowrey in 1985 utilizing both the epidemiologic and cigarette-equivalents approach. At that time, the number of deaths they projected in the population ranged from 500 to 5,000 per year. Their estimates created a great deal of controversy. The controversy continues today. The Environmental Protection Agency (EPA), drawing upon work performed by the U.S. Surgeon General and the National Academy of Sciences (NAS), has undertaken its own risk assessment and has derived, using similar methodology, an estimate of the number of lung cancer deaths expected each year among non- smokers exposed to ETS. (EPA Draft Report) It is my belief that all of these risk estimates suffer from severe methodologic problems. This is true whether they are based on the current epidemiology studies or extrapolation of mainstream smoke to ETS using the daily cigarette equivalents methodology. In this latter regard, I conclude that the key to the adequacy and robustness of ny estimate of risk associated with any environmental pollutant is the ability to measure this environmental pollutant in a sufficiently accurate and precise way to be able to predict a level of exposure and, thus, the dose delivered to the target organ in the population at risk.

Within the context of ambient air quality standards, the EPA uses specific measures of a variety of ambient air pollutants in contact with the free-living population. All ambient air quality standards and, thus, all risk assessments conducted for contaminants of this type, are based upon certified measurement methodology sanctioned by the EPA. Virtually every substance which has been classified as a national environmental hazardous air pollutant has been so classified based upon adequate, extensive, and highly controlled measurements subjected to intense quality control and quality assessment. The EPA Draft Report will make ETS the exception. Below, I will discuss many issues but, clearly, the most important flaw in the EPA risk assessment is the Agency's failure to recognize the fact that no reliable quantitative indicator of ETS exposure underpins the epidemiologic data collected to date. The absence of such data prevents reliance upon the epidemiologic data for assessing risk of ETS exposure. II. Exposure Considerations In its risk assessment, EPA states that it relies upon cotinine levels in exposed populations to estimate the risk of lung cancer due to passive smoking, i.e., exposure to environmental tobacco smoke. On page 4-28, EPA states explicitly, Assuming that lung cancer risk from passive smoking is linearly related to cotinine concentrations at these low doses, lung cancer risk of passive smoking can be estimated at the higher exposure level (background plus spousal smoking, applicable to an exposed person) and at the lower exposure level (background only, applicable to a unexposed person), with both estimates relative to the risk of lung cancer risk from zero exposure to ETS. 2

Throughout the EPA Draft Report, the Agency attempts to justify the use of these urinary cotinine levels to quantify ETS exposure. I strongly disagree with this approach as do most scientists and scientific bodies. In the NAS Report on page 6, after a lengthy discussion relative to the monitoring of ETS, the authors conclude, Absorption, metabolism, and excretion of ETS constituents, including nicotine, need to be carefully studied in order to evaluate whether there are differences between smokers and nonsmokers in these factors. Further epidemiologic studies using - biological markers are needed to quantify exposure- dose relationships in nonsmokers. On page 70, after discussing tracers for ETS and specifying the criteria which must be satisfied in order for a marker or tracer for quantifying ETS concentrations to be adequate, these same authors state, While a variety of measures have been used as proxies or tracers of ETS, no single measure has met all the criteria outlined above, nor has any measure been universally accepted or recognized as representing ETS exposure. After a thorough review of a number of studies estimating the risk of exposure to ETS, a report prepared by Meridian Research to the Office of Health Standard Programs of OSHA concluded on page 31, Thus, considerable uncertainty surrounds the use of urinary cotinine levels to predict the quantitative risk associated with ETS exposure. Subsequent work has not eliminated that uncertainty. Recently (1989), Dietrich Hoffmann opined that, The differences in the elimination time of cotinine from urine preclude a direct extrapolation of cigarette-equivalents to smoke uptake by involuntary smokers. 3

It is likely that if it were in any way possible to utilize nicotine/cotinine as a marker for ETS, Dr. Hoffmann would embrace it. His deprecation of the utility of this marker suggests that its use in a risk assessment context is unscientific at best. The use of cotinine (whether or not it is linearly proportional) to establish exposure and, therefore, an estimate of ETS risk, needs to be abandoned. In my opinion, the best method of assessing exposure to ETS, lacking a specific marker, is to focus on respirable suspended particulate matter (RSP) contributed to the ambient environment by cigarette smoking. This method does have limitations including the presence of ETS constituents of interest in the vapor instead of particulate phase. One thing is clear, however. Historically, in all health-effects studies conducted on mainstream smoking or ambient airborne carcinogens, primary emphasis was placed on particulate matter in general and, more specifically, on the particulate matter which falls into the respirable range (0-10 microns). Most observers have considered this particle size the most likely to be inhaled and retained in the human lung. In the EPA Draft Report, the Agency attempts to tie cotinine/nicotine levels to respirable suspended particulate. The implication fs that there is a quantifiable relationship between cotinine concentrations in the urine and the amount of ETS respirable suspended particle which is inhaled by the passive smoker. No such relationship has been, or is likely to be, identified. It has been demonstrated repeatedly that nicotine in ETS partitions to the vapor phase and, when equilibrated, in no way relates to what the level of ETS particulate is in any ambient environment. Thus, attempts to relate nicotine or cotinine levels to particulate exposure have 4

been futile. What is left, therefore, is the direct measurement of the contribution of ETS to particulate loading. At page 25 in Appendix C of the EPA Draft Report, Table C-4 contains a variety of measurements. Note in Table C-4, reproduced and attached to these comments, that the ETS (RSP) measurements vary from a low of 30 ug/m3 in residences and hospitals to a high of 1000 ug/m3 in coffee houses and restaurants. Since it is known with a reasonable degree of accuracy what percentage of our daily lives is spent in each of these environments, it is neither impossible nor technically indefensible to quantify exposure to ETS utilizing this information. However, RSP measurements provide only an upper bound on the level of ETS in these environments. Many other sources of RSP exist. As early as 1980, authors, such as Repace, were attempting to rely on respirable suspended particulate measurements to quantify ETS exposure. Figure 1 contains a series of measurements made by Repace et 0. of a non- smoker going about his/her everyday life. Note that these RSP measurements vary as a function of where the individual is; indoors vs. outdoors, in the home vs. in the office, and they provide a reasonable estimate of the concentrations of RSP to which this nonsmoker is exposed, both' independent of and in the presence of smokers. Referring again to Table C-4, note that the average residence has roughly 80 Lg/m3 of RSP. Returning to the original studies of Repace and Lowery and other supporting studies, it would not be unreasonable to assume that at least 40 ug/m3 of this amount is particulate matter from sources other than environmental tobacco smoke. For offices, the level is 90 ug/m3; for 5

restaurants, 1000 ug/m3. Any time spent outdoors would involve inconsequential exposure to environmental tobacco smoke. In Table AVC-1, I put forth a simplistic model for estimating the integrated exposure of a non-smoking individual in two scenarios; one in which smoking is allowed in the workplace and one in which smoking is not allowed in the workplace. It is assumed that this individual lives with a smoker and works independent of the smoking spouse. I have estimated that this person spends one hour per day out-of-doors. In many cases, it would probably be much more. However, I have attempted to be conservative. I assume also that one hour per day is spent in a restaurant which, for most people, is not the case. Based on the average individual inhaling 20 m3 of air per day, I calculated the fractional amount of ug/m3 inhaled in each exposure scenario and determined the average amount of RSP inhaled as well as that which is retained under the assumption (which has been verified) that only 11% of inhaled ETS/RSP is retained in the lung. Note that I have not corrected for background RSP. Therefore, all values represent a worst-case estimate and assume all RSP is ETS-derived. Note further that I have assumed one cigarette contains 15,000 ug or 15 mg of tar and the amount of tar retai'ned by the average smoker per cigarette is 12,750 mg or 85%. This has also been documented. From Table AVC-1, it is clear that on an average day, under the worst circumstances, a non-smoker living with a spouse who smokes will inhale and retain 245 ug of RSP. This represents roughly 0.019 equivalents of a cigarette. In short, on a per-day basis, this individual would inhale the equivalent of .02 of a cigarette; per year he/she would inhale the equivalent 6

of .02 x 365 or 7 cigarettes. Adjusting for background exposure will reduce the exposure to about 3.5 cigarettes per year. It is well to recall at this stage that the nonsmoker in the first scenario will work, on the average, only 240 days per year. If this is accounted for, average RSP uptake per day in the workplace will be lower. Furthermore, if I account for ambient background of about 40 ug/m3 RSP (non- ETS), my best estimate is that the average nonsmoker, under these conditions, would inhale between 3 and 4 cigarette equivalents per year. Numerous investigators have attempted to argue that the exposure to ETS varies between .1 and 1 cigarette per day. It is well to recall that when converted to cigarette equivalents under the assumption that only 11% of ETS/RSP is retained in order for an ETS-exposed nonsmoker to inhale and retain the equivalent of 1 cigarette per day (15,000 ug), he/she would need to inhale 136,500 ug because only 11% is retained. Under the assumption that an individual would inhale 20 m3 per day, the ambient concentration of RSP would have to be 6,825 ug/m3, almost 7 mg/m3 of inhalable smoke. The EPA currently has a standard for inhalable particulate matter of 150 ug/m3 which is not to be exceeded over a 24-hour period. In essence, ETS- laden environments would have to be roughly 46 times more concentrated than the worst-case scenario that EPA envisions. Based on my research, I have not found data to support atmospheric loading of ETS particulates anywhere near this range. Moreover, this level must be breathed 24 hours a day. Even one-tenth of a cigarette/day seems far too large. It still requires nearly 700 ug/m3 of RSP in the ambient environment, 24 hours per day. This concentration is 200 ug/m3 above the de-minimus NOEL identified for EPA ~ as severely causing or exacerbating pre-existing chronic obstructive pulmonary ~ I 7

disease, asthmatics and others. To my knowledge, with the exception of crowded restaurants which people occupy over a very short period of time, these concentrations are never reached. In addition to the foregoing, I have also examined the amount of rTS RSP retained daily by a person married to a smoker who follows the activity/exposure pattern seen in Figure 1 from Repace. These data appear in Table AVC-2. Here it is assumed the person has a smoking spouse and works with smokers. Note the difference the smoking spouse adds to retained ETS on a daily basis is ± 25 ug, which is a trivial amount. In cigarette equivalents, this amounts to 0.002 per day or 0.72 cigarettes per year. In the British Journal of Cancer (1988), Darby and Pike examined the effect of ETS exposure on lung cancer risk at age 65 using their risk assessment model. Table I is reproduced from page 828 of their report. Note that the worst case is exposure from age zero to age 65 and that at one-tenth of a cigarette, the relative risk is 1.07. These authors are of the opinion that most non-smoking individuals who are married to smokers will most probably fit into the second category which is exposure from age 20 to age 65 only. In this case, one-tenth of a cigarette will produce a relative risk of only 1.04. However, the risk, if the true exposure is .02 or oless cigarette equivalents per day, is clearly smaller than 1.07 and, in my opinion, is statistically indistinguishable from 1. Keep in mind that Darby and Pike use the multistage model and their results are in general concurrence with other modeling efforts of this type. The use of the multistage model by these authors and others has wide appeal within the environmental regulatory context. Even if their results are accepted at face value and used within this context, it is clear that the 8

carcinogenic effect of ETS exposure to as few as 0.02 or less cigarette equivalents per day is trivial at best. Thus, the risk relative to a non- exposed nonsmoker will be comparable, i.e., virtually 1, implying zero risk. It is well to keep in mind that, at best, extrapolation models, such as the multistage model used*by Darby and Pike and my curve-fitting example referred to below, are mathematical constructs. They incorporate numerous theoretical assumptions regarding the mechanisms of carcinogenesis and rely on statistical methods to validate their applicability to observational data. They do not, however, explain observational data. As such, these models, regardless of their theoretical merits, are not suitable for the purpose of demonstrating causation. My own curve-fitting analysis of the epidemiologic data indicates that the de-minimus NOEL for lung cancer approaches 6 cigarettes per day. This comparison attributes a mortality index of I to never-smokers. Thus, the average smoker would need to inhale in excess of 90,000 ug per day of tar or RSP before any noticeable change in lung cancer incidence is observed. My calculations indicate that the average retention would be 85%. That would mean 76,500 ug need to be inhaled and retained daily by a direct smoker before .any noticeable increase in lung cancer incidence would be observed. My worst- case calculations (Table AVC-1) indicate that the average passive smoker inhales and retains 245 ug per day, when compared to the smoker, this is 312 times less than the NOEL. The usual admonitions that go along with the derivation of causal inference from data which demonstrate weak associations apply emphatically to ETS. It is well to keep in mind that in no epidemiology study thus far conducted with ETS, does the relative risk reach 3.0. In fact, in most cases, 9