Glantz: OSHA Post-Post Hearing Comment


Stanton Glantz, January 23, 1996


Stanton A. Glantz, PhD
Professor of Medicine
University of California
San Francisco, CA 94143-1024 January 23, 1996

This comment responds to issues raised in several post hearing comments that OSHA requested I review. I have not reviewed the entire record submitted during the post-hearing comment period, but only the post hearing comments OSHA asked me to review.

Gio Gori (No. 410)

Gori's post-hearing comment is essentially a restatement of his written and oral testimony and contributes no new information to the record. He continues to restate his two views that scientists who find that ETS is dangerous are somehow "going along with the crowd" and are somehow lacking in scientific integrity and that it is impossible to conclude anything about small risks from epidemiology. He also continues to assert his view that no animal experiments are relevant to humans. If medical science adopted this view, most biomedical research would stop.

Gori's argument that epidemiologists cannot reach conclusions about small risks is based on a news article in Science [1] in which several leading epidemiologists provide a candid discussion of the problems of identifying small risks. Gori (p. 4) states that Prof. Dimitrios Trichopoulos "holds that relative risk or odds ratios values below 3 or 4 are not interpretable." Gori ignores the clarification of this view Trichopoulos published [2] in a letter in Science; Trichopoulos wrote:

"Taubes [1] writes that I have expressed the view that only a fourfold risk should be taken seriously. That is correct, but only when the finding stands in a biological vacuum or has little or no biological credibility. We all take seriously small relative risks when there is a credible hypothesis in the background. Nobody disputes that the prevalence of boys at birth is higher than that of girls (an excess of 3%), that men have a 30% higher death rate compared to women of the same age, or that fatality in a car accident is higher when the car is smaller. [emphasis added]"

The risks that Trichopoulos discusses are similar in magnitude (or smaller) than the risks associated with breathing ETS. Gori's representation of Trichopoulos' position to imply that it is impossible to reach conclusions about small risks from epidemiology is misleading and incorrect. It seems reasonable to assume that Gori was aware of this statement because he published a letter on the same article in the same issue of Science [3]. One wonders why Gori failed to provide an accurate representation of Trichopoulos' views.

It is also important to emphasize that the news article [1] is about the limits of drawing conclusions based on epidemiology alone. As the epidemiologists quoted in Taubes' news article stated in a letter to the editor [4], it is important to consider all the evidence, including epidemiology, biochemical, animal, and human experimental studies. (I made the same point in my opening comments in my testimony (Tr. 402-410).) Considering the weight of all the evidence leads to strong conclusions that passive smoking causes lung cancer, heart disease, and other problems.

Gori continues to insist, without presenting any empirical data to support his assertion, that there are threshold effects for lung cancer induced by ETS, and supports his assertion by citing a low quality and out-of-date study prepared by two economists at the Congressional Research Service [5]. In contrast, the most recent report on ETS by the Congressional Research Service, which was written by life scientists, explicitly recognized that there is no evidence for assuming a threshold effect [6, p. CRS-50]:

"It is important to point out that the threshold illustration is a hypothetical example and does not mean that any lung cancer which might result from ETS exposure would actually exhibit a threshold dose response relationship. While data from some studies have shown such a behavior as seen in the previous chapter, the statistical power of those studies is too weak to conclude that such a behavior exists. The use of a threshold model in these calculations is only to simulate the upper limit of a possible upward dose response behavior in order to bracket the range of consequences of possible dose response relationships. Finally, even if a threshold model were approximately correct, public health officials may still chose to use a model closer to the no-threshold approach in order to build in ensure [sic] that all populations are protected. [emphasis added]

"It is also worth noting that the second, and more recent, CRS report [6], which has been mischaracterized by the tobacco industry and its allies [7,8] as discrediting the EPA report on passive smoking and lung cancer [9], actually provides results consistent with the EPA. The second CRS report [6, p. CRS-2] says:

"Calculations based on data from the Fontham et al study [the largest, best study of passive smoking and lung cancer, published after the EPA report was completed in 1992] and assuming an average exposure for the entire population at risk (a no-threshold model) result in a range of 470 to 5500 annual lung cancer deaths in the U.S. from ETS with a mean value of 2780. This compares to a mean value of 3300 calculated by the EPA under the same assumption. Data from the Brownson et al study, on the other hand, produce no annual lung cancer deaths from ETS under the no-threshold assumption. If a threshold model is used to simulate the upper limit of a possible upward dose response behavior, the mean number of lung cancer deaths is 440 calculated from the Fontham et al data and 530 for the Brownson et al data. Over 70 percent of these deaths calculated in the no-threshold example and all those calculated in the threshold model occur to individuals who are exposed to both spousal and background ETS. The remaining deaths in the no-threshold model would result from exposure only to background ETS."

Given the uncertainties in computing these estimates, there is no practical difference between 2780 and 3300.

The second CRS report also recognizes that the EPA report has received wide support from the scientific community and that virtually all the criticism is coming from the tobacco industry [6, p. CRS-6]:

"The EPA report received widespread support from the public health community and from the larger scientific community. But it has been criticized by tobacco industry researchers and scientific consultants. A few independent statisticians and epidemiologists have also raised objections to EPA's statistical analysis of the ETS epidemiologic studies."

All the "independent" statisticians and epidemiologists that are critical of the EPA that I have seen have ended up having (often quiet) financial ties to the tobacco industry [10-14].

During the Hearing, the tobacco industry placed a report, Choices in Risk Assessment report [15], in the record by the tobacco industry; Choices discussion of ETS was almost entirely based on the first CRS study [5], which has been superseded by the second CRS study [6]. This fact is another reason that Choices does not represent the best available evidence; OSHA should discount this report in its deliberations.

It is also important to emphasize that even the second CRS study [6] took a very narrow view of the issue of ETS and lung cancer and simply looked at the studies published since the EPA report was issued. It would have been better to provide a review of all the data. The bottom line: despite the weakness of the approach in the second CRS report, it reaches essentially the same conclusions as the EPA and is certainly not a repudiation of the EPA.

The California Environmental Protection Agency has released a more complete and careful review of all the evidence on ETS and lung cancer, with special emphasis on studies published after the EPA report [16, p. 7]. This report concludes:

"In 1992, the U.S. Environmental Protection Agency (U.S. EPA) issued a report in which it concluded that ETS is a human lung carcinogen. Similar conclusions were reached in earlier reports by the U.S. Surgeon General (U.S. DHHS, 1986) and the National Academy of Sciences (NAS, 1986). The present document reviews the four U.S. studies that were published since 1991 and were not included in the U.S. EPA document. Results from the largest population-based study, conducted in several metropolitan areas of the U.S. (Fontham et al., 1994), were closest to the pooled estimate from the U.S. EPA report. Of the two other population-based studies, the association found in a Florida study (Stockwell et al., 1992) was stronger and that in a Missouri study (Brownson et al., 1992) was weaker than that indicated by the EPA's pooled estimate. Although the authors of the fourth study, a hospital-based case-control study (Kabat et et al., 1995), interpreted their findings to be unsupportive of an association between ETS exposure and risk of lung cancer, their results (analyzing men and women separately) were based on much smaller sample sizes than the three population-based case-control studies; nevertheless, the odds ratio we calculated from their results (for men and women combined) was in fact very similar to the U.S. EPA result. Despite the compelling biologic plausibility of an effect of ETS on lung cancer, detection has been problematic because a small excess in risk is difficult to establish in a single epidemiologic study. The three population-based studies published since the U.S. EPA report, by design, have successfully addressed many of the weaknesses (i.e., small sample size, possible selection bias, possible misclassification biases, inadequate adjustment for potential confounders) for which the previous studies on ETS and lung cancer have been criticized. The concordance in these study results gives further credibility to the finding of a causal association between spousal ETS exposure and risk of lung cancer described in the U.S. EPA report."

With regard to workplace exposures, the California Environmental Protection Agency concludes that [16, p. 25-26]:

"Studies in which the assessment of workplace exposure to ETS was complete (covering all jobs) with considerable ETS exposure of subjects in the studies are generally supportive of an association between workplace ETS exposure and risk of lung cancer (Wu et al., 1985; Wu-Williams et al., 1990; Fontham et al., 1994) suggested a trend of increasing risks with increasing duration of ETS exposure at the workplace. Compared to women who had no ETS exposure at the workplace, women who reported exposure for 1-15, 16-30, and 30 or more years showed adjusted odds ratios of 1.30, 1.40, and 1.86, respectively (p for trend = 0.001) (Table 7.7)

"In addition to the incomplete assessment of exposure to ETS at the workplace in some studies, respondents, particularly surrogate respondents, may be less able to provide information on the subjects' exposure to ETS at the workplace. In a study in which a test-retest design was used to examine the reliability of passive smoke histories reported in personal interviews, self-respondents more reliably reported residential exposure than exposure at work (Pron et al., 1988). This may be a particularly important problem in studies in which the proportion of surrogate respondents was high (Brownson et al, 1992; Stockwell et al., 1992).

"Despite some of the above mentioned difficulties in obtaining histories of lifetime ETS exposure at the workplace, there is reason to believe that this source of ETS exposure also increases the risk of lung cancer, as does ETS exposure from spouses. The workplace has been a major source of ETS exposure outside the home (Cummings et al., 1989 and 1990; Emmons et al, 1992; Siegel, 1993), although the relative importance of workplace ETS exposure may be declining in California as the result of increasing restrictions on smoking in the workplace. In the International Agency for Research on Cancer (IARC) 10-country collaborative study which correlated urinary cotinine levels to self-reported recent exposure to ETS at home (from spouses), in the workplace, and other social settings, Riboli et al (1990) found that exposure to ETS at the workplace was a significant predictor of cotinine levels, similar to ETS exposure from spouses."

It is worth noting that the tobacco industry's misrepresentation of the second CRS report [6] as contradicting the EPA report is not the first time that the tobacco industry has misrepresented scientific reports in an effort to discredit other scientific reports. (Indeed, the tobacco industry's efforts to "create controversy" where none existed has been a feature of its strategy for dealing with the evidence that passive smoking is dangerous since the earliest evidence that passive smoking caused lung cancer emerged in the early 1980s [17].) For example, in 1986, the US Surgeon General published "The Health Effects of Involuntary Smoking" [18] that concluded that ETS caused lung cancer, was bad for kids, and that simply separating smokers and nonsmokers would "reduce, but not eliminate" the risks of passive smoking. About 6 weeks later, the National Academy of Sciences produced its "Environmental Tobacco Smoke" report [19], which concluded the same things. Even so, the tobacco industry loudly claimed that the NAS report discredited the Surgeon General.

There is another point to keep in mind here: The EPA report, while very important, was not the first time that major scientific bodies concluded that ETS caused cancer. The Surgeon General [18] and NAS [19] reached this conclusion a decade ago. Viewed in this context, the controversies surrounding both CRS reports are another tempest in a teapot being stirred up by the tobacco industry based on hot air rather than scientific substance.

Most of Gori's criticisms of our experimental work on passive smoking and heart disease [20,21] are simply restatements of points he has made earlier and that were addressed in the original publications, my testimony at the hearing, and my post hearing comment. His primary new criticism is that our results are an artifact of increased catecholamine levels due to the stressful conditions in the experiments. As noted in my original testimony (Tr. 422-423, 701-704, 727-728), we have specifically excluded this possibility by showing that administration of beta blocker drugs, which block the effects of catecholamines, did not prevent the ETS effects [22]. Moreover, all our work includes control groups of animals housed under the same conditions eating the same diet; the only difference between the animals in the control and exposed groups is the presence of ETS. In our papers we documented that the exposure conditions were reasonable for this sort of toxicological study, particularly given the short term exposures.

Gori also misunderstands or misstates the presentation of results on the effects of ETS on infarct size [21]; the results are nondimensional (i.e., infarct mass as a percentage of muscle mass at risk). Moreover, new experiments are beginning to indicate the mechanisms by which ETS increases the risk of myocardial infarction. We [23,24] have shown that it is possible to block the effects of ETS on infarct size with L-arginine. The effects of ETS appear to act at a cellular level rather than through changes in hemodynamics. This result suggests that ETS interferes with endothelial function in ways that aggravate the effects of myocardial infarction. As little as one hour a day of ETS exposure (at home or at work) produces significant changes in endothelium-dependent arterial dilation in healthy young humans [25]. Moreover, the effects of passive smoking on endothelium-dependent arterial dilation are similar to the effects of active smoking. This result provides additional biological evidence to support the large (compared to the dose) effects of ETS on the cardiovascular system. These results also indicate that even short-term exposures to ETS produce demonstrable adverse effects on the cardiovascular system. ETS exposure also interferes with normal endothelial responses to acetylcholine [26]. These results point to some of the specific biological mechanisms at a cellular and molecular level through which ETS adversely affects the cardiovascular system.

Gori raises essentially the same criticisms of the work of Penn et. et al. [27,28], which are no more valid than his criticisms of our work and which have been answered by Penn in response to tobacco industry-generated letters to the editor [29]. Indeed, the fact that two different investigators, using different animals and experimental models, reached the same conclusions strengthens the confidence OSHA can have in all the results of these experimental studies. None of the issues Gori raises refute the fundamental result that ETS induces heart disease-like changes in experimental animals exposed to ETS.

Gori's effort to distance himself from his identification of nicotine as an addictive drug (p. 15-18 of his post-hearing comment) has no relevance to the OSHA rule making, but it is amusing. Given the widespread acceptance of nicotine as an addictive drug by authorities including the US Surgeon General [30], the Food and Drug Administration [31,32], and tobacco industry [33], one would have thought that Gori would have been proud to have been one of the early scientists to identify nicotine as an addictive drug. His semantic games and citation of the 1964 Surgeon General's report [34], while ignoring the 1988 report [30] to wiggle out of this position -- which is obviously damaging to his tobacco industry clients -- calls into question the credibility of his entire presentation.

Joseph Wu (No. 422)

Wu did not appear at the hearing, so it was not possible to question him regarding his views.

Like several other submissions by the tobacco industry, Wu recites a long list of technical details about the mechanisms by which different events occur in the cardiovascular system. While this is interesting, it is beside the point. None of this material demonstrates that ETS does not affect the cardiovascular system or provides alternative, plausible explanations for the experimental findings that ETS affects lipid depositions, development of atherosclerotic plaque, size of myocardial infarct, exercise tolerance, platelet activity, ATP production, or any of the other changes that have been demonstrated to occur under controlled conditions during ETS exposure. (For citations, see our review papers [35,36], the American Heart Association position statement on passive smoking and heart disease [37], my testimony or Ford's testimony.)

Wu's statement that my interpretation of the evidence on passive smoking and heart disease is "based on unsound and selective interpretations of existing data that are lacking in scientific validity" (p. 2) is typical of the kind of hyperbole that the tobacco industry loves. While this sounds like a damning statement, Wu does not point to a single experimental or epidemiological paper on ETS that we ignored. Wu's numerous statements that our conclusions are not based on data are also incorrect; our papers [35,36] and my testimony was well supported with citations to the peer reviewed literature. It is also important to emphasize that these publications passed rigorous independent scientific peer review in leading journals (Circulation [35] and JAMA [36]), which apply high standards of proof. If Wu's assertions were true, these journals simply would have not allowed us to publish papers with unsupported statements. It is also worth noting that the review paper we published in JAMA [36] was essentially the same material as presented to OSHA in my testimony. The fact that this material passed the rigors of independent peer review by one of the leading medical journals in the world increases the level of confidence OSHA can have in the findings.

Wu presents a discussion of the effects of lactate on vascular relaxation (p. 3-4). While interesting, it does not say anything one way or the other about the effects of ETS. To the best of my knowledge, no one has yet pointed to lactate as a mechanism for ETS-induced cardiovascular effects. He also provides theoretical arguments as to why ETS should not affect endothelial function (p. 8-9). In contrast to Wu's theoretical arguments, there is direct evidence that ETS impairs normal endothelial function in humans. For example, Adams et et al. [38] recently demonstrated that passive smoking is associated with dose-related impairment of endothelium- dependent dilation in healthy teenagers and young adults, suggesting early arterial damage, which helps explain the observed increase in risk of heart disease in passive smokers. The effects of passive and active smoking were similar in magnitude, providing additional evidence against the practice, advocated by the tobacco industry, of using "cigarette equivalents" to assess the effects of ETS. In contrast to Wu's discussion of side issues, these data provide direct evidence of an effect of ETS on humans.

Likewise, Wu presents a general discussion about why ETS should not cause atherosclerosis, but he fails to explain why our group [20,22,39] and Penn's group [27-29] found these effects. He raises questions about diet and stress (as the other tobacco industry witnesses did). These questions ignore the fact that all the experiments in question included control groups in which the animals were fed the same diet and housed under the same conditions. The only differences between the ETS-exposed animals and the control animals was the presence of ETS. It is hard to see how anything but ETS could be explaining the differences associated with ETS exposure. Wu proposes no plausible explanations.

Wu presents a lengthy discussion of mechanisms of platelet aggregation (p. 14-16) which fails to explain why so many investigators have observed an increase in platelet aggregation in humans and animals with exposure to ETS (see [35] and [36] for references). The only logical explanation is that ETS exposure affects platelet activity. While Wu argued that there might be differences between in vitro and in vivo measurements of platelet activity, he does not provide any positive empirical evidence that this difference can explain the observed effects of ETS on platelet activity. Again, it is important to emphasize that the experimental work in this area included control groups, so any errors or biases that might be associated with the measurement method would be present in all the data. It is hard to believe that there would be differential errors in measuring blood samples taken from people or animals who happened to be breathing ETS compared to the similar people in the control groups. The only plausible explanation for the observed differences associated with ETS exposure in these experiments is the ETS exposure itself.

Wu's comment is another example of the tobacco industry's penchant for grasping at straws for any possible explanation for observed connections between passive smoking (and, for that matter, active smoking) and disease. The simplest and most plausible explanation for the body of evidence on passive smoking and heart disease is that passive smoking causes heart disease.

Peter Lee (No. 356)

Peter Lee is another of the tobacco industry witnesses who did not appear at the Hearing, thereby depriving OSHA and the public of the opportunity to question him regarding his views.

I believe that his comments are mislabelled: Mr. Lee does not have a PhD.

Lee presents a lengthy argument for a threshold effect of ETS on lung cancer risk (p. 5-8). As with the other tobacco witnesses, Lee presents no affirmative evidence to support this assertion. Indeed, as noted above, the second (and most recent) CRS report [6] has explicitly recognized the fact that any analysis based on thresholds is speculative. Even assuming a threshold, however, there would be hundreds of deaths from ETS-induced lung cancer. It is important to emphasize that none of the major national scientific consensus reports [9,18,19] recognizes any evidence that there is a threshold effect for ETS-caused lung cancer. OSHA should not base its risk assessment on the assumption that there is a threshold effect for ETS-induced lung cancer (or, for that matter, heart disease or anything else).

The refusal of tobacco industry consultants to accept the evidence that ETS causes any disease whatsoever is well-illustrated by Lee's meta-analysis of ETS and lung cancer (p. 8). Despite the fact that he reports a statistically significant increase in the risk of lung cancer when analyzing all the studies (RR=1.13, 95% CI 1.05-1.22), he refuses to conclude that ETS causes lung cancer because he fails to find a statistically significant elevation in risk when he limits himself to only the US studies (RR = 1.11, 95% CI = 0.99-1.24). There are two points worthy of comment. First, the fact that the confidence interval is wider when it is computed based on less data is a standard statistical result. Indeed, Lee could obtain a non-significant result by simply limiting his selection of studies for inclusion in order to make the effective sample size small enough. Second, 0.99 is pretty close to 1.00; while it is strictly true that the confidence interval includes 1.00, most investigators would hardly find Lee's US result "negative." At the very least it would be considered as "strongly suggestive" of an association. Lee tries to wiggle out of this conclusion (on p. 9) by repeating the tobacco industry's line that only large relative risks should be considered real (he says somewhere between 1.5 and 3.0). This line of argument was addressed and rebutted above (in response to Gori's comments). As I have stated several times, OSHA should use all the relevant evidence in its deliberations; by this standard, Lee's result supports OSHA's determination that ETS increases the risk of lung cancer.

Lee's comments on one- versus two-tailed tests are silly (p. 9), particularly when he states "I do not believe one can ever be certain any agent might not protect against a disease." Does he really think exposure to ETS could protect people from developing cancer or heart disease? He has certainly never advanced any empirical evidence to support this assertion. The implicit assertion that "a little ETS can be good for you" is another of the silly propositions that the tobacco industry and its consultants have advanced in an effort to avoid accepting the obvious conclusion that ETS causes disease.

Lee's discussion of misclassification (p. 19-26) avoids the key point that misclassification of smokers as nonsmokers (and vice versa) can only be important if active smoking causes disease in smokers. As discussed in my post-hearing comment, tobacco industry witnesses have assiduously avoided admitting that smoking causes disease in smokers. This failure to deal forthrightly with the effects of active smoking makes it hard to take seriously the other arguments that the tobacco industry's witnesses make seriously. (It would have been interesting to cross examine Lee on this point at the Hearing, but he decided not to attend.) Unless and until Lee -- and the other tobacco industry witnesses who are claiming that misclassification explains the observed link between passive smoking and disease -- unequivocally states that active smoking causes lung cancer and heart disease, OSHA should simply ignore their arguments on misclassification errors.

With regard to confounders, Lee noted (p. 27) that I asked LeVois whether he had considered the list of 33 confounders in the paper by Thornton, Lee, and Fry [40]. During the hearing LeVois admitted to having not controlled for these 33 confounding variables in his study (Tr. 5886-5890), yet Lee is not critical of LeVois' study, and indeed, backpedals on the importance of his 33 potential confounders in his post-post-hearing comment (p. 27-38). Lee's inconsistent application of his standards to papers that agree and disagree with the tobacco industry's position (including the dramatic shift in his position regarding the proper interpretation of the Thornton et al. [40] paper from his pre-hearing submission to his post-post-hearing comment) is one more reason that OSHA should not take any of Lee's written comments seriously.

Lee's distinction between active smoking and passive smoking in terms of effects on the heart (p. 58-59) is silly; no one is saying that the effects of ETS and active smoking are so different that they should be considered as totally independent risk factors. Indeed, a great deal of evidence (reviewed in [35] and [36], together with more recent studies [38,41]) demonstrates that in terms of the cardiovascular system nonsmokers exhibit large responses to small doses of ETS, often approaching that of habitual smokers. The evidence indicates a steep dose-response curve at low doses, with decreasing marginal effects as dose increase. Thus, one would not expect to see a large cardiovascular effect of ETS on active smokers.

Lee's unwillingness to accept the scientific evidence that ETS is dangerous is further illustrated when he presents his own meta-analysis of the available data, including the tobacco industry's analysis of the ACS CPS and NMFS data sets (p. 59). (As discussed in detail in my post-hearing comment, OSHA should not use the tobacco industry's analysis of these two data sets.) All Lee's estimates of the risk of heart disease associated with ETS are significantly above one. Lee even admits that our analysis is correct and that the elevated risks are unlikely due to chance. Nevertheless, he chooses to dismiss the finding that passive smoking causes heart disease. This frank refusal to accept the results of his own analysis is another reason that OSHA should not take Lee's comments seriously.

Lee devotes considerable energy (p. 60-63) to justifying the use of the ACS CPS data sets to analyze the effects of ETS on heart disease and urges the ACS to conduct such an analysis. As noted in my post-hearing comment, ACS did conduct such a (preliminary) analysis, which confirmed that ETS exposure is associated with an increase in heart disease.

In contrast to the exacting standards Lee demands of OSHA, the EPA, and other scientific authorities, he is remarkably forgiving of his fellow tobacco witness LeVois. He did not seem the least bit bothered by the fact that LeVois "lost" 13,000 cases in his analysis of the ACS CPS data set (p. 63). Likewise, he does not apply the same demanding standards to the tobacco industry analysis of the NMFS (p. 64) as he demands of legitimate scientific investigations. This is another example of the inconsistent standards that the tobacco industry witnesses apply to evidence that does not further their clients' needs.

It is interesting to note, however, that despite his reinterpretation of the Thornton et al. paper [42] in order to justify LeVois' failure to account for all the confounding variables earlier in his post-post-hearing comment (p. 27-35), Lee has returned to his original interpretation of the Thornton et et al. paper when criticizing our work (p. 67-68). It would be nice if Lee could make up his mind on how to apply this paper.

The remainder of Lee's comments on heart disease are simply a rehash of the arguments he presented in his earlier submissions and do not warrant additional response.

Maxwell Layard (No. 407)

Layard disputes the statement that risk of heart disease drops over time after an individual stops smoking (p. 4) and presents data from a few studies (p. 7-8) that he claims do not show such a drop. Careful examination of the results Layard presents (in his Tables 1 and 2) shows that, except for British doctors over 64, there is a substantial drop in the risk of heart disease death over time following smoking cessation. (The result for elderly British doctors is not surprising, since the excess risk associated with smoking for heart disease falls with age.) As Layard notes, I said that the excess risk declines almost to that of never smokers; the relative risks around 1.1-1.2 observed several years after cessation are almost that of never smokers. More important, Layard does not provide a quantitative analysis to demonstrate that the major misclassification errors in his study do not introduce serious biases.

Layard misses the point of the criticisms of his use of the NMFS data set for his study when he defends the use of death certificate data (p. 8-12); the problems with his use of the NMFS data set are not in terms of identifying the cause of death of the index cases. The problems are in assessing whether or not they were currently exposed to ETS at the time of death. All Layard's computations address the wrong misclassification problem; Layard should have addressed the problem of including currently unexposed people (who were married to former smokers) as exposed to ETS. As Layard admitted (p. 11-12), these issues were carefully elucidated in the Hearing (Tr. p. 6617). It is unfortunate that Layard chose not to provide a careful analysis of the real problems with his study. Given this failure to address the serious problems raised at the Hearing, OSHA should not consider this study in its final risk assessment.

Layard objects to the criticisms of his analysis of the ACS CPS data sets as "gratuitous criticism" (p. 39), yet he does not respond to the specific issues that were raised at the Hearing. In response to the criticism that his paper did not have a "methods" section (Tr. 482), he states that it did. While this is technically true -- there was a section labeled methods -- Layard ignores the point that this methods section was so sketchy that a reader could not reproduce his analysis (which is what the methods section is supposed to provide). More important, major issues, such as the fact that LeVois and Layard's analysis of these data sets seems to have dropped 13,000 cases are not addressed at all. Even Lee (in his post-post-hearing comment on p. 63) expressed hope that LeVois and Layard would resolve this discrepancy "in due course." The fact that Layard has chosen to ignore this problem -- rather than explain it -- discredits his entire analysis. If OSHA elects to include these data sets in its risk assessment, it should rely on the analysis prepared by the American Cancer Society.

Layard's embracing of the small, negative Kabat et et al. study [43], while ignoring larger, better done positive studies (such as Fontham et et al. [44]) is yet another example of how tobacco industry witnesses' willingness to selectively accept and reject data as it suits their needs (p. 41). Moreover, the careful assessment of the Kabat et et al. study conducted by the California Environmental Protection Agency [16] shows that the point estimates of risk, properly analyzed, are consistent with the larger, better done studies. (See discussion of this point in the section on Gio Gori.)

Domingo Aviado (No. 321)

Aviado simply restates the points he made in his written testimony and previous publications. Since Aviado refused to appear at the Hearing and defend his position, there is nothing new to add.

Gordon Newell (No. 333)

Newell cites several papers to support his contention that carbon monoxide at levels present in ETS is not adequate to produce adverse cardiovascular effects, while ignoring the largest study by Allred et al. [45], which shows effects at all levels of exposure. The Allred et al. data shows a continuous, non-threshold effect. There is, however, considerable inter-individual scatter; this scatter may explain why some of the studies which only compared a few levels of exposure did not yield statistically significant results. It is also important to emphasize that no one is saying that carbon monoxide is the sole (or even primary) agent in ETS that produces adverse cardiovascular effects. It is one element among many. OSHA should consider all the evidence.

The evidence of a "controversy" about the Auerbach et al. study is simply a few letters to the editor, including ones published in the popular press. This is hardly evidence of serious scientific disagreement, especially given the tobacco industry's practice of sponsoring scientific letter-to-the-editor writing campaigns [10,12,14].

Newell's statements on the possibility of a threshold effect for ETS are dealt with in my comments regarding Gio Gori.

It is interesting that Newell urges OSHA to consider the animal data (p. 8), whereas Gori states that animal data should not be considered (p. 5).

Helen Hubert (No. 355)

Hubert essentially repeats her criticisms of the epidemiological studies of ETS and heart disease that were presented in her earlier submissions and states that these same criticisms apply to the newer publications. These issues were addressed in my original testimony (which, in fact, served as the basis for our paper in JAMA [36]) and my post-hearing comment; there is nothing to add.

Hubert cites the fact that we [36] did not include the work of LeVois and Layard and that this failure represents evidence of publication bias against publication of negative papers on ETS and heart disease. It is hard to understand how the LeVois and Layard paper is evidence of publication bias since it has been published [46]. More important, there are so many serious problems with their analysis -- which could not be explained during the Hearing -- that this paper should not be used by OSHA at all. (See my post hearing comment for a detailed discussion of the work of LeVois and Layard.)

Jefferey Idle (No. 319)

Idle claims, without presenting any empirical evidence to support this claim, that there is a threshold effect for ETS; see my comments on thresholds in response to Gio Gori's comments.

Idle's continued insistence that dietary nicotine could explain the observed levels of cotinine in the population simply does not add up. As Repace [47] and Jarvis [48] have pointed out, no reasonable diet could produce the levels of cotinine observed in people exposed to ETS. Repace estimated the average daily dietary dose of nicotine in the US diet would be no more than 0.3 g/day, whereas the typical US nonsmoker would have a nicotine dose from passive smoking of 100 g/day. He also questions whether nicotine could actually be "cooked out" of potatoes and other cooked foods. The boiling point of nicotine is 247oC, suggesting that nicotine could be lost as a vapor during cooking. Idle also ignores the fact that the two vegetables with the highest nicotine content -- potatoes and eggplant -- are not eaten raw in America. In addition, most use of tomatoes is generally after cooking. Indeed, the title to the letters by Repace and Jarvis that the British Medical Journal selected says it all: "Dietary nicotine won't mislead on passive smoking ... unless subjects eat 90 kg tomatoes a day."

Indeed, the issue of dietary nicotine is yet another example of a pattern that permeates all the comments advanced by the tobacco industry and its consultants: They ignore obvious and logical connections between ETS and various outcomes and instead advance plausible sounding, but quantitatively unsupportable, explanations for effects associated with ETS.

WashTech (Nos. 334 and 453)

WashTech submitted a paper on the "Identification and Relative Weight of Cardiovascular Risk Factors" [49], published in 1986. I fail to see the relevance of this paper, since it was published before most of the evidence that passive smoking causes heart disease was published, and well before the American Heart Association identified ETS as a risk factor for heart disease [37] in 1992. It is also important to emphasize that no one is saying that ETS is the only cause of heart disease. It is one of many, often interacting, causes.

WashTech also submitted several papers that point out that heart disease is caused by many factors and that there are several pathways between biological insults and the development of clinical cardiovascular disease. The existence of multiple risk factors and pathways in no way is evidence that ETS does not cause heart disease. Indeed, the thrust of my testimony, as well as the review papers we have published [35,36], is that ETS acts synergistically with other risk factors and through many pathways. This fact is why it is not possible to establish a meaningful PEL for ETS.

Butler/Chem Risk (No. 454)

As with WashTech, Butler/Chem Risk submitted several papers which describe other possible risk factors and causes of heart disease. For the reasons just discussed, the presence of these factors in no way undercuts the evidence that passive smoking causes heart disease, cancer, or other disorders.

RJR/ Nelson et al. (No. 451)

This submission consists of copies of slides presenting results indicating that the amount of pollution created by cigarettes varies with several variables related to smoker behavior. These data provide additional reasons why it is not feasible to create a PEL for ETS exposure.

Grossman-Oliver Economic Consultants (for RJ Reynolds) (No. 451)

Grossman-Oliver's critique of the Glantz and Smith study of the impact of smoke free restaurant ordinances on restaurant revenues [50] is similar to the other tobacco industry consultants' criticisms of the scientific evidence that ETS causes disease: it consists of a laundry list of potential difficulties without ever demonstrating that the problems they discuss are actual problems with the study. This is particularly problematic because all the raw data used for our study are a matter of public record. Rather than speculating about possible problems, one would have expected Grossman-Oliver to actually present an alternative, convincing analysis of the data to demonstrate that we were wrong. The fact that no such analysis was presented suggests that no plausible alternative analysis could be produced.

The first speculation in the Grossman-Oliver report is that, while we showed no systematic change in total restaurant revenues associated with smoke free restaurant legislation, we could not exclude the possibility that some restaurants would experience an increase in business and some would experience a reduction in business (p. 3-6). Given that reliable data on individual restaurants are not available, the statement that Grossman-Oliver make is correct. It is also correct that they cannot demonstrate that such an effect exists. Grossman-Oliver also present extensive speculation about what other factors might influence changes in dining habits (e.g., location, ambience, decor, quality, cuisine), but no data. OSHA's rule making should be based on data, not speculation.

Grossman-Oliver complain that our study did not include representative communities (p. 6-7). We included all communities with 100% smoke free restaurant ordinances. These communities included large cities, small towns, rural, urban, and suburban locations, and a wide variety of incomes. (See Table 1 of our paper [50].) The tobacco industry's consultants suggest that "The experiences of these communities might have been more instructive if cities had been randomly selected as part of an experiment to determine the impact of smoking bans on restaurant sales nationwide." One would hope that Grossman-Oliver would recognize that this is a ridiculous suggestion. Basic democratic principles (to say nothing of the tobacco industry's large political organization to oppose such laws [51-53]) would preclude such an experiment. Any introductory statistics course makes the point that such experiments -- such as experimentally manipulating the unemployment rate to examine the effects on the economy -- are simply impossible. This is yet another example of how the tobacco industry demands ridiculous standards of proof.

Grossman-Oliver is correct (p. 8) in stating that OSHA's proposed rule is somewhat more restrictive than some of the communities we studied. While some of the communities included bars in the smoke free ordinance, some did not. In any event, the contribution of bars to total sales is small and would not affect the overall conclusions of the analysis. Indeed, we stated in our paper that analysis of restaurants with and without liquor licences yielded similar results.

Grossman-Oliver present an extensive list of potential problems with our econometric model (p. 9-18) without presenting any positive evidence that such alternative analyses would yield any different results. As noted above, this is particularly surprising because all the data needed to do such an analysis is a matter of public record. Grossman-Oliver could have simply submitted the results of a study that embodied their various suggestions, but they did not.

In fact, we considered many of the issues Grossman-Oliver raised, such as a quarter-by-quarter seasonal adjustment for all cities and alternative ways of allowing for inflation. Doing so yielded more complicated models but no difference in overall conclusions. The price of these more complex models was a reduction in residual degrees of freedom and, so, power. We presented the simplest model possible in order to have the greatest power to detect an effect of the ordinances, if they existed. (We kept a seasonal adjustment for the winter quarter for the Colorado ski resorts because their business is highly seasonal.) It is also important to note that the Claremont Graduate School of Business [54] developed a model which considered many of the factors Grossman-Oliver mention (as well as others) and reached the same conclusion we did: Smoke free restaurant ordinances do not affect restaurant revenues.

The issue of serial correlation is yet another potential problem. Again, one wonders why Grossman-Oliver did not provide evidence that it was a real problem. (In fact, the issue of seasonal autocorrelation was raised during the review of our paper and we satisfied the reviewers that it was not a problem. The American Journal of Public Health, which published our study, did not think it was worth including this detail in the final paper.)

While the issue of a bound dependent variable is of theoretical concern, it is not a practical problem since none of the observations are near the bounds. We examined the residuals for normality and homoskedacity and did not identify any problems.

In conducting our analysis, we were concerned with multicollinearity and examined the variance inflation factors associated with all our analyses. The actual results do not agree with those presented in the Grossman-Oliver analysis. The correct values of the variance inflation factors are: Auburn 2.0, Bellflower 1.2, Beverly Hills 1.1, El Cerrito 2.0, Lodi 3.2, Martinez 1.9, Palo Alto 2.0, Paradise 2.2, Roseville 2.2, Sacramento 1.7, San Luis Obispo 3.6, Aspen 2.9, Snowmass 3.8, Telluride 2.7. Obviously, the paper is in error when it said that all values were below 2; we have submitted an errata to the American Journal of Public Health changing this statement to indicate that all values were below 4. The reason for this error is as follows: There was such a long delay between the time the paper was accepted for publication and when the galley proofs arrived that one or two more quarters of data became available. With the editor's permission, we changed the tables and figures to include the more complete data set. For the reasons Gross-Oliver note, this increased the variance inflation factors. We only corrected the table and figures and missed this number in the text. We apologize for any confusion. It is important to emphasize that this error makes no difference in terms of the interpretation of the data. A variance inflation factor of 10 is generally considered the boundary for concern about multicollinearity [55, p. 219,56, p. 392,57, p. 193]. As noted above, all the variance inflation factors we found are well below 10.

Grossman-Oliver's criticism of our pooled regression analysis (p. 24) based on the observed values of R2 reflect a lack of understanding of what R2 means. Our model included dummy variables to allow for between-community differences. Including these explanatory variables accounts for the increase in the R2, not a problem with variance inhomogenuity.

Given the resources of the tobacco industry and its consultants, it is surprising that Grossman-Oliver stated that "G-O did not have access to the Glantz and Smith data set" (p. 21). After all, the data are all a matter of public record. (Indeed, another tobacco industry consultant, Price Waterhouse (discussed next), did obtain some of the data.) They could have simply demonstrated that our analysis was wrong by presenting an alternative. Instead, they simply listed several hypothetical problems without demonstrating that they were real problems. This is a repeating theme in the tobacco industry's criticisms of the OSHA Proposed Rule and the science that justifies it.

Price Waterhouse (No. 437)

In considering the Price Waterhouse comment, it is important to emphasize that the only material I put in to the record or presented at the Hearing (Tr. 442-449) was the 15 city study published in the peer-reviewed journal, the American Journal of Public Health [50]. Even so, most of Price Waterhouse's comment consists of a rehash of the issues raised in Rupp's cross examination of me at that Hearing (Tr. 566-630) based on preliminary versions of this paper. It is important to emphasize that I am suggesting that OSHA rely on the final, peer reviewed published version of our work [50], not the preliminary versions of the work Rupp and Price Waterhouse devoted so much time to attacking.

As I explained at the Hearing (Tr. 519, 571-572, 623-625) many of the issues the tobacco industry and Price Waterhouse concentrate on reflect a typographical error in one of our preliminary reports and attacks on the accuracy of data provided by the California State Board of Equalization. Price Waterhouse even appears to question the fact that we found an error in the data published by the State Board of Equalization and requested that they correct it (p. 19); we were simply seeking to base our analysis on the most accurate available data. As noted above, much of Price Waterhouse's criticisms deal with preliminary versions of our work which I did not submit to the record. Since all these questions were raised and answered at the Hearing (Tr. 593-630), there is no need to repeat the answers here.

Price Waterhouse again argues that it is not reasonable to use sales tax data for analysis of restaurant sales and total retail sales (p. 8-9), including Figure 1 (p. 8), which was presented to me during cross examination at the Hearing (Tr. 593-608, 623-626). As I noted at the hearing, the restaurant sales and total retail sales appear to exhibit a substantial cross correlation, which goes against the assertions that Price Waterhous makes regarding the inapplicability of using these data. In addition, I raised problems with using percentage changes rather than variations from the mean. It is disappointing that Price Waterhouse did not address any of the issues that I raised at the Hearing when preparing their post-hearing comment. It is also of concern that Price Waterhouse only presented results from selected cities rather than all of them.

The issue of bar sales (p. 9, 28) is discussed above (see Grossman-Oliver).

Price Waterhouse's reanalysis of the data from San Luis Obispo (p. 11-12) is worthy of comment for several reasons. First, their discussion is based on a preliminary version of our work, not the final published version that was submitted to the record. Second, Price Waterhouse presents an analysis based on data which included a typographical error, despite the fact that at the Hearing (Tr. 519, 571-572, 623-626) I explained that this number was wrong and that the correct number was used in our statistical analysis. To knowingly compute statistical results based on incorrect data is irresponsible and unethical. Third, Price Waterhouse implies that I was trying to hide something when they went on at some length about the correction to the numbers the State Board of Equalization published. Quite the contrary; I was the person who contacted the State Board of Equalization about the potential error, obtained the correct value, and released this information to the public.

Price Waterhouse cannot seem to decide whether or not one should use a model based on total sales. On page 11-12 they use a model based on total sales, without any adjustment for inflation or underlying economic conditions. While I used such a model in one of the preliminary reports of our data, I subsequently abandoned this approach in response to legitimate criticism of this approach by the tobacco industry, its consultants, and front groups. Indeed, Price Waterhouse raises these criticisms itself (p. 6), so it is hard to understand why they use the very model they criticize later in their report (on p. 11-12). Indeed, they later criticize me for abandoning the total sales model on page 18, when it was done to satisfy criticisms like the ones Price Waterhouse makes on page 6. They need to make up their mind about what the proper approach is and apply it consistently. My recommendation is to use fraction of retail sales or ratio with a comparison city to study the effects of smoke free ordinances on restaurant sales because these approaches adjust for inflation and underlying economic conditions. As the tobacco industry critics of our original work correctly noted, inflation and underlying economic conditions are important confounding variables, particularly since the California economy went in to a deep recession about the time many of the local smoking control ordinances passed. Given the tremendous concern raised about confounding variables by many other tobacco industry witnesses (and Price Waterhouse itself on page 6), I am surprised to see them continuing to use an analysis based on unadjusted total sales. This is an incorrect approach.

Price Waterhouse presents an analysis of an "extended data set" for Lodi, which includes data through the second quarter of 1994 that Price Waterhouse interprets as somehow discrediting our approach (p. 19-20). Quite the contrary, their analysis confirms our finding of no effect of smoking restrictions on restaurant sales; neither the P value for fraction of restaurant sales (P=.057) nor ratio of sales with the comparison city (P=.538) are statistically significant. The fact that the coefficients (which are not significantly different from zero) changed when more data were added is not surprising; it is precisely what one would expect when adding more data. The conclusion, however, remained unchanged. Price Waterhouse's analysis of total sales without considering underlying economic conditions and other confounders should be ignored for the reasons they state on page 6 of their own report.

In an appendix, Price Waterhouse presents the results of an analysis of their extended data set for three more cities, Bellflower, Beverly Hills, and San Luis Obispo (p. A-17 - A-23). Their results are summarized below:


City                Effect of Ordinance
Change, L          P

Fraction of Sales

Bellflower          2.5 + 0.6          .000
Beverly Hills      -0.1 + 0.8          .905
Lodi               -0.7 + 0.4          .057
San Luis Obispo     1.5 + 0.4          .000

Comparison City

Bellflower          1.0  + 2.6         .697
Beverly Hills      -0.08 + 0.26        .776
Lodi               -4.6  + 7.5         .538
San Luis Obispo     3.0  +15.9         .267


According to Price Waterhouse's analysis, two of the cities (Bellflower and San Luis Obispo) appear to have had increases in restaurant sales when measured as a fraction of total retail sales and none of the cities had a change compared with the comparison cities. (Note that Price Waterhouse did not use the same comparison cities as in our paper [50].) Thus, Price Waterhouse's additional analysis can hardly be cited as evidence that smoke free restaurant ordinances reduce revenues; if anything, their evidence goes the other way.

These data are also consistent with the findings of no change in restaurant revenues associated with smoke free restaurant ordinances that have been reported by several investigators [50,54,58,59,60,Bartosch, 1995 #371,61].

The issue of effects on individual restaurants (p. 25) is discussed above (see Glassman-Oliver). The same comments made concerning unsupported statements by Grossman-Oliver apply to Price Waterhouse as well.

Price Waterhouse raises the same issues of outdoor dining (p. 25, 27) as was raised at the Hearing (Tr. 628-630). The assumption appears to be that everyone in California eats outdoors. This assumption certainly does not square with my personal experience, or the likely situation in Colorado in the winter. In any event, this argument remains one of speculation based on no data.

Price Waterhouse's arguments that experience with local ordinances would not apply to the effects of a national rule is silly (p. 26). Does Price Waterhouse really believe that people would eat at home rather than go to a smoke free restaurant, particularly given the fact that most people are nonsmokers? Indeed, a common argument that the tobacco industry makes against local ordinances has been that people will go to the next town to eat out (so they can breathe second hand tobacco smoke, I suppose). A national standard would eliminate this possibility. In any event, Price Waterhouse fails to present any data to support their assertions.

Price Waterhouse's argument that a failure to affect revenues would not necessarily translate into a change in profits (p. 26) because of differential profit margins is not supported by any data. Restauranteurs have told me that smoke free restaurants experience lower costs (lower cleaning costs, less burn damage to linens, lower fire insurance, fewer sick days for employees). We did not present this anecdotal information, because I sought to base my testimony on hard data.

Finally, I find it amusing that Price Waterhouse was willing to say that one of their models "detected a $1,779,000 (13 percent) decrease in quarterly sales during the presence of a smoking ban at the P = .112 level" (p. 12). Given the loud and consistent demands that results always achieve conventional statistical significance ( at the .05 level), one would expect that the other tobacco industry witnesses would soundly criticize Price Waterhouse for using such a high P value. Does this mean that the tobacco industry is willing to accept P< .112 as identifying a statistically significant increase in risk of disease caused by ETS? I doubt it. This result is yet another example of the inconsistent scientific standards the tobacco industry demands and accepts depending on whether the results fit its purpose at the moment.

National Economic Research Associates (NERA) (p. 417)

NERA submits the results of a survey of national attitudes that, in fact, supports the OSHA regulation. Table 2 (p. 13) shows majority support for smoke free dining in all classes of restaurants and public places, except bars, where the opinion is slightly in favor of separate sections (54% yes). They also show that 56% of respondents say they would be more likely to eat in a smoke free restaurant than one that permitted smoking (Table 6, p. 18). It also reports that 73% would not change the frequency with which they eat out (Table 7, p. 20).

NERA also supplied two affidavits by Albert Nichols (in Talbot Co. Maryland and Thurston Co, Washington) presenting his summary of the literature that creation of smoke free workplaces reduces cigarette consumption. As I noted in my post-hearing comment, this is an important economic benefit associated with OSHA's Proposed Rule that OSHA should consider in its final cost-benefit analysis. These submissions are another way in which the NERA submission supports the OSHA proposal.

National Smokers Alliance (No. 397)

The National Smokers Alliance is a front group established by Philip Morris [62], similar to other front groups such as the Beverly Hills Restaurant Association [51,63], California Business and Restaurant Alliance, and others. They have submitted a series of "reports" and press releases claiming that smoking restrictions have hurt restaurant business. Their "survey," conducted by the Heritage Foundation, was limited to smokers and ignored the fact that three-quarters of the population are nonsmokers. Rather than relying on data concerning what actually happened in cities that required smoke free restaurants, the National Smokers Alliance simply present theoretical models and speculation. They also present a survey that purports to demonstrate big drops in business in New York City since its smoke free restaurant ordinance went into effect.

These "results" directly contradict press reports that the restaurant and tourism businesses were booming in New York City in 1995 [64,65]. Indeed, even the results of a survey conducted by Cornell University [66] that was constructed in a way severely biased towards showing that the New York ordinance would reduce sales [67], showed that most people had not changed their dining habits following the ordinance and that there was no net change between nonsmokers eating out more and smokers eating out less. (The net reported change was -1.5% with a 95% confidence interval from -6.7% to +4.1%, no significant change.)

The material the National Smokers Alliance submitted is rhetoric, not science. OSHA should ignore it.

Vermont Business and Restaurant Coalition (No. 322)

The Vermont Business and Restaurant Coalition is another tobacco industry front group similar to the ones in California [51,53]. The submission is a collection of odds and ends, including yet another Price Waterhouse study predicting disaster if New York City enacted the then-proposed smoke free restaurant ordinance. (This report was done for the United Restaurant, Hotel, and Tavern Association of New York State, another Philip Morris front group [62].) For the reasons discussed above (see NERA), this material, like the other similar reports the tobacco industry routinely produces [68], should be ignored.

San Diego Tavern and Restaurant Association (submitted by Philip Morris) (No. 437)

This submission consists of a press release by an organization named Ferret Research attacking an earlier version of our study on smoke free restaurants, not the one placed in the record [50]. Ferret Research appears to be a couple of smokers rights activists with no record of scholarly research. There are no data here; this submission should be ignored.

Closing Comment

OSHA should consider the views Dr. S.J. Green, head of British American Tobacco Research and Development and a member of its board of directors, expressed concerning active smoking twenty years ago in 1976. In his thoughtful essay entitled "Cigarette Smoking and Causal Relationships," he wrote:

"The public position of the tobacco companies with respect to causal explanations of the association of cigarette smoking and diseases is dominated by legal considerations. In the ultimate companies wish to be able to dispute that a particular product was the cause of injury to a particular person. By repudiation of a causal role for cigarette smoking in general they hope to avoid liability in particular cases. This domination by legal consideration thus leads the industry into a public rejection in total of any causal relationship between smoking and disease and puts the industry in a peculiar position with respect to product safety discussions, safety evaluations, collaborative research, etc. Companies are actively seeking to make products acceptable as safer while denying strenuously the need to do so. To many the industry appears intransigent and irresponsible. The problem of causality has been inflated to enormous proportions. The industry has retreated behind impossible demands for 'scientific proof' whereas such proof has never been required as a basis for action in the legal and political fields. Indeed if the doctrine were widely adopted the results would be disastrous. I believe that with a better understanding of the nature of causality it is plain that while epidemiological evidence does indicate a cause for concern and action it cannot form a basis on which to claim damage for injury to a specific individual.

"In summary, for social policy purposes it is sensible and totally relevant to use the experimental evidence pertaining to large groups and also to select the simplest hypothesis. It may therefore be concluded that for certain groups of people smoking causes the incidence of certain diseases to be higher than it would otherwise be." [69, p. 1, 4] [emphasis added]

While these comments were made concerning the tobacco industry's refusal of the tobacco industry to accept the evidence that active smoking caused disease in smokers, the record of these proceedings before OSHA clearly indicate that the tobacco industry is applying the same strategies to dealing with the evidence that ETS is dangerous.

In reviewing the post-hearing comments by the tobacco industry -- indeed, in reviewing the entire record -- I was struck by the complete lack of substance. There is nothing new here. The tobacco industry continues to apply standards to support work that it likes or attack work it does not like. As shown above, they appear on all sides of an issue, while refusing to consider the evidence as a whole. It is also striking that the tobacco industry has only claimed that the proposed Rule will harm restaurants and bars. This claim is not supported by credible evidence. Even if it were true, this discussion leaves the entire rest of the workforce free from any assertions of potential economic harm, even from the tobacco industry.

It is time for OSHA to close the record and proceed with writing a final Rule to develop and implement a rule to protect American workers from ETS.


1. Taubes G. Epidemiology faces its limits. Science. 1995;269:164-169.

2. Trichopoulos D. The discipline of epidemiology (letter). Science. 1995;269:1326.

3. Gori G. The discipline of epidemiology (letter). Science. 1995;269:1327-1328.

4. Willett W, Greeland S, MacMahon, Trichopoulos D, Rothman K, Thomas D, Thun M, Weiss N. The discipline of epidemiology (letter). Science. 1995;269:1325-1326.

5. Gravelle J, Zimmermann D. Cigarette taxes to fund health care reform. Washington DC: Congressional Research Service, Library of Congress. March 8, 1994.

6. Redhead CS, Rowberg RE. Environmental Tobacco Smoke and Lung Cancer. Washington DC: Congressional Research Service, Library of Congress. November 14, 1995.

7. Price J. Secondhand smoke's dangers doubted: Report critical of EPA, OSHA. Washington Times Washington DC. November 27, 1995.

8. Kaiser J. Another look at secondhand smoke. Science. 1995;270:802.

9. US Environmental Protection Agency. Respiratory health effects of passive smoking: Lung cancer and other disorders. US Environmental Protection Agency. 1992.

10. Rennie D. Smoke and letters. JAMA. 1993;270:1742-1743.

11. Bero LA, Galbraith A, Rennie D. Sponsored symposia on environmental tobacco smoke. JAMA. 1994;271:612-617.

12. Bero L, Barnes DE, Hanauer P, Slade J, Glantz SA. Lawyer Control of the Tobacco Industry's External Research Program: The Brown and Williamson Documents. JAMA. 1995;274:241-247.

13. Bero LA, Glantz SA. Tobacco industry response to a risk assessment of environmental tobacco smoke. Tobacco Control. 1993;2:103-113.

14. Barnes DE, Hanauer P, Slade J, Bero LA, Glantz SA. Environmental Tobacco Smoke: The Brown and Williamson Documents. JAMA. 1995;274:248-253.

15. Regulatory Impact Analysis Project. Choices in Risk Assessment: The Role of Science Policy in the Environmental Risk Management Process. Washington, DC: Regulatory Impact Analysis Project, Inc. 1994.

16. Office of Environmental Health Hazard Assessment. Carcinogenic Effects of Exposure to Environmental Tobacco Smoke: ETS and Lung Cancer (Public Review Draft). Berkeley, CA: California Enviromental Protection Agency. January 1996.

17. Glantz S. The Tobacco Industry's Response to Scientific Evidence on Involuntary Smoking. . Fifth World Conference on Smoking and Health. Winnipeg, Canada: Canadian Council on Smoking or Health, 1983.

18. USDHHS. The Health Consequences of Involuntary Smoking. A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control. 1986.

19. National Research Council Committee on Passive Smoking. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. Washington, DC: National Academy Press. 1986.

20. Zhu B-Q, Sun Y-P, Sievers R, Isenberg R, Glantz SA, Parlmey WW. Passive smoking increases experimental atherosclerosis in cholesterol-fed rabbits. J Am Coll Cardiol. 1993;21:225-232.

21. Zhu B-Q, Sun Y-P, Sievers RE, Glantz SA, Parmley WW, Wolfe CL. Exposure to environmental tobacco smoke increases myocardial infarct size in rats. Circulation. 1994;889:1282-1290.

22. Sun Y-P, Zhu B-Q, Sievers RE, Glantz SA, Parmley WW. Metoprolol does not attenuate atherosclerosis in lipid-fed rabbits exposed to environmentl tobacco smoke. Circulation. 1994;89:2260-2265.

23. Zhu B-Q, Sun Y-P, Sievers R, Shuman J, Glantz S, Chaterjee K, Parmley W, Wolfe C. L-arginine decreases infarct size in rats exposed to environmental tobacco smoke (abstract). JACC. 1996:249A.

24. Zhu B-Q, Sun Y-P, Sievers R, Shuman J, Glantz S, Chaterjee K, Parmley W, Wolfe C. L-arginine decreases infarct size in rats exposed to environmental tobacco smoke. Am. Heart J. 1996 (in press).

25. Celermajer D, Adams M, Clarkson P, Robinson J, McCredie R, Donald A, Deanfield J. Passive smoking and impaired endothelium-dependent arterial dilation in healthy young adults. N. Engl. J. Med. 1996;334:150-154.

26. Hutchinson S, Sievers R, Zhu B-Q, Sun Y-P, Sudhir K, Deedwania P, Parmley W, Chaterjee K. Physiological concentrations of testosterone impair endothelium-dependent vasidilation in hypercholesterolemic rabbits exposed to tobacco smoke (abstract). Circulation. 1995;92:I-68.

27. Penn A, Snyder CA. Inhalation of sidestream cigarette smoke accelerates development of artiosclerotic plaques. Circulation. 1993;88:1820-1825.

28. Penn A, Chen LC, Snyder CA. Inhalation of steady-state sidestream smoke from one cigarette promotes atherosclerotic plaque development. Circulation. 1994:1363-1367.

29. Penn A, Snyder C. Sidestream cigarette smoke; Reply (letter). Circulation. 1994;89:2943- 2944.

30. USDHHS. The Health Consequences of Smoking: Nicotine Addiction. A Report of the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Health Promotion and Education, Office on Smoking and Health. 1988.

31. Food and Drug Administration. Regulations restricting the sale and distribution of cigarettes and smokeless tobacco products to protect children and adolescents. . Federal Register; August 11, 1995:41314-41451.

32. Food and Drug Administration. Proposed rule analysis regarding FDA's jurisdiction over nicotine-containing cigarettes and smokeless tobacco products. . Federal Register; August 11, 1995:41453-41787.

33. Slade J, Bero L, Hanauer P, Barnes DE, Glantz SA. Nicotine and Addiction: The Brown and Williamson Documents. JAMA. 1995;274:225-233.

34. USDHEW. Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service. United States Department of Health, Education and Welfare. 1964.

35. Glantz SA, Parmley WW. Passive smoking and heart disease: Epidemiology, physiology, and biochemistry. Circulation. 1991;83:1-12.

36. Glantz S, Parmley W. Passive smoking and heart disease: Mechanisms and risk. JAMA. 1995;273:1047-1053.

37. Taylor AE, Johnson DC, Kazemi H. Environmental tobacco smoke and cardiovascular disease: A position paper from the Council on Cardiopulmonary and Critical Care, American Heart Association. Circulation. 1992;86:1-4.

38. Adams M, Robinson J, McCredie R, Donald A, Clarkson P, Deanfield J, Celermajer D. Passive smoking is associated with impaired endothelium-dependent dilation in healthy young adults (abstract). Circulation. 1995;92:I-108.

39. Zhu B-Q, Sun Y-P, Sievers RE, Isenberg WM, Glantz SA, Parmley WW. Increased experimental atherosclerosis in cholesterol-fed rabbits exposed to passive smoke: Taking issue with study design and methods of analysis (reply to letter). J. Am. Coll. Cardiol. 1993;22:1752- 1753.

40. Thornton A, Lee P, Fry J. Differences between smokers, ex-smokers, passive smokers and non-smokers. J. Clin. Epi. 1994;47:1143-1162.

41. Howard G, Burke GL, Szklo M, Tell G, Eckfeldt J, Evans G, Heiss G. Active and passive smoking and are associated with increased carotid artery wall thickness: The Atherosclerosis Risk in Communities Study. Arch. Int. Med. 1994;154:1277-1282.

42. Cardador T, Hazan A, Glantz S. Tobacco industry smokers' rights publications: A content analysis. Am. J. Pub. Health. 1995;85:1212-1217.

43. Kabat G, Stellman S, Wynder E. Relation between exposure to environmental tobacco smoke and lung cancer in lifetime nonsmokers. Am. J. Epidemiol. 1995;142:141-148.

44. Fontham ET, Correa P, Reynolds P, Wu-Williams A, Buffler PA, Greenberg RS, Chen VW, Alterman T, Boyd P, Austin DF, IN] eaF. Environmental tobacco smoke and lung cancer in nonsmoking women: A multicenter case-control study. JAMA. 1994;271:1752-1759.

45. Allred EN, Bleecker ER, Chaitman BR, Dahms TE, Gottlieb SO, Hackney JD, Pagano M, Selvester RH, Walden SM, Warren J. Short-term effects of carbon monoxide exposure on the exercise performance of subjects with coronary artery disease. N Engl J Med. 1989;321:1426- 1432.

46. LeVois M, Layard M. Publication bias in the environmental tobacco smoke-coronary heart disease literature. Reg. Toxic. Pharmac. 1995;21:184-191.

47. Repace J. Dietary nicotine won't mislead on passive smoking ... unless subjects eat 90 kg tomatoes a day (letter). Br. Med. J. 1994;308:61.

48. Jarvis M. Dietary nicotine won't mislead on passive smoking ... unless subjects eat 90 kg tomatoes a day (letter). Br. Med. J. 1994;308:62.

49. Hopkins P, Williams R. Identification and Relative Weight of Cardiovascular Risk Factors. Cardiology Clinics. 1986;4:3-31.

50. Glantz S, Smith LRA. Effect of smoke-free restaurant ordiances on restaurant sales. Am. J. Pub. Health. 1994;84:1081-1085.

51. Samuels B, Glantz S. The politics of local tobacco control. JAMA. 1991;266:2110-2117.

52. Samuels B, Begay M, Hazan A, Glantz S. Philip Morris' failed experiment in Pittsburgh. J. Health Politics, Policy, and Law. 1992;17:329-351.

53. Traynor M, M. B, Glantz S. New tobacco industry strategy to prevent local tobacco control. JAMA. 1993;270.

54. Maroney N, Sherwood D, Stubblebin W. The Impact of Tobacco Control Ordinances on Restaurant Revenues in California. Claremont, CA: The Claremont Institute for Economic Policy Studies; The Claremont Graduate School. January, 1994, 1994.

55. Myers R. Classical and Modern Regression with Applications. : PWS Publishers. 1986.

56. Neter J, Wasserman W, Kutner M. Applied Linear Regression Models. : Irwin. 1983.

57. Glantz SA, Slinker BK. Primer of Applied Regression and Analysis of Variance. New York: McGraw-Hill, Inc. 1990.

58. Taylor Consulting Group. The San Luis Obispo Smoking Ordinance: A Study of the Economic Impacts of San Luis Obispo Restaurants and Bars. San Luis Obispo, CA: Taylor Consulting Group (under contract to the City of San Luis Obispo). January, 1993, 1993.

59. Hwang P, Toblas S, Kohout S, Harris M, Satterwhite D, Simpson D, Winn L, Foehner J, Pedro L. Assessment of the impact of a 100% smoke-free ordinance on restaurant sales -- West Lake Hill, Texas, 1992-1994. MMWR. 1995;44:370-372.

60. Bartosch W, Pope G. The Economic Impact of Brookline's Restaurant Smoking Ban. Waltham, MA: Health Economics Research, Inc. November 13, 1995.

61. California Department of Health Services. The Economic Impact of Smoking Restrictions on Restaurants in California. (Press Packet) Sacramento, CA: California Department of Health Services. December 11, 1995, 1995.

62. Smokefree Educational Services. Philip Morris Front Groups. New York, NY: 1995.

63. Fogel B. Testimony before the New York City Council. New York, NY: June 6, 1994.

64. Tharp P. Posh eateries rush to open more sites. New York Post New York, NY. June 13, 1995, p. 26.

65. Lowry T. Apple of world's eye: '95 a record year for tourism in city. Daily News New York, NY. October 23, 1995, 1995 .

66. Enz C, Corsun D, Young C. To Dine or Not to Dine: Restaurant Patrons' Response to the New York City Smoke-Free Air Act. Ithica, NY: Cornell University Center for Hospitality Research. 1995.

67. Glantz S. Letter to Cathy Enz. November 30, 1995, 1995.

68. Americans for Nonsmokers' Rights. Economic Impact Studies Commissioned by the Tobacco Industry: Synopsis and Analysis. Berkeley, CA: Americans for Nonsmokers' Rights. July 8, 1993.

69. Green S. Cigarette Smoking and Causal Relationships. 1976.

This document's URL is:

Return To: OSHA Hearings page
Go To: Tobacco BBS HomePage / Resources Page / Health Page / Documents Page / Culture Page / Activism Page