ENVIRONMENTAL TOBACCO SMOKE AND LUNG CANCER AN EVALUATION OF THE RISK written by J Benítez, Spain, JR Idle, UK and Norway (chairperson), HE Krokan, Norway, PHM Lohman, the Netherlands, M Roberfroid, Belgium, A Springall, UK. This report has been sponsored by Philip Morris Europe S.A., British-American Tobacco Company Limited, and Rothmans International Services Limited.

BY

¹Dept Environmental and Occupational Cancer, The Norwegian Radium Hospital, N-0310 Oslo, Norway, ²Dept of Environmental Medicine, National Institute of Public Health, P.O. Box 4404 Torshov, N-0403 Oslo, Norway, ³101 Fellcia Lane, Bowie, Maryland 20720 USA, ⁴International Agency for Research on Cancer, 150 Cours Albert Thomas, F-69372 Lyon, France.

This rebuttal is submitted to the Norwegian Ministry of Health and Social Affairs. 11 July 1996.

Summary of the criticisms: The report contains the standard tobacco industry misinterpretations of the epidemiology data, ignoring the relevance of the causal association between active smoking and lung cancer for passive smoking, minimizing the similarities between environmental tobacco smoke (ETS) and mainstream and sidestream smoke as well as downgrading the degree of ETS exposure. Several omissions of arguments for the role of passive smoking for lung cancer and of published data have been made. The report does not add new knowledge concerning ETS exposure and risk of lung cancer in humans, and represents another attempt by the tobacco industry to deny the health effects associated with smoking.

Background for the report:

U.S. EPA (1) has concluded that ETS is a Group A (known human) carcinogen. The evidence that ETS is a human carcinogen includes: the dose-related association between active smoking and lung cancer down to low doses with no evidence of a threshold; the qualitative similarities between ETS and mainstream smoke (MS); supporting evidence of the carcinogenicity of MS and ETS from animal bioassays and genotoxic studies; measurable ETS exposure and uptake of ETS constituents by nonsmokers; and the results of 30 epidemiological studies investigating ETS exposure and lung cancer in nonsmokers, including many showing dose-response.

The tobacco industry considers the public attention of the health effects caused by ETS one of the most important threats to the industry.

Main conclusion of the report supported by the tobacco industry is: It is the overall evaluation of the Working Group that there exists insufficient evidence to endorse the view that environmental tobacco smoke (ETS) is a primary lung carcinogen.

The main conclusion of this report is based on the following claims:

• There is insufficient evidence from the epidemiological studies performed to date that ETS is a human lung carcinogen
• There is insufficient evidence from animal studies that ETS is a lung carcinogen in animals
• There is insufficient evidence that the chemical constituents of ETS which are lung carcinogens when administered at high doses to animals reach concentrations sufficient to cause excess mortality from lung cancer in humans
• There is insufficient evidence that the relative small number of known ETS constituents, at the concentrations in which they are found in ETS, adds significantly to the toxic burden of the huge number of environmental chemicals to which human populations are exposed through the diet, fragrance chemicals, agrochemicals, drinking water and outdoor pollution.

The report consists of about 120 pages. Below are given principal comments to the different chapters and conclusions. Only statements or missing information that in our opinion are related to ETS exposure and risk of lung cancer, are considered. The chapters and page numbers given refer to those in the report.

Chapter 1. The nature of environmental tobacco smoke (ETS)

The report (Summary page v): "ETS cannot be equated with mainstream or sidestream tobacco smoke."

The US Environmental Protection Agency (EPA) (1) states: Comparison of the chemical composition of the mainstream smoke inhaled by smokers with that of the sidestream smoke in ETS inhaled by smokers strongly suggest that the toxic and carcinogenic effects would be qualitatively similar.

The report (page 4): Table 1 (and table 2) presenting formal yields of vapour phase (and particulate phase) components assuming 100% undiluted mainstream smoke or 100% diluted sidestream smoke (SS), together with ETS yields of exhaled mainstream smoke (10%) and sidestream smoke (90%).

The authors of the report fail to mention that the relative concentrations of the carcinogenic substances for which data are available, are for many compounds considerably higher in ETS than in the MS (nicotine is used as reference) (see table I) (results calculated for SS/MS show the same ranking). The data used for the calculations presented in Table 1 and 2 of the report are taken from EPA (1) and International Agency for Research on Cancer (IARC) (2).

Table I. Ratio between amount in calculated ETS and amount in MS (mid-point) of nicotine and carcinogenic substances (classified by IARC)

CONSTITUENT	ETS/MS
N-Nitrosodiethanolamine	1.1
N-Nitrosonornicotine	1.6
NNK	2.3
Nicotine	2.66
Benzo(a)pyrene	2.70
Benz(a)anthracene	2.71
1,3-Butadiene	4.0
Cadmium	6.5
Benzene	6.8
N-Nitrosopyrrolidine	16.2
2-Toluidine	17.1
Nickel	19.4
2-Naphtylamine	27.0
4-Aminobiphenyl	27.8
N-Nitrosodiethylamine	~40.0
N-Nitrosodimethylamine	54.0

Chapter 2. A model for the concentrations of smoke components

The report (Summary page v): "a model was developed and calculated concentrations of putative ETS components were derived. This model showed that the concentration of the likely chemical constituents of ETS is 10 to 1,000,000 times lower than permitted levels of exposure in the workplace"

The statement implies that exposure to ETS represent little or no health risk. The exposure levels discussed in relation to regulations in workplaces, refer to exposure for individual substances and not to complex mixtures such as ETS Further, the occupational exposure limits have been set for endpoints other than lung cancer. Repace and Lowrey (3) have on the basis of a pharmacokinetic risk model based on measurements of cotinine in serum and urine of nonsmokers, estimated that "for a substantial fraction of the 59 million nonsmoking workers in US, current workplace exposures to ETS ... appear to pose risks exceeding the de manifestis risk level above which carcinogens are strictly regulated by the federal government". Hammond et al (4) concluded that, based on actual measurements of ETS nicotine in office workplaces which allowed unrestricted smoking, "two thirds of 61 samples were above the de manifestis level," and with respect to OSHA's "significant risk" level (i.e., a 45-working lifetime risk of 1/1000) "more than half were well above the significant risk level".

The report (page 16): "Hardly any of the chemical components of ETS have been detected and quantitated in the human body"

Cotinine measurement in serum, urine, and saliva are commonly used as a reliable indicator of exposure to tobacco smoke. Cotinine is the major metabolite of nicotine and has a half-life of about 16 to 20 hours. Serum cotinine level reflects exposure to nicotine largely from the previous 1 to 2 days. In a recently published report with data from The Third National Health and Nutrition Examination Survey, 1988 to 1991, it was reported (5) that 37% of adult non-tobacco users lived in a home with at least 1 smoker or reported environmental tobacco smoke exposure at home. Serum cotinine levels indicated more widespread exposure to nicotine. Of non-tobacco users, 87.9% had detectable levels of serum cotinine. From Fig 2 in the Survey it is apparent that the most frequent cotinine level among those reporting tobacco use was 400 - 500 ng/ ml. The geometric mean cotinine level among non-smokers with home and work ETS exposure was 0.926 ng/ml. Moreover, among those with 1 smoker at home the level was 0.734 ng/ml and among those with >1 smokers at home the level was 1.24 ng/ml. These numbers correspond to about 0.2% of the most prevalent level among tobacco users. The exposure was in general higher at home than at workplaces. Thus, the geometric mean was 0.651 ng cotinine/ml for home ETS exposure only, compared to 0.318 ng/ml for work ETS exposure only. Jarvis (1987, cited by EPA (1)) found that the cotinine level on average was 0.7% in non-smokers with ETS exposure compared to smokers. In a study of nicotine and cotinine in hair (6) it was found that the cotinine concentration in the hair of pregnant women exposed to passive smoking, was about 10% of that found among pregnant women who smoked.

From the Table I in this rebuttal it is apparent that for most carcinogenic substances in tobacco smoke their relative concentration in ETS is higher than in MS. Several studies have been reported on the haemoglobin-4-aminobiphenyl adducts in blood as a measure of tobacco smoke exposure. It has thus been found (7) that non-smokers living with a smoker may have levels of haemoglobin-4-aminobiphenyl corresponding to 10-20% of that found in smokers. This value is in good agreement with that which could be expected from the data on cotinine, based on the relative level of 4-aminobiphenyl and nicotine shown in the table I.

The report (page 18): "Comparison of ETS levels (modelled or determined, where available data exist) with permitted occupational limits (US Standards, OSHA) demonstrates that, as a mixture of airborne chemicals, each considered separately, ETS does not reach concentrations of concern"

In Table 3 in the report the calculated concentration of respirable suspended particulates (RSP) is 596 mg/m³. The U.S. EPA's 24 hr standard for RSP for the general public (PM₁₀, i.e. particulate matter 10 microns or less in aerodynamic diameter) are 150 mg/m³ not to be exceeded more than once per year. EPA's annual PM₁₀ standard is 50 mg/m³. Thus, since the report supported by the tobacco industry has estimated typical levels of nearly 600 mg/m³, the annual standard would apply, and this is exceeded by a factor of 12. The Norwegian indoor air quality guideline for PM₁₀ is 90 mg/m³ as an 8 hr average.

Chapter 3. Lung cancer: Clinical considerations

This chapter contains no direct discussion of lung cancer in relation to ETS. Further the authors neglect to point out how the risk for lung cancer increases with number of cigarettes smoked. Trichopoulos et al (8) have performed an autopsy study of women who died of causes other than respiratory diseases. An increase in "epithelial, possibly precancerous lesions" was found in the lungs of nonsmoking women who were married to smokers. The authors concluded that their results "provide support to the body of evidence linking passive smoking to lung cancer ...". This study has not been included in the report.

Chapter 4. Epidemiological studies of environmental tobacco smoke and lung cancer

The report (Summary page vi): "The epidemiological studies investigating an association between ETS exposure and lung cancer are at, and many would say beyond, the limits of epidemiological science.

The outcome of 32 epidemiological studies (9) (studies evaluated by EPA and two later studies, Brownson et al, 10 and Stockwell et al, 11) on the relation between ETS exposure of spouses and risk of lung cancer are the following:

• All 4 cohort studies and 22 of 28 case-control studies showed a lung cancer risk >1.

• The increase in the lung cancer risk was statistically significant (p<0.05) in 11 of 19 studies where data were available for calculating statistical significance for the highest exposure groups.

• In 10 of 14 studies (studies evaluated by EPA) which provided sufficient data for a trend test, a statistically significant exposure-response relationship was shown.

The report has considered 17 new studies and 14 of these were included in the risk assessment. The new studies are listed in Table II.

Table II Case-control studies not considered in the risk assessment by EPA (1). Spousal results.

Study	Year	Country	Cases	Crude RR	Adjusted RR	Remarks
Brownson	1992	USA	431	0.97	1.00
Fontham	1994		651	1.26	1.29
Kabat	1995		67	1.10	1.08
Layard	1994		39	0.63	0.58	Not published. Submission to OSHA hearing.
Miller	1990		3			Discounted
Miller	1994		28	4.57	4.57
Stockwell	1992		210	-	1.60
Joeckel	1991	Germany	23	2.27	2.27	Not published. Sympos.. presentation
Knoth	1983		39			Discounted
Du/Lei	1993/4	China	75	1.09	1.09
Liu Q	1993		38	1.66	1.66
Shen X-B	1994		70		0.85	Not published. Sympos. presentation
Sun	1994		230		1.16	Not published. Sympos. presentation
Lan	1993		139	0.86	0.86
Wang F-L	1994		55	0.93	0.93
Wang T-J	1994		135	1.11	1.11	Not published Sympos. presentation
Ger	1993	Taiwan	48			Discounted

In 9 of 14 additional studies included, the relative risk for lung cancer after ETS exposure were >1. In 4 of the 6 studies from USA the risk was >1. The risk was in one study equal to 1. The authors (Brownson et al, 1992) (10) conclude, however "Ours and other recent studies suggest a small but consistent increased risk of lung cancer from passive smoking". The last study (Layard, 1994) represent a submission to an OSHA hearing by a tobacco industry consultant. In this study ETS protected against lung cancer (RR»0.6). The risk was >1 in 4 of the 7 studies from China and in the one study from Germany.

It can be concluded that the additional epidemiological data which has been included in the report, do not give any scientifically sound basis for a change in the conclusions in the EPA report (1). The report fails to recognize the strength of larger vs. smaller epidemiological studies (evident from funnel effect analysis) and it mixes the possibility of bias with the likely occurrence of bias.

In the report special attention is given to occupational exposure. Most studies of ETS in the work-place suffer from the ability to correctly characterize exposure, since people have a much more difficult time remembering whether coworkers smoked than whether spouses smoked. Moreover, as pointed out above, the ETS exposure at home is usually greater than at work. The report has missed giving reference to a paper (12) reviewing health effects including lung cancer caused by ETS exposure in restaurant workplaces.

The report also omitted an updated revision of the study by Fontham et al (13). The revision (14) (528 cases, 1148 controls) selected female respondents who reported ever working outside the home for 6 months or longer, controlling for numerous covariates and adult non-workplace exposures, vs those with no workplace exposure; an overall adjusted odds ratio of 1.6 for all lung cancer and for adenocarcinoma. This study also found that passive smoking in the workplace produced a statistically significant (p<0.001) increasing risk with years of reported ETS exposure, with 30 or more years of work exposure resulting in an odds ratio of 2.08 (95% CI 1.35-3.20) for lung cancer risk.

In the report it is pointed out that "if ETS causes lung cancer, it may be more readily detected from childhood rather than from adult exposure". Determining the effect of childhood exposure is difficult. Most nonsmoking women with lung cancers will probably be in the mid-sixties. The childhood exposure of these women will be very difficult to assess as fewer women smoked at that time and there have been large changes in smoking habits in the last 50 - 60 years.

The most useful epidemiological studies used in the EPA report for the relationship between ETS exposure and lung cancer, comes from the highest exposure groups, since the lower exposure groups suffer more from exposure misclassification bias. High exposure group analyses and exposure-response analyses are proper and important analyses for these types of studies. Such analyses are not reflected in the tobacco industry supported report.

Two general approaches have been used for estimation of lung cancer risk in connection with ETS exposure. One approach analyses the epidemiological evidence as has been done in the EPA report and in the tobacco industry supported report. The other estimates dose-response relationship for ETS exposure extrapolated from active smoking, based on "cigarette-equivalents" determined from surrogate measures of exposure to passive and active smoking (15).

It was pointed out in the discussion of Chapter 2 that nonsmokers exposed to ETS may have a cotinine level corresponding to 0.1 - 1% of a smoker, while the exposure to a number of carcinogens are on a relative basis considerably higher than the exposure for a smoker from the mainstream smoke. Thus, the exposure of the carcinogen 4-aminobiphenyl is 10 - 20% of that of a smoker. If the risk for lung cancer for a smoker is 10 times that of a nonsmoker not exposed to ETS, a 30% increase in the risk of lung cancer will represent a 3% increase in a scale where the risk for a smoker is set equal to 100%. Thus, from consideration of "cigarette equivalents smoked by passive smoking" an increased risk of lung cancer would be substantial.

Chapter 5. Diet as a confounding factor

The report (Summary page vi): "Exposure to ETS is unlikely to be a significant risk factor for lung cancer in humans (particularly women married to smokers) because, compared to such low-level of exposure to chemicals, other common lifestyle risk factors are likely to play more significant roles."

EPA (1) concludes: "In summary, an examination of six non-ETS factors that may affect lung cancer risks find none that explain the association between lung cancer and ETS exposure as observed by independent investigators across several countries that vary in social and cultural behaviour, diet and other characteristics." (The factors considered were: history of lung disease, family history of lung disease, heat sources for cooking or heating (indoor air pollution from other sources, e.g., smoky coal, in part of China may mask any ETS effects in those studies), cooking with oil, occupation, and dietary factors).

Chapter 6. DNA adduct formation, DNA repair and tumour induction - dose response considerations

See discussion on Chapter 8. The importance and impact of genotoxic effects including DNA adducts, of ETS have not been discussed in the report.

Chapter 7. Discussion of specific components

This chapter does not contain any direct discussion of lung cancer in relation to ETS.

Chapter 8. Special consideration for low-dose exposures

The report (Summary page vii): "Thus, processes like DNA repair are likely to play a key role in the human interaction with the internal environment and, indeed, may introduce a threshold below which a chemical exposure poses no risk to humans"

Carcinogenic substances may show differences in dose-response, depending on the underlying mechanisms for cancer development. The available scientific evidence makes it unlikely that genotoxic substances have a no-effect threshold. In contrast to genotoxic carcinogens, some non-genotoxic carcinogens which induce cell proliferation indirectly as a response to toxic tissue damage may show a threshold. ETS is genotoxic and the data on active smokers and lung cancer indicate a sizable risk contribution from even low level exposure to the carcinogens present in ETS.

Chapter 9. Conclusions and risk assessment

The report (page 91): "HAVING CONSIDERED THE WEIGHT OF EVIDENCE, IT IS THE JUDGEMENT OF THE WORKING GROUP THAT ENVIRONMENTAL TOBACCO SMOKE IS NOT A PRIMARY LUNG CARCINOGEN"

IARC (2) concluded already in 1986 that "passive smoking gives rise to some risk of cancer". The EPA (1) report concludes in 1993: "In adults, ETS is a human lung carcinogen responsible for approximately 3,000 lung cancer death annually in U.S. nonsmokers."

The report supported by the tobacco industry does not provide new information which would invalidate the EPA conclusion.

Chapter 10. Postscript

The report (page 92): "Using the EPA's new proposed [Proposed Guidelines for Carcinogen Risk Assessment] "descriptors for classifying human carcinogenic potential", the Working Group would categorise the epidemiological data on ETS and lung cancer as "cannot be determined", but the weight of evidence as demonstrating that ETS is "not likely" to increase the risk of lung cancer".

After studying the EPA's proposed guidelines, it is anticipated based on the experience of the rebuttal authors, that a reevaluation of ETS would also result in placement in the highest category, "known or likely human carcinogen".

The report (page 92): "For example, no longer will the risk assessment be bound by the single default approach of using "the linearized multistage model for extrapolating risk from upper-bound confidence intervals". Inherent in the new approach is the determination of a threshold".

The Postscript falsely presupposes that EPA is advocating that a threshold generally exists for carcinogens. In fact, the proposed guidelines provide a nonlinear default for high-to-low dose extrapolation only for specified cases where there is strong mechanistic data supporting the existence of a threshold. This nonlinear default is irrelevant to the assessment of ETS because 1) ETS is a genotoxic mixture, thus a nonlinear default would not be considered appropriate, and 2) data are available for human lung cancer risk from environmental levels of ETS.

REFERENCES

1. U.S. Environmental Protection Agency. Respiratory health effects of passive smoking: Lung cancer and other disorders. The report of the U.S. Environmental Protection Agency. Smoking and Tobacco Control. Monograph 4. NIH Publication No. 93-3605, Washington DC, 1993.

2. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Tobacco Smoking. Vol 38, 1986.

3. Repace JL, Lowrey AH. An enforceable indoor air quality standard for environmental tobacco smoke in the workplace. Risk Analysis 13: 463-475, 1993.

4. Hammond SK, Sorensen G, Youngstrom R, Ockene JK. Occupational exposure to environmental tobacco smoke. JAMA 274: 956-960, 1995.

5. Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Mauer KR. Exposure of the US population to environmental tobacco smoke. The third national health and nutrition examination survey, 1988 to 1991. JAMA 275: 1233-1240, 1996.

6. Eliopoulos C, Klein J, Knie B, Greenald M, Chitayat D, Koren G. Hair concentration of nicotine and cotinine in women and their newborn infants. JAMA 271: 621, 1994.

7. Hammond SK, Coghlin J, Gann PH, Paul M, Taghizaeh K, Skipper PL, Tannenbaum SR. Relationship between environmental tobacco smoke exposure and carcinogen-haemoglobin adduct levels in nonsmokers. J Natl Cancer Inst 85: 474-478, 1983.

8. Trichopoulos D, Mollo F, Tomatis L, Agapitos E, Delsedime L, Zavitsanos X, Kalandidi A, Katsouyanni K, Riboli E, Saracci R. Active and passive smoking and pathological indicators of lung cancer risk in an autopsy study JAMA 268: 1697-1701, 1992.

9. Jinot J, Bayard S. Respiratory health effects of passive smoking: EPA's weight-of-evidence analysis. J Clin Epidemiol 47: 339-349, 1994.

10. Brownson RC, Alavanja MC, Hock ET, Loy TS. Passive smoking and lung cancer in nonsmoking women. Am J Publ Health 82: 1525-1530, 1992.

11. Stockwell HG, Goldman AL, Lyman GH, Noss CI, Armstrong AW, Pinkham PA, Candelora EC, Brusa MR. Environmental tobacco smoke and lung cancer risk in nonsmoking women. J Natl Cancer Inst 84: 1417-1422, 1992.

12. Siegel M. Involuntary smoking in the restaurant workplace. A review of employee exposure and health effects. JAMA 270: 490-493, 1993.

13. Fontham ETH, Correa P, Reynolds P, Wu-Williams A, Buffler PA, Greenberg RS, Chen VW, Alterman T, Boyd P, Austin DF, Liff J. Environmental tobacco smoke and lung cancer in nonsmoking women. A multicenter study. JAMA 271: 1752-1759, 1994.

14. Reynolds P, Fontham ETH, Wu A, Buffler PA, Greenberg RS. Occupational exposure to environmental tobacco smoke. JAMA 275: 441-442, 1996.

15. Repace JL, Lowrey AH. Risk assessment methodologies in passive smoking. Risk Analysis 10: 27-37, 1990.