Tobacco Documents Online

TABLE'4 Reported relative risks of lung cancer in relation to ETS exposure at work. Study (ref) Sex Index of exposure Relative nsk (95°k conf.limits) Garfinkeli (42) Female Smoke exposure at work in last 5 years 0:88C0,66-1.18d Female Smoke exposure at work in last 25 years 0.93(0:73-1.18) Kabat 1 (71) Female Current exposure on regular basis to tobacco 0.68(0:32-1.47) Male smoke at work Current exposure on regular basis to tobacco 3.27(1.01-10.6) Kabat2 (72) Female smoke at work Exposed to ETS at work (ever) 1.00(0.49-2.06) Male Exposed to ETS at work (ever) 0.98(0:46-2.10) Lee (24) Fem ale Passive smoke exposure at work 0.63(0.17-2.33) Male Passive smoke exposure at work 1.61(0.39-6.60) Shimizu (73) Female Someone at working place smokes 1.20(0.44-1.37) Varela (411) Both 150 person/years smoking in the workplace 0:91(0.80,1.04) Wu (20) Female Passive smoke exposure at work 1.3 (0.5-3.3). a'IH3;R LSSLJFS Extension of risk assessments to workplace ETS and heart disease deaths. While the use of epid'emiological data to estimate the number of deaths from lung cancer among never smokers is dubious, extension of these estimates to other diseases and to workplace exposure is even more so. This highlights the invalidity of the estimates by Kawachi et a! (6) where of a total of 273 deaths per year due to ETS among never smokers, only ~Lare from lung cancer due to at home ETS exposure, while as many as .j 5?. are from ischaemic heart disease due to at work ETS exposure., The fragility of the confidence limits, 112 to 442, for the overall total of 273 is obvious. In no sense can we be confident that the true answer lies in this range. The estimate is cast in an even poorer light when one realises that the factor of 4 used to calculate lung cancer relative risks at work from those at home is also used for heart disease. What is the justification for that? The basis for the factor is relative particuliate matter exposure, widely thought irrelevant to heart disease aetiology. It is notable that their resultant heart disease relative risk estimates for at workk exposure are, implausibly, larger than those generally reported in relation to actiive smoking. Extension of risk assessments to ex•smokers. Wells (5)and Repace and Lowrey (7) estimate numbers of deaths due to ETS among never smokers and ex-smokers combined. They assume risk estimates based on results for never smokers are applicable al'so to~ ex-smokers. Neither paper discusses the problems implicit in this approach. In the first place there is no direct epidemiological evidence on risk in relation to ETS exposure for ex-smokers with the limited exception of the study by Varela (41) which found no evidence of an effect of ETS in either never smokers or long term ex-smokers. Nor is there any evidence on levels of ETS exposure in ex- smokers as distinct from never smokers. Without direct evidence the assumption that nskk increases in relation to level of ETS exposure in ex-smokers to t'he same extent' that it does in never smokers seems remarkably simplistic. Might not effects of ex-smoking interact with those of ETS (if any)? Might not the situation depend' on how long ago the smoker has given up , or why? There seems no scientific justification whatsoever for extrapolating estimates to ex- smokers. Extrapolation from one country to another. Kawachi er al (6)' do not discuss the validity of calculating estimates for New Zealand when all their relevant source data comes from other countries.. Their answer depends heavily on the US based factor of 4 used for relative exposure at work to at home. As noted above a UK study (68, found' a factor less than 11. Which is relevant for New Zealand? Variations in relative risk of lung cancer by age. As discussed by Wells ('5) and' in the NRC report by Robins (4'); if the relationship between ?nl

ETS andldng cancer risk depended on age, it would be appropriate to take this into account in the ri'sk assessment. In fact only the study of Hirayama (18; 67) presents data by age, other investigators implicitly assuming that the relative risk is invariant of age. Using a relative risk estimate of 1.44 as applying to all age groups, Wells calculated there would be 992 deaths per year due to ETS exposure. Wells noted that Hirayama's data actually indicated "a declining relative risk with~ age from 1.87 at approximately age 50 to 1.43 at approximately age 75" and used these data to "develop a second d'eath, calculation assuming a declining relative risk but still normalized'to 1.44" arriving at a slightly lower estimate of 911 deaths per year. Wells" calculations mislead in a number of ways. First, he used as source material data an risk by age of the husband (67) when more appropriate data by age of the wife were available (18). Second; he used data for ages 60-69 and 70-79 combined' as applicable at "approximately age 75", concealing the fact that the relative risk estimate at age 70-79 is actually 0.70: If one uses data in Wells' Table 6 for never smoker death rates, nonsmoker populations and4ractions exposed by age, and one uses Hirayama's actual relative risks by age of the wife (18), then it can be shown (Table 5) that allowing for variation in risk by age very substantially affects estimates. Thus, for the 40- 79 age group; one arrives at an estimate of 858 deaths due to ETS if one assumes age invariance, but one actually arrives at an estimate of 964 deaths saved by ETS if one uses Hirayama s data directly. The relative risk estimate for the M79 year age group is certainly unreliable, being based on only 6 deaths in the Hirayama study (as against 46, 91 and 57 for ages 40-49, 50-59', 60,69), so in Table 5 estimates of deaths are also shown using a combined relative risk for the age groups 60-69 and 70-79. This gives an estimate of 299 deaths due to ETS, substantially less that that assuming risk is invariant of age. While there are many problems in applying the Hirayarna estimates, including the fact that Wells' Table 6 is based on age at death whereas Hirayama s d'ata are based on age at start of the study„ Wells' paper conceals the major problems which have been given detailed attentiom by a number of authors (75;, 76). Reliable data broken down by age are clearly needed. How many lung cancer deaths are there in total among never smokers? In 1985 in the USA, there were a total of 83,854 deaths from lung cancer among males and 38;702 among females (77).In his Tables 6 an& Al, Wells (5Y gives estimates of death~ rates among never smokers which, if applied to the age- specific population estimates of never smokers, yield 1,907 deaths among males and 4,232 deaths among females,, respectively 2.3% and 10.9% of the total deaths from lung cancer. TABLE 5 Numbers of lung cancer deaths per year among US nonsmokers occurring in the population aged 40,79 based on Hirayama s (18) estimates of relative risk by age of wife Risk assumed invariant of age Risk assumed to varyy with age Age Relative risk Deaths Relative risk' Deaths' 40-44 1.45 32 2.76 69 45-49 1.45 40 2.76 85 50-54 1.45 58 1.72 79 55-59 1.45 89 1.72 122 60-64 1.45 119 11.12( 0.97 ). 39(1-11) 65-69 1.45 165 11.12(0.97) 54(i-15 ) 70-74 75-79 1.45 1.45 170 185 0.190.97) 0.19(0.97)) -740(-1:.5)-672(-15). Total 858 -964(299) ` Bracketed items assume common estimates for 60-69 and 70-79 age group. Elsewhere C78!1, I have reviewed Uhe respectively, reasonably close to the proportion of lung cancers occurring among Wells. 0 never smokers in a range of recent Other authors have suggested!there are more or-h epidemiological studies of Western populations. deaths than this. Thus in the 1986 NRC report: (4' i This gave an average of 2.4% for males and 13.2% Robins quoted estimates of roughly 5,200 deathss for females, equivalent to 2,012' and 5,109 deaths for males and 7,000 for femal!es among, U.S. estimates of ~J 204

never smokers in 1985, while Repaca and Lowrey (7) cite Kuller et al (36) for an estimate of 6000 to 8000 lung cancer cases each year in US never smoking women. Three points arise. First, there is consid- erable uncertainty about: the number of lung cancer deaths among never smokers. Second, if the lower estimates, which total about 6,000-7,000 deaths in the two sexes combined; are used, then many of the epidemi- ologically based estimates shown in Table 2 are totally unreasonable. Even if (implausibly) everyone were assumed to be exposed to ETS with risk doubled as a result the estimated number of lung cancer deaths occurring among never smokers would only be 3,000-3,500, and yet the four highest estimates in Table 2 all exceed this. Third, none of the estimates of total lung cancer deaths among never smokers cited above make any adjustment for miselassification of smoking status„ all' taking self-reported smoking habits at face value. Starting with the first estimate cited above of 6,139 deaths for the sexes combined, one can readily calculiate that, if 1% of ever smokers were assumed to deny smoking on interview, this figure would fall by over a thousand to 4,972. This underlines the unreasonableness of the higher estimates in Table 2. DLS(1JSSTON In the USA in 1985 there were some 120,000 deaths from lung cancer. Although estimates of the total number occurnng among never smokers of up to around 12,000 have been cited, more reasonable estimates seems to be about 5,000: to 6,000. I!n attempting to estimate how many of these occur as a result of ETS exposure, one has to decide whether to base one's estimate on the epidemiological evidence on ETS and lung cancer or on the dosimetric evidence on exposure to relevant smoke constituents of ETS exposed nonsmokers and smokers. It is abundantly clear that the two methods of estimation give very different answers. Thus, while estimates based on retained particulate matter give tens of deaths and those based on nicotine or respirable suspended particulates give hundreds, the epidemiologicall'y based estimates all give thousands of deaths. Which answer, if any, one accepts depends to a large extent on the faith one places on the. different types of evid'ence. Wells (5), Kawaehi~ et al (6) and Repace and Lowrey (7) accept the epidemiology essentially at face value and pay little or no attention to its poor quality and very obvious weaknesses. They either ignore the dosimetric evidence (6)„ do not make iit clear that it gives different answers and/or dismiss it as inconsistent with the epidemiology (17); or invoke mechanisms to explain the discrepancy which are scientifically unappealing (5). It seems to this author that the epidemiolbgical evidence is untrustworthy and that, between the two, the d'osimetnc evidence is preferable_ Of course problems remain both in choosing the appropriate index of exposure to use and in selecting the appropriate dose response curve at low doses (with the possibility of a threshold), but it seems clear that this approach is better than one which leads to such implausibly high figures. When one restricts attentiom to lung' cancer, to never smokers and to ETS exposure from the spouse, one is at least operating in an area where the epidemiological evidence indicates an association. When one extends risk assessment to other diseases, to ex-smokers and to ETS exposure in the workplace one is stretching the limits of what is science. There essentially is no evidence on possible effects of ETS in ex- smokers and little reason to expect that any effects, if they exist, will be the same as in never smokers. There is some evidence on ETS exposure in the workplace, but this shows no association at all with Iking cancer risk. The epidemiological evidence on ETS in relation to deaths from causes other than lung cancer is unconvincing, and no scientific authority has claimed' cause and! effect. RFT 'FPM1GES 1. International Agency for Research on Cancer. 1ARC Monobr., EraL Coresnog: Risk Che,n: Hum., 3:8, Tobacco Smoking, I'ARC, Lyon (1986). 2. National Health and Medical Research Council. Report of tho %'orking Partw on the Et1vcts of Passive Smoking on Health., Australia (11986). 3. US Department of Health and Human Services. The Health Consequences of Inuolontan• Smoking: a Report of the Surgeon General; Rockville (1986). 4. National Research Council. Environmental Tobacco Smoke. .'4fensuring Exposure and Assessing Health E(fects,, National Academy Press, Washington D!C. (1986): 5. Wells, A.J. Environ. Int., 14, 249-2fi5 (1988):, I 205

6. Kawachi, I., Pearce, N. and Jackson, R. N:Z. Med J., 102, 337-340 (1989): 7. Repace, J.L. and Lowrey, A.H. Risk Anal., 10, 27-37 (1990): 8. Lee, P.N. A Detailed Review of Epidemiological' Evidence relating Environmental Tobacco Smoke (ETS) to the Risk o)`Cancer, Heart Disease andOt/ur Causes ojDeath cn Adults who have Never Smoked. Ih the press. 9. Inoue, R. and Hirayama, T: in Smoking and'Health 1987, Eds. Aoki, M. et a1., Elsevier,. Amsterd'am, 283-285 (1988). 10. Geng, G.Y., Liang, Z.H., Zhang, A.Y. and Wu, G.L. in Smoking and Health 1987, Eds. Aoki, M. et al., Elsevier, Amsterdam, 483-486 (1988). 11. Trichopoulbs, D., Kalandidi', A and Sparros, L, The Lancet, 2, 677-678' ('1983)'. 12. Akiba, S., Kato, H. and Blot, WJ. Cancer Res., 46, 4804-4807 (1986): 13. Brownaon, R.C., Reif, J.S.,, Keefe, T.J., Ferguson, S.W. and Pritzl, J.A. Am. J. EpidcmcoG., 125, 25-34 (1987). 14. Koo, L.C., Ho;,J.H-C., Saw, D. and'Ho, C-Y.Int. J: Cancer,99, 162-169 (1987). 15. Lam, T.H. and Cheng, K.K. in Smoking and' Health 1987, Eds. Aoki, M. et all, Elsevier, 279- 281 (1988): 116. Hole,, D.J., Gillis, C.R., Chopra„ C. and Hawthorne, V.M. Br. Med. J.,,299, 423-427 (1989). 117. Lam, T.H., Kung, I.T.M., Wong, C.M., Lam; W.K.,, K]eevens; J.W.L, Saw; D., Hsu, C., Seneviratne, S., Lam, S.Y., Lo, K.K., and Chan, W.C. Br. J. Cancer, 66, 673-678 (1987): 18. Hirayama, T. in Lung Cancer: Causes and Prevention„ Eds. Mizell, M. and Correa, P., Verlag Chemie International Inc., Deerfield Beach, 175-195 (1984). 19. Gao, Y-T., Blot, WJ., Zheng; W., Ershow; A.G., Hsu, C.W., Levin, L.I., Zhang, R. and Fraumeni, Ji.F. Int: J. Cancer, 40, 604-609 (1987). 20. Wu, A.H., Henderson, B.E., Pike, M.C. and Yu, M.C. J. Nat. Cancer Inst., 74, 747-751 (1985). 21. Correa, P., Pickle, L.W., Fontham, E.,, Lin, Y. and Haenszel, W. The Lancet, 2, 595-59 (1983). 22. Humble, C.G., Samet, JLM. and Pathak, D.R. Am. J. Public Health, 77, 598-602 (1987). 23. Svensson, C., Pershagen, G. and' Klominek, J. Acta Oncol., 28, 623-629 (1989) 24. Lee,, P.N., Chamberlain, J. and Alderson; M.R. Br. J: Cancer, 64, 97-105 (1986) 25. Buffler, P.A., Pickle, L.W., Mason, T.J. and Contant, C. in Lung Cancer Causes and Prevention, Eds. Mizell, M. and Correa, P: Verlag Chemie International' Inc., fsno-.sFrc)i.rr,, 83-89 (1984). 26. -Chan; W.C. and Fung, S.C. in Cancer Epidemiology, Gustav Fischer Verlag, Vol 6, 199-202 (1982). 27. Lee, P.N. Human To:icol:, 6, 517-524 (1987). 28. Wald, N.J., Nanchahal, K. and Cuckle, H. Br. J. Cancer, 61, 337-344 (11990), 29. Arundel„ A., Irwin, T. and Sterling, T. J. Environ. Sci. Health, Part C, 4, 93-118 (1986), 30. McAughey, J.J., Pritchard, J.N. and Black, A. in Present and Future o(Indoor Air Quality, Eds. Bieva, C.J. et al.,, Excerpts Medica International Congress Series 860, Brussels, 161-168 (1989): 31. Adlkofer, F.X., Scherer, G., von Meyerinck, L., von Maltzan, Ch. and Jarczyk, L, in Present and Future of Indoor Air Quality, Eds. Bieva, CJ. et aL' Excerpta Medica International! Congress Series 860, Brussels, 189-196 (1989). 32. Fong„P. J. Biof. Phys., 10, 65-73 (1982). ~ 33. Wald, N.J., Nanchahal, K., Thompson„ S.G: and' Cuckle, H.S. Br. Med. J:, 293, 1217-1222 ~ (1986). N ~ 34. Repace, J.L. and Lowrey; A.H. Environ. lnt., 11, 3-22 (1985). 35. Wigle, D., Collishaw, N., Kirkbride, J. and Mao, Y. Can. Med Assoc. J:, 136, 945-951 (1987):~ 36. Kuller„ L.H'., Garfinkel, L., Correa., P., Preston-Martin. S., Haley. N.J., Sandler„ D:P. and ~ Hoffmann, D. Environ. Health Perspect_„70, 57-69 (1986). ~ 37. Russell, M.A.H., Jarvis, M.J. and' West, R.J. Br. J. Addict., 81„ 275-281 (11986). ~ 38. Lee, P.N. in Assessment oj' Inhalation F'lazards, Eds. Mohr, U_ et al.,, ILSI Monographs, ~ ~ Springer-Verlag, Berlin, 49-59 (1989). 39. Facci,ni, J. Exp.,Pathol., 37, 177-180 (1989):, 40. Begg; C.B. and Berlin, J.A. J: R. Stat. Soc., Ser. A, 151, 419-463 (1988). 206

41. Varela, L.R. 'Aseessment of the Association between Passive Smoking and Lung Cancer. ', Yale University Doctor of Philosophy Dissertation (i1987)'. 42. Garfinkel, L., Auerbach, 0. and Joubert, L. J. Nat. Cancer Inst.,,75„ 463-469 (4985): 43. Sobue, T:, Suzuki, R., Nakayama, N.,, Inubuse, C., Matsuda, M., Doi, O:, Mon, T., Furuse. K., Fukuoka, M., Yasumitsu, T., Kuwabara, 0., Ichigaya, M., Kurata, M., Kuwabara, M.,, N'akahara„K., End'o, S. and Hattori, S. Gan No Rinsho, 36, 329-333 (1990). 44. Koo, L.C., Ho, J.H-C. and Rylander, R. Soc. Sci. Med., 26, 751-760 (11988). 4'5. Lee, P.N. Misclbssification of Smoking Habits and' Passiue Smoking. A Review of the Evidence. Springer-Verlag, Heidelberg, (1988). 46. Saito, R. in Smoking and'Health 1987„Eds. Aoki, M. et al., Elsevier, 517-519 (1988): 47. Gillis, C.R:, Hole, D.J., Hawthorne, V, and Boyle, P: Eur. J. Respir. Dis.(Suppl.133), 65, 121- 126 (1984). 48. Garland, C.,, Barrett-Connor„ E.,Suarez, L.,, Criqui, M.H. and Wingard, D.L. Am. J. Epidemiol., 121, 645-650 (1985). 49. Helsing, K.J'., Sandier, D.P., Comstock, G.W. and Chee, E: Am. J. Epidemiol:, 127, 915-922 (1988). 50. Svendsen, K.H., Kuller, L.H., Martint M.J. and Ockene, J.K Am. J. Epidemiol., 126, 783'-795 (1987). ~ 51. Martin, M.J., Hunt, S.C: and Williams, R.R. Annual Meeting of American Public Health Association (1986)~ 52. Hrrayama, T. Br. Med. J., 282, 183-185 (1981). 53. Layard, h1.W. and Viren, J.R. in Present and Future of Indoor Air Quality, Eds. Bieva, CJi. et al.„ Excerpta Mediea International' Congress Series 860, Btussels„ 177-180 (1989). 54. Lee, P.N~. Br. Med. J., 283, 1465-1466 (1981). 55. Lee, P.N. (1989) N.Z. Med. J., 102, 539 (1989). 56. Hiravama„ T. N:Z Med. J., 103, 54 (1990):. 5-1. Garfinkel, L. J. Nat: Cancer Inst., 66, 1061-1066 (11981). 58. Burchi, J.D., Rohan, T.E., Howe, G.R., Risch, H.A., Hil'1„ G.B:, Steele, R. and Miller, A.B. Int. J: Cancer, 44, 622-628 (,1989): 59. Kabat, G.C., Dieck, G.S. and' Wynder, E.L. Cancer, 57, 362-367 (1986): 60. Hellberg, D., Valentin,, J. and Nilsson; S: The Lancet, 2, 1497 (1983). 611. Miller, G:H. West J. Med, 140, 632-635 (1984)_ 62. Sandler, D.P., Comstock, G.W., Helsing, K.J. and Shore, D.L. Am. J. Public Health, 79, 163- 167, (1989). 63. Reynold's, P., Kaplan, G..1_ and Cohen, R.D. Society for Epidemiological Research, Amherst„ Massachusetts (1987). 64. Slattery, M.L., Robison, L.M., Schuman, K.L., French, T.K.,, Abbott, T.M.,, Overald, J.C. and Gardner, J.W. JAMA, 261. 1593-1598 ('1989): 65. Layd'e, P.M. JAMA, 261, 1631-1633 (1989). 66. Sandler, D.P., Everson, R.B. and' Wilcox, A.J. Am. J. Epid.emiol., 121, 37-48 (11985).. 67. Hirayama, T. Prev., Med., 13, 680-690 (1984). 68. Caporaso, N.E., Tucker, M.A., Hoover, R.N.,, Hayes, R:B., Pickle, L.W., Issaq,, H.J., Muschik,, G.M.,, Green-Gallo, L., Buivys, D., Aisner, S., Resau, J.H.. Trump, B.F., Tollerud, D., Weston, A. and Harris, C.C. J Nat. Cancer Inst:, 82, 1264-1272 (11990): 69. Sellers, T.A., Bailey-Wilson, J.E., Elston, R.C., Wilson, A.F:, Elston, G.Z., Ooi, W.L. and' Rothschild, H. J. Nnt, Cancer lnst:, 82, 1272-1279 (1990). 70. Fnedman„G.D., Pettiti, D.B. and Bawol, R.D: Ain: J. Public Henlth,73, 401-405 (1983) 711. Kabat, G.C. and' Vr'vnder, E.L. Cancer, 63, 1214-1221 (Q9$4). 72. Kabat, G.C. 19901 Winter Toxicology Forum, W.ashington D.C. (1990). 73. Shimizu, H., ilonshita, M., 1lizumo, K., blasuda, T., Ogura, Y.,,Santo. M.. Nishimura. w1., hunishimo. K., liarasawa, Ii., Nishiwaki, K., Yamamoto, M., Hisanichi, S. and Tomonaga, S. Tohoku J: Exp. Med., 154, 389-397 (1988). 74. Kirk, P.W.W.. Hunter, M., Baek, S.O., Lester, J.N. and Perry;, R. in Ir:door a+id' Ambcent Air QF,alit.v, Eds. Perry, R. and Kirk, P.W. Selper. Lond'on,, 99-112 (1988),. 75. Ahlborn, W. and L''berla, K. in Indoor and :lmhzent Air Quality. Eds. Perry', R.and I+.irk, P.W. Selper, Lond'on, 169=1 78(11988): 30'

76. Kilpatrick, SJ. and Viren, J. in Indoor and Ambient Air Quality,, Eds. Perry, R. and F:irk. P.W., Selper, London, 195-202 (1988). 77. W'orld' Health Organisation. 1988 World Health Statistics Annual, Geneva (1988): 78. Lee, P.N. Udnpublished resullts (1990). 208