Tobacco Documents Online

-Oct-97 04:03P 0181 642 2135 "I5 vRvjlonmental tobi!.C,_C,o ~moke • n evaluation of th(; evidence" Comments on the paper bv M R I.aw. J K Morris and N J Wald in the British Medical Journal (1997. 315.973-980) The claims Law, Morris and Wald claim that "breathing other people's smoke is an important and avoidable cause of ischaemic heart disease [IHD], increasing a person's risk by a quarter." This conclusion is based on a sequence of observations and analyses. First, based on a meta-analysis of 19 epidemiological studies they estimated that. among never smokers, exposure to environmental tobacco smoke [ETS]. as indexed by spousal smoking, is associated with a relative risk of IHD of 1.30 (95% confidence interval [CI] 1.22 to 1.38). Second, based on extrapolation of results for smokers from five US and UK studies of smoking and heart disease, the.,,, estimated that smoking one cigarette a day is associated with a risk of IHD. relative to that in nonsmokers, of 1.39 (!. 18 to 1.64) Third, for both the smoking of one cigarette a day'and for E'rS exposure, they argue that the estimated excess risks (39% and 30% respectively) are much higher than would be expected (4% and 0.8% respectively) based on simple linear extrapolation from the observed excess risk of 78% in smokers of 20 cigarettes per day. Fourth. they estimate that confounding by differences in diet associated with ETS cxposure only explains a rclativc risk of 1.06, a bias insufficient to explain the relative risk of 1.30 estimated from the meta-analysis. The bias due to confounding by diet was estimated by two completely different techniques:

04:03P 018! 642 2135 (i) (ii) 2 "direct estimate" :- based on the magnitude of the association of diet with 1HD and on the magnitude of the difference in diet between nonsmokers ~vho live and do not live with smokers; "indirect estimate" :- based on the excess risk of IHD observed in long term ex- smokers. Fifth, based on a UK epidemiological study relating platelet aggregation to risk of subsequent IHD, and on various short term studies relating smoking and ETS exposure to platelet aggregation, they estimate that ETS exposure is likely to increase risk of IHD by 34%, due to its effects on platelet aggrega.tion. They regard the increase in platelet aggregation as providing a plausible and quantitatively consistent mechanism for the unexpectedly high risks associated with ETS and low dose cigarette smoking. P.03

.30 -~).c t - 97 04:03P 0181 642 2135 P.04 3 Weaknesses of the claim~ The evidence presented by Law, Morris a~d Wald is seriously misleading. As is made clear in section 4, where our comments are elaborated in more detail, major weaknesses of the paper are as follows: E~clusion from the recta-analysis of data from the American Cancer Society [ACS] Cancer Prevention Study-I [CPS-I] is totally unjustified, and seriously distorts the estimated association of ETS with spousal smoking. There is also no good reason to exclude data from the National Mortality Followback Survey [IxlMFS] (see section 4. !. ! ). The fact that the combined epidemiological evidence shows no significant association between ETS and workplace exposure never emerges, partly because Law, Morris and Wald restrict detailed attention to spousal smoking as the index of ETS exposure and partly because they misleadingly cite an out-of-date and erroneous meta-analysis by Wells (1995) (section 4.2). Bias due to misclassification of active smoking status is incorrectly assumed to be negligible. Evidence of high heart disease rates in misclassified smokers is ignored (section 4.3). Both the "direct" and "indirect" estimates of confounding bias are open to question. Confotmding could make a major contribution to the observed association (section 4.4). Publication bias is inadequately considered. Not only is there direct cvidcncc that major data sets have v,aongly been excluded from the recta-analysis, but the analysis of publication bias conducted by Laxv, Morris and Wald is inappropriate. merely attempting to refute the proposition that the whole ofthe association may result from this source of bias (section 4.5).

30~0ct-97 04:03P 018l 642 213b 10. 4 Other potential sources of bias are not considered at all (scction 4.61. The claim that the relative risks from the studies of spousal smoking and heart disease are homogeneous is unjustified; smaller, weaker studies show substantially higher risks (section 4.7). Estimates of risk from smoking one cigarette per day or from ETS exposure obtained by extrapolation from evidence in active smokers are subject to huge uncertainty (section 4.8). The theory, proposed by Law, Morris and WaId, ~vith the excess risk resulting from smoking of one cigarette a day only slightly greater than that from ETS exposure, suggests that, within nonsmokers, there would be little or no discernible dose-response with level of ETS exposure. Ho~vever Law, Morris and Wald do not even consider the evidence on dose-response from the spousal smoking studies. Although these results arc heterogeneous, a number report a statistically significant trend, in apparent conflict with the theory (section 4.9). There are considerable difficulties in interpreting the evidence on platelet aggregation as relevant to the possible effect of ETS on heart disease (section 4.10).

' 30-8ct-97 04:04P 0181 642 2135 P.06 An alternative view of the evi_d_e_I!_e~ Based on the evidence available it is possible to arrive at an alternative interpretation very different from that put forward by Law, Morris and Wald. Exposure to ETS is very much less than is exposure to tobacco smoke. Based on the evidence for active smoking it is not possible to infer with confidence that low exposures are associated with any excess risk of IHD, let alone with an excess risk of 30% or so. When all the epidemiologieal evidence relating ETS to heart disease is considered the magnitude of any association is clearly substantially less than the relative risk estimate of 1.30 cited by Law, Morris ~md Wald. It is quite plausible that the various sources of bias and confounding, when taken properly into account, could explain the whole of the observed association. It is also possible that a true excess risk may exist, much smaller than the increase of a quarter claimed by Law, Morris and Wald.

~0-{)cL-97 04:04P 0|81 4.1 4.1.1 Detailed comments 6 Data and studies included in recta-analysis • Law, Morris and Wald present, in their Figure 1, and use in their morn-analysis, relative risk estimates for spousal smoking adjusted for age and sex from 19 epidemiological studies. Elsewhere, in a paper made available at the 10th World Conference on Tobacco or Health, held in Beijing, Lee (1997) presents results of an independent recta-analysis based on data published by the end of 1996. It is useful to compare the data and studies included in the v, vo recta-analyses, especially since they have a very great effect on the conclusions. Lee (1997) considers data from 23 studies, covering all the 19 studies considered by Law, Morris and Wald, and four additional studies. For two studies (Palmer, 1988; Marmino, 1995) only relative risk estimates (respectively 1.20 and 1.12) and not confidence limits were presented so the data could not usefully be included in recta- analyses. The other two studies are the ACS CPS-I study (LeVois and Layard, 1995) and the NMFS study (Layard, 1995). deliberately excluded by Law, Morris and Wald because Layard and LeVois were consultants to the tobacco industry, because the reported results were inconsistent with those of the other studies considered by Law. Morris and Wald and because the analysis by Layard and LeVois ofdata from the ACS CPS-II study was considered inconsistent with the results of a later analysis commissioned by the American Cancer Society (Steenland et al, 1996). The 19 studies included by Law, Morris and Wald involved a total of 6600 [HD events among never smokers. The NMFS study involved 1389 IHD deaths while the ACS CPS-I! study involved 14,891. Inasmuch as the NMFS study data are publicly available, it was clearly open to Law, Morris and Waid to access the data and conduct their own analyses if they did not betieve the results reported by Layard (1995). Failure to do so limits the extent to which

3~±0ct-97 0¢:04P 0181 64~ 2135 P.08 4.1.2. 7 the data considered by Law, Morris and Wald can be regarded as representative. However, omission of the study is clearly less important than is omission of the huge ACS CPS-I study. Issues relatin8 to this are discussed in more detail in the section that follows. Failure to include data from the ACS CPS-! study The ACS CPS-I study involved more than one million men and women in 25 US states in 1959-60 follosved up until 1972. Its results relating to smoking and health are widely cited and indeed Law', Morris and Wald cite some of its results for active smoking in their paper. Subjects were asked about their own smoking habits but not about smoking by their spouse or about ETS exposure. However, as is also the case for other well-known ETS epidemiological studies (e.g. Hirayama, 1981). inte~ie',vs were conducted on all adults in the household so it was possible to identify spousal smoking status from the responses of the spouse. In 1981 ,the ACS reported results relating to spousal smoking and lung cancer from CPS-I (Garfinkcl, 1981 ), based on a total of 153 lung cancer cases in never smoking xvomen. Since IHD in a never smoker is very. much commoner than is lung cancer in a never smoker, it has been evident for many years that the study has the potential to provide valuable data relating spousal smoking to risk of IHD (and other diseases also). Lee has, on a number of occasions (Lee, 1990. 1991a, 1991b, 1992a, 1992b), made it clear that the failure of the ACS to provide results from CPS-I may have caused severe bias to the published literature on ETS and IHD. He notes (Lee, 1992b) that in about 1987 he visited the ACS in New York and had been told by Gaxfinkel that they had examined the data on spousal smoking and IHD from CPS-I but had found no relationship. At that time Garfinkel had said they were awaiting results from CPS-II before publishing.

30-0ct-97 04:05P 0181 64? Z [35 ~*.u~ To this date, the ACS have never published data from CPS-I though they have published data from CPS-II. In a recently published correspondence in Circulation (LeVois, 1997; Steenland et al, 1997; Glantz and Parmley, 1997), arising c~ut of the Steenland et al (1996) paper, the ACS argue that they did not analyse ETS exposure among never smokers in CPS-I because there were no direct questions on ETS exposure and therefore no information on ETS exposure outside the home, so making it difl"tcult to identify a truly non-exposed comparison group. While clearly, in an ideal world, one would like to have data on all sources of ETS exposure, this hardly .seems a reason for non-publication of the results. After all, much of the published literature relates only to spousal smoking as an index of exposure, and Law, Morris and Wald have restricted attention to this index. Law, Morris and Wald do not actually cite inadequacy of the ETS data from CPS- I as a reason for not including the results of LeVois and Layard (1995) in their recta- analysis. Their reasons relate more to suspicions about the validity of the analyses reported by LeVois and Layard. There are two major points to be made here. First, if Law, Morris and Wald had such suspicions, then surely it was absolutely imperative for them to carry out their own analyses ofthe data. With the study providing information on about twice as many cases of IHD as the rest of the published evidence put together, there is no way that its results should be excluded from any self-respecting overview of the data. Second, there seems no great reason to express doubts regarding the validity of the analyses of LeVois and Layard (1995). For both CPS-I and CPS-II they presented results relating risk ofll ID among never smokers to the smoking habits reported by the spouse as follows:

"03-NOU-1997 15:30 GRLLRHER CORPORRTE RFFRS 01932832532 9 Sr~tise smokinff habits Men Women Men Women ~ever smoked 1.00 1.00 i.00 1.00 F.~-smoker 0.95(0.83-1.09) 0.99(0.93-1.05) 0.81(0.70.0.93) 0.99(0.$6. I. Curren| l-I 9 ¢igr,/day 0.99(0.$9-1.09) 1.04(0.97-1.12) l.~]6(l.10-1.65) 1.14(0.86-l.51 } Current 20-39 rigs/day 0.98(0.85-I.13) 1.06(0.95-1.1 $) 1.26(I,00-1.58) 0,98(0.75-1.29) Current ~)+ cigs/day 0.72(0.4 l- 1.2.8) 0.95(0.78- I. ! 5) !, 13(0.6 !-2. I 1 ) 1.27(0.$0-2.0 I • Pipe/cigar only ! •06(0.99. I.I 4) 0.98(0.79- i.20) Ever amoked 0.97(0.90.1.05) 1.03(0.98-1.08) 0.97(0.87-1.08) Note: Relative risks ~nd Cls are adjusted for age ~nd race. For CPS-ff the most comparable results reported by Steenland et ol (i 996) arc as follows: Spouse smoking habits Men Women Never smoked 1.00 1.00 Ex-smoker 0.96(0.$3- I. 1 I) 1.00(0.88- !. ! 3) Current. ! - 19 eig~,'d~y 1.330.09-1.6 I) I. 15(0.90. I .d 8) Current 20 cigrddgy I.I 7(0.92-I .48) 1,07(0.83- 1.40) Current 20+ (2'1-39) cigs/day !.09(0.77-I.$3) 0.99(0.67-1.47) Current 40+ elf, s/day ! .04(0.67- i .6 l) Current any 1.22(i.07- i.40) I. ! 01~0.96-1.27) Note: Relative fisk~ end Cls tre adjured for ~ge, self-reported hislory of heart disease, hypertension. diabc~s end at~u'llls, body mass index, education, use of aspirin, dbretics, aestrogen and alcohol, .. exercise and employment ~tus. Comparing the LeVois and Layard (1995) CPS4I results with those reported by Steenland e~' m~ 0996) it can bc sccn that, despite differences in adjustment factors, subject exclusion criteria and groupings of amount smoked, there are considerable similarities in the reported findings. Thus: • (i) they both show no evidence of an increase in risk for ex-smoking spouses, though LcVois and Layard show rather lower risks for men, P.02/26

"O~-HOU-1997 15:30 GALLAHER CORPORA%E AFFRS 81932832532 (iii) I0 they both show no evidence of a dose-relationship with amount smoked currently , by, the spouse, and they both show some indication of an increased ri~ for current smoking spouses ovcrall, similar except far women married to heavy smokcrs where LeVois and Layard show rather higher risks. (ii) ' A difference has arisen in the overall interpretation because LeVois ,-rod Layard concentrated on the st,,emery relative risk for ~,..~~ whereas Stccnland el al concentrated on the summary reladvc risk for s~_u~ current smoker. In considering d~is choice it should be noted that: (i) many of the studies considered by Law, Morris and Weld have only presented daia for spouse ever smoked, Jn rp/OspeCtiv¢ $~dJes many spouses who were current smokers at the time of interview will have become ex-smokers before the IHD event subse4uently occurred, and (iii) among those studies which present data for both spouse ever smoked ~d for spousc era'rent smoker, Steerdand et a! (I996) is the only one where the distinction mateti~lly affects the relative risk estimate. W~ile it might bc argued that LeVois and LayaM should also have presented o~erall relative risk estimates for spouse current smoker, one can hardly ~gue that their analYses are flawed. The key point to notc is that LeVois and Layard's analysis for ~7,E,5-_[ (see t~ble above) does not suggest any increase at all in IHD risk for men married to current smokcts (Lee, 1997 estimated 0,98, 95% CI 0.91 to 1.06) and only the most modest increase, for women married to current smokers (I .04, 0.99 to 1.09). P.03/26 It is clear that omission of CP$-I was unjustified and will cause considm, able bias.

~ ,03-NOU-1999 15:30 GALLAHER CORPORATE AFFRS "01932832532 P.O4/2G 4.L3 Data included ~om the studies selected The age adjusted rdativc risks and cordidcnce intervals used in the mcta-a~alysis ar¢only given in Figure 1 of Law, Morris and Waid and not as precise numbers in a t~ble. Also not presented are the individual sex estimates which have been combined together in Figure 1. Comparing the dat~ in Figure | with the data from the source papers, as given in Lec (I997), reveals some apparent ~nomalies: _T.unstall_-P~cdg_¢~¢! al (19955 : The confidence limits s~m too narrow compared with th0s~ estimated by Lee (1987) Who gives 1.32 (1.03-I .69) for unadjusted data and 1.37 • (I.07-1.75) for data adjusted for age and housing tenure. Could they have been calcu[at~.d falsely assuming the relative risks by level ~e independent? _M~uscat and W_~vnder f1995"~ : The relative risk given appears to be about 1.7. This seems inconsistent with the relative risks for spousal smoking, calculated from Table 1 of the sdurce paper which are 1.38 (0.70-2.75) for males and 1.33 (0.59-2.99) for females. S~ecnland et al (L995:} : The relative risk estimate, given in the discussion section of the paper, of 1.21 (I.06-1.3g), is based on a special analysis restricted to subjects concordant f0.r both self-reported current exposure to cigarettes and exposure to cigarettes based on spouse report. This involves only 1606 II-ID deaths compared with 38 I9 I'I,ID deaths in the analyses based only on spouse report. In viewofthe possible unrepresentativeness of the subsarnple of Subjects with the fuller information it would seem possibly better to have based the meta-analyse, on the spouse report data (as did Lee, ] 997). Kawachi et alC1996a.b) :Thc confidence limits again seem too narrow. They appear td have been 6alculat~ by meta-analysing results presented in Kawachi et al (I996a) for occasional ET$ exposure (1.58, 0.93-2.68) and for regular ETS cxposure (I.9I, 1,1 I- 3.28), wrongly assuming ~he e, timatcs arc independent, when they are not, and ignoring the combined estimate of 1,71 (1.03-7..84) given in Kawachi et al (I 996b). CirurA et al (19961 : The .some paper gives an estimate of 1.43 (0.9 to 2.0) which is not

~03-NOU-199~ 4.1.4 15:31 GRLLRHER CORPORATE AFFRS 01932832532 P.05/26 significant. significant. I2 Figure [ wrangly gives a lower confidence limit of" about !. 15 which is Apart from noting that there appears to have be=n a statisical error made in deriv.ing Icvd-combined rclative risks and confidcnc~ intervals from level-specific relative risk3 mad confidcnco intervals, two other general conclusions can be drawn from checking the data in Figure I of Law. Morris and Wald back to source. First, although the relative risks ~re stated to be for spousal smoking, a number ofst0dics used different indices of exposure (s~¢ Table 2 of Lee, 1997). Second, although the r, lative risks are stated to bc adjusted for age and sex, many are riot. For about half the studies it is only possible to derive either totally unadjusted i-elative risks or r~lative risks ~justcd both for age and a variable list of heart dise2se risk famors. Law, Morris and Wald givc no details of how they decided which relative risk to u~ in these circumsmncrs. EffeCt ofchoice of_study and data on the m~ta-ana!vsis relative risk estimate. Compared with Law, Morris and Wald's recta-analysis relative risk estimate of 1.30{1.22 to t.38), Lcc (1997) estimamd substantially lower risks, especially using fixed cffc~ts meta-analysis. index Relative risks (95% CI) Adjmt~ for Fixed cff~ts cov~at~s* Random effects Ever ,smoking No 1.02(0.99-I .06) 1.20(I.07-138) by th9 spouse Yes 1.07(I.03=I.I0) 1.18(I.09-I 32) Current smoking No 1.04(I.00-1.07) 1.20(I .0g-1.38) by th~ spous~ Yes l.Og(1.05-I .12) 1.20(I. t I-1.35) * Yc!~ - using data adjusted for covariatcs whcre possible, unadjusted data otherwise No - using data unadjusted for covariatcs where possible, adjusted data otherwise

"03-NOU-1997 4.2 GALLAHER CORPORATE AFFRS 0193283253R Seven studi~ provide |2 estimates of relative risk of IHD essocJatcd with workpl~,¢ exposure. None of these estimates is statistically significan( after adjustment forcovariates. Fixed-effectsmeta-~alysis ofthese statistically homogeneous estimates gives a relative risk of i .06 (0.95 to 1.19) unadjusted t'or covadatcs and 1.07 (0.96 to 1.I9) adjusted for covariates (Lee, 1997). By failing to conduct their ovm analyses of workplace exposure and restricting attention to spousal exposure Law, Mort.is and Weld give a false picture of the overall evidence. They elso give a false impression by citing an analysis by Wells (1995) submitted toOSHA which claimed a meta-analysis relative risk of 1.36 (1.08 to 1.7l). This is misleading as it: (i) does not cite comments by Lee (199:;) on Wells (1995) also submitted to OSHA which make it clear Wells' analysis was erroneous, and (ii) fails to make it clear the recta-analysis relative risk estimate would have been massively reduced by inclusion of data from Muscat and Wynder (! 995) and particularly from Steenland et ai (t996). P.0~/26 0 0

,03-NOU-1997 15:32 GALLAHER CORPORRTE AFFRS 0193283~532 P.07/26 14 SmokJr~ habit miscIassific~tion AIthou~hHackshaw, Law and Weld (1997)adjust spousal smoking risk estimates for lung cancer downward in an attempt to account for bias due to misclassification of smoking habits, Law, Morris and Weld (1997) rnak¢ no such adjustment when considering the evidence for [HD, arguing it will b¢ of negligible importance since the mlatlv¢ risk of IHD in smokers is so much smaller than that of lung cancer (about 2 cornp~ed ~o 20). This argument is incorrect. Whet affects the magnitude of bias is not the excess Hsk in ~, but the excess risk in misclassificd smokers. For lung cancer, FI~ckshaw# Law and Weld assume that the relative risk for misclassificd smokers will be .markedly less ihan the relative risk for all smokers, on the grounds that mi~¢lassifi¢d smokc~s tend to have. lower ¢otinine levels than average smokers, consisten[ with smoking less. This assumption may not bc um'casonable for lung cancer, but may be to~ly incorrect for heart disease. It is known that: (i) Hsk of heart disease death is high in patients with a previous myocardial infarction, (ii). patients= w~th a myocardial infection are us~ly advised by their doctors to give up.smokiZng, and (iii) miscl~ssification rates of smoking are particularly high in patients advised by their doctor to give up (Lc¢, 1988). It is reasonable tO. expect, therefore, that misclassJfied smokers will conlain a relatively high proportion of people with a previous myocardial infarction, and consequently be a relatively high risk group for [HD. R¢c, cnt results from a Danish prospective study ($uadicaRi ~/~z[, 1997) lend support to this possibility. In this study, self-reported smoking habits were recorded and serum samples.taken for cotinJn¢ deterrnination, the popul~tion then being followed-up fo~ eight years'. Cumulative heart disease incidence in self-reported nonsmokers wid~ cotinlne levels inconsistent with nonsmoking was 17.9%, based on t~vc cases. 3"his o o~ o3

~@3-NOU-19~? 15:32 GRLLRHER CORPORATE ~FFRS 15 incidence w~ not only higher th~n that in s~l f-repor',~ nonsmokcrs with cotininc levels below 100 ng/ml, 3.1% based on 43 cases, but was also higher than that in self-reported current smok~ w~t.h cotin~ne levels above 100 ng/ml, 4.3% based on 72 cruses. Though more ~ m'e ne~ed, [he much higher risk seen in {his sLudy {'or rniscl,',ssificd vs. non- rnisclassified smokers (relative Hsk 4.01, 95% CI 1.76 to 9.13) suggcsts that misc[assillcation bia~ might be at Icast ~s relevant for he.~rt disease zs ~t ~s t'or lung c4~Icer, Cbarly t'a/Jure properly to con.sid,~ smoking habit miscla.ssificadon is a m~.~or Hmit~,tJonol~e paper by L,~w, Morals ~nd Wald. P.08/26

~ ~03-NOU-1887 15:32 GALLAHER CORPORATE AFFRS @1932832532 P.@gx26 4.4 4.4. i 16 Evidence fi'om the ETS/heart dise~¢_studies themselves Although Law, Morris and V/old consider the possible magnitude of bias on theoretical grounds (se~ below) it is surprising flint they do not look in detail at the effect confounding adjustment had in the 19 epidcmiological studies they include in their mesa- analysis. Although many of these studies paid only limited attention to potential confounding vm-iables, with many not even t~Jng into account the cla~sicaJ coronary risk factors of" blood pressure, choleslerol and body m~s index, many of the studies did press.hi relative risk estimates before and after adjustment for age and v~rious other factor. The table overleaf summarizes the relevant data and shows the effect that adj~o'nenl for age, for [he o~hcr factors or for age and the other factors combined hnd on the ~e!ative. risk estimates. A striking feature of the data is how large and variable the effects of adjustment were. While Law, Morris and Wald are daim~ng that adjustment would only have a very modest effect on risk estimates, there were a considerable number of studies where adjustment decreased or increased relative r/sk estimates by a factor of 1.20 or more (highlighted in the table). In the two Chinese studies (He e¢ al, I989; He er al, 1994) adju.C.ment explained as much as $0% of the excess risk. : While the overall pattern of results is not clear, partly due to the diffi.cultics of separa.'ting effects of age adjustznent from effects of adjustment for other factors, and tartly due to variability in which other factors were considered, there is clearly a suggestion that confounding may be more important that Law, Morr/s and Wald suggest. They do note that relative risks adjusted and unadjusted for blood pressure, serum cholesterol, body mass index and a measure of social class were similar, but this conceals the fact that sp¢cific studies found that adjustment had quite iargc effect.s, more than Law. Morris and 9/aid suggest could possibly be due to confounding. 0 o~ O3 03 0 O~

~ "03-NOU-199T 15:33 G~LL~HER CORPORATE ~FFRS 01932832532 P.18x26 H!rzy~'la (1990) ' ~m'l~nd (198.~) ~d1~<i989) HumhleO990) ~h~n 0991) He ¢i Me,in C E C C C C C C C t2, 17 Relallve rbkS ^djumo ~ljusxr.d l'or ~ge corm.lares (K,) (Re) CRy) 1.16 3,51 2.70 2.~ 2.25 0.97 - 0.93 1.47 1.61 -. 3.52 I 1.15 !.3I 0.70 1.19 I.~ ~ 0.97 1.61 - 2.46 1.17 1.31 1.21 2.12 - !,~2 - !.~ I.? I 1.97 !.36 1.0~ - i.15 [.40 4.40 3.40 130 3,7~ Eff~! of adju stmenl For For a~¢ olhu ,lad othu (Kd~,) (X~X~) (R/R,) 1.[6 - ! .05 - 0.93 0.96 - 1,10 - 0.43 - !.14 o 1.70 1.19 - 0.93 - I I. ! 2 0.92 1 .~3 - 0,58 - 1,04 0.87 - 0.76 . 1.2~. - 0.77 - 0.96 - 0.94 Nero: Data Crom I.ee(1997). Bold values inflate adjustment affected risk u'pw+u'd or down,.v~d by • fnclor of 1.20 or more. r~3 o 0"~ ~0 0 ~0

~ "03-NOU-1997 4.4.2. 15:33 GRLLRHER CORPORRTE RFFR~ 81932832532 P.11/26 "Direct" cstimate~ of c0nfoundin~ Law, Morris and Na~d first ~mpt to cstimate ~e extcnt ~o which confounding by o~¢r ~sk factors ~t bi~ the ~ia~on ~een smo~ng estimate that confounding due to smoker/nonsmoker diffe~n~s in fruit and vegcmblc ~ption ~d in LDL cholc~crol would c~h ~ ~ ex~ss risk of3%, wi~ blood pre~¢ ~d o~r risk f~tors For IHD h~ving negligible conFo~ding effect. The conclusions for fi-uit and vegetable consumption arc based on an ~alysis reported in a pap¢~ that is "in press" and cannot be studied in detail. Jt is not made clear whether any adjustments havc bccn made for the fact that the reported relationship of diet • t~ disease (and consequently the estimated confounding effect) is likely to b¢ weaker than it actuldly is, due to the substantial errors inherent in estimating consumplion. Morn scri0usly, in view of the 250 or so risk factors that have been identified for~-- heart disease (KarmeI, 1987; Stokes, 1987; Hopkins and Williams, 1981), one must question whether Law, Morris and Wold have comprehensively considered the possible sources of confounding. As noted by Thornton et a! (1987) smokers have increased cxposur¢to very many lifestyI¢ risk factors. Based on. the smaller differences in fruit and vegetable consumption between nonsn~okers who do and do not live with a smoker than between nonsmokers and smokers, Law, Morris and Wold estimate confounding from this source will be less important for.passive San for act/re smokJng. This assumes that th~ association between diet ar~ ]HD is similar in non.smokers and in smokers, an assumption that has not been Considered. Just as when considering smoker/nonsmoker differences, the discussion regarding differences in relation to spousal smoking is limited to only a few of the risk factors identified for heart disease. It is difficult to accept that their analyses provide a r¢tbbl¢ indication of the true effect of confounding.

~03-NOU-199T 4.4.3 15:34 GALLAHER CORPORATE AFFRS 01932832532 19 ~[/Id.j~irn~tcs ofconfound~ne Law, Morris and Wold arc arguing, cssentiaJly, th.~t the obse~ incre~d IHD ~sk in smokers is p~y cause ~d eff~t ~nd p~ly due m confounding, ~d that a~cr giving up ~oking for m~y ye~s the in~cased risk duc to smo~ng is elimina~d d~e inc--due to ~nfo~ding remains. Bemuse the reladvc risk of he~ disease in long te~ ex-smokers ~ '~timated ~ only ! .06, ~ey suggc~ ~at ibis estimate se~ ~ upp~ limit to ~y eff~t of~nfounding. This argument would only have some credence if the distribution of hca,-t disease ~isk factors was similar in current smokers and cx-smokers and, within ex-smokcrs, was similar regard|ess of amount smoked. This is clearly not true. As Thornton eta/ (! 994) clcady showed, for many lifestyle Hsk factors, current smokers have the highest prevalence, ne'~er smokers the leasl and exosmokers intermediate, with [bc difference between ex-smoke~J and never smokers greatesl for shor~ term ex-smokers and least for tong term ex-smokcrs. Based on their results, any confounding effect for long term smokers would be only a fraction of the confounding effect for current smokers. To some extent Law, Morris and Waid recognize this point when they note that people who give up smoking may change their diet. But this is far From the whole story. Increasing i~gth ofex-smoking is significantly associated, according to Thornton et al (! 994), with higher social class, bet'zr education, higher income, working less in "dirty" • jobs, being more likely to do something to keep healthy, as well as numerous aspects of improvement in diet (including less fried foods, more cereal, more fi'uits, vegetables and P.12/26

" ~ 03-NOU-1997 4.5 15:34 GALLAHER CORPORATE RFFRS 81932832532 2O Law, Morris end Wald note that eigh~ studies found a significant (p<0.05) positive result, a result one would expect to see by ehan~, not in the/.9 studies they consider, but in 320 studies. They ~hen argue that it b implausible that then: should be ~s many as 301 unpublished studies, so that. publication bias cannot account for the Observed .association. Thc.r~ ~re two major weadmesses with this arBument. The first is that it ordy dens with the rather uninteresting hypothesis that publication bias ~x.otmts for all the observed associ~.ion. 'l"hJs is hardly relevant when the studies are likely to b~ affected by various other forms of bias, and does not argue against the possibility that the true association is in fact weber zb, n th~ 0bserved in published studies. The second is that it totally ignores the fact that exis6rtg studies (CPS-I and NMFS) have been omitted lYom analysis, although there is evidence that the relative risks from these studies are substantially less Ihan the calculated morn-analysis estimate. It is in fact difficult to understand how anyone could possibly try. to argue that publicatlon bias is not.an issue given thcir knowlcdg¢ ofth¢ existence of these studies and of various warnings in the literature about the likely bias from failing to consider their results (Lee. 1992: LeVois and Layard, 1995). P.13/26

15:34 4.6 4.6.~! 4.6.2 GRLLAHER CORPORATE ~FFRS P.14/26 2! Other ootentia! sources of_bias Miscla,ssification bias, confounding and publication bias are the only sources of bias referred to by Law, Morris and Wold. There are in fact other sources of bias that might be relevant. Some of these are discussed below. Presence of, or knowledge of, a disease may distort recall of a past exposure. Reca/l biasis a particular concern where th~ respondent may seek to blame the exposure • for the disease. Though most obviously a potential problem in case-control and cross- sectional studies, recall bias may also occur in prospeclive studies of a condition like Ih'D, where subjects at high risk of subscqucnt death as a result of angina or a. previous myocardial infar~'tion might mor~ readily report exposure. In a cross-sectional study, Tunstall-Pcdoc el al 0995) reported no significant association I:x:tween prevalence of either IHD or angina to level of serum cotinin¢ in nonsmokers, but in contrast reported a significant association of both IHD and angina with the level of self-reported ETS cxposure. The authors suggested the self-reported exposure data could be biased, with study participants with. symptoms of disease exaggerating exposure. Errors in diagnosing heart disease arc likely to have occurred, especially since .many of the prospective studWs relied on de, ath certificate diagnosis without any. independent confirmation. In the same study noted above, Tunstall-Pedoe et ai (I 995) ireported that s~rum cotinine level was strongly associated with "diagnosed" [.HD (where the subject ~ported a medical diagnosis of angina, .myocardial infarction, coronary thrombosis or heart attack), but was unassociated with "undiagnosed" IHD (based on tests carried out at the time of interview). Inasmuch as one would expect the two associations to be similar, if hear~ disease is actually affected by ETS exposure, the possibility arises that some of the differency-.s in association occurred because ETS exposure was. associated with the extent to which subjects reported existing, medically

~3-NOU-igg7 15:35 GALLAHER CORPORATE AFFRS 22 diagnosed disease, or reported their symptoms to their doctor. P. 15/26

. ~ .' aO3-NOU-1999 15:35 GRLLRHER CORPORRTE RFFRS 819328~25~2 P.16/26 • 23 I:[¢.~genei~ of relative risks Law, Morris and Waid argue that there is no significant heterogeneity between the relative risks for spousal smoking that they ~mbined together in their recta-analysis. This conclusi0n was .primarily ~ on a chi~u~ed test based on the 26 sex-specific estimates that had been combined, though they also noted that summary estimates were similar in women and in men and in cohort and case-control studies. Them is ~ very limited investigation of heterogeneity with no attempt to look at wh¢thcr results vary by such factors as region, date of publication, study size, study quality and fatal/non.fatal endpoint. Lee (1997) carried out a detailed investigation of:k~ sources of heterogeneity and came up with a conclusion directly opposing that of Law, • Morris a~d W~ld. H¢ noted th=t study size and study quality were both strongly • ssociated with.Hsk of IHD. He divided the 23 studies he considered into three groups: A. 1.4 studies of"'worse quality", including those studies reporting only as abstracts or dissertations and not subject to peer review, very small studies involving less than 100 IHD cams, one srady where the data for eases and controls had been collecied in a clearly non-comparable manner and one study where the exposure index had.been shown to be particularly susceptible to bias. 13. 6 studies of"better quality" with less than I000 cases, and . C.. 3. studies of"better quality" with more than 1000 cases (which included the CP$- I. CP$-ff and lsIMF$ studies). Using relative risks Imsed on currant smoking by the spouse, Lee (1997) reported estimates that were hugely different for each of" the three groups: 1.53 (1.33 to 1.77) for A, 1.22 (i .i ! to !.34) for B and 1.04 (1.00 to 1.08) for C. Clearly part of the reason that Lee found heterogeneity and Law, Morris and W'ald did not is because Lee included the results from thc CPS-I and NMFS studies which Law, Morris and Wald excluded. However, this is obviously not the whole study, with "worse quality" studies reporting markedly higher relative risks than "better quality" studies. The l'ailure of Law, Morris and Wald to consider sources of hctcrogen¢ity

• ' ~03-NOU-1999 15:36 GALLAHER CORPORATE AFFRS 81932832532 24 properly or s~udy quality properly iimiis their interprc~tion o[" t.hc cvictc~ce.

-' -83-NOU-1997 15:36 GALLRHER CORPORATE AFFRS 81932832532 P.18/26 4'.8 25 -- |raoolati, on from Law, Morris ~nd Wa|d us~ da~ ~om five l~gc pro~cfivc sludges conducted in ~c US ~d UK ~ ~dmate ~e relative risk oflHD by ~o~t smoked (age s~d~diz~ to' ag~ 65). ~cy @cn linc~ly extra,late ~c risks for active smokers backw~ds to e~m~te ~e Hsk of [HD for smokers o~ one cig=ett= ~ d~y ~d ~ar ~TS exposed nonsmokers (~s~ to smoke the equivalent of 0.2 cig~e[t~ a day). To illustrate the weakness of this approach consider a hypothetical study with a dose-response relationship similar to that seen in the prospective sf,dies. 3 Here the x-axis is consumption ia cigarett~day ~d the y-axis is the relative risk, The four solid circlcs are the actual data points, the dott~ line is the fitted dosc-respon.~ line, and the open circles are the fitted responses at I cig/day and at 0.2 cigs/day. Assuming the data points are accurate estimators of risk. why should one regard the extrapolated poiat$ as valid7 One knows that the true n:lationship cannot be linear indefinitelyas it must pass through the point (0,1). Why should it remain linear all the way back to 0.2 cigarettes a day? Given the data. there are all so~ of possible alternative dosc-relationships as illustrated bclow.

"~3-NOU-1997 15:36 GALLAHER CORPORATE AFFRS 01932832532 P. 19z26 : On What basis is one to assume, as Law, Morris and Wald do, that the true dose- response relationship is similzr to A rather than to B, C, D or E? The different models imply very different responses at an exposure of O.2 cigarettes a day. There is aJso a problem wi(h the whole argument in that no ~ccount is taken of the fact that.epidemiological studics of active smoking compute relative risks compared to ~, whcrtas epidcmiological studics of ETS compute relative risks Compared to non c~. osed nonsmokers. The extrapolation of Law, Morris and Weld is acttmI!y suggesting t.hat the risk of ETS exposed nonsmokers relative to all nonsmokers is' 1.30. If that is actually true, the risk of. ETS exposed nonsmokers relative to non exposed nonsmokers will be substantially higher, e.g. 1.86 if one assumes half of • nonsmokers ar~ exposed (50@/0 x 0.7 + 50% x 1.3 -- I ~d 1.3/0.7 = l.g6). Since such a risk is highly inconsistent with tl~e ETS epidcmiology, the whole argument of Law, Man'is and Wald is blown out. of the w~tcr.

~03-NOU-1997 15:39 GALLAHER CORPORATE AFFRS 0193283253R P.20/26 4.9 27 Dose-r~_~nse-r~,l~tionsh~ps [.~w, Mon'b and W~Id argue, ba.scd on cot~ninc da~, ~hat averag~ ETS ~xposure ~s P~luiv~Ic.nt ~o I% ofev~ag¢ smoking~ i.m abou{ 0.2 cigs/H~y. Cotinine levels resulting From ETS ~xposore do not exceed 10% of the levels of ~vera~e ET$ exposure. ~uivalcn{ on the same basis to abou~ two cigs/day. Relative to all nonsmokers, the Hsks for nonsmokers exposed to the heaviest leve|s of ETS exposure are, according to the dose- response of' Law, Morris and Wald, 1.41, while the risks for nonsmokers exposed to average levels are 1.30. This implies a relative risk for heaviest compared to average ET$ exposure of 1.05, i.e. an extremely flatdose-response. A major limitation or'the whole paper is that Law, Morris and Waid never look at all at the do/c-response cvidenco from the epidemiological studies of ETS and IHD, let alone at whether if. fits ia with their theories. (i) (ii) Had they done so they would have bbserved (see L~¢, 1997, Table 6): no real evidence of a dose-response relationship in the three I~gest studies (CPS- I, CPS-II and NMFS), and striking evidence of a dose-response relat/onship in the other 9 studies providing data. Of the 12relationships presented in these studies, I 1 were monotonically increasing, 8 showed a statistically significant Co<0.05) trend, 6 had relative risks gr~aterihan 2 in the highest cxposed group, and :5 had rcladve risks greater than 4 in the highest exposed group. These data not only show extreme heterogeneity between the two groups of studies, but also show, in a substantial number of studies, a pattern of dose-response totally conflicting with the flat dose-response predicted by the theories of Law, Morris and Waid.

.~ ~83-NOU-1997 15:37 GALLAHER CORPORATE AFFRS 81932832532 P.21/26 41.10 To be addr¢sse~ separately, by Dr F J C Roe. 0 O~ C.~ CO O3 ..b, 0

'03-NOU-199V 15:37 GALLAHER CORPORATE AFFRS 81932832532 P.22/26 References Butler TL. The relationship of passive smoking [o various health outcomes among Seven{h-day Adventists in California [Dissertation]. Los Angeles: University o[" C,dif'omia, 1988. Ciru~.i M el ah P~sive smoking ~nd the risk of~cutc myoca.rdial infarction. European Hea,,'t .~oumal 1996; I ?:309. Dobson AJ el oL Passive smoking and the risk or'heart atlack or coronary death. Medical J'oum~J of Austl'alia 1991 ;154:793-797. OUfink¢l L. Time trends in lung cancer mortality among nonsmokers and a note on passive smoking. Journal of the National Cancer [nsSrutc 1981 ;66:106l-1066. Garland C el al. Effe, c~ of passive smoking on ischcmic hcart disease mortality of non-smokers. American J'oumal of Epidcmiology ! 985;! 21:645-650. (31antz SA and ParmJey WW. Environmen.ta! tobacco smoke and coronary heart disease [L~tter]. Circulation I997;90:2088-2089. Hackshaw AK .et ol. The accumulated evidence on lung cancer and environmental tobacco smoke. British Medical Journal 1997;3! 5:980-988. He Y. Women's passive smoking and coronary heart disease. Chung Hun Yu Fang I Hsueh Tsa Chih ! 959;23:19-22. He Y el al. Passive smoking at work as a risk factor for coronary heart disease in Chinese women who have never smoked. British Medical Journal 1994;30g:380-384. Hirayama T. Non-smoking wives of heavy smokers have a higher risk of lung cancer: a study from Japma. Bdtish Medical Jouma[ 1981 ;282: i 83-185.

~ ~@3-NOU-lgg7 15:38 GRLLRHER CORPORRTE RFFRS @I~32832532 3O Hiray~m~ T. Passive smoking [Letter]. New Zealand Medical Journal 1990;I 03:54. .Hopkins P and Williams R. A survey of" 246 suggested coronary risk factors. • Ath~rosclerosiS 198 ! ;40:1-52. Humble CO et al. Passive smoking and 20-year cardiovascular disease mortality among • non.smoking wives in Evans County, Georgia. American Journal of Public Health 1990;80:599-601. Jackson R.T. Chapter 6: Passive smoking. Auckland Hear{ Survey [Thesis]. Auckland, New ~¢a[and: .University of Auckland, 1989. Kannel Wet al. Cigarette smoking, and risk ofcardiovascular disease: insighfs from the Framingham study. American Heart Journal 1987:i 13:1006.1010. Kawachi I el al. A prospective study of passive smoking and coronary heart disease [Abstract]. American Journal of Epidemiology 1996a; 143:570. Kawachi I and Coldilz GA. Invited commentary: confounding, measurement error, and publication bias in studies of passive smoking. American Journal of Epidcmiology 1996b; 144:909-915. La Vecchia C er al. Passive smoking and the risk of acute myocardial infarction [l.~r]. Lancet i 993;34 i :50~-506. Layard MW. Ischemi¢ heart disease and spousal smoking in the National Mortality Foll0wback Survey. Regulatory Toxicology and Pharmacology 1995;21 :i 80-I 83. Lec PN el al. Relationship of passive smoking to risk of lung cancer and other smoking-associated diseases. British Journal of Cancer 1986;54:97-I 05. P.23/26

~ -- " 03-NOU-1997 15:38 6RLLRHER CO~PORRTE RFFRS 01932832532 P,24/2~ 31 Lee PN. Misci~Siflcation of smoking habits and passive smoking. A review of" the evidence, lnt~mation,,l Archives of Occupational and Environmental Health Supplemcnl. .Heicl¢lberg: SpHnger-Verlag, 198g: ! 14pp. L¢~ PN. P~sive smoking and cm'diorespirator7 health in Scotland [Letter]. British Medical Journal 1990;300.:120-121. Lee PN: Weaknes.s~ in re.cent risk assessmen~ of environmental tobacco smoke. • Enviroe_,'aental Technology 1991a; [2:193-208. Lee PN. An estimate of adult mortality in the Unit=d States from passive smoking [Leu.cr]. En,;,iroru'nent International 1991 b;i 7:379.397. Lee PN., Environmental tobacco smoke and mortality. Basle: Karger, ! 992a. ~L¢¢ PN. An estimate of adult mortality in the United States fTom passive smoking. Further comment [Letter]. Environment International 1792b;18:315-317. Lee PN. Federal R.cgister/Voi 59 No 6S/Tuesday April S, 1994/?roposed rules. Environmental tobacco smoke. Comments on post-hearing comments by A .~udson Wells .concerning meta~analysis of relative Hsks from ETS exposure in the workplace. Lee PN. A review of the epidemio{ogy oFET$ and heart disease. Unpublished, 1997. Avai|ab[e on request from P.N.Le.¢ Statistics and Computing Ltd, 17 Cedar Road, Sutton, Surrey, SM2 5DA, UK. LeVois ME and Layard MW. Publication bias in the cnvironmental tobacco :smoke/coronary heart disease epidemiologic literature. Regulatory Toxicology and :pharmacology t995;21:184-191. .Levols ME. Environmental tobacco smoke and coronary heart disease [Letter].

~ .... ~3-NOU-1997 15:38 GALLAHER CORPORATE AFFRS 01932832532 Circular/on 1997;96:2086-2087. 32 Mannino DM et al. Health effects of environmental tobacco smoke exposure in US adults: data from the i 991 Natiomd Health/nterview Survey. Epidemioiogy 1995;6:568. Martin M./et al.. Increased in:cidcnce of heart a~cks in non-smoking women married ~o smokers. Paper; presented at annual meeting of American Public Health Association. 1986. Muscat ./E and Wynder EL. iExposure to cnvironmental tobacco smoke and the risk of heart attack, i'ntCmational Journal of Epidcmiolog7 ! 995 ;24:715-719. Palmer JR. el aL Passive smoking and myocardial infarction in women. CVD Newsletter 1.988;43:29. .Sandier DP et al. Deaths from all causes in non-smokers who lived with smokers. American Journal of Public Health 1989;79:163-I 67. Steenl~nd K, el al. Environmental tobacco smoke and coronary heea't disease in American Cancer Society CPS-ll cohort. Circulation 1996;94:622-628. ,Steenland K el ¢I. Environmental tobacco smoke and coronary heart disease [Letter]. Circulation 1997;96:2087-208g. .'Stokes J et.al. The relative importance of selected risk factors for various manifestations of csrdiovascular disease among men and women from 35 to 64 years old: 30 years of follow-up on the Framingham study. Circulation ]g87;75(Supp V):V65-73. P.25/26 0 O~ Suadicatd P e! ai. Mortality and morbidity of potentially misclassified smokers. International Journal 0/'Epidemiology 1~)97;26:321-327.

~. .... 03-NOU-1999 15:39 GALLAHER CORPORATE AFFRS 81932832532 P.26/26 Sv~ndscn K.H e~ ~I. Efl'~t~ of passive smoking in the ~ultiple Risk F~¢tor lntcrven~,ien Trial. American Journal Thornton A el:al. Differences belween smokers, ex-smokers, passive smokers nob-smokers. Journal of Clinical Epidcmiology 1994;47:1 ! 43-1162. Tunstall-Pecl0e H el al. Fassive smoking by self report and serum cotinine and pre~valenca: of rcspirat0~ ~d coron~ h~ dis~e in ~e Scottish h~ heal~ study. ~o~al of Epidemiology ~d Co~uni~ H~I~ 1995;49:l 39-143. V/ells AJ. Post-hearing comment re OSHA hem-ing: recta-analysis of relative risk of he,~t disease from ETS:exposure in ~he workplace. Washington DC: US Dept of Labor .(Docket H-| 22, Docket Offic~ P, oom No2625), 1995.