Tobacco Documents Online

-1- R~EVIEW 729 CONFIDENTIAL Subject ref 8b "Childhood and adblescent passive smoking and the risk of female lung cancer" F-L Wang et ali International Journal of Epidemiology (1994), 23, 223-230 In a paper published in 1989 in the Chinese Jiournal of Preventive Medicine (2_3, 270-273) entitled "Analysis of risk factors for female lung adenocarcinoma iin Harbini: Indoor air pollution", the same leading author, based on the same study, reported that "no significant association with cigarette smoking and indoor passive smoking was found", major attention being given to strong relationships of risk of adenocarcinoma with high coal consumption, indoor smog pollution, and low ceiling height of the living room, for all' of which odds ratios reported exceed'ed 10. No relative risks were presented for passiive smoking. The present paper, results from which had previously beenicite&in a paper presented at the 26th Annual Meeting of the Society for Epidemiol'ogic Research at Keystone, Colorado in June 1993 (American Journal of Epidemiology, 1993, 1'38, 639), reports by contrast a significant relationship between risk of lung cancer and exposure to ETS in childhood, though no relationship between risk of lung cancer and exposure to ETS in adulthood. Among,the 55 case-control pairs who had never smoked, the key findings are summarized~below:

-2- Exposure period (age in yearsJ Relative risk 95% confidence level P 0-6 1.6& 1.1i5-13.33' 0.012 7-1'4 3'.401 1.08-12.69 0.019 15-22 2'.43 0.88-7'.33 0.066 23-30 0.93 0.38-2.25 NS 31-69 0.9'1 0.32-2'.53 NS The paper is based on al 1:1' paired case-control study conducted in H'arbin; in China in 1985-86. There were 114 cases ini all, aged 30-69, with; their hospital based non-cancer controls matche& on hospiital, age (+51 years), residential area,, and lifetime smoking status (ever/never). The 55; pairs who had never smoked are of most importance with regard to assessing possible effects of ETS'. Points to note about the study and its findings include the following: (i) The controls included 18% with cardiovascular disease and 22% with respiratory infections. They will certainly not be representative of the popul'.ation in terms of smoking habits (since smoking,is associated'with prevalence of these conditions) and may not be representative in terms of ETS exposure. It would have been preferable to include only controls with diseases not associlated with smoking. (ii) A remarkably similar relationship between ETS andilung,cancer was seen in the smoking pairs (n-59) and the non-smoking pairs (na55)i- see Table 5. Given the relatively wide confidence intervals, I find it surprising that the results are so similar. Thus, for example, for the exposure period 0-6 years, the ratio of relative risks for non-smoking to smoking pairs is as close to 1 as 1.08 despite the

-3- fact that the 95% confidence interval for this ratio is of order 0.2 to 5.6! (iiii) Althoughiquite large numbers of relative risks are cited, which inter aliia show a dose response relationship with ETS exposure in childhood (see Tables 6 and 7), only the results tabulated above (taken from Tabl'e 5) are based on never smokers, as is conventional for ETS analyses. (iv) Despite the 1',989 paper pointing, out huge relationships of adenocarcinoma risk with high coal consumption, indoor smog pollution,, and low ceiling height, and despite about half the lung cancers cases studied being adenocarcinoma, it is quitee remarkable that no consideration at all is given to the possible confounding effect of these variables. Indeed there is no mention at all in the paper of these variables having been recorded in the study!! (v) The discussion concerning other evidence relating childhood ETS exposure to risk of lung cancer in; adulthood is totally misleading. For example, the statement on p228 that "findings from epid'emiological studies7'9„10,17„27' have provided further evidence for the effect of PS in early life on cancer risk in adulthood" includes reference to some studies (Chilmonczyk; Jarvis) that do not even provide data on cancer risk in adulthood. Furthermore, there is no reference at alil to studies reporting no association, let alone the fact that the overall published evidence provides no indicationiwhatsoever of an association between ETS exposure in

-4- ctlildhood and risk of lung cancer in lifelong non-smokers., The results fromi this study do not affect the overall pictlure., P' N Le e 24.5.94