Peter Lee is an Oxford educated statistician who has consulted to pharmaceutical and tobacco companies. Here is a list of 80 studies into passive smoking and lung cancer. He concludes:
“The overall evidence shows no statistically significant increased risk of lung cancer in relation to ETS exposure from parents in childhood, or in social situations, or to nonspousal ETS exposure at home.
The overall evidence shows that lung cancer risk among nonsmoking women is significantly associated with having a husband who smokes (with a similar association seen in nonsmoking men in relation to smoking by the wife, though based on far less data).
There is also evidence of a dose-response relationship, with risk higher if the husband smokes more cigarettes per day or for a longer period of time. However, there are a number of reasons why this association and dose-response relationship cannot be interpreted as indicating a causal effect of ETS exposure including:
• the association is weak and is not statistically significant in the great majority of studies: over 80% show no statistically significant association between smoking by the husband and the development of lung cancer;
• the combined results vary over time, with the association being significantly weaker in the studies published from 1990 than in those published in the 1980s;
• some of the very largest studies show no association, including four of the five studies involving over 400 lung cancer cases. One of these reported no statistically significant association between lung cancer and any index of ETS exposure, while another even reported a statistically significantly reduced risk of lung cancer for non-smoking women married to smokers;
• almost 20% of the studies have not adjusted for age in the analysis, a standard procedure in epidemiology to avoid bias. These studies report much stronger associations with spousal exposure than are reported by studies that did use age-adjustment. • spousal studies are particularly susceptible to various biasing factors. These include:
i) failure to consider diet, lifestyle, family medical history, education, socio-economic status and other factors believed to differ between smoking and non-smoking households; and
ii) the inappropriate inclusion of some misclassified current and former smokers among the lifelong non-smokers.
• reliance on reported rather than objectively measured ETS exposure data, and failure to publish negative studies.
No-one has yet designed a study in such a way as to eliminate all these sources of bias. Analyses published in 2000-2002 demonstrated formally that the weak association and dose-response relationship between lung cancer and smoking by the husband would essentially disappear were proper adjustment made for age, diet, education and misclassification of smoking habits, a conclusion confirmed based on more recent data.
There is also some indication from the overall evidence that lung cancer risk among nonsmokers might be weakly associated with workplace ETS exposure. However, only seven of 39 relative risk estimates are statistically significant and biases that apply to the spousal data are also likely to apply to the workplace data.
There is similarly weak evidence of an association with overall childhood ETS exposure. Here the five largest studies give risk estimates below 1, but several of the moderate size studies report an increased risk.
Taken as a whole, the epidemiology does not support the claim that ETS causes lung cancer in non-smokers.