-
Views
-
Cite
Cite
Harvey A. Risch, Anthony B. Miller, Re: Are Women More Susceptible to Lung Cancer?, JNCI: Journal of the National Cancer Institute, Volume 96, Issue 20, 20 October 2004, Page 1560, https://doi.org/10.1093/jnci/djh302
Close - Share Icon Share
Extract
In a recent editorial in the Journal ( 1 ) , Blot and McLaughlin criticize our case–control study ( 2 ) and other case–control studies in comparison with cohort studies addressing smoking-related lung cancer risks in males and females. In the same issue of the Journal, Bain et al. ( 3 ) also criticize our study. Both sets of authors point out that a number of carefully conducted cohort studies have shown no measurable excess of lung cancer cases among female smokers compared with male smokers, once amounts of smoking have been controlled, whereas our study showed statistically significantly and substantially higher risks for females than for males at every level of smoking. The authors offer various explanations for this discrepancy, including our use of relative risk measures, potential recall biases, interviews with next of kin, and so on ( 1 ) . However, the answer lies elsewhere. Bain et al. ( 3 ) and the other cohort studies [reviewed in ( 3 ) ] examined risks for males and females, adjusted for amount of smoking. In these studies, the scientific question addressed was, Adjusted for dose of smoking (in the same way for males and females), are males and females, or are male and female smokers, at the same risk? In other words, What is the (main) effect of sex, given adjustment for smoking? By contrast, the scientific question addressed by our case–control study and others was, At a given dose of smoking, is the effect of smoking different for males and females? That is, do males and females have different relationships between their smoking doses and risk of lung cancer (i.e., a different risk sensitivity to the effects of smoking)? This question is not addressed by the analyses of Bain et al. ( 3 ) and the others, which do not involve effect modification of smoking by gender. In a statistical model, if men and women are forced to have the same relationship between smoking and risk of lung cancer, then what is being examined is the intrinsic risk of sex itself, not of smoking. Because the major identified cause of lung cancer is tobacco smoking, we think the effect of sex by itself is of secondary importance.
