Reevaluation of the confounding effect of cigarette smoking on the relationship between alcohol use and lung cancer risk, with larynx cancer used as a positive control

Background. The effect of smoking on lung cancer risk has been well documen ted, while the effect of alcohol remains controversial. We examined the hyp othesis that the apparent association between alcohol intake and lung cance r risk is fully due to the confounding effect of cigarette smoke. Methods. Our sample of hospitalized patients included 2,953 male and 1,622 female lung cancer cases; 521 male and 159 female larynx cancers cases; and 8,169 male and 4,154 female controls, admitted to participating hospitals between 1981 and 1994. All controls had been diagnosed with non-smoking-rel ated diseases. Larynx cancer was used as a positive control for lung cancer . Relative risks were estimated through odds ratios, adjusted through multi ple logistic regression. Results. Although the odds ratios for alcohol had been significantly elevat ed prior to adjustment fbr smoking (OR = 2.4, 95% CI = 2.0-2.8), alcohol ha d no effect on lung cancer following this adjustment (OR = 1.2, 95% CI = 1. 0-1.4). By contrast, the effect of alcohol on larynx cancer remained high e ven after adjustment for smoking (OR = 5.6, 95% CI = 3.7-8.6). Conclusion. The often-reported association between alcohol and lung cancer risk can be fully explained by the confounding effect of cigarette use, (C) 2001 American Health Foundation and Academic Press.