Abstract
The etiology of smoking-associated cancers (carcinoma of the lung, oral cavity, esophagus, pancreas, urinary bladder) remain enigmatic despite a plethora of epidemiologic investigations. We have evaluated family histories through second degree relatives of 461 consecutively ascertained patients with histologically verified smoking-associated cancers. The genetic protocol included detailed questionnaires about genealogy, cancer of all anatomic sites, and environmental factors with major attention given to the cigarette smoking habit. Pathologic verification of cancer was vigorously sought. Our findings disclosed a lack of strong evidence for heritable control of lung cancer per se, when only lung cancer in relatives was considered. Confounding factors, most prominent of which was the effect of cigarette smoking (p <.05), variation of secular trends, and heritability of the smoking phenotype (p <.05), tended to obfuscate identification of an inherited effect presenting itself exclusively as lung cancer liability. However, we observed a significant increase in cancer of all sites among siblings (p <.001) and all relatives of 240 lung cancer probands (p <.0001). Most of these carcinomas were not associated with smoking and thus not greatly influenced by secular trends. Furthermore, no significant excess of cancer of all sites in relatives of probands with other smoking-associated cancers was observed. The number of parents affected with cancer and the number of primary cancers in the proband were both positively correlated with risk (both at P <0.05). These data indicate that necessity of stratifying for host factor effects in studying cancer etiology.