The etiology of penile cancer is poorly understood, with neonatal circumcision being one of the few recognized nondemographic risk factors. Multiple logistic regression was used to analyze interview data from 100 matched case-control pairs; cases of carcinoma in situ (CIS) and invasive carcinoma of the penis were analyzed separately as well as together. Phimosis was strongly associated with invasive carcinoma (adjusted odds ratio [OR] = 16; 95% confidence interval [CI] = 4.5-57) but not CIS (OR = 1.7; 95% CI = 0.32-7.8), and these associations persisted when the analyses were restricted to uncircumcised subjects. Neonatal circumcision was inversely associated with invasive carcinoma (OR = 0.41; 95% CI = 0.13-1.1) but not CIS, and the observed association with invasive carcinoma was weakened appreciably when the analysis was restricted to subjects with no history of phimosis (OR = 0.79; 95% CI = 0.29-2.6). Other factors positively associated with invasive carcinoma or CIS or both were injury to the penis, cigarette smoking, physical inactivity and, to a lesser extent, genital warts and other infections or inflammation of the penis.
Conclusions: Although many effects were imprecisely estimated in this study, the protective effect of circumcision on invasive penile cancer appears to be mediated in large part by phimosis; furthermore, the effects of certain factors such as phimosis and circumcision appear to differ for CIS and invasive carcinoma.