Abstract
The relationship between tobacco smoking and prostate cancer (PCa) remains inconclusive. This study examined the association between tobacco smoking and PCa risk taking into account polymorphisms in carcinogen metabolism enzyme genes as possible effect modifiers (9 polymorphisms and 1 predicted phenotype from metabolism enzyme genes). The study included cases (n = 761 localized; n = 1199 advanced) and controls (n = 1139) from the multiethnic California Collaborative Case-Control Study of Prostate Cancer. Multivariable conditional logistic regression was performed to evaluate the association between tobacco smoking variables and risk of localized and advanced PCa risk. Being a former smoker, regardless of time of quit smoking, was associated with an increased risk of localized PCa (odds ratio [OR] = 1.3; 95% confidence interval [CI] = 1.0-1.6). Among non-Hispanic Whites, ever smoking was associated with an increased risk of localized PCa (OR = 1.5; 95% CI = 1.1-2.1), whereas current smoking was associated with risk of advanced PCa (OR = 1.4; 95% CI = 1.0-1.9). However, no associations were observed between smoking intensity, duration or pack-year variables, and advanced PCa. No statistically significant trends were seen among Hispanics or African-Americans. The relationship between smoking status and PCa risk was modified by the CYP1A2 rs7662551 polymorphism (P-interaction = 0.008). In conclusion, tobacco smoking was associated with risk of PCa, primarily localized disease among non-Hispanic Whites. This association was modified by a genetic variant in CYP1A2, thus supporting a role for tobacco carcinogens in PCa risk.