Background: To examine the risk of smoking on histopathologically-confirmed moderate- and high-grade prostate cancer.
Materials and methods: A population-based case-control study was conducted in Melbourne, Sydney and Perth between 1994 and 1998 in men aged below 70 years. Cases were recruited from cancer registries and controls were selected from electoral registers. 1498 cases and 1434 controls were interviewed and a detailed smoking history obtained. Data were analyzed by unconditional logistic regression, adjusting for age, study center, year of recruitment, family history and country of birth.
Results: The odds ratios (OR) were 1.02 (0.85-1.22) for former smoking and 0.82 (0.65-1.05) for current smoking. The respective ORs were 0.95 (0.78-1.15) and 0.76 (0.59-0.99) for moderate grade tumors, and 1.28 (0.96-1.70) and 1.00 (0.67-1.47) for high-grade tumors (P = 0.2 for test that ORs of the two grades were identical). There was no evidence of a dose-response effect for duration of smoking, amount smoked daily, pack-years of smoking and years since quitting and most ORs for these variables were close to unity.
Conclusions: Smoking was not associated with the incidence of prostate cancer. The widths and upper limits of the confidence intervals for the effects of current and former smoking were consistent with weak effects at most.