OBJECTIVE: The aim of this study was to examine associations of smoking at the time of diagnosis with the risk of prostate cancer death in a population-based cohort of men with prostate cancer. METHODS: Data were from 752 prostate cancer patients aged 40-64 years, who were enrolled in a case-control study and under long-term follow-up for mortality. Cox proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for associations between smoking and prostate cancer-specific and other cause mortality. RESULTS: Compared to never smoking, smoking at the time of diagnosis was associated with a significant increase in risk of prostate cancer-specific mortality. After controlling for demographic characteristics, Gleason grade, stage at diagnosis, and primary treatment, the HR was 2.66 (95% CI: 1.10-6.43). CONCLUSIONS: Smoking at the time of diagnosis, independent of key clinical prognostic factors, is associated with an increased risk of prostate cancer death.