Greater exposure to retinol (vitamin A) may prevent prostate cancer, although under some conditions it could promote cell growth and de-differentiation. The authors prospectively examined prostate cancer risk and serum retinol levels, measured by using high-performance liquid chromatography, at baseline (n = 29,104) and after 3 years (n = 22,843) in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study cohort. Cox proportional hazards models were used to estimate the relative risk of total (n = 2,041) and aggressive (n = 461) prostate cancer by quintiles of baseline and 3-year serum retinol concentrations and by change in serum retinol levels from baseline to 3 years. Men with higher retinol concentrations at baseline were more likely to develop prostate cancer (quintile 5 vs. quintile 1 hazard ratio = 1.19, 95% confidence interval: 1.03, 1.36; P(trend) = 0.009). The results were similar for aggressive disease. Joint categorization based on baseline and 3-year retinol levels showed that men who were in the highest quintile at both time points had the greatest increased risk (baseline/3-year quintile 5/quintile 5 vs. quintile 1/quintile 1 hazard ratio = 1.31, 95% confidence interval: 1.08, 1.59). In this largest study to date of vitamin A status and subsequent risk of prostate cancer, higher serum retinol was associated with elevated risk, with sustained high exposure conferring the greatest risk. Future studies may clarify the underlying biologic mechanisms of the retinol-prostate cancer association.