Alcohol is an established carcinogen but not an established risk factor for prostate cancer, despite some recent prospective studies suggesting increased risk among heavy drinkers. The aim of this study was to investigate the role of alcohol on prostate-specific antigen (PSA) levels and prostate cancer risk. Two thousand four hundred PSA detected prostate cancer cases and 12,700 controls matched on age and general practice were identified through a case-control study nested in the PSA-testing phase of a large UK-based randomized controlled trial for prostate cancer treatment (ProtecT). Linear and multinomial logistic regression models were used to estimate ratios of geometric means (RGMs) of PSA and relative risk ratios (RRRs) of prostate cancer by stage and grade, with 95% confidence intervals (CIs), associated with weekly alcohol intake and drinking patterns. We found evidence of lower PSA (RGM 0.98, 95% CI: 0.98–0.99) and decreased risk of low Gleason-grade (RRR 0.96; 95%CI 0.93–0.99) but increased risk of high-grade prostate cancer (RRR 1.04; 95%CI 0.99–1.08; pdifference=0.004) per 10 units/week increase in alcohol consumption, not explained by current BMI, blood pressure, comorbidities, or reverse causation. This is the first large population-based study to find evidence of lower PSA levels for increasing alcohol consumption, with potential public health implications for the detection of prostate cancer. Our results also support a modestly higher risk of high-grade disease for heavy drinkers, but require independent replication to establish the nature of the association of alcohol with low-grade disease, preferably in cohorts with a heterogeneous case-mix.