AIM: Although prostate-specific antigen (PSA) screening reduces mortality from prostate cancer, substantial over-diagnosis and subsequent overtreatment are concerns. Early screening of men for PSA may serve to stratify the male population by risk of future clinical prostate cancer. METHODS AND MATERIAL: Case-control study nested within the Danish 'Diet, Cancer and Health' cohort of 27,179 men aged 50-64 at enrolment. PSA measured in serum collected at cohort entry in 1993-1997 was used to evaluate prostate cancer risk diagnosed up to 14years after. We identified 911 prostate cancer cases in the Danish Cancer Registry through 31st December 2007 1:1 age-matched with cancer-free controls. Aggressive cancer was defined as ⩾T3 or Gleason score ⩾7 or N1 or M1. Statistical analyses were based on conditional logistic regression with age as underlying time axis. RESULTS: Total PSA and free-to-total PSA ratio at baseline were strongly associated with prostate cancer risk up to 14years later. PSA was grouped in quintiles and free-to-total PSA ratio divided in three risk groups. The incidence rate ratio for prostate cancer was 150 (95% confidence interval, 72-310) among men with a total PSA in the highest quintile (>5.1ng/ml) compared to the lowest (