BACKGROUND. Screening for prostate cancer has not only led to a stage migration, but also to a higher incidence of the disease. A decrease in mortality has occurred in several countries during the same time period. Risk stratification of screen-detected cancers at diagnosis has become more important for the anticipation and interpretation of changing incidence/mortality ratios. METHODS. From 1993 to 1998, 633 men were diagnosed with nonmetastatic prostate cancer in the prevalence screen of the Rotterdam section of the European Randomized study of Screening for Prostate Cancer (ERSPC). The characteristics at diagnosis of men who developed metastatic disease were compared with men without evidence of metastases during follow-up. RESULTS. During the median follow-up of 7.5 years, 41 men developed metastatic disease. After 10 years the metastasis-free survival rate was 89.6%, the overall survival 64.7%. In a Cox-model 21ogPSA (prostate-specific antigen), biopsy Gleason score and the number of biopsy cores with prostate cancer were independent predictors for the development of metastases; the latter only predicted metastases that presented within 60 months of follow-up. CONCLUSION. The metastasis-free survival of men with prostate cancer detected in a prevalence screening was very high. Whether this was related to the beneficial effects of screening or to overdiagnosis due to screening (or both) remains unclear. The prognostic factors known for clinically diagnosed disease also hold for screen-detected disease.