BACKGROUND: Current prostate cancer (PCa) follow-up guidelines do not account for the risk of disease relapse. OBJECTIVE: To examine the annual hazard rate (anHR) of biochemical recurrence (BCR) according to risk strata in patients treated with radical prostatectomy (RP) for localised PCa. These rates might be used to devise a risk-adjusted follow-up. DESIGN, SETTING, AND PARTICIPANTS: From January 1992 to December 2005, 2911 patients underwent RP for localised PCa in one institution. This cohort was used to identify three distinct risk groups for BCR. A cohort of 2875 patients operated on in a second institution was used for validation purpose. INTERVENTION: RP, prostate-specific antigen (PSA) tests. MEASUREMENTS: Cox regression models addressing BCR were used to identify significant predictors and cut-offs for risk group stratification. The anHR for BCR was calculated (number of events divided by number of patients at risk) for each risk group. RESULTS AND LIMITATIONS: Three risk groups could be identified: (1) low risk (23.7%), defined as PSA 22ng/ml or seminal vesicle invasion or pathological Gleason sum >8 or lymph node invasion or clinical stage T3; and (3) intermediate risk (57.4%), defined as all other patients. The anHR for the low-risk groups remained very low throughout follow-up (0-2.6). The anHR in the intermediate-risk group was initially low but remained elevated (1.3-7.2). The anHR for the high-risk group was initially markedly high (up to 32) and remained elevated during follow-up. CONCLUSIONS: Annual hazard rates of BCR differ according to risk strata. These data might be used to devise a risk-adjusted follow-up protocol. Low-risk patients appear to need less frequent follow-up, whereas high-risk patients might need to be followed more frequently, relative to the current recommendations.
