This study reports the outcome of active surveillance in men with PSA screening-detected prostate cancer (PC), and PSA doubling time (PSADT) was evaluated as a predictor of selecting patients to active treatment or surveillance. On December 31, 1994, 10,000 men were randomized to biennial PSA testing. Through to December 2004, a total of 660 men were diagnosed with PC, of whom 270 managed with initial surveillance. Of these 270 patients, 104 (39%) received active treatment during follow-up, 70 radical prostatectomy, 24 radiation and 10 endocrine treatment. Those who received active treatment during follow-up (mean 63 months) were significantly younger (62.6 vs. 65.5 years, p < 0.0001) and had a shorter PSADT (3.7 vs. 12 years, p < 0.0001). PSA relapse was observed in 9 of 70 patients who received RRP during a mean follow-up of 37 months. Seven of these nine PSA relapses were in the patients with preoperative PSADT < 2 years. None of the 37 operated patients with a PSADT > 4 years had a PSA relapse. In a Cox regression analysis adjusted for PSA, ratio-free PSA and amount of cancer in biopsy, only the preoperative PSADT was statistically significant predictor of PSA relapse in p = 0.031. The optimal candidate for surveillance is a man with early, low-grade, low-stage PC and a PSADT > 4 years. In younger men with a PSADT of less than 4 years, surveillance does not seem to be a justified alternative, and patient should be informed about the risk with such an approach. © 2006 Wiley-Liss, Inc.