5Bristol Royal Inf irmary, UK Despite a significant increase of the number of radical prostatectomies (RPs) to treat organ-confined prostate cancer, there is very limited documentation of its oncological outcome in the UK. Pathological stage distribution and changes of outcome have not been audited on a consistent basis. We present the results of a multicentre review of postoperative predictive variables and prostaticspecific antigen (PSA) recurrence after RP for clinically organ-confined disease. In all, 854 patient's notes were audited for staging parameters and follow-up data obtained. Patients with neoadjuvant and adjuvant treatment as well as patients with incomplete data and follow-up were excluded. Median follow-up was 52 months for the remaining 705 patients. The median PSA was 10 ng ml À1 . A large migration towards lower PSA and stage was seen. This translated into improved PSA survival rates. Overall Kaplan -Meier PSA recurrence-free survival probability at 1, 3, 5 and 8 years was 0.83, 0.69, 0.60 and 0.48, respectively. The 5-year PSA recurrence-free survival probability for PSA ranges o4, 4.1 -10, 10.1 -20 and 420 ng ml À1 was 0.82, 0.73, 0.59 and 0.20, respectively (log rank, Po0.0001). PSA recurrence-free survival probabilities for pathological Gleason grade 2 -4, 5 and 6, 7 and 8 -10 at 5 years were 0.84, 0.66, 0.55 and 0.21, respectively (log rank, Po0.0001). Similarly, 5-year PSA recurrence-free survival probabilities for pathological stages T2a, T2b, T3a, T3b and T4 were 0.82, 0.78, 0.48, 0.23 and 0.12, respectively (log rank, P ¼ 0.0012). Oncological outcome after RP has improved over time in the UK. PSA recurrence-free survival estimates are less optimistic compared to quoted survival figures in the literature. Survival figures based on pathological stage and Gleason grade may serve to counsel patients postoperatively and to stratify patients better for adjuvant treatment.