To evaluate the pathological variables predictive of biochemical recurrence after radical prostatectomy and their implications for decision making in the adjuvant setting. METHODS: 684 patients with localized prostate cancer who were treated with radical prostatectomy between 1996 and 2007. Before surgery they were classified according to D’Amico risk groups for recurrence. Following prostatectomy the following variables were collected: Gleason score, pathological stage, capsular invasion, surgical margins, perineural invasion and percentage of involvement in the piece. Univariate analysis was performed and subsequently adjusted using a Cox proportional hazards model (method enter).RESULTS: The median follow up of the series was 61 months. 29.1% of patients had biochemical recurrence. Overall mortality of the series was 4.9% and cancerspecific mortality 1.2%. In univariate analysis the Gleason score of surgical specimens, capsular invasion, perineural invasion, involvement of surgical margins, pathological stage and percentage of involvement of the piece had statistically significant (p < 0.001) relation with biochemical recurrence. In multivariate analysis, a Gleason score ≥ 8 in the surgical specimen (HR = 3.08), existence of affected surgical margins (HR = 2.98), pT3 stage (HR = 1.61) and involvement of more than 50% of the piece by cancer (HR = 3.39) were identified as independent predictors of biochemical recurrence. Stratifying by independent predictors of biochemical recurrence (pT, Gleason score and margin), patients with at least 2 of these factors had an incidence of biochemical recurrence at 5 years exceeding 50%.CONCLUSIONS: Patients who have a Gleason score ≥ 8, positive margins, pT3 tumour or a percentage of >50% after prostatectomy have an increased risk of biochemical recurrence. Patients with at least 2 predictors of relapse have a probability of recurrence over 50% in the first 5 years of recurrence and should therefore be candidates for adjuvant radiotherapy