Population-based analysis of salvage radical prostatectomy with examination of factors associated with adverse perioperative outcomes

Introduction: Salvage radical prostatectomy (SRP) is a potentially curative operation performed for recurrent prostate cancer after radiation therapy (RT). The purpose of this study was to examine perioperative and pathologic outcomes of SRP in a national cohort. Materials and methods: The National Cancer Database from 1998 to 2011 was used to identify 408 patients who underwent SRP. We evaluated the demographic data, clinical variables, and pathologic and perioperative outcomes. Primary outcomes of interest were length of stay (LOS) and positive surgical margins (PSM). A multivariable logistic regression model was employed to identify preoperative predictors of LOS ≥3 days and PSM following SRP. Results: Among patients undergoing SRP, the mean age was 62.5±6.8 years, mean prostate-specific antigen (PSA) level was 12.6±15.6. ng/ml, and 64% were categorized as clinical stage T1. External beam RT (89%), followed by brachytherapy (11%), was the most common modalities for primary RT. Most SRPs were performed at high-volume centers (57%) in metropolitan locations (83%). Most patients underwent a lymph node dissection (75%) and the mean node count was 15.0±30.1. On final pathology, 73% had intermediate- or high-grade disease (Gleason score 7 in 43% and Gleason score 8-10 in 30%), 49% had pT3 or pT4 disease, 6.2% had positive lymph nodes, and 34% had PSM. The mean LOS was 3.5±9.9 days, and 43% of patients had a LOS≥3 days. The 30-day readmission rate was 6.5% after SRP, and 30- and 90-day postoperative mortality rates were extremely low at 0.4% (n = 1). On multivariable analysis of the entire cohort, PSA level>20. ng/ml was an independent predictor of PSM (OR = 3.68, 95% CI: 1.2-10.9, P = 0.018). PSA level>20. ng/ml (OR = 4.37, 95% CI: 1.2-16.2, P = 0.027) and cT2 or greater disease (OR = 2.52, 95% CI: 1.0-6.2, P = 0.046) were associated with prolonged LOS (LOS≥3. d), whereas surgery at an academic facility (OR = 0.30, 95% CI: 0.1-0.8, P = 0.02) reduced the odds of LOS≥3 days. Conclusions: In select patients derived from a population-based cohort, SRP results in effective local cancer control with acceptable perioperative outcomes. PSA level>20. ng/ml is independently associated with PSM and prolonged LOS after SRP. Clinical stage T2 or greater disease is also associated with prolonged LOS, whereas surgery at an academic facility reduced this risk.