Purpose: To critically review perioperative outcomes, positive surgical margin (PSM) rates, and functional outcomes of several large series of retropubic radical prostatectomy (RRP), laparoscopic RP (LRP), and robotassisted radical prostatectomy (RARP) currently available in the literature. Methods: A Medline database search was performed from November 1994 to May 2009, using medical subject heading search terms ''prostatectomy'' and ''Outcome Assessment (Health Care)'' and text words ''retropubic,'' ''robotic,'' and ''laparoscopic.'' Only studies with a sample size of 250 or more patients were considered. Weighted means were calculated for all outcomes using the number of patients included in each study as the weighing factor. Results: We identified 30 articles for RRP, 14 for LRP, and 14 for RARP. The mean intraoperative and postoperative RRP transfusion rates for RRP, LRP, and RARP were 20.1%, 3.5%, and 1.4%, respectively. The weighted mean postoperative complication rates for RRP, LRP, and RARP were 10.3% (4.8% to 26.9%), 10.98% (8.9 to 27.7%), and 10.3% (4.3% to 15.7%), respectively. RARP revealed a mean overall PSM rate of 13.6%, whereas LRP and RRP yielded a PSM of 21.3% and 24%, respectively. The weighted mean continence rates at 12 month follow-up for RRP, LRP, and RARP were 79%, 84.8%, and 92%, respectively. The weighted mean potency rates for patients who underwent unilateral or bilateral nerve sparing, at 12 month follow-up, were 43.1% and 60.6% for RRP, 31.1% and 54% for LRP, and 59.9% and 93.5% for RARP. Conclusion: RRP, LRP, and RARP performed in high-volume centers are safe options for treatment of patients with localized prostate cancer, presenting similar overall complication rates. LRP and RARP, however, are associated with decreased operative blood loss and decreased risk of transfusion when compared with RRP. Our analysis including high-volume centers also showed lower weighted mean PSM rates and higher continence and potency rates after RARP compared with RRP and LRP. However, the lack of randomized trials precludes definitive conclusions.
