Predicting erectile function recovery after bilateral nerve sparing radical prostatectomy: a proposal of a novel preoperative risk stratification.

INTRODUCTION: No multivariable model is currently available for the prediction of erectile function (EF) recovery after bilateral nerve sparing radical prostatectomy (BNSRP). AIM: The aim of this study was to develop a novel preoperative risk stratification aimed at assessing the probability of EF recovery after BNSRP. MAIN OUTCOME MEASURE: The International Index of Erectile Function (IIEF) was used to evaluate EF after BNSRP.
METHODS: This study included 435 patients treated with retropubic BNSRP between 2004 and 2008 at a single Institution. Preoperative data, including age, IIEF, Charlson comorbidity index (CCI), and body mass index (BMI) were available for all patients. Moreover, all patients were assessed postoperatively every 3 months and were asked to complete the IIEF during each visit. Cox regression models tested the association between preoperative predictors (age at surgery, preoperative IIEF-EF domain score, CCI, BMI) and EF recovery. Independent predictors of EF recovery were then used to stratify patients into three groups according to the risk of erectile dysfunction (ED) after surgery: low (age <or= 65 years, IIEF-EF >or= 26, CCI <or= 1; n = 184), intermediate (age 66-69 years or IIEF-EF 11-25,CCI <or= 1; n = 115), and high (age >or= 70 years or IIEF-EF <or= 10 or CCI >or= 2; n = 136). Kaplan-Meier curves assessed the time to EF recovery (defined as IIEF-EF score >or= 22). Predictive accuracy of our proposed classification was quantified using the AUC method.
RESULTS: Of 435 patients, 242 (55.6%) received phosphodiesterase type 5 inhibitors (PDE5-I) either on demand or every day for a period of 3-6 months. Overall, EF recovery rate was 58% at 3-year follow-up. Patients treated with PDE5-I had significantly higher 3-year EF recovery rate as compared with patients left untreated after surgery (73 vs. 37%; P < 0.001). Except for BMI (P = 0.7), all preoperative covariates showed a significant association with EF recovery (all P <or= 0.04). The 3-year EF recovery rate significantly differed between the three groups, being 85, 59, and 37% in patients with low, intermediate, and high risk of postoperative ED, respectively (P < 0.001). Multivariable Cox regression analysis confirmed a highly significant association between the risk classification and EF recovery (P < 0.001). The proposed patient stratification tool showed a 69.1% accuracy. Similar results were achieved when patients were stratified according to the use of ED treatment after surgery (all P < 0.001).
CONCLUSIONS: We report the first preoperative risk stratification tool aimed at assessing the probability of EF recovery after BNSRP. It is based on routinely available baseline data such as patient age, preoperative erectile function, and comorbidity profile.