What's known on the subject? and What does the study add?
Urinary incontinence and erectile dysfunction are the most bothersome sequelae affecting health-related quality of life in patients treated with radical prostatectomy for prostate cancer. While it has been widely reported that a nerve-sparing approach significantly improves postoperative erectile function, the impact of neurovascular bundle preservation on urinary continence recovery is still a matter of controversy.
Our study clearly demonstrates that patients treated with nerve-sparing radical prostatectomy have higher chances of recovering full continence after surgery. The results indicate that, when technically and oncologically feasible, an attempt at a nerve-sparing approach should be planned in order to increase the probability of achieving full continence after radical prostatectomy.
To demonstrate that nerve-sparing radical prostatectomy (NSRP) is associated with higher rates of urinary continence (UC) recovery compared with non-nerve-sparing procedures in patients with surgically treated organ-confined prostate cancer.
Patients and Methods
The study included 1249 patients treated with radical prostatectomy between 2003 and 2010. Patients were divided into three preoperative risk groups: low (PSA < 10 ng/mL, cT1, biopsy Gleason sum ≤6), high (cT3 or biopsy Gleason 8–10 or PSA > 20 ng/mL) and intermediate (all the remaining).
Postoperative UC recovery was defined as the absence of any protection device.
The association between nerve-sparing status and UC recovery was assessed in univariable and multivariable Cox regression analyses after accounting for age at surgery, Charlson Comorbidity Index and preoperative risk group.
At a mean follow-up of 42.2 months (range 1–78), 993 patients (79.5%) recovered UC. Overall, UC recovery rate at 1 and 2 years was 76% and 79%, respectively.
On univariable Cox regression analysis, age at surgery, preoperative risk group, medical comorbidities and nerve-sparing status were significantly associated with UC recovery (all P ≤ 0.001).
On multivariable analysis, age, risk group and nerve-sparing status were also independently associated with UC recovery (all P < 0.003). Patients treated with bilateral NSRP had a 1.8-fold higher chance of full UC recovery.
Patients treated with bilateral NSRP have significantly higher chances of recovering full continence.
Therefore, when oncologically and technically feasible, a nerve-sparing procedure should be attempted.
Currently, radical prostatectomy (RP) represents the only treatment for localized prostate cancer with demonstrated benefit on cancer-specific survival compared with conservative management, as shown by a prospective randomized trial . This surgical treatment represents the most widely performed management option for patients with organ-confined prostate cancer, since a good proportion of patients are cured by surgery alone [2, 3]. Moreover, perioperative morbidity is low when this procedure is performed at tertiary care centres . However, urinary incontinence and erectile dysfunction still represent important issues for patients treated with RP . The continuous refinement of the nerve-sparing (NS) approach has experienced substantial improvements over the last few years. A major contribution in this process has also been given by the introduction of the laparoscopic and the robot-assisted approaches which have led to a better knowledge of the pelvic anatomy . However, while it is has been extensively demonstrated that the extent of neurovascular bundle (NVB) preservation is a major predictor of erectile function recovery after RP [7-9], the association between the NS approach and postoperative urinary continence (UC) recovery is still a matter of controversy. Indeed, several studies have demonstrated that an attempted NVB preservation is strongly related to UC recovery [10-13], while some authors failed to report such association [14-16]. However, the majority of these studies are limited by the low number of patients included. Moreover they did not take into account time to UC recovery which is key in the evaluation of functional outcomes after RP. In this study, we tested the hypothesis that nerve-sparing radical prostatectomy (NSRP) is associated with higher rates of UC recovery compared with non-NS procedures in a wide patient population treated with RP at a tertiary care centre.
Patients and Methods
The study included 1249 patients with prostate cancer treated with bilateral NS (BNSRP), unilateral NS (UNSRP) or non-NS retropubic RP with or without pelvic lymph node dissection at a single tertiary referral centre between January 2003 and July 2010.
All patients had complete preoperative clinical data including age at surgery, PSA at diagnosis, clinical stage, biopsy Gleason sum and medical comorbidities assessed by the Charlson Comorbidity Index (CCI). All patients were treated with RP performed by seven expert urological surgeons. NSRP was performed with the technique described by Walsh and Donker  with minor modifications. NS status at surgery, namely non-NS, UNSRP and BNSRP, was defined by each operating surgeon at the end of the procedure and reported in our prospectively collected database. Indication for NVB preservation was given according to the judgement of each treating surgeon, regardless of preoperative functional status. No patient received either neoadjuvant or adjuvant hormonal or radiation therapy. Moreover, no patient had previously undergone surgery for benign prostatic enlargement. Continence rates were assessed by the patient-reported pad usage over 24 h. Patients were followed up at 1, 3, 6 and 12 months postoperatively and every 6 months thereafter. At each visit postoperative UC recovery, defined as use of no pad, was assessed. For the purpose of the study, all patients were stratified into three preoperative risk groups based on prostate cancer characteristics at diagnosis: low (PSA < 10 ng/mL, cT1, biopsy Gleason sum ≤6; n = 602; 48.8%), high (cT3 or biopsy Gleason 8–10 or PSA > 20 ng/mL; n = 130; 10.4%) and intermediate (all the remaining patients; n = 517; 41.4%). Kaplan–Meier univariable analyses targeted time to UC recovery after surgery in the overall population as well as according to NS status and to preoperative CCI and risk group. The log-rank test was used to compare the rate of UC recovery over time among groups. The association between NS status and UC recovery was then tested in univariable and multivariable Cox regression models. Covariates consisted of patient age at surgery, CCI score and preoperative risk group.
Statistical analyses were performed with the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA), version 17, with a two-sided significance level set at P < 0.05.
Preoperative clinical characteristics of patients included in the study are shown in Table 1. At a mean follow-up of 42.2 months after surgery (median 26, range 1–78), 993 patients (79.5%) recovered complete UC. Overall, UC recovery rate at 1 and 2 years was 76% and 79%, respectively (Fig. 1). Rates of UC recovery at 1 and 2 years according to NS status were 79.5% and 84% for patients treated with BNSRP vs 62.8% and 75.9% for patients treated with UNSRP vs 44.6% and 44.6% for patients treated with non-NSRP (Fig. 2; log-rank P < 0.001). UC recovery rates at 1 and 2 years according to medical comorbidities were 77.2% and 79.2% for patients without comorbidities (CCI = 0) vs 65.6% and 67.8% for patients with CCI ≥ 1 (Fig. 3; log-rank P = 0.007). UC recovery rates at 1 and 2 years were 79.9% and 83% vs 69.9% and 75.2% vs 54.7% and 56.2% for patients in the low vs intermediate vs high preoperative risk group, respectively (Fig. 4; log-rank P < 0.001).
Table 1. Patients' characteristics and descriptive statistics
*Low risk: PSA < 10 ng/mL, cT1, biopsy Gleason sum 6 or less. High risk: cT3 or biopsy Gleason 8–10 or PSA > 20 ng/mL. Intermediate risk: all the remaining patients.
Table 2 shows univariable and multivariable Cox regression analyses predicting UC recovery after surgery. On univariable analysis, age at surgery, preoperative risk group, medical comorbidities and NS status represented significant predictors of UC recovery after RP (all P < 0.001). These results were confirmed on multivariable analysis, where NS status was independently associated with UC recovery (P = 0.002), even after accounting for all other predictors. Likewise, age at surgery and preoperative risk group were significantly associated with UC recovery after RP (all P < 0.01). After accounting for all other variables, patients treated with BNSRP had a 1.81-fold higher probability of recovering UC completely after surgery compared with patients treated with non-NSRP (P < 0.001).
Table 2. Univariable and multivariable Cox regression analyses predicting urinary continence recovery (defined as no pad) after radical prostatectomy
HR, hazard ratio; NS, nerve-sparing.
Age at surgery
Intermediate vs low
High vs low
Charlson Comorbidity Index
1 or more vs 0
Unilateral vs non-NS
Bilateral vs non-NS
The two most common functional side effects of RP are represented by urinary incontinence and erectile dysfunction. Both conditions are associated with a significant negative impact on patient quality of life . Despite several efforts in the identification of preoperative predictors of urinary incontinence [10-16], conflicting results have been reported regarding the impact of an NS approach on UC recovery. In this study, we demonstrated that the preservation of the NVB has a positive effect on UC recovery. Specifically, after adjusting for disease characteristics, patient age and comorbidities, we found that patients treated with a bilateral NS approach had 1.81-fold higher chance of recovering complete continence (P = 0.002), as defined by the use of no pads. Moreover, our analyses demonstrated that age at surgery and preoperative disease characteristics represent also independent predictors of UC recovery.
Other studies addressed the impact of an NS approach on UC recovery, with conflicting results. Burkhard et al.  performed a study on 536 patients treated with RP investigating the impact of attempted NS surgery on UC. They found that the attempt to perform a BNSRP was an important determinant of UC recovery after surgery, being the only independent predictor of UC recovery (odds ratio 4.77, P = 0.001). Takenaka et al.  performed a study investigating UC recovery after laparoscopic RP in 135 patients. They demonstrated that, after accounting for preoperative patient characteristics, an NS procedure had a significant impact on UC recovery. Similarly, Nandipati et al.  demonstrated that patients treated with a bilateral NS technique showed faster recovery of UC compared with patients treated with a non-NS approach. However, all these studies suffer from two main limitations: a relatively low number of patients enrolled and the use of a logistic regression approach, which does not take into account the time to UC recovery. This is key, since a time-dependent analysis is mandatory to adequately assess time trends of functional recovery after RP. We tried to circumvent these potential limitations by evaluating a large number of patients (n = 1249) whose outcome was assessed accounting for time to UC recovery, using both the Kaplan–Meier method and Cox regression analysis. Using this approach, we were able to demonstrate that not only age and disease characteristics but also NS status play a key role in the recovery of UC. On the other hand, several other studies failed to demonstrate an association between the NS approach and UC recovery after RP [14-16]. Marien and Lepor  reported on 1110 patients treated with NSRP addressing the relationship between potency and UC recovery. They found that while roughly 60% of patients had erectile function recovery, almost all patients (97%) recovered UC. Since the 24-month UC recovery rates did not differ between men receiving BNSRP and those treated with UNSRP as well as between potent and impotent men, the authors concluded that a NS approach is not associated with better continence. However, in that study the relationship between NS approach and UC recovery was not specifically addressed. Moreover, all patients had an attempted NS procedure, either bilateral (88%) or unilateral (12%). Therefore, the results by Marien and Lepor should be interpreted with caution since they might also be used to demonstrate the opposite concept, namely that the very high continence rates (97%) after RP could be related to the attempted NS approach performed in all cases (bilaterally in 88% of them). Similarly, Kundu et al.  found no association between an NS procedure and UC recovery after surgery. Again, despite the impressive number of patients evaluated (n = 3477), the vast majority of patients were treated with an NS technique and at least an attempt at NVB preservation was made during the procedure. This would also indirectly support a certain impact of NVB preservation on UC recovery.
Although the association between NVB preservation and UC recovery has not been definitively demonstrated, there are several anatomical studies supporting a rationale for the association between nerve preservation and continence recovery [6, 18-20]. The external urethral sphincter has been shown to be innervated not only by somatic nerve fibres from the pudendal nerves but also by autonomic nerve fibres from the pelvic plexus . An interesting study by Catarin et al. , performed on 44 patients, investigated the relationship between UC recovery and membranous urethral afferent autonomic innervation using a validated questionnaire and a series of neurophysiological tests. They demonstrated that impaired membranous urethral sensitivity was associated with urinary incontinence, particularly in patients with occasional urinary leakage. Damage to the afferent autonomic innervation may thus have a role in the continence mechanism after NSRP . Moreover, it has been reported that RP decreases membranous urethral microcirculation . In patients undergoing NSRP, sparing the NVB may contribute to the preservation of membranous urethral vascular integrity. Therefore, we speculate that an impairment in membranous urethral blood flow in patients treated with non-NSRP may lead to damage to the external urinary sphincter, with a reduction of UC recovery rates. However, all these hypotheses need further more extensive demonstration.
The implications of our findings are several. Since the NS technique has an impact on UC recovery, many patients who would not be considered ideal candidates for NVB preservation due to a deteriorated erectile function already prior to surgery should still be considered suitable for an NS approach if technically and oncologically feasible. For the same reason, also older patients who might not be interested in erectile function might be considered candidates for an NS technique, given that age itself represents a risk factor for urinary incontinence. In other words, preoperative functional status and age at surgery should not be considered as deterrents to performing an accurate NVB preservation, if technically and oncologically feasible. Finally, even in those patients with less favourable prostate cancer characteristics at diagnosis, sparing the NVB at least on one side can be attempted. The risk for side-specific extra-prostatic extension can indeed be calculated and the NVB can be spared based on the individual risk of extra-prostatic extension on that side .
Despite several strengths, our study has some limitations. First, the number of patients who received a monolateral NS approach is limited and therefore the results obtained in this patient category must be considered with caution. Second, preoperative CCI did not achieve independent predictor status in multivariable models, although it was significantly associated with urinary incontinence on univariable analysis. However, only few patients had a severe CCI profile (≥2) and thus the real effect of a severe comorbidity profile could not be correctly evaluated. Third, all cases were performed by seven different surgeons and there might have been some differences in the NS technique among them despite the use of a standardized approach. However, this limitation might also be interpreted as a point of strength, since the NS approach was a predictor of UC despite potential differences in each individual surgeon's technique. Finally, the NS definition was given by the treating surgeon at the end of the operation and this may introduce a potential bias in the analyses. However, this limitation applies to all the studies addressing this issue. Moreover, subjective surgeon interpretation of NS quality during RP has been shown to correlate with functional outcome postoperatively . Despite all these limitations, our study represents one of the largest single-centre series addressing the association between NS status and UC recovery after RP using a proper methodological approach.
Thus we have demonstrated that patients treated with BNSRP have higher chances of recovering full continence, as defined by the absence of any protection device. Other important associated factors are represented by preoperative risk group and age at surgery. Therefore, when technically and oncologically feasible, an attempt at an NS approach should always be planned in order to increase the probability of achieving full continence after RP.