Open Retropubic Nerve-Sparing Radical Prostatectomy

european urology 49 (2006) 38 48 available at www.sciencedirect.com journal homepage: www.europeanurology.com Surgery in Motion Open Retropubic Nerve-Sparing Radical Prostatectomy Markus Graefen a,b, *, Jochen Walz a, Hartwig Huland a,b a Department of Urology, University Medical Center Eppendorf, Martinistr. 52, 20246 Hamburg, Germany b Martini-Clinic, Prostate Cancer Center, University Medical Center Eppendorf, Martinistr. 52, 20246 Hamburg, Germany Article info Article history: Accepted October 17, 2005 Published online ahead of print on November 18, 2005 Keywords: Prostate cancer Radical prostatectomy Nerve-sparing Abstract Retropubic radical prostatectomy is the most commonly used therapeutic option for the treatment of clinically localized prostate cancer. An ongoing stage migration toward organ-confined cancers allows performance of a nerve-sparing procedure in a growing number of patients. Key elements for achieving convincing functional results are a sphincter-preserving ligation of the distal part of the Santorini plexus and the subtle preparation of the neurovascular bundle. This article gives a detailed description of the operative technique, which is demonstrated in the attached DVD. Furthermore, indication, oncologic outcome, and functional results addressing postoperative urinary continence and potency are discussed. # 2005 Elsevier B.V. All rights reserved. * Corresponding author. Tel. +49 0 4042803 1313; Fax: +49 0 4042803 1323. E-mail address: graefen@uke.uni-hamburg.de (M. Graefen). 1. Introduction Open retropubic radical prostatectomy (RP) is the most commonly used therapeutic option for the treatment of clinically localized prostate cancer. Numerous articles have been published demonstrating excellent tumor control and functional results in addition to a low morbidity of the procedure [1 3]. A nerve-sparing modification of the procedure has become standard practice because an ongoing stage migration toward early detected organ-confined cancers allows this technique in a growing number of patients without compromising cancer control [4 7]. Furthermore, as more cancers are detected at a younger age patients demand a high level of functional outcome after therapy. We report on our operative technique, long-term cancer control rates, and functional outcome of contemporary nerve-sparing radical prostatectomy (nsrp). 2. Operative technique 2.1. General recommendations Open retropubic RP is the most commonly used operative technique to treat prostate cancer. Numerous refinements were recently implemented making this procedure into a minimal invasive operation when commonly used definitions for minimal invasiveness are used [8]. With the use of an autoretractor system an optimal and standar- 0302-2838/$ see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2005.10.008

european urology 49 (2006) 38 48 39 dized exposure of the prostate and lymphatics can be achieved. Furthermore, open RP is cost effectiveness because no disposable material is used and the procedure can be performed by 2 surgeons alone. It is helpful to use titanium clips with a rectangular applicator in different sizes and as sewing material Monofil resorbable threads with a 5/8 circumferential needle. We strongly recommend the use of magnification glasses with a 2.5- up to 3.5-fold magnification and a xenon headlight to combine the potential advantages of laparoscopic RP (which are magnification and optimal light) with the advantages of open surgery (which are tactile sensation and 3-dimensional view). Furthermore, we recommend performing nsrp with spinal anesthesia with additional total intravenous anesthesia (TIVA), if necessary, which allows a quick recovery for the patient. In addition, restriction of intraoperative infusion (maximum 500 1000 ml crystalloid fluids) until the prostate is removed substantially reduces blood loss. An additional means to reduce blood loss is to have the patient in the Trendelenburg position to lower the venous pressure of the operation site. In our series of 678 RPs performed in 2004 mean blood loss was 540 cc and transfusion rate was reduced to <1%. 2.2. Surgical technique of open nsrp The patient is placed in a prone position with slightly overstretched abdomen and pelvis. A Foley catheter is placed, which stays in the sterile operation site for further manipulation. Following an 8 10-cm median incision the cavum retzii can bluntly be exposed with a sponge stick. Adjacent fat tissue is removed to expose the endopelvic fascia. The fascia is then incised and muscle fibers of the levator ani are removed blunt from the prostate using either a small sponge stick or blunt scissors (Fig. 1a, b). The puboprostatic ligaments are bluntly exposed and incised close to the symphysis with care taken on the veins of the adjacent Santorini plexus. The apex can precisely be identified when all muscle fibers of the levator urethrae (the most distal part of the levator muscle at the level of the urethra) and the levator ani are bluntly removed from the prostate and the anterolateral aspects of the urethra. It is important that the fixation of the Fig. 1 (a + b) The right endopelvic fascia is incised and muscle fibers of the levator ani are removed blunt from the lateral prostate side using either a small sponge stick or blunt scissors.

40 european urology 49 (2006) 38 48 less traction to the prostate. To keep the operation field clear of blood continuous suction is used. The divided dorsal vein plexus is then distally selectively oversewn with the already placed thread from 10 o clock to 2 o clock without touching the fibers of the striated external sphincter (Fig. 2). This is possible because these veins run between 2 membranes (Fig. 3). The upper membrane is the continuation of the endopelvic fascia, whereas the lower membrane is the fascia of the striated external sphincter. It is important to identify and include these 2 whitish layers that cover the dorsal vein complex on its ventral and dorsal aspect. This guarantees a good control of bleeding and prevents accidently incorporating muscle fibers in the suture. By this procedure almost all fibers of the striated sphincter can be preserved. The proximal part of the plexus is then oversewn on the apex of the prostate to achieve complete hemostasis for the following part of the operation. This should be done without giving any tension on the neurovascular bundle. 2.3. The nerve-sparing procedure Fig. 2 The unsutured divided dorsal vein plexus is then distally selectively oversawn with the already placed thread from 10 to 2 o clock without touching the fibers of the striated external sphincter. levator muscle at the ventral part of the urethra is not disrupted. Usually coagulation is not necessary during this step. If bleeding occurs from the pelvic wall, a bipolar forceps can be used for coagulation. However, no coagulation should be used close to the neurovascular bundle or on the prostate surface to prevent injuries of the nerves. The dorsal vein complex is oversewn on the prostate close to the bladder to prevent backbleeding. A superficial stay suture (3/0 Prolene with an UR6 needle) is set distal from the prostate on the right lateral aspect of the Santorini plexus where the urethra enters into the prostate. Special care must be taken to stay superficial without taking any fibers of the sphincter and the urethra. The thread will be used to oversew the plexus after its incision, so the branches are set on small clamps. The dorsal vein plexus is then divided at the apex of the prostate without any ligation starting with a knife followed by sharp scissors until the muscle fibers and the fascia of the external sphincter are visible. If bleeding occurs, it can usually be reduced by giving The parapelvic fascia on the prostate is incised at the lateroventral aspect of the prostate at 10 o clock and 2 o clock, which is about 2 cm away from the neurovascular bundles (Fig. 4a). It is important to start the incision high up on the ventral aspect of the prostate to preserve a maximum number of nerve fibers. A substantial number of fibers are located at the ventral area and it has been shown that potency rates strongly correlate with the number of preserved nerve fibers (i.e., bilateral vs. unilateral). After a small incision of the parapelvic fascia covering the prostate it will be undermined using a small Overholt clamp for careful dissection (Fig. 4b). Underneath the fascia an areolar space containing fat, connective tissue, veins, and small tethering vessels can be identified (Fig. 5a c). These veins need to be undermined and finally aid in identifying the right plane for dissection. The fascia and the vessels will then be clipped with the rectangular applicator and divided (we prefer 5-mm titanium clips) throughout the whole ventrolateral aspect of the prostate (Fig. 4c, d). The neurovascular bundle is then gently pushed laterally and downward using a peanut sponge stick or blunt scissors. If the right plane is achieved, the surface of the prostate should be completely smooth and reflecting. It is crucial not to use any coagulation or ultrasound dissector during this step because this will lead to nerve damage. If minor venous bleeding occurs we would not perform hemostasis to avoid any nerve damage; if arterial bleeding occurs, this should be controlled

european urology 49 (2006) 38 48 41 Fig. 3 Transverse section of the membranous urethra. with clips or directed selective stitches using 5/0 Prolene. Close to the apex the dissection of the parapelvic fascia is performed at the upper third of the urethra. The direction of incision is antegrade down to the 10 o clock and 2 o clock position of the urethra. The neurovascular bundle is separated from the Müller ligament, which runs laterally from the striated external sphincter and represents a continuation of the ventral fascia of the striated sphincter. The neurovascular bundle will gently be pushed away from the apex and the proximal 1.0 0.5 cm of the urethra. Following this maneuver the neurovascular bundles are lying beside the prostate and the proximal urethra and should not be touched during the rest of the procedure (Fig. 6). There are two advantages for separating the neurovascular bundles early from the proximal urethra. 1. It allows a precise identification of the apex. 2. The neurovascular bundles are a safe distance from the urethra. Prior to the dissection of the circular fibers of the striated muscle of the membranous urethra the muscle fibers covering the apex can bluntly pushed toward the urethra to preserve as much functional tissue as possible. The striated external sphincter runs circular and its fibers are attached at the outside of the apex (Fig. 7a). Once they are pushed away (Fig. 7b) the longitudinal smooth muscle is visible that runs into the prostate. After this maneuver the distal tip of the apex is precisely visible and the i- ncision of the urethra can be done exactly at the end of the apex, preserving the whole length of the m- embranous urethra. The urethra is cut from 10 o - clock to 2 o clock down to the inlaying Foley catheter. Four anastomotic sutures are placed (we use 3 0 PDS and a UR-6 needle) at 9, 11, 1, and 3 o clock. It is important to include only a small aspect of smooth muscle and urethral mucosa and to use the ligated dorsal vein complex to anchor the stitch to give stability to the anastomosis. This guarantees a reliable anastomosis with a minimum of functional tissue incorporated. After retraction of the catheter further anastomotic sutures are placed at 5, 6, and 7 o clock. Care has to be taken that the neurovascular bundles are not included in these sutures. The smooth muscle at the dorsal circumference of the urethra is then dissected. Fibers from the striated sphincter muscle are bluntly removed from the ter-

42 european urology 49 (2006) 38 48

european urology 49 (2006) 38 48 43 Fig. 5 (a and b) An areolar space containing fat, connective tissue, veins, and small tethering vessels lies between the parapelvic fascia and the prostate capsule. In this space there are a mean of 30 nerve fibers on each side. (c) Whole mount section of the prostate demonstrating the areolar space between prostate capsule and parapelvic fascia. minal aspect of Denonvilliers fascia. The prostate is gently lifted up by the reinserted catheter and Denonvilliers fascia is incised. The fascia remains on the prostate after a bilateral V-shape incision is made about 2 cm cranial of the urethra at a safe distance to the preserved neurovascular bundles. The tip of the V is directed to the urethra. The prostate is gently m- obilized from the apex up to the bladder. Vessels or arterial bleedings are clipped or ligated and no coagulation is used. The upper pedicles of the prostate (pillars) are identified. For this we find it helpful to divide the overlying perivesicle fat for two reasons: 1. It allows visualization of the exact border of the pedicle to the base of the prostate to ensure a precise dissection 1 cm lateral of the prostate base. 2. It releases the neurovascular bundle at this level, which virtually falls toward the rectum. With this technique special care can be taken not to injure the neurovascular bundle or to put it under traction. At the seminal vesicles the ventral layer of Denonvilliers fascia is incised and left in situ to protect the neurovascular bundle at its medial Fig. 4 (a) The parapelvic fascia on the prostate is incised on the latero-ventral aspect of the prostate at 10 o clock and 2 o clock in safe distance to the neuro-vascular bundle. (b) After a small incision of the parapelvic fascia, the underlying areolar tissue and veins covering the prostate are undermined using a small overholt clamp for careful dissection. (c) The fascia and the vessels will then be clipped with the rectangular applicator (we prefer 5-mm titanium clips) and divided throughout the whole ventrolateral aspect of the prostate.

44 european urology 49 (2006) 38 48 If a nerve-sparing procedure is performed and intraoperative suspicion of extracapsular tumor growth arises, intraoperative frozen section is recommended. After the prostate is removed a slice from the lateral surface of the prostate should be taken and the area of the prostate capsule that was adjacent to the neurovascular bundle should be i- nked. In addition, the area from which the slice was taken should be inked in a different color for later differentiation of true surgical margin and margin of intraoperative frozen section. The slice should be taken from the apex to the base of the prostate. The frozen section is performed perpendicularly and we recommend removing the neurovascular bundle when cancer reaches the inked surface (Fig. 8). Even thorough preoperative and intraoperative tumor selection is no guarantee for negative surgical margins. However, the prognostic impact of small positive margins undetected during frozen section but recognized in the final pathology work-up remains unclear and may be negligible at least in organ-confined tumors as we have intensively discussed in a recent study [6]. 3. Removal of the prostate and urethral anastomosis Fig. 6 The neurovascular bundles are separated from Müller s ligament (not seen on this graphic) which runs laterally from the striated external sphincter and represents a continuation off the ventral fascia of the striated sphincter (see also Fig. 3). The neurovascular bundles are gently pushed away from the apex and the proximal urethra. Following this maneuver the neurovascular bundles are lying beside the prostate and the proximal urethra and should not be touched during the rest of the procedure. aspect. The tips of the seminal vesicles are identified, and the adjacent vessels are clipped and dissected. The vas deferens is clipped and dissected. Again, no coagulation should be used to preserve the integrity of the nerve bundles that run in close relationship to the tips of the seminal vesicles. There are several reasons for difficulties separating the peripelvic fascia and the areolar tissue from the capsule during nerve preservation: - acute or chronic prostatitis - abandoned veins or crossing vessels - tumor infiltrating the neurovascular bundle The bladder is opened cranial of the prostate and the detrusor is dissected next to the gland down to the trigonum. The prostate is freed from remaining adjacent tissue and taken out. Selective stitches or 5- mm clips can now be used if significant venous or arterial bleeding at the neurovascular bundles occurs. The orifices of the ureters are identified and if necessary elongated from the edge of the bladder neck. The mucosa of the bladder is everted using a 4 0 Vicryl suture and the bladder outlet is narrowed using a tennis racket technique. Hemostasis is verified, and if bleeding occurs on or next to the neurovascular bundle it should be controlled with clips or selective stitches. The anastomotic sutures are placed (Fig. 9) and the anastomosis is tied after the transurethral catheter is inserted. We recommend filling the bladder with saline to recognize eventual leakage. We routinely place one drainage catheter; a suprapubic catheter is not necessary. An intracutaneous suture of the incision gives a good cosmetic result. 4. Indication, cancer control, and functional outcome after open retropubic nsrp 4.1. Indication for a nerve-sparing procedure Tumor selection is a crucial step for indicating nsrp. In organ-confined cancers an nsrp can be performed

european urology 49 (2006) 38 48 45 Fig. 7 (a and b) The striated external sphincter runs circular and its fibers are attached at the outside of the apex (Fig. 7a). Once they are pushed away (Fig. 7b) the longitudinal smooth muscle is visible that runs into the prostate. After this maneuver the distal tip of the apex is precisely visible and the incision of the urethra can be done exactly at the end of the apex, preserving the whole length of the membranous urethra. without compromising tumor control by producing artificial positive margins [6,7]. However, capsular penetration of the cancer harbors a high likelihood of neoplastic cells at the surgical margin when nerve preservation is performed. To decide whether or not and to what extend a nerve-sparing procedure should be performed (unilateral or bilateral) a Classification and Regression Tree structures model (CART analysis) was developed calculating the likelihood of side-specific organ confinement of the diagnosed cancer [9]. This algorithm represents a flexible tool that contributes to the individual tumor extent and tumor localization of each prostate lobe. In low-risk patients, characterized by not more than one positive biopsy with high-grade cancer and a prostate-specific antigen (PSA) level below 10 ng/ml, the likelihood of organ confinement is almost 90%. Reliability of this predictive tool was confirmed in an internal prospective validation study, yet an external validation is missing [9]. Ohori et al. have recently reported an alternative approach, where predictions are generated with a logistic regression-based nomogram, with predictive accuracy of 0.806 [10]. In a recent study we compared predictive accuracy of both models on a patient cohort consisting of 1117 men with clinically localized prostate cancer. Predictive accuracy of the side-specific nomogram by Ohori et al. was 0.84 versus 0.70 for the treeregression model of Graefen et al. [11]. This demonstrates the general superiority of continuously predicting model compared to a risk-group stratification scheme. However, both statistical tools have demonstrated a high reliability and we recommend their use for a reproducible indication of nsrp. Furthermore, patient counseling before therapy is more precise when the likelihood of unilateral or bilateral nsrp and the corresponding potency and continence rates are known prior to surgery. A tremendous stage migration toward organ-confined cancers and the reliability of preoperative predictive tools have led to continuous increase of nervesparing procedures in recent years (Fig. 10). Using such predictive tools in addition to intraoperative frozen section, the cancer control rate is not compromised by a nerve-sparing procedure [6]. 4.2. Cancer control Effective cancer control remains the most important goal of RP. Efficacy of RP is well documented by several authors. In our own series total cancer volume and high-grade cancer volume were investigated in addition to traditional histologic features such as pathologic stage and margin status. Biochemical recurrence-free survival rates 5 years after surgery (PSA recurrence defined as a PSA > 0.1 ng/ ml) were 93%, 61%, and 32% for organ-confined cancers, capsular penetration, and seminal vesicle involvement, respectively. Cancer volume and highgrade cancer volume proved to be the most important predictors of recurrence in our series [6].

46 european urology 49 (2006) 38 48 Fig. 9 The mucosa of the bladder is everted using a 4 0 Vicryl suture and the bladder outlet is narrowed using a tennis racket technique. The anastomotic sutures are placed and the anastomosis is tied after the transurethral catheter is inserted. Fig. 8 If extra-capsular extension is suspected a slice from the lateral surface of the prostate should be taken after the prostate is removed (a) and the area of the prostate capsule that was adjacent to the neuro-vascular bundle should be inked. In addition, the area from which the slice was taken should be inked in a different color for later differentiation of true surgical margin and margin of intraoperative frozen section (b). 4.3. Urinary continence after nsrp Numerous published data address urinary continence after RP. In our own series a nerve-sparing The important question whether a nerve-sparing procedure compromises cancer control was investigated with a multivariate analysis treating the nerve-sparing procedure as a potential risk factor for recurrence. In this analysis, however, nerve-sparing itself in carefully selected patients was not an adverse factor for recurrence, demonstrating that this is a safe procedure [6]. Kaplan-Meier curves of recurrence-free survival stratified by pathologic stage and nerve-sparing procedure are shown in Fig. 11. Fig. 10 Number of radical prostatectomies (RRP), number of organ-confined cancers (pt2), and number of nerve-sparing procedures (NE) performed at the University Hospital Hamburg Eppendorf between 1/1992 and 12/ 2004.

european urology 49 (2006) 38 48 47 Table 1 Postoperative potency rates after nerve-sparing retropubic radical prostatectomy without the use of medical aids (n = 524 preoperatively potent patients (IIEF-5 score >19) at the University Hospital Hamburg, Follow-up >12 months) Nerve-Sparing Age < 55 years Age 55 65 years Age > 65 years Erections Intercourse Erections Intercourse Erections Intercourse Unilateral 73.3% 36.7% 75.9% 20.7% 63.9% 18.0% Bilateral 96.5% 69.0% 90.7% 52.8% 84.3% 37.3% procedure was a protective factor for continence [12]. Ninety-two percent of patients stated they were without any loss of urine following nsrp in a validated questionnaire 1 year after surgery. In contrast, 83% of men following non-nsrp stated they were without any loss of urine at the same time, demonstrating the superiority of nerve-sparing in regard to postoperative continence. In a multivariate analysis including the patient s age and pathologic features, the nerve-sparing procedure had an independent protective association of posttherapeutic continence. In our institution these observations led to the recommendation for a nerve-sparing procedure in early cancers even in men with preoperative erectile dysfunction. 4.4. Potency following RP Data regarding postoperative potency were obtained at a median of 13 months after RP. About 90% of preoperatively potent men undergoing bilateral nerve-sparing and 76% of patients after unilateral nerve-sparing RP report on some erectile function after the procedure. Fifty-six percent of men undergoing bilateral nerve-sparing and 30% of patients after unilateral nsrp were able to have intercourse without the use of any medical aid [13]. In those men who had some erectile function but who were not able to perform spontaneous intercourse the majority could perform intercourse with the use of phosphodiesterase 5 (PDE5) inhibitors. These results were reproducible over several time intervals and are not based on a single-surgeon series but on a series in which several surgeons participated. This underlines the importance of a standardized technique for the nerve-sparing procedure. However, we found that even using a standardized technique a learning curve exists, giving better functional results for the more experienced surgeon [13]. As expected, we could found a strong association of postoperative potency rates, unilateral or bilateral nerve-sparing, and age (Table 1). Patients after bilateral retropubic prostatectomy younger than 55 years or older than 65 years could have unassisted intercourse in 69% and in 37.3%, respectively. This correlation, dependent on patient s age, was also found following a unilateral nerve-sparing procedure [13]. As a result of an ongoing stage migration toward organ-confined cancers, we can offer a nerve-sparing procedure now to about 90% of our patients (Fig. 10). 5. Summary Contemporary nerve-sparing retropubic RP is still the most commonly performed treatment for localized prostate cancer. Refinements of the technique especially for a nerve-sparing procedure allow excellent cancer control in combination with good and reproducible functional results. Fig. 11 Kaplan-Meier analysis of recurrence free survival of 1755 patients operated at the University-Hospital Hamburg, Germany, stratified by pathologic stage and nerve sparing procedure. NS = nerve-sparing radical prostatectomy, non-ns = non-nerve-sparing radical prostatectomy. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/

48 european urology 49 (2006) 38 48 j.eururo.2005.10.008 and via www.europeanurology. com. Subscribers to the printed journal will find the supplementary data attached (DVD). References [1] Hull GW, Rabbani F, Abbas F, Wheeler TM, Kattan MW, Scardino PT. Cancer control with radical prostatectomy alone in 1,000 consecutive patients. J Urol 2002;167(2 pt 1):528 34. [2] Michl U, Graefen M, Noldus J, Eggert T, Huland H. Functional results of various surgical techniques for radical prostatectomy. Urologe A 2003;42(9):1196 2002. [3] Augustin H, Hammerer P, Graefen M, Palisaar J, Noldus J, Fernandez S, et al. Intraoperative and perioperative morbidity of contemporary radical retropubic prostatectomy in a consecutive series of 1243 patients: results of a single center between 1999 and 2002. Eur Urol 2003;43(2): 113 8. [4] Derweesh IH, Kupelian PA, Zippe C, Levin HS, Brainard J, Magi-Galluzzi C, et al. Continuing trends in pathological stage migration in radical prostatectomy specimens. Urol Oncol 2004;22(4):300 6. [5] Noldus J, Graefen M, Haese A, Henke RP, Hammerer P, Huland H. Stage migration in clinically localized prostate cancer. Eur Urol 2000;38(1):74 8. [6] Palisaar RJ, Noldus J, Graefen M, Erbersdobler A, Haese A, Huland H. Influence of nerve-sparing (NS) procedure during radical prostatectomy (RP) on margin status and biochemical failure. Eur Urol 2005;47(2):176 84. [7] Ward JF, Zincke H, Bergstralh EJ, et al. The impact of surgical approach (nerve bundle preservation versus wide local excision) on surgical margins and biochemical recurrence following radical prostatectomy. J Urol 2004;172(4 pt 1): 1328 32. [8] Fornara P, Zacharias M. Minimal invasiveness of laparoscopic radical prostatectomy: reality or dream? Aktuelle Urol 2004;35(5):395 405. [9] Graefen M, Haese A, Pichlmeier U, Hammerer PG, Noldus J, Butz K, et al. A validated strategy for side specific prediction of organ confined prostate cancer: a tool to select for nerve sparing radical prostatectomy. J Urol 2001; 165(3):857 63. [10] Ohori M, Kattan MW, Koh H, Maru N, Slawin KM, et al. Predicting the presence and side of extracapsular extension: a nomogram for staging prostate cancer. J Urol 2004;171(5):1844 9. [11] Steuber T, Graefen M, Perotte P, Haese A, Chun FKH, Huland H, et al. Prediction of side specific extracapsular extension at radical prostatectomy in European patients: accuracy of a novel, internally validated logistic regression based nomogram vs. tree structured regression analysis. Eur Urol 2005, Suppl. In press. [12] Michl U, Graefen M, Haese A, Hammerer P, Huland H. Prospective analysis of continence and micturition following nerve sparing and non nerve sparing radical retropubic prostatectomy. Significant impact of the nerve sparing procedure on continence. J Urol 2001;165(suppl). [13] Michl U, Friedrich MG, Graefen M, Haese A, Heinzer H, Huland H. Prediction of postoperative sexual function following nerve sparing radical retropubic prostatectomy. J Urol. In press.