Predictors of Recovery of Erectile Function after Unilateral Cavernous Nerve Graft Reconstruction at Radical Retropubic Prostatectomy

Predictors of Recovery of Erectile Function afterUnilateral Cavernous Nerve Graft Reconstruction atRadical Retropubic Prostatectomyjsm_1436 166..181

Farhang Rabbani, MD,* Ranjith Ramasamy, MD,* Manish I. Patel, MBBS,† Paul Cozzi, MBBS,‡

Joseph J. Disa, MD,§ Peter G. Cordeiro, MD,§ Babak J. Mehrara, MD,§ James A. Eastham, MD,*Peter T. Scardino, MD,* and John P. Mulhall, MD*

*Urology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; †Department of Surgery, University ofSydney, Sydney, Australia; ‡Department of Surgery, University of New South Wales, Sydney, Australia; §Plastic SurgeryService, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

DOI: 10.1111/j.1743-6109.2009.01436.x

A B S T R A C T

Introduction. Although studies have reported a benefit to bilateral cavernous nerve graft (NG) interposition, therole of unilateral NG interposition in recovery of erectile function (EF) after radical prostatectomy (RP) withunilateral neurovascular bundle (NVB) resection is more controversial.Aim. To determine the probability and predictors of EF recovery after unilateral cavernous NG at RP withunilateral NVB resection.Methods. We retrospectively reviewed the records of preoperatively potent men who underwent RP with unilateralNVB resection and ipsilateral NG without prior radiation or hormonal therapy from 1999 to 2007. Postoperative EFwas defined in two ways: (i) physician interview-based assessment (level 3: erections sometimes sufficient forintercourse; level 2: erections routinely sufficient for intercourse; level 1: normal erections; all with or without oralphosphodiesterase-5 inhibitor use); and (ii) according to the sum Q3 + Q4 on the International Index of ErectileFunction (IIEF) questionnaire.Main Outcome Measures. EF recovery based on physician interview-based scale and IIEF questionnaire.Results. In all, 131 men underwent unilateral NG. Median follow-up was 37.3 months. The 5-year actuarialprobability of EF recovery was 46, 30, and 12% for levels 3, 2, and 1, respectively, and 40, 34, and 22% for IIEFQ3 + Q4 sum �6, �8, and = 10, respectively. On multivariate analysis, patient age, specimen weight, and plasticsurgeon were predictive of EF recovery based on physician-interview whereas patient age, ethnicity, and plasticsurgeon were predictive of EF recovery based on the IIEF questionnaire.Conclusions. The impact of plastic surgeon on EF recovery with unilateral NG would suggest that technical factorsplay a role in EF recovery after unilateral NG. Meticulous surgical technique with proper identification of proximaland distal recipient nerve endings may improve the chance of EF recovery. The variation in recovery rate amongplastic surgeons would imply that there is a benefit to unilateral NG in EF recovery. Rabbani F, Ramasamy R, PatelMI, Cozzi P, Disa JJ, Cordeiro PG, Mehrara BJ, Eastham JA, Scardino PT, and Mulhall JP. Predictors ofrecovery of erectile function after unilateral cavernous nerve graft reconstruction at radical retropubicprostatectomy. J Sex Med 2010;7:166–181.

Key Words. Cavernous Nerve; Nerve Graft; Erectile Dysfunction; Prostatectomy

Introduction

E rectile dysfunction (ED) remains a commonconcern for patients following radical pros-

tatectomy (RP) [1]. After Walsh and Donker’s

description of cavernous nerve anatomy [2], andthe subsequent development of the anatomical RP[3], potency rates of 20–95% have been observedwhen both cavernous nerves have been preserved[3–6]. Recovery of erectile function (EF) is directly

166

J Sex Med 2010;7:166–181 © 2009 International Society for Sexual Medicine

correlated with the extent and number of neu-rovascular bundles (NVBs) preserved [4–6]. Kimet al. [7] have reported no spontaneous erectionsin a group of patients with wide bilateral cavernousnerve resection, and rates of 25–53% have beenreported when only one nerve is preserved [3–6].

The predominant cause of impaired EF follow-ing RP is neurogenic in origin. Quinlan [8] andBall [9] have shown a significant improvement inEF in rat models of cavernous nerve transection,where reconstruction was performed with graftedgenitofemoral nerves. More recently, Lowe et al.have developed a large-animal model in dogs forsurgically induced loss of EF with successful cav-ernous nerve graft (NG) reconstruction with thesural nerve [10]. Results of early animal studieshave been confirmed by human studies at theBaylor College of Medicine [11] and MD Ander-son Cancer Center [12], where men undergoingbilateral wide resection of cavernous nerves, andreconstruction with sural NGs had some recoveryof EF compared with a control group who did notrecover EF.

Since the initial report of bilateral cavernousnerve reconstruction with sural NG [13], there hasbeen extensive debate in the literature regardingthe efficacy of NG [14–16]. Further follow-up onthe initial pilot study [11], as well as other series[12,17], have reported EF recovery rates of34–43% with bilateral cavernous NG, comparedwith historically observed rates of <5% withoutNG. Although efficacy of bilateral NG can bemore easily determined, determining that of uni-lateral NG at the time of RP with unilateral NVBis more challenging given that the observed EFrecovery may be solely because of the preservationof the contralateral NVB. The only randomizedtrial [16] of RP and unilateral NVB resection withor without unilateral sural NG did not identify a20% improvement with NG over a presumed 40%2-year EF recovery rate without NG; however,this trial was somewhat underpowered as there wasa higher than anticipated study dropout and it wasterminated early with 107 patients enrolled, afteran interim analysis met criteria for futility despitethe target accrual of 200 patients required for thestudy to have sufficient power. Further contro-versy surrounds the ideal choice for donor nerve.Sural nerves [13,18–21] and more recently, gen-itofemoral nerves [22–24] have been used forcavernous NG interposition.

Despite, the success of injection therapy,vacuum constriction devices and penile prosthesisin the treatment of post-prostatectomy ED, the

importance of return of spontaneous erectile activ-ity, with or without the assistance of oral medi-cation, cannot be overstated. We report ourexperience with unilateral cavernous nerve recon-struction at the time of RP with unilateral wideNVB resection in patients who had functionalerections preoperatively. We sought to iden-tify predictors of EF recovery in this patientpopulation.

Methods

DefinitionsPreoperative EF was assessed prospectively by adetailed preoperative interview by the surgeon,being judged on a 5-point scale as describedpreviously [5]: (i) full erections; (ii) diminishederections but routinely sufficient for sexualintercourse; (iii) partial erections occasionally sat-isfactory for intercourse; (iv) partial erectionsunsatisfactory for intercourse; and (v) no erec-tions. Although this scale has not been validated,the good correlation of this scale with the Inter-national Index of Erectile Function (IIEF) ques-tionnaire [25] has been previously reported [17].Postoperative EF was determined by physicianinterview as well as administration of the IIEFquestionnaire, specifically, questions 3 (Q3) and 4(Q4) in a subset of patients. The date of recoveryof a given category of EF was recorded as the dateit was first achieved. The degree of NVB preser-vation was recorded by the surgeon immediatelypostoperatively and graded based on close inspec-tion of the prostate to determine the grossamount of periprostatic tissue, especially at theposterolateral aspects: (i) intact; (ii) possibledamage, but grossly intact; (iii) definite damagewith partial resection; and (iv) complete resection[5]. Grades 1 and 2 were considered to be nerve-sparing (NS).

Patient PopulationAfter institutional review board approval wasobtained, our prospective database was reviewed toidentify patients undergoing RP and unilateralNVB resection with unilateral NG betweenJanuary 1999 and June 2007. Inclusion criteriaincluded: (i) biopsy-established diagnosis of pros-tate cancer; (ii) RP with unilateral NVB resection(NVB preservation grade of 1 or 2 on one side and4 on the contralateral side) with unilateral NG;and (iii) documentation of preoperative erectionssatisfactory for intercourse (level 1 or 2). RP with

Potency after Prostatectomy with Unilateral Nerve Graft 167

J Sex Med 2010;7:166–181

unilateral NVB resection and unilateral NG wasperformed in 175 men between January 1999 andJune 2007. Exclusion criteria included: (i) pre-operative erections not routinely satisfactory forunassisted intercourse or better (level 3, 4 or 5)(N = 21); and (ii) previous treatment with radio-therapy (N = 13), androgen deprivation (N = 16),or chemotherapy (N = 6); in this, 44 men wereexcluded. The remaining 131 men comprise thestudy population.

Data collected included patient age at surgery,American Society of Anesthesiologists’ classifica-tion, body mass index (BMI), ethnicity, biopsyGleason score, preoperative prostate-specificantigen (PSA), clinical and pathologic stageaccording to the 2002 tumor–node–metastasisclassification, specimen weight, NG type, urolo-gist and plastic surgeon, estimated blood loss,adjuvant treatment, and comorbidities includinghypertension, hypercholesterolemia, coronaryartery disease, and diabetes mellitus.

Operative TechniqueAll patients had RP and bilateral pelvic lym-phadenectomy performed by one of eight experi-enced surgeons by the method of Scardino [26].The extent of preservation of each NVB wasplanned by the surgeon preoperatively based onthe results of the digital rectal exam, prostateimaging studies, serum PSA, grade, location, andextent of cancer in each biopsy core, and the pres-ence of perineural invasion. A final intraoperativedecision was made depending on the presence ofprostatic induration or fixation of the NVB to theprostate capsule after division of the endopelvicfascia. Unilateral NG was recommended inpatients undergoing unilateral NVB resection whowere capable of achieving and maintaining spon-taneous erections sufficient for vaginal penetra-tion. Cavernous nerves were identified with the aidof the Cavermap™ Surgical Aid (Blue TorchMedical Technologies, Ashland, MA, USA) [27].Confirmation of nerve preservation or resectionwas obtained by the response to Cavermap™stimulation [27].

The NG technique was similar to thatdescribed by Kim et al. [13]. The decision regard-ing which nerve to use (sural vs. genitofemoral)was based on plastic surgeon preference as well ascaliber of the genitofemoral nerve. If the gen-itofemoral nerve was of insufficient caliber, thesural nerve was harvested. All NGs except for sixcases were done by one of three plastic surgeons. A

single donor nerve was used in all patients excepttwo where a double-stranded NG was performed.

EF AssessmentFollow-up visits occurred every 3 months for thefirst year, every 6 months until year 5 and annuallythereafter. At each visit, recovery of EF was evalu-ated by surgeon interview according to the 5-pointscale as described earlier. The date of recovery of agiven EF category was recorded as the date that EFof this quality was first achieved. The outcome endpoints were recovery of level 3 or better erections,level 2 or better erections, and level 1 erections.On-demand use of phosphodiesterase type 5(PDE5) inhibitors as desired by patients wasallowed after RP. Patients were considered to haveED if they required a vacuum erection device,intracavernosal injections or intraurethral therapyfor EF, or if they had received a penile prosthesis. Acohort of patients also completed the IIEF ques-tionnaire. Outcome end points for assessment byIIEF questionnaire were a score of �6, �8, or 10for the sum questions 3 (Q3) and 4 (Q4) of the IIEFquestionnaire. The sum Q3 + Q4 has been previ-ously used in other trials as a valid end point [28].

Statistical AnalysisThe statistical significance of difference in meanswas assessed using the Kruskall–Wallis test. Theprimary end point of the study was the recovery ofEF without the need for any device or injections,with follow-up measured from the date of RP.Curves for probability of recovering EF were gen-erated using the Kaplan–Meier method and com-pared using the log-rank test. Cox proportionalhazards analysis was also performed to determinepredictors of recovery of EF using forward step-wise variable selection to obtain maximum likeli-hood estimates of the hazard ration and 95%confidence intervals (CIs) in multivariate analyses.Statistical analyses were performed using commer-cial statistical software. Patients were censored ifthey received adjuvant or salvage therapy for abiochemical recurrence (BCR) (N = 29) or had apenile prosthesis implanted (N = 6). Variablesassessed as predictors of recovery of EF includedpatient age, ethnicity, BMI, specimen weight, pre-operative potency level, number of comorbidities(0 vs. 1 vs. 2 or more), graft length, donor site ofgraft (sural vs. other), urologist, and plasticsurgeon. Age, BMI, and specimen weight wereanalyzed as continuous variables with a naturallogarithmic transformation for specimen weight.

168 Rabbani et al.

J Sex Med 2010;7:166–181

Results

In all, 131 patients underwent complete unilateralNVB resection with unilateral NG. Medianpatient age was 58.7 (interquartile range [IQR]53.4, 62.6). Median follow-up was 37.3 months(IQR 14.7–64.1 months). Adjuvant or salvagetherapy was given in 29 (22.1%) patients at amedian of 12.8 months (IQR 5.8–30.0 months).Adjuvant therapy consisted of hormonal therapy infive men, radiation to the prostatic bed in two, andhormonal therapy and radiation in one. Salvagetherapy consisted of hormonal therapy in eightmen, radiation to the prostatic bed in nine, andhormonal therapy and radiation in four. The pre-operative clinical, intraoperative, and pathologiccharacteristics are presented in Table 1.

Interposed nerves were sural in 49, genitofemo-ral in 79, and ilioinguinal in three cases. The meanlength of the NG was 5.9 cm (range 3–10 cm) witha statistically significant difference in the graftlength between plastic surgeons (6.4 cm for plasticsurgeon 1 and 5.8 cm for plastic surgeons 2, 3 aswell as the others, P = 0.048).

The cumulative probability of EF recoveryaccording to the various thresholds (levels 3, 2, and1) of the physician interview-based scale and thevarious thresholds (6, 8, and 10) for the sum of Q3and Q4 of the IIEF questionnaire is presented inFigure 1. The 5-year actuarial probability of EFrecovery was 46, 30, and 12% for levels 3, 2, and 1,respectively. The 5-year actuarial probability ofEF recovery was 40, 34, and 22% for IIEFQ3 + Q4 � 6, �8, and = 10, respectively.

The univariate analysis predictors of EF recov-ery defined according to the various thresholds(levels 3, 2, and 1) of the physician interview-basedscale are presented in Table 2. On univariateanalysis, graft length was not a significant predic-tor of EF recovery, but was a significant predictorof recovery of level 2 EF, independent of plasticsurgeon, only if specimen weight was not includedin the model with a 35% (95% CI 1–57%) reduc-tion in likelihood of recovery for every 1-cmincrease in graft length. A subset of 89 men com-pleted the IIEF questionnaire from which ques-tions 3 and 4 were included in the analysis ofpredictors of recovery of EF. This was definedaccording to the various thresholds (6, 8, and 10)for the sum of IIEF Q3 and Q4, the univariateanalysis results of which are presented in Table 3.The number of questionnaires completed rangedfrom 1–15 (mean 3.4). The cumulative probabilityof EF recovery by plastic surgeon according to the

Table 1 Preoperative clinical, intraoperative, andpathologic characteristics

Preoperative parameter N (%)

AgeMedian (IQR) 58.7 (53.4–62.6)

Preoperative potency:Level 1 111 (84.7%)Level 2 20 (15.3%)

EthnicityAfrican American 13 (9.9%)Other 4 (3.1%)White 114 (87.0%)

BMI*Median (IQR) 27.8 (25.5–30.9)

Preoperative prostate-specific antigen (ng/mL)Median (IQR) 6.2 (4.3–9.3)

Clinical stageT1c 33 (25.2%)T2a 35 (26.7%)T2b 31 (23.7%)T2c 13 (9.9%)T3a 17 (13.0%)T3b 2 (1.5%)

Gleason score6 41 (31.3%)7 64 (48.9%)8–10 26 (19.8%)

ASA score1 20 (15.3%)2 98 (74.8%)3 13 (9.9%)

ComorbiditiesHypertension 40 (30.5%)Hypercholesterolemia/hypertriglyceridemia 36 (27.5%)Diabetes mellitus 7 (5.3%)Coronary artery disease 5 (3.8%)

Number of comorbidities0 69 (52.7%)1 40 (30.5%)2 or more 22 (16.8%)

Intraoperative/pathologic characteristic All patients N (%)

Pathologic stagepT2a 11 (8.4%)pT2b 36 (27.5%)pT2c 11 (8.4%)pT3a 51 (38.9%)pT3b 15 (11.5%)pT4 7 (5.3%)

Nodal statusN0 113 (86.3%)N1–N2 18 (13.7%)

Positive surgical margins 20 (15.3%)Pathologic Gleason score

6 17 (13.0%)7 82 (62.6%)8–9 32 (24.4%)

Specimen weight (g)†

Median (range) 45.5 (22–132)Operative time (min)

Median (IQR) 248 (225–278)Estimated blood loss (cc)

Median (IQR) 1,200 (1,000–2,000)Blood transfusion

Autologous 74 (56.5%)Allogenic 32 (24.4%)

Urologic surgeonSurgeon 1 49 (37.4%)Surgeon 2 57 (43.5%)Other 6 surgeons 25 (19.1%)

Plastic surgeonSurgeon 1 35 (26.7%)Surgeon 2 58 (44.3%)Surgeon 3 32 (24.4%)Other 3 surgeons 6 (4.6%)

Graft donor siteSural 49 (37.4%)Genitofemoral 79 (60.3%)Ilioinguinal 3 (2.3%)

*Missing in six patients.†Specimen weight missing in five patients.ASA = American Society of Anesthesiologists’; BMI = body mass index; IQR =interquantile range.


J Sex Med 2010;7:166–181

physician interview-based scale and the sum ofIIEF Q3 and Q4 is presented in Figure 2. Thepredictors of EF recovery on multivariate analysisare presented in Tables 4 and 5.

Discussion

Although there is controversy, at least theoreti-cally, as to the efficacy of using a somatic nerve to

graft an autonomic nerve such as the cavernousnerve [11,29], some investigators have demon-strated successful restoration of autonomic path-ways with somatic NGs [30]. Most would agreethat one of the primary reasons for failure of NGto aid in EF recovery is the inability to accuratelyidentify the proximal and distal ends of the cavern-ous nerve at the time of grafting. Given that thecavernous nerve is a plexus [4,31] rather than a

Figure 1 The cumulative probabilityof erectile function recovery accordingto the various thresholds {[A] level 3,[B] level 2, and [C] level 1) of the phy-sician interview-based scale and thevarious thresholds ([D] �6, [E] �8,and [F] =10) for the sum of Q3 and Q4of the International Index of ErectileFunction (IIEF) questionnaire.

A

B

170 Rabbani et al.

J Sex Med 2010;7:166–181

distinct cable, it may be difficult to identify thesingle recipient nerve endings to which to performthe NG anastomosis. This has led some authors[14] to recommend use of the sural nerve ratherthan the genitofemoral nerve as the wider caliberof the sural nerve may give more “room for error”if the exact nerve endings cannot be accuratelyidentified. Many investigators [12,16,17,20,23]

have used the Cavermap Surgical Aid™ (BlueTorch Medical Technologies) to help identify andtag the cavernous nerve prior to resection of thenerve as we have done in the present study to aid inthis regard.

Since the original description by Kim et al. [13]of sural NG at the time of RP with cavernous nerveresection, a number of studies have evaluated the

Figure 1 Continued.

C

D


J Sex Med 2010;7:166–181

recovery of EF after bilateral [11,12,17] and unilat-eral [16,19,20,23,32] NG at the time of RP.Although most of the larger series have been openprostatectomy series, the technique has been morerecently described in laparoscopic [19,21] androbotic prostatectomy populations [18,33]. Bilat-eral NG series [11,12,17], which have reported EFrecovery rates of 34–43% as defined as erectionssufficient for intercourse with or without PDE5inhibitors, have served as proof of principle that

NG may aid in recovery of EF after NVB resection.However, proof of efficacy of unilateral NG is morechallenging given that the observed EF recoverymay be solely because of the preservation of thecontralateral (unresected) NVB.

Although potency level 3 on our interview-based scale and a score of 6–7 for IIEF Q3 + Q4may not be considered functional erections, wehave included these thresholds in our analysis forcomparison with previously published series of

Figure 1 Continued.

E

F

172 Rabbani et al.

J Sex Med 2010;7:166–181

NG that have used the definition of achievementof an erection sufficient for intercourse, corre-sponding to our level 3. In our study, the 5-yearactuarial probability of EF recovery was 46, 30,and 12% for levels 3, 2, and 1, respectively. The5-year actuarial probability of recovery was 40, 34,and 22% for IIEF Q3 + Q4 sum �6, �8, and = 10,respectively. Our overall results are inferior tosome of the previously published studies of unila-teral NG although there are differences in the

methodology of determining EF recovery. Asrecovery is a time-dependent event, we have usedthe Kaplan–Meier methodology to calculate actu-arial probabilities of recovery and used the Coxproportional hazards methodology for multivari-ate analysis.

Results of unilateral NG series reported to datehave been mixed, with a number of studies sug-gesting a potential benefit. Sim et al. compared 41men undergoing unilateral sural NG at the time of

Table 2 Univariate analysis of predictors of recovery of potency levels 3, 2, and 1

Parameter

Potency level 3 Potency level 2 Potency level 1

HR (95% CI) P value HR (95% CI) P value HR (95% CI) P value

Age 0.95 (0.90–0.997) 0.036 0.96 (0.90–1.02) 0.15 0.91 (0.81–1.01) 0.067BMI 0.98 (0.91–1.07) 0.70 0.97 (0.88–1.08) 0.62 0.95 (0.79–1.15) 0.60Ethnicity: AA/other vs. white 1.46 (0.65–3.30) 0.37 0.57 (0.13–2.41) 0.44 1.02 (0.13–8.31) 0.98Specimen weight* 0.39 (0.15–1.03) 0.056 0.25 (0.071–0.91) 0.035 0.12 (0.009–1.60) 0.11Preoperative potency: level 2 vs. level 1 0.45 (0.16–1.26) 0.13 0.35 (0.082–1.48) 0.15 0.036 (0–84.5) 0.40Number of comorbidities

1 vs. 0 0.77 (0.38–1.58) 0.48 1.01 (0.42–2.42) 0.98 0.88 (0.16–4.80) 0.882 or more vs. 0 0.53 (0.20–1.39) 0.20 0.83 (0.27–2.53) 0.74 1.40 (0.26–7.62) 0.70

Graft length (cm) 1.13 (0.85–1.49) 0.40 0.88 (0.60–1.28) 0.51 1.07 (0.59–1.92) 0.83Nerve graft donor site:Genitofemoral/ilioinguinal vs. sural nerve 1.35 (0.70–2.62) 0.37 1.78 (0.75–4.20) 0.19 4.16 (0.51–33.84) 0.18Urologic surgeon:

Surgeon 2 vs. 1 0.95 (0.49–1.83) 0.88 1.33 (0.60–2.96) 0.48 2.16 (0.42–11.13) 0.36Other vs. Surgeon 1 0.38 (0.13–1.13) 0.081 0.17 (0.022–1.34) 0.093 0.85 (0.077–9.44) 0.90

Plastic surgeon:Surgeon 2 vs. 1 0.52 (0.26–1.06) 0.073 0.23 (0.094–0.57) 0.001 0.19 (0.036–0.97) 0.045Surgeon 3 vs. 1 0.40 (0.16–0.99) 0.048 0.24 (0.080–0.69) 0.008 0.16 (0.019–1.39) 0.097Other vs. Surgeon 1 0.66 (0.15–2.92) 0.59 0.34 (0.043–2.66) 0.30 — 0.99

*Natural logarithmic transformation.AA = African American; CI = confidence interval; HR = hazard ratio.

Table 3 Univariate analysis of predictors of recovery of potency (defined as sum of IIEF Question 3 and 4 �6, �8, and= 10) (N = 89)

Parameter

Sum IIEF question 3and 4 �6


Sum IIEF question 3and 4 = 10


Age 0.93 (0.87–0.98) 0.008 0.91 (0.85–0.97) 0.004 0.87 (0.80–0.95) 0.003BMI 1.02 (0.92–1.14) 0.66 1.06 (0.94–1.19) 0.35 1.14 (0.97–1.34) 0.11Ethnicity: AA/other vs. white 0.75 (0.23–2.49) 0.64 0.30 (0.041–2.25) 0.24 0.52 (0.068–4.04) 0.54Specimen weight* 0.58 (0.20–1.71) 0.32 0.66 (0.20–2.23) 0.50 0.45 (0.087–2.29) 0.33Preoperative potency: level 2 vs. level 1 0.73 (0.25–2.12) 0.57 0.66 (0.20–2.24) 0.51 0.82 (0.18–3.72) 0.80Number of comorbidities

1 vs. 0 0.93 (0.41–2.08) 0.85 0.84 (0.34–2.09) 0.71 0.96 (0.29–3.24) 0.952 or more vs. 0 0.66 (0.22–1.95) 0.45 0.37 (0.084–1.60) 0.18 0.70 (0.15–3.26) 0.65

Graft length 1.21 (0.87–1.67) 0.26 1.30 (0.94–1.79) 0.11 1.12 (0.70–1.78) 0.64Nerve graft donor siteGenitofemoral/ilioinguinal vs. Sural nerve 0.79 (0.38–1.67) 0.54 0.81 (0.36–1.87) 0.63 1.22 (0.38–3.91) 0.74Urologic surgeon

Surgeon 2 vs. 1 0.98 (0.45–2.10) 0.95 1.20 (0.50–2.86) 0.69 2.07 (0.64–6.69) 0.22Other vs. Surgeon 1 0.48 (0.14–1.68) 0.25 0.63 (0.17–2.31) 0.49 0.42 (0.049–3.61) 0.43

Plastic surgeonSurgeon 2 vs. 1 0.50 (0.22–1.13) 0.094 0.26 (0.099–0.66) 0.005 0.20 (0.057–0.67) 0.009Surgeon 3 vs. 1 0.29 (0.097–0.87) 0.027 0.27 (0.087–0.84) 0.023 0.13 (0.023–0.71) 0.019Other vs. Surgeon 1 0.46 (0.059–3.51) 0.45 — 0.98 — 0.99

*Natural logarithmic transformation.AA = African American; BMI = body mass index; CI = confidence interval; HR = hazard ratio; IIEF = International Index of Erectile Function.


J Sex Med 2010;7:166–181

RP with unilateral NVB resection with 49 con-temporaneous control patients undergoing RPwith unilateral NVB resection without NG [20].They used the IIEF and Rigiscan data to assignpatients to groups corresponding to our level 3 and

level 2 potency thresholds. At 24 months, level 3potency was achieved in 28/38 (74%) of NGpatients vs. 25/49 (51%) of control patients,whereas level 2 potency was achieved in 24/38(63%) of NG patients vs. 13/49 (27%) of control

Figure 2 Recovery of (A) potencylevels 3, (B) level 2, and (C) level 1by plastic surgeon. P values are forcomparison with Surgeon 1 as thereference category (log-rank test).Recovery of potency defined as(D) International Index of ErectileFunction (IIEF) Q3 + Q4 �6, (E)IIEF Q3 + Q4 �8, and (F) IIEFQ3 + Q4 = 10 by plastic surgeon.P values are for comparison withSurgeon 1 as the reference cat-egory (log-rank test).

A

Follow-up (months)12010896847260483624120

Pro

bab

ility

of

reco

very

of

po

ten

cy

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Surgeon 1-censoredSurgeon 2-censoredSurgeon 3-censored

Other surgeons-censored

Surgeon 1Surgeon 2Surgeon 3Other surgeons

Plastic surgeon

Recovery of potency level 3

P = 0.071

P = 0.039

P = 0.53

B

174 Rabbani et al.

J Sex Med 2010;7:166–181

patients, suggesting a potential benefit to suralNG. However, the study of Sim et al. [20] wasnot randomized and one significant differencebetween the groups was that the mean age was 3.7years less in the NG group. Given that our studyshows a 5% decline in recovery of level 3 potencyfor each 1-year increase in age, the control groupwould be expected to have a 17% relative decreasein recovery of level 3 potency, so that the adjustedrate of recovery would be 62% if the control group

was on average 3.7 years younger. Porpiglia et al.compared 15 men with a mean age of 66 yearsundergoing unilateral sural NG at the time of lap-aroscopic RP with unilateral NVB resection with14 control patients with mean age 66.7 yearsundergoing laparoscopic RP with unilateral NVBresection alone [19]. By 18 months, 5/12 (42%) ofNG patients had achieved erections with orwithout sildenafil compared with 3/10 (30%) ofcontrol patients. The NG group had significantly

Figure 2 Continued.

D

C


J Sex Med 2010;7:166–181

higher IIEF-5 scores at 12 and 18 months(P < 0.01) than those in the control group. Namikiet al. compared 19 men undergoing unilateralsural NG at the time of RP with unilateral NS with34 men undergoing bilateral NS and 60 menundergoing unilateral NS alone, using the Univer-sity of California, Los Angeles Prostate CancerIndex to assess EF recovery [32]. After 24 months,

there were no significant differences between thebilateral NS group and the unilateral NG groupalthough the patients in the unilateral NG groupwere younger than those in the bilateral NSgroup. At baseline, the unilateral NG group andbilateral NS group had better sexual function thanthe unilateral NS group, making comparison withthe latter group on follow-up difficult.

Figure 2 Continued.

E

F

176 Rabbani et al.

J Sex Med 2010;7:166–181

However, a number of other studies have notdemonstrated any benefit. Joffe and Klotzreported on 22 men with mean age of 62 under-going unilateral genitofemoral NG with EF recov-ery assessed by IIEF questionnaire [23]. Threemen (14%) recovered normal erections (IIEF-EFdomain score 26–30) and 6/22 (27%) had IIEF-EF domain score of �22. Zorn and associatesreported on 23 men undergoing a unilateral suralNG at the time of robotic RP with unilateral NVBresection [33]. With a mean follow-up of 26.1months, 11/23 (48%) patients with a unilateralNG regained potency, compared with 128/228(56%) with unilateral NS without a NG (P = 0.44).Rates of return to baseline sexual function, definedas recovery of �75% of the baseline sexual func-tion score, at 6, 12, and 24 months in the unilateralNG group were 11, 36, and 45%, comparable withthe unilateral NS only group (P > 0.05). In theonly randomized trial performed of RP with uni-lateral NVB resection with or without a unilateralsural NG, Davis et al. did not find a 20% differ-ence at 2 years between those having and nothaving a sural NG [16]. The study was designedwith a 2:1 randomization in favor of grafting.Assuming 40% 2-year potency in the ungraftedcontrol arm and 60% 2-year potency in the graftedarm, the sample size necessary for 80% powerusing a two-sided test at a 5% significance level

was 200 evaluable patients (120 NGs, 80 control).The trial was terminated after interim analysiswith 107 randomized patients met criteria forfutility. In this study, no provision was made forpatient dropout, which the authors acknowledgewas higher than anticipated with 88 of 107 menremaining on protocol. For patients completingthe protocol to 2 years, potency was recovered in32/45 (71%) with a sural NG vs. 14/21 (67%) ofcontrols (P = 0.777). Given the assumptions of thistrial, a sample size of 107 patients would only yielda power of 50% using a two-sided test at a 5%significance level, making the study underpoweredto detect the 20% difference anticipated. Further-more, details of the randomization process areunclear in the publication. If the patients wererandomized prior to the surgery, patients may havedropped out prior to surgery. Additionally, theurologist may have preserved some nerve fibers onthe side of the NVB resection or performed apotentially more complete NS on the contralateralside in patients known to be randomized to thecontrol arm. Alternatively, if the arm to which thepatient was randomized was revealed in the oper-ating room after the prostatectomy had been per-formed, this bias would be avoided.

To our knowledge, the present study is the firstto evaluate predictors of EF recovery with unilat-eral NG as well as representing the largest series of

Table 4 Multivariate analysis of predictors of recovery of potency levels 3, 2, and 1 (N = 126)

Parameter

Potency level 3 Potency level 2 Potency level 1


Age 0.95 (0.90–0.997) 0.036Specimen weight* 0.24 (0.064–0.92) 0.037Plastic surgeon

Surgeon 2 vs. 1 0.23 (0.092–0.56) 0.001 0.19 (0.037–1.00) 0.050Surgeon 3 vs. 1 0.19 (0.058–0.61) 0.006Other vs. Surgeon 1 0.34 (0.044–2.62) 0.30

*Natural logarithmic transformation.CI = confidence interval; HR = hazard ratio.

Table 5 Multivariate analysis of predictors of recovery of potency (defined as sum of IIEF Question 3 and 4 �6, �8,and = 10) (N = 89)

Parameter



Sum IIEF question 3and 4 = 10


Age 0.93 (0.87–0.98) 0.008 0.88 (0.81–0.95) 0.002 0.82 (0.73–0.93) 0.001Ethnicity: AA/other vs. white 0.12 (0.014–1.03) 0.053 0.072 (0.006–0.90) 0.041Plastic surgeon

Surgeon 2 vs. 1 0.34 (0.13–0.92) 0.033 0.26 (0.069–1.00) 0.050Surgeon 3 vs. 1 0.24 (0.075–0.78) 0.017 0.088 (0.015–0.52) 0.008

AA = African American; CI = confidence interval; HR = hazard ratio; IIEF = International Index of Erectile Function.


J Sex Med 2010;7:166–181

patients who have undergone unilateral cavernousNG during RP. We identified age, specimenweight, plastic surgeon, and patient ethnicity aspredictors on multivariate analysis of varyingthresholds of EF recovery after unilateral NG atthe time of RP and unilateral NVB resection.Patient age has consistently been a significantpredictor of the recovery of potency [4–6,34].Specimen weight was a significant independentpredictor of recovery of level 2 (functionally mean-ingful) erections that are routinely sufficient forintercourse. There was a 76% reduction in level 2erection for every e-fold increase in specimenweight. This is a reflection of the decreased like-lihood of graft function with longer length nervedefects to be bridged when the prostate is larger,consistent with similar findings for somatic nervegrafting [35]. Given that graft length and specimenweight are interrelated variables, graft length wasnot a significant independent predictor on multi-variate analysis. Specimen weight may reflect notonly graft length but also increased likelihood ofoccult damage to the contralateral nerve that isspared with larger prostates.

Our results suggest that technical factors play arole in recovery of EF after unilateral NG at thetime of RP with unilateral NVB resection. Toaccount for technical factors and the role they mayplay in the recovery of EF after unilateral NG,urologic surgeon as well as plastic surgeon andgraft donor site have been evaluated as potentialpredictors in this study. Although case selectionmay be a confounder, generally, differences inrecovery rates among the different urologic sur-geons would reflect differences in the quality ofcontralateral NVB preservation whereas differ-ences among the different plastic surgeons wouldreflect differences in the quality of identification ofthe recipient nerve endings as well as quality of theNG and surgeon preference for a given donornerve site. Hence, any differences seen amongplastic surgeons or among different graft donorsites would be compelling evidence in support of abenefit to unilateral nerve grafting. On the multi-variate analysis, compared with plastic surgeon 1,plastic surgeon 2 had a 77 and 81% lower likeli-hood of recovery of potency levels 2 and 1, respec-tively, and a 66 and 74% lower likelihood of EFrecovery defined as IIEF Q3 + Q4 �8 and = 10,respectively. Compared with plastic surgeon 1,plastic surgeon 3 had an 81% lower likelihood ofrecovery of potency level 2 and a 76 and 91%lower likelihood of EF recovery defined as IIEFQ3 + Q4 �8 and = 10, respectively. NG donor

site was not a significant independent predictor ofEF recovery although the genitofemoral and ilio-inguinal nerves had a 35, 78, and 316% higherlikelihood of recovery of potency levels 3, 2, and 1,respectively, on univariate analysis. This is likely areflection of the preference of plastic surgeon 1 forthe genitofemoral nerve as a donor nerve. It ispossible that case selection may account for someof the difference in recovery seen between plasticsurgeons given that some plastic surgeons mayperform grafts in all patients and others may notperform grafts in those with poor identification ofthe recipient nerve endings; however, this assumesthat there is a benefit to nerve grafting and again,the differences observed support efficacy of nervegrafting.

Our finding that ethnicity is a significant inde-pendent predictor of EF recovery defined as IIEFQ3 + Q4 �8 and = 10 is an interesting observa-tion. The racial differences in pelvic anatomy havebeen well documented in early anthropologicstudies [36]. Turner reported that the pelvis wasnarrower in blacks than in Caucasian Europeans[36]. We identified an 88 and 93% lower likeli-hood of EF recovery defined as IIEF Q3 + Q4 �8and = 10, respectively, for African American orother ethnicity vs. white men. This finding may bea reflection of increased difficulty in identificationof the distal cavernous nerve ending or greaterdifficulty in the distal anastomosis in individualswith a narrow pelvis, although it may also reflectincreased difficulty with contralateral NS in anarrow pelvis. Indeed, Rabbani et al. have re-ported a 76% higher apical positive margin ratein African-American men, consistent withincreased difficulty in the apical dissection in menwith a narrow pelvis [37]. Again, the impact ofethnicity on the results of unilateral NG recoverywould further support the notion that the NG maybe contributing to EF recovery.

One of the major strengths of our study is thesize of the patient population and the duration offollow-up, together with use of sexual functionquestionnaires in the majority of patients infollow-up. To our knowledge, this study representsthe largest series of patients who have undergoneunilateral cavernous nerve interposition graftingduring prostatectomy and it suggests that at leastin some surgeons’ hands, there may be improve-ment in recovery of EF with a unilateral NG inpatients undergoing RP with unilateral NVBresection. The limitations of the study include theretrospective nature, the lack of IIEF scores for allof our patients preoperatively as well as lack of a

178 Rabbani et al.

J Sex Med 2010;7:166–181

contemporaneous control group undergoing uni-lateral NVB resection without NG. Without acontemporaneous control group, the present studydoes not provide information on the efficacy ofunilateral nerve grafting. However, the 5-pointrigidity scale that was used preoperatively, corre-lates well with the IIEF-EF domain score [17].Ultimately, to determine if NGs aid in recovery ofEF, a properly designed randomized trial withadequate power would be necessary. As such, in theabsence of a randomized trial demonstratingbenefit, nerve grafting at RP has not become astandard practice.

Conclusion

We identified patient age, specimen weight, eth-nicity, and plastic surgeon to be predictive ofrecovery of EF after RP with unilateral NVBresection with unilateral NG. The impact ofplastic surgeon on EF recovery with unilateral NGwould suggest that technical factors play a role inthe recovery of EF after unilateral NG. Meticu-lous surgical technique with proper identificationof proximal and distal recipient nerve endings mayimprove the chance of recovery of EF. The varia-tion in recovery rate among plastic surgeons wouldimply that there is a benefit to unilateral NG in EFrecovery. Although there is a suggestion that uni-lateral NG after unilateral NVB resection mayimprove potency, an adequately-powered random-ized trial is required to confirm this finding.

Acknowledgment

Supported by the Sidney Kimmel Center for prostateand Uroloeic Cancer.

Corresponding Author: Farhang Rabbani, MD,FRCSC, Department of Surgery, Urology Service,Memorial Sloan-Kettering Cancer Center, 1275 YorkAvenue, New York, NY 10065, USA. Tel: (646) 422-4385; Fax: (212) 988-0760; E-mail: [email protected]

Conflict of Interest: None.

Statement of Authorship

Category 1(a) Conception and Design

Farhang Rabbani; John P. Mulhall; Peter T.Scardino

(b) Acquisition of DataFarhang Rabbani; Ranjith Ramasamy; Manish I.Patel; Paul Cozzi

(c) Analysis and Interpretation of DataFarhang Rabbani; Joseph J. Disa; Peter G. Cord-eiro; Babak J. Mehrara; James A. Eastham; Peter T.Scardino; John P. Mulhall

Category 2(a) Drafting the Article

Farhang Rabbani; Ranjith Ramasamy(b) Revising It for Intellectual Content

Farhang Rabbani; Manish I. Patel; Paul Cozzi;Joseph J. Disa; Peter G. Cordeiro; Babak J.Mehrara; James A. Eastham; Peter T. Scardino;John P. Mulhall

Category 3(a) Final Approval of the Completed Article

Farhang Rabbani; Ranjith Ramasamy; Manish I.Patel; Paul Cozzi; Joseph J. Disa; Peter G. Cord-eiro; Babak J. Mehrara; James A. Eastham; Peter T.Scardino; John P. Mulhall

References

1 Chartier-Kastler E, Amar E, Chevallier D, Mon-taigne O, Coulange C, Joubert JM, Giuliano F.Does management of erectile dysfunction afterradical prostatectomy meet patients’ expectations?Results of a national survey (REPAIR) by the FrenchUrological association. J Sex Med 2008;5:693–704.

2 Walsh PC, Donker PJ. Impotence following radicalprostatectomy: Insight into etiology and prevention.J Urol 1982;128:492–7.

3 Walsh PC. Anatomical radical retropubic prostate-ctomy. In: Wein AJ, ed. Cambell’s urology. 7thedition. Philadelphia, PA: W.B. Saunders; 1998:2565–88.

4 Quinlan DM, Epstein JI, Carter BS, Walsh PC.Sexual function following radical prostatectomy:Influence of preservation of neurovascular bundles.J Urol 1991;145:998–1002.

5 Rabbani F, Stapleton AM, Kattan MW, WheelerTM, Scardino PT. Factors predicting recoveryof erections after radical prostatectomy. J Urol2000;164:1929–34.

6 Kundu SD, Roehl KA, Eggener SE, Antenor JA,Han M, Catalona WJ. Potency, continence andcomplications in 3,477 consecutive radical retropu-bic prostatectomies. J Urol 2004;172:2227–31.

7 Kim ED, Nath R, Kadmon D, Lipshultz LI, MilesBJ, Slawin KM, Tang HY, Wheeler T, Scardino PT.Bilateral nerve graft during radical retropubic pros-tatectomy: 1-year followup. J Urol 2001;165:1950–6.

8 Quinlan DM, Nelson RJ, Walsh PC. Cavernousnerve grafts restore erectile function in denervatedrats. J Urol 1991;145:380–3.

9 Ball RA, Lipton SA, Dreyer EB, Richie JP, VickersMA. Entubulization repair of severed cavernous


J Sex Med 2010;7:166–181

nerves in the rat resulting in return of erectile func-tion. J Urol 1992;148:211–5.

10 Lowe JB 3rd, Hunter DA, Talcott MR, MackinnonSE. The effects of cavernous nerve grafting follow-ing surgically induced loss of erectile function in alarge-animal model. Plast Reconstr Surg 2006;118:69–80.

11 Kim ED, Nath R, Slawin KM, Kadmon D, Miles BJ,Scardino PT. Bilateral nerve grafting during radicalretropubic prostatectomy: Extended follow-up.Urology 2001;58:983–7.

12 Chang DW, Wood CG, Kroll SS, Youssef AA,Babaian RJ. Cavernous nerve reconstruction to pre-serve erectile function following non-nerve-sparingradical retropubic prostatectomy: A prospectivestudy. Plast Reconstr Surg 2003;111:1174–81.

13 Kim ED, Scardino PT, Hampel O, Mills NL,Wheeler TM, Nath RK. Interposition of sural nerverestores function of cavernous nerves resected duringradical prostatectomy. J Urol 1999;161:188–92.

14 Scardino PT, Kim ED. Rationale for and resultsof nerve grafting during radical prostatectomy.Urology 2001;57:1016–9.

15 Walsh PC. Nerve grafts are rarely necessary and areunlikely to improve sexual function in men under-going anatomic radical prostatectomy. Urology2001;57:1020–4.

16 Davis JW, Chang DW, Chevray P, Wang R, ShenY, Wen S, Pettaway CA, Pisters LL, Swanson DA,Madsen LT, Huber N, Troncoso P, Babaian RJ,Wood CG. Randomized phase II trial evaluation oferectile function after attempted unilateral cavern-ous nerve-sparing retropubic radical prostatectomywith versus without unilateral sural nerve graftingfor clinically localized prostate cancer. Eur Urol2009;55:1135–44.

17 Secin FP, Koppie TM, Scardino PT, Eastham JA,Patel M, Bianco FJ, Tal R, Mulhall J, Disa JJ, Cor-deiro PG, Rabbani F. Bilateral cavernous nerveinterposition grafting during radical retropubicprostatectomy: Memorial Sloan-Kettering CancerCenter experience. J Urol 2007;177:664–8.

18 Kaouk JH, Desai MM, Abreu SC, Papay F, Gill IS.Robotic assisted laparoscopic sural nerve graftingduring radical prostatectomy: Initial experience.J Urol 2003;170:909–12.

19 Porpiglia F, Ragni F, Terrone C, Renard J, Musso F,Grande S, Cracco C, Ghignone G, Scarpa RM. Islaparoscopic unilateral sural nerve grafting duringradical prostatectomy effective in retaining sexualpotency? BJU Int 2005;95:1267–71.

20 Sim HG, Kliot M, Lange PH, Ellis WJ, TakayamaTK, Yang CC. Two-year outcome of unilateral suralnerve interposition graft after radical prostatectomy.Urology 2006;68:1290–4.

21 Turk IA, Deger S, Morgan WR, Davis JW, Schell-hammer PF, Loening SA. Sural nerve graft duringlaparoscopic radical prostatectomy. Initial experi-ence. Urol Oncol 2002;7:191–4.

22 Kilgo MS, Howard MA, Kaplan G, Rabbani F, Scar-dino PT, Cordeiro PG. Evaluation of genitofemoralnerve donor site morbidity after radical prostatec-tomy. Ann Plast Surg 2005;55:57–61.

23 Joffe R, Klotz LH. Results of unilateral genitofemo-ral nerve grafts with contralateral nerve sparingduring radical prostatectomy. Urology 2007;69:1161–4.

24 Satkunasivam R, Appu S, Al-Azab R, Hersey K,Lockwood G, Lipa J, Fleshner NE. Recovery oferectile function after unilateral and bilateral cav-ernous nerve interposition grafting during radicalpelvic surgery. J Urol 2009;181:1258–63.

25 Cappelleri JC, Rosen RC, Smith MD, Mishra A,Osterloh IH. Diagnostic evaluation of the erectilefunction domain of the International Index of Erec-tile Function. Urology 1999;54:346–51.

26 Eastham JA, Scardino PT. Radical prostatectomyfor clinical stage T1 and T2 prostate cancer. In:Coffey DS, ed. Comprehensive textbook of geni-tourinary oncology. 2nd edition. Philapelphia, PA:Lippincott Williams and Williams; 2000:722–38.

27 Rabbani F, Patel M, Cozzi P, Mulhall JP, ScardinoPT. Recovery of erectile function after radical pros-tatectomy is quantitatively related to the response tointraoperative cavernous nerve stimulation. BJU Int2009 Mar 30 [Epub ahead of print].

28 Padma-Nathan H, McCullough AR, Levine LA,Lipshultz LI, Siegel R, Montorsi F, Giuliano F,Brock G; Study Group. Randomized, double-blind, placebo-controlled study of postoperativenightly sildenafil citrate for the prevention oferectile dysfunction after bilateral nerve-sparingradical prostatectomy. Int J Impot Res 2008;20:479–86.

29 Rabbani F, McKiernan JM, Disa JJ, Scardino PT.Sural nerve transplant: Preserving potency afterradical prostatectomy. Contemp Urol 2002;14:50–60.

30 Hyochi N, Kihara K, Arai G, Kobayashi T, Sato K.Reconstruction of the sympathetic pathway project-ing to the prostate, by nerve grafting in the dog.BJU Int 2004;94:147–52.

31 Walsh PC, Epstein JI, Lowe FC. Potency followingradical prostatectomy with wide unilateral excisionof the neurovascular bundle. J Urol 1987;138:823–7.

32 Namiki S, Saito S, Nakagawa H, Sanada T, YamadaA, Arai Y. Impact of unilateral sural nerve graft onrecovery of potency and continence followingradical prostatectomy: 3-year longitudinal study.J Urol 2007;178:212–6.

33 Zorn KC, Bernstein AJ, Gofrit ON, Shikanov SA,Mikhail AA, Song DH, Zagaja GP, Shalhav AL.Long-term functional and oncological outcomes ofpatients undergoing sural nerve interposition graft-ing during robot-assisted laparoscopic radical pros-tatectomy. J Endourol 2008;22:1005–12.

180 Rabbani et al.

J Sex Med 2010;7:166–181

34 Geary ES, Dendinger TE, Freiha FS, Stamey TA.Nerve sparing radical prostatectomy: A differentview. J Urol 1995;154:145–9.

35 Roganovic Z, Petkovic S. Missile severances of theradial nerve. Results of 131 repairs. Acta Neurochir(Wien) 2004;146:1185–92.

36 Turner W. The index of the pelvic brim as a basis ofclassification. J Anat Physiol 1885;20:125–43.

37 Rabbani F, Herran-Yunis L, Vora K, Eastham JA,Guillonneau B, Scardino PT, Touijer K. Impact ofethnicity on surgical margins at radical prostatec-tomy. BJU Int 2009 Apr 15 [Epub ahead of print].


J Sex Med 2010;7:166–181

Documents

Predictors of Recovery of Erectile Function after Unilateral Cavernous Nerve Graft Reconstruction at Radical Retropubic Prostatectomy