Predicting participation in and successful outcome of a penile rehabilitation programme using a phosphodiesterase type 5 inhibitor with a vacuum erection device after radical prostatectomy

Authors


Masaki Kimura, Duke University Medical Center, PO Box 2804, Yellow Zone, Durham, NC 27710, USA. e-mail: masaki.kimura@duke.edu

Abstract

Study Type – Therapy (case series)

Level of Evidence 4

What's known on the subject? and What does the study add?

The role of the vacuum erection device (VED) has increased with its use in combined therapy with a phosphodiesterase type 5 inhibitor (PDE5i) for penile rehabilitation after radical prostatectomy (RP) and radiotherapy. The advantages of the VED are non-invasive, cost-effective, and a possibility of preventing shrinkage of penis length. Albeit current widespread use of penile rehabilitation programmes for post-RP erectile dysfunction, independent predictors for the rehabilitation participants, as well as for its treatment success have not been fully investigated.

In the present study, we have added several new predictors for rehabilitation participation, e.g. African-Americans and higher preoperative sexual function. Conversely, higher preoperative PSA concentrations and the presence of positive surgical margins were predictors for avoidance of rehabilitation. Notably, there was a primary surgeon difference, which had a trend for predicting outcome of the rehabilitation among the participants, implying their surgical technique and follow-up might influence success of the rehabilitation.

OBJECTIVES

  • • To investigate baseline demographic and clinicopathological characteristics of men who participate in our penile rehabilitation programme after radical prostatectomy (RP).
  • • To determine predictors for participation in rehabilitation, as well as successful rehabilitation outcome using multivariable logistic regression analyses.

PATIENTS AND METHODS

  • • We analysed data on 2345 consecutive patients who underwent RP between 2001 and 2009 in our institution.
  • • The decision to participate in penile rehabilitation using phosphodiesterase type 5 inhibitor (PDE5i) with a vacuum erection device (VED) was based on the patient's choice after post-RP discussions.
  • • Rehabilitation success was defined using the following criteria: (i) patients who continued the penile rehabilitation programme and did not switch treatment from PDE5i to other erectile aids, (ii) success was noted in men who had an Expanded Prostate Cancer Index Composite (EPIC) sexual function (SF) score of >75% of the patient's baseline EPIC score, and (iii) patients who answered that they achieved adequate erections with a PDE5i.
  • • Logistic regression analysis was used to identify factors associated with treatment participation and its success.

RESULTS

  • • Of 676 patients, 354 (53.2%) men participated in a penile rehabilitation programme. Among 329 rehabilitation participants with available data, 96 (29.2%) had treatment success.
  • • In multivariable regression analysis, African-Americans (odds ratio [OR] 3.47, P < 0.001), and higher preoperative SF (OR 1.02, P < 0.001) were associated with participation in rehabilitation.
  • • Higher preoperative PSA concentration (OR 0.50, P= 0.004) and presence of positive surgical margins (OR 0.68, P= 0.042) were found to be independent predictors for non-participation in the rehabilitation.
  • • For rehabilitation outcomes, being older at surgery (OR 0.93, P= 0.001) and adjuvant therapy (OR 0.34, P= 0.047) had a negative association with successful outcome.
  • • There was a trend in the relationship between primary surgeon and rehabilitation success (OR 1.05, P= 0.053)

CONCLUSIONS

  • • Those patients who have risk factors, e.g. adverse prostate cancer features, need to be carefully counselled and encouraged to participate in the penile rehabilitation programme.
  • • Clinicians could lead patients toward successful outcomes if appropriate surgical techniques and rehabilitation are provided.
Abbreviations
(R)(P)RP

(retropubic) (perineal) radical prostatectomy

ED

erectile dysfunction

HRQL

health-related quality of life

SF

sexual function

PDE5i

phosphodiesterase type 5 inhibitor

VED

vacuum erection device

EF

erectile function

RALP

robot-assisted laparoscopic prostatectomy

EPIC

Expanded Prostate Cancer Index Composite

ADT

androgen-deprivation therapy

PSM

positive surgical margins

OR

odds ratio

INTRODUCTION

Prostate cancer is a major health problem, with 217 730 newly diagnosed cases estimated in the USA in 2010 [1]. Radical prostatectomy (RP) is the ‘gold standard’ for localised prostate cancer with excellent disease control, but adverse outcomes after RP are common. Erectile dysfunction (ED) is one unfavourable comorbid outcome of RP and radiotherapy, and it has a significant negative impact on their sexual health-related quality of life (HRQL) after these radical therapies [2–4]. In this context, penile rehabilitation, defined as the use of any drug or device after RP to maximize the recovery of sexual function (SF), is warranted.

A phosphodiesterase type 5 inhibitor (PDE5i) is commonly used as a first-line primary penile rehabilitation strategy for post-RP ED [5–7]. Using a cavernous nerve injury animal model, which mimics the conditions of a human RP, several studies have investigated therapeutic effects of PDE5i after the injury [8–12]. Based upon those studies, the advantages of a PDE5i for preservation of cavernosal oxygenation and endothelial function along with cavernous nerve regeneration after RP were apparent [13]. However, even if using a PDE5i as a primary rehabilitation treatment, the recovery rate of SF after RP and radiotherapy were not always perfect. Even when both cavernous nerves are preserved, ED rates range widely from 31 to 86% after RP with or without a PDE5i, some of which imply the limitations of this therapy [14–16]. Considering the severity of cavernous nerve injury in these settings, a PDE5i may not play a role as an adequate treatment for men with normal SF before RP. Thus some form of adjunct therapy that uses other mechanisms and molecular pathways may be beneficial.

Currently, several articles report on the expanded role of using a vacuum erection device (VED) as combined therapy with a PDE5i for penile rehabilitation programme after RP and radiotherapy [7,17,18]. The efficacy and safety of the VED has been reported as well. The advantages of the VED are that it is a non-invasive and cost-effective form of drug-free therapy, resulting in excellent long-term potency and patient compliance [19]. The VED has another unique profile for treatment of post-RP ED; that is, preservation of inadvertent shrinkage of the penile length after RP [17]. Because of a lack of natural erections after RP, persistent ischaemic and the hypoxic state of cavernosal tissue may lead to cavernosal fibrosis [20]. In this situation, the VED allows cavernosal tissue to be periodically re-perfused that is similar to spontaneous erections. One basic science study showed that physiological erectile function (EF) measured by intracavernosal pressure/mean arterial pressure ratio improved with VED therapy [21]. Moreover, VED therapy preserved smooth muscle content and endothelial integrity via anti-hypoxia, anti-apoptosis, and anti-fibrosis mechanisms in the cavernous nerve injury rat model [21]. Considering these findings, a penile rehabilitation programme using a PDE5i with a VED could be a promising option as a primary treatment strategy for post-RP ED, although the evidence are still unclear.

Albeit current widespread use of a penile rehabilitation programme for post-RP ED, predictive factors for the participant's willingness or desire to undergo rehabilitation, as well as for its treatment success have not been fully investigated. Therefore, we aimed to study the demographic and clinicopathological characteristics among men who participated in our penile rehabilitation programme. In addition, we further examined independent predictors for willingness to participate, as well as predictors for successful outcome of our primary penile rehabilitation programme by multivariable logistic regression analyses.

PATIENTS AND METHODS

The records of 2345 consecutive patients who were treated with retropubic RP (RRP), perineal RP (PRP), or robot-assisted laparoscopic prostatectomy (RALP) at the Duke University Medical Center from 1 January 2001 to 31 December 2009 were analysed from our prospectively maintained Duke Prostate Center database. In this study, we included 684 patients who had available preoperative Expanded Prostate Cancer Index Composite (EPIC) scores. Those men that had androgen-deprivation therapy (ADT), radiotherapy, or PDE5i before RP were excluded from the analyses. We also excluded men who used other forms of ED aids, e.g. intracavernosal injection or intraurethral prostaglandin as a first-line penile rehabilitation. After applying these criteria, 665 subjects with a mean follow-up 33.4 months remained for the analyses. This study was approved by the Institutional Review Board and waivers for consent or informed consent was obtained from all patients.

SF was evaluated using the EPIC questionnaire, a validated HRQL instrument specific for men with prostate cancer [22]. The EPIC consists of an 80-item survey to evaluate patient function and bother in multiple domains, including urinary, bowel, sexual, and hormonal. The sexual domain has a composite 13-item summary score that is sub-categorised into two subscales: a SF subscale and a sexual bother subscale. Response options for each EPIC item form a Likert scale, and each five-item score is transformed into a scale ranging from 0 to 100. Higher scores represent better HRQL. All preoperative EPIC scores were obtained at a clinic visit several weeks or months before RP and kept under control of a third party data manager.

In our study cohort, the decision to participate in penile rehabilitation was based on the patient's choice after postoperative discussions about treatment options and expectations between the primary surgeon and patient. For those patients who participated in a penile rehabilitation programme, we recommended using a PDE5i three-times per week or on demand, and most of them used a VED daily without the constriction band once urinary continence was achieved. To evaluate outcomes, treatment success for penile rehabilitation was defined using the following criteria: (i) for those patients who continued the penile rehabilitation programme and did not switch treatment from PDE5i to other erectile aids, (ii) success was noted in men who had an EPIC-SF score of >75% of the patient's baseline EPIC-SF score [23,24], and (iii) patients who answered that with a PDE5i they achieved erections adequate for vaginal penetration. Those men who switched primary treatment from PDE5i to other erectile aids, e.g. intracavernosal injection or intraurethral prostaglandin, were considered failures of the primary rehabilitation programme.

For comparison of baseline demographic and clinical characteristics of the analysed cohort, categorical variables including race (Caucasian, African-American, or other), Charlson Comorbidity Index (0, 1, ≥2), biopsy Gleason sum (4–6, 7, 8–9), clinical stage (T1, 2, 3), surgery type (RRP, PRP, RALP), nerve-sparing status (none, unilateral, bilateral), positive surgical margins (PSMs), use of adjuvant therapy, e.g. ADT, radiotherapy or a combination of those, were compared using the Chi-square tests or the Fisher's exact test. Continuous variables, e.g. age at surgery, year of surgery, census tract measurement of annual household income, preoperative EPIC scores, preoperative PSA concentration, and prostate weight, were compared with the Mann–Whitney test.

To determine predictors for participating in a penile rehabilitation programme, logistic regression analyses were conducted. In this model, pre- and perioperative demographics as well as clinicopathological characteristics were assigned as confounding factors, including age at surgery (continuous), year of surgery (continuous), race (categorical), preoperative EPIC-SF scores (continuous), Charlson Comobidity Index (categorical), preoperative PSA concentration (continuous after log-transformation), surgical type (categorical), primary surgeon (categorical; 12 surgeons were randomly assigned into categorical variables), and PSM (dichotomous; none vs present). Secondly, to determine the predictive factors for treatment success of rehabilitation, logistic regression analyses were conducted among men who participated in penile rehabilitation. In this model, adjuvant therapy (dichotomous; none vs underwent) was added as a confounding variable. A P < 0.05 was considered to indicate statistical significance.

RESULTS

BASELINE DEMOGRAPHICS AND CLINICOPATHOLOGICAL CHARACTERISTICS

The study cohort had no significant differences when compared with the total cohort for age (P= 0.409), year of surgery (P= 0.314), preoperative PSA concentration (P= 0.167), and biopsy Gleason sum (P= 0.790). Among the study cohort (N= 665), 354 men (53.2%) were identified as participants of a primary penile rehabilitation programme using a PDE5i and a VED, whereas 311 men (46.8%) were identified as non-participants. Table 1 summarises data of baseline demographics and clinicopathological characteristics between rehabilitation participants and non-participants. In this comparison, men who participated in the rehabilitation programme were more likely to be younger at surgery (P < 0.001), to have recent year of surgery (P= 0.001), to be African-Americans (P= 0.011), to have higher preoperative EPIC-sexual domain, SF, and sexual bother scores (P < 0.001, P < 0.001, and P= 0.005), to have lower preoperative PSA concentrations (P= 0.001), to undergo more RALPs (P < 0.001), to have more nerve-sparing RPs (P < 0.001), and to have a lower incidence of PSMs (P= 0.001) compared with non-participants.

Table 1. Baseline demographics and clinicopathological characteristics of the cohort
VariableParticipantNon-participant P
  • Chi-squared test; ‡Fisher's exact test; §Mann–Whitney test.

N 354311 
Mean (sd):   
 Age at surgery, years58.9 (6.6)61.2 (7.6)<0.001§
 Year of surgery2005.8 (2.2)2005.1 (2.7)0.001§
 Race n (%)  0.011
 Caucasian296 (83.6)280 (90.0)
 African-American53 (15.0)24 (7.7)
 Asian5 (1.4)7 (2.3)
Median (range) income, $ × 1000 (American dollars)42.2 (17.5–86.8)42.9 (23.8–106.0)0.608§
Mean (sd) preoperative EPIC score:   
 Sexual domain summary64.7 (23.1)53.2 (24.2)<0.001§
 Sexual function domain61.1 (23.5)47.5 (26.5)<0.001§
 Sexual bother domain72.6 (28.1)65.7 (30.1)0.005§
Charlson comorbidity index, n (%)  0.749
 0255 (72.0)218 (70.1)
 169 (19.5)68 (21.9)
 ≥230 (8.5)25 (8.0)
Median (range) preoperative PSA concentration, ng/mL5.0 (0.4–86.1)5.5 (0.2–75.9)0.001§
N (%):   
 Biopsy Gleason sum:  0.416
  4–6237 (67.0)193 (62.1)
  796 (27.1)96 (30.9)
  8–921 (5.9)22 (7.1)
 Clinical T stage:  0.132
  T1279 (78.8)260 (84.1)
  T272 (20.3)45 (14.6)
  T33 (0.9)4 (1.3)
 Type of surgery:  <0.001
  RRP119 (33.6)97 (31.2)
  PRP43 (12.2)90 (28.9)
  RALP192 (54.2)124 (39.9)
 Nerve sparing:  <0.001
  None50 (14.2)102 (32.9)
  Unilateral62 (17.6)46 (14.8)
  Bilateral240 (68.2)162 (52.3)
Median (range) prostate weight, g40.0 (15–99.6)38.8 (5–324)0.203§
N (%):   
 Pathological Gleason sum:  0.096
  4–6155 (43.8)127 (40.8)
  7172 (48.6)147 (47.3)
  8–927 (7.6)37 (11.9)
 PSM95 (26.8)119 (38.4)0.001
 Adjuvant therapy32 (9.0)32 (10.3)0.586

DEMOGRAPHICS AND CLINICOPATHOLOGICAL CHARACTERISTICS BASED ON PRIMARY REHABILITATION OUTCOMES

Among the 354 men who participated in the penile rehabilitation programme using a PDE5i with a VED, six discontinued the treatment because of PDE5i-related side-effects including headache, back pain, and flushing. In addition, 19 men were missing data for SF recovery status and were therefore excluded, resulting in 329 men with available SF recovery status data included in the final analysis. Of these 329 men, 96 (29.2%) had treatment success as defined, whereas 233 (70.8%) did not achieve a successful outcome. Of 233 men who failed in primary rehabilitation, 67 switched to intracavernosal injection, 12 to an intraurethral suppository, and 11 ceased PDE5i and kept using the VED alone, while the remaining 143 discontinued penile rehabilitation itself. In this comparison, men who had successful outcome of the rehabilitation were more likely to be younger at surgery (P= 0.003), to undergo RRP (P= 0.003), and to have lower prostate weight (P= 0.011) compared with those who failed in primary rehabilitation (Table 2).

Table 2. Demographics and clinicopathological characteristics based on primary rehabilitation outcomes
VariableSuccessFailure P
  1. †Chi-squared test; ‡Fisher's exact test; §Mann–Whitney test.

N 96233 
Mean (sd):   
 Age at surgery, years57.0 (6.8)59.5 (6.6)0.003§
 Year of surgery2006.1 (2.0)2005.6 (2.2)0.083§
Median (range) income, $ × 1000 (American dollars)39.5 (24.8–96.6)39.3 (23.8–106.0)0.344§
N (%):   
 Race:  0.887
  Caucasian80 (83.3)191 (82.0) 
  African-American15 (15.6)38 (16.3) 
  Asian1 (1.0)4 (1.7) 
 Preoperative EPIC score:   
  Sexual domain summary65.7 (24.9)64.5 (22.2)0.361§
  Sexual function domain61.3 (26.5)61.4 (22.0)0.386§
  Sexual bother domain75.6 (28.5)71.5 (28.0)0.107§
 Charlson comorbidity index:  0.581
  072 (75.0)164 (70.4)
  118 (18.8)47 (20.2)
  ≥26 (6.3)22 (9.4)
Median (range) preoperative PSA concentration, ng/mL4.8 (0.4–86.1)5.1 (0.5–27.3)0.461§
N (%):   
 Biopsy Gleason sum  0.955
  4–666 (67.8)157 (67.4)
  724 (25.0)62 (26.1)
  8–96 (6.3)14 (6.0)
 Clinical T stage  0.721
  T1181 (77.7)75 (78.1)
  T249 (21.0)21 (21.9)
  T33 (1.3)0 (0)
 Type of surgery:  0.003
  RRP46 (47.9)67 (28.8)
  PRP8 (8.3)35 (15.0)
  RALP42 (43.8)131 (56.2)
 Nerve sparing:  0.253
  None11 (11.6)38 (16.4)
  Unilateral14 (14.7)45 (19.4)
  Bilateral70 (73.7)149 (64.2)
Median (range) prostate weight, g37.2 (22–74.2)42.030 (15–99.6)0.011
N (%):   
 Pathological Gleason sum  0.156
  4–641 (42.7)104 (44.6)
  752 (54.2)109 (46.8)
  8–93 (3.11)20 (8.6)
 PSM33 (34.4)56 (24.0)0.055
 Adjuvant therapy6 (9.4)25 (10.7)0.206

PREDICTIVE FACTORS FOR PARTICIPATING IN PENILE REHABILITATION AFTER RP

Uni- and multivariable logistic regression analyses were conducted to identify the predictive factors for participating in primary rehabilitation for postoperative ED (Table 3). In all, 665 men were included in this analysis. In univariable analyses, year of surgery (odds ratio [OR] 1.14, P < 0.001), African-Americans (OR 1.91, P= 0.004), and higher preoperative EPIC-SF (OR 1.02, P < 0.001) had a positive relationship with participation in the rehabilitation, while being older at surgery (OR 0.96, P < 0.001), higher preoperative PSA concentration (OR 0.66, P < 0.001), PRP (OR 0.39, P < 0.001), and presence of PSMs (OR 0.59, P= 0.002) had a negative relationship with participation. In multivariable regression analysis, African-Americans (OR 3.47, P < 0.001), higher preoperative EPIC-SF (OR 1.02, P < 0.001), higher preoperative PSA concentration (OR 0.50, P= 0.004), PRP (OR 0.41, P= 0.012), and presence of PSMs (OR 0.68, P= 0.042) remained as independent predictors.

Table 3. Logistic regression analyses predicting factors for participating in penile rehabilitation after surgery
VariableUnivariableMultivariable
OR (95% CI) P OR (95% CI) P
Age at surgery0.96 (0.94–0.98)<0.0010.98 (0.95–1.01)0.066
Year of surgery1.14 (1.07–1.21)<0.0011.01 (0.91–1.12)0.810
Race:    
 CaucasianReferenceReference
 African-American1.91 (1.17–3.13)0.0043.47 (1.56–4.96)<0.001
 Other0.71 (0.22–2.26)0.5810.30 (0.30–3.88)0.914
Preoperative EPIC-SF1.02 (1.02–1.03)<0.0011.02 (1.01–1.03)<0.001
Comorbidities0.97 (0.78–1.20)0.7971.04 (0.83–1.36)0.616
Preoperative PSA concentration0.66 (0.51–0.84)<0.0010.50 (0.50–0.88)0.004
Surgery type:    
 RRPReferenceReference
 PRP0.39 (0.25–0.61)<0.0010.41 (0.21–0.82)0.012
 RALP1.25 (0.88–1.76)0.2151.14 (0.71–1.85)0.590
Primary surgeon0.98 (0.96–1.00)0.0631.01 (0.98–1.03)0.680
PSM0.59 (0.43–0.82)0.0020.68 (0.47–0.99)0.042

PREDICTIVE FACTORS FOR SUCCESSFUL OUTCOME OF PARTICIPANTS IN PENILE REHABILITATION AFTER RP

To determine predictive factors for a successful outcome of rehabilitation, uni- and multivariable logistic regression analyses were conducted (Table 4). In all, 329 men participated in penile rehabilitation with their available SF recovery status included in the analysis. In univariable analysis, year of surgery (OR 1.12, P= 0.048) and primary surgeon (OR 1.05, P= 0.005) were strongly associated with rehabilitation success, whereas age at surgery (OR 0.95, P= 0.003), PRP (OR 0.32, P= 0.009), RALP (OR 0.47, P= 0.004) were negatively associated with rehabilitation success. In multivariable analysis, age at surgery (OR 0.93, P= 0.001) and adjuvant therapy (OR 0.34, P= 0.047) had a negative association with rehabilitation success. There was a trend in the relationship between primary surgeon and rehabilitation success (OR 1.05, P= 0.053).

Table 4. Logistic regression analyses predicting factors for successful outcome of penile rehabilitation after surgery
VariableUnivariableMultivariable
OR (95% CI) P OR (95% CI) P
Age at surgery0.95 (0.91–0.98)0.0030.93 (0.90–0.97)0.001
Year of surgery1.12 (1.00–1.26)0.0481.15 (0.97–1.35)0.103
Race:    
 CaucasianReferenceReference
 African-American0.94 (0.50–1.85)0.8590.83 (0.40–1.74)0.628
 Other0.60 (0.07–5.54)0.6470.19 (0.02–2.13)0.176
Preoperative EPIC-SF1.00 (0.99–1.01)0.6560.99 (0.98–1.00)0.194
Comorbidities0.77 (0.53–1.11)0.1580.89 (0.60–1.33)0.580
Preoperative PSA concentration1.02 (0.70–1.48)0.9281.11 (0.73–1.69)0.616
Surgery type:    
 RRPReferenceReference
 PRP0.32 (0.14–0.75)0.0090.60 (0.19–1.91)0.382
 RALP0.47 (0.28–0.78)0.0040.78 (0.38–1.62)0.506
Primary surgeon1.05 (1.02–1.09)0.0051.05 (1.00–1.11)0.053
PSM1.66 (0.99–2.77)0.0562.02 (1.13–3.62)0.018
Adjuvant therapy0.54 (0.22–1.35)0.1910.34 (0.12–0.99)0.047

DISCUSSION

A PDE5i together with a VED is a promising treatment option for postoperative ED as a first-line penile rehabilitation regimen [19,20,25]. According to a study that investigated practice patterns for penile rehabilitation, 87% of International Society for Sexual Medicine practitioners performed some form of rehabilitation. As a part of a primary rehabilitation strategy, 95% used a PDE5i, 30% used a VED, 75% used intracavernosal injections and 9.9% used intraurethral prostaglandin [5]. For the participation rate in the present study, 53.2 % of men participated in primary penile rehabilitation, while 46.8 % of men did not participate. This is a similar result when compared with a previous study by Salonia et al. [26] showing that 49% of men freely declined penile rehabilitation.

One important finding of the present study is that African-American race was identified to be a strong independent predictor for participation in the penile rehabilitation programme. One possible explanation for this result might be related to age differences at the time of rehabilitation. The mean age of African-Americans who participated in rehabilitation was 56.6 years, whereas Caucasians averaged 59.3 years, and Asians 59.8 years, which were statistically significant. Considering African-Americans tend to have more aggressive prostate cancer features, younger patients were included in the present study, which may influence our results showing more African-Americans participated in rehabilitation, although further detailed studies are needed [27,28].

In the present study, a higher preoperative SF is one of the independent predictors for participation in penile rehabilitation. Similarly, younger patients showed a trend to be a predictive factor for participation in rehabilitation. These results are rational, considering that patients who are young and have normal SF tend to make an effort to maintain their preoperative EF after RP, although a previous study by Salonia et al. [26] did not find statistical differences in the aforementioned variables due to a small sample size. In contrast, higher preoperative PSA concentration and the presence of PSMs were found to be significant predictors to decline participation in rehabilitation. We have assumed that adverse prostate cancer pathological features are reasonable reasons to avoid rehabilitation by both patients and clinicians, considering patients could have psychological stress about a worse cancer outcome than what the patient thought preoperatively. Moreover, clinicians tend to be reluctant to encourage patients to participate in penile rehabilitation in such a situation as the management of cancer might be the priority. Previous studies have shown that other reasons for avoiding penile rehabilitation were cost, a lack of evidence, and unfamiliarity with the concept, although the present data did not show a difference in household income between the participant and non-participant [5].

Here, concerning penile rehabilitation outcome, being younger had a significant positive association with successful recovery, which is consistent with previous reports and a meta-analysis [29–32]. Additional findings in the present study are that adjuvant therapy, e.g. radiotherapy and ADT, was identified as a predictor for rehabilitation failure. This result was supported by previous series, showing that ADT negatively affected SF owing to hypogonadism that leads to decreased libido and impaired physiological EF [33–35]. While several clinical studies have shown the benefit of PDE5i use for radiation induced ED, the role and its mechanisms of penile rehabilitation for men who underwent post-RP (adjuvant) radiotherapy are currently unknown [36,37]. Given that surgically preserved cavernous nerves will probably be additionally injured by adjuvant radiotherapy to the prostate bed, adjuvant radiotherapy potentially could be negatively associated with recovery of EF among rehabilitation participants [36]. Furthermore, the combination of adjuvant radiotherapy and ADT may introduce a further synergetic negative influence to SF recovery. However, these observations and hypothesis will require a larger study focused on this patient population.

In the present study, we found that primary surgeon differences showed a trend for successful outcome of penile rehabilitation among the participants, suggesting that surgical technique and methods of follow-up might influence outcomes of a rehabilitation strategy. In our institution, most of our primary surgeons continually care for their patients in the period after RP including penile rehabilitation. Thus, the encouragement of penile rehabilitation during several follow-up visits might vary by each surgeon and have an effect on the outcome. In contrast, the surgical procedure differences had no statistically significant association with rehabilitation success after adjusting for several cofounders in multivariable logistic regression analysis. This topic is still controversial as some studies have reported that ED is more likely to be diagnosed in men who underwent RALP than RRP or PRP in long-term follow-up, while other series have reported the advantages of RALP showing lower complication rates than that of conventional RP [23,38,39]. Based on the present results, we considered that the individual surgeon had a greater influence on the preservation of SF rather than the type of surgical procedure.

This study has certain limitations. First, due to the retrospective nature of a single centre cohort, the study may be subject to inherent biases compared with a randomised, blinded, multi-institutional study. Second, determination of rehabilitation success was heterogeneously defined, which might introduces biases in the analyses. The reason several definitions were used in the present study is because follow-up EF information with validated instruments was not available in half of the entire study cohort. Using men comprising half of the cohort could result in losing power in multivariable logistic regression analyses. Third, due to the nature of data extraction by chart review, we could not assess compliance for the rehabilitation programme in the subjects. Although penile rehabilitation should start immediately after RP and be continued consistently, we could not measure actual frequency of VED use as well as the dose of PDE5i administrated, which may have resulted in a possible bias to the outcomes in participating subjects. However, if patients had proper instruction, we consider that VED use does not pose difficulties for most patients [7]. Therefore, we assumed that compliance with the present strategy did not have a significant influence on our results. Despite of these limitations, the present study offers new findings about general predictive factors for participation in a penile rehabilitation programme as well as factors for successful penile rehabilitation outcomes, which may aid the decision-making process in clinical practice.

In conclusion in the present study, African-Americans and patients with better preoperative SF were found as independent predictors to participate in penile rehabilitation, whereas pre- and postoperative higher risk disease, e.g. higher preoperative PSA concentration or PSM, trended toward avoiding penile rehabilitation. For rehabilitation success being older at surgery and adjuvant therapy, e.g. radiotherapy, ADT, or the combination, were identified as predictors for an unsuccessful outcome. Additionally, the primary surgeon showed a trend for association with successful outcomes of rehabilitation. Although treatment for prostate cancer involving RP needs to be balanced between oncological and HRQL outcomes, we postulate that those patients who have favourable factors could have good expectations, whereas patients who have risk factors need to be carefully informed and encouraged to participate in the penile rehabilitation programme. In addition, clinicians, especially primary surgeons, could lead patients toward successful outcomes, if appropriate surgical techniques and penile rehabilitation in the post-RP period are provided.

CONFLICT OF INTEREST

None declared.

Ancillary