Sexual bother and function after radical prostatectomy: predictors of sexual bother recovery in men despite persistent post-operative sexual dysfunction

Authors


Correspondence:

Masaki Kimura, Division of Urology, Department of Surgery, Duke University Medical Center, 207 Research Dr, 152 Jones building, Durham, NC 27710, USA. E-mail: masaki.kimura228@gmail.com

Summary

Changes in sexual bother (SB) following radical prostatectomy (RP) negatively affect health-related quality of life (HRQoL) of prostate cancer survivors. However, post-operative SB tends to be neglected whereas sexual function (SF) is thoroughly assessed in clinical practice and few studies have focused on and evaluated patients’ SB. We retrospectively reviewed 2 345 consecutive patients who underwent RP between 2001 and 2009 at a single institution. SF and SB were assessed using Expanded Prostate Cancer Index Composite (EPIC) questionnaires. We stratified our cohort by SB recovery and post-operative SF status, including a subset of men who recovered SB despite persistent post-RP sexual dysfunction. Multivariable logistic regression analyses were conducted to identify factors for men who have SB recovery. Of 319 eligible patients, 133 (41.7%) recovered their SB at a mean of 20 months after RP. Among the 133 men who demonstrated SB recovery, 109 had post-operative sexual dysfunction. Patients with SB recovery despite post-RP sexual dysfunction were more likely to be old (= 0.004), to have higher clinical T stage (< 0.001), to have more non-nerve-sparing RP (< 0.001), to have lower pre-operative EPIC-SF/SB scores (< 0.001), to have more extracapsular extension (= 0.031) and to be PDE5i non-users after surgery (< 0.001). In multivariable analysis, predictors for this subset were lower comorbidity (OR 0.62, = 0.043), higher clinical cancer stage (OR 2.35, = 0.026), worse pre-operative SF (OR 0.98, = 0.010), SB (OR 0.98, < 0.010) and no PDE5i use (OR 0.37, = 0.002); age was not related (OR 0.99, = 0.555). As SB can influence patients’ overall HRQoL, expectations of SB recovery should be provided to patients in the same way that SF recovery is presented. This study may help clinicians to discuss SB with patients and assess their potential for SB recovery following RP.

Introduction

Radical prostatectomy (RP) is the one of well-established treatment option for localized prostate cancer with excellent disease control, but post-RP sexual dysfunction is a morbidity of major concern to both physicians and patients (Litwin et al., 1999). Besides attempting to cure the patient from prostate cancer, surgeons generally aim to preserve sexual function (SF). To this end, numerous efforts have focused on improving surgical techniques and modifying post-operative treatment regimens; these have resulted in improved post-RP SF compared with decades ago (Mulhall et al., 2005; Ayyathurai et al., 2008; Marien et al., 2009; Muller et al., 2009).

Sexual bother (SB) caused by erectile dysfunction (ED) plays an important role in overall health-related quality of life (HRQoL)(Althof, 2002). Among prostate cancer patients, there have been several studies suggesting the importance and significant impact of SB when evaluating HRQoL following RP (Sanda et al., 2008; Stensvold et al., 2012). In fact, SF levels vary among prostate cancer patients prior to RP (Salonia et al., 2006). Therefore, patients who are not sexually active and have lower SF are expected to have a lower risk of worsening SB after surgery by clinicians. However, even though some prostate cancer survivors are not sexually active, some of them could be concerned about their SF, and this would influence their perceived SB. Taking account the wide variations in an individual's perception of sexual HRQoL, SB after RP remains still unclear. Given that SB is multifactorial and consists of physiological and psychological elements, SB may be more widely influenced not only by SF itself but also by numerous factors such as demographics, pre-operative clinical characteristics and post-operative pathological outcomes.

To date, the knowledge base regarding changes of SB after RP is not sufficient, and few studies have focused and evaluated the specific subset of patients who report recovery to baseline SB despite post-operative persistent sexual dysfunction. Although SB is difficult to analyse and interpret as it includes subjective and psychological issues, we sought to examine variation of SB recovery after RP. In this study, we investigated the demographics, as well as clinical and pathological characteristics of patients who have undergone RP based on stratification by both SB recovery and post-operative SF. We then attempted to identify characteristics and predictors for patients who had SB recovery despite persistent sexual dysfunction following RP.

Materials and methods

Study population

This study was approved by the institutional review board and waivers for consent or informed consents were obtained from all patients. We retrospectively analysed data on 2 345 consecutive patients who underwent RP, including radical retropubic prostatectomy (RRP), perineal radical prostatectomy (PRP), or robotic-assisted laparoscopic prostatectomy (RALP) between 2001 and 2009 at our institution. Inclusion criteria are patients who had both pre- and post-operative SF/SB data with more than 12-month follow-up, and we excluded patients who did not have data of SF/SB status prior to surgery, resulting in a total of 319 enrolled patients. The process of decision-making with regard to use of phosphodiesterase type 5 inhibitor (PDE5i) was based on the patient's choice after post-operative discussions about treatment options and expectations with the primary surgeon. For those patients who agreed to our penile rehabilitation program, we recommended using PDE5i three times per week or on demand, and most of them used a vacuum erection device daily without the constriction band once urinary continence was achieved. For classification of pre-operative comorbidity status, we employed Charlson Comorbidity Index which is widely used measure with several validations that weighs comorbid conditions by the strength of their association with 1-year mortality (Charlson et al., 1987, 1994).

Sexual function and bother assessments

SF and SB were evaluated using the Expanded Prostate Cancer Index Composite (EPIC) questionnaire, a validated HRQoL instrument specific for men with prostate cancer (Wei et al., 2000). 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-categorized into two subscales: a SF subscale and a SB subscale. Response options for each EPIC item form a Likert scale, and each 5-item score is transformed into a scale ranging from 0 to 100. Higher scores represent better HRQoL. All pre-operative EPIC scores were obtained at a clinic visit several weeks or months prior to surgery and kept under control of a third party data manager.

SB and SF recoveries were defined as ≥75% of patient's baseline EPIC scores (Hu et al., 2004; Haffner et al., 2005). Post-operative normal SF was defined as EPIC-SF ≥60 points, whereas poor SF (persistent sexual dysfunction) after RP was defined as post-operative EPIC-SF score lower than 60 according to previous studies (Schroeck et al., 2008; Kimura et al., 2011). In this study, we stratified our cohort based on aforementioned definitions resulted in the following four groups: (i) men who recovered SB and had normal SF after surgery; (ii) men with SB recovery, but still with poor SF; (iii) men who did not have recovery of SB, but had normal SF after surgery; and (iv) men who did not recover SB and had poor SF after RP.

Statistical analysis

Demographics and clinical characteristics of subjects who demonstrated both recovery and persistence of SB were recorded. Categorical variables such as Charlson Comorbidity Index, biopsy Gleason sum, type of RP, nerve-sparing status, SF recovery after RP and PDE5i use after RP were compared using the Chi-squared test. Continuous variables, including age at surgery, year of surgery, pre-operative prostate-specific antigen (PSA), prostate weight and pre-operative EPIC-SF/SB score were compared by the Mann–Whitney test. Similarly, in subset analysis with comparison of the demographics, clinical and pathological characteristics among men stratified by SB recovery and post-operative SF, the aforementioned continuous variables were compared using the Kruskal–Wallis test, whereas categorical variables were compared using Fisher's exact test.

Multivariable logistic regression analyses were conducted to identify predictive factors for men who had SB recovery, as well as predictive factors for men who recovered SB despite post-operative sexual dysfunction. In the first logistic regression model, likelihood of SB recovery was the outcome variable. Pre-operative factors, including age at surgery (continuous), year of surgery (continuous), race (categorical), Charlson Comorbidity Index (categorical), pre-operative PSA (continuous after log-transformation), surgical type (categorical) and pre-operative EPIC-SF/SB scores (continuous) were considered potential predictors. Similarly, post-operative factors such as SF recovery (dichotomous; no vs. yes) and PDE5i use after RP (dichotomous; no vs. yes) were included as candidate predictors. In the second logistic regression analyses, we focused on men who recovered SB despite post-operative sexual dysfunction, and the likelihood of this result was the outcome variable. p < 0.05 was considered to be statistically significant. All statistical analyses were performed with STATA software version 11.2 (Stata Corporation, College Station, TX, USA).

Results

Baseline demographics and clinical characteristics of the cohort

Table 1 summarizes the comparison of baseline demographics and clinical characteristics between our entire 2 345 patient cohort and 319 eligible patients, showing there were no differences in age at surgery (= 0.750) or pre-operative PSA (= 0.934). However, our study cohort consisted of men with lower biochemical recurrence rate (9.0 vs. 13.4%; = 0.009), lower clinical T stage prior to surgery (= 0.002), lower proportion of African-Americans (10.0 vs. 16.3%; = 0.003), lower rate of RRP and a higher rate of PRP (< 0.001) and a higher rate of non-nerve-sparing surgeries as well as a lower rate of bilateral nerve-sparing surgeries (< 0.001) compared with the entire cohort. A comparison in baseline demographics and characteristics of the cohort stratified by SB recovery are summarized in Table 2. Of the 319 patients, 133 (41.7%) recovered their SB at a mean of 20 months after RP. In this comparison, we found that men who had SB recovery after RP were more likely to be higher clinical T stage (= 0.011), to undergo RRP (= 0.047), to have bilateral nerve sparing (< 0.001), to have lower pre-operative EPIC-SF and SB scores (< 0.001), to have SF recovery (< 0.001) and to be PDE5i non-user after surgery (= 0.035).

Table 1. Baseline demographics and clinical characteristics between total cohort and study cohort
VariableTotal cohortStudy cohort p-Value
  1. SF: sexual function, SB: sexual bother, PSA: prostate-specific antigen, RRP: radical retropubic prostatectomy, PRP: perineal radical prostatectomy, RALP: robotic-assisted laparoscopic prostatectomy, EPIC: Expanded Prostate Cancer Index Composite, PDE5i: phosphodiesterase type 5 inhibitors.

No. patients2345319 
Median age at surgery, (range)61 (34–79)60 (40–79)0.750
Median year of surgery, (range) 2006 (2001–2009)2004 (2001–2009)<0.001
Race, N (%)
Caucasian1891 (81.2)278 (87.2)0.008
African-American388 (16.7)32 (10.0)
Other50 (2.2)9 (2.8)
Median pre-operative PSA, (range)5.4 (0.2–92.0)5.6 (0.2–35.8)0.934
Median prostate weight, (range)41.5 (5.0–324)39.3 (15–219)0.048
Biopsy Gleason sum, N (%)
4–61371 (63.9)218 (68.3)0.083
7612 (28.5)87 (27.3)
8–9163 (7.6)14 (4.4)
Clinical T stage, N (%)
T11557 (78.0)268 (84.3)0.002
T2484 (23.6)47 (14.8)
T312 (0.6)3 (1.0)
Type of surgery, N (%)
RRP1134 (48.4)72 (22.6)<0.001
PRP322 (13.7)112 (35.1)
RALP889 (37.9)135 (42.3)
Nerve sparing, N (%)
None353 (19.1)104 (32.8)<0.001
Unilateral312 (16.9)69 (21.8)
Bilateral1185 (64.1)144 (45.4)
Table 2. Baseline demographics and clinical characteristics of the cohort stratified by SB recovery after surgery
VariableSB recoveredSB non-recovered P value
  1. SF: sexual function, SB: sexual bother, PSA: prostate-specific antigen, RRP: radical retropubic prostatectomy, PRP: perineal radical prostatectomy, RALP: robotic-assisted laparoscopic prostatectomy, EPIC: Expanded Prostate Cancer Index Composite, PDE5i: phosphodiesterase type 5 inhibitors.

No. patients133 186  
Median age at surgery, (range)60 (40–75)60 (42–79)0.904
Median year of surgery, (range) 2004 (2001–2009)2004 (2001–2008)0.214
Race, N (%)
Caucasian119 (87.2)158 (85.5)0.448
African-American10 (7.5)22 (11.8)
Other4 (3.0)5 (2.7)
Charlson comorbidity index, N (%)
0103 (77.7)129 (69.4)0.266
122 (16.5)40 (21.5)
≥28 (7.8)17 (9.1)
Median pre-operative PSA, (range)5.2 (0.2–35.8)5.6 (0.4–27.3)0.636
Median prostate weight, (range)38.0 (15–219)40.3 (16–103)0.327
Biopsy Gleason sum, N (%)
4–684 (63.2)134 (72.0)0.189
741 (30.8)46 (24. 7)
8–98 (6. 0)6 (3.3)
Clinical T stage, N (%)
T1103 (78.0)165 (88.7)0.011
T226 (19.7)21 (11.3)
T33 (2.3)0 (0)
Type of surgery, N (%)
RRP39 (29.3)33 (17.7)0.047
PRP41 (30.8)71 (38.2)
RALP53 (39.9)82 (44.1)
Nerve sparing, N (%)
None43 (32.3)61 (33.2)<0.001
Unilateral25 (18.8)44 (23.9)
Bilateral65 (48.9)79 (42.9)
Mean pre-operative EPIC-SF, (SD)44.0 (26.8)59.2 (24.3)<0.001
Mean pre-operative EPIC-SB, (SD)56.4 (32.8)76.9 (25.1)<0.001
SF recovery, N (%)49 (38.3)14 (7.53)<0.001
PDE5i use after surgery, N (%) 57 (42.9)102 (54.8)0.035

Baseline demographics and clinical characteristics of patients stratified by SB recovery and post-operative SF

We further stratified each group into two groups by post-operative SF status, resulting in four groups. Among these four groups, 109 (34.2%) patients experienced SB recovery despite post-RP sexual dysfunction (Table 3). This subset was more likely to be old (= 0.004), to have higher clinical T stage (p = 0.001), to have more non-nerve-sparing RP (< 0.001), to have lower pre-operative EPIC-SF/SB scores (< 0.001) and to be PDE5i non-user after surgery (< 0.001).

Table 3. Baseline demographics and clinical characteristics of patients stratified by SB recovery and postoperative SF
VariableSB recoveredSB non-recovered p-Value
Postoperative SF Normal SFPoor SFNormal SFPoor SF
  1. SF: sexual function, SB: sexual bother, PSA: prostate specific antigen, RRP: radical retropubic prostatectomy, PRP: perineal radical prostatectomy, RALP: robotic assisted laparoscopic prostatectomy, EPIC: Expanded Prostate Cancer Index Composite, PDE5i: phosphodiesterase type 5 inhibitors.

No. patients24109 4 182 
Median age at surgery, (range)54.5 (40–70)62 (43–73)58.5 (54–64)60 (42–79)0.004
Median year of surgery, (range) 2005 (2002–07)2004 (2001–09)2006 (2005–07)2004 (2001–08)0.099
Race, N (%)
Caucasian22 (91.7)97 (89.0)2 (50.0)157 (86.3)0.244
African-American1 (4.2)9 (8.3)2 (50.0)20 (11.0)
Other1 (4.2)3 (2.7)0 (0)5 (2.7)
Charlson comorbidity index, N (%)
018 (75.0)85 (78.0)3 (75.0)126 (69.2)0.736
14 (16.7)18 (16.5)1 (25.0)39 (21.5)
≥22 (8.3)6 (5.5)0 (0)17 (9.3)
Median preoperative PSA, (range)5.2 (0.2–35.8)5.2 (0.8–32.4)4.7 (3.0–5.6)5.6 (0.4–27.3)0.600
Median prostate weight, (range) 42.4 (26–68)40.7 (24–106)45.5 (36–56)40.1 (16–103)0.302
Biopsy Gleason sum, N (%)
4–616 (66.7)68 (62.4)3 (75.0)131 (72.0)0.519
76 (25.0)35 (32.1)1 (25.0)45 (24.7)
8–92 (8.3)6 (5.5)0 (0)6 (3.3)
Clinical T stage, N (%) 
T118 (75.0)85 (78.7)1 (25.0)164 (90.1)0.001
T26 (25.0)20 (18.5)3 (75.0)18 (9.9)
T30 (0)3 (2.8)0 (0)0 (0)
Type of surgery, N (%)
RRP13 (54.2)26 (23.9)3 (75.0)30 (16.5)0.001
PRP3 (12.5)38 (34.9)0 (0)71 (39.0)
RALP8 (33.3)45 (41.3)1 (25.0)81 (44.5)
Nerve sparing, N (%)
None1 (4.2)42 (38.5)0 (0)61 (33.9)0.005
Unilateral5 (20.8)20 (18.4)1 (25.0)43 (23.9)
Bilateral18 (75.0)47 (43.2)3 (75.0)76 (42.2)
Mean preoperative EPIC-SF, (SD)71.0 (18.7)38.0 (24.7)84.5 (14.9)58.6 (24.2)<0.001
Mean preoperative EPIC-SB, (SD)84.6 (24.2)50.2 (31.2)90.6 (14.9)76.6 (25.2)<0.001
SF recovery, N (%)24 (100)25 (24.0)3 (75.0)11 (6.0)<0.001
PDE5i use after surgery, N (%)18 (75.0)39 (35.8)4 (100.0)98 (53.9)<0.001

Pathological characteristics of patients stratified by SB recovery and post-operative SF

Regarding difference of pathological features, patients with SB recovery despite post-RP sexual dysfunction were more likely to have higher extracapsular extension (= 0.031) and tended to have higher seminal vesicle invasion (= 0.069). However, median prostate weight (= 0.302), median tumour involvement (= 0.402), median tumour volume (= 0.270), pathological Gleason sum (= 0.519), positive surgical margins (= 0.562), perineural invasion (= 0.969) and vascular invasions (= 0.962) did not differ among groups.

Predictive factors for men who have SB recovery

In multivariable analysis, we found several predictors for SB recovery after RP such as lower comorbidity scores (OR 0.58, = 0.030) and higher clinical stage (OR 2.47, = 0.016). With respect to pre- and post-operative SF and ED treatment, lower pre-operative EPIC-SB score (OR 0.98, < 0.001), SF recovery (OR 10.1, < 0.001) and non-use of PDE5i for post-RP sexual dysfunction (OR 0.38, = 0.003) were identified as predictors for SB recovery. In contrast, surgery type (RRP/PRP/RALP) did not reach statistical difference as a predictor for SB recovery after RP. Of note, patients with higher clinical T stage were more likely to undergo RRP, whereas patients with lower clinical T stage tended to undergo PRP/RALP (data not shown).

Predictive factors for men who recovered SB despite post-operative sexual dysfunction

A second analysis focusing on the subset of patients with SB recovery despite persistent post-RP sexual dysfunction was summarized in Table 4. Identified predictors were same as the first analysis, except that lower pre-operative EPIC-SF score (OR 0.98, = 0.010) was added as a predictor, whereas SF recovery (OR 1.27, = 0.528) was not associated.

Table 4. Multivariable logistic regression analysis predicting men who recovered SB despite persistent post-RP sexual dysfunction
VariableOdds ratio95% CIp-Value
  1. SF: sexual function, SB: sexual bother, PSA: prostate specific antigen, RRP: radical retropubic prostatectomy, PRP: perineal radical prostatectomy, RALP: robotic assisted laparoscopic prostatectomy, EPIC: Expanded Prostate Cancer Index Composite, PDE5i: phosphodiesterase type 5 inhibitors.

Age at surgery0.990.95–1.030.555
Year of surgery 1.050.85–1.290.664
Race
CaucasianReference
African-American0.730.26–2.010.540
Other0.570.11–3.090.518
Charlson comorbidity index0.620.38–0.950.043
Preoperative PSA1.070.65–1.750.804
Prostate weight 1.000.98–1.020.760
Clinical T stage2.351.11–4.980.026
Type of surgery
RRPReference
PRP0.720.23–2.310.585
RALP0.990.46–2.210.972
Preoperative EPIC-SF0.980.97-1.000.010
Preoperative EPIC-SB0.980.97–0.99<0.001
Recovery of SF1.270.61–2.650.528
PDE5i use after surgery0.370.20–0.700.002

Discussion

Sexual bother (SB) plays an important role in overall HRQoL, although post-operative SB tends to be neglected whereas SF is thoroughly assessed in clinical practice. As SB is difficult to analyse as it implies subjective and psychological issues, few studies have focused on and evaluated patients’ SB in post-RP setting. In general, SB seems to be linearly correlated with SF to some extent, which was reported in several articles as well as this study, demonstrating men who had SF recovery after surgery were more likely to recover post-operative SB (Table 2). However, little is known about characteristics of subset patients who recovered SB despite persistent post-RP sexual dysfunction. In this context, we are the first to elucidate specific characteristics for this subset of patients using the stratification by EPIC questionnaire and revealed these features include men with older age, higher clinical T stage, worse surgical pathological outcomes and men who do not use PDE5is. Especially, lower pre-operative EPIC-SF score was found to have significant association with post-operative SB recovery despite post-operative sexual dysfunction (Table 3), suggesting that sexually inactive men tend to recover SB easily than sexually active men. We speculate that this subset might not satisfy their SF pre-operatively, which leads to relatively higher SB, and even though they lose their SF post-surgery, changes of SB is smaller than that of men who satisfied their SF with lower pre-operative SB. This result can be supported by recent report of Steinsvik, et al., suggesting that patients who were sexually active before surgery were identified to be increased risk of experiencing post-operative SB (Steinsvik et al., 2012).

In multivariable analysis, several important predictive factors such as lower comorbidity, higher clinical cancer stage, worse pre-operative SF/SB and no PDE5i use were found in the subset of patients who recovered SB despite persistent post-RP sexual dysfunction. One noteworthy result from the multivariable and comparison analyses is that we found positive association between higher SB recovery and adverse prostate cancer features, for example, higher clinical stage/extracapsular extension. This finding might be rational, taking into account that patients have psychological stress about their cancer outcome rather than their SF in that situation. In such a situation, the priority of bother might be altered from HRQoL to their cancer control resulting in little changes of SB. We assume that this cancer control related-psychological element can considerably influence the change/recovery of SB after RP. Apart from psychological element, Gore et al. reported that severe SB was associated with higher serum PSA and Gleason sums more than eight, suggesting that patients with higher grade cancers may undergo more aggressive therapies such as non-nerve-sparing RP that may negatively affect post-RP SB recovery (Gore et al., 2010). In this study, we could not find substantial data showing the association between aggressive treatment and lower SB recovery after RP, however, we found a slightly increased rate of bilateral nerve-sparing surgery among men who had SB recovery compared with men with SB non-recovered. Although a total of 11 primary surgeons were involved in this study and preference of surgical procedure and post-RP follow-up can vary in each primary surgeon, we could not find any association between primary surgeon and SB recovery in uni- and multivariable analysis (data not shown). As we have been reported that difference in primary surgeons showed a trend for association with successful outcomes of SF recovery (Kimura et al., 2012), further study needed to explore the relationship between primary surgeon difference and SB/SF recovery after RP.

We also elucidated that pre-operative lower comorbidity scores prior to surgery were identified as strong predictors for SB recovery in multivariable analyses, which can contribute to further understanding about the role of pre-operative comorbidity status towards post-operative sexual health. Our results imply that better general health can accelerate SB recovery and perhaps reduce overall bother after surgery, although comparison of comorbidity status between men with SB recovered and non-recovered did not reach statistical significance (Table 2). Similarly, Le et al. reported that higher pre-operative comorbidities were associated with persistent SB after RP using repeated-measures regression models with adjusted multiple variables (Le et al., 2010). In our data, lower pre-operative comorbidities were not associated with higher recovery of SF after RP (data not shown). Taken together, we assume that pre-operative lower comorbidities status may aid patients’ psychological adaptation for post-RP sexual dysfunction, although this explanation requires further investigation.

Regarding post-RP ED treatment, approximately half of the study cohort used PDE5i based on the patient's choice after post-operative discussions about treatment options and expectations. However, we found that PDE5i use was negatively associated with SB recovery. This is somewhat consistent with a previous study in young patients, demonstrating that the use of sexual aids was associated with greater risk of severe worsening of SB (Wright et al., 2008). In our study, men who use PDE5i post-operatively have lower SB at baseline. However, SB of PDE5i users worsened more than those of PDE5i non-users, suggesting that PDE5i users had increased SB. Based on these results, one possible explanation could be that there is a mismatch between true SF recovery and patient expectations because expectations for better SF recovery and expense for ED treatment may be greater among these men who used PDE5i. Moreover, a recent study demonstrated no differences in SB scores between patients who were potent with and without the use of PDE5i (Punnen et al., 2011). Of note, men with lower interest in sexual intercourse may be less likely to use PDE5i and to complain about their post-operative SF resulting in lower SB. Considering the difficulty in defining a causative factor in this model, PDE5i use cannot be addressed as a cause of lower SB recovery. It was hard to interpret patient preference and expectations based on our methodology. Recently, Namiki et al. reported the relationships between pre-operative sexual desire and quality of life following radical prostatectomy (Namiki et al., 2012). In this study, authors divided patients into low sexual desire group and high sexual desire group, and conclude that higher sexual desire group experienced greater distress about post-RP sexual dysfunction, even though there was a statistically higher rate of post-operative PDE5i use in high sexual desire group. Although we were limited to evaluate causative factors in this study, this study seems to be consistent with our speculation regarding the mismatch between positive expectations for SF recovery and post-operative SF with PDE5i use.

There are limitations in this study. First, this study is retrospective in design, therefore the elimination of potential biases is difficult compared with a prospective study. Second, because of small sample size and the number of eligible patients is small compared with the entire consecutive cohort, only 319 were eligible from an original population of 2 345, this study may not be representative of the original population who underwent RP in our institution. Particularly, year of surgery, race, pre-operative clinical stage and surgical types were statistically different between the original cohort and the study cohort, which can impact SF/SB recovery statuses after RP. Third, only 28 patients (8.8%) had normal SF after treatment in this study. This low rate of normal SF after RP may result from relatively lower pre-operative SF in analysed patients, although no difference was seen in pre-operative EPIC-SF/SB scores compared with entire cohort. Fourth, four patients were defined as men who did not have recovery of SB, but had normal SF after surgery. This is somewhat paradoxical. In fact, EPIC-SF/SB scores of this group seem to be significantly higher than other groups and half of them had positive surgical margin. Although we could not find a good explanation for this group, these results imply that SB recovery may not be associated linearly with recovery status of post-operative SF in certain patients. Lastly, post-RP SB has been reported to continue improving with follow-up as long as 4 years (Parker et al., 2010). Considering SB may take longer than 20 months to stabilize, our follow-up period was not long enough to analyse stabilized SB status after RP. Although, we found a recovery rate of 41.7% at a mean follow-up of 20 months after RP, further validation is needed in a larger sample size with a longer follow-up to support the current results.

Conclusion

A certain subset of patients was identified to recover their baseline SB despite persistent sexual dysfunction after surgery. In multivariate analysis, this subset of patients has several factors, including lower pre-operative comorbidities, higher clinical T stage, worse pre-operative SF/SB and PDE5i non-users, whereas older age did not predict patient's SB recovery. In addition, this subset of patient has higher extracapsular extension and tended to have higher seminal vesicle invasion. Although one third of patients recovered SB even though they had poor SF after RP, we could not conclude that treatments for their post-operative sexual dysfunction are unnecessary. Given the significant correlation between post-operative SF and SB recovery, striving to improve post-operative SF through polishing surgical techniques and offering post-operative rehabilitation regimens should be emphasized. Considering changes of SB post-RP can be attributed to the adaptation for real SF, clinicians should make an effort to reduce the gap between expected outcomes and true SF by realistic patient counselling, which may lead to improve post-RP SB. As SB is quite influential for patient's overall HRQoL, expectations of SB recovery may be provided to patients in the same way that of SF recovery. Thus, this study can help clinicians to discuss risks of SB changes following RP in the pre- and post-operative setting.

Acknowledgements

This work was supported by a funding from the Committee for Urologic Research, Education, and Development (CURED) of Duke University.

Disclosures

There is no conflict of interest.

Author's contributions

M. K. & L. B. designed the research, L. B. & L. G. collected data, M. K. & L. B. analysed the data and wrote the draft, L. B. & T. P. & R. B. contributed the revising it critically, C. R. & C. D. & J. M. approved of the submitted and final versions of this article.

Ancillary