Prospective analysis of penile length changes after radical prostatectomy


  • Boback M. Berookhim,

    1. Male Sexual and Reproductive Medicine Program, Urology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
    Search for more papers by this author
  • Christian J. Nelson,

    1. Department of Psychiatry and Behavioral Sciences, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
    Search for more papers by this author
  • Brian Kunzel,

    1. Male Sexual and Reproductive Medicine Program, Urology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
    Search for more papers by this author
  • John P. Mulhall,

    Corresponding author
    1. Male Sexual and Reproductive Medicine Program, Urology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
    • Correspondence: John P. Mulhall, Memorial Sloan Kettering Cancer Center, 1275 York Avenue, Box 435, New York, NY 10065, USA.


    Search for more papers by this author
  • Joseph B. Narus

    1. Male Sexual and Reproductive Medicine Program, Urology Service, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
    Search for more papers by this author



  • To assess the impact of radical prostatectomy (RP) on penile dimensions.
  • To assess the impact of phosphodiesterase-5 inhibitor (PDE5i) use on penile length changes.

Patients and Methods

  • Men undergoing RP were enrolled in this prospective study before surgery. Demographic, clinical and PDE5i frequency-of-use data were collected.
  • Erectile function was measured using the erectile function domain (EFD) of the International Index of Erectile Function.
  • A single evaluator measured stretched flaccid penile length (SFPL) before RP, and at 2 and 6 months after RP.
  • Repeated measures analysis was used to test differences in SFPL between timepoints. Pearson correlation was used for univariate analyses and multiple regression was used for multivariable analysis.


  • A total of 118 patients were evaluated at baseline, with 76 and 63 patients evaluated at 2 and 6 months, respectively.
  • At 2 months, there was a 2.4-mm mean decrease in SFPL, while at 6 months there was no significant difference.
  • At 6 months, those subjects who took a daily PDE5i had no SFPL loss (n = 36, 1±6.7 mm gain, P = 0.37 compared with baseline), while those subjects who did not consistently take a PDE5i had SFPL loss (n = 27, 4.4±6.6 mm loss, P < 0.002 compared with baseline).
  • In multivariable analysis, PDE5i use at 6 months and 6-month EFD score without on-demand PDE5i were significant predictors of 6-month SFPL loss, suggesting that an increase in these variables leads to SFPL preservation.


  • In this rigorously conducted prospective study of SFPL changes after RP, there was evidence of SFPL loss at 2 months, but not at 6 months after RP.
  • PDE5i use moderated SFPL loss, with patients who regularly used PDE5i having no loss in SFPL.


Significant research has been directed towards the efficacy of treatment options for prostate cancer and its associated effects on patient quality of life, particularly with respect to sexual function and urinary continence [1, 2]. After radical prostatectomy (RP), clinicians often hear complaints from patients about loss of penile length and girth [3]. Loss of penile length is a source of patient bother and distress [3] and this phenomenon has been assessed by a total of eight studies to date [4-11]. The results of these studies have been mixed, with five studies showing a loss of penile length [4, 5, 7, 9, 10], and three showing no changes in length [6, 8, 11]. Measurement of stretched penile length has previously shown a strong correlation with erect length [12]. Methodological weaknesses and study size limit the generalizability of some of the data presented. The objective of the present study was to describe the effect of RP on penile dimension changes. Our secondary objective was to assess the impact of phosphodiesterase type 5 inhibitor (PDE5i) use on penile length changes in patients undergoing RP.

Patients and Methods

Patient Population

After institutional review board approval had been obtained, all patients undergoing RP at our institution, irrespective of approach, were eligible for enrolment in this prospective trial. Inclusion criteria were clinically localized prostate cancer and being sexually active (penetrative or non-penetrative sexual activity). Patients with a preoperative or postoperative history of pelvic radiation therapy or androgen deprivation therapy were excluded from the study. In addition, those men with a preoperative history of Peyronie's disease or previous surgical placement of a penile prosthesis were excluded. Patients gave informed consent for the study before undergoing surgery. Demographics, including patient comorbidities, body mass index, clinical and operative data, were collected.

Penile Length Measurement

Prospective data on stretched flaccid penile length (SFPL) were gathered; a single evaluator measured SFPL before RP and at 2 and 6 months after RP. Using a ruler, measurement of SFPL was performed from the pubic bone to the coronal sulcus with maximum extension of the phallus. Care was taken to ensure maximum compression of the pre-pubic fat pad so that the ruler was positioned firmly against the pubic bone. The extent of intra-observer variance in SFPL recording was measured in eight patients. SFPL was measured with the patient lying in the supine position with the penis placed at a 90° angle to the body and the temperature of the room maintained at ≥22 °C. Patients with SFPL loss within the defined intra-observer variance were recorded as having no change in SFPL.

Erectile Function Assessment

We measured erectile function (EF) using the erectile function domain (EFD) of the International Index of Erectile Function (IIEF). This was administered by a male nurse practitioner at each office visit before penile length measurement was performed.

PDE5i Use

After RP, patients at our institution are offered penile rehabilitation for the first postoperative year, using daily low-dose PDE5i (sildenafil 25 mg), and twice-weekly full-dose PDE5i (sildenafil 100 mg) or intracavernosal therapy to induce penetration hardness erection. Patients quantified their use of PDE5i using a five-point scale: 1: PDE5i use almost never/never; 2: much less than half the time; 3: about half the time; 4: much more than half the time; and 5: almost always/always. A subgroup analysis was conducted of two separate PDE5i groups: Group A included those patients always using PDE5i, and Group B included all other PDE5i use options (whether PDE5i were used never, seldom or frequently). The evaluator was not blinded to the PDE5i use of each patient.

Nerve-Sparing Score

Nerve-sparing was graded intra-operatively by the surgeon using a four-point nerve-sparing score (NSS) for each neurovascular bundle with a combined score range of 2–8: 1: complete preservation of the neurovascular bundle; 2: near-complete preservation, 3: partial resection of a neurovascular bundle; and 4: complete resection.

Statistical Analysis

Mean (sd) values are reported. Repeated measures analysis was used to test differences in mean SFPL between timepoints. The chi-squared test was used to compare differences in percentages. When determining predictors of SFPL (continuous variable, reported in mm), Pearson correlations were used for univariate analyses and multiple regression was used for multivariable analysis. The predictors tested in the model included: age, number of vascular risk factors, nerve-sparing status, baseline EFD scores, PDE5i use at 6 months, 6-month EFD scores without PDE5i, and 6-month EFD scores while using PDE5i. A P value <0.05 was considered to indicate statistical significance. The statistical package used was IBM (Armonk, NY, USA) spss v.21.


Patient Population

A total of 118 patients were evaluated at baseline, with 76 patients evaluated at 2-month follow-up and 63 patients at 6-month follow-up. Patients’ mean (sd) age was 58 (7) years. Patient characteristics are shown in Table 1. In all, 41% of patients underwent open RP; the remaining patients underwent either robot-assisted laparoscopic or straight laparoscopic RP. No patients had Peyronie's disease at baseline or the 2-month follow-up, and 2% (n = 1/63) had Peyronie's disease at 6 months. The mean (sd) baseline EFD score was 25 (7). At 2 months, the mean (sd) EFD score for Group B patients was 7 (7), and for Group A patients it was 12 (9). At 6 months, the Group B mean (sd) EFD score was 7 (7), and for Group A it was 13 (9).

Table 1. Patient characteristics
  1. aFive patients were missing NSS score.
  2. bKnown risk factors for cardiovascular disease, including, hypertension, diabetes, hypercholesterolaemia and history of smoking.
Number of patients118
Mean (sd) patient age, years58 (7)
Mean (sd) Gleason score7 (1)
Mean (sd) preoperative PSA, ng/mL7.6 (14)
NSSa, n (%) 
253 (47)
318 (16)
422 (19)
511 (10)
65 (4)
72 (2)
82 (2)
Vascular risk factorsb, % 

Penile Length

Intra-observer variance was evaluated in eight patients and the mean intra-observer variance in SFPL was found to be 3 mm. Mean (sd) baseline SFPL was 114 (17) mm. At the 2-month assessment, for the 76 patients evaluated, 49% of patients had no SFPL loss, while 5% had at least a 10-mm loss (Table 2). At this timepoint, there was a mean decrease of 2.4 mm in SFPL, within the observed 3-mm intra-observer variation (113 [18] vs 111 [17] mm, n = 76, P < 0.01). At 6 months, for the 63 patients evaluated, 41% had no SFPL loss, while 4% had at least a 10-mm loss. There was no difference in mean (sd) SFPL compared with baseline (114 [17] vs 113 [17] mm, n = 63, P = 0.15).

Table 2. Patients with loss in SFPL
SFPL change2 Months % (n)6 Months % (n)
No change49 (37)41 (36)
≤5 mm30 (23)20 (13)
6–10 mm4 (3)13 (8)
11–15 mm1 (1)2 (1)
>15 mm4 (3)2 (1)

PDE5i Use

PDE5i were ‘always’ used in 7% of men before RP, 70% of men at 2 months, and 57% of men at 6 months. The mean weekly frequency of use in Group A patients was scored as 5 across all timepoints, as these patients always used PDE5i; for Group B patients these frequencies were 2 ± 1.7 and 1.8 ± 1.4. At 2 months after RP, both PDE5i groups showed a mean SFPL loss, while at 6 months, only Group B patients (those patients not consistently using daily PDE5i) were found to have SFPL loss (Table 3). There was no significant difference in the percentage of men with SFPL loss at 2 months between Groups A or B (38 vs 44%, respectively, P = 0.64). At 6 months, fewer patients had SFPL loss in Group A than in Group B (25 vs 52%, P = 0.03).

Table 3. SFPL changes at 2 and 6 months stratified by PDE5i use
 Number of patientsMean change in length (sd), mmP (compared with baseline values)
2 Months   
Group A532.4 (6.8) loss0.01
Group B232.5 (6.0) loss0.06
6 Months   
Group A361 (6.7) gain0.37
Group B274.4 (6.6) loss<0.002

Predicting SFPL Loss

We examined factors predicting SFPL loss at 6 months. On univariate analysis, PDE5i use at 6 months (r = -0.41, P = 0.001) and 6-month EFD score without the use of on-demand medication (r = −0.34, P = 0.05) were both negatively associated with SFPL loss at 6 months, indicating that more frequent use of PDE5i and higher scores on the EFD were both associated with less SFPL loss. Age (r = 0.10, P = 0.45), nerve-sparing status (r = 0.01, P = 0.96), baseline IIEF-EFD scores (r = −0.02, P = 0.91), 6-month EFD scores while using PDE5i (r = −0.16, P = 0.28), number of vascular risk factors (r = 0.18, P = 0.17), and type of surgery (r = 0.11, P = 0.41; open RP vs robot-assisted/laparoscopic RP) were not associated with SFPL loss at 6 months. In multivariable analysis, when entering all the variables tested in univariate analysis (significant and nonsignificant), PDE5i use at 6 months (β = −0.54, P = 0.002) and 6-month EFD score without the use of on-demand medication (β = −0.35, P = 0.05) remained significant predictors (in the negative direction) of SFPL loss. The 6-month EFD scores while using on-demand PDE5i were not entered into the multivariable model owing to high multicollinearity with 6-month EFD scores not using on-demand medication (r = 0.62, P < 0.001 [Table 4]).

Table 4. Multivariate analysis predicting SFPL loss at 6 months
Baseline IIEF-EFD score0.090.550.59
Number of vascular risk factors0.010.040.97
Type of RP (open vs robot-assisted/laparsocopic)−0.03−0.180.86
PDE5i use−0.54−3.430.002
6-month IIEF-EFD (without on-demand medication)−0.35−2.070.05


The present study shows a decrease in mean SFPL at 2 months after RP, but this was minimal and almost certainly not clinically meaningful, with no significant changes noted at 6 months postoperatively. A small percentage (∼5%) of patients, however, were found to have a decrease in SFPL of >1 cm at the 2- and 6-month follow-ups, indicating that some men will ultimately have a significant degree of penile shortening. The present study is among the first to show preservation of SFPL in patients using daily PDE5i as compared with those patients not using regular PDE5i. PDE5i use in a non-daily fashion, and 6-month EFD score without on-demand PDE5i were predictors of SFPL loss in our multivariate analysis. This suggests a possible protective effect of PDE5i-based penile rehabilitation in terms of SFPL, although this cannot be definitively concluded from these data.

Changes in penile dimensions after open RP have been described in several previous studies, in which results have varied [4-10]. Numerous methodological issues are at play in penile length measurement studies [13]. Critical to the measurement of SFPL are the two points used to define length, especially with regard to the proximal aspect of the penis. Using the standard reference point of the pubic bone helps to account for changes in the size of the pre-pubic fat pad over the course of a study follow-up, especially when compared with the pubo-penile skin junction reference point. In addition, the coronal sulcus provides a fixed reference, as measurement to the tip of the penis is hampered usually by a hand grasping the glans. Inter-observer variability plays an important role in some studies, where different practitioners have performed preoperative and postoperative measurements, potentially accounting for some changes in length. Given the relationship of the ambient environment on SFPL measurements, the room temperature should also be controlled.

Among the studies performed describing penile dimensions, disparate results are noted with respect to method of measurement, length of follow-up, and SPFL loss. Six studies do not refer to penile rehabilitation protocols when describing methods or the population [4-8, 10]. Of these, five of the studies are noted to measure penile length from the pubo-penile skin junction to the tip of the glans [4-8], with only a single study using the pubic bone as the proximal point of reference in measurement [10]. Data as to the number of observers making measurements is lacking in all the studies. Study populations range from 31 to 126 patients, with most studies reporting up to 1 year of patient follow-up. Mean SFPL length loss is reported in three of these studies [7, 8, 10], with results varying from no mean loss at 6 months [8] to 2-cm loss at 1-year follow-up [7]. In the single study with follow-up >1 year, with significant attrition to the follow-up population from an initial population of 105 patients, 5-year data showed full recovery of SFPL among the 14 patients evaluated [10].

Two of the studies performed described a penile rehabilitation protocol being used, but data as to adherence to the protocol is lacking [9, 11]. Engel et al. [9] reported results on changes in penile length in 127 patients undergoing robot-assisted laparoscopic RP. Patients were randomized in a 2:1 manner to receive nightly intra-urethral alprostadil or oral sildenafil as part of a penile rehabilitation protocol. SFPL was measured from the pubic bone to the coronal sulcus with a semi-rigid ruler, by one of two examiners. They reported a statistically significant 6.4-mm decrease in SFPL at 1-month follow-up, and a 3.2-mm and 2.6-mm decrease at the 3- and 6-month evaluations, respectively. Evaluations at 9 (105 patients), 10 (99 patients) and 11 (94 patients) months did not show decreases in SFPL from baseline. Changes in SFPL were significantly associated with return of sexual function only at the 11-month follow-up.

In a small prospective study by a group of Turkish investigators, the role of tadalafil rehabilitation (20 mg given three times per week) after RP and its effect on SFPL when compared with patients not receiving rehabilitation were analysed [11]. SFPL was measured from the pubo-penile skin junction to the tip of the glans. Thirty-three patients did not receive rehabilitation (Group 1), and 32 were in the tadalafil group (Group 2). While statistical data are sparse, patients in Group 1 had a statistically significant 8-mm SFPL loss at 3 months, a 3-mm loss (P = nonsignificant) between 3 and 6 months, and a 1-mm loss between 6 and 12 months (P = nonsignificant). No significant loss of length was noted in Group 2, although they reported mean SFPL loss of 5 mm at 3 months, 2 mm between 3 and 6 months, and 1 mm between 6 and 12 months. IIEF scores between the groups were no different at 1 year. The authors conclude that tadalafil rehabilitation is effective in preserving penile length after RP.

In addition, two studies have looked at the use of a vacuum erection device (VED) and its effect on stretched penile length in patients after RP. In one study, 39 men had pre- and postoperative measurements of penile length, with VED use initiated the day after catheter removal and continued for 90 days [14]. Thirty-six men used the device at least 50% of possible days, with one patient having a decrease in stretched penile length of >1 cm. This compared with two of the three men with VED use <50% of possible days having a stretched penile length loss of >1 cm, leading the authors to conclude that early intervention with daily use of a VED preserved penile health and length after surgery. Another study confirmed these results, where 28 patients undergoing RP were randomized to using a VED at 1 month postoperatively vs 6 months postoperatively [15]. Postoperative IIEF-EFD scores were significantly higher in the early intervention group at the 3- and 6-month follow-up. At the last follow-up (mean 9.5 months), 12% of the early intervention group had a >2-cm loss, compared with 45% in the late group (P = 0.044). None of our patients used a VED after RP.

Multiple theories have been proposed for the pathophysiology of penile length changes in the post-RP patient population. Mulhall [16] has described a well-supported hypothesis dividing penile length changes into those in the early postoperative period and those in the delayed period. This postulates that cavernous nerve injury results in sympathetic overdrive, causing smooth muscle contraction and a hypertonic retracted penis during the early phase, usually most pronounced in the first 3 to 6 months after surgery. Animal data support this hypothesis, with a study in dogs suggesting that sympathetic nerve fibres are among the first to regenerate after operative neural injury, a concept known as competitive sprouting [17]. This may help to describe the initial loss of length noted in our study.

With time, delayed structural changes in cavernosal smooth muscle and possibly the tunica albuginea may play a more significant role in penile length changes [18]. Studies have shown that cavernosal nerve injury leads to apoptotic changes in corporal smooth muscle [19, 20]. The rat model has also shown lower penile wet weights in cavernosal nerve injury, with the most significant decreases noted in those subjects with bilateral nerve injury [20]. Cavernosal nerve injury has also been associated with collagenization of the corporal smooth muscle in rats [18]. Penile fibrotic changes have been observed in 41% of men with ED after RP, leading to decreased elasticity and probably decreased stretched length [21]. Furthermore, corpora cavernosa tissue samples obtained 2 and 12 months after RP in men have been shown to have quantitative and qualitative decreases in elastic and smooth muscle fibres, and increases in collagen fibre content [22]. We found that PDE5i had a protective effect on SFPL, which may be explained by data from a small study of patients volunteering for penile biopsy at the time of RP and at 6-month follow-up. Patients were randomized to receive 50 mg of sildenafil daily or 100 mg of sildenafil every other day, and no difference in pre- and postoperative smooth muscle content was found on histological examination of the total study population [23].

Our study has some limitations. While initial enrolment was fairly robust, a number of patients were lost to follow-up, although it is not likely this would have affected the results noted in our study. Loss to follow-up is generally attributable to a myriad of reasons, and it is possible that those patients failing to return were those patients frustrated with a poor recovery of erectile function. It should be noted, however, that all patients seen in our practice are extensively counselled and have realistic expectations as to the natural history of recovery of erectile function after RP, and that those with poorer responses generally are more apt to attend follow-up given their understanding of penile rehabilitation. Additionally, some patients may have been lost to follow-up by opting to follow only with their surgeon, although it is unclear that this would bias the results noted in either direction. Our short follow-up period, although similar to the majority of previously published articles, may not allow potential delayed fibrotic changes over the long term to be accounted for. The lack of a control group of patients not receiving any PDE5i makes it impossible to draw firm conclusions as to the role of PDE5i therapy in the prevention of SFPL loss; however, the absence of a randomized controlled trial to assess the length loss phenomenon 13 years after the introduction of sildenafil and 14 years after the first report on penile length loss after RP, is indicative of the challenges and difficulties in conducting such trials.

Penile length, although measured with a standardized method removing the pre-pubic fat pad as a factor, was not measured in the erect state, which may have provided different results. While statistically significant, the clinical relevance of the 1-cm loss in SFPL that we describe is unclear and requires further study.

As expected from any study including measurements, there is the possibility of inherent measurement error. Despite our attempts to address all potential confounders and to standardize SFPL measurement, we note that 12% of patients at 2 months (none with a gain of >1 cm) and 22% of patients at 6 months had a gain in penile length (with two patients with a gain >1 cm). Additionally, the person evaluating SFPL was not blinded to the patients use of PDE5i, potentially biasing our results. Another potential factor leading to the preservation of SFPL, the number of erections achieved per week, was not evaluated in our patient diary.

As strengths, however, the number of patients followed up at 6 months was greater than most of the previously published analyses, and we used a validated erectile function instrument. The patients in our population are probably representative of the population of patients undergoing RP at large, with a mean age of 58 years (compared with 61 years in a patient population of nearly 90 000 patients in one study), with diverse adherence to a PDE5i regimen and varying NSS [24]. Furthermore, PDE5i use was monitored using a diary, and importantly, all patients had SFPL measurement performed by a single evaluator with minimal intra-observer variance.

In conclusion, the present study describes the changes noted in SFPL after RP. Men were noted to have early loss of mean SFPL, which recovered back to baseline by 6 months postoperatively; however, ∼5% of men had a persistent SFPL loss >1 cm at 6-month follow-up. A regimen of daily PDE5i use predicted preservation of SFPL at the 6-month timepoint. Patients should be counselled as to the risks and natural history of penile length loss before undergoing RP.

Conflict of Interest

None declared.


radical prostatectomy


phosphodiesterase type 5 inhibitor


erectile function


erectile function domain


International Index of Erectile Function


stretched flaccid penile length


nerve-sparing score


vacuum erection device