A prospective evaluation of efficacy and compliance with a multistep treatment approach for erectile dysfunction in patients after non-nerve sparing radical prostatectomy


Paolo Gontero, Clinica Urologica, Università del Piemonte Orientale, Novara, Italy.
email: gontero@med.unipmn.it


Associate Editor

Michael G. Wyllie

Editorial Board

Ian Eardley, UK

Jean Fourcroy, USA

Sidney Glina, Brazil

Julia Heiman, USA

Chris McMahon, Australia

Bob Millar, UK

Alvaro Morales, Canada

Michael Perelman, USA

Marcel Waldinger, Netherlands


To assess the response rate to different erectile aids in a consecutive series of patients treated with non-nerve sparing radical prostatectomy (NNSRP).


Ninety-four potent men were counselled about the different treatment options to restore an assisted erection before they had NNSRP. They were invited to participate in a multiphase protocol involving the sequential use of different erectile aids which aimed at restoring erectile function after surgery. The first proposed treatment was oral apomorphine sublingual. Patients with a positive response to the 1-item overall efficacy question and a minimum score of 3 in both question 3 and 4 of the International Index of Erectile Function were considered responders to oral pharmacotherapy. Treatment with sildenafil was then suggested to those not responding. If patients did not respond to oral pharmacotherapy a trial with a vacuum erectile device was offered; those not responding to this were then offered intracavernosal injection therapy with prostaglandin-E alone as the first option, followed by a mixture of vasoactive agents if needed. In those in whom injections also failed, a penile implant was recommended. At the 1-year follow-up visit all patients were offered a second trial with oral therapy regardless of the treatment currently in use.


Seventy-six patients entered the protocol; there was no response to apomorphine. Five of 59 (8%) patients responded to sildenafil when they first used it at a mean of 7 months after NNSRP, while there were three additional responders in 22 patients who tried it for a second time a year later. Of patients achieving at least a complete tumescence sufficient for vaginal penetration, 52% and 60% were considered responders to the vacuum device and intracavernosal injections, respectively. Overall, 44% of patients enrolled in the protocol chose to use an erectile aid for at-home use. At the 1-year follow-up, only 20% of patients were still using an erectile aid, including two who had had a penile implant.


Up to 10% of patients may achieve a clinically significant erection with sildenafil after NNSRP, but 80% will not be using any erectile aid at 1 year after surgery. In the present study protocol the proposed erectile aids were largely inadequate for treating the permanent erectile dysfunction that follows NNSRP.


(non)(nerve-sparing) radical prostatectomy


erectile dysfunction


intracavernosal injection


International Index of (Erectile Function)




overall efficacy question




visual analogue scale


Radical prostatectomy (RP) is commonly used for organ-confined prostate cancer [1] and has been developed with the addition of techniques to preserve the neurovascular bundles, allowing the prostate to be completely excised while maintaining sexual function [2]. However, a significant proportion of men undergoing a nerve-sparing procedure fail to recover spontaneous erectile function. Also, both oncological and surgical factors mean that many RP are not nerve-sparing (NNSRP). Patients with bilaterally transected neurovascular bundles are not expected to regain any spontaneous erection [3], and permanent erectile dysfunction (ED) remains an important issue in patients with prostate cancer, with a significant impact on their health-related quality of life [4].

Despite the many reports on nerve-sparing series, few have assessed the use of the available erectile aids as permanent treatments after NNSRP. A recent American survey reported that many patients after NNSRP use no treatment for ED, and that the few reports of treatment failed to provide sexual quality-of-life measures comparable with men regaining spontaneous erectile function [5]. By contrast, up to 85% of patients predominantly having NNSRP were still using erectile aids effectively at 1 year after surgery in a series from Israel [6]. To our knowledge no study has prospectively evaluated different treatment alternatives for ED in a selected series of patients after NNSRP.

Studies of the efficacy of oral treatment after NNSRP are based on small series reporting a response rate to sildenafil of none [7] to 15%[8]. Intracavernosal pharmacotherapy still remains the main treatment option for patients who fail to recover spontaneous erections after RP, and high response rates have been reported using mixtures of vasoactive agents [6,9]. The present study was designed to evaluate the efficacy and the patients’ compliance with currently available treatments for ED. A consecutive series of patients who had RP with NNS intent were invited to participate in a multistep trial which offered a sequence of treatments for ED, starting from the less-invasive oral tablets and the vacuum pump, and then moving to second-line therapies, represented by intracavernosal injections (ICIs), and finally to the third-line prosthetic therapy, as outlined in the European Association of Urology guidelines [10].


From January 2001 to October 2002, 152 patients treated at the first author's institution had a NNSRP. Eligibility for surgery was based on a local protocol adopted at our institutions since 1998, and involving extensive indications for retropubic RP, i.e. all patients with prostate cancer confirmed by biopsy, a negative bone scan and a life-expectancy of >10 years were offered surgery, regardless of PSA level at presentation, Gleason score and DRE findings.

Indications for excising nerve bundles were one or more of the following: bilateral positive biopsies, Gleason score >7, a DRE suggestive of extracapsular tumour extension, age > 70 years, PSA >20 ng/mL. In the absence of precise published guidelines on when the nerves should be spared, these criteria represented the personal view of the members of the institutions where the study was conducted. All patients were requested to complete the International Index of Erectile Function (IIEF) questionnaire before RP [11]; 94 reporting an erectile function (EF) domain score on the IIEF of ≥ 25 [12] received early counselling after RP on the different treatment options to restore an assisted erection, including penile prostheses. Based on early results from our recently published data [13], patients were advised that any non-surgical treatment would be of maximum benefit if started within 3 months after surgery. Patients were then given the opportunity to attend a specialized andrology outpatient clinic at any time during the follow-up.

Patients presenting to the andrology clinic at any time after RP (Visit 1) were invited to take part in a multiphase protocol involving the sequential use of different erectile aids which aimed to restore erectile function, provided they had not recovered spontaneous erectile function.

Because oral apomorphine sublingual (SL) had been recently released on the Italian market, the patients’ request for this was significant and we subsequently adopted this drug as the first proposed treatment in the protocol. Four sample doses of apomorphine SL 3 mg were supplied to be taken ‘as required’, not exceeding one dose in any 8-h period. Patients were instructed to take one tablet of apomorphine SL 15 min before attempting sexual intercourse and to allow it to dissolve after placing it under the tongue. Treatment response was assessed with the one-item overall efficacy question (OEQ), ‘Has the treatment you have been taking over the past two to four weeks improved your erections?’, and with the EF domain questionnaire. Those with a positive response to the OEQ and with a minimum mean score of 3 above baseline for both questions 3 and 4 of the IIEF were considered responders.

Those not responding were then offered treatment with sildenafil and assessed using the same criteria for response as for apomorphine. A trial with a vacuum erection device was offered as the first-line option to those patients who had no response to both oral agents. Patients achieving at least a grade III erection, i.e. a complete tumescence sufficient for vaginal penetration, were considered responders. ICI using prostaglandin-E (PGE), followed by a mixture of vasoactive agents if necessary, was regarded as a second-line treatment if the vacuum device failed. If ICIs also failed to produce at least a grade III erection, a penile implant was recommended. Patients were allowed to change to a different treatment at any time.

Data on pathological stage, Gleason score and concomitant adjuvant treatment were also recorded during the first visit. The current status of continence was categorized as: no significant incontinence, mild incontinence (less than one moist pad per day) and marked incontinence (one or more wet pad per day).

Visit 2 was scheduled for 4 weeks afterward; those not responding to apomorphine as documented by the OEQ and the EF domain score were instructed to take a tablet of sildenafil 1 h before sexual intercourse, provided they had no contraindications to phosphodiesterase type 5 inhibitors [10], and were supplied with four doses of 100 mg for a 4-week home trial. Patients were considered for evaluation if they had taken at least two of the four tablets dispensed for any oral treatment.

When sildenafil also failed to produce a positive response, patients were instructed in the use of the vacuum erection device at Visit 3 (Somablue, Atlanta, USA). The erection obtained after placing the constriction band was categorized into four grades by the same doctor, according to objective examination criteria as previously described [14]. To assess the discomfort experienced during the use of the device patients were asked to score the pain on a 10-cm visual analogue scale (VAS), where ‘0’ meant no pain and ‘10’ the maximum pain [15]. All patients were supplied with a device (electric or manual, depending on their preference) and an informative videotape for a 4-week home trial if desired.

When returning the vacuum erectile device at Visit 4, information about buying a device was provided to interested patients, while those not responding were offered ICI; a 20-µg injection with PGE was administered initially and subsequent doses were individually titrated depending on the erectile response obtained after the first injection. When there was an unsatisfactory response to a full dose of PGE (20 µg), a mixture of papaverine 10–20 mg and phentolamine 1–2 mg was added to PGE. Patients with a positive response, i.e. at least a grade III artificial erection, were taught self-injection for home use, and those not responding were recommended a penile implant.

Follow-up visits or telephone calls were scheduled at 6 and 12 months to assess any discontinuation of treatment. A further trial with both apomorphine SL and sildenafil was offered 1 year after visit 1 and the results assessed with the OEQ and the EF domain score.

A paired t-test for matched variables was used to compare the IIEF and the VAS scores before and after therapy, with the chi-square test to compare proportions (i.e. responders vs. nonresponders).


Of the 94 men potent before RP and counselled on the early use of erectile aids, 77 (mean age 65.6 years, sd 6.7) attended the andrology clinic. The mean (sd) time between RP and self-referral to the andrology clinic was 6.5 (3.49) months. One patient who had RP 3 months earlier recovered spontaneous erections and was thus excluded from the study. The 76 remaining patients reported a mean (sd) EF domain score after RP of 4.64 (6.7) and were eligible for the protocol. The patients’ characteristics (pathological stage, Gleason score and adjuvant treatment) are listed in Table 1.

Table 1.  The patients’ characteristics, including pathological stage, Gleason score and adjuvant treatment, for the 76 men potent before NNSRP and referred to the andrology clinic
CharacteristicN (%)
Pathological stage
T250 (66)
T3a14 (18)
T3b 9 (12)
T4 3 (4)
N+ 8 (10)
N020 (26)
Nx48 (63)
Gleason score after RP
4–643 (57)
717 (22)
8–916 (21)
Adjuvant treatment
Radiotherapy10 (13)
Hormone therapy 11 (14)
Radiotherapy + hormone therapy 2 (2.5)

Figure 1 shows the flow chart relative to oral therapy. No patient failed to take the minimum of four tablets required for evaluation. Two of the 76 patients (2.6%) taking apomorphine SL answered ‘yes’ to the OEQ although the improvement for both Q3 and Q4 was below the level for them to be considered responders. Eight of 59 patients (13.5%) receiving sildenafil answered ‘yes’ to the OEQ but only five (8%) were considered responders. Three of them were discharged from the clinic with a prescription and the remaining two requested to proceed to the next step of the protocol. Neither of the two patients reporting marginal benefit after apomorphine responded to sildenafil. The overall mean (sd) EF domain score after apomorphine was 4.65 (1.97), and comparable with that at baseline (P = 0.978). The improvement of mean EF domain score after sildenafil for all patients, at 6.14 (5.08), was statistically (P = 0.01) but not clinically significant.

Figure 1.

Flow chart of 95 men potent before RP who were counselled on early erectile treatment after NNSRP and tried oral treatment.

Side-effects were reported by nine of 75 patients (12%) after apomorphine SL (nausea in six, vomiting in two and headache in one) and by seven of 59 (12%) after sildenafil (headache in all).

Figure 2 shows the flow chart followed by the 74 patients who did not respond to oral therapy; 17% refused a trial with the vacuum erectile device and opted for ICI with PGE. Overall, 98.5% of patients received a full dose of PGE, including all but one of the 33 responding to the vacuum device. The PGE response rate (60%) did not significantly differ from that for the vacuum device (52.5%) (chi square, P = 0.764). The mean VAS score recorded after using the vacuum device, at 2.57 (0.78), was significantly lower than after PGE, at 2.91 (0.96) (P = 0.02). Three patients developed priapism after ICI with PGE. There was a statistically significant correlation between PGE response status and vacuum response status (chi square, P = 0.054). Data on treatment response rate and the percentage of patients who chose an erectile aid for at-home use are summarized in Table 2.

Figure 2.

The flow chart for patients who had NNSRP, in whom oral treatment failed and who were then offered a trial with vacuum erectile device and ICIs.

Table 2.  Response rate, initial choice for at-home use and the 1-year discontinuation rate of the proposed treatment options. Values are n (%) for totals > 20
OptionRespondersChoosing at-home useDiscontinuing
  • *

    Patients originally not responding to oral treatment;

  • †Patients not responding to PGE alone.

Apomorphine 0/76
initial 5/59 (8) 3/5 0/3
1-year* 3/22 (14) 2/3
Vacuum32/61 (53)13/32 (40) 6/13
ICI PGE44/73 (60)15/44 (34)10/15
ICI cocktail 4/5 2/4 2/2

There were three deaths during the 1-year follow-up (two cancer-related, one other causes). All the remaining patients (73) were reviewed either at the clinic or by telephone. Two patients received a penile implant during the first year of follow-up in our institution (one had discontinued PGE, the other was not satisfied with the vacuum device). No patient recovered spontaneous erectile function. The numbers of patients who discontinued treatment after 1 year are shown in Table 2; overall, 17 (23%) were still using some erectile aid at 1 year. Apomorphine was not offered at the 1-year follow-up because of its poor efficacy. Twenty-two patients initially not responding to oral treatment attempted a second trial with sildenafil; three (14%) then responded, according to the previous criteria, and two decided to use the treatment at home.


Sexual bother is becoming a key issue for patients with prostate cancer who choose radical surgery. In a recent study [4] sexual dysfunction was an independent determinant of worse general health-related quality of life after primary treatment for prostate cancer. A retrospective analysis showed that men undergoing NNSRP reported worse quality-of-life scores for sexual function than patients treated with NSRP [5]. Previous questionnaire-based surveys indicated that most patients with ED after RP used no erectile aids, but the reason for this could not be identified [5,16]. The present study focused on a selected series of patients with normal sexual function before RP who then had NNSRP and received standard counselling on the current view of how to optimize sexual response using erectile aids [13,17]. Of the counselled patients, 82% attended the ED clinic and entered the multiphase treatment protocol, a finding which suggests that a large proportion of men potent before RP are interested in regaining normal sexual function afterward.

None of the two types of oral compounds currently used to improve sexual function is expected to be effective if the cavernosal nerves are transected at the time of surgery. Apomorphine SL cannot enhance the neurological signalling to sexual stimuli delivered from the CNS to the periphery if the cavernosal nerves are interrupted [18]. Similarly, no amplification of the nitric oxide cascade should occur after giving a phosphodiesterase-5 inhibitor in the absence of nitric oxide release from intact cavernosal nerve fibres. Despite these physiopathological considerations, the present study design was supported by the lack of sufficient clinical evidence to discourage oral treatment as a first-line option in this patient category, as outlined by the European Guidelines [10]. Also, there is increasing evidence that a few patients may either recover spontaneous erectile function or respond to oral treatment after NNSRP, implying the existence of unrecognized residual nerve tissue or a non-neurogenic mechanism. Zagaja et al.[7] reported that sildenafil was ineffective in all 33 patients who had bilateral transection of neurovascular bundles during RP. Conversely Zippe et al.[8] found four responders among 26 NNSRP patients treated with sildenafil. In the present study all patients accepted the initial treatment with apomorphine SL, an oral agent that significantly improved erectile function in a series of patients with ED of various causes [19]. Apomorphine SL became available when the present study started and many patients were motivated to try it because of its rapid onset of action, lack of cardiovascular contraindications, and safety profile. To our knowledge there is no previous report on its use after NNSRP; however, it was ineffective in this series.

Fifty-nine patients who did not respond positively to apomorphine reported a 8% response rate to sildenafil. This not responding were offered an additional trial during the 1-year follow-up visit, in line with previous observations that sildenafil may not be effective in the first 9 months after surgery [7], and that the regeneration of partially damaged nerves may take up to 2 years [20]. An additional 13% response rate was documented in this small series. The present data suggest that, after NNSRP, treatment with a phosphodiesterase-5 inhibitor may be worthwhile, particularly at 1 year after RP.

The European guidelines suggest that the use of a vacuum erectile device should be regarded as a first-line option when oral treatment fails, and to consider ICI as a second-line treatment [10]. High response rates and long-term efficacy were reported for both of these erectile aids [17]. There was a 60% response rate for PGE, the only drug licensed for ICI, in the present series; 85% success rates were reported in other studies when using a mixture of vasoactive agents [6,21].

A randomized study excluding patients after RP showed a trend favouring ICI because there were fewer side-effects and greater patient satisfaction [22]. Conversely, in the present analysis of this unrandomized series, ICI was judged more painful than the vacuum device. The proportion of patients choosing ICI for domestic use was as low as 16% and comparable with the vacuum device. By contrast, Baniel et al.[6] found that all of the 85% of patients in that series who responded to a mixture of vasoactive drugs used the treatment at home. In the present study almost half the patients discontinued after 1 year for both the vacuum device and ICI. Although these values may depend on several variables not addressed by the current study, there are similarities with previous American series [17,23]. Overall, only 44% of patients enrolled in the protocol chose to use an erectile aid at home. Again, these data are comparable with previous retrospective studies on unselected patients [16].

European Guidelines state that for patients who fail pharmacological therapy or who prefer a permanent solution to their problem, surgical implantation of a prosthesis may be considered [10]. All the present patients received counselling before RP on the excellent satisfaction rate after a penile implant for ED [24]. Those not responding to all the proposed treatments were offered the option of having an inflatable prosthesis implanted at no cost, under the Italian national health system. However, only two patients had a prosthesis implanted within a year from NNSRP. This seems to suggest that the currently available erectile aids are largely inadequate for treating the permanent ED that results from NNSRP.

However, the possibility of a design bias should be considered when interpreting the present results. The overall poor response rate to oral treatments raises doubts about the appropriateness of the sequence of treatments adopted. The use of a drug such us apomorphine, which was then ineffective in most patients with ED, as a first choice in a rehabilitation protocol may have produced a detrimental effect on the overall compliance with treatment. The subsequent administration of another oral compound, sildenafil, with a 90% failure rate, could have compromised patients’ expectations of any subsequent therapy proposed. The primary objective of the study was to assess the efficacy of different erectile aids administered in a sequence that reflected the recommendations in the European guidelines. From the results it appears that oral compounds may not be an appropriate first choice after NNSRP. In our current practice these patients are now offered ICI or the vacuum device as the first option, with a change to oral sildenafil suggested no earlier than 1 year from surgery.

In conclusion, this study provided information that may be useful in the sexual counselling of patients before and after NNSRP. The recovery of spontaneous erectile function is rare after NNSRP; up to 10% of patients may benefit from the use of sildenafil, while apomorphine SL had no role in treating this type of erectile failure. The overall response rate to the available nonsurgical treatments may be influenced by individual motivation but it should not be expected to be more than 50–60%. This study indicates that half of patients are likely to discontinue treatment within a year. Treatment of ED after NNSRP remains a difficult and frustrating task for the urologist. Patients assessed in the present series appeared to be motivated to regain erectile function, as they were sexually active before surgery and chose to seek treatment. Surprisingly, only 20% of the patients were still using an erectile aid after 1 year, a situation that warrants future efforts to seek more effective treatments. Different protocols not using oral compounds as the initial treatment may prove less frustrating for patients and thus produce less disappointing results.


None declared.