Fascrs.org

Randomized, Double-Blind, Placebo-Controlled Trial of Sildenafil (Viagra®)for Erectile Dysfunction After RectalExcision for Cancer and InflammatoryBowel DiseaseIan Lindsey, F.R.A.C.S., Bruce George, M.S., Michael Kettlewell, F.R.C.S.,Neil Mortensen, M.D.
From the Department of Colorectal Surgery, John Radcliffe Hospital, Oxford, United Kingdom PURPOSE: Controlled trials have demonstrated the efficacy (mean difference, 29.5; 95 percent confidence interval, 15.8 to of sildenafil for “mixed etiology” erectile dysfunction, but 43.2; P ϭ 0.003) from precrossover baseline scores. Seven (50 this may not be the case if there is underlying pelvic para- percent) of 14 patients on sildenafil compared with 4 (22 sympathetic nerve damage. We aimed to determine the percent) of 18 on placebo experienced side effects (differ- efficacy of sildenafil after rectal excision for rectal cancer ence, 28 percent; 95 percent confidence interval, Ϫ4.4 to 60.4 and inflammatory bowel disease. METHODS: Patients with percent; P ϭ 0.14), 91 percent of which were mild and well erectile dysfunction after rectal excision were randomly tolerated. CONCLUSION: Sildenafil completely reverses or sat- assigned in a double-blind manner to sildenafil or placebo isfactorily improves postproctectomy erectile dysfunction in groups. After unblinding, placebo patients crossed over to 79 percent of patients. Side effects are usually mild and well open sildenafil. Primary end points were improvement in tolerated. The damage incurred by the pelvic nerves after erectile function on a global efficacy question and erectile proctectomy, less profound than after prostatectomy, is likely function questionnaire scores. Secondary end points were to result in a partial parasympathetic nerve lesion. [Key words: frequency and severity of side effects. RESULTS: Thirty-two Impotence; Rectum; Phosphodiesterase inhibitors] patients were randomly assigned, and two dropped out Lindsey I, George B, Kettlewell M, Mortensen N. Random- before randomization. Fourteen received sildenafil, and 18 ized, double-blind, placebo-controlled trial of sildenafil (Vi- received placebo. Eleven (79 percent) of 14 responded to agra®) for erectile dysfunction after rectal excision for can- sildenafil, on global efficacy assessment, compared with 3 cer and inflammatory bowel disease. Dis Colon Rectum (17 percent) of 18 taking placebo (mean difference, 61.9 percent; 95 percent confidence interval, 34.4 to 89.4 per-cent; P ϭ 0.0009). Sildenafil improved both erectile func-tion domain scores (mean difference, 13.3; 95 percent con-fidence interval, 7.9 to 18.7; P ϭ 0.0001) and total T here has been a significant change in the treat- ment of erectile dysfunction or impotence with International Index of Erectile Function scores (mean dif-ference, 30.6; 95 percent confidence interval, 18.7 to 42.6; the recent development of the selective type-5 phos- P Ͻ 0.0001) from pretreatment baseline scores. Placebo did phodiesterase inhibitor, sildenafil (Viagra®, Pfizer not produce improvement in either erectile function (mean Ltd., Groton, CT). Sildenafil is administered in tablet difference, 1.7; 95 percent confidence interval, Ϫ0.8 to 4.2; form and is well tolerated, with clear advantages over P ϭ 0.16) or total International Index of Erectile Functionscores (mean difference, 5; 95 percent confidence interval, previous impotence treatments in this respect. A re- Ϫ1.1 to 11.1; P ϭ 0.1). Ten (100 percent) of 10 crossover cent double-blind placebo-controlled trial has dem- patients not responding to placebo did respond to sildenafil onstrated an improvement in erectile function in (difference, 100 percent; P Ͻ 0.0001). Sildenafil improvedboth erectile function domain scores (mean difference, males with erectile dysfunction of various etiologies 16.8; 95 percent confidence interval, 9.7 to 24; P ϭ 0.002) or so-called “mixed” erectile dysfunction.1 and total International Index of Erectile Function scores Erectile dysfunction is a special concern for colorectal surgeons and their patients, occurring with a mean fre- Ian Lindsey was supported by the Colorectal Research Fund. Silde- quency of 43 percent in patients after rectal excision for nafil and placebo were supplied by Pfizer Ltd.
Winner of the New York Society of Colon and Rectal Surgeons cancer and 3 percent in patients with inflammatory bowel disease (IBD),2 although rates of up to 100 per- Read at the meeting of The American Society of Colon and RectalSurgeons, San Diego, California, June 2 to 7, 2001.
cent3 and 25 percent,4 respectively, are reported. Al- Address reprint requests to Mr. Lindsey: Department of Colorectal though there may be a contribution made by psycho- Surgery, John Radcliffe Hospital, Headley Way, Headington, Ox-ford, OX3 9DU, United Kingdom.
logical factors, such as the presence of a stoma and the fear of recurrent cancer, there is an increasing recogni- tion of the major influence of damage to pelvic para- tion in identically labeled containers supplied by Pfizer sympathetic nerves, particularly the cavernous nerves Ltd. and consecutive patient prescribing by the trial that lie between the rectum and prostate immediately surgeon. The dispensing pharmacist and trial surgeon outside the mesorectal plane, which are especially vul- were blinded to the randomization code. The primary nerable during anterior rectal dissection.5–7 end point was successful improvement in erectile func- The response to sildenafil after radical prostatec- tion (either sufficient for sexual intercourse or returning tomy in which cavernous nerve damage is causal is the patient to normal presurgery functional state). This modest, ranging from 50 to 80 percent.8–10 Animal was assessed by a global efficacy question (Did the studies have shown that bilateral cavernous nerve therapy lead to an improvement in your erections? Table transection ablates erectile function despite the pres- 2) as well as improvement in both the erectile function ence of sildenafil.11 In the absence of a neural signal, domain score and the total score of the IIEF (Table 1).
sildenafil cannot induce cavernous smooth muscle The secondary end points were frequency and severity relaxation and thus an erection. We aimed to deter- of side effects. Analysis was conducted on an intention- mine whether sildenafil would improve erectile func- tion in male patients with complete or partial erectile Assuming the study hypothesis that the response to dysfunction after rectal excision for rectal cancer and sildenafil would be at the lower end of that seen after IBD. Our hypothesis was that the response would be radical prostatectomy, response rates were anticipated similarly modest as that seen after radical prostatec- at 50 percent for active medication and 15 percent for tomy, in which cavernous nerve damage is causal.
placebo. At a power of 80 percent, 25 patients were Subgroup analysis on the influence of disease group required in each arm to show a significant difference at and severity of erectile dysfunction was planned.
P Յ 0.05. The trial was stopped after interim analysis of32 patients because of the highly significant difference in the response rate between active medication and pla-cebo (more than 3 standard deviations, P ϭ 0.0009); it Inclusion criteria were male patients with exclu- was considered unethical to continue randomizing pa- sively postoperative erectile dysfunction identified on tients. Statistical analysis was conducted using the Stu- a prospectively compiled IBD and colorectal cancer dent’s t-test for paired IIEF questionnaire data and Fish- database by a validated, self-reporting impotence er’s exact test for differences in percentages (Statview, questionnaire, the International Index of Erectile Function (IIEF; Table 1). Exclusion criteria were pre- The trial protocol was as follows. Baseline pretreat- operative erectile dysfunction and medical contrain- ment erectile function was established by deriving a dications to sildenafil. Ethical approval for the trial total IIEF score and an erectile function domain score was granted by the Central Oxfordshire Research Eth- from the IIEF questionnaire. The IIEF consists of 15 ics Committee (COREC Project No. C99.092). Full in- scaled questions relating to 5 domains of sexual function formed consent was obtained from each patient.
agreed on by an international panel of erectile dysfunc- A one-to-one randomization code was derived by the tion specialists12 (Table 1). It has been shown to be pharmacy trials coordinator using a computer-generated psychometrically robust with high internal consisten- random number sequence. Double blinding was cy.12 Severity of erectile dysfunction was classified ac- achieved by using identical active and placebo medica- cording to a validated grading of the erectile functiondomain scores of the IIEF,13 and possible domain scoresrange from 6 to 30. Complete erectile dysfunction was defined as an erectile function domain score less than International Index of Erectile Function (IIEF) 10, and partial erectile dysfunction a score less than 17 Patients were randomly assigned to receive at least four weeks of primary therapy, either sildenafil or pla- cebo control. Those aged 65 and older were com- menced on 25 mg and those younger than 65, 50 mg; the maximal dose was 100 mg. During therapy, compli-ance and efficacy was checked. It was ensured patients had allowed one hour for the medication to take effect EVIDENCE FOR PARTIAL PARASYMPATHETIC NERVE LESION Global Efficacy Question Response Options* No improvement in erectile function at all Improvement in erectile function, but unsatisfactory and insufficient for Satisfactory improvement in erectile function, sufficient for sexual intercourse but not returning to normal presurgery state Improvement in erectile function to normal presurgery state * Question: Did the therapy lead to an improvement in your erections? and that initiation of sexual activity was required for the cancer, five by abdominoperineal resection with per- drug to exert its influence on erectile activity. If normal manent end colostomy and seven by low anterior erectile function had not returned, the dose was in- resection with colorectal or coloanal anastomosis.
creased until maximally effective or clearly ineffective, Twenty patients underwent proctectomy for IBD: 16 and that highest final dose was used to evaluate efficacy.
had an ileal pouch-anal anastomosis and 2 underwent After four weeks, each patient was asked to clearly a proctocolectomy with permanent ileostomy for ul- identify improvement in erectile function or otherwise.
cerative colitis. Two patients underwent proctocolec- Each patient answered the global efficacy question and tomy with permanent ileostomy for Crohn’ s disease.
completed the IIEF questionnaire again to document Eleven patients had a stoma at the time of random- changes in posttherapy erectile function. Responders ization: five a permanent colostomy after abdomino- and nonresponders were defined by answers to the perineal resection, four a permanent ileostomy after global efficacy question (Table 2). Side effects were proctocolectomy for both Crohn’ s disease and ulcer- documented and graded for severity: nil; mild (well ative colitis, and two a temporary loop ileostomy tolerated, therapy continued); and significant (poorly diverting their ileal pouch-anal anastomosis.
Median age at surgery was 52.8 (interquartile range, Unblinding occurred once all efficacy data had 42.1-59.2) years. Time elapsed from surgery to trial been collected. Those receiving placebo were given a randomization was a median of 5.6 (interquartile letter to take to their local physician, who com- range, 3.3–7.7) years. Median age at randomization menced crossover sildenafil. These patients were not was 58.7 (interquartile range, 49.4–65.3) years. The blind to this subsequent therapy, and they were as- median age at surgery for rectal cancer was 58 (inter- sessed as previously by repeat global efficacy ques- quartile range, 50.1–66) years and the median age at randomization 65.1 (interquartile range, 52.9–68.4)years. The median age of IBD patients was about seven years lower than that for rectal cancer patients,with median age at surgery of 50.7 (interquartile range, 37–56.1) years and median age at randomiza-tion of 58 (interquartile range, 42.8–64) years.
Forty-three patients with erectile dysfunction were The sildenafil and placebo patient groups were initially considered for the trial. Nine patients were well matched for age. The median age at randomiza- excluded because they had medical contraindications tion was 59.5 (interquartile range, 51.1–64.9) years in to sildenafil or the patients considered themselves too the sildenafil group and 58.7 (interquartile range, old for the trial. Thirty-four patients were randomly 49.4–67.5) years in the placebo group. The median assigned, and two patients dropped out before ran- age at surgery was 53.5 (interquartile range, 47.8– domization. One patient considered himself to be too 57.8) years in the sildenafil group and 50.7 (interquar- old, and one considered himself to be in too poor tile range, 40.5–62) years in the placebo group.
health as a consequence of pouchitis to participate.
Thirty-two patients completed the trial to unblinding.
Eighteen patients had total erectile dysfunction, and Global Efficacy Question. Of 32 patients complet- Twelve patients underwent proctectomy for rectal ing the trial, 14 received sildenafil and 18 placebo.
Eleven (78.6 percent) of 14 receiving sildenafil had a after sildenafil (mean difference, 13.3; 95 percent confi- significant improvement in erectile function com- dence interval, 7.9 to 18.7; P ϭ 0.0001) but not after pared with 3 (16.7 percent) of 18 receiving placebo placebo (mean difference, 1.7; 95 percent confidence (difference, 61.9 percent; 95 percent confidence inter- interval, Ϫ0.8 to 4.2; P ϭ 0.16; Table 3).
val, 34.4 to 89.4 percent; P ϭ 0.0009 Fisher’s exacttest; Table 3).
IIEF Questionnaire Scores. The sildenafil and pla- cebo patient groups were well matched for pretreat- Global Efficacy Question. Ten patients who origi- ment erectile dysfunction. There was no difference nally received placebo openly crossed over to silde- between groups in either mean baseline total IIEF nafil. No patient had improvement in erectile function score (mean difference, 2.8; 95 percent confidence with placebo, but all 10 had improvement with silde- interval, Ϫ11 to 16.5; P ϭ 0.67, paired t-test) or mean nafil, a 100 percent response (difference, 100 percent; baseline erectile function domain score (mean differ- P Ͻ 0.0001, Fisher’s exact test; Table 3).
ence, 0.2; 95 percent confidence interval, Ϫ5.3 to 5.7; IIEF Questionnaire Scores. There was a significant improvement in mean erectile function domain score There was a significant improvement in mean total over baseline after crossover sildenafil (mean differ- IIEF score over baseline after sildenafil (mean differ- ence, 16.8; 95 percent confidence interval, 9.7 to 24; P ence, 30.6; 95 percent confidence interval, 18.7 to 42.6; ϭ 0.002, paired t-test) not seen after placebo (mean P Ͻ 0.0001) but not after placebo (mean difference, 5; difference, 0.7; 95 percent confidence interval, Ϫ1 to 95 percent confidence interval, Ϫ1.1 to 11.1; P ϭ 0.10; 2.4; P ϭ 0.36; Table 3). Similarly, there was a signifi- Table 3). Similarly, there was a significant improvement cant improvement in mean total IIEF score over base- in mean erectile function domain score over baseline line after crossover sildenafil (mean difference, 29.5; IIEF ϭ total International Index of Erectile Function score (all five domains); EFD ϭ International Index of Erectile Function (erectile function domain score).
* P values between sildenafil and placebo.
† P values for sildenafil (after) over baseline (before).
‡ P values for placebo (after) over baseline (before).
EVIDENCE FOR PARTIAL PARASYMPATHETIC NERVE LESION 95 percent confidence interval, 15.8 to 43.2; P ϭ experiencing side effects between the 50-mg (46.1 0.003), which was not seen after placebo (mean dif- percent) and the 100-mg (63.6 percent) groups, the ference, 1.3; 95 percent confidence interval, Ϫ2.1 to difference did not reach statistical significance (differ- ence, 17.5 percent; 95 percent confidence interval,Ϫ21.8 to 56.8 percent; P ϭ 0.44).
Subanalysis by disease group and severity of impo- tence was performed on 24 patients receiving silde- Until recently, available treatments for erectile dys- nafil either as primary or crossover treatment. There function were limited to vacuum-constriction devices, was a small trend to a better response for IBD pa- intracavernosal injection or transurethral delivery of tients, with 14 (93.3 percent) of 15 patients respond- vasoactive agents, penile prosthetic implantation, or ing compared with 7 (77.8 percent) of 9 patients with vascular reconstructive surgery. With the exception of rectal cancer. However, this did not reach statistical penile prostheses, which have excellent satisfaction significance (difference, 15.5 percent; 95 percent con- and compliance rates, these treatment options have a fidence interval, Ϫ14.5 to 45.5 percent; P ϭ 0.53, variable range of efficacy and satisfaction rates, with a substantial drop-off in compliance rates over time.
There was a smaller trend to a better response to Sildenafil represents a significant advance in the care sildenafil for patients with partial (10 of 11, 90.9 per- of patients with erectile dysfunction, with simpler, cent) rather than complete (11 of 13, 84.6 percent) less invasive administration, good compliance, and impotence. However, again this did not reach statis- tical significance (difference, 6.3 percent; 95 percent It is of considerable interest to those performing confidence interval, Ϫ19.7 to 32.3 percent; P Ͼ 0.99).
pelvic surgery and their patients whether postproc-tectomy erectile dysfunction is likely to respond to sildenafil to the same extent as in the general silde-nafil trials.2 The results from these trials reflected large Study medication was generally well tolerated. Seven numbers of patients attending clinics with various (50 percent) of 14 sildenafil patients compared with 4 causes of erectile dysfunction, so-called “mixed” erec- (22 percent) of 18 placebo patients experienced side tile dysfunction, and thus the patient groups were effects, although the difference did not reach statistical quite heterogeneous. It is therefore difficult to deter- significance (difference, 28 percent; 95 percent confi- mine the response to sildenafil of various subgroups, dence interval, Ϫ4.4 to 60.4 percent; P ϭ 0.14, Fisher’s including those who had undergone pelvis surgery.
exact test; Table 4). Most (90 percent) side effects were No background information is available on the of mild severity and transient. The most commonly re- likely response of colorectal patients, but it might be ported was facial flushing, followed by headache. One expected to be similar to that reported in urology patient on placebo experienced severe diarrhea, but this patients after prostatectomy who similarly have a ma- did not reoccur on rechallenge. No patient discontinued jor element of parasympathetic nerve damage. Recent small studies reporting the use of sildenafil after rad- Of all patients receiving sildenafil (14 primary and ical prostatectomy suggest that response to sildenafil 10 crossover therapy), 13 were dosed up to 50 mg, is less impressive than in patients with mixed erectile whereas 11 were dosed up to 100 mg. Although there dysfunction, with the response greater after bilateral was a trend to a difference in proportions of patients nerve-sparing surgery (80 percent) than after unilat-eral (50 percent) or non-nerve-sparing surgery (0 per- cent).8–10 Animal studies have shown that bilateral cavernous nerve transection ablates, whereas unilat- eral nerve sparing preserves, erectile function in thepresence of sildenafil.11 There is a theoretical explanation for this poor response after parasympathetic nerve damage (Fig.
1). Nitric oxide released from endothelium or cavern- ous nerves acts through guanylate cyclase to trigger a and in fact these patients have an excellent chance ofimprovement in their erectile function with sildenafil.
These findings also suggest that erectile dysfunc- tion after rectal excision either is caused by a largelypartial parasympathetic neural injury or is mainly ex-plained by psychological factors. The strong cross-over response rate, however, is more consistent with Figure 1. Mechanism of action of sildenafil. GMP ϭ
a complex partial parasympathetic neural injury.
guanosine monophosphate; cGMP ϭ cyclic guanosinemonophosphate; PDE ϭ phosphodiesterase.
1. Goldstein I, Lue T, Padma-Nathan H, Rosen RC, Steers rise in cyclic guanosine monophosphate (cGMP), WD, Wicker PA. Oral sildenafil in the treatment of which is accompanied by a reduction of cytosolic erectile dysfunction. Sildenafil Study Group. N Engl calcium, causing smooth muscle relaxation. Type-5 phosphodiesterase inhibitors decrease the break- 2. Lindsey I, Guy RJ, Warren BF, Mortensen NJ. Anatomy down of cGMP to GMP, thus increasing cGMP levels, of Denonvilliers’ fascia and pelvic nerves, impotence, which relaxes cavernous smooth muscle, increasing and implications for the colorectal surgeon. Br J Surg intracavernous blood flow and pressure. However, in the absence of a neural signal or nitric oxide, silde- 3. Weinstein M, Roberts M. Sexual potency following sur- nafil does not induce cavernous smooth muscle re- gery for rectal carcinoma. Ann Surg 1977;185:295–300.
laxation. In vivo injection of sildenafil into the penis 4. Watts JM, de Dombal FT, Goligher JC. Long-term com- fails to induce any spontaneous erectile response or a plications and prognosis following major surgery forulcerative colitis. Br J Surg 1966;53:1014 –23.
5. Lepor H, Gregerman M, Crosby R, Mostafi FK, Walsh PC.
Our hypothesis that improvement in erectile function Precise localization of the autonomic nerves from the would be modest, similar to that seen with postprostate- pelvic plexus to the corpora cavernosa: a detailed anatom- ctomy erectile dysfunction, proved incorrect. It may be ical study of the adult male pelvis. J Urol 1985;133:207–12.
that the damage incurred by the parasympathetic nerves 6. Walsh PC, Lepor H, Eggleston JC. Radical prostatectomy during rectal excision may be less profound than during with preservation of sexual function: anatomical and prostatectomy, owing to the closer proximity of the pathological considerations. Prostate 1983;4:473– 85.
nerves to the surgical plane during prostatectomy.2 Al- 7. Lindsey I, George BD, Kettlewell MG, Mortensen NJ. Im- ternatively, other factors, possibly psychological, may be potence after mesorectal and close rectal dissection for responsible for erectile dysfunction after rectal excision.
inflammatory bowel disease. Dis Colon Rectum 2001;44: However, the low efficacy with placebo and high silde- nafil crossover efficacy would support the notion of a 8. Lowentritt BH, Scardino PT, Miles BJ, et al. Sildenafil citrate after radical retropubic prostatectomy. J Urol If the better response to sildenafil reflects less dam- 9. Zippe CD, Kedia AW, Kedia K, Nelson DR, Agarwal A.
age to the cavernous nerves, then this is an important Treatment of erectile dysfunction after radical prostatec- finding. It would highlight the need for more aware- tomy with sildenafil citrate (Viagra). Urology 1998;52: ness of the precise anatomy of the cavernous nerves and their relationship to the anterior plane of rectal 10. Marks LS, Duda C, Dorey FJ, et al. Treatment of erectile dissection,2 as well as development of new but prac- dysfunction with sildenafil. Urology 1999;53:19 –24.
tical nerve-sparing surgical techniques that would still 11. Carrier S, Zvara P, Nunes L, et al. Regeneration of nitric oxide synthetase-containing nerves after cavernousnerve neurotomy in the rat. J Urol 1995;153:1722–7.
12. Rosen RC, Riley A, Wagner G, Osterloh IH, Kirkpatrick J, Mishra A. The International Index of Erectile Function(IIEF): a multidimensional scale for assessment of erec- Erectile dysfunction after rectal excision for rectal tile dysfunction. Urology 1997;49:822–30.
cancer and IBD is completely reversed or satisfacto- 13. Cappelleri JC, Rosen RC, Smith MD, Mishra A, Osterloh rily improved in 79 percent of patients. Side effects IH. Diagnostic evaluation of the erectile function do- are not uncommon but are mild, well tolerated, and main of the International Index of Erectile Function.
often transient. We have refuted our study hypothesis,

Source: http://www.fascrs.org/gedownload!/2002%20Randomized_Double_Blind_Placebo_Controlled_Trial.pdf?item_id=6966003&version_id=6966004