Oral sildenafil in the treatment of erectile dysfunction 

Authors Goldstein I, Lue T, Padma-Nathan H, Rosen R, Steers W, Wicker P (for the Sildenafil Study Group). 
Source New England Journal of Medicine. 338:1397-404. May 14, 1998. 
Institutions Multi-institutional in the United States; Pfizer Central Research. 
Support Pfizer. 


Erectile dysfunction is a common problem, increasing in prevalence with age. Available treatments include intracavernosal and intraurethral prostaglandin (alprostadil), penile prostheses and vacuum devices. Until the recent introduction of sildenafil (Viagra), no effective oral therapy was available. 

The pathway leading to penile erection includes the release of nitric oxide by nerve endings in the corpus cavernosum, in response to sexual stimulation. This, in turn, leads to the formation of cyclic GMP, which produces smooth muscle relaxation in the corpora cavernosa. Sildenafil, which reaches its peak effect about one hour after oral ingestion and has a half life of 3-5 hours, selectively inhibits the metabolism of cyclic GMP by phosphodiesterase type 5 (the predominent metabolizer of cGMP in the corpus cavernosum).



Men 18 years of age or older, in a stable relationship with a female partner for at least 6 months, with erectile dysfunction of at least 6 months duration. 

Exclusion criteria: Anatomic penile abnormalities, other sexual disorders (such as premature ejaculation), spinal cord disease, major psychiatric disorders, history of alcohol or substance abuse, major medical disorders (uncontrolled diabetes, active peptic ulcer disease, significant hepatic, renal or hematologic disorders, recent stroke or MI), or treatment with nitrates. 


Two separate studies were carried out and reported here, a 24-week dose-response, efficacy and safety trial, and a 12-week dose titration study with an optional 32 week, open-label extension.  In both trials, men were instructed to take the medication one hour before intercourse and not more than once daily. 

  • 24-week trial: Men were randomized to placebo, 25 mg, 50 mg or 100 mg of sildenafil. 

  • 12-week trial: Men were initially randomized to placebo or 50 mg of sildenafil. At follow-up visits the dose of the study drug could be doubled or halved, depending on efficacy or side-effects. 

Efficacy of the drug was assessed with three instruments: 

  • An event log, in which men were asked to record the date medication was taken, presence of sexual stimulation, hardness of erection and success of intercourse. The event logs were reviewed 8 times during the 24-week study and 5 times during the 12-week study.

  • A global efficacy question (did the treatment improve your erections?). This question was asked at weeks 12 and 24 in the 24-week study and at week 12 in the 12-week study.

  • The International Index of Erectile Function. This is a 15 item, self-administered questionnaire which includes questions designed to explore 5 domains of sexual function (erectile function, orgasmic function, sexual desire, intercourse satisfaction and overall satisfaction). Particular emphasis was placed on two specific questions from this index that explore erectile function: question 3 (frequency of penetration) and question 4 (maintenance of erections after penetration). The index was completed at weeks 0, 12 and 24 in the 24 week study, and at weeks 0 and 12 in the 12 week study.




532 men were enrolled in the 24-week dose-response study, 216 to placebo and 316 to the three doses of sildenafil (25, 50 and 100 mg); 87% of them completed the study.  329 men were enrolled in the 12-week dose-escalation study, 166 to placebo and 163 to sildenafil; 93% of them completed the study. 

The baseline characteristics of the patients were largely well matched between the two studies and between the sildenafil and placebo groups. These included: 

  • Mean age: 58 years.
  • Mean duration of erectile dysfunction: 3.2 years in the dose-response study; 4.8 years in the dose-escalation study.
  • Cause of erectile dysfunction organic (vs. psychogenic or mixed): 77% in the dose-response study, 59% in the dose-escalation study.
  • Associated conditions: hypertension: 28% of men; coronary disease: 8%; diabetes 13%.

Sildenafil was significantly more effective than placebo in improving erectile function, at all doses and with a positive dose-response effect.  Results include: 

  • Answers to the global efficacy question (did the drug improve your erections?) in the dose-response study were positive in 25% of men taking placebo, and in 56%, 77% and 84% of men taking 25, 50 and 100 mg of sildenafil. In the dose-escalation study, the answer was yes in 19% of men taking placebo and in 74% of men taking sildenafil.

  • For question 3 of the International Index of Erectile dysfunction (frequency of penetration), at the end of treatment the score increased by 5% in men taking placebo vs. increases of 64%, 84% and 100% for the three doses of sildenafil.  In the dose-escalation study, the increase was 10% for placebo and 95% for sildenafil.

  • For question 4 of the Index (maintenance of erections after penetration), the score increased by 24% in the placebo group and by 121%, 133% and 130% for the three doses of sildenafil. In the dose-escalation study, the increases were 13% for placebo vs. 140% for sildenafil.

  • Looking at the five domains of sexual function covered by the Index, sildenafil improved the score significantly compared with placebo in four domains (erectile function, orgasmic function, intercourse satisfaction and overall satisfaction). There was no change noted with either sildenafil or placebo in the sexual desire score.

  • The results from the logs were similar. The percentage of men achieving erections hard enough to sustain intercourse during the last week of the dose-response study was 50% for placebo and 72%, 80% and 85% for the doses of sildenafil.

  • In the dose-escalation study, the percentages of men taking 25, 50 or 100 mg after the 12 weeks were 2%, 23% and 74% (for the placebo arm the percentages were 0%, 5% and 95%).

  • The effect of sildenafil was similar in patients who were judged to have organic, psychological and mixed etiologies of erectile dysfunction. 

Adverse effects

A number of adverse effects occurred that are presumably drug related. In the dose-response study, these were: 

  • Headache. It occurred in 6% of the placebo patients vs. 14%, 21% and 30% of patients taking the drug at 25, 50 and 100 mg.

  • Flushing. Occurred in 1% in the placebo group and in 13%, 27% and 20% in the sildenafil group.

  • Dyspepsia. 1% in placebo vs. 3%, 11% and 16%.

  • Rhinitis. 2% placebo vs. 1%, 3% and 11%.

  • Visual disturbances (changes in perception of color hue or brightness). <1% in the placebo group vs. 2%, 6% and 9%.
Although these side-effects were relatively common, they led to discontinuation of the drug in only 5 men in the dose-response study (1%).

Priapism was not reported. No abnormalities in standard blood tests were noted.



In this study, self-administered oral sildenafil substantially improved erectile function in men with predominantly organic (but also psychogenic and mixed) erectile dysfunction. Improvement occurred at all doses, but was greater at higher doses, as was the incidence of side-effects.  Side effects were relatively common, particularly headache and flushing, which occurred in 30% and 20% of patients taking the highest dose (100 mg). 

As would be expected from its mode of action, the drug did not cause any increase in sexual desire, only in the efficacy of the erectile response. 

Because phosphodiesterase is also involved in the vasodilator response to nitrates, the drug can potentiate (to a dangerous degree) the vasodilating and hypotensive effects of nitrates, and is contraindicated in patients taking these drugs. It is believed that some of the deaths attributed to sildenafil and reported in the media are due to an interaction with nitrates. 

Although the drug appears to be safe, patients with significant hepatic, renal and hematologic disorders were excluded.  Furthermore, the number of patients studied and the length of follow-up were relatively limited. This medication will probably be taken by a large number of patients, for many years in certain cases. Diligence in seeking and reporting possible adverse drug reactions will be very imortant. 

June 14, 1998 


References related to this article from the NLM's PubMed database. 

Reader Comments

Date: Sat, 20 Jun 1998
From: "Bob Woolery, DC" <bwool@ix.netcom.com>
Organization: State of the Art Chiropractic, Simi Valley, CA

Patient selection to exclude those treated with nitrates seems appropriate, given the probable severe interaction, but one wonders if the target group and their prescribing physicians will be so careful.

There was a general news report 6/19 in either the LA Times or Daily News, with the usual lack of adequate references, which proported to account for some of the greater frequency of high blood pressure in the African american population by the observation of a significant difference in the response to nitric oxide in the black population. It is unpopular to specify race in medical experiments these days, but if there really is a racial difference in response to and metabolism of nitric oxide between races, the variable and its interaction effects should be teased out. If the reporter got it right, and I follow correctly, black folks have less response re: lowering of BP by vaso-dilation than other groups. This should reflect in a lesser response to Viagra. 

The patient groups were certainly large, but the treatment effects were also, and a monotonic dose-response curve demonstrated, so it does not appear that the subject pool was made large to manage a significant result from a small effect, as has been suggested.

The comment about contraindication of nitrates deserves further consideration. In some states, Paramedics, EMT's and other paraprofessionals apparently have nitroglycerine, etc for immediate administration to apparent MI cases. Even if a person were not taking nitrates, a heart attack while indulging in unaccustomed sexual activity could result in him being given nitrates by emergency response personnel. Some clear warning to prevent this is probably in order, but seems unlikely to be implemented.

Date: Sun, 15 Nov 1998
From: GELSUS@aol.com 

It seems that interracial differences in response to Viagra can be taken care of by a dose adjustment.  This information would come out of properly designed clinical trials, except that for ethical reason, we do not do them in one particular racial or ethnic group at a time. 

Nevertheless, a dosage adjustment should not be beyond the average physician and often within the purview of the patient himself. 


November 19, 1998

The September 3, 1998 NEJM has letters to the editor about this article.

Points raised include reports of cases of ventricular arrhythmias, a case of pulmonary hemorrhage, "honeymoon cystitis" in sexual partners, the evaluation of endocrine causes of impotence and some clarification of the improvement achieved in erectile function.

Submit a comment

Journal Club homepage

Site Meter