What the studies counted
The label's controlled studies used the International Index of Erectile Function. Two questions were primary end points: the ability to achieve erections sufficient for sexual intercourse, and the maintenance of erections after penetration. Patients also answered a global question about whether erections improved, and many kept a diary.
The diaries, in studies of about 1600 patients, showed no effect of VIAGRA on rates of attempted intercourse (about 2 per week). Success changed. Per patient, weekly success rates averaged 1.3 on VIAGRA versus 0.4 on placebo, and group mean success rates were about 66% on VIAGRA versus about 20% on placebo. The tablet did not create extra attempts. It changed the share of attempts that worked.
Men in these studies were not starting from zero sexual function. One-third to one-half reported successful intercourse at least once in a 4-week run-in before treatment. That baseline matters when a later review talks about how many men need to be treated for one to improve.
Fixed doses, then a flexible climb
In four randomized fixed-dose studies (1797 patients) of 12 to 24 weeks, erectile dysfunction was organic in 58% (including diabetes, and excluding spinal cord injury in that organic share), psychogenic in 17%, and mixed in 24%. Sixty-three percent, 74%, and 82% of the patients on 25 mg, 50 mg and 100 mg of VIAGRA, respectively, reported an improvement in their erections, compared to 24% on placebo. Titration studies (n=644), with most patients eventually receiving 100 mg, were similar. The label reports an overall treatment p<0.0001.
The 1998 New England Journal of Medicine paper is one of the studies behind that era, and it uses a different cut. Goldstein and colleagues ran a 24-week dose-response study in 532 men (25, 50, or 100 mg versus placebo) and a 12-week flexible dose-escalation study in 329 other men. In the dose-escalation study, 69 percent of all attempts at sexual intercourse were successful for the men receiving sildenafil, as compared with 22 percent for those receiving placebo. Mean successful attempts per month were 5.9 versus 1.5. That 69% is a share of attempts in one flexible-dose study. It is not the 74% of men on 50 mg who said erections improved in the label's fixed-dose pool.
Groups in which the same end point moved less
A diabetes-only flexible-dose study on the label (n=268) started patients at 50 mg and allowed 25 or 100 mg. By the end everyone on the active arm was on 50 or 100 mg. On the global question, 57% of VIAGRA patients reported improved erections versus 10% on placebo. Diary data: 48% of intercourse attempts were successful versus 12% on placebo. The direction matches the broader trials. The percentages do not.
Across trials, the label says VIAGRA improved the erections of 43% of radical prostatectomy patients compared to 15% on placebo. That is the smallest of these global figures. Nerve-sparing detail is not broken out so the 43% is the pooled prostatectomy line, not a promise for one surgical technique.
A spinal-cord study on the label (n=178, flexible dose up to 100 mg, crossover) reported improved erections in 83% on VIAGRA versus 12% on placebo, and successful intercourse on 59% of attempts versus 13% on placebo. The 1999 Annals of Neurology paper listed in this search is a spinal-cord crossover of 178 men as well, and it reports its own cut: 132 of 166 (80%) said sildenafil improved sexual intercourse, compared with 17 of 166 (10%) on placebo. Use the paper's 80% and 10% for that publication, and the label's 83% and 12% for the label. Do not average them into a third number.
Psychogenic etiology, in the label's subgroup of fixed-dose and titration studies, was higher: 84% of VIAGRA patients reported improvement in erections compared with 26% of placebo.
What a 2001 review added, including its limit
The first result for this query is a systematic review of randomised trials of sildenafil for male erectile dysfunction (Burls, Gold, and Clark, British Journal of General Practice, 2001). Twenty-one trials were identified. All of them showed a statistically significant improvement versus placebo. In a meta-analysis of 16 trials, men were 3.57 times as likely to have improved erections on sildenafil (95% CI 2.93 to 4.43). The number needed to treat to have one man with improved erections was two.
The same review then narrows that two. Many participants still had some erectile function at baseline, and the authors write that in ordinary practice, where function tends to be more impaired, the number needed to treat may be higher. The diabetes and prostatectomy rows on the label are what that sentence is pointing at.
Side effects inside the same programs
In placebo-controlled studies, the discontinuation rate due to adverse reactions was 2.5% on VIAGRA and 2.3% on placebo. The fixed-dose adverse-reaction table (25 mg, 50 mg, 100 mg, then placebo) lists headache at 16%, 21%, 28%, and 7%. Flushing at 50 mg was 19%, versus 2% on placebo. Abnormal vision, described as a color tinge, light sensitivity, or blurred vision, was 2% at 50 mg and 11% at 100 mg, versus 1% on placebo. Goldstein's dose-escalation study named headache, flushing, and dyspepsia as the common effects, in the range of 6 percent to 18 percent of the men.
A trial that shows more headaches at 100 mg than at 25 mg is the same evidence base as the efficacy climb. It is not a reason to skip the nitrate rule. The trials did not authorize nitrates, nitrites, or riociguat. How to time a dose and a meal is the how to take sildenafil note. The angina-program history is how Viagra was discovered.
Results in this search that are not ED trials
The second listing is a study of sildenafil for exercise ability in people with heart failure, not a placebo comparison of erections. A cancer-institute page is an index of trials that happen to use the drug. A chemical-vendor page is a catalog of study listings. The ninth result is a pulmonary arterial hypertension trial of scheduled daily doses. Those doses are not the as-needed 25, 50, and 100 mg ED program in section 14, and this note does not convert them into an erectile-dysfunction regimen.
If the question is whether the ED trials beat placebo on erections and on successful attempts, the label, the 1998 journal trial, and the 2001 review agree on the direction and disagree, usefully, on how large the gain is once diabetes or prostatectomy is the whole sample. Buying checks and the imprint on the tablet sit on the sildenafil medicine note. Julian Frost is the PDE5 signer. The margin disclaimer states the limit of a reading note.
Talk with your own clinician before you treat a trial percentage as the result you should expect.