Sildenafil

• Boolell 1996 was a double-blind crossover trial in 12 men with erectile dysfunction of no established organic cause. Single doses of 10, 25 and 50 mg lengthened the time penile rigidity stayed above 80 percent during visual sexual stimulation compared with placebo, and 10 of 12 men reported better erectile activity on sildenafil against 2 of 12 on placebo (PMID: 8813924).
• Goldstein 1998 reported two double-blind trials, one of 532 men for 24 weeks and one of 329 men for 12 weeks, with erectile dysfunction of mixed causes. Erectile function scores rose with dose (p<0.001), and in the flexible-dose trial 69 percent of intercourse attempts succeeded on sildenafil against 22 percent on placebo; headache, dyspepsia and flushing affected 6 to 18 percent of men (PMID: 9580646).
• Rendell 1999 studied 268 men with diabetes for 12 weeks. Improved erections were reported by 56 percent on sildenafil and 10 percent on placebo (p<0.001), and 61 percent against 22 percent had at least one successful attempt at intercourse (PMID: 9952201).
• Nurnberg 2003 was a 6-week trial in 90 men whose depression was in remission on serotonergic antidepressants. The primary sexual function rating was much or very much improved in 54.5 percent on sildenafil against 4.4 percent on placebo (p<0.001) (PMID: 12503977).
• The label summarizes 21 randomized, double-blind, placebo-controlled trials of up to 6 months in more than 3,000 men aged 19 to 87, and sildenafil was statistically better than placebo in all 21 (US prescribing information for Viagra).
• In comparisons with other PDE5 inhibitors, a 2009 meta-analysis found no difference in efficacy or adverse events among sildenafil, vardenafil and tadalafil in four head-to-head trials (PMID: 19884626). A 2015 network meta-analysis of 82 trials and 47,626 men found that, at starting doses, sildenafil 50 mg had the greatest efficacy and the highest overall rate of adverse events, while tadalafil 10 mg had intermediate efficacy and the lowest adverse event rate (PMID: 25817916).

Pulmonary arterial hypertension

• SUPER-1 (Galie 2005) gave 278 adults 20, 40 or 80 mg three times daily or placebo for 12 weeks. The primary endpoint was met: six-minute walk distance rose by 45, 46 and 50 meters more than with placebo (p<0.001), with lower pulmonary artery pressure and better functional class; the incidence of clinical worsening did not differ, and the study was not powered for mortality (PMID: 16291984).
• PACES (Simonneau 2008) was a 16-week trial in 267 patients already on intravenous epoprostenol. Walk distance rose 28.8 meters more than with placebo (95% CI 13.9 to 43.8), and fewer patients had a clinical worsening event by week 16 (0.062 against 0.195, p=0.002); headache affected 57 percent on sildenafil and 34 percent on placebo (PMID: 18936500).
• Study A1481243 was a 12-week trial in 103 adults already on bosentan. Adding sildenafil did not improve six-minute walk distance compared with bosentan alone (US prescribing information for Revatio).
• STARTS-1 (Barst 2012) was a 16-week trial in 235 children aged 1 to 17. The primary comparison, change in peak oxygen consumption for the three doses combined, narrowly missed significance (7.7 percent, p=0.056); the medium and high doses improved oxygen consumption, functional class and hemodynamics, and the low dose was ineffective (PMID: 22128226).
• STARTS-2 (Barst 2014) was the extension study. By August 2011, when all children had at least 3 years of follow-up from the start of STARTS-1, 37 deaths had been reported, and mortality was higher with the high dose than with the low dose (hazard ratio 3.95); the authors noted that multiple analyses raised uncertainty about the dose and survival relationship (PMID: 24637559).
• Study A1481324 (Hoeper 2024) gave 385 adults 5, 20 or 80 mg three times daily. The FDA required this trial after the pediatric finding; it stopped at its first interim analysis when 80 mg proved noninferior to 5 mg for all-cause mortality (hazard ratio 0.51, 99.7% CI 0.22 to 1.21), 78 patients died in all, and the authors report that the FDA then revoked approval of the 5 mg dose for adults (PMID: 38752352).

Trials that found no benefit

• RELAX (Redfield 2013) tested sildenafil for 24 weeks in 216 patients with heart failure with preserved ejection fraction. The primary endpoint, change in peak oxygen consumption, did not differ from placebo (p=0.90), nor did walk distance or clinical status; serious adverse events occurred in 22 percent on sildenafil and 16 percent on placebo (PMID: 23478662).
• Basson 2002 studied 781 women with female sexual arousal disorder, and no patient or partner endpoint differed between sildenafil and placebo in 577 estrogenized and 204 estrogen-deficient women (PMID: 12150499).
• UK STRIDER (Sharp 2018) enrolled 135 women with severe early-onset fetal growth restriction. Sildenafil did not prolong pregnancy (median 17 against 18 days from randomization to delivery, p=0.23) or change live births, deaths or birthweight (PMID: 30169244).
• In the Dutch STRIDER trial (Pels 2020) of 216 women, the primary outcome, perinatal death or major neonatal morbidity, occurred in 60.2 percent on sildenafil and 54.2 percent on placebo (p=0.38), and the trial was halted because of more neonatal pulmonary hypertension in the sildenafil group (PMID: 32585017).

Other conditions studied

• Nurnberg 2008 was an 8-week trial in 98 women with antidepressant-associated sexual dysfunction. In these women, the primary sexual function rating improved more with sildenafil than with placebo (mean 1.9 against 1.1, p=0.001) (PMID: 18647982).
• Fries 2005 was a double-blind crossover trial in 16 patients with Raynaud's phenomenon, with 4 weeks per period. Attacks were less frequent (35 against 52) and shorter in total (581 against 1,046 minutes) on sildenafil, and 2 patients stopped the study drug because of side effects (PMID: 16275885).
• In 14 mountaineers, sildenafil raised maximum workload while they breathed low-oxygen air (172.5 against 130.6 watts) and at Mount Everest base camp (PMID: 15289213). In 10 trained men, sildenafil shortened a 6-km cycling time trial by 15 percent at a simulated altitude of 3,874 m but had no effect at sea level (PMID: 16455814), and in 11 volunteers it did not improve exercise capacity in normal or low-oxygen conditions (PMID: 27414159).

Observational studies of Alzheimer's disease

In a Cleveland Clinic analysis of insurance claims for 7.23 million people, sildenafil use was associated with a 69 percent lower risk of Alzheimer's disease (hazard ratio 0.31), and the authors stated that the association does not establish causality (PMID: 35572351). A Medicare study of patients with pulmonary hypertension that compared PDE5 inhibitors with endothelin receptor antagonists found no reduced risk (PMID: 36330433), and a 2025 Mendelian randomization study found no genetic evidence that PDE5 inhibition lowers the risk of Alzheimer's disease (PMID: 41078087).

Products sold as supplements

Of 353 sexual enhancement products sold as dietary supplements that drew FDA warnings for hidden drug ingredients from 2007 to 2016, 166 (47.0 percent) contained sildenafil (PMID: 30646238). By 2012, at least 46 unapproved structural analogues of sildenafil, tadalafil and vardenafil had been reported in food supplements, and the authors described them as essentially experimental drugs, most with no known efficacy or safety profile (PMID: 22464558).

Blood pressure and the heart

In healthy volunteers, sildenafil caused transient decreases in supine blood pressure, by a mean maximum of 8.4/5.5 mmHg, and the label states that treatments for erectile dysfunction, sildenafil among them, should not generally be used in men for whom sexual activity is inadvisable because of their cardiovascular status (US prescribing information for Viagra). When men took sildenafil before a sublingual nitroglycerin tablet, systolic blood pressure fell about 4 times as much as after placebo (PMID: 10078539). In pooled data from more than 120 trials run from 1993 to 2001, the rate of heart attack or cardiovascular death was 0.91 per 100 person-years on sildenafil and 0.84 on placebo (relative risk 1.08, 95% CI 0.45 to 2.77) (PMID: 14529061). After approval, heart attacks, sudden cardiac death, strokes and other serious cardiovascular events were reported in temporal association with Viagra, mostly in men with existing cardiovascular risk factors, and the label states that it is not possible to determine whether these events were caused by the drug, by sexual activity, by underlying disease or by other factors (US prescribing information for Viagra). In a Swedish registry study of men with coronary disease taking nitrates, combined use of a PDE5 inhibitor (5,710 men) was associated with higher all-cause mortality (hazard ratio 1.39) and with more heart attacks and heart failure; the study was observational (PMID: 38233015).

Sudden vision loss (NAION)

The label states that non-arteritic anterior ischemic optic neuropathy (NAION), a cause of decreased vision including permanent loss of vision, has been reported rarely after marketing with all PDE5 inhibitors, and that its background rate is 2.5 to 11.8 cases per 100,000 men aged 50 or older each year (US prescribing information for Viagra). A case-crossover study with Pfizer authors found about twice the risk of NAION within five half-lives of a PDE5 inhibitor dose among 43 confirmed cases (odds ratio 2.15, 95% CI 1.06 to 4.34), which the authors estimated as three added cases per 100,000 men aged 50 or older each year with weekly use (PMID: 25358826). A similar study with Eli Lilly authors found a rate ratio of 2.27 (95% CI 0.99 to 5.20) over 30 days, and a 2018 meta-analysis found no significant overall association but a significant one in the sildenafil subgroup (risk ratio 2.25) (PMID: 28336289, PMID: 29884471).

Hearing loss

Sudden decrease or loss of hearing, sometimes with ringing in the ears and dizziness, has been reported in temporal association with PDE5 inhibitors, and the label states that it is not possible to determine whether these events are related directly to the drugs (US prescribing information for Viagra). In a cross-sectional US survey of 11,525 men aged 40 or older, self-reported hearing impairment was associated with sildenafil use (odds ratio 2.05) but not with tadalafil or vardenafil use (PMID: 20479381).

Priapism

Erections lasting more than 4 hours and priapism, defined in the label as a painful erection lasting more than 6 hours, have been reported infrequently since approval, and the label warns that untreated priapism can cause penile tissue damage and permanent loss of potency (US prescribing information for Viagra).

Melanoma

In the Health Professionals Follow-up Study of 25,848 US men, recent sildenafil use was associated with a higher risk of melanoma (hazard ratio 1.84, 95% CI 1.04 to 3.22) (PMID: 24710960). A Swedish study of 4,065 melanoma cases found a smaller association with PDE5 inhibitors (odds ratio 1.21) that was seen with a single prescription but not with repeated prescriptions and only for early-stage tumors, a pattern the authors said raises questions about whether the link is causal (PMID: 26103029). A 2017 meta-analysis of five observational studies found a slightly higher melanoma risk with PDE5 inhibitor use (odds ratio 1.12) (PMID: 28728871).

Pregnancy and fetal growth restriction

The Viagra label states that the drug is not indicated for women and reports no adverse developmental outcomes in rats and rabbits given sildenafil during organogenesis (US prescribing information for Viagra). In a Dutch trial of 216 pregnant women with severe early-onset fetal growth restriction, pulmonary hypertension occurred in 18.8 percent of newborns in the sildenafil group against 5.1 percent with placebo (relative risk 3.67), and the trial was stopped early for safety concerns on 19 July 2018 (PMID: 32585017).

Children with pulmonary arterial hypertension

In the long-term extension of a pediatric trial, children randomized to the high dose had higher mortality than those on the low dose (hazard ratio 3.95, 95% CI 1.46 to 10.65), though the authors noted that multiple analyses raised uncertainty about the dose and survival relationship (PMID: 24637559). The Revatio label reports deaths in 9.1, 13.5 and 22 percent of the low, medium and high dose groups of the pediatric studies, and states that a causal link is unlikely in view of a later adult study (US prescribing information for Revatio; PMID: 38752352).

Other labeled warnings

The labels also report nosebleeds in 13 percent of Revatio-treated patients with pulmonary arterial hypertension linked to connective tissue disease, more vaso-occlusive crises with sildenafil in a small, prematurely stopped study of pulmonary hypertension from sickle cell disease, a risk of pulmonary edema in pulmonary veno-occlusive disease, seizures and transient global amnesia after marketing, and hypersensitivity reactions including anaphylaxis (US prescribing information for Viagra and Revatio).