This review indirectly compared phosphodiesterase-5 (PDE-5) inhibitors for the treatment of erectile dysfunction. The authors concluded that common efficacy outcomes were similar between the PDE-5 inhibitors. Poor reporting of review methods, a limited search, and differences between the studies in terms of patient characteristics make it difficult to confirm the robustness of the conclusions.

To compare phosphodiesterase-5 (PDE-5) inhibitors for the treatment of erectile dysfunction through indirect comparisons of interventions.

Previous systematic reviews, PubMed (to June 2005) and the Cochrane Library (Issue 1, 2005) were searched. The search terms were drug names and randomis(z)ed trial. Studies reported in abstracts or posters were not included.

Study designs of evaluations included in the review

Double-blind randomised controlled trials (RCTs) that lasted at least 3 weeks, had at least 10 men per treatment group, and scored at least two points on the Jadad scale were eligible for inclusion. Most of the included studies lasted 4 to 12 weeks.

Specific interventions included in the review

Studies that compared three specified oral PDE-5 inhibitors (sildenafil, tadalafil and vardenafil) with placebo or an active comparator and were conducted in the home setting were eligible for inclusion. Only licensed doses of the respective drugs (as listed in the British National Formulary) were included in the review. All of the included studies were placebo controlled and none compared PDE-5 with an active comparator. The primary studies used fixed dosing and dose optimisation; the review only included the top fixed dose in optimisation studies.

Participants included in the review

Studies of men with erectile dysfunction of any aetiology were eligible for inclusion. The included studies were generally of men with erectile dysfunction for at least 3 to 6 months and aged mid-fifties or older. Most of the included studies were in men with mixed aetiology (organic, psychogenic or mixed). Studies of tadalafil and vardenafil included some patients with erectile dysfunction after prostatectomy (15% and 13%, respectively). Sildenafil studies included men with depression, spinal cord injury, multiple sclerosis, coronary heart disease, radiotherapy for prostate cancer, renal failure, rectal surgery and spinal bifida (0.5% to 5.7% of patients had each of these conditions); studies of tadalafil and vardenafil did not include any patients with these conditions. Most of the tadalafil and vardenafil studies were in men who had been unsuccessfully treated with a PDE-5 inhibitor; none of the sildenafil studies were in this group.

Outcomes assessed in the review

Studies that reported 'useful information' on efficacy or adverse effects were eligible for inclusion. The review sought the following efficacy outcomes: improved erections; erections per week; successful attempts at intercourse (including more than 40% or more than 60 or 75% successful); final score or change from baseline on questions 3, 4 and the erectile function domain of the International Index of Erectile Function (IIEF); and normal erectile function at the end of the study (IIEF total score of at least 22 out of 30). The review also sought withdrawal (all-cause, lack of efficacy and due to adverse events), adverse events, and the relationship between time of taking a dose of the drug and time of attempted or successful intercourse. The review actually assessed efficacy outcomes that were commonly reported: improved erection, and scores and change in score for the erectile function domain.

How were decisions on the relevance of primary studies made?

Decisions about the selection of studies were made by consensus.

Validity was assessed and scored using the Jadad scale, which considers the reporting and handling of randomisation, blinding and handling of withdrawals. The maximum possible score was five points.

The authors did not state who performed the validity assessment.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction.

Data extracted included the number of patients on which calculations were based and the number of events or event rates. The authors extracted data using the intention-to-treat definitions used in individual studies; where this was not reported, the authors used the numbers of men reported as the denominators.

How were the studies combined?

The intention was to pool data from clinically homogeneous studies with similar patients, drug dose, treatment duration, outcomes and comparators, and where the number of events was not small, using meta-analysis. Data were only pooled where there were at least 2 studies with more than 200 patients and more than 30 events. The relative benefit (or risk) was calculated with 95% confidence interval (CI) using a fixed-effect model, with a value of 0.5 added to treatment arms with no events. Where there was a statistically significant difference between treatment arms, the number-needed-to-treat (NNT) or -harm (NNH) was calculated. In the review, the term NNT was used where significantly more beneficial events were reported with PDE-5 inhibitors compared with placebo; the number-needed-to-treat to prevent (NNTp) was used where significantly fewer withdrawals or adverse events were reported with PDE-5 inhibitors compared with placebo. The NNH was used where significantly more adverse events were reported with PDE-5 inhibitors than with placebo. Mean values for continuous data were calculated, weighted by sample size.

The review first analysed outcomes according to each intervention and then narratively compared drugs indirectly taking the comparability of response rates into account.

How were differences between studies investigated?

Statistical heterogeneity was not assessed, but clinical homogeneity was examined graphically. Differences between the studies were also discussed in the text with respect to clinical condition, exclusion criteria and the reporting of outcomes.

Thirty-five RCTs examined sildenafil (n=7,135), 8 RCTs examined tadalafil (n=2,071) and 7 RCTs examined vardenafil (n=3,374).

The quality scores were high: 11 sildenafil studies, 2 tadalafil studies and 3 vardenafil studies scored five points; and 17 sildenafil studies, 1 tadalafil study and 4 vardenafil studies scored four points.

Sildenafil (combined 50- and 100-mg fixed doses and dose optimisation regimens).

Sildenafil significantly improved erections compared with placebo (n=5,467), 76% versus 23%; the NNT was 1.9 (95% CI: 1.8, 2.0). It also reduced all-cause withdrawals (NNTp 23, 95% CI: 17, 37) and lack of efficacy withdrawals (NNTp 25, 95% CI: 21, 34) compared with placebo, but increased withdrawals due to adverse events (NNH 120, 95% CI: 67, 560). The NNH for reporting at least one adverse event was 4.9 (95% CI: 4.2, 6.0). Headache was the most commonly reported adverse effect (17%), followed by flushing (13%) and dyspepsia (7.8%).

Tadalafil (combined 10- and 20-mg fixed doses).

Tadalafil significantly improved erections compared with placebo (n=1,651), 75% versus 24%; the NNT was 1.9 (95% CI: 1.8, 2.1). It also reduced all-cause withdrawals (NNTp 15, 95% CI: 8.8, 46) and lack of efficacy withdrawals (NNTp 24, 95% CI: 14, 69) compared with placebo, but increased withdrawals due to adverse events (NNH 52, 95% CI: 29, 260). The NNH for reporting at least one adverse event was 4.6 (95% CI: 3.4, 7.2). Headache was the most commonly reported adverse effect (13%), followed by dyspepsia (10%) and flushing (4.8%).

Vardenafil (combined 10- and 20-mg fixed doses and dose optimisation regimens).

Vardenafil significantly improved erections compared with placebo (n=2,856), 71% versus 22%; the NNT was 2.0 (95% CI: 1.9, 2.2). It also reduced all-cause withdrawals (NNTp 7.7, 95% CI: 6.0, 11) and lack of efficacy withdrawals (NNTp 11, 95% CI: 9.0, 16) compared with placebo, but increased withdrawals due to adverse events (NNH 65, 95% CI: 37, 250). Headache was the most common adverse effect (15%), followed by flushing (13%) and rhinitis (7.9%).

Comparisons of different treatments. There was consistency between drugs in the percentage of successful attempts at intercourse: 23 to 28% with placebo, 65% for sildenafil, 62% for tadalafil and 59% for vardenafil. There was also consistency in the percentage reporting improved erections: 22 to 24% with placebo, 76% for sildenafil, 75% for tadalafil and 71% for vardenafil.

There was less consistency between drugs for withdrawals: all-cause and lack of efficacy withdrawals were more common with placebo in vardenafil trials (32% and 12%) than in tadalafil trials (19% and 7.5%) and sildenafil trials (12% and 4.4%). Adverse event withdrawals were lower with sildenafil (0.6%) than with tadalafil (1.5%) or vardenafil (1.8%).

Event rates were similar between studies for individual adverse events.

A lack of consistent reporting of outcomes limited the comparisons. Common efficacy outcomes were similar between the PDE-5 inhibitors, with few serious adverse events.

The review question was clear in terms of the study design, intervention, participants and outcomes. Limiting the search to two databases might have resulted in the omission of relevant data. It was not clear if any attempts were made to locate unpublished studies or if any language restrictions were applied. It appears that methods were used to minimise reviewer errors and bias in the selection of studies, but it was unclear whether similar steps were taken in the data extraction and validity assessment processes. Only double-blind RCTs were included, and validity was assessed using established criteria and the results of the assessment reported.

Differences between the studies were discussed and statistical heterogeneity was examined graphically. The treatments were compared indirectly in a narrative and the condition of comparability of placebo response rates was assessed. The lack of complete reporting of review methods, a limited search, and differences between the studies with respect to patient characteristics make it difficult to confirm the robustness of the conclusions.

The authors did not state any implications for practice or further research.

Pfizer Ltd.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.