Five RCTs were included in the review (2,225 patients). The quality of the trials was moderate; randomisation was adequate in two trials and all trials were double blinded. Loss to follow up was balanced, so the risk of bias for incomplete outcome data was low.
Effectiveness: The pairwise meta-analysis of all dosage groups suggested that avanafil was significantly more effective than placebo in improving International Index of Erectile Function–Erectile Function scores (MD 4.47, 95%CI 3.51 to 5.43), and answers to Sexual Encounter Profile question two (SEP-2, MD 17.41, 95%CI 14.03 to 20.79) and question three (SEP-3, MD 20.01, 95%CI 22.98 to 37.22). Similarly, avanafil was significantly effective compared with placebo when different dosages were used (50mg, 100mg and 200mg). There was no evidence of major heterogeneity (results not reported). Higher dosage was significantly more effective than lower dosage, except for the comparison between 100mg and 200mg. The network meta-analysis showed a similar result to pairwise meta-analysis, although some estimates from network meta-analysis failed to achieve statistical significance (results not reported). There was evidence of dose-response relationship among various dosage groups.
Safety: Avanafil was associated with significantly higher incidence of any adverse events and serious adverse events, flushing and headache (full results were reported in the paper). The risks of adverse events were similar between different dosage groups.
The meta-regression analyses did not identify any significant effect modifiers.
There was no evidence of publication bias.