IMPORTANCE: Nalmefene is a newly approved drug for alcohol use disorder, but the risk of harms has not been evaluated from empirical trial evidence.

OBJECTIVE: To assess the harm of nalmefene administered to individuals diagnosed with substance use or impulse control disorders by performing a systematic review and meta-analysis of randomised controlled trials.

DATA SOURCES: A search was performed in Cochrane Central Register of Controlled Trials (CENTRAL, 2014), MEDLINE via PubMed (1950), EMBASE via Ovid (1974), and Clinicaltrials.gov through December 2014.

STUDY SELECTION: This study included only randomised controlled trials with placebo or active controls that administered nalmefene to adult individuals for treating impulse control and/or substance use disorders. Both published and unpublished randomised controlled trials were eligible for inclusion.

DATA EXTRACTION AND SYNTHESIS: Internal validity was assessed using the Cochrane risk-of-bias tool. Published information from the trials was supplemented by contact between reviewers and industry sponsor. Data were combined using two meta-approaches in fixed effects models; Peto Odds Ratios and risk differences were reported with 95% confidence intervals (95%CIs).

MAIN OUTCOMES AND MEASURES: Number of patients with serious adverse events, including specific psychiatric serious adverse events and withdrawals due to adverse events.

RESULTS: Of 20 potentially relevant studies, 15 randomised controlled trials met the inclusion criteria, and 8 of these provided data enabling the meta-analysis. Overall, serious adverse events did not occur more often in the nalmefene group than in the placebo group (Peto Odds Ratio = 0.97 [95% CI 0.64-1.44]; P = 0.86). Risk of psychiatric serious adverse events was slightly elevated, albeit not at a statistically significant level (Peto Odds Ratio = 1.32 [95% CI 0.62, 2.83]; P = 0.47). Withdrawals due to adverse events were significantly more likely to occur with nalmefene compared to placebo (Peto Odds Ratio = 3.22 [95% CI 2.46-4.22]; P<0.001).

CONCLUSIONS AND RELEVANCE: The three-fold increased risk of withdrawal from treatment on nalmefene due to adverse events is a matter of safety concern. The nature of these adverse events cannot be elucidated further without access to individual patients data.