Background and aims - Pharmacologically controlled drinking in the treatment of alcohol dependence or alcohol use disorders (AUDs) is an emerging concept. Our objective was to explore the comparative effectiveness of drugs used in this indication.

Design - Systematic review with direct and network meta-analysis of double-blind randomized controlled trials (RCTs) assessing the efficacy of nalmefene, naltrexone, acamprosate, baclofen or topiramate in non-abstinent adults diagnosed with alcohol dependence or AUDs. Two independent reviewers selected published and unpublished studies on Medline, the Cochrane Library, Embase, ClinicalTrials.gov, contacted pharmaceutical companies, the European Medicines Agency and the Food and Drug Administration, and extracted data.

Setting - Thirty-two RCTs.

Participants - A total of 6036 patients.

Measurements - The primary outcome was total alcohol consumption (TAC). Other consumption outcomes and health outcomes were considered as secondary outcomes.

Findings - No study provided direct comparisons between drugs. A risk of incomplete outcome data was identified in 26 studies (81%) and risk of selective outcome reporting in 17 (53%). Nalmefene [standardized mean difference (SMD) = -0.19, 95% confidence interval (CI) = -0.29, -0.10; I  = 0%], baclofen (SMD = -1.00, 95% CI = -1.80, -0.19; one study) and topiramate (SMD = -0.77, 95% CI = -1.12, -0.42; I  = 0%) showed superiority over placebo on TAC. No efficacy was observed for naltrexone or acamprosate. Similar results were observed for other consumption outcomes, except for baclofen (the favourable outcome on TAC was not reproduced). The number of withdrawals for safety reasons increased under nalmefene and naltrexone. No treatment demonstrated any harm reduction (no study was powered to explore health outcomes). Indirect comparisons suggested that topiramate was superior to nalmefene, naltrexone and acamprosate on consumption outcomes, but its safety profile is known to be poor.

Conclusions - There is currently no high-grade evidence for pharmacological treatment to control drinking using nalmefene, naltrexone, acamprosate, baclofen or topiramate in patients with alcohol dependence or alcohol use disorder. Some treatments show low to medium efficacy in reducing drinking across a range of studies with a high risk of bias. None demonstrates any benefit on health outcomes.