AbstractAimsThe estimated effect of sodium oxybate (SMO) in the treatment of alcohol dependence is heterogeneous. Population severity and treatment duration have been identified as potential effect modifiers. Population severity distinguishes heavy drinking patients with <14 days of abstinence before treatment initiation (high-severity population) from other patients (mild-severity population). Treatment duration reflects the planned treatment duration. This study aimed to systematically investigate the effect of these potential effect moderators on SMO efficacy in alcohol-dependent patients.MethodsNetwork meta-regression allows for testing potential effect modifiers. It was selected to investigate the effect of the above factors on SMO efficacy defined as continuous abstinence (abstinence rate) and the percentage of days abstinent (PDA). Randomized controlled trials for alcohol dependence with at least one SMO group conducted in high-severity and mild-severity populations were assigned to a high-severity and mild-severity group of studies, respectively.ResultsEight studies (1082 patients) were retained: four in the high-severity group and four in the mild-severity group. The high-severity group was associated with larger SMO effect sizes than the mild-severity group: abstinence rate risk ratio (RR) 3.16, P = 0.004; PDA +26.9%, P < 0.001. For PDA, longer treatment duration was associated with larger SMO effect size: +11.3% per extra month, P < 0.001. In the high-severity group, SMO showed benefit: abstinence rate RR 2.91, P = 0.03; PDA +16.9%, P < 0.001. In the mild-severity group, SMO showed benefit only in PDA for longer treatment duration: +23.9%, P < 0.001.ConclusionsIn the retained studies with alcohol-dependent patients, high-severity population and longer treatment duration were associated with larger SMO effect sizes.