Comparative effectiveness, safety and acceptability of pharmacological and psychosocial interventions for the treatment of cannabis use disorder:a network meta-analysis

<b>Aim</b>:<br/>Our primary aim was to conduct a network meta-analysis (NMA) for the effectiveness, safety and acceptability of all psychosocial interventions and all pharmacotherapies for cannabis use disorder (CUD). <br/><br/><b>Methods</b>:<br/>We conducted a NMA of studies identified from two completed systematic reviews examining the effectiveness, safety and acceptability of pharmacotherapies and psychosocial interventions for CUD in people aged ≥16 years. Outcomes were level of cannabis use, abstinence, adverse events and treatment completion.<br/><br/><b>Results</b>:<br/>Fifty-seven studies were eligible for NMA. Depending on outcome, NMAs of psychosocial interventions included 6-16 studies (445-2287 participants), and NMAs of pharmacotherapies included 5-36 studies (260-3106 participants). Results are described relative to minimum clinically meaningful difference [mean difference (MD) ±0.05, odds ratio (OR) ≤0.8 or ≥1.25].<br/><br/>Evidence for pharmacological interventions reducing cannabis use was very uncertain. However, low-certainty evidence suggested that dialectical behavioural/acceptance and commitment therapies [DBT/ACT; MD -0.18, 95% credible interval (-0.26 to -0.09)] and cognitive-behavioural therapy with motivation enhancement (MET-CBT) with contingency management [CM; MD -0.15 (-0.23 to -0.07)] reduced use frequency relative to a nonspecific comparator. Interventions including CM also supported abstinence (very low certainty). Limited, low-certainty evidence indicated that cannabidiol [OR 2.91 (0.45 to 27.98)], N-acetylcysteine [OR 1.30 (0.55 to 3.74)] and varenicline [OR 4.85 (0.65 to 48.26)] promoted abstinence compared with placebo, although cannabidiol was associated with adverse events [OR 1.58 (0.26 to 8.98)]. Very low-certainty evidence suggested mixed-action antidepressants, benzodiazepines, bupropion and buspirone also had more adverse events than placebo, without any indication of effectiveness. Very low-certainty evidence indicated that DBT/ACT and CM improved treatment completion. In contrast, MET-CBT, with/without affect management, reduced completion, relative to a nonspecific comparator. Low-certainty evidence suggested N-acetylcysteine [OR 1.29 (0.76 to 2.19)] and delta-9-tetrahydrocannabinol preparations [OR 1.32 (0.89 to 1.97)] increased treatment completion, while fatty acid amide hydrolase inhibitor [OR 0.80 (0.40 to 1.58)] and selective serotonin reuptake inhibitors [OR 0.56 (0.29 – 1.04)] reduced it, relative to placebo.<br/><br/><b>Conclusions</b>:<br/>There is some evidence that dialectical behavioural/acceptance and commitment therapies, contingency management alone or with cognitive-behavioural therapy with motivation enhancement, N-acetylcysteine, cannabidiol or varenicline may facilitate abstinence, reduce use and/or support treatment completion for cannabis use disorder. Some interventions may have unfavourable effects on those outcomes, relative to placebo or nonspecific comparators. While safety data for psychosocial interventions is lacking, several pharmacotherapies may be associated with adverse events. These findings should be interpreted with caution as all evidence is low- to very-low-certainty due to imprecise or heterogeneous treatment effects and high risk of bias in some study results.