Pharmacotherapies for cannabis use disorder: A systematic review and network meta-analysis.

OBJECTIVE

This study aimed to determine the efficacy and acceptability of pharmacotherapies for cannabis use disorder (CUD).

METHODS

We conducted a systematic review and frequentist network meta-analysis, searching five electronic databases for randomized placebo-controlled trials of individuals diagnosed with CUD receiving pharmacotherapy with or without concomitant psychotherapy. Primary outcomes were the reduction in cannabis use and retention in treatment. Secondary outcomes were adverse events, discontinuation due to adverse events, total abstinence, withdrawal symptoms, cravings, and CUD severity. We applied a frequentist, random-effects Network Meta-Analysis model to pool effect sizes across trials using standardized mean differences (SMD, g) and rate ratios (RR) with their 95% confidence intervals.

RESULTS

We identified a total of 24 trials (n=1912, 74.9% male, mean age 30.2 years). Nabilone (d=-4.47 [-8.15; -0.79]), topiramate (d=-3.80 [-7.06; -0.54]), and fatty-acid amyl hydroxylase inhibitors (d=-2.30 [-4.75; 0.15]) reduced cannabis use relative to placebo. Dronabinol improved retention in treatment (RR=1.27 [1.02; 1.57]), while topiramate worsened treatment retention (RR=0.62 [0.42; 0.91]). Gabapentin reduced cannabis cravings (d=-2.42 [-3.53; -1.32], while vilazodone worsened craving severity (d=1.69 [0.71; 2.66]. Buspirone (RR=1.14 [1.00; 1.29]), venlafaxine (RR=1.78 [1.40; 2.26]), and topiramate (RR=9.10 [1.27; 65.11]) caused more adverse events, while topiramate caused more dropouts due to adverse events.

CONCLUSIONS

Based on this review, some medications appeared to show promise for treating individual aspects of CUD. However, there is a lack of robust evidence to support any particular pharmacological treatment. There is a need for additional studies to expand the evidence base for CUD pharmacotherapy. While medication strategies may become an integral component for CUD treatment one day, psychosocial interventions should remain the first line given the limitations in the available evidence.