A meta-analysis of cocaine use disorder treatment effectiveness

Background: In both the U.S. and U.K., after a period of decline, prevalence of cocaine use has been increasing since 2012 and is now the second leading cause of overdose death from an illicit drug. However, psychosocial treatments for cocaine use disorders are limited, and no pharmacotherapy is approved by regulatory bodies in the U.S. or Europe. We performed a comprehensive meta-analysis to assess treatments impacts on cocaine use. Methods: We performed a meta-analysis of clinical trials that included the word "cocaine" in the title and were published between 31/12/1995 and 31/12/2017. All studies of outpatient adults with active cocaine use and reporting urinalysis results were included. Treatment approaches were clustered into 11 categories. Missing data were imputed using multiple imputation by chained equations. We calculated intention-to-treat log-odds ratios (OR) for the change in proportion of patients testing negative for cocaine at the end of each study and performed multivariate mixed-effects meta-regression. This study was prospectively registered on covidence.org on 31/12/2015, study 8731. Findings: One hundred fifty-seven studies with 15,842 participants were included. Only contingency management was significantly associated with increased odds of testing negative for cocaine (OR of 2.13, 95% CI 1.62-2.80) and remained significant after all sensitivity analyses. Interpretation: This meta-analysis is unique in its broad inclusivity of treatment types and trial designs over a two-decade period of investigation. Our results converge with focused meta-analyses on treatments for cocaine use disorders; thus, research efforts and policy changes that expanded implementation of contingency management programs are expected to reduce cocaine use in active users and the associated individual, community, and societal burdens associated.