In Reply We appreciate the thoughtful letter by Dr Shi regarding our systematic review and meta-analysis 1 of pharmacotherapies for alcohol use disorder. Regarding the metaanalyses that included multiple doses, the analyses were stratified by dose whenever possible (eg, 50-mg oral naltrexone and 100-mg oral naltrexone). 1 More details are provided in the full evidence report. 2 In the case of injectable naltrexone, there were simply too few studies to stratify the analysis by dose. Given the findings from those studies, our conclusions would not change. The same is true for the acamprosate studies noted by Shi. Seven studies used a range of doses, presumably based on individual participants' ability to tolerate the medication. There was an insufficient number of studies for a valid dose comparison. For the 2 studies that had multiple groups using different doses, no dose showed superiority. Therefore, current evidence does not suggest a difference in outcomes by acamprosate dose.Regarding disulfiram, evidence from well-controlled studies does not support efficacy compared with placebo for improving alcohol consumption outcomes. 2 However, some trials did report a reduction in drinking days among those who returned to any drinking and who had a complete set of assessment interviews. It is therefore possible that disulfiram could be beneficial for some patients.There are challenges in designing trials to adequately evaluate a medication, such as disulfiram, with its unique deterrent mechanism. The available trials do not establish efficacy. However, the trials do suggest that combination programs of counseling, support, and coaching along with disulfiram may work for motivated patients who are interested in taking disulfiram and in adhering to the medication.In addition, the reported effects of pharmacotherapy for patients with alcohol use disorder result from a combination of medication and psychosocial interventions. 2 In other words, the benefits observed reflect the added benefit of medications beyond those of counseling and placebo, and there is uncertainty regarding the benefits of the medications when used alone without psychosocial co-interventions.