Comparative Effectiveness of Phenobarbital vs Diazepam for Alcohol Withdrawal Syndrome.

IMPORTANCE: Phenobarbital use for alcohol withdrawal syndrome is increasing, but its effectiveness and safety have largely been compared with lorazepam rather than against more pharmacokinetically similar long-acting, self-tapering benzodiazepines such as diazepam. OBJECTIVE: To emulate a clinical trial comparing patients randomized to phenobarbital vs diazepam for alcohol withdrawal syndrome. DESIGN, SETTING, AND PARTICIPANTS: This multicenter comparative effectiveness research study analyzed data from the Premier Healthcare Database (October 2022 to August 2024), representing 754 US acute care hospitals. Adult inpatients with a primary diagnosis of alcohol withdrawal syndrome present on admission who initiated phenobarbital or diazepam on the day of hospital admission were included. EXPOSURES: Initiation of phenobarbital or diazepam on the day of hospital admission. MAIN OUTCOMES AND MEASURES: The primary effectiveness outcome was hospital length of stay in days. The primary safety outcome was invasive mechanical ventilation use. Secondary outcomes assessed treatment escalation, complications of severe alcohol withdrawal syndrome, and resource utilization. Targeted maximum likelihood estimation with ensemble machine learning was used to estimate marginal mean differences and risk differences between exposure groups. RESULTS: Among 24 643 patients (median [IQR] age, 45 [35-56] years; 18 511 male [75.1%]), 11 919 (48.4%) received phenobarbital and 12 724 (51.6%) received diazepam on the day of admission. There were no differences in hospital length of stay (marginal mean, 4.1 vs 4.3 days; marginal mean difference, -0.21 [95% CI, -0.42 to 0.00] days) or invasive mechanical ventilation use (marginal risk, 3.1% vs 2.7%; marginal risk difference, 0.4% [95% CI, -0.7% to 1.4%]) between initiation of phenobarbital vs diazepam. Secondary outcomes were also not different between phenobarbital and diazepam, except for self-directed discharge, which was more frequent among those initiated on diazepam (marginal risk, 9.5% vs 12.6%; marginal risk difference, -3.1% [95% CI, -4.9% to -1.3%]). CONCLUSIONS AND RELEVANCE: In this multicenter study of patients hospitalized for alcohol withdrawal syndrome, there were no differences in hospital length of stay or the need for invasive mechanical ventilation between initiation of phenobarbital vs diazepam. These results may suggest near therapeutic equivalence between phenobarbital and long-acting benzodiazepines.