Phenobarbital and Diazepam Show Similar Outcomes in Hospitalized Patients With Alcohol Withdrawal

As phenobarbital use for alcohol withdrawal increases, comparisons with long-acting benzodiazepines remain important. This multicenter observational study emulated a clinical trial using the Premier Healthcare Database, representing 754 US acute care hospitals, from October 2022 through August 2024. It included 24,643 adults with a primary diagnosis of alcohol withdrawal present on admission who started phenobarbital or diazepam that day. Primary outcomes were hospital length of stay and invasive mechanical ventilation.

Among participants, 11,919 received phenobarbital and 12,724 received diazepam. Estimated marginal mean length of stay was 4.1 versus 4.3 days, respectively (difference, −0.21 days; 95% CI, −0.42 to 0.00). Estimated ventilation risk was 3.1% versus 2.7% (difference, 0.4 percentage points; 95% CI, −0.7 to 1.4). Secondary outcomes also showed no differences except self-directed discharge, which was less frequent with phenobarbital: 9.5% versus 12.6% (difference, −3.1 percentage points; 95% CI, −4.9 to −1.3).

These findings suggest similar inpatient effectiveness and safety for the two initial treatment strategies. However, treatment was not randomized, and residual confounding remains possible despite statistical adjustment using targeted maximum likelihood estimation and ensemble machine learning. Absence of a detected difference does not establish equivalence. Only the abstract was available for review.