Phenobarbital and/or benzodiazepines for recurrent alcohol withdrawal: A self-controlled, retrospective cohort study

While there is ample data supporting the use of barbiturates and benzodiazepines (BZDs) for the treatment of alcohol withdrawal, there is a paucity of information on treating recurrent withdrawal among high healthcare utilizing patients. The purpose of this study was to assess the efficacy and safety of phenobarbital (PB), with or without adjuvant BZDs, for treatment of acute alcohol withdrawal in the emergency department (ED) in patients with high rates of recurrent withdrawal.

This non-matched, self-controlled, retrospective cohort study evaluated patients seen in the ED of an urban trauma center and safety-net teaching hospital between July 1st, 2018, and July 31st, 2019. Patients treated for alcohol withdrawal were included if they had at least one visit where they received intravenous PB with or without BZDs, then during a separate encounter received BZD only. Each encounter was then assigned to a treatment group based on administration of PB only, BZD only, or the combination of PB and BZD. The primary outcomes were admission to hospital or discharge and return to the ED for any reason within 48 h of disposition.

A total of 137 unique patients were included, with 642 encounters composed of 245 PB only, 293 BZD only, and 104 combination visitations. No significant difference was found between the PB, BZD, or combination treatment groups for rates of admission (36.7%, 38.9%, and 46.1% respectively) or for return within 48 h (17.1%, 15.0%, and 13.5%). There was a significantly longer ED length of stay for the combination group (8.6 h) compared to either the PB or BZD only groups (6.4 and 7.0 h, respectively, p < 0.05) but not between the monotherapy groups. There were significantly higher rates of ICU admission and hypotension when PB and BZDs were used together (8.6% and 15.4%) versus either agent alone (PB 2.9% and 5.7%, BZD 3.8% and 4.5%, p < 0.05).

Among patients with multiple visits presenting with alcohol withdrawal, treatment with PB, BZDs, or both did not result in significantly different rates of admission or readmission within 48 h. Receiving a combination of PB and BZDs was associated with significantly longer ED length of stay, more ICU care, and increased incidence of hypotension as compared to either PB or a BZD alone.