SAEM GRACE: Phenobarbital for alcohol withdrawal management in the emergency department: A systematic review of direct evidence

OBJECTIVES: Alcohol withdrawal syndrome (AWS) is a commonly presenting condition in the emergency department (ED) and can have severe complications, including mortality. Benzodiazepines are first-line medications for treating AWS but may be unavailable or insufficient. This systematic review evaluates the direct evidence assessing the utility of phenobarbital for treating AWS in the ED. METHODS: A systematic search was conducted and designed according to the patient-intervention-comparator-outcome (PICO) question: (P) adults (≥18 years old) presenting to the ED with alcohol withdrawal; (I) phenobarbital (including adjunctive); (C) benzodiazepines or no intervention; and (O) AWS complications, admission to a monitored setting, control of symptoms, adverse effects, and adjunctive medications. Two reviewers independently assessed articles for inclusion and conducted risk of bias assessments for included studies. RESULTS: From 70 potentially relevant articles, seven studies met inclusion criteria: three retrospective cohort studies, two retrospective chart reviews, and two randomized controlled trials (RCTs), one examining phenobarbital monotherapy and one examining adjunctive phenobarbital. Across the retrospective cohort studies, treatment of AWS with phenobarbital resulted in lower odds of a subsequent ED visit. The retrospective chart reviews indicated that phenobarbital was associated with higher discharge rate compared to benzodiazepine-only treatments. For the two RCTs, phenobarbital did not differ significantly from benzodiazepine for most outcomes, although concomitant treatment with phenobarbital was associated with lower benzodiazepine use and intensive care unit admission. The heterogeneous designs and small number of studies prevented quantitative synthesis. CONCLUSIONS: Relatively few studies provide direct evidence on the utility of phenobarbital for AWS in the ED, but the evidence that exists generally suggests that it is a reasonable and appropriate approach. Additional RCTs and other methodologically rigorous investigations are needed for more definitive direct evidence.