Abstract Purpose Benzodiazepines remain the standard therapy for alcohol withdrawal syndrome (AWS), though phenobarbital is increasingly adopted as an alternative due to its predictable pharmacology and potential to reduce escalation of care. High-quality comparative effectiveness data remain limited. We evaluated short-term outcomes of phenobarbital monotherapy compared with benzodiazepines in hospitalized patients with AWS using a large federated electronic health record (EHR) network. Methods We conducted a retrospective cohort study using TriNetX’s Global Collaborative Network (154 health care organizations). Adults (≥18 years) admitted with AWS (ICD-10 F10.x codes) were included. Data were extracted on August 25, 2025. Cohort 1 received phenobarbital without benzodiazepines within 1 day of AWS diagnosis; Cohort 2 received benzodiazepines without phenobarbital. Exclusion criteria included pregnancy, major trauma, or peri-intubation medication use. Outcomes were measured within 10 days of index hospitalization: critical care utilization, mechanical ventilation, aspiration pneumonitis, seizures, delirium tremens (DTs), vasopressor requirement, additional sedative/antipsychotic use, and mortality. Propensity score matching (PSM) 1:1 was performed on demographics, psychiatric comorbidities, substance use disorders, and baseline labs. Risk ratios (RR) with 95% confidence intervals were reported. Results From 5,117 phenobarbital and 199,379 benzodiazepine encounters, 4,712 matched pairs were generated. After PSM, groups were balanced across age, sex, race, psychiatric and hepatic comorbidities, and substance use disorders. Compared with benzodiazepines, phenobarbital was associated with significantly lower risk of seizures (3.3% vs 4.7%; RR 0.70, 95% CI 0.57-0.86, p 0.001) and delirium tremens (2.7% vs 5.3%; RR 0.52, 95% CI 0.42-0.64, p 0.001). Phenobarbital also reduced need for additional sedatives/antipsychotics (29.2% vs 32.9%; RR 0.89, 95% CI 0.84-0.94, p 0.001). There were no significant differences in aspiration pneumonitis (1.1% vs 1.6%; p = 0.075), vasopressor use (0.6% vs 0.8%; p = 0.46), or mortality (0.9% vs 1.0%; p = 0.77). Rates of mechanical ventilation were lower with phenobarbital (2.9% vs 3.5%; RR 0.82, 95% CI 0.66-1.03; p = 0.08). However it was statistically not significant. Unexpectedly, critical care utilization was higher with phenobarbital (16.8% vs 9.0%; RR 1.85, 95% CI 1.66-2.07; p 0.001). Conclusion In this large, propensity-matched cohort, phenobarbital monotherapy for AWS was associated with lower risks of seizures, delirium tremens, and need for adjunctive sedatives compared with benzodiazepines, without increasing mortality, mechanical ventilation, or aspiration. However, higher ICU utilization in the phenobarbital cohort warrants further investigation, potentially reflecting institutional practice patterns or severity selection. This abstract is funded by: None
Le et al. (2026) studied this question.