Real-World Phenobarbital Use, Treatment Escalation, and Clinical Outcomes in Critically Ill Patients With Alcohol Withdrawal Syndrome: A Multicenter Propensity-Matched Cohort Study.

Abstract

Importance: Phenobarbital use in ICU alcohol withdrawal syndrome (AWS) may reflect severe or refractory withdrawal and treatment escalation, making multicenter data important to understand its use within ICU practice and its relationship with clinical outcomes.

Objectives: To evaluate phenobarbital use in real-world ICU AWS as a marker of severe or refractory withdrawal and treatment escalation. We also examined its association with time-to-discharge and safety outcomes compared with nonphenobarbital regimens.

Design, setting, and participants: Multicenter retrospective cohort study using multiple imputation and propensity score matching. Adults with AWS, defined as Clinical Institute Withdrawal Assessment for Alcohol, revised (CIWA-Ar) score greater than or equal to 10, were included if they received a phenobarbital-based or nonphenobarbital-based regimen in ICUs across a large academic health system from June 30, 2017, to July 1, 2024. Postoperative ICU admissions were excluded.

Main outcomes and measures: Primary outcomes were time to ICU and hospital discharge alive from treatment initiation. Safety outcomes were hypotension, new intubation, and aspiration within 48 hours.

Results: Of 1216 patients, 345 (28.4%) received phenobarbital. Before matching, phenobarbital-treated patients had higher CIWA-Ar, Acute Physiology and Chronic Health Evaluation III, and pre-ICU benzodiazepine exposure, consistent with more severe or refractory withdrawal. In adjusted time-to-discharge models, phenobarbital use was not significantly associated with time to hospital discharge alive (discharge rate ratio [DRR], 0.91; 95% CI, 0.74-1.12; p = 0.370) or ICU discharge alive (DRR, 0.77; 95% CI, 0.59-1.01; p = 0.057). Hypotension and intubation were similar, and aspiration was less frequent with phenobarbital.

Conclusions: In a contemporary ICU cohort, phenobarbital use reflected treatment escalation in patients with more severe or refractory AWS. After adjustment, time to ICU and hospital discharge alive did not differ significantly, hypotension and intubation were similar, and aspiration was less frequent with phenobarbital. Future studies should define optimal timing, dosing, and patient selection.