Hospital-Wide Implementation, Clinical Outcomes, and Safety of Phenobarbital for Alcohol Withdrawal

Abstract

Importance: Phenobarbital monotherapy is increasingly used to treat alcohol withdrawal syndrome (AWS) in hospitalized patients, but its implementation outside academic emergency departments (EDs) and intensive care units (ICUs) remains understudied.

Objective: To evaluate the implementation, clinical outcomes, and safety of an electronic health record (EHR) order set for intravenous weight-based phenobarbital loading (10 or 15 mg/kg) across all care locations (ED, acute care, and ICU) of a community hospital.

Design, setting, and participants: This retrospective quality improvement study was conducted at a 281-bed community hospital in Seattle, Washington, including all hospitalized adults treated for AWS 11 months before and 12 months after implementation of the phenobarbital EHR order set (April 1, 2021, to March 31, 2023). Data analysis occurred in May 2024.

Exposure: Initiation of AWS treatment before vs on or after March 24, 2022.

Main outcomes and measures: Implementation was assessed by order set use and cumulative phenobarbital and benzodiazepine exposure. Clinical outcomes included daily maximum Clinical Institute Withdrawal Assessment of Alcohol Scale Revised (CIWA-Ar) scores, AWS treatment duration (time from first to last medication), and time from AWS treatment initiation to hospital discharge. Safety outcomes included prolonged use of physical restraints (>1 day), intubation, and in-hospital mortality. Logistic and negative binomial regression models were used to estimate differences in each outcome, before vs after implementation, adjusting for baseline group differences.

Results: Among 254 patients, the preimplementation (154 patients) and postimplementation (100 patients) groups were similar in terms of baseline characteristics; most were middle-aged (mean [SD] age, 53.0 [14.9] years), male (177 men [66.9%]), and publicly insured (186 patients [73.2%]). After implementation, the phenobarbital order set was used for 67 patients (67.0%), predominantly in ED and acute care locations. AWS symptoms resolved more rapidly after implementation, with a 4.2- to 5.0-point reduction in daily maximum CIWA-Ar scores at 24 to 96 hours from hospital presentation, 30.1-hour reduction in AWS treatment duration (95% CI, 16.7-43.5 hours), and 2.2-day reduction in time to hospital discharge (95% CI, 0.7-3.7 days). Safety outcomes did not significantly differ before and after implementation.

Conclusions and relevance: In this quality improvement study, implementation of a weight-based phenobarbital order set for AWS in all care locations of a community hospital was associated with faster symptom resolution, shorter length of stay, and no differences in safety outcomes. These findings support the broader adoption of phenobarbital-based AWS treatment across diverse hospital settings.

Figure 1.. Daily Maximum Clinical Institute Withdrawal Assessment for Alcohol Revised (CIWA-Ar) Scores Before and After Implementation of the Phenobarbital Electronic Health Record Order Set, Adjusted for Baseline Differences Between Groups

Depicted results are recycled estimates from the negative binomial regression model adjusting for baseline group differences (standardized mean difference >0.15) regarding homelessness, brain injury, other substance use disorder excluding alcohol, and trauma (see eTable 4 in Supplement 1). Error bars denote IQRs.

Figure 2.. Temporal Trends in the Proportion of Patients Treated With the Phenobarbital Electronic Health Record Order Set, the Proportion With Elevated Clinical Institute Withdrawal Assessment for Alcohol Revised (CIWA-Ar) Scores (≥10) at 24 to 48 Hours, and the Median Time From Alcohol Withdrawal Syndrome Treatment Initiation to Hospital Discharge

On September 15, 2022, a national benzodiazepine shortage prompted addition of a best practice alert upon opening the benzodiazepine order set, encouraging use of the phenobarbital order set. Q indicates quarter.