Evaluation of Phenobarbital and Benzodiazepines for Alcohol Withdrawal Treatment in the Medical ICU

Introduction: There is a high prevalence of alcohol use disorder (AUD) in critically ill patients leading to a significant risk of alcohol withdrawal syndrome (AWS). However, despite existing guidelines, there is limited evidence on the use of phenobarbital for AWS as part of a standardized protocol for medical intensive care unit (ICU) patients. This study aims to compare ICU length of stay for patients receiving phenobarbital and benzodiazepines versus benzodiazepines alone when treating AWS.


Methods: This study was a single-center, retrospective, pre-post analysis that compared pre-pathway patients admitted between January 2017 to January 2020 who received benzodiazepines alone to post-pathway patients admitted between August 2022 to September 2024 who received phenobarbital in addition to benzodiazepines for AWS. Patients were included if they were hospitalized with an ICD-10 code for alcohol withdrawal, admitted to the ICU with medical critical care as the primary team, and, for the post-pathway group, must have received phenobarbital from the pathway. The primary outcome was ICU length of stay, while secondary outcomes included hospital length of stay, incidence and length of mechanical ventilation, cumulative dose of benzodiazepines, incidence of delirium, and use of adjunctive medications for AWS.


Results: 132 patients were included in analysis (n=92 pre-pathway; n=40 post-pathway). The median ICU length of stay for the pre-pathway group was 2.95 days compared to 2.60 days for post-pathway group (p=0.86). The median hospital length of stay was statistically significant at 8.9 days for the pre-pathway group compared to 6 days for the post-pathway group (p=0.02). The average lorazepam equivalent dose in the ICU was 22.935 mg for the pre-pathway group compared to 12 mg for the post-pathway group (p=0.013). There was no significant difference in mechanical ventilation, use of anti-psychotics, or use of adjunctive agents for AWS.


Conclusions: There was no significant difference found in the ICU length of stay between the pre-pathway group and the post-pathway group. However, the hospital length of stay for the post-pathway group was statistically shorter than the pre-pathway group. Therefore, phenobarbital seems to be safe and effective for AWS, and benzodiazepine-sparing regimens should be considered.