2026 Southeastern Residency Conference

Background: The American Society of Addiction Medicine guidelines recommend benzodiazepines as first-line treatment for moderate-severe alcohol withdrawal syndrome (AWS). Intravenous (IV) lorazepam is a first-line therapy for AWS at Emory Healthcare (EHC) to prevent complications such as seizures and delirium tremens. The quick onset, half-life, and lack of metabolites make IV lorazepam a desirable agent for AWS. There have been intermittent shortages of IV lorazepam, with the most recent shortage beginning in May 2025. In response, EHC built an electronic alert as a soft stop for all IV lorazepam orders to notify providers of the shortage and recommend alternatives. In addition, the AWS order set was modified to replace IV lorazepam with IV midazolam. The goal of this study was to understand medication utilization patterns during this shortage and provide guidance for future shortages.  

Methods: This study was a multi-center, retrospective chart review of adult patients receiving treatment for moderate-severe AWS from 1/30/2025 to 9/7/2025 in the ICU. Pre-shortage and post-shortage groups were determined based on IV lorazepam shortage alerts, which began in EPIC on 5/20/2025. The primary objective was to examine medication utilization patterns of AWS before and after the alert was implemented. Evaluation of this endpoint was determined by medication name, strength, route, frequency, and duration for 48 hours after the alcohol withdrawal order set was active. Secondary objectives measured safety and clinical outcomes and included respiratory rates, hypotension, and changes in scores of Clinical Institute Withdrawal Assessment Alcohol Scale (CIWA), length of stay, mortality, 30-day readmissions, respectively. Data was analyzed using descriptive statistics, t-test, and Mann-Whitney U test.

Results: There were 127 patients in the pre-shortage group and 152 patients in the post-shortage group. Baseline characteristics were similar between groups. The average baseline CIWA scores were similar with first CIWA ≥8 13.44 pre-shortage versus 12.86 post-shortage. IV lorazepam doses were higher in the first 12 hours of the AWS order set ordering time. In the post-shortage group, there was a greater decrease in IV lorazepam dose 12-24 hours post-alert (63.14% versus 43.72%). IV lorazepam orders trended down (57% versus 47%, p-value 0.097), and patients prescribed oral lorazepam trended up (23% versus 32%, p-value 0.081). More patients in the post-shortage group were prescribed chlordiazepoxide (5 versus 16 patients, p-value 0.038) and midazolam (20 versus 45 patients, p-value 0.006). Phenobarbital use did not change significantly (86 versus 89 patients, p-value 0.115). First CIWA <8, was similar at 4.37 pre-shortage versus 3.84 post-shortage. The average time to CIWA score <8 from the first score ≥8 was greater in the pre-shortage group 9 hours and 13 minutes compared to 8 hours and 37 minutes (p-value 0.205). Average hospital length of stay (12.39 versus 11.07 days, p-value 0.290), average ICU length of stay (5.28 versus 5.44 days, p-value 0.725), and mortality (5 versus 7 patients, p-value 0.784) were similar between the two groups. Readmission within 30 days and respiratory depression were higher in the post-shortage group (12% versus 18%, p-value 0.166 and 67% versus 73%, p-value 0.267, respectively).

Conclusion: Following IV lorazepam alerts, medication utilization shifted towards chlordiazepoxide and midazolam. The average cumulative dose of IV lorazepam decreased more in the post-shortage group after the first 12 hours. Phenobarbital use may not have changed due to provider practice styles and the recommended alternatives. This was a retrospective study, alerts for alternative agents could be bypassed, and medications listed may have been used for other indications. Future direction includes evaluating alternative medication usage patterns with a developed phenobarbital order set. A comparative analysis focused on adverse drug events and efficacy may further shape treatment.