Effect of Phenobarbital-Based Alcohol Withdrawal Protocol on Provider Practice and Patient Outcomes-A Quality...

Introduction: Alcohol is the most common substance use disorder in the United States. Despite this prevalence, there remains significant heterogeneity in medical management of alcohol withdrawal syndrome (AWS). While the 2020 American Society of Addition Medicine continues to recommend the use of benzodiazepines as first-line therapy for AWS, there is increasing use of phenobarbital in patients at high risk of severe AWS. Despite phenobarbital's favorable pharmacologic profile, historically, clinical utilization on general medicine services has been low and often restricted. In this project, we have examined practice patterns and associated clinical outcomes in adult patients experiencing AWS on the general medicine service pre and post implementation of a phenobarbital-based protocol for the treatment of severe AWS at our institution. Methods: This quality improvement study evaluated changes in management of AWS on general medicine units associated with implementation of a phenobarbital-based protocol and order set in the electronic medical record (EMR). Our primary outcome measures were receipt of a phenobarbital loading dose, concomitant benzodiazepine administration, and total benzodiazepine dose. Safety outcomes were also explored to assess clinical impacts of this protocol implementation. The project was determined "not research" by our Institutional Review Board. Results: Phenobarbital-protocol implementation was associated with increased frequency of receiving a phenobarbital loading dose (49.5% vs 9.4%; P < .001), decreased use of concomitant benzodiazepine/phenobarbital (4.3% vs 28.9%; P < .001), and decreased total benzodiazepine dose (7.8 vs 15.5 mg; P < .001). Regarding safety, there was no significant pre/post difference in the rate of ICU transfer, but among those transferred there was a trend toward decreased mechanical ventilation rate (100% vs 28.6%; P = .051), and a significantly reduced ICU length of stay (median 11 vs 3 days; P = .04). There were no pre/post differences in seizures, delirium or use of adjunct medications. Conclusions: This quality improvement study demonstrates a marked change in provider prescribing practices for treating AWS after implementation of an institutional phenobarbital-based protocol. We observed no difference in overall clinical outcomes after protocol implementation, although a larger follow-up study is needed to confirm this and to further explore the shorter ICU length of stay for patients with AWS postimplementation.