Retrospective Evaluation of Phenobarbital Compared to Standard Benzodiazepine Therapy for Alcohol Withdrawal Patients in the Icu at an Academic Medical Center

Abstract

Phenobarbital has long been studied in the setting of acute Alcohol Withdrawal Syndrome (AWS), yet data is lacking for practical protocol utilization in hospitals. The purpose of this study was to determine if the use of phenobarbital decreases benzodiazepine requirement in alcohol withdrawal patients requiring ICU care. We evaluated patients ages 19 and older with AWS who were admitted to the ICU with CIWA scores over 20. Patients who received phenobarbital with benzodiazepines were compared to those who received benzodiazepines alone. METHODS:This study was approved by the Institutional Review Board. A retrospective chart review was conducted of 134 patients being treated for AWS between June 1, 2017 to October 31, 2020 who received a dose of phenobarbital with benzodiazepines (n = 69) or benzodiazepines alone (n = 65) for AWS. The primary outcome assessed is benzodiazepine utilization, measured as benzodiazepine equivalents by converting non-lorazepam doses to an equal dose of lorazepam. Secondary outcomes included hospital length of stay (LOS), ICU LOS, CIWA scores, use of ventilator, delirium (CAM-ICU days), and hours on dexmedetomidine. RESULTS:There was no statistically significant difference in benzodiazepine equivalents in the phenobarbital group vs the benzodiazepine group (64.1 vs. 62.5, p = 0.365 in Spearman's Correlation and 0.379 in Mann Whitney U Test). Statistical significance was found in hospital LOS (10.0 days vs 14.4 days, p = 0.012) and positive CAM-ICU days (1.6 days vs. 3.0 days, p = 0.005) in favor of the phenobarbital group. The study population was mostly male (85.5% in phenobarbital group and 84.6% in control group). Baseline characteristics were not significant between the two groups with the exception of age. Patients who received phenobarbital were younger in median age (47 years vs. 61 years; p < 0.05). When controlled for age, hospital LOS was no longer statistically significant. Differences in favor of the phenobarbital group were found in ICU LOS (7.3 days vs 8.0 days), time on dexmedetomidine (42.0 days vs. 45.5 days), and need for mechanical ventilation (13% vs. 21.5%), however these differences were not statistically significant. Average CIWA scores were higher in the phenobarbital group compared to the control group (16.4 vs 15.6), however this also did not meet statistical significance. CONCLUSIONS:The present retrospective analysis suggests that phenobarbital administration does not decrease the utilization of benzodiazepines in AWS. However, a significant decrease in delirium was shown in the phenobarbital group. Given the evidence that favorable outcomes are shown and safety outcomes are similar, phenobarbital should be explored further for incorporation into institutions' standard AWS protocol.CLINICAL IMPLICATIONS: A quality improvement project with a goal of incorporating phenobarbital use early in AWS hospital stay should be considered.