Comparing length of stay between adult patients admitted to the intensive care unit with alcohol withdrawal syndrome treated with phenobarbital versus lorazepam

Abstract

Objective: Review the use of phenobarbital and lorazepam to treat alcohol withdrawal syndrome (AWS) in the ICU, compare length of stay, examine medication use trends, implement provider training, and evaluate outcomes post-training. 
Methods: Design - Retrospective observational study, Quality improvement. Setting - Tertiary care hospital with 36 ICU beds. Patients - Adults admitted to the ICU and placed on clinical institute withdrawal assessment (CIWA) protocol. Patients with epilepsy were excluded. 
Results: During the 34-month baseline period, 713 patients were admitted to the ICU with alcohol use disorder (AUD) on CIWA, without epilepsy. 189 patients were treated with phenobarbital, 460 patients received only lorazepam, 64 patients received neither medication. All but 2 of the patients who received phenobarbital also received lorazepam. Compared to phenobarbital, lorazepam-only patients had shorter ICU LOS (p<0.001, 95% CI -2.36, -1.32) but higher mortality 13.91% vs. 4.76% (p=0.0008). We then developed and provided a training (which we refer to as an “intervention”) to all ICU providers encouraging consistent use of phenobarbital in the ICU when appropriate. During the 3-month post-intervention period 44 patients were admitted to the ICU with AUD on CIWA protocol. Of 
the 44 patients: 26 received phenobarbital, 12 received only lorazepam, 6 received neither medication. Significantly more patients were treated with only phenobarbital (57.69%)) compared to baseline (1.06%). Compared to patients treated with phenobarbital, patients treated with only lorazepam had a significantly higher mortality rate (33.33% vs. 7.69%, p=0.04).
Conclusions: We found significant variability in the use of phenobarbital and lorazepam for treatment of AWS in the ICU. After a quality improvement training for ICU physicians, there was less frequent concurrent use of benzodiazepines and barbiturates, no difference in ICU or hospital LOS, and significantly lower mortality rate for those treated with phenobarbital.