Alcohol withdrawal syndrome in critically ill patients: protocolized versus nonprotocolized management.

Background: Approximately 18% to 25% of patients with alcohol use disorders admitted to the hospital develop alcohol withdrawal syndrome (AWS). Symptom-triggered dosing of benzodiazepines (BZDs) seems to lead to shorter courses of treatment, lower cumulative BZD dose, and more rapid control of symptoms in non-critically ill patients. This study compares the outcomes of critically ill patients with AWS when treated using a protocolized, symptom-triggered, dose escalation approach versus a nonprotocolized approach.

Methods: This is a retrospective pre-post study of patients 18 years or older with AWS admitted to an intensive care unit (ICU). The preintervention cohort (PRE) was admitted between February 2008 and February 2010. The postintervention cohort (POST) was admitted between February 2012 and January 2013. The PRE patients were treated by physician preference and compared with POST patients who were given escalating doses of BZDs and/or phenobarbital according to an AWS protocol, titrating to light sedations (Richmond Agitation Sedation Scale score of 0 to -2).

Results: There were 135 episodes of AWS in 132 critically ill patients. POST patients (n = 75) were younger (50.7 [13.8] years vs. 55.7 [8.7] years, p = 0.03) than PRE patients (n = 60). Sequential Organ Failure Assessment (SOFA) scores were higher in the PRE group (6.1 [3.7] vs. 3.9 [2.9], p = 0.0004). There was a significant decrease in mean ICU length of stay from 9.6 (10.5) days to 5.2 (6.4) days (p = 0.0004) in the POST group. The POST group also had significantly fewer ventilator days (5.6 [13.9] days vs. 1.31 [5.6] days, p < 0.0001) as well as a significant decrease in BZD use (319 [1,084] mg vs. 93 [171] mg, p = 0.002). There were significant differences between the two cohorts with respect to the need for continuous sedation (p < 0.001), duration of sedation (p < 0.001), and intubation secondary to AWS (p < 0.001). In all of these outcomes, the POST cohort had a notably lower frequency of occurrence.

Conclusion: A protocolized treatment approach of AWS in critically ill patients involving symptom-triggered, dose escalations of diazepam and phenobarbital may lead to a decreased ICU length of stay, decreased time spent on mechanical ventilation, and decreased BZD requirements.

Level Of Evidence: Epidemiologic study, level III; therapeutic study, level IV.