ABSTRACT
OBJECTIVES: Ketamine offers a plausible mechanism with favorable kinetics in treatment of severe ethanol withdrawal. The purpose of this study is to determine if a treatment guideline using an adjunctive ketamine infusion improves outcomes in patients suffering from severe ethanol withdrawal. DESIGN: Retrospective observational cohort study. SETTING: Academic tertiary care hospital. PATIENTS: Patients admitted to the ICU and diagnosed with delirium tremens by Diagnostic and Statistical Manual of Mental Disorders V criteria. INTERVENTIONS: Pre and post guideline, all patients were treated in a symptom-triggered fashion with benzodiazepines and/or phenobarbital. Postguideline, standard symptom-triggered dosing continued as preguideline, plus, the patient was initiated on an IV ketamine infusion at 0.15-0.3 mg/kg/hr continuously until delirium resolved. Based upon withdrawal severity and degree of agitation, a ketamine bolus (0.3 mg/kg) was provided prior to continuous infusion in some patients. MEASUREMENTS AND MAIN RESULTS: A total of 63 patients were included (29 preguideline; 34 postguideline). Patients treated with ketamine were less likely to be intubated (odds ratio, 0.14; p < 0.01; 95% CI, 0.04-0.49) and had a decreased ICU stay by 2.83 days (95% CI, -5.58 to -0.089; p = 0.043). For ICU days outcome, correlation coefficients were significant for alcohol level and total benzodiazepine dosing. For hospital days outcome, correlation coefficients were significant for patient age, aspartate aminotransferase, and alanine aminotransferase level. Regression revealed the use of ketamine was associated with a nonsignificant decrease in hospital stay by 3.66 days (95% CI, -8.40 to 1.08; p = 0.13). CONCLUSIONS: Mechanistically, adjunctive therapy with ketamine may attenuate the demonstrated neuroexcitatory contribution of N-methyl-D-aspartate receptor stimulation in severe ethanol withdrawal, reduce the need for excessive gamma-aminobutyric acid agonist mediated-sedation, and limit associated morbidity. A ketamine infusion in patients with delirium tremens was associated with reduced gamma-aminobutyric acid agonist requirements, shorter ICU length of stay, lower likelihood of intubation, and a trend toward a shorter hospitalization.