Predictors of mortality in patients with delirium tremens

Abstract

Objectives: The objectives were to identify factors that may help predict mortality for patients with delirium tremens (DT). Methods: The authors conducted a 1:1 gender- and age-matched case-control study of patients hospitalized for DT. Using McNemar chi-square tests and conditional logistic regression (CLR), risk factors for death, including demographics, location of diagnosis, vital sign derangements, treatment methods, and comorbid conditions, were evaluated. Crude and adjusted odds ratios (OR) and 95% confidence intervals (CI) are reported. Results: Thirty-five patients with DT died between January 2000 and June 2006. The majority (31; 88.6%) were male with a mean (±standard deviation [SD]) age of 51.7 (±7.6) years. Hyperthermia in the first 24 hours of DT diagnosis (OR = 10.0, 95% CI = 2.3 to 42.7), persistent tachycardia (OR = 24.0, 95% CI = 3.3 to 177.4), and use of restraints (OR = 7.50, 95% CI = 1.7 to 32.8) were associated with increased mortality by univariate analysis, while an emergency department (ED) diagnosis of DT (OR = 0.18, 95% CI = 0.05 to 0.6) and use of clonidine (OR = 0.10, 95% CI = 0.01 to 0.78) were associated with decreased mortality. In the CLR model, restraint use and hyperthermia were the only variables that remained significant (OR = 5.8, 95% CI = 1.0 to 32.2; and OR = 6.1, 95% CI = 1.2 to 30.4, respectively). Conclusions: The use of restraints and hyperthermia is associated with increased odds of death for patients with DT. This study highlights the need for further research into modifiable factors influencing mortality from DT. © 2008 by the Society for Academic Emergency Medicine.