Risk factors for delirium tremens in patients with alcohol withdrawal syndrome in a hospital setting

Abstract Background The clinical expressions of alcohol withdrawal syndrome (AWS) may vary and the factors determining these variations are not well-known. It would be useful to have a set of clinical tools capable of predicting which patients are likely to develop the more severe forms of the syndrome. Aim To analyse the clinical variables associated with the development of delirium tremens (DTs) in patients who were admitted to a general hospital with AWS. Methods Cohort study of AWS patients admitted to the Xeral Hospital in Lugo, Spain, between 1987 and 2003. The characteristics of patients with the syndrome who did not develop delirium tremens were contrasted with those who did. Cases presenting with DTs at diagnosis were excluded. The different clinical, epidemiological and biochemical variables reflective of alcohol consumption habits, basal health status and presentation features of syndrome were all recorded. Results Data from 156 episodes of AWS that coursed without DTs were compared with 147 cases that coursed with it. Three independent variables for development of DTs were identified in a multivariate logistic regression model: number of seizures [1 or 2: OR 2.2 (CI 95% 1.2–3.8), p  = 0.005; 3 or more: OR 2.6 (CI 95% 1.04–6.8), p  = 0.04]; systolic blood pressure > 150 mm Hg [OR 1.9 (CI 95% 1.1–3.8), p  = 0.03] and axillary temperature > 38 °C [OR 1.9 (CI 95% 1.05–3.5), p  = 0.01]. ROC analysis revealed an area under the curve of 0.679. Conclusion Three clinical findings (seizures, blood pressure and temperature) can aid in identifying patients with AWS who are likely to develop DTs. The model's predictive capacity is not high.