To determine the characteristics associated with an increased risk for delirium tremens (DT) we performed a case-control study at the detoxification units of two hospitals. Cases met DSM-IV criteria for DT. For each case (n = 15), 3 controls (n = 45) were chosen. Eligibility criteria were applied equally to cases and controls. Cases were more likely than controls to report a prior complicated withdrawal (DT or alcohol withdrawal seizure) (53 vs. 27%, OR 3.1, 95% CI 0.94-10.55), have a systolic blood pressure greater than 145 mm Hg on admission (60 vs. 27%, OR 4.1, 95% CI 1.21-14.06), and have comorbidity scores of at least 1 (60 vs. 18%, OR 6.9, 95% CI 1.92-25.08). Zero cases (0%) and 15 (33%) controls had no prior complicated withdrawals and no adverse clinical features (systolic blood pressure >145 or comorbidity score >1). Compared to this group, the odds of being a case and having both prior complicated withdrawal and at least 1 adverse clinical feature was 44.8 (95% CI 4.36-460). Elevated blood pressure, prior complicated alcohol withdrawal and medical comorbidity, alone and in combination, are associated with an increased risk of delirium tremens.