This retrospective cohort study sought to identify clinical variables that independently correlate with severe alcohol withdrawal and to quantify risk in a clinically useful manner. The records of 284 inpatients admitted to an acute detoxification unit at a Veterans Affairs teaching hospital were reviewed. Clinical data were recorded on standardized forms at the time of admission and abstracted by a physician reviewer. Alcohol withdrawal severity was prospectively measured with the revised Clinical Institute Withdrawal Assessment for Alcohol (CIWA‐Ar) scale. Seventy‐one patients (25% of cohort) had severe withdrawal. We identified six independent correlates of severe withdrawal: use of a morning eye‐opener (adjusted odds ratio [OR], 5.6; 95% confidence interval [CI], 1.2–25.9), an initial CIWA‐Ar score ≥ 10 (OR, 5.1; 95% CI, 2.4–10.6), a serum aspartate aminotransferase ≥ 80 U/L (OR, 4.2; 95% CI, 2.0–8.8), past benzo‐diazepine use (OR, 3.6; 95% CI, 1.3–9.9), self‐reported history of “delirium tremens”; (OR, 2.9; 95% CI, 1.3–6.2), and prior participation in two or more alcohol treatment programs (OR, 2.6; 95% CI, 1.3–5.6). Significantly higher risk was observed in subjects with three or more independent correlates. In conclusion, several readily available clinical variables correlate with the occurrence of severe alcohol withdrawal. Ascertainment of these variables early in the course of alcohol withdrawal has the potential to improve triage and treatment decisions.