Objective: In a department of hepatology and gastroenterology, a significant number of patients are hospitalized for alcohol withdrawal. The aim of this retrospective study was to identify factors predictive of severe or complicated alcohol withdrawal in order to improve patient management.
Methods: Between June 2002 and June 2005, 182 patients admitted for alcohol dependence according to the DSM-IV classification were enrolled in this study. A unique management protocol for alcohol withdrawal was applied for all patients. The Cushman score was recorded on day 1, 2 and 3 to assess the severity of alcohol withdrawal. We searched for correlations between epidemiological, clinical and biological data and the Cushman score.
Result: The study population included 136 (74.7%) men and 46 (25.3%) women, mean age 47.6+/-10.1 years. One hundred and eighteen patients (64.8%) were referred from a specialized outpatient clinic and 64 (35.2%) patients were referred from the emergency unit. The mean and median Cushman scores on day 1, 2 and 3 were: 5.1 and 5; 3.9 and 4; 2.3 and 2, respectively. Twenty patients (11.0%) and five patients (2.7%) had scores greater than or equal to 8 and greater than 12, respectively. The proportion of patients with Cushman score greater than or equal to 8 on day 1 was significantly greater in patients referred from the emergency unit than in those referred from a specialized outpatient clinic (p=0.002). Mean alanine aminotransferase level on day 1 was significantly higher in patients with a score greater than or equal to 8 than in those who had a score less than 8 (112.1+/-44.4 UI/L versus 78.4+/-11.8 UI/L; p=0.046). Referral via an emergency unit as well as an alanine aminotransferase level greater than 1.5fold the upper limit of the normal range were independent predictive factors for a Cushman score greater than or equal to 8. In conclusion, severe alcohol withdrawal (Cushman score>or=8) is significantly associated with initial management in an emergency unit and serum alanine aminotransferase level greater than 1.5 fold the upper limit of the normal range. These predictors should be monitored in order to appropriately adapt the therapeutic schedule.