Clinical and Laboratory Predictors of All Causes Deaths and Alcohol-Attributable Deaths Among Discharged Alcohol-Dependent Patients


Reprint requests: Jin Pyo Hong, MD, PhD, Department of Psychiatry, Asan Medical Center, 388-1 Pungnap-2dong, Songpa-gu, Seoul 138-736, South Korea; Tel.: 82230103421; Fax: 8224858381; E-mail:



Alcohol dependence is associated with increased mortality, but it is not clear whether certain laboratory abnormalities on admission predict premature death in discharged alcoholic patients. We examined total deaths and alcohol-attributable deaths among discharged alcohol-dependent patients in Korea, and the associated risk factors including laboratory abnormalities.


The subjects consisted of 442 patients who were admitted for alcohol dependence to a general hospital located in Seoul, Korea from January 1989 to December 2006. At the end of 2009, the risk of death among the discharged alcohol-dependent patients was compared with that of the general population of Korea by calculating standardized mortality ratios. Cox proportional hazards regression models were used to analyze the risk factors for total deaths and alcohol-attributable deaths in these patients.


Twenty-nine percent (127/442) of the discharged alcoholic patients died during the study period. Mortality among these patients was 6.67 times higher than in the Korean population as a whole (7.12 times for men and 2.62 times for women). The most common causes of death were alcoholic liver disease and unspecified liver cirrhosis. Recurrent admission and low albumin level at admission were independently associated with both all causes and alcohol-attributable mortality. High bilirubin level at admission was independently associated with alcohol-attributable mortality.


These observations indicate a high risk of death in discharged alcohol-dependent patients, particularly those who are repeatedly admitted to psychiatric wards or who yield abnormal laboratory findings.