Single-Center Experience With Drug-Induced Liver Injury From India: Causes, Outcome, Prognosis, and Predictors of Mortality

OBJECTIVES: Although drug-induced liver injury (DILI) is rare, it may result in significant morbidity or death. The causes and outcome vary according to regions, with acetaminophen and complementary medicines common in the West and the Far East, respectively. This study evaluates the causes, outcomes, predictors, and models for 90-day mortality from DILI from India. METHODS: Consecutive patients with DILI from 1997 to 2008 based on International Consensus Criteria from a medical college hospital setting were studied. RESULTS: Of the 313 patients, 58% were males. Leading causes were a combination of four anti-tuberculous drugs (ATDs) (58%), anti-epileptics (11%), olanzapine (5.4%), and dapsone (5.4%). The overall 90-day mortality of 17.3% was significantly higher for ATD hepatitis (21.5%) vs. those without (11.4%) (P=0.02). The highest mortality was for leflunomide (75%). Seventy-eight percent of patients received more than one drug. Fulminant hepatic failure developed more commonly in females than in males (23% vs. 17%). Of the 66% of cases with jaundice and/or icterus, mortality was 26%. Multivariable models for mortality using a combination of encephalopathy, ascites, and bilirubin, or a combination of albumin, prothrombin time, and white blood cell count yielded a C-statistic of at least 0.86 by recursive partitioning and 0.92 by logistic regression. Model for end stage liver disease (MELD) scores of 38 and 46 yield probabilities of death of 0.90 (confidence interval (CI): 0.71-0.97) and 0.99 (CI: 0.90-1.00), respectively. CONCLUSIONS: DILI results in significant overall mortality (17.3%). ATDs, anti-convulsants, sulphonamides, and olanzapine are the leading causes of DILI. Although common in males, more females developed fulminant hepatic failure. High-MELD score or a combination of ascites, encephalopathy, high bilirubin, prothrombin time, and leukocyte count are predictive of mortality.