Drug induced liver injury: accuracy of diagnosis in published reports.

The diagnosis of drug induced liver injury (DILI) is based primarily on the exclusion of alternative causes. To assess the frequency of alternative causes in initially suspected DILI cases, we searched the Medline database with the following terms: drug hepatotoxicity, drug induced liver injury, and hepatotoxic drugs. For each term, we used the first 100 publications. We reviewed references, selected those reports relevant to our study, and retrieved finally 15 publications related to DILI and alternative causes. A total of 2,906 cases of initially assumed DILI were analyzed in these 15 publications, with diagnoses missed in 14% of the cases due to overt alternative causes. In another 11%, the diagnosis of DILI could not be established because of confounding variables. Alternative diagnoses included hepatitis B, C, and E, CMV, EBV, ischemic hepatitis, autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, hemochromatosis, Wilson's disease, Gilbert's syndrome, fatty liver, non alcoholic steatohepatitis, alcoholic liver diseases, cardiac and thyroid causes, rhabdomyolysis, polymyositis, postictal state, tumors, lymphomas, chlamydial and HIV infections. Causality assessment methods applied in these 15 publications were the CIOMS (Council for International Organizations of Medical Sciences) scale alone (n = 5) or combined with the Maria and Victorino (MV) scale (n = 1), the DILIN (Drug-Induced Liver Injury Network) method (n = 4), or the Naranjo scale (n = 1); the qualitative CIOMS method alone (n = 3) or combined with the MV scale (n = 1). In conclusion, alternative diagnoses are common in primarily suspected DILI cases and should be excluded early in future cases, requiring a thorough clinical and causality assessment.