Antimitochondrial rather than antinuclear antibodies correlate with severe drug-induced liver injury.

Introduction A proportion of patients with drug-induced liver injury (DILI) present with autoantibodies, which has led to the current concept of autoimmune-like DILI. However, no standardized definition exits and the clinical relevance has not been studied in detail yet. Methods 1434 patients with DILI enrolled in a prospective study were analyzed. DILI diagnosis was based on the monocyte-derived hepatocyte-like (MH) cell test and supported by Roussel Uclaf Causality Assessment Method (RUCAM) and expert adjudication. Testing for antinuclear antibodies (ANA) and anti-mitochondrial antibodies (AMA) were performed using immunofluorescence. ANA titers ≥ 1:100 were considered positive and ≥ 1:400 clinically relevant; AMA positivity was considered at titers ≥ 1:100. Results 67 % exhibited ANA ≥ 1:100 and 298 % ANA ≥ 1:400; 102 % were AMA positive. There was no significant correlation between the ANA titers and the causative drug, while AMA positive patients had taken nonsteroidal anti-inflammatory drugs more frequently. No difference was seen regarding clinical characteristics or laboratory parameters in patients with ANA ≥ 1:400, while patients with positive AMA presented with higher aminotransferases, bilirubin and INR. Significantly higher proportions of patients with ANA ≥ 1:400 or AMA positivity exhibited elevated IgG levels were observed for patients with ANA ≥ 1:400 or AMA positivity. Both ANA titers ≥ 1:400 and AMA positivity but not elevated ANA titers correlated with a higher proportion of Hy's law positivity. Conclusion A closer look in a causality proven DILI cohort provided no evidence that presence of ANA titers is specific for DILI by a certain medication. AMA rather than ANA positivity was related to a more pronounced liver injury.