Evaluation of diagnostic and prognostic candidate biomarkers in drug‐induced liver injury vs. other forms of acute liver damage

Aims Detection and characterization of idiosyncratic drug‐induced liver injury (DILI) currently rely on standard liver tests, which are suboptimal in terms of specificity, sensitivity and prognosis. Therefore, DILI diagnosis can be delayed, with important consequences for the patient. In this study, we aimed to evaluate the potential of osteopontin, cytokeratin‐18 (caspase‐cleaved: ccK18 and total: K18), α‐glutathione‐S‐transferase and microRNA‐122 as new DILI biomarkers. Methods Serial blood samples were collected from 32 DILI and 34 non‐DILI acute liver injury (ALI) cases and a single sample from 43 population controls without liver injury (HLC) and analysed using enzyme‐linked immunosorbent assay (ELISA) or single‐molecule arrays. Results All biomarkers differentiated DILI and ALI from HLC with an area under receiver operator characteristic curve (AUC) value of >0.75 but were less efficient in distinguishing DILI from ALI, with ccK18 (0.79) and K18 (0.76) demonstrating highest potential. However, the AUC improved considerably (0.98) for ccK18 when comparing DILI and a subgroup of autoimmune hepatitis cases. Cytokeratin‐18, microRNA‐122 and α‐glutathione‐S‐transferase correlated well with traditional transaminases, while osteopontin correlated most strongly with the international normalized ratio (INR). Conclusions ccK18 appears promising in distinguishing DILI from autoimmune hepatitis but less so from other forms of acute liver injury. Osteopontin demonstrates prognostic potential with higher levels detected in more severe cases regardless of aetiology.