Differentiating Acute-onset Autoimmune Hepatitis From Drug-Induced Autoimmune-like Hepatitis: A Multicenter Study and Score Development.

The incidence of acute-onset autoimmune hepatitis (A-AIH) is increasing, yet diagnosis remains challenging, especially in patients with recent hepatotoxic drug exposure. The clinical presentations of A-AIH and drug-induced autoimmune-like hepatitis (DI-ALH) at onset are often indistinguishable, complicating timely diagnosis. We conducted a three-center retrospective study in China, screening patients with acute liver injury, hepatotoxic drug exposure, and autoimmune features. Patients were ultimately classified as DI-ALH if they achieved sustained remission after culprit drug cessation, or as A-AIH if they relapsed. We compared baseline demographics, laboratory indices, immunological profiles, liver histology, and established AIH diagnostic criteria. We developed and independently validated a multivariable logistic regression model to improve discrimination between A-AIH and DI-ALH. Of 458 patients screened, 238 met inclusion criteria and constituted the final cohort (94 A-AIH and 144 DI-ALH). Compared to DI-ALH, A-AIH showed significantly higher IgG levels, lower platelet counts, and higher autoantibody titers. Histologically, A-AIH exhibited more severe portal inflammation, interface hepatitis, fibrosis, and rosette formation. The discriminatory capacity of the 2022 and 2008 histological criteria was limited and comparable (AUC 0.62 vs. 0.59, p = 0.616). The Dx-AID score, incorporating platelet count, IgG, autoantibodies, and histological features, achieved high diagnostic accuracy in both the derivation cohort (AUC 0.84, 95% CI: 0.77-0.91) and the validation cohort (AUC 0.83, 95% CI: 0.72-0.94). Although A-AIH and DI-ALH share similarities, they differ in immunological and histological profiles. The Dx-AID score integrates these differences into a practical tool for early differentiation and guides personalized treatment decisions.