The efficacy and clinical application of glucocorticoids (GCs) in patients with drug-induced liver injury (DILI) remain controversial.

To determine the efficacy and suitable population of GCs in patients with DILI.

This was a single-center, retrospective study. Patients with moderate-to-severe DILI who met the diagnostic criteria from January 1, 2009 to December 31, 2024 were enrolled. Patients in the GC group and the non-GC group were matched 1:1 by propensity score-matched (PSM), and the reduction of disease severity and biochemical parameters were compared between the two groups. According to the level of TB, the patients with DILI were divided into three groups, and the efficacy of GCs in each subgroup was compared.

Patients with DILI in our study were evaluated by the updated RUCAM causality assessment scale. All patients had RUCAM scores≥6 and causal relationship graded as " probable" or " highly probable". The use of GC treatment differs according to the severity of patients, especially the baseline level of TB (% in patients with TB <5 ULN, % in patients with 5ULN≤TB≤10ULN, and % in TB≥10ULN). After PSM analysis, 163 patients were included in each group. After PSM, the time of severity reduction was faster in the GC group than the non-GC group (P=0.022). The adjusted cumulative rate of severity reduction was 62% in the GC group and 43.6% in the non-GC group. The cumulative rates of reaching 50% reduction in AST, ALP, TB level was higher in the GC group than in the non-GC group (P<0.001, P=0.0086 and P=0.003). Patients were divided into three subgroups according to baseline TB level. We found the cumulative rates of who achieved severity reduction and 50% reduction in liver biochemical parameters at discharge was higher in the GC group than in the non-GC group in patients with 5ULN≤ TB level ≤10ULN but not in patients with TB <5ULN and TB>10ULN. Multivariate analysis showed that sex, age, treatment group and severity were significantly associated with disease severity reduction of patients with DILI. After PSM, there were 56 patients with drug-induced autoimmune hepatitis (DI-AIH) and 270 patients with DILI with no autoimmune features cases in our cohort. The cumulative rate of severity reduction at discharge in the GC group was higher than that in the non-GC group in patients with DI-AIH (P=0.049). Although the incidence of side effects in the GC group was higher than that in the non-GC group, the adverse reactions were basically relieved with the withdrawal of GCs.

GCs accelerated the reduction of disease severity and liver biochemical parameters, especially in patients with 5ULN≤baseline TB level≤10ULN. Sex, age, treatment group and severity were significantly associated with disease severity reduction of patients with DILI.