Antibiotic exposure and indication-specific corticosteroid use differentially modulate outcomes of immune checkpoint inhibitor therapy in hepatobiliary malignancies.

Background: Concomitant medications may influence the tumor-immune microenvironment and potentially affect the efficacy of immune checkpoint inhibitors (ICIs), yet their impact in hepatobiliary malignancies remains poorly defined. We evaluated the associations of antibiotic (ATB) exposure and indication-specific corticosteroid (CS) use with clinical outcomes in hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA). Methods: In this retrospective cohort of 759 ICI-treated patients (511 HCC, 248 CCA), patients were stratified into four groups: no exposure (None, n=251), ATB only (n=135), CS only (n=183), and combined exposure (Both, n=190). ATB exposure was defined as systemic use within ±30 days of ICI initiation, and CS exposure as ≥10 mg prednisone-equivalent daily for ≥3 consecutive days. Inverse probability of treatment weighting (IPTW) and multivariable Cox models were used to adjust for confounding. Additional baseline-only sensitivity analyses restricted exposure to the 30 days before ICI initiation to reduce potential time-related bias. Results: After IPTW adjustment, baseline covariates were well balanced across exposure groups. Combined ATB and CS exposure was associated with significantly worse overall survival (OS) (adjusted HR 3.09, 95% CI 2.31-4.13; p<0.001) and progression-free survival (PFS) compared with no exposure. Objective response rates were comparable across groups (p=0.20), suggesting a greater association with impaired response durability rather than initial tumor shrinkage. Among CS-treated patients, corticosteroid use for immune-related adverse event (irAE) management was associated with improved OS compared with non-irAE indications (p<0.01). Landmark and baseline-only sensitivity analyses demonstrated generally consistent findings. Conclusions: Antibiotic exposure and corticosteroid use for non-irAE indications were associated with inferior survival outcomes in ICI-treated hepatobiliary malignancies. These findings suggest that concomitant medication exposure may influence the durability of immunotherapy responses in real-world clinical settings. In contrast, corticosteroid use for irAEs was not associated with compromised outcomes, supporting the importance of context-specific corticosteroid administration during immunotherapy. Clinically, these findings highlight the importance of judicious antibiotic use and careful consideration of corticosteroid indications during ICI treatment.