Integrating clinical cohort and integrative transcriptomic and computational analysis reveal shared and unique characteristics and targets of triazole antifungal drugs-associated liver injury in patients with pulmonary mycosis

Triazole antifungal drugs (TADs) are widely used for pulmonary mycosis but carry hepatotoxicity risks with incompletely understood mechanisms. This study aimed to characterize TADs-associated liver injury through integrated clinical and comprehensive bioinformatics analysis. A retrospective cohort of 199 pulmonary mycosis patients receiving TAD monotherapy was analyzed. Liver injury was defined by CIOMS/DILIN/ DEWG criteria; adverse events by Common Terminology Criteria for Adverse Events (CTCAE) v6.0. Independent risk factors were identified using Kaplan–Meier curves and multivariable Cox regression, and prediction models were developed and internally validated. Concurrently, TADs targets were predicted from multiple databases and intersected with hepatotoxicity genes. Transcriptomic validation was performed using bulk and single-cell RNA-seq data from cholestatic liver injury patients. Molecular docking, cell–communication analysis, and virtual knockout experiments were conducted to elucidate underlying mechanisms. Incidences of liver adverse events and injury were 31.7% and 11.6%, respectively; voriconazole showed the highest rates. Cholestatic injury predominated. High baseline ALT, AST, GGT, ALP, and immunocompromised status were independent risk factors. The prediction model achieved an optimism-corrected C-index of 0.73. Integrative transcriptomic and computational analysis identified KDR as the only shared target with strong binding to all three TADs. KDR was upregulated in cholestatic liver injury, with enhanced VEGFA–KDR signaling between Kupffer cells and endothelial cells. Virtual knockout implicated apoptosis, hypoxia, inflammation, and VEGF pathways. This study provides a comprehensive characterization of TADs-associated liver injury, identifying key risk factors and KDR as a candidate shared hepatotoxicity target, offering insights for risk stratification and potential therapeutic intervention.