Abstract: Voriconazole is a triazole antifungal agent with broad-spectrum activity against several common yeast and mold species. However, hepatotoxicity is a major adverse effect of voriconazole, and the lack of specific biomarkers for the detection and prediction of voriconazole-induced hepatotoxicity remains an urgent issue. In this study, ultra-high-performance liquid chromatography was coupled with mass spectrometry to analyze plasma and liver metabolites and determine possible plasma biomarkers for hepatotoxicity. Firstly, male C57BL/6J mice were randomly divided into four groups (each n = 7): control (0 mg/kg voriconazole), and 20, 40, and 80 mg/kg voriconazole. Voriconazole was intraperitoneally injected, and the resulting differential plasma and liver metabolites were identified to determine which differential metabolites were released from the liver into the periphery. Additionally, 133 plasma samples obtained from patients with (n = 45) and without (n = 88) hepatotoxicity were collected to further validate the differential metabolites found in the animal experiment. Alpha-ketoglutarate (AKG), 5-hydroxyindole, 7-ketolithocholic acid, 3-methylglutarylcarnitine, uracil, phosphatidylcholine (20:3/20:4), and lysophosphatidylcholine (22:6) were associated with voriconazole hepatotoxicity in a dose-dependent manner. Uracil (area under the curve (AUC): 0.979, 95% confidence interval (CI): 0.958-0.999), and AKG (AUC: 0.877, 95% CI: 0.812-0.941) were biomarkers of voriconazole-induced hepatotoxicity and showed great potential for clinical diagnosis. The combined biomarker composed of uracil, AKG, and alkaline phosphatase reached an AUC of 0.997 (95% CI: 0.993-0.999), showing that the combination of the three is a good choice to judge voriconazole-induced hepatotoxicity.

Aims: The pharmacokinetic (PK) profiles of voriconazole in intensive care unit (ICU) patients is quite different. We aimed to develop a population pharmacokinetic (PopPK) model to evaluate the effects of various biological covariates and the use of extracorporeal membrane oxygenation (ECMO) and continuous renal replacement therapy (CRRT). Methods: The modeling analysis of the pharmacokinetic parameters were conducted using the nonlinear mixed-effects modeling method (NONMEM) using a two-compartment model. Monte Carlo simulations (MCSs) were performed to observe the probability of target attainment (PTA) when receiving CRRT or not under different dosage regimens, different quick C-reactive protein (qCRP), and different minimum inhibitory concentration (MIC) ranges. Results: A total of 408 critically ill patients with 746 voriconazole concentration–time data points were included in this study. A two-compartment population PK model with qCRP, CRRT, creatinine clearance rate (CLCR), platelet (PLT), and prothrombin time (PT) as fixed effects was developed using the NONMEM. Conclusion: The results showed that qCRP, CRRT, CLCR, PLT, and PT affected the PK parameter clearance. The most commonly used clinical regimen of 200 mg q12h is sufficient for the most common sensitive pathogens (MIC ≤ 0.25 mg/L) in China, regardless of whether CRRT is performed, and at what level qCRP is. When the MIC is 0.5 mg/L, 200 mg q12h is insufficient only when qCRP is less than 40 mg/L and CRRT is performed. When MIC ≥ 2 mg /L, a dose of 300 mg q12h cannot achieve ≥ 90% PTA, and a higher dose needs to be explored.