Voriconazole is a triazole antifungal agent with high efficacy against most Aspergillus species (Scott and Simpson, 2007; Pascual et al., 2008). It is commonly used to prevent and treat invasive fungal infections and is recommended as first-line therapy for invasive aspergillosis (IA) (Patterson et al., 2016; Moriyama et al., 2017). However, its use is limited by adverse drug reactions, including visual disturbances, encephalopathy, rash, and liver injury (Levine and Chandrasekar, 2016; Benitez and Carver, 2019).

Voriconazole pharmacokinetics (PK) are complex, with high inter- and intra-individual variability (Purkins et al., 2002; Saleh et al., 2024). However, trough voriconazole plasma concentration (VPC) is correlated with efficacy and adverse events (Hu et al., 2025; Jin et al., 2016). Voriconazole is metabolized by hepatic cytochrome P450 (CYP) isoenzymes, primarily by CYP2C19 and partially by CYP2C9 and CYP3A4 (Hyland et al., 2003; Zonios et al., 2014). Trough VPC may be influenced by various confounding factors, including drug–drug interactions and CYP2C19 gene polymorphisms (Wang et al., 2014; Hu et al., 2023). Therefore, therapeutic drug monitoring (TDM) of trough VPC is recommended (Park et al., 2012; Takesue et al., 2022).

Voriconazole disposition is influenced by liver function, with reduced metabolic enzyme activity observed in patients with liver dysfunction (Tang et al., 2021). General hepatotoxicity from voriconazole use occurs in 69% of patients with liver dysfunction and in only 15.82% of the general population (Solís-Muñoz et al., 2013; Xiao et al., 2023; Hu et al., 2024). The incidence of ALI, which is a severe form of hepatotoxicity, in patients with mild-to-moderate liver dysfunction treated with voriconazole is unknown. These patients are often treated with other potentially hepatotoxic drugs and may have underlying conditions that predispose them to ALI. Thus, understanding the risk factors for voriconazole-induced ALI is essential for improving drug safety.

The aim of this nested case–control study was to: (1) determine the incidence of ALI in patients with mild-to-moderate liver dysfunction receiving voriconazole and (2) identify the risk factors of voriconazole-induced ALI in this population.

Voriconazole is the standard therapy for IA. In addition to that of immunosuppression, the incidence of aspergillosis appears to be increasing in patients outside commonly recognized risk groups, including those with chronic obstructive pulmonary disease (COPD) (Solís-Muñoz et al., 2013). In the present study, hypoproteinemia and renal insufficiency accounted for a large proportion of cases. The primary reason for initiating empiric therapy was suspected IA of the lungs.

Voriconazole is effective against most Aspergillus species, and it reaches high concentrations in the lung tissue. However, frequent adverse effects, such as hepatotoxicity, limit its use. Although the incidence of voriconazole-associated hepatotoxicity is moderate in the general public (Levine and Chandrasekar, 2016), it is more common in patients with liver dysfunction (Hu et al., 2024). In the present study, we evaluated voriconazole-induced ALI, specifically in patients with mild or moderate hepatic dysfunction. Among patients with liver dysfunction (Child–Pugh A or B), 44 (30.6%) developed ALI during voriconazole therapy. Therefore, we sought to identify clinical risk factors to reduce ALI incidence in this population.

Previous research has focused on the safety of voriconazole in patients with severe liver dysfunction (Solís-Muñoz et al., 2013). However, studies on the hepatotoxicity of voriconazole in patients with mild-to-moderate liver dysfunction are lacking. Our study demonstrated that patients with liver dysfunction are at a high risk of developing voriconazole-induced ALI. High trough VPC was identified as an independent risk factor for voriconazole-associated hepatotoxicity. The optimal steady trough concentration for voriconazole, per Infectious Disease Society of America guidelines, is 1.0–5.5 mg/L (Patterson et al., 2016). Similarly, a population PK study in patients with liver dysfunction reported an association between hepatotoxicity and trough VPC levels of <5.1 mg/L. Meanwhile, in the present study, the optimal trough VPC cutoff to prevent ALI was 3.81 mg/L, suggesting that trough VPC levels of >3.81 mg/L in patients with liver dysfunction (Child–Pugh A or B) receiving voriconazole pose a higher risk of ALI. The potential underlying mechanisms are as follows. (1) Voriconazole is primarily metabolized by CYP450. Hepatic impairment decreases CYP450 enzyme activity, leading to accumulation of the parent drug in the liver, which may cause liver injury (Hu et al., 2024). (2) Decreased bile excretion may lead to the retention of drugs or their toxic metabolites (such as voriconazole N-oxide) in the liver (Yuan and Kaplowitz, 2009). (3) Patients with hepatic impairment have insufficient glutathione reserves, impairing their ability to effectively detoxify reactive metabolites (European Association for the Study of the Liver, 2024). The correlation between trough VPC levels and hepatotoxicity highlights the significance of TDM for voriconazole, particularly in patients with liver dysfunction. Adjusting the voriconazole dosage based on trough VPC levels could reduce the risk of ALI without necessitating drug discontinuation for drug-induced hepatotoxicity. The duration of voriconazole treatment is another independent risk factor for hepatotoxicity. Our study demonstrated a significantly increased risk of ALI associated with voriconazole treatment duration exceeding 10 days in patients with liver impairment. Clinicians should monitor liver function in patients receiving voriconazole therapy for more than 10 days. For patients anticipated to require voriconazole therapy for more than 10 days, proactive assessment to minimize other hepatotoxic exposures (e.g., sodium valproate, paracetamol, and macrolide antibiotics) is necessary. Our findings support targeting a trough concentration of <3.81 mg/L for patients with hepatic dysfunction to reduce the incidence of ALI. A trough level of ≥1.0 mg/L is strongly recommended to improve efficacy. Trough levels of ≥2.0 mg/L are suggested for invasive aspergillosis (Japanese Antimicrobial Therapeutic Drug Monitoring Guideline Committee, 2022). If voriconazole trough levels approach 3.81 mg/L, clinicians should monitor liver function more closely. If voriconazole trough levels are within 3.81–10 mg/L, the maintenance dosage of voriconazole is recommended to be decreased by 20%. If the trough VPC is above 10 mg/L, voriconazole administration is suggested to be skipped once, with the maintenance dosage decreased by 50%, followed by dosage adjustment based on the blood concentration (Chen et al., 2018). Considering that tigecycline can independently cause liver toxicity (Jiang et al., 2022), we hypothesized that its concomitant use with voriconazole may contribute to ALI. Accordingly, we included tigecycline use as a variable to assess its relationship with voriconazole-induced ALI. The rate of tigecycline use was higher in the ALI group than in the control group, although the difference was not significant between the groups (p > 0.05).

Certain underlying disease may influence the adverse effects of voriconazole. In a previous study, the risk of liver injury increased in patients with hypoproteinemia (Hirata et al., 2019). In the present study, we systematically collected and analyzed underlying medical conditions in patients, including renal insufficiency, immunosuppressive state, leukemia, hypoproteinemia, and COPD. As patients can suffer from multiple underlying diseases simultaneously. We did not include underlying medical condition as a separate variable in the statistical analysis. Hypoalbuminemia can increase the concentration of plasma-free voriconazole, distributing more free molecules to the liver and other tissues and organs, increasing the risk of adverse reactions.

This study had some limitations. First, only four variables were included in the multivariable analysis, all of which appeared significant (p < 0.1). A larger sample size could enhance the statistical power to identify additional risk factors in patients with liver dysfunction. Second, the diagnostic criteria for drug-induced ALI were inadequate. Although a temporal relationship exists between elevated liver enzyme levels and voriconazole use, biochemical abnormalities during drug administration may not accurately confirm drug-induced ALI. Third, serum voriconazole concentrations can vary in patients with different CYP2C19 genotypes, and this may influence drug safety (Wang et al., 2016). In addition, CYP2C19 polymorphisms vary considerably between races. CYP2C19-poor metabolizer prevalence in China is considerably higher than that in Europe (specifically, Caucasians) and Africa. The applicability of our proposed trough VPC cutoff and hepatotoxicity risk predictions may be influenced by ethnicity-dependent genetic polymorphisms. In our study, the majority of patients did not undergo genetic polymorphism testing. Therefore, we did not incorporate this risk factor for voriconazole-induced ALI. Pharmacogenetic testing should be considered as a future direction to refine risk stratification and dosing. Finally, our study reflects the clinical context of a tertiary center in China. Regional variations in liver disease etiology may limit generalizability; for instance, alcoholic liver disease and metabolic dysfunction-associated steatotic liver disease prevail in Western populations. Future validation in multi-regional patients would strengthen voriconazole applicability.

This study has significant clinical implications for the management of patients with liver dysfunction undergoing voriconazole treatment. Trough VPC was identified as an independent risk factor for ALI, with a toxic threshold of 3.81 mg/L for patients with mild or moderate liver dysfunction. Therefore, TDM of voriconazole is crucial for this patient population to prevent discontinuation owing to ALI. A lower target trough VPC (<3.81 mg/L) is suggested for these patients. If the clinical response is inadequate at lower trough VPC levels, clinicians may consider alternative antifungals, such as caspofungin alone or in combination with nebulized amphotericin B.

Overall, this study revealed that voriconazole-induced ALI occurs in 30.6% of patients with mild-to-moderate liver dysfunction (Child–Pugh A and B). Independent risk factors for ALI include a trough VPC of >3.81 mg/L and a treatment duration exceeding 10 days. Further large-scale prospective studies are required to validate these risk factors and enhance voriconazole management in this population.