According to studies using cell lines involved in tumor transformation, the World Health Organization (WHO) classifies hematological malignant tumors into “myeloid tumours”, “lymphoid tumours”, “mast cell disorders”, and “histiocyte tumours” (Arber et al., 2016; Mertowska et al., 2023; Swerdlow et al., 2016). In patients suffering from a hematological malignancy, due to the fact that their immune system may be suppressed by the disease or treatment, these patients are more susceptible to invasive fungal infections (IFIs). In addition, patients with hematological malignant tumors often require the use of central venous catheters to provide medication, fluids, or nutrition, which also increases the risk of infection. Therefore, these patients are a high-risk group for IFIs and require special prevention and treatment strategies (Jenks et al., 2020; Pagano et al., 2011).

Voriconazole is a broad-spectrum antifungal agent belonging to the triazole class. It inhibits ergosterol biosynthesis in the membranes of fungal cells. Voriconazole is active against a wide range of fungi, including Aspergillus spp, Candida spp, and several others (Fernández Ávila et al., 2021; Purkins et al., 2003; Xie et al., 2023). Voriconazole is used as first-line treatment for invasive aspergillosis. It can also be used for prophylaxis in high-risk patients with an IFI. Voriconazole is active against a wide range of fungi, but it produces adverse effects, including hepatotoxicity, visual disturbances, and rash (Eiden et al., 2007; Yasu et al., 2022; Kim et al., 2017; Mihăilă, 2015; Wu et al., 2021; Xie et al., 2023). Studies have shown that adjustments to dosing regimens for voriconazole based on therapeutic drug monitoring are beneficial for promoting its safety and efficacy (Luong et al., 2016; Valle-T-Figueras et al., 2021).

There are significant differences in physiology and pharmacokinetics between children and adults,but the dosage for children is usually extrapolated from the adult dose (Leroux et al., 2021; Sienkiewicz-Oleszkiewicz et al., 2023; Wei et al., 2019). Children and adults may have different sensitivity and tolerance to drugs, leading to differences in the types and incidence of adverse drug events (ADEs) (Kearns et al., 2003; Pasternak et al., 2019). Therefore, grouping patients in these two age groups can more accurately evaluate the safety and efficacy of voriconazole in patients with hematological malignancies, providing more specific guidance for clinical medication.

The US Food and Drug Administration Adverse Event Reporting System (FAERS) is one of the major databases for the post-marketing surveillance of drugs (Cirmi et al., 2020; Feng et al., 2022). It is updated quarterly and publicly available for free download (Sakaeda et al., 2013). The database is used widely in pharmacovigilance studies to compensate for the limitations of the pre-marketing studies of drugs. FAERS plays an important part in updating drug inserts and releasing information on drug-safety alerts (Raisch et al., 2014).

Herein, we undertook data mining of ADEs with voriconazole in patients with a hematological malignancy using FAERS. We compared the differences between children and adults to provide information on the safety of voriconazole.

Voriconazole underwent rigorous pre-marketing clinical trials to ensure its efficacy and safety in treating IFIs. However, the diversity and complexity of the patient population in actual clinical use suggest that there may be ADEs that have not been identified or evaluated fully. We conducted in-depth signal mining and evaluation to explore differences in ADEs between children and adults treated with voriconazole in patients with a hematological malignancy using FAERS. We aimed to provide important information for future clinical use.

Figure 2 showed an upward trend in the number of ADEs related to voriconazole from 2004 to 2020. This trend may have been related to the increased frequency of voriconazole use and increased awareness of ADEs. In particular, the number of reports peaked in 2020, which may have been associated with the coronavirus disease-2019 pandemic (Salmanton-García et al., 2024; Hlaing et al., 2023; Papakonstantinou et al., 2021). Since 2021, there has been a downward trend, which may be related to the development of individualized use of voriconazole in clinical practice.

At the SOC level, commonalities and differences were presented between child and adult groups. The only positive signal that satisfied all four algorithms simultaneously in children was “psychiatric disorders” (Table 2) whereas, in adults, the positive signals were “endocrine disorders” and “hepatobiliary disorders” (Table 3). The percentages of SOC that were significantly higher in children than in adults were “skin and subcutaneous tissue disorders”, “gastrointestinal disorders”, “psychiatric disorders”, and “musculoskeletal and connective tissue disorders” (Figure 3). The percentages of SOC that were significantly higher in adults than in children were “investigations”, “respiratory, thoracic and mediastinal disorders”, “cardiac disorders”, “eye disorders”, and “neoplasms benign, malignant and unspecified (including cysts and polyps)”. In addition, two SOCs were unique to the children group (“ear and labyrinth disorders” and “product issues”), four SOCs were unique to the adult group (“reproductive system and breast disorders”, “congenital, familial and genetic disorders”, “surgical and medical procedures”, and “pregnancy, puerperium and perinatal conditions”). This phenomenon may have been due to the unique physiological state and drug-metabolism characteristics of children (children show greater systemic metabolism of voriconazole than adults) (Leroux et al., 2021).

At the PT level, “photosensitivity reactions” is a more common dermatological complication of voriconazole than in other azole antifungal agents (Malani and Aronoff, 2008). We found that the ROR signal intensity of this PT was higher in children and significant compared with that in adults (Figure 4). This finding provides additional evidence of the need for caution when prescribing voriconazole in children.

In the children group, “hallucinations, mixed” was the PT signal with the highest ROR signal intensity (Table 4). In an observational study of 72 patients aged 14–76 years treated with voriconazole, hallucinations occurred in 12 cases (16.67%). Half of these patients did not report their hallucinations spontaneously. They showed reluctance to describe them, possibly due to embarrassment and other contributing factors (Zonios et al., 2008). A case report and literature review of voriconazole-induced hallucinations and visual disturbances reported 42 cases, three of whom were children (Zheng et al., 2021). In a recent study, a search of multiple databases on drug-induced musical hallucinations identified 27 cases and 21 triggering drugs. Among them, three patients (11.11%) had musical hallucinations induced by voriconazole (Bakewell et al., 2024; Zonios et al., 2008). Voriconazole treatment-related hallucinations may be overlooked by physicians. There are few reports of hallucinations associated with voriconazole in children. However, given the high ROR in the current study, we suggest that whether children experience hallucinations deserves more clinical attention.

In the adult group, “toxic optic neuropathy” had an unusually high ROR signal intensity and high statistical significance (Figure 4). One study revealed six cases of toxic optic neuropathy induced by voriconazole in pharmacovigilance databases (e.g., VigiAcess), and voriconazole was the only drug suspected in two cases (Orssaud et al., 2021). Mounier et al. reported a case of ophthalmic complications possibly caused by toxic optic neuropathy (Mounier et al., 2018). Understanding the mechanism leading to this neuro-ophthalmic adverse effect is crucial for clinical practice. One meta-analysis indicated a trough concentration >3.0 mg/L to be associated with an increased risk of moderate-to-severe hepatotoxicity, and >4.0 mg/L to be associated with an increased risk of neurotoxicity (Jin et al., 2016). Those data suggest a need for close monitoring of the therapeutic concentrations of voriconazole during treatment. Notably, there were “drug interactions”, “drug level below therapeutic”, and “drug level increased” in children and adults in our study. Studies have shown interactions between voriconazole and carbamazepine, cyclophosphamide, aprepitant, tacrolimus, and letermovir, which are related to the induction or inhibition of metabolic enzymes such as CYP2C19, 3A4, 3a5-2D6. Multiple guidelines recommend monitoring the drug concentration during voriconazole treatment to improve safety and efficacy, preferably with prospective dose optimization based on genotype.

We identified new PTs that were not previously listed in the drug label, such as “generalised tonic–clonic seizures” (Table 5) and “disorientation” (Table 4). This finding: (i) suggests that certain patient groups may be at risk; (ii) demonstrates the importance of ongoing post-marketing surveillance and signal mining for ADEs.

Analyses of time-to-onset of the ADE showed that most ADEs occurred within 0 days and 30 days of dosing (Figure 5). The median time-to-onset of the ADE was 11 days in the children group and 8.5 days in the adult group. These data suggested that close monitoring should be carried out during the initial stage of voriconazole treatment. However, one case occurred in the adult group after 1 year of treatment with voriconazole.

Our study had two main limitations. First, FAERS faces inherent challenges, including incomplete, inaccurate, inconsistent, and delayed reporting of ADEs. These factors may have affected the relevance and accuracy of our results. Second, our analysis was affected by the uneven distribution of cases in FAERS, with more adult patients but a significantly smaller number of children. This uneven distribution of cases may have introduced a bias and limited the applicability of our findings. Further prospective clinical studies are needed to overcome these limitations and provide more reliable insights.

We used four algorithms (ROR, PRR, BCPNN, MGPS) to mine the signals of voriconazole in patients with a hematological malignant tumor. We found similarities and differences in SOC/PT signals between children and adults, but also identified some new PT signals not included in the drug label. In the future clinical use, differentiated pharmaceutical monitoring should be carried out for children and adults, and personalized dosing measures, such as therapeutic drug monitoring, should be combined to optimize the dosage, in order to improve the safety of voriconazole in patients with hematological malignancies.