A 64-year-old Chinese man weighing 70 kg suffering for more than 1 month due to recurrent fever was transferred from an outside medical facility to the Department of Critical Care Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine for further treatment on 26 May 2022. He was diagnosed with hemophagocytic lymphohistiocytosis (HLH) based on splenomegaly, hemocytopenia, elevated ferritin, hypertriglyceridemia, and abnormal coagulation function, according to the criteria of HLH-2004 (Henter et al., 2007). Then, non-Hodgkin’s lymphoma had been diagnosed through bone marrow biopsy, causing HLH. This patient received continuous glucocorticoid treatment with chemotherapeutic agents, zanubrutinib and obinutuzumab, for the treatment of HLH and non-Hodgkin’s lymphoma (Table 1). He subsequently developed hospital-acquired pneumonia. Carbapenem-resistant Klebsiella pneumoniae (CRKP), Aspergillus fumigatus, and Enterococcus faecium were identified in sputum and blood by metagenomics next-generation sequencing (mNGS) and culture, necessitating initiation of antimicrobial treatment. Antibacterial therapy included empirical piperacillin–tazobactam, polymyxin B, fosfomycin, minocycline, ceftazidime–avibactam, and vancomycin. Voriconazole and amphotericin B were administrated for aspergillosis infection.

The plasma trough concentration (C_min) of voriconazole was 0.2 mg L⁻¹, determined by liquid chromatography–electrospray tandem mass spectrometry (detailed in the Supplementary File S1), on 28 June 2022. Twenty-one days of intravenous (IV) voriconazole (400 mg twice on the first day and maintenance dose of 200 mg q12 h) treatment was not satisfactory (Figure 1). On assessing, Aspergillus fumigatus was found again in sputum (mNGS) on 29 June 2022. Before voriconazole initiation, IV dexamethasone 20 mg had been used for 7 days and then gradually reduced. Following dexamethasone, IV methylprednisolone 80 mg was given and then gradually reduced. Subsequently, IV dexamethasone 20 mg combined with etoposide 0.5 mg and vincristine 18 mg was intravenously administrated for chemotherapy for lymphoma from 1 July to 4 July 2022. On 4 July, voriconazole C_min was 0.1 mg L⁻¹ and still below the therapeutic index. His chest CT scan indicated progressive pulmonary aspergillosis with a nodule in the right upper lobe and a thick-walled cavity inside the nodule. Aspergillus fumigatus could be detected in bronchoalveolar lavage fluid (BALF) and blood by mNGS. Double maintenance dose of voriconazole (400 mg q12 h) was given on 5 July 2022, and voriconazole C_min of 1.7 mg L⁻¹ (July 8th) achieved the therapeutic range. In addition, this patient carrying the CYP2C19*1*2 genotype [*2: 681G>A, rs4244285; *3: 636G>A; rs4986893; *17: −806C>T; rs12248560; genotyping adopted using the Sanger DNA sequencing method with an ABI3730xl-full automatic sequencing instrument (ABI Co.)] was an intermediate metabolizer of voriconazole. The interaction between dexamethasone and voriconazole was recognized after checking the other medications of this patient. In addition, the total score of the Drug Interaction Probability Scale (DIPS) was 10 (>8: highly probable) (Horn et al., 2007). Voriconazole C_min increased to 2.6 mg L⁻¹ on 11 July 2022. Voriconazole dose was adjusted to 300 mg q12 h on 12 July, and voriconazole C_min was 2.0 mg L⁻¹ on 18 July. Aspergillus fumigatus still could be detected in BALF by mNGS. A nodule was formed in the apex of the right lung, with a small cavity inside it, as shown by CT scan (18 July 2022). The amphotericin B cholesteryl sulfate complex with dexamethasone 2.5 mg was added on 21 July 2022. Voriconazole C_min was monitored, and its value was 2.0 mg L⁻¹ on 27 July. The patient received the second chemotherapy regimen for lymphoma treatment consisting of dexamethasone 10 mg, etoposide 0.5 mg, and vincristine 18 mg from 1 August to 4 August. Voriconazole C_min was 1 mg L⁻¹ on 3 August. Voriconazole dose was adjusted to 400 mg q12 h, and voriconazole C_min increased to 1.8 mg L⁻¹ on 8 August. Then, voriconazole dose was reduced to 300 mg q12 h on 12 August. The chest CT scan indicated that symptoms of pulmonary aspergillosis improved on 16 August 2022. Voriconazole C_min was 1.9 mg L⁻¹ (August 18th) and effective against aspergillus infection.

To our knowledge, this is the first report of a dose-dependent interaction between voriconazole and dexamethasone in the patient carrying the CYP2C19*1*2 genotype (IM). The co-administration of dexamethasone increased voriconazole clearance with low exposure. Voriconazole C_min was 0.1 mg L⁻¹ (dose: 5.7 mg kg⁻¹ day⁻¹, 400 mg day⁻¹; C_min/dose (C/D) ratio: 0.018), when 20 mg of dexamethasone was co-administrated on 4 July 2022. Progressive pulmonary aspergillosis, indicated by CT scan and Aspergillus fumigatus in BALF, suggested unsatisfactory voriconazole therapy. Voriconazole C_min was 2 mg L⁻¹ (dose: 8.6 mg kg⁻¹ day⁻¹, 600 mg day⁻¹; C/D ratio: 0.23) when 2.5 mg of dexamethasone was used on 27 July. Then, 12.5 mg of dexamethasone was given from 1 August to 4 August. Voriconazole C_min was 1 mg L⁻¹ (dose: 5.7 mg kg⁻¹day⁻¹, 400 mg day⁻¹; C/D ratio: 0.18) on 3 August. Subsequent C_min values were 1.8 mg L⁻¹ (dose: 11.4 mg kg⁻¹day⁻¹, 800 mg day⁻¹; C/D ratio: 0.16) and 1.9 mg L⁻¹ (dose: 8.6 mg kg⁻¹day⁻¹, 600 mg day⁻¹; C/D ratio: 0.22) on 8 August and 18 August, respectively (dexamethasone: 2.5 mg). This patient received effective voriconazole therapy, combined with the amphotericin B cholesteryl sulfate complex, and showed improvement in pulmonary aspergillosis.

Reduced voriconazole C_min was due to increase in clearance, attributing to the co-administration of glucocorticoids (dexamethasone and methylprednisolone). Dexamethasone had a greater effect on voriconazole exposure compared to methylprednisolone (Dolton et al., 2012; Jia et al., 2021). Voriconazole C_min is a good measure of drug exposure, recommended as the pharmacokinetic (PK)/pharmacodynamic parameter for regular therapeutic drug monitoring (TDM) in the clinic (Takesue et al., 2022), because it has a linear relationship with the voriconazole area under the concentration–time curve (AUC_0–12h) (Hope, 2012). However, the results of this interaction in different research studies are inconsistent (Gautier-Veyret et al., 2015; Blanco-Dorado et al., 2020). It may be related with difference in the distribution of CYP2C19 genotypes and glucocorticoid dose employed among the studied population.

Voriconazole is primarily metabolized in the liver by CYP2C19 and CYP3A4 enzymes. Difference in the CYP2C19 gene between individuals can greatly affect voriconazole metabolism. The *2 and *3 alleles were loss-of-function variations. IM with one such variant had significantly lower enzyme activity than normal metabolizer (NM, *1*1) (Li et al., 2023). In the absence of functional CYP2C19, CYP3A4 became the important alternative pathway for voriconazole clearance (Murayama et al., 2007). In CYP2C19 poor metabolizer (PM) liver microsomes, inhibition of voriconazole metabolism by ketoconazole (a specific CYP3A4 inhibitor) was most potent (Hyland et al., 2003). A PK study of healthy participants (Mikus et al., 2006) also showed that co-administration of ritonavir (a potent CYP3A4 inhibitor) led to lower voriconazole clearances in CYP2C19 IM (*1*2) and PM (*2*2), compared with NM. It is inferred that high dose of dexamethasone might affect voriconazole clearance much more in IM and PM of CYP2C19.

The extent of voriconazole–dexamethasone interaction seems to depend on the dexamethasone dose. The induction of CYP450 enzymes, particularly CYP3A4, by dexamethasone is dose-dependent (Bourdin et al., 2023). Co-administration of dexamethasone with high and moderate doses (20 and 12.5 mg) seemed to impact voriconazole C_min. Consistently, CYP3A4 induction was, respectively, >50%–<80%, >20–<50%, and <20% at the dexamethasone dose of >16 mg (high dose), >1.5–16 mg (moderate dose), and ≤1.5 mg (low dose) (Jacobs et al., 2022). Taghvaye et al. (2019) reported the significant interaction between iv voriconazole and iv dexamethasone (8 mg q12 h), leading to failure of antifungal treatment, in a 32-year-old woman with acute lymphoblastic leukemia. Wallace et al. (2016) also found the clinical relevant drug interaction between iv voriconazole and oral dexamethasone (4 mg q8 h) in an 84-year-old male patient. In our case, subtherapeutic C_min of voriconazole (0.2 mg L⁻¹) on 28 June and subsequent detection of Aspergillus fumigatus in sputum on 29 June indicated ineffective voriconazole therapy, associated with the treatment of both dexamethasone and methylprednisolone from 26 May to 27 June. The used methylprednisolone could enhance the CYP induction of high-dose dexamethasone (20 mg, from 1 July to 4 July), related with the 0.1 mg L⁻¹ of voriconazole C_min on 4 July.

A multicenter study about the PK and TDM of voriconazole (Dolton et al., 2012) showed that co-administration of dexamethasone and methylprednisolone reduced voriconazole concentrations to a greater extent than prednisone or prednisolone, associated with their potency of binding to glucocorticoid receptors (GRs). Glucocorticoids induce CYP450s, including CYP2C19 and CYP3A4, through binding to GR. In addition to that, dexamethasone strongly induces CYP3A4 via direct activation of pregnane X receptor (PXR). A recent retrospective observational study of 231 patients with 918 voriconazole C_min (Jia et al., 2021) confirmed that the co-administration of glucocorticoids (dexamethasone, prednisone, prednisolone, and methylprednisolone) reduced the voriconazole C_min/dose, among which dexamethasone caused the lowest median of the voriconazole C_min/dose ratio. However, they did not analyze the effect of glucocorticoid dose on voriconazole. The study of immunocompromised patients (n = 38) (Imataki et al., 2018) found that voriconazole concentrations were significantly decreased in corticosteroid users compared to the non-user (p = .013) (the median doses of corticosteroids: 89.8 mg/d, equivalent to dexamethasone 13.47 mg). Blanco-Dorado et al. (2020) (the median doses of corticosteroids: 50 mg/d, equivalent to dexamethasone 7.5 mg) and Gautier-Veyret et al. (2015) (no data about the dose of glucocorticoids) showed no interaction between glucocorticoids and voriconazole. The inconsistent results could be explained by the heterogeneity of the type and dose of the glucocorticoids, in line with that reported in our case.

In addition, voriconazole has a long-lasting and potent inhibitory effect on CYP3A activity (Huang et al., 2021). Dexamethasone and methylprednisolone are primarily metabolized by CYP3A4 (Blanco-Dorado et al., 2020; Bourdin et al., 2023). The maximum plasma concentration (C_max) and the AUC from 0 h to infinity (AUC_0→inf) of dexamethasone could be increased by 2.44-fold and 2.60-fold, respectively, when combined with voriconazole. For methylprednisolone, C_max and AUC_0→inf were increased by 1.56-fold and 2.23-fold, respectively (Li et al., 2018).Thus, their increased exposure in combination with voriconazole may enhance the interaction. Multiple daily high doses (>16 mg) of dexamethasone increased the risk of clinical significant interactions. High-dose dexamethasone is widely administrated in the treatment of hematological malignancies (Jacobs et al., 2022). Furthermore, moderate-dose dexamethasone (>1.5–16 mg) with long treatment course (>10 days) had a considerable risk of interactions and commonly used in the treatment of palliative care and cerebral edema.