High metabolic N-oxidation of voriconazole in a patient with refractory aspergillosis and CYP2C19*17/*17 genotype

Case report

Voriconazole (VRC) is a highly potent triazole antifungal drug recommended as initial treatment of invasive or pulmonary aspergillosis in immunocompromised patients. VRC is mainly metabolized by CYP2C19 and CYP3A4 and exhibits a non-linear pharmacokinetic with a wide inter- and intra-individual variability, justifying the dose adjustment based on therapeutic drug monitoring (TDM) 1. Particularly, polymorphism of CYP2C19 was reported to be a major factor of the VRC pharmacokinetic variability 2. We present here a case of VRC therapy failure in a homozygous CYP2C19*17/*17 patient.

A 26-year-old Caucasian woman with an acute mixed lymphoid and myeloid leukaemia was treated with cytotoxic chemotherapy. After initiation of the consolidation phase, she developed a neutropenic fever associated with respiratory distress. Detection of galactomannan antigen and β-d-glucan fungal markers in serum were positive (7.1 for index values of galactomannan in patient serum over standard control and 720 pg ml^–1 for β-d-glucan) and chest high-resolution CT scan showed multiple lesions in the lungs. Criteria for a probable pulmonary aspergillosis were met and treatment with intravenous VRC was started with a loading dose of 300 mg twice daily (about 6mg kg^–1 12 h^–1) the first day followed by a maintenance dose of 200 mg twice daily (about 4mg kg^–1 12 h^–1). On the 15th day of treatment, TDM of VRC showed a serum concentration of 1.1 mg l^–1 (therapeutic range 1 to 5.5 mg l^–1 1). At day 17, the patient suffered from cough and haemoptysis and detection of galactomannan and β-d-glucan in serum remained positive (2.7 and 210 pg ml^–1, respectively) and chest CT scan revealed newer lesions in lungs, suggesting a refractory aspergillosis. Intravenous administration of liposomal amphotericin B at 180 mg once a day (about 3.6 mg kg^–1) was added to VRC, which was continued orally. The VRC trough concentration was 0.4 mg l^–1 2 weeks after the intravenous to oral switch. Since no drug interaction was suspected, the low concentration of VRC was attributed to a decrease of oral bioavailability caused by the repeated vomiting the patient had suffered, and VRC was re-administered intravenously. However, VRC trough concentration still was extremely low, at 0.12 mg l^–1 and, surprisingly, despite increasing the dose to 300 mg twice a day, it remained at 0.13 mg l^–1. VRC serum concentrations were then measured at four sampling points after a 2 h intravenous infusion of 400 mg (about 7 mg kg^–1) and showed the following results: 3.1, 0.95, 0.16 and 0.11 mg l^–1 at 2, 4, 8 and 12 h from the beginning of the infusion respectively. These results were confirmed by another laboratory, finding in the same samples that concentrations of VRC were 3.48, 1.10, 0.22 and 0.17 mg l^–1, respectively, and that concentrations of VRC-N-oxide, the main circulating metabolite of VRC in humans, were 10.6, 11.1, 3.47 and 2.13 mg l^–1, respectively (Figure1). CYP2C19 genotyping was performed and revealed a homozygous CYP2C19*17/*17 genotype. VRC was switched to intravenous caspofungin (50mg day^–1) and, in association with liposomal amphotericin B (180mg day^–1), fungal infection was controlled (serum galactomannan index value at 1 and serum β-d-glucan concentration at 125 pg ml^–1, associated with a regression of respiratory disorders) allowing the chemotherapy continuation and an allogeneic haematopoietic stem cell transplantation.

Figure 1.

Serum concentration of voriconazole (VRC) and voriconazole-N-oxide (VRC-N-oxide) over time after a 2 h intravenous administration of 400 mg voriconazole in a CYP2C19*17/*17 genotype patient. Dashed curve is drawn with extrapolated concentration for time 0. VRC, VRC-N-oxide, VRC-N-oxide: VRC ratio

From the 4-points of serum measurement, area under the VRC concentration–time curve was estimated at 11.1 mg l^–1 h. Total clearance was then estimated at 35.9 l h^–1, which was 1.4, 3.1 and 4 times higher than that previously measured in healthy volunteers who also received a 2 h intravenous infusion of 400 mg of VRC and carried the CYP2C19*1/*1 (associated with extensive metabolism), CYP2C19*1/*2-*1/*3 (associated with intermediate poor metabolism) or CYP2C19*2/*2-*2/*3-*3/*3 (associated with poor metabolism) genotypes respectively (3 and online supplementary material, Table 1). Interestingly, the concentration ratio of VRC-N-oxide over VRC increased very quickly after the infusion reaching 10.2 at 12 h which was about 15 to 20 times higher than previous reports after a single dose or at steady-state in cohorts that did not include the CYP2C19*17/*17 genotype (3,4 and online supplementary material, Table 1). The high metabolic ratio observed in our patient could be partly explained by her CYP2C19 genotype corroborating several recent reports of VRC underexposure in CYP2C19*17 homozygotes 5–9. However, diseases (she was suffering from severe sepsis and haematological malignancy) and co-medications may also have modulated the CYP2C19 and CYP3A4 activities and the contribution of each in VRC N-oxidation. Particularly, concomitant treatment with methylprednisolone may have strengthened the enzymatic elimination of VRC as suggested in a recent population pharmacokinetic modelling based on published data 10. Other drug–VRC interactions cannot be excluded since several co-medications were also known to be substrates and/or potential modulators of these cytochromes [online supplementary material, Table 2].

Finally, from a practical point of view, the measurement of serum VRC-N-oxide concentration may help in the TDM of VRC. A low VRC concentration associated with a high concentration ratio of VRC-N-oxide over VRC- in our case, this ratio was over 10 when the VRC concentration was below 1.1 mg l^–1 - may suggest an ultrarapid metabolism of VRC and should prompt adaptation of the treatment by switching to alternative medicines rather than increasing the VRC dosage. Whether the CYP2C19*17/*17 genotype is indeed responsible for an increase in VRC N-oxidation will need to be confirmed prospectively. Moreover, the clinical relevance of a probable interaction between methylprednisolone and VRC deserves to be investigated.

Competing Interests

All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.