Induction of Voriconazole Metabolism by Rifampin in a Patient with Acute Myeloid Leukemia: Importance of Interdisciplinary Communication To Prevent Treatment Errors with Complex Medications

Voriconazole is metabolized by the CYP450 isoenzymes 2C19 and 3A4 and, to a lesser extent, by CYP2C9 (3). Several drug-drug interactions with voriconazole can be expected because many other drugs are also transformed by these enzymes, leading to a contraindication for coadministration of voriconazole with some other drugs. Rifampin (a CYP450 inducer; 600 mg once daily) decreased the C_max (maximum plasma concentration) and AUCτ (area under the plasma concentration-time curve within a dosing interval) of voriconazole by 93% and 96%, respectively (1). However, detailed data about coadministration of voriconazole and rifampin have not been formally published.

In an ongoing study on the steady-state pharmacokinetics of voriconazole, 31 patients receiving voriconazole during hospitalization, as a regular therapeutic drug treatment according to the indications of the drug, were included. A 12-h (dosing interval) pharmacokinetics profile after dosing was obtained from each participating patient when treatment began (first dose, 400-mg oral loading dose) or (if possible and) after he or she had received voriconazole for at least 14 days (200-mg oral dose twice a day). Therefore, blood was taken at certain times. Patients had to be over 18 years old to be admitted to the trial. Exclusion criteria included a hemoglobin level under 7 g/dl in the last laboratory screening (not older than two days) and inability to communicate well with the investigator due to language problems or poor mental development. The study protocol (EudraCT: 2005-005188-27) was approved by the ethics committee of the University Hospital, Heidelberg, and has been authorized by the competent authority (BfArM, Germany). We obtained written informed consent from each patient prior to any study-related activity.

On the day of obtaining the steady-state voriconazole pharmacokinetics profile, we identified one patient as having accidentally been treated with the contraindicated combination of voriconazole and rifampin. Voriconazole was started because of persistent fever of unknown origin after chemotherapy for acute myeloid leukemia and was continued for antifungal prophylaxis. Blood samples from this patient were collected after the first 400-mg voriconazole dose. Treatment with rifampin was suggested by a clinical microbiologist 7 days after the start of antifungal therapy because of a Staphylococcus epidermidis infection proven by cultures of blood samples from the central venous catheter, as well as the peripheral blood. Obviously, the microbiologist wasn't aware of the patient's voriconazole therapy, while the treating physicians unfortunately accepted this suggestion without another check for drug-drug interactions and didn't recognize the contraindicated combination, so that these two drugs were taken together. In obtaining the steady-state pharmacokinetics profile (200 mg dose) after 36 days of voriconazole administration, this treatment error was perceived and the voriconazole prophylaxis was stopped immediately.

The plasma concentrations of voriconazole and three of its metabolites (N-oxide, hydroxyvoriconazole, and dihydroxyvoriconazole) were determined by using a fully validated (4) liquid chromatography-tandem mass spectrometry assay. Nonparametric pharmacokinetic parameters were calculated using WinNonlin 5.2 (Pharsight, Mountain View, CA).

After the starting dose (400 mg), plasma concentrations in this patient were well within the anticipated range, with a C_max of 3.92 μg/ml and an AUC_0-inf of 27.4 h μg/ml (2). After 36 days of voriconazole therapy and 30 days of rifampin therapy, the total voriconazole exposure was dramatically decreased, by 99% (Fig. 1), with a C_max of 0.038 μg/ml and an AUC_0-12 of 0.145 h μg/ml. In addition, the plasma concentrations of the three main metabolites were similar or even increased compared to the concentrations following the first dose without rifampin. In consequence, the mean metabolic ratios in plasma of the metabolites divided by the parent drug (corrected for molecular weight) showed large increases (N-oxide, 45-fold; hydroxyvoriconazole, 178-fold; dihydroxyvoriconazole, 422-fold), suggesting indeed the induction of voriconazole metabolism by rifampin. Because voriconazole itself and not one of its metabolites is the agent which is most responsible for the antifungal effect, there will be a loss of effectiveness of voriconazole caused by rifampin.

FIG. 1.

Plasma voriconazole concentrations after treatment without rifampin (circles; first dose 400 mg voriconazole) and after 30 days of rifampin treatment (squares; 400 mg voriconazole daily).

In conclusion, coadministration of voriconazole and rifampin will result in a loss of the therapeutic efficacy of voriconazole due to the massive reduction of systemic voriconazole exposure due to induced metabolism. Furthermore, this case emphasizes the need for good communication between treating physicians and consulting experts of other departments, as well as the importance of clarifying the reasons for prescribing a drug and of regular reviews of complex medication regimens.