Voriconazole Population Pharmacokinetics

Background: Voriconazole is widely used in the treatment of life-threatening invasive fungal infections (IFIs). The use of voriconazole is problematic due to its highly variable pharmacokinetics and narrow therapeutic index which complicate dosage selection and adjustment; few population pharmacokinetic data are available.

Aim: This study aimed to develop a voriconazole population model incorporating data from multiple studies of voriconazole pharmacokinetics in healthy volunteers and patients. The study also aimed to assess demographic and clinical covariates that may affect voriconazole pharmacokinetics.

Methods: Four studies including 63 healthy volunteers with rich pharmacokinetic sampling were included in addition to sparse sampling data from 146 patients receiving voriconazole for the treatment of IFIs. Non-linear mixed effects modelling was carried out using NONMEM 7.2. Studies were added in a stepwise manner and a base model developed from the rich data; sparsely sampled patient data was then incorporated into the model. Goodness-of-fit criteria and visual predictive checks guided model selection and testing of covariates.

Results: Voriconazole pharmacokinetic data was best described by a two-compartment model with an absorption lag time and Michaelis-Menten elimination; a model incorporating mixed linear and non-linear elimination did not improve the fit. Voriconazole is primarily metabolised by CYP2C19; CYP2C19 heterozygous extensive metabolisers (CYP2C19*1/*2) and poor metabolisers (CYP2C19*2/*2) had a 57% lower Vmax compared to homozygous extensive metabolisers (CYP2C19*1/*1). Voriconazole elimination was significantly faster after a single dose than at steady state, indicating likely auto-inhibition of metabolism. Co-administration of the CYP3A4 inhibitor ritonavir decreased the Vmax by 43%, whereas concomitant administration of St John’s Wort increased the Vmax by 106%. The strong CYP450 inducers phenytoin and rifampicin increased the Vmax 205%. Voriconazole bioavailability was estimated to be 88%.

Conclusions: CYP2C19 genotype is a major intrinsic determinant of voriconazole disposition. Following further development of this model future work will involve dosing simulations to target the narrow therapeutic range and assessment of the feasibility of using this model in a Bayesian dose forecasting tool to aid clinicians in voriconazole dosage selection and adjustment.