Sammendrag
Background:
Cytomegalovirus (CMV) is the most common pathogen responsible for morbidity and mortality in immunosuppressed individuals. One population at high risk is solid organ transplant (SOT) recipients.
Intravenous ganciclovir is the gold standard for treatment of CMV disease. Oral valganciclovir is a prodrug of ganciclovir, with an almost 60% oral bioavailability, and has recently been shown to be non-inferior to intravenous ganciclovir in the VICTOR study.
The purpose of this thesis was to develop a pharmacokinetic population model to describe the ganciclovir plasma concentration observations in the patient population of the VICTOR study, after receiving oral valganciclovir for treatment of CMV disease. Additionally the model was to be validated, in order to make it clinically applicable.
Methods:
Of the 321 patients who were included in the VICTOR study 164 were randomized into the valganciclovir arm, and in 108 was the plasma concentration of ganciclovir measured. The development of the population pharmacokinetic model was based on these 108 patients. The data contained information of demographical, physiological and pathophysiological nature, as well as the measured drug concentrations. The model was constructed by use of the NONMEM version VI computer program, as well as PREDPP and R for subroutines and graphs, respectively.
Results:
A 2-compartment model with first-order absorption was found to be the overall best model for the data set. The ADVAN4 routine was used in combination with the TRANS4 subroutine. After a forward inclusion and backwards deletion procedure creatinine clearance and gender were significant covariates in the model. The average (± SD) parameter estimates for the final model were CL/F = 12±0.05 L/h, V2/F = 59.4±2.76 L, Q/F = 7.21±0.20 L/h, V3/F = 304±7.79 L.
The validation procedure employed a K-fold cross-validation, as recommended by the FDA, and the Jackknife technique. The model appeared stable, but susceptible to variations in the data-set (Central volume had a range of 57.1 – 93.8 L, while peripheral volume had a range of 270 – 357 L.
Conclusions:
The model fulfills the goal of predicting ganciclovir plasma concentrations to a satisfying degree, especially for individually predicted concentrations, although it has room for improvement. It is a good basis for the further development of a PK/PD model that compares ganciclovir pharmacokinetics with both therapeutic effects and adverse events.