| dc.description.abstract | Background: Cyclosporin A (CsA) has been a cornerstone of solid organ transplantation since its introduction to the market in the early 1980’s, and is a major part of the success of immunosuppression in the clinical setting. CsA is like many other immunosuppressive drugs, and has a narrow therapeutic window and large inter-individual variability. Highly variable drug exposure is associated with a high risk of organ rejection, and side effects like nephrotoxicity, infection, hepatotoxicity, and cancer. Obtaining the optimal exposure of the drug will not only prevent acute rejection, but also prolong the survival of the grafts, the organs, and inevitably the patients.
The overall purpose was to develop a pharmacokinetic population model for further use in later studies to improve therapeutic drug monitoring of CsA in renal transplant patients. Specific goals for the thesis include testing different compartment models with different absorption and elimination profiles, screen for possible covariates that may improve the compartment model, and finally validate the model.
Methods: Data was gathered from three separate studies, previously performed by the Department of Pharmaceutical Biosciences, School of Pharmacy, University of Oslo. 49 patients provided a total of 1027 plasma concentration samples and various patient demographics. By using the nonlinear mixed-effect modeling program NONMEM a pharmacokinetic population model was developed.
Results: A 2-compartment model with an absorption lagtime gave the best fit for the cyclosporin A data, with CL/F = 26.1 L/h, Q/F = 20.5 L/h, VC/F = 77.7 L, VP/F = 342 L, ka = 1.88 1/h, and ALAG = 0.452 h. Screening for covariates showed that age (years), body mass index (BMI, kg/m2), creatinine clearance (ml/min.), gender, height (m), lean body mass (LBM, kg), steroid dose (mg), post-transplantation time (weeks), and weight (kg) were significant to varying degrees. The data-splitting as recommended by Food and Drug Administration (FDA) was employed as an internal validation of the model. The resulting objective function value (OFV) was very variable and clearly showed that the model has a serious lack of robustness. The difference between the maximum and minimum value were 253.09, which is significantly more than the maximum allowed value of < 3.84. Inclusion of covariates that proved to be statistically significant may possibly have had their clinical significance overestimated. The criteria set up for inclusion of covariates in to the final model, will need revision to possibly remove some covariates and hopefully stabilize the model.
Conclusion: This model provides a good basis upon which a dosage regimen for cyclosporin A may be designed, though some further refinement may be needed to improve upon the models somewhat lack of robustness | eng |