A Mechanism Based Population Pharmacokinetic-Pharmacodynamic Model for Epoetin Alfa and Darbepoetin Alfa in Chronic Kidney Disease Patients
Amit Garg (1), Anis Khan (1), William Hanley (1), Janet van Adelsberg (1), Amy Ko (1), Elizabeth Quackenbush (1), Michael Crutchlow (1) and Sunny Chapel (2)
(1) Merck & Co., Inc, Whitehouse Station, NJ, USA; (2) Ann Arbor Pharmacometrics Group, Ann Arbor, MI, USA
Objectives: The purpose of this study was to develop a mechanistic longitudinal pharmacokinetic / pharmacodynamic (PK/PD) model for the characterization of the erythropoietic effects of epoetin alfa, darbepoetin alfa and switching of one agent to another, in chronic kidney disease patients on hemodialysis.
Methods: Long term dialysis data for darbepoetin and epoetin was obtained from dialysis center. About eleven hundred male and female subjects (52% males and 48% females) were included in the analysis. The dataset contained laboratory values (hemoglobin, reticulocytes), records of dose adjustments, patient demographics (body weight, age, sex, BMI etc.) and other factors such as creatinine clearance, concomitant medications, etiology of CKD etc. for a treatment duration of 3-12 months for each patient, during the period of 2002 and 2008. The time course of red blood cell production (reported as hemoglobin concentration) on epoetin and/or darbepoetin was described based on the hematopoiesis processes. Darbepoetin and epoetin PK parameters were obtained from published literature [1, 3]. The population analysis was performed using the non-linear mixed effects modeling approach implemented in NONMEM V. The predictive performance of the final model was assessed by conducting a posterior predictive check (PPC) and by external predictive check.
Results: A catenary cell production and life-span based indirect response model was developed to describe the pharmacodynamics of epoetin and darbepoetin alfa. This mechanistic model modified from published work [2, 3] consisted of cell life span of normoblasts, reticulocytes, and red blood cells. A linear concentration-response model was selected to describe the effects of epoetin and darbepoetin on erythropoiesis, as data didn't support nonlinear relationships such as Emax and Power models. The mean hemoglobin values from the observed data were in good agreement with the distribution of hemoglobin values obtained from the simulated data (internal PPC). In addition, the simulated hemoglobin data from PK/PD model were in good agreement with the observed external data, confirming the predictability of the model. At 0.45 μg/kg/week dose (darbepoetin phase 2 dose-finding and dose-scheduling study protocol 960245), the observed change in hemoglobin from baseline at 4 weeks was 1.27 (0.55, 2.00) g/dL - mean (95% CI) while the model predicted change from baseline was 1.06 g/dL (0.3, 1.9), whereas, the observed change from baseline for the darbepoetin protocol 980211 was 1.1 g/dL (0.82, 1.37) while the model predicted change from baseline was 1.05 g/dL (0.77, 1.32) (top 10% of the non-responders were excluded from simulated data to represent clinical study population and exclusion/inclusion criteria).
Conclusion: The PK/PD model adequately described the longitudinal hemoglobin-time data in chronic kidney disease patients. The PK/PD model has the potential to inform future trial design including switching from one agent to the other and to evaluate dose titration strategies.
 Agoram B, Sutjandra L, and Sullivan JT. Population pharmacokinetics of darbepoetin alfa in healthy subjects. Br J Clin Pharmacol. (2007) 63(1): 41-52.
 Ramakrishnan R, Cheung WK, Wacholtz MC, Minton N, Jusko WJ. Pharmacokinetic and pharmacodynamic modeling of recombinant human erythropoietin after single and multiple doses in healthy volunteers. J Clin Pharmacol. (2004) 44(9): 991-1002.
 Krzyzanski W, Jusko WJ, Wacholtz MC, Minton N, Cheung WK. Pharmacokinetic and pharmacodynamic modeling of recombinant human erythropoietin after multiple subcutaneous doses in healthy subjects. Eur J Pharm Sci. (2005) 26(3-4): 295-306.