External Validation of Population PK and PD Models for Characterizing the Impact of CYP2C9 and VKORC1 Genotypes in the Management of Warfarin Therapy in Caribbean Hispanic Patients.
Jorge Duconge(1), Gledys Reynaldo-Fernández(2), Karine Rodriguez-Fernández(3), Leyanis Rodríguez-Vera(4), Stephanie Reyes(1), Camila de las Mercedes de las Barreras(2), Darlene Santiago(1), Víctor Mangas-Sanjuan(3,5)
(1) Department of Pharmaceutical Sciences, School of Pharmacy, University of Puerto Rico – Medical Sciences Campus, 3rd Floor, R 325-14, Pharmacy Building, PO Box 365067, 00936-5067, San Juan, Puerto Rico, USA, (2) Institute of Pharmacy and Foods, University of Havana, Havana, Cuba, (3) Department of Pharmacy and Pharmaceutical Technology and Parasitology, University of Valencia, Valencia, Spain, (4) Center for Pharmacometrics & System Pharmacology. College of Pharmacy. University of Florida, (5) Interuniversity Research Institute for Molecular Recognition and Technological Development, Polytechnic University of Valencia, Valencia, Spain.
Objectives: The role of genetic polymorphisms in partially characterizing interindividual variability (IIV) in warfarin exposure and clinical response has been established. Allelic variants have been shown to increase the international normalized ratio (INR) and reduce warfarin dose requirements [1]. Population pharmacokinetics (PK) and pharmacokinetics/ pharmacodynamics (PK/PD) models accounting for the impact of genetic polymorphisms on PK and PD parameters have been published and serve as the basis for our work [2, 3]. However, their predictive performance and comparability remain unknown. The aims of this study are to (i) evaluate the prediction ability of PK modeling structures to characterize the impact of CYP2C9 on warfarin exposure, (ii) externally validate a PD model of INR after concomitant administration of warfarin in clinical practice, and (iii) quantitatively characterize the impact of individual CYP2C9 and VKORC1 genotypes on PD parameters in a cohort of Caribbean Hispanic patients.
Methods: The study design included 6 to 8 INR measurements per patient for 1037 INR measurements from 138 enrolled patients who received treatment before study initiation. They were primarily male and elderly taking warfarin for various thromboembolic disorders at the Veterans Affairs Caribbean Health System-affiliated anticoagulation clinic (VACHS) in San Juan, Puerto Rico. These patients carry different CYP2C9 alleles (*1, *2, *3, *5, *8) and VKORC1 genotypes (G/A, G/G, A/A). First, two PK models previously published [2, 3] were implemented to retrieve the typical PK parameter estimates for each CYP2C9 genotype sub-groups. Secondly, a previously published PD model for INR was used [3], which relies on an indirect response model coupled with a series of transit compartments to account for the delay between warfarin exposure and INR levels. The effect of warfarin concentration was incorporated using a sigmoid function. A numerical evaluation of the parameter-covariate relationship was performed manually in univariate testing, which included body weight, age, CPY2C9 and VKORC1 genotypes, race, diabetes mellitus, and smoking status. Model evaluation was performed through pc-VPC (n=1000) and GOF plots. Experimental PK and PD data were logarithmically transformed. All data analyses were performed based on the population approach with NONMEM v7.4 software. The study was approved by the Institutional Review Board (#A4070109). The study was conducted following Helsinki’s declaration for human subject protection in clinical surveys.
Results: The simulation-based analysis revealed similar performance across the CYP2C9 polymorphisms when population PK model structure [2] (median Cmax: 4.67-4.81 mg/L) or PK model structure [3] (median Cmax: 4.61-4.86 mg/L) were considered. Since our published PK model [2] incorporates the prediction in 8 out of 9 polymorphisms, it was selected as the PK reference model. CYP2C9 and VKORC1 polymorphisms were statistically significant on baseline INR value and IC50 (p<0.05). The final system-related PD estimates (RSE%) were MTT= 0.53 (18%) days, MTT2= 942 (77%) days, and baseline were 6.92 (7%), 6.82 (15%), 7.95 (21%), 2.30 (9%), 6.93 (53%), 4.95 (39%), 16.9 (31%), and 13.70 (4%) for CYP2C9*1/*1, *1/*2, *1/*3, *1/*5, *1/*8, *2/*2, *2/*3, *2/*5 genotypes, respectively. Drug-related PD parameters were λ=1.57 (26%), γ=0.51 (26%), and IC50 were 3.19 (53%), 3.41 (52%), and 2.04 (30%) mg/L for VKORC1G/A, G/G, and A/A genotypes, respectively.
Conclusions: An external validation of PK and PD modeling structures has been successfully conducted to characterize long-term (>3 years) INR levels after warfarin administration in Caribbean Hispanic patients. The results confirmed previous conclusions about the impact of VKORC1 genotypes on the inhibitory potency of warfarin. In addition, we have been able to identify the effect of the CYP2C9 genotype on baseline INR values, although these values may be influenced by the fact that patients were on active warfarin therapy before study enrollment.
References:
[1] Li J, Wang S, Barone J, Malone B. Warfarin pharmacogenomics. Pharm Ther 2009; 34:422–7.
[2] Reyes-González et al. Genotype-driven pharmacokinetic simulations of warfarin levels in Puerto Ricans. Drug Metabol Pers Ther 2020; https://doi.org/10.1515/dmdi-2020-0135.
[3] Hamberg AK, Dahl ML, Barban M, Scordo MG, Wadelius M, Pengo V, et al. A PK-PD model for predicting the impact of age, CYP2C9, and VKORC1 genotype on individualization of warfarin therapy. Clinical pharmacology and therapeutics. 2007; 81(4):529-38.