Frédéric Gaspar 1,2,3, Léon Desmeules Jules Albert 1, Jean Terrier 4,5, Pauline Gosselin 4,5, Pierre Fontana5,6, Jean-Luc Reny 4,5,7, Youssef Dali 8,9, Chantal Csajka 1,2, Monia Guidi 1,2,3
1 Center for Research and Innovation in Clinical Pharmaceutical Sciences, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland 2 Institute of Pharmaceutical Sciences of Western Switzerland, University of Geneva, University of Lausanne, Geneva & Lausanne, Switzerland 3Service of Clinical Pharmacology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland 4 Division of General Internal Medicine, Geneva University Hospitals, Geneva, Switzerland 5 Geneva Platelet Group, Faculty of Medicine, University of Geneva, Geneva, Switzerland 6Division of Angiology and Haemostasis, Geneva University Hospitals, Geneva, Switzerland 7 Division of Internal Medicine and Rehabilitation, Geneva University Hospitals, Geneva, Switzerland 8 School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Geneva, Switzerland 9Clinical Pharmacology and Toxicology Service, Anesthesiology, Pharmacology and Intensive Care Department, Geneva University Hospitals, Geneva, Switzerland
Objectives: Antithrombotics Platelet receptor blockers are commonly prescribed to lower the risk of cardiovascular disease (CVD), but are also among the most high-risk medications [1]. Over- and sub-dosage have been indeed associated with bleeding and recurrent ischemia, but, unfortunately, no reliable tool to predict the right dosage for a given patient exist [2]. This issue could be solved using pharmacokinetic (PK) modeling. Population pharmacokinetics (PopPK) is widely used today to propose a priori dosages adapted to specific individual characteristics, and a posteriori complementing with drug measurements. [GM1] Recent advances of physiologically-based PK models (PBPK) may be an alternative for a priori personalized treatments [3]. In this study, we compared the prediction of individual drug exposure of the two modelling approaches for ticagrelor, a widely prescribed antiplatelet agent.
Methods: A “bottom-up” PBPK model for Ticagrelor was used to predict drug exposure in two different populations: healthy volunteers included in a single dose PK study and an inpatient population enrolled in a clinical study. Individual physiological, demographic and genetic characteristics (CYP3A4/5 phenotype) were used to create “virtual twins[GM1] ” in SIMCYP® for both populations. Predicted (AUCSIMCYP,pred) and observed (AUCobs, calculated with Pheonix®) systemic exposures were then compared in each subject by mean-fold error (MFE=AUCSIMCYP,pred/AUCobs) to assess the predictability of the PBPK approach. Two validation criteria were applied: MFE 0.5 to 2.0, a criterion commonly used in PBPK modelling, and MFE 0.8 to 1.25, largely used for bioequivalence studies. Meanwhile, a “top-down” popPK model was built using full PK profiles collected in both populations (NONMEM). Several covariates (age, body weight (BW), height, body mass index (BMI), ethnicities, liver function, comedications, and cytochrome (CYP) 3A4 and PgP activity) were tested on Ticagrelor PK parameters. The final popPK model was used to estimate subject exposure (AUCNONMEM,pred), that were compared to AUCobs as previously described for AUCSIMCYP,pred.
Results: A total of 427 plasma concentrations of Ticagrelor from 55 subjects were available for analysis: 160 concentrations from 20 healthy volunteers and 267 concentrations from 35 hospitalized patients. PopPK analysis showed that a two-compartment model with linear absorption and elimination best described the PK of Ticagrelor. Clearance[GM1] , central and peripheral compartment distribution volume, first-order absorption rate constant, and inter-compartmental clearance were estimated at 44 L/h (Between subject variability 22.9%), 68.5 L (71%), 222 L, 0.32 h-1, and 18.7 L/h, respectively. Among the variables tested on PK parameters, only CYP3A4 had a significant influence on clearance (p=0.012), found 10.2% lower in poor and 29.6% higher in rapid compared to extensive metabolizers. Ticagrelor AUCobs was adequately predicted in all subjects using the large criterion, and in 95% of healthy volunteers and 97% of the inpatients with the stricter criterion (MFE of 0.83 IQR [0.76-0.85] and of 1.11 IQR [1.06-1.16],respectively). Following virtual twins model construction in Simcyp®, 100% and 70% of healthy volunteers (MFE of 1.13 IQR [0.99-1.23]) and only 69% and 23% of the inpatients (MFE of 0.79 IQR [0.47-0.92]) [GM2] [MOU3] satisfied the large and the stringent criterion, respectively.
Conclusions: The present study showed significant between-subject [CP1] variability in the PK of ticagrelor of which a large proportion remained unexplained after inclusion of covariates[CP2] . The PopPK model proved to be highly predictive for both tested populations, while PBPK model only for healthy subjects despite the individualization of patient characteristics. An optimization of the model used should be envisaged and in particular in the creation of a virtual population closely matching the patients in terms of physiology, genetics and diseaseAn optimisation of the model used should be envisaged and in particular in the creation of a virtual population more similar to reality. To move the concept of PBPK modeling in precision medicine, clinical relevance in patient care are still required. Despite the uncertainties, these preliminary data demonstrate that PBPK modeling offers future perspectives in personalized medicine.
References:
[1] Terrier, J., et al., Towards Personalized Antithrombotic Treatments: Focus on P2Y12 Inhibitors and Direct Oral Anticoagulants.Clin Pharmacokinet, 2019. 58(12): p. 1517-1532.
[2] Lee, C.W., et al., Optimal duration of dual antiplatelet therapy after drug-eluting stent implantation: a randomized, controlled trial.Circulation, 2014. 129(3): p. 304-12.
[3] Marsousi, N., et al., Usefulness of PBPK Modeling in Incorporation of Clinical Conditions in Personalized Medicine. J Pharm Sci, 2017. 106(9): p. 2380-2391.
Reference: PAGE () Abstr 9564 [www.page-meeting.org/?abstract=9564]
Poster: Clinical Applications