Population pharmacokinetics of ticagrelor and active metabolite during veno-arterial extracorporeal membrane oxygenation
Kyoung Lok Min (1, 2), Ryu Kyong Kuk (1), Soyoung Kang (1, 2), Jongsung Hahn (1, 3), Jin Wi (4, 5), Min Jung Chang (1, 2, 6)
(1) Department of Pharmacy and Yonsei Institute of Pharmaceutical Sciences, Yonsei University, Incheon, Republic of Korea, (2) Department of Pharmaceutical Medicine and Regulatory Science, Yonsei University, Incheon, Republic of Korea, (3) School of Pharmacy, Jeonbuk National University, Jeonju, Republic of Korea, (4) Division of Cardiology, Department of Internal Medicine, Gachon University Gil Medical Center, Incheon, Republic of Korea, (5) Division of Cardiology, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea, (6) Graduate Program of Industrial Pharmaceutical Science, Yonsei University, Incheon, Republic of Korea
Objective: Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is a mechanical method for supporting the cardiopulmonary systems usually used for treating cardiac failure and pulmonary failure in critically ill patients. Ticagrelor is used as dual anti-platelet therapy (DAPT) in treating acute coronary syndrome, which is the main cause of cardiac failure. Ticagrelor has a bleeding risk at excessive exposure and a risk of stent thrombosis In case of underexposure. In addition, the pharmacokinetics (PK) of the drugs change during VA-ECMO and the risk of bleeding is high due to the co-administration of heparin. Additionally, ticagrelor is lipophilic and has high protein binding affinity, so its PKs are highly likely to be changed during VA-ECMO. Therefore, studies on proper dosage regimen to maintain the blood concentration are necessary.
The aim of the study was to evaluate PK parameters and factors that affect ticagrelor’s PKs, develop a population PK model for ticagrelor, and to determine the optimal dosage regimen for ticagrelor during VA-ECMO.
Methods. This prospective observational study was performed at a coronary intensive care unit. Patients taking ticagrelor orally undergoing VA-ECMO were included in this study. Blood samples were collected pre-dose and, 1, 2, 3, 6, 8, and 12 h after ticagrelor administration during and after VA-ECMO. Ticagrelor and its active metabolite, AR-C124910XX, were analyzed by liquid chromatograph-tandem mass spectrometer (LC-MS/MS). Population PK model analysis was conducted using Nonlinear Mixed Effect Model (NONMEM)® software. To develop error model for inter-individual variability and residual unexplained variability (RUV), the exponential random effect model, and combined additive and proportional residual error model were used. PK parameters including systemic clearance of ticagrelor (CL/F), systemic clearance of metabolite (CLM/(fm×F)), central volume of distribution of ticagrelor (Vd/F), central volume of distribution of metabolite (VM/(fm×F)), and first-order absorption rate constant (Ka) were evaluated using first-order estimation with an iterative approach. Additive, proportional, and combined additive and proportional error models were used to assess residual variability. Potential covariates such as sex, age, height, weight, smoking status, and their laboratory data, ECMO flow rates, and dialysis records for the final model were evaluated. The validity of final models was confirmed by visual inspection of goodness-of-fit plot. Visual predictive check (VPC) and nonparametric bootstrap resampling methods were to evaluate the final model.
Results. Ticagrelor was described as a one-compartment model, and its metabolite was also best described by the one-compartment model. Inter-individual variability terms were included on CL/F, Vd/, CLM/(fm×F), and VM/(fm×F). Residual variability was described with an additive and proportional error model. The covariate analysis showed improved model fit by inclusion of ECMO as a covariation on CL/F and Vd/F, and LPM (liters per minute) on Vd/F. The final model was as follows: CL/F (L/h) = 12.2 × 0.35ECMO; Vd/F (L) = 157 × 0.518(LPM/3) × 2.74ECMO; CLM/(fm×F) (L/h) = 8.8; VM/(fm×F) (L) = 29.5; Ka (h-1) = 0.533.
The goodness-of-fit plot showed that conditional weighted residual was evenly distributed around zero, and individual predictions and observations were highly correlated. The visual predictive check (VPC) showed that most data fit within the 90% confidence interval (CI) and 5000 bootstrap 95% confidence interval (CI) results were similar with the parameters of the final model of ticagrelor.
Conclusion. This study was the first population PK study of ticagrelor in VA-ECMO patients. The use of VA-ECMO decreased the CL/F and increased the Vd/F of ticagrelor. Through this study, more evidence could be introduced to optimize ticagrelor dosage in patients undergoing VA-ECMO.
(This work was supported by a grant (No. 2020R1F1A107054913) from the National Research Foundation (NRF) of Korea, funded by the Korean government (Ministry of Science, ICT & Future Planning.)