PK/PD of unfractionned heparin during cardiopulmonary bypass: an individual patient data meta-analysis of two studies
Audrick Gibert (1), Julien Lanoiselée (1,2), Isabelle Gouin (3), Adeline Pontis (3), Xavier Delavenne (1,2), Edouard Ollier (1,2)
(1) Université Jean Monnet Saint-Étienne, Mines Saint-Etienne, INSERM, SAINBIOSE U1059, F-42023, SAINT-ÉTIENNE, France, (2) CHU Saint-Étienne, France, (3) Université Rennes, Inserm, EHESP, Irset (Institut de Recherche en Santé, Environnement et Travail) - UMR_S 1085, Rennes, France.
Introduction:
Unfractionated heparin (UFH) is commonly used during cardiac surgery under cardiopulmonary bypass (CPB) to prevent blood activation and thrombosis in the circuit. The effect of UFH is monitored in real time during CPB using a point-of-care device that returns an activated clotting time (ACT). However, according to current recommendations, UFH is administered empirically to achieve a target ACT of 400 s. This last point leads to varied individual responses depending on patients characteristics. PK/PD modeling of UFH could allow to rationalize UFH dosing regimen and perform real-time individualization. Several studies have already been conducted to model the PK/PD of heparin. However, these studies had some limitations:
- UFH PK could be quantified using anti-Xa or anti-IIa activity, but no studies jointly analyzed both [1,2]. The pharmacokinetics of UFH molecules with anti-Xa activity may differ from molecules with anti-IIa activity [3].
- Studies were conducted using a single type of heparin while different ones are used in routine.
- Few or no covariates were included in these studies.
- PD modeling was not adjusted on time-dependent factors known to influence ACT values (i.e. body temperature and hemodilution).
The present study aims to address these limits by performing an individual patient data meta-analysis of two studies.
Objectives:
- Create a PK/PD model of heparin based on data from two different studies with a joint analysis of anti-Xa and anti-IIa activities.
- Use the PK/PD model to individualize UFH dosing regimen based on patients’ covariates.
Methods:
This work was an individual patient data meta-analysis from two single-center prospective observational studies. The UFH used to managed anticoagulation differed in both studies: i) Panpharma UFH in study 1 and ii) CHOAY UFH in study 2. In both studies, ACT was measured at baseline and then routinely throughout surgery. 5-mL blood samples were collected 5, 15, and 30 minutes after the first bolus of heparin and 5, 30, and 60 minutes after the initiation of CPB. Based on these data, a PK/PD model of UFH was developed using a nonlinear mixed effects approach using MONOLIX® (Lixoft, version 2021R1) [4]. Covariates were added to the model using two different covariate search strategies: COSSAC (COnditional Sampling use for Stepwise Approach based on Correlation tests) [5] and SAMBA (Stochastic Approximation for Model Building Algorithm) [6]. The influence of time varying covariates such as hemodilution due to CPB priming or fluid infusions and body temperature was also studied.
Results:
86 patients were included. The anti-Xa and anti-IIa activities of heparin were best described by a two-compartment model with linear elimination. No differences in terms of PK parameters were found between anti-Xa and anti-IIa kinetics. Concerning covariates effect, UFH clearance was best explained by body weight and type of UFH and volume of distribution was best explained by body weight and pre-operative fibrinogen. PD data were best described by a log-linear model that considered the effect of hemodilution and body temperature. Based on the developed PK/PD model, equations were derived to individualized dynamically UFH dosing regimen in both continuous and intermittent settings based on the targeted ACT level, patient’s covariates and type of UFH. These equations included a correction factor to ensure target attainment in a desired fraction of the population (80% in our analysis) [7].
Conclusions:
The population model adequately characterized the PK/PD of UFH in patients undergoing CPB. This meta-analysis allowed the inclusion of new covariates associated with the PK/PD of UFH, which made it possible to provide personalized dosage regimens. The proposed model may be considered for a posteriori individualization use through Bayesian estimation.
References: [1] Delavenne X, Ollier E, Chollet S, Sandri F, Lanoiselée J, Hodin S, Montmartin A, Fuzellier JF, Mismetti P, Gergelé L. Pharmacokinetic/pharmacodynamic model for unfractionated heparin dosing during cardiopulmonary bypass. Br J Anaesth. 2017 May 1;118(5):705-712. doi: 10.1093/bja/aex044. PMID: 28510738.
[2] Jia Z, Tian G, Ren Y, Sun Z, Lu W, Hou X. Pharmacokinetic model of unfractionated heparin during and after cardiopulmonary bypass in cardiac surgery. J Transl Med. 2015 Feb 1;13:45. doi: 10.1186/s12967-015-0404-5. PMID: 25638272; PMCID: PMC4326208.
[3] Despas N, Larock AS, Jacqmin H, Douxfils J, Chatelain B, Chatelain M, Mullier F. Heparin monitoring: clinical outcome and practical approach. Ann Biol Clin (Paris). 2016 Dec 1;74(6):637-652. English. doi: 10.1684/abc.2016.1198. PMID: 27848915.
[4] Monolix 2019 User guide. Version 2021. http://monolix.lixoft.com/single-page/
[5] Ayral G, Si Abdallah JF, Magnard C, Chauvin J. A novel method based on unbiased correlations tests for covariate selection in nonlinear mixed effects models: The COSSAC approach. CPT Pharmacometrics Syst Pharmacol. 2021 Apr;10(4):318-329. doi: 10.1002/psp4.12612. PMID: 33755345; PMCID: PMC8099437.
[6] Prague M, Lavielle M. SAMBA: A novel method for fast automatic model building in nonlinear mixed-effects models. CPT Pharmacometrics Syst Pharmacol. 2022 Feb;11(2):161-172. doi: 10.1002/psp4.12742. Epub 2022 Feb 1. PMID: 35104058; PMCID: PMC8846636.
[7] Lanoiselée J, Zufferey PJ, Hodin S, Tamisier N, Gergelé L, Palao JC, Campisi S, Molliex S, Morel J, Delavenne X, Ollier E. Pharmacokinetic Model for Cefuroxime Dosing during Cardiac Surgery under Cardiopulmonary Bypass. Antimicrob Agents Chemother. 2020 Nov 17;64(12):e01687-20. doi: 10.1128/AAC.01687-20. PMID: 33020154; PMCID: PMC7674023.