TRACES pilot pharmacokinetic study

Published: 8 October 2020| Version 2 | DOI: 10.17632/z5mhfckr29.2
Sixtine Gilliot,


This dataset summarizes raw data used to build the pharmacokinetic model described in Gilliot et al. The dataset was designed to be ready to be analyzed by the software MONOLIX 2019R1 (Lixoft, Orsay, France). Patients were included if they were undergoing haemorrhagic cesarean section (blood loss > 800 mL) and receiving a single i.v dose of TA (0.5, 1 or 2 g over 1 minute). Non-inclusion criteria used were presented in the TRACES pilot study protocol.[1] The first seven columns on the left of the dataset describe the pharmacokinetic data. The blood TDM samples were obtained from our patients at T0 (inclusion time, when bleeding ≥ 800 mL is diagnosed), T1 (at the end of injection), T15, T30, T60, T120, T180, and T360 (defined as 15, 30, 60, 120, 180 and 360 min after the injection). The urinary samples were collected within 6 hours after treatment. From the eight column, the variables describe the anthropometric, biological and clinical characteristics of the patients included in the study. Vs1 represents the volume of blood loss at the inclusion time in L. SCr1 represents the serum creatinine at inclusion time and SCr2 represents the serum creatinine 360 after the administration of tranexamic acid, in mg/L. Uremia is expressed in g/L.


Steps to reproduce

1. Methods for tranexamic acid assay Data were analyzed by the toxicology laboratory of the Lille centre. Fifty microliters of plasma or pre-diluted (1/10th) urine were mixed and centrifuged (4500 g, 4 °C, 10 min) after addition of 400 μL of methanol containing 7β-hydroxyethyl-theophylline (Sigma-Aldrich, Saint Quentin Fallavier, France) at 20 mg.L-1, as an internal standard. The supernatant (20 μL) was added to water/formic acid 0.1% (180 μL). Measures were carried out on liquid chromatography system coupled with tandem mass spectrometry (Acquity Xevo-TQ Detector, Waters, Milford, MA, USA) following a fully validated method.[8] The lower limit of quantification was 2 mg/L. Linearity was tested using the calibration range 5–200 mg/L (r² =0.995). Intra-day and interday precisions were <3.80% and 5.30%, respectively, for a 20-mg/L-spiked sample and <2.90% and 4.15%, respectively, for a 150-mg/L-spiked sample. The system was equipped with an HSS T3 column (1.8 μm × 2.1 × 50 mm) maintained at +50°C. Compounds were separated using a mobile phase gradient consisting of methanol and formic acid. Ions of each analyzed compound were detected in a positive ion mode using multiple reaction monitoring (MRM). An injection volume of 5 µL was used for all the analyses. Data acquisition and quantification were performed using MassLynx 4.1 Software (Waters). 2. Additional analysis for the pharmacokinetic model The pharmacokinetic parameters were estimated using the parametric NLMEM software program Monolix 2019R1 (Lixoft, Orsay, France) by applying the stochastic approximation expectation-maximization algorithm (SAEM) combined with a Markov Chain Monte Carlo procedure. This SAEM algorithm estimates for each patient the distribution of the fixed parameters. The fixed parameters were computed without any approximation of the model (no linearisation).


Centre Hospitalier Regional Universitaire de Lille