Gastón GarcÃa-Orueta1, Zinnia P Parra-Guillén1,2, Iñaki F Trocóniz1,2,3
1Department of Pharmaceutical Sciences, School of Pharmacy and Nutrition, University of Navarra, 2Navarra Institute for Health Research (IdiSNA), 3Institute of Data Science and Artificial Intelligence, DATAI
Introduction. In pharmacokinetic (PK) studies, the number of samples is essential for building robust models. In pediatric patients undergoing Continuous Kidney Replacement Therapy (CKRT), the main challenge is the limited availability of subjects, further compounded by children’s lower blood sampling tolerance compared to adults [1]. Additionally, the high cost of each sample poses a significant constraint. Variations in the number of patients, samples per patient, or the distribution of samples across different matrices (plasma, prefilter, postfilter, effluent) can significantly impact model development. Therefore, optimizing the study design is crucial to maximizing efficiency while minimizing the required sample size. Objectives. The aim of this study is to propose optimal designs for the development of PK models for teicoplanin, piperacillin and meropenem in critically ill children, including those undergoing CKRT. Methods. For each antibiotic, a dataset comprising 14 individuals per group (with and without CKRT) was assembled, all of them with the median covariate values in the original population (weight = 8 kg, eGFR = 119 mL/min/m2, height = 69 cm, age = 12 months, filter size = 0.2 m2). Within the CKRT group, prefilter, postfilter, and effluent samples were incorporated into the dataset. Parameters from previously established models for teicoplanin, piperacillin and meropenem in critically ill children undergoing CKRT were employed [2,3,4]. The criterion for assessing and refining the designs was the evaluation of the population Fisher Information Matrix (FIM) with block-diagonal modality [5]. The software utilized for this study was NONMEM 7.5 [6], incorporating the $DESIGN option [7]. The number of samples necessary to achieve parameter RSEs lower than 40 % were selected and the sampling times were optimized. Finally, stochastic simulations and re-estimations (SSEs) without and with variations in covariates were conducted to validate the designs and test them for different covariate values. Results. The optimized designs resulted in a significant reduction in the number of samples. For teicoplanin, piperacillin and meropenem model development, a design with three, four and six sampling times per individual gave satisfactory results, respectively. In addition, omitting some prefilter or postfilter samples while maintaining effluent samples did not worsen significantly the expected relative standard error (RSE%) of the parameters. RSEs% predicted from FIM were similar to those from the original models, while the total number of samples needed has been reduced from 11.1 to a mean of 5 samples per patient in average. The conducted SSEs yielded similar RSEs to those predicted by the FIM, validating the study designs. Variations in covariate values did not worsen significantly parameter RSEs. Conclusions. An optimization of the study designs demonstrated that it is possible to develop population PK models for teicoplanin, piperacillin, and meropenem in pediatric patients undergoing CKRT with fewer samples and greater precision than what was achieved in the actual study, saving economic resources and reducing the inconvenience of blood sampling in children.
[1] Roberts, J. K., Stockmann, C., Balch, A., Yu, T., Ward, R. M., Spigarelli, M. G., & Sherwin, C. M. (2015). Optimal design in pediatric pharmacokinetic and pharmacodynamic clinical studies. Paediatric anaesthesia, 25(3), 222–230. [2] Butragueño-Laiseca, L., García-Orueta, G., Riva, N., Trocóniz, I. F., Fernández, S. N., Camacho Vicente, V., Padilla, B., Slöcker, M., & Santiago, M. J. (2025). Population pharmacokinetic analysis of teicoplanin in paediatric patients, including those receiving continuous kidney replacement therapy: a prospective cohort study. The Journal of antimicrobial chemotherapy, 80(3), 868–875. [3] Butragueño-Laiseca, L., Marco-Ariño, N., Troconiz, I. F., Grau, S., Campillo, N., García, X., Padilla, B., Fernández, S. N., Slöcker, M., & Santiago, M. J. (2022). Population pharmacokinetics of piperacillin in critically ill children including those undergoing continuous kidney replacement therapy. Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases, 28(9), 1287.e9–1287.e15 [4] Butragueño-Laiseca, L., Troconiz, I. F., Grau, S., Campillo, N., Padilla, B., Fernández, S. N., Slöcker, M., Herrera, L., & Santiago, M. J. (2024). How to use meropenem in pediatric patients undergoing CKRT? Integrated meropenem pharmacokinetic model for critically ill children. Antimicrobial agents and chemotherapy, 68(6), e0172923 [5] Retout, S., Duffull, S., & Mentré, F. (2001). Development and implementation of the population Fisher information matrix for the evaluation of population pharmacokinetic designs. Computer methods and programs in biomedicine, 65(2), 141–151. [6] Beal SL, Sheiner LB, Boeckmann AJ, Bauer RJ, eds. NONMEM 7.5 Users Guides. ICON plc; 1989–2020. https://nonmem.iconplc.com/nonmem750. [7] Bauer, R. J., Hooker, A. C., & Mentre, F. (2021). Tutorial for $DESIGN in NONMEM: Clinical trial evaluation and optimization. CPT: pharmacometrics & systems pharmacology, 10(12), 1452–1465.
Reference: PAGE 33 (2025) Abstr 11440 [www.page-meeting.org/?abstract=11440]
Poster: Drug/Disease Modelling - Paediatrics