Unit:
Postgraduate Programme Biostatistics & Health Science DataLibrary of the School of Health Sciences
Author:
Papachroni Ioanna
Supervisors info:
Αριστείδης Δοκουμετζίδης, Επ.Καθηγητής, Φαρμακευτική, ΕΚΠΑ
Νίκος Πανταζής, Ε.ΔΙ.Π. , Ιατρική, ΕΚΠΑ
Βασιλική-Αναστασία Σύψα, Επ.Καθηγήτρια, Ιατρική, ΕΚΠΑ
Original Title:
D-optimal design of a pediatric pharmacokinetic study of a fixed dose combination product of rifampicin-pyrazinamide-isoniazid for the treatment of tuberculosis
Translated title:
D-optimal design of a pediatric pharmacokinetic study of a fixed dose combination product of rifampicin-pyrazinamide-isoniazid for the treatment of tuberculosis
Summary:
The objective of this dissertation was to design a sparse sampling pediatric pharmacokinetic study for a fixed dose combination product of isoniazid (Η), rifampicin (R) and pyrazinamide (Z) for the treatment of tuberculosis. A single dose of FDC tablet was supposed to be given into two cohorts of fifty (50) children each. Non-linear mixed effects models were used to describe the structure of each drug model. We determined a unique optimal sampling schedule for the three drugs, such that the parameters of the PK models of each drug are estimated with high precision. We applied a method based on an expression for the Fisher Information Matrix (FIM) for non-linear mixed effects to improve the sampling design so as to obtain efficient parameter estimates. The approach is based on Rao-Cramer inequality which states that the inverse of FIM is the lower bound of the variance-covariance matrix of any unbiased estimators of the parameters. The criterion used for the optimization is D-optimality; a design is considered D-optimal if it maximizes the determinant of the Fisher information matrix. The particle swarm optimization (PSO) algorithm was applied for the optimization procedure while all implementation was conducted in MATLAB. The final design was evaluated by simulations and estimation with NONMEM. Bootstrap, Visual Predictive Check (VPC) plots and Goodness of fit plots were generated.
A pharmacokinetic study with 4 blood samples per subject was eventually designed. The optimal blood sampling times for the first cohort is0.10h, 0.13h, 0.55h and 4.28h and optimal blood sampling times for Cohort 2 is 0.57 h, 1.38h, 2.25h and 6h.The evaluation of the 3 pharmacokinetic models showed accurate results, as regards the parameter estimates.
Finally, if the clinical trial that we designed is implemented, it could be used for taking Marketing Authorization Approval of first-line paediatric fixed dose combination product for the treatment of tuberculosis in Europe and USΑ, which is currently unavailable.
Main subject category:
Health Sciences
Keywords:
D-optimal, Pharmacokinetic, Rifampicin, Pyrazinamide , Isoniazid
