Optimal design for a dexmedetomidine pharmacokinetic trial in mechanically ventilated neonates with single organ respiratory failure.
Sven C. van Dijkman (1), Vincenzo L. Di Iorio (1), Pieter A.J.G. De Cock (3,4), Hugo Robays (4), Peter De Paepe (3) and Oscar E. Della Pasqua (1,2)
(1) Division of Pharmacology, LACDR, Leiden University, The Netherlands; (2) GlaxoSmithKline, Clinical Pharmacology Modelling and Simulation, Stockley Park, UK (3) Heymans Institute of Pharmacology, Ghent University, Belgium (4) Pharmacy Department, Ghent University Hospital, Belgium
Objectives: The European legislation has created an increased demand for clinical trials in neonates and infants. However, paediatric protocols still rely on feasibility and empiricism rather than clinical pharmacology principles. The objective of this investigation was to illustrate how protocol design can be optimised in terms of patient numbers and sampling times to obtain well-estimated pharmacokinetic parameters, whilst ensuring minimal discomfort and limited blood withdrawal. A pilot study with dexmedetomidine has been selected to demonstrate the advantages of the approach. Data from the optimised pilot study will be used to support the design of an efficacy trial with 30 neonates.
Methods: Number of patients and sampling times were optimised with the ED optimality methodology in PopED using a pharmacokinetic model developed by Potts et al. (1). Simulated concentrations at the optimal sample times were used to re-estimate model parameters in NONMEM 7® according to an automated procedure in PsN. The parameters of interest were clearance, volume of distribution and inter-compartmental clearance. The concordance between original and re-estimated parameters was assessed by graphical and statistical diagnostic measures.
Results: Preliminary results show that model parameters and concentration curves over time (with 90%CI) can be accurately characterised with as few as 5 patients. SME and RE of parameter estimates showed acceptable values in all simulated scenarios. NONMEM minimisations were successful in 67-93 % of the cases.
Conclusions: Our analysis showed that as few as 5 blood samples from 6 patients are sufficient to allow accurate estimation of the parameters of interest. In contrast to current beliefs, the use of optimal design concepts is critical for the implementation of pharmacokinetic studies in neonates. Uncertainty and bias in parameter estimates due to empirical sampling may lead to under or overestimation of clearance and consequently yield inappropriate dosing recommendations.
 Potts AL et al. Dexmedetomidine pharmacokinetics in pediatric intensive care – a pooled analysis. Pediatrics Anesthesia 2009;19:1119-1129.