Paediatric drug development: are population models predictive of pharmacokinetics across paediatric populations?

摘要: WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT • In paediatric development, initial estimation of the dose is obtained by extrapolation. This usually performed using dosing regimen in another population as reference. However, no consensus on selection methodologies has been reached so far. • The paradigm compound identified to illustrate these concepts a nucleoside reverse transcriptase inhibitor (NRTI) used treat HIV infection. The recommended abacavir 8 mg kg−1 twice daily up maximum 300 mg daily. WHAT STUDY ADDS • Our findings show that use model may not suffice predict parameter distributions and drug exposure across populations. • Estimation covariate effects critical, but sufficient extrapolate pharmacokinetics from reference population. • Covariate-parameter correlations remain constant beyond range observations. Exponential relationships allometry do correct for discrepancies. AIMS To assess predictive value model-based approach populations early clinical development. METHODS Abacavir was selected data wide age range. (PK) children were analysed separately infants toddlers. Two independent models obtained, systemic (AUC) then simulated based estimates each model. Drug exposures toddlers predicted pharmacokinetic children, other way around. RESULTS (a two-compartment PK one compartment children) accurately described which they built. neither different population: infants, median AUC (95%-CI) estimated at 7.03 (6.72, 7.48) µg ml−1 h, whilst it 5.75 (4.82, 6.26) µg ml−1 h; 6.96 (5.85, 7.91) 6.45 (5.80, 7.01) µg ml−1 h. CONCLUSIONS These suggest assumption an identical (linear or nonlinear) correlation between parameters demographic factors hold true groups. Whilst modelling enables accurate characterization properties, extrapolations such have limited due differences impact developmental growth populations.