Clinical Pharmacokinetics

, Volume 45, Issue 9, pp 931–956

Prediction of the Clearance of Eleven Drugs and Associated Variability in Neonates, Infants and Children

  • Trevor N. Johnson
  • Amin Rostami-Hodjegan
  • Geoffrey T. Tucker
Original Research Article

DOI: 10.2165/00003088-200645090-00005

Cite this article as:
Johnson, T.N., Rostami-Hodjegan, A. & Tucker, G.T. Clin Pharmacokinet (2006) 45: 931. doi:10.2165/00003088-200645090-00005

Abstract

Background

Prediction of the exposure of neonates, infants and children to xenobiotics is likely to be more successful using physiologically based pharmacokinetic models than simplistic allometric scaling, particularly in younger children. However, such models require comprehensive information on the ontogeny of anatomical, physiological and biochemical variables; data that are not available from single sources.

The Simcyp® software integrates demographic, genetic, physiological and pathological information on adults with in vitro data on human drug metabolism and transport to predict population distributions of drug clearance (CL) and the extent of metabolic drug-drug interactions. The algorithms have now been extended to predict clearance and its variability in paediatric populations by incorporating information on developmental physiology and the ontogeny of specific cytochrome P450s.

Methods

Values of the clearance (median and variability) of 11 drugs (midazolam [oral and intravenous], caffeine, carbamazepine, cisapride, theophylline, diclofenac, omeprazole, S-warfarin, phenytoin, gentamicin and vancomycin) were predicted for 2000 virtual subjects (birth to 18 years). In vitro enzyme pharmacokinetic parameters (maximum rate of metabolism [Vmax] and Michaelis-Menten constant [Km]) and in vivo clearance data were obtained from the literature.

Results

In neonates 70% (7/10) of predicted median clearance values were within 2-fold of the observed values. Corresponding results for infants, children and adolescents were 100% (9/9), 89% (17/19) and 94% (17/18), respectively. Predicted variability (95% confidence interval) was within 2-fold of the observed values in 70% (7/10), 67% (6/9), 63% (12/19) and 55% (10/18) of cases, respectively. The accuracy of the physiologically based model incorporated in the Simcyp® software was superior to that of simple allometry, especially in children <2 years old.

Conclusion

The in silico prediction of pharmacokinetic behaviour in paediatric patients is not intended to replace clinical studies. However, it provides a valuable aid to decision-making with regard to first-time dosing in children and study design. The clinical study then becomes ‘confirmatory’ rather than ‘exploratory’.

Copyright information

© Adis Data Information BV 2006

Authors and Affiliations

  • Trevor N. Johnson
    • 1
  • Amin Rostami-Hodjegan
    • 1
    • 2
  • Geoffrey T. Tucker
    • 1
    • 2
  1. 1.Simcyp LimitedBlades Enterprise CentreSheffieldUK
  2. 2.Academic Unit of Clinical Pharmacology, Pharmacokinetics and Pharmacogenetics GroupUniversity of Sheffield, Royal Hallamshire HospitalSheffieldUK