Predicting paediatric PK in order to investigate it
Andreas Groth, Christoffer Tornøe
Quantitative Clinical Pharmacology, Novo Nordisk, Soeborg, Denmark
Objectives: Scaling of PK parameters for a given drug from adults to children is typically done using standard allometric scaling exponents (0.75 and 1 for CL and V, respectively). However, if the drug bears similarities in chemical structure and likely clearance mechanisms to another drug with both adult and paediatric PK data available, assuming an adult-paediatric PK parameter relationship similar to that of the well-known drug may be a better prediction for the new drug than using standard scaling parameters.
The objective of the present work is to investigate the use of prior PK data to predict paediatric PK in order to improve the quality of paediatric PK assessments and hence paediatric dose and therapeutic results [1].
Methods: The population PK analyses were performed in NONMEM VII using adult and paediatric PK data with 7 samples per subject for the 1st generation drug and adult PK data with 14 samples per subject from 2nd generation drug.
The sample size N required to achieve a 95% confidence interval within 60-140% of the point estimate of the population value is calculated as N = (CV/log(1.4)/t(0.975,N-1))^2, where CV is the coefficient of variation of the relevant PK parameter to be estimated.
Results: The estimated allometric exponents for CL and V from the combination of adult and paediatric PK data of the 1st generation drug were 0.577 (95% CI: 0.37-0.79) and 0.77 (95% CI: 0.55-0.99), respectively. These estimated exponents were subsequently used for the estimation of the 2nd generation drug's PK parameters, leading to estimates for unexplained between subject variability of 12 and 10 CV% for CL and V, respectively.
Assuming rich PK sampling in children (7 samples per child), the estimated paediatric sample size required to achieve 95% confidence intervals within 60-140% of the point estimate for the population typical values of CL and V with the 2nd generation drug was 3 children in each age group defined as 2-<6 and 6-<12 years.
A proposed sparser sampling schedule (4 samples per child) focusing on the early part of the plasma profile was predicted to increase the required recruitment to 4 children per age group.
Conclusions: Using prior adult and paediatric PK data, a paediatric PK study of a 2nd generation drug was designed to get adequate PK information with feasible sample size and sampling. The proposed method of designing paediatric PK studies ensures informative paediatric data is acquired to inform dose selection and labelling.
References:
[1] M Lala et al., Implementation of a Pharmacokinetic Quality Standard to Improve Pediatric Trial Design, ACoP 2011
