C.Chavanne(1), N.Frey (1)
(1) Pharma Research and Early Development, Clinical Pharmacology, Roche Innovation Center Basel, Switzerland, F. Hoffmann-La Roche Ltd.
Objectives: Conducting clinical trials in the pediatric population is generally a challenge, particularly when trying to ensure that enough patients are recruited to be able to explore the entire age range. Consequently, the collection of information to characterize the impact of the change in demographics and laboratory parameters with age on the drug PK properties, is generally incomplete. Population PK approaches could help by establishing covariate effect relationships to fill those gaps. However when simulations need to be conducted, the lack of covariate information for some age groups may become an issue. The objective of this work is to create a procedure that artificially increases the number of patients from an existing clinical database to get a more complete coverage of the age range. This work was done especially to support the evaluation of a new dosing algorithm for a drug X.
Methods: A covariate database was built with pediatric patients (age < 16 yrs) sourced from five Roche-sponsored and investigator-initiated trials and treated for the same pathology. The following covariates that influence the PK of drug X were of interest: age, gender, weight, height, CrCl. All sets of covariates from all trial visits were considered. In order to artificially increase the number of patients, each set of covariates at a given visit was considered as being that of a new patient. The Center for Disease Control (CDC) growth charts [1] were used as reference to evaluate if the coverage of the weight and height distribution over the entire age range was optimal. The population PK model of drug X was used to simulate steady-state exposure for the new dosing algorithm. Comparisons were then made between the databases both with and without the artificial increase.
Results: 293 patients were considered from the five clinical trials. Using each set of covariates as being a different patient, increases this number to 1473. The weight and height values were well distributed over the CDC charts. The simulated exposures with the largest dataset allowed for a better evaluation of the properties of the new dosing algorithm across the entire age range.
Conclusions: Considering each set of covariates at a given visit in pediatric clinical trials as being that of a different patient is a very effective procedure to artificially increase the range of covariates that change with age and then conduct simulations in a pediatric population to explore for example the evaluation of a new dosing algorithm.
References:
[1] Center for Disease Control (CDC) growth charts (http://www.cdc.gov/growthcharts/)
Reference: PAGE 24 (2015) Abstr 3364 [www.page-meeting.org/?abstract=3364]
Poster: Methodology - Estimation Methods