Elucidation of the optimal dosing scheme of an antiviral drug in children
Michael Looby, Guenther Kaiser, Beatrice Abrams and Bill Sallas
Novartis
Objectives: Elucidate the optimal dosing scheme of Famvir in a pediatric population between 1 and 12 years
Methods: Very limited prior data pharmacokinetic data of famciclovir/penciclovir was available in children down to 6 years. Using a mixed effects analysis, this data was combined with available adult data to develop a population pharmacokinetic model which predicted the pharmacokinetics of penciclovir in both adults and children down to the age of 1 year. Using this model, studies were designed to assess the pharmacokinetics and safety of Famvir in pediatric patients with Varicella Zoster and Herpes Simplex virus infections. These studies were divided into two Parts A & B. In Part A, single doses of Famvir were administered to fully characterize the pharmacokinetics of pencicolovir and the relationship between systemic exposure and body-weight or age. The criterion for dose selection was to achieve similar penciclovir exposure in the pediatric- as in adult- population. The information collected in Part A was subsequently used to refine the Famvir doses for the 7-day multiple-dose safety Part B. The FDA endorsed the suggested approach.
Study design (Part A – Single dose pharmacokinetics): The model based analysis of the prior data determined the following characteristics as being optimal:
- Dose: 12.5 mg/kg if weight < 40 kg, otherwise 500 mg (adult dose).
- N = 52 (12 in the range 1 to <2 y, 24 in the range 2 to <6 y, 16 in the range 6 to 12 y)
- Up to 6 blood samples per child
A subsequent model based analysis of this data was used for ultimate dose selection.
Results: After an interim analysis in Part A, recruitment into this part could be terminated early due to success of prior predictions. Target 52; attained 41(at interim analysis) A model independent comparison of the exposure metrics of the pediatric population with the adult exposure envelop confirmed that the dosage algorithm (12.5 mg/kg) was successful in attaining exposure that lay within the adult exposure envelope. This correspondence provided an external validation of the model based predictions Nevertheless a trend was apparent in the pediatric data with exposure decreasing with age. This was expected given the linear scaling factor.
Using the newly generated data, the model was updated to simulate the dosing scheme that predicted penciclovir exposure across the pediatric population (age: 1 to 12 years, body-weight: 9 to > 40 kg) corresponding best to adults. Given the model, it is a straightforward task to test a whole range of possible dosing adjustment rules in silico. Finally, an 8-step dosing scheme with 50 mg dose steps was selected as the best compromise between simplicity of dosing and accuracy of target exposure. This optimized dosing scheme was then used in the 7-day multiple dose safety Part B of the pediatric studies.
Conclusions:
- An innovative model based approach was used to design and analyse a pediatric study. Prior data was used to select a dosing scheme used in the initial part of the study.
- Confidence generated in the ability of this dosing scheme to attain adult exposures lead to the early termination of enrollement in the pharmacokinetic part of the trial. This was endorsed by FDA.
- With the new data in hand, it was possible to simulate a wide range of dosing schemes to determine the optimal dosing scheme.
- The final 8-step dosing scheme, endorsed by the FDA, was much simpler than the scheme used in the original study.