Optimizing Pediatric Doses for Weight-Based Dosing Regimens Under Constraints

Objectives: Determining pediatric doses via exposure matching for weight-based dosing regimens involves selection of both suitable doses and relevant weight ranges to confirm its effect on systemic exposures. The doses and weight cutoffs have been elegantly optimized for the narrow therapeutic index tuberculosis combination therapy using a utility function approach that minimized deviations from the target adult exposures [ 1, 2]. Often it is desired to enforce constraints for the dose selection, for example, due to minimally acceptable exposure limits. In this work, we demonstrate how to add such constraints, and the benefit of this approach. Methods: Doses and/or weight cutoffs are optimized by simulating the steady-state area under the curve over the dosing interval (AUC) and penalizing exposure deviations from the target exposure. The World Health Organization growth charts are used to determine appropriate weight-for-age for models including maturation effects. Simulated exposures are maintained within a desired range by using soft constraints, in which deviations for the population average AUC falling outside of the desired range are penalized and added to the sum of squared errors for deviations. The optimization is implemented in R using the Nelder-Mead optimization algorithm in the nloptr package. Results: We illustrate the approach using two motivating examples. First, using a hypothetical drug with one-compartment pharmacokinetics, we demonstrate how weight cutoffs can be optimized with either pre-specified or optimized doses while enforcing that the average population concentration remains above a minimal therapeutic concentration. Second, we apply the methodology to evaluate the optimality of weight cutoffs selected for Triumeq, a once-daily fixed-dose combination containing abacavir, dolutegravir, and lamivudine for which pediatric doses and weight cutoffs in children with human immunodeficiency virus weighing ≥ 6 kg were recently reported [ 3]. Constraints are used to maintain minimal AUC values for each analyte required for efficacy. Via optimization, we show that alternative weight cutoffs than those proposed can more closely match exposures shown to be safe and efficacious in adults without falling below the minimally acceptable AUC values. Conclusions: By formulating pediatric dose selection for weight-based regimens selected via exposure matching as an optimization problem, we demonstrate that pediatric exposures can be more precisely matched to adult exposures even in the presence of constraints. By implementing the optimization in R , these constraints can readily be enforced at the population level rather than on the individual-subject level. Sponsored by Vertex Pharmaceuticals Incorporated