A proposed novel modeling of an oral methylphenidate formulation for attention-deficit/hyperactivity disorder (ADHD) goes beyond the traditional relating of in-vitro dissolution and drug release with in-vivo serum drug levels, to specify the amount and time of drug release which can best evoke clinical benefit. The proposed model-based approach was published in Clinical Pharmacology & Therapeutics.

Roberto Gomeni, President of Pharmacometrica, La Fouillade, France and colleagues describe the application of a "convolution-based" model to link in-vitro/in-vivo measures to hour-by-hour behavioral ratings of ADHD symptoms.

"The objective of the paper was to present how a (computer based) modeling methodology linking drug exposure and clinical response can be used to identify the optimal drug-delivery mechanism appropriate for maximizing a clinical benefit of a treatment for ADHD," said Gomeni.

Gomeni and colleagues undertook the additional step of linking pharmacokinetics to hourly ratings with an objective, observational rating scale of ADHD symptoms, the Swanson, Kotkin, Agler, M-Flynn and Pelham Scale (SKAMP).

The convolution-based approach fo in-vivo/in-vitro correlation modeling was described 2 decades ago by William Gillespie, FDA Office of Clinical Pharmacology and Biopharmaceuticals.

Although a clinical trial to test the predictions is planned, this first investigation used the population parameters established for Concerta (Janssen) 54mg as a reference, and compared clinical benefit in ADHD against "area under the curve" (AUC) of drug in serum over time from 0–12 hours and 0–24 hours after administration.

The modeling results suggest that optimal time for the in-vivo release of the first fraction of the dose is sooner, and the time for delivering the second fraction is longer than what this formulation provides.  The rates of release were lower in the optimized scenario than what occurs. The fractions of the dose released in the first and second steps, however, are the same as predicted for optimal benefit.

The investigators also analyzed the necessary design of the clinical trial to test a product formulated with the new model, projecting that 20 subjects and 7 to 8 pharmacokinetic and SKAMP measures for each subject would be sufficient to provide results with a precision variance of 20%.

"The selection of optimal sampling times is critical when conducting pediatric population pharmacokinetic/pharmacodynamic studies," the investigators indicated. "Attention to the choice of sampling strategies can reduce the cost of the trial and burden imposed on the study's participants."

Gomeni commented on how such a trial, incorporating measures of clinical effect from an instrument like the SKAMP, would better serve the testing of a new extended-release methylphenidate formulation for ADHD than what is now required of New Drug Application (NDA) sponsors. 

"The pharma companies propose novel formulations and try to demonstrate that the clinical response is better than the placebo response. The trick is that a clinical response can be better than placebo but not sufficient to match with the physician and the parents' expectations," he said.