- 1:10 PM

Use of Pharmacokinetic/pharmacodynamic (PK/PD) Modeling to Aid Formulation Selection

Larissa A. Wenning1, Eseng Lai2, Tami Crumley2, Tian-Quan Cai3, Fang Liu4, M. G. Waters3, Shaun Fitzpatrick5, and Julie Stone1. (1) Clinical PK/PD, Merck & Co, Inc, WP75B-100, 770 Sumneytown Pike, West Point, PA 19486, (2) Clinical Pharmacology, Merck & Co, Inc, (3) Cardiovascular Diseases, Merck & Co, Inc, (4) Biostatistics, Merck & Co, Inc, (5) Pharm R&D, Merck & Co, Inc

Background: Niacin (nicotinic acid) has beneficial effects on lipids and reduces cardiovascular events in patients with dyslipidemia. PK/PD modeling was applied to project the probability that 3 candidate extended release niacin formulations would have comparable efficacy on lipids compared to a reference formulation. For the purpose of this projection, the beneficial effects of niacin were assumed to directly correlate with reduction in plasma free fatty acid (FFA) concentration.

Methods: A single dose crossover study was conducted in 32 healthy subjects to compare the pharmacokinetic (PK) profiles of 3 test extended release niacin formulations to a reference formulation. An Emax model was used to relate niacin plasma concentrations to percent inhibition of lipolysis as follows: Effect = Emax*C/(C+EC50); where Effect represents percent reduction in plasma FFA levels, C is niacin plasma concentration, Emax was fixed to 100% based on in vitro data, and EC50 was fixed to 12.4 ng/mL, again based on in vitro data. The PK/PD model was applied to project whether the PK differences between the 3 test formulations vs. the reference formulation, if there were any, would have a clinically meaningful impact on lipid efficacy (changes in LDL and HDL cholesterol) of ER niacin, which was postulated to derive from reduction of plasma FFA level.

Results: All three test formulations had different niacin plasma PK profiles compared to each other and the reference formulation. Despite the differences in PK profiles, all three formulations were projected to have similar effects on plasma FFA levels compared to the reference formulation based on the above PK/PD model. Two of these formulations were carried forward into a subsequent clinical study which showed similar effects on LDL and HDL cholesterol for both formulations compared to the reference.

Conclusions: PK/PD modeling is a useful tool that can aid interpretation of PK differences between formulations and selection of optimal formulations.