Evaluate human PK prediction of compound LY based on allometric scaling with three species

Purpose To evaluate prediction of human pharmacokinetics (PK) for compound LY after the single dose PK study. Methods. Single dose intravenous (IV) and oral (PO) PK data from rats, dogs and monkeys were fit to a one‐compartment model using nonlinear mixed effect model (NONMEM). No species difference in metabolism or protein binding was needed for adjustment of CL. Absolute bioavailabilities were estimated and fixed at 70, 62 and 15% for rat, dog and monkey, respectively. Body Weight (BW) was incorporated into CL and V linearly with exponents either estimated or fixed to theoretical values of 0.75 for CL and 1 for V. A typical human assumed a 70 kg BW. The relationship between dose and AUC was simulated (Pharsight TS2). Predicted AUC vs. dose curves and associated 95% confidence interval were overlaid with observed data. Results. Exponents on BW were estimated to be 0.957 (SE=0.0388) for CL, and 1.02 (SE=0.0774) for V. When these exponents were fixed to theoretical values, human CL was underpredicted by 40% compared to CL predicted based on estimated exponents. Predicted AUC vs. dose curve based on estimated exponents and dog bioavailability encompassed observed. Conclusion. Allometric scaling using PK data from 3 animal species provide reasonable estimates of human PK parameters. Fixing exponents on body weight may not provide more accurate PK parameter estimates in human. Clinical Pharmacology & Therapeutics (2004) 75 , P30–P30; doi: 10.1016/j.clpt.2003.11.114