Calculation of the 90% CI values from pop PK model estimates

From: "Xiao, Alan" alan_xiao@merck.com Subject: [NMusers] Calculation of the 90% CI values from pop PK model estimates Date: Thu, 15 Sep 2005 18:50:49 -0400 Hello, Could anyone share his/her experience/opinion on the following scenario? Assuming a linear PK model with clearance (CL) as a function of weight (WT), drug-drug interaction 1 (DDI1), etc: CL=theta1*{1+theta2*DDI1+theta3*[(WT/WT0)**theta4-1]}*exp(eta1) with parameter estimates as (Mean+/-SE): theta1=30+/-2; theta2=-0.3+/-0.1; theta3=0.8+/-0.3; theta4=2+/-0.8; and eta1=0.1+/-0.03. WT0 is the median value of the weight distribution of the whole population in the data file for model development. The population size is 1000, of which 10 subjects were on DDI1. The task is to calculate the geometric mean AUC ratio and its 90% CI values for the effect of weight: 140 vs 70 and for the effect of DDI1: DDI1 vs. non-DDI1. ANY comments/questions/suggestions are appreciated. Alan Alan Xiao, Ph.D. Clinical PK, MRL Merck & Co, Inc.