Suspicious CL/F vs. CLCR and SGOT values

From: "HUTMACHER, MATTHEW [Non-Pharmacia/1825]" <matthew.hutmacher@pharmacia.com> Subject: Suspicious CL/F vs. CLCR and SGOT values Date: Tue, 1 May 2001 14:47:00 -0500 Hello everyone, I am building a population PK model for a drug based on sparse sampling in an outpatient setting. I am fitting a one-compartment model with first-order absorption with a lag time and elimination to known 2-compartment kinetics in healthy volunteers. A simulation study was used to design the study, and the data appear to be well behaved with reference to the data that was simulated. The simulation study demonstrated that only 5-7% bias is expected in the CL/F population mean parameter, so that inference on CL/F is most likely trustworthy. I am getting some curious results, however, with regards to some covariates. I am finding that increases in calculated creatinine clearance correlate with decreases in CL/F and that increases with SGOT (liver) correlate with increased CL/F. The drug is mainly metabolized by the liver. I need to do a randomization test to see if these results are spurious, but if not, does anyone have any experience or explanation for this? Thanks for your time in advance. Matt