Allometric scaling

From: "Eyas Abu-Raddad" <RADDADE@rx.uga.edu> Subject: Allometric scaling Date: Fri, 30 Jun 2000 00:01:45 -0500 We are trying to scale up PK parameters across species. We have limited data from 3 species (mouse: one observation per individual, rat:one observation per individual, and dog:multiple observations per individual) and would like to extrapolate PK parameters to humans, using allometric models with weight as the covariate, for e.g.: TVCL=THETA(1)*WT**THETA(2) TVV=THETA(3)*WT**THETA(3) My question is, should we pool the data as if all individuals came from one population? Or should we use species-specific error terms? Data might limit adding this many error terms, will the estimated theta's still be of use? Eyas AbuRaddad Department of Pharmaceutical and Biomedical Sciences University of Georgia