Model
Two models were included in the study(16, 17). Both studies used a fluorescence polarization immunoassay method to determined tobramycin concentrations. Both models described data with 2 structural compartments. Parameters in the model developed by Alganhem et al. (model 1) (16) were Cl=4.65 L/h, V1=13.3 L, Q=0.452 L/h and V2=6.62 L. Height and creatinine clearance (Clcr) were set as covariates on Cl and height was set as covariate on V1. Between Subject Variability (BSV) were estimated only on Cl and V1. In the model of L. Crass et al. (model 2) (17), parameters used were Cl=4.90 L/h, V1=15.4L, Q=0.34 L/h and V2=5.56 L. Height and Clcr were set as covariates on Cl and height was set as covariate on V1. BSV were estimated on all parameters. For each model, inclusion of between occasion variability (BOV) on clearance achieved a further improvement of fit.