Pharmacogenetics Into the Model

Despite the improvement in goodness of fit when significant covariates are incorporated into a pharmacokinetic model, in reality, the addition of most covariates reduce the unexplained variability in a model by very little. There are very few silver bullet covariates where their addition to the model has such an effect that residual 

To date, genotyping has been used solely to explain the between-subject variability in clearance with the genotype treated as any other covariate. Kvist et al. (2001) first studied the role CYP 2D6 genotype plays in the clearance of nortriptyline in 20 subjects with depression and 20 healthy volunteers. CYP 2D6 genotype can be classified into four groups based on the number of functional genes  

2 Wild-type genes are always designated as the 1 allele and have normal metabolic function. 

Kirchheiner et al. (2002) used another approach to model the clearance of glyburide, an oral hypoglycemic that is metabolized by CYP 2C9, in 21 healthy volunteers. At the time of the study, the genotype for CYP 2C9 consisted of two inherited functional polymorphisms of three different alleles (denoted 1, 2, and 3) that were known to affect the catalytic efficiency of the CYP 2C9 enzyme. Hence, humans were classified as 1/ 1 (wild type), l/ 2, l/ 3, 2/ l, 21 2, or 21 2. They could have used a categorical model like Mamiya et al. (2000) but instead chose to model clearance a little differently. The final model for glyburide clearance (CL) was modeled as the sum of the partial clearances 

While genotyping may seem useful, the models that are developed may only be applicable until a new nomenclature is used or another defective allele is identified. For example, in 2001, when Kirchheiner et al. 

---