Table 2. Fit Criteria of Selected Models Fit to Our Data Using NONMEM

* Our best model to fit the data.

SBC = Schwarz-Bayesian Criterion; θ^{1}, …, θ^{7}= model population parameters fit by NONMEM; Ĉ^{i}= model predicted plasma TA concentration for sample i; C^{i}= observed plasma TA concentration for sample i;ϵ^{1i}, ϵ^{2i}= random variable error terms for sample i with means = 0 and SD ς^{1}and ς^{2}, respectively. GPE = geometric performance error. In our best model,* 90th GPE = 2.28 means that 90% of our model predictions fell within 1/2.28 and 2.28 times (*i.e.* , 44% and 228% of) the observed concentrations.

Add = additive; CV = coefficient of variation; CPB = cardiopulmonary bypass; wtkg = weight in kg.