Minimal models in meal-like protocols: Simulation studies to assess precision and physiological plausibility of parameter estimates

In the present study we evaluated the impact of different experimental protocols on glucose minimal models identification. We used a realistic comprehensive nonlinear model of the glucose-insulin system to generate unlabeled and labeled glucose data during an IVGTT and a meal-like study. Subsequently, Monte Carlo simulations were used to compare the values and precision of the estimates of the minimal models indices obtained from the two experimental protocols. Our results demonstrate that both the values and the precision of the minimal model parameters are influenced by the experimental protocol. In particular, we found that the minimal model parameters estimated from the meal-like study are less precise but probably more physiologically plausible than those obtained from the IVGTT. The lower precision achieved with the meal-like study protocol is probably due to both the relatively unfrequent sampling schedule and the non-optimality of the glucose and insulin profiles for minimal model analysis.