Structural components of the final model to characterize interactions among gastric emptying, glucose absorption and glycemic control in nondiabetic obese and lean healthy adults. The blue bubble corresponds to observed insulin kinetics, gray bubble to observed gastric emptying profiles, and pink bubbles to the population effect of obese vs. lean. CMT: compartment; Ka, glucose absorption rate; Gast, observed dynamic gastric emptying rates; EXPOGast, effect of gastric emptying rate on Ka. A, amount of glucose after ingestion; Kprod, glucose endogenous production rate; Ins: observed dynamic insulin data; Gluc, glucose concentration; V, volume of distribution; GL0, baseline glucose concentration; CL, insulin-independent glucose clearance; CLIns, insulin-dependent glucose clearance; InsE, insulin concentration in the effect compartment; Ins50, insulin concentration from the effect compartment associated with 50% of CLIns/V; H, Hill coefficient determining the steepness of the insulin-clearance relationship; K1E, equilibration rate constant of the insulin effect compartment.
Individual predictions of glucose absorption rates (Ka) vs. time (A) and individual predictions of glucose absorption rates vs. gastric emptying rates (B) in lean healthy and nondiabetic obese individuals for the different glucose doses. Each curve and color corresponds to one individual. The black curves correspond to smooth lines of all individuals for each category. OGTT, oral glucose tolerance test.
Evaluation of the final model by comparing observed and predicted (A) or simulated (B) glucose values. A: glucose observations are plotted against individual glucose predictions; the red line corresponds to the identity line and the blue line is a linear regression of all points. B: glucose values are plotted against time. Red and blue areas correspond to the simulated confidence interval (95%) of the median, 10th, and 90th percentiles. The red curves are the observed median, 10th, and 90th percentiles (dashed lines). Blue dots correspond to the observed glucose values.
Independence of occasions was assumed and each “individual + glucose dose occasion” was considered as one study individual. Data are presented as median [minimum-maximum] or number of studied individuals (%). GIP, glucose-dependent insulinotropic peptide; GLP-1, glucagon-like peptide-1; HOMA-IR, index for homeostasis model assessment of insulin resistance; NA, not available.
↵* Gastric emptying for lean healthy adults receiving 75 g of glucose has been replaced by the median gastric emptying profile from two previous trials performed in the same conditions (12, 42).
Parameter estimates of the final model
RSE Estimate, %
RSE IIV, %
Glucose absorption rate (TVKa), min−1
Gastric emptying effect on (EXPOGast), dimensionless
Insulin-independent glucose clearance (CL), l/min
Volume of distribution (V), liter
Baseline glucose in lean individuals (GL0Lean), mmol/l
Baseline glucose in obese individuals, (GL0Obese), mmol/l
Insulin-dependent clearance in lean individuals (), l/min
Insulin-dependent clearance in obese individuals (), l/min
Insulin efficacy in lean individuals (), μU
Insulin efficacy in obese individuals (), μU
Sigmoidal Hill coefficient (H), dimensionless
Equilibration rate constant of insulin effect compartment (K1E), min−1
Proportional residual error in lean individuals
Proportional residual error in obese individuals
CV, coefficient of variation; FIX, fixed parameter; IIV, interindividual variability; RSE, relative standard error; Ins50, insulin concentration from the effect compartment associated with 50% of CLIns/V.