Glucose turnover in presence of changing glucose concentrations : error analysis for glucose disappearance

The present studies were undertaken to determine whether 1) the cold- and hot-GINF techniques used with Steele's model provide equivalent estimates of the rates of glucose appearance (R(a)) and disappearance (R(d)) in the presence of physiological changes in glucose and insulin concentrations, 2) the conditions for the best estimation of R(a) are the same as those for R(d), 3) the magnitude of error (if present) differs in diabetic and nondiabetic subjects, and 4) situations exist in which the knowledge of R(d) allows inferences to be made on whole body glucose uptake. To do so we performed experiments in non-insulin-dependent diabetes mellitus and nondiabetic subjects using simultaneous infusions of [6-3H]glucose and [6-14C]glucose; glucose and insulin were infused to mimic normal postprandial glucose and insulin profiles; the infused glucose contained [6-14C]glucose but not [6-3H]glucose. Compared with the hot-GINF method, the traditional cold-GINF method underestimated (P < 0.05) R(a) and R(d) by 10-15% and hepatic glucose release by 25-50% during the 1st h of the study, with the magnitude of error being the same in both diabetic and nondiabetic subjects. Error analysis demonstrated that errors in R(a) and R(d) have different analytic expressions containing common structural but different volume errors. Both R(a) and R(d) can be accurately measured in diabetic and nondiabetic subjects if glucose specific activity is kept constant and the volume of the accessible pool is used to calculate glucose disappearance. The relationship between R(d) and whole body glucose uptake was also derived. Although R(d) can be determined by relying on measurements in the accessible pool only, the assessment of whole body glucose uptake requires a model of the nonaccessible portion of the glucose system. However, knowledge of R(d) can provide useful insights into the behavior of whole body glucose uptake.