The fed state
The fed state is the absorptive period between meals. Carbohydrates are normally mixed with lipids and protein in the diet and make up 40 – 60% of the caloric content. Absorption of dietary carbohydrates is influenced by their chemical form (refined sugars or complex carbohydrates) and by other components of food. Furthermore, disposition of dietary carbohydrate is indirectly affected by fats and protein to the extent that these latter (i) compete with glucose as substrates, and (ii) interfere with glucoregulatory hormones by altering insulin secretion. To circumvent the difficulties of study of glucose ingestion, the regulation of glucose homeostasis during the fed state has classically been investigated with the use of intravenous glucose, which can be administered in formats more suitable for formal analysis.
Glucose can be injected intravenously as a single bolus (0.33 or 0.5 g [1.8 or 2.8 mmol] per kg of body weight) and the decline in plasma glucose concentration after the initial peak followed for 60 – 90 min (intravenous glucose tolerance test or IVGTT). Between 10 – 60 min, glycemia decreases approximately as a single exponential function of time, and a decay constant (k value) can be calculated to estimate tolerance to intravenous glucose. The IVGTT has several drawbacks. First, the time course of glucose fall in reality is a multi-exponential function of time, so that the decay constant takes on different values according to which segment of the curve is used for analysis. Second, over the same time interval, different curves can have similar k values; for this reason, the area under the glycemic excursion is sometimes used instead of the k value. However, the area under the curve is also influenced by the volume into which the injected glucose is distributed. Third, and most important, the shape of the glucose curve is heavily affected by the endogenous insulin response to the acute hyperglycemia caused by the intravenous bolus. Such response is highly irregular, and after an initial peak proceeds in two or three smaller spikes tightly synchronized with similar glycemic spikes .
In general, in the presence of an intact feedback loop between glucose and insulin, glucose tolerance is the integrated outcome of multiple changes in the glucose as well as in the insulin system: distribution of the exogenous glucose, stimulation of peripheral glucose uptake by insulin and hyperglycemia, suppression of hepatic glucose output, and secretion, distribution and degradation of insulin. It is therefore not surprising that the information provided by an IVGTT is somewhat ambiguous, and that the test itself is neither very sensitive nor easily reproducible. A modern version of the IVGTT is that which interprets the changes in glucose and peripheral insulin concentrations (measured at frequent intervals following the bolus) on the basis of a “minimal” mathematical model of the glucose and insulin systems and their interactions . The model generates a parameter reflecting the ability of hyperglycemia to stimulate insulin secretion, another parameter estimating the ability of insulin to stimulate glucose metabolism, and an index of the ability of glucose to promote its own disposal. Although attractive for its simplicity, the minimal model approach generally falls short of its ambitious goal, that is, to describe all aspects of glucose tolerance with a minimum of data. In particular, the minimal model does not work when the endogenous insulin response is scarce, for example in diabetic patients. The model is not minimal in that the data analysis requires a computer program that, like all package deals, deprives users of critical evaluation. Several updates of the technique have been devised. Labeled (tritiated or deuterated) glucose can be co-injected with cold glucose; appropriate model analysis of the tracer data makes it possible to dissect out the effects of insulin on the liver and on peripheral tissues . Because of its complexity of use and interpretation, the latter remains a purely investigative tool.
A secondary injection of exogenous insulin or tolbutamide has been used to circumvent the failure of the minimal model in case of insufficient endogenous insulin response . A simple way to estimate whole-body sensitivity to insulin is to paralyze endogenous insulin release with a constant infusion of somatostatin (at a rate of 0.3 – 0.5 mg h−1 ) while simultaneously infusing glucose (at a rate of 1.35 mmol min−1 m−2 ) and regular insulin (at a rate of 50 mU min−1 m−2 ). With this technique (also known as the pancreatic suppression test) , steady hyperinsulinemia (∼80 mU L−1) is associated with a level of hyperglycemia that is inversely proportional to the ability of whole-body tissues to increase their glucose utilization in response to insulin. This test suffers from the fact that somatostatin inhibits the release of other glycoactive hormones (e.g. glucagon). With this limitation, however, the test is simple and reliable enough for clinical use. The glucose clamp technique has become the reference method to study glucose metabolism . Figure 14.6 (top) exemplifies the euglycemic hyperinsulinemic version of the clamp technique. An exogenous infusion of regular insulin is started at time zero in a format comprising a prime followed by a constant infusion (usually at a rate of 1 mU min−1 kg−1 ); such infusion quickly establishes a hyperinsulinemic plateau of about 60–70mUL−1. A few minutes after starting the insulin infusion, an infusion of glucose is begun at a rate which is adjusted every 5–l0min on the basis of on-line plasma glucose measurements obtained with the same frequency. Over the second hour of a 2-h experiment, euglycemia in the face of constant hyperinsulinemia is maintained by an approximately constant glucose infusion, which in a healthy adult ranges between 20 and 50 μmol min−1 kg−1 (a mean value is shown in the figure). Such a rate equals the overall rate of glucose uptake (also called M for metabolism) in a subject in whom endogenous glucose production is nil. Relative insulin insensitivity or insulin resistance is indicated by a low M value at comparable levels of glycemia and insulinemia. The technique has the following advantages: (a) any preset combination of plasma glucose and insulin levels can be easily achieved; (b) the time course of insulin action can be determined with a time resolution of about 10 min; (c) hypoglycemia with its attendant counterregulatory hormonal response can be avoided; (d) other techniques, such as tracer glucose infusion, muscle biopsy, and indirect calorimetry, can be readily combined with a clamp protocol; (e) the interference of other hormones or substances with insulin action can be quantitated by co-infusing them during a clamp study; (f) high reproducibility.
Although computerized algorithms are available to run a clamp, manual operation with a minimum of experience does just as well. The major drawback of the euglycemic insulin clamp is the need to draw frequent blood samples from an arterialized vein (e.g. a heated wrist or hand vein). The hyperglycemic version of the glucose clamp (schematized in Figure 14.6, bottom) consists of acutely raising plasma glucose to any desired level, and then clamping it at that level with a variable infusion of exogenous glucose  (as in the euglycemic clamp version). The hyperglycemic step evokes an endogenous insulin response that typically is biphasic: an early output (of preformed hormone) that lasts 10–15min, followed by a gradual, continuous rise in insulin levels (I), reflecting glucose-induced triggering and potentiation of β-cell secretory activity. By analogy with the euglycemic insulin clamp counterpart, the hyperglycemic clamp provides an M value, which represents the combined effect of endogenous hyperinsulinemia plus hyperglycemia on whole-body glucose disposal. The M value can be expressed per unit of insulin to provide an index of insulin sensitivity (M/I) that agrees closely with that derived from the euglycemic insulin clamp .