Modeling Gastric Emptying and Intestinal Glucose Absorption

From an Oral Dose to Systemic Glucose Appearance

An oral glucose challenge is not an instantaneous input to the circulation. Carbohydrate must leave the stomach, move through the intestine, and be absorbed before it contributes to systemic glucose appearance. The time course of that input can therefore reflect several processes, not just glucose regulation after absorption. This tutorial builds a mass-balanced representation of those processes. It complements the mechanistic perspective of oral glucose tolerance tests by specifying how the gastrointestinal input can be represented.

The model tracks amounts of digestible carbohydrate expressed as glucose equivalents, in grams. It assumes the dose is available for the modeled gastric and intestinal processes; it does not explicitly represent digestion of complex food, meal structure, or glucose metabolism after systemic appearance. Those processes require additional states and assumptions.

Gastric Emptying With a Variable Rate

Let \(G(t)\) be the amount remaining in the stomach, in grams, and let \(k_g(t)\) be the gastric emptying rate constant, in reciprocal hours. A first-order emptying model assumes that the instantaneous emptying flux is proportional to the amount currently in the stomach:

The rate constant may be constant, piecewise constant, or specified as a nonnegative time-varying function. If it varies with time but is specified independently of \(G\), integration gives the remaining amount and cumulative amount emptied:

Here \(D_g(t)\) is the cumulative amount transferred out of the stomach. This solution makes an important distinction: a changing emptying rate changes the accumulated hazard of emptying, not the definition of the instantaneous flux. If \(k_g\) depends on meal or physiological states, it must be specified as part of the model and the coupled equations may require numerical solution.

A variable rate can also be represented by a constant rate with the same cumulative emptying over a specified interval, but that equivalence applies only to the total emptied by the interval’s end. It does not imply the same emptying profile within the interval. For predictions at intermediate times, the time pattern of \(k_g(t)\) matters.

Representing Intestinal Transit and Absorption

A single intestinal compartment treats all material entering the intestine as immediately available for absorption. A transit chain can represent delay more realistically. Let \(I_j(t)\) be the amount in intestinal transit compartment \(j\), for \(j=1,\ldots,N\). Material enters the first compartment from the stomach and moves forward at rate \(k_{\mathrm{tr}}I_j\). For simplicity, take \(k_{\mathrm{tr}}\) as constant and use an absorption flux from each compartment:

where \(a_j\) is the absorption rate from compartment \(j\). Material leaving the final compartment at rate \(k_{\mathrm{tr}}I_N\) is treated as unabsorbed intestinal loss. This is a simplified representation of transit, not a claim that the gut consists of discrete anatomical compartments. With no absorption or other loss, the chain has an average transit time of \(N/k_{\mathrm{tr}}\): each of the \(N\) sequential compartments contributes an average residence time \(1/k_{\mathrm{tr}}\).

A common saturable uptake function for each compartment is:

Here \(V_{\max,j}\) is the maximum absorption rate, in grams per hour, and \(K_{m,j}\) is the amount at which the compartment’s absorption rate reaches half of \(V_{\max,j}\), in grams. When \(I_j\) is small relative to \(K_{m,j}\), absorption is approximately linear in amount; when \(I_j\) is large relative to \(K_{m,j}\), the rate approaches \(V_{\max,j}\). The systemic appearance rate is then

where \(F_{\mathrm{sys}}\), between zero and one, is the fraction of absorbed oral glucose-equivalent material that reaches systemic circulation in this simplified model. It is a lumped availability parameter, not a model of the processes that determine pre-systemic handling.

Mass Balance, Assumptions, and Model Checks

The model’s bookkeeping provides a direct correctness check. For the gastric state and \(N\) intestinal states, the total modeled amount is \(G+\sum_j I_j\). Adding the differential equations cancels the internal stomach-to-intestine and compartment-to-compartment transfers:

Thus modeled gut content decreases only through absorption and unabsorbed loss from the final compartment. With \(F_{\mathrm{sys}}<1\), systemic appearance is \(F_{\mathrm{sys}}\sum_j a_j\), not the entire loss from gut content. This distinction prevents a common accounting error: treating the systemic appearance rate as if all material absorbed across the intestinal wall necessarily enters systemic circulation.

The equations assume nonnegative initial amounts, nonnegative rate constants, positive \(K_{m,j}\), and no unmodeled input or recycling. Under those conditions the fluxes are nonnegative, and the model does not remove more material from a compartment than is available in the continuous-time equations. Numerical solutions should preserve nonnegative states to within solver tolerance; substantial negative amounts usually indicate a coding, units, or solver-step problem.

Choosing and Diagnosing the Model

A simpler alternative is a single intestinal pool with first-order absorption, \(a=k_a I\). It is parsimonious and may be adequate when data do not resolve transit delay or saturation. A transit chain adds delay structure; saturable uptake adds a rate ceiling. These additions can improve mechanistic plausibility, but they also add parameters. They should be supported by data or prior physiological knowledge rather than included automatically.

Useful diagnostics include plotting predicted stomach and intestinal amounts, emptying flux, total absorption, and \(R_a(t)\), not just the final glucose concentration. Check that the input is delayed relative to ingestion when transit is represented, that absorption does not exceed the specified capacity, and that cumulative absorption plus unabsorbed loss never exceeds the ingested amount. Compare predictions across plausible values of \(N\), \(k_{\mathrm{tr}}\), and the absorption parameters. If a fitted model requires negative rate constants, implausibly large capacity, or substantial negative states, its structure or implementation needs review.

A frequent mistake is to infer gastric emptying directly from the shape of a blood-glucose curve. That curve is downstream of emptying, transit, absorption, and glucose handling; a slow rise could have more than one explanation. The correction is to treat the gastrointestinal model as an explicit input-generating component and assess it jointly with the downstream model, using independent information where available. Even then, parameters may be confounded: a longer transit delay can resemble slower emptying, while a lower \(V_{\max}\) can resemble lower effective availability over a limited dose range.