Gastric Emptying and the Personalized Management of Type 1 Diabetes

“The retention of stomach contents in a diabetic obviously may cause confusion as far as food intake and utilization are concerned.”

—Kassander (1)

It is increasingly advocated that management of type 1 and type 2 diabetes be “personalized” (i.e., targeted to the characteristics of the individual patient). This approach has major implications for health care costs, has proven successful in monogenic diabetes (2), and is practiced widely in other areas of medicine, such as oncology. Management strategies for type 1 diabetes currently focus on the achievement of targeted glycemic control, as assessed by a glycated hemoglobin (HbA1c) value <7% (or perhaps even lower) to diminish the development and progression of microvascular complications (i.e., retinopathy, neuropathy, and nephropathy) without increasing the risk of hypoglycemia (3). The longstanding strategy of normalizing fasting/preprandial blood glucose level, with lesser attention to postprandial blood glucose level, is now recognized as inappropriate. Indeed, this is to some extent self-evident, as in modern societies, many individuals eat three larger meals a day interspersed with snacks and accordingly spend the majority of the day in a postprandial state, with a period of true fasting of only a few hours before breakfast (4). Postprandial excursions make a major contribution to HbA1c (4). In type 2 diabetes treated with or without insulin, mathematical modeling indicates that the contribution of fasting/preprandial vs postprandial glycemia to HbA1c level varies substantially. Thus, fasting glucose level has a greater impact in patients with poor glycemic control (HbA1c >8%), whereas postprandial glycemic excursions predominate in patients with better-controlled values (HbA1c ≤8%) (4). A number of strategies have been developed to prevent major and/or rapid elevations in blood glucose level after eating. Carbohydrate “counting” is the most widely used method for determining mealtime insulin doses, in some cases employing complex algorithms. The potential effects of concomitantly ingested protein and fat on postprandial glycemia are usually not accounted for. There has been increased emphasis on technical advances, including the use of continuous glucose sensors and “closed-loop” insulin pumps; postprandial exercise has also been proposed as useful.

Although this is an area where blinded placebo-controlled studies may be impossible and/or impractical, the efficacy of all these approaches, which are for the main part empirical, appears modest at best. Somewhat surprisingly, the relevance of the rate of gastric emptying, which determines the delivery of carbohydrates to the small intestine for their subsequent absorption, has received little attention. This contrasts with the increasing recognition of this concept in the management of type 2 diabetes, as attested by the widespread use of the “short-acting” glucagon-like peptide 1 receptor agonists exenatide twice daily and lixisenatide alone or in combination with basal insulin, which markedly diminish postprandial glycemic excursions predominantly by slowing gastric emptying (5).

Gastric emptying is a complex, coordinated process that depends on the intricate interplay between the proximal and distal stomach, interstitial cells of Cajal (the gastric “pacemaker” cells), enteric and neurohumoral pathways, and numerous mediators including acetylcholine, vasoactive intestinal peptide, carbon monoxide, and nitric oxide (6). The rate of gastric emptying—as opposed to the time for its complete emptying, which is greater with higher caloric meals—is highly dependent on the composition (liquid and/or solid) and especially the macronutrient content (i.e., fat, protein, and carbohydrate) of the meal rather than the meal volume. Solids and nutrient liquids exhibit different patterns of emptying; liquids are emptied preferentially (about 80% of liquid is emptied before solid emptying commences) and in an overall “monoexponential” pattern for low-nutrient liquids, with a more linear pattern as nutrient content increases. Digestible solids, on the other hand, exhibit an initial “lag phase” (usually ∼20 to 30 minutes, during which solids are ground into small particles) before a linear emptying phase (6). It is not widely appreciated that gastric emptying of nutrients exhibits very wide interindividual variations in health (∼1 to 4 kcal/min), even more so in diabetes, which is associated with a high prevalence of abnormally delayed and, in some cases accelerated, gastric emptying. In contrast, intraindividual variations are relatively modest. In a 25-year longitudinal evaluation of 13 patients (12 with type 1 diabetes), we found that rates of emptying were remarkably stable (7). In other words, gastric emptying is highly variable between individuals but relatively reproducible over time in health and diabetes.

The rate of emptying of carbohydrates profoundly influences postprandial glycemia in a time-dependent fashion (8, 9). Postprandial glucose excursions are determined much more by the rate of carbohydrate delivery to the small intestine than by the total amount of carbohydrates in a meal. For example, we showed a direct relationship between glycemic excursion and gastric emptying of a 75-g glucose drink in healthy individuals as well as in those with impaired glucose tolerance and both type 1 and type 2 diabetes (10, 11). Modulating gastric emptying pharmacologically (e.g., accelerating emptying with the prokinetic drug erythromycin) increases the postprandial glycemic excursion in both type 1 (12) and type 2 diabetes (13). Our studies, in which intraduodenal infusions of glucose were administered to healthy individuals and patients with type 2 diabetes at rates spanning the physiological range of gastric emptying (i.e., 1 to 4 kcal/min), indicate that even relatively minor differences in small intestinal glucose delivery may have a substantial effect on glycemic response (14, 15). Hence, gastric emptying, even when normal, is central to the regulation of postprandial blood glucose level, and individuals with relatively more rapid gastric emptying will, in general, exhibit a higher initial glycemic response to ingested carbohydrates in both health and diabetes. It would not be surprising if individual gastric emptying rates have implications for the risk of developing type 2 diabetes.

Abnormal gastric retention in diabetes was first noted by Boas (16) in 1925. In a seminal paper published in 1958, Kassander (1) described six patients with insulin-treated diabetes (five with type 1 diabetes) who exhibited abnormal intragastric retention in the absence of gastrointestinal symptoms and coined the term gastroparesis diabeticorum. Diabetic gastroparesis, which may be defined as abnormally delayed gastric emptying in the absence of mechanical obstruction, occurs frequently; the outcomes of several cross-sectional studies indicate that 40% to 50% of patients with long-standing type 1 diabetes have gastroparesis, a prevalence comparable to that of microvascular complications (12, 17). For example, a recent follow-up of the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications cohort reported a prevalence of 47% (18). Diabetes is also associated with an increased prevalence of gastrointestinal symptoms that adversely affect quality of life (6). However, there is a poor correlation between upper gastrointestinal symptoms and rate of gastric emptying; abdominal bloating and early satiety are weak predictors of delayed emptying, whereas vomiting usually shows no association (6). Diabetic gastroparesis is also weakly associated with the presence of autonomic neuropathy, the latter usually assessed by standardized cardiovascular reflex tests.

In patients treated with insulin, delayed gastric emptying should intuitively increase the potential for mismatch between intestinal absorption of ingested carbohydrates and the action of exogenous prandial insulin and should influence the timing of the postprandial insulin requirement. For example, in a study involving 11 patients with type 1 diabetes, less insulin was required during the first 120 minutes after a meal in the five patients with gastroparesis and more between 180 and 240 minutes (19). Paradoxically, delayed gastric emptying in type 1 diabetes has been associated with an overall increase in blood glucose level during the day (12). There is preliminary evidence that in type 1 and insulin-treated type 2 diabetes, gastroparesis predisposes patients to postprandial hypoglycemia, so-called gastric hypoglycemia.

The relationship between gastric emptying and glycemia is bidirectional, so the glycemic environment affects gastric emptying acutely and reversibly. Acute hyperglycemia, even within the normal postprandial range, slows gastric emptying, whereas insulin-induced hypoglycemia accelerates it in health and type 1 diabetes (20). Although the magnitude of slowing induced by hyperglycemia remains contentious, it is appropriate that gastric emptying be measured under “euglycemic” conditions or, at a minimum, when blood glucose level is between 4 and 10 mmol/L at baseline (3). In contrast, the acceleration of gastric emptying by acute hypoglycemia is unequivocally substantial; however, given that it is likely to represent an important counter-regulatory response to enhance the absorption of carbohydrate, it has received inappropriately little attention (20).

In the study by Lupoli et al. (21), they report their evaluation of gastric emptying of a solid meal in a consecutive series of 42 tertiary center outpatients with longstanding type 1 diabetes [predominantly female, 80% with disease managed with continuous subcutaneous insulin infusion and with relatively good glycemic control (mean HbA1c value, 7.7%)] compared with 31 age-, sex- and body mass index‒matched healthy subjects. Despite the long duration of diabetes (mean, 19 years), patients did not have evidence of autonomic neuropathy. Consistent with other series, gastric emptying was slower than normal in about one-third of the diabetic cohort (perhaps higher than would be anticipated given the absence of autonomic neuropathy), and there was no relationship of upper gastrointestinal symptoms with gastric emptying. The novel and important observation relates to the demonstration of a substantial effect of gastric emptying on the magnitude and “shape” of the postprandial glycemic response, even after adjustment of preprandial insulin dosing on the basis of glycemic load; specifically in patients with delayed gastric emptying, the time-to-peak glucose was prolonged by an average of 27 minutes.

Limitations of the study should be appreciated. In participants with type 1 diabetes, postprandial glycemic excursions were in the hyperglycemic range which, as discussed, may slow gastric emptying. Conversely, because the caloric load (372 kcal) and carbohydrate content of the test meal were modest, the effect of gastric emptying on glycemia may have been underestimated. Gastric emptying was measured with a stable isotope breath test, which represents a noninvasive and acceptable alternative to the gold standard of scintigraphy. However, it remains to be established how precisely solid and/or liquid emptying can be measured with a breath test, particularly in individuals with markedly delayed gastric emptying. Stable isotope breath tests are certainly useful as a “point-of-care” measurement in the clinic. However, in the study by Lupoli et al. (21), there is no evidence that the octanoate label was bound to the solid meal because it was given in a capsule; it may have emptied preferentially with the water phase. The observations by Lupoli et al. (21) support the measurement of gastric emptying as part of a personalized approach in type 1 diabetes to guide individual prandial insulin dosing. Strategies focusing on a meal’s carbohydrate content exclusively, while ignoring the substantial variation between individuals in carbohydrate absorption, are inherently suboptimal.

Abbreviation:

Abbreviation:
 
• HbA1c

  glycated hemoglobin

Acknowledgments

Financial Support: C.S.M. is supported by a National Health and Medical Research Council Early Career Fellowship. T.W. is supported by a Royal Adelaide Hospital Florey Fellowship.

Author Contributions: All authors contributed to the drafting of and approved the final manuscript.

Disclosure Summary: The authors have nothing to disclose.

References