Gastric emptying of a glucose drink is predictive of the glycaemic response to oral glucose and mixed meals, but unrelated to antecedent glycaemic control, in type 2 diabetes

Gastric emptying (GE) is a major determinant of postprandial glycaemia in both health and type 2 diabetes (T2D), with modest differences in GE rate exerting substantial effects on postprandial glucose profiles. Interventions that slow GE, such as nutrient preloads and GLP-1 receptor agonists, attenuate glycaemic excursions, while acceleration of GE increases postprandial glycaemia. Although GE shows wide inter-individual variability and reasonable intra-individual reproducibility, it is unclear whether GE assessed using a standardised carbohydrate liquid (e.g., a 75 g glucose drink) predicts the glycaemic response to physiological mixed meals. It is also uncertain whether antecedent hyperglycaemia influences GE in T2D. This study aimed to evaluate: (i) whether the glycaemic responses to a glucose drink and to standardised mixed meals correlate with GE (T50) of a 75 g glucose drink, and (ii) whether GE of the glucose drink relates to markers of short-, medium-, and long-term glycaemic control in newly diagnosed, treatment-naive Chinese adults with T2D.

Prior literature establishes that GE typically proceeds at a relatively constant caloric rate (about 1–4 kcal/min) in health, but is frequently disordered in T2D, showing both delays and accelerations with wide inter-individual variation. Slowing GE through dietary strategies (e.g., protein or whey/guar preloads) and pharmacological agents (e.g., GLP-1 receptor agonists) reduces postprandial glycaemia, whereas prokinetics like erythromycin can accelerate GE and increase glycaemic excursions. The baseline rate of GE can predict the magnitude of postprandial glucose lowering with therapies such as GLP-1RAs and DPP-4 inhibitors, with greater benefit when GE is faster at baseline. While GE is relatively reproducible within an individual using the same test meal, it had remained unclear whether GE measured with a 75 g glucose drink predicts glycaemic responses to mixed meals. Additionally, although acute experimental hyperglycaemia can slow GE, evidence for effects of spontaneous or longer-term variations in glycaemia on GE in diabetes has been inconsistent, necessitating clarification in treatment-naive T2D.

Design: Observational study in 55 newly diagnosed, treatment-naive Han Chinese adults with T2D. Ethics approved (Nanjing First Hospital KY20220124-08); trial registered (NCT05284344). Admission and Protocol: Participants were admitted for three consecutive days. Day 1: medical history, assessment of microvascular complications; insertion of a CGM sensor (Medtronic) for interstitial glucose monitoring. Day 2: standardised meals—breakfast at 07:00 (459 kcal; 46% carbohydrate, 33% fat, 21% protein), lunch at 11:00 (615 kcal; 39% carbohydrate, 35% fat, 26% protein), dinner at 17:00 (438 kcal; 58% carbohydrate, 22% fat, 20% protein). Only water allowed besides the meals. Capillary glucose checked four times for CGM calibration. Day 3: after overnight fast, participants consumed 75 g glucose in 300 mL water over 0–5 min, containing 150 mg 13C-acetate. Breath samples collected at baseline and every 15 min for 3 h. Venous blood sampled at 0, 30, 60, 90, 120, 150, and 180 min for plasma glucose; baseline samples for HbA1c and serum fructosamine. Measurements: Plasma glucose (glucose oxidase; Hitachi 7600-120), HbA1c (HPLC; Bio-Rad), serum fructosamine (Glamour 2000; normal 200–285 µmol/L). CGM data from Day 2 (00:00–24:00) used to compute 24-h mean interstitial glucose, mean amplitude of glycaemic excursions (MAGE), coefficient of variation (CV), and time in range (TIR; 3.9–10.0 mmol/L). 13CO2 in breath was measured by HCBT-01 analyser. Gastric emptying half-time (T50) was derived from the 13C-acetate breath test using the Wagner–Nelson method, previously validated against scintigraphy. Sample size and Statistics: Based on prior work, n=20 provides ≥80% power to detect correlations between GE and glycaemic responses; 55 were enrolled. Primary analyses examined relationships between iAUCs for plasma glucose after oral glucose and interstitial glucose after mixed meals with T50 of the glucose drink over prespecified time windows. Secondary analyses assessed relationships between T50 and antecedent glycaemic control: fasting plasma glucose (FPG), 24-h CGM metrics (mean, MAGE, CV, TIR), serum fructosamine, and HbA1c. Normality was confirmed; univariate linear regression was applied (GraphPad Prism 9.0). Data are mean ± SEM; significance at P<0.05.

Participants (n=55) had high HbA1c (9.8 ± 0.2%) and fructosamine (427.6 ± 8.1 µmol/L); small subsets had microvascular complications (nephropathy 3/55; neuropathy 7/55; retinopathy 7/55). Day-2 CGM showed 24-h mean interstitial glucose 13.4 ± 0.3 mmol/L; TIR 19.7 ± 2.8%; CV 22.3 ± 1.0%; MAGE 6.3 ± 0.4 mmol/L. On Day 3, fasting plasma glucose was 11.1 ± 0.3 mmol/L; after 75 g glucose, plasma glucose peaked at 22.1 ± 0.4 mmol/L at 90 min. Gastric emptying T50 of the glucose drink averaged 73.5 ± 3.3 min (range 26.0–134.7). Relationships with oral glucose: Plasma glucose iAUCs were inversely related to T50 at 0–60 min (r = −0.32, P = 0.016), 0–90 min (r = −0.40, P = 0.003), and 0–120 min (r = −0.34, P = 0.012). Relationships with mixed meals: Interstitial glucose iAUCs were inversely related to T50 after breakfast at 0–30 min (r = −0.54, P < 0.001), 0–60 min (r = −0.51, P < 0.001), and 0–120 min (r = −0.34, P = 0.012); and after dinner at 0–30 min (r = −0.34, P = 0.011), 0–60 min (r = −0.36, P = 0.007), and 0–120 min (r = −0.28, P = 0.040). No significant relationship was observed between lunch iAUCs and T50. Glycaemic variability metrics (MAGE, CV, TIR) over 24 hours did not correlate with T50. Antecedent glycaemia and GE: T50 was not related to FPG (r = 0.19, P = 0.155), 24-hour mean interstitial glucose (r = −0.02, P = 0.896), serum fructosamine (r = −0.05, P = 0.704), or HbA1c (r = 0.06, P = 0.686).

GE of a 75 g glucose drink predicts the magnitude of postprandial glycaemic responses not only to the glucose load itself but also to physiological mixed meals, particularly breakfast and dinner, in newly diagnosed, treatment-naive Chinese adults with T2D. The inverse correlations between T50 and early-to-mid postprandial glycaemic increments are consistent with GE being a principal driver of glucose appearance in the circulation. Lack of association beyond 2 hours likely reflects increasing contributions from insulin secretion/action and other disposal mechanisms. The absence of correlation for lunch may be due to residual gastric contents from breakfast, as suggested by higher pre-lunch glucose levels. Importantly, GE was unrelated to antecedent short-, medium-, or long-term glycaemic control, indicating that spontaneously elevated glycaemia does not substantially modulate GE in this population, in contrast to acute experimentally induced hyperglycaemia known to slow GE. These findings support the clinical utility of measuring GE during a 75 g OGTT to elucidate mechanisms underlying postprandial hyperglycaemia and to tailor dietary or pharmacological strategies that slow GE to improve postprandial glucose control.

In newly diagnosed, treatment-naive, Chinese patients with T2D, GE of a 75 g glucose drink predicts the glycaemic response to more physiological meals, especially breakfast and dinner, and is not influenced by spontaneous short-, medium-, or long-term variations in glycaemia. Concurrent measurement of GE during a 75 g OGTT, achievable with a stable isotope breath test, may inform management of postprandial glycaemia in this population. Measurement of GE appears not to be limited by spontaneously developed hyperglycaemia. Future research should evaluate generalisability across broader T2D populations, different levels of glycaemic control, and under various glucose-lowering therapies, including interventional studies manipulating glycaemia to test effects on GE.

- Modest sample size, although sufficient to detect consistent relationships. - Study restricted to newly diagnosed, treatment-naive T2D patients; generalisability to broader T2D populations and to healthy controls is uncertain. - Observational design; causality cannot be inferred. - Lack of detailed dietary habit data, which may influence inter-individual variability in GE.