The Glucagon-Like Peptide 1 Receptor Agonist Exenatide Inhibits Small Intestinal Motility, Flow, Transit, and Absorption of Glucose in Healthy Subjects and Patients With Type 2 Diabetes: A Randomized Controlled Trial

Therapies specifically targeting postprandial glycemia are important in the management of type 2 diabetes, especially in patients with relatively good overall glycemic control (HbA_1c ≤7.5%; 58 mmol/mol) (1). The rate of gastric emptying is an established determinant of postprandial blood glucose (2), a principle illustrated by “short-acting” glucagon-like peptide 1 (GLP-1) receptor agonists, such as exenatide, where the capacity to slow gastric emptying predominates over the insulinotropic effect in the postprandial setting (3).

Small intestinal glucose absorption, predominantly via sodium–glucose cotransporter 1 and GLUT2 transporters, is limited to ∼0.5 g/min per 30 cm (2). Interventions that increase the exposure of luminal glucose to the mucosal surface can therefore augment glucose absorption. We previously reported that the anticholinergic agent hyoscine delays the absorption of intraduodenally infused glucose in humans by decreasing small intestinal flow (4), indicating that modulation of small intestinal motor function can impact substantially on postprandial glycemia. Exogenous GLP-1 has been reported to inhibit both fasting and postprandial duodenal motility in humans (5,6), but its impact on the flow of chyme and on small intestinal transit and glucose absorption have not previously been explored. We therefore examined the effects of the short-acting form of exenatide on small intestinal motor function and glucose absorption in response to an intraduodenal glucose infusion in both healthy subjects and patients with type 2 diabetes.

We studied 10 healthy subjects and 10 patients with type 2 diabetes managed by diet alone, after obtaining written informed consent (Table 1). None had any significant medical comorbidity or used medications known to affect gastrointestinal motility. The protocol was approved by the Human Research Ethics Committee of the Royal Adelaide Hospital and was conducted in accordance with the principles of the Declaration of Helsinki as revised in 2000.

Each subject underwent two study visits separated by at least 5 days in double-blind, randomized fashion. Computerized randomization was undertaken by the Royal Adelaide Hospital Pharmacy, which supplied a 50-mL solution in a masked bag containing either exenatide (40 μL Byetta solution [AstraZeneca, North Ryde, Australia], 0.5 mL 20% albumin, and 49.5 mL 0.9% saline, to a final concentration of 10 μg/50 mL) or 0.9% saline alone (control).

Each subject attended the Department of Nuclear Medicine at 8:30 a.m. after an overnight fast. A multilumen silicone catheter (Dentsleeve, Mississauga, Canada) bound to an impedance catheter (Sandhill Scientific, Highlands Ranch, CO) was introduced through an anesthetized nostril and allowed to pass into the duodenum by peristalsis, with continuous monitoring of its position by measurement of antroduodenal transmucosal potential difference via saline-perfused side-holes (4). The assembly was positioned with six water-perfused manometry side-holes at 3-cm intervals, seven impedance electrode pairs (2 cm between each electrode), and an infusion port in the duodenum.

Subjects remained supine under a gamma camera (7,8). At −30 min, an intravenous infusion of exenatide or control was commenced and continued for 270 min (50 ng/min exenatide for the first 30 min and then 25 ng/min). Between 0 and 60 min, an intraduodenal glucose infusion (45 g glucose, mixed with 5 g 3-O-methylglucose [3-OMG] and 20 MBq ^99mTc-sulfur colloid in water to a total volume of 200 mL) was administered at a rate of 3.3 mL/min (3 kcal/min). 3-OMG is absorbed by the same mechanisms as glucose but is not metabolized; serum concentrations therefore represent an index of glucose absorption (9). A cobalt marker was placed over the right superior iliac spine as a reference (10). Anterior scintigraphic images were acquired every 3 min from 0 to 240 min. Venous blood was sampled frequently for blood glucose, serum 3-OMG and insulin, and plasma C-peptide concentrations, and gastrointestinal sensations were assessed using 100-mm visual analog questionnaires (11).

Both the manometric and impedance data were recorded digitally (InSIGHT stationary system; Sandhill Scientific). The number and amplitude of duodenal waves over successive 15-min periods were analyzed with custom-designed software (Andre J. Smout, Academic Medical Center, Amsterdam, the Netherlands). Flow events were defined as a transient reduction in impedance of ≥12% from baseline in three or more sequential electrode pairs (4). Small intestinal transit time was determined from the scintigraphic data (IDL v6.2 software; RSI, Boulder, CO). A region of interest was drawn around the colon using a composite image and used to identify the cecal arrival time (10). Blood glucose concentrations were measured by the glucose oxidase method (MediSense Optium, Bedford, MA), serum 3-OMG by liquid chromatography and mass spectrometry (10), and serum insulin and plasma C-peptide by ELISA immunoassays (Mercodia, Uppsala, Sweden).

On the basis of our previous study (4), we determined that 10 subjects would provide 80% power to detect a 50% reduction in duodenal pressure waves and 90% power to detect a 50% reduction in flow events, with exenatide compared with control (α = 0.05). The area under the curves (AUCs) for different parameters were calculated using the trapezoidal rule. Paired Student t test and two-factor repeated-measures ANOVA, with treatment and time as factors (adjusted by Bonferroni correction), were used to compare variables within each group. Motility indices (MIs) were calculated as ln[(sum of amplitudes × number of duodenal waves) + 1] (12). Wilcoxon signed rank test was used for intragroup comparisons of nausea scores. Within-subject correlations using univariate analysis, with subject number as a fixed factor, were used to assess relationships between variables. This method allows assessment of “weighted correlations” between variables in the total data pool, with P values adjusted for the total number of subjects (13). All analyses were performed using SPSS 21 (IBM, Armonk, NY). Results are expressed as mean ± SEM; P < 0.05 was considered statistically significant.

During and after intraduodenal glucose infusion (0–240 min), there were marked reductions in the frequency and MI of duodenal pressure waves with exenatide compared with control, in both healthy subjects and subjects with type 2 diabetes (all P < 0.05). Duodenal flow events were mainly (>95%) antegrade and were substantially fewer with exenatide compared with control, in both groups (all P < 0.05) (Fig. 1 and Table 1).

With saline control, the mean small intestinal transit time was 121 min (95% CI 98–144) in healthy subjects and 120 min (95% CI 85–155) in patients with type 2 diabetes (n = 9), without a difference between the groups. The remaining patient with type 2 diabetes had a transit time >240 min. With exenatide, there was a marked slowing of transit, such that the radiolabel did not reach the cecum within 240 min in any healthy subject or subject with type 2 diabetes.

From −30 to 0 min, blood glucose decreased slightly on both days in healthy volunteers (P ≤ 0.05), but only with exenatide (P < 0.001) in subjects with type 2 diabetes. During this period, insulin and C-peptide increased only with exenatide in each group (all P < 0.05) (Fig. 2 and Table 2).

During and after intraduodenal glucose infusion (0–240 min), blood glucose, insulin, and C-peptide concentrations increased to a peak before decreasing to near-baseline values. Both the peak and AUC for blood glucose were lower with exenatide than with control in both groups (P < 0.05), whereas the peak and AUC for insulin and C-peptide were higher with exenatide than with control in patients with type 2 diabetes (P < 0.05), but not in healthy subjects. With exenatide, blood glucose and insulin both decreased slightly during the initial 30 min of intraduodenal glucose infusion in both groups, in contrast to control. Peak and AUC for serum 3-OMG concentrations were markedly lower with exenatide than with control in both groups (all P < 0.005) (Fig. 2 and Table 2).

During (but not before) intraduodenal glucose infusion (0–60 min), nausea increased in each group only with exenatide and returned to baseline after the end of the glucose infusion, such that the mean and peak nausea scores were higher with exenatide than with control (all P < 0.05) (Supplementary Fig. 1).

Even in subjects with a peak nausea score below the median of 27.5 mm (n = 10; four healthy subjects and six subjects with type 2 diabetes), the suppression of duodenal pressure waves (678 ± 137 vs. 1,963 ± 467; P < 0.01) and antegrade flow events (58 ± 9 vs. 106 ± 6; P < 0.001) with exenatide remained significantly different from control and was still associated with lower mean blood glucose and serum 3-OMG concentrations (P < 0.001 for each).

On pooling the data from both groups on both study days, the AUCs for serum 3-OMG and blood glucose concentrations were related directly to the frequency of duodenal pressure waves (r = 0.66, P = 0.001 and r = 0.58, P = 0.006, respectively), duodenal MI (r = 0.84, P < 0.001 and r = 0.67, P = 0.001, respectively), and antegrade flow events (r = 0.75, P < 0.001 and r = 0.59, P = 0.005, respectively) but were not related to small intestinal transit time on control days.

We observed that exenatide, when administered intravenously in both healthy subjects and patients with type 2 diabetes, 1) markedly suppressed duodenal motility and flow, 2) slowed small intestinal transit, 3) decreased 3-OMG absorption, and 4) delayed and suppressed glycemic increments in response to an intraduodenal glucose infusion. The dose of exenatide (7.5 μg) was within the usual clinical range and was infused in a regimen shown previously to achieve steady therapeutic plasma concentrations (14,15).

In addition to manometry, we evaluated small intestinal motor function using the impedance technique (16), which is more sensitive than manometry for assessing flow (17). We observed a direct relationship between absorption of the glucose analog, 3-OMG, and the frequency of both duodenal pressure waves and flow events; the latter is likely to be more important based on our previous study (4).

The inhibition of motility and flow that we observed with exenatide is likely to have reduced the length of small intestinal mucosa exposed to glucose (18), which is supported by our scintigraphic data, and to have diminished the thinning of the unstirred water layer, which occurs with flow of chyme (19); both phenomena would contribute to reduced absorption of glucose, indicated by lower 3-OMG concentrations, despite a possible increase in mesenteric blood flow that occurs with GLP-1 receptor stimulation (20). That GLP-1 receptor agonists are associated with diarrhea in some patients may seem at odds with our observations but could be due to acceleration of colonic transit, which was observed independently of any effect on small intestinal motility with the GLP-1 receptor agonist ROSE-010 (21).

During the intraduodenal glucose infusion (0–60 min), we noted a delay in the rise of blood glucose with exenatide compared with control, associated with initially lower insulin concentrations, suggesting that inhibition of gut motility at first outweighed the insulinotropic effects of exenatide in contributing to lowering of blood glucose. A similar phenomenon has been demonstrated previously in health, when gastric emptying was slowed by infusion of exogenous GLP-1 (22). After the intraduodenal glucose infusion was completed (60–240 min), the insulinotropic effect of exenatide became manifest, which undoubtedly contributed to the reduction in glycemia during this later phase of the study.

Nausea is an established adverse effect of exenatide and is thought to be less with intravenous than subcutaneous administration (23). In our study, nausea tended not to occur with intravenous exenatide until the stimulus of intraduodenal glucose was added, perhaps due to synergy between central and gastrointestinal stimuli. It is unlikely that nausea accounted for the suppression of small intestinal motor function by exenatide, since the latter was evident even in subjects with little or no nausea.

Limitations of this technically challenging, proof-of-concept study include the relatively small sample size and very good glycemic control of our patients with type 2 diabetes. Further studies would be required to evaluate the effects in less well-controlled patients and to determine whether there is tachyphylaxis for the suppression of small intestinal motor function with repeated dosing and whether the effects would be less pronounced if baseline motility were already impaired, as is the case with gastric emptying (24).

Nevertheless, our observations represent a hitherto unrecognized mechanism for lowering postprandial glycemia by GLP-1 receptor agonists in type 2 diabetes.

Acknowledgments. The authors thank Andre J. Smout (Academic Medical Center, Amsterdam, the Netherlands), who provided the software for duodenal motility analysis. The authors particularly wish to acknowledge the contribution of their late colleague Antonietta Russo (Discipline of Medicine, The University of Adelaide) to the acquisition and analysis of small intestinal transit data for this study.

Funding. This study was supported by funding from the National Health and Medical Research Council of Australia. S.S.T. and C.S.M. received scholarship funding from The University of Adelaide, and S.S.T. also received funding from the Rebecca L. Cooper Medical Research Foundation. T.W. has been supported by a Royal Adelaide Hospital Research Committee Early Career Fellowship.

Duality of Interest. This study was supported by funding from AstraZeneca Australia. K.L.J. has received research funding from Merck Sharp & Dohme and Sanofi. M.H. has participated on advisory boards and/or in symposia for AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, Merck Sharp & Dohme, Novartis, Novo Nordisk, and Sanofi and has received honoraria for this activity. C.K.R. has received funding from AstraZeneca, Merck Sharp & Dohme, and Novartis. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. S.S.T. and C.S.M. were involved in the subject recruitment, coordination, data collection and interpretation, statistical analysis, and drafting of the manuscript. T.W. was involved in the data collection, data interpretation, statistical analysis, and drafting of the manuscript. J.C. and J.K. were involved in the subject recruitment, coordination, and data collection. P.K. was involved in the conception and design of the study, subject recruitment, coordination, and data collection. H.L.C., M.J.B., and R.S.R. assisted in recruitment and data collection. B.C. was involved in the design of the study and in the development of the software for analysis of small bowel transit. K.L.J. was involved in the conception and design of the study and data analysis and interpretation. M.H. was involved in the conception and design of the study and data interpretation. C.K.R. was involved in the conception and design of the study and data analysis and interpretation. All authors critically reviewed the manuscript and have approved the final version. C.K.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.