Foregut Exclusion Enhances Incretin and Insulin Secretion After Roux-en-Y Gastric Bypass in Adults With Type 2 Diabetes

Abstract

Introduction

Patients with type 2 diabetes experience resolution of hyperglycemia within days after Roux-en-Y gastric bypass (RYGB) surgery. This is attributed, in part, to enhanced secretion of hindgut factors following exclusion of the gastric remnant and proximal intestine during surgery. However, evidence of the mechanisms of remission remain limited due to the challenges of metabolic evaluation during the early postoperative period. The purpose of this investigation was to determine the role of foregut exclusion in the resolution of type 2 diabetes after RYGB.

Methods

Patients with type 2 diabetes (n = 15) undergoing RYGB had a gastrostomy tube (G-tube) placed in their gastric remnant at time of surgery. Patients were randomized to receive a mixed meal tolerance test via oral or G-tube feeding immediately prior to and 2 weeks after surgery in a repeated measures crossover design. Plasma glucose, insulin, C-peptide, incretin responses, and indices of meal-stimulated insulin secretion and sensitivity were determined.

Results

Body weight, fat mass, fasting glucose and insulin, and circulating lipids were significantly decreased 2 weeks after surgery. The glycemic response to feeding was reduced as a function of total area under the curve but not after adjustment for the reduction in fasting glucose. Oral feeding significantly enhanced insulin and incretin secretion after RYGB, which was entirely ablated by G-tube feeding.

Conclusion

Foregut exclusion accounts for the rise in incretin and insulin secretion but may not fully explain the early improvements in glucose metabolism after RYGB surgery.

It is estimated that more than 700 million people globally have pre- or type 2 diabetes (1). The greatest risk factor for the development and severity of type 2 diabetes is excess body weight as evidenced by the presence of overweight and obesity in ~90% of all type 2 diabetes cases (2). Metabolic surgeries, including Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG), have gained recognition as being more effective than conventional lifestyle and medical therapies for the treatment of type 2 diabetes, with ~50% of patients still in remission 2 or more years following surgical intervention depending on the degree of β-cell function at time of surgery (3). One of the prevailing hypotheses explaining RYGB-mediated glycemic benefits is that exclusion of the duodenum and proximal jejunum from nutrient transport may restrict secretory factors that drive hyperglycemia and insulin resistance (4). Comparing oral vs gastrostomy tube (G-tube) feedings after RYGB surgery allows one to evaluate the effects of enhanced delivery of nutrients to the distal intestine on glucose homeostasis while gastrostomy feeding provides selective stimulation of the excluded duodenal and proximal jejunum (5,6). To test this hypothesis, we placed G-tubes at the time of RYGB surgery and assessed postprandial glucose metabolism in patients with type 2 diabetes in a randomized controlled crossover design. We hypothesized that rerouting nutrients back through the duodenum via G-tube after RYGB would reverse the enhanced insulin secretory effect of RYGB and ablate restoration of postprandial glycemia.

Methods

Study Design

This trial was a randomized controlled crossover substudy within STAMPEDE II (NCT01278823). Eligible study patients were candidates for general surgery, 18 to 60 years old, body mass index of 27 to 43 kg/m², clinical diagnosis of type 2 diabetes, not currently pregnant or breastfeeding, and willing to use reliable birth control for the duration of the study. Patients were excluded for prior bariatric surgeries of any type; prior complex abdominal surgery; abdominal, thoracic, pelvic, and/or obstetrical surgery within the last 6 months; significant cardiovascular disease; kidney disease with a creatinine greater than or equal to 1.8 mg/dL; chronic liver disease except for nonalcoholic fatty liver disease/nonalcoholic steatohepatitis; celiac, malabsorptive, and/or inflammatory bowel disorders; psychiatric disorders requiring 3 or more medications; cancer except squamous or basal cell skin cancer or cancer in remission; anticoagulation therapy that could not be stopped for surgery; clotting disorders; and severe pulmonary disease. The study was approved by the Cleveland Clinic Institutional Review Board, and all participants provided written informed consent in accordance with our guidelines for the protection of human participants.

Surgical procedures

All participants received RYGB with intent-to-treat type 2 diabetes as previously described (7). Briefly, gastric bypass was achieved by the creation of a 15 to 20 mL gastric pouch, a 150 cm Roux limb, and a 50 cm biliopancreatic limb. During the surgery, a G-tube was placed into the gastric remnant to maintain enteral access to the excluded proximal bowel, which remained in place for the 2-week study period (Fig. 1). All bariatric procedures were performed laparoscopically by a single surgeon. Postsurgery, patients were instructed to maintain a 500 to 1000 Kcal liquid diet for 2 weeks in compliance with the standards described by the American Society for Metabolic and Bariatric Surgery (8). Medications for treatment of type 2 diabetes were withheld for 24 h prior to metabolic testing.

Figure 1.

Schematic illustration of gastrostomy feeding tube insertion during RYGB.

Anthropometrics and Body Composition

After a 12-h overnight fast, height and weight were measured using standard techniques in a hospital gown. Systolic and diastolic blood pressure was determined using an automated platform (DINAMAPProcare 400; GE Medical Systems) as described previously (9). Dual-energy X-ray absorptiometry (Lunar iDXA; Madison, WI, USA) was used to determine body fat and lean mass as described previously (10).

Mixed Meal Tolerance Test

Postprandial insulin secretion and sensitivity were determined by a mixed meal tolerance test (BOOST® Original: 150 kcal; 68.3% carbohydrate, 15.0% protein, 16.7% fat) immediately prior to (7 ± 2 days) and 2 weeks (10.6 ± 3.4 days) after surgery. During the test, patients were seated in a semi-recumbent position, and blood samples were obtained from an indwelling intravenous line. Samples were assayed immediately for glucose (YSI 2300; STAT Plus, Yellow Springs, OH, USA), and plasma was isolated by centrifugation at 1000 rpm for 10 min at 4°C and stored at −80° until time of assay. The homeostatic model of insulin resistance [HOMA-IR; (glucose × insulin)/405), disposition index [(insulin₀₋₃₀/glucose₀₋₃₀)/S_I (Matsuda)] (11), and hepatic extraction indices were derived according to the molar ratios of C-peptide:insulin as described previously (12). Total area under the curve and incremental area under the curve (iAUC) for glucose, insulin, C-peptide, glucagon‐like peptide‐1 (GLP-1), and gastric inhibitory polypeptide (GIP) during the mixed meal tolerance test (MMTT) were determined using the trapezoidal rule.

Substrate utilization

Total energy expenditure, whole body respiratory exchange ratio, and substrate metabolism were assessed under basal and post-prandial conditions (~120 min) via indirect calorimetry as previously described (Vmax Encore; Viasys, Yorba Linda, CA, USA) (13).

Biochemical Analyses

Blood metabolite and lipid concentrations were determined at baseline and 2 weeks post-RYGB using an automated clinical diagnostic platform as described previously (11). Samples for plasma insulin and C-peptide were collected in vacutainers containing EDTA and the protease inhibitor aprotonin, and were analyzed by radioimmunoassay (Linco, St. Louis, MO, USA). GLP-1 and GIP were collected in vacutainers containing EDTA, aprotonin, and dipeptidyl peptidase IV inhibitor and analyzed by enzyme-linked immunosorbent assay (Millipore, Billerica, MA, USA).

Statistical Analysis

Normality of distribution was assessed by the Shapiro-Wilk test. Sphericity was validated by Mauchly’s test. Graph Pad Prism 7 and SAS 9.3 packages were used for statistical analyses. Student’s paired samples t-test or Fisher exact test were used to assess differences from baseline to 2 weeks postsurgery. A repeated-measures analysis of variance (ANOVA) was used to assess differences in mixed meal responses over time. The level of significance was set a priori at P < 0.05.

Results

Eighteen patients with type 2 diabetes (7.8 ± 2.4 years) were prospectively randomized to receive oral and G-tube feeding in a randomized crossover design (Table 1). After RYGB surgery, 3 patients were excluded due to surgical and/or medical complications that prohibited G-tube insertion (Fig. 2), and 15 were included for analysis. Body weight, body mass index (both P = 0.001), body fat, and lean mass (both P < 0.001) were reduced from baseline (PRE) to 2 weeks post-RYGB. Diastolic blood pressure but not systolic blood pressure was significantly decreased 2 weeks post-RYGB (P = 0.013 and P = 0.402, respectively). Fasting glucose and high-density lipoprotein were significantly reduced after surgery (P < 0.001 and P = 0.005, respectively), whereas no differences were evident for low-density lipoprotein, very low-density lipoprotein, triglycerides, total cholesterol, alanine amino transferase, aspartate aminotransferase, and creatinine (all P > 0.05).

Table 1.

Participant characteristics at baseline (PRE) and 2-weeks after (POST) RYGB surgery

	PRE	POST	P-value
Medical history
Age, years	46.5 ± 2.9	—	—
Sex, M/F	3/12	—	—
Duration of type 2 diabetes, years	7.8 ± 2.4	—	—
Medication
Insulin, n (%)	8 (53.3)	6 (40.0)	0.165
Metformin, n (%)	8 (53.3)	8 (53.3)	0.999
Thiazolidinediones, n (%)	0 (0)	0 (0)	0.999
Incretins, n (%)	4 (26.6)	0 (0)	0.040
Body composition and blood pressure
Weight, kg	135.6 ± 8.9	126.6 ± 8.9	0.001
Body mass index, kg/m2	50.3 ± 2.5	47.0 ± 2.8	0.001
Fat mass, kg	67.4 ± 3.6	60.0 ± 4.2	<0.001
Lean mass, kg	67.3 ± 3.6	60.6 ± 2.7	<0.001
Systolic blood pressure, mmHg	125.2 ± 3.2	119.1 ± 4.9	0.402
Diastolic blood pressure, mmHg	73.8 ± 2.5	66.1 ± 2.2	0.013
Blood markers
Fasting glucose, mg/dL	177.7 ± 14.0	124.5 ± 10.8	<0.001
High-density lipoprotein, mg/dL	38.5 ± 2.1	31.3 ± 1.8	0.005
Low-density lipoprotein, mg/dL	83.5 ± 6.6	86.3 ± 9.6	0.917
Very low-density lipoprotein, mg/dL	28.8 ± 2.6	24.3 ± 2.0	0.080
Triglyceride, mg/dL	144.5 ± 13.1	122.1 ± 10.5	0.066
Total cholesterol, mg/dL	151.0 ± 8.2	142.0 ± 10.2	0.228
Alanine amino transferase, U/L	19.1 ± 3.0	18.4 ± 4.1	0.819
Aspartate aminotransferase, U/L	24.6 ± 2.7	19.3 ± 2.7	0.359
Creatinine, mg/dL	0.71 ± 0.03	0.73 ± 0.03	0.120

Figure 2.

CONSORT flow diagram of patients through trial phases.

Oral and G-tube feeding significantly reduced total AUC glucose after surgery (PRE: 21 930 ± 3253, RYGB + Oral: 18 040 ± 2318, RYGB + G-tube: 16 855 ± 2065mg/dL × min⁻¹; P = 0.024 and P = 0.018 vs PRE, respectively), with no differences observed between the 2 feeding procedures (P = 0.211). iAUC glucose did not differ among the 3 conditions (Fig. 3A and 3B). iAUC for insulin was increased after oral delivery (P = 0.023 vs baseline; P = 0.009 vs RYGB + G-tube), and this was ablated after G-tube feeding (P = 0.577 vs PRE) (Fig. 3C and 3D). In addition, the oral feeding led to an increase in the iAUC for C-peptide (P = 0.015 vs PRE; P = 0.006 vs RYGB + G-tube) while a significant reduction in the C-peptide iAUC was observed after G-tube feeding (P = 0.0005 vs PRE) (Fig. 3E and 3F). Moreover, oral feeding increased GLP-1 and GIP-1 iAUCs compared to PRE (P = 0.0276 and P = 0.0023, respectively) and G-tube feeding (P = 0.0427 and P = 0.0468, respectively) (Fig. 4A and 4D).

Figure 3.

Assessment of glucose, insulin, and C-peptide concentrations after a mixed-meal tolerance test (A, C, and E), incremental AUC glucose (B), insulin (D), and C-peptide (F). Total AUC glucose and incremental AUC insulin and C-peptide were compared by 1-way ANOVA with Tukey’s post hoc analysis. Data are presented as mean ± SE. *P < 0.05. Abbreviations: AUC, area under the curve; RYGB, Roux-en-Y gastric bypass surgery.

Figure 4.

Gastrointestinal hormonal (A and C) and incremental AUC (B and D) mixed-meal tolerance response after RYGB surgery. Incremental AUC GLP-1 and GIP were compared by 1-way ANOVA with Tukey’s post hoc analysis. Data are presented as mean ± SE. *P < 0.05. Abbreviations: AUC, area under the curve; GIP-1, gastric inhibitory polypetide 1. GLP-1, glucagon-like peptide 1; RYGB, Roux-en-Y gastric bypass surgery.

Baseline HOMA-IR decreased significantly after RYGB surgery in both feeding arms (RYGB + oral: P = 0.02 vs PRE; RYGB + G-Tube: P = 0.01 vs PRE) (Fig. 5A). β-cell function was significantly increased after oral feeding compared to PRE (P = 0.0026) and G-tube feeding (P = 0.0005) (Fig. 5B). Fasting hepatic extraction was significantly increased after RYGB with both oral and G-tube feeding (RYGB + oral: P = 0.0218 vs PRE; RYGB + G-tube: P = 0.0450 vs PRE) (Fig. 5C). However, postprandial hepatic extraction tended to be greater after G-tube compared to oral feeding (P = 0.0643 vs RYGB + oral) (Fig. 5D). Neither between- nor within-arm differences in respiratory quotient and resting energy expenditure (both P > 0.05) were observed (Fig. 5E-F).

Figure 5.

Changes in insulin resistance (A), insulin secretion corrected by insulin sensitivity (B), prehepatic extraction (C and D), and whole-body energy metabolism (E and F) after RYGB surgery. HOMA-IR, diposition index, and hepatic extraction were analyzed by 1-way ANOVA with Tukey’s post hoc analysis. Respiratory quotient and resting energy expenditure were compared by 2-way ANOVA (trial × time) with Tukey’s post hoc analysis. Data are presented as mean ± SE. *P < 0.05. Abbreviations: HOMA-IR, homeostatic model assessment for insulin resistance; REE, resting energy expenditure; RYGB, Roux-en-Y gastric bypass surgery.

Discussion

A fundamental barrier to understanding the mechanisms of diabetes remission following RYGB surgery lies in the challenge of directly measuring metabolic and physiological changes in the human gastrointestinal tract. Using a randomized controlled study design, we tested a prevailing view that foregut derived factors negatively regulate glycemic control by performing G-tube feeding in the early postsurgical period and thereby restore nutrient flow through the bypassed duodenum and jejunum (14). We observed improved glucose homeostasis accompanied by increased postprandial insulin and incretin secretion in response to an oral glucose challenge, congruent with previous reports in mice, rats, and humans (6,15-18). While this result was expected, we did not anticipate glucose responses to G-tube feeding to be similar to the oral feeding. Although a similar response was observed by Hansen et al (19), it is contrary to 2 independent studies in rodents where G-tube feeding in the excluded gastric remnant essentially reversed the antidiabetic effects of RYGB (5,6). This result is even more intriguing in light of the fact that G-tube feeding entirely ablated the effect of RYGB on both postprandial insulin and incretin secretion. These data suggest that enhanced enteroinsular action is surgery dependent but may not fully explain the effects of RYGB on glucose metabolism.

The role of the enteroinsular axis in mediating the effect of RYGB remains elusive. The intestine is one of the most rapid self-renewing tissues in mammals and utilizes glucose and glutamine to fuel these metabolic processes (20,21). There is evidence from both human and rodent studies to support a role for surgery-induced gastrointestinal remodeling in the regulation of glycemia (22-24). In a preclinical model, Saedi et al observed that the jejunal segment, created during the RYGB procedure, directly contributed to decreased circulating glucose and the reversal of type 2 diabetes. They attributed the response to the increased anabolic demands of intestinal tissue growth and maintenance (22). They also observed that the increased intestinal mass (due to hyperplasia and hypertrophy) occurred independent of weight changes following surgery and concluded that intestinal remodeling may contribute to the restoration of glucose homeostasis. Since the glucose response to the MMTT after surgery was not reversed by the G-tube feeding, alternative mechanisms such as intestinal remodeling and hypertrophy may be responsible for normalizing glycemia.

This idea is also supported by data showing that gastrointestinal reprogramming following RYGB results in a substantial increase in intestinal glucose uptake (22,24,25). Moreover, the degree of uptake correlated tightly with resolution of glycemia (24). A number of potential mechanisms could explain this phenomenon. First, increased secretion of GLP-2 after surgery could induce intestinal hypertrophy, ultimately enhancing intestinal blood flow and nutrient uptake (26). Second, glucose excretion into the intestinal lumen via the sodium-glucose cotransporter could increase glucose utilization by enteric cells (24). Third, RYGB may increase glucose transporter (GLUT)-1–mediated uptake into the Roux limb (22). This is supported by observations that GLUT-2 and GLUT-3 expression are unaltered following RYGB (22). The argument favoring a putative role of intestinal GLUT-1 is strengthened by phloretin inhibition, which has been shown to blunt glucose uptake in the Roux limb (22).

Conversely, it remains possible that restoration of glycemia following RYGB is mediated by the magnitude and durability of weight loss. Lean et al demonstrated that ~45% of patients achieved complete remission of type 2 diabetes 2 years after intensive calorie restriction and behavioral weight maintenance (27). These findings are of central importance in that they suggest that dietary restriction alone can induce sustained improvements in glycemic control. Furthermore, the percentage of patients achieving complete remission paralleled what is typically observed after SG and RYGB (28), albeit the surgical patients had poorer β-cell function than those subjected to intense calorie restriction. It has also been reported that both SG and RYGB can bring about similar improvements in insulin secretion ~3 weeks after surgery, and the effect persists for years (29). Since the gastrointestinal tract is unaltered by SG, one may further conclude from these data that the profound effects of RYGB on the enteroinsular axis may be explained by enhanced weight loss. More recently, Yoshino et al (30) demonstrated that weight-matched calorie restriction achieves similar improvements in hepatic and peripheral insulin sensitivity and β-cell function as RYGB. This may imply that calorie restriction can yield similar improvements in glycemic control and peripheral insulin sensitivity in a controlled environment where it is possible to match weight loss achieved by RYGB (30). However, phenotypic similarities between calorie restriction and RYGB in the weight-reduced state are not indicative of a unifying causative mechanism.

In conclusion, foregut exclusion of the duodenum and upper jejunum stimulates incretin-mediated insulin secretion which contributes, in part, to the glucose lowering effects of RYGB. Emerging mechanisms, such as intestinal proliferation and enhanced blood flow, provide new areas of research to explain the weight-independent effects of RYGB surgery on diabetes remission. More broadly, these findings highlight a critical need for a direct comparison between metabolic surgery and intensive calorie restriction, for both short- and long-term improvements in glycemic control and type 2 diabetes remission.

Limitations

Eighty percent of all patients who undergo metabolic surgical procedures are female and as such sex-specific effects cannot be excluded. Postoperative MMTTs were conducted 2 weeks after surgery. It is therefore plausible that foregut-mediated effects occur more rapidly after surgery, or much later, and as such were not captured by our study timeline. G-Tube feeding after RYGB may alter gastric emptying of nutrients and thus may not accurately mimic the pre-RYGB state. The glycemic responses to the mixed meal were dampened due to the relatively small volume that was used to overcome food aversions, tolerance, and potential for dumping syndrome after surgery.

Author Contributions: JPK, SRK, and PRS designed the study. JPK, CLA, EML, SKM, WSD, KP, SAB, and SRK performed data collection, analysis, and contributed to the interpretation. CLA, WSD, and KP drafted the manuscript, and all authors contributed to editing and finalization of the manuscript. JPK 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.

Clinical Trial Information: NCT01278823.

Additional Information

Disclosures: None of the authors have a conflict of interest related to this work.

Data Availability

The data that support the findings of this study will be made available upon reasonable request by the lead contact, John P. Kirwan (john.kirwan@pbrc.edu).