|Home | About | Journals | Submit | Contact Us | Français|
The aim of the present study was to determine the mechanisms underlying Type 2 diabetes remission after gastric bypass (GBP) surgery by characterizing the short- and long-term changes in hormonal determinants of blood glucose.
Eleven morbidly obese women with diabetes were studied before and 1, 6, and 12 months after GBP; eight non-diabetic morbidly obese women were used as controls. The incretin effect was measured as the difference in insulin levels in response to oral glucose and to an isoglycemic intravenous challenge. Outcome measures were glucose, insulin, C-peptide, proinsulin, amylin, glucagon, glucose-dependent insulinotropic polypeptide (GIP), glucagon-like peptide-1 (GLP-1) levels and the incretin effect on insulin secretion.
The decrease in fasting glucose (r = 0.724) and insulin (r = 0.576) was associated with weight loss up to 12 months after GBP. In contrast, the blunted incretin effect (calculated at 22%) that improved at 1 month remained unchanged with further weight loss at 6 (52%) and 12 (52%) months. The blunted incretin (GLP-1 and GIP) levels, early phase insulin secretion, and other parameters of β-cell function (amylin, proinsulin/insulin) followed the same pattern, with rapid improvement at 1 month that remained unchanged at 1 year.
The data suggest that weight loss and incretins may contribute independently to improved glucose levels in the first year after GBP surgery.
One of the major benefits of gastric bypass (GBP) surgery is the remission of Type 2 diabetes in 60–80% of cases.1 The rapidity of onset and the magnitude of the effect of GBP on diabetes remain incompletely explained. In addition to the effect of caloric restriction and weight loss, gut peptides, such as incretins, may play a role in the metabolic improvement after GBP.2,3 The incretins glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) enhance pancreatic insulin secretion (β-cell) and suppress glucagon secretion (α-cell) during meals. The resultant effect is maintenance of normal glucose homeostasis with a reduction in excess endogenous glucose production and better insulin sensitivity. The incretin effect on insulin secretion is impaired in diabetes.4 We have shown that both incretin levels and effects on insulin secretion are markedly increased 1 month after GBP in patients with diabetes,2 but not after a matched weight loss by diet.3 Moreover, although fasting levels of glucose and insulin decrease similarly with surgical and dietary weight loss, post-prandial glucose (PPG) levels decrease more after GBP than after dietary weight loss.3 Some of the effects of GBP on diabetes control may be secondary to caloric restriction and weight loss because they improve similarly after GBP or an equivalent diet-induced weight loss.3 In contrast, the early change in incretins occurs independently of weight loss and may be related to the nature of the surgery. Nevertheless, the relative contribution of changes in incretins and weight loss to diabetes remission is difficult to tease apart because patients lose weight rapidly after the surgery.
The first goal of the present study was to characterize the pattern of changes in hormonal determinants of blood glucose levels in the first year after surgical weight loss by GBP in patients with diabetes. We hypothesized that the early (1 month) increase in incretin levels after oral glucose, and the improved incretin effect on insulin secretion, would remain unchanged up to 1 year after GBP. We also hypothesized that other markers of metabolism and β-cell function, such as glucose, insulin, amylin, proinsulin, glucagon, leptin, and adiponectin, would improve as a function of weight loss.
Eleven obese women with Type 2 diabetes (OBDM) known for less than 5 years (range 1–52 months), HbA1c 6.7 ± 0.8%, body mass index (BMI) 45.1 ± 7.8 kg/m2, and aged 44.3 ± 11.1 years were studied prior to and the 1, 6, and 12 months after GBP. Five women took sulfonylureas and/or metformin and had their medications held 3 days prior to baseline studies and stopped after GBP. Obese women with normal glucose tolerance (OBNGT; n = 8) were used as controls. All participants provided informed consent prior to enrollment in the study. Data at 1 month from eight GBP patients have been published previously.2,3
All OBDM underwent laparoscopic GBP. The jejunum was divided 30 cm from the ligament of Treitz and anastomosed to a 30-mL proximal gastric pouch. The jejunum was re-anastomosed 150 cm distal to the gas-trojejunostomy. The post-GBP diet recommendations included daily intake of 600–800 kcal, 70 g protein, and 1.8 L fluid for the first few weeks, followed by an ad libitum diet with an emphasis on protein, fluid, and vitamin intake in the following year. This was achieved, on an individual basis, with multiple small meals and snacks with various commercial protein supplements. The diet after GBP was monitored by food records, but was neither directly supervised nor controlled for in the few days preceding the testing.
In order to calculate the incretin effect for each patient, all subjects underwent an oral glucose tolerance test (OGTT) followed by an isoglycemic intravenous (i.v.) glucose test (IsoG IVGT). The OGTT was always performed first, followed by, on a separate day, the IsoG IVGT. Each test was performed in the morning after an overnight fast. During the OGTT and IsoG IVGT, the arm used for blood sampling was kept warm with a heating pad.
For the OGTT, after i.v. insertion, at 08:00 h, subjects received 50 g glucose in 200 mL orally. Blood samples were collected over 3 h into chilled EDTA tubes with aprotinin (500 kallikrein inhibitory units/mL blood; Roche Applied Sciences, Indianapolis, IN, USA) and a dipeptidyl peptidase (DPP) IV inhibitor (50 μmol/L; Millipore, St Charles, MO, USA) and centrifuged at 1620 g for 15 min at 4°C to separate the plasma. Plasma samples were stored at −70°C until analysis.
In order to calculate the incretin effect for each patient, an i.v. glucose challenge was given on a separate day. The goal of the IsoG IVGT was to expose the pancreas to blood glucose levels matched to those obtained during the OGTT for the same patient during the same study period. Glucose (sterile 20% dextrose) was infused i.v. over 3 h using a pump (Gemini; CareFusion, San Diego, CA, USA). Blood samples were collected every 5 min using a contralateral antecubital i.v. catheter for immediate bedside measure of glucose levels. The glucose infusion rate was adjusted to match the glucose concentrations obtained for the same patient during the OGTT at each time point over 3 h.
The difference in β-cell responses [insulin total area under the curve (INS AUC0–180)] to the oral and the isoglycemic i.v. glucose stimuli represents the action of the incretin effect, expressed as the percentage of the response to oral glucose, which is taken as the denominator (100%).4 The formula used was:
where INC is the incretin effect and INS AUC is the insulin area under the curve after oral or isoglycemic i.v. glucose, as indicated.
Total GLP-1 was measured by radioimmunoassay (RIA; Millipore) after plasma ethanol extraction. The intra- and interassay coefficients of variation (CV) were 3–6.5% and 4.7–8.8%, respectively. This assay reacts 100% with GLP-17–36, GLP-19–36, and GLP-17–37, but not with glucagon (0.2%), GLP-2 (<0.01%), or exendin (<0.01%).
Amylin and GIP were determined by ELISA (Millipore). The assay reacts 100% with GIP1–42 and GIP3–42, but not with GLP-1, GLP-2, oxyntomodulin, or glucagon. The intra- and interassay CV were 3.0–8.8% and 1.8–6.1%, respectively.
Plasma insulin, C-peptide, proinsulin, and glucagon concentrations were determined using RIA (Millipore) with intra- and interassay CV of 3–8% and 5.5–9%, respectively. The glucagon assay reacts 100% with glucagon, but less than 0.1% with oxyntomodulin.
Glucose concentrations were measured at the bedside by the glucose oxidase method (Beckman Glucose Analyzer; Beckman, Fullerton, CA, USA).
All hormone and metabolite assays were performed at the New York Obesity Research Center, except for the adiponectin assays, which were performed with a Millipore human adiponectin ELISA kit at The University of Texas Southwestern Medical Center.
Total AUCs were calculated using the trapezoidal method. Early phase and late-phase insulin secretion were calculated as the AUC0–30′ and AUC60–180′, respectively. The homeostasis model assessment of insulin resistance [HOMA-IR; fasting serum insulin (μU/mL) fasting plasma glucose (mmol/L)/22.5] was used as an index of insulin resistance, whereas the insulin sensitivity index (ISI) composite was used as an index of whole-body insulin sensitivity and calculated from the OGTT values as follows:
Scatter plots were performed to evaluate the distribution of variables. Differences between OBNGT and OBDM were assessed by anova. GLM with repeated measures was used to detect changes over time during the OGTT in the OBDM group. Because the sample size limited full multiple regression analysis, Pearson correlation and simple bivariate regression analyses were performed to explore relationships between weight loss and/or incretin changes and outcome variables. Two-tailed P < 0.05 was considered significant. All data are expressed as the mean ± SD, except in the figures, where data are presented as the mean ± SEM. Statistical analyses were performed with spss 16.0 (SPSS Inc., Chicago, IL, USA).
Subject characteristics at baseline are given in Tables 1 and and2.2. The rate of weight loss was 2.3 ± 0.5 kg/week in the first month after GBP, decreasing to 0.8 ± 0.2 kg/week from 2 to 6 months (P < 0.0001) and to 0.2 ± 0.1 kg/week from 6 to 12 months (P 0.0001). At 1 year, mean weight loss was 36.0 ± 11.2 kg (range 23.6–63.8 kg), HbA1c had decreased from 6.7 ± 0.8% to 5.2 ± 0.4% (P < 0.0001), and blood pressure and lipid profile (Table 2) had normalized on no or decreased medications. As predicted, HOMA-IR was higher and ISI composite lower in OBDM than in OBNGT prior to GBP and changed as predicted after GBP (Table 2). There was a positive correlation between ISI and adiponectin (r = 0.666; P < 0.001) and a negative correlation between ISI and leptin (r = −0.580; P < 0.001) before and up to 1 year after GBP.
All β-cell markers, namely fasting insulin and glucose-stimulated insulin, fasting proinsulin, C-peptide levels, amylin, and proinsulin, decreased significantly during the first year after GBP (Table 2; Fig. 1). At 12 months, there were no significant differences in fasting and 120 min glucose between the OBDM and OBNGT groups (Tables 1 and and2).2). In OBDM, the blunted early phase of insulin secretion (AUC0–30′) increased significantly 1 month after GBP to remain unchanged at 6 and 12 months (Table 1; Figs 1 and and2).2). This pattern was also seen for proinsulin and amylin levels during the OGTT (Fig. 1). Similarly, the insulinogenic index, or insulin/glucose ratio (AUC0–30′), increased at 1 month and remained unchanged up to 1 year (Table 2). In contrast, the elevated late-phase insulin secretion decreased with continued weight loss (P < 0.0001).
Fasting glucagon did not decrease until 6 months after GBP. Unexpectedly, the increase in glucagon levels after oral glucose, observed before surgery, persisted up to 1 year after GBP (Table 2; Fig. 1).
Fasting incretin levels did not change after GBP (data not shown). The increase in peak GLP-1 and GIP responses to oral glucose, by factors of 10 and 1.4, respectively, shown previously at 1 month,2 did not change further at 6 and 12 months (Table 2; Fig. 2). Glucose levels were matched and did not differ significantly between OGTT and isoG IVGT at each study period (data not shown). The early increase in the incretin effect on insulin secretion (from 22.1% to 46.9%) to the level seen in the control OBNGT (43%) persisted at 6 and 12 months (Table 2; Fig. 2).
There was a positive correlation between body weight and fasting glucose, insulin, proinsulin, amylin, glucagon, and leptin, and a negative correlation between body weight and adiponectin. Only changes in fasting glucose (r = 0.724; P < 0.001), fasting insulin (r = 0.576; P = 0.001), HOMA-IR (r = 0.625; P < 0.001), fasting C-peptide (r = 0.577; P = 0.001), and fasting amylin (r = 0.380; P = 0.046) were correlated with weight loss (Fig. 3). Postprandial glucose (120 min) was negatively correlated with early phase insulin secretion (r = −0.407; P = 0.023), peak GLP-1 (r = −0.562; P < 0.001), and the incretin effect (r = −0.601; P < 0.001), and positively correlated with weight loss (r = 0.519; P = 0.003). Both incretin effect and HOMA-IR were predictors of 120 min glucose (R2 = 0.235; P = 0.024). Peak GLP-1 and incretin effects were correlated with early phase, but not late-phase, insulin secretion. Only fasting glucose after GBP, and not glucose at 120 min, predicted HbA1c at 12 months (R2 = 0.323; P = 0.022).
One year after GBP, all patients went into diabetes remission, were free of medications, were glucose tolerant, had normal lipid profile and blood pressure despite remaining obese. The short duration of the diabetes, diagnosed for less than 5 years prior to the surgery, may explain the 100% remission rate. In diabetes, there is a progressive deterioration in β-cell function and decreased β-cell mass. At the time of diagnosis, islet function is estimated to be approximately 50% that of a lean control and β-cell mass is approximately 40% lower on necropsy studies.5 The impairment of β-cell function, and possibly β-cell mass, appears to be reversible, particularly during the early stages of the disease, by many therapeutic interventions including insulin therapy, incretin mimetics and/or enhancers,6 and weight loss.7,8 Diet-induced weight loss, even of a small amount, decreases inflammation and gluco- and lipotoxicity, and contributes to improvements in β-cell function and insulin sensitivity.8 The contribution of weight loss and the incretins to diabetes remission and glucose control after GBP is not fully understood. Our data demonstrate that the early2 increase in GLP-1 and GIP levels in response to oral glucose is long lasting and sustained 1 year after GBP. This is an important finding, in agreement with cross-sectional and prospective studies showing a robust increase in GLP-19,10 or GIP9 after bypass procedures. In addition to the significant increase in stimulated incretin levels, the incretin effect, which increased rapidly2, remained elevated to the level of non-diabetic controls 1 year after GBP. This persistent incretin responsiveness, still present 4 years after GBP (n = 4; B. Laferrère, unpubl. data, 2009) is a marker of improved β-cell function.
With the increased incretin effect, our data show a shift of the pattern of insulin secretion (and C-peptide), with increased early phase and decreased late phase, in response to oral glucose. This is in agreement with restoration of the early phase of insulin secretion after GLP-1 infusion11 or after i.v. glucose.12 Although the present study is small and limited to 1 year, larger and longer studies have shown that diabetes remission persists years after bypass surgery.1,13 It is possible to speculate that the higher incretin levels after GBP may protect islet cell function and/or β-cell mass.14
Our data suggest that the increase in incretin levels and the effect after GBP, which is accompanied by restoration of early phase insulin secretion, may contribute to the amelioration of PPG. In our previous study, a short-term matched diet-induced weight loss, in contrast to GBP, did not restore the early phase insulin secretion or decrease 120′ glucose in persons with diabetes.3 In addition, the diet did not increase incretin levels or effects.3 A recent study in Goto-Kakizaki rats showed that increased GLP-1 secretion and improved glucose tolerance after duodenal jejunal bypass is reversed by a GLP-1 receptor antagonism, providing direct evidence that improvement of glucose tolerance following a bypass surgery is mediated by enhanced GLP-1 action.15 Our data suggest a beneficial effect of GBP via changes in incretins and early phase insulin secretion to control PPG. This could have substantial clinical relevance because elevated PPG is a marker of cardiovascular risk.16 According to our hypothesis, the change in incretin levels and effect, as well as in early phase insulin secretion after GBP are independent of weight loss and are likely a consequence of the surgical bypass procedure per se, as suggested by animal studies.17
In addition to an incretin-mediated effect on improved insulin secretion and glucose homeostasis, GBP exerts its metabolic effect largely via the significant and sustained weight loss. Our data also show that, in contrast to incretin changes, fasting glucose, proinsulin, C-peptide and insulin levels improve as a function of weight loss in the first year after the surgery.
The elevated proinsulin levels observed before surgery decreased significantly, reflecting the decreased demand on the β-cells and improved islet function, as shown previously in non-diabetic individuals.18 Fasting amylin levels and the response to oral glucose improved at 1 month and did not change with further weight loss. The pattern of levels of insulin, pro-insulin and amylin, three β-cell peptides, in response to oral glucose stimulation changed similarly after GBP, with increased early phase and decreased late phase secretion (Fig. 1). The changes occur rapidly at 1 month and levels then remain unchanged up to 1 year. Amylin, co-secreted with insulin, not only reflects β-cell function, but may also exert an effect on insulin secretion and PPG metabolism.
There are few studies reporting changes in insulin resistance after bypass surgery. Our data show that insulin resistance, estimated by HOMA-IR, decreased at 1 year, when the rate of weight loss is minimal, similar to other studies using clamps.19 In addition to weight loss, factors such as changes in glucose transporters,20 incretins, or adipokines,18,21 and increased intestinal gluconeogenesis, as shown in a rodent model,22 could explain improved long-term insulin sensitivity. In the present study, the increase in adiponectin levels was significant only between 6 and 12 months after surgery. There are several instances in which total adiponectin levels do not change, yet the distribution of the different adiponectin complexes is altered. It will be of interest to see whether the different complexes respond differentially to GBP. Measurement of the different complexes tends to give a more precise indication of general metabolic improvements in response to an intervention.23
The hyperglucagonemia of diabetes decreased with diet-induced weight loss.24 Curiously, our data show that fasting glucagon levels did not change 1 month after GBP. Furthermore, glucagon levels increased more in response to oral glucose 1 month after GBP. The mechanism of this paradoxical hyperglucagonemia 1 month after GBP is unclear and appears independent of GLP-1, GIP, and/or weight loss. It could be related to acute neuronal changes after GBP. At 6 and 12 months, although fasting glucagon levels decrease with further weight loss, glucagon levels are still not suppressed by oral glucose.
In summary, our data suggest a relative role of incretin changes and weight loss on islet cell function and glucose control. Our data show that the weight loss effect seems predominant on fasting glucose and insulin levels, whereas the surgical effect is predominant on incretin changes. In addition, we have shown that the incretin effect, independent of weight loss, may explain the shift of insulin secretion towards the early phase and lower glucose levels at 120 min. Understanding the mechanisms of rapid and sustained diabetes remission after GBP could lead to the development of new therapeutic approaches and newer, less invasive surgical techniques to treat people with diabetes.
This work was funded by grants from the American Diabetes Association (7-05 CR-18), National Institutes of Health (R01-DK67561, 1 UL1 RR024156-03, ORC DK-26687, and DERC DK-63068-05), the Merck Investigator Initiated Studies Program, and the Amylin Investigator Initiated Studies Program. The authors thank the volunteer participants and Yim DAM and Ping ZHOU (New York Obesity Research Center Hormonal Core Laboratory) for their technical help with the hormonal assays.
The authors have nothing to declare.