The Sum of Many Parts: Potential Mechanisms for Improvement in Glucose Homeostasis After Bariatric Surgery

Abstract

Bariatric surgery has emerged as the most durably effective treatment of type 2 diabetes (DM). However, the mechanisms governing improvement in glucose homeostasis have yet to be fully elucidated. In this review we discuss the various types of surgical interventions and the multitude of factors that potentially mediate the effects on glycemia, such as altered delivery of nutrients to the distal ileum, duodenal exclusion, gut hormone changes, bile acid reabsorption, and amino acid metabolism. Accumulating evidence that some of these changes seem to be independent of weight loss questions the rationale of using body mass index as the major indication for surgery in diabetic patients. Understanding the complex mechanisms and interactions underlying improved glycemic control could lead to novel therapeutic targets and would also allow for greater individualization of therapy and optimization of surgical outcomes.

Introduction

Type 2 diabetes mellitus (DM) is a chronic and progressive disease marked by insulin resistance and eventual reduction in insulin secretion [1, 2]. Despite an overall improvement in diabetes care from 1999 to 2010, 47.5 % of diabetic patients still do not meet the targets for glycemic control [3, 4••].

Weight gain and obesity are clearly established causes of type 2 diabetes (DM) [5]. The rise in obesity has been associated with an increase in the prevalence of diabetes [6]. Weight loss improves glycemic control in obese patients with DM, but conventional medical management often fails to achieve sustained weight loss in severely obese patients [5]. Bariatric surgery has emerged as one of the most effective methods of achieving sustained weight loss and inducing long-term improvement in DM and its comorbidities [7, 8••, 9••, 10, 11]. Multiple studies have demonstrated the greater efficacy of bariatric surgery compared with medical therapy in achieving the treatment goals recommended by the American Diabetes Association (ADA) [8••, 9••, 12••, 13–15, 16••]. While bariatric surgery is currently reserved for patients with BMI higher than 40 kg/m² or higher than 35 kg/m² with significant comorbidities, the International Diabetes Federation has recently recommended consideration of bariatric surgery for patients with a BMI of 30–35 kg/m² when traditional medical management is unable to achieve adequate diabetic control [17••]. This review will discuss the impact of bariatric surgery on DM and potential mechanisms for improved glycemic control.

Effect of Bariatric Surgery on Glycemic Control

Rapid improvement in DM following bariatric surgery has been described anecdotally as early as the 1970s [18]. In 1995, Pories et al reported the long-term efficacy of bariatric surgery in achieving diabetes “remission” in 608 morbidly obese patients who underwent gastric bypass, with a 96.3 % follow-up rate over 14 years [19]. A later meta-analysis reported that DM was completely resolved in 76.8% of patients, however, a more recent analysis of randomized controlled trials indicates that figures may be lower [7, 20].

Dixon et al published the first randomized trial to compare the effect on DM of surgical vs medical weight loss interventions. Sixty patients with BMI >30 kg/m² but <40 kg/m² with recently diagnosed DM(less than 2 years) were randomized to laparoscopic adjustable gastric banding (LAGB) vs medical intervention. Of the 92 % patients who completed the 2 year follow-up, 22/29 (75.8 %) in the surgical group vs 4/26 (15%) in the conventional group achieved diabetes remission, as defined by fasting glucose <126 mg/dL and hemoglobin A1c (HbA1c) ≤6.2 % after discontinuation of medications. Remission was associated with amount of weight loss and lower baseline HbA1c levels [15].

More recently, several groups have published randomized controlled trials comparing bariatric surgery to medical management of DM. In a prospective randomized trial, Mingrone et al compared Roux-en-Y gastric bypass (RYGB) and biliopancreatic diversion (BPD) to medical therapy in obese patients (BMI≥35 kg/m²) with DM. At 2 years, achievement of fasting glucose <100 mg/dL and HbA1c <6.5 % after stopping medications, did not occur in any medically treated patients, but occurred in 75 % of the RYGB and 95 % of the BPD patients. There was no association of diabetes remission with age, BMI, sex, or duration of DM [8••]. Schauer et al randomized 150 patients with DM to either medical treatment per ADA guidelines alone or medical treatment plus RYGB or sleeve gastrectomy. One year later, only 12% of the medically treated patients vs 42 % RYGB and 37 % sleeve gastrectomy patients achieved diabetes remission with HbA1c <6 % as defined by the study [9••]. The Diabetes Surgery Study, a prospective, multi-center, and multinational trial randomized 120 obese patients (BMI 30–39.9 kg/m²) with mean HbA1c of 9.6% to intensive lifestyle modification or to a combination of RYGB with intensive lifestyle modification. At 1 year, weight loss was 7.9 % of initial body weight and HbA1c was 7.8 % in the medical arm; in the surgery arm, weight loss and HbA1c were 26.1% and 6.3%, respectively. In regression analyses, improved glycemia was mostly driven by the degree of weight loss.

The variability in remission rate among earlier studies and these more recent studies likely reflect differences in study design and surgical as well as lifestyle interventions, baseline characteristics of patients, duration of follow-up, and the criteria used to define remission [21••, 22]. When the ADA definition of remission (fasting glucose <100 mg/dL and HbA1c <6 % of at least 1 year’s duration in the absence of active pharmacologic therapy) [21••] was retrospectively applied to data collected prospectively in 3 bariatric centers, the total rate of complete remission was 34.4 % for all bariatric surgery, which included RYGB, sleeve gastrectomy, and gastric banding; for RYGB, the remission rate by ADA criteria was 40.6 % [21••, 22]. Preoperative BMI did not correlate with diabetic control postoperatively. Although numbers are relatively low, it appears that outcome is similar in individuals with BMI <35 kg/m². The efficacy of bariatric surgery in achieving “remission” is further complicated by the difficulty of defining “remission” in diabetes, as hyperglycemia exists on a continuum. Also implicit within the definition of remission is the possibility of recurrence, and the period of time afterwhich remission may be considered effectively a cure has been arbitrarily chosen [21••]. A proposed model (DiaRem) for the prediction of diabetes remission gave greatest weight to preoperative use of insulin and has not been validated prospectively [23].

The durability of the surgical effect on glycemic control is less clear due to lack of long-term studies, and in particular with newer procedures such as sleeve gastrectomy. The Swedish Obese Subjects (SOS) survey was the first long-term, prospective, nonrandomized controlled trial to compare bariatric surgery vs conventional medical treatment on diabetes. After 2 years of follow-up, 72 % of the patients with diabetes at baseline were in remission (defined by cessation of diabetes medications or fasting plasma glucose <126 mg/dL) after gastric banding, vertical banded gastroplasty, or RYGB; however, 50 % of the diabetic patients in remission at 2 years had relapsed after 10 years [11, 16••]. Similarly, other groups have found recurrence rates ranging from 19 % to 35.1 % in bariatric surgery patients who had initially achieved remission, defined variably as HbA1c <6 % and fasting plasma glucose ≤100 mg/dL or ≤124 mg/dL with cessation of medications, after follow-up ranging from 3 to 9 years [24–26]. The risk of diabetes recurrence appeared to be inversely correlated with long term excess weight loss (EWL) and directly correlated to duration of diabetes and severity of diabetes preoperatively [24–26].

Proposed Mechanisms for Improvement in Diabetes

Weight loss through bariatric surgery improves DM more effectively than conventional medical therapy, but the mechanisms by which this improvement occurs have yet to be fully elucidated. The changes in glucose homeostasis are likely the result of both weight loss dependent and weight loss independent mechanisms, and the following mechanisms are not necessarily mutually exclusive of one another.

Caloric Restriction

It is well known that significant caloric restriction improves glucose tolerance in diabetic patients [27]. Bariatric surgery has been associated with profound improvements in fasting glucose concentration and insulin action, which often occur early in the postoperative course, before significant weight loss has even occurred [28–30]. Caloric restriction, which occurs immediately after all surgical interventions, likely plays an important role in the favorable metabolic changes observed after bariatric surgery. For example, obese diabetic patients who underwent either RYGB or a 500 kcal/day diet lost an equivalent amount of weight over approximately 3 weeks and exhibited similar improvements in insulin sensitivity, acute insulin secretion, and beta cell function when assessed by an intravenous glucose challenge [31•].

Incretin Secretion and Response

Changes in the secretion pattern of gut-derived hormones from enteroendocrine cells in response to altered nutrient transit have emerged as important potential mechanisms for improvement in DM.

Due to impaired insulin secretion in DM, the incretin effect, which is the greater insulin response after oral glucose compared with an equivalent dose of intravenous glucose, is diminished [32]. Glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) are the 2 incretins responsible for one-half to two-thirds of postprandial insulin secretion. GLP-1 is secreted mostly from ileal L cells while GIP is secreted from duodenal K cells. Both are rapidly inactivated by the enzyme dipeptidyl peptidase IV (DPP-IV). GLP-1 analogues and DPP-IV inhibitors are currently used as antidiabetic agents, stressing the importance of GLP-1 in glucose homeostasis [33, 34].

Multiple studies have shown that postprandial secretion of GLP-1 is substantially and durably increased after RYGB, but not after LAGB or diet-induced weight loss [35–40]. The increase in GLP-1 occurs early following RYGB, before significant weight loss has occurred [38, 39]. Moreover, the postprandial increase in GLP-1 is clearly related to the alteration in gastrointestinal anatomy following RYGB, as illustrated in 2 case reports examining the effect of peroral feeding vs gastroduodenal feeding in RYGB patients with gastric tubes inserted into the bypassed gastric remnant [41, 42]. When RYGB patients were fed via the gastric bypass pathway, exaggerated postprandial GLP-1 and insulin responses were noted. However, when patients were fed through the gastric tube via the gastroduodenal pathway, GLP-1 and insulin secretion were similar to the pattern seen preoperatively. These cases demonstrate that more rapid delivery of nutrients to the distal small intestine is a key factor in the postoperative changes in GLP-1 and insulin secretion. Greater improvements in postprandial GLP-1 secretion and the incretin effect after bypass surgery, independent of weight loss, were also demonstrated in a study of obese diabetic women studied 1 month after being randomized to gastric bypass surgery or diet-induced equivalent weight loss [35].

Early recovery of beta cell function has been associated with exaggerated postprandial secretion ofGLP-1 after RYGB but causality has not been definitively established. Jorgensen et al expanded upon the earlier study by Salehi et al [43] to illustrate the insulinotropic role of GLP-1 in diabetic patients after RYGB. By pharmacologically blocking the GLP-1 receptor (GLP-1R) during a liquid meal tolerance test after surgery using exendin (9–39), they were able to show that improvement in beta cell glucose sensitivity, glucagon suppression, and insulin secretion were all decreased [44••]. However, Jimenez et al found only minimal decrease in glucose tolerance following blockade ofGLP-1 action [45]. GLP-1 and the incretin effect, while important to improved beta cell function, are likely not the only factors altered by bariatric surgery.

Reported postprandial changes in GIP following bariatric surgery have been inconsistent. In theory, exclusion of the duodenum, where most GIP-producing K cells reside, would lead to decrease in GIP secretion after RYGB. Instead, some studies have found an increase in GIP while others have noted lack of change or a decline [35, 37, 46, 47]. It is possible that the inconsistencies in postprandial GIP levels postoperatively are due to variability in surgical technique, particularly differences in the length of Roux anastomoses, composition of test meal stimulus, and the timing of blood sample collections that may overlook early postprandial secretion.

Glucagon and Glucagon-Like Peptide 2

An increase in meal-stimulated glucagon in the early postprandial period after RYGB has been noted in several studies and argues against GLP-1 mediated suppression of glucagon as a mechanism for improved glucose homeostasis after bariatric surgery [35, 46, 48]. However, fasting glucagon appears to decrease by 1 year after RYGB and may potentially explain some of the long-term benefits of RYGB on glucose metabolism [48].

Glucagon-like peptide 2 (GLP-2) is a member of the enteroglucagon family derived from the preproglucagon gene, which also encodes glicentin, glucagon, oxyntomodulin, and GLP-1. GLP-2 regulates gastric motility, gastric acid secretion, intestinal hexose transport, glucagon secretion, and increases absorptive area and barrier function of the gut epithelium via stimulation of crypt cell proliferation and inhibition of apoptosis in the enterocyte and crypt compartments [49–51]. In both murine and human patients, le Roux et al found that GLP-2 levels rose substantially after RYGB, peaked at 6–12 months and correlated with the period of maximal weight loss, before returning to baseline levels. Following RYGB in rats, increased GLP-2 levels were associated with increased crypt cell proliferation as early as 23 days postoperatively [52]. These accommodations in gut anatomy likely account for how well RYGB is tolerated and the relatively minimal degree of macronutrient malabsorption, but the precise role, if any, that GLP-2 may play in the improvement of glucose homeostasis following RYGB remains to be seen.

Changes in Appetitive Hormones

GLP-1 and Peptide YY (PYY) are anorexigenic and are implicated in the decreased hunger and increased satiety that often follows RYGB [53]. Consistent with this notion is the demonstration that GLP-1R agonism enhances weight loss achieved by adjustable gastric banding (AGB) in obese rats [54]. Like GLP-1, PYY is released from intestinal L cells in response to a meal and has anorectic properties, delaying gastric emptying, slowing intestinal transit time, and increasing satiety. PYY levels are dramatically elevated following nutrient ingestion after RYGB but not after LAGB [36, 40, 55, 56]. The marked postprandial elevations in GLP-1 and PYY after bariatric surgery suggest that gut hormone mediated changes in appetite, in addition to improved insulin sensitivity and secretion, may facilitate decreased food consumption and thereby contribute to the attainment and maintenance of improved glycemic control.

Compromised secretion of ghrelin, an orexigenic peptide produced in the fund us and body of the stomach, has also emerged as a potential mechanism for both decreased hunger and improvement in glycemia after RYGB and sleeve gastrectomy [57–59]. Various studies have reported decreases in fasting and postprandial ghrelin levels after RYGB, while others have noted no change or even increases in ghrelin [46, 55, 60–63]. These seemingly discrepant results may be due in part to the postoperative interval, amount of weight loss, the time-points evaluated (single fasting vs meal-associated), and the use of assays measuring “active” octanoylated iso form vs total ghrelin.

Changes in the Direction and Rate of Nutrient Flow

A proposed mechanism for the improvement in DM following bariatric surgery is that early and rapid delivery of unabsorbed nutrients to the distal small intestine activates the “ileal brake” that potentiates the secretion of GLP-1 and PYY (the “hindgut hypothesis”). Consistent with this hypothesis is the observation that procedures with the most consistent improvement in DM shorten the route of nutrient flow from the stomach to intestine, and/or increase the rate of transport of ingested nutrients [14, 58, 64–66].

Further evidence for the role of the distal ileum in improving glucose tolerance is derived from rodent models of ileal interposition (IT), a surgical procedure whereby a segment of ileum is inserted into the proximal small intestine so that there is no gastric restriction or duodenal bypass. IT in different diabetic rodent models results in elevated levels of GLP-1 and PYY and improvements in insulin sensitivity, glucose tolerance, and beta-cell function [67–70]. Inhibition of GLP-1R with exendin (9–39) reversed the improvement in oral glucose tolerance after IT, lending further support to the hypothesis that early and increased activation of GLP-1R leads to subsequent improvement in glucose metabolism [71]. In humans, IT associated with sleeve gastrectomy has shown some promise for the treatment of T2DM in patients with BMI of 21– 34 kg/m² [72, 73••].

Exclusion of nutrients from the proximal small intestine (the “foregut hypothesis”) has also been proposed to restore euglycemia. In animal studies of both non obese and diet induced obese diabetic rats, duodenal-jejunal bypass (DJB) without gastric restriction led to improvement in hyperglycemia independent of food intake and weight reduction [74–76]. Intestinal glucoregulatory hormones and vagal innervation may also contribute to alleviation of hyperglycemia in the DJB model [76]. DJB in humans had only a moderate effect on glucose homeostasis in patients with DM resulting in a change in mean HbA1c from 9.3 % at baseline to a nadir of 6.5% at 3 months followed by a progressive increase to 7.7% at 12 months postsurgery [77••]. It is unclear if the improvements observed were due to altering the intestinal site of delivery of ingested nutrients and/or moderate weight loss.

Novel, less invasive endoscopic therapies are also being explored for the treatment of DM. By using a rat model of the endoluminal sleeve (ELS) to mimic 2 components of RYGB, bypass of the proximal intestine and early exposure of the jejunum to partially digested nutrients, Aguirre et al demonstrated that despite less weight loss than similar rats that had undergone RYGB, rats with the ELS attained comparable improvement in glucose homeostasis [78]. In humans, reductions in HbA1c have been demonstrated after implantation of various types of endoluminal duodenal-jejunal bypass sleeves [79, 80••]. Further investigations into the safety, efficacy, and long-term outcomes of the ELS are ongoing.

“Hindgut” and “foregut” hypotheses have both been proposed as mechanisms that alter glucose homeostasis [64, 81]. Current evidence suggests that multiple factors are at play that are not mutually exclusive, and may in fact act in concert, to achieve the striking improvements observed using different devices and surgical approaches.

Other Mechanisms for Improvement in Diabetes

Adipose Tissue and Adipokines

Weight loss following RYGB consists of a significant reduction in whole body fat including 50 %–60 % reduction in visceral adipose tissue [82]. Adiponectin, a 244 amino acid protein secreted from white adipose tissue, is inversely proportional to body fat mass, BMI, waist-to-hip ratio, serum insulin, and glucose levels; low levels are closely correlated with insulin resistance and suspected to play a role in DM pathogenesis [83]. Most studies have shown an increase in adiponectin after RYGB, which correlates with the improvement in insulin resistance following surgery [60, 63]. Changes in body composition, inflammation, and adipokines contribute to improvement but most are likely not unique to surgery-induced, as opposed to diet-induced, weight loss.

Bile Acids and FGF 19

Increased bile acid reabsorption has also been proposed as a potential mechanism for improved insulin sensitivity after RYGB [84]. Based on rodent studies suggesting that bile acids increase energy expenditure through activation of G-protein coupled receptor TGR5 and thereby type 2 thyroid hormone deiodinase, Patti et al performed a cross-sectional analysis of fasting serum bile acid composition and various metabolic variables in a group of nondiabetic post-RYGB, matched obese, and overweight patients. Total serum bile acid concentrations were 2-fold higher in post-RYGB patients and were inversely correlated with 2-hour postprandial glucose levels as well as fasting triglycerides, and positively correlated with adiponectin and peak postprandial GLP-1 [84]. The altered enterohepatic recycling of bile acids and consequently elevated serum bile acid levels induced by IT in rats may also contribute to improved glucose homeostasis [85••, 86]. Bile acids also act on the farnesoid X receptor stimulating secretion of fibroblast growth factor (FGF)19 secretion with subsequent inhibition of hepatic gluconeogenesis. In prospective human and animal studies, Pournaras et al concluded that RYGB but not LAGB leads to more rapid delivery of bile acids to the terminal ileum, which in turn causes an increase in total bile acid concentrations, plasma GLP-1, PYY and FGF19 [87]. Other studies in humans have also shown increases in plasma bile acids and FGF19 after RYGB [88–90].

Microbiota

Data from mouse studies and recent small studies in humans have provided evidence that gut microbiota may play an important role in energy storage and possibly the development of obesity and associated complications. It remains unclear whether changes in the gut microbiome are a consequence or cause of obesity in humans. Bariatric surgery, in particular, induces environmental, systemic, and anatomic changes that may all impact the composition of gut microbiota. Initial studies have suggested a shift in bacterial flora after surgery toward profiles more similar to that of lean patients [91–94]. Gastrointestinal reconfiguration may play a key role in the changes in microbiota, as changes in the microbial ecology were most notable distal to the site of surgical manipulation in rat models. Moreover, transfer of gut microbiota from RYGB treated mice to non-operated, aseptic mice was sufficient to cause decreased weight and adiposity, possibly due to altered microbial synthesis of short-chain fatty acids [94]. Whether these changes in microbiota or those involving other digestive processes such as the thermic effect of feeding contribute to improved glucose homeostasis remains to be elucidated.

Genetics

Various studies have begun to examine the potential of obesity associated single nucleotide polymorphisms (SNPs) to predict weight loss outcomes following bariatric surgery. Higher pre and postoperative BMI following RYGB has been associated with an increased number of obesity SNPs or homozygous SNP genotypes [95]. The SOS group looked at 11 obesity candidate genes and found that the FTO SNP rs16945088 was associated with maximal weight loss [95, 96].

The melanocortin system is an integral component in the regulation of energy homeostasis. Loss of a single functional copy of the melanocrtin-4 receptor (MC4R) gene is the most common mutation associated with obesity in humans. Mirshahi et al found that carriers of the MC4R I251L allele were more likely to lose weight during dietary and surgical interventions and had less insulin resistance [97]. A later study provided further confirmation of the crucial role MC4R signaling plays in the weight loss effects of RYGB: while mice heterozygous for MC4R remained fully responsive to RYGB, MC4R null mice lost substantially less weight after surgery. By sequencing the MC4R gene in 972 patients undergoing RYGB, Hatoumet al also showed that a single normal copy of the MC4R gene was sufficient to conserve the weight loss effects of RYGB. However, given the few number of diabetics in the study, it is not clear if MC4R mutations have a clinically significant effect on glycemic outcome [98].

Genetic polymorphisms in the gene encoding transcription factor 7-like 2 (TCF7L2) are associated with an increased risk of developing DM. Inheritance of 2 specific TCF7L2 variants, rs1225372 and rs7903146, has been associated with relatively reduced insulin secretion and an increased probability of progression from impaired glucose tolerance to DM, but no impairment of insulin sensitivity [99]. The insulinotropic effect of exogenous GLP-1 in carriers of these variants has been reported as impaired in some but not all studies, so it is unclear if carriers would reap equivalent benefits from bariatric surgeries that enhance GLP-1 secretion [100–102].

More recently, fetal and adult ablation of the transcription factor Foxo1 in mouse enteroendocrine progenitors have given rise to the development of gut cells with the ability to produce biologically active insulin in a glucose-responsive manner. Unlike embryonic stem cell-derived insulin producing cells, the gut-derived insulin producing cells demonstrated a singular plasticity and ability to regenerate and produce insulin [103]. It is feasible that bariatric surgery, through changing hormonal and nutrient cues, may somehow influence insulin secretion from the gut through Foxo1 mediated alterations in Notch and Wnt signaling pathways, both of which have been shown to regulate gut and pancreatic cell differentiation.

Taste Receptors of the Gut

G-protein coupled taste receptors detect gut luminal contents and transmit signals that regulate nutrient transporter expression and nutrient uptake, as well as the release of gut hormones and neurotransmitters involved in the regulation of energy and glucose homeostasis. Sweet taste receptors, TAS1R2, have been shown to be dysregulated in DM and may potentially increase the risk of postprandial hyperglycemia by increasing glucose absorption via Na⁺/glucose co-transporter SGLT1 during hyperglycemia. Rapid delivery of undigested nutrients to the lower small intestine after RYGB may also affect the regulation of taste receptors or glucose transporters on L cells, leading to increased PYY and GLP-1 secretion. In rodent models, it has been demonstrated that DJB leads to decreased TAS1R2 and TAS1R3 expression in the alimentary limb and decreased SGLT1-mediated glucose transport. Given the decreased preference for sweet and fatty foods demonstrated by many patients after RYGB, it has been postulated that increased intensity of sweet perception following surgery leads to changes in gut hormones and ultimately alterations in the reward circuitry and energy homeostasis [104].

Insulin Clearance

The rapid decrease in fasting glucose and fasting insulin after RYGB suggests that early improvement in hepatic glucose production and hepatic insulin sensitivity may account for this change. Bojsen-Moller et al prospectively studied 32 diabetic and 32 normoglycemic patients recruited for RYGB and found that fasting hepatic insulin clearance increased after 1 week and further at 3 months for both groups, regardless of diabetic status [105]. Postprandial insulin clearance increased only in the DM patients, in whom the increase was noted as early as 1 week following surgery and maintained at 3 months and 1 year. Based on this study and others that have noted the rapid improvement in fasting glucose and fasting insulin before significant weight has even occurred, it is quite plausible that nonenteral factors, such as increased hepatic insulin sensitivity and insulin clearance, may be responsible for some of the early improvement in response to energy restriction [106]. Early postoperative increases in plasma free fatty acids are then followed by increased suppression that likely contributes to improved peripheral insulin sensitivity later on as demonstrated during clamp studies in diabetic patients 1 year after RYGB [107].

Intestinal Gluconeogenesis

In mice, enterogastric anastomosis (EGA, analogous to RYGB in humans) but not gastric banding increased gastrointestinal gluconeogenesis. Following EGA, researchers noted decreased food intake in the mice and direct secretion of enteral glucose into the portal vein, suppressing endogenous hepatic glucose production. They illustrated the importance of hepato-portal glucose sensing in mediating appetite by performing EGA in a GLUT-2 knockout mouse and confirming that without the specific glucose sensor, there was no inhibition of food intake in the mice even after EGA. The authors linked increased intestinal gluconeogenesis to improved hepatic insulin sensitivity with subsequent suppression of hepatic gluconeogenesis and improved glucose tolerance overall, asserting that this occurred independently of GLP-1 and is only partially explained by weight loss [108].

Amino Acids

Dietary macronutrients function as signaling molecules to affect feeding behavior, fuel efficiency, enteric hormones, insulin release and insulin sensitivity [109–111]. Obesity and DM are known to alter circulating concentrations of many metabolites [111, 112]. Metabolite profiling in diabetic patients has shown that weight loss after RYGB but not equivalent diet-induced weight loss is associated with a decrease in fasting plasma concentrations of branched chain amino acids and their C3 and C5 acylcarnitine metabolites and correlated negatively with insulin sensitivity [113]. Similar changes were also noted after AGB or RYGB in a nondiabetic group after equivalent weight loss [114••]. It is, therefore, unclear whether the changes in amino acid metabolism, and possibly transport, are unique to RYGB or are causal to the improvements in insulin action. The use of metabolomics will undoubtedly further this area of study [115].

Conclusions

The mechanisms by which bariatric surgeries cause improvement in DM are multifactorial and dependent upon both weight loss associated and weight loss independent factors. Expedited delivery of nutrients to the distal gut, as well as duodenal exclusion and hormonal changes, likely contribute to the improvement in glucose homeostasis. Given the profound metabolic effects of bariatric surgery and the known long-term benefits of glycemic control, it may be worthwhile to reconsider the applicability of “remission” in defining successful treatment of diabetes, and to advocate for improved glycemic control and decreased use of medications as better parameters of treatment efficacy. Evidence indicating that some of the positive outcomes are at least partially independent of weight loss renders use of BMI as a major criteria for surgery in diabetics somewhat arbitrary. Improved understanding of the complex mechanisms and interactions involved in glucose homeostasis after bariatric surgery could identify potential therapeutic targets including less invasive techniques, optimize surgical outcomes, and also encourage greater individualization of therapy.