Endocrine mechanisms mediating remission of diabetes after gastric bypass surgery

Transcription

1 (2009) 33, S33 S40 & 2009 Macmillan Publishers Limited All rights reserved /09 $ REVIEW Endocrine mechanisms mediating remission of diabetes after gastric bypass surgery 1,2,3 1 University of Washington School of Medicine, Seattle, WA, USA; 2 Veterans Affairs Puget Sound Health Care System, Seattle, WA, USA and 3 Diabetes & Obesity Center of Excellence, Seattle, WA USA Bariatric surgery is currently the most effective method to promote major, sustained weight loss. Roux-en-Y gastric bypass (RYGB), the most commonly performed bariatric operation, ameliorates virtually all obesity-related comorbid conditions, the most impressive being a dramatic resolution of type 2 diabetes mellitus (T2DM). After RYGB, 84% of patients with T2DM experience complete remission of this disease, and virtually all have improved glycemic control. Increasing evidence indicates that the impact of RYGB on T2DM cannot be explained by the effects of weight loss and reduced energy intake alone. Weightindependent antidiabetic actions of RYGB are apparent because of the very rapid resolution of T2DM (before weight loss occurs), the greater improvement of glucose homeostasis after RYGB than after an equivalent weight loss from other means, and the occasional development of very late-onset, pancreatic b-cell hyperfunction. Several mechanisms probably mediate the direct antidiabetic impact of RYGB, including enhanced nutrient stimulation of L-cell peptides (for example, GLP-1) from the lower intestine, intriguing but still uncharacterized phenomena related to exclusion of the upper intestine from contact with ingested nutrients, compromised ghrelin secretion, and very probably other effects that have yet to be discovered. Research designed to prioritize these mechanisms and identify potential additional mechanisms promises to help optimize surgical design and might also reveal novel pharmaceutical targets for diabetes treatment. (2009) 33, S33 S40; doi: /ijo Keywords: diabetes; gastric bypass; bariatric surgery; ghrelin; GLP-1 Bariatric operations traditionally have been thought to cause weight loss through gastric restriction and/or intestinal malabsorption. Restrictive operations such as adjustable gastric banding (AGB) create a small pouch in the proximal stomach to reduce functional gastric capacity and retard gastric emptying. Malabsorptive procedures such as biliopancreatic diversion (BPD) leave much of the stomach intact but divert ingested food from the stomach to the ileum; thus, only a small segment of the distal bowel can absorb nutrients. The typical proximal Roux-en-Y gastric bypass (RYGB) creates gastric restriction and also causes food to bypass most of the stomach and a short segment of the proximal intestine, primarily the duodenum, leaving enough intact small bowel in digestive continuity to avoid clinically significant malabsorption. RYGB causes weight loss through gastric restriction and additional mechanisms involving changes in gut hormones and other processes under active investigation RYGB and BPD produce the Correspondence: Dr, Diabetes & Obesity Center of Excellence, University of Washington, Box (mail stop 111), Seattle, WA 98195, USA. davidec@u.washington.edu highest and most rapid rates of remission of type 2 diabetes mellitus (T2DM). Remission of diabetes after RYGB Although bariatric surgery was designed to facilitate weight loss, anecdotal reports of rapid postoperative remission of T2DM emerged as early as the 1970s. In 1995, Pories et al. 4 reported long term results from 608 severely obese patients who underwent gastric bypass, with a 93% follow-up rate over 14 years. Beyond achieving substantial, durable weight loss (37% of total body weight at 1 year and 30% at 14 years), these individuals experienced remarkable remission of T2DM. Of the 146 patients with T2DM at baseline, 83% were normoglycemic without diabetes medications at the time of publication. Most interestingly, glucose levels often normalized within just a few days after surgery, before any significant weight loss had occurred. Subsequent research has shown very similar results. A meta-analysis of 136 bariatric surgery studies including individuals confirmed an overall 84% remission of T2DM following RYGB. 3 The Swedish obese subjects study,

2 S34 a prospective, contemporaneously matched, multicenter trial of bariatric surgery vs medical care for obesity, reported outcomes after 2 years in 4047 patients, and at 10 years in 1703 patients, with outstanding retention rates. 1 At 2 years, no postsurgical patients had developed T2DM, whereas 5% of the medical control group had. At 10 years, the risk of developing T2DM was 43 times lower for surgically treated patients, whereas recovery from T2DM was 43 times more common. In these studies, the relatively small number of RYGB patients who did not achieve complete remission of T2DM were primarily characterized by a longer duration of disease, suggesting that they may have had insufficient residual b-cell mass to achieve euglycemia following surgery. At least seven published reports show significantly decreased overall mortality after RYGB, most notably two large, multicenter studies recently published in the New England Journal of Medicine. 2,15 One of these reported a remarkable 92% decrease in diabetes-related deaths after RYGB. 15 Although bariatric procedures have traditionally been reserved for patients with BMI 440 kg/m 2, or 435 kg/m 2 with significant comorbidities, the remarkable control of diabetes after RYGB raises the possibility that this operation could be a therapeutic option for less obese patients. Trials in people with BMI o35 kg/m 2 have reported complete remission of T2DM in a similar or even higher percentage of cases as in severely obese patients undergoing RYGB. 16,17 Given that less obese individuals lose less absolute and proportionate body weight after RYGB than do severely obese persons, the similarity in their degree of T2DM remission hints that the operation might exert anti-diabetic effects beyond those related to weight loss. Consistent with this concept, experimental studies indicate that variations of RYGB improve T2DM not only in obese but also in nonobese diabetic animals Despite a growing interest in using RYGB to treat T2DM, even among less obese or non-obese patients, the underlying mechanisms of action of this procedure have not been fully evaluated. As gastrointestinal (GI) surgery appears to eliminate T2DM more effectively, rapidly and durably than any existing non-surgical method, and because the usage of bariatric surgery is rapidly increasing worldwide, elucidating the mechanisms by which these operations improve glycemia is an important research priority. Such insights could lead to new pharmaceuticals that might mimic some of the dramatic anti-diabetic effects of surgery, without the need for a surgical procedure. Mechanisms of diabetes improvement following bariatric surgery The mechanisms by which bariatric procedures improve T2DM are yet to be fully determined. Earlier human studies of RYGB have reported the following findings (reviewed in Cummings et al. 10,14 ). The insulin sensitivity index, determined with intravenous glucose tolerance testing and minimal modeling, increases by 4 5 fold. This is associated with increases in the insulin-sensitizing hormone adiponectin, especially the important high molecular weight form, which rises in proportion to the amount of decrease in fat mass and predicts the degree of improvement in insulin sensitivity estimated by homeostatic model assessment. In muscle, the concentration of insulin receptors increases, as does the expression of the mitochondrial transcription cofactor PGC-1 and its target MFN2. As both insulin signaling and PCG-1 activity in muscle stimulate fatty acid metabolism, these changes would be predicted to reduce intramyocellular lipids. Indeed, lipid content decreases in muscle after RYGB, as it does in the liver. Lipid accumulation in both muscle and liver, which are the two biggest contributors to blood glucose levels, causes insulin resistance; hence, a fall in intracellular lipids in these key tissues increases insulin sensitivity. Complementary observations have been made in humans following the malabsorptive bariatric procedure BPD, which has been studied more rigorously to date than RYGB in terms of mechanisms underlying glycemic improvement (reviewed in Cummings et al. 10,14 ). After BPD, favorable changes occur in genes regulating fatty acid and glucose metabolism in muscle. Also increased are insulin-induced glucose oxidation and non-oxidative glucose disposal. This is associated with the expected increase in muscle glucose uptake, as well as whole-body glucose uptake, a measure of overall insulin sensitivity. Insulin sensitivity is improved by 6 months following BPD and remains durably enhanced, showing no further improvements at 10 years after the operation than at 2 years. A critical caveat about all of these observations is that they were made many months to several years following surgery, at which point profound weight loss had already occurred. As most or all of these findings are expected consequences of weight loss achieved by any means, it is not known whether they represent the secondary sequelae of massive weight loss or the direct anti-diabetic effects of the GI operations per se. Evidence for weight-independent anti-diabetic effects of RYGB The following lines of evidence indicate that RYGB improves glycemic control and causes remission of T2DM through mechanisms beyond those related to weight loss and reduced energy intake. Rapid kinetics of diabetes improvement after RYGB T2DM typically resolves within several days to weeks following RYGB, long before substantial weight loss has occurred. 4,5 In the landmark study, by Schauer et al. 5 of 1160 patients undergoing RYGB, among the 240 with T2DM severe enough to be treated with oral medications and/or

3 insulin, 30% were discharged from their initial surgical hospitalization with normal plasma glucose levels off all diabetes medicinesfafter an average hospital stay of only 2.8 days. Most of the remainder discontinued their diabetes treatments within a few weeks, and the eventual complete remission rate was 83%. Similar rapid resolution of T2DM following RYGB has been observed by many investigators and clinicians, and the rate of in-hospital remission (that is, within a few days after the operation) has been reported to be as high as 89%. 22 In contrast, purely gastric-restrictive operations such as AGBFwhich, similar to RYGB, involve peri-operative restriction of energy intake followed by longterm weight lossfameliorate diabetes only after substantial weight loss is achieved. 6,23 One study, in which glucose homeostasis was examined very soon after RYGB, reported that insulin sensitivity (estimated by homeostatic model assessment and the AUC (area-under-the-curve) glucose after an intravenous glucose tolerance testing) was significantly increased at only 6 days following surgery, before any meaningful weight loss had occurred. 22 In that study, the improvement in glucose tolerance appeared to result from reduced basal glucose levels before the i.v. glucose load, suggesting a primary improvement in hepatic insulin sensitivity. On the other hand, some investigators argue that RYGB ameliorates diabetes more through increased insulin secretion. 24 In summary, glucose homeostasis improves very rapidly after RYGB, often causing complete T2DM remission within a few days, long before weight loss occurs. The mechanisms of this early improvement are unknown, including whether they primarily involve increased insulin secretion, sensitivity, or both. Greater glycemic improvement with RYGB than with equivalent weight loss from other means Undoubtedly, RYGB-induced weight loss contributes to diabetes improvement. The key question is whether glucose control improves more after RYGB than after other methods of weight loss, for a given amount of lost weight. Recently, several careful studies have addressed this question and yielded similar results, all indicating a weight-independent anti-diabetic effect of RYGB. LaFerrere et al. studied patients with T2DM who underwent either RYGB or dietary weight loss. Careful glucose-homeostasis phenotyping was carried out shortly before surgery and again postoperatively at an equivalent amount of weight loss (B9.5 kg in both groups). 25 The groups were exquisitely matched for preoperative severity and duration of diabetes, homeostatic model assessment values, oral glucose tolerance, magnitude of incretin effect, BMI, age and gender. After the weight-loss interventions, however, post-rygb patients displayed substantially better glucose tolerance and incretin effect than did patients who had lost the same amount of weight with dieting. In an analogous study of RYGB vs AGB, Pattou et al. 26 found similar results. Fifty obese patients with T2DM, matched for preoperative BMI and degree of impaired glucose tolerance on oral glucose tolerance tests, were studied before surgery and again postoperatively when they had lost exactly 10% of their body weight. Despite equivalent weight loss, the post-rygb group showed markedly better glucose tolerance by oral glucose tolerance testing than did the post-agb group. Similarly, leroux et al. 27 observed better glucose tolerance in people following RYGB than after equivalent weight loss from AGB, and the estimated insulin sensitivity of the post-rygb group was equivalent to that of lean counterparts, despite a significantly higher BMI in the post-surgical patients. Korner et al. 28 also observed better glucose tolerance following a standardized test meal in a cross-sectional analysis of patients who had undergone RYGB compared with AGB, with equal postoperative BMI values. Lee et al. randomized patients with less severe obesity (BMI kg/m 2 ) but poorly controlled T2DM (average hemoglobin A1c (HbA1c) 10% in both groups) to either sleeve gastrectomy or a sleeved version of RYGB, with a similar sized in-continuity gastric pouch after both operations (larger than in a standard RYGB). 29 The groups were matched preoperatively for BMI and HbA1c. Although RYGB ultimately causes greater and more durable weight loss than sleeve gastrectomy does, weight loss in this study was equivalent after both procedures for the first 6 months. At 6 months, however, with 15% weight loss in each group, HbA1c had fallen far more after RYGB than after AGB. Among RYGB patients, 93.3% had HbA1c values o7% at that time, compared with only 46.7% of the sleeve gastrectomy group, despite equivalent pre- and post-operative BMI values. In summary, glycemic control improves more after RYGB than after an equivalent magnitude of weight loss from dieting or purely gastric-restrictive bariatric operations, indicating an additional anti-diabetic impact of RYGB beyond just the consequences of weight loss, most likely related to the intestinal bypass component of the operation. Cases of severe hyperinsulinemic hypoglycemia developing late after RYGB Hints that RYGB can exert weight-independent effects on b-cell mass and/or function come from increasing reports of patients with life-threatening hyperinsulinemic hypoglycemia developing long after surgery. 11,30,31 In 2005, Service et al. 30 reported in the New England Journal of Medicine 11 the details of six such cases, which these authors deemed adultonset nesidioblastosis, and which were severe enough to require partial pancreatectomy. Several other groups have since reported similar observations At a recent meeting devoted to this topic at Harvard, B135 cases of severe, lateonset, post-rygb hyperinsulinemic hypoglycemia were discussed, many with intractable neuroglycopenic episodes necessitating pancreatectomy. An intuitive explanation for this phenomenon is that severely obese patients develop hypertrophy and/or S35

4 S36 hyperplasia of b cells over many years in response to insulin resistance, and with the insulin-sensitizing effects of RYGBinduced weight loss, a mismatch arises between b-cell mass/ function and the degree of insulin sensitivity. This hypothesis, however, predicts that hypoglycemia should occur soon after surgery, before b cells have regressed and during the period of dynamic weight lossfor at the latest, near the body-weight nadir (B1 year post-op), 1 when insulin sensitivity should be maximum. In contrast, the cases of hyperinsulinemic neuroglycopenia reported to date have developed very late, from 1 to 9 years after RYGB, typically at 2 4 years. By this time, maximum weight loss has generally already been achieved, some weight regain has begun (presumably with a degree of insulin resistance), and b cells would be expected to have fully adjusted to the postoperative milieu. The implication of these findings is that RYGB might cause long-lasting stimulation of b-cell mass and/or function, a phenomenon that probably benefits the vast majority of patients with diabetes but occasionally goes too far. Reports claiming that b-cell mass was increased in the pancreata of hypoglycemic post-rygb patients who required pancreatectomy fueled enthusiasm that RYGB might activate b-cell trophic factors. 30,31 Identifying these putative molecules could lead to new medical strategies to reverse the progressive b-cell failure that defines T2DM. Other investigators, however, claimed that pancreas samples from afflicted patients showed no increases in b-cell mass or proliferation beyond those expected from obesity, and therefore that the syndrome must arise from inappropriately increased b-cell function. 37 Still other investigators argue that the early glycemic effects of RYGB result from improved insulin sensitivity. 22 Indeed, the relative contributions from increased insulin secretion and/or sensitivity to the weightindependent anti-diabetic effects of RYGB are unclear. The question of whether or not RYGB stimulates b-cell growth remains controversial, and the debate is difficult to resolve in humans without pancreas samples from appropriately matched obese non-surgical patients. Regardless of the answer, a key point is that improvements in glucose tolerance never overshoot with non-surgical weight loss; thus, these observations constitute additional evidence favoring weightindependent glucose-lowering effects of RYGB. Theories regarding weight-independent anti-diabetic mechanisms of RYGB There are several plausible hypotheses to explain the rapid, weight-independent glycemic effects of RYGB None of these theories is necessarily exclusive of these others, and any or all them may be operational to some degree. patients are not permitted to eat much in the immediate postoperative period, and thus their residual b cells are only minimally challenged. After this major GI operation, most patients are initially deprived of all oral intake, followed by a period of consuming only clear liquids, then full liquids, and finally escalating to solid food. The ability of acute restriction of energy intake to transiently improve glycemia in T2DM is well known, and according to this model, by the time patients return to ad libitum eating, they begin to experience the insulin-sensitizing effects of dynamic weight loss from the bariatric operation. Although this hypothesis is undoubtedly valid to some degree, it fails to explain many observations about post- RYGB glycemic control. First, if this were the only important phenomenon, a similar degree of rapid diabetes remission would occur following all bariatric procedures. For example, AGB is also a major GI operation that is accompanied by a period of acute energy restriction followed shortly thereafter by progressive weight loss. After AGB, however, diabetes remits in only 48% of cases, compared with 84% of cases after RYGB. 3 More importantly, the pace of such remission is far slower after AGB than RYGB, and it is strongly linked to the timing and degree of weight loss. In a randomized trial by Dixon et al. of AGB vs best medical care to treat T2DM, none of the surgical patients had experienced remission of diabetes by 6 months after the operation, even though the disease ultimately did remit by 2 years in 73% of these individuals (who had very mild T2DM at baseline). 6,23 In marked contrast, most patients with T2DM who undergo RYGB experience full remission of their disease within several days to a few weeks after the operation (vide supra). 5,22 Patients undergoing many types of general surgery are subjected to immediate perioperative restriction of energy intake. However, rapid remission of T2DM is not observed in this setting; if anything, glycemic control tends to worsen after major surgery because of inflammation and upregulation of counter-regulatory stress hormones such as cortisol and catecholamines. The remarkable phenomenon of rapid postoperative dissipation of T2DM is unique to RYGB and certain other bariatric operations involving intestinal bypasses, such as BPD. Furthermore, the starvation-followedby-weight-loss hypothesis fails to explain the fact that glycemic control improves more after RYGB than after equivalent weight loss resulting from dieting or purely gastric-restrictive bariatric operations. Neither can this model account for the severe hyperinsulinemic hypoglycemia that sometimes develops late after RYGB. Hence, although the starvation theory is intuitive and has merit, it clearly does not account for the weight-independent antidiabetic impact of RYGB. The starvation followed by weight loss hypothesis A rather prosaic but important hypothesis to consider is that glycemic control improves soon after RYGB solely because The ghrelin hypothesis Our group provided the first evidence suggesting that compromised secretion of the orexigenic, pro-diabetic upper

5 GI hormone ghrelin might contribute to the anorexic and anti-diabetic effects of RYGB. 7 We found that human 24-h ghrelin profiles displayed marked preprandial surges followed by postprandial suppression, and that circulating levels increased in proportion to the amount of weight lost with dieting. These and other observations implicated ghrelin in mealtime hunger as well as in the adaptive increase in appetite that resists non-surgical weight loss. 7 As 490% of ghrelin is produced by the stomach and duodenumftissues directly altered by RYGBFwe hypothesized that ghrelin regulation might be disturbed following this operation. Indeed, we found that ghrelin levels in post-rygb patients were extremely low throughout the 24-h period, a paradoxical response in the face of profound prior weight loss. Since then, eight other groups have reported in prospective studies that human ghrelin levels fall after RYGB (or at least are more suppressed by food intake), and four cross-sectional studies have confirmed abnormally low ghrelin levels in post-rygb patients compared with appropriate controls (reviewed in Cummings et al. 14 ). Ghrelin levels also decrease after RYGB in rats and mice. Three other groups found no significant change in human ghrelin levels after RYGB, but these authors interpreted this as an impairment in the adaptive response of ghrelin, which would increase with weight loss caused by other means. In contrast, four groups have reported normal increases in ghrelin with RYGB-induced weight loss. These heterogeneous findings suggest that differences in surgical techniques, possibly involving treatment of the vagus nerve, 38 might lead to compromised ghrelin secretion after most, but not all, variants of this operation. Beyond possibly contributing to the marked decrease in appetite and food intake that follows RYGB, compromised ghrelin secretion might also help improve glucose tolerance. 14 Ghrelin can stimulate at least three of the four counter-regulatory hormones that oppose insulin action; it suppresses the insulin-sensitizing hormone adiponectin; it blocks intracellular insulin signaling directly in hepatocytes (at the level of PI3 kinase) and it inhibits insulin secretion. All of these actions raise blood glucose levels, as do ghrelin s effects to increase food intake, GI motility and body weight. Thus, to the extent that RYGB compromises ghrelin secretion, this should help lower glycemia. The lower intestinal hypothesis An intuitive hypothesis is that bariatric operations that create intestinal shortcuts to expedite delivery of ingested nutrients to the lower bowel accentuate the secretion of glucagon-like peptide-1 (GLP-1), thereby improving glucose homeostasis. GLP-1 is an incretin that increases glucose tolerance by enhancing insulin secretion, suppressing glucagon secretion, inhibiting gastric emptying, increasing b-cell mass (at least in animals), and possibly improving insulin sensitivity. 39 The peptide is produced primarily in the ileum and colon by nutrient-stimulated L-cells, which also secrete peptide YY and oxyntomodulin. All three of these L-cell products reduce food intake and are therefore implicated as possible contributors to the anorectic effects of some bariatric operations. Consistent with the lower intestinal hypothesis, the two bariatric operations most noted for rapid, high frequency T2DM remissionfrygb and BPDFboth create GI shortcuts for food to access the distal bowel. After BPD, which conducts food directly from the stomach to the ileum, postprandial GLP-1 excursions are unquestionably increased. It is less obvious that this would occur with RYGB because its intestinal shortcut is far less extensive. Moreover, GLP-1 secretion normally arises not only from direct nutrient contact with distal intestinal L cells but also from proximal nutrient-related signals that are transmitted from the duodenum to the distal bowel via neural pathways. As the latter mechanism is silenced after RYGB, which diverts nutrients away from the duodenum, this operation might theoretically lower postprandial GLP-1 levels. Nevertheless, recent studies clearly show that meal-stimulated secretion of GLP-1 and other L-cell peptides such as peptide YY is, indeed, substantially and durably increased after RYGB, whereas this does not occur after AGB. 27,28,40,41 Consistent with elevated postprandial GLP-1 secretion, post-rygb patients display an increased incretin effect. 25,40 As GLP-1 not only enhances insulin secretion but can also increase proliferation and decrease apoptosis of b cells, 39 it is a candidate mediator of the increase in b-cell mass that is claimed by some investigators to accompany post-rygb hyperinsulinemic hypoglycemia. 30,31 Further support for the lower intestinal hypothesis comes from experiments involving ileal interposition. In this operation, a segment of L-cell-rich ileum is transplanted into the upper intestine near the duodenum jejunum boundary, thereby increasing its exposure to ingested nutrients. As predicted, this configuration greatly enhances postprandial GLP-1 and peptide YY surges. This is associated with improved glycemic controlfwith or without concomitant weight loss depending on the rodent model or humans studiedfin the absence of any gastric restriction or malabsorption Whether improved glucose tolerance results primarily from enhanced insulin secretion, as predicted from increases in the incretin GLP-1, or from improved insulin sensitivity is unclear, and findings from different experiments support both possibilities. The upper intestinal hypothesis This hypothesis postulates that exclusion of a short segment of proximal small intestine (primarily the duodenum) from contact with ingested nutrients exerts direct anti-diabetic effects, presumably through one or more unidentified, nutrient-regulated duodenal factors or processes that influence glucose homeostasis. Dr Francesco Rubino was the first to provide strong evidence supporting this model. He developed a gastric-sparing S37

6 S38 variant of RYGB known as the duodenal jejunal bypass (DJB), in which the stomach is left intact, but an intestinal bypass is created that excludes from digestive continuity the same modest segment of proximal small intestine as is excluded in a standard RYGB. 19 In Goto Kakizaki rats, a nonobese model of polygenic T2DM, this operation improved diabetesfrapidly, durably and impressivelyfeven though it caused no reduction in food intake or body weight compared with sham-operated controls. Similar observations have subsequently been made with independent investigations of non-obese diabetic Goto Kakizaki rats 20,21 and obese diabetic Zucker rats. 18 Likewise, several small, ongoing human studies of DJB all show improvements in glycemic control, including among non-obese patients, with little or no weight loss (Cohen et al. 46 and personal communications with Arguelles-Sarmiento J et al., Lakdawala M et al., and Ramos AC et al.). Subsequent work has provided evidence that the antidiabetic impact of DJB in rats is specifically related to exclusion of the proximal small intestine from nutrient contact. 13 A variation of DJB, in which ingested food passes from the stomach into both the proximal jejunum and the proximal duodenum, had no impact on glycemia. Goto Kakizaki rats subjected to DJB with duodenal exclusion followed by DJB without duodenal exclusion, or vice versa, experienced reversible remission and reconstitution of T2DM. Diabetes was eliminated or restored on the basis of the absence or presence, respectively, of nutrient flow through the duodenum, with a fixed and unchanging degree of minor shortcutting for nutrients to reach the lower bowel. 9 These studies indicate that in rats the enhanced delivery of nutrients to the distal intestine is not the only explanation for early diabetes improvement following upper intestinal bypass. The work strongly implicates a role for the excluded proximal intestine per se, identifying a fundamentally novel physiological effect of RYGB and possibly shedding light on the pathogenesis of T2DM. In addition, support for the upper intestinal hypothesis comes from experiments in which a flexible plastic sleeve is implanted in the upper intestine, conducting food from the pylorus to the start of the jejunum without coming in contact with duodenal mucosa. Such endoluminal duodenal sleeves cause only minor or no weight loss, but they markedly improve glucose tolerance. In humans with T2DM, the placement of this device caused substantial improvement in fasting and postprandial AUC glucose levels starting as early as 1 week, long before any weight loss had occurred. 47,48 By 6 months, although minor weight loss (3.8 kg) had recently begun, glycemia was disproportionately improved, with a remarkable fall of HbA1c levels by 2.9%, a result better than would be expected with any T2DM medicine except insulin. A very intriguing and unexpected feature of both DJB and upper intestinal sleeves is that they reduce fasting and postprandial blood glucose levels to approximately the same degree, and consistent with this continuous reduction in glycemia, they have a major impact on HbA1c levels. 13,19,46 48 Thus, although these interventions seem only to re-route the flow of food through the GI tract after meals, they exert salutary effects on glycemia that persist between meals. Summary and conclusions Roux-en-Y gastric bypass causes complete remission of T2DM in a large majority of cases, and mounting evidence indicates that this remarkable phenomenon results from effects beyond those related to reduced food intake and body weight. Weight-independent anti-diabetic actions of RYGB are apparent because of the very rapid resolution of T2DM (before weight loss occurs), the greater improvement of glucose homeostasis after RYGB than after equivalent weight loss from other means, and the occasional development of very late-onset b-cell hyperfunction. Several mechanisms likely mediate the direct anti-diabetic impact of RYGB, including enhanced nutrient stimulation of L-cell peptides (for example, GLP-1) from the lower intestine, intriguing but still unidentified phenomena related to exclusion of the upper intestine from contact with ingested nutrients, compromised ghrelin secretion, and most probably other effects that have yet to be discovered. It is increasingly clear that the gut plays a major role in glucose homeostasis, regulating both insulin secretion and sensitivity, and RYGB probably influences several GI pathways in complementary ways to improve glucose control. Teasing out which of these effects are the most important, as well as identifying potential additional mechanisms, is a compelling research objective that promises not only to help optimize surgical design but also to reveal entirely novel targets for the medicinal treatment of diabetes. Conflict of interest The author has declared no financial interests. References 1 Sjostrom L, Lindroos AK, Peltonen M, Torgerson J, Bouchard C, Carlsson B et al. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004; 351: Sjostrom L, Narbro K, Sjostrom CD, Karason K, Larsson B, Wedel H et al. Effects of bariatric surgery on mortality in Swedish obese subjects. N Engl J Med 2007; 357: Buchwald H, Avidor Y, Braunwald E, Jensen MD, Pories W, Fahrbach K et al. Bariatric surgery: a systematic review and metaanalysis. JAMA 2004; 292: Pories WJ, Swanson MS, MacDonald KG, Long SB, Morris PG, Brown BM et al. Who would have thought it? An operation proves to be the most effective therapy for adult-onset diabetes mellitus. Ann Surg 1995; 222: ; discussion Schauer PR, Burguera B, Ikramuddin S, Cottam D, Gourash W, Hamad G et al. Effect of laparoscopic Roux-en Y gastric bypass on

7 type 2 diabetes mellitus. Ann Surg 2003; 238: ; discussion Cummings DE, Flum DR. Gastrointestinal surgery as a treatment for diabetes. JAMA 2008; 299: Cummings DE, Weigle DS, Frayo RS, Breen PA, Ma MK, Dellinger EP et al. Plasma ghrelin levels after diet-induced weight loss or gastric bypass surgery. N Engl J Med 2002; 346: Cummings DE, Overduin J, Foster-Schubert KE. Gastric bypass for obesity: mechanisms of weight loss and diabetes resolution. J Clin Endocrinol Metab 2004; 89: Cummings DE, Overduin J, Foster-Schubert KE, Carlson MJ. Role of the bypassed proximal intestine in the anti-diabetic effects of bariatric surgery. Surg Obes Relat Dis 2007; 3: Cummings DE, Overduin J, Shannon MH, Foster-Schubert KE. Hormonal mechanisms of weight loss and diabetes resolution after bariatric surgery. Surg Obes Relat Dis 2005; 1: Cummings DE. Gastric bypass and nesidioblastosisftoo much of a good thing for islets? N Engl J Med 2005; 353: Cummings DE, Shannon MH. Ghrelin and gastric bypass: is there a hormonal contribution to surgical weight loss? J Clin Endocrinol Metab 2003; 88: Rubino F, Forgione A, Cummings DE, Vix M, Gnuli D, Mingrone G et al. The mechanism of diabetes control after gastrointestinal bypass surgery reveals a role of the proximal small intestine in the pathophysiology of type 2 diabetes. Ann Surg 2006; 244: Cummings DE, Foster-Schubert KE, Carlson MJ, Shannon MH, Overduin J. Possible hormonal mechanisms mediating the effects of bariatric surgery. In: Pitombo C (ed). Obesity Surgery: Principle and Practice. McGraw-Hill: New York, 2007, pp Adams TD, Gress RE, Smith SC, Halverson RC, Simper SC, Rosamond WD et al. Long-term mortality after gastric bypass surgery. N Engl J Med 2007; 357: Cohen R, Pinheiro JS, Correa JL, Schiavon CA. Laparoscopic Roux-en-Y gastric bypass for BMIo35 kg/m(2): a tailored approach. Surg Obes Relat Dis 2006; 2: , discussion Lee WJ, Wang W, Lee YC, Huang MT, Ser KH, Chen JC. Effect of laparoscopic mini-gastric bypass for type 2 diabetes mellitus: comparison of BMI435 and o35 kg/m 2. J Gastrointest Surg 2008; 12: Rubino F, Zizzari P, Tomasetto C, Bluet-Pajot MT, Forgione A, Vix M et al. The role of the small bowel in the regulation of circulating ghrelin levels and food intake in the obese Zucker rat. Endocrinology 2005; 146: Rubino F, Marescaux J. Effect of duodenal jejunal exclusion in a non-obese animal model of type 2 diabetes: a new perspective for an old disease. Ann Surg 2004; 239: Pacheco D, de Luis DA, Romero A, Gonzalez Sagrado M, Conde R, Izaola O et al. The effects of duodenal-jejunal exclusion on hormonal regulation of glucose metabolism in Goto Kakizaki rats. Am J Surg 2007; 194: Wang TT, Hu SY, Gao HD, Zhang GY, Liu CZ, Feng JB et al. Ileal transposition controls diabetes as well as modified duodenal jejunal bypass with better lipid lowering in a nonobese rat model of type II diabetes by increasing GLP-1. Ann Surg 2008; 247: Wickremesekera K, Miller G, Naotunne TD, Knowles G, Stubbs RS. Loss of insulin resistance after Roux-en-Y gastric bypass surgery: a time course study. Obes Surg 2005; 15: Dixon JB, O Brien PE, Playfair J, Chapman L, Schachter LM, Skinner S et al. Adjustable gastric banding and conventional therapy for type 2 diabetes: a randomized controlled trial. JAMA 2008; 299: Muscelli E, Mingrone G, Camastra S, Manco M, Pereira JA, Pareja JC et al. Differential effect of weight loss on insulin resistance in surgically treated obese patients. Am J Med 2005; 118: Laferrere B, Teixeira J, McGinty J, Tran H, Egger JR, Colarusso A et al. Effect of weight loss by gastric bypass surgery versus hypocaloric diet on glucose and incretin levels in patients with type 2 diabetes. J Clin Endocrinol Metab 2008; 93: Pattou F, Beraud G, Arnalsteen L, Seguy D, Pigny P, Fermont C et al. Restoration of beta cell function after bariatric surgery in type 2 diabetic patients: a prospective controlled study comparing gastric banding and gastric bypass. Obes Surg 2007; 17: le Roux CW, Aylwin SJ, Batterham RL, Borg CM, Coyle F, Prasad V et al. Gut hormone profiles following bariatric surgery favor an anorectic state, facilitate weight loss, and improve metabolic parameters. Ann Surg 2006; 243: Korner J, Bessler M, Inabnet W, Taveras C, Holst JJ. Exaggerated glucagon-like peptide-1 and blunted glucose-dependent insulinotropic peptide secretion are associated with Roux-en-Y gastric bypass but not adjustable gastric banding. Surg Obes Relat Dis 2007; 3: Lee WJ, Lee YC, Chen JC, Ser KH, Chen SC, Lin CM. A randomized trial comparing laparoscopic sleeve gastrectomy versus gastric bypass for the treatment of type 2 diabetes mellitus: preliminary report. Surg Obes Relat Dis 2008; 4: Service GJ, Thompson GB, Service FJ, Andrews JC, Collazo-Clavell ML, Lloyd RV. Hyperinsulinemic hypoglycemia with nesidioblastosis after gastric-bypass surgery. N Engl J Med 2005; 353: Patti ME, McMahon G, Mun EC, Bitton A, Holst JJ, Goldsmith J et al. Severe hypoglycaemia post-gastric bypass requiring partial pancreatectomy: evidence for inappropriate insulin secretion and pancreatic islet hyperplasia. Diabetologia 2005; 48: Clancy TE, Moore Jr FD, Zinner MJ. Post-gastric bypass hyperinsulinism with nesidioblastosis: subtotal or total pancreatectomy may be needed to prevent recurrent hypoglycemia. J Gastrointest Surg 2006; 10: Bantle JP, Ikramuddin S, Kellogg TA, Buchwald H. Hyperinsulinemic hypoglycemia developing late after gastric bypass. Obes Surg 2007; 17: Goldfine AB, Mun EC, Devine E, Bernier R, Baz-Hecht M, Jones DB et al. Patients with neuroglycopenia after gastric bypass surgery have exaggerated incretin and insulin secretory responses to a mixed meal. J Clin Endocrinol Metab 2007; 92: Z Graggen K, Guweidhi A, Steffen R, Potoczna N, Biral R, Walther F et al. Severe recurrent hypoglycemia after gastric bypass surgery. Obes Surg 2008; 18: Kellogg TA, Bantle JP, Leslie DB, Redmond JB, Slusarek B, Swan T et al. Postgastric bypass hyperinsulinemic hypoglycemia syndrome: characterization and response to a modified diet. Surg Obes Relat Dis 2008; 4: Meier JJ, Butler AE, Galasso R, Butler PC. Hyperinsulinemic hypoglycemia after gastric bypass surgery is not accompanied by islet hyperplasia or increased beta-cell turnover. Diabetes Care 2006; 29: Williams DL, Grill HJ, Cummings DE, Kaplan JM. Vagotomy dissociates short- and long-term controls of circulating ghrelin. Endocrinology 2003; 144: Drucker DJ. The role of gut hormones in glucose homeostasis. J Clin Invest 2007; 117: Laferrere B, Heshka S, Wang K, Khan Y, McGinty J, Teixeira J et al. Incretin levels and effect are markedly enhanced 1 month after Roux-en-Y gastric bypass surgery in obese patients with type 2 diabetes. Diabetes Care 2007; 30: Morinigo R, Moize V, Musri M, Lacy AM, Navarro S, Marin JL et al. Glucagon-like peptide-1, peptide YY, hunger, and satiety after gastric bypass surgery in morbidly obese subjects. J Clin Endocrinol Metab 2006; 91: Strader AD, Vahl TP, Jandacek RJ, Woods SC, D Alessio DA, Seeley RJ. Weight loss through ileal transposition is accompanied by increased ileal hormone secretion and synthesis in rats. Am J Physiol Endocrinol Metab 2005; 288: E447 E Patriti A, Facchiano E, Annetti C, Aisa MC, Galli J, Fanelli C et al. Early improvement of glucose tolerance after ileal transpositition in a non-obese type 2 diabetes rat model. Obes Surg 2005; 15: S39

8 S40 44 Patriti A, Aisa MC, Annetti C, Sidoni A, Galli F, Ferri I et al. How the hindgut can cure type 2 diabetes. Ileal transposition improves glucose metabolism and beta-cell function in Goto kakizaki rats through an enhanced proglucagon gene expression and L-cell number. Surgery 2007; 142: de Paula AL, Macedo AL, Prudente AS, Queiroz L, Schraibman V, Pinus J. Laparoscopic sleeve gastrectomy with ileal interposition ( neuroendocrine brake )Fpilot study of a new operation. Surg Obes Relat Dis 2006; 2: Cohen RV, Schiavon CA, Pinheiro JS, Correa JL, Rubino F. Duodenal jejunal bypass for the treatment of type 2 diabetes in patients with body mass index of kg/m 2. Surg Obes Relat Dis 2007; 3: Rodriguez-Grunert L, Galvao Neto MP, Alamo M, Ramos AC, Baez PB et al. First human experience with endoscopically delivered and retrieved duodenal jejunal bypass sleeve. Surg Obes Relat Dis 2008; 4: Tarnoff M, Sorli C, Rodriguez L, Ramos AC, Galvao M, Reyes E et al. Interim report of a prospective, randomized shamcontrolled trial investigating a completely endoscopic duodenal jejunal bypass sleeve for the treatment of type 2 diabetes. Diabetes 2008; 57 (Suppl A): A32.