Almost since the beginning of the era of measurements of the gastrointestinal hormones, it has been observed that changes in intestinal transit might be associated with changes in gut hormone secretion. For instance, it was observed in 1979 that jejunoileal bypass is associated with markedly increased secretion of ‘enteroglucagon’ 1, a collective designation of the peptides, glicentin and oxyntomodulin, N-terminal fragments of proglucagon (Fig. 1), produced and released by the so-called L-cells, epithelial endocrine cells of the gut 2. This was interpreted to reflect unusual exposure to nutrients of the distal small intestine, where the L-cell density is the highest. Indeed, accelerated gastric emptying was also shown to be associated with an increased release of ‘enteroglucagon’ 3, and later, when the C-terminal proglucagon products, glucagon-like peptides 1 and 2 (GLP-1 and 2, Fig. 1) were discovered 2, further experiments indicated that GLP-1 secretion was considerably enhanced in humans after partial gastrectomy, resulting in accelerated gastric emptying 4. Similarly, enhanced GLP-2 secretion could be elicited in experimental animals after jejunoileal transposition, where an L-cell rich segment of the ileum was interposed into the proximal gut 5. For GLP-2, this was believed to have consequences for adaptive growth in the gut, consistent with the role of GLP-2 as a growth factor for the gut 6. GLP-1 is now known to be a powerful insulin-stimulating and glucagon-inhibiting hormone 2 and the exaggerated secretion was suspected, and later proven, to be able to cause reactive hypoglycaemia 7.
After gastric bypass operations 8, the anatomy of the upper gut is considerably rearranged (Fig. 2). Important changes include the creation of a small (30 ml) gastric pouch from the fundic part of the stomach and sectioning of the small intestine 50–100 cm from the ligament of Treitz. The distal part of the severed gut is then anastomosed to the gastric pouch, and this in effect eliminates any retention of ingested nutrients at the stomach level; instead, the nutrients rush without delay to the more distal part of the small intestine as clearly shown by studies measuring absorption of paracetamol added to a test meal 9. The proximal segment of the small intestine, with all the digestive juices secreted from the stomach, pancreas and liver, is anastomosed to the small intestine at a more distal site, creating a Y-like structure (Roux-en-Y) with three limbs: the alimentary limb (pouch→enteroentero-anastomosis), the secretory limb (stomach and duodenum→enteroentero-anastomosis) and the ‘common’ limb (from the enteroentero-anastomosis and further along the lower small intestine).
In agreement with the paracetamol absorption data already mentioned, nutrients are also extremely rapidly delivered to more distal gut segments (both alimentary and common limbs) and are absorbed at considerably accelerated rates 10,11. This applies to carbohydrates and surprisingly also to protein absorption, which shows an equally accelerated absorption rate 11, from which it can be concluded that the absorptive process of ingested proteins must occur in the common limb to allow proteolytic digestion and absorption of individual amino acids, and that this entire digestive and absorptive process is markedly accelerated.
These considerations of absorption are important because it is clear today that nutrient stimulation of gut hormone secretion, rather than being caused by exposure of ‘nutrient receptors’ on the luminal cell membrane, arises secondary to absorption either by the endocrine cells themselves, which seem to share many absorptive mechanisms with the enterocytes, or by neighbouring enterocytes 12. For glucose, the response of the L-cells [GLP-1, GLP-2, pancreatic peptide YY (PYY); see below] clearly depends on glucose and sodium entry by SGLT-1 transporters (followed by depolarization and eventually calcium entry) 13, and for amino acids, a similar sodium-coupled transport seems to be required 14, whereas lipids and bile interact with various G-protein-coupled receptors that are expressed on the basolateral membrane of the endocrine cells, accessible only after absorption 15.
From these considerations, a number of predictions can be made on secretion of the gut hormones after Roux-en-Y gastric bypass (RYGB).
- Gastric hormones: The gastric pouch is not known to produce specific gut hormones (although it contains several), but a major hormone from the stomach is, of course, gastrin. In agreement with the lack of luminal stimulation of the stomach, gastrin secretion is not elevated after the operation, but rather lowered, resulting in lower than normal postprandial peripheral levels 16. Therefore, it may also be assumed that gastric pH is low as a decreased acid secretion would be associated with increasing gastrin levels. Thus, it is difficult to imagine that gastrin plays a major role in the metabolic changes after bypass surgery.
- Duodenal hormones: In general, the duodenal hormones are not restricted to the duodenum proper, but are also found in the upper small intestine. Therefore, one would predict that the length of the secretory limb 17 would be important for postoperative duodenal hormone secretion. If this limb is long, the cells there would not be exposed to nutrients and therefore less stimulated, and conversely, if it were short, there would be endocrine cells in the alimentary limb that would be exposed to accelerated nutrient exposure, resulting in increased secretion. Most importantly, the more distally the enteroentero-anastomosis is positioned (and therefore the start of the common limb), the fewer of the duodenal hormones would be expected to be secreted. It follows that the effect of RYGB depends on the exact pattern of distribution of endocrine cells in the upper gut, but unfortunately, the mapping of the endocrine cells in humans, so far, not been detailed well 18. In addition, it is known that some adaptations may occur after the operation – for instance, the alimentary limb apparently shows significant growth postoperatively, and also expression levels of the various hormones may change, especially in the common limb 19. So what happens? A classical duodenal hormone is glucose-dependent insulinotropic polypeptide, one of the important incretin hormones. As one might expect, postoperative meal responses have been reported to be increased, unchanged or decreased 20. Most importantly, however, the changes are small compared with preoperative responses, suggesting that glucose-dependent insulinotropic polypeptide is not an important player in postoperative metabolic events. Another upper intestinal hormone is cholecystokinin (CCK), which exerts important effects on gastric motility, pancreatic enzyme secretion, bile flow to the duodenum as well as appetite regulation. In the few studies available, CCK responses are increased postoperatively 16, and this positions CCK as a potential factor in the reduction of appetite and therefore food intake. In terms of its potential effects on digestive secretions, very little is known about this after RYGB although, as mentioned, protein digestion is accelerated rather than inhibited, perhaps consistent with the elevated levels. In the single study that seems to have been carried out, the secretion of secretin, the hormone responsible for pancreatic and hepatic fluid and bicarbonate secretion, was also enhanced after surgery in agreement with the location of the responsible S-cells – similar to the CCK-releasing I-cells – to the proximal small intestine in addition to the location in the duodenum (N.W. Albrechtsen and J.J. Holst, unpublished data). Although normal secretin levels do not seem to influence postprandial insulin levels, the higher postoperative levels may actually have this function 21, but supportive evidence for this has not been provided so far. A final duodenal hormone is somatostatin, but although some of the somatostatin-producing D-cells in the gut are of the open type and not exclusively limited to the most proximal small intestine, and therefore would be expected to respond to luminal nutrient stimulation, postoperative changes do not seem to occur 16. This applies to plasma levels of the 14 amino acid form of somatostatin as well as the 28-amino acid form, considered to be the main form released from the intestinal D-cells.
- The distal small intestine: The main hormones secreted from the distal small intestine are GLP-1, GLP-2, PYY and neurotensin, and all of these show markedly increased responses after the operations 16. Up to 30-fold increases have been observed 22. GLP-1 and 2, derived from the same prohormone precursor (Fig. 1), are products of the L-cells that show an increased density in the mucosa of the small intestine towards the ileum, but are, nevertheless, because of the larger mucosal surface area there, also found in rather large numbers in the more proximal parts of the small intestine 23. PYY is a coproduct of the L-cells, but only in the more distal L-cells 14. L-cells are also found in large numbers in the colon, but nothing is known about the contribution of the colon towards the normal secretory responses of these hormones – but it is known that completely normal meal responses are observed in patients after total colectomy 24, suggesting that the normal contribution is small. It has been observed that the rate of entry of chyme into the colon is normal after RYGB, suggesting that the colon is not a major contributor towards the postoperative plasma response 25.
The metabolic outcomes of the large neurotensin response are not yet clear, but experimental data suggest that neurotensin may inhibit food intake 26 and may contribute towards the weight loss after gastric bypass. PYY exerts pronounced effects on food intake both in experimental animals and in humans, indicating that PYY might play a role in the weight loss 27. As indicated in Fig. 1, both GLP-1 and GLP-2 are products of intestinal expression of proglucagon. Additional proglucagon products from the gut include glicentin and oxyntomodulin (Fig. 1). Little is known about the metabolic actions of glicentin, if any (it exerts no effects on insulin secretion in humans), but each of the other products has distinct metabolic actions. Oxyntomodulin is a dual agonist for the GLP-1 and the glucagon receptor 28. It is weaker than each of those peptides, and therefore, hardly has any important metabolic actions under normal circumstances, but after the operation, its concentrations increase so markedly that it may contribute 29. The glucagon action is of interest because this hormone may inhibit food intake and stimulate insulin secretion 30; the GLP-1 actions are discussed below. Currently, the pharmaceutical industry is attempting to develop GLP-1/glucagon coagonists for obesity because of these actions 31. The glucagon-like activity would also be expected to increase hepatic glucose production, but direct studies have shown that postprandial hepatic glucose production is inhibited normally after RYGB 10.
The main action of GLP-2 apparently is to maintain the integrity and promote growth of the small intestinal mucosa 32. The increases in GLP-2 levels are very dramatic 16 and it is conceivable, although not proven, that the postoperative trophic actions on the small intestine, mentioned above, are at least partly because of the actions of GLP-2.
The main actions of GLP-1 are to stimulate insulin secretion (it is one of the incretin hormones) and to regulate appetite and thereby inhibit food intake 2. These effects are so robust that long-acting analogues of this hormone are now being used for therapy of diabetes 33 as well as obesity 34. Therefore, it is natural to assume that the extremely exaggerated postoperative meal responses of this hormone play an important role in the metabolic effects of the operation that include diabetes resolution and weight loss. Indeed, this seems to be the case. For GLP-1, a potent receptor antagonist, exendin 9–39 (Ex-9) has been available for some time 35 and has considerably helped to elucidate the effects of endogenous GLP-1. With the antagonist, it is possible to completely block the actions of exogenous GLP-1 36. In experiments with Ex-9, it has been shown that the improved insulin response to a test meal observed after RYGB, and that is considered essential for the improved glucose tolerance and thereby diabetes resolution, can be completely prevented with this antagonist 37. Consistent with this, all of the improvement in glucose tolerance induced by the operation in obese individuals with type 2 diabetes was lost after antagonist administration. It was also observed that the marked improvement in β-cell glucose sensitivity, representing the primary effect of GLP-1 on the β-cell, was lost after antagonist administration. Indeed, the postoperative enhancement of insulin secretion is so powerful that in some individuals, it results in severe reactive hypoglycaemia, and also in these individuals, the antagonist prevents both hyperinsulinaemia and hypoglycaemia 38. Is enhanced GLP-1 secretion the only antidiabetic mechanism engaged by RYGB? Clearly not. The first important event is the perioperative dietary restriction, which is associated with loss of liver fat and therefore considerably improved hepatic insulin resistance 20. The improvement in insulin sensitivity progresses further as weight is lost, and eventually also comprises peripheral insulin resistance. Undoubtedly, this is a major factor. However, the evidence strongly supports that the exaggerated GLP-1 secretion is responsible for the improved insulin secretion (see Fig. 2).
But what about the effect on appetite and food intake? Early studies showed that good responders (i.e. those with a large postoperative weight loss) also had large GLP-1 (and PYY) responses whereas poor responders had more shallow responses 39. It was also observed that although intravenous GLP-1 and PYY were administered in doses that would somewhat resemble the operated patients’ meal responses, both exerted rather weak effects on food intake; their combination elicited a synergistic and pronounced anorectic response 27. Experiments with the GLP-1 antagonist, Ex-9, were initially confusing: although administration of the antagonist clearly increased food intake before the operation (indicating elimination of an inhibitory effect of GLP-1), there was no effect after the operation 40. Analysis of the concomitant hormone responses showed that the antagonist markedly augmented the already exaggerated GLP-1 responses further (challenging the efficacy of the antagonist). The increase is probably because of a well-known negative feedback effect of GLP-1 on L-cell secretion 41 and, when this is blocked by the antagonist, secretion increases. However, also PYY responses, already considerably increased by the operation, were increased markedly by the antagonist, consistent with secretion of this hormone from the same cells as GLP-1 42. However, the antagonist cannot block the appetite-inhibitory effects of PYY; thus, although the effects of GLP-1 were probably blocked, the inhibitory effects of PYY were correspondingly increased, explaining the resulting lack of effect. To investigate the effects of the associated increase in PYY, a PYY antagonist would have been useful, but PYY antagonists do not seem to exist. However, the anorectic effect of PYY requires conversion of the initially secreted peptide, PYY 1–36, into a truncated form, PYY 3–36, mediated by the ubiquitous enzyme, dipeptidyl peptidase-4, which also truncates (and inactivates) GLP-1. Thus, it is possible to prevent the formation of the active metabolite PYY 3–36 with the aid of dipeptidyl peptidase 4 inhibitors 43, which are used widely for the therapy of diabetes (because of their protective effects on GLP-1). In final studies, it was shown that although neither administration of Ex-9 nor a dipeptidyl peptidase 4 inhibitor alone had any effect on food intake compared with control experiments, the combination of the two, which did, indeed, result in considerably decreased levels of PYY 3–36, increased food intake significantly, by about 20%, sufficient to explain a major part of the decreased food intake elicited by the operation 40.
Thus, it seems clear that the gut hormones play a major role in the results of bypass surgery, with food intake evidently being inhibited by the two hormones, GLP-1 and PYY, in combination (and perhaps as a result of a potentiating interaction), whereas the antidiabetic effect is probably mainly because of the effects of GLP-1 on insulin secretion.
Conflicts of interest
There are no conflicts of interest.
1. Holst JJ, Sorensen TI, Andersen AN, Stadil F, Andersen B, Lauritsen KB, Klein HC. Plasma enteroglucagon after jejunoileal bypass with 3 : 1 or 1 : 3 jejunoileal ratio. Scand J Gastroenterol 1979; 14:205–207.
2. Holst JJ. The physiology of glucagon-like peptide 1. Physiol Rev 2007; 87:1409–1439.
3. Lauritsen KB, Frederiksen HJ, Uhrenholdt A, Holst JJ. The correlation between gastric emptying time and the response of GIP and enteroglucagon to oral glucose in duodenal ulcer patients. Scand J Gastroenterol 1982; 17:513–516.
4. Miholic J, Orskov C, Holst JJ, Kotzerke J, Meyer HJ. Emptying of the gastric substitute, glucagon-like peptide-1 (GLP-1
), and reactive hypoglycemia after total gastrectomy. Dig Dis Sci 1991; 36:1361–1370.
5. Thulesen J, Hartmann B, Kissow H, Jeppesen PB, Orskov C, Holst JJ, Poulsen SS. Intestinal growth adaptation and glucagon-like peptide 2 in rats with ileal–jejunal transposition or small bowel resection. Dig Dis Sci 2001; 46:379–388.
6. Hartmann B, Thulesen J, Hare KJ, Kissow H, Orskov C, Poulsen SS, Holst JJ. Immunoneutralization of endogenous glucagon-like peptide-2 reduces adaptive intestinal growth in diabetic rats. Regul Pept 2002; 105:173–179.
7. Toft-Nielsen M, Madsbad S, Holst JJ. Exaggerated secretion of glucagon-like peptide-1 (GLP-1
) could cause reactive hypoglycaemia. Diabetologia 1998; 41:1180–1186.
8. Ashrafian H, Bueter M, Ahmed K, Suliman A, Bloom SR, Darzi A, Athanasiou T. Metabolic surgery: an evolution through bariatric animal models. Obes Rev 2010; 11:907–920.
9. Falkén Y, Hellström PM, Holst JJ, Näslund E. Changes in glucose homeostasis after Roux-en-Y gastric bypass
surgery for obesity at day three, two months, and one year after surgery: role of gut peptides. J Clin Endocrinol Metab 2011; 96:2227–2235.
10. Jacobsen SH, Bojsen-Møller KN, Dirksen C, Jørgensen NB, Clausen TR, Wulff BS, et al.. Effects of gastric bypass
surgery on glucose absorption and metabolism during a mixed meal in glucose-tolerant individuals. Diabetologia 2013; 56:2250–2254.
11. Bojsen-Møller KN, Jacobsen SH, Dirksen C, Jørgensen NB, Reitelseder S, Jensen JE, et al.. Accelerated protein digestion and amino acid absorption after Roux-en-Y gastric bypass
. Am J Clin Nutr 2015; 102:600–607.
12. Gribble FM, Reimann F. Enteroendocrine cells: chemosensors in the intestinal epithelium. Annu Rev Physiol 2016; 78:277–299.
13. Kuhre RE, Frost CR, Svendsen B, Holst JJ. Molecular mechanisms of glucose-stimulated GLP-1
secretion from perfused rat small intestine. Diabetes 2015; 64:370–382.
14. Svendsen B, Pedersen J, Albrechtsen NJ, Hartmann B, Toräng S, Rehfeld JF, et al.. An analysis of cosecretion and coexpression of gut hormones
from male rat proximal and distal small intestine. Endocrinology 2015; 156:847–857.
15. Brighton CA, Rievaj J, Kuhre RE, Glass LL, Schoonjans K, Holst JJ, et al.. Bile acids trigger GLP-1
release predominantly by accessing basolaterally located g protein-coupled bile acid receptors. Endocrinology 2015; 156:3961–3970.
16. Jacobsen SH, Olesen SC, Dirksen C, Jørgensen NB, Bojsen-Møller KN, Kielgast U, et al.. Changes in gastrointestinal hormone responses, insulin sensitivity, and beta-cell function within 2 weeks after gastric bypass
in non-diabetic subjects. Obes Surg 2012; 22:1084–1096.
17. Mahawar KK, Kumar P, Parmar C, Graham Y, Carr WR, Jennings N, et al.. Small bowel limb lengths and Roux-en-Y gastric bypass
: a systematic review. Obes Surg 2016; 26:660–671.
18. Rhee NA, Wahlgren CD, Pedersen J, Mortensen B, Langholz E, Wandall EP, et al.. Effect of Roux-en-Y gastric bypass
on the distribution and hormone expression of small-intestinal enteroendocrine cells in obese patients with type 2 diabetes. Diabetologia 2015; 58:2254–2258.
19. Hansen CF, Bueter M, Theis N, Lutz T, Paulsen S, Dalbøge LS, et al.. Hypertrophy dependent doubling of L-cells in Roux-en-Y gastric bypass
operated rats. PLoS One 2013; 8:e65696.
20. Madsbad S, Dirksen C, Holst JJ. Mechanisms of changes in glucose metabolism and bodyweight after bariatric surgery. Lancet Diabetes Endocrinol 2014; 2:152–164.
21. Fahrenkrug J, Schaffalitzky deMuckadell OB, Kühl C. Effect of secretin on basal- and glucose-stimulated insulin secretion in man. Diabetologia 1978; 14:229–234.
22. Jørgensen NB, Jacobsen SH, Dirksen C, Bojsen-Møller KN, Naver L, Hvolris L, et al.. Acute and long-term effects of Roux-en-Y gastric bypass
on glucose metabolism in subjects with Type 2 diabetes and normal glucose tolerance. Am J Physiol Endocrinol Metab 2012; 303:E122–E131.
23. Hansen CF, Vrang N, Sangild PT, Jelsing J. Novel insight into the distribution of L-cells in the rat intestinal tract. Am J Transl Res 2013; 5:347–358.
24. Nauck MA, Siemsglüss J, Orskov C, Holst JJ. Release of glucagon-like peptide 1 (GLP-1
[7-36 amide]), gastric inhibitory polypeptide (GIP) and insulin in response to oral glucose after upper and lower intestinal resections. Z Gastroenterol 1996; 34:159–166.
25. Dirksen C, Damgaard M, Bojsen-Møller KN, Jørgensen NB, Kielgast U, Jacobsen SH, et al.. Fast pouch emptying, delayed small intestinal transit, and exaggerated gut hormone responses after Roux-en-Y gastric bypass
. Neurogastroenterol Motil 2013; 25:346–e255.
26. Grunddal KV, Ratner CF, Svendsen B, Sommer F, Engelstoft MS, Madsen AN, et al.. Neurotensin is coexpressed, coreleased, and acts together with GLP-1
and PYY in enteroendocrine control of metabolism. Endocrinology 2016; 157:176–194.
27. Schmidt JB, Gregersen NT, Pedersen SD, Arentoft JL, Ritz C, Schwartz TW, et al.. Effects of PYY3-36 and GLP-1
on energy intake, energy expenditure, and appetite in overweight men. Am J Physiol Endocrinol Metab 2014; 306:E1248–E1256.
28. Baldissera FG, Holst JJ, Knuhtsen S, Hilsted L, Nielsen OV. Oxyntomodulin (glicentin-(33–69): pharmacokinetics, binding to liver cell membranes, effects on isolated perfused pig pancreas, and secretion from isolated perfused lower small intestine of pigs. Regul Pept 1988; 21 (1–2):151–166.
29. Albrechtsen NW, Albrechtsen R, Hornburg D, Toräng S, Svendsen L, Jepsen SL, et al.. A novel immune-based approach for measurement of the anorectic gut hormone oxyntomodulin: changes after gastric bypass
surgery. Diabetes 2015; 64 (Suppl 1):2076P.
30. Bagger JI, Holst JJ, Hartmann B, Andersen B, Knop FK, Vilsbøll T. Effect of oxyntomodulin, glucagon, GLP-1
, and combined glucagon +GLP-1
infusion on food intake, appetite, and resting energy expenditure. J Clin Endocrinol Metab 2015; 100:4541–4552.
31. Day JW, Ottaway N, Patterson JT, Gelfanov V, Smiley D, Gidda J, et al.. A new glucagon and GLP-1
co-agonist eliminates obesity in rodents. Nat Chem Biol 2009; 5:749–757.
32. Estall JL, Drucker DJ. Glucagon-like peptide-2. Annu Rev Nutr 2006; 26:391–411.
33. Madsbad S, Krarup T, Deacon CF, Holst JJ. Glucagon-like peptide receptor agonists and dipeptidyl peptidase-4 inhibitors in the treatment of diabetes: a review of clinical trials. Curr Opin Clin Nutr Metab Care 2008; 11:491–499.
34. Pi-Sunyer X, Astrup A, Fujioka K, Greenway F, Halpern A, Krempf M, et al.. A randomized, controlled trial of 3.0 mg of liraglutide in weight management. N Engl J Med 2015; 373:11–22.
35. Edwards CM, Todd JF, Mahmoudi M, Wang Z, Wang RM, Ghatei MA, Bloom SR. Glucagon-like peptide 1 has a physiological role in the control of postprandial glucose in humans: studies with the antagonist exendin 9–39
. Diabetes 1999; 48:86–93.
36. Schirra J, Sturm K, Leicht P, Arnold R, Göke B, Katschinski M. Exendin(9–39)amide is an antagonist of glucagon-like peptide-1(7–36)amide in humans. J Clin Invest 1998; 101:1421–1430.
37. Jørgensen NB, Dirksen C, Bojsen-Møller KN, Jacobsen SH, Worm D, Hansen DL, et al.. Exaggerated glucagon-like peptide 1 response is important for improved β-cell function and glucose tolerance after Roux-en-Y gastric bypass
in patients with type 2 diabetes. Diabetes 2013; 62:3044–3052.
38. Salehi M, Gastaldelli A, D’Alessio DA. Blockade of glucagon-like peptide 1 receptor corrects postprandial hypoglycemia after gastric bypass
. Gastroenterology 2014; 146:669–680.
39. le Roux CW, Welbourn R, Werling M, Osborne A, Kokkinos A, Laurenius A, et al.. Gut hormones
as mediators of appetite and weight loss after Roux-en-Y gastric bypass
. Ann Surg 2007; 246:780–785.
40. Svane MS, Jørgensen NB, Bojsen-Moller KN, Dirksen C, Nielsen S, Kristiansen VB, et al.. Peptide YY and glucagon-like peptide-1 contribute to decreased food intake after Roux-en-Y gastric bypass
surgery. Int J Obes (Lond) 2016. [Epub ahead of print].
41. Deacon CF, Wamberg S, Bie P, Hughes TE, Holst JJ. Preservation of active incretin hormones by inhibition of dipeptidyl peptidase IV suppresses meal-induced incretin secretion in dogs. J Endocrinol 2002; 172:355–362.
42. Böttcher G, Alumets J, Håkanson R, Sundler F. Co-existence of glicentin and peptide YY in colorectal L-cells in cat and man. An electron microscopic study. Regul Pept 1986; 13 (3–4):283–291.
43. Aaboe K, Knop FK, Vilsbøll T, Deacon CF, Holst JJ, Madsbad S, Krarup T. Twelve weeks treatment with the DPP-4 inhibitor, sitagliptin, prevents degradation of peptide YY and improves glucose and non-glucose induced insulin secretion in patients with type 2 diabetes mellitus. Diabetes Obes Metab 2010; 12:323–333.