Department of Medicine, Division of Metabolism, Endocrinology and Nutrition University of Washington Veterans Affairs Puget Sound Health Care System Seattle, Washington 98108
Address all correspondence and requests for reprints to: David E. Cummings, M.D., University of Washington, VA Puget Sound Health Care System, 1660 South Columbian Way, S-111-Endo, Seattle, Washington 98108. E-mail: davidec{at}u.washington.edu.
Ghrelin, a recently discovered peptide secreted primarily by the stomach and proximal small intestine, is the only known appetite-stimulating hormone (1, 2, 3, 4). Circulating ghrelin levels increase markedly before, and decrease after, every meal in humans and other animals (5, 6, 7, 8). These and other observations (4) have led to the hypothesis that ghrelin participates in the short-term control of pre-meal hunger and meal initiation (5). Moreover, ghrelin exhibits many properties of an "adiposity signal" that communicates the status of body energy stores to the brain and contributes to long-term body weight regulation (4). Circulating ghrelin levels correlate inversely with body mass index and are modulated by alterations in body weight (9, 10, 11, 12). Ghrelin activates neurons in the hypothalamic arcuate nucleus that cosecrete neuropeptide Y and agouti-related protein, both of which are prototypical anabolic neuropeptides that stimulate food intake and promote weight gain (13, 14, 15). Peripheral or central ghrelin infusions increase short-term food intake in humans and other species as potently as any known agent does (3, 14, 16). Chronic ghrelin administration increases body weight not only by stimulating food consumption but also by suppressing energy expenditure and fat catabolism (3, 13, 14, 17). Conversely, the blockade of ghrelin signaling in the brain has been reported to decrease food intake, fat mass, and body weight (13, 18, 19, 20). These findings suggest that ghrelin may be an orexigenic counterpart to leptin and insulin in the long-term control of body weight (4).
Consistent with the hypothesis that ghrelin participates in energy homeostasis, circulating levels increase in response to weight loss resulting from numerous causes. Elevated ghrelin levels have been demonstrated in the context of weight loss arising from low-calorie diets (12, 21), mixed lifestyle modifications (10), chronic exercise (22), cancer anorexia (21, 23), cardiac cachexia (24), hepatic cachexia (25), and anorexia nervosa (26, 27, 28). The implication of these findings is that an increase in ghrelin levels caused by weight loss may help to promote regaining weight. If this is true, methods of weight loss that fail to trigger a compensatory rise in ghrelin levels might help sustain weight loss on a long-term basis. In this issue of JCEM, Holdstock et al. (29) contribute intriguing new findings to the growing debate over whether gastric bypass surgery causes durable weight loss in part by suppressing ghrelin levels.
Roux-en-Y gastric bypass (RYGB) surgery is the most effective approved treatment for morbid obesity (30). The procedure restricts the gastric volume that is capable of storing food, bypassing most of the stomach and all of the duodenum (i.e. the majority of ghrelin-producing tissue) with a gastrojejunal anastomosis (Fig. 1). RYGB typically causes a 3540% loss of body weight, and most of this effect is sustained, in follow-up studies, for as long as 14 yr in the literature (31, 32, 33) and 20 yr according to unpublished observations (Pories, W. J., personal communication). The mechanisms by which RYGB overcomes the adaptive responses that typically constrain weight loss remain enigmatic (4). The most intuitive mechanism is restriction of the gastric volume that can accept food, and there is no question that early satiety and reduced meal sizes do occur after RYGB. Similar changes follow the equally restrictive vertical-banded gastroplasty, yet this procedure is less effective at maintaining long-term weight loss (4, 31, 32). It is unlikely that the difference in efficacy between RYGB and vertical-banded gastroplasty results from either RYGB-induced malabsorption (which is not clinically significant after the proximal procedure) or dumping syndrome (which develops inconsistently). The discrepancy is more likely to arise because patients who undergo RYGB typically experience a generalized loss of appetite that extends beyond the immediate postprandial period. This is associated with a seemingly paradoxical decrease in meal frequency (as well as size) and reduced consumption of calorie-dense foods (4). These unexpected changes in eating behavior suggest that additional anorexigenic alterations beyond simple gastric restriction occur after RYGB.
|
Several other groups have subsequently observed low and/or suppressed ghrelin levels following RYGB. In a prospective study, Geloneze et al. (34) reported that ghrelin concentrations fell significantly in both diabetic and nondiabetic patients at 1 yr after RYGB, when compared with preoperative values. Ghrelin levels declined even in the face of a 38% loss of body weight, a change that would stimulate ghrelin if it had been achieved by most other methods. Similarly, Tritos et al. (35) found that ghrelin concentrations at baseline and in response to an oral glucose load were significantly lower in post-RYGB subjects than in equally obese, nonsurgical controls. Ghrelin values were again lower in the RYGB group despite massive weight loss; and paradoxically, values were 47% lower than those in morbidly obese subjects about to undergo RYGB and 57% lower than those in lean controls. Frühbeck et al. (36) found that ghrelin levels in patients who had undergone RYGB were only approximately one fourth as high as those in body mass index-matched patients who had lost comparable weight following either adjustable gastric banding or biliopancreatic diversion.
In their prospective study, Faraj et al. (37) reported that the effect of RYGB on circulating ghrelin levels depended on the dynamic status of weight loss. Subjects who had achieved a stable postoperative weight had ghrelin levels that were unchanged relative to preoperative values. Given that these people had lost 36% of their body weight, which would normally trigger an increase in ghrelin, the authors interpreted their stable ghrelin values to be inappropriately normal in the context of weight loss. Ghrelin levels in subjects who were still actively losing weight were higher than preoperative values, as would be expected from other modes of weight loss. The authors conclusion was that RYGB in these subjects had disrupted the normal adaptive response of ghrelin to the underweight condition but not to nutritional deficit.
In this issue of JCEM, Holdstock et al. (29) report data that contrast with those outlined above, showing instead that subjects experiencing weight loss after RYGB had higher ghrelin levels than they had before surgery. This excellent prospective study is the largest yet published on this topic. The authors examined 66 morbidly obese individuals before RYGB and at 6 and 12 months after surgery. The average body mass index decreased by 22.3% and 29.7%, respectively, at these two postoperative time points, and average ghrelin levels increased by 43.8% and 61.8%, respectively. Final ghrelin concentrations in a subset of postoperative patients were equivalent to those in equally obese, nonsurgical controls. The authors interpreted this finding to indicate that RYGB did not directly affect ghrelin levels. It could be argued, however, that the postsurgical patients should have had higher ghrelin levels than did the matched obese controls because the former group had experienced substantial weight loss, whereas the latter had not.
In summary, the effect of RYGB on circulating ghrelin appears to differ from one surgical center to another. Results range from postoperative ghrelin suppression, to inappropriately stable values despite massive weight loss, to increased ghrelin as seen in other forms of weight loss. These discrepant results raise important questions with potential clinical implications for surgical design. For example, does the presence or absence of ghrelin suppression after RYGB influence the weight-reducing efficacy of the procedure? This possibility is hinted at by the observation that in the study by Holdstock et al. (29), in which ghrelin levels increased postoperatively, surgical weight loss was less than that achieved in the studies showing low postoperative ghrelin values (12, 34, 35, 36) (29.7% vs. 3638% weight loss, respectively). The magnitude of weight loss reported by Holdstock et al. (29) resembles that typically achieved with vertical-banded gastroplasty, whereas the other studies report greater degrees of weight loss that are typical of RYGB (4, 30, 31, 32). Although surgical differences, such as the size of the gastric pouch, gastrojejunal stoma, Roux limb, and biliopancreatic limb, clearly contribute to heterogeneous results among centers, a critical, unanswered question is whether ghrelin suppression achieved by some types of RYGB contributes to weight loss. If so, determining the mechanism by which this effect on ghrelin occurs will be important, so that it can expressly be sought, either surgically or medically.
We have speculated that ghrelin levels sometimes decrease after RYGB because the majority of ghrelin-producing cells, chronically isolated from contact with enteral nutrients, undergo override inhibition (12). According to this model, the condition of an empty stomach and duodenum, which acutely stimulates ghrelin production, paradoxically inhibits it when present continuously after RYGB. This phenomenon would resemble the paradoxical suppression of gonadotropins or GH by continuous signaling from GnRH or GHRH, respectively. The possibility that ghrelin-producing cells in the gut are subject to override inhibition is suggested by several lines of evidence that we have summarized elsewhere (4). If this model is valid, the location of the staple line that partitions the stomach into upper and lower compartments in RYGB may be a critical determinant of ghrelin suppression. The gastric fundus, which is the richest source of ghrelin in the body, lies immediately adjacent to this dividing line in a vertical-banded RYGB. Positioning the staple line even slightly too far to the left might include some of the ghrelin-dense fundus in the upper pouch. Ghrelin-producing cells in this tissue would remain in intermittent contact with food and, thus, fail to be silenced through override inhibition. Most endocrine tissue has vast excess reserve capacity; hence, even a small residual segment of ghrelin-producing terrain in the upper pouch might sustain relatively normal ghrelin levels. Normal concentrations might also be sustained in RYGB variants with a short biliopancreatic intestinal limb because intermittent retrograde flow of ingested nutrients from the jejunojejunostomy anastomosis could reach and enliven the ghrelin-rich duodenum and stomach. Interestingly, in the study by Holdstock et al. (29), subjects who had high ghrelin levels after RYGB had among the shortest biliopancreatic limbs and widest upper gastric pouches among RYGB patients examined in ghrelin studies, to date.
The override inhibition model predicts that bariatric operations that do not exclude major ghrelin-producing tissues, such as the fundus, from contact with food would be ineffective at suppressing ghrelin. In these settings, levels should increase in response to weight loss resulting from gastric restriction and/or malabsorption. Consistent with this prediction, Frühbeck et al. (36) found that ghrelin concentrations were approximately four times higher in patients who had undergone adjustable gastric banding or biliopancreatic diversion, both of which leave the fundus in contact with nutrients, compared with subjects who had RYGB. Similarly, in a prospective study, we found that ghrelin levels were 53% higher 8 months after adjustable gastric banding than they were before surgery (38). These values remained elevated in a subset of patients followed for up to 3 yr. Persistently increased ghrelin levels after purely restrictive procedures, such as gastroplasty, may help explain why these methods achieve less effective long-term weight loss than RYGB (4, 30, 31, 32). The override inhibition hypothesis also predicts that horizontal-banded bariatric procedures, which do not exclude the fundus from contact with food, would be ineffective at suppressing ghrelin production. It is conceivable that this property contributes to the disappointing results that have led to the widespread abandonment of horizontal-banded approaches.
An alternate hypothesis to reconcile disparate reports regarding the effect of RYGB on ghrelin levels pertains to variable surgical treatment of the autonomic nervous input to ghrelin-producing tissue in the foregut. Rodent data indicate that vagal (parasympathetic) innervation influences ghrelin levels (39), and we have found that sympathetic stimulation can also affect ghrelin levels (Cummings, D. E., T. O. Mundinger, and G. J. Taborsky, unpublished data). Treatment of the vagus nerve in RYGB is not consistent among bariatric surgeons. Although many practitioners strive to preserve vagal fibers, some projections innervating the fundus, especially from the left vagal branch, are almost certainly severed in modern procedures that completely transect the stomach. Increasingly popular laparoscopic approaches often sacrifice the entire vagal input to the majority of ghrelin-producing tissue. Differing degrees of parasympathetic and/or sympathetic denervation of the stomach and proximal small intestine could thus contribute to variable effects of RYGB on circulating ghrelin.
We have found that vagotomy in rats disrupts the long-term adaptive response of circulating ghrelin to reduced body weight, but does not affect the short-term response to ingested nutrients (40). These observations predict that, in cases of RYGB in which substantial ghrelin-producing tissue has lost vagal innervation, ghrelin levels would fail to increase in response to decreased body weight but would increase in states of especially low food intake, i.e. in active negative energy balance. Consistent with this prediction, Faraj et al. (37) reported that post-RYGB subjects who were still actively losing weight had elevated ghrelin levels, whereas subjects who had achieved a new steady state weight had unchanged ghrelin levels, despite their massive weight loss. Similarly, patients in the study by Holdstock et al. (29) appeared to be actively losing weight at the end of the 12-month study period, and these patients also showed increased ghrelin levels. In contrast, most of the subjects in the four studies reporting low and/or decreased ghrelin levels had achieved stable, reduced body weights and had, thus, returned to neutral energy balance (12, 34, 35, 36).
The hypotheses we offer here to help reconcile disparate findings regarding the possible ability of RYGB to reduce ghrelin levels do not account for all of the differences observed. Moreover, it has not been proven that inhibition of ghrelin by any meanssurgical or medicalcauses weight loss in humans. Additional studies are required to clarify these issues and to answer the overarching question of whether ghrelin blockade will prove to be an effective antiobesity approach.
![]() |
Footnotes |
---|
Received April 18, 2003.
Accepted April 18, 2003.
![]() |
References |
---|
![]() ![]() |
---|