Department of Medicine, Columbia University College of Physicians and Surgeons, 650 West 168th Street, Black Building, Room 905, New York, NY 10032. E-mail: firstname.lastname@example.org
Objective: To quantify plasma concentrations of hormones that regulate energy homeostasis in order to establish possible mechanisms for greater weight loss after Roux-en-Y gastric bypass (RYGBP) compared with gastric banding (BND).
Research Methods and Procedures: Four groups of women were studied: lean (n = 8; mean BMI, 21.6 kg/m2); BND (n = 9; BMI, 35.8; 25% weight loss), RYGBP (n = 9; BMI, 34.2; 36% weight loss), and controls matched for BMI to the surgical groups (n = 11; BMI, 34.4).
Results: Fasting total peptide YY (PYY) and PYY(3–36) immunoreactivity were similar among all groups, but the postprandial response in the RYGBP group was exaggerated, such that 30 minutes after the meal, total and PYY(3–36) levels were 2- to 4-fold greater compared with all other groups. Maximal postprandial suppression of total ghrelin was blunted in the BND group (13%) compared with RYGBP (27%). Postprandial suppression of octanoylated ghrelin was also less in BND (29%) compared with RYGBP (56%). Fasting insulin was lower in RYGBP (6.6 μU/mL) compared with BND (10.0 μU/mL). Compared with lean controls, leptin concentrations were significantly higher in BND but not in RYGBP. There was a greater increase in post-meal satiety in the RYGBP group compared with BND and overweight controls.
Discussion: The differences between RYGBP and BND subjects in postprandial concentrations of PYY and ghrelin would be expected to promote increased satiety and earlier meal termination in RYGBP and may aid in greater weight loss. The differences in insulin and leptin concentrations associated with these procedures may also reflect differences in insulin sensitivity and energy partitioning.
The treatment of obesity with conservative measures or pharmacotherapy often fails to produce a permanent reduction in body weight. As a consequence, surgical methods are increasingly employed that produce a greater degree of long-term weight loss (1, 2, 3, 4). Roux-en-Y gastric bypass (RYGBP)1 is one of the most commonly performed bariatric surgery operations. This procedure restricts gastric volume through the creation of a small pouch along the lesser curvature and reroutes nutrient flow from the upper portion of the stomach directly into the mid- to distal jejunum. Adjustable gastric banding (BND) is another more recently designed intervention that consists of a band placed laparoscopically around the upper stomach and connected by tubing to an access port placed outside the abdominal cavity. The inner diameter of the band can be readily adjusted by the addition or removal of saline through the access port. The average reduction in body weight is greater after RYGBP compared with BND, 40% vs. 28%, respectively, with less weight regain over time (1, 2). It is unclear what mechanisms contribute to the greater efficacy, particularly over the short term, of RYGBP. Although absorption of iron and vitamin B12 is decreased after RYGBP, malabsorption of protein, carbohydrate, and fat does not typically occur (5, 6, 7). Recent data suggest that neural and hormonal mechanisms may contribute to the greater efficacy of RYGBP as compared with BND and diet-induced weight loss (8, 9, 10).
Peptide YY (PYY) and ghrelin are gastrointestinal peptide hormones that modulate metabolism and appetite. Secretion of PYY occurs from L cells lining the distal small bowel and colon shortly after food intake, before ingested nutrients arrive in the distal intestine, and subsequently through direct stimulation by nutrients, particularly lipid and carbohydrate. PYY is present in human intestinal extracts and plasma in at least two known molecular forms, PYY(1–36) and a cleavage product, PYY(3–36) (11, 12). The regulation of this processing is not known, and the physiological effects of either form of PYY are complex due to different binding affinities for several Neuropeptide Y receptor subtypes located in both the periphery and central nervous system (13, 14). For example, PYY(1–36), administered centrally is an orexigen (15), whereas systemic injection inhibits gastric emptying and induces emesis (16). Peripheral infusion experiments of PYY(3–36) in rodents have produced inconsistent results on food intake and body weight (17), whereas infusion of PYY(3–36) decreases 24-hour food intake in non-human primates (18, 19) and both lean and obese humans (20, 21, 22).
Ghrelin is produced primarily by A cells in the oxyntic glands of the stomach fundus. A unique post-translational acylation with octanoic acid modulates the bioactivity of ghrelin (23) and may also be necessary for efficient passage of ghrelin across the blood-brain barrier (24). Although originally described as an inducer of growth hormone release, ghrelin has more recently been shown to affect appetite, food intake, energy expenditure, and gut motility (25, 26, 27, 28, 29). In the rat, administration of ghrelin stimulates feeding, increases body weight, and decreases fat use. Intravenous injection of ghrelin in humans induces hunger and food intake (30), yet plasma levels of ghrelin are down-regulated in obese individuals (31, 32). Thus, dysregulation of fasting ghrelin levels appears unlikely to be an etiologic factor for common obesity. Ghrelin concentrations are also regulated acutely in relation to food intake (33, 34). The high level of ghrelin before a meal and subsequent postprandial fall were originally interpreted as an indication that ghrelin may play a role in meal initiation. However, because the spontaneous request for a meal is not predicted by recovery of plasma ghrelin level, it has been proposed that the rapid fall in ghrelin after a meal may act more as a signal not to eat than the recovery of ghrelin acts as a signal to eat (35). Although weight loss by caloric restriction is associated with an increase in circulating concentrations of ghrelin (28, 36), studies that examine ghrelin concentrations before and after surgically induced weight loss differ with regard to changes in basal levels and meal-related oscillations (9, 37).
In a previous study, we showed that weight-reduced subjects after RYGBP exhibit an early and exaggerated postprandial rise in total PYY concentrations, an absence of a compensatory increase in plasma ghrelin levels, and decreased fasting insulin and leptin concentrations compared with weight-matched controls (37). The aim of this study was to compare circulating concentrations of these hormones between RYGBP and BND subjects in an effort to understand the possible mechanisms for the different efficacies of these procedures. This report also differentiates between total PYY and ghrelin and the levels of the respective post-translational modified products, PYY(3–36) and octanoylated ghrelin.
Research Methods and Procedures
Four groups of adult women were studied and consisted of the following: Group 1, lean controls defined as BMI (weight in kilograms divided by the square of the height in meters) from 18 to 25 kg/m2 (n = 8); Group 2, overweight controls (OW) with no history of bariatric surgery who were BMI- and age-matched to subjects in Groups 3 and 4 (n = 11); Group 3, individuals who had undergone the BND procedure (n = 9); and Group 4, individuals who had undergone the RYGBP procedure (n = 9). BND and RYGBP groups were also matched for age, BMI, and the duration of the post-operative period. Subsets from some of these groups have been studied previously: seven from OW, five from RYGBP, and all lean controls (37). All subjects were weight stable, defined as weight change of <5% over the 3-month period preceding the study. Only women were studied because leptin and possibly ghrelin concentrations vary according to gender (38). Subjects were excluded from the study if they had diabetes; uncontrolled hypertension; current or recent tobacco use; use of prescription or over-the-counter weight loss products within the prior 6 months; cardiac, neurological, or renal disease; depression; eating disorder; use of antipsychotics, neuroleptics, opiates, or oral glucocorticoids; allergy to cocoa products; or lactose intolerance. All subjects signed an informed consent form approved by the Western Institutional Review Board or the Columbia University Institutional Review Board.
The surgical procedure for the BND consisted of placement of a 10-mm silastic band (Lap-Band; INAMED Health, Santa Barbara, CA) around the proximal stomach attached by tubing to a port sutured to the anterior rectus fascia. The diameter of the band and, thereby, the circumference through which nutrients must pass was adjusted post-surgically through the access port by removal or addition of saline within the band system. The surgical procedure for RYGBP consisted of creation of a 15- to 30-mL pouch that was divided from the proximal lesser curvature of the stomach and excluded the fundus. The pouch was anastomosed to a Roux limb of jejunum created by division of the jejunum 50 to 100 cm distal to the ligament of Treitz and anastomosing the afferent biliopancreatic limb to the jejunum 100 to 150 cm distally. Division of the vagus nerve and its branches was avoided. All surgical procedures were performed by the same surgeon (M.B.).
Subjects arrived at the Clinical Research Center in the morning after an overnight fast of at least 10 hours duration and were weighed in undergarments and a light gown. Subjects consumed a chilled chocolate-flavored test meal (Optifast; Novartis, Minneapolis, MN; 474 mL, 320 kcal, 50% carbohydrate, 35% protein, 15% fat) within a 15-minute period. Venous blood was drawn premeal and 30, 60, 90, 120, and 180 minutes after meal consumption. Subjects also completed a validated visual analogue scale (VAS) questionnaire at 0 and 180 minutes (39). The VAS consisted of 100-mm lines with words anchored at each end describing extreme sensations of hunger, satiety, and nausea or abdominal discomfort. Subjects were asked to make a vertical mark across the line corresponding to their feelings. Quantification was performed by measuring the distance from the left end of the line to the mark.
Plasma hormone measurements were performed on blood samples in EDTA tubes that were centrifuged for 15 minutes at 4 °C after collection and stored at −80 °C until assayed. Thawed samples were not refrozen for other assay measurements. Plasma to be used for quantification of octanoylated ghrelin was aliquoted immediately after centrifugation into tubes containing 25 μL of 1 N HCl per 0.5 mL of plasma. Leptin was measured with a human radioimmunoassay (RIA) kit (LINCO Research, Inc., St. Charles, MO) using a 125I-iodinated human leptin tracer. Total plasma immunoreactive ghrelin was measured by a RIA kit (Phoenix Pharmaceuticals, Belmont, CA) using 125I-iodinated ghrelin tracer and a rabbit polyclonal antibody against full-length, octanoylated human ghrelin that recognizes the acylated and des-acyl forms of the hormone. The lower limit of detection for this assay was 20 pg/mL, and the coefficient of variation was 8.5% within assays and 11.3% between assays. The octanoylated form of ghrelin was measured with an RIA kit (LINCO Research) using 125I-labeled ghrelin tracer and guinea pig antighrelin antibody that has <0.1% cross-reactivity with des-octanoyl ghrelin. The lower limit of detection for this assay was 8 pg/mL, with 7% intra-assay and 13% inter-assay coefficients of variation. Total plasma levels of PYY were measured using a commercial enzyme-linked immunosorbent assay (ELISA; Diagnostic Systems Laboratories, Webster, TX) that measures PYY(1–36) and PYY(3–36). The lower limit of detection was 12 pg/mL, and the coefficients of variation were 10.1% within and 10.3% between assays. This ELISA has replaced the RIA assay for PYY used in a previous report (37) that required extraction and column purification of larger plasma volumes. Plasma samples have been quantified in both assays, and although the absolute values obtained with the ELISA method are greater than values obtained with RIA, the pattern and significance of the results are the same (correlation coefficient from 77 samples run in duplicate, r = 0.68, p < 0.001). PYY(3–36) was quantified by RIA (LINCO Research) using an antibody with no measurable cross-reactivity to PYY(1–36) according to the manufacturer and tested in our laboratory by spiking plasma with known amounts of PYY(1–36) standard. Sensitivity was 20 pg/mL, and intra-assay and inter-assay coefficients of variation were 7.5% to 10.3% and 8.6% to 14.7%, respectively. The use of Aprotinin, recommended by the manufacturer, did not alter results and was, therefore, not used in this assay protocol. Plasma insulin was measured with the Immulite Analyzer with the lower limit of detection of 2 μIU/mL. Serum glucose was measured by the hexokinase method. Hormone measurements were performed on all subjects except where indicated. All samples were assayed in duplicate.
Significant differences between groups were determined by one-way ANOVA followed by Fisher's protected least difference test. p < 0.05 was considered statistically significant. Insulin resistance (IR) was calculated using the homeostasis model assessment (HOMA) (40, 41). Mean values ± standard error are reported.
Clinical Characteristics of Study Groups
The clinical characteristics of each of the four study groups are shown in Table 1. There were no significant differences in age, body weight, or BMI between the surgical groups and matched controls. The surgery subjects were also matched for duration of the post-operative period: 22.7 ± 2.0 months (range, 15 to 35) for BND and 26.7 ± 1.5 months (range, 22 to 34) for the RYGBP group. The initial BMI was 46 ± 2 in the BND and 55 ± 4 in the RYGBP group. The mean loss of initial total body weight was 25 ± 3% (range, 15 to 36) and 36 ± 3% (range, 25 to 51) in the BND and RYGBP groups, respectively.
Table 1. Clinical characteristics of study groups
Values are ± standard error.
30.6 ± 3.6
21.6 ± 0.7
57.5 ± 1.9
47.0 ± 4.2
34.4 ± 1.4
91.9 ± 5.1
42.4 ± 4.5
35.8 ± 1.9
97.4 ± 6.8
48.3 ± 2.7
34.2 ± 2.1
88.8 ± 5.4
Total PYY and PYY(3–36) Plasma Concentrations
Fasting total PYY plasma concentrations were not significantly different among groups: lean, 84 ± 23; OW, 88 ± 11; BND, 81 ± 13; and RYGBP, 107 ± 14 pg/mL. In response to the test meal, there was an early and exaggerated rise in total PYY levels in the RYGBP group, such that the concentration at 30 minutes was approximately 3- to 4-fold higher in comparison with all other groups (Figure 1A). Similarly, fasting PYY(3–36) immunoreactivity was not significantly different among groups: lean, 59 ± 9; OW, 70 ± 7; BND, 73 ± 6; and RYGBP, 86 ± 8 pg/mL. There was also an exaggerated postprandial rise in PYY(3–36) levels in the RYGBP group that was over 2-fold greater compared with the other groups (Figure 1B).
Total and Octanoylated Ghrelin Plasma Concentrations
Fasting total ghrelin plasma levels were not significantly different among groups (Figure 2). However, the magnitude of postprandial suppression of ghrelin was significantly blunted in the BND (13.1%) and OW (16.5%) subjects compared with the lean (26.6%) and RYGBP (26.9%) groups. Fasting levels of octanoylated ghrelin were higher in the lean controls (220 ± 36 pg/mL) compared with OW (120 ± 13; p < 0.01), BND (100 ± 16; p < 0.001), and RYGBP (163 ± 19; p = 0.068) subjects (Figure 2). In all groups, postprandial suppression of octanolyated ghrelin was more marked than with total ghrelin, but the magnitude of suppression was significantly less in the BND group (29.3%) compared with RYGBP group (56.1%; Figure 2).
Glucose, Insulin, and Leptin Concentrations
Fasting glucose concentrations were similar between BND and RYGBP; however, insulin values were significantly lower in the latter and nearly identical to lean controls (Figure 3). Although the RYGBP group exhibited exaggerated acute phase insulin secretion, levels rapidly decreased, and by 180 minutes, levels in RYGBP subjects (6.9 ± 0.5 μIU/mL) were again significantly less than BND subjects (12.7 ± 2.0; p < 0.05). IR, as determined by HOMA-IR, was significantly less in both BND (2.24 ± 0.32; p < 0.05) and RYGBP (1.46 ± 0.19; p < 0.001) compared with matched controls (3.50 ± 0.37), but the difference between surgical groups did not reach statistical significance (p = 0.129). Fasting plasma leptin concentrations were significantly higher in the BND and OW groups compared with lean controls (Figure 4) and tended to be greater in BND vs. RYGBP (p = 0.079).
VAS measurements were used to quantify feelings of hunger, satiety, and nausea or abdominal discomfort. Results are reported as the difference in rating 3 hours post-meal compared with the premeal value (Table 2). The RYGBP group reported an increase in satiety that was significantly greater compared with the change in satiety in the OW and BND subjects. No differences among groups were detected in hunger or post-meal abdominal discomfort.
Table 2. Change in VAS rating at 3 hours compared with fasting
VAS, visual analogue scale. Values are ± standard error.
Fasting and postprandial concentrations of circulating hormones involved in the regulation of energy balance and glucose homeostasis were examined in this study to establish possible mechanisms for the greater efficacy of RYGBP procedure compared with BND. Several important similarities and differences in hormonal regulation were evident among the four study groups: fasting total and PYY(3–36) concentrations were similar among all groups, but the postprandial PYY response in the RYGBP group was exaggerated, such that 30 minutes after the meal, levels were 2- to 4-fold higher compared with all other groups; maximal postprandial suppression of total and octanoylated ghrelin was blunted in the BND group compared with the RYGBP group; fasting concentrations of leptin were similar between lean and RYGBP groups, whereas BND subjects had significantly greater leptin levels compared with lean controls; and fasting concentrations of insulin were significantly lower in RYGBP compared with BND subjects.
The greater and earlier postprandial rise in PYY concentrations after RYGBP may aid in earlier satiety and meal termination. Indeed, VAS results in this study indicate that RYGBP subjects experience a greater degree of satiety compared with BND and OW. The role of PYY(3–36) as a satiety factor is supported by studies in humans (20, 21, 22) and non-human primates (18, 19) showing that peripheral administration of PYY(3–36) decreases food intake. A significant decrease in body weight was also observed after administration for 2 weeks in monkeys (19). However, discrepant results of the effects of PYY(3–36) on food intake and body weight in rodents have been reported (17). Although it has been reported that obese individuals exhibit lower fasting levels of PYY compared with lean subjects (21, 42), we and others (43) have not demonstrated this difference, suggesting that the postprandial change in PYY concentration may be of greater physiological importance than fasting levels. Of note, two different assays were used in this report to measure PYY immunoreactivity. One assay utilized an antibody that recognizes both PYY(1–36) and PYY(3–36), and another assay used an antibody specific for PYY(3–36) with no demonstrable cross-creativity with PYY(1–36). Characterization of these various peptide products is important because PYY(1–36) and PYY(3–36) have different potencies with regard to food intake in rodents (44) and different affinities for the various NPY receptor subtypes (15, 16). Specifically, we have shown that both total PYY immunoreactivity and PYY(3–36) are increased several-fold in the postprandial state of RYGBP subjects. Interestingly, the levels obtained 30 to 60 minutes post-meal in RYGBP patients are similar to the supraphysiological levels required for acute reductions in food intake in lean male volunteers (22). It is also possible that other post-translational modifications of PYY occur, but our methodology would not be sufficient to decipher such differences.
The mechanism of the early and exaggerated PYY response after RYGBP is unclear but may be due to expedited delivery of nutrients to the small bowel. In fact, partial resection of the small bowel in man results in elevated fasting plasma PYY and a substantially increased postprandial response (45). It is unlikely that the difference in PYY secretion is attributable to greater weight loss in the RYGBP group because this exaggerated response occurs as early as 6 months post-surgery when weight reduction can be matched to post-op BND patients who do not exhibit this PYY response (data not shown). Rodent models have also been studied to evaluate the contribution of gastrointestinal hormones to body weight regulation. Strader et al. (46) have shown that transposition of a segment of ileum to the jejunum results in increased synthesis and release of PYY and another ileal hormone, glucagon-like peptide-1. Interestingly, ileal transposition results in more weight loss and decreased food consumption compared with control rats with intestinal transections and reanastomosis without transposition.
In this study, we have also shown that maximal postprandial suppression of total and octanoylated ghrelin was blunted in the BND group but not the RYGBP group. If the fall in ghrelin is perceived as a signal to terminate a meal, it is conceivable that this signal is impaired in the BND subjects or that greater caloric intake is required to suppress ghrelin to the extent observed in the RYGBP group. The effects of RYGBP on concentrations of total ghrelin (octanoylated ghrelin concentrations have not been reported by others) are inconsistent among earlier reports (9, 10). Fasting ghrelin concentrations have been shown to decrease after RYGBP or be equivalent to either preoperative levels or non-surgical obese controls (36, 47, 48, 49, 50, 51, 52), and in some of these studies, meal-related fluctuations of ghrelin levels were either absent (36) or inconsistent (47). Plasma ghrelin levels have also been observed to increase after RYGBP in patients who were experiencing active weight loss (49, 50), an observation that is consistent with findings in subjects who have undergone diet-induced weight loss (36). Results from studies that examine the affect of BND on fasting ghrelin concentrations are also varied and range from less than (47), equivalent to (53), or greater than (52) BMI-matched controls. Possible explanations for varied results include different surgical techniques, such as location and circumference of the band, location of the staple line that partitions the stomach and determines the size of the upper gastric pouch, or length of the biliopancreatic limb and the site of anastomosis to the jejunum (9). Surgical disruption of the various autonomic nerve fibers that innervate the foregut may also affect both short- and long-term regulation of ghrelin secretion (54).
Fasting levels of insulin were decreased in both surgical groups compared with the weight-matched controls and significantly lower in RYGBP compared with BND subjects; however, HOMA-IR indices were not significantly different between RYGBP and BND. The lower levels of insulin in the RYGBP group may be due to factors that contribute independently to insulin sensitivity, such as a greater degree of weight loss and/or lower energy intake (55). More detailed studies are warranted to determine whether bypassing the foregut confers an additional improvement in insulin sensitivity. Interestingly, although the surgical groups were matched for age, BMI, and post-operative period, the BND group had significantly higher leptin levels, whereas the RYGBP group had similar levels compared with lean controls. The difference between BND and RYGBP approached but did not reach statistical significance (p = 0.079). It is unlikely that the difference in leptin levels is due to a greater reduction in body weight in the RYGBP group because leptin concentrations when normalized to fat mass do not decrease significantly more after a loss of >10% of body weight (55, 56). The exaggerated reduction in leptin levels after RYGBP may be indicative of a restoration of leptin sensitivity; however, analysis of body composition needs to be performed in these subjects to determine whether similar findings are obtained after leptin concentrations are normalized to fat mass.
In conclusion, gastric bypass and banding are effective procedures that induce weight loss; however, gastric bypass usually promotes a greater degree of weight loss and is associated with different changes in a number of hormones that regulate energy homeostasis. The greater postprandial increase in PYY and a greater suppression of ghrelin in RYGBP subjects may promote increased satiety and earlier meal termination, thus producing greater weight loss and maintenance compared with BND subjects. It will be important to decipher whether these hormonal differences are a cause of the greater degree of weight loss associated with RYGBP or are a by-product of anatomical differences and have no bearing on procedural efficacy. Other notable differences between these procedures are lower insulin and leptin levels in the RYGBP group. Further studies are necessary to determine whether these lower levels reflect a greater amelioration of the insulin and leptin resistance often associated with obesity or a difference in energy partitioning. A long-term prospective analysis is warranted to determine whether changes in these hormones correlate with the degree of weight reduction and, in particular, maintenance of weight loss. If such changes are, indeed, associated with long-term efficacy, then evaluation of surgical techniques or pharmaceuticals that optimize or reproduce these changes may be beneficial in the design of obesity therapies.
We thank the participants in this study and the staff of the Irving Center for Clinical Research. We also acknowledge the excellent technical assistance of Robert Sundeen. This work was supported by NIH Grants DK072011 (to J.K.) and RR00645 (to the General Clinical Research Center).
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.