Differential Effects of Laparoscopic Sleeve Gastrectomy and Laparoscopic Gastric Bypass on Appetite, Circulating Acyl-ghrelin, Peptide YY3-36 and Active GLP-1 Levels in Non-diabetic Humans

Laparoscopic Roux-en-Y gastric bypass (LRYGBP) reduces appetite and induces significant and sustainable weight loss. Circulating gut hormones changes engendered by LRYGBP are implicated in mediating these beneficial effects. Laparoscopic sleeve gastrectomy (LSG) is advocated as an alternative to LRYGBP, with comparable short-term weight loss and metabolic outcomes. LRYGBP and LSG are anatomically distinct procedures causing differential entero-endocrine cell nutrient exposure and thus potentially different gut hormone changes. Studies reporting the comparative effects of LRYGBP and LSG on appetite and circulating gut hormones are controversial, with no data to date on the effects of LSG on circulating peptide YY3-36 (PYY3-36) levels, the specific PYY anorectic isoform. In this study, we prospectively investigated appetite and gut hormone changes in response to LRYGBP and LSG in adiposity-matched non-diabetic patients. Anthropometric indices, leptin, fasted and nutrient-stimulated acyl-ghrelin, active glucagon-like peptide-1 (GLP-1), PYY3-36 levels and appetite were determined pre-operatively and at 6 and 12 weeks post-operatively in obese, non-diabetic females, with ten undergoing LRYGBP and eight adiposity-matched females undergoing LSG. LRYGBP and LSG comparably reduced adiposity. LSG decreased fasting and post-prandial plasma acyl-ghrelin compared to pre-surgery and to LRYGBP. Nutrient-stimulated PYY3-36 and active GLP-1 concentrations increased post-operatively in both groups. However, LRYGBP induced greater, more sustained PYY3-36 and active GLP-1 increments compared to LSG. LRYGBP suppressed fasting hunger compared to LSG. A similar increase in post-prandial fullness was observed post-surgery following both procedures. LRYGBP and LSG produced comparable enhanced satiety and weight loss. However, LSG and LRYGBP differentially altered gut hormone profiles.


Introduction
Obesity is a leading cause of morbidity and mortality, placing growing demands on healthcare systems. Bariatric surgery induces significant long-lasting weight loss, ameliorates obesity-associated co-morbidities and reduces mortality [1]. Laparoscopic Roux-en-Y gastric bypass (LRYGBP) is the 'gold standard' procedure, resulting in 65-80 % excess bodyweight loss, decreased appetite, and rapid weightindependent amelioration of type-2 diabetes mellitus (T2DM) [2][3][4][5]. LRYGBP is cost-effective, but technically challenging with associated mortality-albeit low ∼0.09 % [6]-and micronutrient deficiencies risks, necessitating lifelong followup. Laparoscopic sleeve gastrectomy (LSG) (originally undertaken as a first-step in super-obese patients with subsequent conversion to a hybrid restrictive-malabsorptive procedure) is technically less complex, with lower complications and nutritional deficiencies rates than LRYGBP [2]. In light of reports of comparable weight loss and metabolic outcomes to LRYGBP, LSG is increasingly undertaken as a stand-alone procedure [7][8][9][10]. However, its long-term efficacy for weight loss and metabolic benefit remains unclear [8,10].
Understanding the mechanisms mediating the weight loss and metabolic effects of bariatric surgery is key for developing less invasive procedures and medical obesity treatments. Postoperative changes in circulating gut hormones, including ghrelin, peptide YY (PYY) and glucagon-like peptide-1 (GLP-1), are thought to play a key role to the beneficial outcomes of bariatric surgery [2].
Ghrelin is produced by X/A-like cells predominantly located in the stomach fundus and proximal small intestine. Its circulating levels increase with fasting and decrease postprandially. Acyl-ghrelin, the bioactive form, exerts orexigenic properties and is produced by ghrelin octanoylation in serine-3 mediated by ghrelin-O-acyl transferase (GOAT). Acylghrelin is rapidly converted by endogenous esterases to the main circulating form, des-acyl-ghrelin [11]. Moreover, ghrelin is labile, with highest stability in acidic states [12]. Thus, plasma acyl-ghrelin assessment requires specific sample processing by addition of esterase inhibitor and plasma acidification [13].
LRYGBP and LSG differentially alter gastrointestinal anatomy. LRYGBP reduces stomach volume and bypasses the majority of the stomach, duodenum and proximal jejunum, with direct nutrient delivery to the distal gut [2]. In LSG, the gastric fundus (the major source of ghrelin source) is excised, and this accelerates gastric emptying and intestinal transit post-operatively resulting in rapid nutrient delivery to the duodenum and hindgut [25][26][27]. Thus, LSG and LRYGBP produce differential nutrient exposure of entero-endocrine cells, and as such would be anticipated to differentially alter circulating gut hormones. Most studies comparing the two procedures report larger decreases in circulating fasting and/or meal-stimulated ghrelin after LSG versus LRYGBP [21,[28][29][30]; however, findings of their effects on hindgut hormones have been inconsistent, reporting either comparable increases in GLP-1 and/or total-PYY [8,21,28,29,31]; or superior total-PYY [32] and GLP-1 [30] increases post-LRYGBP versus LSG. Technical procedural variations, differences in hormonal isoforms assessed, time-lapse from surgery, subjects' HOMA-IR and glycaemic status, and differences in subject standardization and sample processing may account for these discrepancies.
Several studies have measured total-ghrelin, total-GLP-1 and/or total-PYY post-surgery, which depict hormone production, but may not necessarily reflect circulating levels of their respective bioactive forms. In support of this, DPP-4 activity declines after LRYGBP [33], whereas GOAT expression is altered by caloric restriction and plasma GOAT is BMIdependent and thus may change post-bariatric surgery [34][35][36]. Moreover, the effects of LSG on the anorectic PYY-isoform, PYY3-36, are unknown. Therefore, we prospectively compared the effects of LSG and LRYGBP on anthropometric indices, leptin, acyl-ghrelin, active GLP-1, PYY3-36, and appetite in non-diabetic patients using our established subject standardisation and stringent sampleprocessing protocols.

Surgical Procedures
LRYGBP included an antecolic-antegastric Roux-en-Y construction with 120-cm alimentary/80-cm biliopancreatic limbs and a small vertical gastric pouch. LSG was performed as previously described [37] and calibrated tightly with a 32-Ch bougie with stapling commenced 5 cm from the pylorus.

Study Protocol
Patients attended for a 500-kcal test meal (43 % carbohydrate/ 18 % protein/39 % fat) within 2 weeks pre-surgery and at 6 and 12 weeks (6w and 12w) post-operatively. Subjects maintained similar food intake, refrained from alcohol 24 h prior to each study day, fasted from 9 p.m. and drank only water. At ∼9 a.m. on study days, a peripheral cannula was inserted, and 1 h was allowed for recovery/habituation [13]. At t0 min, a baseline visual analogue scale (VAS) [13] and a blood sample were collected. Subjects consumed the test meal (250 ml Resource2.0+fibre, Nestle Nutrition, Croydon, UK) within 15 min, with blood drawn at t15, t30 and every 30 min thereafter until t180 min post-meal. Coincident with blood sampling, subjects completed appetite VAS [13]. Samples were processed by previously described stringent protocols [13].

Statistical Analysis
Results are expressed as mean ± standard error of the mean. Normal distribution was assessed by D'Agostino-Pearson omnibus normality test. Integrated area-under-the-curve (AUC) t0-t180 for appetite and hormone concentrations versus time was calculated using the trapezoid rule. Non-paired student's t test was used for between-groups comparisons, and paired student's t test and repeated measures one-way ANOVA with Bonferroni post-hoc for within-groups' analysis. p< 0.05 was considered significant.

Preoperative Patient Characteristics
The two groups had comparable age, BMI and adiposity and exhibited similar circulating fasting leptin, acyl-ghrelin, PYY3-36, active GLP-1, glucose, insulin levels and HOMA-IR at baseline (Fig. 1, Tables 1 and 2). In response to the test meal, we observed decreases in hunger and acyl-ghrelin and increases in perceived satiety and in plasma PYY3-36, active GLP-1, glucose, and insulin, with no between-groups differences ( Fig. 1, Tables 1 and 2).

Post-surgery Changes in Adiposity and Plasma Leptin
Both procedures induced comparable, marked reductions in BMI, fat mass, visceral fat area (VFA) and fasting leptin, with similar BMI and %EWL observed at 6w and 12w postsurgery (Tables 1 and 2).

Post-operative Subjective Hunger and Satiety Scores
Fasting hunger significantly decreased post-LRYGBP, but remained unchanged post-LSG. Both procedures comparably reduced hungerAUC 0-180 in response to the standard test meal (Fig. 4, Table 2). At 12w, fasting hunger was lower post-LRYGBP versus LSG, with no other between-groups differences in hunger (Fig. 4, Table 2).

Discussion
LRYGBP and LSG reduced BMI, excess weight, adiposity and plasma leptin at 6w and 12w post-operatively to a similar extent. Hence, the observed differences in appetite and gut hormones were not attributable to differences in weight loss and more likely reflect differences in the surgical procedures per se.
Our study reports the first comparison of plasma acylghrelin in non-diabetic patients' post-LRYGBP and LSG. These procedures produce differential nutrient contact with ghrelin-producing X/A -like cells. Post-LRYGBP, X/A -like cells in the gastric fundus and duodenum remain in situ but are excluded from nutrient contact. Diet-induced weight loss increases ghrelin [19]. Yet, despite marked weight loss, post-LRYGBP fasting acyl-ghrelin non-significantly decreased at 6w but rose towards baseline values by 12w, whereas at t30 post-meal declined from pre-surgery at 6w and 12w. Our results suggest that stomach fundus and duodenal ghrelinproducing cells contribute to circulating ghrelin post-LRYGBP despite their exclusion from nutrient contact and that mechanisms independent from X/A -like cell nutrientsensing, for example the vagus nerve [39], may signal mealinduced ghrelin suppression. Chronaiou et al. provided direct evidence that fundus ghrelin-producing cells remain active post-LRYGBP by demonstrating superior decreases in ghrelin post-LRYGBP+fundus resection versus LRYGBP with fundus preservation [23]. Interestingly Barazzoni et al. reported no change in total ghrelin, but significantly increased acylghrelin at 1, 3, 6 and 12 months post-LRYGBP [24]. Substantial differences in sample handling by addition of esterase inhibitor and plasma acidification in our study may underlie these discrepant findings.
LSG removes gastric fundus ghrelin-producing cells and accelerates nutrient delivery to duodenal ghrelin-producing cells by increasing gastric emptying [25][26][27]. After LSG, 40-50 % decreases in fasting total-ghrelin have been reported, sustained for up to 5 years post-surgery, with reductions in post-meal circulating ghrelin levels [21,28,40]. We also observed reductions in fasting acyl-ghrelin and acyl-ghrelinAUC 0-180 post-LSG. Moreover, as anticipated in view of the stomach fundus excision, LSG induced superior acylghrelin reductions than LRYGBP. Interestingly, despite removing the majority of the gastric fundus, fasting acylghrelin and acyl-ghrelinAUC 0-180 declined by only ∼20-30 % post-LSG. A possible explanation is that although the majority of ghrelin-producing cells are in the stomach fundus, circulating acyl-ghrelin may primarily originate from the duodenum. Alternatively, plasma acyl-ghrelin is highly regulated, and compensatory up-regulation of duodenal ghrelinproduction may occur.
PYY and GLP-1 are released from distal gut L-cells. Postmeal their levels rapidly rise, implicating a yet unknown neural and/or humoral mechanism to this initial release. Subsequent PYY and GLP-1 release results from L-cell nutrient-contact. LRYGBP and LSG increase gastric emptying [25][26][27]41] but have different effects on gut nutrientpassage. LRYGBP excludes nutrients from foregut contact and expedites nutrient delivery to distal gut L-cells, which is suggested to augment hindgut hormone release; the 'hindgut theory'. LSG accelerates gastric emptying, reduces acid production and rapidly transits nutrients into the duodenum and proximal intestine, enhancing foregut stimulation [2]. Few studies have examined the effects of LRYGBP on circulating PYY3-36, the anorectic PYY-isoform. This is the    [32]. However, they are at odds with reports of comparable post-prandial total-PYY following LRYGBP and LSG [21,29,31]. Differences in PYY-isoforms assessed, sampling timepoints, subject standardization, sample handling and in subjects' HOMA-IR pre-surgery may account for these discrepancies.
Similarly to PYY3-36, both procedures augmented nutrient-stimulated active GLP-1 levels, with again greater, more sustained release observed post-LRYGBP. These findings are at odds with reports by Chambers et al. of similar increases in nutrient-stimulated active GLP-1 post-sleeve gastrectomy and gastric bypass in rats [8]. These discrepancies may be accounted for by structural inter-species rodenthuman stomach differences, which potentially affect gastric emptying post-surgery and hence gut hormone responses. Moreover, Chambers and colleagues measured active GLP-1 5 months post-surgery in weight-stable rats, whilst our studies were undertaken during the acute weight-loss phase. Peterli et al. have reported greater meal-stimulated active GLP-1 responses at 1 week and non-significant increases 12 weeks post-LRYGBP versus LSG [21]. In another study, they showed non-significantly greater active GLP-1 peak and active GLP-1AUC following LRYGBP [29]. Again, methodological and subject-related differences may underlie these discrepancies.
Despite greater reductions in the 'hunger hormone' acylghrelin post-LSG versus post-LRYGBP, paradoxically the Fig. 2 The effects of LRYGBP and LSG on plasma fasting, mealstimulated acyl-ghrelin and acyl-ghrelinAUC 0-180 . Plasma acyl-ghrelin temporal profile in response to the test-meal for LRYGBP (a) and LSG (b) groups at pre-surgery (black, solid squares), and at 6w and 12w postoperatively (red, solid circles and green, solid triangles, respectively). c Acyl-ghrelinAUC 0-180 for LRYGBP (black, solid columns) and LSG groups (grey, solid columns) at pre-surgery and at 6w and 12w post-operatively. Results are expressed as mean±SEM. *p< 0.05, withingroup at 6w post-operatively compared to pre-surgery; † p< 0.05, within-group comparisons at 12w post-operatively versus pre-surgery; ‡ p< 0.05, within-group 6w versus 12w comparison. The p values at the right upper corner of c indicate one-way ANOVA within-group analysis. Within-group Bonferroni post hoc and between-group t test significance is indicated over the corresponding bars Fig. 3 The effects of LRYGBP and LSG on fasting, meal-stimulated plasma concentrations and area-under-the curve (AUC 0-180 ) for PYY  and active GLP-1. Plasma PYY 3-36 temporal profile in response to the test-meal for LRYGBP (a) and LSG (b) groups at pre-surgery (black, solid squares) and at 6w and 12w post-operatively (red, solid circles and green, solid triangles, respectively). Plasma active GLP-1 temporal profile in response to the test-meal for LRYGBP (c) and LSG (d) groups at pre-surgery (black, solid squares) and at 6w and 12w post-operatively (red, solid circles and green, solid triangles , respectively). PYY 3-36 AUC 0-180 (e) and active GLP-1AUC 0-180 (f) for LRYGBP (black, solid columns) and LSG groups (grey, solid columns) at pre-surgery and at 6w and 12w post-operatively. Results are expressed as mean±SEM. *p < 0.05, **p <0.01 and ***p <0.001 within-group at 6w post-operatively compared to pre-surgery. † p <0.05, † † p <0.01 and † † † p <0.001 for withingroup comparisons at 12w post-operatively versus pre-surgery. The p values at the right upper corner of e and f indicate one-way ANOVA within-group analysis. Within-group Bonferroni post hoc and betweengroup t test significance is indicated over the corresponding bars Fig. 4 The effects of LRYGBP and LSG on fasting, meal-stimulated and area-under the curve (AUC 0-180 ) for subjective hunger and fullness VAS ratings. Hunger VAS temporal profile in response to the test-meal for LRYGBP (a) and LSG (b) groups at pre-surgery (black, solid squares), and at 6w and 12w post-operatively (red, solid circles and green, solid triangles, respectively). Fullness VAS temporal profile in response to the test-meal for LRYGBP (c) and LSG (d) groups at pre-surgery (black, solid squares) and at 6w and 12w post-operatively (red, solid circles and green, solid triangles , respectively). HungerAUC 0-180 (e ) and fullnessAUC 0-180 (f) for LRYGBP (black, solid columns) and LSG groups (grey, solid columns) at pre-surgery and at 6w and 12w postoperatively. Results are expressed as mean±sem. *p<0.05, **p<0.01 and ***p<0.001 within-group at 6w post-operatively compared to presurgery. † p <0.05, † † p <0.01 and † † † p <0.001 for within-group comparisons at 12w post-operatively versus pre-surgery. The p values at the right upper corner e and f indicate one-way ANOVA within-group analysis. Within-group Bonferroni post hoc and between-group t test significance is indicated over the corresponding bars LRYGBP group exhibited lower fasting hunger; post-prandial hunger was reduced comparably by both procedures. We also observed similar post-operative increments in nutrientstimulated fullness perception in both groups. These findings again are slightly at odds with the accepted notion that PYY3-36 and active GLP-1 mediate satiety, as LRYGBP induced superior increases of these anorectic peptides and thus would be expected to result in greater satiety perception. These findings highlight that additional factors to active GLP-1 and PYY3-36 regulate satiety.
The novel finding of our study is the characterisation for the first time of the effects of LSG on circulating levels of the anorectic PYY-isoform, PYY3-36. Moreover, our study is the first to simultaneously measure bioactive forms of ghrelin, PYY and GLP-1 in the same patient cohort, while concurrently undertaking parallel appetite assessment. The main strength of our study is the use of validated subject-standardisation protocols, stringent sample processing [13], and tight groupmatching pre-operatively for the gut hormone confounders age [42], sex [43] and adiposity [44]. Furthermore, we studied patients without T2DM, dissecting out confounding effects of T2DM on the incretin effect [45], circulating PYY3-36 [5] and ghrelin [46]. The limitations of our study are our small sample sizes, non-randomization and limited follow-up. We studied females only as these represent the majority of patients undergoing bariatric surgical procedures in the UK, and future studies are needed to assess whether gender differences exist in postoperative gut hormone changes. Moreover, studies in patients with T2DM are required to examine whether their inferior weight-loss outcome post-bariatric surgery [47] results from altered/aberrant gut hormone responses. Future larger, randomised studies with longitudinal assessment of gut hormones, intestinal transit, glycaemic and anthropometric indices are required to further elucidate the mechanisms underlying the beneficial effects of LRYGBP and LSG in an attempt to develop novel, less invasive surgical and non-surgical T2DM and obesity treatments.