This was a prospective experimental comparative animal study conducted at the Center for Surgical Education and Research (CERC – Centre d’Enseignement et Recherche en Chirurgie) of the Faculty of Medicine North at the Aix-Marseille University (France). The study design, the care and the handling of animals were approved by the institutional review board of the Aix-Marseille University (Ethical comitee #14) and French Authorities (Ministère de l’Enseignement et de la Recherche, authorization APAFIS #22017033011503900 v3).
Pigs were given by INRA (Institut National de la Recherche Agronomique, Rennes, France). All applicable institutional and/or national guidelines for the care and use of animals were followed. Four groups were predefined: group 1 (G1), consisting of endoscopic bypass with GJA without pyloric closure; group 2 (G2) consisting of endoscopic bypass with GJA and surgical duodenal exclusion; group 3 (G3) consisting of surgical RYGB; and group 4 (G4) consisting of a control group, without any intervention.
The endoscopic procedures were performed by two experts in therapeutic endoscopy (MB and J-MG), while surgical bypass procedures were performed by two surgeons, experienced in bariatric surgery (DB and LB).
Animals handling and anesthesia protocol.
All animals were obese Yucatan pigs, aged of 12 months old, rendered obese with insulin-resistance after hypercaloric alimentation since they were 9 months old at the INRA.
All animals arrived 5 days before the procedures at the CERC for acclimation and were housed individually. They received water and pig chow diet. All of them received a 14 cm double lumen (20 Gauge) venous central catheter (VCC) (Arrow, Kendall Health Care Products, Mansfield, EU) 2 days before intervention, placed in jugular position and immediately tunnelized. They were used for test meals, initially left in place in the first animals, and then removed immediately after the tests because of infection and lethal vascular complications. Feeding was stopped 24h before endoscopic or surgical intervention.
For anesthesia induction, animals received intramuscularly injection of both azaperone 1mg/kg and ketamine 5mg/kg. Anesthesia was maintained with continuous intravenously injection of 100mg/h of propofol 2% and 100 micrograms/hour of remifentanil for analgesia. They were intubated and mechanically ventilated. Perioperative antibiotic prophylaxis was administrated by intravenously injection of Cefoxitine 2g and continued twice daily during postoperative period. Each animal had close monitoring with heart rate and oxygen saturation during the procedure, performed in supine position in the three intervention groups.
Procedures of GJA creation were previously described by our team (11, 12). Briefly, a dual-channel video gastroscope (3.7 and 2.8mm; XGIF-2T180H; Olympus Europe, Hamburg, Germany) was used and the following procedures were performed in the 2 first endoscopic groups. For this study, measurement of the bypassed limb length was done surgically despite the GJA was done endoscopically:
Mini laparotomy for limb selection. A surgical median laparotomy was performed by the surgeons and limb selection at 300 cm from the pylorus.
Endoscopic gastric parietal incision, performed in the horizontal portion of the anterior preantral zone, away from the small and large curvature, using a Hook Knife (Olympus, Japan).
Access to the peritoneal cavity, followed by prehension of the jejunal loop, presented by the surgeon, using a twin grasper forceps (OTSC Twin Grasper; OVESCO), and a 0.035” guidewire was inserted in the limb, after a parietal puncture with a 19 Gauge needle.
LAMS (Boston Scientific, USA) insertion overt the wire, and deployment. First the distal flange was into the jejunum, then the limb was then gently pulled into the gastric lumen, using both distal flange of the stent and the twin-grasper, followed by proximal flange into the stomach.
In the second group, to mimic a surgical bypass-like malabsorptive effect, a laparotomic duodenal exclusion was surgically performed with a stapler placed at the level of the genu superius.
Surgical bypass was performed following standard way, using conventional laparotomy equipment. A classic RYGB was created, with a 300 cm bypass limb, a pancreatobiliary limb and a gastric pouch. The 300cm length was decided because the small bowel length in a pig is twice than humans, that why we decided to double the length of the alimentary loop to mimic the same ratio than applied in humans’ by-passes.
Follow-up and euthanasia.
After procedure, each animal was clinically observed during a period of 3 months. They were maintained nihil per os during the first 24h, followed by water at the 1st postoperative day (POD) and finally progressive re-feeding at the POD3 (they received a quarter of the usual pig chow for 48h, then half pig chow for 48h, followed by 3 quarters for 48h, before being fed normally until the end of follow-up). All animals received a standardized meal. Antibiotic prophylaxis was continued for three days. Different clinical parameters (overall behavior, food intake, temperature, pain, bowel and urinary functions) were monitored intensively the first 2POD, and twice daily after. Intramuscular injection of tramadol 100mg was administered twice daily during the three first days, and after in case of signs of pain. Failure to eat, vocalization and teeth grinding were considered as signs of pain.
At the end of follow-up, animals from endoscopic G1 and G2 had endoscopic LAMS retrieval, followed by GJA evaluation through laparotomy. Then all survival animals were sacrificed with administration of a lethal dose of pentobarbital. Necropsies were performed among animals with premature death, and all surviving animals from G1 and G2, allowing macroscopical and microscopical anastomosis evaluation.
Metabolic and hormonal assessment.
For biological assessment, all samples were taken in CERC, through the VCC, and plasmatic dosages were performed in the research Unit UMR S 1260 of the faculty of Medicine, University Aix-Marseille.
A test meal was performed in each animal of all groups, for following molecules: glucose, insulin, peptide YY (PYY), xylose, FGF-19, FGF-21, GLP-1 and ghrelin.
Before the procedure and after 12 hours of fasting, all animals had a meal containing 1925 kcal for 30 minutes. Blood samples were collected 15 minutes before the meal (t0), and then dynamic dosages were performed at 30, 60 and 120 minutes after the meal, except for the ghrelin (only a fasting sample was performed).
After the end of the FU, survival animals had a second test meal, under exactly same conditions as before procedure.
Endpoints and outcomes.
The primary endpoints of this study were the weight after a follow-up of three months in each animal group. Secondary endpoints included the rate of perioperative mortality and morbidity, histological anastomosis analysis and biological analysis.
Descriptive statistical analysis of quantitative variables were expressed as median (with range) or mean (with standard deviation), while qualitative variables are expressed as a percentage. Non-parametric Mann-Withney tests were used to determine the significance of difference between 2 groups means and non-parametric Kruskal-Wallis test was performed for weight comparison between multiple groups. All tests were two-sided, with significant level determined at 5%. All statistical analyses were performed using SPSS.