Metabolic improvements in obese type 2 diabetes subjects implanted for 1 year with an endoscopically deployed duodenal–jejunal bypass liner

DIABETES TECHNOLOGY & THERAPEUTICSVolume 14, Number 2, 2012 ª Mary Ann Liebert, Inc.
DOI: 10.1089/dia.2011.0152 Metabolic Improvements in Obese Type 2 Diabetes Subjects Implanted for 1 Year with an Endoscopically Eduardo G.H. de Moura, M.D., Ph.D.,1 Bruno C. Martins, M.D.,1 Guilherme S. Lopes, M.D.,1 Ivan R. Orso, M.D.,1 Suzana L. de Oliveira, B.S.,1 Manoel P. Galva˜o Neto, M.D.,2 Marco A. Santo, M.D.,3 Paulo Sakai, M.D., Ph.D.,1 Almino C. Ramos, M.D.,2 Arthur B. Garrido Ju´nior, M.D.,3 Marcio C. Mancini, M.D.,4 Alfredo Halpern, M.D., Ph.D.,4 and Ivan Cecconello, M.D., Ph.D.3 Background: The purpose of this study was to evaluate the effect of the duodenal–jejunal bypass liner (DJBL), a60-cm, impermeable fluoropolymer liner anchored in the duodenum to create a duodenal–jejunal bypass, onmetabolic parameters in obese subjects with type 2 diabetes.
Methods: Twenty-two subjects (mean age, 46.2 – 10.5 years) with type 2 diabetes and a body mass index between40 and 60 kg/m2 (mean body mass index, 44.8 – 7.4 kg/m2) were enrolled in this 52-week, prospective, open-label clinical trial. Endoscopic device implantation was performed with the patient under general anesthesia, andthe subjects were examined periodically during the next 52 weeks. Primary end points included changes infasting blood glucose and insulin levels and changes in hemoglobin A1c (HbA1c). The DJBL was removedendoscopically at the end of the study.
Results: Thirteen subjects completed the 52-week study, and the mean duration of the implant period for allsubjects was 41.9 – 3.2 weeks. Reasons for early removal of the device included device migration (n = 3), gas-trointestinal bleeding (n = 1), abdominal pain (n = 2), principal investigator request (n = 2), and discovery of anunrelated malignancy (n = 1). Using last observation carried forward, statistically significant reductions in fastingblood glucose ( - 30.3 – 10.2 mg/dL), fasting insulin ( - 7.3 – 2.6 lU/mL), and HbA1c ( - 2.1 – 0.3%) were ob-served. At the end of the study, 16 of the 22 subjects had an HbA1c < 7% compared with only one of 22 atbaseline. Upper abdominal pain (n = 11), back pain (n = 5), nausea (n = 7), and vomiting (n = 7) were the mostcommon device-related adverse events.
Conclusions: The DJBL improves glycemic status in obese subjects with diabetes and therefore represents anonsurgical, reversible alternative to bariatric surgery.
because of the increased risk of diabetes-associated morbidity,including heart disease, peripheral vascular disease, stroke, The prevalence of obesity (defined as having a body kidney disease, retinopathy, and neuropathy.
mass index [BMI] of ‡ 30 kg/m2) is increasing in both Treatment of diabetes is focused on glycemic control and developing as well as developed countries.1 Obesity is a major typically proceeds in a stepwise fashion when target glycemic risk factor for type 2 diabetes,2 and thus it is not surprising goals are not reached or maintained.7,8 The initial steps are that the global prevalence of type 2 diabetes continues to in- lifestyle changes aimed at controlling diet and increasing ac- crease.3 In 2010, the estimated global prevalence of diabetes tivity with the goal of reducing body weight, followed by the among adults was 6.4% and is projected to increase to 7.7% by addition of orally active pharmacologic agents (sulfonylureas, 2030.4 In Brazil, the 2006 prevalence of diabetes based on self- metformin) and insulin to the treatment regimen.7,8 reports was 5.3%,5 although higher estimates have been re- Of interest is the observation that obese patients with dia- ported.6 This increasing prevalence of diabetes is alarming betes who undergo certain gastric bypass procedures 1Gastrointestinal Endoscopy Unit, 3Gastrointestinal Surgery Unit, Gastroenterology Department, and 4Endocrinology Department, University of Sa˜o Paulo Medical School, Sa˜o Paulo, Brazil.
2Gastro Obeso Center, Sa˜o Paulo, Brazil.
demonstrate improvement in glycemia, often within days of disease or other GI disease, pancreatitis, symptomatic coro- surgery and before significant weight loss.9,10 The mechanism nary artery disease or pulmonary dysfunction, known gall- responsible for this improvement is not fully understood; stones, severe coagulopathy, upper GI bleeding conditions however, the surgical rearrangement of the anatomy of the (e.g., esophageal or gastric varices, congenital or acquired gastrointestinal (GI) tract changes the location where partially intestinal telangiectasia), congenital or acquired anomalies of digested nutrients first contact the intestine, suggesting that the GI tract (e.g., atresias or stenoses), inability to discontinue correction of dysfunctional homeostatic mechanisms may nonsteroidal anti-inflammatory drugs during the study, pre- contribute to the glycemic improvement.11 vious GI surgery that could interfere with placement of the A nonsurgical method of altering the location of the initial device, or family or patient history of a known diagnosis or contact of partially digested nutrients with the intestine is by symptoms of systemic lupus erythematosus, scleroderma, or use of the duodenal–jejunal bypass liner (DJBL).12,13 The DJBL other autoimmune connective tissue disorder.
consists of a 60-cm, impermeable fluoropolymer liner and anitinol anchor, which reversibly fixes the device to the wall of the duodenum. The DJBL is open at both ends and functions The study was a 52-week, prospective, open-label, single- to prevent the partially digested food from contacting the center clinical trial intended to assess the safety and efficacy of proximal intestine. A pilot study of the DJBL in obese subjects the DJBL in obese subjects with type 2 diabetes. Following with type 2 diabetes resulted in a mean reduction of 50 mg/dL screening and up to 30 days prior to implanting the DJBL, in fasting plasma glucose (FPG) after 1 week that was sus- baseline demographics and medical history were obtained.
tained through 24 weeks.14 Here we describe a 52-week study Excess body weight was calculated as the amount of body that was designed to investigate the effect of the DJBL on weight that exceeded a BMI of 25 kg/m2. In addition, a metabolic parameters in obese subjects with type 2 diabetes.
physical examination, electrocardiogram, chest x-ray, liver,biliary duct, and pancreas ultrasound, upper GI endoscopy, and fasting blood tests were conducted. A schedule of study assessments and activities is presented in Figure 1. In accor-dance with the protocol, all subjects received nutritional This study was conducted in compliance with international counseling at the baseline visit and at all monthly follow-up guidelines and local country regulations. The study protocol visits. Baseline counseling consisted of a 30-min visit with a and subject informed consent forms were reviewed and ap- study nurse to discuss nutritional, behavioral, and lifestyle proved by the Ethics and Research Committee of the Uni- modifications recommended throughout the study. At follow- versity of Sa˜o Paulo, Sa˜o Paulo, Brazil. All subjects or their up visits, subjects were only asked if they were following legal guardians provided written informed consent prior to their diet, and results were recorded. Therefore, nutritional counseling was minimal at the post-implant follow-up visits.
During the period starting 3 days prior to device implantation and ending 2 weeks after explantation of the device, subjects Male and female subjects with type 2 diabetes between the were instructed to take an over-the-counter proton pump in- ages of 18 and 65 years who had a BMI of ‡ 40 and < 60 kg/m2 hibitor (40 mg twice a day). Daily multivitamin and iron were eligible for inclusion in the study. Exclusion criteria in- supplements were also recommended throughout the study.
cluded use of prescription anticoagulation therapy, iron de- Subjects were instructed to follow a liquid diet during the first ficiency or iron deficiency anemia, inflammatory bowel 2 weeks following implantation and to gradually transition to Schedule of study assessments. BL, baseline; BMI, body mass index; GI, gastrointestinal.
DUODENAL–JEJUNAL BYPASS LINER AND DIABETES a normal diet over the next 10 days. Follow-up examinations were to be conducted 1, 3, and 6 months after the explantation All device implantations were performed with the patient under general anesthesia using a minimally invasive endo- scopic procedure.13 At the end of the study period (or earlier if indicated by an adverse event), the DJBL was removed en- doscopically using a custom grasper passed through the working channel of a standard endoscope with the patient under conscious sedation for 21 of 22 subjects; one subject The DJBL (EndoBarrier Gastrointestinal LinerÒ) and the implantation and removal devices were manufactured by GI Several subject populations were included in the analyses.
The safety analysis population included all subjects in whom an implant procedure was attempted. The full efficacy anal- ysis population included all enrolled subjects in whom an implantation was successful. The completer population in- cluded all subjects who completed the 52-week implantation The study had two primary efficacy end points: percentage change in excess body weight and change in type 2 diabetes status, including hemoglobin A1c (HbA1c) and fasting blood glucose and insulin levels. Secondary efficacy end points in- cluded changes in blood lipid levels and blood pressure.
All statistical analyses were performed using SASÒ version 9.2 or later (SAS Institute, Cary, NC). Efficacy variables were evaluated as a change from baseline. Missing values were not imputed. However, last observation carried forward (LOCF) on or before explantation was also used when calculating change from baseline. A P value (calculated using Student’s t test) of £ 0.05 was considered statistically significant.
Twenty-two subjects were screened, and all were enrolled in the study. The baseline characteristics of the study subjects BMI, body mass index; GERD, gastroesophageal reflux disease; are presented in Table 1. At baseline, 17 of 22 subjects were HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low- receiving drugs to control their diabetes, including 11 re- ceiving metformin alone, two receiving glibenclamide alone,two receiving chlorpropamide alone, one receiving metfor-min plus glibenclamide, and one receiving metformin plus weeks after implantation. Subsequent computed tomography acarbose. A DJBL was successfully implanted in all subjects, scan revealed a 30-cm abdominal tumor that was identified as with the procedure taking an average of 26.7 – 11.0 min metastatic ovarian cancer. This event was considered not re- Thirteen subjects completed the 52-week period, and 18 Baseline FPG, fasting plasma insulin, and HbA1c levels are subjects completed at least 24 weeks. The mean duration of presented in Table 1, and changes in these parameters are the implant period for all subjects was 41.9 – 3.2 weeks. The shown in Table 2. Reductions in FPG were seen as early as reasons for early removal of the device were migration or week 1 and reached statistical significance at week 24. At rotation of the device (n = 3; 36, 36, and 48 weeks post- week 52, the completer population demonstrated a decrease implantation), GI bleeding (n = 1; 4 weeks post-implantation), in plasma glucose of - 37.1 – 11.8 mg/dL (mean – SEM, abdominal pain (n = 2; 21 and 30 weeks post-implantation), P < 0.01). A similar reduction was seen in the full analysis and principal investigator request due to subject’s noncom- population using LOCF analysis (Table 2). Mean HbA1c was pliance with study visits (n = 2; 20 and 32 weeks post- statistically significantly decreased by week 24 and remained implantation). The device was removed from one subject who statistically significantly decreased thereafter (Table 2). The presented with fever and mild abdominal symptoms 17 decrease in HbA1c was evident in subgroups of subjects with Data are mean – SEM values (number of patients).
aP < 0.01, bP < 0.001, cP < 0.0001, dP < 0.05 for change from baseline.
HbA1c, hemoglobin A1c; LOCF, last observation carried forward.
baseline HbA1c < 8%, 8% to < 10%, and ‡ 10% (Fig. 2). The Blood levels of total cholesterol, low-density lipoprotein progressive improvement in HbA1c is illustrated in Figure 3.
cholesterol, and triglycerides were significantly reduced At baseline, only one subject (4.5%) had an HbA1c level below during the study (Table 3). Nonsignificant decreases in mean 7%. For final HbA1c measurement, 16 of 22 subjects (73%) had systolic (6.6 – 4.4 mm Hg, P = 0.15, LOCF) and diastolic an HbA1c below 7%. In the nine subjects who experienced an ( - 1.6 – 3.5 mm Hg, P = 0.65, LOCF) blood pressure were ob- early device explantation, HbA1c decreased from 8.9 – 0.5% at baseline to 7.2 – 0.4% at final measurement (median, 28 Sixteen subjects had HbA1c measured 3 and/or 6 months after explantation of the DJBL. These subjects demonstrated a During the study, nine subjects experienced no change in mean decrease in HbA1c during the original 52-week study of their diabetes medications, three subjects reduced their dose - 2.3 – 0.4%. Three and 6 months after removal of the device, of metformin, and one subject stopped other diabetes medi- their mean changes from baseline were - 2.3 – 0.3% (n = 15) cation. One subject who was previously untreated began and - 1.7 – 0.7% (n = 11), respectively.
treatment with metformin, five subjects increased their dose All 22 subjects reported at least one treatment-emergent of metformin, and three subjects added an additional drug to adverse event. Treatment-emergent adverse events that oc- curred in ‡ 10% of the subjects are presented in Table 4. The Figure 4 shows the percentage change in excess body most common adverse events that were deemed to be possi- weight observed during the study. The 13 subjects who bly or probably related to the device or procedure were GI completed the study weighed 121.8 – 7.6 kg (mean – SE) at disorders, including upper abdominal pain, nausea, and vo- baseline and 101.6 – 5.7 kg at week 52. This change re- miting. With one exception, all adverse events were mild or presented a mean loss of excess body weight (percentage ex- moderate in severity. The one exception was the 43.5-year-old cess weight loss) of 39.0 – 3.9% (P < 0.0001). In the full analysis white woman in whom metastatic ovarian cancer was dis- population using LOCF, the mean percentage excess weight covered 17 weeks after device implantation (see above). No loss was 35.5 – 3.1% (P < 0.0001). The reduction in excess body weight (LOCF) was reflected by reductions in BMI and waistcircumference of - 6.7 – 0.7 kg/m2 and - 13.0 – 1.7 cm, re- The present study represents the longest evaluation of the safety and efficacy of the DJBL in obese patients with type 2diabetes. The mean duration of the DJBL implant was 41.9weeks; 18 subjects completed at least 24 weeks, and 13 Change in hemoglobin A1c (HbA1c) in subgroups of subjects. No statistical analysis was performed on thesedata. SE bars are hidden within symbols at some time points.
Distribution of hemoglobin A1c (HbA1c) levels BL, baseline; LOCF, last observation carried forward.
during the study. LOCF, last observation carried forward.
DUODENAL–JEJUNAL BYPASS LINER AND DIABETES Other clinical studies provide anecdotal observations of the effect of the DJBL on type 2 diabetes. In the initial 12-weekstudy of the DJBL, four of the 12 obese subjects had type 2diabetes;15 the authors reported that ‘‘all 4 diabetic patientshad normal FPG levels without hypoglycemic medication forthe entire 12 weeks. Of those 4 patients, 3 had decreased he-moglobin A1c of 0.5% by week 12.’’ A second study enrolled40 obese subjects who were candidates for Roux-en-Y gastricbypass and randomly assigned them to either DJBL (n = 26, 25implanted) or low-calorie diet alone (n = 14) with the objectiveof weight loss before surgery.16 Three subjects with type 2diabetes were included in the DJBL group, and one was in-cluded in the diet group. After 12 weeks, two of the three DJBLsubjects demonstrated decreases in HbA1c, and the third,who had a baseline HbA1c of 5.5%, was able to discontinue Percentage excess weight loss by visit. The number of subjects represented by each point is shown above the x- diabetes medications within 1 week of implantation. A similar axis. The dotted line reflects the loss of excess body weight in 12-week study randomly assigned obese subjects to either the completer population during the study.
DJBL (n = 30) or low-calorie diet (n = 11).13 Eight subjects withtype 2 diabetes were included in the DJBL group and dem-onstrated a reduction in mean HbA1c from 8.8 – 1.7% at subjects completed the entire 52-week study. During the 52- baseline to 7.7 – 1.8% after 12 weeks (P reported as nonsig- week study, substantial improvements in diabetes status were nificant). Six of the eight subjects were able to reduce their seen as shown by the decreased blood glucose and HbA1c dosage of oral diabetes medication or insulin after 1 week, and levels in the study population (Table 2 and Figs. 2 and 3). The at 12 weeks, one subject had completely stopped taking an- improvement in glycemic control is convincingly demon- strated by the results presented in Figure 3, which shows that In the present study, device-related early removal of the the percentage of subjects with HbA1c < 7% at baseline im- DJBL occurred in six of the 22 (27%) subjects after a median of proved from 4.5% to 73.0% at final study assessment. It is 31 weeks. There were three non–device-related early remov- important that the improvements in HbA1c were seen re- als: one after 17 weeks due to a metastatic ovarian tumor and gardless of baseline value (Fig. 2).
two (weeks 20 and 32, respectively) due to principal investi- A single previous study of the DJBL for the treatment of gator request as subjects were noncompliant with study visits.
type 2 diabetes has been reported.14 In that pilot study, 12 In previous studies, device-related early explantation oc- obese patients with type 2 diabetes received a DJBL, and a curred in 15–33% of subjects after a median of 21 days (range, comparable control group of six patients underwent a sham 3–120 days, n = 18).12–16 Although the early explantation rate procedure. After 1 week, FPG had declined significantly in the of the present study was high, these subjects still derived a DJBL group ( - 50 – 18 mg/dL, mean – SE) compared with the metabolic benefit from the DJBL; at their last measurement, sham group ( + 25 – 29 mg/dL, P = 0.042). In addition, post- HbA1c was 7.2% compared with a baseline value of 8.9%.
prandial glucose excursions were improved in the DJBL Based on the safety results of previous clinical studies of the group but not in the sham group. Of interest is the observation DJBL, no unanticipated adverse events were observed. The that at the 1-week time point, the improvement in glucose primary device-related safety issues seen in the present study metabolism seen in the DJBL could not be explained by were GI pain, nausea, vomiting, and back pain (Table 4). All changes in body weight, as both groups had experienced the cases were mild to moderate in severity. Two subjects cited same median weight change from baseline (about - 3.5 kg).
abdominal pain as the reason for requesting early explanta- The improvements in glucose metabolism were sustained tion of the device. It is important that the rates of GI or ab- through the end of the study (mean, 200 – 22 days [28.6 dominal pain adverse events appeared to be similar to those weeks]). At week 24, FPG had declined by 83 – 39 mg/dL reported in previous studies, suggesting that the longer im- (from a baseline of 199 – 71 mg/dL), and HbA1c had declined plant duration of the present study was not responsible for an by 2.4 – 0.7% (from a baseline of 9.2 – 1.7%).
increased incidence of these events.
Data are mean – SEM values (number of patients).
aP < 0.05, bP < 0.01 for change from baseline.
HDL, high-density lipoprotein; LDL, low-density lipoprotein; LOCF, last observation carried forward; TG, triglycerides.
Table 4. Treatment-Emergent Adverse Events durability of the response. Although these results are similar to results reported for bariatric surgery, the study provides noinformation regarding the relative efficacy and safety com- pared with the surgical procedures. Finally, the study results provide no information about the biologic mechanisms that contribute to the improvement in glycemic control.
The study was not randomized and did not include a control group. Therefore, it is possible that behavioral changes in the subjects may have contributed to some of the benefits observed. The pharmacologic treatment of type 2 diabetes was not specified or standardized in this study, and changes in antidiabetes drug regimens may have influenced the re- sults. Finally, the study was small, and device-related early explantation of the DJBL was required in 27% of patients.
The results of this study provide clear evidence that the DJBL can improve glycemic status and cardiometabolic fac- tors in obese subjects with type 2 diabetes and that it therefore may represent a nonsurgical and reversible alternative to bariatric surgery and an effective adjunct to pharmacotherapy GI Dynamics paid for professional medical writing and editorial assistance that was provided by Edward Weselcouch, Ph.D., of PharmaWrite (Princeton, NJ). The au- thors participated fully in the planning and drafting of this manuscript and are fully responsible for its content. GI Dynamics reviewed the final draft of this manuscript to en- sure the accuracy of the data reported from this company-sponsored clinical trial. Finally, the authors would like to thank Dr. Lee Kaplan for his discussions regarding the manuscript.
The mechanism responsible for the improvement in type 2 diabetes in this and previous studies is not well understood.
No competing financial interests exist.
However, the rapid improvement in plasma glucose and in-sulin levels suggests that the reduction in body weight was not solely responsible. This rapid metabolic response has beenobserved in many studies of gastric bypass surgery, sug- 1. Selassie M, Sinha AC: The epidemiology and aetiology of gesting that rearrangement of GI anatomy plays a role.9,10,17 obesity: a global challenge. Best Pract Res Clin Anaesthesiol Hickey et al.18 suggested that dysregulated neuroendocrine signaling between the proximal intestine and pancreas might 2. Nguyen NT, Nguyen XM, Lane J, Wang P: Relationship participate in the insulin resistance of type 2 diabetes and that between obesity and diabetes in a US adult population: bypassing a portion of the foregut may interrupt this abnor- findings from the National Health and Nutrition Examina-tion Survey, 1999–2006. Obes Surg 2011;21:351–355.
mality. Incretins, like glucagon-like peptide-1 and glucose- 3. Rato Q: Diabetes mellitus: a global health problem. Rev Port dependent insulinotropic peptide, have been implicated in this response,19 as has a yet-to-be-identified counter- 4. Shaw JE, Sicree RA, Zimmet PZ: Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin The results of this study leave several questions un- answered. The first is how long does the antidiabetes response 5. Schmidt MI, Duncan BB, Hoffmann JF, Moura L, Malta DC, last after removal of the device? The results from 11 subjects Carvalho RM: Prevalence of diabetes and hypertension indicated that the HbA1c response continued for up to 6 based on self-reported morbidity survey, Brazil, 2006. Rev months after device removal. It is tempting to speculate that Saude Publica 2009;43(Suppl 2):74–82.
the stability of the response is an indication of a ‘‘resetting’’ of 6. Malerbi DA, Franco LJ: Multicenter study of the prevalence glucose homeostasis, but no data from this study support this of diabetes mellitus and impaired glucose tolerance in the concept. Clearly, the weight loss may have contributed to the urban Brazilian population aged 30–69 yr. The Brazilian DUODENAL–JEJUNAL BYPASS LINER AND DIABETES Cooperative Group on the Study of Diabetes Prevalence.
bypass liner for the treatment of type 2 diabetes. Diabetes 7. Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR, 15. Rodriguez-Grunert L, Galvao Neto MP, Alamo M, Ramos Sherwin R, Zinman B: Medical management of hyperglycemia AC, Baez PB, Tarnoff M: First human experience with en- in type 2 diabetes: a consensus algorithm for the initiation and doscopically delivered and retrieved duodenal-jejunal by- adjustment of therapy: a consensus statement of the American pass sleeve. Surg Obes Relat Dis 2008;4:55–59.
Diabetes Association and the European Association for the 16. Tarnoff M, Rodriguez L, Escalona A, Ramos A, Neto M, Study of Diabetes. Diabetes Care 2009;32:193–203.
Alamo M, Reyes E, Pimentel F, Ibanez L: Open label, pro- 8. Turner RC, Cull CA, Frighi V, Holman RR: Glycemic control spective, randomized controlled trial of an endoscopic with diet, sulfonylurea, metformin, or insulin in patients duodenal-jejunal bypass sleeve versus low calorie diet for with type 2 diabetes mellitus: progressive requirement for pre-operative weight loss in bariatric surgery. Surg Endosc multiple therapies (UKPDS 49). UK Prospective Diabetes Study (UKPDS) Group. JAMA 1999;281:2005–2012.
17. Hussain A, Mahmood H, El-Hasani S: Can Roux-en-Y gas- 9. Greenway SE, Greenway FL 3rd, Klein S: Effects of obesity tric bypass provide a lifelong solution for diabetes mellitus? surgery on non-insulin-dependent diabetes mellitus. Arch 18. Hickey MS, Pories WJ, MacDonald KG Jr, Cory KA, Dohm 10. Rubino F, Gagner M: Potential of surgery for curing type 2 GL, Swanson MS, Israel RG, Barakat HA, Considine RV, diabetes mellitus. Ann Surg 2002;236:554–559.
Caro JF, Houmard JA: A new paradigm for type 2 diabetes 11. Rubino F: Is type 2 diabetes an operable intestinal disease? A mellitus: could it be a disease of the foregut? Ann Surg provocative yet reasonable hypothesis. Diabetes Care 1998;227:637–643; discussion 643–634.
19. Bose M, Olivan B, Teixeira J, Pi-Sunyer FX, Laferrere B: Do 12. Gersin KS, Rothstein RI, Rosenthal RJ, Stefanidis D, Deal SE, incretins play a role in the remission of type 2 diabetes after Kuwada TS, Laycock W, Adrales G, Vassiliou M, Szomstein gastric bypass surgery: what are the evidence? Obes Surg S, Heller S, Joyce AM, Heiss F, Nepomnayshy D: Open-label, sham-controlled trial of an endoscopic duodenojejunal by-pass liner for preoperative weight loss in bariatric surgerycandidates. Gastrointest Endosc 2010;71:976–982.
13. Schouten R, Rijs CS, Bouvy ND, Hameeteman W, Koek GH, Janssen IM, Greve JW: A multicenter, randomized efficacy study of the EndoBarrier Gastrointestinal Liner for presurgical weight loss prior to bariatric surgery. Ann Surg 2010;251:236– 14. Rodriguez L, Reyes E, Fagalde P, Oltra MS, Saba J, Aylwin CG, Prieto C, Ramos A, Galvao M, Gersin KS, Sorli C: Pilotclinical study of an endoscopic, removable duodenal-jejunal

Source: http://www.professorsukkar.it/home/1_Year_T2D_DTT.pdf