The Interventional Procedures Advisory Committee has considered this procedure and NICE has now issued a consultation document about its safety and efficacy. This has been issued for 4 weeks’ public consultation (until 1700pm on 27 October 2014).

If you wish to comment on the consultation document please use the link at the bottom of this page.

Through its guidance NICE is committed to promoting race and disability equality, equality between men and women, and to eliminating all forms of discrimination. One of the ways we do this is by trying to involve as wide a range of people and interest groups as possible in the development of our interventional procedures guidance. In particular, we aim to encourage people and organisations from groups who might not normally comment on our guidance to do so.

In order to help us promote equality through our guidance, we should be grateful if you would consider the following question:

Are there any issues that require special attention in light of NICE’s duties to have due regard to the need to eliminate unlawful discrimination and promote equality and foster good relations between people with a characteristic protected by the equalities legislation and others?

Please note that NICE reserves the right to summarise and edit comments received during consultations or not to publish them at all where in the reasonable opinion of NICE, the comments are voluminous, publication would be unlawful or publication would otherwise be inappropriate.If you wish to be alerted to developments regarding this procedure, including the release of the consultation document, please use the register an interest button to the left of this page

 

In this procedure, a plastic sleeve or liner is inserted through the mouth into the bowel, with the help of an endoscope (a thin flexible tube with a camera on the end), so that it lines the upper part of the bowel (duodenum and jejunum). The liner is intended to reduce food absorption by forming a barrier between food and the bowel. The aim is to help people lose weight and improve control of their diabetes. The sleeve is usually removed after a year.

Introduction

The National Institute for Health and Care Excellence (NICE) has prepared this interventional procedure (IP) overview to help members of the Interventional Procedures Advisory Committee (IPAC) make recommendations about the safety and efficacy of an interventional procedure. It is based on a rapid review of the medical literature and specialist opinion. It should not be regarded as a definitive assessment of the procedure.

Date prepared

This IP overview was prepared in April 2014 (25-04-2014).

Procedure name

• Implantation of a duodenal–jejunal bypass sleeve for managing type 2 diabetes

Specialist societies

• Association of British Clinical Diabetologists (ABCD)
• British Obesity and Metabolic Surgery Society
• British Society of Gastroenterology
• Association of Upper Gastrointestinal Surgeons of Great Britain and Ireland.

Description

Indications and current treatment

Type 2 diabetes is caused by insulin resistance with or without inadequate pancreatic insulin secretion. It is most commonly seen in people who are overweight or obese. Presenting symptoms include polyuria, polydipsia, and fatigue. Type 2 diabetes is commonly associated with raised blood pressure, abnormal blood lipid levels and a tendency to atherosclerosis. This combination is often described as the ‘metabolic syndrome’, which is associated with fatty liver and abdominal adiposity (increased waist circumference).

Type 2 diabetes is managed by lifestyle and dietary changes, exercise or oral hypoglycemic medications. If blood glucose levels remain poorly controlled, subcutaneous insulin injections may be needed.

What the procedure involves

Endoscopic implantation of a duodenal–jejunal bypass sleeve (DJBS) is a procedure that aims to improve glycaemic control in people with obesity or who are overweight.

The procedure is done with the patient under general anaesthesia or sedation, using image guidance. The sleeve is positioned endoscopically (via the mouth). Using a delivery catheter, a capsule containing a single-use impermeable DJBS is positioned in the duodenal bulb just distal to the pylorus. It is secured there using an integral spring metal anchor. The sleeve is advanced distally into the jejunum with the aid of a tension wire that is part of the ‘introducer’ device. It extends about 60 cm down the small intestine. The exact mode of action of the DJBS is unclear but the sleeve may reduce the absorption of food by forming a barrier between food and the intestinal wall, so delaying the mixing of digestive enzymes with food.

After the procedure, patients are recommended a diet that typically involves progression from fluids to semi-solid foods and then to solid foods.

After a maximum of a year, the sleeve is removed with the patient under sedation, using image guidance and endoscopy. The anchor has a drawstring mechanism such that it can be collapsed and partly withdrawn into a plastic hood fitted to the endoscope before withdrawal.

Outcome assessment tools

HOMA score and Matsuda Index

The homeostatic model assessment of insulin resistance (HOMA‑IR) and the Matsuda index (10,000/√ [(G₀xI₀) * (G_mean x I_mean)], where G is glucose and I is insulin, are used to estimate insulin resistance and sensitivity.

Insulinogenic index

The insulinogenic index [(Ins₃₀−Ins₀)/(gluc₃₀−gluc₀) or ∆Ins₃₀/∆Glu₃₀], where Ins₃₀ is insulin in 30 minutes and Gluc₃₀ is glucose in 30 minutes, was used to estimate the first phase of insulin release.

Literature review

Rapid review of literature

The medical literature was searched to identify studies and reviews relevant to Implantation of a duodenal–jejunal bypass sleeve for managing type 2 diabetes. Searches were conducted of the following databases, covering the period from their commencement to 25-04-2014: MEDLINE, PREMEDLINE, EMBASE, Cochrane Library and other databases. Trial registries and the Internet were also searched. No language restriction was applied to the searches (see appendix C for details of search strategy). Relevant published studies identified during consultation or resolution that are published after this date may also be considered for inclusion.

The following selection criteria (table 1) were applied to the abstracts identified by the literature search. Where selection criteria could not be determined from the abstracts the full paper was retrieved.

Table 1 Inclusion criteria for identification of relevant studies

Characteristic	Criteria
Publication type	Clinical studies were included. Emphasis was placed on identifying good quality studies. Abstracts were excluded where no clinical outcomes were reported, or where the paper was a review, editorial or a laboratory or animal study. Conference abstracts were also excluded because of the difficulty of appraising study methodology, unless they reported specific adverse events that were not available in the published literature.
Patient	Patients with type 2 diabetes.
Intervention/test	Implantation of a duodenal–jejunal bypass sleeve
Outcome	Articles were retrieved if the abstract contained information relevant to the safety and/or efficacy.
Language	Non-English-language articles were excluded unless they were thought to add substantively to the English-language evidence base.

 

List of studies included in the IP overview

This IP overview is based on 186 patients from 1 randomised trial¹, 6 case series^2-7 and 1 case report⁸.

Other studies that were considered to be relevant to the procedure but were not included in the main extraction table (table 2) have been listed in appendix A.

 

Table 2 Summary of key efficacy and safety findings on implantation of a duodenal–jejunal bypass sleeve for managing type 2 diabetes

Details

Study type	RCT
Country	Netherlands (multicentre)
Recruitment period	Not reported
Study population and number	Patients with type 2 diabetes and obesity n=77 (38 DJBS group versus 39 control group – low calorie diet)
Age and sex	DJBS arm mean 49.5 years, control arm 49 years DJBS arm 38% (13/34) female, sham arm 36% (14 /39) female Mean BMI: DJBS arm 34.6 kg/m2, sham arm 36.8 kg/m2; Mean HbA1c: 8.3%
Patient selection criteria	Inclusion criteria: aged 18–65 years, BMI 30–50 kg/m2, type 2 diabetes for less than10 years with an HbA1c level 7.5–10%. Exclusion criteria: weight loss of > 4.5 kg within 12 weeks before screening, pregnancy or intention to become pregnant, use of NSAIDs, anticoagulation therapy, corticosteroids, weight loss medication, or drugs known to affect GI motility, substance abuse, active Helicobacter pylori infection, probable insulin production failure (C‑peptide level of < 10 ng/ml), iron deficiency, GI abnormalities, or previous surgery in GI, symptomatic gall stones or kidney stones, infection, bleeding disorders, GORD, connective tissue disorders, severe liver or kidney failure.
Technique	DJBS arm: EndoBarrier implanted under general anaesthesia, fluoroscopy and endoscopy guidance. Dose of glucose-lowering medication (except metformin) was reduced by 50% and PPIs given. Liquid diet for the first week, pureed food during the second week and solids thereafter. Recommended intake 1200 calories per day for women and 1500 calories per day for men. Patients advised to increase physical activity. Follow-up every month, after 6 months DJBL was removed and patients were followed up for an additional 6 months. Low calorie diet alone: patients received only dietary intervention.
Follow-up	12 months (including 6‑month follow-up after device removal)
Conflict of interest/source of funding	Study funded by manufacturer. 4 authors are consultants for GI Dynamics.

Analysis

Follow-up issues: In the device arm (n=34) 2 patients were lost to follow-up (at 191 and 272 days) and 1 withdrew consent (at day 10 due to abdominal pain); in the diet arm (n=39) 1 patient was lost to follow-up (at day 267) and 4 withdrew consent (at week 1, 3 months, day 273 and 315).

Study design issues: The method of randomisation was not reported. There was no allocation concealment. There was no significant difference between groups at baseline with respect to age, sex, BMI and comorbidities.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 38 DJBS versus 39 low-calorie diet alone Procedural outcomes in DJBS arm % (n) Implantation success 89.4 (34/38) Implantation failure  7.8 (3/38) Withdrawal 2.6 (1/38)   Change in body weight (mean) Mean body weight % DJBS arm (n=38) Diet arm (n=39) p value Baseline 105.4 110.8 0.29 6 months 94.8 105.5 <0.05 12 months 98.6 106.8 0.07 Excess weight loss (EWL) was more in the DJBS group than in control group at 12‑month follow-up (19.8% versus 11.7%, p<0.05). Change in glycaemic control measured by HbA1c(mean) Mean HbA1c % DJBS arm (n=38) Diet arm (n=39) p value Baseline 8.3 8.3 0.82 6 months 7.0 7.9 <0.05 12 months 7.3 8.0 0.95 Change in fasting glucose (mean) Mean fasting glucose mmol/litre DJBS arm (n=38) Diet arm (n=39) p value Baseline 11.0 11.0 0.87 6 months 8.5 10.0 0.10 12 months 9.0 9.7 0.41   Change in fasting insulin (mean) Mean fasting insulin mU/litre DJBS arm (n=38) Diet arm (n=39) p value Baseline 15.0 17.0 0.11 6 months 11.1 14.0 0.40 12 months 15.0 15.7 0.73 85% of patients in DJBS arm showed decreased postprandial glucose excursions versus 48.7% of control patients (p<0.05). Changes in oral glucose lowering medication At 12‑month follow-up, the daily insulin dos and use of sulfonylureas decreased or discontinued in the DJBS group  than in control group  (p<0.05). Change in cardiovascular parameters The RCT of 77 patients reported that at 12‑month follow-up, blood pressure decreased from 147/92 mmHg to 130/82 mmHg in the DJBS group and from 152/90 mmHg to 140/85 mmHg in the control group (p=0.31 for systolic pressure and p=0.38 for diastolic pressure). The total cholesterol levels were comparable with baseline, 4.4 versus 4.4 mmol/litre, DJBS versus control group, respectively (p=0.79).	Adverse events DJBS arm % (n) Diet arm % (n) At least 1 adverse event 76.5 (29/38) 59 (23/39) Minor GI symptoms, abdominal pain or discomfort 63.2 (25/38) 28.2 (11/39) Nausea or vomiting 23.7 (9/38) 17.9 (7/39) Mild to moderate hypoglycaemia 23.7 (9/38) 25.6 (10/39) Adverse events requiring hospitalisation 8 (5/8 device related: 1 melena and pain in epigastric area; 1 abdominal pain and dehydration both managed conservatively; 1 device blocked with food, removed device early; 1 symptomatic gallstones treated with cholecystectomy 1 procedure related: oesophageal perforation during device removal at 6 months(caused by one of the barbs on the anchor not being covered by the removal hood), treated by endoscopic stenting and feeding tube, after 3 weeks resolved without sequelae) 8 (5 resolved without sequelae; other needed treatment)
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; GI, gastrointestinal; HbA1c, glycated haemoglobin; PPI, proton pump inhibitors; NSAID, non-steroidal anti-inflammatory drugs; RCT, randomised controlled trial

Study 2 Rodriguez L (2009)

Details

Study type	RCT
Country	Chile (single centre)
Recruitment period	2007–8
Study population and number	Patients with type 2 diabetes and obesity n=18 (12 DJBS versus 6 sham endoscopy)
Age and sex	DJBS arm 45 years, sham arm 51 years DJBS arm 67% female, sham arm 50% female Mean BMI: DJBS arm 38.9 kg/m2, sham arm 39.0 kg/m2 Mean HbA1c: 9.1%
Patient selection criteria	Aged 18–55 years with type 2 diabetes for more than 10 years and an HbA1c 7–10%, fasting plasma glucose under 240 mg/dl and BMI 30–50 kg/m2.
Technique	DJBS (EndoBarrier) procedures used fluoroscopy and endoscopy. Endoscopy 3 days and 4 weeks after explantation. Sham procedure: upper gastrointestinal endoscopy. Liquid diet for the first week, pureed food during the second week and solids thereafter. Recommended intake 1200 calories per day for women and 1500 calories per day for men.
Follow-up	24 weeks
Conflict of interest/source of funding	Study funded by manufacturer. Authors are consultants/ shareholder for GI Dynamics.

Analysis

Follow-up issues: 42% (5/12) of patients in the device arm (with explanted devices) and 24% (2/6) of patients in the sham ITT arm were lost to follow‑up at 12 weeks.

Study design issues: The method of randomisation was not reported. There was no allocation concealment.

There was no significant difference between the groups at baseline.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 12 DJBS versus 6 sham endoscopy Change in glycaemic control measured by HbA1c (ITT population) (mean±SD) Mean HbA1c % DJBS arm (n=12) Sham arm (n=6) p value Baseline 9.2 9.0 NS 12 weeks −1.3±0.9 −0.8±0.3 NS 24 weeks −2.4±0.7 −0.8±0.4 NS HbA1c change in population completing treatment is more than 0.05 at all time points between both arms. Change in FPG concentration (ITT population) (mean±SD) Mean FPG mg/dl DJBS arm (n=12) Sham arm (n=6) p value Baseline 193±24 140±38 <0.05 Week 1 −50±18 +25±29 0.042 12 weeks −45±26 −8±35 NS 24 weeks −83±39 +16±42 NS Both arms had equivalent baseline FPG concentrations. Oral antidiabetic medication use   Follow-up DJBS arm % Sham arm % Ceased drug use (ITT group)* Week 12 42 17 Ceased drug use (group who completed treatment)** Week 12 50 25 Ceased drug use (remaining patients)*** Week 24 40 25 *All treated patients. **All patients who completed at least 24 weeks. ***Patients remaining on the study. Postprandial 7-point blood glucose profile Mean postprandial plasma glucose AUC* DJBS arm (n=12) Sham arm (n=6) p value  Baseline mg/dl 31,226± 11,570 27,558±11,480 NS  Week 1 22% decrease 16% increase 0.016 * There was no change in postprandial insulin concentrations in either arm. Weight loss: At 12 weeks mean weight loss was comparable (p>0.05) for both treatment arms (both ITT and completer groups).	Explants during 12 weeks’ follow‑up % (n) Anchor migration (1 turned or migrated) 3 had symptoms: moderate pain (n=1), nausea and moderate vomiting (n=1) and mild abdominal pain and vomiting (n=1) 2 had no symptoms: noted at removal (n=1) and at scheduled endoscopy (n=1). 42 (5/12)   Adverse events (total 64) Adverse events DJBS % (n=episodes) Upper abdominal pain (in 12 patients) 30.8 (20) Vomiting (in 4 patients) 10.8 (7) Abdominal pain 4.6 (3) Nausea 7.7 (5) Symptoms of hypoglycaemia (but blood glucose more than 100 mg/dl in all cases) 7.7 (5) Decreased blood iron 6.2 (4) Flatulence 4.6 (3) Procedural vomiting 4.6 (3) Increased blood cholesterol 3.1 (2) Erosive duodenitis 1.5 (1) Constipation 1.5 (1) Diarrhoea 1.5 (1) Gastritis 1.5 (1) Headache 1.5 (1) Decreased HDL cholesterol 1.5 (1) Esophagitis 1.5 (1) Pain 1.5 (1) All events were mild or moderate.
Abbreviations used: AUC, area under the curve; BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; FPG, fasting plasma glucose; HbA1c, glycated haemoglobin; HDL, high density lipoprotein; ITT, intention to treat; NS, not significant; SD, standard deviation,.

Study 3 de Moura (2011)

Details

Study type	Case series
Country	Brazil
Recruitment period	Not reported
Study population and number	Morbidly obese and type 2 diabetes patients n=81
Age and sex	Mean 50.8 years; 4.4% female Mean BMI: 43.8 kg/m2
Patient selection criteria	Aged 18–65 years with a BMI over 35 kg/m2, type 2 diabetes with or without comorbidities, TG/HDL ratio ≥3.5
Technique	DJBS (EndoBarrier) procedures used fluoroscopy and endoscopy. PPI used in entire study. Liquid diet initially, solid diet in third week.
Follow-up	6 months
Conflict of interest/source of funding	2 authors independent consultants of GI Dynamics.

Analysis

Follow-up issues: 38/54 patients completed the study (26 completed 24 weeks, 12 completed 20 weeks).

Study design issues: 70% (54/77) of the patients had an initial TG/HDL ratio greater than or equal to 3.5 indicating insulin resistance and metabolic syndrome.

Study population issues: Comorbidities: 86% had hypertension, 36.7% had hyperlipidaemia.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 54 Procedural outcomes % (n) Implantation success 96 (78/81) Implantation failure (due to short duodenal bulb) 4 (3/81) Control of diabetes (HbA1c improvement) at 6 months All patients implanted with the device achieved statistically significant reductions in HbA1c (p<0.005). More than 70% of patients had HbA1clevels of less than 7%.   Improvement in insulin resistance and metabolic syndrome at 6 months   Number of patients * Initial average TG/HDL ratio Final average TG/HDL ratio p value Controlled TG/HDL 23 5.15 2.85 <0.001 Not controlled TG/HDL ratio 31 6.2 5.47 0.1641 Total 54 5.75 4.36 <0.001 *Patients presented with insulin resistance and metabolic syndrome. 42.6% of the patients presented a TG/HDL ratio lower than 3.5 at 6‑month follow-up. Weight loss Average weight loss of 12.6% of their initial weight. Relationship between TG/HDL ratio control and weight loss Comparing the patients who lost weight with the patients who controlled their TG/HDL ratio, an association can be observed between a weight loss greater than 10% of initial weight and control of TG/HDL ratio (p<0.01) with an odds ratio of 5.06.	Early explantations Total explants 16 Migration 9 Observation of a free device anchor during endoscopy 4 Bleeding without migration 1 Patient request 1 Investigator decision 1 12 devices were removed at 16 weeks, 2 at 12 weeks and 2 at 4 weeks.
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; HbA1c, glycated haemoglobin; HDL, high density lipoprotein; PPI, proton pump inhibitor; SD, standard deviation, TG/HDL ratio, triglyceride high density lipoprotein cholesterol ratio;

 

Study 4 de Moura (2012)

Details

Study type	Case series
Country	Brazil
Recruitment period	Not reported
Study population and number	Obese patients with type 2 diabetes n=22
Age and sex	Mean 46.2 years; 86.4% female Mean BMI: 44.8 kg/m2
Patient selection criteria	Patients with type 2 diabetes, between 18 and 65 years with a BMI over 40 kg/m2 and below 60 kg/m2.
Technique	DJBS (EndoBarrier) was implanted and explanted after 52 weeks using fluoroscopy and endoscopy. PPI were used until 2 weeks after explantation. Follow-up examinations were done at 1, 3 and 6 months after explantation. Patients were given 30 minutes’ nutritional counselling (on diet, lifestyle and behaviour) at baseline and monthly follow-up visits. Liquid diet for 2 weeks. Daily vitamin and iron supplements were recommended.
Follow-up	52 weeks
Conflict of interest/source of funding	Study sponsored by GI Dynamics (manufacturer).

Analysis

Follow-up issues: 82% (18/22) patients completed 24 weeks’ follow-up. Only 59% (13/22) patients completed 52 weeks’ follow-up.

Study design issues: The drug treatment for type 2 diabetes was not specified or standardised.

Study population issues: 77% (17/22) of patients took drugs for diabetes.

Other issues: 1 patient needed general anaesthesia for explantation. Authors suggest that changes in anti-diabetic drug treatment regimens may have influenced the results.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 22 Implantation success: 100% Changes in metabolic parameters and lipid levels (mean±SD values)   Base-line (n=22) 24 weeks (n=16) 52 weeks (n=13) LOCF* (n=22) Fasting glucose mg/dl 179.4±68.8 −33.4± 9.2 (p<0.01) −37.1±11.8 (p<0.01) −30.3± 10.2 (p<0.01) HbA1c % 8.9± 1.7 −1.5±0.4 (p<0.001) −2.3±0.3 (p<0.0001) −2.1±0.3 (p<0.0001) Fasting insulin U/ml 19.5± 14.7 −5.2±2.8 −10.1±4.2 (p<0.05) −7.3±2.6 (p<0.05) Total cholesterol mg/dl 201± 37 −16.7±6.9 (p<0.05) −28.1±5.6 (p<0.01) −19.7±5.9 (p<0.01) Triglycerides mg/dl 213± 89 −56.8±25 (p=0.05) −62.4±18.3 (p=0.01) −44.8± 17.4 (p<0.05) Diastolic blood pressure (mmHg) 79±10     −1.6±3.5 (p=0.65) *last observation carried forward in all patients analysed on or just before explantation  Improvement in glycaemic control At the end of the study 73% (16/22) of patients had an HbA1c under 7% compared with 4.5% (1/22) at baseline. Glycaemic control after device removal (at 6 months) HbA1c response continued for up to 6 months after device removal in 11 patients (mean change from baseline [8.9±1.7] was −1.7±0.7%). Weight loss (mean±SD values) Mean % EWL at 52 weeks (n=13) 39.0±3.9 (p<0.0001) Mean % of EWL (LOCF, n=22) 35.5±3.1 (p<0.0001) Decrease in mean BMI (kg/m2) (LOCF, n=22) −6.7±0.7 Mean reduction in waist circumference (cm) (LOCF, n=22) −13.0±1.7	Early device explantations % (n) Total explantations  40 (9/22) Device related (median 31 weeks) 27 (6/22) Device migration or rotation (48 weeks after implant) 14 (3/22) Gastrointestinal bleeding (4 weeks after implant) 4 (1/22) Abdominal pain (24 and 30 weeks after implant) 9 (2/22) Non-device related 1 unrelated malignancy (at 17 weeks due to metastatic ovarian cancer) 2 at investigator request (at weeks 20 and 32 due to ‘patients’ non-compliance with follow-up’) 14 (3/22)   Adverse events that occurred in more than 10% of patients Adverse event % (n) Device or procedure related (n) Gastrointestinal disorders 95 (21/22) 12 Upper abdominal pain 91 (20/22) 11 Nausea 50 (11/22) 7 Vomiting 63 (14/22) 7 Diarrhoea 13 (3/22) 1 Procedural and other complications     Procedural nausea 45 (10/22) 4 Procedural vomiting 32 (7/22) 3 Back pain 59 (13/22) 5 All events were mild or moderate, except 1 severe event caused by an unrelated malignancy.
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; EWL, excess weight loss; GA, general anaesthesia; HbA1c, glycated haemoglobin; HDL, high density lipoprotein; LCOF, last observation carried forward; PPI, proton pump inhibitors; SD, standard deviation.

 

Study 5 Cohen RV (2013)

Details

Study type	Case series
Country	Brazil (single centre)
Recruitment period	Not reported
Study population and number	Patients with lower BMI and type 2 diabetes n=23
Age and sex	Mean 49.8 years; 58.3% female Mean BMI: 30 kg/m2, type 2 diabetes duration: 6.6 years
Patient selection criteria	Aged 18 and 55 years with T2DM of <10 years, with oral glucose lowering medications, HbA1c 7.5 -10%, BMI 26–50 kg/m2
Technique	EndoBarrier deployed and removed under general anaesthesia. Nutritional counselling, proton pump inhibitors before implantation and 2 weeks after explantation. Liquid diet in first week and 1200–1500 calories intake thereafter.
Follow-up	52 weeks
Conflict of interest/source of funding	Study funded by GI Dynamics (manufacturer).

Analysis

Follow-up issues: 16 patients completed 1‑year treatment.

Study design issues: Patients with type 1 diabetes, insulin use, autoimmune disease, weight loss of >4.5 kg within 12 weeks, previous gastrointestinal surgeries, active Helicobacter pylori infection, on non-inflammatory drugs, weight loss medication, uncontrolled reflux disease were excluded.

Women either postmenopausal, sterile or on oral contraceptives were included.

Sulfonylurea dosage reduced to avoid hypoglycaemic events.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 23 Procedural outcomes % (n) Implantation success 87 (20/23) Implantation failure (due to unfavourable anatomy) 13 (3/23) Mean implantation duration 348 days Body weight glucose metabolism and plasma lipids during treatment with DJBS   Baseline (n=20) Week 12 (n=19) Week 52 (n=16) p value FPG (mg/dL) 207±61 132±41 155±52 0.012 HbA1c (%) 8.7±0.9 7.0±0.9 7.5±1.6 0.004 Total cholesterol (mg/dL) 221±50 167±38 188±32 NR Low density lipoprotein (mg/dL) 135±40 95±33 108±31 NR HDL (mg/dL) 42±11 39±7 40±10 NR Body weight (kg) 84.0±16.6 79.0±16.8 77.2±17.6 <0.0001 BMI (kg/m2) 30.0±3.6 28.3±3.7 28.5±3.3 <0.0001 62.5% (10/16) patients who completed the study had HbA1c levels <7% at week 52. 4/5 patients with HbA1c >9% at baseline did not show any reduction in HbA1c. Diabetic medications: 7 patients decreased and 4 increased the number of drugs or the doses of antidiabetic drugs. No significant correlation between change in body weight and change in FPG or HbA1c was observed (data not reported in paper).	% (n) At least 1 adverse event (mild or moderate) 96% (22/23) Most common device or procedure related adverse events Gastrointestinal disorders (including abdominal pain, nausea and vomiting) 13/23 Metabolic and nutritional disorders, including hypoglycaemia and iron deficiency 61 (14/23) Early device removals (1 patient at 10 weeks due to noncompliance with follow-up, 1 at 7 months due to recurring abdominal pain, in 2 due to device rotation and/or migration at 6 and 10 months) 80% (4/20)
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; EWL, excess weight loss; FPG, fasting plasma glucose; GA, general anaesthesia; HbA1c, glycated haemoglobin; HDL, high density lipoprotein; PPI, proton pump inhibitors; SD, standard deviation, T2DM, type 2 diabetes mellitus.

 

Study 6 Cohen RV (2013)

Details

Study type	Case series
Country	Brazil (single centre)
Recruitment period	Not reported
Study population and number	Patients with type 2 diabetes and BMI≤ 36kg/m2 n=16
Age and sex	Mean 49.8 years; 37.5% (6/16) female Mean BMI: 30 kg/m2, Mean HbA1c %: 8.6
Patient selection criteria	Patients with oral glucose lowering medications, no insulin,
Technique	EndoBarrier deployed and removed under GA after 52 weeks. Regular anti-diabetic medications were continued during implantation period. They were kept constant and additional medication given by physician only if baseline HbA1c levels were exceeded. Consistent diet, exercise and lifestyle advice throughout. Before each follow-up (1, 12, 24 and 52 weeks), all diabetic medications were stopped for 24 hours and patents fasted overnight and a standard 525 kcal meal given.
Follow-up	52 weeks after implantation and 26 weeks after explantation
Conflict of interest/source of funding	Study funded by GI Dynamics (manufacturer).

Analysis

Follow-up issues: All 16 patients completed 1‑year treatment.

Study design issues: Combined therapies (DJBS and regular anti-diabetic medical therapy) used in the study; results may be confounded by the inclusion of diabetic medications. Four patients had higher HbA_1c levels before implantation.

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 16 Body weight, glucose metabolism during and after treatment with DJBS (n=16)   Baseline Week 1 Week 12 Week 52 Week 78* p value (all) HbA1c (%)^ 8.6 - 6.9 7.5 7.8 <0.001 Mean fasting glucose concentration (mg/dl) 203.3±13.5 138.3±8.2 130.8±10.8 155.1±13.1 150.2±10.06 <0.001 Median HOMA-IR score (IQR) 6.6 (4.2–13.4) 3.1 (1.7–4.8) 3.1 (1.9–4.3) 3.0 (2.2–4.7) 4.4 (2.0–6.1) <0.001 Median Matsuda index score (IQR) 1.7 (1.2–2.3) 3.4 (2.2–6.1) 3.5 (1.9–4.9) 3.2 (2.2–4.6) 2.4 (1.8–4.4) <0.001 Mean fasting insulin (microU min/ml) 16.3 (2.3) 10.7 (1.6) 13.4 (2.5) 11.0 (1.9) 13.2 (1.6) 0.051 Mean insulin AUC (microU min/ml) 5757±606.7 4399±701.3 5280±825.6 5068±713.7 5018±642.8 0.28 Median (IQR) insulogenic index score I30/G30 0.36 (0.2-0.7) 0.6 (0.3-0.8) 0.36 (0.15-0.69) 0.41 (0.23-0.68) 0.53 (0.16-0.73) 0.43 Mean fasting insulin secretion after meal rate (pmol/min/m2) 128.4±12.8 114.1±14.9 120.4±14.4 128.7±16.8 124.8±15.9 0.51 Mean total insulin secretion rate after meal (nmol/min/m2) 23.0±1.7 22.7±2.5 25.7±2.9 24.7±.7 23.4±2.2 0.46 Mean C-peptide fasting (ng/ml) 3.8 ±0.4 3.1 ±0.3 3.2 ±0.4 3.1 ±0.4 3.2 ±0.4 0.024 Mean C-peptide AUC (ng/min/m) 743.3±50.3 689.1±69.5 773.2±75.0 763.3±0.5 744.8±65.9 0.46 Mean body weight (kg) 82.1±4.5 80.8±4.0 77.9±4.3 76.2±4.3 79.7±4.6 <0.001 Mean BMI (kg/m2) 30.0±0.9 29.4±0.8 28.7±0.9 28.1±0.9 29.4±0.9 <0.001 *26 weeks after DJBL removal. ^ 4 patients did not show an improvement during the 52 weeks.	No device related complications. Mild, transient abdominal pain (self-limiting) after first week of implantation: 19% (3/16)
Abbreviations used: AUC, area under the curve; BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; EWL, excess weight loss; GA, general anaesthesia; HbA1c, glycated haemoglobin; HOMA-IR, homeostatic model of assessment of insulin resistance; T2DM, type 2 diabetes mellitus.

 

Study 7 Munoz R (2014)

Details

Study type	Case series
Country	Chile (single centre)
Recruitment period	2009–2011
Study population and number	Morbidly obese patients n=79 (21 with type 2 diabetes)
Age and sex	Mean 35.4 years; 72% (44/79) female Mean BMI: 43±5.6 kg/m2
Patient selection criteria	Between 18 and 55 years with a BMI > 35 kg/m2 if presenting with comorbidities such as hypertension, diabetes, and/or dyslipidaemia; otherwise with a BMI 40–60 kg/m2.
Technique	DJBS (EndoBarrier) was implanted and explanted using fluoroscopy and endoscopy. Patients were advised at baseline to take a liquid and pureed diet for 2 weeks, followed by normal diet and moderate physical therapy for the rest of the study period. Proton pump inhibitors, multivitamins and iron supplements were used during the study period. Surveillance endoscopies were performed at 12, 24 and 36 weeks.
Follow-up	52 weeks
Conflict of interest/source of funding	Study was funded by GI Dynamics (manufacturer). Two authors disclosed a financial relationship with the manufacturer.

Analysis

Follow-up issues: Only 77% (61/79) patients completed 52 weeks follow-up

Other issues: Data on 39 patients were included in a previous publication

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 79 (21 with T2DM) Clinical factors associated with weight loss Univariate analysis shows that fasting glycaemia (r2=−0.303, p<0.013), insulin resistance determined by HOMA-IR (r2 =−0.457, p<0.019) and glycated haemoglobin HbA1c (r2=−0.471, p<0.001) were inversely associated with %EBWL at 52 weeks after DJBS implantation. Multivariate analysis indicates that only baseline HbA1c levels were associated inversely with %EBWL after 1 year of treatment (β adjusted coefficient −0.758, p<0.016). No differences at 1 year in %EBWL were observed between patients with or without T2DM (%EBWL T2DM 46.7±20% versus non T2DM 46.8 ±18.6%, p=0.988). Mean % EBWL At 3 months: 33±12; at 1 year: 46±18	No complications related to implant and explant procedures.   % (n) Early device removal* 26 (21/79) Device migration 8 Device obstruction 5 Abdominal pain 2 Liver abscess (medical treatment) 1 Upper gastrointestinal bleeding 1 Cholangitis 1 Ulcerative colitis 1 Acute cholecystitis 1 Patient request 1 *further information not reported in all cases.
Abbreviations used: AUC, area under the curve; BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; %EBWL, percentage of excess body weight loss; HbA1c, glycated haemoglobin; HOMA-IR, homeostatic model of assessment of insulin resistance; T2DM, type 2 diabetes.

 

Study 8 De Jonge (2013)

Details

Study type	Case series
Country	Brazil (single centre)
Recruitment period	Not reported
Study population and number	Obese patients with type 2 diabetes n=17
Age and sex	Mean 51 years; 18% (3/17) female Mean BMI: 37 kg/m2; HbA1c 8.4%
Patient selection criteria	Aged 18 and 55 years with type 2 diabetes of <10 years, with oral glucose lowering medications, HbA1c 7.5 -10%, BMI 26–50 kg/m2
Technique	DJBS (EndoBarrier) deployed and removed under GA after 24 weeks. Nutritional counselling and proton pump inhibitors given before implantation and 2 weeks after explantation. Liquid diet in first week and 1200–1500 calories intake thereafter. Glucose, insulin, GLP-1, GIP and glucagon responses after a standard meal were studied before, during and 1 week after DJBS treatment.
Follow-up	24 weeks
Conflict of interest/source of funding	Study funded by GI Dynamics (manufacturer).

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 17 Changes in diabetic parameters and gut hormones   Baseline 1 week 24 weeks HbA1c (%) 8.4 ± 0.2   7.0 ± 0.2 (P<0.01) Fasting glucose mmol/litre 11.6±0.5 9.0±0.5 (p<0.01) 8.6±0.5 (p<0.01) Postprandial glucose AUC mmol/litre/min 1999±85 1536±51 (p<0.01) 1588±72 (p<0.01) Fasting Insulin µU/ml 25.5±7.8 22.5±7.8 (p=0.23) 15.1±3.1 (p=0.06) Insulin AUC µU/ml/min 6.603±1100 6688±1164 (p=0.86) 6446±770 (p=0.84) HOMA-IR 14.6±5.8 9.2±3.5 (p=0.06) 6.3±1.8 (p=0.06) Fasting GLP pmol/litre 29.0±2.6 32.5±2.7 (p=0.21) 30.3±2.6 (p=0.70) Postprandial GLP-1 pmol/litre/min 4440 ± 249 6407 ± 480 (p<0.01) 6008 ± 429 (p<0.01) Fasting GIP pg/ml 145.9±23.3 233.1±128.3 (p=0.50) 155.1±29.8 (p=0.79) GIP pg/ml/min 115,272 ± 10,971 99,388±11073 88,499 ± 10,971 (p<0.05) Fasting glucagon pg/ml 105.9±14.9 79.7±15.2 (0.12) 78.7±14.9 (p=0.16) Glucagon AUC pg/ml/min 23,762 ± 4,732 15,989 ± 3,193 (p=0.02) 13,1207 ± 1,946 (p=0.02) 1 week after device removal (at week 25) in a subset of 8 patients, glucose response to a meal remained decreased and no significant changes in insulin response or HOMA-IR were observed. GLP-1 levels decreased but the improved glucagon and GIP response remained the same. Mean weight loss At 24 weeks after implantation, patients lost 12.7±1.3 kg (p<0.01).
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; GLP-1, glucagon like peptide-1; GIP, gastric inhibitory peptide; HbA1c, glycated haemoglobin; T2DM, type 2 diabetes.

 

Study 9 Koehestanie P (2014)

Details

Study type	Case series
Country	Netherlands
Recruitment period	Not reported
Study population and number	Obese patients with type 2 diabetes n=12
Age and sex	Mean 50.3 years; 40% (5/12) female Type 2 diabetes duration 7.4 years, BMI 33.5 kg/m2
Patient selection criteria	Aged between 18 and 60 years, BMI 28–35 kg/m2, T2DM with HbA1c level above 7%. Patients were allowed to take metformin, sulfonylurea derivatives and/or insulin.
Technique	DJBS (EndoBarrier) deployed and removed under GA after 24 weeks. Nutritional counselling, liquid diet in first week and 1200–1500 calories intake thereafter.
Follow-up	24 weeks
Conflict of interest/source of funding	Four authors received consultancy fees from GI Dynamics (manufacturer).

Key efficacy and safety findings

Efficacy	Safety
Number of patients analysed: 12 Fasting plasma levels and gut hormones effect on weight loss and glycaemic control (mean± standard error)   Baseline 1 week 4 weeks 24 weeks p value Fasting glucose mmol/litre 12.1±0.7 9.7±1.2 9.2±0.3 10.6±0.7 0.21 Fasting insulin mU/litre 21.5±6.0 11.4±2.5 7.2±1.1 15.5±2.5 <0.05 HOMA-IR 12.4±3.3 4.8±0.9 4.1±0.5 7.3±1.4 <0.05 HbA1c mmol/mol 73.3±4.5   67.7±3.3 61.3±4.0 0.39 GIP, pg/ml 206.5±37.5 142.9±16.6 136.5±13.4   0.20 GLP-1, pM 6.1±1.2 3.2±0.5 4.8±0.7   <0.05 Gherlin pg/ml 341.2±51.0 651.5±89.5 712.3±95.8   <0.05 Weight (kg) 104.9±3.0 101.3±2.9 99.9±2.9 97.7±3.3 0.39 BMI (kg/m2) 33.5±0.8 32.3±0.8 31.9±0.8 31.2±1.0 0.24 Fat mass (%) 40.3±1.7 40.0±1.9 35.0±1.9 33.1±1.8 <0.05 C-peptide nmol/l 1.3±0.1 1.1±0.1 1.2±0.1 1.1±0.1 (p=.85) 0.52 42% reduction in diabetes medication use (p<0.05) in the first week after implantation of DJBS.	‘No complications due to implantation’.
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; HOMA-IR, homeostatic model of assessment of insulin resistance; GA, general anaesthesia; GIP, gastric inhibitory peptide; GLP-1, glucagon like peptide-1; HbA1c, glycated haemoglobin; T2DM, type 2 diabetes mellitus.

Study 10 Lasle C (2014)

Details

Study type	Case report
Country	Germany
Recruitment period	Not reported
Study population and number	n=1
Age and sex	49-year-old man with BMI 40.9 kg/m2, and HbA1c 9.6%
Patient selection criteria	Not relevant
Technique	DJBS (EndoBarrier) implanted
Follow-up	4 weeks
Conflict of interest/source of funding	None

Analysis

Study design issues: standard clinical protocols were used.

Key efficacy and safety findings

Efficacy	Safety
	Number of patients analysed: 1 After 4 weeks, patient presented to emergency unit with an acute abdomen. Peritonitis in the right epigastric region noted. Radiologic imaging revealed free air in the abdomen suggestive of intestinal perforation. The DJBS was removed endoscopically and this was followed by laparoscopic closure of the perforation in the duodenal bulb using a running suture. The abdominal cavity was rinsed and drained. The patient recovered and was discharged 9 days after the surgery.
Abbreviations used: BMI, body mass index; DJBS, duodenal–jejunal bypass sleeve; HbA1c, glycated haemoglobin.

 

Efficacy

Glycaemic control

The RCT of 77 patients with obesity and type 2 diabetes compared DJBS treatment in combination with dietary intervention (n=38) against only dietary intervention (n=39) for 6 months. It reported that glycated haemoglobin (HbA_1c) levels improved from 8% at baseline to 7% in the DJBS plus diet group but remained at about 8% in the diet-alone group at 6‑month follow-up. The difference between the 2 groups was significant (DJBS plus diet versus diet alone; p<0.05). At 12‑month follow-up (including 6 months of post-DJBS removal), HbA_1c was 7% in the DJBS plus diet group and 8% in the diet-alone group (p=0.95).¹

The RCT of 77 patients reported that, at 12‑month follow-up (including 6 months after DJBS removal), fasting glucose levels had decreased from 11.0 mmol/litre to 9.0 mmol/litre in the DJBS plus diet group compared with to 9.7 mmol/litre in the diet-alone group. The difference between the 2 groups was not significant (DJBS plus diet versus diet alone; p= 0.41) ¹.

The RCT of 77 patients reported that, at 12‑month follow-up (including 6 months after DJBS removal), fasting insulin levels remained the same in the DJBS plus diet group (15.0 mU/litre), and decreased in the diet-alone group from 17.0 mU/litre to 15.7 mU/litre. The difference between the 2 groups was not significant (DJBS plus diet versus diet alone; p= 0.73) ¹.

The RCT of 18 patients with obesity and type 2 diabetes comparing DJBS (n=12) against sham endoscopy (n=6) reported that in the intention-to-treat group, HbA_1c values decreased by −1.3±0.9% for the DJBS group and by −0.8±0.3% in the sham endoscopy group (p>0.05) at 12‑week follow-up. At 24‑week follow-up, the HbA_1c had decreased by 2.4±0.7% in the DJBS group and by 0.8±0.4% in the sham endoscopy group (p>0.05). These differences were not statistically significant. Mean postprandial glucose area under the curve was reduced in the DJBS arm by 22% from baseline, compared with a 16% increase in the sham endoscopy group (p=0.016)².

In a case series of 81 obese patients with type 2 diabetes, after 6 months of DJBL implantation more than 70% of patients had HbA_1clevels of less than 7%. No further details were reported³.

A case series of 22 morbidly obese patients with type 2 diabetes (mean BMI 44.8 kg/m²) reported statistically significant reductions in fasting blood glucose (30.3 mg/dl), fasting insulin (−7.3 mU/ml) and HbA_1c (−2%) at last observation carried forward (LCOF) in all patients analysed. At the end of the study 73% (16/22) of patients had an HbA_1c under 7% compared with 5% (1/22) at baseline⁴.

A case series of 23 moderately obese patients with type 2 diabetes (mean BMI 30 kg/m²) reported a reduction in fasting plasma glucose (from 207±61 mg/dl to 155±52 mg/dl, p=0.012) and HbA_1c (from 8.7±0.9% to 7.5±1.6%,p=0.004) at 1‑year follow-up⁵.

Glycaemic control after removal of device

The case series of 22 patients with obesity and type 2 diabetes treated with a DJBS reported that improved glycaemic control (mean percentage decrease of -17±0.7% in HbA_1c from a baseline level of 8.9±1.7%) continued for up to 6 months after device removal in 11 patients ⁴.

A case series of 16 obese type 2 diabetic patients reported that 1 year after DJBS implantation, the mean HbA_1c level and fasting glucose levels decreased (HbA_1c to 7.5% from baseline 8.6%; p<0.001) (fasting glucose to 150.2mg/dl from baseline 203.3mg/dl; p<0.001) and insulin sensitivity and resistance improved (median Mastuda index score to 2.4 from baseline 1.7; p<0.001), (HOMA-IR score to 4.3 from baseline 6.6; p<0.001). 6 months after device explantation all these variables deteriorated. Fasting insulin levels, insulin area under the curve, fasting C‑peptide, C‑peptide area under the curve, fasting insulin secretion rate after a meal and total insulin secretion rates did not change either during the implantation period or after explantation of DJBS⁶.

Percentage of excess weight loss

The RCT of 77 patients with obesity and type 2 diabetes comparing DJBS treatment plus dietary intervention (n=38) against only dietary intervention (n=39) reported a significantly higher percentage of excess weight loss (EWL) at 12‑month follow-up (including 6 months of post DJBS removal), for DJBS group than for the diet-alone group (20% versus 12% respectively, p<0.05)¹.

Glycaemic factors associated with weight loss

In a case series of 79 patients (including 21 with type 2 diabetes) univariate analysis identified that fasting glycaemia (r²= −0.303, p<0.013), insulin resistance determined by HOMA‑IR (r² = −0.457, p<0.019) and glycated haemoglobin HbA_1c (r²= −0.471, p<0.001) were inversely associated with percentage of EWL at 52 weeks after DJBS implantation⁷.

The case series of 23 patients reported that the change in body weight loss was not significantly associated with a change in fasting plasma glucose at 1‑year follow-up (values not reported)⁵.

Change in cardiovascular parameters

The RCT of 77 patients reported that, at 12‑month follow-up, blood pressure decreased from 147/92 mmHg to 130/82 mmHg in the DJBS plus diet group and from 152/90 mmHg to 140/85 mmHg in the diet-alone group. The difference between the 2 groups was not significant (DJBS plus diet versus diet alone; p=0.31 for systolic pressure and p=0.38 for diastolic pressure). At 12‑month follow-up, total cholesterol levels in the DJBS group were comparable with baseline (4.4 mmol/litre in both groups [p=0.79])¹.

The case series of 22 patients treated by a DJBS reported significant reductions in total cholesterol (19.7±5.9 mg/dL; p<0.01) and triglycerides (44.8±17.4 mg/dL; p<0.05) at LCOF on or before explantation⁴.

Reduction in insulin resistance and metabolic syndrome

The case series of 81 patients reported improvement in insulin resistance, with a significant reduction of the TG/HDL ratio from 5.75 to 4.36 (p<0.001) and 43% of the patients presented a TG/HDL ratio lower than 3.5, 6 months after DJBS implantation³.

Diabetes medication use

In the RCT of 77 patients, at 12‑month follow-up the daily insulin dose and use of sulfonylureas had decreased or the medication had been stopped in the DJBS plus diet group more often than in diet-alone group (p<0.05)¹.

A case series of 12 obese patients reported 42% reduction in diabetes medication use (p<0.05) 1 week after DJBS implantation⁹.

Implantation failure or difficulties

In the case series of 81 patients the DJBS could not be implanted in 4% (3/81) of patients because of a short duodenal bulb³.

Safety

Perforation

Perforation of the duodenal bulb (associated with the DJBL) was observed 4 weeks after implantation of DJBL in a case report of 1 patient. The device was removed endoscopically and the perforation was closed laparoscopically with a running suture. The patient was discharged 9 days after the surgery¹⁰.

Oesophageal perforation (6 cm) during device removal at 6 months (caused by one of the anchor barbs that was not covered by the removal hood), was reported in 1 patient in the DJBS group (n=38) in the RCT of 77 obese patients with type 2 diabetes. This was treated by endoscopic stenting and placement of a feeding tube. The tear resolved without sequelae within 3 weeks¹.

Early device explantation

In the RCT of 18 patients 42% (5/12) of devices were explanted early in the DJBS group, because of device migration (4 because of anchor migration and 1 because of ‘device turning or migration’) during 12‑week follow-up. Three patients presented with symptoms such as pain, nausea and vomiting and 2 were asymptomatic². In the RCT of 77 patients, 1 patient in the DJBS group (n=38) had an obstruction which was resolved by early device removal¹.

Early removal was needed in 40% of the patients (in the case series of 22). This was because of: This was because of: device migration or rotation in 3 patients (at 48 weeks); gastrointestinal bleeding in 1 patient (at 4 weeks); abdominal pain in 2 patients (at 24 and 30 weeks); principal investigator request because of non-compliance with follow-up in 2 patients(at 20 and 32 weeks); and discovery of an unrelated malignancy (at 17 weeks; n=1)⁴.

In the case series of 79 patients, 26% (21/79) of devices were explanted early due to device migration (n=8), obstruction (n=5), abdominal pain (n=2), liver abscess (n=1), upper gastrointestinal bleeding (n=1), cholangitis (n=1), ulcerative colitis (n=1), acute cholecystitis (n=1) and patient request (n=1)⁷.

Nausea, vomiting and abdominal pain

Upper abdominal pain, minor gastrointestinal symptoms or discomfort were reported in 63% (25/38) of patients in the DJBS group and 28% (11/39) of patients in the diet-alone group in the RCT of 77 patients. Nausea or vomiting occurred in 24% (9/38) of patients in the DJBS group and in 18% (7/39) in the diet-alone group, all of whom were managed conservatively¹.

Back pain

Device-related back pain was reported in 23% (5/22) of patients in the case series of 22 patients⁴.

Hypoglycaemia

Mild-to-moderate hypoglycaemia was reported in 24% (9/38) of patients in the DJBS group and 26% (10/39) patients in the diet-alone group in the RCT of 77 patients¹.

Iron deficiency

Metabolic and nutritional disorders, including hypoglycaemia and iron deficiency, occurred in 61% (14/23) of patients in a case series of 23 patients⁵.

Validity and generalisability of the studies

• Most of the studies published were small and implanted the device for a period of 3, 6 or 12 months only. One of the randomised clinical trials (Rodrigues L 2009¹) included a first generation DJBL.
• The evidence comes mainly from studies in South America and Europe (none from the UK).
• The studies included only patients with type 2 diabetes and obesity.
• There is a lack of data on management after explantation.
• There is a lack of long-term data on how long any beneficial effect may last after removal of the device.
• There is a lack of patient-reported outcomes data.
• The majority of the studies are sponsored by the manufacturer.
• There is overlap of patients between the Cohen RV^5,6 studies.
• Limitations of the evidence base, such as uncertainties over method of response, lack of long-term data etc.

Existing assessments of this procedure

A Horizon Scanning Prioritising Summary Report conducted for Australia and New Zealand in 2010 concluded that ‘EndoBarrier appears to have the potential to induce significant weight loss and improve diabetic symptoms’. It is mainly based on evidence from 4 RCTs^1-4. In addition, it concludes that ‘additional comparative studies with appropriate controls are necessary as the evidence base for this device is limited and lacks long-term follow-up results’¹¹.

The American College of Surgeons’ report on endoluminal treatments for obesity in 2010 assessed the DJBS procedure using EndoBarrier. It concluded that ‘the early evidence on the effectiveness of the EndoBarrier was encouraging. In comparison to diet control alone, patients who received the EndoBarrier lost significantly more weight and also experienced considerable improvements in their diabetic symptoms. However, when compared to patients who received sham endoscopy, those who underwent EndoBarrier treatment did not lose significantly more weight compared to the sham controls at 20 weeks’ follow up. Self-limiting nausea (up to 77%) and upper abdominal pain (up to 30%) were common in patients who received the EndoBarrier and some serious complications were evident, with early removal being required in 20% to 40% of patients’. It considered that ‘additional long-term comparative studies (with appropriate controls) are necessary before any firm conclusions can be made regarding the safety and efficacy of the emerging procedures and devices. Until then these procedures and devices should only be used in a clinical trial setting’. In addition, it concluded that ‘future research is necessary to determine if there are any particular patients’ subgroups that may particularly benefit from certain procedures’. It also recommends that ‘these procedures and devices are new and are undergoing active development and should be monitored as refinements will alter their safety and efficacy profiles’¹².

Related NICE guidance

Below is a list of NICE guidance related to this procedure. Appendix B gives details of the recommendations made in each piece of guidance listed.

Interventional procedures

• NICE interventional procedures guideline 471 (2013). Available from ipg471

Clinical guidelines

• NICE clinical guideline 43 (2006). Available from cg43
• NICE clinical guideline 87 (2009) Available from cg87
• NICE public health guidance 35 (2011). Available from ph35

Specialist advisers’ opinions

Specialist advice was sought from consultants who have been nominated or ratified by their specialist society or royal college. The advice received is their individual opinion and does not represent the view of the society.

Mr James Byrne, James Hopkins, Kesava Mannur (Association of Upper Gastrointestinal Surgeons of Great Britain and Ireland); Dr Barbara McGowan, Dr Bob Ryder, Dr Stephanie A Amiel (Association of British Clinical Diabetologists (ABCD).

• Two specialist advisers have performed the procedure at least once and 4 specialist advisers have never performed it but have taken part in patient selection or referred patients for this procedure. One adviser stated that it is performed in specially selected patients or those in clinical trials.
• Four specialist advisers considered the procedure to be novel and of uncertain safety and efficacy, 1 considered it to be the first in a new class of procedure and 1 thought that it was an established practice and no longer new as its been around for a few years.
• Specialist advisers listed the relevant comparators as diet and exercise, medical management of type 2 diabetes, bariatric surgery procedures such as gastric banding, gastric bypass (laparoscopic proximal Roux-en-Y), biliopancreatic diversion (duodenal switch) or laparoscopic sleeve gastrectomy.
• The procedure is likely to be performed by less than 10% of specialists.
• Advisers stated that there is some controversy as to whether the procedure is performed by medical endoscopists or bariatric surgeons and, if widely adopted, there could be tensions between bariatric surgeons practising within established bariatric surgical practice, gastroenterologists and endocrinologists.
• Key efficacy outcomes listed include reduced insulin resistance leading to improved glycaemic control, reduction in HbA_1c, reduction in hypoglycemic medication use, fasting insulin, C-peptide, HOMA-IR, percentage of weight loss, percentage of excessive weight loss, maintenance of weight loss, improvement in hypertension and quality of life both in the short and long term (that is after the implant has been removed).
• Specialist advisers stated that the main concerns relate to variable efficacy between patients, the extent of improvement in diabetes and weight loss, duration of improvement in diabetes and weight loss (including long-term efficacy after device removal), and lack of evidence from randomised controlled trials.
• Theoretical adverse events reported include gastrointestinal tract laceration, oropharyngeal, esophageal, gastric or bowel perforation, peritonitis, bleeding, aspiration, infection; small bowel obstruction (with knotting or kinking of the sleeve), device intolerance, migration, erosion, vitamin and mineral deficiency, dehydration, constipation, belching, bloating, diarrhea, hypoglycemia, hyperglycemia, flatulence, gastro-oesophageal reflux disease, oesophagitis, pseudopolyps, nausea, vomiting, gastrointestinal pain, peptic ulcer disease, duodenitis, local inflammation, back pain and adynamic ileus.
• Anecdotal adverse events listed include difficulties in deploying the device, halting the procedure after endoscopy due to residual food in stomach, nausea, vomiting, gastrointestinal pain, cramping, intussusception, recurrence of previous neurological leg pain, bowel obstruction, perforation, gastric bleeding due to inappropriate prescription of nonsteroidal anti-inflammatory drugs, pharyngeal obstruction during explantation, liver abscess (due to the device being left in-situ for more than 2 months beyond the recommended implant duration) and misplacement of device hood in pharynx during device removal.
• Specialist advisers stated that availability of the bariatric multidisciplinary team for patient selection and follow-up, surgeons with experience in bariatric surgery and good upper gastrointestinal endoscopic skills are needed to implant and remove the device and deal with rare complications. Advisers also stated that good training under supervision by an experienced proctor in the technique and standard operating theatre facilities with fluoroscopy, endoscopy equipment and anaesthetic support are needed. Patient selection and adjustment of diabetes medication need to be accommodated.
• Specialist advisers stated that the National Bariatric Surgery Registry (NBSR) hosts an EndoBarrier registry for UK EndoBarrier cases and the manufacturer is also setting up an international registry. One adviser stated that the ABCD has a presence on N3, the NHS broadband network, and is volunteering to host the international registry
• There are few studies currently in progress (2 from the UK) that are likely to inform practice.
• Two advisers stated that the likely speed of diffusion cannot be predicted at present as it all depends on how effective, safe and cost effective the procedure turns out to be. One adviser stated that it could be a major procedure in people with a BMI of less than 35 kg/m² in the next 5 years and may surpass the surgical treatments.
• One adviser stated that currently the procedure is mainly done in the private sector and is expensive for widespread adoption as costs are similar to bariatric surgical procedures. He suggests that it should be introduced in the NHS after review of long-term evidence.
• Three advisers stated that the procedure is likely to be carried out in a minority of hospitals, but at least 10 in the UK (mainly in teaching hospitals and major district general hospitals) if it is safe and efficacious.
• In terms of patient numbers and use of resources, 3 advisers stated that the impact on the NHS would range between moderate to minor and 2 advisers stated that it would be major because of the large number of people who are eligible for this procedure (obese and have type 2 diabetes, tried previous treatments but have failed treatment or did not tolerate it). One adviser stated that DJBS has the potential as a medium-term aid to significant weight reduction initially to prepare vulnerable patients for definitive surgery but also as a stand-alone treatment to start significant weight reduction with metabolic improvement. Another adviser stated that, if found safe and efficacious, the procedure can be offered as part of the bariatric procedures on offer for obesity and type 2 diabetes, with patients having to fulfill requirements for lifestyle/dietetic programmes before being eligible.

Patient commentators’ opinions

NICE’s Public Involvement Programme sent 35 questionnaires to 2 NHS trusts for distribution to patients who had the procedure (or their carers).

Issues for consideration by IPAC

• The device has not yet received US Food and Drug Administration (FDA) approval.
• Several additional studies are ongoing:
• NCT01114438: Post Marketing Study in Subjects Who Have Type 2 Diabetes Using the EndoBarrier™ Gastrointestinal Liner; type: open-label single-group assignment; location: United Kingdom (Imperial College/St. Mary's Hospital, London; Trafford General Hospital/NOSC, Manchester; Southampton General Hospital, Southampton); estimated enrolment: 45 patients; inclusion criteria: subjects with type 2 diabetes for more than 1 and up to 10 years who are on oral diabetic medications and/or insulin, with an Hb A1c level over 7.5 and up to 10.0 and a BMI over 30 and under 50; primary outcome: HbA_1c at 12 months; estimated primary completion date: January 2013 (status: completed, but not yet published).
• NCT00986349 Study of EndoBarrier Liner for Treatment of Type 2 Diabetes; location: Brazil; type: open-label single-group assignment; estimated enrolment: 20 patients; inclusion criteria: subjects with type 2 diabetes who have been treated for ≤10 years and are on oral diabetic medications, with an HbA_1c level over 7.5 and under 10% and with a BMI over 26 and under 50; estimated study completion date: November 2012 (completed).
• NCT00985114: Safety and efficacy study of EndoBarrier in subjects with type II diabetes and obesity; type: multicentre RCT with crossover (after 12 month washout); location: Netherlands; estimated enrolment: 70 patients; inclusion criteria: type 2 diabetes treated for under 10 years, BMI over 30 and under 50, with an HbA_1c level over 7.5 and under 10%; primary end point: percentage of patients who achieve a greater than 0.5% reduction in HbA_1c at 24 weeks or last visit from baseline; study completion date: January 2012 (status: completed, in press).
• NCT01728116: Safety and efficacy of EndoBarrier in subjects with type 2 diabetes who are obese (ENDO); type: RCT; location: USA; estimated enrolment: 500; inclusion criteria: HbA_1c over 8% and under 10%, BMI over 30 and under 50; primary outcome: improvement in HbA_1c at 12 months; estimated study completion date: June 2015 (currently recruiting).
• NCT01718457: Endobarrier treatment in obese subjects with type 2 diabetes; type: interventional, single-group assignment; location: Israel; estimated enrolment: 45; estimated study completion date: January 2018 (not recruiting).
• NCT02055014: Randomization to Endobarrier alone versus with incretin analogue in sustained diabesity (REVISE-Diabesity). Type: RCT (Endobarrier with continued liraglutide 1.2 mg for 12 months; Endobarrier alone for 12 months; or liraglutide 1.8 mg without Endobarrier); study population: patients with type 2 diabetes and obesity (HbA_1c ≥ 7.5%, BMI ≥ 35kg/m²) despite previous GLP-1RA therapy; outcomes: HbA_1c, mmol/mol; % [24 months ], Weight [24 months ]; location: UK; estimated enrolment:72; estimated study completion date: December 2016 (currently recruiting).
• NCT01724060: Effects of obesity treatments (including EndoBarrier) on food preferences and metabolism (FPS); type: observational case-control study; location: UK; estimated enrolment: 400; estimated study completion date: October 2014 (currently recruiting).
• NIHR EME sponsored randomised study in UK (Southampton and London); Type: RCT; estimated enrolment: 160 patients randomised to either EndoBarrier or best medical treatment; presently in ethics, to recruit Septmeber 2014 and will report in 2017/8.

References

1. Koehestanie P, de Jonge C, Berends FJ et al (2014). The Effect of the Endoscopic Duodenal-Jejunal Bypass Liner on Obesity and Type 2 Diabetes Mellitus, a Multicenter Randomized Controlled Trial. Annals of Surgery: Publication ahead-of-print, doi: 10.1097/SLA.0000000000000794
2. Rodriguez L, Reyes E, Fagalde P et al (2009). Pilot clinical study of an endoscopic, removable duodenal-jejunal bypass liner for the treatment of type 2 diabetes. Diabetes Technology & Therapeutics 11 (11) 725-732.
3. de Moura EG, Orso IR, Martins Bda C et al (2011). Improvement of insulin resistance and reduction of cardiovascular risk among obese patients with type 2 diabetes with the duodenojejunal bypass liner. Obesity Surgery 21 (7) 941-947.
4. de Moura EG, Martins BC, Lopes GS et al (2012). Metabolic improvements in obese type 2 diabetes subjects implanted for 1 year with an endoscopically deployed duodenal-jejunal bypass liner. Diabetes Technology & Therapeutics 14 (2) 183-189.
5. Cohen RV, Neto MG, Correa JL et al. (2013). A pilot study of the duodenal jejunal bypass liner in low body mass index type 2 diabetes. J Clin Endocrinol Metab. 98(2):E279-82.
6. Cohen RV, le Roux CW, Papamargaritis D et al (2013). Role of proximal gut exclusion from food on glucose homeostasis in patients with Type 2 diabetes. Diabetic Medicine 30 (12) 1482-1486.
7. Munoz R, Dominguez A et al (2014). Baseline glycated hemoglobin levels are associated with duodenal-jejunal bypass liner-induced weight loss in obese patients. Surgical Endoscopy 28 (4) 1056-1062.
8. de Jonge C, Rensen SS et al. (2013) Endoscopic Duodenal-Jejunal Bypass Liner Rapidly Improves Type 2 Diabetes. Obes Surg. Mar 23.
9. Koehestanie P, Dogan K et al (2014). Duodenal-jejunal bypass liner implantation provokes rapid weight loss and improved glycemic control, accompanied by elevated fasting ghrelin levels. Endoscopy International Open.2 (1) (pp E21-E27), 2014.Date of Publication: 2014. (1) E21-E27.
10. Lasle C, Laubner K et al (2014). Minimally invasive treatment of a duodenal perforation associated with the EndoBarrier duodenal-jejunal bypass liner. Endoscopy 46 Suppl-2.
11. Perera C and Lee L (2010). EndoBarrier Gastrointestinal Liner for obesity. Australian Safety and Efficacy Register of New Interventional Procedures-Surgical (ASPERNIP-S) on behalf of Australia and New Zealand Horizon Scanning Network.
12. Endoluminal treatments for obesity (2010). Horizon scanning in surgery: application to surgical practice and education. Prepared by Australian Safety and Efficacy Register of New Interventional Procedures – Surgical for the American College of Surgeons.