Methods / Study design: RCT
No information about either random sequence generation or allocation concealment
No blinding
Duration: 7 years (mean Follow-up: 43 months)
No information regarding intention to treat analysis
Loss to follow-up at12 months: 18%, at 24 months: 33% and at 48 months: 67%
Patients / 45 super obese patients (at least 90 kg higher than ideal weight);
73% women; 18-61 years old
Mean preop. BMI=61.6±9 kg/m2
New-Jersey, USA
Intervention / Open surgery RYGB, 1 surgeon
Gastric pouch volume: 20±5 ml.
Roux-limb length:
1) 75 cm
2) 150 cm
Outcomes and results / Mean Weight Loss (MWL) in kg: greater MWL among gp2) vs. gp1) at 24 months (76.4±23.6 vs. 53.2±27.3 kg) and 36 months (75±22.3 vs. 52.3±22.2 kg) (p ≤ 0.02)
Mean postop. absolute BMI (MAB): at 24 months, significantly lower MAB among gp2) vs. gp1)
Mean EWL%: At 24 months, significantly greater mean EWL% in gp1) vs. in gp2)
% Success rate (>50% EWL): At 24 months, success rate=82% in gp2) vs. 50% in gp1) (p≤ 0.03)
Percentages of patients presenting vitamin and mineral deficiencies (Folate, Vit. B12, Iron) and iron-deficiency anaemia: non significant difference
Bias and remarks
Choban 2002 [27]
Methods / Study design: RCT
Stratified randomization according to patients’ initial BMI. No information about either random sequence generation or allocation concealment
No blinding
No information regarding intention to treat analysis
Duration: Not reported
Loss to follow-up: Not reported
Patients / 133 obese or super obese patients with BMI ≥ 40 naïve for any kind of bariatric surgery procedure.
84% women. Mean preop. BMI= 44±6 kg/m2 in gp 1) and2) and 61±2 in gp 3) and 4)
Ohio State, USA
Intervention / Open surgery retrocolic, antegastric RYGB
Gastric pouch volume: 15-30 ml
Roux-limb length:
1) BMI ≤50: 75 cm
2) BMI ≤50: 150 cm
3) BMI≥50: 150 cm
4) BMI≥50: 250 cm
Outcomes and results / MWL: Non-statistically significant difference at 6, 12, 24, 48 months
MAB: Non-statistically significant difference at 6, 12, 24, 48 months
Mean EWL%: Non-statistically significant difference at 6, 12, 24 and 48 months
% Success rate (≥ 50% EWL): At 18 months postop, success rate significantly greater in gp4) vs. gp3) (but no such significant difference in gp2) vs. gp1))
Bias and remarks / No information of loss to follow-up
Flancbaum 1998 [29]
Methods / Study design: RCT
Stratified randomization according to patients’ initial BMI. No information about either random sequence generation or allocation concealment
No blinding
No information regarding intention to treat analysis
Duration: Not reported
Loss to follow-up: Not reported
Patients / 60 obese or super obese patients
Mean baseline BMI = 42.9±0.6 in gp1) 46.3±1.16 in gp2) ; 61±1.6 in gp3) ; 61±1.4 in gp4)
85%women.
Ohio State, USA
Intervention / Open surgery, retrocolic (ante-retrogastric?) RYGB
Gastric pouch of ~30ml
Roux-limb length:
1) BMI 40-51: 75 cm
2) BMI 40-51: 150 cm
3) BMI >51: 150 cm
4) BMI >51: 250 cm
Outcomes and results / MWL in kg at 12 months only: Non statistically significant difference in MWL in gp1) vs. gp2) and gp3) vs. gp4)
Mean EWL% at 12 months only: Non statistically significant difference in EWL% at 12 months between gp1) vs. gp2) and gp3) vs. gp4)
Bias and remarks
Inabnet 2005 [30]
Methods / Study design: RCT
No information about either random sequence generation or allocation concealment
Patients, surgical team, consultants, ancillary staff and operating-room staff blinded
Duration: 20.5 months
No information regarding intention to treat analysis
Loss to follow-up not clearly reported
Patients / 48 obese patients with a BMI < 50 kg/m2;
Mean preop. BMI= 44.9±2.9 kg/m2.
96% women; mean age 35±9.5 years old
New-York, USA
Intervention / Laparoscopic antecolic, antegastric RYGB
Gastric pouch volume: estimated by authors to be ~ 20 ml
Roux-limb length:
1) 100 cm
2) 150 cm
Outcomes and results / MAW in kg at 3 weeks, 3 months, 6 months, 12 months, 18 months, 24 months: Non statistically significant difference MAW (in kg) between gp1) and gp2) at every time points
MWL: deducible, not directly reported by the authors
Mean EWL% (same time intervals as above) Non-statistically significant difference in EWL% between gp1) and gp2) at every time points
MAB (same time intervals as above): Non-statistically significant difference in MAB (kg/m2) between gp1) and gp2) at any time points
Mean values for PTH, Calcium, iron, folate, selenium, zinc, vit. B12, vit. D, albumin at 3-12 months: Non-statistically significant difference in nutritional status at any time point
Bias and remarks / Due to the fact that the academic group changed institution during the study, only part of the included patients (48/78) were available for analysis
Pinheiro 2008 [33]
Methods / Study design: RCT
No information about either random sequence generation or allocation concealment
No blinding
Duration: Not reported; Mean follow-up 48 months (range 6-56)
No information regarding intention to treat analysis
Loss to follow-upNot clearly reported
Patients / 105 super obese patients (BMI range: 50.6-76 kg/m2) and all Type 2 diabetic
71 % women. Mean age, gp1)andgp2): 53 and 56 years old, respectively
Sao Paolo, Brazil
Intervention / Laparoscopic antecolic, antegastric RYGB
No information regarding gastric pouch volume
Roux-limb length:
1) 150 cm
2) 250 cm
Outcomes and results / Mean EWL% at 6, 12, 24, 48 months: Non-statistically significant differences in EWL% at all time points
T2DM improvement, control: T2DM controlled in 93% of gp2) patients vs. 58% of gp1) patients (p < 0.05) (disease control was achieved during the first 12 weeks postop. and lasted throughout the whole study)
Lipid disorder improvement: Lipid disorder improved in 57% of gp1) patients vs. 70% of gp2) patients (p ≤ 0.05) at a mean follow-up of 9 and 8 months respectively
HTN control: Non statistically significant improvement in HTN
Nutritional outcomes (anaemia, Iron, Vit. A, B12, D, Calcium, Zinc, Magnesium, Parathyroïd hormone: All patients necessitated Vit. D supplementation. Vit. A supplementation was necessary in 13% vs. 27% in gp.1) andgp.2), respectively. 16% of gp.2) patients requiredVit. B12 supplementation
In both groups:
GERD symptoms totally controlled
Patients with sleep apnoea no longer uses CPAP (proxy measure for sleep apnoea symptoms prevalence)
Bias and remarks / Measurement bias: GERD assessed either clinically or endoscopically preop. but only clinically postop.
Sleep apnoea assessed via polysomnography preop. but with rate of CPAP use postop
Poor reporting of postop. assessment of GERD and sleep apnoea
No statistical analysis regarding nutritional outcomes
Lee 2006 [31]
Methods / Study design: Prospective cohort study
Duration 41 months
Loss to follow-up: 41.3 %
Patients / 97 obese and super obese patients drawn from an US army population. Mean preop. BMI: 46.6.
87% women. Mean preop. age: 44 years
Texas, USA
Intervention / Open (94%) or laparoscopic (6%) (ante-retro- colic/-gastric?) RYBP
No information about gastric pouch
Steady increase in Roux-limb length according to the following principle: for patients with BMI≤40: 100cm; 5 cm extra Roux-limb length added for every unit of BMI over 40
Outcomes and results / MWL at 12 months: Statistically significant linear regression relationship between Roux-limb length and MWL (p = 0.0025)
Absolute change in BMI at 12 months: Statistically significant linear regression relationship between Roux-limb length and absolute change in BMI (p= 0.0011)
Nutritional status and deficiencies (laboratory data such as albumin, protein, iron): not reported in the results section
Bias and remarks / Highly mobile population, explaining the high dropout rate
Not normally distributed Roux-limb lengths, with a higher number of patients that received either 100 cm or 150 cm Roux-limb length, as these were the minimum and maximum length allowed in the protocol
Ciovica 2008 [28]
Methods / Study design: Prospective cohort study
Duration: 41 months
Loss to follow-up: 15% at 12 months follow-up
Patients / 161 super obese patients; Mean preop. BMI: 61.1 (range 50.1-103)
82.5% women. Age: 44.4 years (range 25-66)
San Francisco, USA
Intervention / Open (34.3%) or laparoscopic (65.7%) antecolic, antegastric RYGB
Gastric pouch of 30 ml
Roux-limb length:
1) 100 cm (n=102, 74.5%)
2) 150 cm (n=35, 25.5%)
Outcomes and results / Statistically significant difference between gp1) andgp2) regarding:
MWL in kg (55.3 (95% CI 52.2-58.5) vs. 68.5 (61.3-73.9)) (p < 0.01)
MAB in kg/m2 (40 (95% CI 39-41) vs. 36 (34-38)) (p < 0.01)
Mean BMI change in kg/m2 (21 (95% CI 20-22) vs. 25 (23-27)) (p=0.03)
EWL% (53% (95% CI 50-56) vs. 64% (58-69)) (p=0.02)
Bias and remarks / Risk of selection bias because:
a) Patients of gp1) were significantly older than those of gp2) (mean age 45.8 vs. 40.1 p=0.003)
b)More open procedure in gp1) vs. gp2) (p < 0.01)
Muller 2008 [32]
Methods / Study design: Prospective cohort study. Random matching for age, sex, and baseline BMI, yielding paired patients of each group
Duration mean follow-up =48 months
Loss to follow-up: at 48 months complete follow-up was 52 % for proximal bypass group and 20% of distal group
Patients / 50 morbidly obese patients. Mean age proximal bypass group: 37.9 y.o. Mean age distal bypass group: 38.8 y.o.
Mean preop. BMI in proximal bypass group: 45.9. Mean preop. BMI in distal bypass group: 45.8
75% women
Zürich, Switzerland
Intervention / Laparoscopic (100%) RYGB (ante-retro- colic/-gastric?)
Gastric pouch of 25 ml
Roux-limb length:
1) Proximal RYGB group: Roux-limb of 150 cm (n=25)
2) Distal RYGB group: jejunojejunostomy 100-150 cm proximal from ileocaecal valve (n=25) (common limb 100-150 cm)
Outcomes and results / Mean operating time: 170 min (SD=55.6) in gp 1) vs. 242 min (SD=86.5) in gp 2) (p=0.004)
Median length of hospital stay (days): 8 (Range=4-43) in gp 1) vs. 9 (Range=6-24) in gp2) (p=0.063)
MAB: Decrease in gp 1) from 45.9 to 31.7 kg/m2. Decrease in gp 2) from 45.8 to 33.1 kg/m2. No statistically significant difference between groups at 4 years follow-up, and throughout follow-up
EWL (no results reported in the results section)
Prevalence of co-morbidities (T2DM, lipid disorders, HTN) decreased in both groups by 2 years, with similar preop. and postop.distribution in the two groups)
No statistically significant differences between the groups in terms of the rates of early or late complications, endoscopic re-intervention, reoperation, complication rates, mortality
Bias and remarks / Patients in the distal group were significantly less likely to complete a 4 year follow-up than patients in the proximal group: 13/25 vs. 20/25, p=0.037)
ESM Table A1 Detail characteristics of the included studies.
BMI body mass index, MAB mean absolute BMI, MWL mean weight loss, MAW mean absolute weight, EWL excess weight loss, RCT randomized controlled trial, RYGB Roux-en-Y gastric bypass, T2DM type 2 diabetes mellitus, HTN hypertension, gp. group
Search strategy used in EMBASE[1]:
1 (superobese OR super-obese OR super obese OR superobesity OR super-obesity OR super obesity).mp. OR malabsorption/ OR obesity/ OR morbid obesity/ OR body weight/ OR intestine bypass/
2 gastric bypass.mp. OR stomach bypass/ OR Roux Y anastomosis/ OR roux en y gastric bypass.mp. OR roux-en-y gastric bypass.mp.
3 (limb lengthening or limb extending).mp. OR long limb.mp. OR short limb.mp.
We then combined these terms in: 1 AND2 AND3
[1]“/” defining MeSH term and “mp.” defining a random keyword term