Banded Versus Nonbanded Sleeve Gastrectomy

A Randomized Controlled Trial With 3 Years of Follow-up

Jodok M. Fink, MD; Andrea Hetzenecker, MD; Gabriel Seifert, MD; Mira Runkel, MD; Claudia Laessle, MD; Stefan Fichtner-Feigl, MD; Goran Marjanovic, MD


Annals of Surgery. 2020;272(5):690-695. 

In This Article


The complete study protocol was approved by the local ethics committee and conducted in accordance with the principles of the Declaration of Helsinki. All patients gave written informed consent.

Study Design

The study was a single-center, open-label, randomized controlled trial including patients eligible for sleeve gastrectomy according to the German guidelines for bariatric surgery. The first patient was included in January 2015. The last follow-up visit was conducted in August 2019. The study was registered with the German Clinical Trials Register (DRKS00007729). The full trial protocol can be accessed at


Eligibility criteria included patient age ≥18 and ≤65 years, history of conservative weight-loss treatment failure and body mass index (BMI) ≥35 kg/m2 with relevant obesity-related comorbidities or a BMI ≥40 kg/m2. Exclusion criteria were untreated psychiatric illness, alcohol or drug abuse, chronic inflammatory bowel disease, liver cirrhosis, pregnancy, expected poor compliance, a <5-year history of malignant disease, and gastroesophageal reflux disease with Barrett esophagus or hiatal hernias >5 cm. Furthermore, patients with previous extensive gastrointestinal or bariatric surgery were excluded.


Patients were randomized using sequentially numbered, sealed envelopes that had been randomly filled with group allocations in a 1:1 ratio. Randomization took place during the final preoperative visit at the institution's outpatient clinic. Both patient and surgeon were informed about the allocated study group.


SG and BSG were performed in a single center by 2 experienced surgeons using the already established technique described earlier.[11] Sleeve formation starts 4 to 6 cm from the pylorus along a 35 Fr bougie applying a linear Endo GIA Tri-Staple device (45/60 mm purple cartridges, Medtronic) until the angle of His. For silicone ring (MiniMizer, Bariatric Solutions, Switzerland) implantation in BSG (Figure 1), the introducer is pushed through the peritoneum of the lesser curvature 4 cm below the gastroesophageal junction and guided along the posterior gastric wall sparing gastric vessels. The ring is closed at a circumference of 7.5 cm. A nonabsorbable suture secures the ring's fixation loop to the anterior gastric wall. Hiatal hernias were not repaired in this study. Adherence to surgical protocol was reviewed using video snippets of the interventions.

Figure 1.

Banded sleeve gastrectomy using a silicone ring (MiniMizer, Bariatric Solutions, Switzerland). The ring is placed 4 cm below the gastroesophageal junction and closed at a circumference of 7.5 cm (modified image – originally provided by International Federation for the Surgery of Obesity and Metabolic Disorders).

Outcome Measurements

Study visits matched clinical routine in the authors' institution and were set at baseline, 1, 3, 6, 12, 24, and 36 months postoperatively. Unless described elsewise, outcomes were determined at every visit.

Primary Outcome. Primary outcome was defined as excess weight loss (%EWL) 3 years after surgery. %EWL was measured as (initial weight – follow-up weight)/(initial weight – bodyweight at a BMI of 25 kg/m2) × 100.

Secondary Outcomes. Type 2 diabetes (T2D) was evaluated recording oral medication, insulin use, and glycated hemoglobin (HbA1c, not determined at 1 and 3 months). Diabetes remission was defined as normal glycemic measurements of at least 1-year duration in absence of antidiabetic medication.[14] Improvement was defined as relevant reduction in HbA1c not meeting criteria for remission.

Number of antihypertensive agents was grouped as 1 to 2, 3 to 4, or ≥5. Quality of life was evaluated using the Bariatric Analysis and Reporting Outcome System (BAROS).[15] Reflux symptoms included heartburn, acidic reflux to the mouth, and retrosternal pain. Regurgitation was considered as regurgitation of undigested food postprandially. Both symptoms were documented as either not present, incidence of ≥1/week, ≥1/month, or <1/month. Dysphagia was graded as 0: no dysphagia, 1: some solid food tolerated, 2: only soft food possible, 3: only liquid food possible, 4: complete dysphagia.[16]

Blood serum values for Vitamin D3, B1, B12, and folic acid were recorded at baseline and yearly visits. Deficiency was defined as serum value below ranges quoted in Table 1.

Postoperative complications were defined as early/late or major/minor following guidelines of the American Society for Metabolic and Bariatric Surgery.[17] Total weight loss (%TWL) was determined as [(initial weight – follow-up weight)/initial weight] × 100.

The original study protocol included MRI sleeve volumetry 1 and 3 years after surgery as well as iron and lipid serum levels. Iron and lipid evaluation was abandoned because of missing baseline values, volumetry due to low patient adherence and high internal costs. Assessment of dysphagia was nondiscriminative (>90% grade 0) and excluded from evaluation.

The study protocol was amended to include gastroscopic assessment of hiatal hernias and reflux esophagitis (baseline and 3 years). Reflux Symptom Index (RSI) recorded at 3-year follow-up was included as well.

Statistical Analysis

Sample size was calculated to detect weight loss difference of 8%EWL 3 years after surgery with a power of 80% assuming a standard deviation of 13%EWL. Based on a 2-sided, 2-samle t test 84 patients were required. Sample size was adjusted to 94 to account for an estimated drop-out rate of 10%.

Continuous variables were characterized using mean and confidence interval (CI). Categorial variables were summarized using frequencies and percent of patients. All patients were included as intention-to-treat. A linear mixed-model for repeated measures (MMRM) approach was used to compare both groups with respect to mean outcome over time. The MMRM model included patient as a random effect. Group allocation, follow-up visit, type of surgery-by-visit interaction, baseline value of the outcome, sex, age, and presence of T2D were interpreted as fixed effects. An unstructured within-patient covariance structure was assumed. HbA1c was analyzed in diabetic patients only.

BSG and SG were compared with respect to primary endpoint at 3 years. Regarding secondary longitudinal endpoints (ie, quality of life and HbA1C), comparisons between the 2 interventions were performed at several time points. P values were corrected for multiple comparisons using the Bonferroni-Holm method. Weight regain from nadir weight and RSI at 3 years were analyzed using a Mann-Whitney test. All statistical tests were performed at the 2-sided 0.05 significance level.

For categorical variables, Pearson chi square was used. Correlations between group allocation and ordinal variables (eg, frequency of regurgitation) were defined with a chi square for trend. Within-group development was analyzed using a McNemar-Bowker test.

Prism 8.3 for macOS (GraphPad Software, LLC) and SAS (SAS Institute Inc.) were used for statistical analysis.