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


Of 189 patients screened, 94 patients were randomized to undergo BSG or SG. In 2 patients randomized for BSG, a (banded) RYGB was performed. This was due to diagnosis of Barret's esophagus after randomization in one patient and to clinically relevant reflux complaints in the other patient. Table 2 contains baseline demographic data. Supplementary Figure 1, depicts flow of patients throughout the study and number of missing patients at each time point. 3 years after surgery, 97% of patients were available for evaluation.

Primary Endpoint

Adjusted %EWL 3 years after surgery was 62.3% (95% CI, 56.2–68.5) for SG and 73.9% (95% CI, 67.8–80.0) for BSG (P=0.0073). The difference in %EWL was 11.6%. Figure 2 displays %EWL throughout the study.

Figure 2.

Time course of percentage excess weight loss of each individual patient as well as adjusted means of each group. One patient in the sleeve gastrectomy group experienced weight regain above initial weight (not depicted).

Secondary Endpoints

Type 2 Diabetes. At baseline, T2D was insulin-dependent in 2 patients (both BSG). Mean duration of T2D prior to surgery was 3 years in both groups (range BSG 0–16 years, SG 1–6 years). At 3 years, diabetes remission was 24% higher in BSG (BSG 91% vs. SG 67%, Table 1). This difference was not significant (P=0.21). One patient with a long duration (16 years) of T2D remained on insulin medication. Mean adjusted HbA1c levels are displayed in Table 1.

Hypertension. At baseline, mean duration of medication use was 6 years in both groups (range BSG 1–15 years, SG 1–21 years). Sixty-four percent (16/25) of patients with BSG were treated with 1 to 2 agents (6 pat. 3–4; 3 pat. ≥5). This distribution was not statistically different for SG (7 pat. 1–2 agents; 8 pat. 3–4; 1 pat. ≥5; P = 0.44). From baseline to 3-year follow-up, number of antihypertensive agents significantly decreased in both groups (BSG, P = 0.0018; SG, P = 0.043). Further group comparison is found in Table 1.

Reflux. At baseline, clinically relevant reflux complaints (≥ 1/week) were present in 9% in BSG and 4% in SG. Three years after surgery, reflux complaints increased significantly in SG only (BSG 16%, P = 0.81; SG 33%, P = 0.024). Frequency of reflux symptoms was not different in both groups until 2-year follow-up (baseline, P = 0.48; 1 year, P = 0.87; 2 years, P = 0.17). It was significantly higher in SG 3 years after surgery (P = 0.004).

Regurgitation. Frequency of regurgitation was significantly higher after BSG (BSG vs SG 1 year P = 0.002; 2 years, P = 0.001; 3 years, P = 0.006). It remained stable from 1 to 3-year follow-up in both groups (BSG, P = 0.43; SG, P = 0.81). Percentage of frequent regurgitation (≥1/week) was significantly higher in BSG at 1 year (BSG 14%, SG 2%; P = 0.051), but not different at 3 years (BSG 13%, SG 4%; P = 0.13).

Vitamins. Details of vitamin deficiencies for each group are found in Table 1.

Quality of Life. Quality of life significantly improved from baseline to 3 years in both groups (BSG and SG, P < 0.0001). Change in BAROS score over the complete follow-up significantly differed between groups (operation × time P = 0.001). At 3 years, adjusted BAROS score was 1.5 points higher in BSG (BSG 6.44; SG 4.98; P = 0.0017).

Morbidity and Mortality

Number of complications is displayed in Table 3. Postoperative hemorrhage following SG was treated conservatively. Sleeve stenosis in SG was successfully dilated to 20 mm during a single endoscopic intervention. There was no revision due to regurgitation in either group. Major late complications occurred in 5.5%. Ring slippage in BSG led to laparoscopic ring removal. There was no mortality in either group.

Post Hoc Analysis

Gastroscopic Evaluation, Reflux Score. At baseline, reflux esophagitis was present in 18% of patients. 70.6% presented with esophagitis Los Angeles grade A (BSG 63.6%, SG 83.3%). One patient in BSG was diagnosed with Barret esophagus. Incidence of reflux esophagitis increased to 29.4% in BSG and 28.9% in SG, which could be majorly attributed to de novo esophagitis (BSG 50%, SG 90.1%). De novo reflux esophagitis was not associated with presence of hiatal hernias at 3 years in either group (BSG P = 0.24; SG, P = 0.9). There was no association between esophagitis and frequency of regurgitation (BSG, P = 0.85; SG, P = 0.41). Further details are found in Table 1.

RSI values were not statistically different between groups (BSG 3.65 ± 5.71 vs SG 4.03 ± 6.34; P = 0.94).

Total Weight Loss and Weight Regain. Adjusted total weight loss at 3 years was significantly higher after BSG (BSG 37.02, 95% CI 33.72–40.32 vs. SG 31.11, 95% CI 27.81–34.42; P = 0.011). Weight regain from nadir weight was significantly lower after BSG (SG 10.6 ± 6.51 vs BSG 5.45 ± 6.51%EWL; P = 0.02). Forty-seven percent of patients after BSG reached nadir weight at last follow-up (SG 30%).