Introduction
Obesity is a major health problem with a significant global burden. Surgery remains
the only recognized treatment for obesity and obesity-associated complications [1]. Endoscopic bariatric and metabolic therapies have been developed to fill the gap
between diet with lifestyle modification and surgery [2]. Among them, endoscopic sutured gastroplasty (ESG) has been associated with significant
weight loss and an excellent safety profile [3].
ESG was initially described using the Overstitch system (Apollo Endosurgery, Austin
Texas, United States). Overstitch has been the most frequently performed procedure
for treating class I and II obesity with good results reported in post-marketing retrospective
studies, while a systematic review and meta-analysis of eight (retrospective) studies
concluded that ESG using the Overstitch system may lead to a 15.6 % total body weight
loss (TBWL) at 12 months post-procedure [4]
[5]. The results of a randomized clinical trial (RCT) recently have been announced with
a 49 % EWL at 12 months, two times higher than lifestyle therapy alone in controls
[6].
Another system used for ESG is the Endomina platform (Endo Tools Therapeutics S.A.,
Gosselies, Belgium), which has been shown to be effective in prospective studies,
including a randomized cross-over trial, with excess weight loss (EWL) being significantly
higher in the treatment group compared to the control group at 6 months post-procedure
(38.6 % vs. 13.4 %; P < 0.001) [7]. However, despite the efficacy of these systems, no exact mechanism through which
ESG induces weight loss has been clearly identified. It has been hypothesized that
reduced gastric volume, effects on gastric emptying, and/or fundus distensibility
could be involved. In this context, we conducted a RCT to evaluate the effect of three
different suture patterns during ESG using the Endomina System (E-ESG) on post-ESG
weight loss.
Patients and methods
Study design
This investigator-initiated prospective, randomized study was conducted at Erasme
University Hospital in Brussels, Belgium in compliance with the Helsinki Declaration.
The hospital’s Institutional Review Board approved the study protocol (ClinicalTrials.gov:
NCT03843801). The devices were provided free of charge by Endo Tools Therapeutics
S.A. All patients provided written informed consent and the study is reported according
to CONSORT guidelines [8] (Supplementary Table 1).
Participants
Individuals were assessed for eligibility during outpatient consultations at our dedicated
multidisciplinary bariatric center. The inclusion criteria were as follows: age 18
to 65 years, obesity class I or II defined (body mass index [BMI] 30–40 kg/m²), ability
to comply with all study requirements for the duration of the study as outlined in
the protocol, ability to provide written informed consent and successful completion
of the prespecified multidisciplinary workup including a consultation with a gastroenterologist,
a psychologist, a dietician, and an occupational therapist. In addition, each patient
was discussed at the medical board meeting with the participation of bariatric surgeons
before inclusion in the protocol. The exclusion criteria were: presence of an esophageal
motility disorder, unstable angina, myocardial infarction within the past year, or
heart disease classified within the New York Heart Association’s Class III or IV functional
capacity, uncontrolled hypertension during the last 3 months, severe renal, hepatic,
or pulmonary disease, or active cancer, gastrointestinal stenosis or obstruction,
pregnancy or breastfeeding or willing to become pregnant within study duration, history
of any type of gastric surgery, impending gastric surgery 60 days post-intervention,
history of weight changes (± 5 % of the total body weight [TBW]) or placement or removal
of an intragastric balloon during the last 6 months, positive Helicobacter pylori status, and active participation in another clinical trial.
Randomization and concealed allocation
We randomly assigned participants (1:1:1) by computer-generated randomization to undergo
E-ESG with one of three different suture patterns (groups A, B, and C). Before starting
the procedure, the site study coordinator opened the concealed envelope to reveal
the group allocation to the endoscopist.
Endoscopic procedure and study groups
The Endomina System is a triangulation platform that can be used with any flexible
endoscope and a dedicated needle (TAPES, Endo Tools Therapeutics S.A., Gosselies,
Belgium) to create gastrointestinal sutures and has been described in detail elsewhere
[9].
The three evaluated suture patterns used for the purpose of this study are depicted
in [Fig. 1]. The first pattern consisted of creating a pouch at the upper part of the body by
placing intraluminal sutures mimicking a gastric band aiming to increase the distention
ability of the fundus. After placing three to five sutures, food was expected to accumulate
in the fundus, giving an early feeling of satiety. The second pattern aimed to drastically
reduce the size of the stomach by placing five to seven longitudinal sutures on the
anterior wall of the stomach, starting from the incisura and going up to the upper
body of the stomach. Finally, the third pattern consisted of making bridges (dams)
in the gastric body. Four to six sutures are placed antero-posteriorly along the greater
curvature of the stomach to create an interruption of normal gastric emptying and
provide a prolonged feeling of satiety (the pattern used in previously published RCT).
All endoscopic procedures were performed under general anesthesia with endotracheal
intubation and patients were hospitalized overnight per protocol for surveillance.
All patients received proton pump inhibitors (40 mg once a day for 3 months) post-intervention.
Patients were kept on a liquid diet for 3 days after the procedure, and then the food
texture was gradually broadened during the second week using mashed food as a transition.
Participants returned to solid food within 10 days.
Fig. 1 Graphic illustration and final endoscopic image of the three suture patterns and
their respective endoscopic results post-procedure. a Pattern “band” aiming to increase distensibility of the fundus. b Gastric “volume reduction.” c Pattern “dams” aiming to slow down gastric emptying.
Patient baseline and post-procedure evaluations
Nutritional follow-up
A dietary evaluation was performed pre-procedure and then at 2 weeks and 1, 3, 6,
9, and 12 months post-procedure. Patients were prescribed a low-calorie, high-protein
diet and lifestyle counseling based on Belgian Association for the Study of Obesity
(BASO) 2020 and American Society of Metabolic and Bariatric Surgery (ASMBS) 2013 guidelines
[10]
[11]. Physical activity was promoted. Psychological support was added to follow-up on
a case-by-case basis. Weight loss medication prescription was not allowed during the
study period.
Satiety test
Patients underwent satiety tests at the baseline evaluation, and at 6 and 12 months
post-procedure. The drinking satiety test (modified from a previously described version
[12]) consisted of drinking, every 60 seconds, 30 mL of a protein-rich and high-calorie
nutritional drink (Fortimel Energy [Nutricia] [1.5 kcal/mL, 5.9 g P/100 mL, 18.4 g
CHO/100 mL, 5.8 g L/100 mL] and Resource Energy [Nestlé Health Science] [1.5 kcal/mL,
5.6 g P/100 mL 21 g CHO/100 mL, 5 g L/100 mL]). Five symptoms (hunger, fullness, nausea,
bloating, and pain) were assessed on a 10-point scale (0 = no feeling, 10 = maximal
feeling) every 5 minutes. The test was stopped, and volume of intake was calculated
when the patient reported fullness to be at a level of 10 or if one of the other symptoms
became unbearable.
An upper gastrointestinal endoscopy was performed at 1, 6, and 12 months of follow-up
to assess the integrity of the suture pattern and the number of sutures remaining
intact.
Gastric emptying scintigraphy was performed at baseline, and at 6 and 12 months post-procedure.
The examination was performed according to the procedure laid down by the joint guideline
of Society for nuclear medicine and the American Neurogastroenterology and Motility
Society. Patients were scanned under a dual-head gamma camera set at a 99 mTc window
(140 keV ± 15 %) in the supine position to visualize the whole stomach; simultaneous
anterior and posterior radioactivity measurements were obtained. After ingesting the
radiolabeled test meal, a static image of 60-second duration was done immediately
post-meal, followed by static serial pictures at 30, 60, 120, and 180 minutes and
a half emptying time was measured.
Outcomes
Weight loss, expressed as TBWL and EWL, among the three groups at 12 months of follow-up
was the primary outcome. Secondary outcomes included the impact of the different suture
patterns on gastric emptying scintigraphy and satiety test, the impact of the remaining
number of sutures at the end of the follow-up on weight loss and assessment of nausea,
vomiting, and abdominal pain at 24 hours post-procedure using the 0–10 visual analog
scale. As an exploratory outcome of our study, we evaluated potential correlations
between the number of remaining sutures during the one year of the follow-up after
the E-ESG. To do that, the number of sutures were counted during the follow-up gastroscopies
(months 1, 6, and 12 post-procedural). Taking into account the fact that the three
suture patterns necessitated a different number of sutures, any change in the number
of remaining sutures was expressed as a percentage of the difference (Δ) between the
number of the initially applied sutures and those remaining in place at each timepoint
([initial number of sutures-number of remaining sutures]/initial number of sutures).
Statistical analyses
Due the absence of existing studies which report the magnitude of effect of the type
we aimed to investigate, we decided to proceed with a 1:1:1 randomization on a total
of 48 patients (16 in each group). Descriptive statistics were computed for all study
variables. Discrete variables were expressed as count (percentage) and continuous
variables as mean ± standard deviation. Intention to treat is considered the principal
approach for all analyses [13]
[14]. For EWL and TBWL, the treatment effect has been assessed using linear mixed models
with time as a repeated factor. Time, group, and interaction were considered fixed
effects in the model. The model accounted for participant correlations over time and
between groups. The fixed effects estimates are reported as means with 95 % confidence
intervals (CIs). To adjust for multiple comparisons, Bonferroni correction method
was used. Multiple imputation was used to investigate the sensitivity of the results
of the primary analysis with regard to the missing data. The mixed model was then
refitted, and the obtained results were compared with those from the previous results.
The association between variables was set by estimating the correlation between them.
All statistical tests were two-tailed and P < 0.05 was considered statistically significant. Data were analyzed using IBM SPSS
Statistics for Windows v27 (Armonk, New York, United States) and R software (R Foundation
for Statistical Computing, Vienna, Austria).
Results
Patients
We randomly allocated 48 patients (mean age 41.9 ± 9.5 years; 40 female [83.3 %],
BMI 33.8 ± 2.7 kg/m2) to the three index study arms from April 2019 to November 2019. [Fig. 2] shows the study flowchart. Baseline BMI was similar for the three groups (34.65 ± 2.55,
33.8 ± 2.69, and 33.34 ± 2.51 for groups A, B, and C, respectively). The detailed
baseline demographic, anthropometric, and laboratory parameters of the enrolled patients
per group are shown in [Table 1].
Fig. 2 Study flowchart.
Table 1
Baseline demographic, anthropometric, and laboratory parameters of enrolled patients.
|
Group A
|
Group B
|
Group C
|
|
Sex, n (%)
|
|
|
14 (87.5)
|
13 (81.3)
|
13 (81.3)
|
|
|
2 (12.5)
|
3 (18.7)
|
3 (18.7)
|
|
Smoking status, n (%)
|
|
|
8 (50)
|
13 (81.3)
|
12 (75)
|
|
|
2 (12.5)
|
1 (6.2)
|
2 (12.5)
|
|
|
6 (37.5)
|
2 (12.5)
|
2 (12.5)
|
|
Alcohol, n (%)
|
|
|
8 (50)
|
12 (75)
|
10 (62.5)
|
|
|
0 (0)
|
1 (6.2)
|
0 (0)
|
|
|
8 (50)
|
3 (18.8)
|
6 (37.5)
|
|
Age (years), mean (SD)
|
42.81 (7.95)
|
36.75 (6.78)
|
44.25 (11.2)
|
|
Weight (kg), mean (SD)
|
93.81 (10.24)
|
91.79 (7.84)
|
90.23 (11.67)
|
|
Height (cm), mean (SD)
|
165 (7)
|
165 (7)
|
164 (8)
|
|
BMI (kg/m2), mean (SD)
|
34.65 (2.55)
|
33.8 (2.69)
|
33.34 (2.51)
|
|
Fat Mass
|
40.9 (4.76)
|
39.23 (6.3)
|
41.04 (19.84)
|
|
Glucose (mg/dL), mean (SD)
|
97.6 (14.9)
|
92.46 (12)
|
94.31 (9.68)
|
|
Total cholesterol (mg/dL), mean (SD)
|
185.69 (50.03)
|
172.64 (44.34)
|
212.25 (48.95)
|
|
HDL cholesterol (mg/dL), mean (SD)
|
51.54 (12.19)
|
57.07 (16.28)
|
57.44 (12.75)
|
|
Triglycerides (mg/dL), mean (SD)
|
146.93 (105.92)
|
96.64 (41.36)
|
127.06 (71.66)
|
|
AST
|
19.4 (11.59)
|
20.13 (5.63)
|
19.13 (6.28)
|
|
ALT
|
22.07 (13.37)
|
19.93 (7.18)
|
22.63 (12.8)
|
|
gGT
|
24.07 (16.42)
|
20.53 (13.48)
|
24.06 (16.12)
|
|
ALP
|
76.67 (28.24)
|
79.25 (10.72)
|
46 (21.31)
|
BMI, body mass index; AST, aspartate transaminase; ALT, alanine aminotransferase;
ALP, alkaline phosphatase; gGT, gamma glutamyl transferase; HDL, high density lipoprotein.
Procedure details
Overall, the mean duration of the procedure was 60.67 ± 19.04 minutes. Procedural
time in Group B (72.94 ± 19.58) tended to be longer compared to Group A (55.13 ± 16.49)
and C (53.94 ± 15.45), but this difference did not reach the level of statistical
significance (P = 0.06). As expected per study design, fewer sutures were used in Group A (3.88 ±
0.62) and C (4.5 ± 0.73) compared to Group B (5.5 ± 0.89; P < 0.001). No severe adverse events related to the E-ESG were reported during the
study period.
Primary outcome
[Fig. 3a] depicts the TBWL (95 %CI) and EWL (95 %CI) per suture pattern group during the 12
months of the study. After 12 months of follow-up, the mean TBWL in all groups was
10.11 % (7.1–13.12), over the recommended 5 %, compared to baseline weight [2]; however, no statistical difference among the different groups was detected. TBWL
was 9.13 % (2.15–16.11) in Group A, 11.29 % (5.79–16.80) in Group B, and 9.96 % (4.58–15.35)
in Group C (P = 0.589). At the end of the follow-up (12 months), the mean EWL exceeded the recommended
25 % (42.56 % (28.23–56.9) compared to the baseline weight in the entire cohort ([Fig. 3b]). Once again, there were no statistically significant differences among the three
groups [34.54 % (6.09–62.99) vs. 44.75 % (23.63–65.88) vs. 46.94 % (16.72–77.15) for
groups A, B, and C, respectively; P = 0.888]. [Table 2] shows the per group TBWL and EWL for the different study time points. The lack of
difference between the three groups was confirmed during the multiple imputation analysis
(Supplementary Table 2); TBWL (95 %CI) was 10.43 % (6.01–14.85) in Group A, 11.5 % (7.79–15.20) in Group
B, and 9.97 % (5.71–14.23) in Group C (P = 0.497) and EWL (95 % CI) was 37.40 % (19.70–55.11) vs. 47.76 % (30.25–65.27) vs.
45.61 % (22.51–68.70) for groups A, B, and C (P = 0.970), respectively, at the end of the 12-month follow-up.
Fig. 3 Percentage of total body weight loss (TBWL%) and excess weight loss (EWL%) according
to suture pattern during follow-up.
Table 2
Total body weight loss and excess weight loss over time.
|
TBWL (%), mean (95 %CI)
|
P value
|
EWL, % (95 %CI)
|
P value
|
|
Group A
(n = 16)
|
Group B
(n = 16)
|
Group C
(n = 16)
|
Group A
(n = 16)
|
Group B
(n = 16)
|
Group C
(n = 16)
|
|
Month 1
|
(n = 15)
5.24
(4.04–6.44)
|
(n = 15)
5.89
(4.50–7.28)
|
(n = 14)
5.74
(4.20–7.29)
|
0.589[*]
|
(n = 15)1
9.46
(14.58–24.34)
|
(n = 15)
23.89
(16.92–30.86)
|
(n = 14)
24.55
(15.81–33.29)
|
0.888[*]
|
|
Month 3
|
(n = 13)
10.35
(8.21–12.48)
|
(n = 13)
9.70
(7.33–12.07)
|
(n = 15)
8.92
(5.67–12.16)
|
(n = 13)
37.54
(29.77–45.32)
|
(n = 13)
39.93
(30.97–48.89)
|
(n = 15)
41.91
(24.12–59.69)
|
|
Month 6
|
(n = 11)
11.93
(7.67–16.19)
|
(n = 13)
10.75
(6.96–14.55)
|
(n = 14)
10.57
(6.27–14.86)
|
(n = 11)
45.15
(27.68–62.62)
|
(n = 13)
41.96
(27.37–56.56)
|
(n = 14)
49.60
(26.12–73.08)
|
|
Month 9
|
(n = 11)
12.55
(7.62–17.48)
|
(n = 13)
9.81
(5.28–14.35)
|
(n = 15)
10.28
(6.08–14.49)
|
(n = 11)
47.55
(26.11–68.98)
|
(n = 11)
37.75
(19.78–55.72)
|
(n = 15)
47.60
(23.90–71.31)
|
|
Month 12
|
(n = 11)
9.13
(2.15–16.11)
|
(n = 13)
11.29
(5.79–16.80)
|
(n = 15)
9.96
(4.58–15.35)
|
(n = 11)
34.54
(6.09–62.99)
|
(n = 13)
44.75
(23.63–65.88)
|
(n = 15)
46.94
(16.72–77.15)
|
TBWL, total body weight loss; EWL, excess weight loss; n: available data in that time;
data are analyzed as ITT using mixed models.
* Comparison among the three groups at month 12.
Secondary outcomes
Gastric emptying scintigraphy
At the end of the follow-up, 25 patients performed the control gastric emptying scintigraphy
(6, 8, and 11 for groups A, B, and C, respectively). At baseline, the half emptying
time was 57.5 min (43.64–71.36) for Group A (n = 16), 59.64 min (46.91–72.37) for
Group B (n = 16), and 68.07 min (59.86–76.27) for Group C (n = 16) with no differences
between the three groups. There was no statistically significant difference regarding
the half emptying time either per group compared to baseline or among the three different
suture patterns at the end of follow-up (54.00 min [32.86–75.14], 68.38 min [48.25–88.50],
and 69.45 min [55.05–83.86]) for groups A, B, and C, respectively; P = 0.577. [Table 3] presents the results of gastric emptying scintigraphy for the three different time
points (baseline, month 6, month 12).
Table 3
Results of gastric emptying scintigraphy over time.
|
Half time emptying (min), mean (95 % CI)
|
|
Group A
|
P value
|
Group B
|
P value
|
Group C
|
P value
|
P value
|
|
Baseline
|
(n = 16)
57.50
(43.64–71.36)
|
0.999[1]
|
(n = 16)
59.64
(46.91–72.37)
|
0.945[1]
|
(n = 16)
68.07
(59.86–76.27)
|
0.999[1]
|
0.577[2]
|
|
Month 6
|
(n = 8)
57.50
(40.48–74.52)
|
(n = 9)
65.56
(45.29–85.82)
|
(n = 8)
61.25
(45.51–76.99)
|
|
Month 12
|
(n = 6)
54.00
(32.86–75.14)
|
(n = 8)
68.38
(48.25–88.50)
|
(n = 11)
69.45
(55.05–83.86)
|
n, available data in that time; data are analyzed as ITT using mixed models.
CI, confidence interval; ITT, intention to treat.
1 Per group comparison between baseline and month 12.
2 Time-group interaction for the three groups at month 12.
Drinking satiety test
At baseline, the mean ingested volume was 375.63 mL in Group A (n = 16), 444.38 mL
in Group B (n = 16), and 513.13 mL in Group C (n = 16). Overall, 12 months after the
E-ESG, the ingested volume to reach the test’s endpoint dropped in all groups [268.57 mL
in Group A (n = 7), 231.43 mL in Group B (n = 7), and 401 mL in Group C (n = 10)].
No statistically significant difference was evident between the three groups (P = 0.875). [Table 4] shows the results of the drinking satiety test for the three different time points
(baseline and months 6 and 12). The feeling of satiety using a 0–10 visual scale was
8 (5–9), 7 (7–8), and 8 (5–8) for groups A, B, and C, respectively and was less intense
in all three groups (4.5 [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8], 6 [2]
[3]
[4]
[5]
[6]
[7]
[8], and 6.5 [2.5–8.5]) for groups A, B, and C, respectively) at the end of follow-up.
There was no difference among the three groups regarding the satiety test at the end
of follow-up (P = 0.933).
Table 4
Results of drinking satiety test over time.
|
Volume to reach endpoint[1] (mL), mean (95 %CI)
|
|
Group A
|
P value
|
Group B
|
P value
|
Group C
|
P value
|
P value
|
|
Baseline
|
(n = 16)
375.63
(299.69–451.56)
|
0.330[2]
|
(n = 16)
444.38
(303.41–585.34)
|
0.077[2]
|
(n = 16)
513.13
(368.10–658.15)
|
0.214[2]
|
0.875[3]
|
|
Month 6
|
(n = 8)
221.25
(138.68–303.82)
|
(n = 9)
253.33
(135.69–370.98)
|
(n = 10)
297.00
(160.72–433.28)
|
|
Month 12
|
(n = 7)
268.57
(130.11–407.03)
|
(n = 7)
231.43
(162.13–300.73)
|
(n = 10)
401.00
(34.78–767.22)
|
n, available data in that time; data are analyzed as ITT using mixed model.
CI, confidence interval; ITT, intention to treat.
1 Hunger, fullness, nausea, bloating, and pain according to protocol.
2 Per group comparison between baseline and month 12.
3 Time-group interaction for the three groups at month 12.
Correlation between the percentage of sutures remaining in place with TBWL and EWL
The mean difference in number of sutures between those initially applied and those
remaining in place 1 year later was 30.21 % (7.27–53.15) for Group A, 17.99 % (3.46–32.52)
for Group B, and 17.42 % (6.99–27.86) for Group C (P = 0.001).
When taking into account the observations from all follow-up endoscopies, a statistically
significant positive correlation between the number of remaining sutures and TBWL
was observed (correlation coefficient [95 %CI] 0.37 [0.18–0.53], P < 0.001). When looking separately at the three suture patterns, this correlation
remained statistically significant for groups B and C (coefficient [95 %CI] 0.55 [0.23–0.76],
P = 0.001 and 0.51 [0.19–0.73], P = 0.002, respectively), but not for Group A (coefficient [95 %CI] 0.14 [-0.24–0.48],
P = 0.480). Similar observations were made regarding EWL both for the entire cohort
(correlation coefficient ([95 %CI] 0.33 [0.14–0.50], P = 0.001), as well as per randomization group (correlation coefficient [95 %CI] 0.11
[–27–0.46], P = 0.568 for Group A, 0.54 [0.22–0.75], P = 0.001 for Group B, and 0.43 [0.09–0.68], P = 0.011 for Group C).
Adverse events
[Table 5] shows evaluation of nausea, vomiting, and abdominal pain (cramps) at 24 hours post-procedure
using the 0–10 visual analog scale. There was no difference among the three groups
for all of these symptoms. Even though abdominal cramps were numerically less intense
in Group C (mean score 3.63 ± 2.53) compared to Group A (4.40 ± 2.59) and Group B
(4.25 ± 2.67), this difference did not reach statistical significance (P = 0.617).
Table 5
Adverse events assessed at 24 hours post-procedure using a 0–10 visual analogue scale.
|
Group A
|
Group B
|
Group C
|
P value
|
|
Nausea, mean (SD)
|
1.27 (1.87)
|
2.25 (2.84)
|
1.88 (2.31)
|
0.690
|
|
Vomiting, mean (SD)
|
1.27 (3.03)
|
1.94 (3.36)
|
0.94 (2.57)
|
0.483
|
|
Cramps, mean (SD)
|
4.40 (2.59)
|
4.25 (2.67)
|
3.63 (2.53)
|
0.617
|
SD, standard deviation.
Discussion
Endoscopic sleeve gastroplasty using Endomina (E-ESG) previously has been demonstrated
to be effective for weight loss when compared to only lifestyle interventions [15]. In the present randomized study, none of the three different suture patterns tested
proved to be superior compared to the others. Still, all three of them led to a ≥ 10 %
TBWL and ≥ 25 % EWL at the end of the 12-month follow-up, achieving the target for
an endoscopic bariatric treatment to be considered efficacious [2].
Currently, different techniques for performing an endoscopic sleeve gastroplasty are
available, with all of them showing promising results for the treatment of obesity.
However, the exact mechanism through which the goal of weight loss is achieved remains
unclear. Volume reduction leads to earlier postprandial satiety and lower calorie
intake, and a decrease in the rate of gastric emptying has also been proposed to be
important.
Although various studies in the literature have tested different suture patterns for
ESG, almost exclusively using the Overstitch system [16]
[17], the major aim of these studies was not to reveal potential pathophysiologic mechanisms
regarding beneficial effects of ESG, but rather, to evaluate different patterns in
terms of feasibility, ease of application, and cost-effectiveness.
Espinet-Coll et al. [18] performed a retrospective study examining three different suture patterns; a transverse
bilinear design starting from the anterior wall then moving to the greater curvature,
to the posterior wall, and then repeated vice versa, a longitudinal pattern with two
rows of parallel sutures along the gastric body, and a transverse monolinear pattern
starting from the anterior wall then moving to the greater curvature and finishing
at the posterior wall. All three patterns were performed using the first generation
of the Overstitch and reflect the evolution of the technique at a single center. All
three designs aimed to achieve a “classic” endoscopic gastroplasty by suturing the
gastric body and reducing the gastric volume.
In contrast, we aimed to create three different gastroplasty types: increased fundus
distensibility, decreased volume, and delayed gastric emptying. This study demonstrated
that patients in all three groups achieved sufficient weight loss with no differences
in the number of the sutures or stitches applied. However, in both studies, no difference
was detected between the evaluated patterns regarding weight loss, meaning that regardless
of the suture pattern used, a similar weight loss was achieved.
Although gastric emptying recently has been associated with the sensation of appetite
in obese patients [19], our results do not confirm this observation. Interestingly, our study did not demonstrate
any difference in gastric emptying. The gastric emptying scintigraphy time did not
differ among the three groups at the end of follow-up. In comparison, there was no
statistically significant difference in scintigraphy between baseline and the end
of follow-up for all three groups. On the other hand, the feeling of satiety was achieved
more quickly at the end of follow-up compared to baseline for groups A and B, but
not for Group C. It is also worth mentioning that the volume needed to be ingested
to achieve satiety did not differ among the three groups at the end of follow-up.
A trend to a lower required volume was detected in Group B between baseline and the
end of follow-up. While acknowledging that our study was not powered to detect these
differences, we should consider that endoscopic gastroplasty may contribute weight
loss in a multifactorial way, beyond volume reduction and delayed gastric emptying.
From a technical point of view, the ease of performing each of the suture patterns
was not assessed in this study. However, all endoscopists reported Group A to be the
most challenging to achieve. This is confirmed because the mean procedural time was
similar for all three groups, while Group A was the one with the smallest number of
sutures used. Moreover, Group A was associated with a higher percentage of displaced
sutures at the end of follow-up, probably due to higher tension between the sutures
applied in the different parts of the gastric wall in this pattern trying to mimic
a gastric band. Even if dislodgement of some of the initially placed sutures did not
lead to loss of efficacy, as shown in our exploratory analysis, we would tend to avoid
pattern A because it was technically more cumbersome.
This is the first RCT evaluating different suture patterns in patients undergoing
ESG. Despite great interest in this topic [5], existing evidence derives either from simple cohort studies or from retrospective
studies evaluating a single center's experience with different suture patterns throughout
the years [16]. While RCTs on ESG are scarce, this is the second RCT evaluating the efficacy of
Endomina for obesity, confirming the previously published results, showing that E-ESG
can lead to a significant weight loss among obese patients [15]. Another advantage of our study is that all patients underwent a detailed pre-intervention
structured bariatric workup guided by a multidisciplinary team. Beyond the standard
workup, all participants underwent gastric scintigraphy and a satiety test, allowing
us to assess the potential impact of our intervention in terms of gastric emptying
and gastric volume.
Our work is not without limitations. Most importantly, this study was not powered
to detect any difference regarding secondary outcomes (impact of suture patterns on
gastric emptying and gastric volume). Thus, even in the absence of statistical significance,
these results should be interpreted cautiously. Beyond that, the unprecedented COVID-19
pandemic inevitably affected our academic projects because participants hesitated
to visit the hospital for non-urgent issues or planned follow-up appointments were
canceled during the first pandemic wave.
Consequently, and despite all our efforts, in terms of the primary outcome (EWL and
TBWL) nine patients were lost during follow-up and were not evaluated at month 12.
To overcome this drawback, we performed an imputation analysis, taking into account
the available data for every patient until the last follow-up, confirming our per-protocol
analysis results. Finally, this study did not aim to evaluate changes in patient comorbidities
(hypertension, diabetes) before and after treatment application. However, we acknowledge
that this could be part of the design of a future study, which would also be warranted
to elucidate potential pathophysiological mechanisms contributing to weight loss following
endoscopic gastroplasty.
Conclusions
To conclude, our study showed that different suture patterns effectively achieved
significant weight loss among obese patients. On the other hand, the differences regarding
gastric scintigraphy results and volume needed to achieve satiety were non-significant
among the three different patterns at the end of follow-up.
