Introduction
It has long been thought that endoscopic approximation of tissues with sutures could
be very useful in minimally invasive treatments, especially for obesity. Despite some
advances in technology and improved maneuvering capability, the endoscopic suture
procedure is still complicated by failures in the postoperative course. In general,
endoscopic sutures are not considered reliably permanent [1].
A study conducted by Hashiba et al showed that for an effective suture between two
gastric walls, the point of contact should be an area of submucosa or muscular propria
(without mucosa) measuring ≥ 4 cm2 that is approximated by at least three sutures [2]. This model allows formation of a tunnel in the gastric lumen and serves as a basis
for this study that can also be applied to obesity treatment. The aim of the study
was to evaluate the feasibility, safety, and efficacy of the basic principles of conventional
surgery applied to endoscopic suturing in animal models.
Material and methods
Animal model
Survival experiments were performed on six live pigs (MinipigBR, Brazil, body weights
35 kg to 40 kg). Intramuscular midazolam (0.2 mg/kg) was used to induce general anesthesia,
maintained by using isoflurane and propofol. This procedure was performed by a veterinarian
and the animals were housed at an animal research facility. During the procedure,
cefazolin sodium hydrate (2 g) was administered and repeated for the first 2 days.
Analgesics were used in the postoperative course.
The animals were on a liquid diet for 2 weeks after the procedure, after which a normal
diet was resumed. The pigs were followed for either 4 or 8 weeks. The study was approved
by the Sirio-Libanes Care Institutional Review Board (approval no CEUA-P 2016-17).
Basic materials and preparation
All procedures were performed by a skilled endoscopist with the pig in the supine
position. A standard gastrointestinal endoscope (Silver Line Karl Storz Video endoscope,
Tuttlingen, Germany) was used. The endoscope was introduced into the gastric lumen
and two wide areas of the gastric body were selected, one in the anterior wall and
another one in the posterior wall. The selected areas were marked with the tip of
a polypectomy snare (Captivator, Boston Scientific, Natick, Massachusetts, United
States). The proximal limit of these areas was approximately 2.0 cm distal to the
cardia level in the anterior and posterior wall; the distal limit was the gastric
incisura.
Endoscopic mucosal resection
Exposition of the submucosa/muscular propria was done using the snare placed around
a transparent cap used for variceal band ligation. When the cap suctioned the gastric
wall to the maximum, the snare was released and subsequently closed, grasping the
aspirated tissue. The suction was interrupted with the snare still grasping the tissue
outside of the cap. A pulling maneuver was applied to avoid the deep layers of the
gastric wall, followed by resection using electrocoagulation ([Fig. 1]). No fluid was injected into the submucosa. The procedure was repeated many times
in the demarcated area, until total exposition of the submucosa/muscular propria layers
in the previously marked area.
Fig. 1 Gastric wall after EMR.
Novel suture device
A novel device termed DASE was used (prototype, G-Flex, Sao Paulo, Brazil), which
comprised a plastic chamber for tissue aspiration and T-tag placement (Tissue Apposition
System (TAS), Ethicon EndoSurgery, Cincinnati, Ohio, United States). The device is
a 3-cm-long, transparent plastic tube that is closed at the distal end by a transparent
wall ([Fig. 2]).
Fig. 2 DASE chamber with the needle appearing in the window.
Laterally, the DASE has a window that is 1.0 × 1.2 cm wide. The endoscope was inserted
into the proximal end. The chamber distal wall has a thin channel, which allows for
insertion of the metallic needle that carries the T-tag suture. The T-tag is a device
consisting of a small magnetic resonancing imaging-compatible steel tag that is loaded
onto a needle.
Suture and plication
The sutures were placed approximating the anterior and posterior wall. The procedure
began after EMR and the steps performed were as follows. The gastric wall was aspirated
through the DASE window until it no longer penetrates into the chamber ([Fig. 3]). Once the T-tag was loaded inside of the needle, the needle was inserted through
the captured tissue. The needle was pushed through the aspirated tissue and the T-tag
was released by the needle stylet ([Fig. 4]). A small amount of traction then was applied to the surgical stitch while the injected
needle was withdrawn, which caused the T-tag to be deployed from the needle and assume
a T-shaped position perpendicular to the thread direction, anchoring it in situ ([Fig. 5]). This procedure was repeated in the posterior wall. At that time, both gastric
walls without mucosa (anterior and posterior) were crossed with a T-tag surgical thread
at the distal tip. The two proximal tips of the T-tag thread were exteriorized through
the mouth. Finally, the two threads were cinched together with a tie-knot that completed
the suture ([Fig. 6]). The thread above the tie-knot was cut with scissors. Plication consisted of a
group of three aligned sutures.
Fig. 3 View of the gastric wall through the DASE chamber window.
Fig. 4 Gastric wall without mucosa inside the DASE chamber.
Fig. 5 Illustration of the gastric walls showing the sutures and plications.
Fig. 6 Joining of the anterior and posterior gastric walls with a suture cinched by a tie-knot.
Procedure completion
Each plication was made with three sutures. Three plications were used to make a pig
model gastroplasty. In each plication, the first step was insertion of the T-tag through
the tissue in both walls, as previously described.
It is worth mentioning that in the central suture of each plication, between the gastric
wall and the T-tag, the two T-tags were approximated using a metallic clip and subsequently
the tie-knot was used to close the suture, as we previously described. The two remaining
sutures of the plication were cinched straight together by pulling the surgical threads
and placing the tie-knot (prototype, Cook Endoscopy, Winston-Salem, NC, USA) that
completed the suture, followed by cutting the thread with scissors ([Fig. 7a], [Fig. 7b], and [Fig. 7c]).
Fig. 7 a Sutures with T-tag caught by a metallic clip. b The sutures are pulled through the tie-knot and the gastric walls are joined together.
c Three sutures together forming one plication.
Two plications were performed in two animals, the first two in this series, and the
remaining four animals received three plications.
Results
Plication placement was successful in all animals ([Fig. 8]). In the second animal in this series, one plication failed 4 weeks later. The remaining
15 plications were stable 4 to 8 weeks later. One perforation occurred during the
EMR and it was closed with one clip. A Veress needle was used to treat the pneumoritoneum.
The procedure was conducted normally after the event. There were no other early or
late adverse events. All animals were submitted to an esophagogastroduodenoscopy (EGD)
and sacrificed.
Fig. 8 The wall created by the plication between: a a tunnel on the lesser curvature side and b the tunnel of the greater curvature.
Four animals were sacrificed within 4 weeks and two animals were followed up for another
4 weeks (8 weeks total). On EGD, the plications remained stable, but the wall retraction
increased, mainly on the side of the greater curvature. The spaces between the plication
were closed with scar tissue, probably due to tissue retraction.
Externally, the stomach showed some deformity with three retraction areas, but the
organ shape was not significantly altered ([Fig. 9]). Necropsy showed only adhesions between the great omentum and the stomach, proving
that the adjacent organs were not affected by the sutures. There were signs of retraction,
even where the serosa was not involved. Histology showed two muscularis propria layers
joined by dense fibrotic tissue ([Fig. 10]).
Fig. 9 a, b Anterior and posterior views of the stomach 3 weeks after the suture. Many retracted
areas were created by the procedure.
Fig. 10 a, b Histologic images of a transverse section of the suture. The muscularis propria layers
are separated by the scar (dense fibrotic tissue).
The present study included 16 plications performed in six animals. The number of plications
per animal ranged from three (n = 4) to two (n = 2). Only one plication failed after
4 weeks.
The hypothetical likelihood of procedure failure or success was 50 % but the observed
incidence of individual suture failure was 6.25 % (n = 16). This was a very significant
difference (P < 0.001, Fisher exact test). Thus, the total incidence of suture failure after 4
weeks was 6.25 %.
Discussion
Development of an effective, fast, and cost-efficient method of endoscopic gastrointestinal
wall apposition and permanent fusion has long been desired. Endoscopic suturing is
a surgical procedure and should be performed based on basic scientific principles
[3].
To ensure the success of suturing, important details should be considered, such as
the area, depth, and conditions of the sutured tissue [1]
[4]. Furthermore, the healing process has not changed since Lembert stated in 1826 that
suturing the mucosal surfaces on both sides does not lead to fusion of the walls [4], as this does not trigger a strong healing process, which only occurs if the submucosa
is extensively involved [5]
[6]. Gross and Halsted’s seminal studies at the end of the 19th century also demonstrated the importance of submucosal involvement [5]
[6]. These and several other studies [7]
[8] have demonstrated that to achieve fusion of the gastrointestinal walls, trauma is
required because it triggers an organized and complex cascade of cellular and biochemical
events, starting with hemostasis and an inflammatory phase, which is characterized
by increased vascular permeability, chemotaxis of cells from the circulation into
the wound, local release of cytokines and growth factor, activation of migrating cells
like macrophages and wound fibroblasts, and a phenotypic change in the dermal fibroblast
with increased collagen synthesis capacity. The proliferative phase, in which fibroblasts
and endothelial cells proliferate, is followed by the maturation phase, which starts
1 week after the injury [7]. This provides necessary resistance to tissue fusion [8]. Sutures involving the deep muscular propria layer would guarantee the initial support
required for all these steps to occur.
The endoscopic suturing currently being performed has not respected the traditional
rules created by surgeons who extensively studied anastomoses [3], given that suturing mucosa to mucosa would explain the findings of suture failure
are 1-year follow-up that have been reported in studies [9]. This means that approximation of tissues without prior trauma does not trigger
a sufficient inflammatory process and, therefore, will not lead to effective tissue
fusion [7]
[8].
The present study demonstrated postoperative fusion of anterior and posterior gastric
walls with dense fibrosis between the muscularis propria layers of the stomach and
showed that the suture material remained in place. These results, verified by endoscopy
and histology, suggest that separation of the sutured tissues can only be done surgically.
Based on established surgical practice as well as studies by Hashiba et al (2), certain
rules must be applied if a permanent suture is desired. The apposition must be in
an area of the submucosa/muscularis propria that measures ≥ 4 cm2, two to five sutures made of non-absorbable material should be used, and the muscularis
propria layer must be incorporated into all sutures. Despite the involvement of the
serosa, detected in the laparotomy post-mortem, the study showed it is not important
to plicate the serosa during suture placement [2].
From a technical point of view, the technique used in the present study is particularly
relevant as it was performed using a conventional gastrointestinal endoscope.
As the aspirated wall lacked mucosa because of the previous EMR, it is reasonable
to assume that the muscularis layer was strongly represented ([Fig. 6]). The author’s previous study, described above, showed that the muscularis propria
layer is always included [2]. When the serosa was reached, no more than 2.3 mm of the suture was exposed. Even
pathological evaluation of the stomach after euthanasia did not allow definitive conclusions,
although the serosa often was seen or encompassed in the suture or the inflammatory
process.
The maintenance of intact plication in 15 of 16 attempts at 4 and 8 weeks was highly
significant. Distention of the stomach with solid foods also should be avoided in
the early postoperative period. As this method is developed, the length of time during
which solid folds are restricted may be reduced, as has been the case with conventional
open surgery.
Some improvements can be made in the devices used in this study to make the procedure
more practical and less time-consuming. Development of the tie-knot to allow for automatic
suture cutting could eliminate the need for use of scissors.
Conclusions
In conclusion, this study showed that endoscopists performing permanent endolminal
suturing need to follow some fundamentals. Our results also suggest that endoscopic
sutures can be used in different ways, such in situations involving large fistulas
in the gastrointestinal tract, for gastroesophageal reflux therapy, and for gastroplasty
and gastric restriction in treatment of obesity. Although novel and effective, the
suturing method as a whole may not be as important as understanding that exposing
the submucosa with EMR and ensuring that suturing is effective by providing adequate
support for the wound healing cascade are fundamental steps for successful tissue
fusion.
