Introduction
Clinical management of upper gastrointestinal tract leakage with septic sequelae has
markedly improved since the introduction of endoluminal sealing techniques. This has
mainly been accomplished either with covered stents or endoscopic vacuum therapy [1]
[2]
[3]. However, both techniques are hampered by inherent shortcomings. SEMS are burdened
by a migration rate of 50 % or more, and because they exert a radial force on the
intestinal wall, they impair microperfusion and wound healing at the leakage site
[4]
[5]. Endoluminal vacuum therapy improves wound healing and drains inflammatory secretions
via the suction-induced negative pressure [6]
[7]. The latter, on the other hand, induces positive back-pressure of the same magnitude
in the opposite direction (Newtonʼs third law of action and reaction), which puts
a strain on the intestinal wall and occludes the lumen. This then might induce continuous
microaspiration. Upper gastrointestinal endoluminal EVT therefore mandates additional
feeding-tube running alongside or through the sponge.
The new VACStent concept combines the benefits of both techniques, sealing the intestinal
wound with a negative-pressure sponge and ensuring patent intestinal passage for fluids
and nutrition. This is accomplished by a fully covered SEMS with a sponge-cylinder
affixed to its exterior, which in turn is connected by catheter tubing to a vacuum
pump. Three initial clinical cases are presented here as examples for the proposed
application of this new endoscopic device.
Patients and methods
The VACStent comprises a self-expanding nitinol stent covered with a silicone-membrane
impermeable to liquid and gas (VacStent Medtech AG, Steinhausen, Switzerland). Affixed
to the exterior of the stent is a polyurethane sponge cylinder (thickness 10 mm) connected
to an external vacuum pump via a fixed small gauge (12 F) catheter ([Fig. 1a], [Fig. 1b]). There are multiple holes at the distal end of the catheter over the length of
the sponge cylinder. The lumen of the 50 mm long sponge cylinder has an inside diameter
of 12 mm and is framed by the flanged ends of the 30-mm diameter stent. Once negative
pressure is applied, only the flanged ends of the stent are in contact with the intestinal
wall, sealing the sponge cylinder against the intestinal fluids. This constant suction
induces a force of the same magnitude – but in opposite direction – as the applied
negative pressure, thereby immobilizing the VACStent on the intestinal wall and preventing
stent-migration. The VACStent is loaded on a flexible introducer system (12-mm diameter),
which is inserted transorally in over-the-wire technique. Positioning and release
is monitored either by fluoroscopy or more simply through a small endoscope paralleling
the inserted introducer system. If the positioning is controlled by fluoroscopy, the
unfolding of the stent can easily be followed on screen. If controlled by direct vision
using a small endoscope, leak position is determined by using the measured distance
from the leak to the front teeth and visualizing proximal deployment of the sponge.
Fig. 1 a VAC Stent: a silicone-coated Nitiolstent covered by a PU sponge cylinder at the outside
connected by a tube to the vacuum pump. b Components of the VAC Stent: a silicone-coated Nitiolstent, a Polyurethan Sponge
cylinder and a suction tube connected to vacuum pump. c After release the nitinol filaments unfold the VACStent to its original shape.
After release, the nitinol filaments unfold the VACStent to its original shape ([Fig. 1c]). Because of the open-pored sponge, a vacuum is created all over the contact plane
between the sponge cylinder and the esophageal wall. This means that the suction functions
independently from the position of the catheter. There are markers at the ends of
the VACStent that facilitate orientation, if applied fluoroscopy.
If the VACStent is deployed too distally, it is easy to reposition the device more
proximally by grasping the circular wire mounted at the proximal bulged end with a
forceps and just pull, because this reduces the stent diameter. The same procedure
is used to extract the device. The repositioned transnasal suction catheter is connected
to a vacuum pump. The applied negative pressure ranges from –80 mm Hg to –120 mmHg.
However, continuous suction of secretions and VACStent fixation on the intestinal
wall is also possible with vacuum pressure down to –50 mmHg [3]
[8].
This study was a retrospective analysis of prospectively collected data from patients
selected for an applicability trial of the new VACStent system [9]. The study protocol was approved of the Institutional Review Board of the University
Witten/Herdecke. Three patients were selected to study in whom the postulated benefits
of this hybrid sponge-stent system were realized in clinical practice. While the main
focus was on the applicability and migration resistance of the VACStent, the study
also investigated sealing of different leak configurations and unrestrained passage
of swallowed liquids through the stent.
Results
The first patient had undergone subtotal esophagectomy with esophago-jejunal circular
stapler anastomosis for adenocarcinoma. On postoperative day 14, endoscopy confirmed
anastomotic leakage with a small abscess cavity ([Fig. 2a]). A VACStent was inserted, which sealed off the leak as demonstrated by contrast
fluoroscopy. Suction with a continuous negative pressure of –120 mmHg was instituted.
Control thoracic computed tomography on day 4 showed the exact VACStent position with
no mediastinal collection ([Fig. 2b]). On day 5, the VACStent was removed and the area treated was examined. Good induction
of granulation tissue was seen, indicating active wound healing ([Fig. 2c]). A second VACStent was inserted and the negative pressure lowered to –80 mmHg.
After another 7 days of continuous suction, the VACStent was removed ([Fig. 2d]). The leakage site was clean and only viable granulation tissue was evident, leaving
a small cavity of 3 to 4 mm in depth, which was completely sealed off as demonstrated
by contrast X-ray. Follow-up endoscopy 2 and 4 weeks later showed no evidence of stenosis.
Fig. 2 a A small abscess at the circular stapler line 14 days after subtotal oesophagectomy
with intrathoracic anastomosis. b Thorax CT shows exact location of the VAC Stent and no signs of mediastinitis or
abscess formation. c The leak at the staple line cleaned up 5 days later. d Good granulation tissue is shown with a shallow cavity completely covered by granulation
tissue.
The second patient, a 66-year-old woman with acute Boerhaave syndrome (esophageal
tear from 29 cm to 36 cm), was treated with a covered SEMS. She demonstrated signs
of sepsis, most probably due to mediastinitis, but without a distinct fluid collection
suitable for percutaneous drainage. In addition, a tube was inserted into her right
thorax. After 14 days the SEMS was retrieved, and because of a stil- persistent esophageal
tear from 30 to 34 cm, management was converted to EVT (Esosponge, BBraun). The patient
improved clinically and 7 days later the sponge was retrieved. However, the still-persistent
leak required continued treatment ([Fig. 3a]). Because of her good physical and neurological status with no need for a ventilator,
a VACStent was inserted to allow oral nutrition ([Fig. 3b]). This was monitored fluoroscopically by upper gastrointestinal tract contrast series
with the patient fully conscious ([Fig. 3c]). This study also confirmed that the esophageal tear was fully sealed off ([Fig. 3d]). At VACStent removal after 5 days, the leak was found to be closed by extensive
granulation tissue, leaving no residual cavity ([Fig. 3e]). This was monitored during 6-month follow-up, during which, the patient had no
residual clinical symptoms ([Fig. 3f]).
Fig. 3 a Superficial rather broad mucosal dehiscence from 7.00 to 9.00 with a persisting small
transmural gap at the top of the tear. b The applicated VAC Stent has already expanded, the suction catheter is seen at 2.00.
c Gastrografin contrast liquid was swallowed and passes through the VAC Stent into
the gastric fundus. d The endoscope is passed through the VAC Stent and the positioning of the distal bulge
of the stent at the esophageal-gastric junction is demonstrated. e The mucosal tear is now completely filled with granion tissue, and no further leak
was found. f One month later no further scar or stenosis was detected.
The third patient was a 75-year-old woman treated for severe dysphagia due to a stenosis
induced by a migrated magnetic LINX band (Johnson & Johnson, United States), which
had been implanted 2.5 years previously. Gastroscopy demonstrated endoluminal invasion
of the metal band over one-third of the circumference of the distal esophagus ([Fig. 4a]). The band was removed surgically, leaving an intraoperative transmural gap ([Fig. 4b]) plugged by insertion of a VACStent with a constant negative pressure of –80 mmHg.
During the completely unremarkable postoperative course, the patient was able to swallow
liquids and also a blended diet. After 4 days, the VACStent was removed, leaving good
granulation tissue within a sealed-off gap. On postoperative day 6, an upper gastrointestinal
contrast series using swallowed contrast fluid demonstrated a sealed situation with
good passage of the gastro-esophageal junction (GEJ). The patient was discharged on
postoperative day 8, and during follow-up, no further dysphagia was reported.
Fig. 4 a The distal esophagus is invaded by the LINX band with a functional stenosis. More
than one-third of the circumference has migrated into the lumen. b After removal of the LINX band a transmural leak sustained at the location where
the LINX band had invaded.
Discussion
These initial clinical cases demonstrates that the VACStent was easy to insert and
able to seal off esophageal leakage and anastomotic failure. The VACStent concept,
with its combined benefits of the EVT and covered SEMS, has potential to improve clinical
management of these otherwise often challenging cases [3].
The VACStent is a hybrid system comprising a fully covered stent and a polyurethane
sponge cylinder connected to a vacuum pump. Clinical experience proves that the sponge
cylinder is able to exert circumferential continuous negative pressure on the intestinal
wall, which becomes the driving force for both elimination of inflammatory secretions
and anchoring of the VACStent. Thus, no migration of any applied VACStent has been
observed to date. This is impressive, since the GEJ is a challenging region in which
most covered SEMS will dislodge [2]
[4]. In line with the basic concept, use of continuous negative pressure, as measured
by the suction pump, resulted in no significant problems in any of these cases. In
other words, no leaks or sponge obstructions were observed in these patients, implying
that the EVT principle also works in this hybrid stent system.
The sponge material employed, open-cell polyurethane foam, is identical to that successfully
used for years in negative-pressure wound therapy (NPWT) [10] and the endoscopic EVT system (BBraun AG, Melsungen, Germany). Thus, we observed
clearing of debris and excellent induction of granulation tissue, reflecting good
wound healing. Given this evidence we believe that the VACStent is able to seal off
the sponge cylinder against the intestinal lumen, thereby allowing continuous negative
pressure to anchor the stent and heal the perforation. Even in case of intermittent
leaks between the ends of the covered stent and the intestinal wall, the suction force
of the negative pressure reapproximates the flexible intestinal wall and closes the
gap. That this indeed does happen in clinical practice is reflected by the good experience
documented with the pressure control via the vacuum pump.
The drawback of endoluminal EVT is obstruction of the gastroinstinal tract passage
[4]
[5]. This causes microaspiration and mandates parenteral nutrition or application of
an additional feeding tube running alongside or through the sponge. Our experience
demonstrates that the VACStent offers the option of oral nutrition. Of course, swallowing
is hampered somewhat, because the covered stent blocks the active peristaltic wave
over the length of the stent and passage is only passive. As evidenced by our clinical
observations and radiographic findings, this may slow down passage but still allow
it.
Clinical success of this new device will be subject to several parameters, which are
and will be addressed by ongoing and future studies [9]
[10]. One major parameter is the selected negative pressure [7]
[8]
[11]. This level will not only impact stent-migration stability and the rate of wound
healing but also determine the extent of ingrowth into the sponge. Due to the cell
size and the consistency of the wound surface, the sponge may be suctioned strongly
to the wound bed and thus adhere there quite tightly. This ingrowth is the reason
why the sponge must be changed every 3 to 7 days [12]
[13]. In other words, the VACStent must be removed at least once a week. Our experience
shows that the duration of the needed VACStent treatment depends on the degree of
leak dehiscence, possible presence of an abscess or larger extraluminal cavity, and
the time between initial diagnosis and sealing of the leakage. In the one case of
iatrogenic intraoperative transmural defect, just one VACStent application for 4 days
was sufficient for successful clinical closure. Similarly, successful brief durations
of treatment were observed in iatrogenic esophageal perforation managed by endoluminal
EVT [14].
There are some limitations of this study, as three cases are not very representative
of the different types of upper gastrointestinal defects and none of the patients
exhibited a large defect or a greatly infected cavity. Larger mediastinal wound cavities
may be treated either directly with a VACStent, when the opening of the cavity is
large enough to admit negative pressure within, or as a first step, by intracavitary
EVT. The next step may involve a VACStent as in the management of our second case.
Conclusions
Future trials should address whether the efficacy of the VACStent is comparable to
standard EVT in complex situations and how cost-effective it is. If this can be proven,
then the VACStent is first choice for early detected defects or anastomotic leaks,
but also for complex cavities, if enough drainage capacity is provided by the sponge
cylinder to allow NPWT.
Most VACStent indications, however, will be for anastomotic leakage following oncological
gastroesophageal surgery or bariatric procedures, e. g. sleeve gastrectomy and gastric
bypass. Current studies will reveal the role that this innovative device will play
in the armory of interventional endoscopists.
