Introduction
Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are
established techniques for the treatment of precancerous lesions and early stage cancer
in the western world [1 ]
[2 ]. The ESD technique was originally developed for the resection of gastric neoplasms,
but is increasingly used in the colon and the esophagus. Compared to EMR, ESD has
certain advantages, including a higher rate of en bloc resection with a precise histopathological
examination. In addition, resection of larger (> 2 cm) lesions is possible with simultaneously
lower recurrence rates [3 ]
[4 ]
[5 ]. On the other hand, ESD is an advanced technique with a long learning curve, even
for experienced endoscopists, has a longer procedure time than EMR and a higher incidence
of adverse events. One of the major complications is perforations, with an incidence
up to 4 – 10 % [2 ]. For that reason, perforation prevention is essential, and the most effective way
to do this is to use an adequate and effective submucosal fluid cushion [6 ]. For this reason, the ESGE guidelines specify an ideal submucosal injection fluid
which provides a sustainable lift, facile en bloc resection, low cost, widespread
availability and few adverse events [7 ].
To date, the most studied injection fluid is NaCl 0.9 %, beginning in the early 1950s
[8 ], and with the advantages of widespread availability and low cost [9 ]
[10 ]. The limitation of NaCl 0.9 % is that it does not provide a long lasting cushion
effect and repeated injections are necessary even after a short period of time, which
prolongs the procedure time [11 ]
[12 ].
Other solutions studied include gelatin, hydroxypropylmethylcellulose (HPMC), hydroxyethyl
starch (HAES) (Hetastarch, Hespan), hyaluronic acid, glucose, dextrose, fructose,
glycerine, Na-alginate, Eleview (Aries Pharmaceuticals, San Diego, California, United
States) and mixtures of all those components [9 ]
[13 ]. Up to now, no clear recommendation exists to support the use of one fluid over
another; however, comments have been made in position and guideline papers (ESGE)
promoting the use of more viscous substances over normal saline in EMR [7 ]
[14 ]. There is an absolute necessity for research in this field.
It is for this reason that a new highly viscous injectate solution (LiftUp, Ovesco
Endoscopy AG, Tübingen, Germany) was developed. LiftUp is composed of distilled water,
and a co-polymer derived from ethylene oxide, propylene oxide and sodium chloride,
and has a thermo-gelling behavior. It is a co-polymer, which is fluid at low temperatures,
and forming micelles above a critical temperature. The micelles connect above the
critical micelle temperature and form a grid, which increases the viscosity of the
fluid. Maximum viscosity is reached at body temperature (37 °C). This is a reversible
effect, which allows for washing out LiftUp by cooling the resection site with cold
water. It builds a gelatinous cushion, which remains stable and therefore enables
subsequent resection without reinjection.
The aim of this study is to evaluate different parameters of the ESD procedure with
the thermo-sensitive injectate LiftUp in comparison to NaCl 0.9 % and HAES 6 % (control
group) in an established ex vivo ESD model.
Materials and methods
This study was exempt from IRB review since no animal or human subjects were included.
It was designed as a prospective, randomized ex vivo study. The experiments were carried
out at the Federal Research Institute for Animal Health in Mariensee, Germany. Stomachs
from pigs (German Landrace, live weight 60 kg) were cleaned and frozen after slaughtering.
The stomachs were thawed before the procedure and placed in the EASIE-R simulator
([Fig. 1 ]). This model has previously been evaluated in different endoscopic procedures and
is an established model for interventional endoscopic training and research [15 ]
[16 ].
Fig. 1 EASIE-R simulator for ESD procedures. The porcine stomach is thawed before the procedure
and placed in the EASIE-R simulator. A water bath at 37 °C was used to simulate the
in vivo environment and keep the porcine stomach homothermal.
All 60 procedures were performed by one endoscopist, who had previous experience in
performing gastric ESD in the ex vivo simulator (> 60 ESDs before this study) and
was blinded to the injection fluid used.
LiftUp was compared to two standard submucosal injection fluids: hydroxyethyl starch
(HAES 6 %, B. Braun, Melsungen, Germany) and normal saline (NaCl 0.9 %, B. Braun,
Melsungen, Germany). In total, 20 ESD procedures were performed for each injection
fluid.
Preparation of LiftUp for submucosal injection
At low temperatures, LiftUp is liquid and forms a gel-like consistency with increasing
temperature. Due to its higher viscosity, an inflation device (Accura Medizintechnik
GmbH, Ref. no. 8610230S) was used to deliver the material through the injection needle
(needle length 6 mm, needle diameter 0.7 mm), providing a pressure of up to 20 bars.
To keep the conditions comparable, the same injection needle was also used for HAES
and NaCl injections. After submucosal injection of LiftUp, the material was then left
in place in the submucosa for approximately 1 – 3 minutes to form the desired gel-like
configuration. A water bath at 37 °C was used to simulate the in vivo environment
and keep the porcine stomach homothermal ([Fig. 1 ]). After 5 minutes, endoscopic resection was initiated. Lifting properties were measured
by the blinded researcher using a subjective score from 1 to 5 points, with 5 points
being the best valuation and representing properties of an ideal lifting injection
with an efficient submucosal cushion. Ovesco is currently in the process of applying
for CE marking for LiftUp.
ESD procedure
Before the interventions, standardized lesions, each measuring 3 × 3 cm, were manually
placed in the corpus of the porcine stomach ([Fig. 2 ]). The stomach was placed in the EASIE-R model, which consists of a plastic hull,
where the esophagus and the pig stomach are secured [15 ]. Six ESD lesions were created per ex vivo stomach. The endoscopist was blinded with
regard to which injection solution was provided for the resection. For randomization,
60 identical lots (20 for each injection solution) were drawn from a pot. Before injection,
the LiftUp solution was stored cooled between 2 °C and 8 °C in a refrigerator. All
lesions were resected using the classic ESD technique. The procedures were performed
sequentially, over a study period of 1 month. ESD was performed using a gastroscope
(EG-530 D Fujinon, Fujifilm, Japan), an ESD cap (DH 29CR ST, Hood short type, Fujifilm,
Japan) and a FlushKnife 1.5 mm (Fujinon, Fujifilm, Japan). Injection fluids were mixed
with Methylene Blue dye (Amino AG, Gebenstorf, Switzerland) to optimize visualization
and differentiation of submucosal fibers during the ESD. The electrosurgical unit
used was the ERBE VIAO 200 (ERBE Elektromedizin, Tübingen, Germany). For the ESD,
the settings on the ERBE device were EndoCut Q 1/1/1.
Fig. 2 Standardized lesions to be resected by ESD were manually placed in the corpus of
porcine stomach, each measuring 3 × 3 cm.
Data collection
An independent observer recorded the following parameters: time required for fluid
injection, amount of fluid injected, time required for circumferential incision, ESD
procedure time, overall procedure time (submucosal dissection time), en bloc resection
rate, and perforation rate. Data concerning all parameters mentioned were collected
during each ESD. The resected specimens were spread and pinned on cork plates after
each ESD procedure and photo documented ([Fig. 3 ]). The ESD procedure time was defined as the time from the beginning of the ESD procedure,
after circumferential incision, to complete resection of the lesion from the porcine
stomach. Each resection site was evaluated visually from the endoscopic site in addition
to an insufflation test to determine the presence of a perforation.
Fig. 3 Steps in the ESD procedure with LiftUp. a Submucosal cushion after LiftUp injection. b – d Submucosal dissection with a FlushKnife. The submucosal cushion remains stable with
LiftUp during the entire procedure. The specimen was spread and pinned on a cork mat
after the ESD procedure.
Statistical analysis
Data analysis was performed using JMP 11 (SAS Institute Inc., 2014) and R 3.3.3 (R
Core Team, 2017) software. Box and whisker plots were used to graphically represent
the spread in the data by showing the quartiles of the data along with the extreme
values. The bottom and top of the box are the 25th (Q1) and 75th (Q3) percentiles,
and the band near the middle of the box is the median (50th percentile, Q2). The ends
of the whiskers represent Q1 minus 1.5 × interquartile range (IQR) and Q3 plus 1.5 × IQR,
respectively.
For continuous variables (injection time, etc.), differences between injection fluid
groups were tested using a Kruskal – Wallis test (one-way ANOVA on ranks), followed
by nonparametric comparisons of each pair using the Wilcoxon method when the global
test was significant. For binary variables (en bloc resection and adverse events),
we used Fisher’s exact test to compare the differences between injection fluid groups.
A P value less than 0.05 was considered to be statistically significant.
Results
Injection time for the submucosal injectates
Due to its more viscous properties, the injection time with LiftUp (median 4 minutes,
range 2 – 7.8 minutes) was significantly longer than HAES (median 3 minutes, range
1 – 6.9 minutes, P = 0.0177), and NaCl (median 3 minutes, range 1 – 6 minutes, P = 0.025).
Lifting properties and required volume of the submucosal injectates
The lifting persistency of all three injection solutions is presented in [Fig. 4 ]. Injection with LiftUp received the best scores with 16 of 20 receiving 5 points.
The median quality score for HAES = 4 vs. median LiftUp 5, P = 0.0005. The median quality score for NaCl = 3 vs. LiftUp 5, P = 0.0002.
Fig. 4 Lifting persistency under simulation circumstances in the porcine model for HAES
6 %, LiftUp, and NaCl 0.9 %.
The overall injection volume needed for adequate lifting during ESD was significantly
lower in the LiftUp group compared to the HAES (median 33.5 mL vs. 231 mL, P < 0.0001) or normal saline groups (median 33.5 mL vs. 299 mL, P < 0.0001) ([Fig. 5 ]). Within 60 minutes of the injection, only LiftUp permitted a stable cushion.
Fig. 5 Injection volume (mL) dependent on the injection substance (mL).
With regard to the time required for the circumferential cut, there was no statistically
significant difference between the injection groups (LiftUp vs. HAES: median 10.5
minutes vs. 8.0 minutes, P = 0.5688/LiftUp vs. NaCl: median 8.0 minutes vs. 7.5 minutes, P = 0.2219).
The variability of the ESD procedure time was lower in the LiftUp group compared to
the other two groups. However, despite a trend towards lower ESD procedure time in
the LiftUp group compared to the NaCl group, the differences did not reach statistical
significance (LiftUp vs. HAES: median 19.5 minutes vs. 22.0 minutes, P = 0.3296/LiftUp vs. NaCl: 19.5 minutes vs. 25.0 minutes, P = 0.0544) ([Fig. 6 ]).
Fig. 6 ESD procedure time (minutes) dependent on the injection substance.
Similarly, there were no statistically significant differences in total procedure
time between groups.
En bloc resection rate
En bloc resection was achieved in 20 out 20 ESD cases when HAES was used, compared
to 19 out of 20 ESDs in the LiftUp group and 16 out of 20 ESDs in the NaCl group (P > 0.05).
Adverse events
Due to the ex vivo nature of the ESD model, peri- or post-ESD bleeding was not able
to be simulated. No perforation occurred during ESD in the LiftUp group (n = 0/20),
whereas one perforation occurred in the HAES group (1/20) and two in the normal saline
group (2/20) (P > 0.05).
Discussion
The purpose of this study was to evaluate the performance characteristics of three
submucosal injection materials for ESD, including a new thermosensitive co-polymer
(LiftUp), in comparison to established injectates. ESD and EMR are minimally invasive
endoscopic procedures increasingly accepted for treatment of dysplastic and early
malignant lesions limited to the superficial layers of the gastrointestinal wall [13 ]
[17 ]. The disadvantages of EMR are a relatively higher recurrence rate, especially when
piecemeal EMR is performed [18 ]. ESD is superior to EMR in achieving en bloc resection for large lesions [19 ].
The submucosal cushion, produced with the injection of a solution into the submucosal
space, is an essential step in ESD with regard to ensuring the safety and efficiency
of the procedure. An ideal injection substance should create a long lasting “cushion”
effect by expanding the submucosal space, thereby preventing perforations and limiting
the need of recurrent injections.
The ESGE guidelines have recommended more viscous injection solutions over the use
of normal saline for endoscopic mucosal resections [7 ]. Special technical recommendations concerning injection fluids in ESD do not exist
yet and most of the commonly used submucosal injectates are used off-label for ESD
at this time [20 ].
To date, there have been a number of trials comparing different submucosal injection
fluids for ESD use [21 ]
[22 ]
[23 ]
[24 ]
[25 ]
[26 ]. Ferreira et al. have investigated a number of existing injection solutions in a
meta-analysis, including 11 randomized controlled trials (RCTs) with two on ESD. They
concluded that sodium hyaluronate and normal saline injectates were clinically equivalent
[13 ]. Recently, Spadaccini et al. published an RCT in a porcine model showing that a
new biocompatible polymer SIC-8000 (Eleview) was more effective for ESD and EMR compared
to saline solution [27 ]. Mehta et al. found that lower injection volumes and fewer repeat injections were
required with HAES and Eleview in comparison to normal saline for the same mucosal
lifting effect [26 ].
In general, normal saline is reported to be a safe solution with regard to adverse
events and cost-effectiveness. In terms of mucosa lifting, it is inferior to more
viscous solutions [13 ]. In an RCT, hydroxyethyl starch has been shown to increase mucosal lifting time
and reduce the need for additional injections in comparison to normal saline solution
[28 ].
A newly developed injection material, LiftUp, is currently undergoing investigation
for its use in ESD and EMR. In this study, we conducted a prospective, randomized
ex vivo study comparing the new fluid with two well-investigated injectates, normal
saline solution and HAES. The results of this study demonstrate that LiftUp had significantly
better lift quality scores and required significantly less volume injected when compared
to HAES and NaCl. LiftUp provided a more long lasting cushion effect, which enabled
a secure submucosal dissection with no perforations in the LiftUp group. As a result
of its higher viscosity, the injection time for LiftUp was statistically longer than
for NaCl 0.9 % and HAES 6 %, but significantly less volume of LiftUp was needed. In
the LiftUp group, we performed one injection at the beginning of the procedure and
no more repeat injections were required for the remainder of the ESDs. Furthermore,
at body temperature, the LiftUp material does not diffuse into the surrounding tissue.
In this study, we found that procedure times and complications were not significantly
different between groups.
This current study does have some limitations. First, the copolymer was only compared
to HAES 6 % and NaCl 0.9 % and not to all injectates available for ESD. However, to
date, HAES 6 % and NaCl 0.9 % are the best studied injectates. A further limitation
is that the research was performed in an ex vivo model. The effects of blood flow,
tissue movement, tissue healing, histopathological evaluation, and other physiological
factors which may affect tissue fluid diffusion and durability of a submucosal cushion
could not be investigated.
We conclude that LiftUp has similar safety and efficiency qualities compared to NaCl
and HAES, but requires significantly less injected volume and has a preferable subjective
quality score to the blinded endoscopist. LiftUp, with its characteristic of a thermally
sensitive transition by body temperature, can effectively maintain mucosal elevation,
and has ideal lifting properties. Further clinical studies are warranted, to analyze
its characteristics in human ESD procedures.
