Introduction
Upper gastrointestinal bleeding (UGIB) is encountered commonly in emergency departments
and can result in hospital admission and significant morbidity and mortality [1]. Currently, endoscopic hemostasis is accepted as a first-line treatment modality
for management of UGIB and has been demonstrated to effectively reduce rebleeding,
surgical intervention, and mortality rates [2]. Endoscopic methods commonly used to control bleeding are injection or thermal or
mechanical devices. However, endoscopic management of UGIB is often challenging due
to anatomic position or a diffuse bleeding lesion, and as a result, endoscopic hemostasis
has a failure rate of 8 % to 15 % [3]. In addition, success of endoscopic hemostasis is dependent on operator skill, thus
success rates may be lower when endoscopic hemostasis is performed by inexperienced
hands [4]. In view of these challenges, more effective and straightforward endoscopic hemostasis
modalities are required.
Hemostatic powders have been developed recently for endoscopic hemostasis and reportedly
have excellent immediate hemostatic rates (93 % – 98 %) in UGIB patients [5]
[6]. These powders have the advantage of being easy to use because application does
not require accurate spraying at bleeding sites, which are frequently difficult to
visualize in endoscopic view and to approach. For these reasons, clinicians not experienced
at endoscopic hemostasis are expected to be able to use hemostatic powders in emergency
cases. However, recently developed powders have high rebleeding rates (33 % – 49 %)
[7]
[8], and their applications may be demanding due to, for example, clogging of the delivery
catheter and impaired visualization due to formation of powder-in-air suspensions.
To address limitations of available hemostatic powders, we developed a hemostatic
adhesive powder (UI-EWD; [Fig. 1]) composed of a biocompatible natural polymer produced using aldehyde dextran and
succinic acid modified ε-poly (l-lysine), which is immediately converted to a highly
adhesive hydrogel in the presence of water. The reaction between UI-EWD and water
leads to formation of a Schiff base and to multiple crosslinks within the hydrogel
and between it and tissues. UI-EWD can be delivered to bleeding loci without catheter
clogging or forming powder-in-air suspensions by liquid coating technology using a
fluidized bed granulator [9]. Clinically, UI-EWD has been shown to achieve high immediate hemostasis rates and
low rebleeding in patients with UGIB refractory to conventional endoscopic procedures
[9]. To obtain further information on practical performance of UI-EWD, the current study
was undertaken to evaluate the feasibility and efficacy of UI-EWD monotherapy for
treatment of non-variceal upper gastrointestinal bleeding (NVUGIB).
Fig. 1 Images of a UI-EWD and b spraying devices.
Patients and methods
Study design and study population
Patients were retrospectively selected from established prospective registries using
the following inclusion criteria: (1) age > 18 years; (2) signs of acute bleeding
(e. g., coffee-ground or fresh blood vomiting and/or melena); (3) acute NVUGIB of
Forrest class Ib, IIa, or IIb caused by peptic ulcer, post-endoscopic therapy, tumor,
or others; and (4) receipt of endoscopic hemostasis using UI-EWD monotherapy. Exclusion
criteria applied were as follows: (1) low gastrointestinal tract bleeding; (2) variceal
bleeding at time of endoscopy; (3) hospitalization for another illness; and (4) receipt
of treatment by other endoscopic or surgical means within 30 days prior to UI-EWD
application. Fifty-six consecutive NVUGIB patients that met study inclusion criteria
underwent UI-EWD monotherapy from January 2017 to December 2018. Patient medical records
were analyzed and information on clinical characteristics, bleeding, clinical outcomes
(including immediate hemostasis success and rebleeding rates), and UI-EWD-related
complications were collected. The study protocol was approved by our Institutional
Review Board (INHAUH 2019-04-014).
Endoscopic procedures
All bleeding was controlled by one of six experienced endoscopists using UI-EWD; each
endoscopist had performed more than 1000 endoscopic procedures per annum and had extensive
experience (6 – 25 years) with therapeutic endoscopy for UGIB. UI-EWD was applied
using a conventional endoscope (GIF-H290 Evis Lucera Elite; GIF-H260 Evis Lucera;
GIF-2TQ 260 M Evis Lucera; Olympus, Tokyo, Japan) to completely cover bleeding lesions
of Forrest class Ib, IIa, or IIb using an 8 Fr catheter and a novel spraying device
under direct endoscopic vision. After positioning the catheter for powder delivery,
the working endoscopic channel was flushed with air (delivered using a 60-mL syringe)
to ensure that the catheter tip was dry. Initially, 3 g of UI-EWD was administered
in onto bleeding lesions, which were then directly observed for 5 minutes using the
endoscope. If bleeding was not controlled, UI-EWD application was repeated up to a
maximum powder delivery of 6 g ([Fig. 2]). In cases of failure, a conventional endoscopic hemostatic modality was used based
on discretion of the endoscopist, and the case was defined as an immediate hemostasis
failure. Standard scheduled second-look endoscopy was not conducted.
Fig. 2 Representative endoscopic images of UI-EWD application for non-variceal upper gastrointestinal
bleeding. a Endoscopic submucosal dissection (ESD) for early gastric cancer was conducted from
the stomach body. b A Forrest-IB bleeding from dissection was noted 1 day after ESD. c Application of UI-EWD at the bleeding site; the bleeding lesion completely covered.
d Second-look endoscopy 72 hours after UI-EWD application, showing ulcer and no bleeding.
Outcome measurements
Rates of technical success, immediate hemostasis, and rebleeding were evaluated. In
addition, all medical records were checked to evaluate adverse events associated with
UI-EWD, such as newly developed symptoms, vital signs, and blood tests (i. e., hemoglobin,
platelet count, and chemistry). Successful immediate hemostasis was defined when powder
application led to hemostasis within 10 minutes by visual inspection. Rebleeding was
defined as clinical evidence of bleeding, such as melena or hematemesis, with an associated
hemoglobin reduction ≥ 2 g/dL within 30 days of the endoscopic procedure. This definition
was applied to all bleeding cases to suspect rebleeding. When rebleeding was suspected,
second-look endoscopy was performed to confirm its presence.
Results
Patient demographics
Fifty-six patients (43 men, 13 women) of mean age 64.6 ± 11.2 years were treated with
UI-EWD monotherapy for UGIB by six endoscopists at two centers ([Fig. 3]). Baseline patient characteristics and indication for UI-EWD are summarized in [Table 1]. Post-interventional bleeding was the most common cause of bleeding and the most
common location was the stomach body. Bleeding was classified as Forrest Ib (n = 36),
IIa (n = 12), or IIb (n = 8).
Fig. 3 Flow diagram of the study design, showing the number of patients at each exclusion
step.
Table 1
Baseline characteristics of patients and bleeding.
|
Characteristic
|
Value
|
|
Patients, n
|
56
|
|
Age, mean (SD), years
|
64.6 (11.2)
|
|
Sex, n (%)
|
|
|
43 (76.8)
|
|
|
13 (23.2)
|
|
Bleeding etiology, n (%)
|
|
|
46 (82.1)
|
|
|
8 (14.3)
|
|
|
1 (1.8 %)
|
|
|
1 (1.8 %)
|
|
Bleeding location, n (%)
|
|
Stomach
|
|
|
3 (5.3)
|
|
|
24 (42.9)
|
|
|
15 (26.8)
|
|
|
13 (23.2)
|
|
|
1 (1.8)
|
|
Forrest classification, n (%)
|
|
|
36 (64.3)
|
|
|
12 (21.4)
|
|
|
8 (14.3)
|
Clinical outcomes
Immediate hemostasis
UI-EWD was used as the initial bleeding control modality for NVUGIB and was successful
in all patients, and in all cases, the amount of UI-EWD used was ≤ 6 g (3 g, 52 patients;
6 g, 4 patients). Primary hemostasis rates achieved by UI-EWD monotherapy are shown
in [Table 2]. Immediate hemostasis was achieved in 54 of 56 patients (96.4 %). Of the remaining
two patients, one with post-interventional bleeding was successfully treated with
thermal therapy and in the other, who underwent jejunal anastomosis, bleeding was
controlled using thermal therapy with endoclips. At second-look endoscopy 24 hours
after procedures, UI-EWD hydrogel remained attached at bleeding sites in 33 of 47
patients (70.2 %) and at 3 days after procedures remained attached in 15 of 38 patients
(39.4 %).
Table 2
Clinical outcomes of UI-EWD for non-variceal upper gastrointestinal bleeding
|
Total patients, n
|
56
|
|
Overall success of immediate hemostasis, n (%)
|
54 (96.4)
|
|
Overall rebleeding in day 7, n (%)
|
2 (3.7)
|
|
Success rate of hemostasis according to etiology, n (%)
|
|
|
45 (97.8)
|
|
|
8 (100)
|
|
|
0 (0)
|
|
|
1 (100)
|
|
Success rate of hemostasis according to Forrest classification, n (%)
|
|
|
35 (97.2)
|
|
|
11 (91.6)
|
|
|
8 (100)
|
Rebleeding
Rebleeding within 7 days occurred in two patients (3.7 %), all of which was ascribed
to post-interventional bleeding, and occurred within 24 hours of procedures. All rebleeding
was successfully treated with a conventional modality, and neither surgery nor interventional
radiology was required to achieve hemostasis.
Spraying catheter clogging and adverse events
Clogging of spraying catheters during UI-EWD spraying occurred in two of 56 patients
(3.6 %), and when it occurred the powder was sprayed using another catheter. No procedure-related
AEs related to UI-EWD application such as intestinal obstruction or perforation occurred.
Discussion
Results of this study suggest that UI-EWD constitutes a reasonable primary treatment
strategy for NVUGIB. The powder was easily applied at all bleeding sites, the catheter
clogging rate was low at 3.6 %, and the immediate hemostatic rate was excellent (96.4 %).
In addition, the rebleeding rate within 7 days of endoscopic procedures was only 3.7 %.
Recently, several hemostatic powders like TC-325 (Hemospray, Cook Medical, Winston-Salem,
North Carolina, United States), Endoclot (AMP; EndoClot Plus Inc, Santa Clara, California,
United States), and Ankaferd Blood Stopper (AnkaferdHealth Products Ltd, Istanbul,
Turkey) were introduced for the endoscopic control of UGIB. These powders have the
advantage of being easy to apply in difficult situations to control bleeding from
multiple sites or large areas and have been demonstrated to allow immediate hemostasis
to be achieved satisfactorily (73.4 – 100 %) [5]
[10]
[11], but they have also been associated with high rebleeding risk. Sung et al. were
the first to describe use of TC-325 in human subjects. They evaluated the hemostatic
effect of TC-32520 patients with NVUGIB (mostly Forrest Ib) from gastroduodenal ulcers
with high-risk stigmata in a manner similar to that used in the current study [12]. The immediate hemostasis rate was found to be excellent at 95 % for TC-325 monotherapy,
but the 7-day rebleeding rate was high (10.5 %). Furthermore, second-look endoscopy
at 72 hours, which was performed in all subjects, showed ulcer healing but no remaining
hemostatic powder [12]. A later series by Smith et al. returned a much more modest rate of immediate hemostasis
(76 %) and a relatively high 7-day rebleeding rate (15.8 %) for TC-325 monotherapy
[13]. Subsequently, some experts recommended hemostatic powders not be used to treat
conditions such as ulcers with high-risk stigmata because hydrogel residence times
are probably less than 24 hours [14].
In the current study, UI-EWD was also found to have an excellent immediate hemostatic
effect (96.4 %), but the rebleeding rate was only 3.7 % (2/54). We attribute this
low rate to use of a Schiff base reaction between the hydrogel and tissue [9]. In previous reports, hydrogels produced using other commercially available hemostatic
powders were not observed by second-look endoscopy 3 days after endoscopic treatments
[13]
[15], whereas we found UI-EWD hydrogel was still present at 70.2 % (33 of 47 patients) of bleeding sites by second-look endoscopy at 24 hours and at 39.4 %
(15 of 38 patients) of bleeding sites at 3 days after procedures. We believe this
extended residence of UI-EWD provides more effective tissue sealing and mechanical
tamponade and increases effectiveness of hemostasis. However, the hemostatic effect
of UI-EWD on spurting arterial bleeding was not investigated because no case of Forrest
type Ia bleeding was included. In a previous study, UI-EWD was found not to satisfactorily
arrest spurting arterial bleeding (50 %), and it was suggested a method other than
UI-EWD be used to address this type of bleeding [9]. We suggest a further well-designed prospective study be undertaken to determine
the efficacy UI-EWD in cases of spurting arterial bleeding.
Ideally, endoscopic hemostasis should immediately stop active bleeding, prevent recurrent
bleeding, and be easily applied to lesions in any location in the gastrointestinal
tract. Although the delivery systems used for currently available hemostatic powders
have been improved, they are probably not optimal. Endoscopists must be careful when
delivering the agent and should avoid contacting tissue with the delivery catheter
tip because of risk of clogging. In addition, if application is conducted too close
to the lesion, powder-in-air suspensions may impair adequate visualization [16]. To address this limitation, we used a fluidized bed granulator to modify the water
absorption capacity of UI-EWD powder and applied it as a liquid coating, which allowed
us to deliver the agent to bleeding areas without catheter clogging or powder scattering
[9]. Actually, the catheter clogging rate was only 3.6 % (2/56) in the current study,
and adequate visualization was achieved during UI-EWD application in almost all cases.
Some limitations of the current study warrant consideration. First, it is inherently
limited by its retrospective design. Second, the majority of bleeding cases were caused
by post-interventional bleeding and the bleeding lesions in the stomach body or antrum,
which is a convenient position in which for the clinician to spray UI-EWD. In particular,
spurting arterial bleeding was excluded, which suggests the possibility that our results
do not well reflect the performance UI-EWD for classic ulcer bleeding. Third, because
the current study was conducted in tertiary care settings, patients may have exhibited
bias toward ost-interventional bleeding (82.1 %). In our clinical experience, post-interventional
bleeding is easier to control than classic peptic ulcer bleeding because the focus
of bleeding can be identified more easily and UI-EWD is able to splay more effectively
against the bleeding focus. This tendency may have affected immediate hemostasis rate
of UI-EWD as well as rebleeding rate. In other words, there would be a risk that the
effect of UI-EWD has been overestimated. Therefore, we suggest that prospective, multicenter
studies be conducted to confirm the efficacy of UI-EWD in daily endoscopic practice.
Conclusion
This study demonstrates the effectiveness of UI-EWD monotherapy for treatment of NVUGIB.
Based on our results, we suggest it is worth considering UI-EWD as the initial hemostatic
method in emergency NVUGIB cases not suitable for traditional hemostatic endoscopic
treatments, such as clipping or thermal therapy.
