Introduction
Acute gastrointestinal bleeding (GIB) is a medical emergency and carries significant
morbidity and mortality [1]
[2]. Endoscopy is the diagnostic and therapeutic procedure of choice and is often indicated
within the first 24 hours [3]
[4].
Over the years, numerous therapeutic options have been developed to provide the optimal
strategy for hemostasis; however, success rates often vary depending on the location
of the lesion, etiology, patient factors, and endoscopist expertise. The existing
armamentarium includes injection needles, thermal devices, electrocoagulation probes,
forceps, clips (through-the-scope and cap-mounted), endoscopic suturing, banding devices,
and hemostatic powders [5]
[6]. However, technical aspects such as difficult anatomic position of the bleeding
lesion (including the posterior wall of the duodenal bulb and the lesser curvature
of the gastric body) and intense or diffuse bleeding may impact the efficacy of some
of these therapeutic modalities [7].
The use of an inorganic hemostatic powder, such as Hemospray (Cook Endoscopy, Winston
Salem, NC) or TC-325, has become popular in the last few years for various GIB scenarios,
including for non-variceal and malignant bleeding lesions [8]. Systematic reviews of observational cohort studies have reported high rates of
initial hemostasis (> 90 %) and > 15 % rate of recurrent bleeding with TC-325 use
[9]
[10]. Recently published outcomes from a 5-year international multicenter registry reported
a 100 % hemostasis rate with Hemospray use in malignant bleeds and suggested that
its use as monotherapy is feasible in routine clinical care [11]. Additionally, a cost-effectiveness analysis found Hemospray economical as a first-line
strategy [12].
The American College of Gastroenterology (ACG), as a conditional recommendation, suggests
using hemostatic spray in non-variceal upper GIB for bleeding ulcers [4]. In contrast, the European Society of Gastrointestinal Endoscopy (ESGE) preferred
its use in cases of refractory GIB [1]. However, both these recommendations were backed by very low-quality evidence. While
multiple prior meta-analyses have been conducted to better address our gap in knowledge
[13]
[14]
[15]
[16], comparative outcomes with standard endoscopic therapy (SET) have yet to be thoroughly
evaluated.
Recently, several randomized controlled trials (RCTs) have reported outcomes of TC-325
compared to SET in non-variceal GIB (NVGIB), especially high-risk peptic ulcers and
malignant bleeding lesions [17]
[18]
[19]. The aim of the study was to systematically appraise the published literature and
compare the efficacy of Hemospray (TC-325) to SET in patients with NVGIB.
Methods
Protocol and registration
This review was designed following the preferred reporting items for systematic reviews
and meta-analyses (PRISMA) statement to identify studies reporting clinical outcomes
of hemostatic power (Hemospray or TC-325) [20].
Eligibility criteria, literature search, and search strategy
A librarian conducted a systematic search of several databases and conference proceedings,
including EBM reviews via Ovid, Ovid Embase (1974 +), Ovid Medline (1946 + including
epub ahead of print, in-process, and other non-indexed citations), PubMed, EMBASE,
Google Scholar, LILACS, SCOPUS, and Web of Science databases in October 2022. Keywords
used in the literature search included a combination of “TC-325”, “Hemostatic powder”,
“Hemospray”, “gastrointestinal bleeding,” OR “GI bleeding”. The search was restricted
to studies on human subjects in English. Duplicates were filtered using EndNote. Reference
lists of identified sources were cross-checked for additional relevant studies by
SD and LK). The complete search strategy is available in Supplementary Appendix-A.
Study selection
[Heading 2]We included RCTs comparing Hemospray to standard therapy in patients with
NVGIB. Studies were included irrespective of whether they were performed in inpatient
or outpatient settings, follow-up time, and country of origin if they provided the
appropriate data needed for the analysis. Our exclusion criteria were as follows:
(1) observational cohort studies reporting outcomes of TC-325 without a comparative
arm; (2) single patient case reports and case series studies; (3) studies with sample
size < 10 patients; (4) studies reported on variceal bleeding; and (5) studies performed
in the pediatric population (Age < 18 years). In cases of multiple publications from
a single research group reporting on the same patient cohort and/or overlapping cohorts,
data from the most recent and/or most appropriate comprehensive report were retained.
The retained studies were determined based on the publication timing (most recent)
and/or the sample size of the study (largest). PRISMA Flowchart for study selection
and PRISMA Checklist are provided in Supplementary Fig. 1 and Supplementary Appendix
B, respectively. Reference lists of evaluated studies were examined to identify other
studies of interest.
Data abstraction and quality assessment
Data on study-related outcomes were abstracted onto an a priori-designed Google sheet
by two authors (NR, JAB), while two authors (SD, SC) completed the quality scoring
independently [21]. The Jadad scale was used to assess the quality of studies. A score of 2 or less
is considered low, 3 to 4 is moderate, and 5 is of excellent quality [22].
Outcomes assessed
The primary outcome was primary hemostasis, defined as endoscopically verified cessation
of bleeding for 3 to 5 minutes or control of bleeding within 30 days of randomization,
defined as achievement of endoscopic hemostasis by the assigned treatment modality
during the first endoscopy and no recurrent bleeding after endoscopic hemostasis.
The second outcomes were as follows. Failure to achieve hemostasis was defined as
recurrent bleeding during index intervention necessitating cross-over to alternative
therapeutic modality. Rebleeding was defined as a drop in hemoglobin of 2 g/dL or
more, a new episode of hematemesis/melena/hematochezia, red blood content from a nasogastric
or oro-gastric tube, or rebleeding identified during second look/repeat endoscopy
up to 30 days after the index procedure. Length of stay (LOS) was defined as the duration
of hospitalization post-index endoscopy. Rescue interventions were failed endoscopic
management necessitating rescue intervention, including arteriography, angiography,
and/or surgery. Mortality was death from any cause within the first 30 days after
index intervention.
Statistical analysis
Meta-analysis techniques were used to calculate the pooled risk ratios (RR) with pooled
estimates and 95 % confidence intervals (CI) and mean difference (SMD) using a random-effects
model. A continuity correction of 0.5 was added to incident cases before analysis
[23] We assessed heterogeneity using Cochran Q statistical test for heterogeneity with
I2 statistics [24]
[25]
[26]. In this, values < 30 %, 31 % to 60 %, 61 % to 75 %, and > 75 % were suggestive
of low, moderate, substantial, and considerable heterogeneity, respectively. Publication
bias was ascertained qualitatively by visual inspection and quantitatively by the
Egger test. When publication bias was present, further statistics using the fail-Safe
N test and Duval and Tweedie's 'Trim and Fill' test were used to ascertain the impact
of the bias [27]. All analyses were performed using comprehensive meta-analysis (CMA) V3 software
(Englewood, New Jersey, United States).
Results
Search results and population characteristics
All search results were exported to EndNote, where 2303 obvious duplicates were removed,
leaving 2342 citations. All titles were extracted and screened, and 145 full-length
articles were reviewed in detail. A schematic diagram demonstrating our study selection
is illustrated in Supplementary Fig. 1.
The final analysis included five RCTs with 362 patients (TC-325 178, SET 184) [17]
[18]
[19]
[28]
[29]. There were 123 females and 239 males (mean age 65 ± 16 years). The most common
etiologies of bleeding were PUD in 173 (TC-325 87, SET 86) and upper and lower GI
malignancies in 123 and 3 patients, respectively (TC-325 71, SET 55). Other etiologies
included Mallory-Weiss in 15 patients, post-sphincterotomy bleeding in four, Dieulafoy
lesions in 24, reflux esophagitis, esophageal erosion, ischemic gastritis, gastric
amyloidosis, diffuse hemorrhage from erosive gastritis, portal hypertensive gastropathy,
angiodysplasia, and antral vascular ectasia in 12 patients. The etiology was not only
reported as others in eight and. The mean peptic ulcer size was 11.61 ± 8.4 mm (TC-325
11.58 ± 8.2 mm, standard therapy 11.63 ± 8.68 mm).
Characteristics and quality of included studies
Two studies were conducted in Brazil [19]
[29], one in Canada [18]
, and two in Singapore [17]
[28]. Bleeding was characterized as Forrest IA in 26 (T-325 13, SET 13), IB in 264 (TC-325
131, SET 133), IIA in 10 (TC-325 4, SET 6), and IIB in two patients (TC-325 1, SET
1). In three studies, hemorrhage was characterized as oozing and spurting [17]
[18]
[29]. All studies reported using TC-325 as the index intervention, while SET included
endoscopic hemoclips, saline adrenaline injections (1:10,000 in four quadrants in
0.5–2 ml aliquots), heater probe, bipolar electrocautery, argon plasma coagulation,
and/or laser photocoagulation. In one trial, a combination of epinephrine injection
and TC-325 was used [29], while TC-325 was used as monotherapy in all the others. Reported time to verify
bleeding cessation during index endoscopy – three minutes [18]
[19]
[29], five minutes [28], and unspecified [17]. Further details of included studies and patient characteristics are presented in
[Table 1] and [Table 2]. The majority of included studies were moderate [17]
[19]
[28] and excellent quality [18] with one low-quality (Supplementary Table 1) [29].
Table 1
Study details and population characteristics.
|
Study
|
Design
|
Standard therapy
|
Total patients
|
Age (mean, SD, y)
|
Gender (male/female)
|
Etiology
|
Ulcer size (mm)
|
|
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
|
Kwek 2017 [28]
|
Prospective, RCT, December 2013 to February 2015, single-center, Singapore
|
Clip + epinephrine injection/coagulation + epinephrine injection
|
10
|
10
|
67.9 (18.4)
|
72.1 (11.4)
|
9/1
|
7/3
|
PU 10 (Forest IA 1, IB, 4, IIA 4, IIB 1)
|
PU 10 (Forest IB, 3, IIA 6, IIB 1)
|
10.3 5.4
|
13.1 3.2
|
|
Baracat 2020 [29]
|
Prospective, RCT, July 2015 to July 2017, single-center, Brazil
|
Hemoclips + epinephrine injection
|
19 (Forest IA/IB)
|
20 (Forest IA/IB)
|
57.2 (16.2)
|
56.5 (15.6)
|
14/5
|
12/8
|
PU 9, Malignancy 4, others 6 (Forest IA 2, IB 16)
|
PU 8, Malignancy 1, others 11 (Forest IA 1, IB 19)
|
NR
|
NR
|
|
Chen 2020 [18]
|
Prospective, RCT, April 2014, dual-center, Canada
|
Heater probe, bipolar electrocautery, APC, and laser photocoagulation & Injection
treatments (Epinephrine injection and Sodium tetradecyl sulfate)
|
10 (Forest IA/IB)
|
10 (Forest IA/IB)
|
68.2 (9.7)
|
66.1 (20.9)
|
7/3
|
8/2
|
Malignancy 10 (Forest IA 1, IB 9)
|
Malignancy 10 (Forest IA 0, IB 10)
|
NR
|
NR
|
|
Lau 2022 [17]
|
Prospective, RCT, September 2015 to December 2018, Multicenter, Hong Kong, Thailand,
Singapore
|
Heater probe or Bipolar probe or Hemoclips with or without prior injection of diluted
Epinephrine
|
111 (Forest IA/IB)
|
113 (Forest IA/IB)
|
68.5 (15.0)
|
66.3 (16.7)
|
77/34
|
73/40
|
PU 68, malignancy 29, other 14 (Forest IA 9, IB 102)
|
PU 68, malignancy 13, other 32 (Forest IA 12, IB 101)
|
11.7 (8.4)
|
11.5 (9.0)
|
|
Costa Martins 2022 [19]
|
Prospective, RCT, August 2016 to February 2020, single-center, Brazil
|
Epinephrine injection, clipping, argon plasm coagulation or others
|
28
|
31
|
55 (15.3)
|
62.1 (12.1)
|
10/8
|
12/6
|
Malignancy 28
|
Malignancy 31
|
NR
|
NR
|
RCT, randomized controlled trial; NR, not reported.
Table 2
Study outcomes.
|
Study
|
Outcomes
|
|
Primary hemostasis
|
Failure
|
Rebleeding (30 d)
|
Length of hospital stay (range)
|
Rebleeding treatment
|
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
TC 325
|
Standard
|
|
Kwek 2017 [28]
|
9/10
|
10/10
|
1/10
|
0/10
|
3/10
|
1/10
|
NR
|
NR
|
2 endoscopy
1 angiography
|
1 endoscopy 0 angiography
|
|
Baracat 2020 [29]
|
19/19
|
18/20
|
0/19
|
2/20
|
5/19
|
3/20
|
11.00 (10.09)
|
5.94 (3.82)
|
1 surgery
|
0 surgery
|
|
Chen 2020 [18]
|
9/10
|
4/10
|
1/10
|
6/10
|
2/10
|
6/10
|
14.6 (9.9)
|
9.4 (6.1)
|
1 angiography
4 radiation
1 surgery
|
1 angiography
3 radiation
2 surgery
|
|
Lau 2022 [17]
|
100/111
|
92/113
|
3/111
|
11/113
|
9/111
|
10/113
|
6 (1–90) (after randomization)
|
6 (1–107) (after randomization)
|
8 endoscopy
2 angiography
1 surgery
|
10 endoscopy
4 angiography
0 surgery
|
|
Costa Martins 2022 [19]
|
22/22
|
NR
|
0/18
|
NR
|
9/28
|
6/31
|
17.4 (± 17.7)
|
12.8 (± 14.1)
|
1 surgery
0 arteriography
12 radiotherapy
|
2 surgery
16 radiotherapy
2 arteriography
|
NR, not reported.
Meta-analysis outcomes
There was no statistical difference in the pooled rates of primary hemostasis between
TC-325, 91 % (CI 85.4–94.6) compared to SET, 79.1 % (CI 53.9–92.5), RR 1.09 (CI 0.95–1.25;
I2 43), P = 0.2, including in patients with oozing/spurting hemorrhage (Forrest IA, IB), RR
1.13 (CI 0.98–1.3; I2 35), P = 0.08 ([Fig. 1]).
Fig. 1 Forest plot of primary hemostasis.
Given the variation in definition of primary hemostasis, as reported by Lau et al,
we performed a subgroup analysis excluding this study. We found no statistical difference
between the rates of primary hemostasis between the two modalities, RR 1.11 (CI 0.83–1.48;
I2 61), P = 0.5. [17].
The pooled rate of failure to achieve hemostasis was higher in SET, 16.5 % (CI 4.4–45.7)
compared to TC-325, 4.3 % (CI 1.9–9.5), RR 0.30 (CI 0.12–0.77, I2 0), P = 0.01, including among patients with oozing/spurting hemorrhage (Forrest IA, IB),
RR 0.24 (CI 0.09–0.63, I2 0), P = 0.004 ([Fig. 2]).
Fig. 2 Forest plot of failure to achieve hemostasis.
There were no statistical differences in the pooled rates of rebleeding between TC-325,
20.8 % (CI 10.6–36.6) compared to SET, 18.8 % (CI 8.8–37.3), RR 1.13 (CI 0.62–2.07;
I2 43), P = 0.7 ([Fig. 3]).
Fig. 3 Forest plot of rebleeding.
There was no statistical difference in the overall LOS between TC-325 and SET, with
standardized mean difference (SMD) 0.27 (CI, –0.20–0.74; I2 62), P = 0.3 (Supplementary Fig. 2).
Pooled rates of rescue interventions necessitating arteriography and/or angiography
were higher in SET, 20 % (CI 1.7–78.4) compared to TC-325, 13.2 % (CI 1.3–16.5), RR
0.68 (CI 0.5–0.94, I2 0), P = 0.02 (Supplementary Fig. 3). There was no statistical difference in the pooled rates
of rescue surgery between TC-325, 4.3 % (CI 1.7–10.7) and SET, 4 % (CI 0.9–15.8),
RR 1.11 (CI 0.3–3.7; I2 0), P = 0.9 (Supplementary Fig. 4).
Pooled rates of all-cause mortality were higher in TC-325, 18.9 % (CI 10.6–31.4),
compared to SET, 14.9 % (CI 10.2–21.3), however, the difference between the two was
not statistically significant, RR 1.14 (CI 0.69–1.9, I2 0), P = 0.6 (Supplementary Fig. 5).
Validation of meta-analysis results
Sensitivity analysis
To assess whether any study had a dominant effect on the meta-analysis, we excluded
one study at a time and analyzed its impact on our main summary estimate. While numerically
higher, we found no statistical difference in the pooled rates of primary hemostasis
between TC-325 and SET groups, RR 1.13 (CI 0.98–1.3; I2 35), P = 0.08. Upon sensitivity analysis and removing the study by Kewk et al, we noticed
that the difference did reach statistical significance. This may be because of several
reasons. Firstly, this manuscript was from a pilot feasibility study where only 20
patients were randomized to TC-325 and SET. Furthermore, only 40 % (8/20) had actively
bleeding Forrest Ia or Ib ulcers. Five of these patients received Hemospray (including
one ulcer that extended into the retroperitoneum), and only three with Forrest Ia
or Ib ulcers received standard dual therapy. Since TC-325 requires active bleeding
to achieve hemostasis and is not recommended for patients with non-bleeding vessels,
data regarding initial hemostasis and rebleeding rates from this study are difficult
to assess [28].
Heterogeneity
We assessed the dispersion of the calculated rates using the CI and I2 percentage values. The overall distribution of effects was minimal to low based on
the 95 % CI and I2% values across included studies.
Publication bias
Based on visual inspection of the funnel plot and quantitative measurement that used
the Egger regression test, there was no evidence of publication bias for hemostasis
and failure (Egger’s 2-tailed z = 0.31, p = 0.81 and Egger’s 2-tailed z = 0.60, P = 0.74) (Supplementary Fig. 6 and Supplementary Fig. 7).
Discussion
Our analysis, based on data from randomized controlled trials (RCTs), shows that among
patients with NVGIB, there is no significant difference in the rates of primary hemostasis,
rebleeding, length of hospital stays, or need for rescue surgery between Hemospray
(TC-325) and standard endoscopic therapy (SET). Our findings suggest that monotherapy
with TC-325 may be a viable option for NVGIB, including actively bleeding peptic ulcers
and malignancies, even though TC-325 is not typically used in this manner and can
be cost-prohibitive.
Current guidelines have suggested using TC-325 as a temporizing measure that should
be followed by a second definitive hemostatic modality [30]
[31]. This is because TC-325 sloughs off the mucosa and is eliminated from the gastrointestinal
tract within 24 hours after application, and subsequent bleeding is common in observational
studies of TC-325. Furthermore, SET may fail to achieve successful hemostasis in 8 %
to 15 % of patients with active peptic ulcer bleeding, and rebleeding occurs in 5 %
to 10 % of patients after initial hemostasis using combined endoscopic therapy [32]
[33]. Due to its ability to be applied to difficult-to-reach sites and treat large areas
where the exact location of bleeding is unknown, TC-325 offers a viable alternative
to SET [10]
[34]. Our pooled analysis shows similar rates of primary hemostasis between TC-325 monotherapy
and SET, adding to the current body of literature on the efficacy of TC-325.
When assessing failure to achieve hemostasis, we found that pooled rates were significantly
higher with SET than with TC-325. These results must be reviewed with caution since
failure rates of SET were notably higher in two studies [17]
[18], which included 62 patients with oozing and/or spurting malignant GIB. However,
after analyzing the data after removing the studies, failure rate was 6.2 % (1.2 %–26 %)
for TC-325 and 8.5 % (2.5%–25.4 %) for SET. It is known that endoscopic therapy for
malignant GIB is generally less successful and can be technically challenging because
of the large surface area of tissue requiring treatment, tissue friability, and underlying
coagulopathy. Data regarding the efficacy of SET for malignant GIB is variable, with
primary hemostasis rates reported between 31 % and 86 % and rebleeding rates between
28 % and 80 % [35]
[36]
[37]. A recent meta-analysis found TC-325 to be highly effective in this scenario with
a success rate of 94 % and rebleeding rates between 11 % to 24 % [8]. In our analysis, 126 patients with malignant GIB were included, of which 71 were
randomized to TC-325 and 55 to SET. While we found no statistical difference in rates
of rebleeding or primary hemostasis between the two modalities, it must be emphasized
that our patient population was heterogeneous, including a combination of patients
with benign and malignant etiologies of GIB. We believe that further studies are needed
to confirm if TC-325 indeed has equivalent efficacy compared to SET when considering
specific underlying etiologies of GIB.
There are several strengths to our analysis that are worth mentioning. First, we only
included RCTs to give us the most robust evidence of efficacy between TC-325 and SET.
Second, as part of our meta-analysis, we performed a comparative pair-wise analysis
of studies where outcomes of patients with oozing and/or spurting hemorrhage (Forrest
IA and IB) were reported and found that the two modalities did not differ in terms
of primary hemostasis and rebleeding. Third, we conducted a systematic literature
search with well-defined inclusion criteria, careful exclusion of redundant studies,
and inclusion of good quality studies with detailed extraction of data and rigorous
evaluation of study quality.
This study also has several limitations, most of which are inherent to any meta-analysis.
First, our analysis only included five RCTs of a heterogeneous population with various
GIB etiologies. In addition, we were unable to report outcomes separately for patients
with PUD and malignant GIB. Second, while a vast majority of patients in our analysis
with malignant GIB had an upper gastrointestinal source, three patients were included
with a lower GIB source. Third, 21 patients required repeat endoscopy to treat rebleeding
episodes (TC-325 10, SET 11), there were insufficient data to calculate pooled outcomes
of comparison between the two groups. Fourth, while all studies defined primary or
initial hemostasis as achieving successful endoscopic hemostasis within 3 to 5 minutes
of intervention during the index endoscopy, this outcome was defined differently in
the study by Lau et al. Fifth, in one of the included trials, only data regarding
rebleeding episodes, LOS, and rescue interventions was reported between the two modalities.
Outcomes regarding primary hemostasis were not reported [19]. Most of the trials included in our analysis were performed in similar geographic
locations, limiting the generalizability of our results. Finally, we were unable to
perform a cost-effectiveness analysis between TC-325 and SET, as the included trials
did not report information on the same.
Conclusions
Overall, our analysis demonstrates that compared to SET, TC-325 monotherapy may be
an acceptable therapeutic option for patients with acute non-variceal GIB, including
those with oozing/spurting hemorrhage from gastrointestinal malignancies and peptic
ulcer disease, and those lesions that are difficult to treat with SET. We found no
difference in the two modalities when comparing rates of rebleeding and LOS.
