Introduction
Tumor-related gastrointestinal bleeding is currently a challenging clinical problem
[1 ]. A retrospective review of upper gastrointestinal bleeding in 55 patients with primary
or metastatic gastrointestinal malignancy revealed that up to 20 % were related to
tumor invading gastrointestinal lumen and causing bleeding [2 ]. Currently, endoscopy is recommended as the main diagnostic tool to locate the site
of bleeding [2 ]
[3 ]. Unfortunately, current endoscopic hemostatic methods including coaptation therapy,
argon plasma coagulation, and mechanical hemostasis do not reliably control active
bleeding, with rates of successful immediate hemostasis as low as 40 % and a significant
short-term rebleeding rate (up to 30 %) [1 ]
[4 ]
[5 ]. Surgery, embolization, and radiotherapy can serve as salvage hemostasis because
they are more effective (50 % – 100 %) with lower rates of rebleeding (0 – 18 %) [6 ]
[7 ]
[8 ]. However, a bridging endoscopic therapy is required during resuscitation and stabilization
of patients.
Hemospray is an inorganic powder not absorbed or metabolized by mucosal tissue. When
in contact with blood, its adhesive properties result in a physical barrier that covers
the bleeding site. The Hemospray barrier is stable because of the effect of accumulation
of clotting factors, and consequently the bleeding point is not exposed to acid, allowing
the healing process to continue [9 ]. Neither luminal nor systemic side effects have been reported with the product [9 ]
[10 ]
[11 ].
Recently, Hemospray has been proposed as a novel way of producing endoscopic hemostasis
for active gastrointestinal (gastrointestinal) bleeding, mostly for ulcer bleeding
[9 ]. To date, the study of Hemospray in tumor bleeding is limited as results have not
focused on this particular patient group [9 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]. This study aimed to compare the hemostatic efficacy of Hemospray with that of conventional
endoscopic treatment in patients who presented with upper gastrointestinal bleeding
from tumor.
Patients and methods
Patients
Between January 2014 and January 2015, patients with a history of upper gastrointestinal
bleeding from primary gastrointestinal malignancy or metastasis who presented at the
King Chulalongkorn Memorial Hospital were enrolled. The inclusion criteria were as
follows: 1) male or female patient, 18 years of age or older; 2) ability to provide
written informed consent; and 3) presence of active bleeding from tumor during endoscopy.
Patients were excluded if they had received any definitive treatment, such as surgery,
embolization or radiation within the first 48 hours (unable to evaluate the hemostatic
result of Hemospray). For the control group, the authors retrieved matched cases from
the hospital electronic database during the previous 5 years using the key words “upper
gastrointestinal bleeding AND cancer”. Tumor location was used as the criterion for
matching and the patients were further divided into those with upper gastrointestinal
tumor and those with hepatico-pancreaticobiliary tumor corresponding to the Hemospray
group.
In both groups, patient gender, age, type and stage of malignancy, Blatchford score,
amount of blood transfusions, number of endoscopies needed, and requirement for additional
interventions including endoscopy, surgery, adjuvant embolization, and radiotherapy
during admission were analyzed. The 14-day rebleeding rates, lengths of hospital stay
(LOS), and mortality rate at 30-day follow up were assessed. The study protocol and
consent form were approved by the Chulalongkorn University Institutional Review Board
(No.092 /58).
Instruments
Hemospray or TC-325 (Cook Medical, Winston-Salem, North Carolina, USA) was used as
the only hemostatic method with the maximum dose of 20 g (1 cartridge).
Endoscopic procedure
Medical resuscitation described elsewhere for upper gastrointestinal bleeding [16 ] was initiated in all eligible patients. Following medical resuscitation, therapeutic
esophagogastroduodenoscopy (EGD) was performed by an experienced endoscopist (RP)
certified in Hemospray endoscopic hemostasis. If tumor bleeding was identified, Hemospray
tube was applied through a 10-Fr catheter in which the powder was sprayed onto the
bleeding site until hemostasis was achieved, but no more than 1 cartridge was used.
Successful initial hemostasis was defined as no further active bleeding seen at least
5 minutes before withdrawing an endoscope after Hemospray was applied ([Fig.1 ]).
Then standard post-endoscopic care with a 72-hr proton pump inhibitor infusion was
prescribed in all patients [16 ]. Re-EGD was done in patients with suspected rebleeding, which was defined as: 1)
3 % of more drop in hematocrit level even after adequate blood transfusion; and 2)
new development of hematemesis or hematochezia or melena.
Statistical analysis
For numerical variables, the results were expressed as a mean ± SD, whereas other
quantitative variables are expressed as percentages. Continuous variables were compared
by the student’s t test. Discontinuous variables were compared by the chi-square (x2)
test. SPSS version 17.0 (SPSS (Thailand) Co., Ltd., Bangkok, Thailand) for Windows
systems was used for statistical analysis. Differences were considered significant
at the level of 0.05.
Results
During the study period, 14 patients experienced blood oozing from tumor and in all
cases, Hemospray was used as the first-line hemostatic therapy. Four of those patients
were excluded because they received either chemoembolization or radiation therapy
within 72 hours after Hemospray, although there was no evidence of rebleeding. Subsequently,
10 patients were eligible for analysis.
Historical controls were selected from the medical record of patients during 2010
and 2014. Seventeen patients were found to have presented with upper gastrointestinal
bleeding from tumor confirmed by endoscopy but only 10 of them were matched in terms
of bleeding location and bleeding activity of blood oozing and selected as a control
group (upper gastrointestinal tumor, n = 5 and hepatico-pancreaticobiliary tumor,
n = 5). Baseline characteristics including age, stage of tumor and Blatchford score
did not differ between the intervention and control groups ([Table 1 ]). The volume of blood transfused was higher in the Hemospray group (10.5 ± 16.0
vs. 4.9 ± 3.8 packs; P = 0.8). The 14-day rebleeding rate in the Hemospray group was 3 times lower than
the control group (10 % vs. 30 %; P = 0.60). Nine out of 10 patients in the Hemospray group did not undergo additional
intervention during the first 14 days. Although there was no rebleeding, on Day 12,
1 patient with bleeding gastric cancer who had previously been scheduled for elective
surgery underwent a partial gastrectomy ([Table 2 ]). Three of 10 patients (30 %) in the control group re-bled and required a rescue
intervention during the first 14 days and another 4 patients (40 %) underwent additional
intervention according to a prescheduled plan ([Table 3 ]). In all 3 patients, rebleeding occurred quite early (within 48 hours) ([Table 3 ]). LOS did not differ between the 2 groups (28.2 ± 21.2 days vs. 23.8 ± 12.5 days;
P = 0.26). The 30-day mortality rate in the Hemospray group was 3 times lower than
in the conventional therapy group (10 % vs. 30 %, P = 0.7) ([Table 1 ]). Moreover, no adverse effects were reported during the follow-up period.
Table 1
Baseline characteristics and results of treatment with Hemospray and with conventional
endoscopic treatment.
Parameters
Hemospray group (N = 10)
Conventional endoscopic treatment group (N = 10)
P value
Age (mean ± SD; years)
63.4 ± 10.4
60.3 ± 13.3
0.90
Sex (% male)
90
70
0.30
Blatchford scores (mean ± SD)
8.7 ± 2.9
7.9 ± 3.3
0.66
Advanced stage (%)
70
30
0.70
Total blood transfusion (mean ± SD and median; unit)
10.5 ± 16.0 and 5
4.9 ± 3.8 and 4
0.80
Number of total EGD (mean ± SD)
1.0 ± 0
1.4 ± 0.6
N/A
Additional intervention performed during first 10-day (%)
0
30
N/A
Additional intervention performed during admission (%)
10
70
0.30
14-day rebleeding rate (%)
10
30
0.60
Length of hospital stay (mean ± SD; day)
28.2 ± 21.2
23.8 ± 12.5
0.26
30-day mortality rate (%)
10
30
0.70
Table 2
Detailed course of patients in the Hemospray group.
Case
Sex
Age
Cancer
Advance stage
Blatchford
Score
Total PRC
Rescue treatment
14-day rebleed
30-day rebleed
Death
(day after spray)
LOS (days)
1
M
74
CCA
Yes
11
7
No
No
No
No
40
2
F
55
CCA
Yes
7
1
No
No
Yes, d 20
Yes, d 20
10
3
M
80
HCC
Yes
11
4
No
No
No
Yes, d 50
50
4
M
56
CA stomach
No
2
4
No
No
No
No
15
5
M
69
Metastatic to stomach
Yes
11
9
No
No
No
No
4
6
M
46
CA stomach
Yes
7
6
No
No
No
No
38
7
M
63
HCC
No
11
4
No
No
No
No
15
8
M
72
CA stomach
Yes
7
55
No
Yes, no EGD
Yes, no EGD
Yes, d 70
70
9
M
56
CA stomach
No
9
13
Yes, Surgery d 12
No
No
No
30
10
M
63
HCC
Yes
11
2
No
No
No
No
10
CA, carcinoma; CCA, cholangiocarcinoma; HCC: hepatocellular carcinoma; PRC, packed
red cells; LOS, length of hospital stay; N/A, no data available
Table 3
Detailed course of patients in the control group.
Control
Sex
Age
Cancer
Advance stage
Blatchford score
Total PRC
Endoscopic Rx
Rescue treatment
14-day rebleed
30-day rebleed
Death (day after admission)
LOS
1
F
43
HCC
Yes
5
5
No
Yes, TACE at d 1
Yes, d 1
N/A
Yes, d 9
9
2
F
57
CA HOP
No
10
3
Yes, Adrenaline injection
Yes, Surgery at d 40
No
No
Yes, d 50
50
3
F
79
CA GB
No
8
4
Yes, APC
No
N/A
N/A
No
10
4
M
57
CA stomach
No
4
4
No
Yes, Surgery at d 20
No
No
No
30
5
M
67
CA stomach
No
2
1
No
Yes, Surgery at d 15
No
No
No
28
6
M
45
CA stomach
Yes
9
15
No
No
Yes, d2, APC
N/A
Yes, d 16
16
7
M
65
CA ampulla
Yes
7
5
No
Yes, RT at d 2
Yes, d2, RT
No
No
21
8
M
83
CA stomach
No
12
6
No
No
No
No
No
24
9
M
50
CA esophagus
No
11
4
No
No
No
No
No
15
10
M
57
CA HOP
No
11
2
No
Yes, Surgery at d 6
No
No
No
36
CA, carcinoma; HCC, hepatocellular carcinoma; HOP, head of pancreas; GB, gallbladder;
PRC, Packed red blood cells; APC, argon plasma coagulation; TACE, transarterial chemoembolization;
RT, radiation therapy; LOS, length of hospital stay; d, day; N/A, no data available
Discussion
Because most incidents of active tumor bleeding are difficult to control with conventional
endoscopic hemostatic techniques such as coaptation, hemoclipping, and band ligation,
[1 ]
[4 ]
[5 ] there is a need for bridging therapy. Hemospray is well suited for treatment of
upper gastrointestinal bleeding from tumor because it provides temporary hemostasis
while allowing time to schedule a patient for more definitive treatment such as surgery,
angiographic embolization, or radiation. The mechanism of tumor bleeding is explained
by erosion of the raw surface of a malignant lesion. In addition, acidic content from
the stomach can promote more bleeding because it dissolves clot and possibly digests
tumor tissue that lacks a barrier of mucous and epithelium [4 ]. Although conventional endoscopic hemostatic techniques can provide immediate control
for some tumor bleeding, it tends to recur in a short period of time [1 ]
[4 ]
[5 ]. It is possible that the conventional endoscopic treatments may fail to protect
tumor tissue from digestion. In addition, if coaptation is selected, tumor necrosis
may progress because of the effect of heat [12 ]. In contrast, Hemospray treatment can prevent tumor bleeding by providing immediate
hemostasis and the powder that remains on the surface for a period of time may protect
the tumor tissue from further erosion by gastric acid.
In recent years, multiple studies have been conducted to prove the efficacy of Hemospray
for many different types of gastrointestinal bleeding [9 ]
[10 ]
[13 ]
[14 ]
[15 ] but only 1 study focused on tumor-related gastrointestinal bleeding is available
[10 ]. In 2012, Chen et al. reported a case series of upper gastrointestinal bleeding
from different types of tumors which was successfully controlled by Hemospray [10 ]. In their series, there were 3 gastric cancers, 1 pancreatic cancer invading duodenum,
and 1 metastatic breast cancer invading duodenum. Four of 5 patients (80 %) subsequently
received chemotherapy or radiotherapy within 5 days after initial hemostasis with
Hemospray although there was no evidence of rebleeding during the 5-day observation
[10 ]. Based on experience in a very limited number of patients, the authors concluded
that Hemospray may be useful for both immediate hemostasis and as a bridging treatment
for further adjuvant therapy.
To our knowledge, this is the first study that compares the hemostatic effect of Hemospray
with conventional treatment for upper gastrointestinal bleeding from tumor. To demonstrate
the longer hemostatic effect of Hemospray in bleeding tumor, the current study excluded
patients who received definitive treatment within 72 hours. We could not, however,
demonstrate statistical significance given the limited number of cohorts but we did
demonstrate a trend toward lower rates of recurrent bleeding in 14 days (10 % vs.
30 %, P = 0.6) and a lower mortality rate at 30 days (10 % vs. 30 %, P = 0.7). In addition, no additional treatment was needed after Hemospray application
in all except 1 patient during hospitalization whereas 30 % of patients in the control
group required additional treatment.
The limitation of the present study is a rather small number of cases, which made
it impossible to make a pair match in other parameters apart from the location of
tumor. In addition, we believe that more patients would have allowed the results to
rise to statistical significance, as we showed a large gap in outcomes between the
2 groups. Furthermore, no patients had bleeding from a lower gastrointestinal tract
tumor. Because little acid is involved in rebleeding from such tumors, the efficacy
of Hemospray in those patients needs to be demonstrated and its differing mechanism
of action in that setting explained.
Conclusions
Hemospray is a promising therapy for initial hemostasis in upper gastrointestinal
bleeding from tumor because it can achieve hemostasis during the first 14 days, which
may allow sufficient time before consideration of appropriate additional intervention.
Fig. 1 Pictures of active tumor bleeding (a ) and post Hemospray (b ).
