Keywords Endoscopy Lower GI Tract - Lower GI bleeding - Diagnosis and imaging (inc chromoendoscopy,
NBI, iSCAN, FICE, CLE...) - Polyps / adenomas / ...
Introduction
Rates of colonic diverticular bleeding (CDB) are increasing in Japan, with patients
hospitalized for bloody stools accounting for approximately 60% of cases [1 ]. Although spontaneous hemostasis is common in this condition and mortality risk
is low, the high rebleeding rate remains a clinical concern. Methods of endoscopic
hemostasis include clipping, ligation, local injection, and cauterization. However,
a recent large-scale Japanese cohort study (CODE BLUE-J Study) has found that, in
real-world clinical practice, clipping (57.2%) and ligation (41.0%) are most commonly
used [2 ]. Clipping can be either direct or achieved via a diverticulum closure [3 ]. Meanwhile, ligation is broadly divided into endoscopic band ligation (EBL) and
endoscopic detachable snare ligation [4 ]. Some recent studies have suggested that EBL may be more effective than clipping
at preventing both early (within 30 days) and late (30 days to 1 year) rebleeding
[5 ]
[6 ]
[7 ], especially in cases involving active bleeding and right colon bleeding [2 ]. Although several previous studies have compared the efficacy of either method alone,
combined use of these methods has not been assessed. We have previously reported on
combined use of EBL and clipping (EBL-C) for a bleeding diverticulum, which involves
hemostasis by direct clipping with a repositionable clip, followed by aspiration and
bowel inversion for each clip [8 ]. Since November 2021, our first-choice method of hemostasis has been EBL-C, except
for patients at risk of perforation due to long-term steroid therapy or maintenance
dialysis. In this study, we aimed to compare clinical outcomes of EBL-C with those
of EBL.
Patients and methods
Patients and study design
From March 2019 to July 2024, we conducted a retrospective review of medical records
from 138 patients with CDB definitively diagnosed with stigmata of recent hemorrhage
who underwent endoscopic hemostasis procedures at our department. Procedures included
clipping (n = 50), EBL (n = 24), EBL-C (n = 56), and over-the-scope clipping (OTSC)
(n = 8). We analyzed 80 cases treated with ligation procedures (EBL or EBL-C). EBL
and EBL-C involve band ligation, and patients with a risk of perforation associated
with ligation (e.g., those on maintenance dialysis or long-term steroid therapy) were
excluded. In addition, cases in which diverticulum inversion was technically challenging
were excluded due to difficulty of performing ligation. These patients were primarily
treated using the clipping method.
We investigated overall treatment outcomes in patients who underwent ligation procedures
(EBL or EBL-C). Subsequently, we compared patient characteristics and treatment outcomes
between the EBL-C and EBL groups. Patient characteristics were comparable between
both groups, including age, sex, underlying conditions, and medication use (e.g.,
antiplatelet agents, nonsteroidal anti-inflammatory drugs) ([Table 1 ]).
Table 1 Characteristics of patients treated with EBL or EBL-C.
EBL (n = 24)
EBL-C (n = 56)
P value
* Recent hemostasis is defined as hemostasis to the same diverticulum within 30 days.
† Definitive cases: Treatment history of the same diverticulum was confirmed.
‡ Possible cases: Treatment history of the same diverticulum could not be ruled out.
§ Past hemostasis is defined as hemostasis to the same diverticulum more than 30 days
prior.
¶ Fisher's exact test.
CT, computed tomography; DAPT, dual antiplatelet therapy; EBL, endoscopic band ligation;
EBL-C, endoscopic band ligation plus clipping; NSAID, nonsteroidal anti-inflammatory
drug; PEG, polyethylene glycol.
Age ≥ 70 years, n (%)
17 (70.8)
40 (71.4)
1¶
Sex, male, n (%)
15 (62.5)
27 (48.2)
0.329¶
Hypertension, n (%)
16 (66.7)
38 (67.9)
1¶
Antiplatelet agent, n (%)
6 (25)
19 (33.9)
0.599¶
Anticoagulant, n (%)
2 (8.3)
6 (10.7)
0.266¶
DAPT, n (%)
1 (4.17)
8 (14.3)
0.266¶
NSAIDs, n (%)
5 (20.8)
8 (14.3)
0.516¶
Bowel preparation (use of PEG solution), n (%)
12 (50)
22 (39.3)
0.461¶
Extravasation on contrast-enhanced CT, n (%)
9/20 (45)
30/49 (61.2)
0.286¶
Recent hemostasis* (< 30 days)† + possible cases‡ ) n (%)
2 (8.3)
3 (5.4)
0.633¶
Past hemostasis§ (> 30 days)
3 (0)
11 (3.6)
1¶
Location, left-side colon, n (%)
9 (26.8)
15 (36)
0.426¶
Use of long attachment cap, n (%)
22 (91.7)
53 (94.6)
0.633¶
Methods and devices
Identifying the responsible diverticulum is crucial. Therefore, performing bowel preparation
within 24 hours and conducting the examination using a long hood attachment are recommended
[9 ]
[10 ]
[11 ]. In addition, the identification rate increases when contrast-enhanced computed
tomography (CT) shows evidence of extravasation [12 ]
[13 ]. Therefore, our hospital prioritizes performing contrast-enhanced CT whenever CDB
is suspected. We recommend endoscopic examination after bowel preparation with polyethylene
glycol within 24 hours. However, when extravasation is clearly visible on contrast-enhanced
CT scans or the patient is hemodynamically unstable, emergency endoscopy may be conducted
without bowel preparation at the attending physician’s discretion.
We used a water-jet scope, fitted with a long attachment cap, a SureClip (Micro-Tech
Co., Nanjing, China), and an EBL device (Sumitomo Bakelite Co., Ltd., Tokyo, Japan).
EBL-C procedure
EBL-C was conducted as follows ([Fig. 1 ]) ([Video 1 ]).
Fit the long attachment cap. Once the bleeding diverticulum is identified, perform
hemostasis by direct clipping (SureClip; Micro-Tech Co).
Withdraw the scope, fit the EBL device (Sumitomo Bakelite Co., Ltd.), and reinsert
it. Suction the bleeding diverticulum containing the clip as far as possible, then
release the O-ring. If suction and inversion are sufficient, ligation with the O-ring
will have caused formation of a neck, indicating effective ligation ([Fig. 2 ]).
Fig. 1 Schema of the EBL-C method.
Endoscopic hemostasis of colonic diverticular bleeding by combining endoscopic band
ligation (EBL) and clipping.Video 1
Fig. 2 Success and failure of the EBL-C procedure.
Statistical analysis
We used Fisher’s exact test for comparisons of categorical variables and a Mann–Whitney
U test of the comparisons of continuous variables between groups. The association
between endoscopic treatment and clinical outcomes was analyzed using univariate and
multivariate logistic regression models. P < 0.05 was considered statistically significant. All statistical analyses were performed
with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan).
Definition of terms
We focused on presence or absence of treatment history for the same diverticulum.
In cases of early rebleeding (within 30 days), identifying the same diverticulum was
often possible based on evidence of prior procedures, such as residual clips or ulcers
formed by EBL.
Conversely, identifying the same diverticulum after a longer period is often more
challenging, but anatomical landmarks like the appendix or Bauhin's valve, as well
as scarring, could aid determination. In addition, contrast-enhanced CT or endoscopy
sometimes confirmed bleeding at a different site, ruling out involvement of the same
diverticulum. Thus, we distinguished between definitive cases, in which treatment
history of the same diverticulum was confirmed, and possible cases, in which involvement
of the same diverticulum could not be ruled out.
Results
Among patients who underwent ligation procedures (EBL or EBL-C) (n = 80), 11 cases
(13.6%) experienced early rebleeding (within 30 days). In multivariate analysis of
risk factors for early rebleeding, failure of neck formation was the only independent
risk factor for rebleeding (adjusted odds ratio [OR] 0.076; 95% confidence interval
[CI] 0.015–0.398; P = 0.0023) ([Table 2 ]). Neck formation was achieved in 66 cases (82.5%) and undergoing EBL-C was the only
independent factor contributing to successful development of neck formation (adjusted
OR 7.01; 95% CI 1.41–34.8; P = 0.0095) ([Table 3 ]).
Table 2 Early rebleeding risks after ligation therapy on logistic regression analyses.
Early rebleeding
No rebleeding
Crude OR
P value
Adjusted OR
P value
CI, confidence interval; DAPT, dual antiplatelet therapy; EBL-C, endoscopic band ligation
plus clipping; NSAID, nonsteroidal anti-inflammatory drug; OR, odds ratio.
Age ≥ 70 years, n (%)
9 (81.8)
48(69.6)
1.97
0.411
Sex, female, n (%)
3 (27.3)
3(27.3)
2.75
0.16
1.92
0.48
Hypertension, n (%)
9 (81.8)
45(65.2)
2.4
0.287
Diabetes, n (%)
1 (9.1)
14(20.1)
0.393
0.392
Antiplatelet agent, n (%)
3 (27.3)
22(31.9)
0.801
0.76
Anticoagulant, n (%)
2 (18.2)
6(8.7)
2.38
0.342
DAPT, n (%)
1 (9.1)
8(11.6)
0.763
0.808
NSAIDs, n (%)
3 (27.3)
10(14.5)
2.21
0.295
Bowel preparation, n (%)
3 (27.3)
31(44.9)
0.46
0.28
Recent hemostasis (< 30 days) (definitive + possible), n (%)
2 (18.2)
3(4.3)
4.89
0.105
4.22
0.322
Past hemostasis (> 30 days) (definitive + possible), n (%)
2 (18.2)
12(17.4)
1.06
0.949
Location, left-side colon, n (%)
5 (45.5)
19(27.5)
2.19
0.236
Use of long attachment cap, n (%)
9 (81.8)
66(95.7)
4.89
0.105
1.93
0.597
Neck formation (+), n (%)
4 (36.4)
62(89.9)
0.0645
0.000225
0.076
0.00227
EBL-C, n (%)
5 (45.5)
51(73.9)
0.294
0.0656
0.59
0.509
Table 3 Factors associated with neck formation after ligation therapy on logistic regression
analyses.
Neck formation (+)
Neck formation (-)
Crude OR
P value
Adjusted OR
P value
DAPT, dual antiplatelet therapy; EBL-C, endoscopic band ligation plus clipping; NSAID,
nonsteroidal anti-inflammatory drug; OR, odds ratio.
Age ≥ 70 years, n (%)
45 (68.1)
12 (85.7)
0.357
0.203
0.297
0.238
Sex, female, n (%)
35 (53.0)
3 (21.4)
0.242
0.0414
0.3
0.195
Hypertension, n (%)
41 (62.1)
13 (92.9)
0.126
0.0526
0.0682
0.0879
Diabetes, n (%)
13 (19.7)
2 (14.3)
1.47
0.639
Antiplatelet agent, n (%)
18 (27.3)
7 (50)
0.375
0.103
0.794
0.796
Anticoagulant, n (%)
5 (7.6)
3 (21.4)
0.301
0.133
0.183
0.114
DAPT, n (%)
8 (12.1)
1 (7.1)
1.79
0.597
NSAIDs, n (%)
11 (16.7)
2 (14.3)
1.2
0.827
Bowel preparation, n (%)
29 (43.9)
5 (35.7)
1.41
0.573
Recent hemostasis (< 30 days)
2 (3.0)
3 (21.4)
0.115
0.0254
0.126
0.131
Past hemostasis (> 30 days)
13 (19.7)
1 (7.1)
3.19
0.284
Location, left-side colon, n (%)
20 (30.3)
4 (28.6)
1.09
0.898
Use of long attachment cap, n (%)
3 (4.5)
2 (14.3)
0.286
0.195
0.394
0.431
EBL-C, n (%)
50 (75.6)
6 (42.9)
4.17
0.0196
7.01
0.0095
The early rebleeding rate (rebleeding within 30 days) was lower in the EBL-C group
than in the EBL group, although this difference was only marginally non-significant
(8.9% vs. 25.0%, P = 0.0776). Over the 1-year follow-up period, the late rebleeding rate (rebleeding
between 30 days and 1 year) was also lower in the EBL-C group than in the EBL group,
but the difference was non-significant (2.4% vs. 8.3%, P = 0.0549). Conversely, the frequency of neck formation, considered a factor potentially
reducing rebleeding risk, was significantly higher in the EBL-C group (EBL-C: 89.3%
vs. EBL: 66.7%, P = 0.0235). There were no significant differences in use of red blood cell transfusion,
treatment time, or hospitalization duration ([Table 4 ]).
Table 4 Outcomes of patients treated with EBL or EBL-C.
EBL
EBL-C
P value
*Early rebleeding: rebleeding that occurs within 30 days following the initial hemostasis.
† Late rebleeding: rebleeding that occurs between 30 days and 1 year (or 2 years) following
initial hemostasis.
‡ Time to hemostasis was defined as time to hemostasis complete after identifying the
bleeding site.
§ EBL procedure time was defined as time to hemostasis complete after initial clipping.
¶ Fisher's exact test.
†† Mann–Whitney U test.
EBL, endoscopic band ligation; EBL-C, endoscopic band ligation plus clipping.
Neck formation, n (%)
16 (66.7)
50 (89.3)
0.0235
¶
Early rebleeding*, n (%)
6 (25)
5 (8.9)
0.0776¶
Late rebleeding within 2 years† , n (%)
2/24 (8.3)
1/27 (3.7)
0.187¶
Late rebleeding within 1 year† , n (%)
2/24 (8.3)
1/41 (2.4)
0.0549¶
Time to hemostasis‡ , median (range), min
20 (7–46)
21 (7–47)
0.871††
EBL procedure time§ , median (range), min
15 (7–25)
15 (4–42)
0.617††
Transfusion, n (%)
11 (45.8)
29 (51.8)
0.808¶
Length of hospital stay after hemostasis, median (range), day
6 (4–23)
5 (1–14)
0.208††
Complications, n (%)
0 (0)
0 (0)
NA
Rebleeding risk was examined in the EBL-C group. A definitive history of treatment
for the same diverticulum over 30 days prior was significantly associated with higher
risk of rebleeding. No other background factors were found to be significant ([Table 5 ]).
Table 5 Patient-related risk factors for early rebleeding after EBL-C on Fisher‘s exact test.
Early rebleeding
No rebleeding
P value*
*Fisher’s exact text.
CT, computed tomography; DAPT, dual antiplatelet therapy; EBL-C, endoscopic band ligation
plus clipping; NSAID, nonsteroidal anti-inflammatory drug.
Age ≥ 70 years, n (%)
4 (80)
36 (70.6)
1
Sex, male, n (%)
3 (60)
20 (39.2)
1
Hypertension, n (%)
5 (100)
33 (64.7)
0.164
Diabetes, n (%)
0 (0)
12 (23.5)
0.574
Antiplatelet agent, n (%)
2 (40)
17 (33.3)
1
Anticoagulant, n (%)
2 (40)
4 (7.8)
0.0836
DAPT, n (%)
1 (20)
7 (13.7)
0.522
NSAIDs, n (%)
1 (20)
7 (13.7)
0.552
Bowel preparation, n (%)
1 (20)
21 (41.2)
0.638
Early colonoscopy, n (%)
3 (60)
36 (70.6)
0.634
Extravasation on contrast-enhanced CT, n (%)
4/5 (80)
26/44 (59.1)
0.636
Use of long attachment cap, n (%)
4 (80)
49 (96.1)
0.249
Recent hemostasis (definitive), n (%)
0 (0)
3 (5.9)
1
Recent hemostasis (definitive + possible), n (%)
0 (0)
3 (5.9)
1
Past hemostasis (definitive), n (%)
2 (40)
0 (0)
0.00649
Past hemostasis (definitive + possible), n (%)
2 (40)
9 (17.6)
0.251
Location, right-side colon, n (%)
3 (60)
38 (74.5)
0.602
Among procedure-related factors, rates of no neck formation and insufficient clipping
were higher in the early rebleeding group than in the comparison group ([Table 6 ]).
Table 6 Procedure-related risk factors for early rebleeding after EBL-C on Fisher’s exact
test.
Early rebleeding
No rebleeding
P value
*Cases of insufficient hemostasis with initial clipping.
† Fisher's exact test.
EBL-C, endoscopic band ligation plus clipping.
Neck formation (-), n (%)
3 (60)
3 (5.9)
0.00661
*
Insufficient clipping)† , n (%)
3 (60)
4 (7.8)
0.0112
*
Among the EBL-C patients (n = 56), 44 (78.6%) successfully achieved both clip hemostasis
and neck formation, with none experiencing early rebleeding (0/44). However, 11 patients
(19.6%) had either unsuccessful clip hemostasis or neck formation, and early rebleeding
occurred in four of them (36.4%). In addition, one patient (1.8%) had both unsuccessful
clip hemostasis and neck formation and experienced early rebleeding. No complications
occurred in either group.
Regarding rebleeding cases, rebleeding from the same diverticulum was observed in
eight cases in the EBL group (early: 6, late: 2) and six in the EBL-C group (early:
5, late: 1). Methods of hemostasis for rebleeding EBL cases were as follows: for the
six early rebleeding cases, one was treated with EBL-C, three with clipping, one achieved
spontaneous hemostasis, and one was treated with OTSC. For the two late rebleeding
cases, one was treated with EBL and one achieved spontaneous hemostasis. For rebleeding
EBL-C cases, hemostasis methods were as follows: for the five early rebleeding cases,
four were treated with OTSC and one with clipping. For the one late rebleeding case,
EBL-C was performed. None of the five rebleeding cases treated with OTSC experienced
rebleeding again.
In this study, there were seven definitive cases of treatment history in the same
diverticulum.
There were five definitive cases of recent treatment history (within 30 days), including
four treated with clipping and one with EBL. Among the clipping cases, the clips had
dislodged in two cases and remained in place in two cases. The bands had dislodged,
resulting in ulcer formation. There were two definitive cases of past treatment history
(over 30 days prior), including one treated with EBL and one with EBL-C. In both definitive
cases of past treatment history (over 30 days prior), clips and bands had dislodged,
and although scarring was observed, the diverticulum itself remained. Both cases were
treated with EBL-C, but early rebleeding occurred in both cases.
Discussion
Unsuccessful neck formation was a risk factor for early rebleeding in ligation procedures.
In EBL-C, which involves performing EBL after placing a clip, the rate of neck formation
was higher compared with EBL alone, suggesting a potential reduction in rebleeding
risk.
EBL-C is a simple procedure with several advantages. First, it involves hemostasis
at two different parts of the vessel using different methods: clipping the exposed
part of the vessel and using EBL on the non-exposed part. This dual approach enhances
the hemostatic force, potentially increasing the effectiveness of rebleeding prevention.
Conventional EBL involves marking the vicinity of the bleeding diverticulum, which
may prove difficult to identify while placing the EBL device and reinserting the scope.
In EBL-C, because the bleeding diverticulum itself is clipped, there is no chance
that it will be misidentified or overlooked, and EBL can be conducted with confidence.
In addition, in conventional EBL, after the vicinity of the diverticulum is marked
with a clip, it continues to bleed while the scope is being temporarily withdrawn
to place the EBL device before reinserting the scope, which can cause hemodynamic
fluctuations and obscure the visual field. In EBL-C, because the bleeding diverticulum
itself is clipped to stop the bleeding, hemostasis is usually achieved, and thus,
the operator can continue the procedure without feeling under pressure.
Depending on the location of the responsible diverticulum, obtaining a frontal view
may be difficult. With conventional EBL, the surrounding mucosa may be aspirated,
making it challenging to invert the diverticulum. By performing EBL after clipping,
we have managed multiple cases in which the clip acted as an axis, making it easier
to identify the diverticulum center and perform aspiration. These experiences formed
the basis for developing the EBL-C method.
Some concerns involve risk that presence of a clip within the diverticulum in EBL-C
might make it difficult to fully pull the diverticulum into the EBL device during
suction. However, in this study, the EBL-C group showed a significantly higher rate
of neck formation (EBL-C 89.3% vs. EBL 66.7%, P = 0.0235). In practice, we seldom had the impression during the procedure that the
diverticulum could not be fully pulled into the device. This suggests that presence
of the clip may support sufficient suction by aligning the axis of aspiration.
Although we used maximum suction strength, we have not experienced a case in which
the clip was dislodged by this suction. This procedure may require use of clips with
a strong gripping force, such as a SureClip. Evidence from comparative studies involving
other clips is required.
EBL-C patients who developed rebleeding had insufficient clipping and unsuccessful
neck formation. Early rebleeding did not occur in any patients without these characteristics,
and their risk of rebleeding can be regarded as extremely low. Conversely, if either
of these characteristics is present, risk of rebleeding is high, and additional preventive
treatment should be considered. Furthermore, both patients with definitive histories
of past hemostasis for the same diverticulum experienced rebleeding. Absence of neck
formation and a history of past hemostasis for the same diverticulum may both reflect
hardness of the diverticulum.
EBL is considered challenging to perform when the diverticulum is hard because it
makes suction and inversion difficult. In such cases, OTSC (Ovesco Endoscopy, Tübingen,
Germany) reportedly is effective [14 ]
[15 ]. In this study, we conducted hemostasis using OTSC for patients who developed early
rebleeding after ligation procedures (either EBL or EBL-C), with good outcomes. Data
on use of OTSC for CDB are scarce and OTSC is associated with risk of complications,
high cost, and extended treatment time. Further studies are required to verify whether
OTSC should be recommended for all patients with risk factors for rebleeding.
There are many reports evaluating risk of rebleeding based on patient characteristics
and diverticular features before treatment [16 ]
[17 ]
[18 ]
[19 ]. However, few studies have assessed risk of rebleeding according to procedure-related
factors, such as post-ligation morphology. The only report examining complete eversion
as a risk factor for bleeding showed a trend toward lower frequency of complete eversion
in the rebleeding group, although no statistically significant differences were observed
[20 ]. If the rebleeding risk could be assessed immediately after the procedure, this
would be significant for CDB treatment, which has a higher risk of rebleeding and
hemodynamic instability compared with other conditions.
Limitations of this study include a retrospective single-center design and small sample
size. In addition, this study may be affected by sampling bias because treatment methods
varied across different time periods. To further prove its efficacy, it will be necessary
to conduct multicenter, randomized, controlled trials and accumulate evidence.
Conclusions
In conclusion, EBL-C for CDB may be more effective in preventing rebleeding than EBL
alone because it facilitates better ligation of the target diverticulum.
Bibliographical Record
