Introduction
Colonic diverticular bleeding is a major form of lower gastrointestinal bleeding [1]
[2]
[3]. Although the bleeding stops spontaneously in most cases, endoscopic, radiologic,
or surgical treatment may be required if it persists [4]
[5]. When the source of bleeding is identified by colonoscopy, endoscopic hemostasis
can be performed to prevent recurrent bleeding and decrease the need for surgery [6]. Several endoscopic hemostatic treatments for colonic diverticular bleeding are
available, including endoscopic clipping, endoscopic band ligation (EBL), epinephrine
injection, and contact thermal therapy [6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]. However, they vary in effectiveness, and there is still no standardized therapy
for this condition. Because of available local expertise, our unit prefers EBL as
first-line therapy. Although endoscopic clipping is considered one of the most effective
treatments, the direct placement of hemoclips on the vessel can be technically challenging
because of a dome location or massive bleeding. Hemoclips can be placed indirectly
in a zipper fashion in these challenging cases; however, complete hemostasis is often
difficult to achieve [12].
Although EBL has been reported to be a safe and effective method [13]
[14]
[15]
[16]
[17]
[18]
[19], no studies on the feasibility of this technique have been published. The use of
EBL is gradually increasing, but reinsertion of the colonoscope after the EBL device
has been attached to its tip is considered cumbersome. In addition, the applicability
and safety of EBL when performed by non-expert endoscopists have not been sufficiently
verified. This study aimed to elucidate the feasibility of EBL when performed by non-expert
endoscopists and the possibility of EBL as a standard method for treating colonic
diverticular bleeding.
Patients and methods
Study population
A retrospective cohort study was conducted at St. Luke’s International Hospital, a
tertiary referral center in Tokyo, Japan, between June 2009 and October 2014. A total
of 108 patients with definite colonic diverticular bleeding and stigmata of recent
hemorrhage (SRH), such as active bleeding, a nonbleeding visible vessel, and adherent
clot [6], were treated during this period. Within the cohort, seven patients who had been
treated with transcatheter arterial embolization, epinephrine injection, or endoscopic
clipping were excluded. Epinephrine injection was used in six patients in whom the
diverticula could not be adequately suctioned into the hood of the endoscopic ligator
because of a small orifice. Two patients treated with epinephrine injection required
additional transcatheter arterial embolization for persistent bleeding. In one patient
with a diverticulum in the ascending colon, the orifice of the diverticulum with SRH
was so large that the endoscope could be passed into the diverticulum. Therefore,
endoscopic clipping was selected instead of EBL. Five patients were excluded because
the procedure time was not recorded. One patient was also excluded because EBL was
performed at two suspected sites of bleeding during the same intervention. A total
of 95 patients who had successful initial hemostasis with EBL were analyzed in this
study ([Fig. 1]).
Fig. 1 Flow diagram for all patients included in a study of endoscopic band ligation for
colonic diverticular bleeding performed by expert and non-expert endoscopists. SRH,
stigmata of recent hemorrhage; EBL, endoscopic band ligation.
Endoscopic band ligation methods
Bowel preparation with polyethylene glycol was performed before each examination to
achieve a high rate of SRH identification. In patients who were hemodynamically unstable
under intravenous fluid resuscitation, colonoscopy to identify the source of active
bleeding was done without bowel preparation. Colonoscopy was performed with a water-jet
scope (PCF-Q260AZI, PCF-Q260JI, or GIF-Q260J; Olympus, Tokyo, Japan). When a diverticulum
with SRH was identified, hemoclips (HX-610-135; Olympus) were applied as markers near
the diverticulum ([Fig. 2 a], [Fig. 2 b]). The endoscope was removed, and a band ligator device (MD-48710 EVL Device; Sumitomo
Bakelite, Tokyo, Japan) was attached to its tip ([Fig. 2 c]). The endoscope was then reinserted to the identified diverticulum. The diverticulum
was aspirated into the transparent hood of the band ligator device, and an elastic
O-ring was deployed ( [Fig. 2 d]) [13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]. Other endoscopic treatments, such as diluted epinephrine (1 : 20 000) injection
and endoscopic clipping, were considered for refractory bleeding when several EBL
attempts by experts had been unsuccessful. Patients were followed on an outpatient
basis at our institution for at least 30 days after EBL.
Fig. 2 a Endoscopic view of colonic diverticulum with active bleeding. b Marking with hemoclips near the diverticulum. c The colonoscope is removed, and a band ligator device is attached to its tip. d The colonoscope is reinserted, the diverticulum is aspirated into the transparent
hood of the band ligator device, and the elastic O-band is deployed.
Comparison of endoscopic band ligation treatments in the groups treated by expert
and non-expert endoscopists
Patients were classified into two groups: those treated by expert endoscopists and
those treated by non-expert endoscopists. The expert endoscopists included institutional
teaching staff of St. Luke’s International Hospital who were also board-certified
members of the Japanese Society of Gastrointestinal Endoscopy. The non-expert group
included trainees who had completed training in routine colonoscopic procedures. None
of the non-experts were board-certified members of the Japanese Society of Gastrointestinal
Endoscopy, but they had performed more than 500 colonoscopies before performing EBL
treatment. Because of the technical difficulties encountered in performing urgent
colonoscopies, trainees in our institution are generally required to complete 500
colonoscopies before performing urgent colonoscopies without on-site assistance by
highly experienced endoscopists. The decision regarding the selection of an operator
(expert or non-expert) was left to the discretion of the staff physician. As previously
stated, patients who were hemodynamically unstable with intravenous fluid resuscitation
underwent colonoscopy without bowel preparation. Because colonoscopy in a patient
without preparation may be difficult, with a low completion rate and impaired identification
of SRH, these difficult cases may have been assigned to the experts. EBL procedure
time (the time between marking the site of bleeding with hemoclips and completing
the O-band release) and total procedure time in these two groups were compared ([Fig. 3]). Safety was determined based on EBL-associated adverse events.
Fig. 3 Time frame of the endoscopic band ligation (EBL) procedure.
Statistical analysis
For bivariate analyses, Student’s t test and Fisher’s exact test were applied for continuous and categorical variables,
respectively. If a non-normal distribution was suspected for a continuous variable,
Wilcoxon’s rank sum test was used.
Linear regression analyses were used to determine possible factors affecting EBL procedure
time and total procedure time. A multivariate linear regression model was created
by including independent variables with P values of less than 0.2 in a simple linear regression model as well as including
clinically important variables. All 95 % confidence intervals (CIs) were two-sided.
P values of less than 0.05 were considered significant. Statistical analyses were conducted
with JMP version 9 (SAS Institute, Cary, North Carolina, USA). This study was approved
by the ethics committee of St. Luke’s International Hospital.
Results
EBL achieved successful immediate hemostasis in all 95 patients in our study cohort
(100 %). Six experts and six non-experts participated in the study. The median number
of years of endoscopy experience was 14.0 (range 5 – 18) for the experts and 3.5 (range
1 – 7) for the non-experts at the time of the EBL treatment. The mean (standard deviation
[SD]) number of EBL cases performed by the experts during the study period was 8 (6),
and the mean (SD) number performed by the non-experts was 8 (4).
The characteristics of the patients in the expert and non-expert groups are shown
in [Table 1]. Experts performed 47 of the 95 procedures (49.5 %). Comorbidities, medications,
shock index, hemoglobin level on admission, location of the bleeding diverticula,
rate of bowel preparation, total procedure time, and EBL-associated adverse events
were evaluated in each group. A significant difference was found between the rates
of bowel preparation in the two groups, with all the patients treated by non-experts
having undergone bowel preparation. The median EBL procedure time was 15 minutes (range
4 – 45) for the experts and 11 minutes (range 4 – 36) for the non-experts (P = 0.03). No significant difference was found between the total procedure times in
the two groups.
Table 1
Characteristics of the patients in a study of endoscopic band ligation (EBL) performed
by expert and non-expert endoscopists.
|
Expert endoscopists (n = 47)
|
Non-expert endoscopists (n = 48)
|
P value
[1]
|
|
Age, mean (SD), y
|
66 (14.2)
|
65 (14.1)
|
0.70
|
|
Male, n (%)
|
32 (68.0)
|
40 (83.3)
|
0.10
|
|
Hypertension, n (%)
|
26 (55.3)
|
23 (46.9)
|
0.54
|
|
Anticoagulant, n (%)
|
18 (38.3)
|
13 (27.1)
|
0.28
|
|
Shock index, mean (SD)
|
0.74 (0.25)
|
0.74 (0.22)
|
0.59
|
|
Hemoglobin, mean (SD), g/dL
|
12.2 (2.0)
|
12.4 (2.0)
|
0.52
|
|
Right-sided diverticula, n (%)
|
37 (78.7)
|
36 (75.0)
|
0.81
|
|
Bowel preparation, n (%)
|
35 (74.5)
|
48 (100)
|
0.0001
|
|
Total procedure time, median (range), min
|
33.0 (19 – 101)
|
33.5 (13 – 72)
|
0.63
|
|
EBL procedure time, median (range), min
|
15.0 (4 – 45)
|
11.0 (4 – 36)
|
0.03
|
SD, standard deviation.
1 Statistical significance was defined as P < 0.05.
The results of a linear regression analysis to investigate factors related to EBL
procedure time are shown in [Table 2]. With use of a simple linear regression model, performance of EBL by experts and
right-sided location of diverticula were factors that significantly increased EBL
procedure time. In addition, multivariate analysis adjusted for the rate of bowel
preparation, location of bleeding diverticula, and expertise in endoscopy revealed
that a right-sided location of diverticula was the factor most significantly affecting
EBL procedure time (β = 3.75; 95 %CI – 0.09 to 7.59; P = 0.05).
Table 2
Simple and multiple linear regression models to investigate factors related to endoscopic
band ligation procedure time.
|
Simple linear regression model
|
Multiple linear regression model
|
|
β
|
95 %CI
|
P value
[1]
|
β
|
95 %CI
|
P value
[1]
|
|
Age
|
– 0.02
|
– 0.12 to 0.08
|
0.68
|
|
|
|
|
Male
|
– 0.04
|
– 3.90 to 3.82
|
0.82
|
|
|
|
|
Bowel preparation
|
– 2.3
|
– 7.24 to 2.64
|
0.36
|
0.07
|
– 5.21 to 5.35
|
0.98
|
|
Right-sided diverticula
|
3.9
|
0.10 to 7.70
|
0.045
|
3.75
|
– 0.09 to 7.59
|
0.05
|
|
Expert endoscopist
|
3.1
|
– 0.10 to 6.30
|
0.06
|
2.95
|
– 0.53 to 6.43
|
0.09
|
β, beta coefficient; CI, confidence interval.
1 Statistical significance was defined as P < 0.05.
The results of linear regression analysis to investigate factors related to total
procedure time are reported in [Table 3]. Simple and multivariate linear regression models revealed no factors that significantly
increased total procedure time.
Table 3
Simple and multiple linear regression models to investigate factors related to total
procedure time.
|
Simple linear regression model
|
Multiple linear regression model
|
|
β
|
95 %CI
|
P value
[1]
|
β
|
95 %CI
|
P value
[1]
|
|
Age
|
– 0.01
|
– 0.12 to 0.08
|
0.91
|
|
|
|
|
Male
|
– 3.42
|
– 11.4 to 4.58
|
0.39
|
|
|
|
|
Bowel preparation
|
– 8.09
|
– 18.3 to 2.13
|
0.12
|
– 7.85
|
– 19.1 to 3.43
|
0.17
|
|
Right-sided diverticula
|
1.89
|
– 6.25 to 10.0
|
0.64
|
1.06
|
– 7.14 to 9.26
|
0.80
|
|
Expert endoscopist
|
2.14
|
– 4.72 to 9.00
|
0.53
|
0.10
|
– 7.32 to 7.43
|
0.98
|
β, beta coefficient; CI, confidence interval.
1 Statistical significance was defined as P < 0.05
Although all patients included in this study had successful initial hemostasis, 15
cases of rebleeding within 30 days occurred in this cohort. Five of these cases were
in patients treated by non-experts. In addition, in four cases more than one EBL attempt
was required because of inadequate suctioning into the endoscopic ligator hood. Three
of these patients requiring multiple EBL attempts were treated by experts, and one
was treated by a non-expert. Two patients had right-sided diverticula and two had
left-sided diverticula. Although up to four attempts were made by experts, one case
with a right-sided diverticulum required 28 minutes for EBL completion. The median
time with multiple EBL attempts was 21.5 minutes (range 6 – 28). There were no adverse
events related to EBL in either group.
Discussion
EBL achieves immediate hemostasis even in cases of massive bleeding. On the other
hand, it requires reinsertion of the colonoscope after a band ligator device has been
attached to its tip, which may be time-consuming. Furthermore, the endoscopic visual
field may become narrow by the attached band ligator device. For these reasons, EBL
can be technically challenging compared with other hemostatic techniques.
As expected, a right-sided location of diverticula significantly lengthened the EBL
procedure time because of the longer time required for reinsertion. Neither bowel
preparation nor expertise in endoscopy affected EBL procedure time in multivariate
analysis. Surprisingly, the expert endoscopists had longer EBL procedure times. It
is unclear why the experts required more time to complete EBL procedures. However,
one reason may be that experts were selected for the difficult cases with hemodynamic
instability and considerable co-morbidities, thus prolonging the procedural time.
As a result of technical improvements in EBL, non-experts are now able to learn EBL
technique faster and more efficiently compared with the initial experiences of experts.
In addition, in some cases experts supervised non-experts while they were performing
procedures. At any rate, reinsertion of the colonoscope for EBL is fast and simple
with less loop formation, so that advanced colonoscopy skills may not be required
if the procedure is done in a patient with adequate bowel preparation. We believe
that the EBL procedure can be performed with skills that are easily acquired.
We had 15 cases of rebleeding within 30 days in this cohort. In five of these cases,
EBL was performed by a non-expert. With this limited number of patients, it is difficult
to make assumptions about the cause of rebleeding; however, younger age, active bleeding
of SRH, and left-sided lesions are considered risk factors for rebleeding [20]. In four cases, more than one attempt at banding was required because of inadequate
suctioning into the endoscopic ligator hood. Even in these difficult cases, the median
EBL procedure time was 21.5 minutes (range 6 – 28). Therefore, repeated suctioning
attempts may not substantially affect EBL procedure time. Moreover, only single-use
band ligator devices were used in this study, whereas multiple-use band ligator devices
may further decrease the time required for repeated attempts at EBL. Across our cohort,
we did not experience any serious complications, such as perforation and infection.
Some limitations of our study merit discussion. First, this was a retrospective study
from a single institution, with a limited number of EBL cases performed by a limited
number of endoscopists. Second, there may have been some selection bias regarding
assignment of the endoscopists. Decisions regarding the selection of an operator (expert
or non-expert) were left to the discretion of the staff physician and difficult cases
may have been assigned to experts. In addition, although no change in endoscopist
was made after the lesion to be treated with EBL had been identified, there may have
been cases in which an expert took over to identify SRH, which may have affected the
total procedure time.
This is the first study to evaluate the possibility of standardizing the EBL procedure.
Our findings suggest that the acquisition of technical skills for EBL is relatively
straightforward and that the procedure can be completed in a short amount of time.
As long as the endoscopists are well trained in standard colonoscopy and the patients
have undergone bowel preparation, EBL is a feasible technique. We therefore conclude
that within the limitations of our retrospective study, EBL can be safely and effectively
performed by non-expert endoscopists.
