Introduction
Depiction of extravasation on contrast-enhanced computed tomography (CECT) is definitive
evidence of colonic diverticular bleeding and provides endoscopists with information
on the region of interest for colonoscopic investigation. Identification of an extravasation
on CECT increases the detection rate of a bleeding diverticulum by colonoscopy to
50 % to 68.3 %, compared to 20.2 % to 36.3 % in the absence of an observable extravasation
[1]
[2]. However, even with positive CECT information, the detection rate of a bleeding
diverticulum remains unsatisfactory when spontaneous cessation of bleeding occurs.
Moreover, as colonoscopy does not provide precise positional information, compared
to computed tomography (CT), detectability of the target diverticulum during colonoscopy
depends on the skill of the endoscopist and on chance.
The “step-clipping” method overcomes this limitation by making artificial signposts
in the colon using endoclips, thus providing the positional relationship between the
clips and the target diverticulum responsible for the bleeding [3]. In this way, the endoscopist can navigate to the target site by referring to the
positional information from each clip, rather than performing a time-consuming search
in cases of spontaneous cessation of bleeding at the target diverticulum. Accordingly,
the aim of our study was to evaluate the usefulness of the step-clipping method for
endoscopic navigation to a target diverticulum.
Patients and methods
Patients
Prospective patients were those presenting with a chief complaint of hematochezia
at our hospital, between June 2017 and January 2019. Of them, those with evidence
of extravasation on CECT, but with absence of active bleeding on subsequent colonoscopy,
were candidates for the step-clipping method. Patients for whom bowel preparation
or colonoscopy examination were contraindicated due to extreme age or multiple comorbidities
were excluded. We also excluded patients for whom CECT was contraindicated due to
renal malfunction or allergy to the enhancement agent.
Step-clipping method
CT imaging was performed using a 64-slice diagnostic imaging technique on a Somatom
Definition Flash system (Siemens Healthineers, Erlangen, Germany). For patients with
an extravasation detected on CECT ([Fig. 1]), an initial colonoscopy was performed using a PCF-Q260AZI endoscope (Olympus, Tokyo,
Japan), under CO2 insufflation, after antegrade bowel preparation using 2 liters of polyethylene glycol
(PEG). When a site of active bleeding was not identified, several clips were placed,
2 inches apart, including at the site of observed extravasation on CECT images. The
range and number of clips deployed was dependent on the certainty of the location,
which was determined in reference to the presumptive location of the cecum, hepatic
flexure, splenic flexure, and junction of the sigmoid-descending colon on transluminal
viewing colonoscopy, with this wide region certainly including the specific region
of interest. The placement of endoclips was performed with minimal search in order
not to change the environment of the colon ([Fig. 2]). After placement, the scope was promptly removed and unenhanced CT imaging was
conducted 2 hours later. On the unenhanced CT image, the positional relationship between
each clip and the target diverticulum, identified on the initial CECT, was evaluated
using 1-mm axial image slices, with coronal or sagittal plane images added as needed
([Fig. 3]). The target diverticulum was tracked and identified on the CT images in reference
to the neighboring symbolic structures (e. g., fecal-impacted diverticulum and large-sized
diverticulum). CT images were reviewed and interpreted by two endoscopists, each of
whom had previously performed more than 1000 colonoscopies, with a conclusive diagnosis
reached by agreement. Based on the evaluation, secondary colonoscopy was conducted
to identify the target diverticulum ([Fig. 4]). All endoscopic procedures were conducted during the day, regardless of the time
of admission. No fluoroscopy was used.
Fig. 1 Extravasation observed on contrast-enhanced computed tomography images obtained on
the day of admission. a Axial image showing the bleeding (target) lesion in the ascending colon, among numerous
diverticula (yellow arrow). b Sagittal image showing the target lesion located in the middle of the ascending colon
(yellow arrow).
Fig. 2 The step-clipping method performed at initial colonoscopy. a Marking clips, placed 2 inches apart, including the suspected site. b The placement scheme of the marking clips.
Fig. 3 Comparison of the sagittal computed tomography images before and after step-clipping.
a Contrast-enhanced image obtained on the day of admission (before step-clipping).
The target lesion is located in the middle of the ascending colon (yellow arrow).
b Unenhanced computed tomographic image obtained after step-clipping, with the target
diverticulum identified opposite to the fourth clip (yellow arrow).
Fig. 4 Identification of the target lesion on secondary colonoscopy. a Based on evaluation of computed tomography images, the colonoscope was immediately
navigated to the suspicious site, located opposite to the fourth clip (blue arrow).
b The target lesion was successfully identified.
Statement of ethics
This study conformed to the principles of the 6th revision of the Declaration of Helsinki (2008), and the study protocol was approved
by the institutional review board (IRB) of the Hiroshima City Asa Citizens Hospital,
which also granted us permission to access the patients’ information on April 19,
2019 (IRB No. 31-1-10). Owing to the retrospective design, the need for patient consent
was waived.
Results
Over the defined study period, 89 patients were admitted to our hospital with a diagnosis
of colonic diverticular bleeding. Extravasation on CECT was observed in 25 of these
patients. Of them, four were excluded for the following reasons: contraindication
to colonoscopy due to extreme age (n = 2) and active bleeding during initial colonoscopy
(n = 2). Therefore, our analysis included data from 21 consecutive patients (14 men
and 7 women; mean age, 73.2 years). Ten patients (48 %) presented with severe bleeding,
based on the need for red blood cell transfusion or a decrease in hematocrit ≥ 20 %
[4]. The average (standard deviation) severity of bleeding score was 3.1 [1.2], according
to the NOBLADS system [5]. Extravasation was located in the ascending colon in nine cases, transverse colon
in four, descending colon in three, and sigmoid colon in five. Fifteen cases (71 %)
had clustering diverticula, with a density of >10 per 2 inches, in the colon. Relevant
characteristics of the study group are summarized in [Table 1].
Table 1
Characteristics of the patient group.
|
Variable
|
N = 21
|
|
Age, mean [SD], years
|
73.2 [12.8]
|
|
Sex (male/female), n
|
14/7
|
|
Severity of bleeding
|
|
|
11.9 [2.1]
|
|
|
9 (43)
|
|
|
4 (19)
|
|
|
3.1 [1.2]
|
|
Distribution of diverticula, n (%)
|
|
|
17 (81)
|
|
|
18 (86)
|
|
Location of the extravasation on CECT, n (%)
|
|
|
9 (43)
|
|
|
4 (19)
|
|
|
3 (14)
|
|
|
5 (24)
|
|
Density of diverticula (per 2 inches), n
|
|
|
15
|
|
|
4
|
|
|
2
|
SD, standard deviation; CECT, contrast-enhanced computed tomography
1 Scoring system based on nonsteroidal anti-inflammatory drugs use, no diarrhea, no
abdominal tenderness, blood pressure of 100 mmHg or lower, antiplatelet drugs use,
albumin level less than 3.0 g/dL, disease score of 2 or higher, and syncope.
The target diverticulum was identified in 20 of 21 cases (95%; [Table 2]). The target diverticulum was not identified in the ascending colon in one case
after 58 minutes of endoscopic examination. In that patient, numerous diverticula
were identified in the ascending colon, but none of them were identified as a potential
source of bleeding. Median endoscopic search time to identify the target diverticulum
was 5 minutes (range, 1 to 58 minutes), with a median total examination time of 40
minutes (range, 20 to 85 minutes), which included insertion of the endoscope, step-clipping,
searching, and treatment, combined for both the initial and secondary colonoscopy.
In all cases, no active bleeding was identified during the colonoscopy, with the target
diverticulum confirmed by evidence of clot attachment in nine cases and non-bleeding
visible vessels in the other 11 cases. Endoscopic band ligation (EBL) was used as
the primary treatment in all cases, with one case requiring subsequent clipping and
another case requiring arterial embolization. Median time of total work-up from admission
to the second colonoscopy was 32 hours (range, 14 to 136 hours). There was no incidence
of adverse events, with no report of rebleeding after treatment, over a 3-month period
of follow-up.
Table 2
Endoscopic management.
|
Variable
|
N = 21
|
|
Number of deployed clips, median (range)
|
8 (3–12)
|
|
Detection of target diverticulum, n (%)
|
20 (95)
|
|
Total examination time, minutes
|
|
|
41.5 [20.1]
|
|
|
40 (20–85)
|
|
Total searching time, minutes
|
|
|
11.2 [16.9]
|
|
|
5 (1–58)
|
|
Type of SRH, n (%)
|
|
|
9 (43)
|
|
|
11 (52)
|
|
Treatment, n (%)
|
|
|
18 (86)
|
|
|
1 (5)
|
|
|
1 (5)
|
|
|
0 (0)
|
SD, standard deviation; SRH, stigmata of recent hemorrhage; EBL, endoscopic band ligation.
Discussion
Colonic diverticulosis is a common disease that is frequently observed during colonoscopy.
The etiology of colonic diverticulosis has not been fully clarified, although aging
is considered as a definite risk factor [6]. Among patients with colonic diverticula, 15 % to 30 % experience some diverticula-related
event during their lifetime, such as inflammation or bleeding [7]. Extravasation on CECT is definitive evidence of colonic diverticular bleeding and
thus, helps in planning of the colonoscopy examination. However, because of the intermittent
nature of colonic diverticular bleeding, the detection rate of the target lesion by
colonoscopy remains at ≤ 70 % [1]
[2]. Moreover, colonoscopy does not provide positional information about the target
diverticulum that is as precise as CT or barium enema [8]. Because of the low detection rate of the target diverticulum, endoscopic examination
can typically be extended to 45 to 140 minutes, with the examination becoming troublesome
[9]
[10]
[11]. Step-clipping has been advocated as a breakthrough technique for the management
of a diverticulum suspected of causing bleeding, detected by extravasation on CECT,
but with absence of active bleeding on subsequent colonoscopy. When active bleeding
is present, endoscopic hemostasis can be directly applied. In the absence of active
bleeding, step-clipping can facilitate identification of the target diverticulum,
especially in the presence of other residual clots in the colon. In our study, the
step-clipping technique was associated with a 95 % detection rate, with a short examination
(mean, 11.2 minutes) needed to identify the target lesion.
As the step-clipping technique is based on an analysis of CT images, the method affords
endoscopists the time needed to examine the colon, free of the stress associated with
extending an endoscopic examination. Even in the presence of intestinal peristalsis
or pooled gas on the CT image, these features do not have significant effects on the
primary colon structure or on the interpretation of the image. Therefore, CT-based
pre-planning eliminates the need for redundant searching for the target lesion during
the second colonoscopy for treatment, even when active bleeding is not present. In
our protocol, unenhanced CT was performed 2 hours after step-clipping, which provided
sufficient time for CO2 to be either completely absorbed or excreted. This step is important to ensure that
the colon is deflated, allowing comparison of the before and after CT images. In our
study, we used 1-mm-axial before and after images, with the 1-mm thick slices providing
a more appropriate method to identify a probable bleeding diverticulum than 5-mm CT
slices, particularly for small-size diverticula.
A previous study reported on the effectiveness of colonoscopy performed within 24
hours from the onset of hematochezia [12]. By comparison, step-clipping can be conducted regardless of the time to endoscopy
from the onset of hematochezia, in a well-prepared environment with PEG bowel preparation.
Although it is true that early colonoscopy may improve outcomes in patients with acute
lower gastrointestinal bleeding, urgent CT before colonoscopy provided 15 % additional
detectability compared to colonoscopy alone [13]. The step-clipping method achieves a high detection rate of the target diverticulum,
even in cases requiring elective colonoscopy. Therefore, with accessibility to CT
imaging, the step-clipping method provides a clinically feasible method to identify
the target diverticulum for treatment, which could be implemented in clinics with
insufficient human or material resources to conduct urgent colonoscopy.
The step-clipping method has two primary limitations that should be acknowledged.
First, step-clipping requires additional costs (about $ 270) and CT examination (an
effective dose of radiation of about 5 mGy), compared to the traditional approach
combining CECT and colonoscopy. Although this is troublesome, consistent detection
and treatment of the target lesion might reduce the rate of rebleeding and readmission,
lowering costs and radiation exposure in the longer term. Further investigation of
the clinical impact of the step-clipping method, through the accumulation of experience,
is warranted. Second, the step-clipping method can only be used in cases where extravasation
is identified on CECT.
Conclusions
Based on our analysis, we believe that step-clipping provides an effective map to
facilitate navigation to the target diverticulum within a short time, even in the
absence of active bleeding, during endoscopy.
Video 1 Placement of the marking clips and subsequent computed tomography provide objective
positional information, which can lead to rapid and precise identification of the
responsible diverticulum on subsequent colonoscopy.
