Gastroenterology 2020;158:168–175

Efficacy and Safety of Early vs Elective Colonoscopy for Acute

Lower Gastrointestinal Bleeding
Ryota Niikura,1,* Naoyoshi Nagata,2,3,* Atsuo Yamada,1,* Tetsuro Honda,4 Kenkei Hasatani,5
Naoki Ishii,6 Yasutoshi Shiratori,6 Hisashi Doyama,7 Tsutomu Nishida,8 Tetsuya Sumiyoshi,9
Tomoki Fujita,10,11 Shu Kiyotoki,12 Tomoyuki Yada,13 Katsumi Yamamoto,14
Tomohiro Shinozaki,15,16 Munenori Takata,17 Tatsuya Mikami,18 Katsuhiro Mabe,19
CLINICAL AT

Kazuo Hara,20 Mitsuhiro Fujishiro,1,21 and Kazuhiko Koike1

1
Department of Gastroenterology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan;
2
Department of Gastroenterology and Hepatology, National Center for Global Health and Medicine, Shinjuku-ku, Tokyo, Japan;
3
Gastroenterological Endoscopy, Tokyo Medical University, Shinjuku-ku, Tokyo, Japan; 4Department of Gastroenterology,
Nagasaki Harbor Medical Center, Nagasaki-shi, Nagasaki, Japan; 5Department of Gastroenterology, Fukui Prefectural Hospital,
Fukui-shi, Fukui, Japan; 6Department of Gastroenterology, St. Luke’s International Hospital, Chuo-ku, Tokyo, Japan;
7
Department of Gastroenterology, Ishikawa Prefectural Central Hospital, Kanazawa-shi, Ishikawa, Japan; 8Department of
Gastroenterology, Toyonaka Municipal Hospital, Toyonaka-shi, Osaka, Japan; 9Department of Gastroenterology, Tonan
Hospital, Sapporo-shi, Hokkaido, Japan; 10Department of Gastroenterology, Otaru Ekisaikai Hospital, Otaru-shi, Hokkaido,
Japan; 11Department of Gastroenterology, Sapporo Century Hospital, Sapporo-shi, Hokkaido, Japan; 12Department of
Gastroenterology, Shuto General Hospital, Yanai-shi, Yamaguchi, Japan; 13Department of Gastroenterology and Hepatology,
National Center for Global Health and Medicine Kohnodai Hospital, Ichikawa-shi, Chiba, Japan; 14Department of
Gastroenterology, Japan Community Healthcare Organization Osaka Hospital, Osaka-shi, Osaka, Japan; 15Department of
Biostatistics, School of Public Health, The University of Tokyo, Bunkyo-ku, Tokyo, Japan; 16Department of Information and
Computer Technology, Faculty of Engineering, Tokyo University of Science, Shinjuku-ku, Tokyo, Japan; 17Clinical Research
Support Center, The University of Tokyo Hospital, Bunkyo-ku, Tokyo, Japan; 18Division of Endoscopy, Hirosaki University
Hospital, Hirosaki-shi, Aomori, Japan; 19Department of Gastroenterology, National Hospital Organization Hakodate Hospital,
Hakodate-shi, Hokkaido, Japan; 20Department of Gastroenterology, Aichi Cancer Center Hospital, Nagoya-shi, Aichi, Japan;
and 21Department of Gastroenterology and Hepatology, Nagoya University Graduate School of Medicine, Nagoya, Japan

early colonoscopy (within 24 hours of initial visit to the hos-

See editorial on page 38.
BACKGROUND & AIMS: We performed a large, multicenter,
randomized controlled trial to determine the efficacy and safety
of early colonoscopy on outcomes of patients with acute lower
gastrointestinal bleeding (ALGIB). METHODS: We performed
an open-label study at 15 hospitals in Japan of 170 patients
with ALGIB randomly assigned (1:1) to groups that underwent
pital) or elective colonoscopy (24–96 hours after hospital
admission). The primary outcome was identification of stigmata
of recent hemorrhage (SRH). Secondary outcomes were
rebleeding within 30 days, endoscopic treatment success, need
for transfusion, length of stay, thrombotic events within 30
days, death within 30 days, and adverse events. RESULTS: SRH
were identified in 17 of 79 patients (21.5%) in the early colo-
noscopy group vs 17 of 80 patients (21.3%) in the elective
January 2020
colonoscopy group (difference, 0.3; 95% confidence interval,
–12.5 to 13.0; P ¼ .967). Rebleeding within 30 days of hospital
admission occurred in 15.3% of patients in the early colonos-
copy group and 6.7% of patients in the elective colonoscopy
group (difference, 8.6; 95% confidence interval, –1.4 to 18.7);
there were no significant differences between groups in suc-
cessful endoscopic treatment rate, transfusion rate, length of
stay, thrombotic events, or death within 30 days. The adverse
event of hemorrhagic shock occurred during bowel preparation
in no patient in the early group vs 2 patients (2.5%) in the
elective colonoscopy group. CONCLUSIONS: In a randomized
controlled study, we found that colonoscopy within 24 hours
after hospital admission did not increase SRH or reduce
rebleeding compared with colonoscopy at 24–96 hours in pa-
tients with ALGIB. ClinicalTrials.gov, Numbers:
UMIN000021129 and NCT03098173
Keywords: Acute Lower Gastrointestinal Bleeding; Early Colo-
noscopy; Randomized Controlled Trial; Stigmata of Recent
Hemorrhage.
A cute lower gastrointestinal bleeding (ALGIB) is a
common condition worldwide, and the bleeding is
often severe enough to require blood transfusion and he-
mostatic intervention.1,2 Early colonoscopy has been defined
as that performed within 8–24 hours of admission ALGIB by
American Society for Gastrointestinal Endoscopy and Amer-
ican College of Gastroenterology guidelines,3,4 although some
studies cite no lower limit,5–8 and is recommended for the
diagnosis and treatment of severe bleeding; however, there is
little supporting evidence of the optimal timing of colonoscopy.
In addition, adequate colon preparation is recommended, but
data on the safety of early colonoscopy is limited. Therefore,
the strategy of colonoscopy for ALGIB has not been stan-
dardized in real clinical practice,9 and there is general concern
regarding the efficacy and safety of early colonoscopy. Three
meta-analyses8,10,11 and 3 single-center randomized
controlled trials (RCTs)6,12,13 of early versus elective colo-
noscopy have been reported; however, controversy remains as
to whether early colonoscopy improves important outcomes
including rebleeding, need for transfusion, and mortality.6,12,13
In addition, 2 of 3 single-center RCTs were stopped during
enrollment because of difficulty reaching the prespecified
sample size, which had a potential risk of type II error and
lacked generalizability. Therefore, there remains a need for a
high-quality multicenter RCT with a large sample size to
evaluate the efficacy and safety of early colonoscopy in ALGIB.
To address these issues, we designed a multicenter
randomized trial to evaluate whether early colonoscopy
improves clinical outcomes compared with elective colo-
noscopy in patients with ALGIB.
Methods
Trial Oversight
In this multicenter, open-label, RCT, patients were allocated
(1:1) to undergo either early colonoscopy or elective colonos-
copy at 15 hospitals in Japan. We obtained written informed
Early vs Elective Colonoscopy 169
WHAT YOU NEED TO KNOW
BACKGROUND AND CONTEXT
Many guidelines recommended early colonoscopy (within
8–24 hrs of admission to the hospital) for patients with
acute lower gastrointestinal bleeding, but there are few
data on the optimal timing of colonoscopy.
NEW FINDINGS
In a randomized controlled trial of patients admitted to the
hospital with acute lower gastrointestinal bleeding,
performing colonoscopy within 24 hrs of admission did
CLINICAL AT
not reduce the proportion of patients with stigmata of
recent hemorrhage or subsequent rebleeding compared
with performing colonoscopy after 24–96 hrs.
LIMITATIONS
This was an open-label trial. However, indirect blinding
was achieved using an independent central review
committee.
IMPACT
These findings indicate that early colonoscopy in patients
admitted to the hospital for acute lower gastrointestinal
bleeding does not increase stigmata identification and
reduce rebleeding.
consent from all patients before enrollment. The trial was
approved by the institutional review boards of all participating
hospitals. The institutional review boards monitored the
progress of the trial. The rationale and methodology of the
study have been published, and the full protocol is available
online.14 All authors had access to the study data and reviewed
and approved the final manuscript. The trial is registered at
UMIN Clinical Trials Registry (UMIN000021129, February 21,
2016) and ClinicalTrials.gov (NCT03098173, March 24, 2017).
Patients
Outpatients aged 20 years who presented with moderate
to severe hematochezia or melena within 24 hours of arrival
were eligible. Major inclusion criteria were (1) 3 occurrences
of hematochezia within 8 hours or (2) hemorrhagic shock or (3)
requiring transfusion according to previous trials.6,12,13
Exclusion criteria included hematemesis or coffee ground
vomitus; upper gastrointestinal bleeding, diagnosed by using
nasogastric tube or upper endoscopy; ingestion of oral bowel
preparation solution deemed not possible; computed tomog-
raphy performed before colonoscopy; diagnosis of peptic ulcer
disease within the previous 10 days; ulcerative colitis or
Crohn’s disease; abdominal surgery within the previous 10
days; polypectomy, endoscopic mucosal resection, or
* Authors share co-first authorship.
Abbreviations used in this paper: ALGIB, acute lower gastrointestinal
bleeding; CI, confidence interval; RCT, randomized controlled trial; SRH,
stigmata of recent hemorrhage.
Most current article
© 2020 by the AGA Institute. Published by Elsevier Inc. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.
org/licenses/by-nc-nd/4.0/).
0016-5085
https://doi.org/10.1053/j.gastro.2019.09.010
 170 Niikura et al
endoscopic submucosal dissection in the lower gastrointestinal
tract within the previous 10 days; suspected perforation or
peritonitis; suspected intestinal obstruction; hemorrhagic
shock refractory to infusion or blood transfusion; history of
total colectomy; suspected disseminated intravascular coagul-
opathy; end-stage malignant disease; severe cardiac failure;
active thrombosis; severe respiratory failure; and pregnancy.
All eligible patients were hospitalized in this trial.
Trial Design and Treatment
Enrolled patients were randomly assigned 1:1 to undergo
CLINICAL AT
either early colonoscopy or elective colonoscopy. Immediate
and concealed allocation was performed using a Web-based
central randomization system with 24-hour access. Randomi-
zation was performed in blocks, with block sizes varying ac-
cording to the allocation sequence generated by the Central
Coordinating Unit at the University of Tokyo. A unique patient
identification number was then entered into the system to
ensure anonymity.
After randomization, early colonoscopy was performed
within 24 hours of the initial visit. Elective colonoscopy was
performed between 24 and 96 hours after the initial visit. All
colonoscopies were performed with an electronic video endo-
scope (Olympus Optical, Tokyo, Japan or Fujifilm Corporation,
Tokyo, Japan) after administration of 2–4 L of oral bowel
preparation solution. Enema was additionally administered to
patients who did not ingest all of the oral bowel preparation
solution and was continued until the effluent was free of fecal
material. Colonoscopy was performed by using both a water-jet
device and attachment cap. If patients used any medications,
such as nonsteroidal anti-inflammatory drugs, low-dose aspirin,
thienopyridine, other antiplatelet drugs (eg, dipyridamole, cil-
ostazol), or oral anticoagulants, these drugs were either dis-
continued or continued and were resumed or not resumed at
discharge at the discretion of the primary physician.
Outcomes
Primary outcome was stigmata of recent hemorrhage (SRH)
identification. Diagnosis of SRH was based on colonoscopic
visualization of active bleeding; nonbleeding visible vessel; and
adherent clot.15–17 A central review of all endoscopic images
was carried out by an independent effect judgment committee
comprising 2 nonstudy investigators. The key secondary
outcome was 30-day rebleeding, which was defined as signifi-
cant fresh blood loss after initial colonoscopy with any of the
following: (1) hemorrhagic shock, (2) need for transfusion, (3)
identification of blood pooling on further colonoscopy, (4) SRH
in the lower gastrointestinal tract, or (5) extravasation identi-
fied in the colorectal region on contrast-enhanced CT. Other
secondary outcomes were endoscopic treatment success, need
for additional endoscopic examinations, need for interventional
radiology, need for surgery, need for transfusion during hos-
pitalization, length of stay, 30-day thrombotic events, 30-day
mortality, bowel preparation-related adverse events, and
colonoscopy-related adverse events. Outcome definitions are
presented in detail in the Supplementary Materials.
Statistical Analysis
We estimated the required sample size assuming an SRH
rate of 9% in the elective colonoscopy group and 27% in the
Gastroenterology Vol. 158, No. 1
early colonoscopy group.7 With an alpha level of 5% (2-sided),
a sample size of 142 (2  71) patients was required to ensure
statistical power of 80% in the chi-squared test without Yates’s
correction. To allow for any observed differences that might be
obscured by patient noncompliance and/or dropout rate, we
recruited 20 additional patients to account for these, and thus
recruited a total of 162 randomized patients.
The primary analysis included a modified intention-to-treat
population, excluding the following patients from the genuine
intention-to-treat analysis set: (1) those who did not satisfy the
enrollment criteria after randomization, (2) those who provided
no postrandomization data on primary outcome, and (3) those
who did not undergo colonoscopy. Data on SRH identification in
the modified intention-to-treat population were compared using
the chi-squared test. A secondary outcome analysis was also
performed in the modified intention-to-treat population to
compare the secondary outcomes. The P value for statistical
significance of the primary outcome was <.05 and was 2-tailed.
All analyses were performed using SAS software, version 9.4
(SAS Institute, Cary, NC). The statistical analysis plan, including
subgroup analysis and per-protocol analyses, is shown in detail
in the Supplementary Materials. We assessed the severity of
bleeding by using hemoglobin level data and vital signs,
including systolic blood pressure and pulse rate, at the initial
visit; these are the most frequent factors appearing in various
severity scoring systems,18 along with blood transfusion in-
dications for lower gastrointestinal bleeding.19 We categorized
severe as hemoglobin <8 g/dL, systolic blood pressure <90 mm
Hg, or pulse rate 100 beats/min; moderate was defined as
hemoglobin 8 g/dL. For the per-protocol analysis population,
the early colonoscopy group had colonoscopy performed within
24 hours of the initial visit, and the elective colonoscopy group
had the procedure performed between 24 and 96 hours.
Results
Of the 170 patients who were enrolled between July
2016 and May 2018, a total of 162 underwent randomiza-
tion; 3 were excluded, and 159 were included in the
modified intention-to-treat population. Finally, 79 patients
were allocated to the early colonoscopy group and 80 to the
elective colonoscopy group (Supplementary Figure 1). Two
of 79 patients in early colonoscopy group and 1 of 80 pa-
tients in the elective colonoscopy group had esophagogas-
troduodenoscopy. No patients were excluded due to these
findings in either group. The patient demographics were
similar in both groups (Table 1). Severe or moderate
bleeding was observed in 31 (39.2%) and 48 (60.8%) of 79
patients in the early colonoscopy group and 21 (26.3%) and
59 (73.7%) of 80 patients in the elective colonoscopy group
(Table 1). Central review was performed in all patients
(100%) in both treatment groups. Colonoscopy was per-
formed at a mean of 13.9 hours after arrival in the early
colonoscopy group and 41.4 hours after arrival in the
elective colonoscopy group. Only 1 patient had colonoscopy
beyond 24 hours in the early colonoscopy group. Cecal
insertion time, total procedure time, and bleeding sources
were similar in the 2 groups (Tables 1 and 2). The diag-
nostic yield of unknown bleeding source was 13 of 79 pa-
tients (16.5%) and 13 of 80 patients (16.3%) in the
early and elective colonoscopy groups, respectively
January 2020 Early vs Elective Colonoscopy 171

Table 1.Baseline Patient Characteristics the elective colonoscopy group (difference, 0.3; 95% confi-
dence interval [CI], –12.5 to 13.0; P ¼ .967). SRH were
Early Elective rarely identified beyond 36 hours, and the last SRH were
colonoscopy colonoscopy identified at 44 hours (Supplementary Figure 2). On central
Variable (n ¼ 79) (n ¼ 80) review, SRH were identified in 20 of 79 patients (25.3%) in
Age, y, mean ± SD 68.8 ± 12.8 71.9 ± 12.3 the early colonoscopy group and 21 of 80 patients (26.3%)
Sex, male, n (%) 52 (65.8) 54 (67.5) in the elective colonoscopy group (difference, –0.9; 95% CI,
Body mass index, mean ± SD 23.0 ± 3.3 23.9 ± 4.1 –14.5 to 12.7; P ¼ .893) (Table 3). These findings remained
Comorbidities, n (%) unchanged in the per-protocol analyses (Supplementary
Previous lower GI bleeding 28 (35.4) 30 (37.5) Table 1). The results for the primary outcome in pre-
Ischemic heart disease 12 (15.2) 18 (22.5)
specified subgroups were consistent, except for patients
Chronic obstructive 3 (3.8) 0

CLINICAL AT
pulmonary disease
without colonic diverticular bleeding, for whom SRH were
Peptic ulcer 6 (7.6) 2 (2.5) identified more frequently in the early colonoscopy group
Liver cirrhosis 1 (1.3) 2 (2.5) (Supplementary Table 2).
Diabetes mellitus 11 (13.9) 16 (20.0)
Chronic heart failure 4 (5.1) 5 (6.3)
Cerebrovascular disease 11 (13.9) 15 (18.8) Secondary Outcomes
Dementia 0 2 (2.5) We applied endoscopic treatment for 15 patients in the
Collagen disease 4 (5.1) 5 (6.3) early colonoscopy group and 15 patients in elective colo-
Chronic kidney disease 8 (10.1) 11 (13.8) noscopy group. Successful endoscopic therapy and the need
Leukemia 0 1 (1.3)
for additional endoscopic examinations and for interven-
Malignant lymphoma 2 (2.5) 0
Solid cancer 7 (8.9) 12 (15.0)
tional radiology and surgery were similar in the groups
Medication (Table 3). Thirty patients in the early colonoscopy group
Low-dose aspirin 16 (20.3) 22 (27.5) and 26 in the elective colonoscopy group required trans-
Thienopyridine 7 (8.9) 6 (7.5) fusion during hospitalization. Length of stay was similar in
Cilostazol 5 (6.3) 2 (2.5) both groups.
Other antiplatelet drugs 5 (6.3) 3 (3.8) Thirty-day rebleeding was observed in 11 of 72 patients
Warfarin 4 (5.1) 6 (7.5)
(15.3%) in the early colonoscopy group and 5 of 75 patients
Direct oral anticoagulants 3 (3.8) 7 (8.8)
NSAIDs 14 (17.7) 16 (20.0) (6.7%) in the elective colonoscopy group (difference, 8.6;
Initial assessment 95% CI, –1.4 to 18.7) (Table 3). These findings remained
Hemodynamic instabilitya 4 (5.1) 1 (1.3) unchanged in the per-protocol analyses (Supplementary
Hemoglobin, g/dL, mean ± SD 11.4 ± 2.7 11.3 ± 2.3 Table 1). One patient in the early colonoscopy group was
Upper GI endoscopy 2 (2.5) 1 (1.3) not included because the initial endoscopic treatment was
before colonoscopy unsuccessful. Among patients who underwent successful
Severity of bleeding
endoscopic treatment in the early (n ¼ 14) and elective (n ¼
Severe bleeding 31 (39.2) 21 (26.3)
Moderate bleeding 48 (60.8) 59 (73.7) 15) colonoscopy groups, 14 in the early group and 12 in the
Endoscopic procedure elective group met the 30-day follow-up criteria. Of these,
Expert endoscopistb 48 (60.8) 48 (60.0) 14.3% (2/14) in the early group and 16.7% (2/12) in the
Cecal insertion 77 (97.5) 77 (96.3) elective group experienced rebleeding within 30 days.
Cecal insertion time, min, 9.6 ± 6.5 9.0 ± 6.5 Thirty-day thrombotic events were observed in 1.3% in
mean ± SD
the early colonoscopy group and in no patients in the
Total procedure time, min, 35.8 ± 19.2 31.3 ± 15.1
mean ± SD
elective colonoscopy group. The 30-day mortality rate was
zero in both study groups (Table 3). These findings are
similar to those obtained in the per-protocol analysis
GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory (Supplementary Table 1).
drugs; SD, standard deviation.
a
Hemodynamic instability was defined as heart rate >100
beats/min and systolic blood pressure <115 mm Hg.24 Safety
b
Expert endoscopists were defined as having performed The frequency of bowel preparation- and colonoscopy-
>1000 colonoscopies, including endoscopic hemostasis. related adverse events was less than 5% except for exac-
erbation of bleeding and nausea in both groups (Table 4).
(Table 2). The per-protocol and modified intention-to-treat Bowel preparation and colonoscopy-related adverse events
analyses were virtually the same, given that only 2 pa- did not differ between the early and elective colonoscopy
tients were excluded from the modified intention-to-treat groups. All adverse events had resolved by the end of the
analysis (Supplementary Table 1). trial, and none led to discontinuation of the trial.

Primary Outcome Discussion

SRH were identified in 17 of 79 patients (21.5%) in the We found that early colonoscopy does not improve the
early colonoscopy group and 17 of 80 patients (21.3%) in rate of SRH identification versus elective colonoscopy in
 172 Niikura et al Gastroenterology Vol. 158, No. 1

Table 2.Bleeding Source

Early colonoscopy Elective colonoscopy

Bleeding source (n ¼ 79), n (%) (n ¼ 80), n (%) P value

Diverticular bleeding (definite)a 10 (12.7) 16 (20.0) .211

Diverticular bleeding (presumptive)b 37 (46.8) 36 (45.0) .816
Rectal ulcer 0 0 N/A
Colorectal cancer 3 (3.8) 2 (2.5) .639
Ischemic colitis 9 (11.4) 5 (6.3) .253
Infectious colitis 1 (1.3) 0 .313
Radiation colitis 1 (1.3) 0 .313
Colonic ulcer 0 2 (2.5) .157
CLINICAL AT

Nonspecific colitis 0 0 N/A

Hemorrhoid 1 (1.3) 2 (2.5) .567
Otherc 7 (8.9) 4 (5.0) .338
Unknown 13 (16.5) 13 (16.3) .972
Upper GI bleeding 0 1 (1.3) .319

NOTE. Data are allowed in duplicate.

GI, gastrointestinal; N/A, not applicable.
a
Definite diverticular bleeding source was defined as diverticula with SRH.
b
Presumptive diverticular bleeding source was defined as diverticula without stigmata of recent hemorrhage.
c
Other included small intestinal bleeding for postendoscopic mucosal resection bleeding (n ¼ 1 in the early colonoscopy
group), colonic diverticulitis (n ¼ 1 in the elective colonoscopy group), angioectasia (n ¼ 3 in the early colonoscopy group, n ¼
2 in the elective colonoscopy group), and bleeding from the small intestine (n ¼ 3 in the early colonoscopy group, n ¼ 1 in the
elective colonoscopy group).

patients with ALGIB. Furthermore, the 30-day rebleeding
rate did not differ between the early and elective colonos-
copy groups. Guidelines and prior studies often recommend
that early colonoscopy be performed within 8 to 24 hours of
admission to increase diagnostic yield10,11 and the likeli-
hood of a therapeutic intervention.3,9,20
Contrary to our hypothesis, SRH identification rate in the
elective colonoscopy group was much higher than our ex-
pected 9%, although the SRH identification rate in early
colonoscopy was consistent with that of previous RCTs
(6%–19%).6,13 We suggest 2 reasons for this discrepancy.
First, in our trial, elective colonoscopy was performed at a
mean of 41.4 hours, a shorter time than expected. A previ-
ous study21 showed that SRH are rarely identified beyond
36 hours, and, indeed, our last SRH were identified at 44
hours from arrival (Supplementary Figure 2). In the elective
colonoscopy group, 48.8% of patients underwent colonos-
copy within 36 hours of arrival at the hospitals. Thus,
relatively early performance of colonoscopy in the elective
group may have resulted in a high rate of SRH identification.
Another reason for the high rate of SRH identification in the
elective colonoscopy arm may be differences in colonoscopy
procedure among the facilities. In this study, 4 of 15 hos-
pitals showed SRH identification of >20% in elective colo-
noscopy, and 2 of these are high-volume emergency
hospitals, where the physicians are more accustomed to the
colonoscopy procedure for ALGIB compared with other
hospitals.
The results of our trial, the largest trial to date to our
knowledge, are in line with 2 of the 3 prior RCTs,6,12,13
which found no increase in diagnostic yield, although
some meta-analyses do show improved diagnostic yield.10,11
Approximately 15% of patients with an unknown bleeding
source have potential small intestinal bleeding sources, and
further endoscopic work such as capsule endoscopy and
balloon-assisted endoscopy, not early colonoscopy, is
needed to identify bleeding sources.
We hypothesized that identification of SRH would enable
clinicians to perform accurate hemostasis with reduced risk
of rebleeding. However, our results showed no beneficial
effect of early colonoscopy to achieve diagnosis, hemostasis,
or reduce rebleeding rate. Although our study had small
sample size, the upper bound of the 95% CI in our study
was 1.4%, suggesting that even a larger study is unlikely to
show a benefit for early colonoscopy in terms of rebleeding.
Various preparation- and colonoscopy-related adverse
events have been reported in ALGIB.22 Bowel preparation–
related adverse events were reported as including 5.5%
volume overload13 and 7% hemorrhagic shock23; thus,
physicians may hesitate to perform bowel preparation for
patients with ALGIB. Perforation and cardiovascular events
have also been reported as adverse events during the co-
lonoscopy procedure.6,12,13,22 However, to our knowledge,
no RCTs or prospective studies have systematically assessed
these adverse events with a meticulous reporting system.
Also, our trial was the first study to show few preparation-
and colonoscopy-related adverse events, with each having
rates of less than 5% in both groups.
Our study has several strengths. First, to our knowledge,
our trial is the first multicenter RCT with the largest sample
size to date. All previous RCTs were single-center studies
with small sample sizes. Second, protocol deviations for
performing colonoscopy were only 1%, suggesting that our
trial showed high-quality protocol compliance. Nevertheless,
there are several limitations to the study. First, a direct-
blinded patient assessment was difficult to perform
January 2020 Early vs Elective Colonoscopy 173

Table 3.Primary and Secondary Outcomes

Early Elective
colonoscopy colonoscopy
Outcome (n ¼ 79), n (%) (n ¼ 80), n (%) Difference P value

Primary outcome
SRH identification 17 (21.5) 17 (21.3) 0.3 (–12.5 to 13.0) .967
SRH identification by central review 20 (25.3) 21 (26.3) –0.9 (–14.5 to 12.7) .893
Secondary outcome
30-day rebleedinga,b 11 (15.3) 5 (6.7) 8.6 (–1.4 to 18.7) .094
Success rate of endoscopic treatmentc 14/15 (93.3) 15/15 (100.0) –6.7 (–19.3 to 6.0) .309
Success rate of endoscopic treatment 8/10 (80.0) 6/8 (75.0) 5.0 (–0.3 to 0.4) .800

CLINICAL AT
for active bleedingd
Need for additional endoscopic examinations 31 (39.2) 23 (28.8) 10.5 (–4.1 to 25.1) .163
Need for interventional radiology 1 (1.3) 0 1.3 (–1.2 to 3.7) .313
Need for surgery 0 0 0 N/A
Need for transfusion during hospitalization 30 (38.0) 26 (32.5) 5.5 (–9.4 to 20.3) .470
Length of stay 7.1 ± 5.7 7.6 ± 6.0 –0.5 (–2.3 to 1.3) .112
30-day thrombotic eventsb,e 1 (1.4) 0 1.4 (–1.3 to 4.1) .306
30-day mortalityb 0 0 0 N/A

N/A, not applicable.

a
Rebleeding was defined as significant fresh blood loss after an initial colonoscopy with any of the following: (1) hemorrhagic
shock, including cold sweats, nausea, syncope, or systolic blood pressure 90 mm Hg; (2) need for transfusion according to
the guidelines of the Japanese Ministry of Health, Labor and Welfare; (3) further colonoscopy identifying blood pooling; (4) SRH
in the lower gastrointestinal tract; or (5) contrast-enhanced computed tomography scan identifying extravasation in the
colorectal region.
b
Rebleeding, thrombotic events, and mortality were evaluated for patients who approximately met the 30-day follow-up criteria
(early colonoscopy for 72 patients and elective colonoscopy for 75 patients).
c
Success rate of endoscopic treatment was defined as achievement of hemostasis (early for 15 patients and elective for 15
patients).
d
Success rate of endoscopic treatment for active bleeding was defined as achievement of hemostasis for active bleeding with
SRH (early for 10 patients and elective for 8 patients).
e
Thrombotic events included acute coronary syndromes, including angina pectoris and myocardial infarction; stroke, including
cerebrovascular infarction, cerebral hemorrhage, and transient ischemic attacks; deep vein thrombosis; and pulmonary
embolism.

Table 4.Adverse Events

because most investigators and patients knew the allocated
treatment based on the time course. However, blinding was Patients, n (%)
achieved by using an independent central review committee
that did not directly examine the patients. The central review Early Elective
achieved results for the primary outcome that were similar colonoscopy colonoscopy
to those obtained by the site investigators. Second, the choice Adverse event (n ¼ 79) (n ¼ 80)
of SRH identification as the primary outcome is a limitation.
Bowel preparation-related
However, our results for rebleeding suggested that early adverse events
colonoscopy does not improve rebleeding outcomes, even if Nausea, vomiting 4 (5.1) 3 (3.8)
we chose rebleeding as the primary outcome. Third, Abdominal pain 1 (1.3) 1 (1.3)
approximately 70% of patients met criteria for moderate Volume overload 0 0
bleeding based on our severity categorization criterion. Aspiration pneumonia 0 0
However, subgroup analysis showed no differences between Hemorrhagic shocka 0 2 (2.5)
Exacerbation of bleedingb 31 (39.2) 22 (27.5)
primary outcomes according to severity categorization
Ileus 0 0
(Supplementary Table 2). Fourth, most facilities enrolled Colonoscopy-related
only a few patients, and our trial has a potential bias in the adverse events
differing numbers of enrolled patients between the facilities. Hemorrhagic shock 1 (1.3) 0
However, subgroup analysis showed no differences between Perforation 0 0
primary outcome at different facilities (Supplementary
Table 2), and the impact on the results seems small. a
Hemorrhagic shock was defined as cold sweats, nausea,
Finally, our results cannot be generalized to patients with syncope, or systolic blood pressure 90 mm Hg.
inpatient-onset ALGIB or patients with severe ALGIB who b
Bowel preparation–related exacerbation of bleeding was
had hemorrhagic shock refractory to blood transfusion defined as more than 1 occurrence of hematochezia during
because these patients were excluded from our trial. preparation.
 174 Niikura et al
This multicenter RCT showed that early colonoscopy,
performed within 24 hours of arrival, did not increase SRH
identification rate or reduce 30-day rebleeding when
compared with elective colonoscopy in patients with ALGIB.
Supplementary Material
Note: To access the supplementary material accompanying
this article, visit the online version of Gastroenterology at
www.gastrojournal.org, and at https://doi.org/10.1053/
j.gastro.2019.09.010.
CLINICAL AT
References
1. Lanas A, Garcia-Rodriguez LA, Polo-Tomas M, et al.
Time trends and impact of upper and lower gastroin-
testinal bleeding and perforation in clinical practice. Am J
Gastroenterol 2009;104:1633–1641.
2. Longstreth GF. Epidemiology and outcome of patients
hospitalized with acute lower gastrointestinal hemor-
rhage: a population-based study. Am J Gastroenterol
1997;92:419–424.
3. ASGE Standards of Practice Committee, Pasha SF,
Shergill A, et al. The role of endoscopy in the patient with
lower GI bleeding. Gastrointest Endosc 2014;79:875–885.
4. Gerson LB, Fidler JL, Cave DR, Leighton JA. ACG clinical
guideline: diagnosis and management of small bowel
bleeding. Am J Gastroenterol 2015;110:1265–1287.
5. Nigam N, Ham SA, Sengupta N. Early colonoscopy for
diverticular bleeding does not reduce risk of post-
discharge recurrent bleeding: a propensity score
matching analysis. Clin Gastroenterol Hepatol 2019;
17:1105–1111.
6. van Rongen I, Thomassen BJW, Perk LE. Early versus
standard colonoscopy: a randomized controlled trial in
patients with acute lower gastrointestinal bleeding re-
sults of the BLEED study. J Clin Gastroenterol 2019;
53:591–598.
7. Nagata N, Niikura R, Sakurai T, et al. Safety and effec-
tiveness of early colonoscopy in management of acute
lower gastrointestinal bleeding, based on propensity
score matching analysis. Clin Gastroenterol Hepatol
2016;14:558–564.
8. Kouanda AM, Somsouk M, Sewell JL, Day LW. Urgent
colonoscopy in patients with lower GI bleeding: a sys-
tematic review and meta-analysis. Gastrointest Endosc
2017;86:107–117.
9. Strate LL, Gralnek IM. ACG clinical guideline: manage-
ment of patients with acute lower gastrointestinal
bleeding. Am J Gastroenterol 2016;111:459–474.
10. Sengupta N, Tapper EB, Feuerstein JD. Early versus
delayed colonoscopy in hospitalized patients with lower
gastrointestinal bleeding: a meta-analysis. J Clin Gas-
troenterol 2017;51:352–359.
11. Seth A, Khan MA, Nollan R, et al. Does urgent colo-
noscopy improve outcomes in the management of lower
gastrointestinal bleeding? Am J Med Sci 2017;353:298–
306.
12. Green BT, Rockey DC, Portwood G, et al. Urgent colo-
noscopy for evaluation and management of acute lower
Gastroenterology Vol. 158, No. 1
gastrointestinal hemorrhage: a randomized controlled
trial. Am J Gastroenterol 2005;100:2395–2402.
13. Laine L, Shah A. Randomized trial of urgent vs. elective
colonoscopy in patients hospitalized with lower GI
bleeding. Am J Gastroenterol 2010;105:2636–2641.
14. Niikura R, Nagata N, Yamada A, et al. A multicenter, ran-
domized controlled trial comparing the identification rate
of stigmata of recent hemorrhage and rebleeding rate
between early and elective colonoscopy in outpatient-
onset acute lower gastrointestinal bleeding: study proto-
col for a randomized controlled trial. Trials 2018;19:214.
15. Jensen DM, Machicado GA, Jutabha R, Kovacs TO.
Urgent colonoscopy for the diagnosis and treatment of
severe diverticular hemorrhage. N Engl J Med 2000;
342:78–82.
16. Kaltenbach T, Watson R, Shah J, et al. Colonoscopy with
clipping is useful in the diagnosis and treatment of
diverticular bleeding. Clin Gastroenterol Hepatol 2012;
10:131–137.
17. Chavalitdhamrong D, Jensen DM, Kovacs TO, et al.
Ischemic colitis as a cause of severe hematochezia: risk
factors and outcomes compared with other colon di-
agnoses. Gastrointest Endosc 2011;74:852–857.
18. Tapaskar N, Jones B, Mei S, Sengupta N. Comparison of
clinical prediction tools and identification of risk factors
for adverse outcomes in acute lower GI bleeding. Gas-
trointest Endosc 2019;89:1005–1013.
19. Kherad O, Restellini S, Martel M, et al. Outcomes
following restrictive or liberal red blood cell transfusion in
patients with lower gastrointestinal bleeding. Aliment
Pharmacol Ther 2019;49:919–925.
20. Nagata N, Ishii N, Manabe N, et al. Guidelines for colonic
diverticular bleeding and colonic diverticulitis: Japan
Gastroenterological Association. Digestion 2019;
99(Suppl 1):1–26.
21. Niikura R, Nagata N, Aoki T, et al. Predictors for identi-
fication of stigmata of recent hemorrhage on colonic
diverticula in lower gastrointestinal bleeding. J Clin
Gastroenterol 2015;49:e24–e30.
22. Zuckerman GR, Prakash C. Acute lower intestinal
bleeding. Part I: clinical presentation and diagnosis.
Gastrointest Endosc 1998;48:606–617.
23. Niikura R, Nagata N, Shimbo T, et al. Adverse events
during bowel preparation and colonoscopy in patients
with acute lower gastrointestinal bleeding compared with
elective non-gastrointestinal bleeding. PLoS One 2015;
10(9):e0138000.
24. Strate LL, Orav EJ, Syngal S. Early predictors of severity
in acute lower intestinal tract bleeding. Arch Intern Med
2003;163:838–843.
Received March 7, 2019. Accepted September 18, 2019.
Reprint requests
Address requests for reprints to: Atsuo Yamada, MD, PhD, Department of
Gastroenterology, Graduate School of Medicine, The University of Tokyo, 7-
3-1 Hongo, Bunkyo-ku, Tokyo 1138655, Japan. e-mail: yamada-
a@umin.ac.jp; fax: 81-3-3815-5411.
Acknowledgments
We thank all the patients and their families. We also thank Dr Hirotsugu Watabe
and Dr Shiro Oka for help with central review; Mariko Takeda (data
January 2020
management), Yumi Tanaka and Ai Okazaki (clinical monitoring planning and
monitoring), Ikue Wada (clinical monitoring), Maki Kobayashi (trial
coordinator), and Yuki Kusaka (audit); Dr Tomohiro Tada for help with
providing anonymization software for central review; and the trial
investigators for their roles in the trial: Dr Yoshihiro Hirata, Dr Nobutake
Yamamichi, Dr Shinya Kodashima, Dr Satoshi Ono, Dr Yosuke Tsuji, Dr
Shuntaro Yoshida, Dr Keiko Niimi, Dr Yoku Hayakawa, Dr Takeshi
Yoshikawa, Dr Hiroto Kinoshita, Dr Yumiko Ota, Dr Takayuki Shinpo, Dr
Yoshiki Sakaguchi, Dr Yu Takahashi, Dr Sozaburo Ihara, Dr Itaru Saito, Dr
Yosuke Kataoka, Dr Ayako Nakada, Dr Hikaru Kakimoto, Dr Chihiro
Takeuchi, Dr Yuta Matsumoto, Dr Seiichi Yakabi, Dr Hiroya Mizutani, Dr
Daisuke Ohki, Dr Kenta Gondo, Dr Mitsuru Konishi, Dr Rei Ishibashi, Dr
Tomonori Aoki, and Dr Itsuko Hirayama of The University of Tokyo; Dr Kohei
Hayashi and Dr Ryousuke Hirata of Nagasaki Harbor Medical Center; Dr
Takashi Ikeya, Dr Kenji Nakamura, and Dr Katsuyuki Fukuda of St. Luke’s
International Hospital; Dr Ryosuke Ota and Dr Akihiro Dejima of Ishikawa
Prefectural Central Hospital; Dr Dai Nakamatsu and Dr Aya Sugimoto of
Toyonaka Municipal Hospital; Dr Takeyoshi Minagawa, Dr Yutaka Okagawa,
Dr Ryoji Fujii, and Dr Masahiro Yoshida of Tonan Hospital; Dr Yusuke Kanari
of Otaru Ekisaikai Hospital; and Dr Hirotsugu Saiki of Japan Community
Healthcare Organization Osaka Hospital. Thane Doss contributed to editing
of the revised manuscript.
Early vs Elective Colonoscopy 175
Author contributions: Ryota Niikura, Naoyoshi Nagata, and Atsuo Yamada
are co-first authors and contributed equally to the design of the study and
writing the manuscript. Munenori Takata established the clinical monitoring
plan and performed the data management. Tomohiro Shinozaki wrote the
statistical analysis plan and statistics section of the manuscript and
performed the statistical analysis. Hisashi Doyama, Yasutoshi Shiratori,
Tsutomu Nishida, Shu Kiyotoki, Tomoyuki Yada, Tomoki Fujita, Tetsuya
Sumiyoshi, Kenkei Hasatani, Tatsuya Mikami, Tetsuro Honda, Katsuhiro
Mabe, Kazuo Hara, Katsumi Yamamoto, Mitsuhiro Fuhishiro, and Kazuhiko
Koike advised on the design of the study and contributed to writing of the
manuscript. All authors read and approved the final manuscript.
Conflicts of interest
This author discloses the following: Mitsuhiro Fujishiro has received lecture
honoraria from Takeda Pharmaceutical, AstraZeneca, Zeria, and Daiichi-
Sankyo; research grants for unrestricted purposes from Takeda
CLINICAL AT
Pharmaceutical and EA Pharma; and collaborative funding from HOYA-
Pentax. The remaining authors disclose no conflicts.
Funding
The authors received a grant from the Japanese Gastroenterological
Association Foundation.
175.e1 Niikura et al Gastroenterology Vol. 158, No. 1

Supplementary Figure 1. Eligibility assessment, intervention, randomization, follow-up, and analysis of the patients.

Supplementary Figure 2. SRH detection

by time from first visit to colonoscopy.
January 2020 Early vs Elective Colonoscopy 175.e2

Supplementary Table 1.Primary and Secondary Outcomes in Per-Protocol Analysis

Early colonoscopy Elective colonoscopy

Outcome (n ¼ 78) (n ¼ 79) Difference P value

Primary outcome, n (%)

SRH identification 17 (21.8) 16 (20.3) 1.5 (–11.2 to 14.3) .813
SRH identification by central review 20 (25.6) 20 (25.3) 0.3 (–13.3 to 14.0) .963
Secondary outcomes
30-day rebleeding, n (%)a,b 10 (14.5) 4 (5.9) 8.6 (–1.4 to 18.6) .096
Success rate of endoscopic 14/15 (93.3) 14/14 (100.0) –6.7 (–19.3 to 6.0) .326
treatment, n/total, (%)c
Success rate of endoscopic treatment 8/10 (80.0) 6/8 (75.0) 5.0 (–0.3 to 0.4) .800
for active bleeding, n/total, (%)d
Need for additional endoscopic 31 (39.7) 23 (29.1) 10.6 (–4.2 to 25.4) .161
examinations, n (%)
Need for interventional radiology, n (%) 1 (1.3) 0 (0.0) 1.3 (–1.2 to 3.8) .313
Need for surgery, n (%) 0 (0.0) 0 (0.0) 0.0 N/A
Need for transfusion during hospitalization, n (%) 30 (38.5) 26 (32.9) 5.6 (–9.4 to 20.5) .086
Length of stay, d, mean ± SD 7.2 ± 5.8 7.6 ± 6.0 –0.5 (–2.4 to 1.4) .122
30-day thrombotic event, n (%)b,e 1 (1.5) 0 (0.0) 1.5 (–1.4 to 4.3) .319
30-day mortality, n (%)b 0 (0.0) 0 (0.0) 0.0 N/A

N/A, not applicable; SD, standard deviation.

a
Rebleeding was defined as significant fresh blood loss after an initial colonoscopy with any of the following: (1) hemorrhagic
shock, including cold sweats, nausea, syncope, or systolic blood pressure  90 mm Hg; (2) need for transfusion, according to
the guidelines of the Ministry of Health, Labor and Welfare; (3) further colonoscopy identifying blood pooling; (4) SRH in the
lower gastrointestinal tract; or (5) contrast-enhanced computed tomography identifying extravasation in the colorectal region.
b
Rebleeding, thrombotic events, and mortality were evaluated for patients who had approximately met the 30-day follow-up
criteria (early colonoscopy for 69 patients and elective colonoscopy for 68 patients).
c
Success rate of endoscopic treatment was defined as achievement of hemostasis (early colonoscopy for 15 patients and
elective colonoscopy for 14 patients).
d
Success rate of endoscopic treatment for active bleeding was defined as achievement of hemostasis for active bleeding with
SRH (early for 10 patients and elective for 8 patients).
e
Thrombotic events include acute coronary syndromes, including angina pectoris and myocardial infarction; stroke, including
cerebrovascular infarction, cerebral hemorrhage, and transient ischemic attacks; deep vein thrombosis, and pulmonary
embolism.
175.e3 Niikura et al Gastroenterology Vol. 158, No. 1

Supplementary Table 2.Subgroup Analysis of the Risk Difference in Identification of SRH

P value
Subgroup (early/elective CS) Early CS, % Elective CS, % Difference for interaction

Prespecified analysis
Patients with colonic diverticular bleeding
Yes (n ¼ 47/52) 21.3 30.8 –9.5 (–26.7 to 7.7) .020
No (n ¼ 32/28) 21.9 3.6 18.3 (2.4 to 34.2)
Withdrawal from the trial because of
inadequate bowel preparation
Yes (n ¼ 8/8) 37.5 12.5 25.0 (–15.6 to 65.6) .207
No (n ¼ 71/72) 19.7 22.2 –2.5 (–15.8 to 10.8)
Colonoscopy performed by an expert
Expert only (n ¼ 48/48) 25.0 22.9 2.1 (–15.0 to 19.2) .716
Including nonexpert (n ¼ 31/32) 16.1 18.8 –2.6 (–21.3 to 16.1)
Colonoscopy performed within 24 h
of onset of hematochezia
Yes (n ¼ 36/1) 27.8 100.0 –72.2 (–86.9 to –57.6) <.001
No (n ¼ 43/79) 16.3 20.3 –4.0 (–18.1 to 10.2)
Each participating site
Site sequence number 1 (n ¼ 35/33) 22.9 6.1 16.8 (0.7 to 32.9) .751a
Site sequence number 2 (n ¼ 6/8) 33.3 37.5 –4.2 (–54.7 to 46.3)
Site sequence number 3 (n ¼ 3/5) 33.3 0.0 33.3 (–20.0 to 86.7)
Site sequence number 4 (n ¼ 7/5) 14.3 40.0 –25.7 (–75.9 to 24.4)
Site sequence number 5 (n ¼ 1/1) 0.0 0.0 N/A
Site sequence number 6 (n ¼ 2/1) 100.0 0.0 100.0 (100.0 to 100.0)
Site sequence number 7 (n ¼ 6/2) 16.7 0.0 16.7 (–13.2 to 46.5)
Site sequence number 8 (n ¼ 0/0) 0.0 0.0 N/A
Site sequence number 9 (n ¼ 1/5) 0.0 80.0 –80.0 (–100.0 to 44.9)
Site sequence number 10 (n ¼ 6/8) 0.0 12.5 –12.5 (–35.4 to 10.4)
Site sequence number 11 (n ¼ 0/0) 0.0 0.0 N/A
Site sequence number 12 (n ¼ 0/0) 0.0 0.0 N/A
Site sequence number 13 (n ¼ 10/10) 20.0 40.0 –20.0 (–59.2 to 19.2)
Site sequence number 14 (n ¼ 2/1) 0.0 100.0 –100.00 (–100.0 to –100.0)
Site sequence number 15 (n ¼ 0/1) 0.0 0.0 N/A
Non-prespecified analysis
Age subgroup
<70 y (n ¼ 30/28) 16.7 25.0 –8.3 (–29.2 to 12.5) .312
70 y (n ¼ 49/52) 24.5 19.2 5.3 (–10.9 to 21.4)
Sex
Male (n ¼ 52/54) 21.2 24.1 –2.9 (–18.8 to 13.0) .467
Female (n ¼ 27/26) 22.2 15.4 6.8 (–14.1 to 27.8)
Medication: low-dose aspirin (n ¼ 16/22) 25.0 22.7 2.3 (–25.2 to 29.8) .998a
NSAIDs (n ¼ 11/11) 27.3 27.3 0.0 (–37.2 to 37.2)
Thienopyridine (n ¼ 7/6) 14.3 0.0 14.3 (–11.6 to 40.2)
Cilostazol (n ¼ 5/2) 40.0 0.0 40.0 (–2.9 to 82.9)
Other antiplatelet drugs (n ¼ 5/3) 40.0 33.3 6.7 (–61.8 to 75.2)
Warfarin (n ¼ 4/6) 50.0 33.3 16.7 (–45.2 to 78.5)
Direct oral anticoagulants (n ¼ 3/7) 33.3 28.6 4.8 (–58.2 to 67.7)
Timing of colonoscopy from initial visit
to colonoscopy performed:
12 h (n ¼ 78/1) 21.8 100.0 –78.2 (–87.4 to –69.0) <.001a
12–48 h (n ¼ 1/57) 0.0 28.1 –28.1 (–39.7 to –16.4)
>48 h (n ¼ 0/22) 0.0 0.0 N/A
Severity of bleeding
Severe bleeding (n ¼ 31/21) 22.6 28.6 –6.0 (–30.3 to 18.3) .578a
Moderate bleeding (n ¼ 48/59) 20.8 18.6 2.2 (–13.0 to 17.4)

CS, colonoscopy; N/A, not applicable; NSAIDs, nonsteroidal anti-inflammatory drugs.

a
Global tests for interaction between multiple subgroups, assessed by the type 3 Wald tests in the linear binomial (risk-
difference) models.
January 2020 Early vs Elective Colonoscopy 175.e4

Supplementary Table 3.Subgroup Analysis of the Risk Difference in 30-Day Rebleeding

P value
Subgroup (early/elective CS) Early CS, % Elective CS, % Difference for interaction

Prespecified analysis
Patients with colonic diverticular bleeding
Yes (n ¼ 43/49) 16.3 8.2 8.1 (–5.3 to 21.6) .856
No (n ¼ 29/26) 13.8 3.8 9.9 (–4.6 to 24.5)
Withdrawal from the trial because of
inadequate bowel preparation
Yes (n ¼ 8/7) 12.5 0 12.5 (–10.4 to 35.4) .744
No (n ¼ 64/68) 15.6 7.4 8.3 (–2.6 to 19.1)
Patients who underwent endoscopic hemostasis
Yes (n ¼ 14/12) 14.3 16.7 –2.4 (–30.3 to 25.6) .389
No (n ¼ 58/63) 15.5 4.8 10.8 (0.1 to 21.5)
Colonoscopy performed by an expert
Expert only (n ¼ 45/46) 22.2 6.5 15.7 (1.6 to 29.8) .043
Including nonexpert (n ¼ 27/29) 3.7 6.9 –3.2 (–14.8 to 8.5)
Colonoscopy performed within 24 h
of onset of hematochezia
Yes (n ¼ 34/1) 17.6 0 17.6 (4.8 to 30.5) .212
No (n ¼ 38/74) 13.2 6.8 6.4 (–5.8 to 18.6)
Each participating site
Site sequence number 1 (n ¼ 35/33) 8.6 6.1 2.5 (–9.8 to 14.9) .751a
Site sequence number 2 (n ¼ 6/6) 0 16.7 –16.7 (–46.5 to 13.2)
Site sequence number 3 (n ¼ 3/5) 33.3 0 33.3 (–20.0 to 86.7)
Site sequence number 4 (n ¼ 2/3) 0 0 N/A
Site sequence number 5 (n ¼ 1/1) 100.0 0.0 100.0 (100.0 to 100.0)
Site sequence number 6 (n ¼ 2/1) 0 0 N/A
Site sequence number 7 (n ¼ 6/2) 16.7 0.0 16.7 (–13.2 to 46.5)
Site sequence number 8 (n ¼ 0/0) 0.0 0.0 N/A
Site sequence number 9 (n ¼ 1/5) 0.0 20.0 –20.0 (–55.1 to 15.1)
Site sequence number 10 (n ¼ 6/7) 16.7 0 16.7 (–13.2 to 46.5)
Site sequence number 11 (n ¼ 0/0) 0 0 N/A
Site sequence number 12 (n ¼ 0/0) 0 0 N/A
Site sequence number 13 (n ¼ 8/10) 50.0 10.0 40.0 (0.7 to 79.3)
Site sequence number 14 (n ¼ 2/1) 0 0 N/A
Site sequence number 15 (n ¼ 0/1) 0 0 N/A
Non-prespecified analysis
Age subgroup
<70 years (n ¼ 25/27) 8.0 11.1 –3.1 (–19.0 to 12.8) .081
70 years (n ¼ 47/48) 19.1 4.2 15.0 (2.4 to 27.6)
Sex
Male (n ¼ 46/51) 17.4 7.8 9.5 (–3.7 to 22.8) .829
Female (n ¼ 26/24) 11.5 4.2 7.4 (–7.3 to 22.0)
Medication
Low-dose aspirin (n ¼ 14/22) 35.7 9.1 26.6 (–1.2 to 54.4) .998a
NSAIDs (n ¼ 14/16) 14.3 0 14.3 (–4.0 to 32.6)
Thienopyridine (n ¼ 7/6) 28.6 16.7 11.9 (–32.9 to 56.7)
Cilostazol (n ¼ 5/2) 0 0 N/A
Other antiplatelet drugs (n ¼ 4/3) 0 33.3 –33.3 (–86.7 to 20.0)
Warfarin (n ¼ 4/6) 25.0 0 25.0 (–17.4 to 67.4)
Direct oral anticoagulants (n ¼ 3/6) 0 0 N/A
Timing of colonoscopy from initial visit
to colonoscopy performed
12 h (n ¼ 71/1) 15.5 0 15.5 (7.1 to 23.9) .490a
12–48 h (n ¼ 1/52) 0.0 9.6 –9.6 (–17.6 to –1.6)
>48 h (n ¼ 0/22) 0 0 N/A
Severity of bleeding
Severe bleeding (n ¼ 29/18) 10.3 5.6 4.8 (–10.5 to 20.1) .513a
Moderate bleeding (n ¼ 43/57) 18.6 7.0 11.6 (–1.8 to 25.0)

CS, colonoscopy; N/A, not applicable; NSAIDs, nonsteroidal anti-inflammatory drugs.

a
Global tests for interaction between multiple subgroups, assessed by the type 3 Wald tests in the linear binomial (risk-
difference) models.
175.e5 Niikura et al Gastroenterology Vol. 158, No. 1

Supplementary Table 4.Baseline Characteristics in Per-Protocol Analysis

Early colonoscopy Elective colonoscopy

Variable (n ¼ 78) (n ¼ 79)

Age, y, mean ± SD 68.6 ± 12.7 71.8 ± 12.3

Sex, male, n (%) 51 (65.4) 54 (68.4)
Body mass index, kg/m2, mean ± SD 23.0 ± 3.3 24.0 ± 4.0
Comorbidities, n (%)
Previous lower GI bleeding 27 (34.6) 29 (36.7)
Ischemic heart disease 12 (15.4) 17 (21.5)
Chronic obstructive pulmonary disease 3 (3.8) 0 (0.0)
Peptic ulcer 6 (7.7) 2 (2.5)
Liver cirrhosis 1 (1.3) 2 (2.5)
Diabetes mellitus 11 (14.1) 16 (20.3)
Chronic heart failure 4 (5.1) 4 (5.1)
Cerebrovascular disease 11 (14.1) 15 (19.0)
Dementia 0 (0.0) 2 (2.5)
Collagen disease 4 (5.1) 5 (6.3)
Chronic kidney disease 8 (10.3) 11 (13.9)
Leukemia 0 (0.0) 1 (1.3)
Malignant lymphoma 2 (2.6) 0 (0.0)
Solid cancer 7 (9.0) 12 (15.2)
Medication, n (%)
Low-dose aspirin 16 (20.5) 22 (27.8)
Thienopyridine 7 (9.0) 6 (7.6)
Cilostazol 5 (6.4) 2 (2.5)
Other antiplatelet drugs 5 (6.4) 3 (3.8)
Warfarin 4 (5.1) 6 (7.6)
Direct oral anticoagulants 3 (3.8) 6 (7.6)
NSAIDs 13 (16.7) 16 (20.3)
Initial assessment
Hemodynamic instability, n (%)a 4 (5.1) 1 (1.3)
Hemoglobin, g/dL, mean ± SD 11.3 ± 2.7 11.3 ± 2.3
Upper GI endoscopy before colonoscopy, n (%) 2 (2.6) 1 (1.3)
Severity of bleeding, n (%)
Severe bleeding 31 (39.7) 21 (26.6)
Moderate bleeding 47 (60.3) 58 (73.4)
Endoscopic procedure
Expert endoscopist, n (%)b 48 (61.5) 47 (59.5)
Cecal insertion, n (%) 76 (97.4) 76 (96.2)
Cecal insertion time, min, mean ± SD 9.6 ± 6.5 8.8 ± 6.3
Total procedure time, min, mean ± SD 35.7 ± 19.3 31.2 ± 15.2

GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs; SD, standard deviation.

a
Hemodynamic instability was defined as a heart rate >100 beats/min and systolic blood pressure <115 mm Hg.25
b
An expert endoscopist was defined as one who had performed >1000 colonoscopies including endoscopic hemostasis.
January 2020 Early vs Elective Colonoscopy 175.e6

Supplementary Table 5.Bleeding Source in Per-Protocol Analysis

Early colonoscopy Elective colonoscopy

Bleeding source (n ¼ 78), n (%) (n ¼ 79), n (%) P value

Diverticular bleeding (definite)a 10 (12.8) 15 (19.0) .291

Diverticular bleeding (presumptive)b 37 (47.4) 36 (45.6) .815
Rectal ulcer 0 (0.0) 0 (0.0) N/A
Colorectal cancer 3 (3.8) 2 (2.5) .639
Ischemic colitis 9 (11.5) 5 (6.3) .252
Infectious colitis 1 (1.3) 0 (0.0) .313
Radiation colitis 1 (1.3) 0 (0.0) .313
Colonic ulcer 0 (0.0) 2 (2.5) .157
Hemorrhoid 1 (1.3) 2 (2.5) .567
Otherc 7 (9.0) 4 (5.1) .337
Unknown 12 (15.4) 13 (16.5) .855
Upper GI bleeding 0 (0.0) 1 (1.3) .319

NOTE. Multiple sources are possible.

GI, gastrointestinal; N/A, not applicable.
a
Definite diverticular bleeding source was defined as diverticula with SRH.
b
Presumptive diverticular bleeding source was defined as diverticula without SRH.
c
Other included small intestinal bleeding for postendoscopic mucosal resection bleeding (n ¼ 1 in the early colonoscopy
group), colonic diverticulitis (n ¼ 1 in the elective colonoscopy group), angioectasia (n ¼ 3 in the early colonoscopy group; n ¼
2 in the elective colonoscopy group), and small intestinal bleeding (n ¼ 3 in the early colonoscopy group; n ¼ 1 in the elective
colonoscopy group).

Supplementary Table 6.The Quality of Colon Preparation

Early CS Elective CS P value

Score Description (n ¼ 78), n (%) (n ¼ 79), n (%) for trend

5 Inadequate (repeat preparation needed) 0 (0.0) 0 (0.0) .1469

4 Poor (semisolid stool could not be suctioned, 0 (0.0) 0 (0.0)
and <90% of mucosa seen)
3 Fair (semisolid stool could not be suctioned, 3 (3.8) 0 (0.0)
but >90% of mucosa seen)
2 Good (clear liquid covering up to 25% of mucosa, 8 (10.3) 7 (8.9)
but >90% of mucosa seen)
1 Excellent (>95% of mucosa seen) 67 (85.9) 72 (91.1)

NOTE. The scores of 1 patient in the early CS group and 1 patient in the elective CS group were missing, but these data were
the same as for the per-protocol analysis population. The trend P value of the Cochran-Armitage test was calculated.
CS, colonoscopy.