Introduction
Diverticulosis is a common disease initially thought to affect elderly individuals,
with a prevalence of 70% in people aged >70 years and an increasing prevalence with
age [1 ]
[2 ]. However, a growing number of younger individuals are affected by the disease [3 ], and the lifetime prevalence is estimated to be 72% [4 ], making diverticula the most common lesion found incidentally on routine colonoscopy
[5 ]. Nonetheless, it is estimated that only about 20% of individuals will develop symptoms
in their lifetime, and <5% will progress from colonic diverticulosis to diverticulitis
[6 ]
[7 ]. Yet, this condition can lead to hospitalization and potentially life-threatening
complications for patients, such as perforation and bleeding [1 ]. Furthermore, the economic burden on health care services is significant, although
most likely underestimated due to a substantial number of uncomplicated cases being
managed outside hospitals [8 ]
[9 ]
[10 ]. When patients are admitted to the hospital, diverticulitis is typically classified
by the modified Hinchey classification based on severity as determined by a computed
tomography (CT) scan [11 ]
[12 ]
[13 ]; if there are no findings indicating the need for further treatment, the patients
are discharged.
The current standard procedure in the Danish health care system, regardless of the
severity of the disease, is that patients are referred for a follow-up colonoscopy
4–6 weeks after discharge following a colonic diverticulitis episode. This aims to
confirm the diagnosis of diverticulitis and exclude malignancies. However, this approach
is being questioned because studies have found very low rates of colorectal neoplasia
in this patient group [14 ]
[15 ]. In colonoscopies performed only for diverticulitis, a low colorectal cancer (CRC)
prevalence of 0.17% has been found [16 ]. An alternative to a follow-up colonoscopy could be colon capsule endoscopy (CCE)
([Fig. 1 ]). This procedure has gained increased acceptance in recent years [17 ]
[18 ], owing to patient preference, a low risk of complications, and the ability to be
performed out of hospital [19 ]
[20 ]. However, no research on the use of CCE as a follow-up procedure in patients with
diverticulitis has been published.
Fig. 1 Colon capsule endoscopy image of colonic diverticulitis.
Therefore, this randomized clinical trial aimed to investigate the patient experience
and clinical performance of using CCE compared with colonoscopy as a follow-up examination
in patients with diverticulitis.
Methods
Trial design
This was a single-center, 1:1 randomized controlled superiority trial conducted at
the Surgical and Emergency departments at Odense University Hospital in the Region
of Southern Denmark. The trial randomized patients presenting with diverticulitis
to either CCE (intervention) or colonoscopy (standard procedure). In cases of incomplete
CCE or when CCE identified significant neoplasia, patients of the intervention arm
underwent a subsequent colonoscopy.
Patients
We included patients aged ≥18 years with in-hospital CT-diagnosed diverticulitis (uncomplicated
or complicated diverticulitis modified Hinchey grades 1–3). Key exclusion criteria
were colonic CT findings requiring biopsy (suspected CRC) or polyp removal, CT-verified
stenosis in the gastrointestinal tract, imaging of colonic mucosa within the last
12 months (therefore no indication for renewed endoscopy), cardiac pacemaker, renal
insufficiency, pregnancy/breastfeeding, allergies to active substances given in the
trial, or inability to provide oral and written informed consent.
Interventions
The intervention group included patients randomized to CCE. Patients underwent bowel
preparation at home, starting 72 hours before CCE. The bowel preparation kit was sent
by mail and contained polyethylene glycol (PEG) solutions as well as instructions
on how to perform the preparation.
At 72 hours and 48 hours before capsule ingestion, patients took two sachets of PEG
and 2 L of water but otherwise remained on a normal diet. The day before and on the
day of the procedure, patients ingested 1 L of PEG and 1 L of water with a clear liquid
diet.
Before capsule ingestion, patients ingested a 2-mg tablet of prucalopride. After capsule
ingestion, patients ingested boosters (sulfate-based saline solution) mixed with water
at receiver signals 1–3. The signals were as follows: signal 1 indicated that the
capsule had reached the small bowel; signals 2–4 – every 2 hours after signal 1. After
the last signal (signal 4), the patients inserted a bisacodyl suppository. The capsule
was typically excreted 4–6 hours after ingestion. The CCE investigation was considered
complete if the capsule was excreted within the battery lifetime or there was a recording
of the anal cushions in patients with acceptable bowel preparation. The bowel preparation
was evaluated using the Leighton–Rex grading scale and was determined acceptable if
it was fair or better for all colonic segments [21 ]. If the CCE investigation was incomplete for any reason, the patient was referred
for colonoscopy.
If the CCE video showed suspected cancer, more than two polyps, or any number of polyps
>9 mm, patients were invited to undergo a diagnostic colonoscopy for biopsies or polyp
removal; colonoscopy was performed according to the protocol described below. Any
findings on CCE showing no neoplasia or low-risk polyps (defined as 1–2 polyps <10
mm) would elicit no further investigation [22 ]
[23 ]
[24 ]; however, individuals with those results were encouraged to participate in future
CRC screening programs if eligible. All CCE videos were assessed by the same expert
gastroenterologist, who had experience of evaluating >4000 CCE videos.
The control arm consisted of patients who, following the local guidelines at the Department
of Surgery, Odense University Hospital, underwent colonoscopy. The bowel preparation
consisted of a PEG+ascorbic acid split-dose regimen, involving two doses: one the
day before the procedure and one on the day of the procedure. Five days before the
procedure, patients were instructed to eat a diet without seeds and kernels, and the
day before the procedure they were instructed not to eat solid foods, and to only
drink clear liquids until 2 hours before the procedure, at which point they began
fasting. The specific dose of medication and the degree of sedation achieved (conscious
sedation using midazolam and pethidine, sedation assisted by anesthesiologists using
propofol, or general anesthesia) were registered.
Outcomes
The primary outcome was patient-reported experienced physical and mental discomfort
from colonoscopy and CCE. Secondary outcomes were 1) patient-reported expected physical
and mental discomfort associated with colonoscopy and CCE (i.e. examination and the
bowel preparation), 2) the discrepancy between the expected and experienced physical
and mental discomfort associated with colonoscopy and CCE, 3) the number of complete
examinations, 4) the number of patients undergoing CCE who were referred for a subsequent
colonoscopy, 5) the prevalence of other abnormal findings, such as polyps and cancers.
To ensure patient safety, complications, adverse events, and investigation quality
were monitored continuously by the trial manager during the study.
The expected and experienced physical and mental discomfort scores were measured through
questionnaires. Enrolled patients were asked to complete two or three questionnaires
before and after their examinations. The first questionnaire (Q1) measured the expected
physical and mental discomfort associated with both CCE and colonoscopy. The second
questionnaire (Q2) measured the experienced discomfort associated with CCE and was
therefore only distributed to those randomized to CCE. The third questionnaire (Q3)
measured the experienced discomfort associated with colonoscopy and was therefore
only distributed to individuals referred for colonoscopy after their CCE and to those
randomized to colonoscopy.
Additionally, patients were contacted by phone by one of the affiliated researchers
7 days after CCE or colonoscopy and asked which examination they would prefer if they
needed to undergo a future endoscopic examination, and to state the reason(s) for
their choice. All questionnaires were qualitatively validated for the CareForColon2015
trial [25 ] and then adapted for the current trial.
Sample size and power considerations
In a previous randomized trial on CCE in a bowel cancer screening population (not
yet published) using the same visual analog scale (VAS) as the current trial, the
interim analysis of patient-reported outcomes found a mean discomfort score of 18.6
for CCE and 32.2 for colonoscopy (SD 26.5). Applying this difference to the current
study (with an SD of 26.5), and with a level of statistical significance of 5% and
a power of 80%, a minimum of 60 patients were needed in each group [26 ]. As this power calculation was performed based on a large sample from the previous
trial, we have learnt that the distribution of the experienced discomfort scores tends
not to follow a normal or log-normal distribution. With no suitable transformations
available, we performed a simulation with 10 000 repetitions of drawing a random sample
of 60 from each group from these data. We compared the medians of the experienced
discomfort between CCE and colonoscopy for each sample. This enabled us to estimate
the power of the median test to detect a difference in medians of at least 17 to be
approximately 77% at the two-sided 5% significance level. The simulation was performed
using Stata IC 15.1 (StataCorp., College Station, Texas, USA) [27 ].
Statistical analysis
The intention-to-treat [28 ] population was defined as patients undergoing their allocated randomized procedure.
All analyses were for complete cases (i.e. no imputation of missing values was planned).
The primary outcome was measured using VAS. The VAS scores were treated as discrete
variables with a range of 101 values (0–100, where a score of 0 means no discomfort,
and a score of 100 means a very high degree of discomfort). As no normal or log-normal
distribution of the recorded VAS scores was present, non-parametric tests were applied.
The analyses were modified intention-to-treat analyses, excluding patients who, by
their own or physician’s decision, withdrew from the study. For the primary outcome,
patients had to have reported their experienced physical and mental discomfort of
CCE and colonoscopy, respectively, to be included in the analysis. Univariate comparison
for the primary outcome was performed using the Wilcoxon rank sum test in SAS 9.4,
TS1M5 (SAS Institute Inc., Cary, North Carolina, USA) [29 ]. Additionally, continuous ordinal regression models were conducted to test differences
in VAS scores [30 ]
[31 ] using the ordinalCont R package in RStudio, R version 4.1.2 (R Foundation for Statistical
Computing, Vienna, Austria) [32 ]
[33 ]. The secondary outcomes of expected discomfort, and discrepancy between expected
and experienced discomfort, were performed using the Wilcoxon rank sum test, while
examination completion rate, and the prevalence of diverticula, polyps, cancers, and
other abnormal findings were estimated as proportions and compared between the study
arms using the chi-squared test. Multiple statistical testing was not considered.
Ethics
The Regional Committees on Health Research Ethics for Southern Denmark (ref. S-20210127)
and the Danish Data Protection Agency (ref. 22/43235) approved the trial protocol.
Our trial protocol paper was published in 2023 [34 ]. All patients provided informed consent.
Results
Patients
We enrolled patients in the trial between 4 December 2022 and 29 January 2024 at the
Surgical Department and Emergency Department of Odense University Hospital in the
Region of Southern Denmark. Of the 159 patients who were randomized, 148 received
their allocated intervention, and 83 completed the clinical evaluation at week 4,
answering either questionnaire 2 and/or 3 ([Fig. 2 ]). The demographic and clinical characteristics of the two groups were comparable
at baseline ([Table 1 ]), except for some imbalance in sex and American Society of Anesthesiologists score.
Fig. 2 Flow of patients through the study. CCE, colon capsule endoscopy; ITT, intention-to-treat.
1 Missed referral, received CCE instead of colonoscopy due to operator error. 2 Two individuals did not receive questionnaire 2 due to operator error.
Table 1 Baseline characteristics of patients randomized to undergo colon capsule endoscopy
or colonoscopy after an episode of colonic diverticulitis (intention-to-treat population
N = 136).
Characteristics
CCE (n = 61)
Colonoscopy (n = 75)
Total (n = 136)
ASA, American Society of Anesthesiologists; BMI, body mass index; CCE, colon capsule
endoscopy; CT, computed tomography.
1 Diverticula, thickening of the wall, increased density of the pericolic fat.
2 >5 cm from inflamed bowel segment.
Female sex, n (%)
39 (64)
54 (72)
93 (68)
Age, mean (SD), years
60.0 (11.1)
60.6 (10.4)
60.3 (10.7)
BMI, kg/m2
0 (0)
1 (2)
1 (1)
16 (31)
15 (25)
31 (28)
36 (69)
43 (73)
79 (71)
9
16
25
ASA score, n (%)
16 (26)
9 (12)
25 (18)
45 (74)
66 (88)
111 (82)
CT classification, n (%)
44 (72)
52 (69)
96 (71)
13 (21)
19 (25)
32 (24)
4 (7)
4 (5)
8 (6)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Primary outcomes
The median VAS score for experienced physical and mental discomfort during the diagnostic
investigation was not higher for the patients who underwent colonoscopy (94, interquartile
range [IQR] 67) than for those who underwent CCE (75, IQR 71), as the difference was
not statistically significant (P = 0.26).
Secondary outcomes
We observed a significantly higher median VAS score for expected physical and mental
discomfort of diagnostic investigations in the patients randomized to colonoscopy
than in those assigned to CCE (95 [IQR 71] vs. 28 [IQR 56]; P < 0.001) ([Table 2 ]). The discrepancy between the median VAS for expected and experienced physical and
mental discomfort was significantly higher for patients allocated to CCE (39, IQR
72) than those assigned to colonoscopy (4, IQR 46; P = 0.002).
Table 2 Outcomes of colon capsule endoscopy and colonoscopy after an episode of colonic diverticulitis
in patients completing questionnaires (n = 83), and in all patients (n = 136).
Outcome
CCE
Colonoscopy
P value1
CCE, colon capsule endoscopy; IQR, interquartile range; VAS, visual analog scale.
1 Chi-squared test.
2 100% had complete capsule colon transit, but 7 had poor bowel preparation, and 3 had
technical errors.
3 Out of the 20 with CCE-detected polyps, 14 underwent colonoscopy, of whom 7 had colonoscopy-identified
polyps.
Patients completing the questionnaire(s), n
39
44
Outcomes, VAS, median (IQR)
75 (71)
94 (67)
0.26
28 (56)
95 (71)
<0.001
39 (72)
4 (46)
0.002
All patients, n
61
75
Outcomes, n (%)
51 (84)2
69 (92)
0.13
20 (33)
21 (28)
0.55
7 (11)3
NA
NA
0 (0)
0 (0)
NA
0 (0)
0 (0)
NA
9 (15)
23 (31)
0.030
17 (28)
NA
NA
The multivariate models showed a significantly lower probability of higher expected
physical and mental discomfort with CCE compared with colonoscopy (odds ratio [OR]
0.12, 95%CI 0.05–0.29), and a considerably higher likelihood of a difference between
expected and experienced physical and mental discomfort in the CCE group compared
with the colonoscopy group (OR 3.92, 95%CI 1.77–8.71). However, colonoscopy was not
associated with a significantly higher probability of experienced physical and mental
discomfort than CCE ([Fig. 3 ]).
Fig. 3 Odds ratios for higher expected physical and mental discomfort, experienced physical
and mental discomfort, and discomfort discrepancy compared with the reference (n =
83).
The rate of complete colonoscopy examination was 92% ([Table 2 ]). All CCEs (100%) showed complete transit; however, seven were associated with inadequate
bowel preparation and three experienced technical errors, giving a complete examination
rate of 84%. In total, 33% of patients undergoing CCE and 28% of patients undergoing
colonoscopy had polyp findings. Only 11% of patients with polyp findings on CCE had
the same polyps detected again on subsequent colonoscopy. Other abnormal findings
were present in 31% of patients undergoing colonoscopy and 15% of patients undergoing
CCE, while 28% of patients in the CCE group had other abnormal findings on CCE and/or
follow-up colonoscopy. There were no findings suspicious for cancer with either procedure.
When asked about preferences for future endoscopic examinations, 49% of patients (48/98
responders) preferred CCE compared with only 13% (13/98) who preferred colonoscopy
([Table 3 ]). Furthermore, CCE was preferred to colonoscopy in patients who underwent CCE only
(68%) and in patients who underwent colonoscopy only (43%). However, 43% of patients
who underwent both investigations would prefer colonoscopy in the future.
Table 3 Preferences for hypothetical future endoscopic examination.
Preference for future endoscopic examination
Sample
CCE, n (%)
Colonoscopy, n (%)
Do not know, n (%)
CCE, colon capsule endoscopy.
All responders (n = 98)
48 (49)
13 (13)
37 (38)
Responders undergoing only CCE (n = 28)
19 (68)
3 (11)
6 (21)
Responders undergoing only colonoscopy(n = 56)
24 (43)
4 (7)
28 (50)
Responders undergoing both investigations (n = 14)
5 (36)
6 (43)
3 (21)
Patients were further asked about the reasons for their future choices. Of the patients
who would choose CCE for future endoscopic examinations, the most common explanation
for their choice was to experience less physical and mental discomfort, as many described
having severe pain, both during and after the colonoscopy. Another common answer was
that colonoscopy could be intrusive, often leading to feelings of vulnerability and
embarrassment, and issues related to personal privacy as the examination involves
an exposure that many patients found to be challenging. Of the patients choosing colonoscopy,
most of the explanations for the choice were unfamiliarity with CCE, whereas colonoscopy
was more familiar and considered to be the safe and easy choice. A few patients also
mentioned the possibility of taking samples during colonoscopy, which is not possible
with the less invasive CCE, and seemed especially appealing for those patients with
particular concerns. Regardless of the examination, a common response was that the
bowel preparation was the worst part of the whole process.
Discussion
To the best of our knowledge, this is the first study to investigate the performance
of CCE compared with colonoscopy as a follow-up procedure after an episode of acute
colonic diverticulitis. No statistically significant differences in experienced physical
and mental discomfort were found between the two examinations. Expected physical and
mental discomfort was lower for patients undergoing CCE compared with those undergoing
colonoscopy. Patients preferred CCE except for those undergoing both procedures (n
= 14).
One major strength of this study is its design as a randomized controlled trial. However,
it is limited by the single-center design and the fact that one reader assessed all
CCE videos, which establishes a risk of detection bias. Furthermore, only 55% (39/71)
of patients who underwent allocated CCE and 57% (44/77) of patients who underwent
allocated colonoscopy completed both baseline and 4-week questionnaires, which presents
a risk of selection bias.
In a systematic review from 2021 investigating patient-reported outcomes and preferences
for CCE and colonoscopy, the tolerability for CCE was consistently reported as being
higher compared with colonoscopy in the included studies [20 ]. However, no concurrent increase in preference for CCE was shown. One of the main
shortcomings of CCE compared with colonoscopy is the lack of interventional options,
as patients with neoplastic findings require a subsequent colonoscopy or sigmoidoscopy
for polyp removal or biopsy. This knowledge will likely tilt the preference toward
colonoscopy. Another drawback of CCE is the need for more extensive bowel preparation,
as peri-procedural cleansing of the bowel mucosa with flushing and suction is only
possible during colonoscopy. The bowel preparation in the present study commenced
72 hours before the CCE procedure, whereas patients allocated to colonoscopy required
only 1 day of preparation before colonoscopy. Despite extensive bowel preparation,
CCE failed to meet the completion rate standard for colonoscopy of >90% [35 ]. However, no statistically significant difference was detected in completion rates
for CCE and colonoscopy in the current study (84% vs. 92%, P = 0.13).
For decades, it has been debated whether colonoscopy is mandatory after an episode
of colonic diverticulitis to rule out an underlying malignancy. In a recent systematic
review of endoscopic findings after CT-confirmed colonic diverticulitis, CRC was detected
in 2.0% of cases, whereas advanced adenomas were detected in 3.8% [36 ]. Interestingly, 15% of detected cancers were incidental findings located at sites
other than the site of diverticulitis on CT. Furthermore, the OR for CRC in patients
with complicated diverticulitis compared with uncomplicated diverticulitis was 9.2
[36 ].
The prevalence of CRC in patients with colonic diverticulitis is 3–6 times higher
than the global population prevalence of 0.3%–0.4%. Besides early diagnosis of CRC,
a follow-up colonoscopy offers detection of advanced adenomas and contributes to addressing
other disease states such as inflammatory bowel disease, lymphoma, and ischemic colitis.
The risk of complications during colonoscopy, although low (<0.5%), needs to be considered
to ensure the benefits outweigh the risks. Moreover, in many countries, endoscopy
units are overburdened with long waiting lists for colonoscopy.
In that context, CCE is a pertinent modality to rule out malignancy after an episode
of colonic diverticulitis. It has a CRC detection rate equivalent to colonoscopy [37 ]
[38 ], no risk of bleeding or perforation, and can be performed out of hospital in the
patient’s home. Larger studies powered to investigate CRC detection rates for CCE
in patients after an episode of diverticulitis are needed. Of course, patients with
significant lesions and incomplete CCE will need a subsequent colonoscopy. In our
study, 34.4% (21/61) of patients were in this category, which needs to be improved
by identifying patients at high risk of poor bowel preparation. Moreover, the need
for a subsequent colonoscopy can be decreased by a “diagnose-and-leave” approach,
meaning that diminutive and small polyps <10 mm detected by CCE are ignored, and the
patient is encouraged to participate in existing bowel cancer screening programs.
In conclusion, no difference in experienced physical and mental discomfort was observed
between patients randomized to CCE and those undergoing colonoscopy in patients completing
questionnaires. CCE was a safe alternative to colonoscopy after an episode of colonic
diverticulitis to rule out an underlying malignancy. It was preferred by patients,
except for those who needed a subsequent colonoscopy after CCE due to incomplete examination
or pathological findings. Bowel preparation for CCE needs to be improved to meet the
standards of colonoscopy.
