Article Text

Download PDFPDF

Original article
Adenoma detection with Endocuff colonoscopy versus conventional colonoscopy: a multicentre randomised controlled trial
  1. SC van Doorn1,
  2. M van der Vlugt1,
  3. ACTM Depla2,
  4. CA Wientjes3,
  5. RC Mallant-Hent4,
  6. PD Siersema5,
  7. KMAJ Tytgat,
  8. H Tuynman1,2,
  9. SD Kuiken3,
  10. GMP Houben2,
  11. PCF Stokkers3,
  12. LMG Moons5,
  13. PMM Bossuyt6,
  14. P Fockens1,
  15. MW Mundt4,
  16. E Dekker1
  1. 1Departments of Gastroenterology & Hepatology, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands
  2. 2Departments of Gastroenterology & Hepatology, Slotervaartziekenhuis, Amsterdam, The Netherlands
  3. 3Departments of Gastroenterology & Hepatology, Sint Lucas Andreas Ziekenhuis, Amsterdam, The Netherlands
  4. 4Departments of Gastroenterology & Hepatology, Flevoziekenhuis, Almere, The Netherlands
  5. 5Departments of Gastroenterology & Hepatology, University Medical Centre Utrecht, Utrecht, The Netherlands
  6. 6Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, Amsterdam, The Netherlands
  1. Correspondence to Dr Evelien Dekker, Departments of Gastroenterology and Hepatology, Academic Medical Center, Meibergdreef 9, Amsterdam1105 AZ, The Netherlands; e.dekker{at}


Background and aims Colonoscopy is the current reference standard for the detection of colorectal neoplasia, but nevertheless adenomas remain undetected. The Endocuff, an endoscopic cap with plastic projections, may improve colonic visualisation and adenoma detection. The aim of this study was to compare the mean number of adenomas per patient (MAP) and the adenoma detection rate (ADR) between Endocuff-assisted colonoscopy (EAC) and conventional colonoscopy (CC).

Methods We performed a multicentre, randomised controlled trial in five hospitals and included fecal immonochemical test (FIT)-positive screening participants as well as symptomatic patients (>45 years). Consenting patients were randomised 1:1 to EAC or CC. All colonoscopies were performed by experienced colonoscopists (≥500 colonoscopies) who were trained in EAC. All colonoscopy quality indicators were prospectively recorded.

Findings Of the 1063 included patients (52% male, median age 65 years), 530 were allocated to EAC and 533 to CC. More adenomas were detected with EAC, 722 vs 621, but the gain in MAP was not significant: on average 1.36 per patient in the EAC group versus 1.17 in the CC group (p=0.08). In a per-protocol analysis, the gain was 1.44 vs 1.19 (p=0.02), respectively. In the EAC group, 275 patients (52%) had one or more adenomas detected versus 278 in the CC group (52%; p=0.92). For advanced adenomas these numbers were 109 (21%) vs 117 (22%). The adjusted caecal intubation rate was lower with EAC (94% vs 99%; p<0.001), however when allowing crossover from EAC to CC, they were similar in both groups (98% vs 99%; p value=0.25).

Interpretation Though more adenomas are detected with EAC, the routine use of Endocuff does not translate in a higher number of patients with one or more adenomas detected. Whether increased detection ultimately results in a lower rate of interval carcinomas is not yet known.

Trial registration number Dutch Trial Register: NTR3962.


Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Significance of this study

What is already known on this subject?

  • Conventional colonoscopy can miss a substantial number of adenomas, with adenoma miss rates from 10% to 24%.

  • A new endoscopic cap with plastic projections, the Endocuff, was developed to improve adenoma detection by straightening the colonic folds during withdrawal and thereby potentially revealing more adenomas located behind those folds.

  • Previous studies described increased adenoma detection rates with Endocuff-assisted colonoscopy (EAC), but the effect on the mean number of adenomas per patient (MAP) is still unknown.

What are the new findings?

  • EAC increases the number of adenomas detected compared with conventional colonoscopy, specifically the number of diminutive and flat adenomas.

  • Increased detection with EAC does not result in a higher number of patients with adenomas or advanced adenomas detected.

How might it impact on clinical practice in the foreseeable future?

  • EAC increases the yield of colonoscopy, especially the detection of diminutive and flat adenomas.

  • Routine use of EAC will not increase the number of patients in whom one or more adenomas are detected.

  • Future studies are needed to explore the association between MAP and the occurrence of interval carcinomas.


Colonoscopy is the current reference standard for the detection of colorectal cancer (CRC) and its precursor lesions, adenomas.1 Several tandem studies, however, have reported a substantial adenoma miss rate of 20%–26%.2 The adenoma detection rate (ADR) is considered the prime quality indicator in CRC prevention because of the association with the risk of interval carcinomas.3–5 However, ADR also has an inherent limitation as it does not measure the total number of adenomas detected within a person, potentially resulting in the ‘one and done phenomenon’.6 In order to assign an appropriate surveillance interval after baseline colonoscopy, identification of all polyps during this colonoscopy is important.7 Although the risk of each diminutive adenoma to develop into invasive cancer is very small, the presence of three or more tubular adenomas at baseline colonoscopy predicts an increased risk of future advanced adenomas and CRC,8–12 ,9 according to the European postpolypectomy surveillance guidelines of the European Society of Gastrointestinal Endoscopy (ESGE).13 Therefore, measurement of the mean number of adenomas detected per colonoscopy (MAP) in addition to the ADR is suggested in the literature, providing additional information about colonoscopy performance.5

Adenomas might be overlooked when located behind folds or in flexures, and thus outside the regular visual field during colonoscopy.14 In order to increase colonic surface visualisation, the use of a transparent plastic cap placed at the tip of the colonoscope was introduced to flatten colonic folds. Several large randomised controlled studies and meta-analyses have compared cap-assisted colonoscopy with conventional colonoscopy (CC) and showed conflicting results and therefore, due to the lack of convincing results, its regular use is not advocated.7 ,15–20

Recently a new endoscopic cap, the Endocuff, was developed to improve adenoma detection and tip control of the colonoscope. The Endocuff (Arc Medical Design, Leeds, England) is a 2 cm long, flexible cap with two circular rows of plastic, hinged projections (‘hairs’) that during withdrawal might straighten mucosal folds. Two recent German studies, performed by one study group in an academic setting, demonstrated improved adenoma detection with Endocuff-assisted colonoscopy (EAC) when compared with CC.21 ,22 Larger studies are needed to confirm a positive effect on the MAP detected during EAC in addition to the ADR and to evaluate the effect in a non-academic setting, more resembling daily practice.

We performed a multicentre, randomised controlled trial comparing the MAP detected during EAC and CC.



This study was performed between August 2013 and October 2014 in five hospitals in the Netherlands: two academic hospitals (Procolo/Bergman Inwendige Zorg Amsterdam (IZA) at the Academic Medical Centre in Amsterdam and University Medical Centre in Utrecht) and three regional hospitals (Slotervaartziekenhuis in Amsterdam, Sint Lucas Andreas Ziekenhuis in Amsterdam and Flevoziekenhuis Almere) in the Netherlands. Patients aged ≥45 years, scheduled for colonoscopy for one of the following indications for colonoscopy, were invited to participate in this study: polyp surveillance, changed bowel habits and/or bloody stools, bowel complaints, a positive family history for CRC or a positive FIT. Before inclusion, patient charts were reviewed for possible exclusion criteria and these included patients with polyposis syndromes, IBD, severe diverticulosis or colonic stricture, previous (partial) colonic resection, previous CT colonography within last 2 years or an acute indication for colonoscopy. Patients were informed about the study during consultation prior to colonoscopy and received an information brochure to read at home. After written informed consent, eligible patients were randomly allocated 1:1 to either EAC or CC by a computerised randomisation programme (ALEA Randomisation Service).23 Randomisation was stratified by centre using random block sizes of block size 2, 4, 6 and 8. Randomisation was performed within 24 h of the colonoscopy and performed by the research staff. This study was approved by the medical ethical committees of all participating centres and the study underwent frequent monitoring by an independent monitor. All the authors had access to the study data and reviewed and approved the final manuscript. The trial was registered in the Dutch Trial Register: NTR3962 (


Colonoscopy variables were directly noted on a case record form by the endoscopist or research staff. All colonoscopies were performed by endoscopists with an experience of at least 500 colonoscopies. They were all trained in EAC (training included an extensive explanation and demonstration by the research coordinator) and had performed at least 15 EACs prior to the start of the study. Colonoscopes used within this study were CF-H180AL, CF-HQ190L and PCF-H180AL series variable-stiffness instruments by Olympus (Olympus Medical Systems, Tokyo, Japan) and Fujinon EC 530 WL instruments (Fujifilm Europe, Düsseldorf, Germany).

Most patients received standard bowel preparation: low-fibre diet and oral intake of 2 L of transparent fluid and 2 L of hypertonic polyethylene glycol solution (Moviprep; Norgine, Amsterdam, The Netherlands) at home. The majority of the procedures were performed under conscious sedation using intravenous midazolam and fentanyl/alfentanil. Administration of antispasmodics (butylscopolamine) was allowed at the discretion of the endoscopist. All administered medications were recorded. Endoscopists intended to intubate the caecum as quickly as possible without performing polypectomies during introduction. Caecal intubation was confirmed by documentation of caecal landmarks (ileocaecal valve, appendiceal orifice or terminal ileum). During withdrawal of the colonoscope, the mucosa was carefully inspected and in case of polyps endoscopic removal of the lesions was attempted. All polyps were retrieved in separate specimen containers and sent for histopathological assessment. All polyp features (colonic segment, morphology: divided in pedunculated, sessile, flat or depressed lesions, estimated size, optical diagnosis and polypectomy technique) were reported by the research staff. Location was considered proximal if proximal to the splenic flexure. Intubation and withdrawal time (including time taken for polypectomies) were recorded by a member of the research staff or by an endoscopy nurse, using a stopwatch.

Patient discomfort was scored by one of the nurses present at colonoscopy on the five-point Gloucester Comfort Score, with scores ranging from no discomfort (1) to severe discomfort (5).24 Bowel preparation was scored per segment using the validated Boston Bowel Preparation Score (BBPS),25 ranging from 3 (complete segment without faeces, good visualisation) to 0 (solid faeces, not evaluable).

Endocuff-assisted colonoscopy

The first generation Endocuff received a CE mark in July 2011, and was approved for human medical use by the food and drug administration (FDA) in 2012. For the EAC, the Endocuff was placed onto the tip of the colonoscope distal to the bending section just before insertion. In all other aspects, the procedural protocol was similar to the CC. It is constructed with two rings of soft, flexible slim projections that are hinged at their bases so as not to interfere with forward movement. There are several Endocuff devices available with varying sizes for different colonoscopes. For the Olympus endoscopes we used the AEC 140, for the Fujinon endoscopes the AEC 120 and for all procedures with a paediatric endoscope the AEC 130 (figure 1). The type and number of the colonoscope and type of Endocuff were recorded on the case record form for each colonoscopy.

Figure 1

The Endocuff AEC 140 and AEC 130 models. The Endocuff (Arc Medical Design, Leeds, England) is a 2 cm long, flexible cap with two circular rows of plastic, hinged projections.


Histopathology was processed and stained using standard methods and evaluated by one of the GI pathologists in each centre. The pathologists were blinded to the study allocation. Polyp histology was evaluated according to the Vienna criteria.26 All lesions were classified as hyperplastic polyp, sessile serrated adenoma/polyp, traditional serrated adenoma, tubular, tubulovillous or villous adenoma or carcinoma. Dysplasia was defined as either low grade or high grade. An advanced adenoma was defined as an adenoma ≥10 mm, ≥25% villous histology or with high-grade dysplasia.


All procedural complications were recorded at the time of colonoscopy. Patients were contacted a week after the procedure for registration of postprocedural complications. They were instructed to contact the hospital if complications occurred to ensure a complete complication record. The trial coordinator was informed about all complications.

Outcome measures and statistical analysis

The primary outcome was adenoma detection, defined as the total number of detected adenomas in each group divided by the total number of patients in that group (MAP, per-polyp analysis). The ADR was also evaluated, defined as the proportion of patients with at least one adenoma (per-patient analysis). Adenoma detection was analysed in an intention-to-treat and a per-protocol analysis. Adenoma detection was compared using a Poisson regression model (per-polyp analysis, for Poisson distributed data). In this model, we adjusted for over dispersion caused by the large variance of the data (generalised linear model; GLM). As we stratified by colonoscopy centre during randomisation, we also compared the MAP using a generalised linear mixed model. In this model, both the randomisation arm and colonoscopy centre were entered as fixed factors and number of adenomas per procedure as dependent factor, to account for the correlation between colonoscopy centre and MAP. ADR was compared between groups using the χ2 test as well as with multivariable mixed model logistic regression, to account for the intracentre correlation.

Secondary outcomes were caecal intubation rate, caecal incubation time, withdrawal time, patient discomfort and complications. We adjusted for poor bowel preparation and colonic strictures to calculate the adjusted caecal intubation rate (as referred to in table 1, footnote). The difference in caecal intubation rate was also evaluated with multivariable mixed model logistic regression to account for the intracentre correlation.

Table 1

Baseline demographics and colonoscopy characteristics

The Mann–Whitney U test statistic was used to compare procedural times, discomfort scores and the BBPS. For the comparison of withdrawal times, only colonoscopies in which no polyps were detected and no biopsies were taken were analysed.

We performed additional per-protocol subanalyses to explore differences in polyp morphology, location and size between groups. In these analyses, we only included patients in whom the colonoscopy could be completed with the allocated procedure. We used generalised linear mixed models for all subanalyses for MAP, except for the evaluation of MAP per endoscopist, as all endoscopists only performed colonoscopies in one of the colonoscopy centres. Differences among endoscopists and centres (academic vs regional) were evaluated in per-protocol analyses.

Two-sided p values of <0.05 were considered to indicate statistically significant differences. All analyses were performed using SPSS V. 21.0 for Windows.

Sample size

The MAP was assumed to be Poisson distributed. We expected to detect 0.7 adenomas per patient with CC. Based on two previous studies in a comparable, symptomatic patient group,27 ,28 we expected the detection rate to increase by at least 20%, to 0.84 adenomas per patient following Endocuff (rate ratio >1.2). With a two-sided test significance of 0.05, we would achieve 80% power in detecting a rate ratio larger than 1.2 when 972 patients are included, 486 per arm. The expected number of adenomas to be detected in the CC group equalled 341 compared with 409 in the Endocuff group. To estimate the overall enrolment goal to ensure 972 evaluable patients for the per-protocol analysis, the number of protocol violations was assessed blinded for study outcomes after inclusion of the first 922 patients. Some patients who were allocated to Endocuff colonoscopy underwent a crossover to the CC group (4.70% of the Endocuff population, because of failure to intubate the caecum). Besides, in both groups incomplete colonoscopies occurred, amounting to 3.48% of the total study population. Based on these proportions, we estimated that the enrolment goal needed to be increased to a total of 1058 patients to ensure 972 completed and assessable patients for the per-protocol analysis (972/(1–0.0348)/(1–0.0470)).


Patient characteristics

Figure 2 shows the study flowchart. Between August 2013 and November 2014, a total of 1065 patients consented. Two patients could not participate because of the absence of a trial endoscopist. Of the 1063 included patients, 530 were allocated to EAC and 533 to CC. As shown in table 1, groups were comparable with respect to age, gender and colonoscopy indication.

Figure 2

CONsolidated Standards of Reporting Trials (CONSORT) study flow chart Endocuff trial. *Three patients were allocated to Endocuff-assisted colonoscopy (EAC), but received conventional colonoscopy (CC). Four patients were allocated to CC, but received EAC. **Colonoscopy was cancelled because patient had a rectal prolapse and was referred to the emergency room. ***In 22 cases, it was not possible to extend past the sigmoid colon because of diverticular disease, and these patients crossed over to CC.

Colonoscopy results

Thirty six patients had an incomplete colonoscopy (see figure 2). In both groups, six colonoscopies were incomplete because of colonic strictures. Five colonoscopies were incomplete because of poor bowel preparation, three in the EAC group and two in the CC group. In 22 EACs (4.2%), the Endocuff was removed (converted to CC) because no intubation beyond the sigmoid was possible. Adjusted caecal intubation rates were therefore significantly lower in the EAC group (94% vs 99%; p<0.001) (table 2). However, after removing the Endocuff, 19 of 22 colonoscopies were completed to the caecum without Endocuff. This resulted in similar adjusted caecal intubation rates of 98% vs 99% for the intention to treat (ITT) analyses (p=0.25). Caecal intubation time for the complete colonoscopies was significantly shorter in the EAC group (7.00 vs 8.30 min; p<0.001), as was the net withdrawal time in negative colonoscopies (median 7.00 vs 8.00 min; p=0.02). No significant differences were detected with respect to bowel preparation scores or Gloucester comfort scores.

Table 2

Colonoscopy results

Adenoma detection rates

Table 3 summarises the results. The MAP was higher in the EAC group: 1.36 (SD (±) 2.10) vs 1.17±1.65. There was an average gain of 0.19 adenoma per patient (95% CI −0.037 to 0.42); this difference was not statistically significant (p=0.08).

Table 3

Polyp detection with Endocuff-assisted colonoscopy versus conventional colonoscopy*

The proportion of patients with at least one adenoma was the same in both groups: 52% in the EAC group versus 52% in the CC group (RR 0.99; 95% CI 0.88 to 1.12). This was also the case when additionally adjusting for intracentre correlation (OR 1.01; 95% CI 0.79 to 1.30).

The proportion of patients with at least one advanced adenoma was comparable between the groups: 21% in the EAC group versus 22% in the CC group (p=0.57), also when adjusting for intracentre correlation (OR 1.10; 95% CI 0.81 to 1.48). The mean number of advanced adenomas per patient was also comparable: 0.28±0.67 vs 0.29±0.62 (p=0.90). Forty nine carcinomas were detected in 47 patients: 24 in the EAC group (4.3%) and 25 in the CC group (4.5%).

Additional analyses

We included 1000 patients in the per-protocol analysis: 486 in the EAC group and 514 in the CC group. The MAP was 1.44±2.16 in the patients undergoing EAC versus 1.19±1.67 in those that had CC, a significant difference (p=0.02), but the ADR was comparable: 54% vs 53% (RR 1.01; 95% CI 0.90 to 1.15). When adjusting for centre, the difference was also not significant (OR 0.96; 95% CI 0.74 to 1.24). There was no significant difference in the proportion of patients with at least one advanced adenoma, or in total number of advanced adenomas between the groups (data not shown).

Size, morphology and proximal location of all adenomas detected in this study are shown in table 4. Only detection of diminutive (<6 mm) adenomas was significantly better for EAC (0.91 vs 0.74; p=0.03). EAC also detected a higher number of flat adenomas per patient (0.27 vs 0.16; p=0.02). There were no significant differences between the groups in the detection of proximal located adenomas.

Table 4

Polyp detection with EAC versus CC: per-protocol analysis*

We also evaluated the number of detected adenomas per colonic segment. There was no difference in adenoma detection for all segments. However, a significantly higher number of patients had at least one adenoma in the descending colon in the EAC group (16% vs 11%; p=0.03), and the MAP detected in the descending colon was also higher in the EAC group (0.23 vs 0.14; p<0.01).

Indication, centre and endoscopist effects

We explored additional factors affecting MAP or ADR. The ADR differed significantly between indication groups (p<0.001). In FIT positive screening participants the ADR with EAC was 73%, when compared with 75% with CC (p=0.52). There was also no difference in ADR or MAP between EAC and CC groups, except for the symptomatic colonoscopies, in which MAP was higher in the EAC group (0.81±1.48 vs 0.60±1.01; p=0.04) (table 5).

We also compared ADR and MAP in the academic versus the non-academic centres. ADR and MAP were significantly higher in the academic centres (68% vs 41%; p<0.001 and 1.92 vs 0.84; p<0.001), also after adjusting for colonoscopy indication. In colonoscopies performed in academic centres the ADR did not differ between EAC and CC (71% vs 66%; p=0.21), but the MAP was higher in the EAC group (2.30±2.63 vs 1.58±1.91; p<0.01). There were no differences between EAC and CC in ADR and MAP for procedures performed in non-academic centres: 0.79±1.42 vs 0.88±1.38 (p=0.48) and 40% vs 43% (p=0.56), respectively.

Nine endoscopists performed most colonoscopies: 851 of 1063 procedures (80%). All contributed at least 50 colonoscopies to the study (table 6). The remaining 212 colonoscopies were performed by 11 endoscopists. Each endoscopist performed a similar number of EACs and CCs within the study (p=0.94, tables 5 and 6). One endoscopist who performed 113 endoscopies had a higher MAP with EAC (2.98 vs1.38; p<0.01), but the ADR was comparable. For all other individual endoscopists, MAP and ADR were not significantly different between EAC and CC.

Table 5

Adenoma detection per different colonoscopy indications (per-protocol analysis)*

Table 6

Adenoma detection of the different endoscopists (per-protocol analysis)*


In both the EAC group and the CC group only minor complications occurred; two postpolypectomy bleedings in each group, a perianal abscess after colonoscopy in the CC group and a deep venous thrombosis with a pulmonary embolism, probably not related to the colonoscopy, in the EAC group.


In this multicentre, randomised controlled trial, we compared adenoma detection with Endocuff-assisted versus CC and observed that using an Endocuff during colonoscopy increased the number of adenomas detected, although the gain was smaller and no longer significant in the total group. More adenomas were detected, but most were classified as non-advanced, and the number of patients with at least one adenoma detected was not significantly different. EAC specifically increased the detection of diminutive adenomas, flat adenomas and adenomas in the descending colon. Adjusted caecal intubation rates were similar in both groups.

This is the first large prospective randomised controlled trial that is adequately powered to compare the MAP between EAC and CC. Colonoscopies were performed in two academic and three regional hospitals, by experienced endoscopists with high ADRs and who were trained in EAC. EAC and CC were consecutively performed in a random order and data on polyp detection, procedural times and bowel preparation scores were prospectively recorded, ensuring accurate and optimal data collection of high quality colonoscopies. We therefore believe that our results are reliable and applicable to daily clinical practice.

Some potential limitations of the study have to be acknowledged. Our sample size calculation was based on the expectation of a MAP of 0.7 per patient with CC. The MAP with CC was 1.19; much higher than we expected. During the inclusion period, the national FIT-based CRC screening programme started in the Netherlands and we included more FIT positive screening participants (approximately 38% of study population) than expected. However, despite the high MAP in the control group, the MAP in the EAC group was 1.44, an increase of 21.4% compared with 1.19.

We did not standardise colonoscopy procedures between the centres, such as colonoscopes, or the use of antispasmodics and sedatives. As this is a large, randomised controlled study, we do not think these factors confounded the comparison. Due to the multicentre, pragmatic character of our study the results are likely to apply to daily practice, with high quality colonoscopies performed in all centres.

In our study, endoscopists could not be blinded for the allocated technique. However, the ADR in the CC group was equally high at 53%, well meeting the established threshold of 20%.3 ,29 The MAP in the CC group was high, making an investigator bias (or a ‘new tool bias’) less likely.

ADR is considered the most important surrogate measure for quality of colonoscopy, but has the inherent limitation that it does not measure the total number of adenomas detected during a procedure and might provoke the ‘one and done’ phenomenon.6 In the literature it has been suggested that besides ADR, the MAP should be reported.5 Besides avoiding the ‘one and done’ phenomenon, identification of every adenoma is necessary to advice a sufficient surveillance interval, and if the surveillance advice less intensive than it should have been if all adenomas would have been detected, this could result in interval carcinomas.6 ,7

In this study, we report high ADRs of 53% in both EAC and CC groups. Several large randomised controlled trials and meta-analyses have compared polyp and adenoma detection with cap-assisted colonoscopy to CC. Although the earliest meta-analyses showed increased polyp detection rates and improved detection of diminutive and small adenomas with cap-assisted colonoscopy,7 none of the meta-analyses detected a convincing increase in ADR.15 ,17 ,18 The most recent randomised controlled trial in over a thousand patients did not detect a significantly higher ADR with cap-assisted colonoscopy.19 ,20 ,30

Two previous studies compared adenoma detection between EAC and CC. They did report an increase in ADR with EAC.21 ,22 The most recent study did not detect an increase in MAP.22 The fact that we did not detect a difference in ADR could well be the result of our high ADR in the CC group, while the two mentioned previous studies had lower baseline ADRs.21 ,22 Because we anticipated a high ADR based on previous studies by our research group, our study was powered to detect an increase in MAP.19 ,27 ,28

In some procedures, the Endocuff had to be removed from the colonoscope because of the inability to proceed with intubation beyond the sigmoid. In these cases, diverticular disease or technical difficulties were reported as reasons for incomplete intubation. After removal of the Endocuff, the caecum could be intubated in 19 of these 22 procedures. For daily practice, this implies that in patients with known diverticulosis or a previous difficult colonoscopy (eg, after abdominal operations) the Endocuff should not be used or might have to be removed.

Use of the Endocuff reduced withdrawal times in negative colonoscopies. While the Endocuff was designed to flatten colonic folds, withdrawal times could be reduced by facilitating inspection of the mucosa. Besides, the Endocuff might provide a more stable position of the endoscope with EAC,31 spending less time reintubating the colon during withdrawal. The Endocuff was originally designed to facilitate caecal intubation by easier straightening of the endoscope and loop prevention. The time to achieve caecal intubation was reduced by >1 min in the EAC group, which was also reported in cap-assisted colonoscopy studies, but not in previous studies comparing EAC with CC.15 ,16 ,19 ,21 ,22 ,32

In line with Biecker’s study,21 we observed an increase in the detection of flat adenomas, which have been suggested to contribute to the development of interval carcinomas, as they are more difficult to detect.33 ,34 EAC could therefore be of clinical relevance in patients with a non-polypoid adenoma.

While previous studies described an increase of polyps detected in the caecum and ascending colon, our study did not confirm this.21 ,22 Endocuff aims to flatten the colonic folds with the projections, and the advantage of EAC might therefore be greater in the left colon, which is less wide than the right colon. We found an increase in MAP and ADR in the descending colon for EAC. Surprisingly, this effect was not seen in the sigmoid, where ADR and MAP were comparable between the groups. The inability of EAC to increase detection of lesions in the sigmoid might be explained by the fact that at this location adenomas are more often pedunculated and more easily detected with CC. Accordingly, in our study significantly more pedunculated polyps were detected in the sigmoid when compared with other colonic segments.

The number of small and large adenomas did not significantly increase with EAC, nor did the number of advanced adenomas. The number of diminutive adenomas did increase significantly with EAC. This is in line with previous studies evaluating Endocuff, which also found a more pronounced increase in the detection of smaller polyps.21 ,22 As EAC only seems to result in the detection of more diminutive adenomas (resulting in an increased MAP), its clinical value might be questioned.

Although no data are available on an association between MAP and interval carcinomas, we anticipate that such a link could be detected in future studies, as a higher MAP expresses a more thorough inspection of the colon. Future large epidemiological studies exploring the association between MAP and interval carcinomas are eagerly awaited.

One should also investigate the interval for surveillance colonoscopy after an index colonoscopy of high quality, including a high MAP. As a result of an increased number of adenomas per patient, surveillance intervals might shorten in accordance to the current guidelines.13 We believe it may be reasonable to prolong surveillance intervals after a high quality colonoscopy. The progression of adenomas into cancer usually takes time; the interval, estimated between 10 and 15 years, provides a long window in which polyp can be detected in a benign stage.35 ,36

Whether the benefits of increased detection of adenomas in patients, without a significant increase in the detection of advanced adenomas, justify an investment in the Endocuff as a standard addition to colonoscopy may vary between centres and settings. The routine or selective use of the Endocuff will be guided by budget considerations and primarily by the relative weight one attaches to improved detection, both as an indicator of colonoscopy quality in itself as well as for the long-term protection for CRC for patients.


We would like to thank Dr M. Dijkgraaf and Dr M. Merkus from the Clinical Research Unit Academic Medical Center (AMC) for their advice and help regarding the sample size calculation. We would like to thank Dr N. van Geloven for her advice regarding the statistical analyses. We would like to thank Karin de Groot for monitoring the study.



  • Correction notice This article has been corrected since it published Online First. The supplementary data information has been removed from the headings of tables 5 and 6.

  • Contributors SCvD, ED and PF developed the idea for the study, designed the study and protocol, were responsible for project organisation, supervision and obtained funding. SCvD, MvdV and ED were involved in further developing the idea, collection of the data and writing the manuscript. ACTMD, CAW, RCM-H, PDS, KMAJT, SDK, HT, GMPH, PCFS, LMGM and MWM were involved in the final draft of the study protocol and contributed by critical revision. ACTMD, CAW, RCM-H, PDS, KMAJT, SDK, HT, GMPH, PCFS, LMGM and MWM contributed to the acquisition of data, and discussed the interpretation of the data with the first author (SCvD). ACTMD, CAW, RCM-H, PDS, KMAJT, SDK, HT, GMPH, PCFS, LMGM and MWM contributed to the manuscript through critical revision and correction of draft versions, and they approved the final manuscript. SCvD was responsible for the statistical analyses of the data and was principally responsible for drafting this manuscript. PMMB revised the manuscript through critical revision from an epidemiology point of view.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Medical ethical committees of all participating centres.

  • Provenance and peer review Not commissioned; externally peer reviewed.