Article Text

Download PDFPDF

Adenoma detection with cap-assisted colonoscopy versus regular colonoscopy: a randomised controlled trial
  1. Thomas R de Wijkerslooth1,
  2. Esther M Stoop2,
  3. Patrick M Bossuyt3,
  4. Elisabeth M H Mathus-Vliegen1,
  5. Jan Dees2,
  6. Kristien M A J Tytgat1,
  7. Monique E van Leerdam2,
  8. Paul Fockens1,
  9. Ernst J Kuipers2,
  10. Evelien Dekker1
  1. 1Department of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, The Netherlands
  2. 2Department of Gastroenterology and Hepatology, Erasmus University Medical Center, Rotterdam, The Netherlands
  3. 3Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, Amsterdam, The Netherlands
  1. Correspondence to Dr Evelien Dekker, Gastroenterologist, Department of Gastroenterology & Hepatology C2-115, Academic Medical Center, PO Box 22700, 1100 DE Amsterdam, The Netherlands; e.dekker{at}amc.uva.nl

Footnotes

  • Funding The study was funded by The Netherlands Organisation for Health Research and Development (ZonMW 120720012) and by the Center for Translational Molecular Medicine (CTMM DeCoDe-project).

  • Competing interests None.

  • Ethics approval Dutch Health Council (2009/03WBO, The Hague, The Netherlands).

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

View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • Conventional colonoscopy can miss a substantial number of adenomas.

  • Cap-assisted colonoscopy (CAC) may improve colonic visualisation and thus may improve adenoma detection.

  • Currently, the possible improvement in adenoma detection by CAC is arguable, as mixed results on polyp detection have been reported.

What are the new findings?

  • CAC does not improve the detection of adenomas, nor the detection of small size, flat or proximally located adenomas.

  • CAC does not improve the detection of adenomas in patients with good bowel preparation.

  • CAC does reduce the degree of discomfort during colonoscopy.

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

  • On the basis of our results, CAC should not be used in daily practice to improve the detection of adenomas.

  • CAC can be used to reduce caecal intubation times and patient discomfort.

Introduction

Colonoscopy is widely accepted as the reference standard for detection of colorectal neoplasia. However, a substantial adenoma miss rate of 20–26% has been reported in tandem colonoscopy studies.1 Forward-viewing colonoscopes cannot visualise the full colonic surface, and adenomas may be missed because they are located outside the visual field, hidden behind folds or flexures.2

The use of a transparent cap attached to the tip of a colonoscope may increase colonic surface visualisation by depressing the colonic folds with the cap. In addition, a better endoscopic view can be created by keeping an appropriate distance between the tip of the colonoscope and the mucosa, preventing a ‘red-out’. This suggests that cap-assisted colonoscopy (CAC) may improve adenoma detection. A disadvantage of CAC, however, might be that the view is blurred if the bowel preparation is poor, as faecal material can remain in the cap.

So far, clear evidence that CAC improves adenoma detection is lacking. Previous studies did not report histopathology of all polyps and/or did not achieve enough power to compare adenoma detection, and the results may have been influenced by investigator bias or by other confounders.3–11 CAC has particularly been studied in Asian populations. Mixed results on polyp and adenoma detection have been reported.3–10 CAC trials in Western populations are limited; one small single-centre tandem study with only two participating endoscopists showed a reduction in adenoma miss rates by CAC.11 However, a recent meta-analysis could not draw any conclusions on the improvement in polyp or adenoma detection by CAC.12 Regarding caecal intubation, CAC studies have shown a shorter caecal intubation time and suggested easier caecal intubation by inexperienced endoscopists.3 ,4 In addition, patient discomfort seems to be less during CAC.5

It is currently argued that an improvement in adenoma detection could possibly be achieved with CAC. We aimed to compare adenoma detection by CAC and conventional colonoscopy (CC) in a large two-centre randomised controlled trial, comprising screening-naïve participants in a primary colonoscopy screening programme. In addition, we compared caecal intubation time and rate, the degree of discomfort during colonoscopy, perceived burden of colonoscopy 2 weeks afterwards, and complication rate. We prospectively recorded all colonoscopy quality indicators that could have affected adenoma detection. Several endoscopists participated in this study mimicking daily clinical practice of the effectiveness of CAC.

Methods

Study population

Data were collected in the randomised, multicentre Colonoscopy or Colonography for Screening (COCOS) trial. The overall design of this invitational population-based colorectal cancer screening programme has been described in detail previously.13 Between June 2009 and July 2010, 6600 asymptomatic people from the Amsterdam and Rotterdam regions were randomly selected and invited for colonoscopy screening.

Subjects who had undergone a full colonic examination in the previous 5 years (complete colonoscopy, CT colonography and/or double-contrast barium enema) were excluded from the screening programme, as well as subjects scheduled for surveillance colonoscopy (personal history of colorectal cancer, colonic adenomas or inflammatory bowel disease) and subjects with end-stage disease. In addition, subjects with a (partial) colonic resection were excluded.

All screening participants scheduled for colonoscopy were invited to this randomised, parallel designed, study. After providing informed consent, eligible participants were randomly allocated 1:1 to either CAC or CC by a computerised randomisation program (ALEA Randomisation Service).14 Randomisation was stratified by age, sex and screening centre using random block sizes of a maximum of six per block. It occurred within 24 h of colonoscopy, and was performed by the research staff. Participants and endoscopists were blinded to the randomisation result until the start of the colonoscopy. Ethics approval was obtained from the Dutch Health Council (2009/03WBO, The Hague, The Netherlands). The trial was registered in the Dutch Trial Register: NTR1888 (http://www.trialregister.nl).

Colonoscopy

Colonoscopies were performed at the Academic Medical Center Amsterdam and Erasmus Medical Center Rotterdam. Scheduled colonoscopies (CC or CAC) were consecutively performed in a morning or afternoon session according to the standard quality indicators defined by the Society of Gastrointestinal Endoscopy.15 All colonoscopies were recorded on DVD. Colonoscopy variables were directly noted on a case record form by the research staff. All colonoscopies were performed by endoscopists with experience of more than 1000 colonoscopies. They were trained in CAC and had experience of at least 20 cap colonoscopies. Colonoscopes were CF-Q160 (140° field of view), CF-Q180 (170° field of view) and PCF-Q180 (140° field of view) series variable-stiffness instruments (Olympus Medical Systems, Tokyo, Japan).

All participants received standard bowel preparation, which included a low-fibre diet and oral intake of 2 litres of transparent fluid and 2 litres of hypertonic poly(ethylene glycol) solution (Moviprep; Norgine, Amsterdam, The Netherlands) at home. Procedures were performed with the subject under conscious sedation in combination with an analgesic if desired using intravenous midazolam and fentanyl.

Endoscopists intended to intubate the caecum as quickly as possible without performing polypectomies. Caecal intubation was confirmed by documentation of caecal landmarks (caecal valve and appendix orifice or intubation of terminal ileum). During withdrawal of the colonoscope, the colonic mucosa was carefully inspected, and all detected polyps were directly removed and obtained for histological assessment. The minimal withdrawal time (minus time for polypectomy) was at least 6 min. The size of all polyps was measured by the endoscopist using open biopsy forceps with a 7 mm span. Localisation was considered proximal if proximal to the splenic flexure.

Discomfort during colonoscopy was scored by the research staff on the five-point Gloucester Comfort Score, with scores ranging from no discomfort to severe discomfort.16 Bowel preparation was scored using the validated Ottawa bowel preparation score,17 ranging from 0 (excellent bowel preparation in all three colonic segments) to 14 (very poor bowel preparation). Good bowel preparation was defined as a total score of 7 or lower, including segment scores of 2 or lower. In the case of insufficient bowel preparation (Ottawa score ≥11), the procedure was interrupted and rescheduled with the same endoscopist using the same allocated strategy, unless the participant refused to undergo repeat colonoscopy.

Cap colonoscopy

For the CAC group, a transparent cap was fitted to the tip of the colonoscope so that it protruded 4 mm ahead of the tip of the colonoscope. We used a cap with a diameter of 13.4 mm (D-201-12704; Olympus Medical Systems) or 15.0 mm (D-201-14304; Olympus Medical Systems) depending on the diameter of the colonoscope chosen in each procedure. Some improvements were made over the cap used in previous CAC studies. A side hole on the cap was created for drainage of fluid and faecal material. In addition, the edge of the cap was rounded off to minimise mucosal damage, and the material was made more transparent (figure 1).

Figure 1

Cap used for the cap-assisted colonoscopy arm. A side hole on the cap was created for drainage of fluid and faecal material, the edge of the cap was rounded off to minimise mucosal damage, and the material was made more transparent.

Histopathology

Histopathology was processed and stained using standard methods and evaluated by two expert pathologists (one in each centre) according to the Vienna criteria.18 All lesions were classified as hyperplastic, serrated, tubular, tubulovillous, villous or carcinoma. Dysplasia was defined as either low grade or high grade. An advanced adenoma was defined as an adenoma ≥10 mm, ≥25% villous or with high-grade dysplasia.

Complications

All acute complications were recorded at the time of the colonoscopy. Subjects were contacted 2 weeks after the procedure for registration of post-procedural complications. They were instructed to contact research staff if complications occurred in the following 2 weeks to ensure a complete complication record of 4 weeks.

Questionnaire

All participants were asked to complete a questionnaire on perceived burden of the colonoscopy (PBQ) 2 weeks afterwards. It had been previously validated.13 It measured the perceived burden and pain of colonoscopy-related items and the full screening procedure (eg, ‘how burdensome/painful did you find insertion of the endoscope?’). All items were scored on a five-point Likert scale (1=not at all; 2=slightly; 3=somewhat; 4=rather; 5=extremely).

Outcome measures and statistical analysis

The primary outcome measure was adenoma detection, defined as the proportion of participants with at least one adenoma (per-patient analysis). The number of adenomas per subject (per-polyp analysis) was defined as the total number of detected adenomas in each group divided by the total number of participants. Secondary outcomes were caecal intubation time and rate, the degree of discomfort during colonoscopy, perceived burden of the colonoscopy 2 weeks afterwards, and complication rate. We performed a subanalysis to investigate the influence of bowel preparation. We calculated adenoma detection rates in patients with good bowel preparation scores.

Adenoma detection was analysed in an intention-to-treat and a per-protocol analysis. Adenoma detection was compared using the χ2 test (per-patient analysis) and Mann–Whitney U test (per-polyp analysis) statistics. The Mann–Whitney U test statistic was used to compare procedural times and discomfort and perceived burden scores. The χ2 test statistic was used to compare caecal intubation rate.

Two-sided p values of <0.05 were considered to indicate significant differences. All analyses were performed using SPSS V.16.0 for Windows. The results were reported using the CONSORT guidelines.19

Sample size

In the CC group, we expected that 20% of all subjects would have at least one adenoma, based on a large colonoscopy screening study.20 We aimed to detect an increase in adenoma detection by 35%, resulting in an expected adenoma detection rate of 27% in the CAC group. A priori, we planned to scope a total number of at least 1250 colonoscopies (625 per arm). With a two-sided test significance level of 0.05, we would achieve a power of at least 81% in detecting the indicated difference.

Results

Figure 2 shows the patient flow. A total of 1380 eligible screening participants consented and were 1:1 randomised to either CC (n=694) or CAC (n=686). After randomisation, 41 subjects dropped out because of withdrawal (n=26), absence of a trained endoscopist on the day of colonoscopy (n=13) or technical problems (n=2). As displayed in table 1, groups were comparable with respect to age, gender and previous abdominal operation.

Table 1

Baseline demographics and colonoscopy characteristics

Colonoscopy results

Almost all colonoscopies (1328 of 1339; 99%) were performed by five endoscopists, who each performed at least 50 study colonoscopies (table 1). Each endoscopist performed a similar number of cap colonoscopies and regular colonoscopies within this study. Colonoscopes with a 140° and 170° field of view were used equally between the endoscopists. Caecal intubation was achieved in 671 of 683 subjects (98%) in the CC group versus 649 of 656 (99%) in the CAC group (p=0.29). Caecal intubation time was significantly lower in the CAC group (7.7±5.0 min) than the CC group (8.9±6.2 min) (p<0.001). No significant differences were detected with respect to net withdrawal time or bowel preparation scores.

Polyp detection

In the intention-to-treat analysis, the proportion of participants with at least one adenoma was the same in the two groups (28% vs 28%; RR 0.98; 95% CI 0.82 to 1.16). The total number of detected adenomas per subject was not significantly different between CC and CAC (0.49±1.05 vs 0.50±1.03; p=0.91). In the CC group, 63 participants (9%) had at least one advanced adenoma versus 51 participants (7%) in the CAC group (RR 0.80; 95% CI 0.57 to 1.17). The total number of detected advanced adenomas per subject was also comparable between the groups (0.12±0.45 vs 0.09±0.36; p=0.27). The per-protocol analysis is displayed in table 2; it showed comparable results.

Table 2

Polyp detection with conventional colonoscopy (CC) versus cap-assisted colonoscopy (CAC): per-protocol analysis

Table 2 also shows size, morphology and location of all detected adenomas. Detection of small size (<6 mm) adenomas was comparable for CC and CAC, as well as the detection of 6–9 mm and large (>10 mm) adenomas. In addition, CAC did not detect a higher number of flat adenomas per subject or a higher number of subjects with flat adenomas. No significant differences between the groups were noted in the detection of proximal located adenomas.

Influence of endoscopist and bowel preparation on adenoma detection

Adenoma detection rates of all endoscopists are displayed in table 3. One endoscopist who performed 54 colonoscopies in this study detected a lower number of subjects with at least one adenoma in the CAC group (55% vs 24%; p=0.02). Adenoma detection rates for all other endoscopists were not significantly different between CC and CAC.

Table 3

Adenoma detection rates of the different endoscopists

We performed a subanalysis in patients with good bowel preparation. In the CC group, 465 (68%) had at least good bowel preparation versus 434 (66%) in the CAC group. The proportion of subjects with at least one adenoma was 30% in the CC group with good bowel preparation versus 31% in the CAC group (p=0.92). The number of detected adenomas per subject was also comparable between the groups (0.55±1.15 vs 0.56±1.06; p=0.82).

Discomfort during colonoscopy and perceived burden 2 weeks afterwards

In both groups, the majority of subjects received a combination of midazolam and fentanyl (table 1). During colonoscopy, 21% in the CC group had ‘more than two episodes of discomfort’ versus 16% in the CAC group (figure 3). Overall, Gloucester Comfort Scores were lower in the CAC group than in the CC group (mean score 2.0±1.0 vs 2.2±1.0; p=0.03).

Figure 3

Discomfort during colonoscopy. This was measured by the Gloucester Comfort Score ranging from no discomfort (1) to extreme discomfort (5). On top of the bars, mean (SD) scores and significance levels between the groups are displayed.

Two weeks after colonoscopy, a total of 467 of 683 (68%) CC subjects returned the PBQ versus 483 of 656 (74%) subjects in the CAC group. The perceived burden and pain for colonoscopy-related items (introduction of the colonoscope and during the procedure including caecal intubation and withdrawal) 2 weeks after colonoscopy were scored comparably between the groups (figure 4). The full procedure was perceived as not or slightly burdensome by 82% of CC participants and by 83% of CAC participants (mean score 1.8±1.0 vs 1.8±0.9; p=0.75) and as not or slightly painful by 78% and 84% (mean score 1.8±1.1 vs 1.7±1.0; p=0.27), respectively. Two weeks after the colonoscopy, women had lower pain scores in the CAC than the CC group during the procedure (mean score 1.9±1.3 vs 1.7±1.1; p=0.04). Patients who did not receive sedation had comparable burden scores in the two groups.

Figure 4

Perceived burden 2 weeks after the procedure. Perceived pain and burden of different items and the full procedure was measured by a validated questionnaire. On top of the bars, mean (SD) scores and significance levels between the groups are displayed.

Complications

One post-polypectomy bleeding and one perforation occurred in the CC group versus none in the CAC group. One patient in the CC group died because of a spinal epidural abscess 23 days after the colonoscopy. In retrospect, this event was probably not related to the colonoscopy. Three non-colonoscopy-related complications occurred in the CAC group: pneumonia, urinary tract infection and atrial fibrillation.

Discussion

We compared adenoma detection rates by CC and CAC in a population at average risk of colorectal cancer. We found that CAC did not improve the detection of adenomas nor the detection of small size, flat or proximally located adenomas. We also performed a subanalysis in patients with good bowel preparation scores showing similar results. CAC reduced caecal intubation times by more than 1 min. In addition, CAC participants showed lower discomfort scores during colonoscopy.

This is the first large prospective randomised controlled trial adequately powered to compare adenoma detection between CC and CAC. Five experienced endoscopists with good adenoma detection rates and who were trained for CAC performed 99% of all colonoscopies. Our study population was uniform, as all included participants underwent a primary screening colonoscopy. In addition, research staff attended all colonoscopies and prospectively recorded all data on polyp detection, procedural times and bowel preparation scores, ensuring accurate and optimal data collection. CC and CAC were consecutively performed in a random order. Therefore we believe that our results are reliable and applicable to daily clinical practice.

Unfortunately, we had to exclude 39 participants after randomisation because of withdrawal from the study or absence of a CAC-trained endoscopist on the day of the procedure. Because of logistics, we had to randomise some participants 1 day before colonoscopy. Allocation to one of the two arms did not seem to be responsible for dropping out, since the result of randomisation was only revealed to the patients and endoscopists just before the colonoscopy. More importantly, no significant differences were observed between the intention-to-treat and the per-protocol analysis. In our study, colonoscopes with different fields of view (140° and 170°) were used. This does not seem to have influenced the results, as both types were used equally between the endoscopists. In line with this, use of wide-angle colonoscopes did not affect adenoma detection rates in previous studies.21 ,22 Lastly, as blinding is not possible because the cap is visible on the monitor during colonoscopy, it is impossible to rule out investigator bias in any study with the cap. Investigator bias would have been more likely if we had detected a higher number of subjects with adenomas in the CAC group, as we aimed to improve adenoma detection by CAC.

According to quality guidelines, adenoma detection rates over 20% are required in populations at average risk of colorectal cancer.15 ,23 Low adenoma detection rates are associated with an increased risk of interval colorectal cancer.23 In our study, the endoscopists fulfilled this quality condition. Adenoma detection rates of the CC group (control group) varied from 23% to 41% for those endoscopists who performed more than 100 study colonoscopies. Good adenoma detection rates in our control group did minimise the risk of investigator bias and did secure a solid comparison with CAC.

In our study, adenoma detection rates in the CAC group were comparable to those in the CC group and varied from 24% to 35%. Only one endoscopist, who performed 54 colonoscopies in this study, detected significantly more patients with at least one adenoma in the CC group than in the CAC group (55% vs 24%; p=0.02), but this difference did not affect the overall results. In the literature, mixed results have been reported on improvement of polyp detection by CAC (table 4).3–11 The majority of these studies did not report histopathology of detected polyps. One Japanese study with a similar design reported a higher polyp detection rate in the CAC group.4 In contrast, a Chinese study reported lower polyp detection rates with CAC, but polyp detection also correlated with withdrawal time.3 Two other parallel randomised controlled trials reported similar polyp detection rates, but these studies did not achieve enough power to compare polyp detection.5 ,8 A limited number of CAC studies did report histopathology. Two tandem studies showed improvement in adenoma detection by CAC.9 ,10 However, adenoma miss rates in the control groups were lower than expected based on miss rates in a meta-analysis, suggesting investigator bias.1 Hewett and Rex studied CAC in a Western population and found that CAC decreased adenoma miss rates, especially for small size adenomas.11 In this study, adenoma detection rates of the participating endoscopists were remarkably high (69% for CC vs 65% for CAC), making these results less applicable to daily clinical practice. Although a tandem design, as used in this study, is generally considered the most reliable, it can lead to investigator bias in studies in which blinding for the technique is impossible. In our study, a large number of CC and CAC procedures were consecutively performed in a random order, through which we aimed to mimic daily clinical practice.

Table 4

Polyp and adenoma detection rates (per-patient and per-polyp analysis) in randomised controlled trials (conventional colonoscopy (CC) vs cap-assisted colonoscopy (CAC))

Caecal intubation times of the CAC group were reduced by more than 1 min in our study, which is in accordance with findings from other CAC studies.3–5 This reduction may be caused by the protruding cap, which may facilitate sliding along folds and flexures allowing quick advancement of the colonocope to the caecum. Furthermore, CAC may be especially helpful in patients with difficult bowel anatomy, such as female patients, old patients, patients with previous abdominal surgery, and patients with left-sided diverticulosis.4 However, the caecal intubation rates were not improved by CAC and were equal to those in other studies in the literature.3 ,5

We showed that participants undergoing CAC had lower discomfort scores during colonoscopy. This finding is in accordance with the literature.5 ,24 ,25 However, after 2 weeks, no significant differences in the perceived burden of the procedure were reported. Discomfort during colonoscopy was scored by the research staff, whereas the burden after 2 weeks was reported by subjects themselves. Investigator bias may be an explanation for these conflicting results. A subanalysis in women found lower pain scores 2 weeks after colonoscopy for caecal intubation and withdrawal of the colonoscope, a finding that is in line with the literature.5

We compared the adenoma detection rate between CAC and CC for experienced endoscopists. It has previously been reported that CAC improved caecal intubation rates in female patients among trainee endoscopists.4 CAC may be useful for improving adenoma detection by less experienced endoscopists, but further studies are needed to confirm this. A possible disadvantage of CAC is the visibility of the cap on the monitor during colonoscopy reducing the visual field. A possibility for improving this could be the development of a cap with an angle that is similar to the field of view of the colonoscope (140° or 170°). In this case, the cap would smooth colonic folds without blurring the endoscopic view. In addition, because of the oblique sides, the chance of faecal residue remaining in the cap may be lower, and manoeuvering the cap to each fold may take less effort because of the extended range of the cap. Another option is to combine CAC with other advanced imaging techniques. A Japanese study combined CAC and autofluorescence imaging and found higher ‘neoplasm detection rates’ (adenomas, carcinomas and carcinoids altogether) compared with white light endoscopy only (1.96 vs 1.19; p=0.02).7 The conclusion that a combination of these techniques improves adenoma detection seems premature. Further studies are needed to verify these results. A recent study showed that CAC improved polyp detection in patients referred for endoscopic mucosal resection of polyps detected during an initial CC that could not be removed with standard biopsy forceps.26 CAC may possibly have a role during a ‘second look’ in patients being referred for removal of (large) colorectal polyps.

We conclude from this large randomised controlled trial that CAC does not improve adenoma detection. It does reduce caecal intubation times and is safe, as no complications occurred. On the basis of the results of our study, we strongly feel that CAC should not be used in daily clinical practice to improve the detection of adenomas. It may be useful in reducing caecal intubation time and patient discomfort. This technique could therefore be used for these indications.

Acknowledgments

We would like to acknowledge The Netherlands Organisation for Health Research and Development (ZonMW) and the Center for Translational Molecular Medicine (CTMM) for their financial support. Furthermore, we acknowledge Olympus Medical Systems (Tokyo, Japan) for providing the caps. In addition, we acknowledge Karin de Groot for professional research support and accurate recording of all colonoscopy items.

References

View Abstract

Footnotes

  • Funding The study was funded by The Netherlands Organisation for Health Research and Development (ZonMW 120720012) and by the Center for Translational Molecular Medicine (CTMM DeCoDe-project).

  • Competing interests None.

  • Ethics approval Dutch Health Council (2009/03WBO, The Hague, The Netherlands).

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

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.

Linked Articles