Effect of Endocuff-assisted colonoscopy on adenoma detection rate: meta-analysis of randomized controlled trials

• Abstract
• Introduction
• Methods
• Literature research
• Inclusion and exclusion criteria
• Inclusion criteria
• Exclusion criteria
• Data extraction
• Quality assessment
• Study end points
• Statistical analyses
• Results
• Study selection
• Quality assessment
• Primary end point: ADR
• Secondary end points
• Polyp detection rate
• Technical outcomes
• Adverse events
• Discussion
• Conclusions
• References

Abstract

Background Yield of Endocuff-assisted colonoscopy (EAC) compared with standard colonoscopy is conflicting in terms of adenoma detection rate (ADR). A meta-analysis of randomized controlled trials (RCTs) appears necessary.

Methods PubMed and Google Scholar were searched in December 2017. Abstracts from Digestive Disease Week and United European Gastroenterology Week meetings were also searched to 2017. All RCTs comparing EAC with standard colonoscopy were included. Analysis was conducted by using the Mantel–Haenszel models. Heterogeneity was quantified using the I ² test.

Results Of the 265 articles reviewed, 12 RCTs were included, with a total of 8376 patients (EAC group 4225; standard colonoscopy group 4151). In the meta-analysis, ADR was significantly increased in the EAC group vs. the standard colonoscopy group (41.3 % vs. 34.2 %; risk ratio [RR] = 1.20, 95 % confidence interval [CI] 1.06 to 1.36; P = 0.003; I ² = 79 %), especially for operators with low-to-moderate ADRs (< 35 %): RR = 1.51, 95 %CI 1.35 to 1.69; P < 0.001; I ² = 43 %). In contrast, this benefit was not reached for operators with high ADRs (> 45 %): RR = 1.01, 95 %CI 0.93 to 1.09; P = 0.87; I ² = 0.0 %). The mean number of adenomas per patient tended to be higher with EAC (mean difference = 0.11 adenomas/patient, 95 %CI – 0.17 to 0.38). Similar results were shown for polyp detection rates (61.6 % vs. 51.4 %; RR = 1.20, 95 %CI 1.06 to 1.36; P = 0.004). Use of the Endocuff did not impact the cecal intubation rate (95.1 % vs. 95.7 %; P = 0.08), or the procedure time compared with standard colonoscopy. Adverse events related to Endocuff were rare and exclusively mild mucosal erosion (4.0 %; 95 %CI 2.0 % to 8.0 %).

Conclusion With moderate-quality evidence, this study showed an improvement in ADR with EAC without major adverse events, especially for operators with low-to-moderate ADRs.

Introduction

Colonoscopy aims to detect and resect adenomas. Adenoma detection rate (ADR) is now the main quality indicator of colonoscopy because of its inverse correlation with interval cancer rate [1] [2] [3]. During the past decade, consistent progress in colonoscopy imaging has been made to improve adenoma detection, including high definition scopes, wide field of view, and chromoendoscopy (flexible spectral imaging color enhancement and narrow-band imaging). New innovative technologies have also been developed to improve ADR, including the Full-Spectrum Endoscopy system (Boston Scientific, Marlborough, Massachusetts, USA) [4], an extra-wide angle colonoscope (Olympus, Tokyo, Japan [5] [6]), a novel balloon colonoscope [7], and endoscopic devices, including distal caps [8].

The Endocuff (Norgine, Rueil Malmaison, France) is a new device that can be attached to the tip of the colonoscope to hold away colonic folds during withdrawal [9] ([Fig. 1]). The impact of the Endocuff on ADR remains controversial, despite strong signals in favor of its use [10] [11]. Numerous retrospective and prospective non-controlled studies have been published, including one meta-analysis comparing Endocuff-assisted colonoscopy (EAC) with standard colonoscopy [10]. Only four randomized controlled trials (RCTs) [12] [13] [14] [15], including one published only as an abstract, were included in this meta-analysis [10]. Another very recent meta-analysis compared the efficacy of add-on devices (cap, Endocuff, and Endorings) with each other or with standard colonoscopy in 25 RCTs that were carried out before 2017 [16]. The authors concluded that the improvement in ADR was only modest with the use of distal attachment devices, especially in low-performing endoscopists; however, heterogeneity of the pooled analyses was high (P < 0.001; I ² = 75 %), and the meta-analysis aggregated different types of add-on device. This year, supplementary outcomes from RCTs specifically comparing EAC with standard colonoscopy have been reported. Thus, it appeared necessary to propose a meta-analysis of RCTs focusing on the potential benefit of Endocuff.

Methods

This systematic review is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) ([Supplementary Table e1], available online) [17].

Literature research

Two reviewers (N.W. and Q.T.) independently conducted a systematic search according to the Cochrane Handbook for Systematic Reviews of Interventions. The search databases included PubMed and Google Scholar from January 2007 to 31 December 2017. For the PubMed search, only “Endocuff” [All Fields] was used as a keyword. For the Google scholar search, we used the free Publish or Perish software [18] with “Endocuff” and “randomized” as keywords. This software is useful for easy retrieval of all corresponding papers and extraction of metrics into a data frame for further analyses. Abstracts from Digestive Disease Week (DDW) and United European Gastroenterology (UEGW) meetings were also searched using “Endocuff” as a keyword in the respective websites (American Gastroenterological Association journals and United European Gastroenterology Journal, respectively).

Inclusion and exclusion criteria

Inclusion criteria

Only texts in English were considered. All RCTs that compared EAC with standard colonoscopy and included more than 100 adult patients undergoing colonoscopy were included. No selection criterion was used regarding patient characteristics or colonoscopy indications.

Exclusion criteria

All retrospective studies, non-controlled studies, reviews, animal studies, and duplicates were excluded. All RCTs evaluating a cap system other than Endocuff or RCTs including fewer than 100 patients were also excluded.

Data extraction

Two independent reviewers inspected the titles and abstracts of the included studies, and screened the selected literature according to inclusion and exclusion criteria. For each RCT included, the two main reviewers independently extracted data regarding ADR, polyp detection rate (PDR), cecal intubation rate, advanced adenoma detection rate, and ADR in the right-sided colon. Additionally, data on the mean number of adenomas or polyps per patient, procedure time (minutes), and proportions of adverse events were collected, if available. In cases of missing data, especially for continuous variables, we estimated the sample mean and SD from the sample size, median, range, and/or interquartile range, if available, according to the different approximation methods previously validated for performing meta-analyses [19]. In cases of disagreements, the final decision was taken by a third reviewer (C.V.) or by consensus decision.

Quality assessment

Based on recommendations from the Cochrane Handbook 5.1.0 software [20], assessment of RCT quality was focused on selection bias, performance bias, attrition bias, publication bias, and other biases. The analysis results were defined as “Yes” (low bias), “No” (high bias), or “Unclear” (bias-related information is not clear or bias cannot be determined). [ Supplementary Table e2 ] (available online) summarizes these recommendations.

Study end points

The primary objective of this meta-analysis was to calculate a pooled ADR. The secondary end points were PDR, number of adenomas or polyps detected per patient, cecal intubation rate, procedure time, and safety. Where heterogeneity of results existed, sensitivity analyses were conducted to identify potential cause(s) and to search criteria that could better select patients or endoscopists who might truly benefit from EAC compared with standard colonoscopy.

Statistical analyses

Pooled analyses were based on the Mantel – Haenszel models with risk ratio (RR), to compare outcomes of EAC with those of standard colonoscopy. Fixed- or random-effects models were used according to the heterogeneity across considered studies in order to calculate pooled estimate analyses for ADR, PDR, cecal intubation rate, and pooled proportion of adverse events related to Endocuff use.

Heterogeneity was assessed using chi-squared and I ² tests. When the analysis results showed no heterogeneity (P ≥ 0.10 and I ² < 50 %), we adopted a fixed-effect model for description of potential publication bias. When the analysis results showed the presence of heterogeneity (P < 0.10 or I ² ≥ 50 %), we chose a random-effect model [21]. Then, multiple sensitivity analyses were performed to identify the reason(s) for this heterogeneity [21]. These analyses were based on: a) exclusion of one study at a time, as recommended by the Cochrane collaboration; b) inclusion of RCTs according to study design (parallel group vs. crossover design); c) inclusion based on study design (multicenter vs. single-center studies); d) inclusion of RCTs according to size of the study population (“small” defined as n < 500 and “large” defined as n ≥ 500); e) inclusion of RCTs focused on screening colonoscopy (positive fecal immunochemical test or personal or familial history of colorectal cancer); f) inclusion of RCTs taking into account the Endocuff generation used (Endocuff vs. Endocuff Vision); g) inclusion of RTCs with ADR as the primary end point; h) inclusion of RCTs according to the level of ADR with standard colonoscopy (control arm): < 20 %, < 25 %, < 30 %, < 35 %, < 40 %, < 45 % and > 45 %; i) inclusion of RCTs according to the mean withdrawal time compared with standard colonoscopy (control arm) (< 10 vs. ≥ 10 minutes); j) inclusion of RCTs according to endoscopist experience (experts and those trained in Endocuff use, respectively); k) according to potential funding bias; and l) inclusion of only fully published RCTs.

If heterogeneity remained unknown and average effect appeared to favor or not favor EAC (by a confidence interval [CI] excluding 0), prediction interval of this effect was computed to assess the individual effect of Endocuff. A minimum of three studies were needed to calculate this interval [21]. Assessment of publication bias was also carried out by funnel plot asymmetry for pooled analyses involving at least 10 studies.

All statistical analyses were performed using R, version 3.2.2 (R project, Auckland, New Zealand) [22] and its metafor package [23].

Results

Study selection

From 265 studies identified using the described search strategy, we included 12 RCTs in the network meta-analysis [12] [13] [14] [15] [24] [25] [26] [27] [28] [29] [30] [31] ([Fig. 2]). The characteristics of included studies are summarized in [Table 3a], [Table 3b]. Seven RCTs comparing EAC with standard colonoscopy were published from 2014 to December 2017 [13] [14] [15] [24] [25] [26] [27]. Five others were reported as abstracts in the 2016 UEGW meeting [28] and in the 2015 [12], 2016 [29], and 2017 [30] [31] DDW meetings. A total of 8376 patients were included: 4225 patients in the EAC groups and 4151 in the standard colonoscopy groups. Mean age of patients across studies ranged from 55 to 67 years. The proportion of men was, in most cases, relatively balanced with that of women (47 % – 57 % male), except for one study (26.6 % male) [27]. These data were not available for three studies [28] [29] [31]. The proportion of colonoscopies performed for colorectal cancer screening was reported in eight studies [13] [24] [25] [26] [27] [29] [30] [31], and ranged from 45 % [30] to 100 % [25]. The number of operators ranged from 4 [24] [25] to 18 [27], and probably more [30]. Colonoscopies were performed exclusively by experts in the majority of cases [12] [13] [14] [15] [24] [25] [26] [30], with some operators having been trained in the use of the Endocuff at baseline [12] [13] [14] [27] [30]. The majority of studies were multicenter studies [12] [13] [14] [26] [30] [31], two were two-center trials [15] [28], and four were conducted at a single center [24] [25] [27] [29]. Two studies were designed as crossover trials [24] [26], for which only data of the first sequence were extracted for this network meta-analysis.

Quality assessment

Quality assessment was performed considering ADR, the primary end point ([Fig. 3], [Supplementary Fig. e4], available online). Overall, studies were associated with a high risk of performance and detection bias due to their unblinded design. Two studies [26] [30] had potentially high funding bias and four had an unclear bias in this respect [12] [15] [29] [31]. As all included studies were RCTs, the risk of selection bias was low. Only two RCTs did not clearly report the modality for randomization [24] [26]. Reported data were satisfactory overall for the primary end point (ADR), but some data for secondary end points were missing, especially in studies that were only reported in abstract format [12] [28] [29].

Primary end point: ADR

All 12 studies (n = 8376) reported ADR in each group [12] [13] [14] [15] [24] [25] [26] [27] [28] [29] [30] [31]. Pooled ADR was significantly higher with EAC (41.3 %; 95 %CI 35.7 % to 47.2 %) compared with standard colonoscopy (34.2 %; 95 %CI 26.6 % to 42.7 %; RR = 1.20, 95 %CI 1.06 to 1.36; P = 0.003, according to the random effect model), but heterogeneity across studies was high (I ² = 79 %) ([Fig. 5a]). The corresponding funnel plot is shown in [Supplementary Fig. e6] (available online) and shows evidence of publication bias. Heterogeneity remained similar in sensitivity analyses ([Table 4]), except when only including single-center RCTs [24] [25] [27] [29] (I ² = 40 %; P  = 0.17), those with potential high funding bias [26] [30] (I ² = 7 %; P  = 0.30), small studies (I ² = 29 % ; P = 0.22) [14] [24] [26] [27] [28], or studies with ADR level < 35 % [14] [15] [24] [26] [27] [31]. For the single-center studies, the yield of Endocuff compared with standard colonoscopy appeared null (RR = 1.01, 95 %CI 0.86 to 1.18; P = 0.91), whereas, in contrast, for studies with potential high funding bias, EAC appeared to be moderately beneficial (RR = 1.15, 95 %CI 1.03 to 1.29; P = 0.01). The best results were reached with small studies: RR = 1.40, 95 %CI 1.22 to 1.61; P < 0.001 ([Supplementary Fig. e7], available online) and more interestingly, for operators with low-to-moderate ADRs (< 35 %) in the standard colonoscopy group (RR = 1.51, 95 %CI 1.35 to 1.69; P  < 0.001). In contrast, for operators with high ADRs (> 45 %), results remained homogeneous (I ² = 0.0 %; P = 0.61) but with no difference between EAC and standard colonoscopy groups (RR = 1.01, 95 %CI 0.93 to 1.09; P = 0.87) ([Fig. 8]). Sensitivity analysis demonstrated that heterogeneity was not attributable to a single RCT or other criteria such as colonoscopy indication (I ² = 78 %), or large studies (I ² = 83 %). If we consider that the cause of heterogeneity remains unclear, the RR for ADR should be interpreted with caution, and the prediction interval should be considered, which ranged from 0.78 to 1.85 across the 12 RTCs ([Fig. 5a]).

Two studies (n = 1594) reported advanced adenoma detection rates [13] [25]. In the pooled analysis ([Supplementary Fig. e9a], available online), corresponding rates were similar between the two groups (19.4 % vs. 20.8 %; P = 0.47; RR = 0.93, 95 %CI 0.76 to 1.13) without heterogeneity across studies (I ² = 0 %). However, ADRs in the control arms were high in these two studies (≥ 50 %).

Three studies (n = 4378) reported the ADRs in the right-sided colon. With very high heterogeneity across studies (I ² = 94 %), ADRs in the right-sided colon were similar (P = 0.26; RR = 1.36, 95 %CI 0.80 to 2.34) in EAC (20.7 %; 95 %CI 12.2 % to 32.9 %) and standard colonoscopy groups (15 %; 95 %CI 8.2 % to 25.8 %) ([Supplementary Fig. e9b], available online). In sensitivity analyses, removal of the RCT by Catalano et al. [31] allowed homogeneity to be achieved (I ² = 0 %; P = 0.50) across the two remaining studies [13] [30]; homogeneity was also achieved when only RCTs with parallel group design or RCTS with expert endoscopists or those trained in the use of the Endocuff were considered, respectively. Therefore, there was no additional yield of EAC over standard colonoscopy for ADR in the right-sided colon (RR = 1.09, 95 %CI 0.94 to 1.26; P = 0.27).

The mean number of adenomas per patient was extracted from a total of nine studies [12] [13] [14] [15] [24] [25] [26] [27] [30] (n = 5795), either directly from the body text with corresponding SD values (n = 5) [12] [13] [25] [27] [30], or by estimation from the sample size, median, and interquartile range (n = 2) [14] [15]. Additional sample means could be calculated from total adenoma number and size of population in two RCTs (n = 2) [24] [26], but data were missing to allow estimation of the SD and so they were not included in pooled analysis. Hence, mean number of adenomas per patient tended to be higher in the EAC group (mean difference = + 0.11, 95 %CI – 0.17 to + 0.38), but heterogeneity was high (I ² = 92 %) ([Supplementary Fig. e10], available online). In sensitivity analyses, homogeneity was reached by including RTCs with colonoscopy for screening colorectal cancer or with a mean withdrawal time of ≥ 10 minutes [25] [27]: I ² = 41 %; P = 0.19. The mean difference of adenomas per patient was statistically significant in favor of EAC (mean difference = 0.08, 95 %CI 0.03 to 0.14; P = 0.002).

Secondary end points

Polyp detection rate

Five RCTs (n = 2529) evaluated PDR [14] [15] [25] [28] [29] ([Fig. 5b]), which was higher (P = 0.004; RR = 1.20, 95 %CI 1.06 to 1.36) with EAC (61.6 %, 95 %CI 56.2 % to 66.7 %) vs. standard colonoscopy (51.4 %, 95 %CI 40.7 % to 62.0 %). But there was high heterogeneity across studies (I ² = 70 %). Homogeneity was reached in sensitivity analyses ([Table 4]) by taking into account the ADR levels across studies. For ADR level < 30 %, the RR for PDR was in favor of Endocuff (RR = 1.39, 95 %CI 1.22 to 1.59; P < 0.001) with no heterogeneity (I ² = 0.0 %; P = 0.58) across the two corresponding studies [14] [15]. The corresponding pooled PDRs were 55.9 % (95 %CI 51.5 % to 60.2 %) vs. 40.1 % (95 %CI 35.9 % to 44.5 %) in the EAC and standard colonoscopy groups, respectively. The only study reporting PDR with ADR > 45 % [25] showed no benefit of Endocuff use in terms of PDR (RR = 1.05, 95 %CI 0.94 to 1.18; P = 0.38). Across multicenter RCTs, PDR was also significantly higher with EAC (58 %) compared with standard colonoscopy (43.2 %; RR = 1.34, 95 %CI 1.21 to 1.49; P < 0.001). In contrast, single-center design or mean withdrawal time ≥ 10 minutes were not statistically associated with a benefit of the Endocuff use (RR = 1.06, 95 %CI 0.97 to 1.15; P  = 0.22). Taking into account level of operators, one additional patient with polyp may be diagnosed after seven EAC procedures for operators with an ADR of < 30 %.

The mean number of polyps per patient could be extracted from six studies [13] [14] [15] [25] [26] [27], either directly from the body text with corresponding SD values (n = 2) [25] [27], or by estimation from the sample size, median, and interquartile range (n = 2) [14] [15]. Additional sample means could be calculated from total polyp number and size of population in two RCTs (n = 2) [13] [26], but data were missing to allow estimation of the SD and so were not included in pooled analysis ([ Supplementary Fig. e11 ], available online). Hence, mean differences of mean number of polyps per patient between EAC and standard colonoscopy groups tended to favor Endocuff (mean difference = 0.31, 95 %CI – 0.01 to 0.62) [24], with high heterogeneity (I ² = 89 %). By considering only studies involving endoscopists trained in the use of the Endocuff [12] [13] [14] [27] [30] or having a low ADR (< 25 %) [14] [27] [31], smaller studies (n < 500), or RCTs having ADR as the primary end point, heterogeneity disappeared (0 %) and the mean differences of number of polyps per patient become statistically significant in favor of EAC: mean difference = 0.22, 95 %CI 0.20 to 0.24; P < 0.001.

Technical outcomes

Eight studies (n = 5309) reported cecal intubation rates [13] [14] [15] [25] [26] [27] [28] [30]. Overall, no difference (P  = 0.08) was found in terms of cecal intubation rates between EAC (95.1 %) and standard colonoscopy (95.7 %) groups. However, heterogeneity was high (I ² = 66 %), but consistently improved (range 0 % – 16 %) by removing the Bhattacharyya study [25] (I ² = 16 %), or by including RCTs with ADR as the primary end point [12] [13] [14] [24] [27] [28] [29] [30] [31] (I ² = 18 %), those assessing Endocuff (I ² = 15 %), those involving endoscopists trained in the use of the Endocuff (I ² = 0 %), or across the ADR level. Hence, cecal intubation rates remained similar between EAC and standard colonoscopy ([Table 4], [Fig. 5c]).

The mean time for cecal intubation and withdrawal were estimated from sample size, median, range or interquartile range in four [13] [14] [24] [27] and seven [13] [14] [24] [25] [26] [27] [29] studies, respectively. We estimated total procedure time, in the vast majority of cases, by adding time to cecal intubation to the withdrawal time, and subtracting the time to valve intubation if available. Mean times were similar between the EAC and standard colonoscopy groups both for cecal intubation (mean difference = – 0.57, 95 %CI – 1.43 to 0.28; I ² = 88 %; [Supplementary Fig.e12], available online), withdrawal (mean difference = – 0.27, 95 %CI – 0.74 to 0.21; I ² = 86 %; P  = 0.27; [Supplementary Fig. e13], available online), and the whole procedure (mean difference = – 0.23, 95 %CI – 1.42 to 0.95; I ² = 87 %; P < 0.01). Despite an improvement in heterogeneity by removing the Van Doorn study (I ² = 24 % – 25 %), or by considering only RCTs with ADRs < 40 % in the control arm (I ² = 24 % – 25 %), mean difference of colonoscopy times was not significant. Level of ADR did not impact the results except in the only one study reporting ADRs of < 20 % and time for withdrawal [27]: mean difference + 0.70 minutes, 95 %CI 0.05 to 1.34; P  = 0.03.

Adverse events

A total of 10 studies reported adverse events in patients who underwent EAC (n = 3087) [13] [14] [15] [24] [25] [26] [27] [28] [29] [30] ([Fig. 5 d]). Excluding subjective complaints such as abdominal discomfort, two bleedings post-polypectomy, and one deep severe thrombosis with pulmonary embolism which was probably not related to colonoscopy [13], mucosal erosion was the only adverse event related to EAC. The corresponding pooled rate was 4.0 % (95 %CI 2.0 % to 8.0 %) with high heterogeneity (I ² = 91 %). Heterogeneity remained even after removing the Wada study, which had a particularly high adverse event rate (23 %). Sensitivity analysis of RCTs testing the new generation of Endocuff (Endocuff Vision) [25] [30] allowed homogeneity to be reached (I ² = 0 %; P  = 0.56) with a corresponding pooled rate of 0.1 % (95 %CI 0.01 % to 0.73 %). Homogeneity was also reached by considering ADR level. The pooled RR was 5.1 % (95 %CI 3.7 % to 6.9 %) for operators with ADRs of < 30 % vs. 0.13 % (95 %CI 0.02 % to 0.9 %) for operators with ADRs of > 45 % ([Table 4]).

Discussion

To the best of our knowledge, this study is one of the first meta-analyses of RCTs assessing the yield of Endocuff-assisted colonoscopy compared with standard colonoscopy in terms of ADR. One previous meta-analysis included retrospective studies and a few prospective non-controlled trials [10], and demonstrated that EAC increased the ADR (odds ratio [OR] = 1.49, 95 %CI 1.23 to 1.80; P  = 0.03) and the sessile serrated adenoma detection rate (OR = 2.34, 95 %CI 1.63 to 3.36; P < 0.001), without any significant adverse event. Another meta-analysis has recently been published but included analyses of several distal attachment devices, and showed only a modest improvement in ADR, especially in low-performing endoscopists [16]. Regarding the specific performance of Endocuff from nine RCTs, authors found low-quality evidence that Endocuff increases the ADR compared with standard colonoscopy (RR = 1.21, 95 %CI 1.03 to 1.41), which is in accordance with our results. However, in our study, we showed for the first time that homogeneity of the Endocuff effect on ADR was reached by considering only studies with low-to-moderate ADRs (< 35 %) or those with ADRs of > 45 %. Hence, it is clearly suggested that the expertise of the endoscopist is inversely correlated with the benefit of Endocuff in terms of ADR vs. standard colonoscopy. Similar results and findings with PDR support this hypothesis. In contrast, no benefit of Endocuff was found for operators with very high ADRs (> 45 %). Thus, the increase of ADR (+ 20 %, 95 %CI 6 % to 36 %; P  = 0.003) by using Endocuff should be interpreted with caution, and should be pondered with the ADR level of operator: + 51 % (95 %CI + 35 % to + 69 %) for operators with ADRs of < 35 % (P < 0.001) vs. no effect for operators with ADRs of > 45 % (P = 0.87). We also showed in the present study, across three RCTs [14] [27] [31], that ADR is increased by 68 % (95 %CI + 46 % to 94 %) in operators with low ADR (< 25 %). This is a very important point because the 25 % threshold of ADR is an emerging minimal target for improving the overall practice of gastroenterologist and reducing the risk of interval colorectal cancer [32] [33]. Moreover, a limitation of these RCTs is the non-blinded design, which was underlined by the positive effect of Endocuff shown in sensitivity analyses that considered only RCTs with a potential high funding bias (I ² = 7.1 %; RR = 1.15, 95 %CI 1.03 to 1.29]; P = 0.01). One of the best results in terms of homogeneity and effect of Endocuff was reached by considering small studies, but the reasons remain unclear. Further research is needed to identify the subgroups of patients that might benefit most, although screening colonoscopy appears to be the most appropriate setting for the use of Endocuff.

The strengths of our study are the number of RCTs included and the exhaustive sensitivity analyses. For the first time, we considered different definitions of endoscopist experience. ADRs of < 20 % [32] or < 28 % [33] are thresholds that have been previously associated with an increased risk of interval colorectal cancer. Considering these cutoffs, heterogeneity was improved. We also confirmed that use of the Endocuff does not seem to increase procedure time, even in low-to-moderately experienced endoscopists. Moreover, cecal intubation rate did not seem to be impacted by the device, regardless of the operator level. Finally, nature (mucosal erosion) and frequency of adverse events were acceptable.

The limitations of our study are the potential bias related to non-blinded assessment. Potential funding bias remains possible. Hence, further studies are needed because Endocuff may impact on the management of patients. In the RCT by De Palma et al., EAC shortened the surveillance interval in 6.6 % and 7.3 % of cases according to European and American guidelines, respectively [24], compared with standard colonoscopy. Furthermore, the Endocuff is cheaper than other detection systems such as balloon or extra-wide angle colonoscopies [6], and could be more accessible in clinical practice. Nevertheless, the impact of EAC on interval colorectal cancer incidence cannot be evaluated from the current literature.

The other limitations include the limited data available on detection of advanced adenomas, adenomas located in the right-sided colon, and adenomas < 5 mm, and the total lack of data on serrated adenoma detection. Indeed, it would be clinically relevant to demonstrate a detection yield for these different lesions in order to support a systematic use of Endocuff for screening colonoscopy. Hence, outcomes regarding advanced adenoma detection or adenoma located in the right-side of the colon remain unclear in this meta-analysis. According to the ClinicalTrials.gov website (https://clinicaltrials.gov), at least three other RCTs comparing EAC with standard colonoscopy are currently active, underlining that these results should be interpreted with caution given their temporary value.

Conclusions

Even if the present meta-analysis of RCTs demonstrated low-quality evidence of Endocuff yield in adenoma detection compared with standard colonoscopy for all operators, its use could be moderately to highly recommended for operators with low (< 25 %) to moderate (< 35 %) ADRs. Overall, routine use of this simple accessory may be considered by operators to improve ADR, especially for screening colonoscopies; adverse events are rare and mild, and procedure time and cecal intubation rates are similar to those of standard colonoscopy. Further research is needed to confirm the impact of the operator’s ADR level for this benefit and to identify patient characteristics associated with the higher benefit of using Endocuff.

Competing interests

None.

• References

• 1 Cooper GS, Xu F, Barnholtz Sloan JS. et al. Prevalence and predictors of interval colorectal cancers in medicare beneficiaries. Cancer 2012; 118: 3044-3052
  CrossrefPubMedSearch in Google Scholar
• 2 Singh S, Singh PP, Murad MH. et al. Prevalence, risk factors, and outcomes of interval colorectal cancers: a systematic review and meta-analysis. Am J Gastroenterol 2014; 109: 1375-1389
  CrossrefPubMedSearch in Google Scholar
• 3 Baxter NN, Sutradhar R, Forbes SS. et al. Analysis of administrative data finds endoscopist quality measures associated with postcolonoscopy colorectal cancer. Gastroenterology 2011; 140: 65-72
  CrossrefPubMedSearch in Google Scholar
• 4 Gralnek IM, Siersema PD, Halpern Z. et al. Standard forward-viewing colonoscopy versus full-spectrum endoscopy: an international, multicentre, randomised, tandem colonoscopy trial. Lancet Oncol 2014; 15: 353-360
  CrossrefPubMedSearch in Google Scholar
• 5 Uraoka T, Tanaka S, Matsumoto T. et al. A novel extra-wide-angle-view colonoscope: a simulated pilot study using anatomic colorectal models. Gastrointest Endosc 2013; 77: 480-483
  CrossrefPubMedSearch in Google Scholar
• 6 Uraoka T, Tanaka S, Oka S. et al. Feasibility of a novel colonoscope with extra-wide angle of view: a clinical study. Endoscopy 2015; 47: 444-448
  Thieme ConnectPubMedSearch in Google Scholar
• 7 Gralnek IM, Suissa A, Domanov S. Safety and efficacy of a novel balloon colonoscope: a prospective cohort study. Endoscopy 2014; 46: 883-887
  Thieme ConnectPubMedSearch in Google Scholar
• 8 Pohl H, Bensen SP, Toor A. et al. Cap-assisted colonoscopy and detection of Adenomatous Polyps (CAP) study: a randomized trial. Endoscopy 2015; 47: 891-897
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Lenze F, Beyna T, Lenz P. et al. Endocuff-assisted colonoscopy: a new accessory to improve adenoma detection rate? Technical aspects and first clinical experiences. Endoscopy 2014; 46: 610-614
  Thieme ConnectPubMedSearch in Google Scholar
• 10 Chin M, Karnes W, Jamal MM. et al. Use of the Endocuff during routine colonoscopy examination improves adenoma detection: a meta-analysis. World J Gastroenterol 2016; 22: 9642-9649
  CrossrefPubMedSearch in Google Scholar
• 11 Pioche M, Matsumoto M, Takamaru H. et al. Endocuff-assisted colonoscopy increases polyp detection rate: a simulated randomized study involving an anatomic colorectal model and 32 international endoscopists. Surg Endosc 2016; 30: 288-295
  CrossrefPubMedSearch in Google Scholar
• 12 Cattau E, Leal R, Ormseth E. et al. The effect of Endocuff-assisted colonoscopy on adenoma detection rate: a randomized trial in community ambulatory surgical centers. Am J Gastroenterol 2015; 110: S602
  PubMedSearch in Google Scholar
• 13 van Doorn SC, van der Vlugt M, Depla A. et al. Adenoma detection with Endocuff colonoscopy versus conventional colonoscopy: a multicentre randomised controlled trial. Gut 2017; 66: 438-445
  CrossrefPubMedSearch in Google Scholar
• 14 Floer M, Biecker E, Fitzlaff R. et al. Higher adenoma detection rates with endocuff-assisted colonoscopy – a randomized controlled multicenter trial. PLoS One 2014; 9: e114267
  Search in Google Scholar
• 15 Biecker E, Floer M, Heinecke A. et al. Novel endocuff-assisted colonoscopy significantly increases the polyp detection rate: a randomized controlled trial. J Clin Gastroenterol 2015; 49: 413-418
  CrossrefPubMedSearch in Google Scholar
• 16 Facciorusso A, Del Prete V, Buccino RV. et al. Comparative efficacy of colonoscope distal attachment devices in increasing rates of adenoma detection: a network meta-analysis. Clin Gastroenterol Hepatol 2017; pii: S1542-3565(17)31318-6
  PubMedSearch in Google Scholar
• 17 Moher D, Liberati A, Tetzlaff J. et al. Pre-ferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. The PRISMA Group PLoS Med 2009; 6: 07e1000097
  PubMedSearch in Google Scholar
• 18 Harzing AW. Publish or Perish. 2007 Available from: http://www.harzing.com/pop.htm
  PubMedSearch in Google Scholar
• 19 Wan X, Wang W, Liu J. et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014; 14: 135
  Search in Google Scholar
• 20 Available from: Higgins JPT, Green S. , eds. Cochrane handbook for systematic reviews of interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. 2011 http://handbook.cochrane.org.accesdistant.sorbonne-universite.fr
  PubMedSearch in Google Scholar
• 21 Riley RD, Higgins JPT, Deeks JJ. Interpretation of random effects meta-analyses. BMJ 2011; 342: d549
  CrossrefPubMedSearch in Google Scholar
• 22 R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2014 Available from: http://www.R-project.org/
  PubMedSearch in Google Scholar
• 23 Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010 36. Available from: https://www.jstatsoft.org/article/view/v036i03
  PubMedSearch in Google Scholar
• 24 De Palma GD, Giglio MC, Bruzzese D. et al. Cap cuff-assisted colonoscopy versus standard colonoscopy for adenoma detection: a randomized back-to-back study. Gastrointest Endosc 2018; 87: 232-240
  CrossrefPubMedSearch in Google Scholar
• 25 Bhattacharyya R, Chedgy F, Kandiah K. et al. Endocuff-assisted vs. standard colonoscopy in the fecal occult blood test-based UK Bowel Cancer Screening Programme (E-cap study): a randomized trial. Endoscopy 2017; 49: 1043-1050
  Thieme ConnectPubMedSearch in Google Scholar
• 26 Triantafyllou K, Polymeros D, Apostolopoulos P. et al. Endocuff-assisted colonoscopy is associated with a lower adenoma miss rate: a multicenter randomized tandem study. Endoscopy 2017; 49: 1051-1060
  Thieme ConnectPubMedSearch in Google Scholar
• 27 González-Fernández C, García-Rangel D, Aguilar-Olivos NE. et al. Higher adenoma detection rate with the endocuff: a randomized trial. Endoscopy 2017; 49: 1061-1068
  Thieme ConnectPubMedSearch in Google Scholar
• 28 Wada Y, Wada Y, Wada M. et al. OP169 Efficacy of endocuff-assisted colonoscopy in the detection of colorectal polyps. Abstr UEGW. 2016
  PubMedSearch in Google Scholar
• 29 Hass D, Jaffe C, Malangone L. et al. Endocuff (EC) increases adenoma detection rates on surveillance colonoscopy and improves efficiency of colonoscopy by shortening of withdrawal times. Gastroenterology 2016; 150: S28
  PubMedSearch in Google Scholar
• 30 Ngu WS, Bevan R, Tsiamoulos ZP. et al. Improving colorectal adenoma detection rate with Endocuff Vision. Results of the Adenoma Randomised Controlled Trial. Gastroenterology 2017; 152: S1313-S1314
  PubMedSearch in Google Scholar
• 31 Catalano MF, Khan N, Lajin M. et al. Increase in adenoma detection rate (ADR) using endocuff (EC) assisted colonoscopy. Gastrointest Endosc 2017; 85: AB495
  PubMedSearch in Google Scholar
• 32 Kaminski MF, Regula J, Kraszewska E. et al. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010; 362: 1795-1803
  CrossrefPubMedSearch in Google Scholar
• 33 Corley DA, Jensen CD, Marks AR. et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014; 370: 1298-1306
  CrossrefPubMedSearch in Google Scholar

Corresponding author

• References

• 1 Cooper GS, Xu F, Barnholtz Sloan JS. et al. Prevalence and predictors of interval colorectal cancers in medicare beneficiaries. Cancer 2012; 118: 3044-3052
  CrossrefPubMedSearch in Google Scholar
• 2 Singh S, Singh PP, Murad MH. et al. Prevalence, risk factors, and outcomes of interval colorectal cancers: a systematic review and meta-analysis. Am J Gastroenterol 2014; 109: 1375-1389
  CrossrefPubMedSearch in Google Scholar
• 3 Baxter NN, Sutradhar R, Forbes SS. et al. Analysis of administrative data finds endoscopist quality measures associated with postcolonoscopy colorectal cancer. Gastroenterology 2011; 140: 65-72
  CrossrefPubMedSearch in Google Scholar
• 4 Gralnek IM, Siersema PD, Halpern Z. et al. Standard forward-viewing colonoscopy versus full-spectrum endoscopy: an international, multicentre, randomised, tandem colonoscopy trial. Lancet Oncol 2014; 15: 353-360
  CrossrefPubMedSearch in Google Scholar
• 5 Uraoka T, Tanaka S, Matsumoto T. et al. A novel extra-wide-angle-view colonoscope: a simulated pilot study using anatomic colorectal models. Gastrointest Endosc 2013; 77: 480-483
  CrossrefPubMedSearch in Google Scholar
• 6 Uraoka T, Tanaka S, Oka S. et al. Feasibility of a novel colonoscope with extra-wide angle of view: a clinical study. Endoscopy 2015; 47: 444-448
  Thieme ConnectPubMedSearch in Google Scholar
• 7 Gralnek IM, Suissa A, Domanov S. Safety and efficacy of a novel balloon colonoscope: a prospective cohort study. Endoscopy 2014; 46: 883-887
  Thieme ConnectPubMedSearch in Google Scholar
• 8 Pohl H, Bensen SP, Toor A. et al. Cap-assisted colonoscopy and detection of Adenomatous Polyps (CAP) study: a randomized trial. Endoscopy 2015; 47: 891-897
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Lenze F, Beyna T, Lenz P. et al. Endocuff-assisted colonoscopy: a new accessory to improve adenoma detection rate? Technical aspects and first clinical experiences. Endoscopy 2014; 46: 610-614
  Thieme ConnectPubMedSearch in Google Scholar
• 10 Chin M, Karnes W, Jamal MM. et al. Use of the Endocuff during routine colonoscopy examination improves adenoma detection: a meta-analysis. World J Gastroenterol 2016; 22: 9642-9649
  CrossrefPubMedSearch in Google Scholar
• 11 Pioche M, Matsumoto M, Takamaru H. et al. Endocuff-assisted colonoscopy increases polyp detection rate: a simulated randomized study involving an anatomic colorectal model and 32 international endoscopists. Surg Endosc 2016; 30: 288-295
  CrossrefPubMedSearch in Google Scholar
• 12 Cattau E, Leal R, Ormseth E. et al. The effect of Endocuff-assisted colonoscopy on adenoma detection rate: a randomized trial in community ambulatory surgical centers. Am J Gastroenterol 2015; 110: S602
  PubMedSearch in Google Scholar
• 13 van Doorn SC, van der Vlugt M, Depla A. et al. Adenoma detection with Endocuff colonoscopy versus conventional colonoscopy: a multicentre randomised controlled trial. Gut 2017; 66: 438-445
  CrossrefPubMedSearch in Google Scholar
• 14 Floer M, Biecker E, Fitzlaff R. et al. Higher adenoma detection rates with endocuff-assisted colonoscopy – a randomized controlled multicenter trial. PLoS One 2014; 9: e114267
  Search in Google Scholar
• 15 Biecker E, Floer M, Heinecke A. et al. Novel endocuff-assisted colonoscopy significantly increases the polyp detection rate: a randomized controlled trial. J Clin Gastroenterol 2015; 49: 413-418
  CrossrefPubMedSearch in Google Scholar
• 16 Facciorusso A, Del Prete V, Buccino RV. et al. Comparative efficacy of colonoscope distal attachment devices in increasing rates of adenoma detection: a network meta-analysis. Clin Gastroenterol Hepatol 2017; pii: S1542-3565(17)31318-6
  PubMedSearch in Google Scholar
• 17 Moher D, Liberati A, Tetzlaff J. et al. Pre-ferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. The PRISMA Group PLoS Med 2009; 6: 07e1000097
  PubMedSearch in Google Scholar
• 18 Harzing AW. Publish or Perish. 2007 Available from: http://www.harzing.com/pop.htm
  PubMedSearch in Google Scholar
• 19 Wan X, Wang W, Liu J. et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014; 14: 135
  Search in Google Scholar
• 20 Available from: Higgins JPT, Green S. , eds. Cochrane handbook for systematic reviews of interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. 2011 http://handbook.cochrane.org.accesdistant.sorbonne-universite.fr
  PubMedSearch in Google Scholar
• 21 Riley RD, Higgins JPT, Deeks JJ. Interpretation of random effects meta-analyses. BMJ 2011; 342: d549
  CrossrefPubMedSearch in Google Scholar
• 22 R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2014 Available from: http://www.R-project.org/
  PubMedSearch in Google Scholar
• 23 Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010 36. Available from: https://www.jstatsoft.org/article/view/v036i03
  PubMedSearch in Google Scholar
• 24 De Palma GD, Giglio MC, Bruzzese D. et al. Cap cuff-assisted colonoscopy versus standard colonoscopy for adenoma detection: a randomized back-to-back study. Gastrointest Endosc 2018; 87: 232-240
  CrossrefPubMedSearch in Google Scholar
• 25 Bhattacharyya R, Chedgy F, Kandiah K. et al. Endocuff-assisted vs. standard colonoscopy in the fecal occult blood test-based UK Bowel Cancer Screening Programme (E-cap study): a randomized trial. Endoscopy 2017; 49: 1043-1050
  Thieme ConnectPubMedSearch in Google Scholar
• 26 Triantafyllou K, Polymeros D, Apostolopoulos P. et al. Endocuff-assisted colonoscopy is associated with a lower adenoma miss rate: a multicenter randomized tandem study. Endoscopy 2017; 49: 1051-1060
  Thieme ConnectPubMedSearch in Google Scholar
• 27 González-Fernández C, García-Rangel D, Aguilar-Olivos NE. et al. Higher adenoma detection rate with the endocuff: a randomized trial. Endoscopy 2017; 49: 1061-1068
  Thieme ConnectPubMedSearch in Google Scholar
• 28 Wada Y, Wada Y, Wada M. et al. OP169 Efficacy of endocuff-assisted colonoscopy in the detection of colorectal polyps. Abstr UEGW. 2016
  PubMedSearch in Google Scholar
• 29 Hass D, Jaffe C, Malangone L. et al. Endocuff (EC) increases adenoma detection rates on surveillance colonoscopy and improves efficiency of colonoscopy by shortening of withdrawal times. Gastroenterology 2016; 150: S28
  PubMedSearch in Google Scholar
• 30 Ngu WS, Bevan R, Tsiamoulos ZP. et al. Improving colorectal adenoma detection rate with Endocuff Vision. Results of the Adenoma Randomised Controlled Trial. Gastroenterology 2017; 152: S1313-S1314
  PubMedSearch in Google Scholar
• 31 Catalano MF, Khan N, Lajin M. et al. Increase in adenoma detection rate (ADR) using endocuff (EC) assisted colonoscopy. Gastrointest Endosc 2017; 85: AB495
  PubMedSearch in Google Scholar
• 32 Kaminski MF, Regula J, Kraszewska E. et al. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010; 362: 1795-1803
  CrossrefPubMedSearch in Google Scholar
• 33 Corley DA, Jensen CD, Marks AR. et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014; 370: 1298-1306
  CrossrefPubMedSearch in Google Scholar

• Supplementary Material