The main indication for video capsule endoscopy (VCE) is small bowel bleeding in patients with previous negative upper gastrointestinal endoscopy and colonoscopy. The utility of enteroscopy has been shown to be similar to that of VCE in small bowel bleeding, albeit false negatives have been described with both techniques, depending on the type of lesion.1 The diagnostic yield of VCE in the context of small bowel bleeding is reported in different studies to be between 32-83%2,3 and said procedure helps define the insertion route for enteroscopy.4 The majority of studies on the agreement between the two techniques in relation to the different types of lesions are small case series with scant scientific evidence. There are few meta-analyses that comparatively analyze their diagnostic yield. Enteroscopy has shown a better diagnostic yield (68.2%) when performed after VCE, compared with enteroscopy alone (43.8%), suggesting that the two modalities are complementary.5 Enteroscopy is indicated when VCE results in positive findings that require intervention or diagnostic confirmation.6

The aim of the present study was to define the levels of agreement between VCE and enteroscopy, according to the type of small bowel lesion and the findings from a prior VCE performed to study small bowel bleeding, in a cohort of patients from a tertiary referral hospital.

All the patients suspected of presenting with small bowel bleeding that had undergone VCE and balloon-assisted enteroscopy at the Hospital Universitario San Ignacio, within the time frame of June 2011 to June 2019, were included in the study. Small bowel bleeding was defined in accordance with the definition of the American College of Gastroenterology,7 and the type of bleeding (overt or occult) was considered. All the patients underwent upper gastrointestinal endoscopy and colonoscopy, as well as other imaging techniques, with no significant findings to explain the bleeding. Patients under 18 years of age and those that did not undergo VCE prior to the enteroscopy were excluded from the study.

Within the time frame of June 2011 to June 2019, 428 VCEs and 313 enteroscopies (201 antegrade and 112 retrograde) were performed at the gastroenterology unit of the Hospital Universitario San Ignacio. Seventy-four (23.6%) VCEs were followed by single balloon or double balloon enteroscopy with antegrade and/or retrograde access in 71 patients suspected of presenting with small bowel bleeding. The mean time between the two procedures was 19.5 (IQR 5.0-94) days.

Mean patient age was 63.9 ± 13.5 years, with a predominance of women (n = 42) (56.8%). The main indications for VCE were obscure overt gastrointestinal bleeding in 46 (62.2%) patients and iron-deficiency anemia in 21 (28.4%) patients (Table 1). Angiectasias presented mainly in the jejunum in 23 (31.1%) patients (Fig. 1), followed by the duodenum in 13 (17.6%) patients, the ileum in 9 (12.2%) patients, and gastric angiectasias in one (1.4%) patient. The tumor was located in the jejunum in 11 (14.9%) patients and in the ileum in 3 (4.1%) patients (Figs. 2 and 3). The majority of the inflammatory lesions were in the ileum, with a total of 13 patients (17.6%). VCE complications presented in 6 (8.1%) patients (Table 2).

Figure 1.

Angiectasia measuring 4 mm in the proximal jejunum. A) VCE, B) Antegrade double balloon enteroscopy.

(0.22MB).

Figure 2.

Moderately differentiated ulcerated intestinal adenocarcinoma. A) VCE, B) Antegrade double balloon enteroscopy.

(0.13MB).

Figure 3.

Well differentiated neuroendocrine tumor. A) VCE, B) Antegrade double balloon enteroscopy.

(0.3MB).

Sixty-three (85.13%) antegrade enteroscopies and 34 (45.94%) retrograde enteroscopies with single or double balloon were performed (some patients required both routes), confirming a total of 43 lesions with the following findings: angiectasias were identified mainly in the jejunum in 18 (24.3%) patients, followed by tumors in 6 (8.1%) patients; angiectasias were identified in the duodenum in 9 (12.2%) patients, followed by inflammatory lesions in 4 (5.5%) patients, and found via the antegrade route. Retrograde enteroscopy identified angiectasias in the colon in 5 (6.8%) patients and diverticula in 2 (2.7%) patients; enteritis was viewed in the ileum in 10 (13.5%) patients, stricture in 2 (2.7%) patients, angiectasias and subepithelial lesion in one (1.4%) patient; and enteritis and subepithelial lesion in the jejunum in one (1.4%) patient (Table 3).

Overall diagnostic yield for the detection of positive findings was greater in VCE than in enteroscopy (86.5% vs. 58.1%, respectively, p = 0.0527), with no statistically significant differences between the two studies (Table 4). VCE detected 64 lesions, with signs of active bleeding in 13 (17.6%). Importantly, there were no signs of angiectasias in the enteroscopies of 14 (18.91%) patients whose VCEs showed them. There were statistically significant differences in the diagnostic yield of VCE, compared with enteroscopy, with respect to angiectasias (41.9% vs. 29.7% respectively, p = 0.0001) (Fig. 1), tumors (18.9% vs. 9.4%, respectively, p = 0.0002) (Figs. 2 and 3), and inflammatory lesions (23% vs. 19%, respectively, p < 0.0001).

Overall, the kappa index for positive findings between VCE and enteroscopy was poor (Ik = 0.17, 95% CI: –0.0097-0.3543). However, the level of diagnostic agreement of the findings was modified, according to the type of lesion (Table 5).

The present study found a similar diagnostic yield between VCE and antegrade and/or retrograde, single or double balloon enteroscopy performed after the VCE, on patients suspected of having small bowel bleeding that had negative upper gastrointestinal endoscopy and colonoscopy (86.5% vs. 58.1% respectively, p = 0.0527). In addition, overall agreement was low, regarding the same findings in the two modalities, with respect to lesions and anatomic site (k = 0.17). However, diagnostic agreement was modified, according to the type of lesion, obtaining a higher frequency of agreement in inflammatory lesions (k = 0.71), angiectasias (k = 0.45), and tumor (k = 0.40).

Due to its location and length, the small bowel used to be considered the Pandora’s box of the gastrointestinal tract for the study of pathologies at that level. The diagnostic approach to the middle intestine improved when the Food and Drug Administration (FDA) approved its examination through enteroscopy in the year 2000 and VCE in 2001. Of those two diagnostic methods, VCE is the method of choice for the study of the small bowel, because of its technical and operative characteristics, being less invasive and less expensive. It indicates the access route if a later invention is required. It is limited by the fact that biopsies cannot be taken, and interventions cannot be performed,5,6,8 which is where enteroscopy becomes essential as a diagnostic and therapeutic method.9–11 The majority of reports describe a similar yield for the two techniques. In the study by Tian Min et al.,12 VCE followed by enteroscopy was performed on 62 patients, finding an overall detection rate of intestinal lesions of 70.9% (44/62), with no significant differences between the 2 methods (p = 0.6739). Likewise, there were no statistically significant differences in the diagnostic yield between the two techniques in the present study.

In a systematic review by Liao Z et al.13 that included 227 retrospective and prospective studies, with a total of 22,840 VCEs for evaluating small bowel lesions, the most frequent indication was occult gastrointestinal bleeding (66%). The diagnostic yield was 59.4% for lesions, in general, and 60.5% for occult bleeding. Angiectasias were identified in 50% of the cases. The complete retention rate was 1.4%, the same as the rate in the present study (n = 1, 1.4%). The diagnostic yield of that study was higher for positive pathologic findings (86.5%) and the main indication was obscure overt gastrointestinal bleeding.

Fukumoto et al.14 evaluated 76 patients suspected of having small bowel pathology. Total enteroscopy was achieved in 77.6% of the patients with VCE and 56.6% with double balloon enteroscopy. Lesions were detected in 42/76 patients with VCE and 46/76 patients with enteroscopy, with no statistically significant differences and a moderate agreement rate (kappa index = 0.57). Differences were observed between VCE and enteroscopy in 16 patients and no lesions were detected through enteroscopy in 6 patients. Different analyses have shown that VCE is limited for detecting certain submucosal lesions due to the lack of air insufflation, especially in the proximal intestine. In the present study, 64/74 lesions were detected through VCE and 43/74 lesions were detected through enteroscopy, with no statistically significant differences, but with poor agreement (0.17, 95% CI: –0.0097-0.3543). No lesions were detected through enteroscopy in 24 patients, in whom lesions were identified in the previous VCE. In a meta-analysis and subgroup analysis of positive findings, lesions were detected in 72.2% of the VCEs but were identified in only 27.8% of the double balloon enteroscopies.15 Unlike that reported by other authors, in our study, the diagnostic yield for tumor was greater with VCE than with enteroscopy (18.9% vs. 9.4%, respectively, p = 0.0002), with moderate agreement (0.40, 95% CI: 0.1217-0.6794). The poor agreement in the interpretation of positive pathologic findings was due to the fact that we determined the agreement between the lesions identified with the same findings and at the same anatomic site in both the VCE and enteroscopy. That was the importance of our study, which differs from previous analyses. A large portion of the literature does not take those relations into account, which could be due to the poor agreement between the two procedures.

A 2011 meta-analysis that included 10 studies, with a total of 651 patients, reported that the diagnostic yield for VCE and double balloon enteroscopy was 62% and 56%, respectively, with no difference in identifying the cause of bleeding.16 The main indication for enteroscopy was overt intestinal bleeding, with an overall yield of pathologic findings in the small intestine of 65.21%.17 In our case series, the diagnostic yield was similar for the 2 procedures (86.5% vs. 58.1%, respectively, p = 0.0527), with higher values than others reported. Nevertheless, diagnostic yield is modified by different factors. Therefore, inconclusive or negative VCEs are indications for performing complementary enteroscopy.18 In the present study, in the 10 patients with negative VCE, lesions were confirmed through enteroscopy in 3 (30%) patients (VCE false negatives). Likewise, the enteroscopy false negatives in 24 (37.5%) patients could be related to the interval of time between the two procedures and the difficulty in performing complete enteroscopies, which were only achieved in 5 (6.8%) patients.

In a retrospective cohort of 418 patients that underwent VCE, Shiani et al.19 evaluated 95 patients with positive findings in VCE that then underwent diagnostic and therapeutic single balloon enteroscopy. There was strong agreement for active bleeding and blood clots, moderate agreement for vascular lesions, and poor agreement for ulcers. There was no correlation for masses and polyps, unlike the present study that showed moderate agreement for tumors.

Marmo et al.20 conducted a prospective, multicenter study on 193 patients and reported good agreement for vascular lesions (0.72, 95% CI: 0.59-0.84) and inflammatory lesions (0.78, 95% CI: 0.58-0.99]) and no agreement for polyps or tumors. Similarly, our study showed that agreement varied, depending on the type of lesion. It was good for inflammatory lesions, but moderate for tumors. Our results concur with reports that the two methods are specific and complementary for the detection of tumors in the small intestine.21

Systematic reviews including 12,823 procedures, describe total enteroscopy, defined as the examination of the entire small bowel, whether by antegrade or retrograde access routes or a combination of the two, in 44% of the patients. One of the limitations for its performance is the technical difficulty involved and the duration of the examination, which can take up to 4 hours.22 The performance rate of complete single balloon enteroscopy has been reported at 26.1%,23 which is lower, compared with the double balloon procedure.24,25 There is quite a difference with the present study, in which total enteroscopy was achieved in 6.8% of the patients, with the limitation of not distinguishing between the single or double balloon procedures. Reports in the literature show no difference in the diagnostic yield, the therapeutic yield, or the failed enteroscopy rate between single and double balloon enteroscopy.26,27

Other limitations of the present study, apart from those already mentioned, were its retrospective design, interobserver variation, and the fact that it was conducted at a referral hospital, meaning there could be a reference bias with an increase in diagnostic yield. In addition, a Pillcam SB3 and SB2 subgroup analysis was not carried out, given that, according to the literature, the SB3 detects a higher number of P1 lesions, without modifying the diagnostic yield of the P2 lesions that were selected as significant, given their high potential for bleeding.28

The present study showed that the overall yield was similar between VCE and single and double balloon enteroscopy for detecting small bowel lesions, with a poor overall agreement for positive findings. The main factor that could modify those results was the type of lesion, with good agreement for inflammatory lesions and moderate agreement for angiectasias and tumors.

The present study was self-funded.

The authors declare that there is no conflict of interest.