Digestive Diseases and Sciences

, Volume 56, Issue 6, pp 1763–1768

Bleeding Lesions Within Reach of Conventional Endoscopy in Capsule Endoscopy Examinations for Obscure Gastrointestinal Bleeding: Is Repeating Endoscopy Economically Feasible?

Authors

    • Hepatogastroenterology Unit, First Department of Medicine, PropaedeuticMedical School, Athens University, Laiko General Hospital, Athens
  • Kostis Papaxoinis
    • Hepatogastroenterology Unit, First Department of Medicine, PropaedeuticMedical School, Athens University, Laiko General Hospital, Athens
  • Nikos Viazis
    • Second Department of GastroenterologyEvangelismos General Hospital
  • Anastasia Kegioglou
    • Second Department of GastroenterologyEvangelismos General Hospital
  • Ioannis Binas
    • Hepatogastroenterology Unit, First Department of Medicine, PropaedeuticMedical School, Athens University, Laiko General Hospital, Athens
  • Dimitrios Karamanolis
    • Second Department of GastroenterologyEvangelismos General Hospital
  • Spiros D. Ladas
    • Hepatogastroenterology Unit, First Department of Medicine, PropaedeuticMedical School, Athens University, Laiko General Hospital, Athens
Original Article

DOI: 10.1007/s10620-011-1592-3

Cite this article as:
Vlachogiannakos, J., Papaxoinis, K., Viazis, N. et al. Dig Dis Sci (2011) 56: 1763. doi:10.1007/s10620-011-1592-3

Abstract

Background

Most tertiary gastroenterology centers currently offer an open-access capsule endoscopy (CE) service, including patients with obscure gastrointestinal bleeding. However, CE may identify lesions missed by conventional endoscopy.

Aims

To determine the incidence of bleeding lesions missed by the preceding gastroscopy/colonoscopy that were revealed by CE and compare potential differences in the rate of identifying such lesions in patients that we investigated as opposed to those investigated elsewhere.

Methods

We prospectively reviewed data from patients subjected to CE for obscure bleeding. We analyzed all cases where a source of bleeding was located in the stomach, duodenum, or colon.

Results

A total of 317 consecutive patients were subjected to CE for obscure gastrointestinal bleeding within 28 months. Prior to CE examination, 174 patients had gastroscopy and colonoscopy in our institutions and 143 were referrals, all with negative endoscopic investigation. We identified 11 (3.5%) cases where the source of bleeding was found in the stomach (n = 4) or the cecum (n = 7). There was a significant difference of extra small intestinal lesions diagnosed by CE between referrals (9/143, 6.3%) and endoscopic investigation performed in our institutions (2/174, 1.15%), (p = 0.026). The estimated cost of re-endoscoping in our institution all CE referrals would be €50,050 (143 patients × €350), to avoid unnecessary CE examinations (9 patients × €600 = €5,400).

Conclusions

Reading the whole CE video is important, because small-bowel CE may identify lesions responsible for obscure bleeding missed by the preceding gastroscopy and colonoscopy. Repeating conventional endoscopy by experts before CE is not a cost-effective approach.

Keywords

Obscure gastrointestinal bleedingCapsule endoscopyGastroscopyColonoscopy

Introduction

Obscure gastrointestinal bleeding (OGIB) is defined as bleeding of unknown origin that persists or recurs after negative initial endoscopic evaluation (gastroscopy and colonoscopy) [1]. OGIB accounts for approximately 5% of gastrointestinal bleeding and is usually due to a lesion in the small bowel [2]. It may be subdivided into obscure-overt with macroscopic blood loss and obscure-occult, presenting as iron deficiency anemia [3].

Capsule endoscopy is a relatively new diagnostic method that enables the complete examination of small bowel through the transmission of video images. Several studies have shown that capsule examination may detect a bleeding site in approximately 45 to 66% of patients with obscure gastrointestinal bleeding [4]. Moreover, it has been shown to be superior to small-bowel follow-through [5], push enteroscopy [6], and computed tomography enteroclysis [7], and equal to double-balloon enteroscopy (DBE) in detecting bleeding small intestinal lesions [8].

Based on solid literature data, most physicians recommend that CE should be the method of choice in patients with obscure gastrointestinal bleeding, following a negative gastroscopy/colonoscopy. Despite the initial negative endoscopic investigation, a couple of case series showed that CE detects bleeding lesions within the reach of conventional upper and lower GI endoscopy that have been obviously missed during the initial examination [9, 10].

The aims of our study were: (1) To determine the incidence of lesions missed by the preceding gastroscopy and colonoscopy, responsible for obscure gastrointestinal bleeding, in patients subjected to CE in our departments, and (2) To investigate whether there was any significant difference in the rate of identifying such lesions in patients that we investigated as compared to patients referred to our departments, offering an open-access CE service.

Patients and Methods

We reviewed prospectively collected databases between September 2006 and December 2008. These included a series of consecutive patients who participated in published and ongoing CE studies of two tertiary care gastroenterology departments offering open-access CE service. Overall, there were 605 consecutive CE studies. Of these, 317 patients had CE examination for occult (120 patients, 38%) or overt (197 patients, 62%) gastrointestinal bleeding of obscure origin and were considered eligible for entry into the study.

Patients with occult bleeding had recurrent iron-deficiency anemia despite treatment with iron supplementation. Patients with overt gastrointestinal bleeding had presented with recurrent episodes of melena and needed blood transfusion or iron supplementation therapy. All patients had undergone at least one upper gastrointestinal endoscopy and colonoscopy the previous 3 months to CE examination, which had failed to establish a diagnosis.

Capsule endoscopy was performed with the Given M2A video capsule system (Given Imaging Ltd, Yogneam, Israel). Patients were given a bowel preparation with a 2-l polyethylene glycol electrolyte lavage solution or phosphosoda the evening before the investigation [11], because as we have shown bowel preparation increases the diagnostic yield of capsule endoscopy [12]. The capsule was ingested at about 9:00 a.m. the next morning. At the end of the recording period, the equipment was removed and video capsule endoscopy images were evaluated by experienced personnel (N.V., K.P.). The suspected blood indicator (SBI) was used to facilitate the screening of possible sites of active bleeding. Findings were considered positive if they could explain the patient’s symptoms, help further management, or were later confirmed by other diagnostic modalities [13].

The study protocol conformed to the principles outlined in the Declaration of Helsinki. All patients gave written informed consent for the CE examination and for use of the data and images for research purposes.

Statistics

We used Fisher's exact test for categorical data. A p value less than 0.05 was considered as statistically significant. All statistics were performed using SPSS 16.0 for Windows (SPSS Inc. Chicago, IL, USA).

Results

We performed capsule endoscopy in 194 men and 123 women with OGIB (mean age 61, range 21–89 years); 143 (45.1%) were referrals, i.e., they had endoscopic investigation elsewhere, that failed to establish a diagnosis (Fig. 1). They had previously undergone a median of 2 (range 1–4) upper gastrointestinal endoscopies and two colonoscopies (range 1–3). The median (SD) time from last endoscopy to CE examination was 3 (3.3) weeks. Small-bowel follow-through had been performed in 114 (36%) and push enteroscopy in 84 (26.5%) patients. The capsule reached the cecum in 276 patients (87%). A definite or likely cause of bleeding in the small intestine was found in 215 patients (67.8%). In 11 cases (3.5%), the lesion was found outside the small bowel and within the reach of conventional upper gastrointestinal endoscopy or colonoscopy. Of these, seven lesions were found in the cecum and four lesions in the stomach or in the duodenum (Table 1). SBI feature detected only three bleeding cases (3/11, 27.3%) and two out of four (50%) actively bleeding lesions.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-011-1592-3/MediaObjects/10620_2011_1592_Fig1_HTML.gif
Fig. 1

Flowchart diagram of the study

Table 1

Lesions diagnosed at capsule endoscopy

Case no.

Diagnosis

Gender

Age (years)

Clinical presentation of bleeding

Duration of bleeding (months)

Transfusions (units)

Prior endoscopies (no.)

Time from endoscopy to CE examination (weeks)

Push endoscopy

CT scan of the abdomen

Colonic lesions

1

Colon cancer

F

65

Occult

9

0

2

3

No

Yes

2

Colon cancer

M

78

Occult

1

2

2

2

No

Yes

3

Colon cancer

M

65

Occult

4

3

3

1

Yes

Yes

4

Angiodysplasia

M

80

Occult

13

8

3

8

Yes

No

5

Angiodysplasia

M

44

Occult

8

14

4

5

Yes

Yes

6

Diverticulum

M

78

Overt

2

5

2

2

Yes

Yes

7

Crohn’s colitis

F

28

Occult

5

0

2

8

No

No

Upper GI lesions

8

Cancer of the cardia

F

70

Occult

7

0

2

3

No

Yes

9

Angiodysplasia of the duodenum

M

79

Occult

12

12

3

11

Yes

Yes

10

Angiodysplasia of the stomach

M

63

Occult

18

3

2

8

No

Yes

11

GAVE

F

82

Occult

16

14

4

3

No

Yes

A carcinoma of the colon was found in three patients, two men and one woman. In all of them, the referral colonoscopy was reported as total; however, in two of them, the cecal carcinoma was not found (Fig. 2), while in the last patient it missed an ulcerated lesion in the ileocecal valve (Fig. 3). All three patients had a barium enema examination that was negative. The SBI detected bleeding in one patient. After the CE examination they were referred for repeat colonoscopy that confirmed the diagnosis. They underwent right hemicolectomy and the prognosis was favorable in all but one who developed liver metastases 9 months later and died of metastatic disease.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-011-1592-3/MediaObjects/10620_2011_1592_Fig2_HTML.jpg
Fig. 2

Cecal cancer revealed by small-bowel wireless capsule endoscopy

https://static-content.springer.com/image/art%3A10.1007%2Fs10620-011-1592-3/MediaObjects/10620_2011_1592_Fig3_HTML.jpg
Fig. 3

Cecal cancer presented as an ulcerated lesion in the ileocecal valve (arrow)

A bleeding diverticulum of the cecum was found in a patient who had overt bleeding while previous endoscopies were not diagnostic. The colonoscopy had found the diverticulum but this was not considered the cause of bleeding as there were no signs of recent hemorrhage and no blood in the lumen. SBI was positive. Two haemostatic clips were placed in the diverticulum in a subsequent colonoscopy with short-term improvement. Two months later the bleeding relapsed and the patient was operated.

Two patients with long-standing iron-deficiency anemia had angiodysplasias of the cecum that were missed in previous colonoscopies. Both had endoscopic treatment with APC and there had been no further bleeding.

The last bleeding case from the colon was a young girl with anemia, weight loss, and abdominal pain. Initial colonoscopy was negative although it was mentioned that the bowel preparation was not optimal. The barium enema was negative but the CE revealed multiple aphthoid ulcers in the cecum. Repeat colonoscopy confirmed the presence of ulcers and histology showed colonic Crohn’s disease.

Regarding the bleeding sites from the upper gastrointestinal tract, angiodysplasias were found in two male patients with long-standing anemia. In one of them, the lesion was identified by CE in the second part of the duodenum. Two previous endoscopies and a push enteroscopy had failed to diagnose it. Colonoscopy was negative. Repeat duodenoscopy established the diagnosis and APC was performed at the same time. The patient improved with a reduction in transfusion requirements. He had a second session of endoscopic therapy with a permanent stop of bleeding. The other patient had an angiodysplasia in the body of the stomach (Fig. 4). The lesion was not considered significant in a previous endoscopy but it was actively bleeding at the time of capsule examination. It was also detected by SBI. A subsequent endoscopy identified the angiodysplasia and the patient was successfully treated with the Gold Probe.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-011-1592-3/MediaObjects/10620_2011_1592_Fig4_HTML.jpg
Fig. 4

Angiodysplasia of the stomach

An 82-year-old lady was found to bleed from gastric antral vascular ectasia (GAVE). She was suffering from scleroderma and in the previous 16 months has been intensively investigated for severe iron-deficiency anemia. She had undergone three endoscopies in different clinics without getting a diagnosis and she had received multiple transfusions. GAVE was described as “antral gastritis” in two out of three endoscopies. CE revealed active bleeding that was verified 2 days later by a repeat endoscopy. The patient had four treatment sessions of APC with gradual restoration of hemoglobin levels and no need for further transfusions.

The last case was a 70-year-old woman with a 7-month history of iron-deficiency anemia who was referred for CE after a non-diagnostic work-up with upper gastrointestinal endoscopy and colonoscopy, performed 15 days before admission. The capsule passed with difficulty from the GE junction and revealed a ring-like tumor around the gastric cardia (Fig. 5). A repeat endoscopy confirmed the diagnosis. The histology of the lesion was consistent with adenocarcinoma of the cardia. A CT scan demonstrated multiple secondary liver lesions. An appropriate chemotherapy was started but unfortunately the patient passed away 2 weeks after diagnosis.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-011-1592-3/MediaObjects/10620_2011_1592_Fig5_HTML.jpg
Fig. 5

A ring-like tumor of the gastric cardia

From these 11 cases, nine have been referred to us using the open-access CE service and two have been initially evaluated in our departments. There was a statistical significant difference regarding the identification of bleeding lesions within the reach of conventional upper and lower endoscopy found by CE in patients referred to us from other centers (9/143, 6.3%) as opposed to patients that we evaluated (2/174, 1.15%), (p = 0.026).

Discussion

In the present study, we reviewed our databases of 317 patients who had capsule endoscopy for the investigation of OGIB. Capsule endoscopy identified several lesions within the reach of conventional upper and lower endoscopies that could explain the GI blood loss in 3.5% of our study population. The rate of finding such lesions was significantly higher in patients referred to us for capsule endoscopy than that of patients investigated in our institutions.

Management of patients with gastrointestinal bleeding of obscure origin is a challenging problem in clinical practice. When a bleeding source is not identified by upper and lower GI endoscopy, small-bowel CE is the preferred next step of investigation [14]. Indeed, the diagnostic yield of CE has been demonstrated in a number of comparative studies, which show that capsule endoscopy is superior to barium follow-through (31 vs. 5%) [5], push enteroscopy (50 vs. 24%) [6], computed tomography enteroclysis (59 vs. 36%) [7], intraoperative enteroscopy (74.4 vs. 68%) [15] and angiography (72 vs. 56%) [16]. Furthermore, CE and double-balloon enteroscopy have a similar diagnostic yield in detecting small-bowel lesions (55.3 vs. 60.5%) [8] and therefore CE is recommended as a first-line investigation over DBE, in view of its non-invasiveness, length of the small bowel examined, the quality of the examination, and safety [17].

In accordance with previous studies [18, 19], CE revealed a definite or likely cause of bleeding from the small intestine in 67.8% of our study population. Interestingly though, in 3.5% of our patients the bleeding site was found in the stomach or colon, despite the fact that all of our patients have been previously subjected, at least once, to upper and lower GI endoscopy. What is the reason that these lesions, revealed by CE, were missed in the endoscopic examination? Possible explanations for overlooking a lesion include the small size or the unusual location of the lesion, such as vascular lesions or Cameron ulcers that preclude their proper visualization. Meanwhile, if a lesion is not actively bleeding at the time of the endoscopic examination, it might not be evident or considered not to be a significant one. This was the case in one of our patients where an angiodysplasia was thought to be “innocent”, while in a second patient GAVE was misinterpreted as gastritis. Furthermore, in patients with significant anemia, a lesion may look less obvious and may be missed, whereas it could be argued that since during CE there is no need for insufflation of air, this could make some lesions more prominent. Moreover, in several cases, busy endoscopists perform a very quick EGD, not even taking enough time for visualizing properly the stomach in retroflexion. Regarding colonoscopy, sometimes endoscopists think they have reached the cecum, although unfortunately this is not always so. We suppose that this was actually the reason for missing the lesions in the cecum (two patients) and in the ileocecal valve (one patient).

In our study, we used the suspected blood indicator (SBI) to facilitate the detection of bleeding lesions by capsule endoscopy. The SBI function identifies red-colored pixels on the screen and creates a red mark, which indicates to the reader that a potential bleeding lesion may be present [20]. However, in accordance with previous studies, the SBI-based identification of bleeding lesions was suboptimal and of limited clinical value, even in patients with active bleeding [21].

Our study has shown that there was a statistically significant difference regarding the identification of bleeding lesions outside of the small bowel in patients referred to us for CE from other physicians/centers as opposed to patients that we evaluated endoscopically ourselves. Both of our institutions participating in this study have implemented the open-access CE service; i.e., patients are subjected to the test when their physicians ask for it, without us being involved in the process of evaluating the patient or judging the reason for the referral. Although it is possible that any operator may miss a significant lesion, it seems logical to assume that this may be more frequent if the endoscopist who performed the initial endoscopy either used suboptimal equipment or is less experienced.

A preferred algorithm for further investigating patients with gastrointestinal bleeding of obscure origin when a bleeding source is not identified in upper or lower GI endoscopy has not been established so far. AGA guidelines recommend repeating EGD or colonoscopy if there is a suspicion of an overlooked lesion [1]. Does this mean that patients referred to tertiary centers for CE should have a routine second-look upper and lower endoscopy before proceeding to further investigations? Although the results of our study indicate that in such cases a lesion in the stomach or large intestine might have been missed more easily, such an approach would increase the cost of the investigation enormously. According to our estimations, the cost of re-endoscoping in our institution all CE referrals is €50,050 (143 patients ×  €350) to avoid nine unnecessary CE examinations (9 patients × €600 = €5,400). Furthermore, overall, the percentage of missed bleeding lesions from the stomach or large intestine identified by CE is relatively small (3.5%), suggesting that routinely repeating the upper and lower GI endoscopy would improperly delay the diagnosis. This is supported by the findings of another study where 50 patients who were referred for the investigation of obscure gastrointestinal bleeding underwent repeat upper GI endoscopy and colonoscopy before capsule endoscopy examination [22]. A probable cause of bleeding was found at gastroscopy in only two patients while colonoscopy did not reveal a source of bleeding in any patient. On the other hand, capsule endoscopy identified a probable bleeding source in 51% of patients and a possible source in another 11%. Therefore, since in the great majority of these patients the bleeding site is within the small intestine, we believe that CE should be preferred, given the fact that it will diagnose both lesions arising from the small intestine as well as lesions missed by conventional gastroscopy/colonoscopy. Repeated endoscopic examination could probably be worthwhile in cases of severe, continued overt bleeding and initial negative findings on upper and lower endoscopy.

The retrospective analysis of a prospectively collected database containing information of all patients undergoing CE in our centers is one limitation of this study. It should also be noted that among the negative CE cases, there might be cases with bleeding lesions outside the small bowel that were also missed.

In conclusion, we have shown that CE revealed definite causes of gastrointestinal bleeding that were missed in previous endoscopic examination in 3.5% of patients with OGIB. Careful review of the capsule images of the stomach and colon might be useful in the detection of lesions missed on prior endoscopies. Early use of CE, instead of repeating endoscopy, would allow more rapid diagnosis and less cost associated with obscure bleeding but this must be further investigated in outcome studies.

Conflict of interest

There are no conflicts of interest in this study.

Copyright information

© Springer Science+Business Media, LLC 2011