Diagnostic yield of capsule endoscopy in patients with obscure gastrointestinal bleeding or unexplained anemia – A monocenter experience with 1006 consecutive examinations

Introduction: The most important indication for capsule endoscopy (CE) is the evaluation of patients with suspected mid-gastrointestinal (GI) bleeding. Data concerning the diagnostic yield of CE differ significantly between the studies published so far.

Methods: A total of 1006 consecutive CEs performed in 929 patients at our centre between 2002 and 2011 were analysed. All CEs carried out because of obscure GI bleeding or unexplained anemia were included in this study. For the examinations the systems Given Imaging® (Yoqneam, Israel, n=943) and MiroCam® (n=63) were used. Lesions detected during CE were classified as having high (P2), intermediate (P1) or no (P0) risk for bleeding.

Results: The most frequent indication for CE was obscure GI bleeding (672/1006, 66.8%), comprising 355 (52.8%) cases of occult, 281 (41.8%) cases of overt and 36 (5.4%) cases of not defined GI bleeding; 418 (62.2%) patients with obscure GI bleeding had transfusion dependent anemia. Anemia without signs of GI bleeding was the referring diagnosis in 170 patients (16.9%). Thus, a total of 842 CE examinations (m: f=421: 421; mean age±SD: 62.0±17.1 years; range: 16–93) were included in this study.

In 322 patients with obscure GI bleeding either no lesion (46.9%, n=315) or a P0 lesion (1.0%, n=7) was detected. P1 lesions were found in 101 patients (15.0%), and P2 lesions in 249 patients (37.1%). In patients with obscure-occult bleeding P0 lesions were detected in 5 patients (1.4%), P1 lesions in 56 patients (15.8%) and P2 lesions in 133 patients (37.5%). The corresponding numbers in patients with obscure-overt GI bleeding were: P0 lesions in 2 patients (0.7%), P1 lesions in 41 patients (14.6%) and P2 lesions in 99 patients (35.2%). The percentages of P0, P1 and P2 lesions did not significantly differ between patients with obscure-occult and obscure-overt bleeding.

In the group of patients with anemia without signs of bleeding the percentage of P2 lesions was significantly lower as compared to patients who underwent CE because of obscure bleeding (22.9% vs. 37.1%, p-value 0.00053) and the percentage of examinations with no detected lesions was significantly higher (62.9% vs. 46.9%, p-value 0.00018).

The most frequent P2 lesions were: Angiodysplasia (n=183), acute bleeding without visible bleeding source (n=36), polyps/polypoid lesions (n=28), ulcer (n=20); submucosal lesions/tumors (n=19).

Discussion: In patients who underwent CE because of obscure GI bleeding P2 lesions were detected in 37.1% and P1 lesions in 15.0%. The diagnostic yield did not differ significantly between patients with obscure-overt and obscure-occult bleeding, but was significantly lower in patients with anemia without signs of bleeding.