PE involves insertion of an endoscope through the oral cavity into the jejunum. PE is a readily available, safe, and effective technique for detecting and treating diseases of the proximal gut. Complications are rare if PE is performed without an overtube. PE with an overtube is generally performed only when a moderate increase in depth of insertion into the small bowel is required. When CE is not available, PE is a reasonable, low-risk option, but produces it produces only a moderate diagnostic yield.6 For example, in a study of 63 patients, after exclusion of all lesions proximal to the ligament of Treitz, the diagnostic yield for PE was 41% in patients with recurrent overt OGIB, 33% in those with persistent overt OGIB, and 26% in those with occult OGIB.7

The benefits of CE include visualization of the entire small bowel, safety, noninvasiveness, and high diagnostic yield. However, CE has some limitations such as the fact that no biopsies are taken to accompany the test, it is difficult to accurately locate the source of the bleeding, and there is a risk of capsule retention.3 CE also cannot be used for therapeutic intervention. Capsule retention has been reported in 1.5% to 5% of patients with suspected Crohn disease and OGIB. CE should be used with caution in patients with known or suspected GI obstruction, fistulas, or motility disorders.3,8

Compared with CE, DBE is more invasive, requires sedation, and can be laborious. It also takes time to learn how to perform DBE. Complications include the potential for pancreatitis and perforation of the small bowel and ileus.9,10,11 However, the major advantage of DBE is that it can be used for therapeutic interventions such as endoscopic hemostasis of bleeding, obtaining tissue biopsies for histological analysis, and marking the location of disease to direct subsequent surgery.12

There are few comparative studies or meta-analyses and no prospective, randomized, controlled trials comparing CE and DBE specifically in OGIB. Arakawa et al.13 reported that the overall diagnostic yield did not differ significantly between DBE (64%) and CE (54%). A meta-analysis also reported a similar diagnostic yield for CE (61.7%) and DBE (55.5%).14 A meta-analysis of 20 prospective studies comparing CE with other diagnostic modalities, such as balloon or spiral-assisted enteroscopy, showed that the diagnostic yield for CE was 56% compared with 26% for PE and 6% for small-bowel follow-through.15 In a recently reported meta- and pooled analysis of 12 eligible studies that included 712 patients with OGIB, the overall diagnostic yields of CE and DBE were similar. In subanalyses, the diagnostic yields of CE and DBE differed significantly for certain causes of OGIB such as blood clots (CE 21.8% vs. DBE 3.3%; p<0.00001) and diverticulum (CE 0.6% vs. DBE 3.97%; p=0.02). Of 205 cases of OGIB, 148 (72.2%) were detected with CE but not DBE and 57 (27.8%) were detected with DBE but not CE.16 Thus, both modalities are important for the detection of OGIB, and their combined use is better than either modality alone.

For patients with OGIB, the diagnostic yield of traditional radiological examination modalities such as small-bowel follow-through or enteroclysis is relatively low (6% to 10%).17 Computed tomography (CT) enterography is a noninvasive technique that visualizes the extravasation of contrast medium into the intestinal lumen to identify the source of OGIB.18 In a study of 26 patients with massive GI bleeding, multidetector CT accurately diagnosed 89%, with a positive predictive value of 95%. The location of the actively bleeding lesions in these patients corresponded exactly to sites revealed by angiograms.19 Of these patients, 56% had small-bowel tumors. CT enterography is more appropriate than multidetector CT when GI obstruction is suspected, as in cases of small-bowel tumors.20 Therefore, CT enterography should be the modality of choice when neoplastic disease is suspected but CE findings are negative.

Angiography is useful for the evaluation of overt OGIB. The diagnostic yield of angiography for lower GI bleeding has been reported to be 27% to 77%, but there are limited data on the diagnostic yield of angiography in OGIB.21 Vascular lesions, mainly small-bowel angiodysplasia, have the highest rates of rebleeding, despite endoscopic therapy, and are associated with comorbid conditions.22

The diagnostic yield of CE allowed the identification of bleeding lesions in 67% of patients with overt OGIB, which was higher than in previous studies.23 The diagnostic yield of CE is improved when it is performed early after the bleeding, varying from 44.2% to 92.3% in patients with OGIB.24 When DBE was performed during active bleeding, the diagnostic yield was 83% to 100%,25,26 which was significantly higher than the 48% to 58% noted when the examination was performed after bleeding ceased. "Emergency" DBE can be performed within 24 hours of onset of overt OGIB.27 With DBE, it is possible to diagnose and treat the lesion simultaneously. However, 45% of the patients were given a nonendoscopic treatment such as surgery, radiology, or conservative medical therapy.23 DBE is an invasive and time-consuming procedure that often requires general anesthesia. In addition, it is unlikely to be feasible in most centers.

Therefore, the Korean Gut Image Study Group and the American Society for Gastrointestinal Endoscopy (ASGE) have recommended CE or tagged red blood cell scintigraphy for patients with massive upper bleeding and negative esophagogastroduodenoscopy (EGD) findings (Figs. 1, 2).18 CE permits clinicians to choose the appropriate therapeutic option for patients. For patients with overt inactive OGIB, the ASGE guidelines recommend CE, deep enteroscopy, PE, and/or colonoscopy.18

Repeated endoscopy (including EGD) and colonoscopy are recommended for patients with occult OGIB. A retrospective review of prospectively collected data on the use of CE showed that 9 of 140 patients with OGIB had lesions that could have been evaluated by using conventional endoscopy or colonoscopy.28 Generally, CE detects bleeding lesions that could have been evaluated by using conventional endoscopy in 3% to 17% of cases and that could have been evaluated by using EGD and colonoscopy in 2% to 4% of cases. Repeated endoscopy should be considered when the initial examination was suboptimal or when there is reason to suspect that lesions have gone undetected.29

CE is recommended as the first diagnostic test if no contraindications exist (Figs. 1, 3). If a lesion is detected, appropriate endoscopic, angiographic, medical, or surgical intervention should be performed. If CE findings are negative, the patient's clinical status should be considered.18 Stable patients may be observed without further testing. In patients with negative CE findings who were reevaluated within a mean of 24 months, the rebleeding rate was 16.4%, which was significantly lower than that of patients with positive CE findings. However, for patients with negative CE findings who are taking anticoagulation medicine, close observation is required and alternative modalities should be considered.4,30 There are currently no clear indications about which alternative technique should be used or the appropriate timing for additional testing. A significant increase in the diagnostic yield of CE was reported on repeating the procedure in patients who exhibited decreased hemoglobin of at least 4 g/dL or in patients converting from occult to overt bleeding.31 Another option is to proceed with DBE instead of repeating CE. DBE may detect the source of bleeding in 30% of OGIB patients after CE yielding negative findings.14 Thus, patients with ongoing or recurrent overt bleeding, or patients with occult bleeding who experience significantly reduced hemoglobin levels, should proceed with either repeat CE or with DBE after an initial CE with negative findings.3