Push enteroscopy and colonoscopy in melena patients with negative esophagogastroduodenoscopy: Prospective multicenter study.

Background and study aims: Proper evaluation of patients with melena, no hematemesis, and nondiagnostic esophagogastroduodenoscopy (EGD) is poorly defined. Guidelines recommend colonoscopy, but the additional diagnostic yield is low. Owing to the high likelihood of proximal small bowel bleeding, push enteroscopy (PE) may be beneficial.

Patients and methods: We conducted a prospective, multicenter cohort study from four referral centers. Consecutive patients with melena, no hematemesis, and negative EGD results underwent PE followed by colonoscopy. For patients with culprit lesions found on PE and who were at risk of undergoing colonoscopy, colonoscopy was not performed and results were presumed to be negative. Diagnostic yields of both investigations were compared.

Results: Of 221 eligible patients who underwent EGD, 77 (34.8%) with nondiagnostic results were included in the analyses. Mean age of participants was 67.8 years and 51.9% were men. Culprit lesions were identified on PE in 27 of 77 patients (35.0%). Colonoscopy was performed in 59 patients and the source of bleeding was found in 10 patients (12.9%). Diagnostic yield of PE was significantly greater than that of colonoscopy ( P = 0.005). Combining PE and colonoscopy increased diagnostic yield to 48%, which was significantly greater than the yields of PE ( P = 0.002) or colonoscopy ( P < 0.0001) alone.

Conclusions: PE is beneficial for patients with melena and nondiagnostic EGD. It should be considered before or in combination with colonoscopy for these patients.