Background: GI evaluation with bidirectional endoscopy is recommended in patients with iron deficiency anemia (IDA). In practice, however, 2 other settings are anemic patients with some clinical clues and patients who have only anemia. Diagnostic yield of bidirectional endoscopy in the latter 2 settings are unknown.

Materials and Methods: All patient who underwent bidirectional endoscopy for anemia during 2011-2016 were reviewed. Patient were divided into 3 groups (group A, definite IDA; group B, anemia with some clues; group C, anemia without definite proof of IDA or any clinical clue). Prevalence of significant lesions and details were analyzed.

Results: Three hundred twenty five patients were enrolled (43 in group A, 95 in group B and 187 in group C). Significant GI lesions were found in 62.8%, 32.6% and 24.1% and cancers were found in 16.3%, 10.5% and 2.7%, respectively. From EGD, 39.5% of patients in group A, 18.9% of group B, and 15.5% of group C had significant GI lesions and the most common lesion was erosive gastroduodenitis. From colonoscopy, 14% of group A, 7.4% of group B, and 7.0% of group C had significant GI lesions and the most common lesion was colonic carcinoma. Dual lesions were found in 9.3%, 6.3%, and 1.6% of group A, B, and C, respectively. Multivariate analyses showed no predictor for significant GI lesions in group B, but reveled hemoglobin < 9 g/dL to be significant predictor in group C (odds ratio 6.07, 95% confidence interval 1.1-33.9, p = 0.04)

Conclusion: Significant GI lesions detected by bidirectional endoscopy in patients with definite IDA, anemia with some clinical clues of GI blood loss, and unconfirmed IDA without any clinical clue of GI blood loss were 63%, 33% and 24%, respectively. Erosive gastroduodenitis and colonic carcinoma were the most common significant upper and lower GI lesions, respectively. Patients with Hb <9 g/dL predicted significant lesions in anemia patients without IDA confirmation.