Can patient characteristics predict the outcome of endoscopic evaluation of iron deficiency anemia: A multiple logistic regression analysis

Background The purpose of this study was to identify clinical and biochemical variables that predict the outcome of upper/lower endoscopy in outpatients with iron deficiency anemia and to determine which endoscopic procedure should be performed first. Methods Ninety-eight patients (74 women, 24 men; mean age 55 years) with iron deficiency anemia referred from the hematology department were interviewed and responded to a questionnaire that included clinical and biochemical variables, and underwent EGD (with biopsies) and colonoscopy. The endoscopic findings were recorded as presence/absence of GI cancer, upper/lower GI tract lesions and bleeding/non-bleeding-associated GI lesions. A multiple logistic regression analysis was applied to identify variables significantly related with the outcome of the investigations. Multiple analyses were performed so that a Bonferroni correction for multiple testing removed significance except where p<0.01. Results A likely cause of iron deficiency anemia was found in 86.7% of patients. The risk factors for GI malignancies were: male gender (OR 7.5: 95% CI[1.7, 31.9]; p<0.01), advanced age (OR 1.1/y: 95% CI[1, 1.2]; p<0.01), and lower mean corpuscular volume (OR 1.1/unit: 95% CI[1, 1.2]; p<0.002). The risk factors for bleeding-related diseases were the following: greater age (OR 1.1/y: 95% CI[1.1, 1.2]; p<0.001), absence of lower-GI tract symptoms (OR 4.7: 95% CI[1.3, 16.6]; p<0.05), and a positive fecal occult blood test (OR 4.1: 95% CI[1.2, 14.3]; p<0.05). The risk factors for non-bleeding-related GI tract diseases were the following: negative fecal occult blood test (OR 4.5: 95% CI[1.16, 20]; p<0.05) and higher Hb level (OR 1.4/unit: 95% CI[1.1, 1.8]; p