Objectives: GI blood loss is the most common cause of iron deficiency anemia (IDA) in postmenopausal women and menstrual blood loss in premenopausal women. We aimed to evaluate the diagnostic yield of endoscopy in women with IDA and to define predictive factors of a GI lesion.
Method: Clinical, biological, endoscopic, and histological data from patients with IDA were systematically collected on a computer. Multivariate analysis (logistic regression) was performed to determine whether these data were associated with a GI lesion.
Results: Between January, 1989 and June, 1999, 241 consecutive women had endoscopies for IDA (mean age = 52.3 +/- 21.8 yr). A substantial GI lesion was detected in 119 patients (49.4%). Ten patients (4%) had both upper and lower GI lesions. A source of IDA was revealed by upper endoscopy in 86 cases (35.6%) and by colonoscopy in 33 (13.7%). The most common upper lesions were peptic ulceration (42/241 [17.4%]), esophagitis (15/241 [6.2%]), and cancer (9/241 [3.7%]). Colonic cancer (15/241 [6.2%]) and polyps (10/241 [4.1%]) were the most frequent lesions detected by colonoscopy. Predictive factors (odds ratio, 95% CI) of GI lesions diagnosed by endoscopy were abdominal symptoms (8.3, 3.9-17.2), age > 50 yr (4.4, 2.1-9.2), and Hb < 9 g/dl (3, 1.5-6.1). Thirty-one women (13%) had none of these predictive factors; in this group only two lesions were identified (one esophagitis and one duodenal ulcer). The positive predictive value of these three independent predictors was 87%, and the negative predictive value was 93.5%.
Conclusion: Endoscopy revealed a source of IDA in 49.4% of cases. Three predictive factors of GI lesion were identified. Endoscopic investigation should be avoided in women without these three predictive factors. Conversely, these factors are strongly associated with a GI lesion.