PTU-033 The Utility Of Narrow Band Imaging Endoscopy In Identifying Potential Causes Of Iron Deficiency Anaemia In The Absence Of Any Overt Gi Cause

Introduction IDA is prevalent in up to 5% of the developed world and endoscopy remains the most utilised investigation. Magnified white light endoscopy has been shown to accurately identify gastric atrophy, H. pylori gastritis and coeliac disease but the role of magnified NBI endoscopy (NBI-Z) in this context has not been evaluated. The study aim was to assess the ability of NBI-Z to make a real time diagnosis of these conditions compared to histology as the gold standard. Methods This prospective cohort study recruited patients undergoing endoscopic evaluation for IDA. All procedures were performed with an Olympus video endoscopy system by clinicians with advanced imaging experience. Systematic NBI-Z imaging in parts of the duodenum and gastric mucosa were taken with corresponding biopsies. A previously validated Nottingham Type 1–4 classification system was used to classify the characteristic gastric mucosal pit pattern, and magnified morphological features were used to describe intestinal metaplasia and villous atrophy. This allowed for a real time diagnosis to be made for villous atrophy, gastric atrophy and H. pylori gastritis. The specimens were examined by a single blinded GI pathologist. Results 105 patients were recruited over 3 years. Excluding those with an obvious cause (n = 11), a total of 94 patients were included in the final analysis. Female: male ratio was 1: 0.7, median age 66 years (range 21–85). 38% had significant co-morbidities. At time of endoscopy 52% were taking iron therapy, 19% aspirin, 30% PPI and 4% NSAIDs. The median (range) anaemia parameters were: Hb 10.8g/dL (7.7–12.6), MCV 82fl (60–97), Ferritin 10g/L (1–379) and iron 7.5 µmol/L (1–22). 73% had the procedure under sedation with median doses of 2.5 mg midazolam and 25 mg pethidine. Abstract PTU-033 Table 1 NBI performance (%) compared to histology with 95% confidence intervals Sensitivity Specificity Positive predictive value Negative predictive value Type 1 predicting normal 100 (81.4–100) 24.2 (18.8–24.2) 27.7 (22.5–27.7) 100 (77.7–100) Type 2/3 predicting H. pylori gastritis 100 (76.9–100) 28.8 (22.6–28.8) 27.5 (21.1–27.5) 100 (78.4–100) Type 4 predicting atrophy 92.9 (72.7–99) 93.3 (72.7–99.4) 92.9 (72.7–99.4) 93.3 (74.5–94.4) Predicting Intestinal Metaplasia 50 (15.8–65.8) 98.1 (94.4–99.9) 75 (23.7–98.7) 94.6 (91–96.3) Predicting Villous Atrophy 66.7 (13.6–97) 98.7 (96.6–97) 66.7 (13.6–97) 98.7 (96.6–99.9) Conclusion In patients with IDA, NBI-Z is highly specific in pr