Diagnostic accuracy of gastrin-17 in relation to severe endoscopy feature of dyspepsia patients

Introduction: Screening populations using endoscopy is impractical and expensive.We need a noninvasive method to choose the patient that really need endoscopy.We evaluate the accuracy of gastrin-17 as a tool of screening patients for endoscopy. Method: Endoscopy finding was classified in 2 category, severe and mild/normal, severe if we found ulcer or tumor in gastric and mild if we found normal, superficial and erosive gastritis. Fasting serum gastrin-17 was determined by standard immuno assays. Receiving operating characteristic (ROC) analysis was used to determine the best cut-off for gastrin-17serum test in severe and mild/normal endoscopic feature. Results: Seventy seven patients underwent endoscopy. Seventy one patients with normal/mild finding and 6 patients with severe finding. Base on non parametric test with Mann Whitney test, we found significant mean different of gastrin-17 between mild/normal and severe group (p = 0.025). Diagnostic accuracy of Gastrin-17 on determining severe finding base on ROC procedure, we found AUC 78% (95% CI: 63%–91%), p = 0.025, with sensitivity and sensitivity are 66.7% and 77.5% at value ? 21.75 pg/ml. Conclusion: In dyspepsia patient, Gastrin-17 has anacceptable accuracy in determining severe abnormality on endoscopy and value ? 21.75 pg/ml is the best cut off value for screening severe endoscopic feature. Key Word(s): 1. Gastrin-17; 2. endoscopy feature