Endoscopic evaluation of patients with dyspepsia: Results from the national endoscopic data repository

Background & Aims: Endoscopy is commonly performed to evaluate symptoms of dyspepsia. The aim of this study was to characterize patients who receive endoscopy for dyspepsia and measure predictors of primary endoscopic outcomes, utilizing a large national endoscopic database. Methods: The Clinical Outcomes Research Initiative (CORI) receives endoscopy reports from a network of 74 sites in the United States. Sixty-one percent of reports come from private practice settings. Patients with reflux dyspepsia and nonreflux dyspepsia were identified from January 2000 to June 2002. Patients with dysphagia and known Barrett's esophagus were excluded. Primary endoscopic outcomes included esophageal inflammation and stricture, gastric ulcer, duodenal ulcer, suspected Barrett's esophagus (greater than or equal to2cm), and suspected esophageal and gastric malignancy. The presence or absence of alarm symptoms (vomiting, weight loss, and evidence of GI blood loss) was determined. Adjusted relative risk (RR) for predicting serious outcomes was calculated in a multivariate model. Results: We received 117,497 endoscopic reports, representing 99,558 unique patients. Dyspepsia, with and without reflux symptoms, accounted for 43% of upper endoscopies. Among dyspeptic patients, 36.5% were younger than 50 years of age without alarm symptoms. Esophageal or gastric malignancy in patients with dyspepsia was associated with increasing age, male sex, Asian race, Native American race, and symptoms of weight loss and vomiting. Suspected Barrett's esophagus (greater than or equal to2cm) was associated with reflux symptoms, male sex, age, and white race. Ulcers were associated with evidence of bleeding, vomiting, male sex, black race, and Hispanic ethnicity. Conclusions: These practice-based data reveal important practice behaviors and outcomes.