BACKGROUND/AIMS: We aimed to estimate the proportion of significant endoscopic findings and their association with dyspeptic symptoms and to evaluate the predictors for significant endoscopic findings. METHODS: Total of 3,872 subjects (58.3% men, mean age 43.6+/-9.3 years) who had undergone endoscopy were enrolled at the health promotion center. Each subject completed validated questionnaires, including data on gastrointestinal symptoms, socio-demographic history and medical history. Significant endoscopic findings were included peptic ulcer disease, reflux esophagitis, gastric cancer, Barrett's esophagus and gastro-duodenal erosions. Multiple logistic regression models were used to assess the predictors for significant endoscopic findings. RESULTS: The proportion of significant endoscopic findings was 39.1%. There was no significant difference of endoscopic findings between the dyspepsia and asymptomatic group (41.0% vs. 37.4%, p>0.05). There was no difference of the incidence of reflux esophagitis or peptic ulcer between subjects with and without dyspepsia. Peptic ulcer was more frequently present in subjects with reflux symptoms than asymptomatic subjects (12.3% vs. 9.0%, p=0.03). Male gender (odds ratio [OR], 3.91; 95% confidence interval [CI], 3.18-4.81) increased the risk for having endoscopic abnormality and having symptoms of functional dyspepsia according to Rome III criteria (OR, 0.75; 95% CI, 0.57-0.97) significantly decreased this risk. CONCLUSIONS: About 40% of subjects with dyspepsia had abnormal endoscopic findings, and the dyspepsia symptoms may not predict the significant endoscopic findings. Diagnostic criteria of functional dyspepsia by Rome III may be useful to predict not having significant upper endoscopic findings.