Aspirin Protects Against Barrett's Esophagus in a Multivariate Logistic Regression Analysis

Background & Aims: Better criteria are needed to identify patients who should be screened for Barrett's esophagus (BE) to reduce overtesting and improve the cost effectiveness. There is evidence that chemopreventive agents such as nonsteroidal anti-inflammatory drugs, particularly aspirin, reduce the risk of esophageal adenocarcinoma (EAC), but little is known about their effects on BE. We analyzed characteristics of patients with BE for factors that might be used in screening and management. Methods: In this case-controlled study, we identified 434 patients with BE diagnosed at the first endoscopy (incident cases) at a single institution (1997-2010). BE cases were matched with controls on the basis of indication for endoscopy, year of endoscopy, and endoscopist. Risk factors analyzed included age, sex, body mass index, medical and social history, and medications. We performed a multivariate logistic regression analysis to identify clinical risk factors for BE. Results: In a multivariate regression model, men had a greater risk for developing BE (odds ratio, 3.2; 95% confidence interval, 2.3-4.4), whereas current aspirin users had a lower risk than nonusers (odds ratio, 0.56; 95% confidence interval, 0.39-0.80). A subset analysis, limited to patients who had endoscopies for symptoms of gastroesophageal reflux disease, yielded similar findings. No interactions were found between aspirin use and smoking or use of acid-suppressive medications. Conclusions: In a case-controlled study of 434 patients with BE, current aspirin use appeared to reduce the risk of BE; previous studies associated aspirin use with a reduced risk of EAC. Although efforts were made to minimize biases in our analysis, the possibility of residual confounding remains.