Clinical and demographic predictors of Barrett's esophagus among patients with gastroesophageal reflux disease: a multivariable analysis in veterans.

BACKGROUND: The subgroup of patients with gastroesophageal reflux disease (GERD) that should undergo endoscopy to rule out Barrett's esophagus (BE) has not been well defined. GOALS: To examine demographic and clinical variables predictive of BE before endoscopy. STUDY: A validated GERD questionnaire was administered to 107 patients with biopsy-proven BE and to 104 patients with GERD but no BE shown by endoscopy. Frequent symptoms were defined as symptoms that occurred at least once or more each week. Severity of symptoms was rated on a scale from 1 to 4 (mild to very severe). Univariate analysis and multivariable logistic regression were performed to determine whether demographic characteristics and the duration, severity, and frequency of GERD symptoms were associated with the identification of BE.
RESULTS: Eighty-five percent of the GERD patients and 82% of the BE patients completed the questionnaire. There was no difference between the groups in terms of race, gender, or proton pump inhibitor use. The BE patients were older (median age, 64 vs. 57 years, p = 0.04). In multivariable logistic regression, an age of more than 40 years ( p = 0.008), the presence of heartburn or acid regurgitation ( p = 0.03), and heartburn more than once a week ( p = 0.007) were all independent predictors of the presence of BE. Interestingly, patients with BE were less likely to report severe GERD symptoms ( p = 0.0008) and nocturnal symptoms ( p = 0.03). Duration of symptoms, race, alcohol, and smoking history were not associated with BE.
CONCLUSIONS: Upper endoscopy should be performed in GERD patients more than 40 years of age who report heartburn once or more per week. The severity of symptoms and the presence of nocturnal symptoms are not reliable indicators of the presence of BE.