Association between obesity and Barrett's esophagus in a Japanese population: a hospital-based, cross-sectional study

Background: The association between obesity and Barrett's esophagus (BE) in the Japanese population remains unclear. The prevalence of BE and its associated risk factors was examined. Methods: A cross-sectional study of 1581 consecutive individuals who underwent upper gastrointestinal endoscopy was conducted. The prevalence of endoscopically suspected BE (ESBE) was evaluated. Obesity was evaluated by body mass index (BMI, >= 25 kg/m(2)) and waist circumference (WC) (males, >= 85 cm; females, >= 90 cm). Because endoscopic diagnosis of ultra-short ESBE (< 1 cm in extent) is difficult and highly unreliable, this type of ESBE was excluded from the study. Results: In proton pump inhibitor (PPI) non-users, the prevalence of ESBE >= 1 cm was 5.6%. In univariate analysis, male sex and reflux esophagitis (RE) were significantly associated with BE, but BMI, WC, and reflux symptoms were not. In multivariate logistic regression analysis, only RE (odds ratio [OR] = 3.48, 95% confidence interval [CI] 1.89-6.41, p < 0.0001) was an independent risk factor for BE; obesity and the other factors were not. In contrast, RE (OR 5.67, p = 0.0004) and large WC (OR 5.09, p = 0.0005) were significant risk factors for ESBE >= 1 cm in PPI users. Only male sex, but not obesity or the other risk factors, was associated with an increased risk of RE in patients not taking PPIs. Conclusions: RE, but not obesity, may have an independent association with the risk of ESBE in the Japanese population. Furthermore, obesity measures were not independent risks for RE. Interestingly, PPI-refractory RE and large WC were risk factors for ESBE >= 1 cm in patients taking PPIs.