The association of body mass index with Barrett's oesophagus

Obesity has been linked to gastro-oesophageal reflux disease symptoms and oesophageal adenocarcinoma; however, there is no published evidence for an association with Barrett's oesophagus.

To investigate the association between obesity and Barrett's oesophagus.

We conducted a retrospective cross-sectional study of patients who underwent upper endoscopy at the Southern Arizona Veteran's Affairs Healthcare System between 1998 and 2004. We examined male patients without malignancy, with available information on weight and height. Based on endoscopic and histological findings, patients were classified as cases with Barrett's oesophagus or non-cases without Barrett's oesophagus. Multivariable logistic regression analysis was conducted to examine the association of body mass index and obesity with Barrett's oesophagus and Barrett's oesophagus length while adjusting for age and race.

There were 65 cases with Barrett's oesophagus and 385 non-cases without Barrett's oesophagus. The mean body mass index was significantly higher in cases than in non-cases (29.8 vs. 28.0, P = 0.03). Cases had significantly greater mean weight than controls (206 lb vs. 190,P = 0.005). The proportions of cases with body mass index 25-30 and body mass index > or =30 were greater than those in non-cases (44.6% vs. 37.7%) and (40.0% vs. 33.5%), respectively (P = 0.08). In the multivariable logistic regression model adjusting for race and age, when compared with body mass index < 25, the odds ratio was 2.43 (95% confidence interval: 1.12-5.31) for body mass index 25-30 and 2.46 (1.11-5.44) for body mass index > or =30. When examined as a continuous variable the adjusted odd ratio for each five-point increase in body mass index was 1.35 (95% confidence interval: 1.06-1.71, P = 0.01). The association between weight and Barrett's oesophagus was also statistically significant (adjusted odd ratio for each 10 pound increase = 1.10, 1.03-1.17, P =0.002). Among the 65 cases of Barrett's oesophagus, there was no correlation between the length of Barrett's oesophagus at the time of diagnosis and the body mass index (correlation coefficient = 0.03, P = 0.79).

This retrospective cross-sectional study in male veterans shows that overweight is associated with a two-and-half-fold increased risk of Barrett's oesophagus. Larger studies of the underlying mechanism are warranted to better understand how and why obese patients are at greater risk for Barrett's oesophagus.