Background: Body mass index is associated with the risk of Barrett's oesophagus (BO). It is uncertain whether this is related to total body fat or other factors that correlate with body mass index. We aimed to quantify the association between total body fat (measured by bioelectrical impedance) and risk of BO and examine if this association was modified by gastro-oesophageal reflux (GOR) and abdominal obesity. Methods: In 2007-2009, we surveyed 235 cases (69% Males, Mean age 62.1 years) and 244 age and sex matched population controls from a population based case-control study of BO. We conducted structured interviews, standard anthropometry and bioimpedance analysis of total body fat. Odds ratios (OR) and 95% confidence intervals (CI) were estimated using multivariable logistic regression analysis. Results: There was a significantly increased risk of BO among those in the highest tertile of total body fat weight (OR 2.01; 95%CI 1.26-3.21) and total body fat percentage (OR 1.86; 95%CI 1.10-3.15). These risks were largely attenuated after adjustment for GOR and waist circumference. There was a significantly increased risk of BO among those in the highest tertile of waist circumference (OR 2.21; 95%CI 1.39-3.51) and this was minimally attenuated after adjustment for total body fat and moderately attenuated after adjustment for GOR. Conclusions: Total body fat is associated with an increased risk of BO but this appears to be mediated via both abdominal obesity and GOR. These findings provide evidence that abdominal obesity is more important than total body fat in the development of BO.