Introduction Oesophageal adenocarcinoma is thought to arise from columnar metaplasia of distal oesophageal mucosa caused by gastro-oesophageal reflux. Obesity is a risk factor for this process. There is some evidence that “normal” cardia may be an acquired mucosa arising through the same pathway and predisposing to adenocarcinoma development at this site. The influence of obesity on the aetiology of cardia mucosa is unknown.
Methods 62 H pylori negative healthy volunteers (age 18–74 years) were recruited. BMI, waist circumference and gender were recorded. MRI (Phillips 1.5T) was performed for quantification of visceral and subcutaneous fat (average of three axial planes; L2, L3 and L4). Upper GI endoscopy was performed with biopsies of the gastrooesophageal junction. Biopsies were taken in a craniocaudal direction and targeted to include enough squamous mucosa to confirm position. Intra-procedure pathological feedback was available and two to three biopsies were taken to optimise accuracy. Junctional biopsies were assessed to determine cardia length, considered measurable provided there was consecutive squamous, cardia and oxyntic mucosal types present. Non-parametric correlations were examined between BMI, waist circumference and cardia length and between fat distribution quantified by MRI and cardia length. Regression analysis (Stepwise method) incorporating age, BMI, waist circumference and MRI total fat was used to determine predictors of cardia length.
Results 37 of 62 volunteers had at least one junctional biopsy including squamous, cardia and oxyntic mucosa; median total length 6.5 mm (IQR 1.6). Median cardia mucosal length was 2.5 mm (IQR 1.5 mm). Length of cardia mucosa increased with age (R=0.457, p=0.004) and with waist circumference (R=0.466, p=0.004). A correlation was also seen with intra-abdominal fat (R=0.374, p=0.027) and total fat measured by MRI (R=0.389, p=0.021) but not with subcutaneous fat (p=0.091). There was no significant correlation with BMI. On regression analysis the independent predictors of cardia mucosa length were waist circumference (Standardised coefficient 0.342, p=0.035) and age (Standardised coefficient 0.322, p=0.046). Intestinal metaplasia at the cardia was seen in only 4 of 62 volunteers.
Conclusion These findings suggest that cardia mucosa may be acquired with increasing age through a process of distal squamous columnar metaplasia accelerated by central obesity. A possible mechanism is opening of the distal portion of the lower oesophageal sphincter and short segment acid reflux.
Competing interests None declared.