Is there any influence of diet on the development of Barrett's metaplasia?
G.W.B. Clark, D.J. Bowrey, M.K. Roy, P.D. Carey (Cardiff)
The increasing incidence of esophageal adenocarcinoma in the Western hemisphere throughout the last two decades  has focused the attention of investigators towards improved understanding of the etiological agents. To date the only unequivocal risk factor is Barrett's esophagus (BE), a condition which represents an abnormal metaplastic healing which occurs in response to chronic gastroesophageal reflux disease. The diagnosis of BE is surrounded by controversy but most would agree that the histological hallmark of this premalignant lesion is the presence of specialized intestinal epithelium in biopsies obtained from the lower esophagus. Based on this definition between 75-100% of esophageal adenocarcinomas are Barrett's carcinomas .
An association between dietary factors and the development of esophageal squamous carcinoma has been recognized for some time. It is only recently that studies have addressed the relationship of diet and esophageal adenocarcinoma. There are no reported studies on the role of dietary factors in the genesis of BE in humans. A high intake of dietary fat, particularly saturates, associated with a low intake of fruit and vegetables has been correlated with an increased risk of gastrointestinal tract cancer, most notably adenocarcinoma of the colon and pancreas [3-5]. In this setting the following studies have attempted to evaluate whether there is a relationship between dietary factors and the development of esophageal glandular neoplasia.
In a detailed study the role of dietary fat in the induction of both BE and esophageal cancer was evaluated in rats which were operated to induce profound reflux of gastric and duodenal juice .
One hundred and thirty five male eight week old rats were entered into one of nine groups shown in Table I. Fifteen rats served as non operated controls and 120 underwent the operation of end to side esophagoduodenostomy with gastric preservation to induce profound reflux of gastroduodenal juice into the lower esophagus. Carcinogen was administered and the tumor promoting effect of different diets was evaluated.
An esophagoduodenal anastomosis was performed in a sterile field through an upper midline incision. The gastroesophageal junction was ligated flush with the stomach. The distal esophagus was transected proximal to the ligature and anastomosed to a 4 mm transverse enterotomy on the antimesenteric border of the duodenum 1 cm distal to the pylorus.
At 10 weeks of age the rats were transferred to a chemical fume hood. Rats were injected weekly for 4 weeks with methyl-n-amylnitrosamine (MNAN) via the intraperitoneal route, at a dose of 25 mg/kg (one-fifth of the LD50).
Rats were initially fed commercial chow diet which was replaced by respective semipurified diets at 14 weeks of age. The diets were fed for 20 weeks until sacrifice. The induction period (carcinogen treatment) was therefore separated from the promotion phase (diet) avoiding the possibility that the diets may have affected the metabolism of MNAN. The semipurified high fat diet contained 24.6% corn oil, the semipurified control diet contained 5% corn oil, and the semipurified calorie restricted diet contained 3.8% corn oil (Table II). The high fat diet provided 420 Kcal/100 g and the control diet 380 Kcal/100 g. The calorie restricted diet was formulated so that the rats received 40% fewer calories (with calories removed from fat and carbohydrate) than rats fed the control diet, but had the same daily intake of protein, fibre, vitamins and minerals.
Sacrifice and handling of specimens
Rats were sacrificed at 34 weeks of age. The presence of mucosal abnormality in the esophagus and the number of papillomas and macroscopic tumors was recorded. Two strips of the length of the esophagus, each with a segment of duodenum, were taken from each rat which included the macroscopic tumors. Sections were stained with H & E and, when tumors were present, with mucicarmine.
Barrett's esophagus was identified when specialized columnar epithelium with goblet cells extended at least 5 mm above the site of the surgical anastomosis (the site of anastomosis was evident in sections by the presence of the non absorbable blue suture material). Columnar cells lining the esophagus were not identified in any of the control rats, where the squamous epithelium of the esophagus is continuous with that of the squamous forestomach, therefore the presence of intestinal metaplasia lining any length of the esophagus was considered abnormal. However, as the specialized intestinal epithelium lining the distal esophagus was continuous with that of the duodenum it was necessary to select a minimum distance of 5 mm of length to be confident that this mucosa was lining the esophagus and not distorted duodenal mucosa at the site of anastomosis. The diagnosis of a columnar lined esophagus could therefore also made when squamous epithelium was present distal to the location of the area of columnar change even when it measured less than 5 mm in length.
The presence of benign pedunculated squamous cell papillomas was recorded. Carcinomas were diagnosed as either squamous cell carcinomas or adenocarcinomas. All adenocarcinomas were confirmed when the glandular elements stained magenta with mucicarmine.
Table III shows the pertinent study findings. Overall, the prevalence of BE was 13% (14/111). Dietary fat content did not influence the incidence of a columnar lined esophagus. Diets with a high fat content were associated with an increased incidence of esophageal carcinoma both adenocarcinoma and squamous carcinoma.
In a retrospective review of risk factors in 255 patients with esophageal cancer we reported an association between adenocarcinoma and increased body mass index . Patients were divided up into those with esophageal adenocarcinomas (62/255, 24%), those with adenocarcinomas of the cardia (79/255, 31%) and those with squamous carcinoma (114/255, 45%). Patients with both esophageal adenocarcinomas and cardia adeno-carcinomas had higher body masses indexes (BMIs) (29.2 ± 4.1 and 29.0 ± 6.4 respectively) compared to patients with squamous cell carcinoma (25.2 ± 5.3), p < 0.01. Further, these high BMIs in patients with distal esophageal adenocarcinoma are above the 75 th percentile for US white males. Of note BE was found significantly more often in patients with esophageal adenocarcinoma as compared to patients with cardia adenocarcinoma.
In a population based case control study of obesity and dietary factors adenocarcinoma of the esophagus there were similar findings of significantly elevated BMI in cancer patients compared to age and sex matched controls . A BMI above the 75% conferred an increased risk of esophageal cancer of 3.1 (CI 1.8-5.3). No significant association was found with total calories from food or level of fat intake but a significant reduction in risk was found in those consuming higher proportions of fresh vegetables and raw fruit.
There are only a few reports relating diet to the development of esophageal adenocarcinoma where there appears to be an association between high body mass index and a significantly increased risk. Whether this risk relates directly to fat intake or is a reflection of the predisposition of obese subjects to develop gastric distention, hiatus hernia and gastro-esophageal reflux remains to be determined. Experimentally, in rats, the high fat content of the diet was significantly associated with an increased predisposition to develop esophageal cancers. There is no published data to support a dietary association with Barrett's esophagus per say although it would be expected that any factors which predispose to the development of gastroesophageal reflux would be associated with an increased prevalence of Barrett's esophagus.
6. Clark GWB, Smyrk TC, Mirvish SS, et al. The effect of gastroduodenal juice and dietary fat on the development of Barrett's esophagus and esophageal neoplasia: an experimental rat model. Ann Surg Oncol 1994;1:252-261.