The present definition of gastroesophageal reflux disease (GERD) is based on clinical criteria that are difficult to reproduce accurately. This study provides a method to develop a histologic definition of GERD based on biopsies obtained from the affected esophagus. Pathology reports from 1655 patients who had upper gastrointestinal endoscopy and biopsy according to a systematic protocol were reviewed. Biopsies were obtained from the esophagus, around the gastroesophageal junction and the stomach: proximal, body, and antrum. Patients who had oxyntocardiac±cardiac±intestinal epithelia between the squamous epithelium proximally and the proximal limit of gastric oxyntic mucosa distally were defined as having a squamo-oxyntic gap. The length of the squamo-oxyntic gap varied from less than 1 cm in 1399 (84.5%) patients to greater than 5 cm in 80 (4.8%) of the patients. Only oxyntocardiac epithelium was seen in 190 (11.5%) of the patients, oxyntocardiac and cardiac epithelia in 898 (54.3%), and intestinal metaplasia in addition to the other 2 epithelial types in 567 (34.2%). The prevalence of intestinal metaplasia was directly proportional to length of the squamo-oxyntic gap, being 24.3% (340/1399) when the length was <1 cm, and 83.5% (147/176) with length 1 to 5 cm. All patients with a length more than 5 cm had intestinal metaplasia. The distribution of the 3 epithelia was constant irrespective of the length of the squamocolumnar gap; intestinal metaplasia, when present, was seen maximally in the proximal region of the gap, cardiac epithelium intermediate and oxyntocardiac epithelium in the most distal segment of the gap. The squamo-oxyntic gap started in a dilated region distal to the end of the tubular esophagus and distal to the proximal limit of the rugal folds and extended into the tubular esophagus. Distal gastric biopsies showed no evidence of significant inflammation, intestinal metaplasia or Helicobacter pylori infection in 1543 (93.2%) of the patients, indicating that the squamo-oxyntic gap was largely independent of gastric pathology. We provide evidence that the squamo-oxyntic gap is equivalent to the columnar-lined esophagus. Its presence is a specific and sensitive indicator of reflux and can be used as a cellular criterion to define GERD. The length of the squamo-oxyntic gap provides an accurate assessment of the severity of chronic GERD. The distal limit of the squamo-oxyntic gap, which is the junction between oxyntocardiac and gastric oxyntic epithelium is the true gastroesophageal junction. The presence of intestinal metaplasia within the squamo-oxyntic gap is the most accurate risk indicator for esophageal adenocarcinoma and defines Barrett esophagus.