Endoscopic and histological correlation of the findings of mucosa of the distal esophagus in nonerosive reflux disease

BackgroundEsophagogastroduodenoscopy (EGD) is frequently requested for patients with gastroesophageal reflux symptoms. When the distal esophagus mucosa shows no erosions, questions are posed: (1) Does the esophageal biopsy contribute to confirm reflux as the cause of the symptoms? (2) Do minimum endoscopic changes have correspondence with histology?MethodsA total of 1901 patients were prospectively interviewed by a standard questionnaire to obtain a clinical diagnosis of gastroesophageal reflux disease (GERD); 150 patients (98 males; median age, 42 years) with clinical manifestations of GERD and an intact esophageal mucosa were assigned to nonerosive reflux disease (NERD) patients, and 49 patients (24 males; median age, 46 years) with no complaints of GERD were assigned to Controls. The EGD images were documented on tape. Four biopsies were performed at the distal esophagus. Experienced endoscopists and pathologists examined twice, at different times and separately, the respective endoscopic and histological examinations. For all comparisons, kappa (K) match, 95% confidence interval, and the P value were determined.ResultsIn NERD patients, the match between endoscopic and histological findings as to the aspect of the mucosa of the esophagus was poor (K = 0.08). EGD showed 17.2% sensitivity and 90.2% specificity, whereas the esophagus histological study showed 39% sensitivity and 73% specificity. NERD patients presented a greater frequency of esophagitis than Controls (38.6% vs. 22.4%), but this value was nonsignificant (P = 0.023).ConclusionAgreement between patient symptoms or endoscopic findings and histological diagnosis is poor; therefore, histology of the esophageal mucosa as a diagnostic test of NERD demonstrates low sensitivity and does not justify routine use.