[Diagnostic value of unselected biopsies during gastroscopy]

In a prospective study 1414 unselected biopsies were obtained during 509 consecutive gastroscopies. Concordance between endoscopic and histological diagnosis was judged by means of the kappa coefficient. Best agreement was obtained with biopsies from the duodenal bulb, at 67.6% (kappa = 0.39). With biopsies from the body of the stomach agreement was only 35.7% (kappa = 0.12), from the antrum 28.8% (kappa = 0.10). Degree of agreement between endoscopic and histological result was independent of the instrument used, but was clearly dependent on the examining doctor and the number of examinations. The results suggest that clinically significant disease is often overlooked during endoscopy alone, unless simultaneously obtained unselected biopsies are examined at the same time. Thus, endoscopic examination alone missed Crohn's disease, chronic atrophic gastritis and various degrees of dysplasias. Unselected biopsy aids in objectifying the gastroscopic findings and in securing the diagnosis of various diffuse gastric mucosal changes.