Errors, rarities and surprises in our urgent upper panendoscopic series

During the last three decades 4412 urgent panendoscopies were performed because of upper gastrointestinal bleeding in our department. In 2562 cases we had the chance to compare the initial results of the emergency oesophago-gastro-duodenoscopy with the findings of a repeated endoscopy, histology, surgery or autopsy. Based on the evaluation of diagnostic end-points, in 46 cases out of 1348 patients followed-up with repeated endoscopy, in 13 out of 139 patients with a definitive histological finding, in 52 out of 635 surgeries, and in 26 out of 440 autopsies a discrepancy or error was discovered in the initial identification or localization of the bleeding source. We report these diagnostic faults divided into groups and listed according to different types. A few rarities are also presented, and these are illustrated with endoscopic, surgical, histological, pathological, radioisotopic, angiographic, capsule endoscopic or endosonographic photos. The diagnostic minimal requirements of emergency endoscopy (establishing the intensity of the bleeding, localizing it to the appropriate segment and determining whether it is circumscribed, diffuse or multiplex) can be fulfilled with the 90–95% accuracy reported in the literature even during the first examination. When investigating the causes of diagnostic errors the following reasons should be considered: (1) the evaluability of the endoscopic examination itself, (2) certain difficulties in localization, (3) subjectivity in describing the findings, (4) the disturbing presence of haemorrhagic gastropathy, and (6) pitfalls resulting from multiplicity of bleeding sources. In any uncertain case or before surgery a repeated upper panendoscopic examination can increase the diagnostic accuracy.