Dieulafoy’s lesion is an uncommon vascular anomaly characterized by a dilated submucosal artery that can cause severe upper gastrointestinal bleeding in the absence of underlying mucosal disease. It accounts for approximately 1–2% of all gastrointestinal hemorrhage.Endoscopic diagnosis relies on specific criteria: (1) active arterial spurting or micropulsatile bleeding from a small (<3 mm) mucosal defect or through normal mucosa; (2) a visible protruding vessel with or without active bleeding from a tiny mucosal defect; or (3) a densely adherent clot with a pinpoint attachment on otherwise normal mucosa. Despite these criteria, overdiagnosis remains under-reported, and the extent of misinterpretation in real-world endoscopy practice is unclear. We aimed to evaluate both diagnosis and overdiagnosis of Dieulafoy’s lesions over a 30-year period in a tertiary hospital.

 We reviewed all gastroscopy reports from our department between September 1993 and January 2024 in which the term “Dieulafoy” appeared. Demographic data (age, sex) were collected, and all reports underwent detailed reassessment. We evaluated stomach contents, completeness of mucosal inspection, and the presence of alternative bleeding sources such as ulcers, varices, erosions, or angiodysplasia. Each report was reassessed to determine whether the standard endoscopic diagnostic criteria for Dieulafoy’s lesion were clearly documented. For confirmed cases, the endoscopic hemostatic modality used was recorded.

A total of 59 reports were retrieved, corresponding to 52 patients (mean age 70 years; 71% male), all presenting with upper gastrointestinal bleeding. Lesions were most commonly described in the fundus (28,8%), followed by the duodenal bulb (21%), the lesser curvature (11%) and the posterior wall of the gastric body (11%). Overall, 42,3% of lesions labeled as “Dieulafoy” did not fulfil the diagnostic criteria, indicating substantial overdiagnosis. Nearly half of the procedures (57,6%) were performed in a stomach contaminated with blood or clots, limiting mucosal visualization. Apart from the lesions described as Dieulafoy, other mucosal abnormalities potentially responsible for bleeding were present in 34,6 % of cases. Several lesions initially labeled as Dieulafoy’s were reinterpreted as ulcers (19%). Hemostasis was achieved using adrenaline injection, ethanolamine injection (used until 2005), APC, thermal coagulation probes, hemoclips, or combination therapy (53%). Successful hemostasis was reported in 96,6 % of confirmed Dieulafoy cases. Five patients underwent second-look endoscopy; outcome data for the remaining patients were not available.

Dieulafoy’s lesions remain a significant diagnostic challenge. Limited visualization and inconsistent application of standard diagnostic criteria contribute to substantial overdiagnosis. This 30-year review underscores the importance of rigorous documentation and strict adherence to the recognized endoscopic criteria to improve diagnostic accuracy and guide appropriate treatment.