The aim of the study was to develop a risk model for upper gastrointestinal tract (GIT) bleeding in patients with brain injury of various etiologies.Material and methods. Case histories of 33 patients were included into a retrospective descriptive study: 22 patients had severe brain injury of various etiologies, and 11 patients after elective surgery for cerebral aneurisms with uneventful postop period were taken for comparison. The patients were grouped in two arms: Group 1 included patients with obvious signs of GIT bleeding (N=11) and Group 2 had no obvious signs of bleeding (N=22). Complaints, life and medical history, comorbidities, specialists’ exams data, results of laboratory and instrumental examinations, therapeutic regimens were analyzed. Presence of disproportionate pathologic sympathetic overreaction to acute brain injury, i.e., paroxysmal sympathetic hyperactivity (PSH), was assessed on admission and on Days 1, 3 and 5 after brain injury. Results. A model for upper GIT bleeding risk assessment was designed using logistic regression. The resulting model gains high quality rating: χ²=33,78, 3; p<0,001; OR=315. The risk of upper GIT bleeding exceeded 95% in patients having combination of 4 symptoms in their medical history (presence of PSH on Day 1 after acute brain injury; Karnofsky performance scale index 75; lack of neurovegetative stabilization in the acute period of brain injury; gastric and/or duodenal ulcer).Conclusion. Determining the risk factors thresholds enables stratification of patients by the risk for upper GIT bleeding. Modification of the identified four risk factors (presence of PSH on Day 1after acute brain injury; Karnofsky performance scale index 75; lack of neurovegetative stabilization in the acute period of brain injury; gastric and/or duodenal ulcer) will probably reduce the occurrence of upper GIT bleeding in patients with acute brane injury of various etiology.