ABSTRACT
Upper gastrointestinal bleeding [UGIB] is a frequent life-threatening emergency resulting in a large number of hospital admissions. Upper endoscopy has a crucial role in the diagnosis and treatment of UGIB, however the characteristics of our patients and the impact of our practice in these cases are still limited. Our aim was to assess, in a prospective way, the predictive factors of mortality in patients admitted to the gastroenterology unit at Hotel-Dieu de France hospital during the years 2002-2003 and to establish predictive factors of prolongation of hospital stay and occurrence of complications. Our study included 96 consecutive patients. The sex ratio was equal to 1 with a mean age of 63.24 +/- 8.72 years. Most endoscopic exams[67.7%] were done after 24 hours of the onset of UGIB. Endoscopic accuracy was a high as 98.95% [95/96 cases] with the use of 2 endoscopic exams to localize the bleeding lesion in only 4 cases [4.2%]. Peptic ulcer was the main cause of UGIB [44.8%], followed by bleeding erosive gastritis or duodenitis [13.5%], variceal bleeding [10.4%], oesophagitis [10.4%], and Dieulafoy's lesions[6.3%]. Endoscopic treatment was performed in 33.3% of the patients. Permanent hemostasis was achieved in 81.3% of the patients. Permanent hemostasis was achieved in 81.3% of the patients at the first endoscopic intervention and in62.5% of the patients at the first endoscopic intervention and in 62.5% of the patients after rebleeding. Rebleeding and/or absence of hemostasis after endoscopic diagnosis were reported in 15.6% of patients. Emergency surgery was rarely necessary [6 cases]. The average number of blood units was 3.37 +/- 2.28 per patient. Coagulation disorders, chemotherapy treatment, shock at admission and absence of hemostasis were predictive of a transfusion higher than 2 blood units on multivariate analysis. The average length of hospital stay was 9.58 +/- 5.97 days. The overall mortality rate of 10.4% was correlated, on the basis of multivariate analysis to 1/ cirrhosis, 2/ creatinin level higher than 110 micro mil/l, 3/ hemoglobin level at admission lower than 5 g/dl, 4/ prothrombin time below 60% and 5/ defective hemostasis after endoscopic intervention. In conclusion, when UGIB occurs in cirrhotic or renal insufficient patients or in the presence of coagulation disorders and when it is massive and uncontrollable it will be associated with a bad prognosis. In these cases a more aggressive treatment may be able to improve their outcome