Objective: To investigate the sources of hemorrhage and use of endoscopic hemostasis in patients with clinically significant upper gastrointestinal (UGI) hemorrhage after admission to the intensive care unit (ICU).
Design and setting: Prospective study, 123 beds of ICU in a 1,629-bed medical center.
Measurements and results: Of the 9,512 consecutive admissions over a 2-year period 105 UGI hemorrhage patients underwent urgent bedside UGI endoscopy. We compared two groups of these patients, one receiving and the other not receiving endoscopic hemostasis. Ulcers with profusely bleeding stigmata occurred in 31 patients (29.5%), ulcers with clean bases or firmly adherent blood clots in 27 (25.7%), stress-related mucosal diseases in 23 (21.9%), esophageal varices in 5 (4.8%), malignancy in 4 (3.8%), and no detectable bleeding site in 15 (14.3%). Endoscopic hemostasis was attempted in 34 patients (32.4%). Primary hemostasis for them was achieved in 67.6% and the rebleeding rate was 30.4%. In-hospital mortality rate was 77.1% and death related to hemorrhage 6.2%. Length of ICU stay before endoscopic diagnosis was significantly shorter in those who underwent endoscopic hemostasis than those who did not (28.2+/-26.3 vs. 41.2+/-57.5 days).
Conclusions: Endoscopic hemostasis may be more beneficial when the period between ICU admission and development of hemorrhage is shorter. Bleeders can be more readily identified and controlled endoscopically in such patients. A significant proportion of bleeding sites cannot be identified by UGI endoscopy. It may indicate higher risk of small bowel hemorrhage in these critically ill patients.