Causes and outcome of upper and lower gastrointestinal bleeding: the Grady Hospital experience.

BACKGROUND Gastrointestinal (GI) bleeding remains a common medical condition, with a mortality rate believed to have remained unchanged over the past five decades. METHODS Over a 50-month period, the gastroenterology consultative service at a large inner-city hospital prospectively evaluated acute upper GI (UGI) bleeding or lower GI (LGI) bleeding in consecutive patients. A number of clinical variables were recorded at admission on a standardized data collection form. The cause of bleeding was determined in most patients by endoscopic examination. RESULTS Of the 796 patients assessed for UGI bleeding, 727 (91%) had upper endoscopy. The most common causes of UGI bleeding were gastric ulcer (32%), duodenal ulcer (28%), esophageal varices (9%), and Mallory-Weiss tear (6%). The rebleeding rate was 14% and 20% of patients had endoscopic therapy. Surgical therapy for bleeding was required in 7% of patients. Of the 165 patients assessed for LGI bleeding, 150 (91%) had colonoscopy. Colonic diverticulosis was considered etiologic in 56% of patients, followed by colonic ulcers in 10%, carcinoma in 7%, and vascular ectasias in 5%. The rebleeding rate in these patients was 20%, and surgical therapy for bleeding was required in 10%. The overall mortality for patients with UGI bleeding was 9% and was independently associated with portal hypertension and rebleeding. In contrast, the mortality rate for LGI bleeding was 4%, and there was little power to determine significant factors associated with death. CONCLUSIONS The causes of gastrointestinal bleeding remain little changed over the past several decades, though in our large series the need for surgical therapy and the mortality from both upper and lower GI bleeding were low.