Colonoscopy: the initial test for acute lower gastrointestinal bleeding

Despite literature showing safety, accuracy, and therapeutic capability of emergency colonoscopy for acute lower gastrointestinal (LGI) bleeding, surgical literature suggests that this examination is difficult to perform in the acute setting. In contrast to currently accepted protocols, we believe that unprepared colonoscopy within 24 hours of presentation can be performed safely with a high rate of success in localizing and often treating the specific cause of LGI bleeding. We report results over a 7-year period in our institution using early colonoscopy as the primary investigative method for the diagnosis and treatment of LGI bleeding. We analyzed 85 consecutive patients suspected of LGI bleeding referred to the surgical service between 1989 and 1996. LGI bleeding was defined as the passage of blood per rectum, distal to the ligament if Trietz. We excluded patients who were only hemoccult positive or had an upper gastrointestinal source by nasogastric aspirate or upper gastrointestinal endoscopy. All patients underwent urgent unprepped colonoscopy by surgical endoscopists relying on the cathartic effect of blood and liberal suction/irrigation to cleanse the colon. Therapeutic maneuvers included Nd:YAG laser or BICAP coagulation. Studies in which active bleeding was found or lesions with endoscopic evidence of recent hemorrhage were considered positive. A total of 126 colonoscopies were performed in 85 patients, 44 males and 41 females, with a median age of 75 years (range, 12-91 years). Fifty-three patients (62%) had hematocrit drops of greater than 5 per cent. Thirty-four patients were transfused an average of 4.5 units of blood per patient. The source of bleeding was correctly identified in 82 of 85 (97%) patients. Ninety-one per cent of sources were colonic, and 9 per cent were small bowel. Fecal residue prevented initial adequate examination in only two patients. Diverticulosis (20%), ischemic colitis (18%), hemorrhoids (14%), and arteriovenous malformations (11%) were the predominant sources of bleeding. Spontaneous cessation of bleeding occurred in 58 (68%) patients. Control of active hemorrhage was achieved endoscopically in 17 of 27 acutely bleeding patients. Significant therapeutic interventions were performed in 26 additional patients, including fulgration, polypectomy, relief of obstruction, and removal of foreign body. One patient with asymptomatic free air was observed nonoperatively, for a complication rate of 0.8 per cent. In-hospital mortality was 3.5 per cent (three patients), all secondary to multisystem organ failure and underlying disease. In-hospital rebleeding rate was 3.5 per cent (three). We conclude that, using colonoscopy, it is possible to identify the source of acute LGI bleeding in more than 95 per cent of cases. Diagnostic and therapeutic capability with colonoscopic intervention to control active hemorrhage is especially appealing. Additionally, the pattern, amount, and location of blood in the unprepared colon all give clues as to source and rate of bleeding. In experienced hands, morbidity and mortality of emergent colonoscopy is very low. High accuracy, safety, and therapeutic capability makes colonoscopy the initial diagnostic test of choice for acute LGI hemorrhage.