Background: Postpolypectomy hemorrhage may warrant intensive care monitoring, transfusions, and surgery. We sought factors predicting significant bleeding requiring blood transfusion and the benefits of critical care monitoring.
Methods: Patients with postpolypectomy bleeding between April 1989 and November 1996 were identified from a comprehensive GI bleeding database. Data included age, gender, medical history, medications, polyp characteristics, and polypectomy technique. Outcomes assessed included bleeding cessation, transfusion requirements, recurrent bleeding, length of stay, and death.
Results: There were 83 patients with a median age of 73 years (range 18 to 88 years; 56 men, 27 women). Comorbid conditions were common (71.1% cardiovascular, 43.4% musculoskeletal, 14.5% hematologic, 6.0% renal). Within 3 days of presentation, 32.5% had taken aspirin, 10.8% nonsteroidal anti-inflammatory drugs, 12.0% warfarin, and 12.0% corticosteroids; and within 1 day, 10.8% intravenous heparin, 7.2% subcutaneous heparin, and 7.2% dipyridamole. Fifty-seven percent of patients were hemodynamically stable. Sessile cecal polyps greater than 2 cm in diameter bled more commonly. The median number of units transfused was equal between critical care and noncritical care patients. Using age in the logistic regression model, no other variable was predictive of transfusion. Eighty patients (96.4%) received endoscopic therapy, 1 required embolization and 2 hemicolectomy. There was no significant difference in outcomes for patients managed in an intensive care unit versus a general medical floor.
Conclusions: Postpolypectomy bleeding appears to have a predictable presentation and outcome. Advanced age seems to be predictive of transfusion requirement. Patient monitoring in an intensive care setting is not absolutely necessary.