Management of Delayed Postpolypectomy Bleeding: A Decision Analysis

OBJECTIVES The benefit of repeat colonoscopy in managing delayed postpolypectomy bleeding is unknown. This study aimed to assess the outcome of repeat colonoscopy to achieve hemostasis. METHODS Endoscopic management of postpolypectomy bleeding is modeled as a decision tree, measuring the expected overall fraction of patients who benefit from therapeutic hemostasis and the number of patients needed to treat (NNT) in order to achieve one beneficial hemostasis. RESULTS A repeat colonoscopy to identify and treat postpolypectomy bleeding is beneficial in about 22% of patients, corresponding to an NNT of 4.5 patients. The outcome of the model is sensitive to assumptions underlying the fractions of patients who need treatment and would benefit from successful endoscopic hemostasis. Varying these probabilities over a broad range changes the fraction of patients benefiting from endoscopy between 3% and 33% and the NNT between 28 and 3 patients, respectively. CONCLUSIONS The expected outcome of repeat colonoscopy justifies the endoscopic attempts at therapeutic hemostasis. The results also suggest that in many patients expectant management aimed at spontaneous resolution of the bleeding remains a valid option.