Predictors of Recurrence Bleeding in Diverticular Hemorrhage. Prospective Study

Background: Diverticular hemorrhage is the most frequent cause of lower gastrointestinal bleeding, but its natural history is not well known. Aim: To know the risk of recurrence and mortality in the diverticular bleeding and its associated factors. Patients and Methods: Design: prospective cohort study. Setting: tertiary teaching hospital. Patients: all patients admitted to our hospital with diverticular hemorrhage from April 2004 to August 2008. Analysis: To evaluate the clinical-endoscopic risk score (Rectal bleeding in the first 4 hours after admission, presence of red blood on early colonoscopy and absence of ulcerative mucosal lesions on colonoscopy -J. Guardiola, DDW 2007-) to predict recurrence of bleeding, and to determine the risk and predictive factors of readmission, need for surgery and mortality. Results: A total of 126 consecutive patients admitted with diverticular hemorrhage were identified with 144 bleeding events. The 12% (15/126) were definitive source of bleeding (active bleeding, visible vessel and/or adherent clot). During the first episode, 20% of patients presented recurrence of bleeding. Some kind of therapy was given to 11 patients (9%): 7 endoscopic therapies, 3 urgent surgeries and 1 arterial embolization by angiography. Independent predictive factors for recurrence of bleeding were: rectal bleeding during de first 4 hours after admission (RR 6.06; 95% CI, 1.80-20.03; p=0.003) and presence of red blood on diagnostic colonoscopy (RR 3.76; 95% CI, 1.02-13.92; p=0.04). If diagnostic colonoscopy was performed within the first of 24 hours (55 patients), the combination of both factors can to predict the recurrence of bleeding (0 factors: 3.4%, 1 factor 10% and 2 factors 66%) The area under de ROC curve was 0.82. Fifteen patients (12%) were readmitted with a new episode of diverticular bleeding. With a median of 421 days of follow-up, the accumulative risk of readmission was 7%, 9% and 12% for 6, 12 and 24 months respectively. One patient required surgery and 3 patients were readmitted a second time. All of them were discharged. Predictive factors of readmission were not identified. Conclusion: The clinical-endoscopic risk score can identify the diverticular bleeding episodes with a high risk for recurrence bleeding. Given the low probability of readmission, and the absence of mortality in our cohort, we don't recommend elective surgery in the second episode of diverticular hemorrage.