Randomized clinical trial of elective resection versus observation in diverticulitis with extraluminal air or abscess initially managed conservatively

Background: The aim of this RCT was to determine whether elective resection following successful non-operative management of a first episode of acute sigmoid diverticulitis complicated by extraluminal air with or without abscess is superior to observation in terms of recurrence rates. Methods: This was a single-centre, sequential design RCT. Patients were randomized to elective surgery or observation following non-operative management and colonoscopy. Non-operative management included nil by mouth, intravenous fluids, intravenous antibiotics, CT with intravenous contrast on arrival at hospital, and repeat CT with intravenous and rectal contrast on day 3 in hospital. The primary endpoint was recurrent diverticulitis at 24 months. Patients with a history of sigmoid diverticulitis, immunosuppression or peritonitis were not included. Results: Of 137 screened patients, 107 were assigned randomly to elective surgery (26) or observation (81), and underwent the allocated intervention after successful non-operative management. Conservative management failed in 15 patients. Groups were similar in age, sex, BMI, co-morbidities and colorectal POSSUM. Rates of recurrent diverticulitis differed significantly in the elective surgery and observation groups (8 versus 32 per cent; P = 0·019) at a mean(s.d.) follow-up of 37·8(8·6) and 35·2(9·2) months respectively. There was also a significant difference in time to recurrence (median 11 versus 7 months; P = 0·015). A total of 28 patients presented with recurrent diverticulitis complicated by extraluminal air and/or abscess (2 elective surgery, 26 observation), all of whom recovered with repeat non-operative management. Conclusion: The majority of patients observed following conservative management of diverticulitis with local extraluminal air do not require elective surgery. Registration number: NCT01986686 (http://www.clinicaltrials.gov).