Observational versus antibiotic treatment for uncomplicated diverticulitis: an individual-patient data meta-analysis

Background Two RCTs (AVOD and DIABOLO) demonstrated no difference in recovery or adverse outcomes when antibiotics for acute uncomplicated diverticulitis were omitted. Both trials showed non-significantly higher rates of complicated diverticulitis and surgery in the non-antibiotic groups. This meta-analysis of individual-patient data aimed to explore adverse outcomes and identify patients at risk who may benefit from antibiotic treatment. Methods Individual-patient data from those with uncomplicated diverticulitis from two RCTs were pooled. Risk factors for adverse outcomes and the effect of observational management were assessed using logistic regression analyses. P < 0 center dot 025 was considered statistically significant owing to multiple testing adjustment. Results In total, 545 patients in the observational group and 564 in the antibiotics group were included. No statistical differences were found in 1-year follow-up rates of ongoing diverticulitis (7 center dot 2 versus 5 center dot 0 per cent in observation versus antibiotics groups respectively; P = 0 center dot 062), recurrent diverticulitis (8 center dot 6 versus 9 center dot 6 per cent; P = 0 center dot 610), complicated diverticulitis (4 center dot 0 versus 2 center dot 1 per cent; P = 0 center dot 079) and sigmoid resection (5 center dot 0 versus 2 center dot 5 per cent; P = 0 center dot 214). An initial pain score greater than 7, white blood cell count exceeding 13 center dot 5 x 10(9)/l and previous diverticulitis at presentation were risk factors for adverse outcomes. Antibiotic treatment did not prevent adverse outcomes in patients at high risk of adverse events. Conclusion Observational management of acute uncomplicated diverticulitis is safe. Some statistical uncertainty remains, depending on the thresholds of clinical relevance, owing to small differences, but no subgroup that would benefit from antibiotic treatment was apparent.