In summary, patients who received antibiotics:

Benefits in NNT

    3.1% lower risk of treatment failure
    1 in 32 were helped (treatment failure, defined as deterioration prompting initial or expanded antibiotic treatment, prevented)

Harms in NNT

    4.1% higher risk of treatment related morbidity
    1 in 24 were harmed (adverse reaction or morbidity related to antibiotic use)

View As:

Source: Chabok A, Påhlman L, Hjern F, Haapaniemi S, Smedh K. Randomized clinical trial of antibiotics in acute uncomplicated diverticulitis. Br J Surg. 2012;99(4):532-9.

Daniels L, Ünlü Ç, De korte N, et al. Randomized clinical trial of observational versus antibiotic treatment for a first episode of CT-proven uncomplicated acute diverticulitis. Br J Surg. 2017;104(1):52-61.

Study Population: 1151 patients from two RCTs

Efficacy Endpoints: Requiring additional treatment or intervention during initial episode

Harm Endpoints: Antibiotic-related morbidity

Narrative: In the United States there are over 2.6 million outpatient visits and 200,000 inpatient admissions for diverticulitis annually.1 Diverticulitis can be divided into uncomplicated and complicated forms. The term complicated is used when diverticultitis is associated with abscess formation, fistula, and bowel obstruction or perforation. About 5-15% of patients develop an abscess or fistula, while bowel obstruction and frank perforation are rare.2 The mainstay of treatment for uncomplicated diverticulitis has been antibiotic therapy with bowel rest. However, recent studies have questioned the role of antibiotics.3, 4 Systematic reviews have examined outcomes of acute uncomplicated diverticulitis treated with or without antibiotics.5, 6, 7, 8

The reviews included randomized trials and observational studies. Because of the high risk of bias and confounding in observational studies, we report only results from the two randomized trials.3, 4 Endpoints included treatment failure, recurrence of diverticulitis, complications, re-admission, and mortality. Follow up was one month in one trial and 50 months in the second.3, 4

The difference between groups was not significant for any major endpoints including rates of recurrence (56/571 vs 54/580, p=0.77), complications (18/571 vs 10/580, p=0.12), readmission (113/571 vs 81/580, p=0.26), or mortality (3/571 vs 1/580, p=0.4).3, 4

Treatment failure, a secondary outcome initially reported in the systematic review by Emile et al. (defined as deterioration prompting initial or expanded antibiotic treatment), was lower in the antibiotic group (Odds Ratio: 0.6, 95% CI; 0.3 to 0.97); absolute risk reduction 3.1%, NNT 32, n=1151).3, 4, 5

While harms were not reported in the systematic review because of inconsistent reporting, the two randomized trials briefly mention adverse events. In the study by Chabok et al. only 3 patients in the antibiotic group experienced adverse events (allergic reactions). Daniels et al. reported 22 adverse events in the antibiotic group (all “antibiotic related without specification) and 1 in the control group (Odds Ratio: 25.68, 95% CI, 3.47 to 190.14; absolute risk reduction: 4.1%; Number-needed-to-harm: 24).3, 4

Caveats: There are several limitations of these data including having only two studies, along with important variations in outcomes and definitions. The Daniels et al. study included patients with small peri-colonic abscesses while the Chabok et al. study patients with pericolonic abscesses, which may have contributed to a higher treatment failure rate in the Daniels study (10.7% vs 3.2%), though the definition of treatment failure also differed between the studies. In Chabok et al., antibiotic usage was guided by C-reactive protein (CRP) levels, while the Daniels study did not describe a defined protocol or guidance for the use of antibiotics in the control group. Antibiotic treatments also differed between studies.3, 4

Both studies had significant limitations and potential for bias. Both were unblinded. Enrollment rates also varied by center, leading to a risk of selection bias, however concealment and randomization may have decreased this risk.3, 4

Based on limited data with high potential for bias we have chosen a recommendation color of Yellow (unclear benefits vs. harms, more data needed). Ongoing studies and future efforts will, we hope, be both more rigorous and better standardized to allow for more reliable pooling and comparison.

This series is coordinated by Christopher Bunt, MD, AFP Assistant Medical Editor, and Daniel Runde, MD, from the NNT Group.

Author: Shiva Poola, MD; Michael Ritchie, MD
Supervising Editors: Shahriar Zehtabchi, MD

Published/Updated: September 21, 2020

  1. The Title Bar

    The title bar is color-coded with our overall recommendation.

    • Green: Benefits outweigh risks.
    • Yellow: Unclear risk/benefit profile.
    • Red: Benefits do not outweigh risks.
    • Black: Obvious harms, no clear benefits.
  2. Tip content...