- Design
- Systematic review and meta-analysis of 10 randomised trials
- Population
- 8,970 immunocompetent children under 18 with non-severe community-acquired pneumonia
- Primary outcome
- Treatment failure, relapse and non-adherence
- Effect
- Failure RR 1.07 (0.93–1.24); relapse RR 1.04 (0.72–1.49); non-adherence RR 0.35 (0.24–0.50)
This meta-analysis pooled 10 randomised trials with 8,970 immunocompetent children under 18 with non-severe community-acquired pneumonia, comparing antibiotic courses of 5 days or less with longer ones.
Treatment failure (RR 1.07, 95% CI 0.93–1.24) and relapse (RR 1.04, 0.72–1.49) did not differ. Non-adherence was about a third as common with short courses (RR 0.35, 0.24–0.50). Adverse events were similar, though trials varied widely.
For family physicians this matters because the longer course is often not finished anyway. Stopping a course the family has already abandoned does no good; prescribing one they will finish does. The analysis was not designed to prove non-inferiority formally, but it agrees with individual trials. It applies to non-severe pneumonia in otherwise healthy children; it does not cover severe disease, hypoxia or complicated pneumonia.
- Consider a 5-day antibiotic course for non-severe community-acquired pneumonia in an otherwise well child.
- Check for signs of severe disease — chest indrawing, hypoxia, inability to feed — before choosing a short course.
- Tell families why the course is short so they do not seek a longer one elsewhere.
- Arrange review if fever persists beyond 48 to 72 hours, whatever the course length.
- Follow national or IAP guidance where it specifies a duration for your setting.
Why it matters
The longer course often goes unfinished, so the short one may be the one children actually receive.
Don't overread it
This meta-analysis was not designed to prove non-inferiority, and it excludes severe pneumonia.
The statistics, in plain English
A risk ratio of 1.07 for failure with an interval of 0.93 to 1.24 crosses 1.0: no difference was detected, and a large increase in failure is unlikely. The non-adherence result (0.35) is the largest and most consistent effect, meaning children on short courses missed doses about a third as often.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for family medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free