Design: Systematic review and meta-analysis of randomised trials
Population: 8,970 immunocompetent children with non-severe CAP, 10 trials
Primary outcome: Treatment failure, relapse and non-adherence
Effect: Failure RR 1.07 (95% CI 0.93–1.24); non-adherence RR 0.35 (0.24–0.50)

A meta-analysis pooled 10 randomised trials with 8,970 immunocompetent children with non-severe community-acquired pneumonia, comparing antibiotic courses of five days or less with longer courses.

Treatment failure (RR 1.07) and relapse (RR 1.04) did not differ, with low heterogeneity. Non-adherence was about a third as common with short courses (RR 0.35). Adverse events were similar, though the estimate varied widely between trials.

The authors note that a pooled analysis like this does not formally prove non-inferiority, but it agrees with the individual trials behind it. The practical advantage is real: a child who finishes a five-day course is better treated than one who stops a ten-day course on day four. For non-severe pneumonia in a child who is improving, a shorter course is a sound default, with clear advice to return if fever persists or breathing worsens.

Consider a five-day course of oral amoxicillin for non-severe pneumonia in an otherwise well child.
Reassess at 48 to 72 hours, or advise return, if fever or fast breathing persists.
Give parents a clear end date for the antibiotic to support adherence.
Keep longer courses for severe, complicated or immunocompromised cases.

Why it matters

Shorter courses fix the adherence problem that undermines longer ones in children.

The statistics, in plain English

A risk ratio of 1.07 for treatment failure, with an interval of 0.93 to 1.24, means short courses were about as effective, though a small disadvantage cannot be fully excluded. I² of 3% shows the trials agreed. The adherence result (RR 0.35) is large and consistent.