Oral amoxicillin matched injectable antibiotics for chest-indrawing pneumonia without danger signs

Design

systematic review and random-effects meta-analysis with GRADE, PROSPERO registered

Population

children aged 2-59 months: 7,699 in four trials for fast-breathing pneumonia, 4,864 in three trials for chest-indrawing pneumonia

Primary outcome

treatment failure at day 6 and day 14, and mortality

Effect

fast breathing, failure risk ratio 0.84 (95% CI 0.75-0.94); chest indrawing, failure 0.94 (0.79-1.13) at day 14 and mortality 0.28 (0.09-0.86)

Two comparisons were pooled from randomised and quasi-randomised trials in children aged 2 to 59 months. Four trials with 7,699 children compared oral amoxicillin against no antibiotic for fast-breathing pneumonia. Three trials with 4,864 children compared outpatient oral amoxicillin against inpatient injectable antibiotics for chest-indrawing pneumonia without danger signs.

For fast-breathing pneumonia, amoxicillin reduced treatment failure at day 14 by 16% (risk ratio 0.84, 95% CI 0.75 to 0.94, I squared 0%, moderate certainty). For chest-indrawing pneumonia, treatment failure was similar between oral and injectable treatment at day 6 (0.96, 0.83 to 1.11) and day 14 (0.94, 0.79 to 1.13), while mortality was 72% lower with oral amoxicillin (0.28, 0.09 to 0.86). Serious adverse events did not differ. The authors are careful to say the absolute risk differences behind the mortality figure are small.

This is directly usable where it matters most. A child with chest indrawing and no danger signs - no inability to drink, no persistent vomiting, no convulsions, no lethargy or unconsciousness, no central cyanosis - does not need admission for injectable antibiotics on this evidence, and managing them at home avoids the transport cost, the bed, and the hospital-acquired exposure that pushed the mortality the other way. The danger-sign assessment is the whole safeguard, and it has to be done properly and repeated.

• Assess and document every danger sign before choosing outpatient management - that is what this evidence is conditional on
• Treat chest indrawing without danger signs with oral amoxicillin at home, with a defined review
• Arrange explicit follow-up at 48 hours and tell the carer exactly which signs mean return immediately
• Do not withhold antibiotics in fast-breathing pneumonia - failure was 16% lower with treatment
• Check the child can actually take and keep down oral medication before sending them home

Why it matters

It moves a large group of children out of hospital admission and injectable therapy without a measurable cost in treatment failure.

Don't overread it

This applies only where danger signs have been actively excluded - it is not evidence for oral treatment of severe pneumonia.

The statistics, in plain English

The two comparisons carry different weight. The fast-breathing result, risk ratio 0.84 with I squared of 0% and moderate certainty, is solid. The mortality finding for chest indrawing - 0.28, with an interval reaching 0.86 - looks dramatic as a ratio but rests on few deaths in either arm, which is why the authors flag that absolute differences are small; a 72% reduction of a small number is still a small number. Note also that the treatment-failure intervals for the oral-versus-injectable comparison span 1.0 in both directions, which is what equivalence looks like rather than superiority.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.