Acute cough in a young child is one of the commonest reasons a parent asks for a prescription, and one of the commonest reasons a child receives a drug with no useful effect and a real side-effect profile. This meta-analysis compared honey directly with pharmacological treatment — not with placebo — across seven randomised trials, six of which contributed to the meta-analysis, totalling 576 children with mean ages between 3 and 5 years and follow-up limited to three days.
The primary analysis showed high heterogeneity and no significant difference on most endpoints. The exception was the cough's effect on the child's sleep, which was better with honey (mean difference -0.73, 95% CI -1.45 to -0.01, p=0.05). After leave-one-out sensitivity analysis, honey came out significantly better across every outcome — cough frequency, severity, bothersomeness, sleep quality for both child and caregiver, and the global score — and the pattern held regardless of which pharmacological comparator was used.
The authors are appropriately careful, and so should we be: the across-the-board superiority emerged from post-hoc analysis and is hypothesis-generating, not established. What the primary analysis does establish is that honey is not worse. Given that the comparators include dextromethorphan and other antitussives with documented harms in young children and no convincing efficacy, 'not worse' is enough to decide with.
So the practical position is straightforward. For a child over one year with acute cough, honey is a reasonable first recommendation: cheap, available in every household, no meaningful adverse effects, and it appears to help the thing parents actually care about, which is whether anyone gets any sleep. The absolute contraindication is age under 12 months, because of infant botulism, and that needs saying explicitly every time rather than assumed. Also worth naming: this addresses symptomatic relief in self-limiting illness, and it is not a reason to skip assessing a coughing child for pneumonia, foreign body or tuberculosis.
- Recommend honey rather than an antitussive for acute cough in a child over one year.
- Never in infants under 12 months — infant botulism risk, and say so to the parent explicitly.
- Frame the benefit around sleep, which is the outcome parents came about.
- Use the consultation to explain why cough medicines are being avoided, or the prescription request will simply move elsewhere.
- Symptomatic advice does not replace assessment for pneumonia, inhaled foreign body or tuberculosis.
The statistics, in plain English
The distinction between the two analyses here is the whole interpretation. The prespecified primary analysis found no significant difference on most outcomes, with high heterogeneity between the trials — the honest headline. The sweeping superiority appeared only in a leave-one-out sensitivity analysis, which removes one study at a time to see how much any single trial is driving the result. That is a legitimate technique for checking robustness, but when the significant findings appear only after it, they are exploratory. The one primary-analysis positive, on the child's sleep, has a confidence interval running from -1.45 to -0.01 — its upper limit is essentially zero, so it only just clears significance.
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 paediatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free