In severe childhood sleep apnoea, tonsillectomy alone often did not cure

Design

Retrospective nested cohort with inverse probability weighting

Population

174 children with pre- and post-tonsillectomy polysomnography for OSA

Primary outcome

Surgical success (post-operative AHI <5 events per hour)

Effect

Success 70.1% overall; OR 0.96 per AHI event; 41.7% success when AHI >30

This retrospective cohort from a US quaternary paediatric centre looked at 940 children who had tonsillectomy for obstructive sleep apnoea (2016–2024). Of these, 174 had polysomnography both before and after surgery. Because children sent for a repeat study tend to be sicker, the authors used inverse probability weighting to correct for who was retested.

Surgery achieved an AHI below 5 in 70% of the 174. Each extra event per hour before surgery was associated with lower odds of success (OR 0.96 per event per hour, P < 0.001). Children with very severe apnoea — an AHI above 30 — succeeded only 41.7% of the time. Obesity appeared to lower the odds of success (OR 0.53) but did not reach significance.

The practical message is about counselling and follow-up. Families of children with very severe apnoea should hear before surgery that tonsillectomy is likely to help but less likely to cure, and these children need a planned post-operative sleep study rather than reassurance alone.

• Tell families of children with an AHI above 30 that tonsillectomy alone is unlikely to be curative.
• Plan post-operative polysomnography for children with severe or very severe apnoea.
• Discuss weight management alongside surgery in children with obesity.
• Consider further assessment — drug-induced sleep endoscopy, CPAP, orthodontics — when apnoea persists after surgery.
• Keep the preoperative AHI in the referral letter; it predicts the outcome.

Why it matters

Families often expect cure from tonsillectomy, and children with severe disease are the ones left untreated if that expectation goes unchallenged.

Don't overread it

A single-centre retrospective cohort; only 174 of 940 children had both sleep studies.

The statistics, in plain English

An OR of 0.96 per event per hour sounds small, but it compounds: a child with an AHI of 40 has much lower odds of success than one with 10. The weighting corrects for which children had a repeat study, but only for factors that were recorded.

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.