Adenotonsillectomy for Childhood Sleep Apnea: What the Trial Actually Showed
The largest randomised trial of surgery for childhood sleep apnea missed its own primary outcome. It still supported the operation — and it found that 46% of the children who had no surgery at all normalised anyway within seven months.
Removing the tonsils and adenoids is the standard treatment for obstructive sleep apnea in children. The best trial of it did not show what everyone expected, and the most useful number in it is about the children who had no operation at all.
This page is about that evidence. It is not a substitute for a paediatric assessment, and nothing on this site is a treatment for a child's breathing.
Why is surgery the first-line treatment in children?
Because the cause is usually different from adults. In an adult, obstructive sleep apnea is typically a matter of airway size, soft tissue, weight and muscle tone. In a child it is most often adenotonsillar hypertrophy — tonsils and adenoids large enough to physically crowd the airway.
That makes it one of the few situations in sleep medicine where removing the obstruction is both possible and curative in principle, which is why the surgery, rather than a device, is where treatment starts.
Adult remedies do not transfer. Why adult snoring remedies do not apply to children.
Orthodontic expansion is often raised alongside it. On pooled evidence it adds to adenotonsillectomy rather than replacing it, and on its own it does not reduce apnea severity at all. Palatal expansion for a child's airway.
What did the CHAT trial find?
The Childhood Adenotonsillectomy Trial, published by Marcus and colleagues in the New England Journal of Medicine in 2013, randomised 464 children aged 5 to 9 with obstructive sleep apnea syndrome — without prolonged oxygen desaturation — to early adenotonsillectomy or to watchful waiting with supportive care. Outcomes were measured at seven months.
The primary outcome was negative. It was attention and executive function on a standardised neuropsychological assessment. Improvement was 7.1 points in the surgery group against 5.1 in the watchful-waiting group, P = 0.16. Both groups improved. The difference between them was not statistically significant, and the baseline score was already close to the population average of 100.
The secondary outcomes were positive, consistently. Behaviour, quality of life, symptoms and polysomnographic findings all improved significantly more with surgery.
And the number worth carrying away: polysomnographic findings normalised in 79% of the surgery group versus 46% of the watchful-waiting group.
What does that 46% actually mean?
Nearly half the children who had no surgery had normal sleep studies seven months later.
That is not an argument against the operation — 79% is still clearly better than 46%, and the symptom and behaviour benefits were real. It is an argument against urgency in a specific group: school-age children with apnea that is not severe and does not involve prolonged desaturation.
Childhood sleep apnea in that band has a meaningful spontaneous resolution rate. Children grow, airways grow with them, and adenotonsillar tissue regresses with age on its own. A period of watchful waiting with proper follow-up is a legitimate clinical choice rather than neglect, and the trial is the reason it can be offered.
What the trial does not license: watchful waiting in a child with severe apnea, significant oxygen desaturation, or failure to thrive. Those children were not in it. The enrolment criteria specifically excluded prolonged oxyhaemoglobin desaturation, so the results say nothing about that group.
Evidence tier 1. A 464-child randomised controlled trial in the NEJM is the strongest evidence available in paediatric sleep medicine. The nuance is in reading which outcome was which, not in the quality.
Why did the cognitive result come out flat?
Several plausible explanations, and the trial cannot distinguish between them.
The children started near the population mean, so there was limited room to improve on that particular measure. Seven months may be too short for cognitive change to appear. Attention and executive function as measured by one instrument may not be the faculty most affected. And both groups received supportive care, which is not the same as nothing.
The honest reading is that the trial did not demonstrate a cognitive benefit on that measure in that timeframe — not that there is no cognitive cost to untreated childhood apnea, which is a different claim resting on different evidence.
What should a parent do with this?
Get it assessed properly, by a paediatric service. Childhood apnea is not diagnosed by a parent listening at a door, and the thresholds are different from adults: in children an AHI considerably below the adult cut-off is abnormal.
Expect tonsil and adenoid size to be examined, because that is usually the obstruction and it is visible. What tonsil grading shows.
Ask which group your child is in. Severe apnea or significant desaturation points toward surgery without delay. Moderate apnea without desaturation in a school-age child is the group where the trial supports a discussion about timing rather than an immediate booking.
Know the symptoms that are not snoring. Daytime behaviour, hyperactivity, bedwetting, mouth breathing and poor concentration are frequently the presenting signs, and parents often do not connect them to breathing. Snoring in children, and when to worry.
The line, stated plainly for children
Nothing sold for adult snoring is appropriate for a child's airway. Mouth tape, nasal strips and positional devices are not paediatric treatments, they are not tested in children, and using one on a child with undiagnosed apnea risks muting the sign that gets them assessed.
If a child snores habitually, pauses, gasps, or sleeps restlessly in unusual positions, the step is a doctor. Can children use nasal strips · Mouth breathing and a child's face.