The Most Popular Kids’ Sleep Aid Was Linked to Less REM Sleep

Melatonin has become the go-to sleep aid for kids. In the US, surveys suggest up to 19% of school-age children have used it in the past month.

It’s sold as a supplement, often as a sweet-tasting gummy. Many parents see it as natural and harmless.

In the US, it’s available without a prescription. Yet experts still recommend changing sleep habits first, before any sleep aid.

But we know surprisingly little about how it changes children’s sleep itself.

A new study used overnight lab sleep tests to find out. Children taking melatonin had slightly less REM sleep, the dreaming stage.

Everything else about their sleep looked the same.

Is melatonin safe for kids?

The short answer from the American Academy of Pediatrics: short-term use looks relatively safe. Much less is known about longer use.

In the US, melatonin is sold as a dietary supplement. So it isn’t regulated or approved by the Food and Drug Administration for this use.

In the UK, it’s prescription-only. Children with longer-term sleep problems usually get it from a specialist, the NHS says.

The most common side effects are morning sleepiness, drowsiness and needing to pee more at night, according to the AAP.

The bigger everyday risk is accidental overdose. From 2012 to 2021, there were more than 260,000 US child poisoning reports involving melatonin.

Most of those children had no symptoms. Some needed hospital care, and two children died, the AAP notes.

The new study adds one more open question: what melatonin does to the structure of a child’s sleep.

The study: 684 children in a sleep lab

Where the data came from

The researchers, from Harvard Medical School, used a large database of sleep tests from Nationwide Children’s Hospital in Ohio.

It holds overnight sleep studies of children seen at the hospital’s sleep lab between 2017 and 2019.

The gold-standard sleep test

These children had polysomnography. That’s a full overnight test that records brain waves, breathing, eye movements and muscle activity.

It’s the only reliable way to measure sleep stages. Most earlier melatonin studies in children used wrist trackers or parent reports instead.

One review of melatonin in autistic children found 18 studies. None reported lab-measured sleep stages.

Matching like with like

The team started with 3,392 children. Of these, 346 had melatonin on their medication list.

Children given melatonin were different to begin with. More had epilepsy, ADHD and other health conditions.

So each melatonin user was matched with a similar child who wasn’t using it, based on age, sex, weight, overall health and key sleep conditions. That gave 342 matched pairs, or 684 children.

The results: a small dip in dreaming sleep

Less REM sleep

The main question, chosen in advance, was about REM sleep.

Melatonin users spent a median of 16.7% of their sleep in REM. Matched non-users spent 19.0%.

That’s a difference of 2.3 percentage points. By standard measures, it’s a small effect.

It’s also within the normal night-to-night variation in children’s REM sleep. And nobody knows how big a REM change must be to affect development or memory.

Why REM sleep matters

REM sleep is when most dreaming happens. In children, it’s thought to support brain development, memory and emotional regulation.

Some medicines that suppress REM sleep, such as certain antidepressants and sedatives, are used cautiously in children for that reason. Melatonin interacts with some of the same brain chemical pathways, which is why the researchers looked at REM first.

Everything else looked the same

The team checked 14 other sleep measures. None differed after accounting for testing so many things.

Interestingly, melatonin users didn’t fall asleep faster in the lab. That’s the effect melatonin is best known for.

Time awake during the night looked a little lower in melatonin users at first. But that difference didn’t survive the strict statistical check.

The authors suggest the lab setting, with its wires and unfamiliar room, may swamp small effects on falling asleep.

How solid is the REM finding?

The REM difference held up when the team matched each user with two or three non-users. It also held when they averaged across repeat sleep tests.

But it shrank when they also accounted for ADHD and other psychiatric diagnoses. When they added insomnia and other sleep medicines too, it was no longer statistically clear.

Age made a difference, maybe

The dip looked largest in children aged 6 to 12. It was smaller in younger children and teenagers.

But the tests for an age pattern weren’t statistically significant, so this needs more study.

In autistic children, melatonin users had a bit more light (stage 2) sleep, but this didn’t hold up statistically. In children with ADHD, the REM difference was small and not significant.

Why the result is hard to read

It might not be the melatonin

The authors are refreshingly frank. The REM difference could reflect melatonin itself.

But it could also reflect differences between the children that the matching missed. And children who naturally have less REM sleep may be more likely to be given melatonin in the first place.

A big gap in the data

The study only knew that melatonin was on a child’s medication list. It didn’t know the dose, timing or how long they’d used it.

It couldn’t even confirm that the child took melatonin on the night of the test. Some non-users may also have been taking it over the counter.

The authors note that each of these gaps would tend to hide a real effect rather than create a false one. The test also captured just one lab night, which may not match a child’s usual sleep.

A special group of children

These were children referred to a hospital sleep lab, often for breathing problems, epilepsy or developmental conditions.

Most had obstructive sleep apnoea: about seven in ten melatonin users and three in four non-users.

Both groups had less REM sleep than typical children. That probably reflects their sleep problems, not melatonin.

The authors say doctors treating children with developmental conditions, who may already have less REM sleep, should be aware of the finding.

What earlier research found

Melatonin in young children

A 2026 systematic review in JAMA Network Open looked at melatonin in children aged 6 and under.

Trials involving 167 young children with neurological conditions found melatonin helped them fall asleep, with few side effects.

But there was no long-term safety data and no evidence for typically developing children. Registry and poison-centre data showed rising prescriptions and overdoses.

What’s really in the gummies

A 2023 study in JAMA tested 25 melatonin gummy products sold in the US.

Their actual melatonin content ranged from 74% to 347% of what the label said. Only 3 of 25 were within 10% of the label.

One product contained no melatonin at all, but did contain CBD.

How it fits together

Melatonin can help some children with specific conditions fall asleep. But the evidence for everyday use in healthy children is thin.

The new study adds a small, uncertain signal about REM sleep. It’s not proof of harm, but it’s a reason not to treat melatonin as sweets.

How much should you trust this?

Early. It’s the biggest lab-sleep study of its kind, but a single snapshot can’t show cause and effect.

What makes it convincing

• It used full overnight lab sleep tests, the gold standard.
• REM sleep was chosen as the main question in advance.
• Careful matching made the two groups very similar on key factors.
• The REM finding held across several alternative analyses.
• The authors reported no conflicts of interest.

What makes me cautious

• It’s cross-sectional: a single night, not a before-and-after comparison.
• Nobody knows whether children took melatonin that night, or how much.
• The difference shrank, and lost statistical clarity, with extra adjustment.
• The children came from a hospital sleep lab, not the general population.
• It’s from one hospital, so it needs repeating elsewhere.

What this means for you

If your child uses melatonin, there’s no need to panic. This study doesn’t show harm.

But it’s a good moment to check how and why it’s being used.

• Start with sleep habits. A regular bedtime, a calm routine and screens off before bed are the recommended first step.
• Talk to your doctor before starting. Ask whether melatonin suits your child, what dose and for how long.
• Treat gummies as medicine. Store them out of reach; they look and taste like sweets.
• Choose carefully. Doses in over-the-counter products can differ a lot from the label.
• Review it regularly. Many children don’t need melatonin long term.
• Watch for signs of a sleep disorder. Loud snoring, pauses in breathing or constant daytime tiredness need a doctor’s check, not a supplement.

The NHS explains how melatonin is used, including for children. The American Academy of Pediatrics also has advice on melatonin and children’s sleep.

In this video, a sleep expert at Boston Children’s Hospital answers whether melatonin is safe for kids:

What we still don’t know

1. Does melatonin itself lower REM sleep? Only a trial that tracks sleep before and after can tell.
2. Does a small REM dip matter for development? No threshold for harm is known.
3. Do dose and timing matter? This study couldn’t measure them.
4. What about healthy children at home? These children were referred to a sleep clinic.
5. Which ages are most affected? The hint of a bigger effect at 6 to 12 needs testing.
6. What happens with years of use? Long-term data are still missing.

My take: a small signal worth respecting

What I like about this study is its honesty. The authors found a result, then showed how fragile it is.

I’m not worried that melatonin is harming children’s brains. The difference is small, and it may not be caused by melatonin at all.

But it does push back on the idea that melatonin is just a harmless sweet. It’s a hormone, and we still know little about years of nightly use in children.

For most kids, better sleep starts with routines, not gummies.