Melatonin: uses, evidence, dosing and safety

Melatonin is sold as a sleeping aid and it is not one. It is a timing signal, and almost everything confusing about it follows from that single fact.

Given at the right moment it will move your body clock. Given at bedtime to someone whose clock is already in the right place, it does very little, which is exactly what the trials show. And then there is a separate problem: the product frequently does not contain what the label says.

What melatonin is

Melatonin is a hormone rather than a nutrient. NCCIH puts it simply: melatonin is a hormone that your brain produces in response to darkness.

The research literature is more precise about the job: melatonin is an important neurohormone, which mediates circadian rhythms and the sleep cycle, circadian meaning the roughly 24-hour cycle the body runs on.

That distinction matters commercially. Most things in the supplement aisle are nutrients or plant extracts. This is a hormone, sold over the counter in doses far above what the body makes.

How melatonin works

The brain releases melatonin as light fades, and the rise is a message rather than a sedative effect. It tells the rest of the body that night has started, and the body then does the things it associates with night, of which falling asleep is one.

Two consequences follow, and they explain most of the evidence below.

Timing matters more than dose. A signal delivered at the wrong hour moves the clock the wrong way, which is why the same product can help one person and leave another wide awake at 3am.

And if your clock is already correct, adding the signal tells the body something it already knows. There is little left to fix.

What the evidence says about melatonin and ordinary insomnia

This is what most buyers want and it is the weakest case.

A meta-analysis of prolonged-release melatonin in insomnia disorder found it efficacious with a small to medium effect size on subjective sleep onset latency, meaning how long people said it took them to fall asleep. That came to a little over six minutes.

Measured objectively rather than by self-report it was smaller still, around five minutes, and sleep efficiency rose by under two percentage points.

Five to six minutes is a real effect. It is also not what someone lying awake at 2am is hoping to buy, and it is worth knowing before spending money on it.

One subgroup did better. For patients aged 55 and over, melatonin was efficacious on sleep efficiency with a large effect size, which fits the fact that natural melatonin production declines with age.

What the evidence says about melatonin and delayed sleep phase

Here the picture changes, because here the problem is the clock.

Delayed sleep phase disorder means the body clock runs genuinely late: sleep does not arrive until the early hours, and waking for work is a daily fight. It is a timing fault, which is what melatonin is for.

Pooling five trials in 91 adults and four trials in 226 children, melatonin treatment advanced mean endogenous melatonin onset by 1.18 hours and the clock hour of sleep onset by 0.67 hours. It also decreased sleep-onset latency by 23.27 minutes.

Then the detail that proves the mechanism. The wake-up time and total sleep time did not change significantly.

It moved the whole night earlier without adding a single minute of sleep. That is a clock being reset, not a sedative working, and the authors conclude that melatonin is effective in advancing sleep-wake rhythm and endogenous melatonin rhythm in delayed sleep phase disorder.

What the evidence says about melatonin and children

Use in children has grown fast, and the evidence is better than the adult insomnia case while the caution is greater.

Pooled trials covered eight RCTs with 419 children and adolescents with idiopathic chronic insomnia, idiopathic meaning no identified underlying cause. Melatonin led to a moderate increase in total sleep time of around 30 minutes.

Thirty minutes is more than adults get, and that is the reason for the growth in use. The pediatric reviewers note that melatonin prescriptions for children and adolescents have increased substantially during the last decade.

Three things sit against it. It is a hormone given to a developing endocrine system, and long-term follow-up does not exist. The product quality problem below applies squarely to children’s gummies. And chronic insomnia in a child usually has a cause worth finding.

What the evidence says about melatonin and migraine

This is the strongest non-sleep result on the page, and almost nobody buys melatonin for it.

A 2026 review of randomized trials found melatonin beat placebo across the outcomes that define whether a migraine treatment is working: fewer headache days a month, less need for painkillers, and a higher share of people whose attacks halved. It also improved sleep quality and disability, meaning how much migraine interfered with work and daily life.

The comparison that keeps it honest is against an actual drug. Amitriptyline, a standard preventive, beat melatonin on several measures: compared with amitriptyline, melatonin was generally less effective for attack duration and severity.

But it was easier to live with, including a lower risk of sleepiness, which is the side effect that stops people taking amitriptyline.

The reviewers land somewhere sensible. Melatonin represents a reasonable preventive option, particularly as an adjunct during titration of first-line agents: something to use alongside a proper drug while its dose is being worked out. Anyone getting migraines often enough to consider this should be having that conversation with a doctor, not self-medicating.

What the evidence says about melatonin and blood sugar

Melatonin is involved in metabolism as well as timing, and there are now enough trials to answer this properly.

A 2026 review pooled 31 randomized trials. Its own title gives away the finding: melatonin improves insulin resistance markers but not fasting glucose. Insulin resistance means the body needing more insulin to shift the same amount of sugar.

Read the two halves separately. The meta-analysis showed no significant reduction in fasting glucose levels. That is the standard blood sugar test, unmoved. But improvements were observed in secondary metabolic outcomes, including the three-month blood sugar average and two measures of how well insulin is working.

That combination is genuinely ambiguous rather than quietly positive. When the primary measure does not move and secondary ones do, the honest reading is a hint, not a result. The authors put it as a modulatory role in metabolic regulation, which is a careful way of saying something is happening and it is not yet a treatment.

Nobody should take melatonin to manage blood sugar on this evidence, and nobody with diabetes should assume it is inert either.

Why it may depend on your genes

A 2026 crossover trial narrowed that ambiguity considerably. The common melatonin receptor 1B gene (MTNR1B) variant increases risk of type 2 diabetes, and the trial tested whether melatonin acts differently in people who carry it.

It does. Melatonin worsened glucose tolerance and reduced early C-peptide responses in risk allele carriers but not in non-carriers, driven by a suppression of glucose-stimulated first-phase beta-cell responsivity, the rapid burst of insulin released within minutes of sugar arriving.

It was 21 people, so it settles nothing about advice. It does suggest the mixed population results have a reason behind them.

What the evidence says about melatonin and cancer care

Melatonin is widely sold to people having cancer treatment, so the Cochrane review matters more here than usual.

Its verdict on the two things most often promised is blunt. The available evidence is of very low certainty, so we are unable to draw conclusions about the effects of melatonin on quality of life and sleep in people receiving cancer treatment.

Two narrower findings survive. Melatonin used alongside standard treatment probably reduces the risk of fatigue and may reduce nausea, with “probably” and “may” doing real work: they are Cochrane’s certainty grades, not hedging.

A separate 2024 meta-analysis found the same fatigue benefit and added the condition attached to it: significantly improved fatigue in studies with treatment durations of 13 weeks or more, but not in shorter ones. A three-month course is a different proposition from a bottle bought for a bad fortnight.

Cochrane’s overall position is the one to carry away: the decision for or against using melatonin as an adjunct to cancer treatment cannot easily be made at the current time. Anyone in active treatment should raise it with their oncology team, because interactions there are not hypothetical.

What the evidence says about melatonin and fertility treatment

Melatonin is given during IVF, or in-vitro fertilisation, on the theory that it protects the egg from oxidative damage as it matures.

A 2025 review of assisted reproduction found the theory holds and the outcome does not. Melatonin supplementation may not significantly improve clinical pregnancy or live birth rate, which are the two results anyone going through fertility treatment is actually counting.

What did improve was the laboratory picture. Melatonin appears to have a positive effect on oocyte and embryo quality, particularly in women with polycystic ovary syndrome, where it increased the number of mature eggs collected.

This is the same shape as the blood sugar section and as half the supplement aisle: a measurable improvement in something on the way to the outcome, and no improvement in the outcome. Better-looking embryos are not more babies until a trial shows they are.

The reviewers ask for exactly that: further well-designed, large-scale studies are needed before it can be recommended for routine use in clinical practice.

What a sleep lab saw in children

A separate 2026 analysis matched 342 pairs of children at a pediatric sleep clinic. Melatonin users had a lower median percentage of REM sleep than matched non-users, 16.7% against 19.0%, while total sleep time, sleep efficiency, non-REM sleep stages, respiratory indices and ten other measures did not differ.

Everything was measured at one point in time, so it cannot separate the drug from the reason it was prescribed. It is the first large check of melatonin against objective sleep architecture in children, and it found something rather than nothing.

Myths about melatonin, and what the evidence says

“It is a natural sleeping pill.” It is a hormone and it is not a sedative. In ordinary insomnia it takes about five minutes off sleep onset.

“More is better.” Signals do not work that way. Trials that succeed generally use small doses, and the retail market sells much larger ones.

“Take it at bedtime.” For a clock problem, bedtime is often too late. Timing relative to your own rhythm is the active ingredient.

“It is harmless because you can buy it anywhere.” Over 71% of products missed their stated content by more than 10%, and a quarter contained serotonin, which nobody chose to buy.

“It works for everyone with sleep trouble.” It works best where the fault is timing. For insomnia driven by pain, anxiety, apnea or a broken routine, it is treating the wrong thing.

Dosing and forms of melatonin

The trials that work sit roughly between 0.5 and 5 mg. Products commonly sell 10 mg, which is far above physiological levels and has no better evidence behind it.

Timing is the part most people get wrong. For shifting a late clock earlier, melatonin is generally taken several hours before the target bedtime, not at it. For jet lag the timing depends on direction of travel. Getting this wrong can push the clock the way you do not want.

Prolonged-release forms are what most of the adult insomnia evidence used, and they are a prescription product in many countries. Immediate-release gummies and tablets are a different product with a different release profile.

Safety, side effects and who should be careful

Short-term use is generally well tolerated. Headache, daytime grogginess and vivid dreams are the common complaints, and grogginess usually signals a dose that is too high or too late. On the liver specifically, despite wide scale use, melatonin has not been convincingly linked to instances of clinically apparent liver injury.

The product itself is the bigger safety story on this page. Melatonin content did not meet label within a 10% margin of the label claim in more than 71% of supplements, and lot-to-lot variability within a particular product varied by as much as 465%. That means the same bottle can behave differently between batches.

Worse, an additional 26% were found to contain serotonin, identified at levels of 1 to 75 μg. Serotonin is a controlled substance in some places and interacts with antidepressants. Nobody is buying melatonin in order to take an unlabeled second compound.

Long-term nightly use in healthy adults is untested. Anyone pregnant or breastfeeding, anyone on sedatives, anticoagulants or blood pressure medication, and any parent considering it for a child should speak to a clinician first.

Interactions

The ones that matter are sedatives, where effects add up, and anticoagulants, where bleeding risk may rise.

Because of the serotonin contamination finding, anyone taking an antidepressant that acts on serotonin has an additional reason to be careful about product choice, quite apart from melatonin itself.

Blood pressure and diabetes medications are also commonly listed, and a pharmacist can settle it in a minute.

Bottom line on melatonin

If your body clock is in the wrong place, from shift work, travel or a genuinely delayed sleep phase, melatonin is a reasonable and well-evidenced tool, and the skill is in when you take it rather than how much.

If you sleep at normal hours and simply sleep badly, expect about five minutes. That may still be worth it to you, but buy it knowing the number.

If it is for a child, that is a medical conversation.

And whatever the use, the product is the weak link. In a category where seven in ten bottles miss their own label and a quarter contain something else entirely, choosing a brand that tests its batches is not fussiness.