You may have taken it most evenings for months and still not be able to say, hand on heart, whether it helps. Some nights you sleep and some nights you don't, and underneath runs a half-worry that a hormone every night must be doing something to you, or that you couldn't sleep now without it. And there's a faint sense, too, that needing something every night is a small failure of your own.

The longest trial speaks to that worry. People with insomnia took a prolonged-release tablet, the kind that lets its dose out slowly, every night for six months. Their side effects, blood tests and hormone measures looked much like those of people on a dummy pill4, and the tablet showed no sign of wearing off. Six months is as far as the evidence reaches, and there isn't good research on years12 of nightly use.

Melatonin is the message your brain sends as the dark comes on, telling the rest of you what hour it is, and it doesn't push you under the way a sleeping pill does. So when you take it matters far more than how much. Once you see it that way, you can start to tell whether your own trouble is the kind it can answer.

What is melatonin doing in my body?

Think of the hour you usually reach for the tablet, when the light outside is going. As it goes, your body clock sets the brain making melatonin, and the rising level tells the rest of you that the day is done25. In the morning, light switches it off again. The steep climb into sleepiness tends to come about two hours after5 your own melatonin starts. A tablet shows its effect most clearly when your own level is low, as it is around midday, and does least once the evening rise is under way5. Swallowed at lights-out, then, a large dose joins a rise that's already under way, and probably tells your body something it already knows.

In one laboratory study, twenty healthy men took melatonin just before noon, when their own level was low, and they grew sleepier at every dose from 0.1 to 10 milligrams6. A later review, written by a researcher tied to a melatonin maker, finds that it isn't sedating5 in the way sleeping pills are, which can sound like the opposite. Read side by side, the two papers fit the hour: melatonin seems to make people sleepy when the body's own supply is low, and to add little once the evening has already done that work.

If you take your tablet beside a bright laptop or under the kitchen lights, that light may be working against it. In a laboratory study of 116 healthy young adults, ordinary room light in the hours before bed delayed the start of their melatonin in ninety-nine percent of them7, and cut about ninety minutes from its nighttime run compared with dim light. There's more on that in how light affects sleep.

The sun descending in the west, The evening star does shine, The birds are silent in their nest, And I must seek for mine,
Night (Songs of Innocence), by William Blake lines 1–4

Why would the timing matter more than the dose?

The clearest picture comes from a laboratory where thirty-four healthy adults lived in dim light and took one pill a day at a set hour, so the researchers could watch when melatonin moved their clocks. A 0.5-milligram dose moved the clock earliest when it came in the afternoon, two to four hours before their own melatonin began to rise8. Taken at bedtime it did relatively little to the clock, and taken near the end of sleep it could push the clock later.

When the same study compared doses, each taken at its own best time, 0.5 milligrams moved the clock by about the same amount8 as 3 milligrams, with the smaller dose a little later in the day. The researchers suggest the smaller one wherever the sleepiness of 3 milligrams would get in the way. They also thought 3 milligrams somewhat more likely to bring a big shift, so a bigger dose isn't wasted every time. If you've drifted from 1 milligram to 5 and then to 10, each step up has likely changed little about what your clock hears. The hour you take it changes a great deal more.

When a small dose moves the clock, counted from your own evening rise

The hours are counted from each person's own melatonin rise, which comes at a different hour for everyone. The study was of thirty-four healthy adults in dim light. A 0.5-milligram dose moved the clock earliest two to four hours before that rise, and did relatively little from the rise through most of sleep8. That second stretch is where sleepiness climbs5 and where a bedtime dose falls. For someone whose clock runs late, the bedtime they want comes well before their own rise, so a dose an hour before it falls earlier on the chart.

Worked out from the figures in the studies cited

See this as a table
The scale runs from 6 hours before to 8 hours after. Hours are counted before and after your own melatonin starts to rise, which is “The rise” on the scale
WhatFromTo
Your own melatonin is up the rise 8 hours after, and on
A small dose moves the clock earliest 4 hours before 2 hours before
Little effect, through most of sleep the rise 8 hours after, and on
Marked on the figure
AtWhat
2 hours after Sleepiness climbs
7 hours after Middle of sleep

It’s often said

It's a natural sleeping pill, so if one doesn't work, I should take a bigger one.

Why it can feel that way

It's on the shelf beside the vitamins and comes in strengths from 1 to 10. Then one night you take a bigger dose and sleep well, and the dose gets the credit.

What’s going on underneath

Melatonin works mostly by telling your body what time it is, more than by making you drowsy the way a sleeping pill does. At lights-out your own melatonin has usually already begun to rise5, so it's uncertain whether extra adds much. And for moving a late clock, a small dose did as much as a larger one8 when each came at the right hour.

A closer look How did researchers work out when melatonin acts?

The study behind the timing picture followed sixteen men and eighteen women aged 18 to 428. Each took one pill a day for three days, at a fixed hour that differed from person to person, while living in dim light so that room light wouldn't muddle the result. The researchers measured the start of each person's evening melatonin rise in dim light before and after, and called it the dim light melatonin onset.

Plotting how far each clock moved against when the pill was taken gives a curve: earliest shifts from an afternoon dose two to four hours before that onset, little change from around the onset through most of the night, and later shifts from a dose near waking. The biggest shifts were about an hour and a half over the three days. The authors wondered whether any dose above 0.5 milligrams might shift the clock in an all-or-nothing way, which is an idea they offered to be tested and is still unproven.

Will it help with my kind of sleep problem?

What melatonin can do depends on what's keeping you up, and the evidence for each kind of trouble comes from different people. If your nights are short because life leaves too little room for them, between a late finish and an early alarm, melatonin can't add the hours that are missing, and getting through the day after a bad night may be more use to you for now.

  • If you lie awake for a long time, or wake and can't get back to sleep

    You may know the long stretch after lights-out, or the waking in the small hours with sleep nowhere near. When researchers pooled trials of people with sleep problems, those on melatonin fell asleep about seven minutes faster and slept about eight minutes longer3 than those on a dummy pill, and the gain didn't fade as they kept taking it. Seven minutes is real, but at 3am you'd hardly notice it. Looking at its own set of trials, the American Academy of Sleep Medicine suggests not using it for chronic insomnia9, though only as a weak recommendation. If nights like these have gone on for months, CBT for insomnia is the treatment the physicians' and sleep-medicine guidelines1011 both strongly recommend first.

  • If you sleep well enough, but hours later than your life allows

    If you're wide awake at 2am and wrecked at 7am, you may be sleeping well enough at the wrong time, and this is where melatonin's timing comes into its own. In an Australian trial, adults with a diagnosed late clock took 0.5 milligrams an hour before the bedtime they wanted, and went to bed at that hour. After four weeks they were falling asleep thirty-four minutes earlier than on placebo12. Sleep came earlier, but in the people whose clocks were measured, the clock itself hadn't measurably moved12. The researchers put the gain down to melatonin's pull toward sleep, together with the bedtime they kept. Guidelines suggest timed melatonin13 for a late clock, and there's more on teenagers' late nights.

  • If you're crossing time zones or sleeping after a night shift

    Is your body still keeping the hours of home while you lie awake in a hotel room? In ten trials of people flying across five or more time zones, those who took melatonin near bedtime at their destination, between 10pm and midnight, had less jet lag in nine14. About one traveler in two was helped, and doses from 0.5 to 5 milligrams worked much alike. The hour matters here too: early in the day it can make you sleepy and slow your adjustment14. After a night shift the evidence is thin. Small, low-quality trials suggest it may add about twenty-four minutes of daytime sleep15, though it didn't help people fall asleep faster. The fuller plans are in getting over jet lag and sleeping on night shifts.

A closer look Where do these numbers come from?

The seven and eight minutes come from nineteen trials of 1,683 adults and children3 with sleep problems, pooled together, so they describe a mix of people rather than adult insomnia alone. Melatonin also brought a small lift in how good the sleep felt.

The late-clock trial ran for four weeks with 116 adults, 29 on average12, whose own melatonin rise came late for the bedtime they wanted. A little over half of those on melatonin improved meaningfully, against about a quarter on placebo. The clock was measured in forty-three of them, which is why its stillness is an early finding.

For jet lag, people fell asleep faster and slept better after 5 milligrams than after 0.5, and doses above 5 did no better14. After night shifts, the trials were small and used 1 to 10 milligrams, and higher doses did no better15.

Is it different if I'm older?

One large trial asked whether melatonin works differently later in life. Researchers gave 791 adults with insomnia, aged 18 to 80, a prolonged-release 2-milligram tablet or a placebo to take each night, and the people kept sleep diaries. Three weeks in, those aged 65 to 80 were falling asleep about nineteen minutes sooner, while on placebo the gain was under two minutes4. Younger people with low melatonin levels gained nothing like it. Across everyone in the trial, the gains held through six months of nightly tablets. In Europe the tablet is now licensed for short-term insomnia in people 55 and over4.

Researchers don't read that trial the same way. The tablet's maker sponsored it, which is reason enough to want it confirmed. The American Academy of Sleep Medicine looked at trials of the same tablet in adults over 55 and judged its effects below clinical significance9, too weak to recommend. So what might you make of it, if you're older and lying awake? The tablet seems a reasonable thing to try for a set stretch, the way the UK prescribes it, usually for up to thirteen weeks16, with your own nights as the judge. That view is about the prolonged-release product that was studied, and what the trial found may not apply to the gummies on a supermarket shelf.

Later in life melatonin may also stay active for longer216 and leave you drowsy in the day, and for sleep problems in people living with dementia, the academy recommends against it213.

How much do I take, and when?

If you turn your own bottle over, the strength printed on it is probably 3 or 5 milligrams, which is what most US gummies tested in 2022 held in a serving17, often as two gummies. The doses that helped in studies were often far smaller. A dose of 0.5 milligrams moved the clock in healthy adults in the lab8, and it brought sleep earlier for adults with a diagnosed late clock12 in the Australian trial. For jet lag, doses up to 5 milligrams14 helped airline passengers, crews and military personnel flying across five or more time zones.

So what the studies point to depends on your problem. For a late clock, the dose that helped was small and came an hour before the bedtime you want, with that bedtime kept, which suits a tablet that peaks in the blood about fifty minutes after you swallow it18. For jet lag, it goes near the new bedtime at your destination. For ordinary sleeplessness at lights-out the gain is small, and the studies are mixed on whether more adds anything. Trials that ran longer at higher doses found somewhat larger effects3 on how fast people fell asleep and how long they slept. Yet in a small study of thirty adults over 50, 0.3 milligrams left them spending at least as much of their time in bed asleep as 3 milligrams did9. That makes a smaller dose a fair experiment, one your own nights can judge.

A closer look How little melatonin does the body need?

Very little is enough to match the body's own nighttime levels. In the 1994 study of twenty healthy men, doses of 0.1 and 0.3 milligrams brought blood melatonin into the normal nighttime range6, and researchers testing supplements still use that study as their yardstick for a dose the body would make itself.

The study of older adults that compared 0.3 and 3 milligrams was small: fifteen sound sleepers and fifteen with insomnia, all over 509, each spending a week on each dose with their sleep measured in a laboratory. No dose shortened the time it took them to fall asleep.

Across the US, meanwhile, the share of adults taking more than 5 milligrams a day more than tripled20 between 2005 and 2018, though it is still fewer than three in a thousand.

Is what's in the bottle what's on the label?

In the US melatonin is sold as a dietary supplement, regulated less strictly than medicines2, and what's inside can wander a long way from the label. When researchers bought twenty-five brands of melatonin gummies in the US and measured what each one held, they found twenty-two that didn't match their labels17. Some held less melatonin than the bottle said and some more than three times as much. One held no melatonin at all, but 31 milligrams of CBD.

An earlier team tested thirty-one supplements bought in Canada and found gaps wider still. A bottle there might hold 83 percent less than its label or 478 percent more21, and one batch of a product could differ from the next by up to 465 percent. About a quarter also contained serotonin, and paying more or choosing a known brand didn't protect anyone from any of it.

So the label in your hand can't tell you everything, though a few things help you choose. In the US, the USP Verified mark is the most reliable sign of dose and purity. The scheme is voluntary, and at the time of the 2022 count only four melatonin products had it1319, all at 3 or 5 milligrams. Outside the US you may meet melatonin quite differently, because in several countries it's a prescription medicine2. In the UK it's prescription-only16, mainly for short-term sleep problems in people 55 and over.

Will I be able to sleep without it?

It’s often said

I've taken it so long that I can't sleep without it now.

Why it can feel that way

After months of a tablet every night, the odd night you forgot it may well have been rough. Lying awake on one of those, it felt like proof that you need it.

What’s going on underneath

In the six-month trial, people switched to placebo for two weeks had no withdrawal effects and no rebound4: their gains faded, but their sleep stayed better than where it had started. That was one prolonged-release product in one trial, and nobody has looked the same way at high-dose supplements, but nothing in what has been studied suggests that stopping leaves you worse off than before you began.

What side effects do people notice?

The morning after a tablet, some people wake groggy or with a headache, or feel tired, sick or irritable216 through the day. In trials, though, few people noticed much. When reviewers pooled thirty-seven trials, daytime sleepiness was what people reported most, and it came to fewer than two people in a hundred1; headache and dizziness turned up less often still. A few people had nightmares, agitation or mood swings, and for most, whatever they noticed faded within days or once they stopped. If you wake groggy after taking it, please don't drive or use machinery until the feeling has passed.

When is it worth checking with a doctor or pharmacist first?

If you take warfarin or another blood thinner, or have epilepsy, please take melatonin only with your doctor's supervision; the concern comes from case reports213 rather than trials, and it's enough to make that conversation worth having. The antidepressant fluvoxamine keeps the body from clearing melatonin, so levels can climb to about seventeen times22 what the dose would usually give. If you take it, your prescriber can help you decide. If you have diabetes, it's worth a word first too, since a single dose worsened glucose tolerance13 in a study of healthy women.

Taking melatonin alongside herbal remedies for sleep can make you very drowsy16, and alcohol or smoking can stop it working as well. In pregnancy and while breastfeeding, no trials have tested it223 for sleep and the human studies there are have been few and small, so it's one to talk through with your midwife or doctor.

Is it okay to give melatonin to my child?

Your child has been lying awake night after night, and another parent swears by gummies. The jar beside the vitamins starts to look like an answer. The good evidence in children comes mostly from those with autism, ADHD or other neurodevelopmental conditions, under specialist care2. In a thirteen-week trial of 125 children and teenagers, almost all of them autistic, whose sleep hadn't improved with behavioral help, prolonged-release melatonin added about thirty-two minutes of sleep a night24 compared with placebo. The maker sponsored that trial, and what it found applies to children like those in it.

For children generally, the right dose and timing aren't known yet, and neither are the effects over long periods or whether a hormone supplement could affect puberty or other development, because the studies have been small and brief213. The American Academy of Sleep Medicine asks parents to talk to a pediatric health professional first19, since many children's sleep troubles are better helped by changes to schedules, habits and routines. There's more on those in helping your child sleep.

If there's a jar of gummies in the house, what helps most is keeping it as you would any medicine, up high and out of children's reach19, with the lid closed, because small children do find it. Over the decade to 2021, calls to US poison centers about children swallowing melatonin rose 530 percent25, and the biggest rise came among children 5 and under who found it at home and swallowed it by accident. Most of those children, about eighty-three in a hundred, had no symptoms at all.

If you're the parent of a teenager, the worry is a different one. Most of the children admitted to hospital25 over those years were teenagers who had taken melatonin on purpose. If you fear yours might have, please get help without waiting.

What the research says

  • Melatonin is the body's own signal that night has begun, switched on by the body clock and off by light. Well established
  • Timing decides what it does to the clock, and a small dose moves a clock about as far as a larger one when each comes at its best hour. Good evidence
  • Across trials of adults and children with sleep problems, it brings sleep about seven minutes sooner on average and adds about eight minutes, and guidelines don't recommend it for chronic insomnia. Good evidence
  • For a late body clock, a small dose before the desired bedtime, with that bedtime kept, brings sleep earlier. Good evidence
  • For jet lag across five or more time zones, it helps about one traveler in two. Good evidence
  • Taken nightly for up to six months, it showed no tolerance and no withdrawal or rebound on stopping, in one manufacturer-sponsored trial. Good evidence
  • US supplements often hold much more or less melatonin than their labels say. Good evidence
  • Taking it every night for years, or in pregnancy or while breastfeeding, hasn't been well studied. Not well studied yet
  • Researchers disagree about older adults: one large sponsored trial found prolonged-release melatonin helped, while a guideline judged the evidence too weak to recommend. Still debated
  • Good evidence in children comes mostly from those with autism or ADHD; for other children, dose and long-term effects aren't known. Not well studied yet

Most trials ran four weeks or less, and several of the most reassuring findings come from a single prolonged-release product studied by its maker, so they may not apply to high-dose gummies bought off the shelf.

So what about tonight's tablet?

The bottle in your palm this evening answers some kinds of trouble far better than others. If your trouble is a clock that runs late, or a flight across the world, melatonin is a signal worth timing well, in a small dose and at the right hour. If you lie awake for other reasons, it may be giving you a few minutes, and when you want something made for the long nights, CBT for insomnia is there. And whether you take it tonight or leave the bottle in the cupboard, your own melatonin will begin to rise as the light goes, as it has done every evening of your life.