Melatonin side effects:
what was measured in people who stopped taking it
Two questions almost always arrive before all the others: is it addictive, and does it make you gain weight. Both have a measured answer, and neither looks like the headlines. Dependence is not a feeling, it is three things tracked separately, and the longest study that tracked them lasted a year and ended with two weeks in which the researchers looked specifically for what happens on stopping. Weight was measured in seven studies, and the pooled result straddles zero. The serious part of this article is neither of the two.
“Dependence”
covers three different things
When someone asks whether melatonin is addictive, they are really asking three things at once, and in pharmacology each is tracked with different tools.
The first is tolerance: the same dose stops doing what it did, and you need more for the same effect. It shows up in the dose, followed month by month. The second is withdrawal symptoms: what appears in the days after the last evening. It is measured with symptom checklists applied before and after, not from memory. The third is loss of control: you take it because you cannot stop, not because you need it. It is assessed with questionnaires about the pattern of use.
The three do not necessarily go together. You can have tolerance without withdrawal, and the other way round.
Melatonin reaches two receptors of its own, MT1 and MT2. A paper published in 2026 in Nature Communications, which solved the structure of one of them at a resolution of 3.0 angstroms, describes them as G protein coupled receptors that respond to a hormone the body secretes on a 24 hour rhythm. In other words, the substance arrives at a place the body uses anyway, day after day.
The practical consequence shows in how doctors are advised to stop each thing. A summary of insomnia recommendations, published in 2026 in a German medical journal and built on a Delphi consensus, splits substances into two groups. In the first, those that can be stopped abruptly because they do not cause withdrawal symptoms, and melatonin is there. In the second, benzodiazepine receptor agonists, antidepressants, antipsychotics and gabapentinoids, which after more than two weeks are tapered by 10 to 25% a week, and after a year of use or at high doses are tapered even more slowly.
A word that describes three different things is not an answer, it is shorthand. The question that has an answer is: what was measured in people who stopped taking it.
| What is tracked | How it is measured | What was found for melatonin |
|---|---|---|
| Tolerance the same dose no longer does what it did | Dose followed month by month, on continuous treatment | No signs of tolerance over twelve months, in an open label study of 244 adults |
| Stopping what appears in the days after | Symptom checklists applied before and after, plus sleep reported daily | No withdrawal symptoms and no worsening of sleep in the two weeks after |
| Control taking it without being able to stop | Questionnaires about the pattern of use | Not measured in the studies above, so neither confirmed nor ruled out |
| The body's own production does the body still make melatonin? | A melatonin metabolite, measured in night-time urine | No drop in the body's own production after twelve months of use |
First, how long the dose is tapered for, according to the recommendations for doctors. Then, and this is the point of the chapter, how long the window lasted in which someone actually looked for something.
The tapering durations come from the 2026 summary of recommendations. The study duration and the window after stopping come from the open label study of 244 adults, published in 2011 in Therapeutics and Clinical Risk Management.
The study that actually measured stopping
lasted a year and ended with two weeks of looking
In 2011, five researchers published the results of a long follow-up in Therapeutics and Clinical Risk Management. The interesting point is not that melatonin helped with sleep, but what happened after people stopped.
There were 244 adults with long-standing sleep problems, who had come from an earlier double-blind dose-finding study. They took 2 mg of prolonged-release melatonin every evening for six to twelve months. Of the 244, 36 dropped out along the way, 112 completed six months and 96 completed twelve. Only then did the part that matters begin: after the last evening, two weeks of follow-up.
Four things were tracked in those two weeks. Sleep, reported daily by each person in a diary. Side effects, vital signs and laboratory tests, at every visit. A symptom checklist built specifically for withdrawal, with 84 questions about physical state. And the fourth, the one nobody usually includes: night-time urinary excretion of a melatonin metabolite, which is how you indirectly measure how much melatonin the body is still making on its own.
The result, in the authors' words: there was no evidence of tolerance to the treatment. Stopping was accompanied neither by worse sleep nor by withdrawal symptoms but, on the contrary, by a residual benefit. And the body's own melatonin production was not suppressed, even after twelve months of continuous use.
This answers directly the sentence repeated on almost every Romanian page about melatonin: that regular use teaches the body to rely on an outside supply and to make less. That sentence circulates with no source at all. The study that measured it measured it in urine, and production did not fall.
And now the honest part, because a good study is not a complete study. It was open label, meaning everyone knew what they were taking, and sleep diaries carry that with them. It used a single form, 2 mg prolonged release, which is not the same as a higher dose released all at once. It was in a single adult population. And 36 of the 244 dropped out, and we do not know why. It is the most careful measurement of stopping that exists, and it is one.
The weight question
has an answer that straddles zero
Here the measurement is smaller and simpler to tell.
In 2017, Current Pharmaceutical Design published a systematic review with meta-analysis that gathered the controlled human studies in which melatonin was given and weight was measured. Seven studies came out, all parallel, all run in clinics, with 244 patients in total.
The authors pooled the differences with a random effects model. The point estimate came out at 0.09 on an effect size scale, with a 95% confidence interval running from minus 0.17 to plus 0.34. Heterogeneity came out at zero, I squared equal to 0.0%, with p equal to 0.66.
Seven controlled studies, 244 people in total. The point estimate of the pooled difference is 0.09, and the interval runs from minus 0.17 to plus 0.34. The square is the estimate, the bar is the interval, the dotted line is zero. When the interval crosses zero, the measurement cannot tell more, less and nothing apart.
Two things to read from these figures, and they are different.
First: the interval crosses zero, and crosses it in both directions. The authors say so directly: none of the included studies found a strong, significant effect on body weight, and among them some reported a decrease and others an increase. The measurement cannot tell more, less and nothing apart.
Second: zero heterogeneity means the studies did not contradict one another. That is unusual, and it makes the null result more informative than it looks at first sight, because it is not an average between two camps.
What it is not: seven studies and 244 people are a small measurement, none was designed with weight as the main question, and the people in them were clinic patients also on other treatments. So the correct sentence is that nobody measured a change in weight, in either direction, and that the measurement that exists is small. That is not the same as “proven not to”, and nowhere near “it makes you gain weight”.
What was actually reported,
across thirty-seven studies
In 2019, CNS Drugs published a systematic review with a good inclusion rule: it took only randomised, placebo-controlled studies that reported both the types and the frequencies of side effects. Many studies report only that it “was well tolerated”, and those say nothing. That left 37.
Daily doses ranged from 0.15 mg to 12 mg. People were followed for up to 29 weeks, but most studies lasted four weeks or less.
The most frequently reported, in order: daytime sleepiness, 1.66%. Headache, 0.74%. Other sleep-related effects, 0.74%. Dizziness, 0.74%. Lowered body temperature, 0.62%. Very few effects were considered serious or clinically significant, and those reported were agitation, fatigue, mood changes, nightmares, skin irritation and palpitations. Most faded on their own within a few days, with no dose adjustment, or straight after treatment was stopped.
And now the thing worth reading twice: the most frequent side effect of a substance taken to fall asleep faster is sleepiness during the day.
It is not an irony, it is the mechanism arriving at the wrong time. Melatonin is not a sedative, it is a time signal. In a study published in 2025 in the International Journal of Molecular Sciences, 66 healthy volunteers, 33 women aged 24 to 38 and 33 men aged 24 to 44, gave blood at 02:00, 08:00, 14:00 and 20:00. Melatonin followed a clean 24 hour rhythm, with a night-time peak at 02:00 of around 98 picograms per millilitre and a daytime low at 14:00 of around 9. A ratio of roughly eleven to one between night and the middle of the day.
If that signal lands too late, it lands in your morning. That is why the timing matters more than the dose, and the light in your home before bed shifts the same signal without you taking anything.
One last note of honesty about the percentages above: the authors rated the 37 papers themselves against risk of bias criteria, and only eight came out good overall, 16 fair and 13 poor, largely because almost half were published before reporting guidelines existed. So these are the best figures we have, not precise figures.
First, how often each effect was reported across the 37 studies. Press the second button to switch the measurement completely, from what was felt to how much of the swallowed milligrams reaches the blood.
The frequencies come from the 2019 review of 37 randomised studies. The absorption percentages come from the 2015 pharmacokinetics review, which gathered 22 studies with doses between 0.3 and 100 mg.
The same milligram,
a different blood level
In 2015, the European Journal of Clinical Pharmacology published a systematic review of the pharmacokinetics of melatonin in humans. It gathered 22 studies, with doses between 0.3 and 100 mg, given by mouth or into a vein.
Two figures are stable. After an immediate-release form taken by mouth, the blood peak comes at about 50 minutes. The half-life is around 45 minutes, by either route. That is a short window, one that does not cover a whole night.
The third figure is the one that changes the discussion. Oral bioavailability, meaning the percentage of what you swallow that actually reaches the circulation, averaged around 15%, but ranged from 9 to 33% from one study to another. And peak concentration, total exposure, clearance and volume of distribution varied, the authors write, extensively between studies.
What moved the figures, according to the same review: age, coffee, smoking, oral contraceptives, whether or not you had eaten, and fluvoxamine. In critically ill patients, absorption was faster and elimination impaired.
Fluvoxamine is the only one of these with a figure of its own, and it is a big one. In a study published in 2000 in Clinical Pharmacology and Therapeutics, five healthy volunteers received 5 mg of melatonin, once on its own and once alongside 50 mg of fluvoxamine. With fluvoxamine, total exposure to melatonin was on average 17 times higher, and peak concentration 12 times higher. The terminal half-life did not change significantly, which told the authors that no more melatonin was being made, but that the route by which it is cleared had been blocked.
Coffee and smoking appear on the same list because they pass through the same place in the liver. Not by a factor of 17, but not by zero either.
| What changes the blood level | In which direction | How well it is measured |
|---|---|---|
| Fluvoxamine | Total exposure 17 times higher, peak 12 times higher | One study, five healthy volunteers, single dose |
| Coffee | Higher level, no pooled figure | Flagged as a factor in the pharmacokinetics review |
| Smoking | Lower level, no pooled figure | Flagged as a factor in the same review |
| Oral contraceptives | Higher level, no pooled figure | Flagged as a factor in the same review |
| Food | Changes the timing and how much is absorbed | Flagged as a factor in the same review |
| Age | Changes both the level and how quickly it is cleared | Flagged as a factor in the same review |
Translated into what you do with the information: the figure on the box is what you swallow. How much reaches your blood depends on your age, your coffee, smoking, your evening meal, whatever else you take, and a personal variation the review could not narrow down. That is why “this many milligrams worked for my colleague” is not information about you. It is the same kind of arithmetic that the magnesium figure on a label needs: the number written and the number that matters are not always the same number.
The risk that is rising,
and it is not in adults
Up to this point, the article has been largely reassuring. This chapter is not, and it has the most solid figures in the whole text.
In 2022, the US Centers for Disease Control and Prevention published an analysis of data from the national poison data system. Between 1 January 2012 and 31 December 2021, 260,435 ingestions of melatonin alone were reported in people aged up to 19. The annual number rose by 530%. In 2021, melatonin accounted for 4.9% of all paediatric ingestions reported to poison centres, compared with 0.6% in 2012.
Hospitalisations and more serious outcomes also rose over those ten years, and the increase came mostly from unintentional ingestions in children aged five and under. Five children needed mechanical ventilation. Two died.
The reason is not that melatonin became more dangerous between 2012 and 2021. It is that much more of it ended up in many more homes. The same analysis notes that US sales rose from 285 million dollars in 2016 to 821 million in 2020, and that by 2020 melatonin had become the most frequently ingested substance of all those reported in children.
It is the only chapter of the article that asks for an action, not reassurance. The box goes up high, closed, where a child cannot reach it. And a child does not get melatonin without a doctor, not 5 mg, not 1 mg.
When melatonin is not a good idea, or not on its own. Our gummies contain 5 mg per serving, and the serving on the box should not be exceeded. After taking it, do not drive or operate machinery: the most frequently reported effect is sleepiness, and it does not check the clock. If you are pregnant or breastfeeding, or if it is for a child or a teenager, ask your doctor first. If you take an anticoagulant, an immunosuppressant, an antiepileptic, an antidepressant or any long-term treatment, ask your doctor or pharmacist, because the level of melatonin in the blood can be shifted a lot by a single medicine taken alongside it. And if you already take a medicine for sleep, do not stop it or replace it with a supplement on your own. Keep out of the reach of children.
Where our gummy
fits
The figures on the box first, because they are the only thing a manufacturer actually guarantees. A serving is two gummies and contains 5 mg of melatonin, 200 mg of chamomile extract at a 25 to 1 ratio and 200 mg of lavender extract at a 10 to 1 ratio. The box holds 60 gummies, which is 30 servings. The flavour is blackberry. Made in Romania.
The sentence that can be said simply is this one: melatonin contributes to the reduction of time taken to fall asleep.
The article above is about something else, about what was measured on stopping, on weight and in reported effects, and that part is not a sales argument, it is context. The gummy is a form of melatonin, not an answer for a whole night: its window in the blood is short, a matter of hours, not of the whole night through to morning.
In practice: take it shortly before bed, not at three in the morning when you have woken up, because then the signal lands in the morning that follows. If your evening also has a problem with restlessness, not just with timing, L-theanine is another chapter, and we have written it separately.
Frequently asked questions
about melatonin side effects
“Dependence” covers three things that are measured separately: needing ever larger doses, symptoms on stopping, and losing control over use. The longest study that tracked them followed 244 adults for six to twelve months and ended with two weeks in which researchers looked specifically for what appears after stopping. Neither a need for higher doses nor withdrawal symptoms appeared. It is a single open label study, of a single form of melatonin.
In the seven controlled studies in which people were also weighed, the pooled difference was 0.09 on an effect size scale, with an interval running from minus 0.17 to plus 0.34. In other words, the interval crosses zero, and crosses it in both directions. The authors write that none of the included studies found a strong, significant effect on body weight. That is 244 people in total, so a small measurement.
In the one year study, stopping was done directly, and the two weeks after were precisely the window in which symptoms were looked for. A summary of recommendations for doctors, published in 2026, places melatonin in the group of substances that can be stopped abruptly, and benzodiazepine receptor agonists in the group that is tapered, by 10 to 25% a week. If you already take a medicine for sleep, do not stop it or replace it on your own, ask your doctor.
Daytime sleepiness, reported by 1.66% of participants in the 37 randomised studies gathered in a 2019 review. Next come headache, at 0.74%, other sleep-related effects, at 0.74%, dizziness, at 0.74%, and lowered body temperature, at 0.62%. Most faded on their own within a few days, with no change in dose, or straight after stopping. The most common side effect is, in fact, the mechanism arriving at the wrong time.
This is the question worth taking to a doctor or pharmacist, not settling on your own. In a study of five healthy volunteers, 5 mg of melatonin taken alongside 50 mg of fluvoxamine gave a total exposure 17 times higher and a peak concentration 12 times higher. The same pharmacokinetics review also finds that age, coffee, smoking, oral contraceptives and whether or not you have eaten change the blood level.
Not without a doctor, and this part has the most serious figures in the whole article. Between 2012 and 2021, US poison centres recorded 260,435 melatonin ingestions in people aged up to 19, with a 530% rise in the annual number. The increase came mostly from unintentional ingestions in children aged five and under. Five children needed mechanical ventilation and two died. Keep the box where a child cannot reach it.
Comments (0)
There are no comments for this article. Be the first one to leave a message!