Melatonin: Dangers, Side Effects and Contraindications, According to Research

★ OUR FORMULA

Optimal Sleep

An evening formula combining 1.9 mg of melatonin with four plants : 100 mg lemon balm, 100 mg hops, 50 mg passionflower, and 50 mg valerian extract. One vegetable capsule daily with a large glass of water, 45 minutes before bedtime.

View Optimal Sleep

Melatonin is a hormone secreted at night by the pineal gland: it transmits the darkness signal and helps regulate the sleep-wake cycle. When taken as a supplement, its oral bioavailability is around 15%, its blood peak occurs approximately 50 minutes after an immediate-release form, and its half-life is about 45 minutes, according to a review of 22 studies; age, caffeine, tobacco, the pill and certain medications modify these parameters.

Immediate-release produces a brief peak; extended-release spreads absorption over several hours and has been studied primarily at 2 mg in those over 55 years old. The results of one do not necessarily apply to the other. <<<17>>> étale la diffusion sur plusieurs heures et a surtout été étudiée à 2 mg après 55 ans. Les résultats de l'une ne valent pas forcément pour l'autre.

The benefit is real but modest. A meta-analysis of 19 studies and 1,683 participants found faster sleep onset by 7.06 minutes and longer sleep by 8.25 minutes compared to placebo, with no loss of effect over the course of treatment. A dose-response meta-analysis of 26 trials places the peak effect at around 4 mg per day and suggests that taking it approximately 3 hours before bedtime could be more effective than the most common regimen, 2 mg 30 minutes before.

Our Optimal Sleep provides 1.9 mg of melatonin per capsule, a dose within the range studied in trials, with lemon balm, hops, passionflower and valerian root extract dosed as complementary ingredients, at levels lower than those in most studies of these plants; no results obtained with other doses or forms are attributed to it here.

The reference review compiled 37 randomized controlled trials versus placebo in sleep disorders, from 0.15 to 12 mg per day, over 29 weeks at most and most often 4 weeks or less. The most commonly reported adverse effects were daytime somnolence (1.66% excess compared to placebo), headaches, other sleep-related effects and dizziness (0.74% excess each), followed byhypothermia, a drop in body temperature (0.62% excess). These excesses are calculated from 1,625 treated participants, and no trial found a significant difference with placebo. Most disappeared within a few days without dose adjustment, or upon discontinuation; no effect threatened vital prognosis. However, only 8 trials were of good quality, 16 moderate and 13 poor.

For high doses, a meta-analysis of doses of 10 mg and higher in adults over 30 years old reviewed 3,861 participants and 79 studies, of which 37% reported nothing on adverse effects. In the 4 trials at low risk of bias, melatonin did not detectably increase either serious effects (hazard ratio 0.88) or discontinuations due to poor tolerance (0.93), but it increased by approximately 40% effects such as drowsiness, headaches and dizziness (ratio of 1.40), with very wide intervals remaining for serious effects and discontinuations.

A critical review of 50 controlled studies found 24 with at least one statistically significant adverse effect, especially on fatigue, mood and psychomotor performance. In monitoring of 653 patients on extended-release over 3 weeks, without placebo, nausea ranked first (1.5%), with no serious effects.

In the review of 37 trials, nightmares rank among the rare adverse effects considered notable, along with agitation, fatigue, mood swings, skin irritation and palpitations. Vivid dreams and nightmares are counted with other sleep-related effects: 21 melatonin participants versus 9 placebo in 6 trials, representing 0.74% excess. These effects disappeared within a few days or upon discontinuation. Verdict: rare. In case of recurrent nightmares, stopping melatonin allows verification of whether it is the cause (practical advice).

Regarding mood, a two-year follow-up in 80 children and adolescents, primarily autistic, on extended-release formulation, reports mood swings related to treatment in 4.2% of the 95 open-phase participants, fatigue and sleepiness in 6.3% each; this uncontrolled follow-up included manufacturer employees among its authors. In adults, a review of 10 studies and 486 patients, at 0.5 to 6 mg daily for 2 weeks to 3.5 years, shows no net effect on depression: in six studies, scores improved under placebo as well, without significant difference, and no serious adverse effects were reported.

Trials that investigated dependence did not find it.

In 40 patients aged 55 and older, randomized between extended-release and placebo for 3 weeks, the 3-week discontinuation produced no rebound. In the monitoring already cited, rebound insomnia (sleep worse than before treatment) affected 3.2% of patients within two days of stopping and approximately 2.0% around two weeks later; this uncontrolled study was also linked to the manufacturer. In children, after two years of treatment, two weeks on placebo reduced sleep scores, which remained better than at baseline, without problems related to discontinuation.

The review of 37 trials acknowledges it: due to the lack of long-term trials, conclusions regarding continuous use remain limited. The longest available data consist of a 6-month trial in adults, a 2-year follow-up in 80 children, and a follow-up of 51 children treated for an average of 3.1 years; beyond that, mostly observational studies.

At the end of 2025, a medical records study presented at a cardiology congress without publication in a peer-reviewed journal associated prolonged use in insomniac adults with increased heart failure at 5 years, hospitalizations for heart failure, and all-cause mortality. An analysis published in 2026 highlights its limitations: melatonin users sleep poorly, are more often depressed, and consult more frequently—factors that are themselves linked to cardiac risk.

Verdict on heart failure: observational signal, causality not established. Regarding arrhythmia, a 2026 publication describes a 63-year-old man taking 10 mg daily presenting with palpitations, premature contractions, and at night a second-degree atrioventricular block, which resolved after discontinuation; the systematic review identified only two other publications, on three men taking 1 to 10 mg daily, and causality remains to be established. Verdict: rare. Liver damage has been described only at very high doses: at 300 mg daily, in multiple sclerosis, three of the four treated participants had elevated liver enzymes, which resolved upon discontinuation and reappeared upon rechallenge in one of them.

In the United States, 260,435 melatonin ingestions in individuals aged 19 years or younger were reported to poison control centers between 2012 and 2021, an annual number rising from 8,337 to 52,563 (+530%). The ingestion was unintentional in 94.3% of cases, and 83.8% of children were 5 years or younger. The vast majority remained without serious consequences, but 5 children required ventilation and two died, aged 3 and 13 months; the authors could not establish whether melatonin alone was responsible. Most hospitalized children were adolescents who intentionally ingested melatonin, but the steepest increase in hospitalizations concerned children 5 years or younger following accidental ingestion.

Between 2019 and 2022, these ingestions accounted for 7.1% of emergency department visits for unsupervised medication ingestion in children 5 years or younger; at least 10 units were accessible in 35.8% of solid form ingestions, and nearly one time in two, the product was intended for adults or family use. In a regional study, melatonin gummies also led to emergency care more often than other gummies (odds ratio of 4.8), with the dominant symptoms being drowsiness, digestive disorders, and hyperactivity.

Another uncertainty: actual content. A Canadian study measured melatonin in 31 supplements: the measured quantity ranged from −83% to +478% of the label, more than 71% of products fell outside a 10% margin, and content varied up to 465% from one batch to another. Eight products contained serotonin, from 1 to 75 µg. The authors of the poison control centers report note that the strongest variation involved a chewable form, though the original study estimated the variability to be independent of the manufacturer and product type.

To help a child sleep, the benefit of melatonin is debated: a meta-analysis of 8 trials in 419 children and adolescents aged 5 to 20 years suffering from chronic insomnia without other illness found 30 minutes additional sleep and 18 minutes faster sleep onset, but more non-serious adverse effects (risk ratio of 3.44), with low to very low certainty. Its authors conclude that it should never be a first-line choice, and short-term, monitored use should only be considered after non-pharmacological approaches have failed.

For puberty, the concern raised in prepubertal children has not been confirmed to date. In 80 children treated for up to 104 weeks with 2, 5, or 10 mg, weight, height, and BMI remained within age-appropriate norms, as did pubertal stage examined by a physician in 31 of them. In 51 children treated for an average of 3.1 years at 2.69 mg, pubertal stage did not differ from that of Dutch children of the same age. A Dutch follow-up 9 to 12 years later, cited by the review of 37 trials, reports that 31.3% of 33 respondents reported delayed puberty, data to be interpreted cautiously according to the authors. These studies remain small, and the review of 37 trials judges the data insufficient: verdict theoretical.

In a large American survey, approximately 4% of pregnant women aged 18 to 40 reported having taken melatonin in the past year. Yet, a review of 15 human studies found no randomized trials on its use for sleep disorders during pregnancy or breastfeeding. Melatonin crosses the placenta rapidly: in a perfused human placenta model, maternal and fetal concentrations equilibrated in 150 minutes.

Three small trials conducted during pregnancy for other reasons, at 8 to 30 mg per day, reported no safety concerns; concerns stem from animal studies, such as decreased birth weight in sheep. During breastfeeding, natural melatonin in milk reaches approximately 35% of its blood concentration; only one case has been described, that of an 18-month-old breastfed child presenting with bleeding, whose platelet aggregation normalized 3 months after the mother stopped taking melatonin occasionally.

The authors consider use probably safe, without firm conclusions possible. Verdict: theoretical risk, which the data do not allow us to rule out. No melatonin during pregnancy or breastfeeding without the advice of a doctor or midwife; our Optimal Sleep, like our Crystallized Glycine, is not intended for pregnant or breastfeeding women.

Melatonin is metabolized by an enzyme, CYP1A2: the pill, which inhibits this enzyme, and fluvoxamine strongly increase circulating melatonin. In 5 male volunteers, 50 mg of fluvoxamine combined with 5 mg of melatonin multiplied by an average of 17 the total amount of melatonin in the blood and by 12 its peak. In 29 volunteers who took 6 mg, exposure was 4 to 5 times higher in oral contraceptive users, with no notable effect on their alertness.

Other interactions add to or counteract effects. In 16 volunteers aged 55 and older, extended-release melatonin alone did not impair any performance, but it worsened the psychomotor, memory, and driving disturbances caused by zolpidem, a sleeping pill. In 47 hypertensives well controlled by nifedipine, 5 mg of melatonin in the evening for 4 weeks increased systolic blood pressure by 6.5 mmHg and diastolic by 4.9 mmHg over 24 hours, and heart rate by 3.9 beats per minute.

For anticoagulants, a review on jet lag reports isolated observations suggesting a risk with warfarin. For alcohol, tranquilizers, and other sleeping pills, with no dedicated study found, the addition of sedative effects remains theoretical, but plausible in light of the trial with zolpidem. For immunosuppressants, the risk is theoretical as well, with no clinical interaction described. On caffeine itself, see our article caffeine and sleep.

Epilepsy. Isolated observations suggest a risk in people with epilepsy, and review of 37 trials judges the data insufficient to evaluate it. Conversely, a meta-analysis of 10 trials where melatonin was added to epilepsy treatment found improved sleep onset and less severe seizures than with placebo, but several trials did not systematically evaluate tolerability. Verdict: theoretical, contradictory data ; melatonin should be discussed with your neurologist.

Diabetes. In 21 healthy young women, 5 mg of melatonin taken 15 minutes before a 75 g glucose load increased the area under the glucose curve by 186% in the morning and 54% in the evening compared to placebo. In 17 women, the morning effect was six times greater in carriers of a common variant of the MTNR1B gene, linked to type 2 diabetes risk. These trials involve a single dose in non-diabetics: they suggest not taking it just before eating and discussing it with your doctor if you have treated diabetes. Our article on elevated blood sugar details the studied interventions.

Hypertension. The trial with nifedipine shows that evening melatonin may interfere with some antihypertensive medications: if on treatment, monitor your blood pressure after starting melatonin; our guide blood pressure 14/9, 15/9, 16/10 helps you read the numbers.

Autoimmune diseases. The concern stems from melatonin's possible action on immunity. In multiple sclerosis, a review of 10 studies and 435 participants found no reduction in relapses, disability, or MRI activity, with low to very low certainty, and adverse event reporting remained incomplete. Verdict: theoretical ; with immunosuppressants, specialist advice remains the rule.

In the zolpidem trial, 2 mg of extended-release melatonin taken alone at bedtime did not impair psychomotor performance, memory, or driving one and four hours later or the next morning. In 40 insomniacs of the same age, daytime psychomotor performance was even better than with placebo.

These results do not cover everything: daytime sleepiness remains the most commonly reported adverse effect, and 3 mg may leave melatonin elevated throughout the day. To our knowledge, high immediate-release doses taken before driving have not been subject to dedicated clinical trials.

For chronic insomnia, the best-supported option is not a supplement. According to a meta-analysis of 20 trials and 1,162 adults, cognitive behavioral therapy for insomnia, which combines bed time restriction, sleep hygiene, and relaxation, shortens sleep onset by 19 minutes and nighttime awakenings by 26 minutes, and improves sleep efficiency by nearly 10%, with effects that appear to persist and no reported adverse effects. It is practiced with a physician or trained psychologist.

Among supplements without melatonin, glycine, an amino acid, takes a different approach. A 2026 review identifies 3 small trials on glycine and sleep in humans, all reporting a benefit on at least one sleep-related measure.

★ OUR FORMULA

Crystallized Glycine

100% pure crystallized glycine, in powder form, with a naturally sweet taste and without melatonin. A dose of 5 g per day, at any time, dissolved in a large glass of water or a hot or cold beverage, for a one-month course that can be renewed.See Glycine

The trials used 3 g of glycine in the evening; our

Crystallized Glycine is dosed at 5 g per day, without melatonin, and is not recommended during pregnancy and breastfeeding. Our guides on glycine before bedtime and its side effects detail the evidence; see also our articles on sugar and sleep and on saffron 12.

Melatonin: proper use in 6 points

For an adult with no contraindications, our Optimal Sleep provides 1.9 mg of melatonin per capsule, 45 minutes before bedtime; it is not recommended before age 12, during pregnancy or breastfeeding, and without medical advice when taking a treatment that could interact. See also our article magnesium morning or evening.

Frequently Asked Questions

Sources

Learn more