Peptide 101: Melatonin (N-acetyl-5-methoxytryptamine)

Most people meet melatonin as a sleep aid, which undersells it and oversells it at the same time. Its effect on falling asleep is real and small — about seven minutes. Its effect on when your body believes night is, which is the thing that actually moves jet lag and shift work, is the stronger act, and it turns far more on when you take it than on how much. This week the column explains both, then gets into the antioxidant claims and the long list of things melatonin has been tried for outside of sleep.

Melatonin is a hormone your pineal gland releases when it gets dark. It is one of the most studied and most consumed supplements on earth, sold in the US as a dietary supplement and prescribed as a drug in much of Europe and Australia.

It is also not a sedative, and almost everything interesting about it follows from that. A sedative works by how much you take. Melatonin is a time signal — a chemical announcement, sent to nearly every cell you own, that the day is finished. Signals work by when they arrive.

Melatonin is your body's internal sunset. When darkness falls, the pineal gland releases it to tell every cell that it is night. Taking melatonin is like reaching for the dimmer by hand when the automatic one has stopped keeping time.

The technical version: melatonin acts through two G-protein coupled receptors, MT1 and MT2, concentrated in the suprachiasmatic nucleus (the hypothalamic cluster that serves as the body's master clock) and spread through peripheral tissues. MT1 activation promotes sleep onset by quieting the firing of those clock neurons; MT2 is the one that phase-shifts the rhythm — moves the clock earlier or later. Separately from its receptors, melatonin is a direct free radical scavenger, neutralizing hydroxyl radicals and stimulating the body's own antioxidant enzymes (superoxide dismutase, glutathione peroxidase, catalase). It also modulates immune function by enhancing T-helper cell activity, and damps inflammation through NF-kB inhibition.

Hold on to the fact that the receptor work and the antioxidant work are separate jobs. It is the reason the dose conversation splits in two.

Overall strength of the evidence, by our read: Strong Evidence. That grade is earned on the clock and on sleep, and it thins out the further you move away from them. Some of what it rests on:

The question people ask is how much. For the circadian half of melatonin, the question that matters is when.

In 2010, Burgess and colleagues at Rush ran the comparison properly: the same protocol, the same lab, 34 healthy adults, 0.5 mg against the 3.0 mg curve they had already published. They mapped the full phase response curve — how much the body clock moves depending on the hour the pill is taken. The result is the useful one. Each dose, given at its own best time, produced advances and delays of similar size. Six times the dose bought no extra clock movement. What the lower dose changed was the schedule: its optimal window arrived later.

For an advance — pulling your clock earlier, which is what a night owl or an eastward traveler needs — the largest shifts from 0.5 mg came from taking it in the afternoon, roughly 2 to 4 hours before the body's own melatonin starts to rise, or 9 to 11 hours before the midpoint of sleep. That is the opposite of how most people use it, which is a large dose at bedtime.

There is a reason more is not better here. The message is "it is night now." A dose big enough to keep blood levels elevated into the next morning is a message that night is still going, which smears the edge the clock is trying to read. The Cochrane review of melatonin for jet lag landed in the same place from a different direction: daily doses between 0.5 and 5 mg were similarly effective, and doses above 5 mg appeared no more effective at all. The one thing the larger dose did do was put people to sleep faster — a sedative-flavored effect, not a clock effect, and worth separating in your head from the phase shift.

So for the circadian uses — delayed sleep phase, non-24, shift work, jet lag — the lever is the clock time, and finding it depends on where your own rhythm currently sits. That is the conversation to have with a clinician who can work out your phase, which is also what the AASM guideline points at: it recommends strategically timed melatonin and does not identify a dose that outperforms the others.

This is the strongest mechanistic story in the file and the weakest outcome story, and both halves matter.

What makes melatonin unusual as an antioxidant is not that it scavenges free radicals; plenty of molecules do. It is the cascade. When melatonin neutralizes a radical, its own metabolites — AFMK and AMK — are themselves scavengers, so one parent molecule can account for several radicals down the chain, where a classical antioxidant is typically spent in a single reaction. Melatonin is also both water- and fat-soluble, which lets it cross membranes that vitamin C and vitamin E are restricted by, including into mitochondria, where most of a cell's radicals are actually generated. And on top of the direct scavenging, it upregulates the enzymes the body already uses for the same job. Reiter's 2016 review in the Journal of Pineal Research lays all of this out, and the title it chose — melatonin "under promises but over delivers" — tells you the tone of that literature.

Here is the part the supplement marketing skips. Almost all of it is mechanism, established in cells, tissue and animals. A review of mechanism is not a demonstration of outcome. There is no large body of human trials showing that melatonin's antioxidant properties change a clinical result a person would notice or care about. The mechanism is not in doubt; the benefit is not established.

The doses are also a different conversation from sleep. The antioxidant and oncology literature typically works at 10 to 20 mg a night, with some oncology protocols going higher — an order of magnitude above a sleep dose and more than an order of magnitude above a clock dose. So someone taking 10 mg nightly "as an antioxidant" is acting on a mechanism, not on a demonstrated endpoint. That can be a reasonable thing to decide with a clinician. It is not the same thing as evidence that it works.

Melatonin has been tested in a lot of places. Four are worth your time, and they do not all point the same way.

Nocturnal blood pressure. Sixteen men with untreated essential hypertension, 2.5 mg an hour before sleep, in a randomized crossover trial published in Hypertension in 2004. Three weeks of nightly dosing lowered blood pressure during sleep by about 6 mmHg systolic and 4 mmHg diastolic, and restored a good part of the normal day-night swing — the amplitude of that rhythm rose 15% and 25%. A single dose did nothing. The effect required repetition, which is itself a hint that what moved was the rhythm rather than the vessel. Sixteen men is a pilot, not a practice change, and it has not been replicated at scale.

Migraine prevention. The cleanest non-sleep result in the file, and at an ordinary dose. A 2016 trial in JNNP randomized 196 people with 2 to 8 migraine attacks a month to placebo, amitriptyline 25 mg, or melatonin 3 mg, and followed the 178 who took a study drug for three months. Melatonin cut migraine days by 2.7 a month against 1.1 on placebo. It matched amitriptyline statistically, beat it on the share of people who improved by more than half, was better tolerated, and was associated with weight loss where amitriptyline produced weight gain.

Cancer supportive care. The most cited melatonin literature and the most misused. A 2012 meta-analysis in Integrative Cancer Therapies pooled 21 randomized trials in solid tumors, usually around 20 mg nightly added to chemotherapy or radiation, and reported a relative risk for one-year mortality of 0.63 (95% CI 0.53–0.74), along with less fatigue, nausea and vomiting, and fewer low platelet and white cell counts. Read flat, that is a dramatic survival effect. Read in context it is not yet a finding: most of those trials came from a single Italian research group, most were unblinded and small, and nothing of that magnitude has been reproduced in a large modern blinded trial. A very large effect from one centre that nobody else has replicated is the specific pattern that tends to shrink or vanish when it is finally tested properly. Note also that every one of these designs added melatonin to cancer treatment; none of them tested it instead of treatment.

Delirium in hospital. Included because this is the one where the proper trials got run. The mechanism was plausible, the early signals looked decent, and then a 2025 meta-analysis in Intensive Care Medicine restricted itself to the trials at low-to-moderate risk of bias — six randomized trials, 2,209 critically ill patients — and found no reduction in delirium (RR 0.89, 95% CI 0.73–1.09) and no effect on mortality. A column that lists only the wins is not reporting.

Note, preliminary or early findings are not the same as proof. Any use beyond a peptide's FDA-approved labeling (where one exists) is described here for educational purposes only and is not a recommendation.

Melatonin is generally very well tolerated. Common side effects include daytime drowsiness, headache, dizziness, and vivid dreams. Less common effects include short-term depressive symptoms, mild anxiety, abdominal cramps, and irritability. Melatonin can lower blood pressure and may interact with anticoagulants, antihypertensives, diabetes medications, and immunosuppressants. It does not appear to cause dependence or withdrawal.

Because the higher doses come up so often, it is worth saying what the safety review actually found. A 2016 review in Clinical Drug Investigation — Reiter among its authors — looked specifically at the question and concluded that short-term melatonin is safe even at extreme doses, with adverse effects limited to mild ones: dizziness, headache, nausea, sleepiness. Randomized long-term use produced side effects comparable to placebo. It named two gaps plainly. Long-term use in children and adolescents has not been settled, and there is too little human data to support use in pregnancy or breastfeeding. Keep the two findings separate in your head: that literature establishes tolerability at high doses far better than it establishes benefit at high doses.

What is in the bottle is its own variable, and it bites hardest at the low doses the clock research points to. A 2017 analysis in the Journal of Clinical Sleep Medicine ran 31 commercial supplements through liquid chromatography. Melatonin content ranged from 83% below the label to 478% above it. Lot-to-lot variation inside a single product reached 465%. More than 71% missed their label claim by more than 10%, and eight products contained serotonin, which has no business being there. A 2023 research letter in JAMA found the same labeling problem in melatonin gummies. Two practical consequences: a mislabeled "0.5 mg" can quietly be a multi-milligram sedative dose, and if you are deliberately taking a large dose, the number on the front is not the number you are getting.

Reasons to avoid it, or to talk to a clinician first:

This is not a complete safety list. It is the set of things worth knowing before a conversation with a prescriber.

Three melatonin papers landed this week, and for once they line up with the column above: one on what melatonin does to children's sleep architecture, one testing the antioxidant rationale in an actual clinical setting, and one on the beyond-sleep claims.

JAMA Network Open · melatonin · cross-sectional, propensity-matched

Melatonin is the most-used sleep aid in children and nobody had checked it against objective sleep recordings at scale. This study did, using overnight lab sleep studies (polysomnography) from a pediatric sleep clinic: 342 children on melatonin matched one-to-one against 342 who were not, on age, sex, BMI percentile, other illness, sleep apnea and epilepsy. The matched children on melatonin spent less of the night in REM sleep — a median 16.7% against 19.0%. Nothing else moved: total sleep time, the non-REM stages, breathing measures, arousals and limb movements were all indistinguishable.

Read it carefully, because the design limits what it can say. This is a snapshot, not a trial — nobody was randomized, dose and timing were not recorded, and children prescribed melatonin differ from children who are not in ways matching cannot fully erase. When the authors added ADHD, anxiety and depression to the model, the REM difference shrank but held. The effect is small, and whether melatonin caused it is genuinely unresolved; the authors say so themselves and call for prospective studies. Still, this is the biggest objective dataset on the most common pediatric sleep intervention, and it points somewhere specific.

Source

Journal of Evidence-Based Dental Practice · melatonin · systematic review and meta-analysis

This is the kind of study the antioxidant section above is missing — someone taking melatonin's antioxidant, inflammation-modulating and bone-building properties and testing whether they change a measurable clinical result. The setting is periodontitis: 18 randomized trials, 828 patients, melatonin added to standard deep cleaning (nonsurgical periodontal therapy) versus the cleaning alone.

The pooled result favors melatonin on gum attachment, most strongly at two months, where the added benefit was about 1.5 mm of clinical attachment gain. The honest shape of it is less tidy. Pocket depth favored the control group at one month, favored melatonin at two and three months, and was no longer significant at six. Bleeding on probing never reached significance at any timepoint. A signal that appears, inverts, and fades across timepoints is usually telling you about heterogeneity between small trials rather than about a stable effect — which is roughly where the antioxidant literature sits generally.

Source

Indian Journal of Orthopaedics · melatonin · narrative review

A survey of melatonin outside its sleep role, collecting the mechanistic case across bone, cartilage, muscle, tendon, ligament and spine. The bone section is the most developed: melatonin pushes bone-building osteoblasts, suppresses bone-resorbing osteoclasts, and acts on the RANKL/OPG signaling pair that governs the balance between them — which is why it keeps coming up in osteoporosis and fracture-healing work. The review also covers cartilage (resynchronizing the clock genes inside joint tissue), muscle (sarcopenia and cachexia), and autoimmune conditions.

Treat it as a map of where researchers are looking, not as evidence. It is a narrative review, much of it built on cell and animal work, and its own conclusion names the gap: translation still needs dosing strategy, better bioavailability, and large-scale validation. Useful for understanding why melatonin's non-sleep claims proliferate, which is that the mechanisms really do reach almost every tissue. Mechanism reaching a tissue is not the same as a result in a patient.

Source

Topics we'll be tracking:

About melatonin

Melatonin is not in our regulatory database. It has not been evaluated or approved by the U.S. Food and Drug Administration for any medical use. Discuss with your healthcare provider before considering any substance not reviewed here.

Medical disclaimer

This letter is for educational and informational purposes only. It does not constitute medical advice, diagnosis, or treatment, and no provider-patient relationship is created by reading it. Peptides and medications discussed may not be FDA-approved for the uses described. Always consult your healthcare provider before making any health decision. Read our full medical disclaimer.