Melatonin
What it is
Melatonin is a hormone produced primarily by the pineal gland, synthesised from serotonin via two enzymatic steps involving the amino acid tryptophan. Its secretion is tightly regulated by light exposure: darkness triggers release, peaking typically between 2am and 4am, while light suppresses it. This relationship with the light-dark cycle makes melatonin the primary hormonal signal through which the brain communicates time-of-day information to peripheral tissues.
The hormone exerts its effects through two G-protein-coupled receptors, designated MT1 and MT2, which are distributed throughout the brain and peripheral organs. MT1 receptor activation is primarily associated with sleep induction and suppression of alertness signals from the suprachiasmatic nucleus (SCN), the brain's principal circadian pacemaker. MT2 activation is more closely tied to phase-shifting effects, meaning the ability to advance or delay the timing of the sleep-wake cycle relative to external time cues.
Endogenous melatonin production declines substantially with age. Levels in older adults may be a fraction of those seen in younger adults, a pattern that is thought to contribute to the increased prevalence of sleep disturbances and circadian fragmentation in ageing populations. Perimenopausal women also show evidence of altered melatonin rhythms, though the relationship between declining oestrogen, circadian disruption, and sleep disturbance in this population involves multiple interacting mechanisms that have not been fully characterised.
Supplemental melatonin is sold widely at doses ranging from 0.5 mg to 10 mg, though the pharmacologically active dose for circadian phase-shifting is considerably lower, generally in the range of 0.5 mg to 1 mg. Many commercially available preparations substantially exceed this range, and the dose-response relationship for different outcomes is not well understood. Immediate-release formulations are most widely studied. Prolonged-release preparations have been evaluated specifically in older adults with insomnia, where they have regulatory approval in some European jurisdictions.
What the evidence shows
The evidence for melatonin is meaningfully differentiated by indication. Its strongest and most consistent support is for circadian rhythm disruption, particularly jet lag associated with crossing five or more time zones, and for delayed sleep-wake phase disorder. For these indications, the mechanism is well understood and the trial data are reasonably consistent. In primary insomnia in otherwise healthy adults, the evidence is more modest and less consistent, with effects on sleep onset latency that are statistically significant but clinically small in most trials.
A 2013 meta-analysis by Ferracioli-Oda and colleagues, published in PLOS ONE, pooled data from 19 trials of melatonin for primary sleep disorders and found a reduction in sleep onset latency of approximately 7 minutes (95% CI: 6.21 to 7.90 minutes) and an increase in total sleep time of around 8 minutes (95% CI: 1.74 to 14.75 minutes). These are statistically reliable findings but the clinical magnitude is modest, and the authors themselves characterised the effects as such. The trials included were predominantly small and of short duration, and small-study effects combined with probable publication bias mean these estimates likely overstate the benefit seen in the broader population. A 2022 meta-analysis in Sleep Medicine Reviews by Choi and colleagues, examining efficacy in chronic insomnia, reached broadly similar conclusions, with statistically significant but small effects versus placebo on objective and subjective sleep measures.
The picture for jet lag is more favourable. A Cochrane review by Herxheimer and Petrie, which examined ten RCTs in airline passengers and personnel, found that eight of ten trials showed melatonin taken close to the target bedtime at the destination reduced jet lag severity. The effect was most pronounced for eastward travel across five or more time zones and was consistent with melatonin's known chronobiotic properties.
For clinical populations, including older adults with insomnia and individuals with neurodevelopmental disorders, the evidence is more supportive than for healthy adults. A 2022 systematic review and meta-analysis examining children and adolescents with neurodevelopmental conditions found significant improvements in sleep onset latency. In older adults, where endogenous melatonin decline provides a plausible mechanistic rationale, prolonged-release preparations have shown modest benefit in several trials.
Five questions
Does low melatonin status cause harm? Declining endogenous melatonin with age is associated with increased sleep disturbance and circadian fragmentation, and there is epidemiological interest in its relationship with other age-related conditions. However, the causal direction is difficult to establish from observational data. Low melatonin is a feature of normal ageing and cannot be characterised as a deficiency state in the same sense as a micronutrient deficiency. The claim that low melatonin directly causes harm in otherwise healthy individuals is not established.
Does supplementation prevent disease? There is no robust clinical trial evidence that melatonin supplementation prevents any disease. Animal and mechanistic data have generated interest in potential antioxidant and neuroprotective properties, and there is preliminary observational interest in relationships between melatonin and bone density in postmenopausal women. None of this constitutes evidence for disease prevention, and these claims should be treated as speculative until supported by adequately powered human trials.
Does it affect biomarkers? Melatonin supplementation has been shown to affect circadian biomarkers including dim light melatonin onset (DLMO) timing, which shifts in the expected direction following administration. It also affects luteinising hormone (LH) and follicle-stimulating hormone (FSH) levels in some trials in perimenopausal women, though the clinical significance of these changes is uncertain and findings have been inconsistent across studies. A 2026 meta-analysis by Du and Tan in Frontiers in Nutrition found a possible increase in bone mineral density at the femoral neck in menopausal women, but high heterogeneity across just two trials prevents confident pooled interpretation.
Does it help clinical populations? The strongest clinical evidence is in populations with identifiable circadian pathology or endogenous melatonin deficiency. Older adults with insomnia, children and adolescents with neurodevelopmental disorders and associated sleep disturbance, and individuals with delayed sleep-wake phase disorder or non-24-hour sleep-wake disorder represent the populations with the most supportive trial data. In these groups, effects on sleep onset latency are more consistent and clinically more plausible given the underlying biology.
Does it benefit healthy individuals? The evidence in healthy adults without circadian disruption or clinical sleep disorder is weaker. Trials in this population show smaller effects that are more variable across studies. The GRADE assessment across multiple systematic reviews has supported only weak recommendations for use in healthy adults with insomnia, reflecting low certainty in the evidence rather than absence of effect. Jet lag in otherwise healthy travellers represents a specific context where benefit is reasonably well supported.
Individual variation
Age is the most clinically relevant source of individual variation. Older adults may be more biologically plausible responders because endogenous melatonin production declines substantially with age, and this provides a rationale for supplementation that does not apply in the same way to younger healthy adults. This population has also been the subject of more targeted trial programmes, including the prolonged-release formulation studied specifically in adults over 55. Evidence is more internally consistent in this group than in the general adult population, though trial durations are generally short.
Direction and distance of travel affects response in the jet lag context. Eastward travel across many time zones produces the strongest and most consistent evidence of benefit, consistent with the known circadian biology of phase-advancing effects.
Chronotype may influence response. Individuals with delayed chronotype, whose natural sleep timing is later than the conventional social schedule, represent the population in which phase-advancing effects of low-dose melatonin are most likely to be clinically relevant.
Perimenopausal and menopausal women experience disrupted sleep through multiple mechanisms, including vasomotor symptoms, altered hypothalamic-pituitary-ovarian axis signalling, and changes in melatonin secretion. The available RCT data specific to this population are limited. A 2021 systematic review by Treister-Goltzman and colleagues in the Journal of Pineal Research examined 24 studies in menopausal women and found mixed results across outcomes including sleep, mood, and hormonal markers. The 2026 meta-analysis by Du and Tan found no significant improvement in sleep quality in menopausal women specifically, though bone mineral density signals were present. This remains an area where more adequately powered trials are needed.
Children with autism spectrum disorder and other neurodevelopmental conditions have a relatively strong evidence base for melatonin's efficacy on sleep onset, with consistent improvements across multiple trials and a favourable short-term safety profile. Long-term safety in this group has not been adequately studied, and use should be under medical supervision particularly for extended periods.
Individuals with shift work disorder represent another population where circadian realignment is the clinical goal, though the evidence for melatonin in shift workers is less consistent than for jet lag, likely because the disruption pattern in shift work is ongoing rather than a single-event phase shift.
Testing and status assessment
There is no routine clinical test for melatonin status that is used in standard medical practice. Dim light melatonin onset (DLMO), measured via saliva or urine sampling under controlled lighting conditions, is a research tool that characterises the timing of melatonin secretion and can guide optimal supplementation timing in specialist settings. It is not widely available outside research contexts.
Serum melatonin levels can be measured but have limited clinical utility because melatonin levels fluctuate substantially across the 24-hour cycle and are highly sensitive to light exposure at the time of sampling. A single measurement does not characterise an individual's circadian melatonin profile meaningfully.
For practical purposes, the indication for melatonin supplementation is usually clinical rather than biochemical: jet lag associated with time zone crossing, delayed sleep timing, or age-related insomnia in older adults. The absence of a measurable deficiency state does not make melatonin supplementation inappropriate where the indication is circadian rather than replacement.
Safety
Melatonin is generally well tolerated at the doses used in most clinical trials, typically 0.5 mg to 5 mg. Commonly reported adverse effects include daytime drowsiness, headache, and dizziness, though these are reported at low rates and are usually mild. Most RCTs are 12 weeks or less in duration; safety data beyond this timeframe are limited, and the term "long-term safety" in the current literature refers to this relatively short evidence window rather than to years of continuous use.
Long-term safety data are substantially more limited than short-term data. A 2022 systematic review and meta-analysis by Menczel Schrire and colleagues in the Journal of Pineal Research examined higher doses and found no significant safety signals in available studies, but noted the evidence base for long-term safety remains thin given the short duration of most included trials. A preliminary observational analysis of health records from over 130,000 adults with chronic insomnia, presented at the American Heart Association's Scientific Sessions in November 2025, found an association between melatonin use for one year or more and higher rates of heart failure diagnosis, hospitalisation, and all-cause mortality. This study has not yet been published in a peer-reviewed journal. The magnitude of the association has not been peer-reviewed and residual confounding is highly likely given that the population studied -- people with chronic insomnia using melatonin long-term -- has substantially different baseline health characteristics from general melatonin users. The finding does not support causal inference, but it warrants attention and should be confirmed or refuted in prospective studies. Until that evidence is available, long-term supplementation in individuals with existing cardiovascular disease should be discussed with a prescriber.
Routine supplementation during pregnancy should be avoided unless specifically advised by a clinician. The concern is absence of adequate safety data rather than proven harm: melatonin crosses the placenta and its effects on fetal development have not been adequately characterised in human studies. Women who are breastfeeding should also avoid supplementation, as melatonin passes into breast milk.
Interactions with other medications are clinically relevant. Melatonin has additive sedative effects when combined with benzodiazepines, z-drugs, and other central nervous system depressants. There is evidence of interaction with warfarin and other anticoagulants, with some reports of altered INR, though the evidence base for this interaction is limited to case reports and small studies. Immunosuppressant therapy may be affected, as melatonin has immunomodulatory properties, and use in individuals on such medication should involve prescriber discussion. Fluvoxamine substantially inhibits melatonin metabolism via the CYP1A2 pathway and may markedly elevate melatonin levels.
Use in autoimmune conditions requires caution given melatonin's immunomodulatory effects, though the clinical implications are not well characterised. Children and adolescents should use melatonin under medical supervision, particularly for extended periods.
The dose sold in many countries substantially exceeds what is required for chronobiotic effects. It is useful to distinguish two dose contexts. For circadian phase-shifting purposes, doses of 0.5 mg to 1 mg may be sufficient in many contexts, consistent with receptor activation at near-physiological levels. However, dose-response evidence for sleep-onset outcomes is more complex: the 2024 Cruz-Sanabria analysis found sleep-onset latency effects appearing to peak at around 4 mg per day, with timing remaining at least as important as dose. Higher doses in the range of 5 mg to 10 mg appear to shift towards a more sedative profile, with effects closer to a hypnotic than a chronobiotic, though both mechanisms likely overlap to some degree at intermediate doses. Common supraphysiological dosing increases uncertainty around long-term safety, since most available safety data come from trials using doses at or below 5 mg. The clinical significance of sustained supraphysiological dosing over extended periods is unknown.
What can reasonably be concluded
Melatonin has a consistent evidence base for specific, well-defined indications: circadian rhythm disruption from jet lag crossing five or more time zones, delayed sleep-wake phase disorder, and sleep disturbance in older adults where endogenous melatonin production has declined. For these indications, the mechanistic rationale is coherent and the trial data are reasonably consistent. The evidence for primary insomnia in otherwise healthy adults is present but modest, with effects on sleep onset latency that are statistically reliable but small in absolute terms. The evidence specific to perimenopausal and menopausal women is limited and inconclusive for most outcomes. Long-term safety questions, particularly regarding cardiovascular outcomes, are unresolved and warrant continued monitoring. The common practice of using doses substantially higher than those shown to be effective for chronobiotic purposes has no clear evidential basis.
Where evidence is limited or outcomes are uncertain, conclusions should be treated as provisional and subject to revision as the evidence base develops.