How much melatonin should you take to sleep better?
For most adults, 4 mg appears to have the greatest effect, but it should be taken 2 to 3 hours before bedtime; if you are older than 55, the extended-release formulation (2 mg) is the only one officially recommended.
A dose-response analysis of 26 randomised studies (more than 1,600 participants) found that melatonin best promotes falling asleep quickly and extends total sleep time at a dose of 4 mg per day. Lower doses also help, but the effect increases up to that 4 mg threshold and then levels off.
The timing of intake matters at least as much as the dose. In practice, people take melatonin about 30 minutes before bedtime, but that same analysis shows that taking it 3 hours earlier noticeably improves the effect. That may seem inconvenient, but it fits with the way melatonin works: it shifts your biological clock, and that takes time.
For people older than 55, the formulation also makes a difference. The European guideline for insomnia recommends extended-release melatonin at a dose of 2 mg, for a maximum of three months. The same guideline explicitly advises against the standard immediate-release form as a treatment for insomnia. A smaller study in older adults found that a dose of 5 mg significantly improved sleep efficiency, while 0.3 mg had no noticeable effect on nocturnal sleep.
For children older than 2 years who do not respond well to better sleep habits and behavioural approaches, a European expert panel describes low-dose melatonin as a possible second step, taken 30 to 60 minutes before bedtime, always under the supervision of a paediatrician. Note that several authors of that advice have ties to a manufacturer, so factor that in.
Melatonin is not the major sleep solution it is sometimes portrayed as in popular media. Production does decline with age, but the scientific evidence that supplementation genuinely restores sleep quality in older adults is limited and inconsistent. Furthermore, higher doses are pharmacological: they produce blood levels far above what the body naturally generates, and the long-term safety of this has not yet been adequately studied.
Based on one dose-response meta-analysis (26 RCTs, n=1689), the European insomnia guideline (2023), a crossover study in older adults, a European paediatric expert panel (with declared conflicts of interest) and several narrative reviews. No large independent RCTs exist for higher doses or long-term safety.