Your melatonin dose is probably too high

Probably. In the trials, melatonin's effect on sleep stops improving at about 4 mg a night, and most bottles sell more than that.¹ The whole benefit is measured in minutes rather than hours. And when you take it matters more than how much.

What the trials gave

Start with the schedule most people are already on. The largest study of dose and timing describes usual practice as 2 mg taken 30 minutes before bed.¹ It gathered 26 trials in people with insomnia and in healthy volunteers, then asked how the dose and the hour changed two things: how long it took to fall asleep, and how long people slept.¹

Both improved as the dose rose. Both stopped improving at 4 mg a day, and nothing was gained above it.¹ The authors' own advice was to move the usual schedule up rather than down: 4 mg instead of 2, taken 3 hours before the intended bedtime instead of half an hour.¹

The shelf goes well past that ceiling, and the space above it is empty. A review of melatonin for jet lag, built on ten trials, found daily doses between 0.5 and 5 mg similarly effective.² People fell asleep faster after 5 mg than after 0.5 mg, and doses above 5 mg did no more than 5 mg did.²

How much sleep it actually buys

The dose argument shrinks once you see the size of the prize. The first review to gather the placebo-controlled trials covered 17 studies.³ Melatonin shortened the time to fall asleep by 4.0 minutes, raised the share of time in bed spent asleep by 2.2%, and added 12.8 minutes of sleep in total.³

A second review the same year reported a bigger average, 11.7 minutes off the time to fall asleep, but that average hid two different groups of people.⁴ In those with delayed sleep phase syndrome, a body-clock disorder in which sleep arrives hours late, the reduction was 38.8 minutes. In those with insomnia it was 7.2 minutes.⁴

Its authors reached two conclusions worth keeping apart. Melatonin was not effective for most primary sleep disorders over 4 weeks or less, and it looked safe over 3 months or less.⁴ The delayed sleep phase result was the one they called clinically important.

Read together, the two reviews describe a hormone that tells the body what time it is and sedates a little as a side effect. Where the clock is the problem, melatonin does something worth noticing. Where sleep is the problem and the clock is fine, it buys minutes.

The one trial that set a small dose against a large one

The case for a very small dose is about blood levels. A third of a milligram brings melatonin into the range the body itself makes at night. Several milligrams pushes far past it.

Whether that translates into better sleep was tested directly, in 24 healthy adults over 55.⁵ Each spent 2 weeks on placebo and 2 weeks on either 0.3 mg or 5.0 mg, taken 30 minutes before lights out.⁵ Sleep was recorded in the lab during both the biological day and the biological night.⁵

The small dose showed only a trend toward better sleep efficiency, and that came from sleep scheduled during the biological day. The large dose raised sleep efficiency during both day and night, mainly by lengthening stage 2 sleep and slightly shortening awakenings.⁵

That is the reverse of what the low-dose argument predicts, and it deserves saying plainly. In older adults, whose own melatonin output has fallen, the larger dose did more for sleep. What survives for the small dose is a different measurement, and the fact that 5 mg sits at the top of the effective range rather than above it.²

What the dose does to the clock

Melatonin's clearer job is moving the body clock, and there the dose and the hour work together. A review of 30 studies in healthy participants found that melatonin and related drugs pulled the sleep-wake cycle earlier according to both the dose and the time it was given.⁶ That happened even though these participants' own melatonin rhythms were normal to begin with.

The shift came to a standardized mean difference of -0.639.⁶ That unit measures a result in standard deviations: 0.2 counts as small, 0.5 as moderate and 0.8 as large.

Timing outweighed dose in the dose-response work too. The gap between swallowing the tablet and going to bed predicted how fast people fell asleep, which is why its authors recommended 3 hours.¹ The jet lag review makes the same point from the other side: taken early in the day, melatonin causes sleepiness and delays adjustment to local time.² A bigger dose at bedtime does not sedate harder. It keeps the signal present longer, and an ill-timed signal is the one thing these trials agree is unhelpful.

What a third of a milligram does that 3 mg does not

The strongest argument on the low-dose side is about blood levels rather than sleep. A French intensive care trial gave 355 ventilated patients placebo, 0.3 mg or 3 mg of melatonin at 9 p.m.⁷ Its first question was which dose produced the better melatonin blood profile a day later. The small dose did, in 50% of patients, against 24% on 3 mg and none on placebo, and the 3 mg arm was dropped.⁷

The low dose then made no difference to delirium: 54.4% against 55.2% on placebo.⁷ That is a finding about delirium in critically ill people, not about sleep at home. What it shows is narrower and still worth having. A third of a milligram reproduces the body's own nightly profile better than ten times as much does.

Where the evidence thins

Every number above has an edge. The dose curve can only describe the doses that were tested, and its authors set their 4 mg optimum against usual practice at 2 mg, not against the largest bottles on sale.¹ The finding that nothing is gained above 5 mg comes from travelers, not from people with insomnia.² The head-to-head test of 0.3 against 5.0 mg had 24 participants and ran for a month.⁵ The blood profile result comes from sedated patients on a ventilator, whose absorption differs from a healthy adult swallowing a tablet at home.⁷

The effect sizes come from the same trials that set the dose. Falling asleep 4.0 minutes sooner and sleeping 12.8 minutes longer is the average across every dose tested.³ So the difference between a well-chosen dose and a careless one is a fraction of an already small number.

What remains unknown

Nobody has compared a third of a milligram against 2 to 4 mg in adults with ordinary insomnia, in enough people and for long enough to settle it. The one head-to-head comparison ran in older adults, favored the larger dose, and had 24 participants.⁵

Whether the effect survives nightly use over months has not been measured. The safety record runs to 3 months and the sleep record is shorter.⁴ And no sleep trial has shown that the blood profile a small dose produces is the profile that matters for sleep rather than for timing.⁷