Melatonin is not a sleeping pill.

What the molecule actually does

Melatonin is the chemical signal for night. The brain releases it as light falls, and its biological job is to announce the internal hour to the tissues that need it. It does not knock you out. It informs.

That distinction isn't vocabulary pedantry: it determines entirely how the supplement should be used. A sedative is taken when you want to sleep. A time signal is taken when you want the body to believe it is that hour. Those are different clock times, and confusing them is the most common mistake.

The effect on sleep: real, modest

A meta-analysis published in PLoS ONE in 2013 pooled the placebo-controlled trials in primary sleep disorders. It concludes there is a significant effect — time to fall asleep shortens, total duration rises slightly, perceived quality improves a little — and that the magnitude is modest. Against what the shelf promises, the gap is wide.

Against behavioural therapy for insomnia, wider still. In established insomnia the guidelines put CBT-I first-line, and no supplement comes close. Melatonin is not a treatment for insomnia: it is a tool for timing.

Where it genuinely works: moving the clock

That is its documented, useful role: jet lag, a late chronotype, preparing for a night shift, coming back to a normal rhythm after drifting.

A study published in the Journal of Clinical Endocrinology & Metabolism in 2010 established phase response curves for melatonin given on three consecutive days, comparing 0.5 mg against 3.0 mg. Both advance the clock comparably — and the low dose leaves fewer residual effects the next day. Put plainly: the 5 or 10 mg tablets on the shelf add nothing on the variable that matters.

Timing, on the other hand, changes everything. Taken several hours before the target bedtime, the supplement lands on the part of the curve where the phase advance is largest. Taken at lights out, it arrives too late for the clock effect and leaves only the small sedative one.

The problem no dose fixes

Sold as a food supplement in many countries, melatonin escapes the quality control applied to medicines. This is not a suspicion: it has been measured.

An analysis published in JAMA in 2023 assayed 25 gummy products sold in the US. Twenty-two were inaccurately labelled, with actual content running from 74% to 347% of the claim — and one contained no detectable melatonin at all. Earlier work, published in the Journal of Clinical Sleep Medicine in 2017 across a wider range of formats, found an even broader spread, plus serotonin — an undeclared active substance — in one product in four.

The practical consequence is simple. If you decide to take it, format matters: prefer a scored low-dose tablet to a flavoured gummy, and be wary of the child-facing presentations, which are precisely the ones that were caught out.

The free version of the same mechanism

Before the supplement there is light — same mechanism, stronger effect, no labelling problem. Morning light advances the clock, a dim evening stops delaying it, and the human clock is very sensitive to evening light. One evening under a ceiling fitting is enough to delay the melatonin you make yourself. Buying gummies while keeping the same brightly lit evening is paying to correct what you keep breaking.

How Sillon handles it

The card "Melatonin, right dose and right hour" carries a grade B, gives the low dose and the time window, and states that the use is a few days rather than continuous. It cites the four references below, including the two on labelling — because a practice whose product isn't reliable has to say so. It is one of the cards whose job is as much to close a door as to open one.