Melatonin — the right dose, the right hour, and why almost everyone uses it wrong

What melatonin actually is, biologically

The mistake most users make starts with the conceptual model. Melatonin is not a sedative. Treating it as one is the source of most of the misuse below.

Melatonin is a hormone produced by the pineal gland, a small structure deep in the brain. Endogenous secretion begins about two hours before habitual sleep — the point researchers call dim-light melatonin onset, or DLMO — peaks during the middle third of the night, and tapers before waking. Total quantity secreted over a typical night is roughly 0.01 to 0.08 milligrams. Peak plasma concentration in healthy adults sits around 60 to 80 picograms per milliliter.

The function of endogenous melatonin is to signal darkness to the rest of the body. It is a circadian phase marker, not a sleep-inducing agent in the pharmacological sense. The brain does not wait for melatonin and then fall asleep. The brain produces melatonin because it is already preparing to sleep. The hormone is the signal of darkness, not the cause of sleep.

The doses sold over the counter — 1, 3, 5, and 10 milligrams — produce plasma levels thirty to three hundred times the endogenous peak. Pharmacologically, taking 5 milligrams of melatonin at bedtime is a different intervention than the body's own melatonin signaling. Not a higher dose of the same thing. A fundamentally different regime, with a fundamentally different response.

Endogenous melatonin secreted over an entire night totals about thirty micrograms. A 5-milligram gummy is roughly one hundred and sixty times that.

The right dose — 0.3 to 0.5 milligrams

The dose-response work was done in the late 1990s at MIT. The findings have been replicated. The supplement industry has ignored them, because the supplement industry has no commercial reason to listen.

Richard Wurtman and his colleagues at MIT published a series of studies between 1994 and 2001 mapping the dose-response curve of exogenous melatonin on sleep onset and on plasma melatonin levels. The clean finding: 0.3 milligrams produces a plasma melatonin level approximating the endogenous nocturnal peak. The effect on sleep onset latency saturates at roughly this dose. Doses above 0.3 milligrams do not increase efficacy. They do increase plasma melatonin by orders of magnitude and prolong the duration over which plasma levels remain elevated — which is the source of next-day grogginess and the chronic-dosing problems described later in this piece.

A 0.5 milligram dose sits a notch above the saturation point with a slightly wider safety margin and is also reasonable. Anything above 0.5 milligrams is exogenous substrate overshoot. It does not work better. It works worse over time, because of receptor downregulation, and worse the next morning, because of the duration of elevated plasma levels.

The over-the-counter market settled on 1 milligram, 3 milligrams, 5 milligrams, and 10 milligrams primarily because consumers respond to higher-dose labelling as if it indicated higher efficacy, because the US supplement regulatory category permitted any dose, and because no commercial incentive existed to release the 0.3 milligram product the evidence supported.

If you are taking 5 or 10 milligrams of melatonin, you are running a pharmacology experiment that nobody has published results for. The dose-response curve plateaus at 0.3 milligrams. Everything above that is excess substrate signaling a brain that does not need any more signal.

Wurtman did the dose work in 1995. The supplement industry has had thirty years to update its bottles. It has not.

The right timing — five to six hours before bed

Dose is the first error. Timing is the second, and they compound.

Endogenous melatonin onset in a healthy adult occurs about two hours before sleep. The natural rise is gradual, beginning in the early evening, peaking during the middle of the night, and tapering before waking. The pharmacological intervention that mimics this curve — and produces the circadian phase-shifting effect melatonin is best validated for — is exogenous melatonin taken in the early evening, four to six hours before target sleep time, on a daily basis.

Melatonin taken at bedtime is something different. The pharmacology is dominated by the absorption-and-clearance curve rather than by the phase-signaling effect. A 0.3 milligram dose at bedtime produces a small sleep-onset benefit — effect size around d = 0.2, small in the clinical literature. A 5 milligram dose at bedtime produces a slightly larger short-term effect on sleep onset, at the cost of receptor downregulation, morning grogginess, and the failure to achieve the phase-signaling effect that is melatonin's actual mechanism.

The phase-shifting use case — five to seven hours before bed — does work and has the strongest evidence base. Used this way, melatonin advances the circadian clock by 0.5 to 1.5 hours per night cumulatively, with the largest effects in the first three to seven nights of consistent dosing. This is the protocol used in clinical management of delayed sleep phase disorder and in jet lag. It is also nearly nobody's use case at home.

Melatonin at bedtime is the most common use of melatonin and the use case the evidence supports least. It works modestly for sleep onset in a small subset of users; it does not work for sleep maintenance; and it is not what melatonin is for. Use melatonin to move your clock, not to sedate yourself.

Melatonin is a phase signal. Treating it as a sedative is like using a thermometer to drive a nail.

When melatonin is the right intervention

Snerva is not anti-melatonin. Used correctly — right dose, right hour, right indication — it is one of the cleanest tools in this category. The right indications are specific and they are not what most users are taking it for.

Delayed sleep phase disorder

0.3 to 0.5 milligrams, taken five to seven hours before desired sleep time, daily, paired with morning bright-light exposure within thirty minutes of desired wake. Four to eight weeks of consistent use produces a measurable phase advance for most patients. The cleanest use case in the literature. Full protocol in delayed sleep phase disorder.

Jet lag

0.3 to 0.5 milligrams in the early evening at destination local time, for three to five nights after arrival. Eastward travel — where the clock has to advance — benefits more than westward, where the clock delays naturally. Full day-by-day protocol in jet lag.

Shift work disorder

Limited but reasonable evidence for targeted melatonin in shift workers — typically 0.5 to 3 milligrams taken on entering the sleep window after a night shift. Mechanistically similar to jet lag, with the complication that shift workers rotate. Specialist guidance is usually warranted.

Free-running in blind individuals

Adults without light perception experience non-24-hour sleep-wake disorder. Established treatments are tasimelteon (a prescription melatonin-receptor agonist) and timed exogenous melatonin. A small, specific clinical population. Outside the scope of self-administered protocols.

Children with autism-related sleep onset difficulty

Distinct evidence base, pediatric-specific dosing, and decisions that belong with a pediatrician familiar with sleep medicine. Mentioned because the use case exists. Not detailed because it is not where the consumer purchase pattern overlaps.

When melatonin is not the right intervention

These are the cases where users take melatonin, do not get better, and conclude that nothing works. The melatonin was the wrong tool. Something else was the right one.

Sleep-maintenance insomnia

If your problem is waking at 3 or 4am unable to return to sleep, melatonin has minimal effect. The maintenance phase of sleep is governed by cortisol, core body temperature, and sleep architecture — not by the dim-light melatonin signal that has long since passed by 3am. The relevant article is can't sleep at 3am. Sleep restriction and stimulus control are the protocol, not melatonin.

Anxiety-driven insomnia

If your insomnia is hyperarousal — racing thoughts, sympathetic activation, sleep effort — melatonin does not address the loop. The loop is cognitive and behavioral. The protocol is in anxiety insomnia. Melatonin is, at best, a placebo for this presentation, and the placebo response decays.

Chronic insomnia beyond three months

Chronic insomnia is a conditioned, behavioral disorder. CBT-I is curative. Melatonin is symptom management at best and, given that most chronic insomnia is not circadian, usually not even that. The trajectory and treatment-by-stage view is in acute vs chronic insomnia. The broader pharmacology comparison is in CBT-I versus sleeping pills.

Perimenopause-driven insomnia

Hormonal sleep disruption is its own clinical category. Melatonin does not address the upstream driver, which is estrogen withdrawal affecting thermoregulation and sleep architecture. The piece is perimenopause insomnia.

Insomnia driven by stimulus-control failures

If you have made your bed a place for reading, work, and worry, that conditioning is the mechanism and stimulus control is the intervention. Melatonin will not unmake the pairing. Stimulus control therapy covers the six rules and the failure modes.

Side effects, real and overhyped

The side-effect profile of melatonin is mostly benign at correct doses and meaningfully worse at the doses most users take. Worth separating the real concerns from the imagined ones.

Real and dose-related

Morning grogginess is the most common side effect at doses above 1 milligram. Vivid dreams and occasional nightmares above 3 milligrams. Mild headache. Lowered core body temperature, which can feel like morning chills. Paradoxical wakefulness — feeling more alert after a dose — occurs in roughly five percent of users and is not well understood mechanistically.

Concerns with chronic high-dose use

Receptor downregulation is the most-cited theoretical concern. Melatonin receptors, like most receptor systems, downregulate under sustained agonism. Long-term users at 3 to 10 milligrams may experience reduced sensitivity over months to years — meaning the dose stops working and the user escalates, which compounds the original problem. The mechanism is plausible. The long-term trial data at supratherapeutic doses is limited because those trials have not been run.

Probably not real, despite popular concern

Melatonin dependence in the addiction sense is not a recognized clinical entity. The receptor pharmacology does not support it. Discontinuation is not associated with withdrawal symptoms beyond a transient return to the baseline sleep difficulty that prompted use. Permanent suppression of endogenous melatonin production is not supported by the evidence; endogenous secretion recovers within days off exogenous dosing.

Caution warranted

Pregnancy: data are insufficient; default conservative. Autoimmune conditions: theoretical concern about immune modulation, not strongly supported, not ruled out either. Children long-term: open question, particularly around pubertal timing, which melatonin may influence. The pediatric clinical use case is real; the chronic-use case in otherwise healthy children is not warranted by the evidence.

Plasma melatonin from a 10-milligram pill stays elevated through breakfast the next morning. The grogginess is not in your head.

What is in the bottle is probably not what the label says

If the dose and timing problems were the only issues, this article would be shorter. There is also a supply-chain problem, and it is not small.

Erland and Saxena published a survey in the Journal of Clinical Sleep Medicine in 2017 testing thirty-one melatonin supplements purchased in Canada and the US for actual content versus label claim. Seventy-one percent varied by more than ten percent from the labeled dose. The highest deviation found was 478 percent above label — a product labeled as containing 1 milligram of melatonin that actually contained 4.78 milligrams. Twenty-six percent of the products tested also contained serotonin contamination, an active pharmacological compound that has no business being in a sleep supplement.

Subsequent surveys in 2020 and 2023 showed similar variance. Gummies are consistently the worst format for dose accuracy. Brand premium correlates loosely with accuracy — the Amazon-bestseller-list products skew worse — but expensive labelling is not a guarantee. The only product feature that reliably correlates with what is in the bottle is third-party testing certification: a USP, NSF, or ConsumerLab seal.

Practical implication: if you are going to take melatonin, buy from a brand that publishes its third-party testing results. Pure Encapsulations, Thorne, NOW Foods, and Life Extension are reasonable starting points. Amazon's bestseller list is not. The brand premium for a tested product is roughly twenty percent over the cheapest options and is the only product feature actually correlated with the dose in the pill.

Melatonin gummies are not a sleep aid

The gummy format deserves its own paragraph. It is the worst form of an already misused substance, sold to people who should not be using it at all, and increasingly to children, for whom the question is whether the format itself is the harm.

Dose accuracy is consistently worst in the gummy format across all surveys of supplement contents. The reasons are manufacturing: gummies are produced from liquid syrup, set, cut, and coated, and the active ingredient distributes unevenly through the matrix. A 5 milligram label on a gummy means somewhere between 2 milligrams and 25 milligrams in the actual gummy you ate, with the variance distributing across a single bottle.

The format also normalizes daily use, treats melatonin as candy, makes accidental over-consumption trivial — a child or an anxious adult reaching for a second is not interrupted by the form factor — and contains sugar that itself disrupts sleep architecture if consumed within the hour before bed. The combination is uniquely bad.

Pediatric ER visits for melatonin overdose increased 530 percent in the United States from 2012 to 2021, according to CDC data. The majority of those overdoses involved gummies the child accessed unsupervised. The product category as currently designed is, at the population scale, a poison-control problem.

Snerva position: do not buy melatonin gummies. If you need melatonin, buy a tablet from a third-party-tested manufacturer at the dose the evidence supports.

Children's melatonin gummies were not a category in 2010. Pediatric ER visits for melatonin overdose have risen 530 percent since.

The 0.3 milligram supply problem

The dose-response curve plateaus at 0.3 milligrams. The shelf, mostly, does not stock 0.3 milligrams. Here is how to get the dose the evidence supports without writing your own clinical trial.

Several manufacturers do sell 0.3 milligram (or 300 microgram) products. Pure Encapsulations 0.3mg, Life Extension low-dose, NOW Foods 300mcg, and a small number of specialty pharmacy brands are reliable starting points. These products are available online and increasingly through health-focused retailers. They are not what is sitting at the front of the typical pharmacy shelf.

If only higher-dose tablets are available — most commonly 1 milligram — a pill cutter divides them into thirds. Imprecise but adequate for non-critical use. Half a 1 milligram tablet (0.5 milligrams) is also within the saturation range and easier to cut accurately. Do not crush gummies; the matrix is uneven and any fragment dose is unreliable in either direction.

Sublingual liquid melatonin allows accurate sub-milligram dosing by drops, and several brands offer 0.5 milligram per drop concentrations. Compounding pharmacies can produce precise low-dose tablets by prescription if absolute dose accuracy is required, which is rarely necessary for self-administered use.

The supplement industry will sell 0.3 milligram melatonin when consumers stop buying 5 milligrams. So far, they have not.

What to do this week

Four reader profiles, four next steps.

If you currently take 3 to 10 milligrams of melatonin at bedtime for general sleep

Stop. The dose is wrong, the timing is likely wrong, and the use case may be wrong. Identify your actual sleep problem before deciding whether melatonin is the right tool — our insomnia hub covers the three patterns and the protocols for each. If the answer is circadian, switch to 0.3 to 0.5 milligrams five hours before desired sleep. If the answer is behavioral, stop melatonin entirely and start the CBT-I work. If you are unsure, default to stopping for two weeks while you sort it out.

If you are considering trying melatonin for the first time

Do not, unless you have a clear circadian use case (delayed sleep phase, jet lag, shift rotation). For garden-variety sleep onset difficulty, the right starting point is stimulus control and sleep restriction. The acute vs chronic insomnia article covers the trajectory; the CBT-I versus sleeping pills piece covers the broader comparison. Melatonin is rarely the first move and almost never the right first move for an anxious sleeper.

If you have delayed sleep phase disorder and are using melatonin correctly

You are in the right place. Confirm 0.3 to 0.5 milligrams taken five to seven hours before desired sleep time, daily, paired with morning bright-light exposure within thirty minutes of desired wake. The light half of the protocol is the half most patients underuse; light therapy covers the dosing curve.

If you are traveling east and need a jet-lag protocol

0.3 to 0.5 milligrams in the early evening at destination local time, for three to five nights post-arrival. Pair with morning light at destination. Avoid melatonin in the destination morning — opposite signal. Full day-by-day protocol in jet lag.

Frequently asked questions