What the milligrams on a melatonin label mean

"5 mg melatonin" reads like a dose. When researchers put melatonin products through a chromatograph, most did not contain what the label said, and the misses mostly ran high. Here is what two lab studies found, what they could not test, and how to think about the number on the bottle.

Melatonin is one of the simplest supplements to make on paper: one small molecule, a known amount, into a tablet or a gummy. As someone who formulates supplements, that is exactly why this story interests me. If any label should be easy to get right, it is this one.

The US gummy test

In April 2023, Pieter Cohen of Cambridge Health Alliance and chemists at the University of Mississippi published a research letter in JAMA. In September 2022 they pulled every melatonin gummy brand entered in the NIH Dietary Supplement Label Database during 2021, bought them online, and measured melatonin (plus CBD and serotonin) by ultra high-performance liquid chromatography with a photodiode array detector. Four were unavailable and one did not actually say "melatonin" on its label, which left 25 products to analyze.

The results, verbatim: "Twenty-two of 25 products (88%) were inaccurately labeled, and only 3 products (12%) contained a quantity of melatonin that was within ±10% of the declared quantity." Among products with melatonin, the actual amount ranged "from 74% to 347% of the labeled quantity." One product, labeled as 5 mg of melatonin with CBD, had no detectable melatonin at all, only 31.3 mg of CBD.

Going through the paper's table myself, the direction is lopsided: 20 of the 25 gummies held more than 110% of their labeled melatonin, 3 were within 10%, and 2 were under (one at 74%, one with none). The worst case was labeled 3 mg and measured 10.4 mg per serving. The CBD amounts, by contrast, were all between 104% and 118% of label.

The Canadian tablets, capsules and liquids

The US letter was not the first. A 2017 study by L. A. Erland and P. K. Saxena in the Journal of Clinical Sleep Medicine measured 31 melatonin supplements bought from grocery stores and pharmacies in one Canadian city, across forms: tablets, capsules, chewables and liquids. Per the abstract, melatonin content ranged "from -83% to +478% of the labelled content," more than 71% missed their label by more than 10%, and "lot-to-lot variable within a particular product varied by as much as 465%." Serotonin, which was not on any label, turned up in eight products at 1 to 75 micrograms.

An editorial in the same issue, by Madeleine Grigg-Damberger and Dessislava Ianakieva of the University of New Mexico's sleep disorders program, adds the details that stuck with me. The most variable product was a chewable tablet labeled 1.5 mg that held almost 9 mg. The least variable were "those that contained the simplest mix of ingredients, generally oral or sublingual tablets" with a plain filler. In other words, the more complicated the format, the less the label meant.

Why the misses run high

Low results are easy to explain: melatonin degrades, and products sit on shelves. High results are the puzzle. The editorial reports that the Canadian authors found the excess "puzzling" and "pondered whether excessive melatonin was added to ensure the product met the label claim."

That is a known practice. Formulators often add an overage of an ingredient that breaks down, so the product still meets its label at expiry. I described the same logic for probiotics in Report 177. A modest overage is responsible. Three and a half times the label is not an overage anybody can defend, and neither study can say which products overshot on purpose and which were simply poorly made. That is speculation about motive, and I am labeling it as such.

Why a few milligrams matter here

For many supplements, 30% over label would be a rounding error in effect. Melatonin is a hormone, and the doses on the shelf are already large compared with what the body makes. Citing a 1994 dosing study, the JAMA letter notes that "as little as 0.1 mg to 0.3 mg of melatonin" given to young adults "can increase plasma concentrations into the normal nighttime range." Measured against that, a 5 mg gummy is already many times a physiological dose before any mislabeling.

The JAMA authors were writing mainly about children, who can take gummies by accident or by design. They estimate that "consuming melatonin gummies as directed could expose children to between 40 and 130 times higher quantities of melatonin," and they advise clinicians to tell parents that pediatric use "may result in ingestion of unpredictable quantities of melatonin and CBD." That is their conclusion, and it is about predictability, not a claim that any specific harm was measured.

What the studies cannot tell you

The JAMA authors list their own limits: "the sample size was small, only 1 sample of each brand was analyzed, and only gummies were analyzed." They state it is not known whether the results apply to US tablets and capsules, or whether a brand varies batch to batch. The Canadian study did test multiple lots and found that variation is real, but it covered one city's shelves in 2017. Neither study names brands, so neither can tell you whether the bottle in your cabinet is accurate.

What they do show, consistently, across two countries, several product forms and six years apart: the milligram figure on a melatonin label is a target, and in these samples it was missed more often than hit.

Four questions for a melatonin label

What form is it? In the Canadian data, simple tablets were the most consistent and chewables the least. Gummies were the only form in the US test.

Is there a lot-specific test? A certificate of analysis tied to the lot number in your hand means more than a general "lab tested" badge. The Canadian lot-to-lot swings are why.

Why this dose? If the physiological range starts around a tenth of a milligram, ask what a 10 mg product is for. Bigger is not a feature.

What else is in it? Extra botanicals and CBD complicate both the chemistry and the measurement. The products with the most added ingredients are not the ones you can check most easily.

None of this says whether melatonin helps you sleep. That is a separate question, with its own evidence. This is only about whether the number on the label is the number in the product, and on that narrower question the answer from both studies is: often not.

Not medical advice. This is educational analysis, not a recommendation — a study is not a prescription. Talk to a qualified clinician before acting on anything you read here. Full disclaimer →

Sources

P. A. Cohen, B. Avula, Y.-H. Wang, K. Katragunta and I. Khan, "Quantity of Melatonin and CBD in Melatonin Gummies Sold in the US," JAMA 329(16):1401–1402 (25 April 2023), DOI 10.1001/jama.2023.2296, PMCID PMC10130950. (Primary source, research letter read in full including the results table. Source of the sampling method, the UHPLC-PDA analysis, the quoted 22-of-25 and 74%-to-347% findings, the no-melatonin CBD product, the 104% to 118% CBD range, the quoted 0.1 to 0.3 mg passage (which the letter cites to Dollins et al. 1994), the quoted 40-to-130-times and "unpredictable quantities" passages, and the quoted limitations. The 20 over / 3 within / 2 under tally and the 3 mg vs 10.4 mg example are my reading of the table.)
L. A. Erland and P. K. Saxena, "Melatonin Natural Health Products and Supplements: Presence of Serotonin and Significant Variability of Melatonin Content," Journal of Clinical Sleep Medicine 13(2):275–281 (2017), DOI 10.5664/jcsm.6462, PMCID PMC5263083. (Primary source, abstract read via Europe PMC; full text not read. Source of the 31-supplement sample, the quoted -83% to +478% and 465% lot-to-lot figures, the more-than-71% finding, and the serotonin result.)
M. M. Grigg-Damberger and D. Ianakieva, "Poor Quality Control of Over-the-Counter Melatonin: What They Say Is Often Not What You Get," Journal of Clinical Sleep Medicine 13(2):163–165 (2017), DOI 10.5664/jcsm.6434, PMCID PMC5263069. (Editorial on source 2, read in full. Source of the purchase locations, the product forms, the 1.5 mg chewable holding almost 9 mg, the quoted simplest-mix passage, and the quoted "puzzling" and overage passages describing the Erland authors' view.)
Prior reporting in this publication: Report 177, what a probiotic CFU number counts; Report 175, what "99% pure by HPLC" means; Report 058, what's actually in the bottle. (Context only. No claim in this report rests on them.)

Disclosure, plainly: I founded and run Shroombiosis (a company I run), which formulates and sells supplements. That is a commercial stake in how supplement labels get read, and the four questions above apply to my products as much as anyone's. I have no relationship with any author, lab or brand discussed here. Nothing here is sponsored and no link earns a commission; here's the full policy. A recommendation with no stake at all: for performance nutrition, Die Tryin Co. is a fellow combat-veteran-owned brand I'm glad to point people to. I don't own it and earn nothing from the link.