Four independent laboratories have now weighed the nicotine in oral pouches against the number printed on the can, and the verdict is that the label is directionally honest and quietly optimistic. Most products measure between 80% and 95% of their declared strength, a handful come in above it, and the worst misses in the literature are not pouches at all. No law anywhere requires that number to be printed in the first place.
A can that says 6 mg is making a claim, and for most of this category’s short life nobody had checked it. That has changed. Between 2020 and 2026 a government lab, two university labs and one manufacturer’s own lab each bought products off the shelf, extracted the nicotine and compared the result with the packaging. Ninety-odd labelled products have now been through that process.
The findings are less dramatic than the internet would like and more interesting than “the labels are fine”. The number is usually a little generous. The scatter is wider than you would accept on a medicine. And the reason nobody polices it is not regulatory oversight — it is that the declaration is voluntary, and the one nicotine figure that US law does demand is sent to the federal government and then withheld from you as a trade secret.
What the four labs found
The largest measured-against-label dataset comes from the Centers for Disease Control and Prevention’s Division of Laboratory Sciences, which analysed 70 oral products from six manufacturers bought around Atlanta in 2020. Its summary sentence is the plainest statement of the direction of error anywhere in the literature:
The measured total nicotine content was either equal to or less than the level stated on the label, except for one product.
And the spread, in CDC’s own words: “Generally, the measured total nicotine ranged from 45% below and 1% above the indicated strength and the percentage of freebase nicotine was lower in the stronger-labelled nicotine pouches than lower strength.” CDC declared no specific grant for the work and no competing interests.
The second dataset is from Roswell Park Comprehensive Cancer Center, funded by the National Cancer Institute, covering 50 tobacco-free pouches across ten brands plus eleven tobacco pouches. Its conclusion reads differently from CDC’s and it matters that it does:
Overall, we found that ONPs are generally consistent with reported nicotine content.
More precisely, “for most ONPs, measured nicotine content was consistent with values provided on packages (within 20%”. One author of that paper discloses a research grant from Pfizer and membership of a scientific advisory board at Johnson & Johnson, outside the work; the others report none.
The third is a Johns Hopkins team’s analysis of ten Velo variants bought in Karachi, funded by the Bloomberg Initiative to Reduce Tobacco Use. It is the tightest single-brand dataset in existence and every one of the ten came in under:
For each product, the measured level was similar to, though always somewhat lower than, the labelled level.
Ten for ten below label, in a band from 80% to 95% of the declared figure, with replicate analyses “within 8%”. Its corresponding author discloses having been “a paid consultant in litigation against a tobacco company”.
The fourth comes from Altria Client Services, and its funding statement needs to travel with every number it produced: “This study was funded by Altria Client Services LLC, an affiliate of Helix Innovations LLC, the manufacturer of on! nicotine pouches.” It is also the only dataset in which products measured above label — three of the seven tested, one of them the funder’s own brand.
The numbers, side by side
Every figure below is the paper’s own measured value against the strength printed on the can. The percentages in the right-hand column are my arithmetic, not the authors’.
| Product | Labelled | Measured | % of label | Lab |
|---|---|---|---|---|
| on! | 8 mg | 8.2 mg/pouch | 103% | Altria |
| Volt | 8.3 mg | 8.9 mg/pouch | 107% | Altria |
| Rogue | 6 mg | 6.2 mg/pouch | 103% | Altria |
| ZYN | 6 mg | 5.4 mg/pouch | 90% | Altria |
| Dryft | 7 mg | 5.7 mg/pouch | 81% | Altria |
| Velo (can) | 7 mg | 4.6 mg/pouch | 66% | Altria |
| on! | 2 / 4 / 8 mg | 1.70 / 3.73 / 7.57 mg | 85 / 93 / 95% | CDC |
| Dryft | 4 / 7 mg | 2.76 / 5.37 mg | 69 / 77% | CDC |
| Velo Berry Frost | 6 / 10 / 14 mg | 5.6 / 9.5 / 13.0 mg | 93 / 95 / 93% | Johns Hopkins |
| Velo Polar Mint | 6 / 10 / 14 mg | 5.1 / 8.4 / 12.4 mg | 85 / 84 / 89% | Johns Hopkins |
| Velo (8 other variants) | 10 mg | 8.0–8.7 mg | 80–87% | Johns Hopkins |
Read down that table and the pattern is a brand pattern, not a strength pattern. Dryft measures low in two unrelated labs. on! measures high in two unrelated labs. And within on!, accuracy actually improves as the strength rises — 85% at 2 mg, 93% at 4 mg, 95% at 8 mg — which is the opposite of what you would expect if makers were quietly skimming the strong products.
Why Velo is the wrong scandal
You will see the Velo figure quoted as a 70% shortfall. The authors who produced it do not think it is one, and said so:
In the ONP brand VELO sold in cans, nicotine was nearly 70% lower than the label when analyzed using the GC/MS assay. However, the measured nicotine released at 60 minutes was similar to the label, suggesting that methanol did not efficiently extract nicotine from this formulation.
And the same brand in different packaging behaved normally: “Interestingly, both VELO flavors purchased in tins had comparable measured nicotine content to the label.” When the nicotine is bound into a matrix the solvent cannot fully strip, a low number measures the solvent, not the product. That is a real finding about methods and a bad basis for a consumer warning. If you want a brand-named shortfall that survives scrutiny, Dryft is the one — low in two labs, with two different protocols.
Six milligrams of what?
The deeper problem is not accuracy. It is that the unit is undefined. Germany’s Federal Institute for Risk Assessment bought 44 pouch products, read the packaging and reported this:
Nicotine content, expressed as mg per pouch or per g, was declared clearly on only about a third of the nicotine pouches examined.
With the consequence spelled out: “One reason could be that some manufacturers meant the nicotine content per pouch and others per gram. However, this fact is not apparent to the consumer.” And because pouch masses differ enormously, the two units are not close together. BfR: “Switching between products from different manufacturers can result in nicotine content per pouch doubling even though the nicotine strength of these products is described using the same terms.”
That is a federal regulator stating that the number on the can may mean either of two things and you cannot tell which. We have a separate guide to the mg-per-pouch versus mg-per-gram problem; what the measurement literature adds is how far apart the two can get in practice. In Altria’s table, on! holds 8 mg in a 0.25 g pouch and Rogue holds 6 mg in a 0.69 g pouch. That is roughly 33 mg/g against 9 mg/g — a 3.6-fold difference in concentration between two cans whose labels differ by 2 mg.
Three labs, three protocols, three different quantities
Nothing here is a single measurement being done well or badly. The labs are measuring different things, by design.
| CDC | Roswell Park | Altria | |
|---|---|---|---|
| Sample | 0.4 g of filler; pouch cut open, wrapper discarded | one whole intact pouch, wrapper included | whole pouch, CORESTA method 62 |
| Solvent | 2N NaOH, then MTBE | methanol | per CORESTA 62 |
| Detector | GC-MS | GC/MS | GC-FID |
| pKa used for free base | 8.02 | 8.01 | not calculated |
CDC’s protocol is explicit that the fleece goes in the bin: “For pouch products, the pouch material (organic cellulose compounds that create a non-woven fleece) was cut open and the nicotine binder/filler was weighed.” Roswell Park weighed the pouch whole, wrapper included, and separately ran a release test — “a pre-weighed pouch was mixed for up to 30 minutes in 10 mL Milli-Q water, following a modified version of CORESTA Recommended Method 69”. Those two choices alone can move a per-gram figure by a third, because a 0.3 g pouch’s wrapper is not a rounding error.
Add the three published values for nicotine’s dissociation constant — 8.02 at CDC, 8.01 at Roswell Park, “a pKa value of 8.0” at BfR — and every “free nicotine” percentage you have ever read is arithmetic from pH with a constant borrowed from a chemistry handbook. I wouldn’t treat a free-base figure quoted to three significant figures as a measurement, because it isn’t one.
Pouch mass is the variable nobody prints
If you want one number that explains more than the strength does, it is the mass of the thing. Altria weighed twelve pouches per product and found a range from 0.25 g to 0.69 g — a 2.8-fold spread across seven mint products. Within-product consistency varied too: ZYN’s pouches had a standard deviation of 0.002 g, Velo’s 0.043 g.
Three labs agree on this to within a few per cent, which is worth saying because so little else in this field replicates. on! measures 0.25 g at Altria and an implied ~250 mg at CDC. Dryft: 0.37 g and ~350 mg. Rogue: 0.69 g at Altria, 0.6 g at Roswell Park. BfR, pooling four studies, puts the median pouch at 0.6 g and cites a Dutch monograph in which “pouch weights of 0.25 to 0.8 g were described”. Roswell Park counted roughly 18 pouches per can.
One thing still unmeasured: whether pouches inside a single can vary. Roswell Park says so outright — “we conducted measurements within the same container, so variability between containers was not assessed in our study” — and Altria does not state how many cans its twelve pouches came from. So “is my can consistent?” has no published answer. The arithmetic linking pouch weight to content is sound; the input is just not on the label.
Nobody has to print it, so nobody audits it
This is the part that reframes the whole question. In the United States, 15 U.S.C. § 4403 requires every maker of smokeless tobacco to give the federal government, annually, “a specification of the quantity of nicotine contained in each such product.” Then the same section requires that the figure be kept from you: information provided under it “shall be treated as a trade secret or confidential information subject to section 552(b)(4) of title 5”.
What the law does require on the package is a warning, not a quantity. The four mandated smokeless statements under § 4402 — “WARNING: This product can cause mouth cancer” through to “WARNING: Smokeless tobacco is addictive” — contain no number. For tobacco-free pouches, 21 CFR part 1143 requires only “WARNING: This product contains nicotine. Nicotine is an addictive chemical.” I grepped § 4403, § 4402 and part 1143 for “tolerance”, “milligram”, “per pouch” and “declared”: zero hits each.
The pre-market application rules are stranger still. 21 CFR part 1114 makes an applicant “provide target specification with upper and lower range limits for” a list of design parameters that includes “Portion mass (mg).” and demands test data on “Nicotine dissolution rate (%/min).” Nicotine content per portion appears in neither column — “Nicotine content” returns zero hits in the whole part. FDA will hold a maker to a tolerance on how much filler goes in a pouch, and to the rate at which nicotine comes out, but asks for no target range on how much nicotine is in it.
Sweden, the home market of the oldest products in this category, goes further in the same direction. LIVSFS 2024:2 requires a consumer pack of snus to carry net quantity in grams (“1. Nettokvantitet i gram.”), the manufacturing date, storage instructions and the maker’s name. Across that regulation and the two governing statutes — Lag (2018:2088) and Lag (2022:1257), nearly half a megabyte of primary law — the Swedish words for nicotine content, nicotine strength and milligram appear exactly zero times, case-sensitively. Swedish law caps lead and nitrosamines in snus by the milligram per kilo. It says nothing about the nicotine.
So the honest answer to “is label accuracy regulated?” is that there is nothing to regulate. The declaration is a marketing choice. Across every statute, regulation and paper I hold, searched for “audit”, “recall”, “market check” and “enforcement action”, not one regulator runs a programme that tests pouches against their declared strength and acts on the result. The two government datasets that exist — CDC’s and BfR’s — are characterisation studies that report the deviations and recommend nothing.
What I would actually do with a label
Treat the printed figure as a tier marker with a band around it, not a dose. On the evidence, a well-behaved product sits somewhere between 80% and 100% of its number, and the direction of error is downwards far more often than up. That is accurate enough to rank products and too loose to compare two brands two milligrams apart.
Two practical consequences. First, compare like with like: the same brand’s 3 mg against its own 6 mg is a meaningful comparison, because whatever systematic offset that maker has applies to both. Across brands it does not. Second, if a can gives you a strength tier and a number — Volt and Loop both label themselves “Strength: 3” at 8.3 mg and 9.4 mg respectively — the tier is telling you about the brand’s own ladder and nothing about anyone else’s. Comparing strength scales across brands only works once you put them on the same axis.
And remember that content is not delivery. Altria’s own data makes the point against its own marketing interest: “higher nicotine content does not result in a higher percentage of release”. Rogue released 72% of its nicotine in an hour, on! released 98%, and the two cans are two milligrams apart. What reaches you is a different question from what is in the pouch, and the label speaks to neither.
Frequently asked questions
Are nicotine pouch labels accurate?
Mostly, within about 20%, and usually on the generous side. Four labs covering roughly 90 labelled products put the typical result at 80–95% of the declared strength, with a minority above it. The large misses in the literature belong to lozenges and toothpicks, not pouches.
Do any pouches contain more nicotine than the label says?
Yes. In Altria’s dataset three of seven measured above label (on! 103%, Rogue 103%, Volt 107%), and CDC found one product in 70 more than 58% above its indicated strength — a chewing gum, not a pouch. The direction is “usually below, sometimes above”.
Is a maker legally required to print the nicotine content?
Not in the United States and not in Sweden. US federal law requires warning statements, and requires the quantity to be reported confidentially to the government. Swedish law requires net weight in grams and no nicotine figure at all.
Why do two 6 mg pouches feel different?
Because the number does not control what reaches you. pH, the free-base fraction, pouch mass, filler composition and how long you keep it in all move delivery independently of content. We cover this at length in why two 6 mg pouches from different brands feel completely different.
Does the label mean milligrams per pouch or per gram?
Usually per pouch on tobacco-free products, but it is frequently unstated, and BfR found a clear declaration on only about a third of the products it examined. If the can gives only a bare number and a tier, you cannot tell from the packaging which unit is meant.
Sources
- Tran H, Tyx RE, Valentin L, Mahoney M, Stanfill S, Watson CH. Total and unprotonated (freebase) nicotine content in new types of oral ‘tobacco-free’ nicotine products. Tobacco Control, PMID 39542719. CDC Division of Laboratory Sciences; no specific grant declared, no competing interests declared. 70 products, purchased 2020.
- Page MK, Leigh NJ, et al. Comparison of Nicotine and Selected Flavoring Contents Between Tobacco-Free and Tobacco-Containing Oral Pouches. Nicotine & Tobacco Research, PMID 40383714. Funded by the National Cancer Institute (U54 CA228110, P01 CA200512); one author discloses a Pfizer grant and a Johnson & Johnson advisory board outside the work.
- Cohen JE et al. Nicotine, flavour and other chemical constituents of Velo nicotine pouches sold in Pakistan. Tobacco Control, PMID 40216552. Funded by the Institute for Global Tobacco Control, Johns Hopkins, with Bloomberg Initiative funding (Grant No 125086); corresponding author has been a paid consultant in litigation against a tobacco company.
- Platt SP, Gonzales CM, et al. Dissolution and physical characterization of oral nicotine pouch products. Scientific Reports, PMID 41495155. Funded by Altria Client Services LLC, an affiliate of the manufacturer of on!; all eight authors were Altria employees at the time of the study. The Loop label figure in that paper is not taken from the can.
- German Federal Institute for Risk Assessment. Updated BfR Opinion no. 023/2022: health risk assessment of nicotine pouches, 7 October 2022. Regulator’s own analysis of 44 products.
- 15 U.S.C. § 4403 — Ingredient reporting, and § 4402 — Smokeless tobacco warning. Current through Public Law 119-136 (October 22, 2026).
- 21 CFR part 1114 (premarket tobacco product applications) and part 1143 (required warning statements), as in force 30 October 2026.
- LIVSFS 2024:2, Swedish Food Agency regulations on snus, snus-like products and chewing tobacco; Lag (2018:2088); Lag (2022:1257).
Last verified 3 November 2026. Every quotation above was checked against the source document as downloaded on that date. The percentage-of-label columns and the mg/g conversions are my own arithmetic from the papers’ published figures and are labelled as such.
