TSNAs in snus: what the numbers actually mean

A TSNA figure for snus is meaningless until you know which two compounds it covers and whether it is measured wet or dry — and since snus is roughly half water, the same product’s number nearly doubles when you switch basis. Sweden’s statutory limit is 2 mg/kg dry weight for NNN and NNK combined; the industry’s own standard is 0.95 mg/kg “as is”; FDA’s unfinalised proposal is 1.0 µg/g dry weight for NNN alone. Those three numbers are not comparable, and only two of them measure the same thing.

TSNAs — tobacco-specific nitrosamines — are the reason anyone discusses snus manufacturing at all, and they are the single most misquoted set of numbers in this subject. This article is about what the figures mean rather than which product wins. For the wider product, see what is snus; for how the units work on the nicotine side of the label, pouch weight versus nicotine content does the same job for a different number.

What NNN and NNK are

Four of these compounds occur in tobacco; two matter. IARC’s monograph describes them as forming “by the nitrosation of nicotine and other tobacco alkaloids,” and is precise about which precursor gives which: “NNN, NAB and NAT are formed primarily from their corresponding secondary amines (namely nornicotine, anatabine and anabasine) in the early stages of tobacco curing and processing whereas the majority of NNK and some NNN are formed from the tertiary amine nicotine at the later stages of tobacco curing and fermentation.” The step that supplies the nitrite is microbial: as leaf dries, cell membranes break down and “microbes generate nitrite as a by-product, and this becomes available to react with alkaloids to form TSNA.”

IARC classifies the two as Group 1, carcinogenic to humans — and the way it got there is worth knowing, because it is the reason the classification is sometimes disputed. Its working group found “inadequate evidence in humans for the carcinogenicity of tobacco-specific N-nitrosamines” and “sufficient evidence in experimental animals” for both NNK and NNN, then upgraded on mechanism, taking into account that the two “are the most abundant strong carcinogens in smokeless tobacco; uptake and metabolic activation in smokeless tobacco users have been clearly observed” and that combined application induced oral tumours in rats “consistent with their induction by smokeless tobacco.”

The United States has not followed it all the way. The National Toxicology Program’s Report on Carcinogens lists both compounds as “Reasonably anticipated to be a human carcinogen” — its middle category, not its top one — “based on sufficient evidence of carcinogenicity from studies in experimental animals,” and states for NNN that “No epidemiological studies were identified that evaluated the relationship between human cancer and exposure specifically to N-nitrosonornicotine.” Both classifications are current. The gap between them is the mechanistic upgrade: IARC allowed it, NTP did not. Anyone quoting “Group 1 carcinogen” at you is quoting IARC correctly; anyone quoting “reasonably anticipated” is quoting NTP correctly.

The measurement basis, which is the whole problem

FDA defines the two bases in a footnote to its proposed rule: “The term ”wet weight” refers to the weight of tobacco as used by the consumer, while the term ”dry weight” refers to the weight of tobacco after the removal of water.” Snus is wet. IARC records that moist snuff “may contain up to 50% moisture,” and a Massachusetts market survey it reproduces found moist snuff averaging 52.6% moisture across 106 products. Remove half the mass and every concentration in what remains roughly doubles.

FDA does that conversion for Swedish snus explicitly, and it is the best-sourced statement of the magnitude anywhere: “The current GothiaTek standard for NNN and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) (combined) in snus is 0.95 µg/g wet weight tobacco, which would be about 2 µg/g (combined NNN and NNK) dry weight tobacco.” FDA also prints the arithmetic: “[NNN µg/g dry weight] × [1-moisture content] = [µg/g wet weight (as used)]”. That one line is the article in a sentence.

A warning on reading the original. In the published PDF of that rule, the micro sign renders as the letter “m” in the body font, so the body text appears to say “0.95 mg/g” and “1.0 mg/g” where the document means µg/g. The tables in the same document extract correctly, with proper µ, and the preamble spells the unit out in words — “1.0 microgram per gram” — which settles it. The independent cross-check is GothiaTek’s own table, which gives the figure as 0.95 mg/kg; 0.95 mg/kg is 0.95 µg/g. It is a thousandfold error waiting for anyone who copies from that PDF without checking, and I have quoted the correct units above.

WHO’s expert group did not simply pick a basis; it worked through five and published its reasoning, which makes it the most useful source on this question that exists. “Per gram as sold” has the advantage that “the sample tested would be what the user actually encounters,” but “Different products have different moisture levels… Therefore, samples of smokeless tobacco would have to be stored at a standardized humidity before measurement, as is done for cigarettes,” and — the sharp objection — “manufacturers might vary the moisture in their products to achieve the desired regulatory level.” Per gram dry weight avoids “the variation created by differences in moisture content among products and by differences due to the humidity of the storage environment.” Per typical dose used “might approximate the amount of toxicant presented to a user,” but suffers from “substantial variation in the amount actually used by individuals” and “no accepted international standard for typical dose used.”

The group chose dry weight, and attached a caveat that almost nobody quotes alongside the number: it “recommended that the amount of toxicant be normalized per gram of dry weight of smokeless tobacco. This metric is intended to be a product performance standard rather than a measure of human exposure to toxicants.”

The four numbers, and why they do not line up

Instrument	What it measures	Limit	Basis	Per gram of
Sweden, LIVSFS 2024:2 § 8 (statutory)	NNN + NNK	2 mg/kg	dry weight (torrvikt)	product
WHO TobReg, 2009 recommendation	NNN + NNK	2 µg/g	dry weight	tobacco
FDA 2017 proposed § 1132.10	NNN only	1.0 µg/g	dry weight	tobacco
GOTHIATEK (industry standard)	NNN + NNK	0.95 mg/kg	“as is” (moist)	product

All four are the same order of magnitude and none is interchangeable with another. One measures a different analyte. One uses a different basis. Two are per gram of tobacco and two per gram of product.

The Swedish number is the only statutory TSNA limit in the EU, and it is worth quoting because it is routinely cited to the wrong instrument. LIVSFS 2024:2 § 8 provides that “Snus och tuggtobak får inte heller innehålla tobaksspecifika nitrosaminer NNN och NNK (summan) i mängder över 2 mg/kg torrvikt,” and then devotes a further sentence to the measurement itself: “Torrvikt ska bestämmas med en etablerad metod som visat sig kunna fastställa denna halt noggrant.” No other instrument I have read specifies the basis and then requires the basis to be determined properly. Two details follow from the drafting. The nitrosamine sentence binds “snus och tuggtobak” only — not the “snusliknande produkter” in the preceding sentence, so tobacco-free pouches are outside this particular limit. And the instrument in force is the 2024 one; Swedish Match’s own GOTHIATEK page still cites the superseded LIVSFS 2012:6.

GOTHIATEK’s basis is stated on its own page: “In the table below, the limits and the average contents are based on snus, ‘as is’,” with the legend “mg/kg = thousandth gram per kilogram product (based on moist snus).” The limit is 0.95 mg/kg and the reported 2025 average is 0.35, with a 95% confidence interval of 0.05–0.64. Swedish Match claims the standard is tighter than both the Swedish limit and the WHO recommendation — which is true only once you convert, and the conversion is the point: 0.95 as-is on roughly half-water snus is, on FDA’s own reckoning, about 2 µg/g dry, which is at the Swedish and WHO limits rather than comfortably inside them. The word “voluntary” does not appear anywhere on that page, incidentally; FDA is the body that calls it “the GothiaTek voluntary standard.”

FDA’s own proposal is the narrowest of the four and the only one that is not in force. Proposed § 1132.10 reads: “The mean level of NNN in any batch of finished smokeless tobacco product must not exceed 1.0 microgram per gram (µg/g) of tobacco (on a dry weight basis) at any time through the product’s labeled expiration date.” Four things in that sentence do work — NNN alone, per gram of tobacco, dry weight, and a batch mean that must hold to the expiry date rather than a single factory-gate reading. FDA explains the batch mean on the grounds that “the cancer risk is due to long term and repeated exposure, and given the variability of NNN in this agricultural product.”

And the status, because it gets misreported in both directions: it was proposed on 23 January 2017 and it is still a proposal. FDA’s own product-standards page says the agency “continues to evaluate the science on this topic and consider appropriate next steps.” It has not been finalised, and no document says it has been withdrawn.

What the measurements actually show

The Swedish regulator’s own laboratory surveyed the market in 2001 and 2002 and reported “Total TSNA levels of between 0.15 and 3.0 microg/g wet weight,” with means of 1.1 and 1.0 µg/g across 14 and 27 products — a twentyfold spread between brands inside one national market in the same two years. That paper covers all four TSNAs rather than NNN+NNK, and the authors were employees of the Swedish National Food Administration with no grant recorded. FDA’s reading of the same paper supplies the historical arc: in 1983, before GothiaTek, average NNN was 3.8 µg/g and NNK 0.8 µg/g wet weight; in 2002, after it, NNN had fallen to 0.49 and NNK to 0.19.

For American products, the University of Minnesota group reported, on a dry basis, that “Total TSNAs averaged 1.97 microg/g dry weight tobacco in Taboka, Marlboro Snus, and Camel Snus, 4.54 microg/g tobacco in Skoal Dry, and 7.42 microg/g tobacco in traditional brands” — NIH-funded, no conflicts recorded. The widely repeated General-snus figure from that paper is not in its abstract and I have not been able to verify it, so it does not appear here.

FDA’s own laboratory is the only study here that reports water content alongside the nitrosamine, which is exactly what makes a basis conversion possible: across 34 US products, NNN ranged “0.64-12.0 μg/g dry weight” while “The water content of the smokeless tobacco products surveyed ranged from 3.92 to 54.8%.” Intramural FDA funding, no conflicts recorded.

The variability is the headline finding rather than a footnote to it. FDA states it directly: “The levels of NNN in smokeless tobacco products on the U.S. market can vary by several orders of magnitude, not only among different subcategories of products, but also among products in the same subcategory.” Its own table bears that out — within moist snuff alone, 0.66 to 12.77 µg/g dry weight across 28 products. IARC makes the international version of the point, noting that “The ranges vary widely and are product- and country-specific,” with home-made toombak from Sudan reported as high as 3,085 and 7,870 µg/g dry weight.

That IARC passage contains an unintentional demonstration of this article’s thesis, and I think it is the best single illustration available: in consecutive clauses it reports one figure as “μg/g tobacco” and the next as “μg/g dry wt tobacco,” without flagging the switch. If the agency that classified the compounds moves between bases mid-sentence, a reader comparing two numbers from two sources has no chance. WHO’s group said as much about the underlying literature: it “includes measurements made over several decades and with various methods in different laboratories… Some of the differences in reported levels may be due to use of different methods by different laboratories.”

The clearest single demonstration that a per-gram figure and a per-pouch figure can move in opposite directions comes from a Minnesota study of American snus pouch sizes: because the pouches got bigger, “the amounts of total nicotine, unprotonated nicotine, and the sum of NNN and NNK present in the large Camel Snus pouches released in 2010 are 1.9-fold, 2.4-fold, and 3.3-fold higher, respectively, than in the original smaller pouches that entered the market in 2006.” NIH-funded. Concentration can hold steady or fall while the dose per pouch climbs.

Storage, and the one place the rule does not apply

TSNA levels are not fixed at manufacture. WHO’s group recommended that smokeless tobacco “be stored and sold under conditions of refrigeration whenever possible,” because “The concentrations of TSNA increase in smokeless tobacco during storage at room temperature due to microbial action.” FDA’s rule devotes a section to the mechanism: as products age “the water content can change, leading to bacterial growth, and the pH and nicotine content can decrease, causing nitrosamine levels such as NNN to rise,” and “Humidity levels during storage can have an even greater influence than temperature on NNN formation in finished smokeless tobacco products.” An independent measurement on cured leaf found “total TSNA content increasing about 215% for both tobacco types” over a year of natural storage, with “the most rapid increase” in the hot season.

Swedish snus is the documented exception, and the provenance of that exception matters. FDA writes that snus “is usually refrigerated by retailers to maintain its quality and taste but refrigeration is not generally required to maintain stability because modern Swedish snus production techniques achieve very low levels of microbial activity and yield no new nitrosamine formation even when held at room temperature.” FDA’s reference for that sentence is a paper written by Swedish Match’s own scientific affairs group. So the position is: a generic refrigeration recommendation from WHO, and a product-specific exemption from FDA citing the manufacturer. I would not call that settled independent evidence, and I have not been able to read the underlying paper’s figures — its publisher is unreachable from here and its abstract carries none.

What about tobacco-free pouches?

The expected answer is “no tobacco, no tobacco-specific nitrosamines,” and it is nearly right but not cleanly right — and the two available studies point different ways, which is itself the finding.

Swedish Match’s own study of seven oral products found that “The two ZYN products contained no nitrosamines or polycyclic aromatic hydrocarbons (PAHs)” and that “No nitrosamines or PAHs were quantified in the ZYN and NRT products.” All four authors are Swedish Match employees or otherwise funded by the company. Note the verb: quantified, not absent. The limits of quantification decide what that sentence means and they are not in the abstract, so I cannot write “zero” on the strength of it.

Germany’s federal risk assessment institute, which has no commercial interest and funded the work itself, screened 44 nicotine pouch products from 20 manufacturers and reported that “TSNAs were detected in 26 products, with a maximum of 13 ng N-nitrosonornicotine/pouch.” Note the unit: nanograms per pouch — a fifth basis, different again from all four in the table above, and convertible only because the paper also gives a median product weight of 0.643 g. On that median, 13 ng per pouch is roughly 0.02 µg/g, which is around two orders of magnitude below Swedish snus NNN on a wet basis and about the limit of quantification reported for smokeless tobacco. But the median is across all 44 products rather than the weight of the one containing 13 ng, so treat that conversion as indicative and mine, not the paper’s.

The honest summary is therefore two-sided: TSNAs in tobacco-free pouches are two to three orders of magnitude below snus, and “tobacco-free” has not meant “nitrosamine-free” in independent testing of 44 products. The BfR authors’ own conclusion is blunter: “Presence of carcinogenic TSNAs in the nicotine pouches is of serious concern.”

What the product number does not tell you

Three things, each of them sourced to a regulator.

There is no threshold. FDA’s review of the General Snus modified-risk applications states that “since, according to the available toxicological evidence, there is no established threshold level for NNN carcinogenicity, the products pose an increased risk of mouth cancer compared to non-use.” The proposed rule puts the same thing as a model: “As NNN appears to have a genotoxic mode of action, FDA followed the U.S. Environmental Protection Agency’s guidance for carcinogen risk assessment and assumed a linear relationship in the low-dose region of the dose-response model… Using this model, the risk of cancer is linearly reduced as exposure to NNN approaches zero.” There is no level at which the modelled risk becomes nothing.

A lower number is not a safety claim, and both bodies say so unprompted. FDA: “the proposed product standard establishing a limit for NNN in smokeless tobacco products is not intended to communicate that such levels are ”acceptable” or ”negligible” from a public health perspective.” WHO goes considerably further, and the passage reads like a warning against writing this article badly: “Mandating levels and disallowing brands with higher levels from the market does not imply that the remaining brands are safe or less hazardous than the brands that are removed… In particular, ranking of brands by the metric proposed in this report could be interpreted by users as reliable differences in the probable exposure or harm that will result from using different smokeless tobacco brands. Communicating such rankings to the consumer or allowing them to be communicated directly or indirectly is likely to influence user behaviour in ways that will cause harm.” That is why there is no league table in this article. And FDA, applied to snus by name: “Although exposure to HPHCs, including nitrosamines, is lower than many other smokeless tobacco products, exposure is still elevated compared with non-use and, therefore, long-term use of Swedish snus is not without health risks.”

And the user-exposure data point the other way from the product data. This is the most surprising thing in the record and the strongest answer to “what does the number mean for me.” In the PATH study’s first wave, among every-day exclusive users, “the geometric mean urinary NNAL concentration was highest for smokeless tobacco users (993.3; 95% confidence interval [CI: 839.2, 1147.3] ng/g creatinine), followed by all types of combustible tobacco product users (285.4…)” — smokeless users carrying about three and a half times the NNAL of smokers. An independent analysis of fourteen years of NHANES data found the same direction: “Exclusive smokeless tobacco users had higher geometric mean concentrations of serum cotinine… and NNAL (583.0 pg/mg creatinine…) than exclusive cigarette smokers (130.6 ng/mL… and 217.6 pg/mg creatinine, respectively).” Both are CDC and FDA work with no commercial funding.

Two caveats, and they are important enough that the point collapses without them. Both datasets are overwhelmingly US moist snuff, not Swedish snus — FDA’s own figures put moist snuff at 94.1% of the US smokeless market — so neither is a measurement of snus users. And NNAL is a biomarker of uptake, which is a different quantity from the concentration in a tin. The conclusion I draw from putting them side by side is narrow but I think solid: a product-concentration figure and a user-exposure figure are different measurements, and they can point in different directions. Neither one substitutes for the other.

On whether pouches help anyone switch, the most recent systematic review is candid about how thin the evidence is: four studies, “total n = 284; 3 independent, 1 industry-funded; 3 at high risk of bias and 1 at unclear risk of bias,” with the one NNAL comparison rated “very low-certainty evidence.” It was funded by the National Cancer Institute and FDA’s Center for Tobacco Products.

FAQ

Is 0.35 mg/kg good? It is GOTHIATEK’s reported 2025 average, “as is,” against its own 0.95 limit. Whether it is “good” is a question the two bodies that set limits both decline to answer, on the record, in the quotes above. What it is not is comparable to a 2 mg/kg dry-weight figure without converting.

Why does the same product have two different numbers? Because one is measured on the product as you use it and the other after the water is removed. Snus is roughly half water, so the dry figure is roughly double. FDA’s formula is “[NNN µg/g dry weight] × [1-moisture content] = [µg/g wet weight (as used)]”.

Are snus TSNA levels falling? Historically, yes, and substantially: NNN in Swedish market surveys went from 3.8 µg/g wet weight in 1983 to 0.49 in 2002. Whether they are still falling is not something the sources I can read establish.

Does keeping snus in the fridge matter? WHO recommends refrigeration for smokeless tobacco generally, for a stated mechanistic reason. FDA says Swedish snus does not need it — citing a paper by the manufacturer. Refrigeration will not make anything worse.

Do nicotine pouches have TSNAs? In independent testing of 44 products, 26 did, at a maximum of 13 ng NNN per pouch. In a manufacturer’s study of two ZYN products, none was quantified. Both results are real; the first is independent and the second is not.

Does the EU set a TSNA limit? No. The Tobacco Products Directive does not contain the word “nitrosamine” anywhere, because Article 17 requires Member States to “prohibit the placing on the market of tobacco for oral use” — Sweden excepted under its accession terms. The only statutory TSNA limit in the EU is the Swedish one.

Sources

• IARC, Smokeless Tobacco and Some Tobacco-specific N-Nitrosamines, Monographs vol. 89 (2007) — part 7 (the Group 1 evaluation and formation mechanism) and part 6 (moisture content); and vol. 100E part 8.
• National Toxicology Program, Report on Carcinogens, Fifteenth Edition — N-Nitrosamines.
• WHO, Report on the scientific basis of tobacco product regulation: third report of a WHO study group, Technical Report Series 955 (2009), §§ 3.6–3.9.
• FDA, Tobacco Product Standard for N-Nitrosonornicotine Level in Finished Smokeless Tobacco Products, proposed rule, 82 FR 8004, 23 January 2017 (document 2017-01030).
• FDA, Tobacco Product Standards — current status of the NNN proposal.
• FDA, Technical Project Lead review of the General Snus modified-risk applications (2019).
• Livsmedelsverket, LIVSFS 2024:2 § 8, decided 16 February 2024; and Kontrollwiki article 887.
• Swedish Match, GOTHIATEK standard — limits for undesired components.
• Directive 2014/40/EU, Article 17.
• Österdahl B-G, Jansson C, Paccou A, “Decreased levels of tobacco-specific N-nitrosamines in moist snuff on the Swedish market,” J Agric Food Chem 2004;52(16):5085-8, PMID 15291479 — Swedish National Food Administration; no grant recorded.
• Stepanov I, Jensen J, Hatsukami D, Hecht SS, “New and traditional smokeless tobacco,” Nicotine Tob Res 2008;10(12):1773-82, PMID 19023828 — NCI and NIDA funded; no conflicts recorded.
• Ammann JR et al., “A Survey of NNN and Total Water Content in Select Smokeless Tobacco Products,” J Agric Food Chem 2016;64(21):4400-6, PMID 27192054 — intramural FDA.
• Stepanov I et al., “Increased pouch sizes and resulting changes in the amounts of nicotine and tobacco-specific N-nitrosamines,” Nicotine Tob Res 2012;14(10):1241-5, PMID 22259150 — NCI and NIDA funded.
• Shi H et al., “Changes in TSNA contents during tobacco storage,” J Agric Food Chem 2013;61(47):11588-94, PMID 24188376 — authors at a Chinese national tobacco cultivation research centre; no grant recorded. Cured leaf, not finished snus.
• Bäck S, Masser AE, Rutqvist LE, Lindholm J, “Harmful and potentially harmful constituents in two novel nicotine pouch products,” BMC Chem 2023;17(1):9, PMID 36869349 — “All authors are either employees of or otherwise funded by Swedish Match AB.”
• Mallock N et al., “Levels of nicotine and tobacco-specific nitrosamines in oral nicotine pouches,” Tobacco Control 2024;33(2):193-199, PMID 38378209 — all authors at the German Federal Institute for Risk Assessment; “Competing interests: None declared.”
• Xia B et al., “Tobacco-Specific Nitrosamines… in the PATH Study Wave 1,” Nicotine Tob Res 2021;23(3):573-583, PMID 32716026 — intramural CDC and NIDA funded.
• Rostron BL et al., “Nicotine and Toxicant Exposure among U.S. Smokeless Tobacco Users: NHANES 1999–2012,” Cancer Epidemiol Biomarkers Prev 2015;24(12):1829-37, PMID 26582044 — intramural CDC and NIEHS funded.
• Hartmann-Boyce J et al., “Oral nicotine pouches for cessation or reduction of use of other tobacco or nicotine products,” Cochrane Database Syst Rev 2025;10(10):CD016220, PMID 41130602 — funded by the National Cancer Institute and FDA’s Center for Tobacco Products.

Last verified 20 October 2026. Figures taken from the 2017 Federal Register PDF are reported in µg/g; its body text renders the micro sign as “m”, and every figure here was cross-checked against the document’s tables or its spelled-out units. Three things I could not verify and have not asserted: the limits of quantification behind “no nitrosamines quantified” in the Swedish Match study and behind the BfR detection figures, the figures in the manufacturer paper FDA cites for snus storage stability, and a never-user NNAL value from either national dataset. The 2 mg/kg Swedish limit was read from LIVSFS 2024:2 as published; I did not check the amendment register for a later change to § 8.

Nicotine is addictive. No tobacco or nicotine product is safe, and completely stopping is the lowest-risk option. This site is for adults who already use nicotine — 21+ in the United States, 18+ or your local legal age elsewhere. Nothing here is medical advice.