Why 5% potassium nitrate: the dose the trials used, and what nobody has tested.

Five per cent is the concentration of potassium nitrate the published toothpaste trials used, from the placebo-controlled trials of the 1990s to a manufacturer-funded four-arm randomised trial published in 2026: it is the tested dose, not the top of a strength ladder, and no trial has ever compared two concentrations of potassium nitrate in a toothpaste¹¹⁰. The only randomised comparison of two potassium nitrate concentrations anywhere used bioadhesive gels instead of paste, ran for fourteen days in 45 patients, and its authors called the result preliminary¹¹. S3 Sensitivity Science™ declares 5% potassium nitrate on its ingredient list, one of three rows in the September 2026 category review whose published ingredient information gives a potassium nitrate level at all, and one of the only two of those three on sale in the UK, the third being a US listing.

What was checked25 peer-reviewed studies, product information as published by each brand, and guidance from the Oral Health Foundation and the NHS

What does "strongest" mean on a sensitivity toothpaste, and can you read it off the pack?

You cannot, on most packs, because the number that would answer the question is not printed. A UK toothpaste has to publish its ingredient list, and most adult packs give their fluoride in parts per million, but the level of a desensitising active appears only if the brand decides to print it. In the September 2026 category review, twelve rows list potassium nitrate somewhere in their published ingredient information. Nine give no figure for it.

This page compares ingredients and stated actions only, not clinical performance, based on what each brand states about its own formula. Prices and formulations may change; always check the pack. Every row below resolves to a public listing through its CR id, and each was read on the date shown.

Two things follow from the table, and only two. The first is that "strongest" cannot be read off this shelf, because for nine of the twelve there is no number to compare. The second is that where a number is given it is always the same number. Not one row anywhere in the scan declares a potassium nitrate level other than 5%. The strength ladder people are looking for does not exist on the shelf, and the rest of this page is about why it does not exist in the literature either.

A shelf-level version of this question, which takes every desensitising active and not just this one, is in the guide to reading strength off a UK sensitivity pack, and the aisle-by-aisle read of the two big chemists is in sensitive toothpaste at Boots and Superdrug. This page stays with one ingredient and one number.

Where does the 5% come from?

From the trials, and from nowhere else. Potassium nitrate has sat in desensitising toothpastes since the 1970s, and placebo-controlled trials of it date from the 1990s. Those trials were run at 5%, and every trial since has stayed there¹²³⁴. In a twelve-week double-blind trial of 36 adults, a 5% paste improved cold-air, tactile and self-rated scores at four, eight and twelve weeks, and 67% of the treated group reported complete relief of symptoms against 6% on placebo¹. A twelve-week randomised trial of 67 adults found significant improvement against a matching placebo at six and twelve weeks². A multicentre randomised trial reported significant reductions against placebo on every index at four weeks⁴. An eight-week randomised trial of 48 adults tested the same 5% with 1500 ppm sodium monofluorophosphate in a precipitated calcium carbonate base and found significant improvement at four and eight weeks³.

That is the origin of the figure, and it is worth being exact about what it is: a formulation choice that was made once, tested, and then copied. The concentration was not arrived at by testing several and keeping the best.

One number in circulation looks like a second rung on the ladder and is not. Three per cent is a mouthwash concentration, used in rinses tested on their own against fluoride rinses and against no active, and a rinse is a different product with a different contact time. It is not a weaker toothpaste, and it does not give a second point on a dose curve. The comparison between the two forms is covered on the sibling page on potassium nitrate mouthwash versus toothpaste.

Has anyone tested a different concentration?

Not in a toothpaste. This page ran the searches rather than assuming, and the terms are here so that anyone can repeat them.

The last of those five searches is the one that found something, and it is the closest anyone has come. A randomised trial published in 2003 gave 45 patients a bioadhesive gel containing either 5% or 10% potassium nitrate, or a placebo gel, and measured the response to a blast of air over fourteen days: at 48 hours the reduction was 35.8 per cent with the 10% gel against 11.8 per cent with the 5% gel and 13.4 per cent with placebo, and the gap between the concentrations reached significance at 96 hours¹¹. On the tactile measure and on the patients' own ratings there was no significant difference between any of the three. The authors describe their own results as preliminary, the groups held fifteen patients each, and a gel left on the tooth is not a paste brushed on and spat out. It is a real dose comparison, and it is the only one; it cannot be carried across to a tube.

The design is not impossible, which is what makes its absence a finding rather than an oversight. A triple-blind split-mouth randomised trial in 31 patients compared potassium oxalate gels at 5% and 10% over twelve months and found the higher concentration held its effect longer, separating at nine and twelve months¹². Somebody ran the dose comparison for a different potassium salt, followed the patients for a year and published it. For potassium nitrate in a toothpaste, the equivalent trial — three arms, the same base, three concentrations, one stimulus, eight weeks — has never been published. Until it is, "how much potassium nitrate is enough" has no evidence-based answer, and any pack that implies one is going beyond what has been measured.

Do two pastes with the same 5% deliver the same potassium?

There is good laboratory reason to think not, and no clinical evidence either way. This is a mechanism question, and everything in this section is laboratory work on extracted teeth or on dentine specimens: none of it says a person felt better.

The clearest sign comes from a manufacturer's own account of its development programme. A 2004 paper from a manufacturer's research centre describes building a 5% potassium nitrate dentifrice intended to improve on the market leader of the time, which also contained 5% potassium nitrate, and reports that the objective was met through what the authors call a new activated silica technology that raised potassium ion activity in laboratory testing¹³. The whole effort went into the base. The concentration was fixed, because the concentration was the one the trials had used.

Two further laboratory results sit behind that. In an in vitro model, potassium nitrate moved quickly through the dentine matrix out of a silica-based paste, and the deposit left on the dentine was a mixture of copolymer and silica rather than the potassium salt, in work by authors at the same manufacturer's research centre¹⁴. And when 50 extracted human molars were treated for thirty minutes with four different potassium nitrate materials and the pulp cavity was measured by spectrophotometry, the amount of nitrate that arrived differed significantly between the formulations, with the toothpaste giving the highest median reading and the most penetrating material the least viscous¹⁵. Extracted teeth have no pulpal pressure and no outward fluid flow, so this is a statement about formulations, not about mouths.

The trial literature has the same variety inside it. A scoping review of 138 randomised trials catalogued 368 toothpaste formulations and found potassium compounds in 68 of them, the largest single group of actives; across the whole set, silica-based and calcium-based abrasives and two fluoride salts, monofluorophosphate and sodium fluoride, account for most of the formulations¹⁶. The same active at the same level is not the same paste, and the reviews that pool those trials are pooling that variety.

One clinical trial comes as close to isolating it as anything published. In an eight-week randomised trial with 118 completers, funded by a manufacturer whose employees are among the authors, an experimental paste combining 3% copolymer with 5% potassium nitrate reduced cold-air and tactile scores more than a paste carrying the same 5% potassium nitrate alone, at every measured point from week two, while the patients' own pain ratings separated none of the four arms¹⁰. Held constant across those arms was the potassium; what changed was everything around it.

What did the 5% trials actually find?

A mixed record that improves on the examiner's measures more often than on the patient's, and syntheses that disagree with each other about what the pooled result is worth.

Read the findings column and the pattern is the one the syntheses argue about. Six randomised trials went into the 2006 Cochrane review and a split verdict came out: significant improvement on the air blast and on the tactile threshold at six to eight weeks, nothing on the patients' own assessments, and reviewers who would not say the pooled result settled the question²⁰. A 2015 meta-analysis of 31 randomised trials put potassium toothpastes ahead of placebo with a standardised mean difference of −1.28, and reported heterogeneity between trials of 86 to 95%²¹. A 2020 network meta-analysis of 125 randomised trials in 12,541 patients found potassium ahead of fluoride paste on the tactile stimulus with moderate certainty²². A 2017 network meta-analysis of nine trials found only a tendency towards relief for potassium nitrate formulations, without reaching significance²⁴. In a 2019 network meta-analysis, potassium and placebo did not separate²⁵. And a 2026 systematic review and network meta-analysis of 93 randomised trials rated potassium's two-week effect small and low-confidence, while the only high-confidence node in its network was stannous fluoride, which S3 does not contain²³.

Those six syntheses read the same underlying trials. Averaging them would be dishonest. The honest summary is that the examiner's measures improve in most of these randomised trials, the patients' own ratings often do not, and the reviewers who have pooled the same evidence disagree about whether what is left is worth calling an effect²⁰²³. Two of the randomised trials in the table had no placebo arm at all⁶⁷, so part of what they record cannot be separated from the act of being treated. Whichever synthesis is right, none of them is about a concentration: they had only one to pool. The full reading of the Cochrane review is on the sibling page on what the Cochrane review found and what it did not, and how the actives rank against one another is on nerve-calming actives compared.

What does S3 declare, and why does declaring it matter more than the number?

S3 declares 5% potassium nitrate, and that is a decision to match the published evidence rather than a claim to beat the tube next to it. Five per cent is the concentration the trials happened to use; nobody has shown it to be better than three or eight or ten in a paste, because nobody has compared them. Anyone reading this page for permission to believe that a declared 5% is stronger than an undeclared one should not take it: the level is the label, not the formulation, and a manufacturer's own laboratory work, set out above, says the base can change how much potassium is actually available from an identical declared level¹³¹⁵. A paste that publishes nothing may hold the same amount, or more, or less. What can be said is narrower and duller: with a published level you can check the tube against the trials, and with a blank you cannot.

The same applies to the other percentages on this pack. Both of the hydroxyapatite figures, 10% and 5%, are inclusion levels of the ingredient as supplied rather than active content, which is lower, and S3 publishes the two numbers side by side. A brand publishing figures to look impressive would leave the second one out.

One more thing this page will not soften. In an eight-week randomised trial in 101 adults that held the 5% constant and changed only the fluoride salt beside it, the arm carrying stannous fluoride came out ahead of the arm carrying sodium monofluorophosphate on tactile and air-blast measures at four and eight weeks; the test product was the manufacturer's own, and the record does not state who funded the trial¹⁷. Two further eight-week randomised trials of that same test paste, published the same year, put it ahead of a paste combining the same 5% with sodium fluoride¹⁸¹⁹. Sodium monofluorophosphate is the fluoride salt S3 uses, chosen for a reason that has nothing to do with the nerve: it does not react with the calcium in hydroxyapatite inside the tube. That comparison is on the record and it is not softened here, because it illustrates this page's argument better than any level on any pack does: what the potassium travels with has moved outcomes, and the potassium itself has never been tested at another level.

Potassium nitrate is also not the whole of what a sensitive tooth needs. Hydroxyapatite occludes and potassium desensitises; neither does the other's job. What either of them does takes weeks, not days: relief from this class of active is judged at two to four weeks, which is set out on the sibling page on how long potassium nitrate takes to work, and the mechanism itself — still unconfirmed in an intact human tooth — on how potassium nitrate is thought to calm a tooth. The Journal keeps a ranked view of the category, which is a different exercise from this one, and how sensitive toothpastes work explains what happens once any of these actives reaches the tooth.

When is severe sensitivity not a toothpaste question at all?

When it fits a pattern that a toothpaste was never tested against. Four patterns on the Oral Health Foundation's list send a person to a dentist rather than to a shelf: pain that is severe, sensitivity that has lasted more than a few weeks, a single affected tooth, and pain that arrived suddenly. Any of them can signal decay, a cracked tooth, gum problems or infection²⁶. A dentist has options a shelf does not: a desensitising treatment applied to the tooth, a high-fluoride toothpaste on prescription, a filling over a worn area at the gumline, or treatment of the gum disease underneath. The NHS puts a plainer limit on waiting: see a dentist for toothache lasting more than two days, or toothache that painkillers do not settle²⁷.

That is worth holding beside the hunt for a bigger number. The randomised trials on this page recruited people with dentine hypersensitivity — the short, sharp response to cold air or a cold drink, usually across several teeth — and measured them with a calibrated air blast and a probe over eight to twelve weeks¹⁶. Pain that lingers, throbs, wakes you or sits in one tooth was not what was being treated in any of them. A reader who has arrived here because the pain is bad enough to go looking for the strongest thing on the shelf may be looking on the wrong shelf entirely, and the fastest route to relief in that case is an appointment.

Frequently asked questions

Why do most packs not state how much potassium nitrate they contain?

Because nothing requires them to. A cosmetic toothpaste must publish its ingredient list, and most packs add a fluoride figure, but the level of a desensitising active is left to the brand. Nine of the twelve rows carrying potassium nitrate in the September 2026 category review publish no level. That is a labelling choice and nothing more: an undeclared level is not evidence of a low one, and this page does not read it as one.

Would 10% potassium nitrate work better than 5% in a toothpaste?

Nobody knows, and anyone who tells you otherwise is extrapolating. The comparison has never been made in a paste. It has been made once, in a randomised trial of two bioadhesive gels, where the 10% gel reduced the response to an air blast more than the 5% gel in the first four days, with no difference on the tactile measure or on what the patients themselves reported, in 45 people over fourteen days¹¹. A gel that sits on the tooth and a paste that is brushed on and spat out are different products, and one preliminary trial in fifteen patients per group is not a dose curve.

Is S3 stronger than a paste that does not state its potassium nitrate level?

No one can say, including us. The other paste's level is not published, so there is nothing to compare; what differs is that S3's is. A declared level does not make a formula work better — a manufacturer's own laboratory work, above, suggests the base can change how much potassium is available from an identical declared level¹³ — it only makes the claim checkable against the trials that used the same figure.

Can a dentist give you something stronger than a shelf paste?

A dentist can do things a paste cannot, which is not the same as a bigger dose in a tube. The Oral Health Foundation lists a desensitising treatment applied directly to the tooth, a high-fluoride toothpaste on prescription, a filling over a worn area near the gumline and treatment of gum disease among what a dentist may do²⁶. The extra strength in a prescription paste is fluoride, aimed at decay risk, not a bigger dose of a desensitising active.