Potassium nitrate mouthwash versus toothpaste: two small studies that disagree, and two literatures that measure different things.

Nobody can tell you whether a potassium nitrate mouthwash works better or worse than a potassium nitrate toothpaste, because only two studies have ever put the two in the same trial and they came to different answers¹². Everything else on the subject compares two separate bodies of evidence across a gap: the toothpaste side and the mouthwash side come from different groups, at different concentrations, against different comparators, and they do not agree about which measurements move. S3 Sensitivity Science™ is a twice-daily toothpaste rather than a rinse, and it carries potassium nitrate alongside two hydroxyapatites and full adult-strength fluoride in one tube.

What was checked25 peer-reviewed studies, S3 product information, Oral Health Foundation and NHS guidance

Has anyone put a potassium mouthwash and a potassium toothpaste in the same trial?

Twice, and the two answers do not match. The first is a four-week clinical study, not randomised as its own abstract describes it, in thirty patients split fifteen and fifteen: one arm brushed with a 5% potassium nitrate toothpaste, the other rinsed with a 3% potassium nitrate mouthwash, both arms improved significantly on a visual analogue scale by two and four weeks, and there was no significant difference between them — with no values published in the abstract at all¹. The second is a one-month clinical study in seventy-five adults with five arms, two of them a potassium nitrate toothpaste and a 3% potassium nitrate mouth rinse; there, every arm improved significantly from its own baseline except the potassium nitrate mouth rinse, which did not².

Read them together and the honest summary is short. One says the two forms are indistinguishable. The other says the rinse did not clear the bar the paste cleared. Both were run in India, neither had a placebo arm, neither says how patients were allocated, neither published the numbers behind the significance, and between them they hold about a hundred people. That is the entire head-to-head literature on the question this page exists to answer.

Two further trials come close without arriving. A four-week randomised study of a calcium sodium phosphosilicate paste against the same active in a mouthwash — one ingredient, two vehicles, one trial — found air-blast scores down 68.53% with the toothpaste and 48.52% with the rinse, cold-test scores down 56.38% and 38.87%, and no statistically significant difference between the two forms in twenty-eight patients³. It is the cleanest design anyone has run for this question, and its active is not potassium. And a four-week randomised trial whose authors are employed by the manufacturer of the rinse it tested put a 1.4% potassium oxalate mouthrinse and a potassium nitrate toothpaste against one shared negative control: the trial found sensitivity down 110% in the toothpaste arm and 80% in the rinse arm relative to that control — a figure that can pass one hundred because it is measured against the control arm rather than against the starting score — with no significance test between the two active arms reported⁴. The salts differ, so the vehicle is still not isolated.

Here is how the searching was done, so that you can repeat it. Four searches of PubMed on the tenth of September 2026, alongside the one the planning session had already run.

Four distinct studies in that whole set carry both a rinse and a paste. Two of the four use potassium nitrate in both.

What has the toothpaste evidence found?

A longer record, over a longer period, with a consistent split between what the instruments show and what the patients report⁵⁶. The 2001 Cochrane review pooled four of the eight randomised trials it found and reported a significant effect on air-blast sensitivity (SMD -1.51, 95% CI -2.09 to -0.94) and on tactile sensitivity, while the subjective assessment at six to eight weeks was not significant; three of the four trials it could not pool showed no effect, and the reviewers concluded that no strong evidence supported the toothpaste⁶. The 2006 update pooled six randomised trials in 390 participants, reported the air-blast effect at SMD -1.25 (95% CI -1.65 to -0.85), again found nothing significant on the patients' own assessment, and again concluded there was no clear evidence⁵. That review has not been updated since.

The individual placebo-controlled trials of the five per cent dose are more positive than the syntheses built on them⁷⁸⁹. A double-blind randomised trial in thirty-six adults found significant reductions on all three measures at weeks four, eight and twelve, with complete relief of subjective symptoms in 67% of the treated group against 6% of the control⁷. A randomised trial in sixty-seven adults found significant improvements against placebo on tactile, thermal and air-blast measures at six and twelve weeks⁸. A randomised trial in forty-eight adults found the same on tactile and thermal measures at four and eight weeks⁹. A double-blind randomised trial in one hundred and twenty adults, comparing three marketed products with no placebo, found the potassium nitrate arm down 35% on air and 34% on water at two weeks and still improving at twelve¹⁰. This is the evidence behind S3's dose: 5% potassium nitrate is the concentration those trials used.

The two most recent large analyses are cooler. A 2019 network meta-analysis of thirty randomised trials found no significant difference between potassium toothpaste and placebo, and a significant placebo effect across the network¹⁴. A 2026 network meta-analysis of ninety-three randomised trials placed potassium with or without fluoride at a mean difference of -0.42 (95% CI -0.77 to -0.07) against a benchmark fluoride paste on the evaporative Schiff score at two weeks, from a single contributing study, at low confidence¹³. A 2015 systematic review and meta-analysis of thirty-one randomised trials had put potassium against placebo at SMD -1.28 (95% CI -2.05 to -0.51), with heterogeneity between 86% and 95%¹². And a 2025 double-blind randomised trial of one hundred and twenty adults funded by Procter & Gamble, seven of whose eight authors are its employees, found a marketed potassium nitrate paste ahead of a non-desensitising control at every visit on the Schiff score, and behind the stannous fluoride paste at two weeks¹¹.

The honest summary is the contradiction itself: the instruments improve in most of this work, the patients' own ratings often do not, and the syntheses disagree about what that adds up to⁵⁶¹²¹⁴.

What has the mouthwash evidence found?

A smaller, younger literature that improves on what patients say and stumbles on the instruments, which is the opposite of the pattern above. The first study, in 1996, was a six-week multi-centre double-blind randomised trial of forty-seven adults comparing a 3% potassium nitrate silica and sodium fluoride rinse with a fluoride and silica rinse without the potassium: both arms improved highly significantly from baseline on the tactile probe, the between-group difference at six weeks was significant on the probe and on subjective air scores, and there was no significant difference between the groups on subjective probe scores¹⁵. The second, in 2001, was a six-week double-blind randomised trial of fifty adults comparing a 3% potassium nitrate and 0.2% sodium fluoride rinse with a 0.2% sodium fluoride control: sensitivity fell in both arms on all four assessment methods and the difference in the fall favoured the potassium rinse significantly¹⁶. Neither published a single value.

Then a manufacturer's programme. Three eight-week examiner-blind randomised trials of the same experimental 3% potassium nitrate mouthwash, all funded by GSK Consumer Healthcare with GSK employees among their authors, all comparing rinse-plus-fluoride-toothpaste against the same fluoride toothpaste alone¹⁷¹⁸¹⁹. In Study 1, funded by GSK, in two hundred and sixteen adults, the rinse arm was significantly better on the evaporative stimulus and on tactile threshold at four and eight weeks¹⁷. In Study 2, also funded by GSK, in one hundred and thirty-five adults, both arms improved on every clinical measure, the rinse arm improved more, and the between-group difference reached significance only for the evaporative stimulus at week four, with no difference at all in quality-of-life scores¹⁸. In the third, an examiner-blinded randomised trial published in 2019 in one hundred and ninety-one adults, tactile sensitivity improved from baseline only in the rinse arm and every between-treatment comparison favoured the rinse; that paper's record carries no funding statement, and its first two authors are the same industry researchers who ran the other two¹⁹.

Set against the programme is the one trial that gave the control group a rinse to use. A 2016 double-blind randomised trial of eighty-two adults compared a mouthwash whose main component was potassium nitrate against a placebo mouthwash and found no significant difference between them on the cold test, the tactile test, the compressed-air test or self-reported daily pain, while both arms improved significantly from baseline on all four²⁰. That is worth pausing on. Comparing a rinse against no rinse at all cannot separate what is in the bottle from the act of using it.

The two pooled analyses of these trials then disagree with each other. A 2023 systematic review and meta-analysis of nine randomised trials of potassium-salt mouthwashes found the rinses ahead of control on the tactile test at two, four and eight weeks and on the evaporative test at four and eight weeks, with no difference in side effects²¹. A 2017 systematic review and meta-analysis of seven trials, five of them potassium nitrate, found a significant reduction in the patients' own reported pain at eight weeks (SMD 0.77, 95% CI 0.23 to 1.31, p=0.005) and no significant effect on tactile or thermal measures²². Two teams, largely overlapping evidence, opposite conclusions about which outcomes move. This page reports that as the finding rather than splitting the difference between them.

Why do the two literatures disagree about which outcomes improve?

Because the two halves of the table below were not built to be compared, and three things differ across them besides the vehicle.

The first is the comparator. The toothpaste work mostly ran against a matched placebo paste, which is a hard test. The largest mouthwash work ran against no rinse at all, which is an easy one. The single mouthwash trial that used a placebo rinse found nothing²⁰.

The second is who holds the instrument and what it measures. A tactile threshold read off a Yeaple probe by an examiner, and a Schiff score called by that same examiner, are not the patient's experience; a visual analogue scale filled in by the patient is. The Cochrane reviewers were explicit that the potassium toothpastes moved the first and not the second⁵⁶, which is a warning about the instruments as much as about the ingredient.

The third is the size of the improvement that happens anyway. In a 1998 double-blind randomised trial of ninety adults whose funding is not stated in its record, a potassium citrate, cetylpyridinium chloride and sodium fluoride rinse was compared with the identical base rinse minus the actives; both arms improved highly significantly, the sensitivity difference between them never reached significance, and the authors present the result as a direct measurement of the placebo response in this field, at 30% to 40%²³. A 2019 network meta-analysis of thirty randomised trials found the same placebo effect across a much larger evidence base¹⁴. When an inactive rinse buys a third of the benefit on its own, every arm improving from baseline tells you very little, and a good deal of what is written about desensitising mouthwashes rests on exactly that pattern.

Are you comparing forms, or concentrations?

Concentrations, at least partly, and there is no way around it with the studies that exist. Every potassium nitrate mouthwash in the table above was tested at three per cent. Every potassium nitrate toothpaste in it was tested at five per cent. Nobody has published a trial of a five per cent rinse, and nobody has published a trial of a three per cent paste. Even the head-to-head studies inherit the confound: the first put five per cent in the tube against three per cent in the bottle, so a difference between its arms, had there been one, could have belonged to either variable¹.

That matters, because there is no dose-ranging trial of potassium nitrate in a toothpaste either. Five per cent is the concentration the published trials used, not a concentration anyone has ever compared with a second one; three per cent is a mouthwash convention with the same absence behind it. Reading the higher figure as the more powerful product is arithmetic, not evidence, because the study that would settle it has not been run.

What is physically different about a tube and a bottle?

Several things, and only one of them has been measured directly for potassium nitrate. In a laboratory study on ninety dentine samples cut from forty-five extracted premolars, potassium nitrate applied as a toothpaste occluded more tubules than the same active applied as a mouthwash, and twice-daily application occluded more than once-daily, at three, seven and fourteen days; the authors state that randomised trials would be needed to confirm any effect in a person²⁴. That is a laboratory observation about a vehicle, and this page will not turn it into a benefit to anybody, particularly since potassium nitrate is not thought to work by occlusion in the first place.

The rest of the table below is either arithmetic anyone can check or an honest blank.

The products behind those two columns are not equally documented, and this page says so rather than filling the gap. The September 2026 category review recorded the ingredient lists of UK sensitivity toothpastes and did not record mouthwashes, so the only product that can be described here by its declared actives is ours.

Should you add a mouthwash to a paste like S3, or does it wash it off?

That pairing has never been put in an arm of its own, so what follows is guidance and measurement rather than evidence about the combination itself. The Oral Health Foundation's advice on sensitive teeth is to spit after brushing and not to rinse with water, and its advice on mouthwash is to use a mouthwash at a different time of day from brushing unless your dental team says otherwise²⁶²⁷. Both lines exist because whatever you put in your mouth after brushing displaces what the brushing left there. That has been measured, though for the wrong ingredient: in a double-blind randomised trial of one hundred and twenty adults, the post-brushing rinsing routine had a statistically significant effect on how much fluoride stayed in the saliva afterwards, alongside the choice of toothpaste and the time elapsed²⁵. The equivalent measurement for potassium nitrate does not exist.

Here is the sentence this page owes you, and it costs us something. The whole head-to-head literature is two small studies that disagree, neither of them telling you how patients were allocated, neither publishing a number, and together they hold about a hundred people¹² — so anyone who tells you a paste beats a rinse, or a rinse beats a paste, is going beyond the evidence, and that includes us. What we can say is narrower and checkable: the toothpaste form is the one with the larger trial record behind it, at the concentration those trials used, and that is the form S3 takes.

The practical answer, then, is not about potassium at all. If you want to use a mouthwash, use it away from brushing — mid-morning, after lunch, whenever suits — and keep the twenty minutes after brushing for the toothpaste. That rule applies to every sensitivity toothpaste on the shelf, this one included, and it is the single change on this page that costs nothing and might alter what you feel.

What should you do, and when is neither the answer?

Give the toothpaste the time the trials gave it. The five per cent dose was judged at four, eight and twelve weeks in the placebo-controlled work, not at day three⁷⁸⁹, and the Oral Health Foundation's advice is to keep using a sensitive toothpaste to maintain the effect²⁶. If you want to add a rinse on top, the manufacturer-funded mouthwash studies that found something used theirs twice daily for eight weeks alongside brushing, and none of them put a rinse in place of a desensitising paste¹⁷¹⁸¹⁹.

Some symptoms should not be handed to either form. Toothache that has lasted more than two days is a reason to book an appointment, on the NHS's own advice²⁸. The Oral Health Foundation adds severe pain, sensitivity that runs past a few weeks, a single affected tooth, and pain that arrives suddenly, any of which can point to decay, a crack, gum trouble or infection²⁶. A lone tooth that has begun aching by itself has a dental cause with a name, and neither a tube nor a bottle is the treatment for it.

Of the sensitivity toothpastes recorded in the September 2026 category review of 51 products, S3 is the only one whose formula combines potassium nitrate, hydroxyapatite and fluoride, which you can check row by row on the category review. For the mechanism behind the two actives paired in that formula, see how sensitive toothpastes work and, for the wider routine, how to stop sensitive teeth pain.

Frequently asked questions

Is a desensitising mouthwash as good as a sensitivity toothpaste?

Nobody knows, and the two studies that tried to find out disagree¹². One found no significant difference between a 5% potassium nitrate toothpaste and a 3% potassium nitrate mouthwash in thirty patients over four weeks¹; the other found the toothpaste arm improving significantly from its baseline over a month while the rinse arm did not². What can be said is that the toothpaste form has the longer record behind it, and that the biggest mouthwash studies, all funded by their manufacturer, gave the rinse to people who were brushing anyway¹⁷¹⁸¹⁹.

Can I use a mouthwash with S3?

No trial has tested a desensitising mouthwash alongside a desensitising toothpaste, so there is no evidence either way and no published rule from us about it. The general guidance is clear enough to act on: the Oral Health Foundation advises spitting after brushing without rinsing, and using a mouthwash at a different time of day from brushing unless your dental team advises otherwise²⁶²⁷. That is about timing, not about any one paste, and it applies whatever you brush with.

Why are mouthwashes 3% and toothpastes 5%?

Convention, as far as the published record shows. The mouthwash trials that exist all used three per cent and the toothpaste trials all used five per cent, and no trial has tested either concentration in the other vehicle. No dose-ranging study of potassium nitrate in a toothpaste exists to settle which concentration to prefer, so the two figures record what was tested.

Does rinsing after brushing wash a sensitivity toothpaste away?

It changes what is left in your mouth, and that has been measured for fluoride rather than for potassium. In a randomised trial of one hundred and twenty adults, whether people rinsed with water after brushing had a statistically significant effect on the fluoride concentration in their saliva afterwards²⁵. Nobody has run the same experiment on potassium nitrate. The advice from the Oral Health Foundation is to spit and not rinse²⁶, and a mouthwash used immediately after brushing is a rinse whatever else is in it.

Which one works sooner?

Neither, because nobody has compared how quickly the two forms act. The five per cent toothpaste work measured change from week two onwards and reported its significant results at four weeks and beyond⁷⁸⁹; the manufacturer-funded mouthwash work read its outcomes at four and eight weeks¹⁷¹⁸¹⁹. A question about the first week has no answer in either literature, and a page that gave you one would be inventing it.