Hall 2019: a 3% potassium nitrate rinse over eight weeks in 191 adults.

Hall and colleagues found that rinsing twice a day with a 3% potassium nitrate mouthwash, on top of brushing with an ordinary fluoride toothpaste, reduced dentine sensitivity more than the same toothpaste alone over eight weeks, in an examiner-blinded randomised trial of 191 US adults sponsored by GlaxoSmithKline, the company developing the rinse it tested¹, though the most recent placebo-controlled trial of a potassium nitrate rinse we found, from a university whose authors reported no conflict of interest, found no difference from placebo on any sensitivity measure². It is a trial of a rinse added to brushing, not of a potassium nitrate toothpaste, and there was no dummy rinse, so the people who rinsed knew they were rinsing. S3 Sensitivity Science™ is a toothpaste containing 5% potassium nitrate, not a rinse, so none of these trials is a test of it. And if the pain in your tooth throbs, lingers or wakes you, it may not be sensitivity at all.

What was checked11 peer-reviewed studies, product information as published by each brand

Is your "nerve pain" dentine sensitivity at all?

Often it is not, and that question comes before any trial. Dentine sensitivity is a short, sharp jab when something cold, hot, sweet or acidic reaches exposed dentine, and the Oral Health Foundation describes it as usually brief, sometimes longer⁶. A pain that behaves differently belongs to a different problem.

The Oral Health Foundation lists three signs that sensitivity should go to a dentist rather than a tube: it has gone on for more than a few weeks, it is confined to one tooth, or it arrived suddenly, because each can point to decay, a crack, gum disease or infection⁶. The NHS toothache page sets a shorter clock for pain in general: book a dentist if an ache is still there after two days, if painkillers do not touch it, or if it comes with a temperature, pain on biting, red gums, a bad taste, or a swollen cheek or jaw⁵. Swelling near the eye or in the neck, or swelling that makes it hard to breathe, swallow or speak, is an A&E matter on the same page⁵.

If "nerve pain" means a tooth that throbs, lingers after the cold drink has gone, or keeps you awake, no toothpaste on this page, ours included, is the treatment. Our page on sensitivity that lingers after the trigger has gone and the Journal's two causes of sensitive teeth cover the difference in more detail. Everything below is about the brief, triggered kind.

What did Hall and colleagues test?

They tested a mouthwash added to brushing. Hall C, Sufi F, Milleman JL and Milleman KR published the trial in the Journal of the American Dental Association in 2019 (PMID 30803493): a manufacturer-sponsored, randomised, single-centre, examiner-blinded, parallel-group trial of 191 adults with dentine hypersensitivity, registered as NCT02226562¹. The registry entry for this manufacturer-sponsored trial names GlaxoSmithKline as responsible party, lists a single site in the USA, and gives the two arms: one used a 3.0% potassium nitrate rinse that also contained 0.02% sodium fluoride, twice a day, with a standard toothpaste holding 1000 ppm fluoride as sodium monofluorophosphate, and the other used that toothpaste alone¹.

Three details matter more than the P values. First, the registry lists no placebo or water rinse, so one group rinsed and the other did not. Second, masking was single, with the investigator blinded; nothing in the abstract or the registry says participants were blind, and they could not have been, because they knew whether they had a bottle of mouthwash. Third, the rinse added a little extra fluoride as well as potassium, so the rinse arm differed from the other arm in more than one ingredient, as the manufacturer's own registry entry records¹.

What we could not find: the rinse volume, how long it was held in the mouth, and whether it was used straight after brushing or at another time. Those are in neither the abstract nor the registry, and the full text on the journal site sat behind an access check we could not pass when we tried. The authors' own funding and conflict declarations are in that full text, so we have not read them.

What happened on each measure?

The rinse group improved more on every measure, and the brushing-only group improved on the air test too. On the Schiff scale, where an examiner scores the reaction to a puff of air from zero (no response) to three (painful, asks for it to stop), the manufacturer-sponsored trial's registry entry reports a mean fall of 1.47 points at week eight with the rinse and 0.37 points with toothpaste alone, a difference of 1.10 points (confidence interval 1.28 to 0.92), and falls of 0.94 and 0.31 at week four¹.

The touch test tells a sharper story. In the manufacturer-sponsored trial the examiner pressed a Yeaple probe on the tooth, starting at 10 g and rising in 10 g steps until the person reported discomfort, and the threshold rose by an average of 31.3 g in the rinse group by week eight and by 3.7 g with toothpaste alone, and by 14.1 g and 0.7 g at week four¹. The abstract of that manufacturer-sponsored paper says tactile sensitivity improved from baseline only in the rinse group¹.

On the visual rating scale, the participant's own mark of how sensitive the tooth felt, scores fell by 3.05 with the rinse and 1.05 without it at week eight, and by 1.64 and 0.86 at week four in the manufacturer-sponsored trial¹. This is the one measure the people doing the rinsing scored themselves, and it is also the measure where knowing you are on the "extra" treatment could move the number most.

The control arm's improvement is worth a second look. People who changed nothing except joining a trial and brushing with a standard toothpaste still reported less air sensitivity at both visits, significant against their own baseline (P < .0001) in the manufacturer-sponsored trial¹. That pattern turns up across sensitivity trials, and it is why a trial with no dummy rinse leaves room for expectation in the difference between the arms.

On safety, the registry entry for this manufacturer-sponsored trial lists at least one adverse event of any kind over eight weeks in 11 of the 95 in the rinse arm and seven of the 96 in the toothpaste-alone arm, none in its most severe class; two people in each arm reported tooth sensitivity as an event¹. The abstract calls the treatments generally well tolerated.

Who ran the other trials of this rinse, and what did they find?

GSK ran both of the others, and the independent placebo-controlled evidence on potassium rinses is smaller and mixed. Among the PubMed records for potassium nitrate with mouthwash, mouthrinse or rinse and sensitivity, searched on 2026-09-13, three are manufacturer-sponsored trials of this experimental 3% rinse, each with Hall as first author and Sufi as second¹³⁴.

Study 1 (Hall, Sufi, Wang and Goyal, American Journal of Dentistry 2017, PMID 29178711) randomised 216 adults and found the rinse added to brushing reduced air and touch sensitivity more than brushing alone at four and eight weeks; the study was funded by GSK Consumer Healthcare, which was developing the rinse, and Hall and Sufi were its employees³. Study 2 (Hall, Sufi and Constantin, American Journal of Dentistry 2017, PMID 29251457), 135 adults, was funded by the same company with the same two employees among the authors and separated the groups on one measure only, the air test at week four; both arms improved on everything, and a quality-of-life questionnaire did not separate them⁴. The 2019 trial's PubMed record carries no funding statement, and the registry names GlaxoSmithKline as sponsor¹.

The older trials used a control rinse rather than no rinse. Pereira and Chava (Journal of Periodontology 2001, doi:10.1902/jop.2001.72.12.1720), a double-blind randomised trial in 50 adults at a dental college in India, found a 3% potassium nitrate rinse reduced sensitivity more than a fluoride control rinse over six weeks on all four methods, with no values in the abstract and no funding stated in the record⁷. Gillam and colleagues (Journal of Clinical Periodontology 1996, doi:10.1111/j.1600-051x.1996.tb00526.x), a double-blind multicentre trial of 47 adults from the Eastman Dental Institute, found the rinse ahead of its control on the probe and on subjective air scores at six weeks, but not on the subjective probe scores; no funding is stated in the record⁸.

Hong, Lim and Herr (Journal of Periodontal & Implant Science 2016, doi:10.5051/jpis.2016.46.1.46) ran a double-blind placebo-controlled trial of a rinse containing potassium nitrate, sodium fluoride and cetylpyridinium chloride in 82 adults at Kyung Hee University; both arms improved on cold, touch, air and daily pain, and the rinse did not differ from placebo on any of them, though a gum-bleeding index fell more with the rinse; the authors reported no conflict of interest². The only later placebo-controlled evidence on rinses that this page draws on is a meta-analysis⁹. That meta-analysis, Molina and colleagues (International Journal of Dental Hygiene 2017, doi:10.1111/idh.12250), pooled seven double-blind placebo-controlled mouthwash trials and found people reported less pain at eight weeks (standardised mean difference 0.77), while the examiner's touch and air tests showed no significant effect; no funding is stated in its record⁹.

The pattern in the table is this: the largest and most consistently positive rinse trials compared rinsing with not rinsing and were sponsored by the manufacturer developing the rinse¹³⁴, while the placebo-controlled evidence is thinner and points both ways²⁷⁸⁹. That is a description of the evidence, not a verdict that the rinse does nothing.

What can a rinse trial tell you about a potassium nitrate toothpaste?

Very little directly, because the form, the dose and the comparison all differ. The rinse was 3%; the toothpaste trials pooled by Cochrane all used a 5% potassium nitrate paste against the same paste without potassium (Poulsen and colleagues, Cochrane Database of Systematic Reviews 2006, six randomised trials, 390 participants), and that review found the paste moved the air and touch tests at six to eight weeks (air standardised mean difference −1.25) while people's own ratings did not change significantly, concluding there was no clear evidence for potassium toothpastes¹⁰.

Later reviews of potassium toothpastes disagree with each other. Bae and colleagues (Journal of Clinical Periodontology 2015, doi:10.1111/jcpe.12347), a meta-analysis of 31 randomised trials, found potassium toothpastes ahead of placebo (standardised mean difference −1.28), with very high heterogeneity between trials¹¹. Hu and colleagues (Journal of Dentistry 2019, doi:10.1016/j.jdent.2019.07.008), a network meta-analysis of 30 randomised trials, found potassium toothpaste not significantly different from placebo, and measured a significant placebo effect across the field¹². Gormley and colleagues (Periodontology 2000 2026, doi:10.1111/prd.70088), a network meta-analysis of 93 randomised trials, rated the potassium effect at two weeks small but important with low confidence, and reported that 76% of the potassium studies it found were industry funded¹³. Our page on the Cochrane review on potassium toothpastes reads that review in full.

There is a curious mirror between the two literatures: in the rinse meta-analysis people's own ratings moved and the instruments did not⁹, while in the Cochrane toothpaste review the instruments moved and people's ratings did not¹⁰; our comparison of potassium nitrate mouthwash versus toothpaste takes that argument further.

The table sets the three forms side by side. S3's declared potassium nitrate level is 5%. The 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite on the tube are inclusion levels of the raw materials as supplied, so the paste holds less active hydroxyapatite than those two numbers imply.

S3 puts potassium nitrate in the toothpaste: does a rinse trial tell you anything about that?

It tells you one thing and leaves the rest open. What the three manufacturer-sponsored trials show, taken at their word, is that adding potassium nitrate on top of brushing moved sensitivity scores over eight weeks when the comparison was brushing alone¹³⁴. Those manufacturer-sponsored trials say nothing about how long potassium stays in contact with a tooth from a paste, nothing about a 5% dose, and nothing about any particular toothpaste¹³⁴.

No published trial on this page tested our toothpaste, and no trial we know of has tested potassium nitrate, hydroxyapatite and fluoride together in one paste. Potassium is thought to quieten the nerve's response rather than close the dentine tubules, which is why the formula pairs it with occluding minerals: hydroxyapatite occludes, potassium desensitises, and neither does the other's job. How that mechanism is understood, and where it is still argued over, is on our page on how potassium nitrate is thought to calm a sensitive tooth.

The honest cost of this page for us is plain. The potassium nitrate trials with the largest recent samples tested a rinse on top of brushing and were run with the rinse's maker; none tested a potassium nitrate toothpaste like ours. The evidence that a 5% paste helps comes from older, smaller toothpaste trials and from reviews that do not agree¹⁰¹¹¹².

When is a different product, a dentist, or no toothpaste the better answer?

The company that makes the toothpaste discussed above wrote this page, which is a reason to read this section with some scepticism. There are situations where the answer is something else.

A pain that is not sensitivity. A tooth that throbs, aches for days, hurts when you bite or sits under a swollen gum needs examining, and a desensitising paste or rinse can hide the warning while the cause gets worse⁵⁶. The Journal's guide on how to stop sensitive teeth pain starts from the same place.

Wanting the highest-confidence two-week evidence. In the 2026 network meta-analysis, stannous fluoride toothpaste reduced air-test scores against a standard fluoride toothpaste at two weeks, the only comparison with that benchmark the review rated high confidence; the review also reports that 96% of the stannous fluoride studies were industry funded¹³. Our paste does not contain stannous fluoride: its declared fluoride is sodium monofluorophosphate. Someone who wants the active with the strongest graded evidence should buy a stannous fluoride paste¹³.

Already using a mouthwash every day. For a person who rinses anyway, a potassium nitrate rinse has trials of a few hundred adults behind it, bearing in mind those trials were sponsored by the manufacturer developing the rinse and compared it with no rinse at all¹³. If you go that way, use it at a different time from brushing so you are not rinsing toothpaste off; our page on rinsing, mouthwash and timing covers the order.

No new product. In Study 2, 135 adults brushing with an ordinary fluoride toothpaste improved significantly on every measure over eight weeks, whether or not they added the rinse; that trial was funded by GSK Consumer Healthcare, which was developing the rinse⁴. Mild sensitivity sometimes settles with a standard fluoride toothpaste, gentler brushing and fewer acidic drinks, and trying that for the eight weeks these trials ran costs nothing extra.

Frequently asked questions

Does potassium nitrate mouthwash work for sensitive teeth?

The evidence is split by who ran the trial and what the rinse was compared with²⁹. Three trials of one experimental 3% rinse, all sponsored by the manufacturer developing it and all comparing rinsing with not rinsing, found it ahead on at least one measure¹³⁴. The most recent placebo-controlled trial we found showed no difference on sensitivity², and a meta-analysis of placebo-controlled mouthwash trials found people reported less pain at eight weeks while the examiner's tests did not move⁹.

Is a 3% rinse the same as a 5% toothpaste?

No. The concentration, the time in contact with the teeth and the way it is used all differ, and the rinse and toothpaste reviews on this page each looked at one form only⁹¹⁰. The rinse trials here cannot be read as evidence for a paste⁹.

Who paid for the Hall trials?

The two earlier studies were funded by GSK Consumer Healthcare, which was developing the rinse, and Hall and Sufi were its employees³⁴. For the 2019 trial the PubMed record has no funding statement, and the ClinicalTrials.gov record names GlaxoSmithKline as sponsor¹. We could not reach the full text to read the authors' own declarations.

Should I use a potassium nitrate mouthwash as well as S3?

No trial has tested a potassium nitrate rinse on top of a toothpaste that already carries 5% potassium nitrate, so there is no evidence either way. If a pain is not settling, see a dentist before adding products; the timing of mouthwash and brushing is covered on our rinsing and timing page.

Can potassium nitrate help a tooth nerve that is inflamed?

Not as a treatment. The trials on this page, including the manufacturer-sponsored one read here, enrolled people with dentine sensitivity, the brief pain from exposed dentine¹². A nerve that is inflamed or infected tends to cause lasting or throbbing pain, and the NHS sends that to a dentist⁵.