Best toothpaste for sensitive teeth in the UK: a scoring rubric, not a shopping list.

There is no best toothpaste for sensitive teeth in the UK that anyone can name from evidence, because the trials that would settle it have barely been run: the largest independent randomised comparison of named commercial sensitivity toothpastes located for this page tested three products in 140 adults over six weeks, and every arm improved, the plain fluoride control included¹. What can be done is to score the pack in your hand on the five things its label publishes and the one thing no label on this shelf can offer, so this page states the weights first and then scores all 49 UK sensitivity toothpastes whose ingredient lists were read in September 2026 — S3 Sensitivity Science™ among them, the only one of the 49 whose list carries potassium nitrate, hydroxyapatite and a fluoride salt together, which you can check row by row on the category review. That is a fact about ingredient lists. It is not a result in anyone's mouth, and this page does not treat it as one.

What was checked16 peer-reviewed studies, product information as published by each brand, the Oral Health Foundation and the NHS

Why can nobody hand you the name of the best sensitive toothpaste?

Because that comparison has never been made across this shelf, and the largest independent attempt at it put only three products in one study¹. Two PubMed searches were run for this page, and their terms, dates and results are below so that anyone can repeat them and disagree with what they found.

Those three comparisons are the whole visible literature on the question people actually ask, and they are worth reading one at a time.

The largest with no manufacturer among its authors is a triple-blind randomised controlled trial of 140 adults in which 128 finished: three over-the-counter pastes and a plain fluoride paste, assessed at one minute, three days and two, four and six weeks¹. All four groups improved significantly from baseline at every assessment¹. The arginine paste and the strontium acetate paste were the two that moved fastest, and the trial found nothing to choose between them¹. By six weeks the bioactive-glass paste had produced the largest fall on the water-jet test, and its authors concluded that class could be the best option for managing the condition long term¹. Two of the three products it tested are not on the UK shelf today in the formulation that was tested.

The other two are older and were run by a manufacturer. In an eight-week double-blind randomised controlled trial of 150 adults, whose author list includes scientists employed by the maker of one of the pastes in it, three named commercial toothpastes were compared and the sponsor's arginine paste came out ahead of the other two³. In an eight-week examiner-blind randomised controlled trial of 97 adults, also written by authors employed by one of the manufacturers involved, two pastes that both declared 5% potassium nitrate separated on tactile and air-blast scores, and the difference lay in the second active and the base rather than in the potassium⁴. That last result is the most useful thing in the three, and it cuts against the rubric below: two packs can print the same active at the same level and still not behave the same way.

Then there is the response that arrives whatever is in the tube. In a six-week double-blind randomised trial of 120 adults, a plain fluoride toothpaste improved cold-air sensitivity significantly from baseline and none of the three pastes tested differed from the others⁵. In a subject-blind split-mouth study of 22 adults, a dressing applied over sensitive teeth reduced pain by 95% to a thermal stimulus and 85% to an evaporative one against no dressing at all, which the authors read as evidence that perception itself shifts when a tooth is being treated⁶. That is why "it worked for me" is not a criterion on this page and cannot be one on any page: a person switching pastes has no control arm.

The pooled reviews can settle something narrower, and they do. They can rank ingredient classes against a benchmark fluoride toothpaste at two weeks and say how confident they are². They cannot rank the tubes on a shelf, because they were never designed to, and a 2026 scoping review and Delphi consensus is blunt about why: across 72 papers the definition of dentine hypersensitivity itself varied, few studies applied every part of the accepted definition, and among 58 recent trials the reporting of randomisation, blinding and effect estimates was inconsistent⁷. A field that still had to agree its own case definition in 2026 is not a field that can tell you which of two tubes at Boots to buy.

What should a rubric score, and what is each criterion worth?

Six things, weighted as below, and the weights are this page's judgement rather than a validated instrument — nobody has published a scoring system for this shelf, so this one is offered with its arithmetic visible so you can disagree with it in a specific place. Five criteria read the pack; the sixth reads the literature and scores zero for every product in the review, S3's row included.

Three things follow from the shape of that table. The maximum any UK sensitivity toothpaste can currently reach is eighty out of a hundred. The twenty points nobody can score are the same twenty points for everybody, so they change no ranking and are printed anyway. And a score describes a label and its evidence base; it does not predict what will happen to a tooth.

Criteria one and two: does the pack name an active for the nerve, and one for the tubules?

Two jobs, two lists of words, and most packs on this shelf do one of them. Of the forty-nine recorded products, eighteen carry a potassium salt, thirty carry an occluding active, six carry both, and seven carry neither in their published list. The Oral Health Foundation's sensitive-teeth page names potassium citrate, potassium nitrate and stannous fluoride as the ingredients to look for, describes these pastes as blocking the tiny channels in dentine, and says the effect has to be maintained by continued use⁸. Read that page with its funding in view: it carries an acknowledgement that Oral-B provided an educational grant for the information, and stannous fluoride is Oral-B's own sensitivity chemistry⁸. The advice is sound and the disclosure belongs beside it.

The separation is not theoretical; it has been photographed. In a four-week double-blind randomised trial of 30 adults with a companion electron-microscope arm, a 5% potassium nitrate paste reduced air-blast and cold-water sensitivity against a non-desensitising control by four weeks, and the microscope found the dentine tubules still open underneath that result⁹. An active can quieten a nerve without sealing a thing, and it can seal without going near the nerve. Quieten the nerve while the channels stay open, and every cold drink still gets through to it; plug the channels while the nerve stays jumpy, and it goes on firing at whatever does reach it. Hydroxyapatite occludes and potassium desensitises, and neither of them does the other one's job.

Why is the tubule criterion weighted above the nerve one? Because that is where the graded evidence sits, not because occlusion is a nobler mechanism. In the 2026 network, stannous fluoride's cold-air result was rated high confidence and nano-hydroxyapatite's moderate, while potassium with or without fluoride came in as a small but important effect at low confidence². The six products carrying both kinds of active are Spotlight's paste, S3, SURI's refillable gel, Nura's paste, Boka's Ela Mint and the Davids sensitive paste, and four of those six have dropped fluoride to do it.

Criterion three: does the pack publish how much of the active is in the tube?

Eleven of the forty-nine do; thirty-eight do not, which makes this the emptiest column on the shelf. The eleven are three stannous fluoride pastes declaring 0.454% w/w, one bioactive-glass paste declaring 5% w/w, one arginine paste declaring 8.00%, two potassium nitrate pastes declaring 5%, three hydroxyapatite pastes declaring a figure for their mineral, and S3, which declares a level for all three of its actives. Everyone else prints a fluoride figure, which is a legal requirement, and stops.

The reason this is worth twelve points and not twenty is that a declared level is a fact about disclosure, not about effect. The published trials of potassium nitrate that anchor the family used 5%, and a pack that prints 5% can be lined up against them; a pack that prints nothing cannot be lined up against anything, and you are trusting the front of the box. That is worth something. It is worth less than having the active in the tube at all, and the 2026 consensus review is a reminder of the ceiling: where the trials themselves vary in how they define and measure the condition, a matching percentage promises less than it appears to⁷.

One caution about the levels that are published, including ours. Where a hydroxyapatite percentage appears on any pack, ask what it is a percentage of: 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite on the S3 tube are inclusion levels of the ingredient as supplied, the active hydroxyapatite content is lower than either figure, and S3 states both. Two brands in the review publish a mineral figure with no such basis attached, so their percentages and ours are not directly comparable.

Criterion four: does it keep fluoride, and at how much?

Thirty-six of the forty-nine carry a fluoride salt and thirteen carry none; twenty-seven of the thirty-six declare 1450 ppm, one declares 1350 ppm, one declares 1100 ppm, one declares 530 ppm bound inside a bioactive glass, and six name a salt without publishing any figure at all. Full marks here need both the salt and an adult-strength figure, because a fluoride-free paste is making a trade and a paste that will not tell you its ppm is asking you to take the trade on trust.

The trade is not about sensitivity at all. In a 36-month double-blind randomised trial of 2,222 children, a sodium monofluorophosphate toothpaste prevented decay as well as a sodium fluoride one at the same nominal level, which is the head-to-head evidence sitting behind the whole fluoride column¹⁰. The NHS's toothache page tells you to see a dentist for pain lasting more than two days and treats brushing with fluoride toothpaste as the baseline, not the upgrade¹¹. Every one of the thirteen fluoride-free pastes in this review has opted out of that baseline in exchange for a mineral, and four of them are among the six products that carry both kinds of desensitising active.

Which salt it is matters far less than whether it is there, with one exception worth knowing. In S3's case the 1450 ppm of fluoride is carried as sodium monofluorophosphate, picked because it will not react with the calcium in hydroxyapatite while the tube sits in a cupboard. Two other pastes in the review use monofluorophosphate beside a calcium-based abrasive for the same chemical reason. That is formulation, not efficacy, and it earns no extra points in the table below.

Criterion five: how strong is the published evidence for that ingredient class?

Stronger for some families than the packs suggest, and weaker for others, and the honest way to score it is to take the reviewers' own confidence ratings rather than the size of the effect. Everything in this column is anchored on one paper: the 2026 systematic review and network meta-analysis that pooled 93 randomised trials in 9,548 adults, set each ingredient class against a benchmark fluoride toothpaste at two weeks, and published a confidence grade beside every number it produced². Its own recommendation is worth quoting against the grain of a scoring page: stannous fluoride and arginine as the first things to try at home, chosen on preference, tolerability and what the shop has, and explicitly not on an expectation that either one wins². The same review records how much of the field is paid for by the people selling it, at 96% of the stannous studies and 86% of the arginine ones².

The classes contradict each other across reviews, and the contradictions are printed here rather than averaged. A 2020 network meta-analysis of 125 randomised trials ranked calcium sodium phosphosilicate highest across tactile, cold and air stimuli¹². That is the same bioactive-glass class the 2026 network rates a small effect at low confidence, and the page prints both rather than splitting the difference². A 2015 meta-analysis of 31 randomised trials found every family beat placebo except strontium, which showed no significant effect at all¹³. The 2006 Cochrane review of six randomised trials found potassium's benefit on air-blast and tactile tests at six to eight weeks but not in the patients' own ratings, and concluded there was no clear evidence supporting potassium toothpastes¹⁴. For hydroxyapatite the picture splits by form: a 2019 systematic review and meta-analysis of six four-week randomised trials rated nano-hydroxyapatite's benefit on evaporative and tactile stimuli high quality by GRADE and found no difference on cold stimuli¹⁵, while a 2023 systematic review and meta-analysis of 44 clinical trials, whose authors include scientists employed by Dr Wolff, a maker of hydroxyapatite toothpastes, pooled every hydroxyapatite form and format together at 39.5% against placebo¹⁶.

MD is the mean difference on the Schiff cold-air score against a benchmark fluoride toothpaste in the 2026 network; SMD is the standardised mean difference against placebo in the 2015 review; the certainty column is the reviewers' own grading and the points column is this page's translation of it.

Criterion six: has this product ever been tested against another one on this shelf?

No, for all forty-nine of them, and that is the finding rather than an omission. The three randomised controlled trials of named commercial sensitivity toothpastes described above cover seven products between them, not one of which is on the UK shelf today in the formulation that was tested, and two of the three were written by authors employed by one of the manufacturers involved¹³⁴. Nothing in the literature ranks the pack in your hand against the pack beside it on the shelf.

This is why the criterion is worth twenty points that nobody scores. Leaving it out would have made the rubric look complete; printing it keeps the shape of the answer honest, because the gap it marks is the whole distance between a well-specified label and a tube that will settle your tooth. Every score in the next table is five-sixths of a rubric.

What do the scores look like across the UK shelf?

They spread from seventy-six down to ten, and the spread is mostly about how many jobs a formula tries to do and whether it kept fluoride while doing them. Prices are as shown online on 2026-09-09, and every row resolves to a public listing by its CR id on the category review page. 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.

Five rows in the review are not in that table, and four more products were searched for and not found on the UK shelf at all. A row with no readable ingredient list cannot be scored on a rubric made of ingredient lists, so it is named and left out rather than estimated. Two of the five state their actives in brand copy without publishing the list behind them, which is precisely the position criterion three exists to penalise.

Four patterns are worth pulling out of the arithmetic. The three stannous fluoride pastes that declare their level tie for second on sixty-eight, eight points behind the top and clear of everything else, on a high-confidence class and full disclosure. The fluoride-free hydroxyapatite pastes lose a fifth of the total in one column, which is what puts several genuinely well-formulated products in the thirties and twenties. The £1.00 own-label tube and a £5.00 branded one score the same 38 because their published lists say the same two things, potassium nitrate and 1450 ppm of sodium fluoride, and neither declares its potassium level. And four products that came into this review as sensitivity toothpastes carry no desensitising active in their published list and still score twenty for keeping adult-strength fluoride, which is the rubric working correctly: a plain fluoride paste is a reasonable thing to own and a poor answer to dentine hypersensitivity.

How does S3 score on this rubric, including where it scores nothing?

Seventy-six of the eighty available, and it is worth being exact about which four points it drops and which twenty it never had. S3 takes criterion one on potassium nitrate, criterion two on two hydroxyapatites, criterion three by declaring a level for all three actives, and criterion four on 1450 ppm of fluoride. On criterion five it takes twelve of sixteen, not sixteen, because its strongest-graded class is nano-hydroxyapatite, which the 2026 review rates a large effect at moderate confidence from two trials, while stannous fluoride is the one class it rates high confidence². Every stannous fluoride paste on this shelf therefore beats S3 in the evidence column, and three of them are within eight points of it overall.

On criterion six S3 scores zero, like everything else in the table. Nothing in the published literature has put it beside another product on this shelf. S3's own two pieces of research do not fill that gap and are not admitted to this table: a consumer panel is not a controlled trial and a laboratory report is not a person, so what S3 earns here rests on ingredients, levels and the published evidence for its ingredient classes, and on nothing else.

Does carrying all three, alone among the forty-nine lists, make S3 the best sensitive toothpaste in the UK? No. The review counts what brands print on ingredient lists; it does not measure what happens inside a mouth, and no independent study has put these products against each other. Among 51 sensitivity toothpastes on sale in the UK in September 2026, S3 is the only one whose ingredient list carries potassium nitrate, hydroxyapatite and a fluoride salt together, and the category review page prints all 51 lists with their sources and read dates so the sentence can be checked rather than believed. The nearest product to it uses potassium citrate rather than potassium nitrate and declares no level for either active.

Two more things the rubric rewards are chemistry rather than merit. A water base is what lets one tube hold potassium nitrate and hydroxyapatite at the same time; an anhydrous formula built around bioactive glass cannot carry that pairing. And its actives are formulated at inclusion levels that match the published studies, which is what earns criterion three rather than any claim about the result. The Journal keeps a ranked list of sensitivity toothpastes, this brand's own paste placed first among them; that is the ranked view, this page is the other kind, and where the two disagree the one with its weights printed is this one.

What should you do with a score, and how will you know in four weeks?

Score two or three packs yourself in the aisle, pick the one whose weaknesses you can live with, and then give it long enough to be judged. Scoring in the aisle takes about a minute per pack: find the desensitising active among eight words, note whether a percentage follows it, find the fluoride line and its ppm, and check whether the list is doing one job or two. A pack that scores well on your reading is a pack that has told you what is in it and belongs to a class with published evidence behind it. It is not a promise.

Then set the clock properly, because the published time points are not the ones the packaging implies. The 2026 network meta-analysis measured its comparisons at two weeks². The Cochrane potassium review measured at six to eight weeks¹⁴. A 2019 systematic review of 74 randomised trials grouped treatments by when their reduction reached significance, and placed hydroxyapatite and stannous fluoride among the classes with a significant medium-term effect by one month and potassium nitrate only in the long-term group¹⁷. Relief from any of these builds as the actives accumulate, so judge a paste at two to four weeks of brushing morning and night, not after a weekend. Use it as your normal toothpaste, spit and do not rinse, and keep using it, because the Oral Health Foundation is explicit that the effect stops when the use stops⁸.

Stop shopping and get an opinion if the pain is severe, if it has run on for more than a few weeks, if it sits in one tooth, or if it started suddenly, because those are the patterns the Oral Health Foundation lists as possible decay, a cracked tooth, gum problems or infection⁸. Plain toothache has a shorter clock: the NHS says more than two days is a dentist's job rather than a GP's, and swelling that reaches the eye or neck needs A&E the same day¹¹. No score on any rubric, including this one, is a reason to keep buying tubes while one of those is true. The Journal's page on how sensitive toothpastes work covers what to expect once you have chosen, and the sibling guides in this section take the shelf apart product by product, work out the cost per day, and set out how to audit the evidence and the funders behind any claim you are shown.

Frequently asked questions

Is S3 the best toothpaste for sensitive teeth?

No page can honestly say that about any toothpaste, this one included. What S3 has is the only ingredient list among the 51 UK sensitivity toothpastes read in September 2026 that carries potassium nitrate, hydroxyapatite and a fluoride salt together, a declared level for each of its three actives, and 1450 ppm of fluoride kept in; all of that is checkable on the category review. What it does not have is a trial putting it against another product on this shelf, and neither does anything else here. On the evidence column of this page's own rubric it scores below every stannous fluoride paste on the shelf, because stannous fluoride is the one class the 2026 network rated high confidence².

Does a higher score mean it will work better for me?

No, and the rubric is built so you can see why. Five of its six criteria read a label and the sixth, worth a fifth of the total, reads a literature that has nothing to say about individual products, so a high score means a formula has been well specified and belongs to classes with published evidence, not that it will suit your teeth. Two pastes declaring the same active at the same level separated in an eight-week randomised controlled trial of 97 adults written by scientists employed by one of the manufacturers, and the difference lay in the rest of the formula⁴. Your own trial has one participant and no control arm, which is the same reason a friend's recommendation carries less than it feels like it does.

Why do the rankings on other sites disagree with each other?

Because none of them is reading the same thing, and there is no shared measurement underneath them to converge on. A ranking can be built from ingredient lists, from the reviewer's own testing, from what sells, from what pays a commission, or from the pooled evidence for ingredient classes, and those five produce five different orders from the same shelf. The only defence a reader has is to ask what a list scored and how much each part was worth, which is why the weights on this page are printed before the scores rather than after them.

How long should I give a toothpaste before switching?

Two to four weeks of brushing morning and night for most of these classes, and give a potassium-only paste longer. The 2026 network meta-analysis took its measurements at two weeks; the Cochrane review of potassium toothpastes took its at six to eight weeks and found the benefit on the instrument tests but not in what patients said about themselves²¹⁴. The 1997 consensus on trial design recommended eight weeks for most trials, with follow-up afterwards to see whether the change held¹⁸. Switching after a few days tells you nothing except that the first days are noisy.

Is the £1.00 own-label tube really scoring the same as a £5.00 one?

On this rubric, yes, and the two tubes make the point better than an argument would. The Boots own-label sensitive paste and one of the mid-priced branded potassium pastes both publish potassium nitrate and 1450 ppm of sodium fluoride in a water base, neither declares how much potassium is in the tube, and both therefore score 38. What the extra money buys on this shelf is a second mechanism, a declared level, a flavour, a brand or a whitening agent, and only the ingredient list tells you which of those you are paying for. The sibling guide in this section works the cost out per day instead of per tube, which is the arithmetic that matters over a year.