Tubule occlusion versus nerve desensitisation: what each approach can and cannot do.

Occlusion and nerve desensitisation are two different physical jobs — one narrows the open channels in exposed dentine, the other acts on the nerve at the far end of those channels — and the division is real chemistry, but it turns out to be a poor guide to how much any given tube will help you. What it does predict, reliably, is the other half — what each approach cannot do — and that half has stayed stable while the rankings have moved around it. Among 51 sensitivity toothpastes on sale in the UK whose published ingredient lists were read in September 2026, one carries potassium nitrate, hydroxyapatite and a fluoride salt together, and it is S3 Sensitivity Science™ — the full category review lists all 51.

What was checked36 peer-reviewed studies, product information as published by each brand, and NHS and Oral Health Foundation guidance

What is the difference between occluding a tubule and desensitising a nerve?

One puts something in the hole; the other works on what the hole leads to. Exposed dentine is threaded with tubules running from the surface down towards the pulp, and the standard explanation of why cold air on a bare root hurts is that a stimulus makes the fluid in those tubules move sharply, and the movement excites nerve endings at the inner end. The account is usually attributed to Brännström, whose 1966 review set it out¹. A 2025 review of the ion channels involved argues that the prevailing hypothesis now folds the fluid movement together with the odontoblast-transducer account, so the signal runs from the tubule to a receptor to a nerve rather than from tubule straight to nerve². That matters here for one reason only: the two approaches sit at opposite ends of the same chain, which is why they are not really rivals, whatever the shelf implies. The Sensitivity Journal's account of the two causes of sensitive teeth is the short version of that chain; this page is about whether the split it describes predicts anything once you weigh the trials.

An occluding active narrows or plugs the mouth of the tubule with material it deposits or precipitates there — hydroxyapatite, bioactive glass, an arginine and calcium carbonate complex, a stannous salt, an oxalate. A nerve-directed active is, in practice, a potassium salt: potassium nitrate, potassium citrate or potassium chloride, which raise the potassium concentration around the nerve ending and are proposed to make it harder to fire. The mechanism has been argued for fifty years and never confirmed in an intact human tooth, which the standard narrative review of the potassium literature says in the same breath as reporting that all 16 of the double-blind randomised trials it identified found significant reductions in sensitivity²⁴.

How can you tell, from the outside, which one a toothpaste is doing?

You mostly cannot, and neither can a single instrument. Three instruments are used, and they disagree with each other often enough that the choice of instrument partly decides the answer.

Electron microscopy shows you a surface. It is the source of most "seals the tubules" claims and it is genuinely good at what it does, but a picture of a plugged tubule is not a measurement of anything a person feels. Fluid-flow rigs measure flow across a dentine specimen under simulated pulpal pressure, which is closer to the mechanism the theory names. In a 2026 in vitro study on 60 bovine dentine specimens, four desensitising toothpastes cut measured flow by between 44% and 70%, comparable with a professionally applied sodium fluoride varnish — and one of the four did not beat distilled water, while distilled water itself cut flow by about 30% while it sat on the surface⁷. Pain measures — Schiff, tactile threshold, visual analogue scales — measure what the reader actually cares about and are the noisiest of the three. The page on how desensitising toothpastes are tested takes the instruments apart properly.

The instruments can disagree inside a single paper. A 2015 in vitro study of nano-hydroxyapatite, oxalate and arginine copolymer rinses found the hydroxyapatite rinse covered the dentine surface best on electron microscopy, while the oxalate and the arginine copolymer reduced fluid flow more⁶. Same specimens, same session, two measurements, two rankings. The same gap turns up between the laboratory and the clinic: an industry-funded 2026 study of a hydroxyapatite-citrate complex occluded 99.7% of bovine tubules in vitro at day seven, while its clinical arm improved the cold-air score by 21.39% against placebo at eight weeks²⁸. Both numbers are correct. Only one of them is about a person.

Which approach acts faster — and is that even true?

The common claim that occluders act sooner is a generalisation, and the best evidence for it is weaker than the claim. A 2019 systematic review of 74 randomised trials grouped every desensitising treatment by the follow-up window at which its reduction reached significance, and found that only in-office treatments — glutaraldehyde with HEMA, glass ionomer cements, laser — got there inside seven days¹⁰. Stannous fluoride and hydroxyapatite appear in the medium-term group, up to one month; potassium nitrate, arginine and hydroxyapatite in the long-term group¹⁰. On that reading, nothing you buy in a tube acts within a week, whichever category it belongs to.

Two later trials complicate it in both directions. In an industry-funded double-blind randomised trial of 120 adults, seven of whose eight authors are employees of the company that makes the stannous paste it tested, potassium nitrate, stannous fluoride and an oxalate paste all beat the fluoride control on the cold-air score at day three, and potassium nitrate was the one that missed on the tactile measure at that point¹¹. By week eight, in that same industry-funded randomised trial, the improvements against control were 57%, 47% and 44%, and the differences between the three actives were not significant¹¹. A randomised trial of 215 completers, funded by the maker of the test product and written with its employees, found a 5% calcium sodium phosphosilicate paste separating from a reference fluoride paste from day three onwards and widening to eight weeks¹². And in a four-week double-blind randomised trial of 30 adults, a bioactive glass paste beat 5% potassium nitrate at two weeks and four weeks, while potassium nitrate's own advantage over the control appeared at four³.

So: an occluder can show something at day three, a potassium paste tends to show it later, the gap between the actives has closed by week eight in the industry-funded trial that measured both ends¹¹, and the largest review to group treatments by follow-up window puts the honest floor for anything in a tube beyond one week¹⁰. The reason relief builds is covered on the page on why relief that builds over weeks is a feature. What no active in either group does is work on the first day. It is worth saying plainly that a large part of what people feel early on is not the active at all. In a split-mouth, subject-blind randomised study of 22 adults, a dressing placed over the sensitive tooth cut thermal pain by 95% and evaporative pain by 85% in a single application²⁹. In a 1997 double-blind randomised trial of 120 adults, strontium acetate, potassium nitrate and a plain fluoride toothpaste all improved over six weeks with no significant difference between the three³⁰. A six-arm randomised trial of 164 adults in 2025 found the same shape again: every active group improved on all three of its measures by week eight, and so did the plain fluoride control, on two of them²⁶.

Where do the two categories stop lining up with the results?

In three places, and they are the reason this page exists.

The pooled reviews do not agree with each other. Nine analyses of the same literature put the actives in different orders, with different certainty ratings, depending on which papers each one included and which comparator it used; two of the nine have manufacturer-affiliated authors and both say so^{131415161718192023}. The table below is the state of the evidence, not a verdict, and every cell is the review's own wording.

Two Cochrane reviews of potassium toothpastes, in 2001 and 2006, found a significant effect on air-blast and tactile measures and no significant effect on what patients themselves said, and concluded there was no clear evidence to support them²¹²². That is the same active the 2020 network rates at moderate certainty on the tactile measure¹⁹, and the same active the standard narrative review reports as significant in all 16 of its double-blind randomised trials²⁴. Nobody is lying. The literature is heterogeneous, the outcome measures are not interchangeable, and the pooling method decides a good deal of the answer. When the two approaches have been put head to head directly rather than through a network, the answer has been a shrug: in a four-week randomised trial of 39 adults, a hydroxyapatite toothpaste and a potassium nitrate toothpaste both relieved air-blast sensitivity and were comparable to one another, in a trial far too small to establish or exclude a difference²⁵.

The active everyone files under "nerve" may occlude. In the four-week double-blind randomised trial of 30 adults above, companion electron microscopy found no occlusion at all from 5% potassium nitrate³. In an in vitro electron-microscopy study of 90 dentine samples from 45 extracted premolars, a potassium nitrate toothpaste produced detectable occlusion at three, seven and fourteen days⁴. In a second in vitro study, 45 human dentine disks under a four-day acid challenge, both dipotassium oxalate and potassium nitrate occluded more than an untreated control, with oxalate faster and the difference gone by day four⁵. And in the 2015 rinse study, a potassium nitrate rinse was used as the negative control for occlusion, on the assumption that it does not occlude⁶. Three laboratory answers, two of them yes, one of them no by design, and one clinical answer of no. This is a contradiction, not a resolution, and it is taken apart at more length on the page asking whether potassium nitrate blocks tubules.

Occlusion and pain have not been shown to move together in the same mouths. One group in Bristol has tried twice with the replica technique, taking silicone impressions of the same teeth before and after four weeks. In the first examiner-blind randomised trial, 19 participants on either an occluding paste or a non-occluding fluoride one, with one co-author employed by a toothpaste manufacturer, occlusion scores did not correlate significantly with pain scores, occlusion fell with both pastes without reaching significance, and the thermal score at four weeks was significantly lower than baseline for the non-occluding paste⁸. In the second randomised trial, 20 participants, the correlation was significant and only the occluding paste produced a significant fall in pain⁹. Twenty is not a large number and neither is nineteen; the two results are the same group, the same method and the opposite finding, and the authors of the second call the technique a proof of concept that still needs refinement. Put honestly: nobody has yet demonstrated that how much of a tooth's dentine is plugged predicts how much it hurts. Worth saying too, because it is the sort of gap a page like this normally glides over: a search of the indexed literature for occlusion and pain measured in the same mouths, with an occluding paste set against a nerve-directed one, comes back empty. The two Bristol experiments, one of them manufacturer-affiliated, each put an occluder against a plain fluoride paste⁸⁹. The comparison at the centre of this page has never been run with both instruments pointed at the same teeth.

What can neither approach do?

Neither grows dentine back. The one paper that looked directly, whose authors include employees of the company developing the peptide it tested, compared two marketed desensitising toothpastes with that experimental peptide in an in vitro and canine model, and found new tertiary dentine in the peptide group alone; the two marketed pastes and the control formed none³⁵. "Repair" in a toothpaste means mineral deposited on a surface and into a channel, which is measurable, and not tissue regrown, which is not on offer. Neither approach touches the cause, either. The Oral Health Foundation's list of what makes dentine become exposed is brushing too hard, gum recession, acid from drinks or stomach, brushing straight after acidic food, gum disease, grinding and cracked teeth — and a toothpaste changes none of those³⁷. It is worth adding that the Foundation's own page carries an educational grant from Oral-B, disclosed at the foot of it³⁷ — exactly the sort of thing this page is asking you to notice about its own sources.

Neither is durable on its own. A 45-patient randomised trial of professionally applied hydroxyapatite showed complete tubule obliteration on electron microscopy and significantly less hypersensitivity than control from one day to four weeks — and by week eight the groups no longer differed²⁷. The 2025 meta-analysis restricted to trials of six months or longer found that calcium-phosphate-based agents did not reach statistical significance over that horizon, while glutaraldehyde and laser did²³. The Oral Health Foundation puts the practical version in one line: you need to keep using these toothpastes to maintain the effect³⁷, which is also the point the Journal makes about why a sensitive toothpaste can seem to stop working. The UK clinical guidance is blunter still, and it is this page's thesis in a dental journal's words: there does not currently appear to be one ideal desensitising agent that can be recommended³⁶.

And neither is a diagnosis. Sensitivity that is severe, that lasts more than a few weeks, that affects only one tooth or that arrives suddenly needs a dentist, because those patterns point at decay, a crack, gum problems or infection rather than at exposed dentine³⁷. The NHS asks you to see a dentist for toothache lasting more than two days, and to treat pain when you bite, red gums, a bad taste or a swollen cheek or jaw as reasons to go sooner³⁸.

What is actually in the multi-active tubes on the UK shelf?

Less than the front of the pack suggests, and much of it is unquantified. What follows is drawn from each brand's or retailer's own published ingredient list, read in September 2026, based on what each brand states about its own formula. This page compares ingredients and stated actions only, not clinical performance. Prices and formulations may change; always check the pack.

The striking thing in that table is the phrase "level not stated". Of the 48 competitor products whose ingredient lists were read for the category review, ten publish a concentration for their desensitising active: three stannous fluoride rows at 0.454%, one potassium nitrate at 5% w/w, one calcium sodium phosphosilicate at 5% w/w, one arginine at 8.00%, three hydroxyapatite rows, and one further potassium nitrate row on a listing whose UK availability is not established. Everyone else publishes the fluoride figure and nothing else. A reader who has correctly worked out that dose matters cannot compare doses, because most of the shelf does not publish them.

Two of these tubes carry an occluding and a nerve-directed active together. That is the category the search phrase "multiple active ingredients" is reaching for, and it is worth knowing what has and has not been tested about it. A UK examiner-blind randomised trial of a paste combining potassium nitrate with hydroxyapatite and aluminium lactate beat a potassium-nitrate-only control at every time point, and found a relative risk reduction in the cold-air score of 88.6% at 14 days³¹. An industry-funded double-blind randomised trial of 127 adults, every author employed or funded by the maker, found potassium nitrate with aluminium lactate beating potassium nitrate alone at four and eight weeks³². A manufacturer-funded trial of 118 completers found a copolymer-plus-potassium-nitrate paste beating 5% potassium nitrate alone³³. Each of those adds a second active to potassium nitrate and each shows a gain. Against them: a triple-blind split-mouth trial of a paste combining 8% arginine, calcium carbonate and potassium nitrate found the occlusion in vitro and no significant difference from its own placebo on any sensitivity measure at any point out to three months³⁴. And no published trial has taken a potassium nitrate and hydroxyapatite paste and tested it against the same paste with each active removed in turn — the design plainly exists, it has been run for other pairings, and it has not been run for this one. The argument for combining is mechanistic, and this page will not dress it as a clinical result. The page on single-active versus multi-active pastes follows that argument the whole way.

Where does S3 sit between the two approaches?

At both ends of the chain, by design, and with the same evidential limits as everything else in the table. Each of the three actives is aimed at a different point on the chain this page opened with: the potassium at the far end where the nerve sits, the nano particle at the mouth of the tubule it is small enough to enter, the biomimetic form at the exposed surface. The formula is water-based, which is what allows it to carry potassium nitrate alongside hydroxyapatite at all; anhydrous formulas built around bioactive glass cannot. The fluoride is sodium monofluorophosphate rather than sodium fluoride, chosen because it does not react with the calcium in hydroxyapatite inside the tube — a piece of chemistry you can check against the ingredient list on the back. The levels are the ones the published studies used: 5% potassium nitrate, and 10% nano-hydroxyapatite as a solution.

Two things belong in the same paragraph as the "only" sentence at the top. The first is that the 10% and the 5% are inclusion levels of the ingredient as supplied, not the weight of active hydroxyapatite in the tube, which is lower — and S3 publishes both figures. The second is Spotlight Sensitivity + Rebuilding Pro, on the Boots shelf at £16.00 for 100 ml when it was read on 9 September 2026: a water-based tube declaring a potassium salt, hydroxyapatite and 1450 ppm sodium fluoride. The entire distance between that formula and the sentence at the top of this page is the word nitrate, which is why the claim is written with "potassium nitrate" in it and never loosened to "a potassium salt" or "a nerve-calming ingredient" — the loosened versions would not be true. A claim that survives someone checking it is the only kind worth making.

Nor has the combination been tested as a finished product. The evidence for each active is ingredient-level evidence, and the pooled reviews above are the honest picture of what that is worth: contested, heterogeneous, and nowhere near as settled on either side of the split as the marketing implies. What can be said is what is in the tube, at what level, and why those two jobs are two jobs.

Frequently asked questions

Is a toothpaste with several actives better than one with a single active?

Nobody has run the experiment that would answer this for the pairing most people are asking about. Three trials have added a second active to potassium nitrate and compared the result with potassium nitrate alone, and all three found a gain; two of them declare funding or authorship from the maker of the product tested, and the third states no funding in its record³¹³²³³. A fourth trial of a paste combining arginine, calcium carbonate and potassium nitrate could not separate it from its own placebo in the mouth³⁴. What has not been done is the same comparison for a potassium nitrate and hydroxyapatite paste, and until it is, the case for combining rests on mechanism, not on a measured result.

Does potassium nitrate block dentine tubules as well?

Unresolved, and the sources disagree in a way that cannot be averaged. Two in vitro studies found occlusion from potassium nitrate, one of them a four-day acid-challenge model on 45 human dentine disks⁴⁵. A double-blind randomised trial of 30 adults with companion electron microscopy found none³. Because the one answer that came from a mouth is the negative one, this page treats potassium nitrate as a nerve-directed active and flags the laboratory contradiction instead of resolving it.

What is in S3, and which of the two approaches does it use?

Both, which is the short answer to why it is on this page at all. The tube carries the best-established nerve active alongside two hydroxyapatites and 1450 ppm fluoride: 5% potassium nitrate, 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite as solution. The division of labour is the one this page has been describing — the potassium at the nerve, the nano particle sized for the inside of the tubule, the biomimetic form for the surface. It suits people who would sooner repair the surface than only numb the nerve, and people who want hydroxyapatite without giving up fluoride.

Does a paste with no fluoride still count as a sensitivity toothpaste?

For sensitivity, yes; for the rest of what a toothpaste is for, you are giving something up. Two of the hydroxyapatite tubes in the table declare no fluoride at all. Hydroxyapatite has its own body of sensitivity evidence, including a meta-analysis of 44 clinical trials whose authors include scientists employed by Dr Wolff, a maker of hydroxyapatite toothpastes²⁰; and the Oral Health Foundation's advice names potassium citrate, potassium nitrate and stannous fluoride as the actives to look for³⁷. Plain sodium fluoride is not on that list because it is not a desensitising active — it is the decay protection, and dropping it is a separate decision from the sensitivity one.

How long before I know which approach is working for me?

Longer than most people give it. The systematic review that grouped 74 randomised trials by follow-up window found nothing outside a dental surgery reaching significance inside seven days¹⁰. The two industry-funded randomised trials that measure at day three and again at week eight both show the effect growing across that span rather than fading¹¹¹². Two to four weeks of twice-daily use is a fair test, and a fair test means not switching in week two. The page on how long to give a new sensitivity toothpaste sets out how to judge it, and the two-week estimates in the table above are the reason the answer is not "a few days".