Alencar 2019: nano-hydroxyapatite for dentine hypersensitivity, and the one stimulus where it made no difference.

The Alencar 2019 meta-analysis found that nano-hydroxyapatite relieved dentine hypersensitivity more than the treatments it was compared with on an air blast and on a touch probe, but made no difference on cold (SMD −0.17, p = 0.61), pooling six randomised trials that each ran four weeks¹. Its authors rated the pooled evidence high quality and asked for longer trials, and they did not explain the cold result in the abstract we could read¹. S3 Sensitivity Science™ contains nano-hydroxyapatite alongside potassium nitrate and fluoride, and this review did not test it.

What was checked20 peer-reviewed studies, product information as published by each brand, NHS guidance

What did Alencar and colleagues review, and how?

They ran a systematic review with meta-analysis of randomised trials in people with dentine hypersensitivity, comparing treatments that contain nano-hydroxyapatite with treatments that do not, or with a placebo or negative control¹. Seven databases were searched on 27 April 2018, the risk of bias was judged with the Cochrane tool, and the certainty of the evidence was graded with GRADE¹. The paper, by de Melo Alencar and colleagues, appeared in the Journal of Dentistry in 2019¹.

A note on the word in your search. When someone asks for hydroxyapatite toothpaste reviews, this page answers with systematic reviews of trials: star ratings and shoppers' comments are not used here, because they cannot separate a paste's effect from a placebo response or from the passing of time.

The full text sits behind the publisher's paywall; Europe PMC lists it as subscription only, and no accepted manuscript was reachable on 2026-09-13¹. Everything below about the review itself comes from its abstract and its Crossref reference list. The PubMed record carries no funding or conflict-of-interest statement, which means the funding is unknown rather than declared as none.

Which trials were pooled, and what did each measure?

We could not confirm the six from the review itself, because the trial list is in the paywalled full text¹. What we can check is the review's reference list, deposited with Crossref (saved 2026-09-13): it cites eight nano-hydroxyapatite clinical studies that already have cards in our library, and the six pooled trials are very probably among them¹. The table sets out what each card records; treat every row as a candidate, not a confirmed inclusion.

Three trials that other pages link with this review are absent from its reference list. Orsini 2010, a double-blind trial of 70 adults that compared a zinc-carbonate hydroxyapatite paste with a potassium nitrate and fluoride paste over eight weeks and found no difference on cold water¹², is not cited. Amaechi 2018, a 51-person double-blind trial of a nano-hydroxyapatite cream in trays tested with ice-cold and air stimuli¹³, is a 2018 paper that a search run on 27 April 2018 may simply have predated. Shetty 2010, an in-office hydroxyapatite trial in 45 patients¹⁴, is also not cited.

Where did nano-hydroxyapatite make a difference, and where did it not?

On the air blast and the touch probe, the pooled trials favoured nano-hydroxyapatite by a margin whose whole confidence interval sat on the helpful side of zero (air −1.09, tactile −0.93)¹. On cold the estimate was small, −0.17, and its interval ran from −0.81 to 0.48, crossing zero, with p = 0.61¹. In plain words: on cold, the review cannot tell nano-hydroxyapatite apart from what it was compared with.

A standardised mean difference puts trials that used different pain scales on one footing, so it describes the size of a gap in units of spread, not points on a scale anyone scored. That is why the numbers cannot be turned into "less pain" in millimetres.

Two limits run through every figure. The comparators were pooled together, so placebo, negative controls and active pastes such as arginine or calcium sodium phosphosilicate sit in one group, and the review cannot say how nano-hydroxyapatite compares with any single one of them¹. And every trial stopped at four weeks, which is also the limit of what the review can say about how long any benefit holds¹.

"High quality" belongs to the pooled overall result. The secondary analyses, the comparisons with other treatments or placebo and the at-home versus in-office split, were rated moderate¹. A high rating for the overall pool does not make the cold result, or any single analysis, strong.

What can be said about why cold was the exception, and what cannot?

Nobody has published why. The abstract reports that cold was the exception and does not give a reason¹, and the full text was out of reach. Two PubMed searches on the day of review came back empty on this point, one for hydroxyapatite with dentine hypersensitivity and cold, ice, thermal or "cold water", one for nano-hydroxyapatite reviews and meta-analyses published since 2019; neither turned up an explanation for a weaker hydroxyapatite result on cold. We therefore give no mechanism in the tooth.

What can be said sits in the trials, and it is about counting and method rather than biology.

How few trials could have fed the cold analysis

Of the eight candidate trials, only Gopinath 2015 records a cold-water test on its card⁶. The two Vano trials, Wang 2016 and Jena 2015 used an air stimulus, which is evaporative rather than thermal⁴⁵⁷¹⁰. The double-blind Anand 2017 trial used an electrical threshold⁸. The review's abstract reports evaporative and cold results separately, so if the air-stimulus trials sat under evaporative, the cold analysis may rest on very few trials, and a pooled estimate from few small trials has a wide interval almost by construction¹. That is an inference from the cards, not a statement from the review¹.

How differently "cold" was applied

Ice held against a tooth, cold water dripped onto it and a jet of room air all get called cold in this field. They are different stimuli, delivered for different times and scored on different scales, and pooling them adds noise to any estimate.

What the comparators were

In Gopinath 2015, a double-blind trial of 36 adults and the one candidate with cold water on its card, nano-hydroxyapatite was set against a calcium sodium phosphosilicate paste, an active treatment, and the two were judged comparable on all three stimuli⁶. A trial against an active paste will find a smaller gap than one against a placebo, and if cold-water data came mainly from active-comparator trials, that alone could pull the cold estimate towards zero⁶. Again, this is a reading of the design, not a finding.

The review group's next trial went in another direction. In 2020 the same group tested nano-hydroxyapatite with and without laser photobiomodulation in 32 people and 83 teeth, and adding the laser did not enhance the effect; nano-hydroxyapatite alone still showed less sensitivity at one month on the tactile and evaporative stimuli, and cold was not among the tests reported¹⁵.

Do later reviews see the same thing on cold?

They do not agree, and the disagreement is mostly about which "cold" is meant. Read side by side, the reviews since 2019 give four different pictures.

The 2020 network meta-analysis of 125 randomised trials held only 16 studies and 1,093 participants in its cold network, and in that network calcium sodium phosphosilicate was the only active node with moderate certainty³. There was no node for potassium combined with hydroxyapatite on cold at all³. Martins and colleagues' review (Journal of Dental Research, doi:10.1177/0022034520903036) carries no conflict-of-interest statement in its PubMed record, which lists research support from an unnamed non-US-government source³.

The 2026 network meta-analysis by Gormley and colleagues (93 randomised trials, Periodontology 2000) found nano-hydroxyapatite toothpaste ahead of a benchmark fluoride paste on the Schiff score at two weeks, mean difference −0.96, with moderate confidence, from two studies². The Schiff score is an air stimulus, so this is not a cold-water result, and the full text shows that no network was reported for thermal cold². The text we hold carries no conflict statement for the review itself; on an earlier network meta-analysis, the University of Bristol team of its co-author N. X. West declared advisory fees, lecture fees or research grants from GSK and Colgate-Palmolive, both makers of sensitivity toothpaste.

One industry trial saw a gap open late for one concentration. Amaechi and colleagues' eight-week double-blind trial, with 85 completers, was industry-funded by Sangi, the Japanese company that makes Apagard toothpastes: every paste reduced ice-cold sensitivity from baseline, but the plain paste with a 10% nano-hydroxyapatite inclusion relieved cold less than the calcium sodium phosphosilicate paste at weeks six and eight, while the 15% paste and the 10% paste with potassium nitrate matched it throughout¹⁶. That is one trial, and a sponsor's own.

A 2023 systematic review of 20 randomised trials of calcium phosphates, with no external funding, found pain fell by about 2.5 points on a visual analogue scale from baseline to four weeks on air, touch and cold water alike¹⁷. That fall is measured against each person's own starting point, not against a comparator: the authors state that comparing calcium phosphates with placebo or other products was not their research question¹⁷.

Two further meta-analyses pooled nano-hydroxyapatite without a stimulus split in what our cards record. Saxena 2026 gave three mainly nano-hydroxyapatite trials a pooled SMD of −0.87 against controls, from a review whose authors declare no competing interests¹⁸. Limeback 2023 pooled hydroxyapatite products against placebo in a review where two of the three authors are employees of Dr Kurt Wolff, a German maker of hydroxyapatite toothpastes¹⁹. Neither can be read for cold.

And every result above is short. A 2025 review of trials with at least six months of follow-up found the calcium-phosphate class did not reach a statistically significant long-term benefit²⁰.

So the contrast stands. The one review that split thermal cold found no difference¹; one network's "cold" is air²; one trial industry-funded by Sangi, the manufacturer of its pastes, saw a gap open late for one concentration¹⁶; a from-baseline review saw cold water fall like everything else¹⁷. None of that supports "hydroxyapatite helps with cold", and none of it supports "hydroxyapatite does not work for cold".

Which hydroxyapatite toothpastes are on the UK shelf, and what do they declare?

The table lists what each brand's own page or retailer listing declared when it was read on 2026-09-09, drawn from the rows behind our UK sensitivity toothpastes by ingredient review. This page compares ingredients and stated actions only, not clinical performance, based on what each brand states about its own formula. Formulas and listings may change, so the ingredient list on the tube you hold is the check.

None of the candidate trials' cards names a hydroxyapatite paste sold on this shelf, so the last column cannot link a trial result to any row^4567891011.

Six of the nine rows give no hydroxyapatite concentration at all, which is the point to keep: nobody can match a paste on this shelf to a trial dose from its pack. S3 publishes both kinds of figure: the 10% and 5% in its row are inclusion levels of the ingredient as supplied, and the active hydroxyapatite content is lower.

Does the cold exception apply to a toothpaste like S3?

Nobody can say, because no trial has tested it. S3 pairs 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite, both inclusion levels as supplied, with potassium nitrate and fluoride. No trial in this review, or since, has tested this toothpaste on a cold stimulus, or on any other. The nearest candidate is Wang 2016, a randomised trial in 28 people whose professionally applied 20% hydroxyapatite paste, and a 10% home paste, both also carried potassium nitrate and sodium fluoride; it measured an air stimulus, not cold⁷. The nearest cold result comes from Amaechi 2021, where a 10% nano-hydroxyapatite paste with potassium nitrate did not differ from calcium sodium phosphosilicate on ice-cold testing over eight weeks, in a double-blind trial industry-funded by Sangi, maker of the test pastes¹⁶.

S3 does not rely on hydroxyapatite alone; it adds a nerve active. Potassium's own cold evidence is no stronger: in the 2020 network meta-analysis its cold estimate could not be told apart from fluoride's³. That is the uncomfortable half of the argument, and it stays on the page. Combining two actives is a design choice with a mechanistic reason behind it; it is not evidence that the combination relieves cold, and we do not claim it does.

For more on nano-hydroxyapatite in general, see does nano-hydroxyapatite toothpaste work for sensitive teeth, and for the physics of a cold drink in a tubule, what happens inside a dentine tubule when you drink something cold.

When is a different toothpaste, or none, the better answer?

This page is written by S3 Science, the company behind the toothpaste named on it, and it has an interest in how you read what follows. There are still clear cases where our paste is not where to start.

If cold is the trigger that bothers you most, calcium sodium phosphosilicate had the firmest cold evidence in the 2020 network meta-analysis, as the only active node there rated moderate certainty³. It is not in our formula. Sensodyne Repair & Protect is one UK paste whose Boots listing names it, next to 1450 ppm fluoride.

If you want the result held with the most confidence in the newest network meta-analysis, stannous fluoride was the only active rated high confidence against benchmark fluoride on the two-week air score, from ten studies; nano-hydroxyapatite's estimate was larger but rested on two studies at moderate confidence². Stannous fluoride is not in our formula either. Sensodyne sells a stannous fluoride paste that declares 0.454% stannous fluoride.

If the pain from a cold drink lingers after the drink has gone, or throbs on its own, a toothpaste is the wrong tool. That pattern points to something a dentist needs to look at, such as decay, a cracked filling or an inflamed pulp; our guide to sharp and brief versus dull and lingering pain sets out the difference. The NHS advises booking a dental appointment when toothache has not settled after 2 days²¹.

And if a hydroxyapatite paste has not helped after the four weeks the trials ran, our page on still getting sensitivity with a hydroxyapatite toothpaste covers what to try next. The Sensitivity Journal's pieces on teeth sensitive to cold and on choosing a toothpaste for sensitive teeth give the wider picture, and for ice cream specifically see ice cream and iced drinks.

Frequently asked questions

Does hydroxyapatite toothpaste help with sensitivity to cold drinks?

It has not been shown to beat other treatments on cold. The 2019 meta-analysis found no difference on the cold stimulus (SMD −0.17, p = 0.61) while finding one on air and touch¹. Later reviews either measured air rather than cold water² or measured the fall from each person's own baseline¹⁷, which is a different question.

Are there reviews of hydroxyapatite toothpaste I can trust?

Yes, if you mean systematic reviews of randomised trials. The Alencar 2019 meta-analysis declares no industry link in its PubMed record, though its funding is not stated there¹; the 2020 and 2026 network meta-analyses place nano-hydroxyapatite among every other active³². Check who funded each one and how long its trials lasted: the Limeback 2023 review includes employees of a hydroxyapatite toothpaste maker¹⁹.

Was S3 included in the Alencar review?

No. The review pooled trials published before April 2018, and it made no statement about this or any other named retail paste¹. S3's formula combines potassium nitrate, two hydroxyapatites and fluoride; no candidate trial tested that formula, though the Wang 2016 randomised trial used hydroxyapatite pastes that also carried potassium nitrate and fluoride⁷.

Is nano-hydroxyapatite better than other sensitivity actives?

The evidence does not settle it. Alencar pooled all comparators together, so it cannot rank nano-hydroxyapatite against any one active¹. Where trials set it against a named active, such as calcium sodium phosphosilicate or arginine, they mostly found it comparable⁶⁸.

How long did the trials last?

Four weeks, in every trial Alencar pooled¹. A 2025 review of trials lasting six months or more did not find a significant long-term effect for the calcium-phosphate class²⁰, so four weeks is the honest horizon for any claim built on this review.