Acidic food and drink triggers: citrus, wine, fizzy drinks and why relief from cold does not last.

The toothpastes with a measured result against cold are the ones built round stannous fluoride, arginine, nano-hydroxyapatite or potassium, and in the 2026 network meta-analysis that pooled 93 randomised trials only stannous fluoride's cold-air figure was rated high confidence¹. Whether that relief lasts is a second question, and it is as much a question about your diet as about the tube: tubule occlusion is a deposit laid on and inside worn dentine, and dietary acid keeps taking it off again. S3 Sensitivity Science™ is built for both halves of that problem, with potassium to calm the nerve, nano-hydroxyapatite to work inside the tubule and biomimetic hydroxyapatite to work on the surface, each doing a job the others cannot.

What was checked29 peer-reviewed studies, S3 consumer trial (ADSL, 2026), the Oral Health Foundation and the NHS

What is the best toothpaste for lasting relief from cold sensitivity?

On the pooled evidence, the shortlist for cold is short and it is not ranked the way a shelf is. The largest synthesis in the field, a 2026 systematic review and network meta-analysis of 93 randomised trials and 9,548 participants, compared dentifrice classes against a benchmark fluoride paste on the cold-air score at two weeks: stannous fluoride came in at a mean difference of −0.85 (95% confidence interval −1.08 to −0.62, ten studies, high confidence), nano-hydroxyapatite at −0.96 (−1.40 to −0.52, two studies, moderate confidence), arginine at −0.78 (low confidence), potassium with or without fluoride at −0.42 (−0.77 to −0.07, one study, low confidence) and bioglass at −0.36 (low confidence)¹. Its authors recommend stannous fluoride and arginine as first-line self-care, and they are explicit that the choice should follow preference, tolerability and availability rather than an expectation of superior efficacy¹. S3 contains neither of those two actives. That is the honest starting point of this page, and it is the reason the rest of it is about durability rather than about rankings.

The word doing the work in the question is lasting, and no toothpaste answers it on its own. Two things end relief. The first is stopping, and that has been measured. The second is the diet, and it has barely been looked at. One caution belongs here before you read on: nano-hydroxyapatite's own 2019 meta-analysis of six four-week randomised trials found it ahead of its comparators on evaporative and tactile stimuli and found no difference on cold stimuli, which is a caution that belongs on any page that promises a cold result⁶. And the Oral Health Foundation puts the durability point plainly in its own guidance: a sensitive toothpaste works by blocking the channels in dentine, and you have to keep using it to maintain the effect⁷.

Why does acid set teeth off, and why does it undo the relief?

Because acid does two separate things to dentine, and the second one takes the deposit away. Dentine is threaded with fluid-filled channels running towards the nerve, and their outer ends are normally partly blocked by plugs of mineral and debris. In a laboratory study on human dentine discs, five minutes each of a cola drink, orange juice, white wine, vinegar and a mucolytic syrup all significantly raised the flow of fluid through the dentine, and the authors attribute it to the acid dissolving those plugs and the peritubular dentine around them⁸. Open the channel and every cold mouthful has a shorter route to the nerve. That is the trigger.

The durability half is the one that rarely gets written down. A desensitising paste works by putting something back into and over that channel, and something you put on a tooth can be taken off a tooth. In a laboratory study on human dentine discs, two brushing cycles left one paste with 85.7% of tubules occluded and another with 88.1% — statistically indistinguishable — and then ten minutes in agitated orange juice left the first with 20.3% and the second with 79.1%⁹. Two pastes that were equally good at occluding were not remotely equally good at staying occluded, and the authors conclude that resisting the acid, rather than the amount of material that goes in, may be what matters.

Where the deposit sits matters as much as what it is made of. In a laboratory study funded by the manufacturer of the bioglass-based desensitiser it tested, human dentine discs went through three ten-minute exposures to lactic acid and two standardised brushings a day for twelve days under a simulated pulp pressure, and the homogeneous layer the product had laid over the surface was gone within the first forty-eight hours of brushing¹⁰. The plugs inside the tubules were still there on day twelve, after thirty-six acid exposures and twenty-four brushing cycles, with silicon from the product detectable in the plugs and almost none in the dentine around them, in that same manufacturer-funded laboratory study¹⁰. A surface coat and an in-tubule plug are not one thing with one lifespan.

And the picture does not run one way. An in vitro study of 120 dentine discs found that a bioactive-glass toothpaste occluded more extensively than an arginine paste or a control, and that its occlusion withstood an acidic soft-drink challenge for comparatively longer, while the extent of occlusion in every group, that one included, went down as the exposure lengthened¹¹. The best-designed test on this question put the same comparison into real mouths. In a randomised, split-mouth in situ study whose only recorded funder is a National Institute for Health Research clinical lectureship, and which carries no funding statement from any toothpaste maker, 28 adults wore human dentine samples in appliances and brushed twice a day, and two pastes that had both beaten water and a plain fluoride paste on occlusion before any acid came near them then separated once two one-minute grapefruit-juice challenges a day were added¹². The 8% strontium acetate paste held its occlusion and scored better than every other treatment; the authors describe the 8% arginine paste as more susceptible to the acid challenge, in that same in situ study with no funding statement deposited¹².

One more finding is worth having, because it is against the grain of how these pastes are sold. In a laboratory test in which human dentine discs were etched with citric acid and then brushed twice daily for two weeks, a hydroxyapatite toothpaste and a stannous fluoride toothpaste both occluded tubules significantly better than water, and none of the three pastes tested made the dentine any harder¹³. Occluding a channel and rebuilding a tissue are different claims, and only the first one has this evidence behind it. The Journal's page on tooth enamel erosion covers the wear itself and what causes it; this page is about the deposit that acid keeps removing, and about the durability question that follows.

Which drinks and foods count, and does a pH league table help?

The list is longer than fizzy drinks and it is not ranked here, for a reason given below. The Oral Health Foundation names fizzy drinks including diet and sugar-free versions, energy and sports drinks, fruit juices, smoothies, hot water with lemon, wine and cider among acidic drinks, and citrus fruit, tomatoes and tomato sauces, vinegar and pickled foods, berries, pineapple and sour sweets among acidic foods¹⁴.

Now the reason for not ranking them. Two properties of a drink pull in different directions and people confuse them. In a study that measured salivary pH in twenty volunteers and enamel and root dentine hardness in situ in twelve, solutions at pH 2.5 drove salivary pH lower whatever their titratable acidity, while solutions with higher titratable acidity took longer to let the mouth return to neutral whatever their pH; adding citric acid raised enamel hardness loss by two-and-a-half to three times and dentine hardness loss by at least five times³. A pH number on its own tells you how hard the first hit is and nothing about how long the mouth stays acidic afterwards. Temperature is a third variable and it behaves unpredictably: in an in vitro artificial-mouth study on bovine enamel, two fruit teas and one soft drink caused significantly more enamel loss when served warm, while a cola, an orange juice and a citric-acid control did not change with temperature at all¹⁵.

What has actually been measured in people is how the drink is taken. In a cross-sectional clinical study of 600 dental-clinic patients, self-reported sensitivity was associated with spending ten minutes or more eating fruit at a sitting (odds ratio 2.72, 95% confidence interval 1.32 to 5.61) and with sipping, swishing or holding drinks in the mouth (2.33, 1.40 to 3.88), and it was not associated with how often acid was consumed or how often the teeth were brushed; the authors conclude that contact time may matter more than frequency¹⁶. That is the most useful sentence in this literature for anybody who does not want to give things up. For scale, a cross-sectional clinical study of 3,551 adults in seven European countries, funded by Haleon, recorded some erosive tooth wear in 97.6% of participants and found sensitivity significantly associated with it¹⁷. A corrigendum to that paper has been published; its content could not be read, so the figures here stand as first published.

Self-report fills in one more row that no trial has: in a cross-sectional survey of dental-clinic patients with sensitive teeth in Kerala, India, the things most often reported were packaged food (70%), pickles (62%), soft drinks (58%) and sweets (56%)¹⁸. That is what people said they ate, in one region, with nobody to compare them against, and it is on this page as a shape rather than as a finding.

Is there really a recovery window, and should you wait before brushing?

There is a window, and the advice built on it is the weakest-evidenced thing on this page. Take the window first. The Oral Health Foundation describes enamel softening after every acidic mouthful, losing minerals, and hardening again from saliva, and advises waiting at least an hour before brushing¹⁴. In people, the one measurement that touches it comes from a cross-sectional clinical study of 350 adults aged nineteen to thirty-four in south-east England, in which the severity of air-stimulus sensitivity on biting surfaces was significantly associated with the time since the person's last acidic food or drink, and the authors read that as clinical support for sensitivity being episodic and driven by active erosive wear². The same clinical study also reports a far higher share of sensitivity among those who had drunk something acidic within sixty minutes of the appointment, 87.2%, than among those who had not, roughly one in eight; the two counts behind those two figures do not reconcile with the study's own prevalence total, and the full text is behind a subscription, so this page states the association and leaves the figures where they were published².

Now the rule. The highest level of evidence on delayed brushing is a 2020 systematic review and meta-analysis of twelve laboratory and in situ studies, and it found no significant difference in the erosive wear of human enamel between delayed and immediate brushing, a significant benefit of waiting in bovine enamel, no difference in bovine dentine, and no studies at all on human dentine⁴. Its clinical message is stated in the paper: delaying toothbrushing on its own, after erosive food or drink, cannot prevent erosive enamel wear⁴. The one positive result in it points somewhere else entirely — a subgroup analysis found that brushing with a fluoridated toothpaste significantly reduced the wear of human enamel after erosion and abrasion⁴.

Where the disagreement comes from is visible in the individual papers. In an in situ experiment in which five volunteers wore human enamel samples through five-day cycles, enamel loss was 45.2 micrometres with erosion alone, 79.3 when brushed with a fluoride-free paste straight after the acid, and 81.7 when brushed two hours later — no protection from the wait — while brushing with a fluoride paste brought the loss to 51.5 micrometres, and a fluoride paste with a gel and a rinse to 41.2⁵. A laboratory study on human enamel found that up to four hours in human saliva neither hardened the softened surface nor reduced the wear from subsequent brushing, and its authors write that the recommendation to postpone brushing after an erosive attack should be reconsidered¹⁹. Against them, an in situ study in which ten volunteers wore bovine enamel blocks did find that waiting reduced abrasive loss compared with brushing immediately — and found no difference between waiting thirty minutes and waiting sixty²⁰. That is exactly the bovine-enamel result the meta-analysis isolated, which is why the contradiction is printed here rather than averaged away.

So be fair to both sides. Waiting an hour costs nothing, has no plausible downside, and is a sensible precaution built on a mechanism nobody disputes. What is not supported is the confidence with which the interval is repeated: nothing establishes an hour, the work that does find a benefit finds no advantage of an hour over half an hour, and the human-enamel evidence does not separate waiting from not waiting at all⁴²⁰. If you keep one habit from this section, keep the fluoride toothpaste rather than the stopwatch⁵⁴. Swapping a fluoride paste for a fluoride-free one because of the acid in your diet is the change this literature argues against.

Which ingredients have evidence, and what happens when you stop?

Two clocks and one gap. The clocks are how long an active takes to work and what happens when you stop; the gap is that neither has been measured in a mouth whose diet keeps re-acidifying the surface.

On the ingredients, three results carry most of the weight. A four-week double-blind randomised trial of 105 adults found a fluoride-free 15% nano-hydroxyapatite toothpaste reduced cold-air and tactile sensitivity more than a fluoride paste and a placebo at both two and four weeks²¹. The 2006 Cochrane review of six randomised trials put potassium nitrate's effect on air-blast sensitivity at a standardised mean difference of −1.25 at six to eight weeks, recorded no significant effect on the patients' own overall assessment, and concluded that there is no clear evidence to support potassium toothpastes — a review whose own conclusion is more cautious than the way it is usually quoted²². And a Procter & Gamble archive meta-analysis of fourteen randomised trials in 1,287 participants, with the manufacturer's own employees among the authors, put stannous fluoride at a 57% cold-air and 142% tactile benefit over plain fluoride pastes and reported an 83% benefit against enamel erosion in a separate six-trial in-situ arm²³. That erosion figure is the plainest thing on this page that does not favour S3: stannous fluoride is the active with the erosion evidence, and S3 does not contain it.

On durability under acid, everything that exists is laboratory or in situ. A five-phase in situ crossover study in which fifteen volunteers wore human dentine samples through an erosion and abrasion regime found no significant difference in surface loss between any of the desensitising pastes and a fluoride-free control, while three of them did significantly reduce the number of open tubules²⁴. Occluding a tubule and protecting the surface are different jobs, and on that evidence the pastes did the first and not the second. A laboratory study on dentine from 42 human third molars found that all four desensitising materials tested cut dentine permeability on the day they were applied, and that after seven days of citric acid six times a day plus brushing three times a day only the two professionally applied varnishes still held their effect²⁵. The nearest thing to good news is a twenty-day in situ randomised study, part-authored by employees of the company that makes the paste, in which the mineral laid down by a 5% calcium sodium phosphosilicate dentifrice resisted a repeated intra-oral dietary acid challenge better than that laid down by a fluoride-only paste, measured as surface hardness at day twenty²⁶.

On stopping, the numbers are better than most people expect. In an industry-funded double-blind randomised trial of 120 adults, run by a company that makes a stannous fluoride paste, everyone moved onto a non-desensitising control paste for three weeks after eight weeks of use; the potassium nitrate group's cold-air advantage over the control went from 44.3% at week eight to 35.8% at week eleven, and its tactile advantage from 104% to 72.5%²⁷. A twelve-week double-blind randomised trial of a hydroxyapatite-and-potassium-citrate paste, funded by the manufacturer with four of its six authors employed there, saw the air-blast advantage still standing at 43.34% four weeks after the product was stopped while the tactile advantage fell from 100.85% to 14.56%, which the authors put down to a superficial mineral layer not being replenished²⁸. Read together with the laboratory work above, that is the same story twice: what is inside the tubule persists and what is on the surface does not.

Two results keep the section honest. In a manufacturer-funded examiner-blind randomised trial that followed 76 adults for 24 weeks, a calcium sodium phosphosilicate paste produced small but significant improvements in the clinician's cold-air score at weeks eight, sixteen and twenty-four — adjusted changes of about half a point — with no significant change in tactile threshold and little or no change in what the participants themselves said about their sensitivity²⁹. And the most recent long-term synthesis, a 2025 systematic review and meta-analysis of 22 randomised trials followed for at least six months, found the largest effects for glutaraldehyde and low-level laser therapy and reported that calcium-phosphate-based agents did not reach statistical significance³⁰. Hydroxyapatite is a calcium phosphate. That review is the most unfavourable piece of evidence this project holds on long-term relief from the occluding side of the formula, and it belongs on a page whose subject is lasting relief.

Where does S3 sit if acid is part of the picture?

On the mechanism, in the place the two problems meet. Hydroxyapatite occludes and potassium desensitises, neither does the other's job, and a formula built for sensitivity has to carry both. S3's declared levels are 5% potassium nitrate with 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite, the last two being inclusion levels of the ingredient as supplied rather than active content; S3 publishes the inclusion level and the lower active figure together.

On acid specifically, the one approved statement is a laboratory one: in laboratory testing, biomimetic hydroxyapatite released calcium when the surrounding conditions turned acidic, at pH 5, and stayed stable at neutral pH. In laboratory testing, the same ingredient formed a mineral layer of about 9.7 µm after a single brushing, thicker than tricalcium phosphate at 7.2 µm or amorphous calcium phosphate at 8.2 µm. Those are measurements on an ingredient in a laboratory, and they are not a statement that anyone is protected from what they drink. Set the laboratory work above beside them: a mineral layer sitting on the surface is exactly the part that went first when a deposit was brushed under repeated acid in the manufacturer-funded laboratory study above¹⁰.

The fluoride is the part of the formula this section is most confident about, and it is confident for a reason that has nothing to do with sensitivity. S3 carries full adult-strength fluoride at 1450 ppm as sodium monofluorophosphate, a salt chosen because it does not react with the calcium in hydroxyapatite inside the tube. In an acidic diet, fluoride matters more rather than less: it was the fluoride, not the waiting, that reduced enamel loss when the two were tested in the same in situ experiment, and the meta-analysis on delayed brushing found its only significant protective effect in the fluoridated-toothpaste subgroup⁵⁴. That evidence is about tooth substance and decay, not about pain, and this page does not stretch it into a sensitivity result.

Three things this page will not claim, and they earn their place here. No trial has tested whether any desensitising toothpaste keeps working in an acidic diet: a PubMed search for a clinical trial of a desensitising toothpaste with an acid stimulus, run in September 2026, returned a single record, an in situ randomised study whose authors include employees of the maker of the paste it tested, and it measures hardness and occlusion on dentine specimens worn in an appliance rather than symptoms in a tooth²⁶. The triple-active combination has no acid-challenge work behind it, which is why its row in the table above reads "not tested" instead of carrying a figure. And the closest published study of a toothpaste combining nano-hydroxyapatite, potassium nitrate and monofluorophosphate is a single-arm, uncontrolled clinical study of self-reported scores with no comparator and no sample size in its abstract, so nobody has yet run a controlled trial of that combination at all³¹.

What the consumer trial can add is what people said, with its method attached. Based on an independent third-party consumer trial of 51 adults with sensitive teeth over eight weeks, run by ADSL in Devon to Good Clinical Research Practice, 88% said their teeth felt protected from sensitivity after four weeks. That is what people said in an unpublished consumer trial, run by ADSL in Devon, that had no control group, and it sits in a different register from the pooled figures earlier on this page. The Journal's page on how sensitive toothpastes work covers the mechanisms in more detail, and the resources page on why relief builds over weeks takes the timing further.

When is acid-related tooth pain something else?

You can tell by how the pain behaves rather than by how much it hurts. Acid-triggered dentine sensitivity is sharp and short and tied to its trigger: the mouthful arrives, the tooth complains, and it is over. Four patterns on the Oral Health Foundation's list belong to a dentist instead — severe pain, sensitivity that has lasted more than a few weeks, a single tooth rather than several, and a start that was sudden — because each of them can mean decay, a crack, a gum problem or an infection⁷. The NHS puts its own line at toothache lasting more than two days, or sooner if painkillers do not touch it, if it comes with a high temperature, pain on biting, red gums or a bad taste, or if the cheek or jaw is swollen — and says to book a dentist rather than a GP, because a GP cannot treat it³². Swelling that reaches the eye or the neck, or that makes breathing, swallowing or speaking difficult, is an accident and emergency matter rather than a dental appointment³².

There is one more question this page has to hand over. If the acid is coming from inside — reflux, or repeated vomiting, both named as erosion causes in UK guidance, and heartburn was among the factors associated with sensitivity in the industry-funded European clinical study above — then no change to the diet and no toothpaste addresses the cause¹⁴¹⁷. Unexplained erosion that keeps progressing is a medical question as much as a dental one, and it is worth saying to a doctor as well as to a dentist. The resources page on being sensitive to hot and cold at the same time deals with the pattern that suggests something other than exposed dentine.

Frequently asked questions

Does S3 give lasting relief from cold sensitivity?

It is designed to be used every day rather than as a course, and the honest answer about durability comes from the trials of its ingredient classes rather than from a trial of S3. Those trials say two things. Relief does not vanish when you stop: in an industry-funded double-blind randomised trial, three weeks after a potassium nitrate group came off its paste the cold-air advantage over a control had fallen from 44.3% to 35.8% rather than disappearing²⁷. And relief measured over months is modest and does not always show up in what people say about themselves: in a manufacturer-funded 24-week randomised trial of an occluding paste the clinician's cold-air score improved significantly while the participants' own sense of their sensitivity barely moved²⁹. S3 is judged at two to four weeks of twice-daily use and then kept going.

Should I really wait an hour after orange juice before brushing?

Wait if you can, but do not treat the hour as a number the evidence produced. The Oral Health Foundation advises at least an hour, and the mechanism behind it is real¹⁴. The trials are another matter: the systematic review on delayed brushing found no significant difference for human enamel between waiting and brushing immediately, the one in situ study that does find a benefit found no advantage of an hour over half an hour, and the one experiment that tested waiting and fluoride side by side found the protection came from the fluoride toothpaste⁴²⁰⁵. Brushing before the acidic meal rather than after it is the version of the advice with the least to argue about.

Is sparkling water bad for sensitive teeth?

Much less than a cola, and it depends on the water. Plain carbonated water is mildly acidic, and fizzy drinks in general are on the Oral Health Foundation's list of acid sources¹⁴. What decides how much a drink actually does is not the fizz but the pH together with the titratable acidity, and plain sparkling water is low on the second: it has no citric acid to keep the mouth acidic after the mouthful, which is the property that multiplied dentine hardness loss at least fivefold when it was added to a test solution³. Flavoured sparkling waters usually do contain citric acid, so read the label. Either way, drinking it with a meal and finishing it rather than sipping it across the afternoon is the change with a measurement behind it¹⁶.

Why did my sensitivity toothpaste stop working?

The likeliest answers, in order. You stopped using it, or started using it less often, and the Oral Health Foundation's own line is that the effect is only maintained while you keep using it⁷. Or your diet is stripping the deposit as fast as the paste lays it down: ten minutes in orange juice removed most of one paste's tubule occlusion in a laboratory test, and a week of acid plus brushing undid the occlusion of three of four materials in another⁹²⁵. Or the tooth is not doing what a sensitive tooth does — one tooth, sudden onset, pain that lingers after the trigger goes — in which case the toothpaste was never the answer and a dentist is⁷.

Does rinsing with water after an acidic drink help?

There is no trial that answers it, and the guidance points at other habits instead. What the Oral Health Foundation recommends is keeping acidic food and drink to mealtimes, swallowing drinks rather than holding them in the mouth, using a straw, finishing a meal with milk or cheese, and chewing sugar-free gum to get saliva flowing¹⁴. Every one of those shortens the contact between the acid and the tooth, which is the one thing that has been measured in people: the habits associated with reported sensitivity were long contact ones, sipping and swishing and picking at fruit, not the number of acidic things consumed¹⁶. A mouthful of water plainly shortens that contact, so it is a reasonable thing to do; it is on this page as sensible, not as evidence.