Start two weeks before? What the evidence says about potassium nitrate before bleaching.

Potassium nitrate before bleaching has a small pooled benefit and a split record: a 2021 systematic review and meta-analysis of 24 studies found a 12% lower risk of sensitivity when a potassium nitrate desensitiser was applied before bleaching, and its own authors called that reduction subtle and clinically questionable¹. Starting two weeks ahead is the schedule of one open-label randomised trial, whose funding is not stated and in which both groups also bleached with a gel that contained potassium nitrate; the other trials of a potassium nitrate toothpaste started it a week or 15 days ahead, and none set two start dates against each other²³⁴. S3 Sensitivity Science™ declares 5% potassium nitrate in a daily toothpaste; no trial has put S3 into a bleaching course, and in the one triple-blind randomised trial of a 5% potassium nitrate toothpaste brushed before in-office bleaching, 38 adults had no less sensitivity than on a placebo toothpaste³. Nobody has tested a potassium nitrate toothpaste started after bleaching pain begins, which is the question most people are actually asking.

What was checked36 peer-reviewed studies and reviews, the NHS and the Oral Health Foundation

What did the 2021 meta-analysis set out to test?

Whether a desensitiser containing potassium nitrate, put on the teeth before bleaching, lowers the risk and the intensity of the sensitivity that bleaching causes, compared with a placebo¹. The authors searched seven databases, among them PubMed, Scopus, Web of Science and the Cochrane Library, added grey literature, combined randomised trials with a random-effects model, judged each trial's risk of bias with the Cochrane tool and graded the certainty of the evidence with GRADE, according to the abstract of the systematic review¹. Twenty-four articles came through screening.

What "before bleaching" meant in practice is the question a toothpaste buyer most needs answered, and it is the one this page could not settle. The review's full text sits behind the publisher's paywall, so the vehicle, the concentration and the lead time used in each of the 24 included articles could not be checked here¹. Its reference list is public and gives a partial view. Most of the potassium nitrate trials it cites applied a gel in the dental chair or in the bleaching tray, it cites university theses and trial-registry entries beside journal papers, and it does not cite the 2005 toothpaste trial that the two-week advice traces back to, according to the Crossref record of the meta-analysis¹. Cited is not the same as included, so treat that as a clue rather than a finding.

How was it done, and what went into it?

It was a systematic review with a meta-analysis, and the table below holds only what can be read from its abstract and its public record¹. Where a cell says a detail was not available, the detail is in the paper, not missing from it.

What did it find?

A benefit that cleared statistical significance and not much else. Across the pooled trials, the meta-analysis found the risk of sensitivity 12% lower with potassium nitrate than with placebo (risk ratio 0.88, 95% CI 0.78 to 0.98), the average intensity 0.77 points lower on the ten-point visual analogue scale and 0.36 points lower on the numeric scale, and no difference in colour change¹.

Its authors did not dress the number up. The abstract's conclusion says the clinical significance of the reduction "is subtle and clinically questionable", and its line on clinical relevance repeats that the benefit is "subtle and maybe clinically questionable"¹. Applied to the in-office randomised trials described below, where nearly nine people in ten felt some sensitivity on the placebo side, a 12% lower risk from the meta-analysis would still leave roughly three people in four feeling it⁶¹.

An earlier meta-analysis produced a bigger number because it asked a looser question. The 2015 review by Wang and colleagues pooled potassium nitrate desensitisers with sodium fluoride ones, in several vehicles, from a search that closed in April 2014, and put the odds of sensitivity at 0.45 (95% CI 0.28 to 0.73) against placebo⁵. The P values printed beside those intervals in the meta-analysis abstract read like tests of heterogeneity rather than of the effect, so the odds ratio and its interval are the figures worth using⁵.

Three syntheses have appeared since 2021, and they do not agree. A 2022 systematic review of 13 randomised trials of desensitisers applied before in-office bleaching concluded that they lowered pain scores, mostly in the first 24 hours, and named 5% potassium nitrate with 2% sodium fluoride as the popular choice; its authors did not pool the trials, and their own table includes two triple-blind potassium nitrate trials whose authors reported no benefit⁹. A 2024 systematic review of desensitising toothpastes, five trials and 387 people, found them effective in two bleaching protocols and not in two others, again without pooling¹⁰. A 2026 umbrella review of 24 systematic reviews of in-office bleaching concluded, at moderate certainty, that potassium nitrate and sodium fluoride desensitisers reduce sensitivity without affecting the colour result; that line rests on four reviews its authors rated moderate or low to moderate for quality, and the 2021 meta-analysis read on this page is not one of them¹¹.

Where does "start two weeks before" come from?

From the schedule of a single 2005 paper, not from any comparison of start dates. In that open-label randomised trial, run across 14 dental offices in Florida and with its funding not stated, 202 people brushed twice a day with a potassium nitrate and fluoride toothpaste or a standard fluoride toothpaste for 14 days before tray bleaching and through the 14 days of it, and in the first week of bleaching 58% of the potassium nitrate group were free of sensitivity against 42% of the fluoride group². Over the fortnight of bleaching the potassium nitrate group also had more sensitivity-free days, and the shade of both groups improved by about the same amount.

The record adds two limits to that result. The trial was open-label, so people knew which toothpaste they were using while they filled in their sensitivity diaries. And the bleaching gel both groups used itself contained potassium nitrate, so the comparison was a potassium nitrate toothpaste plus a potassium nitrate gel against a fluoride toothpaste plus the same gel. Nothing in it tells you whether a fortnight beats a week, or a week beats none: it tried a single schedule.

Nor could this page find a head-to-head of start dates. A PubMed search run in September 2026, for potassium nitrate, toothpaste, dentifrice or gel, tooth bleaching or whitening, and phrases for one week, two weeks, seven days and fourteen days along with duration, prior to and pretreatment, returned twenty-seven records, and none of them set one lead time against another. A broader search for any desensitiser in bleaching studies, one hundred and twenty-nine records, turned up nothing further. The lead times in the trials that exist were each team's own choice: two weeks in the 2005 trial (funding not stated), 15 days in two in-office trials, and one week in two others, one of which used an arginine toothpaste rather than potassium nitrate^2124313.

Laboratory work does not fill the gap either, because it measures minutes. In a laboratory study of 90 extracted molars, potassium nitrate and hydrogen peroxide reaching the pulp chamber were sampled at intervals up to 45 minutes, and a separate layer of potassium nitrate under the peroxide delivered more of it than the same salt mixed into the gel¹⁴. An earlier laboratory study found potassium nitrate from a toothpaste in the pulp cavity of extracted molars within 30 minutes¹⁵. Extracted teeth have no blood supply and no outward flow of fluid, and neither laboratory study measured days, weeks or pain¹⁴¹⁵.

Which trials disagree, and why might they?

Grouped by vehicle, the results fall into recognisable clusters, so check the vehicle before using any of them to judge a toothpaste. The table lists every trial and synthesis of potassium nitrate and bleaching sensitivity that this page cites. Its last row places S3 for orientation only, as a daily toothpaste declaring 5% potassium nitrate, and it is not evidence of anything.

Toothpastes brushed before in-office bleaching give the sharpest split. The largest of those trials randomised 150 adults in Beijing and found lower average pain scores straight after a single session and on the next day with a potassium nitrate toothpaste started 15 days ahead, and no difference from the second day; its published summary gives averages, not the share of people affected, and does not mention blinding⁴. The best-blinded one found nothing: in a triple-blind randomised trial of 38 adults, a 5% potassium nitrate toothpaste brushed for a week beforehand, and again before each session, left the risk and the intensity of sensitivity no different from a placebo toothpaste³. A clinical trial of 45 adults that used a 15-day lead found a difference after the first of two sessions and none after the second¹². Of the three, only the triple-blind trial counted how many people had sensitivity at all, and it found no difference³.

Gels painted on in the chair before in-office bleaching lean towards no effect in the recent trials. In a triple-blind split-mouth randomised trial of 43 patients, a 10% potassium nitrate gel applied for 10 minutes before each session did not lower the risk of sensitivity, which the trial found above 87% on both sides, or its intensity⁶. A randomised trial of 75 patients found a potassium nitrate and fluoride gel before each of three sessions better than no desensitiser at one time point only, and its authors concluded that the gel was not effective²³.

Two more trials often counted on that side test something narrower. In one, every participant had a desensitising gel before bleaching, and what the trial compared was whether a second application afterwards helped: 90% had pain on the before-only side and 93% on the before-and-after side, in a triple-blind split-mouth randomised trial of 90 patients²². In the other, a split-mouth randomised trial of 60 people put 10% or 35% potassium nitrate gel on one side and a 5% gel on the other, found no difference between sides, and left no side of any mouth without potassium nitrate²⁴. Neither can say whether potassium nitrate beats nothing.

Gels held in a tray before home bleaching are where the positive results sit. In a single-blind randomised trial of 40 people recruited because they were at risk of sensitivity, whose funding is not stated, 41% of those using a 3% potassium nitrate and fluoride gel for 30 minutes before each session had at least one day of sensitivity, against 78% on a placebo gel²⁰. A triple-blind randomised trial of 32 people published in 2025 repeated that design with 16% carbamide peroxide and found 37.5% with sensitivity on the desensitising gel against 87.5% on placebo, with no difference in how intense it was²¹. Neither product is a toothpaste, and both kept the gel against the teeth for half an hour at a time, which brushing does not.

Potassium nitrate inside the bleaching gel is the clearest null. Pooling six trials of home gels, a 2020 meta-analysis found no difference in the risk of sensitivity (risk ratio 0.93, 95% CI 0.73 to 1.19), in its intensity or in the colour result¹⁶. A 2019 meta-analysis of any desensitiser mixed into the gel found no difference in risk either, at moderate certainty¹⁷. In a triple-blind within-person randomised trial of 60 adults, potassium nitrate and fluoride in a home gel made no difference²⁵, and in a double-blind split-mouth randomised trial of 60 patients, 5% potassium nitrate in an in-office gel left the risk the same and the intensity 1.32 points lower on a ten-point scale for the first 24 hours²⁶.

Why the disagreement? Four reasons are visible in the table, and none requires anyone to have done anything wrong. Contact time is one: a gel held against the teeth for half an hour is a different exposure from a toothpaste brushed and spat out. The strength of the bleach is another, because when almost everyone on the placebo side feels something, as in the in-office gel trials, there is little room for a difference in risk to show⁶. The outcome is a third: several of the positive results are lower average intensity, while the question of whether a person gets sensitivity at all is unchanged in most of the blinded in-office trials⁴²⁶³⁶. The fourth is design. The trial behind the fortnight was open-label, and in split-mouth trials the same person rates both sides of their own mouth.

Does a potassium nitrate toothpaste help once whitening has already made teeth hurt?

Nobody knows, because nobody has run that trial. A second PubMed search in September 2026, for potassium nitrate, toothpaste or dentifrice, post-bleaching or after or following bleaching, and sensitivity, returned three records, and none started a toothpaste after the pain began; the same search without the toothpaste terms returned seventy-five records and the same answer.

The closest designs sit either side of the question. In an uncontrolled clinical study of 30 people bleaching at home, the one study found that started a potassium nitrate product after sensitivity appeared, 11 of the 12 who put a 5% potassium nitrate and fluoride gel in their trays reported less sensitivity; there was no comparison group, and its funding is not stated⁷. After each home session with 22% carbamide peroxide, a potassium nitrate and fluoride toothpaste, analysed together with a strontium toothpaste, gave lower air-jet sensitivity over ten days than a toothpaste without actives, in a double-blind randomised trial of 60 adults¹⁸. Worn in a tray for four hours after each in-office session, a 5% potassium nitrate toothpaste gave lower scores than sucralose or sodium fluoride toothpastes in a double-blind clinical trial of 48 volunteers that excluded anyone with sensitive dentine¹⁹. Brushed three times a day once the course had begun, a potassium nitrate toothpaste alone did no better than a placebo toothpaste in a double-blind randomised trial of 108 volunteers, and scores were lower only when it was combined with a potassium nitrate gel painted on before each session⁸. All three controlled trials started the toothpaste with the course, not after the pain.

So, as far as a search can find, the toothpaste that shortens bleaching sensitivity once it has started has not been identified. What the trials do show is that the pain is usually brief: in a randomised trial that followed 120 adults for a week after in-office bleaching, scores were highest over the first two days and had almost reached zero on the seventh²⁷.

The guidance is about who to talk to, not what to buy. On the NHS page on teeth whitening, sensitivity to cold or sweet things is one of three side effects listed for treatment by a dental professional, and the advice is to raise any that bother you with the dentist²⁸. The Oral Health Foundation's whitening page calls the sensitivity temporary and usually short-term, and says the dental team can give advice or products to help²⁹. Neither page, as read in September 2026, recommends starting a potassium nitrate toothpaste beforehand. The Journal's page on sensitive teeth after whitening covers the fortnight after a course, and the page on using a sensitivity toothpaste alongside a course covers the routine.

Why does evidence from everyday sensitive teeth not carry over?

Because the two conditions are thought to start in different places, and potassium nitrate's evidence was gathered on the other one.

Its record in dentine hypersensitivity comes from weeks of brushing, scored with cold air and a probe on exposed dentine. A 2006 Cochrane review of six randomised trials found lower air-blast and tactile sensitivity at six to eight weeks, no significant effect on people's own ratings, and no clear evidence overall for potassium toothpastes³⁰. In a double-blind randomised trial of 36 adults, the first changes on cold air appeared from week two³¹. The usual account is that potassium reaches the nerve through open tubules and quietens it, and a 2000 review pointed out that this has never been confirmed in intact human teeth³². The Journal explains the everyday mechanism in how sensitive toothpastes work.

Bleaching pain is thought to begin deeper. On the leading explanation, peroxide crosses the enamel and dentine into the pulp, which is why the ache can arrive in a healthy tooth with nothing touching it; a 2010 hypothesis review set out that contrast with dentine sensitivity³³. In a laboratory study of extracted teeth, peroxide from both a weak varnish and a strong gel reached the pulp chamber³⁴, and a 2024 systematic review of in vivo studies found inflammation in the pulp of human teeth bleached with high-strength gels, without a measured link to what the people felt³⁵. The same hypothesis review predicted that nerve-damping treatments such as potassium salts should help if its idea was right³³. The bleaching trials in the table are, in effect, a test of that prediction, and so far they return a small pooled effect with blinded nulls beside it.

That is why "potassium nitrate takes two weeks to work" does not mean "start it two weeks before bleaching". The first describes how long brushing takes to change cold-air and touch scores on exposed dentine. The second is a question about peroxide in the pulp, and nobody has answered it with a trial. The pages on how potassium nitrate calms a sensitive tooth, on how long it takes to work and on the Cochrane review of potassium toothpastes are about dentine hypersensitivity, and none of them is about bleaching. The page on why whitening makes teeth sensitive takes the pulp explanation further.

Does the 5% potassium nitrate in S3 change any of this?

No. S3 declares 5% potassium nitrate, 10% nano-hydroxyapatite and 5% biomimetic hydroxyapatite, both hydroxyapatites as solution, and fluoride at 1450 ppm. The two hydroxyapatite percentages describe the ingredient as it goes into the tube rather than the mineral itself, whose content is lower, and S3 states both figures. In the formula, potassium is there for the nerve, nano-hydroxyapatite for the inside of open dentine tubules and biomimetic hydroxyapatite for the enamel surface, a division of work designed around everyday dentine hypersensitivity.

None of the trials on this page used S3, and the one that tested the design nearest to it, a triple-blind randomised trial of a 5% potassium nitrate toothpaste brushed before in-office bleaching, found no difference in risk or intensity³. The trial most often quoted for the opposite view gave its toothpaste group a potassium nitrate bleaching gel as well, did not blind anyone, and its funding is not stated². The positive gel trials kept 3% potassium nitrate against the teeth for half an hour in a tray, and one of them carries no funding statement²⁰²¹. The in-gel trials delivered potassium nitrate inside the peroxide itself²⁵²⁶. A toothpaste brushed morning and night is none of those things, and no trial has put one containing potassium nitrate, hydroxyapatite and fluoride into a bleaching course.

What does the evidence not show?

The gaps are specific, and several sit exactly where the everyday advice sounds most sure of itself.

• A lead time. No published trial has compared starting a potassium nitrate toothpaste one week, two weeks or no weeks before bleaching; the fortnight is one trial's protocol, a trial whose funding is not stated, and the others chose a week or 15 days²³⁴.
• Anything started after the pain. No controlled trial has tested a potassium nitrate toothpaste begun once bleaching sensitivity has appeared; the one study of that design used a gel in the tray, had no control group, and its funding is not stated⁷.
• A pooled answer for toothpaste. No meta-analysis has pooled potassium nitrate toothpastes used around bleaching: the 2024 systematic review of desensitising toothpastes described its trials without combining them, and the vehicle split inside the 2021 meta-analysis was not available to this page¹⁰¹.
• The combination. No trial has tested a toothpaste pairing potassium nitrate with hydroxyapatite and fluoride during bleaching; the nearest, a nano-calcium phosphate paste that also carried potassium nitrate and fluoride, made no difference in a randomised trial of 40 patients³⁶. The page on nano-hydroxyapatite and bleaching sensitivity reads the mineral trials.
• People who already have sensitive teeth. Several trials turned this reader away: sensitive dentine or gum recession excluded people from both tray-and-toothpaste trials by the same authors, and a history of dentine hypersensitivity excluded them from the 2025 split-mouth trial¹⁹⁸²⁴. The one trial that recruited people at risk tested a gel in the tray, not a toothpaste, and its funding is not stated²⁰. The hub page on how to whiten sensitive teeth counts how many bleaching trials excluded such people.
• The months afterwards. In the one trial here that followed people for six months, no group reported sensitivity at 24 weeks, so there was no later difference left to find⁸.

What can you ask a dentist instead?

Ask about the two things with steadier evidence behind them: how strong the gel is and how long it stays on. A 2020 meta-analysis of 25 randomised in-office trials found the risk of sensitivity lower with low or medium hydrogen peroxide concentrations than with high ones, at moderate certainty³⁷. A 2024 meta-analysis found fewer sensitivity events when a home gel was worn for less time each day, and the colour came out better with the full wearing time on most measures³⁸. Both trade some colour for comfort, which is a conversation for before the appointment.

Frequently asked questions

How long before whitening should I start a sensitivity toothpaste?

Nobody has tested that, so there is no evidence-based number. The two-week figure is the schedule of a 2005 open-label randomised trial whose funding is not stated, and that trial did not set it against any other start date². Other trials started a potassium nitrate toothpaste a week or 15 days ahead, with results running from no difference to lower pain on the first day³⁴.

Will a sensitivity toothpaste stop my whitening from working?

Not in the trials that checked. The 2021 meta-analysis found colour change unaffected by potassium nitrate desensitisers, and randomised trials of potassium nitrate toothpastes used before or during bleaching reported the same colour result as their comparison groups¹³¹⁸. That is evidence about colour, not comfort.

Is the potassium nitrate in S3 the same as in the bleaching trials?

Only in name: S3 carries 5% potassium nitrate in a toothpaste that is brushed and spat out, beside two hydroxyapatites and fluoride, and no bleaching trial has used it. The trials used a 3% gel held in a tray for half an hour, a 10% gel painted on in the chair for 10 minutes, potassium nitrate mixed into the peroxide gel, or single-active toothpastes brushed for a week to a fortnight beforehand, and the funding is not stated for two of them^2062632.

What should I use if my teeth already hurt after whitening?

Tell the dentist who is supervising the course first, because nobody has tested a toothpaste started after bleaching pain begins, and they can change the gel strength or the schedule. If the pain is severe, keeps returning in a single tooth, arrives out of nowhere or is still there weeks later, the Oral Health Foundation's page on sensitive teeth puts that in the see-a-dentist column³⁹. Most bleaching sensitivity fades on its own: in a randomised trial of 120 adults it had almost disappeared a week after in-office bleaching²⁷.

Does potassium nitrate in the bleaching gel itself help?

Not with whether you get sensitivity. A 2020 meta-analysis of home gels with potassium nitrate added and a 2019 meta-analysis of desensitiser-containing gels both found no difference in risk¹⁶¹⁷. In one randomised trial of 60 patients, 5% potassium nitrate inside an in-office gel lowered the intensity but not the risk²⁶.