12  Fluoride mouth rinse

A small chapter with one large problem in it, and the problem is a kind of mismatch this book has not yet met head-on: evidence gathered in one setting, applied to a completely different one.

The advice

Use a fluoride mouth rinse daily (0.05% NaF; 230ppm fluoride) at a different time to brushing

Strength of recommendation: Conditional

— Delivering Better Oral Health, chapter 2, tables 1e and 1g (1)

Conditional, and aimed only at children from 8 years and adults giving concern because of dental caries risk. Both restrictions are appropriate and the guideline deserves credit for them.

Note the “at a different time to brushing.” That is a sensible piece of reasoning, and it comes from the same place as Chapter 7: do not wash away the toothpaste you just applied. It is also, like most of that chapter, untested.

What the guideline says its evidence is

Recommendation based on moderate certainty evidence from supervised school use in children and adolescents.

— Delivering Better Oral Health, chapter 13 (1)

The citation is the Cochrane review of fluoride mouthrinses (2).

I want to point at that sentence and say: this is exactly the right way to write an evidence statement. It gives the certainty and the setting the evidence came from, in one line, which lets a reader see the gap without going anywhere else. Most of the statements in chapter 13 do not do this. This one does.

The gap it discloses

First, what the evidence shows, since an earlier draft of this chapter argued about its applicability without ever telling you how big the effect was. Marinho and colleagues pooled 37 trials in 15,813 children and adolescents. For permanent tooth surfaces, from 35 trials in 15,305 participants, the prevented fraction was 27% (95% CI 23% to 30%), at moderate certainty. That is a substantial effect, and it is the number the recommendation rests on.

Now the setting. The evidence is from supervised school programs: a teacher or dental worker handing out rinse to a class, watching them use it, on a schedule, every school day.

The recommendation is for daily unsupervised home use by an individual.

Those differ in almost every way that determines whether a preventive intervention works:

Adherence. A supervised school program has near-perfect adherence by construction. Daily home use of a mouth rinse, by a teenager, indefinitely, does not. The effect estimate from the trials embeds a level of compliance that home use will not reach.

Population. School programs are typically population-wide. The recommendation targets individuals already identified as at higher risk, who have more to gain in absolute terms but who may also be the group least likely to sustain a daily routine.

Timing. The school program is at school; the recommendation specifies “at a different time to brushing”, which is a constraint the trials did not impose and which makes the home routine harder to fit in.

None of this means rinses do not work. Fluoride delivered to teeth reduces decay; that is the most robust finding in this book. It means the size of the benefit observed under supervision is plausibly an upper bound on what unsupervised daily use will deliver.

Here I have to give ground, because an earlier draft said the transfer problem went undisclosed and that is not true of either document. DBOH’s evidence statement names the setting, which is why this chapter praises it. And Marinho and colleagues address the question directly in their own abstract:

All trials tested supervised use of fluoride mouthrinse in schools, with two studies also including home use. … Most of the evidence evaluated use of fluoride mouthrinse supervised in a school setting, but the findings may be applicable to children in other settings with supervised or unsupervised rinsing, although the size of the caries-preventive effect is less clear.

— Marinho and colleagues, 2016 (2)

So the reviewers considered the transfer, judged it plausible, and flagged that the effect size does not transfer with it. That is the responsible way to handle it, and my complaint has to narrow accordingly: not that nobody disclosed the gap, but that nobody has quantified it, and that a person handed a bottle of rinse has no way to know whether they should expect 27% or a third of that.

Under GRADE, this is textbook indirectness: a mismatch on the intervention as delivered, and arguably on population, between the evidence and the question. Moderate certainty for supervised school use translates into something lower for unsupervised home use, and the amount of the downgrade is a judgment nobody has published.

But the recommendation is Conditional, and that changes what I can complain about. Conditional means the panel is not confident enough to tell everyone to do this, and expects the decision to depend on the person. That is precisely the right response to evidence that may not transfer. Had this been labeled Strong I would be objecting at length. It is not, so I am not.

The cost side also matters and is favorable: a fluoride rinse is cheap, available over the counter, and its harms are minor, with the practical caveat that it should not be given to children who cannot reliably avoid swallowing it, which is why the age threshold is 8.

PubMed, 21 August 2026, for an update to CD002284. None found; the 2016 version (.pub2) remains current. Its own search ran to 22 April 2016.

I have the abstract for this review, not the full text; it was not among the documents supplied. Statements about the supervised-school setting come from DBOH’s own evidence statement and the review’s abstract, and the strength of this chapter’s argument does not depend on details I have not seen.

Full record in appraisals/searches/ch12-fluoride-rinse.md.

Verdict

Certainty of evidence
Moderate for supervised school-based use, as DBOH states: a prevented fraction of 27% (95% CI 23% to 30%) for permanent tooth surfaces, from 35 trials in 15,305 children and adolescents. Lower for the advice as given, which is unsupervised daily home use. The review judges its findings may still apply there while saying the effect size is “less clear”; nobody has put a number on how much less.
Directness to the advice as worded
Poor on the setting, and both the guideline and the review say so. This is the clearest example in the book of a chapter 13 statement that discloses its own gap, and the review it rests on discloses the same gap in its own abstract. The disclosure is the good practice; the unquantified effect size is what remains.
Is the strength label defensible?
Yes. Conditional is the right call for evidence that may not transfer from supervised to unsupervised use, aimed at a subgroup rather than everyone. Along with Chapter 17 and Chapter 19, this is the guideline handling uncertainty properly.
What would change my mind
A trial of unsupervised daily home rinsing, with adherence measured rather than assumed, in people at high caries risk, reporting a prevented fraction that could be set beside the 27% from supervised use. Also worth knowing: whether “at a different time to brushing” matters, which is the same untested timing claim as Chapter 6 wearing different clothes.

What this means for you

If a dentist has told you that you or your child are at high risk of decay and suggested a daily fluoride rinse, it is a reasonable thing to try. It is cheap and the fluoride will do something.

Expect less than the trials found. Those were school programs where an adult watched every child use it every day, and where the reduction in decayed surfaces was about a quarter. The reviewers think the finding probably carries over to home use; they also say the size of the benefit is less clear there, and nobody has measured it. The honest version of this advice is: this probably helps if you actually do it daily, and daily is harder than it sounds.