Commentary

Professionally applied fluoride varnish (22,600ppm F-) is a common caries preventive measure. Its use is widely recommended; EAPD, AAPD, SIGN, SDCEP and DoH Oral Health Guidance.1,2,3,4,5 In Scotland, its routine use in pre-school children forms a cornerstone of the Scottish Government's Childsmile Programme.6

In their trial, Oliveira and colleagues investigated the use of six-monthly fluoride varnish application amongst high caries risk 1-4 year olds living in Rio de Janeiro, Brazil, (fluoridated public water supply) over a two year period. At enrolment the child's carer attended an oral health educational session, and throughout the trial children had regular toothbrushing instruction (including the free provision of toothbrushes and 1450 ppm F- toothpaste). Following random allocation children received either 22,600ppm F- fluoride varnish or a placebo (identical apart from fluoride content) at six-montly intervals.

Though otherwise methodologically sound, this clearly reported study had a significant issue with its power calculation and sample size. It appears that the study was underpowered, being based on the local caries incidence and an 18% expected difference between the control and fluoride varnish groups, which gave an indicative sample size of 85 participants per group. A post-study power calculation, based on the actual caries incidence within the study population, indicated that a sample size of 332 in each group was required to detect an effect of the fluoride varnish. This makes any conclusions somewhat uncertain; the lack of statistical difference may have been a result of inadequate numbers of participants to detect a difference and should not be interpreted as a lack of effectiveness. This highlights the importance and difficulty of undertaking an appropriate power calculation at the outset of a study.

Cochrane reviews on the topic of fluoride indicate that the Preventive Fraction (PF) for fluoride varnish versus placebo/no treatment to be 43%;7 the PF for fluoride varnish in combination with fluoride toothpaste is reported as an additional 10% over fluoride toothpaste alone.8 Based on the figures, a larger sample size was likely required to detect a benefit from fluoride varnish when used in combination with the additional fluorides the participants were exposed to during the trial. Given the trial's low sample size from the initial power calculation, there is a significant risk of a type II error – failure to reject the null hypothesis and not detect an effect when one exists.

The authors question which surfaces benefit most from the anti-caries effect of fluoride varnish, querying whether it is sound enamel surfaces that derive the most benefit. Their rationale for this is the suggestion from the data that children caries free as baseline showed a greater propensity to remain so at two year follow-up. However, it seems reasonable to suggest that caries free children represent a lower risk group and so a high proportion would be expected to remain caries free at two years.

Whilst the authors obviously took great lengths to conduct a truly placebo-controlled study of fluoride varnish, they included a number of additional preventive interventions (ie oral health education, free toothbrushes, free 1450 ppm F- toothpaste and toothbrushing instruction). It is not unreasonable to include these interventions, particularly if they form part of the routine care received by Brazilian children, and to the trial's credit, they were applied to both groups equally. However, these additional preventive interventions, particularly 1450 ppm F- toothpaste in conjunction with a fluoridated public water supply, are likely to have reduced the estimate of the preventive effect of six-monthly fluoride varnish applications. There was a tendency towards a reduction in the incidence of caries and early enamel lesions in the fluoride varnish arm of the trial. A higher background dmft, lower background fluoride, longer follow-up or an increased sample size may have detected a benefit. This paper highlights why care should always be taken before drawing direct clinical conclusions from an individual study. Fluoride varnishes remain an integral preventive agent when there are no additional sources of fluoride, and may still be of benefit when used in combination with other fluoride interventions.