Reduction in dental caries with four concentrations of sodium fluoride in a dentifrice: a meta-analysis evaluation.

Information on the effects of fluoride concentrations above 1,100 ppm in dentifrices is not extensive in the literature. The objective of this meta-analysis was to examine and compare the anticaries effectiveness (in terms of DMFS scores) of 1,700 ppm, 2,200 ppm and 2,800 ppm F- ion (as sodium fluoride) dentifrices vs. an 1,100 ppm F- ion (as sodium fluoride) control dentifrice based on results from six double-blind, randomized clinical studies, each conducted over a two- to three-year period. The studies each enrolled approximately 1,200-2,000 male and female school children per treatment group in grades I through 8, and were conducted in areas with low fluoride content water supplies in the states of Indiana, Pennsylvania, Ohio and Oregon. Separate meta-analyses were performed on the study results (DMFS increment scores determined by visual-tactile examinations supplemented with radiographs) for the one-year, two-year and three-year examinations. Comparisons of the 1,700 ppm F-, 2,200 ppm F-, and 2,800 ppm F- groups vs. the 1,100 ppm F- group were based on pooling the effect sizes for these comparisons from the individual studies. The effect sizes were calculated in two different ways, reflecting the analyses that were performed in the original studies: 1) effects based on the sample means and variances; and 2) effects based on the adjusted sample means and mean squared error from an analysis of covariance. The results obtained from this meta-analysis provide evidence that the use of a 2,800 ppm F- ion, as sodium fluoride, dentifrice results in statistically significantly lower caries increment than the use of an 1,100 ppm F- ion, as sodium fluoride, dentifrice. This result was noted after one, two, and three years of dentifrice use. The 1,700 ppm F- and 2,200 ppm F- dentifrice groups showed some directional advantages over the 1,100 ppm F- dentifrice group, however the analysis did not establish these groups as statistically significantly better than 1,100 ppm F-. The meta-analysis based on analysis of covariance results was somewhat more sensitive to treatment group differences than the analysis based on sample means and variances, as was expected.