The association between fluoride in drinking water and dental caries in Danish children. Linking data from health registers, environmental registers and administrative registers

Objectives: To study the association between fluoride concentration in drinking water and dental caries in Danish children. Methods: The study linked registry data on fluoride concentration in drinking water over a 10-year period with data on dental caries from the Danish National Board of Health database on child dental health for 5-year-old children born in 1989 and 1999, and for 15-year-old children born in 1979 and 1989. The number of children included in the cohorts varied between 41.000 and 48.000. Logistic regression was used to assess the correlations, adjusting for gender and taxable family income as a proxy variable for socioeconomic status. Results: Fluoride concentration in drinking water varied considerably within the country from very low (< 0.10 mg/l) to more than 1.5 mg/l. Only little variation was found over the 10-year study period. Dental caries in both 5-year-olds and 15-year-olds decreased over the study period. An inverse relation between the risk of dental caries and fluoride concentration in drinking water was found in both primary and permanent teeth. The risk was reduced by approximately 20% already at the lowest level of fluoride exposure (0.125-0.25 mg/l). At the highest level of fluoride exposure (> 1 mg/l), a reduction of approximately 50% was found. Similar findings were found if analysis was limited to children residing in the same place during the entire study period. Conclusions: The study confirmed previous findings of an inverse relation between fluoride concentration in the drinking water and dental caries in children. This correlation was found in spite of the extensive use of fluoridated toothpaste and caries-preventive programs implemented by the municipal dental services in Denmark. Linking Danish health registers with environmental and administrative registers offers an opportunity for obtaining sample sizes large enough to identify health effect, which otherwise could not be identified.