Fluoride and Neurodevelopmental Hazard Modelling: An Assessment of Concentration-Response Analysis

A National Academies Consensus Study report concluded that the evidence did not support an assessment that fluoride is a neurodevelopmental hazard. However, some researchers have undertaken benchmark dose modelling to determine a safe fluoride concentration level in water. Therefore, the suitability of the data for modelling fluoride concentration in urine and water and cognition response using standard criteria was assessed.

Data quality was evaluated using a standard tool. A random-effects meta-analysis of standardised mean difference (SMD) and regression coefficients was conducted to assess effect sizes and heterogeneity. The Environmental Protection Agency (EPA) benchmark dose modelling was utilised to determine the association between fluoride concentrations and cognition scores.

All four maternal urinary fluoride (MUF) studies did not meet the standards for acceptable quality, as identified by the EPA data quality criteria, which are necessary for combining data from different studies for dose-response analysis. The pooled estimate was not statistically significant (β_MUF = -1.06, 95% CI: -3.63, 1.50; p = 0.42; I² = 62%). A meta-analysis of five studies conducted in fluoridated areas showed a pooled SMD effect size of 0.04 (95% CI: -0.06, 0.14; p = 0.42; I² = 0%), favoring higher fluoride. The benchmark dose models did not reveal a functional relationship between MUF or water fluoride concentration and cognitive outcomes (Goodness-of-fit p < 0.1).

The data quality assessment revealed serious flaws that render the maternal urinary studies unacceptable for hazard assessment and benchmark dose modelling. Therefore, more appropriate studies in endemic fluorosis areas are needed to accurately determine whether fluoride is associated with adverse cognitive outcomes in populations with meaningful exposure.