Objective

We aimed to study the association of urine fluoride with intelligence quotient (IQ) in children with a careful consideration of up to 30 potential confounding factors as well as possible heterogeneity of the relation between urine fluoride levels and IQ scores across children with different dopamine receptor-2 (DRD2) Taq 1A genotypes (CC, CT, and TT).

Methods

A school-based cross-sectional study design was applied. A total of 323 children (2014–2015, 7–12 years old) were enrolled from four schools in both historical endemic and non-endemic areas of fluorosis in Tianjin of China using a cluster sampling method. Urine fluoride levels and age-specific IQ scores in children were measured at the enrollment. Polymerase chain reaction-restriction fragment length polymorphism methods were used to genotype DRD2 Taq 1A polymorphism with genomic DNA isolated from whole blood collected at the enrollment. Multiple linear regression models were applied to evaluate the relationship between urine fluoride levels and IQ scores overall and within the DRD2 Taq 1A SNP = CC/CT and TT subgroups. Model robustness was tested through bootstrap, sensitivity analysis, and cross-validation techniques. A safety threshold of urine fluoride levels for IQ impairment was determined in the subgroup TT.

Results

In overall participants, the DRD2 Taq 1A polymorphism itself was not related to IQ scores in children who had a high level of urine fluoride. In the CC/CT subgroup, urine fluoride levels and IQ scores in children were unrelated (adjusted β (95% confidence interval (CI)) = − 1.59 (− 4.24, 1.05), p = 0.236). Among the participants carrying the TT genotype, there was a strong and robust negative linear relationship between log-urine fluoride and IQ scores in children (adjusted β (95% CI) = − 12.31 (− 18.69, − 5.94), p < 0.001). Urine fluoride levels had a stronger association with IQ in children carrying the TT genotype (adjusted β = − 12.31, bootstrapped standard error (SE) = 1.28), compared to that in overall participants (adjusted β = − 2.47, bootstrapped SE = 3.75) (Z = 2.483 and bootstrapped p = 0.007). The safety threshold of urine fluoride levels in the subgroup TT was 1.73 mg/L (95% CI = (1.51, 1.97) (mg/L)).

Conclusions

There is heterogeneity in the relation between urine fluoride and IQ across children carrying different DRD2 Taq 1A genotypes. Large-scale epidemiological studies are needed to confirm our findings.