Bayesian benchmark dose assessment of fluoride exposure and renal function impairment in adolescents

Adolescents are highly susceptible to fluoride toxicity, yet evidence on renal effects and safety thresholds is limited and inconsistent. Data from adolescents aged 12-19 years in the 2015-2016 National Health and Nutrition Examination Survey (NHANES) were used in this study. The sample sizes were 1,042 for urinary fluoride (UF), 1,062 for plasma fluoride (PF), and 916 for drinking-water fluoride (WF). We analyzed associations between fluoride exposure (UF, PF, and WF) and renal function. Methods included survey-weighted linear/logistic regression, restricted cubic splines (RCS), inverse probability treatment weighting, trend tests, and Bayesian Benchmark Dose modeling (BBMD). RCS indicated linear relationships (P-nonlinear >0.05). Weighted linear regression showed significant inverse associations of UF (β = -11.16 (-20.10, -2.22), P = 0.029) and PF (β = -5.04 (-9.45, -0.62), P = 0.036) with estimated glomerular filtration rate; WF was non-significant (P = 0.585). The 10% benchmark concentration (BMC₁₀) for UF was 1.63 mg/L (10% Benchmark Concentration Lower bound (BMCL₁₀): 1.25 mg/L); for PF, BMC₁₀ was 0.46 μmol/L (BMCL₁₀: 0.24 μmol/L). These renal BMC values were lower than bone-derived benchmarks. Future regulations on environmental fluoride exposure should consider non-skeletal health effects in vulnerable populations.