Community Water Fluoridation and Urinary Fluoride Concentrations in a National Sample of Pregnant Women in Canada

Abstract

Background: Fluoride exposures have not been established for pregnant women who live in regions with and without community water fluoridation.

Objective: Our aim was to measure urinary fluoride levels during pregnancy. We also assessed the contribution of drinking-water and tea consumption habits to maternal urinary fluoride (MUF) concentrations and evaluated the impact of various dilution correction standards, including adjustment for urinary creatinine and specific gravity (SG).

Methods: We measured MUF concentrations in spot samples collected in each trimester of pregnancy from 1,566 pregnant women in the Maternal-Infant Research on Environmental Chemicals cohort. We calculated intraclass correlation coefficients (ICCs) to assess variability in MUF concentrations across pregnancy. We used regression analyses to estimate associations between MUF levels, tea consumption, and water fluoride concentrations as measured by water treatment plants.

Results: Creatinine-adjusted MUF values ([Formula: see text]; milligrams per liter) were almost two times higher for pregnant women living in fluoridated regions ([Formula: see text]) compared with nonfluoridated regions ([Formula: see text]; [Formula: see text]). MUF values tended to increase over the course of pregnancy using both unadjusted values and adjusted values. Reproducibility of the unadjusted and adjusted MUF values was modest ([Formula: see text]). The municipal water fluoride level was positively associated with creatinine-adjusted MUF ([Formula: see text], 95% CI: 0.46, 0.57), accounting for 24% of the variance after controlling for covariates. Higher MUF concentrations correlated with numbers of cups of black ([Formula: see text] but not green tea ([Formula: see text]). Urinary creatinine and SG correction methods were highly correlated ([Formula: see text]) and were interchangeable in models examining predictors of MUF.

Conclusion: Community water fluoridation is a major source of fluoride exposure for pregnant women living in Canada. Urinary dilution correction with creatinine and SG were shown to be interchangeable for our sample of pregnant women. https://doi.org/10.1289/EHP3546.

Figure 1.

Fluoride concentrations by trimester in the urine of pregnant women from the MIREC cohort living in fluoridated versus nonfluoridated communities. MUF (maternal urinary fluoride) levels are shown unadjusted and adjusted for specific gravity (${\text{MUF}}_{\text{SG}}$; Equation 1) and creatinine using two different methods (${\text{MUF}}_{\text{CRE}\_1}$ and ${\text{MUF}}_{\text{CRE}\_2}$; Equations 2 and 3, respectively). Box plots display the upper and lower quartiles of the data; the median is marked by the vertical line inside the box. The whiskers show the 5th and 95th percentile, whereas the individual data points represent values that exceed the 95th percentile. Box plots were produced after removing outliers defined as a MUF concentration $\ge 5$.

Figure 2.

Pearson r correlations between pairs of trimesters (T1, T2, T3) and intraclass correlation coefficients (ICCs) across trimesters [with 95% confidence interval (CI)] for ${\log }_{10}\text{-transformed}$ maternal urinary fluoride (MUF) levels without adjustment, with adjustment for specific gravity (${\text{MUF}}_{\text{SG}}$), and with adjustment for creatinine using two methods of adjustment (${\text{MUF}}_{\text{CRE}\_1}$ and ${\text{MUF}}_{\text{CRE}\_2}$). Individual data points represent individual observations, solid lines represent regression lines.

Figure 3.

${\text{Log}}_{10}\text{-transformed}$ maternal urinary fluoride (MUF) exposure levels as a function of water treatment plant fluoride levels. MUF levels are shown unadjusted and adjusted for specific gravity (${\text{MUF}}_{\text{SG}}$) and creatinine using two different methods (${\text{MUF}}_{\text{CRE}\_1}$ and ${\text{MUF}}_{\text{CRE}\_2}$). Individual data points represent individual observations. Solid lines represent regression lines.