Blood metals and vitamin D status in a pregnancy cohort: A bidirectional biomarker analysis

Low 25-hydroxyvitamin D (25OHD), a biomarker of vitamin D status, is associated with reduced immune function and adverse pregnancy outcomes, such as preterm birth. Observational studies indicate that long-term, high level exposure to metals such as cadmium (Cd) and lead (Pb) can impact a person's vitamin D status. However, the directionality of the association is uncertain, particularly for low-level exposures. We used three distinct longitudinal data analysis methods to investigate cross-sectional, longitudinal and bidirectional relationships of Cd and Pb biomarkers with 25-hydroxyvitamin D (25OHD) in a Canadian pregnancy cohort. Maternal whole blood Cd and Pb and plasma 25OHD concentrations were measured in the 1st (n = 1905) and 3rd (n = 1649) trimester and at delivery (25OHD only, n = 1542). Our multivariable linear regression analysis showed weak inverse associations between Cd and 25OHD concentrations cross-sectionally and longitudinally while the latent growth curve models showed weak associations with Pb on the 25OHD intercept. In the bidirectional analysis, using cross lagged panel models, we found no association between 1st trimester metals and 3rd trimester 25OHD. Instead, 1st trimester 25OHD was associated with 9% (-15%, -3%) lower 3rd trimester Cd and 3% (-7, 0.1%) lower Pb. These findings suggest the 25OHD may modify metal concentrations in pregnancy and demonstrates the value of controlling for contemporaneous effects and the persistence of a biomarker over time, in order to rule out reverse causation.