The role of mercury, selenium and the Se-Hg antagonism on cognitive neurodevelopment: A 40-month follow-up of the Italian mother-child PHIME cohort

Despite a 15-year long effort to define the “safety” of fish intake during pregnancy, there remains still uncertainty on this important public health issue. The evaluation of the toxic effects of contaminants, particularly mercury (Hg) in fish-eating populations is complicated by the fact that sea-food is also rich in beneficial nutrients, such as selenium (Se). There is toxicological plausibility of an antagonistic effects between Se and Hg, and some theoretical support for the inclusion of the Se–Hg interaction to better assess the risk linked with fish intake. To assess the effects of exposure to low-level Hg through fish consumption on the developing brain and the interaction between Hg and Se, we conducted an analysis at age 40 months in Italian children, enrolled in a prospective mother-child cohort, comparing additive and multiplicative models. Participant subjects were the 470 children born within the Northern Adriatic Cohort II (NAC-II) cohort who were tested by using the Bayley Scales of Infant and Toddler Development third edition (Bayley-III) (BSID-III) at age 40. Family demographic and socioeconomic information, pregnancy and delivery history, parental and child medical history and food consumption were assessed through questionnaires. Maternal blood samples were collected during pregnancy, cord blood at birth and maternal milk 1 month after delivery. As other exposures of interest, we considered the level of Se in maternal and cord blood and in breast milk and the potential Se–Hg antagonism. Se and inverse of THg (1:THg) concentrations were categorized according to the tertiles of their distributions, in low, medium and high levels of exposure. The lower end of the composite cognitive score distribution closest to 20% was defined as suboptimal development. Multiple logistic regression were applied to assess the association between the dichotomized composite cognitive score and the categorized exposure to Se and 1:THg, and the antagonism between Se and 1:THg. In the recruiting period, 900 pregnant women were enrolled in the cohort; 767 of these remained in the study at delivery and 470 children at 40 months. After excluding preterm births, 456 children were used in the final analyses. The larger difference in risk for suboptimal neurodevelopment was observed for the category with High THg and Low Se with OR = 2.55 (90% CI 1.02; 6.41) under the multiplicative and OR = 1.33 (90% CI 0.80; 1.87) under the additive model. The category High THg and High Se showed a very slightly better fit of the additive model (OR = 1.07, 90% CI 0.65; 1.50) versus the multiplicative (OR = 1.66, 90% CI 0.73; 1.77). A negative – antagonistic – interaction term for this category was estimated under the multiplicative model giving an OR = 1.17 (90% CI 0.42; 3.28). Although this evidence of the effects of the Se–Hg antagonism on the children neuro-development needs to be confirmed, if Se can counterbalance Hg toxicity, the evaluation of the effect on human health of fish consumption, should also consider the diverse ratios between Se and Hg concentration in different fish species.