Conditional neuroprotection: Blood selenium attenuates the cognitive toxicity of DEHP metabolites in older adults.

Background

Phthalates, particularly di(2-ethylhexyl) phthalate (DEHP), are ubiquitous environmental contaminants with well-documented neurotoxic potential. Selenium, an essential antioxidant trace element, may mitigate oxidative stress-induced neuronal damage. However, whether selenium modifies the cognitive effects of phthalate exposure remains unclear in human populations.

Methods

We analyzed cross-sectional data from adults aged ≥60 years in the National Health and Nutrition Examination Survey (NHANES) 2011-2014. Urinary phthalate metabolites and blood metals (lead, cadmium, mercury, manganese) were quantified. Survey-weighted linear regression with interaction terms tested whether blood selenium modified the associations between these exposures and global cognitive Z-scores. Effect modification was further evaluated using restricted cubic splines, simple slope analysis, and joint Wald tests. Oxidative stress biomarkers (alkaline phosphatase, total bilirubin, serum iron) were assessed as potential mediators. To contextualize the main NHANES findings, we performed two supplementary analyses: (i) bidirectional two-sample Mendelian randomization (MR) to test whether selenium has an independent causal effect on Alzheimer's disease (AD), and (ii) an exploratory ecological analysis using Global Burden of Disease (GBD) 2021 data to illustrate the risk of ecological fallacy when environmental co-exposures are unaccounted for. These ancillary analyses do not constitute formal triangulation, as they address selenium's main effects rather than its interaction with phthalates.

Results

Among 573 participants (mean age 69.5 years, 52.2% female), blood selenium significantly modified the associations of four DEHP metabolites with cognitive function, with the strongest interaction observed for MEHP (β = 0.00103, 95% CI: 0.00081-0.00124, P < 0.001). In individuals with low selenium, higher MEHP was associated with lower cognitive scores; this association was reversed in those with high selenium. In contrast, no significant interactions were found for any heavy metal (all P > 0.05). Joint Wald tests confirmed effect modification for phthalates (F = 13.34, P = 0.003) but not for metals (F = 1.03, P = 0.43). Oxidative stress biomarkers did not mediate the observed interactions. Bidirectional MR found no causal effect of selenium on AD (IVW β = -0.008, P = 0.87), ruling out reverse causation as an explanation. An exploratory ecological analysis using GBD data found a counterintuitive inverse association between dietary selenium inadequacy and dementia burden (β = -0.325, P = 0.037). This paradoxical pattern remained consistent in sensitivity analyses and likely reflects ecological confounding. It underscores that population-level nutrient-disease links can be strongly distorted by differences in diagnostic practices and unmeasured co-exposures, such as phthalates.

Conclusions

Selenium reduces the cognitive harm of DEHP metabolites in older adults, with clearer and more consistent evidence for phthalates than for heavy metals. Null MR results rule out an independent causal effect of selenium on Alzheimer's disease, reinforcing that selenium's role is conditional rather than universal. A paradoxical ecological pattern, presented as a cautionary example, highlights the risk of ecological fallacy when co-exposures are unaccounted for. Together, these findings suggest that selenium's neuroprotection is context-dependent-it appears to act mainly under specific toxicant exposure rather than offering universal benefits. These findings caution against indiscriminate selenium supplementation and suggest that further longitudinal studies are needed before targeted interventions can be recommended.