Urinary Thallium Concentration and Cardiovascular Mortality: An Exploratory NHANES Analysis with Disease-Context Single-Cell Transcriptomic Hypothesis Generation

Environmental thallium exposure is a growing public health concern, but evidence regarding its association with cardiovascular mortality remains limited. We evaluated this association in NHANES and used public disease-context single-cell transcriptomic datasets to generate candidate hypotheses for future experimental study. In the fully adjusted original-scale analysis, urinary thallium concentration showed a nominal positive association with cardiovascular mortality (OR = 2.13, 95% CI: 1.21-3.75; unadjusted P = 0.011). The multi-metal screening analyses were exploratory, and P values were not adjusted for multiple comparisons. After natural-log transformation of urinary thallium, the association with cardiovascular mortality was not statistically significant (OR = 1.53, 95% CI: 0.71-3.28; P = 0.264). Exploratory analyses of public disease-context single-cell datasets identified expression patterns of database-derived thallium-related candidate genes across vascular smooth muscle cells (VSMCs), macrophages, and other cell types. These candidate genes overlapped with disease-associated transcriptional programs annotated to VSMC phenotypic state and lipid metabolism. Because neither dataset included donor-level thallium measurements, these observations were used only to nominate hypotheses for future studies and were not interpreted as evidence of thallium-induced cellular changes or regulatory pathways. Because donor-level thallium measurements were unavailable, these transcriptomic observations are hypothesis-generating and do not establish that thallium exposure directly caused the observed cellular changes or activated specific regulatory pathways. In this exploratory cross-sectional analysis, higher urinary thallium concentration was positively associated with cardiovascular mortality in the original scale, but this association was not robust to natural-log transformation. Single-cell transcriptomic analyses of disease-context datasets identified candidate genes and pathways expressed in vascular smooth muscle cells and macrophages, but these findings are hypothesis-generating and require validation in studies with repeated exposure assessment and experimental models. The present results should be interpreted as preliminary and not as evidence of a causal relationship.