Association of multiple metals with lipid markers against different exposure profiles: A population-based...

We sought to evaluate whether essential and toxic metals are cross-sectionally related to blood lipid levels using data among adults from Shimen (n = 564) and Huayuan (n = 637), two counties with different exposure profiles in Hunan province of China. Traditional and grouped weighted quantile sum (WQS) regression and Bayesian kernel machine regression (BKMR) were performed to assess association between exposure to a mixture of 22 metals measured in urine or plasma, and lipid markers. Most of the exposure levels of metals were significantly higher in Shimen area than those in Huayuan area (all P-values < 0.001). Traditional WQS regression analyses revealed that the WQS index were both significantly associated with lipid markers in two areas, except for the HDL-C. Grouped WQS revealed that essential metals group showed significantly positive associations with lipid markers except for HDL-C in Huayuan area, while toxic metals group showed significantly negative associations except for HDL-C and LDL-C in Huayuan area. There were no significant joint effects, but potential non-linear relationships between metals mixture and TC or LDL-C levels were observed in BKMR analyses. Although consistent significantly associations of zinc and titanium with TG levels were found in both areas, the metals closely related to other lipid markers were varied by sites. Additionally, the BKMR analyses revealed an inverse U shaped association of iron with LDL-C levels and interaction effects of zinc and cadmium on LDL-C in Huayuan area. The relationship between metal exposure and blood lipid were not identical against different exposure profiles.