The association between metal exposure and body mass index of preschool children of Shaanxi, China

There are few studies on the association between polymetallic exposure and obesity in preschool children, and the conclusions are inconsistent. The study aimed to explore the association between urine metals and body mass index (BMI).For 342 preschool children in the industrial park of Shaanxi, China, uniformly trained investigators collected basic information of preschoolers and measured height and weight. Morning urine were collected, and BMI were calculated. We measured concentrations of 22 urine metals by quadrupole inductively coupled plasma-mass spectrometry (ICP-MS). The least absolute shrinkage and selective operator (LASSO) regression model was used to screen urine metals larger impact on BMI. In the polymetallic exposure model, generalized linear regression was used to explore the relationship between the selected urine metals and BMI. The overall relationship between urine metals and BMI was analyzed using restricted cubic spline (RCS). The interaction method was used to explore the influence of metal-metal interaction on BMI. After adjusting for covariates, for each unit increase in ln-transformed levels of copper and molybdenum, BMI level decreased by -1.219 (-2.047, -0.391), and -0.797(-1.508, -0.087) kg·m−2 (P<0.05), respectively. Polymetallic model showed similar results. BMI level decreased by - 1.556(-2.488, -0.624) kg·m-2 (P<0.05) for each unit increase in ln-transformed levels of copper (Cu), and for each unit increase in ln-transformed levels of molybdenum (Mo) level, BMI level decreased by -0.915 (-1.817, -0.013) kg·m−2 (P<0.05). The results of RCS visually demonstrate that Cu and BMI are non-linearly correlated (P total < 0.001, P nonliearity = 0.005), Mo and BMI are linearly correlated (P total < 0.05, P nonliearity = 0.099). Urinary Cu was negatively correlated with BMI nonlinearly in preschool children, and there was a borderline interaction between urinary Mo and Cu levels. Further researches are needed to validate these associations and reveal underlying mechanisms