Skip to main content

Table 5 Determinants of U-Cd in models based on U-SG adjustment

From: The physiological determinants of low-level urine cadmium: an assessment in a cross-sectional study among schoolchildren

   U-Cd (μg/l) adjusted to U-SG U-Cd (μg/l) with U-SG as predictor
Model n Ind. variable Regression coefficient (95% CI) P r2 Ind. variable Regression coefficient (95% CI) P r2
With U-A1M 249 U-Zn 0.20 (0.12 to 0.28) <0.001 0.13 U-SG 34.8 (23.5 to 46.1) <0.001 0.51
   Age 0.07 (0.02 to 0.13) 0.01   U-Zn 0.20 (0.11 to 0.29) <0.001  
   U-A1M 0.07 (0.01 to 0.13) 0.03   Age 0.08 (0.02 to 0.13) 0.008  
       U-A1M 0.07 (0.01 to 0.14) 0.02  
       U-Pb 0.06 (−0.01 to 0.14) 0.10  
With U-RBP 249 U-Zn 0.19 (0.10 to 0.27) <0.001 0.12 U-SG 34.0 (22.5 to 45.5) <0.001 0.51
   Age 0.07 (0.02 to 0.13) 0.01   U-Zn 0.18 (0.09 to 0.27) <0.001  
   U-RBP 0.08 (−0.002 to 0.16) 0.06   Age 0.07 (0.02 to 0.13) 0.01  
       U-RBP 0.09 (0.01 to 0.18) 0.02  
       U-Pb 0.06 (−0.01 to 0.14) 0.10  
With U- β2-m 238 U-Zn 0.20 (0.12 to 0.29) <0.001 0.11 U-SG 39.4 (28.1 to 50.6) <0.001 0.50
   Age 0.07 (0.01 to 0.12) 0.02   U-Zn 0.20 (0.11 to 0.29) <0.001  
       Age 0.07 (0.01 to 0.13) 0.02  
       U-Pb 0.07 (−0.001 to 0.14) 0.09  
With U-CC16 182 U-Zn 0.17 (0.06 to 0.27) <0.001 0.09 U-SG 30.4 (17.4 to 43.4) <0.001 0.42
   U-Pb 0.08 (0.003 to 0.17) 0.042   U-Zn 0.19 (0.09 to 0.30) <0.001  
       U-Pb 0.09 (0.006 to 0.17) 0.04  
  1. All parameters except age were log transformed. Independent variables measured in urine were expressed in the same units as U-Cd. For models built with variables adjusted to U-SG, the highest variance inflation factors (VIF) were 1.001 (U-A1M) in the model with U-A1M, 1.053 (U-Zn) in the model with U-RBP, 1.001 (U-Zn and age) in the model with U-β2-m and 1.005 (U-Zn and U-CC16) in the model with U-CC16. For the models built with U-Cd in μg/l with U-SG among predictors, the highest VIF values were 1.98 (U-SG) in the model with U-A1M, 2.05 (U-SG) in the model with U-RBP, 1.86 (U-SG) in the model with U-β2-m and 1.85 (U-SG) in the model with U-CC16. Contrarily to what is observed with U-Creat, there is no residual correlation between SG-adjusted U-Cd and U-SG, which explains that adding U-SG among independent variables does not change the models based on variables adjusted U-SG