Confounders in the assessment of the renal effects associated with low-level urinary cadmium: an analysis in industrial workers

Abstract

Background: Associations of proteinuria with low-level urinary cadmium (Cd) are currently interpreted as the sign of renal dysfunction induced by Cd. Few studies have considered the possibility that these associations might be non causal and arise from confounding by factors influencing the renal excretion of Cd and proteins.

Methods: We examined 184 healthy male workers (mean age, 39.5 years) from a zinc smelter (n = 132) or a blanket factory (n = 52). We measured the concentrations of Cd in blood (B-Cd) and the urinary excretion of Cd (U-Cd), retinol-binding protein (RBP), protein HC and albumin. Associations between biomarkers of metal exposure and urinary proteins were assessed by simple and multiple regression analyses.

Results: The medians (interquartile range) of B-Cd (μg/l) and U-Cd (μg/g creatinine) were 0.80 (0.45-1.16) and 0.70 (0.40-1.3) in smelter workers and 0.66 (0.47-0.87) and 0.55 (0.40-0.90) in blanket factory workers, respectively. Occupation had no influence on these values, which varied mainly with smoking habits. In univariate analysis, concentrations of RBP and protein HC in urine were significantly correlated with both U-Cd and B-Cd but these associations were substantially weakened by the adjustment for current smoking and the residual influence of diuresis after correction for urinary creatinine. Albumin in urine did not correlate with B-Cd but was consistently associated with U-Cd through a relationship, which was unaffected by smoking or diuresis. Further analyses showed that RBP and albumin in urine mutually distort their associations with U-Cd and that the relationship between RBP and Cd in urine was almost the replicate of that linking RBP to albumin

Conclusions: Associations between proteinuria and low-level urinary Cd should be interpreted with caution as they appear to be largely driven by diuresis, current smoking and probably also the co-excretion of Cd with plasma proteins.

Figure 1

Associations of the creatinine-adjusted concentrations of protein HC and Cd in urine with the urinary concentration of creatinine.

Figure 2

Associations between urinary RBP and the concentrations of Cd in urine or blood in workers stratified in never, ever and current smokers. Concentrations of Cd in urine, expressed per g of creatinine, have been further adjusted for the residual influence of diuresis on the basis of the simple regression coefficient between log U-Cd and log urinary creatinine.

Figure 3

Associations between RBP and Cd in the urine of workers according to the integrity of the glomerular filter (panels A and B, U-ALB<27 mg/g creatinine, panel C, U-ALB>27 mg/g cr mg/g creatinine) and between albumin and Cd according to the integrity of the tubular function (panels D and E, U-RBP<220 μg/g creatinine, panel F, U-RBP>220 μg/g creatinine). The panels B and E represent the same relationships as respectively the panels A and D but by limiting the U-Cd scale to that of the panels C and F in order to ensure that differences in these relationships were not due to differences in U-Cd levels. The cut-off values used to stratify the population correspond to the 90^th percentiles of values obtained over the whole population.

Figure 4

Comparison of the associations of RBP with Cd and albumin in the urine of workers.