The significance of the associations between biomarkers of renal effects and cadmium at low environmental exposure

In this thesis work, we evaluate the current risk assessment methodology for cadmium (Cd), which is largely based on studies among populations with high environmental or industrial exposure to the metal. Basically the methodology relies on the dose-response analysis between the urinary excretion of Cd (U-Cd), used as indicator of Cd body burden, and of low-molecular-weight (LMW) proteins used for screening Cd-induced tubular dysfunction. The threshold for Cd toxicity of 10 µg/g creatinine established 30 years ago was confirmed in a cross-sectional study on a large group of industrial workers. Because of the renal effects of smoking unrelated to Cd, dose-response relationships between U-Cd and LMW proteins in urine are shifted to lower U-Cd values, despite the adjustment for cumulative exposure to tobacco smoke. We therefore suggested that the true Cd threshold for renal dysfunction might be derived after exclusion of smokers only. Using the benchmark dose (BMD) approach with associations undistorted by smoking, we estimated the Cd threshold value for a 5% excess in the background prevalence of abnormal LMW proteins around 9.5 µg/g creatinine with a lower confidence limit between 4.2 and 1.7 µg/g creatinine, for U-RBP and U-β2M, respectively. In a large sample of adolescents with very low environmental exposure, we observed similar significant associations between Cd and LMW proteins in urine with no evidence of threshold. The causality of these associations observed at low exposure levels has been questioned. We proposed a mechanism of co-excretion between Cd bound to metallothionein (MT) and the LMW proteins sharing the same tubular binding sites. Therefore, these associations might be unrelated to Cd-toxicity and simply reflect the physiological variations in the renal uptake capacity of proteins affecting Cd-MT and LMW proteins in the same direction. The urinary excretion of Cd was modeled as a function of age in different age-groups from the general Belgian population. We showed that, over lifetime, U-Cd followed a non-monotonous and nonlinear increase with age with large interindividual variations unlikely to reflect parallel variations in the body burden of the metal. Our findings suggest that at these low exposure levels, U-Cd variations are mainly driven by physiological factors influencing the intake and the renal elimination of the metal. In conclusion, this thesis challenges the current paradigm for the Cd health risk assessment when using U-Cd as marker of the Cd body burden in populations with low-level exposure to the metal. The potential influence of confounding factors on biomarkers variations might lead to an overestimation of health risks of environmental Cd. These findings deserve careful consideration before inferring causation and establishing threshold values from epidemiological studies using U-Cd as exposure indicator.(MED 3) -- UCL, 201