Benchmark dose for cadmium-induced renal effects in humans

Our goal in this study was to explore the use of a hybrid approach to calculate benchmark doses (BMDs) and their 95% lower confidence bounds (BMDLs) for renal effects of cadmium in a population with low environmental exposure.

Morning urine and blood samples were collected from 820 Swedish women 53-64 years of age. We measured urinary cadmium (U-Cd) and tubular effect markers [N-acetyl-beta-d-glucosaminidase (NAG) and human complex-forming protein (protein HC) ] in 790 women and estimated glomerular filtration rate (GFR; based on serum cystatin C) in 700 women. Age, body mass index, use of nonsteroidal anti-inflammatory drugs, and blood lead levels were used as covariates for estimated GFR. BMDs/BMDLs corresponding to an additional risk (benchmark response) of 5 or 10% were calculated (the background risk at zero exposure was set to 5%) . The results were compared with the estimated critical concentrations obtained by applying logistic models used in previous studies on the present data.

For both NAG and protein HC, the BMDs (BMDLs) of U-Cd were 0.5-1.1 (0.4-0.8) microg/L (adjusted for specific gravity of 1.015 g/mL) and 0.6-1.1 (0.5-0.8) microg/g creatinine. For estimated GFR, the BMDs (BMDLs) were 0.8-1.3 (0.5-0.9) microg/L adjusted for specific gravity and 1.1-1.8 (0.7-1.2) microg/g creatinine.

The obtained benchmark doses of U-Cd were lower than the critical concentrations previously reported. The critical dose level for glomerular effects was only slightly higher than that for tubular effects. We suggest that the hybrid approach is more appropriate for estimation of the critical U-Cd concentration, because the choice of cutoff values in logistic models largely influenced the obtained critical U-Cd.