Associations between urinary excretion of cadmium and proteins in a nonsmoking population: renal toxicity or normal physiology?

Abstract

Background: Associations between cadmium (Cd) and kidney function have been reported even at low levels of exposure in the general population. Recently, the causality of these associations has been questioned.

Objectives: We examined associations between urinary Cd (U-Cd; a biomarker of exposure) and urinary proteins that are used as biomarkers of kidney effects, based on repeated short-term sampling in healthy subjects.

Methods: Twenty-four hour urine samples were collected on 2 separate days at six fixed times from 30 healthy nonsmoking men and women (median age 39 years). We analyzed the samples (N = 354) for Cd (i.e., U-Cd) and two proteins used as kidney function biomarkers: urinary albumin (U-Alb) and alpha-1-microglobulin (U-A1M). Concentrations were adjusted for creatinine concentration or for specific gravity, and excretion rates (mass per hour) were calculated. Possible associations were assessed within each individual participant, and mean correlations and regressions were evaluated.

Results: We found clear positive mean associations within individuals between the excretion of U-Cd [mean, 0.11 µg/g creatinine (range, 0.01-0.52 µg/g creatinine)] and both U-Alb and U-A1M. The associations were stronger for excretion rates and concentrations adjusted for specific gravity than for concentrations adjusted for creatinine. We also found significant positive associations of urinary flow with excretion of U-Cd, U-Alb, and U-A1M.

Conclusions: Associations between short-term changes in U-Cd and markers of kidney function within individual nonsmoking study participants are unlikely to reflect effects of Cd toxicity. A more likely explanation is that these associations result from normal variation in renal function, including changes in urinary flow, that influence the urinary excretion of both Cd and proteins in the same direction. These effects of normal variability may result in overestimation of the adverse effects of Cd on kidney function at low-level Cd exposure.

Figure 1

Example showing the association, for 1 of 30 individual participants, between Cd excretion rates (A,B) and Cd concentrations adjusted for SG (C,D), and Cd concentrations adjusted for U-Crea concentration (E,F) and U-Alb (A,C,E), or U-A1M (B,E,F), respectively.

Figure 2

Distributions of individual Spearman correlation coefficients (r_S; A,B) and individual regression coefficients (β; C,D) for associations between U-Cd and U-Alb (A,C), and U-Cd and U-A1M (B,D), calculated for 30 individual participants [individual values based on 6 samples per day over 2 days for each participant, 354 total samples (1 participant had 6 samples only)]. Boxes indicate 10th and 25th percentiles; dotted line, mean; solid line, median; and whiskers, 75th and 90th percentiles across all participants, with dots indicating outliers. Excretion rates in µg/hr for U-Cd, mg/hr for U-Alb and U-A1M; SG-adjusted concentrations in µg/L for U-Cd, mg/L for U-Alb and U-A1M; U-Crea–adjusted concentrations in µg/g creatinine for U-Cd, mg/g creatinine for U-Alb and U-A1M. All measures deviate significantly from 0 (p < 0.05).