Bi-linear dose--response relationship in general populations with low-level cadmium exposures in non-polluted areas in Japan.

ABSTRACT

BACKGROUND AND OBJECTIVES: Previous findings suggest that the dose (e.g., cadmium in urine)-response (e.g., tubular dysfunction markers in urine) relationship after environmental exposure to cadmium (Cd) may be in a shape of a hockey stick with a steep slope when Cd exposure exceeds a certain level, i.e., the point of flexion. However, less attention has been paid to the relation below the flexion point whether it is linear or not. The present study was initiated to examine through epidemiological analyses whether the assumption of linearity holds true when Cd exposure is low. METHODS: Data were cited from previous publications of this research group on Cd, α(1)-microglobulin (α(1)-MG), ß(2)-microglobulin (ß(2)-MG), N-acetyl-ß-D -glucosaminidase (NAG), and creatinine (CR) in urine from more than 17,000 adult Japanese women. The cases were selected for age (50-59 years), urine specific gravity (1.010-1.030) and CR (0.3-3.0 g/l), and 5,306 cases were available after selection (the group selected in terms of age and urine density, or the AD-selected group). RESULTS: Statistical analyses of the AD-selected group revealed that the relation was not linear. In case of ß(2)-MG, the slope of the regression line between Cd as an independent variable and ß(2)-MG a dependent variable was steeper at lower Cd level (i.e., <2.5 µg/l) and shallower when Cd was higher, showing a clear trend of leveling off. The same was also the case for α(1)-MG and NAG. When Cd was 2.0-2.1 µg/l or higher, α(1)-MG, ß(2)-MG, and NAG were in excess of the 95%-tile values for the markers in more than 5% of the population. The significance of the observation was discussed in reference to existing criteria for Cd such as 2.5 or 5.2 µg/g cr or for ß(2)-MG of 300 µg/g cr. CONCLUSIONS: When the present observation of bilinear relation at low Cd exposure levels was combined with the previous observation of very sharp increase in responses among heavily exposed cases, it appeared likely that the overall relationship is tri-phasic in Cd dose-response relationship in humans. A substantial fraction (>5%) of the populations had the marker levels in excess of the 95%-tile values when Cd was ≥2 µg/l. Comparison with findings in reported literature suggests that the presence of such cases by itself may not necessarily mean a risk of tubular dysfunction and that other factors including Cd levels in urine should be considered together.