N-acetyl-beta-D-glucosaminidase (NAG) as the most sensitive marker of tubular dysfunction for monitoring residents in non-polluted areas.

Alpha(1)-microglobulin (alpha(1)-MG), beta(2)-microglobulin (beta(2)-MG) and N-acetyl-beta-D-glucosaminidase (NAG) are tubular dysfunction markers often used in cadmium (Cd) epidemiology. The purpose of the present study was to identify the best maker among the above-referred three that correlates most closely with Cd in urine of residents with no known Cd pollution. Survey was conducted in 2007-2008 in three prefectures in Japan. Adult women, 2163 in total, participated in the survey; they provided informed consents, offered spot urine samples and filled questionnaires on possible confounders of Cd burden. Urine samples were analyzed for Cd, alpha(1)-MG, beta(2)-MG and NAG together with calcium (Ca), magnesium (Mg), creatinine (CR) and specific gravity (SG). The analyte levels, as observed (e.g., Cd(ob)) or after correction for CR (e.g., Cd(cr)) or SG (e.g., Cd(sg)) were subjected to simple and multiple regression analysis. Correlation matrix analysis with observed values for total cases showed that the coefficients of correlation with Cd were highest for NAG, followed by alpha(1)-MG, and lowest for beta(2)-MG. Multiple regression analysis by three prefectures either separately or in combination (thus four analyses) disclosed that Cd was the independent variable most influential to NAG (as the dependent variable) throughout the four analysis conditions with high R(2) values (>0.3), whereas the most influential variables were not the same depending on the analysis conditions in cases with alpha(1)-MG and beta(2)-MG. When coefficients of correlation for the three dysfunction markers with Cd were compared among the observed, CR- and SG-corrected values, the coefficients for the observed values were higher than the counterpart values for CR- or SG-corrected values. In conclusion, NAG rather than alpha(1)-MG or beta(2)-MG should be recommended for monitoring Cd exposure-related tubular effects among general populations. Observed (i.e., un-corrected) values rather than CR- or SG-corrected values should be used.