Bias induced by the use of creatinine-corrected values in evaluation of beta2-microgloblin levels

The present study was initiated to examine if the correction for creatinine (CR or cr) is the best approach among the three methods of correction for CR, correction for a specific gravity (SG or sg) and the use of observed values in managing difference in urine density. For this purpose, a database previously developed on 10,753 adult women in 10 non-polluted areas in Japan was re-visited for information on age, urinary levels of Cd, Mg, Ca, Zn, beta(2)-MG, and creatinine, and urine specific gravity as well as smoking habits. Never-smoking women with various urine density counted 8975 cases (the various urine density group). From these cases, 7081 cases with adequate urine density (i.e. 0.5 g/l < or = CR < or = 3.0 g/l and 1.010 < or = SG < or = 1.030) were selected (the adequate urine density group). When a beta(2)-MG level of 400 microg/g CR or 400 microg/l was taken as a cut-off value for beta(2)-MG-uria, both the prevalence of beta(2)-MG(cr)-uria [i.e. cases with beta(2)-MG (as corrected for CR) in excess of 400 microg/g cr] and that of beta(2)-MG(sg)-uria increased as a function of the decrease in Cd(cr) or Cd(sg). The prevalence of beta(2)-MG(ob)-uria also varied as a function of CR and SG, especially of CR, but its range of variation was smaller than the corresponding changes in beta(2)-MG(cr)-uria prevalence. A noteworthy advantage for the use of observed values over that of SG-corrected values was the minimum effect of age. In over-all evaluation, therefore, the recommended approach appeared to be the use of non-corrected observed values (after selection of urine samples for adequate urine density if desired) or correction for SG, rather than correction for CR.