The Precision and Accuracy of Urinary Creatinine Analysis for Correction of Urine Dilution in Biological Monitoring Measurements

Abstract

Biomonitoring has been widely used to assess environmental exposure, and spot urine is commonly studied due to its easy sample collection. The variability of urine dilution has been crucial for assessing the exposure magnitude. Previous studies indicated that creatinine-adjusted concentrations are better correlated with the parent chemical concentrations in blood, plasma or serum than unadjusted concentrations. More recent studies have documented problems with creatinine-adjustment in pregnant women and diverse populations because creatinine varies with age, gender, race, body size, and diet.In addition to the inter- and intra-person variability in creatinine, we found significant variability in the testing methods and the testing laboratories. Two immunoassays gave decidedly different concentrations on the same samples in two different laboratories and even had a greater than 35% bias on NIST Standard Reference Materials. In an attempt to standardize the creatinine measurements, both laboratories adopted the same immunoassay.The inter-laboratory relative percent differences (RPD) ranged from 0.3-13.4% with 5.2% median, and the values determined in each laboratory were highly correlated (Pearson correlation coefficient: r=0.98). Although the results were correlated and relatively precise, 15% biases were found between laboratories with values determined by laboratory B being uniformly higher. The directional significant differences were confirmed by paired t-test (p-value<0.001). Thus, a systemic error can occur among different laboratories which might alter the point estimates of central tendency and percentile concentrations.The results show that in addition to the inherent limitations of biological variation in using creatinine-adjustment in biomonitoring, laboratory error from non-standardized measurements can further introduce variation. Taken together, we suggest that the standardized creatinine measurement or cross-laboratories validation is needed.