CitationWillman, Eric J.; Hertz-Picciotto, Irva; Keller, Jean A.; Martinez, Eugenio; & Charles, M. Judith (2001). A Reproducible Approach to the Reporting of Organochlorine Compounds in Epidemiologic Studies. Chemosphere, 44(6), 1395-402.
AbstractA growing body of research indicates that the most biologically active PCB congeners and organochlorines are not the most abundant components in human and wildlife samples. As researchers attempt measurement on a wider pool of less abundant compounds, they inevitably face quantification problems. To address this problem and enhance comparability across studies, we propose a standardized approach to report organochlorines that is based on a reproducible method to determine the limit of quantification (LQ). Two statistical methods are incorporated into our approach, one by Gibbons termed the Alternative Minimum Level (AML), and one based on determining a region of stable relativestandard deviation in instrument response (RSD). We illustrate our approach using historical samples collected during the 1960s from a cohort of pregnant women enrolled in the Child Health and Development Study. The results are applicable to determining the LQ of any method, and are of utmost importance to environmental scientists conducting trace organic analyses of complex mixtures. Our results demonstrate that: (1) precision as measured by RSD is the most important criterion in determining LQ; (2) the AML routinely isolates a region of constant RSD; and (3) the precision of the instrument detector response as measured with pure standards locates the LQ applicable for real samples - that is, the true limits of quantification reside in the detector, not the matrix effects or analyte recoveries associated with real samples. A corollary of these findings is that bias due to matrix effects and analyte recoveries can be assessed separately from precision and LQdetermination. Previous approaches involved spiking matrix blanks to determine LQ, a problematic strategy for real world, complex matrices. We have now validated the use of pure standards in LQ determination, an approach that is practical and accessible to most analysts.
Reference TypeJournal Article
Author(s)Willman, Eric J.
Keller, Jean A.
Charles, M. Judith