A reproducible approach to the reporting of organochlorine compounds in epidemiologic studies

A 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 (L _Q). 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 L _Q 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 L _Q; (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 L _Q 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 L _Q determination. Previous approaches involved spiking matrix blanks to determine L _Q, a problematic strategy for real world, complex matrices. We have now validated the use of pure standards in L _Q determination, an approach that is practical and accessible to most analysts.