ABSTRACT
Under the sponsorship of the World Health Organization (WHO), an interlaboratory calibration on the analysis of PCDD/PCDFs in human milk and blood was carried out which included 19 laboratories from 14 countries. The study design involved the analysis of three samples of each matrix in triplicate. Selected samples were spiked with native standards of certain 2,3,7,8-substituted congeners at concentrations known only to WHO staff. The study design resulted in approximately 4000 individual pieces of PCDD/PCDF data generated by a variety of analytical methods, at various concentrations, and by laboratories of widely different experience. This was, by considerable margin, the largest study which allowed for the direct comparison of laboratory and method performance. The results of statistical analysis of this data base addresses the effect on data quality of clean up methods, instrumental methods, analyte concentration, laboratory QA programs, and laboratory experience. The study has shown that the laboratory is the single most important determinant of data precision and accuracy. The method of analyte enrichment (sample clean up), analyte measurement [gas chromatography/mass spectrometry (GC/MS) protocol], and analyte concentration have weaker correlations with data quality.
