Comparability of world-wide analytical data of PCDDs, PCDFs and non-ortho PCBs in samples of chicken, butter and salmon

It was the primary objective of the present study to elucidate, from a chemical analytical point of view, the comparability of dioxin determinations and the consequences concerning the data on levels in food reported world-wide. To do so, an international intercalibration study on the determination of polychlorinated dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs) and dioxin-like PCBs in three natural food items was performed between March and June 2000. Samples of chicken, butter and salmon, collected in Norway, where sent to 42 laboratories and analysed for 20 PCDDs/PCDFs/non-ortho PCBs. The study generated congener-specific data for the three food matrixes both on wet weight and lipid basis. The reported data from 37 laboratories, in 15 countries, were statistically analysed in order to determine the congener-specific comparability as well as the comparability based on a toxic equivalency value (TEQ). Based on the reported data the mean relative standard deviations (mean RSD) for the WHO98-TEQDFP measurements were found to range from 36% to 57% the three foods. For chicken, the consensus level was 1.2 ppt TEQ on lipid basis, and the reported levels by 24 laboratories were between 0.53 and 2.2 ppt. For butter, the consensus was 0.55 ppt TEQ. 27 laboratories reported levels from 0.18 up to 1.5 ppt. For the salmon, sample levels of 5.9 to 38 ppt TEQ (by 25 laboratories) was reported where 15 ppt was the consensus value. It can thus be concluded that the world-wide data on dioxins, at background levels in foods, are not direct numerically comparable due to inherent analytical reasons. Up to 50% deviations are expected when interlaboratory dioxin results for foods are compared.