Article révisé par les pairs
Résumé : A cooperative study was conducted under the auspices of the IUPAC Commission on Toxicology to measure total concentrations of selenium in three lots of commercially available freeze-dried human whole blood (Seronorm, NycoMed A/S, Norway). A serum material already examined in two previous studies was simultaneously distributed to assess accuracy of the analytical procedures better. A total of 39 laboratories from 15 countries participated in the trial using seven inherently different methods: acid decomposition-fluorimetry (ADF); electrothermal atomic absorption spectrometry (EAAS); acid digestion-hydride generation/atomic absorption spectrometry (ADHG-AAS); proton-induced X-ray emission (PIXE); instrumental neutron activation analysis (INAA) or instrumental activation analysis with radiochemical separation (RNAA); energy dispersive-X-ray-fluorescence spectrometry (EDXRF) and acid digestion-isotope-dilution mass spectrometry (ADIDMS). The performance of the different analytic methods was critically assessed. For serum, all laboratories using PIXE, RNAA and EDXRF had to be excluded in the final estimation as well as two of the three laboratories using INAA. For whole blood too, laboratories applying PIXE and RNAA had to be excluded, but performances for INAA were acceptable. Better agreement and satisfactory variance components were found between other methods and the exclusion rate was considerably lower. The following mean concentrations [± 68 % confidence intervals for one future observation] were established for whole blood: Batch 904, 96.4 ± 4.4 μg/L; Batch 905, 97.5 ±5.1 pg/L; Batch 906, 96.0 ± 4.4 μg/L. The concentrations in these materials were similar to those found in population studies in most European countries. The material is suitable as quality assurance material for the measurement of typical concentrations of selenium in blood. © 1992 IUPAC