HBM4EU ICI report UV-filters in urine round 1

Within the framework of the HBM4EU project, an interlaboratory comparison was organized and conducted for the analysis of benzophenones (BP) in urine. Benzophenones correspond to 4 biomarkers: 2,4-Dihydroxybenzophenone (BP1), 2,2’,4,4’-Tetrahydroxybenzophenone (BP2), 2-Hydroxy-4-methoxybenzophenone (BP3), 5-Chloro-2-hydroxybenzophenone (BP7). The study was performed from January to February 2020. The HBM4EU QAU had selected three expert laboratories for benzophenones in urine. All expert laboratories were from three different countries in Europe. The participation in this interlaboratory comparison for benzophenones in urine was mandatory for these laboratories. Two different test samples consisting of 5 mL urine mixed from burdened native material to obtain two different concentrations (BPR1A, BPR1B) were prepared and sent to the participating expert laboratories for single analysis. Homogeneity assessment of the control materials confirmed that the materials were adequately homogeneous. Due to technical problems, stability of the control material could not be assessed in time. As it is essential that stability testing be conducted in the same laboratory as homogeneity testing, for this 1st round, no stability could be evaluated. For the next rounds, however, stability will be tested and provided in the respective reports. Consensus values were calculated as the arithmetic mean of the values obtained by the expert labs, provided that the relative uncertainty of the mean was within 17.5%. In order to express the proficiency of the laboratories in a numerical way, Z-scores were calculated using the consensus value and a fixed fit-for-purpose relative target standard deviation (FFP-RSDR) of 25%. For BP7R1A and BP7R1B, consensus values could be calculated from the results of all three experts, the obtained Z-scores were all satisfactory and the relative standard deviation (RSD) was in a range from 7.5% (BP7R1A) to 25.1% (BP7R1B). In case there were only two results for a parameter, the two results were considered comparable if the difference to the mean was ≤35%. Then no Z-scores were calculated. For BP1R1B, BP2R1A and BP2R1B, no Z-scores could be provided, because the relative uncertainty (u) of the mean concentration from three expert labs was too high and no statistical outlier could be detected. However, the results of two expert laboratories (UEL1 and UEL5) showed a difference from the mean of 1% (BP1R1B), 34% (BP2R1A) and 40% (BP2R1B) and were thus in a good or medium comparable range. For BP1R1A, BP3R1A and BP3R1B, only two quantitative results were reported (UEL1 and UEL5). These results showed a difference from the mean of 10% (BP1R1A), 11% (BP3R1A) and 2% (BP3R1B) and were thus in a good comparable range. The final evaluation of the comparability of the respective expert laboratories can, however, only take place upon completion of all interlaboratory comparison rounds.