Report of the interlaboratory comparison organised by the EU reference for laboratory food contact material : ILC02 2009- Bisphenol A in 50% aqueous ethanol (milk simulant) : laboratory performance and precision criteria of a harmonised method

The Institute for Health and Consumer Protection (IHCP) of the European Commission's Directorate-General Joint Research Centre hosts the Community Reference Laboratory for Food Contact Materials (EURL-FCM). One of its core tasks is to organize interlaboratory comparisons (ILCs) among appointed National Reference Laboratories (NRLs). This report presents the results of the second ILC exercise of the EURL-FCM for the year 2009, which focused on the determination of Bisphenol A (BPA) in 50% aqueous ethanol as a food simulant for milk. The general aim was to develop and perform the validation of a method for the analysis of Bisphenol A as model substance for a polycarbonate (PC), since it is a material that has typically been used as baby bottles and therefore typically in contact with mostly milk-type products. The strategy rose from the implementation of the new milk simulant 50% Ethanol (EtOH) in Commission Directive 2007/19/EC that current CEN standards for specific migration have not addressed yet. The first specific intention of the exercise on BPA was to validate an extension of scope of EN13130 Part 13 to be extended to include this new simulant 50% EtOH, i.e. the method for the quantification of BPA in 50% EtOH in the range around the legislative specific migration limit (SML, 0.6 mg/kg). In addition, rather than just validating the method at levels close to the SML as required for compliance purposes, it was agreed that a second validation range would also be studied to allow validation data to be generated for exposure purposes. The test materials used in this exercise were spiked samples prepared by EURL-FCM with several level of Bisphenol A (Sigma Aldrich) in 50% ethanol. In total four 50% ethanol solutions containing different concentrations of Bisphenol A were provided for analysis encompassing concentrations of relevance to exposure determination and compliance determination. The homogeneity and stability studies were performed by the EURL-FCM laboratory. Standard operating procedures (SOPs) for the two approaches were also written. The spiked samples and SOP(s) were sent in August and the deadline for the submission of results was mid-October. Participation of local laboratories under NRLs was encouraged (by producing 60 samples). There were 31 participants from twenty-five countries to whom samples were dispatched and 26 of which submitted results. From the EURL-NRL network 18 laboratories out of 24 reported results. There were 3 guests from Spain and 3 from Germany that provided results as well. Participants were invited to report four replicates measurements in repeatability conditions. This was done by most of the participants. The ILC was closed permanently in the end of October for statistical interpretation. The results of analyses were received and statistically interpreted. The assigned values were obtained as a consensus values after applying the robust statistics to the results obtained from the participants. Laboratory results were rated with z-scores in accordance with ISO 13528 [1]. Standard deviations for ILC comparison (also called target standard deviations) were set based on Horwitz equation and Horrat ratio 0.5. The results and preliminary report were discussed in the plenary of December 2009. The participation of the laboratories was regarded as satisfactory for the aim of the precision experiment with regards of the numbers of received results. Absolute minimum of participating laboratories for conducting a precision experiment is 8. Some of the NRLs communicated the lack of the possibility to follow exactly the method for BPA determination as they used different analytical technique (e.g. LC-MS instead of HPLC/FLD). Since the aim of the ILC was directed towards method validation there was a communication to the participants that different techniques were not fit to the scope of this ILC. That was the reason for slightly lower percentage of reported test results as compared with the other ILC exercise of 2009. As a conclusion of the precision exercise on the quantification of Bisphenol A in the new milk simulant 50% ethanol, this ILC showed that : the participation in the ILC was satisfactory for the purpose of the study towards validation with 24 laboratories. The validation of the method based on HPLC-FLD according to the description based mostly on the previous CEN standard TS 13130-13 was successful with more than 8 valid results thanks to the proactive involvement of the NRLs. The precision that can be suggested were of 15% reproducibility SD and 6% repeatability SD for the 0.0067 mg/kg level, 10% reproducibility SD and 4% repeatability SD for the 0.0.21 mg/kg level, 6% reproducibility SD and 2% repeatability SD for the 0.0.75 mg/kg level, 6% reproducibility SD and 0.2% repeatability SD for the 0.56 mg/kg level. With respect to the scarcity of data previously available in the validation performed as reported in the CEN standard TS 13130-13 (issued version of 2005), this validation also provides a great breadth of valuable detailed and traceable raw data, which should prove extremely relevant for the creation of an extension of the standard from CEN.