Microsoft Word - Candidates for 2017 ERP of the Year

AOAC ERP Report Attachment 1

Address to review panel’s questions

1. “Repeating the spiking study without stripping the matrices.” All data in this report were generated by using matrices without stripping, including both LOD/LOQ and spiking study. 2. “Characterize in-house matrix based reference materials (Guidance in OMA)” Determination of BPA in blank samples and testing samples was performed following the instruction from AOAC Technical Division on Reference Materials (TDRM) and ISO/Guide 80:2014(en): Guidance for the in-house preparation of quality control materials, including homogeneity, stability, preparation and analysis. 3. “LOD study with sample that has low levels of BPA already in it is acceptable.” Following instructions from the review panel, we selected the least BPA contaminated samples (without stripping) to identify the LOD an LOQ. A total of 10 replicates were analyzed for each sample type, and the mean and SD were calculated. Since the blank samples were contaminated with BPA, the blank mean reflect the actual contamination levels of the samples, not noise generated by the sample matrix or instruments. The blank mean was subtracted for the LOD and LOQ calculation. 4. “Units must be stated” All units are stated. 5. “Brand names removed” Brand names are removed. 6. “System suitability” According to email communication with AOAC review panel, “The ERP would like to see some guidance for the user related to understanding and demonstrating that the system is working as expected. This often includes analysis of a standard sample (such as an in-house reference material as described above) and/or a calibrant, with tolerance limits for peak area, resolution, retention time, linearity, etc. in addition to blank checks. As well, the SMPR requires negative controls (blanks) and low and mid-point positive controls”. In this report, standard samples were tested every day the test was performed to generate a standard curve. Negative controls (blanks) and low, mid-point and high positive controls (for example, samples spiked at 0, 1, 4, 16 and 32 ng/mL BPA levels) were tested in three different