AOAC SPIFAN ERP Reviewer Forms (March 16, 2017)

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Supporting Data • General Comment:

The final results for all samples from all participants (except three labs) are presented. The reason to exclude the three labs is explained. The results meet the SMPR (2011.004) requirement on repeatability, reproducibility, and accuracy (using NIST 1849a). The LOD/LOQ data is not included. However, one of the tested samples has vitamin D3 concentration (0.7 ug/hg in sample, which corresponds to 0.08 ug/hg in RTF sample) below the SMPR LOQ. Feedback from the collaborators is included, which shows no critical negative comment.

• Performance Characteristics:

Analytical Range: MLV results covers 0.7 to 10.2 ug/hg sample (powder), which corresponds to 0.08 – 1.1ug/hg of RTF products; The SMPR requirement is 0.12-5.1 ug/hg; LOQ: not data on LOQ, but one of the samples has vitamin D3 at 0.08 ug/hg RTF sample, lower than the SMPR LOQ of 012 ug/hg.

Accuracy/Recovery: Meet SMPR (10.2 vs 11.1+-1.7 ug/hg)

): Meet SMPR (1.9 – 5.8%, less than 11%)

Precision (RSD r

Reproducibility (RSD R

): meet SMPR (6.4 – 12.7% less than 15%)

• System suitability:

- There is no system suitability specified in the method.

Recommendation:

Based on the MLT results, the method performance meet the SMPR. I would recommend this method to be the final action method. However, this method does not detect previtamins, and relies on the assumption that the previtamin isomers’ levels are low and will not significantly affect the overall accuracy and precision. The MLT data seems supporting this assessment with excellent reproducibility from 9 labs on 15 samples. But it would be precautious to detect the previtamins and ensure their contents are below certain level to prevent large error under some rare circumstances.

2