295 / 620
Discussion
Twelve samples were analyzed by thirteen independent laboratories over the course of a
year during two independent MLT collaborations. These laboratories were from
industry, contract research organizations and government institutions. Laboratories were
located in North America, Europe, and Asia.
The initial MLT collaborative results were presented to the AOAC ERP in July of 2014.
This collaboration was conducted using six of the twelve original SPIFAN test materials
in which ten laboratories participated. Upon review of the results by the ERP, the Panel
asked for additional information. As the NIST 1849a was not completely blind and the
reproducibility of three of the six test materials, the Indian vitamin D
2
infant formula
powder, and the partially hydrolyzed milk based infant formula powder and the milk
protein based adult nutritional powder, did not meet the Standard Method Performance
Requirements (SMPR) guidelines, AOAC 2011.004. It was recommended that the MLT
be conducted again using the new SPIFAN test materials, where applicable, and to re-
analyze the test materials that did not meet the SMPR guidelines.
The second MLT collaboration was conducted using five of the new SPIFAN test
materials and five of the original SPIFAN test materials. The new SPIFAN test materials
used in this collaborative were the partially hydrolyzed milk based infant formula
powder, the milk based infant formula powder, the soy based infant formula powder, the
high protein adult nutritional RTF and the high fat adult nutritional RTF. The results of
the second MLT collaboration were presented to the ERP in March of 2015. Upon
review of the results by the ERP, the Panel recommended some additional actions. It was
recommended by the reviewers of the final report that the data for Laboratory A be added
to the data tables and that the method platform be clarified to only state the use of
UHPLC-MS/MS. Initially, Laboratory A’s results were removed as their results for the
SRM NIST 1849a were well below the established range and therefore their data was
suspected to be biased low. However, this was not sufficient justification to reject the
data and therefore the data was not removed and is presented in this report. It was also
recommended that the data from the initial MLT collaboration for the partially
hydrolyzed soy based infant formula and the milk based RTF infant formula not be
included in the final report as it was confusing to determine what data was generated and
when.
For this third submission of the data, several changes have been made to the final report.
As all laboratories utilized UHPLC instrumentation, the use of the HPLC platform for
low level vitamin D
2
samples was removed from the method. Laboratory A’s results
were added to the data tables and used in all statistical calculations. Data from two test
materials from the initial MLT collaboration were included in this report at the request of
the ERP (Tables 1.3 and 1.4). These data are report on a sample basis – not reconstituted
in SPIFAN units. All data from the March 2015 MLT are reported in SPIFAN units.
All of the laboratories’ results are presented in Table 1.1 and Table 1.2. Table 1.5 and
Table 1.6 show the statistical evaluations for all the samples analyzed in the MLT study
from July 2014 and March 2015. The HorRat values for all results were excellent,
ranging from 0.27 – 0.52. The RSD
r
ranged from 2.55 – 13.5% and the RSD
R
ranged
from 8.89 – 16.2%. Of the twelve samples, results from eight of the samples met all of
the SMPR guidelines, while four samples fell slightly out of range. The adult nutritional
2011.11 VITAMIN D MLT-Sept 2015
FOR ERP USE ONLY
DO NOT DISTRIBUTE