139 / 258
Butler-Thompson et al.:
J
ournal of
AOAC I
nternational
V
ol.
98, N
o.
6, 2015
1661
on concentration versus the average peak area of the working
standards. A standard curve must have a correlation coefficient
(r) of at least 0.999 to be considered acceptable for sample
calculations. (
2
) At each working standard concentration, the
peak areas of standards injected at the beginning and end of a set
of samples should not increase or decrease by more than 10%.
(e)
Calculation of vitamin B
12
concentrations in samples
.—
The vitamin B
12
concentration in each injected sample
preparation is extrapolated from the vitamin B
12
standard curve
prepared as described above. The concentration of vitamin B
12
in each product can then be calculated:
C
p
= C
i
× D
1
÷ ss × D
2
÷ V
where C
p
= product concentration in μg/kg; C
i
= vitamin B
12
concentration of the injected sample preparation extrapolated
from standard curve in μg/L; D
1
= volume of the first dilution
in mL (100 mL); ss = sample size in g; D
2
= volume of the
second (final) dilution in mL; and V = volume of filtrate
loaded onto the cartridge in mL (2, 6).
Results and Discussion
Data from all participating laboratories are summarized in
Tables 1 and 2. It should be noted that two of the participating
laboratories (Laboratories 6 and 11) only received half of
the study samples. It should also be noted that data for the
second adult high-fat RTF replicate from four laboratories
(Laboratories 1, 7, 9, and 11) had to be excluded because
of sample mislabeling. Based on chromatographic profiles
and physical appearance of the products, it was determined
that some of the adult high-protein RTF samples were
mislabeled as adult high-fat RTF samples. Chromatograms
of the adult high-protein RTF samples, the adult high-fat
RTF sample, and the mislabeled adult high-fat RTF sample
are shown in Figures 1–4, illustrating the chromatographic
differences between the two products. Although only five
of the nine participating laboratories had duplicate results
for the adult high-fat RTF because of mislabeling, AOAC
recommended keeping the adult high-fat product in the study
and calculating repeatability and reproducibility from the data
that were available. As a result, repeatability was calculated
from duplicate results generated at five laboratories, and
reproducibility was calculated from the single and duplicate
results from all nine laboratories.
Using the AOAC INTERNATIONAL Interlaboratory Study
Blind (Unpaired) Replicates workbook, statistical outliers
from one laboratory were identified for six of the 12 products
analyzed. After removal of outliers, repeatability RSD
r
ranged
from 2.98 to 9.77% and met the SMPR ≤7% for eight of the
12 products analyzed. The RSD
R
ranged from 3.54 to 19.5%
and met the SMPR of ≤11% for seven of the 12 products
analyzed. Although repeatability and reproducibility for several
products exceeded the requirements in SMPR 2011.005, it
was the majority opinion of the ERP that these results were
Table 2011.10F. RP column elution gradient
Mobile phase, %
Time, min
A
B
C
0.00
90
10
0
14.5
90
10
0
14.6
40–60
a
60–40
a
0
27.0–30
40–60
a
60–40
a
0
27.1–30.1
0
10
90
29.90–33.00
0
10
90
a
Appropriate gradient conditions must be established with each column to adequately resolve vitamin B
12
and riboflavin and to elute vitamin B
12
between approximately 24 and 30 min. To establish appropriate gradient conditions with a new column, set the gradient composition at 14.6 and
27.0–30.0 min to the midpoint of the allowable range from the table above. Inject the resolution test solution and calculate the resolution (Rs) between
vitamin B
12
and riboflavin. Adjust the mobile phase composition between 14.6 and 27.0–30.0 min until Rs is >1.5. After vitamin B
12
elutes from the C
18
or phenyl column rinse the column with 90% mobile phase C for at least 2.8 min.
Figure 2011.10C. Example chromatogram showing riboflavin and vitamin B
12
resolution.
139