614 / 670
average of 3 independent testing of the sample, and the 50% and 100% of the background
values of samples were used as the two spike levels in this study. As no VD2 was detected in
the selected samples, lower spike levels were used. Three replicates were done for all the
samples, and the recoveries of VD were calculated. The results in Table 7 showed that the
recoveries of VD2 and VD3 were all within 89-108% at the two spike levels.
Table 7. Contents and recoveries of VD in samples with different matrices
Sample
number
Content
(n=3),
ng/g
VD3
Content
(n=3),
ng/g
VD2
Spike level,
ng
Recovery
RSD
(n=3)
Spike level,
ng
Recovery
RSD
(n=3)
1
˄
powder
˅
82±1
50
101% 7.48%
ND
5.0
90% 2.93%
100
97% 8.18%
10
95% 7.82%
2
˄
powder
˅
93±2
50
99% 4.79%
ND
5.0
89% 2.96%
100
98% 7.36%
10
94% 3.42%
3
˄
powder
˅
91±1
50
108% 2.68%
ND
5.0
93% 3.02%
100
105% 0.95%
10
93% 0.91%
4
˄
liquid
˅
12.3±0.3
70
102% 4.82%
ND
3.5
97% 3.88%
140
102% 2.48%
7.0
91% 2.03%
Note
˖
Samples 1, 2, 3, and 4 are amino acid based, milk protein based, soy protein based, and adult nutritional
liquid (high fat) products, respectively; ND: no detection.
3.3 Limit of Detection (LOD) and Limit of Quantitation (LOQ)
3.3.1 Calculation method
A blank sample was prepared by the same procedure described in 2.2, and analyzed 10 times
by UPLC-MS-MS using the same parameters as samples. The mean value (x) and standard
deviation (SD) of the 10 testing results were calculated, and
the values of “x + 3SD” and “x +
10SD” were determined as the LOD and LOQ of the method, respectivel
y. In this validation
study, the “x” and “SD” values for VD3 determination were 1.34 and 0.83 pg/mL, respectively,
and the LOD and LOQ of this method for VD3 determination were calculated as 0.382 and 0.961
ng/100 g reconstituted/RTF samples, respectively.
For the VD2 determination, the “x” and “SD”
values were 0.782 and 0.734 pg/mL, respectively, and the LOD and LOQ of this method were
calculated as 0.299 and 0.813 ng/100 g reconstituted/RTF samples, respectively.
3.3.2 Dilution method
A blank sample was prepared by the same procedure described in 2.2, and analyzed by
UPLC-MS-MS using the same parameters as samples. The signal/noise ratio (S/N) was then
calculated, and the 3 and 10 times of the S/N were defined as the LOD and LOQ of the method,
respectively. By the dilution method, the 3 and 10 times of S/N for the VD3 determination were
0.0105 and 0.0343 ng/mL, respectively, corresponding to 1.05 (LOD) and 3.43 (LOQ) ng/100 g
reconstituted/RTF samples, respectively. The 3 and 10 times of S/N for the VD2 determination
were 0.0137 and 0.0458 ng/mL, respectively, corresponding to 1.37 (LOD) and 4.58 (LOQ)
ng/100 g reconstituted/RTF samples, respectively.
VitD-17 (February 2016)
FOR ERP USE ONLY
DO NOT DISTRIBUTE