19 / 258
G
ill
et al
.:
J
ournal of
AOAC I
nternational
V
ol
.
99, N
o
.
5, 2016
1329
Bias was evaluated by replicate analyses of the National
Institute of Standards and Technology (NIST) 1849a Standard
Reference Material (SRM). Differences between the measured
value and the certified value were determined with the mean
and SD of the differences, and the test statistic was calculated.
A
p
(α = 0.05)
of 0.25 indicates that there was no bias between the
measured results and the certified value (Table 5). As part of
initial method validation, the LC-MS/MS was evaluated for bias
against an HPLC–UV method based on AOAC
2002.05
(8, 9).
A
p
(α = 0.05)
of 0.09 indicates that there was no bias between the
methods (Table 6). Bias against a certified reference material
or a reference method is not a defined parameter within the
SMPR.
Vitamin D–Previtamin D Interconversion
Although the described method specifically detects vitamin D
and not the previtamin D isomer, the method quantifies an
aggregate result for both previtamin D and vitamin D. This
satisfies the requirement of the applicability statement of the
SMPR, which specifies total vitamin D
2
or vitamin D
3
, including
their previtamin isomers. It was assumed in this analysis, as with
all analytical methods for vitamin D that use calciferol internal
standards, that the previtamin D:vitamin D ratio was equivalent
for the sample analyte and the internal standard. For deuterated
internal standards, the labeled site must be remote from the triene
center because of the difference in interconversion behavior
between the analyte and the internal standard (10). To confirm
this assumption, the effect of temperature on the final results
was evaluated. Experiments were performed with saponification
assessed in three different ways: (
1
) at 70°C for 1 h, according
to the described method protocol; (
2
) at 20°C for 7.5 h; and (
3
)
at 70°C for 7.5 h. A 7.5 h saponification was chosen because this
is the time needed, as previously reported, for a pure solution of
vitamin D to reach equilibrium with previtamin D at 70°C (11).
Samples 1–6 and 13–18, which were saponified at 70°C, showed
significantly lower absolute peak areas for the vitamin D–PTAD
quantifier ion than samples 7–12, which were saponified at 20°C.
This was as expected because a higher proportion of vitamin D is
converted to previtamin D at the elevated temperature. This effect
was seen for both the analyte vitamin D in the sample and the SIL
d6
-vitamin D internal standard, illustrating the appropriateness
of the internal standard to account for any temperature-induced
interconversion between previtamin D and vitamin D (Figure 5).
The final results obtained showed that, within sample error, there
Table 5. Results for the bias experiment against NIST
1849a SRM
a
Parameter
Value
Certified value, μg/hg
11.1
Uncertainty, μg/hg
1.7
Certified range, μg/hg
9.4–12.8
Coverage factor,
k
2
Degrees of freedom, DF
CRV
60
Mean,
x
10.1
SD
0.53
Number of replicates,
n
13
95% Confidence interval, μg/hg
9.8–10.4
T
stat
1.165
Degrees of freedom
63.92
p
(α = 0.05)
0.25
a
SRM=Standard Reference Material.
Table 6. Results for the bias experiment against
AOAC 2002.05
Parameter
Reference method LC-MS/MS method
Mean, μg/hg
10.5
10.8
SD, μg/hg
3.18
3.66
Number of replicates,
n
40
40
95% Confidence interval, μg/hg 10.0–11.0
10.2–11.4
Mean of paired differences
–0.3
SD of paired differences
1.27
T
stat
1.73
Degrees of freedom
38
p
(α = 0.05)
0.09
Figure 5. Effect of saponification time/temperature on vitamin D and
d6
-vitamin D.
19