207 / 363
S
alvati
et al
.:
J
ournal of
AOAC I
nternational
V
ol
.
99, N
o
.
3, 2016
9
Thiamine was a notable exception because of chromatographic
interference in the first transition for the
13
C
4
-thiamine internal
standard (Table 1). This chromatographic interference does not
impact method accuracy because the first transition is not used
for quantitation. However, it does mean that ion ratio suitability
criteria cannot be specified for the thiamine internal standard. Ion
ratios for the lower intensity vitamins (pyridoxal, pyridoxamine,
and nicotinic acid) had a larger degree of variation because of
the lower signal intensity. They averaged 102 ± 12% of the ion
ratio in the standards. During the over-spike studies in which
the signal intensity was higher, the variation in the ion ratio was
reduced and approached the ±3% level of the more abundant
vitamin forms.
Finally, the choice of enzyme is important for method
performance. During method development, two different acid
phosphatases were investigated, one from Roche Diagnostics
and one from Sigma-Aldrich. The acid phosphatase from
Roche did not fully hydrolyze pyridoxamine-5′-phosphate
and generally recovered about 50% of the over-spiked
level. Further, it generated significant amounts of nicotinic
acid during digestion on the order of up to 10% of the
total vitamin B
3
. Although the source of the nicotinic acid
is not entirely clear, it appears to result from conversion
of nicotinamide to nicotinic acid because the total B
3
concentration (sum of nicotinamide and nicotinic acid) did not
increase significantly in the three matrixes studied in detail.
The method was validated using the acid phosphatase from
Sigma-Aldrich. This acid phosphatase contains higher levels
of pyridoxamine and pyridoxal, riboflavin, and nicotinic acid;
but was chosen because it eliminates problems with nicotinic
acid conversion and pyridoxamine-5′-phosphate recovery.
The background vitamin levels in the Sigma-Aldrich acid
phosphatase as a percent of their concentrations in SRM
1849a are 0.1% thiamine, 2.8% riboflavin, 0.2% nicotinamide,
18% nicotinic acid, 6.2% pyridoxal, 0.5% pyridoxamine,
and 0.2% pyridoxine. However, these data need additional
context. Nicotinic acid, pyridoxal, and pyridoxamine in SRM
1849a are virtually absent. From a total vitamin perspective,
the overall contribution of vitamins from the enzyme in
SRM 1849a is 0.1% total B
1
, 2.8% total B
2
, 0.3% total B
3
,
and 0.5% total B
6
. Despite the small contribution from the
enzyme, the standards are prepared as samples to mitigate
any impact on method accuracy. The development work
presented serves as caution: substitution of enzymes for other
than those specified by this method may be deleterious to
method performance. The use of an alternative enzyme would
require significant investigation to the efficacy, digestion,
and background contribution of vitamins to ensure adequate
method performance.
Table 5. Intermediate precision for six independent preparations is expressed as %RSD
Matrix
Total B
1
Total B
2
Total B
3
Total B
6
S01: 1849a
3.0
4.7
4.5
4.0
S02: Infant formula powder partially hydrolyzed, milk-based
3.7
3.9
4.6
3.6
S03: Infant formula powder partially hydrolyzed, soy-based
2.5
4.5
2.3
3.3
S04: Toddler formula powder, milk-based
3.5
5.6
2.4
6.2
S05: Infant formula powder, milk-based
2.8
5.4
2.0
3.0
S06: Adult nutritional powder, low-fat
3.4
9.0
2.8
5.1
S07: Child formula powder
3.6
5.3
2.4
3.6
S08: Infant elemental powder
3.0
5.7
1.4
4.3
S09: Infant formula powder, FOS/GOS-based
a
4.0
4.1
2.3
4.1
S10: Infant formula powder, milk-based
4.1
4.5
1.6
4.5
S11: Infant formula powder, soy-based
3.1
3.7
2.2
3.5
S12: Infant formula RTF, milk-based
4.0
5.8
3.3
5.7
S13: Adult nutritional RTF, high-protein
4.1
5.6
3.6
6.0
S14: Adult nutritional RTF, high-fat
3.2
4.2
3.2
5.0
a
FOS/GOS = Fructo-oligosaccharides/galacto-oligosaccharides.
Figure 1. A chromatogram for the seven vitamin forms in the child
formula powder. The data are unsmoothed, and the intensity of
each peak is normalized to aid visualization. There is more than two
orders of magnitude difference in signal intensity, which makes the
small features such as pyridoxamine and pyridoxal difficult to see
when all are plotted on the same scale.
Candidates for 2016 Method of the Year
206