30 / 258
H
aselberger
&
J
acobs
:
J
ournal of
AOAC I
nternational
V
ol
.
99, N
o
. 6, 2016
1585
10
H
aselberger & Jacobs
: J
ourna
of
AOAC I
nternati nal
V l. 99, No. 6, 2016
the validation work, spanning a period of 10 months. Relative
calibration errors for each standard were calculated and used
as the basis for determining adequacy of fit to the quadratic
calibration model.
Accuracy
Because there are no relevant certified reference materials,
assessment of accuracy was based entirely on recovery of
fructans from spiked samples. The samples used for these
experiments were products included in the Stakeholder Panel on
Infant Formula and Adult Nutritionals (SPIFAN) Sample Kit.
Of these, six were determined to contain fructan in the form of
either scFOS or oligofructose (group 1 fructan, for purposes of
CF assignment). These six samples were overspiked with either
scFOS or oligofructose at ~50 and 100% of the endogenous
level. Two of the unfortified SPIFAN products were spiked at
two levels with either inulin (HP) or an oligofructose/inulin
mixture (Synergy 1). SPIFAN sample overspike and spike
configurations are described in Tables
2016.06B
and
C
. Both of
these sample sets were analyzed in duplicate on each of 3 days.
In addition to the SPIFAN matrixes, in-house low-lactose, milk-
based infant formula containing GOS was spiked at two levels
with three different fructans. Table 1 provides spike information
for the latter.
Precision
The six fructan-containing SPIFAN samples were analyzed in
duplicate on each of 6 days. In addition to the SPIFAN products,
four Abbott in-house materials were also analyzed. Three of
those were analyzed in duplicate on each of 10 days, by each
of two independent laboratories. A fourth in-house (control)
sample was analyzed in duplicate on each of 10 days by three
independent laboratories. Repeatability and intermediate
precision metrics were calculated for each of the 10 samples for
which replicate data was collected.
Specificity
Specificitywas evaluated primarily by analysis of products not
fortified with fructans, in order to determine potential apparent
background (Table 2). Each of these products was analyzed in
duplicate. In addition, chromatograms of multiple maltodextrin
commodities and products containing either hydrolysate protein
or free amino acids were examined for potential interferences
Figure 2. Relative calibration errors (composite from 43 analytical runs).
4
3
2
1
4.00%
3.00%
2.00%
1.00%
0.00%
-1.00%
-2.00%
-3.00%
-4.00%
Standard Level
Relative Calibration Error
Table 2. SPIFAN kit products without added fructan -
(analyzed for specificity evaluation)
Sample type, batch
Total
fructan found
a
Infant elemental powder
ND
b
Adult nutritional RTF high protein
ND
Adult nutritional RTF high fat
ND
Infant formula RTF milk based
ND
SRM 1849a
ND
Infant formula powder partially hydrolyzed milk based
ND
Infant elemental powder
ND
Infant formula powder milk based
ND
Infant formula powder soy based
ND
Infant formula RTF milk based
ND
Adult nutritional RTF high protein
ND
a
Mean fructan level detected in Part II of method analysis.
b
ND = Not detected; No fructose peak could be identified. Detection limit
was not experimentally determined but it is estimated that apparent
fructan in these samples is consistently <0.01%.
30