581 / 696
Page
11
of
39
This equation is the product of two individual factors. The first accounts for the terminal glycosyl loss, either due to the reduction
of the reducing fructosyl moiety of F
n
type fructans or the failure to determine glucose. The second corrects for the water of
hydrolysis. So for example, factor #1 in Table 2-6 is based on the assumption of a fructan with an average degree of
polymerization (DP) of 4.
1.233 =
ሺ
ସ ସିଵ
ሻ
כ
ሺͲǤͻ
Ǥଵ ସ
ሻ
The factors chosen for the three groups listed in Table 2-6 were done in such a way so as to cover the known spectrum of fructan
commodities currently in use. Any factor should then impart a bias of no more than ~5% for the grouping that it covers.
Validation Protocol
1.
Linearity
–
A total of 43 standard curves were collected. Linearity Data analyzed in Microsoft Excel.
2.
Accuracy
–
Triplicate analyses of spiked samples on each of three days. Data analyzed in Microsoft Excel.
a.
Additional testing on SPIFAN matrices of three days in duplicate also performed. Data also analyzed in
Microsoft Excel.
3.
Precision
–
Data was analyzed in Microsoft Excel and Minitab.
a.
Duplicate analyses of products on each of 10 days. Calculated run-to-run %RSD, within-run %RSD, and total
(intermediate) %RSD in 2 separate labs. Additionally, a total of 80 analyses of Abbott internal control were
collected in a total of 3 labs (60 points collected as 10 days of duplicate analyses in three labs, the remaining 20
were single determinations at Abbott Columbus). This data is also included for the sake of
b.
Performed 6 days of duplicate testing on SPIFAN matrices (for the 6 products fortified with fructan material).
Calculated run-to-run %RSD, within-run %RSD, and total (intermediate) %RSD.
4.
Specificity
–
Specificity was demonstrated by analysis of spiked samples containing GOS (known amount of fructan
spiked in). Additionally checked for other chromatographic interferences by analysis of multiple maltodextrin
commodities and multiple products containing either hydrolysate protein or free amino acids. Lastly it was checked by
the analysis of SPIFAN matrices known to not be fortified with fructan material.
5.
MDL/LOQ
–
The LOQ was empirically demonstrated by analysis of a low level spike solution.
Results and Discussion
Method Development
The method described here is based on methods that were previously published. The principle differences are summarized in Table
2-7.
Fos-04 (February 2016)
FOR ERP USE ONLY
DO NOT DISTRIBUTE