145 / 220
Dual Lab Validation of a Method for the Determination of Fructans in Infant
Formula & Adult Nutritionals /
25 Jul 2016
35
are reconstituted by dissolving 25 g of powder in 200 g of water. The SRM 1849a is weighted directly
or it was reconstituted by dissolving 10 g in 90 g of water.
In this validation study three different standards of pure fructan ingredients have been applied, being:
‐
Orafti HP, a long chain inulin ingredient
‐
Oraftie P95, an hydrolyzed inulin ingredient consisting of both GFn en Fm constituents
‐
Nutraflora P95, a short chain fructo‐oligosaccharide based on enzymatic sucrose elongation,
mainly constisting of GF
2
, GF
3
, and GF
4
constituents
The purity of these standards was established by analysis according the AOAC 997.08 method.
5.
Validation design
a.
Introduction
In the SMPR 2014.02 [4] the following demands for the FOS analyses are described:
Minimal requirements for the FOS determination
analytical range
0,03 – 5,0 %
LOQ
≤ 0,03%
Repeatability (%RSD
r
)
≤ 6%
Reproducibility (%RSD
R
)
≤ 12%
recovery
90 – 110%
The above mentioned requirements relate to reconstituted products.
The SLV is organized in accordance with “Appendix L: AOAC Recommended Guidelines for Stakeholder
Panel on Infant Formula and Adult Nutritionals (SPIFAN) Single‐Laboratory Validation” [5].
b.
Calibration fit
The SPIFAN SLV guidance [5] demands for the calibration fit a minimum of 6 concentration levels that
span the entire working range of the calibration. A minimum of 3 independent calibrations have to be
measured with independently prepared standards.
To meet these demands, 3 different, independently prepared, stock solutions of glucose and fructose
are prepared. Each stock solution is used to prepare an independent set of calibration standards of 12
different concentrations. Chitobiose is applied as an internal standard (IS). The 3 sets of 12 calibration
samples have analyzed in different HPAEC‐PAD system. The obtained data will be used to calculate
non‐linear (quadratic) calibration curves and to establish the dynamic range of the calibration curve.
The demand for a good calibration is that the calculated lack of fit for all data calibration points is less
than 5% with the exception of the lowest calibration concentration, which is allowed to be higher than
5%.
VALIDATION REPORT
FOR ERP USE ONLY
DO NOT DISTRIBUTE