152 / 220
Dual Lab Validation of a Method for the Determination of Fructans in Infant
Formula & Adult Nutritionals /
25 Jul 2016
42
Table 5 Summarizing results of the repeatability and reproducibility
investigations in the reconstituted samples.
sample
no
sample
description
number
of
analyses
average
content
% m/m
SD
r
% m/m
RSD
r
%
meet
target
SD
R
% m/m
RSD
R
%
meet
target
1
Child formula powder
5 x 2 + 1 x 1
0,287
0,004
1,4
yes
0,021
7,4
yes
9
Toddler formula
powder milk based
6 x 2
0,223
0,003
1,6
yes
0,013
6,3
yes
10
infant formula powder
milk based
6 x 2
0,263
0,003
1,1
yes
0,017
6,5
yes
12
child formula powder
6 x 2
0,282
0,003
1,1
yes
0,016
6,3
yes
14
infant formula powder
FOS/GOS based
6 x 2
0,036
0,004
8,9
no
0,005
13,8
no
19
Adult nutritional RTF
high fat
6 x 2
0,507
0,013
2,6
yes
0,021
4,6
yes
As can been seen in the experimental data in Annex C, there was an outlier value measured in the
duplicate at day 1 in sample 1. This outlier was detected with Grubb’s outlier test. In the summarizing
statistical evaluation in table 5, this outlying value was rejected. Therefore the number analyses in
sample 1 is just 5 x 2 + 1 x1 instead of 6 x 2. This means that the SD
r
and RSD
r
are based on 5
duplicates instead of 6 duplicates. And the SD
R
and RSD
R
are based on 11 values instead of 12
measured fructan concentrations.
The RSD
r
and RSD
R
of sample no 14 “infant formula powder FOS/GOS based” don’t meet the target
values of respectively of 6 and 12%. However this is not surprisingly because the fructan content in
this sample is very low, in the range of the LOQ of the method. And for that reason the RSD
r
and RSD
R
are somewhat higher than the target values in the SMPR 2014.02.
e.
Accuracy
The same 6 FOS containing samples (after reconstition) in the SPIFAN sample kit as selected for the
precision study have been applied for the accuracy study. The average fructan content in these
samples was already determined in the precision study. Those date are used in the accuracy study as
the fructan content before spiking the samples with a pure fructan ingredient.
The established purity of the fructan ingredients Orafti HP, Orafti P95 and Nutraflora P95, applied for
the standard additions, are given already in table 3.
The addition solutions are prepared by dissolving about 1.1 gram (weighted with an accuracy of 1 mg)
in 100 ml water. For all samples except sample 14, the same amounts of fructan have been added in
the addition experiments. The fructan content in sample 14 (infant formula powder FOS/GOS based)
was very low, near the LOQ value. Therefore in case of this sample a much lower addition level was
applied.
For the samples 1, 9, 10, 12 and 19 the addition at low level a volume of 0.70 ml was added to 4.00 ml
of the reconstituted sample and for addition at the high level a volume of 2.00 ml was added to 4.00
ml of the reconstituted sample. For sample 14 the added volume of the standard fructan solutions was
VALIDATION REPORT
FOR ERP USE ONLY
DO NOT DISTRIBUTE