15 / 41
FOR EXPERT REVIEW PANEL USE ONLY
NOT FOR DISTRIBUTION
2
•
No fructan-dependent correction factor needed, the method avoids pre-analysis
•
Measurement in one run possible (max. 2 in case of low concentrations)
•
Results of dual lab validation is in good agreement
Cons/Weaknesses
•
Many enzymes needed, combined in a kit from Megazyme. If other enzymes are used, they need to be
tested individually for activity and side-activity
•
Less accuracy at the extremes of the desired working range
•
Two runs are needed if GOS/Polydextrose are present to correct for the blank
•
Post-column pump is needed, not commonly available
Supporting Data
•
General Comment:
The set-up of the methods was according to the SLV in both of the lab validations described.
-
Method Optimization:
The method is significantly optimized with respect to AOAC 997.03 and 999.03.
No fructan-dependent correction factor is needed. Fructan can be measured in one run.
•
Performance Characteristics:
Analytical Range:
NRC: 0.03-5.0 %
CCC: 0.03-5.0 %
LOQ:
NRC: 0.018%
CCC: 0.016-0.030%
Accuracy/Recovery:
NRC: 86-119% (After blank substraction)
(For 2 out of 6 matrices spiked at 0.03 g/100g recoveries were
above 110% (114% and 119%). For 1 matrix out of 6 spiked at 5g/100g recovery was below 90% (86%).
Excluding those 3 results, recoveries were in the range 92 – 105%)
CCC: 89-102% (one statistic outlier)
(Spike‐recoveries were between 91 – 102% for all samples except 1,
which had a spike recovery of 89%)
Precision (RSD
r
):
NRC: 0.78-3.2%
CCC: 1.1-8.9%
(for all samples except 1 was between 1.1 – 2.6 %, high RSD of 8.9 % at the lowest level (0.036
g/100g) only)
Reproducibility (RSD
R
):
Not tested in SLV
•
System suitability:
OK
Recommendation: