16 - Overall conclusions of MLT on fatty acids
The majority of results provided for individuals and groups of fatty acids in all samples (
i.e.
dairy,
infant formula and adult nutritional products) were in agreement with expectations (
i.e.
results
gained over the years, proficiency tests, SLV) and limits fixed in SMPR.
Results provided for some fatty acids (who can contribute also to sums) were significantly
influenced by chromatography or analyst (
i.e.
response factors of instrument, co-elution, wrong
peak identification or integration but also error in the reporting).
The instruction to identify and report some fatty acids (
i.e.
TFA) were not carefully applied by some
participants and the real method capability for TFA is impacted. TFA results were seen above the
limits fixed in SMPR.
The removal of outliers in ISO-5725 is based principally on Cochran & Grubbs and considerations
given above (
i.e.
instructions given not followed) were not taken into account in the current
statistical evaluation. The labs affected by the problem can be also identified by Mandel’s H & K.
Once again for this kind of method the chromatography performances is a key parameter, the
results obtained for the majority of labs showed that optimal performances can be achieve with
some efforts.
