We do have someone on this board who also performs analysis on cheese in the EU, so you are in luck. They may be available shortly for more specific help. However, whenever you make any change to a method, you can't just say it makes no difference once, you must validate the new method. Basically, you must formulate an experimental plan, try many known samples -- some standards that are traceable back to another method, possibly even controlled by a regulatory body, and also some known and spiked samples -- perhaps 3 or more separate production lots, and have the statistics on the data to support that the new method gives identical results. This should all be documented, and available for a regulatory body to review at their request.
The sudden result giving double values is very suspicious for me. Perhaps you have made other changes, that even you are not aware of. Small, minor changes in a method should only produce slight changes. A quick review of this method shows many wet chemical steps before the GC analysis, so it might be hard to pin down where the method might go wrong. Logically, failure to silanize tubes should produce less cholesterol signal, because it does tend to bind to plastic tubes. But analytical chemists often don't have the luxury of making determinations on what the "figure" ought to arise logically. Instead, we have to rely on what we can prove empirically.
ChemBuddy | 
Chem Reddit | 
Topic: Over quantification of Cholesterol in Cheese with AOAC method(45.4.10_994.10) (Read 4484 times)