At what step do we consider preventive controls of Allergens?

Dear Collegues! during ISO 22000 introduction (or BRC) at what step it must be? Usually we done at hazard analysis?

Youre asking when should allergens come into play?

Ideally at the time the facility is designed. If you can segregate allergens effectively, you have fewer problems.

As far as hazard analysis goes, then all steps should consider allergens. For example, is your raw material likely to contain allergenic material? Are allergens stored near the raw material?

I think you're talking about HACCP?  When you consider allergens I would consider them at every process step throughout the process and consider them as a separate category distinct from chemical, physical and biological.  Ideally as previous posters have mentioned you will have considered allergens at the build stage of the plant meaning you've built food safety into the process and certainly you must consider them at any NPD stage.  

A few tips with allergen hazards.

• Consider the form of the allergen, the physical form, how it can be spilled and how it can be cleaned.  Dusts and powders can get widely distributed in a production area.  Small seeds like sesame seeds can be very difficult to clean up.  Egg can bake onto surfaces and become very difficult to remove with normal washing processes so these all need special consideration.
• Ideally look at segregation.  If complete segregation isn't possible, try segregating by time (e.g. running allergenic products after non allergenic and before a full clean.)
• Can ingredients be substituted for non allergenic ingredients in a cost effective way?  E.g. mustard for mustard flavour?  Oats for gluten free oats?  Wheat flour for cornflour?  Worcestershire sauce for vegetarian (non fish) Worcestershire sauce?  
• Don't forget mislabelling.  One of the biggest allergen hazards isn't mixing up the allergens but mixing up the labels.  One of the biggest reasons for public recalls.
• It is perfectly acceptable to clean effectively to remove allergens and it's possible to achieve that aim.  Just validate using ELISA and verify.
• BUT always ensure you do a positive control (i.e. a swab and a positive product).  Recovery in ELISA testing can be inconsistent leading to false negatives.  By using a positive control you can at least see the method picks up the allergen.
• Ensure the testing you use is specific for the allergen.  While a positive control would pick it up if it isn't, it's a waste of money.  For example, I remember sending off an allergy test for Worcestershire sauce once (which contains anchovies) but I checked before I sent it what "fish" the ELISA test was sensitive for.  It turned out to only be sensitive for white fish.  Had I not asked the question, I could have easily assumed the anchovies were so broken down over their long maturation time that they were no longer "present".  At the time the only method available for anchovy was PCR which came back as positive much to my surprise but of course PCR doesn't directly measure the allergen itself.
• And of course, train, train, train...  Everyone needs to know what they're doing.

I'd support GMO. Be certain exactly what you're testing for! I'm allergic to fish (but not white fish) so have to find vegetarian Worcestershire Sauce for home use.

Dear Collegues! during ISO 22000 introduction (or BRC) at what step it must be? Usually we done at hazard analysis?

i suggest you have a look at NOTE 3 and clause 7.2.3(g) of iso22000 standard.

And clauses 10.3 and 14.2 of iso22002-1 standard.

If you would like to see an example of (intrinsic) allergen handling in  a "combined" haccp plan -

http://www.ifsqn.com...ge-7#entry50651