Soy Allergen Threshold
Our manufacturing plant is introducing a new product that contains natural flavors made with soy. After sending a sample of the finished product out to an allergen specialized laboratory, the Soy Allergen Analysis showed that it contained 2.3 ppm of soy proteins.
In my correspondence with the laboratory, they advised that most companies would find a level of 5 ppm and below acceptable. Apparently 1100 ppm (220 g of 5 ppm product) is where the most sensitive 1% of soy allergic people begin to feel effects (see attachment).
Does this seem like an acceptable level to test for in post cleaning (5 ppm)?
---We currently have a soy test kit that will show positives of 5 ppm+. Will this suffice?
Thank you,
Doug
This is the problem with zero tolerance testing for allergens as methods get better and better at detecting trace amounts of protein residues.
The hyper-conservative approach would be no, this is not acceptable because the presence of allergens on your swabbed surface suggests that in a non-homogenous contamination of your line you could have residual allergen levels in excess of those values.
The customer compliance approach would be no-depending on the customer. I've had large multi-national companies that have used me as a contract service, and they require negative results on FARRP ELISA testing, which have detection limits of down to 2.5ppm if I recall. Potentially lower depending on assay/dilution.
In the US we've adopted a zero tolerance approach for allergens in general, FARRP has this to say on the subject regarding thresholds for reactions:
Regulatory guidance on thresholds for allergenic foods is sorely needed to improve the quality of life for food-allergic consumers and to focus limited regulatory resources on the genuine risks that exist to these consumers. Of course, regulatory (population) thresholds must be scientifically based but sufficient evidence exists to allow the development of regulatory thresholds for some of the most hazardous allergenic foods including peanuts and milk.
Right now these thresholds aren't recognized, and your datapoint doesn't suggest that you wouldn't have some amount of food which might exceed a threshold. More data would be needed to give statistical power to your quantitiative assessment that all of your food contact surfaces would be below that 5ppm threshold.
Given the actual food safety risk presented by the soy levels you encountered. I would lean towards thinking that unless you have a harborage space in your line that could lead to higher levels of contamination, it may be in the best interest of your business to stick with the 5ppm limit test methods so that you can establish a realistic threshold for allergen removal in your process, and identify any materials/areas/cleaning processes on your line that might pose a problem and get them taken care of before moving onto a more sensitive method.