On January 1st the new regulations governing the use of the term ‘gluten-free’ came into force. As of Jan.1, to be able to call itself ‘gluten-free’ a food must contain less than 20 parts per million (ppm) of gluten, to call itself  ’very low gluten’  it must contain less than 100 ppm of gluten. It has taken years of negotiation, heated discussion and compromise to get these regulations established and although many people are still unhappy with them, they have been welcomed with open arms by manufacturers of gluten-free food and by many coeliacs. So, a bit more detail…

The Food Standards Agency  has written a detailed article for our CoeliacsMatter site explaining the background to the  regulations and what they will mean for both manufacturers and those on gluten free diets – and is a good place to start if you just want to understand the situation as it now is.

However, although these levels of ‘acceptable’ contamination have now been formalised by the regulators, this does not mean that either all medical experts or concerned coeliacs actually agree with them…  See the report on CoeliacsMatter of a symposium held in Oslo last year where two of the leading coeliac experts, Professors Carlo Catassi and Paul Ciclitira, went head to head on this issue. And for a very detailed exploration of the issue, including impassioned, and very interesting, comments from coeliacs who are both for and against the 20ppm level, see Alex Gazzola’s  blog on the 4th January.  For what it is worth, my own view of the situation, which also appears on Alex’s blog, is as follows:

1. There really is no gold-standard expert guidance in this area. There are a number of medical/coeliac experts who quote various trials to support their particular view of what is a safe level of gluten (zero, 20ppm, 100ppm etc) but as anyone who has worked in the medical field knows, the trial that provides the definitive answer this week, may well prove to be fatally flawed by a new trial next week. So while we can and should listen to what they say and be guided by those who seem most reliable – or sensible – we should not take any of their suggestions gospel.

2. Every coeliac, like every allergic and indeed every person suffering from any condition, is an individual and although their symptoms may fall within a general disease pattern, no one person will ever be identical with any other person. So while 20ppm of gluten may fine for coeliac A, it may cause problems for coeliac B – just as one peanut allergic person may react to inhaled peanut dust while another may only react to ingested peanut.

3. We have to be practical. If gluten free food is to be manufactured for those on gluten free diets, manufacturers must have some level to which to work. Manufacturing to zero gluten is not practically possible, so some level has to be set to give manufacturers guidance. If it is not, they will not make the food.
(I know of several manufacturers who will not sell their products as dairy free, although they actually are. Because there is no level set for dairy contamination and they do not want to risk being outside the law, they are not prepared to label the food as dairy free, thus depriving those on dairy-free diets of a good product.)
Given the current state of knowledge and the huge variability in the sensitivity of coeliacs, 20ppm seems to me to be a reasonable compromise, especially since guaranteeing that they will always meet the 20ppm level means that most foods contain significantly less gluten than that.

As for those who are ultra allergic, ultra sensitive coeliacs may not be able to tolerate this but, sadly, I do not really see that there is any other viable option.

However, it does not end there. The law of unintended consequences operates just as effectively in the food world as it does when you start invading other countries…

Just before Christmas a colleague sent me a copy of an alert put out by by an importer of lentils as follows:

We bring to our Lentil customers’ attention a change in EU regulations about labeling of allergens coming into effect in Jan 2012 where the definition of ‘gluten free’ changes from the current maximum of 200 ppm (parts per million) to a new maximum level of 20 ppm, only 1/10th of the previous tolerance.


Lentils are grown in rotation with wheat and other cereals, and grain sizes are also similar, so despite thorough cleaning over most modern plants, it is only possible to supply lentils with an allowed tolerance for cereal grains which may translate to a gluten content on the borderline of the new 20ppm limit, so there will be a risk whether lentils would be compliant with the changed ‘gluten free’ definition.


For example our specification for Dark Speckled Lentils allows a tolerance of 3 cereal grains per kilo.

Our standard specification for Red Split Lentils allows a tolerance of 0.04% cereals but through work with suppliers we can negotiate new tolerances of 1 cereal grain per kilo (which may translate to about 6ppm but this estimate will be influenced by the cereal variety, grain size and variable gluten content).

As it will be expensive and time-consuming to test each batch, we suggest it will be safer to adopt ‘May contain gluten’ labeling for all lentils.



What, effectively, this means is that yet another excellent and nutritious gluten-free staple food item will become unavailable to anyone on a gluten-free diet who will fear that they may be ‘gluten-ed’ when in fact, the chances of them being so are almost vanishingly small.

You cannot blame the lentil importer. They are quite right that it would be unrealistically expensive and time consuming to test every batch of lentils to guarantee that they were under the 20ppm limit – especially when the vast majority of their produce will be consumed by customers who are totally uninterested in whether or not their lentils are contaminated with wheat. But for those on a gluten free diet, for whom lentils had provided an excellent alternative source of nutritious food, it is hugely frustrating and upsetting.

This does not mean that the 20ppm rule is wrong (for all the reasons explored above) –  it just illustrates yet again how desperately we need  some common sense to be applied to  ’may contain’ labeling. A regulation which forbade the use of ‘may contain’ unless it were put in context (‘made in a factory in which peanuts/gluten is used’, ‘made in a dedicated nut/gluten free factory but cannot guarantee the ingredients are totally nut/gluten free’, ‘grown in rotation with wheat so possibility of contamination’ etc) would be a massive help as it would give those who wanted to avoid that ingredient enough information to make their own informed decision according to their own degree of sensitivity.

Obviously, there are practical problems – not the least being that manufacturers do not wish to have to print (or consumers to read) and essay on what food is safe for whom, but there must be a way to improve the current total lack of specificity. My own preference (as I am sure I have said before…) is for a traffic light system: red for containing nuts/gluten etc, orange for not containing any nuts/gluten etc but made in a factory/grown in a field etc where nuts/gluten are also present; green for totally free.

At least that would avoid nonsenses such as ‘may contain gluten’ labels on lentils or the even more ridiculous situation I highlighted last year with Green and Blacks who declare milk as an ingredient (which it is not)  in their dark, milk-free chocolate just in case it should be contaminated…..