Allergen management tools in food production – Professor Clare Mills, Manchester Institute of Biotechnology
The Paul Wilson case. Richard Wright QC, Park Square Barristers
Allergen Management – exposing and analysing manufacturing pitfalls – Julia Pepler – Food Integrity Consulting
Precautionary Allergen Labelling – are we ready to quantify risk? – Chun-Han Chan, Food Standards Agency
FreeFrom – what does it really mean? – Barbara Hirst, RSSL
Food allergy in the real world – Dr Michael Radcliffe, University College London Hospitals NHS Foundation Trust
Allergen management tools in food production
The introduction of mandatory labelling of allergens has helped allergic people enormously – label reading has become one of their life skills and one of the tools that they use to manage their allergies.
But these skills have not helped them navigate the confusion surrounding PAL – Precautionary Allergen Labelling or ‘may contain’ warnings.
EuroPrevall – gathering data
Fifteen years ago there was very little data on allergy and it was this lack that spawned the 2005 EuroPrevall project – a massive international study involving 63 partners across Europe and Asia designed to discover more about the incidence of allergy with the purpose of improving its management.
EuroPrevall consisted of three strands:
Its purpose was to discover:
It also investigated such subjects as diagnosis and the disconnect between self diagnosed and clinically diagnosed food allergy. Around 20% of the population believe they are food allergic while only 4–5% are clinically diagnosed.
EuroPrevall produced massive quantities of data across all fields of allergy – some of it quite surprising. For example:
The EuroPrevall studies also allowed researchers to assess the levels of allergen at which reactions are likely to occur in the allergic population. This was normally in milligrams (around 3mg for peanut protein for example) except for shrimp which required a much higher dose to elicit a reaction.
IFAAM – moving forward to management
However, although the EuroPrevall project had produced enormous quantities of data on allergy it had not offered any procedures as to how to deal with it. For that a new project was initiated – IFAAM – Integrated Approach to Food Allergen and Allergy Management – tasked with developing management procedures.
This essentially meant creating validated risk management tools for industry and effective allergen management – an allergen management tool box. This task was undertaken by Professor Steve Taylor in the US, Professor Jonathan Hourihane in Ireland and Unilever’s food allergen expert Rene Crevel in the UK.
The tool box asks a number of questions which require appropriate answers before continuing to the next stage.
1. Is the allergen present?
If so at what level and can it be identified? This involves differentiating between allergens in a homogeneous mixture and particulates. Where the allergen is found in a homogeneous mixture it is spread evenly throughout the food at a relatively low level. Where the allergen is a particulate (heterogeneous) it means that there is a small piece of the allergen, such as a peanut, in the product. This will deliver a very high level of allergen if it is eaten – but it will not be present in 99.9% of the food. This makes it very hard to identify.
2. An allergen tracking tool.
3. Probabilistic risk assessment.
The IFAAM work has included single dose challenges to validate the proposed thresholds for major allergens. This work is not yet complete but is making good progress and it looks as though the threshold for peanut may end up at around 1.5mg peanut protein while that for hazelnut may be lower at 0.9mg protein. Thresholds for milk and egg now also look to be in sight.
But while these tools are making good progress there are still issues over analysis.
The ELISA tests that are in common usage are good at identifying allergens within the food (although not at very low levels) but cannot quantify the amount of an allergen in a food. There can also be issues as to the food matrix (the food in which the allergen is found). Some (such as chocolate) effectively disguise the allergen within the food.
Mass spectrometry offers an alternative form of analysis which reduces the proteins to peptides and then tests and analyses them as they would a pesticide. However, this system also has problems in identifying proteins in some foods (such as chocolate). It is also very expensive so is not as yet a viable alternative.
The Paul Wilson case.
Richard Wright was the QC who prosecuted Mohammad Zaman – and got him sent down for six years – for the recent and tragic death of Paul Wilson.
For those who had not followed the case, Paul died from an anaphylactic reaction to peanut in an Indian takeaway meal that he had bought from one of Zaman’s restaurants even though he had specifically asked for a meal with ‘no nuts’.
Mr Wright walked us through the case, starting with an episode a week before Paul Wilson died when another peanut-allergic customer, Ruby Scott, had been hospitalised after eating some of a meal from another of Mr Zaman’s restaurants. Ruby’s meal also contained peanuts when she had specifically asked for a nut free meal.
Thanks to the persistence of the North Allerton CID the two incidents were pursued revealing that Mr Zaman had knowingly, but secretly, substituted peanuts for the more expensive almonds in the food in his restaurants. This change had not been reflected in any of his menus. He had moreover completely failed to institute any allergen training amongst his staff and had no procedures for handling allergens – while local authority tests showed that all of his outlets were massively contaminated with peanut proteins.
He was charged and convicted under the Food Safety Act 1990 for selling food that was not what it said it was – peanuts whereas the menu stated almonds. This act long precedes the new allergen regulations but is still one of the primary means of regulating allergens in food.
Allergen Management – exposing and analysing manufacturing pitfalls
Julia looked in more detail at the manufacturing process to identify the points of greatest risk and most crucial areas of concern.
Risk assessment, she believes, is vital. All too often it is weak – not sufficiently detailed or robust – and she recommends the Camden Guide 71 as a good starting point.
Risk assessment falls into two areas:
High risk areas are those where dispensers or equipment are shared; low risk areas are those where there is normal handling or contact with the foods. The risk from airborne contamination is low.
To achieve effective risk assessment it is vital to record the thought processes involved in each and every step and the mechanism by which cross contamination could occur.
Risk assessment criteria
The criteria which need to be considered in every possible risk situation are:
1. Probable – possible – unlikely
Keep it simple
Over complicated controls will not work. Allergen best practice should be part of day to day best practice so there is nothing extra for operators to remember – or forget!
Extra protective clothing is of little practical value but it does serve to remind operators that they need to be careful.
Often thought of as the simplest of allergen controls, it is not the easiest.
Again, cleaning need not be complicated but it requires great attention to detail. Once again, detailed records of every process are essential to validate the cleaning method – and it is the cleaning method that requires both validation and verification, not the cleaning kits.
Clear lines of communication are vital – from supplier to manufacturer – or, in food service, from front of house to kitchen.
When dealing with raw material suppliers, you cannot ask too many questions – establishing whether what they do might involve a risk to your product, whether their processes match yours, whether their understanding of risk matches yours.
Warehouse, packing and labelling
The warehouse is essentially a low risk area in terms of contamination risk but a very high risk area in terms of communication and the possibility of mis-picks.
Very clear identification of ingredients and of packaging is vital. Most recalls do not occur as a result of allergen contamination but as a result of mis-picking, foods being packed into the wrong containers or being wrongly labelled.
Precautionary Allergen Labelling – are we ready to quantify risk?
Chun-Han Chan, Food Standards Agency
Whatever you call it, Precautionary Allergen Labelling, PAL or 'may contain' warnings, along with any 'freefrom' claim, are entirely voluntary. Thye have no legal force so any product which makes a PAL declaration would still be governed by basic food law: it must not be unsafe, injurious to health or have an adverse health effect and you must be able to provide information about it.
What is an 'adverse health effect'?
In terms of the possibility of an allergic reaction, what is acceptable or tolerable? And how do you communicate risk to the consumer?
Right now there is no international agreement on how much is too much. To assess the public risk, this needs to be defined. (See Clare Mills presentation above.)
But do you assess 'how much' on portion size or on consumption? (The former is easier but unreliable as one individual may consume several portions.)
And how do you deal with homogeneous versus heterogeneous contamination. (See Clare Mills above again.)
Quantifying risk - the public health approach
You cannot use absolute zero as your benchmark as we are not currently able to analyse to absolute zero.
You therefore have to use risk management techniques on a population basis to protect, ideally, 99% of the allergic population. In practice in may be closer to 95%. However, you would not anticipate that, if it were only 95%, any reactions would be severe.
European research projects are close to establishing these levels for peanut, milk, egg and hazelnut (again, see Clare Mills above) but what defines allergenicity in a food? Cross reactivity between similar proteins in different foods may present further problems.
(To illustrate her point Chun referred back to the cumin allergy scare last year when almond was thought to have been detected in some Bart Spices Cumin. Forensic examination could find no possibility in the production chain that the cumin could have been contaminated by almond. Eventually, having used a number different analytical techniques, it was found that the cumin did contain traces of mahaleb, an aromatic spice made from the seeds of a particular cherry tree (prunus mahaleb) that is a very close cousin of almond. The initial analysis had identified its proteins as almond.)
FreeFrom – what does it really mean – and how to achieve it?
Barbara Hirst, RSSL
In regulatory and in consumer terms, 'freefrom' is an absolute claim – it means zero. For retailers and manufacturers it has many different meanings. But in regulatory terms the only claim that is covered by regulation is gluten free/20 parts per million; all other claims are covered by food law. (See above)
However, there is great pressure on the manufacturer to supply 'freefrom' food – so how best can they do this?
A dedicated free from gluten/dairy/nut etc site obviously dramatically cuts down the risk. As long as staff are not allowed to bring the allergen on site the only risk is from raw materials. If a non-dedicated site is used the risk management is crucial and the BRC (British Retail Consortium) and FDF (Food and Drink Federation) offer good guidance.
Barbara pointed up these areas as critical:
Food Allergy in the real world
Dr Michael Radcliffe, University College London Hospitals Foundation NHS Trust
To illustrate how complex allergy is to diagnose and treat in the 'real world' Dr Radcliffe outlined a few of his more recent cases.
1. A patient with a history of asthma and anaphylaxis to peanut had an anaphylactic reaction to chicken and chips with onion rings. There were no traces of peanut in the meal but further investigation showed that the batter mix had come from Holland and included lupin flour. (Many peanut allergics also react to lupin.)The lupin had been listed on the batter mix pack but not on the onion rings.
2. A patient who was nut allergic had started to react to seeds - pumpkin, poppy etc. While waiting for an allergy appointment she ate some flatbreads and had a full anaphylactic attack. There were no nuts in the flatbreads but there were pumpkin seeds and a very small amount of poppy seeds – 1.5% in the whole flatbread.
3. A patient had a reaction after a meal at Wagamama's. Initially she thought it was the medication that she was on but analysis of the meal showed that it contained buckwheat. This was the onset of a totally new allergen as an adult but it turned out that she had been sensitised to buckwheat via a pillow which had been stuffed with buckwheat husk – a common stuffing in Japan. She had also been taking Omeprazole, a drug which alters digestion processes and causes an increasingly poor breakdown of proteins. This could have been relevant in her reaction to the buckwheat protein.
Dr Radcliffe explained that it was now possible to analyse proteins to establish cross reactivity and to test for protein fractions linking, for example, asthma with birch, latex and banana. This was allowing the diagnosis of Pollen Food Syndrome which is the commonest new presentation of serious food allergy although it does not lead to anaphylaxis.
See the Anaphylaxis Campaign site for more on their conferences and support material.