See all issues here

Not all gluten is equal...

January 2014

At the start of 2014, it is likely that the implementation of R. 146/2010 is now in full swing. All food industry role players should thus be fully aware of the requirements for allergen labelling, including the stipulations for gluten declarations and for ‘gluten-free’ claims, the crux of which is summarised in Box 1 below.

Box 1: Prominent SA regulations relating to gluten

According to the South African Foodstuffs, Cosmetics and Disinfectants Act (Act 54 of 1972) – Regulations Relating to the Labelling and Advertising of Foodstuffs (R.146/2010):

● "Gluten" means the proteins that naturally occur in a significant cereal to which some persons are intolerant.

● "Significant cereal" means anyone of the following cereals:
- Wheat, meaning any species belonging to the genus Triticum, including varieties such as kamut and spelt;
- Rye, meaning any species belonging to the genus Secale;
- Barley, meaning any species belonging to the genus Hordeum;
- Oats; or
- Crossbred hybrids of wheat, rye or barley

● The claim "gluten-free" shall only be permitted on a foodstuff if the foodstuff does not contain any of the following:
- An ingredient that is any species of the significant cereals;
- An ingredient that is derived from any of the aforementioned significant cereals that has not been processed to remove gluten;
- An ingredient that is derived from any of the aforementioned significant cereals which has been processed to remove gluten so that the use of that ingredient results in the presence of more than 20 mg/kg (ppm) gluten in the end product; or
- More than 20 mg/kg gluten, where the level of gluten is determined by the R5 Mendez Enzyme-Linked Immunosorbent Assay (ELISA) for gluten (as described in Guidelines), or other Codex recommended methods.

In line with the Codex Standard for gluten-free foods, South African legislation specifies that a foodstuff is eligible to bear a ‘gluten-free’ claim if it can be demonstrated that it contains no more than 20 mg/kg (ppm) gluten, a level considered to be generally safe for celiac disease sufferers. What may often be overlooked, however, is that not all gluten in foodstuffs is equally detectable and different analytical methods may be required to reliably assess the true gluten content of food. This Q&A-based newsletter attempts to elaborate on challenges faced in detecting and quantifying gluten in processed foods and to elucidate the most suitable methods to use and when.

Q . What is wheat allergy and celiac disease and where does gluten fit in?

Unlike wheat allergy which is an acute immune-mediated response to any of the myriad of proteins found in wheat, celiac disease is a hereditary inflammatory disorder of the small intestine that is triggered by the ingestion of gluten, which progressively impedes the absorption of nutrients from foods and results in a number of secondary conditions. Although the term ‘gluten’ pertains to wheat, it is normally used in the medical literature to jointly refer to the proteins (particularly the prolamins) in wheat, rye, barley and possibly oats that are responsible for causing harmful effects in celiac sufferers. Celiac patients consequently need to avoid all of the aforementioned grains, whereas a wheat allergic individual would need to avoid wheat, but would likely tolerate other grains. The collective reference in R.146 to ‘significant cereals’ is therefore aimed at protecting both wheat allergic and celiac sufferers from adverse reactions to the listed grains.

Q . What is the prescribed method for gluten testing?

In 2006, the R5 Méndez sandwich ELISA was collaboratively tested and endorsed by the Codex Committee as the type I method for gluten determination and this method was subsequently adopted in R.146/2010 as the prescribed method for gluten testing to support gluten-free claims (see Box 1).

Q . What are the advantages of the R5 Méndez sandwich ELISA for gluten testing?

The R5 Méndez sandwich ELISA has been validated for the quantification of native (intact) and heat-treated gluten in wheat, barley and rye, with no reported interference from oats, corn (maize), etc.

Technical background: The R5 Méndez sandwich ELISA method is based on the R5 monoclonal antibody raised against the potentially celiac toxic epitope (antibody-binding site) QQPFP (glutamine-glutamine-proline-phenylalanine-proline) and the use of a specialised cocktail extraction procedure improves the response to the heat-treated proteins.

Q. What are the drawbacks of the R5 sandwich ELISA for gluten testing?

Apart from potentially overestimating barley hordeins in barley-contaminated foods, the main criticism of the R5 sandwich ELISA is that it is unable to accurately quantify the smaller protein fragments associated with hydrolysed and fermented gluten, even though these may remain dangerous to celiacs.

Q. Why can sandwich ELISAs underestimate gluten?

The sandwich ELISA relies on the binding of antibodies to at least two linked epitopes (binding sites) of the protein (i.e., for the R5 sandwich ELISA, two QQPFP epitopes would be required). However, when gluten is broken into smaller protein fragments during enzymatic/acid hydrolysis or fermentation, the resulting peptides may no longer contain two epitopes in close proximity for antibody binding.

Consider a protein where ‘QQPFP’ is the epitope and ‘a’ represents other amino acids:

aaaa QQPFP aaaaaaaaaa QQPFP aaaQQPFP aaaaaa QQPFP

If the protein is hydrolysed/ fermented, these fragments may result:
1. aaaa QQPFP
2. aaaaaaaaaa QQPFP QQPFP

The sandwich ELISA would be unable to measure the first or third protein fragment as these contain only 1 QQPFP epitope. Since only the second fragment would be measured, the gluten level would be underestimated, even though the toxic sequence QQPFP is present in all three fragments.

Q. Which method should be used to measure hydrolysed and fermented gluten?

A second generation competitive R5 ELISA was recently developed and validated for the specific detection of gluten-derived peptides from wheat, rye or barley in those foods that have undergone fermentation or hydrolysis. This assay is also based on the R5 antibody and has a low limit of detection, but it differs from the sandwich format in that it only requires one epitope for binding, thus allowing the detection of small or degraded fragments.

One disadvantage of the competitive format is that it is currently only compatible with ethanol extraction, which is unsuitable for the detection of heat-treated proteins. However, a cocktail extraction compatible with this method is under development.

Technical background: A mixture of hydrolysed prolamins from wheat, rye and barley are used as the calibrator for this test, which permits results to be expressed in mg/kg (ppm) gliadin (which is multiplied by a factor of 2 to convert to gluten). This still allows a correlation to the ‘gluten-fee’ thresholds set by Codex and our local regulations.

Q. Which types of food and beverages might contain hydrolysed and/or fermented gluten?

Partially hydrolysed gluten will likely be found in any food ingredients made from wheat and other starch hydrolysates, so look out for terms such as hydrolysed wheat protein, wheat-based dextrin and maltodextrin, barley malt extract and starch/glucose syrup. Fermentation also leads to gluten degradation and this is relevant in foods such as beer, (alcohol-free) malt drinks, hypoallergenic food based on gluten, soy sauce and sourdough. Brewer’s yeast may be a by-product of the beer brewing process and as such may be contaminated with malt and grain peptides. It is also important to note that “yeast extract” could contain gluten peptides if spent yeast arising from the beer manufacturing process is the source – so always make sure that you get all the facts from the supplier.

Q. What are the expected discrepancies in results using the different methods?

Two major studies have evaluated gluten levels in a range of beer products from around the world using both the sandwich and competitive R5 ELISAs and both have reported that the sandwich format considerably underestimated the gluten levels compared to the latter. In some instances, gluten levels fell below the quantification limit with the sandwich format, but were close to 20 ppm when the same products were tested with the competitive ELISA ( Dostá al., 2006; Haas-Lauterbach et al., 2012).

To conclude, the degree of confidence that can be placed in a producer’s assertion that a food is ‘gluten-free’ will be based on him knowing his product, its ingredients and the processes it undergoes, as well as him appreciating that not all gluten is equal and that the best available method must be chosen for analysis on a case by case basis.


Dostá al. (2006). Food Addit. Contam. , 23, 1074–1078.
Haas-Lauterbach et al. (2012). J. AOAC Int., 95, 377-381.
Haraszi et al. (2011). J. AOAC Int., 94, 1006-1024.
Mena et al. (2012). Talanta, 91, 23-40.