No blood test can detect food intolerance, not even the IgG/IgG4 ELISA tests that check for ‘hidden food intolerances’ to over a hundred common foods.
How do I get help, then?
Depending on where you live, it may be difficult to find a doctor or registered dietitian who has experience diagnosing food intolerance. Provided that your doctor has already ruled out other health problems, you can always contact the professional dietitics association in your country or visit their website. These associations maintain lists of dietitians by location and specialty. Depending on your medical history, some dietitians may be willing to consult with you from a distance.
Remember, the only way to diagnose food intolerance is through an elimination diet and food challenges. Even if you are concerned about a dairy or gluten sensitivity, IgG blood tests are of no value. (1-3)
In fact, allergy and immunology associations around the world have issued position statements warning against IgG blood tests, which they see as leading to potentially dangerous dietary restrictions, overlooked conditions (including true allergies), and unnecessary costs. Immunologists have found no evidence that IgG antibodies cause delayed food allergies or intolerances. As we’ll see later, it also turns out that the main demonstration of ‘successful’ IgG-based exclusion diets was not really so successful.
Still, you’re probably here because you have heard (from testing companies or the alternative medicine community, no doubt) that science supports not only the assumption behind IgG blood tests – that IgG antibodies are linked to delayed food sensitivities and chronic symptoms – but also the reported ability of IgG-based exclusion diets to relieve chronic illness. You may have also seen these tests in pharmacies or in the offices of doctors who practice functional or integrative medicine. You may be wondering, “Why are these tests being sold if they don’t work?”
“Why” involves motivation, so I’d rather focus on “how.” In the case of IgG blood tests, “how they can be sold” is an unfortunate melding of a few facts taken out of context plus some flawed research, with a few misrepresentations about how the immune system works thrown in. I must be clear, though – the problem isn’t one with consensus in the medical community. As I said, immunologists have not found any evidence of IgG-based food sensitivities, and it’s not as if they wouldn’t have noticed – for sure, IgG antibodies come up in research on immunotherapy, and this is a hot topic these days.
The alleged medical credibility of IgG blood tests is really coming from researchers in other specialties who hear about these tests and decide to give them a try for whatever condition they study (migraines, IBS, ADHD, etc.). This doesn’t seem particularly wise, but under the tenets of evidence-based medicine, a high-quality study should weed out bad ideas. However, in debunking the myths used to support food intolerance blood tests, we’ll see that not all studies are high quality.
Antibodies and allergies
Before we dive into the myths, let’s start with some background on antibodies and allergies. Antibodies, also known as immunoglobulins (Ig), are proteins produced by the immune system that help protect our bodies from foreign materials. There are several different classes of antibodies. For example, IgE antibodies bind to allergens (proteins from plants, animals, and fungi that should otherwise be of no threat to the body) and trigger histamine release from mast cells in what we think of as classical allergies (like to pollen or peanuts). Specialists refer to classical allergies as IgE-mediated reactions.
Allergy, intolerance, or sensitivity?
You may see different terms used to describe IgG blood tests – some say they detect food intolerance, others say delayed allergies, and others say food sensitivities. Since IgG antibodies are not involved in any adverse reactions to food, it doesn’t matter which label you choose – they’re all incorrect. Historically, though, these tests were advertised as detecting food intolerance, but the recent trend is to use the term food sensitivity. By medical convention, however, if IgG-based reactions were real, they would actually be a type of allergy. I prefer to speak of them as IgG-mediated reactions, but I do vary it a bit depending on the context.
The symptoms of IgE-mediated allergy come on rather abruptly after eating an offending food, often within 30 minutes to an hour. Other types of reactions – like cow’s milk protein intolerance – involve different aspects of the immune system and take longer to develop. Supporters of food intolerance blood tests believe that IgG antibodies cause yet another delayed type of allergic reaction, one which the medical community has failed to notice. Often this reaction is incorrectly labeled as food intolerance.
The IgG antibody class has several specialties, one of which is protective immunity. Protective immunity refers to our immune system’s ability to recognize and remove invaders like bacteria or viruses. This is also what keeps us from getting chicken pox twice. Blood is routinely screened for IgG antibodies to check for prior infections or vaccinations, but this is not the kind of blood test we’re talking about here.
IgG antibodies also support tolerance. Tolerance occurs when the immune system remembers to suppress its reaction to a foreign substance or to the body itself – in other words, tolerance is why the majority of us can eat whatever we want without fear of an allergic reaction. There are different mechanisms behind tolerance, but in many cases we have regulatory T cells to thank. Regulatory T cells invoke several helpers to suppress the immune response, and these include IgG antibodies. (4) IgG blocks IgE antibodies from combining with allergens and producing an allergic reaction, although the precise mechanics are under debate.
Even with this brief introduction, we can start to see that IgG antibodies are unlikely to be behind adverse reactions to food. The most advanced scientific knowledge points to the conclusion that food-specific IgG antibodies in our blood indicate exposure and tolerance, not in-tolerance, to those foods. (5) In other words, IgG antibodies are just a normal part of life. Unfortunately, commercial laboratories and alternative medicine still perpetuate the notion of IgG-mediated food reactions.
Taking on the myths behind IgG blood tests
We’ll debunk six myths that are used to support food intolerance blood tests:
- High levels of IgG antibodies mean that you have a hidden or delayed reaction to food.
- Food-specific IgG antibody levels correlate with chronic symptoms.
- IgG antibodies degranulate basophils.
- IgG antibodies cause inflammation.
- IgG blood tests detect Type III hypersensitivities.
- Diets based on IgG levels have been shown to successfully treat symptoms.
Myths 1 and 2 cover the (lack of a) relationship between food-specific IgG antibodies and symptoms. Myths 3 through 5 get technical as they explore the mechanisms by which IgG antibodies (supposedly) cause illness. Myth 6 is the big one – it covers the claim that exclusion diets based on an individual’s food-specific IgG levels can successfully treat chronic symptoms. We’ll become amateur scientists to deal with this myth, and, in the process, we’ll see that sometimes it only takes a basic understanding of the scientific method and a little logic to evaluate medical studies.
Myth 1: High levels of IgG antibodies mean that you have a hidden or delayed reaction to food
Everyone produces IgG antibodies to food. Even though food intolerance blood tests rank your IgG antibody concentrations for various foods as low, medium, and high, there is actually no such thing as a ‘correct’ level. IgG concentrations vary from person to person and depend on diet – perhaps even on how one was fed as an infant (4) – so even healthy people will have high IgG levels for some foods. This means that a healthy person could get the same diet recommendations from an IgG blood test as a person with symptoms.
The same thing happens in classic food allergies, where there is no question that IgE antibodies are responsible. Some allergy-free people have elevated IgE levels, so doctors will not diagnose someone as having a classic (IgE-mediated) allergy without some sort of secondary evidence. This evidence could include a personal history, a physical exam, a skin-prick test, or an elimination diet and food challenge; of these, the strongest evidence is the elimination diet and food challenge.
Some supporters of IgG blood tests do acknowledge this situation indirectly by saying that the tests should only be used to “guide” a standard elimination diet and food challenge. So, then, how well do these tests perform as guides? Promotional materials almost always include testimonials by people who believe that an IgG blood test helped them find the foods behind their symptoms, and this is not surprising. As Brostoff and Gamlin (6) point out, alternative therapies could not make viable businesses if there were no successes. But these authors go on to explain that success stories are far from evidence that a test is worth your money:
“Given that the most common sources of food intolerance are wheat and milk, such therapists can achieve a reasonable success rate by diagnosing sensitivity to these two foods in all their patients. If eggs, oranges, chocolate, tea, and coffee are added to the list, they may well achieve success with 50 percent or more, and some patients will benefit from the placebo effect alone.” (p. 141)
So it seems that common sense would have about a 50% chance of finding at least some of the relevant foods for people who, we assume, actually have a food sensitivity. That’s the same odds as flipping a coin – any blood test would certainly need to do better than that. But in a 2001 survey of UK residents who had taken the YorkTest IgG blood test, only about 50% saw significant improvement in symptoms after eliminating their reported foods. (7) [The survey was repeated in 2007 (8), but its write-up contains a hole: it does not give us enough information to calculate an analogous success rate to compare with the first survey (9).] All in all, it looks like IgG blood tests, common sense, and luck offer about the same amount of help for guiding an elimination diet.
Myth 2: Food-specific IgG antibody levels correlate with chronic symptoms
You may be surprised to learn that no one has shown that food-specific IgG levels correlate with symptoms. The few studies that looked for a relationship gave mixed results – for example, testing of over 5,000 people with self-reported food reactions (10) found positive correlations for antibodies to certain foods and negative correlations for others. A negative correlation means that high antibody levels are associated with health, which certainly goes against pro-IgG claims. Most importantly, only one of these studies ensured that the participants actually had food sensitivities in the first place, and this small study found no relationship between IgG levels and, in their case, irritable bowel syndrome. (11)
Confirming up-front that participants have food sensitivities (and to which foods) is a fundamental requirement of any study on food allergy or intolerance – without it, all conclusions are meaningless. Moreover, the verification of food sensitivities must be done in some way that is independent of what is being researched. After all, you can’t diagnose someone as having food sensitivity using the IgG blood test when it is the blood test itself that you are investigating. And you can’t just ask someone if they experience adverse reactions to food, because more people think that they have a food sensitivity (up to 10 times more) than actually do. Food sensitivities can only be verified through careful elimination diets and food challenges.
In scientific terms, the strictest elimination diet/food challenge protocol is called the double blind placebo controlled food challenge (DBPCFC), because there is an additional requirement that both participants and researchers be ‘blind’ to whether a person is given a placebo or a real food challenge. The DBPCFC is the gold standard of food allergy and food intolerance diagnosis, which means that it is the standard to which the validity of any new test, like a blood test, must be compared. All good research on food sensitivities requires the DBPCFC, but most studies that look for IgG-mediated reactions fail to use it. This will come into play again when we consider Myth 6.
Myth 3: IgG antibodies degranulate basophils
Basophils and mast cells are closely-related immune cells that lie at the heart of allergic reactions. These cells are first ‘sensitized’ when IgE antibodies attach to their surface; later, when multiple IgE antibodies link up with an allergen, the cells ‘degranulate’ to release histamine and other chemicals that are responsible for allergy symptoms. Over 30 years ago, researchers entertained the possibility that IgG antibodies could also operate in this manner, but this was based on an observation that has since been discounted (12). That’s the short answer — IgG antibodies do not degranulate basophils. The long answer is more complex.
In 1982, Fagan et al. (13) observed that a subclass of IgG antibody (IgG4) degranulated basophils in vitro – this is why some food intolerance tests look at IgG4 levels specifically. After Fagan’s observation, IgG antibodies became a hot research topic. As the 1990s rolled around, immunologists had accepted that IgG was not a direct cause of allergic reactions, and IgG4 in particular was cleared of its alleged involvement; still, though, the initial observation needed to be explained. (14-16) In 1992, Lichtenstein et al. (17) revisited Fagan’s work and uncovered the reason why IgG had appeared to be a reagin.
I included this myth for people who want to pull out the old research and claim that IgG antibodies degranulate basophils. However, you shouldn’t read too much into this explanation of anti-IgE antibodies: there is nothing here to imply that IgG is a sign of classical (IgE-mediated) allergies.
It turned out that IgG did not degranulate the basophils directly. Using the blood of allergic donors, Lichtenstein showed that IgE antibodies were really responsible, as one would expect. However, the IgE antibodies had IgG antibodies attached to them, and this IgG had hidden the IgE in earlier experiments. It may sound strange, but it is possible to have antibodies against antibodies, and that is what these IgG antibodies were – anti-IgE antibodies. Hidden IgE antibodies are not uncommon: in certain tests, the presence of IgG anti-IgE antibodies can give the appearance of increased IgG levels and decreased IgE levels for the same allergen. (18)
Myth 4: IgG antibodies cause inflammation
To say that IgG antibodies cause inflammation is like trying to name a tune from just one note. There are four subclasses of IgG (IgG1 through IgG4), each with different roles. From the study of protective immunity, we know that some IgG antibodies have pro-inflammatory effects while others are anti-inflammatory. (19) However, the protective immune response involves a finely choreographed balance between these players, along with many other antibodies and cells. IgG4 antibodies help to wrap things up at the end of the immune response and have an anti-inflammatory effect. (19) Overall, IgG antibodies are necessary to keep our immune system in check, and singling out one type of IgG to conclude that IgG antibodies cause inflammation is a gross oversimplification – and just plain wrong.
Myth 5: IgG blood tests detect Type III hypersensitivities
Promotional materials from some laboratories will try to convince you of IgG’s role in food sensitivities by bringing up an unrelated point – that IgG antibodies are involved in Type III hypersensitivities. That is true, but it has nothing to do with what we are talking about here.
Type III hypersensitivities occur when immune complexes, made from IgG antibodies bound to other proteins, deposit in tissues like the kidneys, the joints, or blood vessel walls. This activates the immune system and leads to tissue damage. Type III hypersensitivities are caused by chronic infections, by inhaling dusts from hay or mold, or by your own body in autoimmune disorders, but not by foods. If you have a chronic Type III hypersensitivity reaction, you’re sick, you know it, and your doctor knows it – you might have a type of arthritis, breathing problems, or lupus.
Myth 6: Diets based on IgG levels have been shown to successfully treat symptoms
In debunking Myths 1 through 5, we have seen that there is no evidence to support the existence of IgG-mediated food reactions. In spite of this, a handful of clinical studies have attempted to determine whether diets based on IgG levels can reduce symptoms by looking at two specific groups of people – patients with migraine or patients with irritable bowel syndrome (IBS). Some studies found no benefit (see reference 20, for example), while others saw mild effectiveness (see reference 21).
These diet studies compare the effectiveness of test diets – ones that exclude the foods for which an individual has high IgG levels – against “sham” diets that serve as placebo controls. The best known of these studies, and the one regarded by alternative medicine as the pivotal study for legitimizing IgG blood tests, is a randomized controlled trial conducted by Atkinson et al. in 2004 (21). In order to blind participants to the nature of their diet, both diets in the Atkinson study excluded the same number of foods, but the control diets excluded foods for which a person did not have high IgG levels. Here, the group of IBS sufferers that received the test diet saw a 26% improvement in symptoms over the group that received the control diet.
The Atkinson study may look promising, but we’ll see that it suffers from inherent design flaws that essentially nullify its results. Before we get into this, though, we need to discuss how the scientific method is supposed to be applied to clinical studies. The scientific method is a procedure used to ensure that we make valid conclusions about the world around us. Observations are used to formulate a hypothesis about the way things work, and the hypothesis leads to predictions of cause and effect. This is where we hit the first problem with the Atkinson study – no one has observed that food-specific IgG antibodies are related to symptoms, so there should really be no hypothesis to proceed with.
When there is a reasonable hypothesis, it can be validated by testing its predictions in an experiment. Experiments are controlled situations where one and only one factor is varied and the outcome is recorded – if two or more factors are varied at one time, you would not know which is responsible for the outcome. If the outcome is the same as the prediction, and if the same result is obtained when other researchers repeat the experiment, then the hypothesis is true for that situation. In our case, if the Atkinson study really did have a valid hypothesis, it would only be true for IBS sufferers.
The scientific method is harder to apply in medicine than in branches of science like chemistry or physics where experimental conditions can be completely controlled. Some medical experiments that might seem ideal in terms of controlled conditions are unethical to perform on humans because the procedures might do harm. Moreover, people are people – participants in clinical studies have different histories, different environments, and different genes, and they don’t always follow directions. They are also susceptible to the placebo effect, where their own expectations of success or failure influence the outcome of a medical intervention.
To deal with human variability and unpredictability, medical science has developed its own arsenal of experimental techniques, the most notable of which is the use of a control group to serve as a reference for interpreting results. People in the control group might not have the illness in question or might not be subjected to the factor being studied. The selection of the control group can make or break an investigation, so care must be taken to match the characteristics of the people in the control group to those in the test group. When people are randomly assigned to each group, which is the ideal situation, the experiment is called a randomized controlled trial.
Getting back to the Atkinson study, we can see that it was not a well-designed experiment because multiple foods were excluded from the test diet or included in the control diet – in other words, more that one factor was varied at a time. Some might counter that this was necessary in order to see the full effects of the diets, and maybe that’s true, but the proper procedure would have been to conduct a DBPCFC on multiple and individual foods. While the test and control diets were both problematic, the most significant problem came from the control diet, as three independent researchers pointed out in letters to the journal that published the Atkinson results (22). One commenter noted, “regardless of IgG antibody status, the dietary restrictions in one group are not controlled for by the other group, and hence the conclusion may not be valid.” (23) Ironically, the control group added more uncertainty to the experiment than it took away.
Here is an example of this uncertainty. Most participants had high IgG levels for wheat and milk, so the test diets ended up being wheat-free and milk-free while the control diets generally contained these foods. This difference between the diets is significant because wheat and milk are known to aggravate IBS. Was the control group accidentally sabotaged by being given unfriendly foods? We just don’t know. It might be tempting to wonder whether IgG antibodies are the reason why wheat and milk aggravate IBS symptoms, but remember, there is no proof that IgG levels are related to any adverse effects. Moreover, Hunter (11) pointed out that significantly more participants in the Atkinson study had high IgG levels for milk than had been previously observed, so IgG levels are most likely not a factor in IBS.
The main lesson here is that experiments like the Atkinson study are too fraught with uncertainty for us to draw any conclusions from their results. One of the study’s investigators countered criticism of the control diets by arguing that the control diets did successfully compensate for the placebo effect because the test group improved on their diet to a statistically greater extent than the control group. (24) Statistical results are only as good as the experiment, though. The improvement of the control group may not have all been due to the placebo effect – their diet may have inadvertently removed some REAL food intolerance triggers, although we will never know because no one bothered to check the participants for food sensitivities using a DBPCFC.
Experimental flaws aside, it is also worthwhile to get a sense of just what a “26% improvement in symptoms” means for an IBS study. In a different measure of success, called the number needed to treat (NNT), the test diets in the Atkinson study performed much worse than other dietary interventions used in IBS. (11) The NNT is the number of people that need to be treated in order to find one who benefits from the treatment, calculated with respect to the control group. The NNT for this study was 9, while for most IBS diet studies, the NNT is around 2. Again, as we saw in Myth 1, diets based on IgG blood tests just don’t measure up.
In order to prove that food-specific IgG antibodies cause delayed reactions and chronic symptoms, one fundamental question would need to be answered: “do high levels of IgG against a food predict an adverse reaction to that food” (11). In debunking the myths used to justify food intolerance blood tests, we have seen that no research has provided a positive answer to that question. The evidence actually points to there being no association between IgG antibodies and adverse reactions, making IgG blood tests useless.
Some might personalize the argument against IgG-mediated food sensitivities and see it as dismissing their symptoms or delayed reactions in general. This is not true. The issue here is whether IgG blood tests are worth 500 to 1000 USD and the inconvenience, risk, and expense of modifying one’s diet – all possibly for nothing or for less improvement than could have been gained using a proper elimination diet and food challenges.
Remember that elimination diets and food challenges are already reliable means of diagnosing food sensitivities, even though spending a month or so tracking and testing your diet may not seem as attractive as a single blood test. Fortunately, the diet investigation process is not a shot in the dark – an experienced doctor or dietitian can use your personal history and your own suspicions to guide you through the process. Even though testing companies use rhetoric about ‘hidden food intolerances,’ there is usually nothing ‘hidden’ about food sensitivities at all.
Last updated February 7, 2016
1. Hochwallner H, Schulmeister U, Swoboda I, Twaroch TE, Vogelsang H, Kazemi-Shirazi L, et al. Patients suffering from non-IgE-mediated cow’s milk protein intolerance cannot be diagnosed based on IgG subclass or IgA responses to milk allergens. Allergy. 2011 Sep;66(9):1201–7. PubMed
2. Philpott H, Nandurkar S, Lubel J, Gibson PR. Alternative investigations for irritable bowel syndrome. J Gastroenterol Hepatol. 2013 Jan;28(1):73–7. PubMed
3. Carroccio A, Mansueto P, D’Alcamo A, Iacono G. Non-Celiac Wheat Sensitivity as an Allergic Condition: Personal Experience and Narrative Review. The American journal of gastroenterology. 2013;108(12):1845–52. PubMed
4. Akdis CA, Akdis M. Mechanisms of allergen-specific immunotherapy. Journal of Allergy and Clinical Immunology. 2011 Jan;127(1):18–27. PubMed
5. Stapel SO, Asero R, Ballmer-Weber BK, Knol EF, Strobel S, Vieths S, et al. Testing for IgG4 against foods is not recommended as a diagnostic tool: EAACI Task Force Report. Allergy. 2008;63(7):793–6. Available from: http://onlinelibrary.wiley.com/doi/10.1111/j.1398-9995.2008.01705.x/abstract (Archived by WebCite® at http://www.webcitation.org/6OjZbB9va).
6. Brostoff J, Gamlin L. Food Allergies and Food Intolerance: The Complete Guide to Their Identification and Treatment. Inner Traditions/Bear; 2000. 486 p.
7. Sheldon TA. Audit of the York Nutritional Laboratory Survey. Townsend Letter for Doctors and Patients [Internet]. [cited 2014 Mar 4];2002(August/September). Available from: http://webcache.googleusercontent.com/search?q=cache:QMG_H0UxZqIJ:www.drbralyallergyrelief.com/baf-report.doc
8. Hardman G, Hart G. Dietary advice based on food-specific IgG results. Nutrition & Food Science. 2007;37(1):16–23.
9. Are we missing something? 5,286 people responded to the 2007 survey of YorkTest users. Of these respondents, 75.8% indicated that they had experienced a ‘noticeable improvement’ in their condition after excluding foods from their diet. This may be true, but there are two things that are misleading about this figure. First, the definition of ‘noticeable improvement’ was expanded in the 2007 analysis to include people with only moderate improvement – this was not the case in the 2001 survey, so the 2007 results look better. More importantly, the 2007 survey does not tell us the total number of people who originally received the survey; in other words, we know how many people responded, but we don’t know how many people didn’t. When survey results are analyzed, it is important to have some idea of how non-respondents would have answered, because these people are more likely to have a negative attitude about the survey topic. As in the 2001 survey, the 2007 survey did include phone follow-ups with a sample of non-respondents, and, as expected, these people were less successful than the respondents with their diets after the YorkTest. In 2001, the inclusion of non-respondents in the analysis gave an overall success rate that was lower than the rate for respondents alone. The same would be true for the 2007 survey, but since the report did not tell us how many people failed to respond, we have no way of calculating the overall success rate. In other words, the 75.8% figure does not mean what we are led to think it means, and the true percentage would be lower.
10. Zeng Q, Dong S-Y, Wu L-X, Li H, Sun Z-J, Li J-B, et al. Variable Food-Specific IgG Antibody Levels in Healthy and Symptomatic Chinese Adults. PLoS One [Internet]. 2013 Jan 3 [cited 2014 Feb 16];8(1). Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3536737/
11. Hunter JO. Food elimination in IBS: the case for IgG testing remains doubtful. Gut. 2005 Aug;54(8):1203. PubMed
12. Hamilton RG. Relevance of (IgG anti-IgE)-IgE complexes, IgG subclass and modern IgG antibody autoanalyzers in the dying IgG reagin story. Allergy. 2009 Feb;64(2):317–8. PubMed
13. Fagan DL, Slaughter CA, Capra JD, Sullivan TJ. Monoclonal antibodies to immunoglobulin G4 induce histamine release from human basophils in vitro. J Allergy Clin Immunol. 1982 Nov;70(5):399–404. PubMed
14. Van der Zee JS, Aalberse RC. The role of IgG in immediate-type hypersensitivity. Eur Respir J Suppl. 1991 Apr;13:91s–96s. PubMed
15. Shakib F, Smith SJ. In vitro basophil histamine-releasing activity of circulating IgG1 and IgG4 autoanti-IgE antibodies from asthma patients and the demonstration that anti-IgE modulates allergen-induced basophil activation. Clin Exp Allergy. 1994 Mar;24(3):270–5. PubMed
16. Aalberse RC, Van Milligen F, Tan KY, Stapel SO. Allergen-specific IgG4 in atopic disease. Allergy. 1993 Nov;48(8):559–69. PubMed
17. Lichtenstein LM, Kagey-Sobotka A, White JM, Hamilton RG. Anti-human IgG causes basophil histamine release by acting on IgG-IgE complexes bound to IgE receptors. J Immunol. 1992 Jun 15;148(12):3929–36. PubMed
18. Jensen-Jarolim E, de Weck AL, Stadler BM. Are allergen-specific IgG mainly IgG anti-IgE autoantibodies? Int Arch Allergy Appl Immunol. 1991;94(1-4):102–3. PubMed
19. Collins AM, Jackson KJL. A Temporal Model of Human IgE and IgG Antibody Function. Front Immunol [Internet]. 2013 Aug 9 [cited 2014 Feb 24];4. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3738878/
20. Mitchell N, Hewitt CE, Jayakody S, Islam M, Adamson J, Watt I, et al. Randomised controlled trial of food elimination diet based on IgG antibodies for the prevention of migraine like headaches. Nutr J. 2011 Aug 11;10:85.
21. Atkinson W, Sheldon TA, Shaath N, Whorwell PJ. Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised controlled trial. Gut. 2004 Oct;53(10):1459–64. PubMed
22. Gut – eLetters for Atkinson et al., 53 (10) 1459-1464 [Internet]. [cited 2014 March 30]. Available from: http://gut.bmj.com/cgi/eletters/53/10/1459 (Archived by WebCite® at http://www.webcitation.org/6Ojb0PjlN).
23. Sewell WAC. IgG food antibodies should be studied in similarly treated groups. Gut. 2005 Apr;54(4):566. PubMed
24. Whorwell PJ, Bentley KJ, Atkinson W, Sheldon TA. IgG antibodies to foods in IBS. Gut. 2005 Aug;54(8):1204. PubMed