Are IgG food sensitivity tests accurate?

By Jason · Updated August 2026

What the test actually measures

You send blood. A laboratory measures immunoglobulin G antibodies against a panel of typically 100 to 200 foods, sometimes reporting the IgG4 subclass specifically. You get a report sorting foods into red, amber and green, or into numerical reactivity scores, usually with a suggested elimination plan.

Everything in that process works. The laboratory really is detecting antibodies, the numbers are real measurements, and the report is internally consistent. The failure is not technical. It is that the thing being measured does not mean what the report says it means.

Food-specific IgG is found in healthy people. It is a normal, expected consequence of eating. Your immune system encounters food protein constantly across the gut wall and produces antibodies to it as a matter of routine. Finding IgG against a food is roughly as informative as finding that you have eaten it.

The immunotherapy problem, which is the whole argument

There is a reason immunologists are unusually confident about this, and it is worth understanding because it is much stronger than "the evidence is weak".

Consider allergen immunotherapy: the months-long desensitisation given to people with genuine IgE-mediated allergies, to pollen or venom or peanut. It works, and it is well studied. As treatment progresses and patients become able to tolerate the allergen, their allergen-specific IgG4 rises.

That rise is not incidental. IgG4 is understood to be one of the mechanisms of the cure. It acts as a blocking antibody, competing with IgE for the allergen and preventing IgE from cross-linking on mast cells and basophils, which is the step that triggers an allergic reaction. Its production is driven by interleukin-10 alongside regulatory T cells, the same tolerance-promoting machinery that makes the therapy work. IgG4 is generally described as an anti-inflammatory antibody with limited ability to provoke an immune response.

So the picture is this. In the one setting where we deliberately and measurably convert a food or allergen reaction into tolerance, IgG4 goes up as the person gets better. The EAACI position follows directly: IgG4 against a food indicates repeated exposure and is a marker of immunological tolerance, not of hypersensitivity.

A test that reports high IgG4 to bread as evidence you should stop eating bread has the sign backwards.

Why the results feel uncannily accurate

This is the part that keeps the industry alive, and dismissing it is a mistake, because people are genuinely observing something.

Since IgG rises with exposure, the foods flagged tend to be the foods you eat most. For most people in Western diets that means wheat, dairy, eggs, yeast, and whatever else appears daily. Read a report listing those and it feels like recognition, because it is: the test has described your diet.

Then two further effects compound it. Eliminating wheat and dairy removes a large fructan and lactose load, so a meaningful share of people genuinely improve, for reasons that have nothing to do with IgG. And people typically buy the test during a bad stretch, which is followed by a better one regardless, the regression to the mean problem covered on the supplements page.

Real improvement, correctly observed, misattributed to the mechanism. That combination is far more persuasive than any marketing claim.

The strongest evidence in favour, taken seriously

Proponents cite one study more than any other, and a page that ignores it is not worth reading.

Atkinson and colleagues published a randomised controlled trial in Gut in 2004. IBS patients were assigned either a diet eliminating foods their IgG test had flagged, or a sham diet eliminating the same number of foods. At 12 weeks the true diet produced a 10% greater reduction in symptom score, a difference that reached statistical significance.

That is a real trial with a real positive result and it should not be waved away. Here is why it does not establish what it is used to establish.

• The two diets were not comparable. Correspondence published in Gut afterwards made the central point: the arms excluded significantly different foods. Matching the number of excluded foods does not match their effect. The IgG-guided diet disproportionately removed milk and wheat, which are among the most common symptom triggers in IBS for entirely separate reasons.
• The effect is consistent with a simpler explanation. Removing lactose and fructans helps a substantial fraction of people with IBS. The trial is compatible with IgG having done nothing except point at commonly eaten foods, which is what IgG does.
• Blinding a diet is extremely difficult. Participants can tell which foods they have been asked to give up, and expectations affect symptom reporting strongly in gut disorders.
• It has not been convincingly replicated in a design that fixes the comparison problem, in more than twenty years.

The honest summary is that the best trial in favour of IgG testing demonstrated that eliminating milk and wheat helps some people with IBS. Nobody disputes that. It is a different claim from the one the test makes.

What the professional bodies say

Four continents, six independent bodies, none of which sells a competing product. That degree of agreement is rare in medicine, and it is the closest thing to a settled position this site reports on any test.

Related tests with the same problem

IgG panels are the largest category but not the only one. These share the pattern of a real measurement interpreted in a way that has not been validated:

• ALCAT, and leukocyte activation or cytotoxic tests. Measure changes in white blood cells exposed to food extracts. Different technology, same absence of validation against clinical outcomes, and poor reproducibility on repeat testing.
• Mediator release tests (MRT / LEAP). Proprietary, and the underlying method has not been independently validated.
• Hair analysis for food intolerance. Hair contains no relevant immunological information about food reactions.
• Applied kinesiology and electrodermal or "bioresonance" testing. Tested under blinded conditions and found not to distinguish real from sham samples.

What is genuinely validated, and where IgG is legitimate

Being fair to the antibody, since the criticism is of a use rather than of the molecule.

IgG4 measurement is a real and clinically useful tool in the right place. It is used to monitor the immunological response to allergen immunotherapy, and IgG4 concentrations matter in the diagnosis of IgG4-related disease, a genuine and serious condition. Food-specific IgG also has legitimate uses in research. The objection is specifically to selling it as a way of finding out which foods to stop eating.

For food-related symptoms, these are the tests with actual diagnostic performance behind them:

• Coeliac serology (tTG-IgA), with total IgA. Validated, and it must be done while you are still eating gluten. Cutting gluten first is the single most common way people make their own diagnosis impossible.
• Specific IgE testing and skin prick testing for suspected true allergy, interpreted alongside your history rather than as a standalone panel. Component-resolved diagnostics can refine this further.
• Hydrogen breath testing for lactose and fructose malabsorption.
• Faecal calprotectin to separate inflammatory disease from functional symptoms. See which gut tests are worth it.
• Structured elimination and planned reintroduction, ideally with a dietitian. Slower than a blood test, and it is the method that actually identifies food triggers.

If you have already taken one

A reasonable position, given that a lot of people arrive here after the fact.

1. Do not treat the list as a verdict

   Especially not a long one. The number of flagged foods reflects how varied your diet is more than how reactive you are.

2. Keep any genuine observations

   If eliminating something clearly helped, that observation is worth keeping even though the test that suggested it was not valid. It just needs confirming properly.

3. Reintroduce systematically

   One food at a time, spaced, logged, and compared against your usual symptom rate rather than your memory. Most people find that the majority of the flagged foods are fine, and get a large part of their diet back.

4. Get the validated tests done

   Particularly coeliac serology, and particularly before you cut gluten.

That confirmation step is exactly what Minthe is built for: it establishes your personal base rate first, schedules spaced challenges with washout, and declines to report a result when the data cannot support one. The method is published in full.

This page carries no affiliate links, and neither does any page on this site.

Minthe is a wellness and self-tracking tool, not a medical device. It does not diagnose or treat any condition and isn't a substitute for professional medical advice. If you have red-flag symptoms such as blood in your stool, unintentional weight loss, difficulty swallowing, persistent vomiting or a fever alongside gut symptoms, see a clinician rather than tracking them.

Sources

• Stapel SO, et al. Testing for IgG4 against foods is not recommended as a diagnostic tool: EAACI Task Force Report. Allergy, 2008;63(7):793-796.
• American Academy of Allergy, Asthma & Immunology. Choosing Wisely: unproven diagnostic tests in the evaluation of allergy.
• Atkinson W, Sheldon TA, Shaath N, Whorwell PJ. Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised controlled trial. Gut, 2004;53(10):1459-1464.
• Correspondence: Food elimination in IBS, the case for IgG testing remains doubtful. Gut, 2005.
• Shamji MH, et al. The role of IgG1 and IgG4 as dominant IgE-blocking antibodies shifts during allergen immunotherapy. Journal of Allergy and Clinical Immunology, 2023.
• ASCIA, BSACI, Choosing Wisely Canada and SIAAIC position statements on unproven food allergy and intolerance testing.

Frequently asked