How an elimination diet is supposed to work, where it has real clinical backing, and the test you should not use to decide what to cut.
An elimination diet is a structured experiment rather than a way of eating. You remove a defined list of foods for a few weeks, then bring them back one at a time and watch what happens. Done properly it produces something genuinely useful: a personal, tested list of what does and does not agree with you.
Done on the basis of a commercial sensitivity test, it produces a long list of foods you did not need to avoid. This page has been rewritten around that distinction, and one risky variant has been removed from it.
Key Takeaways
- Food IgG sensitivity panels are not a valid way to choose what to eliminate (1).
- The strongest clinical use is eosinophilic oesophagitis, where it is standard therapy (4).
- For IBS, the diets with trial evidence are specific ones, not a general elimination (3).
- Test for celiac disease before removing gluten, not after (8).
- We have removed the water-only fasting variant that used to be listed here (9).
How It Works
SaveTwo phases, and neither is optional.
Elimination runs two to six weeks. The suspected foods come out completely, because a partial removal gives you a result you cannot interpret. You plan replacements so the diet still covers your nutrient needs, and you write down what you eat and how you feel.
Reintroduction is the phase that produces the answer. One food group at a time, in a normal portion, with a few days before the next, watching for the symptoms that sent you here in the first place. No reaction, and the food goes back on the list. A reaction, and you have a candidate.
The food diary is what makes this work. Without a written record you are comparing a memory of last Tuesday against how you feel today, which is how people end up avoiding things that never caused a problem.
The Part Most Articles Get Wrong
Most people arrive at an elimination diet holding a printout from a food sensitivity test, usually an IgG panel covering 90 to 100 foods, sometimes far more.
The AAAAI’s position on those panels is blunt. The test has never been scientifically proven to do what it reports to do, the studies offered in support are often out of date or published in non-reputable journals, and the presence of IgG is likely a normal response to eating a food. In fact, higher levels of IgG4 to foods may simply be associated with tolerance of those foods. The AAAAI and the Canadian Society of Allergy and Clinical Immunology both recommend against using IgG testing to diagnose food allergy or intolerance, and it appears on the Choosing Wisely list of unproven tests (1).
That is worth sitting with, because it inverts the premise. A panel that flags 25 foods may be describing the foods you eat most often.
There is also no single test that identifies food sensitivities (1). The elimination-and-reintroduction method is a reasonable answer to that gap. Choosing the list from a panel is not.
Allergy is a different question
A food allergy is an immune response and can be dangerous, up to anaphylaxis, which is a medical emergency (2). An intolerance is a digestive problem: uncomfortable, not dangerous (1).
If a food has ever caused swelling, hives, breathing difficulty or a sudden drop in blood pressure, that is not a candidate for a home reintroduction. It is a reason to see an allergist before you plan anything.
Where It Has Real Backing
Save| Where it is used | What the evidence looks like | Verdict |
|---|---|---|
| Eosinophilic oesophagitis | Dietary elimination is an established therapy here, used to identify the trigger food and return the person to as normal a diet as possible. Elemental, allergy-testing-directed and empiric versions have all been studied (4) | Established clinical use |
| Irritable bowel syndrome | A network meta-analysis of 28 randomised trials in 2,338 patients ranked a low FODMAP diet fourth of the interventions studied more than once for global symptoms, and it was the only diet superior to habitual eating for bloating (3) | Good evidence, specific diets |
| Food allergy and intolerance, in general | Elimination and supervised reintroduction is the method. The trigger list should come from proper assessment, not from a commercial sensitivity panel (1, 2) | Method is sound, the inputs matter |
| ADHD | A review of the field concluded that elimination diets produce a small but reliable aggregate effect, while calling for updated trials and noting that identifying who responds is still unresolved (7) | Small effect, needs supervision |
| Migraine | The trial usually cited is a crossover study in 21 patients whose exclusions were chosen by an IgG panel against 270 foods (6) | Weak, and rests on IgG testing |
| Skin conditions such as acne and psoriasis | We have removed this claim. It was not supported by trial evidence of elimination diets in those conditions | Claim removed |
The clearest case
In eosinophilic oesophagitis, dietary elimination is not an alternative approach, it is one of the main treatments. Its effectiveness is part of how the condition is understood. Studies have shown symptomatic and histological response in adults comparable to the response reported in children, and the explicit long-term goal is to identify a small number of trigger foods so the person can return to as regular a diet as possible (4).
That last phrase is the model for the whole method: restriction in order to end restriction.
IBS: pick the right diet, not a general purge
For irritable bowel syndrome the evidence favours particular, well-defined diets. A 2025 network meta-analysis pooled 28 randomised trials in 2,338 patients across 11 dietary interventions. A starch-reduced and sucrose-reduced diet ranked first for global symptoms, though on only two trials; a low FODMAP diet ranked fourth across 24 trials, and for bloating or distension it was the only diet that beat habitual eating. For bowel habit, no dietary intervention was superior to control (3).
The low FODMAP diet is itself a phased elimination with structured reintroduction, and it has the trial base that a general elimination diet does not.
Migraine and IBS: read the fine print
The two studies most often cited for elimination diets in migraine and IBS both selected the excluded foods using IgG antibody panels.
In the IBS trial, 150 patients were randomised to exclude either the foods they had raised IgG antibodies to or a sham list of the same size. The true diet produced a 10% greater reduction in symptom score than the sham at 12 weeks (5). The migraine study was a crossover trial in 21 patients, with exclusions chosen from a panel of 270 foods (6).
These are real published trials, and they are also small, and built on a test that allergy bodies advise against (1). That is why this page no longer presents migraine relief as an established benefit.
The Variant We Removed
The previous version of this article listed a fasting version, nothing but water for five days before food was slowly added back.
We have taken it out. Restoring nutrition after a period of starvation can trigger refeeding syndrome, a shift in fluids and electrolytes with low blood phosphate as its hallmark, causing cardiac, respiratory and neurological complications, carrying a high mortality rate and requiring immediate medical treatment (9). Starvation itself is one of the named risk factors.
A five-day water fast is not an elimination diet. It belongs in a hospital protocol or nowhere.
Test for Celiac First
If gluten is on your suspect list, the order matters. NIDDK states that doctors do not recommend starting a gluten-free diet before diagnostic testing, because a gluten-free diet can affect the test results (8).
Someone who removes gluten, feels better, and never gets tested has not ruled celiac disease in or out. They have made it harder to find, and celiac disease is a diagnosis with lifelong consequences for how they need to eat.
Doing It Without Doing Harm
- Get the suspect list from a clinician, not from a panel (1).
- Keep it short. Two to six weeks, then reintroduce. The end date is part of the design.
- Replace what you remove. Cutting dairy, gluten, eggs and nuts at once removes a large share of your calcium, fibre and B vitamins.
- Skip it entirely if you have any history of disordered eating. A forbidden-foods list plus daily symptom tracking is the wrong combination.
- Children need supervision. Restriction during growth is not a home experiment.
- Never reintroduce a food that has caused an allergic reaction outside a medical setting (2).
If your symptoms are digestive, the low FODMAP diet is the version with the trial evidence behind it. The autoimmune protocol is a stricter relative aimed at autoimmune conditions, the arthritis diet guide covers joint symptoms, and the fibre guide helps you keep intake up while the allowed list is short.
Frequently Asked Questions
What is an elimination diet?
A short-term, structured experiment. You remove a defined set of foods for a few weeks, then reintroduce them one at a time and record what happens. The goal is a personal list of triggers, not long-term restriction.
Should I use a food sensitivity test to decide what to cut?
No. The AAAAI says the food IgG panel has never been scientifically proven to do what it claims, and that higher IgG4 levels to a food may simply reflect tolerance of that food. The AAAAI and the Canadian Society of Allergy and Clinical Immunology both recommend against using IgG testing to diagnose food allergy or intolerance.
What is the difference between a food allergy and an intolerance?
An allergy is an immune response to the food and can be dangerous. An intolerance means the body is not digesting or processing the food properly, which is uncomfortable rather than dangerous. That distinction decides whether you are dealing with a nutrition question or a medical one.
Can I test myself for celiac disease this way?
No, and doing so causes a problem. NIDDK advises against starting a gluten-free diet before diagnostic testing, because it affects the results. If gluten is one of your suspects, get tested while you are still eating it.
How long should it last?
The elimination phase usually runs two to six weeks, followed by reintroduction over several more weeks. It is designed to end. An elimination phase that stretches on for months is no longer a diagnostic tool, it is an unsupervised restrictive diet.
The Bottom Line
An elimination diet is a good method pointed at a hard question, and its value depends entirely on how the suspect list is chosen. Allergy bodies advise against the IgG sensitivity panels most people use for that, on the grounds that the test is unproven and that IgG to a food may indicate tolerance rather than trouble. Where the method has genuine clinical standing is eosinophilic oesophagitis, and where the trial evidence sits for gut symptoms is in specific diets such as low FODMAP rather than a general purge. Keep it short, keep a written record, test for celiac disease before you remove gluten, and do the reintroduction phase, because that is the only part that produces an answer.
This article is for informational purposes only and is not a substitute for professional medical advice. Consult a doctor or registered dietitian before starting an elimination diet, and see an allergist rather than self-testing if a food has ever caused an allergic reaction.
Related Articles
FODMAP Diet: The 3 Steps and the List Errors
AIP Diet: What the Evidence Shows and Who Should Skip It
Arthritis Diet: Foods That Help, Foods to Limit and Evidence
High Fiber Foods List and How Much Fiber You Need