The 8-week elimination diet trial, step by step

This is the first thing worth understanding, because it saves money. Blood and saliva tests marketed as "dog food allergy tests" or "cat food allergy tests" are not a reliable way to diagnose a food allergy. Veterinary dermatology consensus is that serum testing does not correlate well with what an animal actually reacts to when fed — studies have found such tests flagging ingredients in dogs that had never eaten them, and returning positives on samples that did not come from a dog at all.

The diagnostic standard is a dietary elimination trial followed by a deliberate rechallenge. You remove everything the animal has previously eaten, wait for symptoms to resolve, then reintroduce ingredients one at a time and watch for the reaction to return. It is slow and unglamorous, and it is the only method that produces an answer you can act on.

That is also why this guide is long. The trial fails far more often from procedure than from biology.

Owners routinely stop at three or four weeks, conclude "the diet did not work," and go back to the old food. The published guidance is that a trial should run at least eight weeks, because that is the window in which more than 90% of food-allergic dogs and cats show a clear improvement. Stopping at four weeks means a meaningful share of genuinely food-allergic animals look like non-responders.

Skin takes time. An inflamed, secondarily-infected skin barrier does not settle the moment the trigger is removed — existing lesions have to heal, secondary bacterial or yeast infections usually need treating alongside, and the itch-scratch cycle has its own momentum. Improvement is typically gradual and partial before it is obvious.

Plan for eight weeks from the start and treat anything shorter as an incomplete test. If you cannot commit to eight weeks right now — a house move, holidays, anything that puts the food out of your control — it is better to start later than to run a trial you will not be able to trust.

This is where most trials quietly break. For the duration of the trial, the elimination diet must be the only thing that crosses your animal's lips. Not "mostly." Everything.

That includes the things nobody thinks of as food: flavoured heartworm and flea preventatives, flavoured or coated medications, dental chews, toothpaste, training treats, the crust of toast from a toddler, licking a plate in the dishwasher, food scavenged on a walk, and treats given by a neighbour, dog walker, or anyone who feeds the neighbourhood cats. A single beef-flavoured chewable given weekly is enough to keep symptoms going and make the whole eight weeks uninterpretable.

Two practical measures do more than anything else here. First, tell every person who has access to your animal, in writing, that a food trial is running — housemates, family, walkers, daycare. Second, switch flavoured medications to unflavoured equivalents for the trial period, which your vet can advise on.

There are two workable options, and one that looks workable but is not.

A hydrolysed diet uses protein broken into fragments small enough that the immune system generally does not recognise them. A novel protein diet uses a single protein and single carbohydrate the animal has genuinely never eaten — which, given how widely beef, chicken, dairy, lamb and fish appear in commercial food and treats, is harder to find than it sounds and depends entirely on an accurate diet history.

The option that does not work is a supermarket "limited ingredient" or "hypoallergenic" food. These labels are marketing terms, not regulated definitions. Testing has repeatedly found over-the-counter diets containing proteins not named on the label, which is fatal to a diagnostic trial: you cannot draw a conclusion from a food whose contents you cannot rely on.

This choice is your vet's call, and it depends on a complete diet history. Writing that history down before the appointment — everything the animal has eaten, as far back as you can reconstruct — makes the appointment substantially more productive.

If symptoms improve over the eight weeks, you have learned something important but incomplete: the animal improved on a new diet. You have not yet learned what it reacts to, and seasonal change or concurrent treatment could explain some of the improvement.

The rechallenge resolves that. You reintroduce the original diet and watch. If a food allergy is driving the signs, they typically return within one to three days, though it can take up to two weeks. A return of symptoms on rechallenge is what converts "improved" into "confirmed."

Owners understandably hate this step — the animal is finally comfortable and you are deliberately making it itch again. It is short, it is reversible, and skipping it means you will spend the next several years avoiding a long list of ingredients without knowing which one matters. From there you reintroduce single ingredients one at a time, holding each for one to two weeks, to build a specific list rather than a vague one.

1. 1

   Weeks 0–1 · Transition

   Move gradually onto the elimination diet. Remove every other food source, including flavoured medications and dental chews.

2. 2

   Weeks 1–8 · Strict elimination

   The diet is the only thing consumed. Log meals and symptoms daily — improvement is gradual and easy to miss without records.

3. 3

   Week 8 · Assess

   Compare symptom scores against week 1. Meaningful improvement means proceed; no change means discuss other causes with your vet.

4. 4

   Weeks 8–9 · Rechallenge

   Reintroduce the original food. Signs typically return in 1–3 days, sometimes up to two weeks. Return = confirmed.

5. 5

   Weeks 9+ · Single-ingredient testing

   Reintroduce one ingredient at a time, 1–2 weeks each, to identify the specific trigger rather than avoiding everything forever.

Eight weeks is long enough that recall becomes unreliable, and the improvement you are looking for is gradual. Owners frequently under-report progress because they are comparing against yesterday rather than against week one — the itch is a little better every week and never feels like a milestone.

The useful minimum is a daily line: what was eaten (including treats and medications), a scratch or itch score out of ten, and any ear, paw, skin or digestive signs. Photographs of affected areas once a week are worth more than any written description when you sit down with your vet, because skin change is visual and slow.

Bring the log to the follow-up appointment. A vet looking at eight weeks of daily scores can tell in seconds what a verbal summary takes ten minutes to approximate badly.