September 29, 2026 4 min read

You leave the appointment with a bag of prescription food and one instruction: feed only this, for eight weeks.
Then, six weeks in, you remember the flavoured heartworm chew you hand over on the first of every month. Or the kids have been sharing crusts. Or you used a pill pocket to get the antibiotics in, because how else were you going to do it.
The trial is gone. Not damaged; gone. You start again.
This is not a failure of attention. An elimination diet for dogs falls over this way constantly, because the diet and everything else your dog swallows get managed in two separate conversations.
Your dog eats one thing, and only that thing, for a set period. The diet is either hydrolysed, where proteins are broken into fragments the immune system is less likely to react to, or novel-ingredient, built on a protein your dog has never eaten. VCA Animal Hospitals sets out the split: a hydrolysed diet suits dogs whose full food history is unknown, while a novel-ingredient diet makes more sense when you can account for everything your dog has eaten.
The proteins worth ruling out are usually the familiar ones. A review by Mueller, Olivry and Prélaud pooled the reported allergens across 297 dogs with confirmed reactions: beef came first, then dairy, chicken, wheat and lamb. Read that list again and notice it is the ingredient deck of the average supermarket food.
If the signs settle, that is not yet a diagnosis. Confirmation comes from the rechallenge: the old food goes back in, signs return, the elimination diet goes back on, signs settle again. As Today's Veterinary Practice explains, this remains the most reliable tool available, and serum, saliva and hair tests are unreliable enough to cause a misdiagnosis.
Longer than most people are told at the desk.
A critically appraised topic by Olivry, Mueller and Prélaud in BMC Veterinary Research reviewed the evidence on duration. By five weeks, more than 80% of dogs had achieved remission of clinical signs. Extending to eight weeks lifted that above 90%.
Different signs also clear at different speeds. The Purina Institute notes that gut signs often improve within one to four weeks and skin signs within four to eight, but complete resolution of skin signs can take up to 12 weeks. That is why most vets quote eight to 12 weeks rather than a flat eight.
So a five-week trial is not a shorter version of the same test. It is a test that misses roughly one in five dogs, and the miss looks exactly like a negative result. Shortened protocols using oral prednisolone can compress things in suitable dogs, so if your vet has given you a different timeframe, follow theirs.
Only ones your vet has approved, and for most trials that means none from outside the diet.
Clinical guidance is unambiguous here. VCA is blunt that allergens hide in treats, rawhide chews, supplements, toothpastes and oral medications, because those products are usually flavoured with protein. Chewable joint supplements count. Flavoured toothpaste counts. Flavoured monthly parasite preventives are usually flavoured with pork, soy, beef or chicken, so they need swapping for an unflavoured oral or topical alternative for the duration.
And the one that catches almost everybody: a scrape of peanut butter, a corner of cheese, one pill pocket. Owners assume a small amount will not register. It can be enough to trigger a reaction, and it will interfere with reading the result. If medication has to go down, use a small amount of the elimination diet itself.
Two things to plan for. Other pets: free-fed bowls and dropped kibble count, so feed separately at set times. Other people: the trial only works if everyone with access to your dog knows it is running.
Start with length. A single ingredient treat has one place for something to go wrong. A twelve-ingredient treat has twelve.
Then read for vague terms, because that is where undeclared proteins live. "Animal fat", "meat meal", "digest" and "natural flavour" do not name the animal, and if the animal is not named you cannot rule it out.
Then the uncomfortable part: labels are not always complete. A PCR study published in Veterinary Dermatology tested commercial products and found DNA from at least one undeclared animal species in nine of 10 over-the-counter diets. A later real-time PCR analysis of commercial dry dog foods found undeclared chicken DNA in 65% of products and undeclared pork in 41%. Those studies looked at foods, not treats, but the same caution applies to anything else going in the bowl.
Ask a manufacturer three things: which animal and which cut, what else runs on the same equipment, and whether the person answering actually makes it.
Not us, and not the internet. Your vet chooses the diet, sets the duration, treats the concurrent ear and skin infections that would otherwise muddy the result, runs the rechallenge and interprets it. We can tell you exactly what is in ours. That is the extent of what we are qualified to say.
If your vet has approved a specific protein, single ingredient treats are the only kind you can genuinely verify. Our air-dried chicken is Australian chicken breast. Our freeze-dried kangaroo is kangaroo. All of it is made in our own Bendigo kitchen, so when you ask what is in it, you get an answer from the person who made it.
If you want to see what a readable label actually looks like, the whole single protein treats collection is one animal per bag: air-dried or freeze-dried, nothing added, no coatings, no binders. Take the ingredient list to your vet and let them tell you which ones fit your dog's plan.
Eight weeks is a long time to give something. It is a much shorter time than eight weeks plus another eight weeks.
General information, not veterinary advice. Elimination diet trials should be run under veterinary supervision. Do not start, stop or alter a diagnostic diet without talking to your vet.
Comments will be approved before showing up.











