How to tell if your dog is allergic to its food: two steps and four possible trial outcomes

A food allergy is proven by two steps, and only together: on the new diet the signs go away, and after the old food is brought back they return. The 2015 ICADA guidelines consider dietary involvement confirmed only after such a flare, and according to Olivry and Mueller (2020), 90% of these flares occur within 14 days. The second step is rarely taken: in a Canadian survey, 8 owners out of 77, or 10%, brought back the old food (Tiffany et al., 2019).

Key points

• Only a flare after the old food is brought back counts as proof (ICADA, 2015; CAT 6, 2018).
• Half of flares occur within 5 days, 90% within 14 days: 234 dogs, 315 challenges (Olivry, Mueller, 2020).
• After improvement, 8 owners out of 77 (10%) brought back the old food, and 33% discussed this step with their vet (Tiffany et al., 2019).
• In a double-blind trial, 6 of 12 allergic dogs “reacted” to a portion containing no test ingredient at all (Sofou et al., 2026).
• Partial improvement is also a result: 33–57% of dogs with a food allergy have another itchy disease as well (Loeffler, 2004 and 2006; Lewis et al., 2025).

What actually proves the food is to blame?

A flare after the old food is brought back. The ICADA guidelines (Hensel et al., BMC Veterinary Research, 2015) put it in one line: “Dietary involvement is confirmed if there is a relapse of clinical disease when the original diet is re-introduced.”

See also: Dandruff in dogs

A challenge means bringing back the old food after the signs have gone away. The 2025 FEDIAF guidelines require the same in section 7.6.4: “ideally this should be confirmed by a challenge” after the clinical signs have disappeared.

A food trial means at least eight weeks on a novel or hydrolyzed protein, with the dog getting nothing else. Eight weeks give you half the answer. The other half comes from a bowl of the old food.

Why does “better on the new food” prove nothing yet?

Dogs without a food allergy also get better on a new food. In Lewis and colleagues (2025), 57 itchy dogs went through two eight-week trials, and dogs with atopy and no food component also improved: because of the change of season, omega-3 in the new food, or temporary medication. The CAT 6 criteria (Mueller, Olivry, 2018) classify improvement without bringing back the previous food as weak evidence.

Take a 12 kg French bulldog that had been licking its paws for six months. The owners switched on their own from a chicken food to a lamb and rice food, the itching settled after five weeks, and they decided chicken was to blame.

There is no way to test this conclusion. The five weeks on the new food coincided with the end of the season, and the new food has a different fat: without bringing back the old bowl, these three causes cannot be told apart.

How long to wait after bringing back the old food?

Up to 14 days. In the CAT 9 series (Olivry, Mueller, 2020), covering 234 dogs and 315 challenges, half of flares occurred within 5 days and 90% within 14 days.

The same picture appears in the survey by Tiffany and colleagues (2019): in all 8 owners who brought back the old food, the signs returned, and in four dogs this happened within the first three days.

Fourteen days is the threshold after which the result is read. When to bring back the food and what counts as a flare is up to the vet; your job is not to skip this step.

What does each of the four trial outcomes mean?

There are only four outcomes.

Our own chart based on primary sources, 2026-09-30. The highlighted row is the bulldog from the example: until the old food is brought back, there is no knowing which of the first two rows it belongs in.

Can you find out at home which ingredient your dog is allergic to?

Not reliably. In a double-blind trial by Sofou and colleagues (2026), 12 dogs with atopy and a confirmed food reaction were given 40 g of a test ingredient per day for a week: eight ingredients and placebo twice, meaning the same diet food mixed with water. Reactions were seen in 11 of 12 dogs to one to six ingredients, but 6 of 12 also “reacted” to placebo.

Data from Sofou et al., Veterinary Dermatology, 2026; per-challenge shares are our own calculation, 2026-09-30.

The authors attribute this to fluctuations in the severity of atopy and to the owner's subjective assessment: the difference between ingredients and placebo was not significant. A home “chicken test” shows roughly the same as a portion without chicken.

The full old food caused a flare in all 12 dogs: the whole bowl works as proof, a single scrap from the table does not.

What does it mean if the itching has eased but not gone away?

Partial improvement means the food is one of the causes of the itching. 33–57% of dogs with a food allergy have another itchy disease as well, most often atopy: 9 of 18 in Loeffler and colleagues (2004), 11 of 21 and 20 of 35 in their 2006 study, 7 of 21 in Lewis and colleagues (2025).

In these dogs, the new diet removes only part of the itching. ICADA warns that the owner may not notice a partial response at all, so the trial is continued until the vet examines the dog. In Lewis 2025, skin scores in 4 of 21 dogs even rose during the first four weeks.

Why does only one owner in ten take this step?

Bringing back the old food means a deliberate flare, which is why it is rarely done. In the survey by Tiffany and colleagues (2019) of 93 owners, 8 of 77 brought back the old food, 33% discussed this step with their vet, and for “diagnosis” 38% had blood tests and 3% saliva tests (why these tests don't work is covered in the article on itching without fleas).

Not everyone finishes the trial either: 10 of 28 owners abandoned a home-cooked diet halfway (Tapp et al., 2002), 17 of 63 dropped out in Loeffler 2004, 10 of 57 in Lewis 2025. In Udraite Vovk and colleagues (2026), two of 11 owners refused further challenges.

The cost of the second step shows up in grams and money.

How much food will an eight-week trial take?

A 12 kg bulldog with normal activity needs 722 kcal a day using the feeding-by-weight method. The grams depend on the calorie content of the diet food: 206 g at 350 kcal per 100 g, 182 g at 397 kcal and 172 g at 420 kcal. Over the 56 days of the trial, that is 9.6–11.6 kg of food.

DELICATE is a dry food from Dnipro for adult dogs: 397 kcal per 100 g (calculated per NRC 2006) and, per the label, 32% crude protein, and a 10 kg bag costs UAH 1,188 with delivery. For a 12 kg bulldog, that is 182 g and about UAH 22 a day, and a bag lasts 54 days. The range is not suitable for the food trial itself: ICADA requires a novel or hydrolyzed protein, and chicken dog food is fine once chicken has been tested.
See the food — UAH 1,188

What to do with your dog?

• If the signs went away on the new food, ask your vet to plan bringing back the old food and wait up to 14 days: that flare is exactly what will serve as proof.
• If nothing came back within 14 days on the old food, keep looking.
• If things got better but the itching remains, tell your vet: a partial response means a second disease alongside, not a failed trial.
• If you want to find a specific ingredient, do it only with a vet and with a repeat, because even a portion with no test ingredient produced a “reaction” in 6 dogs out of 12.
• If you need a portion size for the trial period, enter the weight in the food calculator and divide the kilocalories by the calorie content of the bag: for the bulldog, that is 722 kcal and 181 g at 400 kcal per 100 g.

Questions and answers