He’s been on six foods. Grain-free, salmon, hypoallergenic, one with insect protein, and two that cost more than your own dinner. He’s still itchy, or still loose, or both.
Somewhere in that process the words “food allergy” got used, and from then on every decision was made on that basis. Quite often it was the wrong basis, because three separate things get bundled under the same heading and only one of them is fixed by changing the food.
Three different problems
Food allergy is an immune reaction to a specific protein. The immune system has decided a harmless food component is a threat. It’s non-seasonal, it’s consistent, and it often produces both digestive and skin signs together. Small amounts can trigger it, because immune reactions aren’t proportionate to dose.
Food intolerance isn’t immune at all. The dog simply can’t digest something properly — lactose being the classic. It produces wind, loose stools and sometimes itching, and it is dose-dependent: a little causes a little trouble, a lot causes a lot.
Environmental allergy, or atopic dermatitis, is a reaction to pollen, grasses, dust mites or storage mites. This is by a wide margin the most common allergic skin disease in dogs. It’s often seasonal at first and becomes year-round. And critically: changing the food will not fix it.
That third category is where most of the wasted years go. Owners work through eight proteins for a dog reacting to the grass in the garden.
There’s a fourth possibility that gets overlooked entirely, which is that the gut population itself is the problem rather than any specific ingredient — a dog who reacts to everything sometimes has low-grade dysbiosis rather than eight separate intolerances. That’s one of the situations where daily Probiotics for Dogs alongside a consistent diet is worth a proper four to six week trial before you start eliminating proteins, because it’s cheaper, faster and occasionally the whole answer.
Which foods, and why
The proteins most often implicated are chicken, beef, dairy, wheat, lamb, egg, soya and fish.
Worth understanding why that list looks like that. It isn’t that chicken is inherently more allergenic than kangaroo. It’s that allergy requires prior exposure, and those are the proteins dogs have eaten most. A dog cannot be allergic to something he’s never encountered.
Which is also why “novel protein” diets work — not because venison is hypoallergenic, but because his immune system has never met it.
And why grain allergy is much rarer than the market implies. Genuine wheat allergy exists and is uncommon; the great majority of food-allergic dogs react to an animal protein.
The tests that don’t work
Blood and saliva tests for food allergy are widely sold, widely advertised online, and regarded by veterinary dermatologists as unreliable.
Studies repeatedly show poor agreement between these tests and actual clinical reactions — including tests returning positive results for foods the dog has never eaten, and for samples taken from healthy dogs and even from tap water in one memorable case.
They cost real money and they routinely send owners down the wrong path for months. If someone offers a hair or saliva test that will identify your dog’s food allergies, that’s not a diagnostic tool.
Skin testing has a legitimate role for environmental allergens. Not for food.
The elimination trial, which is the only real test
Not glamorous, and it’s the whole diagnostic method.
Choose the diet with your vet. Either hydrolysed — proteins broken into fragments too small for the immune system to recognise — or a genuinely novel protein based on his actual dietary history. Over-the-counter “hypoallergenic” foods are usually unsuitable, because manufacturing cross-contamination is common and the ingredient list may not reflect trace content.
Eight to twelve weeks. Digestive signs improve in two to three weeks. Skin takes eight or more, because that’s how long the skin takes to reflect any dietary change.
Nothing else. At all. No treats, no chews, no dental sticks, no table scraps, no flavoured medication, no licking plates or high chairs, no training treats, no toothpaste with flavour in it.
Then challenge. Reintroduce the original food and watch for signs returning, usually within two weeks. This confirms it. If nothing returns, the improvement wasn’t the diet.
The challenge is the step nearly everybody skips, and skipping it means you finish twelve weeks with a strong suspicion rather than an answer.
Why trials fail
Almost never because the diet was wrong.
A flavoured wormer. A dentastick because his breath was bad. Half a sausage from a grandparent who thought once wouldn’t matter. Licking a toddler’s fingers. A treat at training class. The cat’s food. Scavenging on a walk.
If you’re doing this properly: tell everyone in the household and write it on the fridge, ask your vet for unflavoured medication where possible, use the trial food itself as treats, keep him on lead where he scavenges, and feed the cat somewhere he can’t reach.
And tell your vet about slips honestly. A trial with three lapses tells you nothing, and pretending otherwise costs you three months.
Grain-free, and what’s actually known
Worth being straightforward, because the marketing is loud in both directions.
Grain allergy is uncommon. Most grain-free foods replace the grain with peas, lentils or chickpeas, which are highly fermentable and make a good number of dogs windier and looser than they were on rice.
There’s also an open question about diet-associated dilated cardiomyopathy, a heart condition that was investigated in relation to grain-free and legume-heavy diets. The picture is genuinely unresolved — no causal mechanism has been established, and the research continues. It isn’t a reason for alarm and it is a reason not to assume grain-free is automatically the better choice.
If your dog does well on a grain-free food, there’s no need to change. If you’re choosing one on the assumption that grains are the problem, that assumption is probably wrong.
How long each sign takes to respond
Useful to know before you start, because mismatched expectations are why trials get abandoned early.
Digestive signs — loose stools, wind, vomiting — usually improve within two to three weeks if food is the cause. This is the fastest indicator you’ll get.
Itching and scratching take longer. Four to six weeks before you’d expect a meaningful change.
Skin condition and coat quality are slowest of all, at eight to twelve weeks, because that’s simply how long the coat takes to reflect any dietary change. Judging a food trial on coat at week four tells you nothing.
Ear infections, which are frequently the main sign of food allergy in dogs, often improve alongside the skin rather than the gut.
So a dog whose stools firm up by week three but who’s still itchy at week five is not a failed trial. He’s a trial in progress.
Score it rather than eyeballing it. A number out of ten for itching, written down once a day, plus a stool score and a weekly weight. Photograph the affected skin monthly. Skin changes slowly enough that you genuinely cannot judge it by memory, and a good proportion of dogs written off as non-responders were improving perfectly well.
Treats during a trial, without going mad
The most common practical obstacle, and it’s solvable.
Use the trial food itself. Set the day’s ration aside in the morning and use it as treats. If it’s a wet or tinned diet, spread it thin on a baking sheet, bake it low until dry and cut it into squares — that gives you a legitimate training treat that can’t break the trial.
Ask about matching treats. Most prescription hypoallergenic ranges include a treat version, and it’s the simplest answer if you can get it.
For chewing, which dogs genuinely need, ask your vet rather than assuming there’s nothing. A stuffed and frozen Kong filled with the trial food does the job for most dogs.
Tell training class. Instructors are used to this and will let you use your own food.
What doesn’t work is deciding one dentastick a day won’t matter. It will, and you’ll have wasted the twelve weeks.
Reading labels properly
“With chicken” means almost nothing. Under UK and EU labelling rules, a food described that way needs only a small percentage of it. “Rich in chicken” means more. Neither tells you a food is free of anything else.
Look for what isn’t there. For a trial you need a full ingredient list and ideally a manufacturer statement about cross-contamination.
Watch the hidden animal proteins. Chicken fat, animal digest, chicken meal, unspecified “meat and animal derivatives”, and flavourings. Any of them can carry the protein you’re avoiding.
Check the treats and chews too. Rawhide, pizzles and dentasticks all contain something.
After a diagnosis
Manageable, and usually for life.
Once you know the trigger, you avoid it. That’s easier than the trial, because you can use any food that doesn’t contain it rather than being locked to a prescription diet.
Some dogs tolerate small amounts and some don’t, and finding out is a matter of careful experiment rather than assumption.
Keep a written list of what he reacts to and give a copy to anyone who looks after him — kennels, sitters, family. And check the ingredients every time you buy, because manufacturers reformulate without notice and a food that was safe last year may not be.
When it isn’t food
If a properly conducted trial produces no improvement, food is probably not the answer, and the useful thing is to stop looking there.
Next questions: has flea control been rigorous for every animal in the house for three months? Have skin scrapes and cytology been done to check for mites, yeast and bacterial infection? Does the seasonal pattern suggest environmental allergy? Has anyone tested for parasites with pooled faecal samples?
Environmental allergy is managed rather than cured, and there are genuinely effective modern treatments — considerably better than steroids for long-term use. That’s a conversation worth having rather than continuing to change the food.
And if you’ve been at this for more than six months without a clear answer, ask for referral to a dermatologist. It’s a specialism, one appointment often reframes the whole case, and it’s cheaper than another year of expensive food.