Is Grain-Free Dog Food Bad for Dogs? What the Evidence Actually Says

Updated September 11, 2026 · first published August 27, 2026Rewritten in full against the current regulations

There is a version of this question that gets asked in a panic, usually late at night, usually by someone who has just read a headline and looked over at a dog asleep on the floor.

So here is the honest answer before anything else, and then the evidence behind every part of it.

The problem was never the missing grain. Seven years of investigation has pointed steadily at something else in those foods — peas, lentils, chickpeas and beans, collectively called pulses. It is an association, not a proven cause. It affects a very small number of dogs. And in the dogs it does affect, it often improves when the diet changes, particularly if it is caught early.

If you are feeding a pulse-heavy food right now and your dog seems fine, you are not in an emergency. You are in a conversation you should have with your vet at the next visit.

This article is what is actually known, what is only suspected, who funded which study, and how a marketing category with no nutritional basis got built in the first place.

This is general information, not veterinary advice. Dilated cardiomyopathy is a serious heart condition. If your dog has any sign of heart disease — coughing, tiring easily, fainting, breathing hard at rest — that is a call to your veterinarian today, not an article.

The short answer

Grain-free is not automatically bad, and grain is not the issue. The evidence points at diets high in pulses, which most grain-free foods use to replace the grain. A food can be grain-free without being pulse-heavy, and a food with grain in it can still be pulse-heavy. The word on the front of the bag does not tell you which one you have. The ingredient list does.

Is grain-free dog food bad for dogs?

Not inherently, and not because of the missing grain. The concern is with diets high in pulses — peas, lentils, chickpeas and beans — which are what most grain-free foods use instead. Tufts veterinary cardiologists state the condition “does not appear to be caused by a lack of grains.” The link to pulses is a strong association across seven years of case reports, but causation has never been proven in a controlled trial. Most affected dogs improve when the diet is changed, especially when it is caught early.

What the FDA actually said, and what it did not

This is where most of the internet goes wrong, in both directions.

Between January 2014 and April 2019, the FDA received 515 reports of canine dilated cardiomyopathy, including 119 deaths. Ninety-three percent of the reported products contained peas and/or lentils.

That number is why the story exists. But the agency attached a caveat to it that almost nobody quotes:

“Adverse event numbers can be a helpful component… but by themselves, they do not supply sufficient data to establish a causal relationship with reported product(s).”

Reports are not a study. People who had read the news were far more likely to file a report, which inflates the count for named brands and deflates it for everything else. That is a known and unavoidable weakness of any reporting system.

The FDA has not issued a substantive public update since December 2022, stating it “does not intend to release further public updates until there is meaningful new scientific information.”

Both halves of that matter. The silence is not an exoneration, and it is not a cover-up. It is an agency saying it has nothing new to report.

Did the FDA ban grain-free dog food?

No. The FDA has never banned, recalled or restricted grain-free dog food over dilated cardiomyopathy. It published investigation updates between 2018 and 2022, named brands that appeared frequently in reports, and stated plainly that report counts alone cannot establish causation. There has been no substantive public update since December 2022.

The word that changed: pulses, not grains

The current scientific position comes from the cardiology group at Tufts University’s Cummings School of Veterinary Medicine, which has done more work on this than anyone. Their June 2026 update is the clearest statement available, and nothing newer exists:

“Diet-associated DCM appears to be associated with diets high in pulses (peas, lentils, chickpeas, beans).”

“This disease does not appear to be caused by a lack of grains or a deficiency of taurine or other nutrients.”

“Diet-associated DCM can slowly improve if diet is changed (especially if caught early).”

That second quote retires two things at once. The grain theory, and the taurine theory that dominated the early coverage. If you read about this in 2019 and have been supplementing taurine ever since, the science has moved.

And the third quote is the one to hold onto. This is not, for most affected dogs, a one-way door.

What are pulses in dog food?

Pulses are the edible seeds of legume plants — peas, lentils, chickpeas and beans. In dog food they appear as whole ingredients and as fractions such as pea protein, pea starch, pea fiber and lentil flour. They are used mainly to replace grain as a carbohydrate and protein source, which is why grain-free formulas tend to be pulse-heavy. They are the ingredient group associated with diet-associated dilated cardiomyopathy, not the absence of grain.

Why pulses, if not taurine? The honest mechanism

Nobody knows. But there is a candidate worth understanding, because it is specific and testable.

Pulses are high in protein, which is why they work so well as a grain replacement — they let a food hit an impressive protein number on the label. But that protein has a shape.

Reilly and colleagues, writing in the Journal of Animal Science in 2020, found that “methionine was the first-limiting amino acid for all pulses.” The measured methionine content of yellow peas is 0.23 percent of dry matter, green lentils 0.21 percent, navy beans 0.23 percent — against an NRC adult minimum of 0.26 percent.

Mansilla and colleagues put it plainly in 2019: the amino acid profile in these ingredients is “generally high in lysine and low in methionine.” And then, to their credit, the caveat: whether this has any link to dietary pulses and heart disease “has yet to be directly demonstrated.”

A separate line of work from the Tufts group has identified a urine biomarker suggesting a process called phospholipidosis may be involved (Freeman et al., American Journal of Veterinary Research, 2024).

Grade: plausible mechanism, not demonstrated. Both papers say so themselves.

The one place it was proven — and it reversed

There is a subset where the picture is much clearer, and it is worth knowing because it is the most hopeful thing in the whole subject.

Kaplan, Stern, Fascetti and colleagues published 24 Golden Retrievers with taurine-deficient dilated cardiomyopathy in PLOS ONE in 2018:

“Twenty-three of 24 dogs diagnosed with taurine deficiency and dilated cardiomyopathy were fed diets that were either grain-free, legume-rich, or a combination of these factors… Twenty-three of 24 dogs had significant improvement in their echocardiographic parameters and normalization of taurine concentrations following diet change and taurine supplementation.”

Twenty-three out of twenty-four improved. On echocardiogram — measured, not owner-reported.

Two honest qualifications. Golden Retrievers appear to have a breed-specific vulnerability to taurine deficiency that most breeds do not share, so this does not generalize to all dogs. And Tufts’ 2026 position is that taurine deficiency does not explain the broader diet-associated DCM picture. This was a real subset with a real, reversible problem — not the whole story.

Can dilated cardiomyopathy from diet be reversed?

In many cases it improves substantially, particularly when it is caught early. In a published series of 24 Golden Retrievers with taurine-deficient DCM, 23 showed significant improvement on echocardiogram after a diet change and taurine supplementation. Tufts cardiologists state that diet-associated DCM “can slowly improve if diet is changed.” This is a veterinary process involving cardiac imaging and monitoring, not something to attempt by switching foods on your own.

The other side of the argument, stated properly

An article on a subject this serious that only presents one side is not worth reading. Here is the strongest case against the whole thing.

McCauley, Clark, Quest and colleagues, in the Journal of Animal Science in 2020, argued that the FDA reports represent roughly 0.05 to 0.1 percent of the US dogs estimated to have DCM at all, that no controlled trial has ever demonstrated the effect, that breed genetics are a powerful confounder, and that publicity drove a reporting spike rather than a disease spike.

Disclosure, because it matters and because you should not have to go and find it: the authors of that paper were affiliated with BSM Partners, a pet food industry consultancy. That does not make the arguments wrong — the reporting-bias point is legitimate and the prevalence arithmetic is checkable. It does mean you should read it knowing who wrote it.

There is also a randomized trial. Morris and colleagues ran an 18-month prospective, double-blind study in 60 healthy adult dogs, published in 2025, and found both grain-free and grain-inclusive diets supported normal cardiac function.

Disclosure again: that trial was conducted at Hill’s Pet Nutrition, a company that sells grain-inclusive food. It is a genuine randomized controlled trial, which is more than the other side has. It also studied 60 healthy dogs for 18 months, which is a small sample and a short window for a disease that develops slowly.

And an independent voice, which is the rarest thing here. Mornard, Massara Brasileiro and Marcondes-Santos published a review in Veterinary Sciences in 2025 that states “this research received no external funding.” Their finding: “a strong link between diet and DCM, particularly regarding legumes,” while conceding that further study is required.

Grading all of it honestly

The honest bottom line: strong association, unproven causation, small absolute numbers, and a reversible course in many affected dogs. Anyone who tells you it is settled — in either direction — is telling you something the evidence does not support.

So where did grain-free come from?

Not from veterinary medicine. This is the part that explains everything else.

Food allergy affects roughly one to two percent of all dogs (Olivry & Mueller, 2017). Not one to two percent of itchy dogs. One to two percent of dogs.

And when researchers looked at what actually causes those reactions — 297 dogs with confirmed food reactions, reviewed by Mueller, Olivry and Prélaud — the list does not say what the marketing says:

Corn is four percent of cases, in a condition affecting one to two percent of dogs. That works out to roughly one dog in a thousand.

Meanwhile the top three — beef, dairy, chicken — are animal proteins, and grain-free foods generally contain more animal protein, not less.

Is grain a common cause of allergies in dogs?

No. Food allergy affects only about 1–2% of dogs overall, and in a review of 297 confirmed cases the leading culprits were beef (34%), dairy (17%) and chicken (15%). Wheat accounted for 13% and corn for just 4%. The proteins most often implicated are animal proteins, which grain-free foods typically contain more of, not less.

And why “lamb and rice” stopped meaning anything

Lamb was the novel protein of the 1980s — novel precisely because almost nothing used it. Once half the aisle used it, it stopped being novel. The industry moved the target and kept selling the old idea.

Why elimination diets so often fail

If you have ever run a strict elimination trial on a retail limited-ingredient food and got an inconclusive result, there is a published reason that may not be your dog’s fault or your vet’s.

Olivry and Mueller reviewed 18 studies in BMC Veterinary Research in 2018. Their conclusion:

“The mislabeling of pet foods appears rather common, even in those with ‘novel’ or ‘limited ingredients’ proposed for elimination diets. Unexpected added ingredients are more frequently detected than those missing from the label.”

Undeclared ingredients were found in 0 to 83 percent of diets tested, with a median of 45 percent. In novel and limited-ingredient diets specifically — the exact products sold for elimination trials — the range was 33 to 83 percent.

A 2020 Canadian study found undeclared species DNA in nine out of nine raw canine diets tested.

Read that again in practical terms: an over-the-counter “no chicken” diet can contain chicken. Every inconclusive elimination trial run on a retail limited-ingredient food may have been the label failing rather than the dog being a mystery.

This is exactly why veterinary dermatologists insist on a therapeutic diet — hydrolyzed or genuinely novel protein, made on controlled production lines — for a real elimination trial. Not because the retail versions are worse food, but because they cannot be trusted to contain only what they say.

If you have a dog with suspected food allergy, this is a conversation for your veterinarian. It is the single most useful thing in this article and it is not something to act on alone.

The part that is about you, not your dog

This is where it gets uncomfortable, and it is the best-evidenced section in the article.

Banton and colleagues at the University of Guelph surveyed 3,298 dog owners across Canada, the USA, Germany, France and the UK, published open-access in PLOS ONE in 2021. They asked what actually predicts buying grain-free.

The single strongest predictor is a belief about a condition that affects one to two percent of dogs.

And look at what the rest of that list has in common. “No fillers.” “No by-products.” Your own dietary rules. Four of the five strongest predictors are not about your dog at all. The researchers’ own phrase for it: consumers use “ingredients as a proxy for nutrition.”

Where people get the information

• Veterinarian: 57.6% · Online: 40.9% · Pet store staff: 29.3% · Friends and family: 27.9%
• Using online resources made someone 1.6× more likely to buy grain-free (P<0.0001)
• Advice from pet store staff: 1.3× more likely (P=0.006)

The veterinarian is the most-consulted source, and consulting one did not push people toward grain-free. The two sources that did are the two with something to sell.

And yes, there is a gender finding

A reader asked whether the marketing is aimed at women. Here is what can actually be proven.

The Guelph study found that male respondents were significantly less likely to select “no grain” when choosing a pet food. It held across three separate statistical models: the allergy model (P = 0.015), the diet model (P = 0.046), and the purchasing habits model (P = 0.002). The authors connect it to the established finding that women purchase organic food more frequently than men.

Women are measurably more likely to buy grain-free. That is proven.

What is not proven is that packaging is deliberately designed to target them. Those are two different claims and only the first has evidence behind it. We are not going to assert the second just because it feels true.

What can be said is this. The industry’s own trade press ran a piece in February 2025 in which an industry figure described the sector’s own framing of its customers:

“In the industry, we call these people blanket ‘pet parents,’ which is the lamest, laziest piece of marketing.”

The same piece reported that 79 percent of consumers feel pet product advertising is misleading and 82 percent believe stricter label regulations are needed, with only 25 to 49 percent understanding common marketing terms.

Disclosure: that consumer data comes from BSM Aperture, the research arm of BSM Partners — the same consultancy behind the industry critique quoted earlier. Which makes the number stronger rather than weaker: this is the pet food industry’s own consultancy reporting that four in five of its customers think its advertising is misleading.

The rise, and the fall

Grain-free was a market position, and it moved like one.

Grain-free dry dog food grew 9.6 percent year on year through September 2018. The FDA alerts landed in July 2018, February 2019 and June 2019. By the end of 2019, grain-free in pet specialty was down 5.9 percent, grain-free dry dog food specifically had fallen 11 percent, and the sixteen brands named in the FDA’s June 2019 update were down 13.3 percent for the year.

A category does not grow at ten percent a year on the strength of canine biology and then fall eleven percent on the strength of a press release. It moved as a marketing axis, because that is what it was.

Who owns the boutique brands

The other thing worth knowing is that “boutique” mostly is not.

This is not an accusation. A large parent company often means better quality control, more testing and more nutritionists on staff, not less. It is simply worth knowing that the small-batch feel on the front of the bag is frequently a brand asset owned by one of the largest food companies on earth.

What to actually do

1. Read the ingredient list, not the front of the bag. The claim you care about is not “grain-free.” It is how far up the list peas, lentils, chickpeas, beans, pea protein and pea starch appear, and how many of them there are. Reading a dog food label covers the ingredient-splitting trick that makes this harder than it should be.

2. Do not panic-switch. An abrupt diet change causes real, immediate gastrointestinal problems. The risk being discussed here is slow and uncommon. Talk to your vet, then transition gradually.

3. Ask your vet at the next routine visit. Not an emergency appointment. Bring the bag or a photo of the ingredient panel.

4. If your dog is a breed predisposed to DCM — Doberman Pinschers, Great Danes, Boxers, Irish Wolfhounds, Cocker Spaniels, and Golden Retrievers for taurine specifically — that conversation matters more, and it should happen sooner.

5. Know the signs. Coughing, tiring easily on walks the dog used to manage, faster or harder breathing at rest, weakness, fainting, or a swollen abdomen. Any of those is a veterinary appointment, regardless of what is in the bowl.

6. Do not supplement taurine on your own. Tufts is explicit that taurine deficiency does not explain the general picture, and supplementing can mask a problem you would rather find.

Which dogs are most at risk from diet-associated DCM?

Breeds already predisposed to dilated cardiomyopathy carry the most concern: Doberman Pinschers, Great Danes, Boxers, Irish Wolfhounds and Cocker Spaniels. Golden Retrievers appear separately vulnerable to taurine-deficient DCM specifically. Large and giant breeds are over-represented generally. That said, cases have been reported in breeds with no known predisposition, and the absolute number of affected dogs remains small relative to how many eat these diets.

Should I switch my dog off grain-free food?

That is a decision for you and your veterinarian, based on your dog’s breed, age and health rather than on a headline. What the evidence supports is checking how prominent pulses are in the ingredient list, raising it at your next routine appointment, and never switching abruptly — an abrupt change causes immediate digestive problems, while the risk under discussion is slow and uncommon. If your dog has any sign of heart disease, that is a same-day veterinary call.

The reusable test

The thing worth taking from this beyond the specifics:

Ask what the claim on the front of the bag is protecting your dog from, and how common that thing actually is.

“Grain-free” protects against a grain allergy. Grain allergy is a fraction of a condition affecting one to two percent of dogs. The claim was solving a problem that barely existed, at a higher price, using an ingredient we now spend our time worrying about instead.

That is not a conspiracy. It is a rare problem, a frightened owner, an unregulated adjective, and a margin.

—

Related reading: What working and service dogs actually need to eat is the full picture this article sits inside. How to read a dog food label is how to check your own bag. How much to feed covers the one dietary factor ever proven to extend a dog’s life. And how I review things explains why no product is named in this article.