Grain-Free Dog Food and Heart Disease: How the Narrative Was Built
THE SHORT VERSION:
“Grain-free food causes heart disease in dogs” became a clinical talking
point. It was never what the data showed. Seven years later, the FDA closed its
investigation without establishing causality, a $2.6 billion lawsuit has raised serious questions about how that investigation was triggered, and the peer-reviewed literature remains genuinely contested. This article covers how the narrative was built: the investigation, the data quality problems, and the legal aftermath. Part 2 covers what the peer-reviewed science actually shows and what it means for your dog.
In June 2018, before the FDA’s first public alert, Dr. Lisa Freeman published a Petfoodology blog post titled “A broken heart: Risk of heart disease in boutique or grain-free diets and exotic ingredients.” This post warned about a possible connection between grain-free diets and dilated cardiomyopathy (DCM) (Freeman 2018) . In the prior year, a small group of academic veterinarians had begun reporting a trend they felt warranted further investigation. Cardiologist Joshua Stern at UC Davis was seeing an apparent cluster of Golden Retrievers with DCM, low taurine, and similar grain-free diets, while Lisa Freeman and colleagues at Tufts were observing DCM in dogs eating BEG diets (Wood 2018).
Between January and July 2018, prior to formal announcement of an investigation, the FDA received 25 case reports of DCM that mentioned a possible diet link. Of those 25 reports, seven came from Freeman’s group at Tufts. Internally, Tufts had circulated guidance instructing its clinicians to report DCM cases to the FDA only if the dog was eating a BEG or non-“reputable” diet. This means that grain-inclusive diets from major brands were not reported by design, even though cases existed.
In July 2018, the FDA announced it was investigating a potential link between certain dog food diets and DCM (Adminstration 2018) . Within months, the story had simplified dramatically: veterinarians warned clients off grain-free diets, brands lost market share by the hundreds of millions of dollars, and a generation of dog owners began associating peas and lentils with cardiac death.
That narrative was reinforced on December 1, 2018, when a commentary article by Freeman and colleagues titled “Diet-associated dilated cardiomyopathy in dogs: what do we know?” was published in the Journal of the American Veterinary Medical Association (JAVMA) (Freeman, Stern et al. 2018) . This was published as a viewpoint piece, not a peer-reviewed original study, but was frequently cited and shared as though it were primary evidence.
In a November 2018 Petfoodology update titled “It’s Not Just Grain‑Free: An Update on Diet Associated Dilated Cardiomyopathy,” Freeman re‑framed the concern around boutique, exotic, grain‑free (BEG) diets, argued that most diet‑associated DCM cases in her hospital were eating such diets (Freeman 2018) . She recommended that owners switch to diets from large, “well‑established” manufacturers using common ingredients including grains, despite acknowledging that the underlying cause of diet‑associated DCM was still unknown.
Combined with the Petfoodology blogs and the FDA alert, the JAVMA and Petfoodology blogs effectively functioned as a de facto position statement for the veterinary profession. This narrative ballooned and solidified quickly, spreading far beyond what the underlying evidence could support. This article walks through what the data and investigation record actually show. This included the parts obscured by both the alarm camp and the grain-free advocacy camp, neither of which has given pet owners the complete picture.
Then in June 2019, the FDA took the unusual step of publishing the names of 16 pet food brands that appeared most frequently in its adverse-event reports, while still stating that it had not established causation (Administration 2022) . Naming brands at this early signal-detection stage implied a level of certainty the data did not support, and media coverage amplified the brand list while largely stripping away the caveats.
What Is DCM?
DCM is a disease in which the heart muscle weakens and the chambers enlarge, reducing the heart’s ability to pump blood effectively (Tidholm and Jönsson 2005) . Over time, untreated DCM can progress to congestive heart failure, arrhythmia, or sudden death. It is the second most common form of heart disease in dogs (McCauley, Clark et al. 2020) . DCM has two clinically distinct categories. Primary or hereditary DCM is genetically driven and disproportionately affects certain large and giant breeds (e.g. Doberman Pinschers, Great Danes, Boxers, Irish Wolfhounds and Golden Retrievers). Athough it’s important to know that DCM has not been widely studied across most breeds (O’Grady and O’Sullivan 2004) . In fact, the natural history of the disease has been described in greatest detail in Doberman Pinschers, where longitudinal and genetic studies show a very high lifetime prevalence and a well-defined progression from occult disease to heart failure. For most other affected breeds, available data on true incidence, age of onset, and typical progression are much more limited and thus extrapolated at least in part from Doberman Pincher data.
Secondary DCM, the kind at the center of this debate, develops because of something else. This could be an infection, a medication, a toxin, a nutritional problem, a combination of these, or even another factor we are not aware of. What separates secondary DCM from primary DCM is that can improve or resolve when the underlying cause is corrected; hereditary DCM does not. This is clinically important and tends to get lost in public discussions. When someone says “we
see lots of heart disease in dogs on grain-free diets,” that could reflect primary hereditary DCM that would have occurred regardless of diet, secondary DCM, or something else entirely. Diet type alone is not a diagnosis.
Clinically, DCM describes how the heart looks and behaves. It does not describe why it ended up that way (O’Grady and O’Sullivan 2004) . Multiple different processes can produce the same dilated, poorly contracting heart: inherited myocardial disease, chronic tachycardia, myocarditis, endocrine disease such as hypothyroidism, certain toxins, and in some cases specific nutrient deficiencies like taurine. In many dogs, especially in general practice, the underlying cause is never definitively identified, because the tests required (e.g. myocardial biopsy, extensive infectious and metabolic workups) are rarely performed. That uncertainty is part of why it is so easy to over-attribute DCM to whatever risk factor is most visible at the time.
How the FDA Investigation Started — and Their Data Quality
The FDA’s investigation formally launched in July 2018 after an unusual cluster of DCM cases was observed in dog breeds not typically prone to the disease. The common thread reported by the investigating veterinarians: the dogs were eating diets labeled “grain-free,” many of which were high in peas, lentils, or other pulses. The FDA began soliciting reports from veterinary professionals and, by December 2022, had collected 1,382 canine DCM case reports. At first
pass, more than 90% from dogs eating grain-free diets, 93% from dogs eating diets containing peas and/or lentils.
That sounds damning. But the data collection methodology has since come under serious scrutiny. Let’s look at why.
The Tufts Reporting Protocol
Among the earliest and most prominent contributors to the FDA’s diet-associated DCM case pool was Dr. Lisa Freeman of Tufts University’s Cummings School of Veterinary Medicine. FDA adverse-event records obtained under the Freedom of Information Act show that Tufts clinicians submitted 7 of the first 25 diet-linked DCM cases that informed the agency’s initial July 2018 communication (Santoro 2022) . In the same FOIA production, a June 2018 internal protocol from Tufts instructed veterinarians to report DCM cases to the FDA “if patient is eating any diet besides those made by well-known, reputable companies or if eating a BEG diet”. This means that dogs eating grain-inclusive diets from major manufacturers were, by design, not reported from this institution.
What we can say with confidence is that the early data submitted to the FDA were not a representative sample of all dogs with DCM. They were a selected subset, shaped by an internal reporting rule that favored one kind of answer over others.
The Surge After Public Announcement
The FDA made its investigation public in July 2018, and case reports surged immediately. Between 2014 and mid-2018, the FDA received only a handful of DCM reports per year (Administration 2022) . However, after the announcement, that number jumped into the hundreds annually, before tapering off. This is a well-described phenomenon in pharmacovigilance: publicizing a suspected link dramatically increases reports of that link, because veterinarians who previously would not have associated diet with DCM began actively looking for and reporting cases that fit the new narrative.
For context, there are an estimated 87 million dogs in the U.S. as of 2025 (Association 2025). Scientific literature sugges that somewhere between 0.05–1.1% of the canine population is living with DCM at any given time (Fioretti and Carri 1988, Sisson, Thomas et al. 1995) .Based on this, simple math estimates approximately 435,000 to 957,000 dogs living with DCM in the U.S. at any given time. And these numbers are before diet enters the picture. Reporting bias does
not mean those dogs were not sick, rather it means the data cannot tell us whether they were sick because of what they ate.
This is why “we see lots of heart disease in dogs on grain-free diets” is not, by itself, evidence that grain-free diets cause that heart disease. Without standardized diagnostics, case definitions, and a proper denominator, it is impossible to separate background hereditary DCM from any true dietary signal.
The FDA Reports Contextualized
It’s important to keep the scale of the problem in mind. Using conservative estimates, somewhere between 435,000 to 957,000 dogs with DCM in the U.S. at a given time. Considering this the FDA’s reported case pool of 1,382 canine DCM reports since 2018 represents a fraction of a percent of dogs estimated to have DCM, regardless of diet. A voluntary adverse-event reporting system that captures only a tiny fraction of background disease, and that was further filtered at some institutions to exclude grain-inclusive diets from the outset, does not constitute surveillance data. It constitutes a selected case series and those selected case series cannot establish prevalence, incidence, or causation.
Unlike human medicine, which benefits from centralized public-health surveillance systems such as the CDC that track disease trends across large populations, veterinary medicine has no comparable, comprehensive database for canine heart disease. This absence of systematic surveillance makes it much harder to accurately quantify risk, identify true patterns, or distinguish signal from noise when evaluating potential dietary and disease relationships in dogs.
More importantly, something like this is detected in a small, highly selected subset of cases, it is dangerous to extrapolate that signal to the entire dog population. Overgeneralizing from such data can inflate perceived risk, stigmatize entire categories of diets or ingredients, and misdirect research and clinical resources away from the much larger burden of background DCM and other cardiac disease.
This Is What Voluntary Reporting Systems Do
The pattern of the FDA’s DCM reports over time: a low baseline, a sharp surge following the July 2018 announcement, and a subsequent plateau, is not unique to this situation. It closely resembles a well-described artifact of voluntary pharmacovigilance systems known as the Weber effect. This is the tendency for voluntary adverse-event reports to spike after media coverage or regulatory announcements and then taper, regardless of whether the underlying event rate has truly changed. Combined with notoriety bias (the tendency for clinicians and owners to report cases that fit a publicized pattern while overlooking those that do not) and the denominator blindness inherent in a system with no defined reporting population, the shape of the FDA’s DCM data is entirely consistent with a reporting artifact layered on top of a large background burden of disease.
The Reversibility Argument Does Not Mean What It Is Claimed to Mean
One of the most frequently cited pieces of evidence for diet causation is that some dogs improved after switching away from grain-free diets. In most reported cases where clinical improvement was documented, dogs were not simply switched to a different diet and observed. They were simultaneously started on standard cardiac medications which are typically pimobendan, ACE inhibitors, or both. These patients also often received taurine and L-carnitine supplementation. In other words, multiple interventions were introduced at once. When a dog improves under those conditions, attributing that improvement specifically to diet change is not scientifically defensible. The reversibility argument, used throughout the diet- association literature as implicit proof of causation, is hypothesis-generating at best, and even that is a reach. Without controlled data that separates these variables, it cannot carry the evidentiary weight placed on it.
The Diagnosis Problem
Compounding the data collection bias is a data quality problem that received far less attention: the cases themselves were not uniformly diagnosed to the same standard. Dukes-McEwan and colleagues specifically recommended that a DCM diagnosis require echocardiographic evaluation by a veterinary cardiologist to be considered reliable (Dukes-McEwan, Borgarelli et al. 2003) . What’s problematic in this instance is that a meaningful portion of the FDA’s reported cases were submitted by general practitioners. When echocardiographic data were present (most cases did not have this), they were not standardized across submissions.
In some instances, dogs with soft murmurs, dogs with radiographic cardiomegaly but no documented systolic dysfunction, or even dogs eating grain-free diets without any confirmed structural or functional heart disease were all reported under the umbrella of DCM. As a result, findings that would not meet formal diagnostic criteria in a cardiology referral setting were reported as DCM, and any analysis built on the assumption that all reports represent equivalent DCM diagnoses will overstate the problem, especially when this underlying data is unverified or incomplete.
The FDA’s Own Conclusion
On December 23, 2022 which happened to be the Friday before Christmas, the FDA announced that it was ending routine public updates on the investigation (Administration 2022) . The agency’s stated reason: adverse event reports alone do not supply sufficient data to establish a causal relationship. After more than four years and 1,382 case reports, the FDA had not identified a mechanism, had not established causality, had not issued any recalls or dietary prohibitions, and concluded it lacked sufficient scientific basis to take further regulatory action.
Industry Guidance Was Never Retracted
In the period following the FDA’s initial July 2018 announcement, major veterinary
organizations and key opinion leaders which included the World Small Animal Veterinary Veterinary Association (WSAVA) Global Nutrition Committee and American Animal Hospital Association (AAHA)-linked nutrition resources began advising clinicians to avoid grain-free/BEG diets in dogs at risk for, or diagnosed with, DCM. However, these were largely applied to all dogs, and even cats and not ‘at risk’ cases. In fact, these recommendations still appear in many current veterinary and client-facing resources, often presented as practical, evidence-based guidance to clients in the clinic rather than clearly labeled as provisional interpretations of an evolving dataset.
These recommendations have not been systematically revisited considering the FDA’s 2022 decision to stop public updates for lack of causal evidence. They continue to shape practitioner conversations and consumer behavior in ways that no longer fully reflect the state of the evidence. This is how preliminary signals become clinical dogma: the announcement travels fast and wide; the quiet closure does not. The asymmetry between how this investigation opened and how it ended is itself part of the story.
The Hill’s Lawsuit: Allegations and Unanswered Questions
In February 2024, KetoNatural Pet Foods filed a $2.6 billion Lanham Act lawsuit against Hill’s Pet Nutrition (a Colgate-Palmolive subsidiary), the Morris Animal Foundation, the Mark Morris Institute, and veterinary researchers Drs. Lisa Freeman, Joshua Stern, Darcy Adin, Ryan Fries, and John Rush (2024) . The complaint alleges that Hill’s and several associated researchers made false and misleading marketing claims about grain-free diets and DCM to the economic
detriment of competing pet food brands.
The suit claims that Hill’s, facing substantial market share losses as grain-free competitors grew, orchestrated a coordinated campaign to weaponize the FDA’s reporting system. Specific allegations include:
- that Hill’s-affiliated veterinarians selectively submitted grain-free-associated DCM cases to the FDA while withholding cases in dogs eating Hill’s-brand and similar grain-inclusive diets;
- that the resulting data created a distorted picture of the association;
- that Hill’s and its associates then amplified the resulting panic through academic publications, veterinary channel marketing, and social media, allegedly including a private Facebook group used to suppress dissenting perspectives;
- and that the financial result was a near-total reversal of Hill’s market decline, with the company becoming arguably the fastest-growing pet food brand in the country in the years following the investigation.
The suit is built around the Lanham Act, a U.S. statute governing false advertising and unfair competition. The legal theory is that Hill’s and its associates made false and misleading statements about the scientific evidence surrounding grain-free diets, disseminated through marketing materials and academic papers, thereby harming competitors economically.
As of the date of this article, the case is on appeal. In November 2024, the district court granted Hill’s motion to dismiss on legal grounds, holding that many of the challenged statements were not actionable “commercial speech” under the Lanham Act and that the complaint had not adequately pled its claims. KetoNatural has appealed that decision to the Tenth Circuit.
In plain language, the court’s ruling addressed whether the alleged statements could legally be treated as advertising claims. It did not rule on whether the underlying scientific narrative around grain-free diets and DCM was accurate. These are two different things. A controversy that began as a scientific question about diet and heart disease has, at least so far in the courtroom, been decided on the technicalities of marketing law rather than on a full evaluation of the evidence.
What the lawsuit means for your interpretation of the science:
Serious. The FOIA documents describing the Tufts reporting protocol are real. The financial conflicts of interest involving the named researchers and the major grain-inclusive pet food companies (Hill’s, Nestlé Purina, Mars) are documented. The market share data is real.
What it doesn’t mean: The lawsuit does not establish that no dogs have developed DCM on grain-free diets. Clinical cases of diet-associated DCM are real. Dogs got sick, some improved on diet change, and that signal is meaningful regardless of whether Hill’s manipulated the reporting pool. The lawsuit, if it proceeds, will adjudicate commercial law claims — not definitively resolve whether legume-heavy diets contribute to DCM in some dogs.
The honest answer is: both things can be true. The investigation may have been shaped by financial conflicts of interest and biased data collection, AND a real but narrower signal may exist in a subset of susceptible dogs on specific high-legume formulations. These are not mutually exclusive.
The information in this article is for educational and informational purposes only and is not a substitute for individualized veterinary medical advice, diagnosis, or treatment. Always consult your veterinarian or a board- certified specialist regarding your pet’s specific health needs, diagnostic testing, and dietary recommendations. Never disregard professional veterinary advice or delay seeking it because of something you have read here.
