Dilated Cardiomyopathy and Grain-Free Diets: Thoughts on the FDA Update
July 3, 2019 · Dr. Justin Shmalberg
Fatal heart disease from pet foods? This is the alarming suggestion circling the internet again following the release of more information from the FDA. It’s important to view such information in a rational context and to use the opportunity to think about your pet’s nutrition broadly.
What do we know?
Veterinary cardiologists suggested that a heart condition known as Dilated Cardiomyopathy (DCM) was occurring more commonly in Golden Retrievers and other breeds not commonly thought to be genetically predisposed (breeds in that latter category include Dobermans, Boxers, Great Danes). DCM has of course been reported in other breeds in smaller numbers, generally other medium- to large-breeds, but there was a concern that the overall number of atypical cases was higher than previously recognized. Some cardiologists suggested that many atypical dogs were fed grain-free diets. This led to interest from veterinarians, breed groups (specifically Golden Retriever owners), the public, and the FDA. The FDA has been accumulating data for months at the present time, and in the last week provided some more details than previously available.
What is DCM?
DCM is broadly a condition characterized by poor contraction of the heart muscle, which can lead to heart failure, a potentially fatal condition. Dogs may have no signs at all, nonspecific signs (poor appetite or energy level), or in advanced stages coughing due to fluid in the lungs or enlargement of the heart, fluid in the belly, severe weakness, collapse, or even death.
Is DCM always a nutritional problem?
Not usually, as the genetic predisposition is thought to be the cause in most cases. However, deficiency of taurine is known to cause a potentially reversible form of DCM (non-nutritional DCM is managed but not reversed). So what is taurine? Taurine is an amino acid made by dogs (but not cats) from other precursor amino acids, specifically methionine (an essential amino acid) which is then converted to another amino acid, cysteine. We can group these together and call them sulfur-containing amino acids (SCAA) since they, you guessed it, contain sulfur. If you give a cat a diet with little or no taurine, DCM is a real possibility. If you give a dog a diet without taurine that is low in SCAA or has SCAA that are not available (not well digested), DCM is also possible. This is what happened nearly 2 decades ago when low-protein diets without adequate or digestible SCAAs (lamb is naturally low) were fed to dogs. It occurred in cats fed diets low in natural taurine without supplementation. Now, most pet foods contain supplemental taurine for cats, and either supplemental taurine or SCAA for dogs (unless they are very high protein diets with naturally occurring amounts). It’s important to note that measuring amino acids or nutrients in general isn’t the whole story because it only tells you what theoretically can be digested, not what is actually unlocked from the food and absorbed. Basically, a number can look good but not produce real-world effects.
What is the scope of atypical DCM?
Unfortunately, this is not known. The FDA has reviewed information from 560 dogs and 14 cats between January 1, 2014 and April 30, 2019, with 119 and 5 deaths in those pets, respectively. My cardiologist colleagues tell me that some owners are not reporting their animals, especially now since the problem is more widely publicized - many probably feel it’s not worth the effort or won’t help to unlock what is happening. There is no streamlined or central reporting system for veterinary cardiologists to input their suspected cases and so the FDA is likely a lower estimate. Some of the dogs reported were in breeds with predispositions (25 Great Danes, 15 Dobermans, 11 Boxers). Unfortunately detailed information is not always available for each dog, and so if a dog was of a predisposed breed and showed signs of reversal, we would think it more likely to not be genetic DCM which isn’t known to dramatically improve.
Not all of the cases above have been confirmed to have DCM - the FDA reported that 202 dogs (35% of total reports) and 6 cats (42.8%) had both medical records reviewed and met the DCM definition. Of those reviewed, 59.4% had findings consistent with DCM while others were classified as non-DCM (which included those with decreased left ventricular function (which might be pre-DCM) or other heart changes which may not be related to DCM). Based on a comment below, this paragraph was edited, and I also would emphasize here as I did elsewhere that there certainly are some number of pet parents who did not report cases similar to what the FDA has collected.
It’s important to remember that the total number of atypical DCM cases in previous years also isn’t known so we have no baseline to which we can compare the current numbers. It’s still safe to say that DCM is very uncommon in dogs fed any specific pet food.
What about breed?
Golden retrievers were the most common breed identified (95 cases). This could be because they are truly over-represented or just because their owners were very active in mobilizing as a group and reporting cases. If Goldens are truly more frequently affected, then it’s possible that they have a unique genetic predisposition in some families - previous reports have found this for taurine deficiency in Goldens. There could be previously identified genetic predispositions in the other breeds on the list (like Labradors and other retrievers). Or perhaps, it’s a combination between genes, diet, the microbiome and a host of unknown factors, especially in those Goldens with normal taurine levels.
It’s important to note that almost all of the breeds reported are medium to large (as has been seen in atypical cases before). The only small breeds were Shih Tzus (5 cases) and potentially ‘unknown’ (13 cases) or maybe some mixed breed dogs (62 cases). We have no frame of reference to compare to previous years in atypical DCM breed distribution. This would only be identified through a consortium of cardiologists working to identify all previous cases - something that is very challenging to do.
Is taurine to blame for the current cases?
Taurine levels were only available for 64% of dogs with confirmed DCM, and less than half of those had low taurine (42%). Generally speaking, it would be safe to say that low taurine was less common than a normal or high taurine across the dogs reported.
For Golden Retrievers, it’s a bit different. The FDA lists taurine results for confirmed Golden DCM cases (24 total). 79.2% had a low taurine, 16.7% normal taurine, and 4.2% high levels. In normal Golden Retrievers, 7 dogs had low taurine and 4 normal. In Goldens with other heart changes (non DCM), 13/17 had low values and 4/17 had normal values. Across the board, with or without DCM, the reported Goldens seem more likely to have low taurine. It is previously known that Goldens can have issues with taurine causing DCM on a familial basis.
The type of sample submitted can influence results. It is known that taurine will yield potentially different results if run on whole blood (the blood that comes straight from the vein) or on the fluid portion of the blood (plasma) and there were differences in what type of samples submitted that could be significant. Some cardiologists are now also suggesting a new taurine reference interval just for Golden retrievers.
In cats, there’s really not enough information to make any conclusions.
Taurine was a part of treatment in many dogs that improved (in addition to diet change, medications, and other supplements like fish oil). Taurine is known to help encourage normal contraction of the heart (which is why DCM is caused with true taurine deficiency - remember that DCM is characterized by the heart not contracting well). Taurine may help even genetic DCM to some degree - this has never been really studied to my knowledge. As so many treatments were initiated at the same time, it’s nearly impossible to tell if taurine alone was helpful. Certainly taurine is not harmful but the evidence is presently inconclusive that taurine is to blame in many cases. Additional work is needed in Golden retrievers to determine relationships between DCM and taurine.
Is diet to blame at all?
There is no proven association between any dietary strategy or specific diet and DCM. The FDA is testing reported foods but has released limited information. They have stated that the macronutrients, SCAAs, and taurine are similar between tested grain-free and grain-containing foods.
There are a number of guesses, which are just that - guesses, about what could be going on IF a dietary link to these cases were proven. These include:
Poor availability of SCAAs and taurine in the diet, meaning they are there but not well absorbed
Legumes or legume proteins at high amounts providing less digestible amino acids, antinutritional factors, and soluble fiber that interacts with digestion and the bacterial flora in the gut (microbiome)
Poorly digestible ingredients especially those providing protein (meat meals with high bone content (ash), vegetable sources of protein). Remember that poor quality, high ash meals are less expensive than others and that for some meats that may be all that’s commonly available.
Processing changes to SCAAs, especially during extrusion (kibble formation)
Lack of testing for taurine and SCAAs (although the FDA is saying levels are similar in their testing)
High fiber amounts that impact the recycling of taurine in the gut (taurine is a part of bile salts found in the bile that get resorbed at different levels depending on the diet)
Genetic predispositions to differences in handling nutrients or in production of taurine from SCAA, such as previously identified in Golden retrievers
Unfortunately, the information about brand does nothing for nutritionists to be able to look into these potential factors. A detailed list of complaints, some of which contain specific products within brands, is found here. It will take some time to aggregate the information by specific food, and it’s not clear that will be done by the FDA.
Is there a type of food more commonly reported?
Kibble is the most common food type reported (452 reports) as compared to those feeding multiple types (24), unknown (26), and raw (9). One instance each of home-cooked, refrigerated, semi-moist, and tubbed are reported. We do know that kibble is by far the most commonly fed type of pet diet, so the data may just reflect that. However, it’s interesting that canned and fresh diets are not generally reported here (especially canned which would be second most commonly fed). This may be because there is no true association between diet or if there is, that it is not as likely to cause issues because of different ingredients or because of different cooking/processing.
The brand list - should they always be avoided?
Every brand produces a number of foods and this is where the incidence of specific diets would be helpful. In looking through the reports of DCM, many lack complete information about the specific product being fed. Many owners were feeding multiple diets together or rotating (even between brands) which complicates the picture. Examining those brands with 50 or more reports, which are Acana, Zignature, and Taste of the Wild, I searched the reports for specific products. Remember that the data is not easy to search or manipulate in the form provided by the FDA so please interpret these findings with caution and apologies in advance for inaccuracies identified. This said, for Zignature, the kangaroo product appeared 43 times (when reported as the only diet fed), trout 6 times, and turkey 4. Other products were below that. For Acana, lamb appeared 19 times, followed by pork 5, duck 5, and a combination of only duck and pork flavors 5 times. For Taste of the Wild, Pacific Stream appeared by itself 8 times (fish), High Prairie (poultry) 8 times, and Pine Forest (venison) 6 times, followed by other products. This was a very cursory glance, but you can appreciate the diversity of products and the distribution. This data also would be best compared to the sales of the particular brands and individual diets. If a company is more popular, it might appear more frequently in any list and the same would be true for a particular food.
The absence of generally less expensive kibble diets (by this I mean the cheapest diets you can find on grocery store shelves) may be a bias because of the fact that some pet parents feeding those diets may be less likely to see a veterinary cardiologist for screening of heart disease. (Note (added after publication): Please see my commentary below as this was not intended to suggest that people who feed diet X don’t care about their pets or won’t see a cardiologist but rather that I do not know what the background diets are that are fed to pets that do see a cardiologist - they could be different than the general population. This would be important to know in comparing the relative risk for any diet contributing to DCM. The diets fed in my nutrition practice are very different than those fed in the general population, as an example).
So should you avoid particular brands? Likely not, based on the information so far. If there is an association, it will likely be related to food type (processing) and ingredients more than a particular brand. Also remember the vast majority of dogs on all these foods don’t have any issue. Most companies are working with nutritionists to continuously evaluate their products especially in light of the present concerns. As nutritionists, we have very limited information to tell them except for those related to ingredients as described below.
Are foods which have been evaluated through feeding trials the answer or foods meet meeting ‘WSAVA standards’?
This was added based on a comment you can read below (as well as my reply which contains more detailed information), but I do not believe there is evidence that feeding trials would be preventative for designing a food which avoids this issue if and when causation is determined. Feeding trials conforming to only AAFCO standards are a very low bar to clear, and foods that meet feeding trials can actually have nutrient values outside AAFCO standards. You can read more in the comments below about my take on feeding trials at this stage. WSAVA guidelines for selecting a pet food are helpful, and the majority of listed foods likely meet WSAVA recommendations. Remember the WSAVA provides very few actual recommendations except for ensuring that a food is formulated to meet AAFCO standards OR has been tested with feeding trials for the appropriate life stage of your pet. There is a list of questions to ask, without any opinion given by WSAVA as to what the answer should be. It is my opinion as a nutritionist that foods should, feeding trial or not, have been testing for a nutrient profile after they are formulated and then produced. Remember that there is no evidence that listed foods have a nutrient deficiency, which is consistent with FDA’s reported testing.
What about ingredients - any learnings from the FDA data?
The FDA approach to ingredients has not done any favors for the analysis. They include main ingredients as those before vitamins and minerals - as a result, these ingredients could be present in 1% to say 80% of the diet. A big difference nutritionally. It of course is very difficult to determine the concentration of any ingredient based on the ingredient list. They are listed in descending order as-fed (which means with water). A fresh meat will contribute fewer calories than a meal when first on an ingredient list - this is an example of how complicated this becomes. The FDA does say 90 percent were labeled grain-free and 93 percent had legumes (peas, lentils, chickpeas). Potatoes and sweet potatoes in a lesser number - 42%. But unfortunately we don’t learn anything from the recent FDA filing on the actual percentage of calories provided by these ingredients - something admittedly difficult to get from manufacturers.
We also need to keep in mind that grain-free diets are extremely popular in the pet-owning population, especially I suspect in those that have regular veterinary care. So we should expect in any list to see a high background of grain-free diets. Certainly if an association exists, the high prevalence of legumes may be important. But the inclusion rate is even more important - there’s a big difference between 3% peas and 50% peas or pea derivatives (pea protein, pea starch). Low amounts of legumes have been fed to pets for a long time without any observed issues, and modest amounts of potatoes have also been included in diets, including kibble, for some time.
The animal proteins in the diets appear to be distributed across a range (chicken most common, followed closely by lamb, then salmon, whitefish, kangaroo, turkey, beef, pork, venison, etc.). Of course, for kangaroo, this is mostly reflective of a couple of specific diets as compared to chicken which is a wider range (but of course not many companies make a kangaroo diet nor is kangaroo sold as commonly as chicken).
Take-home messages: if you’re avoiding foods based on the list, the ingredients appearing most frequently are legumes. Here again, the amount of legumes isn’t addressed which is a big deficiency in our current knowledge of these diets when it comes to ingredients. It would be difficult to believe that 5% or less would be a problem, but in a larger amount, that’s a relatively recent phenomenon in pet food production driven by the push for grain-free diets.
What about detailed nutritional information?
Detailed information about the nutritional profiles of the diets has not yet been released from FDA testing. The guaranteed analyses are available for many of the diets, but keep in mind some have changed in the past few years. It’s also worth considering that the guaranteed analysis is not a good way to compare pet foods, a subject I and other nutritionists have talked about extensively.
At the present time, we’ll have to wait on this data. Of course, if there ends up not being an association between diet and DCM, it may not be particularly relevant anyway.
Why are some dogs improving with diet change?
Individual reports within the FDA data as well as reports from veterinary cardiologists suggest that pets improved following a diet change. This is a bit of an over-simplification because of the fact that many dogs were also given supplemental taurine at very high doses (higher than in pet foods), carnitine (a nutritional supplement which may have benefits in DCM), fish oil, and in severe cases, drugs. Even if a new diet helped, did it help because a particular dog or breed has a higher requirement for a nutrient than an ‘average’ dog? This would not be the manufacturer’s fault per se since they used the best available scientific information to formulate. The effect of diet alone is very difficult to isolate. In addition, the diets to which dogs were changed are often unknown and not compared to the previous diet.
This is a very complicated scientific problem with the need for careful analysis, and we don’t yet have the level of analysis required to draw any conclusions or even to compare dogs with and without DCM, of similar breed and age, fed the same foods. It’s also important to compare how often DCM occurs on a particular diet within the entire population fed that food as compared to the background incidence (how often DCM occurs in all dogs (or by breed)). For example, if 1 out of 10,000 dogs get DCM when fed any diet, and 1 out of 1,000 develop DCM on diet X, then diet X is certainly worth investigating. The total number of cases reported needs to be compared by the FDA to the number of total dogs exposed. If diet X is 10 times more commonly fed than diet Y, the total number of DCM cases would be expected to be 10 times higher for diet X than Y (even if diet X did not contribute to DCM - more dogs are exposed in the general population). In veterinary studies, we also talk of confounding variables, which basically means other factors that could complicate the ability to establish cause and effect, and this data that the FDA has obtained is complicated by confounding variables like drugs, supplements, breed, age, duration of feeding, calorie intake, microbiome differences, genetic mutations, etc. (which is not entirely their fault since the quality of data from these reports can be difficult to standardize and the amount of data needed is enormous). This is one of the most complex questions to answer that I’ve seen be asked in pet nutrition.
BEG (Boutique, Exotic, and Grain-free) - is this an appropriate guide?
This is a term being used by some veterinarians to describe the diets in question. I think any attempt at simplification at this point is misplaced. We just don’t have enough information. There are brands on the list that I wouldn’t consider to be boutique - and this really has no meaning anyway. Brands are often owned by a large pet food company, or their foods manufactured by a larger company, which is the case in the foods reported (including some of the biggest names in pet nutrition). Nevertheless, because a brand is smaller or sold in independent retailers or direct to consumer, are they inherently less reputable or capable? I think not, and some positive innovation in pet food is driven by smaller companies (some of which the large companies purchase or invest in when they recognize this). In full disclosure, I’ve consulted with companies of all sizes as a nutritionist.
What about exotic - the ingredients aren’t particularly exotic looking at the list of proteins and the individual diets being fed - certainly kangaroo diets may appear more frequently than we might expect since they are not as commonly sold as others but this needs to be examined with real data (rather than my impression). Is kangaroo itself likely to be a cause if there was an association between DCM and food? Doubtful if used appropriately and analyzed for nutritional content; if instead high ash meats (or any animal) were being used, it’s possible the protein and amino acids were not as digestible as they seem on paper. Remember that many meats and meals used in many pet foods contain cuts or trimmings of meat as well as bone after processing and can vary significantly from supplier to supplier and even from batch to batch. It’s also possible kangaroo being expensive was mixed with more legumes - again, we don’t have the data. In any respect, I encourage clients to reserve exotic proteins for food trials (testing for food allergies when necessary). But looking at the data, there are certainly a number of poultry, lamb, and fish diets - far more than exotic proteins - so it doesn’t seem if there was a definite link that it would be related specifically to the protein source (unless that protein source were not balanced with other proteins or that supplemental amino acids were not added to account for deficiencies in a final formula). Remember, however, that the FDA has said so far the testing of SCAAs and taurine appears similar in the diets tested.
Grain-free diets have been around for some time, long before it was a frequent appearance on the label. Remember the duck and potato diets for allergy testing - these have been fed for decades. There have been significant changes in how grain-free diets are formulated including the use of legumes and legume protein in higher amounts than done historically. This could be another area where the values look good in testing, but where there isn’t optimal absorption of nutrients (or possible interactions between nutrients which were unpredictable). Admittedly, there is no evidence that grain-free foods are superior to grain-containing foods when nutritional composition is similar, but we don’t yet have enough information to say all grain-free diets are harmful (and I’m absolutely positive this won’t be the case).
Nutrition isn’t simple, and the best diet for any pet is usually one that is individualized. Applying any cute label to a complicated and unproven problem does a disservice to the attempt to characterize the problem fully and give pet parents facts, not slogans.
How am I advising pet parents?
The first is to approach the situation as calmly as possible. The vast majority of pets are fed many of these diets without any reported DCM (whether DCM is related to food or not in the end) - it’s heartbreaking when any pet is sick and certainly every precaution should be taken as information becomes available. It’s understandable that trust in pet food has been eroded because of contamination, recalls, and controversies. Talking to your veterinarian about your pet, which is not the same as any other pet, is helpful as again, the best diet is individualized. I underscore that we haven’t yet proven a link between diet and certainly not what part of the diet or diet-pet interaction is to blame. These tips may prove helpful, however:
Consider feeding a higher protein diet. Dogs have requirements for essential amino acids (protein) and fatty acids (fat) but not for carbohydrates (although carb-containing ingredients, including peas and potatoes, provide energy along with a host of phytonutrients and other benefits in moderation). A diet with more than 75 grams of protein per 1000 calories is a good place to start, especially if your pet is overweight, doesn’t eat a lot of food, has a known heart condition, or is active to give pertinent examples. This is determined by calling the company or estimating from a guaranteed analysis. Feeding more protein generally provides more SCAAs, those taurine precursors I mentioned earlier.
Consider a diet with additive taurine. We again don’t yet know that taurine is a preventative or treatment factor in atypical DCM cases, but it has a high margin of safety. Diets with supplemental methionine can also be helpful and represent an alternate option (since that’s a taurine precursor in dogs). You can always ask a company how they determined if additional taurine or methionine was or wasn’t required - but you may not get an answer to such a technical question.
Avoid diets at the current time which appear to rely heavily on legumes to meet the protein content of the diet. This is difficult to evaluate on the label. If a fresh meat is first on the ingredient list, followed by a legume or legume protein, it’s likely the legume that provides more contribution to the diet since it’s dry whereas the meat is mostly water. Similarly, if there’s a number of legumes listed and only one meat, that could also be a sign. Or if a legume is the first or second ingredient. Legumes can afford some nutritional benefits, but their history as a significant protein source is less well known. It may end up being perfectly safe, but the reason for higher inclusions (vs. meat) is generally for sustainability, processing, or cost. There’s no reason to avoid legumes or potatoes entirely, and I’ve formulated diets of all types with both ingredients but at low inclusion rates, in the interest of full disclosure.
Consider varying the diet. I’m not a proponent of feeding the same diet for the life of any dog or cat. Nutritional variety helps to overcome any particular issues with a certain formulation. Also consider different types of diets - if you’re feeding kibble, other options can be explored in addition or alone (canned foods, pasteurized foods, fresh foods, balanced home-prepared diets, etc.).
Talk to pet food companies about their testing and formulation process. Do they work with nutritionists to formulate? Do they test the final product to ensure it meets the requirements? Both are preferable.
If you have a Golden Retriever on any diet, screening is an option both in terms of an echocardiogram and taurine level until we know more. There certainly are some Golden retrievers that may have DCM from taurine deficiency, and this may be due to genes and not diet alone.
Remember that vets and other industry experts (including those at companies) aren’t trying to be evasive when not giving answers about the potential DCM association to diet. Simply put, no one has conclusive answers and there are a number of people working to sort this out but it will unfortunately take time.
Comments on this blog below:
I’m an admittedly reluctant blogger, but felt it important to express my thinking on this issue for my clients. I will try to respond to comments as I’m able, and remember that the information above is my opinion and my opinion only based on the published data accessible to everyone (I know there is an unknown amount of unpublished data circulated in parts throughout the internet). I hope, like all pet parents, that definitive answers are coming. I believe the tips above are sound advice generally, and I admit that I do not have the answers (but also think no one yet has those). If you draw any conclusion, it’s that we need to be talking about diets, not brands, nutrients not impressions, and specifics rather than generalizations. It’s always easier to make a sweeping generalization, but in nutrition, that’s when we’re often wrong.
I am a certified pet food nutrition specialist. My concern working in a pet retail establishment was the amount of misinformation our customers received from their veterinarians concerning this matter.
Most provided their clients with unclear or misleading information and sent them back to retail stores with no idea how to proceed other than to add lots of grains back into their pets diets. Handing clients a warning letter from the FDA gave them little to go on and left us unraveling the small amount of information they were given.