New DogRisk Study Compares Risk of Allergies in Dogs Associated with Raw and Processed Foods

The subject of raw diets is a constant and controversial one in veterinary medicine. The controversy stems mostly from the fact that some people are passionately committed to the idea that raw diets have health benefits, but there is no scientific evidence to support this and plenty of evidence showing uncooked food increases the risk of parasitic and infectious disease.

Just this week, there was another series of cases in which people suffered serious foodborne illness due to contact with raw pet food. This is a frequent phenomenon, and there have been many studies in the last 10 years showing raw diets present real health risks to people and pets. There is no comparable body of evidence showing any health benefits. There are, however, a few studies like this that would claim to do so:

Hemida M, Vuori KA, Salin S, Moore R, Anturaniemi J, Hielm-Björkman A. Identification of modifiable pre- and postnatal dietary and environmental exposures associated with owner-reported canine atopic dermatitis in Finland using a web-based questionnaire. PLoS One. 2020;15(5). doi:10.1371/JOURNAL.PONE.0225675

This study comes from the DogRisk group, which I have written about previously-

This is a group of veterinarians and nutrition researchers led by Dr. Hielm-Bjorkman, a faculty member at the University of Helsinki. The organization is independent of the university, and it appears to focus on generating evidence to support raw diets, supplements, and a few other alternative health practices (such as acupuncture). The team is composed mostly of academic researchers, both vets and PhDs, working on issues of animal nutrition. Individually, they have legitimate credentials and research topics, but as a group they seem pretty dedicated to proving a set of pre-existing beliefs about raw diets rather than investigating the subject impartially. 

This paper appears to fit into the group’s campaign to validate their pre-existing beliefs about raw diets. The stated goal was to evaluate potential causal factors associated with the presence of allergies in dogs, including the diet of dogs with or without allergies and the diet of the mothers of these dogs. The data come from an online survey of owners, and it is currently only available in Finnish, confining the study to a narrowly limited population not representative of pets or pet owners in other places. 

Such online owner surveys are deeply unreliable. The diagnosis of allergies and the details of diet history are based entirely on the recollections and reports of pet owners, most of whom have establishes beliefs about pet health and nutrition and little or no scientific training in these subjects. Many such studies, in human and veterinary medicine, have shown that such information is often biased and not consistent with findings in more controlled research. The authors of this study actually report that almost 13% of the responses had to be discarded because they appeared to be “robot answers,” though it isn’t clear how this was determined. Clearly, this is a dubious source of data to begin with.

The authors compound the problem of reliability by imposing their own arbitrary and biased definitions on the variable they claim to be examining. They divided diet into two extreme categories: 1. A non-processed meat-based diet (NPMD) and 2. An ultra-processed carbohydrate based diet (UPCD). These are artificial categories that reflect the ideological bias of the researchers, not the reality of the kinds of foods dogs are actually fed. 

“Non-processed” is meaningless unless the dogs are basically eating whole prey and plants, which it is highly unlikely many of them are doing. The term “ultra-processed” is also a value judgment that implies a negative health impact that has not actually been shown to exist in pet foods. And commercial kibble, which is likely what the authors are referring to here, is not accurately described as “carbohydrate-based.” The carbohydrate content of commercial dry diets varies dramatically, as does the type of carbohydrate (from simple sugars to complex fibers). The nutritional value and potential health effects of commercial diets will also vary greatly based on these differences, and lumping them all together in this artificial category is not a legitimate scientific research approach; it is simply a way of stacking the deck in favor of the result the authors want to get.

The authors claim they are using a validated system of nutritional categories from human nutrition science, known as the NOVA Classification. This system, however, is not meant to apply to pet foods, and the definitions of “processed” used in the system have no relationship to how pet food is prepared. Here’s what the WHO says about the NOVA classification:

A number of commonly consumed processed foods and drinks [are] certainly or probably implicated in obesity and various chronic non-communicable diseases. These include energy- dense food products, “fast foods”, “convenience foods”, soft drinks, sugary drinks, various refined starchy foods, processed meat and salt-preserved foods. [note, none of these categories is relevant to properly formulated commercial pet diets]

Food processing in itself is not the issue. One obvious reason is that nowadays, practically all food is processed in some sense and in some way. The term ‘processing’ (like the term ‘industry’) is very general and therefore not helpful, and so judgements of foods simply because they are ‘processed’ are not meaningful. Food scientists and technologists and food manufacturers are right to emphasize the benefits of originally ancient and also relatively novel processes such as drying, non-alcoholic fermentation, chilling and freezing, pasteurization, and vacuum-packing. 

Confusing commercial pet diets with human convenience and snack foods because they are both produced by large industries and come in packages is misleading. The intent and formulation methods are vastly different, and the superficial similarities are not a legitimate foundation for assuming they have the same health impacts.

The NOVA classification also makes no reference to raw or uncooked meat and, in fact, states that, “These methods and processes are designed to preserve natural foods, to make them suitable for storage, or else to make them safe or edible or more pleasant to consume. Many unprocessed or minimally processed foods are prepared and cooked.”

Beginning this study with a distorted misinterpretation of terminology not intended to apply to commercial pet diets is a reflection of the underlying intent and bias behind the research. Other evidence of the authors pre-existing convictions come in the form of discussions of the “natural” diet of dogs as “carnivores” (myths used to promote raw diets which I have addressed before) and irrelevant reference to celiac disease in humans as a way to imply grains are unhealthy for dogs.

There is no apparent attempt to control for these biases in the methodology. The researchers designed the survey, defined the variables of interest, and had total freedom to create and conduct the regression analysis as they chose. It is well-established that such a lack of methodological control for bias leads to research which reflects the beliefs of the researchers more than the underlying reality of nature. 

This study was also funded in part by a company selling raw dog food. Funding bias is a known problem in medical studies, and while it is not an excuse for ignoring or dismissing research out of hand, it does require methods built into the study design to limit the impact of funding source on results, which were quite limited in this study. It is also common for proponents of raw diets to claim most veterinary nutrition research is useless because it is funded by pet food companies. It will be interesting to see if this community is consistent in ignoring this paper, which supports their belief, or if they prove interested in funding bias only when it helps dismiss research they don’t agree with.

Bottom Line
This study is a great illustration of ideologically motivated research. A group of scientists committed to a pre-existing belief design and conduct a study tailor-made to provide evidence supporting that belief. The authors do acknowledge some limitations to their project, but they consider them minor and accounted for, and they ignore some of the most significant problems I have mentioned here. 

This study fits into a series of efforts by this same research group to validate their belief in the benefits of raw diets, and unfortunately it doesn’t contribute much to actually understanding the pros and cons of raw foods for pets because it doesn’t contain much objective, reliable evidence. It will undoubtedly be useful as marketing r propaganda for proponents of raw diets, and much like research studies done on homeopathy, it confuses rather than clarifies the issue.