Friday, October 16, 2009

Kibble Can't Cause Illnesses When All Pets Eat Kibble

Here comes a scientific lesson that may be an A'HA experience.

Suppose that wheat gluten causes allergies in people (and it does cause allergies in some). Suppose that 100% of the population studied consumed the same amount of wheat gluten every day. We want to determine causes of allergies in this population. We look for differences between those who have allergies and those that don't. Even though wheat gluten is the cause of allergies in some members of this population, we won't find it, because consumption of wheat gluten does not vary in this population. Constants cannot be detected as causes of differences.

Instead, we find that the observed allergies run in families, that allergic parents are more likely to have affected children than non-allergic parents. Thus, the study will find that allergies are due to genetic differences among individuals, which is also true. Individuals do vary genetically in susceptibility to most or all major illnesses, but they do not become ill unless exposed to the illness-producing agent. We will not find the major culprit, wheat gluten, because there was no variation among individuals in the consumption of wheat gluten. To find the culprit, wheat gluten, we need to study a population in which consumption varies.

Even in a population where wheat consumption varies, there is a problem: Amount of wheat gluten consumed may be negatively correlated with allergic responses to it -- people who get sick from eating breads, pastas, breakfast cereals, etc.. are likely to consume less of these foods. We may find that people with allergies in this population eat less wheat gluten, not more. Based on the correlation between lower wheat gluten consumption and more allergies, some (scientifically ignorant) investigators might recommend that allergic people eat more wheat gluten.

Now let's look at pet disorders. More than 90% of dogs today are fed exclusively on commercial kibbles -- the same cooked carbohydrates at every meal for years on end. Even though kibble causes most canine diseases (and it does), studies of dog illnesses will not find kibble to be a cause, because there is too little diet variation in the population to be detected as a cause of illnesses. Even though cooked starches are the major cause of dogs' illnesses, the only differences between sick and not-yet-sick animals are genetic differences and non-dietary environmental differences. Effects of the culprit kibble cannot be found when all animals eat kibble.

It is easy to imagine how good studies of dietary effects on canine heath would be designed. For one, let's gather health data on three diet groups: (1) commercial kibble, (2) home-cooked BARF, and (3) raw-meaty-bones. We start with puppies at weaning and assign them randomly to one of the three diet conditions. For at least two years, and preferably longer, at frequent intervals we measure a variety of growth, behavioral, and health outcomes. Puppies are brought to a data collection site, where data collectors are blind to their diet assignment. Unlike people, dogs do not have hypotheses about their feeding conditions, so a single blind study is probably sufficient. At the end of data collection, we sort the measurements into the three diet groups and look at how well or poorly dogs fed on different diets scored.

Another study could sample existing groups of raw-fed, BARF-fed, and kibble-fed dogs, with histories of prior feeding documented. Histories could include the feeding background of the dogs' parents to detect trans-generational effects. Doubtless, there would be differences among the groups, other than feeding practices, such as amount of exercise, grooming, human contact, and so forth -- all of which could have health implications. Using naturally occuring groups to study the effects of dietary differences requires a sophisticated statistical approach to control for other differences that will occur, but the evidence against kibble diets may be so overwhelming, it is impossible to explain away with other "causes".

Such studies are "not rocket science" but a run-of-the-mill, scientific approach to a $15 billion a year question about the harm done to pets' health by commercial pet foods. Do you wonder why such straight-forward studies have not been done? Two problems: Funding and finding investigators. Pet food companies fund pet nutritional research, but they would not fund a study to compare their kibble concoctions with raw-meaty-bones or even home-cooked BARF diets. They already know kibble is a bad diet for pets, which is why they keep tweaking its nutrients and textures. Veterinary research investigators are so indebted to the pet food industry for research support, they are unlikely to risk that support and their careers to conduct such studies. So, without veterinary collaboration and funding, honest pet diet studies are not done.

An important aside: kibble fed animals are the baseline for all pet research studies. When dogs with infected mouths and long-term stressed immune systems are included as healthy subjects in a study of bowel or heart disorders, how valid can the results be?

So, what implications can we draw from this lesson in scientific follies? If you want to know what effects a junk-food diet has on human health, you'd better find a population that varies in its consumption of such foods. Very few people today eat only the recommended diet of fresh fuits and vegetables, lean meats, and whole grains, so junk foods probably do not get as bad a rap as they should. If you want to assess the effects of kibble on pets' health, you need a population that includes a substantial percentage of non-kibble eaters -- virtually impossible to find in developed countries today. Experiments and small scales studies of naturally-occuring diet variation can be designed and conducted, but they won't be.

No comments:

Post a Comment