Cats and dogs: what is really in the bowl

The cat is an obligate carnivore. That is not a matter of taste, it is a list of enzymes it lacks. It has no amylase in its saliva, and its pancreatic and intestinal amylase is far lower than a dog's: starch is not its fuel, and its body was never equipped to handle much of it.

The other way round, four nutrients a human or a dog makes for itself have to arrive ready-made in its food:

• Taurine. It synthesises it poorly, and loses it constantly: it attaches its bile acids almost exclusively to taurine, where we also use glycine. A diet short of it brings dilated cardiomyopathy and retinal degeneration, shown in 1987 and reversible if corrected in time.
• Arginine. It cannot build it from the amino acids we start from. A single meal without it is enough to raise the ammonia in its blood.
• Preformed vitamin A. It does not turn plant beta-carotene into retinol. The carrot is a colour for the buyer, not a source of vitamin A for the cat.
• Arachidonic acid. Its liver makes very little of it from vegetable oils, lacking the enzyme that does the job in us. It is found in animal flesh and fat.

And in the opposite column, a requirement that does not exist: the cat has no dietary need for carbohydrates. It makes its glucose from protein, continuously. That does not mean a little starch poisons it, it means it pays for room in the bowl for something it can make on its own.

Putting the dog and the cat in the same sentence is the first mistake. In 2013, sequencing dogs and wolves showed that domestication selected three genes of starch digestion, among them AMY2B, the pancreatic amylase gene. The wolf carries two copies; dogs carry 4 to 30 depending on the individual, and each extra copy comes with higher amylase activity in the blood.

In other words, the dog spent several thousand years beside our granaries and its body adapted. Grain in its bowl is not an absurdity: the question becomes how much and which ones, not why. The nuance is worth knowing the other way too: some Arctic dogs and the dingoes do not have that multiplication, the adaptation did not reach every population.

The answer is known, it has been measured. Researchers gathered seven studies on feral cats, the ones living outdoors and hunting, and rebuilt the nutrient profile of their thirty diets. The result, as a share of the energy taken in over a day:

Two per cent. That is the stomach contents of its prey, nothing else. On a dry-matter basis it comes to 63 % protein, 23 % fat and 2.8 % carbohydrate.

Across from that, commercial dry kibble draws on average 36 % of its energy from carbohydrates, and up to 40 % for some. The gap is not a matter of premium brand against cheapest shelf: it is in the process. A kibble has to hold its shape, and extrusion needs starch for that.

Here is the quirk that changes everything: in Europe, a pet food label is not required to show carbohydrates, and no brand does it of its own accord. It shows the "analytical constituents": crude protein, crude fat, crude fibre, crude ash, moisture.

Those five lines, added up, never make 100. What is missing to reach 100 is the carbohydrate. The calculation is one subtraction, and it works on any bag, in the shop, before the till.

Here it is on a grain-free premium product, a fish kibble at a premium kibble price:

Twenty-three per cent of the bag. On a product that holds not one gram of grain and whose list opens with four fish. It is no scandal, it is the physical price of kibble, but it is a figure the buyer has to work out alone while paying not to have to.

"With salmon", "rich in chicken", "beef menu", "duck flavour": these are not free advertising words. Each one matches a minimum amount, set by the labelling code the industry applies in Europe.

Second rule to know: ingredients are ranked by weight at the time of manufacture, before cooking. Fresh meat is weighed with its water, about 70 % of it, and loses that water on drying. So it can take the first line of the list and weigh, in the finished kibble, far less than the ingredient behind it.

An example taken from a supermarket bag, kibble for neutered cats sold "with salmon":

Nothing in there is illegal or even hidden: it is all written down. But the word that fills the front of the bag weighs 4 % of what is inside, and the main ingredient is the one the cat has no need of at all.

The market's answer to all of the above fits in two words on the front of the bag: grain-free. Wheat, maize and rice have left the lists. Something had to replace them, since kibble needs starch to hold together: pulses. Peas, lentils, chickpeas, beans.

This is not a rumour, it is written on the premium bags themselves. The cat ranges of Orijen and Acana, two names often quoted as the best, list whole red lentils, whole peas, chickpeas, sometimes pinto beans, right after their fish and their meats.

What actually changes, measured across 77 cat kibbles: the grain-free ones hold 64 g of carbohydrate per 1,000 kcal, against 86 g for those with grain. Less, then. But not zero, and a long way from the 2 % of a hunting cat.

There was a less known sequel, on the dog's side. In 2018 the American drug agency opened an investigation after reports of dilated cardiomyopathy in dogs fed grain-free kibble: 515 dogs reported between 2014 and April 2019, and in the products named, peas or lentils in 93 % of cases, a grain-free claim in 91 %. The agency's conclusion deserves the same rigour: no cause-and-effect link has been established, it said so repeatedly, it never advised changing diet, and it stopped publishing in December 2022 for want of new scientific evidence.

Five checks, all doable on the bag, with no app and no expert:

• The first five ingredients. A named species (chicken, herring, turkey) is worth more than "meat and animal by-products", which is an open category and can change from one batch to the next.
• The subtraction. 100 minus the five analytical constituents. For a cat, under 20 % is already a good kibble, and over 40 % the bag is half starch.
• The real percentage of the animal in the name. 4, 14 or 26 % depending on the wording used. It is always written, in small type, next to the ingredient.
• The share of wet food in the day. Wet food is 75 to 80 % water. The cat descends from a desert animal, it drinks little and takes its water from its prey; a dry-only diet leaves it with less water overall.
• The word "complete". A complementary food, treats and most pure-meat pouches, does not cover the needs and should not be the meal.

From there, three ways of feeding hold up, in this order of simplicity: wet food of good composition, which settles the water and the carbohydrate in one go; kibble chosen on the five points above, possibly alongside the wet food; home-cooked or raw feeding, which gives the most control and demands the most rigour. That last one is not improvised: an unformulated home diet drifts quickly on taurine, on the calcium to phosphorus ratio and on vitamin A, and raw meat calls for the same hygiene as in our own kitchen. A recipe worked out by a veterinary nutritionist costs one consultation, once.

Read the sections above again with "muesli" in place of "kibble": you find, word for word, what our page on hidden carbs says about the breakfast aisle.

• "Premium" on the front, cereal as first ingredient on the back.
• "Rich in meat" guaranteeing 14 %, and fresh meat weighed with its water.
• A product rated excellent elsewhere, and a quarter of carbohydrate nobody shows.

The mechanism is identical because the reader is the same: they look at the front of the pack, and the front of the pack answers the question they are asking, not the one that matters. The difference is that the cat cannot turn the bag over.