Feeding Your Dog — What the Peer-Reviewed Science Actually Shows

What We Feed and Why

Every puppy that leaves Northwest Lagotto goes home eating Royal Canin Medium Puppy. That is not a sponsorship. It is a decision I made on my veterinarian’s recommendation eleven years ago and have not had a reason to change.

Royal Canin is one of the oldest pet nutrition companies in the world. They conduct ongoing research. Their product line supports breeders specifically — from D42 prenatal food through mother-and-puppy starter food to size-specific puppy formulas — which means the nutritional profile is calibrated to each stage of development rather than offering a single “puppy food” and hoping it covers everything.

Other brands make strong claims. Some may be excellent. But what matters more than the brand label is whether the company meets the criteria that the World Small Animal Veterinary Association’s Global Nutrition Committee established for evaluating pet food manufacturers. The WSAVA guidelines are the closest thing to an industry standard, and they are deliberately unglamorous: does the company employ a full-time, board-certified veterinary nutritionist (diplomate of the ACVN or ECVCN); do they own and operate their own manufacturing plants rather than contract out; do they conduct AAFCO feeding trials rather than relying solely on formulation; and do they publish research in peer-reviewed journals?

A handful of companies meet all four criteria. Royal Canin, Purina Pro Plan, Hill’s Science Diet, Eukanuba, and Iams are the most commonly cited. Most boutique and “premium” brands do not — not because they are malicious, but because they are small operations without the infrastructure to employ full-time nutritionists or conduct formal feeding trials. When grain-free diet-associated dilated cardiomyopathy began emerging around 2018, almost all of the implicated brands were outside this WSAVA-compliant group. That is a pattern worth paying attention to.

Whatever you feed, it should carry an AAFCO “Growth” or “All Life Stages” statement for puppies, have a correct calcium-to-phosphorus ratio (1.1:1 to 1.4:1), and be grain-inclusive unless your veterinarian has specifically directed otherwise for a medical reason. These three things are non-negotiable. Everything else is preference.

The Calcium Rule — the Most Important Thing in This Essay

This is not opinion. It is settled physiology, demonstrated in a series of elegant Dutch studies by Hazewinkel and colleagues at Utrecht University beginning in the late 1980s and continuing through the 2000s. The key paper is Tryfonidou and colleagues’ 2002 study in the Journal of Nutrition, which showed that growing dogs absorbed approximately 53% of dietary calcium through passive, non-regulated mechanisms — meaning the more calcium you feed, the more the puppy absorbs, with no mature feedback system to shut the process down. A follow-up study by Mack and colleagues in 2015 confirmed that adult dogs develop the ability to regulate calcium absorption; puppies do not. This difference is the entire reason puppy food and adult food are formulated differently, and it is the reason why calcium supplementation in puppies causes the specific, irreversible skeletal diseases that every veterinary orthopedic textbook describes: osteochondrosis dissecans, hypertrophic osteodystrophy, angular limb deformities, and contributions to hip and elbow dysplasia.

Ca:P ratio
1.1:1 to 1.4:1
NRC / AAFCO
optimum

A complete and balanced puppy food already contains the correct ratio — typically 1.2–1.8% calcium on a dry matter basis, with a calcium-to-phosphorus ratio between 1.1:1 and 1.4:1. This range is backed by the National Research Council’s nutrient requirements and reflected in AAFCO’s nutrient profiles. Böswald and colleagues’ 2019 factorial analysis in PLOS One refined these numbers further and is the most current primary source. Adding calcium on top of a complete food disrupts the ratio and overwhelms a system that has no mechanism to protect itself.

If someone recommends adding calcium to your puppy’s diet — a breeder, a trainer, a website, a neighbour — politely decline. If a website recommends raw meaty bones or bone meal as a “natural” supplement for growing puppies, close the tab. This is one area where the science is settled and the consequences of getting it wrong are permanent.

Lean Is Healthy — Body Condition Scoring

Those numbers come from three papers: Kealy and colleagues’ 1992 paper in the Journal of the American Veterinary Medical Association reporting the hip dysplasia findings at age two, Kealy and colleagues’ 2002 follow-up reporting the lifespan data, and Lawler and colleagues’ 2008 paper in the British Journal of Nutrition synthesising two decades of observations across the cohort. All three are open-access or widely available and every claim above is verifiable at its primary source.

One honest caveat: Labradors are a high-risk breed for hip dysplasia. The absolute effect size of body condition on hip dysplasia in a breed with much lower baseline risk — Lagotti, Border Collies, most terriers — is likely smaller. But the direction of the finding has been replicated across every breed studied. The lifespan and osteoarthritis delays are not breed-specific. A 2019 study by Salt and colleagues using Banfield Pet Hospital records examined over 50,000 dogs across 12 breeds and confirmed that overweight dogs had shorter lifespans. A 2024 retrospective of 4.9 million dogs (Salt et al., Preventive Veterinary Medicine) found that dogs overweight during the growth phase had nearly twice the odds of being overweight as adults (odds ratio 1.85). The Purina study is not a curiosity. It is the closest thing canine nutrition has to settled science.

Your dog should live between a 4 and a 5 on the 9-point body condition scale — what the scale calls “ideal.”

The most common feedback I get from families is that they think the puppy looks too thin. Visitors comment. Neighbours comment. If you can feel the ribs easily and see a waist, the puppy is not too thin. It is exactly where it should be. Trust the science on this one — the Purina study is as close to definitive as nutritional research gets, and the twenty years of replication since have only strengthened it.

Feeding Schedules and Portions

Measured meals, not free-feeding. This is non-negotiable. Dogs fed from a bowl left full all day eat substantially more than dogs fed measured portions on a schedule, and according to the Association for Pet Obesity Prevention’s 2024 survey, approximately 59% of dogs in the United States are classified as overweight or obese — 37% overweight, 22% obese. Portion control is one of the simplest and most consequential interventions you can make.

Measured meals also let you notice appetite changes — one of the earliest signs of illness in any dog. A dog that skips a measured meal tells you something. A dog that grazes from a bowl all day tells you nothing until the problem is advanced.

Seven-month body condition predicts adult weight. Serisier and colleagues’ longitudinal study of female Beagles, published in BMC Veterinary Research, identified body condition score at seven months as a direct predictor of adult obesity. The larger 2024 Banfield retrospective of 4.9 million dogs confirmed the effect at scale. What you set in the first year is what you will manage for the rest of your dog’s life.

Feeding as a Mind Game

Meals are the biggest mental event in most dogs’ days. If you pour food into a bowl and walk away, you are wasting a resource that could be doing real enrichment work, and for working breeds like Lagotti — but also for Border Collies, Retrievers, Poodles, and any other dog bred to think hard — this matters more than most owners realise. A dog that has to work for its food uses its brain. A dog that uses its brain is calmer, more satisfied, and better behaved.

I tell families to feed at least one meal a day through something that requires effort. The specific object is less important than the principle: slow down, work for it, engage.

Meals become the day’s richest enrichment, not an afterthought. Dogs that work for their food are calmer, more satisfied, and less prone to the destructive behaviours that come from understimulated intelligence. If you have a working breed, this is not optional; it is the difference between a settled dog and a frustrated one.

Joint Supplements — What the Evidence Actually Supports

The joint supplement market is enormous, largely unregulated, and — this matters — mostly built on marketing claims that do not match what the peer-reviewed veterinary literature actually shows. Before I tell you what works, I want you to understand the kinds of claims you will encounter when you start shopping. Most of them will not survive scrutiny.

With that context, here is what the evidence actually shows — ranked honestly by strength of peer-reviewed support.

The Prevention Question

I want to be direct about something that the supplement industry obscures. The entire canine joint supplement literature is a treatment literature. Every trial cited above was run on dogs with diagnosed, radiographically confirmed osteoarthritis — usually middle-aged to senior dogs with clinical lameness. There is essentially no peer-reviewed evidence that supplementing a structurally sound, healthy two-year-old dog prevents joint problems later in life. None of the omega-3, green-lipped mussel, UC-II, or glucosamine studies were run on healthy young dogs as a prevention protocol.

This means when you see a supplement marketed for “preventive joint support” in young dogs, that is an extrapolation from treatment data, not a finding. It is the most unsupported claim in the joint supplement market, and also one of the most common. The best-evidenced prevention for young dogs is exactly what the Purina study showed: stay lean, protect growth plates, feed a complete diet. Everything else is optional.

The Raw Feeding Question

Raw feeding — unpasteurised raw meat-based diets (RMBDs), whether commercial frozen-thaw products or homemade — is one of the more contested areas in companion animal nutrition. Proponents cite ancestral diet arguments, coat and energy improvements, and mistrust of kibble processing. The peer-reviewed literature on risks is unambiguous; the peer-reviewed literature on benefits, beyond short-term coat and stool appearance, is essentially absent.

I do not feed raw, I do not recommend it, and I want to explain why with specific numbers rather than blanket warnings.

Pathogen contamination

The FDA’s Veterinary Laboratory Investigation and Response Network (Vet-LIRN) conducted a surveillance study published by Nemser and colleagues in 2014 in Foodborne Pathogens and Disease. They tested 196 samples of commercial raw pet food and found 15 positive for Salmonella, 32 positive for Listeria monocytogenes, and 9 positive for both — a total of 38 contaminated samples, or approximately 20% of raw products tested. For comparison, the same study tested 860 samples of dry kibble, semi-moist food, and treats and found one positive for Salmonella. Raw pet food was between 15 and 24 times more likely to carry foodborne pathogens than conventionally processed food. Similar or worse contamination rates have been found in European studies (Vecchiato 2022; van Bree 2018).

Nutritional adequacy

Freeman & Michel (2001, JAVMA) evaluated 5 raw diets — 3 homemade and 2 commercial — and found that every single diet had multiple nutritional imbalances or deficiencies that would cause long-term health problems. Twenty years later, Vecchiato and colleagues (2022, Animals) analysed 44 commercial raw meat-based diets. Every product in the study had at least one nutritional concern, and 96% failed hygiene quality thresholds applied to products for human consumption. Davies, Lawes & Wales reviewed the field in the Journal of Small Animal Practice in 2019 and summarised that “investigations of both home- and commercially-prepared raw diets commonly have identified nutritional problems.” The patterns repeat across decades and across countries.

Professional position statements

Every major veterinary organisation — the American Veterinary Medical Association, Canadian Veterinary Medical Association, World Small Animal Veterinary Association, American Animal Hospital Association, plus the FDA and CDC in the United States — has issued a position statement recommending against feeding raw meat-based diets to companion animals, primarily because of the zoonotic transmission risk to humans (especially children, immunocompromised individuals, and elderly people in the household). That level of consensus across independent bodies is uncommon in veterinary nutrition and worth taking seriously.

If you want the purported benefits of home-prepared food without the raw risks, the evidence-based alternative is cooked home-prepared diets balanced using a veterinary nutritionist service. BalanceIT.com (University of California-Davis) is the most commonly cited resource; the recipes are built from USDA food database values, balanced to AAFCO profiles, and adjustable to individual dogs. Home cooking is more work than kibble, but it removes the raw pathogen risk and removes the adequacy problems that come from ad-hoc recipes.

Grain-Free, Legumes, and DCM

This is an area where the evidence has continued to develop, and I want to be careful to represent it accurately because much of what you will read online is either oversimplified or actively misleading.

In July 2018 the FDA announced it was investigating reports of dilated cardiomyopathy (DCM) in dogs — including in breeds without known genetic predisposition — eating diets labelled “grain-free” that contained peas, lentils, other legume seeds (pulses), and potatoes as main ingredients. In December 2022 the FDA stopped publishing public updates, citing insufficient data to establish causation. This is not the same thing as exoneration. Independent peer-reviewed research has continued since then, and the picture that has emerged is nuanced:

• The link is specifically to high-pulse, legume-rich diets, not to grain-free as a category. The active ingredient of concern is peas, lentils, chickpeas, and similar pulses appearing in the top 10 ingredients, particularly as primary ingredients.
• Multiple 2023–2025 studies continue to find cardiac differences. Owens et al. 2023 (JVIM) found lower ejection fraction and larger left ventricular volumes in healthy dogs on non-traditional diets versus traditional. Coppinger et al. 2024 (JVIM) found significantly more ventricular premature contractions in Irish Wolfhounds eating high-pulse diets. A November 2025 narrative review in Veterinary Sciences summarised the accumulated evidence and concluded that “these findings suggest a strong link between diet and DCM, particularly regarding legumes.”
• Some industry-funded studies have found no cardiac effect (Leach et al. 2023, Morris et al. 2025). These are real studies and should not be dismissed, but they have shorter follow-up, different methodologies, and industry sponsorship. They demonstrate that not every dog develops detectable change on a grain-free diet; they do not disprove the link.
• Many affected dogs improved after switching to grain-inclusive, low-pulse diets (Freid et al. 2021, Haimovitz et al. 2022, Walker et al. 2022). The condition appears at least partially reversible when caught early.

There is no demonstrated health benefit to grain-free diets for dogs without a diagnosed grain allergy. And as I will show in the next section, true grain allergies are uncommon. The grain-free category exists because it sells well, not because it represents a nutritional advance.

Food Allergies — What’s True and What’s Marketing

Food allergies are one of the most over-diagnosed conditions in pet nutrition. Two facts, well-supported in the peer-reviewed literature, tell you most of what you need to know.

First, true food allergies are rare. The petMD reference figure, based on a range of clinical studies, puts the prevalence of food allergies in dogs at roughly 0.2% — one dog in five hundred. Most “my dog has allergies” cases are environmental (flea, pollen, dust mite) or are something other than allergy entirely — yeast overgrowth, parasitic, or dermatological conditions unrelated to diet.

Second, when food allergies do occur, the offending ingredients are not what the marketing suggests. Mueller, Olivry & Prélaud’s 2016 systematic review in BMC Veterinary Research analysed the peer-reviewed literature and identified the most common food allergens in dogs by prevalence:

Note what this table says. Animal proteins — beef, dairy, chicken, lamb — occupy the top five slots. Grains as a category appear once (wheat, 13%). The entire grain-free category is premised on the idea that grains are the common allergens, and the peer-reviewed data does not support that. If your dog genuinely has a food allergy, it is overwhelmingly more likely to be to a protein source than to a grain.

If you suspect your dog has a food allergy, the only reliable diagnostic is a veterinarian-supervised elimination diet trial — typically 8 to 12 weeks on a novel or hydrolysed protein with no other food sources allowed, followed by controlled reintroduction. Blood and saliva tests for food allergies are not supported by the veterinary dermatology literature. They produce reliable-looking results that do not match diagnostic elimination trials.

Treats That Earn Their Place

Treats should make up no more than 10% of daily caloric intake. That is the standard figure from the American Animal Hospital Association’s weight management guidelines, and it is more restrictive than most families realise when they actually count what they give.

I use treats strategically rather than affectionately. The best treats are the ones your dog will work for but that also fit your body condition management: small, low-calorie, easy to deliver. Pieces of the puppy’s own kibble count. So does a frozen carrot. So does a sliver of lean chicken breast. What does not count is a continuous drip of high-calorie training treats, cheese, peanut butter, and leftovers from dinner plates — those add up to far more than 10% of the daily calorie budget in most households, and they are the single biggest driver of the 59% obesity figure the APOP reports.

Things to avoid: rawhide (choking and obstruction risk, variable manufacturing), cooked bones of any kind (splinter risk), pig ears and bully sticks in quantity (calorie-dense and often contaminated), xylitol-containing anything (see toxic list below), grapes and raisins (see toxic list below), fatty scraps from the table (pancreatitis risk).

Things that work: the puppy’s own kibble, frozen vegetables (carrots, green beans), small pieces of apple (no seeds), blueberries, plain cooked lean meat, commercial training treats that list a single protein and that you count against the 10% budget. Nothing fancy. Nothing that breaks the pattern.

Toxic Foods — The Complete List

Some foods are harmless to humans and dangerous to dogs. The mechanism varies — some are toxic at small doses, some cause damage only in quantity, some are individually variable. Print this list or save the interactive version below and make sure everyone in the household knows it. Paediatric accidents in children are a good analogy: the hazards are not hidden, but they do need to be known by everyone who interacts with the dog.

If your dog has eaten any of these, the two numbers you want saved in your phone are your regular veterinarian and the ASPCA Animal Poison Control Center at (888) 426-4435. Poison Control is staffed 24 hours a day by veterinary toxicologists; there is a consultation fee but they are the people who actually wrote the textbook chapters on this stuff, and they can guide you through whether the amount your dog ate warrants emergency care.

Nine Rules for Feeding Your Dog

Everything in this essay distilled into a reference card. Save it, print it, share it.