Can dogs eat salmon? Cooked salmon, yes, and it is one of the better fish to put in a bowl. Raw salmon is a separate question with a separate answer, and the reason is not the one most owners expect. It is not mercury, it is not fat, and it is not the bones. It is a flatworm the size of a grain of sand that carries a bacterium capable of killing a dog in under a fortnight, and it exists in exactly the part of the world where salmon is most abundant.
The distinction between cooked and raw matters more here than it does for most foods, because the two versions of the same fish differ in risk by an order of magnitude while looking almost identical on a plate. A fillet that has been through a pan is a different object from a fillet that has not, and the gap between them is categorical rather than gradual.
Take a plain fillet, cook it through, let it cool, flake it, and you have a food that sits comfortably alongside chicken or white fish in a dog's diet. It is a complete protein, it is digestible, and it carries a fat profile that is unusual among meats.
The safety of cooked salmon rests on heat, and heat does two jobs at once. It denatures the proteins that parasites and bacteria depend on, and it destroys the organisms themselves. A fillet that has reached 63 °C at its thickest point and held there for a moment is not a risk anymore. It is dinner.
What cooking does not do is remove salt, and it does not remove fat. Those two survive the pan, and both matter for a dog, which is why the cooking method and the cut of fish you start with decide more about the outcome than the species does.
A note on the word cooked: it means heated through, not seared, not warmed, not "done on the outside". A fillet that is opaque at the centre is cooked. One that is translucent in the middle is raw fish wearing a cooked surface, and the parasite question has not been addressed at all.
Farmed Atlantic salmon and wild Pacific salmon are not nutritionally identical, and the gap is larger than most labels suggest.
Per 100 g, rawFarmed AtlanticWild kingTinned pink, drainedEnergyabout 208 kcalabout 180 kcalabout 139 kcalProteinabout 20 gabout 20 gabout 20 gFatabout 13 gabout 10 gabout 6 gSodiumabout 59 mgabout 50 mg300 to 400 mgEPA and DHA combinedabout 1.8 gabout 1.3 gabout 1.0 gThe farmed fish is fattier because it is fed to be, and the extra fat carries extra omega-3 as well as extra calories. For a dog, that cuts both ways: more of the nutrient you want, and more of the energy you have to account for.
The tinned column is the one that surprises people. It is lower in fat and lower in omega-3 than either fresh column, and much higher in sodium, because salt is added during canning. It is still a reasonable food, but it is a different food.
Salmon's reputation rests on two fatty acids, EPA and DHA, and the reason they matter to a dog is that dogs are poor at making them. A dog can elongate and desaturate plant-derived alpha-linolenic acid into the longer forms, but the conversion is inefficient enough that the pathway is not a reliable supply.
What the long-chain forms do in the body is structural. DHA is a component of cell membranes in the retina and the brain. EPA is the substrate for a family of signalling molecules that modulate inflammation, and that is the mechanism behind most of the clinical interest.
The evidence in dogs is strongest in two areas. In osteoarthritis, diets supplemented with EPA at around 40 mg per kilogram of body weight per day have been shown to improve weight-bearing and reduce the dose of pain relief a dog needs. In certain skin conditions, particularly those driven by allergic inflammation, higher intakes of EPA and DHA are associated with reduced itching and improved coat quality.
What salmon does not do is treat anything on its own. A fillet twice a week is a good food. It is not a therapeutic dose, and the gap between the two is wide enough that a dog with a diagnosed condition needs a measured supplement rather than a fish supper.
Owners tend to think in portions. The veterinary literature thinks in milligrams per kilogram, and translating between the two is the part that goes wrong.
The range that appears repeatedly in canine studies runs from about 20 to 50 mg of EPA and DHA combined per kilogram of body weight per day. A 20 kg dog therefore sits somewhere between 400 mg and 1,000 mg a day depending on what is being treated. Maintenance sits at the low end; osteoarthritis trials sit near the top.
Now put that against the table above. A 100 g piece of farmed Atlantic salmon carries roughly 1,800 mg of EPA and DHA. For a 20 kg dog, that single piece covers the therapeutic range for several days. Which means the dose question for food is not "how much can I give" but "how often", and the answer for a healthy dog is once or twice a week rather than daily.
There is a ceiling as well as a floor. Very high intakes of fish oil can cause loose stools, and in dogs with a bleeding disorder or on certain medications the antiplatelet effect of high-dose EPA is worth discussing with a vet before increasing anything.
This is the part of the salmon story that has no equivalent on any other food page, and it deserves the space.
Salmon poisoning disease is caused by Neorickettsia helminthoeca, a bacterium that lives inside a flatworm called Nanophyetus salmincola. The worm is a trematode with a three-host lifecycle: a snail, a fish, and a mammal that eats the fish. Dogs are that mammal often enough for the disease to have a name.
The bacterium does not survive outside the worm, and the worm does not cause much harm on its own. It is the passenger that kills. Once the worm encysts in the wall of a dog's small intestine, the bacterium is released into the lymphatic system, and from there it spreads to the lymph nodes, the liver, the spleen and the gut-associated tissue. The result is an acute, febrile, haemorrhagic illness.
The geography is specific. The lifecycle needs a particular freshwater snail that lives in rivers draining into the north-eastern Pacific, so the risk zone is the Pacific Northwest of the United States, British Columbia, and northern California. Salmon caught in the Atlantic, in Norway, in Chile or in a Scottish river do not carry it, because the snail is not there.
The timeline is fast. Signs typically appear five to seven days after the meal, sometimes as late as two weeks. Fever comes first, then vomiting, then diarrhoea that may become bloody, then profound lethargy and enlarged lymph nodes. Without treatment the mortality rate is high, often quoted above 90 per cent, and dogs can die within a week or two of the first sign.
Treatment exists and it works. Tetracycline antibiotics, given early, are effective. The difference between a dog that recovers and a dog that does not is almost entirely the delay between the meal and the diagnosis, which is why the single most useful thing an owner can do is remember the date of the fish.
Most advice about raw fish says freeze it first and the parasites are dealt with. That advice is correct for roundworms and tapeworm larvae, and it is wrong here, and the reason is worth understanding.
Standard parasite control through freezing follows a temperature and time schedule: minus 20 °C for seven days, or minus 35 °C for fifteen hours, which is the schedule food safety authorities publish for fish intended to be eaten raw. Anisakis larvae and Diphyllobothrium larvae are inactivated by it.
Nanophyetus metacercariae are not reliably inactivated by it. The encysted form is hardy, and the bacterium inside it is protected by the worm tissue around it. Reports of salmon poisoning in dogs fed fish that had been frozen are part of the veterinary literature, which is why the advice in the risk zone is heat rather than cold.
Cooking to 63 °C destroys both the worm and the bacterium. That is the intervention that works, and it is the reason the answer to the raw question in the Pacific Northwest is a firm no while the answer to the cooked question is yes.
Outside the salmon poisoning zone, raw salmon still carries the ordinary parasites of fish, and they are worth knowing by name.
Anisakis is a nematode that lives in the viscera of marine fish and migrates into the flesh after death. Dogs that eat raw marine fish can pick up larvae, which attach to the stomach or intestinal wall. Most infections are mild or silent. Some produce vomiting within hours of the meal, and occasionally a larva is found embedded in a piece of stomach lining during an endoscopy that was done for another reason.
Diphyllobothrium is a tapeworm with a fondness for fish that spend part of their life in fresh water, which includes Pacific salmon and most Atlantic salmon that return to rivers. The adult worm can reach several metres in a dog's intestine. It is usually asymptomatic, and it is usually found as segments in the stool or as an incidental finding.
Both are treatable with the standard anthelmintics, and both are prevented by the same thing: heat. Praziquantel covers the tapeworm; the nematode is less reliably covered by routine worming, which is another argument for cooking.
Smoked salmon is the version most likely to be offered, because it is in the fridge and it looks like a treat. It is also the version with the least in common with a plain fillet.
Cold-smoked salmon is not cooked. The process cures the fish in salt and exposes it to smoke at a temperature below 30 °C, which flavours it and dries it slightly without ever reaching a temperature that would kill a parasite. So cold-smoked salmon is raw fish plus salt, and it inherits both problems at once.
Hot-smoked salmon is cooked, so the parasite question is answered, but the salt is not. A typical cold-smoked product carries roughly 1,700 to 2,000 mg of sodium per 100 g. A 10 kg dog needs somewhere in the region of 100 to 200 mg of sodium a day. Two thin slices can therefore deliver a substantial fraction of a day's allowance in a food that provides almost nothing else.
There is a third issue, which is that smoked salmon is usually sliced thin and eaten as a delicacy, and delicacies come with a habit of being shared in quantities nobody counts.
Tinned salmon is a genuinely useful food for a dog, with three caveats that are easy to manage.
The first is salt. Most tins are salted, and the sodium figure sits around 300 to 400 mg per 100 g. Tins sold without added salt exist and are worth looking for if the fish is going to be a regular item rather than an occasional one.
The second is the bones. Tinned salmon is packed with the skeleton, and those bones are soft enough to be crushed between a thumb and forefinger, which means they are soft enough not to splinter. They are also a real source of calcium, and a dog that gets a lot of home-prepared food without bone may benefit from them. The catch is that they must be broken up. A whole vertebral segment swallowed is a choking hazard regardless of how soft it is.
The third is the oil. Tins packed in oil are higher in fat than tins packed in water or in their own juice, and the oil is usually a vegetable oil rather than fish oil, which changes the fatty acid profile in a direction that has nothing to do with salmon.
Drain the tin, break up the bones, and two spoonfuls make a reasonable addition to a meal for a medium dog.
Salmon skin is where the fat is concentrated, and the numbers reflect it. A piece of roasted salmon skin carries in the region of 40 g of fat per 100 g, which puts it close to pork belly on the fat scale and well above the fillet it came off.
That does not make it dangerous in a small quantity. It makes it a rich food, and rich food in a dog whose pancreas is easily upset, or who has a history of loose stools, is a predictable way to produce a bad night. The fat is also where a good share of the omega-3 sits, so there is an argument for leaving the skin on the fillet rather than stripping it off.
The version to avoid is the one from a takeaway, which has been salted, possibly battered, and cooked in oil that has been through several cycles. That combination delivers salt, fat and a coating in one bite.
Grilled skin from a fillet you cooked yourself, cut into strips, is a high-value training treat. It is not a meal, and it is not something to hand over whole.
Mercury is the first thing many owners ask about, and salmon is close to the best answer in the fish market rather than the worst.
Methylmercury accumulates up the food chain, so the fish with the highest levels are the long-lived predators at the top: shark, swordfish, marlin, orange roughy, bigeye tuna and tilefish. Food safety agencies in the United States publish a three-tier guide for people who are pregnant or feeding young children, and salmon sits in the lowest tier, the one described as best choices, alongside prawns, cod, tilapia and tinned light tuna.
The reason is lifespan and position. Most salmon are harvested at two to five years old and feed low on the chain, on krill, small fish and, in farms, on formulated feed. They do not have the decades of accumulation that a swordfish has.
For a dog the calculation is more favourable still, because the exposure is intermittent and the body weight is not the same as a human child's in the way that matters for a single meal. The mercury question is real for fish-eating species and for daily fish-based diets. It is not the limiting factor for a salmon dinner twice a week.
There is a food poisoning syndrome associated with fish that has been badly stored, and it is worth separating from the parasite question because the mechanism is different and the onset is much faster.
Scombroid poisoning is caused by histamine, which forms when bacteria in fish tissue convert histidine into histamine during spoilage. The conversion is not reversed by cooking, so a fish that has sat too long is a problem even if it is subsequently cooked properly. The classic culprits are the scombroid family, which includes tuna, mackerel, bonito and mahi-mahi, because those fish have high levels of free histidine in their muscle.
Salmon is not a scombroid fish, and it is not a classic cause of the syndrome. It is still a fish that spoils, and spoiled fish of any species can accumulate enough histamine and other biogenic amines to cause vomiting and diarrhoea. The signs come on within minutes to a couple of hours of the meal, which is much faster than anything infectious.
The practical rule is the same one that applies to any raw protein: keep it cold, cook it promptly, and do not rescue a fillet that has been out on a counter for an afternoon by cooking it harder.
Food allergy in dogs is less common than owners assume and more specific than they expect. The proteins most often implicated in canine adverse food reactions are beef, dairy, wheat, chicken, egg, lamb, soy and fish, in roughly that order of frequency across published case series.
Fish is therefore on the list, and it has a particular place on it because fish is often used as a novel protein in elimination diets. A dog that has been fed a fish-based hypoallergenic diet for months and then develops itching has a different problem from a dog that has never eaten fish.
The signs of a food allergy are not dramatic in most cases. They are itchy ears, itchy paws, recurrent skin infections, and sometimes vomiting or loose stool. Anaphylaxis is rare in dogs and is more often associated with insect stings or vaccines than with food.
Diagnosis is by elimination trial, not by blood test. The protocols that work involve feeding a single novel protein and carbohydrate for eight weeks with nothing else at all, then reintroducing the old diet and watching. A vet sets the protocol; a bag of salmon treats from a shop does not count as a controlled trial.
One of the standard warnings about feeding raw fish to dogs concerns thiaminase, an enzyme that breaks down thiamine, and salmon is not the fish that carries it.
Thiaminase is found in several freshwater and estuarine species, including carp, goldfish, smelt, and various clams and mussels. A dog fed a diet heavy in raw fish containing the enzyme can develop thiamine deficiency, which in dogs presents as neurological signs: wobbling, weakness, head tilt, and in severe cases seizures. The enzyme is destroyed by heat, so cooked fish is not a risk.
Salmonids are not a significant source. The warning gets attached to salmon because it gets attached to "raw fish" as a category, and category-level warnings tend to name the most familiar member. If the concern is thiamine, the fish to look at are the ones in the carp family, not the ones with a finned back and a pink stripe.
This is worth knowing because it changes the answer to a common question. A dog on a home-cooked diet that is heavy in fish needs a thiamine source, and the fish to be careful about are not the expensive ones.
The bone question for salmon depends entirely on how the fish was prepared, and the difference between the three common forms is not small.
A whole cooked salmon has a spine, ribs and a dense set of intramuscular bones, and those are not safe to hand to a dog. Cooked bone is brittle in a way raw bone is not, and a splinter off a cooked spine is an emergency waiting to happen. The fillet bones are small and sharp and easy to miss.
Tinned salmon is the opposite case. The canning process softens the skeleton to the point where it crushes between fingers, and those bones are safe once broken up.
A fillet, checked by hand and flaked, should have no bones at all. The check takes a minute and is worth doing twice, because the intramuscular bones in a salmon fillet are fine, flexible and exactly the colour of the flesh around them.
A puppy can have cooked salmon, in the same way a puppy can have any plain cooked meat: as a small part of a diet that is otherwise complete.
The constraint that matters is not the fish, it is the arithmetic. A growing puppy has a narrow energy budget and a high requirement for calcium, phosphorus and the full set of micronutrients. A food that provides protein and fat but no calcium displaces food that provides all three, and the displacement is the problem rather than the fish.
So the portion for a puppy is a taste, not a serving. A teaspoon flaked into the growth diet, or a small piece used as a reward, is the right scale. Give a puppy a bowl of salmon in place of its dinner and you have fed it a meal with the wrong mineral balance.
Raw salmon is a different answer again, and for a puppy the answer is no regardless of where in the world you live. The margin for a treat that carries an infectious risk is smaller in an animal whose immune system is still learning what to do.
Four kinds of dog need a different plate, and for none of them is it a question of portion size.
The first is a dog with pancreatitis or a history of it. Salmon is a fatty fish, and fat is the trigger for pancreatic inflammation in a predisposed dog. A lean white fish is the safer choice in that household.
Next comes a dog on a low-sodium diet for heart or kidney disease. Fresh cooked salmon is fine, because its sodium is low. Smoked salmon and most tinned salmon are not, because their sodium is not.
The third is a dog with a diagnosed fish allergy, which removes the food entirely rather than reducing it.
The fourth is a dog on a hypoallergenic elimination trial, where the entire value of the trial depends on nothing else being eaten. A piece of salmon during an eight-week trial does not spoil the trial slightly; it resets the clock.
Beyond those four, a dog with advanced kidney disease may be on a protein restriction that a fillet counts against, and a dog with a bleeding disorder may need the high-dose omega-3 discussion before any supplement is added.
Salmon oil occupies an odd position: it is the most convenient way to deliver what salmon contains, and it is the least like salmon.
A capsule or a pump delivers EPA and DHA without the protein, the fat load, the bones or the salt, which makes dosing precise and makes it usable in a dog that cannot have the fish. That is a real advantage, and it is why veterinary dermatologists and nutritionists reach for oil rather than fillets when a therapeutic dose is needed.
The trade is that the oil is a processed product with its own stability question. Polyunsaturated fatty acids oxidise, and oxidised oil is at best useless and at worst a source of free radicals. A bottle that smells sharp or fishy in the wrong way, or a capsule that has been in a warm cupboard for two years, is not a bargain.
The other trade is calorie density. A gram of oil is nine calories, and a therapeutic dose for a large dog is a meaningful fraction of a meal's energy. That has to be subtracted from somewhere, and the subtraction is usually the point owners forget.
The useful way to think about a portion is by calories rather than by weight, because a piece of fish is not a meal and the meal it is added to has its own budget.
A rough guide for a healthy adult dog is that treats and extras together should stay under about ten per cent of daily energy intake. For a 15 kg dog eating roughly 800 kcal a day, that is 80 kcal of extras, which is around 40 g of farmed salmon or 60 g of tinned.
For a small dog the numbers shrink fast. A 5 kg dog has a budget of roughly 25 kcal for extras, which is a piece the size of a thumbnail. That is not a stingy portion; it is the correct portion, and the reason it looks small is that a dog is small.
A simpler frame that works for most households: salmon as a meal topper twice a week, in a quantity that would fit in a tablespoon for a medium dog, or as a high-value reward in pieces smaller than a pea.
Cooked salmon is a protein-rich, moist food, and it behaves like every other protein-rich moist food once it leaves the pan.
The rule that covers it is the one food safety agencies publish for people: cooked food should not sit at room temperature for more than two hours, and in warm weather the window is one hour. After that, bacteria have had enough time to multiply to a level where reheating does not undo the damage.
In the fridge, cooked fish keeps for three to four days. Beyond that the risk climbs even if it still smells acceptable, because the organisms that cause trouble in a dog's gut are not always the ones that announce themselves through smell.
Freezing cooked salmon is fine and extends the window to about three months for quality, longer for safety. Thaw it in the fridge rather than on a counter, and do not refreeze a portion that has already been thawed.
The one thing not to do is leave a bowl down. A dog that grazes on fish for an afternoon has eaten a meal that spent the afternoon warming up.
The response depends on which wrong plate, and the three scenarios are different enough to separate.
If a dog ate cooked salmon off a plate, including a piece with butter, salt and pepper, the likely outcome is nothing at all. Watch for loose stool and nothing more. A large quantity of heavily salted fish in a small dog is the one version worth a call to a vet, because sodium load in a 4 kg animal is arithmetic that can go wrong.
If a dog ate raw salmon and you live in the Pacific Northwest, British Columbia or northern California, the advice is to call a vet the same day and say what happened and when. Salmon poisoning is treatable and the treatment is time-sensitive, so the conversation is worth having before there is a symptom rather than after. Prophylactic treatment is sometimes used in the risk zone.
If a dog ate smoked salmon, the concern is salt and richness rather than infection. Watch for thirst, vomiting and diarrhoea, and expect a soft stool.
If a dog ate a salmon bone or a piece of spine, watch for gagging, drooling, pawing at the mouth, or a change in swallowing. A bone that has passed into the stomach is usually fine. One stuck in the oesophagus is not, and a dog that cannot swallow water needs to be seen.
Fish is unusual among foods in that a single, memorable rule covers most of the risk, and the rule is that the fish a dog eats should have been cooked by someone in the household.
That rule handles the parasite, the bacterium, the salt, the oil and the bones in one stroke, because a fillet cooked at home is a fillet you can see, touch and check. It also handles the version of the problem that no list can cover, which is the plate left within reach at a party.
The second rule is about quantity, and it is the one owners break. Salmon is rich, and rich food in a dog that is not used to it produces a loose stool that gets blamed on the fish rather than on the volume.
The third rule is about frequency. Twice a week is a good food. Every day is a diet that has stopped being balanced, and the imbalance is in the minerals rather than in the fish.
PawAtlas keeps these three rules in one place because they answer the question in the order owners actually ask it: is it safe, how much, how often.
Cooked fillet, yes. It supplies complete protein and the two long-chain omega-3 fats that a dog's own metabolism barely produces from plant oils. Treat it as a quality ingredient rather than a treatment, because the clinical doses studied in arthritis and skin disease sit well above what a piece of fish provides.
The answer depends on where the fish came from, and along the north-eastern Pacific coast from northern California up through British Columbia it is no. A flatworm common in salmon there passes on a bacterium that causes an acute and often fatal illness. In other waters the raw-fish concerns are the usual nematode and tapeworm larvae.
For the nematodes and tapeworm larvae, yes: a week at minus 20 °C, or just over half a day at minus 35 °C. The flatworm that carries the salmon poisoning bacterium survives that treatment, so cold is not a substitute for heat in the risk zone.
Better not. The cold-smoked product never reaches a cooking temperature, so what you are holding is uncooked fish under a heavy salt cure, and the load carried by two or three slices dwarfs what a meal of that size ought to contain.
Yes, once drained and once the softened skeleton has been mashed. Choose a water-packed tin over an oil-packed one, and read the label for added salt, which varies a lot between brands.
Plain roasted skin, cut into narrow strips, is fine. It is the fattiest part of the fish and will upset a sensitive gut in a quantity that would be modest for the fillet, and the battered takeaway version should not be given at all.
Think in energy rather than in grams. Extras of every kind belong inside a tenth of the day's calories, which for a medium dog works out to a small flake or a couple of spoonfuls, and a toy breed gets a fraction of that.
It is one of the lowest-mercury options in the market, because the fish is short-lived and eats low in the chain. The species that accumulate the most are the old predators — swordfish, shark, marlin and bigeye tuna — not salmon.
As a taste, yes. The reason to keep it small is displacement: a puppy's bowl has to stay complete, and filling part of it with fish takes away the calcium and phosphorus a growing frame needs.
The crumbly ones in a tin, after mashing, are the only kind that qualify. A spine or a rib from a cooked whole fish splinters, and a splinter in the throat is a veterinary emergency.