The Pembroke Welsh Corgi is the sixteenth most registered breed in the United States. It is also the breed most likely to have a frightening statistic forwarded to you before you have finished reading about it.
That statistic is real, and it is not what it looks like. So we will take it first, slowly, with the arithmetic showing.
The gene test everyone will send you
There is a DNA test for degenerative myelopathy, a disease of the spinal cord in older dogs. It looks for one change in a gene called SOD1. Here is what the Orthopedic Foundation for Animals records for this breed, from 5,425 tested Pembrokes:
| Result | Share of the 5,425 |
|---|---|
| Two copies of the variant (“at risk”) | 53.8% |
| One copy (“carrier”) | 31.7% |
| No copies (normal) | 14.5% |
So 85.5% of tested Pembrokes carry at least one copy, and on the two-copy share this breed ranks second of every breed the registry lists. Read as a disease rate, that would be one of the worst numbers on this site.
It is not a disease rate. It is a genotype: a gene result, not a symptom. The registry itself says so: “not all dogs testing as A/A have shown clinical signs of DM”. The same page mentions dogs that reached 15 years old before any sign appeared.
How many two-copy Corgis actually get ill
This is the number that matters, and almost nobody quotes it.
The breed club’s own genetics write-up does. It reviews the one controlled comparison of clinical rates across breeds, a UC Davis analysis of the Veterinary Medical Database for 1990 to 1999. That database holds dogs taken to American veterinary teaching hospitals rather than pet Corgis at large. Hold on to that. In it, 0.6% of Pembroke Welsh Corgis were diagnosed with the disease. Set 0.6% against the 52.7% testing at risk and you get the write-up’s conclusion: “only 1.1% of Pembroke Welsh Corgis testing homozygous ‘at risk’ for the sod1 mutation will become clinically affected with DM.” Or, as it prefers to put it, 98.9% never develop it.
Now the caution, because that sum is doing a lot of work. It divides a clinical rate from one population by a gene rate from another, and the clinical half is a hospital caseload: the share of Corgis brought to a teaching hospital who were diagnosed, which is not the share of Corgis who fall ill. That count also took in Corgis of every age, and this illness arrives at a mean of about eleven years. Young dogs sit in the denominator having had no chance to show it. Diagnosis in life works by ruling other things out, and certainty comes only after death. All of it pushes the true figure up.
A different review of the same American database found 1.51% of Pembrokes affected (again, of those brought to US veterinary teaching hospitals), which would put the share of two-copy dogs falling ill nearer 3%. Take the widest reading and it is still a small minority.
Nobody has done the study that would settle it: enrol two-copy Pembrokes, follow them to the end of life, and count. Until that exists, the honest sentence is that the gene is nearly universal here, the illness is not, and the gap has never been measured properly. Cornell puts it more gently: even in a genetically predisposed dog, “their overall chance of becoming clinically affected may still be low”.
So what is the test good for
Two jobs, both real.
Choosing a mating. Two carriers bred together are predicted to produce 25% at-risk puppies. But the breed club argues against deciding on this result alone, and the reason is the frequency. It counts 87.8% of Pembrokes carrying at least one copy, a different set of tested dogs from the registry’s 85.5% above, which is why the two numbers do not match. Either way, breeding only from clear dogs would gut the gene pool to fix something that rarely happens. So the club advises breeders not to use the SOD1 result “unless there are confirmed DM affected close relatives”.
The American Kennel Club’s own list agrees by omission. For the Pembroke it asks for a hip evaluation and an ophthalmologist evaluation. The degenerative myelopathy DNA test appears on the Cardigan Welsh Corgi’s list, not this one.
Narrowing a differential in an old dog. If a twelve-year-old Corgi starts scuffing a back foot, the useful question is what else it could be. A clear result argues away from this disease. A two-copy result does not confirm it: the UC Davis laboratory is blunt that “other more common and treatable causes of progressive hind end weaknesses should still be considered even in homozygous DM/DM dogs”. Disc disease, arthritis and nerve compression all start the same way, and unlike this one they can be treated.
A second test exists for a modifier gene called SP110, offered only to Pembrokes already homozygous for the SOD1 variant. Only 71 dogs appear in that database: treat it as new rather than settled.
What it looks like if it does happen
It is not painful. It starts as wobbliness and weakness in the hind legs, and it progresses. Following 21 affected Pembrokes to the end, vets recorded a median of 19 months of clinical signs and a median age at euthanasia of 13 years. Every dog had lost the use of its back legs by then, and 15 of the 21 had front-limb weakness too. Owners are rarely warned about that last part.
One thing may help, on thin evidence. A Swiss team studied 50 dogs with suspected degenerative myelopathy, of which only 22 had follow-up. Among those, mean survival was 255 days with intensive daily physiotherapy (9 dogs), 130 days with moderate (6 dogs) and 55 days with none (7 dogs). No Corgis were in it, only four of the 50 were confirmed after death, and owners picked their own group. A lead, not a protocol.
One country ran the experiment
Japan shows what testing does at population scale. Researchers looked at SOD1 results from 5,512 Pembroke Welsh Corgis and watched the at-risk share collapse: from 14.5% (95 of 657) of dogs tested in 2019 to 2.9% (24 of 820) in 2022. Back in 2017, when adult testing began, it was 39.4% (163 of 414).
Two caveats. Who gets tested changed over that window (adults early, puppies later), so this tracks the tested dogs, not every Corgi in Japan. And the same team checked inbreeding and genetic diversity and found “no obvious changes”. Breeders moved the variant without narrowing the gene pool.
Thirty pounds on ten-inch legs
The panel above comes from the breed standard, so treat it as a specification rather than a survey of weighed pets.
The American standard asks for 10 to 12 inches at the withers and weight “not exceeding 30 pounds for dogs and 28 pounds for bitches”. It then names a ring ideal below both ceilings: about 27 pounds for a male and about 25 for a female “in show condition”. Carry those two to the vet. The panel’s upper numbers are caps, not targets.
The sentence that matters most is about shape. The standard asks that “the distance from the withers to the base of the tail should be approximately 40 percent greater than the distance from the withers to the ground.” That is a back roughly 1.4 times the height of the dog. It is not a description of an unlucky individual. It is the requirement.
The spine that comes with the legs
Short legs on a normal-length body have a cause, and in this breed it is known.
An extra copy of a gene called FGF4 (a retrogene, meaning a gene copy reinserted somewhere new in the genome) sits on canine chromosome 12. It does two things at once. It shortens the limbs, and it makes the cushioning discs between the spine bones harden years earlier than they should. A hardened disc can burst upward into the spinal cord rather than bulging slowly, which is why these episodes arrive without warning.
That single retrogene carried an odds ratio of 51.23 for disc disease, with a confidence interval of 46.69 to 56.20. The disc risk is inherited as a dominant, so one copy is enough to raise it.
Corgis carry a lot of it. The UC Davis laboratory measured the variant’s frequency in Pembrokes at 0.8323, with a confidence interval of 0.8109 to 0.8517, from between 500 and 999 dogs tested, about five copies in six.
A second FGF4 retrogene, on chromosome 18, shortens legs without the disc effect. This breed carries that one too. The team that mapped their separate effects named the pattern: “the breeds with the shortest legs, for example the Dachshunds and Corgis, have both retrogenes at high allele frequencies.” Two copies of the chromosome-18 version cut forearm length by 28%; two copies of the chromosome-12 one cut it by 10%.
The prevalence figure that does not exist
Here is where we stop, because the literature does. We could not find a published rate of disc disease in Pembroke Welsh Corgis. The largest survey of the question, covering 43,517 US dogs, reported 1.2% across all dogs, and the breed figures it puts forward are for Dachshunds and French Bulldogs. We could not open its full breed table, so we cannot tell you whether Corgis were counted separately at all.
A UK conformation study comes closest, and only in passing. Having measured 700 dogs, its authors wrote that some long-backed breeds “were relatively rare in this study population… they should still be considered high-risk due to their morphology”, and named the Pembroke Welsh Corgi as one. That is a judgement about shape, not a measurement of Corgis, and its risk curve was fitted to Miniature Dachshunds.
What that study did find across all 700 dogs is that heavier body condition went with higher risk. That part you control. Our guide to judging weight by hand covers how, and working out daily calories covers the arithmetic.
The Cardigan comparison, briefly
The other Welsh Corgi is the obvious control, and its numbers run the other way. On the same DNA test, 15.5% of 1,019 Cardigans tested at risk against the Pembroke’s 53.8%. Yet Cardigan hips are worse: 24.2% dysplastic of 4,126 evaluations, 36th of 219 breeds against the Pembroke’s 42nd. Two breeds built to the same brief, with the risks stacked differently.
Hips: a large-dog test on a small dog
The registry figure looks alarming for a 27-pound animal. Of 17,398 Pembroke hip evaluations, 21.8% were graded dysplastic, 42nd of 219 breeds, worse than the Rottweiler and the German Shepherd.
Two things pull in opposite directions on that number.
It is probably an underestimate. Owners submit voluntarily, and one study found normal hips were 8.2 times as likely to be sent in as abnormal ones. So the registry’s percentage is a floor.
And it overstates the trouble. Dysplasia is graded on a radiograph, and a dog can be graded dysplastic and never limp. This breed carries the best demonstration of that gap we have seen. Researchers measured 399 Pembroke Welsh Corgis using a laxity score called the distraction index, where a higher number means a looser joint. Every single Corgi scored above 0.30, with a mean of 0.66, laxity that goes with arthritis in large breeds. Yet only 6.8% had arthritis. Subluxation, where the joint sits partly out of place, showed up in 18%.
The authors’ own conclusion is the sentence to remember. Corgis “had a low frequency of conventional OA despite having hip laxity that has been shown to correlate with hip OA and hip dysplasia in large-breed dogs.”
Two markers did track trouble within the breed. A rim of new bone around the head of the femur raised the odds of subluxation 8.7 times and of hip dysplasia 8.9 times. A curved spur at the back of the femoral neck raised the odds of arthritis 4.6 times and of dysplasia 2.2 times. Subluxation itself raised the odds of arthritis 15.4 times.
So ask a breeder for the hip result. Then ask your vet what the grade means for this shape of dog rather than for a Labrador.
Living with a small dog bred to move cattle
One finding limits every sentence anyone writes about breed character. A 2022 genomics paper put behaviour questionnaires to the owners of 18,385 dogs and genotyped 2,155 of them. Its verdict: “breed explains just 9%” of the behavioural differences between individual dogs. Whatever a breed page tells you covers a tenth of the answer at best.
For this breed, even that tenth is barely measured.
The largest owner survey of breed differences in aggression collected two samples: 33 breeds recruited online, and 11 through breed clubs. We read both breed tables rather than assuming, and the Pembroke Welsh Corgi is in neither. The aggression rankings people quote from that paper do not contain this dog, in either direction.
A Finnish survey of 9,270 purebred dogs did include it, as one of 22 breeds analysed. The Pembroke has its own bar in that paper’s breed figure, with its own error bars. What we could not open is the supplementary table setting each breed against each other breed, and the discussion picks out other names.
Read its headline number carefully, though. Those 9,270 dogs are not a census. They are a contrast drawn from 13,715 surveyed: 1,791 with frequent aggressive behaviour toward people, set against 7,479 with none, and every dog in between left out on purpose. So the fraction you can make from those two numbers is a study design rather than a rate. The risk factors the survey landed on were age, being male, fearfulness, small size, having no other dog at home, and being an owner’s first dog. Not one is a breed.
On heel-nipping, we found nothing measured. The job involved moving cattle by pressing at their legs, so the folklore has an obvious origin. But no study we could open has counted nipping, chasing or herding behaviour in Pembrokes, and we will not convert a job description into a temperament claim.
What is documented is what the ring asks for: an “Outlook bold, but kindly. Never shy or vicious”, with judges instructed to dismiss an excessively shy dog. Selection has run against timidity, not toward it.
The first eighteen months
Growth itself is the easy part. This is a medium dog by weight, well under the sizes where the research on feeding for slower growth applies, and it finishes close to a year old. Keeping the puppy lean matters for the spine here, not for the joints.
Behaviour is where the year gets interesting. Researchers following young dogs recorded a dip in trainability around puberty. Their questionnaire arm covered 285 dogs at 5, 8 and 12 months (golden retrievers, Labradors, German shepherds and crosses), and carers scored trainability lowest at 8 months. A separate group of 93, goldens and Labradors only, was tested on a sit command at 5 and 8 months. That smaller group produced the strange finding: obedience fell for the dog’s own carer and not for a stranger. Two cohorts, two findings, often welded together. They should not be.
A UK study of the same stretch found complaints peaked later. Owners reported one or more unwanted behaviours in 42.1% of dogs at twelve months (513 of 1,219), falling away after. Barking rose from 7.9% at six months to 13.4% at twelve. Plan for the loudest stretch around the first birthday, and do not read it as a breed trait.
Neither study included Corgis. This is what happens to adolescent dogs, not to adolescent Pembrokes.
The rest of the check-up list
None of this is a diagnosis. These are things worth raising with your vet.
Bleeding, before any surgery. A DNA test for type 1 von Willebrand disease, a clotting disorder, shows 8.3% of 724 tested Pembrokes carrying one copy and 0.4% two copies, fifth among the breeds screened. The label needs care. UC Davis classifies this variant as dominant with incomplete penetrance, so a one-copy dog “may be affected”. That is not what carrier means on a recessive test. Signs run from nothing at all to prolonged bleeding after surgery or whelping, so mention it before a spay or a dental.
Collapse after hard exercise. The DNM1 variant behind exercise-induced collapse turns up here too: 11.0% carriers of 246 tested, none two-copy. It is not only a retriever finding: the team that surveyed the variant across breeds found carriers and affected dogs with a collapse phenotype in Pembroke Welsh corgis. Their striking figure, that 83.6% of two-copy dogs had collapsed by four years old, was measured in Labradors. Do not attach it to this breed.
Eyes, on the schedule. An ophthalmologist exam is one of the two things AKC asks Pembroke breeders for, and 5.6% of 10,271 exams came back abnormal, 80th of 219 breeds, unremarkable. No study we found names which eye conditions drive that figure here, so treat it as an appointment rather than a warning. Elbows sit at 5.5% abnormal of 2,488, 77th of 219.
Weight. In a UK sample of 22,333 dogs, vets recorded overweight status in 7.1% over a single year, and the highest odds went to the Pug, Beagle, Golden Retriever and English Springer Spaniel. The Corgi is not among them. Given the spine underneath, we would rather say no breed figure exists than repeat the internet’s certainty.
Lifespan, from one country. Japanese pet insurance records give the only measured breed figure we could open: 12.4 years for the Pembroke Welsh Corgi against 13.6 years for dogs overall in the same tables. Insured Japanese pets are not a random sample of the world’s Corgis. The big UK longevity work keeps its per-breed numbers in supplementary files we could not open, so we have one country’s answer and no second opinion.
What we could not verify
A disc disease rate for this breed. The gene is measured, the shape is measured, the consequence is not. Anyone quoting a Corgi percentage for back problems is almost certainly quoting a Dachshund.
How often the SOD1 result becomes illness. The 1.1% figure above is arithmetic on two populations, not a followed cohort. It is the best estimate published, and it is not a measurement.
A primary-care disorder study. For several breeds here we can give you what UK vets actually wrote down across a whole practice population. No such paper exists for this breed, which is why this page leans on a voluntary registry.
That Corgis are unusually greedy. The POMC deletion linked to appetite in Labradors was reported “absent from dogs of 38 other diverse breeds”, in a supplementary table we could not open. So we cannot even say whether Corgis were screened, and no equivalent work in this breed turned up.
Nipping, herding drive and children. Nothing measured, in any paper we could open.
A month-by-month Corgi weight chart. No published growth study exists. The table here is our own model run on the standard’s adult weights: a projection, not a record of real puppies.
Exercise in minutes a day. The usual figures trace to nobody. What the research supports is keeping the dog lean.
An eye disorder behind the registry figure. A French clinical and genetic study of persistent pupillary membranes in this breed exists; its publisher blocked us, and we will not summarise a paper we have not read.