Two von Willebrand tests, two different questions

A common scenario: a Doberman Pinscher's DNA report comes back "at risk" for von Willebrand disease (vWD) type 1, and before her spay the veterinarian also asks for a von Willebrand factor antigen (vWF:Ag) blood test. If the DNA report already found the variant, why run a second test? Because the two tests answer different questions, and neither one on its own tells you whether a particular dog will bleed.

The DNA test asks a permanent, inherited question: Does this dog carry a specific, known variant in the VWF gene that it can pass on to puppies? The antigen assay asks a current, measured question: How much von Willebrand factor is circulating in this dog's plasma at the moment the sample was drawn? Neither of those is the same as "will this dog bleed during surgery," and the antigen test measures the amount of factor, not how well that factor works.

The DNA test is typically run on a cheek (buccal) swab, and some laboratories also accept an EDTA whole-blood sample. Because genes do not change with age, health or heat cycles, one valid result lasts the dog's lifetime. Its limits are just as fixed. Each test detects only the specific variant it was designed for, and the UC Davis Veterinary Genetics Laboratory (VGL) states that test selection must be breed-appropriate. A clear result therefore clears only the tested variant. An at-risk result does not forecast bleeding either: VGL says dogs with one or two copies of the type 1 variant "may be affected," and not every dog with the variant shows signs. Cornell's Riney Canine Health Center advises that genetic testing identifies at-risk individuals and should be followed by blood tests that measure the amount of vWF.

The vWF:Ag assay measures the concentration of von Willebrand factor in a citrated plasma sample and reports it as a percentage of a pooled 100% standard. It works for any dog, purebred or mixed, and it is the test most commonly used to diagnose vWD. But it is a snapshot. Cornell's comparative coagulation laboratory notes that levels fluctuate from day to day, more so during pregnancy, heat or systemic illness, and that clotted or hemolyzed samples can give inaccurate results. Values between 50% and 69% fall in a borderline zone that cannot classify a dog as carrier or clear. A concentration figure also says nothing about the structural defect behind type 2 disease.

What von Willebrand disease is - and why routine clotting panels miss it

Von Willebrand disease is the most common inherited bleeding disorder in dogs. The Merck Veterinary Manual describes it in nearly all breeds and in mixed breeds, and VCA and Cornell's eClinPath resource note that it has been identified in more than 50 breeds. It is caused by a shortage or a defect of von Willebrand factor (vWF), a large multimeric glycoprotein made by the cells lining blood vessels (endothelial cells) and by megakaryocytes, and stored in endothelial organelles called Weibel-Palade bodies.

According to eClinPath, vWF has two jobs that matter here. First, in fast-flowing small vessels it is the initial "sticky tether" that captures circulating platelets at an injury, bridging the platelet glycoprotein Ib-IX-V receptor and exposed subendothelial collagen. Second, it carries coagulation factor VIII in the circulation and protects it from breakdown. The largest vWF multimers are the most effective at capturing platelets, which is why a selective loss of large multimers causes more severe bleeding.

Cornell's comparative coagulation laboratory classifies canine vWD into three forms defined by the quantity and structure of plasma vWF, and notes that within each breed a single form predominates:

• Type 1: low vWF concentration with normal structure, meaning all multimer sizes are present but in reduced amounts. Today's Veterinary Nurse reports that type 1 accounts for more than 95% of cases. Cornell's affected-breed table lists the Airedale, Akita, Bernese Mountain Dog, Dachshund, Doberman Pinscher, German Shepherd, Golden Retriever, Greyhound, Irish Wolfhound, Manchester Terrier, Schnauzer, Pembroke Welsh Corgi, Poodle, Shetland Sheepdog "and others." Severity is variable, typically mild to moderate. eClinPath points out a striking breed difference: Airedales rarely bleed despite a high prevalence of the trait, whereas Dobermans bleed quite commonly.

• Type 2: low concentration plus abnormal structure, with a relative loss of the high-molecular-weight multimers. Cornell classes it as severe and Merck as moderate to severe. It is described in German Shorthaired and German Wirehaired Pointers. eClinPath notes that bleeding in type 2 does not track the antigen value, because antigen levels do not correlate with multimer composition.

• Type 3: vWF markedly reduced or absent, the most severe form. Cornell lists familial type 3 in the Chesapeake Bay Retriever, Dutch Kooiker, Scottish Terrier and Shetland Sheepdog, and sporadic cases in the Blue Heeler, Border Collie, Bull Terrier, Cocker Spaniel, Labrador Retriever, Pomeranian, mixed-breed dogs and others. Affected dogs typically have vWF:Ag values of 1% or less. VGL and eClinPath describe type 3 as autosomal recessive.

Signs follow severity. Cornell describes spontaneous bleeding from the nose, mouth, and urinary, reproductive or intestinal tracts in severe cases, uncontrollable bleeding after surgery, and excessive bleeding in affected puppies after dewclaw removal or during teething. eClinPath adds that less severely affected dogs may bleed abnormally only after surgery or trauma, and that pinpoint petechial hemorrhages are rarely seen in vWD; if they appear in a predisposed breed, a platelet-count problem such as immune-mediated thrombocytopenia should be considered first.

A routine pre-anesthetic profile does not screen for vWD. eClinPath states that most dogs with vWD have normal platelet counts and normal coagulation profiles (PT, aPTT and fibrinogen), so specific vWF assays are needed; VCA likewise states that coagulation panel results are normal in dogs with vWD. Our guide to PT and aPTT coagulation tests in dogs explains what those panels measure and what they miss. Because vWF carries factor VIII, low vWF can lower factor VIII activity, but eClinPath notes this effect tends to be smaller in dogs than in people, which fits with routine panels usually reading normal. A complete blood count is still useful: VCA uses it to confirm that platelet numbers are normal and that another cause is not driving the bleeding. Acquired problems are a separate pathway. Merck explains that anticoagulant rodenticide poisoning, for example, reduces the liver's production of clotting proteins, which is a different mechanism from an inherited vWF deficiency.

Many clinics have a quick in-house screen, the buccal mucosal bleeding time (BMBT). VCA describes it as a small, precise cut in the upper lip made with a spring-loaded device, timing how long bleeding takes to stop. eClinPath notes it can be useful in the clinic, for example in a Doberman with unknown vWF results before surgery, but that it is neither sensitive nor specific for vWD and does not replace specific vWF assays. Today's Veterinary Nurse adds that the test is operator-dependent, does not predict how much a dog will bleed, and should be used cautiously in dogs suspected of type 2 or type 3 disease because of the risk of prolonged bleeding. VCA notes that relatively large decreases in vWF are needed before the BMBT lengthens significantly. A normal BMBT is therefore not reassurance.

Which breeds have a DNA test - and which do not

Canine vWD DNA tests look for specific known variants in the VWF gene. Cornell's comparative coagulation laboratory lists the described mutations, and VGL offers a separate test for each vWD type, each with its own list of breeds for which the test is appropriate. Cornell's Riney center adds an important caveat: not all causative variants for every predisposed breed are known. The breed lists below are VGL's as accessed on October 25, 2026; laboratory lists change, and other providers' lists differ.

• Type 1, c.7437G>A: a splice-site variant at the end of VWF exon 43. VGL lists the test as appropriate for 25 breeds or breed groups: Australian Cobberdog, Barbet, Bernese Mountain Dog, Brazilian Terrier, Cardigan Welsh Corgi, Coton de Tulear, Doberman Pinscher, Drentsche Patrijshond, Dutch Partridge Dog, Dutch Shepherd, French Water Dog, German Pinscher, Havanese, Irish Red and White Setter, Irish Setter, Kerry Blue Terrier, Kromfohrländer, Manchester Terrier, Papillon, Pembroke Welsh Corgi, Phalène, Poodle and Doodle/Poo crosses, Stabyhoun, Toy Manchester Terrier and West Highland White Terrier.

• Type 2, c.1657T>G (p.Trp553Gly): a missense variant. VGL lists the test for the German Shorthaired Pointer and German Wirehaired Pointer, with autosomal recessive inheritance.

• Type 3, three breed-specific variants: Int16G>A, a splice-site variant in intron 16, in the Dutch (Deutsch) Kooiker; c.255Cdel, a deletion in exon 4, in the Scottish Terrier; and c.735Tdel in the Shetland Sheepdog. VGL lists its type 3 test for those three breeds only, with autosomal recessive inheritance.

Set those lists against Cornell's affected-breed tables and the gap is obvious. Of the breeds Cornell names for type 1, the Airedale, Akita, Dachshund, German Shepherd, Golden Retriever, Greyhound, Irish Wolfhound, Schnauzer and Shetland Sheepdog do not appear on VGL's type 1 list. For type 3, the Chesapeake Bay Retriever is a familial breed in Cornell's table but is not covered by VGL's type 3 test, and sporadic cases occur in breeds no test lists. Other laboratories may cover different breeds; eClinPath, for instance, mentions type 1 genetic testing in Rottweilers and Airedales. Before ordering, check that the specific laboratory lists your dog's breed for the specific test.

The table below brings the two sources together. It is a coverage map, not a statement of prevalence.

The practical rule: a DNA "clear" result is specific to the variant tested and to the breeds the laboratory lists. If a Golden Retriever is tested for c.7437G>A and comes back clear, that only shows the dog does not carry that one variant; it does not rule out von Willebrand disease, because the cause in that breed may be a different or not-yet-identified variant. Mixed-breed dogs are in the same position. Merck notes that vWD occurs in mixed breeds, and mixed-breed dogs are among those most commonly submitted to Cornell's laboratory for vWF:Ag testing. For any dog without a breed-appropriate DNA test, measured vWF:Ag, interpreted by a veterinarian, is the route to an answer.

Reading a vWF antigen result: what the percentage bands mean

The vWF:Ag test is an immunoassay, usually an ELISA, run by a reference laboratory on citrated plasma. VCA notes that vWD is most commonly diagnosed this way. The result is a percentage of a 100% standard made from pooled plasma from healthy dogs, so a result of 40% means the sample contained roughly 40% of the reference concentration.

The bands below are the diagnostic ranges published by Cornell's comparative coagulation laboratory, with the additional reference points eClinPath and Cornell describe. They are one laboratory's reporting conventions; other laboratories may use different units or intervals, so read your dog's result against the interval printed on its report.

The borderline band causes the most confusion. Cornell and eClinPath describe it as an overlap interval: some dogs in it are genetically clear and some are carriers, and a single result cannot tell which. eClinPath lists three ways to clarify the picture:

• A second vWF:Ag test. On retesting, some borderline dogs fall into the normal or abnormal interval, which allows a prediction of genetic status. Timing the retest is a decision for your veterinarian.

• Offspring screening (a test mating). If a borderline dog bred to a high-testing clear mate produces only clear pups, the borderline parent is predicted to be clear; one or more pups in the abnormal range indicates the borderline parent is a carrier.

• A genetic test, where one is listed for the breed. eClinPath describes the genetic test as useful for confirming carrier status in dogs with questionable vWF:Ag results, but says it should not be used to diagnose the disease. Today's Veterinary Nurse likewise calls genetic testing particularly valuable for borderline results. This option only helps if the laboratory's test covers the dog's breed.

Physiology changes the number too. Cornell's laboratory notes that plasma vWF levels fluctuate somewhat from day to day even in healthy dogs, and that the fluctuation is exaggerated during pregnancy or heat and in any dog with a systemic illness, especially liver disease or inflammatory disorders. Cornell describes results as most useful for predicting genetic status when collected at physiologically "quiet" times from healthy dogs that are not pregnant or in heat; Today's Veterinary Nurse adds lactation to that list. Puppies can be sampled from as young as 6 to 8 weeks of age.

Sample quality matters as much as timing. Cornell warns that samples containing clots or hemolysis (broken red cells) are likely to give inaccurate or irreproducible results, and Today's Veterinary Nurse emphasizes atraumatic collection from a peripheral vein. Collection and shipping follow the receiving laboratory's own instructions and are handled by the clinic. For owners, the useful step is simply to tell the veterinarian if the dog is in heat, pregnant, nursing or unwell when a test is planned.

When the DNA result and the factor level disagree

Owners and breeders are often surprised when the two results seem to contradict each other. Two illustrative patterns, both consistent with the published evidence: a Doberman with two copies of the type 1 variant whose antigen result is borderline and who has never had a bleeding problem; and a dog with one copy whose antigen result is low. Neither result is necessarily wrong. They are measuring different things.

Peer-reviewed work in Doberman Pinschers shows why. Brooks and colleagues (2001, American Journal of Veterinary Research) studied 102 client-owned Dobermans and found strong associations between plasma vWF concentration and genotype at two VWF marker loci. The 25 dogs with substantially reduced vWF and multiple bleeding events all shared the associated marker genotypes, and those genotypes were found only in dogs with low plasma vWF, yet some dogs with the same genotypes did not have abnormal bleeding. The study used linked markers rather than today's direct c.7437G>A assay, involved a single breed, and is summarized here from its published abstract.

Crespi and colleagues (2018, Journal of Veterinary Diagnostic Investigation) genotyped the c.7437G>A variant in Doberman Pinschers in the Buenos Aires region of Argentina and reported a mutated-allele frequency of 0.41 in the dogs tested. They describe type 1 inheritance as autosomal dominant with incomplete penetrance. That frequency is a single-region figure, not a global prevalence. Incomplete penetrance simply means that carrying the variant does not guarantee the disease will be expressed, which matches the variable bleeding seen in the Brooks study and eClinPath's observation that bleeding tendency differs between breeds carrying type 1 disease.

The authorities do not describe type 1 inheritance in the same way, and readers will meet all three descriptions:

• UC Davis VGL lists type 1 as autosomal dominant with incomplete penetrance: dogs with one or two copies of the variant may be affected, and not all dogs with two copies show signs.

• Cornell's Riney Canine Health Center describes the type 1 variant as recessively inherited, with dogs at risk if they carry two copies, while noting that carriers with one copy can also have reduced vWF levels.

• Cornell's comparative coagulation laboratory states that in many breeds one abnormal VWF gene appears sufficient to cause abnormal and variable bleeding, a pattern it calls incomplete dominance, while breeds with type 2 and type 3 vWD appear to follow recessive patterns.

This article does not choose between them. What the three descriptions share is practical: one copy of the type 1 variant does not guarantee normal vWF levels, two copies do not guarantee bleeding, and a measured factor level plus the dog's history are needed to judge severity. Cornell's Riney center accordingly says that a dog found at risk on genetic testing may then be tested for the amount of vWF in the blood to characterize severity at that point in time, and that a veterinarian may also recommend measuring vWF in a carrier. If a breeding decision depends on the inheritance model, ask the laboratory or your veterinarian which model their guidance assumes.

The reverse mismatch, an antigen result that looks reassuring in a dog that bleeds, is most relevant to type 2 disease. eClinPath notes that type 2 antigen values do not correlate with multimer composition, and Today's Veterinary Nurse states that the vWF:Ag assay cannot truly identify the type 2 variant, so function must also be assessed. The vWF collagen-binding assay is a functional test: VCA describes it as measuring how well a dog's vWF binds collagen, and Today's Veterinary Nurse calls it a valuable screen for type 2, which also helps distinguish type 1 from type 3 alongside the antigen result, although it is not widely available. A normal-looking antigen value in a pointer breed with bleeding signs is therefore a reason to discuss function testing, not a reason to stop looking.

Which test fits which decision

Matching the test to the decision avoids most of the confusion. The situations below are the ones the retained sources address; how they apply to an individual dog is for the veterinarian to decide.

• Breeding selection. For breeds a laboratory lists, the DNA test is the primary tool; VCA describes genetic testing as typically reserved for screening dogs used in breeding programs. Cornell's laboratory publishes breeding guidance that works with either vWF:Ag ranges or genetic tests: dogs in the normal range are ideal; matings between two test-clear parents are predicted to produce only clear pups; a carrier that does not bleed may be bred to a test-clear mate; carrier-to-carrier matings are undesirable because they are likely to produce severely affected offspring; and no dog that shows abnormal bleeding should be bred. Today's Veterinary Nurse adds that a positive bitch should not be bred because of the risk of hemorrhage during whelping or a cesarean section.

• Planning surgery in an at-risk breed. Here the question is how much factor the dog has now, so measured vWF:Ag is central, even when a DNA result is already known. Today's Veterinary Nurse encourages testing before anesthesia when vWD is suspected and reports a recommendation that all Doberman Pinschers have coagulation testing before anesthesia and surgery; that is a practice recommendation, not a universal standard. Whether blood products or medication are prepared in advance is a clinical decision. Cornell's Riney center notes that management can include avoiding cosmetic surgeries.

• Unexplained bleeding. VCA describes a stepwise workup: a CBC to check platelet numbers, possibly a BMBT, a coagulation panel to rule out other clotting-factor problems, and then specific vWF testing by immunoassay, with a collagen-binding assay when function is in question. Today's Veterinary Nurse stresses that no single test detects all vWD variants and that both concentration and function are needed for a definitive diagnosis. A DNA test is not the tool for working up active bleeding.

• A borderline (50% to 69%) antigen result. The options are a repeat measurement, offspring screening, or a genetic test if one is listed for the breed, as described above.

• A mixed-breed dog or a breed no laboratory lists. Measured vWF:Ag is the applicable test; a DNA result for a variant listed only for other breeds does not answer the question.

What a result does not license: safety boundaries

A test report is information, not a treatment plan. Whatever the DNA or antigen result says, decisions about medications, procedures and blood products belong to the veterinarian who knows the dog. Owners should not start, stop or change medicines on the strength of a test report.

For context only, these are the tools the sources describe veterinarians using:

• Desmopressin acetate (DDAVP). Cornell's laboratory notes it can improve hemostasis in dogs with type 1 vWD by releasing stored vWF. It does not work in type 3, because those dogs lack stored vWF, and Today's Veterinary Nurse describes it as less effective in type 2. eClinPath cautions that not all type 1 dogs respond and that repeated doses lose effect, so it should not be relied on alone to achieve surgical hemostasis.

• Blood products. Cornell's Riney center explains that cryoprecipitate or fresh frozen plasma temporarily raise vWF levels, and both are used for bleeding episodes and before necessary surgery. Today's Veterinary Nurse notes that cryoprecipitate delivers vWF and factor VIII in a smaller volume and was found more effective in a clinical study, while eClinPath advises that whole blood alone does not supply enough vWF to stop hemorrhage. Product choice and timing are clinical decisions.

• Medications that affect clotting. Cornell's Riney center says veterinarians may need to avoid NSAIDs, which interfere with platelet function, in dogs with vWD. Cornell's laboratory lists aspirin, sulfa-type antibiotics and heparin among drugs affected dogs should not be given, and Today's Veterinary Nurse advises that such drugs be used with caution or avoided. If your dog takes any of these, raise it with your veterinarian rather than stopping a prescription yourself.

Cornell's Riney center also notes that there is no cure for vWD, but with proper management many dogs live a normal lifespan. Severity, and therefore the outlook, depends on the type.

Finally, a species boundary. Everything above is canine evidence. VCA notes that cats are rarely affected, and eClinPath records type 3 vWD reports in cats, but the breed lists, variants and interpretation bands in this article do not transfer to cats or to people. For feline transfusion questions, see our guide to cat blood typing.

Sources

• Cornell University College of Veterinary Medicine - Riney Canine Health Center: Von Willebrand Disease (vWD) in Dogs

• Cornell University Animal Health Diagnostic Center - Comparative Coagulation Laboratory: Canine von Willebrand Disease (Clinical Topics)

• Cornell University College of Veterinary Medicine - eClinPath: von Willebrand disease

• VCA Animal Hospitals: Von Willebrand's Disease Testing

• UC Davis Veterinary Genetics Laboratory: Von Willebrand Disease I (vWD Type 1) - DNA test page

• UC Davis Veterinary Genetics Laboratory: Von Willebrand Disease II (vWD Type 2) - DNA test page

• UC Davis Veterinary Genetics Laboratory: Von Willebrand Disease III (vWD Type 3) - DNA test page

• American Journal of Veterinary Research (PubMed): von Willebrand disease phenotype and von Willebrand factor marker genotype in Doberman Pinschers

• Journal of Veterinary Diagnostic Investigation (PubMed): von Willebrand disease type 1 in Doberman Pinscher dogs: genotyping and prevalence of the mutation in the Buenos Aires region, Argentina

• Merck Veterinary Manual: Bleeding Disorders of Dogs (Dog Owners)

• Today's Veterinary Nurse: A Veterinary Nurse's Guide to von Willebrand Disease