Usefulness of Breed Information

Historically, dog breeds were groups of dogs with similar behavioral characteristics, such as hunting or herding ability, and members of a breed did not necessarily resemble each other physically.¹ Since the turn of the 20th century, however, there has been a shift from purpose-based breeding to breeding on the basis of appearance.^1,2

For certain dogs breeds, breeding on the basis of appearance has inadvertently selected for specific genetic diseases,³ in essence adding these diseases to the phenotype of the breed.^4,5 Cocker Spaniels and Chihuahuas, for example, have a tendency to have hydrocephalus,³ German Shepherd Dogs are more likely than dogs of many other breeds to have hip dysplasia,⁶ and Bulldogs commonly have reproductive disorders, with > 80% of all Bulldog puppies in 1 study⁷ delivered by caesarean section. Thus, for many purebred dogs, information on breed may provide insight into the predisposition for certain diseases.

On the other hand, breeding primarily on the basis of appearance has, at least in some cases, altered the classic breed-associated behaviors. A Swedish study,¹ for example, found that some behavioral characteristics classically associated with individual breeds have not survived under current breeding practices. In that study, an individual dog's behavioral characteristics correlated more closely with the use of its relatives in the preceding few generations (eg, show dog, working dog, or companion dog) than with the traits classically associated with the dog's breed. Similarly, a genetics study⁴ involving dogs with specific behavioral characteristics found only a loose correlation between genes known to influence the nervous system or behavior and certain types of behavior, such as boldness, pointing, and herding. Thus, breed identification gives limited insight into the behavioral characteristics of individual animals,¹ and behavioral characteristics can differ considerably within a breed.⁸

Breed Identification and Misidentification

Approximately 44% of dogs in the United States are of mixed breeding,⁹ and for many of these mixed-breed dogs, lineage is unknown. For mixed-breed dogs of unknown lineage, breed has often been assigned on the basis of physical appearance.

However, physical appearance does not appear to be a good indicator of breed when dealing with dogs of mixed breeding. For example, in a classic study by Scott and Fuller³ that was designed to differentiate learned behavior from genetically based behavior, the behavioral characteristics of several generations of Cocker Spaniel–Basenji crossbred dogs were compared with the behavioral characteristics of purebred dogs of each breed. The overall conclusion was that behavior is an extremely complex process that involves environmental factors such as socialization and heritable factors and that breed of the parents is not a good predictor of the behavioral characteristics of mixed-breed puppies. Two ancillary findings, however, were that the second (F2) generation of the Cocker Spaniel–Basenji crosses took a “great variety of form and color” and that none of the 72 F2-generation puppies closely resembled either parental breed (Figure 1). Without knowledge of their lineage, these F2-generation puppies would be as likely to be identified as predominately Labrador Retriever, Beagle, or Springer Spaniel mixes as they would be to be identified as Cocker Spaniel or Basenji mixes.

View Full Size

Photographs of 6 second (F2)-generation Cocker Spaniel–Basenji crossbred dogs. Notice the wide variations in color and conformation and the substantial differences in appearance from the breeds of the parents. (Reprinted with permission from Scott JP, Fuller JL. Genetics and the social behavior of the dog. Chicago: University of Chicago Press, 1965. ©1965 by the University of Chicago.)

Citation: Journal of the American Veterinary Medical Association 241, 9; 10.2460/javma.241.9.1163

• Download Figure
• Download figure as PowerPoint slide

More recently, Voith et al¹⁰ compared, for dogs from multiple shelter locations, results of breed identification made on the basis of visual inspection alone with results of DNA analysis of breed. Although the number of dogs was small, the major breed determined on the basis of visual inspection matched the predominant breed identified by means of DNA analysis for only 25% of the dogs. This suggests that there is a high potential that results of visual identification of breed for shelter dogs of unknown lineage will differ from results of DNA analysis.

Importantly, misidentification of a dog's breed can lead to a variety of negative consequences. Some of these are a direct result of the owner's expectations related to the dog's perceived breed; others are a result of breed-based policies adopted by certain entities, such as insurance companies, legislative bodies, and even human adoption agencies.^11–14 For example, an owner who adopts, from an animal shelter, a puppy that has been identified by shelter workers as a small-breed mix will naturally expect the dog to be relatively small when it reaches adulthood. If, instead, the dog grows to be a medium- or large-breed dog, the owner may elect to relinquish the dog as a direct result of this breed misidentification.

Recently, some insurance companies have increased homeowner insurance rates or denied coverage altogether for owners of dogs of certain breeds alleged to be particularly dangerous and mixes of such breeds.^15,16 Although some states have legislation pending that would disallow such practices,¹⁷ concerns regarding liability insurance coverage could adversely affect willingness to own dogs misidentified as belonging to these breeds.

Similarly, some local legislatures have banned ownership of various dog breeds, some human adoption agencies and foster home agencies will not place children in homes that own certain breeds of dogs,¹⁴ and some airlines have started to refuse to transport dogs of certain breeds.^18,19 While all of these actions are based on the mistaken notion that dogs of these breeds are inherently dangerous or pose a greater threat to people than dogs of other breeds, they also all point out the negative consequences that can occur for dogs that are mistakenly identified as being related to these breeds.²⁰

An Alternative to Breed Identification

It is possible to describe dogs without assigning a breed. In fact, in previous times, dogs were characterized not so much by breed as by their personality or function. During the time of the Roman Empire, for instance, dogs were simply described as hunting dogs, guard dogs, sheep dogs, and lap dogs.¹ Even as recently as the 1800s, any dog used for tracking was considered a bloodhound.²

Throughout the years, a variety of terms—mutt, cur, mongrel, and potlicker, for example—have been used to describe dogs of unknown parentage.²¹ More recently, the term American shelter dog has been suggested for use in shelter environments,^a and the American Kennel Club has adopted the term all-American dog to refer to mixed-breed dogs that compete in obedience, agility, and rally competitions.²² Currently, the most consistently and most widely used non–breed-based term for dogs of unknown parentage is mixed-breed dog.

To avoid problems associated with breed misidentification, we recommend that veterinary practices adopt a single non–breed-based term to describe all dogs of unknown parentage, rather than attempting to identify the predominant breed on the basis of visual inspection. For new clients and for new dogs acquired by existing clients, we recommend including a short statement on the new client or new patient sheet expressing the hospital's position. A statement such as the following could be adopted:

Because new scientific evidence has called into question the accuracy of visual breed identification of dogs, our hospital has adopted a policy to not identify canine patients by predominant breed unless the dog is purebred, the predominant breed of the dog's parents is known, or the dog's lineage has been established through the use of DNA analysis.

Over the past 6 months, 2 of the authors have successfully used various versions of this statement in their own practice. We have also used pictures of the F2 generation of the Cocker Spaniel–Basenji crossbred dogs (Figure 1) to explain to owners the reasoning behind the hospital's policy.

Adoption of this policy places the onus of breed identification on the owner, rather than the veterinarian. Some owners may have actual knowledge of their dogs' lineage or may choose to identify their dogs on the basis of perceived predominant breed, and in these instances, the breed reported by the owner should be recorded in the dog's medical record, along with any documentation regarding the dog's lineage. However, efforts should be made to ensure that the owner has read and understood the statement, particularly in instances when owners report their dogs to be of unusual breeds or unusual breed combinations, and to ensure that owners are not simply reporting breed identifications provided by an animal shelter or other source.

For mixed-breed dogs for which the predominant breed of the parents is known, we recommend including this information in the dogs' records, as this information may be clinically important. For example, there is a higher incidence of the multidrug-resistant gene in dogs that are mixes of certain breeds, such as Collies.²³ That said, we strongly recommend the consistent use of a single non–breed-based term to identify all dogs of unknown lineage, regardless of appearance and regardless of any previous identifications that might have been made. If, for example, a non–breed-based term were used only for dogs that might be mixes of breeds thought of as dangerous (as defined by various breed-specific bans), the veterinarian could potentially be accused of intentional misrepresentation.

In addition, because most practice management software systems have the capability to add a picture of the patient to the electronic medical record, a picture of the patient should be used for identification purposes, rather than an assigned breed. Some cities require picture identification for dog-licensing purposes,²⁴ and a picture along with a non–breed-based descriptor is much more descriptive than the traditional visually based predominant breed identification.

Finally, it is ill advised to alter a dog's medical record, including changing the breed identification, without documenting the change and explaining why the change was made.²⁵ Therefore, we advise not applying the non–breed-based nomenclature policy retroactively to any existing patients, unless a DNA analysis has been done to substantiate the change in breed identification and a copy of the test results can be included in the patient's medical record.

Use of DNA Analysis for Breed Identification

Use of DNA analysis for identification purposes has been around for many years,²⁶ and the courts have used DNA analysis for a variety of purposes, including verifying paternity, identifying murder suspects, confirming the potential source of an HIV infection, and even identifying the source of black market caviar.²⁷ Canine DNA analysis, unlike human DNA analysis, is still in its infancy, although results of canine DNA analysis are slowly becoming admissible in court.^26,28–31 However, because of the more limited genetic diversity of dogs, compared with humans, canine DNA analysis lacks the individual identification accuracy associated with human DNA analysis.³²

It has been suggested that DNA analysis correctly identifies the breed of between 86% and 99% of pure-bred dogs.³² However, between 2007 and 2012, the stated mean accuracy of DNA analysis of first-generation crossbred dogs of known parentage increased from 84%¹⁰ to 90%,³³ and breed identification accuracy of DNA analysis is expected to continue to improve.

In those instances when owners ask their veterinarians to identify the predominant breed of their dogs, we recommend that the veterinarians state that they cannot be certain of breed on the basis of visual inspection and offer to submit samples for DNA analysis instead of offering a guess. If a guess as to breed must be offered, the veterinarian should preface it as such and should not record the guess in the medical record.

Conclusion

Given the hundreds of recognized dog breeds, it is unreasonable to think that veterinarians can visually identify all of these breeds or the parentage of mixes of these breeds. Therefore, we believe that it is time for a paradigm shift in the veterinary profession's approach to canine breed identification. Specifically, we believe that veterinarians and animal shelters should adopt a policy to avoid visual breed identification of any dog of unknown lineage and should train their staff on the rationale for this policy, including the drawbacks of visual breed identification and misidentification. Owners should be made aware of the hospital's policy against visual breed identification before any guesses are made regarding the possible breed of any dog of unknown lineage, and a single non–breed-based term should be used in the medical records for all dogs of unknown lineage. Note, however, that this does not preclude including breed information in the medical records of dogs when the owner has actual knowledge of the dog's lineage or when breed identification has been determined by means of DNA analysis. Veterinarians have for years attempted to visually identify the breed of dogs of unknown lineage; it is time to rethink this and consider the ramifications that doing so may have.