This is a response to this prompt on the Canine Debate Forum: “What in your opinion is the most genetically healthy breed? The least healthy? What major health concerns does your breed of choice suffer from and how would you as a breeder go about fixing them?”
I got carried away and ended up writing a massive overview of why it’s difficult to classify which breeds are the most unhealthy and the reason why in the grand scheme of things it doesn’t really matter anyway due to the way genetics work. Since it got too long for me to post in good conscience on FB, I’m turning it into a blog post instead. Apparently I get too excited when writing on this topic! :P
The following overly long spiel is me trying to speak from an objective (scientific) rather than a subjective (anecdotal) standpoint like most of the responses on this thread, so it will get long and complicated and thick with genetics terms. :P
I think it’s misleading to label any breed “most/least genetically healthy.” A more accurate term would be “phenotypically diseased.” The former label wrongly suggests that some breeds simply have fewer “bad” genes than others, and so they are naturally more healthy. It also leads to the overly simplistic, scientifically inaccurate conclusion that in order to eliminate congenital disease from a breed, all you have to do is create a test for each disease and then breed away from disease-causing genes. Genetic health is much more complex than this, and not so easily measured.
Even when altering the label to “phenotypically diseased,” it’s a virtually impossible question to answer because it’s so nuanced. What makes a breed more or less healthy than others? Is it the overall number of known disorders in the breed? Is it the average prevalence of these disorders? Do we weigh different diseases differently since some are more lethal and untreatable than others (i.e. DCM is a much graver health concern than color dilution alopecia)? Does average lifespan and litter size figure into this equation somewhere? Are polygenic and monogenic disorders rated differently since the latter is easier to test for and avoid? Are the diseases with available genetic tests weighted differently than those without them since breeders can more accurately identify carriers? Should immune mediated disorders and diseases which become “fixed” be weighted more heavily since they’re impossible to “breed out” without crossbreeding? How do you rate diseases with high breed-specific prevalence but overall low heritability (eg hip dysplasia)?
If you want a truly objective answer to that question, you would have to create a complex rating system taking all these factors into account AND have access to data that doesn’t really exist. Most of what we know about breed health is due to self-reports by breeders to breed clubs and registries. These tend to have tiny sample sizes and sometimes people will deliberately falsify or exclude data to make themselves look better. Many breeders don’t complete these surveys at all so it gives only an incomplete picture of a breed as a whole. While these data certainly are better than nothing, their accuracy should not be taken for granted.
On the other hand, empirical studies on breed-wide prevalence of health problems are few, and a number also suffer from the problem of having a very limited sample size. Some studies have been indirectly biased by receiving funding from breed clubs which encourages research toward a specific goal (most often developing a genetic test for a given disease) rather than keeping an open mind about the best methods to tackle a breed health problem (such as avoiding certain phenotypes altogether), and this affects both the direction of the study and what conclusions the researchers draw from their data. The UC Davis Standard Poodle genetic diversity test is one notable example. The test created a means of measuring relatively tiny differences in genetic diversity between individuals, proposing this sort of testing as the solution to widespread disease, rather than going with the simplest solution that makes the most scientific sense: creating a well-designed crossbreeding program to restore diversity to the breed before it’s too late. In general, genetic diversity tests are an attempt to cater to the mainstream dog breeding community’s desire to “breed out” health problems while maintaining blood “purity,” even as this flies in the face of the ideal solutions as determined by applying basic population genetics. There’s profit to be made in placating the desires of a group by providing them the solutions they want, but not so much by proposing ideas which may be objectively more effective but go against the status quo.
Even more worrying is when researchers falsify data to support their presuppositions about breed health. One study of genetic diversity in Nova Scotia Duck Tolling Retrievers actually contained intentionally misleading, incomplete data. One of the researchers involved in the study, Clare Wade, is a Toller breeder and consequently has a strong (but unreported) conflict of interest, desiring to make her breed’s health appear better than it really was. Her study was almost certainly a response to another recently published study that reported levels of inbreeding to be dangerously high in Tollers and found that the gene pool was too shallow for long term survival of the breed. Clare Wade’s study claimed that the opposite was true and that Tollers were remarkably genetically diverse and healthy, basing her conclusions on incomplete, incorrectly compiled pedigree data. It wasn’t long after the study’s publication that other professionals publicly criticized her work as intentionally misleading. (Summary of this case of academic fraud here: http://www.border-wars.com/2011/10/academic-fraud-in-toller-research.html).
All this goes to say: if even professional scientists are willing to cast an unrealistically optimistic light on empirical data (or even falsify data entirely), it’s clear that data regarding breed health must be taken with a grain of salt whether it’s self-reported or the result of a research study. The overwhelming tendency is to under-report health problems, so it’s typically safe to assume that any breed’s health is less optimistic than may be presented or reported.
You also have to consider the fact that the more common breeds will be better studied than the less common breeds, which can easily give the misleading impression that one is very sick and the other is very healthy when in reality, the difference in reported health problems is merely a matter of insufficient data. According to one study (http://www.instituteofcaninebiology.org/blog/what-does-health-tested-really-mean) GSDs were found to have 77 genetic disorders; by contrast Japanese Spitzes are only reported to have 1. Does this mean that Japanese Spitzes are overall healthier than GSDs, with fewer genetic disorders? Possibly, but it’s also quite likely that a very popular breed like the GSD has a lot more data available for analysis and a lot more public interest in genetic health research which leads to a greater number of genetic diseases being identified and reported. There may well be dozens of unreported (or currently unidentified or latent) genetic disorders in Japanese Spitzes, but there’s no way to tell without reliable data.
Because of these factors, when comparing breed health problems, rather than lining breeds up along a spectrum of most to least healthy, I find it more helpful to split breed health problems into general categories like: “Will suddenly drop dead before age 10” (like the Doberman with high rates of DCM); “Will die young but with a longer, more drawn out death” (Flat-coated Retrievers with cancer, Cavaliers with congestive heart failure); “Will probably live a long time but will suffer from at least one health issue that has a moderate to strong negative effect on quality of life” (Italian Greyhounds, which often live to 15+ years but suffer from chronic allergies, epilepsy, hypothyroidism, and various immune disorders while their teeth rot out due to genetic factors); “Will have an overall poor quality of life directly due to their physiology” (many extreme brachycephalic breeds, and also Shar Peis with familial Shar Pei fever); “May not suffer from obvious serious health issues but won’t live beyond age 6” (Irish Wolfhounds, Great Danes); “Will likely suffer a chronic health problem that can be somewhat managed with medication” (Standard Poodles and Addison’s disease); “May have serious behavioral problems due to genetics that may or may not be effectively managed due to the owner’s circumstances and experience” (Rage Syndrome in English Springer Spaniels, OCD in Bull Terriers, extreme noise sensitivity in Border Collies); “High chance of being born with or having a propensity for developing a disability that will decrease quality of life unless effectively managed in the right environment” (Dalmatians and deafness, Australian Cattle Dogs and congenital blindness and deafness, Dachshunds and IVDD). I’m sure I could come up with more if I kept thinking about it.
Each of these categories will be more or less acceptable to different people, so classifying some as objectively worse than others is pointless. Some people may prefer having an especially short-lived dog over a longer-lived dog with a disability or chronic health issues, and vice versa. Every breed will fit into one or more of these categories which highlights the nature of a disease rather than focusing on poorly known prevalence estimates. This is important because while disease prevalence can and will fluctuate over time, few diseases will ever be truly eliminated from a breed via selective breeding. No breed should be considered inherently healthier as a rule, but at risk for varying diseases.
The fact is that every organism has numerous deleterious alleles, regardless of how phenotypically healthy they are. In humans, the average is 0.29 recessive lethal alleles and about 6 disease-causing alleles. Because of this, breed health is not so much related to “genetic health” (i.e. this breed has few or no disease-causing genes) so much as it’s related to how well a breed’s genetic diversity has been managed. Poor management can increase the risk for health problems at any time in any breed, as well as specific lines.
If you’ve got a breed with very low genetic diversity due to things like small founder size, popular sire syndrome, numerous bottlenecks etc. you will inevitably run into salient health problems more quickly. The least genetically diverse breed in existence, the Norwegian Lundehund, is fixed for a severe, fatal immune-mediated bowel disease that affects 40% of the population and the entire population carries the genes for the disease. Less extreme but still concerning examples are Italian Greyhounds, Bernese Mountain Dogs, Standard Poodles, and Nova Scotia Duck Tolling Retrievers. As genetic diversity decreases, the frequency and expression of diseases will increase, and dogs will also be affected in more subtle ways like slightly reduced lifespan, reduced fertility, smaller litters with less ability to thrive, slightly reduced immune function etc.
These negative effects are often subtle and not very salient when only judging health on an individual basis or via anecdotal horizontal analysis (casually comparing all dogs in the same generation) rather than empirical vertical analysis (systematically comparing all dogs over many generations). Because of this, a breed can slip into poorer health and lowered reproductive fitness gradually with no breeders noticing this steady loss of vitality and increase of disease until genetic diversity is too severely depleted to be restored within a closed studbook. The breed gets stuck in a bad situation that can only get worse unless drastic measures are taken in time. This is the main way most dog breeds have become unhealthy, despite the best efforts of responsible breeders and passionate fanciers.
Because genetic diversity always decreases in a closed system over time, it’s not uncommon to see newer breeds that appear to be very healthy currently. The Alaskan Klee Kai is generally regarded as having few health problems, and has only had a closed studbook for a couple decades. However, on a genetic level, UC Davis’ genetic research on the breed has only found 9 DLA haplotypes in the entire breed, which is less than half the known haplotypes of breeds with very low genetic diversity and a high amount of disease as a result (namely Standard Poodles and Italian Greyhounds). While the distribution of DLA haplotypes in Klee Kais is currently is not as uneven as in some other breeds like the IG, the inevitable loss of alleles over time due to selection and genetic drift essentially dooms them to become an unhealthy breed given enough time, unless they periodically accept new founders to reverse gene loss and impose strict guidelines limiting the number of litters per sire to balance allele frequencies.
The longer a breed remains in a closed system, the lower genetic diversity will inevitably become (https://drive.google.com/open?id=0B3ry0Kv4xqooRm5YX2NnOFUySTA2c3czVHhReG9QdXdFczlJ), and the more prone to health issues.
Just as low genetic diversity produces more health problems, higher genetic diversity avoids them. Therefore, as a general rule, any breed with greater, more evenly distributed genetic diversity will suffer from fewer expressed genetic diseases at a lower prevalence.
(As an aside, it’s important to note that genetic diversity must be fairly evenly distributed throughout a population in order to be beneficial, as disease expression boils down to allele frequencies. When a large number of alleles have a fairly similar level of distribution in a population, there’s a lower probability of two alleles “doubling up” in a homozygous state or of genes developing strong linkage disequilibrium which can be dangerous for additive disorders. For instance, IGs have a decent amount of genetic diversity breed-wide overall, but it is very poorly distributed due to heavy use of popular sires and the genetic isolation of IG populations in different countries due to differences in breed standards that discourage interbreeding. The majority of DLA haplotypes in the breed have a prevalence of <5%, and many are very rare with a prevalence of <1%. Of their 20 known DLA class 1 haplotypes, only 4 have a prevalence >19%. Their immune systems are compromised and very dysfunctional breed-wide as a result.)
Breeds with greater phenotypic diversity will typically tend toward greater genetic diversity, as long as these varying phenotypes are not segregated into isolated subpopulations (eg working vs show lines which don’t interbreed). Dogs bred for work tend to be slightly more genetically diverse than those bred solely for conformation (https://drive.google.com/open?id=0B3ry0Kv4xqooTlc2cVBRQWFKRDQ), likely because behavioral traits generally have a lower level of heritability than physical traits so this calls for different breeding strategies and generally allows for greater phenotypic variation.
Landraces will have fairly high levels of genetic diversity overall, as their phenotypes have a relatively wide range of geographic variations. Even with this variety, they still maintain a measurable amount of genetic homogeneity as a group which indicates common ancestry and warrants a shared classification into a spectrum of types within one landrace rather than distinct, separate breeds like those found in Western registries.
Regarding purebreds specifically, breeds like Salukis where COO dogs are still regularly imported and bred with registered dogs will have greater diversity as a breed as well, and Salukis have the most DLA haplotypes of any known breed as a result. Of course, breeders that reject COO dogs due to concerns of breed “purity” won’t benefit from this greater diversity. Without taking the genetic contributions of COO dogs into account, the gene pool of “pure” Western Salukis is rather small and inbred, and they are predisposed to hemangiosarcoma, DCM, and thyroid problems.
Ultimately, the greatest level of diversity in any type of dog is that of the humble, indigenous street dog. Wild animal populations tend toward heterozygosity under ideal conditions, and the only selecting factor for pariah dogs is their ability to survive and reproduce. Dogs that get sick easily from faulty immune systems will die. Dogs born with a serious hereditary illness will be naturally culled. Dogs that are born with deformities or any physical trait that reduces reproductive fitness will not pass their genes onto the next generation. Dogs that are infertile, bad mothers, or produce small litters, all of which are consequences of high levels of inbreeding, will have few offspring if any at all. On a genetic level, allele frequencies will be much more balanced as the sexes contribute fairly equally and popular sire syndrome can’t skew the distribution of genes so wildly as in domestic purebreds. The only significant factors preventing any two dogs from mating are geography and reproductive fitness; there are no distinctions between breeds resulting in genetically isolated populations that are arbitrarily distinguished like in purebreds.
However, there is a notable downside to this high level of genetic diversity: the lack of specific selection that many people desire in a dog, whether a working dog or a companion, which is necessary for a predictable appearance and temperament. With stronger, tighter artificial selection and greater predictability comes lower genetic diversity. This is inevitable due to the direct, inverse relationship between homozygosity and heterozygosity.
This is the deepest, most significant conflict of interest that’s at the heart of all dog breeding. Both breed health and breed predictability are highly desirable traits that breeders strive for. Achieving success in both goals is a complex task of striking a balance between breeding for enough homozygosity to produce a dog with a specific appearance and function while maintaining enough heterozygosity to prevent high levels of crippling disease.
There are means of achieving this, like choosing assortative mating over linebreeding; breeding dogs with faults that don’t impede health or function even though they may be considered “undesirable;” and the introduction of new blood via outcrossing or crossbreeding periodically.
However, some current breeding practices are simply unsustainable due to significantly favoring homogeneity over health: linebreeding heavily to develop a distinct look for your lines or attempt to “fix” a trait in as few generations as possible; maintaining a closed studbook merely over concerns of blood purity which have no scientific founding; utilizing popular sires to the point where a dog may father hundreds of puppies which will widely distribute both his desirable traits and any latent diseases throughout the breed; narrowing the allowed phenotypes of a breed due to politics rather than science, function, or basic logic (eg disqualifying colors and markings which are already present in a breed’s gene pool, and strict selection for minor aesthetic features like the presence of a ridge or a perfectly curled tail). Wide use of these methods has created the health problems found in most dog breeds today, and in breeds and lines where these techniques have been used especially often, it’s fair to expect their health to eventually deteriorate whether quickly or over the course of many generations.
The bottom line is that it’s impossible to single out specific breeds as objectively being the “most unhealthy” due to the complexity of categorizing and measuring the impact of genetic diseases with the current lack of quality data. Furthermore, no breed should truly be considered “genetically healthy,” though they may perhaps be phenotypically healthy; this incorrectly suggests that some breeds have few or no disease-causing alleles, and that achieving genetic health is a simple matter of identifying every faulty gene and eliminating it from the gene pool. The reality is that every breed has health issues, and every breed is capable of becoming highly diseased when their genetic diversity is poorly managed. Every single dog has numerous deleterious genes and must be considered a carrier for at least one disease even if they’re phenotypically healthy and are tested clear for the minuscule fraction of diseases with genetic tests. When genetic diversity is managed properly and an ideal balance is struck between selection for homozygosity to maintain type and function and allowing for enough phenotypic variation and providing a steady flow of new blood to maintain heterozygosity and reduce risk of congenital disease, THAT’S where you’ll find the healthiest breeds. Anything else is just splitting hairs.