Several toy breed dogs are predisposed to syringomyelia (SM), a spinal cord disorder, characterised by fluid-filled cavitation. SM is a complex trait with a moderately high heritability. Selective breeding against SM is confounded by its complex inheritance, its late onset nature and high prevalence in some breeds. This study investigated the early outcome of existing SM breeding guidelines. Six hundred and forty-three dogs, 550 Cavalier King Charles spaniels (CKCS) and 93 Griffon Bruxellois (GB), were identified as having either one (454 dogs) or both parents (189 dogs) with MRI-determined SM status. Offspring without SM were more common when the parents were both clear of SM (SM-free; CKCS 70 per cent, GB 73 per cent). Conversely, offspring with SM were more likely when both parents had SM (SM-affected; CKCS 92 per cent, GB 100 per cent). A mating of one SM-free parent with an SM-affected parent was risky for SM affectedness with 77 per cent of CKCS and 46 per cent of GB offspring being SM-affected. It is recommended that all breeding dogs from breeds susceptible to SM be MRI screened; that the SM status at five years old is established; and all results submitted to a central database that can be used by dog breeders to better enable mate selection based on estimated breeding values.
Statistics from Altmetric.com
SYRINGOMYELIA (SM) is characterised by fluid-filled cavities within the spinal cord and occurs secondary to obstruction of cerebrospinal fluid especially if that obstruction is at the foramen magnum. The most common predisposing cause in the dog is Chiari-like malformation (CM) (Rusbridge and others 2006), a condition characterised by disparity in size between the brain (too big) and the skull (too small) such that the cerebellum and brainstem are herniated into or via the foramen magnum (Cross and others 2009). Several toy breed dogs are predisposed to CM and SM. Studies into the inheritance of SM associated with CM in Cavalier King Charles spaniels (CKCS) have shown it to be a complex trait with a moderately high heritability (h2=0.37±0.15 se) (Lewis and others 2010). It has a varying age of onset, there is 46 per cent prevalence in asymptomatic breeding CKCS but prevalence may be as high as 70 per cent in dogs over six years of age (Parker and others 2011). MRI, which is financially prohibitive, is currently the only means of confirming SM and its potentially late onset nature means that the optimum age to MRI screen for breeding purposes is unclear. In addition, the high prevalence means that it can be difficult for breeders to find suitable SM-free dogs. There is also a concern that breeding restrictions may result in further genetic bottle necking particularly as the CKCS is predisposed to many other inherited disorders (for review, see Gough and Thomas 2010). At the request of the CKCS breeding fraternity, interim breeding guidelines for SM were formalised at a international Cavalier club round table in 2006 (Cappello and Rusbridge 2007). Between September 2004 and 2006, similar guidelines had been used by a small number of CKCS breeders particularly in the Netherlands. The 2006 breeding guidelines had the aim of reducing early onset and/or clinical SM while attempting to maintain genetic diversity; they were implemented as a ‘common sense’ approach but without there being any evidence of the appropriateness or effectiveness. Dogs were graded according to MRI status and age as follows: A, an asymptomatic dog aged 2.5 years or older where SM is absent or with a central canal dilatation (CCD) with a transverse diameter less than 2 mm; C, an asymptomatic dog aged younger than 2.5 years without SM or CCD; D, an asymptomatic dog aged 2.5 years or older with SM which has a transverse diameter equal to or greater than 2 mm; E, an asymptomatic dog aged younger than 2.5 years with SM or CCD; and F, a symptomatic dog of any age with SM. For historical reasons, there is no B grade. The Round Table Working Group suggested that dogs be bred according to the guidelines in Table 1. These breeding guidelines have also been used by many breeders of Griffon Bruxellois (GB) although it is yet to be determined if the pathogenesis of SM in the GB is the same as in the CKCS.
Several groups of breeders and pet owners of CKCS and GB from the UK, Netherlands, Finland, Canada, USA, South Africa and Australia have participated in the research into the genetic basis of SM. Phenotypic and pedigree information for both breeds have been collated as part of a DNA collection programme used for genomewide linkage studies carried out in collaboration with the University of Montreal.
Lewis and others (2010) used a mixed linear REML model to investigate the mode of inheritance for SM associated with CM and identified two families for which the parents and offspring had SM status confirmed between three and four years of age. The criterion by which a dog was designated an SM-affected dog was the presence of SM or CCD that had a transverse diameter of 2 mm or more. This 2 mm ‘cut-off’ is not age dependent, that is, a dog deemed SM-free at three years of age will not necessarily be SM-free at five or six years of age. CM is defined as overcrowding of the caudal cranial fossa so that the cerebellum does not have a rounded shape, that is, there is indentation by the supraoccipital bone and/or the cerebellar vermis is impacted into or herniated via the foramen magnum.
This study seeks to investigate if the 2006 breeding guidelines are appropriate, that is, achieving a goal of reducing early onset SM and also aims to identify factors associated with increased or decreased risk of SM, which might be useful for future breeding guidelines to be issued in conjunction with a proposed British Veterinary Association/Kennel Club CM and SM MRI screening scheme. Although a computerised mate selection programme (Lewis and others 2010) is shortly anticipated for CKCS, this is not yet available to other breeds with a tendency for SM and associated CM. Until such programmes are developed and their accuracy verified, some sort of breeding guidance is useful.
Materials and methods
The CKCS CM and SM Microsoft access database constructed for use in the genome studies currently holds over 12,200 dogs linked by their pedigree. Of these, 1282 dogs have full MRI and clinical phenotypical data with stored DNA extracted from blood, saliva or toenail clippings. The GB CM and SM Microsoft Access database currently holds 1394 dogs linked by their pedigree with 244 dogs with full MRI and clinical phenotypical data and mostly with stored DNA. All the MR images have been reviewed by one of the authors (CR) or in consultation with her to ensure consistency and all dogs were assigned an SM grade according to the current breeding guidelines (Table 1). In addition, to estimate the influence of late onset SM, an *A grade was assigned to dogs that had an MRI scan aged five years of age or older and were without SM or with a CCD with a transverse diameter less than 2 mm. Within the CKCS group, 92 dogs had a second MRI scan in order to ascertain any change in SM status or progression of the condition. For the analysis, results of the most recent MRI scan were used. Presence or absence of CM was noted but not analysed in the study because CM is not included in the 2006 breeding guidelines as this trait is almost ubiquitous in the CKCS breed. The offspring (outcome) of a mated pair was considered as an independent event and no account was taken of familial relationships within the cohort.
All dogs included in the study were reported as asymptomatic for SM when presented for MRI screening to ascertain their or their descendants' suitability for breeding purposes. According to the 2006 breeding guidelines, clinically affected (F grade) dogs should not be used for breeding and therefore are excluded from this study by default. Owners of dogs submitted for MRI screening were required to submit limited clinical information by answering questions regarding the common clinical signs of SM, which might be observed in their dog (for the specific questions, see www.veterinary-neurologist.co.uk/mriform/form.html). All dogs were permanently identified with a microchip or ear tattoo and the appropriate KC registration certificates were obtained giving details of dam and sire.
Using the above database, CKCS and GB dogs with MRI-determined SM status that also had one or more parents with MRI-determined SM status were identified. Dogs with inadequate records or with MRI that was of insufficient diagnostic quality were excluded. The decision of which dogs to use in the breeding programme was entirely the choice of the breeders. Thus, although the 2006 breeding guidelines (Table 1) were available, they were not necessarily followed.
The following variables were recorded: SM affectedness in offspring in all possible parental crosses, that is, *Ax*A, *AxA, *AxC, *AxD, *AxE, AxA, AxC, AxD, AxE, CxC, CxD, CxE, DxD, DxE and ExE and crosses that had one parent with unknown (U) SM status, that is, *AxU, AxU, CxU, DxU and ExU; and age at which the dog was MRI scanned.
A total cohort of 643 dogs comprising 550 CKCS (358 females and 192 males) and 93 GB (54 females and 39 males) were identified within the databases having one (392 CKCS and 62 GB) or both parents (158 CKCS and 31GB) with MRI-confirmed SM status. The breakdown of the 643 SM grades, sex and parental status is demonstrated in Table 2. In the rescanned CKCS group (92 dogs), 12 of 26 (43 per cent) previously grade A dogs were now SM-affected. The mean age at repeat MRI was 3.4 years (range 2.5 to 4.3 years) and the mean interval between MRI scans was 2.2 years (range 0.9 to 4.1 years). In addition, 20 of 66 (30 per cent) previously grade C dogs were deemed SM-affected on repeat MRI. These findings suggest that the SM status of young adult C grade dogs and even young A grade dogs can be ambiguous, that is, the SM status can change as the dog ages.
The parents of the offspring were not all novel as the cohort included a variety of relationships: siblings, half siblings, parent/grandparent/great grandparents. There were no data for CxD breeding combinations in CKCS and in the GB there were no data for *AxE, AxA, AxC, AxE, CxC, CxD, CxE, CxU, DxE, ExE and UxC breeding combinations.
The MRI status of the offspring are tabulated in Tables 3 (CKCS) and 4 (GB). The most common breeding combination in both breeds was UxD, that is, an SM-affected dog with a dog of unknown status (34 per cent CKCS and 38 per cent GB parental combinations). The second most common breeding combination was *A or AxU (30 per cent CKCS and 28 per cent GB parental combinations). The proportion of breeding combinations that included at least one *A or A parent was 50 per cent for CKCS and 55 per cent for GB. Unfortunately, the proportion of SM-affected versus SM-free offspring cannot be tested statistically because it is not possible to estimate the significance of breeders' bias with respect to their pedigree knowledge in their breeding choices, for example, selecting breeding dogs that are closely related to an A grade dog or avoiding dogs that are clinically affected but not MRI confirmed. However, taking into account this lack of statistical robustness, a number of observations could be made. Grade A or *A offspring (145 CKCS and 41 GB) generally only occurred when there was at least one A grade parent. When dogs with one unknown parent and ambiguous C grade dogs were excluded, there were only three grade A CKCS that had both parents who were SM-affected. These three dogs were younger than 2.7 years of age when MRI scanned, that is, their lifetime SM-free status was not assured. The study suggests that using at least one *A grade parent increases the likelihood of SM-free offspring. In the CKCS group (Table 3), there are 131 (49+82) offspring where at least one parent was *A grade and if the authors exclude the ambiguous C grade offspring (33) there are 98 offspring. Of these, 54 (18+13+23) were either *A or A, hence 54/98 (55 per cent) were SM-free. For the same group of dogs, when at least one parent is A grade, there were 145 (64+81) offspring (not double counting the *AA) and of these, 59 were C grade leaving 86 offspring. Thirty-five of the 86 (41 per cent) were either *A or A. Hence the majority of offspring (55 per cent) were SM-free from at least one *A parent as opposed to only 41 per cent in the case of only one A grade parent. If the GB data (Table 4) are considered in a similar way, excluding the C grade offspring, 26 of the 45 (58 per cent) offspring from at least one *A parent were *A or A, compared with two of the four (50 per cent) offspring from at least one A grade parent albeit from a very small sample. Once again, the offspring from a parental cross where at least one parent is *A increases the likelihood of them being SM-free.
However, using only *A parents was not a guarantee of the offspring being SM-free. In the GB cohort, three of eight dogs (37.5 per cent) were SM-affected from *Ax*A parental crosses. Of these, two dogs were uncle and nephew from a GB family, which did not have CM. The third had SM and CM and resulted from a mating of a *A dam with CM and CCD and *A sire that had neither CM nor SM.
In Table 5, the ambiguous grade C dogs (154 CKCS and 10 GB) have been removed together with all offspring involving a C grade parent (24 CKCS and one GB) and there is a more straightforward comparison between the breeding of SM-free (*A or A with *A or A) and SM-affected dogs (D or E with D or E) with all possible combinations including a parent of unknown status (U). This is represented in percentage of offspring in Fig 1. It can be seen in both the CKCS and GB groups that offspring without SM were common when the parents were both clear of SM (SM-free; CKCS 70 per cent, GB 73 per cent). Conversely, offspring with SM were likely when both parents had SM (SM-affected; CKCS 92 per cent, GB 100 per cent). In the CKCS group, the 8 per cent SM-free offspring were young dogs (younger than 3.1 years old) and therefore are still at risk for developing SM. One of these young SM-free dogs had three affected siblings. There were only two dogs in the GB group, that is, the figure of 100 per cent SM-affected is unreliable. A mating of one SM-free parent with an SM-affected parent was risky for SM affectedness with 77 per cent of CKCS offspring and 46 per cent of GB offspring being SM-affected.
SM associated with CM in the CKCS was first reported in 1997 (Rusbridge 1997). The timing coincided with the availability of spinal MRI for animals and it is unlikely that the emergence of SM was because it was a ‘new’ disease but more a reflection of the ability to diagnose it (Rusbridge and Knowler 2004). It became apparent that SM was prevalent in the CKCS and was a disease that affected some other toy breed dogs (Rusbridge and Knowler 2004, Rusbridge and others 2009, Lewis and others 2010, Parker and others 2011). The spectrum of disease varies from asymptomatic to dogs with severe neuropathic pain and/or neurological deficits relating to spinal cord damage (Rusbridge and others 2007). The significance of asymptomatic dogs is that breeders may be unaware and the dogs' offspring may be symptomatic (Rusbridge and others 2005). With the increasing availability of MRI and with the provision of veterinary services offering SM screening at comparatively low cost, many breeders now ascertain the SM status of their dogs. However, guidance was sought on how to select suitable breeding combinations (Cappello and Rusbridge 2007). A recommendation to breed only unaffected dogs is too simplistic and may even have a negative impact as it fails to take account of the problem that the majority of dogs are bred as young adults and SM can occur as a late onset disease. Also in the CKCS, the prevalence of SM may be as high as 70 per cent in older dogs (Parker and others 2011) and there is a danger of further genetic bottlenecking. Therefore, a group of veterinary neurologists offered a system that took account of age and also made provision for breeding asymptomatic SM-affected dogs that were free of other known inherited diseases (Cappello and Rusbridge 2007). However, since the mode of inheritance of SM was unknown, this guidance may have been inappropriate for the CKCS or other breeds with a tendency for SM. The 2006 breeding guidelines have been adopted by several GB breeders. However, unlike the CKCS where CM appears to be ubiquitous, the GB can have SM independently from CM (Rusbridge and others 2009).
This paper serves to report the early outcome of using the 2006 breeding guidelines for dogs contributing to a genome project and whose pedigrees were known. The study suggested that SM-free offspring were more likely if both parents were SM-free and SM-affected offspring were more likely if both parents had SM. There was a trend that older SM-free dogs (*A) only resulted when at least one parent was *A. This finding seems logical but since *A and A grade dogs are in a minority it is problematic for breeders to select SM-free dogs especially when the SM status of the dog can alter with age. Often, the true SM status may not be known until after the dog is used for breeding. In 2010, there were 2136 CKCS litters registered with the UK kennel club (8095 puppies). Of these, 33 per cent had a sire aged younger than 2.5 years on the day of mating and 36 per cent had a dam younger than 2.5 years on the day of mating. This translates as 55 per cent of the KC registered CKCS puppies having one parent younger than 2.5 years and 14 per cent having both (Grahame Ford, personal communication). Thus, a recommendation that only A or *A grade parents be used is unlikely to be practical for breeders. If a younger SM-free dog (C grade) is mated with an older dog (A grade), then there is greater risk for SM than with two A grade parents. The data in the study (Tables 3 and 4) suggested that 63 per cent (12/19) of offspring were SM-free from A or*AxC parents. This is likely to be an overestimation of the proportion of SM-free dogs since these figures include C grade offspring, which may develop SM as they become older. However, it could be argued that a recommendation to include at least one A grade dog in any proposed mating is a reasonable, practical alternative since it allows breeders to use their younger dogs. However, a far better proposal for prevention of SM and maintaining genetic diversity is a proposed mate select programme using estimated breeding values (EBV). EBVs are a statistical estimate of the genetic risk for a given disease in an individual dog and a measure of the likelihood of their passing on the disease to their offspring (Lewis and others 2010). An EBV is the best method of genetic evaluation available for complex traits and can be calculated for most dogs even if they have not undergone an MRI scan, as long as they are related to dogs that have been MRI scanned. In other words, knowing the EBV may help breeders choose safer breeding combinations. However, the validity of EBVs depends on the collection of accurate populationwide data. Moreover, for the scheme to be successful, it is important that MR images from every MRI scanned dog be submitted for assessment, including clinical cases and whether or not the animal is required for breeding. Since dogs were shown to be SM-affected when MRI was repeated at five years or older and after they had been used for breeding, it is advantageous for breeders to ascertain the SM status of the dog when at least five years old. Even if the dog is no longer being used for breeding purposes, this information is likely to be valuable for breeding decisions for the dog's descendants and for determining accurate EBVs.
This study also looked at the risk of using dogs of unknown status. Many breeders are reluctant to subject their dogs to an MRI scan citing a number of reasons including cost and risk of anaesthesia or sedation. Undoubtedly, fear of an unfavourable result also plays a part. Not all unscanned dogs are equal and experienced breeders may select a dog on the basis of it having relatives that are SM-free or select a stud dog that is known to have produced a high proportion of SM-free offspring: one unscanned champion stud dog was mated with 16 bitches resulting in 11 C grades, 10 A grades and only three SM-affected offspring. Results from this study suggested that using an SM-affected dog (grade D) with a dog of unknown status resulted in higher numbers of SM-affected dogs in the older population (75 per cent SM-affected CKCS and 59 per cent affected GB). Using an SM-free (*A or A) CKCS with an unknown dog still resulted in 43 per cent of older offspring being SM-affected (50 per cent GB) and therefore, it is recommended that all breeding dogs from breeds susceptible to SM be MRI screened.
The most controversial part of the 2006 breeding recommendations was that it permitted older clinically asymptomatic SM-affected dogs to be bred to older SM-free dogs. This was because there were only a few A grade dogs identified when the breeding guidelines were first proposed and overuse of these might limit genetic diversity, reducing the effective population size. This investigation showed that a proportion of the offspring from such a cross can be SM-free, however, in the CKCS, the number of SM-affected dogs was far greater than SM-free dogs (77 per cent SM-affected). Moreover, where these offspring had siblings, the ratio of SM-free to SM-affected was 1:1 compared with 3.5:1 for *A or Ax*A or A parents or 1:1.7 for D or ExU parents (Knowler and Rusbridge, unpublished data). It is debatable whether it is ethical to knowingly breed a dog with an inherited disease especially when the majority of the offspring may be destined for the pet-owning public, and this study suggested that it is not advisable to use SM-affected dogs for breeding. However, if the prevalence of SM is as high as 70 per cent in the CKCS (Parker and others 2011), this will have dire consequences for the effective population size of the breed. Again it is hoped that using an EBV mate select programme may allow safer parental crosses, maintaining genetic diversity while decreasing the number of SM-affected offspring.
One problem noted with the 2006 breeding guidelines was that if the dog was deemed asymptomatic but SM-affected when first scanned over 2.5 years old it was attributed a grade D status. However, in some cases it was suspected, due to the size of the syrinx, that if the dog had been scanned when younger then the syrinx would have been apparent, in other words, the dog was actually an E grade. It was also possible that some dogs may have been clinically affected (ie, an F grade) as the early behavioural signs of pain from SM and/or CM can be subtle and/or intermittent and may not be detected in a routine clinical examination. In addition, some breeders fail to recognise or acknowledge clinical signs of SM and/or CM. Since the aim of the guidelines is to reduce the incidence of early onset and clinical SM, a future recommendation is that a D status (or equivalent) will only be appropriate if the dog was first proved by MRI to be SM-free before 2.5 years of age. One encouraging observation was that the proportion of grade E (SM-affected younger than 2.5 years of age) dropped from an average of 15 dogs per year in years 2004 to 2006 to 11 dogs per year in years 2007 to 2009, that is, the breeding guidelines may be achieving the goal of reducing the risk of early onset SM, however, this hypothesis would be needed to be tested in a more vigorous study.
This report has a number of limitations but particularly that it is not an independent and controlled breeding programme. The source of information is from that submitted at the time of MRI screening and the breeders made the decisions with regard to which dogs to breed and which dogs to submit for MRI screening and at what age. Although participants were worldwide, the study has been biased towards identifying as many SM-free dogs as possible as a control cohort for the genome project. Due to the lack of independence of the individual offspring from each other, the authors have confined themselves to making observations and have avoided testing for statistical significance given all appropriate testing procedures require independent observations; thus resultant p values would have been meaningless. It is possible these observations are misleading given this lack of independence, the short study timescale and the limited numbers in the GB cohort. Despite these limitations, however, the authors believe that it is useful for breeders to have some feedback on the breeding programmes they have been using, even if this is simply a starting point for a more formal study.
The study is also limited by the potentially late onset nature of SM. In the study, 46.9 per cent of the CKCS and 20 per cent GB were scanned when younger than 2.5 years (ie, C/E grades). Given the natural history of SM, it is likely that a proportion of the A and C grade dogs will ultimately prove to be SM-affected. The cost of scanning older dogs, often retired in pet homes, means data collection is difficult. Continuing to monitor the outcome of parental crosses and proportion of SM-affected offspring should be an important future goal in any proposed MRI screening programme. A further limitation of the study is that it does not take account of the SM affectedness of grandparents, whose status breeders may or may not know.
The study also suggests that, although the 2006 breeding guidelines might increase the proportion of SM-free dogs, a reservoir of affected dogs remain. An EBV mate select programme may help in this respect but in the near future this is only anticipated for the CKCS breed. For other breeds, it is recommended that as much information about the ancestors is ascertained as possible and that EBV mate select programme be developed as rapidly as possible.
The relationship between CM and SM is complex (Rusbridge and others 2009, Driver and others 2010). This study does not take account of CM since this is not included in the 2006 breeding guidelines. This is partly because CM is almost ubiquitous in the CKCS (Cerda-Gonzalez and others 2009). In the past 24 months, one of the authors (CR) reported 564 MRI scans from breeding CKCS and there were no dogs without CM and only six dogs (1.1 per cent) with mild CM. The tendency for CM may be a fixed genetic trait in the CKCS. The subsequent development of SM may be dependent on other modifying or protective genes that influence severity and age of onset. It is possible that both the 2006 breeding guidelines and an EBV mate select programme may merely select for protective traits, reducing the incidence of clinical and early disease but without reducing the prevalence of the main gene(s) for SM associated with CM. Epistasis (gene interaction that perturbs the normal Mendelian ratios) is a common feature of complex traits and different breeds with SM may have their own unique genomic characteristics (modifier genes) with any gene mutation. In the GB cohort, 67 of 93 dogs (72 per cent) had CM. Of the CM-free GB dogs, nine of 26 (35 per cent) had SM. This suggests in this breed the genetic aetiology of SM may be different from CKCS. It will only be possible to understand the mode of inheritance when genome studies are completed. It is expected that the use of genomic breeding values will further reduce the incidence of this disease.
In conclusion, the results from this study suggest that it is appropriate to continue using the breeding guidelines for both the CKCS and GB until a more robust system based on EBV or genetic testing is available. The following modifications are suggested but it should be realised that these are the recommendations that are based on limited data and consequently should be subjected to further prospective vigorous study: To increase the number of SM-free offspring, at least one parent should be ascertained to be SM-free by MRI as a young adult. In ideal circumstances, both parents would be SM-free. According to the study by Parker and others (2011), the optimum age for this early MRI screening is 36 months. If an SM-affected dog is used, for example, to preserve desirable traits or to increase genetic diversity then ideally the chosen mate would either be selected on the basis of its EBV and/or would be an older SM-free dog (five years or older). The offspring of the proposed mating should also be MRI scanned and ideally bred to older SM-free dogs. The SM status of the dog when at least five years old should be established. SM has a complex inheritance and an EBV mate select programme should allow breeders to select safer breeding combinations. To ensure success, the programme requires a substantial collection of accurate populationwide data. Consequently, all breeding dogs from breeds susceptible to CM and SM should be MRI screened and these results should be submitted to a central source. Pedigree and clinical history from dogs presenting with clinical signs of CM and/or SM should also be submitted to this central system. ‘D’ status (or equivalent) will only be appropriate if the dog was first proved to be SM-free before 36 months of age. Future breeding recommendations should also take account of dogs with CCD less than 2 mm. Guidance for other breeds like the GB will need to take account of any additional research findings with regard to the relationship between CM and SM.
The study is underpinned by the owners and breeders who have paid for their dogs' MRI scans and the authors acknowledge this and also the many veterinarians who have given their time and effort towards the genome research. Many have gone to extraordinary lengths to send DNA samples to the University of Montreal and MRI scans to CR to ensure phenotypic consistency. The authors also thank the CKCS and GB carers who have donated to Syringomyelia DNA Research that helped fund the study. Special thanks are also due to the For the Love of Ollie Fund for funding MR images of both CKCS and GB, and the Cavalier Club of Canada and UK and Rupert's Fund that funded the MR imaging of many CKCS. The study also included data generously provided by the following institutions and individuals: Guelph University (Dr Roberta Poma) 25 CKCS; University of North Carolina (Dr Natasha Olby) 10 CKCS; University of Ohio (Dr Carly Abraham) 7 CKCS; Matheson Blvd Veterinary Services, Mississauga (Jane Sykes and Karen Kennedy) 12 CKCS; University of Georgia (Dr Simon Platt) 22 GB. The authors also acknowledge the comments from Mr Grahame Ford, who has produced an annual compilation of CKCS statistics each year since 1993. Finally, thanks to the Goddard Veterinary Group and Dr Erik Noorman of Dierenkliniek den Heuvel, Netherlands for their commitment and support.
Provenance not commissioned; externally peer reviewed
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.