The prevalence of syringomyelia was investigated in a sample population of 555 Cavalier King Charles spaniels. All dogs, which were declared by their owners to be showing no clinical signs of syringomyelia, underwent MRI to determine the presence or absence of the condition. Data were analysed by logistic regression to determine the effects of sex and age on the prevalence of syringomyelia. Only increased age was found to have a significant effect. The prevalence of syringomyelia was 25 per cent in dogs aged 12 months, increasing to a peak of 70 per cent in dogs aged 72 months or more.
Statistics from Altmetric.com
SYRINGOMYELIA is an increasingly well-recognised condition affecting Cavalier King Charles spaniels (CKCS) due to the recent widespread availability of MRI in veterinary practice. The disease is characterised by the formation of fluid-filled cavities (syrinxes) within the parenchyma of the spinal cord and, in human beings, it has been associated with histological evidence of degeneration and ultrastructural abnormalities within the spinal cord (Milhorat and others 1995). Clinical signs of cervical syringomyelia in CKCS may be clinically undetectable but range to apparent severe dysaesthesia of the neck and chest region, although scratching at the neck and shoulder region (often without making contact with the skin) is most common. Clinical signs are more common in association with syrinxes that are asymmetric or of a large diameter (Rusbridge and others 2007).
In general, syringomyelia is thought to arise because stenosis of the subarachnoid space alters the flow of the cerebrospinal fluid such that pressure gradients develop between the subarachnoid space, central canal and parenchyma of the spinal cord (Rusbridge and others 2006). The intraparenchymal fluid may thus derive from the central canal or from blood vessels within the cord, depending on the mechanical model. In the CKCS, syringomyelia develops most commonly within the cervical spinal cord, and protrusion of the caudal cerebellum through the foramen magnum is thought to cause the stenosis of the subarachnoid space necessary for syrinx formation. This cerebellar herniation appears to result from foreshortening of the caudal skull, termed Chiari-like malformation, and it appears to be almost ubiquitous in the CKCS breed (Cerda-Gonzalez and others 2009). However, many affected dogs do not develop concurrent syringomyelia and it is therefore possible that the aetiology of syringomyelia involves factors in addition to, or even independent of, Chiari-like malformation.
The prevalence of syringomyelia within the CKCS breed is currently unknown. A recent publication suggested that it varies with age and may be over 60 per cent for specific and small cohorts (Lewis and others 2011). The British Veterinary Association and the Kennel Club are preparing to implement a screening programme to identify syringomyelia in CKCS. Syringomyelia has a significant heritable component (Lewis and others 2010) and it is intended that the results will be used to direct selective breeding to reduce or eliminate the condition. To enable the efficacy of this scheme to be assessed in future, it is imperative to define the current prevalence of the disease within the population. Pedigree CKCS breeders in the UK and abroad have been informally screening their dogs for several years, providing sufficient data for this analysis.
The aim of the present study was to establish the prevalence of syringomyelia within the CKCS breed and to determine whether age and sex had a significant effect on prevalence.
Materials and methods
The sample population comprised pedigree CKCS from the UK and the Netherlands, which had been presented to an MRI facility between 2004 and 2010 for screening to determine the presence or absence of cervical syringomyelia. All animals were declared asymptomatic by their owners, defined by the lack of typical clinical signs such as air scratching, spontaneous vocalisation and scoliosis.
Each dog was sedated before scanning, using a protocol that varied between centres; a typical combination was 0.3 mg/kg butorphanol and 10 µg/kg medetomidine. The dogs were positioned in sternal recumbency and T1- and T2-weighted sagittal and transverse MR images were obtained extending from the cerebellum to C5 vertebra. The resultant scans were read by diplomats of the European College of Veterinary Neurologists (CR and NDJ) at the Stone Lion Veterinary Hospital, London, and at the Queen's Veterinary School Hospital, Cambridge. In the present study, syringomyelia was defined as a T1-weighted hypointense area within the spinal cord that was greater than 2 mm in diameter at its widest point (Fig 1). Screening outcome was described in binary terms, as either positive or negative for the condition. Many dogs were scanned twice but only the result of the first scan was used.
Initially, the prevalence of syringomyelia was expressed as a proportion of the entire scanned population. Then, because there is evidence that symptomatic syringomyelia is more common in older dogs (Lewis and others 2010), the effect of age on the odds of a positive diagnosis of syringomyelia was analysed. Two methods of analysis were used to derive different measures of the prevalence of syringomyelia. First, the dogs were categorised according to age into six groups: group 1, 0 to 12 months; group 2, 12 to 24 months; group 3, 24 to 36 months; group 4, 36 to 48 months; group 5, 48 to 60 months, and group 6, over 60 months. The data were used initially to derive a table of odds of a positive diagnosis at each age and were tested using a chi-square test for measuring the trend in odds. Subsequently, the data were entered into a logistic regression analysis in combination with sex to determine the odds ratio for the diagnosis of syringomyelia between sexes, within each age group and for differences in the odds ratio between age groups. Finally, the age in months was entered directly into another logistic regression analysis together with sex to derive a coefficient to express the effect of each month of increasing age on the probability of a positive diagnosis of syringomyelia. All tests were carried out using Stata 11 for Windows, using P<0.05 to indicate rejection of the null hypothesis.
A total of 555 dogs were available for study. Their age ranged from 0.37 to 128 months (mean [sd] 32  months), and 356 (64 per cent) dogs were female. Overall, 255 of 555 (46 per cent) dogs were given a positive diagnosis of syringomyelia. The prevalence of syringomyelia in dogs in different age groups is given in Fig 2; there was an observable trend that the prevalence of syringomyelia increased with age, which was confirmed by a highly significant chi-squared test for the trend in odds ratios (ORs) (chi-squared 26.9, P<0.0001).
Logistic regression detected no significant effect of sex (OR 0.949, 95 per cent confidence interval [CI] 0.658 to 1.370), but a highly significant effect of age (OR per month 1.022, 95 per cent CI 1.013 to 1.031, P<0.0001). Further analysis demonstrated that for all age groups that comprised dogs younger than 36 months, there was a significantly lower OR of positive diagnosis of syringomyelia (Table 1). The overall increase in OR of positive diagnosis of 1.022 per month corresponds to an approximately 1.3 increase in odds ratio per year.
The present study confirms the widely held belief that syringomyelia is a highly prevalent condition within the CKCS breed, with an overall prevalence within this tested population of 46 per cent. There is evidence from this data to suggest that the lifetime risk of developing syringomyelia may be even higher, because in CKCS of six years and older the prevalence was 70 per cent; however, because this was a cross-sectional study, this risk cannot be determined. Sex did not influence the likelihood of a positive scan for syringomyelia; the genetic basis of the disease therefore does not appear to be sex-linked.
There are, however, two caveats to the conclusion regarding a gradually increasing prevalence with age. First, the confidence limits for the OR in the oldest group of animals were wide (because of the relatively small number of animals included within this group) and therefore must be considered less reliable. Secondly, although specifically excluded in terms of the present study, it is possible that the breeders may have included their dogs for the screening procedure while knowing that they expressed clinical signs of syringomyelia. There is currently no action to prevent this type of fraudulent behaviour, since the clinical signs rarely have an objective test that can be applied (Harcourt-Brown and others 2011). On the other hand, it is probable that most dogs showing obvious clinical signs of syringomyelia were excluded from the present analysis, yet contribute an additional factor to the overall prevalence in the breed that is not included in the prevalence calculated here.
The evidence for a lower prevalence in younger animals is more reliable (because of the higher numbers included in the present study and the lower likelihood of false inclusion) and this effect lasts until dogs are at least three years of age. This finding has important implications for the design of a screening test procedure and may conflict with the current recommendations that the optimum age for screening should be 30 months. These data would imply that it is probable that dogs aged up to three years may yet have reduced odds for the diagnosis of syringomyelia. However, there is a need for the dogs to be screened when they are reasonably young so that breeders can decide at an early stage whether their animals are suitable for breeding; many breeders would consider 36 months unduly old.
The prevalence of asymptomatic syringomyelia in the oldest age group was 70 per cent (Fig 1). Syringomyelia is therefore likely to be one of the most prevalent heritable conditions affecting CKCS; estimates for the prevalence of heart murmur, widely considered a common defect of the breed, range between 8.6 per cent (Häggstrom and others 1992) and 50 per cent (Lewis and others 2010) for middle-aged dogs. The prevalence of syringomyelia in CKCS is also high, relative to other heritable diseases that are perceived to be common in other breeds of dog. For example, hip dysplasia has been reported to affect up to 23 per cent of German shepherd dogs (Coopman and others 2008). An exception is the very high incidence of dystocia in some brachycephalic breeds, in which approximately 80 per cent of litters are delivered by caesarean section (Evans and Adams 2010).
The high lifetime prevalence of syringomyelia raises concerns for the welfare of the CKCS breed and also suggests that eliminating the genetic risk factors for the disease by selective breeding may be difficult, because the heritability has previously been shown to be complex (Lewis and others 2010) and the prevalence of the determinant genes within the population is therefore likely to be high. The true prevalence of syringomyelia in the general CKCS population is expected to be higher than that found in this sample population because symptomatic dogs were specifically excluded.
The authors are grateful to the pedigree CKCS breeders who participated in the voluntary syringomyelia screening programme. J. E. Parker also thanks the Wellcome Trust for funding his Integrated Training Fellowship and An Vanhaesebrouk for her support during the preparation of the manuscript. The authors also thank Annette O'Connor for her advice during preparation of the final manuscript.
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.