Subaortic stenosis (SAS) is a cardiac disorder with a narrowing of the descending aorta below the left ventricular outflow tract of the heart. It occurs in several species and breeds. The Newfoundland is one of the dog breeds where it is more common and usually leads to death at early adulthood. It is still discussed to which extent SAS has a genetic background and what its mode of inheritance could be. Extensive pedigree data comprising more than 230,000 Newfoundland dogs from the European and North American population reaching back to the 19th century including 6023 dogs with a SAS diagnosis were analysed for genetic factors influencing SAS affection. The incidence and prevalence of SAS in the analysed Newfoundland population sample were much higher than those reported in previous studies on smaller population samples. Assuming that some SAS-affected dogs remained undiscovered or were not reported, these figures may even be underestimated. SAS-affected Newfoundland dogs were more often inbred and closer related to each other than unaffected dogs, which is an indicator for a genetic background of SAS. The sex had no significant impact on SAS affectedness, pointing at an autosomal inheritance. The only simple mode of inheritance that fitted the data well was autosomal codominant with lethal homozygosity and a penetrance of 1/3 in the heterozygotes.
Statistics from Altmetric.com
Aortic stenosis is a cardiac disorder with a narrowing of the aorta. Constriction of the ascending aorta is a dominant feature of supravalvular stenosis in man, an inherited vascular disorder, which can be caused by mutations in the elastin gene (ELS) on human chromosome 7 (Ewart and others 1994). For subaortic or subvalvular stenosis, the constriction occurs on the descending aorta below the left ventricular outflow tract of the heart. So far, it has not been possible either to map it to a chromosome, or to define its mode of inheritance. Apart from the dog, subaortic stenosis (SAS) has been described in several species such as the pig, the cat, the cow and man (Patterson and Detweiler 1963, Pyle and others 1976, Masci and others 2008, OMIA 2011). SAS in the dog has been reported in several breeds (Baumgartner and Glaus 2003). Breed-specific high prevalence was found in the Newfoundland dog (Jones and others 1981, Freedom and others 2005), the Boxer (Patterson and Detweiler 1963, Swift 1996, Bussadori and others 2001, Freedom and others 2005, Menegazzo and others 2005, Linde and Koch 2006), the Rottweiler (del Palacio and others 1998) and the German shepherd dog (Patterson and Detweiler 1963). In the dog, SAS often leads to premature death. This is especially true for large breeds such as the Newfoundland.
In human beings, SAS is reported to occur in 0.5 per cent or less of infants. In dogs, SAS is estimated to occur at a rate of about 0.89 per cent in purebred dogs and 0.26 per cent in mongrels (Mulvihil and Priester 1973, Freedom and others 2005). Patterson (1968) reported 39 cases of subvalvular stenosis of 290 total dogs having one or more cardiovascular disorders, which corresponds to 13.4 per cent. In the Newfoundland, the prevalence for all kinds of congenital cardiac disease is 0.78 per cent, SAS being the most common lesion (Freedom and others 2005). Among active breeders, the prevalence of SAS in the Newfoundland population is currently perceived as being much higher.
Inherited versus acquired
There has been much discussion as to whether stenosis is inherited, acquired or both. Freedom and others (2005) gave a good overview of this discussion with a special focus on human beings and the Newfoundland dog, claiming that there is little evidence of fixed SAS in newborn babies or puppies (Pyle and others 1976, Pikula and others 2005, Markovic and others 2009). This could lead to the conclusion that SAS has an acquired aetiology. However, SAS has also been observed in neonates, in which case it would be a congenital heart defect (Kleinert and others 1993, Pikula and others 2005). Familial clustering of SAS and SAS-affected siblings descended from unaffected consanguineous parents was reported, indicating a genetic background (Digilio and others 1993, Fatimi and others 2006).
A current hypothesis postulates that the presence of morphological abnormalities within the outflow tract could lead to an elevation of septal shear stress. In the presence of a genetic predisposition and cellular proliferation in response to septal shear stress, this could lead to the known fixed and stenotic lesion (Muna and others 1978, Cape and others 1997, Cilliers and Gewillig 2002, Freedom and others 2005, Piacentini and others 2007). This hypothesis supports the argument of SAS being inherited, but only developing later (Freedom and others 2005, Markovic and others 2009).
Mode of inheritance
Freedom and others (2005) have not only shown strong similarities in the expression of SAS between human beings and dogs, but also in its mode of inheritance. It has been proposed that SAS in human beings and/or dogs is inherited in an autosomal-recessive mode (Digilio and others 1993, Piacentini and others 2007). Other authors have suggested an autosomal-dominant mode (Piacentini and others 2007) or a complex polygenic mode (threshold model) of transmission (Fatimi and others 2006, Ahmad and others 2007, Piacentini and others 2007). Up to the present time, all speculations on the heritability and mode of inheritance are based on very small pedigree information or cross and backcross trial schemes (Patterson 1968, Pyle and others 1976, Freedom and others 2005).
In most cases reported in the literature, affected individuals (dogs and human beings) are born to unaffected parents. This fact strongly supports the theory that SAS follows a recessive mode of inheritance. As such, it makes it extremely difficult for breeders to exclude carriers of SAS because in the most informative case (monogenic bi-allelic recessive inheritance), only one-quarter of all offsprings will be homozygous and thus show clinical signs of SAS. Yet, another half of the offsprings will be heterozygous carriers, showing no clinical signs.
SAS screening practice
SAS often leads to death soon after detection in early adulthood. SAS can only be diagnosed by cardiologists and involves extensive examinations. Usually, unaffected SAS carriers or affected dogs whose hearts have not been examined by a cardiologist cannot be clearly identified at present and are therefore not excluded from breeding on a phenotypic basis. Unfortunately, stud breeders often do not declare a diagnosis of SAS as they fear a negative impact on their reputation. The number of unreported cases may therefore be high. Hence, breeding aimed at eliminating the often lethal disease in the Newfoundland dog is extremely difficult and inefficient, even in Switzerland, where cardiac tests for stud breeding animals are compulsory. Leroy and Baumung (2011) discussed the effects of popular mating practices on the dissemination of genetic disorders in detail, focusing in particular on dog breeding.
The aim of the study was to show the spread of SAS in the context of the development of the Newfoundland population and to analyse possible risk factors for SAS in the Newfoundland dog breed.
Materials and methods
Extensive pedigree data on more than 230,000 Newfoundland dogs from the European and North American populations, extending back to the 19th century and including many founder animals of the breed, were analysed. After data validation, 229,854 dogs remained in the pedigree, with 6023 dogs (2.66 per cent) having a cardiac examination sufficient for SAS status determination, that is, auscultation by stethoscope and continuous wave-Doppler echocardiography for peak velocities (Vmax) in the aorta. Dogs were classified as grade 1 (unaffected, Vmax < 1.8 m/s), grade 2 (suspected, mildly affected Vmax 1.8 to 2.0 m/s) or grade 3 (severely affected, Vmax > 2.0 m/s). Classification was done by an experienced cardiologist (Dr Michael Deinert, Collegium Cardiologicum e.V., 35578 Wetzlar). The pedigree structure was analysed with regard to several important population parameters in relation to SAS and their interactions, such as sex, inbreeding coefficient F, effective population size Ne calculated from the rate of inbreeding ∆F (Ne = 1/(2*∆F)) and average relationship. The parameters were estimated using the programs CFC (Sargolzaei and others 2009) and PopRep (Groeneveld and others 2009). Interactions of multiple variables were tested using a log-linear model as described by Zar (1999). Log-linear analysis is a version of chi-square analysis for multidimensional contingency table data. The null hypothesis is tested by computing chi-squared as,
with f being the frequency in row i, column j and tier k.
Differences in sex within affected and unaffected dogs were tested by applying a chi-squared test applied on the 2 x 2 contingency table. Differences between the percentage of inbreds and the average inbreeding coefficient of the two groups were tested with a two-sample t-test for independent samples. Interaction modelling and significance testing were done with the VassarStats online programs (Lowry 2011). Average relationship within affected and unaffected dogs was tested with the non-parametric Wilcoxon rank-sum test using the software ‘Statistical Analysis System’ Release 9.1 (SAS 2011).
Spreading of SAS
The Newfoundland dog population evaluated in this study had a registration rate of less than 1000 dogs per year of birth from the beginning of the 20th century until the 1970s (Fig 1). This low registration rate is basically due to the incomplete information available for these early years. Between 1977 and 2007, the period of time during which all dogs having a cardiac examination sufficient for SAS status determination were born, the number of registered puppies born every year fluctuated between 4000 and 7000. The pedigree contained 32,630 full-sib groups with an average family size of 6.39 (2 to 55 progeny). The average number of offsprings per sire and per dam was 17.7 (1 to 500) and 8.7 (1 to 79), respectively (results not shown). The real family sizes and number of offsprings per sire or dam may be higher, as often not all offsprings are registered.
Of the 6023 dogs having a cardiac examination sufficient for SAS status determination, 138 were severely affected, 143 were mildly affected and 5792 were considered unaffected (Table 1). Both the total number of dogs with cardiac examinations and the number of SAS-affected dogs increased strongly between the early 1980s and the turn of the millennium (Fig 2). While the incidence increased to 18 per cent by 1991, it has since stabilised at about 4 to 5 per cent. Fig 3 shows that the trend of the average inbreeding coefficient of the newborn puppies increased every year until the end of the 1960s, when it dropped prominently. The effective population size has increased slightly in the past 30 years, but has always been lower than 200 animals. With the tremendous increase in the real population size in the 1970s and 1980s, the number of highly inbred dogs (>15 per cent) increased proportionally (Fig 1). The percentage of inbred dogs was 82.2 per cent in the dataset, with inbreeding coefficients of up to 56.22 per cent (Table 1).
Risk factors for SAS
Most of the dogs examined for SAS (96.7 per cent) were inbred to some degree, with maximum inbreeding coefficients of 38.02 per cent. SAS-affected dogs were significantly more often and more highly inbred than unaffected dogs (Table 1). The average relationship among SAS-affected dogs was significantly higher than among unaffected dogs (Table 2). There was no significant difference in the distribution of males and females among SAS-affected and unaffected dogs (Table 1). The log-linear analysis for interactions between SAS, sex and number of inbreds (INBRED) gave significant results for the three-way interaction (Table 3). The two-way interactions removing the effects of the third variable and the two-way interactions with nested effects confirmed the results of the chi-squared test in Table 1: Significant interactions were found for SAS x INBRED and SEX x INBRED, whereas SAS x SEX interactions were not significant.
All possible combinations of one or two affected parents and unaffected offsprings and vice versa were found in the Newfoundland population (Table 4). The only constellation which never occurred was two affected parents with severely affected offsprings. There was no evidence for a simple Mendelian inheritance mode such as dominant, recessive or co-dominant inheritance.
Spreading of SAS
In total, 4.6 per cent (281) of 6023 dogs with a cardiac examination sufficient for SAS status determination displayed clinical signs of SAS (Table 1). If the dataset is considered representative, it indicates that the prevalence of SAS in the Newfoundland population is higher at present than SAS prevalence reported in the literature (Freedom and others 2005). Recent studies (Swift 1996, Bussadori and others 2001) suggest an SAS prevalence of up to 33 per cent for the Boxer. The incidence in the Newfoundland population increased strongly until the year of birth 1991, when it reached its highest peak. Of all puppies having a cardiac examination sufficient for SAS status determination and born in 1991, 18.5 per cent were SAS affected (Fig 2). The perception of active breeders that the number of SAS-affected dogs is increasing was confirmed, since the highest absolute number of SAS-affected dogs was born in 1997 and 2000 (29 cases each year). The sharp decline in the number of both diagnoses and incidence in the past two years does not necessarily reflect a decreasing trend, but is rather due to lack of information, as the cardiac examination sufficient for SAS status determination should not be done until the dogs are at least one year old (Bussadori and others 2000). Hence, only few of the dogs born in 2006 and 2007 were already having a cardiac examination sufficient for SAS status determination.
If SAS has a genetic background, mating relatives (with resulting inbreeding) would lead to higher incidences of SAS in the population. Most dogs were only slightly (<6 per cent) or moderately (6 to 10 per cent) inbred, although inbreeding levels of up to 56.22 per cent were found (Table 1, Fig 1). Such extreme inbreeding values can only occur when very closely related dogs are mated over several generations. As an example, offsprings of full-sib matings have an inbreeding coefficient of ‘only’ 25 per cent. Today, many breeders' associations (eg, in Holland, Switzerland and Germany) do have regulations to limit very close inbreeding. More than 80 per cent of all dogs in the evaluated population were inbred to some extent (Table 1), which is very high compared with the values cited in the literature (Leroy and Baumung 2011), but may be due to the deep pedigree studied. Since 1983, the average pedigree completeness has been more than 90 per cent over three generations and more than 75 per cent over six generations (results not shown). The average inbreeding level increased until the end of the 1960s, partially due to the steadily increasing pedigree completeness (Fig 3). There may be different reasons for the subsequent pronounced decline in the inbreeding coefficient. First, many dogs of Canadian origin but unknown pedigree and born in the late 1960s through the 1970s are included in the database. Secondly, genetic exchange between North America and Europe was increasing at that time. Thirdly, regulations on colours and patterns were eased. Fourth, awareness arose that very high inbreeding may lead to health problems, for example, a higher incidence of inherited diseases. Fig 3 shows the effective population size, deduced from the annual rate of inbreeding ∆F. It is rather low (<200 animals) throughout the whole century. The only exceptions were in 1924 and 1994, when, due to a sharp drop in the average annual inbreeding coefficient to very low values, the effective population size was greatly overestimated. Hence, although the actual population size is several thousand dogs, the gene pool in the population is equivalent to the gene pool of 200 unrelated dogs, that is, a very limited number. The consequence of this small gene pool is rapidly increasing relatedness between dogs, unless relationships are carefully monitored. Inbreeding coefficients in the offsprings would also rapidly increase, bringing many potentially negative ramifications. In livestock, effective population sizes below 50 to 100 are considered as critical for the survival and health of a breed (Meuwissen 1999, FAO 2007). Therefore, inbreeding levels in the Newfoundland breed should be kept under close observation.
Risk factors for SAS
SAS-affected dogs were significantly more often and more highly inbred (Tables 1 and 3) than unaffected dogs. The slightly lower pedigree completeness of unaffected dogs compared with affected dogs (results not shown) may explain part of this difference. Average relationship among SAS-affected dogs was also significantly higher than among unaffected dogs (Table 2), which indicates a genetic background of SAS. The occurrence of SAS did not differ significantly between the genders (Table 1). Similar proportions of SAS-affected dogs occurred in males (43.17 per cent) and females (56.83 per cent), respectively. When looking at inbreeding as a second factor (Table 3), significant interactions were found between the number of inbreds and SAS (more inbreds among SAS-affected dogs) and the number of inbreds and sex (more inbreds within the males), whereas SAS and sex had no significant interaction (not more SAS-affected males). The higher number of inbreds in the males is most probably because breeding males usually are stronger selected and do have more offsprings than the breeding females. Hence, as in other studies (Patterson 1968, Baumgartner and Glaus 2003, Petric and Cvetko 2009), no evidence was found of sex having an influence on SAS affection.
From the pedigree information, there was no evidence of a simple mode of inheritance for SAS such as co-dominant, recessive or dominant (Table 4). Several studies suggest complex inheritance (Freedom and others 2005, Markovic and others 2009). It should be kept in mind that SAS cannot usually be diagnosed at birth but only later in a dog's life. Development of SAS may be controlled by a complex genetic background, by the environment or both. More information on environmental factors and loci/alleles (SNP data) is needed to gain further insights into the influence of the environment or multiple alleles/loci on SAS.
If analysis is restricted to monogenic, bi-allelic inheritance for SAS, the only mode of inheritance that matches the data of this and other studies (Patterson 1968, Markovic and others 2009) reasonably well is co-dominant inheritance with lethal homozygosity for an allele S causing SAS. Hence, all SAS-affected dogs would be heterozygous. Offsprings that are homozygous for the SAS causing allele would not be viable and would die before or soon after birth. As a consequence, litters of two affected parents would on average be smaller than those of two unaffected parents, as one in four of the siblings would be homozygous for the SAS allele and not viable. Unfortunately, it was frequently the case that not all siblings - and especially no stillborn puppies - of a litter were registered in the database, so that no further investigation on this hypothesis was possible with the given dataset. Further, lethal homozygosity could explain why it is extremely rare to observe SAS in puppies (Freedom and others 2005, Markovic and others 2009) and none in fetuses (Freedom and others 2005): homozygous fetuses may be aborted at a very early stage of pregnancy.
If the mode of inheritance was monogenic and homozygotes (SS) were viable, Mendelian inheritance would lead to 50 per cent (one affected parent SS or Ss) and 75 per cent or more (two affected parents SS or Ss) affected offsprings, respectively. These figures are much higher than the observed number of affected offsprings in the study population (Table 4). Moreover, when homozygotes are viable and reproduce, offsprings of two affected parents are much more often affected by SAS than offsprings of one affected parent. This was not true in the study population (Table 4).
Based on the data presented in Table 4, it may be hypothesised that the SAS causing allele is not only lethal when homozygous, but also that its penetrance (or rather the detection rate of affected dogs) is about one in three. This hypothesis is in keeping with the percentage of affected and unaffected offsprings of affected parents (Table 4). Given the fact that there are unreported and undetected cases of SAS - which biases the number of affected offsprings towards unaffected offsprings as shown in Table 4; co-dominant inheritance with lethal homozygosity and a penetrance of 33.3 to 100 per cent seems the most appropriate simple mode of inheritance for SAS in the Newfoundland. Yet, polygenic inheritance cannot be ruled out.
If one assumes that the offsprings of parents with unknown SAS status are representative for the whole population, the frequency of a hypothetical co-dominant, lethal SAS allele S in the population is about 1.4 to 4.2 per cent (Table 4). This frequency is in the same range as allele frequencies of other lethal genetic disorders (Manatrinon and others 2009, Leroy and Baumung 2011).
The analysed Newfoundland population was highly inbred and had an effective population size of less than 400 animals. SAS-affected dogs were more often inbred and more closely related to each other than the unaffected dogs, which suggests a genetic background of SAS. In the analysed Newfoundland population the incidence and prevalence of SAS were much higher than those reported in earlier studies on population samples that are smaller. Given the fact that some SAS-affected dogs remained undetected or were not reported, these figures may even be underestimated. Further inbreeding may lead to a further increase in the number of SAS-affected animals, if carriers of a probable SAS causing allele are not (or cannot be) excluded from breeding.
Sex had no significant impact on SAS affectedness, indicating that SAS inheritance is autosomal. Given the data of the studied population, it may be hypothesised that SAS follows a monogenic co-dominant mode of inheritance with partial to full penetrance (33.3 to 100 per cent). However, this hypothesis should be confirmed by complementary approaches, for example, a classical segregation analysis.
The authors kindly acknowledge the language editing of A. Rigby. The project was funded by the Federal Veterinary Office (FVO).
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.