- Email: [email protected]
Breed predispositions in canine mast cell tumour: A single centre experience in the United Kingdom
The Veterinary Journal 197 (2013) 496–498
Contents lists available at SciVerse ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Short Communication
Breed predispositions in canine mast cell tumour: A single centre experience in the United Kingdom James Warland ⇑, Jane Dobson Queen’s Veterinary School Hospital, Department of Veterinary Medicine, University of Cambridge, Madingley Road, Cambridge CB3 0ES, United Kingdom
a r t i c l e
i n f o
Article history: Accepted 21 February 2013
Keywords: Dog Mast cell tumour Breed Incidence
a b s t r a c t Genetic factors play a major role in carcinogenesis. Many breeds have been reported to be predisposed to mast cell tumour (MCT) development using various methods and diverse control populations. A database of 222 dogs with MCT seen at a UK university referral hospital was compared to three control populations, namely, an insured population of UK dogs, registrations with the UK Kennel Club and other dogs seen through the same hospital. Odds ratios were calculated for each breed. Boxers, Labradors, Golden Retrievers and Staffordshire Bull Terriers appeared predisposed to MCT development. English Springer Spaniels, English Cocker Spaniels, German Shepherd Dogs, West Highland White Terriers and Cavalier King Charles Spaniels were underrepresented. Ó 2013 Elsevier Ltd. All rights reserved.
Mast cell tumours (MCTs) are common tumours of the canine skin, representing up to 20% of all skin tumours in dogs (Dorn et al., 1968). In the United Kingdom (UK), MCT is the second most common canine malignancy (Dobson et al., 2002). Cutaneous MCTs are typically solitary lesions, but their clinical appearance can be variable and dogs can develop more than one unrelated MCT (Murphy et al., 2006). Breed predispositions have been reported in many different canine cancers. A recently published American study considered the breed distribution of various canine cutaneous tumours, including MCT (Villamil et al., 2011). The aim of the present study was to document which breeds were over- and under-represented for the development of MCT in a single referral centre in the UK. The study consisted of a retrospective analysis of all dogs with MCTs entered into the oncology database of a UK university referral teaching hospital between 1st January 1997 and 31st December 2008. The data were compared to a population of insured dogs, Kennel Club registrations and to the breeds seen at the same hospital through all departments. The insurance population was a previously published database of 130,684 insured dogs in the UK (Edwards et al., 2003). The UK Kennel Club registration data used were published online and covered a 10-year period (2002– 2011).1 The hospital breed list was established by considering new registrations to the hospital over a similar 11-year period (2000– 2010).
⇑ Corresponding author. Tel.: +44 1223 337621. E-mail address: [email protected] (J. Warland). See: Kennel Club, 2013. 10 Yearly Breed Statistics. http://www.thekennelclub. org.uk/item/1128 (Accessed 10 January 2013). 1
1090-0233/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tvjl.2013.02.017
For the analysis, the incidence of MCT in each breed was compared in breeds presenting more than five MCT and for the ten most frequently seen breeds at the hospital. Odds ratios (OR) were calculated, with 95% confidence intervals (CI)2 as to whether a breed was more likely to develop MCT compared to the hospital population, insurance population and Kennel Club registrations. The Peto OR3 was used if no dogs of that breed were represented. Where OR 95% CI did not cross ‘1’, the result was deemed statistically significant at the 5% level. In total, 222 dogs were included with 42 different breeds represented in the dataset. Five breeds included more than five cases: Labradors (n = 66), Golden Retrievers (n = 29), Boxers (n = 22), Staffordshire Bull Terriers (SBT) (n = 16), and Jack Russell Terriers (JRT) (n = 6); there were 23 crossbreed dogs. Compared to the insurance, Kennel Club and hospital populations, the Labrador Retriever, Boxer, SBT and Golden Retriever were overrepresented. The English Cocker Spaniel, German Shepherd Dog (GSD), West Highland White Terrier (WHWT), and Cavalier King Charles Spaniel (CKCS) were underrepresented in the MCT group, and this was found in comparison to all three control populations. The Springer Spaniel (all varieties included) also appeared underrepresented against the hospital population and the Kennel Club populations, but this was not found to be a statistically significant difference in the insurance population. With the Border Collie, the OR appeared to be reduced in all analyses, but the CIs were wide and therefore the results were not statistically significant other than for the hospital population. The full results are presented in Table 1. 2 3
See: http://www.hutchon.net/ConfidOR.htm. See: http://www.hutchon.net/ConfidORnulhypo.htm.
497
J. Warland, J. Dobson / The Veterinary Journal 197 (2013) 496–498
Table 1 Odds ratios (OR) and 95% confidence intervals (CI) for the development of mast cell tumour disease in different breeds of dogs, by comparing those affected to a previously published database of insured animals, to the other dogs seen by the same veterinary hospital, and to registrations with the United Kingdom (UK) Kennel Club. OR significantly >1 are highlighted in bold and OR significantly <1 are highlighted in italics. Breeds are highlighted in the same way if the majority of the population analyses support an under- or over-representation. n = total number in group. Number affected (n = 222) n (%)
Odds ratio vs. insurance population (n = 130,684)
Odds ratio vs. hospital population 2000– Odds ratio vs. UK Kennel Club registrations 2010 (n = 12,112) 2002–2011 (n = 2,554,018)
n (%)
OR
Upper 95% CI limit
n (%)
OR
Lower 95% CI limit
Upper 95% CI limit
n (%)
OR
Border Collie
1 (0.45)
0.24 0.03
1.74
0.02
0.91
3.54
2.44 1.57
3.80
2.08 1.33
3.25
3.44 2.21
5.33
Cavalier King Charles Spaniel English Cocker Spaniel Crossbreed
0 (0)
0.35
0.17
0.73
0.35
0.15
0.85
23,014 (0.9) 79,246 (3.1) 101,079 (4.0)
0.50 0.07
22 (9.9)
417 (3.4) 609 (5.0) 285 (2.4)
0.13
Boxer
2378 (1.8) 5628 (4.3) 4529 (3.5)
0.35
0.18
0.69
0.20
0.05
0.82
0.06
0.95
2.45
0.81 0.53
1.26
195,388 0.11 0.03 (7.7) Not applicablec
0.44
1.59 1.03
3 (1.4)
0.13
0.04
0.42
0.21
0.07
0.67
2.33
3.11 2.09
4.64
1.10 0.49
2.49
123,283 0.27 0.09 (4.8) 93,775 3.94 2.67 (3.7) Not applicablec
0.84
1.58 1.07
66 (29.7)
2.90 2.17
3.89
2.79
2 (0.9)
0.23
0.06
0.92
0.16
0.04
0.62
Staffordshire Bull Terrier West Highland White Terrier
16 (7.2)
3.49 2.07
5.89
1.77 1.06
2.94
0.35
0.87
429,546 (16.8) 140,560 (5.5) 107,322 (4.2) 80,547 (3.2)
2.09 1.57
Springer Spaniela
452 (3.7) 1509 (12.5) 728 (6.1) 558 (4.6) 299 (2.5) 1541 (12.7) 464 (3.8) 258 (2.1) 276 (2.3)
0.23
German Shepherd Dog Golden Retriever Jack Russell Terrier Retriever
5568 (4.3) 8851 (6.8) 12,157 (9.3) 11,348 (8.7) Data not
0.36
0.76
Breed
2 (0.9) 23 (10.4)
29 (13.1) 6 (2.7)
0 (0)
16,259 (12.4) 4308 (3.3) 2844 (2.2) 6134 (4.7)
Lower 95% CI limit
availableb 2.97 2.23
3.97
0.27 0.06
1.07
3.49 2.09
5.81
0.35
0.65
0.19
0.14
Lower 95% CI limit
0.17
Upper 95% CI limit
5.82
a Both dogs that presented with mast cell tumours were English Springer Spaniels (ESS), and this was compared to the ESS data in the insurance population and Kennel Club registrations, but to all Springer Spaniels in the hospital population, as specific data relating to which variety of Springer Spaniel were not available. b Data were taken from Edwards et al. (2003), which did not include data on Jack Russell Terriers. c The UK Kennel Club does not register Jack Russell Terriers or crossbreed dogs.
The results of this work support findings from previous studies that Boxers, Labrador Retrievers and Golden Retrievers are predisposed to MCT (Murphy et al., 2006; Villamil et al., 2011). These results also agree with a previous study showing that GSD are less prone to MCT (Villamil et al., 2011). The same study found several other breeds apparently less likely to develop MCT. Interestingly, Bronden et al., (2010) documented an underrepresentation for overall cancer development in the GSD population of Denmark but this is the first study to report that the English Cocker Spaniel, WHWT and CKCS are less frequently affected by MCT than other breeds. It is inevitable that no sample population can act as a genuine surrogate for the UK dog population and it is important to note that there are differences between the groups in terms of breed proportions. Analysis of the breeds affected was carried out against three separate populations to try to account for weaknesses in each control group. Findings in certain breeds that are consistent across all three control populations are particularly compelling, and the finding of apparent breed differences in MCT incidence (e.g. English Cocker Spaniels, WHWT and CKCS) is further supported by the consistency of the findings. All three control groups have the same limitation in that they do not exclude dogs that have MCT or will develop the condition in the future. However, given the large size of these control populations and likely low prevalence of MCT in these breeds, statistical analysis using OR is still valid and unlikely to bias results. For example, Dobson et al. (2002) demonstrated a prevalence of 129 MCT in 130,684 dogs in the insurance population used, but the breeds affected were not specifically analysed. The insurance population has previously been used to look at the general incidence of neoplasia and lymphoma specifically
(Dobson et al., 2002; Edwards et al., 2003). It is likely to be a good surrogate to the population that would attend the referral clinic, as most dogs are insured. However, it is possible that different breeds of dogs will have different insurance patterns, given their role in society and this may bias the insurance population. Owners of different dog breeds may be more or less inclined to insure their dog depending on many factors, including their relationship with the dog or perceived health status of the breed. In this population, certain breeds such as the GSD showed increased rates of insurance (9.3% of total insured population) compared to the hospital population (6.0%) or Kennel Club registrations (4.8%), whereas others, such as the Border Collie and crossbreed dogs showed lower insurance rates compared to the hospital population (1.8% vs. 3.4% and 6.8% vs. 12.5%, respectively). Similarly insurance patterns will vary with age and the insurance population was noted to have a high proportion of young animals (Dobson et al., 2002). The UK Kennel Club allows voluntary registration of pedigree dogs and is the largest organisation registering dogs in the UK. Crossbreed dogs are not included in the database, and certain breeds such as the Jack Russell Terrier are not recognised by the UK Kennel Club, so comparisons could not be made with these populations. Nevertheless, this group provides the most comprehensive comparison group for pedigree dogs, but there are a few limitations. Firstly, registration is voluntary, so it is possible that some breeds are less likely to be registered than others, particularly working dogs. Secondly, the database contains new puppies registered and thus may contain a different population seen in adult dogs developing MCT if breeds are becoming rapidly more or less popular. By comparing the numbers of dogs seen with MCT to the hospital population, it was possible to see whether particular breeds are
498
J. Warland, J. Dobson / The Veterinary Journal 197 (2013) 496–498
more likely to have a MCT than another disease whilst accounting for likelihood of referral. However, other breed disease predispositions could artificially lower the relative prevalence of MCT. General agreement about dogs over- and under-represented in data obtained from each analysis supports the findings of the other, and suggests that the findings are genuine. The analyses were not completely concordant in only three breeds, namely, crossbreed dogs, Border Collies and English Springer Spaniels (ESS). Of these groups, the agreement appeared good for the ESS as two groups showed statistically reduced risk of MCT, and the final group, the insurance population, showed a tendency towards a reduced OR (OR 0.27; 95% CI 0.06–1.07). In the crossbreed dogs, the insurance population comparison suggested a predisposition, but the hospital population was neutral to any effect. Given the diverse genetic background of crossbreed dogs, it seems irrelevant to attempt to draw conclusions in these animals. The differences between the analyses may reflect a lower rate of insurance in crossbreed compared to pedigree dogs. No analysed breeds had discordant results (i.e. some suggesting over- and others underrepresentation) between the different populations. Previous studies have suggested that certain breeds, such as the Boxer (Bostock, 1986) and Pugs (McNiel et al., 2006) develop low grade tumours. The population were all referred for treatment, so may represent higher grade or difficult to manage tumours. Breeds considered underrepresented may simply develop low grade tumours and not be referred. This weakness might be alleviated as a reasonable proportion of animals were referred for incompletely excised, grade II MCTs, as the centre is one of a few offering radiotherapy in the UK. In conclusion, these data indicate that Boxers, Labradors, Golden Retrievers and Staffordshire Bull Terriers appeared predisposed to MCT development, while to our knowledge the finding of decreased risk in English Cocker Spaniels, WHWT and CKCS has not been previously reported. The population studied represents a single referral population in the UK. Hence, these breed predispositions may not be representative in other countries or other UK regions if the dog populations are genetically distinct. Although significant effort has been placed on studying breeds predisposed
to many tumours, future work considering those breeds underrepresented for MCT may uncover insights into the aetiology of this common condition. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper. Acknowledgements The authors would like to thank Richard Parker, Centre for Applied Medical Statistics, University of Cambridge, for his time, expertise and guidance, in reviewing the statistical methods used. References Bostock, D.E., 1986. Neoplasms of the skin and subcutaneous tissues in dogs and cats. British Veterinary Journal 142, 1–19. Bronden, L.B., Nielsen, S.S., Toft, N., Kristensen, A.T., 2010. Data from the Danish Veterinary Cancer Registry on the occurrence and distribution of neoplasms in dogs in Denmark. Veterinary Record 166, 586–590. Dobson, J.M., Samuel, S., Milstein, H., Rogers, K., Wood, J.L.N., 2002. Canine neoplasia in the UK: Estimates of incidence rates from a population of insured dogs. Journal of Small Animal Practice 43, 240–246. Dorn, C.R., Taylor, D.O., Schneider, R., Hubbard, H.H., Klauber, M.R., 1968. Survey of animal neoplasms in Alameda and Contra Costa counties, California II, cancer morbidity in dogs and cats from Alameda County. Journal of the National Cancer Institute 40, 307–318. Edwards, D.S., Henley, W.E., Harding, E.F., Dobson, J.M., Wood, J.L.N., 2003. Breed incidence of lymphoma in a UK population of insured dogs. Veterinary and Comparative Oncology 1, 200–206. McNiel, E.A., Prink, A.L., O’Brien, T.D., 2006. Evaluation of risk and clinical outcome of mast cell tumors in pug dogs. Veterinary and Comparative Oncology 4, 2–8. Murphy, S., Sparkes, A.H., Blunden, A.S., Brearley, M.J., Smith, K.C., 2006. Effects of stage and number of tumours on prognosis of dogs with cutaneous mast cell tumours. Veterinary Record 158, 287–291. Villamil, J.A., Henry, C.J., Bryan, J.N., Ellersieck, M., Schultz, L., Tyler, J.W., Hahn, A.W., 2011. Identification of the most common cutaneous neoplasms in dogs and evaluation of breed and age distributions for selected neoplasms. Journal of the American Veterinary Medical Association 239, 960–965.
Contents lists available at SciVerse ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Short Communication
Breed predispositions in canine mast cell tumour: A single centre experience in the United Kingdom James Warland ⇑, Jane Dobson Queen’s Veterinary School Hospital, Department of Veterinary Medicine, University of Cambridge, Madingley Road, Cambridge CB3 0ES, United Kingdom
a r t i c l e
i n f o
Article history: Accepted 21 February 2013
Keywords: Dog Mast cell tumour Breed Incidence
a b s t r a c t Genetic factors play a major role in carcinogenesis. Many breeds have been reported to be predisposed to mast cell tumour (MCT) development using various methods and diverse control populations. A database of 222 dogs with MCT seen at a UK university referral hospital was compared to three control populations, namely, an insured population of UK dogs, registrations with the UK Kennel Club and other dogs seen through the same hospital. Odds ratios were calculated for each breed. Boxers, Labradors, Golden Retrievers and Staffordshire Bull Terriers appeared predisposed to MCT development. English Springer Spaniels, English Cocker Spaniels, German Shepherd Dogs, West Highland White Terriers and Cavalier King Charles Spaniels were underrepresented. Ó 2013 Elsevier Ltd. All rights reserved.
Mast cell tumours (MCTs) are common tumours of the canine skin, representing up to 20% of all skin tumours in dogs (Dorn et al., 1968). In the United Kingdom (UK), MCT is the second most common canine malignancy (Dobson et al., 2002). Cutaneous MCTs are typically solitary lesions, but their clinical appearance can be variable and dogs can develop more than one unrelated MCT (Murphy et al., 2006). Breed predispositions have been reported in many different canine cancers. A recently published American study considered the breed distribution of various canine cutaneous tumours, including MCT (Villamil et al., 2011). The aim of the present study was to document which breeds were over- and under-represented for the development of MCT in a single referral centre in the UK. The study consisted of a retrospective analysis of all dogs with MCTs entered into the oncology database of a UK university referral teaching hospital between 1st January 1997 and 31st December 2008. The data were compared to a population of insured dogs, Kennel Club registrations and to the breeds seen at the same hospital through all departments. The insurance population was a previously published database of 130,684 insured dogs in the UK (Edwards et al., 2003). The UK Kennel Club registration data used were published online and covered a 10-year period (2002– 2011).1 The hospital breed list was established by considering new registrations to the hospital over a similar 11-year period (2000– 2010).
⇑ Corresponding author. Tel.: +44 1223 337621. E-mail address: [email protected] (J. Warland). See: Kennel Club, 2013. 10 Yearly Breed Statistics. http://www.thekennelclub. org.uk/item/1128 (Accessed 10 January 2013). 1
1090-0233/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tvjl.2013.02.017
For the analysis, the incidence of MCT in each breed was compared in breeds presenting more than five MCT and for the ten most frequently seen breeds at the hospital. Odds ratios (OR) were calculated, with 95% confidence intervals (CI)2 as to whether a breed was more likely to develop MCT compared to the hospital population, insurance population and Kennel Club registrations. The Peto OR3 was used if no dogs of that breed were represented. Where OR 95% CI did not cross ‘1’, the result was deemed statistically significant at the 5% level. In total, 222 dogs were included with 42 different breeds represented in the dataset. Five breeds included more than five cases: Labradors (n = 66), Golden Retrievers (n = 29), Boxers (n = 22), Staffordshire Bull Terriers (SBT) (n = 16), and Jack Russell Terriers (JRT) (n = 6); there were 23 crossbreed dogs. Compared to the insurance, Kennel Club and hospital populations, the Labrador Retriever, Boxer, SBT and Golden Retriever were overrepresented. The English Cocker Spaniel, German Shepherd Dog (GSD), West Highland White Terrier (WHWT), and Cavalier King Charles Spaniel (CKCS) were underrepresented in the MCT group, and this was found in comparison to all three control populations. The Springer Spaniel (all varieties included) also appeared underrepresented against the hospital population and the Kennel Club populations, but this was not found to be a statistically significant difference in the insurance population. With the Border Collie, the OR appeared to be reduced in all analyses, but the CIs were wide and therefore the results were not statistically significant other than for the hospital population. The full results are presented in Table 1. 2 3
See: http://www.hutchon.net/ConfidOR.htm. See: http://www.hutchon.net/ConfidORnulhypo.htm.
497
J. Warland, J. Dobson / The Veterinary Journal 197 (2013) 496–498
Table 1 Odds ratios (OR) and 95% confidence intervals (CI) for the development of mast cell tumour disease in different breeds of dogs, by comparing those affected to a previously published database of insured animals, to the other dogs seen by the same veterinary hospital, and to registrations with the United Kingdom (UK) Kennel Club. OR significantly >1 are highlighted in bold and OR significantly <1 are highlighted in italics. Breeds are highlighted in the same way if the majority of the population analyses support an under- or over-representation. n = total number in group. Number affected (n = 222) n (%)
Odds ratio vs. insurance population (n = 130,684)
Odds ratio vs. hospital population 2000– Odds ratio vs. UK Kennel Club registrations 2010 (n = 12,112) 2002–2011 (n = 2,554,018)
n (%)
OR
Upper 95% CI limit
n (%)
OR
Lower 95% CI limit
Upper 95% CI limit
n (%)
OR
Border Collie
1 (0.45)
0.24 0.03
1.74
0.02
0.91
3.54
2.44 1.57
3.80
2.08 1.33
3.25
3.44 2.21
5.33
Cavalier King Charles Spaniel English Cocker Spaniel Crossbreed
0 (0)
0.35
0.17
0.73
0.35
0.15
0.85
23,014 (0.9) 79,246 (3.1) 101,079 (4.0)
0.50 0.07
22 (9.9)
417 (3.4) 609 (5.0) 285 (2.4)
0.13
Boxer
2378 (1.8) 5628 (4.3) 4529 (3.5)
0.35
0.18
0.69
0.20
0.05
0.82
0.06
0.95
2.45
0.81 0.53
1.26
195,388 0.11 0.03 (7.7) Not applicablec
0.44
1.59 1.03
3 (1.4)
0.13
0.04
0.42
0.21
0.07
0.67
2.33
3.11 2.09
4.64
1.10 0.49
2.49
123,283 0.27 0.09 (4.8) 93,775 3.94 2.67 (3.7) Not applicablec
0.84
1.58 1.07
66 (29.7)
2.90 2.17
3.89
2.79
2 (0.9)
0.23
0.06
0.92
0.16
0.04
0.62
Staffordshire Bull Terrier West Highland White Terrier
16 (7.2)
3.49 2.07
5.89
1.77 1.06
2.94
0.35
0.87
429,546 (16.8) 140,560 (5.5) 107,322 (4.2) 80,547 (3.2)
2.09 1.57
Springer Spaniela
452 (3.7) 1509 (12.5) 728 (6.1) 558 (4.6) 299 (2.5) 1541 (12.7) 464 (3.8) 258 (2.1) 276 (2.3)
0.23
German Shepherd Dog Golden Retriever Jack Russell Terrier Retriever
5568 (4.3) 8851 (6.8) 12,157 (9.3) 11,348 (8.7) Data not
0.36
0.76
Breed
2 (0.9) 23 (10.4)
29 (13.1) 6 (2.7)
0 (0)
16,259 (12.4) 4308 (3.3) 2844 (2.2) 6134 (4.7)
Lower 95% CI limit
availableb 2.97 2.23
3.97
0.27 0.06
1.07
3.49 2.09
5.81
0.35
0.65
0.19
0.14
Lower 95% CI limit
0.17
Upper 95% CI limit
5.82
a Both dogs that presented with mast cell tumours were English Springer Spaniels (ESS), and this was compared to the ESS data in the insurance population and Kennel Club registrations, but to all Springer Spaniels in the hospital population, as specific data relating to which variety of Springer Spaniel were not available. b Data were taken from Edwards et al. (2003), which did not include data on Jack Russell Terriers. c The UK Kennel Club does not register Jack Russell Terriers or crossbreed dogs.
The results of this work support findings from previous studies that Boxers, Labrador Retrievers and Golden Retrievers are predisposed to MCT (Murphy et al., 2006; Villamil et al., 2011). These results also agree with a previous study showing that GSD are less prone to MCT (Villamil et al., 2011). The same study found several other breeds apparently less likely to develop MCT. Interestingly, Bronden et al., (2010) documented an underrepresentation for overall cancer development in the GSD population of Denmark but this is the first study to report that the English Cocker Spaniel, WHWT and CKCS are less frequently affected by MCT than other breeds. It is inevitable that no sample population can act as a genuine surrogate for the UK dog population and it is important to note that there are differences between the groups in terms of breed proportions. Analysis of the breeds affected was carried out against three separate populations to try to account for weaknesses in each control group. Findings in certain breeds that are consistent across all three control populations are particularly compelling, and the finding of apparent breed differences in MCT incidence (e.g. English Cocker Spaniels, WHWT and CKCS) is further supported by the consistency of the findings. All three control groups have the same limitation in that they do not exclude dogs that have MCT or will develop the condition in the future. However, given the large size of these control populations and likely low prevalence of MCT in these breeds, statistical analysis using OR is still valid and unlikely to bias results. For example, Dobson et al. (2002) demonstrated a prevalence of 129 MCT in 130,684 dogs in the insurance population used, but the breeds affected were not specifically analysed. The insurance population has previously been used to look at the general incidence of neoplasia and lymphoma specifically
(Dobson et al., 2002; Edwards et al., 2003). It is likely to be a good surrogate to the population that would attend the referral clinic, as most dogs are insured. However, it is possible that different breeds of dogs will have different insurance patterns, given their role in society and this may bias the insurance population. Owners of different dog breeds may be more or less inclined to insure their dog depending on many factors, including their relationship with the dog or perceived health status of the breed. In this population, certain breeds such as the GSD showed increased rates of insurance (9.3% of total insured population) compared to the hospital population (6.0%) or Kennel Club registrations (4.8%), whereas others, such as the Border Collie and crossbreed dogs showed lower insurance rates compared to the hospital population (1.8% vs. 3.4% and 6.8% vs. 12.5%, respectively). Similarly insurance patterns will vary with age and the insurance population was noted to have a high proportion of young animals (Dobson et al., 2002). The UK Kennel Club allows voluntary registration of pedigree dogs and is the largest organisation registering dogs in the UK. Crossbreed dogs are not included in the database, and certain breeds such as the Jack Russell Terrier are not recognised by the UK Kennel Club, so comparisons could not be made with these populations. Nevertheless, this group provides the most comprehensive comparison group for pedigree dogs, but there are a few limitations. Firstly, registration is voluntary, so it is possible that some breeds are less likely to be registered than others, particularly working dogs. Secondly, the database contains new puppies registered and thus may contain a different population seen in adult dogs developing MCT if breeds are becoming rapidly more or less popular. By comparing the numbers of dogs seen with MCT to the hospital population, it was possible to see whether particular breeds are
498
J. Warland, J. Dobson / The Veterinary Journal 197 (2013) 496–498
more likely to have a MCT than another disease whilst accounting for likelihood of referral. However, other breed disease predispositions could artificially lower the relative prevalence of MCT. General agreement about dogs over- and under-represented in data obtained from each analysis supports the findings of the other, and suggests that the findings are genuine. The analyses were not completely concordant in only three breeds, namely, crossbreed dogs, Border Collies and English Springer Spaniels (ESS). Of these groups, the agreement appeared good for the ESS as two groups showed statistically reduced risk of MCT, and the final group, the insurance population, showed a tendency towards a reduced OR (OR 0.27; 95% CI 0.06–1.07). In the crossbreed dogs, the insurance population comparison suggested a predisposition, but the hospital population was neutral to any effect. Given the diverse genetic background of crossbreed dogs, it seems irrelevant to attempt to draw conclusions in these animals. The differences between the analyses may reflect a lower rate of insurance in crossbreed compared to pedigree dogs. No analysed breeds had discordant results (i.e. some suggesting over- and others underrepresentation) between the different populations. Previous studies have suggested that certain breeds, such as the Boxer (Bostock, 1986) and Pugs (McNiel et al., 2006) develop low grade tumours. The population were all referred for treatment, so may represent higher grade or difficult to manage tumours. Breeds considered underrepresented may simply develop low grade tumours and not be referred. This weakness might be alleviated as a reasonable proportion of animals were referred for incompletely excised, grade II MCTs, as the centre is one of a few offering radiotherapy in the UK. In conclusion, these data indicate that Boxers, Labradors, Golden Retrievers and Staffordshire Bull Terriers appeared predisposed to MCT development, while to our knowledge the finding of decreased risk in English Cocker Spaniels, WHWT and CKCS has not been previously reported. The population studied represents a single referral population in the UK. Hence, these breed predispositions may not be representative in other countries or other UK regions if the dog populations are genetically distinct. Although significant effort has been placed on studying breeds predisposed
to many tumours, future work considering those breeds underrepresented for MCT may uncover insights into the aetiology of this common condition. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper. Acknowledgements The authors would like to thank Richard Parker, Centre for Applied Medical Statistics, University of Cambridge, for his time, expertise and guidance, in reviewing the statistical methods used. References Bostock, D.E., 1986. Neoplasms of the skin and subcutaneous tissues in dogs and cats. British Veterinary Journal 142, 1–19. Bronden, L.B., Nielsen, S.S., Toft, N., Kristensen, A.T., 2010. Data from the Danish Veterinary Cancer Registry on the occurrence and distribution of neoplasms in dogs in Denmark. Veterinary Record 166, 586–590. Dobson, J.M., Samuel, S., Milstein, H., Rogers, K., Wood, J.L.N., 2002. Canine neoplasia in the UK: Estimates of incidence rates from a population of insured dogs. Journal of Small Animal Practice 43, 240–246. Dorn, C.R., Taylor, D.O., Schneider, R., Hubbard, H.H., Klauber, M.R., 1968. Survey of animal neoplasms in Alameda and Contra Costa counties, California II, cancer morbidity in dogs and cats from Alameda County. Journal of the National Cancer Institute 40, 307–318. Edwards, D.S., Henley, W.E., Harding, E.F., Dobson, J.M., Wood, J.L.N., 2003. Breed incidence of lymphoma in a UK population of insured dogs. Veterinary and Comparative Oncology 1, 200–206. McNiel, E.A., Prink, A.L., O’Brien, T.D., 2006. Evaluation of risk and clinical outcome of mast cell tumors in pug dogs. Veterinary and Comparative Oncology 4, 2–8. Murphy, S., Sparkes, A.H., Blunden, A.S., Brearley, M.J., Smith, K.C., 2006. Effects of stage and number of tumours on prognosis of dogs with cutaneous mast cell tumours. Veterinary Record 158, 287–291. Villamil, J.A., Henry, C.J., Bryan, J.N., Ellersieck, M., Schultz, L., Tyler, J.W., Hahn, A.W., 2011. Identification of the most common cutaneous neoplasms in dogs and evaluation of breed and age distributions for selected neoplasms. Journal of the American Veterinary Medical Association 239, 960–965.