In the present study, 0.16% of all canine tumours submitted to the diagnostic pathology laboratory at the Norwegian School of Veterinary Science during the period March 1998 to December 2009 were gastric carcinomas. This proportion corresponds to what Arnesen et al. (2001) found in their study of the Norwegian Canine Cancer Register from 1990–1998 .
Breed predisposition to gastric carcinoma was detected based on the PMR. The highest relative proportion of gastric carcinoma was observed in Tervuren, Bouvier des Flandres, Groenendael, collie (rough or smooth not specified), standard poodle, and Norwegian elkhound. Tervuren had a PMR of 56, indicating that this breed is 56 times more likely to be diagnosed with gastric carcinoma than the average breed in this database. Belgian shepherd dogs are bred as four separate types – Tervuren, Groenendal, Mallanois, and Lakenois – but breeding between these types is permitted by the Norwegian Belgian shepherd dog club (NBFK) when specific conditions are fulfilled and with limited frequency. Registration of offspring is determined by the length, colour, and structure of the coat . Even though the four types are closely related, there are no reports of increased risk of gastric carcinoma in the Mallanois or the Lakenois. It is unknown whether this is due to the Mallanois and Lakenois being less frequent types of Belgian shepherd dog or whether there is a different prevalence between the types.
Belgian shepherd dogs, both Groenendael and Tervuren types, and rough collies have previously been reported to be at increased risk of developing gastric carcinoma [3, 7, 8, 25]. Single cases of gastric carcinoma have previously been reported in Norwegian elkhound, standard poodle, and Bouvier des Flandres [5, 6, 12], but the present study is the first time that these breeds have been found to be at risk of this type of cancer. In this study, the breed Bouvier des Flandres was found to have an increased risk of developing gastric carcinoma, but the wide 95% CI for PMR renders the result uncertain, and it may not be clinically significant.
It is worth noting that the breeds boxer, Rottweiler, and flat-coated retriever were not among the cases of gastric carcinoma in our study. These are all common breeds in Norway, and in previous studies based on the Norwegian Canine Cancer Register they were reported to have the highest relative risk ratio (RR) for developing tumours in general . The database utilized for the current study included 616 tumour cases in boxers, 514 in Rottweilers, and 1,052 in flat-coated retrievers. A further interesting observation is the lack of mixed breeds among the cases of gastric carcinoma, although there are 1,932 tumours from mixed breeds in the database, making ‘mixed’ the most frequent ‘breed’. Despite being the most common pure-bred dog in the database, with 1,407 tumours, the English setter is only represented with a single case. Chow chow and Staffordshire bull terrier breeds have previously been reported to have an increased risk of gastric carcinoma [3, 11]. Both breeds were found to be infrequent in the database, with 14 tumour cases in chow chows and 22 tumour cases in Staffordshire bull terriers during the period 1998–2009. Neither of these breeds are represented in the current study material. This may be because these breeds are relatively uncommon in Norway, or that they have a different genetic composition compared to dogs from the same breeds in other countries.
Males were found to have significantly higher odds of gastric carcinoma than females in our study. This finding is in agreement with results from other studies and similar to what has been described in humans [2, 3, 6–10].
All dogs with gastric carcinoma in the current study, where clinical information was available, experienced vomiting. This, together with weight loss and anorexia, were the most commonly reported clinical signs. The complete clinical symptomatology may have been underreported, as a complete veterinary medical record was not available for all cases. In several cases, ‘vomiting’ was the only clinical sign reported, as this probably has been the main reason for diagnostic work-up, even though other clinical signs may have been present. The described clinical signs are in accordance with those reported in earlier studies [3, 5, 6, 8, 9].
Dog breeds arise from a limited number of ancestors and combined with a frequent use of popular males, each domestic breed is an isolated population [26, 27]. The genetic diversity within breeds is reduced, but there is a greater genetic divergence between breeds . This makes the performance of genetic studies of dogs easier than those conducted in complex populations. Dogs share environments with their owners. Furthermore, because they naturally develop spontaneous cancers, in contrast to mouse models, they constitute an excellent comparative model for cancer . The results from the study, identifying several breeds at risk, suggest that a genetic predisposition is of greater significance in dogs than humans, who have experienced a rapidly declining incidence of gastric carcinoma due to recent changes in the environment . The dog may thus be a useful model for further comparative genetic studies of gastric cancer.
If differences in the ability of veterinarians to diagnose a gastric tumour and the willingness of dog owners to pursue and pay for diagnostic work-up are related to dog breed, they may influence the results of our study and be a source of selection bias. Underestimation can occur as a result of false negative biopsies. There might also be an underestimation of the prevalence of gastric cancer, as advanced disease and the age of an affected dog can result in owners being reluctant to pursue further diagnostic examinations.
A strength of the present study is the use of PMR to assess the proportion of all the affected animals in the population that have a particular disease under question. This allowed comparison with the background population of submitted cases, as opposed to the case series commonly reported in veterinary literature. However, a weakness of PMR is its sensitivity to variations in breed popularity and uncertainty in the estimates related to uncommon breeds. This is further influenced by the low prevalence of the disease. This might explain why some breeds previously described to have a higher risk are not represented in this study. It is also possible that a larger study would have been able to detect more breeds at increased risk among Norwegian dogs.
In the Norwegian Canine Cancer Register there is a higher proportion of females than males. This is probably due to mammary tumours dominating the register, as reported in previous studies . Of the 44 cases that were diagnosed with gastric cancer in the Norwegian Canine Cancer Register, 22.7% were given a new histopathological diagnosis when they were re-evaluated in 2011. The proportion of cases given a new histopathological diagnosis corresponds to that found in previous studies based on the Norwegian Canine Cancer Register .
The breed predisposition observed in the current study indicates a genetic susceptibility to gastric carcinoma. This is in contrast to gastric cancer in humans, which in developed countries has declined significantly in incidence over the past 50 years, attributed largely to the eradication of H pylori infection . A reasonable interpretation is that environmental risk factors play a smaller role in the development of gastric cancer in dogs compared to humans.