Trends in the phenotypic hip status of selected breeds of dog as measured by the New Zealand Veterinary Association Hip Dysplasia scheme (1990-2008).

Abstract

AIM: To determine whether there has been improvement in the phenotypic hip dysplasia status in four susceptible dog breeds as measured by the New Zealand Veterinary Association (NZVA) Canine Hip Dysplasia (CHD) scheme. METHODS: A retrospective analysis of the NZVA CHD database was performed using records of all German Shepherd dogs, Labrador Retrievers, Golden Retrievers and Rottweilers that had undergone evaluation for hip dysplasia between 1990 and 2008. The effect of date of birth on the total hip score was analysed using linear regression, including the covariates of age and gender. When a significant effect of date of birth on total score was noted, ordinal logistic regression was performed to determine the probability of different grades of the Norberg angle and subluxation scores by year of birth; these categories being most indicative of laxity of the coxofemoral joint. Given the known heritability of hip phenotype, determined using radiological measurements, the hypothesis was that if sufficient selection pressure has been applied there would have been a trend towards a lower total score over time. RESULTS: For Labrador Retrievers (n=1,451), Golden Retrievers (n=896) and Rottweilers (n=313), there was no effect of date of birth on total score over the period of the study (p>0.1). For German Shepherd dogs (n=1,087), there was a significant trend to a lower total score over time (p=0.0003). However the actual size of the effect was small. Ordinal logistic regression on the Norberg angle and subluxation scores for German Shepherd dogs demonstrated a significant lowering of grade in both of these measures of hip laxity. CONCLUSIONS: This study failed to show significant improvement in the phenotypic hip status of three out of the four most populous large-dog breeds in the NZVA CHD database. Even in the German Shepherd dog, the trend towards a lower total score did not represent a substantial change. Lack of evidence of phenotypic improvement may be due to insufficient selection pressure over the course of the study, selective usage of the scheme (and thus a biased sample), or deficiencies within the NZVA CHD scoring method itself. CLINICAL RELEVANCE: Greater improvement might be possible if use of the scheme (or an equivalent) is made a compulsory requirement for registration of pedigree breeding stock, if greater selection pressure is applied and/or if pedigree data are included to enable estimations of breeding value.