Study of Changes in Hip Joint phenotype of dogs

[KingstonTodd]

Article on hip certification and improving hip scores in multiple breeds. The tables are formatted a bit funny from the copy/paste, but the first colum of numbers is the first breed listed above, and so on.

"Update of a Retrospective Cohort Study of Changes in Hip Joint Phenotype of Dogs Evaluated by the OFA in the United States, 1989–2003

JOHN B. KANEENE, DVM, MPH, PhD 1 , ULREH V. MOSTOSKY, DVM, MS 1 , and ROSEANN MILLER, MS 1
1 Center for Comparative Epidemiology and the Department of Small Animal Clinical Sciences, College of Veterinary Medicine, Michigan State University, East Lansing, MI 48824-1314
Correspondence to Address reprint requests to John B. Kaneene, A-109 Veterinary Medical Center, Michigan State University, East Lansing, MI 48824-1314. E-mail: kaneene@cvm.msu.edu.
Copyright © 2009 American College of Veterinary Surgeons
ABSTRACT


Objective—To determine whether there has been improvement in canine hip joint phenotype classifications of dogs whelped from 1989 to 2003 by the Orthopedic Foundation for Animals (OFA), by examining results of radiographic evaluations and identifying any trends in percentages of dogs classified as having desirable hip joint phenotypes.

Study Design—Retrospective cohort study.

Sample Population—OFA radiographic classifications (n=431,483) on dogs whelped between 1989 and 2003.

Methods—Numbers and percentages of dogs classified by hip joint phenotypes were determined for 2-year cohorts. Differences between breeds and sexes were assessed using the Fisher's exact test, and odds ratios with 95% confidence intervals were calculated to express associations. The Cochran–Armitage test for trend was calculated to identify significant trends over time.

Results—There were statistically significant (P<.05) increases in the proportion of all breeds of dogs evaluated as excellent and good from 1993 to 2003, controlling for gender and age at evaluation. Labrador Retrievers, Bernese Mountain Dogs, and Rottweilers had the highest proportions of excellent and good scores, and the highest rates of improvement in excellent and good scores were seen in Bernese Mountain Dogs and Rottweilers.

Conclusions—Results support the contention that there have been improvements in hip joint phenotype classifications in dogs in the United States since the previous study (1989–1992), through increases in the proportion of dogs receiving excellent and good classifications.

Clinical Relevance—Hip joint phenotype classifications can be used by dog breeders to develop breeding programs to improve the hip joints of future generations of dogs.


--------------------------------------------------------------------------------

Submitted April 2008; Accepted July 2008

DIGITAL OBJECT IDENTIFIER (DOI)
10.1111/j.1532-950X.2008.00475.x About DOI



INTRODUCTION


CANINE HIP dysplasia (CHD) is recognized as a polygenetic disease, in which phenotype is not a direct indicator of genotype. The condition is developmental, and adult dogs with normal hip joints may carry genes associated with CHD. To reduce the proportion of CHD-associated genotypes in dogs, selective breeding has been successfully implemented using radiographic hip evaluations to determine whether a dog has signs of CHD. Radiographic evaluations of canine hips, at a minimum of 24 months of age, can be correlated to the occurrence and severity of CHD in later life.1–3

The Orthopedic Foundation for Animals (OFA) has been evaluating hip joint phenotypes of dogs since 1966 to provide information to be used for selectively breed dogs with the goal of reducing CHD. The OFA database contains information based on owner-submitted radiographs, and their evaluation system has been shown to be over 80% accurate in the identification of CHD,4 and agrees with other canine hip evaluation protocols.5 Hips are classified as normal or as having signs of CHD—normal hips are further classified as excellent, good, or fair, whereas hips with signs of CHD are classified as mild, moderate, or severe.6 The OFA database is made available to owners and veterinarians for guidance in selecting animals for breeding, based on the OFA evaluation of the dog of interest and evaluations of any ancestors that were registered with the OFA. The OFA website, which provides access to this information, averages >2 million page requests per month.7

There have been reports of mixed results in effectiveness of selective breeding in reducing CHD,8–10 and a variety of studies have been undertaken to address this issue. Studies have been conducted to determine the prevalence of CHD in different canine populations using different diagnostic protocols and criteria,11,12 whereas other studies have examined the diagnostic capacity of different hip classification protocols.4,13,14 One area of concern is in the use of data from registries which rely on voluntary submissions, as does the OFA.11,12 Such databases are subject to self-selection bias, in that owners with clinically or radiographically abnormal hip conformation are less likely to submit radiographs for a potentially unfavorable OFA classification.11 One study12 estimated the prevalence of CHD in radiographs taken for reasons unrelated to hip conformation from a veterinary teaching hospital from 1991 to 1995, and found rates of CHD of 30% in Golden Retrievers and 35% in Rottweilers from this population, using a combination of OFA grading, Norberg Angle, and percent coverage of the femoral head as diagnostic tools. These rates were lower than those from a study based on dogs referred to a veterinary teaching hospital from 1991 to 1993, for hip classification using the PennHip method,11 in which Golden Retrievers had rates of CHD from 53% to 73%, whereas rates of CHD in Rottweilers were estimated from 41% to 69%. Both the Rettenmaier and Paster studies had sufficient numbers of Golden Retrievers (99 and 200, respectively) and Rottweilers (99 and 140, respectively) from which to draw statistical inferences, but both studies covered relatively short periods of time and could not address the question as to whether there has been change over time. Additionally, these studies involved animals from single veterinary teaching hospitals, whose canine populations may not be representative of the population of the United States.

To determine whether selective breeding had improved hip joint classifications over time in the population of dogs evaluated by the OFA, a study was undertaken to compare hip joint phenotypes of dogs whelped from 1972 to 1980 with dogs whelped from 1989 to 1992.15 Given the likelihood of self-reporting bias in cases where owners of possibly dysplastic dogs are less likely to submit radiographs for evaluation than owners of dogs with normal hip conformation, the 1997 study focused on changes at the high end of the OFA classification scale for normal dogs. Based on information from 270,978 radiographic evaluations, the study found that the percentage of dogs having hips classified as excellent increased from 7.8% in the 1972–1980 cohort, to 10.6% in the cohort whelped from 1989 to 1992. These increases were also greater in males than females, with the percentage of excellent males increasing from 7% to 11%, whereas females increased from 8% to 10%. Differences in the proportions of excellent classifications and the rates of improvement were also seen between breeds: Labrador Retrievers had the highest proportions of excellent classifications (10% and 15% from 1972 to 1980 and 1989 to 1992, respectively), and Rottweilers showed the highest proportions of improvement in excellent classifications (4–7% from 1972–1980 to 1989–1992). This study was conducted to update our previous cohort study15 to determine whether there has been improvement in hip joint phenotype classifications done by the OFA, by examining results of radiographic evaluations for dogs whelped between 1989 and 2003, and determining whether there were any trends in the percentage of dogs classified as having excellent or good hip joint phenotypes.



MATERIALS AND METHODS



Data
Data for the current study were collected from a database compiled by the OFA in 2005, from dogs whelped between 1989 and 2003 that had OFA examination for hip joint phenotype. In the OFA submission process, hip radiographs taken by a veterinarian, following American Veterinary Medical Association guidelines for positioning, are mailed by the owner to the OFA. These radiographs were independently evaluated by 3 randomly selected board-certified veterinary radiologists from a pool of 20–25 consulting radiologists in private practice and academia: the pool includes a core group of 10 radiologists, of which 9 have been readers for the OFA for over 25 years. Each radiologist evaluated the animal's hip status considering breed, sex, and age, and made classifications based on deviations from breed normal, scored as Excellent, Good, Fair, Mild CHD, Moderate CHD, and Severe CHD.6 As reported by the OFA, when results of 1.8 million classifications by 45 radiologists were analyzed, it was found that all 3 radiologists agreed 73.5% of the time on the same hip phenotype, and 21% of the time 2 radiologists agreed on the same hip grade, with the 3rd radiologist off 1 hip grade of the other 2.16

Analysis was conducted on dogs whelped from 1989 to 2003. The year of birth was divided into 7 cohorts: 1989–1992 (cohort represented in the previous study), 1993–1994, 1995–1996, 1997–1998, 1999–2000, 2001–2002, and 2003. Because dogs are evaluated by the OFA at a minimum of 24 months of age, the cohort of dogs whelped in 2004 and 2005 were likely to be underrepresented, and were excluded from analysis. Information from the OFA dataset included breed, sex, year of birth, age at evaluation (in months), and OFA classification score. To determine the rate at which OFA-evaluated dogs registered by the American Kennel Club (AKC), the number of dogs registered per year by breed from 1989 to 2003 was gathered from the AKC.



Statistical Analysis
Given the number of breeds available, separate analyses were conducted for common breeds of interest in the OFA dataset: Labrador Retrievers (22% of all records), Golden Retrievers (11%), German Shepherd Dogs (9%), Rottweilers (8%), and Bernese Mountain Dogs (1%). These 5 breeds make up over 50% of all records in the dataset, and records for all other breeds were combined for analysis. Percentages of dogs registered with the AKC that were evaluated by the OFA were calculated. Numbers and percentages of dogs classified as having excellent, good, fair, and combined mild/moderate/severe CHD hip joint phenotypes were determined for each year and for 2-year whelping cohorts. Differences in the numbers of dogs classified by hip joint phenotype between breeds were assessed using the Fisher's exact test (2-tailed analysis), controlling of age at evaluation and gender, and odds ratios with 95% confidence intervals were calculated to express the association between breed and year with classifications.

The Cochran–Armitage test for trend (Z) was calculated to determine if there were any significant trends over time in the percentages of dogs classified with different ratings. Statistical significance was set at P<.05. The likelihood of a dog receiving a given rating within a given whelping cohort was reported as odds ratios with 95% confidence intervals. Odds ratios >1.0 indicated that dogs whelped during the whelping cohort were more likely to have a specific rating as dogs whelped during other times in the study, whereas odds ratios <1.0 indicated that dogs whelped during the whelping cohort were less likely to have a specific rating compared with dogs whelped during other times.

To attempt to assess potential changes in submissions over time that might be because of increasing self-selecting or selfcensoring by owners as they become more aware of the consequences of negative OFA classifications, patterns of submissions were examined, and if the age at which dogs were evaluated changed significantly (P?.05) over time, this might be an indicator of changing bias. To estimate the percentage of AKC-registered dogs evaluated by the OFA in a given year for a specific breed, the number of classifications for dogs whelped in a given year was divided by the number of AKC registrations for that year. The age at which dogs were evaluated was divided into 4 categories (24–30 months of age, 30–35 months, 36–47 months, and over 47 months of age), and the proportion of dogs evaluated by age class was computed for each birth year. The Cochran–Armitage test for trend (Z) was calculated to determine if there were any significant linear trends over time.



RESULTS


A total of 431,483 classifications from 1993 to 2003 were available for analysis. The mean age of dogs in the study at the time of OFA evaluation was 29.5 months. The mean age of dogs at evaluation for those whelped in 1993–1995 was 33.1 months, 33.0 months in 1996–1998, 31.6 months in 1999–2001, and 27.8 months in 2002–2003.

The 5 breeds of interest in the dataset were Labrador Retrievers (n=102,960), Golden Retrievers (n=46,634), German Shepherd Dogs (n=36,793), Rottweilers (n=23,894), and Bernese Mountain Dogs (n=6878; Table 1). The numbers of OFA classifications declined over the course of the study for most breeds, whereas classifications of Bernese Mountain Dogs have increased over time. On average, the OFA evaluated ?5% of all AKC-registered dogs during the course of the study (Table 2); however, when viewed by the anticipated number of AKC-registered dogs expected to produce registered litters, the OFA classification rate was ?14%. Of the breeds of interest in the study, the highest classification rates for AKC-registered dogs were for Bernese Mountain Dogs (30% overall), and the lowest rates seen were for Rottweilers (4%) and other breeds combined (2%).


Table 1. Numbers of Dogs Evaluated by the Orthopedic Foundation for Animals (OFA), by Whelping Year (n=431,483 Dogs)


--------------------------------------------------------------------------------

Whelping Year Labrador Retriever Golden Retriever German Shepherd Rottweiler Bernese Mount Dog Other Breeds

--------------------------------------------------------------------------------

1993 8052 4342 4277 4701 492 19,133
1994 9624 4900 4365 4080 543 20,660
1995 10,534 5168 4356 3003 562 21,197
1996 10,405 4797 3810 2493 571 21,432
1997 10,133 4602 3557 1986 568 20,944
1998 10,004 4267 3103 1601 535 19,415
1999 10,075 4208 3137 1463 571 20,078
2000 9712 4028 2783 1283 628 19,843
2001 9071 3818 2763 1201 793 18,891
2002 8836 3636 2574 1193 877 18,546
2003 6514 2868 2068 890 738 14,185


--------------------------------------------------------------------------------





Table 2. Evaluation Rate of Evaluations of Hip Joint Phenotype Performed by the Orthopedic Foundation for Animals (OFA), for American Kennel Club (AKC) Registered Dogs Whelped from 1993 to 2003 (n=431,483)


--------------------------------------------------------------------------------

Labrador Retriever Golden Retriever German Shepherd Rottweiler Bernese Mount Dog Other Breeds

--------------------------------------------------------------------------------

Number registered with the AKC
1993–1994 251,292 132,447 158,935 206,756 3052 2,016,018
1995–1996 281,556 133,100 157,164 183,523 3531 1,851,733
1997–1998 316,302 135,839 140,503 130,498 3895 1,801,276
1999–2000 327,738 128,952 114,916 79,131 4354 1,640,002
2001–2002 320,586 118,621 98,588 51,465 5041 1,445,854
2003 155,896 52,520 43,938 18,214 3132 641,741
Evaluation rate (%)
1993–1994 7.0 7.0 5.4 4.2 33.9 2.0
1995–1996 7.4 7.5 5.2 3.0 32.1 2.3
1997–1998 6.4 6.5 4.7 2.7 28.3 2.2
1999–2000 6.0 6.4 5.2 3.5 27.5 2.4
2001–2002 5.6 6.3 5.4 4.7 33.1 2.6
2003 4.2 5.5 4.7 4.9 23.6 2.2
Percentage of AKC registered that produced registered litters, 1999–2003 (%)
Mean 19.6 16.1 18.7 — — 23.2
Evaluation rate, based on expected number of dogs with registered litters (%)
1999–2003 31.7 41.2 27.6 — — 14.1


--------------------------------------------------------------------------------




A total of 62,078 dogs whelped between 1993 and 2003 were rated as having excellent hip phenotypes by the OFA, 252,623 rated good, 65,741 rated fair, and 51,041 rated as mild, moderate, or severe CHD. There were statistically significant (Cochran–Armitage Z-test statistic <0, at P<.05) increases in the proportions of the 5 breeds of interest evaluated as excellent (Table 3a) and good (Table 3b) from 1993 to 2003, whereas proportions of fair (Table 3c) and mild/moderate/severe CHD (Table 3d) significantly decreased during this period. Labrador Retrievers had the highest proportions of excellent and good classifications (19.4% and 59.8%, respectively), Golden Retrievers had the highest proportion of fair classifications (25.6%), and Rottweilers had the highest proportion of mild/moderate/severe CHD classifications (16.9%).


Table 3a. Percent of Dogs with "Excellent" Classifications of Hip Joint Phenotype Performed by the Orthopedic Foundation for Animals (OFA), by Whelping Year


--------------------------------------------------------------------------------

Whelping Year Labrador Retriever
(n=102,960) Golden Retriever
(n=46,634) German Shepherd
(n=36,793) Rottweiler
(n=23,894) Bernese Mount
Dog (n=687 Other Breeds
(n=214,324)

--------------------------------------------------------------------------------

1993 18.4 4.5 3.5 10.6 10.6 14.9
1994 16.9 4.8 3.4 9.7 8.7 14.5
1995 18.2 4.6 3.7 9.5 12.8 15.3
1996 19.3 5.2 4.9 11.9 9.6 16.5
1997 21.5 5.8 5.8 14.8 18.8 18.6
1998 20.9 5.8 4.8 15.9 18.1 18.5
1999 20.7 5.8 4.8 12.3 16.8 17.6
2000 19.7 4.6 3.9 10.8 13.7 15.8
2001 19.4 4.6 4.4 11.7 16.1 15.2
2002 18.3 5.4 4.4 11.3 15.6 14.4
2003 19.9 5.3 5.1 12.5 14.1 14.7
Cochran–Armitage Z-test for trend (P) ?4.0 (<.0001) ?1.6 (.1090) ?3.3 (.0009) ?4.1 (<.0001) ?3.9 (.0001) .7 (.474


--------------------------------------------------------------------------------





Table 3b. Percent of Dogs with "Good" Classifications of Hip Joint Phenotype Performed by the Orthopedic Foundation for Animals (OFA), by Whelping Year


--------------------------------------------------------------------------------

Whelping Year Labrador Retriever
(n=102,960) Golden Retriever
(n=46,634) German Shepherd
(n=36,793) Rottweiler
(n=23,894) Bernese Mount
Dog (n=687 Other Breeds
(n=214,324)

--------------------------------------------------------------------------------

1993 59.5 52.6 54.4 55.7 58.1 60.5
1994 59.0 50.5 51.8 54.1 59.3 59.2
1995 58.4 48.6 51.8 54.6 54.6 58.2
1996 57.5 50.8 53.5 54.7 52.4 57.7
1997 57.5 53.2 55.9 54.1 57.0 58.2
1998 57.7 52.4 53.5 54.7 56.1 58.8
1999 59.2 53.8 56.7 56.8 56.0 60.1
2000 62.1 55.4 59.8 60.3 58.6 62.7
2001 61.8 52.3 60.5 59.1 61.4 62.6
2002 63.9 55.2 58.3 60.3 59.4 62.4
2003 62.8 55.6 56.4 61.0 61.1 62.8
Cochran–Armitage Z-test for trend (P) ?10.4 (<.0001) ?6.7 (<.0001) ?8.5 (<.0001) ?5.3 (<.0001) ?2.6 (.0094) ?13.2 (<.0001)


--------------------------------------------------------------------------------





Table 3c. Percent of Dogs with "Fair" Classifications of Hip Joint Phenotype Performed by the Orthopedic Foundation for Animals (OFA), by Whelping Year


--------------------------------------------------------------------------------

Whelping Year Labrador Retriever
(n=102,960) Golden Retriever
(n=46,634) German Shepherd
(n=36,793) Rottweiler
(n=23,894) Bernese Mount
Dog (n=687 Other Breeds
(n=214,324)

--------------------------------------------------------------------------------

1993 10.1 25.0 24.8 15.8 17.9 15.0
1994 12.2 27.0 27.2 18.2 16.4 16.6
1995 11.6 27.6 25.5 17.5 13.9 15.8
1996 10.6 26.6 24.1 15.9 22.2 15.2
1997 9.8 24.5 21.9 14.5 13.0 13.4
1998 10.0 25.0 23.6 14.8 13.6 12.9
1999 9.3 24.2 21.9 15.6 15.2 13.3
2000 9.0 25.6 21.3 13.8 15.8 12.8
2001 8.9 27.9 20.1 14.2 11.7 13.2
2002 8.0 23.8 20.0 13.7 12.8 13.7
2003 8.0 23.6 19.7 11.3 11.4 12.9
Cochran–Armitage Z-test for trend (P) 11.9 (<.0001) 2.8 (.0046) 9.8 (<.0001) 5.5 (<.0001) 4.5 (<.0001) 13.4 (<.0001)


--------------------------------------------------------------------------------





Table 3d. Percent of Dogs with "Mild/Moderate/Severe" Dysplastic Classifications of Hip Joint Phenotype Performed by the Orthopedic Foundation for Animals (OFA), by Whelping Year


--------------------------------------------------------------------------------

Whelping Year Labrador Retriever
(n=102,960) Golden Retriever
(n=46,634) German Shepherd
(n=36,793) Rottweiler
(n=23,894) Bernese Mount
Dog (n=687 Other Breeds
(n=214,324)

--------------------------------------------------------------------------------

1993 12.1 17.9 17.3 17.9 13.4 9.6
1994 11.9 17.8 17.6 18.0 15.6 9.7
1995 11.8 18.2 18.9 18.4 18.7 10.7
1996 12.5 17.5 17.4 17.5 15.8 10.5
1997 11.3 16.5 16.5 16.6 11.1 9.8
1998 11.4 16.9 18.1 14.5 12.1 9.8
1999 10.9 16.2 16.5 15.2 11.9 9.0
2000 9.2 14.3 15.0 15.1 11.9 8.7
2001 9.9 15.2 15.1 15.0 10.7 9.0
2002 9.7 15.6 17.2 14.7 12.2 9.4
2003 9.3 15.5 18.9 15.2 13.4 9.6
Cochran–Armitage Z-test for trend (P) 10.0 (<.0001) 6.6 (<.0001) 2.0 (.0489) 5.2 (<.0001) 3.1 (.0019) 5.1 (<.0001)


--------------------------------------------------------------------------------




Based on the odds ratios for the likelihood of a dog whelped between 1993 and 2003 receiving a given OFA rating (Fig 1), the odds for dogs in the current study receiving excellent classifications were significantly better (odds ratio >1.0) than dogs evaluated during the previous study period (1989–1992), and the odds for good classifications increased significantly for dogs whelped after 1998. The odds ratios for fair and mild/moderate/severe CHD were significantly lower for all dogs in the current study, and these odds steadily declined during the study period.


Fig 1. Odds ratios for dogs whelped between 1993 and 2003 receiving excellent, good, fair, or mild/moderate/severe hip dysplasia classifications by the Orthopedic Foundation for Animals (OFA), compared with dogs whelped from 1989 to 1992, adjusted for age and gender.

[Normal View ]

Mean age of all dogs at evaluation was 32.4 months (interquartile range 25–35, minimum 24, maximum 18. The mean ages of dogs receiving excellent, good, fair, and mild/moderate/severe CHD classifications were 32.7, 32.2, 32.6, and 33.2 months, respectively. The mean age declined slightly during the course of the study, from a mean age of 33 months in 1993 to 26.5 months in 2003.

Over 63% of hip classifications by the OFA were for females (Table 4), and the proportion of females to males increased over time from 62.6% in 1993 to a peak of 64.9% in 2002 (Cochran–Armitage Z=7.9, P<.0001). The only statistically significant differences by gender were for excellent and fair classifications, where males had slightly higher proportions than females (14.6% versus 14.3% excellent classifications, respectively, age-adjusted odds ratio for males=1.02, 95% CI=1.01–1.04; 15.2% versus 15.1% fair classifications, respectively, age-adjusted odds ratio for males=1.03, 95% CI=1.01–1.05), whereas females had statistically significantly higher proportions of mild/moderate/severe CHD classifications than males (12.0% versus 11.8%, respectively, age-adjusted odds ratios for males=0.95, 95% CI=0.93–0.97).


Table 4. Numbers of Dogs Evaluated by the Orthopedic Foundation for Animals (OFA), by Gender (n=431,483 Dogs)


--------------------------------------------------------------------------------

Breed Total Females Males

--------------------------------------------------------------------------------

Labrador Retriever 102,960 68,904 34,056
Golden Retriever 46,634 31,791 14,843
German Shepherd 36,793 21,802 14,991
Rottweiler 23,894 14,206 9688
Bernese Mount Dog 6878 4239 2639
Other Breeds 214,324 133,456 80,868


--------------------------------------------------------------------------------






DISCUSSION


In our current study, the percentages of classifications of excellent and good hip phenotype improved in dogs whelped from 1993 to 2003 (Fig 1, Tables 3a–d). This agrees with findings from several studies that have indicated that selective breeding using this information might improve hip conformation in several breeds of dogs,3,15 although other studies have failed to support this.8,9 Whereas the analysis was limited to dogs whelped before 2004 to ensure that the majority of eligible animals would have been evaluated by the OFA, it is possible that these later years were not sampled as completely as dogs from earlier whelping year cohorts. As noted above, the mean age of dogs at evaluation remained relatively steady until 2000, when it began to decline noticeably. This may indicate that dogs whose owners choose to evaluate them at later ages are not being captured in the last years of the study, which may bias results for these years and censor results from older dogs not yet evaluated, and reduced reporting may be a factor in the slower improvement in desirable hip classifications seen in the 2002–2003 cohort. There may also be issues associated with the subjectivity of radiographer assessments of hip conformation. As the OFA uses a pool of expert radiologists with considerable experience in hip evaluation, and uses evaluations from 3 radiologists randomly selected from this pool for each radiograph submitted, this will result in fairly standardized radiograph interpretations that should maintain continuity over time.

This study identified statistically significant, but very small associations, between gender and OFA classifications, with males more likely to achieve excellent and fair classifications, and less likely to achieve mild/moderate/severe CHD classifications than females. Whereas some studies have reported this,17 others have found no association between gender and CHD.11,18,19 It is important to note that whereas these associations are statistically significant, the miniscule amount of influence (odds ratios very close to 1.0) gender has on OFA classification might be clinically insignificant.

One interesting finding was the changing breed demographics of the canine population, whelped from 1993 to 2003, which had hip phenotype evaluation by the OFA. Declines were seen in overall numbers of submissions all breeds of interest, with the exception of Bernese Mountain Dogs (Table 1). The largest declines in numbers were seen in Rottweilers (loss of 81% from 1993 to 2003) and German Shepherds (loss of 53% from peak in 1992 to 2003). Most of these findings were reflected in AKC registrations statistics (AKC, 1981–2005), in that Rottweilers went from 44th place to break into the top 5 breeds in 1990, and have declined to 15th place in 2003. In addition to breed popularity, breed organizations may push for radiographic hip evaluations to improve breed standards, which could result in increases in submissions to the OFA over time.

Studies have reported mixed results in effectiveness of selective breeding based on phenotypic hip classifications in reducing CHD, particularly between different breeds of dog.8–10 These studies suggested that, as CHD is a multifactorial condition that is not only genetically based, environmental and other extrinsic factors (e.g. diet,20 activity levels) might play a critical role in the development of CHD. They also indicated that adherence to breeding standards alone may not result in breeding selection criteria levels necessary to effect change, given the polygenetic nature of CHD.9 One study found significant associations between hip classifications of progeny with dam and sire hip classifications, with excellent/excellent dam/sire breedings resulting in progeny with the most desirable hip classifications,18 which supports the use of this approach in the development of breeding programs. As the rate for OFA evaluations of AKC-registered dogs continues to be low, as previously reported,15 this indicates that many breeders base their decisions on which dogs to breed solely on the phenotypic traits of the individual dogs being bred, without information on phenotypic characteristics of other related dogs or previous offspring. Dog breeders are usually involved in breeding dogs for only 5 years, which means that there is continuous replacement of experienced breeders with new breeders that may be unaware of CHD and the importance of evaluation programs. Improving evaluation rates and providing breeders with more information, such as results of progeny testing, can make selective breeding for desirable phenotypes more efficient.

As indicated by other researchers,4,10,11 there are biases associated with the use of data from the OFA or other registries with voluntary submissions to determine changes in hip phenotypes to the canine population at large. As the decision to submit radiographs to the OFA is voluntary and based on the owner, there is self-selection bias, which may preferentially include submissions from dogs with apparently normal hip phenotypes, and miss dogs with CHD, thereby underestimating the actual proportions of CHD. While the evaluation rates in this study were less than desired for extrapolating results to the general population of all dogs being bred, results of this study on the subset of the population of breeding dogs can be used to indicate how aggressively managed breeding may be affecting rates of desirable hip conformations, and additional work is needed to determine how these results can be applied to the general canine population at large.

It was not the goal of the original study15 and this study to estimate the rates of CHD, but to determine if changes in proportions of excellent/good (where underreporting would be less likely) had occurred over time, and inferences about the prevalence of CHD from these findings would be strongly discouraged. If owners were intentionally withholding obviously dysplastic submissions to the OFA, this practice would not be expected to change over the period from 1993 to 2003, which would make the levels of underreporting consistent during this time. It is also possible that, as more owners and veterinarians are aware of the impact of OFA ratings on the potential value of their dog for breeding in the future as the beginning of the rating system, that self-selection bias may be increasing over time. However, the age profiles of dogs with submissions in this study were relatively consistent over time, with a slight decline over time may be because of the fact that dogs whelped in later years have not aged to the extent that dogs whelped earlier have, and have yet to be evaluated.

Statistics from the AKC for the study period reported that the major breeds registered <47% of the puppies whelped from registered litters as adult dogs (Table 2). Further, for the same time period, 15.5% and 23% of registered male and registered female (respectively) Labrador Retrievers, 12.1% and 20.2% of registered male and female Golden Retrievers (respectively), and 14.9% and 22.5% of registered male and female German Shepherd Dogs (respectively), had one or more registered litters (AKC, personal communication, 2007). Owners of non-AKC registered purebred dogs, and owners of AKC-registered dogs not actively participating in breeding, have few incentives for acquiring OFA certification and are less likely to have their dogs screened for hip certification.

In conclusion, we found improvements in desirable OFA hip phenotype classifications of dogs from 1993 to 2003 in comparison with dogs from the earlier study period of 1989–1992. Continuing to collect and assess data from OFA into the future should eliminate age-related data censoring in the later study years, which will answer the question as to whether or not the results of the last years of the study (2000–2003) are true or an artifact of the database: information from 2000–2003 showed less improvement in hip classifications, which may be because of data censoring when classifications for dogs whose owners who chose to have their dogs evaluated at older ages have not yet entered the dataset. This information and continuing examination will give future investigators the ability to determine how effective the use of OFA classifications are in selective breeding programs. Communicating and educating veterinary practitioners and the dog breeding public on the value of classification is a necessity to improve evaluation rates, and consequently allow more accurate assessments of the utility of different approaches to dog breeding.




ACKNOWLEDGMENTS


The authors would like to thank the OFA Inc. and the AKC Inc. for providing data for this study.



REFERENCES


1. Smith GK, Popovitch CA, Gregor TP, et al: Evaluation of risk factors for degenerative joint disease associated with hip dysplasia in dogs. J Am Vet Med Assoc 206:642–647, 1995
2. Lust G: An overview of the pathogenesis of canine hip dysplasia. J Am Vet Med Assoc 210:1443–1445, 1997
3. Swenson L, Audell L, Hedhammar A: Prevalence and inheritance of selection for hip dysplasia in seven breeds of dogs in Sweden and benefit: cost analysis of a screening and control program. J Am Vet Med Assoc 210:207–214, 1997
4. Corley EA, Keller GG, Lattimer JC, et al: Reliability of early radiographic evaluations for canine hip dysplasia obtained from the standard ventrodorsal radiographic projection. J Am Vet Med Assoc 211:1142–1146, 1997
5. Comhaire FA, Snaps F: Comparison of two canine registry databases on the prevalence of hip dysplasia by breed and the relationship of dysplasia with body weight and height. Am J Vet Res 69:330–333, 2008
6. Orthopedic Foundation for Animals. Hip Grades. http://www.offa.org/hipgrade.html. Accessed May 27, 2008.
7. Dziuk E: The OFA—40 years of dedication to the advancement of canine health. Available at: http://www.offa.org/40yeararticle.pdf Accessed April 11, 2007.
8. Willis MB: A review of the progress in canine hip dysplasia control in Britain. J Am Vet Med Assoc 210:1480–1482, 1997
9. Leppanen M, Saloniemi H: Controlling canine hip dysplasia in Finland. Prev Vet Med 42:121–131, 1999
10. Genevois JP, Remy D, Viguier E, et al: Prevalence of hip dysplasia according to official radiographic screening, among 31 breeds of dogs in France. Vet Comp Orthop Traumatol 21:21–24, 2008
11. Paster ER, LaFond E, Piery DN, et al: Estimates of prevalence of hip dysplasia in Golden Retrievers and Rottweilers and the influence of bias on published prevalence figures. J Am Vet Med Assoc 226:387–392, 2005
12. Rettenmaier JL, Keller GG, Lattimer JC, et al: Prevalence of canine hip dysplasia in a veterinary teaching hospital population. Vet Radiol Ultrasound 43:313–318, 2002
13. Puerto DA, Smith GK, Gregor TP, et al: Relationships between results of the Ortolani method of hip joint palpation and distraction index, Norberg angle, and hip score in dogs. J Am Vet Med Assoc 214:497–501, 1999
14. Kapatkin AS, Gregor TP, Hearon K, et al: Comparison of two radiographic techniques for evaluation in hip joint laxity in 10 breeds of dogs. J Am Vet Med Assoc 224:542–546, 2004
15. Kaneene JB, Mostosky UV, Padgett GA: Retrospective cohort study of changes in hip joint phenotype of dogs in the United States. J Am Vet Med Assoc 211:1542–1544, 1997
16. Orthopedic Foundation for Animals. Hip Dysplasia Radiograph Procedures. Available at http://www.offa.org/hipproc.html#accuracy. Accessed May 27, 2008.
17. Janutta V, Hamann H, Distl O: Complex segregation analysis of canine hip dysplasia in German Shepherd Dogs. J Hered 97:13–20, 2006
18. Reed AL, Keller GG, Vogt DW, et al: Effect of dam and sire qualitative hip conformation scores on progeny hip conformation. J Am Vet Med Assoc 217:675–680, 2000
19. Pederson NC, Pool R: Canine joint disease. Vet Clin N Am 8:465–494, 1978
20. Smith GK, Paster ER, Powers MY, et al: Lifelong diet restriction and radiographic evidence of osteoarthritis of the hip joint in dogs. J Am Vet Med Assoc 229:690–693, 2006"

[Jac]

Thanks, Todd, that is a lot of info to digest...and, frankly, my non-scientific brain struggled with some of it...but I am really pleased to see that they are studying this and that it appears that because of testing there is a significant improvement in scores...or, perhaps, that is just because only the breeders who test actually care enough to do something to protect the breed, once they know of the results.
Too bad that there are so many breeders who don't want to know, so they bury their heads in the sand...hoping that by not knowinw, they aren't responsible...still, it is commendable that the numbers are improving.