Publishing Authors : Minna LeppaÈnen, Hannu Saloniemi
Date Published : 05/05/99
The aim of the study was to evaluate the Finnish Kennel Club’s hip-dysplasia screening and
control programs. As a retrospective study, records of hip-dysplasia screening of 69,349 dogs in 22
breeds that were born in 1988±1995 were analyzed and compared to data from prior to 1988. In
most breeds, no significant changes in dysplasia prevalence could be found. In English cocker
spaniels, golden and Labrador retrievers and Rottweilers a significant decrease ± but in boxers,
Dobermans, German Shepherd dogs and rough collies a significant increase ± in prevalence was
detected. In flat-coated retrievers overall prevalence increased ± but the prevalence of severe hipdysplasia
decreased significantly during the study period.
The present control program has not resulted in fast progress. Selecting against hip-dysplasia
cannot be expected to be very effective, when based only on mass selection on phenotypic
observations. Predicted breeding values based on progeny testing would probably give better
results. Also, breeders’ compliance and commitment to programs is not always high and other
selection criteria in breeding are thought to be more important.
Modern society has high demands for animal welfare and consumer issues, and breeders and
kennel societies should pay more attention to health issues in breeding pedigreed dogs. # 1999
Elsevier Science B.V. All rights reserved.
Keywords: Dog; Hip-dysplasia; Control program; Effectiveness
Canine hip-dysplasia was reported over 60 years ago (Schnelle, 1935). Research on hipdysplasia
has been ongoing ever since and there is now wide consensus that hip-dysplasia
Preventive Veterinary Medicine 42 (1999) 121±131
* Corresponding author. Tel.: +358-9-70851; fax: +358-9-27476450
E-mail address: email@example.com (M. LeppaÈnen)
0167-5877/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved
is a quantitatively inherited trait. Variation in heritability estimates between breeds and in
different studies has been discovered; reported heritabilities vary from 0.17 to 0.6 (Fisher,
1979; Lingaas and Klemetsdaal, 1990; Swenson et al., 1997; Tomlinson and Mclaughlin,
1996). Environmental factors and their role in the development and severity of hipdysplasia
have also been widely researched (Aichinger, 1997; Brass, 1989; Schwalder et
al., 1996; Swenson et al., 1997; Tomlinson and Mclaughlin, 1996).
There are different opinions of the clinical importance of hip-dysplasia and
possibilities of predicting the development of clinical signs from radiographs. Joint
laxity is important in predicting hip-dysplasia and degenerative joint disease.
Susceptibility for degenerative changes has important between-breed variation (Lust,
1997; Popovitch et al., 1995; Smith, 1997). Hip-dysplasia is commonly considered to
cause important disability and to decrease ability to work. However, a recent study of
military working dogs showed no significant decrease in working ability or time in dogs
suffering from hip-dysplasia (Banfield et al., 1996).
Many countries and kennel organizations have established control programs for hipdysplasia.
All the present official programs are based on radiographic evaluation of the
phenotype (Brass, 1989; Corley, 1992; FluÈckiger, 1994; Anon., 1994). Schnelle (1954)
proposed the first classifying system in 1954; currently, there are many different systems
in use. In Finland the Federation Cynologue Internationale’s (FCI) five-scaled system is
used (Appendix A).
The first official hip-dysplasia-control program in Finland was started for German
Shepherds in 1963. Then, negative hip-dysplasia radiographs were required before a dog
could gain a champion’s title (Paatsama, 1978, 1979). The first of the currently used
control programs were established in 1984, when breeding animals of retriever breeds
were required to undergo official hip screening as a prerequisite for litter registration.
Currently, 71 breeds have an official control program for hip-dysplasia under the Finnish
Kennel Club and about 7500 radiographs are screened annually in this scheme (Liman,
A., personal communication). For some breeds in the program, only screening for hipdysplasia
is mandatory ± but for the others, there are threshold values for hip-dysplasia
for breeding stock.
Control programs and their effectiveness in reducing hip-dysplasia have shown
variable results. FluÈckiger et al. (1995) found in about 3700 dogs Swiss an overall
dysplasia prevalence of 42%; however, significant between-breed variation was observed.
In some breeds, prevalence decreased clearly ± but in others, it increased during the study
period. In most breeds, however, a decrease in severity of dysplasia was noticed. In that
study also, some correlation was observed between breeding restrictions and the decrease
of dysplasia prevalence: the milder the restrictions, the slower the decrease in prevalence
(FluÈckiger et al., 1995). Also, English and American studies have found remarkable
variation in prevalence and changes in different breeds (Corley, 1992; Willis, 1997). The
best progress was found in Sweden, where disease prevalence decreased during the study
period in all breeds investigated (Swenson et al., 1997).
Hip-dysplasia control programs have been widely discussed since they have been in
use. The slow progress in reducing the disease has been especially criticized; this has
been stated to show futility of the programs and to justify giving them up. On the other
hand, slow progress in decreasing dysplasia prevalence has also been considered a
122 M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131
justification for more-advanced measures and stronger restrictions on breeding animals.
There has also been a lot of debate about the costs of the program to breeders and dogowners.
Hip-dysplasia screening and restrictive programs are also stated to cause a
decrease in genetic variance (because of single-trait selection), which can lead to other
more-severe problems (Bouw, 1982).
The objectives of this study were to evaluate the effectiveness of the Finnish hipdysplasia-control
program and changes in disease prevalence. The final aim of this study
was to discuss the reasons for the effectiveness or futility of the program.
2. Materials and methods
2.1. The data
This study included hip-dysplasia records for 64 349 dogs in 22 breeds (Table 1).
Nineteen of these breeds have had an official control program since 1993 and the number
of registered puppies in each breed has exceeded 700 in the years 1993±1997. Rottweilers
were included in the study; although they only joined the program in 1994, the breed club
has campaigned strongly for voluntary screening for a long time before this. Saint
Bernards were included as a model of a breed that is commonly thought to have a bad
hip-dysplasia situation and for which the breeders commonly oppose the breeding
restrictions. Although there were not >700 registered short-haired Saint Bernards during
the study period, they were included, because in Finland the pedigree is open between
both coat-types, and thus breeders can freely mate short- and long-haired dogs, although
matings typically are between similar coat types. Also, in the same litter there can be
dogs of both coat variations that are registered accordingly. Nine breeds in the study have
established breeding restrictions, that only dogs whose hips are screened as better than the
breed’s threshold value are accepted for breeding and have their offspring registered in
the Finnish Kennel Club. Because the German pointer became a restricted breed only on
January 1, 1996, it is considered a non-restricted breed in the study.
All official hip-dysplasia radiographs are judged by panelists that have the
authorization of the Finnish Kennel Club. There have been five panelists altogether.
Until 1988, one panelist screened all the radiographs. In 1988±1995 three new panelists
were appointed; since 1996, the fifth panelist has screened all radiographs. Each panelist
had screened his share of radiographs individually.
This study consisted of the records of dogs that were born in 1988±1995 and were
screened before January 1, 1997. The material was compared to the Kennel Club’s
dysplasia data that were recorded before 1988. To minimize the effects of annual random
variation, breed-specific dysplasia prevalence was calculated from all dogs that were born
in 1988±1995 and this was compared to the breed’s prevalence in dogs that were screened
before 1988. However, this material was available only divided according to the breed
and screening year, so a direct comparison between years of birth could not be made.
Also, the classification varied slightly (Appendix A). In some breeds, the number of
screened dogs before 1988 was also very small (which in part makes comparisons
difficult). The data do not allow any distinction according to the sex or dogs that were reexamined;
neither was the age at examination included.
M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131 123
Description of programs and of records available
Breed Year Threshold value Total registrations % of registrations Number of screened
Joined the program Restrictions begun
from 1988 to 1995 screened before 1988
Bearded collie 1989 1989 C 1550 22 545
Boxer 1988 1994 C 2798 40 1844
Bullmastiff 1989 1989 D 1525 34 243
Dalmatian 1991 NO NO 1694 24 55
Doberman 1986 1993 C 3874 31 1025
English cocker spaniel 1990 NO NO 5943 18 307
English springer spaniel 1990 NO NO 2833 22 364
Flat-coated retriever 1984 1991 C 1418 64 426
German Shepherd dog 1986 1989 C 23 385 36 14 189
Golden retriever 1984 1991 C 16 414 32 3847
Irish setter 1986 NO NO 1462 29 99
Labrador retriever 1984 1991 C 13 327 36 5682
Longhaired Saint Bernard 1992 NO NO 3307 23 674
Nova Scotia duck tolling retriever 1984 1991 C 979 45 11
Rottweiler 1994 NO NO 6672 17 1844
Rough collie 1986 NO NO 10 708 26 1457
Samoyed 1988 NO NO 1979 36 980
Shorthaired German pointer 1989 1996 C 1725 26 153
Shorthaired Saint Bernhard 1992 NO NO 253 42 3
Smooth collie 1986 NO NO 857 32 9
Welsh springer spaniel 1990 NO NO 1130 31 233
Wirehaired German pointer 1989 1996 C 1977 22 227
Total 105 810 34 217
Threshold value the most-severe grade of dysplasia allowed for breeding.
NO no restrictions or threshold value, only obligatory screening before breeding.
124 M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131
2.2. Statistical analysis
The ÿc2 test (Statistix Version 4.1TM, Analytical Software, Tallahassee, FL, USA) was
used in analyzing the heterogenity between dogs screened before 1988 and dogs born in
1988±1995 and screened before 1997. Fisher’s 2-tailed exact test with alpha 0.05 was
used for comparisons between the breeds with breeding restrictions and no restrictions, as
well as between the breeds that established breeding restrictions in 1991 or earlier, or the
breeds that had restrictions since 1993.
3.1. Disease prevalence
The breed-specific proportions of registered dogs that were screened varied from 18%
(English cockerspaniel) to 64% (flat-coated retriever). Breeds varied greatly in dysplasia
prevalence; it varied from 2% (smooth collie) to 80% (long-haired Saint Bernard)
(Table 2). In every breed, annual variation in prevalence was also noticed. Figs. 1±3 show
dysplasia prevalences according to the year of the birth in those breeds with significant
changes. The figures for the year 1987 present the prevalences in all dogs of a breed that
Fig. 1. Overall hip-dysplasia prevalence (straight line) and prevalence of severe dysplasia (broken line)
according to the year of birth in English cocker spaniels, Dobermans and rough collies. Year 1987 represents the
prevalence of all the animals in the breed screened before January 1, 1988.
M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131 125
were screened and that were born before 1988. No significant difference was found
between the breeds with or without breeding restrictions (p0.20), or between those
breeds that started restrictions earlier and those that established restrictions recently
In Finland, approximately 95% of pedigreed dogs are registered with the Finnish
Kennel Club. Within-breed proportions of dysplasia-screened dogs are quite stable every
year, and in most cases the radiographs are submitted to the screening despite a bad
preliminary evaluation. However, the presented frequencies of screened animals are only
Dysplasia prevalences (%) within the study breeds
Breed Dysplasia prevalence % Prevalence % of severe dysplasiaa
Before 1988 In 1988±1995 c2 (df1) Before 1988 1988±1995 c2 (df1)
Bearded collie 22 24 0.69 11 12 0.39
Boxer 28 40 0.00 18 16 0.20
Bullmastiff 45 53 0.04 26 29 0.72
Dalmatian 16 20 0.52 5 4 0.59
Doberman 10 16 0.00 4 4 0.66
English cocker spaniel 23 20 0.21 16 8 0.00
English springer spaniel 28 29 0.58 14 14 0.95
Flat-coated retriever 16 20 0.00 7 4 0.03
German Shepherd dog 33 46 0.00 18 21 0.00
Golden retriever 41 36 0.00 21 18 0.00
Irish setter 33 40 0.02 14 19 0.49
Labrador retriever 32 29 0.00 22 17 0.00
Longhaired Saint Bernard 79 82 0.24 70 67 0.15
Nova Scotia duck tolling
0 25 0.06 0 14 0.13
Rottweiler 45 38 0.00 31 20 0.00
Rough collie 13 16 0.00 9 10 0.00
Samoyed 36 34 0.11 23 19 0.06
Shorthaired German pointer 16 7 0.90 1 1 0.63
Shorthaired Saint Bernard 67 59 0.80 33 42 0.78
Smooth collie 11 2 0.06 0 0
Welsh springer spaniel 41 35 0.19 20 17 0.29
Wirehaired German pointer 16 16 0.21 5 5 0.97
a Severe canine hip dysplasia grades D or E or >3.
126 M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131
estimations, because all dogs are not registered in the same year as they are born. At least
all breeding animals ± as well as a reasonable number of pet animals ± in control-program
breeds are screened. Thus the hip-dysplasia-control program gives a quite good overall
picture of a breed’s hip-dysplasia situation. Also, because almost all pedigree dogs are
registered, registration restrictions for offspring of dysplastic parents are an effective
measure that prevents breeding with affected animals. Slight bias might be caused by
dogs that have been screened more than once (because then one dog has multiple results
in the data).
In nine breeds included in this study ± (boxer, Doberman, English cocker spaniel, flatcoated,
golden and Labrador retriever, German Shepherd dog, rough collie and
Rottweiler) significant changes in hip-dysplasia prevalence were detected. However in
four breeds ± boxer, Doberman, German Shepherd dog and rough collie ± the prevalence
increased. Our study could not support the previous studies which found significant
overall decrease in disease prevalence (Brass, 1989; Swenson et al., 1997); instead, our
findings support those of the studies that found distinct between-breed variation
(FluÈckiger et al., 1995; Willis, 1997). The effectiveness of strong official breeding
Fig. 2. Overall hip-dysplasia prevalence (straight line) and prevalence of severe dysplasia (broken line)
according to the year of birth in boxers, Rottweilers and German Shepherds. Year 1987 represents the prevalence
of all the animals in the breed screened before January 1, 1988.
M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131 127
restrictions is also questionable: German shepherds have had restriction the longest time
and showed significant increase in dysplasia prevalence ± but Rottweilers joined the
program only recently and prevalence had decreased already before joining the scheme.
Also, we could not show significant differences in effectiveness between the breeds with
or without breeding restrictions. In another study (LeppaÈnen et al., 1999), however, we
reported that the number of German shepherd parent animals graded worse than excellent
(B or C) had actually increased during the study period ± which gives reason to doubt that
the selection has been strong enough. Our current data set does not give information of
parents’ hip-dysplasia statuses, but at least the parents of all the dogs born after the breed
has joined the program had to have been screened. Also, for most breeds involved, it was
already a common procedure to screen at least breeding stock before the official program
was begun ± so it is justified to suppose that most animals in this study have screened
parents. This conclusion is also supported by findings in the Swedish data (Swenson et
There are several possible reasons for detected minimal improvement in many
breeds. It can be assumed that before joining the program, radiographs with
dysplastic changes detected by veterinarians who took the radiographs were not
forwarded for the official screening ± and yet the parents still could have been used for
Fig. 3. Overall hip-dysplasia prevalence (straight line) and prevalence of severe dysplasia (broken line)
according to the year of birth in flat-coated, golden and Labrador retrievers. Year 1987 represents the prevalence
of all the animals in the breed screened before January 1, 1988.
128 M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131
breeding ± which may have biased earlier prevalences to look better than the real
situation. Each breed club’s own code of ethics as well as the authority a breed club had
amongst the breeders will have had an influence on breeders and their actions. Also,
various breeds have other selection considerations concerning health, conformation or
performance abilities (which might dilute the effort to select against hip-dysplasia).
Finland removed previous strict quarantine regulations for imported dogs in 1988.
In many breeds (especially in those of German origin), this had led to a marked
increase in the numbers of imported breeding dogs whose hip-dysplasia background is
Willis (1997) found a smaller observed decrease in dysplasia prevalence than was
expected, and suggested that heritability might be lower than estimated.
Overestimation of heritability could have been caused by assortative mating or
strong in-breeding. The heritabilities in that study were estimated with parental halfsib
analysis. This type of error can be avoided with more-advanced methods for
heritability estimation such as the restricted maximum-likelihood (REML) method
(Patterson and Thomson, 1971; Groeneveld, 1997). Slow progress can also be
explained by the fact that when selecting against a quantitative trait with moderate
heritability ± like hip-dysplasia ± mass selection does not allow fast progress (Hutt,
1967). Advanced use of progeny results and individual breeding indices would give better
reliability of selection criteria (Leighton, 1997; Lingaas and Klemetsdaal, 1990; Willis,
Because the present screening system is based on subjective classification of
radiographic findings, panelist-dependent variation also could be possible. Studies have
found significant within- and between- examiner variation in classifications (Smith et al.,
1997; Stur et al., 1996). When dealing with large populations, good correlation of
dysplasia results and classification variables between the examiners was, however, found
(Stur et al., 1996).
Lately, there has been a steady increase in discussion and demands for advanced
animal welfare (Anon., 1995, 1996). Discussions of pets’ role in modern society and their
meaning to human health has increased (Allen, 1997; Beaver, 1997; Glickman, 1991).
Pets are often seen more as a family member than an also discussions animal. Thus, dog
breeders and kennel societies have created health programs although direct economic
benefits could not be shown. Also, Finnish consumer legislation treats animals as any
other article and thus a producer ± that is a breeder ± is held responsible to compensate
pet owners for possible production errors (such as inherited diseases). In spite of this,
demonstrable progress in reducing the prevalence of canine hip-dysplasia is scarce in
This study was financially supported by the Research Fund of the Finnish Club. We are
grateful for the personnel of the breeding department of the Finnish Kennel Club for all
their help in data management. Mr Jarmo Juga, Ph.D., gave us advice and comments
during the preparations of this manuscript.
M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131 129
Classification of hip-dysplasia
Present Past Interpretation
A Normal No signs of dysplasia
B Borderline Healthy, with sligh changes in conformation
C 1 Mild dysplasia
D 2 Moderate dysplasia
E 3±4 Severe dysplasia
Aichinger, O., 1997. Rassenrein aber krank ± Hilfe bei HuÈftgelenkdysplasie. TieraÈrztliche-Umschau 52(3), 153±
Allen, D.A., 1997. Effects of dogs on human health. J. Am. Vet. Med. Assoc. 210(8), 1136±1139.
Anon., 1994. Joint WSAVA/ Kennel Clubs Meeting to discuss the control of hip-dysplasia in pedigree dogs. June
17, 1994, London.
Anon., 1995. European convention for the protection of pet animals. Resolution on breeding in pet animals.
Council of Europe.
Anon., 1996. Animal welfare act 247/96. Finnish law.
Banfield, C.M., Bartels, J.E., Hudson, J.A., Wright, J.C., Montgomery, R.D., Hathcock, J.T., 1996. A
retrospective study of canine hip-dysplasia in 116 military working dogs. Part II: Clinical signs and
performance data. J. Am. Anim. Hosp. Assoc. 32(5), 423±430.
Beaver, B., 1997. Human-canine interactions: a summary of perspectives. J. Am. Vet. Med. Assoc. 210(8),
Bouw, J., 1982. Hip-dysplasia and dog breeding. Vet. Quarterly 4(4), 173±181.
Brass, W., 1989. Hip-dysplasia in dogs. J. Small Anim. Pract. 30, 166±170.
Corley, E.A., 1992. Role of the orthopedic foundation for animals in the control of canine hip-dysplasia. Vet.
Clin. N. Am: Small Anim. Pract. 22(3), 379±593.
Fisher, T.M., 1979. The inheritance of canine hip-dysplasia. Mod. Vet. Prac. 60(11), 897±900.
FluÈckiger, M., 1994. The standardized analysis of radiographs for hip-dysplasia in dogs. Objectifying a
subjective process. Eur. J. Comp. Anim. Pract. V(2), 39±44.
FluÈckiger, M., Lang, J., Binder, H., Busato, A., Boos, J., 1995. Die BekaÈmpfung der HuÈftgelenksdysplasie in der
Schweiz: Ein RuÈckblick auf die vergangenen 24 Jahre. Schw. Arch. fuÈr Tierheilk. 137(6), 243±250.
Glickman, L.T., 1991. Implications of the human/animal bond for human health and veterinary practice. J. Am.
Vet. Med. Assoc. 201(6), 848±851.
Groeneveld, E., 1997. REML VCE ± a multivariate multimodel restricted maximum likelihood (co)variance
component estimation package. Version User’s guide. Inst. Of Anim Husbandry and Anim. Ethology, Feder.
Res. Center of Agric., Germany (Mimeo).
Hutt, F.B., 1967. Genetic selection to reduce the incidence of hip-dysplasia in dogs. J. Am. Vet. Med. Assoc.
Leighton, E.A., 1997. Genetics of canine hip-dysplasia. J. Am. Vet. Med. Assoc. 210(10), 1474±1479.
LeppaÈnen, M., MaÈki, K., Juga, J., Saloniemi, H., 1999. The factors affecting hip-dysplasia in German Shepherds,
prediction of breeding values and genetic change in breeding programmes. In press.
Lingaas, F., Klemetsdaal, G., 1990. Breeding values and genetic trend for hip-dysplasia in the Norwegian
Golden retriever population. J. Anim. Breed. Genet. 107, 437±443.
Lust, G., 1997. An overview of the pathogenesis of canine hip-dysplasia. J. Am. Vet. Med. Assoc. 210(10),
130 M. LeppaÈnen, H. Saloniemi / Preventive Veterinary Medicine 42 (1999) 121±131
Paatsama, S., 1978. Koiran lonkkanivelen kasvuhaÈirioÈn vastustaminen (Breeding against canine hip-dysplasia).
Koiramme no. 6±7, 1±4.
Paatsama, S., 1979. Hip-dysplasia in dogs: a controlled restrictive breeding programme in Finland. Twenty First
World Veterinary Congress, Summaries, vol. 4. Sect. IX, pp. 42±43.
Patterson, H.D., Thomson, R., 1971. Recovery of interblock information when block sizes are unequal.
Biometrika 58, 427.
Popovitch, C.A., Smith, G.K., Gregor, T.P., Shofer, F.S., 1995. Comparison of susceptibility for hip dysplasia
between Rottweilers and German Shephard Dogs. J. Am. Vet. Med. Assoc. 206(5), 640±648.
Schnelle, G.B., 1935. Some new diseases in dog. American Kennel Gazette 52, 25±26.
Schnelle, G.B., 1954. Congenital dysplasia of the hip (canine) and sequalea. AVMA. Proceedings of the Annual
meeting. Seattle, WA, pp. 253±258.
Schwalder, P., Spreng, D., Dietschi, E., Dolf, G., Gaillard, C., 1996. Die HuÈftgelenkdysplasie im Umfeld von
sekundaren Einflussen und ektopischen Ursachen. Kleintierpraxis. 41(9), 625±638.
Smith, G., 1997. Advances in diagnosing canine hip-dysplasia. J. Am. Vet. Med. Assoc. 210(10), 1451±1457.
Smith, G.K., LaFond, E., Gregor, T., Lawler, D., Nie, R., 1997. Within- and between-examiner repeatability of
distraction indices of the hip joints in dogs. Am. J. Vet. Res. 58(10), 1076±1077.
Stur, I., Koppel, E., Schroder, K., 1996. Populationsgenetische Aspekte der HuÈftgelenkdysplasie (HD) ±
diagnostik bei Hund ± Bewertung unter Berucksichtigung differierender HD-Befunde. Wien. TieraÈrztl.
Mschr. 83, 91±97.
Swenson, L., Audell, L., Hedhammar, AÊ ., 1997. Prevalence, inheritance and selection for hip-dysplasia in seven
breeds of dogs in Sweden and a cost/benefit analysis of a screening and control program. J. Am. Vet. Med.
Assoc. 210(2), 207±214.
Tomlinson, J., Mclaughlin, R., 1996. Canine hip-dysplasia: developmental factors, clinical signs, and initial
examination steps. Vet. Med. 91(1), 26±33.
Willis, M.B., 1997. A review of the progress in canine hip-dysplasia control in Britain. J. Am. Vet. Med. Assoc.